Peter Borgesen

Recrystallization and Precipitate Coarsening in Pb-Free Solder Joints During Thermomechanical Fatigue

The recrystallization of β-Sn profoundly affects deformation and failure of Sn-Ag-Cu solder joints in thermomechanical fatigue (TMF) testing. The numerous grain boundaries of recrystallized β-Sn enable grain boundary sliding, which is absent in as-solidified solder joints. Fatigue cracks initiate at, and propagate along, recrystallized grain boundaries, eventually leading to intergranular fracture. The recrystallization behavior of Sn-Ag-Cu solder joints was examined in three different TMF conditions for five different ball grid array component designs. Based on the experimental observations, a TMF damage accumulation model is proposed: (1) strain-enhanced coarsening of secondary precipitates of Ag3Sn and Cu6Sn5 starts at joint corners, eventually allowing recrystallization of the Sn grain there as well; (2) coarsening and recrystallization continue to develop into the interior of the joints, while fatigue crack growth lags behind; (3) fatigue cracks finally progress through the recrystallized region. Independent of the TMF condition, the recrystallization appeared to be essentially complete after somewhat less than 50% of the characteristic life, while it took another 50% to 75% of the lifetime for a fatigue crack to propagate through the recrystallized region.

The effect of Sn grain number and orientation on the shear fatigue life of SnAgCu solder joints

A study of the dependence of room temperature shear fatigue lifetime of SnAgCu solder joints on Sn grain number and orientation was conducted. Both essentially single Sn grain and multi (two or three) Sn grain samples are found in many SnAgCu solder joints in the field, and these Sn grain morphologies were examined here. The mean fatigue lifetime was found to be significantly longer for samples with multiple Sn grains than for samples with single Sn grains. For single grain samples, correlations between Sn grain orientation (with respect to the loading direction) and lifetime were observed, providing insight on early failures in SnAgCu solder joints.

Microstructure and Damage Evolution in Sn-Ag-Cu Solder Joints

The implementation of no-Pb soldering is progressing rapidly, often without immediately obvious problems. However, our current understanding of pertinent materials issues is far from sufficient to prevent surprises, and potentially serious miscalculations. The fundamentally different nature of no-Pb from SnPb solder joints has serious consequences for the design and interpretation of accelerated tests. The present work addresses the behavior of ball grid array (BGA), wafer level chip scale package (WL-CSP) and flip chip assemblies with SnAgCu solder joints in long term thermal cycling. Components were assembled onto either Cu or Ni pads on high-Tg printed circuit boards and cycled between 0 o C and 100 o C. Samples were removed at various stages for cross sectioning and microstructural characterization at various stages. Careful examination of the formation of intermetallic compounds at the pad surfaces, the size and orientation of the Sn grains, the morphology and number of precipitates, and the growth of cracks, revealed the sensitivity of the damage evolution to the solder microstructure. The consequences of the observed pattern of damage evolution are discussed.

Assessing the risk of “Kirkendall voiding” in Cu3Sn

Abstract A common location of sporadic, usually unpredictable, premature failures of both SnPb and Pb-free solder joints is within the intermetallic bond to one of the contact pads. Cu pads tend to be less prone to such defects than Ni/Au coated pads, but one phenomenon is a cause of growing concern for high reliability applications. Sporadic intermetallic problems are notoriously difficult to screen for, but at least a number of them are detectable right after assembly. However, the occasional formation of a network of voids within the Cu3Sn layer forming when soldering to Cu pads is rarely noticeable until much later. To make matters worse, efforts to establish practical engineering tests for purposes of identification and screening have usually not been successful. The root cause of the so-called ‘Kirkendall voiding’ has been shown to be associated with the quality of the electroplated Cu, apparently as determined by the incorporation of minute amounts of specific organic impurities. The present work describes and discusses a series of potential tests that allow the sorting of electroplated Cu in terms of its propensity for Cu3Sn voiding during soldering and subsequent aging. The times and efforts involved in these tests differ greatly, as do their effectiveness and accuracy. Individual tests may therefore be useful at various stages from supplier qualification to monitoring of batch to batch variations.

Accelerating the effects of aging on the reliability of lead free solder joints in a quantitative fashion

The properties of lead free solder joints continue to change over a very long period of time in service before the microstructure becomes stable. The quantitative assessment of long term service life by accelerated testing invariably misses this significant effect, and may thus end up seriously misleading. The long term goal of the present work is to establish a protocol for preconditioning of lead free solder joints before thermal cycling or mechanical testing. For this purpose, the state of a solder joint at any given time was characterized in terms of three different room temperature properties, shear strength, shear fatigue resistance, and micro hardness. These properties were measured before and after aging for different lengths of time at different temperatures. Three common lead free alloys were selected for the present study: 98.5Sn-1.0Ag-0.5Cu (SAC105), 96.5Sn-3.0Ag-0.5Cu (SAC305), and 95.5Sn-4.0Ag-0.5Cu (SAC405). The present study did not address effects of solder volume, pad size, pad finish or reflow profile, focusing on 30 mil (760µm) diameter solder spheres reflowed onto solder mask defined OSP coated Cu pads with a typical lead free profile. Isothermal aging was conducted for up to 3,000 hours at temperatures of 70°C, 100°C, and 125°C respectively. As expected, the resulting room temperature properties all decreased with aging time, and faster so for higher aging temperatures. Some of the acceleration factors extracted for the evolution of the individual properties did, however, differ greatly for a given alloy. The only way to establish the same microstructure, and thus the same combination of properties, faster by annealing at a higher temperature is thus to fully stabilize it. This takes thousands of hours even at 125°C, i.e. it is not practical for real assemblies.

