SANTA CLARA, Calif.—South Korea's Samsung Electronics Co. Ltd. has increased the stakes in the leading-edge semiconductor business by rolling out a 20-nm process.
Samsung joins an elite list of chip makers that have talked about a 20-nm process, including Intel, Globalfoundries and TSMC.
Reportedly geared for logic and foundry applications, Samsung's 20-nm process features a bulk CMOS technology, 12 metal layers, copper interconnects, ultra low-k, stressors and a high-k/metal-gate scheme.
At 32- and 28-nm, Samsung plans to roll out a gate-first, high-k technology. But at 20-nm, the company will use rival gate-last technology.
The company will continue to push 193-nm immersion to 20-nm-with the help of source-mask optimization (SMO), double patterning and other tricks.
During a presentation, D.K. Sohn, vice president of the R&D Center at Samsung, said the company's 20-nm process is 30 percent faster than its 28-nm low-power process.
Samsung's first 20-nm technology will be a low-power process. The bulk CMOS, 1.8/1.5-volt process features a dual-gate oxide technology, thin and thick dielectics and other features.
Samsung will offer process design kits next month. Risk production is slated for the second half of 2012, with early production due in the first quarter of 2013.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.