See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/233780013

Thermodynamic description of the Cu-Sn system

Article · November 2007

DOI: 10.1557/jmr.2007.0396

CITATIONS READS

36 1,366

3 authors:

Wojciech Gierlotka Sinn-Wen Chen

National Dong Hwa University National Tsing Hua University
95 PUBLICATIONS   835 CITATIONS    230 PUBLICATIONS   5,103 CITATIONS   

SEE PROFILE SEE PROFILE

Shih-kang Lin
National Cheng Kung University
103 PUBLICATIONS   1,176 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Lead-free solder database View project

Effect of electric currents on materials phase stability View project

All content following this page was uploaded by Shih-kang Lin on 29 December 2014.

The user has requested enhancement of the downloaded file.

 Thermodynamic description of the Cu–Sn system
Wojcieh Gierlotka
Department of Chemical Engineering, National Tsing Hua University, Hsin-Chu 300, Taiwan; and
Non-Ferrous Metals Department, Akademia Górniczo-Hutnicza (AGH) University of Science and
Technology, Mickiewicza 30, 30-059 Krakow, Poland
Sinn-wen Chena) and Shih-kang Lin
Department of Chemical Engineering, National Tsing Hua University, Hsin-Chu 300, Taiwan

(Received 2 January 2007; accepted 25 July 2007)

The Cu–Sn binary system is important for various applications, especially for recent
developments in the electronics packaging industry. The ⑀-Cu3Sn and ␩-Cu6Sn5
(␩⬘ phases) phases are frequently encountered in electronics products. However, the
two phases have been described as line compounds in previous thermodynamic
modeling, and their compositional homogeneities were not considered. In this study,
the thermodynamic properties of the Cu–Sn binary system are modeled and the phase
diagram is calculated by the CALPHAD method, using experimental information
reported in the literature. The ⑀ and ␩ (␩⬘) phases are described using compound
energy models with two and three sublattices, respectively, so that their compositional
homogeneities could be calculated. Good agreement was observed between the
calculated result and the existing experimental data.

I. INTRODUCTION descriptions of the Cu–Sn-based multicomponent sys-

The binary Cu–Sn system is the basis for various tems, a sound description of the Cu–Sn binary system is
emerging Pb-free solders and for most of the important necessary. The binary Sn–Cu system has been assessed
solder-related materials systems.1–3 With the increasing by various investigators.13,14 However, in the previous
demand for Pb-free solders and more advanced applica- literature,13,14 the ⑀-Cu3Sn and ␩-Cu6Sn5 phases are de-
tions of soldering technology in electronic products, the scribed using the line compound models without compo-
Cu–Sn system has become one of the most important sitional homogeneity. In contrast, the current study ther-
material systems.4–6 There are seven intermetallic phases modynamically models the Cu–Sn system using the
in the Cu–Sn system. Of these, the ⑀-Cu 3 Sn and CALPHAD approach. For accurate descriptions of the
␩-Cu6Sn5 (␩⬘-Cu6Sn5) phases are on the more Sn-rich binary system and the higher-order systems, the interme-
side and are most frequently encountered in the soldering tallic compounds are described as phases with solubili-
studies. These two compounds are also observed in other ties using compound energy models.
solder-related systems, such as Sn–Ag–Cu, Sn–Cu–Ni,
and Sn–Ag–Cu–Ni systems.1–10 The compositional ho-
II. THERMODYNAMIC DESCRIPTION OF
mogeneity ranges of these two intermetallic compounds
THE PHASES
are significant in some of the Cu–Sn-based higher-order
systems.7–10 The following phases are considered in this work:
Phase diagrams of multicomponent systems are valu- FCC_A1, BCT_A5, BCC_A2, liquid, ␥, ␨, ␦, ⑀, ␩, ␩⬘.
able for a basic understanding of materials processing Detailed information about these phases is given in Table
and materials properties. However, thorough experi- I and below.
mental determinations of the phase diagrams of multi-
component systems are very time-consuming; it is much A. Substitutional solution—FCC_A1, BCC_A2,
more feasible to determine the phase diagrams with the BCT_A5, liquid
CALPHAD approach.11,12 To have good thermodynamic
The Gibbs free energies of pure elements with respect
to temperature oGi(T) ⳱ Gi(T) − HSER
i are represented by
Eq. (1):
a)
Address all correspondence to this author.
e-mail: swchen@che.nthu.edu.tw
o
Gi共T兲 = a + bT + cT ln共T兲 + dT2 + eT −1 + fT 3 + iT 4
DOI: 10.1557/JMR.2007.0396 + jT7 + kT −9 . (1)

3158 J. Mater. Res., Vol. 22, No. 11, Nov 2007 © 2007 Materials Research Society
 W. Gierlotka et al.: Thermodynamic description of the Cu–Sn system

