Under the general umbrella of “Enabling processes” we have brought together seven topics, each of which has a range of applications from component packaging through to box build. As well as describing the processes, the materials and the options for each, we have also included a short list of applications, although these should not be regarded as exhaustive.
As we explain in the video version of the introduction, we decided to group together selected “enabling processes”, rather than fragmenting our treatment of key topics such as printing by splitting up the material over a number of different Units. This approach:
This last is highly significant. After all, the choice of process has implications both for unit cost and set-up costs, and for the investment needed in tooling and jigs. In many cases, more than one process would be possible, and the choice will be affected by the materials chosen, the volume of manufacture, the application of the end product, and of course by the availability of the process. This final consideration may be a limitation whether one is carrying out a process in-house or contracting it to a manufacturing partner – Is the correct equipment available? Is there sufficient capacity? Are there sufficient trained personnel?
[ back to top ]
The term patterning applies to many activities involved in electronics manufacture. Before reading further, make a list of the tasks that you can remember where patterning processes of any kind might be used. Look at some typical products, and think as widely as you can, before referring to our list.
Having made a list a wide range of different applications, review the requirements for the patterning processes, and try to identify the key considerations that affect the choice of process for a particular application. Do some lateral thinking before looking at our answer and then reading further.
From these general considerations, we can select appropriate approaches from the range of different methods available. But note that we do not necessarily have to employ a patterning process just because it happens to have been used in the past – when you come to look at advanced manufacturing methods, you will find many occasions when perceived wisdom has been “turned on its head”.
The approaches that we can use fall into broad categories, ‘negative’ and ‘positive’ processes. In the first, the substrate surface is removed selectively; in the latter, materials are added to the surface, either in the pattern required, or subsequently patterned by a negative process.
An example of a direct negative process is the use of airbrasion or laser machining to code a semiconductor package, by abrading or burning away the surface. A truly positive method is seen in solder paste application to a printed circuit board, though positive methods are not as common as might be expected in making the board itself. In fact, printed circuit fabrication gives many examples of processes that are first additive (for example using plating or applying solder mask) and then negative, removing unwanted material to create a pattern.
Re-look at the list of patterning applications that you created for the first activity, and try and classify each according to the approach used. Negative? Positive? A combination of the two? Look at our analysis before you go further.
The actual creation of the image itself is occasionally direct, as in laser marking, PCB direct imaging, inkjet printing, or dispensing. But in most other cases there will be some photoimaging process involved. Key applications here are in semiconductor manufacture, board fabrication, and in the manufacture of screens and stencils for marking and solder paste application. The basic method for photoimaging involves three stages:
We will see later how this is applied to board fabrication, but the general principle applies over a wide range of applications, even though the materials and processes differ widely, depending on the definition required and the area of coverage.
For this topic, go to the link on Photoimaging.
[ back to top ]
Photoimaging is a mechanism that allows us to pattern a substrate or a layer on a substrate, but we need first to get photo-sensitive material in place on the substrate. For this we might use a coating process (as we discuss on the next section), or apply a liquid or film, as we did when photoimaging. But there are some key processes that allow us to deposit material that is already in the desired pattern.
We have already referred to dispensing, but a much more common technique is one of the many varieties of printing. From pre-school days you will have been familiar with printing on paper and similar materials, and a number of printing processes have been developed for this application since printing’s introduction into Europe in the 1400s. These offer excellent definition and registration, but have limitations for electronic assembly for three reasons:
There is of course one exception to the first two of these, in the shape of the process usually referred to as ‘Xerox’, after the company that first developed it. Here the finely-divided particles of toner have a coating that is fused to the substrate by the application of heat after printing. A life-changing process, but one that has as yet few applications in the electronics arena. However, when we want only a thin coat, as we do for coding, then traditional printing processes, such as offset printing have been used.
The printing process does not need to be sophisticated, even for thicker prints of material. One example which is conceptually very simple is pin transfer (also referred to as ‘dip and dab’ or ‘stamping’), where a tool is dipped first in the adhesive or flux to be applied and then onto the work-piece. Whilst the first few applications may be slightly short of material, the process soon reaches equilibrium. All that is needed is a constantly renewed surface of known adhesive thickness into which to dip the tool, and consistent timing for the application part of the process. On typical machines, the even layer of material is produced using a doctor blade held over a rotating pot – the blade retains surplus fluid in the well behind it.
Though apparently crude, the process has been used extensively for applications where only a small amount of adhesive is needed. It is particularly useful when glue dots of substantially different sizes are to be applied simultaneously, and in its automated form is most commonly found on die placement equipment. Reduced to its absolute essentials, the ‘dip and dab’ method will be found useful when trying to apply very small amounts of resin prior to component placement during manual assembly.
The key to pin transfer is to have a material of the right viscosity. Pin transfer, usually automated, has also been widely used for applying flux to small assemblies, such as flipchips and Chip Scale Packages. The success met with has been somewhat variable, with the thickness deposited being variable.
