Luke Wentlent

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.

Improving the thermomechanical behavior of lead free solder joints by controlling the microstructure

Effects of solder alloy, volume and pad finishes on various aspects of microstructure and the corresponding thermomechanical properties of SnAgCu solder joints were investigated. Particular attention was focused on the behavior of solder joints with interlaced Sn grain morphologies. Crossed polarizer microscopy and scanning electron microscopy (SEM) were used to characterize Sn grain structures. Precipitate sizes and distributions were measured using backscattered scanning electron microscopy and quantified using image analysis software. Mechanical properties including hardness and indentation creep were measured. Results show that the amount and frequency of interlacing increased as the joint size decreased, as the amount of Ag in the solder increased, and if the joint was reflowed on ENIG substrates. The interlaced structure was harder and more creep resistant compared to the common beach ball morphology. Image analysis results showed this to be related to much higher densities of secondary precipitates in the interlaced regions. A mechanistic understanding of the microstructure is discussed and recommendations are made as to the design of more reliable solder joints.

Dependence of SnAgCu solder joint properties on solder microstructure

It is well known that variations in the microstructure of lead free solders greatly affect their thermomechanical properties. Sn grain size, orientation and number, as well as secondary Ag3Sn and Cu6Sn5 precipitate sizes and numbers, are all seen to influence the mechanical response of solder joints during isothermal and thermal cycling. The solidification temperature of a SnAgCu solder joint dramatically affects its microstructure. Generally, smaller solder balls (e.g. CSP) undercool more, and thus their microstructure and properties are very different than larger solder balls (e.g. BGA). We report results of a study of the effects of solder joint volume, and pad sizes, on the microstructure and thermomechanical properties of solder joints. Solder joint shapes and dimensions spanned the ranges typical of BGA and CSP assemblies. Temperatures of solidification during cool-down were quantified by differential scanning calorimetry. Sn grain structures were characterized by crossed polarizer microscopy and scanning electron microscopy with electron backscattered diffraction. Precipitate sizes and distributions were measured using backscattered scanning electron microscopy. Corresponding properties, including hardness, strength and fatigue resistance were measured before and after aging for various lengths of times at temperatures up to 125ºC. Smaller solder joints on smaller pads were shown to be harder and stronger than larger ones, but to age faster and eventually end up softer and weaker.

Recrystallization behavior of lead free and lead containing solder in cycling

The present work addresses the effects of thermomechanical history on the recrystallization behavior of lead free and backward compatible solder joints. 30 mil SAC305 balls were reflowed onto BGA pads using either a SAC305 or a eutectic SnPb paste. Systematic variations of the rate of recrystallization with precipitate coarsening as well as with cycling and annealing parameters were characterized and correlated with observations from thermal cycling experiments on lead free BGA assemblies. The introduction of a few percent of Pb has been seen to not only affect the distribution of secondary precipitates but also add minute inclusions of Pb within the individual Sn dendrites. These also act as barriers to dislocation motion but do not coarsen as rapidly as the precipitates, and we find recrystallization in cycling to be strongly delayed.

Controlling the Superior Reliability of Lead Free Assemblies With Short Standoff Height Through Design and Materials Selection

We have recently demonstrated a significantly longer life in accelerated thermal cycling for Land Grid Arrays (LGAs) assembled only with SAC305 solder paste than for the corresponding SAC305 based BGA assemblies. This superior performance was shown to be a direct effect of the solder microstructure. The final Sn solidification temperature strongly affects the initial microstructure of a SnAgCu solder joint, including the Sn grain morphology, and thus the thermomechanical behavior of the joint. Right after reflow, larger BGA joints of SnAgCu alloys, which solidify at higher temperature, reveal either a single β-Sn grain or three large grains with clearly defined boundaries formed by cyclic twinning. The orientations of the highly anisotropic Sn grains are not yet controllable in manufacturing, leading to substantial statistical scatter in the performance of the solder joints. Typical LGA solder joint dimensions, however, tend to facilitate greater undercooling and the formation of an alternative interlaced twinning microstructure.A systematic study was undertaken to identify the parameters that control the interlaced twinning microstructure. Sn grain structures were characterized by crossed polarizer microscopy and electron backscatter diffraction (EBSD). Precipitate sizes and distributions were measured using backscattered scanning electron microscopy and quantified using image analysis software. Systematic effects of solder alloy, dimensions and pad finishes were identified. Recommendations are made as to design and materials selection. The practicality of controlling the desired microstructure, as well as potential disadvantages for certain applications is discussed.Copyright

Effects of variable amplitude loading on lead-free solder joint propoerties and damage accumulation

Current extrapolations of accelerated test results to life in long term service may be greatly misleading when it comes to lead free solder joints. Notably, realistic service conditions are almost never well approximated by cycling with a fixed amplitude. However, recent results have demonstrated the consistent breakdown of common damage accumulation rules. In isothermal cycling tests the remaining life, after a step-down in amplitude, was invariably shorter than predicted by such a rule, while a step-up tended to have the opposite effect. The present work offers a mechanistic explanation for this and a basis for the development of a practical approach to the assessment of life under realistic service conditions. Realistic BGA joints were cycled individually in a micromechanical tester, monitoring the solder stiffness and the inelastic energy deposition. Cycling was seen to first cause rapid hardening, followed by leveling off in a “cyclic saturation” stage and eventually the initiation and growth of a crack until failure. A temporary increase in amplitude during cycling caused a lasting reduction in hardness, and thus enhanced inelastic energy deposition and damage evolution, afterwards as well. This would dominate during repeated increases and decreases, eventually shortening the remaining life dramatically.

