Materials for a Magnetic Memory in Three Dimensions

This project concerns data storage in three dimensions (3D). Conventional planar (2D) devices are expected to reach the limits of scaling within less than a decade. We will develop the materials needed for 3D memories based on magnetic shift-register devices, namely dense arrays of vertical magnetic wires in a matrix (race-track memory, IBM patent). In this concept series of bits are shifted along each wire, requiring only one read/write element per wire. Synthesis will rely largely on bottom-up routes (self-organized anodization, atomic-layer deposition, electroplating) to minimize costs. In order to minimize risks, three different strategies will be explored for coding bits: domain walls in continuous wires; solitons with transverse magnetization in a series of magnetic disks; a hybrid route with discontinuous wires with longitudinal magnetization. Two route will be explored for data shifting: magnetic field, or current (spin transfer torque). Addressing each wire will be demonstrated, using writing methods such as thermally-assisted writing, and reading methods such as a TMR junction embedded in the wire. We address call targets density (5-50Tbit/in2), and reasonable cost per Tbit (2-20€), going beyond the scalability of all-planar devices while remaining very competitive in terms of speed and energy consumption (1-10GHz with zero seek time; 10-100 pJ/bit). In all four targets, 3D magnetic memories promise to outperform Hard Disk Drives, providing more storage capacity with less energy consumption. The project brings together the relevant leading academic research groups in Europe. It aims to set Europe ahead of Asia and the USA in 3D storage, a largely unexplored area with a high potential for innovation. Two SMEs are partners, one for material development (SmartMembranes, world leader in self-organized anodized products), and the European leader in Magnetic-RAM development, Crocus Technology.