You are on page 1of 9

Materials Science and Engineering A 530 (2011) 402410

Contents lists available at SciVerse ScienceDirect

Materials Science and Engineering A


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

Microstructure and tensile behavior of Sn5Sb lead-free solder alloy containing


Bi and Cu
M.J. Esfandyarpour, R. Mahmudi
School of Metallurgical and Materials Engineering, College of Engineering, University of Tehran, Tehran, Iran

a r t i c l e

i n f o

Article history:
Received 25 May 2011
Received in revised form 27 August 2011
Accepted 27 September 2011
Available online 5 October 2011
Keywords:
Lead-free solder
SnSb alloy
Tensile properties
Stress exponent

a b s t r a c t
Tensile deformation behavior of Sn5Sb, Sn5Sb1.5Bi, and Sn5Sb1.5Cu alloys was investigated at
temperatures ranging from 298 to 400 K, and strain rates ranging from 5 104 to 1 102 s1 . Addition
of Bi and Cu into the binary alloy resulted in an increase in both ultimate tensile strength (UTS) and ductility. The improved strength of the Bi-containing alloy can be attributed to the microstructural renement,
uniform distribution of the SnSb intermetallic particles, and solid-solution hardening effect of bismuth
in the -Sn matrix. The enhanced strength of the Cu-containing alloy was ascribed to the presence of the
Cu6 Sn5 intermetallic particles and structural renement. The ductility of both ternary alloys was, however, improved by the structural renement, caused by the addition of alloying elements. The results of
tensile tests indicate that the strength of all three alloys increase with increasing strain rate and decrease
with testing temperature. The variation of ductility with strain rate showed a descending trend, while it
exhibited a minimum at medium testing temperatures. Based on the obtained stress exponents and activation energies, it is proposed that the dominant deformation mechanism in Sn5Sb is dislocation climb
over the whole temperature range investigated. For the ternary alloys, however, grain boundary diffusion
and dislocation climb are the deformation mechanisms at high and low temperatures, respectively.
2011 Elsevier B.V. All rights reserved.

1. Introduction
Tinlead solders have long been used in the microelectronic
packaging industry. The widespread use of these solders is primarily due to their low cost, good wettability and satisfactory strength
properties for lifetime performance [1,2]. Despite the long-term
utilization of SnPb alloys, recent concerns about lead toxicity have
resulted in serious restrictions on their use. Therefore, there has
been a great deal of research work to develop alternative lead-free
solders [35]. Accordingly, many lead-free Sn-based alloy systems
with different alloying elements such as; Ag, In, Cu, Zn, Bi, Ni and Sb
have been developed and their microstructures, mechanical properties, creep resistance, and wettability have been studied.
Among the developed lead-free solders, the near-peritectic
Sn5Sb alloy with a relatively high melting point of 518 K and a solder substrate contact angle of about 43 has been developed as one
of the potential alternate materials for replacing lead containing
solder alloys. The Sn5Sb alloy has the advantages of good creep
resistance [69] and superior mechanical properties, particularly
strength and ductility [10,11]. To further improve the mechanical
properties of the binary SnSb alloys, third alloying elements such

Corresponding author. Tel.: +98 21 8208 4137; fax: +98 21 8800 6076.
E-mail address: mahmudi@ut.ac.ir (R. Mahmudi).
0921-5093/$ see front matter 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.msea.2011.09.103

as Ag, Cu, In, and Au have been employed. Most of these studies are
concerned mainly with the creep resistance rather than strength
of the developed materials. Recently, El-Daly et al. [12] studied the
effect of 0.7% Cu and 0.7% Ag on the tensile behavior of a Sn5Sb
alloy. They showed that both Ag and Cu rened the microstructure and formed new intermetallic compounds. Addition of Cu was
found to be more effective than Ag in increasing the yield stress and
ultimate tensile strength (UTS) of the base material. The strength of
all tested alloys increased with increasing strain rate and decreasing testing temperature. Similar trends in strength have also been
reported when Ag and Au have been added to the Sn5Sb base alloy
[13].
The present paper examines the inuence of adding Bi and Cu
on the tensile properties of a wrought Sn5Sb solder alloy. It is well
established that Sn5Sb is a composition that can provide a wide
range of desirable soldering and mechanical properties. Addition
of alloying elements is usually aimed at improving some of these
properties without impairing the other ones. The amount of added
Bi and Cu was limited to 1.5 wt.% in order to increase the strength
and ductility without damaging the wettability and other soldering
characteristics. The tensile deformation and fracture behavior of
Sn5Sb1.5Bi and Sn5Sb1.5Cu solder specimens at various strain
rates in the range 5 104 1 102 s1 , over the wide temperature
range of 298400 K are compared with those of the Sn5Sb base
solder material.

