SOLID STATE PHASE TRANSFORMATIONS2 295.00 SOLIO STATE PHASE TRANSFORMATIONS by W Raghavan 10 soley wth the pubsaher November, 2019 Foreword ships between structure and propertiesconstitutea central theme in physical metallurgy and materials science, and phase transforma tions offers one of the major methodologies through which structures can be changed and controlled. According ject of phase transformations wey and , both of which are important types of id on phenomena characteristic of phase ion of the subject continue to pr treatises, But the book at hand fils a different need, namely as a. sned specifically for teaching purposes. Cambridge, MA USA. MORRIS COHENPreface ‘This book is the outcome of the author's experience spread over a period ‘of 20 years in teaching a course on Solid State Phase Transformations to Metallurgy and Materials Science majors and to junior postgraduate students, The material was fist put together in the form of lecture notes, when the author was teaching the course in association with Professor Mortis Cohen at the Massachusetts Institute of Technology in the years 1969-71, ‘The notes have since undergone updatings as found necessary. ‘The opening chapter is brief introduction to the various types of phase transformations. In the second chapter, the principles of diffusion are dealt with, as many transformations occur through diffusional processes. The principle of the nucleation and growth kinetics are discussed in some detail in two separate chapters. The applications include continuous precipitation, particle coarsening, the pearlitic reaction and the martensitic transformations. ‘The last chapter discusses spinodal decomposition as an example of @ homo- seeepted view. The student surely acquaint himself with the Suggest of each chapter. The meaning of symbols is explained in the text, when they are introduced fist. In addition, a list of symbols is included at the end of the book for ready reference. Grateful thanks are due to Professor Mortis Cohen for making it possible for the author to teach this course at MIT for 2 years. When it became clear that he isnot going to be involved in the book-writing, Professor Cohen ‘most graciously agreed to the author doing it on his own, using the jointly- prepared lecture notes. Thanks are due to the Indian Institute of Techno~ logy, Delhi for partial nancial assistance in preparing the manuscript and the diagrams. V. RAGHAVAN Now Delhi, July 1986.COLLEGE XEROX ,ENGG COLLEGE. OPP,SALEM-11)9865324908 L Contents INTRODUCTION 1.1 Definition of a Phase Change 12 Atom Movements in Phase Transformations Long-range and short-range diffusion, 2 Diffusionless changes, 1.3 Types of Phase Transformations 1.4 Homogeneous Versus Heterogeneous Transformations 15 Buerger's Classification DIFFUSION IN SOLIDSvill 2.4 Atomic Theory of Diffusion Fick's first law fro fomic model, 22 Mechanisms of diffusion, 22 ‘Temperature dependence of D, 23 2.5 Other Diffusion Processes Diffusion in ionic erystals, 26 Diffusion along grain boundaries, 28 3. THERMODYNAMICS OF TRANSFORMATIONS. 3:1 Free Energy of Elemental Crystals 3.2 Free Energy of Solid Solutions The interaction energy P, 34 Clustering or short-range order parameter 2, 35, Free energy as of temperature, 35 = 0,2 =0,36 Tendency for ordering: P <0, Tendency for clusteriag and imi P>0,2>0,37 3.3 Thermodynamic Order of Transformations First order transformations, 38 Second order transformations, 39 3.4 Driving Force for First Order Transformations ional change, 42 ial change, 43) 3.5 Second Order Transformations » NUCLEATION KINETICS 4.1 Homogeneous Nucleation ‘The homogeneous nucleation barrier, 48 Rate of hmogeneous nucleation, 49 ‘The temperature dependence of the nucleation rate, 51 ‘The time dependence of the nucleation rate, 51 42. Homogeneous Nucleation with Compositional Change Becker's model, 53, 2 26 2 44 an a race, 55 volume change, $6 shape change, $7 coherency, 58 [Nucleation barrier with strain energy, 60 4.4 Heterogeneous Nucleation Nucleation on grain boundaries, 63 Nuc present), 65 on grain edges and corners, 66 Rate of heterogeneous nucleation, 69 GROWTH KINETICS 5.1 Interface-Controlled Growth Experimental measurements, 74 5.2 Diffusion-Controlled Growth A general expression fo Growth of spherical part Growth of small spherical particles, 78 Plane front growth, 80 OVERALL TRANSFORMATION KINETICS 6.1 Empl 6.2 Transformation Kinetics for Interface-Cont Growth6.9 Transformation Kinetics for Diffusion-Controlled Growth Kinetics of continuous pi Time exponents in ditfusion-cor 1. PARTICLE COARSENING 7A Driving Force for Coarsening ‘The Thomson-Freundlich equation, 94 Interaction between two pa 7.2 ‘The Kinetics of Coarsening (Greenwood’s Model) General features, 97 growing par The cube-root PEARLITIC TRANSFORMATIONS 8.1 Continuous Versus Discontinuous Precipitation 8.2 Experimental Charact ies 8.3 The Mechanism and the Kinetics of Nucleation ‘The Smith hypothesis, 108 The time dependence of the nucl 8.4 The Mechanism and the Kinetics of Growth Long-range diffusion of carbon, 110 The growth kine 8.5 Ioterlamell Spacing 8.6 Effect of Alloying Flements ). MASSIVE TRANSFORMATIONS 9.1 Experimental Characteristics 9.2 The Driving Force for the Massive Transformation 9.3 The Mechanism and the Kinetics of Massive Growth Growth rates from pulse heating experiments, 120 ‘The mechanism of growth, 122 ‘The temperature dependence of the growth rate, 123 contents 9 94-103, 94 ” 104-116 104 10s 108, contents 10, RECOVERY, RECRYSTALLIZATION AND GRAIN GROWTH 10.1 Recovery "8 of point imperfections, 127 Dislocation rearrangements, 128 102 Recrystallization Effect of Eflect of orientation on boundary migration, 133 The solute drag effect, 134 10.3 Grain Growth Driving force for grain grow! ‘The parabolic law of grain gi Pinning by second phase particles, 11, MARTENSITIC TRANSFORMATIONS Diffusionless nature, 140 Displacive nature, 141 ‘The habit Strain energy due to shape change, 1d growth processes, 143 behaviour, 143, 2 Classification of Martensitic Transformations 11.3 Morphological Characteristics ‘The lath martensite, 145 ‘The plate martensite, 145 ‘Transition in morphology, 147, 11.4 Kinetic Characteristics ‘The athermal Kinetics, 148 ‘The isothermal kinetics, 148 The burst transformati Factors influencing the kinetics, 149 126-139 127 129 136 140-163, 136 143 144 147xii 11.5 Models for Isothermal Transformation Kinetics ic plates, Computation of the isothermal transformation ‘curves, 152 11.6 The Dimensions of Martensitic Plates 11.7 Crystallographic Features shear, 157 Theories of martensite crystallography, 157 12. SPINODAL DECOMPOSITION 12.1 The Spinodal Curve 12.2 Free Energy of Compositional Fluctuations ‘The gradient energy, 166 Helmholtz free energy of sinusoidal fluctuations, 167 Coherency strains, 169 12,3 The Kinetics of Spinodal Decomposition Comparison with experim LIST OF SYMBOLS INDEX ‘conrents 150 155 164-173, 164 165 170 174-182 182-184 COLLEGE XEROX jENGG COLLEGE.OPP,SALEM -11)9665324905 1 Introduction ‘Understanding the relationship between the microstructure and the pro- gists, ceramists and other ‘or a thermomechanical tes of aggregation te of aggregation, if they have res. For the same composi- onic structure are also efined as the change ‘one or more other fram the above defini tion of a phase, that a phase transformation involves changes in the 1 state of agere 2 composition, yn and ‘reaction’ are often used to have wider usage, whereas the rmodynamically higher order changes such as electronic transitions,2 SOLID STATE PHASE TRANSFORMATIONS 1.2 ATOM MOVEMENTS IN PHASE TRANSFORMATIONS. ‘As we are concerned with solid state phase transformations, we shall not ‘consider changes in the state of aggregation while discussing the atom move- ments involved in transformations No atom movements take place during changes in the electronic struc- ture. Changes in composition and crystal structure in the solid state clearly Fequire the movement of atoms within the solid. The nature and the extent of such atom movements differ widely in different transformations and f under the following three categories: 1 movements over a large number of interatomic distances: 2. movements over one oF two interatomic distances: and 3° movements over a fraction of an interatomic Long-range and Short-range Diffusion The atom movements of the first two categories above are brought about by the process of diffusion in the solid state, Diffusion is the mass flow process by which atoms change their positions relative to their acighbours by “random-walk”, The phenomenon of diffusion in the solid state i discussed in Chapter 2. Changes in composition require atom movements over a large number of interatomic distances and the corresponding process is called Tong-range diffusion. ‘When a change in crystal structure takes place, the atom movements may be only over one or two interatomic distances. The atoms may be simply transferred from the parent crystal structure to the product structure across an interface which is one or two interatomic distances in thickness. ‘This process is called short-range diffusion Diffusionless Changes In category 3 above, the atoms may move only through fraction of an ucture changes. The product eryst ly when the atom movements oveut in a coordinated fashio amorphous pro duct \dom-walk diffusion referred to earlier ean be com pared coordinated movement of atoms has been called a military transformation by Christian. In the absence of interchange of atom positions by random walk, the military transformations are said to be diffusionless 1.3. TYPES OF PHASE TRANSFORMATIONS. ‘Transformations with a Change in Composition Consider the miscibility gap in the Al—Zn phase diagram, illustrate Fig. 1.1. Above 352°C, an AI-39.5 at. % Zn alloy is a single phase, x, the FCC crystal structure. On cooling below 352°C, say, to 300 NTRODUCTION Weight Percent Zine 10 0 10 couse $5606 70 75 80 85 90 9S ro sol 7 Pa e pelts i z “te 7 dl ‘tte 30— 30103060 Atemie Percent Zine The Al-Zn phase di region. The doted te lower temperate Jifferent zinc contents. to two phases ay and 2 There is no change in crystal structure, a> yt Structure: FCC FCC FCC Zoat%: 39S 2 $6 A transforming region of the « phase wi ‘a, after the transformation as shown schematically in Fig, 1.2. Clearly, requires