The rapid takeoff of e-readers has sparked the imagination of consumers. There is a rising clamor for advances that will deliver the ability to read a newspaper on a piece of plastic that rolls up and fits in a pocket. So the technologies closely associated with such futuristic products are getting a lot of scrutiny.
Two of the most high-profile technologies in this camp are flexible-circuit and display substrates, and ink jetting. The promise of ink jetting is that it can potentially serve as a way to deposit exotic semiconductor materials in environments that don’t need the high temperatures and vacuums associated with making conventional semiconductors. And indeed, ink jetting has come along way from its original role in document printers. Increasingly, ink jets are a way of putting materials ranging from nanotech inks to genetic material on unusual substrates. A recent news item highlights the trend: Wake Forest University researchers are experimenting with the technology to apply skin grafts by “printing” skin cells directly onto a damaged area. They hope the new method will work better than skin grafts, which can be difficult to apply on severe burns.
Though the details are sketchy, ink-jettable versions of transistors are said to be in the wings. And ink jetting is a technology of interest for fabricating the next generation of flexible displays and smart ID tags.
Nevertheless, it will probably be a few years before ink jetting is ready for prime time as a means of creating active electronic components. Flexible displays now in the prototype stage typically use a combination of traditional circuit-deposition techniques modified for use on flexible substrates. Researchers working in flexible displays say ink-jetting techniques are too immature to play a role in the displays just yet. “There is considerable concern about whether you will be able to get the kind of throughputs you’ll need to get the kind of cost advantages that have been promised,” says Arizona State University’s Flexible Display Center, Tempe, Ariz., director Nick Colaneri. “No one can yet make a usable display just using printing processes. You still need conventional metal deposition and patterning steps.”
A case in point are prototype flexible displays created at Colaneri’s labs. To get units as close to commercial technology as possible, ASU researchers started with conventional processes for making amorphous silicon (a-Si) thin-film transistors, then adapted them for use on plastic substrates. “We basically had to make flexible plastic look like a piece of glass, because the processing equipment was accustomed to seeing rigid materials,” says Colaneri.
What resulted was sophisticated work in materials engineering that resulted in techniques for temporarily attaching the plastic substrate to a glass plate for processing, then peeling it away without disturbing the electronics that had just been deposited. “No one currently makes the production tools for either of these steps. We are in discussions with three Asian manufacturers that want to license the technology and a key part of the discussion concerns bringing in suppliers to make the tools for the processing step,” says Colaneri.
ASU chose to work with a-Si because it has become the material of choice for the active layer in thin-film transistors (TFTs) widely used in liquid-crystal displays (LCDs). One attraction of a-Si is that it can be deposited at temperatures as low as 75°C. This allows for deposition on plastic over large areas by plasma-enhanced chemical vapor deposition, a technique that can deliver low production costs.
Researchers are also investigating the use of organic semiconductors for display electronics. Organic semiconductors based on plastics hold an allure for flexible displays, but, “The jury is out on whether those materials will deliver the kind of performance that would be worth the headache of moving away from a-Si,” says Colaneri. “And it is not clear how you would deposit them in a mass-production environment.”
Zinc-oxide semiconductors are getting attention as well. The material already sees use in displays because it is transparent. Its high-electron mobility and wide band gap also make it a candidate for thin-film transistors and light-emitting diodes. However, technology capable of producing ZnO semiconductor circuits in high volume seems to be way off. “The benefit of ZnO is that you can get more performance than with a-Si and it looks as though the deposition processes can be similar,” says Colaneri. “Most people deposit ZnO with conventional sputtering at the moment, but atomic-layer deposition may be promising. There the challenge is to develop tools that can run the process much more quickly. Otherwise, it is an economic nonstarter.” Colaneri also says printing techniques are one of the methods under investigation for ZnO deposition.
Ink jetting is increasingly viewed as a more “green” alternative to conventional printed-circuit technology. That’s because ink jetting is additive; laying down conductors and insulators only where needed. This contrasts with subtractive wet processes normally employed to plate conductors onto a circuit board. The electronics industry increasingly views as liabilities the etching, stripping, metallization, and washing that define ordinary printed circuits because these processes entail use of strong acids and other chemicals that need special handling and disposal. Ink jetting looks like a way to eliminate many of the environmental complexities that now accompany such processes.
