Addressing retention and write bottlenecks for sub-10nm MRAM magnetic memory cells

Magnetic random access memories (MRAMs) are now on the market, in FPGAs or cache memory blocks. The concept of Perpendicular Shape Anisotropy MRAM (PSA) has been proposed to expand the MRAM market: the usual ultra-thin storage layer being replaced by a vertical magnetic pillar, allowing retention to be maintained at <10nm size and at high temperature. Now that its read-write proof-of-concept has been achieved, the objective of this PhD is to address the bottlenecks that prevent its technological adoption, related to dense manufacturing and low-current writing. We propose three concepts to globally solve this problem, concerning the manufacturing technology method, and the shape and material engineering of the magnetic information retention cell. The project consists in exploring the fundamental physics of spintronics in these memory points, materials science with innovative techniques and their shaping in state-of-the-art clean rooms. This involves mostly experimental work, but also numerical simulation to refine their design. The thesis aims to magnetically qualify the pillars, and provide proof of concept of read/write operations. More broadly, it will provide knowledge and technology suitable for the implementation of other spintronic concepts in 3D architecture in the future, for example, for logic or artificial intelligence purposes.

Facebook

Twitter