1 Introduction
BGA packaging technology has been widely used in high I/O and surface mount devices. The solder joints of BGA packaged components often suffer shear damage due to thermal mismatch, external force of device assembly, etc. Good solder joint shear strength is an important guarantee for the high reliability of the device. The shear strength of BGA solder joints is related to the solder ball material and the pad metal. Traditional lead-tin solder is gradually being replaced by lead-free solder because it contains lead, which is seriously harmful to the human body and the environment [1, 2]. Sn-Ag-Cu is a type of lead-free solder commonly used at present [2]. Due to the differences in reflow soldering process and materials, the shear strength of lead-free BGA solder joints is different from that of lead-tin BGA solder joints. On the other hand, the shear strength of solder joints is inversely proportional to the thickness of the intermetallic compound (IMC) at the solder pad metal and solder ball welding interface. The Ni-P coating on the Ni-P substrate can effectively block the diffusion of Sn, reduce the thickness of the IMC layer, and improve the shear strength of the solder joints. The thickness of the IMC is also related to the thermal process experienced by the device. The test of the shear strength of the solder joint after different aging at a certain temperature can be used as a basis for evaluating the reliability of the solder joint. From a microscopic perspective, changes in the microstructure of the welding interface material directly affect the shear strength of the solder joint [3, 4]. Combining the results of macroscopic shear strength tests with the study of material microstructure changes has a reference value for a deeper understanding of solder joint reliability.
This paper improves the solder joint shear test method and studies the shear strength of Sn-3.5Ag-0.7Cu lead-free solder on Cu and Ni-P substrates. The effect of aging on the shear strength of Sn-3.5Ag-0.7Cu/Cu and Sn-3.5Ag-0.7Cu/Ni-P solder joints is discussed. Combined with the study of the shear strength of lead-free solder, the micromorphology of the solder joint fracture on different substrates was observed and analyzed using a scanning electron microscope (SEM), and the relationship between the micromorphology, structure and shear strength of the solder interface was studied.
2 Shear strength test method
2.1 Two test methods
Hwa-Teng Lee [5] soldered the two ends of two 15 mm long copper wires together with solder, fixed the copper wires with a special chuck, and tested the shear strength of the solder joint, as shown in Figure 1 (a). In Lee's method [5], due to the small size of the sample, the quality of the copper wire welding has a greater impact on the shear strength of the solder joint. In addition, the thickness of the solder joint is also a major factor affecting the shear strength test results. Figure 1 (b) is a photo of the deformation of the solder joint after the test. The deformation of the solder will cause the measurement results to deviate from the true value, resulting in the shear strength results measured by this method [5] being 20 to 60 MPa larger than the experimental results of this paper and the data reported in other literature [6]. Akio Hirose et al. [6] proposed another test method for solder shear strength. The sample was obtained by soldering two pure copper coaxial cylinders with diameters of 3 mm and 10 mm respectively. The copper cylinder with a larger diameter was fixed, and a shear force was applied to the other copper cylinder until the solder joint broke, thereby obtaining the shear strength of the solder joint. Since the sample size used in this method is larger than that of the previous method, it is easier to fix and the maximum shear force is increased, which reduces the accuracy requirements for the mechanical sensor. However, the solder thickness is still an important factor affecting the accuracy of the shear strength test. The data variance varies from 5 to 10 MPa[6].
2.2 Test method and sample preparation in this paper
By comparing the above two shear strength test methods, this paper proposes a method for directly testing the shear strength of solder joints on solder columns. The test sample and experimental device (Figure 2, Figure 3) can effectively reduce the deformation of the solder column welding end, making the measurement result of solder joint shear strength more accurate. The selection of sample size draws on the experience of literature [6], and the sample size is appropriately enlarged to have a larger welding area and a larger maximum shear force to reduce the impact of test errors on the measurement results. Figure 4 is a sample that has been shear tested. There is no obvious deformation near the welding end of the solder column. Experiments have shown that the measurement data obtained by this method are reliable, and the data convergence is good, with a variance of only 1.5 to 5 MPa.
The solder alloy used in the experiment is made by melting each high-purity solder component in a predetermined ratio under high vacuum and high temperature conditions. The copper substrate used in this experiment is a pure copper plate and a Ni-P substrate obtained by chemical plating on its surface, and the thickness of the Ni-P layer is 2μm. During the preparation of the sample, the solder is melted and cast into a mold so that it is welded to the surface of the substrate in a cylindrical shape. Subsequently, the samples were thermally aged at 150°C for 145, 300, 500 and 1000h respectively. Finally, the shear strength of the solder joint was measured by a CMT6104 tensile testing machine. Five samples were measured under each condition, and the average value was used as the test result. In order to investigate the influence of the IMC at the interface and the microstructure at the fracture on the reliability of the solder joint, a part of the tested samples was selected, the microstructure of the sample fracture was observed by SEM, and the main components of the IMC were determined by EDX.
3 Results and discussion
3.1 Shear strength test results
Figure 5 is a curve showing the shear strength of the solder joints of Sn-Ag-Cu solder on Cu and Ni-P substrates as a function of aging time. During the first 145 hours of aging, the shear strength decreased rapidly, and after 300 and 500 hours of aging, the shear strength of the solder joints increased slightly, after which the shear strength of the solder joints slowly decreased to the minimum value. After 1000 hours of thermal aging, the shear strength of the solder joints on the Ni-P substrate decreased more than that of the solder joints on the Cu board. The reason for this is related to the difference in the location of the shear fracture of the solder on the Ni-P substrate and the Cu substrate after long-term thermal aging.