M.J. Esfandyarpour, R. Mahmudi / Materials Science and Engineering A 530 (2011) 402410

2. Experimental procedure

Sn-5Sb-1.5Cu

Sn
SnSb
Cu6Sn5

2.1. Materials and processing

2.2. Tensile tests


Tensile specimens were taken along the rolling direction. The
parallel gage length was 28 mm long and 6.5 mm wide. The specimens were pulled to fracture at the temperatures of 298, 320,
340, 370, and 400 K and at initial strain rates of 5 104 , 1 103 ,
5 103 , and 1 102 s1 using a SANTAM universal tensile testing
machine equipped with a three-zone split furnace. Tensile specimens required 20 min to equilibrate at the testing temperature
prior to the initiation of straining. Load-extension curves were

Sn-5Sb-1.5Bi
CPS

The materials used in this investigation were Sn5 wt.% Sb,


Sn5 wt.% Sb1.5 wt.% Bi and Sn5 wt.% Sb1.5 wt.% Cu leadfree solder alloys. They were prepared from high-purity Sn
(99.95%), Sb (99.90%), Bi (99.95%) and Cu (99.95%) melted in
an electric furnace under inert argon atmosphere and cast into
120 mm 30 mm 13 mm slabs. The cast slabs were rolled to a 91%
reduction at room temperature (T > 0.5Tm ) to generate a homogeneous ne-grained material without the initial as cast structure. In
order to ensure a stabilized microstructure as the initial state of the
materials, the samples were stored at room temperature for more
than 6 months. The microstructure of solder alloys was revealed
using scanning electron microscopy (SEM). The samples were polished with 0.25 m diamond paste, followed by polishing on an
abrasive-free microcloth. Etching was carried out using a 5% HCl, 2%
HNO3 , 3% FeCl3 , and 90% ethanol solution at room temperature for
several seconds. X-ray diffraction (XRD) and energy-dispersive Xray (EDX) analysis were performed on selected samples to identify
the phases.

403

Sn-5Sb

20

30

40
50
22 , Degree
Degree

60

70

Fig. 1. XRD patterns of Sn5Sb, Sn5Sb1.5Bi, and Sn5Sb1.5Cu alloys.

obtained over the whole gage length, from which the true stress
at the peak load was calculated as the ow stress needed for the
calculation of the stress exponent (n). Fractography was performed
on the fractured surfaces in order to study the effect of Bi and

Fig. 2. SEM micrographs of: (a) Sn5Sb, (b) Sn5Sb1.5Bi, (c) Sn5Sb1.5Cu, and EDX analyses of (d) SnSb, (e) -Sn, and (f) Cu6 Sn5 particles in the respective alloys.

404

M.J. Esfandyarpour, R. Mahmudi / Materials Science and Engineering A 530 (2011) 402410

70

T = 298 K

50

A:

=1X10-2 s -1

40

B:

=5X10 s

30

-3

-1

-3

-1

-4

-1

C:

=1X10 s

D:

=5X10 s

20

55

45
1: Sn-5Sb-1.5Bi
2: Sn-5Sb-1.5Cu
3: Sn-5Sb

35

10
0
0

B C

50

100

150

(a)

25

200

250

0.005

170

70

T = 298 K

(b)

60

0.01

Strain Rate, s

Strain, %

50

A:

=1X10-2 s -1

40

B:

=5X10 s

30

C:

=1X10 s

D:

=5X10 s

-3

-1

-3

-1

-4

-1

20

120
95
1
2

70

10
A

0
0

50

100

150

(b)

45

200

250

0.015

-1

1: Sn-5Sb-1.5Bi
2: Sn-5Sb-1.5Cu
3: Sn-5Sb

145

Sn-5Sb-1.5Bi
Elongation ,%

Stress, MPa

65
UTS, MPa

Stress, MPa

60

75

(a)

Sn-5Sb

0.005

0.01

Strain Rate, s

0.015

-1

Strain, %
Fig. 4. Effect of strain rate on mechanical properties at 298 K: (a) tensile strength,
and (b) elongation.

Stress, MPa

70

(c)

60

Sn-5Sb-1.5Cu

50

A:

=1X10-2 s -1

40

B:

=5X10 s

C:

=1X10 s

30

D:

-3

-1

-3

-1

-4

-1

=5X10 s

20
10

0
0

50

100

150

200

250

Strain, %
Fig. 3. Stressstrain curves of: (a) Sn5Sb, (b) Sn5Sn1.5Bi, and (c) Sn5Sb1.5Cu
alloys, obtained at 320 K and different strain rates.

