Intel moves to 0.13 micron, copper
First working parts roll off the line
Intel has built its first semiconductors using a 0.13 micron (130nm) process. The chip behemoth has working static RAM and microprocessors using transistors approximately one thousandth the width of a human hair.
The process will enter volume manufacturing next year with a die shrink of both the Pentium III and 4 and has the capability to deliver microprocessors containing more than 100 million transistors, running at multi-GHz clock speeds. Mobile Celerons are also likely to be amongst the first products to benefit from the new process.
The first prototype products feature a 70nm transistor gate width, 1.5nm gate oxide thickness, copper interconnects and low-k dielectrics. Intel plans to present details of the process at the International Electron Devices Meeting (IEDM) in December 2000.
"This accomplishment reaffirms our faith in Moore's Law," said Sunlin Chou, Intel vice president and general manager of the Technology and Manufacturing Group. "It is a credit to our development teams who have repeatedly overcome rising technical challenges to accelerate the arrival of new generations of silicon technology. We believe that our 130 nm process will be the earliest to ramp into volume production and to deliver products with leading edge performance, density, and power efficiency."
As well as boasting power requirements as low as one volt, Intel claims that the 130nm process will run 65 per cent faster than its current 180nm technology.
Production on the 130nm process starts next year on 200mm wafers, moving to 300mm wafers in 2002. Fabrication costs on 300mm wafers are generally reckoned to be at least 30 per cent lower than those of 200mm wafers.
The prototype 18Mb SRAM currently operates faster than 1.6GHz, with higher speeds expected as the process is enhanced. Similarly, the first 130nm microprocessors operate significantly faster than their 180nm equivalents, says Intel. ®