Simplification of Low-Temperature Sintering ... - Springer Link

2 downloads 0 Views 922KB Size Report
Apr 13, 2013 - YUNHUI MEI,2 GANG CHEN,1,3 YUNJIAO CAO,1 XIN LI,2 DAN HAN,1 and XU CHEN1. 1.—School of Chemical Engineering and Technology, ...
Journal of ELECTRONIC MATERIALS, Vol. 42, No. 6, 2013

DOI: 10.1007/s11664-013-2561-8 Ó 2013 TMS

Simplification of Low-Temperature Sintering Nanosilver for Power Electronics Packaging YUNHUI MEI,2 GANG CHEN,1,3 YUNJIAO CAO,1 XIN LI,2 DAN HAN,1 and XU CHEN1 1.—School of Chemical Engineering and Technology, Tianjin University, Tianjin, People’s Republic of China. 2.—Tianjin Key Laboratory of Advanced Joining Technology and School of Materials Science and Engineering, Tianjin University, Tianjin, People’s Republic of China. 3.—e-mail: [email protected]

Conventional solders cannot meet the requirements for high-temperature applications. Recently, a low-temperature sintering technique involving a nanosilver paste has been developed for attaching semiconductor chips to substrates. Sintered nanosilver joints showed high reliability in high-temperature applications. We used the nanosilver paste to attach 10 mm 9 10 mm chips by introducing a pressure as low as only 1 MPa during drying at 185°C. Die-shear tests showed that shear strengths of higher than 50 MPa could be generated by applying 5 MPa at 225°C for only 10 s or 1 MPa at 150°C for 600 s, followed by sintering for only 60 s at 275°C. The sintering temperature could be reduced to 250°C in most applications with a slight reduction in shear strength. As a result of good bonding, significant plastic flow and ductile fracture of the sheared silver joint could be observed by scanning electron microscopy (SEM). SEM also showed that the fracture of the sheared silver joint was a cohesive failure. Key words: Nanosilver, sintering, pressure-assisted, electronic packaging, shear strength, microstructures

INTRODUCTION Power electronics systems have wide industrial applications, ranging from inverters for hybrid electric vehicles to power converters in wind generators. Multichip modules consisting of large chips attached to substrates are some of the most important units in these power systems. Today, chips in such modules are bonded to the substrate by using conductive adhesives1 or lead-free and high-lead solders by applying a solder reflow process at temperatures from 260°C to over 300°C to solidify the solder alloy.2 However, there are many disadvantages of solder alloys, e.g., low electrical and thermal conductivity. In addition, soldered chip joints are susceptible to fatigue failure under cyclic loading due to creep and accumulation of inelastic strain.3 Therefore, if the junction temperature is (Received October 9, 2012; accepted March 1, 2013; published online April 13, 2013)

raised to 175°C, or even higher as envisioned for some future systems, the reliability of the soldered joint will become an even greater concern. In the quest for 175°C power modules,4,5 the low melting temperature of solder attachments is one of the foremost challenges to overcome. The increased operating temperature approaches the melting point of conventional solders, making the joint susceptible to a host of thermomechanical and metallurgical problems. The European power electronics industry is leading the way in introducing a superior lead-free die-attach technology to the marketplace.6 Unlike the widely used soldering or adhesive bonding technologies,1 this new technology, often referred to as low-temperature joining technology (LTJT),7–12 is based on sintering of micrometer-sized silver powder at temperatures below 300°C. Usually, a screen- or stencil-printed layer of micron-scaled silver flakes is used as the interconnection material. To get such low sintering temperatures, pressure of 1209

