Formation and morphology of the intermetallic compounds formed at the 91Sn-8.55Zn-0.45Al lead-free solder alloy/Cu interface

Article abstract of DOI:10.1016/j.jallcom.2004.08.009

Formation and morphology of the intermetallic compounds (IMCs) formed at the 91Sn-8.55Zn-0.45Al lead-free solder alloy/Cu interface has been studied using X-ray diffraction (XRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM)

Full text of DOI:10.1016/j.jallcom.2004.08.009

Journal of Alloys and Compounds 389 (2005) 133–139
Formation and morphology of the intermetallic compounds formed
at the 91Sn–8.55Zn–0.45Al lead-free solder alloy/Cu interface
Moo-Chin Wanga,b,∗, Shan-Pu Yuc, Tao-Chih Changc, Min-Hsiung Honc
a
Department of Materials Science and Engineering, National United University, 1 Lien-Da, Kung-Ching Li, Miao-Li 36003, Taiwan
b
Department of Mechanical Engineering, National Kaohsiung University of Applied Sciences, 415 Chien-Kung Road, Kaohsiung 80782, Taiwan
c
Department of Materials Science and Engineering, National Cheng Kung University, 1 Ta-Hsueh Road, Tainan 70101, Taiwan
Received 15 June 2004; received in revised form 4 August 2004; accepted 5 August 2004
Abstract
Formation and morphology of the intermetallic compounds (IMCs) formed at the 91Sn–8.55Zn–0.45Al lead-free solder alloy/Cu interface
has been studied using X-ray diffraction (XRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM) and electron
diffraction (ED). The XRD analysis indicates that the ␥-Cu5Zn8 IMC is found at the solder/Cu interface in the as-wetted sample. When heat
treated at 423 K for 100 or 400 h, the IMC phases still remain as ␥-Cu5Zn8 and Cu6Sn5. As the aging time increases from 400 to 1000 h,
besides ␥-Cu5Zn8 and Cu6Sn5, ␥ꢀ-Cu9Al4 also appears. Planar-shaped ␥-Cu5Zn8 and Cu6Sn5 are found near the solder and the Cu substrate,
respectively. However, scallop-shaped ␥-Cu5Zn8 is found at the eutectic 91Sn–8.55Zn–0.45Al solder alloy/Cu interface when aged at 423 K.
© 2004 Elsevier B.V. All rights reserved.
Keywords: Intermetallic compounds; 91Sn–8.55Zn–0.45Al lead-free solder alloy; Planar-shaped; Scallop-shaped; ␥-Cu5Zn8; Cu6Sn5; ␥ꢀ-Cu9Al4
1. Introduction
to oxidation and corrosion. Al has been incorporated with
Zn to enhance the atmospheric corrosion resistance of the
conventional galvanizing coating for steels. Al forms a solid
solution with Zn and Sn and has an eutectic point at 470 K as
reported by Sebaoun et al. [6], who have discussed the diffu-
sion path of various Sn–Zn–Al systems at various isotherms.
In addition to Al, certain transition metals such as Cr, Ti and
Zr also assist in improving the oxidation and corrosion re-
sistance of the alloys in view of the passivation behavior of
these elements. Nevertheless, these elements have high melt-
ing points and do not form low melting eutectic alloys with
Sn and Zn, and thus are excluded from consideration as Pb
substitutes.
Variouslead-freesolderalloyshot-wettedonCusubstrates
have been reported by Yu et al. [7–11]. This method offers the
benefit that optimal bulk quantities such as adhesion strength
[7], process parameters [8], wetting characteristics [9], com-
position and heat-treatment effects [10], and IMCs formation
[11] can be obtained. The present study is a continuation of
our work to evaluate the formation and microstructure of the
intermetallic compounds (IMCs) layer formed at the eutetic
For electronic parts and devices, solder joints provide
electrical conductivity and suitable mechanical strength
[1]. Although a lot of solder alloys can be chosen, eutectic
Sn–Pb solder alloys with 37–40 wt.% lead are widely used
for microsoldering in electronic assembles because of their
excellent wettability and other necessary properties, such
as appropriate melting temperature and surface tension.
