Journal of Alloys and Compounds 544 (2012) 152158

Contents lists available at SciVerse ScienceDirect

Journal of Alloys and Compounds

journal homepage: www.elsevier.com/locate/jalcom

Analysis of phase formation in multi-component alloys

R. Raghavan, K.C. Hari Kumar, B.S. Murty
Department of Metallurgical and Materials Engineering, Indian Institute of Technology Madras, Chennai, India

a r t i c l e i n f o a b s t r a c t

Article history: An attempt has been made to predict phase formation using a CALPHAD-based approach for a large num-
Received 22 June 2012 ber of compositions that are known to form FCC, BCC and a mixture of FCC and BCC phases. The stable
Accepted 22 July 2012 phase is assumed to be the rst phase that is formed upon cooling from liquid state with the highest driv-
Available online 7 August 2012
ing force. The driving force for other phases at the transition for various compositions is also presented. A
comparison between the parametric approach of phase prediction by study of thermodynamic and topo-
Keywords: logical parameters on one hand and the CALPHAD approach on the other is also presented. CALPHAD
Multi-component alloys
approach appears to predict BCC phase formation much more accurately than the FCC phase formation.
High entropy alloys
Thermodynamics
The results indicate that solid solution formation in multicomponent alloys is favored when the ratio of
CALPHAD DScong/DSfusion is greater than 1 and 1.2 for equiatomic and non-equiatomic alloys, respectively. The
Phase prediction results also point out that BCC phase is favored when the atomic size difference is larger, which is
reected by a higher value of DSr/k. Formation of FCC phase appears to be governed mainly by DHmix,
while BCC phase governed by DSr/k, which is representative of strain in the structure.
2012 Elsevier B.V. All rights reserved.

1. Introduction
Multi-component nearly-equiatomic alloys, also known as
high-entropy alloys (HEAs) constitute a recent class of alloys rst
developed by Yeh and his co-workers [1] in 2004. The most signif-
icant aspect of this class of alloys is that they yield mostly a simple
solid solution or a mixture of solid solutions based on FCC or BCC
phases. The presence of multi-principal elements in nearly equi-
atomic composition gives rise to large entropy of mixing (congu-
rational entropy), which increases with the number of elements if
the composition remains close to equiatomic. Consequently, alloys
can be formed with simple crystal structure by suppressing forma-
tion of complex phases and intermetallics.
Further, alloys with these simple structures also possess several
promising properties such as high strength, wear resistance, oxida-
tion resistance, etc., which in turn closely depend on the phase(s)
formed. In this regard, to obtain better homogeneity in properties
and to truly call the resulting alloy a high-entropy alloy, the tar-
get is to obtain a single solid solution. In this work, formation of
FCC and BCC phases, the two most important phases seen in this
class of alloy systems is studied using CALPHAD-based and para-
metric thermodynamic and topological analyses.
with their references [223]. These are in turn divided according to the predomi-
nant phase that is observed (a) FCC, (b) FCC + BCC and (c) BCC. In total, 93 com-
positions were studied that fall into either FCC or BCC or a mixture of FCC and
BCC phase forming compositions. The exact split is given in Table 3.
3. Background of the CALPHAD-based approach
The CALPHAD technique makes use of the principle that the
Gibbs energy of a phase, described as a function of temperature
and composition, is enough to obtain a complete thermodynamic
description of the system since almost all thermodynamic proper-
ties can be derived from the Gibbs energy function. Central to nd-
ing the stable/metastable state of the system is the Gibbs energy
minimization. Equilibrium conditions are described as a set of
non-linear equations involving the Gibbs energies of the corre-
sponding phases.
According to the CALPHAD approach, the Gibbs energy of a
phase u is expressed as [28]
Gu ref Gu id Gu ex Gu 1
The rst term is the contribution of the pure components of the
phase, the second is the ideal mixing contribution and the third is
the excess Gibbs energy of mixing. The second term is given by:

2. List of compositions
id
Gu RT Ri xi ln xi 2
For a liquid phase, the excess Gibbs energy is described as:
For the present work, the systems from literature used for the analysis are listed
in Table 1 (equiatomic systems) and in Table 2 (non-equiatomic systems) along ex
Gu Ri Rj>i xi xj Rv L/ij xi  xj v 3

Corresponding author. The L/ij model parameters are the RedlichKister parameters that
E-mail address: murty@iitm.ac.in (B.S. Murty). need to be determined by an assessment procedure that relies on