that the atoms diffuse at least through a distance comparable to the smallest dimension of the transforming region. A typical transforming region in the solid state is 1 ym, which means that has to occur over several thousands of interatomic distances to bring. about the compositional change, Thus, long range diffusion is required here. be split into two regions #1 and Fig. 1.2. Schematic illustration ofa transformation with a composition change.4 SOLID STATE FHASE TRANSFORMATIONS ‘Transformations with a Change in Crystal Structure ‘occur in pure iron BCC crystal of iron changes to the FCC form on cooling t in to BCC on cooling through ‘composition is fixed, there is a change in the crystal structure only. coal coo! bre > oY Fe a Fe ire r Structure: BCC FCC Bcc As already pointed out range diffusion or b tomic dis these changes can be brought about by either short ‘coordinated movement of atoms over a fraction of ‘Transformations with both Composition and Crystal Structure Changes Consider the Al— Cu phase diayram shown in Fig. 1.3. At 525% 4.5wi% Cu in the form of an. solid solution (FCC) just ‘with copper. On rapid cooling to 250°C, the solid solution becomes super- saturated, as the equilibrium solubility of copper decreases {0 about Weight Percent copper Fh boaiy aaa dss Rate Tine ses —— a arene 5 eerie rue © maw Sarees 0 ease 2 Amie Fig. 1.3 The aluminiumerich end ofthe Al-Cu phase diagram intropuction 5 ion of metastable transition table solvus boundaries corres- ponding to these precipitates are shown in Fig. 1.3. Note that the solubility of copper inc s the precipitate becomes increasingly metastable in the order: # > 6° —> 6” —» GP zones, Austenite of 0.8% C decomposes on cooling through 727°C to pearlite, ture of BCC ferrite of 0.02% C and orthorhombic cementite of cool Austenite © Ferrite Cementite FCC BCC Orthorhombic different phases: BCC « (5 at. % Ni) and FCC Fe—Ni phase diagram in Fig. 1.4 Substitutional diffusion is very slow as compared to interst s0 the above phase change rarely occurs at ordinary cooling rates. the 7_at.% Ni alloy on further cooling below 600°C may simply from the FCC to the BCC structure wi such a change occurs by means of sh Massive transformation (Chapet 9). Ifthe change ours by the coordinated movements of atoms, i is The 0.8% C steel in the previous example might also transform martensite ina similar fashion on rapid quenching: austenite “> martensite Structure: FCC BCT Cw 08 086 SOLID STATE PHASE TRANSFORMATIONS INTRODUCTION 1 ‘of 769°C serves as an example of hich are randomly oriented in the paramag- in a parallel fashion on cooling through the re Fe wwe sramagnetic ferromagnetic THtth = ttttttttttttt when paramagnetic manganese oxide is cooled of 157°C, the product phase being the paramagnetic antiferromagnetic, tion: SLANT THM tt 1.4 HOMOGENEOUS VERSUS HETEROGENEOUS ‘TRANSFORMATIONS8 SOLID STATE PHASE TRANSFORMATIONS the product phases. The compositional fluctuation grows in inten to finally yield the equilibrium phases. Thi ted in Fig. 1.6(a). The 39.5 at. % Zn alloy discussed earlier may occur as & homogeneous transformation, the compositional fluctuations being brought about by long range diffusion, with corresponding changes in composition (and erystal structure) behind the advancing front. This -1.6(b). The long range diffusion Distance curs in the surround: necessary to bring about the comp. changes 0% the ing matrix, The principles of nucleation and growth are discussed in Chaps. 4 and 5. A large number of transformations discussed in this book are of the nucleation-and-growth type. LS BUERGER’S CLASSIFICATION Buerger has given ‘tions based on changes 1 Transformations of secondary coordinat Displacive rapid Reconstructive sluggish ification of solid-state transforma- in and bond type. His categories are: INTRODUCTION ’ rapid ns of first-coordination rapid sluggish ns of bond type (usually sluggish) of secondary coordin sms of low ygement of the change in the primary coordi tion ean be brought about in two ways: respect to one another the rest of the structure and so induce disorder. Likewise, interchanging joms in a random fashion can cause disorde! respond to the rotational and substitutic wnsformations requiring long range diffusion are included in Buerger’s classification, FURTHER READING ast). M4. Buerger, Phase Transformations in Solids, p. 183, John Wiley, New York (1951), EXERCISES 1.1 Give two examples of pHase tr of the transformation can result wrmations where the proper in different combinations of10 SOLID STATE PHASE TRANSFORMATIONS 1.2._List the possible differences in the nature of short-range diffusion in a polymorphic transformation versus an order-disorder transformation 1.3. What types of phase changes require long-range diffusion? 1.4 In what type of transformation is a composition change not possible? 1.5 Can a ferromagnetic to paramagnetic change be described as an order-disorder transformation? Explain COLLEGE XEROX ,ENGG COLLEGE.OPP,SALEM-11,986532490F 2 Diffusion in Solids ‘A number of phase transformations in solids are brought: about by the process of diffusion, The study of diffusion is the understanding of cules change their positions relative to their the influence of thermal energy and a gra re, electric field oF gr tion gradients only. 2.1 FICK’S LAWS OF DIFFUSION The conce chemical potential dered. Fick’ The first law of Fick states de de wpa a where Ft isthe numer of moles ofthe difsion specie pasting through « al area perpendeslar tothe Gifsion divestion ion of the specie in mol m-?, and Dis the proport factor known as the diffusivity or difusion coefficient2 SOLID STATE PHASE TRANSFORMATIONS indicates that the mass flow property that depends on ‘The negative sign on the right side of Ea, the diffusing specie, the composition of the medium in which diffusion ‘occurs and the temperature. Iti easy to show that its units are mst ‘The first lw can also be expressed in terms of flux J: 1 dn de G2 This form of Fick's law is identical to that of Fourier’s law for heat flow or Ohm's law for electrical charge flow. ‘Steady state flow x obtained wien the fx instant, the flux direction are not equal DIFFUSION IN SOLIDS B situation most frequently met in applications of the diffusion lay Consider an elemental volume of length 4x along the diffusion x and of unit cross-sectional area (perpendicular to the diffusior Fig. 2.2. The volume of such an element is dx. The rate of accumulation (or vais gas aot aS Tat is elemental volume is (Ge/é1) 4x 10 and out of the volume: Be 4 fea aa Substituting Jayas = Je + (2y/éx) dx into Eq. 2.4, we get ge. a xe-¥ as) Using Eq, 2.2 in 2.5, be E-- ee £( 03) eo is Fick’s second law for unidirectional flow under nonsteady state If D is independent of concentra de _ Be ar Pe 5. Eq. 2.6 simplifies te14 SOLID STATE PHASE TRANSFORMATIONS , Solutions of the diffe- , because of the 2:2. SOLUTIONS TO FICK'S SECOND LAW ‘The Thin Film Solution This solution uses the simplified form of the second law given by Eq. 2.7. It can be used to determine the diffusion coefficient D experimentally {A thin dise or film of material rich in the diffusing specie (concentration <2) is welded as a sandwich between two rods of thick discs containing a low (or zero) concentration ¢: of the diffusing specie. The thickness a of the sandwiched film is much smaller than the diffusion distance to be encoun- tered in the experiment. Under such conditions, the concentration-distance curves obtained after different diffusior te aset of symmetrical hin im Concrtraton ¢ Fig. 23 Concentrations (28) eta) = Fate where a=(a-aa DIFFUSION IN SOLIDS being high afer 1s Qs) es en and cmt. fhe volume of the thin film is negli he concentration-distance profiles ime are as shown in Fi different lenge "y16 SOLID STATE PHASE TRANSFORMATIONS The cross-over point is the position corresponding to @ = (cr + «a)/2. The origin for the diffusion distance is taken as the cross-over point. This is the of the diffusing specie decreases on one (the right side), ge number of thin ‘e from the origin. If thin film solution would yield: eb Dem GE oxy { HP 10) For the entire erystal, the solution is then given by eG.) - a= exe exp {5M ae @ Let (x — k)/2V Di = 7. Then, —2V/Di dy = dk. Using these relationships eG.) - a= ase - J exp (=) dn wayBi - J exp (—v) 4 | - 23 2[1-e1(S5)] am The term ‘erin Ea, 212 sands fret funtion, By definition, wf (aig) - vs | eecre a where 9 is an integration variable. In Fig. 2.5, is plotted against exp (—7?). ‘The area under the curve from 7 = 0 ton = +00 is +-/x]2; from 9 = Oto = —0 is —V7/2. So, we have the following results relating to the error function: 1 etloy= ex YEH 2 ef (-a) = Fox Yen DIFFUSION IN soLiDS ” “The hatched area is equal in Eq, 213. 3 ef ()=0 4 et (-sJm:) ~~ (27m) ‘The initial and boundary conditions of the diffusion couple set-up are: 1 Atx=0, forall, o= $8 m6 21-0 3K, ema Din 20 ema 4120 xe-8, ema $im0, wate, ose 6 t=, forallx, e=—$% =e 1e of the error function from Eq Table 2.1,8 SOLID STATE PHASE TRANSFORMATIONS TABLE 2.1 ‘The Error Funetion : ert (2) 2 0,000 0.0000 oss 025 0.0282 0.90 0.05 0.0564 09s 0.10 0.1125 10 0.8427 os 0.1680 0.8802 0.20 0.2227 0.9103 0.25 0.2763 0.9340 0.30 0.3268 0.9523 035 0.3794 0.9661 0.40 0.4284 0.4755 0.5205 0.5633 account the dependence of D on concentration. the more general form of the Fick's second law siven by Ea. de _ af pee ar = eA OB] Boltemann showed that, if¢ isa function ofa single variable such as 2 = Eq. 26 can be transformed into an ordinary homogeneous diferem: on a A de 8f pd) 4 £-S(o% 14) By introducing the initial and boundary conditions fora ifusion couple 11=0, x<0, cma, DIFFUSION IN SOLIDS 19 % z>0, em cy and 130 rede, Boao, fora fixed annealing time 1, the solution of Ba. 2. 