Effect of Ni on growing of intermetallic compound in interface of Sn2·5Ag0·7Cu0·1RE/Cu solder joint during aging

Abstract

Aging treatment was implemented for the Sn2·5Ag0·7Cu0·1RExNi/Cu solder joint under various temperatures simulating the service environment. The growing behaviour of the intermetallic compound at the soldering interface during aging was investigated. In the case of adding 0·1Ni aged solder joint, the maximum growth activation energies of (Cu,Ni)₆Sn₅ and Cu₃Sn are 86·8 and 93·7 kJ mol⁻¹ respectively. The results reveal that the growth of intermetallic compound in the soldering interface can be suppressed, and the solder joint can be strengthened by adding proper quantities of Ni in the solder.

Keywords

Solder joint Interface Intermetallic compound Aging

Introduction

Pollution caused by Pb and Pb containing compounds in microelectronic devices has attracted much attention in academia and industry.1 Hence, it is necessary to develop environmentally friendly and highly reliable lead free solder.1,2 SnAgCu system lead free solder, which is considered as one of the most potential succedanea of SnPb solder, has been adopted now and is expected to be extensively used in the future due to its excellent comprehensive properties.2^–4 The preceding studies reveal that adding 0·1 rare earth (RE) in the solder can enhance the physical properties and creep strength of SnAgCu solder alloys.5 In order to further improve the wettability, tensile strength and elongation of SnAgCuRE lead free solder, minor Ni has been added into the solder during manufacturing.6^–8 As is well known, the formation of intermetallic compound (IMC) at the solder/substrate interface would affect the reliability of the solder joint. Especially, the sustained growth of IMC at the interface would degenerate the solder joint quality under service environment.9,10 Thus, understanding the growth characteristics of IMC at the interface of the solder joint during its service process is essential to restrain the growth of IMC so as to obtain a sound joint. However, few studies have been reported on the effect of Ni on IMC's growth characteristics at the solder/substrate interface during aging so far.

In the present study, to simulate the service environment, aging treatment was applied for the Sn2·5Ag0·7Cu0·1RExNi/Cu solder joint under different temperatures. The growing behaviour of IMC at the joining interface during aging was explored in order to provide some fundamental information for improving the reliability of the solder joint. Results indicate that adding appropriate Ni into Sn2·5Ag0·7Cu0·1RE solder can limit the growth of IMC at the interface of Sn2·5Ag0·7Cu0·1RExNi/Cu solder joint.

Experimental

The Sn2·5Ag0·7Cu0·1RExNi solder alloy was fabricated in a non-consumable melting furnace. Figure 1 shows the shape and dimension of solder joint; Fig. 1a shows the shape of the copper belt base metal. The flux was a water solution containing 22ZnCl₂ and 2NH₄Cl, with a soldering temperature of 270°C. The aging was achieved for the solder joint in the mediator of molybdenum disulphide to avoid the oxidation of the soldering seam surface, with temperatures of 80, 120 and 150°C. At each temperature, four aging times were considered, namely 50, 100, 200 and 500 h.

The solder joint was polished and etched with 4HNO₃+1HCl+95 alcohol solution for 3 s. The microstructure and composition analysis were evaluated using scanning electron microscopy (SEM), energy dispersive spectroscopy (EDS) and X-ray diffraction (XRD). The thickness of the IMC layer was measured using AutoCAD. In order to observe the top view morphology of IMC, most of the solder on the specimens was first ground away, and then the specimens were etched with a solution made up of 13HNO₃ and 87 alcohol under ultrasonic wave for 5 min to allow the remaining solder to dissolve. The average intercept length (L) of the IMC grain was measured according to the quantitative metallography theory11 to quantify the IMC grain dimension. The shear strength of the aged solder joint was tested under a crosshead velocity of 1 mm min⁻¹ at room temperature.

