Novel Whisker Mitigating Composite Conformal Coat Assessment

Dr. Stephan Meschter | BAE Systems

WP-2213

Objective

An electronic assembly is created by integrating thousands of parts from multiple suppliers utilizing a host of circuit card manufacturing processes. With the lead elimination from electronics resulting from the European Union Reduction of Hazardous Substances (RoHS) legislation, many of the heritage aerospace and defense commercial – off – the – shelf solder materials with tin – lead have become obsolete. Most notably there has been increasing cost and schedule pressure to use commercially available pure tin part finishes and lead – free solders in Aerospace and Defense electronic systems.

Since these materials have an increased tendency to grow tin whiskers that can cause short circuit failures and decrease reliability, original equipment manufacturers are relying on multiple mitigation methods to manage the tin whisker risk. One of the main mitigations is the use of conformal coating to prevent tin whisker electrical shorting. Current widely used low cost coating spray processes used for humidity protection generally provide incomplete coverage and insufficient coating thickness for whisker mitigation or no coating at all on many tin surfaces. The most challenging areas are the back side of the leads, vertical surfaces and edges. There are vapor deposited coatings that provide complete coverage, but have yet to demonstrate complete whisker mitigation, are difficult to repair, and have a high manufacturing cost.

There continues to be increasing risk to Department of Defense (DoD) systems reliability and affordability due to the growing gap between consumer electronics designs and DoD electronics needs. The DoD benefits greatly by using consumer parts directly or using them with modification. Environmental compliance, cost, and miniaturization continually drive consumer designs, reducing unnecessary design margins, as long as the relatively short consumer warranties are satisfied. As a result, significant DoD risk continues to accumulate and evolve with lead-free, conversion to greener chemistry, and other design changes accompanying each generation of consumer electronic parts and materials.

The main objectives of project were to (1) develop and evaluate nanoparticle filled conformal coatings designed to provide long term whisker penetration resistance and coverage on tin rich metal surfaces prone to whisker growth in modern DoD electronic assemblies containing commercial lead-free consumer electronics, (2) utilize enhanced liquid coating application processes to improve coverage, (3) perform fundamental studies into the mechanisms by which conformal coatings provide tin whisker penetration resistance and inhibit nucleation/growth, and (4) evaluate coating reworkability. An objective added later was to evaluate the impact of long term simulated power cycling thermal cycling on coating whisker mitigation.

Back to Top

Technical Approach

Desired coating qualities that will exhibit superior whisker mitigation include complete coverage, high strength to buckle whiskers and/or high elongation prior to breaking to capture whiskers within the coatings. The key elements being evaluated are (1) coating and particle chemistry, (2) coating microstructure and mechanical properties enhancement, (3) coverage quality, (4) layered coating application, (5) coating rupture resistance on whisker test samples, and (6) reworkability. Coating microstructural evaluation, nanoindentation, tensile, and adhesion testing were used to characterize the coatings and select candidates for lead – free electronic assembly coating and testing. Finite element modeling was used in conjunction with the experimental observations and the measured coating properties to determine when coatings are expected to rupture from tin whisker or tin nodule growth.

Nanoparticle suspension and coating formulation: For the coating development activity, the base coating was selected to be a polyurethane based chemistry that meets the requirements of IPC-CC-830 and MIL-I-46058. The nanoparticle filled coating development began with making a suspension with functionalized nanoparticles which could covalently bond to the polyurethane matrix. A basic challenge with functionalized nanoparticles is preventing particle agglomeration and keeping them dispersed. After the suspension was mixed into the base material and cured, the nanoparticle surface modification bonding sites would tightly bind the particles to the resin matrix and have uniform dispersion without agglomeration. The initial evaluation of the suspension chemistry, particle size and loading was based on viscosity, transparency and pot life. Select suspensions were mixed with the base coating and cured films were fabricated for microstructural evaluation using the atomic force microscope, microtome sectioning, cryo-fracturing, scanning electron microscope examination, and transmission electron microscopy to assess nanoparticle agglomeration and resin adhesion. Then sample coupons were developed to facilitate nanoindentation and mechanical testing in order to determine the particle loading that yielded optimal tensile and elongation properties for whisker mitigation.

