稀有金属(英文版) 2015,34(11),783-788

收稿日期：9 April 2013

Modification of Sn–1.0Ag–0.5Cu solder using nickel and boron

Jun-Feng Qu Jun Xu Qiang Hu Fu-Wen Zhang Shao-Ming Zhang

National Engineering Research Center for Nonferrous Metal Composites, General Research Institute for Nonferrous Metals

Technology Department, Beijing COMPO Advanced Technology Co. Ltd.

Abstract：

Effects of Ni and B additions on the microstructure and growth behavior of the intermetallic compound(IMC) of Sn–1.0Ag–0.5Cu alloys(SAC105) were investigated in this study. Results show that microadditions of Ni and B result in volume fraction of primary Sn increasing and the grain size decreasing observably. It is found that a large number of fine reinforcement particles with network-like shape are found in the solder, and the thickness of interfacial IMC layer in the solder joint is grew less than that of SAC105 with longer aging time. Shear test results reveal that as-soldered solder joints of microalloyed SAC105 have better shear strength than that of SAC105 solder alloy.

Keyword：

Lead-free solder; Sn–1.0Ag–0.5Cu; Interfacial reaction; B addition; Intermetallic compound;

Author： Jun-Feng Qu e-mail: armytwo@163.com;

Received： 9 April 2013

1 Introduction

Lead-free solder application in the electronic industry was developed to substitute traditional tin–lead solders [1]. The Sn–Ag–Cu (SAC) family of alloy is one of the most favorable systems because of its advantages in mechanical properties and its good solder ability [2]. However, there are still some drawbacks in these alloys, such as the intermetallic compound (IMC) formed at solder joints is more than that of tin–lead solder joints due to their higher reflow temperature and higher tin content [3, 4], which could increase the fragility at the solder joints.

Solder joints have to endure much more electrical and thermal load, which make the solder joints the weakest link in the architecture of the electronic packages. A thin, con- tinuous, and uniform IMC layer is an essential requirement for good bonding between solder and substrates [5]. The reliability of solder joints particularly depends on the mor- phology and thickness of IMC layers. The excessive growth of the IMC in the SAC solder system will be harmful to the long-term reliability of the solder joints [6]. Therefore, the prevention of excessive IMC growth in solder joint becomes an important task for materials researchers.

For further improvement of its performance, many efforts were made through the additions of other elements. Theoretical analyses showed that the addition of Ni into molten Sn can substantially decrease the Cu solubility in Sn [7] and boron benefits precipitation strengthening, solution strengthening, and grain boundary strengthening [8]. It was reported previously that B-doped low-carbon steel [9], Ni₃Al [10–13], Fe Al [14], in which boron migrated to the grain boundaries and strengthened them, simultaneously reduced the rate of grain growth. Our pre- vious work indicated that adding a trace amount of B into SAC solder can visibly improve the mechanical properties and the microstructure of the solder [15].

2 Experimental

2.1 Materials and processing

Four SAC solder alloys were prepared, and the solder alloy compositions are shown in Table 1. After weighing the inpidual pure metals (Sn, Ag, Cu), they were mixed and melted in resistance furnace under the covering agent (Li Cl:KCl = 1.3:1.0) at 400 °C. When the pure metals melted, the Ni–B master alloy was added into the alloy melt. The Ni–B master alloy was prepared in a vacuum medium frequency furnace at 1,200 °C. After mixing uni- form, the alloy melt was cast into block specimens.

The experimental substrates were oxygen-free copper plates with 40 mm 9 40 mm9 2 mm (Fig. 1a). They were polished with diamond powders and degreased in acetone so as to remove surface oxides and impurities. The lead- free solders were placed on the Cu plates to be welded at a temperature of (250 ± 1) °C for about 1 min. The com- mon rosin flux was used in the welding experiments.

2.2 Aging tests

Isothermal aging studies were conducted on solder joint samples (Fig. 1a). Thermal aging of solder joints was performed in heat-treatment furnace at (150 ± 1) °C for 100, 200, 300, and 400 h, then the specimens were air- cooled to room temperature.

For the observation of the interfacial microstructure, the aged joints were etched by using 95 % C₂H₅OH-5 % HNO₃after mechanical cutting and polishing. The mor- phology and composition of interfacial IMCs at the solder joints were observed using optical microscopy (OM) and scanning electron microscopy (SEM). The average thick- ness of the IMC layers in the solder joints were calculatedthrough piding the integrated area by the length of the IMC layers.