TABLE I. Phase of the Cu–Sn system. The oGi(T) functions are taken from Dinsdale16 and from
Phases Models Structure/space group SGTE Unary (Pure elements) TDB v.4.17 The thermo-
dynamic functions of pure elements are listed in Table II.
Liquid Substitional, RKP
The main differences between the functions by Dinsdale
FCC_A1 Substitional, RKP A1 Fm3̄m
BCT_A5 Substitional, RKP A5 I41/amd and SGTE Unary TDB v.4 are the descriptions of tin in
BCC_A2 Substitional, RKP A2 Im3̄m the FCC structure. It was decided to use the newest in-
␥ Two sublattice model D03 Fm3̄m formation taken from the SGTE database. Solid- and liq-
␨ Stoichiometric compound P63 uid-solution phases (FCC_A1, BCC_A2, BCT_A5, Liq-
␦ Stoichiometric compound F43m
uid) are described by the regular solution model11:
⑀ Two-sublattice model Cmcm
␩
␩⬘
Three-sublattice model
Three-sublattice model
B81 P63/mmc
⭈⭈⭈
Gm共T兲 = 冱x G 共T兲 + RT 冱x ln共x 兲
i
o
i i i

冉 冊
i i

RKP, Redlich–Kister polynomial for binary substitutional solutions. + 冱 冱x x 冱 L 共x − x 兲 i j

␯
ij i j
␯
, (2)
i j⬎i ␯

The oGi(T) data are referred to the constant enthalpy

where the
value of the standard element reference HSER at 298.15 K
冱 冱x x 冉 冱 L 共x − x 兲 冊
i
and 1 bar as recommended by Scientific Group Thermo- ␯ ␯
data Europe (SGTE).15,16 The reference states are: i j⬎i
i j
␯
ij i j

FCC_A1 (Cu), BCT_A5 (Sn). The oGi(T) expression

may be given for several temperature ranges, where the part is the Redlich–Kister polynomial for excess Gibbs
coefficients a, b, c, d, e, f, i, j, and k have different values. free energy.

TABLE II. Gibbs free energies of pure elements.16,17

Function Temperature range Expression

GHSERCU 298.15 < T < 1357.77 −7770.458 + 130.485235*T − 24.112392*T* ln(T) − 0.00265684*T**2 + 1.29223E-07*T**3
+ 52,478*T**(−1)
1357.77 < T < 3200.00 −13,542.026 + 183.803828*T − 31.38*T* ln (T) + 3.64167E + 29*T**(−9)
GHSERSN 100.00 < T < 250.00 −7958.517 + 122.765451*T − 25.858*T* ln (T) + 5.1185E-04*T**2 − 3.192767E-06*T**3 +
18440*T**(−1)
250.00 < T < 505.08 −5855.135 + 65.443315*T − 15.961*T* ln (T) − 0.0188702*T**2 + 3.121167E-06*T**3 −
61,960*T**(−1)
505.08 < T < 800.00 +2524.724 + 4.005269*T − 8.2590486*T* ln (T) − 0.016814429*T**2 + 2.623131E-06*T**3
− 1,081,244*T**(−1) − 1.2307E + 25*T**(−9)
800.00 < T < 3000.00 −8256.959 + 138.99688*T − 28.4512*T* ln (T) − 1.2307E + 25*T**(−9)
GCUBCC 298.15 < T < 1357.77 −3753.458 + 129.230235*T − 24.112392*T* ln (T) − 0.00265684*T**2 + 1.29223E
−07*T**3 + 52,478*T**(−1)
1357.77 < T < 3200.00 −9525.026 + 182.548828*T − 31.38*T* ln (T) + 3.64167E + 29*T**(−9)
GSNBCC 100.00 < T < 250.00 −3558.517 + 116.765451*T − 25.858*T* ln (T) + 5.1185E − 04*T**2 − 3.192767E −
06*T**3 + 18440*T**(−1)
250.00 < T < 505.08 −1455.135 + 59.443315*T − 15.961*T* ln (T) − 0.0188702*T**2 + 3.121167E − 06*T**3 −
61960*T**(−1)
505.08 < T < 800.00 +6924.724 − 1.994731*T − 8.2590486*T* ln (T) − 0.016814429*T**2 + 2.623131E −
06*T**3 − 1,081,244*T**(−1) − 1.2307E + 25*T**(−9)
800.00 < T < 3000.00 −3856.969 + 132.99688*T − 28.4512*T* ln (T) − 1.2307E + 25*T**(−9)
GSNFCC 298.15 < T < 505.08 −345.135 + 56.983315*T − 15.961*T* ln (T) − 0.0188702*T**2 + 3.121167E − 06*T**3 −
61,960*T**(−1)
505.08 < T < 800.00 +8034.724 − 4.454731*T − 8.2590486*T* ln (T) − 0.016814429*T**2 + 2.623131E −
06*T**3 − 1,081,244*T**(−1) − 1.2307E + 25*T**(−9)
800.00 < T < 3000.00 −2746.959 + 130.53688*T − 28.4512*T* ln (T) − 1.2307E + 25*T**(−9)
GCULIQ 298.15 < T < 1357.77 +5194.277 + 120.973331*T − 24.112392*T* ln (T) − 0.00265684*T**2 + 1.29223E −
07*T**3 + 52,478*T**(−1) − 5.8489E − 21*T**7
1357.77 < T < 3200.00 −46.545 + 173.881484*T − 31.38*T* ln (T)
GSNLIQ 100.00 < T < 250.00 −855.425 + 108.677684*T − 25.858*T* ln (T) + 5.1185E − 04*T**2 − 3.192767E − 06*T**3
+ 18,440*T**(−1) + 1.47031E − 18*T**7
250.00 < T < 505.08 +1247.957 + 51.355548*T − 15.961*T* ln (T) − 0.0188702*T**2 + 3.121167E − 06*T**3 −
61960*T**(−1) + 1.47031E − 18*T**7
505.08 < T < 800.00 +9496.31 − 9.809114*T − 8.2590486*T* ln (T) − 0.016814429*T**2 + 2.623131E − 06*T**3
− 1,081,244*T**(−1)
800.00 < T < 3000.00 −1285.372 + 125.182498*T − 28.4512*T* ln (T)

