Bs Murthy Thermodynamics

INDEX
S. No Topic Page No.

1 Basic definitions 1
2 Free energy, Stability, equilibrium in a unary system 9
Effect of Pressure on equilibrium transformations: Clausius Clapeyron
3 equation, phase diagram for unary system 32
4 Free energy of solutions, free energy-composition diagrams 57
5 Solution models, chemical potential 82
6 Phase rule, free energy-composition diagrams and phase diagrams 107
7 Evolution of phase diagrams 130
8 Evolution of phase diagrams, miscibility gap 156
9 To concept, partition less solidification 178
10 To concept, partition less solidification (cond...) 203
11 Eutectic solidification, glass formation 226
12 Kauzmann paradox, order of a transformation, glass forming ability 251
13 Eutectic solidification, coupled growth, heterogeneous nucleation 280
14 Peritectic solidification, metastable phase diagrams 307
Errors in drawing phase diagrams, Fe-C vs. Fe-Fe3C phase
15 diagram 332
16 Free energy of undercooled liquid, shape of nucleus 358
17 Solid state phase transformations - Precipitation 380
18 Precipitation 411
19 Precipitation – quasicrystals 437
20 Precipitate coarsening, stability of a phase, spinodal decomposition 465
21 Spinodal decomposition 491
22 Eutectioid reaction 522
23 Eutectioid reaction (cond...) 549
24 Bainitic transformation 577
25 Kinetics of eutectoid transformations 605
26 Martensitic Transformation 630
27 Martensitic transformation, order-disorder transformation 655
28 Miscibility gap in phase diagrams 676
29 Phase diagram calculations 694
30 Thermodynamics of heterogeneous systems 716
31 Thermodynamics of heterogeneous systems (cond...) 740
Advanced Metallurgical Thermodynamics
Prof. Dr. B. S. Murty
Department of Metallurgical and Materials Engineering
Indian Institute of Technology, Madras
Lecture No. # 01
Basic Definition
Today, I want to just know how much we know about free energy. I want to see how
many of you have heard this term, all of you must have heard. Let see what is your
understanding of free energy?
Student: The energy of the system which is available to perform some work.
(Refer Slide Time: 00:36)
Some work, so that is the reason - why we call it as a free energy. And Gibbs has defined
it as you all are aware as G = H - TS. This is something every one of you should know
and must have gone through it; where H is the enthalpy and T is the temperature and S is
the entropy. And this also coupled with the other equations, this comes actually from the
second law of thermodynamics all of you are aware of it. And because the entropy,
concept of entropy is actually comes from the second law.
1
First law basically talks about what is, what is first law?
Student: Conservation of energy
Conservation of energy is the first law, for example, H = E + PV - this is what we call it
as a first law. G = H – TS is let us say one way of calling the second law, if you want to
say. And then we also have something called the third law of thermodynamics, how
many of you are aware of third law? What is third law?
Student: Every substance has 0 entropy at 0 K.
Correct. So, entropy S  0 as T  0 - this is what is third law, this is what we say, am I
right? And there one point that you all have to remember, whenever we say entropy it is
the degree of randomness in a material. And this degree of randomness can come into a
material because of two important factors, one is called thermal; another is called
configurational. For example, at any temperature there are vibrations inside a material
and these vibrations would cause some kind of randomness inside the material and that
particular randomness that you see is what is referred to as thermal entropy.
And we define that particular entropy in terms of, how many of you know this? How is
entropy defined?
Student: ΔH/Tm
Yes. In fact, if you see the second law definition of entropy is dS = dH/T = dq/T.
Actually this is how the first at the beginning people have defined the entropy. And
because we know under certain conditions, under what conditions dq = dH?
Student: Constant pressure
Constant pressure, so we take it as dH and dS = dH/T. And enthalpy is related to what?
Student: Internal energy
2
H = E + PV we have already written, but how do we measure H? From the specific heat.
So, H=∫ C p dT
0
Basically Cp is the heat capacity of a material at constant pressure. How does one define
the heat capacity? Tell me.
Student: The amount of heat a material can hold for a degree raise in temperature.
Correct. Not hold, the amount of heat that is required to raise the temperature. That is
how we defined and once we define H=∫ C p dT ,

0
T
Cp .
S=∫ dT
0 T
So, these are the basic equations that one needs to remember when we are talking about
the entropy and the enthalpy. And in principle one can even find out how the entropy
increases with the heat capacity. In fact, if you carefully observe this, the heat capacity
of a material itself is not a constant term. How does heat capacity change as a function of
temperature? Do you know?
Student: Increases with temperature.
3
Increases with temperature, what kind of a shape it has, if I plot let us say heat capacity
as a function of temperature, how does the plot look like?
Student: Hyperbolic, like S shape
Yes, it is start like this, from zero, (Refer to the graph in above image) and then saturates
somewhere and this value where it saturates is 3R. One can derive that and show. And as
you can see that Cp actually goes to 0 at absolute 0, because there are no vibrations at
absolute 0. Heat capacity comes because of the vibrations and because the atomic
vibrations vanish as you go to 0 the heat capacity goes to 0 and that is why you will see
the entropy also actually goes to 0 and that is what the third law of thermodynamics.
So, you can see that the Cp more or less is very, very small at and then it increases and as
a result because this increases if I try to see what is dS/dT; that means, the rate of change
of entropy with temperature that can be given as what? dS/dT = Cp/T. So, that means,
you can see that the rate of change of entropy with temperature is C p/T and that means
the entropy always increases with increase in temperature. When I say dS/dT how does
the entropy change as you increase the temperature is it positive, is it negative, or is it
does not change.
If you look at it, if Cp was 0 then we can say dS/dT = 0, it means what? That entropy
does not change with temperature, at any temperature material will always have the same
entropy. And again I am talking in terms of the thermal entropy here. We have not yet
come to what is called as configurational entropy, configurational entropy comes into
picture when you take more than one type of atoms and try to put them together in one
place. Something like you take a box and try to put red balls and the white balls together,
certain number of them and depending on how you can arrange these red and white balls,
there is some randomness associated with that particular configuration and this is what
we call it as configurational entropy.
And you will see slowly as we go along, materials which are ordered will have very low
configurational entropy, and in fact, one can even prove that if a material is perfectly
ordered, have you heard of ordered compounds? For example, sodium chloride is an
ordered compound. You will see every sodium will have a chlorine atom around it, and
every chlorine will have a sodium atom around it, because of the ionic bonding. You
4
have also in metallic systems also you have intermetallic compounds, things like if you
have heard of compound such as Ni3Al. Do you know where Ni3Al is very important?
Student: Super alloys.
Super alloys - the strengthening in super alloys basically comes from that particular
phase which is called γ prime phase Ni3Al. Ni3Al is ordered and if it is in a perfectly
ordered condition, an ordered material also can have some disordered inside it - we call
it as partially ordered. If it is perfectly ordered then in such a case we can prove as we go
along we will do in this course that we can prove that it is actually will have 0
configuration entropy. And in that case the material can only have one more entropy
which is the thermal entropy. Thermal entropy is bound to be there whether it is a
ordered or disordered structure. Only difference between the two is, if it is perfectly
ordered it will not have a configurational entropy; it will have only one type of entropy
which it is thermal entropy. Otherwise if it is a disordered, it will have both of them.
And for a pure metal, if I am considering in principle I will not talk about configurational
entropy because there is no second element there. In fact, if you go deeper people even
talk in terms of the vacancies inside the material and talk about configurational entropy
arising because of the presence of vacancies. But in general for normal study we usually
consider for pure metals that only thermal entropy becomes dominant and
configurational entropy does not come into picture, only when I add a second element - a
binary system or a ternary system or a quaternary system and so on, we start talking
about configurational entropy.
Student: In case of NaCl, existing at room temperature will that be perfectly ordered?
It is ordered, of course, In fact, particularly in ionic solids if to introduce a disordered

into them is not so easy. For example, if you want to remove one of the chlorine atom
the bondings are disturbed, you need to maintain what is called the charge balance. And
if you cannot maintain the charge balance you are in, the system is at a higher energy
state. So, in order to to keep the system at the lowest energy state, system prefers to have
every sodium atom would like to have a chlorine atom around it because of , basically
the exchange of electrons in that particular material.
5
So, you would see that in when I am talking about thermal entropy, you would see that
the thermal entropy because Cp is positive. Cp is never negative we have seen from here
Cp the lowest value Cp can take is 0. So, that is the absolute 0, at any other temperature
Cp has a positive value. If Cp/T is positive, then dS/dT is also positive. Is it not? So, this
particular rate of change of entropy with temperature is always positive, so entropy
increases with increase in temperature. And what is the rate, at which it increases,
depends on the material. For some material which has a high C p you would see that the
rate at which it increases is higher you will have a steeper entropy.
If I plot entropy verses temperature you would see a steeper curve there which we will
see it possibly in the next class as we go along. So, this is something which you need to
understand before you start discussing about the free energy. I think with that we will
stop for today.
Student: Can we induce configurational entropy by increasing the temperature as in like

NaCl if you increase the temperature, bonds start becoming weaker, right. So, the order
might not prevail at a certain temperature.
True, but that particular entropy that you are getting there can also be talked in terms of
thermal entropy. So, that is why it is a little tricky issue to say this particular entropy that
is generated in a material, is it because of thermal or not?
Student: But because of thermal we are induce configurational entropy.
Configurational entropy is assumed to be constant, once you have fixed the amount of
the second element as you go along you will see. How many of you remember the
configurational entropy expression?
Student: k lnW
k lnW. And the k lnW if you derive it further it turns out to be -R∑XilnXi.
6
For binary system we can easily show that it is

ΔSconfig = -R(XAlnXA + XBlnXB)
.Some of you might have seen this and you can see that this is a function of only
composition. And once you fixed that, and you can in fact, you can clearly see that this
would obviously go to 0 when you go to a pure metal side. If you put any one of them as
one either XA = 1 or XB = 1, you would see immediately that the function goes to 0. So,
that is why configurational entropy is 0 for a pure metal and it comes into picture only
when you have more than one element. And in principle, if it is more number of
elements, we write it as
n
∆ Sconfig=−R ∑ X i lnX i
i =1
So any number of elements you can continue that expression. We will talk about multi
component systems as we go along and there you will understand as we go along. We
will stop now.
7
8
Lecture No. # 02
Free Energy, Stability, Equilibrium in a Unary System
Last class we have started the basic equations in thermo dynamics. We started with this,
H = E + PV
all of you are aware of it, then we introduced the second law, which is basically we said
it is this G = H-TS
.And that is what introduces two new parameters, which are one is the entropy (S) and
another is the free energy (G). And we said the stability of all the phases is actually
related to the free energy, and let us go a little details into this and see how to represent
free energy as a function of temperature? And how does that give us information about
the stability of phases? To understand that, first let us look at how to evaluate the H and
S. If you look at it, we all know that
dH =Cp.dT
So, as a result I can write it as
T
H=∫ C p dT , where T goes from let say 0 to T.

0
9
In thermo dynamics we have what are called standard states. All these are actually relative
numbers. We have so, you need to set up some particular condition which we call it as a standard
state, and that condition that we set up is at a temperature called 298 K which is room
temperature. So, we define H298 = 0. For most of the materials, you can even change that
depending on the particular condition, you can change this number depending on the
particular circumstances, but usually we assume H298 = 0, and if you assume that, and then try
to plot.
Student: Here H298 = 0, so if we integral from 0 to 298?
So, that is why this H become negatives so that is what we are going to see it right now.
So, if you plot Cp as a function of temperature, we know that C p goes something like this
a typical s curve (Fig. 1 above), all of you are aware and C p goes to 0 at absolute 0, this
is 0. And we also know that the maximum value we reach is approximately around 3R.
And now, if I try to plot H as a function of temperature, you would see the H as a
function of temperature would go something like this (Fig. 2 above), where here at 298 it
is 0 and below 298 K actually it is negative. So, in principle if I define H at 0 K as 0 then
obviously this everything will be in the positive quadrant. So, this is a typical way of
representation of the H and once I know H is represented that way, I can also start
defining S.
Because S, the way we define the second law the entropy term is
dS = q/T and q at constant pressure is known to be as dH. So, actually
10
dS = dH/T
and that also leads you to the other equation dH = TdS
which is a very useful as we go along, we see this because this becomes useful when we
are defining the free energy more easily.
So, you can see that dS = dH/T. then I can simply write
dS = CpdT/T
Student: Why don’t we take H at 0 as 0?
You can always, because 0 K is a term which is an ideal term, which you rarely reach
and if somebody has to experimentally measure anything be rare excepting a few
physicists. Most of us who deal with, either chemical engineers or metallurgical
engineers or material scientist most of us deal with room temperature. So, as result, we
start at room temperature. So, most of the applications demand temperatures beyond
room temperature not really sub 0 to a large extent. So, as a result, understanding this
and in principle even if you take it as that it is only a scaling factor, things get changed.
So, as long as one thing you know then you can start calculating what it is at as a
temperature, it is not a big problem it is only changing some kind of number some
reference is required. So, this is a more easier reference for us to deal with, room
temperature is a easier reference.
Once I understand this, I can even write now as
11
T
S=∫ C p dT /T
0
And this is, where we define a third law of thermo dynamics, where S = 0 at T = 0. So,
that integral we assume it to be 0 and because we also know that the C p itself goes to 0
there are no vibrations, when there are no vibrations there is no thermal entropy. There is
no thermal entropy, please remember. We will come to the configurational entropy a
little later and this configurational entropy need not have
12
to go to 0 at absolute 0. Thermal entropy is 0, because there are no atomic vibrations. If
there are no atomic vibrations, there is no randomness created by atomic vibrations. So,
the crystal is prefect, atoms remain where they are and they do not move. And as a result
they do not cause any randomness inside the material.
(Refer Slide Time: 06: 36)
So, this is how we basically define this and in principle I can also plot, if I plot S versus
temperature. S versus T if you plot, S versus T also would have a curve like this (graph
shown in image above) basically it comes from the Cp type of curve. It would be
something similar to Cp curve and it goes to 0 at absolute 0. And then keeps on
increasing and not only keeps on increasing, if you carefully observe dS/dT.
dS/dT = Cp/T
So, because it is Cp/T, in principle at very high temperatures this slope should come
down as you increase the temperature, but there is a small hitch here that C p is also
increasing with increasing temperature. If Cp was constant then I can say the slope keeps
on decreasing as the temperature increases. Entropy increases with increasing
temperature, but what we are talking is the slope of this entropy; that means, the rate at
which this entropy increases with increasing temperature is going to decrease as the
temperature increases.
So, initially it will be a steeper curve and later it becomes a kind of a shallower curve.
And in principle this curve will never become let say 0; that means, the slope of this
curve will never become 0. It will become 0 only, when Cp become 0 and Cp never
13
becomes 0 at higher temperature, we know pretty well. So, it will still have a low slope,
but it will maintain certain positive slopes. So, not only S is positive, dS/dT is also
positive this is a very important thing that you need to understand. The rate of change of
entropy with temperature is always positive, because C p is positive. The lowest value
that it can go is 0 and that is only at absolute 0 and at absolute 0 we do not really talk
about rate of change, because it is one single temperature.
So, as result you can see that, but what is important is dS /dT decreases as T increases,
because Cp does not increase at the same rate as T. If you remember some of you might
have gone through, how we express Cp = A + BT + CT-2.... So, there is an expression
which is used for Cp to fit that particular curve, which we have drawn before how does
the Cp change with temperature. So, it fits to almost at a in a very small range when I
look at it, it can be fitted as A + BT, but if you take the whole range then you see
particularly at low temperatures the Cp change with temperature is very small.
14
So, this is one important point that you need to remember and will keep bringing this
concepts as we go to the free energy. Now, let us look at the free energy and that is the
most important thing for us because as we said the whole stability is dependent on free
energy. Now, let us try to expand this. I said just now
G = H - TS and H = E+ PV. So, I can write it as
G = E + PV - TS
Now, let us differentiate this, and we get
dG = dE + PdV + VdP – TdS – SdT.
Now, if you look at it dE + PdV= d H. We all know that dH is define in terms of that
and TdS =d H, So, as a result these two get cancelled and you end up in an expression
that
dG = VdP - SdT.
This is a very very profound Maxwell’s equation. And that leads you to a lot of
interesting things such as what is called anyone knows, we talk about Clausius
Clapeyron equation. Clausius Clapeyron equation comes from there; we will see how
does it come.
So, this (above equation) tells us that

dG/dT|p = -S
And S is never negative, we have already defined S is what the degree of randomness
and a degree of randomness can never be negative, it can be minimum, it can be 0. The
lowest value of S can be only 0 and then it will be more than 0 at all other temperatures.
As a result, because S is positive dG/dT, because there is a negative sign before the S.
And if the S would have been negative then the dG/dT would have become positive. So,
I can write as S is positive, dG/dT, is always negative. What does that tell me, it tells me
that, if I plot G versus T. Let say any material like this, I try to calculate, what is the free
energy of this material and how does it change as a function of temperature?
15
If we now plot G versus T, how should the free energy curve look like?
Student: It decreases
So, it decreases. Is this fine (Graph 1)? I would say it is not fine.
Student: S increases and not decreases
Yes, when will this be fine?
Student: S remains constant.
Yes, when S remains constant this plot is perfect. What is this tell you, it tells you that it
is a linear relation. A linear relation means what? Slope is constant, what is the slope of
free energy versus temperature? Slope is minus of entropy and this is decreasing fine.
So, the minus is taken care, but the slope is constant there, if the slope is constant it tells
you that entropy is not changing as a function of temperature, but we know pretty well
that the entropy changes with temperature.
So now, can you tell me how should it change? So, this (graph 1) is wrong. So, what is
the best way?
Student: Hyperbolic
16
Yes, it should decrease like this (Graph 2). To an extent that as I come closer and closer
to absolute 0, the slope should tend towards 0, because at absolute 0 entropy is 0. So, the
slope of G versus T should go to 0 at absolute 0. This slope is 0 at any other temperature,
the slope is higher and the slope keeps on increasing for example, if I look at slope here
(point A), I draw a tangent let say here, and I draw another tangent here (point B). You
can see the slope here (at point B) is higher, the slope here (at point A) is lower. So, at a
at any higher temperature the slope of G versus T is going to be higher and higher in a
negative sense keep that in mind, definitely the whole curve is the slope is negative, at
all temperatures slope is negative, at only once temperature the slope goes to 0, at no
other temperatures slope is either 0 or negative.
This is the temperature (T=0) where the slope goes to 0, at any other temperature slope is
negative. So, that is what you see dG/dT negative and that absolute value of the slope is
the entropy. And that absolute value of the slope, because it is higher at higher
temperatures, we will see this is the typical nature. And now, if I find out what G at
absolute 0, depending on what is there on the right side, you would?
Student: H
H, very easy G = H at T = 0, isn’t. So, this value is nothing but H whatever it is. And at
any temperature if I draw a tangent and extrapolate this tangent to the absolute 0, for
example, if I extrapolate this tangent (at point A) to this value, this intercept is what?
Student: H
H, because G = H -TS and slope is the S. So, you will always get the intercept as the H
at, so at different temperatures and in principle you can see as I increase the temperature
the slope is increasing. As a result the intercept also will increase, which also gives you
an idea that the H increases with increasing temperature. So, where this particular
tangent is going to intersect the vertical axis is going to keep on increasing. And that
gives you an idea that the H of any material, H of any phase that we are talking is going
to increase with increasing temperature and at the same time. So, not only H increases
with
17
temperature, we also see S increases with increasing temperature and G decreases with
increasing temperature, fine.
And in principle, if I write both G and H in the same thing you would see (see graph
above), we have just now seen H is H goes something like this (increases with
temperature) and you see G would go something like this (decreases with temperature).
this difference between this two is what?
Student: TS
TS, very easy. And this difference keeps on increasing as we go to higher and higher
temperature that tells you that both the T is increasing and S is increasing as a result you
would see the difference keeps on increasing. So, in principle we can plot like this. Now,
let us try to understand, is this all clear? So, from now onwards, whenever you draw free
energy be very clear, that G cannot be a straight line as a function of temperature, which
you would have possibly drawn a number of times before. So, this has to go into your
mind that G, the slope of it being nothing but entropy it can never be constant and
entropy is never constant.
18
So, once this is clear let us try to understand what is a necessity of G? Why should
people introduce this concept called G? Earlier people thought the stability of phases,
when do I say something is more stable, I compare let say two things. Let say simple
case, let say I am talking about aluminium. Aluminum it can be a liquid, it can be a solid,
it can be a gas. Let say I am comparing a liquid and a solid, if I consider the enthalpy of
liquid and the enthalpy of solid. What is higher? Liquid is higher. Why is it higher?
Because the Cp of the liquid is always higher than the Cp of the solid, heat capacity of a
liquid is always higher than the heat capacity of the solid why? The vibrations are more
T
in a liquid, when compare to a solid. As a result, Cp is higher and H=∫ C p dT .
0
Thus, you can say Hliq > Hsolid. So, as a result, if I say that at melting point solid is
changing to liquid, it is going from a lower enthalpy state to a higher enthalpy state.
Now, a reverse transformation, where the liquid is transforming to a solid, it is going
from a higher enthalpy state to a lower enthalpy state. So, as a result, I am seeing that
below the melting point a phase with a lower enthalpy is stable that is the solid, above
the melting point a phase with a higher enthalpy is stable. So, if I use enthalpy as a
stability criterion, I have difficulty in the sense that I cannot say a phase with a lower
enthalpy is always stable, I cannot say this. Because I myself I am
19
able to see that the liquid is stable above the melting point and liquid always has a higher
enthalpy. So, how can I say that this phase having a lower enthalpy is always stable.
Same problem comes if I take entropy, if you see entropy of a solid and entropy of a
liquid, Entropy of a liquid is again higher than that of the solid. S liq > Ssolid. Again, I am in
the same situation that I can never say the phase with a lower entropy is always the
stable or the phase with a higher entropy is always stable. I can see this reverse
phenomena happening, above the melting point the phase with a higher entropy is more
stable and below the melting point a phase with a lower enthalpy is more stable. And if I
take this lower entropy phase or a lower enthalpy phase which is a solid and take it
above the melting point I cannot retain as the solid, it immediately starts transforming
into a liquid though it has a lower entropy and lower enthalpy, it does not want to
remain. So, that is where people thought what is that, that will help us to be able to
understand the stability in terms of one parameter and that turns out to be free energy
and how to understand that.
If I try to now, plot G versus T for this liquid and solid. Let us try to super impose the
free energy of the liquid and the free energy of the solid, on this curve and see what
happens (Graph 1). For example, I can draw a curve something like this and then say this
is the free energy of the solid. It is satisfying whatever I have said before, that its slope
goes to 0 at absolute 0 and its slope keeps on increasing in a negative sense as the
temperature is increasing,
20
Is it fine. Now if I super impose a liquid free energy curve on this, how should that be?
What should be the characteristics of that with respect to this curve for solid?
Student: At 0
At 0
Student: Higher free energy
Higher free energy
And, why a higher free energy?
Student: Higher enthalpy
Because, it has a higher enthalpy, at absolute 0 free energy is equal to enthalpy. And the
liquid enthalpy is always higher than the solid enthalpy. Next, second thing is that, its
slope should be higher. So, if the slope is higher than I will see that the curve will be
something like this (refer to the liquid curve in graph 1). Whenever you have two curves
one with a higher slope than the other; obviously, they all they both have to intersect. It
is not possible to have two curves with a different slopes not intersecting. So, as a result
you would see that this is nothing but H liq at absolute 0, this is H solid at absolute 0 (see the
vertical axis of Graph 1).
When I look at this curve, it gives me very interesting phenomena, what is that
interesting phenomena? There is as intersection here (point P). And, if I look at what is
there above the intersection, what is the nature of the curves above the intersection and
what is the nature of the curves before below the intersection, what you see is that, above
the intersection the liquid curve is always below that of the solid curve; that means,
liquid has a lower free energy than the solid above that intersection.
And below the intersection, if I look at it the solid is always below that of the liquid, .
that means the solid has a lower free energy than the liquid below the melting point and
above the melting point. That particular
21
intersection of the two free energy curves is we called it as a melting point, why we call
it as a melting point? Because below beyond that temperature, I see that the liquid free
energy at all temperatures is lower than the solid free energy and because it is lower I
can now say the liquid is more stable above the melting point, because it has a lower free
energy. And similarly, when I come to the temperatures below the melting point I can
see that the solid free energy is lower than the liquid free energy at all temperatures. So,
at all temperatures below melting point, I can always say the solid is more stable because
it has a lower free energy.
Now, we have one parameter which tells you that if that is lower the phase is stable and
that parameter is this free energy (G), Gibbs free energy. So, if the free energy is lower
than it is stable. Is the free energy is higher than it is what we call it as, what do we call
it? It could be, it could be meta stable or unstable this is what that is where we end up
into the regime of kinetics. Always we say thermo dynamics tells you what is feasible it
does not mean that it would happen. For example, if I look at this is (Graph 2), this is
some state called A, this is another state called B let say, and if I am plotting this in
terms of the free energy. The free energy of the A is higher than the free energy of the B,
but it does not mean that this A will immediately go into B.
For example, a simple example could be even this, if you look at this duster, if I leave it
this duster on this table and I keep this duster in this form (horizontal), what is having a
more lower free energy? This horizontal form; obviously, because the CG is lower, if the
CG of it is lower I can say yes this has lower energy state. And if I am keeping it here
(vertical), it has a higher energy state, but I can keep this like this for ever, until I push it,
I can keep it forever; that means, this higher energy state is able to remain in a higher
energy state forever almost until somebody gives you.
Student: Activation
Activation, we called it as activation that is what this is called activation barrier. This is
what we call it as ΔG* (Graph 2); we will talk about it more later. And how to
understand this, if you look at it, if you take this state, this is the duster (D1) and if I say
that I join both the diagonal. And I say this (point C1) represents the free energy that is
the CG of that, if I say that represent the energy of the system and from there it comes to
this level (D2), where I join again and say this (point C2) represent that energy and
between
22
these two states, when this is going from D1 to D2. You come to a state like this, which
is incline and that state if you carefully observe (D1’) and that state if you join you see
that the CG (C1’) of that is higher.
So, it is going from a lower energy to a higher energy and then further coming down that
is what we call it like an activation hill and so, it goes to and this particular state is what
we call it as unstable state (see graph 2), because if you keep this in this state, it will not
remain suddenly it falls of either this direction or that direction. So, that is the interesting
about unstable state either it come to a metastable state this is the metastable state, this is
the stable state (see graph 2). So, it can either shift this direction or this direction
depending on in which direction your force is. So, you can see that this is the difference
between thermodynamics and kinetics. So, thermodynamically, we say this (state B) is
more stable (than state A). In principle it should transform, but it is not happening. Many
of you might have heard of, all of you know about steels I suppose, if I look at the
microstructure of steel, what would I see in microstructure of steel.
Student: Pearlite
Pearlite, what is pearlite?
Student: It is a mixture of ferrite and cementite
Ferrite and cementite, do you know that this cementite is not a stable state?
Student: Metastable
It is a Meta stable. In principle, if I take this cementite to a higher temperature and heat
it, it becomes ferrite or alpha or iron plus carbon, graphite. In principles cementite is not
stable, but we see the steel at room temperature cementite exists forever, for ages
nothing happens to it. The cementite remains as cementite. You make a mild steel and
then just leave it and it remains why because it as a large activation barrier for
Fe3C  3Fe + C and unless you provide that activation barrier, it would not change.
And, as a result that activation barrier you are providing by heating the sample to a
higher temperature. In fact, that is what we do, when we convert one of the cast irons
into another cast iron have you heard of that. There is a treatment, what is that treatment
called?
23
Student: Malleablization
Malleablization, where you take a white cast iron which has cementite network,
cementite phase and you heat it to about 1000 ºC where that cementite splits and gives
you a graphite. So, in principle one can even people when they do spheroidization of
steels, if they do spheroidization of steels for a very long time. You would see that
cementite spheroids that are forming can even split and you get ferrite plus graphite
people have done this and seen. So, in principle, if you can provide that particular
activation any metastable state would like to go to a stable state otherwise it would
remain in a metastable state forever. So, this is the difference between thermo dynamic,
thermo dynamics tells you that this is the stable state.
In principle the system would like to be here (state B, graph 2), but it can still remain
here (state A, graph 2). So, it is like local minima. People talk about local minima this
(state B) is kind of a global minima. So, it remains in a small trust there and if you push
it, it will go up and then come down. So, it is like that. So, you can see that this is the
basic concept of the free energy.
So, we have now defined that this free energy is a concept, which we can use for
understanding what is stable and what is metastable. I can say for example, I have to tell you
this again clearly that if I am at this temperature (T1); that means, I have a taken a solid
heated to this temperature, what should happen when I heat the solid
24
to this temperature. It should become a liquid why? It should become a liquid because
solid has this free energy (Gs) at that temperature; liquid has this free energy (Gl) at this
temperature, T1. So, Gl < Gs .
In principle solid should become a liquid it is possible that solid can still remain as solid
which we call it as a metastable solid. If the conditions are not ripe enough for the solid
to become liquid in case of solid to liquid this does not happen, because solid to liquid
transformation is much faster than liquid to solid transformation. You see the reverse
happening more easily, if I take a liquid cool it to the temperature like this (T2) you can
see that liquid has a higher free energy Gl is higher, Gs is lower. In principle liquid
should become a solid, but there are many- many cases, where liquid does not become a
solid and remains as what is called an under cooled liquid and in fact, that is the
principle behind getting a glass we will slowly see a little later. For example, may be you
know the time has come so let us see it right now or no- no let us leave it unnecessarily
we are deviating, we will talk about liquid to glass transition a little later we will see.
So, you can see that this liquid can remain as liquid the reason is from a liquid if you
imagine for a solid to nucleate, if you have a whole liquid bunch of liquid and you have a
solid coming out of the liquid (see Fig. 1 above). You have a problem what is the
problem?
25
Student: Surface
It has to create a new surface. This surface did not exist before isn’t. That surface (S)
has to be created, though the liquid has a higher free energy than the solid and liquid
wants to transform into a solid. You would see that because this solid this new surface
which we call it as a interfacial area between the liquid and solid and that is represented
by what is called interfacial energy called γsl. And every surface, every interface is
associated with certain energy that energy has to be spent that energy.
You have to you have to create that particular surface and to create that surface you need
to have certain energy with you and if that creation of the surface is difficult the liquid
does not become a solid. And if you have seen what are called cooling curves at some
stage. Have you seen how does a pure metal liquid become a solid? (see Fig. 2) If I take
a furnace melt a solid put a thermo couple there inside and then switch off the furnace
and then watch how the temperature of this liquid is changing as a function of the time
and I plot that T versus small t. And I usually see this kind of a thing (Fig. 3) for a pure
metal many of you might be aware of this and why does this happen.
In principle I expect that it should be like this (Fig. 4) liquid, this is the liquid and once I
bring it to melting point, this is the melting. This liquid should transform into a solid at
that temperature, but the liquid does not transform to solid. This liquid to solid
transformation does not occur suddenly here at the melting point, but the liquid have to
be cooled to below the melting point to some temperature before this liquid actually
starts transforming why is that, that is because when you cool this liquid from this
melting point to a certain lower temperature (Fig. 3) this what we call it as under
cooling, this is what we call it as ΔT under cooling, when you provide that under
cooling, what is happening to the free energy let say from the melting point I have
brought to this temperature
Once I bring it to this temperature (T1, Fig. 5) there is certain, what is called driving
force (ΔG). At the melting point, what is the driving force 0 and that is what we define it
as equilibrium. So, in a in a single component system equilibrium is defined as that,
where the free energy of both the phases is equal that is what it is at the at the
intersection. At the intersection Gs = Gl this is the definition of equilibrium and so, we
say this is the melting point, where the free energy of the both the phases is equal and if
the free energy
26
of both the phases is equal for any of the transformation either from liquid to solid or
solid to liquid there is no driving force.
If I take a liquid keep it at the melting point in principle. It will not become a solid
similarly, if I take a solid heat it up to the melting point and just hold it there. In
principle, it should not become a liquid excepting if you heat it by a very very small
amount. It could be let say third decimal in terms of the temperature. You need some
small driving force for this solid to become a liquid similarly, for the liquid to become a
solid you need a driving force the only difference between a solid and a liquid can be
easily understood. If you take an ice cube from a fridge let say somehow you have taken
a ice cube without holding it and you are just leaving it in the air let say and the
atmosphere is above the 0 degrees. So, this ice cube has to melt how does it start
melting, where does the melting start? Surface, anything?
Student: Corners
corners first is it starts that is why, you will see that the ice cube the corner starts
becoming rounded first why is that so, because than the next stage is actually the edges
after that actually it comes the surface.
The reason is the surface energy, the surface energy is highest at the corner why is that
so? Surface energy comes from what is called broken bonds? If you take a material for
example, a solid if I take, inside the solid ice for example (Fig. 6), if I am taking ice, ice
has a structure do you know what is the structure, what is the crystal structure of ice?
Student: h c p
h c p hexagonal, it is a hexagonal structure. So, this hexagonal structure if I think of the

atoms which are on the surface, in a hexagonal structure you remember what is the
coordination number, twelve; that means every atom should have twelve nearest
neighbours. So, you will have six in the plane three above the plane; three below the
plane, am I right. Now, if you imagine that there is an atoms sitting here (point P, Fig. 6)
that atom has six atoms in that plane three atoms below that plane, but there are no three
atoms above the plane so that means, that particular atom there has certain number of
broken bonds in principle, it is 3/12, 3 out of the 12 bonds that it should have three bonds
are not there.
27
As a result that particular atom is at a higher energy state, because it is not having all the
atoms which it suppose to have where as if you think of an atom which is at the centre
somewhere here (point Q, Fig. 6). It has all the required number of atoms around it as
per the coordination number. So, the atom here (point Q) is in a much better state when
compare to atom here (point P). So, this is at a higher energy state so similarly, if I take
this corner. This corner would have more number of broken bonds and if I now think of
a solid. Think of a solid and now consider these corners and this edges and this surface
(Fig. 6). You can easily see the corner would have more number of broken bonds the
edge would have next level than the surface than the atoms inside.
So, energy levels go by that hierarchy. Any atoms sitting at the corner of a cube would
have the highest energy and that atom that particular region, because it is at highest
energy it wants to transform first to something, which is a lower energy state why
because I am at a temperature above the melting point, I am at room temperature I said,
at room temperature the liquid is more stable than the solid. So, ice cannot remain as ice
it wants to transform to liquid. So, that particular region which is the highest energy
immediately transforms then comes the edges, edges starts become rounded and then the
surface and so, the actual melting starts from the surface and goes into the interior why
because the surface area is available for it.
Now, if I take a liquid and try to make it a solid, where is the surface available for the
liquid until you think of I pore this liquid into a mold, sir the mold is there has a surface
for the solid to nucleate and that is the reason why. In most of the casting process the
solidification starts at the, at the mold interface mold liquid interface and there are
interesting people who do what is called container less solidification. In our department
there is a professor I do not know how many of you have heard of Phani Kumar he has
set up what is called a levitation facility is very interesting to see please go and see. This
is the first levitation facility in the country, where you take a piece of metal like this.
You levitate it, it is standing in the air now and then you heat it using a let say an
induction heating or any heating let say electron beam heating any heating you can heat
it and melt it. Once it is molten, it becomes liquid and this liquid droplet is floating in the
air now and you switch off that heating source.
Now, it starts cooling and you can actually observe how does it solidify, where does the
nucleation start, very very interesting phenomena a lot of people have struggled to
28
understand this whole concept of nucleation, where does the nucleation start and in fact,
long back there is a great man by name Turnbull who has done, what is called droplet
immersion technique, what he has done is he has taken a liquid metal allowed it to go
through a nozzle and made it into a stream of droplets an emulsion of droplets by kind of
atomization, when you do this droplets of different sizes form. And different sizes if you
look at it the smaller droplets will have lesser number of heterogeneous nucleating sides,
because people knew for quiet sometime that this kind (Fig. 4) of a solidification is
happening only because of heterogeneous nucleation.
Otherwise in principle, it should be this (Fig. 3) why because, when I under cool I am
providing sufficient energy. And when this energy is equivalent to the surface energy
that is when the solidification will start, until the energy that I have in my pocket, its like
you want to spend something you should have something in your pocket isn’t. Unless
you say I can use my credit cards sir always so, otherwise you should have something.
So, here also is the situation so, unless you have a driving force some energy available
with the liquid excess energy that is available with liquid. It cannot spend that it should
have sufficient excess energy that is sufficient for creating a new surface that is when
solidification will start. So, the liquid will keep on under cooling, until you get sufficient
energy that is available.
We are going to slowly may be next class relate this to this (Fig. 3 to Fig. 5). There is the
relation between ΔG and ΔT will see that also, what is the relation? And you will see
that once I provide sufficient under cooling and at least graphically you can see (Fig. 3)
as I increase the ΔT because as I go more and more away from the melting point. This is
increasing, I can see that, but in principle one can calculate that also. So, I say that at a
given ΔT there is a certain value of ΔG and if that ΔG is equivalent to that of the surface
that is why if you remember anybody, who has done some nucleation.
You would write this expression for a spherical nucleus that is coming out we say
ΔG = 4πr2γ + (4/3)πr3ΔGv
When I am creating a spherical nucleus inside a liquid this is the energy that is given out,
this is the excess energy of the liquid (second term in above equation) where as this is
the energy (first term in above equation) that you need to create for a spherical particle..
So, you would see that unless these two are equal. You would never see the free the
transformation taking place in such way that this overall ΔG has to become
29
Negative. When this is negative than solidification will occur. If that is 0 or positive
solidification will not occur.
So, that is the reason, why we need certain under cooling that is the reason why people
saw that usually you do not see this, this is what we call it as homogenous nucleation
(Fig. 3). This is what we call it as heterogeneous nucleation (Fig. 4). Heterogeneous
nucleation occurs, when there are already certain particles available in the liquid. In fact,
that is one of the concepts of what is called grain refinement, inoculation, people
regularly use this term. So, if I add certain particles into my liquid. And pour it into the
mold. Now, these particles if they have a similar crystal structure as that of the solid that
is coming out then this surface energy (γ) is very low, whenever the structure is similar.
You must have heard of what is called epitaxy anything can grow very easily if the
structure is similar. So, you would see that if I add particles, which have the same or
similar crystal structure as that of the solid that is coming out. You would see that it
would easily nucleate. So, that is why there are specific nucleating sides for specific type
of solids.
You do not add everything for everything for example, in aluminium alloys. We know
for example, there are something like TiB 2 titanium boride, titanium carbide. TiB2 is
hexagonal. If you carefully observe there is a similarity between hexagonal and
aluminium, what is the crystal structure of aluminium? FCC. So, what is the similarity
between FCC and HCP?
Student: Closed pack structure
Closed pack structure yes anything in terms of any planes, is there any similarity atomic
planes.
Any plane of HCP does it have any similarity with the any plane of FCC, because why I
have to say this is nucleation always occurs on a plane. It cannot occur anywhere, it
happens on a surface that is why planes are important. There is a certain plane in HCP,
which is exactly similar to another plane in FCC, how much of crystallography you
know, 1 0 0 sorry, what is the closest pack plane in HCP? What is the indices of it, (0 0 0
1) - the 0 0 0 1 plane is similar to another plane in FCC, can you think of it, (1 1 1)
prefect. So, the aluminium can nucleate on on that particular (0 0 0 1) of TiB2 very easily
30
with it is (1 1 1) plane parallel to that the basal plane of the TiB2, and then the nucleation
can occur. I think will stop with this, thank you very much.
31
Lecture No. #03

Effect of Pressure on equilibrium transformation: Clausius Clapeyron equation,
phase diagram for unary system
But the only thing when we are talking about stability, we are always talking about
stability with respect to some other thing, as far as that particular phase is concerned, it is
becoming more and more stable with increasing temperature.
Student: Entropy is also increasing? System is becoming more random, then how is
becoming more stable?
Yes. Why should you think that more randomness means it should not be stable, do you
think you are perfect. So, we are all stable. So, it does not mean that imperfection means
lack of stability that is the reason why we are saying that either enthalpy or entropy alone
cannot talk about stability. A combination of these two, which is the free energy alone,
can tell you. So, if I simply use only entropy concept, if you say that the more the lower
the entropy, the more stable it is, that is not the right way of looking the lower the
entropy, the more perfect it is, it does not mean that perfect means stable, may be
spiritually. So, in materialism at least we cannot say that. So, so that is why, we are
saying that when we say stability it is only in terms of the free energy that we try to look
at it enthalpy; for example, if that is a case even the enthalpy wise the lower the enthalpy
the phase is more stable in principle.
For example, look at case like the electrons around a nucleus in various orbitals, the
closer the electron to the nucleus the more stable it is, there it is only enthalpy that we
are looking at. And that is why you know, we are talking about x ray generation that
when you knock of one of the electron from a higher shell the electrons falls into a lower
shell. So, there we are basically looking at the energy aspect of it, the bonding energies,
isn’t? So, basically when you look at convert all these things into free energy term, it is
the free energy that dominates, and that is what we have seen in the last class, and let us
try to continue from there.
32
So, we looked at this (see graph above) and we said, if this is let us say liquid and if this
is solid, we say that there is a temperature called T m, you can call it as T m or T f
freezing point or melting point depending on how you look at it. And below which you
say the solid has a lower free energy than the liquid and above which you have the liquid
having a lower free energy and that is why, we always can say that any phase with a
lower free energy is more stable, when you are comparing two phases, this is comparing
two phases. And in principle, if a particular system has a large number of phases
possible, I should be able to put all of them. For example, if I am taking a metal like
aluminium, what are the possible phases, that I can think of in aluminium?
Student: Alpha
Alpha
Student: Beta
What is beta aluminium?
33
If I take aluminium and keep on changing the temperature, increasing the temperature,
what are all the possible phases that I can expect in aluminium? Anybody knows.
Aluminium as a solid has only one structure, up to its melting point. So, it is only one
single phase up to its melting point and from the melting point up to its boiling point, it
is one single phase, which is a liquid and then from the boiling point onwards, you can
call a gaseous aluminium. So, if I want to bring the third phase, you can call it a third
state of matter and thermodynamically any state of matter can be even called as a phase,
how do you define a phase by the way?
One of you talk.
Student: Chemically homogenous, physically distinct
Distinct. Correct. So, as long as it is chemically homogenous and physically it has a

boundary, to distinguish it from something else, we can call it as a phase. So, whether
you call gas as a phase or a state, it does not matter. You can separate a gas from a
liquid; there is a boundary between them. So now, if I want to put gas on to this free
energy diagram, how does it look like, how should it look like.
34
One is now; I am trying to compare solid, liquid and gas for enthalpy and entropy.
Compare all of them, among all these, which one should have the highest enthalpy.
Student: Gas
Gas; obviously, because ∫CpdT, if I look at it the Cp of the gas should be higher than the
Cp of the liquid and in turn higher than the C p of the solid. And similarly, entropy also,
entropy of the gas should be higher than the entropy of the liquid and that of the solid.
So, as a result, if I now, I am plotting the, the free energy of liquid and solid, I have
already plotted (Refer Slide Time: 02:29) . Now, if I want to super impose free energy of
the gas on this, what should it be the nature of that curve with respective to these two
curves .
Student: At 0 T, it will be on top
Top. Yes and then what about its slope.
Student: More slope
It will have, it will be steeper, it will be steeper curve. So, and as a result and also now, if
I look at, it should intersect the liquid somewhere and that is what will be the
temperature where the gas intersects the liquid, what would you call that?
Student: Boiling point.
As the boiling point.
35
So, if you look at it, so you should have, if I extend that, it would be something like this
(see graph above), where it intersects the liquid curve at some point. It will be steeper, in
principle, I am all I am just simply, you know doing schematic diagrams one can actually
calculate all of them, what you need to calculate the free energy?
Student: Cp
Yes, what is the fundamental quantity that you need C p, as long as you know Cp as a
function of temperature. One can simply integrate, if I know the C p function (A + BT +
CT-2), what are the A B C values for a given material over a certain temperature ranges
interestingly, if you carefully observe it. If I take C p of a solid and that particular
function with the A B C values, these A B C values will not remain the same for the
same metal when you when I go from solid to liquid. It will have different A B C values,
because the rate at which Cp changes with temperature in a liquid will not be same as the
rate at which it changes with temperature in a solid, because atomic vibrations in a solid
will be different from atomic vibrations in a liquid so similarly, in a gas.
So, the same that is why, whenever you go to a hand book and look for a C p values.
There are large numbers of handbooks, which will talk about Cp values. For example,
Hultgren is one of the very popular handbook, which will give you and CRC handbook
is another handbook, which will give you. The thermodynamics parameters C p
parameters even if you go to simple book like Smithells metals handbook, smitten metals
handbook also gives you for all metals the A B C values for the C p and of course, if
36
necessary D value also and these values they always tell you this is the range over, which
these values are valid temperature range and once you cross that particular temperature
range these are not valid, you have to use another values of A B C.
So, as long as, I know these values of A B C and I know the C p function, I can happily
integrate it get the H, get the S and once I get these two, I can calculate the G and plot it
as a function of temperature. So, whatever schematically I am drawing these are all
calculatable and people have been doing in fact, there is a journal which is a dedicated
for this have you heard of journal were people do a lot of thermodynamics. There is the
journal called CALPHAD calculation of phase diagrams. So, where all the phase
diagrams that we know in standard handbooks have been calculated using
thermodynamic principles, what is the connection between the thermodynamic principles
and phase diagram, we will see as we go along in this particular course itself.
One can in principle any phase diagram that you see in a handbook by the way how do
you people get phase diagrams? Have you seen phase diagram like iron carbon phase
diagram, how did people get that phase diagram?
Student: Heat it to a particular temperature, Get that XRD, we can see the phase
How do you do, you mean you will do a high temperature XRD.
Student: Thermal analysis, DSC
DSC, how can it tell you what phases are there?
Student: XRD can tell
But XRD you are doing it by cooling it back to room temperature only isn’t. Or are you
doing XRD at different -different temperatures,
Student: Change in resistivity
Change in resistivity anything else?
Student: We can go different -different temperature, we can heat the sample also and we
can take the temperature
Yes that’s what you can heat the sample to different temperatures no problem, but at that
temperature what phase is present, if you want to know the only way to know is doing a
37
high temperature XRD. If you cool back and then do a room temperature XRD you do
not
38
know whether that phase has transformed to something else during this cooling, in most
of the phase diagrams we see this happening, most of the time the phase diagrams that
we see regularly in our phase diagram handbooks. Particularly the liquidus and solidus is
obtained by a particular method forgetting about solid state transformations.
Student: Cooling curve
Cooling curve, cooling curve that is what is called thermal analysis. So, if you draw a
cooling curve temperature versus time heat the sample to a higher temperature above the
melting point, cool it and then you see what is called a changes the slopes at the liquidus
and at the solidus and from that we can, we get if I can do this for various compositions.
I will be able to get what is the liquidus, what is the solidus then coming to what about
room solid state transformations.
Let us say in iron carbon diagram, we have a eutectoid transformation, we also have
gamma becoming alpha, how do I get these things that is where you can use X-ray
diffraction, if you have read X-ray diffraction there is a chapter in Cullity which talks
about how to get phase diagrams. Particularly the solvus lines, how to get solvus lines
using the phase using the XRD. So, one can do it, but what I am saying is whatever you
have generated from the cooling curve analysis, which is what we see in all the phase
diagram handbook that you see, most of them excepting in cases where experiments are
difficult.
For example, if you are talking about a radioactive element is not so easy to do an
experiment; you have to do it under very controlled atmosphere. If an a element has a
tendency to oxidize, you can still do a phase diagram measurements by you know
melting and casting and doing cooling curve analysis in vacuum or in argon atmosphere,
but some of the elements, which are not stable elements, which have half life being very
small some other radioactive elements, how do I do that phase diagrams or in some
cases, where the melting point of one of the metal is so high.
For example, think of tungsten I want to generate aluminium tungsten phase diagram,
how do I do it, I cannot melt tungsten I do not have crucibles available, where I can melt
tungsten because tungsten melting point is 3400 ºC or so. So the only way to do in such
cases is this kind of calculations. So, what we do is I can generate the phase diagram on
the aluminium side of the phase diagram because aluminium can be easily melted no
problem. So, aluminium tungsten phase diagram
39
on the aluminium side, you can generate experimentally do thermodynamic calculations
and see if the phase diagram that you get from the thermodynamic calculations fits well
with the experimental phase diagram. And if the fit is good then you get a confidence on
the thermodynamics and then assume that it I can extrapolate it towards the tungsten side
lot of people do this the whole calphad journal is dedicated for this kind of thing.
And people do various and we are talking of simple binaries, life is not just binaries even
steel for that matter very very rarely you use a binary steel, which is simply iron carbon.
We have so many elements in steels; you go to a stainless steel. You have additional
elements, you talk about a super alloy you have so many elements there, Ni-based super
alloy. So, then how do I do experiments is not easy to find out the phase diagram for a
four component system doing experimentally. If you want to take every one percent of
the element and do a thermal analysis, you can imagine your whole life will go for that
that is when multi component systems particularly they try to do experimentally and
thermodynamically, we will see it as we go along.
So, this is what we talk about, what is called the boiling point (Tb) (see graph above). So,
because the free energy curve for the gas will be different. This is I can call it as a gas
and I can talk about wherever this gas intersects that of the liquid, I can call this as Tb.
Now, if you carefully observe there is one more interesting intersection here, what is
other intersection where the gas is intersecting the solid. In principle this tells me that if I
can take a solid keep on heating it and this solid does not become a liquid for some
reason, that solid, in principle the moment you reach their beyond that temperature the
solid will have a higher free energy than that of the gas. So, that solid should
catastrophically start becoming a gas sublimate.
Some for some materials this sublimation temperature can be below the boiling below
the melting point. So, that’s the solid directly sublimate without even becoming a liquid
there are many case like that. So, where these intersections are there that is going to give
you idea of this. So, this is what is called a metastable state, metastable transformation
there are cases were things like that can happen. Particularly, when you come to nano
materials many cases you can see things like that happening we will see as we go along.
40
So, you can see that kind of so, if at all I will I have to call it, I would call it as a
sublimation temperature Ts were if solid is remaining. You may say, why should a solid
remain without melting for some reasons if I am let us say heating the solid very -very
rapidly there is no time for the liquid to nucleate. All these are hypothetical cases of
course, there is a possibility and of course, this whole phase the free energy diagram can
be affected by pressure, which is what we will see right now and then all these
temperatures can even change with pressure and that is what is what is called the
Clausius Clapeyron equation, we will try to see, how does it come to.
All of us have seen in the last class that

dG = VdP - SdT
And what is the definition of equilibrium in a single component system that we have
defined in the last class.
Student: Liquid equal to 0
Correct, the definition is

Gl = Gs or ΔG = 0
For any unary system, we call them as unary systems single component systems. We can
use this and if this is the case, if let us assume that solid is becoming a liquid then I can
write this expression for both solid and liquid. So, I can write as
41
dGs = VsdP - SsdT
dGl = VldP - SldT
And at the transformation temperature, where the solid is becoming liquid both of them
should be equal and if I equate both of them
dGs = dGl
then you would come up with an expression that
(Vs-Vl)dP = (Ss-Sl)dT
So, I can now write it as
dT/dP = ΔV/ΔS
where ΔS = Ss-Sl and ΔV = Vs-Vl and what is this dT/dP means, dT/dP is the rate of
change of transformation temperature as a function of pressure.
And in principle the ΔS can also be correlated to ΔH, we can write it as
ΔS = ΔH/Tf.
So, what it tells us is the rate of change of melting point in the, in our case it is the
melting point it can be for any transformation. It can be alpha iron transform into gamma
iron, it can be gamma iron transform into Δ iron, it can be a liquid transform into the gas.
For any such transformation the rate of change of transformation temperature with
pressure is going to be dT/dP = ΔV/ΔS
. And now depending on the sign of it and the magnitude of it, you would see the, the
sign of dT/dP can be either negative or positive or highly negative or highly positive.
For example, take up this particular case of aluminium. In case of aluminium, if I look at
ΔV = Vs-Vl for aluminium is it positive or negative? At the melting point, if I look at
what is the volume of the solid and what is the volume of the liquid.
Student: Negative
It is negative. In fact that is why, if you carefully observe, we would have, you would
have seen, if I plot V verses T, you would see a step like this (Graph 1), have you seen
things like that and that is what, in fact all those foundry people talk about shrinkage,
when a liquid becomes a solid, there is a shrinkage, because this is liquid, this is solid
and this difference is the ΔV and in fact, we give what is called a shrinkage allowance
for the patterns excepting in few cases, where we know that the there is no shrinkage or
there is
42
even a possibility of an expansion.
Student: Grey cast iron
Grey cast iron, why in a grey cast iron there is no shrinkage.
Student: The Graphite has a lower density
Yes, graphite has a lower density in a grey cast iron, when a liquid is solidifying, you get
gamma plus graphite during the solidification and graphite has a much lower density
than that of the iron. As a result it occupies more volume. So, there is a volume
expansion in fact, depending on how much of graphite is coming, you may have a 0 a
lower shrinkage or 0 shrinkage or even an expansion, depending on what is your
composition of the cast iron. And how much of graphite is coming one can in principle
see even an expansion that is why, we even some cases of grey cast iron, we give what is
called a negative allowance, when pattern allowances, you would have heard about it.
So, this is what is it as far as ΔV is concerned.
Similarly, you can also plot H verses T and see a similar thing (Graph 2), this (upper
line) is for the liquid, this (lower portion) is for the solid and what we call this difference
(ΔH)? What is that it, latent heat of fusion that is why, whenever liquid is transform into
a solid. Solid at that temperature, at the melting point cannot take all the enthalpy that is
contained in the liquid and liquid has an excess enthalpy. This excess enthalpy has to be
given out then only the liquid can transform into a solid, otherwise liquid cannot
transform into a solid. Unless you remove that excess heat and that is the reason why,
when you pour a liquid metal into a mould, it’s the mould, is the one which extracts all
this heat and that is the reason why, if you carefully observe, if I draw this T verses t
(graph 3), you would have seen this plots, why does, why do you see this.
It is because there is certain Δ H associated with a liquid to solid transformation, when

liquid is becoming a solid and this (horizontal portion, graph 3) where liquid transform
to solid. And there is certain ΔH associated with this transformation and depending on
the rate at which the mould is extracting the heat. It needs certain amount of time, you
can simply see that, if I want to know, what is the Δt, time required for the solidification
that can be found out by,
43
Δt = ΔH/(dH/dt)
If I know the heat extraction rate (dH/dt) that the mould has and which basically depends
on the conductivity of the mould. I can find out, what is the heat extraction rate.
44
In thermodynamics always capital T is temperature, small t is the time. So, in principle
in any solidification process, what is the time of solidification, when somebody talks
about it, the time of solidification depends on two parameters, what are the two
parameters? One is heat of fusion, second is the rate of heat extraction. So, if I pour the
same liquid metal into a sand mould or a permanent metallic mould. You would see the
solidification time comes down, when I pour it into a metallic mould basically because I
am increasing the denominator; I am not changing the numerator because the same metal
is being solidified because I am increasing the numerator it changes (in above equation).
So, you can see this happening.
So, we know that for aluminium ΔV is negative. What about ΔS, what is ΔS in this
particular case?
Student: Change in entropy
Change in entropy, from where to where?

Student: Solid to liquid
So, it is Ss - S l, we have already defined it. So, what it is?
Student: Negative
Negative, it cannot be positive, so, it is negative and because both are negative, this (dT/
dP) has to be positive, the ratio has to be
45
positive that what does it tell you. It tells you that if I increase the pressure, the melting
point of aluminium will increase; if I decrease the pressure, the melting point of
aluminium will decrease; that means the same aluminium instead of melting it in at
atmospheric pressure. If I melt it at 10 -6 Torr vacuum, it should melt at a lower
temperature than six sixty unless I know this Clausius Clapeyron equation. I would not
be able to know that, this is a very- very crucial thing.
And all those people, who do what is called secondary steel making, they should know
this; otherwise they will not know at what temperature they should heat the steel to melt
all the secondary steel making is vacuum refinement. So, all those people, who do
vacuum refinement of any metals and alloys. They should know, what is the rate of
change of the melting point with pressure so, that at the particular vacuum level that they
are maintaining, what is the melting temperature to which; otherwise either if you
assume that the melting point is same as the atmospheric pressure. You are unnecessarily
wasting energy because you know pretty well that it would be lower excepting in some
cases, where it will not be lower that is one case can be just now, we have seen grey cast
iron, where the ΔV will be positive, but ΔS is negative. So, the ratio is negative; that
means, in case of grey cast iron so, there is a difference between a melting a white cast
iron and a grey cast iron.
If I am taking a grey cast iron and melting it, the rate of change of melting point of grey
cast iron with pressure is different and it is negative there, whereas in case of white cast
iron, the rate of change of melting point with pressure is positive there. The melting
point will increase with increasing temperature, whereas here it decreases with
increasing pressure or it increases with decreasing pressure. So, if I melt a grey cast iron
in vacuum, its melting point will be higher than its usual melting point whereas, if I melt
a white cast iron in vacuum, its melting point will be lower than the usual melting point.
Similar thing, we also see in case of water and a few other metals. For example, few
metals such as you know antimony, bismuth they all have ΔV being negative solid to
liquid, whenever I say ΔV, you need to know, which minus what.
And in thermodynamics, we always take the product minus the reactant, whenever we
are talking of any ΔV, any change in any property, any thermodynamic property it is
product minus the reactant. So, you need to know, which direction is the reaction taking
place and the product minus reactant. So, in our case, let us say, you are talking of
46
liquid to solid transformation then it is Vs - Vl, if it is solid to liquid transformation, it is
Vl – Vs, but because we are consistent in our direction, whatever you take here, if you
take the same thing here, there is no problem even for example, here you interchange
here (in numerator) and also interchange here (denominator, there is no problem the sign
is going be the same.
So, you will see that this is very crucial and if you want to see from another point of
view, whenever I take a solid and start heating it under pressure. Let us say, we know
pretty well that pressure and volume are always inversely proportional. So, under
pressure always the phase that is more stable is the one, which occupies low volume.
Thermodynamically, whenever I apply a pressure the phase that occupies low volume,
less volume is more stable, what among the liquid and solid in case of aluminium, I am if
I am considering aluminium, what has low volume? It is a solid, so solid is more stable,
it wants to be stable up to higher and higher temperatures as I put more and more
pressure and what does it mean, it means the melting point increases and that is why, if
you look at this curve once again.
I can, if I take this as liquid; this as solid, when I apply pressure on the solid, you would
see the solid becomes more stable under pressure. So, the free energy of the solid
decreases and when the free energy of the solid decreases, imagine it is decreasing
(dashed line in above graph). This (bold line) is at atmospheric pressure, this (dashed
line) is let us say at high pressure two atmosphere, three
47
atmosphere whatever it is and what does it influence, you will see the intersection
temperature has changed (moves to the right), what has happened? The melting point has
increased, when I apply a pressure. So, you can see, you can use free energy diagrams to
explain all these things.
And as we go along, we will see in this course that, we can use free energy to explain
every phenomena that we see in, in our material science and we will see and the same
thing holds good, if you take a graphite and apply pressure, why do you see that diamond
comes is just because of that. And diamond occupies less volume when compare to
graphite. Its more closed pack structure; diamond cubic is much more closed pack
structure than graphite. And that is the reason why, under pressure diamond is more
stable and that is why you find diamond that deep inside of course temperature is also
important, high temperature, high pressure in fact, people have done experimentally.
There are many- many physicists, who have developed high pressure die. And they have
taken materials and applied very high pressures and pressures of the order of few giga
Pascal and things like that. And could show that the structures of materials change and
we can in principle talk in terms of a phase diagram where you have both temperature
and pressure. And if I want to see such a phase diagram the simplest phase diagram that
we all know about is that of water.
48
Let us look at water phase diagram, all of you might have seen it, but let us look at it
from this thermodynamic angle right now. For example, if I plot temperature verses
pressure, I can show this is solid, this is liquid, this is gas (see the graph above). And,
these lines that you are seeing are the transition temperatures from one state to the other
at various pressures. For example, if I look at this line (T s), what does this line tell you?
This line is where the solid is transforming gas; that means, it is a sublimation
temperature Ts. How does Ts change with pressure? You can see from here as I increase
the pressure, Ts is increasing why because if I again go back
dT/dP = ΔV/ΔS
And for a solid to gas transformation, ΔV is positive; ΔS is positive, so the whole ratio is
positive, so the slope has to be positive.
Similarly, if I look at this line (Tb), it is a liquid going to gas; that means, it is a boiling
point and for the same, I can say liquid going to gas. Again you would see that the dT/dP
has to be positive, that is why you see a positive curve, excepting this one single line
(Tm), which is negative, which has a negative slope, the reason is when a solid is going to
a liquid ΔV is negative, whereas ΔS is positive for water, this is for water we are doing
or I would not call it as water, I will simply call it as H 2O, because water is one state of
matter of H2O. So, if I am talking of the phase diagram of H2O it is, it looks like this.
And in principle, I can draw a phase diagrams of every metal, every single component.
This is the phase diagram of single component, by the way, you may ask me, sir why I
should call H2O as single component, there is H and O there, there are two components
there, but why should I call it as a single component phase diagram?
Student: Single phase
Yes, when it is melting or solidifying or evaporating, it does all these as a single entity
H2O does not split into H and O, when a ice is melting, it does not split into H and O and
then melt, it melts as a single entity. This is what you would see even when we come to
normal phase diagrams also many inter metallic compounds, when they are melt, they
melt as one single compound. For example, if I take as Fe 3C, Fe3C actually melts as a
single compound, that is the reason why phase rule, when we are doing P+F = C+2, we
will try to see it in a minute. And this phase rule has to be looked at it carefully, because
there, there is a term called C comes, number of components
49
comes, this number of components you have to understand it carefully, otherwise you get
in to troubles in terms of the phase rule. So, we will try to see that in a minute now.
Student: In liquid to gas transition
Liquid to gas transition
Student: dP is negative?
dP? Why we are talking of dP? We are not talking of dP negative, what we are talking is,
if I change pressure, if I keep on increasing pressure, the pressure is increasing in this
direction (along X-axis), whether the melting point would decrease or increase, or rather
the transformation temperature would decrease or increase is what we are looking at and
that, depends on this two isn’t. So, what I am saying is, if I increase the pressure,
whether the liquid to gas transformation temperature, which is the T b boiling point, this
is the Ts, this is the Tm, whether the Tm would decrease or increase here you can see T m is
decreasing with increasing pressure.
This is the temperature, which is increasing in this direction (Y-axis). This is the
transformation temperature, I would call it as Tt, this is not simply temperature, it is the
transformation temperature, we are talking about, the temperature at which one phase is
changing to another phase, that is what we are plotting here, if you look at it then you
would see that the, Yes.
Student: FCC or BCC which phase will form
Which, which phase will be there? Yes, it is a question of you know, I do not know
whether you got worried about this any time, whenever you take a metal like aluminium.
I think day before yesterday, somebody was asking I was talking about it, if you are ever
bothered about, why should a liquid when it becomes a solid aluminium know that it
should become an FCC aluminium. And, if I take a iron, iron when it becomes a solid, it
becomes a BCC iron. And, if I take a zinc, the moment I melt zinc and solidifies zinc, it
becomes a HCP zinc, how does this liquid know? Liquid is random atoms, atoms are
randomly arranged such a random arrangements of atoms suddenly at the melting point
changes into an ordered structure, ordered arrangements of atoms.
50
And then that ordered arrangement of atoms has a fixed crystal structure and this
structure changes from material to material. And how does this liquid come to know
about it and that boils down to a problem in thermodynamics, what we call it as, what is
the ground state for a particular material at any given temperature? The ground state for
a particular material ground structure are; that means, the stable structure is the one,
which is what you will obtain, when a liquid transform to a solid, but this structure, what
is the structure, again is going to depend on the electronic configuration.
So, in principle people have done, if you go to a book by name Kaufman and Bernstein
not very sure about the spelling here, but Kaufman is very famous. Look for Kaufman,
the title of the book is calculation of phase diagrams and if you look for in that book, the
first chapter is dedicated to this, what we call first principle calculations of, what should
be the structure of a particular metal at a given temperature. For example, if I try to
calculate, what should be the free energy of FCC aluminium, HCP aluminium, BCC
aluminium as a function of temperature, if I do this calculation knowing, how the
electronic configuration in aluminium is though, it is easy to say, but it is more difficult
to calculate because all first principle calculations for more and more number of
electrons if you have the, the calculations become more complicate because you have to
take multi electron interactions.
People are able to do it for hydrogen, helium, but beyond that, the calculations are more
and more difficult. Aluminium if you know the atomic number we know, so the number
of electrons we know, so each electron interacts with other electron. So, you have to
consider that, but if I can calculate this, I would be in a position to find out, what should
be the structure, what should be the free energy of different structures. And look at those
free energies, you find out that for aluminium FCC structure would have the lowest free
energy, it is ultimately the free energy, the only way is how do I calculate the free energy
for a particular metal.
As long as, I can calculate it, you would see that it is, what happens then the question is,
how does this liquid know Yes, the FCC has the lowest free energy for aluminium, why
should this liquid become FCCaluminium then? It has a low free energy, but inside
liquid there is no FCCstructure, how does suddenly at melting point f c c structure comes
in, that is where you need to understand that liquid always has what is called a short
range order. This is different from, how do I differentiate a solid from a liquid and a gas,
this is what
51
differentiates. A solid has a long range order, a liquid has a short range order and a gas
does not have.
In a gas, you do not see a short range order, clusters. In a liquid you always have
clusters, within a cluster you have a periodicity, you have a unit cell and these clusters
are of different -different sizes, you have different sizes clusters, this is what we call
them as embryos, inside the liquid. And then one of the cluster attain that size of a
nucleus at the melting point, that nucleus starts growing. And all other clusters are still
there inside the liquid, but they do not grow only that, which has a size of the nucleus, is
the one which starts growing. So, in every liquid, in fact one can even calculate thermo
dynamically, what is the size of that cluster as the function of temperature, all this you
must have done in a nucleation kinetics. Yes?
Student: When I say gas, there is no orderness?
When I say no orderliness molecules are there, I am talking of in case of metals let us
say there is single atom. So, there are only single atom molecules inside the gas of
aluminium and there is no clusters at all that, I can talk about as a as a particular unit
cell.
Student: So the gasesous aluminium is sublimated to solid aluminium
Correct.
Student: So can I come from FCC to BCC Aluminium
In principle, there are cases where people have done, when people have done metastable
transitions, when you quench a gas, there is a possibility that you, you over take one of
the stable state, possibly whenever some of the stable structures, if they are very
complicated structures. For example, there are cases, which as we go along we will see,
what are called metastable phase diagrams. For example, a simple thing like Fe 3C,
which we have seen last time, why does Fe3C form, why not graphite form because this
splitting takes a long time. So, the system prefers to that and there are cases, where some
of the intermetallic compounds have a complicated crystal structure. And as a result, if I
take a liquid and quench this liquid, this liquid instead of forming that inter -metallic
compound can form some other simple structure or can even form a glass, that is what
we see a glass formation, we will see as we go along. So, this
52
kind of metastable states are possible and I will show you a few metastable phase
diagrams, where instead of a stable state, which is available in the phase diagram
something else comes out.
And that is what, I have shown you earlier also in case you can stabilize solid beyond the
melting point, it would become a gas and this gas is now forming at a temperature below
the boiling point. In principle below the boiling point, it should not have gas. So, that is
what you would start seeing. So, this is possible and if you go to that, one can find out
that, so in principle one can talk about, what are the states that are possible and if I
continue here and look at, we have now seen a slope.
We have only talked about the, the positive or negative of the slope only, we are not
talked about what is the absolute magnitude of the slope. Absolute magnitude depends
on ΔV/Δ S, as long as, I know what is the ΔV; what is the ΔS at different pressures then;
that means, I should have a knowledge of how V of solid changes with pressure, how V
of liquid changes with pressure. And, if I know that, in principles at different pressures,
if I know what is ΔV and what is ΔS. In principle, I can calculate that, so this though I
am plotting it as a straight line, it may not be a straight line. And similarly, when we
draw these two as straight lines, they may not be straight lines. In principle, V as a
function of temperature might increase not as a straight line.
53
Student: How do we understand the triple point?
Yes that is what, we are going to come to possibly today, we will not discuss, if you
know the phase rule. So next class, we will talk about the phase rule, where we will be
able to understand this particular triple point because at this triple point, you would see
the three phases are coexisting and when the three phases are coexisting, you will see
that, what is called the degrees of freedom becomes 0. So, this is fine for a single
component system.
Now, what happens, if I take a binary system, let us say this is for aluminium phase
diagram accepting that for aluminium, this is not the correct phase diagram, what should
be the correction that I should incorporate for aluminium.
Student: Liquid solid line should be upwards
Yes, this line should be like this. That is the only change, of course when I say that is the
only change the actual slopes in water in H2O will be different from the actual slopes in
the aluminium, but the aluminium, all the three slopes will be positive, all the three
slopes will be… So, aluminium, iron most of the metals, where you do not have that
anomaly of the ΔV, the volume of the solid being higher than the volume of the liquid.
As long as, that anomaly does not exist, this is true, we can do that.
And now the moment, I say a binary; I have to add something more there. If, I am
talking of two component systems, let say aluminium, copper. I have additional element
there, so this phase diagram, I have to add one more axis there, which is nothing but, the
percentage of copper. So, it becomes a ternary kind of diagram. A three dimensional
diagram and in that three dimensional diagram, we will start working on it, we will talk
about it in the next class, how to do such a thing.
I am stopping because it has to come to 0 there, can we continue even after it goes to 0,
no problem? We will stop. So, that is why. I will follow that, this is the problem for a
teacher, no teacher wants to stop. Yeah tell me.
Student: Specific heat, Cp , do we have different expressions at higher and lower

temperatures? Because, above debye temperature, it is 3R, but below debye temperature,
it decreases to zero.
54
Correct correct, see because the Cp as you knows below the debye temperature, at very
very low temperatures, it is the electronic configuration contribution come into picture.
55
At a higher temperature, it is more of atomic vibrations, very low temperature you know
closer to 0 degree 0 Kelvin actually, we do not actually have much of atomic vibrations,
but this is the electrons inside the atoms, they tend to vibrate and that is the reason why,
you get but that value is too small. And as a result, Cp is very small, but the rate of
change of Cp with temperature will be different, because those vibrations are going to be
different from that of the atoms.
So, there are you know different -different expressions that are available for the C p at
low temperature, at higher temperatures, one can calculate it, they are available. You
have to go to the standard handbooks, and for definite materials that are, but most of the
cases because we deal with metals, above room temperature. We do not actually talk
about that contribution at all, that contribution is insignificantly small when compare to
atomic vibrations, at temperatures above the atmosphere, room temperature. And
excepting physicist, a typical metallurgist or a mechanical engineer or a chemical
engineer is not really concerned about such low temperatures, we will stop.
56
Lecture No. #04

Free energy of solutions, free energy-composition diagrams
In the last class we talked about what is called the equilibrium for single component
systems, we talked even about a phase diagram for a single component system, such as
H2O. Today we will see what happens if I add another component to it. How does the
free energy, and concepts of equilibrium start changing? For example if you remember,
we have seen a diagram of G vs T (see graph above), and then drawn the free energy
curves for any two phases, I call it as phase one, phase two, and then I say based on this
in principle, I can say which phase is stable under what temperatures.
And in addition actually in principle, this is not sufficient; we should have actually added
one more axis there, which is for pressure. In principle, this is a three-dimensional axis,
three-dimensional diagram with a pressure axis going like this, and I should have a G as
a function of P and T, that is what you have seen G equal
dG = VdP-SdT, and
57
from that you should be able to find out the free energy at a particular temperature and
pressure for different, different phases and based on that see which has a lower free
energy and decide, what is more stable at a particular condition. But because most of our
processes or metallurgical processes in general, and any chemical processes are done
usually at atmospheric pressure, what we do is in that three-dimensional diagram.
We take one section at a particular pressure, which is atmospheric pressure. And then
now look at a 2-D diagram, which is the G verses T diagram and then we are
representing the whole free energy in that. So, at the back of our mind, we should always
keep, that actually it is a 3-D diagram, and I have taken a2- D section out of the 3-D
diagram, at a fixed pressure. In principle, if I change the pressure the whole thing
changes, this particular temperature would shift this way or that way, the free energy
curves may go up or down, depending on what pressure it is and that is what we have
actually seen based on the Clausius Clapeyron equation in the last class, that if I change
the pressure, the melting point will change, increase or decrease depends on the
particular condition the particular system, whether the ΔV is positive or negative based
on that. We will know, whether it will increase or decrease, but we definitely know, that
the melting point is never the same, when will it be the same, can you say somebody tell
me, when will the melting point not change as a function of pressure?
Student: When volume doesn’t change.
Yes perfect, if there is a system, where ΔV is 0. If there is a hypothetical system, where a

solid becomes a liquid, where there is no change in the volume. In such a case ΔV is 0,
then dT/dP = ΔV/ ΔS = 0, then you can say that melting point never changes with
pressure. You keep on increasing the pressure or decrease the pressure, the melting point
would remain the same, only under that special condition. Otherwise in principle, you
would always see, it would either increase or decrease depending on the sign of the right
hand side, whether it is positive or negative, yes.
Now, what we see is, this is ok for a pure, pure metal system or a single component
system. Again I should emphasize again and again to you that, when I say pure metal
system, even H2O, we take it as a single component system, so instead of saying pure
metal system, I should say single component system. And as we go along, when we
58
go to binary phase diagrams, you start seeing some of the intermetallic compounds also
we start dealing with them as a single component, when I come to the phase rule. You
will see that, how we deal intermetallics as single component system. And if you do not
deal with them as single component, you get into trouble in terms of the phase rule; you
will see that as we go along.
So now, looking at a single component system this whole thing is fine, I simply need to
think of a 3D diagram of G verses T and P, and then look at different- different sections
depending on, what pressure I am working with and then get to know, how the free
energy curve is and then find out. In principle in a G-T-P diagram, what I am seeing is
not actually the curves but surfaces, isn’t it. G as the function of P and T surface and any
surface, if you in take a section, the section of a surface, when a plane intersects the
surface, what you see is a line. And that is what we are seeing now; all these lines that
you are seeing are intersections of this plane. So, there is a G verses T and P plane for
the phase one and for the phase two, and in that plane I am taking sections.
Now, what I am saying is I am introducing something else additional, I am saying there

is a second component now, if I say there is a second component, whenever I say a two
component system, in this two component system, only one of the component is actually
a variable. For example, if I am talking of steel iron carbon steel, I say it has one percent
carbon then I need not have to say that, it has ninety nine percent of iron. As long as, it is
a binary system, so only one of the variable, one of the element is a variable, the second
one is a dependant variable, one is independent variable. So, that independent variable
the composition also, I have to add and then what I will see is that, this particular
function is also a function of X. X is the composition or C different- different books you
will see them. So, I would treat it as let us say X B, B is my solute let us say, solute and
solvent again change depending on which end you are in, if you are on the A side B
becomes solute; if you are on the B side A becomes solute.
So, I can say that G now, has to be considered as a function of G = (P, T, X B). So, it is a
four-dimensional diagram actually, and as you know it is not easy to visualize,
mathematically one can represent, I do not know whether you have ever heard of what
are called quasicrystals and in quasicrystals people know many of you know when you
59
are talking about miller indices, what are the number of indices that we use for typical
systems like cubic?
Student: 3
Three, but the moment you go to a hexagonal you go to four, and the moment you go to
a quasicrystal, you go talk about six, and then people actually believe that a quasicrystal
is a crystal in a six-dimensional space and when you project that, six-dimensional solid
into a three-dimensional real space, it actually does not retain its periodicity and it
becomes a quasi periodic. So, people actually do a lot of mathematics in quasicrystals,
the whole crystallography in six dimensions imagining. So that, here it is a similar
situation, so because one can do all the mathematics, one can calculate how G varies
with P, T and XB, but the thing is when you want to represent it; obviously for the sake
of representation, we need to take sections. We have already taken one section, which is
the section of pressure, constant pressure and now, now because you have taken the
section of constant pressure.
Now, you take another section, you add another axis here, which is the X B axis now;
that means, composition axis and in that composition axis, I what I do is, I now take
constant pressure free energy curves, where I show G verses X curves (Graph 2). So, I
will have a plot now, where I have G verses X, where one side is pure A; another side is
pure B. And what I am looking at is a constant temperature, pressure is already kept
constant, now I am looking at constant temperature, and at that constant temperature I
see how the phases are stable in which composition range, which phase is stable.
At that temperature and if I do many -many sections at different- different temperatures,

I basically come up with a phase diagram, which is actually for us, what is called T-X
diagram (Graph 3). A phase diagram is nothing but a T X diagram for us in a binary
case, where A will be on one side; B is on other side. And in this diagram what I am
trying to show is, at each temperature, which phase is stable in which composition
domain, and where from that I get that information, I get that information from here
(Graph 2). So, if I draw various sections (in Graph 3), various horizontal lines, at these
horizontal lines I can say from point P to Q phase 1 is more stable; from Q to R, phase 2
is more stable and from R to S phase 3 is stable whatever it is. And if I get that
information from here and I put all that in this diagram, and I can now say that what are
the domains for the phase
60
stability, which phase is stable in which temperature and composition domain, keeping
again at the back of mind that I have already kept the pressure constant. And the
moment, I bring a pressure into it, you would see that it would be a T verses X and
pressure. So, again here also, I am talking of a pressure being taken constant.
So now, we try to look at, how the free energy would change with X at a different
pressures, pressure is already kept constant as a function of temperature, how does it
change. To know that, we have already done earlier, what is G for a pure metal for a
single component system? We said
G = H - TS, so for any pure metal at any given temperature, I can calculate the G, what
is the only input that I note C p, as long as I know how, what is the C p and how Cp
changes as a function of temperature, I can calculate the G. So, in principle because this
diagram (Graph 2) is at a fixed temperature, so if I am talking of free energy of pure A,
what would it B, where would it I represent free energy of pure A in this diagram (Graph
2)?
Where, where should we represent free energy of pure A in this G-X diagram, will it be a
curve? It will be a point and where? On the on this pure A line, it will be a simply one
point somewhere (GA). At a given temperature A has a fixed free energy, once I change
the temperature, this free energy will go up or down, if I increase the temperature will
this free energy (GA) go up or go come down, can you tell me. If I increase the
temperature will this free energy point will go up or come down?
Student: come down
Come down, because we know G is this, because G = H-TS and dG/dT is always
negative. So, the free energy should come down as I increase the temperature, this is for
pure A. Similarly, if I do it for pure B, I will see another point somewhere here, this is I
call it as GB; I call this as GA.
Now if I am talking of an alloy, a phase which consist of A and B, then how do I talk
about the free energy of this particular phase, which has both A and B inside it. One way
to think about it is that this phase is nothing but just a mixture of A and B. If I think, it is
mixture of A and B, the free energy of such a mixture can always be represented by
simply drawing a line joining the two. All of you must have done, what are called rule of
mixtures, many of us do in composites, whenever you want to know the strength of a
composite or young’s modulus of a composite, where two phases are there, we use rule
61
of mixtures, volume fraction of one phase multiplied by the modulus of the other phase.
Of course, you may say whether it is plane strain condition or plane stress condition. So,
the formula will slightly change but otherwise we know, what is called a rule of mixture.
So, if I think, it say a mixture and having certain composition XB, let us say, when I say
XB then XA = 1 - XB, it is automatic and so, for that composition I can say, if I have
mixture, that mixture should have a free energy of that, but in principle this is not always
the case, because A and B atoms cannot be simply considered as a mixture. There is
always an interaction between A and B, even if there is no interaction, the moment I take
A and B, it is like red balls and white balls, I put them into a box. There is an additional
entropy that comes into the system, because of the way I arrange A and B atoms, this is
what we call it as a configurational entropy. So, far we talked of one entropy here, this is
actually a thermal entropy, which is coming only because of atomic vibrations.
Now, the moment I go to a second element or a second component, I need to bring in

configurational entropy. So, even if A and B atoms are not interacting with each other,
they don’t neither attract each other nor repel each other, they are neutral to each other.
Even in spite of that, there will be an additional free energy term that comes into picture
because of simple configurational entropy, that configurational entropy term has to be
added to this mixture to be able to know, what is the actual free energy of such an alloy
and this configuration entropy, you all must have gone through statistical
thermodynamics at some stage.
62
We define it as ΔSconf = klnW. what is the difference between k and K?
Student: small k is Boltzmann constant
Capital?
Student: Kelvin temperature
Capital K is the Kelvin in temperature. So, whenever you use capital K in
thermodynamics it is always fixed, it is a temperature Kelvin. So, you should never use
that, so this is the Boltzmann constant. So, where w is what?
Student: Number of configurations
Number of configurations, number of ways of arranging a B atoms in a mixture of A and

B atoms let us say, and this if you derive it, it turns out to be
∆ Sconf =−R( X A ln X A + X B ln X B )
And in principle, this can be extended to any number of components, where we write it
n
as ∆ Sconf =−R ∑ X i ln X i
i=1
In fact that is where possibly, if the time permits I will talk as we go along some time,
what are called a multi component systems that people are trying now, where they use
this configurational entropy as a very important tool, to generate new type of materials
which we call them as high entropy alloys.
63
If you look at this particular equation, this equation has a very, very systematic way of
change of configurational entropy as a function of composition. You see that it is a very
64
symmetric function and because it is symmetric function and by the way, look at that and
tell me, what would be the sign of it, what would be the sign of configurational entropy?
Student: negative
It is negative?
XA and XB are fractions, ln of a fraction is negative , ln of a fraction is negative, so and

there is a minus there. So, configurational in fact, this is a very simple thing that you
should know that, when I put A and B atoms I am creating randomness, and randomness
is always positive; there can never be a negative randomness. I can only say randomness
is not there, when randomness is not there, we say it is 0, but you can never say there is a
negative randomness there is either perfection or randomness. So, perfection is where we
say randomness is 0 that is where we say the configurational entropy goes to 0. So,
where does the configurational entropy goes to 0? Either on this side or on that side, for
both the pure metal sides configurational entropy should go to 0, and that is why you see
configurational entropy term does not come into picture, when I am talking of a pure
metal. Only when I am talking of an alloy, I talk about configurational entropy;
otherwise I would not talk about configurational entropy.
And now, this configurational entropy also you would see, it would it is positive first and
foremost, and it is 0 on both the sides and it is maximum at the center of the phase
diagram. It is very easy also to understand, when I have 10 atoms of B and 90 atoms of
A, the number of possible configurations. If I compare this with 30 atoms of B and 70
atoms of A, you will have more possible combinations of how to arrange B atoms, if you
have more and more number of B atoms and this would; obviously, go to maximum at
50 -50, when I am coming to other side then automatically A becomes a solute and B
becomes a solvent.
And this is such that, it is always symmetric, whether I am taking 70-30 case or 30-70
case configurational entropy is the same, and also one more interesting important point
that you should note is that configurational entropy does not bother about which
elements that I am considering, whether it is a steel, iron carbon steel or
65
whether it is a aluminium copper alloy, configurational entropy is always the same,
provided the amount of the elements the percentage of elements are the same. If I am
taking iron with one percent carbon alloy and aluminium with one percent copper alloy
both of them will have the same configurational entropy.
Student: Does pure metal exists?
Yes, very interesting question, thermodynamically no, when we do this, after doing
configurational entropy, after doing the free energy function, we will prove this for you,
just may be at the end of this class; you all remind me or may be in the tomorrow’s class.
So, once I put the free energy function then we will be able to prove it, we will prove
that it is not possible, thermodynamically it is not possible, only thing is you can, you
can keep increasing the purity. We can we can talk about 4N purity, 5N purity have you
heard of these things, when I say 4N purity, what does it mean?
No never heard, what is this N?
Student: 4-nines
Four nines, if when I say four N purity it is 99.99%. I have recently attended a
conference where they were talking about a copper which they made 7N purity copper.
So, 99.99999, so and for certain applications for example, if you want to have a
exceptionally high conductivity, you know conductivity is one thing, where the any
impurities, I do not know whether you have heard of the, all the conducting copper wires
that we use have a special name anybody knows?
What, what copper is this, it is a special name.
Student: Electrical
Electrical
Student:
OFHC
66
Yes, they are called OFHC copper, oxygen free high conductivity copper. So, they want
to eliminate oxygen as much as possible, because oxygen is not suppose to be good for
the conductivity any way, we will come back to that later.
So, what we are now looking at, is now the free energy of any phase α, which has two
elements in it, can be represented as a combination of the two pure metals components
plus an additional term, which comes because of two things one either A and B have an
interaction between each other, that they either like each other or do not like each other,
they attract or repel each other, even if that part does not arise that they are neutral to
each other. You have an additional term called the configurational entropy, and that
particular thing is what is called ΔGmix = ΔHmix - TΔSmix or the same ΔSmix is also called it
as called as ΔSconf.
So, I can write the free energy expression as

Gα = XAGA + XBGB + ΔGmix
This is the free energy of any n component system, I mean here I have done it for two
component system; you can extend it to any n component system, excepting that these
initial terms. So, these (first two terms) are pure metal terms this is simply a mixture,
pure metal mixture, and I am adding a further term, which is
ΔGmix = ΔHmix - TΔSmix
67
this ΔHmix can be 0, negative or positive depending on whether there is an attraction
between A and B or repulsion between A and B or in ideal solutions it is 0. All ideal
solutions ΔHmix
68
is 0; that means, the A neither gets attracted to B nor gets repelled from B, but in most of
the cases it is not like that.
So whereas, so in case of an ideal solution, I can write this expression as

Gαid = XAGA + XBGB - TΔSmix = XAGA + XBGB + RT(XAlnXA + XBlnXB)
This is the expression for an ideal solution.
And that, if I want to represent it on the on the free energy curve, I can show it as
something like this (Graph 1), because if I look at the entropy the entropy is always
positive like this (Graph 2). This is ΔSconfig, this is 0; this is pure A; this is pure B and at
50 -50 it is maximum. And because we are going to be talking about -TΔS mix, , I have to
multiply this with T. So, if I draw now, what is -TΔS mix, so this would be something like
this (Graph 2) of course, you have to remember that this scaling (upper portion) will not
be same as this scaling (lower portion), because here, I am multiplying with T, so;
obviously, this value is much larger negative.
And the higher the temperature the more this is, but because I am simply taking this
configurational entropy and multiplying with some constant number called temperature,
the nature of the curve does not change, it still remains symmetric excepting that from
the plus side it has come to the negative side. And now, this side if I take and add
69
it to this (Graph 1), what I get is the G of any phase called α as a function of
composition. At a given temperature, this is for a given temperature; the moment I let us
say again the same ideal solution and I keep on increasing the temperature, what should
happen? This curve should become more and more negative and anything else should
happen.
Student: G will also decrease
Yes, GA and GB also will decrease, where this curve intersects on the two, two Y axis on
the pure A and pure B side also, it is also going to change, because these two quantities
(GA and GB) are going to change as a function of temperature clear. So, this is how, we
can talk about any free energy curve and once I have that kind of a curve, I should be
able to find out whether now, if I try to talk about two different phases, and talk about
equilibrium, we can start talking about equilibrium provided, I can calculate this, but this
is very simple for an ideal solution and life is not really ideal isn’t it. So, I should talk in
terms of non ideal solutions and the moment I say non ideal, I have to bring in a term
called ΔHmix.
And that term, if I add to free energy, I get,

Gα = XAGA + XBGB + RT(XAlnXA + XBlnXB) + ΔHmix
And this ΔHmix can be positive or can be negative depending on whether there is
attraction or repulsion. And now, looking at this people have come up with different
70
models. One of the simplistic model, people have talked about is what is called regular
solution model, without this term it is a ideal
71
solution. So, what is a regular solution? In a regular solution, we define
ΔHmix = ΩXAXB.
And this Ω is called interaction parameter. So, depending on how strong or how weak is
the interaction, if this Ω goes to 0, what does it mean, we are talking of a ideal solution
again. If the interaction is zero; that means, there is no interaction between A and B; that
means, it is an ideal solution. So, depending on how strong or how weak is the
interaction we can talk about ΔHmix. And now, this Ω can be positive or negative,
whenever Ω is positive this is a repulsion and that means, ΔH mix is positive and
whenever Ω is negative, it is an attraction and ΔHmix is negative.
And usually you see this kind of a second case, more and more in case of inter metallic
compounds, whenever there are compounds. In all the cases of whenever you see a phase
diagram with compounds; that means, in that phase diagram A and B have an attraction
like cementite Fe3C or Ni3Al, NiAl or Mg2Si in a magnesium silicon phase diagram. All
these are indications that there is a strong attraction. But there are also phase diagrams,
which are something like about let us say copper gold or any other let us say silicon
germanium, wherever you see an isomorphous phase diagram, where there are no
compounds, that kind of a situation is very close to an ideal situation, where there is
neither attraction nor repulsion.
So, the moment you see attraction between A and B, there is a tendency for the
formation of inter metallic compounds. We will slowly start seeing, how phase diagrams
evolve based on the free energy concepts and wherever there is a repulsion, you start
seeing that there is what is called something like a eutectic or a peritectic. You start
seeing that the solubility of one of the element in the other starts getting restricted,
whenever there is a repulsion; that means, A does not want B to be dissolved into it, to a
large extent its like we have, we are selective in our friends, we want somebody, we like
somebody, we do not like somebody else. Similarly, A does not want B to a large extent,
so you start seeing the, the lesser and lesser solubility of one element in the other, that
you see, it gives you an idea that greater and greater repulsion between A and B.
For example, you might have seen eutectic phase diagrams, where there is some
solubility of A and B on one side, on both the sides or one side. You will also see that, as
72
we go along, we will talk about what is called sub regular solution model within a
minute, where you will see that there could be a possibility that on one side there is
solubility, on the other side there is no solubility.
For example, I give you an idea of copper zinc phase diagram, how many of you have
seen copper zinc phase diagram. Every metallurgist should have seen, because it is
supposed to be brasses, do you know anything about copper zinc phase diagram,
something about solubility of copper in zinc and zinc in copper. Do you any of you
remember? How much of zinc is soluble in copper at room temperature?
Student: 40 %
Approximately around 40 %, somewhere between 35-40, what about copper in zinc, is
less than one percent, less than one percent. How does this… I mean happen, one
element wants to dissolve into another or one element accepts the other element very
easily up to forty percent, but the other one does not accept; that means, it is like one
sided love, something like that. So, you see this, that copper accepts zinc, but zinc does
not want to accept copper.
Student: Sir, high valence atom
See this is, we explain this from a valency point of view.
All of you Hume-Rothery rules if you have gone through, we say that because copper
and zinc have different valency. We say whenever an element goes into another, because
of the metallic bond present; a metallic bond accepts an increase in the e/a ratio and not a
decrease in the e/a ratio. For example, if you take this particular case copper has a
valency of 1; zinc has a valency of 2, when I add copper to zinc. You see the total
valance electrons are three and total number of atoms are two, because one copper atom;
one zinc atom, two atoms are there, and the numbers of valance electrons are two are
already there, I am adding one more, so the total number becomes 3 and the e/a ratio is 3
by 2, which is 1.5.
So, if there was a pure zinc, it had two e/a ratio, two valance electrons for each atom two
by two by one is two, whereas the moment I have added one atom of copper to zinc, I
have seen that from 2 it has fallen to 1.5, so there is a decrease in the e by a ratio; that
means there is the decrease in the valance electrons per unit atom. And this is something
a metallic bond does not like, whereas on the other side, you see when you add zinc to
copper, two electrons are added. So, the total number of electrons become again three
one plus two becomes three and there are two atoms, it is again 3/2 = 1.5, but here this
73
time from one it has gone to 1.5. But then you may say sir, why can’t then hundred
percent of zinc atom go into copper, that is where we talk about other
74
things. If you want 100% solubility of one in the other, there are number of other Hume-
Rothery rules should be satisfied.
And the most important one among them is what?

Student: Same crystal structure
Same crystal structure. If you want 100% solubility, the first and foremost is same
crystal structure, it does not mean that if the structure is the same, that it will have
hundred percent solubility then the rest of the things come into picture.
Yes, very easy example is copper silver and in fact, copper silver has led to a entirely a
new field of metallurgy called RSP. Who knows what is RSP, rapid solidification
processing. A person by name Pol Duwez in 1959, he started worrying about, why does
copper and silver do not get mixed with each other. If you look at the three elements
copper, silver, gold. Three elements and look at the phase diagrams of three of them,
copper and gold is an isomorphous diagram; silver and gold is an isomorphous diagram
75
whereas, copper silver is not an isomorphous, it is a eutectic diagram. Silver dissolves
into copper up to approximately let us about 9% or so. And similarly on the other side, it
is a eutectic why is this so? When we look at the valency, they all have the same
valency; they all have similar electro negativity. Atomic sizes are slightly different, but
based on atomic size itself.
76
People are not able to understand this for for many- many years. This is not clear then he
thought let me do an experiment, where I melt this alloy, and try to cool this alloy very
rapidly, and see what would happen. So, what he did was, he developed a technique
called gun-quench technique (see above image). What he did was, he took this alloy in a
crucible (C) with a small hole in it, and then melted everything, and then put before the
experiment itself, put some kind of a diaphragm (d) of a higher melting metal. So, that it
does not melt, when this alloy is melting and then after everything is molten. He applied
a high pressure of argon such that this diaphragm gets pierced and the liquid comes out
of this in the like a bullet from a gun, that is why it is called gun quench technique. And
allowed this to fall on a slanted copper plate, copper having a high conductivity. He
expected that this liquid the moment it falls on it, it gets quenched very fast.
And in fact when people calculated the cooling rates that he could achieve are the order
of 106 K/s, a million second/s a very, very high. And just for a comparison, if you use a
permanent mould casting, you take a liquid metal and pour it into a cast iron mould, you
usually get around 100 or 200 ºC or maximum, if you have water cooled copper mould
may be you can get about 300, 200 ºC. So, in comparison he achieved a very high
cooling rates and what happened was each droplet of water that is falling has become
like a small flake and when he took these flakes and put it into an X ray machine to do
the X ray diffraction, what would be the difference between… you tell me, if a copper
and silver are existing as a eutectic mixture.
Let us say, I take a phase diagram all of you are aware of eutectic phase diagram (see
phase diagram in above picture). Let us say, I take this alloy (eutectic composition) at
room temperature, what should it contain?
Student: Solid solution
Of?
Student: Eutectic mixture
Eutectic mixture of what, α + β you call one side as α, and you call other side as β only
problem that comes is usually in crystallography, we use this α and β for different crystal
structures, whenever I use the term α, it basically means one crystal structure, β means
some other crystal structure, but incidentally here both of them
77
have the same crystal structure, but whatever it is, we will assume it as α + β. And if that
particular sample containing α + β and if you send X rays through it, what would be the
XRD pattern? How many of you have gone through X-ray course or a one chapter in X
ray on X-ray at some stage, X ray diffraction? So tell me, what should be the XRD
pattern look like? What should it contain XRD pattern, XRD pattern what are, what do
you see whenever, I show you a XRD pattern something here, what will you see in the
XRD pattern.
Student: Peaks
Peaks, one thing at least, some of you have seen, you see peaks provided it is a powder
diffraction pattern. Again different -different types of diffractions are there, powder
diffraction gives you peaks. Now, if I take a α + β mixture and then send an X ray, I get
peaks and what should these peaks correspond to, correspond to?
Correspond to the α and β phases separately. α should give its characteristics whatever,
h k l peaks depending on its crystal structure and β should give its, and incidentally
because both of them have the same crystal structure FCC. And again, if I probe you a
little more for FCC, what should be the type of peaks that I should see, many of you
have a B tech metallurgy background. So, you should know what peaks I should see in a
XRD pattern of an f c c metal?
Student: 1 1 1
111
Student: 2 0 0
2 0 0, what is the condition? Now, let us go to the condition.
Student: h+k+l
Is?
Student: unmixed
78
Unmixed, h k l are unmixed not h+k+l, h k l should be unmixed, what does it mean that
all three of them should be either odd or even. So, 1 1 0 is not possible, 1 0 0 is not
possible, because 1 is odd and 0 is even. So, the first peak that is possible is 1 1 1 and
then you have 2 0 0 and the next one is what?
Student: 2 2 0
2 2 0 then the next one will be 3 1 1 and 2 2 2, and so on. So, but interestingly here,
because it is a mixture of coppers based solid solution and a silver based solid solution,
what you see is two sets of FCC peaks, two 1 1 1 peaks, two 2 0 0 peaks, two 2 2 0 peaks
and so on. One corresponding to copper based solid solution, another corresponding to
silver based solid solutions, what he has seen interestingly when he did this gun
quenching technique is that instead of two sets of peaks, he simply saw one set of peaks,
giving an indication that it has become one single solid solution rather than a two phase
mixture, that was the beginning of what is called rapid solidification processing. And he
started doing this with different eutectic systems, thinking that every eutectic should
become from a two phase mixture to a single phase mixture.
And what he proved from that is inherently there is a tendency for A and B to mix with
each other, to dissolve into each other, but because of some reasons in the solid state
there is some kind of a repulsion and in the liquid state there are all soluble completely.
And, if I somehow cool it rapidly, I am able to somehow prevent this separation, and this
it is like quenching from a higher state. Unfortunately in a steels, we see when you
quench from austenite, you get a martensite different structure, but there are also cases
where, when you quench from the higher temperature the same structure remains.
Now, when he started doing it with various eutectics, in one case which is called a Au-Si
eutectic, he found that suddenly the pattern rather than showing you a pattern like this,
where I verses θ, you see peaks like this (Bottom image, (Refer Slide Time: 37:08))
usually, if it is the solid solution, this is the 1 1 1 and let us say and I have to be careful,
if it is a f c c, you would see a typical f c c will become like this, that is how we actually
differentiate an f c c from a b c c, whenever you have you are looking at an f c c, you
will have two plus one plus two kind of peaks, two peaks will be closer that is because
one is 1 1 1, another is 2 0 0. The h 2 + k2 + l2 is three and four, and the next one will be 8
then 11 and 12 that is three 1 1 1 and 2 2 0 will be closer and so on and so
79
forth. So, whereas, in a b c c, there will be equally spaced, that is the difference, so here
instead of two peaks he simply got one single peak, I mean two 1 1 1 peaks he got, in
case of a copper silver.
In case of a Au-Si, what he saw was a peak like this (Fig. 1, upper image) , that was the
first so called metallic glass in 1959 Pol Duwez has discovered. The first metallic glass
till then people knew about glasses in silicate systems, we know that, what you see in
front of you is a glass there, and people always believed that only silicate glasses,
silicates can become glasses easily. Why silicates become glasses easily, anybody
knows? When I take a liquid of silicate and I simply cool it very slowly, it becomes a
glass why?
Student: Crystal structure is very complex
Crystal structure is very complex in a, in a in a silicate solid, the crystal structure is a

network 3- dimensional network structures. They are so complex for the liquid to really
crystallize into that crystal structure, instead the liquid it would easily become a glass. In
fact, long back turn bull, there is a person, who said every liquid wants to become a
glass, unless crystallization intervenes. Because for a liquid to become glass is easier,
because the structure of glass is same as that of the liquid, whereas crystals once they
start coming from the liquid, they have a different crystal structure, when compared to
the liquid. So, they will be different, whereas here, you would see they would, it is more
difficult for the crystal to come out, whereas in simple metals like aluminium or silver or
copper, FCC kind of structures are simple structures can easily come out of liquid, that is
why, if you really want a metals like this to become glasses.
You need to cool the liquid at such a high speed, that the nucleation of this FCC crystal
is made very difficult. What is the nucleation means somehow atoms have to come
together and form a unit cells of a cluster, which is of the size of r* and then only growth
starts, if that is difficult then it does not… So, by using such high cooling rates of 10 6
K/s, he was able to suppress crystallization and get a glass. And now, there are so many
people, who have been working on this and got many many glasses. Now we are talking
even about, what are called bulk metallic glasses? People are able to take liquid and pour
it into a three inch mould, and get the whole rod three inch rod, which is a glass.
80
Again basically what they have done is they have added a large number of elements into
the alloy, and made the crystallization very difficult, when there are too many elements,
when a liquid becomes a crystal, it does not know where these atoms should sit. So, there
is a some kind of a confusion, we call this is as confusion principle. As a result, liquid
instead of becoming a crystal, it would remain as liquid as similar to a silicate glasses.
People are able to get glasses, which are very, very thick glasses by this kind of a process
what is called bulk metallic glasses. So, like that there are number of glasses available,
we will talk more about it as we go along, thank you.
81
Department of Metallurgical & Materials Engineering
Lecture #05
Solution Models, Μ
If a material has a Tg, in fact, my PhD thesis, one of my chapters was to do

thermodynamics of glasses. So, I tried to calculate what should be the glass transition
temperature for pure metals and when I calculated for some of them it turns out to be
negative in terms of absolute temperature, something like -200 K. If that is the case, then
in principle, that particular metal by some way if I can bring it to absolute 0, before it
actually crystallizes, see all this is assuming, that the liquid is not crystallizing, if liquid
crystallizes during this cooling, then it is not glass. We are talking of if you can retain it
as a super cooled liquid at absolute 0, then it would have that. The question then is how
do I measure that? So, that is why, enthalpy also, we say, enthalpy is 0 at 298 where
simply, because we do not know what is the base, I should start with, so I assume
something. So, that is why, all these standard states have come into picture. O.k., we
start.
82
Yeah, in the last class we have started talking about two component systems and we
started plotting, what are called, G-X diagrams. And we talked about a simple model
called regular solution model and also a much simpler model, called ideal solution
model. In which G of any phase, let us call it as an α, G α, where you put this α as a
subscript or a superscript, depends on, you stick to something, many people, we usually
use here because subscript is used for the component. For example, G A, GB when I say
for example, as you go along you will know, maybe today’s class itself, when I, many
times I have to write GAα and GBα, we will see what does it mean as we go along.
So, here for example, we know, that this

Gα = XAGA + XBGB + ΔGmix
ΔGmix = ΔHmix - TΔSmix
83
And as a result, for an ideal solution we have written this expression as

Gα = XAGA + XBGB + RT(XAlnXA + XBlnXB)
And we said, the moment it is not ideal, non-ideal solution, the one of the simplistic
model because all this, whenever I am talking of a free energy of solid solution, I need to
start thinking about how the A and B atoms are interacting and then give a model to it
and then do some measurements. How do I do measurements in terms of, if I want to
measure for this, how do I measure it? Is there any way?
Student: DSC
Yes.
If you do, let us say, you take A and B metals, you mix them and then do a DSC. If there
is a positive interaction or a negative interaction, you would see it in terms of what is
called heat of formation. This heat of formation, we use calorimetry and more
sophisticated thing is a DSC, differential scanning calorimetric, and if you use DSC in
principle, one can do. So, there are large number of people who have been measuring
this ΔHmix and from that we come to know, whether they are inter-metallic compound
forming or whether there is a repulsion between them because of which there is a ΔH
being positive and from that, you try to fit from the measurements to the model and say,
yes, my model of regular solution or sub-regular solution, things like that are working.
84
So, in a regular solution model we assume ΔH mix = ΩXAXB, where Ω is, what is called,
interaction parameter. So, I can now write the the free energy of a solid solution as
Gα = XAGA + XBGB + RT(XAlnXA + XBlnXB) + ΩXAXB (1)

if at all it is following a regular solution model and this boils down to ideal equation.
when Ω is 0. When Ω is 0, that means interaction between A and B is 0 and that is what
causes, what is called an isomorphous solid solution. You will see, as we go along, when
we go to the phase diagrams.
And whenever you see the Ω being positive, we say there is a repulsion, A and B do not
want to mix into each other and that is what causes, what is called miscibility gaps in
phase diagrams and these miscibility gaps slowly lead to, what are called, the eutectics,
peritectics, even to the extent of monotectics. We will see in possibly the next class,
what is called evolution of phase diagrams, how do phase diagrams evolve just based on
thermodynamics? We will see that, till then have some patience.
So, if Ω is negative, that means, there is an attraction between A and B, any attraction
always causes the formation of compounds. And here also, depending on how strong is
the attraction, you have compounds with different melting points. Some of the
compounds may have very low melting point and some of the compounds may have very
high melting points. If you look at some phase diagrams, you see, the two pure metals
have very low melting points, but the compound has extremely high melting point, that
is because the two elements have such a strong attraction, that this compound, the bonds
in the compound cannot be broken easily. Melting point is basically an indication of
bond breaking and that is why, you can easily correlate the melting point of a metal to
the bond strength within that particular metal.
85
So, you can see, so Ω negative means attraction that means formation of compounds. Ω
positive means repulsion, repulsion means, what is called immiscibility or I can call it as
miscibility gaps, which finally, ends up into phase separation.
Now, people have seen as, possibly we have touched upon in the last class, that this
cannot be always true. What I have done in this is, I have assumed that the interaction
parameter is the same throughout the phase diagram, from one end of the phase diagram
to the other end of the phase diagram, I assumed it to be the same, which is actually not
always true. When I simply look at a number of phase diagrams, I see the solubility of
one element in the other is different from two ends. We took the example, possibly in the
last class on copper zinc.
86
Zinc dissolves in copper very much, whereas copper does not dissolve in zinc and we try
to give a logic for it based on the valancy, and if that is the true, then the interaction
between A and B atoms is different on one end of the phase diagram and it is different
on the other end of the phase diagram and this has brought in, what is called sub-regular
solution model, where people say, that
ΔHmix = XAXB(XAΩA + XBΩB)
.That means, now we are talking of two interaction parameters, one on the A side, one on
the B side, that means, the interaction between A and B atoms on the A side and the
interaction between A and B atoms on the B side and which could be different.
The moment both of them become equal, then what should happen to this equation? It
will become same as regular solution because assume, that ΩA = ΩB. If ΩA = ΩB, then
ΔHmix = XAXBΩ(XA + XB)
. Since, XA + XB = 1, So, this equation will become same as the equation for regular
solution model. So, that is how we can visualize this and this is a kind of a next level of
non-ideality.
And then people observed that this is also not really fitting in very well, why? For
example, if you look at phase diagram such as magnesium-silicon phase diagram, how
many of you have seen magnesium silicon phase diagram? Anybody who knows how
does it look like? Mg2Si is a very important compound in aluminium alloys. No, if you
87
remember, if you have not seen it, I will not bother about the actual temperatures much,
but give you a schematic (see the phase diagram above). Obviously, this should be
silicon side, this should be magnesium side and do not bother about it, this is at the
center of the phase diagram, but this is actually Mg2Si, so that means, this on the
magnesium side. So, you see a phase diagram like this, it has two eutectics and one
compound.
So, that means, suddenly at the composition of Mg2Si, when 2 atoms of magnesium
come together with 1 atom of silicon, suddenly there is a strong attractive interaction
between the two and everywhere else. No. For example, on the, on the silicon side,
magnesium does not dissolve in silicon, that means, ΔH is positive, if it is negative, it
would have dissolved, there should have been solubility.
And similarly, on the magnesium side also, silicon does not dissolve into magnesium.
There is almost negligible solubility on the both sides if you carefully observe the phase
diagram. But suddenly, at this position, there is an attraction and its melting point is
higher than that of both the pure metals. This means that the interaction at certain
compositions suddenly becomes attractive. For example, there are many, many inter-
metallic compounds; our famous one that we know in steels is Fe3C. So, that means,
when you bring 3 atoms of iron and 1 atom of carbon together and there is a strong
attraction, but if you have 2 atoms of iron and 1 atom of carbon there may not be. So,
this is based on variety of factors.
People have, you know, classified them into types of compounds, electrons compound,
size factors compounds, let us not go into details of them, but there is a possibility, that
the interactions can be specific to certain specific compositions and if I come to either
this side or this side of Mg2Si composition, that interaction vanishes. That means, the
free energy of that is strongly negative at this composition and this is the reason why,
people have to use some other models beyond the sub-regular because sub-regular
cannot account for this. Simply, If you assume that there is some interaction on the A
side, some interaction on the B side and then, simply try to find out the ΔH mix as a
function of that, you cannot actually predict, that there is a compound in the center.
So, people have brought in what are called cluster models, cluster variation models. We
will not go deeper into this, just to give you a flavor, if you are interested come to me,
we will talk more. Lot of people, all this, so they assume that there are lot of clusters
88
inside a material and these clusters have, depending on what is the stoichiometry of the
cluster, you have certain free energy or certain, certain interaction, certain interaction
parameter and based on those clusters try to find out, which cluster has the lowest and
then find out based on that, whether there is a possibility of a compound. If all clusters
have the same enthalpy, then you assume that there is no possibility of a compound
formation in this particular system.
So, find out various possibility of such clusters and compute, based on the electronic
configuration and based on the atomic sizes, you see one can calculate that kind of a
thing and then find out, which cluster has the lowest and from that one can find out what
are the possible compounds. And in some systems there are, in fact more number of
compounds. If you take for example, aluminium nickel phase diagram, you have Al 3Ni,
you have AlNi, you have Ni3Al and a number of other compounds too.
So, that means at specific positions, you suddenly start getting inter-metallic compounds.
That means, those clusters at that particular composition have a lowest free energy or the
lowest, highest bond energy and as a result, that cluster becomes stable there and at that
composition, if for example, if I take an isomorphism system and if I choose all these
possible combinations of 3 is to 1, 1 is to 1, 1 is to 2, you know, 1 is to 3 and things like
that, and try to calculate for each composition what is the value of this interaction. You
would find that in an isomorphous system, they are all equal at all compositions; that
means, such compounds cannot form in such a phase diagram. So, people do all these
calculations and from that find out what is the possibility and once you know, that that is
possible, you treat it as a phase and then start calculating the free energy for that phase
too and then incorporate it into phase diagram and then check it experimentally, whether
you are able to see such a compound or not in the experimental verification.
So, that is how people have been, all the CALPHAD calculations are based on this. In
addition, CALPHAD people also have gone a little further because these are also, you
know, to some extent simpler, some cases, none of them fit, so that is why, sometimes
CALPHAD calculations are more kind of fitting parameters. So, instead of just one
parameter which is in regular, two parameters in sub-regular, people are nowadays using
in a four, five parameters.
89
Basically, take some experimental data that is already available, fit it with a certain
model, whatever number of parameters is like curve fitting for you and once you fit it,
try to extrapolate that into the other parts of the phase diagram, where you do not have
experimental data available or it is difficult for you to do experiments. I gave you an
example of copper tungsten, for example, I said, you simply do experiments on the
copper side and tungsten side you extrapolate it and so whatever, you know, free energy
function, that fits that particular experimental region, you try to find out that function
and try to see again, a bit of an assumption here, that I can linearly extrapolate or some
other extrapolation method I can use to see, that on the other side also the same thing can
be used. Otherwise there is no clue for us to be able to calculate the phase diagram
because you have to start with some assumption and people have done that kind of a
things as people went on calculating phase diagrams.
We will have a special lecture on how to calculate phase diagrams as we go along. Now,
this is all various models for finding out the free energy. After finding out the free
energy you would see, depending on the value of Δ H mix, whether it is positive or
negative, the free energy curve is going to look different. For example, this is one
something that we have drawn for an ideal solution (bottom image), where the Ω is 0 and
if that is the case, these two parameters are GA and GB.
90
I need to here tell you what would happen if I am talking about a material or a system,
where both A and B have, do not have the same structure. For example, take an example
of magnesium silicon itself. If I am talking of magnesium silicon system, magnesium,
anybody knows what is the crystal structure?
Student: HCP
It is HCP; silicon ?
Student: Diamond cubic
Yes, it is diamond cubic, DC. Now, because these two are different, when I am talking of
solid solutions for example, I want to do thermodynamics of this system. So, I am trying
drawing a free energy curve. Now, what is this free energy curve for, it should be if in
this particular phase diagram what are all the possible phases? One is, there is a
possibility of magnesium base solid solution, where silicon dissolves into magnesium.
Second, there is a possibility of silicon base solid solution, where magnesium dissolves
into silicon. So, I can call it as α magnesium or α silicon, whatever it is, solid solutions,
two solid solutions and there is a possibility of a liquid, there is a possibility of a
compound; so, four phases possible in that system at various temperatures.
Now, if I want to calculate the free energy of α magnesium and say, that this curve
belongs to that, the problem that I have is because of this parameter X AGA + XBGB. This
GA is that of magnesium, GB is that of silicon, but magnesium is HCP, silicon is diamond
cubic, how do I put the free energy of both, two different crystal structures into one
equation and talk about free energy of α magnesium?
So, whenever I am talking of free energy of an α magnesium, that is, free energy of a
particular structure, whenever we talk of any phase, we are talking of free energy of that
structure as a function of composition and temperature. If that is the case, I need to have
both these in the same structure. If G A and GB are two different crystal structures, then I
cannot put it into that equation. That means, what I need to do is, I need to find out what
is a stable state in this particular system. For example, magnesium stable state is HCP, so
I should find out GAα,. But I know GB, which is silicon, silicon is not HCP structure, but
for the equation of free energy, in order to get the free energy of α magnesium, I need to
have silicon also; free energy of pure silicon in the HCP form.
That is where I told you, if you go CALPHAD book by Kauffman, the first chapter is
dedicated for this on what is called lattice stability. So, people try to calculate what
would be the free energy of silicon if at all it exists as HCP; what would be the free
91
energy of silicon if at all it exists as FCC and that ΔG. So, there, it gives actually ΔG of
for different
92
metals FCC to HCP, similarly ΔG of silicon from diamond cubic to FCC and so on and
so forth. Once these values are calculated from 1st principles, I simply add this value to
the actual free energy of the diamond cubic silicon.
Free energy, what I can calculate from the CP is a free energy of diamond cubic silicon
only; I cannot calculate free energy of anything else from the C P because CP is that of the
diamond cubic silicon. What I am doing, calorimetry and finding out the heat capacity is
for a stable state. I cannot find out the heat capacity for something, which is a meta-
stable, which does not exist. If I put this sample into a calorimeter and measure the C P of
it, I would get the CP of that particular structure, what it is stable at those temperatures. If
I put silicon into the DSC and measure the heat capacity, I would get the heat capacity of
only the diamond cubic.
So, from that if I calculate H and S and get G and that G would only be that of diamond
cubic, I would not get free energy of anything else there. So, that is where I need to use
these things and add them to the actual free energy of silicon that I have calculated and
get the free energy of HCP silicon and put that into this equation, then only I can go
forward.
So, you can see, free energy calculations, they look very simple when you can draw
them this curve, but to get this curve it is not easy. So, so this is how we go ahead and
that is the reason why you need to know these lattice stability criterions. And for various
metals, people have done this kind of calculations and some methods are available to be
able to do that. So, now, if that is the case and try to draw that particular free energy and
if, then after taking care of the 1st part (of equation (1)), then I have to go to the 2nd part.
In the 2nd part, if it is a non-ideal solution, then I have to add this interaction term and
then, if this term is negative, what should happen to the free energy curve?
93
So, if you look at this equation (1)