Cu additions on the fracture mechanisms during tensile loading of


the samples. This was done by using a CamScan-MV2300 scanning
electron microscope.
3. Results and discussion
3.1. Microstructural observations
Fig. 1 shows the XRD patterns of the investigated alloys. It can be
seen that the -Sn matrix and the SnSb second phase are the main
constituents in the near-pretectic Sn5Sb base alloy. The results are
in agreement with those reported by Mahidhara et al. [10]. Examination of the XRD pattern for the ternary Cu-containing material,

however, shows some new peaks corresponding to the Cu6 Sn5


phase. On the other hand, in the Sn5Sb1.5Bi alloy, no new peak is
detected. This implies that the Bi content of the alloy predominantly
dissolves in the -Sn matrix to form a solid solution.
The microstructural evolution of the investigated alloys
together with the EDX analyses of the constituent phases is demonstrated in Fig. 2af. As can be seen in Fig. 2a, the microstructure
of Sn5Sb consists of equiaxed -Sn grains with an average size
of about 3 m, and SnSb intermetallic particles uniformly distributed in the matrix. The EDX analysis of the small bright particles,
depicted in Fig. 2d, indicates that Sn and Sb are the main constituents of such particles. It is to be noted that the solubility limit
of Sb in Sn decreases form 1.2% at 400 K to 0.4% at 370 K and to
almost zero at 340 K [14]. Fig. 2b reveals that after Bi addition, the
microstructure has signicantly been rened to a grain size of about
1.5 m and the volume fraction of the second phase has increased.
The corresponding EDX pattern of the Sn matrix, shown in Fig. 2e,
exhibits that Bi has dissolved in the -Sn matrix. Consistent with
the XRD results, addition of 1.5% Bi does not result in the formation of any new phase. It seems that bismuth dissolved in the -Sn
matrix prohibited Sb dissolution into the matrix, and thus, resulted
in a higher volume fraction of SnSb particles, as compared to the
base alloy.
It is noteworthy that the ne and well-dispersed SnSb particles
are located in the grain boundary areas. On the other hand, addition
of Cu to the base alloy, results in a slightly ner structure containing
some rod-shape Cu6 Sn5 and the same SnSb intermetallic particles,
as shown in Fig. 2c. The achieved rened structure with a grain size
of about 2 m could have been caused by the fact that Cu6 Sn5 with
the higher melting point of 415 C, is the rst phase to form during

M.J. Esfandyarpour, R. Mahmudi / Materials Science and Engineering A 530 (2011) 402410

50

80

(a)

Sn-5Sb

10

298KK
1:1:
298

320Kk
2:2:
320

340Kk
3:3:
340

370 K
4:4:
370
K

5: 400 K

UTS, MPa

Stress, MPa

20

60

0
0

50

100

150

200

250

20
280

300

Strain, %

1:Sn-5Sb-1.5Bi
2:Sn-5Sb-1.5Cu
3:Sn-5Sb
320

. =

120

(b)

Sn-5Sb-1.5Bi

1: 298 K

Elongation, %

Stress, MPa

40
2

1: 298 K

30

2: 320 k

2: 320 K

3: 340 k

3: 340 K

20

4: 370 K

10

360

4: 370 K

5: 400 K

5: 400 K

50

100

150

200

250

300

Strain, %
50

Stress, MPa

30

20

80

40

1:Sn-5Sb-1.5Bi
2:Sn-5Sb-1.5Cu
3:Sn-5Sb
320

360

Temperature, K

400

440

1: 298 K

2: 320 K

3.2. Effect of strain rate on mechanical properties

3: 4:
340
370KK

Tensile engineering stressstrain curves of all three solder alloys


obtained at the constant temperature of 320 K and at various strain
rates in the range 5 104 1 102 s1 , are shown in Fig. 3ac. It is
to be noted that three different specimens were tested in each case
and in all cases, the curves were almost identical. As can be inferred
from Fig. 3, irrespective of the material type, the overall level of the
stressstrain curves rises and the elongation to failure decreases as
the strain rate increases. Almost in all cases, the maximum strength
reached after a relatively small amount of strain, after which the
strength decreased sharply with increasing strain. This means that,
similar to most materials, there is a competition between work
hardening and softening, in which case the pronounced dynamic
recovery of the present soft solders is dominant.
Typical variations of UTS with strain rate for the three solder
alloys at 298 K are depicted in Fig. 4a. It is evident that tensile
stress values are strongly dependent upon the applied strain rate.
The enhancement of tensile strength with increasing strain rate
is in good agreement with the result reported for SnSb [12],
SnCu [16], SnAg [17] and Sn9Zn [18] lead-free solders. The
higher strength levels obtained at higher strain rates imply that
all materials have reasonably high positive strain rate sensitivities
(SRS), usually observed in soft solder alloys. Concerning the comparison of the effect of alloying elements on the strength of the
tested alloys, it can be observed that at all strain rates the ternary
Sn5Sb1.5Bi and Sn5Sb1.5Cu solders have higher strength levels than the binary Sn5Sb alloy. The signicant increase in the UTS
of the Bi-containing solder can be attributed to the microstructural