1 (ax) ft D ~ ala’) | =e Qs) subject to the condition that 2.16) fed by choosing the origin x = Osuch that the hatched 2.6 are equal. The cross-sectional plane led not necessa integral in Eq, 2.15. From experimental data, (x) is known, so that D can be determined sraphically oncentetion € ERE REIT ° Dittsion Distance x Fig. 26 Illustration ofthe Matano-Boltzmana solution, hn aA Af‘This curve can result from 1 the B specie diffusing in one direction, or 3. both species diffusing simultaneously in directions opposite to each other. Whether we plot ca or ca against x, we end up with the same diffusivity as fa function of concent ly a property of the jon, without regard to which is the diffusing specie. This diffu- ‘compensated by a bulk flow of zone) in the opposite direction, That side of the diffusion zone, while they are being destroyed on the other side. that this bulk flow: is quite a different 1g bulk flow carries the ms Phenomenon from the diffusion provess ‘In many cases, porosity is observed on the lower melting component side, indica the bulk flow does not fully compensate for the Gifference in diffusivities of the two species. In cases where porosity is negligible, we can assume that the compensation by bulk flow is complete For such a si face as computed graphically will coincide with the i eto the ends of the couple, of the weld interface. gy of gaseous interdifusion aids in the understand- effect, Let hydrogen and argon at the same pressure in the tube separat interdiffuse. The lighter gas, hydrogen, DIFFUSION IN SOLIDS 2 bressure difference that will tend to shift the piston in the same direction, as the slower diffusing argon is moving, Darken’s Analysis Darken derived a relationship between D, Da and Ds, using the follow- ing assumptions: he molar volume of the solid solutions is independent of con: 2. porosity developed during diffusion is negli Position of the weld interface or the Matano inter- face, the net flux of atoms is zero. That (-04) + (-atp)-0 aw ‘The fluxes of A and B are due to their intrinsic diffusivities as well as to the bulk flow. Therefore, we can write (-2. $84 41) + (-Ds $2 + oa) <0 ew) where © is the vel arranging Eq. 2 ix) + Da(den/dx) 220) yy of the markers, placed at the original weld, Re- cx + ex = I/V, where V is the molar volume assumed to be independent of io, So, we have ~ pa 3X0 Da) ay 21 ¢ mole fraction of B equal to Vew and Xa = 1 ~ Xv. Since (2.22) 229) 2.202 SOLID STATE PHASE TRANSFORMATIONS De — B = (Dn ~ Da)Xn (2.25) = DaXe + DaXs 226) Which isthe relation between the interdiffusion coefficient and the intrinsic 4iffusivities derived by Darken. It can be determined from the experimental shift of the markers, noting that the shift is proportional to the square root b ined from the Matano-Boltzmann analysis. Using 14 Dp can be calculated. Eqs. 2.21 and 2.26, 24 ATOMIC THEORY OF DIFFUSION Fick's First Law from the Atomic Model backward direction), Sinz = dnt (2.27) aie em ons = Wt — ex a= ie am) Comparing Eq, 2.30 with Fick's first law, we see that D= By ean DIFFUSION IN SOLIDS ase by several orders of magnitude at lower temperatures. In the i ley mechanism, Fig. 2:7(@), an atom from a regular site into an interstitial site and displaces another atom from a regular site ‘equal to exp(—4G//RT), where 4G; is the free energy of formation of a mole of vacancies. The number of successful jumpDIFFUSION IN SOLIDS 2s TABLE 22 Approximate Do and Q Values for Some Diffusion Processes 24 SOLID STATE PHASE TRANSFORMATIONS attempts by an atom is given by v exp (—4Gn/RT), where ¥ is the lattice vibration frequency and 4Gw is the free energy maximum (per mole) along the path tothe vacant site, 4G is called the free energy of main of 8 vacancy. The frequency wth which an atom exchanges positon with ay neeniee De @ ofthe neighbouring sits ib thea gtea by lot mts! 15 mot" /= zy 4Gn) 4G) Cu in Cu 0.20 1 v = zvexp (~462) exp (-4% ax) cuincu 920 96 where Z is the coordination number. Substituting Eq. 2.33 into Eq. 2.32, we a ees i” nie a al GeinGe 93 28 1 Sn + 4G) sins s400 ” p= fez em [- (2 3729)] 39 aw s a Cin graphite 7 en en eee Cin ga 7 oy “ a Nin Fe 005 * yuna 29) Cin Fe @) 22 i CinFeO) 02 te where oy i the latte parameter of the cubic crystals In such cass, cae cs pe therefore, Moin Fe 40 5os 3G + 46) NiinFe 26 25 p= cbvexp[- (-2%)] a0 znincy 07 17 Niin Cu 20 20 Using 4G = 411 — TAS, we can write Ea, 2.6 as onal tas ial roan (2524-45 ttn | AH) p= dian (S£45) en [- (28555 4)] aay Experimental data show that - 2.) p= Deep (—-& as) Values of Do (called the frequency factor) and @ (called ‘or diffusion) are obtained from measurements of Dat different for example, by using a diffusion couple yields a straight line. The slope of the ‘cept of the y-axis is In Do. Comparing Eqs. 2.37 and 2.38 is equal to —Q/R and the inter- seen that Q = 4m + SHy and bom abexp (52-49) cw “Approximate values of Do and @ for some diffusion processes are listed in Table effect observed in solid solutions, where the two different species se at different rates%6 SOLID STATE PHASE TRANSFORMATIONS TABLE 22 444m for Different Dlfesion Mechanisms in Copper a 4He Mechanism KS mot! 0 Inte #0 Vacancy 120 ring mechani 370 ble 2.4 compare Of Q for self di fer and gold. TABLE 24 Calculated and Experimental Aci by the Vacancy Mechani In Energies for Diffusion Element any Aim 4H; + Sm 2 Silver 97 80 i” 18 2.5 OTHER DIFFUSION PROCESSES Diffusion in Tonie Crs A pair of one cation vacancy and one ai defect. & pair of a vacancy and a ca defect. When Schottky defects dominate in an ionic erystal, the cation vacancy s the diffusional flux. When Frenkel defects are abundant, the carries the flux. nto these defects in thermal equilibrium, ionic erystals may have defects generated by impurities. If a Ca cation’ replaces two Nat [Nat vacancies carry the flux here. A fraction of produce cation vacancies far in excess of can also be generated by deviations from the diffusional flux. carried by catio ichiometric compounds such as Zn.O, where x > 1 have a larger Iniron oxide (FeO), the Schottky defects dominate and the flux is, carried by cation vacancies. Off-stoichiometric compounds, such as Fe,O, where » <1, have a larger number of cation vacancies than the stoichio ‘metric compound, Fig. 2.8(b). ano es Zl Fat Re OP pet OF ant Reno ot et ol et? 2g an ot ant Po Go eet See oF ag oh ad gh alto BOF zn zal OP Qt GP re? * b ) ZnO and (b) FeO, owing to ichiometry. Vacant cation FeO are indicated by Cl Measurements of the electrical conduct diffusivity from the Nerns ty of ionic crystals yield iT (2.42) where is the electrical cond the number of diffu the diffusing defect, lefects per unit volume, the electronic charge. ture 2.9 is a plot of conductivity versus 1/T for a NaCl erystal doped ‘arying atomic percents of CdCl. At temperature above 650°C, the ion (Na*) vacancies dominate and the conductivities ped NaCl crystals are the same. At lower temperatures,2s SOLID STATE PHASE TRANSFORMATIONS. Tempersire1Pc | ion of the degree of doping. At any one tempera- west conductivity. With high y lattice diffusion. The cross-sectional a ‘occur (see Eq. 2.1) is usually much for lattice diffusion. “The diffusion coefficients for self-diffusion determined for and for polyerystalline silver are shown in Fig. 2.10 as a function of DIFFUSION IN SOLIDS » Temerature "6 900100700600 30046000350 P.G, Shewmon, Diffuse C. Matano. Japan, J mer. Inst. Min. Met. Engrs. Met. Engrs., 174, 184 (1948). 315 (1942). EXERCISES of Fick's second law for the thin-film case is Fan °? (ani) where D is not a function of the conece the diffusing. specie per mv jon, a is the number of moles of form of a very thin layer20 SOLID STATE PHASE TRANSFORMATIONS at the junction of a sandwich type diffusion couple and cr is the concentra of the diffusing specie initially present outside the thin layer. 1 Show that when f > 0,¢-» c1, a8 x» + © or asx — 2 Prove by differen of the Fick's second lav a solution 3. Prove that se Terme otal quantity of moles 1 surface is 2.2 Show that, in a carburization experiment of carbon diffused in time # per unit area of the [rr 0 JB where cris the constant surface concentration of carbon and cr is the initial carbon concentration in the steel 23. A rod of pure copper was joined toa rod of a Cu-29.4 at.%Zn alloy. After annealing for 360 nr, the % Zn versus distance data came out 19 be as follows: at.% Zn x mm a1% Zn x, mm 03 5.005 235 3.655 1s 4315 250 3.405 4a 4.645 26.5 3.075 88 4.495 29 2315, 147 4315 28.8 1.895 206 3.965 29.1 1.495 Distances were measured from a fixed point outside jon zone. the position of the Matano interface and calculate D(c) at 5, 15 sion experiment analyzed by the Matano fare obtained for the metal atthe cross-section Diffusion time 1 = 200 hes Marker movement $= 144mm Interdiffusion coefficient Ba 10 mst DIFFUSION IN SOLIDS 3 ‘Slope of the penetration curve at the markers ‘Atom fraction of A Compute Dy and opera 1000C, cleat the eulibsium actions of ana .cancies, using Hunting values for 4Hy nd ian) = 12043 mol heated to 380°C during heat eatment and ic aimed after the gues, the difason us found to be 26 An Al4% Cu quenched to room tempe rate of copper (which proceeds by a vacancy mechs chan what would be expected from the wuenching? The enthalpy of motion of vacancy in idfor thesame amount of i) the grain boundary the grain boundaries Do = 0.710% m? s-* Q = 188 KF mot For grain boundary diffusion: Dy = 0.09 10-* m? s*# = 90k mol" For lattice diffusion: 2.8 What do you understand by intrinsic and extrinsic regions of diffu alent cation impurity? fons can be Explain how the ent conductivity from the experimental determination of el versus temperature, rarnmrerann % 4 >» % Ds adhCOLLEGE xEROX-ENGG COUFE OF SALE SES. 3 Thermodynamics of Transformations tion ean occur spontaneously only when the free energy ied for a transformation, we need to know the free t phases. The free energies of eleme here as a function of compos ‘The Gibbs free energy G of the eryst G=H-TS where H and Sate the enthalpy and the absolute temperature (in kelvin). The en wo met fora oa ‘where Ho isthe residual enthalpy at O K and cy is the heat capacity at cons- Tr peerure. The entropy ofthe eral is given by oft 3) se | ‘4 "THERMODYNAMICS OF TRANSFORMATIONS. 