Gα = XAGA + XBGB + RT(XAlnXA + XBlnXB) + ΩXAXB
, if you look at the TΔS term (entropy part, third term), this is -TΔS mix is drawn and now,
you have the free energy, the pure metal free energy some value, whatever it is and then
in such a case, if I know that this is 0, then my free energy curve exactly becomes that of
ideal solution, but if it is non-zero and if it is negative, let us assume that. This,
interestingly, regular solution model, the ΔHmix is again symmetric, similar to that of
ΔSmix, you can see it is ΩXAXB and that function, by looking at that function, we can say
it is symmetry.
It reaches either a maximum or a minimum depending on whether Ω is positive or

negative at 50-50. If I take XA = 0.1 and XB = 0.9, 0.1*0.9 = 0.09 and if I take it as
0.5* 0.5 = 0.25. So, 0.25 is the highest.
So, wherever composition I take, I will see that the maximum of XAXB is at 50-50. So, as
a result, if the Ω is negative, then it will have a symmetric curve and if it is a symmetric
curve, my free energy function would be, this is ΔH and this is G α (see the G-x graph
drawn above). In principle I should call this as ΔG mix, this is actually ΔGmix. Addition of
these two gives me ΔGmix because this is 0, this is 0. And this ΔGmix, if I add it to my
pure metal components, if my pure metal components have a free energy values of these
two, this is GA (on A side), this is GA (on B side), then I would know, that my value will
be something like this, something like this. This is the actual
94
Gα (indicated in the graph above) because this ΔGmix has to be added to the pure metal
component, which is the first part of the equation (1), then I will get the G.
Now, when I am talking of this G, depending on whether the Ω is positive or negative, as

long as Ω is negative, your free energy curve will always be plunging downwards. The
moment Ω is positive then you would see that there is the possibility of situation. Let us
say take a particular case again; I will reuse this itself. So, I say, that entropy of mixing is
not changing. Obviously, it will not change, isn’t? Any system, that I choose, entropy of
mixing is the same. We said before, that this entropy part is unbiased by what element it
is because there is no element coming in there, it is simply composition.
So, that is why, whether I take nickel aluminium system or magnesium silicon system, it
will have the same configurational entropy, whenever I take a 50-50 composition. It does
not depend on whether it is nickel aluminium or magnesium silicon. The system is
represented only in Ω and of course here, the actual pure metal values.
So, the interaction between two elements is comes only in this and If that is positive, if it
is following a regular solution model and if it is positive, then what should happen to the
free energy curve? (See the graph above) This is ΔH mix now and -TΔSmix, combination of
these two would give me ΔG mix. Now, depending on how much they are relative to each
other, you are going to see the effect.
95
If TΔSmix is larger than ΔHmix, this ΔGmix will be still negative.
And accordingly, actual Gα would turn out to be something like this.
But now, if you see, ΔHmix is larger than TΔSmix, then you would end up in situations
where the actual free energy curve may be something like this (upper curve). And you
may ask me, why it should go down and then come up, why cannot we have a free
energy curve which goes like this (lower curve)? That is where the question that
Aneesha has asked in one of the last classes is comes into picture.
Now, if you take this particular equation,

and try to find out the slope of it, that means, differentiate this equation with respect to
XB knowing pretty well, that XA + XB = 1 and try to differentiate this and find out that
∂y
value of ∂ x
B
|
X B=1
X A =1
do. I want you give this as an exercise to you, please do it in your
rooms and come back to me in the next class. You would see that if you do this, you
would find this value is -∞; please do it for yourself.
96
So, the slope of free energy versus composition G-X diagram, the slope of the free
energy curve as the XA 1 or XA 1 is -∞. That means, free energy curve always has
to go down. That means, free energy has to, that means, when I add a small amount of B
to A, free energy decreases because this is a function (eq. (1)) and for this function, you
see, when you add a small amount of B to A or small amount of A to B, you would see
that free energy can never increase.
Later, as you add more and more B to A, it may start increasing because of this being
positive, but towards the two pure metal ends, this, however positive Ω is, however
positive Ω is because of the nature of this lnXA and differential of lnXA brings you that,
you would see as we go along. If you calculate that, if you do this differentiating, you
would see that, you would tend to see, that whatever may be value of Ω, however
repulsive the interaction between the two metals is, but on the two pure metal ends the
free energy has to decrease when you add the two. And that is the reason, why you see it
is very, very difficult to purify a metal.
You can talk about four 9s, five 9s, six 9s, but you can never talk about 100% of A or a
100% of B. And the more nines you want, the more difficult it is because you are
somehow pulling the system away from its nature. It wants to have the elements inside it
and such metals are not really stable also. Any possibility, they would try to grab up a
little bit of oxygen from the atmosphere something like that. So, that is why, you would
see, one can calculate this and then see.
This is one aspect, now the second aspect that is important for us to know and we will
come back to this when we are talking about phase diagrams.
97
Second aspect is, whenever I have a free energy curve like this, there is one parameter,
which we define called chemical potential (potential). All of you must have come across
this how do you define a μ?
Student: Change in free energy of the whole system by the small addition of a
component
Correct; change in free energy as a function of composition. Basically, if you want to

relate this, if I draw, if you take this particular curve (see the G-X curve above) and draw
at any particular composition (point P) a tangent to the free energy curve, wherever this
intersects, these two values give me on one side μAα, If this is α phase, this is μBα (on B
side).
potential indicates what is called the activity because there is a relation between μ and
activity. You must have seen RTlna, that equation you must have seen. So μ is always
proportional to the activity or activity is proportional to the chemical. So, whenever you
see a μ of, for a particular element in a phase is higher, we say, that particular element is
more active in that particular phase. Here for example, if I draw this and I show you, that
μBα > μAα, I can say, that for that composition at that temperature, B atoms are more
active in α than A atoms.
98
So, activity and μ are directly related and this is what we are going to use it to
understand, what is called the equilibrium between two phases. When you bring two
phases together and keep them together what would happen?
So, at any composition and these μ’s are not fixed for a given temperature, they depend
on the composition. For example, the same phase I come to here (point Q), on this side
of the free energy diagram and again draw a tangent, you would see that this is μ Bα, this
is μAα. That means, these two compositions, if I take α of, let us say, 30% B and α of 70
% B and take two of those pieces of these two α, you would see, the α on this side (point
Q) would have higher activity for pure A when compared to B and on this side (point P)
higher activity of pure B.
So, in principle, if I bring these two pieces of, let us say, two α’s, let us say, one has 30%
of zinc, whatever it is B element, and one has 70% of B element, bring them, weld them
together, hold it at that particular temperature as a function of time. You would see,
because one has B atoms being more active, another has A atoms being more active, you
would see from this side (P) B atoms will diffuse towards this side (Q), A atoms will
diffuse from this to this (Q to P) and both of them will come to an equilibrium at some
stage, so that the composition of both the mixtures will become same of some
equilibrium composition, if I take the same phase.
99
If I now take two phases with two different compositions, what would happen, you
would see that in a minute. For example, if I take two phases, I can always represent two
phases with two free energy curves. The only, deliberately I am drawing these points
different because I do not know, and carefully if you can observe now, let us take a
system of, let us say aluminium-silicon or magnesium-silicon again.
If you draw these two free energy curves (see the image above), one is for that of α-
magnesium; let us say, another is for α-silicon. You will see, on the pure magnesium side
and pure silicon side, you would see two points here (on Mg-side) and two points here
(on Si-side), what does those two points mean? On the two pure metal ends you have
two free energy points, what are those two free energy points?
Student: Free energy of Magnesium and free energy of silicon
This, these two points, can you tell me first, what are these two points (on pure Mg)?
Student: Pure Magnesium
And what is the second one?
Student: Pure Si
I am talking of this side, do not again bring…
Student: Magnesium in second phase
Yes, this is the free energy of magnesium in the α form; this is the free energy of
magnesium in the diamond cubic form or HCP (G MgHCP) another is G of magnesium in
the diamond cubic form (GMgDC) . And obviously, the free energy of magnesium in the
diamond cubic form has to be higher. And similarly, on pure Si side, if I see that this is
free energy of silicon in the diamond cubic form (G SiDC) and this is free energy of silicon
in the HCP form (GSiHCP) (please refer to the free energy points indicated on pure metals
side). And if I carefully observe for one particular curve, I am using free energy of the
same structures pure metals for calculating that free energy. If I take this particular curve
(α-Mg), I am using HCP, HCP; if I am using this curve (α-Si), I am using diamond
cubic, diamond cubic, otherwise you cannot not draw the free energy curve. So, this is
how free energy of two different phases.
100
If I assume these are the two at, at room temperature in principal, let us say, I am
forgetting the existence of a inter-metallic. Now if you want even to bring in inter-
101
metallic into the picture, you can also draw that, but for today’s class we do not need it;
in next class we will bring that.
Now, if I look at two compositions of magnesium, you know, magnesium-silicon alloys,

one, let us say, containing about 20 % silicon, one piece containing 20 % of silicon in
magnesium, another piece containing, let us say, 80 % of silicon in magnesium and then
put them. Let us say, this is one composition (point P), this is one another composition
somewhere here (point Q), if I draw the free energy curves and draw the tangents.
So, that means, basically, what I am doing is, I am taking two pieces of alloys, welding
them together, one containing 20 % silicon, another containing 80 % silicon, weld them
together, hold them at that temperature and see, what happens?
You would see actually, that whenever I do this, there is, because the activities are
different and which is represented by these free energy curves and the tangents to them.
For example, if I draw a tangent at this particular composition (point P), I will see, that
this tangent go like this. And similarly, if I draw a tangent (point Q), this tangent goes
like this. And now, if I look at it, what is what are these values? The tangent to the α-Si
curve gives μMgα-Si and μSiα-Si, whereas the tangent to the α-Mg curve gives μSiα-Mg and μMgα-
Mg
So, the two μ’s, always remember, we cannot say this is the μ of a particular phase, we
always say this is the μ of a particular element.
μ is always for a given element; activity is always for a given element. Activity is never
for a phase; activity of elements we talk about. Similarly, μ of elements, element in
which phase, that is what we look at. So, if you look at it, here is μ Mgα-Si, this is the μMgα-
Mg
, what does this tell you? This tells me, that the μMgα-Mg is higher than the μ of μMgα-Si.
So, in my mixture, in my piece (see the bottom of the image above), that I have taken, two
pieces that I have taken, one is that this is α magnesium side, this is α silicon side, 20 %
magnesium, 80 % silicon; I said, one is the α silicon side, another is α magnesium side. On
the
102
α magnesium side, if I look at it, the, the activity of magnesium is high and if you look at
it, α magnesium side, α activity of silicon is lower. So, this, here (α magnesium side) you
have higher activity of magnesium, lower activity of silicon. Here (α Si side), you have
higher activity of silicon, lower activity of magnesium. As a result, you see that the
silicon atoms will diffuse this direction (α-Si to α-Mg), magnesium atoms will diffuse in
this direction (α-Mg to α-Si) because you have a higher activity there.
So, if I look at for a given element, because the activity in the α magnesium is higher, so
this element moves from α magnesium side to α silicon side. Similarly, silicon moves
from α silicon side to α magnesium side and in the process what happens? In the process,
the composition of the two starts changing; α magnesium would keep on getting more
and more silicon and α silicon would start getting more and more magnesium. So, these
two compositions will start moving towards each other, when more magnesium is added
to α silicon, the composition will move this direction; when more silicon is added to α
magnesium, the composition will move towards this direction. So, the alloy composition
of the two phases starts moving towards each other.
And now, at these two different new compositions, if I draw the tangents I will see, that
these tangents now will start moving. In fact, I will show you that these tangents would
start moving like this. If the composition changes from here to here, I mean here to here,
this is the 1st case, this is the 1st case, from there you have to case number 2. From 1 to
2 if you have moved, you see what has happened is this difference has come down
(indicated by arrows). What does it mean?
That means, the activities of the two pure metals in the two phases are coming closer and
as this keeps on happening, at some stage you would see that you reach A particular
composition of α magnesium and A particular composition of α silicon, such that when I
draw a tangent for one, that tangent becomes this tangent also for the other phase, this is
what we call it as common tangent. So, that means, you would reach a situation where
the tangents are the same.
When the tangents are the same, what does it mean? This particular composition is mu
μMgα-Mg is also μMgα-Mg , activity of magnesium in α magnesium and activity of
magnesium in α
103
silicon becomes the same, and once the activity on both the sides is the same, then atoms
cannot moves and that is the definition of equilibrium.
In any multi-component system, equilibrium is defined not simply by which has a lower
free energy, it is defined by equality of μ’s and when the μ’s become equal, then one
element cannot move from the other side, from one side to the other then you see
wherever the elements are they remain there and that is an equilibrium thermal
equilibrium at that temperature. And within that material, there may be thermal
vibrations because the temperature is not absolute 0. Thermal vibrations are there in α
magnesium; also, thermal vibrations are there in α silicon also. Atoms are moving in
that, but they are not moving in across the inter phase and because there is no activity
gradient, the gradient in the activity has became 0 and once the…
That is the reason why, whenever, you remember a long back, if you remember Fick’s
first law, J = -D(dc/dx). All of you, in one voice you have said dc by dx, actually this is
not dc by dx. You would see this difference if you have seen, taken any ternary system.
For example, if you take, people have done this interesting experiment. They have taken
a cast iron with about 3% carbon in it, but about 2 % of silicon in it. This cast iron, they
have welded it to another steel containing about 0.4 % carbon, what do you expect when
these two are kept at high temperature? You expect that the carbon should diffuse from
the cast iron side to the steel side, isn’t it? One side is 3
104
% carbon, another side 4, 0.4 % carbon; you actually see the reverse. Carbon will go
from the steel side to the cast iron side. People started calling this as uphill diffusion.
And the reason is, here is, though the concentration gradient is there, there is no μ
gradient. The reason is silicon, the moment it is present, it locks up the carbon and
reduces the activity of carbon atoms. There are certain elements which have this
tendency, for example, compound forming elements, carbide forming elements, the
moment I add to an alloy steel, all this carbide forming elements will lock up the carbon.
Let us say, I add chromium, vanadium, titanium to a cast iron or a steel, you see that all
these lock up the carbon and reduce the activity of carbon. Once the carbon forms a
compound, the carbon atoms cannot move further because they are strongly bonded with
some other element. So, once this activity of carbon comes down, now you have to
actually see the activity gradient and not the concentration gradient and people have
actually done this.
And this problem does not arise if I am taking two binary systems. Let us say, one steel
of 0.4 % carbon, other steel of 1 % carbon with no other alloying element present in it. If
I take, definitely from, carbon will flow from 1 % carbon side to 0.4 % carbon side, there
is no problem. So, because there, the μ is driven by the concentration gradient, μ is the…
Whereas, in ternary cases, the μ’s are changed because of the presence of ternary
interactions and that is what we have to remember and that is how we understand the μ.
So, we should say that it is μAα = μAβ as the equilibrium concept.
With this we will stop.
105
106
Prof. B. S. Murty
Lecture # 06
Phase rule, Free energy-composition diagrams and Phase diagrams
107
Yeah, last class we kind of defined the equilibrium concept in two component systems
and I said, in principle, we can extend this to any number of components. And the
definition there was, that
μAα = μAβ
then we can say, the two phases (α and β) are in equilibrium at that particular
temperature and pressure. So, this is the definition.
And in principle, if you want to extend it to a more number of phases, you can even say,
μAα = μAβ = μAγ
for example, when we say, an eutectic reaction, a L  α + β. And when I say, all the
three phases are in equilibrium, it means this, the chemical potential of one of the
elements, let us say A, is equal in the liquid, in the α, in the β, that is the meaning of
eutectic when I say a L  α + β and all the three are in equilibrium.
And similarly, we can also extend this to another component and say
μBα = μBβ = μBγ
and so on and so forth, depending on whatever number of phases that are present. And
this can be written to all the components that are present in the system. In a binary
system we have only two components, in a ternary you can extend it to a 3rd component,
in a quaternary four component and so on and so forth.
108
And this actually brings us to a concept called phase rule, which Gibbs talks about, this
is, that is why we call it as Gibb’s phase rule, which is again related to this. In fact, when
we talked about this phase diagram (see the phase diagram above), which is the
temperature-pressure diagram for the solid, which is the ice, the water and the steam, we
drew this diagram if you remember. And there, you had a doubt about this triple point,
how do we talk about it and once you understand this phase rule then understanding that
become very easy.
So, what is this phase rule? Phase rule talks about what are the degrees of freedom that
are available for any system, independent variables that are available for any system at
any given condition of temperature and pressure. If I look at, that any system when I talk
about it, the total number of variables in a system are always related to the number of
components. For example, if I say I have a phase here, I want to describe this phase.
When I want to describe this phase, I describe it like the way we describe a human being
in terms of his emotions.
So, we talk about the, this particular phase in terms of the number of components present
in it and additional variables are the pressure and the temperature. So, this phase is
present at this pressure and at this temperature and this has so many components. So, for
each phase, I can actually define a certain number of components, which we can call it
as, let us say, call the components as C and call the phase as P. The total number of
components or the total number of variables are always P*(C + 2); this will be the actual
total number of components.
But one thing, that you have to remember is, whenever I am talking about a phase, I need
not have to talk about all the components present in the phase. For example, if I am
talking of a binary alloy, let us say a steel, iron-carbon alloy and I am referring to what is
the composition of a α ferrite, let us say, if I talk about how much of iron is there in it or
how much of carbon is there in it, automatically I need not have to tell about the 2nd
element.
So, that means, whenever a certain number of components are present in any phase, I
need to talk about only the components - 1 as the independent variables, one of them
becomes automatically a dependent variable. If it is a three component system, if I talk
about how much of A, how much of B is there, I need not have to say how much of C is
there; automatically it is 100 - (A+B). So, as a result, in principle, I need
109
not have to describe all the components present in a phase. So, that is why, we define it
as P*(C-1) + 2 as the total number of variables to describe a system. If I have P number
of phases, each phase has C components, for each phase I need to define how much. For
example, you think of a eutectic reaction, a eutectic reaction, where a L  α + β. If I
want to describe this liquid, I need to describe what is its composition, how much of A is
there, how much of B is there in this, but if I tell how much of A is there, automatically I
am telling how much of B is there.
Similarly, I need to tell, what is the composition of α? Composition of α is not same as

composition of liquid; similarly, composition of β is not same as that. So, for each of the
phase, I need to define how much of A is there, how much of B is there. And because
this is a binary system, if I tell about how much of A is there, that is enough or if I tell
about how much of B is there, that is enough.
So, that means, for each of the phase I need to define what is the, the number of
components, that are independent variables. And at the same time, because this is a
eutectic, all of the phases are at the same temperature; we are going to come to it, why it
should be at the same temperature. It turns out from the phase rule that it has to be, all
the three have to be at the same temperature, otherwise eutectic reaction will not take
place. So, they are all at the same temperature and of course, because I am doing this
whole transformation at atmospheric pressure, they are also at the same pressure.
So, in principle, pressure and temperature are constant for this reaction, but it need not
be the case for all the cases. For any reaction, that is happening, you need not have to
have temperature and pressure of both the, all the phases being at the same. So,
considering, that we in principle talk about, whenever we talk about the total number of
variables to define the whole system containing P number of phases is this way, that
each phase, I need to define C - 1 components and all these phases are at a particular
temperature and pressure. So, I need to define what is that temperature, what is that
pressure? So, I need to have two additional variables for temperature and pressure and if
I have P number of phases, P* C - 1 will tell me that.
And if I say, that all these phases are in equilibrium with each other, then equilibrium
demands, that certain variables become automatically constant. What is that, equilibrium
tells me that for each component if I define, for example, one chemical
110
potential for pure A μAα, I need not have to define what is μAβ and μAγ and so on. Once I
define this they are all fixed. That means, for each component P - 1 gets fixed once I
define the chemical potential. So, for every component, I need to basically define only
one chemical potential. For the remaining P - 1 phases, I need not have to define, they
are all automatically fixed. That means, for each component I need to define a chemical
potential in any one of the phases; if I define in any one of the phases, the remaining
phases I need not have to define.
So, that means, for each component, P - 1 variables gets fixed once I define the chemical
potential, so that means, C(P-1) becomes kind of fixed variables. Thus,
Number of variables that need to be defined for a given system = Total variables – No.
of fixed variables
F = P(C-1) + 2 – C(P-1)
F=C–P+2
F is called as degrees of freedom.
We also usually define it as P + F = C + 2.

This is what we call it as the Phase rule P + F = C + 2. We used to remember, when we
were students of your age, police force is constables + 2, something like that. So, it is, so
it is easier to define, remember.
So, anyway, what is important is, this is a general phase rule and from this general phase
rule we get, what is also called, a condensed phase rule, where we, we keep, one of these
two is the pressure constant because most of our system, most of our processes occur at
atmospheric pressure. I need not have to define pressure, pressure is fixed and once I fix
111
up the pressure, this becomes P + F = C + 1, that is what is called condensed phase rule.
And if I look at this phase rule P + F = C + 2 and then look at this phase diagram now
(water phase diagram in above image), it becomes clear to me, why do this, the melting
point for example, ice to water melting point should change as a function of pressure.
For example, you look at this, whenever an ice is becoming water, how many phases are
involved in a process? There are two phases, one is changing into other; so, there are two
phases. And how many components are there here? There is one component, so P = 2
and C = 1, you put that into this, what happens, F = 1.
So, there is one degree of freedom for the system. That means if I fix up that particular
degree of freedom, then everything else gets fixed. That means, if I fix up the pressure,
then the temperature at which ice transforms to water gets fixed because that is the only
degree of freedom that I have. Once I fix up that degree of freedom, then I do not have
any degrees of freedom in my hand, there are no more variables that are possible. The
system has to be everything, has to be fixed, there are no more variables possible and
that is how we define this as this.
And similarly, if I fix up the temperature, the pressure at which the ice transforms to
water again gets fixed. So, that is how we look at it and any, any of these lines you can
talk from that point of view. And now, if you come to this point, where the three phases
are in an equilibrium, you put P = 3 and C = 1, you would see F = 0. That is where we
say, there are no degrees of freedom for this.
If at all you want all the three phases to co-exist, they can co-exist only at a fixed
temperature and pressure and that is not in your hands. You cannot make all the three
phases co-exist at your wish, that at any temperature and pressure of my choice, no and
that is fixed by the, by the thermodynamics, that they all three can be together only at a
fixed and this is what you see in all the phase diagram.
112
For example, if I draw a phase diagram, what is this phase diagram (Fig. 1)?
Student: Isomorphous
Isomorphous; so, in this isomorphous phase diagram, what are these two points (P and
Q)? Melting points of A and B, and if I look at the melting points of A and B and apply
my phase rule there at that temperature at that point, what is happening at this point?
What is happening? Solid is becoming liquid or liquid is becoming solid depending on
which direction I am looking at. And when solid is becoming liquid or liquid becoming
solid, how many phases we have?
Student: 2
We have two. And how many components we have? We have one. And then, how many
degrees of freedom we have? One. That’s where the problem is.
When we are talking about isomorphous diagram of metals, melting or solidifying, we

always keep the pressure as constant and once we keep the pressure as constant, because
this diagram has come only when you have fixed the pressure, this is at atmospheric
pressure, if I vary the pressure, then this diagram changes. So, all the melting and
solidification that we come across in our day to day life, we all do at atmospheric
pressure.
So, as a result, we do not use the generalized phase rule, but we use a condensed phase
rule. And once I use the condensed phase rule because P + F = C + 1, then you see, that
113
F = 0 there and that is the reason why the, the melting and solidification occurs at a fixed
temperature.
114
If you remember this cooling curve (Fig. 2) temperature versus time, why does this
solidification occur at a fixed temperature for a pure metal? It is because of this, the
Gibb’s phase rule tells, that the melting and solidification has to occur at a fixed
temperature. Thermodynamically it is not possible, that the melting can occur over a
range of temperature, but if I take an alloy of this composition (C0), let us say, there I am
not having components as one, now components is two, when I consider C = 2, in the
same equation P = 2 and C = 2, then F = 1.
So, when, when I see F = 1, that means melting can occur over a range of temperature
and that is what you see. That is why, we call liquidus and solidus; for a pure metal,
liquidus and solidus are equal, there is no separate liquidus and solidus for a pure metal.
So, that is why you see, that both the liquidus and solidus merge into each other when
you come to one end of the phase diagram or the other end of the phase diagram because
phase rule says that they have to have the same melting point and that is why, if I draw a,
what is called, a cooling curve for an alloy T versus t, you would see this kind of a curve
(Fig. 3) and this is what we call it as liquidus, this is what we call it as solidus; liquidus
and solidus. And here it is liquid, here liquid is changing to solid (l  s), here it is solid.
And because there is an enthalpy change, there is an enthalpy released when a liquid
becomes a solid that is why you see the slope changes that take place (see Fig. 3). Why
should the slope change the moment you reach the liquidus is because, the moment you
reach the liquidus, solid is coming out of it and when solid is coming out of it, solid
cannot take all the heat that is there inside the liquid. So, this excess heat, which is there
inside the liquid, has to be given out. When it is given out, this excess heat also has to be
extracted by the mould into which you have poured this liquid and as a result, because
the, heating rate is, cooling rate is fixed, you have an excess heat that is available. So,
automatically, the cooling rate, that dT/ dt is nothing but the cooling rate. The slope of
this tells you the cooling rate. This is a higher cooling rate and the moment solid starts
coming because excess heat is available, the cooling rate suddenly drops and that is the
reason why you see this change of slope.
And once the solidification is over, again it starts cooling by the normal Newtonian
cooling. In a, any single phase material when it cools, what is the rate at which it cools?
Anybody knows rate is proportional to what?
115
Student: Fourth power of the Temperature
Fourth power of temperature; isn’t it. So, that is why you see and that is why this slope
(liquid slope, Fig. 3) has to be higher than this slope (solid slope, Fig. 3 because the
temperature is already low here. So, because the cooling rate is fourth power of
temperature, so in the liquid state, the material cools faster, in the solid state it cools
slower and that is how we can find out, whereas in a pure metal it is like this.
And now, if I look at a eutectic alloy again you see, the thing will be exactly like this
(Fig. 2). Why? Because I have three phases there, liquid giving α + β; there are three
phases that are involved and how many components? Two components; it is a binary
alloy.
So, you imagine this eutectic diagram (Fig. 4), you would see that again, whenever you
are drawing, be careful, that these two have to meet each other at that because that is the
melting point of pure B, this is the melting point of pure A. And so, this horizontal that
we are talking about and that horizontal, if you take an alloy of this composition and do a
cooling curve analysis, you will exactly get this kind of a thing (Fig. 2)
You see an under cooling. Whenever the solidification is difficult, whenever you do not
have, what is called the heterogeneous nucleating sites available, you see for a
homogeneous nucleation, you see an undercooling. You see the same thing for a pure
metal, you see the same thing for a eutectic alloy because if here, because there are three
phases in equilibrium, liquid is giving α + β, there are 2 components, put P = 3 and C =2
and use condensed phase rule and you would see, that F = 0.
So, that is how you see in a eutectic, that F = 0 and once F = 0, then solidification occurs
at a fixed temperature and that is the reason why, we call this as an invariant reaction,
whereas this is a univariant reaction. Solidification in a, in a binary alloy, in an
isomorphous binary alloy is a univariant reaction, whereas solidification of a eutectic
alloy is a invariant reaction. But again, you have to remember, it is not always eutectics
are all invariant reactions.
116
The moment I go to a ternary system, there are also a number of ternary systems and if I take
the same eutectic reaction giving  α + β in a ternary alloy, let us say you have a cast iron,
let us say, iron-carbon-silicon alloy and in a iron-carbon-silicon cast iron, if the same
eutectic reaction is taking place, where, what is the eutectic reaction in iron carbon system?
Liquid  Austenite + cementite .
That reaction if it is taking place, instead of a iron-carbon alloy, if it is taking place in an

iron-carbon-silicon alloy, same reaction, that would never be an invariant reaction, why?
Because in this reaction, in the condensed phase rule, I have to put P = 3 and C= 3, then
you will see F = 1. So, that means, the same horizontal, that you see at 1130 ºC, that
temperature that you see as a horizontal in a iron-carbon diagram, it will be an inclined
line, it will not be horizontal in a ternary.
So, that is why, you know, there is a standard way of asking people, ask, binary eutectic
in a ternary system is univariant. But at the same time, you can have other type of
eutectics, where liquid gives 3 phases. In many ternary systems, such kind of eutectics
do exist, where one L  α + β + γ and the moment you have such a reaction, that
reaction will be definitely invariant reaction in a ternary system because there P = 4, C =
3, then you will see F = 0.
So, one has to be careful in which system, that you are talking about; what kind of a
reaction, that you are talking about; whether it is a binary or a ternary or a quaternary,
accordingly this reactions will be either invariant or univariant or bivariant, whatever it
is. And so, phase rule can really give us a lot of idea. Then, there are also interesting
things, interesting phase diagrams; I will just show you one such phase diagram if you
have ever seen it.
117
Have you ever seen a phase diagram like this (Graph 1)? It is called as pseudo-eutectic
phase diagram. You might also have seen another phase diagram something like this
(Graph 2); that is a phase diagram. These two have a lot of meaning in terms of
thermodynamics; we are going to see how phase diagrams evolve.
And based on this kind of concept, whenever you see this kind of a phase diagram
(Graph 2), it means, solid is trying to be stable up to higher temperatures. Here (Graph
1), it means solid stability is much lower when compared to liquid; solid is not stable.
This happens when the two elements do not attract each other, they do not have a strong
attraction. So, that means, if Ω is close to 0, but positive, then you see this kind of a
situation (Graph 1); when the Ω is negative, you see this kind of a phase diagram (Graph
2).
That means, there is an attraction between A and B, the A and B bonds are stronger and
they do not want to break at low temperatures. So, they retain their bonds even at higher
temperature. So, as a result, the melting point of the alloy is higher than the two
individual pure metals and you see, in all such diagrams, there is a tendency for
formation of inter-metallic compounds.
Copper-gold system, if you look at it, you will see something similar. If you see nickel-
aluminium system, you see a suddenly an inter-metallic compound at the center of the
phase diagram like that. So, this is what is called congruent melting system (Graph 2)
and whereas, this is what is called as pseudo-eutectic (Graph 1). But if I
118
look at both these cases, how do I explain this, here (Graph 2, C0) this alloy is melting at
a fixed temperature, am I right; it is melting at a fixed temperature.
If it is melting at a fixed temperature, means, F = 0 and whenever melting is taking

place, there is a liquid and a solid. P, number of phases are 2 and if I P = 2, and if F = 0,
then C has to be 1. Whereas, I am saying, that I am talking about a binary alloy, how do I
reconcile with this? That is why, I said sometime back to you that you get into problems
like this, how do I visualize this now? That is where we treat this as a, as a, as a simple
component system with, which is nothing, but a cluster which is a strongly bonded
cluster.
So, I cannot, it is like treating a H 2O. When H2O is melting or solidifying, we do not
treat H2O as two components, we treat H2O as one single component. Similarly, we treat
this also as one single component in all such cases. So, otherwise, you will not be able to
explain this. So, these are the things that you need to remember as you go along.
Now, let us come out of this phase rule, we have had enough of it and try to understand,
how to use free energy composition diagrams in various phase diagrams. Let us look at a
typical phase diagram, such as an isomorphous diagram (Graph 1). Whenever I am
drawing an isomorphous diagram and what is this diagram, it is an equilibrium diagram,
am I right?
119
When I say an equilibrium diagram, that means, all the phases, that are present there are
all equilibrium phases at that particular temperature and at that particular pressure.
And the concept, that we have defined just a little while ago where we are talking about
equality of chemical potential, that is also an equilibrium concept, so anything, that we
see and where did it come from? For example, if you remember, we have drawn (Graph
2), let us say, two free energy curves and then drawn a common tangent, and we say that
α of this composition is in equilibrium with β of this composition at that particular
temperature and pressure. We will not talk about pressure because we assume that
pressure is already constant, but it is at that temperature.
So, that means, if I draw a line (at temp T, Fig. 1)), a constant temperature line like a, what
we call such lines in phase diagrams? They are called tie lines; tie lines. So, if I draw a tie
line there and then, what are the phases here? Liquid (L), α and L + α. So, that means, if I
have a liquid here and I have an α here (Fig. 2), in principle, if this is a schematic free energy
composition diagram for this system, then these particular compositions ( Cα and Cl) should
represent these compositions (at temp T, Fig. 1). There has to be a one-to-one
correspondence between the free energy composition diagram and the phase diagram and
this is the simple concept, which people use when they try to calculate this phase diagram
based on free energy curves.
For example, what I can simply do is, at each temperature I can find out, how the free
energy of α and free energy of liquid are changing. For example, if I draw a free energy
curves at this temperature (T1), at that temperature what is stable?
Student: Liquid
Liquid, so how should the free energy curves should look like? Yes, liquid should be
below that of solid. So, if I draw G versus X (Graph 3), I should see liquid curve should
be like this and the solid curve should be something like this. Anything more than this,
that we can get information?
Student: The difference between them is also important
Yes, yes, these differences are also important; why they are important? For example,
look at this, this side on the pure A side and a pure B side if I look at it. Pure A side, the
melting point is very low (Fig. 1); pure B side (Fig. 1), the melting point is high. That
means, on the pure A side
120
at T1 because that is much higher than the melting point, the liquid is much more stable
than the solid.
So, the difference in the free energy of the liquid and solid for the pure A should be
larger when compared to the difference in the free energy of liquid and solid on the pure
B side because here, the temperature of our consideration (T1) and the melting point are
closer to each other.
If you calculate using whatever formula, that we have given,

G = XAGA + XBGB + RT(XAlnXA + XBlnXB) + ΩXAXB
If you put the whole thing and what you need in this whole formula is you need to know,
the T what temperature I am calculating, I need to know what is the Ω, what else I need
to know in that equation? Molar fractions is, actually when I am calculating free energy I
am calculating as a function of composition, so there is nothing to know about molar
fractions. I calculate Gα as a function of XB, fixing the temperature this is what and I try
to plot. So, you need to know, what is the temperature; you need to know, what is the Ω
and anything else? You need to know, the GA and GB.
And also, keeping in mind, in systems which are not isomorphous, in case of
isomorphous there is no problem for me because A and B has the same crystal structure.
If A and B do not have the same crystal structure, I need to go into that lattice stability
criterion and find out the free energy of one phase with respect to the free energy of the
stable phase of that particular pure metal, and in this case, that problem does not arise.
So, this is what I need to know.
Once I have GA and GB at that temperature, I need to find out G A and GB at that
temperature. That means, G = H - TS. I need to calculate H at that temperature, S at that
temperature and then get this, then I can say I have the Gα or Gl as a function of
composition. This is how we calculate. And if I calculate it at a fixed temperature like
that, this is how it will look like.
Now, if I come to lower temperature (T2) like that, what would happen to the free
energy curves? what is happening? I mean, are the free energy curves going up or free
121
energy curves are coming down, what is happening?
Student: α will come down
α will come down and liquid will go up?
Student: As the temperature comes down, both will go up.
He is right; as the temperature comes down, both will go up, why? Remember our 1st
class, where we said, the dG/dT = - S. So, free energy decreases as the temperature
increases; free energy increases as the temperature decreases. So, if I am decreasing the
temperature from T1 to T2, then the free energy should increase. So, what happens, the
free energy of the liquid and free energy of α, both of them will increase.
The only question is what is the rate at which they increase? The rate at which they
increase, because liquid has a higher entropy, the rate at which the liquid curve will go
up is going to be faster than the rate at which the α curve is going to go up. So, slowly it
is like a catching game, two fellows are running, the other fellow is running faster, tries
to catch up with the other fellow.
And exactly, that is what happens. So, liquid is trying to go up and α is also going up,
but α is going at, at a slower pace and liquid is going up at a higher pace. And as a result,
if I now draw at T2, you would see a situation like this (Graph 4).
You would see that this difference (between free energy at pure elements side) comes
down, both the differences come down, but these two come down closer, much closer (at
pure B side), because this temperature (T2) is very close to the melting point (of B). And
the moment you chose, now another temperature, which is exactly the melting point,
then you would see that these two will merge with each other (Graph 5). So, both of
them (Liquid and α) will meet each other at this point (pure B side) indicating, what does
it indicate? That means at that particular composition of B, α and liquid have the same
free energy that means, α and liquid are in equilibrium.
Remember, for a pure component, equilibrium is defined in terms of equality of free

energy. In a multi-component system, equilibrium is not defined in terms of equality of
free energy; it is defined in terms of equality of chemical potential. So, you can see that
122
that means, that is the, the moment you see at pure B, that these two free energy curves
come together and become equal, that temperature, whatever, if I keep on varying the
temperature by every 1 degree and do the same calculation. In fact, large number of
people have done, in fact, in one of the thermodynamics class like this, we have at least
generated 10 phase diagrams, simply calculating it and proving, that what you see in a.
in a Masalski’s hand book, let us say, is what you actually can calculate.
And if you vary the temperature every 1 degree and start calculating, you will see,
exactly when you reach that melting point, these two will meet there and now, if you
come to a lower temperature (T3, Graph 1), the liquid is now going up with respect to
the α and as a result, You will see a situation (Graph 3) where a liquid curve will
intersect the α curve and when it intersects the α curve, when I draw the common tangent
here, these two compositions, this is the liquid curve, this is the α curve and what does it
tell you, what are these two points? This is the free energy of liquid B (G Bl), this is the
free energy of α B (GBα) and you can see here the α has a lower free energy than the
liquid.
On this side (pure A) you see, the liquid has a lower free energy than α, why? Because
on this side it is still liquid that is stable (Graph 1). On this side (pure B side), once I
come to this temperature, it is already α.
123
At T3, pure B side I am already below the melting point. And because I am already
below the melting point, the solid should have a lower free energy than that of the liquid.
On the other side, you will see the reverse and wherever these are intersecting and if you
draw a common tangent to these two free energy curves (Graph 3), you get these two
points (Graph 3, Cα and Cl). And those 2 points are nothing, but these 2 points (Graph 1,
Cα and Cl).
And if I keep doing it at various temperatures, at each temperature I can generate these 2
points, which I will call it as a liquidus point (C l) and a solidus point (Cα) . And if I keep
on joining all these points, that I get, I can generate a phase diagram and in fact, in an
experimentally, how are we doing? We are doing cooling curves, for each composition,
there we are not considering temperature, we are choosing, let us say, A to B, every 1
percent. Let us say, I make an alloy, 100 alloys I make, for each alloy I do cooling curve
analysis, I will be able to generate liquidus and solidus points for each alloy and join all
those liquidus points, join all those solidus points, you will get a phase diagram like this
(Graph 1).
Instead of doing those laborious experiments, I can actually do this calculation and do it
at various temperatures, just one single equation and just only thing is, put it into a do
loop and ask it to calculate at every 1 degree of increment of temperature or 1 degree of
decrement of temperature, whatever you want. Simply, the program can calculate and at
each of that temperature, you ask it to find out, which are the two points, find out where
the common tangents? So, draw the tangents, ask it to draw the tangents and find out
where the slopes are equal and find out that and that is how we can actually do it and get
this whole thing.
124
And in this process, you would also come across one point (Graph 3, point P), which is
here. There is an intersection point, what is that intersection point tells us?
Student: Both free energies are equal
Both free energies are equal, what is the significance of this, that is not equilibrium,
equilibrium is this (common tangent). For example, where is that particular point, that
point is somewhere between these two (Cα and Cl) , am I right, it has to be between these
two and so that means, at a given temperature T3, when I look for this point, that point
falls somewhere here (Graph 1, point P), somewhere between these two compositions
(Cα and Cl) that point will be there, where, which is the intersection point, where you are
calling it as, where Gα = Gl.
So, that means, there is a particular composition where the α having that composition
will have the same free energy as the liquid having the same composition, and this is
125
what we call it as a point, where the equi-free energy point. And if I do this exercise for
various temperatures, at each temperature, whenever I see a common tangent, whenever
I see two compositions corresponding to equilibrium α and equilibrium liquid
compositions, you will also find out an intersection point and all the intersection points
will always be between the two equilibrium compositions of liquid and solid.
And so, that means, you will get the points like this, which if I join, you would get a
curve like this (dashed line, Graph 1), which goes like this and on the two pure metal
ends, this point will join with the melting point because the two pure metals, the
definition of melting point itself is Gα = Gl. So, as a result, our this point is also nothing,
but locus of all those compositions, which have the same free energy, where the α and
liquid have the same free energy. For example, if I choose an alloy of a particular
composition , at that composition, what is the temperature at which G α = Gl? And that
particular point is like P.
And this line is called T0 curve, and what is the significance of this T0 curve? We will
see as we go along, it is a very profound concept and this is the one, which brings you
into what are called metastable phase diagrams, and we will see as we go along.
So, this is how we can, in principle, calculate phase diagrams and the more complicated
it is, that means, if you have more, for example, if it is a eutectic, you have to add one
more phase, L  α + β. So, that means, in principle, I should calculate the free energy
of liquid, free energy of α and free energy of β and in principle, for example, you may
say, that Sir, at this temperature why should I calculate the free energy of α, it does not
exists there?
But at any temperature, at any pressure, whether the phase exists or not, you can always
calculate the free energy, excepting that you will be able to show that it is definitely
having a much higher free energy than the stable phase. The only question, that comes is,
how do I calculate it? For example, there are problems, for example, I take this pure
metal A, bring it to this temperature (T4). I have talked about this difference (free energy
difference between solid and liquid), what is this difference?
126
I call this as ΔG. ΔG = Gs - Gl. If I want to find out this ΔG, what do I do? I need to find
out the Gs, Gl, am I right? You may say, Sir, why should I find out this ΔG?
It is this ΔG, which is the driving force for solidification, when a liquid is and this
driving force in our nucleation equation if you remember, ΔG*, where does this driving
force come in the activation barrier? Anyone remembers nucleation kinetics? ΔG* star
is?
Student: ΔG = 4πr2γ + (4/3)πr3ΔGv
Yeah, correct, from that equation you come to what is called a activation barrier delta G
star. What is the expression for ΔG*?
ΔG* = (16π/3)*(γ3/ΔGv2)
And this ΔGv is this ΔG that you see here in G-x diagram, because that is the volume
free energy change. And so, as a result, if I know ΔGv , I would know ΔG*, of course, I
also need to know γ.
So, this influence has the activation energy. So, the higher the driving force, the smaller
the activation energy. If I take the liquid, bring it to below the melting point, to very
much below the melting point because it is much below the melting point, there is a large
driving force for the solid to come out. And because there is a large driving force, the
activation barrier will be very small and if the activation barrier is very small, if you
remember your ‘I’, nucleation rate, anybody knows how do we define nucleation rate?
I = A.exp(-ΔG*/RT)
127
So, the smaller the ΔG*, because this is negative and it is exponential, you would see the
larger the nucleation rate.
So, that means, if I take a liquid, bring it to below the melting point, to a larger extent I
am providing a large driving force because of which I need only a smaller activation
barrier and because of which I have a higher nucleation rate. And what is the use of it?
When I have a higher nucleation rate, I will get fine grain structure.
This is exactly what you see when you take a liquid metal, pour it into a metallic mould
and take the same liquid metal pour it into a sand mould. When I pour it into a metallic
mould, because of the cooling rate being higher, the liquid does not solidify at the
melting point, but gets under cooled to lower temperature and get solidified at below the
melting point.
And because of that you see all these happening, and all this if I have to calculate, what
is the nucleation rate? And then, without doing an experiment and finding out, getting a
128
micro structure, polishing it, putting it under microscope and measuring the grain size. If
I want to really calculate the grain size without doing an experiment, what I need is this
(ΔGv). If I have ΔGv and if I have γ, I can in principle calculate. We will talk about it
later.
Thank you.
129
Prof. B.S. Murty
Lecture No. #07

Evolution of Phase Diagrams
So last class, we were talking about free energy composition diagrams. We also looked at
how to get free energy composition diagrams, calculate them, also correlate them to phase
diagrams. We took for example, an example of an isomorphous diagram. And then, we
looked at how at various temperatures G-X diagrams can be obtained by calculation of the
free energy as a function of composition at a given temperature, and from that, we can obtain
what is called the solidus and the liquidus compositions at each temperature, because at each
temperature you can get what is called the equilibrium between the two phases - the liquid
and the α - and from that, in principle, we can generate the whole liquidus curve and the
solidus curve, basically taking all these liquidus compositions at various temperatures, getting
the trace of all these liquidus compositions at various temperatures; similarly the solidus
composition at various temperatures.
130
Today, let us try to see, what is the correlation between these G-X diagrams; particularly
looking at one particular parameter which is the, let us say in any G-X diagram. For example,
if you draw the free energy equation, if you remember, we talked about
Gα = XAGA + XBGB + ΩXAXB + RT(XAlnXA + XBlnXB) +
for a regular solution model of course. This is what we thought is then expression for the free
energy of any phase and in this, what is known is this is absolutely no problem, at any given
temperature, I can easily compute the last part; similarly, these two or the first two are
basically the pure metal components, so it is this Ω which decides to a large extend. What is
the type of phase diagram? For example, when you say an isomorphous diagram or a eutectic
phase diagram or a peritectic phase diagram or even a monotectic phase diagram, they are all
to a large extent decided by Ω. And let us see today, how this influences the evaluation of
phase diagrams and see how we can see various phase diagrams, with that particular Ω as
one of the component, for example, if you take an isomorphous system.
All of you know there are certain conditions for you to get an isomorphous system, what are the
conditions?
Student: Same crystal structure
It should have same crystal structure and anything more?
Student: Atomic radius should not be greater
Atomic radius difference should be smaller, and?
Student: Electro negativity
131
Electro negativity and anything else, other than electro negativity, atomic size, structure, valance.
So, these are what we call them as Hume-Rothery rules, which decides that you get an
isomorphous system and if you look at from the thermodynamic point of view, whenever we are
talking about isomorphous, we have
132
two phases: a liquid and a solid. So whenever I say a liquid and a solid, you can call it as an α
phase let us say, for both of these phases I can write an free energy expression like above
Excepting that all parameters will be liquid. So, when you write two expressions for liquid
and solid, you come across two Ω s: one for the solid, one for the liquid, and they need not be
always the same. For example, if you take a example of a eutectic system, in a eutectic
system, how do you define a eutectic?
Student: Liquid giving two solids
Anything beyond that?
Student: Invariant reaction
It’s an invariant reaction; let me see what are all the possible ways of expressing a eutectic.
One is as you said liquid giving two solids and it is an invariant reaction anything more.
What is a eutectic? Melting point decreases anything beyond, when I say system A B system
forms a eutectic, what do I mean? No, no let us not bring the thermodynamics at the moment,
whatever let whatever they are immiscible.
Student: They are not completely mixed
Mixed in what state, and what about liquid?

Student: They are completely soluble in liquid state and completely insoluble in solid state.
Yes, need not be completely insoluble, so the definition of a eutectic is A and B are
completely immiscible in the liquid state, they are either partially or completely insoluble in
the solid state. I hope at some stage you have read this, so if you have forgotten please brush
it up. This is the definition of a eutectic, yes am I right? So, in the liquid state they are
completely soluble, whereas in the solid state they are either partially or completely
immiscible, so what does that mean; that means the Ω for that particular A B system.
In the liquid should be different from the Ω in the solid, isn’t? In case of a eutectic they are
completely immiscible or they are partially miscible and that means, there is a soluble limit B
dissolves into A only up to certain extent, A dissolves into B only certain extent, that means
there is a positive Ω there, whereas in the liquid we are saying it is completely miscible; that
means, possibly the Ω in the liquid state is either 0 or possibly close to negative or close to
positive, but it is close to 0 because they are completely miscible. Whenever we say
completely miscible, it means the Ω is close to 0. Whenever we say they are completely
miscible, that means, they are neither repelling each other nor forming inter metallic
compounds. Whenever they form inter metallic compounds, ΔH is highly negative, that
133
means Ω is highly negative. So, whenever we see an isomorphous system, an isomorphous
system basically indicates you this, it indicates you this, that the
134
Ωl ~ 0 and the Ωs ~ 0, that is when you get a perfect isomorphous system.
The moment you see a phase diagram like that, you should understand that this is satisfied.
The moment you see that, one of this is 0, the other one is negative, for example, let us take a
case of Ωs = 0 and Ωl < 0. What does that mean? It means the liquid is more stable with
respect to the solid, liquid is more stable because liquid the the free energy becomes more
negative; that means, liquid is more stable with respect to the solid, that basically makes the
liquid as plunge down.
What you see the pseudo eutectic kind of phase diagrams (right side graph) that, were you see
this kind of a phase diagram is a result of this. It remains still isomorphous, but the liquid
stability regime is larger now. So the liquid exist over a range of higher range of temperature,
when compared to what you see in a normal case like isomorphous. So, the liquid as goes
down and that gives you an indication that liquid is more stable with respect to the solid.
And if you take another case, were Ωl = 0 and Ωs < 0. Then you get a congruent melting, this
is what you see, you get a situation something like this (left side graph). The solid is more
stable with respect to the liquid, so solid tries to be existing up to higher and higher
temperatures and you immediately see a phase diagram like this. So, the moment you see a
phase diagram, it gives you an idea even without doing calculations, it gives you an idea of
135
what would have been the thermodynamic parameters for the both the liquid and solid; now
we have looked at only the negative aspect of it.
Now if you think of a situation where, the Ω l = 0, but the

Ωs > 0 S is positive, that actually means that both the A and B there is no attraction between
them, there is some repulsion between them. And this basically means there is a tendency for
segregation A does not want to dissolve B to a large extent and B does not want to dissolve
A. And this actually leads you to what is called a miscibility gap in a phase diagram. For
example, if I draw a isomorphous you suddenly start seeing a miscibility gap (see the graph
above). If I try to correlate this with respect to what we talked about Hume-Rothery
parameters let us take one parameter such as a size factor.
You know that isomorphous actually happens when of course, this is more of a empirical
rules, but in a large number of systems this is valid, when the size factor which is called the δ,
δ = |(rsolvent – rsolute)/rsolvent|
if δ < 8 % that is when you have a complete solubility.
When you have, let us say a particular AB let us take a case of simple case of let us say,
copper gold they are completely isomorphous, instead of copper gold, I add copper
aluminium let us say both are fcc, but you do not see a isomorphous between copper and
aluminium. The reason could be because of many factors; and if we think that it is the size
factor which is affecting it, let us assume a case of A-B system with the size factor of around
136
10 % let us say. Then what you suddenly see is that, because the size factor is larger, the A
dissolves, whenever a solute goes into the solvent as long as the solute size is closer to that of
the solvent size, the strain that is generated is very -very small. Whenever you imagine a
case, where you have a lattice structure like this (see the atomic arrangement in the figure
above), imagine that these are all the atoms of the solvent and suddenly I want to put a solute
inside this.
Assume that these are all of the same size, though some of these do not look like that, ignore
that and now you think of a solute being put here. If this solute is of a smaller size than that
of the solvent then what you see is, it tries to in order to maintain the bond, it tries to pull the
atoms which are around it and that creates certain strain. Every 1 % of the solute that you
keep on putting you will keep seeing that, there is certain amount of strain that is generated
and as you keep on increasing the amount of solute, the strain that generated gets
137
keeps on increasing; and any structure can accommodate up to certain amount of strain and
the more closed pack the structure is the less is the amount of strain that it can accommodate.
That is why bcc structures can easily accommodate more strain whereas, the fcc structures do
not of course, again it depends on whether we are talking of what is called substitutional solid
solution or interstitial solid solution. Many of you remember that, carbon goes more into an
fcc structure than into a bcc structure; though fcc is more closed packed than the bcc you see
the amount, what is the solubility of carbon in fcc iron?
Student: 2 %
2 %. What is it in a point?
Student: 0.025
0.025 is the maximum in a bcc, so you see almost 80 times higher. This large quantity is
basically because you see the voids, because it is a interstitial solid solution the void size in
an fcc which is octahedral voids are much bigger than the biggest void, that is available in the
bcc which is the tetrahedral void; and as a result because the voids being bigger the carbon
can go and sit.
And in fact, if you look at the actual carbon atomic size, the size of the carbon atom is much
bigger than the void size and that is the reason why you do not see 100 % solubility of
carbon, in even fcc iron, only 2 %. And that if you compare with many substitutional solid
solutions, where you have isomorphous phase diagrams, you have you do not have an
isomorphous phase diagram with an interstitial solid solution. It is impossible, because the
only four elements that we talk about which can actually go into the interstitial solid
solutions, which are these elements?
Student: Carbon, Nitrogen, Oxygen, hydrogen
Yes.
These elements that we talk about none of them have their sizes, which are so small that they
can actually completely dissolve by 100 %. This is strain generated is sufficient enough that
even if you put large quantity of this, the strain can be accommodated. So, if you look at the
strain that it is, if you can think that there is a critical strain, ϵ* which a structure can
accommodate; and if you think that there is particular strain called ϵ1 which is the amount of
strain that is generated for every 1 % of solute being going into the solvent, you can say
ϵ1 x %solute = ϵ*
138
This ϵ* is the maximum that the solute can take. So if ϵ 1 is larger; that means if the strain
generated for every 1 % of the solute is larger than the solubility has to be smaller for a given
structure.
For example, if I take a copper as a solvent and keep on adding element such as aluminium,
lead, and all other fcc elements, depending on the size factor between the copper and any of
these solute atoms, you see the one which has a higher size factor, the solubility
139
limit of that particular solute will be smaller. Because the moment you cross ϵ*, then what the
solvent does is, it cannot take the solute anymore, so it starts rejecting it out and once it is
rejected, you get various precipitates. For example, in copper aluminium you get CuAl 2 and
things like that. So, in different -different phase diagrams we have different -different phases
coming out.
So, that limits what is called the solubility limit, so there is a critical strain any structure can
accommodate and in fact this is what, is also a reason, when you keep on putting more and
more defects into a system, a crystal structure can even break and become an amorphous. A
large number of people have studied this, what are the conditions under which a crystal can
become amorphous. For example, there are experiments where people have done, taking a
crystal and putting it into an electron microscope and keep on bombarding it with electrons,
high energy electron beam particularly people have done these experiments with a high
voltage microscopes.
I do not know whether you are aware there are microscopes which are 1 million volt
microscopes, 3 million volt microscopes, 3000 kv in comparison to what we regularly see
100 kv or 200 kv or 300 kv microscopes. We are talking of 3000 kv microscope, and when
you have such a microscope where the electron beam of that energy falls onto a sample
obviously, it disturbs the atoms. And as a result, you can have a large number of vacancies
that are generated inside and if this concentration of vacancies keeps on increasing beyond a
limit, the structure becomes unstable and it becomes amorphous.
So, people have seen cases where even an ordered structure, people have a very interesting
example, people have taken is Ni3Al which is an ordered structure, do you know what kind of
an ordered structure it is, anybody knows?
Student: FCC
It is an fcc structure which is ordered and when it becomes ordered what do you call that
structure, it’s called L12 structure, some of you who know about structures it’s called L1 2
structure and this L12 structure were you have the nickel atoms occupying the the face center
positions and the aluminium atoms occupying the body corners; because it is Ni3Al, 3 atom
of nickel and 1 atom of aluminium.
This structure when they have taken and put it into a such a high voltage microscope and with
more and more dosage of this electron beam, they saw that slowly this ordered structure
becomes disordered structure; that means, the super lattice changes into a disorder structure.
140
That means whatever super lattice reflections that you usually see in an ordered structure they
slowly
141
starts vanishing. If you in-situ you can see this in a TEM. So, keep on observing how the
diffraction pattern changes, you see the super lattice reflection slowly vanish and only the
fundamental reflections you see; that means, the ordered structure has become a disordered
crystal structure, but still it is, it is structurally ordered whenever you talk of disordering
many of you might know that there are two types of disordering we talk about - structural
disorder, chemical disorder.
So, whenever I talk about ordered compound I am talking of a chemical order there. When I
say structural disorder it basically means a crystal structure becoming into an amorphous
structure that is a structural disorder. When I say chemical disorder that basically means in a
in a chemically ordered compound, you have specific positions for each of the atom, let us
say in this Ni3Al just now I told you, that nickel occupies the face centered positions, the
aluminium occupies the body corners. If the aluminium and nickel randomly occupy the
atomic positions inside the crystal structure, the crystal structure still will be fcc, but it will
not be called an ordered crystal anymore, it will be a disordered compound, because nickel
and aluminium the probability of finding it at any place will be equivalent to 0.25 and 0.75
depending on the amount of it, but I cannot pin pointedly say that this is where nickel is
sitting, this is where aluminium is sitting, but in a ordered structure I can pin pointedly say,
this is where aluminium is there, this were nickel is there.
And so such a structure is a disordered structure chemically, but structurally still it is ordered
in the sense it has a unit cell, it has a fcc unit cell. So the difference between this chemically
ordered and chemically disordered is only in terms of the diffraction patterns, you can
recognize, you will see a extra spots called super lattice reflections in a chemically ordered
structure and in a chemically disordered structure you will not see these extra spots; and now
if you take that and if you continue to bombard it people have seen that disordered Ni 3Al,
finally becomes an amorphous structure with 75 % of nickel atoms and 25 % of aluminium
atoms, which are randomly their sitting inside without any unit cell. So the unit cell breaks
from an fcc unit cell to a structure which is similar to that of a liquid, and people have seen
this happening and that is all because there is a critical strain that the structure can take, so if
you look at from a physical point of view it is like this.
But if you look at from the thermodynamics point of view we are looking at only from this
and you know that the more the strain is, when will the strain be more, when the size
142
factor is more, so the larger the size factor you will see the larger will be the Ωs positive.
So, you can see that the moment size factor increases if the δ increases then you would see Ωs
increases, higher and higher repulsion between the A and B as because as the size factor
increases then the strain increases, if the strain increases; obviously, the repulsion increases.
So now, we do not look at this part we simply look at what happens if the Ω s increases, so the
moment Ωs is positive what you see is suddenly you see a miscibility gap (Graph 1), as long
as the size factor is less than 8 % you will never see this miscibility gap, the moment size
factor crosses 8 %, you start seeing a miscibility gap; that means what, that means, there is a
limit on the amount of B that can go into A.
So, only up to this much the B (point P, Graph 1) can dissolve into A and only this much is A
can dissolve into B (point Q, Graph 1). And whenever you see a solubility limit of A in B and
B in A, you would see that this solubility limit keeps on changing as you change the
temperature, why?
Student: Thermal energy vibrates the atoms lattice position, because of which it can
accommodate
So, yes the ϵ* would change other function of temperature, the strain that the structure can
accommodate because a lattice expands as the temperature increases, isn’t it? Because the
lattice has expanded, so it can actually accommodate more and more strain and as a result,
you would see as you increase the temperature, the solubility limit keeps on increasing; that
means, the amount of B that can dissolve into A keeps on increasing upto a particular limit
(point R, Graph 1), beyond which they are completely soluble, A becomes completely
soluble in B and B becomes completely soluble in A. Similarly, on the other side also the
solubility limit of A in B also keeps on increasing and you would see that, it will finally meet,
and this is what is the final temperature (T R), above which, you can say completely they are
miscible and for a given Ω TR is going to be fixed.
Now, for example, if you think of a particular value of Ω let us say talk about 10 kJ/mol and
now you increase it to 20kJ/mol, both being positive what should happen now? When you
increase the Ωs from 10 kilo joules to 20 kilo joules what should happen?
Student: widen
Yes the miscibility gap should widen (see the graph in below image), why? Because the the
A and B are more and more repulsive, they are no more attract, so as a result the solubility of
A in B and B in A should further go down, because A and B they repel each other to a larger
143
extent and as a result, what you see is that this miscibility gap increases and the moment
miscibility gap
144
increases, I mean in another words that when I say that means, size factor (δ) has further gone
up, when the size factor further goes up this is what happens, if the size factor goes up here
what happens is ϵ1 will go up.
If ϵ1 will go up for the same structure because we are assuming that our solvent is the same, if
the solvent is the same, solvent has a fixed structure, so the total strain that it can
accommodate is going to be fixed, so the ϵ* gets fixed the moment I fix up my solvent, what
I am changing is only the solute, I am adding different -different solutes each of them have a
different size factor that means, each of them have a different Ω positive. I am adding a
different solute every time with a higher size factor such that it is Ω s is more positive and that
means, the moment this is higher because ϵ* fixed the solubility has to be lower.
Student: Sir then why the temperature is increased?
That’s what, so we will come to it that means, what you would see is that, because ϵ* and at
the same time you would see that the solubility limit and both the sides has come down that
means, the strain that that each % of A, when it is added to B or each % of B when it is added
to A is increasing, ϵ1 value is increasing as a result for the same amount of this structure,
when you are increasing the temperature this ϵ*fixed, but as you increasing the temperature
because this strain is larger, the temperature to which you have to heat, so that the strain gets
accommodated has to increase, as you keep on adding an element with higher and higher size
factor, the temperature above which they completely dissolve into each other, should keep on
increasing and that is why you would see that not only that the solubility of A in B and B in
A decreases, but also you would see, the miscibility gap will go to a higher and higher
temperature and you would see that, the moment you start seeing bigger -bigger miscibility
gaps, it basically indicates that this is the thing, but simultaneously.
145
You would see something else also happens in the phase diagram, whenever you consider
Ωl = 0 and Ωs > 0, you would see that not only that you develop a miscibility gap, but
something happens to the liquidus curve also, what you would see is because the Ω s is
positive and Ωl = 0, so with respect to the solid the liquid is more stable and as a result what
you would see is that, in addition to this you would see that this liquidus curve starts plunging
down, because with respect to the solid, now the liquid is more stable, , liquid is stable over a
larger temperature range and once this happens, you would see that there is a particular
condition at which both these miscibility gap and the liquidus merge into each other, and
once that happens you would see a situation, where the phase diagram actually becomes and
you would see both of them merging into each other and once they merge into each other
what you end up is nothing but a eutectic (see the bottom phase diagram), you would see this
is.
146
This is nothing but, the eutectic point, these two are the compositions where you have
actually the on the maximum solubility of B in A and maximum solubility of A in B. So you
can see an isomorphous system suddenly has become a eutectic system, just because I have
increased a Ωs to a critical value and this at what Ω s this happens actually depends on system
to system, each system you have to calculate and, but at a particular Ω you would definitely
see this happening for everything and now, if you keep on increasing it further, more and
more Ω what should happen to this? You would see two things happening; let me see what
are your answers.
Student: Liquid curve will still come down
Yes, so what does it mean, it means that the eutectic temperature goes down, this is what we
call them as deep eutectics, whenever you see a deep eutectic, what we mean by deep
eutectic, in fact people even quantify this eutectic, because it is very important when people
talk about glass formation, I will tell you an example for example.
147
If you take two cases, take a two cases, one were the eutectic is like this (Graph 1), another
were the eutectic is like this (Graph 2), we call this (Graph 2) as a deep eutectic when
compared to this (Graph 1) and in fact, people try to quantify this in terms of what is called
taking the two pure metal melting points, calculating drawing them by a straight line and
finding out, what is the liquidus temperature as a rule of mixtures of these two that means, at
the eutectic composition (point P, Graph 1).
What should be the liquidus temperature, if a liquid A and liquid B mixture is available let us
say and that with respect to this, so we call it as (T lmix - Te) / Tlmix, as a parameter which
people define it as a glass formation parameter, why? It is important is when does a glass
form, we said glass forms whenever a liquid is under cool to an extent, that it is brought
below the Tg (See graphs 1 and 2). In any system the T g is somewhere here, glass transition
temperature can always be calculated for a given system and usually a glass transition
temperature is somewhere around 200, 300 ºC depending on the particular system of course,
there are some systems it can be even lower, when the liquid is very highly stable and then
you would not see the glass formation easier, so if you think of a glass transition temperature
somewhere if you have a phase diagram like this (Graph 2), it means the liquid for it to be
under cool to get a temperature to Tg, the the temperature that you need to under cool is much
smaller than here (Graph 1) where the under cooling that you need to provide is very large;
that means, the cooling rate that you need to provide for the liquid to under cool to the T g is
much larger here, when compared to the cooling rate that you need to provide for the liquid
to under cool to Tg, that is why in a deep eutectic systems, we can always say glass formation
148
is easier, so that is why these are called easy glass forming systems, these (shallow eutectics)
are not so easy glass forming systems.
So, even with slower cooling rates people are able to achieve glasses. For example,
something like a silicate glass that you see, a silicate glass can easily form a glass, even at a
very -very low cooling rate as low as even 0.1 K/s whereas, a normal metallic glasses you
need to go to 106 K/s, to in order to be able to get that, that is the reason is because most of
them, but even there also if you look at phase diagrams look at a number of phase diagrams,
wherever you see a deep eutectic for example, the first metallic glass, anyone knows which is
the system where people have got first metallic glass? Gold silicon is the first metallic glass
in 1959, Paul Duwez has obtained it and this gold silicon if you compare it with simply like
something like an aluminium silicon, today evening if possible go to the phase diagrams
handbook such as Massalski and try to compare gold silicon with aluminium silicon, you
would see that Au-Si is the much deeper eutectic than this, and that is the reason why you get
a glass more easily Al-Si than that. You take a simple Al-Si liquid and of the eutectic
composition simply quench it, you would not be able to get a glass there whereas, Au-Si if
you quench that liquid, you can get a glass, so that is basically because this is deeper eutectic
and people have given a condition that, if (T lmix - Te) / Tlmix > 0.2 by looking at a number of
systems, people have said that if this value is greater than 0.2, than you can say it is a easy
glass forming system, otherwise it is not so easy glass forming system of course, these are
thumb rules like Hume-Rothery rules, more of empirical rules, based on looking at a number
of systems and people have come across certain values like that.
So, this is one of the glass forming criteria, there are number of them we will talk about them
as we go along a little later. So just because this concept came I had to tell you this, so as you
can see, that if I keep on increasing Ω you would see this (liquidus) goes down and as a result
you would see that the liquidus temperature goes to lower and lower value that means,
eutectic temperature goes to lower and lower value and anything else that happens?
Student: Solubility goes down
So that means the solubility limit of A in B and B in A also starts decreasing further because
the miscibility gap becomes wider and wider,
149
So solubility limits keep on decreasing to an extent that at some value, you would see that the
solubility limits vanish something like this, you would see a phase diagram become
something like this, where there there is no solubility of B in A or A in B, it is like
L  A + B instead of L  α + β, there is the difference between the two isn’t? When I say L
 α + β that means the two are terminal solid solution that means, there is a solubility of A
in B and B in A, so instead of a phase diagram like this (with limited solubility), you end up
into a phase diagram like this (with zero solubility).
Something like this, where there is no solubility of A in B. So the moment you see these
things, you can get an idea of these and people actually even do the calculations to be able to
find out whether it really happens or not, and a number of people have already done this and
as I told you, even we have a journal which is available called CΑLPHAD which regularly
does this kind of calculation. And now let us see what happens, if the Ωl is positive.
Student: Pure iron goes from BCC to FCC, what exactly happens there?
It is because of the magnetic transitions, you know in principle that iron is bcc at room
temperature you know and then it changes to fcc and then. In fact, if you carefully observe
most of the actual metals, when you go to a higher temperature the phase that is more stable
is the one with a more open structure, you regularly see this happening in in. For example
titanium, high temperature phase is a bcc phase, in most of the structure the high temperature
phase is always a bcc phase even if you look at things like ceramics, barium titanate you take
150
high temperature phase is always a cubic phase, low temperature phase is a tetragonal phase,
even if
151
you take the zirconia, titania and yttria all these things you see this kind of things were as
here.
Because there is a magnetic transition, that interferes with that, so as a result you get that and
because that phase is actually not a closed pack structure is not really stable at high
temperature, so when you further heat it to high temperature it comes back to an open packed
structure, which is the bcc structure. So the intermediate fcc structure is a result of actually
magnetic transition in this material, otherwise you would not have seen this.
So now, if you look at this case of, if I assume a case of like this which I cannot actually
assume, this actually looks absurd to anybody, but why should it look absurd, when I say
Ω s = 0, Ωl > 0. You say that when you say Ω l > 0 that means, A and B are repelling each
other in a liquid state. If A and B repel each other in a liquid state automatically they actually
repel each other in the solid state also, it’s very -very difficult to imagine that A and B are
having a ideal kind of a situation, they neither repel nor attract in a solid state, but in the
liquid state they repel each other, because liquid is a more open structure. In a more open
structure actually A and B should dissolve more easily, that is why we see whenever I say
eutectic by the way when we talk about this, the eutectic formation from an isomorphous, you
can also get a peritectic in the same fashion, excepting the difference is you get a peritectic
when the melting points of the A and B are largely different.
152
Same way you can construct and you can get a peritectic diagram from an isomorphous phase
diagram. For example, whenever you see a peritectic a good example is copper zinc let us
say, copper zinc is a classic example of what people call it as cascade of peritectics. You have
a number of peritectics in a copper zinc, copper tin is also another such example. The reason
is one of the element copper melts at a much higher temperature than the tin which melts at a
let us say 232 ºC or so. So when you see this large differences in the melting point, then
instead of a eutectic you see a peritectic happening. And you would see if you go back to any
of the phase diagrams, you will see this happening. Now coming back here, you see that if
this has to happen hypothetically anything can be assumed, if you assume this hypothetically
what it means is, if the Ωl is positive that means, there is a miscibility gap in the liquid, have
you heard of miscibility gaps in the liquid? Where do you see miscibility gaps in the liquid?
Monotectic phase diagram is one; of course the extreme of this is a syntectic. So monotectic
for example, if I draw a monotectic phase diagram, what is a monotectic phase diagram, this
is the monotectic.
That there is a liquid and here you have l1 plus l2, so what does this reaction means l 1 gives
you, what is a monotectic reaction, how many of you remember?
Student: liquid gives another liquid and solid.
Yes l1  l2 + α, this solid can be an α, this can be a pure metal depending on the situation,
most of the cases the solubility of B in the solid in many of the monotectic systems is almost
0, so we can simply say that this gives you a solid A let us say, A or B depending on which
side the monotectic is so that means, this phase diagram can be replaced by something like
this. That means there is no solubility on this side and as a result you can say l 1  l2 + A and
now whenever, you say liquid 1 gives you a solid plus another liquid that means, below the
reaction and again this is also is an invariant reaction like what we have seen earlier, why it is
an invariant reaction?
Student: 0 degree
Why 0 degree of freedom? Three phases we have and it is a binary system. It is a binary
system, numbers of components are two and you have three phases involved, why I say three
phases because there are two liquids, what is the difference between these two liquids?
Student: Solute concentration
Concentration, the composition of l1 is different from composition of l2 both having the same
structure, both have the same structure, they are all both of them are liquids, so they have
random arrangements of atoms, but they have two different compositions.
153
And the moment I say two different compositions, there is a boundary between them and as a
result I have to define it as a separate phase, like when I say a miscibility gap in the solid
state, what is there between in within the miscibility gap? Two solids, we call them as α 1 plus
α2 and interestingly both of them have the same structure, because on the both the sides two
pure metals have the same structure otherwise you would not have got a isomorphous system,
the fact that it is a isomorphous it indicates that A and B are the same structure, we are only
talking about A and B having the same structure, but with different atomic sizes such that you
are getting a miscibility gap or with whatever may be the parameter, whether it is atomic
sizes or no valency or electro negativity, but ultimately giving to a positive Ω s. We have only
taken a size factor as one of the parameter because it is easy to see, because it is easy to
understand based on the strain whereas, the other two I can also choose electro negativity or
valency, but again the same thing has to basically everything gets reflected to this, if this
becomes positive you get though A and B have the same structure, you get a miscibility gap
and in within that, both the α1 and α2 that you see have the same structures, but excepting that
their compositions are different that is why, if I draw at any given temperature a tie line
across the miscibility gap, the intersection points will represent the composition of α 1 and
composition of α2 exactly similarly, if I draw at any given temperature here, these two (points
P and Q in above diagram) represent this is the composition of l 1 (P), this is the composition
of l2 (P), and at the peritectic at the monotectic temperature, liquid of this particular
composition (P) is giving you another liquid of this composition (Q). And because below the
monotectic there is a liquid that is available, because one liquid is giving you another liquid,
this liquid which is coming out of the monotectic has somehow to solidify, so as a result
many monotectic phase diagrams, you would see in addition to a monotectic reaction another
additional reaction where the liquid, which comes out of the monotectic solidifies by either a
eutectic reaction or by a peritectic reaction. So, that is why you would see here, an extension
of it, you would see a either a eutectic reaction so that means, here the l2 that is coming out,
so here you can say it is A + l 2 below the monotectic reaction. So, liquid gives you and this is
what, what is this phase field? A + l1, so you will have A + l2 and this l2 will undergo a
eutectic reaction or a peritectic reaction in this particular case it is a eutectic reaction. And as
a result below that eutectic reaction, what will you have, A + B simply, that is how we can
see whereas, if it is a eutectic reaction or a peritectic reaction, you need not have to have
another reaction. There can be phase diagrams where you can have a number
154
of various reactions like iron carbon where you have a eutectic, a peritectic and eutectoid, but
there are also systems where you have only one simple reaction for example, Al-Si, it has one
single eutectic nothing else, Mg-Si you have two eutectics nothing else.
So, there are also systems like that, so you will see whenever you see a monotectic you will
see things like this, and this monotectic is caused by what, the Ω l > 0, whenever Ωl is positive,
there is a miscibility gap inside the liquid and once there is a miscibility gap inside the liquid,
you lead to a monotectic reaction and that is how, we can understand and the more the Ω l
becoming more and more positive it finally, leads you to a syntectic kind of a case, case
where there are completely insoluble in the liquid state and also insoluble in the solid state,
you rarely can think of a situation, where they are completely insoluble in the liquid state, but
soluble in the solid state, reverse is possible, that you can have a complete insolubility in the
solid state, but solubility in the liquid state which is what is our eutectic, but you would not
see a situation that is why whenever there is a syntectic there is a complete insolubility in the
liquid state, you also see that in the solid state, there is complete insolubility and that is how
we can see just based on these two parameters we can evaluate these diagrams will stop now.
155
Prof. B.S. Murty
Module No. # 01
Lecture No. # 08
Evolution of Phase Diagrams, Miscibility gap
Last class, we were talking about the evolution of phase diagrams based on
thermodynamic principles. We took up two parameters, which are basically the Ω s and
the Ωl, which is nothing but the interaction parameter in the liquid between A and B and
interaction parameter in the solid between A and B, and used them to basically see how
phase diagrams evolve from a isomorphous system. And you get an isomorphous system
when both of them are close to 0, and as you start varying one, keeping the other one
fixed, how does the phase diagram evolves, so we will try to see that.
And today, I just had brought you an image from one of the reference as you can see
here, this is from T.B. Massalski, published in Metall. Trans. in 1999, which gives you a
flair of how really the phase diagrams evolve; yesterday I was talking about that. You
can see
156
here, on the left axis you have the Ωs and on the top, which is the X axis is the Ω l, and
you see how as you start varying these things, how various from phase diagrams come.
And you see, when you take both of them as 0, that is what ends up as the isomorphous
system. And if you keep the Ωs as 0, keep decreasing the Ωl, make Ωl more and more
negative, you see when it is -10 and when it is -20, what you see? You start seeing the
pseudo eutectic phase diagram evolves, so the liquid becomes more and more stable, so
the liquidus curve starts plunging down.
And, as you start seeing, you see more and more of liquid stability there. And on the
other side, keeping the same, the Ωs = 0, and let us say, increase the Ω l to positive side,
again you see that, you start seeing that, because the moment Ω l is positive and Ωs = 0,
that means solid is more stable than the liquid, and as a result solid stability regime
increases and you start seeing a congruent melting case develops. And similar thing can
also happen when Ωs is negative. For example, keeping Ωl as constant at 0, and if you
make the Ωs negative, this is what you see in the diagram below, where you see the
moment you have -15 for the Ωs, for the same Ωl = 0, so you see the congruent melting
starts evolving.
So, You can have a pseudo eutectic type, when the liquid is more stable; when the solid
is more stable, you start getting what is called the congruent melting, which is an
indication of the formation of inter metallic’s. And if this becomes stronger, that means,
if the Ωs becomes more and more negative, you would actually see that inter metallic
starts actually evolving inside the phase diagram. And this is as far as either Ω l being
negative or Ωs being negative, but if you think if Ω s is positive, what is going to happen?
Keeping the Ωl = 0 and if you start making Ωs positive, for example you see, this is +15
and everything is kJ/mole, all the numbers that we are talking are basically the enthalpy
units because Ωs is nothing but, the enthalpy units, because it is an interaction parameter
and that multiplied by XA and XB will give you the actually enthalpy of mixing.
So, it will have the same unit as the enthalpy which is nothing but kJ/mol. So, you can
see here, if you keep Ωl = 0 and then make Ωs = +15 kJ/mol, you
157
start developing a eutectic. And we saw yesterday how this eutectic actually develops
because of the miscibility gap. And the miscibility gap and the liquidus plunging down
side downwards and both of them meeting, both the liquidus and solidus and the
miscibility gap meeting. Basically, what you see down is this is the miscibility gap and
this is the liquidus both of them merge into each other, and you will get that. And now, if
you keep Ωs as constant and let us say make Ωl more and more negative, you see that
liquid is becoming more and more stable with respect to the solid and as a result the
eutectic temperature starts falling.
You will see the same thing can be also seen, if you want to increase the Ω s keeping the
Ωl the same. For example, if you keep the Ωl = 0 and increase the Ωs = +30 kJ/mol, what
you see here is that this, the terminal solubility on both the ends vanished and you will
also see the liquidus coming down. Like this we will be able to see, how various and
similarly, if you keep Ωs = 0 and make Ωl = +30 and make this more and more negative,
similar thing happens. So, as the liquid becomes more and more stable by making the Ω l
more and more negative, you see that the liquidus line comes down, that means the
eutectic becomes more and more deep eutectic.
Yesterday, I also gave you an idea of how to quantitatively measure, how deep the
eutectic is with respect to basically, we take the melting points of the two pure metals
and then find out what is the melting point of a mixture of the two pure metals at the
eutectic composition, and take down what is the eutectic temperature; and the difference
between these two will give you a measure of the deeper the eutectic it is, and of course,
it has to be normalized with respect to the melting point of the mixture and that is what
we have seen in the last class.
Now, if you see that, if the Ωl becomes positive, so far we have seen whether the Ωs is
positive and the Ωl is either 0 or negative, and if you make the Ω l, start making the Ωl
positive, then you start seeing the development of the monotectic phase diagram, That
you start seeing a miscibility gap inside the liquid, not only the solid also will have the
miscibility gap, but also the liquid also will have the miscibility gap and you start
developing either the monotectic or the syntectic kind of phase diagrams.
And that is how you see all the phase diagrams can be, in fact though this is
schematically shown in principle, one can calculate all these phase diagrams simply by
just varying the
158
two parameters. And whole calculation is based on what we have seen, the equilibrium
principle, which is nothing but, the μAα = μAl.
And if you use this as the starting point for your equilibrium, from that, you can find out
at each temperature, what is the liquidus composition, what is solidus composition, and
from that, we can come across all this. And in case there is a third phase to be evolved,
for example, in case of eutectic or in case of a monotectic, you have three phases coming
into picture, and you would see how to deal with those three phases. By considering third
phase, and that is what we are going to see today, what happens, how do I consider the
third phase?
At the moment, we have been talking about only two phases, Ω s and Ωl, and how do I
consider a third phase. And that is what we will see today, and that also give you an idea
of how to understand the miscibility gap.
159
Yesterday, somebody was asking me, sir if this is the miscibility gap, how does this gets
closed? What is thermodynamics behind it? We will also see today, how what is
thermodynamics behind this, the miscibility gap getting closed at higher temperature and
why does it get closed. To understand this, what we see is, how does the free energy
curve looks like at that temperature?
For example, you take a system which is an isomorphous system (Graph 1) and draw a
free energy curve at that temperature (T1), how does the free energy curve look like? If I
draw it, this is G
160
verses X (Graph 2) we are talking and at this temperature (T1), what is stable? Solid α is
stable, this is liquid, this is liquid + α , so α is more stable, very easy. So, as a result, I
will see a free energy curve for α which looks like this and of course, there will be a free
energy curve for the liquid which is something like this, this is liquid, this is α (Graph 2).
And this free energy curve is expressed in terms of the free energy expression that we
have given earlier, where
So that expression which we have given can be used to calculate this free energy. This is
all fine, the moment you see a miscibility gap here, how does this free energy curve is
going to change?
The difference between isomorphous and the system with miscibility gap is that, in the
latter the Ωs is positive, am I right. So, there is a positive interaction between the A and
B; that means, A and B do not attract each other, they repel each other. The moment Ω s
is positive, ΔHmix is positive. Because, ΔHmix = ΩXAXB, as long as I assume that regular
solution model is valid. So, if I plot the ΔH mix as a function of composition, you would
see the ΔHmix curve will be something like this (Graph 3).
So, once it is like that, and you have - TΔS mix, which is the second part of the free energy
of mixing will always be negative (Graph 3). The only difference is, depending on the
temperature, this value is going to depend, how large absolutely it is, or how small
absolutely it is, depends on what is the T. Because, ΔS mix is only a function of
composition, it is a symmetric function we have seen that
ΔSmix = -R(XAlnXA + XBlnXB)
So, as a result, this is going to be solely dependent on the temperature and as a result, at
room temperature if I am considering, where the temperature is not too high, you would
see that possibly this particular value, TΔS mix is lower than the ΔHmix. For a certain case,
as long as the maximum value of ΔHmix is smaller than the minimum value of TΔSmix,
you would see that the ΔGmix = ΔHmix - TΔSmix. If ΔHmix is positive, which is what we
have seen here, but its absolute value if it is smaller than TΔS mix, then the ΔGmix is still
negative. If the ΔGmix is still negative
161
then the free energy curve will be having a curve downwards as we have drawn here
(Graph 1), it will be like that. The moment ΔH mix value is higher, the free energy curve
would look something like this (Graph 3). Assuming that the G A and GB are 0, we can
always assume there is the standard state, G A and GB is 0, it’s an extreme case of course,
it may not be 0. So if I assume that GA = 0, GB = 0, that XAGA + XBGB term vanishes for
schematic case, we can just assume it to be 0. So, once that vanishes, then I can write
this as Gα (middle curve, Graph 3) , but at the same time you may say, sir why not it go
the free energy curve go like this (Graph 4), when this is going to be higher than this,
why should it have an inflection like that?
Why not it goes like this (Graph 4)?
That is where, we have seen in one of the pervious classes, that the free energy
expression which is here, even if you assume that the pure metal terms are not there,
Gα = RT(XAlnXA + XBlnXB) + ΩXAXB
and differentiate this and find out
α
∂G
∂xA | =−∞
X A=1
X B=1
162
So you see that it is -∞ and as a result, that means the slope of the free energy curve as a
function of composition, irrespective of what is this Ω value. However, large positive or
however large negative it is, this slope is bound to be negative and it is - infinity. And
because, it is - infinity on both sides of the free energy diagram, in terms of the
composition you see that the free energy has to fall because, the free energy curve should
have a slope which is - infinity towards this end and towards this end, the only difference
that you see when you keep on increasing Ω positive. If G α is negative, there is no
problem at all; if this is negative, what is going to happen? This curve is going to
become more and more deeper down.
You will see that this curve will go down more and more, if Ω is negative. Ω being
negative is not a problem at all the free energy curve is not going to change much.
Expecting that the shape remains the same, the absolutely values are going to change.
But, if Ω value is positive and depending on how large this value is, you would see that
this free energy curve could in an ultimate case, could even be like this or could even
become drops a little and then goes becomes like this (see right side graph). If the Ω
value is highly positive, then you see it drops by a small extent and then starts rising and
this drop is because of slope being -∞. So, you would see in all the cases, it will never
start from 0 and starts rising.
163
So, it has to have a drop and how small it is, how measurable it is, depends on the sensitivity
of your measuring instruments. But, in principle, thermodynamically and in fundamentally it
has to come down and that is what you see here. And the moment you see a curve like this, I
can start talking in terms of what is called a common tangent because, this curve is one
single curve but it has an inflection (see G α curve). The slope of it is changing at these
points. In fact, we will come to these points later, they are called spinodal points and when
we talk about spinodal transformation, we will try to understand that in greater detail. But,
the moment you see a free energy curve like this, then I can start talking about.
If I start drawing the tangents (PQ), you will see at one stage, there can be a tangent,
which is something like this, which is a tangent for both the ends. And this particular
tangent where, where you see the 2 points there, these 2 points (points P and Q) are
nothing but, the equilibrium compositions of the original phase which is now
decomposing into 2 phases. Because of this limitation of the Ω, Ω being positive, the A
and B are not able to dissolve into each other. So, they are force to segregate the A
dissolves into B to some extent. B dissolves into A to some extent and they are not able
to completely dissolve to the extent to which A dissolves into B and B dissolves into A
can be obtain from this common tangent. So, that means, basically what you see, these 2
points at this temperature (points P and Q, left side image) are nothing but, these 2 points
in this free energy curve.
And that is why we call this as composition of α1 and composition of α2, incidentally you
call these as α1 + α2, and you do not call them α and β basically because, both are same
structures. It is actually one single phase which is decomposing into two, with the
structure remaining the same, excepting the composition being different. So, as a result,
the α is splitting into two phases, one having a A rich composition, another having a B
rich composition; and that is how you see that this is happening. We will know more
about this as we go to the spinodal and try to understand why spinodal occurs and things
like that. Now, if you look at what should happen if that is the free energy curve which is
leading to this situation of α 1 having this decomposition, α 2 having this composition
at this particular temperature.
And mind you, I have to also say that if I am on the terminal sides (where G α is
negative), what does it mean? You remember that, maybe I will draw this in a greater
detail for you to understand, why this splitting has to occur?
164
For example imagine a case, something like this and you have a common tangent, which
is something like this (Graph 1). And mind you, common tangent does not mean it has to
be always horizontal, the free energy curve could be like this (Graph 2); though
generally, for most of the in the text books they show like this (Graph 1). It is not that, it
has to be a horizontal.
(For the following discussion please refer to Graph 2)

So, if you take a case of this composition (C0), composition to the left of it (C1), and
draw a tangent (k), you would see that the whole free energy curve, the whole free
energy curve is above this tangent. And in comparison, if I take a composition here (C2)
and draw a tangent (l) to that. You see that if this is the tangent (l) for that particular
composition (C2) I am taking a α of that particular composition, You see that, beyond
this particular composition (C3), you see that there are certain compositions (hatched
region) which have a lower free energy than that corresponding to this tangent (l).
That means, these compositions are stable compositions because their free energy is
lower. So, they can in principle come out of this particular composition (C2). So, if I
take a α of this composition, in principle α of that composition is not stable. Why
because, in principle, I can even show what is the free energy of this is this, free energy
of this is this and if I take a 2 phase mixture of α1 + α2 of the same composition, the 2
phase mixture has a free energy which is given by this (m). Because this is the common
165
tangent, if I draw that line (m), any 2 phase mixture, whenever you draw a 2 phase
mixture, the free energy of a mixture can be always found out by the line joining the two
free energy curves, free energy points. So, if this is the free energy of the α 1 of the
equilibrium composition and if this is the free energy of the α2 of the equilibrium
composition, then the line joining these two will give you the free energy.
If I have a phase mixture of α1 and α2 with different amounts of α1 and α2, what is the
free energy of such a mixture can be always obtained by this line. It is a simple rule of
mixtures and if that is the case the 2 phase mixture here has a lower free energy than this
(C2) and because 2 phase mixture has a lower free energy than this, you would see
instead of a single phase existing for this particular alloy composition it would like to
exist as a 2 phase mixture. And that is what you see if you draw a tangent, you see that
the phase here (hatched region) which is a B rich phase, having compositions all these
compositions, anyone of the composition can in principle nucleate from this particular
alloy (C2) because, this alloy has a higher free energy than the 2 phase mixture. Which
one will actually nucleate we will see when we talk about what is called the spinodal.
We will know more details about it, which phase should actually form and in principle
usually the one that forms is the one which has the highest driving force. And the highest
driving force can be obtained by the difference between P and C4; this will give you the
highest driving force. As a result, that is the composition which would try to nucleate.
We will know more about it when we go to spinodal, but at the moment you can
understand that, if I take this particular alloy composition (C1), a composition of this,
you see that none of these particular compositions have a free energy lower than this line
(k). As a result, none of them can nucleate from an α of this composition (C1) that means
what? The single phase α is stable at this particular composition (C1), but if I take this
particular composition (C2), single phase α is not stable, it wants to split into two, which
is α1 + α2 and α2 will nucleate out of that α .
And whatever α is left out is what we start calling it as α 1 because, the moment you see
that α2 of this particular composition nucleates (C4), then immediately there has to be an
equilibrium between that α2 and α . For example, if you have an α from which α2 has
nucleated this α2 has to be in equilibrium with α . What particular composition of α that
can be in equilibrium with this α2 if you try to look at. If this is α 2 (C4), the composition
of α that can be in equilibrium with α2 is only
166
this composition (C0) because, the 1 that can come as a common tangent. When you
draw the common tangent, you would see where this common tangent touches the α ;
that is, this composition (C0) that alone can be in equilibrium with this α 2. You cannot
have this particular composition (C2) in equilibrium with this α2 because, I cannot have a
common tangent between that and this, it is not possible to have a common tangent.
That means what the moment α2 comes out of α , the composition of α surrounding this
particular α2 changes (see right side image). And actually it becomes poorer in B and
richer in A. Why does it become poorer in B? Because, it is α 2, which is rich in B is
coming out of α . As a result, the α which is surrounding that will loss the B atoms
because, all the B atoms are coming together and forming a α 2 phase. And because α22
phase is coming out the α , which is around that will lower its B atoms and as a result,
the composition of α shifts towards pure A side such that, that particular α becomes in
equilibrium with the α2. So, that α which you see it around this α2 we will start calling it
as α1. So, now that α , which is so far away from this α1 and α2, you may still have the α
.
Having this particular composition until the whole reaction is complete. At the beginning
of the reaction, you will see the α 2 will come out and around this α 2, always you will
have a α having a composition which is, which is decided by this common tangent. And
that is what we will call it as α 1 and this is what we call it as α 2; α2 is B rich and α1 is A
rich and this is how you see this happening. And similarly, if I choose a composition
here for the α , again you will see that if I draw a tangent there, you will see all these free
energy points, that means free energy of any of these composition is higher than this
tangent, so as a result, nothing can come out of that particular α . So, single phase α is
stable at that, everything we are talking at a particular temperature T.
Because free energy composition diagram is always drawn at a fixed temperature, we are
talking at at a particular temperature. So, at this temperature to the left of this point (C0)
single phase α is stable, to the right of this point again (C4), single phase α is stable.
That is what this free energy curve tells me and between these 2 points (C0 and C4)
single phase α is not stable and that is why whether I choose this composition (C2) or
choose this composition (C5), if I draw a tangent (o) at C5 you will see all the
167
compositions to the left of this can precipitate out of this (C5). So, that means, from α ,
α1 will precipitate and everything will be inverse of what we have talked so far. Once α 1
comes out, α1 is A rich phase as a result the A atoms from α have to come together and
form the A rich α1 phase, as a result the A atoms in the α surrounding the α1 will deplete
in the A atoms and that becomes the α2.
So, you will have α1 surrounded by α2; in the previous case, it is α 2 surrounded by α1.
So, depending on whether you are taking a A rich composition or a B rich composition,
this is what would happen. This keeps on happening until you reach the whole
composition of α1 and α2 reach these two compositions, which is what the free energy
concept tells you. At that temperature that these are the only two compositions which can
be in equilibrium, no other compositions can be in equilibrium. So slowly, you see that
the α composition changes and the α 2 composition remains what it is, according to the
phase diagram, according to the free energy composition diagram, and the remaining
whole α vanishes and you will have α2 growing at the expense of α . And you would
have α2 and α1 at the end of this.
We will know more about it when we go to the spinodal, but at the moment, what you
need to understand is that this is how we will know that these. Any composition that I
choose between C0 and C4, you will see all these compositions have a higher free
energy than a 2 phase mixture. The 2 phase mixture composition is given by this (C0 and
C4), and as a result, a 2 phase mixture is more stable in this, between these two points
and these two points (C0 and C4) are called bimodal points. We will talk about this
again, when we come to spinodal decomposition. So, what you see is that, you will see
between these two there is the α1 and α2. And we call it as α2 and α1 because both of
them have the same structure; we do not call them as α and β. The moment you bring in
a β, that means, you are talking of a different structure. So, this is how you see this
happening. Now, what should happen if I increase the temperature, what should happen?
If I increase the temperature, this ΔHmix, is it a function of temperature? Actually yes, and
no, according to this it is actually no. But, we if you carefully look at
ΔHmix= ΩXAXB, Ω is what interaction parameter; it tells you actually about how strong is
the A B bond; whether A and B attract each other or repel each other and these bond
energies are in principle function of temperature. So, in principle, there are models that
are available where Ω is consider as a temperature function and if you go to CALPHAD
journal’s
168
and CALPHAD, the book that I have told you Kauffmann and Bernstein book, you will
see where Ω is also taken as a function of temperature. But for most of the basic
thermodynamic understanding, people consider Ω to be constant. If ΔHmix is constant, the
only temperature function is TΔSmix and this being negative,
So, you would see that this keeps on going more and more negative as I increase the
temperature. As it becomes more and more negative, what you see, because
ΔHmix is fixed and TΔSmix is becoming more and more negative; you see you will reach a
stage where this free energy curve changes from this shape (Graph 1) to this shape
(Graph 2). You would actually see that this free energy curve (Graph 2) slowly starts
becoming like this (reduced convexity) and then it further becomes like this because, this
hump starts decreasing.
Because, the negative value of the entropy of mixing is becoming more and more, and
you would see at some stage, the free energy curve becomes like this (completely below
x axis). And beyond that temperature, the free energy curve becomes more and more
negative again because, TΔSmix is much more negative, you can see that just looking at
the TΔSmix term we can understand how this free energy curve, which has a concave and
convex mixed, is getting converted to one single type of free energy, which indicates
stability. We will also talk about how stability is related to this, when we come to the
point of spinodal. So, this is an indication of stability and this is an indication of
instability for free energy. We will see why it is so, when we come to
169
spinodal decomposition. At the moment, you just leave it. So, the moment you see that
this free energy curve (graph 2) that means, if you take any composition and
draw a tangent, at any given composition, free energy of all these composition is above
that of tangent. So, there is no chance for any other composition to precipitate out of α
of this composition. I can see it from all the composition, anywhere I choose a
composition and draw a tangent, you would see all the whole free energy curve is above
that of the tangent. And because it is above that of the tangent, there is no chance for
this. That is why this free energy curve indicates its single phase is stable and that is the
reason why, above this particular temperature (above the miscibility gap), you will see
the free energy curve becoming like this (graph 2). And that basically means that, this
miscibility gap has to merge.
If you see this (Graph 1) is, the convexity of free energy curves keep decreasing, and
draw again a common tangent, you would see these points (P’and Q’) would be inside
these two points (P and Q); that means the equilibrium
170
composition, the bimodal points are shifting towards each other as the temperature
increases, and to an extent that you would see that, slowly they start coming closer and
closer to an extent that at some stage, they both become merge into each other and
become one point.
And that is this point that you see here (point K, Graph 1), and that temperature in
principle. How do I calculate that temperature, can you tell me? Can I calculate that
temperature; it should be very easy, what is that temperature? Where this should happen?
Student: We can equate TΔSmix and ΔHmix
You simply equate ΩXAXB = RT(XAlnXA + XBlnXB)
you can find out what is the T and again that is not sufficient; something else you have to
do. Equate it and put one boundary conditions XA = XB because, both these functions
being symmetric, this actually happens at 50-50.
And once you take that, you would be able to find out this T, which is this. In fact, you
would see when you calculate this, later you would see this will become equivalent to
T* = Ω/2R, one can do this calculation and so, the higher the Ω value is, the higher the Ω
value is the higher will be the T*. That is what yesterday when we are talking about, as I
increase the Ω more and more or in other words, as I increase the size factor more and
more, you would see that this temperature increases (as reflected graph 2).
So, like this, we can in principle find out how this miscibility gap closes the temperature
at which the miscibility gap closes can be easily found out by equating this with this and
then finding it out In fact, it is not so simple also, we will see as we go along, when we
go to spinodal, there is also another way to find out by taking the second differential of
the free energy, how to do that we will see when we go to the spinodal at the moment
you take it this way.
Student: α2 is a B rich phase and it is surrounded by α1 , so according to the principles of
diffusion, it should have tendency to diffuse into α1.
171
Let us see, this is all α (Fig. 1) , starting with α we are talking of a situation where a free
energy curve is like this (Fig. 2). We are talking of a composition like this (C0), this is
the starting composition. Let it be 20% or 30% B that is α and in this α now, α 2 having
this particular composition (C2) has nucleated. This is α2 which is nucleated and this has
much more B content isn’t it. And how does this nucleation occur? That is where another
question is, you say sir, α is containing only 30% B and if it 30% B, then how can in a
30% B, which is uniformly 30% B, suddenly, you will have a nucleus of the α 2 which has
a 70% B coming out. This is where you need to understand that every liquid or solid,
you at any given temperature, there is always vibrations. There are always atoms moving
and there are always clusters of varying compositions though, over all the composition is
30%. You can have a small region which is a few nano meter region, or even a less than
nano meter region, where the atoms are more concentrated in terms of either A or B. But,
the question is, according to thermodynamics, at a given temperature, if this α is not
stable, then those clusters which have B content of this would get stabilize suddenly and
that particular clusters will start growing; that is what we call it as r*, this is how even
liquid to solid
172
transformation, you may say, sir liquid is completely random, how does suddenly solid
nucleate? Because, there is a short range order inside the liquid and inside the short
range order, you see some clusters having a particular crystal structure. They may be 100
atom clusters, 10 atom clusters, 1000 atom clusters, different -different sizes do exist,
and once a particular size of the cluster has the value of the r* for that particular
temperature, that particular cluster starts growing. So, let us assume that there are
clusters.
So, that B has nucleated. How it has nucleated, let us not bother about it. Now once it is
nucleated, this is α2 (Fig. 1), now this particular α2, can it be in equilibrium with this α
(C0) ? It is not possible, because I cannot draw a common tangent between this two,
impossible.
Student: So That free energy is responsible for diffusion in α2
Yes, so this.
Inside α2 is because this α2 is stable, it has a low free energy with respect to that α , with
respect to that α , because it has a low free energy. So, it tries to come out, but once it
comes out, because this cannot be the only way this α 2 can be stable is either to dissolve
back into the matrix or change the α around its composition such that it can exist with
that. It is like you know if you want to be friendly with somebody, you change your
attitude or change your friend. So that, you are in kind of harmony with the other friend,
exactly the same way here. So, if you want α 2 to be to be alive, it can be alive only if α
around it changes its composition such that you can have a common tangent between
those two, that α (C1) alone can be in equilibrium with α 2.
So, as a result, because, already from the α B atoms have come together to form α 2,
then obviously, this α has given out some B atoms. Because it has given out some B
atoms, its B atoms will decrease. So, in a small domain like this (Fig. 1), the B atoms
have decreased and when they have decreased, that means, the composition of α has
actually started moving in this direction, only at least a small region, it might not have
happen in the whole α , but at least just around α2 it would happen. And that particular
small region around this, I will call it as α1 and that composition will reach this particular
composition so that this α2 can be alive, otherwise, α2 cannot be alive. It has to dissolve
back this is the major problem, that is why we will see α1.
173
And now, what should happen in future? This α1 α2 is there, then α is also there. So,
finally, at this temperature in principle, what is stable is only α1 +
α2
174
and what is the driving force for that? The driving force is this (ΔG(C0) – ΔG(P)), that is
the ΔG for the α to give you α 1 + α2, that is the driving force. So, driving force is
available, the only thing that you need is, you need diffusion so the diffusion takes place
then what should what you will see is, this α slowly vanishes, the α having this
composition (C0) because, it is not stable. It slowly starts giving out the B atoms more
and more so that the α2 starts growing.
And as α gives out more and more B atoms, what happens? The α becomes poorer in B
and so that the whole of α , so after sometime you would see a situation like this that the
α2’s have grown (Fig. 3). These are all α 2s because, the B atoms are coming
continuously and flowing into the α2 from the α , and as they flow into this, you would
see α2 increases and the remaining α completely becomes α1. So, you will have a α1
matrix and α2 precipitate inside and this can be seen very easily if you look at there is the
gradient between this and this. Why does there is a gradient come, for example, you can
see the moment you see a situation, I see a case where everything is α (Fig. 4),
everything is α with a certain composition initially.
(Refer to Fig. 4 above for this discussion) Suddenly, you see α 2 comes out; if the α2
comes out, α2 has a much higher composition then the α isn’t. So, as result, this is, if I
call this as α2 which is coming out, the remaining is α , but this α 2 cannot be in
equilibrium with this α . So, what is that, it can be in equilibrium with? So, it is this
value. So, you will see that α2 will be in equilibrium with α having a composition which
is much lower than this α . So, you will see that, this will be actually, if I draw the
concentration profile it looks like This is the α , this is the α 1, this is the α2. So, you
would see a concentration profile something like this, just adjacent to α2, α will have a
lower composition.
But away from the α2, α will have the same composition with which you have started.
Now, what happen is, if you see, there is a gradient here, B atoms are more in this.
Everything I am talking in terms of percentage B as a function of distance. So, initially
everything is α and inside that α 2 has come out and α2 has a higher B content. And
because this α2 has to be alive; it can be in equilibrium with α having a lower B content
than what it is. That what you see this is the starting α , this is a α composition which is
lower than the thing. So, now, you see there is a gradient, because there is a gradient you
would starts seeing that the B atoms are flowing
175
from this α towards this. Then what should happen? Slowly, this will grow; this will
start growing more and more.
But will this α1 change, will this composition change? This composition is equilibrium
composition of α1, in equilibrium with the α2, so that cannot change. Then, as this
grows, something else should happen. What should happen? We are talking about in
terms of concentration anything should happen. Of course, amount of α will decrease
because, the α2 is increasing, α will decrease; composition of α will go down. So, as a
result, you see that this curves as this grows you would see a situation when, it becomes
this. This B content in the α has to come down, because it is giving away.
It is like a reservoir, giving the B atoms to the α 2. As it keeps on giving more and more
atoms to α2, the amount of B atoms in the α should come down. So, this particular value
comes down and as this comes down, you will see a situation where finally, the whole α
around this α2 will have the α1 composition. Then, we will call this whole thing as α1,
this as α2. So, now it is α2 and α1. Of course, kinetically if I look at it, because if this is
happening as the α composition is decreasing, the gradient is coming down. That means
dc/dx is coming down; you know that J = - D dc/dx, if the dc/dx is coming down, then
what should happen? So the reaction will slow down, so the growth of α 2 will slow
down with time; that is fine.
But thermodynamically, finally, it should come to this (α 1) and that is how you see that
remaining α will be called now, not as α , it will be called as α 2. So, that is how you see
one single α has become α2 + α1. And the same thing can happen in a different fashion
in a spinodal, we will know how it happens, when we come to spinodal. You can see
that, this is how a splitting takes place, and as you increase the temperature, this is how
you see this gets closed up. I think with this we will close for today because I wanted to
start another concept, which we will start in the next class.
Student: Mass balance will be maintained
Obviously, that is mass balance has to be maintained. So, because the mass balance has
to be maintained, the only thing is, we can write an expression now.
176
fα1Cα1 + fα2Cα2 = C0
with which I have started. So, the amounts of α 1 and α2 will be adjusted in such way
that, these compositions will give you a particular amount. So, we can, this is how
actually our lever rule comes. How to do get lever rule? It is only this.
177
Prof. Dr. B.S. Murty
Indian Institute of Technology Madras
Module #01
Lecture #09
T0 concept, Partition less solidification
In the last class, we have talked about how the miscibility gap closes and the
thermodynamics behind it. How does the entropy of mixing basically helps as you
increase the temperature that the miscibility gap which is created by an inflection in the
free energy curve closes, because at high temperature you have higher TΔS mix term, and
which dominates the ΔHmix term and as a result, you will see a free energy of mixing,
which is completely negative. And, whenever the free energy of mixing is negative, you
will get a curve which will lead you to a single phase state rather than a two phase
mixture. This is what we have seen yesterday.
Today, let us take up the free energy composition diagrams again and see an additional
feature that this free energy composition diagram gives us; and that gives us a lot of
understanding about phase transformations, particularly what we call it as polymorphous
transitions.
178
Let us look at that aspect. We are all now aware of this free energy composition
diagrams, that you can draw something like this for a particular phase, and if you have
another phase, let us call it as a β phase, you can call something like this (see the graph
above).
But one point that all of you should remember is, whenever I draw a free energy curve
like this, it basically represents one issue that, that this particular phase is existing over a
wide range of composition. Usually, we use this kind of free energy curves for solid
solutions or liquids, for that type of cases, where these phases exist over a range of
composition. Whenever, you see a case, such as an intermetallic compound, where it
exists over a very short range, particularly for example a line compound in principle, the
free energy of a line compound in a free energy curve. Let us say, I am talking of an iron
carbon phase diagram and I have a compound there. What is the compound that we have
in an iron carbon, prerna?
Student: Fe3C
Fe3C. Fe3Chas a fixed composition? What is the composition of carbon content in Fe3C?
Student: 6.67
6.67, that is the weight %, atomic %? How much?
Student: 25 %
25 %. So, 25 atomic % and 6 point 6 7 weight % it is a fixed composition. So, if I have
to represent that particular phase in a free energy composition diagram.
179
Let us say, I am talking about iron carbon diagram at room temperature and I want to
draw the free energy curves for the phases that are present at room temperature. At room
180
temperature, what are the phases that are present in an iron carbon diagram?
Student: Ferrite and cementite
Ferrite and cementite, that is it.
You have only two phases present. Whatever composition that you chose, you have only
these two phases and as a result, the ferrite phase we draw a free energy curve like this
(see graph above), whereas, cementite, what you do? Cementite is at if I am talking in
terms of, let us say atomic % here, when I say X of carbon mole fraction or atomic
fraction of carbon. So, in most of our free energy diagrams, we usually use either a mole
fraction or the atomic fraction, rather than the weight fraction.
You can always convert it into a weight fraction and show it. So, let us say if I am
talking of atomic fraction. At 25 % of atomic % of carbon, you should have F e 3 C and
in principle, the free energy of that F e 3 C if you represent on this G X diagram. G X
diagram is always drawn at a fixed temperature, am I right? So, this is at a fixed
temperature and of course, pressure is already fixed. So, at a fixed temperature and
pressure, we are drawing this.
And at a fixed pressure and temperature, the free energy of F e 3C is fixed. Because it is a
fixed, it is like a pure metal. Like you have 1 point here for a pure A, and another point
here for pure B. Similarly, F e 3C is a single; it is almost treated as a single component.
So, it has a fixed free energy at a particular temperature. So, we treat the whole F e 3C as
1 entity and calculate what is the free energy of this intermetallic compound as a
function of pressure and temperature. And at atmospheric pressure we can calculate as a
function of temperature and if you do that, it would turn out to be a nothing but a single
dot, 1 point on this free energy curve, on the free energy diagram.
So, that point can be above or below ferrite. But, because we know that, at in this
particular phase diagram, that both of them coexist; almost from what level, what
composition of carbon they start coexisting?
Student: 0.025
0.025. Wrong. I am talking of room temperature point 0.06 onwards. So, that means, I
should have, in fact, this is not a possibly the right way of drawing this.
181
If you want to really show that and then you should have at 0.25. If I call this as a 50- 50,
and this is somewhere 0.25, let us say, somewhere here you should have a dot (point P).
That is a free energy of cementite. It is 1 single point, it is like 1 single point for the pure
metal, pure iron and pure carbon on the other side, you will have 1 point and because
showing 1 point is difficult, what we actually do is, we draw a very steep curve like this,
just to indicate that at any composition on the left or the right of 25 % of carbon, the
Fe3C phase is highly unstable; it cannot exist as a single phase.
So, it is free energy increases drastically on both the sides. That gives you an idea that, it
is actually a line compound. In principle, this width is very -very small, just for you to
see it, I am showing it. Actually, it is a line compound. So, the width does not exist at all
and now, if I draw a common tangent between the 2, between this tip (point P) of that
curve and the α phase, I will be able to see that this is the composition (C α) at room
temperature. If I am doing it at room temperature, this should correspond to 0.006. Of
course, scale wise, it is not showing that way in this particular figure; but we can take it.
If you actually calculate it, you would see that this common tangent, the tangent of the
Fe3C curve will touch the α curve at room temperature only at 0.006.
Otherwise, thermodynamics is not satisfying the phase diagram. You will see that it will
touch at 0.006 and you can say that here is the α + Fe3C. Of course, we do not
182
know what is there on the other side because we actually do not talk about iron carbon
diagram beyond the Fe3C because of the instabilities that exists. And to the left of this
(Cα), I can say you have a single phase α. Obviously, at any composition, if you chose
less than 0.006, you have actually a single phase α.
In fact, we have a number of steels. I do not know how many of you are aware of a new
type of steels that are coming up, which have very small amount of carbon.
Student: HSLA
No, No HSLA is still has almost 0.2 % carbon. Have you heard of any other steel where
you can have almost 30 ppm of carbon? They are called IF steels, interstitial free steels. I
mean, in though the name is misnomer, it is not interstitial free, but extremely low
interstitials. The carbon content and nitrogen content are kept such a low value that they
are 30 to 40 ppm and in fact, people even represent them as the y axis steels.
Because, if you want to locate that particular steel composition on the phase diagram,
actually, you know if I draw an iron carbon phase diagram and I want to show 0.2 %
carbon, I can show it. I want to show 2 % carbon, I can show, draw a vertical line and
show it. How do I show a 30 ppm carbon content on the iron carbon diagram, because it
actually goes into the vertical line that you have on the left which is nothing, but the pure
iron line.
So, they are almost, call like that, and why do we make such steels? Can you tell me?
Why should I go to such a low carbon content and in fact, if you want to really go to
such a low carbon content, what should you do if you remember when, how do we get a
iron. All of you should know by now, how do you get iron, from where you get iron,
from?
Student: Reduction process
From reduction process, reduction of what to what?
Student: Iron oxide
Iron oxide you reduce it, in what?
Student: Blast furnace
In a blast furnace for example, and what you end up in?
Student: Liquid iron
Liquid iron, is it going to be liquid iron simply, of course, with a lot of carbon, with a lot
of carbon and what do you call that?
Student: Pig iron
183
We call it as pig iron.
And, that is not steel all of us know. And from there, you have to come to a steel level.
And what do we do here?
Student: Oxidized
You oxidize the carbon; that carbon which is there in that iron you start oxidize. So, you
have to first reduce it and then again oxidize it. That is what you have a LD converters,
Bessemer converter, all these things. So, you basically oxidize this and if you want to
really oxidize from a 3.6 % level to about 30 ppm level, you can imagine how much of
oxidization that you need to do, of this carbon.
184
And in a process, obviously, when you are trying to oxidize, even some iron also can get
oxidized. So, you get a lot of slag too. So, you need to really go to, there are what are
called special, what is called secondary steel making processes, where you can actually
reduce the carbon to a very small extent and people are using thus. Why do people go,
take all this trouble?
Student: To avoid strain
Not really that, partly yes, to have a high formability. This is you know, people want to
make automobiles for example, with 1 single sheet so that you can simply stretch it the
way you want.
So that the whole body instead of making it different -different components and welding
it. And another problem while welding steel, you all know that the higher the carbon
content, the more the problems is in welding. So, if you have a very low carbon content,
even welding of steel also is easier. But more importantly, they want to make in one
single go, as much as the body of the automobile. Particularly, these are used for
automobiles, Japan has come up now, even Tata steel is working hard towards that,
which are called IF steels, lot of work is going on.
But, the problem is, we know that when I reduce the carbon, I lose something. What is
that, I lose the strength that is the major sacrifice I am doing. So, how do I overcome that
problem? That is where people are coming up with what are called nano steels now. So,
they are adding some elements into the iron which precipitate inside the iron. For
example, copper. When you add to iron, copper does not dissolve in iron, iron copper
phase diagram if you see, copper does not dissolve into iron. In fact, though there is a
peritectic, there is actually a submerged miscibility gap in that particular phase diagram.
So, copper does not dissolve into iron so easily. So, copper actually precipitates as nano
particles in iron and that gives you a high strength. A lot of people are working on it. In
fact, a such a similar precipitation of a metal which is dissimilar with iron, which does
not dissolve with iron. People have been doing it very regularly earlier. Have you ever
heard of any such element which is added to iron specifically to precipitate out as a pure
metal and give you something, some good property? Have you ever read this? They add
an element which does not dissolve, they pretty well know that it does not dissolve into
iron.
185
But, they know that this precipitates out and gives you a good property. This is known
for 30-40 years.
Student: Niobium
No, that is all HSLA you are talking there HSLA does not come, niobium does not come
out as pure niobium, it forms carbide. We are talking of some element which comes out
as a pure element. It does not form any carbide’s. It is a lead. Have you heard of free
cutting steels? If you have heard of free cutting steels; the lead is added to basically
improve the machinability. It acts almost like a solid lubricant. It is like graphite in in
cast irons which acts as a kind of a lubricant. That is why, cast irons are with graphite,
are easily machinable.
So, similarly, here lead is basically added because you do not have graphite in steels. So,
what is an equivalent of it? So they thought let us take a low melting metal and
interestingly, iron and lead are immiscible. They do not mix with each other. So, this is
where you know the alloy design we talk about. The excitements in metallurgy come in
that kind of a way; that how to really play with the phase diagrams and get what we
want. So, similarly, people are using copper and we ourselves have done a lot of work
where we add small amounts of copper which precipitates as nano particles and gives
you extremely high strength; hardness levels of as high as 5 GPa we have seen in this
kind of materials, very high strength.
Anyway, so, that is what you do on a low carbon. It is an off shoot of what we have been
talking on the free energy, though it has nothing to do with the free energy. Now, let us
come back to this kind of a phase diagram and talk about instead of putting a β there, put
a liquid there.
186
So, I am talking of a liquid and solid and when you see a free energy composition
diagram like that (Graph 1), I can always talk about a common tangent, and talk about
these two compositions (Cαl and Clα). What are these two compositions?
Student: Binodal term
There is no, binodal term, you have to use only when there is a miscibility gap. Please
remember, there is no miscibility gap here, we are talking of a liquid and a solid and
remember that we are drawing two free energy curves. Because there are two different
phases, two different structures, whenever you see, there is 1 single structure.
Have you ever seen, did we talk about pure iron phase diagram? For example, I want you
to take this as an assignment. Please, I want you to draw the free energy curves as a
function of temperature for pure iron. What are the solid phases that you have? α, δ and γ
iron; 3 irons you have. So, I want you to draw the free energy curves to show the
transitions between α to γ, γ to δ, and δ to liquid.
So, that means, also draw the free energy curve for liquid and see what happens and that
is where something which I just now mentioned, that structure should play a role there,
and you will see that it becomes very important. Of course, we will also talk possibly in
the next class; just to check those who have not been able to do. We will try to correct
that, but I want you to do this weekend work on this.
187
So, draw the free energy verses temperature for α iron, for γ iron, for δ iron and for
liquid iron and then see where they are intersecting of course, schematic though, and
show these temperatures; where we see these phase stabilities changing from 1 phase to
the other; and see if there is any difficulty that you find there; and you would find some
difficulty, let us see how you will overcome that difficulty. Coming back to here, let us
say a simple case as a simple isomorphous. I should draw the other way because this
liquid is on this side.
So, let us say, we are talking at this temperature (T, Graph 2). At this temperature, if I
draw this free energy composition diagram, these two points (Cαl and Clα )are nothing but
the equilibrium composition of α and the equilibrium composition of liquid at that
temperature. This is how we write, composition of α in equilibrium with liquid, C αl and
this is composition of liquid in equilibrium with α, Clα.
These terminologies have to be clear, so that we understand them well. As we go to

