The theoretical background of

GTT-Technologies

The theoretical background of

FactSage

The following slides give an

abridged overview
of the major
underlying principles
of the
calculational modules
of
FactSage.
GTT-Technologies

Table of Contents

1 Gibbs energy tree and Maxwell relations

2 Properties derived from G
3 Reaction equilibria and the Law of Mass Action
4 Electro-chemical cells
5 Complex Equilibria
6 From G(T,P,ni) to Phase Diagrams
7 Choice of Axes and Types of Phase Diagrams

Note: Click on the icon to return to ToC.

GTT-Technologies

The Gibbs Energy Tree

Maxwell Phase Diagram
H, U, F µi,cp(i),H(i),S(i),ai,vi

Gibbs-Duhem
Partial Derivatives
Legendre Transform. Equilibria with Respect to
x, T or P

Minimisation

Gibbs
Energy Mathematical
Method
Mathematical methods are used to derive more Calculational
information from the Gibbs energy ( of phase(s) result derived
or whole systems ) from G

Page 1.1
GTT-Technologies

Thermodynamic potentials and

their natural variables
Variables
Gibbs energy: G = G (T, p, ni ,...)
Enthalpy: H = H (S,P, ni ,...)
Free energy: A = A (T,V, ni ,...)
Internal energy: U = U (S,V, ni ,...)

Interrelationships:
A = U − T⋅S
H = U + P⋅V
G = H − T⋅S = U + P⋅V − T⋅S

Page 1.2
GTT-Technologies

Thermodynamic potentials and

their natural variables
Maxwell-relations: S
S U VV

H A
∂H
=V
∂P PP G TT
∂G
= −S
∂T

 ∂G   ∂H   ∂U   ∂A 
and µi =   =   =   =  
 ∂ni T,P  ∂ni S,P  ∂ni S,V  ∂ni  T,V

Page 1.3
GTT-Technologies

Thermodynamic potentials and

their natural
Equilibrium condition:
G = min
dG = 0  for const. T,p,ni, ...

H = min
dH = 0  for const. S,p,ni, ...

A = min
 for const. T,U,ni, ...
dT = 0 
U = min 
dU = 0   for const. S,V,ni, ...

S = max 
 for const. U, V,ni, ...
dS= 0 

Page 1.4
GTT-Technologies

Thermodynamic properties
from the Gibbs-energy
Temperature  ∂G 
S = − 
 ∂T  p,ni
 ∂G 
H = G + TS = G − T 
 ∂T  p,ni
 ∂H   ∂ 2G 
cp =   = − T  2 
 ∂T  p,ni  ∂T  p,n i

Composition
Integral quantity: G, H, S, cp

Partial Operator Gibbs-Duhem integration

Partial quantity: µi, hi, si, cp(i)

Use of model equations permits to start at either end!

Page 2.1
GTT-Technologies

Thermodynamic properties
from the Gibbs-energy

 ∂G 
J.W. Gibbs defined the chemical potential
µ = 
of a component as: i  ∂n 
 i T,p

With G = (∑ ni )Gm (G is an extensive property!)

∂
one obtains µi = (∑ n )G
∂ni
i m

∂
= Gm + (∑ ni ) Gm
∂ni

Page 2.2
GTT-Technologies

Thermodynamic properties
from the Gibbs-energy
Transformation to mole fractions : ni → xi

∂ ∂ ∂ ∂
µ i = Gm + Gm − ∑ xi Gm 1+
∂xi
− ∑ xi
∂xi
= partial operator
∂xi ∂xi

∂ ∂
hi = Hm + Hm − ∑ xi Hm
∂xi ∂xi

si Sm Sm Sm

cpi Cpm Cpm Cpm

Page 2.3
GTT-Technologies

Gibbs energy function

for a pure substance
• G(T) (i.e. neglecting pressure terms) is calculated from the
enthalpy H(T) and the entropy S(T) using the well-known
Gibbs-Helmholtz relation: G = H − TS

T
• In this H(T) is H = H298 + ∫ cp ⋅ dT
298

T
• and S(T) is S = S298 + ∫ cp T ⋅ dT
298

• Thus for a given T-dependence of the cp-polynomial (for

example after Meyer and Kelley) one obtains for G(T):
G(T)= A + B⋅ T + C ⋅ T ⋅ lnT + D ⋅ T2 + E⋅ T3 + F T2

Page 2.4
GTT-Technologies

Gibbs energy function

for a solution
• As shown above Gm(T,x) for a solution ϕ consists
of three contributions: the reference term, the
ideal term and the excess term.
• For a simple substitutional solution (only one
lattice site with random occupation) one obtains
using the well-known Redlich-Kister-Muggianu
polynomial for the excess terms:
nij

Gmϕ (T, xi ) = ∑ xi Gio,ϕ + RT∑ xi ln xi + ∑∑ xi x j ∑ L(ijν ) (T)( xi − x j )ν

i i i j ν =0

+ ∑∑∑ xi x j xk ( xi Lijki (T) + x j Lijkj (T) + xk Lijkk (T)) /(xi + x j + xk )

i  j k

Page 2.5
GTT-Technologies

Equilibrium considerations
a) Stoichiometric reactions
Equilibrium condition:
G = min or dG = 0

Reaction : nAA + nBB + ... = nSS + nTT + ...