Stamping processes are also used to transfer ink for component coding. A typical process uses a steel blade to wipe a thin layer of ink into etched cavities, from which it is removed using a silicon tampon, which is then pressed onto the surface to be marked. Clearly in this case the material needs to have an appropriate viscosity and surface tension which will allow it to end up in the right place! Typical marking inks contain either adhesives or surface etches to aid this.
Ink-jet printing is another direct non-contact process that seems to offer considerable potential – we will come across this later in relation to direct patterning techniques for board prototyping.
Read this paper which shows how inkjet printing can be used for applying solder mask. Reflect on the complex interaction between process and materials, and the importance of choosing the most appropriate combination for a particular purpose.
Finally, we come to the two patterning ‘work horses’ used in electronics manufacture, screen and stencil printing. Often these processes get confused, because in most factories stencils are used on what is used on what is referred to as a “screen printer”!
There are so many issues associated with screen and stencil printing that we have pulled together information on this topic as a separate section, available at this link. Please read this before you go any further, and tackle the SAQs within it.
Whilst we have described the stencil printing process primarily in relation to printing solder paste, stencils are also used for depositing other materials. The only provisos are that the compounds used should have sufficient working life, that they contain no solvents which might adversely affect the screen or stencil, that the particles of filler used are of appropriately small size, and that there exists some means of cleaning which is compatible with screen or stencil.
Printing through patterned screens is the traditional method by which printed circuits were made, by printing etch resist. Nowadays in board fabrication the main uses of this sort of printing are in depositing legend and peelable solder mask. The process parameters are different in these two tasks: the first requires a thin, well-defined print; the second needs a very thick deposit. In each case, the materials are designed to exhibit shear thinning and thixotropy, so that they will flow well during the print stroke, yet retain their shape once the screen is removed. Particularly in the case of peelable solder mask, cosmetic reasons suggest that the material be chosen so that it will ‘level’ after printing, so that it will form a slightly flatter surface without mesh marks, but not of course spread too far over unwanted areas.
Printing through blank screens, to give an overall coat of medium, is increasingly being used for solder mask deposition on account of its cost-effectiveness. Carried out with vertical board and vertical screens, and some means of confining the medium in contact with the screen, it is possible to print simultaneously on both sides of the board. However, this creates problems in handling boards until the solder mask has been cured.
Printing is also used for patterning polymer conductors and resistors, both for making thick film circuits and for embedding components within conventional boards. Yet other materials are deposited on glass as part of an LCD display.
Printing is increasingly used for applying chip attach adhesive to assemblies which are to be wave soldered, though the process is not without problems. There are significant differences between glue and solder paste, resulting from the fact that the particles of filler in the adhesive are much smaller than in solder paste, and glue generally has a lower viscosity. There may also be issues about life on screen, if the glue is a two-part system that starts to cure once it has been mixed. Such materials change viscosity increasingly rapidly as cure proceeds, which is why most assemblers prefer one-part systems that are cured by the application of either heat or UV light.
Ideally one wants a ‘pile’ of resin which can be deformed during device placement to ensure that both chip and board are in contact with the glue, but stencil printing typically leaves a flat top surface. A further complication is that, for component mounting, one ideally one wants to have deposits of different thickness determined by the size of the component. For example, parts such as SO-packages require a deposit with a higher profile than for small resistors purely in order to ensure that contact is made between the peak of the glue dot and the underside of the component.
To overcome this, a number of proprietary techniques have been employed such as printing with the stencil slightly off-contact and in recent years, innovative use of a different stencil material has enabled a conventional printer to be used to create a glue deposit with varying heights.
Applying chip attach adhesive has traditionally been the preserve of dispensing, and there has in consequence been considerable commercially-orientated debate between suppliers of dispensers, such as Camalot, and the manufacturers of screen printers! The inescapable fact is that, for volume use, printing all the glue dots on a heavily-populated board in a single pass has to be quicker than applying each dot separately, which is the dispensing approach. And this is despite attempts made by manufacturers of dispensing machines to achieve a dot rate of over 40,000/hour.
But this does not sound the knell for dispensers, because dispensing as a process is more flexible, requires less set-up, and has no associated stencil cost. Particularly for small volumes, for boards with few components, for use with assemblies needed a large deposit volume, for situations where the design requirement may change, or additional material is needed to accommodate component changes, then the dispenser may be the preferred solution. And dispensers continue to be the only way of tackling tasks such as flip-chip underfill and applying large volumes of support resin for radial components.
The original idea of dispensing used a syringe containing the fluid to be dispensed, with air pressure supplying the force to eject the fluid through the nozzle. However, as anyone who has ever attempted to ice a cake knows, it requires significant skill to apply just the right amount of pressure to achieve the required result, without any “dribbling” from the nozzle after the pressure has been removed.
