Solder Interface Research Articles

Solder Strength Characterization Using Hot Bump Pull Testing

Hot and cold solder ball pull testing has emerged as an attractive alternative to the traditional ball shear method for characterizing the attachment strength of solder interconnection. This paper describes the effects of aging with relative humidity and temperature on solder strength using hot bump pull test method. The results show high dependence of solder interface strength due to aging.

Mechanism of the Delayed Growth of Intermetallic Compound at the Interface between Sn-4.0Ag-0.5Cu and Cu-Zn Substrate

A Cu−Zn wetting layer was very effective to depress the excessive growth of intermetallic compound (IMC). The effect of Zn addition to the Cu layer on the IMC growth and microvoid formation in the solder interface was similar to the effect of Zn addition into the Sn-Ag-Cu (SAC) solders. In this study, the mechanism of slow IMC growth at the SAC/Cu−Zn interfaces was investigated. As the aging time increased, Zn atoms accumulated at the Cu6Sn5/Cu interface and formed a Zn-rich layer. By adding Zn into the Cu wetting layer, the IMC growth was delayed due to the retardation of the formation of the Cu3Sn layer. Since the low driving force for the formation of Cu3Sn became smaller by adding Zn, the diffusion of Cu in Cu−Zn into SAC solder was delayed. Also, the CuZn phase formed at the Cu6Sn5/Cu−Zn interface plays a role as a diffusion barrier of interdiffusion of Cu and Sn.

Ultra Heat-Shock Resistant Die Attachment for Silicon Carbide With Pure Zinc

An ultra heat-shock resistant die-attach structure for a silicon carbide power device was developed. A silicon carbide die with a gold/titanium nitride coating was soldered with pure zinc onto a DBC with a silicon nitride insulator plate. This die-attach structure could resist severe thermal shock in air temperatures between -50°C and 300°C. No substantial change in microstructure was observed at the soldering interface, while conventional high-lead solder forms severe cracking. This die-attach structure can be also applied to other types of high-temperature device applications.

First-principles Investigation of Bi Segregation at the Solder Interface of Cu/Cu 3Sn(010)

in Cu3Sn slab was found to be the most energy-favored at an adhesion energy of 1.96 J/m 2 . Based on this construction, flve possible segregation sites were examined, and the most likely segregation site was determined with adhesion energy as low as 1.06 J/m 2 , which was almost half of the initial one. Comparing with other sites 0 adhesion energies, it was concluded that size efiect took a large part in embrittlement. The analyses of atomic structure and electronic density revealed that the slabs shifted away from interfaces due to bismuth segregation, and the atoms around Bi atom were pressed away. This calculated work agreed qualitatively with reported experimental results.

Electromigration in WLCSP solder bumps

Electromigration in two different bumping technologies for WLCSPs has been investigated. The UBM and the structure of the pads on the PCBs were found to have a major influence on the failure mode. A Ni metallization layer serving as a diffusion barrier drastically reduces the intermetallic growth in the solder ball compared to samples assembled on Cu-OSP pads. In the former the principal failure mode is electrical open due to voiding at the Ni–solder interface. In the latter Cu diffuses into the solder at a high rate creating Cu–Sn intermetallics and eventually leading to an electrical open in the Cu track on the PCB being the principal failure mode. Additional experiments showed that the lifetime of the Cu-OSP mounted samples can be extended by using via-in-pad structures.