However, it has been found that when lead or its compound
is inhaled, its toxicity is very harmful to human health. The
use of Pb alloys is hence prohibited [2–4], resulting in an
emergent research on lead-free solders in electronic industry
for substituting the Pb–Sn system.
McCormack et al. [5] have pointed out that the non-
toxic binary Pb-free solder close to the eutectic composition
(91Sn–9Zn) has a melting temperature of 471 K. The Sn–Zn
alloy has excellent mechanical properties but is susceptible
∗
Corresponding author. Tel.: +886 37 381710; fax: +886 37 324047.
E-mail address: mcwang@nuu.edu.tw (M.-C. Wang).
0925-8388/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2004.08.009

134
M.-C. Wang et al. / Journal of Alloys and Compounds 389 (2005) 133–139
91Sn–8.55Zn–0.45Al lead-free solder alloy/Cu substrate in-
terface.
The XRD patterns of the IMCs were obtained using an X-
ray diffractometer (Model Rad II A, Rigaku, Tokyo, Japan)
with Cu K␣ radiation and an Ni filter, operated at 30 kV,
200 mA and a scanning rate (2θ) of 0.25◦/min.
The thickness of the IMCs layer was determined by
scanning electron microscopy (SEM, JEOL 840, Tokyo,
Japan). The sample was cross-sectioned and the segment was
mounted and prepared to determine the thickness of the IMCs
layer. Twenty thickness measurements were conducted for
each run and the average value was obtained.
The morphology of the IMCs at the lead-free solder al-
loy/Cu interface was observed by SEM (JEOL840, Tokyo,
Japan) and transmission electron microscopy (TEM, H700H,
Hitachi, Tokyo, Japan). The specimen for TEM observation
was prepared by cutting with a diamond saw, grinding and
polishing to 50 ␮m thick, and finally ion milling with Ar+
ions. Electron diffraction (ED) examination was performed
on the carefully thinned specimen.
2. Experimental procedures
2.1. Sample preparation
The 91Sn–8.55Zn–0.45Al lead-free solder alloy was pre-
pared by melting a Zn–5Al master alloy with tin metal. Ac-
cordingly, the solder composition is Sn–9(5Al–Zn), where 9
represents the weight percentage of Zn–5Al.
A Cu plate (about 99.9% pure) of approximately 65 mm ×
20 mm × 2.5 mm in size, was degreased in an alkaline solu-
tion of NaOH (5 wt.%) for 15 s, followed by rinsing in deion-
ized (DI) water for 10 s. Then the Cu substrate was pickled
in an HCl solution (5 vol.%) for 10 s, followed by rinsing in
DI water again. The substrate was then dipped in a dimethy-
lammonium chloride (DMAHCl) flux (2.5 g DMAHCl/100
C2H5OH) for 10 s. After being fluxed, the sample was im-
mersed into the 91Sn–8.55Zn–0.45Al solder bath at 573 K
for 5 s as shown in Fig. 1. All samples were aged at 393, 423,
and 453 K for various aging times, respectively, in a furnace
capable of maintaining the temperature to 3 K.
3. Results and discussion
3.1. Phase identification of the IMC layer formed at the
91Sn–8.55Zn–0.45Al lead-free solder alloy/Cu interface
Interfacial phase transformation at the solder al-
loy/substrate joint affects the IMC growth significantly. At
the interface between the lead-free solder alloy and copper
substrate, the formation of the IMC is accompanied by in-
terfacial reactions and spalling [12–14]. Complex reactions
occur at the interface, and the precipitates in the lead-free
solder near the interface are accumulated to form the inter-
metallic layers.
2.2. Sample characterization
The IMCs were exposed by etching out the un-reacted sol-
der with a solution of 10% 100 g/l FeCl3/6H2O, 10% 100 g/l
CrO3/6H2O, 40% ethanol, 10% HNO3 and 30% HCl, and
were analytically determined by X-ray diffraction (XRD).