Machine manufacturers have tackled the challenge of dispensing the right amount of material, without any defects, in a number of ways, of which the Archimedean screw principle is the most widely used. in our paper this link.
More facts about this topic at Dispensing.
[ back to top ]
If we want to cover a surface, whether the coating is for cosmetic reasons, for protection, or to perform an electronic function, a range of coating processes is available, only some of which have been considered in our section on patterning. As an example, we can apply paints or varnishes to the surface by brush, roller or spray, creating a relatively thin coating whose adhesion to the substrate depends on the materials used, the cleanliness of the surfaces and subsequent processing. Such techniques are used not only for exterior protection, but also for applying process materials such as flux.
When we consider the wider range materials used for protective and decorative purposes at an enclosure level, then the available processes extend to include powder coating and similar methods for applying resilient polymeric materials.
At the component and board level, where materials are more frequently metallic, there are the options of evaporation and sputtering that we will come across in Unit 10 when we consider thin film processes. Carried out in vacuum and often with some heating of the substrate, these provide thin and adherent films of metals and metal oxides, but the thickness is generally limited to around 1µm, and many of the films used for resistors are thinner than this. However, for materials that are easy to evaporate, such as aluminium, evaporation is an excellent way of metallising, whether the end product is a balloon or the high-quality film used in a wound capacitor!
A more common way of applying a metallic coating is the range of processes that we refer to as “plating”. We will find these used to apply solderable coatings to lead frames, to provide decorative and protective finishes for mechanical components that are part of the enclosure, but most particularly in the board fabrication process.
The choice of plating process depends on two factors, on whether the part to be plated is conducting or non-conducting, and on the size of the component. The former dictates the process chemicals; the latter determines the way in which they are deployed.
Although much of our attention in later units will be dedicated to the plating of printed circuit boards, we should not forget the need to protect many of the smaller items used in equipment assembly. These use one of the methods known collectively as “mass finishing”.
The most widely-used system has what is referred to as a “plating barrel”, in which the parts are rotated within the plating solution, electrical contact being made between the barrel and the parts.
Barrels are perforated and may be cylindrical or polygonal in shape, rotating round either a horizontal or inclined axiss. For optimum performance, a barrel will be loaded with a single type of component whose shape allows the load to tumble readily during rotation, rather than rotating en masse as the barrel turns.
Before loading, the components will have been treated either mechanically or chemically to clean the surface, and the filled barrel will then be transferred through the various stages of treatment, processing and post-treatment. Although the impression given by the name is that barrels will be very large (and indeed some are), the principle can be applied to small components, using barrels no more than 5cm in dimension.
A different plating challenge is presented by wires and tubes, which are plated in so-called “continuous plating” plants, in which the material to be plated is immersed in plating fluid and moves past a row of anodes (or between two rows of anodes). The deposition rate is high, and the thickness of deposit is a function of line speed and the length of the plating tank. The process can provide good results and can be highly automated. When continuous plating a metal strip, one potential cost reduction comes from using two rows of anodes and allowing the coating on one side to be thicker than the other, concentrating the metal where it is needed.
Of course, most items to be plated are neither continuous nor small enough to be “tumbled” by barrel plating. The process applied to these is generally referred to as “rack plating”, where the components to be plated are attached to racks, which are fixtures suitable for immersion into the plating solution. Both the rack and the attached parts are subjected to cleaning and pre-treatment, and then plated.
Rack plating is thus a batch process, with individual racks passing through the various baths that make up the final process until the sequence has been completed. Attaching the components to the racks is usually done by hand; flat parts such as boards can be clipped to a carrier, but smaller components may need to be attached by means of copper wire. Allowances have to be made for the shadowing effect that the attachment will have on the deposit, and of course manual rack plating is labour-intensive and hence relatively expensive.
In a typical semi-automatic plating plant, racks once assembled are moved from tank to tank by an overhead conveyor, and the system is set up to control both the sequence and duration of each process step. More sophisticated equipment will also monitor the process parameters.
More facts about this topic at Plating. This paper contains information you will need to grasp in order to understand many of the issues associated with the board fabrication process.
No consideration of plating would be complete without a mention of electroforming, because this process is used for making foil for laminates, some types of lead frame, and stencils for solder paste. Conceptually, electroforming involves the deposition of metal by electroplating onto a substrate from which it is subsequently stripped, so that the plating itself becomes the item of interest. So electroforming needs a considerable thickness of build-up, so that the electroformed part has sufficient mechanical strength to be physically peeled away from the substrate on which it is built.
The substrate on which the temporary layer is formed is typically highly polished to make the stripping process easy, and in consequence the electroformed part will have a smooth surface, mirroring the substrate. We have seen in Unit 1 the use made of this process in producing a foil for use in making PCB laminate, where one side is smooth, and the other relatively rough.
