Search

A recent publication (Jagtap et al. Scr Mat 182:43, 2020) reported real-time observations of stress-driven whisker nucleation and growth in a thin film of Sn subjected to pressure on part of its surface. This paper describes a model of the experiment. Predicted whisker volume in the film adjacent to the loaded area is in good agreement with experimental measurements. The model also predicts correctly the variation of whisker density with distance from the punch at the end of the experiment but underestimates the rate of whisker nucleation during the first 50 h after the film is loaded. Comparison of model predictions with experiments provides clear evidence for a threshold stress for whisker growth, with an initial value of 17 MPa, which increases with whisker length to a saturated value of 22 MPa after whiskers reach a length of 19 µm. Implications for whisker growth in Sn films on Cu substrates are discussed.

A model for the kinetics of residual stress evolution is described that can be applied to both the relaxation during a growth interruption and the recovery following resumption of growth. The stress is attributed to simple atomistic processes based on reversible diffusion of atoms in and out of the grain boundary. Each of these processes is considered separately to understand how the rates differ during relaxation and recovery and how they depend on grain size, film thickness, temperature, deposition rate, and initial stress. The model is compared with experimental data obtained by others for several materials (Fe, Ni and Au).

We investigate whisker nucleation and growth kinetics in Sn coatings induced by the application of mechanical pressure. The pressure was applied to the samples using a clamping fixture that enables a constant pressure to be maintained over long duration of time. The fixture is small enough so that it could be transferred inside a Scanning Electron Microscope (SEM) chamber to observe the stress driven whisker nucleation and growth in real-time while the applied pressure is maintained on the Sn films.

Stress evolution during reactive DC magnetron sputtering deposition in TiN, ZrN and TaN films at different growth rates, pressures and relative gas flow rates was measured by a wafer curvature technique. The results indicate that, for a given growth rate, the stress in each material is tensile at higher pressures and becomes more compressive with lower pressures. At lower pressures, the stress tends to become more compressive with increasing growth rate. Increasing the gas flow rate of N-2 relative to Ar is observed to enhance compressive stress generation in TaN but has a weaker effect on TiN and ZrN. At lower pressure, the compressive stress magnitude scales with TiN < ZrN < TaN. The stress at different processing conditions was analyzed in terms of a kinetic model that includes the effects of growth kinetics and energetic bombardment. The resulting parameters are compared with calculations from binary collision Monte-Carlo simulation codes.

Stress in thin films is a critical issue that affects the performance and lifetime of coatings. Multiple parameters have been shown to affect the stress evolution, including the growth rate, temperature, grain size and gas pressure. In addition, the deposition method (e.g., evaporation, electrodeposition, sputtering) is known to have an influence. To develop a unified picture of the processes controlling stress, a kinetic model has been developed that includes the role of different processing parameters and the microstructural evolution. Results are shown from applying the model to numerous published results for transition metals deposited by different methods. Trends in the fitting parameters are discussed in terms of the values expected from the physical mechanisms included in the model.

The growth rate and pressure dependence on the intrinsic stress in sputter deposited Cu thin films has been investigated and compared to a kinetic growth model, which contains both growth and energetic contributions to stress in its description. Since microstructure also has a strong effect on intrinsic growth stress, we have been able to systematically control a fixed grain size over multiple growth conditions spanning 0.012 nm/s to 2.4 nm/s deposition rates and 0.267 Pa to 2.667 Pa pressures using a seed layer prior to film deposition. At high deposition pressures, the stress became more tensile as the growth rate increased. In the low deposition pressure regime, the stress became more tensile with increases in deposition rate until a critical cross-over point where upon further increases in deposition rate resulted in the stress becoming more compressive. This cross-over has been explained in terms of the energetic contributions to the stress and the intrinsic high mobility of Cu to reveal this behavior. The fitting parameters and corresponding stress contributions from the kinetic model were extracted and compared to other films and deposition techniques. Though caution should be used in comparing absolute values, the kinetic model revealed the correct trends in predicting energetic trapping of defects between low and high mobility films as well as similar growth stress values, which are independent of energetic contributions, between sputtering and electrodeposition. These results suggest that the kinetic model shows promise in fitting different materials and deposition techniques intrinsic stress behavior.

