Developers have made major progress in the technology to manufacture printed or flexible circuits, sensors, batteries and displays. But frankly it’s been hard to build applications with much market pull without logic or memory as well, and those have been much harder to make. However, now printed memory and solutions for integrating conventional silicon die into flexible systems are edging into production, to potentially improve performance for a wider range of applications. On the display side, easily integrated printed or flexible transparent conductive films for touch screens are starting to see some market traction.
Yole Développement projects the market for printed and flexible electronics will remain a modest ~$176 million this year, but will see 27 percent CAGR to ~$950 million by 2020, driven largely by printed layers integrated into large OLED displays.
Thinning patterned die makes flexible silicon on polymer
One interesting solution to add performance to flexible electronics could be an open platform for making flexible silicon die. American Semiconductor proposes drastically thinning conventional fabricated silicon wafers, and coating them with a combination of polymers. The resultant silicon-on-polymer approach protects and eases handling of the ultra-thin die, says CEO Doug Hackler, who will discuss the technology in a program on such hybrid solutions in the emerging market program series at SEMICON West in San Francisco in July. He reports user interest for large area distributed sensing systems that include ICs within structural composites in aircraft bodies to monitor stress, for bio sensors that conform to the body, for RF for wireless data transmission from printed sensors, and for drivers for flexible displays.
The company has qualified TowerJazz’s 130nm process to make SOI CMOS for its initial flexible standard microcontroller, and has worked with the foundry to establish design rules to make an open platform for other designers to create their own flexible chips. American Semiconductor thins these fabricated wafers by standard methods down to about ~40µm. “And then from <40µm it gets trickier and more proprietary,” says Hackler. But once these flexible silicon-on-polymer die are diced and released, they can be handled pretty much like standard chips. “The dicing and release are a little different, but once the die are on tape, then it appears feasible to do traditional pick and place,” he says, noting the company intends to use printed connections instead of bonding wire or solder bumps. After assembly on a flexible substrate, perhaps by a pick-and-place module integrated on a roll-to-roll printing tool, the devices would typically be laminated or overcoated for additional protection. The company plans to follow its flexible microcontroller with a standard analog/digital converter to take in sensor data, and an RF IC to send out the data.
Innovative solutions for assembling silicon on flexible substrates move towards production
Packaging and assembling tiny thin die on flexible substrates remains a challenge, but multiple suppliers are making progress towards solutions that are starting to edge into commercial production. One approach particularly suitable for attaching sensors to the body is the spring-like stretchy wiring developed by MC10 for attaching thin silicon die to flexible substrates, for everything from wearable heart rate and fitness monitors to sensor membranes that can be implanted directly on organs inside the body. VP of R&D Kevin Dowling reports the company’s first commercial application is in a soft skullcap from Reebok that uses flexibly connected motion sensors to measure impacts to the head.
Tiny die size could also help with both cost and attachment of rigid die to conformable substrates, although handling and assembling them then becomes more of an issue. Terepac Corp. CTO Jayna Sheats notes that plenty of logic for simple controls could be very tiny and low cost — microprocessors with ~8000 transistors like the Z-80 generation currently used for many embedded control applications take up <70µm2 of silicon with 90nm design rules, for millions on a wafer. But the die are too tiny to make the input/output connections or to handle with traditional pick and place for packaging and assembly. So Terepac proposes a photochemical assembly process instead, picking up an array of thinned and diced chips with a sticky printhead, positioning the chips over the substrate with a tool similar to a proximity aligner, and vaporizing the proprietary polymer/adhesive behind each selected chip with a combination of heat and UV so it falls into the desired position. Chips can then be attached to the flexible substrate by conductive adhesive, electroplating, or printed connections. The company is working with equipment manufacturing partners including Rockwell International to construct manufacturing facilities for customers with products for the Internet of Things.