3 Combining Eqs. 3.1 through 3.3, we obtain at constant pressure (%),--s oa) Integrating Eq. 3.4 within appropriate limits, we can write [eo- [ser 6s) o=m- [ser =m | (fae )er eo ‘From the thermodynamic laws, we ean show that Ge)-¥ on Gibbs free energy n with pressure ata constant temperature has a posi in Fig. 3.10). He 3 3| 4 ql a |4 soup start SE TRANSFORMATIONS ‘The above description of the free energy of elemental crystals also applies to chemical compounds with ideal stoichiometry, for which the configura- tional entropy is zero, 3.2. FREE ENERGY OF SOLID SOLUTIONS. Here, the composition is a variable. So, the contribution to the free ‘energy from the configurational entropy cannot be ignored. ‘The Internction Energy respectively. When B-B bonds are broken to create new A-B bonds: AA AA >to BoB BoB ‘The interaction energy P is defined as the energy change that occurs when ACA and B-B bonds are broken to produce one A-B bond. As above, when one A-A and one B-B bond are broken, two A-B bonds are produced. So, we have = ras — (ME) on ire intraon ene see, ve ons (AB bn) te ee getically favoured. If 7 has a large negative value, compound formation is magnitude: - Veo — Ysisict Vow P= Vso + = = 369 — 1(-176 ~ 139) = -211.5 KI mol ira jon energy fe bonds (A-A and B-B bonds) are value, the two component energetically favoured. THERMODYNAMICS OF TRANSFORMATIONS 35 the tendency will be'for clustering of like atoms in the solid solution. Clus- tering refers to the formation of small regions in the solid solution, where ACA or B-B bonds are preferentially present. (Clustering or Short-range Order Parameter = Let Paia be the probability that a B atom will have an A atom as the nearest neighbour on a specified atomic site. Ina truly random solid sol there is no preference for any type of bonds. So, we have here =m 6% jon of A in the solution. Where the tendency is the B atom will prefer to have A neighbours. So, >is G.10) ‘Where there is a tendency for clustering, the B atom will prefer B neighbours, 4s B-B bonds lower the energy. So, here Pray < Xn eu) ‘The local ordering (or clustering) parameter « is defined as @.12) = Xa, « = 0, we have a truly random solution, > Xa, a is negative, the tendency is for ordering, Naw = PaZNoXe G13) where Z i the coordination number (number of nearestneighbours of an atom), ‘No is Avogadro’s number, and Xq is the mole fraction of B. Here, PaaZ gives the number of A atoms on the Z sites surrounding a B atom. NoXe is the number of B atoms in one mole of the solid solution. from Eq. 3.12, we obtain Nan = ZNoXaNa(1 — a) G1 Free Energy as a Function of Temperature At O K, the bond energy Ao will be the sum of the energies of the three36 SOLID STATE PHASE TRANSFORMATIONS ‘types of bonds in the solid solution: Ho = NanVan + NosVan + NaaVan where Naa, Nox and Nasare the number of A-/ order to te Ho, we can consider the bond energies of pure A and pure B before mixing and the energy change that occurs on mixing Ho = AZNoXAVan + 4ZNOXaVan + ZNOXAXo( — 2) P ‘energy before mixing ~~ creray of mining yp + 2XANa(l — «)P) G16) se8 on mixing, 3 is truly random: Se = =R(Wa In X4 +Na In Xo) ‘This assumption is an approx! for short range ordering or clus ering in the solid ‘based on this assumption is known as the regular solution mode! ‘The Gibbs free energy G of the solid solution at temperature Tis then iven by: G =H MSserms + Seon = HZNANAV an + Xal The configurat (on the assumption at the sol tn) a + 2XaNa(l ~ 9)P) +] ser- 7 SAT RTA IN Xa+ Xaln Xo) G.18) where ci is the heat capacity of the solid solution. We can assume that the heat capacity of the solid solution varies according to the Newmann-Kopp's valle: cpa Xa + cnXn G19 ‘A straight line relationship results, when cj is plotted as a function of ‘composition. "The entropy of mixing Se and —TS, at constant temperature vary as shown in Fig, 3.2. The —7S. curve has a U-shape. As the composition ‘approaches that of one of the pure components, the magnitude of the slope of the — TS. curve tends to infinity. Téeal Solid Solution: P = 0, 2 = 0 the term containing the interaction energy P is zero for this ‘case, Also, at 0 K, the terms containing the configurational entropy and the heat capacity are zero. So, the Gibbs free energy G, as given by Eq. 3.18, THERMODYNAMICS OF TRANSFORMATIONS 37 4 T 14 (0) — TS. composition ina binary Py of composition as shown in Fig. 3.2. When bond energy terms and the heat ‘capacity terms in Eq, 3.18, a U-curve still results as shown in Fig. 3.3. As the temperature increases, the curve becomes more and more U-shaped, increasing dominance of —TS., Fig. 3.3 th ‘Tendency for Ordering: P <0, x < 0 AtT = 0K, by virtue of the term containing the negative interaction energy P in Eq. 3.18, the free-energy composition curve has shallow U= shape, Fig. 3.4. This becomes more and more U-shaped, as the temperature increases and the configurational entropy term —TS. becomes increasingly dominant, Fig. 3.4 For such a binary system, at very high temperatures, disorder pre so that the configurational entropy term is maximized. On lowering temperature, the tendency for short-range ordering increases, Event ‘long-range order sets in and an ordered phase becomes stable ‘Tendency for Clustering und tai >o2>0 AAs the interaction energy P in Ey. 3.18 is positive in this ease, the freeFy SOLID STATE PHASE TRANSFORMATIONS Ginbe Fre nergy -U (A) shape at OK, Fig. 3.5. At T > OK, the config fopy tere has & U-shape: the ad of these two contributions to the free energy results in a curve with two minima and one maximum, Fig. 3.5. Asthe temperature increases, the ~7Se term in Eq, 3.18 dominates and the free-energy composition curve takes on 48 Usshape, see Fig. 3.5. For such a binary and the tendeney for cl he temperature is lowered. Eventual Separation of two phases, one poor in B component and the other rich in B component, occ 3.3 THERMODYNAMIC ORDER OF TRANSFORMATIONS, ‘Transformations can be classified as frst order, second order, etc. depend- ing on how the Gibbs free energy G changes with the external parameters of temperature and pressure. First Order Transformations Here, the Gibbs free energy changes continuously through a phase trans- "THERMODYNAMICS OF TRANSFORMATIONS 39 ium temperature,To and pressure Po as illustrated in Fig. 3.6 for an « —> 8 transfor the free energy curve ‘changes stope abruptly at To or Po. That ivatives of G with respect to temperature and pressure have From Eqs. 3.4 and 37. we note acy _ (8,--s and ea Fe), md So, first order transformations are associated with a volume change and an entropy change. ‘Second Order Transformations40 SOLID STATE PHASE TRANSFORMATIONS Gibbs Free Energy o% 30 ase Gate Free EnerayG Gone Frenne Te sg. 36 ibbs fee eergyG of te «and f phases asa function of) tm Bs [perature at constant pressure, and (b) pressure at constant {temperature "THERMODYNAMICS OF TRANSFORMATIONS 4 temperature Tz and pressure P., but the second derivatives have disconti- puities. Through the relationsbips es) -- 8), om ($8)- @, om iscontinuities in the second derivatives represent abrupt changes in heat ca eat capacity, smal Fig. 3.7. Heat capacity of p-brass as function of temperature through the order-dsorde temperature (477°O)a SOLID STATE PHASE TRANSFORMATIONS 3.4 DRIVING FORCE FOR FIRST ORDER TRANSFORMATIONS. iccompanied by & is the free energy per sponding free enerey wus transformation at constant duct phase B, and ge then during a sponta temperature and pr fe g2 = dg <0 (23) The negative of the chemical free energy change (~4g) is regarded as the driving force for the phase change. Even when the energy of a BCC steve lcture, as the temperature increases, the two crystal modifi HCP phase is stable; at higher temperatures, the BCC phase is stable, At 882°C, guce ~ gace = dg = 0, below 882°C, gucr — gece < 0, and above 882°C, gucr ~ guce > 0. ‘Other examples of such phase changes are those that occur ‘metals. For example, sodium which is BCC at room temperature transforms “THERMODYNAMICS OF TRANSFORMATIONS. a to HCP structure at —237°C (36K). However, this situation may not always be evident, if there are other contributions to the free energy. For example, the stable form of iron at low temperatures is the BCC form, as the free energy equations are here mi ‘The application of hydrostatic pressure at con: the enthalpy of a phase, since H = E+ PY. This increas free energy of the existing phase sufi Wn from an open to a close packed structur BCC iron (2) at room temperature transforms to a close packed HCP ‘under a hydrostatic pressure of about 12 500 MPa at room per “The differences in the free energies of elements in different crystal forms ‘are called the lattice stability parameters, These data for a number of elements hhave been computed and compiled by Kaufman. ‘Transformations with Compositional Change Consider a binary system of A and B. The free energies of two phases = in Fig. 3.8 plotted against the sperature and pressure. In the -gment ab represents the chemical free energy change per mole transformation without a change in composition, Ginbs Fee Enarsy 6aa SOLID STATE PHASE TRANSFORMATIONS Now consider the formation of a s1 of composition 2. a small change shown that the formed is ange in chemical free energy 4G per mole of the B phase 46 = 04 ~ G2 (a~ 9 4 6.29 the segment mn energy curve Fig. 3.8, drawn the matrix composi e, the segment mn must be be free energy curve with the tangent fap in a binary system, the free-energy/ us. between 3.24 is easily modified to be ap} of a typical second ord: reference to 0 the associated exchange interact aan increasing number of electrons set progress ization of the electron spins offsets the increase in H so as to minimize the free energy at any given temperature, These el ygements are so rapid that the system is essentially in equilibrium on Jing that temperature, “The disordering of the spins isan example of a second order transforma- tion and the cooperative phenomena involved, At low temperatures, when all spins are parallel, a large increase in H is needed to reverse the spin of ‘ne electron against the combined restraints of all its neighbours. But as more and more electrons reverse their spins with increasing temperature, it becomes easier for others to do the same, since those already with reversed Spins ty to poll others into their orientation. When sufficient numbers of Spins point in both directions, the change in H with further disordering Becomes negligible and general disorder sets in: this occurs at the transition temperature or the Cutie poiat 7s. The cooperative action of the ele "THERMODYNAMICS OF TRANSFORMATIONS 4s oat sermeliedSotuation Mapretzotin 47h (0) of oe Temperature 1 Fig. 3.9 Saturation magn of nickel a8 function of temperature, spins causes the saturation magnetization curve in Fig. 3.9 to drop off preci- pitously as the temperature approaches ond order vansformations occur over a