0925-8388/$ - see front matter 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jallcom.2012.07.105
 R. Raghavan et al. / Journal of Alloys and Compounds 544 (2012) 152158 153

Table 1 Table 2
Equiatomic compositions studied in literature used for this work. MA stands for Non-equiatomic compositions studied in literature used for this work.
synthesis by Mechanical Alloying.
System Synthesis Route Refs.
System Synthesis route Refs.
(a) FCC
(a) FCC CoCrCuFeNiTi0.5 Casting [2]
CoCrCuFeNi Casting [2] CoCrCuFeNiTi0.8 Casting [2]
CoCrCuFeNiTi Casting [2] Cu2.5NiCoZnAlTi MA [4]
FeNiCrCuMo Casting [3] Cu5NiCoZnAlTi MA [4]
CuNi MA [4] Ni2.5FeCrCoMnW MA [5]
CuNiCo MA [4] Ni5FeCrCoMnW MA [5]
CuNiCoZn MA [4] NiCrCuFeMo0.25 MA [14]
CuNiCoZnAl MA [4] NiCoCrFeMo0.25 MA [14]
NiFe MA [5] Al0.5CoCrCuFeNi Casting [15]
NiFeCr MA [5] Al0.375CoCrFeNi Casting [8]
NiFeCrCo MA [5] Al0.5CrFeCoNi Casting [16]
NiFeCrCoMn MA [5] CoCrFeNiAl0.3 Casting [17]
FeNiMn MA [5] CoCrFeNiMo0.3 Casting [17]
NiCoCrCuFe MA [6] CoCrFeNiAl0.3Mo0.1 Casting [17]
(b) FCC + BCC FeCoNiCrCu0.5 Casting [18]
FeNiCrCuAl Casting [3] Ti0.5CrFeCoNiAl0.25Cu0.75 Casting [19]
FeNiCrCuMn Casting [3] Al0.5CoCrCuFeNi Casting [20]
AlCoCrCuFeNi Casting [7] Cu0.5NiAlCoCrFeTiMo MA [21]
AlCoCrFeNi Casting [8] (b) FCC + BCC
AlCrFeCoNiCuMn Casting [9] Ti0.5CrFeCoNiAl0.5Cu0.5 Casting [19]
AlCrFeCoNiCuTi Casting [9] Al0.5CoCrCuFeNiTi0.2 Casting [20]
CrCuMnNi MA [10] Al0.5CoCrCuFeNiTi0.4 Casting [20]
AlCrCuFeNiZn MA [11] Al0.5CoCrCuFeNiTi0.6 Casting [20]
AlCoCrCuNiZn MA [11] Al0.5CoCrCuFeNiTi0.8 Casting [20]
(c) BCC Al0.5CoCrCuFeNiTi Casting [20]
AlFe MA [12] Al0.3CrFe1.5MnNi1.5 Casting [22]
AlFeTi MA [12] Al0.3CoCrFeNi Casting [23]
AlFeTiCr MA [12] Al0.3CoCrFeNiMo0.1 Casting [23]
AlFeTiCrZn MA [12] Cu0.5CoNiCrAl Casting [24]
AlFeTiCrZnCu MA [12] Ti0.5CrFeNiAlCo1.5 Casting [25]
FeTiCrZnCu MA [5] Ti0.5CrFeNiAlCo2 Casting [25]
CuNiCoZnAlTi MA [5] Ti0.5CrFeNiAlCo3 Casting [25]
NiCoZnAlTi MA [5] Al0.45FeNiCrCuCo MA [26]
NiFeCrCoMnW MA [5] (c) BCC
FeCrCoMnW MA [5] Al2.5CoCrFeNi Casting [3]
FeNiMnAl MA [5] Al3CoCrFeNi Casting [3]
FeNiMnAlCr MA [5] Al0.45FeTiCrZnCu MA [5]
FeNiMnAlCrCo MA [5] Al2.5FeTiCrZnCu MA [5]
NiMnAlCrCo MA [5] Al5FeTiCrZnCu MA [5]
AlCoCrCuFe MA [6] AlFe0.45TiCrZnCu MA [5]
AlCuCrFe MA [6] AlFeTi0.45CrZnCu MA [5]
AlCrCoFe MA [6] AlFeTiCr0.45ZnCu MA [5]
TiCrFeCoNiAl Casting [13] AlFeTiCrZn0.45Cu MA [5]
AlFeTiCrZnCu0.45 MA [5]
Cu0.45NiCoZnAlTi MA [5]
Ni0.45FeCrCoMnW MA [5]
statistical techniques to obtain the best t to available experimen- Fe0.45NiMnAlCrCo MA [5]
tal data. The binary excess terms are combined using the Muggianu Fe2.5NiMnAlCrCo MA [5]
Fe5NiMnAlCrCo MA [5]
geometrical extrapolation scheme [29].
AlCrCuFeMo0.25 MA [14]
The entire thermodynamic description of phases, which pre- AlCoCrCuFeMo0.25 MA [14]
dominantly includes the Gibbs energy functions as a function of AlCoCrCuFeMo0.5 MA [14]
temperature and composition using a suitable model is contained Al2CoCrCuFeNi Casting [15]
in a TDB le. Thermo-Calc is the software used here, which reads Al2CoCrFeNi Casting [8]
Ti2CrFeCoNiAl Casting [13]
the thermodynamic description from the TDB le and performs Ti3CrFeCoNiAl Casting [13]
the equilibrium calculations. The TDB le contains the unary, bin- CrFeCoNiAlCu0.25 Casting [16]
ary and many ternary interaction parameters of liquid, FCC, BCC Ti0.5CrFeCoNiAl0.75Cu0.25 Casting [19]
and HCP phases for all the elements that are listed in the various Al0.5CoCrCuFeNiTi1.2 Casting [20]
Al0.5CoCrCuFeNiTi1.4 Casting [20]
compositions under study (as outlined in Tables 1 and 2). These
Al0.5CoCrCuFeNiTi1.6 Casting [20]
elements are Al, Co, Cr, Cu, Fe, Mn, Mo, Ni, Ti, W and Zn. Al0.5CoCrCuFeNiTi1.8 Casting [20]
Al0.5CoCrCuFeNiTi2 Casting [20]
4. Implementation and results: driving force calculations using Al0.5CrFe1.5MnNi1.5 Casting [22]
Ti0.5CrFeNiAlCo Casting [25]
Thermo-Calc Al2.5FeNiCrCuCo MA [26]
Al5FeNiCrCuCo MA [26]
Using Thermo-Calc, calculations were performed to nd the AlCoCrFe0.6Mo0.5Ni Casting [27]
rst phase that would crystallize from the liquid state, after incor- AlCoCrFeMo0.5Ni Casting [27]
AlCoCrFe1.5Mo0.5Ni Casting [27]
porating all the required interaction parameters in the TDB le. The
AlCoCrFe2Mo0.5Ni Casting [27]
most stable phase is assumed to be the rst phase to nucleate out
154 R. Raghavan et al. / Journal of Alloys and Compounds 544 (2012) 152158

