Microstructure and Mechanical Properties of Sn 0 7Cu Flip Chip Solder Bumps Using Stencil Printing Method

Microstructure and Mechanical Properties of Sn–0.7Cu Flip Chip Solder Bumps

Using Stencil Printing Method

Dae-Gon Kim

_{and Seung-Boo Jung}

Department of Advanced Materials Engineering, Sungkyunkwan University, 300 Cheoncheon-dong, Jangan-gu, Suwon 440-746, South Korea

The interfacial microstructure of Sn–0.7Cu solder with ENIG (electroless Ni/immersion Au) was studied using scanning electron microscopy (SEM), electron probe micro analyzer (EPMA) and transmission electron microscopy (TEM). (Cu,Ni)6Sn5intermetallic compound

(IMC) layer was formed at the interface between the solder and Ni–P under bump metallurgy (UBM) upon reflow. The thickness of the (Cu,Ni)6Sn5IMC layer increased with isothermal aging time. Two distinctive layers, P-rich and Ni–Sn–P, were additionally found from TEM

observation. Analytical studies using energy dispersive spectrometer (EDS) equipped in TEM revealed that the composition of the P-rich layer is close to that of a mixture of Ni3P and Ni, while that of the Ni–Sn–P layer is analogous to the P-rich layer but containing a small amount of Sn in

it. The shear force of the joints decreased during isothermal aging. The decrease of the shear force should be mainly due to the microstructural degradation within the solder.

(Received May 20, 2005; Accepted August 30, 2005; Published November 15, 2005)

Keywords: flip chip, tin–0.7 copper, electroless nickel–phosphorous, interfacial reaction, shear test

1. Introduction

Flip chip technology has attracted a great deal of attention because it can accommodate a larger number of I/O interconnections, and at the same time it can deliver a uniform voltage distribution over the entire chip.1–5) Al-though the flip chip technology has the various advantages rather than conventional wire bonding technology, there is a reluctance to embrace the flip chip technology for integrated circuit packaging because of the high cost of the flip chip process. A mask-less chemical wafer bumping technology which combines electroless Ni plating and immersion Au process with stencil solder printing solved the high cost problem of the flip chip process, since the process bypasses the high cost process of under bump metallurgy (UBM) sputtering and photolithography using thick photoresist.2,6) Based on this low cost solder bumping technology, the flip chip technology has received much interest nowadays.

Sn–Pb solder alloys have been widely used in electronics as interconnection materials on account of a low melting temperature and good wetting properties on such substrates as Cu, Ni/Au and Pb.7–10)However, due to concerns about the environmental pollution of lead, lead-free solders are being popular. Currently the common lead-free solder alloys being investigated include Sn–Bi, Sn–Ag, Sn–Ag–Cu and Sn–Cu, all of which have a relatively high Sn content in comparison to the traditional Sn–Pb alloy varieties. Among various lead-free solders, Sn–Ag–Cu solders are regarded as the most promising compositions for replacement of Sn–Pb solders.11,12) _{They have superior mechanical properties of} strength, elongation, creep and fatigue resistance than Sn–Pb solders. However, even with these advantages, the Sn–Ag– Cu solders are relatively expansive because of Ag addition in the solder composition.

Sn–0.7Cu solder composition has been suggested in order

to overcome the high cost of the Sn–Ag and Sn–Ag–Cu solders. The cost of this solder is lower than the Sn–Ag and Sn–Ag–Cu solders. The Sn–0.7Cu solder alloy has the longest thermal fatigue life among all the lead-free solder/ UBM interconnect structures, but melting temperature of the Sn–0.7Cu solder is slightly higher than those of the Sn–Ag and Sn–Ag–Cu solder alloys.9) In 2000, the national electronics manufacturing initiative (NEMI) recommended to replace eutectic Sn–Pb alloy by eutectic Sn–Ag–Cu alloy in reflow processing and eutectic Sn–0.7Cu alloy in wave soldering.13) _{However, the metallurgical behavior of Sn–} 0.7Cu solder joints with electroless Ni–P and suitability of the Sn–0.7Cu for flip chip applications have not been sufficiently studied yet. A detailed study to correlate the microstructures and mechanical properties of a solder joint with the Sn–0.7Cu alloy as a function of aging time or reflow number is highly needed.