eutectics and peritectics, you will see more interesting things coming up.
Now, I said there is also a composition here, where there is an intersection (point P,
Graph 1). What does this intersection basically mean and what kind of an information
that it gives you? We said that at this intersection, G α = Gl, am I right? And if that is the
case, what is it going to happen if I am on either left side or on right side of the phase
diagram? For example, at this temperature, on this phase diagram, this is a G-X diagram
and the phase diagram (Graph 2), if I want to locate this particular point, I think in 1 of
the classes, we have seen that it usually falls between these 2 compositions (C αl and
Clα ) , because, obviously, it has to fall between these compositions.
So, let us identify a particular composition, call it as that we call it as X 0 or C0, call it as
because, we are talking in terms of composition as C. See, composition can be talked
either as C or X; I mean I will keep on interchanging in this class may be
188
once in a while, but please forgive me, you can either use X which is a mole fraction or
C as a composition; but both are both are equally accepted in all the books. In different -
different books, people use different -different terminologies.
So, if I call this as C0 point, now what is it if I chose a composition on the left of this?
What I am saying is, I am taking a liquid heated it to a high temperature, where it is
completely liquid. Now, taken this liquid of that particular composition which I will call
it as a C1 composition (Graph 2), brought it to that particular temperature where we are
talking about it, this is the temperature we are talking now. I have brought it to that
temperature, what should happen to this liquid? At that temperature what does the phase
diagram tell you? The phase diagram tells you that that is the temperature which is a two
phase mixture.
It is a liquid plus α mixture at this particular temperature, this whole region is liquid plus
α. So, if I am taking a composition like this, this is the C 1 composition for me, which is
on the left side of the C0 (Graph 1) and I have heated to this temperature, brought it to
this temperature. At that temperature, in principle, liquid is not stable as a single phase
liquid and you can easily see it in the free energy composition diagram also. If I look at
this particular, what is the free energy of the liquid? Free energy of the liquid of that
composition is given by this point (Q), am I right? Because, this is the liquid free energy
curve.
So, at any composition, I can get the free energy of that particular liquid by choosing that
particular point. So, free energy of liquid of C1 composition is this (point Q). This is not
stable because the two phase mixture having a two phase free energy as that (represented
by point R), as that is more stable than the liquid. So, there is a driving force for the
liquid to become a two phase mixture. So, liquid would split as follows:
lC1  αCαl + lClα
and for that you have a driving force, which is given by this particular quantity. This is
what is the ΔG.
That is why, liquid does not remain as liquid. The moment I bring the liquid to that
temperature, it will split into α + L. But if I carefully observe, for this split to happen,
you need something to happen; why because here, this is the liquid of C 1 composition,
this is α of Cα composition, this is the liquid of Cl composition.
189
All the three are three different compositions; that means, from a liquid of C 1
composition, now you have a solid with some other composition and a liquid with some
other composition has to come out.
This means, you can see (Graph 1) from this alloy, this composition and this
composition (Cαl and Clα ) should come out. That means, you need a lot of diffusion to
happen, you need α to first nucleate out of it and α will have some other composition.
Again, how do I know α will nucleate? I simply draw a tangent. The moment I draw a
tangent (k), I can easily see that what is the α that can nucleate out of this liquid; I can
easily. Wherever the tangent intersects the α curve , it is intersecting somewhere here, let
us say, I mean very close to C0, let us not bother about it.
So, it is intersecting somewhere here. On the right side of this, the free energy of α is
higher than this tangent. So, no α of that composition can nucleate out of this liquid
because, the free energy is higher on the left side of this intersection
If I look at it, the whole α curve is below the tangent. So, as a result, all these
compositions in principle (hatched area), this whole domain can, any composition can
nucleate. But, the composition that nucleates will be the one which has the highest
driving force usually. So that, they it has a greater tendency to form and as a result, you
would see some composition will come out of it.
190
Let us not bother which composition comes out of it. Usually, whatever composition that
comes out of it, it shifts again to the equilibrium composition which is possible. For
example, that is also interesting, now that I am saying, let us say, a case that this particular
composition (C2) has the highest driving force. But interestingly, that α cannot be in
equilibrium with the liquid, with any liquid because this composition, if I take it, I cannot
draw a common tangent between this α composition (C 2), with any liquid. Because, if I draw
a tangent, that tangent will be, you know, below the liquid curve; you can see that (Tangent
m).
It is not possible for that, I can draw. So, what happens is, this α which is coming out of
the liquid, because it is not stable, because it has to remain in the liquid environment and
because it cannot remain stable in the liquid environment, it changes its composition
slightly such that, comes to a composition which can coexist with the liquid of some
other composition; whatever is possible under equilibrium conditions at that temperature.
That is why there will be a small change in the α composition that would occur.
So that now a new α composition will come out of the initial nucleus, initial nucleation
will be this (C2). But, that nucleation cannot be in equilibrium with the liquid. So, some
more rearrangement of atoms will occur so that the nucleus composition will change to
this value so that, that particular α can be in equilibrium with liquid (C lα). But now, with
what composition it can be in equilibrium? It cannot be in equilibrium with the C 1 with
which I have started. So, it will have a some other composition. So, this is what in the
last class also we have seen that when a α having a lower B content comes out of the
liquid, there we talked about α and α one, the same argument you can give here.
From a liquid, when an α of a lower B content comes out of it, the surrounding liquid
composition will change.
191
So, for example, if I take this kind of a liquid reservoir and a small α is coming out of it,
and because this α has lower B content, then the liquid which is surrounding that will
have, obviously, more B content. And that is why, you will see the equilibrium will be
with that liquid; not with the liquid which is far away from it. There is no equilibrium
between this α and the liquid which is far away; but with the this is what we call it as
local equilibrium.
So, whenever a nucleation occurs, there will has to be a local equilibrium to be

established. If there is no local equilibrium, they that particular phase, that is nucleating
out of any phase, any parent phase cannot be stable. For it to be stable, it has to establish
a local equilibrium with the parent phase and the way it can establish, is by changing the
compositions locally domain. It can be a even a monatomic layer; very thin 1 nano meter
or even a less than a 1 nano meter layer of liquid changes its. And in a 1 nano meter thin
layer of liquid, the composition change can be easily happen at that temperature.
Temperatures are high enough; we are not talking of absolute 0 or anything like that.
So, because of which, you will see that there will be an equilibrium.
Student: But then how this liquid can nucleate
Yeah the now, now the remaining liquid that is there, has a different composition and this is
this is happening at the beginning of the nucleation. And that liquid, because it cannot
remain, because liquid of C 1 composition cannot remain at that temperature, you see at this
192
temperature, if I bring this liquid as C1 composition to this temperature, liquid cannot remain
as single
193
phase liquid. So, what happens is, slowly now like what we have seen in the last class,
there will be a diffusion of B atoms from this liquid towards that and you would see that
because of the diffusion, slowly liquid composition starts changing and becomes that of
this composition(Clα).
And amount of α will slowly increase from the initial stage such that finally, you will
have α of this composition Clα of certain volume fraction, certain weight fraction.
Similarly, liquid of that composition (Clα), if you draw the same composition profile like
what we have done in the last class, I can draw that composition profile for you once
again if you want.
For example, I am starting a situation like this. Initially, everything is liquid. You have
let us say, a small α that is nucleated. I am talking of α, initially everything is liquid, that
is this is the composition C1.
Now, the α that is nucleated has what composition; has a much lower composition than
C1. You can see α has a lower composition than C1. So, as a result, the liquid in
equilibrium with the α will have a higher composition. So, that is why, at the interface,
this is the interface between α and liquid. At the interface, you will have two
compositions, one the composition of α (Cαl), another, the composition of liquid (Clα).
194
Now, you see that there is a gradient and in fact, this is what we call it as a diffusion
zone (Marked with circle). The chemical potential of α of this composition, chemical
potential of liquid of that composition, will be equal. That is what is the meaning of
common tangent. So, that is why they are in equilibrium, otherwise they will not be. So,
α of this composition cannot be in equilibrium with C 1 because, chemical potentials are
not the same; but α can be in equilibrium with liquid of that composition (C lα) because,
chemical potentials are equal.
So, now, they will be in equilibrium. Now, the only way it can happen is the C 1
composition has to slowly change. The only way the C 1 composition changes is, it keeps
on giving more and more atoms to the α. How does it give, this is in terms of %B?. If I
draw the same thing in terms of % A, you would see that it is a reverse kind of a trend.
So, A atoms in the in the C1 are higher than the A atoms here. So, I will actually see that
there is a gradient from the this one. Let me draw that also for you.
If I draw as a function of A, now you see again, I draw this. Now you see that this is the
situation, this is the C1 composition and you see this. % A in α is higher, Clα has a higher
A content than Clα.
And the C1 composition, you see C1 composition is here and Clα is this side. So, it has
very less amount of A, when compared to Cl. So, you
195
can see A atoms will start diffusing from here to here. So, more and more α, α needs A
atoms, because α is a A rich phase. For it to grow, you need A atoms. In fact, you will
see this becomes a problem, when you are talking of eutectic. We will talk about it; very
interesting problem. For example, in a eutectic, a L  α + β, liquid is in contact with α
and β.
α is a A rich phase, β is a B rich phase. How is that liquid is able to separately give A
atoms to α and B atoms to β so that both of them can grow together? That is what is a
lamellar growth. If both of them have to grow together, you need to see some mechanism
by which liquid is somehow, carefully giving A atoms to α and B atoms to β. It is again,
this can be understood from the thermodynamics. We will see it when we come to
eutectics. At the moment, we are only talking about isomorphous; let us stick to that. So,
you can see that A atoms, though B atoms there is a gradient in one direction, A atoms
there is a gradient in opposite direction.
So, A atoms gets supplied from C 1 and in the process, what happens, the the A atoms in
the liquid comes down. So, you will see at as this grows to next level like this, you
would see that composition shifts like this (see the graph above). As α grows much
more, then you would this composition remains the same. This is the C lα that is fixed at
that temperature. Here, we are talking only as a function of time now, temperature is
fixed, I have brought the liquid to that temperature, and now as a function of time, what
is happening? As a function of time, α will grow and in the process, the composition of
liquid which is far away from that of the α also starts coming down, because, it is
supplying more and more A atoms.
As it supplies A atoms, A atoms have to come down inside that. So, it is like a reservoir
supplying the A atoms. So, slowly the A content in the liquid comes down to an extent,
that at some stage, it will it will come to like that (step like curve in above graph). So,
you will have the composition profile exactly becoming like that. Then, you will see that
the remaining liquid, all the liquid will have that composition (Clα), and the remaining α
will have this composition (Clα).
When this whole process has been finished, we say now, the whole equilibrium has been
196
Established. At that condition, after whatever time T that it takes for that particular
process. After that time t, if I now look at what is the volume fraction of α and the
volume fraction of liquid or weight fraction of α and weight fraction of liquid.
And if I do this,
fαClα + flClα = C1
So, you will have some amount of α, some amount of liquid in such a way that this
multiplication will give me this. And if I change this C 1, what would happen? The
amounts of these two will change; but the equilibrium compositions will not change. If I
take instead of C1, if I take another alloy and bring it to the same temperature. These two
compositions are fixed for that temperature, what will happen is only the fractions will
change such that, this mass balance is maintained and this is how the whole solidification
takes place.
Now, we are actually not looking at this, this is something which is very easily known.
We are trying to see something which is not that easily known. That is what is a new
field that has emerged, that is what we are going to see as, If I now draw this once again,
these 2 equilirbrium compositions, now we are talking of this. Now again, I draw your
C1 composition for you. Look at this C1 composition once again (see the graph above).
What I said just now is that, this is the
197
free energy of the liquid of C1 composition (GC1l), and because it is higher than that of
the free energy of the two phase mixture, which is this (G C1l+α), that there is a driving
force for the liquid to become a two phase mixture.
But, at the same time if you carefully observe the free energy composition diagram, there
is something else there, that there is an α. If I look at what is the free energy of α of C 1
composition, where do I locate it? If I look at what is the free energy of α having C 1
composition where does that fall? Between those two which is this point (G C1l+α). That is
the free energy of α, having the same composition as C1.
So, you see three points there and what is the meaning of this? It means, the liquid has
two choices now, it can take one of the following two paths:
lC1  αC1
lC1  α(Cαl)+ l(Clα)
Why? Because, liquid of C1 composition has a higher free energy than both of them.
Now, which path the liquid will chose? Which path the liquid chooses depends on your
process. If I am cooling the liquid very fast, let us say a non equilibrium processing and
in that condition, the time available is very small let us say then, this process, for it to
take place, that is a second path for it to occur you need a lot of diffusion, you need a
long range diffusion. In fact, diffusion, there are two types of diffusion, some of you
might have heard of. For example, if a pure metal liquid, liquid aluminium changes to
solid aluminium or take an example of pure iron changing from an FCC structure to a
BCC structure.
When I am cooling from γ iron becoming a α iron, all these transformations need short
range diffusion. Why, it is only 1 structure changing to another structure. So, structural
changes from 1 structure to another structure, 1 unit cell changes to another unit cell,
FCC unit cell becomes a BCC unit cell. That means, the atoms which are sitting now, at
the moment, at the phase centered positions will simply move to body centered
198
positions. That means, what is the distance by which the atoms are moving? They are
moving within a unit cell distance.
The actual distance by which atoms move, do not actually cross more than 1 or 2 unit
cells. This is what we call it as short range diffusion. Wherever there is a compositional
change, for example precipitation, super saturated solid solution of aluminium copper
alloy, precipitating what is called a θ phase or a GP zones, where the composition of the
precipitate is entirely different from that of that. So, that needs a long range diffusion,
both the nucleation and growth of this. I am talking of the whole process, both of them
need, actually nucleation process does not need a long range diffusion.
Nucleation process occurs in a very small range though of course, they are also many a
times the nucleus size, is at least about hundred atoms size. There, you need diffusion
over the whole thing. So, this is what we call it as long range diffusion and when you
have long range diffusion, it needs more time, a short range diffusion does not. You also
have another type of process. For example, think of the case as austenite giving you
martensite. What is that type? We call it as diffusion less. There is no diffusion there,
neither short range nor long range and interestingly, there also what is happening is only
structural change. One structure is becoming another structure; whether it is BCT or
BCC, that is the another problem.
All martensites are not BCT, all steel martensite are not BCT for example. There can be
iron base alloys martensites which are not BCT, do you know this? For example, pure
iron also can become martensite and pure iron martensite is not BCT, it is BCC. When
we come to martensite; we will talk about it. If you take a iron nickel alloy, iron nickel
alloy also, you go to austenite quench it, it becomes martensite, is again a BCC. BCT
comes because of something else; because of the presence of carbon, presence of carbon.
We will see it later, when we talk about martensite any way. So, that is where you have
to mix up thermodynamics with something else when, I teach you all these things.
So, the interesting thing is, because this is difficult (first path) whenever I do a non
equilibrium processing, the liquid will tend to do this (second path). In this first process,
what is happening? Liquid is changing to α of the same composition. So, it does not need
any long range diffusion. Only a structure has to change; from the liquid structure to
whatever is the
199
α structure, it may be FCC, it may be BCC, whatever it is, and this is what is called
polymorphous transformation or partition less solidification.
We call it as partition less solidification. Solidification is occurring, but there is not

partitioning of atoms. Partitioning means, the atoms being distributed to 2 phases; that is
not happening. So, the compositions are remaining the same and the general term for this
is called polymorphous transformation, because, such a transformation can occur in even
solid state. Just now I told you, if it is happening in pure metals, we call it as allotropic
transition. For example, pure iron going from FCC to BCC, we do not call it as
polymorphous transition; but, if it is happening in alloys, for example, if you have heard
of β brass or there are you know α β brasses, where if you take a α β brass mixture and
cool it very rapidly.
This phase, which is a high temperature phase, is a β phase transforms to a low

temperature phase, which is α phase, without any change in the composition and at the
same time, not by a martensitic transformation. And that kind of transformations are
called massive transformations. This is we will also possibly come across that later, as
we talk about more and more phase transformations.
So, these are what are called polymorphous transformations. That means, only structure
is changing, composition is not changing. You can also have cases where the
composition and structure both change or only composition changes, structure does not
change. Can
200
you tell me that case where composition changes but structure does not change? Any
example, structure does not change, but composition change?. We have done this, to just
give you a clue, we have done it yesterday. A transformation that we talked yesterday is
that spinodal or even a miscibility gap, a simple binodal. One α having a some structure,
changes to α1 + α2, with two different compositions but the structure remaining the same.
So, this kind of bimodal transition, miscibility gap, whatever decomposition, we call that
is why we call such things as decomposition. It is a simple; 1 thing decomposing into 2,
with two different compositions. Anyway, come back. So, this is what is called
polymorphous. Why this becomes important is, because, on this side of the phase
diagram (right side of intersection of liquid and α free energy curves) , what I have said
just now cannot happen. Why? Because, now on this side of the free energy composition
diagram on the right side of the C0 point (intersection point), if I look at it, liquid has the
free energy GlC2 and α has the free energy GαC2.
Clearly, GαC2 > GlC2. There is no chance for the liquid to transform to α of the same
composition because, there is a increase in the free energy there whereas, it can split into
2 phase structure, which is Gl+αC2. The two phase structure still has a lower free energy.
So, that means, on the right side of this thing, polymorphous transformation is not
possible, or partition less solidification is not possible. That means, if I chose any alloy
of right side of this C0 composition, I cannot have a partition less solidification. You
may say, Sir, what is the advantage of a partition less solidification? You just hold on for
a 1 or 2 days, when we are going to talk about eutectics, I will show you what is the
advantage of it.
One thing is, you can produce alloys with compositions which are much more solute
content than what is by equilibrium allowed. For example, under the normal conditions,
if I take a liquid, and solidifiy under slow cooling conditions, what you see is that liquid
will split into two phases, and finally, you would see the α and liquid will have certain
composition and if I cool it very rapidly, I can retain whatever is solute content that is
there inside the liquid.
201
Of course, in isomorphous, that is not important because finally, when the solidification
is taking place, the α will have the same composition as that of the liquid. Of course, it
is, you will have what is called a non equilibrium cooling coring; all these things are
going to be there. So, considering the coring is going to be there, you have actually
compositional differences, in any dendrite you have that coring. But, here, we are talking
of a solidification where, absolutely there is no coring and you have a composition of the
same as that of the liquid coming out, and particularly if I look at compositions like this,
in a eutectic kind of cases.
Let us take a case of like this. If I take this alloy (C0), in a normal condition this alloy
when it is solidified and brought to room temperature, you will have the α of this
composition (P); β has this composition (Q) isn’t it? And similarly, if I just come out of
the eutectic, at the end of eutectic, α will have this composition (R), β will have this
composition (S). Now, if I do partition less solidification, what I have is the α that comes
out of the liquid, will have the same composition as C0.
So that, now at room temperature, I have α having this much of B content (M) than what
I can have under normal equilibrium conditions. That can give you an advantage that you
can do precipitation later things like that, we will talk about it in the next class.
202
Lecture No. #10

To concept, partition less solidification
Last class we were talking about, what is called a concept, which we usually called it as T 0
concept, we did not use this term in the last class. We will use it today, what we were talking
about is taking a free energy composition diagram, for two phases call it as, let us say a α
and liquid, two phases.
203
In a phase diagram which is something like this, this side is liquid, so this is wrong (Graph
1). So, the phase diagram should be like this (Graph 2), can you tell me why did I say this is
wrong, the previous one (Graph 1).
Student: Free energy graph shows liquid stable on right side
So, you have to have liquid stable on pure B side liquid should be more stable at this
temperature at which we are drawing the. And our temperature is something like this (T,
Graph 2)) let us say, so at that temperature, if I extend this temperature line to both the sides,
I should see that on the pure B side, liquid should be stable at that temperature, that is the
meaning
204
of the common tangent. If I draw a common tangent (Refer Slide Time: 00:16), the two
points that you are going to get here, within these two compositions, which are called the
Cαl, Clα. Of course, some people even called it as equilibrium, Cαeq, Cleq; obviously here,
because when I say Cα, it is obviously in equilibrium with liquid only, there is nothing
else, but when there are three phases, then you have to specifically say it is in
equilibrium with what?
For example, we are going to come to eutectics in a few minutes and once we come to
eutectics, there you have to clearly say, α composition that you are talking is it in
equilibrium with liquid or is it in equilibrium with β. In this particular phase diagram,
because there is only liquid that is left out, other than α. If you simply say C αeq, it means
it is the equilibrium composition of α at that temperature and whenever we say
equilibrium, it is obviously with respect to some other phase, what is the second phase?
It is the liquid. So, when I say Cαeq or Cαl basically I am talking of the same, in this
particular context.
In this context of where there are only two phases, so basically they are kind of
synonymous, so we can call either Cαeq or Cαl. So now, if you take these two points
(Refer Slide Time: 00:16) between these two (C αl and Clα), it is α + L that is in
equilibrium and to the left of it, we had already seen, that it is α in equilibrium and to the
right of it, it is the liquid in equilibrium; that means at that temperature, if I extrapolate
that tie line, towards the pure A and pure B.
205
You should see that to the left of this (Cαl), there should be α in equilibrium and to the
right of this (Clα), there should be liquid in equilibrium. And between these two, it should
be α + L, that is how, there is a one to one correspondence between the free energy
composition diagram and the phase diagram. Exactly, you should see these two
compositions (on free energy curve) should match with these two compositions (on
phase diagram) and on the left side, it should be α and on the right it should be the liquid.
And that is where, you will see, if I draw the phase diagram like this (Graph 1, (Refer
Slide Time: 04:24) this would not correspond to with free energy curve, otherwise you
can always do, because this is anywhere schematic, I will simply replace l with α, α with
l, change the free energy curves then also it would do good. So, once you draw this free
energy composition diagram first (like in Refer slide time. 00:16) then it should be this
phase diagram (Graph 2, Refer slide time. 04:24) rather than that phase diagram (Graph
1, Refer slide time. 04:24), so this has to be remembered.
Now, we were talking about this particular composition, which we will called it as C 0
composition and we saw what would happen, if I choose any alloy and that C 0 composition, I
can put it on this phase diagram somewhere here, need not be exactly at the centre, depends
on how the free energy curves are, but it is definitely between Cαeq and Cleq composition,
(see the figure above). And if I choose any alloy to the left of it and bring it to that
temperature, from it is high temperature from the single phase liquid state, I melt this liquid,
bring this liquid to that temperature and hold it there.
206
And think what should happen, that is where we have said, that particular liquid has two
possibilities. One this liquid of that C1 composition, can become α of C1 composition
because, if I choose a composition like this, C1 I call it (Refer Slide Time: 00:16)
, that C1 composition will have a free energy of this (point P1), this is what we have seen
in the last class. And that can come to either α having this composition (Q1) or α + L
having this composition (R1).
And then we said, if you are cooling the alloy under equilibrium conditions; that means,
the time available is very large then the second possibility will hold good. The liquid
having that composition will split the composition, will split into these two; that means,
we call it as partitioning taking place; that means the composition B atoms, which are
there inside the liquid get partition between α and liquid. And that is how, you get α
having a different composition as that of the starting composition, liquid having a
different composition as the starting composition. This would come out provided you are
close to equilibrium. If you are far from equilibrium; that means, the time available is
very short then in such a case, you would see that, you would get the liquid having C 1
composition, would try to become α of C1 composition. This is what we called it as,
partition less solidification, there is no partitioning taking place during this, because the
composition of the liquid, composition of the α is exactly the same excepting that there is
only a structural change that is taking place and such transformations are usually called
as polymorphous transformations, it is a general term, whereas as far as solidification is
concerned, we call it as partition less solidification.
Now, we also saw, that if I take an alloy on the right side, what would happen; that means,
look at again, this free energy composition diagram (like in Refer slide time. 00:16). And
look at the composition to the right of the C 0; that means, let us call it as C 2 composition, I
am just trying to recapitulate before we go ahead because it is a slightly, you know more
difficult concept for you to understand. So, let us try to see if we take up a C 2 composition
and a C2 composition, if I take I can clearly see the liquid of C 2 composition has this free
energy (P2), this I would call it as G l of C2 composition. And this, if it has to transformed to
α of the same composition, it is not possible because G α of
207
C2 composition is higher than that of the liquid (Point Q2). So, there is no chance for the
liquid to spontaneously, spontaneously, you may always say transformations are possible
from a lower free energy state to a higher free energy state always, when do they happen,
when you somehow energize the low energy system.
For example, you can one best example is how are X rays generated. X rays are
generated, when a high energy beam or a particle hits an atom and knocks off an electron
from let us say a k shell, which is close to the nucleus and goes out. And this electron
from a k shell going out is going from a low energy to a high energy because it does not
want to go out, unless you knock out. So, it is similarly, all excited states that we talk
about, are all created from a low energy state to a high energy state by by putting in
energy into the system. For example a cold worked state, is a high energy state and you
have putting cold worked into the material so, from a low energy state, it has gone to a
high energy state. So, everything, so that kind of transformations are not spontaneous
transformations, remember, so you have to put in a lot of energy.
Of course, even from a meta stable state to come to a stable state, you need to put in some
energy, which is what we call it as an activation energy, but though, you have to remember,
there is a difference between transformation in this direction (KL) and transformation in this
direction (MK). If you see that this is a meta stable state (K) going to a stable state (M)
because this has a higher free energy going to a lower free energy. So, this there is a driving
force for it, though there is a driving force, we say there is an activation barrier for it, this is
what in the first class itself we have seen. So, but if you want to go from this state to this
state (M to K), which is a low energy state to a high energy state, the actual energy that it has
to scale up is not only this activation energy, but also this free energy difference. So, it will
have a much bigger hill that it has to cross, for it to go from here (M), so it has to climb this
and then fall here (K).
So, if you really want ground state material, a phase which is which is in a low energy
state to take it to a high energy state, you need to provide much more energy. It is not
straight away, just giving this much energy is not sufficient, it will not straight away go
like that. So, you will see, that is why, it is more difficult for systems, which are in a
ground state, which are at a low energy state to be taken actually to a high energy state.
Otherwise, you will see that this transformation can regularly occur, that is what we are
talking about, whenever we are talking of liquid transforming to α, we are talking of
208
this going to this (K to M) because at that temperature, we are talking liquid is not stable.
So, it is in a meta stable state, but for that also, you need an activation barrier.
(See the activation hills drawn in image above for the following discussion) And we are
saying that, in this case (liquid splitting into α+L), the activation barrier is, if I now
compare something like this, this is lC1  αC1, here it is lC1  αCαeq + lCleq, it is
something like this, here the both of them, if you look at it. This (N) is at a high energy
state than this (M), can you see that here, α of C 1 composition has a higher energy than
the two phase mixture (see in the free energy curve). So, you would but it is lower than
the liquid of C1 composition, so from here, it is going to this, why should it go to this
rather than going to this because this activation barrier is smaller, because that is smaller,
why it is smaller because it is a short range diffusion, clear.
So, you can see that though the driving force is larger and this is a smaller driving force,
I call it as ΔG1, I call this as ΔG2, ΔG2 > ΔG1. For the path one (lC1  αC1) , the driving
force is smaller; for the path two (l C1  αCαeq + lCleq) , driving force is larger, but the
activation barrier for the path one is smaller, this is what you can call it as ΔG 1*, you
call this as ΔG2*. So, the activation energy is larger there because of the large diffusion
that is necessary. So, as a result, you would see this is happening in this particular case,
where compositions are to the left of the C0 composition, if you are on the right side of
the C0 composition then the first path is ruled out only the choice that is available with
the liquid is only second path and that is what happens and what we do is, if we choose
compositions at different -different temperatures if you draw this free energy
composition diagrams, and get these compositions which are called the C0 compositions.
209
For example, if I draw free energy composition diagram at some other temperature, I can
get another composition, which will be again between these two points. If I draw at some
other temperature, I will again get one more composition, which is between those two
points like that, if I get and join all of them, I will get a line like this. And incidentally,
this line would meet with pure metal melting point here and again meet with pure metal
melting point on this side, why, because those two points are where, the definition of a
pure metal melting point is what? What is the definition of a melting point? Free energy
of both the phases the liquid and the solid is equal. And we are talking of C 0 is nothing
but the where the free energy is equal, free energy of liquid and solid is equal. So;
obviously, this particular composition or this particular temperature, we now call it as a
T0, this curve is called T0 curve.
210
Though in isomorphous this is not very useful concept, you will see as you go to eutectic
this becomes very useful. For example, if I now take an alloy composition called C1,
what is this line the top line?
Student: Liquidus
Liquidus very good, So this is called Tl, Tl the bottom is solidus Ts and this is called T0, .
So, why we say 0 because there the ΔG is zero, Δ G for the transformation is 0, that is
why, we call it as a T0 curve and this T0 curve, what is the significance of it now. The
significance is this, if I take any alloy and I under cool this liquid make it make it
completely liquid and then I am cooling this liquid, but if I under cool the liquid to a
temperature, below the T0. Let us say, I bring it to this temperature (T), there is a
difference between bringing a liquid to that temperature and under cooling the liquid to
that temperature. I repeat this, there is a difference between bringing the liquid of C 1 to
this temperature you call it as, let say T or under cooling the same liquid to temperature
T, what is the difference between the two. It is not the cooling rate.
211
I am saying that, I have under cooled C1 to temperature T, I am also saying I am cooled
liquid C1 to temperature T, what is the difference between cooling the liquid to C1
temperature, what is the temperature between…
Student: Liquid remains stable
Yes, when I say under cool the liquid to T temperature; that means, until you reached T,
the liquid remained as liquid, it did not transform. Otherwise in principle, when you are
cooling this alloy, the solid should start forming below liquidus, under equilibrium
conditions. If it is normally cooled, by the time you reach this temperature T already
some solid should have form and some liquid is left out. So, if you simply are saying
that, I am cooling the liquid to T, temperature T it means solidification has started, but
the moment you use this term under cooled. It has a special significance; that means,
until you reached that temperature, nothing has happened to the liquid, just only the
temperature has come down, liquid remained as liquid, when does this happen? It
happens only when there is a non equilibrium solidification. Under equilibrium this will
not happen, only under non equilibrium condition; that means, I have cooled it rapidly or
there are also other conditions. Do you know, what is the other condition by which you
can under cool a liquid. In fact in 1950’s, there is a great man by the name Turnbull, who
did special experiments to see, how he can under cool the liquid, anybody has heard of it.
Yes, anybody. No, they are called droplet emulsion technique. Have you heard of that
word, droplet emulsion technique anytime? What he said is most of the time, the
nucleation occurs by, what is called heterogeneous nucleation isn’t it. All of us know,
most of the time, in fact, if you want homogenous nucleation is very difficult, that is why
people are doing, what is called you know drop tube experiments. Peoples are also
doing, what is called levitation experiments, what is called container less solidification.
Have you heard of container less solidification? If you want to see containers less
solidification, go to Professor Phani Kumar’s lab, you can actual see it happening. So,
how to do container less solidification, levitate a liquid droplet and allow it to solidify.
So, that it floats in the
212
air and then solidifies, so that you can see possibly, there is no mould effect. So,
solidification has to start may be at the surface of the liquid or within the liquid.
So, what Turn bull has thought is, most of the time liquids always have some inclusions
inside them and these inclusion can act as a nucleating sites, these are what we call
foreign bodies and things like that. So, he says if I can somehow reduce these foreign
particles by making the liquid droplets smaller and smaller. The probability of finding a
foreign particle reduces, if you make liquid into smaller particles. So, he did some kind
of an atomization of the liquid droplets, so when the liquid comes out of an orifice try to
basically put a gas, so that it forms the whole liquid into the form of a jet stream.
And then different sizes of particles are formed and he very painstakingly took each
particle of the solid that has formed. And then polished each particle may be embedded
each particle into something and then polished it, to look at the micro structure of this
particle in the cross section. And when he looked at different sizes of particles, he saw
that the smaller the size of the particle, the finer the micro structure, whether it is a
eutectic lamellar spacing that he is talking about or the grain size that he is talking about,
he saw that the finer the smaller the particle is finer, that basically means the nucleation
rate of that is higher. And that he calculated back and said this is happening because you
have a larger driving force, where from the larger driving force has come, because the
liquid got cooled to a lower temperature; that means, liquid got under cooled to a higher
and higher under cooling.
So and people have later with various techniques confirmed that, yes this is what has
happened and so now, people actually do this droplet emulsion very regularly whenever
you want to study this, what is called the nucleation experiments, under cooling
experiments. One of the most popular techniques of under cooling experiments is this.
So, one way is to cool fast, here there is no fast cooling in the in the droplet emulsion the
cooling rate is normal, but you have deliberately made the particle smaller. So that, the
foreign particles reduce and if there is no heterogeneous nucleation, the liquid has to
keep on under cooling until a particular temperature called homogenous nucleation
temperature. You all must have gone through nucleation kinetic somewhere. So, in the
nucleation kinetics we defined something as TN, nucleation temperature homogenous
nucleation temperature, how do you get that homogenous nucleation temperature,
anybody knows?
213
How do you calculate that, all of you in second year B tech should have heard this, if
not go back to Reedhill today and read Reedhill, nucleation chapter in Reedhill. I will
not go into the details because, we are talking of thermodynamics, I do not want to go
deeper into the kinetics, I can take one extra class on that, right now, but I want to avoid.
So, that you read a little. So go back and read there is a way to calculate, what is called
homogenous nucleation temperature, which people called it as ΔT N, under cooling
required for the homogenous nucleation. So, if heterogeneous…
Student: From Turnbull’s drop emulsion
Yeah, basically from Turnbull, you can come, of course everything has come from
Turnbull, Turnbull is the father of under cooling. So, anything that you talk of under
cooling of liquid, you have to start with Turnbull it was Turnbull who started. So,
basically try to see, what is the barrier for nucleation and how to overcome that barrier,
what is the barrier for nucleation?
Student: Surface energy
Surface energy. So, you need to provide sufficient driving force to overcome the surface
energy and from that driving force, you calculate the ΔT. We will come to it possibly
after this; I will tell you how to calculate that ΔT. So, we will see that so basically you
are to get that ΔG, once you calculate that ΔG, from the ΔG will get ΔT, so coming back
here. So, if you are at this temperature (T) because you are below the T0, what does it
mean by below T0, if I draw a tie line at that temperature, if I draw a tie line at that
temperature, my composition that I am choosing is to the left of T 0 curve am I right; that
means, partition less solidification is possible. If I take the same alloy under cool this
alloy to this temperature, I called it as T1, let us call this as T2. At that temperature; that
means, this the under cooling of this alloy is smaller now, this is I can call this as ΔT 1,
this is ΔT1 and this is ΔT2.
So, I have given a smaller under cooling and at that temperature now again, if I draw a
tie line, I see that my alloy composition now, is to the right of the T 0 corresponding to
that, at that temperature my C0 composition is to the left, of my alloy composition, if I
am on the right side of the C 0 composition. I have already seen here that there is no
driving force for partition less solidification, only if I am on the left side of the C 0
composition, at any given temperature, there is a driving
214
force for the partition less solidification. So that means what, effectively saying that, if I
can bring an alloy below the T 0 curve, for any alloy composition there is a T 0 point. If I
take this alloy, for this alloy this is the T 0, if I can come below the T0, if I can under cool
my liquid below the T0 then there is a possibility for partition less solidification, if I
cannot under cool my liquid below T0, for whatever may be the reason, all those droplet
emulsions, wherever the droplet has under cooled below the T0.
You would see a spontaneous nucleation everywhere, otherwise it has to wait for the
surface to nucleate and the nucleation will go inwards. In fact the way, the nucleation has
happened also you will know, whether it is a heterogeneous nucleation or nucleated on a
particular foreign particle somewhere inside or it has nucleated everywhere and then
started growing and that is what, you would see, if you are below the T 0. So that is the
significance of being below T0 or being above T0. So, you can say that, if I can bring the
alloy below the T0 then I can have partition less solidification, if I am above the T0 then
there is no partition less solidification.
And this becomes more important, when we come to eutectics; let us just look at a
eutectic phase diagram. Let us look at a eutectic phase diagram and now, if it is a
eutectic phase diagram, I will have not simple one T 0 curve, I will have two T0 curves,
why, because I
215
will have one T0 curve between liquid and α, I will have another between liquid and β.
In fact, I can even talk about another between α and β, we will not talk right now,
because these of not that much significance, because we are talking of solidifications.
Let us look at the two curves liquidus (T l) and solidus (Ts) and if I draw a horizontal line
at some temperature there and then extrapolate that across the phase diagram, if I want to
draw a free energy composition diagram at that temperature, how will it look like?
It will be similar to what I have drawn before, but with some additions, what is the
addition here, we will have one more additional curve, instead of two curves, I have to
now draw three curves. Before we even draw that, let us identify the phase fields here. If
I draw that tie line there (at temperature T) and this tie line tells me that, wherever it is
intersecting the phase boundaries, it defines certain phase boundaries, certain domains
for phases. For example, this one (point P), this tells me that to the left of that, I have
single phase α . Between these two (point P and Q), I have α + L; between these two (Q
and R), I have a liquid.
Student: L+ β
L + β (between R and S). This is known to all of you, fine. So, if that is the case then
when I am drawing the free energy composition diagram a schematic one then liquid
should come somewhere in the middle am I right, α should be on one side; β should be
on the other side.
216
Now, let us draw that, I have a liquid curve like this, I have an α curve; I have a β curve.
I have three free energy curves. Now, you will have not just one single common tangent,
you will have two and in fact one can also talk about the third one, let us not talk about it
right now. This is one common tangent (KL), the other common tangent is this (MN), so
you can talk about two compositions here and two compositions here, what is this
composition (point K)?
Student:Cαl
Yes this is Cαl perfect, what is this composition (point L)?

Student: Clα
Clα composition of liquid in equilibrium with α, what is this (point M)?
Student: Clβ
Cl β
and this (point N) is?
217
Student: Cβl
Cβl, perfect you are all getting it right. So now, once I know this, again I have two T 0,
two C0s here, C0αl and C0βl am I right. I will have one for the α and liquid, I will have one
for the β and the liquid. So; that means, on this temperature line (T, Refer Slide Time:
28:06), I can put two points here, one point between the α and liquid; one point between
the β and the liquid. And this liquidus mind you, there are two liquidus here, α liquidus
and β liquidus. Similarly, this is α solidus; this is β solidus, just to differentiate because
between that, on this side it is liquid in equilibrium with α, on this side it is liquid in
equilibrium with β, just the terminologies, I want you to know.
Now, if I continue this at various temperatures, I can keep on drawing similar curves, at
different -different temperatures and at each temperature, I can get one C 0αl; one C0βl and
if I join all of them, I get two lines, one I call it as T 0αl, another is called T0βl. These two
are very important concepts, when you are talking about non equilibrium solidification.
The whole micro structure development in non equilibrium solidification just depends on
this concept, so please try to understand this. So now, if I take this, if this is the T 0, this is
a calculated value based on the free energy concept, whereas this whole phase diagram
other than these two T0 curves, you do not need free energy calculations to get that, how
can you get the whole phase diagram without free energy? By thermal analysis, by doing
thermal analysis one can get the phase diagram, but by doing thermal analysis, you will
not get the T0 curves. T0 curves alone are fully a free energy concept, so T 0 curves can be
obtained only if you have a free energy equation.
So, you need to have a Gα equation, you need to have a Gl equation and Gβ equation and
from that, find out at various temperatures, where are the free energies equal, finding out
the intersections. So, basically equate Gl = Gα, Gl = Gβ at a particular temperature and
then find out where these two are equal, which composition of liquid and solid, where
you have this equal and that if you can calculate you will get this. So, these are actually
calculated, you may also calculate the phase diagram, I don’t say no, the whole phase
diagram can be calculated from free
218
energy, but the whole phase diagram in a large number of cases can also be obtained
experimentally, by doing thermal analysis, but by doing thermal analysis or any other
experiment you will not be able get T0, this has to be calculated from free energies.
Of course, you possibly need some experimental data to get the free energy, what is the
experimental data that you need? Cp and one more, for a solid solutions you need one
more data.
Student: Ω.
Ω is the ΔHmix, so ΔHmix also comes from experiments; that means, when I take A and B,
whether this A and B they have, they have an attractive interaction or repulsive
interaction, how much is that the interaction, again can be obtain from the calorimetric,
that is calorimetric. So, for example, for this most of the time, you do not use a DSC
even a DTA would be sufficient for, if you want to get a liquidus and solidus. Liquidus
and solidus even a DTA is sufficient or in a simple thermal analysis can be sufficient
depends on of course, how shallow or how deep it is, So but here, any free energy
calculations, what you need is Cp data and ΔH data, these two data are required. If you
want really calculate the free energy for anything, if you have these two data available,
you can always calculate.
And now, Let us use an alloy of C 1 composition (Refer Slide Time: 28:06). This alloy
the moment I under cool it below this temperature (T 0αl) to some temperature, then at
that temperature if I look at it, this is to the left of the T 0 or the C0 corresponding to that
temperature, you have to be very crisp, when you start defining technically. So, this
composition is to the left of the C0 composition for that particular temperature. For that
temperature this is the C0, this whole curve is called T0 because why, because this is
giving the locus of all temperatures or locus of all compositions at various temperatures,
where the free energies are equal, locus of all composition at various temperatures,
where the free energies are equal that is what is the meaning of T0 curve.
So, now because we are on the left side then I can see partition less solidification can
occur, what does it mean; that means, liquid having this composition C1
would simply become α of C1 whereas, this particular alloy under normal condition,
what it is suppose to happen, when I am starting cooling this alloy, the moment you
reach this temperature, it should start forming α having this composition, you all
remember. So, the moment I reach here (liquidus) you have to draw a tie line there
219
where it hits, that will tell you the α composition that is going to come out of the liquid.
And then as you slowly come to lower and lower temperature, by the time you reach the
eutectic, you will have more and more α and at the eutectic whatever is the liquid that is
left out will undergo a eutectic reaction.
And in this alloy, once I reach the room temperature, you will have primary α + a
eutectic mixture. This is the typical micro structure for that particular alloy, which we
usually call it as a hypoeutectic alloy, whereas now, this same alloy when I cool it
rapidly, to an under cooling below T0αl , I do not have any eutectic, nothing I will have
singly simply as single phase α at room temperature, with the same composition as that
of the liquid, what does that mean; that means, I have the moment I reach this
temperature, at this temperature I have instead of a two phase mixture α + β, I have a
single phase alloy of simple α and what does that mean; that means, that I have now got
a super saturated solid solution.
Actually in principle, at room temperature α can contain only this much of B (point O)
this is what defines the solubility limit, this is the B amount that can be there in α, but
now I have an α, which has this much of B (equivalent to C 1 composition), so single
phase α, that is why, you know T0 concept becomes very important, that you can
completely change the way solidification occurs. If you can really under cool an alloy
below the T0, you can have solid the solidification taking place in such a way that instead
of a normal regular eutectic solidification, primary phase and eutectic mixture, you
would simply have, what is called a single phase partition less solidified α.
And on this side (pure B side), if you have, for example, if I take this alloy (C2
composition, Refer slide time: 28:06), now under cool it a temperature below T 0βl. Now,
I have to consider the other way, here I am on the right side of the T 0, C0, so; that means,
where I am on this side, I am on the right side, what is the meaning of right side. Now
you see, the liquid has a higher free energy than the β (Refer slide time: 30:37), when I
am talking of α and β, the right and left get inversed. So, on this side, you would see
liquid has a higher free energy than β. So, partition less
220
solidification is possible, if you are on the right side of the C 0, if I am on the left side of
the C0 β has a higher free energy than liquid. So, partition less solidification is not
possible. So; that means, again whether you are right or left, one thing is true, if I under
cool my liquid below the T0, I can get partitionless solidification.
So, what is the under cooling that I need to give? I should give a minimum under cooling
of ΔT0 = Tl - T0, this is the under cooling that I need to provide. For a given alloy
composition, If I can calculate what is the T0 and if I know what is the T l for that
particular alloy, now the difference between this; that means, T l - T0, will give me what
is the minimum under cooling that I should provide, if you give only exactly that, it is
not sufficient because you need some driving force because, if you are exactly there at
the T0, free energies are equal. Free energies are equal does not mean that reaction will
happen. You have to be slightly below the T0 so that, the liquid has a higher free energy
than the α, So, this is the minimum ΔT 0 that you should give, that is where we see that, if
you go deeper this is also not sufficient, what is the problem? So this is what is the whole
T0
221
concept and this T0 concept can give you, how to get extended solid solutions, these are
what are called extended solid solutions.
One of the basic advantages of rapid solidification processing, which Pol Duwez, I mention
to you in 1959 has discovered is this. For example, what he has done, he has simply taken a
copper silver phase diagram, copper silver and he took a eutectic alloy and cooled it rapidly,
when he cooled it rapidly, he found that this eutectic alloy will give only a single phase
structure rather than a two phase structure. It is supposed to give a two phase α + β whereas,
he got a single phase structure, that is because, he has under cool to an extent, that if I draw
this, whatever it is, if you have under cool there are other further questions that may come in
to your mind, we will address those things also sir, what would happen if I under cool below
this or below this. So those questions also we will try to answer as we will go along.
So, if I under cool below this (T0αl) or below this (T0βl), I would get either single phase α
or single phase β. So, I can in principle get a single phase structure, just by under cooling
it, what I should do is, what I need to provide is this much of under cooling (Tl – T0); that
means, my cooling rate should be such that, it gives that much under cooling, if I get that
much under cooling then in principle, I can get a partition less solidification, a eutectic
alloy gets converted to almost like a isomorphous alloy.
222
So this is what we call it as meta stable phase diagrams. So, eutectic vanishes
completely, when you do a rapid solidification, but there is a small other additional
problem. What is the additional problem? The additional problem is this. The moment
you take a liquid, in which a partition less solidified α has nucleated (see Fig. 2 in above
image), what happens then? The moment α gets nucleated, there is a heat release am I
right, what do we call this?
Student: Latent heat release
Latent heat release, we called this with a particular term, what is this called, this latent
heat release is called something.
Student: Heat of fusion
Heat of fusion its the same, the process. This process of latent heat release is called
something; it is called recalescence. This is nothing but heat release when a liquid
becomes a solid, why? Because the solid is not able to contain all the heat which is their
inside the liquid so, the moment this particular area, this particular volume gets
converted in to liquid into solid then immediately, that particular volume releases some
heat, because the earlier this volume had all liquid and it had a high enthalpy. Now, it
has got converted to a solid, which can contain only a low enthalpy. So, their balance
enthalpy is released, where is it released, it is released into the liquid which is
surrounding it, that is why the moment you see a solid forming, there is a increase in the
temperature, that is what you see in the cooling curves (Fig. 3 in image above).
You have a recalescence, and the moment you have a recalescence then if I say, I have
under cooled the liquid to below T0 temperature and if this recalescence will raise the
223
temperature of the liquid above that of the T 0 then you are in soup, you are in liquid, not
in soup, liquid is also a soup. So, you are in to the liquid back again. So, you again are at
a temperature somewhere above T0, where partition less solidification cannot occur. So,
the remaining liquid cannot continue to do partition less solidification. So, you are in
trouble, so that is why, people now call that the minimum under cooling that you need to
give is not just Tl – T0.
224
Simply, if I know, what is the amount of enthalpy that is released, so if I know that, ΔH f
= ΔCp. ΔT, if I can put this. In principle, I can put that, so, H = Cp.dT. So, similarly . So,
ΔHf = ΔCp. ΔT, where ΔT is the temperature raise because of the heat of fusion. So, if I
I can calculate ΔT recalescence, as ΔT R = ΔHf/ΔCp . So, if I know the ΔC p, that is where
the whole problem comes after the moment, we complete this, we are going to a another
concept possibly tomorrow, where you will see what is the problem with this ΔC p, there
is the major problem the whole another new field of whole thermodynamics has evolved,
just because of this ΔCp problem, tomorrow I will address that.
So, you see that this ΔT R can be calculated, so the actual undercooling that you need to
have is ΔT = ΔT 0 + ΔTR this is what is called hyper cooling concept. This concept is
called hyper cooling. if you can give this much of under cooling, then there is a no
problem at all, because it takes care of even the recalescence also, so you are
undercooling up to here (point F, Refer slide time: 28:06). So that even after
recalescence the temperature does not cross T0, so you basically under cool it to such a
temperature, that even after recalescence the temperature does not cross the T 0. And that
is what is what we call hyper cooling concept, we will talk more about it as we go along.
Thank you.
225
Prof. B.S. Murty
Lecture No. #11

Eutectic Solidification, Glass Formation
Activation energy, that is where you are missing, that any nucleation event and a growth
event needs diffusion; this aspect, if you miss out, we will think that the whole nucleation is
just driving force related, is it not? That is the reason why the whole of TTT diagrams come.
Why do you have a TTT diagram, which has a C shape? The C shape of the TTT diagram
comes from the diffusion problem, is it not?
If you remember, this C shape, which for example, we talk about austenite to pearlite
transformation, why is this shape? This is what we call it as start of transformation (first
curve), this is what we call it as an end of transformation (second curve); so, that means this
is nucleation; this is the end of growth. Why is this, it has a C shape? Basically, because the
nucleation and growth, even for that matter nucleation itself, is a combination of two - one
the driving force ΔG, the ΔG is important and diffusivity D is important.
226
In fact, if you go a little deeper - nucleation kinetics, unfortunately we are not really dealing
with nucleation kinetics; you remember, you have an expression like this; you have an
expression like that; this part (ΔGD) you people are forgetting; that is very -very important
part (see the expression above). Please read Reed Hill, Physical Metallurgy Reed Hill
nucleation, there is a wonderful chapter in Reed Hill; I thought all of you should have read at
the B.Tech level, if not please go through that chapter, and this (ΔGD) is very very crucial.
And if you look at only this part (ΔG*), then you are right; that as I keep on going to higher
and higher under cooling, I get higher and higher driving force, if the driving force is higher
in principle, my activation energy should be lower.
So, the same thing when you are talking about liquid going to single phase α, or liquid going
to two phase mixture, we are assuming that diffusion does not play a role. But we know that,
if liquid has to go to a two phase mixture it needs long range diffusion, but if liquid goes to a
single phase structure of the same composition, then it does not need long range diffusion, it
needs only short range diffusion.
And a short range diffusion always needs lesser activation barrier; that means this part (ΔGD)
is smaller for short range diffusion and larger for long range diffusion, and that is the reason
why ΔGD dominates, you are right, , this is also there, Δ G* is going to be higher, if the ΔG is
going to be lower. If the driving force, so if I look at
227
lC1  αC1 or lC1  lCl + αCα
If this ΔG is smaller then ΔG* is larger here, this is ΔG* only for the nucleus, that’s it. So,
this is what is, in fact, if you carefully look at it, what is ΔG* ? It is ΔG for r*, so actually it
is speaking, what is the change in free energy, when a liquid goes to a solid, with the solid
having a size of r*; that is what we call it as ΔG* , but because that it is a higher we call it as
an activation energy.
The overall activation energy for nucleation, you have to consider a combination of these two
(ΔG* and ΔGD), is this clear? So, as a result, we need to consider that aspect, and definitely
this transformation ( l  α of same composition) will have a lower activation barrier. This is
even true when we come to solid state transformations we will see, why is that when I take a
let us say age hardening, or precipitation hardening, you take Al-Cu standard example, in an
Al-Cu when we are doing precipitation what comes out first
Student: GP zones
GP zones we say.
We don’t, what is the stable state there?

Student: θ
θ is the stable state, so definitely the ΔG, the driving force for the super saturation to
transform to θ must be the highest, because that is the equilibrium state, but it does not
happen. Even you see in cases of steels, when you take a steel martensite and temper the
martensite, I do not know how many of you remember, what is the phase that comes out
during tempering? It is ϵ carbide, it is not the cementite which comes out, why is that?
Obviously, ϵ carbide does not stay, if you do the tempering at higher temperature, or for
longer times this ϵ carbide vanishes.
Similarly, GP zones vanish, if you do aging for a longer time, GP zones turn to θ”, then θ’
then finally the θ, what does that mean; that means, all these initial stages are only transient
phases, transient metastable phases, and the stable phase is that the θ. Similarly in case of
steel, it is a cementite which is the stable state, but cementite is not able to come, we will see
when we come to solid state.
What are the further complications that you get, for precipitation or for the phase
transformation in case of liquid to solid, which is what we are now dealing with, in a liquid to
solid, there is only one barrier for transformation, what is that barrier? It is the surface
energy, and when you talk about the solid state you get additional barrier, that is the strain
228
energy we will talk about that, that is when you realize, why you need this to come out, we
will talk about precipitation later.
229
So, you need to understand that, it is just not the driving force that always leads to, so when
we come to, for example, we will definitely talk about this later. When we are going to talk
about solid state transformations, and go in a later little detail on that, at the moment we will
continue with our liquid to solid, that we are talking about.
Student: In the above expression, can I consider ΔG* as the activation energy?
Yes, that is what I was telling; ΔG* is the activation energy for the nucleus to form, without
considering the diffusion, even for the nucleus. What is the nucleus? Nucleus is nothing but,
a cluster of atoms with an r* size. this is what we call it as a nucleus. And for a cluster of
atoms of r* size to evolve, from the liquid, you need atoms to move, atoms to come together
and form that r*, this atoms to come together to form r*, you need some diffusion that is
required, and that is ΔGD.
Student: Then, when ΔT is more, how do I get fine grains?
That is what I am telling.
So, when I say ΔT is more you get fine grains, we are always making an assumption, that we
are in this domain (near the nose of TTT curve), there is an inherent assumption, which one
has to very careful, that I am in this domain, if I am in this domain as I keep on increasing the
ΔT; that means cooling rate, if I increase the cooling rate, I am increasing the under cooling,
if I am increasing the under cooling, I m increasing the ΔG.
230
So, you can see, the cooling rate is proportional to ΔT, which is proportional to ΔG, and as a
result ΔG* decreases, and the temperature here in this domain (near the nose of TTT curve) is
high enough for the ΔGD not to be really dominating, so because we are still not in a low
temperature domain, for the diffusion to start dominating, so diffusion still can occur because
we are at high temperatures.
And as a result, it is this ΔG* which now decides, what is your nucleation rate, not the ΔGD
which decide the nucleation rate, but it is ΔG* which decides the nucleation rate in this portion
(near the nose of TTT curve). And that is the reason why you will see that finer grain size, and
most of the cases the cooling rates that is we adopt in when I talk about sand mold, or a
permanent mold, or a water cool Cu mold, there the under cooling’s that we are talking are not
really the kind of under cooling’s that can really bring, the liquid below this domain. If you really
want to get the under cooling below that domain, you need to use much higher cooling rates; that
is where you will always see the diffusion dominates and that is the reason why you will actually
see, even glasses forming. The fact that a glass forms from a liquid is an indication that
nucleation is not able to happen, the atoms are not able to come together to form that nucleus, and
the nucleus is not able to grow, if that can happen then why will glass come out.
In fact, long back turn bull has made a profound statement, that every liquid would like to
become a glass, unless crystallization intervenes, it is this crystallization which is coming in
the way, otherwise liquid likes to become a glass, because liquid and glass are of similar
231
structure, nothing need to change for a liquid to become a glass, but it is crystal formation
which is commonly observed, why? Because in most of the metals and alloys, the crystal
structures being is so simple that they can easily nucleate, the atoms coming together forming
an FCC cluster or a BCC cluster is easy, but if your cooling rate is so high that, even that
cluster is not able to come together, for forming of a nucleus, even you do not see, pure
metals also people have seen glasses formation, pure silicon people have quenched at very
high rate, and they have seen it becomes a glass, pure nickel people have shown at as high
rates as 1012 K/s.
Do you know, how do we reach this kind of cooling rates, did we talk any time? 10 6 is what
you get when you do, what is called melt spinning or gun quenching, how do reach this kind
of cooling rates?
Student: Melt spinning
That is what I said, melt spinning is the one, where we talked about it, where you take a
wheel, which rotates at a certain speed like 3000, 4000 rpm and you pour the liquid metal on
it, and then the moment a droplet falls, so, because this wheel is spinning, it pulls this droplet
along with the wheel, so the droplet sticks to the wheel and then gets pulled, as a result the
droplet gets converted into like a sheet of metal, it is like a rolling without a second roll there.
So, and as a result you get flakes, every droplet becomes a flake, and if you have a
continuous flow of liquid in principle, you can have a continuous sheet of, people have
produced even up to almost like a one feet width; I mean it is almost like a continuous
casting, excepting that, in a continuous casting you always have two wheels, it comes out as a
slab, the liquid metal goes and then it is like rolling, here there is no such kind of a rolling
because there is no second wheel on top of it, it is the liquid. Only thing is you control the
flow rate of the liquid, and control the speed of this in such a way that it comes out as a
continuous ribbons. And people have made kilometres of ribbons like this; a particular type
of alloy which is an iron based metallic glass, where large numbers of people have worked on
it. So, here you get a cooling rates of the order of around 10 6 and 1012, anybody knows?
Laser, if you use laser for example, have you heard of picosecond laser? No, picosecond
laser, nanosecond laser, pulsed lasers.
232
So, in pulsed laser, a picoseconds laser basically means, 10 -12 seconds it is active; so, imagine
a solid surface on which you have a laser which falls on it, and it is active only for such a
short time, within that time the laser will melt that surface very small volume, may be a few
micro meters of volumes, and then immediately it solidifies. And if you assumes that, this
melting and solidification is happening within this particular period, it melts, of course there
will be a time lag, so actually it may not happen exactly within that time, but if you assume
that, if it melts and think of an Al being melted let us say, and an Al if it is melted, what is the
temperature to which it goes of approximately let us say 1000 K and then from 1000 K, it
cools back within that time.
Because, the laser is active only that period and it is shut off, so that is what is a pulsed laser.
So, then within that period again it has to cool back, and how does it cool? The whole metal
itself acts as a quenched, in a slab of a metal only a few micrometer cube of a region is
getting actually melted; a small pool is melted, so rest of the metal quenches that particular
liquid.
So, you get the liquid brought back from 1000 degrees to room temperature within this
particular period, and if you consider that, if you think that 10 3 is the temperature change that
is taking place, and 1012 is the time, and you would see the ratio gives you 1015 K/s.
233
In principle, in principle theoretically, one can get this, of course practically it is difficult, that
is why I said 1012, if you use a picosecond laser, because there will be a small time difference,
it has to heat and then cool. So if you consider a small difference, people have been able to
achieve this cooling rates (1012 K/s)
So, in principle you can get that cooling rates and with that people have shown that even pure
nickel can be made into a glass, people have made glasses like that; that means, your time
available is so short for the atoms to even form an r*; it does not mean that clusters do not
exist in the liquid, clusters are there, but their size is not equivalent to an r*, they are like an
embryos. So, they are not able reach the r* size, and if they do not reach the r* size,
obviously you will not have a nucleation, so that is what is the whole concept of the glass
formation and in any way we talk about the glass formation later, and this aspect also we will
talk about it, when we are going to talk about glass formation.
Let us continue where we have left last class; in the last class, we were talking about the T 0
curve, and we talked about a curve such as T0, and we said any alloy that you choose, if you
can bring it below the T0, then we are at a condition, where there is a possibility of
partitionless solidification, I said there is a possibility of. So, whenever I say you bring it to
any under cooling below that, if your under cooling is so high, you
234
bring an under cooling up to this level, and you come actually below the T g, there is even a
possibility of glass formation, we will talk about it after a while.
Student: T0 curve is the free energy of liquid and solid.
Correct.
Student: Here, are we extrapolating that curve?
Which curve, this one?
Student: No, it should be in l + α region.
Yes, that is what I am saying, when you say L + α region so for example, if I take this liquid
(line 1), and under cool to this temperature (point P, below T0), what should happen to that
liquid? Under normal conditions, if the liquid is simply cooled, it would have first formed the
α, and then by the time you reach this temperature, eutectic reaction should have occurred.
Now, you have brought this liquid to that temperature, and your T g for that alloy is
somewhere here (point Q) let say, you are above the Tg, but below the T0, so liquid has no
chance for it to become a glass, you have not yet reached the temperature where glass can
form, I will also talk about glass formation possibly in the next class, we will know a little
more about this Tg concept. But, at the moment if you understand that this is above that, so
liquid has to, and at the same time liquid is not stable, at this temperature according to the
phase diagram, what is stable? In principle, it is α + β that is stable at that temperature, so
liquid is no way, it has a chance to be stable there, so liquid wants to transform now. So, how
does it transform, to what it will transform, is what we are looking at.
Again the same concept if because, we are at this temperature, if I extrapolate the T 0 curve, I
will see that my composition at this temperature (point P) is to the left of the T0, so there is a
possibility that, that liquid which is staying there at that temperature has a chance to become
solid by partitionless solidification. But, if the same liquid is somewhere here (above T 0) no,
or if I choose another alloy (line 2), and bring it to the same temperature (point P’), this alloy
has no chance for it to become a partitionless solidification, because now it is on the right
side of the T0 curve at the same temperature, and whenever we are on the right side, there is
no driving force for partitionless solidification.
So, alloy number 1 has a chance for becoming a single phase α by partitionless solidification,
and that is what it happens, and now if you look at compositions like alloy 2 what should
happen to these compositions? It has no driving force for partitionless solidification, at the
same time it is at a temperature, where the liquid is not suppose to be stable, because we have
235
brought it to this temperature (point P’), at this temperature liquid cannot remain as liquid, so
the only chance for it is and at the same time, it is also above the Tg.
236
So, liquid has only one solution, that it has to become a solid by partitioning, when the liquid
is taken, the liquid  liquid + solid. At this temperature once I bring this liquid, it has to give
you liquid + solid, either liquid + α or liquid + β, depending on the situation, and depending
on which side of the phase diagram you are.
And now which will be composition of that α, which will be the composition of liquid, how
will I know at that temperature? I have brought this to liquid to this temperature (point P’), at
this temperature it has no solution, but to undergo a partitioning and it should have a
solidification along with partitioning, and such a partitioning when it is taking place, what
will be the composition of liquid, what will be the composition of α? How will I get? Let say,
if I, if this alloy is at this temperature, this alloy, alloy 1 is at that temperature (R), at that
temperature also it is above the T0, is it not? Above the T0 and if I draw a line at that
temperature horizontal line, it is on the right side of the T 0, so that means, the partitionless
solidification is again ruled out.
So, that alloy will now undergo a solidification by partitioning, what will be the composition
of liquid, what will be the composition of α at that temperature? How do you see?
Student: Lever rule
By not the lever rule, lever rule talks only about amounts of liquid, amounts of α, it is that tie
line and wherever the tie line intersects the α and the liquid that would be the composition of
α, the composition of the liquid.
So, α of this composition (Cαl) will come out of the liquid at that temperature and the
remaining liquid will get this particular composition (C lα), and that is how you will have a
two phase mixture of α and liquid. Now, if I am considering another alloy, alloy 2 and I am
saying that this alloy 2 is brought to this temperature (point P’), again the situation of this
alloy 2 at this temperature (point P’) is similar to the situation of alloy 1 at this temperature
(point R), there is no difference. Now, what will be the composition of liquid and α? The only
difference is, I am now below the eutectic temperature, earlier I was above the eutectic
temperature. Now, what will be the composition of the liquid and the composition of α?
Student: Eutectic
Why? Liquid is not eutectic, liquid composition different from eutectic composition.
Student: Again draw a tie line.
Again draw a tie line; yes, Now, I drew the tie line (through PP’) now, now this tie line, when
I extend this tie line, it is intersecting α and β, but I am talking of liquid and α, liquid is
giving me liquid and α, liquid is not giving me α + β.
237
Student: The α composition is that, liquid composition cannot go beyond that.
238
So, how do you find out the composition of α and liquid now? To find out you need to look at
whenever I am talking of liquid and α, I need to look at what indicates the equilibrium
between liquid and α? What is that particular composition which talks about liquid α
equilibrium, liquid α equilibrium is decided by the liquidus and the solidus, these are the two
which talk about liquid and α equilibrium; the solidus gives you the α composition, the
liquidus you the liquid composition. If you want to know, what is the composition of α and
what is the composition of liquid at any temperature, what you need to do is, you need to
extrapolate those curves. So you extrapolate the liquidus, you extrapolate the solidus, and the
compositions because what I mean to say is, at this temperature (point P’), if α has to come
out of liquid, imagine there is a liquid, inside that α is coming out, if this liquid has to be in
equilibrium with α, it can be in equilibrium with α only by this, solidus is the only line which
tells me what is the composition of α that can be in equilibrium with liquid, no other α
composition can be in equilibrium with liquid.
So, this whole domain (indicated by hatched area) talks about liquid α equilibrium and so if I
want to know what is the composition of α that can be in equilibrium with liquid? It is only
this composition. Because if I draw the free energy curves, common tangent between α and
liquid if I draw, the solidus tells me the composition of the common tangent for the α, and the
liquidus tells me the composition of the common tangent for the liquid. And so at some other
temperature, at a lower temperature if you want to look at it basically I extrapolate that, and
these extrapolations will tell me, what will be the composition of α, what will be the
composition of the liquid.
So, it is the α that comes out at this temperature will be having that composition (point K),
and the moment α of that composition comes out the liquid which is remaining, will tend to
have this composition (K’). And once that happens now you see the liquid composition is
beyond that of the eutectic composition actually. So, because the liquid composition is
beyond that of the eutectic composition, there is a possibility that, the remaining liquid which
has a richer in B atoms can start nucleating the β, and there is a possibility that the remaining
liquid gets converted to β, because we are on the B side now. The liquid composition is on
the B side now, so because it is rich in B, the liquid is rich in B, it can easily nucleate β,
which is the β B rich phase.
239
So, you will see that in all these cases, though the α that comes out initially will be of this
composition (point K), at some stage you will see the remaining liquid gets converted to β, so
you will have a situation of β matrix, and an α phase. And as the, as you hold for a longer
time as you cool to lower and lower temperature, this α slowly changes to the equilibrium
composition of α, because the moment β comes out, α of this composition (K) cannot be in
equilibrium with β; if α has to be in equilibrium with β, only α of this composition (K”) can
be in equilibrium with β.
So, as a result, the α composition shifts from here to here, so that it can be in equilibrium with
β, and you will see that slowly and because of the composition is becoming lower in B
content, so the excess B atoms which are available in α, will be rejected out of α, and will go
and join the β, so that the β quantity will increase and then α composition will get shifted
towards richer A and that is how the whole process happens. This is all assuming that
partitionless solidification is not taking place, if we have a situation where partitionless
solidification is not taking place, and you are at temperature below the eutectic.
If you are above eutectic, all this problem does not happen, here at this temperature (point R)
you will the α, and as you cool to lower and lower the liquid composition changes, and then
finally, it reaches eutectic and that liquid will give you a eutectic mixture like an like any
other solidification. But, if I have undercooled the liquid to a temperature below the eutectic,
and solidification is now about to start, what should happen to this liquid? It cannot straight
away give you the eutectic mixture, because the liquid composition is not a eutectic
composition, you know please remember eutectic reaction is an invariant reaction.
What is the meaning of invariant reaction? Invariant reaction means everything is fixed, the
composition of liquid that can undergo a eutectic reaction is fixed, every liquid cannot
undergo; you may say, sir the alloys left and right to eutectic points also will have eutectic
reaction, why do you say that liquid should have this eutectic composition only? The alloys
to the left or right is undergoing eutectic reaction, but the liquid of that alloy which is
undergoing the eutectic reaction always has only eutectic composition, please remember that.
If I take this alloy (line 3), this is the alloy I am choosing, if I slowly cool under normal
conditions you will have the α coming out, and as I come to lower and lower temperature,
more and more α comes out, and the remaining liquid, its composition shifts towards the
eutectic, and then
240
comes to the eutectic composition, so the moment I reach this temperature the liquid that is
remaining inside the liquid + α mixture will always have that eutectic composition, and it is
that liquid which is undergoing a eutectic reaction, not the liquid with which you have
started. So, the eutectic reaction always the liquid which under goes the eutectic reaction will
have this eutectic composition, because it is A rich phase that is coming out liquid becomes
richer in B, if you are on this side (B side) , it is β that comes out first, and as you come to
lower and lower temperature, more and more β comes and because β is B rich the liquid gets
depleted in B.
All the B in liquid goes in to the β and as a result the liquid composition changes, and then
once you reach eutectic, the liquid composition is always that, so whether if you choose this
any alloy within the eutectic domain, the eutectic reaction always happens between the liquid
of a fixed composition, and that composition is the eutectic composition. So, that is the
meaning of an invariant reaction, you cannot have any liquid giving you a eutectic reaction,
so because we are now talking about this liquid… yes, it is again the free energies.
If you draw for example, if I draw if I draw, what is the meaning of eutectic?
L  α + β, that means, one phase is giving you two phases, there are three phases in equilibrium,
it is not? So, if at that temperature eutectic temperature, if I draw a free energy composition
diagram, I will see that you will have one liquid curve, one α curve, and if I
241
draw a common tangent, you will see at that temperature, if I draw at eutectic temperature T e,
if I draw the free energy composition diagram, I will see a free energy composition diagram
like this (see the image above), where all of them having the same common tangent, and what
does that mean? It means that, where ever this tangent intersects here, on the y axis (B-side)
I can call write
μBl = μBα = μBβ.
Similarly, on this side, I can write

μAl = μAα = μAβ
that is what is the definition of equilibrium here, the chemical potential of A in α, β and liquid
all of them are equal, similarly chemical potential of B in all the three phases is equal. And
now, what decides that composition? It is this composition, where ever this common tangent
is taking place, the composition corresponding to that common tangent will be the
composition of the liquid.
So, if you draw these free energy curves at that temperature, draw a common tangent
whatever composition corresponds to the liquid composition of that common tangent will be
the eutectic composition, and similarly α composition and β composition will be where
common tangent intersects respective free energy curves, at the eutectic temperature; and
these three compositions are reflected in the phase diagram at eutectic temperature.
So, everything is fixed, like we have we drew a unary phase diagram for H2O, we saw that
the pressure and temperature at which all the three phases can exist is fixed, is not in your
hands, similarly the compositions of α, liquid, β is fixed and it is not in your hands, it is
decided only by the temperature. Now if I come to a lower temperature, at a lower
temperature in principle liquid is not stable.
What should happen I, when I decrease the temperature what should happen to this free
energy curves? They will go up, which one will go up, α and β will not go up, everything will
go up, this has to be remembered, it is not that liquid only goes up, α, β do not go up, it is
only the rate at which they move up that matters for us. Yes,
Student: Coming to the equation of l  l + α, isn’t the concept of tie line when there is an
equilibrium?
correct, see the question is whenever we talk of equilibrium there is always something called
local equilibrium which has to be maintained, unless you are a at such a highly non
equilibrium states, that where atoms cannot move at all like a martensite transformation,
242
diffusion less transformation that is taking place, or a liquid giving you a glass, that you are a
such a temperatures, where no diffusion can occur, atoms cannot move, then local
equilibrium fails
243
at all other temperatures liquid tend to have, that is why the moment the α forms what
happens is a thin, may be it can be even a mono atomic layer, it can be a two atomic layer
thickness, but that thickness within that thickness of liquid around it will have a composition
which is decided by the tie line.
And if you think that the temperature is like very small like absolute 0, or very close to room
temperature, even room temperature whenever we talk of temperature, we should remember,
what that we are talking about. For example, you many of you know that deformation at
room temperature for lead we call it as high temperature deformation hot deformation.
So, with respect to what we are talking; that means, with respective which material where
diffusion is easier or faster or not easier; that means, with respect to the melting point of that
metal we are talking about. Similarly, if we are talking about a material which has a
sufficiently high melting point and if we are talking about a temperatures, which are close to
room temperature, where the liquid, inside liquid atoms cannot able to move, to be able to
come to that temperature, that composition decided by the local equilibrium, then what would
happen is liquid will not give you that α.
244
Even for liquid to give you that α atoms have to come together, and form a structure which is
an fcc or a bcc structure having this composition, because now at this temperature (Point P,
Refer slide time 17:02), if I simply draw the free energy composition diagram, I have liquid,
now this is a liquid of certain composition that I am choosing (vertical line in the above
image), this liquid what should happen to it, if I look at it? Then I should look at what all the
other phases that are possible, can I draw the free energy composition diagrams for them?
And if I draw the free energy composition diagram for α let us say, I may see that yes there is
a driving force now. Just because there is a driving force will it form or not is decided by the
structure of that particular phase, how complicated or how easy it is, whether it is able to
nucleate or not. If it can nucleate, that means, the diffusion is able to take place, even if it is
localized diffusion and if localized diffusion can happen, then local equilibrium has to be
established.
If local diffusion also is difficult; that means, short range diffusion is also difficult, then there
is no chance of even local equilibrium.
Student: Will the phenomenon like coring be accentuated if you undercool the liquid?
Yes, coring definitely happen, you see, but the thing is the coring we are assuming, that you
know you have a α at each temperature as you go below of different different compositions
nucleating, this is assuming that our temperature is such that, we are above the T 0, if you are
below the T0 there is no question of coring; the α that comes out will be same composition as
that of liquid, absolutely there is no coring, ok, clear.
245
So, coming back to this problem, if we are at a lower temperature, then you would see that all
of them are going up, only thing is liquid will go up faster, why should it go up faster?
Student: Because more entropy
Yes remember G verses T curve, the slope of it is entropy and as a result, so if you change
the temperature by 1 degree, how much the G will go up is decided by the slope. So, you will
see that, if I decrease the temperature by 1 degree, the liquid will go up faster when compare
to solid because both the solids have a shallower free energy curves, free energy verses
temperature curves, so because the entropies are lower, so they will not raise at the same
level as a liquid; so, as a result when I decrease the temperature, this starts going up.
So, that means, and all of that are going up, let us say they go up like this (indicated by
dashed lines), but liquid goes up faster. If it goes up faster, then you will see a situation, now
I cannot maintain a common equilibrium between all the three, common tangent between the
all the three is not possible, the only tangent that is possible, of course you can have three
tangents; you can have a tangent between α and liquid, you can have a tangent between β and
liquid, you can also have a tangent between α and β, and if you look at which of the tangent is
the lowest, which tangent has a lower free energy? You would see, the α to liquid will be like
this (PQ), the β to liquid will be like this (QR), and α to β will be like this (PR), and because
α β free energy I mean, common tangent is below that of the other two, we say that α-β is
most stable state. That is why when I come to a lower temperature below eutectic, it is α and
β which is stable, and that is
246
what you will see; come back, you see we our interest is not this,
our interest is to see what happens if I take an alloy of this eutectic composition, I have
under cooled it to a very low temperatures, and to an extent that as I keep on under cooling it,
the T0 curves of the α and β are such that, their slopes are such that, this composition is not
intersecting any of them, if that is the case when you reach a low temperature, at that stage
where we are just now talking about it, even local equilibriums cannot be maintained because
atoms are not able to move, and at such stage, this liquid will have a tendency to form a glass.
And there is a temperature below which this will happen and that temperature is what is
called a glass transition temperature.
247
And that comes from a concept of if you draw the viscosity as the function of temperature,
viscosity of the liquid, you will see that liquid viscosity above the melting point is usually of
the order of around centipoise (10-2). Most of the metal liquids metallic liquids just above the
melting point, if you measure their viscosity, they are all of the order of centipoise where as if
you take silicate glass liquid above its melting points, take any silicate glass type of materials
any silicates, and take them above the melting point and measure there viscosities, the
viscosities of them will be the order of 102 - 103 poise. We are all talking of above the melting
point, whatever may be melting point, so that what does it tell you?
It tells you that silicates are highly viscous when compare to metallic liquids, the reason is the
3-dimensional network kind of clusters that are there in those liquids, the silicate clusters that
are there in the liquids will make this liquid very viscous, and thats the reason why we always
say glass formation in these liquids easier, because for them to nucleate a silicate solid,
crystalline solid is more difficult when compared to metals. Atoms can easily move because
they are less viscous, atoms can move because this is more viscous, atoms are not able to
move, and now if I plot how this viscosity changes with temperature, you will see the
viscosity changes almost like this (see the graph above), this is starting from about melting
point, 10-2 to it goes a temperature level of around 1013 - 1014 poise. And this is a kind of
viscosity of solids. The mechanical
248
definition of a liquid is that liquid, which are shear modulus 0 is what we call it as a liquid, it
can easily flow if shear modulus is 0. So, a solid is the one which shows some shear modulus,
so that is actually related to the viscosity, so the moment the viscosity reaches that value, then
immediately this liquid has to become a solid. So, it is a mechanically a solid, it has a shear
modulus now, but it structurally is a liquid, because you are at a temperature were there are
no atomic arrangements can take place, atomic rearrangements, so atoms remain were they
are, that is why we called it as a configurationally frozen liquid. So, glass is what we call it as
a configurationally frozen super cooled liquid, so the movement you bring the liquid to this
temperature, which is what we call it as T g, you suddenly see that the liquid freezes into a
solid, but into a solid which has structure similar to that of a liquid, and that new type of solid
which we have not seen before, because of the most of the solids that we know are all
crystalline solids, now you have a solid which is non crystalline, which has a structure similar
to that of a liquid, we call that as a glass. This is what is one way of looking at glass
formation, we will also possibly look at the other way of looking at how the glass forms, and
I can achieve this particular thing by basically a very fast cooling, if I cool because we are all
talking of under cooling the liquid, please remember I am not simply cooling the liquid, I am
under cooling.
So; that means, I am able to achieve, if this is the Tg for this phase diagram (see the image
above), if I can under cool the liquid below this T g and at the same time, not being able to
intersect the 2 T0
249
curves, then I will see that particular liquid will because at this temperature, partition less
solidification cannot occur, because I am above the T0 for both α and β.
So, the liquid is able to remain as a liquid, it is not able to become a partial solidification to
single phase α or single phase β, now what should this liquid can happen? The liquid can only
undergo solidification by partitioning, or for that to happen if diffusion is difficult the only
choice for it is to become a glass. And because we are at temperatures where diffusion is not
able to take place, because liquid has already configurationally frozen, because atoms are not
able to move, so as a result that particular liquid at that temperature would become a glass. If
I am at any temperature here (above Tg), if I have under cooled the liquid to that temperature
it can still become solid by partitioning, it can start giving a α or β whichever has a lower
activation barrier for nucleation, it will start forming that and then it will continue, but the
moment I bring it below the Tg, this liquid has no chance, but to become a glass.
And that is where we start defining what are called deep eutectics and shallow eutectics.
There are some phase diagrams which are like this lets say, were the T 0 curves look like this
(see the image above), they intersect with each other, and if the T g is somewhere here (below
the intersection point) by the time the liquid eutectic liquid is brought below the Tg it already
intersects the T0, either the α T0 or β T0, and once that happens, we will see either α or β
comes out, we will continue this later and see more of how to understand the glass formation.
250
Prof: B.S. Murty
Module #01
Lecture #12
Kauzmann paradox, Order of a transformation, Glass forming ability
We will continue where we have left last class. And, we were talking about glass
formation. We basically talked about the T0 concept, basically gives two things. One,
what is called extension of solid solubility. As you can see, if you choose any
composition, for example, if I redraw this T0 and choose a composition of this sort (line
1 in above phase diagram) and can bring the liquid below the T 0 in such a way that, I
take care of even the temperature rise, during recalescence also. That, if I can undercool
the liquid sufficiently so that, it can undergo what is called partitionless solidification.
Then, I have an alloy here, which has the same composition as the liquid that which,
which with I have started. And, this is a single phase solid solution now. So, I have a
single phase solid solution, which has much higher solute contain than it can actually
contain according to the phase diagram.
251
So, that means, basically what I have done is, I have got extended solid solution. So, this
is the most crucial aspect of the T0 concept. And, that is what you get, whenever you
undercool. So, undercooling can give you variety of things. One thing that we have seen
earlier, in the beginning itself that undercooling can lead to a higher nucleation rate and
can give you finer grain sizes. So, fine grain size that we usually see at faster cooling
rates is basically coming from higher undercooling. But, yesterday if you remember, we
talked about it, this particular undercooling and higher nucleation rate; we are only
concerned with the upper part of the C-curve. When we actually come to the C-curve,
later, we will look into it. And, if you are below that, then, basically this concept that,
higher undercooling leading to higher nucleation rate is actually not right; because it is
below the half part of the C-curve. The bottom half part of the C-curve, it is diffusion,
which takes control. And, as a result, just because you have a higher undercooling, does
not mean that you will have higher nucleation rate.
So, basically, as that top portion of the C-curve is concerned, then we are right. We can
say that the higher faster cooling rate or higher undercooling will give you fine grain
sizes. But, that has nothing to do with the T 0 concept. Just, higher undercooling leading
to a higher nucleation rate and as a result, finer grain size. And the T 0 concept, basically
tells us that, yes, we can get extended solid solutions. And, if you can choose your
composition as an appropriate composition, you may end up in getting a glass; provided,
for example, if you have two T0 curves like this (see the diagram above) and if you are
choosing an alloy composition (line 2) and in fact, we call this range (PP’) as the glass
forming composition range, wherever the two T0 curves intersect the Tg line. More or
less, Tg is not a strong function of the composition of the alloy in any system. So, as a
result, you see it more or less, as a horizontal. And so, wherever the two T 0s intersects
that Tg that range will give you the range of compositions; where liquid is more stable
and liquid has no chance to undergo partitionless solidification. As a result, this liquid,
once it reaches the Tg will undergo glass formation. So, that is why we call this as glass
forming composition range.
But, all this you have to keep it in mind; is only we are talking from the
Thermodynamics point of view. In principle, you can even have a larger glass formation
composition range, if you have used much faster cooling rates. That kinetics takes over
that you are somehow preventing the crystal to form. Ultimately a glass will form,
252
whenever a crystal is suppressed. So, if you can suppress the crystal formation from the
liquid, you can always get a glass. So, but this (PP’) gives you a tentatively glass
forming range. So, by just looking at the phase diagram we will be able to tell which
phase diagram has a higher glass forming ability, which does not; provided, I can plot
the T0s. And, I have also mentioned to you sometime back. We will also talk about it
when we continue with glass formation. What is a glass forming criterion? One of glass
forming criterion involves the phase diagram itself. How deep is the eutectic or how
shallow is the eutectic?
So, that you can see from the phase diagram itself and we have talked about it, a criterion
based on that, what is called the deep eutectic criterion? We will come to it, as we go
along. So, this glass formation is a second concept that you get from the T0.
And, yesterday in the last class, we started talking about it and we looked at, how the
glass forms as one of the parameter, which is the Viscosity. We looked at how Viscosity
changes with temperature. And, we said there is a particular temperature at which the
Viscosity of the liquid reaches almost that of 1013-1014. And, that particular temperature
defines what is called the Tg because below that temperature, liquid cannot be under
cooled; because by the time, the atomic movements in the liquid freeze and as a result,
this liquid become a glass. Now, let us try to understand this glass formation from the
Thermodynamics point of view and what we have seen, viscosity is mostly like a kinetic
point of view. And, look at it from the Thermodynamics, how do
253
we understand glass formation? Glass formation, we understand from the Entropic
criterion. If, I look at how the entropy changes as a function of temperature, we all know
that the entropy of a liquid is higher than the entropy of the solid. Am I right? So, S l > Ss.
But, at the same time, we also know that the rate at which the entropy of liquid changes
with temperature is also higher than the rate at which the entropy of solid changes with
temperature. Why? Because we know pretty well that S = ∫CpdT/T and, as a result, the
dS/dT = Cp/T. And because we know that Cp of liquid is always higher than the Cp of the
solid.
So, I can expect the slope of the entropy versus temperature curve to be steeper for the
liquid, when compared to the solid. This is understood. If that is the case, if I plot
entropy versus temperature for any metal, for example, I can show that the entropy
changes something like this. Of course, there is a mistake in this curve. What is it?
Student: These are not straight lines
These are not straight lines. These two portions, which I showed them as straight lines
are actually not straight lines. If I say that they are straight lines, basically I am saying
that the slope is constant. The slope is constant means C p/T is constant; Cp/T is not
constant, it varies with temperature.
So, as a result, in principle, that these two portions, this portion and that portion are
actually not straight lines. One has to remember that. But, for the present understanding,
we do not need to bother about that, that much. But, what we need to understand is that,
upper slope is higher than lower slope. If that is the case and what is this that, at the
melting point, in principle at the Tm, there is something called ΔSf which is called
entropy of fusion. So, in fact, entropy of fusion is known to be ΔSf = ΔHf/Tf; because
dS = dH/T. If you look at the second law of Thermodynamics, basic definition of
entropy, where dS = dH/T; so, you can always write that ΔSf = ΔHf/Tf.
So, if I know the heat of fusion and if I know the melting point, in principle, I can find
out what is the entropy of fusion. So, once I know that now, let us look at the nature of
the two curves on the liquid side and on the solid side. Because the slopes are different,
if I now undercool the liquid below the melting point; that means, the liquid remains as
liquid below the melting point and its entropy can be treated as nothing but an
extrapolation of this liquid curve. So, if I want to know
254
what is the entropy of the liquid at some temperature below the T m, I simply, basically
extrapolate this (dashed line). And say that, at a particular temperature here (Line 1), the
entropy of liquid is this, if it remains as liquid.
So If I keep extrapolating that to a very low temperature I will come to a situation where
at some particular temperature, suddenly I start seeing that below that particular
temperature the entropy of the liquid becomes lower than the entropy of the solid. This is
which is an absurd. It is like a paradox. And, this particular paradox is what people call it
as” Kauzmann paradox”. It was Kauzmann in around 1950s, possibly 1954, who pointed
out at this paradox; he says there is a temperature below which the entropy of the liquid
actually appears to be below that of the entropy of the solid, as it keeps on undercooling.
And, that particular temperature at which you see this happening will be the temperature,
where liquid wants to avoid that kind of a paradoxical situation. And, the only way it can
avoid is to simply become a glass. And, that is what we called it as the ideal glass
transition temperature (Tg). The Thermodynamic definition of a glass transition is that,
where liquid freezes into some solid structure in order to just avoid, situation of lower
entropy than that of the solid. So, just to avoid this particular paradoxical situation, liquid
becomes a glass. And, that particular temperature is what we call it as the glass transition
temperature.
So, this is a Thermodynamic definition of the glass transition temperature. And, in

