Abstract
Tensile and low cycle fatigue properties of Sn-5Sb (mass%) and Sn-10Sb (mass%) were investigated using miniature size specimens, and fracture behaviors of the specimens were observed. Tensile strength and 0.1% proof stress of both alloys decrease with increasing the temperature. The tensile strength and 0.1% proof stress of Sn-10Sb are higher than those of Sn-5Sb at 25°C. Elongation of Sn-5Sb decreases with increasing the temperature except for a strain rate of 2 × 10−1 s−1, while Sn-10Sb increases with increasing temperature. Although elongation of Sn-10Sb is lower than that of Sn-5Sb at 25°C, the difference between them is small at 150°C. Chisel-point fracture was observed in both alloys regardless of conditions of the tensile test. The low cycle fatigue lives of Sn-5Sb and Sn-10Sb alloys obey the Manson–Coffin equation, and the fatigue ductility exponent, , was 0.54 for Sn-5Sb and 0.46 for Sn-10Sb in the temperature range from 25°C to 150°C. On the basis of the observation of fractured specimens and the investigation of α, it was clarified that the crack progress can be delayed by the formation of coarse SbSn compounds in the Sn-Sb alloy, and thus the fatigue properties can be improved.
1. Introduction
Due to environmental regulations such as the Restriction of Hazardous Substances (RoHS) directive, lead-free soldering has spread to the world. Lead-free solder is classified broadly into three types depending on the temperature range: high-temperature solder [1–3], medium-temperature solder [3], and low-temperature solder [3, 4]. In general, high-temperature lead-free solder has been used for power semiconductor devices as a die-attach material. Pb-rich solder containing from 5 to 10 mass% Sn has been used for such applications. However, lead-free high-temperature solder alloys have not been regulated by RoHS directive yet, and such solders are expected to be developed [1, 5–8]. These solders need to be equipped with a combination of good mechanical properties, high thermal fatigue properties, and high heat resistance.
Sn-Sb alloys are one of the candidates for the substitution of the Pb-rich solders [9–11]. It has been reported that Sn-5Sb (mass%) has excellent thermal fatigue behaviors and relatively high fracture strength [12, 13]. In addition, in Sn-Sb alloys with high concentration of Sb, mechanical strength is improved by solid-solution of Sb in -Sn matrix and dispersion of SbSn compounds [14].
The aim of this study is to compare tensile and low cycle fatigue properties of the Sn-5Sb (mass%) and Sn-10Sb (mass%) alloys using miniature size specimens, which have microstructures similar to real solder joints [15, 16]. Moreover, fracture behaviors of the specimens after tensile and fatigue tests were investigated.
2. Materials and Methods
2.1. Specimen Preparation
Ingots of Sn-5Sb and Sn-10Sb alloys were prepared. Melting properties of both alloys were investigated by differential scanning calorimetry (DSC). When the DSC was measured, the weight of the specimen was adjusted around 10 mg. The measurement was conducted at a heating rate of 10°C/min in the temperature range from 30°C to 300°C. From the DSC measurement result, the solidus temperature and the melting finish temperature of the alloys were estimated to be 240°C and 248°C for Sn-5Sb and 248°C and 254°C for Sn-10Sb, respectively. Solder wire with 1.2 mm diameter was fabricated by drawing the ingot. As-cast miniature size specimens with 2.0 ± 0.2 mm gage length and 0.50 ± 0.05 mm diameter were fabricated from solder wire by casting [15]. A divided metal mold made of the 2219 aluminum alloy was prepared. The metal mold with solder wire was heated from 20°C to casting temperature. Casting was conducted at the temperature that is 10°C higher than the melting finish temperature. The maximum cooling rates in casting that were measured with a thermocouple put in the metal mold were 3.5°C/s and 3.6°C/s for Sn-5Sb and Sn-10Sb, respectively. In this study, as-cast miniature size specimens were used for the tensile test and the low cycle fatigue test.
2.2. Microstructural Observation
To observe microstructures of as-cast miniature size specimens, specimens were embedded in epoxy resin and cross sections were made perpendicular to the longitudinal direction. Afterwards, the cross sections were polished with #500–#4000 SiC papers and 1 μm alumina powder suspension. After polishing, the microstructural observation for the cross section of the specimen was conducted using a laser microscope and an electron probe X-ray microanalyzer (EPMA).