1210

G. Chen, Cao, Mei, Li, Han, and X. Chen

about 40 MPa on a 100 mm2 chip is required. Under the applied pressure, the silver die-attach layer undergoes significant densification to density of 80% at 250°C for a few minutes. However, when the pressure is applied, even the slightest irregularities can lead to cracking of the brittle silicon chips or ceramic substrates.13 The sintered silver joints were reported to have excellent thermal conductivity and electrical conductivity, high shear strength in excess of 30 MPa, and high reliability.14–17 Recently, we demonstrated18–22 a strategy of replacing the high pressure with a chemical driving force by using nanosilver powder to lower the sintering temperature. Silver is considered as the best conductor and can be stably dispersed in solvents.23 So, the nanosilver powder was made into a paste with solvents to offer a one-to-one replacement for solder or epoxy paste. The introduction of the nanosilver paste significantly simplifies low-temperature joining or sintering technology (LTJT or LTST) and has paved the way for widespread adoption of nanosilver paste by power electronics manufacturers. Some previous research focused on the reliability of nanosilver joints, e.g., mechanical behavior,24–29 transient thermal impedance,30–32 and silver migration.33–35 However, there have been few reports on how to improve the bonding quality of nanosilver joints for the attachment of large-area chips such as insulated gate bipolar transistors (IGBTs) and fast recovery diodes (FRDs). The purpose of this study is to evaluate the effect of different LTST parameters such as drying pressure and sintering time on the shear strength and microstructure of nanosilver joints for bonding large-area (10 mm 9 10 mm) semiconductor power devices.

Fig. 1. Typical load versus displacement trace from a sintered sample.

EXPERIMENTAL PROCEDURES Nanosilver paste was made by adding selected organic surfactant, binder, and thinner to 30-nm nanosilver particles.36 A transmission electron microscopy (TEM) micrograph of nanosilver particles is shown in Fig. 1. The nanosilver paste can be readily processed by common surface-mount techniques, such as screen/stencil printing or syringe dispensing. The silver paste containing spherical nanoscale particles for die attachment was obtained from NBE Tech, LLC.37 Figure 2 shows thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) traces of the nanosilver paste heated in air.2 Weight loss started immediately upon heating because of evaporation of solvent from the paste. After much of the solvent evaporates, a further loss in weight results from an exothermic reaction with a peak at around 150°C. A second exothermic reaction, which also produces a weight loss, shows a peak at about 256°C. The exothermic reaction is caused by the burn-out of the organics added into the nanosilver paste. Finally, all the organics appear to have been removed, as evident from the flat TGA trace beyond

Fig. 2. TEM micrograph of nanosilver particles.

313°C. The silver content in the paste is approximately 85 wt.% after the third exothermic reaction. Based on the TGA results, the heating profile can be divided into three stages. The first stage not only removes some solvents but also ensures good contact of the first dried silver layer with the chip. The first exothermic peak occurs at about 150°C. Therefore, the drying temperature cannot exceed 150°C. We put the printed nanosilver paste on different hot plates with different temperatures. When the hot plate temperature was 50°C, the paste remained wet and soft. When it was 90°C, the paste was too dry to press. When it was 70°C, the

Simplification of Low-Temperature Sintering Nanosilver for Power Electronics Packaging

paste become hard but retained some wettability. As a result, the drying was carried out at 70°C for 10 min. The second stage aims at removing most of the binder, which can be decomposed and burned off at 185°C. Besides, a low pressure is applied on the chip to maintain good contact between the second dried silver layer and the chip. Thus, the drying pressure, P, and the pressing time, tp, were treated as primary parameters in this study. The last stage is to remove all the organics and complete the sintering to form a satisfactory bond that is electrically and thermally conductive. In this paper, the influence of the pressure, P, the pressing time, tp, the temperature during pressing, Tp, the sintering temperature, Ts, and the sintering time, ts, on the shear strength of the sintered silver joint are discussed based on die-shear testing results. For die-shear testing, a copper substrate with dimensions of 30 mm 9 22 mm 9 1.5 mm was electroplated with a 10-lm-thick silver film. The nanosilver paste was stencil-printed onto the copper substrate with thickness of 90 lm before drying. Then, the specimen was subjected to the temperature profile shown in Fig. 3a to form an interconnection between the chip and the substrate. Figure 3b shows a schematic of the sintering process. The specimen was first predried for 10 min at 70°C on a hot plate. A 10 mm 9 10 mm copper dummy chip, which was also electroplated with a 10-lm-thick silver film, was then placed on the