principle, we can even calculate this glass transition temperature very easily. How do
you calculate the glass transition temperature? Basically, find out the temperature at
which ΔS = 0. So, the definition of a glass transition is ΔS = 0; ΔS between the solid and
the liquid. So, calculate the entropy of the solid and liquid using their respective C p
values. And, once you know this, to find out that particular temperature, where this
entropy of the liquid becomes equivalent to the entropy of the solid and that is the
temperature, where you can say, one can call it as Kauzmann temperature.
People also called this Tg as Tk; Kauzmann temperature or the glass transition
temperature, ideal glass transition temperature. And, so this is how one can in principle,
calculate what is this glass transition temperature. And theoretically, in a number of
cases, if you are just simply looking at pure metals, this temperature is actually very-very
close to absolute 0. That is the reason why, most of the pure metals you do not see that
they
255
become glasses. But, in case of alloys, yes, you will see that it is easier to become a glass
in case of alloys; because in case of alloys, the entropy is still higher; because of the
configurational entropy being added in the liquid. So as a result, it is the curve is much
steeper. So, as a result, when you see that the temperature at which entropy of the liquid
intersects that of the solid is higher; for example, if you imagine that this is steeper (line
2), you will see that it intersects somewhere here (Tg’).
So, that means the actual Tg is higher. So, you will see that is the reason why, you will
see most of the bulk metallic glasses, glasses which are very-very stable glasses, their
glass transition temperatures are actually higher. So because they form glasses very
easily and, there are number of parameters to look into that. We will see it, as we go
along. So, this is one of the very important concepts, which comes from
Thermodynamics.
But, if you look at this particular thing a little more carefully, what you see is that, there
is a situation; whenever a glass is being formed the ΔS = 0 and what about ΔG? What
about ΔG? ΔG, for any phase transformation, in principle has to be 0; otherwise, there is
no phase transformation. If you treat liquid and glass as two different phases, at any
given phase transformation ΔG = 0. So, if that is the case, if ΔS and ΔG are 0, in
principle, obviously, ΔH = 0. So, that is a very-very special transformation. In fact, this
is what we called it as a second order transition. Have you ever come cross
256
first order and second order transitions? Anybody who has heard what is the first order
transformation?
Student: First order involve latent heat
First order involves latent heat; second order does not involve latent heat. Anybody, yes
you are partly right first order transition involves a latent heat of fusion.
That means that there is a certain ΔH associated with the first order transition. Whereas,
the second order transition is the one, where there is no ΔH associated with it. That
means there is a transition, but there is no ΔH associated. For example, if you take a
sample, let say a metal, an aluminium and keep hitting it in a DSC, how do you
recognize it is melting in a DSC?
Student: By seeing Endo-peak
Endo-peak you will see, am I right? A solid, when it becomes liquid, obviously it
absorbs heat at a particular temperature. So, you see a peak corresponding to this
melting. And, the area under the peak actually is what is related to of course, with some
scaling factor is related to the ΔHf.
So, any melting transformation or any transformation, for example, even a simple
example transformation, such as a BCC Fe changing to an FCC Fe or an FCC Fe
changing to a BCC Fe. All these transitions, they all will have a ΔH associated with it.
And, we can easily see them. Only thing is solid-solid transformations will have a
smaller ΔH; solid to liquid transformations will have a larger ΔH because the two phases
have largely different enthalpies. And, as a result, the ΔH is larger. Similarly, when a
solid is actually becoming a gas by sublimation, you will see that the ΔH is much larger
because the enthalpy of the gas is much different from the enthalpy of the solid. So, the
enthalpy difference actually is related to how much is the change in the structure that is
taking place from one to the other. Particularly, if it is simply an FCC to BCC
transformation, the ΔH is usually very small.
So, this is how we can see the ΔH; whereas in a second order transition, you won’t see
this happening. That is why, for example, if I take a glass and simply heat a glass in a
DSC, what we usually see is that, if I heat a glass, I see a plot like this (see the bottom
graph in above image); where here is temperature, where here is some ΔH, heat evolved
or heat absorbed, whatever it is; depending on the direction. So, you see, as I heat, you
257
see a step somewhere here and then you would see a peak. What does this step and what
does that peak correspond to?
Student: First one is glass transition.
Yes, the first one is what is called the glass transition called Tg.
258
And, the second one, where here at this point, you start seeing a peak, we call it as onset
of crystallization. Then, what is happening at the glass transition? When I am heating a
glass, at the glass transition what is actually happening? And, what is that, that is
becoming a crystal later? That means, between this T g and this, if I call it as T x, what
exists between the two, two temperatures, what is it? Something is, see we say, when I
say glass transition; that means, glass is changing into something and that something is
crystallizing at Tx. What is it changing it to? If you remember, what has changed it to
glass in the previous plot that we have drawn (top curve in above image), previous curve
that we have drawn, where you see the entropy? So, this is what is T g, what is it above
the Tg? It is liquid, it is supercool liquid, It is a supercool liquid, which is transforming
into a glass at Tg.
So, when I start heating the glass, this glass will again transform to a supercooled liquid.
It will transform to a supercool liquid above this temperature (T g), when I start heating
the glass. And, this supercool liquid because it is not stable, it is much below its normal
melting point. And because that liquid is not stable, it starts transforming into a crystal.
Depending on, whichever crystal can easily nucleate, it can nucleate to some crystal and
that is what we call it as crystallization. So, what you have here between these two is a
supercooled liquid. In fact, that the temperature difference between the T x and Tg is one
of the measures of how strong is a glass.
259
The larger this difference, which we call it as ΔT x = Tx - Tg. This ΔTx is one of the
measures of, how stable the glass is; that means, the glass does not want to crystallize. In
fact all, what are called silicate glasses; they do not want to easily crystallize. You will
see a large difference between Tg and Tx. Similarly, all bulk metallic glasses are all like
that. That, they can become a supercool liquid and they do not to want to become
because the viscosity of the liquid, the atomic movement in the liquid is so difficult.
Imagine a bulk metallic glass with some six elements in it, the atomic movements, the
diffusion is so difficult. So, this atoms which are present in this liquid, do not want to
crystallize to form a unit cell. So, formation of a nucleus and its growth is so difficult in
such liquids. As a result, they remain as liquid for a longer time or a higher temperature.
That is the reason, why you see this ΔT x can be as high as about 100 K. Most of the bulk
metallic glasses have a large ΔTx. This is one of them. So, we can easily see, whether the
glass is stable glass or not, based on this. So, but what is more important for us for the
discussion at the moment, you see that, you see a step there. The moment you see a step
(circled in the plot above) there; that means, it is not associated with any ΔH. You do not
see either an exothermic peak or an endothermic peak there. So, you will only see a step
and that step indicates that the ΔH = 0.
And, if the ΔH = 0, then what is that, that is happening, what is the change that takes
place there. To understand that, we need to understand what is the meaning of first order
and second order. Let us look at that. We call something as a first order transition, when
the first differential of the free energy is discontinuous at the transformation temperature.
260
What do I mean by first differential of free energy? If you look at free energy, and look
at dG = VdP - SdT, I write this Maxwell’s equation;
∂G
|
∂T p
=−S
∂G
∂ P|
=V
T
So, these two will be discontinuous at the transformation temperature; that means, if I
plot S versus T, you will see a situation like this (Graph 1). At the transformation
temperature, there is a discontinuity between the entropy of the liquid and the entropy of
the solid. I mean, I am talking of liquid solid as a transformation; any transformation you
take, product and the parent phase. So, between the product and the parent phase, there
will be an entropy difference. But, if I plot the G versus T, you do not see such a
discontinuity.
For example, if I plot, you remember this, this plot we already done (Graph 2) where we
have the liquid free energy curve and the solid free energy curve. And, if I want to see
what are the stable states beyond this temperature (melting point, intersection point,
Graph 2), it is the liquid that is stable; below this temperature, it is the solid that is stable.
So, if I look at free energy of the stable states, you see that this is the free energy of the
stable state below the melting point; this is the free energy of the stable state above the
melting point. So, if I simply look at the free energy curves of the two stable states and at
261
the transformation temperature, there is no discontinuity there. There is only a change in
the slope.
262
So, you will see the curve is continuous; excepting that, at the transformation
temperature, the slope of liquid and slope of solid are different and the slope actually is
entropy. That is the reason, why you will see, at the transformation temperature there is a
step in entropy curve because there is a slope change in the free energy curve. So, you
can see the first differential of the free energy, free energy itself is continuous at the
transformation temperature, but the first differential of the free energy is discontinuous.
And similarly because if S is discontinuous because S is related to H, S = H/T so, H also
becomes discontinuous. That is the reason why, if I plot H versus T, I will again see this
same discontinuity. So, because entropy is discontinuous, automatically enthalpy
becomes discontinuous.
And similarly, If I plot V versus T, you will also see a discontinuity of V versus T. That
is the reason why, we always talk about a volume expansion or volume contraction
during the solidification. Why is that because the volume of the liquid is different from
that volume of the solid at the transformation temperature. This volume can change with
temperature because we talk about volume expansion. So, there is a volume expansion
that can occur in the solid state, there is a volume expansion that can occur in the liquid
state. But, at the transformation temperature, if I look at it, there is a difference in the
volume. And, that is why we talk about shrinkage in a foundry. For example, we talk
about shrinkage allowance all that.
So, if I plot S, H, V, all three of them have discontinuity so that, those are all the first
differentials. Now, if I say a second order transition; second order transition is that
transition where the first differential is continuous, while the second differential becomes
discontinuous; that means what? That means, the first differential is continuous; that
means H, S and V, all of them are continuous; that means, ΔH = 0 for the
transformation; ΔS = 0 for the transformation; ΔV = 0 for the transformation.
In fact, this last one is a very-very important property which people are using in bulk
metallic glasses. To make any shapes that you want, people are making what are called
micro gears, the world’s smallest gear, if you can go to the Google and see, which is of
the order of
263
100 microns size has been made by bulk metallic glass. The basic reason is that the ΔV =
0. Once the ΔV = 0, when a liquid is transforming into a glass, there is no volume
expansion or volume contraction. So, as a result there is no actually the shrinkage or the
expansion that we talk about. So, whatever is the mould that you create exactly the liquid
will fit into that. So, what if you can create a small mould by whatever means that you
want; for example, Lithography you can use, there are variety of techniques that are
available nowadays to make very small moulds. And, you put liquid into it and allow
that liquid to solidify into a glass, you will see that exactly it comes with that. So, people
are making very precise dimensions of glasses using this particular one single property
that the ΔV = 0.
And the ΔV = 0 comes because it is a second order transition. So, all second order
transitions will have ΔV = 0. And if, ΔV = 0, then what is the second differential? Look
at what is the second differential from this. If S is the first differential of free energy
versus temperature what is the second differential?
Student: Cp/T
Cp/T.
2
∂ G
So, 2 =C p /T dou square G. So, it is Cp which becomes actually discontinuous. So, the
∂T
Cp of a glass is different from the Cp of the liquid at the transition temperature. If you
really want to find out that what is this glass transition, what property I should measure
264
to see that actually there is a transition that is taking place. For example, whenever you
want to use Thermal analysis to find out
265
solid-solid transformation, you find it very difficult. Why because the enthalpy that is
released or absorbed is so small that you actually do not see in the cooling curve, a major
slope changes.
Whereas, when a liquid to solid is taking place, for example, if I plot a cooling curve, all
of you are aware of this cooling curve (T vs t). For example, if I plot T verses t for an
alloy, you see that, it shows something like this. Am I right? This is for a typical solid
solution formation from a liquid. But, these slope changes that you are seeing are only
because of ΔH differences. I mean enthalpy differences that the moment this liquid starts
forming a solid, suddenly the slope becomes shallower. Why, because heat is evolved.
When a solid forms out of the liquid, the heat is evolved.
So, there is an excess heat for it to be removed. As a result, cooling rate becomes lower.
And once, the cooling rate becomes lower, the cooling curve becomes shallower. And,
that is why you see; and if the ΔH is very-very small and you do not see a difference
here. So, a typical cooling curve we, that is why a normal Thermal analysis, we do not
use to simply find out the solid state transformation. That is why we use what is called
differential Thermal analysis because we are going to, looking at difference between a
reference and a sample, where the reference does not undergo any transition at the
temperature, whereas the sample is undergoing. So, because the differences we are
observing, then that is why you start seeing the peaks. So, if you do not use a DTA or a
DSC,
266
simply using a normal thermal analysis, putting a thermocouple inside the sample and
seeing how this thermocouple is responding as a function of time.
You will not be able to really see solid transformations because the ΔH is very small. So,
similarly if you want to see a liquid to glass transition, the only way to see is that there is
a small step, that you see; very-very small step. Unless, your S C is very sensitive, you
will not be able to see that small step. There is a small step and that step indicates that
there is and if you can measure the C p as a function of temperature, then you will
actually see there will be a discontinuity in the Cp; suddenly, the Cp will jump from a
glass to a supercool liquid. And, that jump will give you an indication that, yes, there is a
transition. So, a second order transition always will have a C p change. For example,
order disorder transitions or most of the order disorder transitions are second order
transition. In fact, people, some of them, call them as λ transitions.
Because there Cp, if you look at it, plot Cp versus temperature, you will see something
like this (see above curve). So, this is a disordered state (DO), this is an ordered state
(O). From an order state, when you are going to a disorder state, you suddenly see a C p
change. And, that is what usually people call them as lambda transitions. So, there are
number of order disorder transitions that are second order transitions; where you will see
a Cp change. Similarly, if you look at as a function of pressure, the second differential of
free energy with
267
pressure, will be what? What is that parameter, if you ever heard of it?
∂G
∂P T |
=V
2
∂ G
So, 2 =dV /dP
∂P
So, dV/dp is what? Rate of change of volume with pressure; what do you call that? It is
called compressibility. How, a phase changes its volume as a function of pressure.
So, the compressibility; this is the parameter which is very important for liquids.
Particularly when people talk about liquids, liquid compressibility is one of the important
parameters. Solids have very small compressibility; they do not change their volume
much with pressure. But, liquids change and gases, more importantly they change their
volume to a large extent with pressure.
So, the dV/dp is very large for the gases and smaller quantity for liquids and for solids is
much smaller. So, wherever there is a second order transition, if you really look at the
compressibility values, you will see a discontinuity in the compressibility. So, that is
how we can talk about second order transition. So, glass transition is one such second
order transition.
Student: Sir but if there is any change in the structure, compressibility won’t change
Correct. That is why, most of the time change in the structures are simply first order
transitions. So, that is why you do not see. In a first order transition, you do not see any
change. Cp also, you will not see because ΔH itself, there is a major change you will see.
268
Student: In Order disorder?
In Order,
269
disorder also. If you really, if it is a second order transition and if you measure the
compressibility from the ordered structure to disordered structure, you will see a small
change. But, may not be like a liquid to glass transition, but you would see that there is a
change; it depends on how sensitive your instrument is, but you would definitely see this
theoretically.
We are only saying that there is a discontinuity; we are not talking about the magnitude
of the discontinuity. So, that has to be remembered. So, the magnitude depends on the
particular system. Similarly, ΔH also; when we say there is a discontinuity in the H,
when we are talking about first order transition, the discontinuity can be very-very small.
For example, solid to solid transformation, still it is first order, the FCC to BCC
transformation of a pure Fe is also a first order transition. But, ΔH is very small. So, it
depends on how sensitive your instrument is. So, we are not talking about the quantity,
but we are only talking about the possibility of discontinuity.
So, this is how order of a transformation can be talked about, first order. Though, it may
not have any practical significance, but theoretically one should understand how the
transformation is taking place. Particularly, physicists are very concerned about what is
the order of transition. Particularly, whenever they talk about order disorder transitions,
many of the order, disorder transitions are all second order transitions. So, one has to be
remember.
Student: Why glasses are brittle?
Why glasses are brittle? Basically, it is because they are very hard, there modulus is very
high. So, when I
Student: Diamond is also hard
See the question is not that. You know, the moment you talk about, even diamond also is
brittle. I do not think diamond is very ductile because first of all the bonding itself is
covalent bonding there.
So, once a crack nucleates, when you say a material is brittle; that means the crack
propagation is very-very easy. Is not it? So, similarly any brittle material from the Griffith's
criterion, if you look at it, as long as the crack propagation becomes very easy. In a glass,
whenever a crack is nucleated, there is nothing to obstruct the crack. In a normal
polycrystalline material, grain boundary is an obstacle. There can be precipitates which can
be obstacle; there can be lot of obstacles. Whereas in a glass, no. That is the reason why, in
fact, people are now trying to add a small amount of nanoparticles into glass by, actually in
270
fact, this is very interesting experiment people have done. That, you think of a windshield of
a car, the moment there is a crash, it just shatters into pieces. Is not it?
271
So, people are now trying to put nanoparticles into the windshield. So that whenever
there is a crash, the windshield actually bends and goes back. So, you are making a glass
into a tough material. In fact, nowadays people talk about tough glasses. There are a
number of tough glasses. This tough glass concept is all coming basically because of
presence of the nanoparticles; that means, if you take a glass, crystallize it and do not
crystallize it at higher temperatures. Crystallize it at very close to the crystallization
temperature, control the temperature and time of crystallization in such a way that, very
fine nanoparticles are dispersed inside the liquid. Then, whenever a crack is propagating
this nanoparticle can be an obstacle for it, the, what is called shear band formation that
occurs. Every shear band gets stop by one of these nanoparticles, that is one of the
reason why many of these bulk metallic glasses, per say bulk metallic glasses may not be
of great use for structural applications because they are brittle, but if you can nano
crystallize them, then I can make a rod of bulk metallic glass like this very easily and
then crystallize it, and get nano crystals in it; so that, it is like a composite now. It is a
glass nano crystal composite and such a nano composite can be of great use.
Student: What are the processing techniques?
So, when I say, see that is, I think, we have been talking about it for quiet sometime.
When, you look at how glasses were first made, glass was first made by what we call it
as a Gun quenching technique. You remember. Where we have taken a liquid, cooled it
rapidly to get, liquid to crystal transformation we suppress to get a glass. Why we had to
really suppress the liquid to crystal transformation is because the crystal formation is
very easy in most of the metals and alloys; because most of the metals and alloys have
simple structure such as FCC, BCC. And, as a result, liquid to crystallize into those
structures is very easy.
So, to suppress that, you need really very high cooling rates. So that atoms are not able
to move to form that particular nucleus for it to grow. As a result, you need very-very
high cooling rates. Whereas, in a silicate glass, you do not need high cooling rates, why
because crystal formation is more difficult there. We know that the crystals are more
complicated structures. So, from that concept, what people have now gone is that, if I
can somehow make crystallization more difficult, then I can form glass very easily. So,
what they have done in these bulk metallic glasses is they have added a large number of
elements. Instead of just a binary alloy, they made a five component or a six component
alloy. Once, you have these many elements inside this, inside the liquid, for a unit cell to
272
form and for a nucleus of an r* size to form and grow, you see that all these six elements
have to sit inside the unit cell. This is what we called it as a confusion principle.
So, because of that, the liquid for it to nucleate a solid becomes more difficult. So, as a
result, you can easily undercool the liquid. Earlier, undercooling the liquid is more
difficult because crystal tries to form at the melting point itself. So, here because crystal
is not forming, you can keep on cooling the liquid without the crystal forming, so that
you can bring that liquid to the glass transition temperature; so that, this liquid becomes
a glass. So, such glasses are called bulk metallic glasses because they are able to be
cooled to a glassy state at very slow cooling rates.
For example, as slow as cooling rate as 10 -1 K/s. So, what people have simply done is,
they have taken liquids of such glasses, glassy compositions and poured into a mould of
almost three inch thickness.
And then, basically cut this cross section and then looked at it in the microscope and then
they saw that, throughout this cross section from the center to the surface, it is all glass.
It is almost similar to your hardenability parameter. When you talk about hardenability
in steels, we talk about a critical diameter. What is this critical diameter? The minimum
diameter in which, the center and the surface, both of them you will have martensite.
Why because the center cooling rate is different from that of the surface cooling rate. As
a result, you
273
may have a situation where the surface may give you a martensite, but the center is not
giving you the martensite. So, that particular diameter, where you would get martensite
even at the center, is what we actually call it as a critical diameter.
Similarly, when I pour a liquid metal into a mould like this, the surface cooling rate and
the center cooling rate is different. So, if I have the center also having a cooling rate,
which is higher than the critical cooling rate. How do I talk about critical cooling rate in
this particular case? For understanding that, I can draw a TTT diagram for a liquid to
solid. The way you draw a TTT diagram for the austenite to pearlite transformation, one
can always draw a TTT diagram for a liquid to solid. And here, what was the
temperature that we put in case of austenite to pearlite?
Student:A1
That, what is called your A1 temperature. Here, we will put melting point (MP in TTT
diagram above) there. That is the melting point in this case. So, you will have a TTT
curve like this, where I can now talk about what is called critical cooling rate (line 1,
curve C1).
The way I talk about critical cooling rate in austenite to martensite, exactly the same way
I can talk about critical cooling rate and, this critical cooling rate for a normal metal or a
normal binary alloy such as a steel let us say, is so high that I need to use a cooling rate
as high as something like 106-108 K/s. And, if it is a pure metal, I may even have to use
to 108 K/s to be able to avoid the crystallization. And, if I use such cooling rates, I can
get a glass. In principle, for any material, as long as I have a TTT diagram generated like
this, for a liquid to solid transformation, I can always find out, what is that cooling rate
that is sufficient for me to suppress the crystal to formation; so that, I can get that liquid
remained as liquid, until you reach a temperature here, which we call it as M s there in
case of martensite transformations and we call here as Tg.
You see the concept is exactly more or less similar. So, the moment that liquid is brought
below the Tg the liquid becomes a glass. So, this particular critical cooling rate for a pure
metal is very-very high. Similarly, you see, even in martensite also, we say, when you
add an alloying element, the curve shifts. Why does it shift? Because diffusion is made
more difficult; the pearlite formation is made more difficult. Whenever, you add a
chromium or any other alloying elements, you see that TTT diagram shifts. The reason it
shifts basically, is because the diffusion is more difficult; that means the time required
274
for the pearlite to nucleate is longer. And, if the time required for the pearlite to nucleate
is
275
longer it takes, similarly, if I add more and more alloying elements into liquid, this curve
shifts like that (Curve C2).
To an extent that this cooling rate is (Line 2, Curve C2), now of the order of 10 -1 K/s in
most of the bulk metallic glasses. And, in silicate glasses also is almost like that. So, just
0.1 degree per second is good enough for you to be able to get a glass. And, that means
that you can get a glass, even simply just by pouring the liquid metal into a mould, into a
metallic mould, so that the center also you can get a glass. And now, people call a metal
a bulk metallic glass are those, where the glass forming ability is such that, you get the
critical diameter of 1 mm. If you have a critical diameter of greater than 1 mm, we
usually call them as bulk metallic glasses.
In comparison, if you remember the earlier cases of the what is called gun quenching
technique or melt-spinning technique; the thicknesses were only of the order of 30 to 40
microns. The first metallic glass that people have made Pol Duwez what was the system?
Gold silicon, so, the gold silicon, the metallic glass had a thickness of the order of 30-40
microns. So, melt-spinning technique does not really give you thicker than this. The
moment it is thicker, then you loose cooling rate. The moment it is thicker, it will not
have a high cooling rate. So, that is why, only by reducing the thickness. That is the
reason why most of the old systems, where this concept of bulk metallic glass did not
exist, people were forced to use high cooling rates to get a good glass.
And, because they were forced to use high cooling rates, they were forced to have
thinner sections because thicker sections you cannot get high cooling rate. So, that is
why people had only just thin films sheets of the order of that kind of thickness. But,
such materials, such components are not really useful in all applications. I cannot use the
thin foil for all applications. But, if particularly, if I want to make a structural material,
so I would like to have thicker samples and as result, now bulk metallic glasses promise
you, thicker sample.
276
The glass for example, the one which has a highest glass forming ability is
Pd40Ni20Cu20P20. This has given almost 75 mm thickness; that is the best glass forming,
glass forming alloy so far.
And, there are so many others with different-different thicknesses of the order of 30 mm,
most of the zirconium base alloys have very high glass forming ability of 30 mm critical
diameter. So, like that, we have a number of systems. So, you can see the whole concept
of glass forming ability comes from the same TTT diagram, which we are all familiar in
case of steels. We will stop here.
Student: In bulk metallic glasses, the criteria for choosing elements is diffusivity?
Yes, to a large extent diffusivity.
That is the reason why we, when we take a metal, of course there are many concepts.
One is diffusivity, where you trying to add metals with a higher melting point. So, that
the diffusion is slower. You take, let us say an Al, add Fe to it, add Ti to it or add V to it,
which are all high melting metals add to Al, the tendency is that the diffusion becomes
slower.
The second concept, which is also important, is the ΔH. You should add those elements,
which have a large ΔH because when you stabilize the liquid, then only a glass gets
stabilized. Stabilizing a liquid means the ΔH of the liquid should be highly negative. We
will talk about it possibly in the next class.
277
Student: Are these high entropy materials?
We will talk about it. They need not be very high entropy. Structurally, yes.
278
There entropy is high, but configurationally; their entropy need not be very high. The
reason is, for example, if I take bulk metallic glass, for example, with this particular
composition, which I have told you 40, 20, 20, 20 obviously, can have a higher entropy.
But, there are a number of systems, where one of the elements has a very large quantity,
let us say 70%. Remaining all elements put together are about 30%, you do not expect a
very high entropy because configurational entropy, all of you know goes to a maximum
at the center of the phase diagram. So, high entropy alloys are those alloys, where you
take equiatomic compositions. Let us say a five component system is 20% of each of the
element; that really gives you a very-very high entropy.
Student: What is glass ceramic
Glass ceramics are those again like nano composites, where you try to crystallize a little
bit and get crystals inside the glass; so that, it is like a composite. So, that is again
basically, particularly for the kind of applications that I was talking about that where you
want to make a tough glasses.
279
Department of Metallurgical and Materials Engg
Lecture No. # 13
Eutectic Solidification, Coupled Growth, Heterogeneous Nucleation
In the last few classes, we concentrated on the T0 concept and then its utility for us.
In that connection, we spent sufficient time, quite a significant amount of time on glass
formation. And, all of you know glass formation, you have seen with respect to a eutectic
phase diagram. Usually, you see glass formation only when there are eutectics. Why?
Because eutectic is where the liquid is stable. The meaning of eutectic is the melting point
depression. Isn’t it? Many of you know, eutectic means low melting. So, with respect to the
two pure metal melting points, the eutectic always has a lower melting point. Lower melting
point basically means liquid is stable. Isn’t it? So, if the liquid is stable, then only you would
get a glass; if the solid is stable you would not get a glass. So, in that connection, we talked
about glass. Today, let us go a little more detail into the eutectic; not talk about glass
formation, but about eutectic itself. So, see how the eutectic solidification occurs.
280
For example, if you take a phase diagram something like this (Fig. 1 above), you take a
eutectic composition (Ce). And, the moment you come below the melting point, below
the eutectic temperature (Te), bring this liquid, hold it at this temperature (T). What
should happen?
Student: α + β.
Liquid  α + β. And, this should happen by a nucleation and growth isn’t, because the
liquid composition is different from the two solid compositions, whatever the two solid
compositions are. So, this is a liquid, this is α, this is β, this is liquid + β, this is liquid +
α. All these you people know. So, now, if I take the liquid and bring it to a temperature
T; below the Te, this liquid should give you α + β and this α + β microstructures, there
are variety of these microstructures. The most common microstructure is what is called a
lamellar microstructure. You also might have heard about what is called divorce eutectic.
If possible, today’s class itself we will talk about why various types of eutectics come.
That is again related to what is called surface energies, which is a part of the
thermodynamics; we will look at it. But, if you simply look at this α + β (Fig. 2), which
is nucleated inside a reservoir of liquid, so you have a liquid here. And this nucleus, a
couple, which has nucleated; α and β always nucleate together if it is a lamellar eutectic.
If it is not lamellar eutectic, they will not nucleate together that also we will see it later.
281
But, assuming that they nucleate together, how do they grow? If they want to grow, what
you see is that α is a A-rich phase, β is a B-rich phase. So, A-rich phase and a B-rich
phase, both of them are growing together into the liquid. How do this two grow? They
can grow only when the liquid supplies atoms to α and β. Isn’t it? Atoms are actually
coming from the liquid to α and β; they are joining α and β. And, as a result, the α and β
are growing together at the expense of the liquid, so that at the end of the reaction, if I do
what is called the cooling curve (Fig. 3), temperature verses time, at this particular time
(t0), this has nucleated. And, as I grow, by this time (t f), all the liquid should vanish and
everything should become α + β.
Of course, there are a number of such couples nucleate at various positions, a number of
nuclei and all of them together grow, such that they form what are called certain
boundaries. What do we call those boundaries as?
Student: Grain boundaries
They are not grain boundaries. In a eutectic, you do not call them as grain boundaries,
because a grain boundary is related to one single phase. There is no one single phase
here. So, what do you call when the eutectics one lamella nucleates here and somewhere
else, you have another lamella nucleating α + β, both of them come together and meet
each other; what do called that boundary?
Student: Interface
Of course, it is an interface; even this is also an interface. Isn’t it? So, there is a name,
because this (one lamella) is called a eutectic colony. And, what you get them are called
colony boundaries, eutectic colony boundaries. So, you have a number of such eutectic
colonies grow and you will have eutectic colony boundaries coming in, like grain
boundaries. They are analogous to grain boundaries and you would see in most of such
eutectic microstructures, boundaries of different α, β growing in different-different
orientations. Between each colony, the difference is the orientation of α-β couple. So, the
α-β couple orientation in one and other lamella, orientation is going to be different. So, it
is similar to grain boundaries, where you have grain orientations are different at the
boundaries.
So basically, these α β grow at the expenses of the liquid that means, what? For α-β to
grow together, you need liquid to supply more A atoms to α and more B atoms to β. This
is a very tricky problem. How is that the liquid, which is there in front of it is selectively
able to supply more A atoms to α and more B atoms to β, such that they grow together. It
282
is, it is one single liquid. One single liquid is able to supply more A atoms to α and more
B atoms to β. Unless
283
you do that If only certain atoms, let us say, A atoms are coming uniformly to both the
things, only α will grow β will not grow. Then, it will not be a coupled growth. So, if
you want a coupled growth to happen, this has to. So, that means, selective supply of
atoms to α and β should take place. And, how does this happen? It happens if you
understand the thermodynamics a little bit.
So let us try to see what happens when we draw a free energy composition diagram at a
temperature T. I have taken a liquid of eutectic composition; brought it to this particular
temperature T. At that temperature, this whole thing is happening and I am trying to look
at what is happening. So, if I draw the free energy composition diagram at that
temperature, how will it look like? What is stable at that temperature? Can you tell me?
Student: α + β
α + β, so, that means α and β curves, free energy curves should be below that of the
liquid. By now, all these should be clear to you.
If that is the case, I can easily draw (Fig. 4). So, I can draw a liquid and I can draw an α
and a β and draw a common tangent between them. And, these two compositions (C αβ
and Cβα, Fig. 4), where you have the common tangent of α and β should correspond to
these two compositions in our phase diagram (C αβ and Cβα, Fig. 4). So, because there has
to be a one to one correspondence between free energy composition diagram and the
phase diagram and, this is the G-x diagram, free energy composition diagram (Fig. 4).
Now, if I look at again our couple once again, α and β, you have a liquid here (Fig. 2).
So, this α is in equilibrium with liquid; this β is also in equilibrium with liquid. Now, let
us look at, in this free energy composition diagram, what represents the equilibrium
between α and liquid and β and liquid? Whenever I say equilibrium, it is represented by
what? It is a common tangent. So, when I say an equilibrium between α and liquid, I
should draw a common tangent between α and liquid. So, how do I draw a common
tangent between α and liquid? See something like this which gives two compositions C αl
and Clα, (Fig. 4).
284
Now, similarly, let us do the same thing for the β. So, if I do this, some two compositions
again you get Cβl and Clβ What am I getting out of all these? Just simply certain six
compositions we have got. But, if I carefully observe these six compositions; forgetting
above the first two compositions which are between the α and β, look at other four
compositions. If I compare the composition of liquid in equilibrium with α and the
composition of liquid in equilibrium with β, I get some information.
What is the composition of liquid in equilibrium with α? It is C lα. This is the composition
of liquid in equilibrium with α; this (C lβ) is the composition of liquid in equilibrium with
β. What does that tell you? That tells you that, the liquid which is in equilibrium with α
has a higher B content when compared to the liquid which is in equilibrium with β. This
is the liquid, which is in equilibrium with α; that means, the liquid, which is just in front
of the α and the liquid, which is just in front of the β if I look at the compositions of
those two liquids, they are not the same. And, not only that they are not the same, the
common tangent generation itself makes it such a way that the composition of liquid in
equilibrium with α has more B content; C lα is higher than Clβ in terms of % B; that
means, what? Just in front of α, you have more B atoms; and in front of β you have more
A atoms. And, this means that liquid is not having the same composition everywhere.
If I go for away from this α + β interface, what is the composition of the liquid? It is
same as that of the eutectic composition. It is only because of the local equilibrium that
has to be maintained for this α. If this composition is not maintained, then α and β cannot
exist in equilibrium with liquid. They will not lie; they will possibly dissolve back into
the liquid. If α and β have to exist with liquid, the only way that is possible at this
temperature T; at this temperature T, we are talking about. Remember, if I am talking of
at this temperature Te, situation would be different. At Te, it is the same, because it is a
common tangent between all the three. But, we are talking of a temperature below the
eutectic. I have brought the liquid below the eutectic and now, looking at what would
happen. If I
285
am looking at that, you would see because of this, here in front of α, you have more B
atoms than β. Then, what you see is that B atoms will start travelling in this direction and
A atoms will start travelling in this direction (indicated by arrows, Fig. 2).
And, the gradient is decided by what is called this (PQ, Fig. 2). What do we call it as?
Inter lamellar distance. Actually, this distance, can you tell me what is this distance?
Between the half of the α and half of the β; between the center of the α lamellae and the
center of the β lamellae. Will it be the inter lamellar distance? Or, what do you mean by
inter lamellar distance in a eutectic? You must have heard of this word inter lamellar
distance. When I say inter lamellar distance, what distance I am talking about? From
where to where?
Student: Between two lamellae
Between two lamellae of what?
Student: α
Correct.
So, when I say if I have an α here, β here, another α here. So this to this (PR, in fig.
above); this is the inter lamellar distance λ. So, basically, PQ = λ/2 we are talking about
(Refer Slide Time: 14:53). And, this movement of B atoms going towards the β and the
A atoms going towards the α will make sure that the α will get more A atoms and the β
286
will get more B atoms, so that both of them grow together. And, this is driven by what?
It is driven by the difference between the composition of Clα and Clβ ,
287
which is what we call it as ΔC = Clα - Clβ. The concentration gradient can be written as
(Clα - Clβ)/(λ/2).. And, if you say diffusion
J = - Ddc/dx, the dc/dx for the flow of atoms comes from expression above. Once the
nucleation has taken place, the growth is decided only by diffusion. Isn’t it? And, for the
diffusion, you need a gradient. If there is no gradient, if the gradient is 0, there is no
diffusion. So, this gradient is provided by ΔC/(λ/2). And, how do I see this in the phase
diagram? Can I see these two points in the phase diagram?
Now, Clα and Clβ – where are they in the phase diagram?
Student: Extrapolate
Extrapolate what?
Student: Liquidus
Yes. So, you have two liquidus here: one is called the α liquidus another is called the β
liquidus (Fig. 1, Refer Slide Time: 01:31). And, what does α liquidus tell you? α liquidus
tells you, at all temperatures, what is the composition of liquid in equilibrium with α.
And, the β liquidus tells you the composition of liquid in equilibrium with β at various
temperatures. So, I want to know at this temperature T, what is that?
So, if I want to know what is the composition of liquid in equilibrium with α and liquid
in equilibrium with β at that temperature T, the only way I can find out is by
extrapolating the two liquidus (See the phase diagram above). So, if I extrapolate the two
288
liquidus, I get Clα and Clβ. And, the difference between these two is actually the driving
force for the growth of eutectic. If
289
you want the eutectic to grow and that is the reason why, if you carefully observe, at the
eutectic temperature, this difference is 0. Clα = Clβ, because if I draw at the eutectic, it is
actually a common tangent. If you remember, I can draw it for you once again if you
want (See G-x diagram on left side of image above). So, this is liquid, this is α and this is
β; let us say liquid, α, β. So, if I am trying to look at what is the composition of liquid in
equilibrium with α, what is the composition of liquid in equilibrium with β, they are the
same at the eutectic.
Once they are the same, ΔC = 0. And, if Δ C is 0, the gradient is 0. If the gradient is 0,
the flux is 0. So, in principle, at the eutectic, you cannot have a growth; you need a small
under cooling, at least a small under cooling. It can be very small if there is a
heterogeneous nucleation available; otherwise, it can be a slightly large. But, you need a
small under cooling, so that you have a driving force. And, that is why you can again see
that if I extrapolate these liquidus to lower temperatures, the larger the ΔC. And, that
means the larger the gradient. And, as a result, in principle, you can see the growth rate
has to be higher. But, there is a catch here. What is the catch? The catch is the ‘D’ there.
So, we are forgetting that there is D there also. And, this D incidentally, See if you
carefully look at these two liquids curves, if I assume that these are linear, then I can
assume that the ΔC is linearly dependent on the temperature. Because this difference can
linearly increase with temperature if these two are linear.
290
If these two are straight lines, not curved line, if I assume them to be straight lines, then
you will see if I plot ΔC verses ΔT; ΔT is under cooling, I will see that this will be linear
(see graph above). And, at ΔT = 0, ΔC = 0. Now, if I find out the D as a function of
under cooling, as a function of under cooling if I look at the D, how does it change? How
does the diffusivity change with temperature? It decreases of course; in what fashion?
Student: Exponential
Exponential, be confident whatever you want to say; you are an IITian. Tell me? How
does diffusivity change with temperature? That is what we are questioning. All of you
have read from your B. Tech level, at various stages, how does diffusivity change with
temperature? There is a particular expression which we always read. What is that
expression called?
Student: D = D0.exp(-Q/RT)
Correct. What is that expression called?

Student: Arrhenius equation
It is an Arrhenius equation. So, this tells you that the D is exponentially related to the T. You
can see, the D = D0.exp(-Q/RT)
.So, you will see that the actual, if I plot D, on the same plot, if I plot D, you would see D
would decrease (see in the graph above). I have actually drawn parabolic; it is not actually
parabolic, it is exponential. So, it decreases very rapidly as you increase the ΔT or decrease
the temperature. Increase the ΔT basically means decreasing the temperature. So, our ΔT is
increasing in this direction (along positive X-axis). So, that means there is a particular
temperature, critical temperature beyond which you will see the diffusivity plays a major
role. So, that is where you will see, In fact, we are not really talking about kinetics here
otherwise, I can even talk about the whole growth rate, how does it change.
In fact, λ itself changes as a function of temperature. You cannot have the same λ as a
function of temperature. All these things when you study any kinetics course, you will
know more about it. So, you can see, until certain temperature, we can always say
growth rate is proportional to this under cooling. Because the under cooling is what
decides the ΔC and, the ΔC decides the dc/ dx. And, as a result, I can say at temperatures
closer to the eutectic temperature, the growth rate is always decided by the ΔT, under
cooling. So, the higher the under cooling, the better, that is the reason why any eutectic
291
mixture, If I pour the liquid metal into a metallic mode or to a sand mode, you would see
the growth rates are different.
292
In fact, you will also see not only growth rate, the λ is also different. You will get a finer
λ. Again basically because the nucleation rate is higher, the r* is smaller. So, all those
things become important and you will see this. So, you can see that this problem of how
liquid is able to supply atoms to α and β; we are able to simply understand it, just if we
understand G-x diagram. So, you can see, from thermodynamics very easily we can
explain how does this come.
Now, let us go to the second concept of what should be the nature of the eutectic. It does
not always give you a lamellar eutectic. Can you tell me what is the type of eutectic in a
steel, in a cast iron? Sorry, we should not say steels; you do not have eutectic in steels. In
fact, that is what differentiates steels and cast iron. Have you ever known, what is the
difference between a steel and a cast iron? When do you call an iron carbon alloy as
steel; when do you call it as a cast iron?
(Refer Slide Time: 24:49) Student: Carbon is less than 2%
Less than 2; why? What is so sacrosanct about this 2%?
Student: Because I am not entering into the eutectic regime.
Yes, the cast iron – that name suggests they are easily castable , easy castability basically
related to eutectic. So, cast irons are all those where you have a eutectic reaction. And, in
the phase diagram if you look at it, below 2 %, there is no eutectic reaction. Isn’t? So,
remember the iron-carbon diagram, iron-cementite diagram; you would see that it is only
beyond 2% you start seeing the eutectic reaction. And, that is the reason why we call
beyond 2%. It is not just a magic number; it is basically the phase diagram; the eutectic
reaction wherever it starts, we call them as the cast irons; below that we actually call
them as steels anyway. So coming back to that,
293
And, that also reminds me that we should also try to understand why an iron-carbon
diagram and an iron-cementite diagram are different. Today, we will try to understand
that also, before that, we will try to see, there are different types of eutectics we said. For
example, we are coming back to our first question – the iron, carbon cast irons; what is
the type of eutectic you have?
Student: Lamellar
Lamellar? You people have not seen microstructures. You have what is called a second
year physical metallurgy lab. Many of you B. Tech. should have.
What is the eutectic? If I take a white cast iron, White cast iron is a typical cast iron; the
other cast iron, where you get graphite is not actually decided by the iron carbon; it is a
graphite, which comes from the carbon; whereas, the iron cementite diagram, which we
look at it, the actual eutectic reaction is liquid giving the γ + Fe 3C. That is the eutectic.
And, where do you see that eutectic? You see that eutectic only in the white cast iron;
you do not see that eutectic in any other cast iron. Whether you take a grey cast iron or
nodular cast iron, you see the graphite there, which is not a γ + cementite eutectic. γ +
cementite eutectic, which is called the ledeburite is seen only in a white cast iron. So,
what is the morphology of this γ + ledeburite? What is the morphology of the ledeburite?
Is it lamellar? It is not lamellar. It is you have cementite networks you see;
294
Ledeburite – I am not asking you what is the cementite network you will actually see.
When you come to the cementite being precipitated from austenite, hypo-eutectoid,
hyper-eutectoid steels is what you are talking about. We are talking about cast iron and
the eutectic mixture in the cast iron not what is the microstructure at room temperature. I
am not asking you what is the microstructure at the room temperature, after the eutectic
reaction is over, if you see the microstructure, how should the microstructure be? You
have basically cementite particles in an austenite matrix. That is why if you see white
cast iron, you see chunks of cementite particles; go back and look at atlas of
microstructures ASM hand book; you will get all these.
So similarly, if you look at Al-Si eutectic – Al-Si again a popular eutectic; why is it
popular Al-Si? Where do you use? Where do you use the Al-Si?
Student: Automotive
Automotive, pistons particularly so, there again you would see Al-Si eutectic; you have
Si needles in an α matrix, α Al matrix. So, you do not see it as a alternate lamellae of α
and Si. Can you tell me which eutectic has actually a lamellar eutectic? Now, I have
given two examples of non-lamellar eutectics. Give me any one example of a lamellar
eutectic that you know.
(Refer Slide Time: 29:21) Student: Pearlite
It is not eutectic madam.
You are talking of eutectoid. I am asking eutectic. Which eutectic is lamellar? You have
never come across? All of you are metallurgist; at least many of you, majority of you.
Al-Cu is one; Pb-Sn is another. These are all lamellar eutectics. If you take Al Si and if
you take iron-carbon, these eutectics are non-lamellar and people call them as divorce
eutectics. Divorce in the sense, they are not together; the α and β are not together; they
are kind of separate. Why does that happen? To understand this, we need to understand
the interfacial energies.
295
Whenever I talk about α, β and liquid, I have three interfaces. Am I right? What are the
three interfaces?
(Refer Slide Time: 30:26) Student: Liquid to α, liquid to β and α β
Correct. Liquid to α, liquid to β and α-β
So, if you look at these three interfaces, α-β interface, liquid-α interface, liquid-β
interface, all of them are associated with certain interfacial energies. Isn’t it? There
296
is a α-β interfacial energy; there is a liquid-α interfacial energy; there is a liquid-β
interfacial energy. And, we have already seen that whenever you talk about a liquid to
solid transformation, the barrier for the transformation is always the interfacial energy.
Whenever I bring the liquid below the melting point, it wants to become a solid, because
there is a driving force in terms of the volume free energy. This liquid has a higher free
energy than the solid below the melting point as a result, solid wants to form. But, why
solid does not form is because of the interfacial energy.
So, as a result, whenever you have this kind of three interfaces, the system would choose
the one which has the lowest interfacial energy. If a particular system has α-β interfacial
energy as the lowest, the system would like to have that particular interfaces maximized,
so that the overall interfacial energy, overall interfacial energy is what?
Aγ, this is what is actually preventing. That is why if you remember ΔG, we always
write this.
ΔG = (4/3)πr3ΔGv + 4πr2γ
Do you remember this? Of course, there we have assumed as spherical nucleus; we are
not talking about eutectics there. But, if you talk about eutectics, the same thing holds
good excepting that these geometric factors will change. Basically, it is I can write
ΔG = VΔGv + Aγ
297
. Am I right? So, it is this Aγ to be minimized, so that even at a smaller under cooling,
the liquid can become a solid. The
298
liquid wants to become a solid at as small under cooling as possible. So, it will choose
whenever it has, for example, if it is having let us say pure metal solidifying, it has no
choices of interfacial energies. It is one single solid nucleating out of that. There also, it
will search for if there are any heterogeneous nucleating sites. The moment there are
heterogeneous nucleating sites, the liquid is very happy, because the γ is very small
there, because that is what is the meaning of a heterogeneous nucleating site.
A heterogeneous nucleation site is that site, where the interfacial energy between the
solid that is nucleating and the particle that is available is very small. And, that is the
reason why you see heterogeneous nucleation occurring. If you just add any type of
particle to a liquid, it does not give you heterogeneous nucleation. Simply take let us say
Al add some sand particles to it, it does not give you though you may say sir there are
particles available. That particle that you are adding should have a structural
compatibility with the solid that is nucleating such that, the interfacial energy between
the solid and the particle that is available should be very small, so that Aγ is minimized.
And, that is why if you remember, again, I will take you back to your heterogeneous
nucleation.
Think of a particle here and a solid nucleating and there is a liquid. So, you should know
that if when the solidification has not taken place, if the solid has not formed then what
you have is only particle liquid interface. Isn’t it? You have a particle, you have a liquid,
solid has
299
not formed on it. So, you have a γpl. The moment a solid has formed on it and this γ p l
has some surface area so, you have a A.γ pl, will decide what is the total interfacial energy
of that particular system. There is a particle inside a liquid. Now, the moment there is a
solid that nucleates on the particle, now you have different-different interfaces; you have
a particle solid interface; you have a solid liquid interface also. So, a particle liquid
interface, which was in this area (between the vertical bars), has been now replaced with
particle solid interface and solid-liquid interface. So, you will have Asl.γsl + Aps.γps.
These two are replacing Apl.γpl. If , Apl.γpl > Asl.γsl + Aps.γps, then only the heterogeneous
nucleation will take place, because you can see here this interface is replaced with two
interfaces. There is a particle liquid interface and particle solid interface, which is
actually a circular interface, because it is a plain surface on which you have a hill
forming; a bubble kind of solid forming let us say. So, this interface is circular whereas,
this is like a spherical cap. The solid-liquid interface is a spherical cap. Whatever
geometrically, you can calculate what is the area of that. You calculate the area; multiply
with the γ of that particular thing. Similarly, calculate this PS area, which is this. So,
actually, APS = APl before the solid has nucleated, these two surfaces are the same.
300
It is something like a flat surface. And, in that circular area, the solid has nucleated. And,
solid has formed like a like a balloon. And so, this area (circular) before the solid has
nucleated, it was actually a particle liquid interface. So, I can say that this is equal to this
APS = APl. So, considering these two total surface energies, this overall surface energy
has to be smaller than that. Then only, you would have heterogeneous nucleation. So,
whenever I say heterogeneous nucleation is taking place, when does that happen? It
happens when γps is very-very small. γSl does not depend on what particle I am adding.
γSl is between liquid and solid. That has nothing to do with which particle I have added.
But, it is γps, is decided by actually the structural compatibility between the solid and
particle.
Whenever you see the particle has the same structure, has a very similar structure that is
a reason why when I take an Al alloy, I add a TiB 2 particle; it acts as a good nucleating
site. Why? Because, the TiB2 has a structure, which is hexagonal and, whenever you
have a hexagonal structure, there is some similarity between hexagonal structure and Al.
What is the structure of Al?
Student: FCC
FCC, what is the similarity between these two?
Student: Basal plane
Yes. Some plane of FCC has a similar packing to that of another plane in HCP; the basal
plane of HCP and the 1 1 1 plane of FCC have exactly similarity. So, on the basal plane
of TiB2 particle, actually a Al can grow. And, that is how the actual nucleation takes
place. And, just because structures are same; I mean there is a compatibility that is not
sufficient. What you have to also look is what is the inter-atomic distance in the basal
plane and what is the inter-atomic distance in the 1 1 1 plane is also important.
Otherwise, there is strain during this nucleation. So, that is why actually when we talk
about different-different particles being added, we look at whether first of all there is a
structural compatibility.
For example, if I add a you know cubic particle into a Al; for example, TiC is a
wonderful nucleating site for Al, because it is cubic. If I add a cubic particle because Al
is cubic, it has a easy compatibility. But, if I keep on adding different-different cubic
particles, but with different lattice parameters and at some stage, you see that lattice
parameter of this cubic particle that I am adding is entirely different from that of
301
the Al lattice parameter. Though it is cubic, it may not nucleate. So, all that is reflected
in terms of one single parameter called γPS. So, that is why, we talk in terms of what is
called coherent interfaces, semi-coherent interfaces all that comes from here. When there
is a compatibility, lattice parameter compatibility. And, you can see that when this is
much smaller, so that this is satisfied, then only you have nucleation.
Similarly, here when you have these three interfaces, the interface which has the lowest,
In fact, when I want to talk in terms of this ΔG = VΔGv + Aγ
for example, in case of eutectic, I have to split this Aγ into three parts.
Aγ = Aαl.γαl + Aβl.γβl + Aαβ.γαβ
So, I have to split this into three parts. And, out of the three parts whichever is the
lowest, the system would like to choose that. And, that is the reason why wherever you
see a lamellar eutectic, it means actually α-β interfacial energy is the lowest. You can
actually calculate it or see a hand book, where the interfacial energies are available. And,
you would see that this is true.
And, if one of these two, γ αl and γβl is smaller, then what should happen? When one of
these two is smaller, for example, if I assume that γαl is the smallest, then you see
302
inside the liquid, the α will nucleate first. The moment liquid-α interfacial energy is
smaller than α + β, α + β will never look together as a couple. You will see α nucleating
first. And then, the remaining liquid slowly converts to the β, so that you will have a β
matrix and α particles; it can be α particles, it can be α needles, it can be α plates
depending on growth conditions. Certain structure of, depending on what is the structure
of α, for example, if you look at cementite, cementite is a particle in a ledeburite
whereas, Al-Si eutectic if I look at it, the Si are needles. Why? Because the diamond
cubic structure of Si is such that some directions can grow much faster than the other
directions.
Maybe if you have read about solidification, the growth directions are different, because
in all directions the growth rates are not the same. Have you ever heard about it? For
example, if I take a cubic structure, if I take Al nucleating inside, we always say it is
spherical nucleus. Actually, first and foremost it is never spherical. Never a nucleus is
spherical, excepting a situation where a liquid is nucleating out of a gas, because there is
no crystal structure there. The moment you are talking of solid nucleating out of a liquid
or a solid nucleating out of another solid, you can never ever have a spherical, because
solid nucleus is composed of unit cells.
Student: It would tend to become spherical
It would tend to minimize the surface energy, not tend to become spherical. It would
tend to minimize the surface energy by choosing such a shape, which has the lowest
surface energy. That which you think is sphere, but it is actually not that.
Coming back to that point, here you understand this, that it would be α that nucleates and
grows and then rest will be β. And, in other case where liquid β interfacial energy is
lower, you will see β nucleating and then growing. So, depending on what is the
situation, you will have either different morphologies of α, but you will always see that
there is no lamellar nature there. You will always see something growing inside the other
one. So, when the eutectic reaction is finished, when you look at the microstructure, you
will see one phase embedded in the other phase. The moment you see embedding of one
phase in the other phase, you can easily understand for example, Si inside Al or
cementite inside austenite. It clearly tells you that the cementite has nucleated first.
Similarly, there also, Si has nucleated first than the α Al. So, by looking at the
microstructure we can know which has nucleated first. And, that gives you an idea
303
whose interfacial energy. Not only interfacial energy, I have to also caution you here,
interfacial energy is one aspect. The second aspect that decides which will nucleate first
is a structure. It is also the structure that decides, because One of the two phases has a
very complicated structure let say; its nucleation is more difficult. So, the kinetics will
force that one of the phase nucleates first rather than the other phase. So, that is also
another point.
Now, coming back to this point of shape of the nucleus, lot of people have worked on it,
many -many. Have you heard of this gentleman by name this? Of course, whole
thermodynamics that we talk about; the free energy that we talk about is Gibbs. Gibbs
has worked on it for quite some time to understand what is the shape of the nucleus?
304
And, that is when he has realized that when I look at a particular a cube let say, in a
cube, there are different-different phases , different planes in a cube. You have the 1 0 0
type, 11 0 type, 1 1 1 type variety of phases. Each of them has a different surface energy.
Do you accept this? The 1 0 0 plane will not have the same surface energy as 1 1 1 plane.
In fact, that is the reason why when you take a polycrystalline metal and put an etchant
on it and then see it in a microscope, you will see different shades; some grains look
black, some grains look bright, some grains look some grayish. You must have seen this.
And, what is the reason that we say?
Student: Difference in reactivity with etchant
Yes, if you simply look at a sample with different grains let us say and, let’s this is grain
1, grain 2, grain 3. And, if a light falls on it and goes back and, you are seeing the light,
which is coming through the eye piece and, depending on what is, and, this happens only
when I have an etchant there. I put an etchant and, what is this etchant doing? It is
selectively dissolving different-different regions. Whichever regions are more reactive it
dissolves more, whichever regions are less reactive it does not. And, this reaction
depends on the surface energy. That is the reason why we start seeing the grain
boundaries first. Why? Because grain boundaries are high energy regions, so, they get
attacked first. So, you see a pit formed there at the grain boundary. Once a pit is formed,
the moment light enters there, it does not get reflected back towards the objective; it goes
305
out. So, the light that is coming through, it goes out like that in different direction. Once
it goes out, it does not reach the objective. And, if it does not reach the objective, you see
it as black. That is why grain boundaries look black for you.
Now, similarly, looking at these grains, these grains also dissolve in a different fashion.
Why? Because if I look at the planes which are parallel to this surface, the planes that are
parallel to this surface for grain 1, grain 2, grain 3 are not the same; the atomic planes I
am talking about. So, for example, in grain 1, maybe the plane that is parallel to the
surface may be 1 1 1 plane. In grain 2, it is impossible to have 1 1 1 plane parallel to the
surface why? It is a different orientation. If the same 1 1 1 plane is parallel to the surface,
then I do not call it as grain 2; it would both of them are same grain. So, as a result, you
will see some other grain, some other plane; maybe 1 0 0 is parallel and, maybe here 1 1
0 is parallel.
Now, if you look at these, each plane has a different surface energy. And, what does that
surface energy depend on? It depends on actually the number of broken bonds. If you
think of a plane on surface, there are atoms below it, but there are no atoms above it. Per
unit area, how many number of atoms are not there decides the surface energy of it. The
closest packed plane will always have a higher surface energy that way, because in a
given area, you have more number of atoms supposed to be there, above it, there are no
atoms; below it you have atoms. So, that is why, you see different-different planes have
different surface energy. And, as a result, you will see the etchant reacts.
Now, what Gibbs has done is he tried to look at these surface energies and tried to form a
shape of what is called the nucleus based on the surface energies. And, I think we will
look at that possibly in the next class.
306
Prof. Dr. B. S .Murty
Department of Metallurgical and Material Engineering
Module No.01
Lecture No. 14
Peritectic Solidification, Metastable phase Diagram
We have been so far seeing the free energy concepts utilized in solidification
particularly, during the liquid to solid transformation. We looked at various aspects such
as how phase diagrams evolve, particularly looking at liquid and solid. We also saw how
the isomorphous solidification takes place, how free energy composition diagrams
evolve, as a function of temperature in isomorphous.
And, we also introduced a concept called T0, which is again related to liquid and solid.
And then, we extended that T0 to Eutectic. And then saw what are the kind of
information that we get, particularly looking at glass formation and also extension of
solid solubility. And, we also saw in the last class particularly on, how the Eutectic
solidifies? Particularly, the coupled growth of the α + β, how do α + β together grow in a
liquid matrix? How does the liquid supply atoms to α and β?
307
Let us look at Peritectic today. Peritectic solidification, pertitectic solidification is a very

interesting solidification process, where, if you remember, what is the Peritectic?
Student: Liquid + solid gives solid
Liquid + solid one gives you second solid, that is what is a Peritectic.
So, l + α  β, this is what is the Peritectic reaction and, if you look at this Peritectic
reaction, if you imagine that there is a liquid, inside which there is a α. And, this liquid is
reacting with this α and β is coming out. Where should this β come out? Where will it
nucleate? When, β comes out of the reaction between liquid and α, where should this β
come? It should be at the interface. So, in principle, you will have the β forming at the
interface as a thin layer. This is what is β. The moment, you have a β form at the
interface between liquid and α, there is no more interface between liquid and α.
308
So, how does this Peritectic reaction continue? Because Peritectic reaction is
l + α  β and, the moment, the initial reaction starts at the Peritectic temperature; let us
remember our Peritectic phase diagram. This is a typical Peritectic. Many of you should
know this because it is a part of the Iron-Carbon diagram, which all of you are aware of.
So, let us use exactly the Iron-Carbon itself for understanding. In Iron-Carbon, what is
the reaction, Peritectic reaction Vijaylaxmi, in Iron-Carbon?
Student: l + δ  γ
L + δ  γ. .
So, where are the phases then? This is the liquid; where is δ? δ is on the left side here
and where is γ? γ is the outcome of the reaction. So, γ should be at the bottom. So, that
means this is liquid + δ; this is δ + γ and this is liquid + γ. Am I right? And, at
thePeritectic temperature, if I take the liquid, cool it to a temperature just below the
Peritectic, the moment the liquid starts solidifying, it gives out δ. It is this top portion of
the phase diagram. Initially, δ comes out of liquid and that is how you see δ inside the
liquid. And, as you continue solidification to lower and lower temperature, more and
more δ comes out. And, we can find out the volume fractions of this, not really the
volume fractions, the weight fractions because phase diagram does not give you volume
fractions. I think this, you must be aware of by now.
309
Phase diagram cannot talk about volume fraction because phase diagram, the x-axis is
basically a weight fraction of x or atomic fraction of x. So, volume does not come into
picture. Though, how do you get then the volume fraction of a phase? By using the
density, if I know the density and if I do the microscopy, what is that I will get? For
example, if I have a two phase mixture and I do a microscopy, what is that fraction that I
get?
Student: Volume fraction
Volume fraction, you get actually the area fraction. Because you are seeing a 2D section,
you are actually not seeing a 3D. So, any microscopy will only give you the area
fraction. And, again from the area fraction, you need to convert it to volume fraction.
Again, certain assumptions; there are ways to convert, there is a book called
“Quantitative Metallography”. If you are interested, please read that.
So, where they talk about, how to convert area fractions to volume fractions? Certain
assumptions, of course are involved. If you assume it to be spherical particles, then it is
easy to convert area fraction to volume fraction. Otherwise, it is very difficult. For
example, you think of a plate of a second phase in a matrix. And, if you cut the plate like
this, then what you see it is a needle. Is not it. You are not actually seeing a plate; you
see it actually as a needle. So, this needle area that you are seeing, if you want to convert
it into volume fraction of the particular second phase, it is not so easy. If it is spherical
particle, it is easy because whenever you cut it, you basically get circles. Depending on
where you cut it, you get different diameters of the circles. So, that is why, all
conversions of area fractions to volume fractions are much easier, if you have a spherical
particles.
But real life, you do not see spherical particles all the time. So, that is why, one need to
do some geometric approximations and people regularly do it. There are mathematical
ways of doing all this. So, you should remember that, our microscopy always gives you
area fraction and our phase diagram always gives you weight fraction. And so, the
moment you reach this temperature, you have basically liquid + δ of certain amount of
liquid, certain amount of δ, depending on that particular phase diagram, which can be
easily found out by what? How do you find out the weight fractions in a phase diagram?
Student: Lever rule
Lever rule.
310
So, you use the Lever rule and from the Lever rule, one can find out. Now, this liquid +
δ is undergoing a reaction and giving you γ
311
because I am at that temperature, just below the Peritectic. Now, what should happen in
a Peritectic reaction? Liquid + δ should completely vanish at the end of the Peritectic
reaction and only γ should be left out that is what is a Peritectic reaction. But, what we
are seeing here is (see the image below)
(Refer Slide Time 08:25)
L + δ  γ, the γ that comes out covers the whole of δ because that is where the reaction
is occurring. So, the moment the δ is completely covered, there is no possibility of this
further reaction between liquid and δ to take place, so that liquid slowly shrinks and δ
also slowly should shrink such that, the γ should encompass both the liquid and the δ.
And, for that to happen, you need the reaction to continue. How does the reaction can
continue, when there is no interface at all, available between δ and liquid?
This whole problem of Peritectic is what actually leads you to cases where, Peritectic
reactions are very-very slow. That is why sometimes you may, if you cool this steel
faster, you may even end up in some δ left out at the end of the solidification because the
reaction, Peritectic reaction has not actually gone to completion. There are certain phase
diagrams, which are full of Peritectics. Do you remember any such phase diagram? We
usually use a term called “Cascade of Peritectics”.
Student: Uranium
You have gone too far off, I do not know about Uranium. But, something which
metallurgists are commonly aware of, which is a very important system for
commercially Copper zinc, Copper Tin. There are a lot of Peritectics.
312
In fact, the system has only Peritectics; Copper Zinc if you look at it. It has nothing else
in the liquid to solid phase transformation. Of course, in the solid state you may have
something else, but in the liquid to solid, you have the various inter metallics that you
see; β phase, γ phase, ϵ phase. All these phases come out of the Peritectic reaction and
because of that, you will see many of these reaction are very slow. Now, try to
understand how does this reaction proceeds? To understand that again we let us use our
free energy concept. Yes, at this temperature (just below the peritecticif I draw the free
energy composition diagram; all of you try to draw it. Let us see how it should look like.
What you have? You have a δ here; you have a liquid + γ, γ and liquid + γ.
313
How many phases are involved in this?

Student: Four phases
Four phases, how many phases are there at that temperature (below the peritectic)? If I
want to draw a free energy composition diagram, I need to know how many phases are
there.
Student: Three phases
Three phases, what are the three phases?
Student: Liquid, δ and γ
Liquid, δ and γ, so, let us try to draw (see Fig. 2 in the image above). So, there is a δ,
there is a liquid, there is a γ. Is not it? So, in principle, you have a δ + γ equilibrium. You
have liquid + γ equilibrium. So, this is the δ phase, this is the γ phase, this is the liquid
phase. And, these compositions that I am getting are what? The composition of δ in
equilibrium with γ (Cδγ), the composition of γ in equilibrium with δ (Cγδ), the
composition of γ in equilibrium with liquid (Cγl) and the composition of liquid in
equilibrium with γ (Clγ). And all these four compositions can be seen on the phase
diagram (see Fig. 1 above). If you draw everything to scale, you would exactly get these
compositions reflected in the two diagrams. I have not drawn to scale.
That is why, whenever you draw a free energy composition diagram, it should
completely reflect the phase diagram. This has to be, at least at the end of this course, it
should be very clear to you. So, now if I get these things, how do you understand this?
How is it useful for us to understand the Peritectic Reaction?
314
What you see here is, if I now draw a small segment of this (Fig. 3 in above image), take
a finite element from this and try to look at it. What you have is you have a δ. If I draw a
something like distance, as a function of distance, you have δ phase, you have a γ phase,
you have a liquid phase (See Fig. 4 in above image). Am I right? So, liquid and δ are
separated by a γ. And now, if I look at what is the % of B; that means composition of
each of these phases, if I try to look at, how do they look like?
For example, there are two interfaces here. One is a δ-γ interface; one is a γ-/liquid
interface. At each of the interface, there are two phases, which are in equilibrium. At first
interface, δ and γ are in equilibrium. At second interface, γ and liquid are in equilibrium.
If I look at first, δ and γ equilibrium, I need to look at what is the composition of δ in
equilibrium with γ and what is the composition of γ in equilibrium with δ. If I look at
that, I get that compositions from here.
So, if I look at those compositions, composition of δ in equilibrium with γ (Cδγ) and

composition of γ in equilibrium with δ (Cγδ), this is an increasing % B (reflected in Fig. 1
above). So, that means, the composition of γ in equilibrium with δ is higher than the
composition of δ in equilibrium with γ, in terms of % B. So, in such a case, I can put
these two compositions somewhere here on Fig.4. One, I will call Cδγ, and other Cγδ.
Whatever scale that you choose, I am just schematically drawing them.
Now, if I look at the other interface, which is the γ liquid interface, I get two other
compositions. One is, Cγl, another is
315
Clγ . And, those two compositions and if you carefully look at it, C γl is higher in B
content than the Cγδ. So, as a result, this composition C γl is going to be higher than this
value (on Fig. 4 above). And then, the next one, Clγ , is further higher.. Now, if I do a
kind of a profile, I can see the composition profile is something like this.
I really do not know, whether this (Cδγ and Clγ) is continuous horizontal into this. Let us
not bother about it. In principle, it may not. Why? If you carefully observe, what is the
composition of liquid with which we have started? We have started with a Peritectic
composition. At the Peritectic temperature, what is the composition of the liquid? It is
decided by the Peritectic horizontal (See Fig. 2). And the Peritectic horizontal tells me
that, this is the composition of liquid (Clp), which is reacting with this δ (Cδp) and giving
you this γ (Cγp). So, that means, this is the composition of liquid (Clp); in principle, which
is actually reacting with δ. And, at this temperature, the composition of actual liquid is
this (Clγ). And, that is different from the liquid (Clp), which is actually interacting with δ .
So, in principle, this liquid composition throughout may be exactly that or might fall that
far from the equilibrium, far from this interface, the liquid composition may be this
composition (Clp). If you have a very slow cooling conditions that you will see all the
liquid will have the same composition as that of this (Clγ). This is what we called it as
equilibrium.
If not, you may see that there is, that is why we talk about coring. So, if there is no
equilibrium, then you will see the liquid composition at the interface will be this (C lγ),
but far away from the interface will be this (Clp). But, we are not bothered about that
particular part more. What we are bothered about is, here (Circled part in Fig. 4). We are
saying that if Peritectic reaction has to happen, what should happen? If Peritectic
reaction has to complete, then what should happen? This γ should grow like this. Should
grow in both the directions such that, it absorbs all the δ. It also, absorbs all the liquid.
That is when we can say Peritectic reaction is complete. And, how does this grow? We
are saying that, there is no interface between liquid and δ left out. If there is no interface
316
between liquid and δ for the γ to grow, the only way is by diffusion of atoms. How does
this happen?
To understand this, we look at each of the interface. For example, if I look at the liquid-γ
interface; at the liquid-γ interface, if I look at it, liquid has higher B content than the γ.
So, what happens is, in order to, for the Peritectic transformation to complete.
Because, why this has to happen because at that temperature T, for that alloy
composition liquid is not stable, Is not it? If I am at this temperature, at this temperature,
for that alloy composition, with which I have started with. What is stable is, only γ.
Thermodynamically, that I can clearly see here, my alloy composition is somewhere here
(see in in the G-x diagram). I am choosing the alloy composition this (C γp) in Fig. 2
above) and if I carefully observe (Fig. 1 above), this is the δ phase, this is the δ + γ
phase, this is the γ phase, this is γ + liquid and this is the liquid phase. And, my alloy
composition is falling within the γ. That is what it is. It is falling within the γ because it
is falling within the single phase γ domain; at that temperature what is stable is, only γ,
in principle.
So, whatever liquid that is there, whatever δ that is there, at that temperature, at a
particular point of time, they are only metastable; they are not stable. So, in principle,
that liquid wants to get transformed to γ. Similarly, δ also wants to get
317
transformed to γ. And, how does this happen? This happens by the process of this liquid,
it wants to transform to γ and γ cannot take more % of B. γ can take only this much of B
(Cγl). According to the equilibrium, the composition of γ in equilibrium with liquid is
only this much (Cγl) and liquid has higher B given by (Clγ).
So, what happens is, the liquid gives you γ and the excess B atoms are given out;
lγ+B
that means, what happens is, this liquid wants to transform to γ, but in the process it has
excess B atoms. It is something like, you have seen earlier, when a liquid transforms to a
solid, there is an excess enthalpy that is available in the liquid, this enthalpy is given out.
This is what; we call it as Recalescence. Enthalpy of fusion; exactly similar, here the
number of B atoms in the liquid are higher than what a γ can take.
So, these excess B atoms are rejected out of the liquid. And, where do they go? Because
there is a gradient inside the γ, the B atoms will flow towards the γ/δ interface. So, these
excess B atoms that are available will flow towards this direction because there is a
gradient that is available. And now, what happens is, once these B atoms reach this,
before that what has happened? A small volume fraction, some small volume of liquid,
small Δx of liquid has got converted into γ of the same composition as that of γ.
And, whatever excess B atoms that the liquid has, it has rejected them out. And, those B
atoms have started moving, towards the γ/Δ interface and this small region has got
converted to γ. And, what has happened in that process? The γ has got extended into the
liquid. Some liquid has got converted into the γ. And now, what happens is, once these B
atoms reach this interface, if you look at the δ, δ composition is this Cδγ. δ has a smaller
B content than the γ.
it needs more B atoms because this is the difference in the B atoms between δ and γ (Cδγ -
Cγ δ
So, if the δ takes B atoms, δ has certain structure. In our particular phase diagram, δ has
a BCC structure. This BCC structure, takes away Carbon atoms. In our case, B atoms
means Carbon atoms because we are talking of Iron-Carbon phase diagram. So, this δ
takes Carbon atoms and the Carbon content in the δ rises to that of the
318
γ at that particular temperature, according to the equilibrium. The moment, the Carbon
content of the δ reaches that of the γ; suddenly that particular δ gets converted into γ. It is
only a structural transformation now. Composition of the δ becomes equal to the
composition of the γ. And, once the compositions reach same, suddenly there will be a
BCC to FCC transformation. And, that means, a small amount of δ got converted into a
γ. And, as a result, you will see a small volume here of δ, gets converted.
And, what is the process? The δ takes B atoms and gets converted to γ.
δ+Bγ
For the δ to get converted to γ, it needs excess B atoms because it has low B atoms, low
Carbon content. So, on that δ, once Carbon atoms are provided to it, it can get converted
to a γ. And in the process, if I add these two reactions, what you see is, basically is that
L + δ  γ; because the B atoms, which are here on this side and the B atoms that are
here on this side will get cancelled. So, you see that basically, the same Eutectic reaction
is taking place in two steps. And, this is what we call it as Peritectic transformation. We
do not call it any more, as a reaction because there is no reaction here.
So, the Peritectic reaction that we traditionally call is only at the beginning of the
reaction. At the beginning, you have a reaction because there is a real interface between
the liquid and a solid phase. And, there is a reaction that takes place. And, once the
reaction takes place, there is the third phase, which is the product phase of the Peritectic
reaction comes and covers the solid that is already available.
And, once this covering takes place, then the further reaction can only occur by
diffusion. And, diffusion where? Diffusion, inside that product phase that has formed the
diffusion is taking place inside the product phase and which is a solid phase. Is not it?
And, we all know, diffusion through solids is more difficult. Diffusion inside the liquid,
Eutectic reaction is very easy. Why? Because all the diffusion of atoms that is
happening, α + β  L, all the atom movement is taking place inside the liquid. And then,
they both are moving faster.
That is why Eutectic reaction is definitely much faster than a Peritectic reaction. A
Peritectic reaction needs a diffusion of atoms through that solid layer. And, once you
understand this, you would see that it is very easy to understand the Peritectic
319
transformation. And, that is the reason why, sometimes you actually see, if I cool a
system very fast, you may not see a Peritectic reaction taking place. For example, there
are, what are called metastable. Is this very clear to you now, how does this happen? So,
you can see, at the liquid/γ interface, liquid, a small fraction of liquid transforms, to the γ
by releasing B atoms; because γ cannot take as many B atoms that are available in the
liquid. So, the excess B atoms are rejected out and small volume of liquid gets converted
to γ.
And, in the process, now the B atoms that are rejected out, travel through the γ and then
reach the γ/δ interface and get added to the Δ which is available and the Carbon content
of the δ rises or the B content of the δ rises, in such a way that some part of the δ gets
transformed to that. And, all this is driven because of the diffusion. And, again we all
know, J = -DdC/dx. Now, what is the dc/ dx here, what is the dx? dx is nothing but the
thickness of the γ. dx is the thickness of the γ and what is the dC? dC is this composition
difference; ΔC = Cγl - Cγδ
That is what actually drives. And, again if you look at this difference, this difference at
the Peritectic temperature is, what? It is 0. At the Peritectic temperature, this difference
is 0 because both of them are meeting here at the peritectic point. At the Peritectic
temperature, all the three are in equilibrium. So, there is a single common tangent for all
the three phases. So, the composition of γ in equilibrium with δ and the composition of γ
in equilibrium with liquid is always the same. And, if that is the same then dC is 0. If this
is 0, the gradient is 0. If the gradient is 0, the flux is 0. So, there is absolutely no
possibility of any Peritectic reaction to occur at the Peritectic temperature. You need to
come below the Peritectic temperature.
And again, the way we have talked in the case of Eutectic, the lower the temperature
that you come, below the Peritectic the more this widens. So, the higher under cooling
that I provide is useful. Keeping again at the back of our mind, diffusivity always plays a
role. So, as long as I remember that, so I know that the higher under cooling increases
this. And, when it increases, you will see definitely ΔC increases, if ΔC increases, flux
increases. But, there is one more issue here. The second issue is, as the transformation
progresses as a function of time, what is
320
happening? Now, as a function of time, if you carefully look at it, the γ is becoming
thicker.
The γ is growing in both the directions because a small amount of liquid has got
converted to γ and the small amount of δ got converted to γ. So, what is now the γ
thickness? γ thickness is thicker than what it was before. So, if the γ thickness is
increasing, in our diffusion flux equation, the denominator is increasing; dx is increasing.
If the dx is increasing, obviously the gradient decreases. And, in the gradient, if the
transformation is taking place at a fixed temperature, dC is always constant. The ΔC,
which we have written here, is only dependent on the temperature. Because at a given
temperature, the phase diagram tells you, what is this composition, what is this
composition. Once, I fix up my temperature of transformation; that means I have under
cooled the liquid to a certain extent and there the reaction is taking place.
Then, if that is taking place at that temperature, you would see that in principle, at that
temperature these two compositions are fixed. And, that means the numerator (in
diffusion flux equation) is fixed. But, the denominator is now increasing as a function of
time; that means, Peritectic reaction slows down as a function of time. So, you see that,
initially it starts at a faster rate, but as a function of time, you will slowly see it
decreases. That is the reason why, at very high under cooling if you give, there may not
be any Peritectic reaction, significantly occurring; because at that temperature D itself is
slow. D is slow and dx is larger. So, both of them will lead you. And, this is something
which we need to understand. Now, this leads us to what are called metastable phase
diagrams.
321
What is this Metastability that you get in? Because this reactions are slower, for
example, liquid to react with δ to give you γ is not easy because liquid has a different
composition, δ has a different composition. And, the resulting γ has still different
composition. And, the structures are different and also, this reaction needs essentially a
diffusion through the solid state. And, because of which, there can be situations where
the reaction may not occur at all. And, that can be easily understood, if we can look at
this phase diagram, a little more closely.
If I simply look at this phase diagram and extrapolate these two curves (1 and 2), this,
what are these two curves? What is this curve (1)? Can you tell me?
Student: Solidus
Solidus, sure and, what is this curve (2) then?
Student: Solidus
Solidus? Both are solidus?
What is the definition of solidus, Prerna? I am sorry to ask you at such a late stage of
your life. What is the definition of solidus?
Student: Boundary
Boundary
Student: Forming a phase, solid
Can you tell it in the form of a sentence, put your words together and then frame a
sentence and tell me what is solidus? Anybody else who wants?
Student: A line that separates one solid and solid plus liquid
322
And, if I ask you to think of some other definition for it?
Student: The composition which is in equilibrium with solid and liquid
Composition, you are saying equilibrium with liquidus. Then, what is the liquidus? It is
also the same. Yes again.
Student: Given composition lowest temperature at which any liquid exists.
The temperature at which the solidification is complete; the temperature below which
there is no liquid that exists is a Solidus; where the solidification is complete. And, what
about liquidus then?
Student: Solidification starts
Starts, where the solidification starts, the temperature below which you start seeing the
solid first, from a liquid. If that is the case now, what are these phase boundaries? Again
look at it. This is liquid, what is
323
this? This is δ and this is γ. What is this?
Student: Solvus line
It is a solvus line. It is not a solidus line.

Please understand.
And, what is the solidus here? Can somebody draw it? Let me see. Yes. You see yes and
then, why don’t you put down the phases? Then it becomes easier for you. What is there
on the right side of γ?
So, what about this portion of the Peritectic (right to the peritectic point)?
Student: That cannot be called solidus
That cannot be called as solidus. You are right. This is what differentiates between a
Eutectic and a Peritectic. Let us draw again a Eutectic (see the eutectic below Refer Slide
Time 39:41) for the sake of argument. Now, on this, Bala can you come and draw what
is the solidus here? Not γ, α, Good. That’s it, perfect. In a Eutectic phase diagram, the
whole Eutectic horizontal becomes a part of the solidus. Whereas in a Peritectic, only a
part of the Peritectic falls into the solidus and the remaining part does not. Because
below this line (right half of peritectic line) you still have the liquid left out. If some
liquid is left out, then obviously it is not the solidus line. And liquidus line is easier; the
top portion. Similarly, here also (in case of eutectic) the liquidus is easier. The only
confusion that comes always is the, with the solidus. Now let us come back to our
question, with which we wanted to start, again I may have to erase this a little. Let us do
that.
324
If I extrapolate these two curves (Left phase diagram) ? This is γ + liquid. So, this is the
liquidus. We call it as, what liquidus is this?
Student: γ liquidus
It is called γ liquidus. This liquidus is called the δ liquidus. By now, you should, because
we are talking more and more about equilibrium of liquid with one phase. So,
325
you should slowly start understand because possibly, earlier when you are talking of
phase diagrams, you never talked about γ liquidus and δ liquidus.
Similarly, when we talk about this liquidus (Right phase diagram); this is the α liquidus,
this is the β liquidus. Though both are liquidus, but the liquid on this side is in
equilibrium with α; liquid on that side is in equilibrium with β. So, these things have to
be engrained into you, as you go along. So, what we see is that, whenever this reaction l
+δγ
is very difficult, the system chooses any one of the two cases. Either, it will become like
this or if I draw; may be here itself I will do.
You will reach a situation something like this (refer to the short dashed lines in Fig. 1
above). Depending on which phase is more difficult to nucleate. From the liquid,
depending on the structure of that particular phase, whether γ or δ is more difficult to
nucleate, the system will choose that particular phase which is easy to nucleate. And,
when you cool the liquid rapidly, only that phase comes out rather than one phase
coming out first and that reacting with the remaining liquid, giving you the second
phase. You will see that the whole phase diagram gets converted into an isomorphous
type of a phase diagram.
You see that, for example, here if I unfortunately, I do not have color chalks, if I have I
would have been easy, you will see that this gets converted into a kind of a isomorphous
(short dashed lines form isomorphous type, Left phase diagram). You say this is liquid Δ
isomorphous. So, the liquidus of the δ liquidus, I will extrapolate. The solidus of the δ, I
extrapolate and when both of them meet at some place that gets converted into a liquid δ.
And, now if you cool this liquid rapidly, you will see simply, only δ coming out. And, if
I can calculate what is called the T0 curve for the liquid δ T0 curve. And, if I do my under
cooling such that, I am below the T0, I will get a partitionless solidification; where liquid
transforms to only δ. And, in principle, after the reaction is over, I will see only δ, there
is no γ at all. Provided, my cooling rate is fast enough that there is no time for such a
reaction to occur.
And, in principle, if γ formation is more easier than a δ formation; if the structure of that
is such that; I am simply giving some example. I do not know, what is γ, what is δ. Let
us not simply relate it to Iron-Carbon phase diagram. They can be any phases. If that is
326
the case, then what you will see is that, there will be an extrapolation of γ liquidus (long
dashed lines). Such that, when I am cooling this liquid, what I simply get is the γ
327
directly. That means I need to under cool the liquid, below this equilibrium liquidus, to a
lower liquidus, which is the metastable liquidus between the liquid γ phase equilibrium;
that means, this is the γ liquidus if I extrapolate, wherever this γ liquidus goes.
And, if I can bring the liquid below the γ liquidus, if my under cooling is such that, I
come to below the γ liquidus, then in principle, γ can nucleate out of the liquid. And, if I
under cool to such an extent that I am below the T 0 curve, this is now, what is the T 0?
This is the T0 for γ for me. The earlier T0 that I have drawn here, is the T0 for the δ.
If you want, I will draw it a little more carefully. Too many curves possibly, confusing
you. That is why; I should have some color chalks. This is what is the T 0γ.. And, if you
extrapolate this, what you get is T0δ. So, depending on where you are, you can get
basically any of those phases. And, if I can under cool below the T 0γ, then in principle,
yes, I can get a partitionless solidification of liquid giving you single phase γ.
So, you can see clearly. Now this is extrapolation of what is the γ liquidus for me, and,
as a result, this would be for me and being very careful these two lines (dashed yellow
line) have to meet at a point. Ultimately, what is the left axis? It is a pure metal for me.
That is why, if you remember in the question paper that you had; possibly, in the next
class we
328
will take up that question paper and try to solve that. So, you should see that these two
(yellow dashed lines) ultimately should reach the same point. And, this is the T 0 that we
are talking about.
And now, if I look at the other phase, this is the, and you can see another that is the T 0δ
(pink dashed lines). So, depending on my under cooling and also depending on which
phase is easy to nucleate, I may get a partitioning solidification; solidification with
partitioning giving you either δ or γ or solidification without partitioning giving you
either δ or γ. So, you can see that a normal phase diagram now, some of them gets
suppressed.
For example, one simple easy similar thing, I can even see this. There is a very famous
phase diagram called Magnesium-Silicon phase diagram. Anybody knows about it? This
is an important phase in Magnesium-Silicon phase diagram, which is very-very
important for, you know Aluminium alloys particularly.
What is that?
Student: Mg2Si
Mg2Si. Good. So, Mg2Si, if you carefully look at it, there is a Mg 2Si here And, you have
a Eutectic between these two; this is the typical Magnesium-Silicon phase diagram with
Mg2Si. And, now if I take this alloy Mg 2Si composition and if theMg2Si crystal structure
329
is such a complicated crystal structure that I am cooling this liquid very rapidly. And so,
the nucleation of this Mg2Si is
330
very difficult. Then, what should happen to that liquid? That liquid will not give you
Mg2Si.
And, as a result, what happens actually is that, this liquid gets under cooled to such an
extent that if I extrapolate this liquidus curve (see the dashed lines in above image), I get
another liquid there; another Eutectic there. And, this Eutectic that you generate is a
metastable Eutectic. If this Mg2Si phase does not form, if it cannot form because of any
reason, then you see that the liquid gets under cooled and becomes stable. And, to such
an extent this liquid can easily become a glass. Because that liquidus now, is closer to
the Tg let us say. Then, it can easily become a glass. So, there are many-many phase
diagrams where, particularly when there are inter metallic compounds, they do not form
and suddenly that liquid becomes a glass. So, this is what is metastable phase diagrams,
which is very important. Thank You.
331
Department of Metallurgical and Material Engineering
Lecture No. # 15
Errors in drawing phase diagrams, Fe-C vs Fe-Fe 3 C phase
So, let us try to this time, today try to address some of the questions that we have
recently come across in the quiz one. We will start with the last question; the last
question is basically to identify what are the errors in the phase diagram.
And I thought that should be easier, because you must have gone through such phase
diagrams before, it goes something like this, let me see the phase diagram once again for
you and then. This is what is the phase diagram that is given to you, and people have
been asked to identify what are the errors? Some of the errors are very -very obvious,
because most of the errors that you see here, can be easily addressed using one single
principle called the phase rule.
All of you know, the phase rule says P + F = C + 1 this is the condensed phase rule and
we deal with only condensed phase rule in the whole of solid state phase transformations
or solid to liquid phase transformations. As a result, you can see that if this is the phase
rule that we are talking about, this phase rule tells us that whenever you
332
have a transformation taking place at the pure metal end where the number of
components is, C = 1. Then, the phase transformation should deal with 0 degrees of
freedom, because if C = 1. And if there is a phase transformation where, one phase is
transforming to another phase, P = 2 and automatically F =0. So that is what that
particular phase rule tells you.
So, if C =1 and P = 2 obviously, F =0. So that means, the degrees of freedom for any
phase transformation for a pure metal, where 1 phase changes to another phase, whether
it is a solid state phase transformation, solid to liquid or liquid to gas whatever. And only
thing is when you are talking of liquid to gas cases, we usually use not the condensed
phase rule, we use the generalized phase rule, where P + F = C + 2 and bring in the effect
of pressure. Whereas, all these phase diagram that we talk about, pressure is always kept
constant, because pressure is kept constant, we use the condensed phase rule.
And once, we use the condensed phase rule, for any pure component it is like for a pure
metal, this has to hold good and if this holds good then, automatically you will see that
this is one of the major error there (E1 in above figure). You see that here, solid to liquid
transformation is taking place over a range of temperature, it cannot occur over a range
of temperature. Because solid to liquid phase transformation again there is a solid
involve, there is a liquid involve, so there are 2 phases that are involved.
And when we are talking of a pure metal there is number of components is one and if
number of components is one, obviously, F has to be 0; that means, solid to liquid
transformation for any pure metal should occur at a fixed temperature, it cannot occur
over a range of temperature. But it can occur over a range of temperature for an alloy,
that is what we see in a isomorphous, that is what we see even for an alloy of this
composition (shown by vertical line) let us say. Any alloy of that composition you will
see there is something called a liquidus, there is something called a solidus. And there is
a temperature range between the two and this temperature range between the two
basically means that there is 1 degree of freedom.
And that 1 degree of freedom comes, because C = 2, the moment C = 2 and P =2 then, F
has to be 1. So, if you put C = 2 and P = 2 then F =1 and that
333
is what you see here (E1). So, here (E1) you cannot, because we are talking of pure
component A on one side, pure component B on other side, here you need to see that this
actually has to join each other (see the pink dashed lines in the image below).
This is first, the second thing that you see is that the eutectic horizontal that you see here
(E2 in above figure), where a l  α + β, you see that it is an inclined eutectic whenever
you see inclined means, that eutectic for various compositions in this alloy is occurring
as a function of temperature. So that again is against phase rule, again phase rule tells
that whenever you have three phases involved in a reaction, eutectic is one such a
reaction where l  α + β. So, there are three phases are involved. So, whenever liquid
gives α + β the P =3 and the number of component C, C is how many, we are talking of a
binary alloy. So, C = 2 then, F = 0, F = 0 basically means that everything that is
associated with that reaction is invariant, is fixed. That means, temperature is fixed, the
composition of liquid that undergoes the eutectic reaction is fixed and the composition of
the two solids that are coming out of the liquid are also fixed.
Here you see that the composition of the two solids (C α and Cβ) is fixed but the composition
of the liquid which is undergoing the eutectic reaction is not fixed. You see that, the eutectic
reaction appears to be occurring over a range of composition (indicated by arrow), as if any
composition that you choose, that all of those liquids are going to give you α + β and that
again defines this phase rule where you see that the degrees of freedom are 0
334
when I say degrees of freedom are 0, only one single composition of the liquid can
undergo a eutectic reaction. You cannot have any liquid composition and think that it
will undergo eutectic reaction and that is the reason why I told you earlier also.
For example, if you have a eutectic reaction, if you ask me what would happen if I
choose a hypo eutectic composition. Whether this hypo eutectic composition, the liquid
the alloy composition that I am starting with is this composition (C1), which is different
from the eutectic composition. Then, how do you say Sir, this alloy also there is a
eutectic reaction taking place in this alloy, how do you say that liquid can have only one
single composition. In principle I can choose any alloy here, alloy here, alloy here, alloy
here (represented by vertical lines, all of them will undergo eutectic reaction, am I right?
From this point to this point every composition will undergo eutectic reaction. But
incidentally all the compositions that you are choosing the liquid that undergoes the
eutectic reaction has only one single composition that is this.
Even if you choose this composition called C1, by that time this liquid reaches the eutectic
temperature that liquid will have only this eutectic composition, it will not have the
composition with which you have started with. You are starting with your alloy composition
as that, the moment α comes out of this, the liquid composition changes to such an extent
that, by the time you reach the eutectic temperature the liquid that is left out, after some α
has precipitated out that liquid will have this eutectic composition and it
335
is that liquid which undergoes the eutectic reaction. So, which ever composition you
choose between that eutectic horizontal, in all the cases whether it is a hypo eutectic or a
hyper eutectic the liquid that undergoes eutectic reaction is always this eutectic liquid
only and that is the reason, the reason for that is that the eutectic is invariant.
So, that is the reason why this is the, if you think this is the second error in this, this is
the first error, the third error is this (indicated in figure above). And both of them are
related to only one single phase rule related to the eutectic, the phase rule for the
eutectic, when you apply the phase rule to eutectic it tells you that any binary system.
That is why we always talk about a binary eutectic in a binary system, what sorry a
ternary eutectic in a binary system when I say a ternary eutectic basically, it means one
liquid giving you two phases there are three phases that are involved. So, you can say
simply a eutectic in a binary system is invariant, but a eutectic in a ternary system need
not be invariant.
336
But there are special eutectics in a ternary system where one liquid gives you three
solids; one liquid giving you three solids there are eutectics like that available in ternary
phase diagram. Such a eutectic again would be invariant, because there the P will be
equal to how many?
Student: 3
4 not 3, there are 4 phases involved and the number of components C = 3, and P = 4
then, F = 0 again. That means that particular eutectic reaction will happen at a fixed
temperature and all the 4 phases will have fixed compositions they cannot have any
composition of their choice. The liquid of a particular composition containing, so much
of A, so much of B, so much of C will give you α + β + γ and each of them will have
certain fixed amounts of A, B, C inside them and that is what is a ternary eutectic in a
ternary phase diagram. So, this is something which one has to remember.
337
So, that is the third one, then there is a fourth issue which comes with respect to this, you
see a miscibility gap there (α1 + α2), the miscibility gap basically means that 1 phase is
splitting into 2 phases isn’t it. 1 phase splitting into 2 phases means a miscibility gap
should always be enclosed in a single phase domain only. You cannot talk of a
miscibility gap in a 2 phase domain, what you see here is a 2 phase domain α + β, in a 2
phase domain you cannot have a miscibility gap, because miscibility gap basically means
α1 + α2. You can have a miscibility gap here, possible inside the α, inside the β if you
want you can have a miscibility gap, but not inside a 2 phase field. Because that defies
the whole concept of miscibility gap, miscibility gap basically means one single phase
which is a single phase at high temperature.
338
When I bring it to lower temperature there it is splitting into 2 phases basically because
the ΔH is positive. And at low temperatures the ΔH dominates the TΔS terms and as a
result, the free energy curve will have a tendency like that (upper graph in above image).
And this if I take it to higher temperature, at higher temperature TΔS dominates and as a
result the free energy curve will not be like this. But the free energy curve will become
something like this and the moment you become, you get a free energy curve like that
then it becomes a single phase (lower graph in above image). So that is why at higher
temperature in any 2 phase field, in any miscibility gap at high temperature you get a
single phase.
And this single case phase at low temperature, because TΔS term being smaller because
T itself is smaller. So, ΔH dominates and this you will get only when ΔH is positive. ΔH
for the solid should be positive to be able to get a miscibility gap and it should be
sufficiently positive to be able to overcome the TΔS. That is why in some cases, you
may have what is called a miscibility gap at sub 0 temperatures? At up to room
temperature you may see everything is solid solution, but if you take it to further lower
temperatures where, actually the TΔS is much smaller, because T is very smaller. And
the ΔH is positive though positive is not highly positive, the amount of the actual
absolute value of ΔH is going to decide how big is the miscibility gap. This temperature
that you are going to see, where the miscibility gap closes.
339
The miscibility gap can close at higher or it lower temperature, below the room
temperature depends on the particular system and it can happen when the ΔH is very
small. And if it is very close to 0, you may not have any miscibility gap at all. So, that is
what you see that. So, the moment you see a miscibility gap it has to be there inside a
single phase, because the meaning of closure of miscibility gap means, once I cross that I
have a single phase and in this phase diagram I do not have a single phase.
So, that is why that is the 4th error in this. The fifth error is a little more difficult for you
to see, and I do not know how many of you have come across. We always say a
boundary between any 2 phases should enter into a single into a 2 phase field rather than
entering into a single phase field what does that mean? Let us try to look at it from the
free energy concept. In fact, this is one of the ways to understand, why an iron carbon
diagram is different from iron cementite diagram? If I draw 2 phase region at a given
temperature (horizontal below eutectic), let us now, draw at this temperature the free
energy composition diagram (see image below).
340
If I draw the free energy composition diagram at that temperature, what you see is that,
you have an α, you have a β. So, if you look at this, this common tangent that I am going
to draw is going to decide that this is the composition of α in equilibrium with β (C αβ);
this is the composition of β in equilibrium with α (C βα). Now, if you look at this, if you
extrapolate this to a higher temperature, if this is the type of, we will erase some of these
so that it becomes more clear to you. So, we will draw it everything perfectly now,
excepting this particular problem.
341
So, we have now a phase diagram, where you have a horizontal, the eutectic reaction is
horizontal then the second thing is you have the eutectic composition is also a fixed
composition then, we have a situation that the melting of the pure A happens at a fixed
temperature. So, that is also taken care then, we do not have a miscibility gap here,
excepting that we have a situation like this that is the last error that we are talking about.
And this if I extrapolate this and what is this line (drawn in pink color)? What is that
line, can you tell me?
Student: Solidus of α
Solidus of α, very good, what does that give you? What is this? What does that tell you I
will draw with a color. What does this tell us?
Student: common tangent between liquid and solid
Tells you, common tangent between liquid and solid and anything beyond that?
Student: We don’t have liquid below that temperature.
Fine
So that is the definition of solidus, solidus anything else?

Student: Composition of α in equilibrium with liquid
Composition of α in equilibrium with liquid at all temperatures at various temperatures,
it gives you the composition of α whatever you have said is right I do not say wrong. The
other attribute of that line is that it gives you the composition of α at any given
temperature in equilibrium with liquid. So, now if I look at a particular temperature draw
a temperature, something like this (dashed yellow line), we are looking at this
temperature, at that temperature if I am looking at I draw a free energy composition
diagram. In principle if I draw it, at that temperature what is stable? α up to a
composition, α + liquid and if you go further you will get a β.
342
So, in such a case now, let me draw that this we know. So, if I draw this is an α phase,
you have a liquid and you have a β phase and α + β is not stable at that temperature. So,
as a result the liquid curve should be below the α and β that has to be understood. So, if I
draw β here. So, this is the α curve, this is the liquid curve, this is the β curve (indicated
above). The way I draw will be in such a way that, the stable states will be α and liquid,
and liquid and β not the α + β, if this liquid curve is above than these two curves that
means, that I am at a temperature where the liquid is actually not stable both α + β are
stable. Am I right? So, now if this is the case now, if I draw a common tangent between
α and liquid and another common tangent between liquid and β now, I get these four
compositions. Each of these compositions we have already talked about it earlier I will
still just for the sake of completion I will say this is composition of α in equilibrium with
liquid (Cαl), this is composition of liquid in equilibrium with α (C lα), this is composition
of liquid in equilibrium with β (Clβ), this is composition of β in equilibrium with liquid
(Cβl). So, if that is the case now, at the same temperature if I am extrapolating this curve
(drawn in yellow color in image below), what does this curve tell me, now tell me, we
have said already what does this curve solidus of α tell us.
Now, if I take this curve (drawn in yellow color), what does this give us? With, can you
repeat what you said?
Student: composition of α in equilibrium with α + β
343
In equilibrium with β, it cannot be in equilibrium with α + β what is the meaning of that,
composition of α in equilibrium with β, am I right. So, this line always that is what we
call it as Solvus line, Solvus line always gives you this, on the left
344
Solvus line gives you composition of α in equilibrium with β, the right Solvus line will
give you composition with β in equilibrium with α. So, now if I am extrapolating this
solvus line to a higher temperature such that I come to this temperature of my interest, at
that temperature this point is composition of α in equilibrium with β (Cαβ). Now, let us
see at that free energy composition diagram, what is the composition of α in equilibrium
with β? If I want to know that, what should I do, I draw a common tangent between α
and β, draw it let us draw that.
So, let us draw a common tangent between α and β. And see where these two touch each
other, when I see that, you will see that this the common tangent which is drawn from β
to α will touch the α at some composition like this. And that I would call it as C αβ and the
geometry of these curves is such that, this common tangent point where which is C αβ will
always be higher in B content than Cαl, why? Because β is at a higher level, β curve is at
higher level than the liquid curve, liquid is at a lower level than the β, why? Because at
that temperature liquid is more stable, because at that temperature liquid being more
stable, liquid curve has to be lower than the β curve. And because one curve is lower and
another curve is higher now, when I am drawing common tangents, it is not possible for
the common tangent between β and α to have a composition to the left of the α and
liquid, it has to be always to the right of the α- liquid. So that means what C αβ > Cαl in
terms of %B.
345
It has to be that way, looking at the free energy composition diagram and this you will
know only if you look at the free energy composition diagram. Now, so that means, that
what do you call this composition Cαβ? This means, the composition of α in equilibrium
with a Metastable phase, because β is a Metastable phase at that temperature. The
composition of α in equilibrium with a Metastable phase is always higher than the
composition of α in equilibrium with a equilibrium phase. So, this is true for any type of
phase diagrams, the composition of any one single phase in equilibrium with a stable
phase is always lower in terms of the solute content when compared to the composition
in equilibrium with a Metastable state, this we are going to immediately see with an
example of iron carbon diagram. So, now if that is the case this composition Cαβ, what
does this tell you?
If I extrapolate this solvus line (in yellow color) composition, it is telling that C αβ is for
me, this is Cαl at that temperature. At that temperature this is the C αl, this is the Cαβ. So,
this tells you that Cαβ < Cαl in terms of solute which is wrong. Whereas, if I draw the
solvus line like this (in white color) and if I extrapolate the solvus line, I can clearly see
that this particular composition now, which is nothing but the extrapolation of the α
solvus; that means this is the composition of α in equilibrium with β (falling in the two
phase region, to the right of α solidus). So, now Cαβ > Cαl and that satisfies the free
energy composition diagram for that particular temperature.
346
As a result this kind of curve is not possible. So, you cannot draw a Solvus line like this
(as shown in yellow color) you have to draw a Solvus line like this (as drawn in white
color). That is why in other easy fashion, what people say is that, the extrapolation of a
Solvus line to higher temperature, should always end up into a 2 phase field rather than
ending up into a single phase field, whenever this extrapolation goes into a single phase
field, you get into a problem. So, the extrapolation should always lead into a 2 phase
field, like liquid + α field. So that is what, you will see from this and for that you need
to understand this free energy composition diagram. And now, if you look at this I use
the same concept to iron carbon diagram, we are all aware of iron cementite phase
diagram.
Let us say, let us draw the iron cementite phase diagram. How many of you remember,
can somebody come and draw iron cementite phase diagram? Suppose, to be the bread
and butter of metallurgist, draw it.
My dear friend, you are right in the diagram, but can you look at the scale and give me
some idea, what is this composition in a iron carbon diagram?
Student: 2 percent.
2 percent, what is this composition in iron carbon diagram?
347
Student: 0.35
0.5.
0.5 not 0.35, 0.5.
So, you have put 0.2 percent very close to 0.5 that is wrong, that is the only mistake that
I see in this phase diagram. And second mistake that I also see is what is this
composition?
Student: 0.8.
0.8 and your 0.5 is greater than 0.8. So, always you have to whenever though you are
drawing a schematic diagram, you have to have some idea of the scale, it need not be
exactly to scale, but you should have some feeling for it. That is fine, why do you; do not
shrink it.
You are not using engineering drawing, forget it I understand that you know, but you are
not able to draw it to scale, I will try to do that.
This is horizontal if it does not look like please forgive me. Another important thing that,
you should also remember is, when you drew this, two segments of peritectic were equal
in your diagram, they are not and left is 0.1, peritectic is 0.18, right is 0.5 so that also has
to be kept in mind. So, simply just do not draw anyway you want after that I am you
know
348
one has to be careful in drawings. So, then comeback to 0.8 again. So, fine to some
extent we have tried; we cannot really try to a great extent. So, this is what is the
diagram that we are talking about and we have here (see the phase diagram above), again
Fe3C phase we call it. So, this is liquid, this is liquid + γ, this is γ, this is liquid + Fe 3C
and this is γ + Fe3C and this is α + Fe3C, this is α and this is δ and this is liquid + δ, this
is δ + γ.
Now, if I look at how an iron carbon diagram should look like, why we should bother
about it? We should bother about it, because we know certain grey cast irons when I
solidify I get carbon, I get graphite instead of Fe3C is it not. So that means, I need to
understand, what is the equilibrium between this liquid which is giving you the graphite?
The graphite that is coming out should be in equilibrium with the liquid so that means,
some equilibrium between the graphite and liquid also should be understood and that
will it affect the phase diagram. So, for example, if I take a eutectic where l  γ + Fe3C
and look at the other eutectic where this is l  γ + C. Will these two eutectics have the
same temperature, will these two eutectics will have the same eutectic composition or
not, what does it depend on? That depends on exactly the free energy composition
diagrams.
If I look at some temperature where for example, if I am talking about at a

temperature,T1, let us look at this temperature for that case, if I am looking at that
temperature, where I have a γ + Fe 3C. And l  γ + Fe3C, I have brought the liquid to
this temperature I am looking at that or I can also at this temperature, where liquid +
Fe3C is there, any temperature you can consider, let us look at this temperature T1, let us
say, if I am looking at that temperature I will see what would happen. At T1 temperature
I have what is the equilibrium; what do I have I have γ + Fe 3C, but Fe3C is not a stable
phase. So, in principle what is the equilibrium phase mixture at that temperature, it
should be γ + graphite.
Actually it should be γ + graphite instead of that I have γ + Fe 3C. Now, let us look at the
equilibrium between γ, Fe3C and graphite and if I look at that what I see is that I have
three phases. So, I should be able to draw a free energy curve with the three phases.
349
If I draw that, one is γ, this is γ. Now, what about Fe 3C? Fe3C is an inter metallic
compound, line compound. Line compound with 6.3 percent carbon so, and the other end
I will put it as carbon, pure carbon let us say, that is my graphite, this is pure iron, this is
pure carbon 100 % carbon that means, graphite. Whereas, Fe3C is somewhere here, let us
call Fe3C somewhere here let us say (dashed vertical line), though it is not exactly to
scale let us look at that. If that is the case then, if I am looking at the free energy curve of
Fe3C, how should it be? It should be a narrow free energy curve; so that means, it should
be a free energy curve which should look like this at that composition of 6.67 (see the
figure above).
Now, in comparison to the Fe 3C where should be the free energy for graphite, it should
be lower, because graphite is more stable than cementite and where will it be? It will be
on the right extreme where carbon is, this right axis is the pure carbon. So, the free
energy of graphite will be simply nothing but one point somewhere on this (G C), like
whenever we are talking about free energy of pure metals it is nothing but this; what is
this point? This is nothing but free energy of pure iron in the γ form at that temperature
(GFe). Similarly, free energy of graphite at that temperature pure carbon is some point
here (GC), if I take that now, if I draw common tangents I will see that, there is a
common tangent between γ and Fe3C (AB), there is also a common tangent between γ
and graphite (AC).
350
And if I look at this I will see that these common tangents will not be the same. That
means, the composition of γ in equilibrium with Fe 3C is different from the composition
of γ in equilibrium with graphite and it is always the composition of γ in equilibrium
with graphite will have a lower solute content, lower carbon content when compared to
the composition of γ in equilibrium with Cementite, because Cementite is a metastable
phase.
So, the composition of γ in equilibrium with the stable graphite will have a lower carbon
content when compared to the γ in equilibrium with the metastable Fe3C. That means, at
that temperature T1 (Refer Slide Time: 30:18), if I want to find out, what is this
composition (point P in image below, Refer Slide Time: 39:10) ? What is this point at
that temperature? What is this point?
Student: γ in Fe3C
Yes, not in Fe3C in equilibrium with Fe3C, γ is not inside Fe3C, So, this line (see the
yellow colored line in below image, Refer Slide Time: 39:10) tells me the composition
of γ in equilibrium with Fe3C, am i right? So that means, this point P (in phase diagram
in image below) is point B on the free energy composition diagram (drawn in the image
above) at that temperature. Now, if I want to locate the composition of γ in equilibrium
with the graphite, it will be somewhere on the left side, somewhere here (point P’ in
image below, Refer Slide Time: 39:10).
351
And if I do it for various temperatures, you will see that at all temperatures it will be to
the left only. So that means, if I find out for various temperature these points and join
352
them as a line, you will see that this is what happens (drawn in pink color). Is it a solvus
line or a solidus line?
Student: Solvus
Solvus, this (drawn in pink color) is a γ solvus in equilibrium with graphite, this (drawn
in yellow color) is the γ solvus in equilibrium with cementite so, you have a different
solvus now. And similarly, if I do the same thing here (in liquid + Fe 3C region), for the
liquid and Fe3C and instead of liquid + Fe3C I want to consider liquid + graphite let us
say. If I consider liquid + graphite I exactly get the same thing excepting that this γ curve
is replaced with liquid curve, cementite will be there and graphite will be there. And
graphite will always be at any given temperature, graphite will always have a lower free
energy than a cementite at any given temperature, because graphite is more stable than
Fe3C.
So, this curves, these common tangents are not going to really get affected, the exact
positions will get affected, but the relative positions is never going to be affected. The
relative positions always says that the composition of any phase liquid or γ or α in
equilibrium with graphite will always have a lower carbon content when compare to in
equilibrium with the cementite, if that is the case then I will see that this will be the
liquidus in equilibrium with the graphite (pink color) and this liquidus in equilibrium
with cementite (yellow color). All the pink curves that you are seeing, are all either the
solvus lines or the liquidus line in equilibrium with graphite, all the yellow lines that you
see here are all in equilibrium with cementite.
Now, let us look at what about the reaction temperature, eutectic reaction. if I look at this
reaction l  γ + Fe3C and if I look at this reaction l  γ + C, which reaction should be
lower temperature, which reaction should be higher temperature? Which one?
Student Graphite
Graphite, because when I am thinking of the reverse reaction of this, cementite + γ
giving you liquid, because cementite is a Metastable phase, it wants to convert into a
stable state at a lower temperature.
So, the melting point of the cementite γ eutectic is lower than the melting point of the
graphite γ eutectic. So, graphite γ eutectic will have a higher eutectic temperature,
cementite γ eutectic will have a lower eutectic temperature, because it is a Metastable
eutectic and that
353
is the reason why you would see this (pink horizontal is higher than yellow). The pink
curves represent Fe-C equilibrium, while yellow curves represent Fe-Fe3C equilibrium
and their schematics are shown in the above image. And similarly, you would also see
that, the maximum composition of carbon soluble in γ and carbon soluble in graphite
both of them will be different. And the same thing holds good here also, you will see that
the eutectoid also will shift and you will see that eutectoid composition will also shift to
the left.
So, both the eutectoids will change and what does not change is these lines (boundary
between γ and l + γ), these lines will not change, can you tell me why and this whole
peritectic portion does not change, because there is no cementite involve there, there is
no graphite, there is no cementite there. Wherever cementite and graphite come into
picture that portion alone gets modified, in this it is only giving  γ + δ so, there is no
other reaction. So, the; that means, the eutectic reaction gets modified, the eutectoid
reaction gets modified but the peritectic reaction does not. So, this is how you can
understand that thermodynamics particularly free energy composition diagrams can help
us to understand this. So that, you know that whether the eutectic reaction in a iron
graphite occurs at a lower temperature or at a higher temperature and the eutectic
composition also.
What is the composition that I should choose, if I want to get 100% eutectic mixture in a
iron graphite system or in a iron cementite system. Obviously, that will be different and
unless, you really calculate this thermodynamically and get these compositions, you will
not be able to get a 100%eutectic mixture if you are interested in it. The first question
asked you is about nano, what happens if I take a nano particle and melt it? I think most
of you know that the melting point will not be the same as a normal metal.
354
And that can be easily seen from a G verses T diagram, if I draw it. If I draw a G verses
T diagram for a solid and a liquid, the moment I say instead of a bulk solid I am taking a
nano particle of a solid we though know that nano particle of a solid has a higher surface
energy. And the moment I add this surface energy to the bulk volume free energy, the
free energy overall free energy of this solid is higher with respect to the bulk solid. So
that means, I will have a free energy curve of the solid shifted to a higher value of the
free energy (upper yellow dashed line). So, the nano crystalline solid will have a higher
free energy, the moment I put a higher free energy, you would automatically see that the
intersection is now, liquid is not changing liquid, there is we are not talking of a nano liquid
here. So, liquid free energy curve will not change, the moment you change the solid free
energy you would see that it intersects at a lower temperature and that is the reason why you
would see a nano particle melts at a lower temperature. But there can be a different situation
which is what, is the second part of that question that you see. If this nano particle is
embedded in a matrix, which has a higher melting point, if the matrix is let us say imagine a
situation, I have a lead particle inside an iron matrix or an aluminium matrix. And now I am
heating this and this lead is a nano particle, nano particle of a lead embedded in a aluminium,
lot of people are doing research on this, what happens? What is called super heating? Once,
you start heating this, will this lead melt at 321 ºC
355
which is supposed to be its normal melting point or some other temperature. Will it melt
at lower temperature than its normal melting point or will it melt at a higher temperature
than its normal melting point.
That is where you see that, the moment I start heating it both this solid nanoparticle and
this solid matrix as a function of temperature there is an expansion. Every solid expands
as you heat it, there is something called coefficient of thermal expansion and this
coefficient of thermal expansion when you look at it, a material which has a higher
melting point will have a lower coefficient of thermal expansion. As a result it apply
certain pressure on the solid, on the nano particle, the moment a pressure is applied you
see that under high pressure, if you look at clausius clapeyron. Whenever you apply a
pressure, whenever the ΔV between solid to liquid, when a solid is going to liquid if
there is a volume expansion under that conditions whenever a pressure is applied you see
a solid is more stable, because solid occupies less volume than the liquid. So, you will
see the stability becomes actually this (lower yellow dashed line).
Once that happens you will see that the melting point goes up, always you see this
happening. Whenever you take an embedded nano particle, an embedded nano particle
melts at a higher temperature than its normal melting point it is because of the pressure
effects. And in principle people can actually calculate, what are the pressure effects?
How does the free energy change as a function of pressure?
356
Basically, we all remember G is a function of temperature, pressure, concentration. In a
normal free energy expression that we talk about, we do not talk about Pressure, because
we are fixing it. If you do not fix it will change and once, you do that you will actually
see this happening and that is an explanation for that particular question.
Student: Bismuth instead of lead
Yes, anything, any metal inside another higher temperature metal, if the outside metal
matrix is a lower meting point then this problem does not come, because by the time you
reach the melting point of nanoparticle, matrix has melted already. So, there is no
problem, the problem comes that you have reached the melting point of that, but the
outside has not melted, that is still solid and that applies pressure on this and that
pressure influences the free energy and once it influences the free energy the melting
point gets influenced.
357
Module #01
Lecture #16
We will take up today, one aspect, which was related to the free energy associated with
under cooled melt. We talked earlier that the, “I” which is the nucleation rate is basically
related to, I would say, I will not even put the equation there, is related to Δ G *. All of
you know the equation so, I will not put the equation; of course, there is also, another
term, which we always remember, which is the ΔGd term, which one has to consider and
ΔG* is related to ΔGv. That expression also many of you know, for a spherical nucleus.
We can talk about it.
And if, that is what controls, ΔG*, for example, I can put the equation for this case.
ΔG* = (16π/3).(γ3/ΔGv2), let us say. So, that means ΔGv controls ΔG*, which to a large
extent controls I, at least, at the temperatures above the nose of the TTT diagram. At all
temperatures above the nose what controls the nucleation rate is, only ΔG*.
358
So, if that is the case then, the nucleation rate to a large extent is dependent on ΔG v and
how do I know that value? What is ΔG v? ΔGv is a driving force for solidification. If I
draw this G verses T, so this is, let us say solid; this is let us say liquid. And this is our
melting point Tm (see the graph above). And, at any temperature below the T m, T < Tm,
liquid has a higher free energy than the solid. And, this difference in the free energy
between the liquid and solid which is what we call it as ΔG v. That is our ΔGv and that is
what drives the liquid to become a solid.
And, the higher the under cooling, usually ΔGv is higher, as you can see from that curve
itself. This two, start deviating more and more, as you approach lower and lower
temperatures and at Tm the ΔGv = 0. That is why, in principle at the melting point, there is
no transformation. You do not have any transformation at all and you need the liquid to
be under cooled. You may say that even for heterogeneous nucleation for that matter,
you need a small under cooling; however, small that could be, you need a small under
cooling for the liquid to transform to solid. We are not simply talking in terms of the
barrier for nucleation. Barrier for nucleation is any way there that is the surface energy.
I may say that I am providing you the surface energy, why not the transformation occurs
at the melting point? Because there is because when a surface is already provided, what
is the problem? Then, we say, if I want the transformation to occur at the melting point
359
ΔGv = 0. And, if the ΔGv = 0, that there is no driving force at all. So, even if you provide
a surface, which is a perfect wetting surface, let us say the θ = 0 contact angle is 0, even
then you need an under cooling.
But, usually because such surfaces are not available, you need a higher under cooling
one, to give you a driving force; second, to take care of the surface energy problem. If
you have certain driving force that is available, part of this driving force can be spent to
create the surface, so that the surface energy problem can be taken care. So, that is why,
we always write this expression ΔG, which is the overall change in free energy, for
liquid to solid transformation is given by
ΔG = (4/3).πr3.ΔGv + 4πr2.γ. Remember, assuming that γ = 0; ΔG = (4/3).πr3.ΔGv
So, the driving force for the nucleation for the solid to form, if you assume that there is
no surface energy at all; that means, there is a perfect nucleating sight available and the
surface energy is very-very small. So, even if you assume that then ΔG = (4/3).πr3.ΔGv
and at Tm, ΔG = 0. So, in principle, you need a higher driving force. So, that this takes
care, of the necessary driving force. So, only when ΔG is negative, a transformation will
occur. Any transformation, for it to occur, the driving force has to be negative. This ΔG
is the overall driving force, which takes care of both surface energy and ΔG v. And in
these two terms, we know pretty well that this is, at T < T m, ΔGv is negative term
whereas γ is positive term. This also we know.
And, as a result, first term has to be higher enough, to take care of driving force. This is
what, we are talking about. So, you can see that from all these arguments, we come to
one conclusion that it is ΔGv, to a large extent decides ΔG*. Of course, γ is also very
important. But, γ is not in my hands. γ is an interfacial energy between solid and liquid,
and once I say define a particular metal that I am solidifying, its γ is fixed. So, that is not
in my hand, but this is in my hand; this ΔGv, I can change it.
So, many-many people have tried to understand, how to calculate, this ΔG v and if, you
want to calculate that ΔGv basically, it can be broken into two parts. One is ΔH we will
not talk about v any more now because whenever, we say ΔG, it is basically the volume
free energy that we are talking because the moment, it is surface we use actually a
different term γ. So, the moment, I say ΔG it is always the volume. So, if I assume that
ΔG=ΔH-TΔS
360
I want to know, ΔG at a temperature T < T m. This is what, I want to know. And if, I want
to know that what I need to know? I need to know what is ΔH at temperature T. What is
ΔS at temperature T < Tm? If I know these two, I can calculate ΔG. How do I know these
two? To know that, that is where we see that ΔH at any temperature T, can be expressed
as an expression, in terms of ΔHf. I need to have a reference point. What is the reference
point that I know, is the melting point. At the melting point, I know what is the Δ H, any
standard hand book will give you, what is the heat of fusion at the melting point. But, I
do not know, what is this heat released between liquid at a temperature lower than T m,
whenever I say Δ, Δ means liquid to solid. In our case it is liquid to solid, some other
case it can be some solid to another solid.
So, whatever I am talking of is a change of enthalpy, when a liquid is changing to solid

that change is what, we are talking. And, that change at the melting point is known to us,
at the melting point it is nothing but ΔH f, latent heat of fusion and if, I know that then I
can say it is
Tf
∆ H T =∆ H f −∫ ∆ C p dT
T
this I can say. So, ΔH, at any temperature can be found out, if I know the ΔC p. What is
the ΔCp? Heat capacity difference between liquid and solid; that means, if I can have an
expression of heat capacity, of type Cp = a + bT + cT-2
And, if I know these a, b, c constants for both liquid and solid, in principle, I can plug
them into above equation and then simply do the integration, get its value and ΔH f is
known to us and I can find out ΔHT. Similarly, I can say
Tf
∆C p
∆ ST =∆ S f −∫ dT
T T
This appears to be such a easy thing. But, in practice, it is not easy. Where is the catch
here? The catch here is trying to find out, when I want ΔCp, what is ΔCp?
ΔCp = Cps - Cpl.
Remember, whenever I say Δ, I always write product minus the reactant.
The product for us is the solid, the reactant is the liquid. If I want ΔCp at temperature
T < Tm, I need to know the Cp at the temperature T, Cp of solid at temperature T, Cp of
liquid at temperature T, then, I can find out ΔC p. Now, the question that comes is to, find
361
out Cp at temperature T, is not very difficult. I have an aluminium; I want to find out the
heat
362
capacity of that aluminium at temperature 600 ºC, which is below the melting point.
Melting point is 660 ºC, measuring this heat capacity of aluminium solid, at 600 ºC, is
not a problem. I can put it into a DSC and happily measure the heat capacity.
But, if I want to measure the heat capacity of the liquid at 600 ºC, how do I measure? I
want to know the heat capacity of liquid aluminium, at 600 ºC. To know the heat
capacity of liquid aluminium, at 700 ºC is not a problem for me, no problem, but at 600
ºC, liquid does not exist and if, I somehow bring the liquid aluminium to 600 ºC
centigrade and start measuring, by the time you measure it may not be liquid, it will
become a solid.
So, practically, it is very difficult to measure the heat capacity of the liquid, at
temperatures below the melting point. That is why, if you have ever seen a hand book
such as, Smithells hand book, they give you, C p data in the form of Cp= a + bT... and
things like that. And, they say that this particular thing, the a, b, c values that they are
giving are valid only in this temperature range they say. For a liquid, they say above 660
ºC for aluminium it is valid, below 660 ºC that particular a, b, c are not valid because
nobody has measured them, nobody can measure them, unless there is a liquid which can
be easily under cooled.
And, still kept as liquid before it actually crystallizes, something like a glass forming
liquid, like a silicate glass, you can happily under cool. Yes.
Student: Nano aluminium we can’t?
Nano aluminium, No I am talking of measuring the liquid. Nano aluminium actually
melts at lower temperature. Is not it. We are not measuring the heat capacity of the solid.
We are talking of heat capacity of the liquid.
Student: Without any difficulty, we can measure heat capacity for nano Aluminium at
600 ºC, Aluminium is the same.
So, but then, but then the question is that for that particular Nano aluminium, the melting
point you have already crossed. So, you cannot, see you cannot compare apples with you
know bananas.
So, the moment, I say Nano aluminium and its melting point, let us say is 600 ºC, if I
say, I am measuring the heat capacity at 600 ºC, for the liquid, but that is a stable liquid
already, it is not a metastable liquid. It is not a under cooled liquid. And then, if we are
measuring of that is of no use to you because that is of a different type of material. I am
363
talking of a bulk liquid, under cooled to below the melting point. And, I am measuring
the liquid, what is called the Cp below the melting point and maintain that liquid as a
liquid, during this measurement.
364
And, this is possible only, when the solidification of the liquid into a crystal, is very
difficult that liquid can remain. There are, so many, you know polymer liquids, which do
not want to crystallize easily. All, that is why, if you go to any, you know modelling
fellow for solidification; they all do modelling on polymers. You, if you have ever read a
book on solidification like a “Bruce Chalmers” or if you have read a book “Mallin
Zacatecas” book or any other books, you will see that all the modelling works that they
have done Physical modelling, not really Mathematical modelling.
Physical modelling works are all under cooling experiments, for example, how do I
know a material has under cooled? I say that under cooling and micro structure is
related, why “under cooling” and “I “are related. So, I can actually measure the “I”. How
do I measure the “I”? I need to be able to see, how many nuclei are forming? And if I
want to physically see these nuclei forming, I should have something which is a
transparent, a liquid in which actually nuclei are coming.
So, people have done a large number of experiments on transparent liquids, which are
polymer kind of liquids, Cecilio Nitrate and there are so many such things, where people
take such liquids, which are more difficult to actually crystallize. And, in such liquids
one can do this. In normal metals this is very difficult you cannot under cooled a metallic
liquid and then, retain it to be liquid and that is where, the problem comes is
measurement of this and that is where, many people started making assumptions.
365
The first fellow to assume, something is a person by name” Turnbull”. Turnbull said, “let
us assume, ΔCp = 0. As there is a crudest assumption, one can think of and we know,
definitely it is not close to reality. But, there are so many things we assume for example,
Ideal solution Thermodynamics is not close to reality, but we still assume. So,
assumptions one can assume anything that he wants.
So, he assumed ΔCp = 0; that means difference between the heat capacity of solid and
liquid is zero. Once, you assume that what happens these integral terms vanish in ΔH
and ΔS expressions.. Then, I can say ΔGT = ΔHf - TΔSf. Why? Because ΔH = ΔHf,
ΔS = ΔSf. And, that is where; you see a typical expression that you see in all standard
text books. And, ΔSf= ΔHf/Tf. Only, at Tf it is valid. Is not it. So, because from the basic
definition of ds itself, ds = dH/T.
So, if I use that then I can simply say that

ΔGT = ΔHf - TΔSf
ΔGT = ΔHf - TΔHf/Tf
ΔGT = ΔHf.ΔT/Tf
ΔGT = ΔSf.ΔT
So, you can now, say ΔG is the above expression. And, this is what all standard text
books talk about it, but this you would not have known has come from such a easy
assumption, such an assumption which is actually not valid. So, but this is what, all text
books talk about and then, after that there are large number of people, who have tried to
make various assumptions because measurements are difficult and after assuming bring
out an expression, for ΔG.
And then, try to take certain you know experiments, try to conduct certain experiments
on those systems, where you can actually under cool the liquid and still, measure the C p
of the liquid, which are not definitely metallic systems, but other type of systems. And
see if that ΔG that you have calculated, based on that particular formula or derived based
on the assumption matches with the actual ΔG that is measured. How
366
do you measure the ΔG? By measuring this ΔC p, I can get this ΔH and I can measure this
and I can get the ΔS. Once these two can be measured, I can actually measure the, find
out what is the ΔG, at any given temperature.
And that if it matches with whatever assumptions that we have made. And what are the
various assumptions? One assumption, where people have said, let us assume ΔCp is not
0, but it is constant and ΔCp = ΔCpf . This is one of the assumptions made, which is what
we call it as, Thompson and Spaepen.
Student: Can we extrapolate it
Correct, so that is what, one possibility could be if I know how Cp is changing of the
liquid, you extrapolate it. That is one way, but the thing is, we, for us to know, whether
the extrapolation is valid or not, you need some experimental you know any
extrapolation somewhere you need to have some, you know Experimental validity.
And, the experimental validity is actually coming from non-metallic systems. So, just
because this extrapolation is working in non-metallic systems, can I assume that it would
work in metallic systems, is another thing, so one can always argue, but people are trying
to do that. So, people are trying to do that and this is one such assumption. And, once
you take that ΔCp is constant then; again this integration becomes easy for me. I can take
the constant out. So, and then, there are a variety of, if you are interested there are some
recent papers on this I would get you a reference, a paper by, as we have also done a lot
of work on this.
In Applied Physics letters, 2003, by the first author is K. Mondal, you can see Mondal
and Murty, where we have tried to do certain other ways of doing the same thing. I do
not want to go deeper, unless you are really serious about it, to find out an expression,
which would fit for variety of liquids because most of the assumptions that people have
done each of them have certain limitations.
It would fit for, you know binary liquids, but it would not fit for ternary and quaternary
because now, we are talking about bulk metallic glasses. Luckily, in case of bulk
metallic glasses, people are able to do certain experiments because bulk metallic glasses,
one of the advantage is that liquid does not want to crystallize easily because the TTT
curve is shifted to the right, to a great extent; that means, the nucleation is more difficult
in such liquids because the liquids are viscous.
367
So, in such liquids, there is a possibility of making measurements. So, in such cases,
people are able to measure and see that whatever long back people, I can tell you there
were the major assumptions one was, by this man Hoffman. If you read this paper, you
will get all these details. Then Thompson and Spaepen, I have already given and Jones
and Chadwick. One has said let us assume the Cp as that Cp where ΔH = 0.
So, Hoffman’s assumption is that so, ΔCp is that at ΔH = 0. In principle, at some

particular temperature that ΔH actually goes to 0.
It could be a theoretical temperature. The way I have showed you earlier, if you
remember this, when, you plot at ΔS f or simply S entropy itself (left graph). You
remember there is, this kind of a thing and we said if I extrapolate this (dashed line, left
graph), at some time it intersects that, and at that temperature ΔSf = 0. Similarly, if I
draw H verses temperature, I will again see a similar situation and that H also, if I
extrapolate at some stage it would go to 0 (Right side graph_.
And, in principle, actually if ΔSf = 0 and ΔG = 0, automatically at the same temperature,