The XRD pattern of the sample heated at 573 K for 2.5 min
at a rate of 11.8 mm/s and then etched to eliminate the un-
reacted solder alloy is shown in Fig. 2. It indicates that the
␥-Cu5Zn8 is found at the 91Sn–8.55Zn–0.45Al lead-free sol-
der/Cu interface. The ␥-Cu5Zn8 is also found both at the eu-
tectic Sn–Zn/Cu interface [8] and in the Sn–Zn solder alloy
on the Cu substrate [15,16]. However, no Cu–Sn compound
Fig. 2. XRD pattern of the sample dipped at 573 K for 2.5 min and a rate
of 11.8 mm/s, and then etched to eliminate unreacted solder alloy. (ꢀ) ␥-
Cu5Zn8; (♦) Sn; (ꢁ) Cu6Sn5; (+) ␥ꢀ-Cu9Al4; (᭹) Cu.
Fig. 1. Schematic diagram of the apparatus for wetting test.

M.-C. Wang et al. / Journal of Alloys and Compounds 389 (2005) 133–139
135
forms at the interface between the Sn–Zn–Al solder alloy
and copper substrate, which differs from the previous reports
[16–21].
Since the Gibbs free energy of formation for ␥-Cu5Zn8 is
−12.34 kJ/mol, which is much lower than those of ␧-Cu3Sn
(ꢀG = −7.78 kJ/mol) and ␩ꢀ-Cu6Sn5 (ꢀG = −7.42 kJ/mol)
[22], it would be more stable. Moreover, the diffusivity of
Sn in Cu–Sn alloys is given by DSn = 1.90 × 10−10 cm2/s
at 573 K, and that of Zn in Cu–Zn alloys is DZn = 2.70 ×
10−10 cm2/s at 573 K [11]. Those data justify the formation
of ␥-Cu5Zn8 instead of Cu–Sn compounds.
The XRD patterns of the eutectic 91Sn–8.55Zn–0.45Al
solder alloy hot-wetted on the Cu substrate at 573 K and heat
treated at 423 K for various times are shown in Fig. 3. As heat-
treated for 100 h, the ␥-Cu5Zn8 and Cu6Sn5 phases are found
at the solder alloy/Cu interface as shown in Fig. 3(a). As heat
treated for 250 and 400 h, the IMCs phases still remain as
␥-Cu5Zn8 and Cu6Sn5, as shown in Fig. 3(b). In Fig. 3(c),
when the heat treatment time increases from 400 to 1000 h,
besides, ␥-Cu5Zn8 and Cu6Sn5, ␥ꢀ-Cu9Al4 also appears.
3.2. Morphology of the IMCs formed at the
91Sn–8.55Zn–0.45Al lead-free solder alloy/Cu interface
The SEM micrographs of the 91Sn–8.55Zn–0.45Al lead-
free solder alloys as-wetted on the copper substrate and
heated at 423 K for various times are shown in Fig. 4. The
IMC is observed at the interface of the lead-free solder al-
loy and copper substrate. Comparing with Fig. 4(a), the ␥-
Cu5Zn8 IMC layer of the sample heated at 423 K for 100 h
(Fig. 4(b)) is thicker than the as-wetted one (Fig. 4(a)). More-
over, we observed growth of the ␥-Cu5Zn8 and Cu6Sn5
IMC phases. The scallop-shaped IMC grains and discon-
tinuous IMC layer are shown in Fig. 4(b). As aging time
increases to 250 h (Fig. 4(c)), the interaction between the
91Sn–8.55Zn–0.45Al lead-free solder alloy and Cu substrate
should be enhanced, and the growth of the ␥-Cu5Zn8 and
Cu6Sn5 IMCphasesisobserved, whichinducestheformation
of the continuous scallop-shaped IMC layer. When heated for
400 h, the scallop-shaped IMC grains and Kirkendall voids
in Fig. 4(d) are bigger than that in Fig. 4(c). Especially, af-
ter heating at 423 K for 1000 h (Fig. 4(e)), the cracks formed
from the interconnected Kirkendall voids are more obvious
between the ␥-Cu5Zn8 and Cu substrate. On the other hand,
the cracks also form between the IMC layer and Cu substrate.
According to Fig. 3(c), for more than 400 h, the ␥ꢀ-Cu9Al4
IMC grows and results in microcracks [11]. Simultaneously,
Zn diffuses to expedite the growth of ␥-Cu5Zn8.
