Micro-scale Solder Research Articles

In-Line Test Structures and Non-destructive Characterization of Electromigration-Driven Phase Evolution in Microscale Solder Joints

In-Line Test Structures and Non-destructive Characterization of Electromigration-Driven Phase Evolution in Microscale Solder Joints

A study of micro-scale solder bump geometric shapes using minimizing energy approach for different solder materials

Demand for more interconnection joints between semiconductor devices can be realized with solder bump technology. Surface tension and density are usually material properties related factors that affect solder bump geometric shape. Therefore, to cope with this fast-changing microarchitecture design in semiconductor technology, a better understanding of the solder bump geometric shape is needed. This study used a static equilibrium force approach to integrate the surface tension and gravitational energy into the solder energy content. Surface Evolver software was used to perform calculations and deliver the final solder bump shape. Perfect agreement with less than 10 % comparison between previous studies and the current Surface Evolver results was found. According to statistical analysis using SPSS, the maximum width of solder shape is closely related to the surface tension. In contrast, the maximum standoff is highly correlated with the solder density. By changing the solder volume, the solder bump changes from standard flip-chip bump to Cu pillar bump with consistency in maximum width to maximum standoff height ratio of 1.5. This study shows that the bumping technology can produce various sizes of solder bumps to meet new electronic packaging requirements.

Effect of the anisotropic characteristics of β-Sn on current-induced solder evolution

The miniaturization trend of three-dimensional integrated circuits poses great challenges to solder reliability under high current stressing since anisotropic issues tend to be enhanced in microscale solder bumps. To address these challenges, the solder evolution at the microscale and the corresponding mechanism should be addressed. In this study, by considering the anisotropy of β-Sn in the electrical, thermal, mechanical and diffusional characteristics, the current-induced solder evolution of a typical solder bump was comprehensively explored over time. The results confirmed that the anisotropy of β-Sn aggravated the nonuniform distribution of current-induced stresses, further causing multiple migration modes coupled in the solder bump. The spatial inconsistency between the distribution of β-Sn unit cells and the bump geometry led to accumulation of torsion forces, which became the direct reason for solder rigid rotation. The electron-dislocation interaction promoted dislocation slip, leading to lattice rotation and homogenization of intragranular deformation. In addition, current-induced grain merging via detwinning was observed, and its velocity was faster than imagined. Our study is conducive to thoroughly understanding anisotropic evolution and establishing accurate damage mechanisms for micro solder bump design under current stressing.

A dual-phase crystal plasticity finite-element method for modeling the uniaxial deformation behaviors of thermally aged SAC305 solder

Because typical solder joints undergo thermal aging during use, models simulating the behaviors of solder under mechanical stress must accurately reflect the changes in properties caused by thermal aging effects. The uniaxial deformation of SAC305 (96.5Sn–3.0Ag–0.5Cu wt.%) solder under different aging conditions indicates that the mechanical properties of solder are weakened with aging. In this study, the nanoindentation method was used to evaluate the effect of aging on the mechanical properties of the β-Sn dendritic regions and eutectic regions of SAC305. Based on this evaluation, a dual-phase (DP) representative volume element (RVE) crystal plasticity finite-element method (CPFEM) model was developed to simulate the thermal aging effects of SAC305 solder. In situ tensile tests performed with scanning electron microscopy (SEM) showed that the tensile deformation was concentrated in the β-Sn dendritic regions of the solder. The DP CPFEM model was established based on the microstructures shown by electron backscatter diffraction (EBSD) analysis performed during the in situ tensile tests. This model was used to verify the accuracy of the DP CPFEM in simulating the deformation localization by comparison with the strain field distribution observed via SEM–digital image correlation (DIC) technology. The developed DP CPFEM model allows the prediction of the potential failure locations of a microscale solder joint.

Influence of Microstructure Evolution on the Current Density and Temperature Gradient of Line-Type Cu/Sn58Bi/Cu Microscale Solder Joints Under Current Stressing

Influence of Microstructure Evolution on the Current Density and Temperature Gradient of Line-Type Cu/Sn58Bi/Cu Microscale Solder Joints Under Current Stressing

Experiment and Numerical Simulation of Shear Creep of Cu/SAC305/Cu Microscale Solder Joints

Experiment and Numerical Simulation of Shear Creep of Cu/SAC305/Cu Microscale Solder Joints

Role of Nominal Stress State on Cyclic Fatigue Durability of SAC305 Grain-Scale Solder Joints