3: 340 k

4: 5:
370
400KK

5: 400 K

0
0

0
280

1: 298 K

2: 320 k

10

(b)

Fig. 6. Effect of test temperature on mechanical properties at a strain rate of


1 102 s1 : (a) tensile strength, and (b) elongation.

(c)

Sn-5Sb-1.5Cu
40

440

1 10-2

0
0

400

Temperature, K

50
1

-2

3
40

5: 400 K

(a)

= 1 10

40
30

405

50

100

150

200

250

300

Strain, %
Fig. 5. Stressstrain curves of: (a) Sn5Sb, (b) Sn5Sn1.5Bi, and (c) Sn5Sb1.5Cu
alloys, obtained at a strain rate of 5 103 s1 and different temperatures.

cooling from the liquid state, and thus it can act as a nucleation
site for SnSb and -Sn with the respective melting points of 350
and 232 C [15]. The EDX analysis of the rod-shape particles, given
in Fig. 2f, indicates that they have a composition of 64.6 wt.% Sn
and 35.4 wt.% Cu. This corresponds closely to the stoichiometric
composition of the Cu6 Sn5 intermetallic phase, already detected
by the XRD analysis. Concerning the chemical composition of the
second phase particles, it should be noted that the quantitative EDX
results slightly overestimates the Sn percentage in all cases. This
can be a result of a large interaction volume of electron beam with
the Sn matrix, which leads to the emission of X-ray from material
beneath the particles.

406

M.J. Esfandyarpour, R. Mahmudi / Materials Science and Engineering A 530 (2011) 402410

renement, uniform distribution of the SnSb intermetallic particles, and solid-solution hardening effect of bismuth in the -Sn
matrix [1922]. The addition of Cu to the base alloy also improves
the strength, mainly due to the presence of the addition of Cu6 Sn5
intermetallic particles. However, it is evident that the strengthening effect of Cu is not as strong as that of Bi.
Fig. 4b shows the variation of total elongation with the applied
strain rate for the investigated solder alloys at room temperature. It
can be observed that elongation to failure decreases continuously
with increasing strain rate. This consistent behavior is in contrast
to those reported for a cast Sn9Zn base alloy containing Ag and
Cu [18], in which total elongation had shown a pronounced uctuation with strain rate. It seems that in the present wrought alloys a
more uniform and rened microstructure excludes the possibility
of microstructural inhomogeneities, which can result in inconsistent elongation values, usually observed in the cast structures.
The observed strain rate dependency of ductility can be interpreted based on the interaction of dislocations with intermetallics
and other barriers. These barriers can effectively obstruct the fast
mobile dislocations at low strain rates. At high strain rates, however, the obstacles cannot lock the moving dislocations any more
because of the faster dislocation. This means that at low strain rates
dislocation generation is less favored, and thus, thermally activated
processes such as dynamic recovery and recrystallization have sufcient time to act to extend the total elongation. Similar to the
strength data, shown in Fig. 4a, the Sn5Sb1.5Bi alloy displays
the highest elongation, followed by Sn5Sb1.5Cu and Sn5Sb
over the whole range of strain rates studied. The elongation values of 153, 138 and 113%, found respectively for the Sn5Sb1.5Bi,
Sn5Sb1.5Cu and Sn5Sb alloys at the strain rate of 5 104 s1 ,
are indicative of the excellent ductility of the tested alloys at room
temperature. The superior ductility of the ternary alloys relative to
the base binary material is attributed to the structural renements
caused by the addition of Bi and Cu.
3.3. Effect of temperature on mechanical properties
In order to determine the temperature dependence of deformation behavior of the tested solder alloys, their tensile properties
were investigated at temperatures in the range of 298400 K. Typical stressstrain curves of the alloys obtained at a strain rate of
5 103 s1 are shown in Fig. 5ac. It is evident that the overall strength of the alloys shows a consistent declining trend with
increasing temperature. This is in contrast to the variation of
total elongation, which exhibits a more complicated pattern. For
a more complete view on the overall temperature dependence of
the mechanical properties, the UTS and elongation to failure values were determined for Sn5Sb, Sn5Sb1.5Bi and Sn5Sb1.5Cu
alloys, and plotted against testing temperature at a constant strain
rate of 1 102 s1 in Fig. 6. It is clear from Fig. 6a that tensile
strength of all three solders dramatically decreases with increasing
temperature. This is a consequence of the activation of softening
processes, which are encouraged at high temperatures. In such a
case, the rate of decrease in the dislocation density (annihilation)
increases, leading to continuously falling of stress level. This implies
that at high temperature substructure strengthening becomes inactive and subgrain boundaries are no longer dislocation barriers. It
is to be noted that the room-temperature strengthening effects of
Cu and Bi also persist at high temperatures. Fig. 6b shows the effect
of temperature on the elongation of the three investigated solder
alloys. It is evident that the ductility of all materials passes through
a minimum at about 340 K, after which it increases with temperature. Furthermore, the Bi- and Cu-containing alloys possess higher
ductility than the base alloy at all tested temperatures in such a
way that at 400 K the elongation of these alloys are 1.6 and 1.2
times higher than that of the base alloy, respectively.