Therefore, the present study was carried out to investigate the interfacial reaction kinetics with electroless Ni–P metallization during aging in various temperatures for Sn– 0.7Cu solder. Scanning electron microscopy (SEM), electron probe micro analysis (EPMA) and transmission electron microscopy (TEM) were employed to investigate the micro-structure and phase analysis of the solder joints. The shear test of a solder bump was performed to investigate the effect of microstructural variation of the solder joint on the joint strength and failure mode of the test specimens.

2. Experimental Procedures

The solder composition used in this study was a Pb-free Sn–0.7Cu (in mass%). Silicon wafers of 4 inches in diameter were sputtered with Ti of 0.2mm-thickness and Cu of 0.8mm -thickness. The Ti and Cu layers are an adhesion layer and interconnection pad layer, respectively. The metallized wafers had area array pad arrangements at 500mm pitch with rectangular pad opening of 130mm in width. The bumping for the flip chip devices was performed using an

*1_{Graduate Student, Sungkyunkwan University}

*2_{Corresponding author, E-mail: sbjung@skku.ac.kr}

immersion Au/electroless Ni–P in thickness of 0.15 and 6mm, respectively. The Ni–P serves as a diffusion barrier between the Cu pad and solder, while Au layer is required to enhance the wettability of the solder to the metallization and protect the Ni–P layer from oxidation before solder applica-tion. In the Ni–P layer, the concentration of P is about 15 at%. The lateral overlappment of the Au/Ni–P UBM on the chip passivation layer, SiO2, was about 5mm. Solder paste was applied by stencil printing method and subsequent reflow employing an IR four zone reflow machine (RF-430-N2, Japan Pulse Laboratory Ltd. Co.) with a maximum temper-ature of 530 K for 60 s. After reflow, the average solder bump height and diameter were about 145 and 190mm, respective-ly. Then the wafers were cleaned to remove flux residues. The reflowed specimens were isothermally aged at 353, 373, 393 and 423 K for 72, 144, 720, 1440 and 2400 h in air furnace to observe the interfacial reactions of a solid state solder. After aging, the specimens were mounted in epoxy, and then ground with SiC papers followed by subsequent polishing with 1 and 0.3mmalumina powders. Nital etching solution was employed for the purpose of metallographic observation. The microstructural observation was conducted with SEM, and the resulting IMCs were measured by energy dispersive spectrometer (EDS). Especially, for more detailed analysis of the interfacial reaction, cross-sectional TEM observation was carried out with analytical studies using EDS (Noran) equipped in TEM (HF-2000 operated at 200 kV, Hitach Co.). TEM samples were prepared by ultra-microtomy with a Leica Untracut T Ultramicrotome by (i) cutting a cross-sectioned cubic sample followed by molding with commercial epoxy, (ii) setting the molded sample in ultramicrotomy arm and trimming the cross-section surface with diamond knife, and (iii) slicing the trimmed sample by 15 nm in sectioning thickness using a boat-type diamond knife.14)Shear test of a solder joint was conducted using a global bond tester (PTR-1000, Rhesca Co., Ltd.) with

increasing aging time at various temperatures to examine the effect of interfacial microstructural change on the mechanical joint strength. Each test site was examined using SEM after shear testing to evaluate the mode of failure.

3. Results and Discussion

Cross-sectional studies were conducted to investigate the interfacial reactions between the Sn–0.7Cu solder and immersion Au/electroless Ni–P/Cu UBM. Figure 1 shows the cross-sectional SEM images of the solder joints before and after isothermal aging at 423 K for 72, 144, 720, 1440 and 2400 h. At the as-reflowed interface between the solder and electroless Ni–P layer, a layer of (Cu,Ni)6Sn5 IMC was formed in thickness of about 1.1mm. The composition of this phase, determined by wave length dispersive spectroscopy (WDS) equipped in EPMA, was (Cu0:64Ni0:36)6Sn5. It is known that the thin Au coating is dissolved and consumed away quickly into the Sn-based solder bulk during soldering. The (Cu,Ni)6Sn5 IMC is based on the Cu6Sn5 structure but with a certain amount of Ni dissolved in it, thus, the shape and properties of the IMC are very analogous to those of Cu6Sn5. One of the critical factors that greatly affects solder joint reliability is the formation of Cu6Sn5 and Cu3Sn IMC that evolves at the solder/substrate. Basically, the presence of a thin interfacial Cu–Sn IMC layer is essential for the bondability of the metallized substrate because the IMC layer promotes good wetting between the solder and pad.9) However, too thick IMC layer may weaken the joint due to its brittle nature and lattice mismatch with that of the substrate and solder. Therefore, from the joint reliability point of view, it is very important to study the Cu–Sn IMC layer growth kinetics when high temperature storage or extended bake-out creates thick IMC layers.15,16)As can be seen in the Fig. 1, thickness of the (Cu,Ni)6Sn5 IMC layer increased with aging time, and finally reached about 3.2mm

b)

d)

f)