J. Mater. Res., Vol. 22, No. 11, Nov 2007 3159

 W. Gierlotka et al.: Thermodynamic description of the Cu–Sn system

B. Stoichiometric compounds ␦ and ␨ E. ␩ and ␩ⴕ phases

Binary stoichiometric compounds ␦-Cu41Sn11 and The ␩⬘ phase is hexagonal and has a superstructure
␦-Cu10Sn3 are described as the line compound using the based on a NiAs-type structure. The ␩⬘ transforms to the
following: ␩ phase, which has the NiAs structure between 459 and
502 K.16,19 Similar to the ⑀ phase as mentioned previ-
Gm共T兲 = a + bT + 冱x GHSER
i
i i . (3) ously, although the compositional homogeneities of the
␩ and ␩⬘ phases are small in the binary Cu–Sn binary
system, thermodynamic models describing their com-
C. ␥ phase positional homogeneities are important since their
The ␥ (D03) phase is formed by an ordering process of compositional homogeneities might be significant in
the BCC_A2 phase. A description of the Gibbs free en- the Cu–Sn-based higher-order systems. Both phases
ergy of this phase was proposed by Shim13 as a two were modeled using a three-sublattice model
sublattice model, (Cu,Sn)0.75(Sn,Cu)0.25, and is repre- (Cu)0.3333(Cu,Sn)0.3334(Sn)0.3333. This kind of model was
sented by Eq. (4): used by Liu et al.20 for modeling the Ni3Sn2 high-
temperature phase. Ghosh21 used a different sublattice
G␥m共T兲 = YCu
I II o ␥
YCu GCu:Cu + YCuI II o ␥
YSn GCu:Sn model for the description of the same Ni3Sn2 phase. In
II o ␥ II o ␥ his work, ␩-Ni3Sn2 phase was described by using a three-
+ YSnYCu GSn:Cu + YSnYSn GSn:Sn
I I
sublattice model (Ni)1:(Sn)1:(Ni,Va). Take into account
+ 0.75RT共YCu I I
ln YCu + YSnI I
ln YSn 兲 the existence of the ␩ phase in a larger composition range
+ 0.25RT共YSn ln YSn + YCu ln YCu兲 +
II II II II xs
G␥m , and the more recent literature information, the model
(4) proposed by Liu et al.20 is adopted. The Gibbs free en-
ergies of both phases are given by the following equa-
where xs
G␥m represents excess Gibbs free energy: tions:
xs
G␥m = YCu
I I
YSn II ␥
YCu LCu,Sn:Cu + YCu
I I
YSn II ␥
YSn LCu,Sn:Sn G ␩m共T兲 = YCu
I II IIIo ␩
YCu YSn GCu:Cu:Sn + YCu
I II IIIo ␩
YSn YSn GCu:Sn
II II ␥ II II ␥
+ YCu
I
YSn YCuLCu:Sn,Cu + YSnI
YSn YCuLSn:Sn,Cu + 0.3334RT共YCuII II
ln YCu + YSn
II II
ln YSn 兲
II II ␥ xs ␩
+ YCu
I I
YSn YSn YCuLCu,Sn:Sn,Cu , (5) + Gm , (7)