Coating application process enhancement and whisker mitigation evaluation: For the process evaluation effort, dip coating and spray coating process were used as the baseline coating application method since it provided good coverage and manufacturability. The experimental and analytical assessment evaluated various thicknesses and layer combinations of filled and unfilled coatings on idealized coupons and real electronic assemblies. The coverage and thickness on various metal features were assessed using scanning electron microscopy and cross-sectioning. In addition, whisker penetration resistance was assessed using high whisker propensity coupons and the lead – free assemblies developed in SERDP project  WP-1753 in traditional high temperature/high humidity and thermal cycling. These assemblies were also evaluated for coating reworkability after the layered coating tin whisker environmental high temperature high humidity testing.

Extended power cycling thermal cycling (PCTC): The approach used to evaluate the extended power cycling thermal cycling test assemblies from the SERDP  WP-1753 project was to first perform an inspection of the assemblies remaining at the of the original 1,797 cycle test (note that the longest whiskering samples were cross – sectioned). The initial inspection recorded whisker growth statistics on uncoated and coated assemblies as well as the condition of the conformal coating. Additional thermal cycling was performed until a total of 6,000 cycles was accumulated, then a final inspection was performed. The whisker statistics will contribute to the current database of over 120,000 whiskers counted in the total  WP-1753 project.

Back to Top

Results

Nanoparticle suspension and coating formulation: Two IPC-CC-830 conforming base coatings that had polyurethane isocyanate functionality were selected. The first was a moisture cure solvent thinned polyurethane (PU – PC18M from Henkel). The second was a moisture and ultraviolet (UV) dual cure 100% solids low volatile organic compound (low VOC) polyurethane acrylate (PUA – PC40-UMF from Henkel). The nanosilica and nanoalumina particle suspension were evaluated. The nanosilica suspension with a 20 nm particle size (XP2742 from Covestro) was superior in many respects to the nanoalumina suspension. It had good shelf life, particle size uniformity, particle isocyanate functionality, particle size uniformity, and particle dispersion. Upon mixing into the base coatings, the nanosilica had good particle dispersion with little agglomeration, good particle bonding to the polyurethane matrix, and it increase the coating strength. The disadvantage of using the XP2742 is that the addition of silica brings a certain amount of N3300 Hexamethylene diisocyanate, HDI, trimer along with it.  And while this HDI functionality easily reacts with both the PC18M and PC40-UMF systems, it brings physical properties that are different from the base formulations, which may or may not be desirable. When the XP2742 was mixed with the PC18M, some Dabco T 12 was needed to accelerate the cure slightly to maintain a 2 hour cure. This was also added to the PC18M and called PC18M-mod.

The nanoalumina particles were not functionalized, had poor pot life after mixing with the PU, a broad particle size distribution and upon mixing with the PU had particle agglomeration and gave a limited increase to the PU strength. Even though the nominal particle size was 40 nm according to the specification data, the average particle diameter was confirmed to be about 70 nm, with a range from 20 to 200 nm. During spray trials the nanoalumina had somewhat better coverage than the nanosilica or the unfilled PU. The nanoalumina supplier was a smaller company and after a few improvement attempts, was not interested in further developing the nanoalumina suspension to be isocyanate compatible. Mechanical testing identified the optimal suspension concentration to obtain an increase in strength without impacting the elongation significantly. Mechanical testing was used determine the optimum suspension concentration for the assembly spray coating evaluation.

Coating application process enhancement and whisker mitigation: A baseline dip/partial cure/ 50 micron spray was performed first with a thickness goal of 25 to 75 microns on a flat surface. The following coatings were evaluated: PC18M-mod, PC18M+20%XP2742 nanosilica, PC18M+20%X11102PMA nanoalumina, unfilled PC40-UMF, PC40-UMF+10%XP2742, and PC40-UMF+30%XP2742. The PC18M base coatings exhibited better coverage than the PC40-UMF.  Some, but not all, baseline spray coated PC18M assemblies had extensive coating bubbling after curing. The PC40-UMF based spray coatings had spreading and adhesion issues on the cured PC40-UMF dip coating, which were not resolvable by using plasma etching or reduced UV exposure partial curing. Whisker testing of the baseline process samples showed no whisker growth or coating rupture where the coating was greater than approximately three microns (the coating thickness the SEM beam can penetrate through and examine the underlying metal). Where the coating thickness was excessive, the coating cracked and/or pulled-way from the part features. Some thin areas had corrosion. The reference coatings, vacuum deposited Parylene C and the Parylene C with the Adpro plus adhesion promoter, had excellent uniform coverage and did not have any corrosion, whisker growth, or coating ruptures. The acrylic 1B31 reference coating test did grow whiskers where the coating was approximately three microns for the exposure times tested.