range of temperature under equilibrium conditions. "The anos excess hea capacity sores pondingly spread over temperature range edn Fg 31. Ths Tehaviour is in contrast fo theca oa fst ode ransfooation, where an entropy change occirs athe transformation temperature and the Xs heat apaity becomes infite at that temperature, FURTHER READING “omputer Calculations of Phase Die~ ‘grams, Academic Press, New York (1970),46 SOLID STATE PHASE TRANSFORMATIONS EXERCISES thermodynamic laws, prove the following relation- Pam _ Pais) To Xe ‘where the terms have the same meaning as in the text. 3.3. Derive the f a mole of a binary s ZN0Xa Ne, + Xee) whereas for a mechanical mixture of pure A and pure B of the same overall composition ZN Naa = 35 Show jum in a binar corresponds mixture of phases of 3.6 Prove the rel the free energy of a phase is higher, COLLEGE XEROX JENGG COLLEGE.OPP SALEM -11)9865324908 4 Nucleation Kinetics we referred to the category of transformations rowth type. In nucleation, tiny volumes of the product phase al wn 10 the product phase, form first. These subsequen ‘The nucleation process oceurring by thermal fluctuations is also bryo of another phase forming in the iquid phase, form in the supersaturated parent (vapot ‘small droplets, which may contain of the order of 100 atoms, to have the same surface and thermodynamic properties as the bulk liquid. ‘The assumptions of the Volmer-Weber theory are open to ques48 SOLID STATE PHASE TRANSFORMATIONS face will vary significantly arbitrary. Moreover, the surface envrgy of the for small particles of ‘The Homogencous Nucleation Barrier If AG jis the free energy change accompanying the formation of a new Particle of the transformation product, we can write: 4G = Vdg + Sue a where V and Si denote the volume and the surface area of the particle, 4g the chemical free energy change per unit volume of the product phase. 0 is the specific surface energy of the interface separating the product and the parent phases. When there is supersaturation in the parent phase, dg is nega- tive. The surface energy term in Eq. 4.1 is, however, positive. In the a Of strains, the particle tends to have a spherical shape to minimize i area, For a spherical particle of radius r, ag = Sandy + Ane aa) ‘This function passes through a maximum wit negative. as the new phase particle starts to for system increases, as the surface energy term jon with r of the surface energy term and the exactly balance each other. TI term becomes dominant. As th dependence of 4G is also shown ‘at the 4G-r curve increases cot ym. ALT < To, dg particular tem tritical size are called nuclei. The particle ‘Cal nucleus. As 4g becomes more negative with a lowering of the temperat the eritical condition occurs at smaller values of 4G* and NUCLEATION KINETICS L 0 Tetpheteh Free erry change BG ‘As data on the heat of reaction (the enthalpy change during the transfor~ mation) are more readily available than on the Gibbs free energy change, the anar To as) where 7 is the transformation temperature and 4h is the enthalpy change per unit volume of the product formed, taken to be independent of the tempera- ture of the teansformation. 47 is the degree of undercooling or supercooling. Combining Eas. 44 and 4.5, we obtain ao Equation 4.6 indicates that the nucleation barrier varies inversely as the square of the degree of supercooling, Rate of Homogeneous Nucleation In the’lassical theory, the process of nucleation is defined as he addition (of one atom to a0 SOLID STATE PHASE TRANSFORMATIONS Let there be s* atoms in the parent phase facing the across the interface, Fig. 4.2 phase to the product phas to have nucleated. Any fi ced to be the subsequent 8 particle any one of these jumps from the parent and is said isconsider- ess of growth. -s N* is approximate we = wre (— 42 (4s) where » is the lattice vibration frequency (which is ~ Debye temperature) and 469 interface. Then the rate of he so above the n across the the number NUCLEATION KINETICS a Tene = Wrst exp[— 4" 460) as [Asa first approximation, in homogeneous nucleation, Nr can be taken to be jume (~ 10° m7). Then the reaction 8 me by increasing the driving force temperature, J can be increased by a factor of 10% i. ‘The Temperature Dependence of the Nucleation Rate ready noted in Eq, 4.6, 4G* is a function of temperature. Its value im tenaperature To ite. Therefore, the rate [as given by Eq, 49 is zero at the equilibrium temperature. IT > 0 and AG* isfinite. As AG* decreases with es. At some degree of supercooling, reaches e decrease in the thermal energy kT’ domi- the nucleation rate decreases with decreasing murs can be found jing. expression for 446% _ (aG* + 4G») ar Tins G10) Figure 4 es a graphical method of determination Of Tia,, When AG? is known as a function of temperature. ‘The Time Dependence of the Nucleation Rate ‘quenching a mater temperature Ty. Ifa SOLID STATE PHASE TRANSFORMATIONS Temperature tion may be complete, even before the steady state nucleation tained, Nucleation Rote J a 1 Tie Fig. 44. The time dependence ofthe nul NUCLEATION KINETICS 33 4.2, HOMOGENEOUS NUCLEATION WITH COMPOSITIONAL CHANGE Becker's Model Homogeneous nucleation with a compositional change in systems having ‘wo or more components has been considered by several authors, Becker used an expresion similar to Eq, 324 forthe chemial fee ebange during When there is an attendant compositional change. The sumed to have a unifgym and fixed composition lattices ofthe parent aching at the inter Giei) adC othe abrupt compositional change ted the bai ofthe nearest neighbour bond the diference in ues ofthe chemical corresponding to that of the stable phase. and the product phases are assumed ta of the tangent at 2 with the free energy curve. The nucleation barrier disappears at and beyond this point, Hobstetter’s Model +r corresponds to the saddle point on the free energy surface plot n and radius as the two axes. There is a critical comp jel radius r* corresponding to the saddle point Even though the Hobstetter model is an improvement over the earliers4 SOLID STATE PHASE TRANSFOR °“1ONS Give Free Energy 6 nergy. In martensitic transformations, besides any change in volume, insforming region also undergoes a change of shape, resulting in nergy. So, the types of possi= 3t be understood, before estimates of the steain energy for different interfaces can be made, NUCLEATION KINETICS. 58 ‘The Structure of the Interface “The structure of the interface between the parent and the product phases letermined by the degree of matching between the two crystals across the faces ate clasifed into three groups depending on the degree tures of the parent and the of the product phase will be greatly facilitated cerfacial energy, recall from Eq. 4.4 that the nuclea- to the cube of th~ interfacial energy. smatch between the parent and the product phases is appre- associated elastic strains due to coherency may be excessively large. regions of good coherency in the of severe distortion compris facial dislocations, trated tn Fig. 460). Her, along the interface, the spacing between56 SOLID STATE PHASE TRANSFORMATIONS seven atomic planes in the « phase is seen to be equal tothe spacing between ‘eight planes in the B phase. Accordingly, every eighth plane in the P phase forms an edge disloc sper, the corresponding \- Such an interface is of the elastic and core ie energy of an energy of a semicoberent interface les in the range of es are very different from each other, rela- distorted incoherent interface The energy of this interface 1000 erglem). Strain Energy due to Volume Change Nabarro considered the uniform expansion or contraction which may ‘occur during the formation of a new phase having an incoherent with the matrix. Assuming that all strains are stored in the matrix, Nal showed that, for oblate a the strain energy « per uni = 2,,(42 4 (2 «= Fen FY'4(6) the shear modulus of the matrix (akea to be el AV ia the fractional volume. change, accompa transformation, and 4c) ia function that dependson the partie iMustrated in Fig. 47 face { ' for(£) = 1 (phere) 4(é)- | 075 for(£) > 1608) Ude(S) = () < 1aie «= 4u9(2) 9) in energy factor due to dil ins. For a dise «has a value of where Aas is the with (lr) = (~15 MI m? for typical sol [NUCLEATION KINETICS 7 ny ‘Strain Energy due fo Shape Change isher, Hollomon and Turnbull have estimated the shear strain energy from a change of shape of a transforming Such a change The shear sein within the sphere is tanger aay) Substituting for 8 from Eg, 4.14 into 4.15, we have tang ang 6.16) “The shear strain energy per particle stored inthe spherical volume i given by a bee tment.“TRANSFORMATIONS NUCLEATION KIC “The actual misfit during the nucleation ev mt ie ba eae as tn ‘Then the coherency strain in the matrix i nerky for diflerensuations. His anabyts + batch on ual step: the region which is to be transformed into the product 3 Apply surface tractions (forces) to restore the segion to is originat form layer of body force at the interface to annul the traction ‘This force will strain the matsix and the particle. 