6th International Symposium on Tin Whiskers

Citation

Goosey, M. (2013), "6th International Symposium on Tin Whiskers", Circuit World, Vol. 39 No. 2. https://doi.org/10.1108/cw.2013.21739baa.015

6th International Symposium on Tin Whiskers

Article Type: Conferences and exhibitions From: Circuit World, Volume 39, Issue 2

Henry Ford College, Loughborough University, UK27 and 28 November 2012

At the end of November 2012, Loughborough University’s Henry Ford College was host to the 6th International Symposium on Tin Whiskers. The Symposium was opened with a welcome by Dr Geoff Wilcox from Loughborough University and he began by welcoming Mike Osterman from CALCE, Maryland. Geoff gave an overview of Loughborough University, which had been around in various forms for over 100 years. The symposium had been organised by the university’s department of materials which currently had around 80 PhD students. Mike Osterman then gave the formal symposium opening remarks. He highlighted how the move to lead-free assembly had generated new interest and concerns about tin whiskers. He also outlined the work being undertaken within CALCE as part of their broader interests in electronics reliability. Mike highlighted how the RoHS Directive was evolving to include new categories of equipment that would have to be lead-free and how there were major concerns about tin whiskers across a wide range of industrial sectors. He then gave an introduction to tin whiskers and how they caused problems via their conductivity and the number of failure mechanisms that were possible. There were a large number of factors influencing whisker growth.

The first paper was given by Bob Gregory of Rolls Royce and this was on tin whisker risk management for high-reliability systems. He began by giving an overview of the work of the PERM group – this was the lead-free electronics risk management consortium and it had produced a number of documents on lead-free electronics reliability and risk management. He then stated that industry best practice was to implement formal tin control plans and that the aerospace and defence sector used GEIA-DTD-0005-2 which was initially issued in 2006. There were various levels of tin control that needed to be employed and these varied by product type and the degree of reliability required. The use of component level mitigations did not count toward required mitigations, but preference must be given to lower risk finishes during the component selection process. Bob also stated that the use of barriers was a mitigation strategy rather than a cure. However, replacing tin with tin-lead could provide a cure and one such approach was to apply a hot tin-lead finish. Direct verification of finish composition was also vitally important and the most common tool for non-destructive analysis of plating was X-ray fluorescence. Energy dispersive spectroscopy in conjunction with scanning electron microscopy was also used.

Mark Ashworth from Loughborough University then presented a paper entitled “manufacturing and in service tin whisker mitigation strategies; the co-deposition of particulates and conformal coating”. Mark’s presentation covered work undertaken in the IeMRC funded Whiskermit project which included a number of industrial partners. He began by covering the co-deposition of particulates and the ones used in this work had been silver and silica. Two sizes had been selected and the aim was to evaluate the effect of particulate size and type on whisker formation. Deposition conditions had also been evaluated and this included the use of pulse plating. Pulse plating and reverse pulse plating had been used with silica deposition and these gave a more uniform distribution of the particles within the deposit. The influence of both 0.5 and 5.0 μm silica particles on whisker formation was also evaluated and compared to pure tin samples. It was found that the incorporation of silica particles into the deposit actually accelerated whisker growth. The second part of the presentation was on the use of conformal coatings. A number of commercially available organic coatings had been benchmarked on bright tin deposited on brass substrates. Examples of tin whiskers penetrating acrylic conformal coatings of different thicknesses were shown. Sometimes the whiskers were actually covered by the coating material. Further work was planned where particulates would be incorporated into the coating materials. To date, no whiskers had been observed when silicone and UV cured materials were used. The effect of electrochemically oxidised films on whisker growth had also been investigated and this had reduced whisker growth compared to samples with naturally formed passive oxides.