ΔH = 0 because ΔG = ΔH - TΔS. So, if Δ G = 0, ΔS = 0 and automatically, you will see
ΔH = 0. So, there are lot of theoretical calculations to see, whether S and H are exactly
the same or not. And, to a large extent, they come closer to 0. So,
368
whereas, when Hoffman talked about it, he does not bother about entropy. He says,” let
us assume the ΔCp and take the ΔCp absolute value of it as that that you will get when
ΔH goes to 0. At that value, find out, what is the C p of the liquid, what is the Cp of the
solid.
And, in principle, one can calculate, from the equation
∆ H =∆ H f −∫ ∆C p dT
So, you will get some value of ΔCp, in terms of ΔHf. And, that value, you plug it into that
particular equation and then, try to do it that is one way of doing. And Jones and
Chadwick said, let us assume, the ΔCp is that value, where ΔSf = 0. This is another
assumption. A variety of assumptions, I want you to actually go through because there
are lot of derivations I do not want to really go into these derivations, but for those who
are interested, if you read this paper, you will know most of them.
So, this is one problem that all of us face when you talk about under cooling. So, we do
not really know. And to a large extent, everybody uses ΔGT = ΔSf.ΔT
because it is a very easy expression. And in fact, I should also tell you that in this
connection, there is a professor by name professor P Ramachandra Rao. How many of
you have heard of him, anybody who has heard? Who is
369
this man? He was a former Vice Chancellor of BHU; he is a man who did a lot of work
on under cooled melts.
And, he has two theories credited to him on the same thing. One he tried to expand the
ΔG by a Taylor series expansion. He simply took it as a mathematical expansion and
when you try to expand ΔG, as a function of Taylor series, then you will get different-
different series order, the First order, Second order, Third order, at some stage you
truncate it and try to calculate the ΔG, that is another way. Again, in this (Mondal and
Murty, APL, 2003) paper you will see that also.
So, there is again another thing called Hole theory. Assume that the liquids are
composed of a free volume that there is always certain holes as if because definitely,
there is much more free volume in a liquid than a solid because solid has definite
crystalline structure, crystalline solids only I am talking about. And, there you talk about
tetrahedral voids and octahedral voids and things like that whereas, in a liquid, you have
much bigger voids that we can talk about.
And, that is what, Hole theory talks about it. And, from that they tried to derive, what
would be the Cp of a liquid at various temperatures, assuming that these holes kind of
change their dimensions as a function of temperature and so, there are certain theoretical
assumptions there again. So, when they tried to do it, they got one expression, which is
based on that, that is again another expression that is used.
So, there are variety of these expressions, which are available in the literature and people
depending on, what suits a particular system. In all the cases, they suit a particular
system, but they may not really suit because that is where, each liquid behaves
differently, like the each of us behaves differently. So, each liquid would be a different
because of its nature. Though, we can more or less assume that liquids are all similar
because there structure is similar, but if I take an iron based liquid and an Aluminium
based liquid, both of them, though, we say it is liquid, this Iron, when it solidifies it gives
you a BCC, Aluminium when it solidifies it gives you FCC. Is not it.
So; that means, there must be something different in these two liquids. Though, we say
they both are liquids, that is where you know, you would see the clusters are different, in
both of them. And based on these clusters, its behavior would be different. So, that is
how people see that some of these expressions are valid for certain things in some
370
cases, they do not. So, now let us see, yes.
Student: Validation how do we do?
How do we do? Yes. That is another question, Validation how do we doThe reason why
people bothered about all this to a large extent is after 1959.
After 1959, when Pol Duwez showed that it is possible to get a glass from metallic
liquids, that is the first time somebody talked about a metallic glass. Till then, people
knew about glasses, but they are all silicate glasses that we know a silicate can be easily
made into a glass. But, a metal can be made into a glass is something, which for the first
time people have seen in 1959. Then when they are able to make so, people wanted to
know, what was the under cooling that is necessary for making a liquid into a glass.
So, that is where, the lot of calculations took place to find out, what is that ΔG, how does
it change the driving force for solidification, how does it change as a function of the
temperature. And there, they tried to always take the glass forming liquids. The
advantage of glass forming liquids is, glass forming liquids, the liquid is a stable liquid.
It does not want to crystallize easily, that characterizes a glass forming liquid from
another liquid. That is why, when I said, if you take Eutectics, we talk about a deep
eutectic and a shallow eutectic. In one of the classes, I think we talked about it.
And, In fact, we even, tried to give how do we measure this depth, in terms of a certain
expression. We basically, take the melting points of the two metals. Consider a kind of a
linear relation between the two for a particular alloy and differentiate that temperature. I
mean, take the difference of that temperature, with respect to the eutectic temperature.
And, we call it as ΔTlmix, something like that. And we said if it is smaller than certain
number, it is a shallow eutectic; if it is greater than a certain number, it is a deep eutectic.
And, these are all more of empirical because they have seen that wherever this crosses
certain number they saw a glass formation, in that particular eutectic. Wherever it is less
than a particular number, they did not see a glass formation. So, from that people started
talking about shallow and deep eutectic and in all these deep eutectics, where glass is
easy to form. And, in such cases, people were able to make those measurements. So,
when you say, how do you validate, people were able to validate in those cases, where it
is easy to; that means, you can under cool the liquid and still this liquid is not able to
become a solid, crystal.
371
And, during that whatever little time that is available, you know for example, if you have
heard about you know, what is called micro gravity solidification. Interesting, I would
like to tell you, all those who are interested in solidification, November 20th to 23rd here
we have an International Conference, which we are organizing, where all the people who
have been working on solidification, who are working on solidification, the top grass
people come. It is a conference, which we organize once in three years and bring
everyone, who does modelling on solidification, who does experiments on solidification.
So, that is where, you will see a lot of people and the best person who works on glasses,
is one Prof. Inoue he is coming. So, there are lot of people who come, it is a small
conference about 100 people, but it is a very-very focused conference, everything on
solidification and we have been doing it this is the fourth conference we are doing, last
10 to 12 years we have been doing this.
So, what I wanted to tell you is that you will see that all those people who have been
doing this under cooling, is one branch of this under cooling which people call it as, you
know solidification under micro gravity. When, I try to solidify liquids without gravity
or at very small gravities, how does how does this liquid solidify? Whether the micro
structure, that evolves is going to be same as what you get under gravity or not, we are
all compelled to have a gravity on the earth.
So, whatever micro structure that is coming, we accept it that this is possibly the type of
micro structure because there is always, we are aware of what is called you know,
gravity segregation in most of the alloy solidification. But, this segregation is something,
which we have to accept because gravity is there. So, in case, there is no gravity;
obviously, you would not see this gravity segregation takes place. Similarly, dendrite
growth, dendrite nucleation, is it a function of the gravity.
So, there are lot of people, who do experiments. There are people who have been doing
experiments on ISS, the International Space Station. In fact, Kalpana Chawla, who died,
many of you have possibly are aware. In one of the Space Shuttle, they were doing at
that time experiment on solidification, and incidentally again, this is a experiment
designed by another Indian professor from IIT Kharagpur, there is a professor by name
372
Dhindaw who does a lot of work on micro gravity solidification, this was an experiment
that was going on there.
Basically, they want to do, see if you take a composite, let us say Aluminium Silicon
Carbide composite, you want to know where these silicon carbide particles segregate.
Do, they go to the grain boundaries of aluminium or they are inside this Aluminium, that
depends on whether the growing front, will push the silicon carbide particles or will it
engulf. Once it engulfs, then you will see it curves inside. This engulfment or pushing,
there are lot of theories that are available. People wanted to prove it in a micro gravity.
So, they are doing experiments on that.
So, there are lot of people, who do micro gravity experiments, who will come here and
make their presentations. You will see them. So, what is important there again is, how
does various forces influence the solidification? We have been only talking about, only
temperature as one of the force. We even, do not consider pressure at all. So, the moment
you bring in the pressure aspect, how does it influence? So, this is, pressure can be just a
vacuum, you can talk about it and, the other thing that you can think of, how the micro
gravity brings in pressure? That is another area, lot of people are interested in that.
So, the point that I want to make here is that many people who are interested in under
cooled liquids, whether this under cooling is achieved because of micro gravity because
when you do it under gravity, many people have observed the liquid does not become a
solid immediately at the melting point. But, it under cools and then, becomes a solid. So,
and they were able to see that this extent of under cooling is a function of the extent of
micro gravity that you provide. But, in a normal earth, it is very difficult to provide a
micro gravity excepting under the cases, where people have designed for example, in
Germany there is, what is called a drop tower. You take a liquid droplet and from a
hundred feet, you drop that liquid droplet.
Let us say, a furnace, a small furnace you keep it, liquid is you know metal is liquefied
and this liquid, through an orifice it comes out and then simply falls in a drop tower,
which is evacuated and kept in high vacuum. So, that there is no oxidation and things
like that. And, such liquids, when they took after it solidified, took that solid piece and
then try to, solid when I say solid piece, it is just like a flake because what you are
dropping is a one millimeter droplet possibly. And, they have seen that from the micro
373
structure it is all indirect. You see the micro structure and from the refinement that you
have seen in the micro structure, you quantify, what could have been the nucleation rate.
And, from the nucleation rate, back calculate what would have been the ΔG v and from
the ΔGv again, back calculate and say this could be the under cooling. So, this is how
people go because there is no way directly, for you to measure the under cooling.
Excepting, if there is a way, that it is a transparent liquid, that I am under cooling and I
can see with a with a video, very fast video that as I under cool I can for example,
professor Phani Kumar does experiments on Levitation.
If I can levitate a liquid, and then, where there are no surfaces available now, this liquid
is now being allowed to solidify and I keep on reducing the temperature and then, catch
up if it can show where the nucleation is coming. And, in a normal metallic material,
they are all opaque. So, I cannot really see, if the nucleus is coming inside the liquid
somewhere, at some particular temperature, catching that temperature is not very easy.
But, in other organic liquids, where there is a possibility of under cooling, people have
done these experiments and to see, where exactly, what is the temperature at which the
nucleus form during solidification.
Of course, again the problem is, nucleus is very small, of the order of few Nanometers
how do I catch that? Because my camera may not have that resolution, obviously, it is an
optical camera it would not have a resolution of an Electron microscope. So, it has to
grow to a few micron size for you to be able to be actually catch it. So, these limitations
are there within these limitations, there are lots of people, who are trying to do those and
find out that temperature, where they have seen the first solid. And, back calculate that if
this is my resolution of let us say ten microns or so.
And, for if it, I visualize it at this particular temperature, at what temperature it would
have nucleated? Again with certain assumptions, one can back calculate it. So all these,
there are lots of people, I mean, it’s a wonderful field solidification, lot of people are
working on it. So, again all of them are driven by this simple thing that how do I actually
predict? Before even doing the experiment can I predict that if I cool this liquid at this
rate, this will be the grain size that I will get. That is what; everybody wants to know that
can I predict the grain size that I get or a micro structural refinement that I can get.
374
And, or what should be the cooling rate that I should use to get a glass in this particular
liquid. So, for all this, you need to have the models that are available, where you have a
relation between the cooling rate and the under cooling. Cooling rate and under cooling
to some extent we can understand. So, a relation between cooling rate, under cooling and
the ΔG, so that ΔG to I again the relations are more or less clear to us. ΔG to I relations
are clear.
But, the ΔT to ΔG relations, need to be established. And, that is where, all these
modelling has been happening. And, if you are interested, I can give you a large number
of papers, but any way, we will close that discussion at that. There is another aspect,
which I thought, I will tell you again because we are talking of again nucleation. We
started it, but we did not go to a large extent on that earlier, what is the shape of the
nucleus? You asked me this question some time back. What should be the shape? We
always easily assume sphere.
So, what is the shape of that? So, I was telling you that each plane in a crystal because
when I say, a solid, I am not talking of a glass here, I am talking of a crystalline solid,
that is coming out of a liquid and if it is crystalline solid that is coming out of a liquid, it
has a unit cells. It has planes and different planes have different surface energies. If
different planes have different surface energies then, and how do I assume that a liquid
will have a
375
spherical particle because the moment, I say spherical, what kind of a crystal structure
that I can imagine, which would give me a spherical particle?
Because basically, what is a nucleus? Nucleus is nothing but a bunch of unit cells. There
are a number of unit cells so; that means, I take a unit cell, I extend this unit cell with a
number of other bricks, it is like a wall with a number of bricks in all the three
dimensions that becomes a nucleus. And if, I take a brick with certain shape, we have
seven shapes that are known to us, which are crystal systems and 14 bravais lattices. Out
of these 14 bravais lattices, I chose any one of the bravais lattice and that unit cell I add a
number of those bricks, in all the three directions. Can I ever get a sphere out of it? It is
not possible.
It is not possible that I can think of a nucleus which is composed of unit cells and the
unit cells, none of the unit cell is a sphere. If the unit cells are not spheres then;
obviously, this nucleus cannot be a sphere then, what should be the shape of it? Then
what people said is, now let us assume that the unit cell, that the nucleus is composed of
certain planes. The moment, you think of the outside planes, at this moment, we are
thinking that it is the sphere, it is a curved. It is not possible that atomic planes are
curved, atomic planes are not curved. So, as a result, a actual nucleus would be
something like this, with certain planes on the outer surfaces in a 2D it is like this (see
the figure above), in a 3D it will have other faces.
Now, if that is the case, then the actual shape of the nucleus is that which minimizes the
overall surface energy. It chooses those planes, for its enclosure in such a way that the
overall surface energy is minimized, for that Wulf, there is a Gibbs plot people talk it.
Gibbs, long back talked about it, what people call it as Gibbs plot. What he has done is,
let us assume, a crystal structure. Let us say, a cubic, let us say, FCC aluminium is
solidifying, a liquid aluminium is solidifying into FCC aluminium, what should be the
shape of the nucleus for the FCC aluminium that is coming out?
If, I want to know, let’s simply think that every plane has a certain surface energy. There
is a γ111, γ110, γ100, and so on. And, this surface energy depends on what? It depends on the
number of broken bonds. And, one can easily calculate this, in principle and so, these
surface energies are there. And, if I try to find out a polar plot, where if I draw take a dot
376
and then, draw a direction (see the figure above) and call this as 100 a perpendicular
direction, I call this as 010 and this is 001 let us say (out of the blackboard).
And I say in these direction, I know these values let us say γ’s and in each direction the
length of this particular vector, I take it as a scaling of one of the surface energy, let us
say and so, let us say, if I take a vector like this, from center to 100; that means, this
represents γ100, let us say and similarly from center to 110 is γ 100; that means, every
surface energy, for every surface I show it in the form of a vectors. And then, once I
show that I can talk about γ in any direction, say γ1-10
and perpendicular to will be γ110. And, the value of this γ, if I take, any scaling factor.
And once, I take the scaling factor, I can talk about this is the value. And, once I put all
those vectors then what you do is, you simply draw perpendicular to those vectors and
this perpendicular now, represents that plane. Any vector for a particular plane,
perpendicular to that is actually, that particular plane.
For example, if I draw a perpendicular surface like this (coming out of the board, at 100),
that surface actually is a 100 surface. And similarly if I draw a surface here (coming out
of the board, at 110), I can think of that as 110 surface. Now, what you do is, there can
be some plane, some plane which could be 210 and its surface energy is so high let us
say that this vector goes out to a large extent and now, I draw a perpendicular to that
also. And for, each of these vectors you draw perpendiculars, in a 2D we are doing, in a
3D basically you are drawing surfaces, which are perpendicular to that vectors. And if,
you do this, find out that particular enclosure, which has a minimum area in a 2D, in a
3D the minimum volume.
That will be minimization of the surface energy, for that particular structure. If, I know
for a particular structure and this γ111 for an FCC will be different from γ111 for a BCC.
For a given structure, I need to calculate these values because 111 plane in an FCC is
different from 111 plane in a BCC. Their packing factors are different. So, as a result, for
a given structure, if I can calculate these things and do this and then find out a particular
a shape, which gets formed, which has a minimum area is the one that actually forms
during nucleation.
The moment, a plane, one of the plane, the vector is going out; that means, it has a large
surface energy, I am avoiding all those large surface energies and taking only that
particular area, that is the system would choose that particular shape, in such a way that
377
it is, it need to spend minimum energy because surface energy is something, which you
need to spend, and, where from your spending? It is from the Δ G v that is available. So, it
would like to spend the minimum energy, before it actually solidifies into a shape.
So, it would choose that particular shape. This is what is called Gibbs plot which is, what
has been used to find out and people have done this kind of calculations, to find out,
what should be the, that is what, people talk about tetrakaidecahedron is the shape, that
comes out and variety of names are given, to such shapes depending on, which structure
that I have. If, I say BCC, I can actually, in principle, calculate what should be, but will
that remain after it grows, is another question.
The point is, once it grows, the shape could be different. Why? Now, if I take a particular
shape, I will continue a few more minutes, when it grows the rate at which 100 plane
grows and the rate at which 110 plane grows is different. Why because the packing
density of 100 plane is different from packing density of 110 plane, whenever, 100 has to
grow the rate at which it grows, depends on how many atoms, this plane needs to fill up
that plane. When that plane is filled, it grows by a small Δx. So, if a 110 plane needs less
number of atoms than 100 plane, 110 plane will move faster than 100.
The plane which is loose packed plane will always move faster than the plane, which is
close pack. And, as a result, after sometime, you may see that this if 110 is not growing
faster. If, other four are growing faster, slowly you will see that the final shape, later
after the growth, you will see the shape would have changed, these 110 planes will not
exist there. So, you will see a different shape.
So, the nuclear shape is possibly, is a why possibly? I should say it is definitely different
from the shape after growth. But, the problem is, we can only see the shape, after it has
actually grown. We will not be able to see the shape of the nucleus, unless you freeze the
nuclei somehow and put it into an Electron microscope and be able to see it. So, that is
another problem with all these calculations, though, we are able to calculate, what is the
shape of the nucleus, We do not know, whether this is going to remain and the actual
shape that I am seeing is a growth, shape modified by the growth.
378
That is the reason, why you will always see, for example, in Aluminium Silicon alloy,
you see Silicon needles. In an Iron Carbon alloy, the cementite is like a blocky particles
chunky particles. Why does this happen, because certain directions have a faster growth
rate in a silicon. So, it grows faster there. So, this is another issue that one needs to
consider. But, for the sake of academics, one can calculate the shapes like this. That is
why, this you do not see in every text book this information because this has its own
limitation because this shape, that it can predict. You may not be able to actually, seeing
this shape after it has grown. So, it has; obviously certain limitations, we will stop.
379
Module #01
Lecture #17
Solid state phase transformations – Precipitation
Today, we will start a new chapter, on our topics. The new chapter is - Solid State Phase
Transformations. We have been so far talking about mostly on liquid to solid. You have
seen this so far; we talked about solidification, melting. We even talked about
isomorphous solidification; we talked about eutectic, we talked about peritectic
solidification. We even went to the extent of how a liquid becomes a glass, and we also
brought in a T0 concept in the meanwhile, and so understood what is called metastable
phase diagrams, related to that.
So, today, let us bring in solid state. So the moment you talk about solid state; that
means, one solid giving you another solid or one solid gives you, S1  S2 or S1  S2 +
S3; variety of transformations. What are the solid-state transformations that you are
aware of?
Student: Eutectoid
380
Eutectoid is one, eutectoid is something like this one. You can call this as
381
a eutectoid transformation S1  S2 + S3. So, the well known eutectoid, Aneesha which
is the eutectoid that you are aware of? What is the type of transformation?
Student: Fe plus cementite
What gives Fe plus cementite?
Student: Α plus
What gives α plus?
Student: Fe carbon diagram
What is the eutectoid reaction in a Fe-carbon diagram?
Student: Austenite gives cementite plus ferrite
So, the austenite gives you ferrite plus cementite. This is the eutectoid in the Fe-
cementite phase diagram this is we are aware of. Any other solid-state transformations
that you know, Prerna? Solid-state, where something in the solid state gives you
something else in the solid state. Anybody else?
Student: α to γ , γ to δ
α to γ, γ to δ, this is something like this S1  S2. What do you call them as? There is a
name given to such transformation, when α goes to γ, let us say α-Fe becomes γ-Fe, they
are called allotropic transitions, they are called allotropic. And the same thing if it is
happening in alloys, we call them as, polymorphous transformations. Allotropic
transition is a term, which we use in case of pure metals or even, for example, in
ceramics.
You must have heard of alumina. Alumina exists in different-different structures, γ-

alumina, α-alumina. These are all allotropic transitions, even silica, Cristobalite and
variety of forms. Titania, TiO2, many people are aware of nano titania, particularly. Nano
titania, do you know what is the particular purpose, people use, who want to use nano
titania? Self-cleaning; self-cleaning is all nano titania, nowadays. So that is just one of
the applications. Nano titania, in Titania, there are allotropic transitions.
What are the most important allotropic transitions in TiO 2? Anybody knows? Anybody,
who is attending any ceramics class?
Student: Rutile to Anatase
Rutile-to-Anatase, Anatase-to-Rutile that is another transition, these are all allotropic
transitions. There is no compositional change during this, only structure is changing. We
382
call all of them as allotropic transitions, and in some alloys also, you can have transitions
where composition does not change, only structure changes. We call them as
polymorphous transformations, in general and usually in practice, we use the term -
Massive Transformation.
383
Why we call them as Massive, is because such a transformation is very fast. Why it is
very fast, because, it does not involve.
Student: Long range diffusion
It does not involve long-range diffusion. You cannot say that it does not involve
diffusion. Are you aware of any transformation where it does not involve diffusion?
Student: Martensite
Martensite - so that is the next type of transformation, which is where you have
martensite.
384
So in fact, if you want to talk about solid-state transformations, solid-state
transformations can be basically divided into, diffusion based and diffusionless. This is
385
simply one-way of talking and diffusionless is what is called martensite. And in
diffusion, you have short-range diffusion and long-range diffusion.
All those short-range diffusions are these: allotropic transitions and you have one you
can talk about- massive. Any other transformation that you are aware of, which are
short-range diffusion, where the diffusion distances are not more than a few inter-atomic
distances, may be one unit cell, may be a few unit cells.
Student: Order-disorder
Order-Disorder, very good order-disorder, so, these are certain transformations, which
are basically short-range diffusion transformations.
Anything else, that you are aware, which is again short-range? Spinodal, Spinodal
transformation again is a short range, at least at the nucleation stage, and as it progresses,
it grows. So, spinodal also more or less, can be at least initially, at the nucleation stage it
is very short-range.
And then long-range diffusion, what are the things that you are aware of? One you just
now said, what did you say Aneesha, there is a transformation you talked about?
Student: Eutectoid
We are talking of, eutectoid. Do not say eutectic, eutectoid. So, eutectoid this is the one
and anything else that you are aware of?
Student: Peritectoid
Peritectoid OK fine, let me take it.
But there is something which you are commonly aware of.

Student: Precipitation
Precipitation, very good, where an α, super saturated solid solution, gives you α plus a
precipitate. You call it as let us say, θ, in case of Al-Cu, in case of something else; you
can talk of something else. There are so many types of, even in steels we have a
precipitation.
What is the precipitation that you are aware of in steels? In a plain carbon steels, what is
the precipitation that you know in plain carbon steels? Which is regular, I mean all of
you in second year B.Tech- Metallurgy, have come across, when you are talking about
heat treatment of steels.
Student: Lead
386
In heat treatment of steels, lead is not related to that. Yes, somebody murmured some
word, just now. Yes, what is the precipitation that you know in steels- plain carbon steels
during heat treatment, one particular type of heat treatment it happens, it is not in alloy
steels.
Student: Tempering, the ε carbide
Tempering, yes, the ε carbide formation in tempering is again that, where martensite
with a particular carbon content gives you initially ε carbide and finally, cementite. Is it
not?
387
So, that is also a precipitation where one structure with a certain composition, leads you
to a precipitate with a different crystal structure and a different composition. The ε
carbide is what is the composition of ε carbide anybody knows?
Student: Fe2.4C
Fe2.2C, 2.4 or 2.2 C, this is what people talk about.
So, that means definitely its carbon content is different from that of the steel martensite,
from which you are starting. And its structure, what is the structure of ε carbide? Yes,
somebody said. Tell confidently, whatever you say, say confidently that is important;
you are an IIT-ian. Yes, somebody said it. You only said possibly. Somebody just now, I
heard the answer, but he was not confident. Whoever said, he has said right answer, but
he is not confident. Can you repeat what you have said? You said, No?
Student: HCP
HCP; so it is hexagonal. We will talk about this why hexagonal should come out, we will
talk about those.
So, that is all the precipitation. So it is a Martensite, which is let us say body-centered
tetragonal structure gives you a BCC α plus the ε carbide which is hexagonal. And then
this ε carbide as it grows, slowly it become orthorhombic cementite. Same thing happens
even in Al-Cu. We are going to see now that it first gives you GP zones and then slowly
it gives you θ double prime, and θ prime, and finally the θ. We will see why this
sequence has to be maintained, that also we will try to see.
So, these are all the long-range diffusion, which definitely of the order of a few
nanometers or a few microns, depending on the grain sizes that we are talking about. So
these are all, if you want to generally classify solid-state transformations. We will try to
take some of them and then see how we go about it. We will start with precipitation. This
is a transformation, which commercially also a very important transformation. All age
hardening is all due to precipitation hardening, precipitation that we are all aware of.
388
So, let us look at precipitation, and take the example of Al-Cu itself, the most popular
precipitation hardening. If you remember Al-Cu phase diagram, particularly on the Al
rich side, if you look at it, it looks something like this (see the diagram above). This is
how it looks like. And on this side, so there is the θ phase. This is how the Al-Cu phase
diagram looks like.
Of course θ is what, what stoichiometry is it? What is the composition of θ?

Student: CuAl2
CuAl2, so it is basically, Al2Cu. So, Al2Cu basically means, if I am writing % Cu on this
side, how much is this? How much Cu?
Student: 33.3
33.3% what? It is weight %? It is atomic % correct, so this is atomic % or mole %.
Whatever you want to call it, because Al 2Cu , one atom of Cu and two atoms of Al, and
if you convert it into weight % actually it comes close to about 50 %, because atomic
weight of Cu is almost, even more than double of that of Al. Al is 27 I think 63.5 is Cu.
So, if you consider that, it comes out to be much higher.
Anyway, so this is pure Al on this side (left axis), and when we are talking about
precipitation, what we are talking about is taking an alloy of some composition like this
(point P), and what is this composition (point Q), anybody? I am getting different-
different answers.
Student: 5.6%
389
Yes approximately, anyway, different books will talk differently, approximately 5.6%
but whatever we are talking here is weight %, remember. So, this is all now we have to
convert it into weight %, whatever it is. So most of the commercial phase diagrams that
we talk about are all in weight %, so, this is 5.6 %.
And what is this reaction? Aneesha, can you tell me what is this reaction here?
Student: Eutectic
Eutectic, what gives you what?
Student: Liquid gives two solids
Liquid gives, what are the two types?
Student: θ
θ and, little louder, half of the word is only coming out of your mouth.
Student: θ and α
α, so, liquid α + θ, this is the eutectic reaction there. And we are not at the moment
bothered about that eutectic reaction, which we are talked about it enough in the previous
classes. So, let us not bother about the eutectic reaction as such. Incidentally, this
particular eutectic is a lamellar eutectic. Al - Cu is a standard lamellar eutectic. Let us
not bother about it. We are now looking at the solvus part of it.
So, look at this line, which we call it as the solvus (refer to curve drawn with bold chalk),
and this solvus if I look at it, and take an alloy of this composition (C1), let us say 4 % of
390
Cu, it is a standard alloy that we talk about. And if this alloy is heated to this temperature
(T), what should you get if you heat it to that temperature?
Student: α
It should be α, completely α, if I heat it to that temperature. Now if I take it to that
temperature (T1), cool it slowly, what should happen?
Student: α come out
Yes, by the time you reach this temperature T1 and
391
once you reach a temperature below that, θ should start precipitating out of α. Why?
Because the moment, I come to a temperature like this, (T1) if I come to that
temperature, at that temperature the alloy composition is 4 %, and the solubility of Cu in
the α at that temperature if I draw a tie line, solubility of
Cu in α is this, this is this value(C1’), and because alloy has more Cu content than what
it can dissolve, as a solid solution; this excess Cu should be rejected out. Is it not? That is
the whole principle of precipitation and where does this θ come out?
Student: At the grain boundaries
At the grain boundaries, that is what it is.
So If I look at the microstructure (Fig. 1 above), if I say this is α, all α grains, then you
will see that, in principle, the θ would come out, at the grain boundaries like this (yellow
color lines in Fig. 1). It will nucleate at basically at the triple junctions. These are the
junctions, where the nucleation starts. Again, we will try to understand, why should the
nucleation start at these triple junctions? That is where we need to understand, that
whenever we are talking of a precipitation; α giving you a θ phase, this precipitation
always should have certain driving force, and then because any transformation has a
activation barrier, it has an activation barrier too. Similar to liquid to solid, it has an
activation barrier too.
392
And in this particular case, whenever we are talking of a liquid to solid, the activation
barrier that we talked always is what is called the interfacial energy. Is it not? So that is
why, If you remember we wrote an expression, the ΔG for the transformation liquid to
solid as, for a spherical nucleus
ΔG = (4/3)πr3ΔGv + 4πr2γ
This is what we have written for the solidification, and we said second term is the barrier
for nucleation, and whereas first term is the driving force for the nucleation. You
remember all this. Now, when I am talking of solid-state phase transformation, you have
another barrier, which comes out, that is called strain energy (ΔGe) .
When a structure, which is different from that of the parent phase, is nucleating inside
the parent phase, there is a structural incompatibility between the two and as a result,
certain strain is generated. Even if the structure is exactly the same, if the lattice
parameter of the precipitate is different from that of the lattice parameter of the matrix,
you will see certain strain and that strain is more of a volume in nature, it is volumetric.
So that is why free energy expression is written as
ΔG = (4/3)πr3ΔGv + 4πr2γ (4/3)πr3ΔGe
People call it as e or ε, because strain is usually represented by elastic energy. So, people
put it as either e or ε there. So ΔGe is positive, γ is also positive, and whereas, ΔGv is
negative.
393
For example, for this particular alloy(C1, (Refer Slide Time: 13:54)), if I draw G versus
temperature diagram; let us say if I draw. I can say, this is the α phase, and this is the θ
phase, let us say. If I say, in principle, θ is at low temperatures, let us say. So in
principle, so this is the θ, this is the α if I say, below a certain temperature (intersection
point) there is a driving force from α to θ. For every alloy, there is a particular
temperature, that temperature is what we call it as? What do we call that temperature?
For every alloy, there is a particular temperature, below which the solid solution, wants
to give a precipitate, wants to precipitate out another solid phase. What do you call that
temperature? It is called solvus temperature as simple as that.
So this (intersection point) is the solvus temperature for that 4%Cu, if I am plotting G
versus T for both α and θ, that particular temperature is the solvus temperature, which is
this temperature (T1, Refer Slide Time: 13:54), in principle.
And if I write, what is called free energy composition diagram, which is G versus X
diagram. And we have again two phases: α and θ. So, you have an α phase, free energy
of α phase, and you have a θ phase and θ being an intermetallic compound, it’s more
narrower, it’s free energy composition diagram, and it is suppose to be somewhere, at
33.
394
So, somewhere you can put a θ phase, and if I draw a common tangent between the two.
So, this is one composition (point P) this is another composition (point Q). What does
this composition (point P) correspond to?
Student: Composition of α in equilibrium with θ
Composition of α in equilibrium with θ (C αθ), where is it on the phase diagram? Which
point?
Student: It is Solvus composition
It is the solvus composition. It is nothing but, one dot on the solvus line, at any
temperature at which I am drawing this.
Before you ask me, what exactly that composition is? You should ask me what
temperature you are talking about? Depending on it, if I am talking about this particular
temperature, (T1, Refer slide time 13:54) this point P I above G-x diagram is nothing,
but the point C1’ in the phase diagram (Refer slide time 13:54). So, these compositions
that you get at various temperatures which is nothing but the equilibrium between α and
θ are nothing but this.
In principle, if I can draw α and θ free energy composition diagrams at various

temperatures, and get these compositions, this composition is not going to change,
because this is more or less fixed, because it is like a line compound, its composition is
fixed. Though it is not really a line compound, it is very narrow range, composition
range it has, I have just expanded for the sake of only our easy understanding, otherwise
it is not a large composition range it is possibly less than about 1%. So, that is the kind
of range composition range that we talk about .
So, θ composition does not shift, but α composition is going to shift from here (solvus
composition at room temperature) at room temperature to here (solvus composition,
point Q, refer slide time 11:39) at eutectic temperature. And once I reach this eutectic
temperature, you will have to even talk about liquid also until that temperature is
reached. Where is liquid now, in our free energy composition diagram? It is above α and
θ, somewhere here. You have liquid also, but that liquid composition is somewhere up
like this (see the G-x diagram above), liquid free energy curve is, it is above the two.
So, the stable state is α + θ, and the moment you reach this eutectic temperature,
theliquid free energy curve comes and touches the common tangent between the α and θ.
So that all three of them will have the same common tangent, and this is what is a
395
eutectic temperature. Until you reach that eutectic temperature, you will have liquid
curve always above that of the common tangent
396
between α and θ, and the common tangent composition at various temperatures is
nothing but this composition (Refer Slide Time: 24:52).
And, if you draw, what is called the locus of all this compositions, what you generate is
the solvus line. What are the other methods of finding out solvus line? How do you find
out solvus line?
Student: By XRD
XRD, I hope all of you have done it, otherwise some of you possibly are doing this XRD
course, there is one separate chapter in Cullity, please read it, which will talk about how
to find out solvus lines from that, so one of the most common method is XRD. Of
course, there are other methods too we can talk about, so that is how you talk about it.
Now, what we are talking about is not this liquid. Forget about this liquid, so we are now
talking about an alloy which is, having a composition between Cαθ and Cθα let us say at
temperature as T1 (same as shown in phase diagram , Refer Slide Time 13:54)
And at T1, if I extrapolate the tie line (in the phase diagram, Refer Slide Time 13:54),
wherever this tie line intersects θ phase boundary, that particular composition; is Cθα
(indicated on G-x diagram above) and this composition (C1’ Refer Slide Time 13:54),
and my alloy composition (C1) is somewhere between these two. Am I right? So, that
means, I am talking about a composition like this, a composition between C αθ and Cθα .
So, this particular
397
composition, the moment I look at it, this composition, if it exists as a single-phase α at
that temperature, what is its free energy?
How do I know? If this alloy C1 exists as a single-phase α at that temperature T 1, what

would be its free energy? How do you get that? Can I get it from the free energy
composition diagram?
Student: Intersection point on the α curve
The intersection point of this vertical line C1 on the α curve will give you the free energy
of α, having that composition and what is that point? That point is this (point L in the G-
X diagram above) that is the free energy of α of whatever composition, 4 % Cu, I will
call it as C1 composition, let us say, for your benefit. So, we call this as C1 composition
and so that is the free energy of α. I can call it as GC1α.
And now this, if you carefully look at it, that value G C1α is higher than the free energy of
the two-phase mixture, which is given by this point (M, in above G-X diagram).What is
that? The line joining the common tangent line, which is the straight line, it always gives
you at any composition between the two it gives you what is the free energy of the two-
phase mixture.
The free energy of the α is this (point P), free energy of the θ is that (point Q), and a line
joining the two is something like a rule of mixtures. This is what you do in composites
for example, I know the strength of the matrix, I know the strength of the precipitate or
the second phase and I want to know the strength of the composite, I simply draw a line
joining the two, and call it as a rule of mixtures, and find out.
Similarly, we can use the rule of mixtures and this point (M) on the common tangent at
any given composition will tell you what will be the free energy of that alloy containing
two-phase mixture.. So, this (point M) is the two-phase mixture free energy, so this value
is lower than that value (L). So that means there is a driving force for the α to precipitate
α + θ. Am I right? So, what is that driving force? That driving force is LM.
This is what I can call it as ΔG, and the same ΔG , I showed you just a while ago in the
form of G versus T diagram. So, in a G-X diagram, you can see if this is the alloy (C1)
and you can see that if I increase the alloy composition, take it more and more towards θ
composition, what you will see? You will see larger and larger driving force and that is
what you see here (in the phase diagram, Refer Slide Time 13:54), that if I
398
come to, if I keep on increasing the alloy composition, I would see more and more
driving force or for the same alloy if I change the temperature, instead of this
temperature T1, I come to lower temperature. The lower the temperature, the more the θ
is stable.
You would see that this point what is called the Cαθ, shifts to the left as you can see in the
phase diagram. Shifts to the left to such an extent that this difference will keep on
increasing, the difference between your alloy composition and the composition of α; if
that increases, then you will see the driving force will increase.
So, you have a driving force and which is what is ΔGv term. This is all the driving force
that you have. So, at any given temperature, I can talk about what is the driving force.
And but though you have a driving force, you have a barrier. And the barrier in this case
is the surface that has to be created, the new surface that has to be created and the strain
that the phase the new phase that is coming out. Depending on which phase that is
coming out, you have different-different strain, depending on what is the structural
difference between the parent phase and the product phase; the strain is going to depend
on.
And that is the reason why actually the choices of various structures come into picture.
That is the reason why, why the system chooses the triple junctions, and the grain
boundaries for the nucleation is because the triple junctions are the regions, which are in
a high-energy state. First and foremost, they are in a high-energy state, as a result the
system can easily relax to a low energy state if the nucleation can occur there. Second is
that, these are the regions which have more random structure. As a result, the strain can
be easily accommodated much more easily, at that grain boundaries.
On the grain boundaries and triple junctions, because of the randomness that is available,
because of the open structure that is available, the strain can be easily accommodated.
And as a result, you will see most of the nucleation occurs at the grain boundaries or at
the triple junctions. In fact, when you go by the hierarchy of nucleation or hierarchy of
the energy, you will always see the triple junctions are the highest, then come the grain
boundaries, of course, the surface comes first.
399
The surface is the first, this is where the first nucleation will occur; surface is the first,
and then number two will be the triple junction, and number three will be the grain
boundaries. Grain boundaries are where two grains meets; triple junctions are where
three grains or more than three grains meet, so that is what it is.
So, that is how, these are the regions where you have a nucleation taking place, but the
problem that we have is, fine, nucleation is occurring but because of the nature of the θ
phase, once it nucleates, it grows along the grain boundaries. And once it grows along
the grain boundaries; it forms a thin network at the grain boundaries; it is like if you
imagine the grain to be something like spherical you would see a kind of a layer forming
on the sphere.
And once this happens you will see because this is an intermetallic phase, which is a
brittle phase, any crack that is nucleated at this place can easily grow through the
intermetallic phase. And as a result, what you get is an intergranular cracking. And that
makes the alloy very brittle, and this is what we do not want, and that is the reason why
in high carbon steels, you see the presence of cementite at the grain boundaries of the
austenite, makes the high carbon steels very brittle. And that is the reason why we do
certain treatment.
400
What is the treatment we do?
Student: Spheroidization
Spheroidization we do or even we do another treatment. Other than spheroidization, any
other treatment which does not take as long as spheroidization, high carbon steels, plain
carbon high carbon steels, what is the treatment that we do to take care of this problem
of continuous networking of cementite, at the grain boundaries?
Student: Normalizing
Normalizing, as simple as normalizing, you simply heat the steel to the austenizing
temperature, cool it rapidly.
So, that there is no time available, for all the cementite, which can come under the
equilibrium conditions, which can come according to the lever rule. You do not have
sufficient time for all the cementite to come and it nucleates, and there is not sufficient
time for this nucleated cementite to grow and cover the whole grain boundaries. So, what
you get is a discontinuous network of cementite rather than the continuous network.
And that discontinuous network, which is what, looks something like this (as shown
above). You have discontinuous network of cementite, something like this. Once you
have that kind of thing, there is no continuity of the brittle phase at the grain boundaries.
So, intergranular fracture can be controlled. So this is one way, we can control that. The
other way to do is, if I can somehow prevent the precipitation at the grain boundary, at
all. That is the whole crux of precipitation hardening.
401
I somehow try to avoid precipitating at the grain boundaries. How do I avoid this? The
only way I can avoid is, allow precipitation to occur not near the solvus line, but at
402
temperatures much below the solvus line.
For example, if the precipitation is occurring at this temperature, let us say instead of at
this temperature. Let us say I call this as a T2. Instead of T1, the precipitation is
occurring at T2. What is the difference between the two? Ultimately phase diagram
shows only θ phase. θ has to nucleate, but the θ nucleation at the grain boundaries, at
temperature T2 is not easy. Why? Because, if you want nucleation at the grain boundary
to happen, the Cu atoms because you remember what is the α phase, α phase has 4 % Cu.
What is θ phase? θ phase, if I talk in term of weight % it is approximately 55 % or so. Is
it not?
So basically, if you want to convert it into atomic %, it is 63.5/(2*27+ 63.5) = 63.5/120 ~

55%; Whatever it is, approximately 55 that is what I said. That is the composition.
So, that means, you need Cu atoms to move from the centre of the grains, to the grain
boundary, such that the Cu gets enriched at the grain boundary to the level of 55 %, so
that the θ can nucleate. That means, from the centre of grains you need the carbon Cu
atoms and Cu atoms are substitutional elements; they are not
403
interstitial elements. So, their diffusion rates are not very high and you are talking of
temperature, which is very low, and because of this, you see that the diffusion of Cu
atoms to that particular grain boundaries is not, because we are talking of grain sizes of
the order of let us say a few microns.
Most of the alloys, whenever you do solutionizing, and cool back, they are of the order
of 100 microns. So, if you think 100 microns is a grain size, 50 microns is the diffusion
distance; 50 microns is very-very large diffusion distance for the atomic diffusion. And
If that is the case then you see that at this temperature T1, there is no problem, because
the temperature is high enough, where the diffusion rates are high enough.
Whereas, at this temperature T2, it is difficult, and because it is difficult for the atoms to
diffuse up to the grain boundary to nucleate at the grain boundary, but at the same time,
there is a driving force. α does not want to remain as α. For example, if I take this alloy
C1, heat it to this temperature above solvus (indicated by arrow), make it a single-phase
solid solution, bring that single-phase solid solution to this temperature T2 and I am
holding let us say.
I under cool it; I quench this alloy to this temperature T2 and hold it there, like what we
talked about under cooling during liquid to solid transformation. Similarly, if I under
cool that α to this temperature T2 and now precipitation has to happen, at that
temperature there is a large driving force, for the precipitation. According to G-x
diagram (Refer slide time 25:31), there is a driving force and this is at T1, if I draw it at
T2, the same free energy composition diagram, you would see the driving force is still
larger. There is driving force, so α cannot remain as α, but the problem is there the
diffusion is slow.
Then the only solution for the alloy is that precipitation should occur inside the grain.
That is the only solution and if precipitation has to occur inside the grain, then you need
to consider γ and ΔGe very carefully.
Strain energy becomes very important; surface energy also becomes very important. If
the precipitation has to occur inside the grain, a new interface has to be created. At the
grain boundary you can say the grain boundaries are already high-energy regions, so,
nucleation can occur, so this energy γcan be reduced, but, if it is within the grain, the
nucleation has to occur somewhere inside. Imagine then,
404
you need to create a new surface. That is where the system will choose that particular
structure, which has a lower interfacial energy and lower strain energy.
And that is where; you would see that instead of α giving you θ, where the θ crystal
structure is orthorhombic, am I right? The θ is orthorhombic and θ has to come out of α,
which is a FCC. From an FCC an orthorhombic structure should come out and that is not
easy.
So, the system will choose and that is not easy, because the interfacial energy for such a
transformation between α to θ (γα-θ); interfacial energy is very large. So for that
interfacial energy being very large, the system will try to choose, is there any alternative
for me, or there any possible transient phases, which can have a lower interfacial energy.
So, it starts looking for transient phases, which are metastable phases, of course, but
looks for those, which have a lower interfacial energy, and that is where you come across
something called GP zones. So, the α precipitates, what are called GP zones, inside the
grains at different-different regions and the only heterogeneous nucleating sites are
what?
Student: Dislocations
Yes, dislocations, anything else?
405
I was talking to you earlier on the hierarchy of nucleation, I said - surface first, second is
the triple junction, then the grain boundary then, then comes what you have just now
said, stacking faults. After that comes, your, what is the previous answer that you gave?
Dislocations, after that comes what?
Student: Vacancies
Vacancies, perfect; these are all the hierarchy of nucleation and the surface energies
actually decrease in this order.
And that is why, in fact, if your sample has a lot of quenched-in vacancies; vacancies can
themselves act as nucleating sites, for such nucleation to occur or dislocations can be
there, or staking faults can be nucleating sites. So, this is how you see a GP zones come
out. The reason GP zones come out is GP zones are nothing, but, Cu rich clusters. They
are Cu rich clusters and Cu rich clusters have, Cu has what structure? FCC.
So, that means, inside an FCC matrix another FCC phase is nucleating and when an FCC
nucleates inside another FCC, obviously, the surface energy is very low. The only
problem now, is the strain energy. Why? Because, the lattice parameter of Cu is different
from the lattice parameter of Al, there is definitely some strain. So, it will choose that
particular shape, which has the lowest strain energy; so shape determines to a large
extent the strain.
406
I do not know, whether you have come across. For example, if I look at different shapes,
if I take a spherical shape, if I take a plate like shape, if I take a needle like shape. Either
it could be square or circular cross section.
If you look at these three shapes, each of them have different strain energies. Strain
energy basically is what is the compatibility of the precipitate with respect to the matrix,
that what determines the, if whenever you have a spherical shape, spherical precipitate
sitting inside a matrix, the strain energy is always the maximum. Because the unit cells
of the matrix cannot have compatibility, because they cannot have continuity. For
example, whenever you look at a matrix, inside which you have a precipitate, if the
lattice planes inside the precipitate and those inside the matrix; if there is continuity, then
we will see the strain energy is the minimum.
When do you have this? You have this, when the lattice parameter of the precipitate and
the lattice parameter of the matrix are exactly the same. Whenever you see the lattice
parameter of the matrix and the precipitate are exactly, the same, and this is what we call
it as misfit. It is called the misfit, and this is given as Misfit = [|aP – aM|/aM]*100
407
And obviously, we look at the absolute value of it. It can be negative, it can be positive,
we look at the absolute value and when this is very-very small, you see that there is
continuity. And even if it is a sphere at that, and whenever you have a misfit very small,
strain energy is very small, and when the strain energy is very small, the system will
choose that particular shape, which will reduce the surface energy. What is the shape that
will have the lowest surface energy?
Student: Sphere
Sphere, that is the reason why in some of the systems, you get spherical precipitates. Are
you aware of any system where you have spherical precipitates?
Any solid-state transformations where you get spherical precipitates? γ-γ’; Ni 3Al, when
it nucleates inside the nickel base matrix, you always see spherical precipitates initially,
and when they grow, they become cuboidal. Initially they are. Similarly, Al-lithium
phase diagram, if you look at it, Al3Li, precipitates inside the Al, again spherical, Al-
silver phase diagram, you have a phase called Al 2Ag, again spherical. It is all because
there the strain energy is minimum, so it chooses that particular shape which can
minimize that. If the strain energy is large, then what it chooses is, it chooses that
particular shape, which has the minimum strain energy, by choosing the shape and that
particular shape is.
If you plot the strain energy ΔGε as the function of shape aspect ratio, if you plot you
will see that, you get something like this. This is where the aspect ratio is 1 (at sphere);
408
this is aspect ratio very large, and this is aspect ratio is very small (left end of curve).
This is what we called it as disk, this is what is sphere, and this is what is called a rod
(left, centre and right respectively). Aspect ratio means the length verses the width of it.
If I look at it, sphere obviously will have one, and disk, the length the height is very
small whereas, the length is very large. So, it is less than one.
So, you will always see wherever the strain energy is more, the system will choose a disk
shape precipitate and that is what GP zones are. All the precipitates in Al-Cu system, the
θ double prime, θ prime; all of them are disk shaped or plate like. So, plate can have a
best kind of compatibility, when compared to either a sphere or a rod. So that is the
reason why you see this happening, so the shapes of the precipitates depend on these
two, because ultimately the term (4πr2γ + (4/3)πr3ΔGe) has to be minimized.
And if you can minimize this, the driving force ΔGv that is available, can be utilized for the
precipitation. As long as, (4πr2γ + (4/3)πr3ΔGe) > ΔGv, you can never have precipitation.
And that is how you see the GP zones come out first in the Al Cu, and these GP zones as
they grow, because the final equilibrium is θ, and GP zones are nothing but Cu rich, Cu rich
precipitates are on the right end of the phase diagram. We are only up to 55 atomic % here in
the phase diagram drawn. We have to go up to 100 % to get to Cu, and this Cu rich
precipitates slowly change their composition and their structure towards the structure of the
θ in steps, and that is how you see the GP zones.
409
The super saturated solid solution becomes

SSSS GP zones θ”  θ’  θ. We will stop here and take up later.
410
Lecture No. 18
Precipitation
We have been talking about solid state transformations, and we started with
Precipitation. And we looked at the basics of precipitation, where is the driving force
coming from.
And also we started looking at where is the barrier for precipitation, we said there are
two barriers, one is the surface energy, and two is the strain energy. These are the two
barriers, when we are talking of solid state transformation, any solid state transformation
in principle will have both these barriers, and precipitation is just one of them.
And when you are talking of liquid to solid, this part (strain energy) will not be there,
this we have already seen it, and only the surface energy is what we talk there. Now
when we are looking at both these things, the system wants to somehow minimize the
overall energy,
411
that it has to spend; a summation of these two energies, is the energy that it needs to
spend for creating a precipitate.
If that is the case, the system would like to choose that particular precipitate, which has a
minimum combination of strain and surface energy; this is what actually we were
looking at in the last class. And that is what leads you to the formation of what are called
transient phases for example, in a number of systems, the final product may have a
structure which is entirely different from that of the parent phase, particularly let say an
example of Al Cu itself.
The θ phase, which is the orthorhombic, is entirely different from its structure there is no
real similarity, in terms of any of the atomic planes. If you look at any of the atomic
planes in orthorhombic structure, there is no real compatibility with any of the planes in
the FCC type of structure, which is Al is FCC, so the matrix is FCC. So, in such a case,
when you have such a phase trying to precipitate out inside the grain, all the problem
comes only because the precipitation has to occur inside the grain.
Why it has to occur inside the grain, because the temperature is low, our precipitation is
happening at a temperature which is very low for the diffusion to happen, so that the
atoms move from the centre of the grain to the grain boundary and nucleation to occur
there. As long as nucleation is occurring at the grain boundary, there is absolutely no
problem. Any kind of precipitate, whether it is compatible with the matrix or not
compatible with the matrix, can precipitate at the grain boundaries.
Because, grain boundaries are the regions where you can accommodate, strains can be
accommodated and already grain boundary has a higher surface energy, so with respect
to that this surface energy can be accommodated. So, because we are now talking about
nucleation within the grain, somewhere inside the grain that is where we have to worry
about these energies, this is what we started looking at in the last class. And we said
basically it looks first of all, what is the misfit between the two, and also the system also
looks for , is there any possibility of a phase, which can have very low interfacial energy
with the matrix?
412
And such phases are what we call them as coherent interfaces, a coherent interface is that
where, there is a continuity of lattice planes across the interface. For example, if you
think of a matrix and inside that you have a precipitate (see the image above), and
precipitate let us assume, that the precipitate has the same structure as that of the matrix.
And it has certain atomic planes, if I draw certain atomic planes here, and then also the
matrix also has certain atomic planes; and these are all one single grain let us assume,
that inside one single grain a precipitate has form and this precipitate is trying to have a
compatibility with the matrix.
And when is this compatibility possible, when the atomic planes that are there in the
matrix, have similar interplanar distance with respect to the precipitate, then what you
see is, there is a continuity of the atomic planes. Whenever there is a continuity of
atomic plane, any plane it can be a (1 1 1) plane, (1 1 0) plane, some plane, if it can have
a continuity from the precipitate to the matrix, then we call it as a coherent interface.
And for this to exactly have a matching what you need, you need to see that the misfit
should be actually 0 % that means what, the lattice parameter of the precipitate should be
equal to the lattice parameter of the matrix, a P = aM. Because, interplanar distances are
related to ‘a’, all of you know in cubic what is a relation between d and a in cubic?
Prerana, do you remember, what is the relation between d and a? ‘d’ which is the
interplanar distance, ‘a’ which is a lattice parameter in cubic.
Student: a/ 2.
413
a/2.
414
Student: a/√(h2+k2+l2)
Correct, so we know that d = a/√(h2+k2+l2), this is only for cubic. So, if that is the case I
know that if a of matrix and precipitate are the same, then the d for any plane whether, it
is (1 1 0) or (1 0 0) or (1 1 1) for them you would also see the d also will be the same;
and if the d is same the interplanar distance being the same, then you would see there is a
perfect matching.
And this is a very ideal situation, very difficult to really see and there are very few cases
where such a thing happens, one such case is γ-γ’, where γ is the matrix which is the
nickel base super alloy, and γ’, what is γ’, Ni 3Al. And there also another situation
something like Al3Li, which is called δ’, and in an Al matrix. There are a few such
examples; Al2Ag is another example, Al Ag phase diagram.
So, whenever you have such a situation, you will see that there is a matching, in case the
lattice parameter of the matrix is different from that of the precipitate, what should
happen? You will see some kind of, for example, let us assuming that, the precipitate is
like this.
This is the d spacing for a particular plane (see the spacing between precipitate planes)
and, for the same plane here, you would see that, what I would like to say is that this
spacing is slightly larger (see the spacing between matrix planes) than this spacing by
about
415
a few % let say. If the d spacing is slightly larger what happens is, there will be some
kind of a bending at the interface of the planes, in order to maintain the continuity the
atomic planes will have a kind of a strain that comes in because, there is a bending of the
atomic plane at the interface and that causes what is called coherency strains. In order to
maintain the coherency, the system tends to have some kind of a strain, and this strain
can be accommodated up to a certain value.
We talked about sometime back whenever, you put a solute into a solvent also a strain
can be accommodated up to certain extent beyond that, the solute is rejected, similarly
up to certain extent it maintains. And if let say the distance between the d spacing here
(precipitate), and the atomics planes here (in the matrix) and the distance between the
atomic planes becomes much larger difference (Refer slide time: 10:44), then what
happens.
Then you would see for example, let say this is the situation, that the d spacing in the
matrix is much larger than the d spacing there in the precipitate, then what happens is, in
a unit distance the number of atomic planes any particular plane that we are talking
about, is different for the matrix and the precipitate. So, that means, across the interface
if I look at it, here below the interface you have certain number of planes in a unit
distance of x, and above that you have a less number of planes. So, as a result you do not
have one to one matching for every plane which is there in the precipitate, you do not
have a matching plane in the matrix.
Then what happens, what happens is some of the planes you will have a continuity, and
some of the planes are left out and this is exactly similar to something like a dislocation.
What is a dislocation? You say you have an extra plane of atoms is below, you have a
extra plane of atoms above a particular plane. So, exactly you see, that means you are
creating dislocations, these are called interfacial dislocation.
416
And these are also called misfit dislocations, because the dislocations are coming
because of the misfit, and that is why they are called misfit dislocations, and the moment
you generate misfit dislocations, we cannot call this anymore as a coherent interface. We
call now this as a semi coherent interface, partially coherent and if this keeps on
increasing, then what you see is that the number of dislocations will keep on increasing.
If the misfit keeps on increasing that means, for any given number of atomic planes in
the precipitate, the number of the atomic planes on the matrix is continuously keeps on
decreasing let say, because the spacing is increasing, one of them is increasing with
respect to the other, we are only looking at the relative change. So, if that is the case, you
will see that as this (d-spacing in matrix) increases more and more with respect to this
(d-spacing in precipitate), you will see more and more planes within the precipitate are
left out on the other side, you do not have a matching planes.
So, that means, you have more and more dislocations generated, and every dislocations
is associated with certain energy that means, the energy of the interface is increasing,
every dislocation has an energy. And that is the reason why we actually considered
dislocations also as one of the heterogeneous nucleating sites. Why? Because if you can
have nucleation there the energy can be reduced, because
417
dislocation has an energy, where is this energy coming from for the dislocation, because
there is a distortion above and below the plane.
If you have atoms like this let us say, you see that there is a strain associated there (see
image above), and this strain is what causes the energy of the dislocation and so that is
what is going to increase as more and more dislocations come. And at some stage the
overall energy is so high, that the system prefers not to have any coherency, and that is
what we call it as Incoherent precipitates. Usually you get incoherent precipitates, when
the structure of the precipitate is entirely different from that of the matrix.
As long as the structure is the same, the system the only difference now is only lattice
parameter; the system tries to have either coherent or semi coherent, you get incoherent,
when the misfit which we call it as δ, δ crosses about 15 - 20 %, assuming that the
structure is the same. The moment structure is different, then it is not possible to have
any coherency, so it automatically irrespective of what is the lattice parameter of the
precipitate and the matrix, you get a incoherent precipitate automatically.
The moment structure is same, then we look at δ and usually if the δ is less than about
5%, you get completely coherent interfaces, and γ-γ’ such a cases, are all something like
that, where you have the δ very small, and such a case you get this. So, this is in general,
this has nothing to do with thermodynamics, but
418
coming because of the strain, and because strain energy is also part of one of the thermo
dynamic parameter, we need to discuss this.
So, this is also going to decide, what is the type of precipitate that is going to come out?
Now what we are talking is we are assuming that the coherent precipitate is coming out,
and what is the shape of this coherent precipitate, this is what we were talking in the last
class. What should be the shape of the coherent precipitate, assuming that it is coherent,
if we think that it is coherent that means, strain is not too large.
If the strain is not to large, then the strain energy is too small that means, ΔGϵ is small, if
δ is small. Once that is the case, then what decides is the surface energy, once for a given
precipitate and matrix if strain energy is small the system will choose that particular
shape, which has the lowest surface energy. And coherent interfaces always will have the
lowest surface energy, because there is continuity it is as if like, you know if I give you
an analogy you, it is like low angle grain boundaries with respect to high angle grain
boundaries.
A low angle grain boundary has a lower energy, because the extent of misorientation
between the two grains is very small, and the more the misorientation, the higher the energy
of the grain boundary, exactly similarly because, there is a continuity of the planes at the
interface, coherent precipitates always have the lowest energy. And incoherent will have the
highest, and the semi coherent will have somewhere in the middle.
419
So, as long as you say a coherent precipitate is forming, it will always have a lower
surface energy. So we are saying the strain energy is already low, because the δ is small,
then what should be the shape, that is where you will see the shape automatically turns
out to be a sphere. Because, you know whenever we are talking of overall energy you
have to remember, it is VΔGϵ + Aγ, that is what has to be minimize, this is the overall
thing that has to be minimize.
And once ΔGϵ is small and if γ is also small, it would try to have a smaller area, and that
is what is the sphere, sphere has the smallest surface area. So, once it chooses a spherical
particle you can see this overall can be minimized, and that is how you see spherical
precipitates come; only when the δ is very small. When the δ is large and even you will
see that in some of the cases where, where the spherical precipitates come, when they
grow to bigger sizes, you would not see this spherical precipitate maintaining their
spherical nature, in a liquid to solid there is a possibility that if you have spherical
particles forming from the liquid to solid.
They can still grow as spheres, there is no strain associated there because liquid
everything can be accommodated, all the strain can be accommodated inside the liquid,
whereas if you have a spherical precipitate which is growing inside a solid, then in order
to maintain the coherency becomes very difficult for a spherical precipitate. If you think
of a curved interface, a curved interface if you have atomic planes, curved interface to
maintain the coherency on a curved surface is not easy, maintaining a coherency on a flat
surface is easy; as a result what happens is, as it grows to bigger and bigger, this
coherency becomes more difficult.
So, it slowly changes into shapes which are like this (square drawn in above image), that
is why γ’ are cuboids. When they grow they become cuboids, at the nucleation stage they
are all spheres, if you look at them at the initial stages of precipitation, they are all
spheres and then they become cuboids. These are all in cases where the misfit is small, if
the misfit is large that is what yesterday we have seen, if the misfit is large then the
system, anyway the strain energy with a sphere is always higher, that is what we have
seen, the highest is the with the sphere.
420
So, sphere is avoided and then the other two possibilities are one is a plate like
precipitate, and another could be a rod like precipitate, and you see that a plate like
precipitate has a better compatibility and as a result in a most of the cases, you see plate
like precipitates. And that is what you see in most of the cases, that plate like precipitates
come because, the δ is larger. So, this is all more of somehow minimizing the overall,
but this is all at the starting only we are talking of a starting, and where you have only
come to an intermediate stage.
Why intermediate stage, because if you start with a super saturated solid solution for it to
give you α super saturated solid solution, to give you α + θ though there is a large
driving force, the activation barrier is also large (see the activation hill plotted in above
image). So, what the system does is system chooses that particular structure, which has a
lower activation barrier that means lower combination of these two (strain and surface
energy). What is the activation barrier it is this, which is the activation barrier,
combination of these two is the activation barrier.
And that it tries to minimize that, and then gives you another precipitate with a lower
activation barrier, but at the same time you see the energy of this metastable phase will
be higher than the energy of this, because that is after all a metastable phase. So, a
metastable phase comes out of this, and there is definitely a driving force for this two
come, because this has a lower energy than this. So, from a super saturated solid solution
a another phase which has a
421
lower energy than this comes out, and that is what in our case is what we call it as GP
zones; α + GP zones, and that goes through next stage and then the next stage and then
finally you reach this level (α+θ) So, it goes to θ” to θ’ and finally to θ. As the
precipitate starts growing to bigger and bigger sizes, it slowly changes from the original
state of super saturated solid solution to the final equilibrium shape, which is the θ in a
number of steps.
Because, at each of the steps if it wants if GP zones want to jump to θ, again the
activation barrier will be large. So, it tries to see that the same thing happens in steps,
this is what people call it as Ostwald step rule; Which basically talks about whenever an
activation barrier for a particular transformation is very large, the system tries to achieve
that final equilibrium stage in a number of steps, where at each of the step the activation
barrier is much smaller than the original activation barrier.
And that is why you see the free energy shifts from here to here to here to here (GP
zones to θ) finally, and the number of steps that you need to go through, basically it
depends on the extent of difference in the crystal structure between the parent and the
product. In some cases it can happen it directly in one step that is what γ-γ’ case is,
directly γ’ comes out, you do not have any steps. In case of cementite case you see you
have one step which is the ϵ carbide from the martensite you get first the ϵ carbide, and ϵ
carbide directly gives you the cementite.
So, like that you can see depending on the type of system, depending on the crystal
structure complication you can have a various number of I mean Al-Cu is one classic
example of a large number of steps. Otherwise, there are many systems where it can
happen in directly one step.
And now, if we try to look at this from the thermo dynamic angle
Student: Is the sum of the activation energies should be less than the total activation
energy?
If you look at, see the point is, ya interesting question, but the point is once you have
come to this stage (GP zones), it does not bother what is the activation energy here, now
you are starting state is this, the past is gone. So, now, you have to look at from here (GP
zones) if I want to come to here (θ) what is the activation energy, can I now reduce it by
going to some other
422
stage, where the activation energy can be lower. So, it is not that each of the individual
activation energies if I calculate and add it up will it be lower than this, it can be lower, it
can be higher, but that is not the question here so but, for each step definitely the
particular activation energy will be smaller than it going directly to θ, that is what is an
important, is not it. So, so final stage what you are coming is not really important, at
each of the step it tries to look at how do I minimize the activation energy.
Now, let us look at the Al-Cu, we said right now that α super saturated solid solution
goes to GP zones first, and then θ”, then θ’, then finally θ, this is what we said. And GP
zones and α have the same crystal structure, FCC, the only difference is, what is the
difference?
Student: Lattice parameter
Anything, where is this lattice parameter difference coming from?
Student: The percentage, radii of copper
The composition is different as simple as that, the GP zone composition is different from
α, α super saturated solid solution has 4 % Cu, and 96 % of Al where as GP zones are
possibly the reverse, we call them as Cu clusters. But, unfortunately there is no easy
technique to find out what is the exact composition of it, because the clusters are so
small, have you heard of what could be the technique
423
that you could use to find out the composition of such a cluster? Let us say, you have a
cluster of about you know 100 - 1000 atoms.
Electron
Student: Electron mass spectrometer
Electron mass spectrometer
Student: Atom Probe
Atom Probe, a large number of people use three-dimensional atom probe where you
basically can, you will see each atom. How many of you know about atom probe? Atom
probe is a technique which is basically a field-ion microscope, where you take a sample
in the form of a tip. And then basically sputter away the atoms, when you are sputtering
away the atoms, depending on the weight of the atom, the velocity of the atom is going
to be depend on, this is what we call time of flight measurements we call them.
So, you put a detector at a certain fixed distance, and try to see what is the once you
applied a pulsed voltage and you sputter atoms, you try to measure what is the time that
a particular atom has taken to reach here, once you have done the pulsing, then you start
your time t = 0. And then you see from time t = 0 when an atom is detected by the
detector, and from that particular time difference you can find out that calculate, and find
out this could be an element of this particular mass, if this is the mass of this element it
would take this much of time.
So, all this once you know that you can the moment each atom reaches there, you can
start assigning a particular mass number to it; and that is how if you do layer by layer
sputtering and keep on detecting atoms, and then store them. You can actually, when you
reconstruct the whole thing you will be able to see how the atoms are distributed inside
the tip, in a nano metric level one can see that and there you can actually see, if there are
any small clusters of Cu or of iron whatever it is.
And people regularly see this for example, carbon atoms going and sitting near the
dislocations, we call them as Cottrell atmosphere; this is the only technique which can
424
actually prove that. There is no way you can see by let us say TEM-EDX or any other
technique you will not be able to see, because the regions are so small, so this is the
wonderful technique where large number of people, in the whole country there is only
one place where we have it, DMRL has this. Incidentally I work for 2 years on that
instrument in Japan anyway, so that is a good technique where you can see that.
So, the point is if this (GP zone) is the Cu rich cluster, and this is a Al rich cluster ( α), let
us look at how is the free energy diagram is going to be. Let us draw the free energy
diagram, on one side you have Al, on the other side you have Cu, you start with an α, let
us say this is an α and somewhere here is the midpoint let us say, and so 50, 50. So, θ has
approximately around 55 % this is what we have said, so you have a θ phase somewhere
here, and let us put it slightly down (See the free energy curve above).
We are talking of at a temperature let us say somewhere here, may be we can even
consider room temperature why not, at this temperature I am talking of the precipitation,
I have taken the alloy to high temperature, quenched it why do I quench it, to get this
super saturation; if I do not quench it then what happens?
Student: θ will precipitate at grain boundary
425
Yes, θ will precipitate at the grain boundaries, if I do not quench it you will have, in fact
this is what people observe whenever you are doing age hardening, unless you really
quench it fast, you do not see this whole thing. Otherwise, during the cooling itself you
can have precipitation taking place, so then super saturated solid solution does not form.
So once the super saturated solid solution come, and at this temperature, if you look at
the free energy composition diagram, you will see this is the common tangent (P1Q1,
Refer slide time 31:21), where this composition (C αθ) is what, is this composition (point
P1, Refer slide time 30:36) .
Now we are talking about that this is all under equilibrium, but we are saying that
equilibrium is not happening, θ is not able to come out of α, then what is coming out of
α, it is what is called the GP zones. Which are nothing but, Cu rich and because, they are
also FCC, if I want to draw the free energy composition diagram for the GP zones, how
will I draw it?
It will be the same curve extended towards the pure Cu side, there will be no separate
free energy curve for the GP zones, because they have the same structure. Whenever
there are two phases with the same structure, they will have the same free energy curve;
this is what somewhere we have talked about it. So, you can see that this, this will look
something like that, pure Cu rich phase, so this is the GP zones free energy (Point Q2)
that is the case, now if I draw a common tangent between the two, then how will the
common
426
tangent look like (P2Q2), so this is a composition now (P2), what do I call that
composition now?
Student: Composition of α in equilibrium with GP zones
Correct, it is the composition of α in equilibrium with GP zones, what is this

composition (P1)?
Student: In equilibrium with θ
Yes, so now, let us say that this is the composition of α in equilibrium with θ (Cαθ), this
is the composition of α in equilibrium with GP zones (CαGPZ) and both are different, why
they are different because, the free energies of the two phases are different. And this is
similar to what we have done earlier, if you remember when we are talking of iron
carbon diagram, and iron cementite diagram, we did exactly something similar, and this,
this is the metastable phase (GPZ), this is the stable phase (θ).
The composition of any phase in equilibrium with metastable phase is always higher in
the solute content, when compared to the composition of the same phase in equilibrium
with an equilibrium phase; that is what exactly you see here. That means, at this
temperature if I now want to show, what is the composition of α in equilibrium with the
GP zones, where do I show on the phase diagram? It will be to the right of this point. So,
that means, somewhere here (point P2, Refer slide time 30:36) I do not know exactly
where it is, but that can be calculated once I have the free energy curves.
Now, if I say where is then the next phases, what are the next phases? θ“ and θ’, and θ“
and θ’ they will definitely have lower free energy than GP zones, but higher free energy
than the θ. So, as a result I can say that, and they will also have the same composition as
that of the θ, that is the reason why we still call them as θ“ and θ’; otherwise, we could
have given some other name, this difference is only structure. Structure does not still
become same as that of the orthorhombic.
So, the final structure evolves, so the structure slowly evolves from the FCC to the
orthorhombic. And now, we can draw common tangents for θ’(P3Q3) and θ’’(P4Q4),
which give the composition of α in equilibrium with θ’ (Cαθ’) and θ’’ (Cαθ”).
427
So, GP zones, θ”, θ’, θ, it will always be in that order, the composition of α in
equilibrium with GP zones will have the highest solute content.
The composition of α in equilibrium with θ will have the lowest and it goes on
increasing as we go towards θ.
Now, if I put all these points on the free energy (P1 to P4, Refer slide time: 31:21), on
the phase diagram (see the image above) you will see four different points, one is the
equilibrium composition of α in equilibrium with θ.
Student: Should the free energy of these phases be lower than α?
428
We are not talking of lower than α or higher than α. For example, if that is the case this
the way I have drawn (GP zones curve) is higher than α, what is important is not whether
that phase has higher; when I am taking an alloy of a particular composition, what
precipitates is more important.
In individual phases what is important is, you may be saying that, sir free energy of the
phase should be lower than α for that to precipitate, is not like that. For example, if I
have this kind of a free energy curve (α), the moment I am choosing an alloy of this
composition (point K, in image above), and what will come out for this alloy, when I
bring this alloy to that particular temperature room temperature, hold it there.
What comes out, depends on when I draw a tangent (KL) for that particular alloy, what
where does the tangent intersect (point L), that is what is going to decide. So, now, once
I draw a tangent for that particular composition, and you will see that this tangent goes
like this, and then the driving force is decided by this now (LM), not with respect to α.
Once I draw a tangent for the alloy composition, the α that I am starting with I am
starting with α, that α for that particular alloy, I draw a tangent.
And where does the tangent goes that is going to decide, how much driving force is there
for each phase to come out. And
429
depending on now this driving force, now which phase will come out, all four of them
have driving forces let us say, then what will come out is going to be decided by what is
the activation energy for that phase.
But, now if I choose another alloy composition because, you have raise that question let
us look at this, if I choose this particular alloy composition (N) and draw a common
tangent, I will come to a situation like this, that this particular alloy composition is such
that, it can have a driving force for θ, θ’, θ”, it cannot have a driving force for GP zones.
That means, if I choose that particular alloy, I will not be able to get GP zones, you are
going to see it within a minute when you look at this particular thing.
For example, now if I look at these four compositions, this is composition of α in

equilibrium with θ’ (Cαθ’), this is composition of α in equilibrium with θ” (Cαθ’’); this is
composition of α in equilibrium with GP zones (CαGPZ) (Please Refer Slide Time: 41:59).
Now, if I do this exercise at different-different temperatures, these are all being that at
one temperature, if I do it at different temperatures, what would be happen, at each of the
temperature I will have these common tangents, at each of the temperature I will be able
to generate these four points.
And then that will lead you to different curves, you will four different curves, what are
these curves?
430
They are all solvus lines, they are different solvus lines we call them as the θ solvus, θ’
solvus, θ” solvus, these are the GP zone solvus. And now if you look at these solvus, and
carefully observe what I have been just saying in answer to what Anisha asked, if I take
this alloy (point P) bring this alloy to a temperature below the solvus line for all four of
them, when does a precipitation takes place, whenever my alloy composition comes
below a solvus, then only precipitations happens.
For example, if I take this alloy, heat it to this temperature (point Q), keep it there will a
precipitation take place? No, why because when at that temperature, if I draw a
horizontal my alloy composition is having less solute content then what is the solubility
limit (point R), what is solvus line, solvus line tells you the solubility limit at each
temperature, am I right.
So, if that is the case the my alloy is containing less salute content, than what the solid
solution can contain, and if that is the case then there is no chance for the precipitation;
the amount of solute is smaller than what it can content, but the moment I bring this
alloy to this temperature let say (Q’), at that temperature again, if I draw a tie line. Now,
my alloy composition is higher than what the solute can contain (R’), and my alloy
having higher than what a solute can contain; obviously, it has to precipitate.
So, precipitation can happen only when the solute content of the alloy is greater than the
solubility limit and that when is it happening in terms of the temperature if I look at it; so
far we talked about composition in terms of temperature, when I bring it below the
solvus line. Now, if I am at this particular temperature for this alloy composition (point
P), my alloy at that temperature is below the solvus of all the four of them that means, in
principle all four of them can precipitate.
Instead if I choose another alloy of this composition (P’), instead of 4 % Cu, I choose 2
% Cu let say, if I choose 2 % Cu, then if I carefully look at it this particular composition
at this room temperature is below the solvus of θ and θ’, but it is above the solvus of θ”
and GP zones.
431
That means, the GP zones and θ” nucleate, from this alloy if it is brought to that
temperature, and that is what I was trying to explain. If I choose an alloy of that
composition draw a tangent, I will see that the GP zones and θ” will be above the
tangent; once they are above the tangent, there is no driving force for them to precipitate.
So, if I keep on shifting the composition towards the left side, you will see one by one
each of the phases start dropping off, and that is why if I choose an alloy of this
particular composition (P”), the only phase that can come out is θ; if I bring that alloy to
room temperature. And that is why we do not use such alloy compositions for age
hardening, you will never think of age hardening of a 1 % Cu alloy, because for such an
alloy the only precipitate that can come out is only θ, and that θ anyway you know that it
is not going to give you higher strengthening.
Because, it is basically an incoherent and again I should tell you because, we are talking
of strengthening, strengthening depends on all of you know about precipitation
strengthening. Whenever you have coherent precipitates, the strain energy associated
with a coherence coherent precipitate what is called coherency strain can add to the
strength, not only the particle. And whenever a particle gets sheared by dislocation, the
new surface that is generated also has to be a new energy has to be spent.
And as result the, that is why the smaller the energy of the precipitate that means, the
more coherent it is, the more difficult it is for the cutting. Because for cutting you are
creating an extra surface, the system is already in a lowest energy state, so it does not
want to create a new surface. So, the stronger are those alloys which have coherent
precipitates.
But, again there is a other issues coming into picture, coherent precipitates being very- very
small, they can also be sheared because, the size itself is very small, so the total area that you
are creating, when you are cutting it is not very large. So, you have to look at a weighing
both the cases that is why, usually when you do age hardening the peak aging comes when
you have a semi coherent interface, GP zones do not give you peak aging.
432
For example, if I draw hardness verses time, you will see this kind of a curve, all of you
are aware of it, this peak comes this is the super saturated solid solution (αSS), α super
saturated solid solution and GP zones are somewhere here, and θ” is somewhere here, θ’
is here and then finally, θ is here, something like that. So, somewhere between the θ”
and θ’ is where you get the peak aging, that is where you get a semi coherent interfaces,
so strengthening aspect you do not want θ.
Now, but at the same time you know, you will always remember that this value (HαSS)
and this value (Hθ), if you carefully observe Hθ > HαSS, because the strengthening here is
coming from what, in this case?
Student: Solid solution
It is only solid solutions strengthening, whereas here you have precipitates, and there the
precipitates are larger and the only strengthening that comes is by formation of the loops
Orowan loops; the by the bending of the dislocations not by the cutting of the
dislocations, so that is the another aspect. Let us not go in to the details of it, what is
important for us is to understand that depending on the alloy composition, and depending
also on the temperature at which I am doing the age hardening, now I am talking of this
alloy (point P) age hardening being done at room temperature.
433
Now, if I increase the age hardening temperature, why do I increase the age hardening
temperature? Diffusion is faster. So, the kinetics will be faster, the precipitation will
happen at a shorter time for example, if this is one of the age hardening curve.
And if this is at room temperature, if I am doing at let us say 100 ºC, how will the age
hardening curve look like?
Student: It happens at short time
434
Short time that means the peak is reached at a shorter time.
Now, can you tell me anything more?
Student: Some zones will disappear
No, I am not talking with respect to this simply I am increasing the temperature, let us
not bother about what phase is forming, the peak time peak time decreases and what
about peak hardness, peak hardness comes down, that is very interesting. So, if you look
at this, this is the kind of scenario you find, as the temperature increases in this direction
(indicated by arrow), the peak time comes down and the peak hardness comes down,
why does the peak hardness comes down, can somebody tell me?
Student: Coalescence
Coalescence is one question, something more, more fundamental which you can see
from the phase diagram, if you look at the phase diagram ignore all these solvus lines,
just think of only one θ solvus. And then tell me, what would happen if I keep on
increasing the temperature?
Student: Particle size increases
That’s you have said, more than that
435
Student: Solubility is more
Solubility is more, so
The amount of precipitates will be lower.
As I keep on going to higher and higher temperatures, the extent of super saturation is
decreasing, and it is this which tells you what is the if you use the lever rule, it which
tells you what is the amount of precipitate that is forming;. And the moment I reach
solvus line, it will be 0, so that is why the higher the temperature you would see the
lower the amount of phase; and that is the one and the size being higher both of them
will cause this, will stop now.
436
Module #01
Lecture #19
Precipitation – Quasicrystals
We will continue where we left in the last class on, “precipitation”. We were talking
about one of the solid-state phase transformations - precipitation. When you look at
precipitation, and talk about nucleation of a precipitate, we look at the free energy
composition diagram; let us say. We talked about an α phase, which is nucleating a
phase called; let us call it as β or θ, whatever it is. In an Al-Cu phase diagram, we call it
as θ let us say.
And if you look at the equilibrium between the two, this is the common tangent (PQ,
Left figure), which talks about equilibrium, and the corresponding phase-diagram is
somewhere here; let us say, and let us say we are talking about room temperature. At
room temperature, this particular composition (P’, right figure) that you are seeing is this
composition (P, left figure). This we know. So, whatever
437
composition that we are talking about here (P’), is this composition (P). And if you take
an alloy of a composition something like this (1, right figure) that is what is the whole
age hardening process that we talk about. So, the alloy is somewhere here (1, left figure);
is on the right side of the equilibrium composition, equilibrium composition of α in
equilibrium with θ, and that alloy at this temperature; room temperature, this is where we
are talking about, at this temperature, if it starts precipitating θ, the driving force for the
nucleation of θ, is always given by the line which is a tangent (R1S1), for the α at the
alloy composition.
At the alloy composition, if I draw a line, which is the tangent to α free energy curve,
and wherever, it gets extended, at the precipitate phase composition if you draw a
vertical line, this difference (S1Q) indicates the driving force for the nucleation. We call
it as ΔG for the nucleation (ΔGN) So, anytime you see an alloy, if you want to calculate,
what is the driving force for nucleation of a particular phase, any phase it could be; in
this particular case, I am just simply taking an example of θ. And If I am taking θ,
simply take that alloy composition, it could be 4% Cu, it could be 2% Cu, it could be 3%
Cu, whatever alloy composition it is and for that alloy composition draw a tangent.
And from the tangent position to the free energy of that particular phase at that
temperature, what you get is the driving force for the nucleation. And in such a case, if I
change my alloy composition if I choose another alloy composition here (R2), in
principle, if I draw another tangent here you would see that, that tangent would lead you
to a different driving force (S2Q) for the same phase. For the same θ to precipitate out of
α, alloy 1 and alloy 2 will have different driving force.
And the larger the composition of the alloy, with respect to the equilibrium composition,
the more away the alloy composition with respect to the equilibrium composition, the
greater will be the driving force. So, the ΔG N α ΔX0, we call it, where ΔX0 = X - X0.
What is X0? X0 is the equilibrium composition. This is very clearly seen here.
So, if I choose alloys of different-different compositions, that is the reason why a 2% Cu

alloy, and a 4% Cu alloy will have different driving forces at the same temperature. This,
I am taking the example of room temperature, I can change the
438
temperature. If I change the temperature also if I choose some other temperature, there
also you would see the same thing. And you can also see it in terms of the phase diagram
also. What tells you about the driving force is this (AB, in right figure). This is the actual
driving force I mean not in terms of ΔG, for that matter.
You are at a temperature, for example; if I am considering the same room temperature
alloy 1 and alloy 2. For alloy 1 you are at a temperature away from the equilibrium
transformation temperature. The moment you are below this equilibrium transformation
temperature (T1), there is a driving force for the second phase to precipitate. If I am at a
temperature above T1, obviously there is no driving force for the second phase to
precipitate. If you want second phase to precipitate, I have to come below this
temperature T1, which is what is called the solvus temperature.
So, the farther away, the temperature is, from the solvus temperature, the higher the
driving force, for the precipitation. So for any alloy composition, if I am at different-
different temperatures; so the lower the temperature is, the higher the driving force for
that alloy. Now, if I am choosing two different compositions and choosing the same
temperature, fixing the temperature but changing the alloy composition. You see by
changing the alloy composition, what is changing here, solvus temperature is changing.
Whenever I change the alloy composition, for example, alloy 2 has a different solvus
temperature,.
So, because the solvus temperature is changing, is increasing as I increase the solute
content and If I am fixing my transformation temperature, where the actual
transformation is going on, as a fixed temperature, now you see the difference between
the solvus temperature to the temperature at which the transformation is taking place is
larger in case of alloy 2 and is smaller in case of alloy 1 and that means again, you can
see that alloy 2 has a greater driving force for the precipitation when compared to alloy
1. So, you can see that by choosing an alloy of appropriate composition, and that is the
reason why people want to choose as larger solute content as possible, but that is limited
by something. You cannot choose any composition of your choice, limited by what?
Student: Maximum solubility
It is decided by what is the maximum solubility, of the solute in the alloy. That means if
I extend this phase diagram somewhere here (right figure), I mean, let us not bother
about what is on
439
the other side. Let us only look at this particular composition (point C, right figure). This
is what decides, what is the limit to which I can go. That is the maximum of, if I go
beyond this, then what I have, I have a situation where I cannot actually solutionize the
alloy.
I cannot take it into a single-phase solid-state, single-phase solid solutions condition.

Because unless I take the alloy to the single phase there is no, the age hardening does not
start. Age hardening process, the first step itself is to take it into a single phase, then
quench it, then you try to do the aging to precipitate the second phase.
So, in case you cannot get, for example, if I choose an alloy of this composition (3), this
is one of the typical case of a 6% Cu alloy; let us say. A commercial alloy-2219 alloy; if
I take a Al-2219 alloy with about 6%, 6.5% of Cu. 6.5 % of Cu, the moment I start
heating it, actually in principle as I keep on heating it. What should happen to this alloy,
this alloy? This has been solidified, brought to room temperature, now I am heating this
in a furnace to different-different temperatures. What should happen to this alloy (3)?
Student: θ has to dissolve
θ has to dissolve. Why θ has to dissolve? Because the solvus line, if you carefully
observe, the solubility of Cu in α Al, is increasing with increasing temperature that is
what this solvus line tells you.
So, because the solubility is increasing and at each temperature, if I calculate what is the
volume fraction or the weight fraction of the second phase, the weight fraction of the
second phase keeps on decreasing, if you use the lever rule. Because, what is decreasing
is, this horizontal length is decreasing. Is it not? And from that, you can feel that from
the lever rule the amount of θ is decreasing, and that amount of θ, where should it go. It
gets dissolved into the matrix, in such a way that the matrix composition gets richer in
Cu. Because the θ, which is a Cu rich phase, is dissolving into the matrix, the matrix
becomes enriched in Cu content and that is what you keep seeing, and until the eutectic
temperature, you would see that this will keep on happening.
And even at the moment you reach the eutectic temperature; you will still have some θ
and some α, at the temperature of eutectic, the moment you reach this eutectic
temperature, let us say you are just below the eutectic, melting has not yet started. If you
are just below the eutectic, then you will see the α
440
composition will be this (C) and you will still have some θ, which is decided by this
length of the line (C-3), divided by the total from that to the θ phase, wherever θ phase
is. That is how we calculate.
So, you can see this tells you that there is still some amount of θ that means complete
solutionizing has not happened. But still, you can take such an alloy, and in such an
alloy, where is the θ? θ is still at the grain boundaries. Though it has dissolved, some
amount of θ is still left out at the grain boundaries. Because during solidification,
solidification happens at a slow condition; so as a result θ is precipitating at the grain
boundaries only. So, you will see that the θ is there at the grain boundaries.
But the matrix if you look at it, matrix got enriched but still you have θ here (at grain
boundaries) and so this, If from that temperature let us say if I quench it, then what
should happen? So, α which is there here, will remain as super saturated, because I am
quenching it, but the θ which is there at the grain boundaries, will remain as θ;
depending on what is your alloy composition, some amount of θ will be there, according
to the phase diagram.
And this matrix α, which is now super saturated α which has formed at the room
temperature, when you quenched it, that particular super saturated α is the one, which
will start precipitating out. So, in such an alloy, you still have age hardening taking
place; you will still have precipitation taking place, but in addition to the
441
precipitate, you will also have the θ, which has been formed during solidification, at the
grain boundaries. So, you would still have some θ, which is at the grain boundaries. And
in fact this θ, how does this θ form? Can somebody tell me, during solidification? The θ
which you see, it forms from the eutectic reaction.
If you look at this alloy, and then extrapolate this phase diagram, you have θ somewhere
here (see phase diagram above). So, this alloy (4), if I am solidifying it how does it
solidify first? It is a hypoeutectic alloy, so what comes out first? α; α dendrites come out
first. So, you have the α dendrites and once you reach the eutectic temperature, it is the
eutectic mixture which comes out of the liquid. What is the eutectic mixture here? α + θ,
so at the dendritic boundaries, and these dendrites will get converted into grains later.
So, at the grain boundaries, where you have actually interdendritic regions and those regions
are the ones where you have the θ. So, θ will be present in the form of not as I precipitated.
So, if I look at this alloy (3), it will be a different alloy. This alloy (3) will be entirely
different from this alloy (4) in terms of what is the morphology of the θ. In this case, you see
that this alloy (3) solidifies completely as single-phase α, am I right? Solidification is only
single-phase α, there is no eutectic there, and this α, as I am cooling to room temperature; it
starts precipitating the θ and that θ comes out at the grain boundaries. Whereas, in this alloy
(4), the θ that comes out, the θ
442
does not come out by precipitation from the α, it comes out as a eutectic mixture. But
you can still ask me, Sir, “after the eutectic reaction, when I still cool it from the eutectic
temperature to room temperature, what is suppose to happen during solidification?”
After solidification is completed, but that solid is now cooling to room temperature,
should anything happen in that alloy?
Student: θ should increase
It should increase. Why it should increase? Because the solubility of Cu in α is
decreasing and where should that increased θ appear?
Student: It cannot nucleate inside.
It cannot nucleate inside. So, in principle it would possibly nucleate on the existing θ,
which is there during solidification. Whatever, because already θ is there in the eutectic
mixture, that θ gets precipitated over that, because it does not need further nucleating
sites.
So, you will usually see the microstructure, would look like the dendrites and in the
inter-dendritic regions, you will have the, whatever may be the morphology, in this
particular case, actually you see a lamellar morphology. So, you will see actually a
lamellae of, there can be boundaries, which are called eutectic colonies and you will see
lamellae where you have that. And after the solidification as you, further cool if you
want to look at where does this form, so that extra θ phase will start forming on the θ,
443
which is already there close to the dendrites. Because the excess Cu is coming out of the
dendrites actually, it is the dendrites, which are releasing all the Cu.
444
So as a result, nearer to these dendrites, you will see the precipitation happens. So, in
principle you can also see the θ nucleating either at the grain boundaries, which are the
dendrite boundaries or on the θ which is already there; one of the two.
Student: Wont α and θ be alternating?
That is what I am saying they are alternate layers, the white and the black, white and
black (on the image) what you see are the α, θ; α, θ; so that is what you are seeing there.
And on this, for example, if Cu that is coming out, will not go and precipitate on this θ
(in the lamellae), where it has to travel a farther distance. So, in principle if there is a θ,
which is already available here near the dendrites, that θ becomes thicker or the θ will
form as a layer on the existing α.
So, you will have two morphologies of θ, one θ which is already present in the eutectic
mixture and second the θ which is there at the dendrite boundaries if you want to call
them. Whereas, in this case (alloy 3), you have the α coming out as the grains here, after
the solidification and the θ precipitation is only there. And when I am heating this (alloy
4) back, now both the θs start dissolving, one the θ at the dendrite boundaries and the θ,
which is there within the eutectic mixture.
Both of them start dissolving and as a result, once you reach this eutectic temperature
you will have some θ that is left out. And now, when you quench it, this particular
because the θ has dissolved, these dendrites now become richer again in Cu, to the extent
of what it would have had, just after the eutectic reaction. Just after the eutectic reaction,
α had this composition (point C in Refer slide time 13:10). So, if you have heated this
alloy back to that eutectic temperature, you have more or less reached that particular
condition where just after the eutectic reaction, what was the existing α.
And that α when you quench it, now you will have that α, retaining all the Cu content
and when you start heating it, for the age hardening, that particular α precipitates inside
it. Some phase which is depending on what is the under cooling that we are talking
about, because in the last class we talked about whether GP zones comes out or θ” comes
out, or θ’ comes out or θ comes out, depends on what is the temperature at which we are
doing the age hardening, for a given alloy. That is what you have seen different-different
solvus lines, we have drawn and depending on that we can see that.
445
Student: When θ is already there, why it has to nucleate inside?
Because again this Cu which is there inside the dendrites, because we are doing age
hardening at this temperature, let us say. So, if it has to nucleate on the θ, which is there,
so it has to travel a larger, and these are all 100 microns or so, generally grain sizes. So,
it has to travel a long distance and at that temperature it is not easy for it to happen, and
that is the reason why, it avoids because the time available is shorter.
Otherwise, in principle, if you leave it forever, you will see slowly the diffusion can
happen and then some precipitation can happen at the grain boundaries also. But
otherwise, because the time available is shorter, and the temperature is very small and
the diffusion is difficult, and at the same time the α is super saturated, and it is not stable,
because we can see here (G-x, diagram, Refer Slide time 13:10) it is not stable. Single-
phase α is not stable. It wants to precipitate; it has a higher free energy than a two-phase
mixture.
And as a result, it wants to come out and bring out the precipitate, and that can happen
only by precipitating within the grain. So, this is how you will see, if you have a
composition of that sort. But if you are talking of compositions, which are to the left of it
and most of the precipitate hardening, we are only talking about those compositions. We
are actually not talking about this kind of composition (alloy 3), because this kind of
compositions are not ideally suited for you know age hardening, because you will
somehow have some prior θ, which has been formed during solidification.
And as a result, the full advantage of age hardening, which is the combination of
ductility and strength, that you want to have, is not possible to be achieved, because of
this kind of samples. And as a result, you always try to restrict to, let us say, 4%, 4.5%,
5%. In fact, though I say this is the limit (point C, Refer Slide Time 13:10), in principle
you do not even reach there, because there is another problem that is associated with it.
What is the problem? For example, if I choose this alloy (alloy 5, Refer Slide Time
13:10). You may say sir, this alloy is still to the left of the maximum solid solubility,
why cannot I use that alloy?
Student: Range between the solvus and solidus
446
Yes, the range between the solvus and solidus is very small. When that is very small,
when you want to do homogenization treatment, what we call in steels as austenitization
treatment, so here we call it as you know homogenization treatment. So, during that
homogenization, if you do not maintain your furnace temperature carefully, you end up
into this two phase domain. Once you end up into that domain, some liquid forms, and
the liquid usually forms at the grain boundaries.
So, the moment liquid forms, immediately you will see, the liquid composition will be
something of this sort (point Q, Refer Slide Time 13:10) and that liquid when you start
quenching it, it starts precipitating out some θ at that grain boundaries. So, you cannot
prevent that liquid to solidify. So, that is the reason why, we do not want to go into a
liquid state. Because you are not in a single-phase state, you are into a two-phase region
and the whole purposes of solutionizing or whole purpose of homogenizing is to get into
a single phase state; you do not want to come to a two phase state.
So, that is the reason why you have to have a reasonable margin, considering the
furnaces, that are used in the industries where the temperature ranges, I mean accuracies
would not be of the order of one degree or so. They could be of the order of 10º or 20º.
So, you need to consider that. So as to define what is the maximum composition to
which I can go. So, the reason why I am saying all this is, the higher the composition the
higher the driving force, and for the same alloy composition, the driving force can be
447
further changed, by simply changing what?
Student: Temperature
Temperature; so once I fix up my composition, the only way I can change the driving
force is by changing the temperature; so a higher temperature or a lower temperature,
which is preferable?
A lower temperature will have a higher driving force. So, that means, you can say that
the ΔX0 α ΔT. Am I right? So, the higher the under cooling; so instead of choosing a
temperature of this sort (T1), if I choose the temperature of this (T2); so obviously, you
see here, this is the ΔT1 (T0-T1) and in this case, this is the ΔT2 (T0-T2). The higher the
ΔT, the higher is the driving force. The driving force, you can see, the extent of super
saturation is larger, and here (ΔT2) the extent of super saturation is smaller and you can
also see on the free energy diagram. How does the free energy diagram change with
respect to this temperature and this temperature? Which will shift up? Which will shift
down?
Let us say for example, instead of T1, I have increased to T2, what would happen to the
free energy curves?
Student: Both will go up
Both will go up, and then.
Student: Relative movement of θ will be larger
Relative movement of θ will be larger, and then. How does it shift the ΔG? We are all
bothered about that. How does the ΔG change? Will the ΔG change? ΔG, I mean for the
moment I fix up my alloy composition. This is my alloy composition (alloy 1 in G-x
diagram above), this is the ΔG that I am talking about, this is the overall driving force
and whereas this is, let us say, I am talking of this alloy. So, this is the ΔG for nucleation
(ΔGN).
We have two ΔG‘s. One is the overall driving force for the precipitation (ΔG). This is
the driving force for the nucleation (ΔG N), just for the nucleus to form, at the start of
nucleation. But this ΔG is for the complete transformation. What is the ΔG. And this ΔG,
how will it change? By changing the temperature, it will decrease. Why it will decrease?
Student: Because the θ goes faster
Yes, so the moment you see that θ goes up, with respect to α, what should happen? So,
for example, let us assume that this θ has shifted to this (see θ free energy curve drawn in
yellow color) .
448
Now, alloy composition is fixed, your alloy has not changed. What has changed is this
common tangent composition (see common tangent drawn in yellow color); that means,
the composition of α in equilibrium with θ has shifted and that is shifted to a higher Cu
content. Once that is shifted to a higher Cu content, this value (ΔG) decreases and that is
how
449
you can say. So, the higher the temperature, the lower the stability of θ because θ is
dissolving, at high temperature θ is not stable. So, as a result θ grows faster. θ free
energy rises faster, and once that happens with respect to α you would see the common
tangent change and that is how we can say the driving force changes. So, the driving
force can be easily seen as a function of both composition of the alloy and the
temperature.
Both of them decide this, and at the same time, we also know that this driving force,
there is something which has to be considered, which is something, which is a strain
energy ΔGϵ, which is also a volume free energy, which has to be deducted (from ΔG)
whenever you talk about the total driving force.
In case of a liquid to solid transformation, we never considered this ΔGϵ. So, here in case
of solid-solid transformation, we have already said that, whenever I say, I have this much
of driving force, the realistic driving force that is available is ΔG - ΔGϵ, and that has to
be sufficiently enough to take care of your surface energy term.
Surface energy term is in terms of the per unit area, whereas, these two (are per unit
volume. So, that is why, if you look at, for example, a case like, you remember this (see
G vs r plot above), this is the γ term; this is the ΔG v term. The moment you consider
ΔGϵ, the ΔGv term will be higher. ΔGv is actually negative and extent of negative
decreases
450
because ΔGϵ is a positive term. And as a result, the activation barrier that you are going
to get is going to be higher activation barrier.
So, the moment, I consider a strain energy term in a solid-state transformation, so this
activation barrier, which is the ΔG* that you get, if you again calculate using the ΔG,
which is a summation of the overall energies the way we do is put these values V and A,
in terms of the r, and then differentiate free energy with respect to r and then equate that
differential to zero and that is how you get these values, is not it? r* and ΔG *, and if
you try to look at this ΔG *, the ΔG* value is going to be larger here, because this ΔG
has got decreased.
So, in principle r* expression will be modified as,

r* = 2γ/(ΔG - ΔGϵ)
So, as a result, the r* is going to be larger, because the denominator is decreased. So, you
will have a bigger nucleus and then if you again put this back into the ΔG* term and
calculate the ΔG*, it will be
ΔG* = (16π/3).(γ3/(ΔGv - ΔGϵ)2)
Again, because this denominator value is decreased (Refer Slide Time: 31:35), ΔG*
value will be higher. So, you will see the moment we talk about solid-state
transformations, you need to consider that these things, the strain energy particularly,
which is the new term that is coming up and that new term will decrease everything.
451
And because of that, in principle if you even consider, if I take an alloy of this
composition (C0), this is the actual transformation temperature (T0) for that particular
composition, which is what we call it as solvus line. This is assuming there is no strain
energy term. All phase diagrams when you draw, we do not consider any strain energy
term. We do not even consider surface energy term also.
So, in principle now, the moment you consider the strain energy term, you will see the
transformation temperature comes down (T1). Transformation does not happen there, it
happens at a lower temperature, because only when you come to a lower temperature,
you have sufficient driving force, which can overcome the strain energy term. So, that
you have the actual transformation taking place.
So, if you now, consider the solvus line, considering the strain energy, you would see a
new solvus line coming up (dashed line). This has nothing to do, with the θ“ solvus or θ’
solvus that we have talked about it. Simply, just whatever may be the phase that is
coming out, as long as ΔGϵ is a non-zero term. When will this be a zero? When you have
a perfect coherency, when the misfit is zero, whenever the misfit is zero, and misfit
being zero is a too idealistic situation.
So, when the misfit is zero, in principle ΔGϵ is zero and when that is zero, then
everything changes. r* and ΔG* values will be similar
452
to a liquid to solid transformation. We need not have to bother about it, and then this
solvus line gets shifted to the original value. So, you can see that all these things are
shifted just because of one term called the strain energy term, which comes in and that is
why all the solid-state transformations are somehow trying to see how do we minimize
this strain energy and the surface energy.
This is what we have talked in the last class also, where the system tries to minimize the
strain energy by choosing that particular phase, which can be coherent. So, that the
surface energy is minimum and at the same time that particular shape, which can have
the minimum strain energy. This is how we have seen. Now let us look at.
Student: When θ is precipitating as a eutectic mixture, coming out as eutectic mixture,
can we say that θ is precipitating at the grain boundaries, when there is α also.
There is no precipitation, where is the grain there, when you still have a liquid there, why
do you talk about grain?
Let us say, how is the solidification (see the right figure)? This is all the liquid, and then
inside the liquid, you have the α dendrites, which have come out. These are all the dendrites;
of course, you can have side branches and things like that. So, now the liquid that is left out,
is somewhere between these regions (shown with dots). This is all the liquid; there is no
grain boundary still here. We can only talk of regions which are inter dendritic regions,
453
between the dendrites, and the liquid that is left out, between the dendrites is the one which
is undergoing the thing. Then the question is, where does this eutectic nucleate, and in fact
454
another question also people always bothered for quite some time, is what nucleates in
the eutectic.
If you have an α + θ in a typical eutectic phase diagram, this was and in the eutectoid, lot
of people studied eutectoid in much more detail, because it is the solid-solid, you can
easily study, because you can have a XRD, going to that temperature and also you can
quench that, after just the precipitation has started, reaction has started eutectoid reaction
you can quench it, and then see what is the phase that is coming out.
But from a liquid to solid, is not easy to find out, what is actually the first one to nucleate
in a eutectic mixture. Because whenever you say there are two phases, which are
nucleating simultaneously, how do they nucleate simultaneously, because both of them
have two different structures. One is α; another is θ, in this particular case, let us say.
How do α which is an FCC structure, θ which is an orthorhombic structure, nucleate
simultaneously. It is not an easy thing. That is where people have seen, that if it is a
hypoeutectic or a hypereutectic, life is very easy, because if it is a hypoeutectic, you
already have an α dendrites.
So, in eutectic mixture, when the eutectic mixture is forming, you can have the
nucleation that can occur on the α which is already there. And because it is α that is
already there, it will be the α of the eutectic mixture which will nucleate first, and the
moment α forms, because α is a pure Al kind of a composition, the regions which are
next to it, become rich in Cu. So, θ comes out. So, it is like one follows the other, the θ
follows the α.
Though, nobody has actually seen the nucleation. It is not easy to see, because that
happens in such a fraction of a second and in such a small scale, a few nanometer scale,
so you really need very-very high resolution microscope to be able to catch that
nucleation event, and also catch in situ. If you can keep a liquid, at that temperature, in a
high resolution electron microscope, let say, and then cool it and then see the
solidification happening, you can really see at a high resolution state, but it is not so
easy.
So, that is easy to understand on a hypo, on the hyper we already know, it is a θ that has
already come out. So, you can say the θ of the eutectic mixture nucleates on the θ which
is already existing, but, if you are exactly the eutectic composition, it is very-very
difficult to say. That is where people say, the phase which is easier to nucleate, which
455
has a lower activation barrier, whose structure is similar to that of the liquid structure, is
what nucleates first. So, that is where people always say FCC kind of phases are more
easy to nucleate than orthorhombic kind of phases and that too it is an intermetallic
compound.
So, intermetallic compound nucleation is more difficult, than a solid solution nucleating.
Because solid solutions, it is all random. Chemically, it is random. Structurally not
random, but chemically it is random, whereas, intermetallic compound you have an
ordered structure. So, you have fixed positions for the Cu and Al sitting there. So that is
why it is more difficult to nucleate. Anyway those are the issues. So, here there is no
grain boundary that we talk about.
So, the nucleation starts, let us say here (near the dendrites). So If I enlarge it, let us say
this is the thing. So, you will have things like that. So, the moment a small nucleus of, let
us say, α nucleates, next to it becomes the θ, and both sides the θ. Once the θ comes, the
region next to it again, becomes poorer in Cu. So, α comes out so that is how a small
colony forms, and that colony starts growing, and there will be another colony
somewhere else; another nucleation event.
You see the whole nucleation is actually on the α, and then grows into the liquid. This is
how it happens.
Student: But then if melted, where are the atoms of Cu and Al,
Inside the liquid, all Al and Cu
456
Student: Does the short range order of Al and Cu will correspond to FCC and
457
orthorhombic?
See, this is the interesting thing, when you say short range order, so inside the liquid,
when Al and Cu are there, it is not that all Al atoms are segregated somewhere, all Cu
atoms are segregated somewhere else. It is all a mixture; it is a liquid solution, where
inside the liquid, Cu atoms are statistically there, whatever is the %.
So, the probability of finding a Cu atom at any particular point will be, let us say 0.04,
because 4% of Cu let us say. So, that is how you should look at it.
Student: But when you cool it, how does the copper atoms assumes orthorhombic
structure
Yes, because, there are, you see the question is the phase is stable.
So, the moment α comes out, the remaining Cu atoms that are there; because the phase
that is most stable there has a lowest free energy is the orthorhombic; no other structure
has a lower free energy. The atoms get rearrange themselves to that particular structure.
Student: Why Cu cannot become FCC?
Because it is not stable, as simple as that; It is not stable.
According to the phase diagram if you look at it, for that alloy composition, θ is the
stable. You cannot get anything else, the question is why, for example, Al, why FCC?
Why not BCC? The same argument holds good here, though Cu atoms are there, the Cu
atoms could, in sometimes, you can get what are called metastable phases. For example,
we have seen in number of cases where, when you are cooling, even in precipitation
hardening also, we are seeing the GP zones, θ”, θ’. These are all metastable phases. They
are not the stable phases. The ultimate stable phase is the θ phase, am I right?
Similarly, during liquid to solid transformation also, sometimes you may get particularly,
when you are cooling it rapidly. For example, assume that this alloy is being cooled very
rapidly, then the θ phase which is the orthorhombic, you said it is more difficult to
nucleate. So, it is very difficult for it to nucleate. So, the system will try to choose that
particular phase, which can easily nucleate, if there is a metastable phase, but that would
be remain always as a metastable phase, provided you heat it, it will come back to the θ
phase.
There are quasi-crystalline phases also, which form generally, when you cool it, quench
it. Many alloys people have seen, particularly Al alloys: Al-Fe, Al-V, Al manganese, a
458
number of cases people have seen quasi crystals form, when you take the liquid and
quench it, because the quasi-crystalline
459
structure, the clusters, the short range order inside the liquid is similar to that of the
arrangement in the quasi-crystalline structure.
So, as a result you will see the liquid prefers to give you this, rather than the equilibrium
inter-metallic compounds; but these phases are metastable. Obviously, when you heat the
quasi-crystal, it gives back the equilibrium phase.
Student: How is a quasicrystal different from amorphous structure, because we are
quenching a liquid, wont there be a certain short range order?
Yes correct, the difference between the two is it has a structure. The quasi-crystal, as the
name suggests, it is a crystal. It is one type of a crystal. The only thing that is missing in
a quasi-crystal is what is the translation periodicity, I do not know how many of you read
about crystal.
Any Crystal, I mean we are deviating, but maybe, we will talk at some other time. Any
Crystal has two types of symmetries, one is called rotational symmetry, another is called
translational symmetry. Have you heard of this? A translational symmetry means if you
take an atom here and go in a particular direction by a particular unit distance x; let us
say, you find another atom, and in the same direction, move by another x, you should
find the third atom. The moment you find it, you say, yes there is a translational
symmetry, that is what we call it as periodicity. And what is this rotational symmetry?
That means, around a particular atom, if I rotate the unit cell in a particular angle, that
unit cell you would see, would merge into itself.
460
No new spots are generated. For example, if you take a cube and I rotate along a axis,
which is the z axis, let us say by a 90º. You will see that in that unit cell,
461
the new positions that you are creating, you are not going to differentiate from the old
positions. Take any square rotate around centre axis, you will see this part goes here, this
part goes here, this part goes here; 90º, and that is what we talk about rotational
symmetry.
A quasi-crystal has rotational symmetry, but does not have a translational symmetry. A
crystal has both; an amorphous does not have both. So, it only has a short-range order. It
does not have a long-range order, but a quasi-crystal has a long-range order, long-range
crystalline order, in terms of its rotational symmetry. Only thing is it has special
rotational symmetries. What are those special? For example, a normal crystal has a two-
fold, a three-fold, a four-fold and a six fold. You cannot have any other symmetry other
than these. Of course, you can also have a one fold symmetry, which is what we call it as
a trivial symmetry in a crystallography.
So, these are the possible symmetries, everything else is called forbidden symmetry. For
example, a five-fold symmetry cannot exist in a crystal. Any of you want to know why;
you should go to solid state physics by Kittel. He tries to prove it. I can also prove it in a
minute, and basically that such a rotational symmetry, a five-fold symmetry does not
guarantee a translational symmetry. That is the problem, the moment you have a five-
fold symmetry, you cannot have a translational symmetry, whereas, all these symmetries,
you can simultaneously have a translational symmetry. You can just go back to any
crystallography book and try to see the proof of it.
So, that is the reason why a five-fold, seven-fold, eight-fold, any other things other than
these, are all forbidden. For the first time in 1984, a person by name Schechtman saw in
an alloy of Al 14% manganese, when he prepared this alloy by fast cooling, the
technique that Paul Duwez used- rapid solidification, and took that the foil that came out
of it, and put it into a TEM, and he observed a diffraction pattern, which has a five-fold
symmetry. That was the first time anybody has actually seen a five-fold symmetry in a
diffraction pattern.
462
So, the diffraction pattern looked like this: a spot central spot with some 10 spots,
around it, at equidistant. 3 + 3, whatever it is equidistance. So, 10 spots on a circular
ring, let us say, which actually looks like a ten-fold and if you know a little more
crystallography, you can know, how to differentiate between a ten-fold and a five-fold.
There is something called Fridel’s Law.
If you read electron microscopy you will come across that, where we can differentiate
between how to see a five-fold and a ten-fold. But basically, even if it is considered as a
ten-fold, ten-fold is not possible under normal conditions. So, the moment he saw that,
then he tried to repeat it a number of times and every time he got the same thing, that is
when people started accepting, yes possibly this can exist. And that is when, they said
that we cannot call them as crystals. So, let us call them as quasi-crystals.
There is only difference between this alloy and amorphous alloy, when you do an x-ray
diffraction. An amorphous alloy, what do you see in an x- ray diffraction?
Student: Broad peak
You see a broad peak whereas, this quasi-crystal, you see sharp peaks, exactly the same
way as the crystal. So, you get sharp peaks and that is where recently around 19 end of
90s or so, people, till then the definition of a crystal is the crystal is that which has both
rotational and translational symmetries; they use to define. Now, they say a crystal is
that, which shows a sharp diffraction pattern.
463
So after people have discovered quasi-crystals, people have to change the definition of a
crystal we always use to, If you ever remember in the second year B.Tech, when you
start knowing crystal structures, you said crystal is that which has periodicity, is not it.
Something, which is periodic, is what is a crystal. This (quasicrystal) is not periodic, but
this is also one type of a crystal. So, people had to really change the definition, and the
crystallographers had to accept that there are alloys like this, materials like this, which
can give you rotational symmetries, which are forbidden under normal conditions.
But if you try to, It is simple, you imagine if I have to show you in a 2d a pentagon, if I
keep on putting pentagons next to each other, you will always have some gaps left out
and that is what is the difference. If you take square, you can keep on joining squares.
You take rectangles. You can do that. You can take an equilateral triangle, you can keep
on. This is what we call them as tiles for example, you can see a square tiles there. You
cannot have pentagonal tiles and still have no gaps left out, that is where the problem
comes.
So, if you take such a quasi-crystalline unit cell, and keep joining them, you will always
have some voids left out, which we call it as a frustration in crystallography. So, you will
always have such voids left out. And that is why people say these phases have a higher
energy state, they are at a high energy state, because of those voids that are available,
they are at a higher energy state, and that is why they are more metastable when compare
to stable structures. We will stop now.
464
Module No. # 01
Lecture No. # 20
Precipitate coarsening, stability of a phase, spinodal decomposition
So, last class we were talking about precipitation, there is one concept in precipitation,
which we have not yet covered.
I thought today we will talk about it, that is Coarsening, particle coarsening let us say,
people also call it as, there is another name for this, anybody has heard of it? Any
precipitates, when you heat them to higher temperature, take the alloy to higher
temperature the precipitates coarsen. This phenomena there is a name generally used in
hardening, what is it called?
465
A phenomenon, yes somebody said something, no. It is called Ostwald ripening, so this
is both are basically referred to the same concept. So, if you look at why should particles
coarsen all of you know, yes whenever the temperature is high diffusion is there, so there
is a growth. But, at the same time, if you look at if all the particles are of the same size;
if you assume that there are number of particles inside the material all precipitates, and
all of them are of the same size, then you would never see any particle coarsening.
Particle coarsening happens only, when there are particles of different sizes and in fact,
some particles dissolve decrease in size and some particles increase in size, this is the
basic phenomena of coarsening. And we will understand that, if you understand again
the same free energy concepts. If you look at a material, let us say this is one grain and
you have particles of different sizes, something like this (see the figure above)
You may ask me sir, how do you get particles of different sizes? When do you get
particles all of them are of the same size, can you tell me?
Student: Homogeneous nucleation
If there is a homogeneous nucleation at some time t = 0, and there is no further