Frear et al. [23] have pointed out that the diffusional pro-
cess which forms an intermetallic layer requires all elements
in the solder joint (i.e. Cu and Zn) to diffuse. If one ele-
ment diffuses more quickly than the other, excessive vacan-
cies will be formed within it, which eventually accumulate
and coalesce in a compositional plane defined by the rela-
tive diffusional fluxes. This effect is known as Kirkendall
voids which reduce the effective cross-sectional area and
Fig. 3. XRD patterns of the eutectic 91Sn–8.55Zn–0.45Al solder alloy hot-
wetted on the Cu substrate at 573 K and heat treated at 423 K for (a) 100,
(b) 400, and (c) 1000 h. (ꢀ) ␥-Cu5Zn8; (♦) Sn; (ꢁ) Cu6Sn5; (+) ␥ꢀ-Cu9Al4;
(᭹) Cu.

136
M.-C. Wang et al. / Journal of Alloys and Compounds 389 (2005) 133–139
Fig. 4. SEM micrographs of the cross-sectioned samples for as-dipped and aged at 423 K for various times: (a) as-wetted, (b) 100, (c) 250, (d) 400, and (e)
1000 h.
Table 1
Comparison of the morphology and phases of the intermetallic compound layer formed at the interface between solder alloy and Cu substrate
System
Process condition
Heat treatment
Layer near solder
Layer near Cu
Reference
(temperature/time)
91Sn–9Zn/Cu
91Sn–9Zn/Cu
91Sn–9Zn/Cu
91Sn–9Zn/Cu
62Sn–36Pb–2Ag/Cu
91Sn–9Zn/Cu
91Sn–9Zn/Cu
Wet test at 523 K/30 s
Wet test at 523 K/30 s
Wet test at 523 K/30 s
As soldered at 543 K
≥462 K, ≥0.5 min
Bulk solder
Bulk solder
Hot-dipped
Hot-dipped
Hot-dipped
373 K/20 h
403 K/20 h
433 K/20 h
–
␧-Cu3Sn (scallop)
–
␥-Cu5Zn8 (planar)
␥-Cu5Zn8 (planar)
␩-Cu6Sn5 (inverted trigonal)
␤ꢀ-CuZn (planar)
␧-Cu3Sn (planar)
␩-Cu6Sn5 (isolate)
␩-Cu6Sn5 (inverted trigonal)
␩-Cu6Sn5 (inverted trigonal)
–
[24]
[24]
[24]
[15]
[25]
[8]
[8]
[8]
This study
This study
This study
This study
␥-Cu5Zn8 (scallop)
␥-Cu5Zn8 (scallop)
␩-Cu6Sn5 (scallop)
␥-Cu5Zn8 (planar)
␥-Cu5Zn8 (planar)
␥-Cu5Zn8 (scallop)
␥-Cu5Zn8 (scallop)
␥-Cu5Zn8 (scallop)
␥-Cu5Zn8 (scallop)
␥-Cu5Zn8 (planar)
–
423 K/300 h
423 K/600 h
423 K/600 h
423 K/100 h
423 K/250 h
423 K/400 h
423 K/1000 h
91Sn–9Zn/Cu
91Sn–8.55Zn–0.45Al/Cu
91Sn–8.55Zn–0.45Al/Cu
91Sn–8.55Zn–0.45Al/Cu
91Sn–8.55Zn–0.45Al/Cu
–
Hot-dipped
Hot-dipped
␥ꢀ-Cu9Al4 (planar)
␥ꢀ-Cu9Al4 (planar)

M.-C. Wang et al. / Journal of Alloys and Compounds 389 (2005) 133–139
137
cause mechanical weakening [11]. In addition, after heat-
ing at 423 K for more than 200 h, the adhesion strength de-
creases obviously from above 7 MPa to below 5 MPa for the
91Sn–8.55Zn–0.45Al solder alloy/Cu substrate system [11].
In general, one of the main causes of breakdown in solder
joints has been attributed to the excessive growth of IMCs
formed at the interface.
The comparison of the morphology and phases of the IMC
layer formed at the interface between the various solder alloys
and the copper substrate is listed in Table 1. Lee et al. [24]
have pointed out that in the wet test of the 91Sn–9Zn/Cu sys-
temat423 Kfor30 sandheatingat373 Kfor20 h, thescallop-
shaped ␧-Cu3Zn and planar ␥-Cu5Zn8 are found near the
solderandthecopper, respectively, butno␧-Cu3Znwasfound
when heating at 403 K for 20 h. After heating at 433 K for
400 h, the scallop-shaped ␥-Cu5Zn8 and inverted trigonal ␩-
Cu6Sn5 are found near the solder and the copper, respectively.