While electroforming is used to produce foils both for laminate manufacture and as a base material from which stencils are laser-machined, the process was originally devised for building a small patterned structures, such as lead frames. This is carried out by depositing a photoresist on the substrate, then exposing and developing it, to leave only selected areas free to be plated. However, because of the fragility of the resist, this cannot be done as a continuous process, as is the case with foil manufacture, where the substrate is a slowly-rotating drum or mandrel.
As a process for making stencils, electroforming gives the metallurgical benefits of the electroformed nickel, including its mirror finish, and also allows any arbitrary number of apertures to be defined to high precision and with an excellent finish, depending only on the quality of the side walls of the resist. Where there are many apertures to cut, a fully electroformed stencil may turn out cheaper than one that is laser cut.
A second process that is closely related to plating, and usually used in combination, is what is actually a range of different processes referred to as “etching”. Originally used to describe the process of patterning engraving plates using acid, and then printing designs from them, the term has taken a wider meaning within electronics manufacture. We speak of ‘etchants’ as materials that will attack a surface, gradually removing material into solution, and changing the surface structure and topography. In electronics, etching is not limited to the use of acid on metal, but is applied equally to the use of hydrogen fluoride to etch silica and to various proprietary (and rather nasty) chemicals used for the surface preparation of smooth polymeric surfaces prior to gluing.
Etching may be carried out over the whole surface, or part of the surface can be protected by a ‘etch resist’. Sometimes etch resists are photopolymers; in other processes photolithic methods are used to define more resistant barriers, such as the tin (formerly tin-lead) used in board fabrication.
With the exception of some etches used on crystalline material such as silicon, most of the materials will etch at equal rates in all directions, so etching is often accompanied by undercutting. However, additives can be incorporated to make etchants act preferentially, in the same way that plating baths contain additives to improve throwing power.
Not only is etching used to remove considerable quantities of material, as with the spaces between copper tracks on a board, but as “microetching” for surface preparation. Here the focus is rather on removing a very small amount of material in order to prepare the surface for subsequent processing, giving a mechanical key to the next layer to be applied. Microetching may use some of the same materials as etching for total removal, but at lower concentrations, and with a more tightly controlled process.
[ back to top ]
Despite the variety in placement processes, we have included placement as a “enabling process”, because there are general issues to be considered, regardless of the application and irrespective of whether the process is carried out manually or has been automated to a greater or lesser extent. These general issues are worth reflecting on, because they affect the cost of the activity, and the quality of the end result:
[ back to top ]
Despite the low regard with which many engineers from other disciplines regard “mere mechanicals”, the truth is that all the items we assemble have had at least some input from mechanical processes, either in the way that they are made, or indirectly from the tooling used. In this brief section, our aim is to indicate the wide range of mechanical processes used to provide components, and the need to consider the processes by which the new components we devise may be made
The processes that can be used are determined by the material, by the design, and by requirements for surface finish, but many of them can be carried out over a wide range of scale, from billet to microcomponent. Conceptually, drilling a hole in an engine block for a large marine engine is no different from drilling a printed circuit board – the scale is different, and the laser option only works for the board, but controlling the bit profile, drill speed and feed rate are all considerations, as is the relative accuracy of hole positioning.
Thinking still from the concept point of view, “mechanical patterning” can be carried out in a relatively small number of ways, the first of which mirrors the negative-working processes commonly used for boards. Working from the solid generally means starting with more material than the final product, and removing that which is not needed. The range of machining approaches available is considerable: one can reduce the size and shape of a lump of material by sawing, milling and grinding; parts with axial symmetry can be turned; apertures can be drilled, bored, punched or reamed. Almost every conceivable shape can be created, though designs with considerable interior complexity may need to be assembled from several machined parts.
All the processes have in common the facts that:
It should be no surprise that, when back-grinding a silicon wafer, a high speed is used to remove the bulk of material, reducing to a much slower speed for the last 10% of the approach to nominal thickness, to improve the surface and reduce the amount of surface damage. Yet the quality of the result, and the number of microscopic cracks and scratches remaining, means that the surface still needs to be chemically polished. For similar reasons, bear in mind whenever purchasing machined metal, that the more perfect the finish, the greater will be the effect on your bank balance!
Has it ever occurred to you that most of the terms associated with mechanical engineering are short, single-syllable words; we have already mentioned many of them, and we have more to come. It was once suggested that this derives from the antiquity of many of the processes, and the use of Anglo-Saxon rather than Norman terms. An uncharitable colleague muttered about “monosyllabic mechanicals”, but we believe these words have an honourable history.
All comments and feedback to the collection being made by Martin Tarr.
Machining processes work from a solid block and reduce the amount of material, but there are other mechanical processes that work by distorting the material, rather than necessarily subtracting any. The first of these that is important to electronics is drawing, a continuous process by which a billet of material is gradually reduced in diameter by passing it through a series of dies, during which it becomes a much longer and thinner wire.
The drawing process is complicated by the fact that drawn material exhibits work hardening, as a result of which it may be necessary to anneal the material between successive draws.