Abstract Solder joints in micro-electronic assemblies experience a multiaxial combination of extensional and shear loads due to combinations of thermal expansion mismatch and flexure of printed circuit assemblies during thermal cycling or during vibrational loading of constrained printed circuit assemblies. Although a significant amount of research has been conducted to study cyclic fatigue failures of solder joints under pure-shear loading, most of the current literature on cyclic tensile loading of solders is on long dog-boned monolithic solder coupons. Unfortunately, such specimens do not capture the critical interactions between key microscale morphological features (such as grain orientation, grain boundaries, intermetallic compounds, and substrates) that are believed to play important roles in the fatigue of functional solder joints under life-cycle loading. Therefore, this paper uses a combination of experiments and finite element analysis to investigate the differences in mechanisms of cyclic fatigue damage in Sn-3.0Ag-0.5Cu (SAC305) few-grained (oligocrystalline) microscale solder joints under shear, tensile and multiaxial loading modes at room temperature. Cyclic fatigue durability test results indicate that tensile loads are more detrimental compared to shear loads. Tensile versus shear loading modes are found to cause distinctly different combinations of interfacial damage versus internal damage in the bulk of the solder (transgranular and intergranular damage), which correlates with the differences observed in the resulting fatigue durability. The test results also confirm that the traditional approach of assuming a power-law dependence on equivalent deviatoric strain amplitude is inadequate for modeling cyclic fatigue durability of solder interconnects experiencing multiaxial loading. Instead, multiaxial fatigue damage results are seen to be affected not only by the cyclic equivalent strain amplitudes but also by the severity of the stress-triaxiality, as hypothesized in models such as Chaboche model.

Effects of Graphene Nanosheets on the Wettability and Mechanical Properties of Sn-0.3Ag-0.7Cu Lead-Free Solder

Low silver Sn-0.3Ag-0.7Cu (SAC0307) composite solder reinforced with graphene nanosheets (GNs) was prepared by powder metallurgy techniques, and the effects of GN content on the melting characteristics, wettability, microhardness and shear strength of the composite solder were investigated. The wettability and microhardness of the composite solder increased significantly, and the initial melting temperature decreased slightly with increasing doping content of GNs, the wettability and microhardness of the composite solder increased significantly, but the initial melting temperature decreased slightly. When the GN content reaches 0.01 wt.%, the wettability angle of the composite solder decreases by ∼ 29%. When the content of GNs reaches 0.09 wt.%, composite solder of microhardness is up to 15.3 HV, approximately 1.12 times larger than that of the un-doped SAC0307 solder. At the same time, it is also found that the shear strength of the composite solder increases with the increase of GN content. When reaching 0.05 wt.%, the shear strength is the highest (24.70 MPa) and then presents a downward trend. In addition, the fracture morphology of microscale solder joints exhibits a typical ductile fracture character to the brittle one with increasing GN content.

Experimental and numerical simulation of mechanical behavior of micro-scale SAC305 solder joint based on joint height

Lead-free “copper wire/Sn-3.0Ag-0.5Cu (SAC305) solder/copper wire” sandwich-structured micro-scale solder joints with a constant diameter (d = 400 μm) and various heights (125, 225, and 325 μm) were adopted to carried out the mechanical behavior research. Experimentally, quasi-static shear and tensile experiments were conducted by using a precision dynamic mechanics analyzer (DMA Q800). Then, the micro-scale solder joint model was built to simulate shear and tensile behavior of solder joints using the finite element analysis software ABAQUS. The experimental results show that, when the solder joint diameter is fixed, the smaller the joint height is, the stronger the shear strength and tensile strength. Under the same test conditions (225 μm), the shear strength of micro-scale solder joints of the same size is lower than the tensile strength, which indicates that the service environment of electronic packaging interconnection solder joints under shear stress is more severe. In addition, the micro-scale solder joints fracture at the interface of the solder joint and copper wire under shear force, but in the middle of the solder joint under tensile force. Different fracture failure mechanisms are found by observing the cross sections under different stress states.

Effects of Sn addition on the microstructure and properties of Bi–11Ag high-temperature solder

High-temperature lead-free Bi–11Ag solder with a Sn addition of 1, 3, and 5 wt% was investigated in this paper. The results show that the melting temperature of Bi–11Ag solders dropped from 265 to 255 °C after the addition of Sn, which is suitable for the high-temperature reflow process. When the addition of Sn was 3 wt% or less, dendrites in the microstructure would be refined, however, further additions of Sn show an opposite effect. The contact angle of Bi–11Ag–xSn (x = 0, 1, 3, 5 wt%) solders on the Cu substrate was measured, and the value decreased from 62° to 41° with the increasing of Sn content. The tensile strength of the Cu/Bi–11Ag–xSn solder/Cu sandwich structured microscale solder joint continues to increase from 43.3 to 53.9 MPa and the fracture position evolves from the solder/Cu interface to the middle of solder. In addition, according to EDS analysis of the solder/Cu interface, a Cu3Sn intermetallic compound (IMC) layer can be identified. The combination of this IMC layer and grain boundary grooving at the solder/Cu interface can effectively improve the strength of the Bi–11Ag–xSn solder joint.