in which YN i denotes the site fraction of element i on

xs
G ␩m = YCu
I II II III ␩
YCu YSnYSnLCu:Cu,Sn:Sn . (8)
sublattice N, a colon (:) indicates separation of elements
on the different sublattices, and a comma (,) indicates III. EXPERIMENTAL INFORMATION
separation of elements on the same sublattice.
A. Phase diagram data
D. ⑀ phase A phase diagram of Cu–Sn system was recently sug-
gested by Saunders and Miodownik18 and assessed by
The ⑀ phase in the Cu–Sn binary system is character- Shim et al.13 and Liu et al.14 Various investigations on
ized by a low solubility limit; however, experimental re- phase equilibria were conducted22–30 by using x-ray
sults show that this limit is not neglected.18 More impor- analysis, micrographic analysis, thermal analysis, etc.
tantly, the ⑀ phase is frequently encountered in the Cu– Most of the earlier works are summarized by Hansen and
Sn-based systems, such as in the Sn–Ag–Cu, Sn–Cu–Ni, Anderko,31 while a more recent work on the ␩⬘ phase is
and Sn–Ag–Cu–Ni systems,4–10 and its compositional by Larsson et al.32 Most of the experimental results are in
homogeneities are usually significant in these systems. good agreement, although inconsistencies were found in
Two sublattices could be assumed for the phase, with Cu the boundaries of the ␥ phase.
atoms occupying the first sublattice and second sublattice
is occupied by Sn atoms. To model the stoichiometric B. Thermochemical data
range, it is assumed that Sn atoms occupy Cu sites and
The enthalpies of mixing of liquid Cu–Sn alloys were
Cu atoms occupy Sn sites as antisite defects. The
obtained using calorimetry by Kleppa,33 Hultgren et
⑀-Cu3Sn phase is thus described by two-sublattice com-
al.,34 Takeuchi et al.,35 Yazawa and Itagaki,36 Iguchi et
pound energy formalism [Eq. (6)]:
al.,37 Pool et al.,38 and Lee et al.39 Kleppa33 determined
the enthalpy of solutions of Cu in liquid alloys at 723 K.
G⑀m共T兲 = YCu
I II o ⑀
YCu GCu:Cu + YCuI II o ⑀
YSn GCu:Sn
II o ⑀ II o ⑀ Hultgren et al.34 measured the enthalpy of mixing at
+ YSnYCu GSn:Cu + YSnYSn GSn:Sn
I I
1400 K and analyzed earlier works. Takeuchi et al.,35
+ 0.75RT共YCu I I
ln YCu + YSnI I
ln YSn 兲 Yazawa and Itagaki,35 and Iguchi et al.37 made calori-
+ 0.25RT共YSn ln YSn + YCu ln YCu兲 +
II II II II xs
G⑀m . metric investigation at 1363, 1373, and 1393 K, respec-
(6) tively. Pool et al.’s work described the application of the

3160 J. Mater. Res., Vol. 22, No. 11, Nov 2007

 W. Gierlotka et al.: Thermodynamic description of the Cu–Sn system

SETARAM high-temperature calorimeter to determine For this optimization, thermodynamic data for the liquid
the mixing enthalpies of liquid alloys, with the Cu–Sn phase, invariant reactions, and liquidus/solidus were
alloy was one of the examined systems.38 Lee et al.39 used. Each piece of the selected information was given a
determined the enthalpy of mixing of liquid Cu–Sn sys- certain weight based on personal judgment. The optimi-
tem at 997 K. Additionally, he suggested the existence of zation was carried out step by step. First, the optimiza-
associate Cu3Sn in melt, although this was not confirmed tion of the liquid phase was performed, and then the solid
by other authors. Kleppa33 also measured enthalpy of phases were assessed. All parameters were finally evalu-
formation of solid alloys for the compositions between ated together to provide the best description of the sys-
0.074 and 0.255 mol fraction of Sn at 723 K. tem. The calculated interaction parameters are shown in
The activities of Cu and Sn were measured by Alcock Table III.
et al.,40 Hager et al.,41 Ono et al.,42 Oshi et al.,43 and
Sengupta et al.44 Of these studies, Alcock et al.,40 Hager V. RESULTS
et al.,41 and Ono et al.42 used a Knudsen cell technique
Figure 1(a) shows the calculated Cu–Sn phase dia-
and measured the activities at 1300, 1593, and 1573 K,
gram. Compared with the experimental data,22,25,28–30 as
respectively. Electromotive force (EMF) measurements
shown in Fig. 1(b), the calculated diagram is in good
with the electrochemical cell technique were used by
agreement with the experimentally determined phases
Oshi et al.43 between 1173 and 1373 K and Sengupta et
boundaries. In addition, the invariant reaction tempera-
al.44 at 1073 K. All authors obtained activities of Cu and
tures calculated in this study are in good agreement with
Sn with negative deviation from Raoult’s Law. Alcock
the experimental data listed in Table III. As mentioned
and Jacob45 used a gas–solid equilibration technique to
previously, the ⑀-Cu3Sn and ␩/␩⬘-Cu6Sn5 phases are
measure the chemical potential of Sn in the terminal solid
relatively stable and are usually formed at Cu/Sn alloy
solution of Cu-side at 1000 K. Sommer et al.46 and
interfaces.1–10 However, in the previous optimizations of
Predel and Schallner47 used an EMF technique to meas-
the Cu–Sn system11–13,18 the important intermetallic
ure the same variable at 1000 K.
compounds, ⑀-Cu3Sn and ␩/␩⬘-Cu6Sn5, were treated as
C. Invariant reactions line compounds. This brought certain difficulties in the
assessment of the corresponding higher-order systems
The Cu–Sn phase diagram reported by Saunders and
since those phases exhibit considerable ternary solubili-
Miodownik18 has three peritectics, two peritectoids, four
ties in many systems.4–10 The Cu–Sn system was suc-
eutectoids, one metatectic, one congruent transformation,
cessfully optimized with suitable homogeneous ranges in
and one eutectic reaction. All data concerning invariant
the ⑀-Cu3Sn and ␩/␩⬘-Cu6Sn5 phases.
reactions used for the assessment procedure are listed in
The liquidus line of Cu–Sn system was fitted well with
Table III.
the previous optimization of liquid phase by Shim et al.13
and Liu et al.14 However, both proposed models have
IV. MODELING PROCEDURE independent on temperature enthalpy of mixing, and they
The thermodynamic parameters for all phases in the do not match the thermodynamic properties experimental
system were optimized using ThermoCalc software.48 measured at different temperatures.33–44 Therefore, the
TABLE III. Invariant reaction data and the calculated results.

