Estimated reading time: 7 minutes

The A-Z of Lead-Free Reliability: SMART Group workshop at NPL

The National Physical Laboratory (NPL), located in Teddington on the outskirts of London, UK, is a government funded research organisation carrying out an enormous range of materials metrology projects. Dr. Chris Hunt's Electronics Interconnections group specialises in conducting collaborative projects with the industry and the group is recognised world-wide for their work in characterising and measuring the reliability of printed circuit materials and assemblies. The group has wide-ranging interests in the technology of lead-free soldering and the development of new test methods, guidelines and procedures.

The NPL recently hosted a workshop entitled "A-Z of Lead-Free Reliability," organised under the auspices of SMART Group, which provided the opportunity for members of Hunt's team to update a full-house audience of electronics industry professionals on a series of current NPL investigations and studies, many of which were Joint Industry Projects with external partners.

Dr. Davide Di Maio gave two presentations, the first being an update on the collaborative project aimed at developing a monitoring technique for quantifying the rate of copper dissolution by molten solder--which is known to be exacerbated under lead-free processing conditions--to further the understanding of the influence of copper dissolution on the reliability of lead-free solder joints, and to provide a best-practice guide. The investigations considered the effects of temperatures and flow rates for a range of solder alloys on copper substrates with different grain structures and surface finishes, with reference to copper-tin metallurgy, diffusion kinetics and fluid dynamics. A meaningful test-piece had been designed, together with an electrical method to determine rate of dissolution under conditions representative of real industrial processes. There was evidence of some extremely complex interactions between parameters, certain of which remained to be clarified.

Later in the day, Dr. Maio's second presentation summarized an investigation of factors influencing the occurrence of "tin pest," and how they could be measured. Tin pest is a phenomenon resulting from allotropic transformation of the crystal structure of tin from its tetragonal beta-phase, the shiny and ductile material which exists above 13°C, to the dull and brittle cubic alpha phase at lower temperatures. By seeding with cadmium telluride, the transformation, which normally happens very slowly, had been accelerated and captured by time-lapse photography, enabling a better understanding of the layer-by-layer "peeling" mechanism. (A remarkable video showing the effect can be seen at http://uk.youtube.com/watch?v=vVtCPw7RsW4). An electrical resistance test had been developed to monitor the transformation, which was particularly useful for studying tin alloys. It had been observed that, compared with pure tin, the propensity for transformation decreased successively in the order tin-silver, tin-copper, tin-indium, tin-nickel, tin-zinc, tin-lead and tin-bismuth.

Ling Zou's first paper described three electrochemical techniques for predicting circuit reliability in the presence of contamination from fluxes and process residues. Potential-dynamic scanning was a quantitative method for measuring metal corrosion, by comparing corrosion currents and polarization curves. The surface insulation resistance technique used DC voltage to measure leakage current between two metal electrodes on a substrate surface, resulting from contamination. It gave a direct measurement of circuit reliability and could be used to qualify production processes.

AC impedance measurement over a wide frequency range provided useful information on conduction processes, and could be used as the basis of a non-destructive test for predicting dendrite formation and corrosion for different flux residues.

Zou's second topic was an investigation of the performance of conformal coatings, using a technique based on surface insulation resistance measurement, developed by NPL and now incorporated into IEC and IPC standards. The test vehicle was a PCB populated with a wide range of component packages. Coatings included polyurethane, silicone, epoxy fluoroacrylate and various acrylics, solvent based and water-based. Sea water, flux, acid rain, hydraulic fluids, de-icing fluids and surfactants were used to contaminate the assemblies. The ability to form a uniform coating was critically important; the protection was seen to fail at weak points, particularly the sharp corners of high-aspect-ratio conductors, and corrosion could initiate from these points. Regarding the ability of a coating to protect an assembly from water, it had been observed that conformal coatings were not completely impermeable to moisture, but could offer protection against water droplets and condensation.

