|
Your Assemblies
The Main Characteristics Soldering is defined as a joining technology employing two base metals and a third metal or alloy to complete the union. In practical terms we nowadays usually encounter a Printed Wiring Board (PWB) and components with their respective metallization that are joined to make an electrical connection. There are many other applications even in the electronics industry that range from wire harnesses to connectors and lamps. However, the basic principle of soldering remains the same in all of these cases. When changing to ‘lead-free’ it is the product that in the final analysis dictates all restrictions. It is the cost (or rather price, which can be obtain in the market place) of the product and its reliability that enforces certain choices and the parameterization of processes. By examining those details we will be able to ensure that in future the product will meet the required standard – and hopefully the required price. We start with two very important aspects: the printed wiring board and the components.
The Printed Wiring Board The new EU Directive concerns the PWB in two ways. One acts directly because of the elimination of halogenated flame retardants in the laminate material and then, because lead is eliminated from the solder alloy, one will have to consider metallized surface finishes and solderability. The other is indirect and we will experience it because of impacts on process parameters. The increased melting point of alloys may force the user into more expensive material. 1. The Laminate Material Solving the first legal condition can only be accomplished by purchasing PWBs without the prohibited halogenated flame-retardants. The answer is found by directly specifying the laminate material to be free of such additives. What needs to be done is a direct cooperation with the board manufacturer. Board material that complies with the legislation is already on the market – using other (phosphorous-based) flame-retardants. Although they do not seem to be as efficient as the halogenated ones and they seem to suffer from some other ills (e. g. absorption and retention of humidity) they do meet the legal constraint. A close cooperation between the user and the manufacturer of the boards makes a lot of sense since, as we will see in a moment, the choice of alloy will affect the properties of the board as well. When choosing a lead-free solder there are really only two directions to go: either one chooses an alloy with a substantially lower melting point than eutectic SnPb or one with a substantially higher melting point. Should the user finally decide that all relevant restriction allow to go with a bismuth-solder then there are few changes necessary to the process. Naturally the halogenated flame-retardant has to be eliminated from the board material. However, from a point of view of thermal exposure, the situation has become better than even with SnPb37. The standard FR-4 and FR-3 materials would meet the temperature constraints as they have in lead-bearing solder processes. If, however, an SnAg, SnAgCu or even SnCu solder must be the choice, then much higher temperature are encountered in the process. Even pre-heat temperatures (flow and reflow)– dependent on the strategy pursued – will be moved upward. In this case traditional FR-4 material will be thermally stressed beyond its capabilities and will have to be replaced by a board that is halogen-retardant-free and that can ‘take the heat’. Typical glass transition temperatures [Tg] for FR-4 lie in the neighborhood of 140 ºC. This property of the laminate material will have to move upward by another 30 – 40 K if we intend to keep the chance of thermal damage to the board at a minimum. For such cases we now speak e. g. of ‘hi-temp’ FR-4. Such boards are offered by a number of suppliers but the actual properties vary greatly from one manufacturer to the next. In this area we feel that there is still a lack of standardization. 2. The Solder Mask Material Present-day solder mask material has its share of problems even under the conditions applied during lead-bearing soldering – in particular during flow soldering. Solder balls (or beading) is just one of these predicaments. If we have to raise the temperature even higher we may come to the end of their useful range. Deplorably, no better material seems to be coming over the horizon. The mid-term solution (until improved material enters the market) may be a more controlled application and curing (polymerization) of the substance. We do not want to be settled with additional inconveniences during our changeover to lead-free. 3. The Metallized Surface Finish For most applications – clearly not for all- ‘lead-free’ actually means ‘lead-free’. In those cases any coating on the PWB cannot contain lead. Several studies have shown that even minor traces of lead in the joint cause a phenomenon now referred to in the literature as ‘low melting phases’. (Some investigators also have related lead in the metallized coating to the phenomenon of ‘fillet lifting’ but in this case there is still some discussion going on about the actual cause.) ‘Low melting phases’ describes the occurrence of areas within the joint where (eutectic) pockets of alloys can be found that melt at substantially lower temperatures than the lead-free solder used. Such pockets will lower the reliability of the joint, particularly its resistance at elevated temperatures during use, and thus are not acceptable. We do want to emphasize, though, that for some solders, such as e. g. SnAgCu small amounts of lead may be ‘acceptable’. However, it is not yet clear how the limits for such small amounts can be defined safely. Furthermore, in this case again, there are obvious differences between flow and reflow soldering. Luckily we have a number of alternative surface finishes in use for the last number of years that are lead-free.
Progress has also been made with lead-free Hot Air Leveling. There is, however, the concern about the higher process temperature necessitated by the higher melting points of the new alloys. As the entire PWB enters into the solder the thermal stress to the board – particularly for multilayers – is considerably increased. The reduced wetting ability of such solders may also impact on the aggressiveness of process fluxes – hence potential residues. Since such information about our PWBs will be essential during the decision making process on the replacement alloys, we should already now collect all the relevant data. At the same time we should use the opportunity to approach our different board suppliers/manufacturers to discuss available alternatives and their price structure. For the technical details we would like to meet with technical personnel as there are very few sales people with whom one could solve some of the problems.