Barbara Horvth from Budapest University then presented work on the development of shape variants of whisker formations on bright tin layers. She began by confirming that electroplated tin had become a commonly used component finish, especially since the implementation of the RoHS Directive. Tin whiskering was a major risk when electroplating tin because the smaller grain size in brighter tin could cause greater residual stresses. Examples of a range of whiskers that had grown on electroplated tin (8-10 μm) on copper substrates were shown and these had been divided into different whisker types such as filaments, hillocks and nodules. The steps defining whisker growth in bright tin samples were then discussed in detail for each type of whisker. In the case of hillock/nodule fragments, additional impurity atoms created sub-grains with large abnormal grains. The mass migration of copper into whiskers was also reported and this was related to the development of voids beneath the whisker grains. Voids appeared in the Cu₆Sn₅ intermetallic due to the diffusion of tin into the forming whiskers. The copper, and tin, then moved up along the grain boundaries of the tin into the whiskers.

The second session began with a paper on “tin whisker and hillock growth via grain boundary sliding coupled with shear induced grain boundary migration” and was given by Pylin Sarobol from Purdue University. This presentation was a summary of PhD research work on tin whisker growth from lead-free solder and electroplated films. Whiskers and hillocks had been studied and they both grew from surface grains as a mechanism of stress relaxation. The mechanism of compressive stress formation and its relief were then described and sources of stress included the plating process, the formation of intermetallics and expansion coefficient mismatches. The origin of surface grains and how some became defects was discussed in terms of grain structure and homogeneity. A key factor influencing whisker growth was thought to be grain boundary sliding and a proposed steady state growth model was reviewed. The model was used to explain the formation of different types of whiskers. Hillock growth was then discussed in terms of grain boundary formation, shear and migration. Grain orientation effects were reviewed in terms of the hypothesis that crystallographic orientation had a significant role in the nucleation and growth of whiskers and hillocks. A key recommendation for mitigating whisker formation was to engineer a texture with the lowest density of high-elastic strain/strain energy density locations.

The next paper was on the thermal cycling of whiskers and the influence of atmosphere and this was given by Katsuaki Suganuma from Osaka University. It was reported that there was a difference in whisker formation after 500 thermal cycles between samples in air and those in a vacuum. Oxidation had an impact on whisker morphology and examples of samples exposed to 2,000 cycles in air were presented and the presence and influence of tin oxide discussed. A possible mechanism for whisker growth during thermal cycling was then described and the influence of the tin oxide on the “winding” growth of the whisker detailed. The influence of plating thickness on whiskers had also been investigated. With thinner coatings (2 μm) the whiskers were longer and thinner than those growing from a thicker deposit (5 μm) after 1,000 cycles between −20°C and 80°C in air. The thicker deposits also had a larger grain size (1.94 vs 4.69 μm) and grain size coincided well with whisker width for both air and vacuum grown samples. It was also noted that grain boundary cracking was more severe in air. Another interesting finding was that whisker growth in vacua was faster than in air and this clearly had ramifications for space applications.

Barrie Dunne from the European Space agency then began the final presentation of the morning and this was on increased shorting in tin whiskers due to electric fields and contact pressure. He started by showing some early examples of tin whiskers that had been found in the 1970s and some more examples found on relays used in aerospace applications. A lead-free control plan working group had been established and ESA guidelines had recently been published. Martin Wickham then continued the presentation and he began by giving an overview of NPL’s activities in this area. A special plating chemistry had been used that enabled the relatively rapid formation of whiskers. Whiskers had been harvested and their electrical properties determined. It had been found that larger whiskers had higher resistivities due to defects. Whiskers were known to deform in an electric field and thus the likelihood of shorting between adjacent conductors increased if the whisker moved under the influence of an electric field. The contact resistances of whiskers had been found to be highly variable due to variations in contact areas, contact angles, and the breakdown voltages of whisker oxides.