Table 3  It is seen that the CALPHAD predictions for the systems

Distribution of systems according to phases. seen to form BCC phase are highly accurate. For the 37
System type No. of phases Total No. of phases non-equiatomic compositions showing BCC phase, the
FCC FCC + BCC BCC prediction is correct for 31 cases. For equiatomic systems,
BCC is correctly predicted for 16 of the 18 cases.
Equiatomic 13 9 18 40
Non-equiatomic 18 9 27 53
 In case of FCC, the prediction is only about 50% accurate.
For 18 systems showing FCC in case of non-equiatomic sys-
tems, only seven cases are predicted correctly. For equi-
from the liquid. For this purpose, for every composition, calcula- atomic systems on the other hand, out of the 13
tions are performed in such a way that the alloy is cooled from a compositions, in seven compositions FCC is predicted cor-
temperature of 4000 K (where, there is complete assurance of li- rectly. For majority of the cases where FCC is not predicted
quid alone). The temperature at which the primary solid crystalizes correctly, HCP is the predicted phase.
out is denoted as Tpc (temperature of primary crystallization). One  However, the driving force for other phases to form at the
can also estimate the driving forces for other phases at Tpc, which is transition temperature sheds a lot of light on the next eas-
nothing but a relative measure of the ease of formation of the iest phase that would form. For 13 out of 17 cases where
respective phases under consideration. The most stable phase, FCC is not predicted, the driving force closest to zero is
viz. the phase formed at Tpc is then compared with the results from seen to be for FCC indicating that FCC is at least the sec-
literature and presented in Table 4 for equiatomic compositions ond-most favorable phase to form at the transition tem-
and Table 5 for non-equiatomic compositions. It is to be noted that perature for those compositions.
the driving force for the formation of equilibriated phase is 0 kJ/  For the systems that show a mixture of FCC and BCC (14
mol at the transition temperature. The phase seen experimentally, for non-equiatomic and nine for equiatomic), HCP phase
viz., FCC or BCC or a mixture of FCC and BCC is listed above each is rarely predicted with driving force of FCC and BCC being
subsection in the tables. The observations from these calculations more close to zero than HCP. For equiatomic systems that
are the following.

