Freeing electronics from chips and boards

A breakthrough in inkjet printing could precede the manufacture of electronic circuitry at a low enough cost to justify the claim that it can be considered disposable. Feature sizes measured in microns, even on uneven substrates, allow the achievement of useful computational speeds and powers. The first application is expected to be flat panel displays, including products as thin and flexible as paper. Next will come smart tags and labels, followed by a wide range of disposable and semi-disposable devices. A University of Cambridge student materials practical back in the 1960s used photoresists, evaporation and vacuum deposition to lay down electronic circuitry on photographic slides and film, the intention being to show how integrated circuitry could be deposited on anything vaguely suitable. Sometimes the circuits worked and sometimes they didn't. Sir Nevill Mott, the Cavendish Professor of Physics, was around that time giving his brilliant if hard to understand graduate lectures on amorphous semiconductors. Research has continued since, and it should be no surprise that Plastic Logic, a small Cambridge company which promotes Sir Nevill’s and subsequent ideas, is the brainchild of Richard Friend, the present Cavendish Professor, and Cavendish Laboratory colleague, Dr Henning Sirringhaus. The key breakthrough is not the printing of polymer transistors, which has been possible for the last decade, but finding a way to make some parts of a substrate hydrophobic (water hating) and other parts hydrophilic (water loving). The intention is to allow inkjet printing to lay down transistors with channel widths (usually described as 'lengths') of only a few microns or less. The switching speed of a transistor is proportional to electronic mobility but inversely proportional to the square of the channel length, so this needs to be as short as possible. One method is to first lay down a film of polyimide, which is hydrophobic, on a substrate, which is hydrophilic. The polyimide is coated with photoresist, and then etched through using an oxygen plasma. The polyimide line that defines the transistor channel is protected by the photoresist and remains hydrophobic. In a typical development, the source-drain and gate electrodes are produced by inkjet printing a water-based ink of poly(3,4-ethylenedioxythiophene) doped with polystyrene sulfonic acid, (PEDOT/PSS) 'Baytron P' made by Bayer at Krefeld in Germany. A second breakthrough is in the use of a specially-developed intelligent printer, which uses a commercially available print head plus a camera to look for registration marks or previously printed layers in order to deposit droplets in the right places. It can align itself to the substrate pattern to cope with distortions. The droplets of PEDOT are deposited sufficiently close to the hydrophobic line of polyimide that they will reach it yet be repelled by it. Channel line widths or lengths can be routinely laid down as small as 2 microns across, and sub-micron line widths have recently been produced experimentally. A continuous 150 to 300 Angstrom thick layer of poly(9,9-dioctylfluorene-co-bithiophene) (F8T2) is then spin coated on from a solution in xylene, followed by a 0.4 to 0.4 micron thick layer of polyvinylphenol (PVP) from a solution in iso-propanol and a gate electrode line of PEDOT above the channel. Mechanically rubbing the polyimide allows it to act as a template to line up the rigid rod F8T2 molecules so as to improve field effect mobility in the vertical direction. It is also possible to print resistors and capacitors, and via holes can be made by etching right through the substrate with a suitable solvent and then filling the holes with a suitable conductor. The manufacturing process is conducted at low temperatures, uses no hazardous solvents or other nasty chemicals, and is potentially very low cost. Switching speed of the experimental devices is currently around 1kHz, which is not enough for a wearable PC, but more than enough for the intended applications. The first of these is expected to be to provide active matrix backplanes for niche flat panel displays, taking advantage from the fact that the circuitry is both transparent and very thin. Devices shown to visitors have been laid down on sheets of glass, but the team has already carried out experiments on laying down circuitry on plastic. The intention in this case is to produce smart labels for packaging. Tagging of products for supermarket checkouts, mooted in Eureka's December 1999 edition, is now a practical possibility. The inkjet printing process can be used to lay down arrays of circuits, each of which is different, incorporating a different encrypted code for product tracking and to combat counterfeiting. As Plastic Logic’s marketing manager Tracey Stephens puts it: "Whatever amorphous silicon can do, we can do cheaper". Tags could be powered by induction loop or printed on batteries, such as those developed by Power Paper in Israel. Other target market areas are in wearable electronics, smart cards, and disposable medical diagnostic systems. Displays based on substrates as cheap, thin and flexible as paper are considered a practical possibility. And it is not inconceivable that future greeting cards will not only be able to play a tune, but display a short video sequence. No date has been set for incorporating the technology into the pages of Eureka but we look forward to incorporating it when it becomes available. However, still to be done is lifetime testing. Meanwhile, Plastic Logic comprises around 25 people and has plans to expand. The first request for funding in April raised £6.3 million in a round led by PolyTechnos Venture-Partners (Munich, Germany), Amadeus Capital Partners (London and Cambridge, UK) and The Dow Chemical Company (Midland, Michigan,USA). Shareholders include the University of Cambridge, Cambridge Device Technology and Seiko Epson, with the initial licence income from products expected in 2003/2004. Plastic Logic Tracey Stephens, marketing manager Pointers Inkjet printing can be used to lay down electronic circuitry at extremely low cost Channel lengths can be made short enough to allow kHz switching speeds First commercial products using the technology are expected in 2003/2004