Generally : ∑ν iBi = 0
i

For constant T and p, i.e. dT = 0 and dp = 0,

and no other work terms:

dG = ∑ µ i dni
i

Page 3.1
GTT-Technologies

Equilibrium considerations
a) Stoichiometric reactions
For a stoichiometric reaction the changes dni are
given by the stoichiometric coefficients ni and the
change in extend of reaction dξ.

dni = ν idξ

Thus the problem becomes one-dimensional.

One obtains:
dG = ∑ µ iν idξ = 0
i

[see the following graph for an example of G = G(x) ]

Page 3.2
GTT-Technologies

Equilibrium considerations
a) Stoichiometric reactions
T = 400K

Gibbs energy G
T = 500K

T = 550K

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Extent of Reaction ξ

Gibbs Energy as a function of extent of the reaction

2NH3<=>N2 + 3H2 for various temperatures. It is assumed,
that the changes of enthalpy and entropy are constant.

Page 3.3
GTT-Technologies

Equilibrium considerations
a) Stoichiometric reactions
dG
Separation of variables results in : = ∑ν i µi = 0
dξ i

Thus the equilibrium condition

for a stoichiometric reaction is: ∆G = ∑ν i µi = 0
i

Introduction of standard potentials µi° and activities ai

yields: µi = µi + RTln ai

One obtains: ∑ i i + RT∑ (ν i ln ai ) = 0

ν
i
µ

Page 3.4
GTT-Technologies

Equilibrium considerations
a) Stoichiometric reactions
It follows the Law of Mass Action:

∆G = ∑ν i µi = −RTln ∏ aνi i

i i

where the product

K = ∏ ai νi  ∆G 
or K = exp − 
i
 RT 
is the well-known Equilibrium Constant.
The REACTION module permits a multitude of
calculations which are based on the Law of Mass Action.
Page 3.5
GTT-Technologies

Electro-chemical equilibrium

The mass balance equation is now:

να eα + ν A A + νB B + ... = νM M + νN N + ... + νβ eβ

Including the effect of the electrons being

on different electrical potential one obtains:

∑ν µ + ∑ν
i i j z j Fϕ j = 0

Page 4.1
GTT-Technologies

Separation of the electrons from the other

species leads to:
- F (ν eβ z e ϕ eβ - ν eα z e ϕ eα ) = ∑ ν i µ i

Finally one obtains:

ν e F (φeβ − φeα=
) ∑ν i µ=i ∑ i i +ν i RT ln ai )
(ν µ o

ν i µio RT
E ∑ ln ∏ aνi
Nernst‘s
=
Or: + i

νeF νeF equation

Two extreme cases can be distinguished.

Page 4.2
GTT-Technologies

A simple concentration cell

The sketch of a lamnda-probe

(for control of a car-catalyst).

Page 4.3
GTT-Technologies

The cell scheme:

α | O2 (exhaust gas) | ZrO2 | O2 (air) | β

The cell reaction:

4eα + O2 (exhaust gas) = O2 (air ) + 4eβ

The Nernst-equation:
pO2 ( air )
4 F (ϕ β − ϕα ) = µ − µ + RT ln
o
O2
o
O2
pO2 ( exh. gas )

The simple end result:

RT
E= ( ln pO2( air ) - ln pO2( exh. gas ) )
4F
Page 4.4
GTT-Technologies

The Hall-Heroult cell: Production of liquid

aluminium from solid alumina

The decisive
phase diagram:

Page 4.5
GTT-Technologies

The Hall-Heroult cell

1.5 C + 3 O 2- = 1.5 CO2 + 6 e β

The cell sketch:

1.5 C + 3 O 2- = 1.5 CO2 + 6 e β

6 eα + 2 Al3+ = 2 Al

6 eα + 2 Al3+ = 2 Al

Page 4.6
GTT-Technologies

The Hall-Heroult cell

The cell scheme:
α Al2 O 3 , C molten electrolyte Al, CO2(G) β
The two half cells:
6 eα + 2 Al3+ = 2 Al 1.5 C + 3 O 2- = 1.5 CO2 + 6 e β

