TORONTO — There are a number of emerging memory technologies that have been in that state for some time, with the jury still out as to whether they will ever see mainstream adoption or simply fade away as other technologies solve the same problem more effectively.
Phase-change memory (PCM) is certainly one of those technologies, with a history dating back to the 1960s. A non-volatile, random-access memory, it also known as PRAM, among other names. Companies including Intel, Samsung, and Micron have all developed PCM technology, with the latter claiming to offer the first PCM in high-volume production back in mid-2012 with the availability of a 45 nm multi-chip module for mobile devices.
One of the ongoing challenges for PCM has been reliability, a topic of discussion as recently as this year. Ron Neale, a regular EE Times contributor, noted in July 2014, “The history and literature of the development of PCM devices is replete with examples of claims of levels of performance and characteristics that, while achieved in the laboratory, are never reproduced in production.”
Gartner analyst Brady Wang tells me the research firm regards PCM as a replacement of NOR flash, which is a declining market. PCM is challenging to compete or replace NAND flash or DRAM, he says. But while PCM can provide better performance than NAND flash in speed and simplify controller design, its cost is still much higher than NAND's. A critical factor is that 3D technology will extend NAND’s lifespan for several more years.
As for replacing DRAM, Wang says PCM is still much slower. Overall, Gartner expects the PCM market to be limited even if some vendors continue to work on it for niche applications. But with Micron having decided to stop further development of PCM, Wang says it’s highly likely that PCM will be a niche memory and even disappear from the market.
Jim Handy, principal analyst with Objective Analysis, offers a similar prediction: “PCM is more expensive than NAND, and once NAND goes 3D, PCM will be relatively too expensive.” While PCM may offer some performance benefits, it’s price that gets market share. Handy says Samsung and Micron abandoning major PCM initiatives is a clear indication that neither company sees the technology going mainstream.
Even the niche applications for PCM may be threatened. Handy notes that BAE Systems has a PCM (in this case, C-RAM) device, which tolerates radiation extremely well, making it useful for systems in space, an environment that NAND and DRAM don’t handle very well.
However, Adesto Technologies recently announced a new type of memory that has been tested and proven to endure the sterilization required for medical applications. While the company’s focus has been on low-power memory, it has also spent a number of years investigating the tolerance of its proprietary memory to radiation. Initial applications for Adesto’s memory are aimed at medical uses where wearable devices must be sterilized with radiation, but Handy believes it could easily supplant PCM for space applications.
But while some analysts may not be bullish on the future of PCM, IBM still sees potential. Earlier this year, it demonstrated how to integrate PCM into a solid-state memory hierarchy. Big Blue’s proposed server system eliminates the hard-disk drive (HDD) in favor of a memory hierarchy using both flash-based arrays and PCM-based arrays.
And PCM might have a future with optoelectronics. A team at the University of Oxford and the University of Exeter have created an “optoelectronic framework” to employ PCM materials in digital displays, optoelectronic memory devices, optical switches, and modulators for intra- and inter-chip communications.