> Academic opportunities > PHD positions

High-performance and low-power ferroelectric memory for artificial intelligence applications using in-memory computing

Technological challenge: Emerging materials and processes for nanotechnologies and microelectronics (learn more)

Department: Département Systèmes et Circuits Intégrés Numériques (LIST)

Laboratory: Laboratoire Fonctions Innovantes pour circuits Mixtes

Start Date: 01-09-2022

Location: Grenoble

CEA Code: SL-DRT-22-0326

Contact: bastien.giraud@cea.fr

So-called "ferroelectric" memories, or FeRAM (Ferroelectric Random Access Memories), are currently emerging as a promising alternative to replace fast-access (<1µs) non-volatile memories, necessary in all computing architectures, particularly those using Flash or MRAM / RRAM. Unlike existing memories, FeRAMs encode logic states as variations in the remanent polarization of the ferroelectric oxide. In addition to demonstrating significantly better performance (high endurance, low write energy, etc.), FeRAMs present challenges (destructive readout, multi-state, etc.) which make it possible to consider breakthrough solutions, in particular for IMC (In Memory Computing). For several years, CEA has been developing FeRAM memory technology and studying the impact of microelectronic materials and processes on their performance. The current maturity of these studies at the level of the "single device" makes it possible to envisage an exploration at the memory circuit level. The successful candidate will participate in the design of memory circuits (memory cells and arrays, decoder and partitioning, etc.), analog circuits (generation of specific voltages and currents for programming, erasure and readout), and digital circuits (management of control sequences for analog and memory circuits). The designed memory circuit will be integrated into an IMC-enabled embedded system for AI (Artificial Intelligence) applications. This will enhance the scope of the work but will also enable the improvement of system performance metrics through the use of FeRAM memory.

See all positions Download the offer (.zip)

Email Bookmark and Share