4 per unit volume of the prodct T+ vn) = 200GN=*NUCLEATION KINETICS 6 1. B2ndto? 3 high degree of coherency,Cy SOLID STATE PHASE TRANSFORMATIONS E inasmuch the coh ix, We shall review here for the process taking. Nucleation on Grain Boundaries ‘Clem and Fisher have developed a model for the nucleation kinetics on 18 assumptions: hips between the nucleating particle L_ Any special orientatios ‘and the matrix are neglected coheres Fig. 4.10. fut 4.4 HETEROGENEOUS NUCLEATION ‘As alteady indicated, nucleation in the sol quite often hetero- geneous or nonrandom. Amo: fe 1s are grain Kovndaries, grain edges and grain comers, dislocations and surfaces of Note that the surface tension ha he same dimensions (Nm!) as surface energy (J m-2), as J = Nm. ple Let R be the radius of curvature of the spherical caps and c the semi-os SOLID STATE PHASE TRANSFORMATIONS ‘thickness of the double lens. ris th following results hold good. ius of the circle of intersection, The r= Rsind 439) c= R(l ~ con 8) as) ‘The volume of the particle V can be shown to be V = Zama —3 0086 + cos?) 4.36) and the surface area of the particle Sis Sam 4 R21 — c08 6) aan ‘The free energy change during this type of nonrandom nucleation should now take into account the energy gained in the elimination of grain boun- dary area equal to rr: Amu = 2 2B 300804 co8 Og + ARAL — 08 ous — me rm Rosin 8 and cus = eas cos 8 in Eq. 4.38, we obtain (438) dora = (ZRdg + 2eRey)Q—Son 0 +e) — 3” ting d(4Gra)/AR = indicated by superser the following critical values of the parameters are derived: om 20a pen 2 (340) oa dewsin€ 7 a aa Broly , Gas = $8 2 — e080 + cos (442) = 14Ghomo(2 — 3 cos 8 + cos? #) (4.43) the last equality arising out of substituting from Eq. 4.4 for homogeneous nucleation. — tion of the ratio AGi/4G homo as & function ‘The lens here degenerates into a the interface without any nucleation (40 and 4.42, that (4.44) NUCLEATION KINETICS 65 Jn Energy Present) account during grain boundary the surface tension forces ate in fe the particle to be an oblate ta — thon std 448) the saddle point of the free energy surface are! ct = — eas = fous) (4.46) ag pe = Hea hea) an A6tg = 22h = A 4.48) saa sans66 SOLID STATE PHASE TRANSFORMATIONS om 0, Ba, 448 becomes Eq. 4.31, previously derived for the homo: geneous case. The special cases are: . : " ages and Corners ines along which three neighbouring grains meet. A vce spherical surfaces: o [NUCLEATION KINETICS for nucleation on grain edges and grain corners (alongwith that for grain boundary nucleation) ‘Nucleation on Inclusions surface of a foreign particle isa spherical eap or a hall 1 = 14Ghomo (2 — 3 cos # + cos? 8) 449) Note that the free energy of nucleation here is one-half of that of the Tens ease given in Eq. 4:43 “The following special cases may be noted 1 6 = 180°, cos @ = —I, AGhe: = AGhowe. The product ‘makes only point contact with the Wn, which plays no role in the nucleation process. 2 8 = 0%, cos @ = 1, 4Ghu = 0. The produ le completely ‘wets the inclusion surface and spreads out as a vanishingly thin film. Here, there is no barrier to nucleation. _e 108 9 = 0, AGher = $4Ghomo. The product partiel hemispherical in shape. ‘Nucleation on Dislocations Cah has described the nucleation process on dislocations based on the following assumptions: 1 The product le is cylindrical in shape extending along a 2. The strain energy of the dislocation assisting the nucleation process is taken to be removed over a distance from the dislocation equal to the radius r of the cylindrical particle. 3 An incoherent interface is formed 4 The strain energy due to the particle formation can be neglected. ‘Then, per unit length of the dislocat nrdg + 2nre — Ain (4.50) In r approximates the strain encegy of the dislocatién per unit 1 we have 2erdg + 2n0 — A'Ir = 0 Which is a quadratic equation in r forthe ert a= —gh(1 + [Tea asp~ VT wer os SOLID STATE PHASE TRANSFORMATIONS [NUCLEATION KINETICS o When (1+ 24°4g)/20? is positive, two real roots exist, From the sign ofthe second derivative, d%(4G)/dr2, we note supersaturation of vacancies energy due to volume increas possible by quenching from The extra driving force i+ 246 2) cane) may 1 order So 46m) = ~ 35-( 1 — | Bw) am Se ttey sation ny be of he ett oe Alfor =-= 1+ 24% ae) 1¢ transformati an under- ier td == ap(1+ fa) tan vr muiics prs nae se hereafter, a barrier maximum the 4G versus r curve, Fig. 4.15. of Heterogeneous Nucleation ‘The rate of heterogeneous nucleation can be expressed in a form similar it of Eq. 4.9. In addition to the difference in the 4G* term as dis- sed above, the pre-exponential term will include only the number of preferred 1 3m sites, Which is many orders of magnitude smaller than the number of atoms per unit volume used in the homogeneous case. In order of magnitude, the number of nucleation sites for various situations is occurs at r= ri > w2x's/tter?s0 ibution of sites of varying potencies (the greater the potency, the lower is the corresponding 4G*), If Ni is the Nucleation energy BG bs ch kT 6) 1 own (GE) <1 68) and 460 = exp (482 ume (-4) =P 69) Thus the growth rate increases approximately linearly small degrees of supercooling. However, as the supercool Ds term, which depends exponentially on temperature, decrea ‘The growth rate after reaching a maximum, decreases exponentially wi temperature, as given by Eq. 5.9. Experimental Measurements ‘The measurement of grow nucleation rates. The size particle is measured meta transformed specimens. These measureme samples, where the fraction transformed is the growing particles on one another gible up to this the transformation. The radius of the largest sphere seen on the plane of observation ean be taken to be equal to the true radius of the three: dimensional particle. A plot of the maximum size against transforms time yields a straight line for interface-controlled transformations. Figure 5.2 shows the temperature dependence of the linear growth rate for the transformation of white to gray tin (Ty = +13°C). The maximum ‘the growth rate (~1 mrn/hr) occurs at ~32°C, where ST = 48°C. There- fier, the growth rate decreases exponentially with temperature as the boundary diffusivity Ds decreases, as given by Eq. 5.9 5.2 DIFFUSION-CONTROLLED GROWTH Diffusion-controlled growth refers to the growth process, where there is GROWTH KINETICS Wy change in composition and the rate is controlled by the long-range diffu- sion of one of the component atoms through the matrix phase. L 1 Loiiitiiiitis rr ts se 35 %8 0% x" re dependence ofthe growth rate for the trans © ara tn. Me Fig. 52. The temper formation A General Expression for the Growth Rate les from a super- n of « in a binary system of A and B. Let the compo- ed in terms of c, the concentration of B atoms in units of ‘mol m™, Referring to the schematic phase diagram in Fig. 5.3, at tempera~ ture To, the « soi nposition 2 is just saturated. On cooling 1m becomes supersaturated with respect to Fig. 5.4. Along the distance axis, at y = 0, the precipita tion starts in the matrix of initial composition 2. The precipitate has grown toa size r, where the interface separating the precipitate from the matrix is located, The composition within the precipitate particle from y = 0 toy = r is uniform and is equal to ca, Equilibrium is assumed to prevail atthe inter~ in contact with the precipitate6 SOLID STATE PHASE TRANSFORMATIONS. Pracistte Sams | ‘ ¥ | Matrix is | al ie =| ‘| Due to the formation of the Brrich precipitate, there is a immediately surrounding it. The matrix composition becomes the initial value of 2 only at large distances. The B ‘atoms diffuse down the concentration gradient set up by the depletion, ft the interface and cross it into the precipitate, As the precipitate particle fgrows (o larger sizes, the region of the surrounding matrix depleted of B ‘Moms extends to larger distances, as shown by the dotted-line profile in | GROWTH KINETICS n the interface due to the concent a general expres sphere (5.12) ‘This yields $0, by differentia- (5.13) ‘As we have assumed here Eq. 5.15 yields8 SOLID STATE PHASE TRANSFORMATIONS = Dil = eat Pee ee) a) rena in very small, cap does depend on r. The Thomson-Freundlich ‘equation describes this dependence: cin = cts x (FE) (19) where exp and cZy are the solute concent with a f particle of radius r and co, respectively ‘We ean recalculate the growth kinetics when Eq. 5.18 is applicable. IF 2Va/RTr 1, we can approxi (3.19) If we take that cap > € when r= r*, where nucleation, we have (5.20) (521) U is negative, when r 7738 ium with such particles are stable and tend to grow, ‘OROWTH KINETICS 9 ‘The maximum growth rate, Unan is obtained by setting dU/dr = 0 and cocours at r = 2r%: (5.22) sn of U with ris schematically shown in Fig. 5.5. Atr > r*y ‘growth rate, derived for the case where cag is not a function Gromtnrate these in conjunction with the exponential dependence on temperature Una © (To ~ TP exp (8) 625) Rr ‘Thus, Unas = O at T= Ta and is also very small at low temperatures.80 SOLID STATE PHASE TRANSFORMATIONS ‘The temperature dependence of the maximum growth rate can be obtained by setting 4U,2,+/4T = 0. This yields the condition 2497 Pao 326 Plane Front Growth oe (529) and for d erf (3)/dz from Eq, $.28 into 5.29, we obtain det ‘wt (ia) one (530) al So, differentiating ei 5.27 with respect to 7, we get &),- Bo : (30) (532) parabolic growth rate, as in the case of growth of spherical rer FURTHER READING tions in Metals and Alloys, Growth, Pergamon Press,6 Overall Transformation Kinetics srmation characterized by diff 6.1 EMPIRICAL EQUATIONS. of the exponential form or volume fraction transformed X as a function of X= 1 exp [-(K1 on where K is the reaction rate constant (in units of s!) and i is the time ‘exponent (dimensionless). Taking logarithms in Eq. 6.1, we have 62 ‘OVERALL TRANSFORMATION KINETICS C3 ‘Taking logarithms again, we have = nlog K+ nlogt (co) us time can be plotted as log ‘ponent n. K can be obtained ‘Such a plot is shown for the 8 —> = man- ‘at four transformation temperatures. Valves ‘pond to X — 0.9 and 0.02, respectively. T T T T > 9 4 aos fsor fore a08y 9 4 aL a 4 e L 1 1 a 30 5 rr a 20 3s (for > transformation in log es) Fig. 61 Loplog 1 = 2) verse manganese (after Chistian): Differentiating Eq. 6.1 with respect to time, we obtain the transformation rate: $F = exp(— Cerne 1 =X) from Eq, 6.1, we have Keen = 20, 6a “The nature of X versus ¢ curves depends on the value of m. If m= 1, 4x/ée 4 land the curve becomes a straight line at small values of X anSOLID STATE PHASE TRANSFORMATIONS, Fraction transtormes & Fig. 62. The tear when a9 ye exponent 3) 66) rder terms in Eq. 6.6 can be neglected, so th Co) ‘This is known as the power form describing the transformation kinetics, The corresponding transformation rate is ax _ SF — nkeen 68) ‘Comparison of Eqs. 6.4 and 6.8 shows t these two empirical rate laws is the factor (1 — tion of the untransformed phase. The factor ‘ment factor, because ittakes into account t fof the transformation produ 1¢ only difference between the latter being the frac> =X) is called the impinge power form is not. 62 TRANSFORMATION KINETICS FOR INTERFACE- CONTROLLED GROWTH “The Jobnson-Mehl Model ‘The Johnson-Mehl model of the transformation kinetics for nucleation and growth processes is based on the followi (OVERALL TRANSFORMATION KINETICS 8s mn rate J is equal to 1/(1 — X) ANs/dr, when MN, is the it volume of the system and the factor fe to a unit volume of the a spheres until impingement occurs. th is valid, provided the growth rate is independent