nucleation after that. Nucleation is, is a transient phenomenon and I told you that; there
are different hierarchy of sites where nucleation can happen. For example, nucleation
can happen on the surface first, and then we talk about grain boundaries, is it not, then
you talk about stacking faults, dislocations and vacancies. So, there are variety of places
where nucleation can take place, and once certain nucleation sites are occupied, then you
start seeing the other nucleation sites, start becoming active.
The nucleation sites with the highest energy, once they are all occupied, then the
nucleation sites with the lower energy, lower with respect to the highest energy will start
working. For example, if the surface is, there is for some reason, it is not possible to
nucleate on the surface or if the surface is fully covered with nuclei, then as a function of
time the nuclei which have formed on the surface will start growing.
And at the same time, further new nuclei you will start seeing them either at the triple
junctions or at the grain boundaries; and if once they are covered, then you may see that
as they start growing you may see further nucleation taking place, may be at the stacking
466
faults or at the dislocations. So, you will see that nucleation is not a kind of a phenomena
where everywhere it happens at time t = 0, and all of them start growing at the same rate;
if that happens, then you would see a situation this is what we call it as site saturation.
When all the sites are saturated at the same time, at time t whatever time it is, and then
all these precipitates start growing at the same rate, then you can assume that all the
precipitates are of the same size, which is a very very very ideal situation, it never
happens that way. So, you will see that as a precipitate which is nucleated at some sites
start growing, there will be further precipitation at some other sites; and as a result you
will always see a distribution of particle sizes.
Whenever, you see precipitate particle size distribution, if you try to measure, you will
always see a distribution, there will be smaller particles, there will be bigger particles,
and you can talk about what is the average particle size; that is why we always talk about
average particle size. Because, we know pretty well, that they are all of not the same
size. Once you talk about this kind of a distribution, and then you are heating this to a
higher temperature, let us say all this kind of precipitation has taken place at some
temperature.
Let us say during natural ageing or artificial ageing, and in service this material is being
exposed to a higher temperature now. If it is being exposed to a higher temperature, this
kind of a micro structure how does it grow? We know pretty well that at high
temperature you are providing sufficient thermal energy, and thermal activation for the
diffusion to occur. So, there is a possibility that all particles can grow, but do you see all
the particles grow or not.
So, that is where, if you try to understand smaller precipitate which is in equilibrium
with matrix, and this bigger precipitate which is also in equilibrium with matrix; if you
look at the matrix composition, which is in equilibrium with smaller precipitate and this
larger precipitate it is never the same. How does it come out? I would repeat that the
matrix composition, the composition of the matrix which is α in our case, and let us say
this is you call it as a θ, whatever precipitates it is.
467
The composition of α in equilibrium with this smaller precipitate, it is the same θ, larger
is also the same θ, the composition of α in equilibrium with a smaller θ is different from
the composition of α in equilibrium with a larger θ, why does this happen?
Student: The composition around the θ compare to this is less
Why?
Student: Because the diffusion is occurred for longer time
It has nothing to do with diffusion
Student: Curvature effect
It is the curvature effect because of the curvature effect you would see actually the free
energy of that two particles is different. You always see whenever, we are talking of a
free energy diagram, we always talk about volume free energy only, we do not bring in
the surface free energy term, that is why, if you look at all the free energy composition
diagram that you draw, you assume that the precipitate size is infinite.
Actually precipitate size is not infinite, and different -different precipitates will have
different -different free energies, though they all have the same structure, same
composition. Because, the surface energy has to be added to the volume free energy, so
that you actually talk about the free energy; so a smaller particle always will have a
higher free energy, when compared to a larger particle.
468
And if that is the case, if you draw that G-X diagram, let us say this is the free energy
composition diagram (see figure above=, this is of α, and this is let us say θ phase with a
larger particles and θ phase with a smaller particle. One you can call it as a smaller
particle, the other is for a larger particle. So, a smaller particle will always have a higher
free energy, when compared to the larger particle.
And if that is the case, then if I draw now a common tangent, you would see that this
concept which I just now said, that the composition of α in equilibrium with a larger
particle is different from the composition of α in equilibrium with a smaller particle. So,
this is the composition of α in equilibrium with θ, which is let us say larger (C αθlarger) and
similarly, this is the composition of α in equilibrium with θ which is smaller (Cαθsmaller).
469
The moment that is the case, then if I now simply look a smaller segment of where you
have one small particle, one large particle, and there is α in the middle, the composition
of α which is just in front of this larger particle, the composition of α which is just in
front of this small particle is different.
And which is higher?
Student: The smaller
The smaller is the higher, so if I now draw what is called the compositional profile, how
would the compositional profile look like, the θ phase the θ phase, what is the
composition of θ phase, in aluminium Cu?
Student: CuAl2
Al2Cu or CuAl2, am I write and what is the composition of α?
Student: 4% Cu
That is, that is the composition with which you have started.
470
And we are talking about the starting composition, alloy composition is somewhere here,
that is the alloy composition (indicate by vertical line between free energy curves) with
which you have started, which will be a two phase mixture and what we are talking
about is a equilibrium composition of α. That means, if you look at the phase diagram
(bottom figure in above image), this is let us say α + θ here, and this is α.
So, we are talking at any given temperature this composition, this composition that you
are looking at of course (on solvus line, indicated in pink), this is a temperature versus
composition diagram or T-X diagram, if you want to call it because, we called G-X
diagram. This line (pink color) what is this line called?
Student: Solvus
Solvus line, this solvus line is always assuming that, the θ has an infinite size, any phase
diagram, any solvus line that you see, the assumption is that the second phase and the
first parent phase; parent phase is anyway matrix.
So, you cannot talk about the size of it, and there also in principle if the parent phase is
also, a nano grained α let us say, it would definitely have a different free energy. So,
there also, we always assume that the grain size of α is infinite, and also that the
precipitate is also almost an infinite size.
The moment you bring in size effects, this is what we call it as size effects then you will
start seeing that the free energies start shifting. Α also instead of a an α of a very large
471
grain size, if I now start looking at nano grained α, trying to precipitate out of the θ, then
the α free energy curve also will shift. So, nano grained α
472
will have a definitely higher free energy with respect to an α of a much larger grain size,
so all those things are what we call them as size effects. So, this solvus is in principle for
an infinite, and the moment I assume that the θ has a smaller particle size, you would see
that particle that composition will be different.
And that if I try to draw, you would see that if I draw composition, so let us imagine that
this smaller precipitate has certain, certain composition, this (left axis) is the composition
in terms of %age Cu if I plot, this is some value, let us call it whatever Al 2C u
composition is, in terms of a weight % it is approximately let us say 55 %.
Now, this is the composition of α somewhere here, which is usually just for the sake of
understanding, let us again take this the composition here (near the larger precipitate)
also, this composition again this precipitate, inside the precipitate it is the same
composition, it is the same θ, this whole thing larger particle is θ. So, the composition of
the θ is this value, so inside the θ the composition is this (55% Cu). And then, if you look
at this is α, the α composition here (at P) and the α composition here (at Q) will be
different.
So, there is a gradient here, so if I look at the concentration profile it looks like this
(drawn in yellow color), if I start from the center of the particle to the center of the
particle, let us not bother about the other sides; because on either sides there may be
other particles, we will not bother. So, we will simply look at from the center of this
particle to the center of this particle, if you
473
simply look at the composition profile, this is how you will see. Inside the α there is a
gradient here, the smaller particle the α which is in equilibrium with smaller particle will
have a higher composition, the α in equilibrium with a larger particle will have a smaller
composition.
And once you have a gradient, you always have a diffusion. You are always have a flux,
at any given temperature, because a flux is j equal to?
Student: -DdC/dx
Correct -DdC/dx, and that dC/dx you can see here, the x is nothing but the inter particle
distance, once you know the inter particle distance, you can talk about what is the dC/dx
and at any given temperature D can be known. So, you can actually see and what is the
result of this diffusion? You see that Cu atoms which are here (near smaller particle),
want to diffuse towards this (near larger particle).
But, you need a constant supply of Cu atoms for the Cu atoms to go down the drain,
down this particular decreasing concentration gradient. Where are the Cu atoms coming
from? The Cu atoms come from this particular particle. The smaller particle provides the
reservoir of Cu atoms, so that the Cu atoms start moving to this direction (towards larger
particle), because the θ has a higher Cu content.
So, this θ (smaller particle) rejects some Cu atoms by dissolving by a small amount, a
small amount, small Δx of this θ if it dissolves let us say, what happens some Cu atoms
are released. And those Cu atoms will moved down this concentration gradient, and
come to this region (near larger particle) and once that they reach that region, this
particular θ needs Cu atoms for it to grow.
So, it provides Cu atoms, once the Cu atoms come towards this, they join this (towards
larger particle) particular θ and the θ grows by a small Δx, depending on how many Cu
atoms have come and joined. So, this gradient will more or less remain the same,
excepting that the concentration values this value (point P) and this value (point Q), what
do these two values depend on?
Student: radius.
474
Radius, so, as the radius of these two particles change, in principle these two values also
will change, as this smaller particle, becomes smaller and smaller as it shrinks what
happens, so you can see as this particle shrinks the free energy goes up. If this free
energy goes up the common tangent starts touching it, at a composition which is farther,
which is higher in Cu content, so that means, you would see that this value (C αθlarger)
keeps going up.
And what happens to this larger particle, as this starts becoming larger and larger its
value, free energy comes down and that means, the common tangent again, the Cu
content of α comes down, that means, you actually see that the gradient is increasing.
Because, the ΔC = Cαθsmaller - Cαθlarger starts increasing as the smaller particles start
becoming smaller, the larger particle starts becoming larger. So, in the process what
happens is slowly smaller particles will dissolve, at the expense of these smaller
particles, the larger particles start growing; it is like you see that in a typical life, the
bigger once try to grab the smaller ones. So, exactly the same thing is happening here,
the bigger particle is just growing at the expense of the smaller particle; and the whole
driving force for such a thing is coming because of this, free energy.
Once you understand this free energy that the free energy of different particle sizes will
be different, and you can easily understand why this is happening. And this gradient
starts dC/dx increasing and of course, the distance you may say, what is going to happen
to the dx, , when you say this larger θ is growing and this smaller particle is decreasing.
So, in principle they may remain the same or can change by to some extent, but we do
not know, what is the rate at which one decreases, what is the rate at which other
increases.
So, there may be a small change in the d x also, but d c you can see a significantly a
large d c, and because of which more and more, so as the time proceeds at that
temperature; you will see the flux, the rate at which the atoms move is going to be higher
and higher. This is usually against the normal case of precipitation, in a normal
precipitation, how does a precipitation happen?
475
If you look at, the precipitation concentration gradient in precipitation, so the moment
the α, if you take α, inside α the moment a small nucleus precipitates, a small θ
precipitates, how does the further growth happen there? The further growth happens
again, because of the gradient, where is this gradient coming from. So, the α that you
have is the starting α is this α (C1, Fig. 1 in above image), am I right that is 4 % Cu, the
moment θ is precipitated, inside the α, the α which can be in equilibrium with that θ is
not this α.
This 4 % α cannot be in equilibrium with θ why, because if I take the, forget about the
smaller particle, let us look at the larger particle as equilibrium particle or you can
choose any of them. So, in principle if you look at this particle, this particle it cannot be
in equilibrium with C1 because, you cannot draw a common tangent between this point α
of that composition C1 with the θ.
The only equilibrium that you can have is with the θ with some other α, and that is what
we call it as composition of α in equilibrium with the θ (C αθ); and that composition is
different from this starting α. And as a result, if I now draw again a concentration profile
(Fig. 2 in above image), so if I take a small θ, this is let us say θ, whose composition, if I
plot % Cu, whose composition is again the 55 % let us say has nucleated.
476
And now, Cαθ that means, the α composition which is adjacent to the θ has a different
composition, than the α composition which is far away from the θ. The composition of α
which is far away from the θ is this composition C1, the starting composition. The
composition of α which is adjacent to θ is of different composition.
Which is lower, which is higher?
Student: Adjacent to θ is lower Cu content
Yes, adjacent to θ is a lower Cu content, because of which you will see, something like
this. This is the % of Cu the initial α (indicate as αinitial in Fig. 2) the α with which you
have started, you have taken an alloy of 4% Cu made it into a solutionized condition and
then, quenched it and from there, the actual precipitation has started.
And that particular α is that 4%, but the α which is in touch with the θ is actually smaller
and because it is smaller, now you see there is a gradient here. And because of the
gradient what happens is, this Cu which is there in αinitial starts moving towards the
interface, and once it starts moving towards the interface, there are some Cu atoms get
accumulated at this and once the amount of the Cu atoms is sufficient for the θ, because
θ is 55 %.
So, as the Cu atoms keep coming towards θ, once in a small region the Cu atoms reach
55 %, that small region gets converted. Now, it is like, almost like α 55%Cu  θ55%Cu, it is
like almost like a polymorphous transformation, allotropic transformation kind of. So,
once the Cu atoms reach that concentration level, it is only a structural transformation
now; a FCC structure changes to a θ structure. So, as the Cu atoms get accumulated in
that small region, that small region gets changed to the θ.
477
And as a result, now the θ grows by a small distance, but in the whole process, because
the whole precipitation is happening at a fixed temperature, this value (point P, in image
above) will never change, because this value is coming from the free energy diagram.
And in the free energy diagram, the free energy diagrams are always drawn at a fixed
temperature. At a fixed temperature, the Cαθ is fixed.
Provided the θ size has not changed to such a significant extend, that its value can
change, if the θ size changes, we can expect that the, this value also can change
otherwise, you can say this value Cαθ is fixed. Once this value is fixed, what is that that is
changing as a function of time, in this whole process? You have atoms of Cu coming
from this region towards this and then θ is growing, and this is fixed.
And then what is, what changes
Student: α composition
The composition of α in the matrix what happens to it?
Student: Decreases
It decreases, because the Cu atoms are being released from here and coming here, so the
α concentration decreases and once it decreases what happens to the gradient? Gradient
decreases and as a result the process becomes slower. The reason why, I had to
478
tell you is that is which is reverse of coarsening, so as a result you can see as a function
of time you will see.
Student: Sir, when θ starts growing, still nucleation takes place
Yes I am, I am not talking about that situation, you know this is we are talking of one
particular θ with respect to its α, how does this grow? There will be another nucleation
somewhere else, at that region also similar thing will happen, as this θ wants to grow, the θ
can grow only when Cu atoms come and join to θ otherwise, θ cannot grow, because θ needs
55 % of Cu atoms. So, unless they are supplied to the α-θ interface, the θ cannot grow.
And where do these Cu atoms come from, they cannot come from the thin air, they have
to come from the α only. And as the α supplies more and more θ Cu atoms, α gets
depleted in Cu.
Student: How it can cause further nucleation
No, this is we are talking of a small region, very very small region and there can be some
other region where, see this whole thing is happening in a region of few 100 nano
meters. And in a alloy the grain size itself is very large 100 microns or so. So in some
other corner of the grain you can have another nucleation even can happen.
So, we are talking of that particular, because the θ in the beginning is about 10, 20
nanometers and that particular θ how is it growing, that is what we are looking at. So, as
a result after sometime the whole thing would become like this (tf). That θ would have
grown to a certain extent and then the α composition comes to that of the equilibrium
composition . That means, there is no more excess Cu left out in α to be supplied to the
θ, so now, the growth stops.
Once this happens, the gradient is 0, once the gradient is 0 then there is no flux, and if
there is no flux there is no growth. So, this is how the actual precipitation takes place,
this we saw with the next topic that we are going to cover what is called Spinodal, you
will see there is a lot of difference between this and the spinodal, we will talk about it in
a minute.
So, is this clear now? So how do precipitates grow, yes
Student: Sir, I did not understand why that θ, when α in equilibrium with θ at that point
is lower compare to the.
479
Why Cαθ is lower compared to αinitial , am I right?
480
Imagine I have started with an α of certain composition (C1), this is a 4 % Cu, I started
with this alloy and I am holding this alloy at some temperature. I am holding this alloy
which is solutionized brought to this temperature (T) quenched, and I am holding at that
temperature, at that temperature is α stable? It is not stable.
So, α as a single phase, it is not stable be very-very specific when you say α is still
stable, you still have the under in the equilibrium phase diagram, you have α + θ; we
cannot say α is not stable, α as a single phase it is not stable. α can coexist with θ for that
particular composition, for that temperature, and now if it is not stable as a single phase,
it wants to precipitate.
How do you understand this precipitation, simply draw a tangent, if I draw a tangent at
that composition, then what are all the phases, whose free energy is below that tangent
all those phases can in principle nucleate. Let us not now consider about all the various
phases that are possible, just look at only one phase, call it as θ phase, and that θ phase
the moment I look at this, this particular θ at that temperature has a certain free energy,
and the moment there is a driving force (LM), this θ will nucleate.
The moment the θ nucleates inside the α somewhere, either grain boundary or where
ever it is, the moment it nucleates, this θ cannot be in equilibrium with this α, it is
impossible that you can have a common tangent between the two.
481
So, then what happens is, you will see that this θ will find out some composition of α
with which it can be in equilibrium, and as a result the α which is adjacent to the θ, will
have a different composition. In fact, this can be even you know physically understood,
see when you think of a θ being nucleated inside an α, θ having a higher Cu content,
what happens is Cu atoms get accumulated in a small region, and that region gets
converted to θ, that is what if you want to call it as a nucleation event.
Then they are getting accumulated from α, so the region adjacent to that obviously, is
depleted, that is what is this depleted region. And this depleted region is having a
different composition. So, the α which is in equilibrium with θ is having a composition
which is lower, and the α which is far away from this nucleation site is actually having
the starting composition. The starting composition is this C1, and the composition in
equilibrium is this Cαθ, this is what we are talking of the gradient; if this difference was
not there, there is no gradient, clear.
Student: Sir, but the θ immediately after it nucleates grain size is very small
Yes
Student: Then the free energy should be higher
In principle you are right, very right, so in principle if I consider, that the θ is very- very
small, you can actually draw the free energy curve of θ, which is much higher and see
that the common tangent with that particular θ, can be almost like, if this is the θ (upper
free energy curve for θ) , θ have at the nucleation site stage, you can actually draw a
common tangent between that θ and the starting α itself, this is assuming that that
particular small amount.
But, this is only at the nucleation’s event, but when you want to talk about what is the
overall driving force for the nucleation, overall driving force you should always consider
the θ after it has grown. After the whole precipitation process is over, when you are
looking at the micro structure, what is the θ, and that particular θ if you compare, that θ
which is what we calculate only from the bulk free energy point of view where surface
energy is not considered, that is turns out to be this (LM).
482
This is only the volume free energy, this and the surface energy, so this whole thing you
can call it as γ, if you want to call it corresponding to that particle size of whatever is the
initial particle size. So, the free energy gets lifted, because of the surface energy and at
that particular moment, you can say yes it is in equilibrium, but the moment it grows to
even a smaller extent, it cannot be in equilibrium with that.
So, we are now looking at only that is the reason why, you have to see that most of the
free energy composition diagrams are always drawn only considering the volume free
energy; capillarity effects or the surface effects are not considered. And once you start
considering, this kind of concepts come into picture.
Student: Sir in this case we have seen that we have considered size factor
Correct that is, that is because there is a size distribution. If there was no size
distribution, we would not have considered size effect. Because, there is a size
distribution you wanted to know which one will grow, this will grow or this will grow,
so only if you understand that, you will understand this (Refer Slide Time: 33:45).
Otherwise, at when you are looking at one individual particle, you need not have to
consider its size effect.
Now, let us come to the next concept, which is similar to a precipitation, which is what
we call it as spinodal decomposition. Let us understand, try to understand what
483
are the free energy concepts involved in this. As you can see this, we call it as
decomposition, we do not call it as a precipitation; there is the difference between a
precipitation and decomposition. What is the difference? When you say decompose,
what does it English term mean?
Student: Same phase but different composition
Yes, something is splitting into two, that is what we call it as a decomposition. When
you say precipitation basically, something is precipitating out of something else. So,
there is a matrix from which some precipitate is coming out, here when I say
decomposition a matrix is splitting into two, that is the meaning of decomposition. Let us
try to understand that, and you will see that spinodal decomposition always occurs, when
the phase diagram shows a miscibility gap. When there is no miscibility gap in the phase
diagram, you will never see a spinodal decomposition.
So, let us try to consider a phase diagram with a miscibility gap, something like this α
splitting into α1 + α2 and whenever, you talk about spinodal decomposition, you need to
consider one aspect of stability, what is this stability?
Whenever a free energy curve looks like this (Fig. 1), this phase is called a stable phase,
why because if I take an alloy of this composition C1, if this alloy
484
wants to decompose into two different compositions of the same structure, that is what is
the decomposition means that means, α  α1 + α2. I we call this as free energy of α in G
versus X plot, if it splits into two, one with a lower composition, another with a higher
composition. So, that the overall composition remains the same that means,
Cα = Cα1f α1 + Cα2f α2
this is what is the mass balance.
As long as the mass balance is maintained, some the α splits into some part of α1, some
part of α2, for such a process to happen, what is the overall free energy of such a
mixture? The overall free energy of such a mixture, if this is the free energy of α1(G α1),
this the free energy of α 2 (Gα2); then the mixture will always have a free energy which is
decided by the line joining those two.
Any two phases having two different free energies, if you want to know the free energy
of the mixture, you simply draw a line joining the two, and for the composition C1 that
of interest to you if I choose this alloy composition having these two phases, the free
energy of such a mixture is this (G mix). This is the two phase free energy that is what we
have been showing in all the phase, even earlier case also precipitation also whenever, an
α phase splits into α + θ. So, the α-θ common tangent that we consider, and that
particular point on the common tangent gives you the two phase free energy.
And the single phase free energy, if it is higher than the two phase free energy, then only
precipitation will take place, we have seen that. So, this G mix is the free energy of this
mixture, and this is higher than the free energy of a single phase, as a result such a
decomposition to two is not usually observed; because there is no driving force for such
a thing to happen. So, for a single phase α to split into two phases in this kind of a
situation, does not happen.
Because, there is no driving force, because of such a decomposition, the free energy is
actually increasing rather than decreasing, we know that every process where there is a
decrease in free energy that is favourable, and increase in free energy is not favourable,
is it not. But, if your free energy curve is like this (Fig. 2), then this α can split into two
phases of
485
α1 + α2 and which have a free energy of this (G mix’). So, it can split and it can not only
split like that, it can keep on splitting more and more that means, the α1 and α2
composition will keep on deviating more and more from the α; and this causes further
decrease in the free energy.
That means, if I start with a an α having a uniform composition and if this α is having a
free energy curve something like this (Fig. 2, Refer Slide Time 36:11), at certain
temperature, then what happens is suddenly this α is not stable, because its free energy
curve is like that. So, what happens is there will be a small fluctuation, even a small
thermal fluctuation, temperature whenever you say this particular alloy is at 500 ºC,
when you say this 500 ºC is a bulk temperature that you are talking about.
In a small region of a nano region, you can always have a small changes and this can
cause a compositional fluctuations (Fig. 1 in above image), in such a way that there can
be a small region, where the composition can decrease, and another region where the
composition can increase. Though I have drawn it with such a scale, this can happen in a
less than a nanometer scale. So in less than a nanometer scale if you see a small region
where the atoms, because there are always atoms moving at any given temperature,
unless you are at absolute 0.
486
Because of this, atomic fluctuations you can see a region where the composition goes
slightly below the nominal composition in some region and obviously, those atoms have
to move somewhere else. So, they go to some other region where the atoms are more
than the nominal composition. Once this happens, this kind of a fluctuation is a stable
fluctuation in this, because the moment such a change happens, then suddenly this region
with a lower Cu content, this region with a higher Cu content let us say will be a stable
couple.
Because, that has the resulted in a decrease in the free energy, but if that happens in this
alloy (C1, Fig. 1 in Refer Slide Time 36:11) it will not be stable, what happens is the
atoms will come back the atoms which have the excess atoms here, will come back and
join the α, so that again the α will have the same composition. Where as in this kind of a
situation the atoms, once they come to this situation they will not go back, because that
is a stable situation.
And now, what happens is this particular fluctuation starts growing, why it grows,
because by further growing you will have a further decrease in free energy. So, this
fluctuation starts with the same wave length, the wave length gets fixed which may be of
the order of few nano meters, and within that wave length the amplitude starts changing.
So, that means, one region becomes more rich in Cu, the other region becomes less in
rich in Cu, but the overall composition satisfies this mass balance, the overall
composition of the alloy is the same.
But, you will see that this composition start increasing and decreasing, and to an extent
that finally, if I take that alloy and if I am holding at room temperature this alloy, and it
will keep on growing until the extent that α1 achieves this composition (C α1, Fig. 2) α2
achieves this composition (Cα2, Fig. 2). And in this fluctuations (drawn in Fig. 1) the
lower is α1 let us say the higher is α2.
So, this this wave or this compositional wave will keep on changing its amplitude to such an
extent that finally, you reach this value and this is what happens whenever, you have this and
I have directly correlated G-x diagram to phase diagram, we will see in a minute, why I
should correlate that to this, at the moment this and this are not really co related. So, you will
see that whenever I take a phase of this kind of a free energy diagram, you will actually see
this kind of a free energy diagram, in this miscibility gap, we will see it within a minute.
487
So, but if that happens you can see this will keep on increasing. This in comparison to
the precipitation, in a precipitation the composition of the precipitate is always fixed;
you start with certain fixed composition. And that during the whole precipitation process
as a function of time, that precipitate composition is not changing, only that is changing
is your the α composition will keep of changing.
Whereas here the, if you want to call these as precipitates, they are actually not
precipitates, but these are composition, only when the whole growth has completed.
Then you we will see that this α1 would be a different composition, this α2 would be a
different composition and this happens homogeneously throughout the sample. Whereas,
a precipitation does not happen homogeneously, it happens heterogeneously, in some
heterogeneous nucleating sites; whereas, here this happens throughout, why it happens
throughout because, throughout wherever it can happen this is stable.
So, you can see a homogeneous nucleation can happen in this kind of a situation, and
that keeps on growing, and this is in contrast to the what I have shown you in
precipitation, where the gradient will keep on decreasing as a function of time, here the
gradient is actually increasing as a function of time. If you look at, at the interface, this is
the if you want to call this as one phase, this as another phase at the interface, if you look
at the gradient the gradient is actually keeps on increasing(see the slope at the point of
intersection of compositional wave with x-axis, Fig. 1).
Whereas, in the precipitation, normal precipitation at the interface between the θ phase
and the α, if you draw a gradient and that gradient will slowly decrease, because if you
remember, this is the one this slowly becomes like this finally, it is like this (Fig. 3). So,
this gradient keeps on decreasing whereas, here in spinodal decomposition the gradient
initially is 0, gradient is initially 0, there is no compositional difference and the slowly
the gradient starts building.
And that is why some people want to call this as uphill diffusion, because the gradient is
actually increasing, and this how do I understand from the free energy point of view.
488
Let us now try to draw, the free energy curve at this temperature T for this. If I draw the
free energy curve, the free energy curve actually looks like this (Fig. 2), am I right
because when do you get a miscibility gap, when the ΔH is positive. Whenever, the delta
H is positive and that is higher than the TΔSmix term, then you will see the free energy
curve will have this kind of a shape.
And then if that is the case, if I draw a common tangent you get these two points (P and
Q, Fig. 2) and what are these two points, these are these two points (P and Q in Fig. 1),
we call them as the binodal points; they are usually called the binodal points, they they
are nothing but, the equilibrium composition of α1 and α2 at that temperature. Now, if
you look at this, this particular free energy curve appears to be a mixture of two types of
free energy curve ; one which is a curve like this (K to L), again a curve like this (M to
N) and another curve which is like this (L to M).
If you look at these two curves, concave upwards and concave downwards, so you have
two parts here, which are of similar in nature, which are related to stable phase condition
(hatched region), and there is one region here in the middle, and that is the region, where
there is a change that takes place of slope. If you look at the slope of this free energy
curve, as a function of composition, the slope here is what is it at the two at the pure
metal end? It is -∞
So, this -∞ slowly starts increasing and when you reach this value (point P), what is the
slope?
489
Student: 0
It is not 0, it is low, it need not be, it will be 0 only when the tangent is horizontal,
tangent need not be horizontal.
So if it reaches a value, higher value, so from -∞ if I now plot. So, from the -∞ it keeps
on increasing, and then again you see the free energy curve, what is here? It is positive, it
is positive and then it keeps on increasing and at some point (point L and M, Fig. 2,
Refer Slide Time 46:38), this slope is changing.
So, can you draw these slopes, next class when we come, I want you to see how this free
energy, take any such curve, which has an inflection inside, draw the ∂G/∂x and then
also draw ∂2G/∂x2 . Then you will see something interesting happening. Next class, when
we start we will start with that, and by the time I want you to draw and see what
happens.
You will see there is something which happens, when you look at this part and that is
actually 0 at these two positions; we will try to see that. So, please draw a free energy
curve and for that free energy curve just below that, draw this for the same scale and
below that you draw this for the same scale and see how this changes, we will stop with
that.
490
Module #01
Lecture #21
Spinodal decomposition
So last class we started talking about Spinodal decomposition, we will continue there.
We said, you will see a Spinodal decompositions whenever, you have a miscibility gap
in the phase diagram- this is what we have said. If there is a miscibility gap in a phase
diagram (Fig. 1), it will be reflected in the free energy composition diagram something
like this (Fig. 2), Am I right, at any given temperature within that.
So, if you draw at a particular temperature, you will see a free energy diagram which
looks like this (Fig. 2). And we always know that, towards the two pure metal ends, the
slope of free energy verses composition is - ∞ . At pure A, it is - ∞ if you are looking
from solute side, if you take B as a solvent and ‘A’ as a solute, then it will be - ∞ , but if
you are looking it from other angle, it is + ∞ .
So now, if you look at this, we will understand the Spinodal decomposition better, if we
try to see what is the first and second derivatives of this free energy. Let us try to see the
491
first derivative. Last class we just started talking about it. So if you try to look at the first
derivative, of this free energy curve, basically what we are looking at is, what is the
slope of free energy curve, as a function of composition (Fig. 3)? So try to see, how the
slope is changing, as a function of composition. Start from one end, and start looking at
how the slope is changing. If you look at it, when you start at pure A, it is - ∞ .
And as you increase the concentration, slowly, the slope is increasing or decreasing?
Student: Increasing
(Please refer Fig. 2, 3 and 4 marked above for the discussions below)
Slope is increasing. And increases and reaches around 0 here at point K. Isn’t it? And
then, it continues to increase, in a positive direction, up to K it is in the negative
direction, it is increasing. And then it continues to increase, on a positive side. And the
only difference is the rate of this increase, changes at some point L. So, you can see that,
the slope of this is higher and higher as you go, and at some stage, again it starts
becoming lower and lower.
The slope is positive, but the only thing is, the rate of change of that particular slope, is
going to be different. So, as a result you will see that, as you go along, the slope is
increasing and at a particular point L the slope reaches a peak, and then again starts
decreasing and again reaches 0, at point M. And that particular point L, where it reaches
a peak, is a point of interest to us. We will see what that point means.
And, in the second direction again, from point M, you will see the slope now starts
decreasing, the slope is negative now, and again up to certain extent, the slope keeps on
increasing, beyond a particular point N, the slope again starts decreasing and then, comes
back to 0 somewhere here at point O. And then, again, it is positive and it is positive and
goes to ∞ . It will be on + ∞ if you are looking at from this angle.
492
So, you can see that basically, in the domains KM and MO, there is a change in the
slope. In KM, the slope is always one sign only, slope is always positive, but, that
positive is reaching a peak, and then coming back to 0. In MO, the slope is always
negative, but the negative is reaching a peak, or you can call it as a trough, instead of a
peak, and it is reaching a trough and again going up. Now, this slope, if we call it as ∂G/
∂x and then, now, look at the second derivative of free energy (Fig. 4), second derivative
of free energy is nothing but the slope of ∂G/∂x. Now look at the slope of ∂G/∂x , how
does this slope of this change? And you can see it is positive slope up to L, and this
positive slope keeps on decreasing: initially it is + ∞ and it keeps on decreasing and
reaches around 0 at point L.
And then, this particular portion LN, this whole portion, the slope (second derivative,
Fig. 4) is negative. Again after O, the slope is positive. So, somewhere at point N, again
this is a 0 slope. And then it becomes positive. But in this composition range LN alone, it
is negative slope. And the negative slope again, is in such a way that, at this point M the
negative slope changes its curvature, in such a way that, you see that negative slope; it
reaches a kind of a trough here at point M.
So, you can see the slope in the absolute term, is actually increasing and then again the
slope is decreasing, in such a way that in the whole domain LMN, the slope is negative.
So you can see that in this particular domain LMN of composition, the second derivative
of the free energy, ∂2G/∂x2 , is negative. And that, basically, boils down to two points L
and N, where, we say these are the two points, where, the ∂2G/∂x2 = 0.
493
Because, those are the two 0 points and those two points where it is 0, indicates the
points, which we call them as Spinodal points. And why are they significant? Because,
within those, in the composition range, characterized by these two Spinodal points,
within that, the free energy curve is, convex (Left figure above). Beyond those two
points, the free energy curve is concave (Right Figure above). So, to the left of point L,
to the right of point N, the free energy curve is of concave shape. And the moment you
have a free energy curve of this shape, yesterday we have seen, in the last class, that, it is
a stable state; whereas the convex one is an unstable state. The moment you take any
composition, the composition would like to split into two phases, of two different
compositions. Whereas if you are in the concave regime, it would not like to split,
because by splitting, it is increasing the free energy (see the right figure). In convex
regime, by splitting, it is decreasing the free energy.
So, the points which differentiate these two phenomena: one with a stable state, of a
single phase being stable and the other region where the single phase is not stable, are
those points which we call them as Spinodal points. And to understand where these
points are, you need to come to the second derivative, and see, where the second
derivative is 0 and between those two compositions, the second derivative is always
negative. So that is why we say, Spinodal decomposition occurs, when the second
derivative of the free energy is negative.
494
So, all those compositions, where the second differential to the free energy verses
composition is negative, that whole regime, is called Spinodal regime. How do I, now,
identify this on the phase diagram? So the moment, I draw a particular temperature, let
us say at room temperature, let us assume that, this free energy composition diagram is
drawn at room temperature; at room temperature we know these two points (K and O,
Fig. 1). What are these two points? We call them as binodal points. And these two
points can be seen as minimum in the Free energy curve (Fig. 2)
Now, between these two points, you have two compositions, which are the, so called the
Spinodal compositions, and those points are L and N. And obviously, the Spinodal
points have to lie within the binodal points. They cannot be outside the binodal points,
because from the nature of the curve itself, free energy curve you can see, that the two
Spinodal points are always, within the two binodal points.
If that is the case, somewhere in between K and O on phase diagram (Fig. 1), awe have two
Spinodal points L and N. And if I draw the same free energy curve, at different-different
temperatures, at each temperature, you will get two binodal points, and two Spinodal points
until, you reach a temperature where the two binodal points actually merge into one single
binodal point. And obviously, when the two binodals themselves merge into one single
binodal,
495
then obviously, you will see that the Spinodal points also merge, and that is, this
temperature (TMG), where the miscibility gap vanishes.
And so that means, at various temperatures, I would get different-different compositions

if I draw and finally, if I join all of them, I get a curve like this (pink colored, Fig. 1 in
(Refer Slide Time: 09:54)) This is what we call it as, Spinodal curve. One can easily
calculate this, by simply knowing, what do you need to know, to draw this free energy
curve? Okay.
Student: GA, GB
Yes you are right; you need GA, GB. What else?
Student: Ω
Ω, once you know these three, provided, there is a assumption there when I say I need to
know Ω, what is assumption?
Student: Ω is greater than 0.
Of course, Ω > 0, fine. Anything beyond that, the moment I say I need Ω, I am assuming
something.
Student: It’s a regular solution model
It is a regular solution model, somebody has said that! That the system is following a
regular solution model.
496
If the system is not following a regular solution model, one Ω is not sufficient for you.
You need to consider, if it is sub regular, so Ω A, ΩB and so on. So, if I assume that it is a
regular solution model, then simply, once I need know these three, at any given
temperature, I can simply calculate the whole free energy curve. And once I calculate the
free energy curve, simply I ask the same software to even calculate what is the ∂2G/∂x2 .
Find out that composition where that is 0.
497
And once you do that, you will be able to get these two spinodal points. And once you
do it as a function of temperature, at regular intervals of temperature, you will be able to
easily generate the spinodal curve. And this is something, which you will never be able
to get, if you simply do, let us say a thermal analysis. Miscibility gap, you may be able to
get, using thermal analysis, or to some extent, the more easier technique rather than
thermal analysis is what, if you want to see miscibility gap?
Student: X-ray
X-ray, X-ray diffraction because one single phase, splits into two phases, how will you
find out through x-ray diffraction?
Student: Superlattice
Superlattice? There is no superlattice here. There is no ordered structure here.
What will you see? If I take a single solid solution, which is splitting into two solid
solutions, when I do X ray diffraction, what will I see? One of you, Yes?
Student: 2 sets of peaks
Two sets of peaks you will get, of the same crystal structure, because the crystal
structure is not changing here. It is an α, of an FCC structure, splitting into two α, both
having the FCC structure, then what is the difference Aneesha? What is the difference,
between these two phases that are coming out?
Student: Composition
It is the composition. Once the composition is different, then obviously the lattice
parameter is different; if the lattice parameter is different, then the d spacings are
different; if the d spacings are different, then the θ is different- λ = 2dsinθ.
So, basically you would have got a set of peaks, whatever it is. If it is an FCC, you will
get 2, 1. 2, that kind of a pattern (see the white colored peaks in the image above). So
this is let us say, 1 1 1 peak, this is 2 0 0 peak, then 2 2 0 peak, what is the next one?
How many have you done X ray course?
Student: 3 1 1
3 1 1. What is the next one?
Student: 2 2 2
Yes. Perfect. 2 2 2. So, this is how, you get 2 1 2. I mean in fact, the moment anybody
gives you an X ray diffraction pattern, and wants you to check whether it is really FCC
or not, you will always see 2 peaks of them together, and one. And why does that
happen? Because h2 + k2 + l2 is 3, 4, 8, 11, 12.
498
So, 3 and 4 will be together, 8 will be separated; 11 and 12 will be together, and so on
and so forth. You will have again, 16, 19, and 20, and so on. So you will see always two-
two peaks will be together, separated by one peak, this is a characteristic X-ray. Now
this if it is single α, now it is splitting into two α’s, then you will see two peaks there
(Shown in pink color). One with a lower lattice parameter, another with a higher lattice
parameter, and similarly, this would also be, it may come as two individual peaks, or it
could be over lapping, because most of the time there will be
499
over lapping, because there the lattice parameters are so close, that they actually will be
over lapping.
And you actually need to do, what is called deconvolution, if anybody has done X ray
diffraction they will know. So, you need to do deconvolution of the peaks to be able to
see them separately. Or even do a very slow scan, of XRD to be able to see the two
peaks separately. So you will be able to see every peak, will split into two peaks, and we
will be able to clearly see, yes this is the two phase structure. And that is how you will be
able to see that the Spinodal has taken place. Now let us come to, again, back to our
Spinodal curve, and see what happens if I am inside the Spinodal, or outside the
Spinodal. In a Spinodal, we always talk whether you are inside the Spinodal or you are
outside the Spinodal.
What does that mean? It means that if I have 2 alloys, 1 alloy, I chose this composition,
1 (Fig. 1 above), another alloy I chose this composition, 2 (Fig.1 above). And both of
them are brought to this room temperature. They are cooled to that temperature. What
should happen, in both these alloys? First of all, in both these alloys, single phase α is
not stable. Because, both these alloys are inside the miscibility gap, so that means, α has
to split into, α1 + α2. The only question is, how it splits is the difference.
If I look at this alloy- alloy number 1. Alloy number 1, if I look at it, where does this
alloy number one lie, in the in the free energy diagram?
Student: Between binodal
500
Yes. It is between the binodal
501
and Spinodal. Am I right? So, that means, if I chose this composition on free energy
curve (Fig. 2 above), that is call it as alloy 1; it is between the binodal and Spinodal.
Now if I take this alloy, at this temperature, which is room temperature, the free energy
of single phase α is G1, which is higher than a two phase free energy, which is G1’?
If α  α1 + α2, if single phase α exists at that temperature, it would have free energy
G1; if it splits into α 1 + α 2, the free energy of that mixture will be G1’. So, it clearly
tells you that there is a decrease in the free energy, when single phase α splits into two
phases.
502
So that means, the single phase α will not remain as single phase α, it would like to split.
But how does this split takes place? If you look at, so, let us assume the way we have
seen in the last class, that if α is of a particular composition, we call it as alloy 1
composition, and if there is a small fluctuation in the composition that takes place,
because of the thermal vibrations, if such a composition fluctuation takes place (see the
Fig. above), what does it mean? In a small region, the composition if goes up beyond the
alloy composition nominal composition (1+), in some other region it comes down (1-).
That means, if I show that as one composition which has decreased in terms of the B
content, another composition which has increased in terms of the B content, if I look at
those two compositions, and then take now, if I treat 1 + as α1 and treat 1- as α2, and take
what is the free energy of the mixture of this α1 + α2, that free energy mixture will
always be identified by joining a line, between 1+ and 1-. So, this straight line which is
joining these two spots: one is α1 composition, another is α2 composition. The free
energy of such a mixture is G1”. So if α splits into two compositions, of α1 and α2, the
free energy, there is an increase in the free energy. So obviously, this α will not give α1
+ α2, by this process of splitting.
503
Then what is the way it can finally give you, α1 + α 2? Because it cannot remain as α, α
is not stable at that temperature. So, α somehow wants to become, α1 + α2. How does
that happen? To understand that, what you have to do is, at this particular composition 1,
if I draw a tangent: let us see I draw a tangent, for that alloy composition 1. Once I draw
a tangent, I see that, the compositions that can nucleate from α, that can come out of α,
which is stable, are only those compositions, which are beyond this point N.
Once it is beyond that point, all the free energies are below the tangent. That means,
there is a driving force a, so that means, the only composition that can, and if you take
any composition, which is near to that alloy composition 1: all the compositions near to
the alloy composition are always above the tangent. That means, the only way the α can
become a stable state is, that, it can precipitate something else, whose composition is far
off from its original composition.
And if something is coming out which has a composition, far off from the original
composition that can only happen by a nucleation and growth. So that means the α2, of
different-different compositions, can actually nucleate. And then the question is which
one you will nucleate? Always the one, with the highest driving force will try to
nucleate.
504
So that means, inside the α, there will be compositional fluctuations taking place at a
particular temperature. And in a small region, the B atoms come nearer, in such a way
that, once their composition becomes high enough in B, suddenly, that particular region
becomes α2.
And that α2, which gets nucleated here and there, now, it starts growing (Left figure).
Growing how? By the atoms, solute atoms, joining from the matrix to the precipitate and
this is exactly similar to our precipitation process, where we are talking of θ coming out
or β coming out from α. Exactly similarly, here, the only difference between this
process, and what we talked about precipitation, is what? There, the second phase that is
coming out would possibly have a different structure. θ, or some β, coming out of α, is a
different structure.
Here, the second phase that is coming out, will have the same structure as the first phase,
so, there is not major problem, in terms of coherency problems. The only coherency
problem comes is, only because, the lattice parameter is different. So, the strain will be
different. Structure is the same; lattice parameter is different. So the strain is still
involved. And, the interfacial energy problems will not be there much, because the
structure is the same. So it is α nucleating inside another α, so FCC nucleating inside
another FCC. Excepting the strain energy, surface energy problem will not be
505
there for it. So, it will nucleate at some regions: could be grain boundaries, could be
within the grain.
Usually, it occurs at the grain boundaries, because they are a high energy regions. The
moment α2 nucleates, then the α which is adjacent to that α2, can only have this
composition corresponding to α1. Because the moment α2 nucleates inside, the α which
is surrounding that α2 has to be in equilibrium with α2. So that is why, that α will
immediately change its composition to α1. And this is also physically easy to
understand, because, we said all the atoms, the solute atoms from a small region have
come together, and become rich in the solute, and then α2 has precipitated out.
So, surrounding that α2 obviously, the solute atoms get depleted. And that is how, you
see, that this α2, the moment it comes out, the α around it, will have a lower composition
in comparison to, the α which is far away from here. The α which is far away from α2
will have the starting composition which is the, we called it as alloy 1.
And this is exactly similar to precipitation that we have seen. The only difference being,
that the structure of α2, is same as that of α1. And, in fact, if you look at the G-P zones,
G-P zones situation is exactly like there. So there also, the structure of the G-P zones is
same as that of the matrix. So that, there the problem is, because the equilibrium phase is
not able to come out, being a different structure so, G-P zones which are transient phases
that are coming out. Here the equilibrium phase itself comes out, but the equilibrium
phase has a composition far away from the parent phase that you are starting with.
So, if you see one more precipitate somewhere at a different place, you will see, this kind
of a situation (shown in right figure in image above). So you have one precipitate,
another precipitate, between these two the α will have a higher composition, similar to
that of the alloy composition, but near the precipitate the composition is small. And if
this happens at time t = 0, as a function of time if you see, slowly the atoms, solute atoms
start moving in this direction (as indicated by the arrow).
And, what would be the result? There are two results: one, α2 size increases; second is
the composition of α, within the matrix will come down. So that means, you will see,
slowly α2 would grow, and this composition α1 would not change; it is
506
fixed for the temperature. And, only thing is, you would see, that this (composition of α
in the middle) would decrease.
And, slowly, after sometime, it would become only α1 and α2 (as shown in right figure
above).
So, the α, with which you have started, has actually got converted to what is called α 1
now. So between the two α2’s, what would remain is what is called α1, which was
originally called α. So, α1 would have a lower composition, α2 would have a higher
composition. So you will have a matrix of α1, inside which, you will have precipitates
of α2. That is how the whole micro structure look like after the whole process is finished.
This is for the alloy 1. If I now chose alloy 2; alloy 2, would undergo the same
decomposition in an entirely different fashion. The reason is simple. Alloy 2 is falling
within the Spinodal. So, the composition that I am choosing is between the 2 Spinodal
points. So, then I can say that this alloy is, within the Spinodal, whereas alloy 1 is
outside the Spinodal, for that particular temperature (indicated in Fig. 1, Refer slide time
18:12)
So, the alloy composition 2 is between the two spinodal points. If that is the case, in the
free energy curve, then this one, for it to split into 2 phases, is very easy. That means, if I
draw a tangent, at that composition 2, if I draw a tangent at that composition, the
compositions which are adjacent
507
to that alloy composition 2 can nucleate, because there is a driving force, any
composition.
So if I draw a tangent here, to that alloy composition 2, any small deviation from this
alloy composition has a lower free energy, with respect to the tangent. And that is the
reason why, a small fluctuation, with one region becoming depleted in solute, another
region becoming enriched in solute, can be stable: because that would cause a decrease
in the free energy with respect to the alloy.
But if you remember the overall driving force, if I again redraw this, for your sake.
These are the two Spinodal points, L and N; we have chosen alloy 1, in which, we said
the driving force is ΔG1. In the other alloy 2, which we have chosen inside the spinodal,
the driving force for decomposition is ΔG1. The overall driving force is very large. Not
only that overall driving force is very large, but also, because the composition is falling
within the Spinodal, a small fluctuation in the composition, can be a stable.
508
And then, this fluctuation will start growing further and further, until, the composition
reaches α1 and α2. So, it starts with a small fluctuation, and the both the fluctuations
start becoming bigger and bigger, at every stage there is a decrease in the free energy,
and the final driving force is reached, as a function of time.
And, as a result, you can see, this fluctuation grows and grows (see the bottom figure in
above image).
Student: Sir there why wavelength is constant?
The wave length is, whatever is the initial wave length. Once there is some wave that is
created, that wave depends on that diffusivity at that particular temperature. Once a
certain fluctuation wave is created, that wave length gets fixed and the only the
amplitude changes. Because the atomic jumps, at any given temperature, are fixed.
Depends on the structure, and also depends on the temperature.
So, once the atomic jumps are fixed, then this wave length gets fixed. It depends on that
particular temperature and for the particular structure, whatever structure that we are
talking about. And once that happens, you will see that amplitude keeps on icnreasing. If
you compare this with the previous one, the final stage is the same in both the cases. The
only difference is, the distance between the two precipitates in a normal case of alloy 1,
could be much larger.
And this, the precipitate, I mean if I want to call them as precipitates, they are not really
precipitates, the second phase that is nucleating out of it, can be precipitating, can be
509
forming at the grain boundaries or some high energy regions. Whereas here for alloy 2,
everything
510
happens homogeneously within the material. And, there is no activation barrier for that.
That is why we say, Spinodal decomposition there is no activation barrier. Because it is a
small fluctuation, thermal fluctuations at any temperature, will take place so, and as a
result, there is no activation barrier for spinodal decomposition. And in fact, there is a
way to calculate the activation barrier also, I mean measure the activation barrier, people
do it regularly.
One way to show the activation barrier is: if you have a region like this, if I take an alloy
draw a tangent, for this alloy, this alloy is, let us call it as alloy 1, which is outside the
Spinodal. And, if you want to see, what is the activation barrier, for the nucleation for
this alloy: activation barrier, usually people obtain, by drawing a parallel, to this tangent.
If you draw a parallel to this tangent, so that you get another tangent which is, touching
one of the composition the distance between, the vertical distance between these two
tangents, would indicate the activation barrier. One can derive this, okay, let us not go
into the greater details of it, but, that gives you an idea of what is the activation barrier.
And as you shift your composition, towards the Spinodal: At the Spinodal, if you draw a
tangent, okay, again try to find out the parallel to that, you would see that the distance
vanishes to 0.
511
So, as a result, both the tangents would actually merge into each other. So, you would
see that, this vertical distance between tangents, as I go from alloy 1 to inside spinodal, it
will keep on decreasing and finally it merges. That is one way to graphically show that.
But physically we can easily understand, that activation energy does not, whatever
activation energy that is there for diffusion, will still remain, Okay because diffusion has
to take place. So, that activation energy will be there, but there is no activation energy
for the nucleation of the phases, because there is nothing nucleating. This is not a
nucleation event, unlike the alloy 1 case, where, something has to nucleate. And, its
composition is entirely different from this. So, all the atoms have to travel. So, there is a
lot of diffusion that is involved, long range diffusion that is involved. Here in spinodal it
is all a short range diffusion. And as a result, the activation energy for it is 0.
And as a result, even in the micro structure also, would look entirely different. A
Spinodal decomposed micro structure, if you carefully observe, it will be
homogeneously distributed; different you know gray levels. One is slightly higher
composition, another is, if you really try to etch the surface let us say, with some etchant,
and if this etchant is sensitive to the composition, you will be able to see the gray and
darker regions, inside the thing on a very, very, very small scale, nano scale. Obviously,
you do not have an optical microscope to show that, so, the only way we do is, we go to
an electron microscope. In an electron microscope, how do I different regions of
different compositions?
Student: SEM
Of course, yeah, if you can use a SEM, scanning electron microscope, luckily now we
have SEMs that are available which can really go the nano meter levels, which are called
FEG- SEMs. Otherwise the older generation SEMs, you would not be able to see really a
10 nano meter kind of a modulation inside, you will not be able to see whereas, in a FEG
SEM it is possible.
But, if you want to go to a, transmission electron microscope, and try to see, how do you
see? Where does the contrast come from, in a transmission electron microscope?
Anybody did any course on microscopy? No? Nobody knows? Why do you see some
contrast in a TEM? Why some regions look black, why some regions look white?
Students: Scattering of electrons
There are two types of scattering, in an electron microscope, in a TEM. One is called
diffraction contrast, okay; another is called phase contrast, okay. Diffraction contrast
512
comes, wherever the atomic plane, in that particular phase, is properly oriented with
respect to the beam, such that the Bragg’s law
513
is satisfied, it gives you a dark region, because it gets diffracted. Otherwise, there is
another way to look at it, which is what is called the phase contrast, where you see the
regions, which have a different composition. The moment you have a different
composition, obviously, it also means, that it has different concentration of electrons.
For example, if you take an aluminium copper region. One is aluminium rich region,
another is copper rich region. Copper has different valence electrons, aluminium has
different valence electrons. So, as a result, the region where you have, copper rich
regions, will have different electron concentration. If there is a different electron
concentration, the electron beam which is going through, gets scattered by the electrons
present inside that. So, as a result, wherever you have more electrons, that region looks
darker, wherever you have less electrons, that region looks brighter. So, you will be able
to see that, on a very small scale.
So, in TEM, one can easily observe such kind of modulations. Even at an early stage,
you will be able to see, before the whole the wave grows to the final stage. But once it
goes to the final stage, it almost looks like precipitates inside a matrix. But these
precipitates, the only difference between the alloy 1 and alloy 2 is, they will be
uniformly distributed, within the grain. It is almost like, a normal precipitation
happening within the grain. Excepting that, a normal precipitation and this also are
different. In which way they will be different?
A precipitation that we talked about it, where G-P zones first come, and then you have a
θ”, θ’ and finally it becomes θ. If you look at that micro structure and a Spinodal micro
structure, would you be able to see some difference?
Student: Shape
Shape, perfectly, yes. Most of the cases, this kind of Spinodal decomposition would be,
more or less like, you would not be able to see the borders. The interface is not very,
very sharp. The interface is sharp there, because there is a different structure that comes
out.
So, you will be able to see the differences very easily, because the interphases are not
sharper, and the structures are same. You can easily do a diffraction from both the
regions, and be able to see, that both the phases have the exactly the same structure.
Anyway, in case of actually, a normal precipitate, the moment G-P zones grow to θ”,
immediately the structure changes. And, at the G-P zones stage,
514
you will not be able to see anything. The only way to see is, by the strain contrast.
Because there is a lattice strain associated with that, and that is what we will see in a
minute here also, there is also a strain that is associated.
The moment you talk about a different composition, different composition means
different lattice parameter. So, that means, in a particular region, R1: this region will
have a different lattice parameter; in fact, that lattice parameter in R1 also is changing as
a function of distance, because the composition is changing.
Similarly, in region R2 also the composition is changing. So, as a result R1 and R2, there
will be always, at the interface, some kind of strain. This is what we call it as coherency
strains, similar to what we talk about coherency strains in precipitation. And because of
this coherency strains, though you have a driving force, for the α to split into two, the
coherency strains would be a barrier for this split. As a result, usually, just because I
have brought an alloy within the Spinodal, does not mean that always this alloy will
undergo Spinodal decomposition.
So, you need an extra energy. That means, you need a higher under cooling. If you
provide sufficient under cooling then, you will have a sufficient driving force, for the
alloy to split into two, by a Spinodal decomposition. And as a result, you would always
see that there is another curve, which we usually refer to as Coherent Spinodal curve.
That
515
means, if I take the alloy 2, in principle, the moment I reach below this particular point
(P, see the image below), Spinodal decomposition should start. But just because I have
reached this point, Spinodal decomposition does not start unless, you come to another
point, here (P’), where you have sufficient driving force to be able to take care of the
coherency strains.
So, unless you come below this particular point P’, you will not be able to see the
Spinodal decomposition taking place. So, there is a separate curve, a Spinodal curve,
which takes into account the coherency strains and that particular Spinodal curve, is
what we call it as a Coherent Spinodal curve (drawn in yellow). And, in contrast, the
other curve is what is called Chemical Spinodal (drawn in pink). Because, there, you are
only looking at composition, you are not considering any strains there. So, this Chemical
Spinodal is something which comes simply from ∂2G/∂x2 = 0.
So, that is the Chemical Spinodal, and there, if you add the strain energy contribution,
then, you will see that the free energy expression itself you have to change now. The free
energy curve actually, because of the strain goes up, and then you would see, the point P
comes down. And, as a result, in principle, there may be cases for example, if I now
carefully chose an alloy between the Chemical Spinodal and Coherent Spinodal. That
516
alloy, if I bring it to room temperature, though I have brought it to room temperature,
because it is between the two Spinodals, it may not undergo
517
Spinodal decomposition, until you bring it to a sub 0 temperature, so that, it crosses the
Coherent Spinodal, and then only Spinodal decomposition will take place.
Then, what should happen to this alloy? This alloy also, would undergo decomposition
like alloy 1. Though, you are inside the Spinodal, it will not undergo a Spinodal
decomposition, because of the strains that are involved. And If the strains are very large,
then it would prefer, rather than going through a Spinodal decomposition, it would prefer
nucleating an α2, which has a different composition and then growth of that α2, rather
than a Spinodal decomposition.
So, that is why we need to have an understanding of this Coherent Spinodal. And this
Coherent Spinodal, basically, the extent of which, whether it is very close to the
Chemical Spinodal or far off from the Chemical Spinodal? Because in some systems, the
Coherent Spinodal can be far away from the Chemical Spinodal. Why? It depends on,
how much is the strain energy. For each system, as the lattice parameter changes the
strain changes. The rate of change of strain, with the lattice parameter decides the
position of the Coherent Spinodal.
In some systems, the strain energy may be very small. In such a case Coherent Spinodal
and Chemical Spinodal will be almost close to each other, you will not be able
differentiate them. So, that is how, we need to have an understanding of the strain energy
term, in a particular system and it depends on different-different crystal structures it will
be different, and that gives you an idea of what is the coherency Spinodal. So, and once
you are within the Spinodal, the Spinodal decomposition takes place. Okay.
We will stop with this, and then continue later. Any doubts here? So far whatever we
have said. You may ask me, what is the use of this Spinodal decomposition? In fact,
people do not want Spinodal decomposition to take place. The reason is the moment you
have Spinodal decomposition, the alloy splits into two compositions. So, people want to
avoid Spinodal decomposition. So, they want to choose those systems, where there is no
miscibility gap, or choose alloy compositions, which do not fall into the Spinodal
decomposition. So that you can have a precipitate taking place, rather than, overall
composition is splitting into two.
That means, in a small scale, , the properties are going to be different. So, you have an
inhomogeneous property inside the material.
Student: But that take for strengthening mechanism
518
It depends. Of
519
course, one it can be useful, but then strengthening mechanism, as a Spinodal
decomposition may not be really a strengthening, because it is also a solid solution. It is
basically α is splitting into α1 + α2, both are FCC. So, the α2 that is coming out is not a
hard phase any way. It is a solid solution; does not give you an exceptional high strength.
So instead, you chose a particular system, where you have a precipitate with a different
structure coming out that can give you a higher strengthening.
Student: In the first case, how does it choose the composition for decomposition?
Yeah. How does it choose the composition? So, as I told you, when you draw a tangent,
it can have many different compositions possible (which have lower Free energy). Out of
these compositions, there are two problems. One is it wants to choose, let us say, the one
with the highest driving force (point C1). But there is a problem for this to come out. In
principle, a phase which is close to this composition (C0), should come out, because, the
amount of deviation from the alloy composition that is needed, for this phase (C2) to
come is smaller, than the amount of deviation that is required for the composition C1.
That means, if you are starting from an alloy composition of let us say C0 = 10 % B, if
C2 is let us say 30% B, and if C1 is 50% B, let us say.
So that means, as the atoms accumulate in a small region, and only the moment 30%
accumulate, in principle that is enough for the second to nucleate. Here for C1, you need
more accumulation to happen. That means atoms have to come together more, to be able
to do that. But, it will always have the highest driving force, because the driving force is
higher. And, thermal fluctuations always anyway take place. So, because of the driving
force is higher, if you remember, the nucleation kinetics, the ΔG* is inversely
proportional to the square of the driving force. So, as a result, you would see that this
would tend to form.
Student: Cant we expect this 30% to have a low activation energy?
Obviously, if you look from the point of view of only the thermal movement, the atomic
vibrations alone, this is true.
So, if you simply look at only from the atom view of atomic movement, it is fine. But if
you look at, what is the r*, for this thirty percent; what is the r*, for the 50%? r* for the
50% will be smallest, because the ΔG is larger, r* is smaller. So, you need a bigger r*, if
30% has to nucleate. You need a smaller r*, if 50% has to nucleate. Though you know it
is like a, you know kind
520
of a managing, so you need a greater deviation, but the extent of the greater deviation, a
smaller is enough. So, the system would choose that one. So, it is a very, very interesting
situations if you look at it. I do not know how really system chooses, at least to
understand it is really interesting. Okay, we will stop here.
521
Lecture. #22
Eutectoid reaction
So, we have been talking so far on basically solid state phase transformations you have
been seeing. We took up the precipitation and dealt with it so far, both the precipitation
per se and also coarsening of precipitates this also and of course nucleation. We did not
go too much into the growth of it, because the growth stage basically it is more of
kinetics. So, we did not talk much about that, but if somebody is interested may be as we
go along we will talk about those issues also, but because this course is more on
thermodynamics, I thought I would stick to thermo dynamics aspects of it.
Then, we went to the other kind of extension of precipitation which is spinodal. So, we
have been talking on spinodal so far, spinodal decomposition, and how it is different
from normal precipitation and how does an alloy which is outside the spinodal behave
and how does an alloy which is inside the spinodal behave, in terms of their
decomposition characteristics.
522
We also looked at what is called coherent and chemical spinodal, and looked at the
differences between the two. And we also talked to some extent on the activation energy
for the spinodal decomposition, and looked at how the compositional profiles change as
a function of time, whether it is the alloy which is inside the spinodal or outside the
spinodal, this is what we have talked so far. Any clarifications that you need on that so
far, otherwise we will now go to the other type of solid state phase transformation.
For example, eutectoid, the most common phase transformation is eutectoid for all of us
mostly metallurgist, because we deal with it in steels all of you know. What is the
eutectoid transformation in steels? What happens during eutectoid transformation?
Student: Solid transforms to two other solids
What is it, when I say steels that means
Student: 0.07% C
What happen, changes to what?
Student: Austenite to Ferrite + cementite
Yes cementite good. So, it is austenite which is refer to as γ transforms to ferrite, which
is refer to as α to cementite, this is what we see and in general her statement is right.
γ  α + Fe3C
That one solid gives you two solids and that is what is a eutectoid transformation and
this transformation if you carefully observe in comparison to the precipitation. In
precipitation you have a matrix from which a precipitate is coming out and this
523
precipitate has either a different structure or the same structure. For example, if you are
talking of spinodal decomposition where the alloy is outside the spinodal, then you have
basically a precipitate coming out of a matrix of the same structure
524
excepting that the structure the composition is different. You can have a situation where
the structure is the same or a situation where the structure is different, but definitely the
precipitate has a different composition when compared to the matrix in all precipitation
you will see this.
The only difference could be either the structure is same or structure is not same with respect
to the matrix. If the structure is also different, then the question comes is whether the
precipitate is able to nucleate easily in the matrix or not and if it is not able to nucleate easily
in the matrix then you have transient precipitates. If the structure of the precipitate is exactly
same as the structure of the matrix, you do not need transient precipitates excepting, again in
some cases you may need transient precipitates, when do you need it? When the structure of
the precipitate is same as that of the matrix, when the lattice parameter is significantly
different. Whenever the lattice parameter is significantly different, there can be a necessity
of a transient phases, but you rarely see this happening.
Wherever you have the structure of the for example, simple example that we can take is
a Ni3Al in a nickel base super alloy or Al 3Ni. In fact, if you carefully look at aluminium
lithium phase diagram, what is special about aluminium lithium? Tell me?
Student: High specific strength
High specific strength what is special about lithium? Who knows?
Student: Very light
High, very, very light, yes, it is the lightest metal its density is 0.53 if I remember right.
So, it floats in water in principle, so in principle so when you add lithium to aluminium,
aluminium itself we use it, because of low density because all aerospace materials you
want a lower density material. So, if you add lithium obviously, the density can further
come down and incidentally people have also observed when you add lithium to
aluminium the young’s modulus increases.
As a result, so the strength is going up the density is coming down. So, the ratio of this
which is what he talked about as specific strength, specific modulus is going to be much
higher. The numerator has gone up the denominator has come down. So, obviously the
modulus by density or strength by density if you calculate you are going to have higher.
525
But there also if you look at aluminium lithium phase diagram, where you see a phase
called AlLi, the stable phase in an aluminium lithium diagram, in fact, large number of
people have worked on aluminium lithium, aerospace industries to see, if they can have a
replacement for aluminium copper alloys. So, that you can have much lighter, but there
are problems. First problem that comes is the lithium. Lithium being the lightest metal
and it has a very small atomic size. First and foremost is lithium is reactive. Handling
lithium is not easy; you cannot simply keep lithium outside the atmosphere. So, that is
one, but once you put it inside the aluminium you make a alloy out of it, but even after
the alloy is made the lithium has a tendency to diffuse out being a smaller element.
So, if you take an alloy of a rod like this aluminium lithium rod after a month, if you see
the surface concentration of lithium is going to be higher than the concentration inside.
So it has diffused to the surface and as a result and once it diffuses to the surface there is
a possibility that it can react with the environment, but more importantly your
composition has changed, alloy composition. So obviously, whatever precipitates that
you have inside the matrix they are going to be reduced, the precipitate concentration or
precipitate volume fraction comes down, because the lithium content has decreased.
But, people try to overcome that by coatings, by cladding, variety of things people have
done, aluminium cladding and things like that. But more importantly the problem was
what is called, how many of you read about aluminium lithium? There is a problem, in
526
aluminium lithium alloys particularly which is associated with deformation
characteristics. Have you ever heard of a planar slip? Whenever you have the
precipitates inside a matrix and the moment a precipitate is cut by the dislocations, there
is a tendency for the rest of the dislocations to find a easy path. This is what is called as
planar slip.
And this tendency people have seen it in aluminium lithium alloys to a great extent and
as a result once some of the precipitates are cut you can see easy deformation taking
place and the fracture toughness of this alloy, people found out to be very-very low. So,
processing of the alloy was very difficult and to some extent again there that is where
you know the whole alloy design becomes so interesting. How do I now prevent this
precipitate to cut, if the planar slip is taking place or what people call it as slip
localization. So, if slip localization which is happening because of the precipitate being
cut, how do I avoid it. The best way is to make the precipitate in such a way that it does
not get cut, but how do I change a particular precipitate. So, that it does not get cut.
So what people have done is they try to put something inside that precipitate. So, what
they have done is they have added certain precipitate, certain you know heterogeneous
nucleating sites. For example, Al3Zr is used as a nucleation site for the Al 3Li.
Incidentally when I started I said in aluminium lithium phase diagram, AlLi is the stable
phase, but this is not the one that you get when you do age hardening of aluminium
lithium alloys.
What you get is this phase Al3Li exactly for the same reason why you get GP zones in
aluminium copper alloys that Al3Li is exactly similar to Ni3Al. It is what we call it as
L12 structure. L12 structure some of you have heard of basically, it is an ordered structure
based on FCC where you will see that three atoms, the three lattice points which are the
face centered lattice points are occupied by one type of atom and the body corner lattice
points are occupied by another type of atom , Ni 3Al. So, you can easily see. So, here
aluminium will occupy the face centers, lithium will occupy the body corner and that is
how you get an ordered structure and what they saw is because Al 3Zr is also cubic it can
easily nucleate Al3Li and once Al3Li nucleates on that. So, inside the Al3Li you have a
Al3Zr. So, every Al3Li particle will have almost like one Al3Zr particle inside and once
you have that kind of a situation, you would see this planar slip or the slip localization
becomes difficult.
527
Because Al3Zr, the bond between aluminium and zirconium is much stronger there, the
ordering energy here is much stronger than the ordering energy here first of all Al 3Li is
a metastable phase, whereas Al3Zr is not a metastable phase. In fact, how do I know
about the ordering energy? You will see that, whenever you see an ordered phase usually
an ordered phase at high temperature is disordered. Why does it get disordered?
Student: More entropy
More entropy, from the thermodynamic point of view, we can easily talk in fact we will
have one session on ordered structures and how to understand them from the
thermodynamic point of view, we will go into it.
So, once you look at an ordered structure usually it transforms to a disordered structure
at higher temperature. There is a particular critical temperature, where this happens
which is what we call it as ordering temperature or disordering temperature whatever it
is and this temperature is a function of how strongly bonded the two elements are, which
is basically related to the interaction parameter the omega. If the omega between
aluminium and other the solute element is very-very largely negative, then you would
see this disordering temperature will be very high that basically means when an ordered
structure becomes a disordered structure means the bonds are broken, the strong bonds
which are the ordered bonds are broken. So that means, when you have a lower bond
energies then you will see.
For example, copper zinc is a very easy example have you heard of β brass. β brass for
example at lower temperature is ordered and above something like 450 ºC or so it
becomes disordered. β’ and β that is what we talk about and this again if I compare NiAl
or Ni3Al these structures are stable as ordered structures come to very high temperature.
In fact, NiAl it becomes disordered only after it melts. So, up to its melting point it is
both chemically and structurally ordered.
Once it becomes liquid obviously, in a liquid there cannot be an order, the only order
that you can see in a liquid is what, is a short range order that is it. So it is structurally
also disordered and obviously, chemically also becomes disordered, but within the short
range there can be a chemical order is possible, otherwise if you take a NiAl, 50 50
liquid and when you make it solid that means solidify it, it automatically becomes
ordered NiAl. That means, there is a chemical order inside the short range clusters and
those clusters start growing the nucleus forms and then grows all these things.
528
So that means, there could be in a short range order a chemical ordering also, but any
way we do not really call it as a big order, because whenever we refer to ordering we
always mean a long range ordering. Short range order is not really considered as a very
important thing particularly basically, because it is not contributing to a large extent to
the mechanical properties of a material. But it decides what kind of structures will come
out of it, the short range order inside a liquid is going to decide whether an FCC is going
to come out or BCC is going to come out when it solidifies ultimately.
So, what I mean to say is when you look at structures like this the fact that Al 3Li is a
metastable phase indicates that this ordered structure and if you heat it to higher
temperature Al3Li vanishes and finally, gives you the AlLi; that means, Al 3Li in
principle would split into Al Li + Al. So obviously, you can see for every lithium atom
there are three aluminium atoms. So, two extra aluminium atoms are there so that two
extra aluminium atoms will come out. This is something like Fe 3C splitting into iron and
carbon at higher temperature.
So, Similarly, this can happen at high temperature; that means Al 3Li is not really a stable
phase. So, whereas Al3Zr is a very stable phase and because it is stable phase and that is
the reason why whenever I take a disordered structure and an ordered structure and try to
deform both of them you see a ordered structure, the strength is much higher; the stress
that is necessary to deform an ordered structure is very high. Why?
It is a simple example I take a nickel base solid solution and a Ni 3Al compound and try
to deform both of them or take the same Ni3Al compound and make it disordered. And
once it becomes disordered or for example, take copper zinc β brass and take a
disordered β brass and an ordered β brass and try to deform both of them. You will see
that the ordered β brass would have a higher strength than the disordered β brass. Both
are exactly the same composition, both have the same structure, the base structure is the
same then what?
Student: Difference in bonding
Difference in bonding so, how does it reflects? In slip what is that which is a very
important parameter that decides what is the yield strength?
Student: Number of slip systems
Slip systems beyond that?
Student: Critical resolved shear stress
Critical resolved shear stress and anything else.
529
Student: Atomic distance should be minimum
What do you call that in during slip?
Student: Close packed plane
There is a, there is a particular parameter that we use.
Student: Burgers vector
Burgers vector, whenever the burgers vector is smaller the slip is easier of course, the
other parameters are there. When you talk about critical resolved shear stress, Critical
530
resolved shear stress will be higher obviously; because bonds are stronger, where does
this critical resolved shear stress comes? Because a dislocation motion is decided by
bond breaking and bond reconstruction, unless the two bonds break the dislocation
cannot move. And then this bond gets attached to another atomic plane. So, it is if you
look at how a dislocation moves it is all bond breaking and reconstruction. Am I right?
So, for that you need to consider the bond energies as Aneesha has pointed out, but in
addition if you look at the second issue, which is the burgers vector in a ordered
structure, the burgers vector is much larger than a disordered structure. I want you to
think about it and come back in the next class and tell me why it is so? Why burgers
vector in an ordered structure is long larger? For example, if I take a FCC structure, what
is the burgers vector in a FCC structure?
Student: a/2<110>.
a/2<110>., Am I right? So, now you look at a/2<110> in an ordered FCC structure, in a
disordered FCC structure and tell me what would it be. Think about it, I will leave it to
you. Anyway, so more importantly that is one of the reason, why you see the
deformation in ordered structures is more difficult. So, the burgers vector being larger
the reconstruction of the bonds becomes difficult. So, dislocation motion becomes more
difficult. So, the moment you convert a disordered structure into an ordered structure you
make it stronger any way. So, and on top of that if the bonds are stronger, you will get
much higher strength. Anyway so this is all related to the precipitation. So, the reason I
had to come to this is Al3Li kind of metastable phases can come out of it provided a
stable structure (AlLi) is more difficult to nucleate and you have in numerous examples
of it, one example is aluminium copper itself.
531
So, now if you come out of a normal precipitation, and look at this case
γ  α + Fe3C
In this particular case where the difference between this, and precipitation that we talked
about so far is that in this the two phases that are coming out or having different
compositions, and in a number of cases they are different crystal structures. In fact, in
most of the cases they are different crystal structures. There can be some cases where the
structure of one of the phase could be the same as the matrix, but in a number of cases
you will see the structure is also different. So, obviously that means structure of
products, and composition of products is different from the matrix or the parent phase,
we call it as a parent phase. The moment you have this situation obviously, you will
know that the activation energy for such a transformation is going to be higher.
So, when compared to a normal precipitation you would see that this is more difficult.
So, that means you need more under cooling and so obviously that means, if you talk in
terms of a precipitation; we the system tries to avoid going through a process of higher
activation energy by choosing something which has a lower activation energy. That is
where transient phases are come into picture, GP zones and things like that, here you
have no choice. Here your phases are fixed; your eutectoid phases are fixed. So, the
product phases are fixed.
Once the product phases are fixed, automatically their structures are fixed, their
compositions are fixed. So, as a result the system has no choice of choosing something
532
else. So, it has to go through only that process if it has to go through that process then
that means the activation energy for such a process depends on what? Can we still
change the activation energy for such process? How do we change the activation energy?
Can I reduce the activation energy for such a transformation?
Student: Having the nucleation, having the same structure of all
Nucleation? Ya, one way is to have certain nuclei already available. When do I have
such a nuclei already available, if I am choosing an alloy which is a hypo peritectoid,
eutectoid or hyper eutectoid.
Once, I chose compositions which are hypo or hyper that means some one of these two
phases is already there inside the structure obviously. So, in a hypo you have α that
already forms, ferrite forms from the austenite, in case of hyper cementite form from the
austenite. So, before you reach the eutectoid you already have one of these two phases
already being present. So, that phase which is already there can nucleate this and as a
result the activation energy can be lower, but if I am choosing the eutectoid composition
exactly I do not have this choice with me there is nothing, before the only thing that is
before is austenite. So, then if I want to reduce the activation barrier, how do I reduce the
activation barrier?
Student: Giving some other alloying elements to austenite
Giving, some other alloying elements, interesting very good sometimes yes in fact
people have seen that you all heard about hardenability.
Anyway, the extension of this is going to be hardenability we will talk about it possibly
today or in the next class. Hardenability there are different-different elements which
either aid or do not aid in increasing hardenability. There are certain elements which we
say that they decrease hardenability, have you heard of such cases. Cobalt, why does it
decrease?
Student: Shifts the TTT curve to left
Correct. Why does it shift? These are the issues.
There are a number of cases where the TTT diagram can shift. For example, there are
cases where if I take a high alloy steel, let us say vanadium or chromium or any such
steels which are carbide forming elements are present in steels. You always see that such
an alloy precipitates; I mean such an alloy undergoes eutectoid transformation very
easily. Why? For example, in such a case, for example in tool steels which are suppose
533
to be very hard. In tool steels you use a very high austenization temperature. Do you
know why?
Student: To dissolve carbides
To dissolve? Carbon? Carbides. Why? Why do you want to dissolve all the carbides?
Student: Carbon has to go into the solution
Carbon need not have to go into the solution, the reason is the moment carbides are there
then austenite can undergo eutectoid reaction easily
534
because, these carbides can act as nucleating sites for the cementite. So, whenever you
have carbides present that is why, we say any element increases hardenability or shifts
the TTT diagram to the right, only when it dissolves inside the austenite. If it does not
dissolve inside the austenite, you will never see the TTT diagram shifted to the right.
When it dissolves inside the austenite, why does the TTT diagram shift to the right? The
same alloy, same you know high speed steel containing tungsten and so on. I heat it to a
very high austenization temperature sometimes it can be even 1100 ºC.
So, when you heat it to such high temperatures everything dissolves into it then so what
then, when I am cooling this austenite. How does it prevent nucleation of the eutectoid
mixture?
Student: Reduces the diffusion coefficient
Reduces the diffusion coefficient, whenever you have alloying elements particularly
alloying elements with a higher melting point then iron present inside it the diffusivities
the self-diffusion or the chemical diffusivities are going to be slower. Because, every
atom that is present in a substitutional or an interstitial site is going to make the diffusion
of iron more difficult.
Because, the moment the solute is present there is always a strain around the solute, why
because the solute does not have the same atomic size as that of the matrix. Obviously, it
is impossible to have a solute which has exactly the same atomic size as that of the
matrix. Because the moment you have the atomic number being different, then atomic
size is different. Once the atomic size is different then there is either a compressive strain
or a tensile strain depending on whether it is smaller or bigger than the atom.
So, you have always have a strain and this strain affects the diffusion. So, at any given
temperature the diffusivity is going to be lower and why diffusivity is important, because
nucleation needs diffusion that is the reason why you actually see the TTT diagram
shifted. So, but whenever I add elements which form carbides and remain as carbides
they can easily nucleate the pearlite and that is the reason why if you want a high
hardenability and if you want to get martensite in such high speed steels, you are bound
to go to a higher temperature and that has its own problems. What are the problems?
Student: Oxidation, grain growth
535
Oxidation, grain growth, but grain growth also sometimes is useful for you. What is the
use of grain growth? Tell me, what happens if I have a large grain size. We are talking
about hardenability.
536
In this context do you know?
Student: Less no. of nucleating sites
less number of nucleating sites for the pearlite. So, a coarse grain size is always useful if
you want a high hardenability, but that is not the whole life. A coarse grain structure
always have problems also in terms of strengthening, room temperature strengthening is
going to be lower. Obviously, if you want to use the alloy at a lower temperature, it is
going to be a lower and a coarse grain structure also has other problems related to the
toughness. Toughness is going to be lower for a coarse grain structure.
And in addition even quench cracks also are much more in a coarse grain structure. Of
course, that is not really related to coarse grain structure, it is because of higher
temperature to which you have gone. So, when you are quenching it, obviously and that
is the reason why we do not actually quench such alloys, they are all mostly air
hardened. So, because the TTT diagram shifts to such an extent, that they are all air
hardenable and because they are air hardenable you do not really see a significant
quench cracks.
But, if you want if you do not want to go to higher temperature, to dissolve all of them
and still get a martensite that means you need to use a higher quench rates, because the
TTT diagram has already shifted to the left. And if you want to avoid the TTT curve and
get a martensite you need to use a higher cooling rates and that can lead to problems of
quench cracks. So, these are all the issues that one has to worry and understand then only
we become a good physical metallurgist. So, as you can see that this is one of the major
issues that this difficulty in the precipitation can be to some extent overcome, by as they
said putting things which can nucleate or giving a higher under cooling.
The higher under cooling can lead you to a higher driving force and once you have a
higher driving force at least up to the nose. We will talk about it as we go along that
there is a possibility that if you under cool it. If you can come to a lower and lower
temperature you can have a higher driving force.
537
And if you can have a higher driving force obviously you have a lower activation
energy, because we know that the activation energy for nucleation, ΔG α 1/ΔGv2, and we
also know that ΔGv α Δ T. All this we have done and the principles remain the same
whether it is liquid to solid or solid to solid. Excepting that in a solid to solid we need to
replace ΔGv part with the combination of volume free energy and strain energy.
Yes and the surface energy that we talk about it is going to be a combination of various
surfaces, because various interfaces can be possible depending on where this particular
phase is nucleating. So, as long as we know that one can consider this. So, these are
some of the issues which are not directly related to thermodynamics, but one need to
understand them, if you want to understand the whole eutectoid transformation.
538
Now, let us go into the thermodynamics of it. If I look at a phase diagram having a
eutectoid transformation,. So, this is γ  α + Fe3C, cementite is there somewhere.. What
is this eutectoid temperature?
Student: 723
723 ºC and what is this eutectoid composition?
Student: 0.8
0.8, by the way have you ever thought about you have an α and you have a γ, you do not
have a β here.
Student: It is a non magnetic form
It is a non-magnetic form of α. So obviously, it is there and at what temperature it goes
off?
Student: 768 ºC.
So, 768 ºC is the temperature where you see a α to β transition which is ferromagnetic to
paramagnetic transition. So, that is why in some phase diagrams you see a dotted line
around 768 ºC. So, that indicates you of that anyway.
So, now we are talking of eutectoid transformation, if you take a grain of γ let us say, γ
and an α + Fe3C is nucleating (See the image above) and we can draw a lamellae of α +
Fe3C. There is always a botheration what nucleates first. So, many people broke their
heads in this particular context, because always when you talk of two phases coming out
whether it will be α first or Fe3C first?
Student: What difference does it make?
539
That is true, but is ultimately you may say that as far as the properties are concerned, I
am not bothered about what comes out first.
Because properties depend on what in this particular case, in a pearlite structure what do
the properties depend on?
Student: Inter lamellar spacing
Inter lamellar spacing anything else? Only the inter
540
lamellar spacing that is enough?
Student: Amount of Fe3C
Amount of Fe3C is fixed in a pearlite,. In pearlite, once I say pearlite you can do a what
is called lever rule and then find out and that turns out to be approximately around how
much 12%. Something like that 12 % of cementite
Student: Shape
Shape of? Once, I say pearlite I am defining the shape, pearlite is always lamellar you
can have a non lamellar iron cementite mixtures, what are they called non lamellar iron
cementite mixtures, ferrite cementite mixtures.
Student: Spherodite
Spherodite, very good any other non lamellar ferrite cementite mixtures,
Student: Banite
Banite, again you have two types of banites upper and lower, we will talk about them
may be as we go along.
So, they are all non lamellar. So, when I say pearlite I define that it is lamellar. So, once
it is lamellar. So, the other point which decides the strength is the pearlite colony size.
How is that decided? What decides a pearlite colony size?
Student: Number of nucleating sites
Number of nucleating sites and number of nucleating sites are decided by your under
cooling, your under cooling and of course, the grain size also, the prior austenitic grain
size all these things control. So, if your under cooling is higher, under cooling controls
both the factors, one the pearlite colony size it controls because the number of nuclei.
541
So, if you imagine that in one grain if you have different-different nucleating sites. So,
each of them will have a combination of a couple, ferrite- cementite -cementite or
cementite- ferrite- ferrite. So, once you have a cementite obviously, you will have ferrite
on both
542
sides or once you have ferrite you will have cementite on both side. We will look at what
should come first in a minute.
So, once these things grow ultimately they become separate colonies. So, these are all
what we call them as pearlite colonies. This is different from grains. So, within one
austenitic grain you would have a number of pearlite colonies possible, depending on the
number of nucleating sites that have formed. So, if you have a higher under cooling, you
would see a possibility of a number of nucleating sites. Because, nucleation rate, I is
going to be higher and in addition within the pearlite colony the interlamellar spacing is
again going to be a function of that.
We will see what is the relation between the interlamellar spacing and the under cooling
as we go along. So this is one thing which is going to ultimately define the properties,
because always the eutectoid has been so interesting for people. Because, you can really
control the mechanical properties just by heat treatment and the whole crux of heat
treatment revolves around this eutectoid transformation. So, we can modify the micro
structure the way we want simply by controlling the way, the temperature to which you
heat it or the rate at which you cool from that temperature. So, this is how we are going
to modify it.
So, now if you look at this aspect of, what should come out first? After a lot of debate
what people have thought about is the moment, I have a let us say a situation like this,
543
you have two γ grains and you are keeping at a temperature just below the eutectoid
which is around 723 ºC (See the image above). What can diffuse more easily at that
temperature?
Student: Carbon
Carbon, among the iron and carbon, it is the carbon can diffuse very easily.
Because, it is an interstitial atom, smaller atom. So, it can diffuse more easily and as a
result there is a possibility that at the grain boundary, carbon atoms can accumulate.
Why do they accumulate? Because, I am at a temperature at which γ is not stable. How
do I again look at it, I can draw the free energy composition diagram for it.
If I draw the free energy composition diagram, I have the γ, α and the cementite,
cemenite should be in principle a dot and if you want to show it you can show something
like shown above in Fig. 1. And if I am at a temperature T < 723 then only this free
energy composition diagram is valid. If I am at temperature higher than 723 then, what
should happen then austenite? So, austenite free energy curve will be below. So, you
would have a situation something like this (Shown in purple color)
So, something like a curve like this and then if I draw a common tangent between these
two, the common tangent between those two is going to be higher than the free energy
curve of γ. As a result, in principle so if the same alloy I heat it to a higher temperature
or keep it at a temperature higher than eutectoid, at that temperature in
544
principle you would see that the only way, I can see is I would have a common tangent
between α and the γ.
And I would have another common tangent between γ and Fe3C and these common
tangent compositions, are what?
Student: Composition of γ in equilibrium with α
Cαγ, Cαγ, CγFe3C and leftmost is Fe3C composition. And these compositions if I want to see
at that temperature in the phase diagram, simply I draw a line at that temperature, T
(Fig.2 in the image above) and we can mark C αγ Cαγ and CγFe3C. So that means this line
A3 is nothing, but locus of all γ compositions in equilibrium with α and this line AC m is
nothing but the locus of all γ compositions in equilibrium with cementite.
So, if I am at a temperature above the eutectoid, I can easily see it that this is how it
looks like. But the moment I come to a temperature below the eutectoid as I decrease the
temperature what happens to free energy, as I decrease the temperature free energy
increases. So, free energy increases that means, these free energy curves of of all three of
them start increasing. But because at lower temperature the γ is not stable, the rate at
which the γ free energy goes up is going to be higher. So, when the γ goes up.
545
So, you would come to a situation like this now (as shown in Fig. above), where γ curve
is shifted relatively more than α and Fe3C and if that is the case then I can draw a
common tangent between these two and if I take an alloy of 0.8 % carbon let us say(see
the Fig. above).
That 0.8 % carbon alloy now will have a higher free energy than a mixture of the α +
Fe3C which is given by point P. The free energy of the two-phase mixture is always a
point on the line joining the two free energies. Whenever, I identify any two phases, find
out the free energies of it draw a line joining the two free energies and any point on that,
would give you a free energy of a mixture having a overall composition of 0.8. The
overall composition is 0.8, but the individual phase compositions are different.
So, the only way is we can say that

fαCα + fFe3CCFe3C = Cγ
This is how it is, so the composition of α and cementite are definitely different from that
of γ, but their fractions will be adjusted in such a way that this mass balance is
maintained and that is what leads you to lever rule. But in lever rule we also use one
more thing, what else we use, just with that equation you can you get a lever rule?
What else you need?

Student: The fractions have to equal to 1
The fractions have to be equal to 1. Because at after the transformation is over you know
that fα + fFe3C = 1, that is the second boundary condition. If you do not have that boundary
condition you cannot get the lever rule. So, lever rule these two equations are the crucial
equations which ever phase that we are talking, here we talk about α and cementite. You
can call it α β, you can call it α γ anything that you call it. So, the same equation and for
liquid to solid also, we use the same thing. So, this is what you see that the moment I
come to a temperature which is below eutectoid now you have a situation here.
546
So, there is a driving force. Because, there is a driving force γ does not want to remain as
γ. It wants to precipitate α + Fe3C, the question now we are talking about is what
happens? How does this initiate? The γ wants to transform, how does this transformation
get initiated? That is where we say the moment you have a γ, γ- γ interface at the grain
boundary the atoms at that temperature have a chance to move, because the temperature
is high enough for the atoms to move.
547
And the moment in a small region, you see the carbon atoms coming to a level of 6.67 %
weight % and in terms of atomic % is how much?
Student: 25
Why are you so doubtful? It is 25, Fe 3C very easy, three atoms of iron, one atom of
carbon. So, that means in a small region if 25 % of that particular region has carbon
atoms immediately that region get converted to Fe3C.
And it does not have any problem of nucleation. Because, it is nucleating at the grain
boundary, it is not nucleating within the grain. So, there is no major problem of surface
energy problem or strain energy problem for it, because the nucleation is at the grain
boundary. So, a cementite will nucleate and the moment cementite nucleates obviously,
the regions adjacent to that gets depleted in carbon, because carbon has come closer to a
particular place from both the regions.
And as a result, the region next to it becomes ferrite and for becoming ferrite it also
rejects out carbon in both the directions. So, if I draw it in a bigger fashion, the carbon is
rejected out of this region in both the directions. Once it is rejected out then again you
will see a cementite coming out next to itand this happens and it grows laterally in a,
with alternate layers of cementite and α and this whole thing starts growing
perpendicular to the grain boundary, and for that growth again you have to remember,
what we talked about eutectic. In case of eutectic, we said how does liquid supply atoms
to both the α and β, in such a way that α and β grow together, and we talked about it we
understood that if I look at the composition of α, composition of α in equilibrium with γ,
and composition of cementite in equilibrium with γ is going to be different. We will talk
about it later; I think we will stop now.
548
Lecture No. # 23
Eutectoid reaction
Last class, we started talking about eutectoid transformation. (No audio from 00:14 to
00:24)
This is different from precipitation in the sense; here you get two phases out of a parent
phase when compared to precipitation where you get one phase out of the parent phase.
So, here we take a typical example of iron carbon system, where you have the
γ  α + cementite, where γ is the austenite, α is the ferrite and cementite is the Fe 3C
phase. And this is what we started looking at, and as you clearly see, when you compare
with the previous case of precipitation, where you have an αssss  α + θ.
So, here you have the parent phase remaining with a different composition of course, and
a product phase comes out, which precipitate out of the excess solute that is available in
549
the parent phase. So, the parent phase is the super saturated and because it is super
saturated, the excess solute comes out as precipitate. So, here the parent phase does not
vanish after the transformation. Excepting that, the volume fraction of the parent phase
which is initially 100 % would decrease to certain extent where the product phase takes
over some part of the parent phase. So, that is why so you would see that the parent
phase remains; whereas, in a eutectoid transformation you would see in principle, there
is no parent phase left out.
Excepting in cases, where you are talking about a hypo eutectoid or a hyper eutectoid,
where there is a proeutectoid phase. In this particular case, it could be a proeutectoid α or
a proeutectoid cementite depending on whether we are talking of hypo eutectoid or a
hyper eutectoid. So, as a result you may have the pro eutectoid phase; but not really the
parent phase. If you take the parent phase of the γ, γ does not remain after the
transformation. Excepting again if the transformation is not complete, that is where you
need to understand that, if you go to the TTT diagrams.
For example, look at a diagram like this, where this temperature is what we call it as the
A1 temperature. What is A1 temperature?
Student: 723 ºC
723 ºC. What happens at A1 temperature? Yeah, tell me.
Student: Transformation
Transformation of what to what?
Student: Pearlite to austenite
550
Pearlite to austenite; this is while heating or while cooling?
Student: While heating
While heating. So, we are now talking about the cooling. What happens while cooling?
Student: Austenite transforms to pearlite
Austenite transforms to pearlite; this (α + Fe3C) is what we call it as the
551
pearlite. So, that is what you call it as A1 temperature and in principle, if you have a
hypo eutectoid or a hyper eutectoid. In principle, you have another temperature, which
we call it as either A3 temperature in case of hypo eutectoid and what ACm temperature
in case of hyper eutectoid. So and that is obviously above the A1.
So, you have another temperature, which you can call it as A3 for example, in a hypo
eutectoid. And then, if you can observe, you would see another additional curve that
comes in and that is related to the formation of what does it form? Proeutectoid α from γ
so,
γ  α in this domain (between A1 and A3). So, everything is γ above that temperature
A3 and between A3 and A1, you will have γ + α provided the time that you are
providing for the transformation is longer than what is decided by the TTT diagram. For
example, if it is if I am holding the alloy at a particular temperature, I have to hold the
alloy for time longer than what is decided at that temperature; so that, I can nucleate α.
If I am holding for a shorter time, then α does not get nucleated. And that is why you say
that anything to the left of this TTT diagram is basically what? What do you have to the
left of the TTT diagram?
Student: Unstable γ
It is metastable γ. Do not call it is as unstable γ. There is lot of difference between
metastability and instability whatever it is. So, this is metastable. So, this whole domain
(hatched region) is metastable γ. The reason I have to tell you the difference is, unstable
it immediately transforms to something else. The moment I say metastable, for it to
transform to something else, you need to provide an activation barrier. That is the
difference between a metastable and unstable. For example, if you take a situation like
this (Fig.1, in the image below), I told you also an example possibly sometime back.
552
If you take a duster in vertical position, it is in a metastable state. If I keep it horizontal,

it is in a stable state. But from this metastable to come to the stable state, I need to push
it. If I do not push it, it remains like that forever. So, as long as an activation is provided,
it does not come back to the stable state and that activation is this (Q, Fig. 1). So it has a
higher energy definitely, when compare to the stable state. And infact one can also show
in case of the duster also, if it is vertical (Fig. 2a). If you draw the diagonals and
represent the center as the CG and that is at a higher state, when compare to a state when
it is horizontal (Fig. 2c); where CG in Fig. 2a is higher than Fig. 2b. If you represent CG
as the energy, energy is higher in vertical than in horizontal position.
But when it is pushed, it goes through an intermediate stage which is something like this
or possibly if I put it on the same base (Fig. 2b) and then finally goes to this. So, here if
we join these two, you come to a slightly higher level. So this, the CG which is obtained
by joining the two diagonals is at a higher position than in Fig. 2a.. So, it is at a higher
energy state here, when compared to position in Fig. 2a and this is what, we call it as
unstable; because at that stage, it does not need a push for it to fall. The question is
which direction it falls? Whether it falls this direction or that direction? It has a driving
force to fall in both the directions. Only thing is in the direction of the stable state, it has
a higher driving force.
Because Q’ is the driving force for stable state and whereas, Q is the driving force for
metastable state. Now, that depending on the direction of your push that is given, it will
fall
553
in a particular direction. So, there is definitely a driving force for an unstable state, either
to become a metastable state or to become a stable state and that driving force to a large
extent. And the direction of the force that is available decides whether the unstable state
becomes a metastable state or a stable state. And if you are starting from the metastable
state and you have provided an activation energy, then it would go to a stable state. This
is the example of Fe3C, we talk about it. When you take Fe 3C  α + graphite, this is
what we talk about.
So, you need to provide sufficient activation barrier for this Fe 3C to split into the α +
graphite or γ + graphite. What does it depend on, whether you get α + graphite or γ +
graphite?
Student: Temperature
Temperature, whether you are above A1 temperature or below A1 temperature is what
decides whether you get α + graphite or γ + graphite. And of course, for the α + graphite
and γ + graphite, the time required is going to be different. The activation barrier is
going to be different. Why? Because α + graphite occurs at a lower temperature and γ +
graphite occurs at a higher temperature. So, the reaction rates are going to be higher for
at a higher temperature when compared to a lower temperature.
So, the time required for such a transformation either α + graphite or γ + graphite can be
worked out; looking at what is the diffusivity that is available at a particular temperature,
because it is the diffusion which controls the whole thing. So, anyway, so this is all
slightly beyond the thermodynamics. But one needs to have always a
554
knowledge of this kinetics, otherwise you are always having a half knowledge. So,
thermodynamics without a kinetics knowledge will not give you any information about
the whole phase transformations; because phase transformation is always a combination
of the two.
Many times, thermodynamics may say, yes this phase will form; but the phase may not
form. But at the same time, one important thing that the thermodynamics tells us that if it
tells that if phase cannot form; kinetics cannot make it form. So, this is important. For
example, if you are at a temperature, let us say below the melting point a solid will not
become a liquid. Thermodynamically, it is not feasible. So, however fast reactions you
do it; however you put a catalyst there. It will not be able to make a solid into a liquid,
until you raise the temperature. The raising the temperature can be by thermal means by
any other means. You can do irradiation. For example, any energy can be converted to
any other energy.
So, irradiation also can lead to such kind of transformation of a solid into an amorphous
structure, which you can call it as a liquid. It is almost equivalent to raising the
temperature, where the solid has become a liquid. So, you need to provide energy for it
to transform into the stable state. So, thermodynamics definitely gives you clues, what is
possible; what is not possible. But at the same time, just because it says it is possible; it
does not happen. That is why, you take a liquid bring it below the melting point, it does
not mean that solid has to form. The liquid can remain as liquid. If that solid needs a
large activation barrier for it to form, the liquid does not form a solid at all.
And that is the reason why, in a number of cases we get glasses. Because the liquid is not
able to transform into a crystalline solid and that is a standard example is our own
silicate glasses, where you have brought the liquid below the melting point. But silicate
crystals are not able to nucleate; so, as a result, liquid remains as liquid until you reach
the temperature of Tg and where, this liquid becomes a glass. So, there are number of
numerous examples were just because there is a driving force it does not mean that, the
transformation will take place. At the same time, you also should remember that driving
force also helps us to understand kinetics to some extent.
555
For example, we have been talking all these days that the ΔG v controls to some extent
ΔG* star. How does it control? Prerna you know the relation?
Student: Inversely proportional to ΔGv2
Square, perfect; ΔG* is inversely proportional to ΔG v2. But, again that is only looking at
the nucleation, without considering any diffusion. So, if you start considering the
diffusion aspect that is to a large extent controls the kinetics. There you would see that
this activation barrier that we are talking, you need to add some other activation barrier
to really know the nucleation kinetics or growth kinetics.
That second activation barrier that we are talking about is an activation barrier for
diffusion. If you add that, then you get a full picture. Otherwise, this is not giving the full
picture. That is the reason why, you will see that in a TTT diagram, the curve is a C
shape. The C shape basically means what? It tells you that as I decrease the temperature
what is happening? What does this time tell you at any given temperature? Time needed
for nucleation; this is what, we call it as incubation period.
556
All of you know that, when any phase transformation has a typical S shape; if I plot
fraction transform, F as a function of time, t between 100 % and 0 %. So, if I am talking
of γ to pearlite transformation, it would have a typical curve like this. And this particular
time (ts), where the transformation is starting is what we call it as nucleation and t f is
what is completion; And this is what, you will see in a typical S curve and this period (t s)
is what we call it as incubation period. What happens during incubation?
Student: Diffusion to form nuclei
Nucleate, that means what? An r* is something which evolves during that period. The
atoms come together and form the size of the r*, critical nucleus for decided by that
particular temperature. Every temperature has a certain r*. Because every temperature
has certain ΔG and that ΔG decides, what is r*? You know that,
r* = - 2γ/ΔGv, assuming a spherical nucleus, we have done this possibly. All of you must
have done at the second year B. tech level, when you are talking about a nucleation.
557
So, this is something which we know. So, for a given temperature, ΔG v is fixed. Driving
force is fixed for a given temperature, given under cooling. And if that is fixed, r* is
fixed for a given nucleation. Because once you decide, what is the product coming out of
which parent phase, then γ gets fixed; γ depends on that. So, once γ gets fixed, then r*
gets fixed and that r* once for it to come out, you need certain time. Atoms have to come
together, form that r* and this time (ts), if you look at it if you carefully see here, the time
is decreasing. Why is it decreasing?
Student: r* coming down
Yes, because r* is coming down.
558
As I decrease the temperature, my under cooling is increasing (shown by double arrows
in TTT curve above). Under cooling is always measured with respect to the equilibrium
transformation temperature. For us, the equilibrium transformation temperature for this
particular transformation is A1; γ to pearlite. So, with respect to the A1 if you measure,
you would see the under cooling is increasing and if the under cooling is increasing, you
will see the ΔGv is increasing and then r* decreases.
But this does not continue forever, that is where you would see the, the second part of
the C curve. The lower half of the C curve comes into picture because of that. Though
the r* is continuously decreasing, in principle if I use that, you will see the r*
continuously decreases. But though the r* is continuously decreasing, because the
diffusivities are decreasing as a function of temperature D. You all know D = D 0exp(-Q/
RT), this is known to us. So, as a result you can see as the T decreases, this quantity,
Q/RT increases. This quantity is negative.
So, the negative quantity inside the exponential increases and once that increases, D will
decrease. So, the D is decreasing. If the D is decreasing, the time taken for any particular
size to form; any particular size of r* to form should keep on increasing. If you consider
a particular r*, let us call it as 10 nanometer size. Because the diffusivity at some
temperature T1 and another temperature T2, where T1 > T2. Because the diffusivity at T
1 is higher than T 2, the 10 nanometers can form in a shorter time. At T2 because the
diffusivity is lower, the flux is slower; because ultimately, the flux J =- Ddc/dx.
The flux will be slower and if the flux is slower, then the time taken is larger. So, if I put
that, you will lower part of C curve coming. That is the time taken for a given r*, at any
given temperature. And that you would see, because of these two combinations; you
would get a C curve. So, at temperatures which are lower, the time taken for the r* to
form is so large, because of the diffusion. And that is the reason why, you would see
finally, a C curve. (No audio from 19:54 to 20:02) So, the combination gives you a C
curve (pink colored).
And this is only starting, start of transformation and you need to consider the end of
transformation also; that means, this part. So, for each temperature…
Student: Sir, as we decrease the temperature, wont the r* also decrease
Yes.
Student: So even the time increases, the r* keeps decresing
r* keeps
559
decreasing. But though the r* is decreasing, even for that r* to however small it is for it
to form, you need atoms to come together. For example, this is a case is this is a more
difficult case; because two different phases are nucleating. When I say r*, r* of what? I
need to actually talk about a λ*, in case of this kind of a thing. Because a critical inter
lamellar distance forms because it is a two phase mixture (No audio from 21:05 to 21:44)
Infact, if you compare for example in the liquid to solid transformation, if this is the
melting point let us say (mp). This is let us say is C curve for the liquid to solid
transformation. Even for liquid to solid also, you will have a C curve. This critical
cooling rate (R1), that we are talking about for an alloy, it would be may be slower (R2);
because an alloy would be something like this. For a pure metal, it would be something
like this. For a bulk metallic glass, it would be something like this, the C curve (R3).
Something like a silicate glass, the C curve is shifted so much to the right that you need a
very slow cooling rate for it to become a glass. So, it is that diffusion and the structure
formation.
Basically, it is the r* and the D combination will decide, what is the critical cooling rate that
you need to adopt. So that is why, though the r* is decreasing, for that r* to form you need
time. And if the temperature is very low, the atoms move so slow; that even that r* to form
is very difficult; that r* could be even few nanometers. But even then for it to form; because
atomic movement is so sluggish. But in this particular case of γ to Fe 3C here when you talk
of diffusion, diffusion of what? Here,
560
there are two elements that we are talking about, the iron and the carbon. Diffusion of
iron and diffusion of carbon, both are important. And most of the cases, in this particular
transformation actually, what matters is diffusion of carbon? Because carbon can move
much more easily than iron.
As a result, infact all the transformations if you carefully observe, pearlite formation and
the bainite formation, it is all controlled by carbon diffusion. For example, upper bainite
and lower bainite, when you talk about it; it is all carbon diffusion. For example, if you
look at the upper bainite.
Student: Sir, the shape of C curve, if I have a metastable structure such that I have more
vacancies and we have the diffusion controlled by vacancies.
Correct, this will shift obviously. It will shift to the left.
Student: And this will extend up to the lower temperature.
Up to the lower temperature, until you have, you come to another structure which can
form in this particular case, let us say martensite. So, you can keep on under cooling the
austenite, until you come to a stage where you can have a martensite transformation,
which is a shear transformation that is what you will see that. If I keep on under cooling
the austenite, let us say carbon initially can diffuse faster at higher temperature. But iron
cannot diffuse faster.
So, what happens is in this particular case when you are cooling γ at certain temperature, you
suddenly see this γ; because the temperature you have comedown is such an extent that the
atomic diffusion is becoming difficult for α + Fe 3C to nucleate is becoming very difficult;
561
because α has entirely a different structure and γ Fe 3 C has a different structure,
compositions are different. So, that means you need to have a partitioning taking place.
Carbon atoms have to go both
562
into the Fe3C and get depleted in the α. So, that means you need a long range diffusion to
occur. So, instead what happens is, there will be a short range diffusion and
transformation takes place.
What is that transformation? You will see γ transforms to α.

Student: Sir, even though we are talking of iron carbon diagram, in practical alloys, its
not only iron carbon is there, other alloying elements also are there
Yeah, there could be. So, they can also shift the TTT diagram. We will come to it as
after once we understand iron carbon. Whenever you deal with it, you start with a simple
diagram with a binary effect.
Student: But carbon concentration is also less, so it is not the carbon alone that decides
these things.
Yes, whenever you are talking of a transformation of γ  α + Fe3C, let us say you have
alloying elements something like let us say vanadium or chromium or nickel. Now the
question is where do these elements go during this transformation? Before the
transformation, they are all inside the γ including the carbon. They are all inside the γ. It
is this γ, which is now splitting into these two.
If it is splitting into these two depending on their affinity towards a particular structure,
they will dissolve into it. For example, chromium, vanadium have a tendency to form
carbides let us say. They either dissolve into Fe3C and form a carbide such as (Fe,Cr)3C
or (Fe,V)3C or it can form its own carbides.
Student: Where do they go?
Go means? Obviously at the grain boundaries of your austenite, they nucleate at the
grain boundaries of austenite, the new carbides;
Student: They will stay there till whole transformation
till the whole transformation is over. And once carbon comes out of the γ by the
formation of a special carbides like, vanadium carbide or let us say Cr 23C6 kind of
carbides, the carbon content of γ has come down. Once the carbon content of γ comes
down, again this whole thing gets affected, the TTT diagram.
How does it get affected? Tell me. What happens? Let us say I have vanadium or
chromium, or any other carbide forming element and am cooling this now. So, the
moment I am cooling it, because there is a greater affinity;
Student: Shift to right
yeah, it will shift to right. Sure. Why?
Student: Because the diffusivity comes down
563
Why?
Student: Concentration of carbon decreases
So what?
Student: dC/dx
Student: Depends on the atomic sizes.
The Question is which forms first? That is important. If you think that there is a greater
affinity for carbon to form a vanadium carbide or a chromium carbide rather than
forming a iron carbide, these carbides are the things that will come out first from the γ.
The moment they come out first, they can act as a nucleating side for the pearlite.
As a result, you would actually see that it shifts to the left and that is the reason why, we
always talk about. Whenever you take an alloy steel with has carbide formers, if you
want to have a higher hardenability, you need to go to a temperature where they are all
564
completely dissolved. If they are not completely dissolved, you will see that they act as a
nucleating sites and as a result, the TTT curve shift to the left. The time required for the
nucleation is shorter, because there is already nucleating sites that is available. So, these
are the effects that you get, when you have elements such as Cr, V available. If you have
no carbide former now no carbide former that means, something like a nickel present.
Nickel does not form any carbides. Then, what should happen? That is where you see the
diffusion plays a role because of the presence of nickel which is sitting in the substitution
inside the γ. This prevents the diffusion of both carbon atoms and iron atoms. Because
wherever nickel is sitting, there is a strain. Because nickel atom size is not same as iron
atom size, it is sitting as a substitutional atom and because of which, the movement of
both iron atoms and carbon atoms get restricted. And you will see the curve is shifted to
the right. That is why, we always say all dissolved elements in iron shift the TTT to the
right. Whenever you have undissolved elements in the form of either elemental for
example, cobalt is one such example.
Cobalt does not form any carbide. But at the same time, cobalt also does not want to go
into iron; it comes out. And that is what people have seen that it forms, because cobalt
interestingly has a HCP structure which has some similarity with cubic structures. Is not
it? At least one plane is there, which has a similar thing and more or less the packing
factor also is similar. So, there are certain similarities with HCP to cubic. So, cubic
structure, which is a α structure can nucleate on that and as a result, the pearlite
formation becomes easier. Any one of the two either α or cementite, if it can nucleate
easily on something which is already available, then the whole pearlite formation
becomes easier; because pearlite is combination of the two.
The moment one comes, the second automatically comes out.