This idea of gradual extension is also seen in the “deep drawing” process by which tall cases are made, gradually stretching the material over a series of dies of progressively greater depth and sharper radius. [Bear in mind that deep drawing is not capable of producing totally vertical sides or right-angled corners, a limitation it shares with all processes where a former has to be extracted from a component after moulding. You will see this if you look at a cast aluminium box, which will always have very shallow “draught” to make release easy.]
Work hardening is also seen with products that are worked from the solid by coining or forging, processes that impact both on the form of the product and its surface finish. Typically after forging parts become tougher and more resistant. A parallel process, applied only to the surface rather than distorting the bulk, is “peening”, where repeated blows are used both to decorate and harden the surface. In both coining and forging, considerable distortion is made within a block of material.
Closely related to these are forming, commonly found in mechanical parts used in components and enclosures, and stamping, as in lead-frames. What distinguishes these two processes is that forming merely changes the contours of a surface, whereas stamping also changes the plan profile, with surplus material being removed. Whether stamping or etching is appropriate for a lead-frame will depend on the volume of production, the dimensions and complexity of the part, and the material. It is not uncommon for lead-frames to be etched in small volume, converting to stamping for longer runs. However, it must be borne in mind that the surface finishes produced by the processes are different, and the dimensional tolerances are considerably worse for the stamped part.
Forming and stamping are frequently combined to produce parts with a three-dimensional shape and cut-outs, and the process can be combined with folding. However, many designs cannot be made in one hit, and require several sequential forming and stamping operations to achieve their final shape. In order to carry this out in a single press, “progression tools” are used, where a strip of material is moved in increments through the tool, with different parts of the tool carrying out the successive stages of forming, stamping and folding. In order to make this possible, as shown in Figure x, the first stage in the process is usually to create a set of guide holes or slots that may be used for “indexing” the strip to the next stage.
Whilst the forces involved in forming and stamping are significant, they are much less stressful to the part than coining or forging, so work hardening is not normally an issue. However, there is distortion of the surface, and this has two effects. Firstly, the thickness of the material sheet is modified locally, becoming considerably thinner at the corners as the material is distorted to accommodate the shape of the press tool. Secondly, while forming and stamping are possible using coated materials, care has to be taken to avoid surface damage during the pressing operation, and allowance made for the fact that the coating is confined to the surface, so any holes produced by stamping will not have a coating. Despite the second limitation, it is quite common to use a coated material in enclosure building, because of its reduced cost and improved aesthetics, compared with painting after assembly. Using precoated material gives a cost-effective way of enhancing the appearance of a unit, but demands a finish that is flexible enough to be resistant to modest deformation.
In most mechanical processes where material is removed, we should not forget the environmental and economical impact of having to deal with surplus material. What was already a hazardous waste becomes more hazardous by being contaminated by cooling and lubricating fluids. There are also potential health and safety issues, because milling, grinding and reaming processes may produce swarf with sharp edges or chips that are eye-threatening.
In our processes so far, we have always started with a piece of bulk material, fashioning it into the desired outline. An alternative is to build up the part from liquid or powder. Particularly in volume manufacture, casting can be most effective. A typical material is an aluminium alloy injected into a mould, a process capable of producing quite heavy castings. At the other extreme, lost wax casting is used for components with fine features.
An alternative way of building up a product starts with a suitably treated powder which is pressed and sintered to form a structure that is mechanically sound, though it may be somewhat porous. However, the spaces within the matrix can be filled with other materials, including molten metals. Consideration of such composites is beyond the scope of this module, but these are materials that you will come across in AMI4817 Design for Thermal Issues.
Whilst we have reviewed most of the main mechanical processes, we have probably not covered all the ways in which metal parts can be made, let alone the processes involved in making plastics and ceramics. To reinforce the point that many processes are involved, take a detailed look at an available piece of equipment, such as a personal computer, and try to establish which processes were used to make the metallic components of the total assembly.
Your view should not be limited to structural members – consider fastenings (forged or screw cut?) and the metal parts of components all the way from semiconductors to connectors (not forgetting any modules such as filters). If you really want a challenge, consider how the ‘nail heads’ in a MELF package might be made. But don’t spend too much time on this task, once you have been convinced that real electronic products have a high mechanical content and that many different manufacturing processes are used – otherwise “you may be some time” . . .
There are many methods for converting compounded plastic to the finished product. Which processing method is best suited to a given material is dictated primarily by the class of polymer (thermoplastic, thermoset, or elastomer), as this determines its physical form and cure mechanism. However, within each class, there will be processing variants which depend on differences in the thermal and melt properties of the compounded polymer.
Thermoplastics are, with few exceptions, supplied as pellets, about 3mm across. Many require drying before processing to eliminate absorbed water, which might otherwise result in bubbles in the melt which would weaken the finished product. However, as was implied in the ‘time and time again’ in polymer basics, any surplus material or scrap can be remelted or returned for reprocessing.
