Materials Science & Engineering A 561 (2013) 232238

Contents lists available at SciVerse ScienceDirect

Materials Science & Engineering A

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

Nanoindentation study on the creep resistance of SnBi solder alloy

with reactive nano-metallic llers
Lu Shen a,b, Zheng Yu Tan b, Zhong Chen b,n
a
b

Institute of Materials Research and Engineering, AnSTAR (Agency for Science, Technology and Research), 3 Research Link, Singapore 117602, Singapore
School of Materials Science and Engineering, Nanyang Technological University, Nanyang Avenue, Singapore 639798, Singapore

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 24 August 2012
Received in revised form
22 October 2012
Accepted 25 October 2012
Available online 10 November 2012

SnBi alloy is an attractive soldering material for temperature-sensitive electronic devices. With its
excellent yield strength and fracture resistance, SnBi alloy has become one of the promising candidates
to replace Pb-based solders. However, due to the low melting temperatures of this alloy, the prominent
time-dependent deformation at service temperatures hinders its wide applications. In this study, low
concentration (no more than 4 wt%) of reactive nano-metallic llers, i.e., Cu and Ni, have been added
into the Sn58Bi alloy aiming to enhance its creep resistance. The elastic, plastic and creep properties
are characterized by nanoindentation constant strain rate (CSR) technique. The addition of the llers
has rened the microstructure of the solder matrix leading to moderate strengthening and hardening.
The creep resistance of the Sn58Bi alloy has been improved with the ller addition. Two regions of
stress-dependent creep rates were found in the alloys with and without llers. An optimum ller
concentration for creep resistance enhancement is identied at which there is a balance between the
effects of particle pinning and microstructure renement.
& 2012 Elsevier B.V. All rights reserved.

Keywords:
Nanoindentation
Creep
Nanocomposite
Tinbismuth

1. Introduction
SnBi, with its superior yield strength, fracture resistance and
comparable solderability with SnPb alloy [1], has drawn great
research focus as potential replacement for Pb-containing solders
in the microelectronic industry. Its low melting point (Tm 139 1C
at eutectic concentration) could effectively reduce the thermal
stress built-up at the electronic joints during the multiple reow
processes. However, the high homologous temperature at service
environment or even at room temperature causes large creep
deformation in the alloy. As the creep properties are closely
related to fatigue life which is a function of the accumulated
plastic deformation, enhancing the creep resistance of the low Tm
solder becomes one of the chief objectives of the works carried
out with such alloys. Composite solder, as one of the effective
ways to enhance the hardness and strength of the alloy, has been
investigated extensively. However, relatively less work has been
done for the creep enhancement. In what have been reported so
far, it was suggested that creep resistance is enhanced by
stabilizing the ne-grained microstructure so that homogenized
deformation is realized [2,3].
To serve as suitable ller materials, similar density of the ller
as the constituent materials of the base solder is required as it

Corresponding author. Tel.: 65 6790 4256; fax: 65 6790 9081.

E-mail addresses: [email protected], [email protected] (Z. Chen).

0921-5093/$ - see front matter & 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.msea.2012.10.076

promises a good distribution of the ller in the matrix. Moreover,

it requires good wetting between ller and base solder as well as
a similar Tm and solderability between the composite alloy and its
non-modied counter-part. Ceramic particles, such as Al2O3 [4,5]
and TiO2 [4], carbon nano-tubes (CNTs) [6,7] and metallic llers
including Cu [2,8,9], Ni [9], Sb [10] and Mn [11], etc., demonstrated good potential in improving the mechanical properties of
solder alloys. Lin et al. [8] has reported a 3040% increase in
hardness of the eutectic SnPb with less than 5% of Cu addition.
Marshall et al. [12] reported superior mechanical properties
(including Youngs modulus, yield and tensile strength) of the
Cu6Sn5 powder modied SnPb (60/40) alloy for up to 40% of the
ller concentration. Guo et al. [9] has investigated the creep
resistance of Sn3.5Ag alloys with low concentration of Ni and Cu
addition. Cu was reported to moderately enhance the creep
resistance of SnAg alloy at both room temperature and elevated
temperature. Ni composite alloy showed much higher creep
resistance than Cu composite at room temperature, but insignificant effect at elevated temperatures. The differences of the Ni
and Cu llers were attributed to their solubility and diffusively in
the base alloys. 1 wt% Cu addition was reported as a microstructure stabilizer for SnBi alloy in which insignicant grain size
variation was found in the matrix alloy after 30 days of annealing
at 80 1C [13]. Such ne grain microstructure is benecial in
strengthening the alloy matrix. However, it is not ideal for creep
resistance as it promotes grain boundary sliding especially at low
strain rate. Therefore, it is expected that an optimum ller