The total cell reaction:

6eα + Al2O3 + 1.5C = 2 Al + 1.5CO 2(G ) + 6eβ
The Nernst-equation (full and simple):
E=
∆ G o RT
+ ln
1.52
pCO 2 a Al ∆ Go
1.5 E=
6F 6F a Al2O3 a C 6 F
Page 4.7
GTT-Technologies

Equilibrium considerations
b) Multi-component multi-phase approach
Modelling of Gibbs energy of (solution) phases

(
Gϕm = Gϕm T,nϕi ,p )
Pure Substance Gϕm = µ o,ϕ = Go,ϕ (T, p) (stoichiometric)

Solution phase Gmϕ = Gmϕ ,ref

+ Gmϕ ,id (= −T∆S ) id
m
+ Gmϕ ,ex
Choose appropriate reference state and ideal term, then check for deviations from ideality.
See Page 2.5 for the simple substitutional case.

Page 5.1
GTT-Technologies

Equilibrium considerations
b) Multi-component multi-phase approach
Complex Equilibria
Many components, many phases (solution phases),
constant T and p :
G = min

with (
G = ∑ ni µ i = ∑ ni µ io + RTln ai )
i i

 mp ϕ  ϕ
or G = ∑  ∑ nm  Gm
ϕ  i 

Page 5.2
GTT-Technologies

Equilibrium considerations
b) Multi-component multi-phase approach
Massbalance constraint

∑a n = b
i
ij i j j = 1, ... , n of components b

Lagrangeian Multipliers Mj turn out to be the

chemical potentials of the system components at
equilibrium:
G = ∑ bj Mj
j

Deviding the above equation by Σbj yields the

equation of the well know common tangent !

Page 5.3
GTT-Technologies

Equilibrium considerations
b) Multi-component multi-phase approach
Phase Components System Components
Fe N O C Ca Si Mg
aij j Gas Fe 1 0 0 0 0 0 0
N2 0 2 0 0 0 0 0
O2 0 0 2 0 0 0 0
C 0 0 0 1 0 0 0
CO 0 0 1 1 0 0 0
CO 2 0 0 2 1 0 0 0
i Ca 0 0 0 1 0 0 0
CaO 0 0 1 0 1 0 0
Si 0 0 0 0 0 1 0
SiO 0 0 1 0 0 1 0
Mg 0 0 0 0 0 0 1
Slag SiO 2 0 0 2 0 0 1 0
Fe2O 3 2 0 3 0 0 0 0
CaO 0 0 1 0 1 0 0
FeO 1 0 1 0 0 0 0
MgO 0 0 1 0 0 0 1
Liq. Fe Fe 1 0 0 0 0 0 0
N 0 1 0 0 0 0 0
O 0 0 1 0 0 0 0
C 0 0 0 1 0 0 0
Ca 0 0 0 0 1 0 0
Si 0 0 0 0 0 1 0
Mg 0 0 0 0 0 0 1

Example of a stoichiometric matrix for the gas-metal-slag system Fe-N-O-C-Ca-Si-Mg

Page 5.4
GTT-Technologies

Equilibrium considerations
Multi-component multi-phase approach

Use the EQUILIB module to execute a

multitude of calculations based on the
complex equilibrium approach outlined
above, e.g. for combustion of carbon or
gases, aqueous solutions, metal inclusions,
gas-metal-slag cases, and many others .

NOTE: The use of constraints in such calculations (such

as fixed heat balances, or the occurrence of a
predefined phase) makes this module even
more versatile.

Page 5.5
GTT-Technologies

Phase diagrams as projections

of Gibbs energy plots
Hillert has pointed out, that what is called a
phase diagram is derivable from a projection of a
so-called property diagram. The Gibbs energy as
the property is plotted along the z-axis as a
function of two other variables x and y.

From the minimum condition for the equilibrium

the phase diagram can be derived as a projection
onto the x-y-plane.

(See the following graphs for illustrations of this principle.)

Page 6.1
GTT-Technologies

Phase diagrams
as projections of Gibbs energy plots
γ
α
β µ

β γ
γ
β α+ T
β
α

T
P P

Unary system: projection from µ-T-p diagram

Page 6.2
GTT-Technologies

Phase diagrams
as projections of Gibbs energy plots

1.0

0.5

0.0
G
-0.5

-1.0

300

400

500

600 T
1.0 0.8 0.6 0.4 700
0.2 0.0
Ni
xNi Cu

Binary system: projection from G-T-x diagram, p = const.