Modeling the failure of intermetallic/solder interfaces

Solder joint integrity is recognized as a key issue in the reliability of flip chip and ball grid arrays in integrated circuit packages. Significant reductions in the solder joint interconnect size results in both the increased volume fraction of brittle intermetallics in the joint and joule heating due to high current density. Previously, the prevalent failure mode in a solder joint was the ductile thermo-mechanical fracture of solder material due to repeated thermal cycling. In addition to this mode of failure the joints were also found to fail by brittle fracture near the solder–intermetallic interface. To predict and reduce the failure of solder joints, a model that incorporates both plastic damage in bulk solder and solder–intermetallic interface failure is timely and useful to the electronics industry. Based on cohesive fracture theory, a 3D finite element model has been developed to predict the interfacial damage of Sn–Ag–Cu eutectic lead-free solder interconnects. Unified creep-plasticity theory is incorporated in the model considering the creep and hysteresis effects in solder bulk. Using the unified creep-plasticity-cohesive finite element model, the intermetallic layer growth effect has been researched for different solders. The failure of solder is a complex, coupled thermo-electric-mechanical problem. To understand better the phenomenon, a thermo-electric finite element analysis has been conducted to predict the electrical concentration and joule heating effects on the failure of Sn–2.5Ag–0.8Cu–0.5Sb solder under different applied current densities. The temperature and current density distribution in a solder joint with a crack that propagates near the interface of the bulk solder and intermetallic layer has been predicted. Pronounced temperature and electrical current concentration are observed near the crack tip. Although the Pb-free solder is usually regarded as having a higher melting temperature compared with traditional Pb–Sn eutectic solder, the concentration of heat at the crack tip caused by joule heating may still melt the solder material under high current density stressing.

Electromigration Performance of Pb-Free μPILR™ Flip-Chip Packages

The μPILR interconnect is a copper pillar manufactured as a part of a substrate pad. In this paper, we discuss the electromigration (EM) performance of Pb-free μPILR interconnects in a multi-pair daisy chain within 150μm pitch flip-chip packages. Electromigration performance of μPILR interconnects has shown a significant improvement and noticeably delayed electromigration induced failures. Voids initially begin to appear at Cu6Sn5 and solder interface on the die side, with eventual open failure due to excessive void formation along with a severe depletion of Cu Under Bump Metallization (UBM). No failure was observed on the substrate side of the interconnect regardless of the current direction. The enhanced performance of the μPILR interconnect along with other reliability benefits makes it an excellent alternative to conventional solder joints including thin film stack UBMs, thicker copper UBM as well as copper pillar on die.

Effect of Bi addition on the activation energy for the growth of Cu 5Zn 8 intermetallic in the Sn–Zn lead-free solder

The growth kinetics of Cu 5Zn 8 phase has been investigated under long-term thermal exposure conditions using single shear lap joints. The Cu 5Zn 8 phase was formed by reacting Sn–Zn and Sn–Zn–Bi lead-free solders with Cu substrate. A scanning electron microscope (SEM) was used to observe the morphology of the phases and energy dispersive X-ray (EDX) was used to estimate the elemental compositions of the phases. The morphology of the Cu 5Zn 8 phase is rather flat for all the aging temperatures. The Cu 6Sn 5 isolated phase was observed after long-term aging. Intermetallic thickness measurements show that the thickness of the Cu 5Zn 8 phase in the Sn–8Zn–3Bi/Cu system is lower than in the Sn–9Zn/Cu system. Subsequently the activation energy for the growth of Cu 5Zn 8 intermetallic is higher in the presence of Bi. Therefore, the presence of Bi in the Sn-matrix lowers the Zn diffusion and hampers the higher-growth of intermetallics in the Sn–8Zn–3Bi/Cu system compare to the Sn–9Zn/Cu solder interface.

Interfacial interaction of solid cobalt with liquid Pb-free Sn–Bi–In–Zn–Sb soldering alloys

