Project:
Study of properties of silicon-oxide-based RRAM devices for neuromorphic engineering
Description:
One of the main problems of modern microelectronics and computer systems dominated by ubiquitous embedded systems is the undesired increase in power dissipation in integrated circuits (IC). It results from the growth of the complexity and operation frequency of IC. By processing large amounts of complex data, modern processors require almost constant communication with external Random Access Memory (RAM). Thus the effective speed of IC is limited and power consumption is increased. Interest in different computing paradigm, allowing for increased performance and reduced power consumption, has been growing for many years. The most efficient processing unit known to mankind is a human brain. In real-time it can perform complicated tasks such as image or speech recognition, or other operations that allow the body to function, with a relatively low power consumption of approximately 20 W. In order to achieve such superior performance, different novel computing paradigms have been tested, e.g., in-memory computing, stochastic computing, and neuromorphic computing. Resistive RAM is an electronic component that shows properties necessary to build such systems. One can alter the structure’s resistance by changing the voltage between electrodes, which is reversible, and the stored information in non-volatile. RRAM structures can be relatively easily fabricated in a crossbar array configuration. This project aims to develop fabrication technology the RRAM structures with silicon oxide as a resistive switching material. Devices will be studied using electrical characterization based mainly on static, small-signal and impulse measurements. On this basis, device models will be developed, its underlying physical phenomena will be investigated and the potential of this technology will be verified.
Funding:
Project funded by the National Science Center
![]()
Leader:
Politechnika Warszawska
Project no.:
2023/51/D/ST7/02706
Project’s value:
999 729,00 zł