Since 1986 - Covering the Fastest Computers in the World and the People Who Run Them

Language Flags
July 16, 2012

Developing Diminutive Transitors is a Fight Against Physics

Robert Gelber

The fight to keep Moore’s Law alive has been stacked with physical limitations. Fittingly, Technology Review published an article on Friday the 13th, discussing the behind-schedule technology required to manufacture the next generation chips. Known as extreme ultraviolet lithography (EUV), the process was expected to fabricate 22-nanometer chips on the market today. Unfortunately that didn’t pan out, leaving the silicon to be manufactured by tweaking the aging process of standard lithography. 

The problem has prompted Intel to invest a sum of $4 billion into ASML, a Dutch company that makes the equipment used to fabricate chips. Both Intel and ASML are trying to get large chipmakers to join an effort keeping silicon alive.  While the tweaked version of lithography has been used to manufacture 22-nanometer chips, this process is only viable for the next two generations. The stopgap solution should cover the creation of 14 and 11 nanometer chips, which is expected to suffice until 2013.

After that, the only process expected to keep Moore’s law going is EUV. While current lithography employs 193-nanometer ultraviolet light to makes chips, EUV uses higher energy UV rays with wavelengths around 13 nanometers. The process involves writing a pattern into a chemical layer on top of a silicon wafer. The layer is then etched into the silicon using a chemical process.  Unfortunately, the technology is not yet feasible for a production environment.              

A main roadblock to EUV technology is the need for powerful light sources. Because the wavelength is so short,, it gets absorbed by all types of matter. The EUV machines attempt to alleviate this issue by passing the beam through a vacuum, but it becomes too weak by the time it hits the silicon wafer. Changing the direction and focus of the light has been tried, but ends in similar results.

At this point, ASML’s most advanced EUV prototype can generate beams only half as strong as chipmakers required to make the process viable. The company and Intel are investing in second-generation EUV, but the money is also being used get the first-generation technology working. ASML spokesman Ryan Young expressed the urgency about getting EUV off the ground, saying, “Clearly, there is no next generation if we don’t get this generation working.”

Full story at Technology Review