Mannerisms

“I believe floating gate will be severely compromised below 45nm”, says Cambou, “it may do 40nm but if you go below that there’ll need to be big trade-offs.”

By contrast, Mirrorbit will scale to 25nm and beyond, reckons Spansion, which has a roadmap of 45nm production in 2008/09, 32nm in 2011 and 25nm in 2013.

Spansion has had a hard struggle getting Mirrorbit, an MNOS technology developed with the Israeli company Saifun, to reach the density and speed levels of traditional floating gate flash technology.

Now, Spansion feels it has not only surpassed floating gate in performance, it offers a scaling route which is not available to Spansion’s main rivals, Samsung and Numonyx (the Intel-STMicro joint venture) which rely on traditional floating gate technology.

Numonyx and Samsung have lined up phase change memory based on chalcogenide materials as an alternative to floating gate, but that approach far from yielding a mainstream device in terms of density and speed.

“Five years ago, Mirrorbit was new”, says Cambou, “it’s not an easy technology to make. Now it’s as dense and as fast as floating gate, if not denser and faster.”

Of all Saifun’s licensees, Spansion is the only one to have got the technology to reach the point where it can produce mainstream products. Other licensees Infineon, Macronix and SMIC have not managed to get the technology to produce competitive products.

One measure of Spansion’s mastery of the technology is that its product development times have dramatically shrunk over the years. According to Spansion COO Jim Doran: “The first Mirrorbit products took four years to develop. Now we’re doing concurrent product and technology development it took three years at 90nm, 31 months at 65nm and it will be two years at 45nm.”