Table 1 Solder alloys and their composition (wt%)  下载原图

Table 1 Solder alloys and their composition (wt%)

2.3 Shear test

The shear tests were conducted on as-soldered solder joint samples, at room temperature and at shear strain rate of 5 9 10^-4s^-1. At least five samples were tested and the results were averaged. Fracture surface characterization studies were carried out using SEM. Figure 1 shows the schematic diagram of the solder joint sample.

3 Results and discussion

3.1 Microstructure and intermetallic morphology

The results of the present work indicate that the grain size of primary Sn decreases by the additions of Ni and Ni–B, and results in a higher volume fraction of primary Sn. As illus- trated in Fig. 2a, microstructure of unalloyed Sn–1.0Ag– 0.5Cu is coarse and less-packed. In the Ni-doped samples as shown in Fig. 2b, the primary Sn is finer and more spherical than unalloyed SAC105. Figure 2c and d displays that the area of eutectic regions reduces with Ni–B addition and with the increase of B content, the grains are further refined.

As shown in Fig. 3, with Ni–B addition, a mass of tiny precipitates distribute on the phase boundaries with net- work-like shape. From the binary phase diagrams for B–Ag, B–Cu and B–Sn, it can be seen that B is not dis- solvable in Ag, Cu, and Sn at low temperature. When soldering was carried out at around 250 °C, the B and Ni–B compound from Ni–B master alloy would be remained in a solid state. In this case, part of the B and boride particles might act as heterogeneous nucleation sites for primary Sn upon solidification; the other part of precipitates migrate to the phase boundaries and grain boundaries in the solder alloy, as shown in Fig. 3a and b, and might strengthen the boundaries and impede grain growth.

Fig. 1 Schematic diagrams of solder joint sample: a aging study and b shear test

Fig. 2 OM images of specimens: a Sn–1.0Ag–0.5Cu, b Sn–1.0Ag–0.5Cu–0.05Ni, c Sn–1.0Ag–0.5Cu–0.05Ni-0.01B, and d Sn–1.0Ag–0.5Cu– 0.05Ni–0.02B

Fig. 3 SEM images of Sn–1.0Ag–0.5Cu–0.05Ni–0.02B: a solder alloy and b solder joint

Fig. 4 SEM images of intermetallic in: a Sn–1.0Ag–0.5Cu, b Sn–1.0Ag–0.5Cu–0.05Ni, c Sn–1.0Ag–0.5Cu–0.05Ni–0.01B, and d Sn–1.0Ag– 0.5Cu–0.05Ni–0.02B

3.2 Microstructural evolution of IMCs in solder joints

Figure 4 shows the SEM images of interfacial IMCs in solder joints. It can be seen that after soldering, a contin- uous IMC layer forms between solder alloy and the Cu substrate. The thicknesses of IMCs in the interfacial reac- tion for the four soldering systems are 3.5, 2.9, 2.7, and 2.2 lm, respectively. It is remarkable that IMCs mor- phology of Sn–1.0Ag–0.5Cu–0.05Ni–0.02B/Cu is saw- tooth-like (Fig. 4d), and this structure is beneficial to the strengthening of the solder joints because the interface IMC layer combines the solder tightly.

Different aging times will affect the precipitation, dif- fusion, and combination of the interfacial IMCs, and then the interface morphology. Figure 5 shows the SEM images of Sn–1.0Ag–0.5Cu–0.05Ni–0.01B/Cu interfaces aged under various durations at (150 ± 1) °C. For 100 h of aging (Fig. 5a), the average thickness of (Cu_x,Ni_1-x)₆Sn₅layer for Sn–1.0Ag–0.5Cu–0.05Ni–0.01B/Cu solder sys- tem is 4.1 lm. For 200 h of aging (Fig. 5b), a bit more planar intermetallic layer is found, and the average thick- ness of (Cu_x, Ni_1-x)₆Sn₅layer is 4.7 lm. In the two stages above, the intermetallic growth at the valleys is more rapid than the peaks; as a result, intermetallic layer is trans- formed from the worm-like (Fig. 4c) to layer shape (Fig. 5a). With the holding time longer, the average thickness of intermetallic layer increases to 5.1 lm at 300 h (Fig. 5c) and 5.4 lm at 400 h (Fig. 5d), and the morphology of interface varies inconspicuously after the layer-type IMC formed.

The interfacial IMC growth results of the SAC105 and micro-alloyed SAC105 solder joints are presented in Fig. 6a. The results indicate that after subjected to iso- thermal aging for 100, 200, 300, and 400 h, the interfacial IMC layers of the SAC105 solder joints are observed to grow more significantly than that of the microalloyedSAC105 solder joints. The curves showing the average total thickness of the interfacial IMC layer with respect to different aging times are shown in Fig. 6b. In Fig. 6b, we can see that the growth of IMCs on the solder joint inter- face follows parabolic kinetics approximately.