But ink jetting as a circuit-board process is still not in the mainstream. More typical of how flexible circuits are made today are the processes in place at 3M, St. Paul, Minn. 3M produces flex circuits on polyimide and PET substrates for use in medical displays and high-volume consumer products such as cell phones and cameras. 3M makes flex circuits having up to two metal layers on roll-to-roll handling equipment. Material in the form of flat sheets can have up to four layers.
3M puts layers onto its flexible substrates via sputtering and copper plating. Despite the fact that material can be just 12-microns thick, 3M says it can image traces as fine as 15-microns wide on a 35-micron pitch. 3M also uses a special process to etch into the polyimide, rather than punching, to make features such as cantilevered leads which can then be attached to dies.
It can be tricky getting to such dimensions on superthin flexible films. “As substrates get thinner it is more of a challenge to keep alignment and registration between layers. Tensioning on the material web is also an issue when you need a substrate that is within single digits of microns to features designed on the circuit,” says 3M Flex Global Lab Manager Robin Gorrell.
Nevertheless, waiting in the wings are ink jettable materials poised to help eliminate some circuit board processing steps. For example, Dow has developed a platform of inkjet inks for PCB processing under the trade name of LithoJet. The firm has etch resist inks, a dielectric ink and a UV blocking ink for use with photo-imageable coatings, like solder mask. There are also a series of inks under the trade name Enlight specifically for photovoltaic applications (crystalline silicon cells). These are for selective patterning applications. The Dow roadmap includes metal inks for PV and printed electronic applications, says Dow Electronic Materials Emerging Technologies R&D Director Tom Sutter.
But the ink jet mechanism that deposits circuitry looks nothing like the printer that sits near your desktop. The same pressures on IC makers to continually reduce integrated circuit size are at work in ink-jetted circuitry. So makers of production-grade ink jetting equipment are continually devising ways of depositing smaller, more precisely positioned drops of working material, and doing so at ever higher speeds.
“There are two issues here: inkjet print arrays (heads) and integrated systems (printers),” says Dow’s Sutter. “The head technology is geared towards smaller drop sizes (for resolution), more nozzles per head (for throughput), and greater reliability (reduced clogging and downtime). Even at 1-pL drop sizes, the printed resolution probably would not be in the 1 μm range, unless special tricks are used. Either the ink must deposit with extremely high aspect ratio (as with a hot melt ink) or the surface energy of the substrate must be modified such that the ink ‘beads up.’ This trick is used to image high resolution transistor elements, but it requires a double printing process. For flexible displays, resolution of 5 to 10 μm or smaller will be required for active elements like TFTs. This may be in the range of 1-pL heads, but the ink characteristics and the substrate properties all need to be tuned together. The problem with going to small drop sizes for resolution is that you need more nozzles and/or more heads to print at a reasonable speed. Reliability is paramount, since one defect in an electronic device could be fatal. Recirculating ink supplies in the head, drop monitoring during printing and redundancy in nozzles can help with this.”
When it comes to throughput, Sutter thinks high-volume for ink-jetted PCBs would be two 18x24-in, two-sided panels per minute on a single printer. He says this is what’s needed to compete with other techniques such as laser imaging. Because the substrate size is large for PCBs, the application would need a printer with lots of heads to cover the entire substrate in a reasonable time. For ink-jetted solar cells, production machines would have to produce about 1,500 to 1,800 6x6-in. cells per hour.
Dow’s Sutter expects to see the ink jet industry make progress on a variety of fronts in the next few years. “What I would like to see, and I think the roadmaps are probably on the same track, includes print heads that can do 130°C, have upwards of 1,500 nozzles per head, are capable of sub-1 pL drop sizes, never clog and can jet at high frequencies,” he says. “For print systems, it’s all about cost and print time per unit area. In addition, the systems should be able to dynamically scale patterns to match the piece being printed (for congruency of patterns), register to substrate datums on the fly, have a self calibration routine for the print heads, automatically inspect print quality or nozzle accuracy, and be capable of quick and easy exchange of print heads.”