Expectations of MRAM in Comparison with Other Non-Volatile Memory Technologies

Magnetic Random Access Memory (MRAM) is often presented as the ideal, all-purpose solid-state memory of the future. Indeed, it is expected that MRAM may provide a combination of properties that at present are only found distributed over different types of memories (SRAM, EEPROM, Flash, DRAM, etc.). This is certainly a very attractive perspective; if two (or more) different memory technologies could be replaced by a single memory technology (c.q. MRAM), the development costs for a next generation of a product comprising embedded memory could decrease dramatically. However, for this to come true, it is important that the specifications of MRAM will not merely be a compromise, only suitable for certain niche markets, but that they will really be able to compete with the (future) state-of-the-art "conventional memories". In order to obtain a better insight in this matter, we have looked at (expected) properties of different non-volatile memories for one and the same technology node (0.18 microns). From this, the weaknesses and strengths of MRAM have been identified. Special attention was paid to a comparison with ferro-electric RAM (FERAM), which seems to be the most competitive alternative memory technology.