An analytical model for the evolution of residual stress in polycrystalline thin films is used to analyze numerous previously reported wafer curvature measurements obtained for a variety of materials and processing conditions. The model, which has been described in previous publications, considers stress-generating mechanisms that occur at the grain boundary as it forms between adjacent grains and stress due to the subsurface grain growth in layers that have already been deposited. Current work extends the model to include different types of microstructural evolutions. A set of parameters for each dataset is obtained by non-linear least square fitting. Model parameters that are not expected to depend on the processing conditions are constrained to have a common value when fitting the multiple datasets for each material. The dependence of the fitting parameters on the material and process conditions is evaluated and compared with the physical mechanisms implemented in the model. Published under an exclusive license by AIP Publishing.

We report closed-form formulas to calculate the incremental-deposition stress, the elastic relaxation stress, and the residual stress in a finite-thickness film from a wafer-curvature measurement. The calculation shows how the incremental deposition of a new stressed layer to the film affects the amount of the film/wafer curvature and the stress state of the previously deposited layers. The formulas allow the incremental-deposition stress and the elastic relaxation to be correctly calculated from the slope of the measured curvature versus thickness for arbitrary thicknesses and biaxial moduli of the film and the substrate. Subtraction of the cumulative elastic relaxation from the incremental-deposition stress history results in the residual stress left in the film after the whole deposition process. The validities of the formulas are confirmed by curvature measurements of electrodeposited Ni films on substrates with different thicknesses.

To understand how and why Sn whiskers nucleate, measurements were performed in which the thermal expansion mismatch was used to induce strain in the samples. The current work focuses on the effect of grain size by studying samples with a thickness of 8 μm to compare with previous studies on 2.5 μm samples. The stress evolution during cycles of heating is determined by monitoring the resulting curvature in the sample. At the same time, the number of whisker nuclei are measured with optical microscopy which enables the relationship between film stress and nucleation rate to be quantified, where the stress relaxation kinetics depend on the grain size. The rate of stress relaxation is greater in the thick samples than the thin samples for the same heating cycle and the final stress after relaxation is smaller for the thicker samples. The whisker volume evolution was determined by using short sequences of heating and then measuring the size of different features after each interval in the scanning electron microscope. This enabled quantification of the relationship between film stress and whisker growth rate. The growth rate depends on the stress above the final stress, similar to our previous results for the thinner samples. The whisker nucleation kinetics in the 8 μm samples can also be explained with the same model used for the 2.5 μm samples. However, the stress-dependent activation energy is shifted to lower values.

Molecular dynamics (MD) simulations are performed to analyze the evolution of defects and stress caused by low-energy (25-400 eV) Ar bombardment of polycrystalline Mo thin films. Simulations were performed under different conditions to explore the role of grain boundaries (GBs), Ar-atom's kinetic energies, and incident directions on defect generation. The results show that the GBs enhance the production of interstitial defects, producing compressive stress at much larger depths than the implantation range. This is attributed to sequences of atomic collisions that knock atoms into GBs instead of a diffusional process. Decreasing the grain size or increasing the kinetic energy of the incoming particles increases the number of interstitials in the film, which increases the compressive stress. The incident angle has little influence on the number of interstitials in the film, but the sputtering yield depends on the polar angle. The stress distribution can be modeled by a superposition of the different defect distributions with the appropriate relaxation volumes.

Although Sn whiskers have been studied extensively, there is still a need to understand the driving forces behind whisker nucleation and growth. Many studies point to the role of stress, but confirming this requires a quantitative comparison between controlled stress and the resulting whisker evolution. Recent experimental studies applied stress to a Sn layer via thermal cycling and simultaneously monitored the evolution of the temperature, stress and number of nuclei. In this work, we analyze these nucleation kinetics in terms of classical nucleation theory to relate the observed behavior to underlying mechanisms including a stress dependent activation energy and a temperature and stress-dependent whisker growth rate. Non-linear least squares fitting of the data taken at different temperatures and strain rates to the model shows that the results can be understood in terms of stress decreasing the barrier for whisker nucleation.