Following a networking lunch, the first presentation of the afternoon session was entitled whisker growth in low and high-stress environments: metallurgical and statistical analysis. Polina Snugovsky from Celestica, Toronto gave part 1 of the presentation. She began by outlining the typical occurrences of whiskers in lead-free solder joints. The rougher surfaces of lead-free solder joints were an important factor in whisker formation. Other well-known factors were the presence of compressive stress in the tin and impurities. The whisker formation model that had been proposed by Paul Vianco was then reviewed and this stated that whisker growth was the result of dynamic recrystallization. A specific four year tin whisker testing and modelling research project (WP1753) was then described and the testing programme outlined. This had looked at wide number of variables including contamination, different solders and board designs and a range of accelerated testing conditions. It was reported that the use of clean components, boards and post-assembly cleaning significantly reduced the propensity for whisker formation in lead-free products. For a given tin grain size, it had been found that, with tin, stresses that were either too high or too low, whiskers were not expected to grow. It had also been found that RoHS compliant tin plated components may have microstructural characteristics that could affect whisker formation and these included uneven tin plating, voids in the plating, exposed lead materials and contamination. The results of accelerated testing at 85°C/85 per cent RH for 1,000 h were also reported and it had been found that the whiskers were shorter on clean assemblies. Another key conclusion from this work was that whisker growth was particularly pronounced on SAC305 alloy fillet edges. In summary, Polina presented a huge amount of information and data related to the dependence of whisker growth on component type and lead material, component defects and solder joint microstructure. The second part of this presentation was then given by Stephan Meschter from BAE Systems in Endicott, New York. This part of the presentation had a focus on the statistical analysis aspects of the project. It was shown that the most significant factor was the lead material and this was followed by the contamination level. Many whiskers grew from SAC305 alloy solder joints. Alloy 42 leads with both pre-soldering chlorine contaminated parts and post-soldering contaminated assemblies grew the longest whiskers under power cycling conditions. Also, in simulated power cycling, Alloy 42 had greater whiskering than copper and rework flux residues also encouraged whisker growth. In high-temperature/high-humidity conditions the longest whiskers grew from solder joints with copper-based alloy leads and corrosion was again contamination driven.

Mark Walmsley from Micross Components then gave a talk entitled “A solution for tin whiskers by hot solder dipping”. The main theme of the presentation was to give an overview of the commissioning, testing and qualification of an automated hot solder dipping process for non-hermetically sealed components. A seven axis multifunction robot-based automatic hot solder dipping process had been developed which met the requirements of GEIA. The equipment offered control of depth, dwell, entry and exit speed, solder angle and exit angle. Two projects related to this process had been undertaken with the University of Greenwich and with NPL. The Greenwich project undertook a modelling assessment of the impact of the hot solder dip process and was carried out by Stoyan Stovanov. The study was able to confirm that the refinishing process did not have a significant impact on the electrical performance of the ten different types of components tested. The work undertaken with NPL was to develop a methodology for process qualification. All components were micro-sectioned for optical analysis to determine coverage and solder and intermetallic thicknesses. The overall findings were that solderability was equal to or better than the original components and ball shear test results for BGAs were also acceptable. Additionally, thermally cycled solder joint reliability was improved for re-terminated components compared to tin finished originals.

Mike Swanick from Rolls Royce then presented “a cure from tin whiskers by re-passivation”. He began by giving an overview of the problems associated with pure tin finishes. It was important to remove all of the original tin and to provide 100 per cent coverage of the new solder while ensuring that there was no damage to the parts. Components needed to be dipped for around three seconds to ensure that there was a full exchange of the tin coating. Qualification was by package type and by electronic technology type for the specific package. The qualification had two parts; package damage and reliability assessments which included cycling for aerospace and steady sate high-temperature testing for submarine systems. Testing was carried out with devices biased in simple circuits.