The Components Everyone seems to come to the conclusion that components will produce the greatest challenge when converting to lead-free technology. There are a number of reasons for this assessment and we will just mention a few: 1. Thermal sensitivity may represent the greatest obstacle and thus should be mentioned first. Many component manufacturers specify particular values for the maximum temperature that their component may be exposed to or certain dwell periods at process temperatures that are lower. Such values are derived either from the body material of the component or from internal properties such as solder joints or even from the thermally conductive adhesives that are used to heat-sink the chip. For other components it is the temperature gradient during temperature increase or decrease that is specified. We can assume that the reason for this precaution is the difference in thermal expansion factors of some of the materials used (e. g. ceramics and metal oxides in condensers). Whereas reflow emphasizes such aspects as gradient and dwell time, flow soldering concentrates frequently on the thermal jump from pre-heat into the liquid solder. For many components this critical parameter is limited to <100 K. For most of the lead-free alloys this limitation would require the increase of board and component temperature at the end of the pre-heat section and thus substantially higher pre-heat temperatures. Many component manufacturers have done a lot of work in this area and progress has been noted. Nevertheless, there are many that either neglected to develop technology to facilitate lead-free conversion or are lagging badly behind a schedule that users would like to have seen.
2. The effect of ‘popcorning’ is feared by industry since the discovery of early signs of such damage is difficult. ‘Vapor Dome’ and ‘internal delamination’ are neither called out by standard x-ray equipment nor are they recognized by visual inspection. As they are a substantial factor in ‘catastrophic failure’, their occurrence is closely monitored. Popcorning has occurred more frequently in lead-free applications. The limits for humidity and chemical absorption (usually by weight) that have been established for SnPb-soldering will have to be re-examined and maybe newly set at a lower level. JDEC may issue a new classification for lead-free applications. Plastic material that may absorb less humidity does exist, however the ones identified so far have shown other shortcomings, particularly during components extrusion processes.
3. Protective coatings of the leads – as in the case of the PWB and because of the same reasons – have to be specified as lead-free. Again there is a fair choice of finishes available that meet these demands but it may be extremely difficult to receive reliable information about them. Many components are not purchased directly from the manufacturer but rather through retailers and agents. These intermediate dealers either do not have the information or find it an unnecessary expense to obtain it for their customers. Traditionally purchasing has not specified this parameter and thus the user may ‘face a wall of silence’ with regard to this crucial piece of information. Naturally it is entirely possible to identify the coating in the laboratory. But for the number of different components and the fact that different deliveries may vary in this regard as well, the amount of work required could be enormous and costly. Hence this latter approach again appears to be unrealistic.
4. Whiskers seem to be the ‘spontaneous’ result of some crystalline growth. Tin and silver are notorious candidates for it. Lead, on the other hand, was used in the component industry to check and reduce ‘dendritic’ growth with considerable success. Whiskers are unwelcome as they lead to intermittent malfunctioning of the product. And this sporadic behavior makes them very difficult to trace. The last few years have seen a fair amount of effort going into identification of the phenomenon. In anticipation of the elimination of lead it is particularly the industry where such intermittent failure of product would be especially troublesome (air line industry, space, automotive telecommunication and other security and safety related branches) that has done the research. Investigations have shown that certain countermeasures (Ni-coating; reflowing coatings) can contribute to a minimization of the problem. What gives room for thought is the fact that the ideal solution for the time after lead has not been identified yet.
5. There are certain features of components that play a major role in the long-term reliability of the assembly. The flexibility of the leads is to name. Especially for component types such as SCCs {Solder Column Connectors] the increased stiffness of the replacement solders compared to the malleability of SnPb solders will come as an unwelcome property. Other component types are also affected, which has seen a major increase in the consideration and use of ‘undercoating’. Besides the positive effects of such measures (increased reliability) there are, however, negative ones as well, as may be expected: repair after application and cure of undercoats is practically impossible. Compensating for insufficient information on certain component properties is a problem on that we will spend a lot of valuable time.
Collection of Data We start with a listing of all the different assemblies that are manufactured. We add to them the number of boards assembled (per day or month or year). Then we identify the PWBs and their place of origin (not the European Asian supplier but manufacturer – if at all possible). The more sources we identify the greater the task to collect the necessary information and to control it later on. If at all possible, we add already at this point additional pieces of information e. g. certain differences in quality (number of complaints that had to be launched etc.) Once these specifics have been tabulated we try to add other details: type of laminate material, the type of solder mask, or final surface finish. Layout and design specifications may provide us with the details on component specification that belong to each one of the assemblies. Once we have their generic identification we can proceed with the details such as actual manufacturer(s). Obviously such a task can take on enormous proportions if the work has to start from practically zero. But the least we should have for the critical components is:
|