Fig. 7. Three-dimensional plots of the ductility as a function of strain rate and


temperature for: (a) Sn5Sb, (b) Sn5Sb1.5Bi and (c) Sn5Sb1.5Cu alloys.

To have a more integrated view of the ductility data, the total


elongation values were plotted as a function of temperature and
strain rate in three-dimensional plots, as shown in Fig. 7ac. In
these plots, the ductility data are expressed as surfaces that can
provide the elongation values for any combination of temperature

M.J. Esfandyarpour, R. Mahmudi / Materials Science and Engineering A 530 (2011) 402410

13.0

ln ( T/E),K/MPa- s

Table 1
Creep characteristics of the tested alloys.

(a)

T = 298 K

11.0

Sn5Sb
Sn5Sb1.5Bi
Sn5Sb1.5Cu

10.0
Sn-5Sb,
n = 6.1
Sn-5Sb-1.5Cu, n = 4.5
Sn-5Sb-1.5Bi, n = 6.5

8.0
14.0

14.5

15.0
ln (

15.5

16.0

E)

13.0

ln ( T/E),K/MPa- s

Stress exponent (n)

Low temp.

High temp.

298 K

320 K

340 K

370 K

400 K

55
73
67

55
45
44

7.1
7.8
6.2

6.6
7.3
5.8

6.1
6.5
4.5

5.1
4.2
3.9

4.4
3.7
3.4

recrystallization process [23]. The present alloys with relatively


ne-grained structures of 24 m are susceptible to grain boundary
sliding if they are deformed in the suitable ranges of temperature
and strain rates. As the temperature passes 340 K, the grain boundary accommodated deformation is facilitated and the elongation is
improved. The occurrence of deeper troughs at lower strain rates
is in accord with this argument.

(b)

T = 340 K

3.4. Stress exponens and activation energy

12.0

It is widely believed that the mechanical behavior of alloys at


homologous temperatures higher than 0.5 can be fairly expressed
by the power-law creep over a wide range of strain rates [24]. Thus,
and the tensile stress,
the relationship between the strain rate, ,
, is given by [25]:

11.0
10.0
Sn-5Sb,
n = 7.1
Sn-5Sb-1.5Cu, n = 6.2
Sn-5Sb-1.5Bi, n = 7.8

9.0
8.0
13.5

14.0

14.5
ln (

15.0

15.5

E)

13.0

(c)

T = 400 K

ln ( T/E),K/MPa- s

Q (kJ mol1 )

Material

12.0

9.0

407

12.0
11.0
10.0
Sn-5Sb,
n = 3.4
Sn-5Sb-1.5Cu, n = 3.7
Sn-5Sb-1.5Bi, n = 4.4

9.0
8.0
12.5

13.0

13.5
ln (

14.0

 T

  n

=A

 Q

exp

(1)

RT

where T is the temperature, E is the elastic modulus, A is a material parameter, n is the stress exponent, Q is the creep activation
energy, and R is the universal gas constant. Since A is a constant, it

is possible to obtain the stress exponent n from a plot of ln(T/E)


against ln (/E) at constant T. Similarly, the activation energy Q can

be obtained from a plot of ln(T/E)


versus (1/T) at constant /E.
It is worth noting that the elastic modulus of materials is highly
sensitive to test temperature. Therefore, when this parameter is
involved in the calculation of other mechanical behavior parameters such as stress exponents and activation energies, the effect
of temperature-dependence of elastic modulus (E) must be taken
into account. This procedure has almost no effect on the n-values
but affects Q-values, the correct values of which is necessary for
the identication of creep mechanisms. For the present materials,
the temperature dependence of the elastic modulus of pure Sn can