Dissolution kinetics of cobalt in liquid 87.5%Sn–7.5%Bi–3%In–1%Zn–1%Sb and 80%Sn–15%Bi–3%In–1%Zn–1%Sb soldering alloys and phase formation at the cobalt–solder interface have been investigated in the temperature range of 250–450 °C. The temperature dependence of the cobalt solubility in soldering alloys was found to obey a relation of the Arrhenius type cs = 4.06 × 102 exp (−46300/RT) mass% for the former alloy and cs = 5.46 × 102 exp (−49200/RT) mass% for the latter, where R is in J mol−1 K−1 and T in K. For tin, the appropriate equation is cs = 4.08 × 102 exp (−45200/RT) mass%. The dissolution rate constants are rather close for these soldering alloys and vary in the range (1–9) × 10−5 m s−1 at disc rotational speeds of 6.45–82.4 rad s−1. For both alloys, the CoSn3 intermetallic layer is formed at the interface of cobalt and the saturated or undersaturated solder melt at 250 °C and dipping times up to 1800 s, whereas the CoSn2 intermetallic layer occurs at higher temperatures of 300–450 °C. Formation of an additional intermetallic layer (around 1.5 μm thick) of the CoSn compound was only observed at 450 °C and a dipping time of 1800 s. A simple mathematical equation is proposed to evaluate the intermetallic-layer thickness in the case of undersaturated melts. The tensile strength of the cobalt-to-solder joints is 95–107 MPa, with the relative elongation being 2.0–2.6%.

Direct-soldering 6061 aluminum alloys with ultrasonic coating

In this study, the authors applied furnace soldering with ultrasonic coating method to solder 6061 aluminum alloy and investigated the effects of both coating time and soldering temperature on its properties. The following results were obtained: firstly, the solder region mainly composed of four kinds of microstructure zones: rich Sn zone, rich-Pb zone, Sn–Pb eutectic phase and rich Al zone. Meanwhile, the microanalysis identified a continuous reaction product at the alumina–solder interface as a rich-Pb zone. Therefore, the joint strength changed with soldering time and soldering temperature. Secondly, the tensile data had significantly greater variability, with values ranging from 13.99 MPa to 24.74 MPa. The highest value was obtained for the samples coated with Sn–Pb–Zn alloy for 45 s. Fractures occurred along the solder–alumina interface for the 6061 aluminum alloy with its surface including hybrid tough fracture of dimple and tear ridge. The interface could initially strip at the rich Bi zone with the effect of shear stress.

Development of a Novel Microimpact-Fatigue Tester and Its Application to Impact Reliability of Lead-Free Solder Joints

Drop-impact forces cause portable electronic devices to fail. Assessment of resistance against drop-impact force is required to predict life of an electronic package. Several methods have been used to evaluate impact-fatigue life and other advanced methods have been proposed. However, such conventional impact tests require excessive time and cost. In this paper, a novel micro-impact-fatigue tester is developed to overcome such drawbacks of conventional methods. A newly developed impact-fatigue apparatus directly applies impact force to solder joints and measures deformation of the solder joints. The impact-fatigue test apparatus consists of an electromagnetic actuator, an impact-pin, a load-cell, a displacement sensor, and a main frame. Electromagnetic actuator produces a repeatable impact force with a changeable amplitude and pulse duration. Impact-fatigue apparatus was used to test reliability of a lead-free solder (96.5Sn4.0Ag0.5Cu). An evaluation of impact-fatigue life was performed over a wide range of 1-10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">4</sup> cycles with various applied forces ranging from 40 N to 110 N. Two failure modes were observed in a section inspection. First type of failure was a mixed mode failure, where a bulk solder failure and interface failure coincide and the relation between the impact load and fatigue life is almost linear. Stress-based life-prediction model is proposed for the mixed mode failure. The second type of failure, interface failure between the Ni(P) layer and the solder, occurs under a high-load condition. Fatigue life is shorter in the second type of failure than in the mixed mode failure. Brittleness of the interface reduces the impact-fatigue life.

Reliability challenges of automotive power electronics

A high reliability of power electronic modules is an essential requirement for hybrid traction applications. This includes a high capability to withstand the stress of repeated active and passive thermal cycles in order to meet the lifetime requirements. Active power cycling requirements are not especially severe for hybrid traction applications compared to many industrial applications. The lifetime for passive thermal cycles by a change of ambient conditions in contrast is defined by the materials and the architecture of a power module. The classical module design with Cu base plates is limited in lifetime particularly with respect to passive temperature cycles due to CTE mismatch. The advanced pressure contact design eliminates the base plate together with the base plate solder and the terminal solder interconnections and thus enhances the thermal cycling capability. As a synergy effect, this design establishes a very balanced static and transient current distribution for paralleled chips. Finally, the last remaining solder interface – the chip solder layer – can be replaced by an Ag diffusion sinter technology. The presented cycling test results will confirm, that the first 100% solder-free module shows an improved performance in passive and active cycling tests.