Local shear stress-strain response of Sn-3.5Ag/Cu solder joint with high fraction of intermetallic compounds: Experimental analysis

Current trend in miniaturization of microelectronic devices is the motivating factor for size reduction of joints and interconnects. The electronic industry is expecting to reduce the scale of interconnects in the next generations microelectronic devices. Mechanical behavior of these joints will significantly be different from traditional solder joints due to the effects such as geometrical and microstructural constraints, anisotropy caused by the reduction in a number of grains, and the presence of brittle intermetallics (IMCs). Mechanical experiments at this scale are very challenging. This study was focused on investigating the effect of different volume fraction of IMCs on the shear behavior of micro-scale solder joints with a 50 μm stand-off height. A single lap-shear specimen was designed to conduct the study. Two Copper (Cu) substrates were soldered together with Sn-3.5Ag solder foil and a thickness of 50 μm was achieved. The soldering temperature of 260 °C and soldering time of 10, 30 and 60 min were utilized to achieve approximately 40%, 60% and 80% IMCs of total joint thickness. The experiments were conducted using a micro-tensile tester which is integrated with an optical microscope for monitoring and observing deformation in the testing materials. To investigate the local shear strain behavior, an optical technique along with a developed image processing computer program was used. The single lap-shear tests were conducted with the shear strain rate of 0.015s−1 and 0.15s−1 to observe its effect on shear stress-strain behavior. The study shows that the ultimate shear strength and strain to failure are largely dependent on the volume fraction of IMCs. On the other hand, the far-field or nominal shear strain is totally different than the local shear strain behavior at solder joint region. Fractography analysis was conducted to investigate the possible failure mechanism. A ductile to brittle fracture mode transition along with dominant interfacial delamination is witnessed when the solder joints had a higher volume fraction of IMCs than bulk solder. Effect of different shear strain rates on the shear strength, failure strain and fracture mode were also investigated. Different shear strain rates tend to have an insignificant effect on the shear strength of the joint whereas the failure strain decreases considerably at higher shear strain rate.

Thermomigration-induced asymmetrical precipitation of Ag3Sn plates in micro-scale Cu/Sn–3.5Ag/Cu interconnects

Understanding the microstructural changes induced by thermomigration during thermal bonding process is important to micro-scale solder interconnects. In the present work, the precipitation of Ag3Sn plates in 50μm Cu/Sn–3.5Ag/Cu interconnects under thermomigration has been in situ characterized using synchrotron radiation real-time imaging technology. The precipitation of Ag3Sn plates is found to be asymmetrical, i.e., Ag3Sn plates preferentially precipitate on the cold end rather than the hot end in the interconnects; even if precipitate on the hot end, the Ag3Sn plates tend to dissolve into the molten solder during the subsequent cooling. It is shown that the asymmetrical precipitation of Ag3Sn plates is induced by directional thermomigration of Ag atoms toward the cold end under the temperature gradient across the interconnects, and the simulated temperature gradient is 172°C/cm in the present work. The asymmetrical precipitation of Ag3Sn plates is analyzed and discussed from the viewpoint of thermomigration-affected nucleation and flux-controlled growth.

In situ fixture for multi-modal characterization during electromigration and thermal testing of wire-like microscale specimens

A fixture has been designed and fabricated to facilitate accelerated electromigration (EM) testing with in situ imaging capability of wire-like specimens. This fixture design has enabled microstructural evolution studies in microscale (100μm–500μm in dimension) solder volumes. The fixture allows application of thermal and electrical stimulus to wire-like solders in situ during multi-modal characterization. The design is compatible with surface (2D) characterization techniques including optical microscopy (OM) and scanning electron microscopy (SEM) based analyses. SEM characterization using the fixture has also been coupled with electron back-scatter diffraction (EBSD). Grain orientation image maps (OIMs) were obtained for in situ crystallographic microstructural analysis of the test volumes. The fixture is also compatible with X-ray computed tomography (non-destructive and 3D), or XCT, analysis of microstructure volume. To apply 2D techniques, a facet has been polished into the solder volumes prior to mounting within the fixture, though this surfacing requirement is not required for XCT characterization. A demonstration of the fixture's successful EM testing function is provided. The compatibility of the fixture with these characterization tools is also demonstrated. Both functional thrusts of the new in situ fixture's design have been accomplished through the fixture's form which is provided in reproducible detail.