Thermoset materials are generally supplied as liquids. Some rely on chemical initiation of polymerisation, and are supplied as two components which have to be mixed immediately prior to use. However, for the moulding processes below, a more common form is a partially polymerised powder, referred to as a ‘moulding compound’. which is melted during the processing, but subsequently cures. Unlike thermoplastics, when thermosets solidify, chemical reactions during processing convert the material to a cross-linked, non-remeltable product, as we saw in polymer types.
Elastomers can be either solid or liquid and are processed in a similar fashion to thermoplastics and thermoset plastics.
Processes for making polymeric products are described at the links below. The main processes involve moulding, although the extrusion of plastics is also important for making structural members and coating wires. Which process is appropriate will depend on the volume of manufacture, the size of the part and the material with which it is to be made. As you read the resource materials, think about the applications for which each technique is suitable.
Most major components will be made by injection moulding, whereas components are typically encapsulated by transfer moulding. These are conceptually very similar processes, which it is easy to confuse, and we recommend you read our paper on Making Mouldings.
Whilst the transfer moulding process is appropriate for semiconductors, it is not always the most cost-effective method, nor is it suitable for every type of component, either because of the shape of the internal element or its ability to withstand the process. A range of protection processes are used, including dip encapsulation, powder coating, impregnation, casting and potting. More description of these in the first part of our paper on Humidity protection.
Not all polymer materials are used for making solid blocks: many polymer materials are produced as films, and processes of lamination, moulding and forming are important. These are briefly described in our paper Film processes for polymers.
Finally, in relation to the handling of liquid polymers, there is a need to dispense these materials, and we will find a number of applications in electronics manufacture that we described in our paper on dispensing which was referenced in the section on Patterning.
Particularly when polymeric materials are only partially cured, there are risks to the operators to be considered, and some of the processes described above are associated with other risks that need to be assessed and managed. A brief outline of these is contained in our paper on Polymer health and safety issues.
The processes which are described at the links above apply broadly to all classes of polymers. However, there are differences between what is nominally the same material from different suppliers, and the supplier should always be asked to provide data and guidance in both the design and processing of polymers for specific applications.
Polymers are not just used for making things, but they have a range of applications for protection purposes; transfer moulding, potting, casting and similar methods make a product, but their primary purpose is protection. There are two other ways in which polymers are used to enhance the reliability of assemblies. The most common of these is to afford environmental protection by coating a complete assembly. The issues associated with this are covered in the second section of our paper on humidity protection, under the term Conformal coating.
At the high-tech end of the market, dispensing is also the preferred method for applying underfill to small area arrays. The process is required in order to reduce failure caused by the TCE mismatch between component and substrate. This mismatch is made worse with flip-chips because they are made of silicon rather than mounted on a section of laminate, and the stand-off distance is smaller. More information on the materials and methods in our paper on Underfilling.
[ back to top ]
Electronic products are all about joining components together, the process we refer to as electronic interconnection, and frequently we restrict the scope of our attention to soldering. In fact, many different joining techniques are used, even at the component and board assembly level. As with mechanical processes, we are inviting you to look at a real assembly and identify the different joining techniques used in it.
But before we look at specific techniques, there are points to make that concern all types of bonding. First of these is whether the joint is intended to be permanent (short of unintended failure), or expected to be “demountable”. That is, the joint is intended to be disassembled, either on a limited number of occasions, for example for the purposes of servicing or upgrading, or is designed to be made and remade regularly, as with connectors. In order to choose appropriate ways of joining, we need to establish the intended life profile for any demountable connection – our experience with connectors suggests that specifying an unnecessary number of operations may add needlessly to the cost of the part.
We have already suggested that some types of joint may be designed for disassembly, but even permanent joints may need rework, either during manufacture or for subsequent repair or upgrade. The vital consideration here is restoring the joint to its original quality after the rework operation. For example, in even such a simple case as a screwed connection, the rework operation needs to apply the same torque to the screw as was originally applied, and any screw-retaining adhesive needs to be replaced.
Other considerations for the joining method are the ease of application of the method, and the degree to which the process can be automated, both of which reflect in the process cost.
Of the styles of mechanical fixing, probably the most permanent and cheapest to apply is the tubular rivet, provided that it is sized correctly and the force applied to collapse the rivet does not distort the material. Rivets are particularly good for holding large plates together.
However, especially at the enclosure level, assemblies may need to be accessed or removed. Screws seem a good choice for a demountable assembly, but the choice is not trivial. For example, a self-tapping screw fitted into an undersized hole will give a secure fixing for only a limited number of times, and the hole in a relatively soft metal will need to be reinforced to allow repeated use. The life expectancy is even shorter when using a self-drilling type of screw.