Sn whiskers are a re-emerging reliability issue for Pb-free manufacturing. Although stress is considered as the most likely driving force for whisker formation, there is little quantitative understanding of how stress and whisker growth are correlated. To address this issue, we have developed an analytical model which accounts for strain generation (via thermal expansion mismatch) and multiple relaxation processes (rate-dependent plasticity and diffusion-controlled whisker growth). This model is used to analyze the results of thermal cycling experiments in which the evolution of stress and average volume of whiskers/hillocks were tracked as the features grew on the Sn surface. Predictions of the model are found to be in a good agreement with the time-dependent whisker growth and stress relaxation behaviors that have been observed in the experimental measurements.

The quality of a Sn film deposited by the aerosol process is compared against the quality of Sn films deposited with traditional electroplating. Using the aerosol additive deposition technique, a Sn film was deposited on a brass substrate and exposed to room (25°C) temperature environments for 30 days, followed by a laser photosintering process. The film characteristics and content, formation of intermetallic compounds, residual stress distribution, grain texture, and the tendency of the film to grow Sn whiskers were analyzed. The preliminary results show a successful deposition of Sn film with an aerosol jet process and tensile residual stresses, whereas it was compressive in nature for electroplated Sn film. X-ray diffraction results also show the absence of intermetallic compound (IMC) formation in the aerosol jet-deposited film, while electroplated Sn film has a significant presence of IMC. The aerosol jet-deposited Sn film has the potential to resist nucleation of Sn whiskers under the operating conditions used in this study.

Lifting off gold films A method for growing and removing single-crystal gold films can be used to create a flexible and transparent substrate for devices. Mahenderkar et al. grew gold films on the face of a silicon wafer and then used photoelectrochemistry to undergrow a sacrificial silicon dioxide layer. This layer allowed the gold film to be peeled off with adhesive tape. A 28-nm-thick gold foil showed a minimal increase in sheet electrical resistance after 4000 bending cycles. Flexible films of single-crystal cuprous oxide and of zinc oxide nanowires were then grown on the gold foils. Science , this issue p. 1203

Thin films can develop large residual stresses during their growth that significantly impact their performance. Therefore, there is a need to understand how the stress is related to the developing film structure and underlying kinetic processes. In this work, we describe measurements of stress and the corresponding grain structure during electrodeposition of Ni and Cu films. For Ni deposition, the grain size stays nearly constant during growth and the stress reaches a nearly constant steady-state. For Cu deposition, the grain size grows as the thickness increases and the microstructural evolution affects the evolution of the stress. To remove the effect of subsurface grain growth on the stress, measurements were also done with periodic pauses that allowed the stress induced by grain growth to saturate. We interpret the results in terms of a kinetic model for stress evolution that focuses on the developing boundary between adjacent grains while the film is deposited. The effect of grain growth on stress for different types of microstructural evolution is also discussed. After accounting for the stress from subsurface grain growth, the results are consistent with the model for the dependence on growth rate and grain size at the surface. Keywords: Thin films, Thin film growth, Polycrystalline films, Residual stress, Microstructure, Electrodeposition

We report the stress and microstructural evolution for a series of Ni thin films sputter deposited over a range of rates (0.076 and 0.250 nm/s), pressures (0.27, 0.67 and 1.33 Pa), and substrate temperatures (ambient, 100, and 200 °C). In general, as the sputtering pressure increased, the stress-thickness product, measured by wafer curvature, became tensile and trended with an increase in pressure regardless of the deposition rate. However, at the lowest sputtering rate and highest substrate temperature, the films exhibited a compressive growth stress. The collective data was then fitted to a kinetic model that accounts for the stress generation at the grain boundaries, from grain growth, and from the creation of defects within the film. The model's predicted fitted parameters matched well to the experimental measurements except for films deposited at the lowest deposition pressure. These particular films exhibited a bimodal grain size distribution which could not be accommodated by the model's use of a singular grain diameter. Nevertheless, in monomodal grain sizes, the results do provide support for the kinetic model in helping to ascertain the various contributing factors for the intrinsic stress development and their microstructural relationship in context to the Thornton zonal morphology descriptors for thin films and coatings.

