Edited by Stephen J. Mraz
South Coast Photonics, www.southcoastphotonics.com
See the Camara Oscura in action at www.youtube.com/watch?v=CzGVbaQUPMo
For more information on roll-to-roll manufacturing: people.ccmr.cornell.edu/~cober/mse542/page2/files/Schwartz R2R Processing.pdf
Flexible, thin-film circuits and electronic devices are widespread in the computing and telecom industries and can be found in virtually all commercial and military vehicles across the world. They are a major technological wave that is already changing the way we live, work, and travel.
Thin-film devices formed or printed on flexible substrates include solar cells (amorphous silicon, copper-indium-gallium diselenide (CIGS), cadmium telluride (CdTe), organic photovoltaics (OPV), and dye-sensitized or Grätzel cells), lighting (OLED and PHOLED), displays and signs (LCD, OLED, and PHOLED), and computer memory (TFTS, OTFTS). So far, solar cells have been deposited on cloth used for tents, clothing, awnings, sails, and semitransparent films for window glass. And flexible displays have been developed that could serve as e-newspapers, while OLED lighting panels have been designed for walls and ceilings.
The market for printed flexible circuits and electronics, including organics, inorganics, and composites, is predicted to go from $1.9 billion this year to $57.1 billion by 2019, according to IDTechEx.com.
For flexible displays alone, market potential is $12 billion by 2017. With continued developments enabled by new FOLED (flexible organic light-emitting device) features, the potential may even be much greater, says a market report from Displaybank Universal Display Corp.
One of the most common methods used to make thin-film electronics is roll-to-roll (r2r) manufacturing, and there are currently two basic r2r manufacturing methods. Both place electronically-active materials on flexible substrates which can be metal or plastic foils. These substrates can be yards wide and miles long. Electronic-printing methods evolved from traditional printing technologies, such as those used to make newspaper. The substrate is drawn off a roll and pulled under a series of spray or inkjet heads which coat it with solutions of active materials (thin films). Features like circuit traces are inkjet printed with resolutions of 0.0008 in. or better. The coated substrate is taken up on another roll at the other end of the r2r process line. For complex electronics, the roll may undergo several such processes.
We’re talking some serious rolls
Current r2r machines use rolls of stainless-steel substrate that average 1.5-miles long, 14-in. wide, and 0.005- in. thick. Its density is about 0.28 lb/in.3, so it weighs in at more than 1,800 lb. This roll or bolt increases by 15 to 30% in weight as electronic devices are printed on it (assuming an average density of 0.036 lb of devices per ft2). But a bolt made of Kapton (polyimide), one of the thinnest substrates at 3 mils, and measuring a mile long and 15-ft wide, with a density of 0.029 lb/ft2 would tip the scales at 2,280 lb. If you load it with devices using r2r to the same density (0.036 lb/ft2), the loaded weight is 5,130 lb, or 2.25 times that of its unloaded weight.
A roll of Dow Saranex, a common substrate, is also only 3 mils thick, Still an 80-in.-wide, 9,000-ft-long roll has a 21-in. OD. This represents one of r2r’s problems: Substrates sizes are not standardized, which makes compatibility among equipment and vendors difficult.
The second method uses vacuum-based material depositions, such as chemical-vapor deposition borrowed from semiconductor manufacturing, to form active layers. These two methods can also be overlapped and hybridized.
Until now, all r2r machines have been built to make only one device, such as a solar cell, and can only handle the same-sized roll of one specific substrate, a severe limit on their versatility. A new piece of equipment, the Camara Oscura (Spanish for camera obscura, which is a darkened box or room into which an outside image is projected through a small hole) will overcome that obstacle and handle a variety of substrates and deposition methods.
The Camara Oscura (CO) will make electronic devices out of virtually any film on a wide variety of substrates. Hardened housings let it use high temperatures and pressures, as well as vacuums, which some newer thin-film processes and devices require. For example, newly developed OLED and double and triple-junction OPV call for temperatures and pressures far above or below ambient, which prevents using current r2r machines.