The final paper of the first day of the symposium was given by Michael Osterman from CALCE, University of Maryland, and he talked about the “effectiveness of photosintering in mitigating tin whisker formation”. Photosintering used a high-energy pulse from a flash lamp to rapidly heat and melt the device surface. Due to the rapid nature of the exposure the penetration of the heat was very shallow and limited. A typical pulse would be for 0.5 ms and 1.4 MW. For a thin electroplated tin layer on copper the photosintering induced an even IMC layer. The ability of photosintering to mitigate whisker formation had been evaluated via whisker growth studies and XRD stress measurements. The test coupons were crested using an acid electroplating process and the subsequent whiskers formed were characterised by scanning the coupon surfaces at 400x magnification and counting the whiskers. Accelerated testing data for treated samples was shown and compared. X-ray diffraction had been used to measure the lattice spacing, since this could be used to give an indication of stress in the tin. Interestingly, the stress after annealing was low but after about a year it was highly compressive. One of the key findings from this work was that photosintering at elevated temperatures indicated a long term reduction in film stress but, without elevated temperatures, it did not reduce the compressive stress in tin films plated on copper.

The second day of the symposium began with a presentation on electrical aspects of zinc and tin metal whisker induced failures in electronic equipment and this was also given by Mike Osterman from CALCE. Mike began by giving an overview of tin and zinc whiskers and he described where they could be encountered. Zinc whiskers were similar to tin whiskers, as were those from cadmium. Zinc whiskers were often found on the zinc coated raised floor tiles used in computer data centres. The electrical failures caused by whiskers were then detailed and the electrical breakdown of whiskers was related to the thickness of the dielectric layer, contact resistance and surface roughness of the whiskers. The results of work to measure the breakdown voltages of whiskers were detailed and zinc whiskers were found to have higher breakdown voltages than tin whiskers. The effect of contact force on breakdown voltage had also been measured by using harvested tin and zinc whiskers attached to a copper substrate. The ability of tin whiskers to induce a vapour arc was also discussed and it had been found that, at reduced atmospheric pressure, the voltage required to strike an arc was lower than in a normal air atmosphere. Experiments had been conducted to determine the conditions required to strike an arc in both tin and zinc whiskers. It was found that the resistance of the test specimen whisker was the strongest indicator of whether or not a vapour arc was likely to occur. An arc metric was proposed that could be used to predict the likelihood of an arc forming and this held for both tin and zinc whiskers. Conditions were described under which it was possible to strike and sustain an arc. The proposed arc metric and arc threshold information would be useful for providing a guide for circuit designers and to minimise the vapour arc propensity of whiskers.

The second paper of the day was given by Stephan Meschter from BAE Systems, Endicott, USA and was on novel nanoparticle enhanced conformal coatings for whisker mitigation. The aim of the project work reported was to develop and assess nanoparticle filled conformal coatings designed to improve whisker penetration resistance and coverage on tin-rich metal surfaces prone to whiskering. The coating materials used were based on polyurethane and polyurethane-acrylate resins, because they possessed a range of required properties such as jet fuel resistance, coupled with toughness and flexibility. Methods had been developed to assess whisker penetration and these had included the addition of rare earth elements to tin in order to facilitate whisker growth. A key part of the project had been the development of the nanoparticle suspensions and the integration of the nanoparticles within each coating solution. An example shown had a 12 per cent nanosilica content, homogenously dispersed in a polyurethane resin. Coatings had been characterised using dynamic nanoindentation and nanoscratch techniques as well as by a range of other standard testing techniques. The project had started in 2012 and was due to continue until 2015.