Experimental results Calculated results

Composition (at.%) Composition (at.%)
Reaction Type T(K) Cu Sn T(K) Cu Sn
FCC_Al + L ↔ BCC_A2 Peritectic 1071 87.6 12.4 1070.65 87.0 13.0
BCC_A2 + L ↔ ␥ Peritectic 1028 83.3 16.7 1024.46 82.6 17.4
⑀↔␥ Congruent melting 949 75.0 25.0 949.45 75.3 24.7
␥+⑀↔␨ Peritectoid 913 77.5 22.5 911.50 76.9 23.1
␥+⑀↔L Metatectic 913 57.5 42.5 914.08 57 43.0
␥+␨↔␦ Peritectoid 863 79.7 20.3 866.32 78.9 21.1
BCC_A2 ↔ FCC_A1 + ␥ Eutectoid 859 85.6 14.4 858.82 85.6 14.4
␨↔␦+⑀ Eutectoid 855 78.3 21.7 854.94 77.5 22.5
␥ ↔ FCC_A1 + ␦ Eutectoid 793 83.2 16.8 786.43 84.0 16.0
⑀+L↔␩ Peritectic 688 54.5 45.5 688.60 56.7 43.3
␦ ↔ FCC_A1 + ⑀ Eutectoid 623 79.6 20.4 623.90 78.9 21.1
L ↔ ␩ + BCT_A5 Eutectic 500 1.6 98.4 497.98 1.7 98.3
␩ ↔ ␩⬘ Simple/Long-range order 459–462 55.5–54.9 44.5–45.1 461.05–461.08 55.4–54.2 44.6–45.8

Source: Ref. 14.

J. Mater. Res., Vol. 22, No. 11, Nov 2007 3161

 W. Gierlotka et al.: Thermodynamic description of the Cu–Sn system

TABLE IV. Interaction parameters obtained during optimization

procedure.

Phase Parameter

Liquid LCu,Sn ⳱ −15,583.768 + 40.6239T −

0 liquid
5.6172T ln(T)
LCu,Sn ⳱ −27,739.285 + 41.9270T −
1 liquid
4.4641T ln (T)
LCu,Sn ⳱ −13,228.943 + 0.906T
2 liquid

FCC_A1 LCu,Sn ⳱ −10,309.845 + 1.158T

0 FCC_Al

LCu,Sn ⳱ −16,190.032 + 6.495T

1 FCC_A1

BCC_A2 LCu,Sn ⳱ −41,774.195 + 47.877T

0 BCC_A2

LCu,Sn ⳱ −12,316.502 − 50.445T

1 BCC_A2

BCT_A5 0 BCT_A5
GSn ⳱ GHSERSN
0 BCT_A5
GCu ⳱0
LCu,Sn ⳱ 0
0 BCT_A5

␨ ⌬G␨m ⳱ −7080.194 − 0.955T

␦ ⌬G␦m ⳱ −6498.086 − 1.064T
0 ␥
␥ GCu:Cu ⳱ GCUBCC
0 ␥
GSn:Sn ⳱ GSNBCC
0 ␥
GCu:Sn ⳱ −5912.468 − 2.439T + .75 GHSERCU + .25
GHSERSN
0 ␥
GSn:Cu ⳱ 126,652.750 − 2.987T + .25 GHSERCU + .75
GHSERSN
0 ␥
LCu:Cu,Sn ⳱ −2672.875 − 1.792T
1 ␥
LCu:Cu,Sn ⳱ 57.375 − 3.767T
0 ␥
LCu,Sn:Sn ⳱ 46,820.857 − 42.261T
0 ⑀
⑀ GCu:Cu ⳱ 5534.454 − 0.241T + GHSERCU
0 ⑀
GSn:Sn ⳱ 14,564.708 + 0.100T + GHSERSN
0 ⑀
GCu:Sn ⳱ −8406.738 + 0.100T + .75 GHSERCU + .25
GHSERSN
0 ⑀
GSn:Cu ⳱ 15,470.377 + 7.731T + .25 GHSERCU + .75
GHSERSN
0 ⑀
L*:Cu,Sn ⳱ 217.494 − 1.831T
0 ⑀
LCu,Sn:* ⳱ 0.095 + 0.1T
0 ␩
␩ GCu:Cu:Sn ⳱ −5570.949 − 1.1108T + .6667 GHSERCU +
.3333 GHSERSN
0 ␩
GCu:Sn:Sn ⳱ 930.597 − .4804T + .3333 GHSERCU +
.6667 GHSERSN
0 ␩
LCu:Cu,Sn:Sn ⳱ −7841.425 − .7127T
0 ␩_prime
␩_prime GCu:Cu:Sn ⳱ −4297.175 − 3.9531T + .6667 GHSERCU + FIG. 1. (a) Calculated phase diagram of the Cu–Sn system. (b) Cal-
.3333 GHSERSN culated phase diagram of the Cu–Sn system and the superimposed
0 ␩_prime
GCu:Sn:Sn ⳱ 5334.887 − 10.2934T + .3333 GHSERCU + experimental data.
.6667 GHSERSN
0 ␩_prime
LCu:Cu,Sn:Sn ⳱ −21100.152 + 28.6728T