In situ wafer curvature measurements have been conducted on sputter-deposited nitride films (TiN, ZrN) to investigate the effect of growth rate and pressure on stress. For each growth rate, the stress becomes more compressive at lower pressure. At high pressure the stress shows a weak dependence on growth rate, while at low pressure, it becomes more compressive at higher growth rates. The results are interpreted in terms of a kinetic model that includes the effects of growth kinetics and energetic particle bombardment. The overall agreement suggests that the model can reproduce the dependence on processing conditions in multiple nitride systems.

Minimization of stress-induced mechanical rupture and delamination of conducting polymer (CP) films is desirable to prevent failure of devices based on these materials. Thus, precise in situ measurement of voltage-induced stress within these films should provide insight into the cause of these failure mechanisms. The evolution of stress in films of polypyrrole (pPy), doped with indigo carmine (IC), was measured in different electrochemical environments using the multibeam optical stress sensor (MOSS) technique. The stress in these films gradually increases to a constant value during voltage cycling, revealing an initial break-in period for CP films. The nature of the ions involved in charge compensation of pPy[IC] during voltage cycling was determined from electrochemical quartz crystal microbalance (EQCM) data. The magnitude of the voltage-induced stress within pPy[IC] at neutral pH correlated with the radius of the hydrated mobile ion in the order Li(+)Na(+)K(+). At acidic pH, the IC dopant in pPy[IC] undergoes reversible oxidation and reduction within the range of potentials investigated, providing a secondary contribution to the observed voltage-induced stress. We report on the novel stress response of these polymers due to the presence of pH-dependent redox-active dopants and how it can affect material performance.

Polycrystalline thin films often have residual stress that is affected by the growth conditions (temperature, deposition rate, etc.) and material parameters. Real-time wafer curvature measurements from many groups have enabled the stress evolution with thickness and its dependence on processing conditions to be quantitatively determined for a large number of systems. Based on these results, we have developed a model for the stress evolution that focuses on the processes that occur at the evolving grain boundary between adjacent grains. The model includes parameters to predict the dependence of the stress on the diffusivity, growth rate, grain size and morphology. In this manuscript, we review the physical basis for this model and compare its predictions with two sets of measurements. In the first case, we consider the evolution of stress in films grown on patterned substrate in which the morphology of the film is known precisely, allowing for direct comparison with the model. In the second case, we consider the steady-state stress as a function of the deposition rate in electrodeposited films.

Recent experiments and atomic scale computations indicate that the standard continuum models of diffusion in stressed solids do not accurately describe transport, deformation and stress in Li–Si alloys. We suggest that this is because classical models do not account for the irreversible changes in atomic structure of Si that are known to occur during a charge–discharge cycle. A more general model of diffusion in an amorphous solid is described, which permits unoccupied Si lattice sites to be created or destroyed. This may occur as a thermally activated process; or as a result of irreversible plastic deformation under stress. The model predicts a range of phenomena observed in experiment that cannot be captured using classical models, including irreversible changes in volume resulting from a charge–discharge cycle, asymmetry between the tensile and compressive yield stress, and a slow evolution in mechanical and electrochemical response over many charge–discharge cycles.

Analytical solutions are derived for the distribution of Li concentration, chemical potential, and stress in a spherical particle of an active Li insertion material containing a propagating phase boundary. Results are presented both for elastic and inelastic host materials. We focus in particular on understanding the coupling between the solution thermodynamics, transport, and mechanical forces in the particle. It is shown that solution chemistry can have a strong influence on the stress state; and conversely, the stress state may significantly alter the equilibrium conditions at the phase boundary and the driving forces for phase boundary migration.