2.3. Tensile Test
Tensile tests were conducted at strain rates ranging from 2 × 10−3 s−1 to 2 × 10−1 s−1 at 25°C and 150°C. The test was conducted with a micro load test system (Saginomiya Seisakusho LMH207-10). In the test system, a linear servo motor that has an air-bearing system was used for an actuator to avoid mechanical contact. A linear encoder was used for a displacement detector, and the displacement resolution of the detector was within ±0.1 μm. Five specimens were tested under each condition. After the test, appearances and fracture surfaces of specimens were observed with the EPMA.
2.4. Low Cycle Fatigue Test
A low cycle fatigue test was also conducted with the micro load test system. The test was conducted at a strain rate of 2 × 10−3 s−1 at 25°C and 150°C. In the test, strain was controlled using a symmetrical triangle wave, and the total strain range, , was controlled in the range of 0.4% to 2.0%. Figure 1 shows an example of the stress-strain hysteresis loop at the tenth cycle in the fatigue test. The inelastic strain range, , was defined as shown in Figure 1. The fatigue life was defined as the number of cycles in which the maximum load reduced by 20%. After the test, appearances and fracture surfaces of specimens were observed by the laser microscope and the EPMA. Moreover, the surface roughness of the fracture surfaces of the specimens were tested at the total strain range of 0.8% and 2.0%, which were measured at 20 fields by using a laser microscope, and then the average of maximum height, Rz, was calculated.
3. Results and Discussion
3.1. Microstructure of Miniature Size Specimen
Figures 2 and 3 show back-scattered electron (BSE) images of as-cast miniature size specimens of Sn-5Sb and Sn-10Sb, respectively. In both alloys, very fine dark-gray and bright-gray phases with submicrometer sizes were observed.
Figures 4 and 5 show optical images of as-cast miniature size specimens and the EPMA mapping analysis results of Sn-5Sb and Sn-10Sb, respectively.
Figure 6 shows the Sn-Sb binary phase diagram [17]. From the phase diagram, both Sn-5Sb and Sn-10Sb consist of β-Sn and SbSn phases. On the basis of the EPMA mapping analysis result and the Sn-Sb binary phase diagram, the dark-gray phase and the bright-gray phase, as shown in Figures 2 and 3, were inferred to be the β-Sn phase and the SbSn compound, respectively.
In Sn-10Sb, the distribution of 20∼40 µm sizes of particles was observed in the optical image. From the EPMA mapping analysis result, the particles were inferred to be the SbSn compounds. Therefore, the formation of coarsened SbSn compounds occurs in the Sn-Sb alloy with increasing the Sb content.
3.2. Tensile Properties and Fracture Modes
Figures 7 and 8 show typical stress-strain curves of miniature size specimens of Sn-5Sb and Sn-10Sb, respectively. Although the tensile strength of Sn-5Sb tends to increase with increasing the strain rate regardless of the test temperature, this tendency is not observed in Sn-10Sb at room temperature. Elongation of Sn-5Sb at 150°C and Sn-10Sb at both temperatures tends to increase with increasing the strain rate.
(a)
(b)
(a)
(b)
Figure 9 shows the effects of the strain rate and the test temperature on mechanical properties of Sn-5Sb [18] and Sn-10Sb alloys. Tensile strength and 0.1% proof stress of both alloys decrease with increasing the temperature. The tensile strength and 0.1% proof stress of Sn-10Sb are higher than those of Sn-5Sb at 25°C. It has been reported that the Berkovich hardness of the SbSn phase (approximately 1200 BHV) is four times that of the β-Sn (approximately 300 BHV) [19]. An increase of SbSn compounds in the Sn-Sb alloy with an increase of the Sb content results in increase of the tensile strength and 0.1% proof stress of the alloy. However, differences in them between both alloys are negligible at 150°C. From the binary Sn-Sb phase diagram, as shown in Figure 6, SbSn compounds decompose at high temperature so that the ratio of the SbSn compound decreases, and that of the β-Sn phase increases when the temperature rises to 150°C. Elongation of Sn-5Sb decreases with increasing the temperature except for a strain rate of 2 × 10−1 s−1, while that of Sn-10Sb increases with increasing the temperature. Elongation of Sn-5Sb is approximately 50% at the strain rate of 2 × 10−1 s−1, regardless of the temperature. Although elongation of Sn-10Sb is lower than that of Sn-5Sb at 25°C, the difference between them is small at 150°C.
