The status of flexible printed wiring as we approach the year 2000

Citation

Stearns, T. (1999), "The status of flexible printed wiring as we approach the year 2000", Circuit World, Vol. 25 No. 1. https://doi.org/10.1108/cw.1999.21725aaf.002

The status of flexible printed wiring as we approach the year 2000

The status of flexible printed wiring as we approach the year 2000

Keywords Electronics industry, Flexible circuit, Market, Printed circuit boards

Introduction

What is the state of the flexible printed wiring (FPW) industry as we approach the year 2000? Commercialized in the mid-1950s and exploited from the outset in glamour applications like satellites, medical implants, and inter-planetary exploreration, it is only in the last ten years that this interconnector has become the baseline harness in all types of high-volume commercial electronic packages. What is new in FPW, and where next?

My work in FPW engineering suggests that the main thrust is to finer lines, exotic environments and new materials. Seminars and magazine articles on the potential for FPW technology in MCM modules, "smart cards" and similar high-density products support this idea, but a look at market analysis figures for 1997 tells a different story. For the most part, the industry soldiered on with a standard mix of product and process.

Current situation

Size of the industry

Worldwide, FPW is a multi-billion dollar business. In the USA, which is not the biggest producer, more than $600 million in circuitry was produced last year (mostly based on polyimide films) with roughly another $120 million added in assembly labor and component value. This is not a small business, and it has grown steadily each year by more than 15 per cent. A survey of major producers concludes that expansion will continue past the year 2000, although at a reduced rate of about 12 per centannually.

Not surprisingly, Japan is the largest producer (and consumer) with much of the production coming from efficient, roll-to-roll equipment.

Materials

Polyimide films, produced in several varieties in the USA by Du Pont, and in Japan by Ube Chemical and others, is the dielectric of choice for about 85 per cent of the sales volume of FPW. They are relatively expensive at $0.80/sq. ft per mil, but polyimides are very tough and resistant to thermal or chemical damage, so they are the obvious choice where reliability and immunity to damage are important. Polyester films are similar in most ways to the polyimides but have lower heat resistance, which prevents their use in soldered-assembly applications. So polyesters are found in high-volume, low-cost, pressure-connected or conductive-adhesive attached applications. They are the preferred dielectric for most of the remaining 15 per cent of FPW sales, much of it produced on roll-to-roll equipment.

Nothing new here; FPW materials have been built by adhesive lamination of metal foil on dielectric film for many years and used in approximately this ratio of "high performance" to "cost effective".

Polyimide-based metal-clads for flex are traditionally assembled with thermosetting adhesives which require a certain period of sustained pressure and elevated temperature to "cure" the adhesive and form a stable and strong foil attachment. As a result, polyimide-based metal clads are produced in sheet form, which in turn means they are only usable in a panel-based FPW factory. (There is another reason why press lamination is used, and that is dimensional stability: If a roll-to-roll process is used to join dielectric film to foil, the web tension which is required for control of web position results in excessive after-etch shrinkage which is not desirable.)

In summary, polyimides are expensive materials and polyimide-based flex is usually produced by panel process (also expensive), while polyesters are roughly one tenth of the cost and are used in roll- to-roll produced process (less expensive).

The rising newcomer in flex materials is the "adhesiveless" class. These are composites of metal on dielectric film without an identifiable adhesive layer. When you eliminate the adhesive from FPW materials you remove the degradations which traditional FPW adhesives inflict, i.e. lower thermal performance, reduced chemical resistance, poorer dielectric properties, increased "Z" axis expansion. Because the production processes used for adhesiveless composites are roll-to-roll, these high-performance polyimide-based materials permit use of roll-to-roll production for polyimide FPW.

Adhesiveless materials constituted about 7 per cent of the value of base materials in 1996, but rose to 10 per cent in 1997, showing a healthy penetration into this market. Utilization in 1998 and beyond should be even higher, with ever-more rapidly expanding use as per sq. ft cost drops with higher volumes and increased competition. The dollar value of 1997 adhesiveless material sales was around $13 million.

Material cost in FPW

As production efficiency rises, the percentage of FPW value which is represented by the purchased material content has slowly risen and stands today at just under 25 per cent. This figure includes, of course, scrap losses and other inefficiencies.

Circuit types

FPW is manufactured in single, double and multilayer forms; multilayer circuitry is also built with PWB-type outer layers, and is then called rigid-flex. The largest selling type is, not surprisingly, single-layer, constituting 46 per cent of polyimide-based circuitry and probably even higher percentages in polyester or other dielectrics. Second in usage is double-sided circuitry at approximately 21 per cent; multilayer and rigid-flex both account for around 9 per cent of sales volume.