Table 4
Driving force calculations for equiatomic compositions using CALPHAD approach.

System Phase predicted Temperature of transition (K) Driving forces (kJ/mol)

FCC BCC HCP
(a) Phase seen experimentally FCC
CoCrCuFeNi HCP_A3 1463 0.03 0.09
CoCrCuFeNiTi HCP_A3 1419 0.09 0.12
FeNiCrCuMo BCC_A2 2168 0.12 0.07
CuNi FCC_A1 1585 0.20 0.17
CuNiCo FCC_A1 1653 0.20 0.08
CuNiCoZn FCC_A1 1450 0.20 0.08
CuNiCoZnAl BCC_A2 1313 0.03 0.57
NiFe FCC_A1 1718 0.13 0.41
NiFeCr FCC_A1 1669 0.03 0.24
NiFeCrCo FCC_A1 1539 0.01 0.18
NiFeCrCoMn BCC_A2 1481 0.02 0.30
FeNiMn FCC_A1 1513 0.09 0.50
NiCoCrCuFe HCP_A3 1463 0.03 0.09
(b) Phase seen experimentally FCC + BCC
AlCoCrCuFeNi HCP_A3 1306 0.11 0.08
AlCoCrFeNi BCC_A2 1252 0.11 0.18
AlCrFeCoNiCuMn BCC_A2 1276 0.08 0.11
AlCrFeCoNiCuTi BCC_A2 1312 0.10 0.10
FeNiCrCuAl BCC_A2 1360 0.08 0.06
FeNiCrCuMn FCC_A1 1442 0.00 0.02
CrCuMnNi BCC_A2 1506 0.11 0.30
AlCrCuFeNiZn BCC_A2 1368 0.06 0.37
AlCoCrCuNiZn BCC_A2 1355 0.01 0.64
(c) Phase seen experimentally BCC
TiCrFeCoNiAl BCC_A2 1329 0.18 0.29
AlFe BCC_A2 1592 0.32 1.54
AlFeTi FCC_A1 1187 0.33 0.29
AlFeTiCr BCC_A2 1319 0.11 0.71
AlFeTiCrZn BCC_A2 1244 0.02 0.59
AlFeTiCrZnCu BCC_A2 1269 0.00 0.62
FeTiCrZnCu BCC_A2 1466 0.09 0.27
CuNiCoZnAlTi BCC_A2 1397 0.05 0.66
NiCoZnAlTi BCC_A2 1199 0.31 0.10
NiFeCrCoMnW BCC_A2 1967 0.33 0.58
FeCrCoMnW BCC_A2 2128 0.39 0.54
FeNiMnAl FCC_A1 1325 0.07 0.63
FeNiMnAlCr BCC_A2 1400 0.07 0.53
FeNiMnAlCrCo BCC_A2 1287 0.07 0.39
NiMnAlCrCo BCC_A2 1322 0.17 0.45
AlCoCrCuFe BCC_A2 1279 0.29 0.00
AlCuCrFe BCC_A2 1435 0.22 0.37
AlCrCoFe BCC_A2 1236 0.30 0.39
 R. Raghavan et al. / Journal of Alloys and Compounds 544 (2012) 152158 155

Table 5
Driving force calculations for non-equiatomic compositions using CALPHAD approach.

System Phase predicted Temperature of transition (K) Driving forces (kJ/mol)