L. Shen et al. / Materials Science & Engineering A 561 (2013) 232238

concentration exist for each compatible ller-solder matrix system, in which the effect of pinning and restricting of the grain
boundary motion is maximized by the ller addition.
In the present study, nano-Cu and Ni particles were added into
the Sn58Bi alloy. The creep behaviour of the composite alloys were
studied by nanoindentation constant strain rate (CSR) method [14].
The stress and strain rate relation is obtained by capturing the stress
responses at eight controlled indentation strain rates. The indentation creep behaviour and the effect of ller distribution and the
microstructural change are discussed. The current work contributes
to the limited literatures on the effect of ller type, loading, size, etc.
on the creep behaviour of low-melting SnBi solders.

exponent which can be derived as the slope of the linear t of a

series of stress and strain rate pairs plotted in logarithm scale.
The indentation tests were performed with the MTS Nano
Indenters XP (TN, USA) system using a sharp Berkovich tip. Load
was applied at eight constant strain rates from 0.0005 s  1 to
0.1 s  1 to a depth of 4 mm. During loading, a small oscillated force
with known frequency and amplitude was imposed on the
nominal load to extract a series of contact stiffness (S) along
indentation loading. Continuous modulus (E) and hardness (H)
proles can thus be calculated in respect of indentation depth
from Eqs. (3) and (4) with a single loading/unloading cycle.
p
E
p 1
pS
3

2
1n2
Ac

2. Experimental
H
2.1. Materials and sample preparation
Sn-Bi solder paste with the eutectic composition of 42%Sn
58%Bi and nano-Cu and Ni particles were obtained from ESL
Europes and US Research Nanomaterials, Inc., respectively. Two
types of composition mixture were prepared, i.e., SnBi solder
paste blended with nano-copper powder (average particle size
100 nm) and SnBi solder paste blended with nano-nickel powder
(average particle size o70 nm). The composite solders with
several weight percentages of the nano-llers ranging from
0.5 wt% to 4 wt% were prepared. The mixture was homogenously
blended in an aluminum dish rst before melt in a furnace
maintained at temperature of 300 1C for an hour. The mixture
composition was then air cooled to room temperature.
The solidied samples were prepared for subsequent morphology study and mechanical testing. The samples were moulded
using mixture of epoxy and hardener. Sample preparation
involved grinding using progressively ner grade of silicon
carbide papers with copious amounts of water. After which, ne
polishing was carried out using 6, 3 and 1 mm diamond paste with
polishing cloth. Mirror nishing of the sample surfaces was
achieved using silica suspension as the nal polishing step.
2.2. Scanning electron microscope (SEM) and energy dispersive
X-ray (EDX) analysis
The polished and indented samples were examined using JOEL
SEM 6360 for the microstructure study. The accelerating voltage
of 10 kV was used for the secondary electron (SEI) and backscattered electron images (BEI) collection. Elemental analysis of
the samples was done with an Oxford EDX instrument, attached
to JEOL FESEM 7600 F system. The accelerating voltage of 15 kV
was used to examine the nano-particle distribution in the alloys.
2.3. Indentation constant strain rate (CSR) test
The indentation strain rate, e_ i , is dened as the instantaneous
displacement rate of the indenter, dh/dt, divided by the instantaneous displacement, h:
   
dh
1
U
1
e_ i k1 U
dt
h
where k1 is a constant. The indentation stress (s) is dened as
instantaneous load divided by projected contact area. The stress
and strain rate follows the phenomenological description as
Eq. (2) when the steady state deformation process is achieved.

e_ i C 0 sn

where C0 is a material constant including temperature factor and

activation energy for the creep deformation. n is the creep stress

233

P max
Ac

where n is Possions ratio and set to be 0.35 for current analysis.

