
ReRAM could perform as battery, says Leti
Leti has developed devices that can operate as energy-storage elements as well as memory, depending on the applied bias. Such hybrid dual-behavior ReRAMs can be compatible with CMOS fabrication, Leti said.
In-memory computing is a key area of research, particularly for neural network implementations and because ReRAM batteries are highly scalable and dynamically allocatable, they could be used to augment peak performance demands.
“Our work attempts to tackle the modern demand of energy-efficient computation with an innovative approach, radically different from state-of-the-art solutions” said Gaël Pillonet, a senior engineer at CEA-Leti.
The research results were reported in a recent paper in Advanced Electronic Materials, entitled ‘In Memory Energy Application for Resistive Random-Access Memory’. The ReRAM material is tantalum pentoxide sandwiched between copper telluride (CuTe2Ge) and tungsten electrodes.
The paper derives energy storage capabilities that amount to 3.5 picojoule per square micron.
Future work would aim to quantify the output voltage, coloumbic efficiency, and power-density capability in discharge mode as well as the optimized peripheral circuitry to support dual operation as memory and storage. Designs in which memory and energy storage performance are co-optimized will be explored, Leti said.
Related links and articles:
News articles:
Weebit Nano puts silicon-oxide ReRAM on 28nm FDSOI
Researchers want to ‘copy and paste’ the brain into memory
SureCore to help UK ReRAM startup come to market
Unisantis unveils vertical, capacitor-less DRAM
Chinese ReRAM startup raises $100 million