FCC BCC HCP
(a) Phase seen experimentally FCC
Al0.5CoCrCuFeNi HCP_A3 1384 0.10 0.10
Al0.375CoCrFeNi BCC_A2 1371 0.00 0.14
Al0.5CrFeCoNi BCC_A2 1334 0.02 0.14
CoCrFeNiAl0.3 FCC_A1 1403 0.00 0.15
CoCrFeNiMo0.3 FCC_A1 1523 0.04 0.17
CoCrFeNiAl0.3Mo0.1 FCC_A1 1397 0.01 0.14
FeCoNiCrCu0.5 FCC_A1 1464 0.05 0.10
CoCrCuFeNiTi0.5 HCP_A3 1452 0.08 0.13
CoCrCuFeNiTi0.8 HCP_A3 1434 0.09 0.13
Ti0.5CrFeCoNiAl0.25Cu0.75 HCP_A3 1371 0.07 0.06
Al0.5CoCrCuFeNi HCP_A3 1384 0.10 0.10
Cu0.5NiAlCoCrFeTiMo BCC_A2 1748 0.22 0.33
Cu2.5NiCoZnAlTi BCC_A2 1342 0.03 0.60
Cu5NiCoZnAlTi FCC_A1 1282 0.02 0.44
Ni2.5FeCrCoMnW BCC_A2 1806 0.21 0.51
Ni5FeCrCoMnW FCC_A1 1591 0.07 0.31
NiCrCuFeMo0.25 HCP_A3 1768 0.07 0.10
NiCoCrFeMo0.25 FCC_A1 1526 0.03 0.17
(b) Phase seen experimentally FCC + BCC
Al0.3CrFe1.5MnNi1.5 FCC_A1 1495 0.05 0.38
Al0.3CoCrFeNi FCC_A1 1403 0.00 0.15
Ti0.5CrFeCoNiAl0.5Cu0.5 BCC_A2 1323 0.08 0.00
Al0.3CoCrFeNiMo0.1 FCC_A1 1397 0.01 0.14
Cu0.5CoNiCrAl FCC_A1 1313 0.01 0.20
Ti0.5CrFeNiAlCo1.5 BCC_A2 1266 0.16 0.20
Ti0.5CrFeNiAlCo2 BCC_A2 1237 0.13 0.16
Ti0.5CrFeNiAlCo3 BCC_A2 1219 0.06 0.11
Al0.45FeNiCrCuCo HCP_A3 1392 0.09 0.10
Al0.5CoCrCuFeNiTi0.2 HCP_A3 1382 0.11 0.11
Al0.5CoCrCuFeNiTi0.4 HCP_A3 1376 0.12 0.10
Al0.5CoCrCuFeNiTi0.6 HCP_A3 1367 0.12 0.08
Al0.5CoCrCuFeNiTi0.8 HCP_A3 1356 0.00 0.05
Al0.5CoCrCuFeNiTi HCP_A3 1342 0.00 0.01
(c) Phase seen experimentally  BCC
Al2CoCrCuFeNi FCC_A1 1328 0.27 0.37
Al0.5CrFe1.5MnNi1.5 FCC_A1 1459 0.05 0.40
Al2CoCrFeNi BCC_A2 1199 0.25 0.41
Al2.5CoCrFeNi BCC_A2 1177 0.30 0.56
Al3CoCrFeNi BCC_A2 1155 0.33 0.71
CrFeCoNiAlCu0.25 BCC_A2 1248 0.11 0.16
Ti0.5CrFeCoNiAl0.75Cu0.25 BCC_A2 1313 0.12 0.19
Ti0.5CrFeNiAlCo BCC_A2 1305 0.16 0.24
Ti2CrFeCoNiAl BCC_A2 1330 0.18 0.35
Al0.45FeTiCrZnCu FCC_A1 1329 0.14 0.24
Al2.5FeTiCrZnCu BCC_A2 1221 0.11 0.73
Al5FeTiCrZnCu BCC_A2 1130 0.13 0.42
AlFe0.45TiCrZnCu FCC_A1 1268 0.26 0.77
AlFeTi0.45CrZnCu BCC_A2 1349 0.13 0.58
AlFeTiCr0.45ZnCu FCC_A1 1277 0.05 0.62
AlFeTiCrZn0.45Cu BCC_A2 1338 0.11 0.65
AlFeTiCrZnCu0.45 BCC_A2 1273 0.04 0.62
Cu0.45NiCoZnAlTi BCC_A2 1397 0.37 0.60
Ni0.45FeCrCoMnW BCC_A2 2046 0.37 0.57
Fe0.45NiMnAlCrCo BCC_A2 1300 0.12 0.42
Fe2.5NiMnAlCrCo BCC_A2 1336 0.01 0.36
Fe5NiMnAlCrCo BCC_A2 1479 0.00 0.36
Al2.5FeNiCrCuCo FCC_A1 1311 0.00 0.51
Al5FeNiCrCuCo BCC_A2 1262 0.09 1.25
AlCrCuFeMo0.25 BCC_A2 1708 0.17 0.27
AlCoCrCuFeMo0.25 BCC_A2 1582 0.12 0.23
AlCoCrCuFeMo0.5 BCC_A2 1768 0.20 0.25
Al0.5CoCrCuFeNiTi1.2 BCC_A2 1340 0.06 0.03
Al0.5CoCrCuFeNiTi1.4 BCC_A2 1341 0.03 0.07
Al0.5CoCrCuFeNiTi1.6 BCC_A2 1342 0.00 0.11
Al0.5CoCrCuFeNiTi1.8 BCC_A2 1341 0.14 0.15
Al0.5CoCrCuFeNiTi2 BCC_A2 1339 0.14 1.07
Ti3CrFeCoNiAl BCC_A2 1303 0.16 0.39
AlCoCrFe0.6Mo0.5Ni BCC_A2 1301 0.16 0.24
AlCoCrFeMo0.5Ni BCC_A2 1281 0.11 0.22
AlCoCrFe1.5Mo0.5Ni BCC_A2 1269 0.05 0.21
AlCoCrFe2Mo0.5Ni BCC_A2 1276 0.02 0.21
156 R. Raghavan et al. / Journal of Alloys and Compounds 544 (2012) 152158