Student: The carbide formed need not be coherent
True. So, that is the important. So, in such a case what you see is that if for example,
vanadium carbide or titanium carbide is available. So, when vanadium carbide or
titanium carbide is available, obviously around that carbide all the carbon content has
come down. So, it could even not only nucleate Fe 3C, could even nucleate ferrite;
because all the carbon inside the γ in a small region has been occupied by the vanadium.
So, it has taken all the carbon. So, the γ around the Fe 3C gets depleted in carbon. That
could lead to α precipitation just around the vanadium carbide or titanium carbide.
565
And once that comes, then because of the thermodynamics that at that temperature, α +
Fe3C is stable. So, once α comes out, Fe 3C comes out. It is not that the carbides are
nucleating a cementite, need not. But if the carbide has a structure which is similar to
that of cementite, cementite may nucleate. But more interestingly most of the non
cementite type of carbides in iron carbon diagram, they are all cubic. Vanadium carbide,
tungsten carbide, you take niobium carbide; all of them are all mostly cubic in nature.
So, they would have an effect the other way. But ultimately, one thing is true that the
formation of pearlite becomes faster. That is what, we are talking about.
And this is what is going to lead to a second part of the diagram and which is where you
will see that you see a C curve. Now let us look at this particular transformation,
γ  α + Fe3C which we started talking about. Let us say, I am cooling it fast and brought
it to a temperature which is below the nose of the C curve. At that temperature, the
diffusion of iron and carbon both are becoming slower. So, this kind of a splitting and
again we are talking about alloy which does not contain any other alloying elements. Let
us say in such a case, even if it contains alloying elements, if they are in the dissolved
form, there is no problem. So, in such an alloy when I come to…
Student: Dissolved means
Dissolve means in the solid solution,

Student: Exactly it substitutes iron
566
it could be either substitutional; it could be interstitial, depending on which element you
are talking about. If it is nitrogen present, it is present in the place where carbon is also
present; that means either octahedral voids or
567
tetrahedral voids. If it is nickel let us say, it is sitting in the iron positions…
Student: Sir I am taking still more higher size of transition elements
Yeas, as you keep on adding elements with a higher and higher, what happens is only the
solubility decreases; the maximum solubility
Student: But still the case is different
could there the strain is more. The strain is more and that is the reason why, solubility is
less. That is what, we talked about when we talked about miscibility gap formation in a
phase diagram.
Student: But still, how is it possible to assume for an atom to replace with such a bigger,
smaller atom
It is not replacing, you people have to understand that there is always vacancies inside a
structure.
Student: But still, how will it go and sit in a smaller place, such a big atom
If it cannot sit, it will precipitate with something else;
Student: But we are still talking about dissolved atoms
that is what I am saying. If the dissolution takes place as long as, it can dissolve. If it
cannot dissolve, it will precipitate. As long as there is let us say carbon available; let us
say iron and tungsten, tungsten cannot dissolve into iron. What happen is iron and
tungsten will come out, is like iron and lead. Lead does not want to dissolve into iron.
Lead precipitates out and again that is another problem, such a precipitation is going to
be again as far as pearlite formation is concerned is going to aid in the pearlite
formation; because you have a second phase sitting there.
If the element cannot dissolve, it will precipitate out. There are so many immiscible
systems; iron copper is one such immiscible system. Infact people are using copper as a
strengthener, nowadays. I do not know how many of you know about it. There are what
are called IF steels now. What are these IF steels? Interstitial free steels, where the
carbon content is 30 ppm, 30 to 40 ppm. Why do we want such a low carbon? Tell me.
Formability, stretchability you want the whole car to be made of one single sheet without
welding a number of sheets. So, you want to basically stretch it to the whole body of the
car, let us say and which you have a high stretchability, then the only way is to reduce
carbon. But when you reduce carbon, strength goes down.
So, how do I get the strength? That is where, people are adding a small amount of
copper. Copper precipitates as nano copper with a lot of twins inside. And such a nano
568
copper gives a high strengthening. The reason again is because it is not soluble. So, not
being soluble is also a blessing in disguise. We all know, I know for example, very good
example of a lead. Lead is deliberately added to steels for many years, for what?
Student: Machinability
For machinability, free cutting steels we call them. We know pretty well that, it does not
dissolve. So, the concept of immiscibility is used to our own advantage. So, infact people
also use this, for even in powder metallurgy components many times as what are called
Self lubrication.
569
You add lead in leaded bronzes why lead is added? They are all bearing alloys. So, in
bearing alloys you deliberately add lead, because lead is soft. So, it acts as a lubricant,
when the bearing is there inside the component. So, the lead which is available on the
surface acts as a lubricant. So, you know pretty well, lead does not dissolve; still you add
it, because it is it does some purpose. In case of steels, it is for machinability. In case of
bearings, we add it deliberate. Copper and lead do not dissolve into each other. So, you
deliberately add lead to copper. Understand. So, it is a question of whether it can
dissolve or not dissolve. If does not dissolve, we do not talk of dissolution at all.
Whenever I say it dissolves all these elements that we are talking, they all can dissolve
including tungsten. The only reason is tungsten can dissolve at a higher temperature.
That is a reason why, when you whenever we talk about high speed steels, the
austenisation temperature is almost 1200 ºC, 1100 ºC. Because tungsten carbide for it to
dissolve into γ, it you need very high temperature. The stronger the carbide, the higher
the temperature you need for it to dissolve. It has to first break and then and then only
dissolve both carbon and tungsten. So, the stronger the carbide, that is why chromium
carbide can dissolve at a much lower temperature.
So, if you have a chromium inside the alloy is no problem 900 ºC is good enough. But if
you had vanadium, niobium, titanium, tungsten, the more you add molybdenum all these
elements the moment you add, you need to go to a higher temperature. And now once
you have them, now I am quenching it. Let us say, I am not providing sufficient time; I
have brought it to a lower temperature here. At that temperature, now I have to see what
happens to this steel?
Student: I have gone to the solutionizing state. Again I want to form Niobium carbide.
Since I have got a strong affinity, it has to form at low temperature.
Correct. I do not need to go to high temperature. I am quenching it to a lower
temperature. But at the temperature, now if I want to see whether niobium carbide would
form or not?
Definitely, niobium carbide has a higher affinity; higher affinity means larger driving
force; larger driving force means smaller r*. So, in principle it would like to form. So,
the r* for niobium carbide would be smaller than the r* for iron carbide Fe 3C, let us say.
Now, the question is whether the diffusion rates at that particular temperature allow for
that r* to come out or not? If it not, if it does not. Then, what happens is γ  α of the
same composition as that of γ; whatever is the alloy composition. Let us say, I am
570
talking of the alloy with about 0.8 % carbon, let us say and + whatever elements that you
are talking about.
571
So, if I take the alloy we are talking about; this alloy 0.8 carbon + some alloying
elements, all of them are in dissolved state and the whole alloy is quenched to low
temperatures, into the second half of the C curve. Once I come there, there because the
diffusion of carbon also takes time. Because the temperature is low enough, iron itself is
difficult. So, carbon also takes time.
So, what happens is this γ instead of giving you α + Fe3C or α + some other carbide, it
would give you γ  α. This is what kind of a
572
transformation is this? It is a partition less transformation which is a polymorphous
transformation; not an allotropic transition. I mean in one way, it is allotropic transition;
but there is no we are not talking of a pure metal. So, once this happens now this α is
containing a lot of carbon, what is the solubility of carbon in α? 0.025 at 723; 0.006 at
room temperature. But I am having 0.8 % of carbon that means it is such a highly
metastable state.
This α which has formed where only a FCC structure changes to a BCC structure which
means it only needs a short range diffusion. So, that short range diffusion has taken
place; because short range diffusion does not need a very high activation barrier. It needs
only a short activation barrier. So, atoms can move by a shorter distance and once this
happens, so γ0.8  α0.8. Now you need to see that this α of 0.8 is not stable. It wants to
somehow reject the carbon which is inside. Now the question of where does this carbon
get rejected?
That is where, you see that this α and because of only a structural change is taking place.
If I have a γ/γ interface, the α that comes out will come out as a accicular form. Because
you are the nucleation is taking place at a lower temperature. So, as a result, the α that is
coming out will form on certain planes of γ in order to get a coherency. So that, you will
see there is there is a habit plane between the γ and the α. So, you see that precipitation
forms like a platelets rather than at forming
573
at a grain boundary as what is called grain boundaries allotriomorphs. This is what
people call it as Widmanstatten structures.
Widmanstatten structures come at low temperature; because at low temperature, in order

to have the strain compatibility or the interfacial energy compatibility, you would see
that it would form on a specific planes. So that, you can have a coherency and once this
form, now the question is this particular α has a high carbon, it wants to precipitate out.
Now, that is where depending on whether you are at high temperature or at low
temperature. Now, the carbon the diffusion has to occur not within the whole of γ; but
within this α only and that region is smaller. So, as a result now if the temperature is high
enough, the carbon can diffuse from the center of α to the edge of α.
And then, form as precipitates along this interface of γ/α interface. And this is what, we
call it as upper bainite where the carbon can diffuse up to the interface between α and the
γ in the accicular structure of α and forms at the boundaries of accicular structure of α.
And if the temperature is low, it precipitates inside α and that is what, we call it as a
lower bainite; because it does not have time for moving to this much distance. So, it is
almost like our Al2Cu coming out; θ phase coming out of α. So, it cannot go and
precipitate at the grain boundaries. It precipitates inside the grain.
Similarly, here you will see it precipitates inside and if you go to further lower
temperature where the carbon is not able to move at all. And as a result, γ does not
transform to α by a short range diffusion. It transforms to α by diffusion less process that
is where you get martensite. So, you see a clear cut difference between a pearlite
formation which involves a long range diffusion and an upper bainite and lower bainite.
Both of them needs a short range diffusion excepting that carbon diffusion becomes
difficult for longer distances. So, it either forms at the boundaries or within the grain and
if you go to below the Ms, you see suddenly you will have a diffusion less
transformation. This is similar to like a glass transition.
574
I mean, if you want to compare with the liquid to solid; so, you quench it in such a way
come to below Tg, you get a glass. The liquid transforms to a glass. Similarly, the
austenite if I quench it below the Ms, the only difference is one is a solid to solid
transformation; whereas, the other is a liquid to solid transformation. Liquid to glass is a
liquid to solid transformation; whereas, γ to Martensite is a solid to solid transformation
and where there is a structural change involved. Remember, γ has a different structure
and the martensite has a different structure and this different structure is coming out
because of what is called a shear.
We will talk about it, when we talk about martensite. Whereas, in case of liquid to glass,
there is no structural change involved. You have come to such low temperatures, where
the liquid none of the atoms inside the liquid are able to move and as a result, the liquid
gets frozen. That is why, we call it as configurationally frozen liquid; configurationally
frozen super cooled liquid. So, this is the basic difference between glass and martensite
transformation. So, there is a shear involved in Martensite transformation. So, any solid
to solid transformation has to occur either by diffusion or by shear.
The moment you say that product phase has a different structure than that of the parent
phase. Otherwise the transformation is not possible; whereas, in liquid to glass, you see
that there is no atomic movement. The atom simply gets frozen, the way liquid is;
excepting the difference is that the viscosity is so high; that you cannot call this as a
liquid any more. So, glass is mechanically a solid, structurally a liquid;
575
whereas, martensite is structurally itself different. Excepting in cases, where you do not
get let us say BCT kind of structures and infact, even if you do not get BCT structure.
For example, there are so many iron based alloys, where the martensite is BCC in nature.
We will talk about it, when we talk about martensite. Most of the cases what you see
BCT, BCT comes only because of the carbon. If carbon is not there, iron based
martensite is always BCC. It does not mean that, all martensite. There are so many
martensites, nickel, titanium you have a martensite; copper based alloys you have
martensite. Infact shape memory effect, the whole shape memory effect is based on
martensite formation. So, you have other martensites, but if you talk about iron based
martensite, it can be either BCC martensite or BCT martensite. We will talk about it
later.
576
Module #01
Lecture #24
Bainitic transformation
So we will continue with our discussion on the eutectoid transformations. we took up the
case of the steel which is aN easy and more commercial example for us, where the
γ  α + Fe3C. and in this we see that depending on the temperature at which this
transformation is taking place, where the γ  α + Fe3C, there are different types of
products that are possible; different morphologies that are possible and each morphology
has certain name given to it. The lamellar morphology is what we call it as the pearlite.
And you actually get this usually at higher temperature. At temperatures closer to the A1
temperature; closer to the transformation temperature when the under cooling is smaller.
577
When the under cooling is larger, you end up into the domain which is, if this is the A1
and this is the start (S, Fig. 1 above) ; this is the finish (F) of the pearlite transformation,
if you are somewhere, if your under cooling is somewhere here, below the nose (point
P), then you end up into another type of morphology which is what is called the acicular
morphology. Where, you call them as either upper Bainite or lower Bainite. In fact, if
you take some hypo eutectoid steels, it is possible that if your under cooling is high, then
you may not even get let us say the pro eutectoid phase.
We all know that if I take a steel with, let us say, a composition like (C1, Fig. 2 above).
This is 0.8 our eutectoid composition if I take something like point 0.6 %: in principle
this alloy as you cool from γ, it gives you α + γ first. And then this α + γ, as you cool
down, the amount of α will increase and once you reach this eutectoid temperature you
will have some α + γ and the γ will undergo this transformation. This is what is known to
us.
But if you have cooled this γ very fast to some temperatures low, then you may see the
situation that you actually may not even get α, directly the γ may undergo a transformation.
For example, if you look at this, mind you if you carefully observe this particular start of
transformation and the end of transformation, both of them as you come closer to the
transformation temperature they almost become parallel to the transformation temperature.
What does that indicate?
Student: Incubation time is higher
578
Yes, it indicates that the incubation
579
period increases significantly. In fact, the incubation period, as you reach the A1
temperature it actually becomes infinite.
That means the start of transformation itself takes an infinite time. Why does this
happen?
Student: Undercooling is less
Under cooling is less. When the under cooling is less what is less?
Student: Driving force is less
The driving force is less. And if the driving force is less then, the activation barrier is
high. So, at transformation temperature the under cooling is actually 0. That means the
driving force is 0.
For the γ to transform to the pearlite, in fact, at A1 if I draw the free energy composition
diagram, how does it look like? You have a γ; you have α and this is the free energy of
the cementite let us say (Fig. 3 above). So, at the A1 which is a eutectoid temperature, in
principle this is the type of the situation. And once that is the situation if I am
considering an alloy of eutectoid composition, if I am choosing the alloy of eutectoid
composition there is no driving force for the α + cementite to come, because all of them
are on the same tangent. So, once there is no driving force obviously, the activation
barrier is infinite.
So, in principle there is no transformation possible. And that is indicated by the TTT
diagram going parallel to the transformation temperature. This is true for start line or
proeutectoid line. This particular line (pink colored line in Fig. 1) which we we have
drawn, this indicates the formation of α from the γ as you cool the pro eutectoid phase in
this particular case. So, this is γ + α. That is the reason why in fact, if you carefully
observe. Have you ever checked what is the difference between a steel which is annealed
and a steel which is normalized? In terms of the micro structure can you tell me? A 0.6
% carbon steel I am not take, let us say take a 0.8 % carbon steel and tell me what should
be the difference?
Yes.
Student: The time we are allowing
No, I am asking you micro structure difference. What will be the difference in the micro
structure of a 0.8% carbon steel which is annealed and which is normalized? Prerna, do
you know?
580
Student: Interlamellar spacing is greater in annealed sample than in the normalized
sample
Normalized sample, anything more? The interlamellar spacing in the pearlite within the
pearlite is coarser in the annealed sample and it is finer in the normalized sample. Why?
Because the nucleation rate is higher. Why nucleation rate is
581
higher in a normalized steel? Whenever I normalize a steel basically I am cooling it
faster.
You imagine that this is the TTT diagram (Fig. 1 above). Forget about the α + γ just for
the time being and I am taking a steel above the A1 and I am cooling it. Let us say these
are the two different cooling rates. If these are the two cooling rates: one is let us say,
corresponds to the annealing, Ann; one is corresponds to the normalizing, Norm, let us
say. What is the difference between these two cases? You will see that the one which is
the annealed one, intersects the TTT diagram at this point P, the normalized one
intersects the TTT diagram at this point P’.
Once this happens then you would see the difference between these two is the extent of
under cooling. If you look, the extent of under cooling that the normalized steel has gone
through is different from the annealed steel. The under cooling is smaller for an
annealing and larger for a normalizing.
Even physically also you can understand that when you are cooling it very slow, the
transformation occurs near the transformation temperature. When you are cooling it very
fast the system gets under cooled, because of the inertia in the system. Why does this
inertia come into picture? Because a product phase has to be nucleated; it needs
diffusion. α + Fe3C has to nucleate. So that means, a critical radius has to form. That
means you need certain time. If you are not providing that time the γ will tend to keep on
remaining as γ until a certain lower temperature, where it starts transforming. That is the
reason why you will see a normalized one will have a higher under cooling than
annealed. The moment you have a higher under cooling, automatically a lot of nucleation
kinetics come into picture. We have discussed this a number of times.
So, when you have a higher under cooling obviously, the activation barrier is smaller
because the driving force is larger; the higher the under cooling the driving force is
larger. In fact, what would happen to Free energy curve? The higher the under cooling
the γ curve will be much above the α + cementite. So as a result for the same alloy
composition I can see that the driving force will keep on increasing. γ curve will get
shifted. Why? Because as I come to lower and lower temperature below the A1, γ is less
and less stable and the α + Fe3C mixture is more and more stable, because we are below
the eutectoid transformation temperature.
582
So if at all γ is remaining below the eutectoid transformation temperature it is only
metastable γ. That is it. So, it cannot in principle remain as a stable γ it is only metastable
γ. So, that is why all this γ that you see (hatched region, Fig. 1), until you reach this Ms
temperature, this whole left of this curve is all metastable γ. γ is there, but it is not a
stable γ.
So in principle when I when I simply just use two different cooling rates, I am getting
two different under coolings. And because of the two different under coolings obviously,
the activation barrier is low; the nucleation rate which is related to the activation barrier
exponentially is also going to be very high. So I = Aexp(-ΔG*/RT), So, as a result you
will see that if the activation barrier is small then the nucleation rate is higher. If the
nucleation rate is higher then you will see inter lamellar spacing will come down. So that
is the reason why, in a normalized steel you see a smaller inter lamellar spacing, in a
annealed steel you will see a larger inter lamellar spacing.
In addition you would also see one more difference. What else?
Student: Pearlite colony size
Pearlite colony size, again it is going to be smaller, the number of nuclei. Here when I
say nucleus it is a pearlite colony gets nucleated. It is not a simple one single phase
getting nucleated; one pearlite colony is getting nucleated. So, when I say nucleation rate
is higher a number of pearlite colonies are getting nucleated within a particular grain. So,
one grain is broken into a number of colonies if the under cooling is higher. If the under
cooling is smaller you will see it is going to be less number of nuclei; that means, bigger
colony size. So, colony size and the interlamellar spacing, these are the two basic
differences that you see in a normalized and annealing. This is as far as it is a 0.8 carbon
steel.
The moment I make it point six what will be the difference now? These two are any way
going to be there, whatever two differences that we just now talked about it, in addition
something else would be different?
Student: Amount of hypoeutectoid phase
Amount of hypoeutectoid phase, is going to be smaller, in a normalized case when
compared to annealed case. That means, basically if I am cooling it faster this particular
temperature difference (between A3 and A1), which is what we call it as what is what do
we call that? This? There is a name given for that range. This, what is it called? From
583
here to here, for a given steel, It is called intercritical range. Inter because these are the
two critical temperatures, what is this temperature?
Student: A3, A1
A3 and A1.
584
Whether it is c or r depends on whether you cool or heat, let us not bother. So, this is the
A3 temperature this is the A1. So, between A 3 and A 1 is what we call it as intercritical
range. And that range there is certain amount of what is called some temperature
difference between A3 and A1. If I am cooling it faster, the time taken for crossing this
range is going to be shorter. And once that is the case the probability of γ coming out,
because you are crossing that very fast, that particular inter critical range if you are
crossing it very fast the time available for the α to come out of γ is going to be smaller.
So, when you are cooling it faster the amount of pro eutectoid ferrite or if you are on the
other side pro eutectoid cementite to come out will be smaller. So in fact, if you keep on
cooling it faster and faster you would even see a situation where you may not get the α at
all. For example, if I use a cooling rate something like this, R1 you do not intersect the α
+ γ region at all. If you do not intersect the α + γ region, this alloy which you are cooling
at R1 would not give you any pro eutectoid phase; it will straight away give you some
pearlite.
But at the same time, because this cooling curve, R1 is not intersecting the pearlite
finish, then you will see only some amount of γ will transform to pearlite the remaining γ
will be left out and once you reach below the Ms that remaining
585
γ will undergo a martensitic transformation. So, such an alloy would not give you any
pro eutectoid α it would simply give you pearlite + martensite.
And instead if I cool it to a high temperature like this (T1, Fig. 1 above) and hold it there
then the same γ would undergoa bainitic transformation whether it is upper bainite or
lower bainite. And if I hold it for a shorter time and then cool it then you will get a what?
Student: Pearlite + martensite
Bainite + martensite, Not pearlite +. martensite If you cool if you hold it here at T2 and
the quench it then you get a pearlite + martensite. If you hold it T1 and then quench it
you will get bainite + martensite, because below the nose we have already discussed
yesterday that the morphology of the ferrite is different. You do not get lamellar.
So, because the of the diffusion problem the γ first gives you α of the same composition
as that of the γ. So, it is a polymorphous transformation, followed by precipitation.
Whereas at T2, it is first actually it is cementite coming out and then α coming out. So,
that is the difference between T1 and T2. So, that is why you get a lamellar α +
cementite mixture at T1, and if you have not crossed the end of transformation and if you
are. In fact, one can draw a number of other isotherms here. Other, if you call this as a
for example, call this as start of transformation, S (Fig. 1), it means 0 % transformed. If
you call this as end of transformation, F (Fig. 1) it means it is100 % transformed.
between the two we can draw a number of other C curves which indicate different
percentage of transformation.
All of you possibly know how we get this: TTT diagram. You basically take a steel of a
number of samples which are all austenized and then brought to a particular temperature.
At that temperature at different regular intervals, you take out the steel and then quench
it. And then look at the micro structure. Or indirectly you measure the hardness which is
an indicator of the micro structure. So, in principle the micro structure is much more
accurate. So, you keep on seeing the micro structure for example, if a steel is quenched
from here itself (point Q), what would you get?
Student: Martensite
You get a 100 % martensite. If a steel is held up to this time (Q’) and then quenched,
then you would have some of the γ would have transformed to pearlite and the remaining
γ would give you the martensite.
586
And like that if I keep on holding for different-different times like, you would see once I
cross this line (F) I would not get any more martensite. Why not? I am crossing the
587
ms temperature while quenching,
Student: No γ
because there is no γ left out. And we know pretty well that pearlite cannot transform to
martensite. Why pearlite cannot transform to martensite?
Student: It is already stable
pearlite is already stable. No that is; austenite is also stable. Of course, at a different
temperature. it is not that the reason.
Why pearlite cannot give you Martensite.

Student: Martensite structure forms by shear
So, martensite forms by a shear transformation of one structure to the other. That is why
the FCC structure of the austenite shears and gives you martensite. Whereas, when you
are talking of pearlite it is already a two phase mixture. with two different structures and
two different compositions. So, such a two phased mixture cannot shear and give you
one single phase structure. So, that is the reason why a pearlite cannot give martensite.
So, if I have a pearlite and I want to get a martensite what do I do? I have a pearlite and I
want to get a martensite out of it.
Student: Reaustenize
Again Reaustenize it and quench it. That is the only solution. If I have a pearlite and I
want to get a martensite out of it the only way is Reaustenize it. Heat it back to above A
1 and then hold it until all the pearlite gets converted to the austenite, and then from
there you quench it and depending on your quenching rate, you either get a partial
martensite or a full martensite. Depending on whether it is intersecting the start of
transformation or not intersecting the start of transformation. But if intersects the cooling
curve, if it intersects the end of transformation, there is no chance for you to get a
martensite.
Once the cooling curve, any cooling curve that you adopt if it intersects the end of
transformation, the moment it has intersected the end of transformation that means, the
transformation is complete: γ to pearlite. When we say transformation, we are talking
about γ to pearlite transformation. That is complete. And once that is complete then
obviously, even though I am cooling it and then coming it coming in below ms its of no
use, already transformation has taken place.
588
This is similar to saying that, sir I take a liquid, bring it below Tg why glass is not
forming? It depends on whether will before you bring it to Tg whether it has formed
crystal or not. If it has already crystallized then that crystal cannot give you a glass.
Crystal can still give you not a glass, but an amorphous structure under different
circumstances. Can you tell me when? A crystal can become amorphous. Have you ever
589
heard of such situations?
Student: Ball Milling
Yes, ball milling is one: irradiation is one. Basically you have to break the crystal
structure. You have to destabilize the crystal structure.
How do I destabilize? I introduce lot of defects into it. So that, this crystal with lot of
defects, becomes unstable. And at some stage it cannot remain as a crystal and under
goes a transformation. We call that as an amorphous material, we do not call it as a glass.
Glass is a term which is specifically, you know, reserved for a liquid to solid
transformation. Because the reason is when you take such a glass, because it has come
out of the liquid when you are heating this glass you start seeing first a glass transition,
that is what possibly you have seen when we talked about it earlier. That for example, if
I take a glass and then and do a DSC, this is what you get (see the image below). I am
heating a glass, ΔH verses temperature; enthalpy evolved, verses temperature.
If you do that the glass first gives you a liquid. At this is what kind of a transformation?
We already talked possibly.
Student: Second order
it is a second ordered transition. And in a second order transition what do you see ΔH=
0. .Because ΔH = 0 you do not see any peak there? You just simply see a step. And now
590
once this super cool liquid is crystallizing you get a peak, and the area under the peak
tells you how much is the ΔH evolved or absorbed during any particular transformation.
591
And this particular region, from here to here is what we call it as ΔT x, which is called
super cool liquid regime. So, where you have, the glass is no more a glass it has become
a liquid, but the liquid is still remaining as liquid; it has not yet crystallized. So, it is like
a liquid state. That is the reason why many people are doing a lot of deformation in this
domain. Metallic glasses and then getting super plastic deformation. Glass we all know it
is a brittle, but the same glass can be super plastically deformed if you are in this
domain, and get any shape that you want. So, many people do that. In fact, all the shapes
of normal glasses that we talk about also are more or less done like that, in a liquid state
they do it, and then freeze it.
So similarly here you can do that shape it and then further heat it if you want so that you
can crystallize. And in fact, you can control that crystallization by controlling the
temperature and time of crystallization so that you can induce nano crystals in a glass.
Lot of people have shown that, when you induce nano crystals into a glass, you can
make the glass very tough. People have shown that a typical wind shield of a motor
vehicle like a car will not crash into pieces when there is a crash whereas, it simply
bends and goes back.
There are examples like that where, basically you induce small nano crystals so that a
small shear bands form and then there is no growth of that. So, it because each particle
will act as an obstacle for the shear band to grow. So, as a result you have some kind of
an obstacles for any crack propagation and that gives you higher fracture toughness.
Fracture toughness is always related to either crack initiation or crack propagation. If
you can make the crack either initiation more difficult, how do you make the crack
initiation more difficult in a typical metallurgical scenario? We do some things. In steels
let us say, you want to make the crack initiation more difficult what do you do?
Student: Reduce the inclusions
Reduce inclusions.
Student: Harden the surface
Harden the surface. Yes, shot-peening. There are many things where people you know
try to see compressive stresses on the surface.
Hardening is one way; even inducing compressive stresses on the surface so that there is
no; fatigue life for example, everybody knows fatigue life increases when you do shot-
peening. So, it is simply that. So, you either make the crack initiation more difficult or
592
crack propagation more difficult. One of the two, if you can do that you can make a
material tougher.
Student: Can we achieve Martensite at higher temperature?
yeah. This is, when we come to martensite we will talk about it.
593
So, there is always something called deformation induced martensite. You can, the Ms
temperature can go up. You, all of us know that Ms temperature whenever you add
alloying elements it always goes down. Why does it go down? Again when we come to
martensite we will talk more about it. Basically again shear is made more difficult.
Because martensitic transformation is a shear transformation. Whenever you have
alloying elements present there because of the strains that are there, around each of the
element that is sitting, you would see this shear becomes more difficult. and when the
shear becomes more difficult Ms goes down. we will talk about it when we came to
martensitic transformation.
Any way coming back here, so there are possibilities where you may not even get a
ferrite a pro eutectoid phase and straight away get a pearlite. And depending on the
actual cooling rates that we are adopt (No video). So, to start supplying carbon atoms to
Fe3C and α, the Fe3C needs more carbon atoms. So, you need carbon atoms to diffuse
from the γ which is here, into this. At this temperature that diffusion can be very
difficult, because you need a long range diffusion, because from the whole of the γ grain
the atoms have to start moving. And that, because is difficult at lower temperature this
pearlite whatever has formed will not grow much.
It may grow, but at a very slow rate. But at the same time, the γ that is there is not stable
at that temperature (T1), because you are far away from the A1 temperature. So, this
594
γ has to somehow transform. So as a result new nucleation events will occur. So, this γ
which is now, in fact, this γ will be already depleted in carbon. You are starting from
0.8% carbon let us say, but this γ will not be point eight carbon, because already some
pearlite has come out of it. Because some pearlite has come out of it, some carbon has
come out it so, there is a possibility that some depletion might have taken place.
Otherwise if you think of it, you know other way to look at it pearlite though there is
cementite and ferrite, the overall composition of pearlite is always 0.8. Isn’t it? When I
say, you have pearlite coming out what is the composition of pearlite if somebody asks
though it is a stupid question, because it is not one single phase. If you say what is the
composition of pearlite then basically it is 0.8 only. Only thing is
fαCα + fFe3CCFe3C = 0.8
And if you call this as the pearlite where α and cementite is there so, the overall
composition of the pearlite is 0.8. So, it is because γ of 0.8 is giving you this. So, we can
say that this pearlite would have point eight, but some regions of γ will be left out;, that γ
because it is not stable at that temperature (T1) it wants to nucleate something. And that
is where you will see this γ would undergo a polymorphous transformation and gives
you α. In some other regions of of γ; and that particular thing will be basically in a
acicular nature and depending on where you have actually held it that may give you
either lower bainite or upper bainite.
Though academic exercise a number of people who have tried this out and see a mixture
of lamellar and acicular structures. This is possible.
Student: What is upper bainite and lower bainite?
We have talked about it yesterday; a upper bainite is one where; if you look at if you
think of an acicular structure, an upper bainite is one where all the cementite particles
will be at the grain boundaries. Whereas, in a lower bainite the same acicular ferrite will
be there, in that acicular ferrite you will have cementite nucleated on a certain habit
planes. They are like plate like cementites, formed inside the ferrite and this ferrite is
formed by polymorphous transformation.
595
So, at the time of transformation this ferrite will have the same composition as that of the
γ. If γ is 0.8 α will be 0.8. And that carbon, excess carbon will be rejected out and then
will be given in the form of cementite so that at the end of the transformation is over
then the α will come back to the original α composition decided by the phase diagram.
Whatever is the α composition at that particular temperature you will get that. Whereas,
the remaining carbon, the excess carbon will be converted into the cementite. But if this
process is not complete you will always have possibly some excess carbon left out in the
ferrite so that it will remain as what is called a super saturated ferrite. So, you will
always have some super saturated ferrite, because if the temperatures are very low let us
say and if you have a lot of alloying elements in the steel, the diffusion rates being very
slow, you may have always a situation where some super saturated ferrite will be left
out. The complete transformation has not taken place.
But provided, if you have a plane carbon steel and if the transformation is done at a
relatively higher temperature for a long time, then you would see complete excess carbon
will come out as cementites, and you would see the ferrite having let us say 0.025
composition and the cementite will be there according to the volume fraction that it is
suppose to be there. This is how you would see transformations taking place.
596
Second thing is, if we look at this we also talked there is a possibility that if I take an
alloy of this composition (C1), we just now said, there is a possibility that it will directly
give you pearlite instead of giving you a ferrite; that can happen provided if I extrapolate
this these two lines, A3 and this is AC m, which has to be steeper. Why that is steeper?
ACm line is steeper than A3 line. Can you tell me why?
Student: A3 is the phase transformation, ACm is solubility
That is also phase transformation. γ is giving you a cementite; there is a phase
transformation. Simple geometry. That is it.
You should know that here A3; what is this temperature (pure Fe end) ?
Student: 910 ºC
910 let us say. So, from 910 it is coming to 723 in 0.8 %. And where does this ACm go?
Student: 1130 ºC
1130 ºC. So, from 1130 to 723 it is happening in how much?
Student: 1.2
1.2 %. as simple as that. So, (910-723)/0.8 < (1130-723)/1.2, therefore slope of A3 is
lower than that of ACm.
So as a result that is the reason why we always say in the hyper eutectoid steels, if you
want to go the single phase γ region you need to go to very high temperatures. Whereas,
in case of a hypo eutectoid steel, you need not have to go to very high temperatures. The
597
highest you have to go is only 910 ºC, whereas here as you go closer and closer to 2 %
you need to go beyond 1100 ºC at 2% actually it is 1130 ºC.
598
So, if you are something like 1.4 % carbon steel something like that tool steel is there:
1.4% carbon or 1.5% carbon, such a steel you need to go to very high temperature and
that has its own problems. it has certain advantages certain problems. What is the
advantage tell me? If I heat a steel to very high temperatures, what is the advantage?
Physical metallurgy if you are thorough you will know. Yes?
Student: Coring will reduce
Coring. Why are you talking of coring, why? Coring will be decreased, you are talking
about, let us not bother. We are talking about heat treatment; in terms of heat treatment;
in terms of the micro structures.
Student: Coarser grains
Coarser grains so?
Student: Toughness will be better
Toughness will be better coarser grains?
Student: Hardenability
Hardenability is higher. You will get a higher hardenability, if I take it to a higher
temperature. Because the grains are coarser; the number of nucleating sites for pearlite to
form are lower, obviously, the pearlite formation it is like: when we talked about glass
formation we said every liquid would form a glass until and unless crystallization is in
intervened- this is what we said. Turn bull has made this statement long back. Liquid
wants to become a glass because it has no problem in becoming a glass; because its
structure is the same. But unfortunately some crystallization is coming in the middle.
If I can suppress that crystallization, I will get the glass. Similarly, here if I can suppress
the pearlite formation I can get the martensite. So, in order to get a martensite, you need
to suppress pearlite formation. So, whatever mechanism methodologies that you can
adopt to somehow suppress the pearlite formation is good for you. So, one such
methodology is increasing the grain size. But there are problems. What are the
problems?
Student: Grain boundary oxidation
Grain boundary oxidation, fine. If I am doing my annealing or austenizing in a reduced
atmosphere you would not have problem of oxidation.
But still there is a problem.
599
Student: Strength decreases
Strength decreases coarse grains strength decreases, fine. Toughness decreases. One
more, you get more quench cracks, when you are quenching to get martensite. Because
you are quenching from a higher temperature; thermal shocks are more. So, that is the
reason why. So That too, tool steels. So, there is a possibility of quench cracks. That is
the reason why tool steels you do not quench them. Luckily, because the carbon is
higher, the hardenability is higher already.
600
Why hardenability is higher if the carbon is higher? Again, diffusion problem.
Whenever you have a high carbon steel there is a more stress inside. if you take an
austenite with a 1.5 % carbon, austenite with a 0.4% carbon if you are comparing these
two, you would see that because of a large amount of interstitials are occupied by these
carbon atoms, the diffusion is more difficult. Each carbon atom would be an obstacle for
another carbon atom to move. It is like having too many dislocations; you have a work
hardening taking place.
So similarly here, you have diffusion gets basically, any alloy diffusion or chemical
diffusion nor self diffusion, chemical diffusion is going to be more difficult if you have
the solute content is higher. That is why chemical diffusivity is always a function of, if
you read about diffusion, is always a function of composition. The dilute the alloy is, the
higher the diffusivity is. the more concentrated the alloy is the lower the diffusivity is.
So, that is one of the reasons why diffusion come rate comes down. And once the
diffusion rate comes down the hardenability is higher.
Of course, there is always something which is counteracting this. What is that, that is
counteracting? Because when you take a high carbon steel there is, a pro eutectoid
cementite coming out. And the pro eutectoid cementite, the moment it comes out it can
nucleate pearlite.
601
So, there are always issues like that. That is why, you know interestingly, if you ever
have seen the TTT diagram, in principle. If this is the TTT diagram for a 0.8 % carbon
steel I am only drawing the start of transformation (see image above).
If I take a let us say 0.4 % carbon and 1.2% carbon. So, I decrease by point four; I
increase by point four. Look at the TTT diagram. How will it look like? How will the
start of transformation, that curve will look like?
Student: Both will have extra lines
You forget about that. I am not going to talk about pro eutectoid. I am simply going to
ask you the eutectoid C curve. The C curve, related to the eutectoid, will something
happen to it? Will it shift for that matter? If it shifts in which direction it will shift and
how will it shift? I want you to think about it and may be draw it and bring it in the next
class. Do not consider any alloying elements, nothing. I am not even concerned about the
pro eutectoid. Simply 0.4 and 1.2 - what would happen to the TTT diagram?
Student: In both cases, a shift will be there.
Both sides a case, I mean.

Student: Till 0.8 it shifts to the right
Till 0.8 it shifts to the right, yes.
Student: And they don’t observe that shift after 0.8 .
They do not observe the shift after 0.8.
Student: Shift to the lift.
It is very interesting, from, from almost 0% carbon if you start looking at, and start
increasing the carbon content, I can give you the answer here itself. So, if you look at
from zero % carbon, if you start looking at I mean 0 we do not consider, let us say 0.1 %
carbon or something like that, you would see slowly the TTT curve shifts to the right:
until 0.8, because you are increasing the carbon content; hardenability is increasing.
And also in another words also, amount of ferrite is also decreasing. So, the number of
nucleating sites are also decreasing, isn’t it? Amount of pearlite as you go towards the
point 0.8% carbon, the amount of pro eutectoid ferrite is decreasing. Once the pro
eutectoid ferrite decreases, the number of nucleating sites for the pearlite also decreases.
So, that is the reason why you will see may be, a point one % carbon would be like this
(see the image above); and point eight would be like this. Now if I am increasing it
further what would happen?
Student: Again decrease
602
It would actually start coming back. But there is, this is where the question is, there you
have two counteracting effects.
So, in this particular case (0.1% Carbon) you do not have that. In the hypo eutectoid you
do not have that problem. If I decrease the carbon content, I am decreasing the carbon
content I am
603
increasing the ferrite amount. Both are detrimental for hardenability. Whereas, if I am
increasing the carbon content beyond 0.8, I am increasing the carbon content which
should actually increase the hardenability, but I am getting cementite, which should be
detrimental for the hardenability.
So, you have two counteracting effects. And as a result, actually, you do not see
significant shifts. What Sri Ram said is right. You do not see significant shift, because,
whatever you loose because of one you gain because of the second one. So, you do not
significantly see shifts. Whereas, you see significant shifts on the hypo side. So, that is
something which you need to understand. So, as if you do not understand that you would
not be able to understand the hardenability effects in different compositions and this is
very important.
So, what I am trying to say is, if you take this alloy of C1 composition (Refer slide time
33:39) : if the moment this alloy intersects extrapolated ACm; if your under cooling is
sufficient such that this alloy is brought below that intersection; then, you would
suddenly get a pearlite. Because, now, you are below that cementite ACm line. So,
cementite can nucleate out of that. The moment cementite nucleates immediately, you
will get a ferrite next to it and that is what is pearlite. So, you can get things like that and
if the temperature is too low there, you would get a bainite instead of a pearlite. But you
would get one of the two phase mixture: either a pearlite or a bainite, you would get if
you are below ACm, provided nothing has happen before it. We will stop here.
604
Lecture No. # 25
Kinetics of Eutectoid transformation
We started talking about eutectoid, last two three classes. We will try to continue today
and then go into martensitic transformation, because that is a kind of an offshoot of
eutectoid transformation that you get. So, this one aspect about eutectoid transformation,
which we have been seeing yesterday, is about, what is the kinetics of it? Yesterday, we
considered it partially, last class.
Let us try to continue this. As you all know that eutectoid transformation as long as you
take a eutectoid sort composition. We have seen that the driving force that you get for
such a transformation can be seen from a free energy composition diagram (PQ, Fig. 1
above)
Let us say, you take a case like that, that you have a certain driving force at a given
temperature below the eutectoid transformation. The under cooling, which translates to
this driving force (PQ) is actually drives the whole reaction. This we have seen it even in
the eutectic; so I thought let us just recapitulate once again. Because we if you look at
605
eutectoid transformation as some phase; let us say call it as α, β and coming out of a γ
(Fig.. 2 above). We can take it as an iron carbon system or any other system. If you take
it, we see that this growth of this lamellae depends on the atoms joining from γ to both β
and α.
So, when atoms from γ come and join the β and we know that, β is let us say B rich
phase, α is A rich phase. So, you need excess B atoms to come and join β and excess A
atoms to come and join α and then, you will see the whole thing will grow far. This how
a γ can supply excess atoms of B to β and excess atoms of A to α is what you can see,
again if you draw a common tangent between γ and α and γ and β. So, let us say I am
drawing a common tangent between γ and α (K1L1 in Fig. 1).
That means, the moment I have γ in equilibrium with α. What is the composition of γ,
which is just in front of α? The composition of γ which is just in front of α, which is in
equilibrium with α, is given by this common tangent (K1L1) composition (at K1). So,
where γ is in equilibrium with α; let us not bother about. What is the composition of α in
equilibrium with γ? But if you look at, what is the composition of γ in equilibrium with
α, this (at K1) is what you will get. And similarly, if you look at, what is the composition
of γ in front of β?
Again we draw another common tangent (K2L2) between the β and the γ and you get
another composition here (at K2). We call the compositions at K1 and K2 as C γα and Cγβ,
respectively. We can see now, that the Cγα has more B atoms than Cγβ; So, you have more
B atoms in front of α (point K1, Fig. 2 above) than in front of β (point K2, Fig. 2). As a
result, you will see that the B atoms will start moving in this direction (K1 to K2) and
that gradient, if you want to consider it that gradient can be given as
α β
dC C γ −C γ
=
dx λ /2
λ is what we call it as interlamellar distance, the center of one β to center of another β is
what the interlamellar distance and this diffusion distance, that we are talking about is
half of that interlamellaer distance. Again, certain notations one has to remember, X
usually means mole fraction. So, put x in dC/dx
606
let us say. In fact, some people preferred to write X rather than C for composition. Either
you can use C or X depending on whatever you want, as long as you are consistent in
one of the notation.
So, if this is the case, then we know the flux is

J = -DdC/dx.
So, again coming back to what we have done in eutectic is exactly similar. So, it is the
composition gradient which drives the diffusion and at a given temperature T, if this
eutectoid transformation is happening, if I say that I have taken an alloy γ brought it to T
and holding it there at that temperature, till the transformation is over. If I am saying that
at that temperature Cγα and Cγβ are fixed; that means, dC is fixed and that if it is fixed and
if the λ is fixed, λ basically depends on the under cooling.
It is under cooling, which decides the λ. And once the λ is fixed, the concentration
gradient is more or less fixed and once the concentration gradient is fixed and D is also
fixed for that particular temperature. You would see this grows at a particular rate. So,
the growth rate is dependent on the temperature; that means, the under cooling in two
ways. One, the under cooling decides the D; the under cooling also decides the dC/ dx.
Because the under cooling is what decides the Cγα - Cγβ and how do I see that? I can see it
easily. If I extrapolate the transformation lines of γ-β and γ-α (as shown in pink colour in
image above), the points where it intersects temperature T, gives the values of Cγβ and
Cγα at that temperature.
607
So, if you use the same logic as we have used in eutectic, so this difference C γα - Cγβ
decides the ΔC and because that decides ΔC, in principle we can say the dC/dx is
decided by that. And once you identify that, you can say that the growth rate of this
eutectoid depends on that. But keeping in mind that, there are two parameters here,
which kind of counteract each other. One is that the dC/dx in principle increases with
under cooling.
In dC/dx also, there are two. One is dC, another is dx; dx = λ/2. What happens to λ/2 as I
increase under cooling? It decreases and what happens to ΔC? Increases. So you see that,
the numerator is increasing; the denominator is decreasing. So that means, this fraction
dC/dx increases with under cooling. But at the same time, D, again the diffusivity; as I
increase the under cooling, diffusivity should go down. So, this growth rate is again a
combination of the temperature, which is where diffusivity decides that and the dC/dx.
So, a combination of these two will tell you that, the growth rate will reach a maximum
somewhere in the intermediate temperature range. It will not be highest near at the
eutectoid transformation temperature in principle, growth rate will be 0; because dC = 0.
Because at the eutectoid temperature this difference Cγα - Cγβ = 0, and it is that, which is
driving the gradient. And if there is no gradient, there is no diffusion.
608
So, if I now plot the growth rate v as a function of under cooling, ΔT. You would see the
under cooling would be something like this (as shown above). What I mean by 0 under
cooling? At the eutectoid transformation temperature, ΔT = 0. And v would again go
down, when you go to higher under coolings; because the temperature is going down,
where the D is going down. Though ΔT is increasing, so that is why again it would be a
combination of two curves; one, which is nothing but the diffusivity, D. So, diffusivity is
decreasing with increasing ΔT and the dc/dx is increasing with increasing ΔT .
So as a result, a combination of these two, you would see the growth rate, v will be like
this (as shown above). Now, in addition we also talked yesterday about, in the last class
about the morphology of α.
609
(Refer Slide Time 12:46)
You see, if you take an alloy of a particular composition (C1), we also said yesterday that
if I take this alloy under cool it below T01temperature. You
610
directly get a pearlite or a bainite depending on where you are with respect to that
temperature. So, if I am closer to the eutectoid temperature.
So that means in a TTT diagram, you are closer here (T2) to the eutectoid temperature.
So, because you are closer to the eutectoid temperature, this particular alloy (C 2), if I
under cool it below this particular temperature T02 (See Refer slide time 12:46) then it
would give me a pearlite rather than give me bainite. But if I take this alloy (C 1) which is
farther from the eutectoid composition; this alloy, if I can under cool of course under
cooling that alloy to that temperature (T01) is also not easy. Because the extent of under
cooling, that I need to provide is very large.
If I can somehow provide that under cooling, and hold it at that temperature below that
particular temperature (T01). You would end up in a bainite and that is what you see. When I
under cool it, hold it there (T1, in image above); I would end up in a bainite there. Though
there is proeutectoid phase, no proeutectoid phase will come out during this under cooling.
Because you are cooling it fast, you are not allowing time for the proeutectoid phase. So,
instead of any proeutectoid phase coming, straight away you come to a region.
And that is why we said, if your cooling rate is higher than what is decided by, where the
proeutectoid component reaches the TTT curve of the eutectoid. And if your cooling rate
is higher than that, obviously you will never get a proeutectoid phase. If your cooling
rate (R1) is smaller, then you will cut both the curves; then you will get a proeutectoid
phase. And depending on, where it is cutting and what is the under cooling that you are
611
achieving, the proeutectoid phase also the fineness of the proeutectoid phase is going to
be decided.
For example, if I use a cooling rate like this (R2) with respect to the proeutectoid
transformation temperature, what is called the A3, Now with respect to that, if I consider
the under cooling, the under cooling for R2 is smaller; the under cooling for R1 is larger.
So, because the under cooling is larger for R1, if I am using R1 instead of R2, here I will
get a finer α, We are now talking about the proeutectoid phase. So, you would get a finer
α at R1, instead of what you get using R2. So, that is why whatever we have talked in the
last class about, what is the difference between annealing and normalizing.
If we extent this to, a hypo eutectoid alloy; in a hypo eutectoid alloy you will see another
difference that, you will have a coarser α in a hypo eutectoid steel, if you do annealing;
you will have a finer α, if you do a normalizing. That is the second thing that we talked
about it. The proeutectoid phase amount will keep on decreasing as you increase the
cooling rate; that is what we have seen there. You will see that the amount of
proeutectoid phase will keep on decreasing, until you reach a cooling rate faster than the
nose of TTT curve.
So, if you have a cooling rate something like this (see the bottom right figure in above
image), then you would only get eutectoid phase.
Student: Can we control the amount of the phase that is formed?
In principle, one can provided you have the CCT diagram; one can talk about what
612
is the amount of eutectoid phase? And the higher the cooling rate, that is the reason why
by looking at the microstructure, if you want to find out, what is the composition of
alloy? Infact in B.tech labs possibly you would have seen people asking you, see the
microstructure of the steel and tell me what is the composition? In principle, if it is a
plain carbon steel assuming that, it is a plain carbon steel
In principle, we can do this provided we have a annealed microstructure. If you have a

normalized microstructure, it will not represent the composition of the alloy. Basically,
because whatever you get from the lever rule; because basically you are going to use the
lever rule and from the lever rule, find out the composition of the alloy. So, if you have
seen the microstructure, certain area fraction you assume it to be equivalent to a volume
fraction. Convert it into a weight fraction and finally, find out what will be the alloy
composition corresponding to that weight fraction using the lever rule.
This is how, you have to actually calculate. If you want to really calculate, based on the
microstructure what is the alloy composition? Because the microstructure only gives you
area fraction. It will not give you weight fraction and what the lever rule gives you only
weight fraction. It does not give you any area fraction or volume fraction. So, you need
to correlate these things. So that finally, you get the weight fraction and that, if you want
to do it. The lever rule can be applicable only under equilibrium conditions, because it is
coming from the equilibrium phase diagram.
So, that means you need to use a very slow cooling conditions to be able to use a lever
rule. So, if I have used a normalizing and got a normalized steel microstructure, it will
not correspond to… Let us say 0.6 % carbon steel if I take, in principle lever rule tells
me how much of ferrite? 0.6 % carbon steel; how much is the ferrite?
Student: 25 %
25 %; it is one fourth; three fourth. Because 0.6/0.8 that is what it is. So, if it is about 75
% of eutectoid mixture and 25% of ferrite. And if I now take the same steel and do
normalizing 0.6 % carbon, you will not see 25 % of ferrite.
You may see 15 %; you may see 20 % depending on the cooling rate that you have
adopted. And if I see 20 % and go back and calculate from the lever rule and then, say 20
%. Because it is 20 %, its composition is let us say 0.5 or so. You are doing an error. The
actual composition if you do chemical analysis, it will not correspond to that. The reason
is I have taken a normalized steel. In a normalized steel, the
613
volume fraction of the ferrite phase will not correspond to what is decided by the phase
diagram.
So, these are the minor issues, which you have to be very clear, when you want to
actually utilize the microstructures to understand the composition; because many people
may not directly have chemical analysis facilities to do a composition. So, we tend to do
simply see the microstructure and get an idea of the composition and also one more thing
that you need to know is the microstructure also is decided by… If the alloy is the not a
binary but alloying elements present. The moment you have alloying element present,
the whole phase diagram changes.
The phase diagram is driven by the presence of ternary and quaternary alloying
elements. For example, you must have seen, when you are talking about alloy steels,
how chromium changes the phase diagram, how nickel changes the phase diagram.
Particularly, you must have seen the γ zone shrinks, when you add chromium.
Student: Why?
Basically, because it stabilizes ferrite; it does not stabilize austenite. So, it wants to
dissolve into ferrite; it does not want to dissolve into austenite. So, austenite shrinks
infact around 18 % or so. The austenite region is very small. What is the austenite region
in our phase diagram?
Austenite region, if I want to draw; this is the austenite domain (see top left figure
above) and what do you call this temperature (A 4)? What is it called? What happens
there?
614
Student: δ
γ to δ. What is γ to δ transformation temperature denoted as?
615
Student: A4.
A4, we also have A3, A1, and obviously, there must be A2 somewhere. So, that A2 is the
magnetic transition. So, that A4 and that is what you can actually see, how carbon
stabilizes the austenite. The moment you add carbon, you see A 3 is decreasing; A4 is
increasing.
As I add more and more carbon, you see the A3 is decreasing and the A4 is increasing.
That means, a pure iron γ is stable only between 910 to 1400 C; whereas, in around 0.8
% carbon are so you could see it extend to much larger domain. So, once I know that,
basically tell you that this field is increasing and that means, whenever a phase field
expands. That means that particular element stabilizes that structure.
And this is basically, because the carbon wants to be in an FCC structure rather than in a
BCC structure. So, that is why the α it tries to depress; γ, I mean δ also it tries to depress.
So, the stability range of the δ gets reduced and the stability range of α gets reduced. So,
that you get γ. So, that clearly tells you that carbon is a γ stabilizer and the same thing
now, if you add the chromium you will see that γ field phase shrinks to a small domain,
around 18 % or so or 20 %, if you add.
So, that is how a nickel if you add it, γ phase field goes down further which is an
austenite stabilizer. So, any austenite stabilizer will keep on shifting it down. So, you can
know from that, how each one stabilizes. So, these things also are decided by the
composition, we need to know about it. The second issue is, if I take this alloy (C 1, Refer
slide time 24:35, for the discussions below) and then, I am cooling it. The α that is
forming here in the pro-eutectoid region, what would be the morphology of that α?
For example, if I have under cooled this alloy C 1, to this temperature T1, still you will
have an α forming; because it has not yet intersected this T0 1. So, you will have an α
forming, even if you have cooled it below the eutectoid temperature. If you have under
cool it below eutectoid temperature not simply cooled it. So, if you have simply cooled it
by that time, eutectoid transformation is already over. So, if you under cool it to that
temperature T1, you can still get α. What would be the morphology of that α, when
compared to that α which is coming somewhere here in the proeutectoid region?
616
And this is what, a large number of people have studied and they see that the moment the
transformation temperature is closer to the A3, you will see γ is there. A grain boundary
between γ, you see this kind of precipitation (see Fig. 2 above) which is what is called
grain boundary allotriomorphs which are more or less spherical kind of precipitates of α.
The moment, the under cooling is very high and let us say somewhere here (T1), you see
the α coming out. In such a case, you see the α comes out as platelets (Fig. 3) with semi
coherent interfaces. Here (Fig. 2), these are all incoherent interfaces. The reason is if you
look at the growth rates of a semi coherent interface and coherent interface.
617
If you think of velocity of a semi coherent interface, v semicoh and velocity of an incoherent
interface, vincoh. At high temperature, the velocity of both of them is more or less similar;
because the diffusion rates are very high at high temperatures. But at lower temperature,
you would see that, the incoherent interface moves faster. Because if a semi coherent or
a coherent interface has to move, it has to maintain the coherency with the matrix before
it grows. For example, whenever you look at a plate, a plate AB and CD are coherent or
semi coherent interfaces; whereas, BC and AD are incoherent interfaces.
You will see that BC grows faster when compared to AB. The incoherent interface can
grow faster. So, that is why plates grow laterally, they never grow. So, the thickness of
the plate does not change during precipitation reaction. For example, if I take a
aluminum copper, you get θ” disk. The disk thickness rarely changes much. But the disk
grows laterally; because of this, BC growth is much faster than AB growth; because if
you want AB growth to occur, then you need to maintain the coherency continuously.
The continuity of the atomic planes has to be maintained; whereas, here there is no need
of such thing.
So, as a result you see that, at higher temperatures if you take the ratio of… If I plot the
ratio of
Student: Plate like precipitate, why it has incoherent interface on one side?
618
Yeah because whenever you are talking of coherency, coherency is on any particular
plane. When nucleation is taking place, a plate is nucleating on a matrix. Let us say, θ”
or something like that nucleating on a aluminum matrix; so, the
619
111 plane of the aluminum matrix acts as a nucleating site. So, it is only that plane which
is oriented with respect to some plane of the precipitate, whatever plane it is. And once
you see that, so this (CD) is the precipitation and then, then if you look at both the sides,
it has to maintain coherency with another plane of this.
There may not be a plane, which is perpendicular to CD; which is also coherent with the
precipitate, you may not see that. So, you will see that as long as, there is an orientation
relationship; the interfacial energy is reduced. So, over all the system wants to reduce the
interfacial energy. So, it reduces; so the nucleation has taken place. Now, how about the
growth? So, there is no such coherency on other sides and because there is no coherency
on this sides. Any of the plate like precipitate will always have incoherent component +
either a coherent or a semi coherent.
Initially, it could be coherent; slowly it may lead up to semi coherent and finally, you
would see that, this whole thing grows as a plate. AB retains that, instead of getting
converted into an incoherent interface here itself. If it has grown this way, then it will
finally get converted to incoherent; instead the system would prefer a growth in
incoherent direction. That is why in many cases, you see plates coming. Even for
example, aluminum silicon alloy, the silicon grows like needles. Even in the liquid to
solid transformation also, you see this happening; not only solid to solid. That is because
whenever you look at any precipitate.
For example, a nucleus has formed, you imagine it is like a cube. Because silicon let us
say is cubic in nature; diamond cubic. So, some nucleus has formed. This nucleus has to
now grow. The growth depends on the rate of atoms joining from the liquid to that
silicon and not only it depends on that; it also depends on which plane of these needs less
number of atoms for it to grow. Each plane here, if I consider we have talked about Wulf
plot that there is shape for the nucleus. Once I consider there is a shape for the nucleus,
there are so many planes which are covering that nucleus.
Out of all these planes, whichever plane needs less number of atoms per unit area for it
to grow; that will be the one which grows faster. That means those are the planes which
are loose packed planes will grow faster than the close packed planes. Close packed
planes need more number of atoms. So, at a given temperature, the liquid is supplying
620
atoms. When the liquid is supplying atoms, the rate at which it can supply atoms depends on
the diffusivity. And once this diffusivity is fixed at a given temperature, then it cannot
preferentially give more atoms to a particular plane and less atoms to another plane.
So, all the planes are receiving same number of atoms at in a given time. But one
particular plane, which needs less number of atoms for it grow by a small Δx will grow.
Because if you have planes like this, let us say if this plane (LM in image above) needs
less number atoms for it to grow. Once all the necessary number of atoms come and join
that, within a particular unit area. It grows by a small Δx, which is may be one mono
layer. Similarly, within that particular time, the number of atoms that are required for
this particular plane MN to completely get filled. If they have not reached, then MN will
not grow. Is it not?
So, like that you will see the growth morphology; sometimes will be entirely different
from the nucleation morphology. So, that means this LM and similar oriented planes
keeps on growing; these atoms keep on growing whereas, MN does not grow. Finally,
you will see that the growth morphology may become like a square. The nucleus
morphology is octahedron; the growth morphology is square. So, you can see that, at the
nucleation stage, what you see is different from after certain growth. This is because
certain planes grow faster. Same thing in case of plates, you will see some of the
interfaces grow faster and some of the interfaces grow slower.
621
So, if I plot vincoh/vsemicoh, this ratio increases with higher under cooling. At very small
under cooling, which is close to let us say 0. Then, the ratio is close to around 1. At high
temperatures, the ratio is close to 1; that means both incoherent and semi coherent can
grow. And once that happens, you end up in spherical. All directions, the growth is like a
isotropic growth whereas, at high under coolings, such an isotropic growth is not
possible. And as a result, you see that one direction it grows, which is incoherent
direction.
The semi coherent directions, it does not grow and as a result, it gets elongated. So, that
is what, we call it as Widmanstatten structure. So, Widmanstatten structures come in to
picture, basically because of that. So, this is how, you would see different types of α
coming out of the γ just because of the growth rates being different at different
temperatures; because of the necessity for a semi coherent interface to maintain the
coherency. An incoherent interface need not have to maintain coherency. So, as a result
it can grow faster.
Student: Why you said incoherent in this case and semi-coherent in this way
That is what I am saying, if you are looking at a plate, the nucleation is only on longer
side.
The nucleation is not perpendicular to it. Let us think of a case like aluminum alloys, we
add TiB2 to nucleate aluminum, which is called inoculants. Have you heard of it? Grain
refiners, we called them. So, once I add TiB2 to aluminum, aluminum is nucleating on
TiB2. Where does it nucleate?
Student: Basal plane
Basal plane The basal plane of TiB 2 has the same orientation, the same atomic packing
with respect to the 1 1 1 plane of aluminum. So, that means if I think there is plate of
TiB2, on which aluminum has nucleated.
Aluminum will nucleate like this in such a way that, Al(1 1 1) plane of aluminum is on
the TiB2(0001) (see the top figure in above image). Once it has nucleated, what about the
other planes? If aluminum has nucleated on a large TiB 2 surface the plane, which is
perpendicular to 1 1 1 plane. The plane which is perpendicular to 1 1 1 plane is some
other plane. You can look at the crystallography and find out, what is that plane which is
perpendicular to the 1 1 1 plane?
622
If you look at that particular plane and consider what is perpendicular to basal plane of
TiB2 which are basically 1-101 type of planes, you can look at them. So, the planes
perpendicular to Al(111) and those perpendicular to TiB 2(0001) may not have
coherency.
As a result, on the two sides of aluminum plate, you may not see any coherency unless
both of them are cubic. If the surface is cubic and Al is cubic, both are FCC, then the
plane which is perpendicular to 1 1 1 plane will definitely have coherency because both
are FCC you will have this happening. If not, the planes perpendicular to Al(111) and
those perpendicular to TiB2(0001) may not have coherency and as a result, the sides of
Al plate will be incoherent. So, it will nucleate there and then, grow; but in actually in a
liquid to solid transformation, the aluminum which is nucleating on the TiB2.
You do not see it as a growing as plates, basically because of v incoh/vsemicoh; because this
nucleation is and growth is taking place at temperature just below the melting point. At
that temperature, the diffusivities are very, very high. So, as a result ultimately aluminum
grows like a grain, equiaxed grain on TiB2 particles; whereas if you look at, at low
temperatures where such a nucleation has taken place. At a low temperatures, it is easier
for it to grow in incoherent direction rather than it to grow in coherent direction. If it has
to grow in that direction continuously, you need to have a maintenance of the coherency.
With that more or less, we have understood eutectoid.

Student: Sir, eutectoid in iron carbon alloys, we are getting chunky carbides because of
the interfacial energy of cementite and liquid state
Yes, that is not in eutectoid. Please, eutectic. Yes, tell me.
Student: The same interfacial energy, that logic has to be applied for eutectoid also. Why
we are getting plate like?
The reason is, because it is a solid state. You have to consider the strain also here; you
have to consider the strain also and if you consider the strain and particularly, because
there you are talking of nucleation of cementite inside the liquid. Is not it? The cementite
if you consider a eutectic composition; let us say 4.3 % composition, it is forming inside
the liquid. Here, it is forming inside the γ. If it is forming inside the γ, then you need to
consider, what is that that interfacial energy?
623
So, that means I have to consider the γ/α interfacial energy and γ/Fe 3C interfacial energy
and then α/Fe3C interfacial energy. These are the three interfacial energies, I should
bother. In eutectic case, what I have bothered is L/Fe3C interfacial energy and then, L/γ
interfacial energy and γ/Fe3C interfacial energy. These are the three there. So, in eutectic
case if you consider L/Fe3C was lower and here in eutectoid case, if you consider α/Fe 3C
is lower, that is it as simple as that. It looks at which is the lower interfacial energy and
so whichever is a lower, it would happen.
So and one more thing is, once it nucleates how it grows; it grows as chunky or not. It
depends on the growth rates that is what I have just now said. Some of them would grow
as needles. There are number of cases in eutectics, where you have needle like
microstructures or rod like microstructures. Aluminum silicon is an example. There also
examples of lamellar eutectics. Is not it? You have aluminum copper is an example,
where you have lamellar eutectics; so, where solid-solid interface is lower energy, than
others so as a result you would see that happening.
That means two solid interfacial energies are lower. In aluminum copper, two solid
interfacial energies are lower. Basically, because both are if you consider, one is a θ
phase which is actually orthorhombic and α is the cubic. But there is if you look at high
temperature, the interfacial energies of both of them are close to each other. And as a
result, they can come out as infact, θ is not even orthorhombic; it is
624
BCT; it is tetragonal; sorry orthorhombic is Fe 3C. So, tetragonal structure has co-relation
with a cubic structure excepting the c axis, everything else is similar. Infact you take a
cubic, pull it; that is how, you get a tetragonal structure.
So, that is how you can see that, there is a possibility of α-θ lamellar eutectic coming.
That is how, you get a lamellar eutectic whereas, here you cannot get lamella eutectic; as
γ and orthorhombic, there is no correlation, and compositions are also entirely different.
As a result, you will not see a possibility of a lamellar there. When do you say interfacial
energy is lower, when the structures are similar. For example, Ni 3Al nucleating in a γ
phase, we say it forms very easily; because the structures are similar. If the structures are
not similar, the interfacial energies cannot be.
Infact if you see that eutectic, you would see both the phases that are coming out are
cubic let us say. You would definitely invariably end up in lamellar eutectics, because
the interfacial energies will be similar. For example, if I take a copper silver eutectic,
copper silver eutectic will definitely be a lamellar. If you go by you can see, because
both are FCC and there are no inter metallics in that phase diagram at all. This is a
simple eutectic phase diagram.
Student: Do we see divorce eutectoid?
If you want to call it as divorce eutectoid, in principle bainite is a divorce eutectoid. Is
not it? There is no lamellar structure.
Student: Is the bainite formation a eutectoid reaction?
yeah actually interesting question.
We do not call it actually as eutectoid; because it is not happening. Infact if that is the
case, the cementite formation eutectic reaction, where cementite is coming out first and
the remaining liquid is becoming γ. Can I call this as eutectic? Because both of them are
not forming at the same time. So, ultimately you have to see the final microstructure,
liquid giving two solids finally. So, if you look at it, the γ giving two solids finally; one
is a ferrite, another is pealite. Infact we talked about it even in peritectic reaction.
Peritectic reaction is finished; the moment a peritectic layer is formed.
There is no more peritectic. You think of this case of liquid + δ giving you γ, the
moment γ forms on the δ; where is the peritectic reaction? But at the same time, liquid
and δ will vanish and finally, you will get γ. How does it happen? We said this happens
625
because of the diffusion. Do I call this as peritectic reaction? So, we call it as peritectic
transformation and solve our
626
problem. Is not it? So, that is what it is. So, in principle a bainite can be called as you
know eutectoid. It is only semantics; it is only how you look at it.
Finally, you have; but if you look at it, the whole transformation is first the γ is giving
you α of a same composition, which is a polymorphous type of a transformation and
from polymorphous transformed α, cementite is precipitating out; that is what is that
actual transformation that is taking place. So, in principle it is from that point of view, it
is not actually eutectoid. But it is the ultimately the mixture is α + cementite mixture
with a different morphology. Anything else?
Student: In bainite formation, the Ferrite is the active nucleant
Yes
Student: And in the eutectoid, in the pearlite formation it is cementite. Why is there a
difference
Yes, again it is not very difficult to understand.
See, if you look at the two transformations, γ  pearlite; γ  bainite; in γ  bainite, we

have under cooled to such an extent that the carbon is not able to move much. Because
carbon is not able move much, long range diffusion short range possibly it can still do;
long range diffusion is difficult. As a result, γ first becomes α of the same composition
polymorphous. In γ  pearlite, the carbon is able to move. So, what happens is carbon
would come to the grain boundaries of γ and form a small regions of enriched carbon
(see image above). Once that enriched carbon region reaches that of 6.67 or 25 atoms out
of 75.
627
Initially, the number of atoms of carbon is very, very small. It is 0.8; that is weight %. If
I convert it into atomic %, you would get still possibly about 2 or 3; so, 3 atoms or 4
atoms of carbon in 100 atoms of iron, something like that. That is the initial
628
state. From that initial state, more carbon atoms come from a different region and so that,
that region becomes 25 atoms. Then, you can easily form nucleus. That is the reason
why and carbon can happily move at temperatures closer to 723 ºC. So, that is the
reason why cementite is a nucleus there; whereas, here in γ  bainite, cementite is not
able to be a nucleus.
Carbon is not able to move and if you further go down, you get Martensite. Because
even that small short range diffusion that we can see there is not able to happen and
finally, you get a diffusion less. Tomorrow, we will talk about this. Anything else?
Student: Sir, to improve the diffusivity of an atom, of a particular atom in a system, other
than thermal activation, is there any other possibility?
Yes, you introduce certain let us say vacancies; there are people who have done by…
Somebody may say that all are inter related. For example, if I do irradiation or if I do
ball milling, let us say I take γ austenite powder and do ball milling. I introduce lot of
defects in to it, dislocations. So, the moment there are large number of dislocations, you
have what is called pipe diffusion.
Have you heard of it? Pipe diffusion is diffusion through the dislocation core.
Dislocation is nothing but a row of vacancies. Is not it? So, if you introduce large
number of dislocations, then the diffusivity can be increased. So, that is the reason why
many in ball milling, many reactions that can happen; that usually are supposed to
happen at high temperatures can happen at low temperatures. Another thing is you make
them nano; you introduce large surfaces and reactions can happen. For example, a simple
example I can give you if you take iron oxide + aluminum, there is a reduction reaction.
What do we call this?

Student: Metallothermic
Metallothermic aluminothermic reduction; thermite reaction people call it. This usually
occurs above the melting point of aluminum. The same thing can occur much below the
melting point of aluminum, the moment you make both of them nano. This is what, we
see in ball milling. You take iron oxide and + aluminum mixture, put it into a ball mill;
ball mill it. After sometime, at room temperature you see this reaction happen at low
temperatures. Just because you have enhanced the diffusivities; because surface diffusion
can happen now, and diffusion through dislocations can happen; these are other ways
will stop now.
629
Prof. Dr. B.S.Murty
Module #01
Lecture #26
Martensitic Transformations
Last class we talked about the eutectoid transformation; the logical extension of that is
the martensitic transformation. Let us look at that today.
(no audio between: 00:25-00:38)
As you all know, this is one such transformation where diffusion is not involved. It is a
diffusionless shear transformation. When all diffusional mechanisms fail, that means you
come to a temperature where diffusion is not able to happen inside the alloy. At the same
time the parent phase is not stable, so it wants to transform to something, and there is no
other go, but to go by some mechanism, by which it can transform. That mechanism is a
shear mechanism.
So, You see this happening in many alloy systems. The most classic alloy system is what
we know the steels; the iron-carbon alloys. But, you will also see not only in that the
shape memory
630
alloys if you have heard of. Shape memory alloys the basic mechanism that gives shape
memory effect is the martensite. You see the nickel-titanium; people call it as Nitinol is a
very classic example. Iron-Carbon is another example, even in Cu-Ni-Al, kind of alloys
or Cu-Zn-Al kind of alloys, there are number of copper based alloys where you see. For
example, copper zinc also, you can see martensite formation when you quench it from
high temperatures.
So All these are can be grouped basically into one class of transformation, which we call
it as martensitic transformation, where you have a diffusionless, shear; it is diffusionless
transformation and it is a shear transformation. Let us see many of you know it, let us see
how this shear takes place? If you look at the iron-carbon example, if you see the parent
phase that is undergoing the transformation is the austenite. It is the austenite which
gives you martensite. In fact, in the last class also we have talked about it.
If the austenite, the cooling rate is such that the austenite already gives you the pearlite
or maybe you have cooled it to a temperature, where you have held it for a long time to
g, get a bainite. Under such conditions, how do you get martensite? We said the only
way to get martensite from a bainite or from a pearlite is to take it back to the
austenization temperature and see that all the pearlite and all the bainite get transformed
to austenite. Then the transformation to martensite is possible. The reason is that it is
only one structure transforming to another structure by a shear. Whereas, if you consider
pearlite or bainite, pearlite or bainite are two phase mixtures, so it is not possible for a
two phase mixture to shear, to give you a one single phase. So, Whenever, we are talking
about shear transformation, it is one structure where you have a shear along a particular
direction taking place to give you a martensite.
If you start thinking of lattice points in a FCC unit cell (at the corners and face centers)
and then extend it to one more unit cell like and look at these two unit cells together (as
drawn in the Fig. above). If we join the lattice points of two adjacent unit cells as shown
in Fig. above, in principle what you have generated now is? What kind of a structure unit
cell is that?
Student: Tetragonal
631
It is a tetragonal unit cell and it is not only a simple tetragonal; it is a BCT, because there
is one lattice point at the center. The face centered lattice point of FCC unit cell becomes
a body centered lattice point for the body centered tetragonal.
So, now you have a BCT generated out of an FCC. Any FCC unit cell can be thought
like this. You do not need a shear there. What is a unit cell? Unit cell is an imaginative
way of representing lattice points I mean the way atoms are arranged. I have atoms
arranged in a in a 3 dimensional structure like this and that I can visualize it as a FCC,
and the same FCC can be visualized as a BCT. The Why do not I call this cell as BCT?
Why should we call it as FCC?
Every FCC unit cell can be thought as a BCT unit cell; every FCC unit cell. Why are we
are calling it as FCC? Why cannot we call it as BCT? Have you ever thought about it, at
the second year B tech level, when people are teaching you unit cells? Every FCC can
be, thought of as BCT if we join the lattice point in a different way.
Student: You cannot have a continuous network, symmetry cannot be there
You can have continuous; see when I am saying joining two unit cells, there are no unit
cells inside a lattice. In a 3-dimensional arrangement of atoms, there is no unit cell. The
unit cell that you are imagining that this is a unit cell.
632
I give you a simple example; a two dimensional unit cell. Imagine all the atoms are of
the same size, it is same type of atoms and you have a two dimensional unit cell (See the
arrangement of atoms above). Now, the question is I can choose this as a unit cell (A, in
the Fig. above). I can choose this as a unit cell (B), I can choose this as a unit cell (C).
Student: Sir, it should reflect the symmetry
So, I have shown you three possibilities. One can imagine innumerous possibilities.
Which one should I choose as a unit cell?
Student: Lattice parameter
Student: The smallest unit cell which represents the three dimensional symmetry
Symmetry; the smallest unit, which represents the symmetry of the unit cell; you may
say that what is the symmetry of it. How do I know? The smallest unit which has the
highest symmetry, among those arrangements of lattice points, which has a highest
symmetry among those and it should be the smallest.
Now, comes to this. If you look at all this, what is the symmetry of unit cells A, B and C
above? If you look at it, A has a fourfold symmetry. B does not have a fourfold
symmetry. A rhombus cannot have fourfold symmetry. When will you have fourfold
symmetry? Can you tell me? In a two dimensional object, when will you get a fourfold
symmetry? What is the condition of that geometry which gives you fourfold symmetry?
Student: a = b and angle should be 90º
Yes, So, angle should be 90º and a should be equal to b then only you will get fourfold
symmetry. Just because angle is 90º , you need not have to get fourfold symmetry if a is
not equal to b. That is what is a rectangle is where a ≠ b, but angle is 90º. So, that will
give you what kind of a symmetry?
Student: Two fold
It is a twofold symmetry. This is what the B has.
And If I now look at this particular unit cell B alone; among A and B and if I compare it,
B has a twofold symmetry not only that, if I consider the number of lattice points per
unit cell, is how much in B? For that rectangular unit cell; how many lattice point per
unit cell?
Student: six
Six; lattice points per unit cell; 2, 3, I am getting all kinds of numbers. Let us see.
633
Student: Sir it is 2
It is 2 and why it is 2?
Student: Each corner will contribute 1/4, so 4 x 1/4
Corners, one-fourth, wonderful
Student: + center, 1/2 + 1/2
1/2 + 1/2. We are looking at a two dimensional lattice. So, 4 x ¼ + 2 x ½ = 2
634
In fact, that is the reason why if you have ever seen a face centered tetragonal there is no
face centered tetragonal. Have you ever thought why there is no face centered
tetragonal? Because the moment I have a face centered tetragonal, I can draw the unit
cell exactly like this (as shown in Refer slide time 00:24) excepting that the c-axis is a
little longer and that is it. longer or shorter than a and b and that is what is a tetragonal
lattice is. Now, if I join top face centers and bottom face centers to the respective
corners, I can see a BCT. Every FCT, I can see a BCT. As a result, we cannot say that
there is FCT, because FCT can be represented as a BCT and as a result FCT does not
exist. Otherwise, in principle, if you have ever really broke your head for this problem
that there are 7 crystal systems. And each crystal system can have 4 unit cells, 28 unit
cells should have been there; 28 Bravais lattices should have been there but, how many
Bravais lattices we have Prerna?
Student: 14
14 only. Why only 14? Why not 28? It boils down like this; one by one you can take
each unit cell and star discarding it with this kinds of arguments. So, FCT is not possible.
FCT can be thought as a BCT. Now, the question is why FCC cannot be thought as
BCT?
So, if you want to remember the basic choice of a unit cell, we just now said that it
should have less number of lattice points; it should be the smallest unit cell with the least
number of lattice points per unit cell. And It should have the highest symmetry among
all possible units that you can think of. For the atomic arrangement shown in (Refer
Slide time 07:10), A is such a unit cell. For some other atomic arrangement you may
have something else. For that particular atomic arrangement where the atoms are equally
spaced, and axes are equal, if that is a case then A is the best unit cell. If axes are not
equal then you may get a different unit cell. That is how orthorhombic comes into
picture; the angles may be 90º, but a ≠ b. So, you would see you would have a different
type of unit cell where the a ≠ b.
So, this is one what else? There is One more, when you consider the unit cells; I do not
know whether you have thought about it. This comes into picture mostly when you have
more than one element in the picture. For example, if you have an ordered crystal
635
structure, a simple example is sodium chloride. You have sodium atom; every sodium
atom is surrounded by chlorine atoms and every chlorine atom is surrounded by sodium
atoms. Now what is the unit cell if I have to consider. I do not consider the composite
unit cell which has both sodium and chlorine atoms. I should consider that unit cell
where only sodium atoms are considered and another unit cell where only chlorine atoms
are considered. because a unit cell is always a unit cell of like atoms . That is why
whenever we talk about ordered structures we talk about sub lattices and a super lattice.
A super lattice is that which is composed of both the elements.
For example, the moment FCC unit cell or a BCC unit cell, for example, easy to
understand, imagine a BCC unit cell with one lattice point at the center. What is the
example of an ordered BCC unit cell? I mean ordered unit cell based on BCC,
Student: Copper Zinc
Copper-Zinc,β brass, another example
Student: Cesium chloride
Cesium chloride, so NiAl; these are all of the type. In that the moment it becomes
ordered, I know pretty well that if I am taking a copper-zinc, copper occupies, let say the
corners, zinc occupies the body center. Now, what is the unit cell? Once in all the unit
cells, all the corners atoms are occupied by copper and all the body center atoms are
occupied by zinc. I cannot call it any more BCC.
It is no more BCC, because the body centre position is not copper; the body centre
position is zinc, whereas body corners are copper. So, we say actually there are two unit
cells. One unit cell made of copper atoms; another unit cell made of zinc atoms. What
will be the unit cell of copper atom? It is simple cubic cell. What is the unit cell of zinc
atoms? It is again a simple cubic. So, you can see two simple cubic unit cells; two simple
cubic unit cells and each one displaced with respect to each other by a 1/2, 1/2, ½ (see
the ordered unit cell drawn in top left corner, Refer Slide Time 07:10).
Join all the body centers of eight unit cells, which are surrounding this, you will see that
eight body centers put together, you will get one unit cell. That would be a zinc unit cell
and zinc unit cell is shifted with respect to the copper unit cell by a 1/2, 1/2, 1/2 distance.
So, these are two unit cells. One is copper unit cell, another is a zinc unit cell and both
are simple cubic unit cells. And put together this one is what we call it as a supper
636
lattice. These are the two sub lattices. Copper sub lattice, zinc sub lattice; we need to talk
in terms of that. So, that is
637
why you should never say that copper zinc ordered copper zinc has a BCC structure; it is
wrong. It does not have a BCC structure.
In fact, because of this, I now tell you that if you start considering of what we talk
sometime back that why ordered structures have low deformability; I said the Burgers
vector changes. If I consider this unit cell, the Burgers vector a in a BCC unit cell; what
is burgers vector?
Student: 110
1 1 0, BCC unit cell; 1 1 1, 1/2 <1 1 1> . 1 1 1 direction is the closest packed and that
changes now the moment it becomes ordered that it changes to the simple cubic, burgers
vector becomes [1 0 0]. So, 1/2 [1 1 1] < [1 0 0]. The burgers vector becomes longer.
So, you would see suddenly the moment it becomes ordered, the deformation occurs on
different planes. Now, it starts searching for different directions where slip can easily
occur. And As a result, some of the directions; obviously, you will never find a shorter
Burgers vector once you are a super lattice. So, as a result you will see the deformation
becomes more difficult. And You will see many other things. The reason for why we do
not call FCC as a BCT is what now? Based on whatever we have discussed, FCC, why
do we not call it as a BCT? It is symmetry; the symmetry of a cubic. What is the
symmetry of a cubic system?
Student: Four
Four, Did we talk about it? No, Did somebody talk about it at some stage? What is the
symmetry of a cubic system? In fact, you know, always you might have learnt cubic is
a = b = c; α = β = γ = 90º. That is a very primitive way of learning, I would say.
The best way to know about crystal systems is to know the symmetry of it. In fact, then
you will know when you say cubic, tetragonal, orthorhombic. Why do we actually learn
them in that order? It is because the symmetry decreases in that order. Cubic has the
highest symmetry, triclinic has the lowest symmetry. That is why we talk about triclinic
at the end and that is why you do not see most of the elements actually occupying a
triclinic system. Every element, tries to have as close packed structure as possible as
long as its electronic configuration allows it. That is why you see if you see the whole
periodic table, out of some 105 elements or so, at least 80 to 85 elements fall into cubic
or hexagonal. It could be FCC or BCC and hexagonal.
638
So, you see these are the three, which are the most popular. It is basically, because they
are the most close packed; I mean not just close packed, they are the most symmetric.
So, if you look at cubic, what is the symmetry of cubic?
Student: Threefold
Threefold; how many three folds? Four three folds it has. Where are these four three
folds?
Student: Diagonals
Diagonals, body diagonals, you have eight corners. Join any two opposite eight corners
of the eight corners; join any two opposite corners out of the eight corners, then you get
a direction, which is a [1 1 1] direction , and that [1 1 1] direction is a threefold direction
and because there are eight corners, so four such [1 1 1] directions are possible.
In fact, actually how many 1 1 1 directions are there total?

Student: Eight
Total eight are there, eight 1 1 1. In fact, there are eight 1 1 1 planes and that is why we
call it as octahedral planes. 1 1 1 planes are called octahedral planes and the reason is
because they are eight of them, but we are only considering four here because if I
consider 1 1 1 and -1 -1 -1; they are exactly the same excepting that they are in reverse
direction.
So, both fall on each other except that the directions are different. So, they can be
considered as one axis only. So, out of the eight you will see 4 pairs like that. If I take
-1 1 1 and it will be parallel to what?
Student: 1 -1 -1
This will be parallel to 1 -1 -1.So, these two directions considered as together. They will
be basically same direction opposite to each other. So, like that you will see four sets and
that is why we call them.
And In addition to four threefold axis, what else you have? in a cubic? There is another
important symmetry.
Students: Three fourfold
You have three fourfolds. You see every face is a square; every face is a square in a
cubic. So, the x y z, all the three directions are fourfold axis in a cubic. So, you have
three fourfold axis and four threefold axes. This is the highest symmetry that any crystal
can have. That is why the moment you take cube and then pull it what happens to the
639
cube? It becomes tetragonal. Once it becomes tetragonal, what happens to the crystal
symmetry?
Student: Threefold disappears
Threefold disappears; perfect, threefold completely disappears.
640
If I consider (1 1 1) plane (drawn in pink colour), that is the one, which has the threefold
symmetry. What is the shape of that plane? A triangle, a triangle will have a threefold
symmetry when ?
Student: All sides are equal
all the three sides are equal; when it is equilateral triangle. If it is not an equilateral
triangle you will not a have threefold symmetry. In a cube, you have a (1 1 1) plane and
is an equilateral triangle. In a tetragonal, (1 1 1) plane is not a equilateral triangle,
because you have pulled it, the two sides will be longer than the third one in a tetragonal
and because of which you do not have an equilateral triangle.
Once you do not have an equilateral triangle there is no threefold symmetry. So, all the
(1 1 1) planes have completely got converted into non equilateral triangles; isosceles
triangles. They are all isosceles triangles and once they are isosceles triangles; obviously,
you do not have a threefold. So, you do not have threefold symmetry at all, the four
threefolds that we talked about in a cubic, just by pulling it by some few percent, you
have changed it completely, you have vanished. I mean you have destroyed the
threefold symmetry. Then what about fourfold symmetry? Only the top and bottom faces
are squares now. The remaining four faces are no more squares because you have pulled
it. The remaining four faces, what have they become?
Student: Rectangles
They have become rectangles and rectangle cannot have a fourfold symmetry.
641
So, only the [0 0 1] direction, which is a c axis, that alone will have a fourfold symmetry,
and [1 0 0] and [0 1 0], they will not have a fourfold symmetry. So Whereas in a cubic
642
both [1 0 0] and [0 1 0] and [0 0 1], all three of them had fourfold symmetry. In
tetragonal, you will have only one fourfold symmetry. You see from cubic to tetragonal
what a degradation in terms of symmetry. You simply have one fourfold symmetry and
that is it. You may say that it also has a twofold symmetry, because rectangles are there.
Two twofold symmetries are there, then why am I not talking about them?
Student: It is contained inside fourfold
It is not contained inside.
The twofold axis the [1 0 0] is a twofold, [0 1 0] is a twofold, these two are in two
different directions, which are at 90º to [0 0 1], which is a fourfold. They are not
contained in it.. But why do we not talk about it. Because whenever we talk about the
symmetry of a unit cell, we talk about what is the highest possible symmetry of that unit
cell. Because, fourfold is possible, we do not talk about twofold though twofold does
exist inside the lattice; fourfold is possible that is why in cubic also, we say three
fourfolds, we do not actually talk about four threefolds, four threefolds are as a result of
three fourfolds, because, a = b = c and all the angles are 90º, the three fold is also coming
in the way.
So, you can see that it has one fourfold and now from there if you go to orthorhombic
you will see then again there is a further degradation. There is no more fourfold at all in
orthorhombic. Why? Why there is no fourfold?
Student: All the faces are rectangles.
All the six faces are rectangles. Why because a ≠ b ≠ c. But all the angles are 90º.
α = β = γ = 90º, because all the angles are 90º, you have rectangles. If there is no 90º
then you would not even get rectangles . And because you have got rectangles, you have
some symmetry. What is that symmetry? in a rectangle, you have a twofold symmetry
and so what is the symmetry of an orthorhombic?
Student: Twofold
Twofold?
Student: Three twofolds
Three twofolds, like the way we said for cubic there are three fourfolds, here there are
three twofolds.
Three mutually perpendicular twofolds; a, b, c, all the three are twofolds, because all of them
are rectangles. So, like that you would see from one to other if you start going
643
orthorhombic, you go to rhombohedral. How do you get rhombohedral structure? from a
cubic structure. If somebody gives you a cube and ask
Student: Pull along the diagonal
Pull along the diagonal once you pull along the diagonal. What happens now? All the
directions, all the sides may remain equal, but angles will not be any more 90º. Once the
angles
644
do not remain 90º, you cannot have a twofold symmetry you cannot have a fourfold
symmetry.
So, there is no twofold, no fourfold in a rhombohedral, because angles are not 90º. For
example, if you have kind of a structure, like unit cell B in (Refer Slide Time: 07:10)
there is no twofold here. Rhombohedral is almost an extension of this rhombus in 3D
and it cannot have a twofold symmetry, because this angle is not 90. What is the
meaning of a twofold symmetry? When I rotate a particular lattice point around another
lattice point by 180º do I come back to another lattice point inside, that is what is a
twofold symmetry.
So, that is possible only when the angles are 90º. So, you will not be able to get that. So,
once you do not get that then obviously, you have no twofold, no fourfold, but you have
one threefold symmetry in rhombohedral. You can geometrically prove it there is one
threefold symmetry. Similarly, when you go to a hexagonal, you have a six-fold
symmetry there, because the basal plane has a six-fold symmetry and at the same time
the rectangular planes, which are surrounding the basal plane. I mean the top and bottom
basal planes are connected by six rectangles.
So, those rectangular planes also will have a two fold symmetry, but we do not talk
about that. We say that hexagonal has a six-fold symmetry, because twofold symmetry is
much less than a six-fold symmetry. But Though we say six-fold symmetry is higher
symmetry than a four-fold symmetry, we do not put it above cubic, because cubic has
three fourfold symmetries, whereas, in hexagonal only one of the axis has a higher
symmetry than fourfold. The remaining axis have a lower symmetry than the cubic that
is why cubic is always above the hexagonal in terms of symmetry.
So, whenever certain element wants to choose a structure, the first structure it would like
to choose is cubic and if that is not possible then it will choose something else. So, that is
how it is then you go to monoclinic. In monoclinic, you have one 90º angle. If you
remember, a ≠ b ≠ c, β is 90º the remaining two are not 90º. So, because there is one 90º
angle, you will get one twofold symmetry. So, monoclinic has one twofold and then you
come to triclinic. What is the symmetry?
Student: Zero
Zero symmetry ?
645
Student: One symmetry
It has one fold symmetry. There
646
is nothing like has zero symmetry. Every one of us has one fold symmetry; you rotate
around yourself, 360º, you will be able to generate your own self once again.
Every object has to have a one fold symmetry. One fold symmetry is what we call it as
trivial symmetry; it is a symmetry, which is supposed to be there everywhere, because it
is a 360º rotation. Any object, when you rotate it by 360º comes back to itself. Am I
right? So, coming back here, we say FCC is not BCT, because cubic, the same
arrangement I can also think of as a FCC. And FCC has a higher symmetry than a BCT.
So, I will consider these atomic arrangements as FCC rather than a BCT; same atomic
arrangements, when there is a choice between two unit cells possible. I will choose that
particular unit cell, which has the highest symmetry because that is one of the criteria to
choose unit cells.
So, that is why it is called a FCC. Now, when there is a shear that takes place, the shear
actually occurs along the [1 1 0] (drawn with pink arrows in above image), which is a
closed pack direction. So, along that direction you have a shear and once the shear takes
place the FCC permanently becomes BCT. Because once the shear has taken place, it is
no more FCC; it has become BCT, because of the shear; not just because of the atomic
arrangements, No, because of the shear that is taking place along that. If the shear would
have taken on any other plane or in any other direction, you would have got some other
structure possibly, but because it occurs in the closest pack direction, atoms are close
pack there, you would see that it becomes BCT.
Once FCC becomes BCT, what is the c by a ratio?

Student: 1.414
1.414, perfect, why?
cBCT = aFCC (Refer to figure with two unit cells in the image above)
Only thing is now a and b of the BCT are what when you compare with the original FCC
structure, ?
Student: 1/2 the face Diagonal
1/2 of the face diagonal, face diagonal is a/√2
647
So,
aBCT = (a/√2).aFCC
and cBCT = aFCC
So, (c/a)BCT = √2
But such a BCT structure with a c/a of 1.4.14, it is not stable at room temperature,
because in the phase diagram, I do not see such a structure.
So, what happens is, the system what is stable at the temperature in the phase diagram is
a BCC structure. At room temperature, in a iron carbon system what is stable is a BCC
structure; BCT is not stable. So, what happens is this particular BCT structure wants to
transform to a BCC structure. That is why the whole martensitic transformation is
composed of two processes; one is a shear, which converts the FCC into a BCT, then
another process whereby this BCT wants to get converted to BCC. What is that that can
convert a BCT into a BCC? If you want a BCT to become BCC, what should happen?
Student: Carbon
Forget about carbon, we are talking of only structure;
Student: c has to become a
c axis has to shrink and a and b have to elongate, so that the c/ a = 1. If c can shrink
completely without a and b elongating, I have no problem. Ultimately, I want the c = a.
That is where the whole problem is in iron carbon diagram. In iron carbon steel, if you now
look at this whole structure and see where carbon atoms are
648
sitting, you will see very interesting phenomenon. Carbon atoms in FCC sit where?
Student: Octahedral sites
octahedral sites. Let us look at what are the octahedral sites. One of the most common
octahedral site in a FCC is what? is a body centered position.
Body centered position is a octahedral site. In this (Fig. 1 in above image), the body
centered position is this (point B1 and B2). The body centered position of the FCC are
these two points that is where the carbon atom will sit. And What are the other
octahedral positions?
Student: Edge center
Edge centers; so that means, E1, E2, E3 and E4 are edge centers.
I am not considering other points, I am only considering that corresponding to the BCT,
so that means, once the shear has happened, carbon atoms have not moved anywhere,
they are where they are. That is the meaning of martensitic transformation. If the carbon
atoms can move, why will shear occur; martensite would have not found. It would have
found bainite; if carbon atoms could have moved it would have formed a bainite.
Because carbon atoms are not moving they are where they are exactly, they will remain,
once the FCC becomes BCT, the carbon atoms are actually sitting at the body centers
and edge centers indicated above. Are there any other positions possible for the carbon
atom in this particular unit cell? No, that is it, those are the edge centers those are the
body centers. Once you consider that, out of all these six positions, 4 of them are at c/ 2
positions of the BCT. They are at the center of the c axis, four of them, and because
carbon is sitting at the center of the c axis; for the c to shrink is not easy; c cannot shrink
now, because carbon is sitting there; it cannot shrink and because it cannot shrink, BCT
cannot become a BCC.
That is why whenever you look at a martensitic transformation, in fact, if you carefully
observe a typical martensite like a 0.8% carbon steel martensite, what is the c/a ratio in
that martensite ?
Student: 1.6
1.6? You have zapped me. The maximum possible is this √2. When a FCC shears, this is
what is the maximum. You cannot have beyond that and the actual martensite will not
have that. Go back to any of your physical metallurgy books, look at the c/a ratio, and
we always say c by a ratio increases with
649
increasing in carbon content; if u remember c by a of martensite increases with
increasing carbon content. Why? because the more and more carbon content the
probability of these voids occupied by carbon is going to be more and more.
So, there is more and more difficulty in the shrinking to occur for the c axis and that is
the reason why c/a ratio does not decrease. Usually, c by a ratio is in the range of 1.1 –
1.2 or something like that 1.05 also people observe, depending on what is the carbon
content. So, some shrinkage will occur, but not all. As a result c/a will not come to one.
When will it come to one, when carbon is not there that is why any of the iron based
alloys where carbon is not there then you can have a martensite, which is a BCC
martensite. It will be a soft martensite. All martensites need not be hard; BCT martensite
is hard because of its structure and because of the strains associated with the shear. There
is a strain associated with the shear even in BCC that alone will give you an additional
strengthening, but otherwise as such it is a BCC.
Student: Why do we call it Martensite?
It is martensite because it formed by shear that is it.