A novel assembly technology of aluminum alloy honeycomb structure

In this study, the authors applied the low-temperature ultrasonic soldering method to solder aluminum alloy honeycomb structure and investigated the effects of soldering time and soldering temperature on the properties of the solder. The following results were obtained: The wetting angle changed with applied time and applied temperature. The solder region is mainly composed of four kinds of microstructure zones: rich Sn zone, rich Pb zone, Sn–Pb eutectic phase, and rich Al zone. Microanalysis identified a continuous reaction product at the alumina–solder interface as a rich Pb zone. The wet angle data had significantly greater variability, with values ranging from 10° to 90°. The highest value was obtained for the samples soldered with Sn–Pb–Zn alloy for 30 s. Fractures occurred along the solder–alumina interface for the 6061 aluminum alloy. Fracture surface included hybrid tough fracture of dimple and tear ridge. The interface could firstly strip at the rich Bi zone with the effect of shear stress.

Influence of Thermal Cycling on the Thermal Resistance of Solder Interfaces

The thermal resistance of Si/Sn-52In/Cu sandwiched samples was measured by the laser flash method after various stages of thermal cycling. The thermal resistance was found to increase with increasing thermal cycles after 120 cycles. Cross-sections of the samples were examined by scanning electron microscopy. Cracks were observed in both the solder bulk and the interface between the intermetallic compound and solder. The increase of the thermal resistance was related to widening of the crack segments that were inclined to the interface.

Reliability study of board-level lead-free interconnections under sequential thermal cycling and drop impact

The sequential thermal cycling (TC) and drop impact test are more reasonable to evaluate the reliability of lead-free solder interconnections compared with separate TC test or drop impact test. In this paper, sequential TC (−40 °C/125 °C, 13 min of soak time, 12 min of dwell time, totally 50 min of cycle time) and drop impact test (a sine impact pulse with a peak acceleration of 1500 g and a duration time of 0.5 ms) were conducted to study the failure mechanism of solder interconnections under sequential TC and drop impact test. The TC load has larger effect on the Cu/solder interface at the PCB side than that of Ni (P)/solder interface at the component side. For the thermally cycled samples, the failure location of solder interconnections under drop impact has changed from initiation and propagation along the thin reaction layer (mode 1) between intermetallic compound (IMC) layer and Ni (P) pad at the component side to initiation at the bulk solder and propagation along the Cu 3Sn IMC layer (mode 2) or entirely through the bulk solder (mode 3) at the PCB side. The failure mechanism has also changed from the entirely brittle crack to the mixture of fatigue crack and brittle crack.

Nanoporous Metal Components Bonded with Sputtered Solder for Equation-of-State Laser Targets

Targets were fabricated at Lawrence Livermore National Laboratory and were shot on the Omega laser to study the equation of state of nanoporous copper. The targets had a planar configuration and consisted of a 25-μm-thick beryllium ablator, a 70-μm-thick brominated-polystyrene preheat shield, and a 38-μm-thick aluminum baseplate. A quartz window and a 30-μm-thick nanoporous copper sample were bonded to the baseplate. The interface between the nanoporous copper and the aluminum baseplate was required to be as thin as possible so that it would not disturb the shock as it passed through the target. A process for bonding the nanoporous copper was developed that did not compact it or otherwise degrade its structure. An acceptable bond was achieved by sputtering a layer of indium-based solder onto the surface of the nanoporous copper and on the aluminum baseplate. The components were assembled and heated to melt the solder. The resulting solder interface had a thickness of ~1.5 μm. The targets performed as expected in the experiments, and the interface between the nanoporous copper and the baseplate did not appear to significantly affect the shock as it passed through the target.