Whiskers and hillocks that grow out of Sn-based coatings are a critical reliability issue in Pb-free electronics. Although their growth is widely regarded as a stress-relaxation mechanism, quantitative understanding of the relationship between the stress, growth kinetics, and strain relaxation is still lacking. In this work, the well-controlled strain induced by thermal-expansion mismatch was used to study the whiskering behavior of electroplated Sn films. Stress was quantified by monitoring wafer-curvature and the density of whiskers and hillocks was measured simultaneously by use of optical microscopy. Evolution of the volume of individual features was also measured by scanning electron microscopy after different periods of heating. The measurements were used to develop a model for temperature-dependent and stress-dependent growth kinetics of whiskers and hillocks and to determine the amount of strain relaxation which occurs as a result of their formation.

Sn whiskers/hillocks are believed to form due to stress in the layers, but the dependence on the stress has been difficult to quantify. We therefore used the thermal expansion mismatch between Sn thin films and Si substrates to induce controlled stress by heating. This enables us to measure the average stress in the layer (using wafer curvature) at the same time as we monitor the nucleation rate (using optical microscopy). Scanning electron microscopy of the surface after intervals of heating is also used to quantify the whisker volume as a function of stress and time. The results allow us to determine the dependence of the whisker nucleation rate and the growth rate on the applied stress. They also show that whisker formation is not the dominant mode of plastic strain relaxation in the Sn layer.

To explore the mechanisms controlling residual stress in thin films, we have measured the stress evolution during electrodeposition of Ni on lithographically patterned substrates with different pattern spacings and growth rates. Studying films with a controlled island geometry allows us to relate the stress (measured using wafer curvature) to the evolution of the morphology. We analyze the measurements with a model that focuses on the stress that develops where adjacent islands grow together to form new elements of grain boundary.

We have simultaneously measured the evolution of stress and formation of whiskers/hillocks in Sn layers using stress induced by thermal-expansion mismatch. The formation kinetics suggest that whisker initiation is controlled by a nucleation process. We use measurements of the nucleation rate at different stresses and temperatures to determine quantitatively how the activation barrier for nucleation is decreased by the stress in the layer.

Measurements of stress evolution during low-energy argon ion bombardment of Si have been made using a real-time wafer curvature technique. During irradiation, the stress reaches a steady-state compressive value that depends on the flux and energy. Once irradiation is terminated, the measured stress relaxes slightly in a short period of time to a final value. To understand the ion-induced stress evolution and relaxation mechanisms, we account for the measured behavior with a model for viscous relaxation that includes the ion-induced generation and annihilation of flow defects in an amorphous Si surface layer. The analysis indicates that bimolecular annihilation (i.e., defect recombination) is the dominant mechanism controlling the defect concentration both during irradiation and after the cessation of irradiation. From the analysis, we determine a value for the fluidity per flow defect.

Sn whiskers are believed to form in response to stress in layers used as protective coatings. However, what makes them form at specific sites on the surface is not known. We have used thermal expansion mismatch to induce stress and observe the resulting whisker formation. Cross-sectional measurements of the region around whiskers show that there are oblique grain boundaries under the whiskers that are not seen in the as-deposited columnar structure. The kinetics also suggest that the whiskering sites may be formed by a nucleation process. Based on these results, we propose a nucleation mechanism in which the boundaries of the surrounding grains migrate due to strain energy differences and create oblique boundaries at which whiskers can form. A simple model is developed to predict the stress-dependence of the nucleation rate.

