Samsung to start mass producing 16 gb Nand flash

Samsung Electronics will be the first to begin mass producing 16 gigabit (Gb) NAND flash, so far the highest capacity memory chip now available. The company said it will fabricate the devices in 51 nanometers (nm), the finest process technology to be used in memory mass production to date.

“In rolling out the densest NAND flash in the world, we are throwing open the gates to a much wider playing field for flash-driven consumer electronics,” said Jim Elliott, director, flash marketing, Samsung Semiconductor. “To minimize production costs and improve performance, we have applied the finest process technology a ‘half generation’ ahead of the industry, which is introducing 55nm and higher.”

Samsung’s 51nm NAND flash chips can be produced 60 percent more efficiently than those produced with 60nm process technology.

The new 16Gb chip which has a multi-level cell (MLC) structure can facilitate capacity expansion by offering 16 gigabytes (GBs) of memory in a single memory card. Furthermore, by applying the new process technology, Samsung has accelerated the chip’s read and write speeds by approximately 80 percent over current MLC data processing speeds.

Intel's phase-change memories

Intel is working on new type of memories called phase-change memory technology, called PRAM by Intel and PCM. This technology is set to sample in the first half of this year. They plan to ship the first PRAM modules as a straight-ahead NOR flash replacement, so that they can test the design before trying to move it up the ladder in the memory hierarchy.

The memories are claimed to have much higher number of read-write cycles (100 million) than flash, as well as a potential 10 years' worth of data retention. NOR flash is typically used as program storage memory for mobile devices like cell phones, while more durable but slower (for random read access, but not sequential bursts) NAND flash is used for mass storage in devices like the iPod nano.

The phase-change memories claimed to be potential flash-killer are actively researched by Intel, IBM, Samsung, Hitachi and many others. PCM's advantages over flash are numerous, and in fact the technology gets us closer to the Holy Grail of computer memory, i.e., a nonvolatile medium with a small cell size and fast access times, sort of like DRAM but without the volatility and the refresh circuitry. However, PRAM doesn't quite get us all the way there.

Competing technology from Hitachi boasts a 20ns read latency. This is much better than the 50ns to 90ns read latency typical of flash memory, but it's not even close to DDR2's ~3ns latency. If Intel's PRAM is in the same ballpark as Samsung's technology, then it won't be used as the main memory on your computer anytime soon.

Samsung has already demonstrated a 512Mbit PRAM chip, and has announced plans to offer a full lineup of PRAM products sometime in 2008. The company began sampling 256Mbit and 512Mbit parts produced on a 90nm process to mobile phone manufacturers at the end of last month.