Electromigration effect on pancake type void propagation near the interface of bulk solder and intermetallic compound

Due to the reduction in the size of electronic devices and solder interconnects, their current densities are rapidly increasing and the effect of electromigration becomes more critical to interconnect failure. Traditional methods are not sufficient to determine the failure mechanisms of solder interconnects under such high current densities. An approach to predict the electromigration effect upon interconnect failure including void propagation is timely and useful to the electronics industry. The stress gradient in a 95.5Sn-4Ag-0.5Cu eutectic solder with a void defect is analyzed numerically to determine the electromigration and stress migration forces. A kinetic mass diffusion model is developed to predict void width and propagation speed near the interface between the intermetallic compound and bulk solder caused by electromigration. A three-dimensional finite element analysis has been performed to analyze the current crowding effect in the interconnect. The proposed kinetic model gives a reasonable prediction for the void width and propagation velocity as compared with experimental results.

Effect of Ag micro-particles content on the mechanical strength of the interface formed between Sn–Zn binary solder and Au/Ni/Cu bond pads

In this study, addition of Ag micro-particles with a content in the range between 0 and 4 wt.% to a Sn–Zn eutectic solder, were examined in order to understand the effect of Ag additions on the microstructural and mechanical properties as well as the thermal behavior of the composite solder formed. The shear strengths and the interfacial reactions of Sn–Zn micro-composite eutectic solders with Au/Ni/Cu ball grid array (BGA) pad metallizations were systematically investigated. Three distinct intermetallic compound (IMC) layers were formed at the solder interface of the Au/electrolytic Ni/Cu bond pads with the Sn–Zn composite alloys. The more Ag particles that were added to the Sn–Zn solder, the more Ag–Zn compound formed to thicken the uppermost IMC layer. The dissolved Ag–Zn IMCs formed in the bulk solder redeposited over the initially formed interfacial Au–Zn IMC layer, which prevented the whole IMC layer lifting-off from the pad surface. Cross-sectional studies of the interfaces were also conducted to correlate with the fracture surfaces.

The growth kinetics of intermetallic layers at the interface of a solid metal and a liquid solder

During soldering, intermetallic layers can occur at the interface of a solid metal and the saturated or undersaturated solder melt. In systems with a considerable solubility in the liquid state, dissolution causes a manifold drop in layer thickness. Mathematical equations are proposed to evaluate the thickness of any intermetallic layer formed under conditions of simultaneous dissolution in the undersaturated solder melt. The main features of reactive phase formation at the solid metalliquid solder interface are illustrated using the Co-Sn couple with the growing CoSn3 (250°C) and CoSn2 (350°C and 450°C) layers as examples.

Formation and growth of intermetallic phases in diffusion soldered Cu/In–Bi/Cu interconnections

The paper presents microscopy observations of Cu/Bi–22 at.%In/Cu interconnections obtained as a result of diffusion soldering process. The choice of the material as well as technological process allowed getting thermally stable joints dedicated to the electronic equipment thanks to controllable growth of intermetallic phase(s). The θ[Cu 11In 9] phase was present in the Cu/In–22Bi/Cu joint manufactured in the temperature range of 85–200 °C. Two sublayers of θ were identified; the first one, adjacent to copper substrate, contained (except for In and Cu) up to 6 at.% Bi, the second one appeared as the so-called scallops growing into liquid solder. The θ phase enriched in Bi transformed into η[Cu 2In] phase. The η showed also two morphologies. The homogeneous layer of η grew at the Cu/In–Bi solder interface at 300–325 °C. On the other hand, islands surrounded by unreacted solder were present after soldering at 350 °C. The third intermetallic phase δ[Cu 7In 3] coexisted with η at 350 °C. Moreover, the solid solution of In and Bi in Cu is formed at the δ/copper substrate. All the three phases belong to the binary Cu–In equilibrium phase diagram, while Bi appears in the form of separated areas within the θ or η phase.