14.5

15

E)

Sn-5Sb: ( / E =1.0010 -3)

14

ln (

and strain rate. Obviously, all three surfaces exhibit a minimum


in the vicinity of 340 K and a maximum at the highest temperature and the lowest strain rate. The comparison of the elongation
data for Sn5Sb (Fig. 7a), Sn5Sb1.5Bi (Fig. 7b) and Sn5Sb1.5Cu
(Fig. 7c) indicates the better performance of the Bi-containing alloy,
as deduced from its higher overall level of the surface when considered at the same temperature and strain rate levels. Another
feature of these diagrams is that although the elongation data of
all materials are presented by a single minimum at high strain
rates, the surfaces at low strain rates show a maximum at temperatures around 320 K. The occurrence of a trough in the ductility
in the intermediate temperature region has been recognized in
many alloys. This takes place at temperatures high enough for
grain boundary sliding to initiate grain boundary cracks but not
so high that the cracks are isolated from propagation by a rapid

T/E), K/MPa- s

Fig. 8. Relationship between true strain rate and true stress at different testing
temperatures: (a) 298, (b) 340 and (c) 400 K.

Sn-5Sb-1.5Bi: ( / E =1.2210 -3)


Sn-5Sb-1.5Cu: ( / E =8.2010 -4)

13 45 kJmol -1
12
11

55 kJmol -1

44 kJmol -1

10
9
67 kJmol -1

73 kJmol -1

7
6

2.2

2.4

2.6

2.8

3.2

3.4

3.6

3.8

1000/T, K -1
Fig. 9. Relationship between the tensile stress and the reciprocal temperature for
three solder alloys.

408

M.J. Esfandyarpour, R. Mahmudi / Materials Science and Engineering A 530 (2011) 402410

Fig. 10. SEM fractograph of the solders deformed at 340 and 400 K and at the constant strain rate of 5 104 s1 .

be used. This has been proposed to be obtained from the following


equation [26]:
E(MPa) = 76087 109 T (K)

(2)

According to Eq. (1), plotting the variation of the temperaturecompensated strain rate rates against normalized true tensile
strength on a double logarithmic scale, would yield a straight line
with the slope of stress exponent, n. This is depicted in Fig. 8
for the three lead-free solders at various temperatures. The stress
exponents were determined according to Eq. (1) by plotting the
temperature-compensated strain rate against normalized tensile
strength at different temperatures, as shown in Fig. 8ac. The

results, given in Table 1, show that the stress exponents decrease,


respectively, from 7.1 to 4.4, 7.8 to 3.7, and 6.2 to 3.4, for Sn5Sb,
Sn5Sb1.5Bi, and Sn5Sb1.5Cu, respectively, as the temperature increases from 298 to 400 K. A signicant decrease in the
stress exponent with increasing temperature has been related to
the instability of the microstructure, which occurs during hightemperature deformation [27]. It is worth noting that the results
of conventional creep tests on a wrought and aged Sn5Sb alloy
are indicative of n values of 7.04.0 at 300403 K [28]. Other similar results obtained by indentation creep testing of rolled Sn5Sb
alloy indicated n values of 4.8 to 2.9 in the temperature range of
298423 K [9].

M.J. Esfandyarpour, R. Mahmudi / Materials Science and Engineering A 530 (2011) 402410