Student: What exactly causes the shear?
Because the atoms are not able to move. The atoms are not able to diffuse and FCC is
not stable at the temperature. So, it wants to finally, come to an equilibrium shape or
equilibrium structure, which is a BCC structure. A normal pure iron, when you take a
pure iron at high temperature, above 910 ºC, it is FCC, below it is BCC. How does this
FCC to BCC transformation take place when I am cooling up from above 910 ºC to
below 910 ºC? How does the transformation occur?
Student: Short range diffusion
It is by short range diffusion. The atoms, which are sitting at the body centered face
centered positions move to the body centered positions and as a result you generate a
BCC structure. From every one FCC unit cell you generate two BCC unit cells. Because
obviously there are four atoms in FCC, four lattice points per unit cell. In BCC, you have
only two lattice points per unit cell. So, by this kind of a atomic re arrangements in a
short range, which is of the order of one unit cell dimension movement of atoms; atoms
do not move beyond that and that is what is a short range . But, the same FCC and if I
quench it to room temperature, the same pure iron; FCC γ iron, I quench it to the room
temperature, then this iron wants to become BCC, because FCC is not stable at room
temperature. It now forms a BCC by a shear; it first does a shear then it gets converted to
650
BCT and this BCT, because there are no carbon atoms, because BCC is the stable
structure, not the BCT, it transforms to a BCC.
651
So, ultimately the structure is BCC, but this BCC has been formed by a shear process
that is why this BCC will be stronger than the original BCC, which has been obtained by
a normal cooling. So, by quenching FCC pure iron, you would get a strengthening that
martensite will be stronger than the original BCC iron. But, it would not be as strong as a
BCT martensite in the steels.
And for example, if I take iron nickel alloys; iron nickel alloys also you see the same
type of transformation taking place.
Student: FCC is the room temperature stable phase
No, in our iron carbon diagram what is the low temperature state? You remember the
phase diagram?
Student: Because of the free energy
Student: Entropy of
There is nothing like FCC is a stable state; depends on which particular electronic
configuration of that element. For example, there are metals which are forming BCT. Is
it not? There are so many, like Sn is a BCT structure; it is not even a cubic structure. So,
different elements get different crystal structures depending on what their electronic
configuration, which allows it to have. If it allows it to have FCC, yes. We say that cubic
is the most popular; among the cubic some can be FCC some can be BCC.
One thing I want to tell you FCC and BCC, the symmetry is the same; whenever we talk
of symmetry we do not talk of symmetry of a unit cell, we talk of symmetry of a crystal
system.
Student: Sir, But the entropy of the BCC is more, because of loose packing, which is
supposed to stabilise?
If you look at the entropy, just based on loose packing. I do not think that is the right
way of looking at it. Yes, you can say that because it is loosely packed, the heat capacity
will be higher could be higher. That much only I can say, because atoms can vibrate
more easily ,
Student: Will that doesn’t result on entropy
It need not, it will also result to higher enthalpy. Enthalpy is also ∫CpdT. So, it is not that.
To the some extent there is more open space, but the thing is, it is not that more open
space does not mean more disorder. Is it not? It does not mean that you cannot say that
BCC is more disordered than FCC; you cannot say that.
It is as ordered as an FCC is in terms of atomic ordering.
652
Student: But we say the open space structure is more stable
Yes it is like for example, liquid that is where the problem is. In principle, that is why
actually FCC again becomes BCC at higher temperature; FCC again becomes BCC. It is
magnetic transitions that are there in iron, which make this happen BCC to FCC and F C
C again going to BCC. But there are structures where you would see for example, if you
take cobalt; cobalt is HCP at room temperature and becomes FCC at high temperature,
but there are also examples where the HCP at low temperature becomes a BCC at high
temperature; titanium is an example.
653
But there are also examples where it remains HCP up to the melting point; magnesium, zinc
is HCP up to melting point. So, these structural transitions are more fundamental issues. You
need to look at what the actual issues related to whether this material is ferromagnetic or not.
That again influences to certain extent.
Student: Ni, Fe, Co are ferromagnetic
Ferromagnetic, remaining are not. But in all these three, if you take nickel, it is FCC up to
melting point; iron has a BCC to FCC to again BCC, cobalt you take it is HCP to FCC. So,
all the three though the three of them are ferromagnetic they do not have the same type of
structures. That means, so it is again because their electronic configurations are different.
Student: One one electron
It could be, I mean 1 1 electron can make copper and zinc, you see they are entirely
different; it also 1 1 electron, they are entirely different. So, it is more fundamental issue.
We need to understand. I do not think physicists also have understood this that well, we
still do not know. I at least do not know still.
Student: But with shear stress, wont there be any change in lattice positions?
No, lattice positions do not because that is what we say shear. So, shear basically means
the one lattice structure changes to another lattice structure by a shear that is it and once
that happens that wherever lattice positions are there, this is what we call it as bainite
distortion.
We will stop now.
654
Prof Dr B S Murty
Department of Metallurgical And Material Engineering
Module-01
Lecture-27
Martensitic Transformations, Order-Disorder Transformation
Last class we were talking about martensitic transformation. We talked about

Martensitic = Shear + Compression along c and expansion along a and b axis.
This total is the actual martensitic transformation, so austenite to martensite
transformation if you say it is a combination of the two steps. Depending on to what
extend this second process can occur is going to decide what the c/a ratio is. Just after
shear, c/a is 1.414 we have seen, and in principle it would like to come to 1.0, which is a
BCC structure which is a stable structure at room temperature. So, it would like to come
from 1.414 to 1 depending on the carbon content present in the alloy, and depending on
what is the extent of compression that can take place along the C axis. If there is no
carbon it would in principle be1. So this is what is the whole structural aspects. Let us
now look at the thermodynamic aspects where is the driving force is coming from.
655
We are already seen that at room temperature, if I take an austenite and then quench it
the stable structure is BCC, am I right? at room temperature the stable structure is BCC.
So the austenite wants to transform to this BCC structure that means the FCC structure
wants to transform to this BCC structure, that is the whole process, but in principle under
normal conditions FCC can transform to BCC structure by a polymorphic transition
involving what kind of a diffusion?
Student: Short range diffusion
Short range diffusion. In principle FCC to BCC transformation can happen by a short
range diffusion, but because you are at a temperature where short range diffusion is not
able to happen then you see a shear taking place. So, if I draw the free energy versus
temperature diagram you would see a transformation something like this (see the Fig.
above), at higher temperature it is the γ that is stable, at lower temperature it is the
martensite that is stable. Martensite in principle it could be a BCC martensite or a BCT
martensite depending on the type of situation. So, there is a temperature which we call it
as the T0 for this γ to martensite, but actually the transformation never occurs at that
temperature because the transformation needs some shear; you need a certain shear to
occur inside the material, so that you would see the γ transforms to martensite. If there is
no shear γ would not transform to martensite, and this shear for it to occur you need to
provide certain driving force for the shear to occur.
As a result actually transformation occurs at a temperature below T0 at some

temperature which is what we usually call it as MS temperature which is actually below
656
the T0. This, T0- MS, is the extent of under cooling that you need from the T0 to be able to
have the transformation, and this MS temperature we all know is dependent on a number
of factors one is carbon content, and the other alloying elements present and also the
grain size for example, grain size can also control the MS . So, there are many factors
which will shift this temperature to a lower temperature and avoid transformation at the
T0. When do you have actually the transformation occurring very close to T0, when the
extent of shear that is required for the transformation is very small.
In any particular system where the extent of shear that is required, the shear stresses that
need to be provided for the shear to occur if that is very small then you can say that in
principle MS comes close to T0, otherwise you would always see some temperature
which is below this T0 where the transformation occurs. This is where we need to
understand the influence of stresses on martensitic transformation. We always say there
is something called stress induced martensite, you have heard of it?
Student: Hadfield steels
Hadfield steels is one of the wonderful examples of it. You take an austenite and quench
it and get the austenitic structure and deform it and you would see during this
deformation this austenite transforms to martensite. But the reason why you see austenite
transforming to martensite during deformation which is what we call it as deformation
induced transform martensite and we even give a temperature called MD temperature for
that temperature where the martensitic transformation occurs during deformation, and
usually MD > MS. The reason is that whenever you deform a metal there is always shear
stresses which come into the picture, whether you do tensile deformation, or you do
extrusion; you do wire drawing most of the processes excepting one deformation process
which is called isostatic compression. For example if you have heard of HIPping and
CIPping, if I do HIPping for example, I have a isostatic hot compression or CIPping cold
isostatic compression. So, when you do isostatic compressing; that means you are not
inducing any shear stresses inside the material, the moment you do not induce any shear
stresses in the material the M S would actually go down. The moment, you have always
some shear stresses which get induced into the material during deformation then you
would see the MS will be higher.That basically means that you would see that the γ
transforms at a higher temperature to martensite rather than a lower temperature, the
reason is whenever I apply stresses inside the material; shear stresses,
657
the γ which is the parent phase will have a higher free energy when compared to the
original γ which is un deformed.
A deformed γ will always have a higher free energy, and if I use that Let's say, I use a
different free energy curve for that now, this is call it as a deformed γ (drawn in pink
color in Fig. above), γ which is deformed will have a higher free energy than the normal
γ, and obviously you will see now the intersection has happened at a higher temperature,
that means the T0 has shifted to a higher temperature, the austenite would transform to
martensite at a higher temperature. If the T0 is shifted to a higher temperature then
obviously the MS also will shift to a higher temperature.
So, you would see that the austenite transforms to martensite at a higher temperature
which is what we call it as M D temperature when you have deformed. But now how do I
understand about the other aspect, if I subject austenite to a compression; isostatic
compression, then the situation is different. Then the situation is, if you look at any time
any material, if you subject it to a compression; that means you are trying to reduce the
volume, you are compressing it. Whenever you are compressing it the phase that is
always stable is a phase that occupies less volume. This is always the case for example,
if you want a graphite to transform to a diamond you need to apply a compression,
compression and also a higher temperature.
So, a high temperature and compression would make a graphite form, and the higher the
pressure that you apply the lower the temperature at which the diamond would form
from graphite. So, the reason is the diamond is a much more close pack structure than
graphite. So whenever, this is true even for any liquid to solid transformations, if you
take a liquid to solid transformation, if you take a liquid which is trying to transform to a
solid and you apply a compression on the liquid, some kind of a isostatic compression,
you would see the phase that gets stabilized would be the phase which is always which
can occupying lower volume; usually most of the cases solid occupies the lower volume,
so solid become stable. Solid become stable means what should happen to the melting
point,
Student: Increase
it should increase.
658
So whenever you pressurize, it is basically clausius clapeyron, so if you go back to the
clausius clapeyron, whenever I apply a vacuum you would see that because the pressure
is reduced, at a reduced pressure you would see the melting point occurs at a lower
temperature. Whenever I apply a pressure you will see the melting point will be higher,
the solid melts at a higher temperature or the liquid does not solidify at the normal
melting point but it will solidify at a higher temperature. So the melting point would go
up when you apply the pressure excepting in the cases where the solid occupies more
volume then the liquid. So when the solid that is examples of like water to ice
transformation, or say any other transformation,
Student: Sir, but in terms of free energy effect, whether it is tension or compression
correct this is an important point. So, now if I look at in terms of stresses, any stress in
principle should raise the free energy of a system, so that means the stresses would
definitely raise the free energy of γ. But because of the fact that the martensite is a more
open structure than that of the γ, the relative free energy changes will be such that the γ
free energy because of this pressure would actually go down with respect to the
martensite when compared to what you have seen if there are no hydrostatic pressures.
So, there is definitely going to be some raise in the free energy of γ because of the
stresses that are there. But at the same time the second factor which counteracts that is
the fact that you are putting a hydrostatic pressure, and then you would see if you simply
look at the clausius clapeyron equation which says that
dT/dP = ΔV/ΔS
So if you look at that, you would see that as the pressure increases; if both are positive, T
should increase. In our case if you look at austenite to martensite transformation
Vγ < VM
ΔVγ-M = VM - Vγ = +ve
And if you look at the ΔS part of it,
Sγ > SM because it is a high temperature phase when compare to martensite. So,
ΔSγ-M = SM - Sγ = -ve
So, one is positive another is negative, so ΔV/ΔS is going to be negative, the ratio. So,
higher the pressure the temperature should be lower, and if that is the case you would see
that affect will come into the picture, and as a result you would see this would have a
overall impact of reducing the free energy of the γ in such a way that the T 0 curve shifts
to a lower temperature.
659
So, two types of pressures, two types of stresses can lead to two different effects, one
which is a compressive stress which is an isostatic compression can lead to a lower M S if
you apply, I mean nobody does this for an austenite, nobody has done, I mean excepting
in case of powder metallurgy people have done, if you take powders of let us say
austenite and then do a HIPping of it; isostatic pressing is usually done only in powder
metallurgy otherwise you do not use it in case of normal steels which are produced by
casting route and then deformed to give whatever shape it is. You go through all other
processes such as rolling, forging, and other deformations, but you rarely use a kind of
isostatic pressing, isostatic pressing is more in terms of powder metallurgy. So, in case
you take iron powder, iron or steel powders in the γ state, and then do a hipping, and
then you quench it and you compare it with another steel which is produced by a normal
case you would see that the martensite transformation occurs at a much lower
temperature when compare to a normal steel.
So, only in such case you can actually confirm that and under normal steels we do not
need to use HIPping at all. So, it is this is of more concern for powder metallurgy people,
ok. So, in case you are dealing with powders and then doing HIPping or CIPping, and
then trying to see whether I get matensitic transformation, you may say why should I
bother about what temperature it should, it will transform. I need to bother about it
because I need to know to what temperature I should bring it so that if I want to get
hundred percent martensite, so if MS itself is low, MF will also be low. So, for you to be
able to get hundred percent martensite you need to go to much lower temperature,
sometimes may be even a sub-0, so that is an aspect which you need to understand. So,
the pressure brings in this effect and that changes the free energy, and as a result you
would see the whole transformation changes. So this is another aspect which, regularly
people do not talk about it, people always say deformation will lead to an increase in the
martensitic temperature, that deformation is only a deformation which induces shear. If
there is no shear it is induced inside the material, and it is only a deformation which is
just a compressive deformation, you would not be able to see that. I think with that we
will close martensite, I think we have more or less talked everything about martensite
including the crystallography aspects of it, I will not go too detailed into the kinetics of
it.
Student: Non ferrous also, is it same?
660
Yes thermodynamics will be similar excepting that you would see in some cases the
shape memory effects particularly in the non-ferrous nickel titanium, and those are
related to whether you have the martensite is a thermo elastic martensite, what is the
extent
661
of the transformation strains that are involved. And also for example, I do not know
whether you have ever thought about it, why steel martensite does not have a shape
memory effect, you do not see shape memory affect in steel though there is a martensitic
transformation,
Student: The elastic component is less
Less, anything more. One of the primary criterions for a shape memory effect is that the
transformation should be reversible austenite to martensite, and martensite to austenite
should be reversible.
For example in a steel, typical steel when I start heating the martensite what happens?
Student: Tempering
It starts tempering and then because of the presence of carbon inside, the carbon starts
coming out in the form of carbides and you do not see martensite directly transforming
back to an austenite, there are a number of intermediate phases that come into picture,
isn’t it. So, you unless you go back to the A3 temperature you would not actually get the
austenite transformation. Where as in the non-ferrous nickel titanium and a number of
other cases like copper base alloys these are all reversible transformations, there is no
other phase involve and only when you have a reversible transformation you will be able
to see the shape memory effect, so that is another thing. Of course, as he mentioned
about it the extent of strain is also important, you need to have a reasonably good amount
of strain which can be able to be recovered during either heating or during deformation,
so we will not go into the details of it that is more of a kinetics and not really
thermodynamics may be in some other consideration we will see that. Now let us look at
the other aspects, the other type of transformation which we have not covered so far is
the order disorder.
662
Before we actually go into the thermodynamics let us look at how do we recognize this order
disorder transition in a material. Whenever we say order and disorder, I think all of you are
aware when we talk about order there are two types of order in a material, we talk of
structural order and we talk of chemical order. So, whenever I say this is an ordered material
we are usually using the term chemical order there, when I say structural order, then a
structural order it can be a crystal, it can be a glass or an amorphous, or it can be even a third
one which is a quasi crystal. These are the three types of structural orders and the
differentiation is, a crystal is one where you have both rotational symmetry and translational
symmetry, and amorphous is where you do not have both, and quasicrystal is where you
have one of them, the rotational symmetry exits but the translational symmetry does not
exist, that is how we talk about it.
When I say chemical order first and foremost is it is structurally ordered, and it can be
chemically disordered or ordered, there is no possibility of looking at a structurally
disordered material and look for a chemical order there. And similarly, again when I say
chemical order we are again talking about long range chemical order, we are not talking
about short range chemical order here. Short range chemical order and short range
structural order will always be there in every material until you go beyond the boiling
point. When you are above the melting point you still have a short range order, you still
have a short range structural order, if that is a chemically also order there may be a short
range chemical order also inside the material. But if you go beyond
663
the boiling point there you have individual molecules and individual atoms, so you do
not have a chemical order or neither a structural order. So, whenever we are talking
about order; this material is ordered, we always talk about long range order, we do not
talk about short range order that is people have to keep it in mind. And when I am saying
the long range order, so it is not possible to have a structural disorder and a long range
order, these two will not go together. You cannot have an amorphous alloy which is
chemically ordered because it is already structurally disordered, so because it is
structurally already disordered there cannot be a chemical order inside, chemical order
we always look inside the unit cell. The moment you talk about chemical order, that
means how atoms are arranged inside the unit cell is what represents the chemical order.
If there is no unit cell there is no question of talking about how the atoms are arranged
inside the unit cell, so as a result, because amorphous materials do not have a unit cell at
all, so we will not talk about structural chemical order there. So, chemical order basically
implies a structural order, so it is a subset of structurally ordered materials; that means it
is a subset of crystals. So, chemical order or a chemical disorder are two sets inside a
structurally ordered material. So if I consider, first of all that it is a crystalline that it has
certain crystal structure then I talk about whether it is ordered or disordered, and this
question of ordered or disordered comes into picture again for the structurally ordered
materials only when you have more than one element inside the material. If you have
pure metal we do not talk of a chemical order; obviously, so we have only one element
means it is simply structurally ordered, that is it. You do not talk about chemical order;
chemical order comes into picture only when we talk about alloys, more than one
element.
When you have more than one element present then only we talk of chemical order, then
we talk about how each element is arranged inside the unit cell if they are the randomly
arranged. If there are no preferred sites for any one of the element we call it as it is
chemically disordered, and such material are usually called as solid solutions or
disordered compounds.
664
In fact, it is very difficult to differentiate between a solid solution and a disordered

compound. Both of them will have two elements if it is a binary alloy, two elements
which are randomly present inside the crystal structure, whatever may be the crystal
structure. So, a disordered compound is usually referred to as a solid solution. But
excepting that, for example if I am talking of a nickel aluminium alloys, when I take
compound such as Ni3Al. Ni3Al is a FCC based compound inter metallic with what we
call it as L12 structure. So, here you have three atoms of nickel occupying the face
centered positions, and the aluminium occupies the body corners, and now this becomes
let us say disordered; that means inside the unit cell if you have hundred unit cells, in
hundred unit cells you would have 75 atoms of nickel, and you know 25 atoms of, in
fact, we should not take like that, in a hundred unit cells actually you have much more
because every unit cell effectively has 4 lattice points, ok
So, if I take let us say about 25 unit cells with about 100 atoms, in these 25 unit cells the
nickel atoms will sit wherever they want to sit, the 75 nickel atoms would sit wherever
they want to sit, and the similarly the remaining 25 atoms would sit wherever they want
to sit. There is no preferred place for that, so that means out of the three face centered
positions let us call them as 1/2 1/2 0, 0 1/2 1/2, and 1/2 0 1/2. If we consider these three
positions, one of them may be occupied by the aluminum, another may occupied by
nickel. So there is no preferred, you know site for either nickel or aluminum, but the
665
structure is still FCC. The structure is still FCC excepting that aluminum and nickel are
randomly sitting wherever they want and that is what a disordered compound is. And if
you want to call it as solid solution also there is no problem in calling it as the solid
solution excepting that there can be a solid solution of aluminum in nickel with about 5%
of aluminum or 10% of aluminum. You cannot call it as a disordered compound because
it is not a compound at all, it is a solid solution of aluminum in nickel excepting that
Ni3Al if you want to call it as a solid solution this is the one which will have 75-25
exactly.
So, the composition would correspond to inter metallic composition, but the structure
would correspond to a solid solution structure, clear. So, the composition will be exactly
a stoichiometric composition like 75-25 , but the structure will be a disordered structure;
disordered chemically and ordered structurally, so that would the difference. So, there
can be a innumerous solid solutions with different amounts of solute content, but the
moment I say disordered compound the disordered compound will have a stoichiometry
corresponding to that particular compound, so that would be the difference.
666
Now, if I want to look at how do I find out this extent of disorder when I take an Ni 3Al
and make it disordered in principle, if I take a Ni 3Al and start heating and talk in terms of
what is called order parameter, they refer it to as L which is a long range order
parameter. Order parameter can change in two different ways, it can become like this
(curve 1, above) or it can become like this (curve 2, above). What do you mean by order
parameter being one, that means the probability of finding a aluminum atom at the 0 0 0
position; what are the positions here in FCC, 0 0 0, 1/2 1/2 0, 0 1/2 1/2, and 1/2 0 1/2.
These are the four lattice points, and 0 0 0 position is suppose to be aluminum in an
ordered structure, other three are suppose to be the nickel positions. If the probability of
finding aluminum at 0 0 0 is 0.25 then we can say it is perfectly ordered, that means out
of the twenty five percent of aluminum atoms inside the alloy all the twenty five percent
of atoms are sitting at this, ok. So, that means if you look at probability of, if you call the
corners as A site and you call face centers as B site, if all aluminum atoms are occupying
A site then I call it as a perfectly ordered. If some of the nickel atoms are also occupied
at the A site then it is not completely ordered, it is partially ordered and the disorder can
be when the probability of finding aluminum at the A site is one then we call it as L =1,
that means, all aluminum atoms are there at this. If you see partially aluminum is sitting
there, and when you say that the probability of finding aluminum at this A site is 0.25; if
I call it as A site, if the probability p = 0.25 then we will say L = 0. That means it is as if
the total amount of aluminum is
667
25%, and the 25% is sitting everywhere, and 25% of the aluminum atoms are sitting at
corner, that basically means it is as if there is no preferred site for aluminum. So, if you
find the probability of aluminum atom sitting at the A site as point two five then we say
long range order is 0. So, this is how we can find out provided you know where the
atoms are sitting. But It is not easy to find out where the atoms are sitting, but there is an
indirect way we can find out the long range order by looking at the x r d patterns. If you
take a perfectly ordered structure Ni3Al let us say which is perfectly ordered, which
people have made let us say by some process and do an x r d of it, and then find out
where; what is a basic difference between ordered structure and disordered structure if
you do an x r d,
Student: Extra spots
you get extra spots or extra peaks which we call them as super lattice peaks; super lattice
reflections.
In case of FCC what are the fundamental reflections, ’

Student: 3, 4 , 7
I am asking what are the reflections you are telling me what is h2 + k2 + l2,
Student: 1 1 1
1 1 1, 2 0 0, 2 2 0, 3 1 1. What is the rule it follows, all should be unmixed, ok so that is
how as she says this is h2 + k2 + l2 is being 3, 4, 8, 11, and 12, 12 is 2 2 2. These are what
are called the fundamental reflections; that means allowed reflections, and what are the
668
super lattice reflections, you can have 1 0 0, you can have 1 1 0, you can have all others
whatever is not possible. For example, you can have 2 1 1, 2 1 0,
669
2 1 1, all of them which do not exist in fundamental reflections list. Each of them will
have certain intensity depending on the structure factor associated with that reflection;
every reflection would have a certain structure factor which we can calculate. And in
principle if one has prepared a perfectly ordered thing in JCPDS files, if you go you can
take any one of the superlattice peaks and find out what is the intensity, I of it. The long
range order parameter, L is defined as:
I 100
I=
√ ( )
I 111
I 100
( )
I 111
sample
standard
The reason why we take square root is, the intensity is always square of structure factor,
so square root of I is basically I am getting the structure factor with respect to my sample
and the standard, standard is something which is a perfectly ordered.
Student: Only why those peaks?
Anything you can choose any peak, any fundamental peak and any super lattice peak, we
choose usually the most intense super lattice fundamental peak, in FCC the most intense
is always 1 1 1.
For example in BCC the most intense has to be 1 1 0. As you go to the higher and higher
θ the intensity goes down, if you have done XRD you will know this. And in addition
only in certain cases you would see for example, 2 0 0 being more intense than 1 1 1,
when do you see that when there is a texture. If there is no texture inside the material
then in principle 2 0 0 and all others will be smaller than 1 1 1. So, this is how will, so
we choose the most intense here and among these whatever is the most intense you can
choose that and take the ratio of. So, you can choose any one of them, in principle there
are cases where you do not even choose 1 1 1, when do you not choose 1 1 1 let us say,
think of a hypothetical situation where I would not choose the 1 1 1 peak as the
fundamental peak for finding out the long range order parameter. It is FCC, the phase is
FCC, she says if it is not FCC that is a trivial point. It is a still FCC, but I am not
choosing 1 1 1 why, when there is a second phase inside the material. If you have some
other phase whose some some peak of it overlaps with 1 1 1 then obviously the intensity
that you are looking at that peak is a combination of the intensity coming from the 1 1 1
670
peak of your ordered phase and some other peak of some other phase unless you do what
is called the deconvolution, and if they overlap is very good then you
671
cannot do deconvolution. If there is a small shift something like this (see the image
above) then you can do deconvolution, if the two peaks are not actually shifted at all,
they are exactly falling one above other you cannot use that, then you should go to some
other peak which does not overlap with any peak of the second phase that can be any
peak, so you have to start looking. As long as it is a single phase we choose the 1 1 1, as
long as we have more than one phase we do not choose that, so anyway. So, this is all
about how to recognize this is ordered or disordered, and how to calculate what is called
the long range order parameter.
Now, let us look at the thermodynamics of it. The moment we say it is ordered or
disordered we can talk about the free energy of the ordered phase, free energy of the
disordered phase.
GO = HO - TSO
GD = HD - TSD
The moment I say this when I look at the ordered phase, the entropy is always a
combination of two entropies; thermal and the configurational. The thermal entropy will
always exist in any material at any given temperature excepting at the absolute 0, but the
configurational entropy for an ordered structure is 0. In fact, if it is a perfectly ordered
structure, if L = 1, then the S = 0 because the number of ways of arranging atoms is one,
ΔSconfig = kln w, where w is the number of ways of arranging atoms. If the number of
672
ways of arranging atom is one then ln w is 0, so, the configurational entropy is going to
be 0.
So, in principle this particular term, TSO, is going to be much smaller when compared to
this term TSD . So, as a result if I now plot G versus T, and also look at H O and HD.
Whenever we talk of ordered structure A and B; when do you get an ordered structure?
Student: When Ω is negative
when Ω is negative. Whenever Ω is negative, whenever there is a strong attraction
between A and B you get the ordered structure. So, that means H O < HD. Disordered
structure would not have the same bonding, so, as a result, you would see the ordered
structure there is more attraction between A and B, H O is more negative than HD. And
similarly if I look at entropy, SO < SD, once I know these two I can actually talk about
how the G changes with T. You would see two values of H in the G-T curve (drawn in
above image), here HO will be lower, HD will be higher,. And if I look at how G changes
as a function of temperature, G decreases with increase in temperature or increases with
increase in temperature?
Student: Decreases
Decreases.
So, now the question is the rate at which it decreases is going to be different for both of
them. One of them which is having a higher entropy will have a steeper slope. At any
given temperature if I look at the entropy, S O < SD, you would see that the disordered
structure should have a steeper slope when compared to the ordered structure. If that is
the case you would see at some temperature these two will intersect, and that is what we
call it a TC, what we call it as a critical temperature T C beyond which you would see the
disordered structure will be more stable than the ordered structure, and you would see an
ordered structure transforms to a disordered structure. It is almost like a liquid solid
transformation that we have seen earlier; exactly similar to that you would see this
happening, ok.
So, only thing is then there are two ways in which it can decrease, that is what we call it
as first order and second order transitions. It is one where the order parameter remains
almost one up to the TC and suddenly at TC it changes to 0, this is
673
something like a first order transition where you have sudden change in the phase
structure at a particular temperature. There is a discontinuity, so the solid remains a solid
up to a melting, suddenly at a melting it becomes liquid, it is not that the extent of
disorder in the solid keeps on increasing and at the melting point it becomes liquid, it is
not a continuous transformation it is a discontinuous transformation. Whereas there are
certain order disordered transitions which are continuous where the order parameter
continuously changes with TC and finally becomes 0 at TC. These are what are called
continuous transformations, it is more of mechanism of order disorder, let us not much
bother too much about it, but as far as the thermodynamics is concerned we can talk
about this. And at the transformation temperature I can equate GO = GD.
HO – TSO = HD - TSD
So, TC = (HO – HD)/(SO – SD)

And as a result this difference, H O – HD which is nothing but related to the Ω, so larger
the what is called the interaction parameter between A and B the more negative the
interaction parameter is, the higher will be the T C . So, the order disorder transition
temperature would be higher if this difference is higher. If I treat disordered structure as
that where the interaction parameter is 0 because in principle we can take it as an ideal
solution it is disordered structure A and B, there is no preference they are all sitting
everywhere. So, I can call a disorder structure almost like a solution which always will
have the interaction parameter as close to 0, and that is what we say that isomorphous
kind of phase diagrams you get when a interaction parameter is 0 where liquid gives you
solid which is a disordered or chemically disordered; structurally ordered of course but it
is chemically disordered. So, HD ~ 0, whereas HO is decided by the Ω. And H O – HD is
negative and
SO – SD so finally you will get a positive number for TC.
674
And the value of the TC is again going to be decided by HO – HD to a large extent than SO
– SD because S of the disordered does not depend on anything, it does not depend on
what is A and B it is simply configurational entropy. If I am taking 50-50 lets say copper
zinc, it has a certain value, if I am taking Ni 3Al let say 75-25, it has certain value of
configuration entropy. It does not depend on whether it is nickel or aluminum or copper
or zinc, configuration entropy does not depend on that.
So, you can see it is this Ω which decides it, that is why in certain compounds the
transformation occurs at a lower temperature, in certain compounds transformation does
not occur until you reach the melting point. In copper zinc 450 ºC is where it transforms,
whereas in NiAl it transforms, the order disorder transition occurs at the melting point.
When the solid is actually melting then the disorder takes place, chemical disorder, so
structural disorder and chemical disorder occur together in such a material, so this kind
of things need to be understood, so everything is related to Ω.
We will stop now.
675
Lecture #28
Miscibility gap in phase diagrams
We just looked at first quiz answer scripts. I am generally happy because people did
reasonably well excepting that two questions which I thought people could have done
better. First question was about an expression to find out the Miscibility gap, the
temperature at which the Miscibility gap closes. In fact, only three who have done it
correctly. One is Vijayalaxmi, Prerana Raghukiran. Very happy to see these three doing
it right. Rest of them basically tried to equate.
What we are talking about is Miscibility gap. I want to do this today. So we are looking
at this phase diagram (top Fig. in above image) and we are interested to find out what is
the temperature at which the Miscibility gap closes. That means, the temperature above
which you have single phase α, below which you have α 1 + α2. If you look at the what is
called the free energy composition diagram, at any temperature within the Miscibility
gap, we know that the free energy composition diagram looks like this (bottom Fig. in
above image) and we have, we can talk of a common tangent and these two points (P and
Q in free energy curve) are what we call them as binodals.
676
If this is done at room temperature, P and Q basically corresponds to points P’ and Q’ on
phase diagram. And we said at temperature above miscibility gap the same free energy
curve would look differently.
And If you draw it once again, the free energy curve looks something like this (Fig. 1
above). Am I right? So, that means, there a single phase is stable whereas, in the former
case a single phase is not stable. It splits into two phases and that brings us to two more
additional points which we discussed, which we call them as spinodal points (P1 and Q1,
Fig. 2 above) and in principle, whatever we are discussing here, the temperature for
Miscibility gap closes is also the temperature at which the spinodal closes. Obviously
because if I draw a spinodal curve, let us say chemical spinodal. Mind you, I am not
talking about coherent spinodal. If I draw a chemical spinodal, chemical spinodal also
closes at the same temperature (Fig. 1 in above image).
So, in principle whatever temperature that we are talking about, the temperature at which
of Miscibility gap closure, where the Miscibility gap ends is same as where the spinodal
ends. What we call is the critical temperature for the spinodal decomposition at
temperature at below which you have the spinodal occurring and what is the maximum
temperature where the spinodal can occur? In fact, for every alloy, there can be a
different spinodal temperature.
Similarly, for every alloy composition, there is a different temperature at which the
Miscibility gap is no more there, but if you look at where is the maximum. The
677
maximum usually comes at 50-50, provided we assume that the system is a regular
solution model. That is what this question also clearly says assume that the system is
following a regular solution model. The moment you assume that it is following a
regular solution model, then we know the ΔHmix = ΩXAXB and this ΩXAXB is a
symmetric function. If I plot ΩX AXB , if the Ω is positive, then ΔH mix will like this (Fig. 3
above), perfect symmetric and it would go to maximum at 50-50.
So, and Ω is a single Ω for the whole composition range, and that if you give a particular
value, any value to it, then you can see that the maximum value of ΔH mix is at 50-50. For
example, if I take a simple case like I take X A = 0.2 and XB = 0.8 . Then, XAXB = 0.16.
If I take XA = 0.4 and XB = 0.6, then, XAXB = 0.24. Only when I take XA = 0.5 and XB
= 0.5, you would say it is XAXB = 0.25. Only then this is the highest.
So, that is why the maximum value or the minimum value of ΔH mix depending on
whether Ω is positive or negative is only at 50-50. At 50-50, the ΔH mix will either go to
maximum or go to minimum, provided you will have Ω positive or negative. When does
it not go to a maximum or minimum? When Ω = 0. As simple as that. When you have an
ideal solution, then when Ω = 0. Then, obviously, ΔHmix = 0.
So, that is a very remote situation. We are not considering that. So, we are looking at Ω
being positive. First and foremost is when do you get a Miscibility gap? When Ω is in
the solid is positive. This is very clear. I think this is very clear. Then, the next question
is, what is a temperature at which you get this maximum? If you want to look at that,
then obviously, we said one easy way to look at it which is what many of you, have
actually done, but at the same time, I said that correct way of doing that is another way
which many of you have not done.
678
So, I said the easy way is to see what is the free energy .
G= X A G A + X B G B +Ω X A X B + RT ( X A lnX A + X B lnX B ) (1)
This is the final formulation that we have. If that is the case, we said in principle it is a
combination of ΔHmix and TΔSmix which is going to decide. As a result, when TΔSmix
dominates the ΔHmix; that is when you would see that Miscibility gap closes. And, we
said that if you can assume , G A and GB as standard states and assume it to be 0. If we
assume GA = GB = 0, then we get,
Ω X A X B=−RT ( X A lnX A + X B lnX B ) (2)
And T can be determined. This is a very simple way of doing, but that is not the accurate
way of doing.
The accurate way of doing is when the binodals meet at a particular point the spinodals
also meet. So, you should find out that particular composition and temperature, where
the 2 spinodals are basically meeting. And What is a definition of a spinodal? Yea,
second derivative of the free energy is 0. So, basically take this expression (eq. (1)), find
out the second derivative and equate it to 0 and you would find that particular and again,
put one boundary condition there that I am assuming X A = XB because I know that this
would happen at 50-50. Of course, under situations where it is following a regular
solution model, because we have already assumed that it is a regular solution model.
Once we assume that and then, put a condition that it
679
is happening at a XA = XB = 0.5. Then, you would get an expression which is what
actually Tc = Ω/2R. If you do like eq. (2), you will not get T c = Ω/2R . Many people have
done this up to eq. (2) and then, put
−Ω X A X B
T= (3)
R ( X A lnX A + X B lnX B )
T equal to this divided by this and then finally, said Tc = Ω/2R
How do you get from (3)? You will not get. Just because you put XA = XB = 0.5, what
are you going to do with this ln(0.5)? So, ln 0.5 does not give you the value that you are
talking about. So, let us try to do that for your benefit today and then, see whether this
really happens and you would see very easily that it happens. Let us look at the first
derivative. I hope you people still remember your maths. So, first derivative for this.
Before you all do that, we should try to see there is one more boundary condition in this.
What is the condition that we know in case of a binary? XA + XB = 1.
So, I can write XA = 1 - XB .
Substituting this in eq. (1), we get
G=(1− X¿¿ B)G A + X B G B + Ω(1−X ¿¿ B) X B + RT ¿¿ ¿ (2)
Now you do first derivative, we do it with respect to B and you can do with respect to A
also and there wherever X B is there, you can put XB = 1 – XA
680
- It does not matter. Usually, we consider B as a solute, A as a solvent and that we are
doing with respect to XB.
So, taking the first derivative of eq. (2), we get
∂G
=−G A +G B +Ω(1−2 X ¿¿ B)+ RT ¿¿
∂ XB
Using product rule of differentiation, it can be simplified as:
∂G
=−G A +G B +Ω(1−2 X ¿¿ B)+ RT ¿¿ (3)
∂ XB
Now,
∂ ∂
ln (1− X¿¿ B)= ¿
∂ XB ∂ 1
∂(1−X ¿¿ B)ln(1−X ¿¿ B). (1−X ¿¿ B)= .−1¿ ¿¿
∂ XB 1−X B
Substituting this value in eq. (3), we get
∂G
=−G A +G B +Ω(1−2 X ¿¿ B)+ RT (−ln ( 1−X B ) −1+1+lnX B ) ¿
∂ XB
681
682
So, we get
∂G
=−G A +G B +Ω(1−2 X ¿¿ B)+ RT (−ln ( 1−X B ) +lnX B ) ¿
∂ XB
Once I know this is the first derivative, I can easily do the second derivative,
2
∂ G 1 1
2 =−2 Ω+ RT
∂ XB 1− X B
+
(
X B
)
2
∂ G
Now equating second derivative, 2 = 0
∂ XB
2 Ω=RT
( 1−X1 + X1 )
B B
1 1
2 Ω=RT
(( .
1− X B ) BX )
683
Thus,
2Ω X B . ( 1− X B )
T c=
R
Putting XB = 0.5
Ω
T c=
2R
What are the assumptions in this? Whole derivation,
Student: Regular solution
regular solution model and one more assumption.
Student: XA = XB
Yes, we are assuming that the spinodal or the Miscibility gap closes at the middle of the
phase diagram. It is an assumption. It need not, but in a number of cases, you see that,
provided your phase diagram is a simple isomorphous and there is nothing else on the
other sides. In number of cases, you have a Miscibility gap. If you go back and see some
phase diagrams, you have Miscibility gaps only in a certain region. Rest of the region,
you have possibly intermetallic compounds and things like that.
This is really a fascinating thing that you have an intermetallic compound in a phase
diagram, you still have a Miscibility gap in certain region. There are a number of cases. I
want to you to simply go back, if you are interested see Massalski for example, 3
volumes of phase diagram. Just go through. That basically tells that you have regular
solution model itself is wrong. Regular solution model basically assumes that the
684
interaction parameter is the same from one end of the phase diagram to the other end of
the phase diagram, which is actually not true. In fact, it is not true in all
685
the cases where you have intermetallic compounds. Whether you have Miscibility gap at
all or not, the fact that you have an intermetallic compound in a phase diagram indicates
that at a particular composition, there is a strong negative interaction between A and B.
Otherwise, you would not get intermetallic compound.
So, that means, at some other places, other two sides for example, you take iron carbon
Fe3C. At 6.67 suddenly you get Fe 3C . You do not have any other compound in the
whole of phase diagram up to 6.67. So, that means, that there is something special
happening at 6.67, where you have 3 atoms of iron and 1 atom of carbon. Suddenly, such
a configuration you see there is a strong attraction. This is true with other compounds
also. If I take gold, copper, C u, A u, you have 3 compounds we talk about. AuCu,
Au3Cu, Cu3Au. That means, at 3:1, 1:1 and 1:3. These 3 compositions, you have an
intermetallic suddenly, at no other compositions you have.
That means, at other compositions, the interaction between A and B is not as strong as at
these compositions. So, that basically tells that Ω is not really constant at all
compositions and in other cases also which I have told you earlier. For example, copper,
zinc. If you take an example, we see that zinc dissolves in copper to a large extent;
whereas copper does not dissolve in zinc less than 1%, whereas zinc solubility in copper
is almost 40%, 38% or so.
So, this basically means, the interaction between the copper and zinc, on the copper side
is different and on the zinc side is again different. That is the reason why people had to
go to sub-regular solution models and wherever inter-metallics are coming into picture,
people do not use, neither of these regular or sub-regular, we have what are called other
models called cluster models.
So, you assume that there are clusters inside the material and try to look at what is you
assume all kinds of clusters. 1:1, 1:2, 1:3, all kinds of possible clusters that I can think of
and then, try to calculate the free energy of each other clusters and see which clusters
have a lower free energy. And Those clusters will be the one’s which are stable. If all
clusters have the same equal free energy, that means, we say that particular system
would prefer to be an isomorphous system, because there is no special attraction for any
particular cluster. In a particular system, where let us say it follows a ideal solution
model let us say.
686
In such a case, if I calculate all the possible clusters, you can find out in numerous range
stoichiometries. You can consider 1:1 to 1:100, whatever you want to consider. So, all
possible combinations if you start calculating the free energies and if you find that all of
them have the same equal free energy, that means in principle, you do not have any
compounds. If there is a particular compound which gives you a much lower free energy,
that means that particular compound is the one that you will observe. Rest of them are all
equal free energy.
So, that is how people actually calculate using cluster variation model and then, see
whether you get compounds or not get compounds. So, that is how we do it. So, this is a
way to calculate and this is the same solution whether you are calculating the T c for the
spinodal because you know Tc for spinodal basically means the two spinodal points are
meeting at that particular temperature. So, that means, again it is the same. Clear? So, I
want you to remember this and when you want to do it next time, you should be clear
about it.
Student: Sir, Ω fundamentally represents what? Is it the elasticity mismatch?
Ω fundamentally represents what is the attraction between A and B. For example, it is

more related to the valency mismatches. For example, sodium and chlorine, one with one
electron less and another with one electronic extra. So, you have an ionic bond coming
into a picture. So, that means, because of this, a particular sodium wants to stabilize its
configuration by taking some other electron let us say. So, it tries to combine with an
element which has one excess element electron. It try to look for which are the one’s
which have excess electron and once you put the other element in combination with
sodium, you suddenly form a compound. H2O is another example of that. So, all
compounds basically are related to their electronic configurations.
At the same time, when you are talking about intermetallic compounds, in addition to
electronic configuration, there is also a size effect that come into picture. They may want
to form a compound, but when the second element sitting inside the lattice, there may be
so much of a strain that, that is going to counter act the chemical affinity between A and
B. So, as a result, you may see such a compound may not form, though electronically, it
would like to form. So, it is a combination of both the factors. We cannot just say it is a
size factors. In fact, there are number of what are called size factors
687
compound also. They have nothing to do with the electronic configuration. Just because
the size of the B element is such that it fits into the voids. For example, there are number
of A2B type of compounds. Most of the A2B type of compound are all to what we call
them as size factor compounds, where the second element fits into the lattice of the first
element somewhere, so that it forms a good compound.
So, there are different types. So, you have to consider each element carefully and then,
look at it and when you look at different elements for example, Mg2Si is one such
example is a size factor compound. You do not see that in every other things. Al2Cu is also
another which is to some extent a size factor type of a compound. So, it looks like a
aluminum and copper atomic sizes, where they are going. Of course, they are more of
substitutional, but their atomic sizes are such that aluminum is much bigger when
compared to copper.
So, you have to look at in fact, even the whole bulk metallic glasses also people are
talking about in terms of the size factor. They say that if I can take one element with a
bigger atom, another element with a smaller atom, then this smaller atom fits into the
bigger atom lattice and makes the whole packing very dense and the moment dense
packing is obtained then, the structure is more stable
So, whichever makes such a dense packing possible, you would see such a thing is
stable. For example, the best zirconium base bulk metallic glasses are ZrBe. The best
iron based bulk metallic glasses are FeB. One is a very small element, another is a bigger
688
element. Zirconium beryllium is another. PdNiP is the bulk metallic glasses with the
highest what is called critical diameter, almost 70 mm people have found palladium,
nickel, copper, phosphorous. Again, phosphorous which is very small, there it goes and
sits into the palladium lattice, and then, makes that stable.
There also there is a critical amount of the boron which when it goes into the lattice, the
structure gets stabilized and that is what people call it for example, iron-boron, if you
look at it, 20 of boron and 80 of iron is the one, where you actually see this is happening
and that is the reason why, if you look at iron boron phase diagram, you have a deep
eutectic at 80-20. Same thing palladium phosphorous, you go to and see a palladium
phosphorous phase diagram at 80-20. You have a deep eutectic. That means, for such a
configuration, the liquid will have a high density and such a liquid gets stabilized and
once it is stabilized, obviously once liquid is stabilized, glass is stabilized.
Student: But will that not lead to intermetallic compound?
It need not because the question is whether it would like to form a regular configuration,
or it would like to have a randomly sitting and then giving you high dense random
packing, we call it as dense random packing. It is dense, but it is a random packing.
So, such kind of structures are all there. There are so many liquids. That is why we say
bulk metallic glass that where viscosity is very high. Why is it high? Because it has a
dense random packing. So, you put those elements. That is why, not every 2 elements
just because for example, I take some multi-component system 5 elements. It does not
mean that every 5 elements if I choose, I will get a glass. For example, we are working
on what are called high entropy alloys. In high entropy alloys, we do not get a glass. We
still have 6 elements. We can even put 8 elements. It does not mean that it will give a
glass, so just because you have more elements.
So, that is where you have to look at whether they have a strong attraction first of all
because liquid gets stabilized only when the ΔH mix is negative. So, that is one you should
look at whether A and B have all these elements. When you put them what is the overall
ΔH. See for a binary, I can easily talk about one Ω. How do I talk about a Ω for a
quinary system or hexanary system? So, I have to consider individual interactions, multi-
component interactions. So, this is a little more involved calculation,
689
but one can do it if he is interested. Second is look at size factors again. We know that
you would get a glass whenever the size factor is greater than a certain number, so that
solid solution is not preferred. We know from Hume Rothery that when the size factor is
less than something, you get a solid solution.
So, all those conditions which would prefer a solid solution formation would lead you to
a high entropy alloy without being amorphous and all those conditions which do not
encourage, which are not conducive for a solid solution formation are the one’s which
would lead you to a glass, provided you have multi-component system. So, that is why,
it involves a lot of thermodynamics to understand. Why? For example, in a multi-
component system, in a 5-component system, how do I know which is the best
composition for me to get a glass? In a binary also for example, if I look at the phase
diagram, I can say oh at this composition, I have a eutectic. Because I have a eutectic, I
can say yes it would form a glass. In fact, there are many interesting cases I would show
you one.
If you have a phase diagram which looks like this (see the image above) Take a case like
this. It has 2 eutectics. Any example that you know? Mg 2Si very simple. If I take this
system, there is a Mg2Si at the middle. Let us say if I cool this very rapidly and Mg 2Si
being a compound, you need atoms to sit in specific positions and that is not time is not
allowed for such a thing to happen. So, what happens is Mg2Si cannot come out. Then,
690
what should come out? The liquid would remain as liquid and it would remain as a liquid
until you, if you look extrapolate, this liquidus curve (dashed lines) to a particular point,
where you have another eutectic which is what we call it as metastable eutectic.
So, in a phase diagram if there is a eutectic, a normal eutectic like iron-boron, I can say
yes that is a composition where you can get a glass, but if you see things like shown for
Mg2Si, if you start thinking that intermetallic is not forming what is, then going to form
that liquid at that particular composition is going to be under cooled continuously until
you reach this temperature. So, that means, in this alloy though the 2 eutectics are
shallow eutectics, still this particular phase diagram can give you a glass. Though I may
say that, sir if I do the T0, the T0 is not really very deep, it is shallow, but in spite of that,
you would see that you will get a glass because if the intermetallic is removed, then you
get a different metastable phase diagram. And in that metastable phase diagram, the
eutectic may be very deep.
So, this is how you can see the moment you see a deep eutectic, whether it is in the
stable phase diagram or in a metastable phase diagram. Then, there is a tendency for the
formation of a glass, but this is easier to see in a binary. How do I see it in a ternary.
How do I see it in a quaternary or a quinary or in a hexanary? So, that is where you need
to do certain calculations to find out which are the compositions, where the ΔH mix for
example, is highly negative. So, you try to calculate ΔH mix for the whole multi-
component system and find out where it is negative and also find out which is the
composition where you have the size factor being highly positive.
Where do you have the highest size factor in a normal binary system? Where do you usually
get a highest size factor? It is usually at 50-50. Whenever I put B into A the highest, misfit
comes into at 50-50, provided both are substitutional. That is where you will have the
highest strain. That is the reason why you would see Miscibility gap more or less will go up.
You need a higher temperature to overcome that misfit, to accommodate that misfit.
691
That is the reason why the Tc is lower at lower composition and higher at 50-50. So, at
50-50, you have more strains. So, you need higher temperature to accommodate those
strains, so that you get a single phase solve solution. So, similarly you would see that in
a binary, you can say it is at the highest. What about in a multi-component system? So,
you try to calculate where is this size factor going to be the highest and then, look at
these two combinations.
One is which is simply chemical interaction, another which is a topological interaction,

which is a size factor related and see a combination of these two. And People do all this
calculation. We ourselves have been doing it and if you are interested, it is a very
fascinating field and then do an experiment later. Identify that composition, do an
experiment later and then, see whether that works out correctly or not. So, this is how
people regularly do to identify that. Any other question? The other point I think which
most of you, the only point in the errors in the phase diagram which we discussed I think
in one of the classes, where the phase diagram which looked like this.
692
So, this error (marked in circle in above phase diagram) many of you could not by that
time possibly you did not know, but after the class I think we discussed. So, now, I think
by now you know this. That is it. I think rest of the things were easier and many of you
have answered it also. I think we will leave this. If there are any further questions, please
feel free.
693
Prof Dr B S Murty
Module #01
Lecture #29
Phase diagram calculations
So looking at this quiz questions, the first question derive an expression for order
disorder transition temperature. Many of you did it excepting that; some of you forgot
the next statement which says, which follows a regular solution model. So that part you
people did not bother, you simply equated ΔHmix = TΔSmix, without bothering that it is
following a regular solution model, so that means, I need to bring in ΩX AXB and then try
to get an expression in terms of the Ω. So that is the only mistake some of you have
done; I think it is not a major problem.
Simply only thing is, the moment you say regular solution model you know
ΔHmix = ΩXAXB, and TΔSmix = -RT(XAlnXA + XBlnXB). But, if you want to consider
ordered and disordered now carefully, so for the case of ordered and disordered
basically, what you are going to do is you are going to equate G of the ordered phase and
the G of
694
the disordered phase GO = GD. That is what you are basically going to do, the
temperature at which the both the free energies are going to be equal is what is of
importance to us. That is the temperature at which, you will see one of the phase starts
becoming more stable with respect to the other phase for example, if I simply plot G of
the ordered and disordered, so the ordered phase will be something like this, the
disordered phase will be something like this (see the graph in the above image);So, at the
intersection temperature G has to be equated. So if you want to get that G, basically G =
H - T S.
So, when we are considering that H; the ΔH part of it you would see the ΔH mix for the
ordered will be more negative, so ΔHOmix = ΩOXAXB . For the disordered, ΔHDmix =
ΩDXAXB and this ΩO will be more negative than the ΩD; that has to be remembered. And
when I am talking of ΔSmix term, ΔSmix = 0, if it is perfectly ordered and
ΔSmix = -R(XAlnXA + XBlnXB). This part if you take it then you can easily equate it; I
think this part some of you have not considered, so you have to consider when it comes
to ordered case, only the ΔHmix term will be there, in the ordered case in the G, in the
disordered case you will have both. And you may say where is thermal entropy is
coming into picture; the thermal entropy comes into picture when you are considering
the individual elements.
Because whenever I am writing G0 = XAGA + XBGB. We start from here, and thermal
entropy would come into GA and GB, so that has to be remembered. So, you will have
thermal entropy coming in case of the disordered, and along with the thermal entropy
you will also have a configurational entropy in the disordered whereas, in ordered case
you will not have a configurational entropy, you will only have the thermal entropy. So,
once you consider that then the expression becomes easier and expression becomes more
correct; some of you have not considered it I want you to remember this in future. Next
we said pressure effects; I think all of you have done this, this is very easy, so there is no
point in going in detail. But when you say pressure affects, so some of you have ignored
that that there is also a deformation which can induce pressure in to the sample which
would give you a more of a shear kind of stresses inside the material and that aspect we
are not considering here, you have considered only the isostatic pressure; all of you have
done that I think that is fine . Then the thing is what is the need for
695
the coherent precipitates at high under cooling. This Frankly speaking I am not really
satisfied with any of the answers. When you say what is the need you have to think about
why should coherent precipitates come at high under cooling, why they are not able to
come, why they are not coming at high temperature, why there is a necessity for the
coherent precipitates to come at a low temperature.
If you consider that, if you look at any alloy, if the precipitation is taking place at that
temperature T1, we already said that at T1, if you are just below the solvus line you can
easily have the nucleation of the second phase happening at the grain boundaries.
696
So grain boundary; if you consider an alloy at that particular condition (Temperature, T,

Refer slide time 05:37), so the alloy would be something like this, let us say α phase
(Fig. 1 in above image). This α the moment I bring it below solvus wants to precipitate
out the second phase, and the second phase can actually nucleate at grain boundaries and
triple junctions and then grow. And for the growth of that and for the nucleation of that
what you need is a diffusion of Cu atoms, because the second phase is a Al 2Cu phase
which has a higher Cu content when compared to the matrix.
So, Al2Cu is almost one third is Cu whereas the matrix has only 4% Cu if I am
considering Al- 4Cu alloy which is a traditional alloy. So, if I am considering such an
alloy then if this has to nucleate and grow you need Cu atoms to come from the matrix to
this precipitate, and you have seen this happening . If we plot % Cu what it called
concentration profile; concentration profile looks like this (Fig. 2 in above image).
Because if you remember the free energy composition diagram, we drew a free energy
composition diagram (Fig. 3 in above image) where you have the Al 2Cu (θ) phase, the α
phase and this is our alloy composition 4% Cu (C0), if you join a common tangent, this
composition (Cαθ) is where the composition of α in equilibrium with the θ phase or Al2Cu
phase. And the composition of α far away from the θ phase will be the starting 4% Cu,
and there you say, you see that there is a driving force, there is a gradient for the Cu to
move towards it (Fig. 2) and
697
that will drive the θ phase to grow. These are all possible at high temperature because
there is a thermally driven diffusion that can happen, because you can easily have atomic
diffusion taking place at high temperatures. The moment I undercool this to a
temperature like this (T2, Refer slide time 05:37), then the problem is more of a thermal
diffusion; thermal diffusion is more difficult. And when the thermal diffusion is more
difficult, nucleation and growth of the θ phase or any phase for that matter precipitate
phase to nucleate there, any phase when I say in this particular system it is θ, in some
other system some other phase may be. In this system the equilibrium phase is the θ
phase, so that θ phase to nucleate at the grain boundaries is more difficult, so the system
starts searching for ways of the second phase formation in a way which does not involve
nucleation and growth at the grain boundary.
That is when the system starts looking for what are the other ways where I can have the
super saturation being reduced because the α is super saturated. So, the super saturated α
is not stable because it has a free energy of this (point P, in Fig. 3 above), so this is a
much higher free energy than this (point Q, Fig. 3) which is a two phase mixture, so it
wants to precipitate out. Then it sees that such a precipitation at the grain boundary is
difficult, then it starts looking for what are the other ways of nucleation, then you see
that the only way of nucleation is nucleating within the grain. If the grain boundary is not
possible because for the atoms to diffuse from the center of the grains to the grain
boundary is very difficult, so nucleation and growth has to occur within the grain. So
when you say what is the need, this is the need. We should know that the nucleation at
the grain boundary and growth is difficult because of the diffusivities being low, then the
system has to look for ways of nucleating within the grain, and if it is within the grain
there is no possibility of θ nucleating within the grain. θ cannot nucleate within the grain
because θ has a structure which is entirely different from that of the matrix structure
which is a tetragonal structure where as the matrix is an f c c structure.
So, because the θ now cannot nucleate, now the system starts looking for what are the
other things which can be transient phases which can nucleate within the grain, so that is
how you see the coherent phases come into picture. Only when the phase; second phase
is coherent with the matrix it can nucleate within the grain, if it is not coherent with the
matrix it cannot nucleate within the grain. So, the need for coherent nucleation
698
comes into picture because of the difficulty in nucleation at the grain boundaries and the
necessity for nucleating within the grain, and the moment there is a necessity for
nucleation; where is the necessity coming from because it is super saturated it cannot
remain as α any more. Once you bring it to that temperature T2, this much of super
saturation is not allowed, so it wants to give away Cu atoms. If it wants to give away Cu
atoms then it has to nucleate within the grain, and if it has to nucleate within the grain
the only phases that can come out are either coherent or semi coherent and that too again
depends on as you people already have written some of you, depends on what
temperature you are, because now we know different solvus lines.
We have a θ solvus, a θ’ solvus, θ’’ solvus and a GP zone solvus let us say (Indicated in
the Fig. 1 above). So, depending on where is your alloy and the temperature so that
coordinate the temperature coordinate and the composition coordinate depending on
where you are, you will have that particular phase. Having a sufficient driving force for,
you see for a something to nucleate you also need a driving force. This alloy (C0) for
example, if I bring it and keep it here at temperature T1, it cannot precipitate G P zones
though they are the most coherent, because there is no driving force for the G P zones to
nucleate. So, you need to have a driving force for the phase to nucleate and among the
phases that have the driving force the system will choose the one, that one which has the
best coherency which needs the lowest energy for it to nucleate. When I say lowest
energy it involves both strain energy and the surface energy. So, something which can
699
nucleate and grow with the lowest combination of strain and surface energy is the one
that would nucleate among the phases that are possible. So, for example, if am at this
temperature T2 you can see that at T2, all these three are possible, θ is possible, θ’ is
possible, θ’’ is possible.
So among this;
Student: Sir, the grain, inside the grain, we have also got quenched in vacancies, doesn’t
it accommodate the lattice diffusion faster
yes, see basically if you look at a quenched in vacancies you have some of them, but to a
large extent many of these vacancies; because vacancy migration is faster, they actually
migrate to the grain boundaries very easily. In fact, that is the reason why you see what
is called precipitate free zones. In fact, if you look at how the vacancies are and plot if
you think there is a grain boundary (see Fig. 2 in above image). So, there is α and α, and
if you look at the vacancy concentration, plot vacancy concentration as a function of
distance you would see this kind of a plot (see Fig. 2 above) , the vacancy concentration
is very low at the grain boundary because it has already migrated to the grain boundary
and within the grain you have it. And if you say that there is a critical vacancy
concentration that is required for the nucleation to take place as you have mentioned
also, we need a critical concentration of vacancies so that the vacancies can act as a
nucleating sites and put a certain value of critical concentration (horizontal line in Fig.
2). Then you see that within this domain PQ, the vacancy concentration is lower than
that critical value, so in principle within that distance there can be no nucleation.
That is the reason why we get what is called p f z and in fact, depends on whether you
are quenching very fast or quenching slowly. Sometimes if you are quenching slowly
then what happens is this region PQ is going to be larger. So, you would see the same,
you may see something like this (P’Q’), so that means there is enough time for the
vacancies which are farther from the from the grain boundary also to move, migrate
towards the grain boundary and then get merged into the grain boundary or get
annihilated into the grain boundary, and that is the case then you would see the p f z will
be larger. That is the reason why there is a necessity to actually quench very fast if you
want to avoid p f z. If you do not quench very fast that is why most of the Al alloys they
actually quench into you know either into iced water or sub-zero temperature. If you
simply quench into normal water, that also depends on how much time you are spending
to take the sample out of the furnace. So, when you are actually doing experiments all
700
this become important how much time you actually you know spend on taking the
sample out and then before it reaches the actual quenching medium; these are
701
all important. And if the time is longer then you see a larger domain like P’Q’ and if that
is the case you get a larger p f z.
So, quenched in vacancies are also important, most of the time you can have nucleation
taking place there, but you should have sufficient quantity of that. There is a critical
quantity that is required. because if you have sufficient number of vacancies, sufficient
number of Cu atoms can come. See basically, Al 2Cu or G P zones has to nucleate let us
say, GP zones means is mostly a pure Cu. So, and a pure Cu, I mean almost pure Cu let
us say of a critical size has to form; r* is important. So, pure Cu of a critical size to form;
that means you need to have sufficient vacancies where Cu atoms can come and occupy
there. If you do not have then you would not see any nucleation on that then the only
other possibility of nucleation is it looks at certain crystallographic planes and the
nucleation occurs on those crystallographic planes. If it is nucleation on the vacancies
there is no crystallographic planes, it can be again random nucleation, is it clear; which
one, Cv is concentration of vacancies in principle, I think you all know concentration of
vacancies is an exponential function of temperature.
We can always write Cv = Nexp(-Q/RT) where N is number of atoms and Q is heat of

formation of vacancies, the amount of enthalpy that is required for the formation of
vacancies. So that is the
702
reason why when I am at a higher temperature I have certain vacancies, if I am at a
lower temperature I have certain lower number of vacancies, the lower the temperature
the lower the vacancies, and if I am quenching from a higher temperature to a lower
temperature, because I am not providing sufficient time for the difference of the
vacancies to get annihilated, then they get retained that is what we call them as quenched
in vacancies. Whenever you quench, there is a lack of sufficient time for these
dislocations which are excess dislocations that are available at high temperature which
cannot be retained at lower temperature, this difference has to basically go to the either
surfaces or to the grain boundaries. These are the only two possibilities and that if it does
not happen then you will have quenched in vacancies left out.
I think that is it. Explain why uniform precipitates cannot grow, half of you have done
this I think half of you could not do it. Basically, you all should know that the driving
force for particle coarsening comes only from the size difference.
All of you know that if we draw the free energy composition diagram (right Fig. in
above image), for α, and for a coarse particle, and for a fine particle let us call it as θ
phase. Then if I draw a common tangent, , you see the difference in the equilibrium
composition that means, if I have a fine particle, if I have coarse particle, the α which is
in front of the coarse particle and the
703
α which is in front of the fine particle have different composition and the fine particle
has a higher composition than the coarse particle. So that means there is, if I draw a
concentration profile you may get something like this (Left Fig. in above image), you
have a higher concentration near fine particle, you have a lower concentration near
coarse particle.
The θ phase will have the same concentration that does not change, so whether it is
smaller θ or a larger θ it will have a fixed value of its stoichiometry. So, the α phase will
only have the concentration gradient and the moment they have equal sizes, when I say
uniform precipitates that means they all have the same size, if they all have the same
size; do not bother how did they come we need not have to bother about that as long as
all precipitates have the same size, then this difference between equilibrium
compositions (in Free energy diagram) does not exist. If this difference does not exist
there is no gradient, if there is no gradient then they cannot grow, dissolution of one
precipitate growth of another precipitate is not possible. Only thing is we say that ideally
if all precipitation has occurred at the same time, and if it is a homogenous precipitation
not a heterogeneous precipitation if it is homogenous precipitation then in principle
everywhere nucleation have taken place and this is what we call it as a site saturation
case that there is no more nucleation taking place. At time t=0 all the nucleation has
taken place and all of them are trying to grow, when they are trying to grow you get into
this problem.
And But if you consider a case where you have different levels of heterogeneous
nucleating sites that are available with different hierarchy and some where the nucleation
takes place first, and somewhere else the nucleation takes little later. And during this
process the one which has already nucleated starts growing and as a result you would see
that these differences in the sizes exist, and once there are differences of sizes you will
definitely see such a Ostwald ripening taking place.
Student: Sir, in the precipitation of ferrite from austenite, at higher undercooling we get
Widmanstatten ferrite.
Correct, again the same situation as that of what we see the coherent precipitates,
Widmanstatten precipitates are all either coherent or semi coherent.
Student: But they still precipitate at the grain boundaries
704
No, different places not only always at the at grain boundary you will also see them
nucleating and in fact, even if they have nucleated at the grain boundary they grow in a
particular direction, not
705
all grains you will see them growing. So, those grains which are favorably oriented for
this planes to grow you will see them growing in a particular direction, for example, if
you have a grain boundary you will not see Widmanstatten structure growing in both the
directions, you will see growing only in one direction. So, this is basically because you
need to have a proper plane which can nucleate this phase and then growth can take
place there. Nucleation occurs even at stalking faults, nucleation of Widmanstatten can
also occur at the dislocations, but what is more important is how does it grow.
The growth occurs on certain planes where as a normal grain boundary allotriomorphs
then can grow in all directions that is the difference. Now there is one aspect which we
have not considered earlier which is about phase diagram calculations, we talked about
it, but not really gone further so I thought today we will take it up.
If you want to talk about phase diagram calculation let us take a simple phase diagram
such as an isomorphous (see phase diagram above) and then later go into the other type
of phase diagram. If you look at the l+α region, we said at each temperature you have
equilibrium between liquid and solid. So, if there is an equilibrium between liquid and
solid you can write an expression of free energy for the liquid, and an expression of free
energy for the solid which is can be written like, let us say if it is a regular solution,
G =X A G A + X B G B +Ω X A X B + RT ( X A lnX A + X B lnX B ) (1)
α
706
I can also express Gα in another way in terms of chemical potential. Why I have to bring
in chemical potential here is because our equilibrium definition is from the chemical
potential, all of you now that in a multi component system more than a single component
system whether it is binary or ternary or quaternary the equilibrium is always defined by
equal what is called chemical potential. When I say the chemical potential of a particular
element in a particular phase Ǵ αA ∨μαA , whatever number of phases that are present, if the
chemical potential of a particular element in different phases is equal then we say they
are in equilibrium, the reason is once the chemical potential is equal then there is no
driving force for the atomic movement. This we have seen that whenever Δμ is zero then
there cannot be movement of atoms from one side of the phase to the other side of the
phase, and if there is no movement of atoms we say they are in equilibrium, so that is the
reason why the chemical potential has to be equal. So, we can write any free energy
expression as
α
G =X A μ A + X B μB (2)
So, a free energy of any phase can be written as because what is μ, μ is nothing but rate
of change of free energy with composition, so as a result we can easily write it as above.
If that is the case then basically we can find out what is μA and μB from equation (1) as
follows:
Ω X A X B=Ω X A X B ( X A + X B ) (Since XA + XB = 1)
2 2
Ω X A X B=Ω X A X B + Ω X B X A
Therefore, eq. (1) can be re-written as:
α 2 2
G =X A G A + X B G B +Ω X A X B +Ω X B X A +RT ( X A lnX A + X B lnX B )
G =X A (G¿¿ A +Ω X B +RT lnX A )+ X B (G¿¿ B+ Ω X A + RT lnX B )¿¿ (3)
α 2 2
Comparing eq. (1) and (3), chemical potential for A and B in α can be written as:
α α α 2
μ A =G A +Ω A X B +RT lnX A
α
(4)
μB =G B +Ω B X A + RT lnX B(5)
α α α 2 α
Why I have to say α with X also because when you say X A and XB it is nothing but the
composition of a phase so that means, for example, if I am talking of this particular point
P, it is the composition of α. The composition of the α can be expressed either as X A or
as XB and we know that XAα + XBα = 1. Similarly, I can also write μA and μB for the liquid
as
707
l l l 2 l
μ A =G A +Ω A X B +RT lnX A (6)
μB =G B +Ω B X A + RT lnX B(7)
l l l 2 l
Once we write these, we can equate μαA =μlA and μαB =μlB
What is that that I want to actually find out, whenever you want to find out a phase
diagram when I say I am talking about calculation of phase diagram what is our interest,
Student: Equilibrium compositions
equilibrium compositions at different temperatures that means, at this temperature T1
(see phase diagram above) want to find out what is this in terms of X Aα or XBα one of the
two, similarly XAl or XBl . Once I know one of them the second one is known to me so
that means, these two compositions are to be found out at each temperature; that is the
interest of us when we say I want to calculate the phase diagram.
708
So, that means that at each temperature I have two unknowns, one is the composition of
the liquid another is the composition of the α. So, if that is the case in these eq. (4) to (7)
wherever there is a XB I can write it in terms of XA. So, basically there is only one
compositional variable for the α, and one compositional variable for the liquid. There are
no two compositional variables for a particular phase because XAα and XBα are connected
with each other. So, once I know that they are connected with each other I can replace
XAα = 1 - XBα or I can replace XBα = 1 – XAα , I can do it any everywhere. So, what I
need to find out if I want to equate (4) and (6), in that equation I have basically two
unknowns XAα and XAl ; these are the two unknowns. And what should be then the
known things for us, we should know what is the temperature at which we are doing it,
we should also know Ωl and Ωα. If I know Ωl and Ωα that is nothing but the interaction
parameter of the liquid and interaction parameter of the α, then only I can do this
calculation. If I assume that the system is let us say ideal, then these terms will not exist.
Then simply it will be
α α α l l l
μ A =G A + RT lnX A and μ A =G A + RT lnX A
So, Ω terms will not be there and what I need to know is only temperature and the pure
metal free energies at that temperature, that means what is G αA at temperature T, what is
G A at temperature T. How do I find out the pure metal values?

l
709
Student: H - T S
G = H - TS simply, and
T
H=H 298 + ∫ C p dT (8)

298
So, what I need to know is simply Cp, if I know the Cp I can find out the G αA , similarly the
TS. But there again the problem that comes into picture now for example, I want to
calculate G lA and G αA at this temperature T (see the phase diagram). At that temperature
what is stable?
Student: Only liquid
Only liquid is stable, α is not stable, and because α is not stable you will not know what
is the Cp of the α (or pure solid A) at that temperature T , so that problem still exists for
us. What will be the heat capacity of the α (pure solid A) at that temperature is still not
known, that is where you need to use what are called various thermodynamic models that
are available, which we talked about it one of the class, where we talk about what is the
change in the free energy between α and liquid.
If I know what is ΔG α-l as a function of temperature for a given metal A by some model;
there are various models. One such model is what is called Turnbulls model, where we know
ΔGα-l = ΔH.ΔT/Tf. And if I know this and if I want to find out the α free energy at this
temperature I simply have to calculate what is the liquid free energy at that temperature, and
to that I add this ΔGα-l value then I will find out what is the α free energy at that temperature;
which will be higher than that of the liquid. Similarly, at this temperature T, if I want to find
out what is the G αB and G lB , what is stable it is only solid α (on pure B side), liquid is not
stable. Again I use this ΔG α-l, so I calculate the free energy of the pure metal B in the α form
with the help of eq. (8), this is not a problem for me because C p of the solid (pure B) at that
temperature is available, and from that I calculate G αB , and for finding out G lB I use a model
like Turnbulls. So, I can always write for example,
ΔGα-l = Gl – Gα
Gl = ΔGα-l + Gα
. If I know Gl and I do not know Gα I simply do
Gα = Gl - ΔGα-l
So, one of them if you know and the ΔG if you know then
710
second one you can always find out at that temperature. So, what we need for any such
phase diagram calculation is pure metal free energies I need to know at that temperature,
and if at all the system is following not an ideal solution but if it is a following a regular
solution model I need to know the interaction parameters. And once I know the
interaction parameters I can happily plug those numbers, and then equate (4) and (6), and
equate (5) and (7), so you will get two equations and two unknowns.
Two equations, two unknowns can be solved provided rest of the things are known.
What are the rest of the things I need to know, I need to know pure metal values and the
Ωα and Ωl; then I can plug in those things and easily calculate and write a program that at
each temperature keep equating and calculate equilibrium compositions. And when you
solve the two equations, so that means at that temperature I am able to find out X Aα and
XAl . And now I change the temperature, ask it to again calculate because one more
important thing is when I change the temperature what changes,
Student: Cp
Cp changes; that means these values will change. But one thing that we always assume in
all these calculations is that Ω is not changing; again there is an assumption. In fact, if
you go to a basic CΑLPHAD calculations people have actually done it, but I mean by
problem here is what is Ω, Ω is nothing but the interaction between A and B let us say.
711
In fact, Ω comes from quasi chemical theory where we actually calculate

ΔHmix = ϵAB – ½( ϵAA + ϵBB)
where ϵAB is the energy of the AB bonds , ϵAA is the energy of the AA bonds, ϵBB is the
energy of BB bonds. If you have A and B elements, what are all the types of bonds that
are possible, there are three types of bonds possible AA bonds, BB bonds, and AB
bonds. If the AB bond energy is lower than AA bond energy and BB bond energy; this is
I am considering half, so basically you are normalizing. So, if
ϵAB < ½( ϵAA + ϵBB) , then the ΔHmix will be negative. Smaller means what, they are
strongly bonded, if ϵAB > ½( ϵAA + ϵBB) then you see ΔHmix will be positive.
If you look at this equation (1); this equation is a very interesting equation, the moment
you say that it is a regular solution model what are you saying, a regular solution model
is that where the A and B have some interaction; it could be a positive interaction, it
could be a negative interaction, but it is never zero. If it is zero there is no regular
solution model, its only ideal solution. But when you consider the entropy of mixing,
entropy of mixing is calculated based on a random; everything is random, there is no
specific interaction between A and B, so that problem always exists in any free energy
calculations. There are models which are available, which try to consider what is an
excess entropy that you get when there is an interaction between A and B. For example,
what is called short range order, when there is a short-range order how does the entropy
712
change, so there are detailed models available, but for our class I think it is a little too
much. So, now what we can also do is we can also find out; after finding out this kind of
the phase diagrams, you can also find out how the solubility changes with temperature.
For example, if you look at let say take a phase diagram like this (see top fig. in above
image) where L  α + B instead of β, you can also put β if you want. And now this α
composition changes with temperature, the solubility is changing with temperature. How
do I consider this solubility change, again do the same equation. For example, take this
equation
α α α 2 α
μB =G B +Ω B X A + RT lnX B
and try to draw a free energy composition diagram here (lower fig in above image). If I
draw a free energy composition diagram how does the free energy composition diagram
go, so one for α, and you have a β. β actually is a B; pure B, I can simply write it as β
here, but which is almost like a pure B and draw a common tangent between both of
them. So, on right hand side we have G αB ; pure metal B in the α form and one end of the
common tangent, we have μαB . So, the difference between these two is also can be
written as what is the free energy difference between α and β for the pure metal B.
713
So, if I say ΔG for the B which is α to β; β has a different structure, and α has a different
structure, this can be written as
α −β α α
∆G =G B −μB
From the above two equations, we can write
α −β α 2 α
∆G =−Ω B X A −RT lnX B
α 2 α
∆ H −T ∆ S=−Ω B X A −RT lnX B
And once you show this, then what we can actually write is find out how X B changes
with temperature.
α α 2
lnX B =( ∆ H−T ∆ S+Ω B X A ) / RT
Substituting X αA =1−X αB
you can show that how XB changes as a function of temperature. If you want we can
derive this at some stage, but you can go back and see how it can be derived, and
basically what you can see is how X B changes as a function of temperature because
temperature comes in the denominator, and in fact you will get a term which will be an
exponential term. If you assume that X B is very small, that means the solubility is very
small if you consider,
α
1−X B ≈ 1
and therefore,
α α
lnX B =( ∆ H−T ∆ S+ΩB ) / RT
So, so you can write
−Q
X αB =A . exp ⁡( )
RT
You will see that Q=∆ H+ Ω αB and A = exp (ΔS/R)

so you will see that the XB is a function of the temperature. The higher the temperature
the smaller the Q/RT, so that means the solubility increases. So, how solubility changes
as a function of temperature can be easily seen from this kind of an expression. So, we
can do all these calculations just by knowing what is the chemical potential at a given
temperature, and writing a chemical potential equation and finding out how the
composition changes as a function of temperature. So, all calculations of phase diagrams
are easier simply knowing one equation (5). This equation if you remember everything
714
else can be done, and this equation also comes basically by equating one basic free
energy equation G α =X A μ A + X B μB if you do that you will get that.
We will stop now.
715
Advanced Metallurgical Thermodynamic
Prof: Dr. B.S. Murty
Module #01
Lecture #30
Thermodynamics of heterogeneous system
So far, we have been talking about systems which are liquid and solid systems mostly.
Solidification is something which we have talked about to solid to liquid
transformations, we also talked about solid to solid transformations, variety of solid to
solid transformations. Let us start looking at a liquid gas and solid gas interactions.
These are what we call them as heterogeneous systems, where the moment gas comes,
pressure comes into picture.
So far, we did not actually bother much about the pressure excepting in one case, we
talked about Clausius-Clapeyron equation and how the melting point can change as a
function of pressure. We have talked about it if you remember.
716
Basically, we can write dG = Vdp - SdT. This is one of the Maxwell’s equations, the
most popular Maxwell’s equation. Do you know how does this come? You know how
this can be derived?
We can write
G = H – TS
Since, H = E + PV
Therefore, G = E + PV – TS
The, we can do partial differentiation
dG = dE + PdV + VdP – SdT – TdS
From the first law of thermodynamics, we know
dE + PdV = dQ, dQ is the heat content.
From the basic definition of second law of thermodynamics, we know
dS = dQ/T or dQ = TdS
Therefore, dE + PdV = TdS
Hence, dG = VdP - SdT
This is a very profound thing and we can express every phase with this expression.
So, basically, it connects the free energy with the state variables, volume, pressure,
temperature and entropy. Out of them, some are extrinsic, some are intrinsic. You people
know what are the extrinsic variables and intrinsic. Out of this, which are the extrinsic
and which are the intrinsic? entropy? Which are state variables
717
Student: Entropy
entropy and ?
Student: Pressure and temperature
yes pressure and temperature are the state variables, volume and entropy are not. So, we
can do that. So, any way, so once we can derive this today, let us look at and we use this
expression basically looking at for example, we can write this expression for every phase
transformation. We have done this thing earlier for liquid to solid.
718
For example, we can write
dGl = VldP – SldT
dGs = VsdP - SsdT
And we know that at equilibrium, both the free energies are equal. So, when we equated
these two, that is dGl = dGs
we got expression as
(Vl – Vs)dP = (Sl – Ss)dT
dT/dP = ΔV/ΔS = T. ΔV/ ΔH
Where ΔH is the heat of transformation, ΔV is the volume change during transformation.
So, the change in the transformation temperature, any transformation temperature for
that matter is related to the pressure by this expression. This is what we have talked
about it in Clausius-Clapeyron. Today, let us look at sublimation because when we want
to talk about heterogeneous reactions gaseous, things have to be involved. So, let us start
with sublimation as a reaction, where there is a gas that is involved before we actually
start dealing with, solid gas reactions or liquid gas reaction.
So, let us look at a system where a solid is sublimating into a gas. If that is the case, then
we can actually start writing the expressions for that in terms of the gas or for the solid.
719
So, basically if you can write this expression for the gas
d G = VdP - SdT and if you imagine that there is a chamber where there is a solid which
is sublimating at a constant temperature, if you assume that the temperature is constant,
then in principle, then dT = 0. So, if that term vanishes, then we can talk dG = VdP.
So, if you want to consider this Clasisus-Clapyeron expression in detail.

dP ∆ S ∆H
= =
dT ∆ V T ∆ V
When we talk about this in terms of ΔV here, ΔV = Vg – Vs.
And Usually, Vg >>Vs
Student: Should it be Vs - Vg
It depends on which direction you want to look at it. You can look at that way or this
way. Then, both ΔS and ΔV becomes negative, and minus will get cancelled.
So, when we assume Vg >>Vs , then ΔV = Vg and when there is a gas, gas, we can
always use the ideal gas equation where we can write R T = PV. So, we can write that
V = RT/P.
720
So, if you do that, then we can write
dP ∆ H . P ∆ H . P
= =
dT T . RT RT2
dP ∆ H . dT
=
P R T2
dlnP −∆ H s
=
d (1/T ) R
Then we can basically say that if you see how the vapor pressure changes with
temperature. It is actually always related to the ΔHs of sublimation. So, in principle, we
can say now, if you try to integrate this, you can always write that
lnP = -ΔH / RT + A
So, in principle, that means if you plot ln p verses 1/T, you will always see an expression
something of this sort where the slope of this line (upper graph in above image) turns out
to be ΔHs/R. So, this is how if one can measure the partial pressure of any gas, where a
solid is sublimating and there is a gas inside the chamber and as you change the
temperature, how this partial pressure of the gas is changing. If you try to
721
measure that partial pressure of the gas as a function of the temperature, you would see
that these two are related to this and the higher the sublimation ΔH s, the higher ΔHs /
RT , the smaller will be the partial pressure. This is expected because if the sublimation
enthalpy, the ΔHs associated with sublimation is very large, then at any given
temperature because you need more energy for the sublimation.
So, the amount of gas molecules generated from a solid are always going to be lower.
So, that means, particularly when we are looking at metals with a higher boiling point, a
metal with a higher boiling point basically means, its ΔH s is very large. That is when
actually you will have a boiling point being very high. So, whenever the ΔH s is very very
large, you would say that this material is more difficult to sublimate. So, you need to
provide more energy for the sublimation to take place for any solid molecule to be
converted into a gaseous molecule. So, if that is a case, the partial pressure of the gas
that is generated at any given temperature is going to be , inversely proportional in
principle.
So, as a result, we can say that whenever the sublimation enthalpy is higher, this is also
similar to saying that whenever the heat of fusion is very large, we can say the metal to
melt is more difficult. Isn’t it? As simple as you need to because heat of fusion basically
is something to be provided to the solid, so that a solid gets converted into a liquid. So,
the larger this value, the more difficult is for the solid to melt. Similarly, the larger the
heat of sublimation is, the more difficult for a solid to sublimate. So, the partial pressure
of the gas that you generate because of the sublimation of a solid is always inversely
proportional to this and the larger this value, the smaller will be the partial pressure.
So, by measuring the partial pressure as a function of temperature, even if you do not
know this ΔHs one would be able to get at 4 or 5 different temperatures. You measure
the partial pressure of a particular metal vapour in a chamber. There are so many
methods of the chromatography methods are available, spectroscopic methods are
available to measure the partial pressure of the gases and if you can do those things, then
at various temperatures plot, it get a least square fit for that and from that, you will be
able to measure this sublimation temperature.
722
Or What is the other method? Does anybody know? If I want to know what the enthalpy
of sublimation is?
Student: DSC
DSC is a direct method? So, take a sample in a DSC, provided the DSC can go to the
temperature, where your sample sublimates. Unfortunately, most of the DSC’s go up to,
let us say 1500 ºC or so. The one that we have in our lab goes to about 1500 ºC. So, if
the sublimation temperature of the material is below 1500 ºC in principle, you would see
a strong peak associated with the sublimation (see the lower curve in above image). Of
course, whether it is up or below depends on what is the polarity that we are talking
about, so but basically sublimation being an endothermic reaction, you get an
endothermic peak at the sublimation temperature and this (point A) basically represents
the onset of the sublimation. And What does this (point B) represent? Anybody knows
in any DSC what does the peak indicate?
What does the peak indicate in a DSC? You always know that there is a temperature
difference between A and B. They are not the same. So, there are 2 temperatures here
associated with in principle, there are 3 temperatures (point A, B, C) associated with any
phase transformation or any reaction in a DSC. What are these 3 temperatures? What is a
physical meaning of these 3 temperatures? I take anything. Let us say I am melting, a
metal solid I get a peak again similar and there will be 3 temperatures.
Student: Onset represents starting
Onset represents beginning of the melting,
Student: Peak represents the melting point
Then, onset is also a melting point
Student: Where it completes
where it completes that is a peak.
Then what is third one? So, obviously, the first and third are more easy to answer the
onset and the completion.
Student: Maximum rate
maximum rate. The peak represents always the temperature, where the rate of
transformation is the highest. Rate of transformation is the highest. So, one can, but the
thing is in principle, whenever you want to do what is called reaction rates, for example,
any rate of transformation if you are interested and what is called kinetics of the
transformation. In fact, if you have any phase transformation taking place, let us say I
723
take a metallic glass. I am crystallizing a glass. The moment I am crystallizing a glass, I
get a peak. I have mentioned to you as possibly numbers of the times.
724
We get a peak like this. Am I right? A glass when you start with and then, you are
heating it, you get something like this. What do we call this? What is that?
Student: Tg
T glass transition temperature? So, why this glass transition temperature is like this?
Student: Second order transition
It is a second order transition? So, glass transition is second order transition and what is
the characteristic of the second order transition?
Student: Second derivative is discontinuous
Second derivative is discontinuous. What about the first derivative?
Student: Continuous
They are continuous. What does that means in terms of thermodynamic parameters
enthalpy?
Student: There is no enthalpy involved
There is no enthalpy involved? There is no enthalpy change involved. Please be careful.
So, the ΔH for the transformation is 0 and the ΔS for the transformation is 0 and the ΔV
for the transformation is 0.
So, that is a second order transition. A second order transition will be that where the ΔV
is 0. There is no change in the volume during the transition; there is no change in the
enthalpy during the transition, whereas in the first order transition, all these three are
discontinuous. So, there is a change in enthalpy and that is what we call it as heat of
725
fusion. There is a change in the entropy and that is what we call it as entropy of fusion
and there is a change in the volume, which is what we call it as volume expansion or
volume change during melting or solidification. So, whereas, here no and then, this
(point A) is what we call it as crystallization and if any time you do a DSC, many a
times, if the ΔH involved in the reaction, if it is a very large, what kind of
transformations are those where Δ H is involved is a very large? Tell me? When will be
the ΔH involved in any
726
particular change is going to be large
Student: explosive reaction.
Explosive reaction, that is a reaction. Let us not talk about reactions here. We are talking
of phase changes,
Student: Change in the state
change in the state.
Whenever there is change in the state of the matter, then the Δ H is large solid to liquid,
solid to gas, liquid to gas. When within a particular state if there is a transformation
taking place, solid to solid let us a precipitation, a martensitic transformation, it could be
order disorder transition. So, all these things usually excepting order disorder because in
order disorder, there is large ΔH negative that is involved between order state and
disorder state. As a result, many of a times the order and disorder transitions ΔH is very
large. So, the peak is going to be very large, a big peak whereas, all other.
For example, if I am taking iron, pure iron and transforming a BCC iron to FCC iron, the
ΔH is very small and when the ΔH is very small, many a times this peak is going to be
very broad. Particularly, if you are talking about a cold worked material, some of you if
you have ever done any re-crystallization of a cold worked material and do it by DSC let
us say, you get a usually a very broad peak because it happens, usually thhe recovery and
re-crystallization tend to overlap and you see a broad peak rather than a very sharp peak
and when you get a broad peak, then you identify the onset become very difficult. The
way we identify an onset in any peak is you extrapolate the base and left arm of the peak,
wherever these two intersect (point P) is what we take it as the onset of transformation.
So, and whenever the peaks are very broad, if you a get broad peak like this, it becomes
very difficult for you to identify. You may still say that I will do this and then, find it out
some places. It is more difficult to see and particularly in some places where you have
two peaks overlapping. Let us say, so you will see a situation where something like this
(shown in bottom figure in above image). In such a case, what is the onset of this? It is
very difficult to find out. Then, what I should do? In principle is, I should do
deconvolution and then, make it as two peaks and then, find out the onset for left and
find out the onset for right.
So, under such conditions, people usually consider the peak temperature representing the
transformation. Though we know peak temperature is not the temperature where the
727
transformation has actually started, we definitely know the peak is not the temperature
where the transformation has started, but in case where it is more difficult to find out
what is onset, we should do this. Otherwise, for all, theoretical purposes onset should be
728
actually taken as the transformation temperature and in case of melting kind of things,
the onset and offset what is called offset are going to be very close unless you really do
very fast heating.
When you do fast heating, then the onset to offset there will be differences. Otherwise,
usually they are very close to , to each other and using this information in principle, one
can get a lot of kinetic information. For example, doing this (glass transition DSC) if you
do this at different heating rates, what will you observe? Let us I have a glass. I take the
glass, put it in a DSC and heat it at a particular rate and take another piece of the same
glass and heat it at different rate. Do I get the same pattern? If I do not get the same
pattern, what will be the difference in the DSC traces that I get at two different rates?
Let us see, one is let us say R1, another is R2 where R2 > R1. Now, tell me what will be
the difference that will you get?
Student: Peak will broaden
Peak will broaden anything More?
Student: Peak will shift
Peak will be shifting towards what? In which case the peak will be of at a higher
temperature?
Student: R2
R2. What is the logic?
Student: Kinetics, Rates of transformations, greater uniformity
729
It is basically the inertia of the system. Whenever you heat a sample at a higher rate, the
sample does not transform at its equilibrium transformation temperature and it
transforms at a temperature higher than its equilibrium transformation. This is similar to
taking a liquid and under cooling it is very fast. It does not start melting solidifying at the
normal regular equilibrium transformation
730
temperature. Instead, it under cools to a lower temperature and the transformation starts
at a lower temperature.
Exactly similarly, whenever I am solid, in principle melting point also shifts. The only
difference is the shift in the melting point with the heating rate is not as much as the shift
in the freezing point with a cooling rate. The reason is also very obvious that in case of
freezing, a new surface has to be created. So, there is a activation barrier and because of
the activation barrier, the system when it is being cooled very fast, you are not providing
sufficient time for this nucleus to be created. As a result, it under cools where as if you
are heating it a solid into liquid. Then, you see that because there is already a surface
available, you would see that it does not need that much of an activation as u need it in
case of a freezing.
So, as a result, the shift is not as much as possible, but here it is again a solid to solid
transformation. A glass crystallizing is not exactly like a liquid I know being freezing.
So, as a result, you would see there is a significant shift to higher temperature if the
heating rate is higher. So, by basically heating it at different rates, 4 or 5 different rates
what you get is different temperature, starting temperature for different heating rates R 2,
R 3 and so on. You will get what is called T X1, TX2, TX3 and so on. So, identify for each
heating rate, what is the transformation temperature. If the transformation temperature is
sharp enough identify, what is the onset. Otherwise, consider the peak temperature and
once you have that, you can in principle plot what is called a Kissinger plot which is
nothing, but
lnβ/Tx2 versus 1/T (see the lower graph in image below), where β is the heating rate and
TX is the crystallization temperature or each of the heating rate versus 1/TX.
731
If you do, the slope of it can be taken as Q/ R, where Q is the activation barrier exactly
similar to what we have done few minutes back for a vaporization, where we have taken
ln p by 1/T and plotted it. So, similarly, here you simply take the β and T X, tabulate the
things at least four or five different temperatures if you do, because basically to be able
to get consistent plot do a least square fitting and this is how for any such reaction,
whether it is a polymorphous transformation or an allotropic transition or a
crystallization of a glass or any other transformation, precipitation, all of them you can
use this, this is so-called Kissinger plot. Of course, by the name of the person who has
talked about it which is a very famous.
So, all phase transformations, kinetics are done with this and this is all what is called
thermal kinetics. What do we mean by athermal kinetics? It means, as the functions of
heating at different rates you are doing it, you can also have what is called isothermal
kinetics. That means, at a given temperature, you try to have the transformation taking
place in principle. If the same transformation I do it in a DSC, what we plot is ΔH versus
temperature. The heat evolved versus the temperature is what we are plotting. You can
also plot the heat evolved versus time, basically once I know the heating rate β . What is
β ? β = dT/dt. Am I right?
So, this if I take the temperature, T divide it with β ,it (T/β) should become actually time
for me. Somewhere, it should be related to time. Isn’t it? So, in principle heating rate if I
know
732
and if I know the temperature axis, the temperature axis can be converted into time axis.
Hold it at a given temperature and as a function of time, you see how the phase is
evolving and in this process, you can talk about how much percentage of transformation
has taken place. So, we have onset temperature, completion temperature and at any given
time in between, you can talk about percentage transformation. It is similar to our T-T-T
diagram.
At a given temperature, you are talking of onset. I mean start of perlite formation, end of
perlite formation. Exactly similar you can do a DSC which is called isothermal DSC, and
find out how is the fraction transformed and that fraction transformed can be correlated
to activation energy. We will not talk about it at the moment because this is more too
much of kinetics, but some time, if you are interested will talk about it.
So, this is all about how we can see that partial pressure ln p of it is related to. We have
seen that ln p = -(ΔH/ RT) + A. And In principle, if that is the case, one can also write if
we can write it in the exponential form. p = A.exp(-ΔH/ RT).That means, partial pressure
is exponentially related to the temperature, partial pressure of any gas. That is the reason
why the temperature is very very important in terms of this. We can also see whether
how the external pressure influences the partial pressure.
733
Let us say, I have a solid which is sublimating and I put some pressure, the chamber
where I have the solid which is sublimating. I introduce some gas, let us say an argon gas
at a certain pressure P and I want to see how this influences it. If you want to see that,
then what we can do is let us try to do. Let us write the expression here. In this case, we
have a condensed phase, we have a gas phase. For both of them, if we write the dG
expression,
dGcond = VconddP - SdT
and because the temperature is constant, dT = 0
dGcond = VconddP
So, if there is any external pressure with change of pressure, the dG changes as above.
Now, if you do the same thing with a gaseous thing,
dGgas = Vgasdp
At a given transformation temperature, both of them should be equal, the free energy of
the gaseous phase and the free energy of condensed phase should be same. So, we can
equate dGcond = dGgas
and so VconddP = .cdp
In gaseous case, we are using a small p because we are talking of the gas, partial pressure
of the gas and when we are talking of the external pressure, we use P. So, we can write
the rate of change of the partial pressure of the gas with the external pressure as:
734
dp/dP = Vcond / Vgas
Now, Vcond < < Vgas

Usually, the ratio Vcond / Vgas ~ 10-4 . So, if that is the case, then you can show that the
partial pressure of the gas that we have the sublimated gas that is forming, the change of
it with the external pressure that we are introducing is actually very small.
Why this becomes important is whenever you are trying to sublimate something.
Particularly for those people who want to make nano materials by gas condensation
techniques. Let us say, you take a metal. You want to evaporate and form gas and then,
condense it and get nano particles and you want to do it at various pressures. You want
to know how this sublimates because the rate of formation of nano crystals is related to
the rate of sublimation and rate of sublimation is related to this. So, if I have the certain
gas inside the chamber, the gas pressure inside the chamber, how does it affect the partial
pressure of the gas that is generated because of sublimation.
You can see that the partial pressure of the sublimated gas does not get changed by the
external pressure to a large extent, unless this pressure P is very very large. So, if I take P
as one atmospheric pressure, at one atmospheric pressure this p changes only by 10 -4, not
much. So, this is something which you have to see, whereas if you look at the previous
thing, the partial pressures changes exponentially related to the temperature. While the
partial pressure is exponentially related to temperature, the partial pressure is not related
significantly with the pressure, so these are the two aspects. One has to remember
whenever we are talking of gases coming out of the solids and how the temperature and
pressure which are the two important state variables, how do they influence the
formation of gases from a solid.
So, this is the first concept that we need to understand when we are talking of
heterogeneous reactions involving gases. How pressure and temperature influence the
735
sublimation influence the partial pressures. Now, also we understand this. Let us go
ahead with reactions and start talking about how the reaction rates and the
thermodynamics of reactions depend on pressures, temperatures and things like that.
For that matter, let us simply take up a simple case, where

A + 1/2O2 AO
If you are taking such a reaction, how do we understand thermodynamics of this kind of
reactions? To understand the thermodynamics of that kind of reactions, one or two
concepts which we should already know, which we have possibly done it earlier. If I
want to write G free energy of anything, we have written it in different ways.
736
One we have written is

G= X A Ǵ A + X B Ǵ B (i)
We have also written
G= X A G A + X B G B +∆ G mix (ii)
And we can also write
0
Ǵ A =G A + RTln a A
G0 is the standard state. So,
0 0
G= X A (G A + RTln a A )+ X B (G B + RTln aB ) (iii)
If it is an ideal solution. aA = XA. If that is the case, then

∆ G mix =RT ( X A lnX A + X B lnX B )
where ΔHmix = 0. So, basically if you put in eq. (iii) a A = XA, so it becomes actually
equivalent to equation (ii) excepting that in eq. (i) it is XAGA.
The moment I say it is not an ideal solution, then I can write a A = γAXA, where γ is the
activity coefficient. Then, this γ actually brings out the ΔHmix. I can put in eq. (iii)
ln aA = ln γAXA
ln aA = ln γA + ln XA
The ln γA is what is related to the ΔH. So, the ΔH mix is always related to the activity
coefficient gamma.
737
So, depending on how you express this equation (iii), whether it is ideal solution or a
non-ideal solution, you can express the whole thing. But What is more important is you
can take Ǵ i=G 0i + RTln ai. If you remember this, we can do a lot of things. You hold on
there we will come back to that. As we go along if you take this expression, where one
solid is reacting with another solid and then giving you and reacting with a gas to give
you an oxide, assume that both (metal and oxide) of them are solids and reacting gas is a
gaseous phase in such a case.
We can write what is called the ΔG of this reaction as

0
∆ G=∆ G + RTln a
When we say the a of the reaction, you have to take it as individual activities of each of
them. So,
aAO
a= 1/2
aA pO 2
So, you can write this whole expression,
0 a AO
∆ G=∆ G + RTln 1/ 2
a A pO 2
and in this case, we always know that the activity of solids in their standard states is
equal to 1, as long as they are
738
pure. So, if this reaction, if a pure A reacts with oxygen and gives you a pure oxide
without any impurities there, then aAO = aA = 1 and we can write
1
∆ G=∆ G 0 + RTln
p1/
O
2
2
and for any such a reaction at the equilibrium transformation temperature, ΔG = 0.
So, you can write this

1
0=∆ G 0 + RTln
p1/
O
2
2
This is where we say

1
∆ G 0 =−RTln
p1/
O 2
2
So, this is how, so if I know what is the partial pressure of oxygen during the reaction, I
can actually find out what is the ΔG of the reaction without knowing or if I know the ΔG
of the reaction by some means, if I can find out the Δ G of the reaction by, by let us say
doing a DSC.
If I do DSC, what do I get from the DSC?

Student: ΔH
ΔH of the reaction I will get. If I know the temperature of the reaction and if I know the
entropy of the reaction from the Δ H basically, in principle I will be able to get the Δ G.
So, either measure the ΔG or measure the pO2 and you can see both of them are related
to each other. This whole reaction kinetics, I mean reaction thermodynamics are all
related. Here
aAO
K= 1/2
aA pO 2
K is equilibrium constant or reaction constant K, where it is the activity of the products

divided by the activity of the reactants.
We will take up a few more examples in the next class where we have not simply one
gas. There can be a number of gases involved and most of our actual reactions in
metallurgy, we deal with a number of gases that are involved and try to see how these
reactions change as a function of the free energy and as a function of temperature and
also see as a function of pressure how things change with that. We will stop now.
739
Lecture No. # 31
Thermodynamics of heterogeneous system
Yeah we started talking about reactions. Let us continue and know a few aspects about
various reactions. We start with first of all the relation between the G and a. This is well
known, but still I thought it is better; we look at it once carefully before we go ahead. If
we look at basically the expression dG = VdP - SdT; this is what we all know, am I
right? So at constant temperature this is dG = VdP, and if we assume that there are gases
involved in the reaction, we can use the ideal gas relation, where RT = PV. And then,
use this V in terms of RT and P, and write V = RT/P. Once we write that we can write
dG = RTdP/P = RTdlnP. So, that is how you actually get this expression, and in this now
if I simply look at two pressures, let us say.
I can say
dG1 = RTdlnP1
dG2 = RTdlnP2
Or if we can write it in terms of G alone, you can simply even not write this; simply
write G = RTlnP, the d can be removed. So, you can write
G1 = RTlnP1 and G2 = RTlnP2
740
So, whenever you want to look at
741
differences in the free energy at various pressures, you can simply take as
G2 – G1 = RTlnP2/ P1
So, like this we can simply look at how the free energies change as a function of pressure,
and if one of them, let us say P 1 is a standard pressure, what is the standard pressure for us?
Atmospheric pressure.
I will take G 1 as related to atmospheric pressure and then

G – G0 = RTlnP
Because if P1 = 1, then P1 goes off in this equation. I will call it as standard free energy at
atmospheric pressure. So, that I will refer it to as G 0; so we always refer free energy at
atmospheric pressure as G0. So I can write G – G0 = RTlnP
and the same thing if you are talking in terms of activities, I can write the same thing as
G – G0 = RTlna . If we are talking of a mixtures of it, , you can write G i – G0 = RTlnPi for a
particular gas. If there are mixtures of a number of gases, for a particular gas we can write i th
gas; you can always write this expression. So this is how, we can express the free energy
changes in terms of activities. And once we know this, we can start looking at various
equations; various reactions that take place and look at, how we can correlate the free
energies with the activities. And then,
742
from that get to know information about whether the reaction will go in the forward
direction; whether the reaction will go in the backward direction. All this can be
accumulated, once we are aware of it.
So, you can see starting from a simple Maxwell’s equation which is dG = VdP - SdT and
this I hope all of you know, how we derive it. I will not need to derive it. So, we are
simply assuming the temperature to be constant and from there, we have arrived at this;
for a given atmospheric pressure, one pressure being atmospheric pressure. And from
there we have come to the activity. Now, let us look at a reaction where let us say
A + B  C + D. And here, to make it more specific, we will call it as
aA + bB  cC + dD
And you can see this kind of reaction variety of reaction, where certain numbers of
moles of a particular component are involved in the reactions. For example, if you take
Mg simple example, Mg + O2  MgO, you need to balance the equation as
2Mg + O2  2MgO
When I say small a b c d, small a b c d refer to those moles and if that is the case, then
the free energy change for such a reaction is nothing but the free energy of the products -
the free energy of the reactants.
So, I can say the G of the reaction is

ΔG = cGC + dGD – [aGA + bGB]
I can also write the same thing at a standard state as:
ΔG0 = cGC0 + dGD0 – [aGA0 + bGB0]
Everything remains the same excepting that, we are talking at the standard state.
If I take the difference of this two, it can be written as:

ΔG - ΔG0 = c(GC -GC0) + d(GD-GD0) – [a(GA-GA0) + b(GB-GB0)]
And we already know that G – G0 = RTlna
743
So I can write ,
ΔG - ΔG0 = cRTlnaC + dRTlnaD– [aRTlnaA + bRTlnaB]
Basically, now we have got the whole free energy change in terms of activities. Any
confusions? We simply looked at the free energy change of the reaction. We took it as
products - reactants and then, we looked at the same free energy change at the standard
state. And the difference between the two brings out that, there is a component, which is
a free energy in at any given condition - the free energy in the standard state and this
difference is always RTlna.
And so from this, I can write

ΔG - ΔG0 = RTln[aCc. aDd/ aAa. aBb]
The bracketed item is what we call it as K, equilibrium constant. And we learn a lot of things
from this as we go along; we will see, how does this influence the
744
whole reaction and this, the whole reaction, whether it will go in the forward direction or
the backward direction always depends on that.
So, I can write this as

ΔG - ΔG0 = RTlnK
If there is equilibrium that is established between the reactants and the products, then this
ΔG = 0. Whenever there is an equilibrium between anything, we have seen liquid to
solid transformation ΔG = 0. So, the same argument I can hold even for Mg + O reacting
to give you MgO. The change in free energy for such a reaction should be 0, if all the 3
components whether it is products and the reactants; if everything is under equilibrium,
then the ΔG= 0.
And if that is the case, then you can say that ΔG0 = -RTlnK
This is a very crucial equation; which will tell you that, we can correlate change in the
free energy of a reaction in a standard state to the equilibrium constant; which is nothing
but activities of various components. And this is how, we will be able to see how to
understand reactions and maybe, I will just take one or two examples just to give you.
Whatever we have done so far, is it clear?
Let us say, you are melting steel in a MgO crucible at some temperature 1600 ºC or so.
You want to know, whether this MgO crucible will be intact at that temperature, or will
it dissociate into Mg + O. If it dissociates, obviously
745
you cannot melt the steel at in that crucible; it is no more stable. So, if you want to melt
anything in a crucible, that crucible material should be stable at the temperature. So if
that is the case, how do I find out that? Simply, what you do is find out what is the
reaction that goes on?
2Mg + O2  2MgO
If this is the case, then find out what is the ΔG 0 for this reaction? If the Δ G 0 for the
reaction is available, usually ΔG0 = A + B T. Why it is written in the form of A + B T is
basically, because ΔG0 = ΔH -TΔS. So, whether the B is negative or positive, let us not
bother about it. So basically, because it is ΔH -TΔS. So, you would always see the
expression for the ΔG0 is always given in form of two constants and linear dependence
on temperature.
Once you know the A and B terms, for that particular reaction which are available based
on basically, people have done such reactions at various temperatures and found out,
how the free energy is changing. How do you find out the free energy change for a
reaction?
Student: From change in Cp
from Cp change anything else? Look at particularly, the thing that is easily measurable
for any reaction is the enthalpy change. Because that is something you can measure
using a calorimeter. So measure the enthalpy change at various temperatures for a
reaction.
And once you measure the enthalpy changes at various temperatures for a reaction and
keep the pressure constant; because we are all talking of ΔG 0 there. At atmospheric
pressure, do this reaction at various temperatures. At various temperatures, measure how
the H is changing and from that, you can actually get ΔG 0. In fact, people have done this
and for various reactions and these standard values are available. If you go to any
standard hand books of, on thermodynamics, you will see for various reactions.
What is this in terms of temperature is what are they? Particularly, what are the A B
values? For various oxidation reactions, various reactions where sulphites form; sulphide
form; sulphates form and such reactions they are all tabulated. You go to any
thermodynamics hand book, you will see it. For example, the book one of the good one’s
is what is called Kubaschewski . If you go to this particular book, it is thermo chemistry.
746
A book on thermo chemistry, which gives you towards the end so many appendices and
each appendix, gives about various such reactions, what is the values of A and B.
Student: Are A and B function of temperature?
Once I know this, A itself is a function of temperature you mean? No, A and B are
constants I mean if they are the functions of temperature, then obviously these functions
becomes more complicated. So, A and B are usually fixed. Because that is how in
principle, if you go a little detail, if you think these two are nothing but ΔH - TΔS. These
two are also functions of temperatures, in principle yes. But for in such a case, this whole
expression becomes a little complicated expression where in each of them, you will get
temperature dependence. But to a large extent, people have seen that A and B are
constants for most of the reactions.
You simply take A and B as constants and do it and that works out for most of the
reactions.
Student: Sir, we already have Magnesia crucible, then why are we considering that
reaction?
Yeah why are you see the question is, if this MgO crucible at that temperature will it
dissociate? That is what, we are interested in. At any temperature, whether the reaction
will go in forward direction or whether the reaction will go in backward direction; we do
not know. What does it depend on? Whether the reaction will go in forward direction or
in forward direction depends on what is called partial pressure of O. For this reaction,
there is a equilibrium partial pressure.
How do I calculate the equilibrium partial pressure? Simply, find out what is the K? So, I
know K of this reaction. I can write it as,
K = a2MgO/(a2Mg.pO2)
In case of gases, we call yeah
Student: The dissociation reaction is
MgO  Mg + O2.
Student: That will be happening, since we already have MgO
We are we are trying to find out, whether that will happen or not?
Student: Dissociation has to occur, means MgO should dissociate
Correct. The point is, at that particular temperature and pressure that we have whether
this will happen or not. For example, this melting is taking place in a furnace let us say,
where there is certain O partial pressure.
747
My furnace is under certain O partial pressure. Now, for that O partial pressure, which is
available in the furnace whether this reaction will happen or not? Whether the whether
the MgO will dissociate or whether the MgO is stable? How will I know? If I know what
is the partial pressure of equilibrium partial pressure of O for this reaction at that
temperature; for example, if I know the ΔG0 as a function of temperature, by putting
1600 ºC in to it. Mind it; you should remember that, whenever we are talking of any
thermodynamics, we talk of temperature in Kelvin.
748
So, come let into Kelvin; put it into ΔG0 = A + B T . So, you will know ΔG0 and
ΔG0 = -RT ln K. So, from once I know ΔG 0, I will know K and once I know K, I will
know pO2; P, partial pressure of O, which is the equilibrium partial pressure of O at that
temperature, which will keep on changing at different temperatures. So, once I know that
and what is the actual pressure inside the furnace if I know, and if the actual pressure of
O is higher than the equilibrium partial pressure of O, then what would happen?
Student: Reaction will go in forward direction
Reaction should happen in the forward direction. Whenever O content is more, the Mg
should get oxidized And It should go in forward direction. If the O content inside the
furnace is lower than what is the p O2 of the reaction, then what happens is the system
would go in such a way that, O is created in the process. So that, it maintains the O
partial pressure inside the furnace equivalent to the equilibrium O partial pressure. So, let
us say you are doing this in vacuum. Let us say at 10-4 torr.
So, if the melting is taking place in vacuum and then you find, what is the p O2 for that
reaction at this temperature, 1600 ºC? In fact, if you do it, what you would get is that
partial pressure of O will be actually 10-22. You will get as 10-22 or 10-23, it comes out. So,
this value is much smaller than this. As a result at this particular vacuum level, at this
particular pressure level, because the actual O partial pressure is much larger than the
equilibrium partial pressure of O. The reaction would go in forward direction; because
O is available than the O partial pressure of the reaction.
If the O that is available is less than the partial pressure of O that is required to maintain
the equilibrium, then what happens is MgO will dissociate and will give you O, by the
way. So, that this O joins the atmosphere to maintain the equilibrium. So, this is how, I
can simply calculate, what is the partial equilibrium partial pressure and find out what is
the actual partial pressure inside the thing? And then, find out whether this Mg O is
stable or not stable? This can be done for any of the things when see all meltings are
done in certain crucibles; whether it is alumina crucible or silicon carbide crucible or
whatever kind of crucibles, so you need to make sure that, these crucibles are stable.
749
Particularly, when I am why I have to be bothered is whenever I am doing melting in
vacuum; vacuum means lower pressure. Whenever the pressure is low, the volume has to
be high. P and V are inversely proportional. So, as a result, the reaction will go in such a
direction, which will increase the volume of the system. So, what can increase the
volume of the system, if the O can come out? So, as a result, there is always a tendency.
Whenever I am melting a material in vacuum, the oxide which is there in the crucible
may dissociate to give you O provided that particular pressure in the furnace is smaller
than the O partial pressure under equilibrium conditions, only if it is smaller than and if
it is larger, then it will not happen. So, this is one such example, we can think of many
many such examples. Lets say, you are trying to anneal Cu in a atmosphere, which is
containing hydrogen and H2O. Most of the time annealing is always done in a in
furnaces, which are actually heated by oil; various oils. So where, there is always
hydrocarbons; these hydro carbons will break and give you some hydrogen and some
moisture is also expected in the reaction.
So, there is what is called a partial pressure of pH 2/pH2O. If I know what is the actual
pH2/pH2O ; that is being maintained in the furnace. One can do what is called mass
spectroscopy and things like that, to measure the actual partial pressures I mean various
gases in any furnace atmosphere. And if I can measure and find
750
out, what is the actual value inside the furnace. Now, I would like to know whether this
Cu, which is being annealed in that furnace, is going to get oxidized or not. How do I
know that? To know that, first of all look at the Cu oxidation reaction.
4Cu + O2  2Cu2O (1)
and the other reaction, that is possible is

2H2 + O2  H2O (2)
Now, look at a combination of these two reactions. If you subtract this two and
rearrange, you can say that,
4Cu + 2H2O  2Cu2O + 2H2 (3)
So, that means H2O can oxidize Cu and to give you Cu2O and H2 can reduce Cu oxide to
give you Cu. What does that depend on?
It depends on what is the actual values of H2 and partial pressures of H2 and H2O. For
example, if I know the free energy change for this reaction (1), ΔG 0 and I also know the
ΔG0 for this reaction (2), let us say; as a function of A + B T kind of thing. If I know
that, by subtracting one from the other, I will be able to know the ΔG 0 for the reaction
(3). And because you can see, there are 4 moles and 2 moles in both the sides, I can
actually make it simplified;
2Cu + H2O  Cu2O + H2 (4)
Whenever I am dividing with 2, obviously the ΔG of the reaction also has to be divided
with 2; this has to be kept in mind. So, whatever ΔG 0 I am getting by, subtracting ΔG 0
of reaction (1) and (2) that has to be divided by 2; So that, you know the free energy
change for this particular reaction (4). And for that reaction (4), if I now want to write
down what is K?
K = (aCu2O.pH2)/(aCu2.pH2O)
And if we assume that, this reaction is giving you a pure Cu oxide and a pure Cu is
getting oxidized in this particular condition.
There are no impurities in that, then whenever it is pure, the activity is = 1. So, there is
no second component. Whenever there is a second component, then we have to consider
what is the activity of it. So, activity basically becomes 1, when it is a pure component.
So, if pure Cu is getting oxidized and giving you pure Cu2O,
751
then the activities are 1 for these two. Then, I can write
K = pH2/pH2O
So, I can find out if I know the Δ G of this reaction. What is the equilibrium partial
pressure ratio of H2 to H2O? This can be easily found out.
Once I know the ΔG0, I can find out K using ΔG0 = -RTlnK and K = pH2/pH2O. So, I will
know what is the equilibrium and once I know, what is the actual value of pH 2/pH2O in
the furnace and if that actual value of pH 2/pH2O is greater than this equilibrium value,
then what does it mean? I have an excess hydrogen in the furnace. Remember always use
partial pressure as small p of the standard state. If (pH 2/pH2O) > (pH2/pH2O)0 then I have
excess H2, If there is excess H2, then that means what? the reaction (4) will go in
backward direction. If (pH2/pH2O) < (pH2/pH2O)0, then the reaction will go in forward
direction; that means Cu will get oxidized. So, this is very, very important, whenever
you are doing annealing. For example, steel you want to do hot rolling of steel. Before
hot rolling of steel, you have to heat the steel at 900 ºC or 1000 ºC for some time, before
you start rolling it. And you do not want this whole steel to get oxidized at that
temperature.
So, if that is the case you need to know what is the pH 2/pH2O ; that I should maintain
inside the furnace; so that, the steel does not get oxidized. So, you can see that, all these
the actual practical applications of the thermodynamics come. When you want to use
simply hydrogen annealing of Cu or you know hot rolling of any of the steels or
whatever it is aluminium alloys that you are talking about. So all is related to the
presence of pH2/pH2O; of course, depending on the actual equation, if I am talking of Fe,
then I have to actually see what is the type of oxide that is forming is. Is it FeO or is
Fe2O3 or Fe3O4?
And accordingly, the actual numbers will change a little bit and then, so here also in
equation of K the values will be slightly different. Otherwise you will see that, but as
long as the gas molecules are not changing, then this K will not change. Because we are
not bothered about, what is the metal and oxides much because we are assuming that,
they are But then, this may not be always true for alloys. For example, if I am looking at
752
steel. Steel is not pure iron. So, I need to consider when I say activity of iron, there I
need to find out how much of iron is there in the steel?
Of course, luckily for a plain carbon steel, we are talking of mild steel. Let us say, is
about 0.2% carbon. So it is more or less can be treated as a pure iron. So, activity of iron
actually is 0.9998. If you are considering a steel containing 0.2% carbon, the activity of
carbon is accordingly the value 0.0002 and activity of iron will be 0.9998, if it is a binary
alloy. So, if it is a you know multi-component alloy, if it is a stainless steel that I am
talking about then accordingly I have to consider what is the activity of?
So, the actual activities that we are talking here, which we are equating it to 1; this would
not be true, if I am considering alloys. If I am talking of let us say super alloy, where
there are various elements present. And I need to look at, let us say oxidation of each of
the alloying element; I need to consider; for example, in a super alloy again super alloy
also I am heat treating at certain temperature to do hot rolling of super alloy. Let us say, I
want to find out whether nickel of that super alloy will oxidize; whether aluminum of
that super alloy will oxidize or not. If you want to find out, you have to actually find out
for each of them equations like this.
And the activities of the oxide and the metal that, you are putting in this equation should
be accordingly the compositions of those. Once I know the composition of it mole
fractions of it, activity is equal to mole fraction And again, this issue activity equal to
mole fraction also is not so easy. When is activity = mole fraction? when it is an ideal
solution. If I am assuming it be an ideal solution, and ideal solution works out only when
in a very dilute conditions. Whenever it is a very dilute alloy, then I can think of that; but
if I have alloying elements of the order of 20, 30 %, the ideal solution may not work out.
So, that is where you have to consider, then activity is activity coefficient into the mole
fraction. So, all these complications will come into picture, when you are talking about
materials, where you do not have pure components. So, as long as pure components are
there; these equations turn out to be very easy. So, you can see like this, we can talk
about many many examples, where this kind of knowledge of the equilibrium
753
constant and the knowledge of (No audio from 37:07 to 37:22) our basic equation with
which we have started.
Take dG = VdP - SdT and this time let us assume that, we are doing at atmospheric
pressure. The whole reaction is happening at atmospheric pressure. Then, obviously
dP = 0; pressure is constant. So I can write, at standard state
d(ΔG0) = -ΔS0.dT (5)
And we can always write
ΔG0 = ΔH0 – TΔS0 . If that is the case, then I can reframe this equation and then say
that, ΔS0 = (ΔH0 - ΔG0)/T (6)
From equation (5) and (6)
d(ΔG0) = -[(ΔH0 - ΔG0)/T].dT
Rearranging the terms, we get,
Td(ΔG0) - ΔG0dT = -ΔH0dT
So, if you divide with T2 on both sides,
(T.d(ΔG0) - ΔG0dT)/ T2 = -ΔH0dT/T2
∂( ΔG0/T) = ΔH0. ∂(1/T)

∂( ΔG0/T)/∂(1/T) = ΔH0 (7)
754
We know ΔG0 = -RTln K.
If you put into equation (7), you will see that

∂ lnK/∂(1/T) = - (ΔH0/R)
that means, if I have the K values what is called the equilibrium coefficient values as a
function of temperature. And if I plot ln K as a function of 1/ T, I should be able to get a
straight line; where the slope is ΔH0/R, where the ΔH0 is the heat of the reaction. So,
one can find out the heat of the reaction and the K and 1/T are related by the heat of the
reaction.
This is similar to almost like a you know Kissinger plot and things like that; that we
talked about it in the last class. So, any activation energy more or less, we do the same
thing. So, you can see the equilibrium constant is related to the temperature through the
ΔH of the reaction. So, that means in principle, equilibrium constant is not constant.
Though we use the word equilibrium constant, so it is it is a fixed value for equilibrium
755
at certain temperature. But as a function of temperature, if the same reaction is occurring
at different different temperatures at the same pressure; pressure being constant.
You will be able to see that, this follows such a linear behaviour. Of course, it depends
on how accurate your measurements are to see, whether this will be a good fit, or not a
good fit depends on your actual measurements. And as long as the measurements are
right, this gives you a good fit if I do it; let say at 4 different temperatures and find out
what is K at all these 4 different temperature. How do I find out what is the K? Basically,
I find out ΔG0 at various temperatures; at different temperatures, find out the ΔG0 and
once I know the ΔG0, from that I can find out the K and plot that K with 1/T.
So, if you have 3 or 4 points in principle, you can plot a straight line and as long as the
data is reliable, you will get a good straight line. Because the equation, there is nothing
wrong in the equation. If all the points are not falling on the straight line that means, only
your ΔG0 measurements are wrong. Because if the ΔG0 measurements at various
temperatures are wrong, then the K that you get out of it will be wrong. And If that is
wrong, you will not see a good fit. That is why, whenever you talk of fit like that, in any
statistical we talk of what is called confidence level.
So, whether the R2 value of the of the fit, whenever the R 2 value of the fit is atleast more
than 95%, that is when we say the actually the fit is good. So, that again depends on how
reliable all these K values. And once you have this, one can find out this K and from
that, we will be able to find out at various temperatures, how the reactions take place. So,
your point that actually ΔH and ΔS are related to temperature, actually comes here.
Because K is related to ΔG0 and K itself is a function of temperature; that comes
basically, because ΔG is a function of temperature.
I think with this, we will stop. If you have any queries, feel free. We will again have one
more class, where we will see a few more reactions and look at, what is called an
Ellingham diagrams. Where we are actually going to talk about various stabilities of
various oxides, which oxides is more stable than other oxides. And if one oxide is more
stable than another oxide in principle, we may be able to reduce one oxide with a metal.
756
For example, if you take a very standard example everybody knows:

Fe2O3 + 2Al  Al2O3 + 2Fe. Whether this reaction will go in forward direction or not
depends on whether the ΔG0 of this reaction is negative or not at a given temperature.
what is ΔG0? It is the free energy of the products minus the free energy of the reactants.
So, if the free energy of the products is smaller than the free energy of the reactants, then
only ΔG0 will be negative and if that be the case, the products have a lower free energy
than the reactants, it is like a liquid going to a solid, it happens only when the solid has a
lower free energy than the liquid.
So, you can see such reactions happening and for that, what you need to do is basically,
find out the free energy of
4Al + 3O2  2Al2O3 (7)
and the free energy of
4Fe+ 3O2  2Fe2O3 (8)
and then, find out a combination of these two reactions by subtracting and dividing by 2
as:
Fe2O3 + 2Al  Al2O3 + 2Fe (9)
So, I have the ΔG0 of the reaction (7) and (8)

I will be able to find out the ΔG 0 of this reaction (9) at a particular temperature and at
that temperature, if it turns out to be negative, then I can say this reaction would happen.
If this turns out to be positive, then actually the reverse will happen. So,
757
whether these reactions are such reactions are feasible; if that is feasible, then what is
called, metallothermic reductions are possible. Even for that matter, if you want in a
blast furnace, Fe2O3 to be reduced by C; that reaction of
2Fe2O3 + 3C  4Fe + 3CO2; that ΔG0 of reaction should be negative.
If ΔG0 is not negative, you will not see that reaction happening. Similarly, if you want
iron oxide to be reduced by carbon monoxide to give you carbon dioxide and iron as,
2Fe2O3 + 6CO  6CO2 + 4Fe
this reaction also should have negative ΔG0.
Again here also, you can simply find out ΔG 0 for this reaction by using combination of
reactions
4Fe + 3O2  2Fe2O3 and 2CO + O2  2CO2
That is one good thing about all these reactions is that, you can simply sum them up, and
the free energy being extrinsic property, you can simply sum them up. So, you can sum
them up and get the actual values. And if the actual value becomes negative at a
particular pressure and temperature, we are already talking about at a fixed pressure.
Because we are talking of Δ G 0 and temperature at a given temperature we do this. We
will stop here.
758
THIS BOOK IS NOT FOR SALE
NOR COMMERCIAL USE
PH : (044) 2257 5905/08 nptel.ac.in swayam.gov.in

Bs Murthy Thermodynamics

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bs Murthy Thermodynamics

Uploaded by

Copyright:

Available Formats

INDEX

S. No Topic Page No.

(Refer Slide Time: 00:36)

Student: Conservation of energy

Student: Every substance has 0 entropy at 0 K.

Student: Constant pressure

Constant pressure, so we take it as dH and dS = dH/T. And enthalpy is related to what?

Student: Internal energy

how we defined and once we define H=∫ C p dT ,

(Refer Slide Time: 05:36)

Student: Increases with temperature.

Student: Hyperbolic, like S shape

Student: Super alloys.

It is ordered, of course, In fact, particularly in ionic solids if to introduce a disordered

Student: Can we induce configurational entropy by increasing the temperature as in like

Student: But because of thermal we are induce configurational entropy.

For binary system we can easily show that it is

(Refer Slide Time: 14:36)

(Refer Slide Time: 00:17)

H=∫ C p dT , where T goes from let say 0 to T.

Student: Here H298 = 0, so if we integral from 0 to 298?

(Refer Slide Time: 02:23)

Student: Why don’t we take H at 0 as 0?

(Refer Slide Time: 05:36)

Once I understand this, I can even write now as

(Refer Slide Time: 06: 36)

(Refer Slide Time: 10:07)

So, this (above equation) tells us that

Student: S increases and not decreases

Yes, when will this be fine?

Student: S remains constant.

(Refer Slide Time: 18:22)

(Refer Slide Time: 23:07)

Student: Higher free energy

Higher free energy

And, why a higher free energy?

Student: Higher enthalpy

Pearlite, what is pearlite?

Student: It is a mixture of ferrite and cementite

(Refer Slide Time: 33:39)

(Refer Slide Time: 35:39)

h c p hexagonal, it is a hexagonal structure. So, this hexagonal structure if I think of the

Student: Closed pack structure

Lecture No. #03

What is beta aluminium?

One of you talk.

Student: Chemically homogenous, physically distinct

Distinct. Correct. So, as long as it is chemically homogenous and physically it has a

(Refer Slide Time: 05:16)

Student: At 0 T, it will be on top

Top. Yes and then what about its slope.

Student: More slope

Student: Boiling point.

As the boiling point.

Student: Thermal analysis, DSC

DSC, how can it tell you what phases are there?

Student: XRD can tell

Student: Cooling curve

(Refer Slide Time: 17:46)

All of us have seen in the last class that

Student: Liquid equal to 0

Correct, the definition is

Student: The Graphite has a lower density