Improving Copper Electrodeposition in the Microelectronics Industry

Sporadic voiding within the interfacial Cu3Sn intermetallic compound (IMC) layer-sometimes referred to as ¿Kirkendall voiding¿-has been found to lead to degradation of solder joint reliability in board level, mechanical shock testing. It has been suggested that the voiding phenomenon is a result of the incorporation of organic impurities in the copper (Cu) deposit during electroplating. In the present study, Cu samples were electroplated galvanostatically from a generic solution, containing Cl- ions, as well as a suppressor [polyethylene glycol (PEG)], and a brightener (bis(3-sulfopropyl) disulfide, SPS) as additives. Overpotential transients were measured during electroplating with a range of current densities (0.5-40 mA cm-2) in baths with various compositions. Effects of the bath chemistry on the Cu surface morphology, as well as on the propensity for voiding after soldering, were also investigated. Elemental analysis of selected samples was performed by SIMS. Plating at 10-20 mA·cm-2 with an optimized bath composition led to Cu with a fine-grain structure and smooth appearance. Solder joints formed from these deposits remained void free after soldering and thermal aging. Lower current densities, ran in the same plating bath, led to a significant propensity for voiding, apparently because of incorporation of, principally, SPS and its breakdown products into the growing layer. Continuous plating at 10 mA cm-2 for up to 18 hours without replenishment revealed a strong dependence on bath aging, with Cu changing from ¿void-proof¿ to clearly ¿void-prone.¿ These trends were attributed to the different rates of consumption for PEG and SPS and changes in the contaminants being incorporated in the deposits. In general, differences in the voiding behavior of the plated Cu could be predicted by monitoring a set of characteristic overpotential transient signatures.

On the nature of pad cratering

Intent was to study the effects and interactions of the main process and environmental factors which are likely to affect pad performance. Real time observation of crack propagation under the pad achieved by cross-sectioning very carefully to the middle of the pad and then subjecting it to cyclic loading revealed some important location dependencies. Experimentation also involved cyclic bending of prepared samples with different preconditioning levels of reflow, moisture and cycling, followed by measurement of the crack area under the pad, post cycling. This shed light on the complex degradation mechanisms at play which did not necessarily have linear dependencies nor were they as basic intuition may have led us to believe. An effort was made to develop an analytical model with the help of Multiple Regression techniques.

Assessment of PCB pad cratering resistance by joint level testing

Cracking of the laminate under the solder connect pads, known as pad craters, is a reliability issue related to mechanical stresses generated from either mechanical or thermal loading. The current study aims to establish a mechanistic understanding of pad cratering through the use of joint-level testing techniques. Both strength and cyclic loading lifetime are considered, and the results indicate that these two loading modes are not correlated. The individual crack paths within a crater are found to differ with laminate material, glass style (if any), and micro-via details. Various degradation mechanisms are found to have a significant effect on this failure mode with both thermal and moisture exposure showing decreased laminate performance. Finally, the relationship between joint-level testing is compared to the performance in board level drop test. Here we see that joint-level testing is far more general, in terms of qualifying the robustness of the laminate.

Effects of assembly process variables on voiding at a thermal interface

Too often, the effects of assembly process parameters are not sufficiently accounted for in the optimization of thermal interface performance. This becomes increasingly critical as demands on this performance continue to grow and alternative processes are developed. Notably, stencil printing is proving a competitive alternative to the traditional dispensing of thermal interface materials (TIMs), with potentially significant gains in units processed per hour (UPH) for some applications. The two techniques may, however, pose quite different challenges in terms of material flow, the resulting filler particle distribution and the risk of air bubble entrapment. Another part of the adhesive attachment process certain to affect void formation and growth, as well as possibly filler particle distribution, is the final cure. In addition, voids may severely reduce assembly robustness and reliability. The present work offers a discussion and a first case study to identify and illustrate voiding mechanisms for a particular TIM between a heat spreader and the back of a flip chip. Pronounced differences were observed between stencil printing and dispensing in terms of initial void formation, apparently related to the specific properties of the material. Measurements of the effects of heat ramp rate and peak temperature showed the subsequent evolution and final void size distribution to be determined by the initial part of the cure profile up to the material gelling temperature.

Aspects of the structural evolution of lead-free solder joints

Studies of the formation of intermetallic compounds at some lead-free solder/metallization interfaces are briefly reviewed in this article. SnAgCu/Ni and SnAgCu/Cu interfaces are examined in particular. It has been found that (Cu,Ni)6Sn5 forms at SnAgCu/Ni interfaces until copper is depleted from the solder matrix. This article also contrasts the formation of (Au,Ni)Sn4 and related compounds in PbSn/Ni solder joints and lead-free solder joints.