The activation energy was calculated from the semi-logarithmic


plots of temperature-compensated strain rate versus reciprocal
temperature at a certain stress for each solder, as shown in Fig. 9.
It can be observed that for the Sn5Sb alloy, the curves exhibit a
single linear behavior over the whole temperature range tested
with the activation energy value of 55 kJ mol1 . For the Bi- and
Cu-containing alloys, however, an abrupt change in the values of
activation energy is evident, indicating a change in the rate controlling mechanism. Activation energy values of 73 and 67 kJ mol1 are
obtained for the low temperature, and 45 and 44 for the high temperature regimes of Sn5Sb1.5Bi and Sn5Sb1.5Cu respectively.
The activation energy data for the tested materials are summarized in Table 1. Incidentally, the obtained activation energy values
are in agreement with the activation energies of 42 kJ mol1 and
46.8 kJ mol1 obtained respectively in impression creep testing of
Sn [29] and wrought Sn5Sb alloy [9].
The deformation of polycrystalline materials at temperatures
above 0.5Tm can take place by different mechanisms, associated
with different stress exponent and activation energy values [8]. It
has been suggested that grain boundary diffusion is characterized
by a stress exponent values of 34 [30,31]. Climb of edge dislocations becomes the rate controlling mechanism in pure metals
at high temperatures with n-values in the range 46 and activation energy that is equal to the activation energy for self-diffusion
through the lattice [32]. The activation energy of 55 kJ mol1 ,
obtained for Sn5Sb, is close to the activation energy of about
60.3 kJ mol1 reported for the lattice self-diffusion of Sn. According to the obtained n and Q values, it seems that the dominant
deformation mechanism of Sn5Sb alloy in the investigated range
of stress and temperature is dislocation climb, controlled by latticediffusion. The activation energies of 45 and 44 kJ mol1 , obtained
for the Sn5Sb1.5Bi and Sn5Sb1.5Cu alloys at high temperatures, are very close to the activation energy of about 40 kJ mol1 ,
reported for the grain boundary diffusion of -Sn [33]. Therefore,
it can be inferred that grain boundary diffusion could be the dominant deformation mechanism in the high-temperature regime of
the ternary alloys. On the other hand, at low temperatures, the
activation energies of 73 and 67 kJ mol1 are close to that of lattice self-diffusion of -Sn, and thus, it can be suggested that the
operative deformation mechanism in this condition is dislocation
climb.
3.5. Fractography
SEM fractographs of the samples tested at 340 and 400 K and
the costant strain rate of 5 104 s1 are shown in Fig. 10af. For
340 K, which had already shown the lowest level of elongation in
the form of a trough in the three-dimensional ductility plots in
Fig. 7, the fracture surface of all alloys display ductile dimples, that
are oriented in the loading direction. In the soft materials containing hard second phases, it is quite possible that the very ne
intermetallic particles along grain boundaries act as void nucleation
sites, and thus, stress or strain concentration can take place around
these second phase particles during uniaxial tensile test. Accordingly, the fracture of the present alloys most likely occurs by the
growth and coalescence of voids. The conical shape of the dimples
in many locations, is indicative of the higher plasticity of the Sn-rich
areas, in which the tips have formed. A comparison of Fig. 10a with
Fig. 10c and e indicates that adding Bi and Cu to the binary Sn5Sb
alloy renes the dimples size, consistent with the improved ductility of the ternary alloys. The fractographs of the alloys tested at
400 K, which had already exhibited the highest levels of ductility in
Fig. 7, show distinct differences with those tested at 340 K. As can
be observed, there is a general growth in the dimple size, which is
believed to be caused by a more advanced state of grain boundary
sliding and recrystallization. The coalescence of the voids is clear

409

in the fractographs of the ternary alloys shown in Fig. 10d and f.