Cu

Sn

Electroless Ni

(Cu, Ni)

Sn

a)

_Cu

6

Sn

(Cu, Ni)

Sn

e)

Electroless Ni

(Cu, Ni)

Sn

c)

Electroless Ni

(Cu, Ni)

Sn

Cu

Sn

after the aging time of 2400 h. Some particles of Cu6Sn5 which should be reaction products between Sn and Cu atoms within the bulk solder were found to distribute at the bulk solder region.

Cross-sectional image of the solder joint aged at 423 K for 1440 h and the corresponding EPMA line profiles are shown in Fig. 2. It could be seen that the reaction layer could be clearly divided into two regions which were the (Cu,Ni)6Sn5 and a thin dark layer in the SEM image. EPMA line analyses also confirmed the coexistence of (Cu,Ni)6Sn5 and a thin layer in the interface between the solder and electroless Ni–P layer. As can be seen in the EPMA line profiles, the thin dark layer has relatively higher content of P, so from now, we can name it a P-rich layer. According to previous studies, the thin layer is composed of two layers which are the P-rich layer and extremely thin Ni–Sn–P layer.14,17–19)However, because of the resolution limit of the SEM and EPMA, those layers were not visible in the figure.

For more detailed analysis of the thin layer, cross-sectional TEM micrographs were also taken. Figure 3 shows the cross-sectional TEM micrograph of the soldered specimens aged at 423 K for 720 h. The interface between the solder and electroless Ni–P layer was composed of three distinct phases which were the (Cu,Ni)6Sn5, Ni3SnP and P-rich layer. The lower two phases are planar and upper one phase has a relatively rough morphology. Table 1 summarizes the chem-ical compositions of the electroless Ni–P and reaction products by analytical studies employing EDS equipped in TEM. It is noted that the composition of the P-rich layer is not stoichiometry of pure Ni3P compound which was reported elsewhere. This was already shown in our previous study concerning the interfacial reaction of Sn–Ag–Cu/Ni–P

combination.20)Huang and Suganuma also reported that the P-rich layer was consisted of Ni3P and Ni in the Sn–3.5Ag and Ni–P UBM system with their selected area diffraction (SAD) pattern studies.14)From our TEM-EDS analyses, the averaged composition of the layer is more close to that of mixture of the Ni3P and Ni than that of pure Ni3P compound.

Sn

Ni

Cu

P

Ni-P

_{(Cu, Ni)}

6

Sn

P rich layer

Si wafer

(Cu, Ni)

Sn

P rich layer

Fig. 2 Cross-sectional SEM image of a specimen aged at 423 K for 1440 h (a) and corresponding EPMA line profiles (b).

100 nm P-rich layer

Ni-Sn-P layer (Cu, Ni)₆Sn₅

Fig. 3 TEM cross-sectional view of a specimen aged at 423 K for 720 h.

Table 1 Chemical compositions of electroless Ni–P and reaction com-pounds (in at%) analyzed with TEM-EDS.