Fig. 5 SEM images of Sn–1.0Ag–0.5Cu–0.05Ni–0.01B/Cu interfaces aged under various durations at 150 °C: a 100 h, b 200 h, c 300 h, and d 400 h

Fig. 6 Relationship between thickness of IMC and aging time: a four solder alloy systems under different aging conditions, b aging time, and c square root of aging time

Generally, the intermetallic layer growth behavior of the intermetallic compound for the solder joint interface can be described by means of one-dimensional growth parameter [16]:

where Y(t) is the thickness of the intermetallic layer at time t, Y₀is the initial thickness of intermetallic compound layer, A is the constant, n is the time exponent, and Q is activation energy.

Kim and Jung [17] reported that during the aging, the growth of intermetallic compounds generally follows linear or parabolic kinetics. The growth rate is controlled by the reaction rate at the growth site when the intermetallic growth follows the linear kinetics, whereas parabolic growth kinetics implies that the intermetallic growth is controlled by volume diffusion. For the diffusion-con- trolled mechanism, the growth of intermetallic compound layer thickness after aging should follow the square root of time power law relationship that can be expressed as:

where K is the intermetallic growth rate constant.

Fig. 7 Shear test results of solder joints in as-reflow condition

Figure 6c shows the thickness of the intermetallic layer as a function of square root of the aging time, and the slope of the curves gives the K values for the three alloy systems. There are noticeable differences between the three systems. It can be seen that: the growth rate of IMC layer at the Sn–1.0Ag–0.5Cu–0.05Ni/Cu,Sn–1.0Ag–0.5Cu–0.05Ni– 0.01B/Cu and Sn–1.0Ag–0.5Cu–0.05Ni–0.02B/Cu inter- face has little difference, but much lower than that of Sn–1.0Ag–0.5Cu/Cu interface; and the growth rates of IMC layer of the four solder alloy systems are 0.130, 0.136, 0.140, and 0.210, respectively.

Fig. 8 SEM images of fracture surfaces : a Sn–1.0Ag–0.5Cu, b Sn–1.0Ag–0.5Cu–0.05Ni, c Sn–1.0Ag–0.5Cu–0.05Ni–0.01B, and d Sn–1.0Ag– 0.5Cu–0.05Ni–0.02B

3.3 Shear test

Solders not only provide an electrical path between the components, but also subjected to mechanical loadings. Figure 7 shows the shear strength of the solder joint sam- ples in as-reflow condition which are subjected to different solder alloy systems, and the thickness of solder layers is about 800 lm. Shear test results on soldered joints show that microalloyed (Ni, Ni–B addition) solder alloy yields the better ultimate shear strength compared with Sn– 1.0Ag–0.5Cu sample.

The improvement in mechanical properties can be attributed to the following strengthening mechanisms: (1) the addition of grain refining elements (Ni, B) makes the microstructure finer, so as to improve the plasticity and fracture toughness of the alloy; (2) micron and submicron tiny precipitates distribute in the solder alloy (Fig. 3), and as the capacity fraction of the precipitate in matrix is fixed, the smaller the precipitate’s size is, the larger its amount is, and the smaller the average spacing is, so that the strengthening effect is more significant; (3) the network- like structure formed by the intermetallic compounds strengthens the phase and grain boundaries, so as to improve the mechanical properties.

Figure 8 shows the fracture surfaces of the four solder joints under condition of as-reflow at temperature of 250 °C. It should be mentioned that in as-reflow condition, all the joints exhibit the ductile shear fractures, and the fracture surface shows a well-developed shear dimples along the loading direction. Compared with the other two alloy systems, shear dimples in Fig. 8c and d become denser and thinner, and it is considered that B has signifi- cant effect on structure refinement. In essence, results of IMC layer growth and shear tests demonstrate that micro- alloyed Sn–1.0Ag–0.5Cu with Ni–B has desirable joint reliability.

4 Conclusion

Microstructure and growth behavior of the interfacial IMC layers for SAC alloy systems added Ni and B with Cusubstrate were investigated. The Microstructures are finer and interfacial IMC layers are thinner at the solder joints for micro-alloy SAC by Ni and B additions, which implies that the mechanical properties of the solder alloy could be improved by Ni and B additions. For long-time aging, the layers of interfacial IMC at solder joints do not become thick evidently for Sn–1.0Ag–0.5Cu–0.05Ni–0.01B, which indicates that Ni and B possess the function to hinder the growth of IMC layers. In as-soldered condition, all the joints fail with a ductile fracture mode and compared with Sn–1.0Ag–0.5Cu sample, solder alloys with Ni–B addition exhibit better shear strength.