Although the PC18M+20%X11102PMA had the best coverage, based on suspension availability, the PC18M and the PC18M+20%XP2742 were selected for the enhanced layered coating development. The enhanced layered coating used a dip/partial cure/25 micron spray/partial cure/25 micron spray/cure combination with reduced individual spray layer thicknesses so the overall target thickness was still 25 to 75 microns on a flat surface. There was an issue with the dip coat viscosity increase resulting in excessive coating thickness and pooling around the part bodies. The layered spray coating surface was very rough probably due to a spray coating process that was too “dry”. The lead surface had craters where microbubbles appear to have burst and the surface did not flow and level out. The layered assembly coating coverage was better than the baseline dip/partial cure/spray/cure in that there were no vertical edges and fewer lead tips with thin coating. There was no whisker growth on the layered polyurethane samples.

A key finding was that processing coatings where the viscosity depends on the solvent concentration requires monitoring and tight controls. The PC18M had high viscosity dependence on solvent concentration. As the solvent concentration decreased, the PC18M viscosity became like molasses. During dip coating, the coating viscosity can increase with time due to solvent drag out causing excessive coating build up. During spray coating, either bubbles or a rough surface can result. However, this could be useful attribute to control coating thinning on vertical surfaces and increase corner thickness. Future application techniques might leverage methods to secure deposited material and reduce gravity and surface tension thinning effects.

Unique tin nodule microstructure evolution under coating and the importance of coating adhesion: During this project, several unique tin whisker/nodule growth observations were made. In the presence of a sufficiently thick rigid coating, either growth was suppressed or the tin formed nodules rather than whiskers. During the high temperature high humidity (HTHH) whisker growth testing of bright tin plating over copper after ~2,500 hours of 60°C/60%RH, the presence of a rigid coating suppressed whisker growth, but tin nodule growth developed instead. The following conditions were observed (1) no growth under the coating where the coating was ~100 micron thick and well adhered, (2) where the coating was three to 30 microns thick nodules were observed with larger Cu6Sn5 intermetallic (IMC) islands forming in the tin under the coating, and (3) the seven to 30 micron thick coating regions formed domes that stretched the coating, but where the coating was three microns the coating ruptured. Interestingly, the tin nodule dome diameter decreased as the coating thickness decreased which inspired energy based models for coating rupture and coating adhesion during nodule growth. The models, in combination with the measured coating mechanical properties, were in reasonable agreement with the rupture and dome formation observations.

Another set of observations related to coating adhesion. In the 20 micron polyurethane 85°C/85%RH whisker test, bright tin plated samples with high coating adhesion did not form nodule growth, but samples having lower adhesion regions formed nodules and developed tin/coating separation. Like the 60°C/60%RH samples, large Cu – Sn IMC islands formed near the coating surface, that were not attached to the Cu substrate. The higher magnification SEM imaging suggested that the Cu was transported from the Cu substrate interstitially through the tin and along the grain boundaries toward the coating interface and nucleating Cu6Sn5 IMC grains. Larger Cu6Sn5 IMC islands probably were formed from Cu6Sn5 particle coarsening during HTHH aging. Notably present was a complex layer of tin, Cu, and Cu – Sn IMC which formed where the original plated tin layer was located. In samples with greater aging, the coating redirected the complex tin growth laterally along the coating/tin delamination zone.

In some of the delaminated areas, tin was still attached to the coating indicating that coating adhesion to the tin depends not only on the coating/tin reaction layer but the strength of the tin near the coating, which could be impacted by sub – surface phenomena or increased stresses. The stresses in the coating area increased tin oxide layer growth and the accumulation of Cu6Sn5 IMC near the coating and increasing interfacial stress due the volumetric increase these compounds compared to tin.

Systematic annealing studies of tin and adhesion promoter (e.g., silane treatment) may be useful to assess the nature of this relationship between the adhesion and the tin whisker/nodule growth.  