Sn Ni Cu P

(Cu,Ni)6Sn5 43.21 26.46 30.33 —

The crystallization of electroless Ni–P layer into Ni3P and Ni was induced by the solder reaction with Ni.3) _When molten solder comes in contact with the original electroless Ni–P layer, (Cu,Ni)6Sn5 IMC is formed which induces crystallization of the adjacent layer of the original amorphous electroless Ni–P deposit. Due to consumption of Ni to form Cu–Ni–Sn IMCs, P is expelled to the remaining Ni–P layer and forms a P-rich layer. Because of the high P content, the Ni3P is initially crystallized, and then Ni is crystallized beside the nanocrystalline Ni3P where the P content is lower. Therefore, we can consider the P-rich layer as a mixture of Ni3P and Ni. Another layer of about 100 nm in thickness could be also found between (Cu,Ni)6Sn5 reaction product and P-rich layer in Fig. 3. This layer was identified as a ternary compound, which is consisted of Ni, Sn and P. This layer must result from interaction between the P-rich layer and liquid Sn atoms. In our analytical studies, the compo-sition of the layer is similar to that of the P-rich layer but containing a small amount of Sn, and this is supporting the above assumption. This layer was also indexed to be Ni3SnP by Huang and Suganuma.14)Considering all these data, this ternary phase is believed to be Ni3SnP compound. It is noted that this layer was also found in the study about the interfacial reaction between the Sn–Ag–Cu/electroless Ni–P.20) This means that the Ag element addition to the Sn–Cu solder has not significant influence on the interfacial reactions between solder and Ni–P layer.

Figure 4 shows the variations of the IMC layer thickness as a function of square root of aging time for different aging temperatures. In this study, the growth of only (Cu,Ni)6Sn5 IMC was investigated because the Ni–Sn–P layer is too thin to be examined with SEM. As shown in the Fig. 4, the thickness of the IMC layer increased with increasing aging time for all temperatures while a faster growth of the IMC occurred at higher temperatures. This behavior is very general, and previously reported by others numer-ously.17,20–22)

In Fig. 4, it can be also seen that the growth of intermetallic compounds followed a parabolic law, implying that the growth of the intermetallic layer is diffusion

controlled. The growth rate constant was calculated from a linear regression analysis ofdvst1=2_{, where the slope =k.}

The following Arrhenius relationship was used to deter-mine the activation energy for the (Cu,Ni)6Sn5IMC:

k2¼k2₀exp Q

RT

ð1Þ

wherek2₀ is the constant,Qis the activation energy,Ris the gas constant (8.314 J/molK) and T is the aging temperature (in absolute unit). The activation energy was calculated from the slope of the Arrhenius plot using a linear regression analysis. The activation energy for the (Cu,Ni)6Sn5 layer growth is obtained to be 83 kJ/mol.

To investigate the effect of the increase in thickness of the (Cu,Ni)6Sn5 IMC layer on the mechanical properties of the solder joints, shear test of a solder bump was conducted with various aging conditions.

Figure 5 shows the variation of the shear force with aging time at 353, 373, 393 and 423 K. The shear force was significantly decreased at the first stage of the isothermal aging and then did not change or slightly decreased. The sudden decrease of the shear force at the first stage and the slight decrease in later stage of the aging period should be mainly due to the coarsening effect of the microstructure within the solder. Furthermore, the higher shear strength value of the sample aged at a higher aging temperature is surprising. Although the exact reason for this is not clear at this point, it is presumably due to the discrepancy of Cu solubility in the Sn matrix at different aging temperatures. To verify the variation of the shear force, fracture surfaces of the shear tested specimens were examined using SEM. As shown in Fig. 6, all of the test specimens showed a ductile failure mode consisting of mere solder at the fracture surface. This implies that the IMC thickness of about 3mmis not affecting the mode of failure, and the shear force variation is not much related to the thickness of the IMC layers at the current aging conditions. Generally, in the solder bump-shear test, a fracture occurs at the interface or in the solder region with the lowest strength. It was known that excessively thick reaction layers formed between solder and substrate could significantly degrade the mechanical properties of the solder

0 500 1000 1500 2000 2500 3000

1 2 3 4 (Cu, Ni) 6 Sn 5 IMC thickness, d /m µ

Aging time, t 0.5 / s 0.5

353 K 373 K 393 K 423 K

Fig. 4 Variations of (Cu,Ni)6Sn5 IMC layer thickness as a function of

square root of aging time.

0 500 1000 1500 2000 2500 3000

1200 1400 1600 1800 Shear str ength, F / mN

Aging time, t 0.5 / s 0.5

353 K 373 K 393 K 423 K

joints.23)However, in this study, fracture analysis indicated that the shear strength could not be related to the thickness of the IMC layer formed at the interface between Sn–0.7Cu solder and immersion Au/electroless Ni–P/Cu UBM. The results of mechanical test showed that the Sn–0.7Cu flip chip solder joint has the desirable joint reliability.