Page 6.3
GTT-Technologies

Phase diagrams
as projections of Gibbs energy plots

Ternary system: projection from G-x1-x2 diagram,

T = const and p = const
Page 6.4
GTT-Technologies

Phase diagrams generated with

FactSage

Use the PHASE DIAGRAM module to generate a

multitude of phase diagrams for unary, binary, ternary or
even higher order systems.

NOTE: The PHASE DIAGRAM module permits the choice of

T, P, m (as RT ln a), a (as ln a), mol (x) or weight (w)
fraction as axis variables. Multi-component phase diagrams
require the use of an appropriate number of constants, e.g.
in a ternary isopleth diagram T vs x one molar ratio has to
be kept constant.

Page 6.5
GTT-Technologies

N-Component System (A-B-C-…-N)

 ∂U 
qi φi =  
 ∂qi qj
Extensive S T Corresponding
variables V -P potentials
nA µA
nB µB
⋅ ⋅
⋅ ⋅
⋅ ⋅
nN µN

dU =TdS − PdV + ∑ µi dni = ∑ φi dqi

Gibbs-Duhem: SdT + VdP + ∑ ni d µ=

i ∑ q dφ=
i i 0

Page 7.1
GTT-Technologies

Choice of Variables which always

give a True Phase Diagram
N-component system
(1) Choose n potentials: φ1, φ 2, … , φ n (n ≤ N + 1)
(2) From the non-corresponding extensive variables
(qn+1, qn+2, … ), form (N+1-n) independent ratios
(Qn+1, Qn+2, …, QN+1).

qi
Example: Qj = N+ 2
(n + 1 ≤ i ≤ N + 1)
∑q
J= n+1
j

[φ 1, φ 2, … , φ n; Qn+1, Qn+2, …, QN+1] are then the (N+1) variables

of which 2 are chosen as axes
and the remainder are held constant.

Page 7.2
GTT-Technologies

MgO-CaO Binary System

Extensive variables Chosen axes variables
and corresponding and constants
potentials
S T φ1 = T for y-axis

V -P φ2 = -P constant

nMgO µMgO q3 = nMgO

 nCaO
Q3 =
nCaO µCaO q4 = nCaO  (nMgO + nCaO)
for x-axis

Page 7.3
GTT-Technologies

Fe - Cr - S - O System
S T f1 = T (constant)

V -P f2 = -P (constant)

nO2 µO 2 φ3 = µO x-axis
2

nS2 µS 2 φ4 = µ S x-axis
2

nFe mFe
q5 = nCr 
 nCr
nCr mCr  Q5 =
q6 = nFe nFe
(constant)

Page 7.4
GTT-Technologies

Fe - Cr - C System
improper choice of axes variables
S T φ1 = T (constant)

V -P φ2 = -P (constant)

nC µC φ3 = µC → aC for x-axis
and
nCr Q4 for y-axis
nFe µFe Q4 =
( nFe + nCr + nC )
(NOT OK)
nCr
nCr µCr Q4 =
( nFe + nCr )
(OK)
dQ j
Requirement: = 0 for i ≤3
dqi

Page 7.5
GTT-Technologies

Fe - Cr - C System
improper choice of axes variables
1.0 This is NOT a true phase
0.9
M23C6
diagram.
0.8
Mole fraction of Cr

0.7 Reason: nC must NOT be

M7C3
0.6 used in formula for mole
0.5 fraction when aC is an
0.4 axis variable.
0.3

0.2
bcc NOTE: FactSage users
0.1
are safe since they are
fcc
0
not given this particular
-3 -2 -1 0 1 2 choice of axes variables.
cementite
log(ac) See next slide!

Page 7.6
GTT-Technologies

Fe - Cr - C System
proper choice of axes variables

Calculation is done
in Phase Diagram
module with
X = mole Cr/(Cr+Fe)
and y = log a(C).

Axes are swaped

in Figure module.

Page 7.7
GTT-Technologies

Only three types of 2D-phase diagrams possible

- Type 1 or Potential Diagrams:

Two potentials, Φi vs Φj.

- Type 2 or Mixed Diagrams:

One potential and one ratio of

conjugate extensive properties, Φi vs QΦj/ΣQΦk.

- Type 3 or Extensive Property Diagrams:

Two ratios of conjugate extensive properties,

QΦi/ΣQΦk vs QΦj/ΣQΦk

Page 7.8
GTT-Technologies
Interrelationship of corresponding Type 1 (b), Type 2 (a,d), Type 3 (c) phase
diagrams. (Φ4, Φ5 , ... are held constant; Φ3 is unspecified)

Pelton and Schmalzried, 1973

Page 7.9