Surface oxide analysis of lead‐free solder particles

PurposeThe composition and thickness of surface oxide of solder particles is extremely important to the quality of interconnect and reliability of packaged system. The purpose of this paper is to develop an observable measurement to research the issue.Design/methodology/approachAES (Auger electron spectroscopy), XPS (X‐ray photoelectron spectroscopy), TEM (transmission electron microscopy) and STEM (scanning transmission electron microscopy) were employed to examine the oxide layer on microscale solder powders. Conventional techniques and FIB (Focus Ion Beam) were employed for the TEM sample preparation. High angle annular dark field (HAADF) pattern was applied to distinguish the oxide layer and the solder matrix by the contrast of average atomic number. The results were confirmed by AES and XPS measurement.FindingsThe solder powders were exposed to air (70% relative humidity) at 150°C for 0, 120 and 240 h for the accelerated growth of oxide. The surface oxide thickness was 6 nm and 50 nm measured by TEM for 0 h and 120 h samples, respectively. It was found that the increase in surface oxide thickness of solder particles is proportional to the rooting of time. The elemental distribution along the oxide was quantified by line scanning using STEM and the atomic ratio of Sn to O in the oxide layer nearer to the outer, the middle, and the inner (adjacent to the solder matrix) was found to be 1:2, 2:3 and 1:1, respectively. The result was validated using XPS which gave Sn to O ratio of 1:2 at 5 nm depth of surface oxide.Originality/valueThis is the first time FIB technology has been used to prepare TEM specimens for solder particles and TEM pictures shown of their surface oxide layer. Though requiring more care in sample preparation, the measurements by TEM and STEM are believed to be more direct and precise.

Effects of Current Stressing and Isothermal Aging on the Tensile Strength of Microscale Lead-Free Solder Joints with Different Joint Volumes

Adopting an accurate micro-tensile method based on dynamic mechanical analyzer (DMA) instrument, the tensile strength of three kinds of copper-wire/solder/copper-wire sandwich structured microscale lead-free solder joints that underwent current stressing with a direct current density of 1.0×104 A/cm2 and loading time of 48 hours were investigated, and compared with those solder joints isothermal aged at 100 0C for 48 hours and as-reflowed condition. These three kinds of microscale columnar solder joints have different volumes, i.e., a same diameter of 300 μm but different heights of 100 μm, 200 μm and 300 μm. Experimental results show that both current stressing and isothermal aging degrades the tensile strength of microscale solder joints, and the solder joint with smaller volume obtains higher tensile strength under same test condition. In addition, current stressing induces obvious electromigration (EM) issue under high current density of 1.0×104 A/cm2, resulting in the decreasing of tensile strength and different fracture position, mode and surface morphology of microscale solder joints. The degree of strength degradation increases with the increasing of joint height when keep joint diameter constant, this is mainly due to that electromigration leads to voids form and grow at the interface of cathode, and solder joints with larger volume may contains more soldering defects as well.

The effect of joint size on the creep properties of microscale lead-free solder joints at elevated temperatures

Solder joints in electronic packaging systems are becoming smaller and smaller to meet the miniaturization requirements of electronic products and high density interconnect technology. Furthermore, many properties of the real solder joints at the microscale level are obviously different from that of bulk solder materials. Creep, as one of the key mechanical properties at elevated temperatures, can impair the reliability of miniature solder joints in electronic devices. However, there is a lack of knowledge about the comparative creep properties of microscale solder joints of different sizes. Most previous studies have focused on the creep properties of bulk solder materials or solder joints of the same size. In this research, to determine whether a size effect exists for creep properties of solder joints or not, we characterized the creep behaviors of Sn–3.0Ag–0.5Cu lead-free solder joints under tensile loading modes using microscale butt-joint specimens with a copper-wire/solder/copper-wire sandwich structure with two different sizes. Also, the creep failure mechanisms were investigated. Experimental results show that the creep activation energy and creep stress exponent are very similar for both sizes of solder joint. However, under the same testing conditions, the joints with a larger size exhibit a much higher steady-state creep rate and a shorter creep lifetime than the smaller joints.