Many enclosures use machine screws with a parallel thread, in combination with either a pre-tapped hole, or a captive nut. Such nuts, within cages, are essential components in standard equipment practice racks, as they allow a degree of positional latitude to allow for tolerances in the structure. Avoiding using a nut altogether, or making it captive, both get around the assembly nightmare of trying to align a bolt with washers and nut, and tighten the assembly adequately. Why is it that designers never allow sufficient space for access with a spanner or even a stubby thumb and index finger?!
One difficulty with a screwed assembly is that it is prone to loosen due to vibration, and there are a number of strategies to counter this. Lock nuts and lock washers are non-preferred, as a self-locking nut with a polymeric insert is much easier to assemble, being a one-part assembly. Of course, for a permanent assembly, a thread-sealing material (generally known by the trade name “Loctite”) is very effective. Finally, remembering that not all rotation is unintentional, your design may have to consider using an anti-theft screw of the type that can be applied with a normal screwdriver but only disassembled with the correct special tool.
Holding parts so that they don’t fall out is something a circlip is good at, particularly shafts, but the application can be fiddly, and this approach is best for a ground shaft in a hard material. So you will probably only see circlips in applications such as printers. By contrast, other forms of spring retention are much more common. These are occasionally conventional helical springs, but flat spring forms are more usual, and these are the way that boards are secured within connectors.
An alternative to mechanical fixing is the use of adhesives. We saw in Unit 1 that these are polymeric materials with the ability to wet and adhere to surfaces, and sufficient cohesion after cure to create a satisfactory bond. Detailed consideration of adhesives is beyond the scope of this module, but you need to be aware that adhesives are very much tailored for the application, in terms of the material of the surfaces to be joined and the environmental specification of the bond.
Of most practical importance to the assembler is the form in which the adhesives come. Many are liquids, but tapes and preforms are also used. Liquid adhesives may be one-part types, cured either by high temperature or by exposure to UV radiation. These have a defined shelf life, which can often be extended by storage at very low temperature. The advantage of a one-part adhesive is that the formulation and mixing is carried out by the supplier, and is more likely to be accurately done than in the case of a two-part system, where materials have to be mixed by the end user, before putting into position. You are probably familiar with Araldite for home repairs – a two-part system, it is easy not to get exactly equal amounts of the two materials, or not to mix them completely. Also, you may have experienced a problem with pot life, where the resin is formulated to solidify in a time that is a compromise between the needs of those with a lot of parts to assemble, and those who will have to hold assemblies together during setting. Somewhat faster-acting systems can be adopted using mixing dispensers, where the resin and hardener meet at the dispense head. However, there are obvious practical problems in getting control and ensuring cleanliness.
An alternative liquid adhesive method that is appropriate for some structural applications uses ‘hot melt’ adhesives, supplied in solid form (generally sticks) and extruded into position by a combination of heat and light pressure.
Many adhesives used in electronics are in the form of films or tapes. Even the label on your PCB has an adhesive film to attach it, and tapes are frequently used for thermal bonding. Of course, tapes do not conform to surface irregularities in the same ways as liquids do, so the materials used are typically fairly thick rubber-like formulations. With tapes, the decision has to be made as to whether the parts need to be re-positioned after assembly, in which case a tape of the “instant grab” type would not be easy to use.
Information on how things stick together, why bonds break down, and the importance of surface preparation can be found in our paper on Adhesion.
We are a long way into joining techniques, and you will have seen how many of these are used in electronic products. Yet until this point we haven’t talked about any of the processes that involve metal joints. Soldering is not the only process, but brazing, welding and diffusion bonding all share the common characteristic that they create an interface area between the parts being joined that is significantly different from the bulk, either in terms of its composition or its structure. And all are processes that take place at elevated temperature.
In both welding and diffusion bonding no additional material is brought to the bond site, but the bond is formed from the metals present. In the case of welding, energy is applied locally to the interface to create a volume of material that is the melted product of the two materials. In a spot weld, the interface is not very deep, and only present locally in a few places across the entire interface, whereas in a structural weld a substantial nugget of weld material is formed along the whole of the bond line. Being a high-temperature process, extensive welds would normally only be carried out in an inert atmosphere, such as argon. Which is why the very cheap (and nasty) spot weld is usually associated with discolouration of the surface at the weld point.
But welding is not just a process carried out using ferrous metals and serious amounts of energy – every integrated circuit has a number of small welds, diffusion bonds, formed between the bond wire and metalisation by the application of heat and ultrasonic energy. This creates a bond line that is typically only a micron thick, although the region tends to grow by continued solid state diffusion between the two materials.
Our final type of joint is that which is most familiar, formed by placing a fusible material between the surfaces to be joined, applying heat, and allowing the interface material to melt and wet the surfaces. Carried out at high temperature, this process is referred to as brazing; at temperatures under 425°C it is generally referred to as “soft soldering” . We will see the use of a high-melting point eutectic of gold and tin in some varieties of semiconductor, but most electronics soldering processes use low-melting materials with a high tin content.