4. Conclusion

This study detailed the microstructural evolution of the IMC layer formed between the Sn–0.7Cu solder and their joint strength variation during isothermal aging. The cross-sectional interfaces were examined using SEM, EPMA and TEM. The results were used to make following conclusions. (1) In the initial reaction, the reaction product was (Cu,Ni)6Sn5, while the Au layer appears to have dissolved or reacted with the solder leaving no observable Au at the interface. In a more detailed analysis of the reaction layer using TEM, a layer of a Ni–Sn–P compound could be observed. The chemical composition of the Ni–Sn–P compound is close to Ni3SnP which was reported in previous studies. (2) A P-rich layer could be seen between the Ni–Sn–P

compound layer and the electroless Ni–P layer in the TEM micrographs. We found that the composition of the P-rich layer is more close to that of a mixture of Ni3P and Ni.

(3) The thickness of the (Cu,Ni)6Sn5 IMC layer increased with increasing aging time and temperature. There was a linear relationship between the growth of the intermetallic layer thickness and the square root of the

aging time.

(4) The shear force of the joints decreased during iso-thermal aging. The decrease in shear force should be mainly due to the microstructural coarsening within the solder, while the effect of the increase of the IMC thickness is not of much importance. Only ductile failure mode was detected after the shear test of the solder bumps.

Acknowledgements

This work was supported by grant No. RTI04-03-04 from the Regional Technology Innovation Program of the Ministry of Commerce, Industry and Energy (MOCIE).

REFERENCES

1) J. H. Lau:Flip Chip Technologies, (McGraw-Hill, New York, 1995) pp. 25–59.

2) J. H. Lau:Low Cost Flip Chip Technologies for DCA, WLCSP, and PBGA Assemblies, (McGraw-Hill, New York, 2000) pp. 1–17. 3) J. W. Jang, P. G. Kim, K. N. Tu, D. R. Frear and P. Thompson: J. Appl.

Phys.85(1999) 8456–8463.

4) T. Y. Lee, K. N. Tu and D. R. Frear: J. Appl. Phys.90(2001) 4502– 4508.

5) N. Duan, J. Scheer, J. Bielen and M. Van Kleef: Microelectron. Reliab.

43(2003) 1317–1327.

6) J. Kloeser, P. Coskina, R. Aschenbrenner and H. Reichl: Microelectron. Reliab.42(2002) 391–398.

7) J. W. Nah, K. W. Paik, J. O. Suh and K. N. Tu: J. Appl. Phys.94(2003) 7560–7566.

8) K. N. Tu, T. Y. Lee, J. W. Jang, L. Li, D. R. Frear, K. Zeng and J. K. Kivilahti: J. Appl. Phys.89(2001) 4843–4849.

a)

Shear dir

ection

b)

Shear direction

c)

Shear dir

ection

d)

Shear direction

10) D. Li, C. Liu and P. P. Conway: Mater. Sci. Eng. A391(2005) 95–103. 11) K. Zeng, V. Vuorinen and J. K. Kivilahti: IEEE T. Electron. Pack.25

(2002) 162–167.

12) A. Sharif, M. N. Islam and Y. C. Chan: Mater. Sci. Eng. B113(2004) 184–189.

13) J. Lau, W. Dauksher, J. Smetana, R. Horsley, D. Shangguan, T. Castello, I. Menis, D. Love and B. Sullivan: Solder. Surf. Mt. Tech.16

(2004) 12–26.

14) C. W. Hwang and K. Suganuma: J. Mater. Res.18(2003) 2540–2543. 15) J. W. Kim and S. B. Jung: Mater. Sci. Eng. A397(2005) 145–152. 16) J. W. Kim, D. G. Kim and S. B. Jung: Met. Mater-Int.11(2005) 121–

17) D. G. Kim, J. W. Kim, J. G. Lee, H. Mori, D. J. Quesnel and S. B. Jung: J. Alloy. Compd.395(2005) 80–87.

18) M. He, Z. Chen and G. Qi: Acta Mater.52(2004) 2047–2056. 19) M. He, W. H. Lau, G. Qi and Z. Chen: Thin Solid Films462–463

(2004) 376–383.

20) D. G. Kim and S. B. Jung: Mater. Trans.46(2005) 1295–1300. 21) A. Sharif and Y. C. Chan: J. Alloy. Compd. 393 (2005) 135–140. 22) W. C. Luo, C. E. Ho, J. Y. Tsai, Y. L. Lin and C. R. Kao: Mater. Sci.

Eng. A396(2005) 385–391.