dried silver layer. The specimen was heated at Tp. Meanwhile, a pressure, P, was applied on the specimen to assist the formation of good bonds by hot pressing. The pressing time at this stage, tp, was varied to study its influence on the bonding quality. For the sintering stage, the specimen was rapidly moved to the hot plate, which had already been preheated to 275°C, and maintained at 275°C for ts minutes. The drying and sintering conditions are summarized in Table I. At least four specimens were prepared for each test. Figure 4 presents a specimen in which a largearea dummy chip is attached to a substrate by the nanosilver paste. Figure 5a shows that the bondline thickness of the nanosilver joint is about 50 lm after the sintering process. Figure 5b shows the strong bonding between the silver layer and copper. The shear strength of a specimen was measured using a bond tester (CONDOR 150; XYZTEC) at a velocity of 100 lm/s. A schematic diagram of the dieshearing setup is shown in Fig. 6. RESULTS AND DISCUSSION Figure 7 shows a typical load versus displacement plot obtained from a sintered sample. From the maximum load at which the sample broke, the die-

Table I. Summary of processing conditions of specimens for die-shear testing Test No.

Fig. 3. TGA and DSC traces of nanosilver paste heated in air at 10°C/min.

1211

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

P (MPa)

Tp (°C)

Ts (°C)

tp (s)

ts (s)

5 5 5 5 5 5 5 5 5 3 3 3 3 3 3 3 3 3 3 3 3 3 1 1 1 1 1 1 1 1

225 225 225 225 225 225 225 225 185 225 225 225 225 225 225 185 185 185 185 150 150 150 225 225 225 225 225 225 225 185

275 275 275 275 275 275 275 275 275 225 250 275 275 275 300 275 275 275 275 275 275 275 275 275 275 275 275 275 275 275

600 60 30 10 10 10 10 10 60 60 60 120 60 10 60 600 120 60 10 600 60 10 600 60 10 10 10 10 10 60

600 600 600 600 180 60 30 0 600 600 600 600 600 600 600 600 600 600 600 600 600 600 60 60 60 0 30 180 600 600

1212

G. Chen, Cao, Mei, Li, Han, and X. Chen

Specimen

Dummy Cu chip Silver layer Chip

Shear Direction

Pusher

Substrate

Fig. 4. Sintering process of the nanosilver paste: (a) temperature profile of nanosilver paste sintering, (b) schematic of the sintering process.

Gas pipe

Clamp

Fig. 6. Bond-line of a sintered nanosilver joint: (a) metallographic micrograph, (b) SEM micrograph.

Fig. 7. Die-shear tester and clamp.

shear strength is calculated to be around 43 MPa. The large displacement obtained in the test is around 2.03 mm. Effect of Applied Pressure, P

Fig. 5. Sample in which a dummy chip is attached onto a substrate by nanosilver paste.

First, the samples were preheated at 70°C for 10 min. Then, they were heated at 185°C or 225°C for 10 s at different applied pressures, i.e., 1 MPa, 3 MPa, and 5 MPa. Finally, they were placed on the hot plate, which was preset to 275°C, for 10 min.

Simplification of Low-Temperature Sintering Nanosilver for Power Electronics Packaging

Figure 8 shows how the shear strength of the nanosilver joint changes with the applied pressure, P, which ranges from 1 MPa to 5 MPa. At 225°C, the shear strengths are 11.6 MPa, 28.4 MPa, and 52.5 MPa at P of 1 MPa, 3 MPa, and 5 MPa, respectively. At 185°C, the corresponding pressures are 24.5 MPa, 34.2 MPa, and 42.2 MPa. The shear strength at 1 MPa is obviously higher at 185°C than at 225°C because most of the solvents and binders in the nanosilver paste were probably burned out at 225°C, leaving behind a hard layer of nanosilver, which an applied pressure of 1 MPa cannot deform,

Fig. 8. Effect of applied pressure, P, on shear strength.