Dr. Chris Hunt reported his work on the measurement and prediction of solder joint reliability, particularly in respect of the significant differences in mechanical characteristics of lead-free solders compared with tin-lead. Modelling, based on analytical methods, constitutive law, fatigue law and damage mechanics, was chosen as the preferred methodology. Although some simplifications had to be accepted, modelling could still account for assembly and environmental conditions, but was critically dependent upon good materials data. NPL had developed a testing machine to measure stress and strain in a single solder joint engineered to a precise and repeatable geometry. Dr. Hunt presented creep, fatigue and relaxation figures and damage model results for a SAC305 alloy, which could be related directly to effects within actual joints in printed circuit assemblies.

The second project reported by Dr. Hunt was evaluating thermoplastic substrate materials as potentially recyclable alternatives to FR4, in the light of WEEE legislation. Liquid crystal polymer, polyether-imide and polyether-ether ketone materials all offered possibilities as printed circuit substrates. Subtractive, additive and direct-write technologies were being investigated as feasible routes towards circuitisation. Work was proceeding to determine the fitness-for-purpose of various material and processing combinations. Assemblies had been built using isotropic conductive adhesives, and subjected to stress-screening, damp heat and thermal cycling. The new systems had performed relatively well under harsh testing conditions. A particular failure mechanism was the deterioration of the interface between conductive adhesive and component terminations where the latter were had a tin finish, although it was made clear that this was not a substrate issue, and the effect would be similar with traditional FR4 circuits.

Martin Wickham had three projects to discuss. The first was on substrate reliability with a study of hole size and board thickness effects on plated-through-hole performance in thermal cycling, using a series of different FR4 materials--standard and high Tg, dicy and phenolic cured--and 10-layer test boards 2.5 mm thick. Barrel cracking was the predominant failure mechanism. In a separate experiment, high Tg phenolic-cured FR4 was compared with polyimide. General conclusions were that reliability could be improved by reducing via aspect ratio, reducing peak reflow temperature, reducing board thickness, reducing the number of thermal excursions and increasing the Tg of the substrate material. Polyimide outperformed high-Tg phenolic-cured FR4 as a consequence of its glass transition being above reflow temperature, resulting in significantly lower Z-axis expansion.

Wickham's next topic was an investigation of the dimensional stability of components at the high processing temperatures associated with lead-free soldering, where warpage could result in dry joints and consequent reliability problems. Using a 64-way surface-mount connector, a 1156-pin BGA, and certain electrolytic and polyester capacitors as examples, and he presented case-studies of Z-axis distortions using laser profiler, dilatometer and dynamic mechanical analysis as measurement techniques. Dimensional stability of the connector appeared to be related to the design of the moulding tool and significant stress relaxation was observed during the initial heating cycle. Distortion of aluminium-electrolytic and polyester capacitors usually occurred during initial reflow, even at low temperatures. Tin-lead-rated and SAC-rated BGAs behaved differently during heating, and work was proceeding to examine the mechanical effects of thermal distortion of BGA packages on the integrity of solder joints.

Wickham concluded the workshop with details of work on mitigating tin whisker growth by conformal coating. The introduction of RoHS had led to component manufacturers changing from tin-lead to pure tin finishes on component terminations. The propensity of pure tin to generate whiskers was well known and although new chemistries had been developed that greatly reduced the risk, a guaranteed "whisker-free" finish had yet to found. (Some examples of the disastrous consequences of tin whiskers can be seen on the NASA Web site http://nepp.nasa.gov/whisker/.)

Work at NPL had demonstrated that conformal coating could suppress the formation of whiskers. An electroplating process known to give a tin deposit particularly prone to whiskering had been used to prepare copper test-pieces which could be assembled into a parallel-plate array and monitored electrically for the propagation of whiskers. Various combinations of uncoated and conformal-coated tin finishes had been studied. Paraxylene coating had proved to be most effective at preventing whisker growth. Certain acrylic coatings would prevent whiskers growing from within, but could be penetrated by whiskers from an external source.

The manufacture of high reliability lead-free circuit assemblies presents many challenges, and an understanding of the properties and limitations of materials is a key factor in defining quality. This SMART Group workshop addressed a range of relevant issues, provided guidance to designers and engineers, shared a wealth of information and provided the stimulus for much interactive discussion.


To contact Pete Starkey, click here.