RAM wars: RRAM vs. 3D NAND flash, and the winner is…us

Within a few years, you’ll likely be carrying a smartphone, tablet or laptop with hundreds of gigabytes or even terabytes of hyper fast, non-volatile memory, thanks to two memory developments unveiled this week.

First, Samsung announced it is now mass producing three-dimensional (3D) Vertical NAND (V-NAND) chips; then start-up Crossbar said it has created a prototype of its resistive random access memory (RRAM) chip

Three-dimensional NAND takes today’s flash, which is built on a horizontal plane, and turns it sideways. Then, like microscopic memory skyscrapers, it stacks them side-by-side to create a vastly more dense chip with twice the write performance and 10 times the reliability of today’s 2D, or planar, NAND.

The most-dense process for creating silicon flash memory cells to store data on planar NAND is between 10 nanometer (nm) and 19nm in size. To give some idea of how small that is, a nanometer is one-billionth of a meter — a human hair is 3,000 times thicker than NAND flash made with 25nm process technology. There are 25 million nanometers in an inch.

NAND flash uses transistors or a charge to trap (also known as Charge Trap Flash) to store a bit of data in a silicon cell, while RRAM uses tiny conductive filaments that connect silicon layers to represent a bit of data – a digital one or a zero.

In RRAM, the top layer of silicon nitrate creates a conductive electrode, while the lower layer is non-conductive silicon oxide. A positive charge creates a filament connection between the two silicon layers, which represents a one; a negative charge breaks that filament, creating a resistive layer or a zero.

So, which memory tech wins?

Which of the two memories will dominate the non-volatile memory marketplace in five years isn’t certain, as experts have mixed opinions about how much 3D (or stackable) NAND flash can extend the life of current NAND flash technology. Some say it will grow beyond Samsung’s current 24 layers to more than 100 in the future; others believe it has only two to three generations to go, meaning the technology will hit a wall when it gets to 64 layers or so.

By contrast, RRAM starts out with an advantage. It is denser than NAND, with higher performance and endurance. That means RRAM will be able to use silicon wafers that are half the size used by current NAND flash fabricators. And, best of all, current flash fabrication plants won’t need to change their equipment to make it, according to Crossbar CEO George Minassian.

“It will cost maybe a couple million dollars in engineering costs for plants to introduce it. That’s what it is in our plan,” Minassian said. “It’s about the same cost as introducing a new [NAND flash] node, like going from 65 to 45 nanometer node.”