Jabil reports progress in low temperature attachment technologies for use with heat-sensitive flexible substrates. And Sandia National Lab reports it’s come up with a solution for the common researchers’ problem in this field of how to build prototypes of flexible systems when the necessary ultra thin chips only come in costly wafer-level volumes. Researchers there have figured out how to thin off-the-shelf single die for developing flexible systems.
Printed memory targets low-cost, high-volume applications
Thin Film Electronics, meanwhile, is developing systems that use its simple, low cost printed memory. The company’s 20-bit memory can be produced in volume for under ~$0.05, targeted at applications like consumer packaging, with volumes of billions of units a year where roll-to-roll printing makes most sense, says Chandrasekhar Durisety, assistant director, North America, who will give an update on the company’s progress towards commercialization at the session. Thin Film is introducing a next generation of passive array memory, in 4×4 (16 bit), 5×5 (25 bit) or 6×6 (36-bit) options, a more conventional format with fewer pads at higher density for easier addressing than its initial 20-bit in-line architecture.
The company is working with a global consumer product maker on using low-cost printed memory to make brand authentication cost effective for a wide range of lower-priced products. It’s also working with major flexible packaging supplier Bemis Company Inc. on sensor labels for food, healthcare and consumer products that can collect and wirelessly communicate sensor information at roughly the same low cost as current color-changing chemical indicators. The digital system under development — with Thin Film’s printed memory, an electrochromic display from Acreo, and printed logic technology from PARC — stores data when the temperature exceeds a certain range, to indicate more clearly than a color gradient can whether the product is usable or not.
Thin Film aims to add electronics to applications that currently don’t use them, to add simple intelligence at prices far below those possible with silicon, such as low-cost brand authentication, temperature sensors on packaging, or simple electronics in toys. “Silicon die could add significant capability to printed electronics. But with fabrication and assembly it would likely be more expensive than either silicon or printed electronics alone,” suggests Durisety.”
Market starts to develop for printed/flexible ITO replacements
Another key potential market for printed/flexible electronics is next-generation transparent conductive film to replace brittle and expensive indium tin oxide in touch screens and displays, lighting, and photovoltaics. Touch Display Research says the market for non-ITO transparent conductors will be about $206 million this year, and grow to some $4 billion by 2020. “High demand for touchscreens for notebook and PC size displays has created a shortage of ITO touch sensors since the end of last year to drive more interest in these technologies, and the more flexible and potentially cheaper replacement technologies are getting more mature,” notes Jennifer Colegrove, president and analyst, who will speak at the FlexTech workshop on transparent conductors. She notes that Atmel, FUJIFILM, Unipixel and Cambrios are all in some phase of production.
There is, however, a confusing range of contending options for processes and materials for these films. Applied Materials has interesting progress in its roll-to-roll deposition technology, while FUJIFILM Dimatix targets ink jet printing the materials, and NovaCentrix offers rapid thermal curing that doesn’t heat the substrate. Materials options range from nano metal wires at Cambrios Technology, Carestream and Sinovia, to embossed and metalized patterns from Unipixel, to carbon nanotubes at Brewer Science and graphene at Nanotech Biomachines.
These and other speakers will talk about the challenges and solutions to move printed/flexible electronics into real markets at SEMICON West’s emerging technology programs, July 9-11 in San Francisco.
· Mon, July 8: Market Symposium, SF Marriott Marquis, Keynote: “New Directions in Flexible and Printed Electronics,” Dr. Ross Bringans, VP at PARC (1:00-5:30pm)
· Tue, July 9: Materials Growth Opportunities at Both Ends of the Spectrum (1:30-3:30pm)
· Wed, July 10: FlexTech Alliance Workshop: Emerging Materials and Processes for Transparent Conductors, SF Marriott Marquis (10:00am-5:00pm)
· Thur, July 11: Integrating Conventional Silicon in Flexible Electronics at the Extreme Electronics TechXPOT, South Hall (10:30am-1:10pm)
For more information, visit www.semiconwest.org/SessionsEvents/PlasticElectronics
Paula Doe is an analyst for advanced technologies for the global trade association SEMI.