SMIC: Sustained Development of Cutting-Edge Technology in IC Design and Manufacturing
By Gang Yao, SEMI China
SEMI China’s Gang Yao interviewed SMIC’s Li Xuwu, senior VP of technology R&D, in the fall 2012.
The Semiconductor Manufacturing International Corporation (SMIC) had net profit in Q2 of this year reached $7.1 million, returning to the profit line despite the limited growth. As shown in Q2 financial report, the sales revenue is $421.8 million, increasing by 26.8 percent compared with Q1 and by 19.7 percent compared with the same period in last year. One of the key highlights for this growth is that the quarterly growth of the sales revenue from 65/55nm technique reached 87 percent. Li Xuwu, senior VP of SMIC technology R&D, explains that the design of the 65/55nm logic technology is well received by users thanks to its performance and energy-saving advantages, along with its low cost.
According to Li Xuwu, SMIC is currently accelerating to catch up with the progress of international advanced mainstream techniques at the speed of 0.5 years for each generation. SMIC has now finished the research and development of a full set of 45/40nm technological processes, and passed the validation of international mainstream customers. It still lags behind the leading contract manufacturers by about 2.75 years. 32/28nm process technology has already moved into the substantial development stage, and has made a significant breakthrough of independent innovation in the core field of technology. The development of the full set of 32/28nm technology is expected to be completed by late 2013 and then goes for small batch trial production, shortening the gap from the leading contract manufacturers to two years. Li Xuwu says that SMIC’s preliminary work on the 20nm node is in full swing as well, and is estimated to be finished by the middle of 2015, thus shortening the gap to 1.5 years. SMIC will continue to smaller line width such as 16/14nm, Li Xuwu said with determination.
“Semiconductor technology is developing rapidly. I believe the technology node will continue to evolve driven by the innovation spirit,” Li Xuwu pointed out. In the micron era, the technology node is proceeding slowly; while in the 90nm era, the techniques will be faced with substantial technical innovation and breakthroughs after every two technology nodes. Actually, in the aspect of the technological development of 28nm node, SMIC will realize major leaps on the three key techniques, including HKMG, SiGe embedded into PMOS and ultra-low energy-saving material interconnection.
This year, SMIC’s technology R&D team has been significantly strengthened — building the foundation for SMIC to create a world-class technology R&D team and progress to advanced technology nodes. Li Xuwu explained that based on the development characteristics of the nanometer era, SMIC had three R&D teams for the development of three advanced node techniques simultaneously. While the 28nm team is accelerating the development of core technology, the 22/20nm and 16/14nm development teams were separately set up in Q4 of 2011 and Q2 of 2012, entering into the preliminary development stage.
Li Xuwu stated that during its evolution to the next generation of advanced technology, in addition to the challenges confronted together in the industry such as R&D speed and expensive equipment, SMIC will be faced with challenges that other chip contract manufacturers abroad may never see: the restrictions of international export regulations that SMIC has to use the parameters-reduced units to make products of high-technology nodes. These non-technical barriers can only be conquered through their own efforts. Of course, under the premise of independent R&D, active cooperation with foreign IP developers and bringing in some necessary IP are also important.
To avoid technical gaps, SMIC makes the best use of R&D talents in all aspects of personnel, management and organization, according to Li Xuwu. It organically combines general talents with specialized talents, and systematically integrates the R&D resources at each technology node to assist in sharing of the reusable and general resources, thus speeding up the research progress of process technology. As for the R&D pattern, SMIC has established a systematic working method to approach the technology target values.
According to Li Xuwu, the reusability of previous technology generations sometimes reaches up to 70 percent, but the reusability in the nanometer process is decreasing. If most parts of the thousands of processes need to rely on retests, especially some of which will have to go through tens of thousands of tests before completing the technological parameters setting, it would not only require large amounts of R&D economic investment, but also have delayed the time to implement new technology. Usually the time investment cost is higher than the economic cost. Hundreds of SMIC R&D personnel would spend several years to finish. Advanced process theoretical basics and device simulation help to reduce a large amount of tape-out tests, thus not only saving costs but also shortening the time required by R&D.
He stated that independent technologies are critical. On the 28nm node, SMIC adopted the “gate-last” process route, while IBM chose the “gate-first” process route. Early this year, SMIC successfully signed a cooperative development agreement with IBM on the cooperation of 28nm node’s “gate-last” compatibility technology development. Compared with the previous contract on the 45nm technology, this cooperative agreement comes with substantial changes. Li Xuwu explains that although SMIC has gained IBM’s authorization on the 45nm node process and has finished 90 percent of the technological process updating, it is still forbidden to be used on any 28nm or other node techniques. Today’s 28nm technical cooperation agreement has reduced the cooperative costs greatly. Above all, SMIC would not be restricted, and could apply these cooperative research achievements to the development of much more advanced technology nodes.
Apart from independent technical capabilities, the solidarity, effective management and rational use of the R&D team, as well as the cooperation with colleges and research institutes, and great support from the central and local governments constitute the guarantee and driving forces to advance technology at SMIC, Li Xuwu said confidently.
November 6, 2012