It’s good to see an attempt being made by IBM and TDK to get MRAM into the mainstream by jacking up its density into the multi-Gbit domain, currently enjoyed by tradition floating point flash memory technology.
The IBM-TDK joint R&D initiative to develop Gbit density MRAM could just do the trick. But, with flash densities already at 16Gbit, and IBM suggesting the programme may take four years, it’s going to be touch and go whether the IBM-TDK project will manage to intersect the evolutionary progress of flash, which is doubling in density every couple of years.
The project is certainly needed, however. The highest density MRAM currently on the market is the 4Mbit MRAM being sold by Freescale for about $25 which is a bit-for-the-buck rate which is far from a mass market proposition when you can buy 8Gbits of traditional floating gate NAND flash for around $10..
Another technology being developed by as a flash replacement is the chalcogenide, stored-charge approach called Ovonic Unified Memory (OUM) being adopted by Numonyx, the Intel-STMicroelectronics flash joint venture, among others.
The highest density chip made using the OUM approach is 512Mbit, and that’s not on the market yet.
Memory technologies do not really make the grade unless they’re dense. For years Ramtron has persisted with ferroelectric technology but, though FeRAM has attractive characteristics, its lack of density has prevented it becoming mainstream.
The NROM approach to flash memory , the MNOS-based flash memory technology developed by Saifun and licensed to Spansion and Macronix, has never quite made it to the leading density level of traditional floating gate flash memory, though its new Quad-bit technology, offering four bits to the gate, may take it up to leading edge density.
The moves to look for alternative to the traditional floating gate flash have been prompted by two things: first the strong patent position of Toshiba and SanDisk in the technology, and second an expectation that this technology will start to become difficult, or impossible, to manufacture at the sub-45nm geometry level.
However recent research at Samsung suggests that traditional flash is possible on 30nm processes, which makes the search for a replacement a bit less urgent.
Subscribe to this blog's feed