3.2 Micro-area observation of solder joint fracture structure
3.2.1 Study of Sn-Ag-Cu/Cu solder joint
Figure 6 shows the typical microstructure of Sn-Ag-Cu solder on a Cu substrate, and it can be seen that the IMC at the interface is divided into two layers. EDX analysis shows that the layer of solder close to the Cu substrate is Cu3 Sn, which is thin, but will increase significantly with thermal aging. The layer tightly covered on Cu 3Sn is Cu6Sn5, which has an abrupt and uneven surface morphology, and the grains are mostly irregular hemispherical. Some literature also describes it as scallop-shaped, with a large thickness and slow growth with thermal aging.
In the solder joint shear test, according to the statistics of the fracture positions of multiple test samples, after the Sn-Ag-Cu/Cu solder joints were thermally aged, most of the fractures of the solder joints occurred in the solder. Figure 7 is a cross-sectional SEM photo of the Sn-Ag-Cu/Cu solder joints after shearing after 1000h thermal aging. In the figure, there is a thick layer of solder residue above the IMC layer. Due to the inhomogeneity of the Sn-Ag-Cu solder, the cracks are rugged in the solder.
3.2.2 Study on Sn-Ag-Cu/Ni solder joints Sn-Ag-Cu solder usually forms two phases of Ni3Sn4 or three phases of (Ni1-x,Cux)6Sn5 and (Ni1-x,Cux)3Sn4 at the interface with the Ni-P substrate. From the phase diagram analysis [7], it can be seen that the IMC generated at the interface when the welding is completed is (Ni1-x,Cux)6Sn5, and its crystal structure is similar to Cu6Sn5 [8]. (Ni1-x,Cux)3Sn4 is formed later. Only when the Cu in the solder in the interface area is consumed by (Ni 1-x,Cux)6Sn5 during the thermal aging process and its content is less than 0.6%, will (Ni 1-x,Cux)3Sn4 be generated at the interface. In this experiment, most of the samples have undergone thermal aging, so the IMC at the welding interface is (Ni1-x,Cux)3Sn4.
According to the growth process of Sn, Cu, and Ni three-phase IMC, the (Ni1-x, Cux)6Sn5 generated first after welding is on the top, and the (Ni1-x Cux)3Sn4 generated later is on the bottom, as shown in Figure 8.
Figure 9 is a top view of the Sn, Cu and Ni three-phase IMC. After corrosion, the (Ni1-x, Cu x)6Sn5 on the upper layer of the IMC is exposed first (Figure 9(a)). Compared with the Sn and Cu IMCs mentioned above, the morphology of this IMC compound is not as dense as that of the Sn and Cu IMCs. In addition, its grain shape is not scallop-shaped, but a short polygonal prism or pyramid, and the surface is abrupt and uneven, and there are certain gaps between the grains, indicating that the IMC composed of (Ni 1-x, Cux)6Sn5 is not arranged tightly. Figure 9 (b) is a top view of the (Ni1-x, Cux)3Sn4 IMC. It can be seen from the figure that the appearance of the (Ni1-x, Cux)3Sn4 grain is a polyhedron. This layer of IMC grows close to the welding interface, its surface is relatively flat, and the grains are closely arranged without obvious gaps.
Compared with the solder joints on the Cu substrate, the shear fracture surface of the solder joints on the Ni substrate is more in the IMC layer. The reason is that the loose and disorderly arrangement of the (Ni 1-x, Cux)6 Sn5IMC on the upper layer at the interface of the Ni substrate solder joint makes cracks easier to generate. Figure 10 is a cross-sectional view of the fracture surface of the Sn-Ag-Cu /NiP solder joint in the IMC layer after 300h of thermal aging. It can be seen that there is a thick IMC layer and some solder residue, and the fracture surface is relatively flat. From the analysis of multiple experimental samples and combined with the data of IMC thickness, it can be seen that when the thickness of IMC is less than 6μm, the shear fracture of the Sn-Ag-Cu/NiP solder joint mostly occurs in the solder; when the thickness of IMC is greater than 6μm, the shear fracture of the Sn-Ag-Cu/NiP solder joint mostly occurs at the interface between IMC and solder.
4 Conclusion
This paper improves the test method of solder shear strength, and uses this method to measure the shear strength of Sn-Ag-Cu lead-free solder joints on Cu and Ni-P substrates after different aging. The microstructure of the fracture surface is studied by SEM, and the following conclusions are drawn:
①The solder shear strength test method used in this paper can reduce the test error, and the device is simple and easy to operate;
② The shear strength test results show that after thermal aging, the shear strength of Sn-3.5Ag-0.7Cu/Cu and Sn-3.5Ag-0.7Cu/Ni-P solder joints decreased significantly in the first 145 hours, and then tended to be stable;
③The study of the microstructure of the solder joint fracture structure shows that the IMC formed by Sn-Ag-Cu solder at the Cu substrate interface is Cu6Sn 5 and Cu3Sn, and the intermetallic compounds formed at the Ni-P substrate interface are (Ni1-x, Cu x) 6Sn5 and (Ni1-x, Cu x) 3Sn4;
④Analysis of experimental results shows that for Sn-Ag-Cu/NiP solder joints, when the IMC thickness at the welding interface is less than 6μm, the shear fracture of the solder joint mostly occurs in the solder; when the IMC thickness is greater than 6μm, the shear fracture of the solder joint mostly occurs at the interface between the IMC and the solder.
提升卡
变色卡
千斤顶
京公网安备 11010802033920号