Layered coating rework: Assemblies from the enhanced layered coating whisker experiments were used to evaluate rework. The old coating could be removed with standard shop practices. Then a low viscosity thinned PC18M coating was syringe applied, cured, and then a higher viscosity PU coating was applied and cured. Cross-sections of assemblies showed that the reworked coating had good adhesion to the new replacement part and to the old coating adjacent to the rework site. The adhesion was confirmed on a flat section of a board with a tape test. The brush coating applied excessive material and cracking was observed after environmental testing. The coating materials have the basic adhesion characteristics for successful rework, but further work is needed to refine the method.

Coating rupture modeling: A unique energy based coating rupture model was developed with a combination of classical analysis and finite element modeling. The model predicted that larger nodules are more likely to rupture the coating layer than smaller diameter whiskers. A coating delamination model versus nodule diameter and coating thickness was also developed. The model results were in reasonable agreement with observed coating rupture and delamination by nodules.

Extended simulated power cycling thermal cycling of the WP1753 samples: The coating on the WP1753 cards was an ultraviolet and humidity dual cure polyurethane acrylate (Humiseal UV40) applied with an automated spray coating machine. After a total of 6,000 power cycling thermal cycling (PCTC) cycles, no whiskers penetrated coating where it was thicker than approximately three microns (e.g the approximate coating thickness where the SEM beam cannot penetrate through the coating and image the underlying solder). Although the coating obscured the majority of the tin surface from SEM examination, it was thin enough (less than approximately three microns) on the knee and at the toe to examine the metal. The solder areas with thin coating did have some whisker nucleation on the Alloy-42 (Fe42Ni) leads. Some coating cracking was also observed. On assemblies without coating, the whisker length did increase with additional cycling, but still were less than ~40 microns. A number of unique thin wide tin growths were observed. These appeared mainly between the lead-free solder and the copper board pads on both copper and Alloy-42 lead material terminations. While these growths were short, they could impact coating integrity.

Back to Top

Benefits

The knowledge developed in the current project on whisker mitigation will result in increased aerospace and defense electronic assembly reliability while hopefully preventing overly conservative design practices. The approach used to characterize, analyze and characterize these coatings can be extended to any coating intended for tin whisker mitigation applications. Including a particle enhanced coating to the tin whisker mitigation tool box will be useful because if the particles are used for strengthening, thinner layers could be used. Alternatively, the particles can be used to alter the liquid coating rheology and improve corner coverage, as was observed in nanoalumina case. Layering with a filled layer over an unfilled layer could provide optimal adhesion to the tin while providing a stronger more penetration resistant top coat. In addition, by alternating nanoparticle filled and unfilled layers, the nanoparticles can make it easier to validate the thickness of the various layers applied. A list of publications and briefings generated from this project are provided in Appendix A. Over the five years of study, many references were collected and compiled in the reference section here and in the appendices.

The results of this work have been briefed to both the IPC and SAE committees responsible for lead – free risk management and conformal coating of Aerospace and Defense electronics. Recommendations were made for improvement to GEIA-STD-0005-2, Tin whisker mitigation for Aerospace and High Performance Electronics and discussion have been underway on potential changes to IPC-CC-830 and IPC J-STD-001 when conformal coatings are used for whisker mitigation.

The results of this work were also flowed into ESTCP –  WP–201573-T2 project training webinars to help educate the program management and systems communities and available on line for future use. This project highlights a means of effective systems risk mitigation by having research in place to understand how and if consumer parts changes impact DoD systems reliability. The immediate reflex is to flow down of reliability requirements all the way down the supply chain to make sure “high reliability” is obtained. Unfortunately these requirements end up at suppliers who are furthest from the DoD integration tier, which are often the weakest links in the supply chain. The suppliers in the lower supply chain levels have narrow margins with little or no R&D money. The main way R&D happens in these communities is leveraged investments and consortia activities. Consortia often have foreign participation, making it more difficult for the DoD interests satisfied. Achieving year over year product cost reductions can negatively impact long term reliability without significant research. The DoD and all levels of its supplier chain must actively collaborate on research to evaluate new consumer electronic parts and materials to ensure proper design and manufacturing standards are in place. The methods developed here can be applied to evaluate and monitor fielded assemblies containing lead-free parts and materials for continued reliability improvement.

Back to Top

Points of Contact

Principal Investigator

Dr. Stephan Meschter

BAE Systems

Phone: 607-429-8828

Fax: 607-755-6615

Share