Effect of microstructure evolution on Pb-free solder joint reliability in thermomechanical fatigue

The performance of solder joints in an accelerated thermal cycling (ATC) test depends on the accumulation of fatigue damage, which is affected by various simultaneously evolving microstructure features. A mechanistic understanding of microstructure evolution and damage accumulation in Pb-free solder joints will help us better interpret thermal cycling test results. In this study various components, including Ball Grid array (BGA), Land Grid Array (LGA), Thin Small-Outline Package (TSOP), Quad Flat No-Lead (QFN) and surface mount resistors, were compared for their failure mechanisms. The focus was on the evolution of microstructure, particularly recrystallization, and its correlation to fatigue crack propagation. A general understanding of damage accumulation was proposed, based upon the observed recrystallization behavior and the fatigue life dependence on ATC parameters.

Statistical Variations of Solder Joint Fatigue Life Under Realistic Service Conditions

Common damage accumulation rules fail to predict the fatigue life of solder joints under realistic service conditions where cycling amplitudes vary over time. A modification of Miners rule of linear damage accumulation has been proposed that accounts for effects of amplitude variations in, for example, the vibration of microelectronic assemblies with lead-free solder joints on the average or characteristic fatigue life. We are, however, obviously much more concerned with the first failure across a very large sample set. Prediction of, say, the first failure out of 10000 or a million would require the extrapolation of experimental failure distributions and the assumption of a shape of this distribution. Even qualitative comparisons of accelerated test results and their scatter should account for effects of amplitude variations. We have argued that the long-term life of solder joints in vibration or cyclic bending is limited by the accumulation of inelastic work, and that much can be learned from the low cycle fatigue behavior in shear. Individual ball grid array scale SAC305 and SAC105 solder joints were cycled in shear at room temperature with combinations of two different stress amplitudes. Relying on our modified Miners rule and the associated understanding of the effects of amplitude variations, we show that the statistical uncertainty in the fatigue life of solder joints under a specific set of realistic service conditions must be significantly greater than measured in fixed amplitude cycling tests. The predicted failure distribution was best fit by a Weibull distribution over a limited range, but we argue that the assumption of such a distribution is likely to be increasingly conservative when it comes to the prediction of earlier failure. Estimates are provided for the potential errors.

A Mechanistic Thermal Fatigue Model for SnAgCu Solder Joints

The present work offers both a complete, quantitative model and a conservative acceleration factor expression for the life span of SnAgCu solder joints in thermal cycling. A broad range of thermal cycling experiments, conducted over many years, has revealed a series of systematic trends that are not compatible with common damage functions or constitutive relations. Complementary mechanical testing and systematic studies of the evolution of the microstructure and damage have led to a fundamental understanding of the progression of thermal fatigue and failure. A special experiment was developed to allow the effective deconstruction of conventional thermal cycling experiments and the finalization of our model. According to this model, the evolution of damage and failure in thermal cycling is controlled by a continuous recrystallization process which is dominated by the coalescence and rotation of dislocation cell structures continuously added to during the high-temperature dwell. The dominance of this dynamic recrystallization contribution is not consistent with the common assumption of a correlation between the number of cycles to failure and the total work done on the solder joint in question in each cycle. It is, however, consistent with an apparent dependence on the work done during the high-temperature dwell. Importantly, the onset of this recrystallization is delayed by pinning on the Ag3Sn precipitates until these have coarsened sufficiently, leading to a model with two terms where one tends to dominate in service and the other in accelerated thermal cycling tests. Accumulation of damage under realistic service conditions with varying dwell temperatures and times is also addressed.

Effects of Amplitude Variations on Deformation and Damage Evolution in SnAgCu Solder in Isothermal Cycling

Although apparently simpler than in thermal cycling, the behavior of SnAgCu (SAC) solder joints in cyclic bending or vibration is not currently well understood. The rate of damage has been shown to scale with the inelastic work per cycle, and excursions to higher amplitudes lead to an apparent softening, some of which remains so that damage accumulation is faster in subsequent cycling at lower amplitudes. This frequently leads to a dramatic breakdown of current damage accumulation rules. An empirical damage accumulation rule has been proposed to account for this, but any applicability to the extrapolation of accelerated test results to life under realistic long-term service conditions remains to be validated. This will require a better understanding of the underlying mechanisms. The present work provides experimental evidence to support recent suggestions that the observed behavior is a result of cycling-induced dislocation structures providing for increased diffusion creep. It is argued that this means that the measured work is an indicator of the instantaneous dislocation density, rather than necessarily reflecting the actual work involved in the creation of the damage.