show formation of FCC + BCC, BCC is the predominant where xi is the mole fraction of atom i. Another parameter, used
phase predicted. widely for analysis of bulk metallic glasses and used here, is the
 These results also indicate that most of the multicompo- mismatch entropy (DSr/k) formulated by Mansoori et al. [32]. Their
nent equiatomic alloys prefer to form more open BCC empirical relation for the mismatch entropy, which is a measure of
phase rather than the close packed FCC or HCP phases. the difference in sizes of the constituent atoms, is calculated by mis-
t term DSr from the solution of the PerkusYevik integral equation
5. Coupling the approaches of thermodynamic and topological and is given as:
parametric analyses and CALPHAD    
DSr 3 3 1
f2  1y1 f  1y2  f  1f  3 lnf 1  y3
k 2 2 2
Traditionally, thermodynamic analysis of phase formation in al-
loys has involved parameters like enthalpy of mixing, congura- 6
tional entropy, d parameter, etc. This approach has been where,
extended to multi-component alloys here. Enthalpy of mixing is
calculated based on the regular melt model using Miedemas ap- 1
f 7
proach [30,31] as shown below. 1f

DHmix Rni1;ij Xij ci cj 4 n is the packing fraction (equal to 0.64 for a dense random pack-
ing).The dimensionless parameters y1, y2 and y3 are calculated
Xij 4DHmix; AB is the regular melt-interaction parameter between according to following equation:
ith and jth elements, and DHmix; AB is the mixing enthalpy of binary
1
liquid alloy AB. DHmix values for binaries are obtained from y1 RnjPi1 di dj di  dj 2 ci cj 8
[31].The molar congurational entropy (DScong) of a phase is an
r3
important parameter for phase formation, given by: r2 n
y2 RjPi1 di dj di  dj 2 ci cj 9
DSconfig RRni1 xi ln xi 5 r3 2

r2 3
y3 10
r3 2
and,

rk Rni1 ci dki where k 2; 3 11

Here, di are the atomic diameter of the ith component. The values
for atomic radii are taken from the Smithells Metals Handbook.
In addition to the above parameters, the d parameter
 q
ri 2
Rini1 c i 1  R where ci is the mole fraction of atom i, ri is

the radius of atom i and R is the average atomic radius of the alloy
in
system and DSfusion(=Ri1 DSf ;i ) for all the compositions were esti-
mated with a view to study the range of values exhibited them. The
DSf ;i values are the entropy of fusion values for individual elements
of the alloy and are obtained from the Smithells Metals Handbook.
This parametric study was also done with a view to segregate
FCC and BCC phase forming compositions based on the range of
values of these parameters. Some observations of this study are
as follows:

1. It is clear from Fig. 1(a) that most of the equiatomic solid solu-
tions have DScong/DSfusion > 1, conrming the fact that high
value of DScong aids in the solid solution formation. A few equi-
atomic alloys that have their DScong/DSfusion less than 1 are
actually binary and ternary alloys (Fig. 1(a)). All non-equiatomic
solid solutions studied in this work have (DScong/DSfusion) > 1.2,
indicating that equiatomic alloys have greater tendency to form
solid solutions than the non-equiatomic alloys (Fig. 1(b)).
2. Lattice strain in multi-component alloys is inherent due to the
presence of many atoms with varying sizes. The parameters
DSr/k and d are the two topological parameters that describe
this property. Fig. 1(a) and (b) show that d parameter cannot
be used to demarcate the phases that can form in these alloys.
However, DSr/k, on the other hand, is much more sensitive to
the size difference and is much higher for BCC phase forming
compositions in both equiatomic and non-equiatomic alloys
(Fig. 2(a) and (b), respectively). Due to greater strain, which is
reected by a higher value of DSr/k, a more open structure,
Fig. 1. DScong/DSfusion vs. d plot for (a) equiatomic and (b) non-equiatomic systems. i.e., BCC is preferred than FCC. Fig. 2(a) and (b) clearly show that
 R. Raghavan et al. / Journal of Alloys and Compounds 544 (2012) 152158
Fig. 2. DScong/DSfusion vs. mismatch entropy (DSr/k) plot for (a) equiatomic and (b)
non-equiatomic systems.
we can delineate a window of DSr/k values that can lead to the
formation of BCC phase in most of the alloys under consider-
ation. Fig. 2(a) and (b) also indicate that the window of DSr/k
over which BCC solid solution forms is wider in equiatomic
alloys in comparison to non-equiatomic alloys.
3. Since multicomponent systems quite often show some devia-
tion from ideality, it would be interesting to study how DHmix,
calculated using Miedemas approach (Eq. (4)), and the one esti-
mated using the CALPHAD approach (Eq. (3)) differ from each
other. This is brought out in Fig. 3(a) and (b) for equiatomic
and non-equiatomic compositions, respectively. The values cal-
culated using the Miedemas approach (Eq. (4)) are denoted as
DHMiedema and those found using the CALPHAD approach are
denoted as DHCALPHAD. Linear ts for each type of phase formed
are plotted. It is seen that for FCC and FCC + BCC phase forming
compositions, DHMiedema and DHCALPHAD have better linear rela-
tionship than for BCC forming compositions. This is possibly
due to large strain energy in BCC forming compositions. The lin-
ear t for (FCC + BCC) has R2 value of 0.92 and 0.97 and slope of
1.22 and 1.24 for equiatomic and non-equiatomic compositions,
respectively. The linear t for BCC phase is poor (R2 = 0.64 for
equiatomic and 0.68 for non-equiatomic compositions)
whereas the t for FCC phase is much better (R2 = 0.83 and
0.91 for equiatomic and non-equiatomic cases, respectively).
157
Fig. 3. Comparison of DHMiedema and DHCALPHAD for (a) equiatomic systems and (b)
non-equiatomic systems.
4. Since DHmix is the predominant thermodynamic parameter and
DSr/k the predominant topological/strain parameter governing
phase formation, these are plotted for equiatomic compositions
in Fig. 4. In Fig. 4(a), DHMiedema is considered and in 4(b),
DHCALPHAD is considered. Interestingly, these graphs show that
DH and DSr/k have an approximately inverse relationship with
each other. It is seen that FCC phases are preferred for systems
having a high DHmix value (and correspondingly low DSr/k) and
BCC phases preferred for systems with high DSr/k and relatively
low DHmix. This clearly shows that strain is the predominant
parameter for BCC-forming compositions and enthalpy of mix-
ing is the one for FCC-forming compositions. Interestingly, sys-
tems that show a mixture of FCC and BCC phases have
intermediate values of DH and DSr/k, as expected.
6. Conclusions
1. The CALPHAD approach has been used to study the formation of
FCC and BCC phases in multicomponent alloys. It predicts BCC
forming compositions with a very good accuracy.
2. Prediction of FCC phases using CALPHAD is poor. It is believed
that this may be because of a greater presence of kinetic effects
than in BCC, which is a more open structure.
158 R. Raghavan et al. / Journal of Alloys and Compounds 544 (2012) 152158

4. BCC phase is favored when the atomic size difference is larger,

which is reected by a higher value of DSr/k. The window of
DSr/k over which BCC solid solution forms is wider in equi-
atomic alloys in comparison to non-equiatomic alloys
5. Formation of FCC phase is governed mainly by DHmix, while BCC
phase governed by DSr/k, which is representative of strain in
the structure.