Kinetic Monte Carlo simulations have been used to model the evolution of the morphology and the stress in a polycrystalline film. Stress is incorporated into the simulation by considering the processes that occur at the grain boundary as the film grows, in the same way that has been used for analytical models. A section of a grain is simulated by assuming the grain boundary can act as a sink and a source for atoms on the surface; the relative rates of diffusion into and out of the boundary depend on the surface kinetics and the stress. This approach allows the complex interaction between the film growth kinetics, surface morphology evolution, and stress to be investigated. For a decreasing growth rate, the growth mode changes from being dominated by nucleation of new clusters to step-flow. This affects the flux of atoms into the grain boundary and makes the corresponding stress more compressive. The effect of growth interrupts and a barrier to interlayer diffusion are also investigated.Kinetic Monte Carlo simulations have been used to model the evolution of the morphology and the stress in a polycrystalline film. Stress is incorporated into the simulation by considering the processes that occur at the grain boundary as the film grows, in the same way that has been used for analytical models. A section of a grain is simulated by assuming the grain boundary can act as a sink and a source for atoms on the surface; the relative rates of diffusion into and out of the boundary depend on the surface kinetics and the stress. This approach allows the complex interaction between the film growth kinetics, surface morphology evolution, and stress to be investigated. For a decreasing growth rate, the growth mode changes from being dominated by nucleation of new clusters to step-flow. This affects the flux of atoms into the grain boundary and makes the corresponding stress more compressive. The effect of growth interrupts and a barrier to interlayer diffusion are also investigated.

Compressive stress is believed to be the primary driving force that makes Sn whiskers/hillocks grow, but the mechanisms that create the stress (e.g., intermetallic compound growth) are difficult to control. As an alternative, the thermal expansion mismatch between the Sn layer and the substrate can be used to induce stress in a controlled way via heating and cooling. In this work, we describe real-time experiments which quantify the whiskering behavior and stress evolution during cyclic heating. The density of whiskers/hillocks is measured with an optical microscope, while the stress is measured simultaneously with a wafer-curvature-based multi-beam optical stress sensor. Results from three thermal cycles are described in which the samples are heated from room temperature to 65 °C at rates of 10, 30, and 240 °C/h. In each case, we find that the whisker/hillock formation is the primary source of stress relaxation. At fast heating rates, the relaxation is proportional to the number of hillocks, indicating that the stress is relaxed by the nucleation of many small surface features. At slower heating rates, the whisker/hillock density is lower, and continual growth of the features is suggested after nucleation. Long whiskers are found to be more likely to form in the slow heating cycle.

Sn whiskers are thin filaments that grow spontaneously out of the surface of coatings on Cu and have become a critical reliability problem in Pb-free electronics. In this review, we focus on what creates the driving force for whiskers (or more rounded “hillocks”), and what determines where on the surface they will form. Experimental studies are reviewed that quantify the relationship between the Cu–Sn intermetallic (IMC) formation, stress in the layer and whisker/hillock density. Measurements of the mechanical properties show how stress relaxation in the Sn layer is intimately related to how much stress develops due to the IMC formation. Real-time scanning electron microscope (SEM)/focused ion beam (FIB) studies are described that illustrate the whisker/hillock growth process in detail. Whiskers are found to grow out of a single grain on the surface with little lateral growth while hillock growth is accompanied by extensive grain growth and crystallite rotation. Electron-backscattering detection (EBSD) shows the grain structure around where the whiskers/hillocks form, indicating that whiskers can grow out of pre-existing grains and do not require the nucleation of new grains. This has led to a picture in which stress builds up due to IMC growth and causes whiskers/hillocks to form at “weak grains”, i.e., grains that have a stress relaxation mechanism that becomes active at a lower stress than its neighbors. FEA (finite element analysis) calculations are used to simulate the evolving stress and whisker growth for several different mechanisms that may lead to “weak” grains.

Voltage-induced ion motion in conducting polymers (CPs) and the associated stress that evolves play an important role in determining the performance and lifetime of devices fabricated from these materials. In this work, a multi-beam optical stress sensor (MOSS) coupled with potential cycling was used to study films of polypyrrole (pPy) doped with indigo carmine (IC), a redox-active anion. We observe different magnitudes of stress in these films in different electrolytes (NaCl vs. HCl) due to the pH-dependent redox activity of IC. These results indicate that the redox activity of a dopant introduces an additional mode by which stress develops in films of CPs, which in turn could impact device performance.