The third paper of the first session was given by Mark Ashworth from Loughborough University, UK and was on the mitigation of tin whiskers by optimisation of electroplating process methodologies. This was a further report on the IeMRC supported Whiskermit project. The effect of current density on metal deposit characteristics had been evaluated and images were shown of coating surfaces deposited using current densities from 5 to 50 mA.cm⁻². As current density increased, there was a reduction in whisker length and a decrease in whisker density, but there was an increase in the size of the eruptions that did form. It was also found that there was a reduction in whisker growth as the thickness of the deposit increased. Whisker growth on copper substrates was much lower than on brass. The effect of temperature and humidity on whisker formation from tin on brass had also been studied. As part of the project, pulse plating had been investigated as a potential whisker mitigation strategy. The technique offered the possibility of manipulating the grain boundaries and orientation of the tin deposit to promote improved whisker resistance. The results of a study of the effect of duty cycle and pulse frequency on deposit properties were reported. It was found that whisker growth on pulse plated deposits increased as the duty cycle was reduced and pulse frequency increased. Whisker growth on initial pulse plated samples was greater than for those plated using DC. However, whisker mitigation could be achieved by using a combination of increased deposit thickness and high-current density. Low whisker growth rates had been achieved for tin deposits on copper, even after 4,000 h storage at 55°C and 85%RH.

The penultimate presentation of the symposium was given by Jacob Wang, who reported more of the work that had been undertaken at Loughborough University on tin whisker growth as a part of the IeMRC supported Whiskermit project. The focus was on an investigation into the role of lead as a suppressant for tin whisker growth and work to help guide research for effective replacements for lead by modifying the tin deposition process and by substituting other elements for lead. The ability of lead additions to mitigate against tin whisker growth had been established in the 1950s but the underlying mechanisms were still not fully understood. The effect of plating bath composition on deposit lead content had been studied, along with an assessment of the effect of lead co-deposition on deposit grain structure. It had been found that the grain structure transitioned from a columnar to an equiaxed structure which was considered better for suppressing tin whisker formation. The addition of lead also resulted in a more uniform and planar IMC formation, which had been reported to have a lower strain generating ability and which enabled more uniform creep. The presence of lead also resulted in a deposit texture that had rotated towards a (220) orientation that was less prone to generate whiskers.

The final presentation of the symposium was given by George Milad of Uyemura, Southington, Connecticut, USA and was on work carried out by his company in Japan on the elimination of whiskers from electroplated tin. George began by reviewing the numerous factors that led to the formation of tin whiskers. The mechanism for tin whisker formation under ambient conditions was considered to be the most serious one. An approach for reducing the likelihood of tin whisker formation was then detailed and this included control of the tin crystal structure and controlling the substrate surface roughness in order to decrease the internal stress. Work had been carried out on tin plated copper lead-frames and the surface roughness measured using laser profilometry. The influence of surface roughness on whisker formation had been studied and it had been found that tin deposits on rough copper had reduced whisker formation at ambient conditions. This was because the rougher copper formed a uniform IMC layer which prevented the localisation of internal stresses. The crystal structure of the deposited tin had also been studied in terms of its influence on whisker formation. Three different tin plating solutions had been used to give different types of tin deposit crystal structures. Whisker length vs storage time data for tin films from these three solutions was shown. Compared with large grain size tin deposits, those with smaller grain sizes had reduced tin whisker formation under ambient conditions. Also, tin deposits which had crystal structures similar to those found in tin-lead deposits effectively restrained tin whisker formation. The crystal structure in the tin deposit was found to be one of the most important factors for restraining tin whisker formation. One of the solutions studied had been shown to effectively prevent whisker formation; this was the one that deposited tin with a columnar and equiaxed mixed crystal structure rather than the conventional columnar structure alone. Data for a number of accelerated testing/storage conditions was used to confirm these findings.

This symposium had an excellent combination of presentations which covered all aspects of tin whiskers from fundamental research to understand better the science and mechanisms of whisker formation, through whisker prevention to mitigation strategies. What was clear from the symposium was that there was a large number of factors, often with complex interactions, that played an important role in the formation of tin whiskers. Although much good work has been undertaken over many years to develop the knowledge base and to solve the problem of tin whiskering, it is clear that there still remain many issues to be resolved. There will thus be a need for this work to continue for some time to come and there will certainly be a need for future meetings of this kind. Overall, this was an excellent international symposium with high-quality presentations detailing progress in the global effort to overcome the problem of tin whiskers.

Martin Goosey

November 2012