Results and discussion

Interfacial IMC microstructure of solder joint during aging

Figure 2 presents the SEM images of cross-sectional micrographs of Sn2·5Ag0·7Cu0·1RExNi/Cu interfaces aged under various durations at 150°C. It can be seen from the graph that when the solder joint was not aged (aging time was 0), a reaction layer was observed in its interface. The reaction layer had a relatively flat front in the Cu matrix side and scallop type morphology in the solder bulk side. Figure 2 illustrates that the thickness of the reaction layer varied with the changes in Ni content in the solder (to be mentioned). Judging from the results of EDS and XRD, the reaction layer was mainly made up of Cu₆Sn₅ phase. As the Sn content in the solder alloys is >95, the interfacial reaction product should also be Cu₆Sn₅ according to the Cu–Sn phase diagram.12 In addition, the XRD analysis results confirmed that no sulphide was detected at the solder joint. The results show that the diffusion of molybdenum disulphide to the joint does not occur during aging. In the case of the aged solder joint, two reaction layers were discovered in the interface. Layer 1, beside Cu matrix, displayed a laminar appearance, whereas layer 2, close to solder bulk with light contrast, had a scallop type appearance. Like the joints without aging treatment, the thickness of reaction layers also varied with content changes of Ni in the solder. Figure 3a shows the results of XRD analysis obtained from the cross-section of the aged Sn2·5Ag0·7Cu0·1RE/Cu solder joint (500 h). As is shown in the graph, reaction layers formed at the interface were mainly comprised of Cu₆Sn₅ and Cu₃Sn. Figure 3b and c shows the EDS results of layers 1 and 2 respectively. It can be seen that no Ni was detected in layer 1, and layer 2 only contains a small amount of Ni. Based on the XRD results, layer 1 was Cu₃Sn, and layer 2 was (Cu,Ni)₆Sn₅, a solid solution based upon Cu₆Sn₅.13 As mentioned above, no Cu₃Sn was detected at the interface of the solder joint without aging. It is likely that the driving force for Cu₆Sn₅ phase formation was higher than that of Cu₃Sn based on thermodynamical consideration;14 therefore, the Cu₆Sn₅ formed at the initial stage of soldering. However, limited diffusion of Sn towards the Cu substrate via the intermetallic layer would result in the formation of Cu₃Sn in the process of aging.

Reaction layers that formed at the interface of aged Sn2·5Ag0·7Cu0·1RExNi/Cu solder joint exhibited the following characteristics in comparison with that of the solder joint without aging. First, the thickness of reaction layers rose with the increase in aging time. Second, the boundary between (Cu,Ni)₆Sn₅ layer and soldering seam was also made of irregularities orientated towards the soldering seam; however, these irregularities were found to decrease with the increase in aging time. In other words, the (Cu,Ni)₆Sn₅ morphology changed from rough scallop type to smooth planar type. This can be attributed to two reasons. One is the IMC grain growth, which results in large grains merging small grains. The other is the solution of top IMC (away from the interface) into solder bulk. Otherwise, some cracks could be seen in the IMC, which may be caused by high stress concentration existing in the thicker hard brittle IMC. Moreover, After 500 h of aging, the layer of (Cu,Ni)₆Sn₅ formed at the interface of the solder joint where 0·1Ni was added was thin. As the growth of IMC is mainly determined by the diffusion mechanism during aging, it was likely that the addition of Ni would somehow reduce the driving force for the diffusion of Sn and Cu through the intermetallic layer,12,15 decrease the diffusion rate of Sn and Cu and consequently suppress the growth of (Cu,Ni)₆Sn₅. The (Cu,Ni)₆Sn₅ layer could be over thick under the condition that excessive Ni was added, but the reason for it stays unclear. In addition, adding trace Ni to solder alloy could also suppress the growth of Cu₃Sn during aging, and the inhibiting effect was especially obvious with a Ni concentration of 0·1. The formation of Cu₃Sn is affected by the phase stability of Cu₆Sn₅ according to the reaction Cu₆Sn₅+9Cu→5Cu₃Sn. The thermodynamic affinity in Cu₆Sn₅ is stronger between Ni and Sn than between Cu and Sn. Thus, the phase stability is greater for (Cu,Ni)₆Sn₅ than for Cu₆Sn₅.16 This finding suggests that the addition of Ni to the solder effectively suppresses the growth of Cu₃Sn compounds.