Because this is such an important process, we have a number of supplementary documents that explain what soldering is. However, as one purpose of this unit is to open your minds to alternatives, it is worth pointing out afresh that there are a number of ways of creating a traditional solder joint. For example, we can take surfaces coated with solder, and merely heat them in contact. This diffusion soldering works best with the application of flux, but some companies would say that it is possible to create good joints without flux, given an inert atmosphere for heating and some pre-cleaning of the surfaces to remove oxides.
The mainstream processes, however, involve applying solder and flux to the interface in the form of paste, or applying flux separately and adding liquid solder. In all cases, liquid solder from the soldering iron, the wave solder pot, or the reflowed solder paste, contacts the surfaces to be joined, and sufficient time is allowed first for wetting to occur, and then for the joint to cool and solidify. The only difference is that hand soldering and wave soldering bring liquid solder to the joint, whereas in reflow soldering the solder in the paste melts whilst in contact with the surfaces.
With all these soldering processes, we need to design the conditions so that the only relative movement between the parts is the result of surface tension processes whilst the solder is molten. Physically holding parts together while the joint is made is hardly practicable, yet it is important that the joint is not disturbed during its solidification phase, or the structure will be modified and the reliability of the joint impaired.
The two key resources on this topic are:
Soldering is a topic that will be important in Unit 6, so we have provided here links to two topics that are relevant, but which you will probably decide not to read at this stage, Reflow soldering and Wave soldering.
[ back to top ]
Some use of cleaning is inevitable, although any cleaning is a process that adds cost rather than value! So why is cleaning carried out? In the case of board fabrication, there is an easy answer, that the natural oxide layer on the board, combined with any contamination, needs to be removed, because it would otherwise interfere with subsequent processing. Similarly with board assemblies that are intended for conformal coating, as the materials used are incompatible with flux residues.
The rationale for cleaning becomes weaker when we look at the washing of board assemblies for cosmetic purposes, or in the belief that flux residues will promote corrosion failures. Unfortunately, much unnecessary cleaning is carried out – a no-clean paste will leave some residues, particularly if the reflow process is not optimised, but the residues themselves should have little impact on the reliability of most assemblies.
Assuming however that we are intent on cleaning, whether it is needed or not, we need to be aware that there are a number of different types of contamination. We have already mentioned the possibility of oxide build-up on a surface, but there are also less specific “soils” on the average assembly. [Note the terminology preferred by the cleaning profession, which prefers “soiled” to “dirty”!] Some contamination will be ionic in nature, and will generally merit removal, as it may genuinely impact on the potential for corrosion and electromigration. Other material is organic, often degradation products of the materials used in assemblies. But many other soils come from accidental contact with the environment and the people in it, whether fingerprints, skin flakes, or dust particles. Many of these need to be removed mechanically, rather than attempting to dissolve the problem away. In fact, when it comes to any attempt to hold soils in solution or suspension, one must always remember that a cleaning process is one of dilution, and total removal of the soil is very difficult.
Cleaning processes will always have a mechanical component in order to bring the cleaning agent in contact with the surfaces to be cleaned. This might take the form of agitation, aeration, or the application of ultrasonics, all with the aim of breaking down the bond between soil and substrate, and allowing the solvent to wet the substrate and thus loosen the soil.
Generally the materials used for cleaning should be as benign as possible – deionised water is a good starting point – but some kind of surfactant (wetting agent) will generally be used to enhance the wetting. For breaking down organic materials, such as flux residues, more powerful surfactants may be needed. Alternatively, one might use an organic solvent of some kind. Unfortunately, a wonderful solvent widely used as a cleaning agent (CFC-113) was proven to be damaging the ozone layer, so it has been banned. In consequence, a number of replacement processes have been developed, as indicated in the paper at this link.
Note that all the processes we have discussed so far require some method of re-circulating the cleaning agent used, because a totally open-loop process would be uneconomic.
Where the cleaning is required to remove an adherent and resistant coating, then water plus surfactants or organic solvents may not suffice. For oxides, for example, chemical processes may be needed involving the controlled application of etchants; for the pre-process treatment of polymeric surfaces, it might be necessary to use relatively aggressive methods, such as exposure to oxygen plasma.
Irrespective of the method used, there are two considerations common to all cleaning processes: how to protect the freshly cleaned surface, and how to dry the part. Drying without leaving drying marks is so difficult that in cases where there is subsequent water processing, such as in board fabrication, the best course is actually to proceed immediately to the next process, and circumvent the need for drying. Otherwise, the best practice should be to remove droplets of surface water from active areas, and dry them as fast as possible, leaving any drying marks (effectively soils at low concentration) in non-critical areas of the circuitry.
A final comment about cleaning that is not always remembered is that polymeric materials absorb water, and this may have an impact on subsequent processing. In order to remove as much as possible of the water absorbed during cleaning, baking processes are normally required, but this should not be carried out after the very last cleaning process.
[ back to top ]