Page 39

EETE JUN 2014

Fine prints and gigabits of data: on paper By Julien Happich Printed electronics is about reaching the most cost effective way to integrate and mass produce electronic components and systems into very thin flexible substrates. Often, the idea is to get away from process-intensive hardsilicon designs and adopt well established printing processes such as inkjet or roll-to-roll to apply electronics onto plastic foils or even paper, ready for innovative smart packaging applications. A lot of printed or flexible electronic components such as sensors, displays, some logic blocks and passive components, or conductors have been demonstrated in research, some are even being commercialized. But for more complex system designs, or simply for cheap and flexible data storage, memory is still a challenge. A PhD student from the Institute of Photonics and Optoelectronics & Department of Electrical Engineering at the National Taiwan University, Der-Hsien Lien is proposing to convert paper into a truly dual memory medium. According to his research, a single sheet of paper could combine printed text or graphics as for any conventional reading paper, together with embedded printed electronics memory for more information storage than would normally fit even in very fine prints. Currently a visiting student at UC Berkeley, Der- Hsien Lien uses inkjet printing to stack a specially formulated insulator material including TiO2 nanoparticles, between two electrodes, namely silver and a carbon layer. The so-called Resistive Random Access Memory (RRAM) bit obtained for each stack is operated by changing the resistances of the insulator material. Applying a voltage across the memory dots, one can turn the resistive states on and off for “0” and “1” binary values. These resistive states and the switching voltage window can be tuned to a few volts based on the insulator material’s thickness. With individual dots being tested for endurance, the memory’s retention property was proven at temperatures up to 150°C and with mechanical flexures at a bending radius of 10mm. So how many bits of memory this printing process could yield on a A4 sheet of paper? At the current printing resolution he is using, with a dot size of circa 50μm and a pitch resolution of 25μm, Der- Hsien Lien estimates that one bit occupies a square of about 100μm on each side. That would translate in 104 bits/cm2. With that density, a fully printed A4 sheet of paper could hold 6.237x106 bits. With state-of-the art printing resolutions such as super fine inkjet (SIJ) technology and providing that the required specialized Fabrication and geometry of paper RRAM: (a) Schematic diagram of the fabrication process for the resistive paper memory device. (b) A close-up photograph showing fine prints and arrays of memory dots. (c) A zoom-in optical image from b. (d) A cross-sectional scanning electron microscopy image of the paper-based memory. inks would be compatible, one could achieve a dot resolution of 1μm and nearly the same pitch width, says Der-Hsien. This could increase the density by roughly 2500 times, yield up to several Gbits per sheet of A4 paper or tens of Mbits per square centimeter. Based on materials and process manufacturing costs, the PhD student figures out that a paper-based RRAM would cost about 0.0003 cent/bit, versus 3 cent/bit for a fully printed RFID tag (1-bit) fabricated on PET substrate, or versus a 96-bit creditcard sized ID paper tag at around 0.021 cent/bit. In future research, Der-Hsien envisages connecting arrays of dots with cross-bar electrodes. Such a memory architecture would allow the researcher to stack memory layers into 3D arrays to further increase density. Other activities of the lab include the development of memory drivers to design a complete all-printed memory solution. This sort of printed paper memory would combine very well with other printable devices, for example to be integrated with RFID for ticketing. But Der-Hsien also thinks that single dots of memory could well be used as arrays of individually pigmented dots to be read like QR codes. By combining the pigments (dark and light) and their resistive states (high and low), the stored information on such codes could be doubled and secured, he explained EE Times Europe. A specially designed reader would have to be designed to access the memory dots. Safe and secure disposal of the memory is as simple as shredding or recycling a stripe of paper. Several patents are pending in Taiwan and several others will be applied for in the US too. This research will be one of the key highlights of the 2014 Symposia on VLSI Technology and Circuits to take place in Hawai, June 9-12. www.electronics-eetimes.com Electronic Engineering Times Europe June 2014 39


EETE JUN 2014
To see the actual publication please follow the link above