Size and constraint effects on interfacial fracture behavior of microscale solder interconnects

The fracture behavior of solder joints has long been an important issue in the reliability evaluation of electronic components and packages. In this study, experimental and finite element methods were used to characterize the interfacial fracture behavior of “Cu-wire/solder/Cu-wire” sandwich-structured microscale Sn–3.0Ag–0.5Cu solder joints with different diameters (100–575μm) and thicknesses (35–325μm) under quasi-static micro-tension loading, with systematic comparison to Sn–37Pb solder joints. Dynamic stress intensity factors for the microcracks in Cu, Sn–3.0Ag–0.5Cu, and Sn–37Pb under pulse tensile loading and high speed propagation were calculated. The numerical simulation results showed that the interface stress intensity factors KII and KI for the crack set at the solder/IMC interface increased with increasing the thickness of the joints from 35μm to ∼125μm, and thereafter decreased and then became stable. With increasing the diameter of the joints, basically KI increases while KII decreases. Under the quasi-static micro-tension loading, the crack driving forces in Sn–3.0Ag–0.5Cu solder joints are lower than that in Sn–37Pb joints, and this means that fracture is less likely to occur in Sn–3.0Ag–0.5Cu solder joints than in Sn-joints. The rapid expansion of a high hydrostatic pressure region may enhance the mechanical performance of interconnects when the diameter-to-thickness ratios of the joints are very large. As the diameter of the solder joints increases, the evolution of the energy release distribution results in a change of the fracture position and mechanism in the interconnects. Compared to Sn–37Pb solder, the Sn–3.0Ag–0.5Cu lead-free solder shows a lower resistance to rapid crack propagation under constant tensile loading.

Undercooling Behavior and Intermetallic Compound Coalescence in Microscale Sn-3.0Ag-0.5Cu Solder Balls and Sn-3.0Ag-0.5Cu/Cu Joints

The microstructure of microscale solder interconnects and soldering defects have long been known to have a significant influence on the reliability of electronic packaging, and both are directly related to the solidification behavior of the undercooled solder. In this study, the undercooling behavior and solidification microstructural evolution of Sn-3.0Ag-0.5Cu solder balls with different diameters (0.76 mm, 0.50 mm, and 0.30 mm) and the joints formed by soldering these balls on Cu open pads of two diameters (0.48 mm and 0.32 mm) on a printed circuit board (PCB) substrate were characterized by differential scanning calorimetry (DSC) incorporated into the reflow process. Results show that the decrease in diameter of the solder balls leads to an obvious increase in the undercooling of the balls, while the undercooling of the solder joints shows a dependence on both the diameter of the solder balls and the diameter ratio of solder ball to Cu pad (i.e., Ds/Dp), and the diameter of the solder balls has a stronger influence on the undercooling of the joints than the dimension of the Cu pad. Coarse primary intermetallic compound (IMC) solidification phases were formed in the smaller solder balls and joints. The bulk Ag3Sn IMC is the primary solidification phase in the as-reflowed solder balls. Due to the interfacial reaction and dissolution of Cu atoms into the solder matrix, the primary Ag3Sn phase can be suppressed and the bulk Cu6Sn5 IMC is the only primary solidification phase in the as-reflowed solder joints.

Effect of surface oxide on the melting behavior of lead-free solder nanowires and nanorods

Lead-free nanosolders have shown promise in nanowire and nanoelectronics assembly. Among various important parameters, melting is the most fundamental property affecting the assembly process. Here we report that the melting behavior of tin and tin/silver nanowires and nanorods can be significantly affected by the surface oxide of nanosolders. By controlling the nanosolder reflow atmosphere using a flux, the surface oxide of the nanowires/nanorods can be effectively removed and complete nanosolder melting can be achieved. The complete melting of the nanosolders leads to the formation of nanoscale to microscale spherical solder balls, followed by Ostwald ripening phenomenon. The contact angle of the microscale solder balls formed on Si substrate was measured by direct electron microscopic imaging. These results provide new insights into micro- and nanoscale phase transition and liquid droplet coalescence from nanowires/nanorods to spheroids, and are relevant to nanoscale assembly and smaller ball grid array formation.

Isothermal low-cycle fatigue testing of microscale solder interconnections

A series of isothermal low-cycle fatigue studies of small to extremely small-volume solder joints has been conducted. These solder microjoints were designed and fabricated using processing which duplicates the microelectronics interconnection structures that might be used in high density, highly integrated flip-chip packaging and all fatigue tests were conducted in fully reversed simple shear both with and without dwells at maximum strain. Results of low-cycle fatigue tests of both single and double-bump 95/5 Pb/Sn solder microjoints in the form of Manson-Coffin (plastic strain amplitude versus fatigue life) plots and post-test failure mode analysis (FMA) carried out using scanning electron microscope (SEM) fractography are presented and evaluated.