Ac is the projected contact area at the maximum load Pmax.
The reported modulus and hardness were averaged values in
the depth range from 2 4 mm where a constant values of the
properties are achieved. Five indentations were made at each
strain rate. The indenter area function describing the actual
contact area as a function of indentation depth was calibrated
on a standard SiO2 sample.

3. Results and discussion

3.1. Microstructure study
Figs. 1 and 2 show the SEM images on the cross-section area of
Sn58Bi alloys with and without the nano-particles. The dark
regions represent the Sn-rich phase while the bright regions
represent the Bi phase (note that by the SnBi phase diagram,
at 25 1C there is approximately 3 wt% of Bi dissolved in Sn
forming the Sn-rich phase, but there is no Sn dissolution in Bi
phase). The pristine alloy demonstrates a lamellar structure
where the Sn-rich phase and Bi phase inter-locked with each
other (Fig. 1(a)). The average inter phase spacing of the alternate
layered arrangement of this alloy is approximately 2.62 mm. With
small amount of nano-llers addition, the microstructure of the
composite solder becomes ner (Fig. 1(b)(d) and Fig. 2(a)(d)),
which is likely to be caused by heterogeneous nucleation on the
inter metallic compounds (IMC) formed by Cu and Ni llers with
the Sn in the matrix (elaborated later). The existence of the IMCs
may also play the role of preventing the growth of the lamellar
structures. As the weight percentage of the nano-llers increases,
the size of inter phase spacing (IPS) generally decreases. Table 1
summarizes the evolution of IPS in SnBi composite alloys with
increased ller concentration. It is noticed that the ller addition
as low as 0.5 wt% could cause great reduction of IPS in SnBi alloy.
Further increasing the ller concentration up to 4 wt% only gently
rened the microstructures. EDX was used to verify the distribution of the nano-llers in the alloy matrix. Figs. 3 and 4 show the
elemental mapping of Cu- and Ni-added SnBi alloys, respectively.
Based on existing literatures [2,8,9,15,16], Cu and Ni form Cu6Sn5
and Ni3Sn4 IMCs with Sn upon solder melting. Therefore in the
subsequent analyses we assume the presence of these two IMCs.
It is noticed that Cu element seems to be slightly segregated in
the Sn-rich phase (Fig. 3(d)), which indicates that the CuSn IMC
phase is mainly distributed in Sn-rich phase. Ni distribution, on
the other hand, appears to be quite homogeneous across both
phases (Fig. 4(d)). Further clarication is necessary but beyond
the scope of this paper. More detailed analyses using TEM and
electron diffraction are under the way in our lab.

234

L. Shen et al. / Materials Science & Engineering A 561 (2013) 232238

Fig. 1. Morphology of the constituent phase distribution of (a) Sn58Bi, and the composite alloy with (b) 1%, (c) 2%, and (d) 4% nano-Cu.

Fig. 2. Morphology of the constituent phase distribution of the SnBi composite alloy (a) with (a) 0.5%, (b) 1%, (c) 2%, and (d) 4% nano-Ni.