The enhaced ductility of the alloys at 400 K is consistent with these
observations.
4. Conclusions
Tensile properties of Sn5Sb, Sn5Sb1.5Bi and Sn5Sb1.5Cu
lead-free solders were investigated at various strain rates ranging
from 5 104 to 1 102 s1 over the wide temperature range of
298400 K. The results are summarized as follows:
(1) Addition of 1.5% Bi and 1.5% Cu to the Sn5Sb solder improved
both strength and ductility of the base alloy, the effect which
was more pronounced for the Bi-containing alloy. This is an
important issue, which affects the impact resistance of the
joints manufactured by the solders.
(2) The achieved strength improvement is caused by solid solution
hardening effects of Bi and formation of Cu6Sn5 intermetallics
in the respective alloys. The enhanced ductility of the ternary
alloys can be attributed to structural renement caused by the
addition of the alloying elements.
(3) For all alloys, UTS increased with increasing strain rate and
decreased with increasing temperature. The ductility decreased
with increasing strain rate, but it exhibited some uctuation
with temperature. The lowest ductility levels were achieved
around 340 K, mainly due to limited grain boundary sliding,
which can initiate grain boundary cracks leading to a reduction
in the total elongation.
(4) According to the obtained stress exponents and activation energies, it is proposed that the dominant deformation mechanism
in Sn5Sb is dislocation climb over the whole temperature
range investigated. For the ternary alloys, however, dislocation climb and grain boundary diffusion are suggested to be
the dominant deformation mechanisms in the low- and hightemperature regimes, respectively.
References
[1] M. Abtew, G. Selvaduray, Mater. Sci. Eng. R 27 (2000) 95141.
[2] K.N. Tu, K. Zeng, Mater. Sci. Eng. R 34 (2001) 158.
[3] R. Mahmudi, A.R. Geranmayeh, H. Noori, N. Jahangiri, H. Khanbareh, Mater. Sci.
Eng. A 487 (2008) 2025.
[4] R. Mahmudi, A.R. Geranmayeh, H. Khanbareh, N. Jahangiri, Mater. Des. 30
(2009) 574580.
[5] R. Mahmudi, A.R. Geranmayeh, M. Salehi, H. Pirayesh, J. Mater. Sci.: Mater.
Electron. 21 (2010) 262269.
[6] M.D. Mathew, H. Yang, S. Movva, K.L. Murty, Metall. Mater. Trans. 36A (2005)
99105.
[7] R. Mahmudi, A.R. Geranmayeh, M. Allami, M. Bakherad, J. Electron. Mater. 36
(2007) 17031710.
[8] R. Mahmudi, A.R. Geranmayeh, A. Rezaee-Bazzaz, Mater. Sci. Eng. A 448 (2007)
287293.
[9] R. Mahmudi, A.R. Geranmayeh, M. Bakherad, M. Allami, Mater. Sci. Eng. A 457
(2007) 173179.
[10] R.K. Mahidhara, S.M.L. Sastry, I. Turlik, K.L. Murty, Scripta Metall. 31 (1994)
11451150.
[11] R.K. Mahidhara, S.M.L. Sastry, K.L. Jarina, I. Turlik, K.L. Murty, J. Mater. Sci. Lett.
13 (1994) 13871389.
[12] A.A. El-Daly, A. Fawzy, A.Z. Mohamad, A.M. El-Taher, J. Alloys Compod. 509
(2011) 45744582.
[13] A.A. El-Daly, Y. Swilem, A.E. Hammad, J. Mater. Sci. Technol. 24 (2008) 921925.
[14] Thermo-calc, ver.L. Stockholm, Sweden: Foundation for Computational Thermodynamics, 1996.
[15] S.K. Kang, S. Purushoth, J. Mater. Sci. 27 (1998) 11991204.
[16] W.J. Plumbridge, C.R. Gagg, J. Mater. Sci.: Mater. Electron. 10 (1999) 461468.
[17] I. Shohji, T. Yoshida, T. Takahashi, S. Hioki, Mater. Sci. Eng. A 366 (2004)
5055.
[18] A.A. El-Daly, A.E. Hammad, Mater. Sci. Eng. A 527 (2010) 52125219.
[19] M.L. Huang, L. Wang, Metall. Mater. Trans. 36A (2005) 14391446.
[20] R. Mahmudi, A.R. Geranmayeh, S.R. Mahmoodi, A. Khalatbari, Phys. Stat. Sol. (a)
204 (2007) 23022308.
[21] R. Mahmudi, A.R. Geranmayeh, S.R. Mahmoodi, A. Khalatbari, J. Mater. Sci.:
Mater. Electron. 18 (2007) 10711078.

410

M.J. Esfandyarpour, R. Mahmudi / Materials Science and Engineering A 530 (2011) 402410

[22] R. Mahmudi, A.R. Geranmayeh, H. Noori, H. Khanbareh, N. Jahangiri, Mater. Sci.


Technol. 26 (2010) 10011007.
[23] G.E. Dieter, G.E. Dieter (Eds.), Workability Testing Techniques, American Society
for Testing Metals, 1984, p. p. 11.
[24] P.M. Sargent, M.F. Ashby, Mater. Sci. Technol. 8 (1992) 594601.
[25] A.K. Mukherjee, J.E. Bird, J.E. Dorn, Trans. ASM 62 (1969) 155179.
[26] L. Rotherham, A.D.N. Smith, G.B. Greennough, J. Inst. Metal 79 (1951) 439454.
[27] F. Kabirian, R. Mahmudi, Metall. Mater. Trans. 40A (2009) 116127.

[28] A.A. El-Daly, Y. Swilem, A.E. Hammad, J. Alloys Compd. 471 (2009)
98104.
[29] S.N.G. Chu, J.C.M. Li, Mater. Sci. Eng. 39 (1979) 110.
[30] T.G. Langdon, Mater. Sci. Eng. A 283 (2000) 266273.
[31] R. Mahmudi, A.R. Geranmayeh, A. Rezaee-Bazzaz, J. Alloys Compd. 427 (2007)
124129.
[32] O.D. Sherby, E.M. Taleff, Mater. Sci. Eng. A 322 (2002) 8999.
[33] M. Fujiwara, M. Otsuka, Mater. Sci. Eng. A 319321 (2001) 929933.

You might also like