1213

resulting in a decrease in the contact area of the joint. Figure 9a–c show fracture surfaces of sintered nanosilver joints obtained at 225°C under applied pressures of 1 MPa, 3 MPa, and 5 MPa, respectively. As shown in Fig. 9a, no significant plastic flow, i.e., a kind of deformation without the ability to be restored, was observed on the fracture surface of the sheared nanosilver joint, indicating that the shear strength of the nanosilver joint in this case is low. This observation is consistent with the results of the shear strength of the nanosilver joint in the case of 1 MPa (Fig. 8). However, significant plastic flow presents in Fig. 9b and c, implying that higher shear strength of the sintered joint could be obtained by applying 3 MPa or 5 MPa as a result of ductile fracture in the joints. On the basis of the Mackenzie–Shuttleworth sintering model,38,39 the densification rate during sintering depends on the applied pressure. Therefore, the sintered joint obtained under an applied pressure of 5 MPa should have the highest shear strength because of the highest densification rate. Figure 8 shows that, the higher the applied pressure, the higher the shear strength. Since the shear strengths in the range of 30 MPa to 55 MPa of sintered joints obtained under applied pressures of 3 MPa and 5 MPa are comparable to those of conventional solders,32–34 there seems to be no need to apply a pressure higher than 5 MPa. The process of applying 3 MPa pressure at 185°C seems to be preferable for attaching 10 mm 9 10 mm chips.

Fig. 9. SEM micrographs of fracture surfaces of sintered nanosilver joints under different applied pressures P: (a) 1 MPa, (b) 3 MPa, and (c) 5 MPa.

1214

Fig. 10. Effect of pressing temperature, Tp, on shear strength.

Fig. 11. Effect of sintering temperature, Ts, on shear strength.

G. Chen, Cao, Mei, Li, Han, and X. Chen

Fig. 12. Effect of pressing time, tp, on shear strength under applied pressures of 5 MPa, 3 MPa, and 1 MPa.

for 60 s is enough to obtain shear strengths comparable to those of conventional solders.32–34 When the temperature is increased from 185°C to 225°C for 60 s of pressing time, there is a slight drop of the bonding strength. This is because, when the temperature exceeds a certain level, the silver particles will grow large enough to reduce the sintering driving force, and most of the solvents and binders in the nanosilver paste were probably burned out at 225°C, leaving behind a hard layer of nanosilver, which would result in a decrease in the contact area of the joint. This will reduce the bonding strength. We also investigated the effect of sintering temperature on the shear strength of the sintered silver joints, which were heated sequentially as follows: (1) 70°C for 10 min, (2) 225°C for 60 s under a pressure of 3 MPa, and (3) Ts for 10 min. The shear strength of the sintered silver joints increases with the sintering temperature, Ts, as shown in Fig. 11. However, shear strength higher than 30 MPa can still be obtained by sintering at 250°C for 10 min. It seems that sintering at 250°C for 10 min is good enough for most applications.

Effect of Pressing Temperature, Tp, and Sintering Temperature, Ts

Effect of Pressing Time, tp

Figure 10 shows the dependence of the shear strength of the nanosilver joint on Tp for different pressing times tp (10 s, 60 s, and 600 s) under an applied pressure of 3 MPa. It can be seen that the shear strength is significantly improved by prolonging the pressing time to 600 s at a pressing temperature of 150°C. The shear strength increases with Tp when the pressing time is 10 s. This increase is much smaller at 185°C and 225°C than at 150°C. It seems that pressing at 185°C or 225°C

Figure 12 shows the dependence of the shear strength of the nanosilver joint on the pressing time, tp, in the range from 10 s to 600 s. The shear strength of the sintered nanosilver joints increases with increasing tp. However, at the applied pressure of 1 MPa, we could not significantly improve the shear strength by prolonging tp. Based on Ivensen’s sintering theory, the densification in the sintering process is considered to eliminate crystal defects.40 The decrease in the

Simplification of Low-Temperature Sintering Nanosilver for Power Electronics Packaging

1215

Fig. 13. SEM micrographs of fracture surfaces of sintered nanosilver joints under an applied pressure of 5 MPa for different pressing times, tp: (a) 10 s, (b) 30 s, (c) 60 s, and (d) 600 s.