Table 1
Evolution of inter phase spacing (IPS) in SnBi composite alloys with increasing
ller percentage.
Filler concentration (wt %)

Nano-Cu ller (lm)

Nano-Ni ller (lm)

0 (Sn58Bi)
0.5
1.0
2.0
3.0
4.0

2.62 7 0.18
1.73 7 0.26
1.53 7 0.13
1.51 7 0.23
1.39 7 0.13
1.21 7 0.19

2.62 70.18
1.77 70.14
1.74 70.36
1.28 70.23
1.25 70.10
1.18 70.10

3.2. Elastic modulus and hardness

Fig. 5(a) and (b) shows the modulus (E) and hardness (H) of the
Sn58Bi and its nano-particle composites with ller concentration up to 4 wt%. The values of E and H are listed in Table 2. The
pristine Sn58Bi is shown to possess E and H of 38.43 GPa and
287.3 MPa, respectively. The two types of particles demonstrate
similar strengthening effect, in which 4 wt% ller addition
increases the elastic modulus of the SnBi solder slightly
(8 12%). The increment is believed to be attributed to the stiff
nature of IMCs particles in the composite. The moduli of the

L. Shen et al. / Materials Science & Engineering A 561 (2013) 232238

235

Fig. 3. Elemental mappings of cross-section of SnBi with 4 wt% nano-Cu addition. (a) Low angle backscatter electron (LABE) image and EDX mappings of (b) Sn, (c) Bi,
and (d) Cu.

Fig. 4. Elemental mappings of cross-section of SnBi with 4 wt% nano-Ni addition. (a) Low angle backscatter electron (LABE) image and EDX mappings of (b) Sn, (c) Bi,
and (d) Ni.

composite can be expressed by the empirical Eq. (5), which is the

modication from the iso-strain arrangement of the llers and
matrix [17]:
Ec V m Em kV p Ep

where Ep,m and Vp,m are the IMC particle and matrix modulus and
volume fraction, respectively. k is empirically determined falling
in the range of 0 to 1. The higher the value of k, the greater the
enhancement effect from the IMC particles. The slope of the linear
t, s, of the composite modulus vs. IMC volume fraction (Vp) can

be expressed as Eq. (6):

s k  Ep 38:43

For Cu6Sn5, the reported modulus of 90 GPa [18] is used as Ep,

while the literature modulus value of 134 GPa for Ni4Sn3 is used
[19] for Ni composite prediction. As such, the k values of the
nano-Cu and nano-Ni modied SnBi alloy are both found to be
approximately 0.3. The composite modulus of the SnBi(Cu) and
SnBi(Ni) alloys is thus predicted by following equation provided

236

L. Shen et al. / Materials Science & Engineering A 561 (2013) 232238

Fig. 5. (a) Modulus and (b) hardness of Sn58Bi alloys with and without nano-Cu
and Ni additions.

Fig. 6. Strain rate and stress relation of Sn58Bi solder comparing with its
composite solders with (a) 4 wt% Cu addition, and (b) 4 wt% Ni addition.

Table 2
Modulus and hardness of Sn58Bi and its composites with nano-Cu and nano-Ni
llers.

Table 3
Stress exponents of Sn58Bi and its composites with nano-Cu and nano-Ni llers.

Sn58Bi
SnBi0.5%Cu
SnBi1%Cu
SnBi2%Cu
SnBi3%Cu
SnBi4%Cu
SnBi0.5%Ni
SnBi-1% Ni
SnBi2% Ni
SnBi3% Ni
SnBi4% Ni

Modulus (GPa)

Hardness (MPa)

38.43 7 0.57
38.29 7 1.75
40.347 0.74
39.43 7 1.98
39.12 7 0.64
41.67 7 0.98
38.81 7 0.26
39.19 7 1.01
38.89 7 1.31
39.69 7 2.30
43.23 7 1.08

287.3 720.0
299.9 725.9
317.4 711.9
311.9 717.0
327.7 717.0
333.8 734.1
298.7 712.8
297.5 710.3
326.5 78.0
318.3 730.5
354.1 77.6

0 (Sn58Bi)
0.5
1.0
2.0
3.0
4.0

the solder materials consist of continuous phases without defects.

