Coating and deposition: Key for FPOE
Developers of flexible, printed, and organic electronics (FPOE) products face significant challenges in scaling to true volume production. And coating and deposition play a key role, regardless of the technique used to deposit layers of more conventional thin films or new types of nanoscale functional inks.
By Tom Cheyney
Printing numeric minidisplays for smart cards, RFID tags, and other relatively simple electronic devices on flexible substrates is one thing. But developing and eventually manufacturing integrated, low-power, full-color flexible display modules, plastic memory and logic circuits, and other more sophisticated components is a much more daunting task. Production schemes range from roll-to-roll to batch/sheet toolsets; operate in vacuum and/or atmospheric conditions as well as elevated or ambient process temperatures; use plastic, stainless-steel-foil, or even paper substrates; and employ semiconductor-like equipment or gear adapted from the print industry.
Whatever the approach, many developers and manufacturers of flexible, printed, and organic electronics (FPOE) products face significant challenges in scaling to true volume production and realizing the kind of multibillion-dollar markets that a growing number of forecasters predict. Just as in other thin-film manufacturing industries, coating and deposition play a key role in the creation of FPOE devices, whether the process uses CVD, PVD, or spray tools, or incorporates inkjet-head or other graphic-arts techniques to deposit layers of nanoscale functional inks.
The flexible display vision
The Flexible Display Center at Arizona State University, where academic, government, and industrial research partners are developing manufacturable high-information-content flexible displays, faces a difficult and complex challenge matrix. Yet since its inception in early 2005, the center has made great strides. Materials have been developed and integrated, and the process has largely been stabilized on its low-temperature amorphous-silicon (a-Si) TFT pilot line. Shawn O’Rourke, the center’s director of operations, discusses some of the coating and deposition challenges that come with the flexible display territory.
Aveso uses printed electronics techniques and electrochromic inks to make its plastic display.
“The architecture that you’re going to employ and the process temperature that it can be done at play a major role in the processing defectivity and performance of the materials that go down. Some of things we see in the bond/debond approach that we have are stability of the adhesives with temperature, distortion of the PEN plastic substrates as a function of temperature, as well as the quality of the thin films you can put down at lower temperatures. You have to develop new recipes for those because step coverage may not be as good, so you wind up with higher defectivity, lower yield, and lower- performing devices overall.
“A more global challenge has been the adhesive development that’s gone on here, the coated adhesive that we have for bonding stainless steel foil onto our carrier substrate that allows us to get 100% handling yield through the fab. That’s been huge. On our six-inch line, we use a standard bowl coater from Rite-Track, and for our Gen II line, we’re evaluating the EVG spray coating system for the adhesives. Another challenge is doing high-quality device films at lower temperatures (180° to 185°C) and getting a stable, high-performing transistor process that’s capable of driving multiple different electro-optics.”
Polymer Vision is already producing rollable e-paper display modules “by the thousands” at its Southampton, UK, fab, in its hybrid semiconductor process employing a carrier substrate approach perfected at the company’s R&D line in Eindhoven, the Netherlands. “The most critical layers are, of course, the dielectric and semiconducting layers,” according to CTO Edzer Huitema. “Both are organic and applied from solution and are subsequently patterned by photolithography. The key for both layers is uniformity and very low defectivity rates.”
The technologist cites Polymer Vision’s low temperature budget as a big advantage. “This enables processing on the thin (25-micron) plastic substrates without registration issues. Our feature size is 5 micron with a registration that is better than 1 micron over the complete display area. We have changed the materials of a number of critical dielectric and semiconductor layers, thereby changing all PECVD steps into spin-coat (or spray coat) steps. The key objective when changing these materials was low-temperature processing and the use of materials that can handle more strain (mechanical flexibility), but the elimination of PECVD is a nice extra win.”
The promise of printed electronics
Seeking the promised land of lower manufacturing and facility costs, higher throughput, and larger area processing, a host of companies have made significant progress in fabricating printed devices through the use of nanoparticle inks deposited by gravure, flexographic, and other offset techniques. Many difficult challenges remain though, as Dan Gamota, Motorola’s director of printed electronics, explains: “Key challenges are printing registration and printing resolution at high throughput [not inches per day but feet per minute] and substrate stability over temperature [coefficient of thermal expansion, modulus, glass-transition temperaturesTgetc.]. Organic materials typically are easier to solution-process, but suffer from humidity, air stability, etc. Inorganic materials suffer from higher temperature processing and therefore require Tg materials.”
Gamota points to some success stories as well as areas for improvement, particularly in the production equipment infrastructure. “In 2003, Motorola printed over three miles of ring oscillators using a convention printing platform. Since then we have worked with printing equipment suppliers to reduce resolution of features to below 50 microns [goal is 5 microns] and to improve registration parameters below 75 microns [goal is 1 micron].
“To achieve these dimensions we need more equipment suppliers to participate in the development of printing platforms, inline, offline testing hardware/software, and finer resolution printing consumables, such as gravure cylinder and flexography plates. The most promising manufacturing platform would be modular in nature, comprising a flexo cell with a gravure cell. Today, very few combination platforms are offered. We are engaging internal and external customers for development of prototype products and concepts.”