CO’s controlled environment should create and manage pressures from ultrahigh vacuum (less than 0.01 Torr) on through to high pressures (200 psi or greater), as well as temperatures from below ambient (21°C) to greater than 600°C.
Solar power goes r2r
Taken to the extreme, some calculate that with moderately efficient solar modules covering 0.4% of U.S. (with 1.2 to 6.2 acres generating a gigawatt-hour per year) could supply all the nation’s electricity.
Roll-to-roll manufacturing will be key to meeting this demand as it is well poised to produce next-generation cells and do it economically. It’s estimated that rst-gen cells cost $6/W, while third-gen are at $1/W.
First-generation or traditional crystalline silicon single-junction solar cells are approaching their theoretical limit of 33%. But they are expensive to make. Second-generation amorphous silicon devices aren’t as costly, thanks to r2r manufacturing and how little raw material it uses, but their efficiency now stands at 6 to 12%, with perhaps a jump to 15% expected in the coming years and potentially 20% or more in the long term. Third-generation cells, which will also take advantage of low-cost r2r, are up to 30% efficient, which could be improved to 50 to 60%. (There are concerns in the industry about exceeding the theoretical Shockley-Queisser limit of about 31% for single-junction solar-cell efficiency.)
Third-generation photovoltaics include: CdTe (cadmiumtellurium), CIGS (copper-indium-gallium-selenide), nanocrystal solar cells (quantum dots made with dye-sensitized colloidal TiO2), and organic polymer cells that convert visible and IR light to electricity.
The basic CO machine consists of a cartridge, rail, skate, heads, and station. The cartridge holds a roll of substrate on a spindle and contains all the mechanisms and guides needed to unwind the substrates and pass it across the deposition zone — a gap that exposes the substrate. The cartridge is much like a cassette tape with an open area that exposes the magnetic tape to a read/write head. The cartridge has a second spindle on which to roll up processed substrate. The skate carries cartridges to and from stations on the rail. And stations contain one or more heads which can use any of several deposition techniques to lay down film on the substrate through the deposition gap. After deposition, the substrate is rolled up on the cartridge’s second spindle.
A CO using organic vapor-phase deposition (OVPD) to make multijunction organic solar cells, for example, would work in the following way:
Once a loaded cartridge is coupled to a head in the station via the skate and rail, the machine would create a vacuum of less than 5 mTorr. Bucky Balls or fullerenes would be transformed from solid to gas (sublimed) and the C60 vapor entrained with 400°C nitrogen. The head would hold all the equipment and gas sources needed for this task. The C60-laden nitrogen is sent by the head through the deposition gap, which would measure about 5 × 1 ft, and onto the substrate. The substrate travels stepwise horizontally across this gap. Each 5 x 1-ft section has film deposited on it at 10 Å/sec as C60 condenses on the substrate, a rate consistent with fault-free OVPD films. It would take 200 sec, assuming 10 Åsec/layer, for that 5 × 1-ft section to form. If this process were to continue 24/7 for a year, a CO tool could turn out 0.8 MW of solar cells annually, assuming 10 W/ft2.
If more than one layer is needed, the sealed cartridge and head are flushed of contaminants between deposition processes. The skate then moves the cartridge to another station or rotates it to another head in the same station, and the second film layer is deposited at all the places it is needed on the roll of substrate. Once the device strand (an array of electronic devices on the substrate) is built, including moisture and oxygen barriers, the skate can transport the cartridge to a specific shipping dock where the cartridge is closed, environmentally sealed, and shipped to a vendor’s postprocessing facility.
The technical sophistication and resources needed to design and build a CO machine are not inconsiderable, but there are no technical or material challenges that can’t be met with current state-of-the-art know-how. Current gigawatt tools cost about $1.7 million, with entire plants coming in at $50 million. This is our current target for developing and building a plant with one CO tool capable of turning out a gigawatt of photovoltaics a year. And while some plants may be devoted to making a specific device, others could be regional centers for producing different devices for several thin-film-device firms.