of exystal di = Uy =U, =). The number of particles nucl dr is 1dr. So, the volume of al the corresponding volume fraction transformed: ~via 69) ids for ‘extended” and dei formed (or mpigent of te powing hte Inodang ech een (10 Combining xs. 69 and 6.10, we ave for he corse to: = fev ore oan 4 aur yydr =f fevers86 SOLID STATE PHASE TRANSFORMATIONS CEMALL TRANGPORS(ATION KIARTICS 81 =i exp [—- Ze 6.14) ret -ep (-$1) wa) coum aiaea! ime exponent here Here, the time exponent n is 4 and the reaction rate constant K is E19)". th vale af = 4nd 7 mation product Imation and Eq. 6.14 due to Avrami may apply. The term Nz in Eq. 6.14 will s then be equal to the number of grain corners per unit volume of the parent 3 Phase 5 In the ease of grain boundary and grain edge nucleat , 4 o yo °oe 0008 testa tea 400 Teo The Avrami Model ‘Avrami assumed they decrease exponent = Nyexp (—w the number of preferred nucleat per unit volume and 1) les left over at time ¢ and ar system. IF w is very process. Under such conditions, the fraction transformed X varies with time ras follows:88 SOLID STATE PHASE TRANSFORMATIONS ‘These geometries lead to the following kinetic equations, which are signi- Xa 1 = exp (~rLv22) length of grain edges in unit volume. ata grain boundary ing to a time exponent of I with constant U and 1 ~ 0: ~ exp (-25.U) (6.16) the grain boundary area per unit volume. The effect of grain size fbed by a simple kinetic law: D und 6.17) ‘where (isthe time for virtual completion of the transformation and D is the average grain diameter. ‘Time Exponents For interface-controlled growtl growing linear dimension of the particle. For a const ‘an additional exponent of 1 is added. For zero rate of the time exponents for TABLE. 6.1 ‘Time Exponents for Interfuce-controlled Growth Description of the transformation Time exponent >4 4 34 3 Rod-like particles growing radially Grain-edge saturation, zero 2 Slab-like particles thickening Grain boundary saturation, zero 1 ‘OVERALL TRANSFORMATION KINETICS 89 63. TRANSFORMATION KINETICS FOR DIFFUSION: CONTROLLED GROWTH grain boundary area per unit volume for finer grain sizes, as the precipitates are nucleated at the grain boundaries of the parent phase. 4 Soeomen Geonsize Spars? | 2 x= gta 6.18) ‘where Vo isthe total volume of the system, Vo is the volume of # particles at time 1, and terms have their usual meaping. Now the extended volume fraction is given by90 SOLID STATE PHASE TRANSFORMATIONS 6.19) herical precipitate particles and Ny is the volume (6.20) where forr > 6.22) where a [see particles occ wever, the diffusion fields of neighbouring B par Fig. 6.6. The concentra , Which oveupy the discussed in Section 6.2, om OVERALL TRANSFORMATION KINETICS “na srenannig impinge (6.26) This leads to the exponential form of the transformation Kinetics (see Section 6.1): weet ern [~(F0.n2%9)""sa] (62) the time exponent in both Eqs. 6.25 and 6.27 is 3/2 and the react constant K = $ Devi period’ ‘each dimension of the particle grows as the square root of time and the ‘volume varies as 3/2, just as shown above for a sphere. 1 1 Codeteieeieba Dada.oy SOLID STATE PHASE TRANSFORMATIONS. 34 ‘of Wert and Zener compared with the in EQ, 627 ‘Time Exponents in Diffusion-controled Growth For diffusion-controlled growth 1¢ exponent of 4 arises for each growing dimension of the particle. For a constant rate of nucleation, an additional exponent of 1 arises. Table 6.2 summarises the time exponents for different situations, TABLE 6.2 ‘Time Exponents for Diffuslon-controlled Growth Description of the transformation Time exponent Particles of all shapes growing in all three dimensions 1 increasing 7 > 2. con: 4 FURTHER READING ian, The Theory of Transformations in Metals and Alloys, Chapter 12, Formal Theory of Transformation Kinetics, Pergamon Press, Oxford, p. $25 (1975). OVERALL TRANSFORMATION KINETICS 93 EXERCISES different tem- as follows; fe constant K and the time exponent n for the following. transformation Time, (3) x 38 0.046 45 O19 63 0.31 69 0.49) 16 0.60 9.5 0.80 10.0 0.93 of transformation at X = 0.5. Is this the maxi Johnson-Mehl equation is = 10" mm s-!, caleulate the nucleation rate, 5. e data of Problem 6.2, calculate and pl per unit volume between parent and product phases and (b) as a function of fraction transformed. Wh: for the shape of the above plots? he interfacial area function of time,7 Particle Coarsening such an alloy has to be suficie I occur due to coarsening within the it made from that alloy is chapter, we first describe hi ing the Thomson: 7.1 DRIVING FORCE FOR COARSENING reduction in the total interfacial energy of the system provides driving force for coarsening. The surface area-to-volume radius, So, for a given volume f particle size. Hence, coarsening decreases the “The Thomson-Freundlich Equation “The equation basic to the kine Freundlich solubility relationship equation): lysis of coarsening is the Thomson- Iso known as the Gibbs-Thomson PARTICLE COARSENING. 95 tan Ben on where is the composition of the « matrix in equilibrium with a spheri- cal particle B of radius r (or a particle with the local radius of curvature 7), the matrix composition in equilibrium with a particle of very ge radius, Vis the molar volume of the precipitate, and © is the specific interfacial ereray. ; ono. Win Vai, 20RD O08 Taney aken tb all compared Wh Vagrant Eq heen besplted . 2s dan a1 +B 0 Equations 7.1 and 7.2 indi free-energy/composition relationships. the free energy of the B particle includes its surface energy. It incres cS size. The common tangents drawn at the (a) yield the matrix compositions as 1e corresponding phase diagram. Note the & phase increases with decreasing precipitate size, Fig. 7A (a) Sehe for particles of varying sizes in equilibrium (©) The corresponding phase diagram96 SOLID STATE PHASE TRANSFORMATIONS write from Eq. 7.2: ca~ y= og (LD te — ee = 08 BEG — (7.3) For ri Ira and ce > e~. Thus a concentration local L occur down > gradient, ic., from the smaller particle to the larger parti- strated in Fig. 7.2. As the flow occurs, some part of F particle dissolves to compensate for the decrease in concentration terface due to the flow. Likewise, some amount of the solute will the bigger particle to remove the excess solute arriving as the “no i § between two average radius F of the two particles is (ru + ra)/2. Note that F decreases, a8 rz increases and r; decreases, with the total volume kept 1. But when ri becomes zero and hence is no longer counted in the average, ? jumps from (n> 0 + r2)/2 to ri. Therefore, the coarsening is ‘mainly due to the decrease in the number of particles. PARTICLE COARSENING ” 7.2. THE KINETICS OF COARS! (GREENWOOD'S MODEL) ING long range diffusion of the solute through the matrix or ferface, Note that Eq. 7.2 can be applied only when the process is not sface-conteolled, asthe equat face. If diffusion through the mate extremely difficult, Here, a simple mo Greenwood developed a model fo ‘on the following assumptions: 1 the particles are spherical in shape he particle-coarsening Kinetics based independent of the matrix ‘coarsening, i., the volume frac= oa) particle, From the general growth equation led processes, we can write:98 SOLID STATE PHASE TRANSFORMATIONS. corresponding to an infinite radius of curvature arises, as all the other Dates hve nit radi of curvatre, Following Eas. 8-12and$.13, wecan 08), Using els = ets Bre - cae 2tt) <0 + 2¥0/RTr) in Eq. 7.12, ir Ss) = Nyets 2H Ea) = Noein 2FE em2Veo SRT 3) where F is the mean radius. PARTICLE COARSENING. Rewriting Eq. 7.13, we have Vo e= (1497) 7.4) the Thomson-Freundlich equation, we conclude ion that a particle sees at large distances is jum with a particle of “The concentration profiles near particles of different sizes are shown iy in Fig, 7.3. A particle of mean radius F sees no concentration [A particle of size ry F ces an upward concentration gradient ahead of it, which favours the arrivial of fresh solute atoms at face and growth of that particle. The Giffusion directions are indicated by arrows in Fig. 7.3. ep = Porte aye = Distance ig. 72, Concentration gradient ahead of partes of inserted in Eq. 7.7, we ot When the expression for e’ (Eq. 7.14) fornF100 SOLID STATE PHASE TRANSFORMATIONS. Fastest Growing Particles ‘The maximum growth rate Ums can be obtained from av, _ A flit ty) _ #@-i26-3]-° 1, we find that the particle that grows fa particles is that which has a radius r= = 2F. Sut wwe obtain drm _ p, oe Une = GE = De ‘Thus the maximum growth ra jon in growth the coarsening anneal are shown in Fig. 7.4, The locus of the radius test growing particle is also shown. Geowthiate U it Fig. 74 Grow with cre ‘The Cube-root Relationship The rate of coarsening can be expressed as a functional between P and 1. Consider particles of radius rm = 2F. wwe obtain (7.18) [tom = PARTICLE COARSENING 101 where / = 2° at the beginning of the coarsening process, Solving for F, we get = [oe + 3, 22 J” = [oo + do. o ‘The mean radius P can be determined by quantitative metallography from the Ful eon article og 7, pm (7.20) for the coarsening of copper pr ‘shown in Fig. 7.5.1f 7? is not small enough to be § Yo) = UH) TO) x=coarsening of MayAls (8) precipitates in an From quate mtopy, sat where fy is the volume fraction of the B precipitates. Using Eq, 7. eet r r-1- 7G Substituting for P(t) from Eq. 7.19 in 7.24, we have PARTICLE CoaRaRWINE FURTHER READING G.W, Greenwood, Acta Metall, 4, 243 (1956). EXERCISE 7. Calculate and plot the composition of ferrite coexisting with int fred De ae ches steno teen cata ieee8 Pearlitic Transformations pearlitic transformation in stels. In this chapter, the distinction between continuous and discontinuous precipitation is frst brought out. After describing the experiment the 8.1 CONTINUOUS VERSUS DISCONTINUOUS PRECIPITATION In continuous precipitation, particles of the product 8 phase rich in solute nucleate and grow with a continuous decrease in solute concentration of the supersaturated matrix 2, by solute diffusion across a concentration gradient ‘ahead of the growing particle: apart > Haat +B ‘The parent « phase, except for retains its crystal structure and mar developed in Section 6.3 In discontinuous precipi lamellae or rods lying gene ‘colony), precipitate from cept > B+? (62) “The average composition of the two (or sometimes more) product phases G+ 7) within each cell is equal to the overall composition of the parent spersaturated solution: 08 PEARLITIC TRANSFORMATIONS, 10s phase a. As the transformation proceeds, product phases advance sideways, fe. The cellular configuration grows atthe expense ofthe parent growth of a colony of par rated in Fig. 81 we ‘rain boundary Fig. £1 Sidewise and lengihwise growth of e ‘A typical example of the discontinuous preci f accompanied by ‘morphology. 