Ec 38:43  V SnBi 0:3V IMC EIMC

Filler concentration (wt %)

Fig. 5(b) shows the hardness of the Sn58Bi alloy steadily

increases with nano-Cu and Ni concentration. The enhancement
in hardness is more signicant, in which 4 wt% of Cu and Ni
addition has increased the hardness of the pristine Sn58Bi by
16% and 23%, respectively. The hardening effect from the metal
llers is believed not only coming from the hard IMC phase, but
also the strengthening of the base alloy due to renement of the
microstructure. The hardness behaviour of the composite alloy is
thus not able to be predicted by simple relationship between the
properties of ller and the matrix.
3.3. Creep behaviour
The addition of ller increases the creep resistance of the unmodied sample by shifting the strain ratestress curve towards
higher stress level at the same strain rates. Fig. 6(a) and (b) shows

Cu ller

Ni ller

n1

n2

n1

n2

5.20
5.58
7.04
7.23
7.92
6.11

2.35
2.42
2.51
3.24
3.31
2.14

5.20
6.20
7.45
5.89
5.29
5.16

2.35
3.73
3.85
3.60
2.95
2.93

a representative stress-dependent strain rate behaviour of the

Sn58Bi alloy with 4 wt% of metal llers. The indentation creep
test results show two stress regions in the strain ratestress
curves of Sn58Bi alloys with and without metallic llers. The
transition stress is approximately unchanged at 170 MPa in spite
of the moderate enhancement of the creep resistance by metal
ller addition. Table 3 lists the stress exponents of the samples
measured at the two stress regions. n1 is dened as the slope of
strain ratestress curved at higher stress region (4170 MPa);
while n2 is the one measured at stress region lower than 170 MPa.
n1 for the two types of metal ller samples is found in the range
from 5.16 to 7.45, which implies the dominant creep mechanism
in this stress range is related to dislocation movement in the bulk.
To compare the creep resistance among various samples with
different ller loadings, the strain rates at a particular stress,
selected at 200 MPa and 100 MPa to represent the high stress and
low stress regions respectively, are shown in Fig. 7(a)(d). The
pristine SnBi alloy shows the highest strain rate (lowest creep
resistance) at both stress levels. With addition of nano-Cu or Ni,
the creep rate reduces signicantly and achieved best creep
resistance when ller concentration reaches 3 wt% for Cu ller
composite and 1 wt% for Ni ller composite.

L. Shen et al. / Materials Science & Engineering A 561 (2013) 232238

237

Fig. 7. Strain rates of SnBi(Cu) alloys deformed at (a) 200 MPa and (b) 100 MPa. Strain rates of SnBi(Ni) alloys deformed at (c) 200 MPa and (d) 100 MPa.

Pores

Fig. 8. SEM images of Sn58Bi with Ni 4% deformed at stress of (a) 354 MPa; and (b) 116 MPa.

Fig. 8(a) shows the SEM image of SnBi with 4 wt% Ni deformed
at 354 MPa. The indentation imprint displayed a regular triangle
with evaginated boundaries indicating discernible pile-up of the
material during the indentation process. It is believed that the
indentation induced deformation is accomplished by the bulk
deformation inside Sn-rich and Bi phases. As a result, the
hardened individual phases push the surrounding material outwards causing signicant pile-up around the indented area.
However, when material is deformed at a lower stress of
116 MPa where a lower strain rate is applied, signicant number
of pores are formed at the boundaries of the adjacent phases as
shown in Fig. 7(b). Prominent delamination at phase boundaries
is formed at the surrounding area of the indent. It is believed that
at a lower stress condition, the deformation of the material is
mainly channelled through slipping or shifting at the phase
boundaries between Sn-rich and Bi phases. Comparing to the
indentation creep results, the stress exponent n2 at lower stress
region ( o170 MPa) is in the range from 2.42 to 3.85, which
corresponds well with the phase boundary sliding mechanism.
As shown above, the creep resistance reaches a maximum at
3 wt% Cu and 1 wt% Ni ller concentration, and then starts
decrease when more ller is added. With the nano-ller incorporation, it is believed there are two major strengthening
mechanisms from the stiff metal particles. One comes from the