Fig. 14. SEM micrographs of fracture surfaces of sintered nanosilver joints under an applied pressure of 1 MPa for different pressing times, tp: (a) 10 s, (b) 60 s, and (c) 600 s.

concentration of crystal defects reduces the densification rate, since the macro flow is restricted by crystal defects. In the case of nanosilver joints, no

matter what the type of defect is, the concentration of crystal defects is apt to be reduced by the sintering time and the applied pressure because of

1216

defect restoration in the sintered nanosilver joint.21 The increase in shear strength with increasing tp at 5 MPa or 3 MPa can be explained by the reduction in concentration of crystal defects. The longer the tp, the easier the elimination of crystal defects. However, the shear strength does not change with

G. Chen, Cao, Mei, Li, Han, and X. Chen

increasing tp at 1 MPa, probably because an applied pressure of 1 MPa is too low to eliminate the crystal defects. Therefore, obvious plastic deformation can be observed in the sheared joint under an applied pressure of 5 MPa (Fig. 13). On the other hand, Fig. 14 shows no significant plastic flow in the joint under an applied pressure of 1 MPa. It is concluded that 1 MPa is too low for the formation of strong metallic bonding at the interface between die/substrate and nanosilver joint. As a result, only adhesive failure occurs in the sheared joints under an applied pressure of 1 MPa, and no significant improvement in shear strength can be obtained by increasing the pressing time under 1 MPa. Figure 12 shows that a shear strength of about 30 MPa can be obtained at 185°C under an applied pressure of 3 MPa for over 60 s. Therefore, the parameter combination of tp of 60 s, P of 3 MPa, Tp of 185°C, and Ts of 275°C can be another preferred way for attaching 10 mm 9 10 mm chips. Effect of Sintering Time, ts

Fig. 15. Effect of sintering time, ts, on shear strength under applied pressures of 1 MPa and 5 MPa.

Figure 12 also shows that good metallic bonds could be generated in the sintered joint by applying a pressure of 5 MPa for only 10 s. Figure 15 shows the dependence of the shear strength of the nanosilver joint on the sintering time, ts, in the range from 0 s to 600 s. The shear strength increases with increasing ts, but prolonging ts beyond 60 s has no significant effect on the shear strength.

Fig. 16. SEM micrographs of fracture surfaces of sintered nanosilver joints under an applied pressure of 5 MPa for different ts: (a) 0 s, (b) 60 s, (c) 180 s, and (d) 600 s.

Simplification of Low-Temperature Sintering Nanosilver for Power Electronics Packaging

1217

Fig. 17. Macrograph of fracture surfaces of nanosilver joints under two different applied pressures.

According to the Mackenzie–Shuttleworth sintering model,38,39 the densification rate during sintering depends on the sintering temperature and sintering time. The particle size increases as ts increases, as shown in Fig. 16. Moreover, no sintered silver is left on the copper dummy chip in the sheared joint under an applied pressure of 1 MPa, as shown in Fig. 17. Since roughness of the substrate has a significant effect on the reproducibility of low-temperature sintering nanosilver41 and the surface of the copper chip or substrate is rough to some extent, the nanosilver comes into contact with the copper dummy chip or substrate only at the protrusions of the surface. As a result, pressure is needed to increase the contact area of the interface. The higher the applied pressure, the higher the shear strength.42 When the pressure is low, the contact area of the interface is limited, resulting in a weak interface.43 Consequently, prolonging the sintering time has no effect on the shear strength because of the weak interface under an applied pressure of 1 MPa. It is concluded that an applied pressure of 1 MPa is too low to assist in forming good bonding for low-temperature sintering nanosilver paste. Fracture occurs mostly at the interface between the sintered silver and the substrate in the case of 1 MPa.