8.2. EXPERIMENTAL CHARACTERISTICS is usually depicted inthe form the transfor Cecurve that region. The in size and also in the presence of ‘The nucleation of pearlite occurs predominantly at the austenite grain boundaries. The finer is the grain size of austenite, the more is the boundary area per unit volume and also the number of potential nucleation sites. This explains the higher transformation rate observed in finer-grained austenite. Figure 83 depicts the sigmoidal transformation curves for three different106 SOLID STATE PHASE TRANSFORMATIONS. PEARLITIC TRANSFORMATIONS nie coarse Fig, 6.3 The pea Yaa108 SOLID STATE PHASE TRANSFORMATIONS PEARLITIC TRANSFORMATIONS 109 the grain on the other side of the boundary, facilitating strain-free growth ‘The morpho 1 growth of nodules of pe ships observe colony of nodule is found to be indepen 8.5 shows the pearlite nodule ‘The Time Dependence of the Nucleat s provided a simple picture for the development of lamel n. The nucleation and growth of either phase in the transforma tion product alter the local composition of the surrounding matrix in such 1m of 0.5% molybdk fA way as to favour the nucleation and growth of the other phase. For by about two orders of magnitude. example, nucleation of a cementite particle depletes the surrounding mat of carbon, which promotes the muc 8.4. The growth rates are in the range of 10 the presence of sul T T T 7 T ons ough this does not affect the subsequent argument, Fisher assumes fon an austenite grain boundary. The fee austenite interface. The interfacial area bet- 070 particle grows. The : . terface also increases. Once the t g Tn developing. model along these lines Bo" assumptions: ; |The cementite particle is spherical in shape inthe in 2 4 2. Its growth in the early stages is diffsion-controlled and ° parabolic. ion rate of cementite per unit area of the austenite grain nt temperature 05 area ofthe austenite-cementite perature “The surface area A of a cementite patisle of radius ris 80 Aa Ant 6 “The initial parabolic growth of the particle yields Fig. #5 The peatite nodule radi me atthe transformation 4) the boundary. This results in an incoherent interface between the nucleus and ph 6u10 SOLID STATE PHASE TRANSFORMATIONS ~ Pyirar 67 Substituting for 17 from Eq. 8.6 in 8.7, we obtain — Pyaee «8 Rearranging and =In( grating, we have Innate = exp (=m) 69) where m = J/2m2, From Eq. 8.9, we have -0, P=0; =0,P=1 the rate of pearlite nucleation per cementite particle, then nade a = Amt exp (—m per unit area ofthe aust 4h is obtained by 1g Eq. 8.10 to include all cemen Ducleate per u fe This yields (6.12) Thus the pei stages of the react nucleation rates wi measurements of 4.4. THE MECHANISM AND THE KINETICS OF GROWTH Long-range Diffusion of Carbon In plain carbon steels, the diffusion of carbon to effec ‘changes necessary for the transformation is the rate cont ‘compositional ing step. In an PEARLITIC TRANSFORMATIONS am Jp gmvctes m3 sion domi diffusion. here. A ratio of Ds/Di (where ! stands © Fequired to make the boundary diffu12 SOLID STATE PHASE TRANSFORMATIONS. ‘The Growth Kinetics ‘Assuming that the diffusion of carbon is through the austenite just ahead of the austenite-pearlite interface. we derive here an expression for the ‘growth rate U of a pearlite colony. The geometry of the growth is illustrated schematically in Fig. 8.7. Seis the trve interlamellar spacing. cvs is the con- centration of carbon in austenite that is in contact with ferrite across the interface. cy-req is the carbon concentration in austenite in contact with ‘cementite. These two concentrations at the reaction temperature are obtained by extrapolation of the corresponding phase boundaries into the metastable region, as illustrated in Fig. 8.8. As seen in the figure, cra > ¢r-Fexs Sothat Lengths ‘youth rene re engin eon Fee ya ae —4} Fig. 88 Huligren's extrapolation of the phase ‘the metastable region of an Fe~C phase diagram, PEARLITIC TRANSFORMATIONS a) ‘a concentration gradient exists in the austenite for carbon diffusion from regions which would transform to ferrite to regions that would form cemen~ tite. The directions of carbon di are shown by arrows i Fig. $7. a Fe ETON equation for fusiononled row ven by 1 legree of supercooling below the eutectoid tempe Eq. 4.5, | 4g | c 47. By substituting this and Eq. 8.18 i 19) (8.20) So, we note from Eqs. 8.20, 8.19 and 8.18, ue ex (--£ are 21) NIM |4 SOLID STATE PHASE TRANSFORMATIONS (ATA, just below the eutectoid temperature. After reaching a maximum, the rate decreases again, as the exponential term in Eq. 8.21 dominates. Figure 8.4 compares the experimental measurements of growth rates as function of temperature with those obtained from Eq, §.21 8. INTERLAMELLAR SPACING In the previous section on growth kinetics, it was pointed out that the diffusion of carbon in austenite can be rate controlling. The diffusion distance is approximately half of the interlamellar spacing. As the spacing increases, the diffusion distance also increases and the concentration gradient decreases, resulting in a decrease in the growth rate. Ifthe spacing is small, the growth rate is not necessarily large. With a decrease in spacing, the atea of the a-FesC interface per unit volume of pearlite increases. So the net driving Force for the reaction decreases 'A net driving force Jou for the transformation is defined by taking into ergy of the 2-Fe,C interfaces that lie within a pearlite colony dg + Ao (8.22) fea of the ferritexcementite interfaces per unit volume of specific surface energy. In a u perpendicular to the a-FesC ‘area each. So, we have jum possible interlame! where the net driving foree is zero. © spacing Sin Corresponds to the case ag (829) Substituting for 4¢ from Eq. 8.25 in 8.24, we have toa ene = 2455 5) 626 ‘The optimum value of So corresponds to the maximum growth rate Uns Substituting for 4g from Eq. 8.26 in 8.19, we have for the growth rate FPEARLITIC TRANSFORMATIONS us Setting 4U/ASe = 0, the condition forthe maximum growth rates obtained: 142 Sse 7% =0 So = 2S ria (6.28) “The interlamellar spacing corresponding to the maximum growth rate is thus twice the minimum interlamellar spacing corresponding to the net driving force becoming zero. Since | dg | 47 and Sain = 450, we see 1 So should vary inversely as the degree of supercooling. log-log plot of the experimental values of interlamellar indicating an interfacial Figure 8.9 is spacing So as a function of 47. The slope is seen to be —1 inverse linear relationship between Soand 47. The ferrite-ceme! energy @ can be evaluated from this plot and Eq. 8.27, and ‘about 3 J m2, This value is about an order of magnitude higher than that 49) T T T T testsoin dd Ve ve r 18 36 2 og a7 Fi 89. The interiameta spacing Sa funtion of he dere soperoaing116 SOLID STATE PHASE TRANSFORMATIONS 86 EFFECT OF ALLOYING ELEMENTS ‘The free energy change 4g for the auste by the presence of alloying elements in the steel. Most all fat ‘growth kinetics. The ratio of joundaty and lattice diffusion, is expected to ‘be much larger than 10? for substitutional diffusion. Correspondi diffusion short circuit provided by the austenite-pearl jing of the alloying element ferrite and cementite FURTHER READING J.C. Fisher, Eutectold Decompos.tions, Amer. Soc. Metals, Metals Park, OH, p. 201 (1950). JW. Cahn and W.C. Hagel, Decomps science, New York, p. 131 (1962). ‘of Ausienite by Diffusional Processes, of Austenite by Difusional Trans., EXERCISES. te interfacial energy for a steel forma 1500 A, (b) diffusion coefficient of carbon D? = 3% 10-" m? st the wide variation in the time exponents |) — X) versus log for pearlite transformed ransformations, where site saturation occurs in the time dependence of the nucleation rate exponents obtained from the transformation 9 Massive Transformations has been established that, in a number of alloy ‘can transform to the product phase by short-range involves a relatively rapid uncoordinated transfer of ferfuce, without a change in composition. Such trans- 1920, Philips was the first to observe that. in Cu-Zn alloys, drastic quenching converted the high temperature B phase of the % phase, n. Ten years later, Greninger used the term blocky appearance of FCC « crystals obtained by quenchi :nowledge concer in recent years, in an dalloy during conti- 1 avoided. The rate of cool right range, If the cooling rate is to0 slow, a competing nal change ean oceur by means of long range on the other hand, can induce a diffusion- jc transformation’ by shear displacement. A ie rate plot in Fig. 9.1 for pure at the higher transformation temperature corresponds to the msus SOLID STATE PHASE TRANSFORMATIONS =~ Fe-2at7an Fig 92 Nucleation of the masive product atthe parent grain boundaries, (Courtesy. The lestitute of Meals, London ) MASSIVE TRANSFORMATIONS us no particular orientation relationship between the product and the parent phases. The massive product grows roughly ‘spherical front with, irregular interfaces and may sweep across the phase. ‘The growth rate of the massive product is quite high, about 10 mm s~*. the growth rates observed in other transformations formati ‘A growth rate of 10 mm s~! means sconds 10 grow { is technique, a specimen che high temperature phase TABLE 9.1 Approximate Growth Rates for Different Transformations ee Growth rate, Type of transformation pep ee by long range diffusion of 10> te by long range diffusion of 10 —- 107 109 — 107+ 10 108 9.2. THE DRIVING FORCE FOR THE MASSIVE ‘TRANSFORMATION ‘The free energy-composition curves for the x and 8 phases of a binary 9.3. The common tangent xy’ to the anges corresponding to 1¢ driving force for the transformation without a compositional change decreases, as the overall composition shifts to the right in Fig. 9.3, tuntil it becomes zero at the point of intersection of the two free energy