pinning of the dislocation movement through effective particle

dislocation interactions. Rosler et al. [20] demonstrated that such
interactions are primarily responsible for the excellent creep
resistance enhancement of carbide dispersion-strengthened aluminum alloys. Secondly, the obvious microstructural renement
due to ller addition strengthens the matrix due to effect such as
HallPetch hardening. However, such microstructure renement
promotes diffusional creep (e.g. Nabarro-Herring, Coble) and
grain boundary sliding, adversely affecting creep resistance especially at lower stress regions. With the two effects functioning
simultaneously, it is believed that at certain low ller concentrations, the metal ller pinning effect dominants the creep deformation until an optimum ller concentration is reached, after that
creep rate starts to increase due to signicant amount of diffusional creep and phase boundary sliding is involved, which causes
the stress exponent shift to a lower value.

4. Conclusions
The elastic, plastic and creep properties of Sn58Bi alloy have been
studied with varying amount of nano-metal ller additions. Moderate
stiffness enhancement is found in the two types of composite alloys.
The formed hard particles and the microstructure renement due to

238

L. Shen et al. / Materials Science & Engineering A 561 (2013) 232238

nano-ller addition promote the hardening of the alloys. The creep

resistance of the Sn58Bi is found to increase with addition of nanometal llers. At higher stress range (4170 MPa), the creep mechanism is dominated by dislocation movement in the bulk; while at
lower stress range, it is dominated by phase boundary sliding. 3 wt%
Cu and 1 wt% Ni are found to be the optimum concentrations for the
creep resistance enhancement for Sn58Bi alloy.
References
[1] F. Hua, Z. Mei, J. Glazer, in: 48th Electronic Components and Technology
Conference, 227 (1998).
[2] F. Guo, J. Mater. Sci. Mater. Electron. 18 (2007) 129145.
[3] J. Shen, Y.C. Chan, Microelectron. Reliab. 49 (2009) 223234.
[4] H. Mavoori, S. Jin, J. Electron. Mater. 27 (1998) 12161222.
[5] X.L. Zhong, M. Gupta, J. Phys. D: Appl. Phys. 41 (2008).
[6] K. Kumar, V. Kripesh, L. Shen, A. Tay, Thin Solid Films 504 (2006) 371378.
[7] S.M.L. Nai, J. Wei, M. Gupta, J. Electron. Mater. 35 (2006) 15181522.

[8] D. Lin, G.X. Wang, T.S. Srivatsan, M. Al-Hajri, M. Petraroli, Mater. Lett. 53
(2002) 333338.
[9] F. Guo, S. Choi, K.N. Subramanian, T.R. Bieler, J.P. Lucas, A. Achari,
M. Paruchuri, Mater. Sci. Eng., A 351 (2003) 190199.
[10] F. Guo, G. Xu, H. He, J. Mater. Sci. 44 (2009) 55955601.
[11] K.S. Kim, S.H. Huh, K. Suganuma, Microelectron. Reliab. 43 (2003) 259267.
[12] J.L. Marshall, J. Calderon, J. Sees, G. Lucey, J.S. Hwang, IEEE Trans. Compon.,
Hybrids, Manuf. Technol. 14 (1991) 698702.
[13] L.E. Felton, C.H. Raeder, D.B. Knorr, J. Miner. Met. Mater. Soc. 45 (1993)
2832.
[14] B.N. Lucas, W.C. Oliver, Metall. Mater. Trans. A 30 (1999) 601610.
[15] M. He, Z. Chen, G.J. Qi, C.C. Wong, S.G. Mhaisalkar, Thin Solid Films 462
(2004) 363369.
[16] M. He, Z. Chen, G.J. Qi, Metall. Mater. Trans. A 36A (2005) 6575.
[17] T.H. Courtney, Mechanical Behavior of Materials, second ed., McGraw Hill,
2000.
[18] J.H. Westbrook, R.L. Fleischer, Structural Applications of Intermetallic Compounds, Wiley, New York, 2000.
[19] Z. Chen, M. He, B. Balakrisnan, C.C. Chum, Mater. Sci. Eng., A 423 (2006)
107110.
[20] J. Rosler, R. Joos, E. Arzt, Metall. Trans. A 23 (1992) 15211539.