(a)
(b)
(c)
Figures 10 and 11 show secondary electron images of Sn-5Sb and Sn-10Sb fractured specimens, respectively. In both alloys, chisel-point fracture was observed regardless of conditions of the tensile test. The tip of the break region was a straight line. It has been reported that a miniature size specimen consists of a few grains or a single grain so that the mechanical properties of the specimen strongly depend on the crystal orientation of the grain [15]. Since the slip systems of Sn that has tetragonal crystal structure are limited [20], the fractured area is not deformed uniformly, and thus, it is easily reduced in a straight line.
(a)
(b)
(a)
(b)
3.3. Low Cycle Fatigue Properties and Fracture Modes
Figure 12 shows the results of the low cycle fatigue tests with miniature size specimens of Sn-5Sb and Sn-10Sb, respectively. Typically, low cycle fatigue life of a solder alloy obeys the Manson–Coffin equation as follows:where is plastic strain range, is fatigue ductility factor, is fatigue life, and is fatigue ductility exponent.
From Figure 12, it was confirmed that low cycle fatigue lives of both alloys obey the Manson–Coffin equation, and fatigue ductility exponents, α, are 0.54 for Sn-5Sb and 0.46 for Sn-10Sb, regardless of the temperature. In general, the isothermal fatigue life of a solder alloy degrades with increasing temperature [15]. However, Sn-Sb alloys exhibit no temperature dependence and have excellent isothermal fatigue properties in temperatures ranging from 25°C to 150°C.
Compared with Sn-5Sb and Sn-10Sb, α of Sn-10Sb was a little smaller than that of Sn-5Sb. This means that Sn-10Sb has superior fatigue properties compared with Sn-5Sb.
Figures 13 and 14 show fracture surfaces of specimens of Sn-5Sb and Sn-10Sb after the low cycle fatigue test. In Sn-5Sb, the fracture surface mostly formed with an angle of 45 degrees with respect to a tensile-compression axis direction regardless of strain rate and temperature. In Sn-10Sb, the fracture surface relatively becomes the shape with many irregularities. Table 1 shows the maximum height Rz of the surface roughness of the Sn-5Sb and Sn-10Sb fracture surfaces. From this table, the surface roughness of the Sn-10Sb fracture surface was larger than that of Sn-5Sb. Figure 15 shows the cross-sectional microstructure of the fractured specimen of the Sn-10Sb alloy after the low cycle fatigue test. In the figure, coarse SbSn compounds are observed in the fracture surface area. Since SbSn compounds in the Sn-Sb alloy are hard to break, the crack progress turns around the outer circumference of the SbSn compounds. Thus, the crack growth in Sn-10Sb, in which coarsened SbSn compounds occurred in as-cast microstructure (refer to Figure 5), is slower than that in Sn-5Sb, and the fracture surface of Sn-10Sb was very rough. As a result, it was clarified that the crack progress can be delayed by the formation of the coarse SbSn compounds in the Sn-Sb alloy, and thus, the fatigue properties can be improved.
(a)
(b)
(a)
(b)
4. Summary
In this study, tensile and fatigue properties of Sn-5Sb and Sn-10Sb alloys were compared using miniature size specimens. The results are listed as follows.(1)β-Sn phases and SbSn compounds with submicrometer sizes were observed in Sn-5Sb, while the distribution of 20∼40 µm size of SbSn compounds was also observed in Sn-10Sb.(2)Tensile strength and 0.1% proof stress of both alloys decrease with increasing the temperature. The tensile strength and 0.1% proof stress of Sn-10Sb are higher than those of Sn-5Sb at 25°C, while differences in them between both alloys are negligible at 150°C. Elongation of Sn-5Sb decreases with increasing the temperature except for a strain rate of 2 × 10−1 s−1, while that of Sn-10Sb increases with increasing the temperature. Elongation of Sn-5Sb is approximately 50% at the strain rate of 2 × 10−1 s−1, regardless of the temperature. Although elongation of Sn-10Sb is lower than that of Sn-5Sb at 25°C, the difference between them is small at 150°C.(3)The low cycle fatigue lives of Sn-5Sb and Sn-10Sb alloys obey the Manson–Coffin equation, and fatigue ductility exponents, , were 0.54 for Sn-5Sb and 0.46 for Sn-10Sb in the temperature range from 25°C to 150°C.(4)On the basis of the observation of fractured specimens and the investigation of α, it was clarified that the crack progress can be delayed by the formation of the coarse SbSn compounds in the Sn-Sb alloy, and thus the fatigue properties can be improved.
Conflicts of Interest
The authors declare that they have no conflicts of interest.