Yu et al. [8] have also pointed out that in the 91Sn–9Zn/Cu
system, planar and scallop-shaped ␥-Cu5Zn8 are found near
the solder, and the isolated inverted triangular ␩-Cu6Sn5 is
found in the copper substrate. The morphology of the IMCs
is affected by the process condition and heat treatment. In
the present study, after prolonged heating at 423 K for times
longer than 1000 h, the ␥ꢀ-Cu9Al4 IMC (Figs. 3(c) and 6) con-
tinuously grows. When heating at 423 K for 1000 h, the planar
␥-Cu5Zn8 is found in the solder alloy and the ␥ꢀ-Cu9Al4 is
found near Cu. However, the scallop-shaped ␥-Cu5Zn8 is
found in the eutectic 91Sn–8.55Zn–0.45Al solder alloy/Cu
system when heated at 423 K for 100 and 250 h, respectively,
but no ␥ꢀ-Cu9Al4 is formed in the heat treatment process.
3.3. Microstructure at the interface of
91Sn–8.55Zn–0.45Al lead-free solder alloy/Cu substrate
Fig. 5 shows the TEM micrographs (bright and dark fields)
and ED pattern of the interface region of the as-wetted sam-
ple. The ED pattern also provides the criteria for the presence
of the ␥-Cu5Zn8 in this system. Fig. 5 shows the individual
grains with subangular and somewhat bell-shaped morphol-
ogy. This result may be partly due to the impingement of
small crystallite, and the enhanced isotropy is aided by Zn
which diffuses more quickly than others.
The TEM images and ED pattern of the 91Sn–
8.55Zn–0.45Al lead-free solder alloy wetted on Cu substrate
Fig. 5. TEM micrograph and ED pattern of the 91Sn–8.55Zn–0.45Al lead-free solder alloy as-wetted on the Cu substrate for: (a) BF image, (b) DF image, and
(c) ED pattern of ␥-Cu5Zn8.

138
M.-C. Wang et al. / Journal of Alloys and Compounds 389 (2005) 133–139
Fig. 6. TEM images and ED patterns of the 91Sn–8.55Zn–0.45Al lead-free solder alloy wetted on Cu substrate and aging at 423 K for 1000 h, for (a) BF image,
(b) DF image, and (c) ED pattern of ␥-Cu9Al4.
and aged at 423 K for 1000 h are shown in Fig. 6. Fig. 6(a)
and (b) are the BF and DF images of ␥ꢀ-Cu9Al4. This result
indicates that the heat treatment enhances the Al enrichment
[7] at the ␥-Cu5Zn8/Cu interface and Al interacts with Cu to
form ␥ꢀ-Cu9Al4 IMC [11]. Fig. 6(c) shows the ED pattern
of the ␥ꢀ-Cu9Al4. In conjunction with the XRD analysis, the
BF and DF images of TEM together with the correspond-
ing ED pattern provide the evidence for the formation of the
␥ꢀ-Cu9Al4 IMC at the 91Sn–8.55Zn–0.45Al lead-free solder
alloy/Cu interface.
alloy/Cu interface has been investigated with XRD, SEM,
TEM and ED. The following conclusions are obtained:
(1) The ␥-Cu5Zn8 IMC is found at the 91Sn–8.55Zn–0.45Al
lead-free solder/Cu interface by the XRD analysis of the
as-wetted sample.
(2) As heat treated at 423 K for 400 h, the IMC phases still
remain as ␥-Cu5Zn8 and Cu6Sn5. When the aging time
increases from 400 to 1000 h, besides ␥-Cu5Zn8 and
Cu6Sn5, the ␥ꢀ-Cu9Al4 also appears.