Figures 4 and 5 show the top view morphology of (Cu,Ni)₆Sn₅ in the solder joint unaged and aged for 500 h. From Fig. 4, cobblestone type (Cu,Ni)₆Sn₅ grains with size of 3–4 μm were observed in the unaged solder joint. However, the (Cu,Ni)₆Sn₅ grains changed to polyhedron type and grew in size when aged for 500 h. The morphology transformation violates the law of thermodynamics. Though the reason for it is unclear, the result is interesting. Figure 6 represents the relationship between average intercept length of (Cu,Ni)₆Sn₅ grain and aging time. After being aged for 500 h, both the average intercept length and the number of edges of (Cu,Ni)₆Sn₅ grain dropped to the bottom. Both the average intercept length and the number of edges of (Cu,Ni)₆Sn₅ grain increased if excessive Ni was added. It can be concluded that adding proper quantities of Ni can suppress the growth of (Cu,Ni)₆Sn₅ during aging.

Intermetallic compound growth kinetics

The growth of the IMC layer is a function of temperature and time (see equation (1)) (1) where X is the thickness of the IMC layer, X₀ is the initial thickness, K is the IMC growth rate constant, t is the aging time and n is the growth index.

Generally, the thickness of the IMC layer and the square root of aging time present a linear relationship, which can be seen from Figs. 7 and 8 as well. The gradient of the line is regarded as the IMC growth rate constant K. As shown in Figs. 7 and 8, the growth rate constant of (Cu,Ni)₆Sn₅ and Cu₃Sn reached the minimum at all temperatures when 0·1Ni was added. The growth rate constant of Cu₃Sn decreased but that of (Cu,Ni)₆Sn₅ increased when 0·5Ni was supplemented. Proper quantities of Ni in the solder can reduce the growth rate constant of (Cu,Ni)₆Sn₅ and Cu₃Sn.

The Arrhenius relationship concerning the IMC growth rate constant during aging can be expressed by equation (2) (see Ref. 17) (2) where K₀ is the diffusion constant, Q is the activation energy, R is the gas constant and T is the aging temperature in absolute unit.

Equation (3) can be deduced from equation (2) (3) If ln K was plotted against 1/T, the gradients of the two straight lines represented the growth activation energy of (Cu,Ni)₆Sn₅ and Cu₃Sn. Then, the effect of Ni content in the solder on the growth activation energy of (Cu,Ni)₆Sn₅ and Cu₃Sn is illustrated in Fig. 9. It can be seen that the maximum growth activation energies of (Cu,Ni)₆Sn₅ and Cu₃Sn were 86·8 and 93·7 kJ mol⁻¹ respectively when 0·1Ni was added to the solder joint. Compared with no Ni solder joint, the growth activation energy of Cu₃Sn ascended and that of (Cu,Ni)₆Sn₅ descended when 0·5Ni was added. In a word, an appropriate concentration of Ni can increase the growth activation energy of IMC.

Shear strength of aged solder joint

Figure 10 exhibits the influence of aging time on the shear strength of the solder joint at an aging temperature of 150°C. The curves could be divided into three stages. In the first stage, the shear strength decreased dramatically, and then in the second stage, the drop rate slowed; finally, the line flattened in the subsequent stage. The reason is that the growth rate of IMC differs at distinct aging times. Under the same aging time, the shear strength of the solder joint where 0·1Ni was added took the lead in comparison with that of the other two curves. Consequently, adding a certain amount of Ni during aging can increase the reliability of the solder joint, but the reinforcement effect could be weakened when excessive Ni is added.

Conclusions

The present study has examined the interfacial evolution and shear strength of Sn2·5Ag0·7Cu0·1RE xNi/Cu solder joints during aging. The main conclusions are listed as follows:

The thickness and morphology of the IMC layer formed at the soldering interface vary with the changes in aging time.

The growth of IMC in the soldering interface can be suppressed by adding proper quantities of Ni.

Adding certain amounts of Ni during aging can increase the reliability of the solder joint, but the reinforcement effect could be weakened when excessive Ni is added.

Footnotes

Acknowledgements

The present work was supported by the Foundation of Henan Province Outstanding Youth Scientist (grant no. 074100510011) and the Natural Science Foundation of China (grant no. 50774029).