temperature dependence of enthalpies in liquid phase

was considered in the current study. Figure 2 shows the
calculated and experimental determined enthalpies of
mixing in the liquid phase of the Cu–Sn system. The
experimental values obtained by Yazawa and Itagaki at
1373 K36 and Lee et al. at 997 K39 exhibit slight tem-
perature dependency and are well fitted by the proposed
model. Figures 3 and 4 show the liquid phase activities of
Sn and Cu, respectively. The experimental values ob-
tained from Ono et al. at 1573 K42 and Sengupta et al. at
1073 K44 show very little dependency on temperature
and are again well simulated by the calculation results.
Precise descriptions of the temperature dependency of
thermodynamic properties are of essential importance, FIG. 2. Calculated enthalpies of mixing of the Cu–Sn system at 997
since this precision not only improves the accuracy of and 1373 K and the superimposed experimental results. Both compo-
optimization for thermodynamic properties, but it is also nents referred to liquid.

3162 J. Mater. Res., Vol. 22, No. 11, Nov 2007

 W. Gierlotka et al.: Thermodynamic description of the Cu–Sn system

FIG. 3. Calculated activities of Sn at 1073 and 1573 K and the su- FIG. 5. Calculated activity of Sn in solid state at 1000 K where Sn
perimposed experimental results. Both components referred to liquid. referred to liquid and Cu referred FCC_A1 structure and the superim-
posed experimental data.

FIG. 4. Calculated activities of Cu at 1073 and 1573 K. Both com-

ponents referred to liquid.
FIG. 6. Calculated enthalpies of formation at 723 and 237 K where Sn
referred to BCT_A5 and Cu referred FCC_A1 structure and the su-
perimposed experimental data.
beneficial to have good descriptions of mixing enthalpies
in higher order systems.
The activities of Sn in solid state at 1000 K were
experimentally determined by Alcock and Jacob45 and work of Shim et al.13 Liu et al.’s results shows similar
Sommer et al.46 As shown in Fig. 5, the activities calcu- good simulation of activity of Sn in solid state as in the
lated in this study, using the liquid phase and FCC_A1 present optimization. However, more parameters were
structure as the reference states for Sn and Cu, are in used in Liu et al.’s optimization.14 It is better to use as
good agreement with the experimental values. Compared few numbers of parameters as possible to fit every ex-
with the calculations based on previous optimization perimental result. The modification of thermodynamic
done by Shim et al.,13 the present calculation is better modeling in Cu phase in this study improved the simu-
fitted with the results of Alcock and Jacob,45 while the lation of activity of Sn in solid state by limited param-
former shows lower values toward both experimental re- eters as shown in Fig. 5. In addition, this study adopts the
sults.45,46 The thermodynamic description of the Cu ␤/␥ first-order transformation as suggested by most pre-
phase was also assessed by Liu et al.14 following the vious research groups,13,15,18 but not the A2–B2–D03