Design

In FPW, "fine-line" is defined as "less than 4 mil lines and spaces", a limitation based on the use of dry-film resist for circuit imaging and the lower limit of etch in 1oz (0.0014in thick) copper claddings. There is a lot of talk about "fine-line" circuitry and the rising utilization of smaller features to increase circuit functionality. But all of this discussion hardly translates into product; 94 per cent or more of current FPW is in designs that are 4 mils or larger, only 6 per cent is produced below this "fine-line" boundary. There is a great deal of future improvement to be gained in this area. My discussions with engineers at many flex producers suggests that there is considerable activity in prototype designs that are finer than 4 mil, and all of them expect that the future will include steady movement towards "semiconductor-level" feature sizes.

Market segmentation

Producers

Previously, most FPW was produced by an in-house department for internal consumption and little was purchased from independent producers. This ratio has inverted dramatically with the rise of accountancy as a controlling force in company management, with the result that almost 95 per cent of FPW is produced by independents specializing in this product. Some of the change is driven by management worries about pollution from wet-process and plating chemistries. Most independent producers, in turn, have expanded into assembly work, i.e. they attach connectors and other components, plus perform electrical test on the completed assembly. This is all in line with the popular trend to outsource everything possible and to concentrate on "core capabilities".

Applications

The largest single usage of FPW, more than a third of all sales volume, is in computers, a category which includes peripheral equipment, such as printers. There is a large consumption of FPW in ink-jet cartridges and similar throw-away items.

The second biggest application is automotive, where FPW is found in many locations from the dashboard to engine controls, ABS controls and other "black box" items which give the motoring public the sophistication for which it is apparently willing to pay. In 1997 this class consumed about one fifth of FPW output.

Communications, the cell phone, pagers, all sorts of switching and data transmission equipment, etc., make up another 16 per cent for third place on the list.

Government applications, principally high-tech, in the form of very expensive but indispensible rigid-flex circuitry, constitutes the fourth-place market (slightly more than 10 per cent) in dollar value. Rigid-flex circuits are, essentially, the only interconnector which will fit into the crowded US avionics packages, where they provide dense and reliable circuit paths between connector and motherboard. Defense procurement is down in 1998, but moves along steadily; these are solid applications of FPW technology and it is unlikely that alternative can be found to get the same efficiency in space, weight and reliability.

Other significant uses for FPW are instruments, medical, consumer (e.g. watches, clocks and cameras)

Assembly

As noted, assembly labor and component value accounts for something like 20 per cent of the total value of FPW sold in 1997. Assembly work, as a value-added function, is rising in military circuitry and dropping in commercial product. These changes are probably a result of constant emphasis on documented quality in the military field (with stable to lowering circuit costs) and improving cost effectiveness in commercial work, where automation in component placement, use of conductive adhesives, increasing implementation of SMA techniques and changes in inspection philosophy contribute to lower total cost.

Trends

Materials

The printed wiring board (PWB) industry producers of "rigid circuitry" is about ten times as large as the FPW industry. There is a fundamental difference in the materials of construction, in that PWB laminates consist of a thermosetting resin with fibrous reinforcement (glass cloth, various papers, so-called "non woven matts", etc.) while FPW materials, in general, are based on polymeric films. PWB materials are manufactured of several plies of material,in a press-lamination process, while at least some FPW materials polyesters, adhesiveless polyimides, pressure-sensitive systems are manufactured in single pass by roll-to-roll process. Hence, there is an inherent cost advantage in FPW materials: they can be made in a single pass, and they can be made by roll-to-roll process.

An enormous technical advantage accompanies roll-to-roll production, because this method allows economical production of deposited-metal adhesiveless materials. Since the conductive layer is applied to the dielectric support in a progressive build-up, rather than as a bonded-on foil layer (as used in PWB and traditional FPW materials), it is easy and even inexpensive to create a laminate that has extremely thin copper claddings microinches of thickness, rather than "ounces" or "mils".

Adhesiveless FPW materials, therefore, are ideally suited for production of the highest density circuitry, because they contain the thinnest possible dielectric layers, which means many layers can be built into a relatively thin construction, and because these super-thin claddings allow extremely fine features, inasmuch as no allowance need be made for etch undercut.

The PWB industry is firmly wedded to thick, multi-plied, stiff materials, and culturally resists any change to weak, limp, flimsy FPW materials. But the unavoidable advantages of adhesiveless material technology will, undoubtedly, lead to its use in both PWB and FPW constructions and a blurring (or total elimination) of the distinction between these two industries; it is only a matter of time and familiarity. When that change occurs, there will be, literally, an order-of-magnitude increase in the utilization of these materials simply because of the enormously larger PWB market. Because equipment for producing adhesiveless metalclads is roll-to-roll, there is already enough production capacity in place to serve this increased market for the first few years. The result is that once the emotional and historic barriers are overcome, both industries (or their successor) will enjoy significant improvements in cost/performance.

New polymers

Polyimides and FPW developed together; properties of these highly developed polymers are tailored specifically to the needs of the flex circuit industry. So it is going to be very hard for a new material to significantly penetrate the market. However, there are always new materials in development which offer improvements in performance over traditional material. Success in the commercial world depends on acceptance and the complicated calculus of capitalization, acceptance and cash flow.