(3) When aged at 423 K for 1000 h, the planar ␥-Cu5Zn8
is found in the solder alloy and the ␥ꢀ-Cu9Al4 is found
nearthecoppersubstrate. However, thescallop-shaped␥-
Cu5Zn8 is found in the 91Sn–8.55Sn–0.45Al solder/Cu
system when heated at 423 K for 100 and 250 h, respec-
tively, but no ␥ꢀ-Cu9Al4 is formed in the heat treatment
4. Conclusions
The formation and morphology of the intermetallic com-
pounds formed at the 91Sn–8.55Zn–0.45Al lead-free solder

M.-C. Wang et al. / Journal of Alloys and Compounds 389 (2005) 133–139
139
process.
[6] A. Sebaoun, D. Vincent, D. Treheux, Mater. Sci. Technol. 3 (1987)
241.
[7] S.P. Yu, M.H. Hon, M.C. Wang, J. Electron. Mater. 29 (2000) 237.
[8] S.P. Yu, H.J. Lin, M.H. Hon, M.C. Wang, J. Mater. Sci.-Mater. Elec-
tron. 11 (2000) 36.
(4) The ␥-Cu5Zn8 near the solder alloy and ␥ꢀ-Cu9Al4 near
the Cu substrate are found by TEM observation at the
interface between the 91Sn–8.55Zn–0.45Al solder alloy
and the copper substrate.
[9] S.P. Yu, C.L. Liao, M.H. Hon, M.C. Wang, J. Mater. Sci. 35 (2000)
4217.
[10] S.P. Yu, H.C. Wang, M.H. Hon, M.C. Wang, JOM 52 (6) (2000) 36.
[11] S.P. Yu, M.C. Wang, M.H. Hon, J. Mater. Res. 16 (2001) 76.
[12] D. Yao, J.K. Shang, Metall. Mater. Trans. A 26 (1995) 2677.
[13] A.J. Sunwoo, J.W. Morries Jr., G.K. Lucey, J.K. Shang, Metall.
Mater. Trans. A 23 (1992) 1323.
[14] D. Yao, J.K. Shang, IEEE Trans. Component B 19 (1996) 154.
[15] K. Suganuma, K. Niihara, T. Shoutoku, Y. Nakamura, J. Mater. Res.
13 (1998) 2859.
Acknowledgements
This work was supported by the National Science Coun-
cil, the Republic of China under Contract No. NSC85-2216-
E-151-005 and NSC86-2216-E-151-006, which is gratefully
acknowledged.
[16] S.W. Yoon, J.R. Soh, B.L. Lee, H.M. Lee, Acta Mater. 45 (1997)
951.
[17] Z. Mei, J.W. Morris Jr., J. Electron. Mater. 21 (1992) 73.
[18] Z. Mei, J.W. Morris Jr., J. Electron. Mater. 21 (1992) 599.
[19] P.T. Vianco, D.R. Frear, JOM 45 (7) (1993) 14.
[20] J.W. Morris Jr., J.L. Frear Goldstein, Z. Mei, JOM 45 (7) (1993) 25.
[21] L.F. Felton, C.H. Reader, D.B. Knorr, J. Met. 45 (1993) 28.
[22] R. Hultgren, P.D. Desai, D.T. Hawkins, M. Gileiser, K.K. Kelley,
Selected Values of the Thermodynamic Properties of Binary Alloy,
ASM, Metals Park, OH, 1973.
[23] D.R. Frear, W.B. Jones, K.R. Kinsman (Eds.), Solder Mechanics: A
State of the Art Assessment, TMS, Warrendale, PA, 1991, p. 57.
[24] H.M. Lee, S.W. Yoon, B.J. Lee, J. Electron. Mater. 27 (1998) 1161.
[25] M. Schaefer, R.A. Fournelle, J. Liang, J. Electron. Mater. 27 (1998)
1167.
References
[1] G. Ghosh, M. Loomans, M.E. Fine, J. Electron. Mater. 23 (1994)
741.
[2] Environment Protection Agency, Strategy for Reducing Lead Expo-
sure, 2 February 1991.
[3] P.J. Walitsky, F.G. Yost, The relevance of current environmental issue
to solder joints in microelectronic applications, in: Proceedings of
the NEPCON West, Anaheim, CA, 23–27 February, 1992, p. 1671.
[4] H. Reid, D. Moynihan, J. Leiberman, B. Bradley, Toxic Lead Re-
duction Act of 1990, p. S-2637.
[5] M. McCormack, S. Jin, H.S. Chen, J. Electron. Mater. 23 (1994)
687.