IBM, Infineon to research new high-speed memory

By Steve Malone

Posted on 25 May 2005 at 10:31

IBM has teamed up with chip developers Macronix and Infineon to form a joint research project to investigate the possibilities of a new form of computer memory technology known as phase-change memory (PCM).

PCM stores data by changing the state of a special material from an amorphous to a crystalline structure, rather than storing data as an electrical charge. Although still in the experimental stage, the technology could eventually be used for high speed, high density data storage. As with Flash memory, the data is maintained even when power is turned off.

Phase change technology may prove the key to reducing the size of memory cells to sub micron sizes. In theory, phase change technology suffers less from the problems of heat dissipation and quantum physics than today's designs

The companies have announced that they will contribute a total of 20-25 staff to the research work that will take place at the famous IBM's Thomas Watson Research Center and IBM's Almaden Research Lab in San Jose, CA.

While the announcement might sound ground breaking, Philips announced its own discoveries concerning phase-change memory in the journal Nature back in March.

As Philips points out, phase-change materials are already in use in such things as rewriteable DVDs. However, in the case of DVDs, it is the reflectivity of the material that changes not the state. Here the laser heats the material to the required temperature in order to switch it between its amorphous and crystalline phases and reads the change in its reflectivity back as data.

The Philips solid-state memory cell employs similar phase-change materials laid down as an ultra-thin film on the surface of a silicon chip. Instead of using a laser, it uses an electric current to switch it between phases and to detect the resultant change in its electrical resistance.

Phase change memory has been held back because of the high voltages required to make the switch. Company researchers found that by using a new doping material, Antimony/Tellurium, it lowers the field strength to 14V/μm, sufficiently low for phase change memory to start looking a practical proposition for future memory devices.

It is likely to be several years before any developments make it out of the lab into commercial applications. However, the fact that two of the world's biggest technology companies think it is worth investigating shows that phase change may have a future