What does Gamota want in an ideal turnkey equipment solution for R2R processing of printed electronics? “The turnkey system must have an acceptable resolution at a reasonable throughput to provide the economies of scale necessary to commercialize this technology and products. Investments must be made to develop new equipment that meets the requirements of printed electronics inks and design features. Inline process control and end-of-line high-volume circuit (printed ICs) and product testing are absent.”
Printing for smart cards
One company already manufacturing products using a largely printed electronics approach is Aveso, a spin-off of Dow Chemical. President and CEO Dennis Brestovansky describes the company’s process for fabricating plastic display modules for use in next-generation smart credit cards and other applications: “Our displays comprise a single layer of its proprietary electrochromic ‘ink’ sandwiched between a conductive backplane and clear conductive top plane [ITO-coated polyester]. The manufacturing process, material set, and printing capacity used to produce the displays are nearly identical to those utilized for printed electronic devices, such as electroluminescent lamps and membrane switches.
“The company uses standard printing techniques to deposit the ink. For products such as our Primero 6/7 module, sheet-fed processing is used. Layers are either screen- or stencil-printed, dispensed, or laminated. Ink thickness is less than 100 microns in a typical device. In addition to stencil printing, the ink can be printed via gravure and flexography in roll-to-roll processes.”
The model 5200 stepper from Azores Corp. provides photo exposure with automated distortion compensation for flexible substrates. Photo courtesy of The Flexible Display Center at Arizona State University
Brestovansky notes that, like the semiconductor and flat-panel display industries, reducing defectivity, controlling the process, and enhancing yield are also crucial in the manufacturing of Aveso’s printed displays. “As always, yield is a critical determinant of cost. In our case, the most critical challenges that we overcame were related to defect-free printing of a continuous, void-free layer of electrochromic material, and ensuring good electrical contact with top and bottom electrodes during subsequent lamination steps used to seal and bond active layers of the device together after ink deposition.”
Lessons from PV production
One industry where sophisticated roll-to-roll manufacturing has become relatively mature is thin-film photovoltaics. PowerFilm and Uni-Solar/ECD have fabricated kilometers of flexible single and multijunction a-Si-type PV cells and modules, and both companies are leveraging their experience in FPOE programs.
“PowerFilm’s 18 years of experience gave us a tremendous advantage in starting up our vacuum deposition processes for a-Si TFTs,” says Carl Taussig, program manager for Hewlett-Packard’s joint project with the solar company to develop a R2R process for manufacturing active-matrix electrophoretic display backplanes on plastic, using plasma processing and self-aligned imprint lithography (SAIL). “However, there are a couple of requirements for SAIL-fabricated TFTs that are different from those of a solar cell. TFTs are intolerant of mega-ohm shunt defects, whereas these cause no problem to PV. Also, the level of particulate contamination that can be tolerated for a R2R imprint process is much less than what is acceptable for PV.
“PowerFilm has built a drum coater that has greatly reduced particulates by eliminating sliding contact. In less than 6 months, they brought the coater online and deposited the first functioning TFTs. The R2R imprint process requires extremely uniform coating of the UV curable polymer. This requires extremely high precision in the gravure coating process that we use and inline measurement of a number of key process parameters. At this point we are able to control uniformity on a 100mm web to approximately 100nm.”
Because HP’s process, like many others, is a vacuum/nonvacuum hybrid, Taussig’s team faces the challenge of how to “hand off” between atmospheric and in-vacuum steps. “Plastic substrates are hygroscopic. Although the SAIL process is immune to problems with alignment, large swelling of the substrate can result in delamination of the film stack. It is not a huge problem; we have developed a time X temperature budget for immersion in wet processes. It is most important to completely dry the substrate before vacuum processing, particularly before the initial deposition of the TFT stack. The partial pressure of water in the PECVD system can affect the properties of the devices.”
Although photolithography remains a big challenge for R2R on flexible substrates, CVD and PVD processes are “pretty straightforward” on Uni-Solar/ECD’s 300-ft-long continuous-feed production lines, according to Vincent Cannella, senior scientist at Uni-Solar. “We’ve done depositions on flexible webs for more than 20 years, and we do these multilayer, triple-junction solar cells with nine active layers in roughly a micron of total thickness, and we have great control. From our vantage point, we really don’t see problems with high-volume control of depositions, whether it’s sputtering or PECVD.”
Uni-Solar has also found a way to deal with the aforementioned “hand-off” issue, enabled in part by a proprietary air-to-vacuum gas gauge. “We worked with another company to develop a process that starts in atmospheric wet coating and goes on to high-vacuum metal deposition, then comes back out of vacuum to atmospheric pressure lamination,” explains Cannella. “The front surface of the web material, the active surface, is not touched at any point in that process until it’s laminated.
“Again, we don’t see these issues as big problems,” he concludes. “We think that the solutions exist for them today.”