J. Mater. Res., Vol. 22, No. 11, Nov 2007 3163

 W. Gierlotka et al.: Thermodynamic description of the Cu–Sn system

two-stage ordering reaction by Liu et al.14 The ␤/␥ first- mechanical properties of Sn–In and Sn–Bi solders. JOM 45(7), 25
order transformation mechanism agrees better with ex- (1993).
8. W. Koster, T. Godecke, and D. Heine: The constitution of the
perimental data,49 and the proposed description is rela- copper–indium–tin system in the range from 100 to 50 at.% Cu. Z.
tively easier for extension to higher-order systems as Metallkd. 63, 820 (1972).
well. Figure 6 shows both the calculated and measured 9. H. Enoki, K. Ishida, and T. Nishizawa: Phase equilibria in cobalt-
formation enthalpies of various phases at 723 and 237 K. rich portions of the Co–Si and Co–Ge systems. J. Less-Common
The experimental data was obtained from Kleppa33 and Met. 160, 153 (1990).
10. C-N. Chiu, Y-C. Huang, A-R. Zi, and S-W. Chen: Isoplethal
Gangulee et al.19 The calculation results are in good
sections of the liquidus projection and the 250 °C phase equilibria
agreement with the experimental data. The enthalpies of of the Sn–Ag–Cu–Ni quaternary system at the Sn-rich corner.
formation were also calculated and well fitted by former Mater. Trans. 46, 2426 (2005).
optimization at 723 K.13 However, the present calcula- 11. N. Saunders and P. Miodownik: Calphad, Calculation of Phase
tion demonstrates very good simulation for both 723 and Diagrams a Comprehensive Guide (Elsevier Science, Oxford,
UK, 1998).
273 K as shown in Fig. 6.
12. H. Okamoto and T.B. Massalski: Guidelines for binary phase
diagrams assessment. J. Phase Eq. 14, 316 (1993).
13. J-H. Shim, C-S. Oh, B-J. Lee, and D.N. Lee: Thermodynamic
VI. CONCLUSIONS assessment of the Cu–Sn system. Z. Metallkd. 87, 205 (1996).
Phase equilibria of the Cu–Sn binary system have been 14. X.J. Liu, C.P. Wang, I. Ohnuma, R. Kainuma, and K. Ishida:
Experimental investigation and thermodynamic calculation of the
modeled on the basis of experimental data taken from phase equilibria in the Cu–Sn and Cu–Sn–Mn systems. Metall.
literature. The Cu–Sn phase diagram was calculated with Mater. Trans. A 35(6), 1641 (2004).
suitable homogeneous ranges in the ⑀-Cu3Sn and ␩/␩⬘- 15. The S.G.T.E. Substance Database, version 1997, SGTE Group
Cu6Sn5 phases. Thermodynamic descriptions of every (Grenoble, France, 1997).
phase in the Cu–Sn system were reoptimized, and the 16. A. Dinsdale: SGTE data for pure elements. Calphad 15, 317
(1991).
calculated properties are in good agreement with the ex- 17. PURE 4.4 SGTE Pure Elements (Unary) Database (Scientific
perimental determined results. Especially for the liquid Group Thermodata Europe, 1991–2006).
phase, the temperature dependency was considered, 18. N. Saunders and A.P. Miodownik: The Cu–Sn (copper–tin) sys-
which improves the flexibility for assessments on higher- tem. Bull. Alloy Phase Diagrams 11, 278 (1990).
order systems. The modifications result in very good 19. A. Gangulee, G.C. Das, and M.B. Bever: An x-ray. diffraction and
calorimetric investigation of the compound Cu6Sn5. Metall.
matching between calculated and experimental measured
Trans. 4, 2063 (1973).
thermodynamic values in both liquid and solid states. 20. H.S. Liu, J. Wang, and Z.P. Jin: Thermodynamic optimization of
the Ni–Sn binary system. Calphad 28, 363 (2004).
21. G. Ghosh: Thermodynamic modeling of the nickel–lead–tin sys-
ACKNOWLEDGMENT tem. Metall. Mat. Trans. A 30, 1481 (1999).
22. M. Hamasumi and N. Takamoto: A further research on the Cu–Sn
The authors gracefully acknowledge financial support system. Nippon Kinzoku Gakkaishi 1, 251 (1937).
from the National Science Council of Taiwan (NSC95- 23. M. Hamasumi: On the equilibrium diagram of Cu–Sn system.
2221-E-007-205). Nippon Kinzoku Gakkaishi 2, 147 (1938).
24. C.T. Heycock and F.H. Neville: Complete freezing-point curves
of binary alloys containing silver or copper together with another
REFERENCES metal. Philos. Trans. R. Soc. London, Ser. A 189, 25 (1897).
25. O. Bauer and O. Vollenbruck: Thermal analysis of the Cu–Sn
1. J. Takemoto, A. Matsunawa, and M. Takahashi: Tensile test for binary system. Z. Metallkd. 15, 119 (1923).
estimation of thermal fatigue properties of solder alloys. J. Mater. 26. D. Stockdale: The alpha-phase boundary in the copper–tin system.
Sci. 32(15), 4077 (1997). J. Inst. Met. 34, 111 (1925).
2. C.B. Lee, S.B. Jung, Y.E. Shina, and C.C. Shur: The effect of Bi 27. A.R. Raper: The equilibrium diagram of copper-tin alloys con-
concentration on wettability of Cu substrate by Sn–Bi solders. taining from 10 to 25 atomic percent of tin. J. Inst. Met. 38, 217
Mater. Trans. 42, 751 (2001). (1927).
3. W. Yang and R.W. Meddler: Microstructure evolution of eutectic 28. J. Vero: Aufbau der Zinnbronzen (in German). Z. Anorg. Chem.
Sn–Ag solder joints. J. Electron. Mater. 23, 765 (1994). 218, 402 (1934).
4. Y. Takatu, X. Liu, I. Ohnuma, R. Kainuma, and K. Ishida: Inter- 29. C. Haase and F. Pawlek: Copper–tin alloys. Z. Metallkd. 28, 73
facial reaction and morphology between molten Sn base solders (1936).
and Cu substrate. Mater. Trans. 45(3), 646 (2004). 30. B.D. Bastow and D.H. Kirkwood: Solid/liquid equilibrium in the
5. C-H. Lin, S-W. Chen, and C-H. Wang: Phase equilibria and so- copper-nickel-tin system determined by microprobe analysis.
lidification properties of Sn-Cu-Ni alloys. J. Electron. Mater. J. Inst. Met. 99, 277 (1971).
32(9), 907 (2002). 31. M. Hansen and K. Anderko: Constitution of Binary Alloys
6. X.J. Liu, H.S. Liu, I. Ohnuma, R. Kainuma, K. Ishida, S. Itabashi, (McGraw-Hill, New York, 1958).
K. Kameda, and K. Yamaguchi: Experimental determination and 32. A-K. Larsson, L. Stenberg, and S. Lidin: The superstructure of
thermodynamic calculation of the phase equilibria in the domain-twinned ␩⬘-Cu6Sn5. Acta Crystallogr., Sect. B 50, 636
Cu–In–Sn system. J. Electron. Mater. 30, 1093 (2001). (1994).
7. J.W. Morris, J.L.F. Goldstein, and Z. Mei: Microstructure and 33. O.J. Kleppa: A Calorimetric investigation of some binary and

