Researchers Build World's Smallest SRAM Memory Cell

(PhysOrg.com) -- IBM and its development partners -- AMD, Freescale, STMicroelectronics, Toshiba and the College of Nanoscale Science and Engineering (CNSE) -- today announced the first working static random access memory (SRAM) for the 22 nanometer technology node, the world's first reported working cell built at its 300mm research facility in Albany, NY.

SRAM chips are precursors to more complex devices such as microprocessors. The SRAM cell utilizes a conventional six-transistor design and has an area of 0.1um2, breaking the previous SRAM scaling barriers.

Researchers achieved this breakthrough at CNSE of the University at Albany, State University of New York. CNSE's Albany NanoTech is the world's most advanced university based nanoelectronics research complex. IBM and its partners do much of their leading-edge semiconductor research at CNSE.

A nanometer is one one-billionth of a meter or about 80,000 times smaller than the width of a human hair.

"We are working at the ultimate edge of what is possible -- progressing toward advanced, next-generation semiconductor technologies," said Dr. T.C. Chen, vice president of Science and Technology, IBM Research. "This new development is a critical achievement in the pursuit to continually drive miniaturization in microelectronics."

22 nm is two generations away in chip manufacturing. The next generation is 32 nm -- where IBM and its partners are in development with their leading 32 nm high-K metal gate technology that no other company or consortium can match.

Traditionally, an SRAM chip is made more dense by shrinking its basic building block, often referred to as a cell. IBM-alliance researchers optimized the SRAM cell design and circuit layout to improve stability and developed several novel fabrication processes in order to make the new SRAM cell possible. The researchers utilized high-NA immersion lithography to print the aggressive pattern dimensions and densities and fabricated the parts in its a state-of-the-art 300mm semiconductor research environment.

SRAM cell size is a key technology metric in the semiconductor industry, and this work demonstrates IBM and its partners' continued leadership in cutting-edge process technology.

Key enablers of the SRAM cell include band edge high-K metal gate stacks, transistors with less than 25 nm gate lengths, thin spacers, novel co-implants, advanced activation techniques, extremely thin silicide, and damascene copper contacts.

Additional details of this achievement will be presented at the IEEE International Electron Devices (IEDM) annual technical meeting to be held in San Francisco, CA, December 15-17, 2008.

Provided by IBM


Explore further

Toshiba, IBM, AMD Develop World's Smallest FinFET SRAM Cell with High-k/Metal Gate

Citation: Researchers Build World's Smallest SRAM Memory Cell (2008, August 18) retrieved 21 October 2019 from https://phys.org/news/2008-08-world-smallest-sram-memory-cell.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.
0 shares

Feedback to editors

User comments

Aug 18, 2008
well if we soon start getting CPUs with transistors this small, there will literally be no more room for improvement--i mean, you physically can't get any denser.

They will have to focus their attention at logic optimization, and possibly alternative materials that allow for better heat dissipation so they can crank up the clock speed.

or something completely different like optical processors, since they can multiplex wavelengths to do many instructions over the same logic gate simultaneously.

Aug 18, 2008
well if we soon start getting CPUs with transistors this small, there will literally be no more room for improvement


Of course you can. Multigate transistors, stacked dies(can use optimial type of silicon for each task, short signal distance due to tighther 3d integration, integrated DRAM with ridiculous bandwidth as L3 cache...), wider chips(many more cores, slightly lower clocks) with better power management(individual core clocks, unused processors in sleep state), heterogenous computing(different types of processors suited for different tasks), integrated FPGA's(reconfigure hardware to do exactly what is needed; any kind of task that requires extensive bit-twiddling would gain enormously)...

They will have to focus their attention at logic optimization, and possibly alternative materials that allow for better heat dissipation so they can crank up the clock speed.


Active power goes as the cube of clock-speed(because it is quadratic in voltage and linear in frequency and the maximum frequency you can attain is proportional to the voltage); it seems more likely that they'll go for many more cores or wider cores rather than higher clocks.

Aug 19, 2008
More cores are useless thou for a CPU. There are only a very few instances that can be multi-threaded.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more