Residual stress is a longstanding problem in thin film growth that affects film performance and limits the range of applicability. A better understanding of why this stress develops and how it depends on the growth conditions and material properties would improve our ability to predict it and control it. We describe a kinetic model we have developed that explains stress evolution in terms of a dynamic competition between tensile and compressive stress generation mechanisms. The balance between them changes as the microstructure evolves from isolated islands, through coalescence and into a uniform film, leading to the sequence of stress states observed during growth. The model further predicts that the stress in the steady-state depends on the dimensionless parameter D/LR where D is the diffusivity, R is the growth rate and L is the grain size. Calculations from the model in both the high and low mobility regime are compared with real-time measurements of stress using wafer curvature in different systems The discussion addresses limitations of the model and other issues that are not covered such as surface roughness and energetic particle deposition.

This chapter focuses on describing the primary driving forces and mechanisms causing the growth of whiskers and hillocks. It devotes to describing how the growth of intermetallic phase (IMC) leads to stress in the Sn layer. It also discusses the impact of Sn whiskers on reliability and the importance of suppressing their formation. The chapter discusses results from simultaneously measuring the growth of the IMC between the Cu and Sn whiskers, the stress in the Sn layer, and the whisker density to understand how these different processes interact. It describes experimental studies that are designed to identify the other processes that play a role in whiskers forming, for example, plasticity, diffusion‐mediated creep, and prevention of surface diffusion by the Sn oxide. The chapter also describes results from real‐time studies of whisker growth in the scanning electron microscope/focused ion beam milling (SEM/FIB) that show how whiskers and hillocks evolve in time.

Experiments, numerical simulations, and analytical calculations are used to calibrate a diffuse-interface Cahn-Hilliard model of Li-Sn thin film electrodes that experience a transformation from Sn to Li2Sn5. It is shown that a concentration-gradient dependent mobility can be used in the Cahn-Hilliard equation to give the interface a finite mobility and capture its nonequilibrium behavior. Comparing experiments and simulations, the free-energy of Li-Sn, diffusivity of Li in Sn and Li2Sn5, the exchange current density for the surface reaction, and the mobility of the Sn/Li2Sn5 interface are extracted. The implications of finite interface mobility for practical battery electrodes are discussed., Comment: 25 pages, 14 figures

Stress is believed to be the main driving force for whisker formation in Sn coatings on Cu. This suggests that whiskering can be suppressed by enhancing stress relaxation in the Sn layer, which is believed to be the reason why Sn-Pb alloys do not form whiskers. However, Pb is no longer acceptable for use in electronics manufacturing. As an alternative, we used pulsed plating to create Sn-Bi coatings with an equiaxed microstructure similar to that of Sn-Pb alloys. An optical wafer curvature technique was used to measure stress relaxation kinetics in Sn, Sn-Pb and Sn-Bi alloy thin films during thermal cycles. The results show that Sn-Bi films have significantly enhanced stress relaxation relative to pure Sn films. Comparison between Sn-Bi samples with equiaxed and columnar microstructures shows that both microstructure and alloy composition play a role in enhancing the stress relaxation.

Sn whisker and hillock formation is a reliability risk that has become increasingly important as the electronics industry has moved toward Pb-free manufacturing. To prevent them, we would like to understand what makes specific sites susceptible to deform into whiskers. We have used in situ scanning electron microscopy (SEM)/electron backscattering diffraction (EBSD) to monitor simultaneously the evolution of surface morphology and grain orientation in Sn surfaces in order to correlate whisker/hillock initiation with the underlying microstructure. Because rough films are difficult to measure with EBSD, we developed a unique procedure to make Sn-Cu samples with ultra-flat surfaces so that a large fraction of Sn grains can be indexed over repeated scans. We find that whiskers/hillocks grow from existing grains (not re-nucleated grains) with orientations close to (001). They often rotate from the as-deposited structure so that the orientation after growth does not indicate the orientation from which the whisker initiated. We measured the interface structure after removal of the Sn layer by chemical etching and found that there is no excessive accumulation of intermetallic compound around the whisker/hillock roots. Cross-sectional measurements revealed that a large fraction of the whiskers/hillocks have oblique boundaries underneath the surface, supporting the idea that these allow whiskers/hillocks to grow with lower stress.