3164 J. Mater. Res., Vol. 22, No. 11, Nov 2007

 W. Gierlotka et al.: Thermodynamic description of the Cu–Sn system

ternary liquid alloys rich in tin. J. Phys. Chem. 60, 842, 852–858 erties of the Cu–Sn and Cu–Au systems by mass spectrometry.
(1956). Metall. Trans. 1, 415 (1970).
34. R. Hultgren, P.D. Desai, D.T. Hawkins, M. Gleiser, and K. Kelly: 42. K. Ono, S. Nishi, and T. Oishi: A thermodynamic study of liquid
Selected Values of the Thermodynamic Properties of Binary Al- Cu–Sn and Cu–Cr alloys by the Knudsen cell–mass filter combi-
loys (American Society for Metals, Metals Park, OH, 1973). nation. Trans., JIM 25, 810 (1984).
35. S. Takeuci, O. Uemura, and S. Ikeda: On the heat of mixing of 43. T. Oshi, T. Hiruma, and J. Moriyama: Thermodynamic studies of
liquid copper alloys. Sci. Rep. Tohoku Imp. Univ. A 25, 41–55, molten Cu–Sn alloys using zirconia solid electrolyte. J. Japan
(1974). Inst. Met. 36, 481 (1972).
36. A. Yazawa and K. Itagaki: Heats of mixing in liquid copper or 44. A.K. Sengupta, K.P. Jagganathan, and A. Ghosh: Thermodynamic
gold binary alloys. Trans., JIM 16, 679 (1975). measurements in liquid Cu–Sn alloys. Metall Trans. B 9, 141
37. Y. Iguchi, H. Shimoji, S. Ban-Ya, and T. Fuwa: Calorimetric (1978).
Study of heat of mixing of Cu alloys at 1120 °C. Tetsu-to-Hagane
45. C.B. Alcock and K.T. Jacob: Solute–solute and solvent–solute
63, 349 (1977).
interactions in ␣-solid solutions of u + Sn, Au + Sn, and
38. M.J. Pool, B. Predel, and E. Schultheiss: Application of the
Cu–Au–Sn Alloys. Acta Metall. 22, 539 (1974).
SETARAM high temperature calorimeter for determination of
mixing enthalpies of liquid alloys. Thermochim. Acta 28, 349 46. F. Sommer, W. Balbach, and B. Predel: Thermodynamic investi-
(1979). gation of Cu–Sn alloys with solid state galvanic cells. Thermo-
39. J.J. Lee, B.J. Kim, and W.S. Min: Calorimetric investigation of chim. Acta 33, 119 (1979).
liquid Cu–Sb, Cu–Sn and Cu–Sn–Sb alloys. J. Alloys Compd. 47. B. Predel and U. Schallner: Thermodynamic investigation of the
202, 237 (1993). Cu–Ga, Cu–In, Cu–Ge and Cu–Sn Systems. Mater. Sci. Eng. 10,
40. C.B. Alcock, R. Sridhar, and R.C. Svedberg: A mass spectromet- 249 (1972).
ric study of the binary liquid alloys, Ag–In and Cu–Sn. Acta 48. ThermoCalc v. Q. Foundation Computational Thermodynamic
Metall. 17, 839 (1969). (Stockholm, Sweden, 2000).
41. J.P. Hager, S.M. Howard, and J.H. Jones: Thermodynamic prop- 49. S-W. Chen, unpublished work.

J. Mater. Res., Vol. 22, No. 11, Nov 2007 3165

View publication stats