References

[1] J.W. Yeh, S.K. Chen, S.J. Lin, Adv. Eng. Mater. 6 (2004) 299.
[2] X.F. Wang, Y. Zhang, Y. Qiao, G.L. Chen, Intermetallics 15 (2007) 357.
[3] C. Li, J.C. Li, M. Zhao, Q. Jiang, J. Alloys Compd. 475 (2009) 752.
[4] S. Varalakshmi, M. Kamaraj, B.S. Murty, Metall, Mater. Trans. 41A (2010) 2703.
[5] S. Varalakshmi, Ph.D. Thesis, IIT Madras (2008).
[6] S. Praveen, B.S. Murty, Ravi S. Kottada, Mater. Sci. Eng., A 534 (2012) 83.
[7] S. Singh, N. Wanderka, B.S. Murty, U. Glatzel, J. Banhart, Acta Mater. 59 (2011)
182.
[8] H.P. Chou, Y.S. Chang, S.K. Chen, J.W. Yeh, Mater. Sci. Eng., B 163 (2009) 184.
[9] B.S. Li, Y.P. Wang, M.X. Ren, C. Yang, H.Z. Fu, Mater. Sci. Eng., A 498 (2008) 482.
[10] A. Durga, K.C. Hari Kumar, B.S. Murty, Trans Ind. Inst. Metals 65 (2012) 375.
[11] K.G. Pradeep, M.Tech.Thesis, IIT Madras (2010).
[12] S. Varalakshmi, M. Kamaraj, B.S. Murty, J. Alloys Compd. 460 (2008) 253.
[13] Z. Yong, Z.Y. Jun, H.X. Dong, W.M. Ling, C.G. Liang, Sci. China, Ser. G 51 (2008)
427.
[14] S. Praveen, Private Communication.
[15] J.M. Wu, S.J. Lin, J.W. Yeh, S.K. Chen, Y.S. Huang, H.C. Chen, Wear 261 (2006)
513.
[16] Y.J. Zhang, J.P. Zhou, Lin, G.L. Chen, P.K. Liaw, Adv. Eng. Mater. 10 (2008) 534.
[17] T.T. Shun, C.H. Hung, C.F. Lee, J. Alloys Compd. 493 (2010) 105.
[18] C.M. Lin, H.L. Tsai, J. Alloys Compd. 489 (2010) 30.
[19] F.J. Wang, Y. Zhang, G.L. Chen, J. Alloys Compd. 478 (2009) 321.
[20] M.R. Chen, S.J. Lin, J.W. Yeh, S.K. Chen, Y.S. Huang, C.P. Tu, Mater. Trans. 47
(2006) 1395.
[21] Y.L. Chen, Y.H. Hu, C.A. Hsieh, J.W. Yeh, S.K. Chen, J. Alloys Compd. 481 (2009)
768.
[22] S.T. Chen, Y.F. Kuo, S.Y. Chen, C.H. Tsau, T.T. Shun, J.W. Yeh, Mater. Sci. Eng., A
527 (2010) 5818.
[23] T.T. Shun, C.H. Hung, C.F. Lee, J. Alloys Compd. 495 (2010) 55.
[24] J.W. Yeh, S.K. Chen, J.Y. Gan, S.J. Lin, Metall. Mater. Trans. A 35 (2004) 2533.
[25] F.J. Wang, Y. Zhang, Mater. Sci. Eng., A 496 (2008) 214.
[26] R. Sriharitha, B.S. Murty and Ravi Sankar Kottada, Intermetallics, Under review.
[27] C.Y. Hsu, T.S. Sheu, J.W. Yeh, S.K. Chen, Wear 268 (2010) 653.
[28] M. Palumbo, L. Battezzati, CALPHAD 32 (2008) 295.
[29] Y.M. Muggianu, M. Gambino, J.P. Bros, J. Chem. Phys. 72 (1975) 83.
[30] A.R. Miedema, P.F. de Chatel, F.R. de Boer, Physica B 100 (1980) 1.
[31] A. Takeuchi, A. Inoue, Mater. Sci. Eng., A 304306 (2001) 446.
Fig. 4. Comparison of DSr/k with (a) DHMiedema and (b) DHCALPHAD for equiatomic [32] G.A. Mansoori, N.F. Carnahan, K.E. Starling, T.W. Leland Jr., J. Chem. Phys. 54
systems. (1971) 1523.

3. Solid solution formation in multicomponent alloys is favored

when the ratio of DScong/DSfusion is greater than 1 and 1.2 for
equiatomic and non-equiatomic alloys, respectively.