Flexible re-programmable 256-bit transistor-based and 1-kbit diode-based non-volatile memory

January 10, 2013 // By Julien Happich
The EU-funded eMbedded Organic Memory Arrays (MOMA) project has announced the largest re-programmable non-volatile memory arrays yet produced on flexible substrates.

Its 256-bit transistor-based and 1-kbit diode-based arrays exhibit state-of-the-art performance in numerous key memory parameters, plus excellent production yields, according to the project team.

In recent years, flexible electronics technology has developed rapidly and is now on the verge of commercialization, promising new applications from smart food packaging to wearable health monitors. All these applications require programmable non-volatile memory (similar to the well-known Flash memory), and the more complex the application, the more memory it needs. However, flexible memory development hadn't kept pace with other areas of flexible electronics, limiting market opportunities. So the European Union launched the MOMA project in 2010 to develop low-cost, reprogrammable polymer memories and integrate them with thin-film transistors on flexible plastic substrates.

The MOMA consortium brings together commercial and research partners from across the memory value chain. These partners are chip maker STMicroelectronics, materials supplier Solvay Specialty Polymers, innovation centers imec and Holst Centre plus Catholic University of Louvain and University of Groningen.

The project focuses on ferroelectric memories based on soluble ferroelectric polymers and organic and oxide semiconductors. It has successfully developed new grades of memory materials that can be tailored to specific applications and deposition techniques (including spin-coating, inkjet printing and imprinting). The project partners used these materials to develop memory arrays based on both transistors and diodes, optimizing production processes to deliver high yields and excellent memory performance.

"We're looking to minimize the total cost of the application. The reprogrammable polymer memories currently available require expensive silicon MOSFET technology. It makes no sense reducing cost-per-bit if you then need expensive external silicon components to get it working. Hence we concentrated on ferroelectric memories that are based on resistance rather than charge as these allow for simpler read-out electronics, reducing overall costs" said Albert van Breemen, Senior Researcher at Holst Centre and member of the MOMA project team.

With its 1-kbit (32x32) array, MOMA has delivered the largest flexible memory arrays to