Embedded MRAM Technology

Avalanche’s next generation embedded MRAM technology replaces eFlash and eSRAM which are increasingly facing scaling issues beyond 28nm. While eFlash and eSRAM are widely adopted in modern SoCs, both technologies are facing severe scaling challenges. eFlash scaling is increasingly limited by channel length and endurance beyond 28/40nm nodes, whilst eSRAM suffers from leakage and large cell size (100-300F2).

Avalanche’s embedded STT-MRAM Macros are enabled using standard manufacturing processes in existing CMOS fabs on top of the metal layer with two additional masks.

Avalanche’s Embedded Perpendicular STT-MRAM Technology Benefits

  • Flexible MTJ placement between any two metal layers (including CT and M1) w/o process change below/above MTJ layers – scalable to sub-10nm nodes
  • Thermally robust MTJ – no TMR and Vc degradation @400°C/150 min
  • Excellent MTJ data retention – including 260°C solder reflow support
  • Immunity to high stray magnetic field
  • Chemical-damage-free MTJ etch with single mask – improved performance and low cost

Avalanche’s Embedded Perpendicular STT-MRAM Technology Attributes

Avalanche’s Perpendicular Spin Torque Transfer MRAM technology delivers the next generation scalable embedded unified memory architecture with the benefits of low latency, low power, infinite endurance, high performance and scalability to lower geometry nodes.

Avalanche’s embedded Unified memory Architecture Perpendicular STT-MRAM enabling a unified Code and Data space solutions are available via to foundry partners enabling The next generations of GPUs, MCUs, DSPs, ASSPs and ASICs as eFlash and eSRAM replacements.