References

Liu

, Yao

, Liu

: ‘Effects of multiple reflows on interfacial reaction and shear strength of SnAgCu and SnPb solder joints with different PCB surface finishes’, J. Alloys Compd , 2009, 470, 188.

Kobayashi

, Sasaki

: ‘Experiments and simulations of uniaxial ratchetting deformation of Sn–3Ag–0·5Cu and Sn–37Pb solder alloys’, J. Mater. Sci. Mater. Electron. , 2009, 20, 343–353.

Zhang

, Shi

, Guo

, Lei

, Xia

, Chen

, Tian

: ‘Comparison of drop performance between the Sn37Pb and the Sn3·8Ag0·7Cu solder joints subjected to drop test’, J. Mater. Sci. Mater. Electron. , 2011, 22, 292–298.

Teo

JWR

, Sun

: ‘Spalling behavior of interfacial intermetallic compounds in Pb-free solder joints subjected to temperature cycling loading’, Acta Mater. , 2008, 56, 242–249.

Zhang

, Wang

, Fan

, ”http://www.scientific.net/author/Jie_Yiang„ J. Yiang, Y. F. Yan, ”http://www.scientific.net/author/Xin_Zhang_1„ X. Zhang. Research on Reliability of SnAgCuRE Lead-Free Soldered Joints for Surface Mount Technology.’ Key Engineering Materials . 2007, 4, 2912–2915.

Zhang

, Wang

, Fan

, Zhao

, Yan

, Zhang

: ‘Sn–2·5Ag–0·7Cu–0·1Re–xNi lead-free solder alloy and its creep properties of solder joints’, Mater. Sci. Forum , 2010, 650, 91–96.

Kim

, Huh

, Suganuma

: ‘Effects of fourth alloying additive on microstructures and tensile properties of SnAgCu alloy and joint s with Cu’, Microelectron. Reliab. , 2003, 43, 259–267.

Wang

L. F.

, Sun

F. L.

, Shen , L.Y.X.: ‘The properties of Sn3·0Ag0·5CuxNi solder and solder joints’, Transactions of the China Welding Institution. , 2009, 30, 9–12.

Meng

, Takemoto

, Nishikawa

: ‘Correlations between IMC thickness and three factors in Sn–3Ag–0·5Cu alloy system’, Trans. Nonferrous Met. Soc. China , 2007, 17, 686–690.

10.

Liu

, Yao

, Liu

: ‘Evolutions of the interface and shear strength between SnAgCu–xNi solder and Cu substrate during isothermal aging at 150°C’, J. Alloys Compd , 2009, 486, 474–479.

11.

Ren

: ‘Metallographic experiment technology’; 1986, Beijing, Metallurgical Industry Press.

12.

, Vuorinen

, Kivilahti

: ‘Solder/substrate interfacial reactions in the Sn–Cu–Ni interconnection system’, J. Electron. Mater. , 2007, 36, (2), 136–146.

13.

Lin

, Shih

: ‘IMC formation on BGA package with Sn–Ag–Cu and Sn–Ag–Cu–Ni–Ge solder balls’, J. Alloys Compd , 2008, 452, 291–297.

14.

Prakash

, Sritharan

: ‘Interface reaction between copper and molten tin-lead solders’, Acta Mater. , 2007, 49, 2481–2489.

15.

Laurila

, Hurtig

, Vuorinen

, Kivilahti

: ‘A Effect of Ag, Fe, Au and Ni on the growth kinetics of Sn-Cu intermetallic compound layers’, Microelectron. Reliab. , 2009, 49, 242–247.

16.

Yoon

J.-W

, Noh

B.-I

, Kim

B.-K

, Shur

C.-C

, Jung

S.-B

: ‘Wettability and interfacial reactions of Sn–Ag–Cu/Cu and Sn–Ag–Ni/Cu solder joints’, J. Alloys Compd , 2009, 486, 142–147.

17.

Vianco

, Hlava

, Kilgo

: ‘Intermetallic compound layer formation between copper and hot-dipped 100In, 50In–50Sn, 100Sn, and 63Sn–Pb coatings’ J. Electron. Mater. , 1994, 23, 583–594.