Examples of polymers which have shown major improvements in dielectric or physical properties include Avatrel (BFGoodrich); PBO and PIBO (Dow), and liquid crystal polymers (LCPs) such as Vectra and Superex. Lower moisture absorption and dielectric constant with even higher tensile strength and tensile modulus are some of the advantages of these materials, but they will not be a factor in the materials equation unless sufficient interest is shown in them to achieve (in sequence) at least pilot production capacity at tolerable prices, specification approval, willingness on the part of the FPW producer to process, and end-customer acceptance.

Covercoats: the non-coverlayer process

Traditional FPW circuits have a second dielectric layer overlying and protecting the etched circuit runs. This layer, bonded on to the circuitry in a lamination process using equipment and adhesives which are very similar to those used to produce copper-clad laminates, must be provided with apertures or termination areas for access to the conductors.

There are a number of techniques for making these openings. Before the layer is attached to the circuit, it can be drilled or punched. When this technique is used, the coverlayer must be carefully registered over the etched pattern and held in register as the adhesive is heated and flowed by lamination process to form a tightly bonded layer. This process has many locational and dimensional uncertainties, with the consequences that aperture size and location can vary considerably. The method is simply not accurate enough to be used on termination patterns that are common in SMA technique or in chip-on-flex.

Creating openings after the coverlayer is bonded on to the circuit improves the accuracy of size and location substantially, but at considerable expense, because of the (chosen and desired) toughness and durability of FPW coverlayer materials. So a new technique called covercoating has evolved and is rapidly taking over wherever finely detailed and accurately registered openings are needed. This method relies on photoimageable or accurately printed liquid or film coatings, rather than laminated adhesive-cum-film coverlayers. Photoimageable coatings and films (exposed and developed after they are in place on the conductor pattern) provide best accuracy, but very good results can be had with careful screen printing as well.

Overall cost is reduced because the materials are applied only where needed, the labor and material consumption involved in press lamination is eliminated, and the inherent cost of the coatings is less than the cost of a dielectric film with adhesive coating.

A hidden advantage is that higher thermal performance is possible because there is no traditional FPW adhesive in the system. If an adhesiveless base material is used to create the conductor pattern, and a polyimide covercoat is used for the outer dielectric layers, the resulting FPW circuit could, it is arguable, be considered "adhesiveless" and certainly would have excellent thermal endurance, limited, primarily, by oxidative destruction of a copper circuit pattern.

Design

Significant improvements in circuit density are in sight and will become real as more producers become comfortable with adhesiveless materials and the technology for creating microvias by laser, plasma or chemical etch matures. It is not uncommon today to find FPW with 2 mil lines and spaces, at least over small areas in short-run circuits. Such fineline circuitry demands equally tiny vias and landless interconnects to maintain the efficiency of circuit area utilization. Features less than 2 mils will become common over the next few years while increased use of AOI will reduce quality anxiety.

Liquid resists are likely to be used in increasing quantities in print-and-etch process as this shift occurs, because circuit pattern resolution is inversely limited by resist thickness. Semi-additive techniques, based on dry-film resists and a thin copper-clad base material, has its advantages, and will continue to be be used where it is appropriate. This technique utilizes a thin copper clad base material; adhesiveless clads are ideal. On a negative resist image, open areas correspond to the desired circuitry is applied, the circuit is then plated-up to desired conductor thicknesses, the resist image is stripped off, and a brief "selective etch" step is used to "singe" off the interconnecting background copper, with minimum adverse effect on the image (which is much thicker). A semi-additive process provides nicely controlled sidewalls and good definition, because the circuit replicates the resist image. It also allows overplate with any desired electrolytic finish, because the circuit is fully interconnected until the selective etch step.

Summary

The FPW industry is strong and growing, but at a reduced rate compared with the recent past. There is a steady shift to high quality commercial products and away from government/military production. Line width and spacing continue to reduce and use of vias smaller than traditionally-drilled (0.013in diameter) sizes increase.

Materials of construction continue to be primarily polyimide with rolled copper foil cladding and highly modified or plasticized adhesives, but metal-clads produced by so-called adhesiveless methods now account for over 10 per cent of sales and should extend their market capture based on improved performance, potential to increase circuit density and roll-to-roll availability. Polyester based FPW continues at roughly 15 per cent of the market; may increase with expanded use of conductive adhesive assembly, which avoids elevated temperature exposure.

Assembly activity remains fairly constant, with a slight drop in use in the military area and offsetting rise in commercial circuitry. Added value averages around 20 per cent.

Continuing efforts at the molecular level have produced new polymers which may eventually appear in the FPW arena, offering lower moisture absorption and dielectric constant, two critical characteristics for high-density interconnnects.

Liquid and film-type photoimageable systems replacing traditional adhesive-bonded films provide improved alignment and smaller, better defined apertures and are taking over in FPW for SMA.

Tom StearnsBrander International ConsultantsNashua, New Hampshire, USA