2. With increasing applied pressure P, pressing temperature Tp, pressing time tp, sintering temperature Ts, and sintering time ts, the shear strength of the sintered silver joint mostly increased. 3. A sintering temperature of 250°C was high enough to obtain shear strength comparable to that of conventional solders. 4. A pressure of 1 MPa was too low to form good bonding at the interface between the sintered nanosilver and chip or substrate. Higher pressures were preferred for generating shear strength higher than 40 MPa for most power electronic applications. 5. The optimum sintering process consisted of three stages: (i) drying at 70°C for 10 min, (ii) heating at 185°C for 120 s under a pressure of 3 MPa, and (iii) sintering at 250°C for 10 min. This process was much simpler than the conventional LTJT, which required a pressure of 40 MPa during sintering. ACKNOWLEDGEMENTS The authors gratefully acknowledge financial support from the National Natural Science Foundation of China (No. 51101112, No. 10802056, No. 11172202, No. 11072171, and No. 51175375). REFERENCES

CONCLUSIONS In this study, we developed a low-temperature sintering technique to be used instead of lead-free soldering for attaching large-area (>100 mm2) chips, e.g., IGBTs, for high-temperature applications and obtained the following conclusions: 1. For attaching large-area chips, a low applied pressure of 3 MPa was necessary to generate strong shear strengths higher than 50 MPa.

1. Y.H. Mei, X. Chen, and H. Gao, J. Electron. Mater. 38, 2415–2426 (2009). 2. T.G. Lei, J.N. Calata, G.-Q. Lu, X. Chen, and S. Luo, IEEE Trans. Compon. Packag. Technol. 33, 98–104 (2010). 3. Q.K. Zhang, Q.S. Zhu, H.F. Zou, and Z.F. Zhang, Mater. Sci. Eng. A 527, 1367–1376 (2010). 4. X. Cao, K.D.T. Ngo, and G.-Q. Lu, International Conference on Electronic Packaging Technology & High Density Packaging (ICEPT-HDP, 2009). 5. X. Cao, T. Wang, Z. Tan, K. Ngo, S. Luo, and G.-Q. Lu, 2nd IMAPS Conference on Automotive Microelectronics and Packaging (Dearborn, Michigan).

1218 6. P. Beckedahl, T. Grasshoff, and M. Lederer, International Exhibition & Conference for Power Electronics Intelligent Motion Power Quality (Germany: Nuremberg, 2007). 7. U. Scheuermann and P. Wiedl, Workshop on Metal Ceramic Composites Functional Application (Vienna, Austria,1997). 8. H. Schwarzbauer, Method of Securing Electronic Components to a Substrate (U.S. 4810672, 1989). 9. R. Amro, J. Lutz, J. Rudzki, R. Sittig, and M. Thoben, IEEE International Symposium on Power Semiconductor Devices and IC’s (ISPSD) (Naples, Italy, 2006). 10. C. Go¨bl and J. Faltenbacher, 6th International Conference on Integrated Power Electronics Systems (CIPS) (Nuremburg, Germany, 2010). 11. Z. Zhang and G.-Q. Lu, IEEE Trans. Electron. Packag. Manuf. 25, 279–283 (2002). 12. A.D. Albert, M.F. Becker, J.W. Keto, and D. Kovar, Acta Mater. 56, 1820–1829 (2008). 13. M. Knoerr, S. Kraft, and A. Schletz, 12th Electronics Packaging Technology Conference (EPTC) (Singapore, 2010). 14. G.-Q. Lu, J.N. Calata, G. Lei, and X. Chen, International Conference on Thermal, Mechanical and Multi-Physics Simulation Experiments in Microelectronics and MicroSystems (Berlin: Germany, 2007). 15. T. Wang, G. Lei, X. Chen, L. Guido, K. Ngo, and G.-Q. Lu, International Conference on Electronic Packaging Technology & High Density Packaging (Beijing, China, 2009). 16. X. Cao, T. Wang, K.D.T. Ngo, and G.-Q. Lu, IEEE Trans. Compon. Packag. Manuf. Technol. 1, 495–501 (2011). 17. X. Li, X. Chen, and G.-Q. Lu, 60th Electronic Components and Technology Conference (Las Vegas, US, 2010). 18. J.G. Bai, Z.Z. Zhang, J.N. Calata, and G.-Q. Lu, IEEE Trans. Compon. Packag. Technol. 29, 589–593 (2006). 19. J.G. Bai, J. Yin, Z. Zhang, G.-Q. Lu, and J.D. van Wyk, IEEE Trans. Adv. Packag. 30, 506–510 (2007). 20. Z. Zhang, J.N. Calata, J.G. Bai, and G.Q. Lu, Materials Processing and Manufacturing Division Fifth Global Symposiun (Charlotte, NC, 2004). 21. T. Wang, X. Chen, G.-Q. Lu, and T.G. Lei, J. Electron. Mater. 36, 1333–1340 (2007). 22. J.G. Bai, T.G. Lei, J.N. Calata, and G.-Q. Lu, J. Mater. Res. 22, 3494–3500 (2007). 23. B.J. Kang and J.H. Oh, Thin Solid Films 518, 2890–2896 (2010).

G. Chen, Cao, Mei, Li, Han, and X. Chen 24. D.J. Yu, X. Chen, G. Chen, G.-Q. Lu, and Z.Q. Wang, Mater. Des. 30, 4574–4579 (2009). 25. G. Chen, X.-H. Sun, P. Nie, Y.H. Mei, G.-Q. Lu, and X. Chen, J. Electron. Mater. 41, 782–790 (2012). 26. X. Li, G. Chen, X. Chen, L. Wang, and Y.H. Mei, Solder. Surf. Mount. Technol. 24, 120–126 (2012). 27. X. Li, G. Chen, X. Chen, G.-Q. Lu, L. Wang, and Y.-H. Mei, Solder. Surf. Mt. Technol. 24, 120–126 (2012). 28. X. Li, G. Chen, X. Chen, G.-Q. Lu, L. Wang, and Y.-H. Mei, Microelectronics Reliability (2012). 29. Y. Mei, G. Chen, G.-Q. Lu, and X. Chen, Int. J. Adhes. Adhes. 35, 88–93 (2012). 30. X. Cao, T. Wang, G-Q. Lu, and K.D.T. Ngo, Proceedings of IEEE International Power Electronics Conference (IPEC) (Sapporo, Japan, 2010). 31. G. Chen, D. Han, Y.-H. Mei, X. Cao, T. Wang, X. Chen, and G.-Q. Lu, IEEE Trans. Device Mater. Reliab. 12, 124–132 (2012). 32. Y. Mei, T. Wang, X. Cao, G. Chen, G.-Q. Lu, and X. Chen, J. Electron. Mater. 41, 3152–3160 (2012). 33. Y.H. Mei, G.-Q. Lu, X. Chen, S.-F. Luo, and D. Ibitayo, IEEE Trans. Device Mater. Reliab. 11, 316–322 (2011). 34. Y.H. Mei, G.-Q. Lu, X. Chen, S.-F. Luo, and D. Ibitayo, IEEE Trans. Device Mater. Reliab. 11, 312–315 (2011). 35. Y. Mei, G.-Q. Lu, X. Chen, C. Gang, S. Luo, and D. Ibitayo, J. Electron. Mater. 40, 2119–2125 (2011). 36. J.G. Bai, J.N. Calata, and G.-Q. Lu, IEEE Trans. Electron. Packag. Manuf. 30, 241–245 (2007). 37. http://www.nbetech.com/, 2010. 38. J.K. Mackenzie and R. Shuttleworth, Proc. Phys. Soc. 62, 833–852 (1949). 39. J.N. Calata, G.-Q. Lu, and T.-J. Chuang, Surf. Interface Anal. 31, 673–681 (2001). 40. V.A. Ivensen, Densification of Metal Powders during Sintering (New York/London: Consultants Bureau, 1973). 41. C. Buttay, A. Masson, J. Li, M. Johnson, M. Lazar, C. Raynaud, and H. Morel, High Temperature Electronics Network (HiTEN) (Oxford, UK, 2011). 42. E. Ide, S. Angata, A. Hirose, and K.F. Kobayashi, Acta Mater. 53, 2385–2393 (2005). 43. A. Sunwoo and R. Lum, Scr. Metall. Mater. 33, 639–644 (1995).