Mar

11

Is Hitachi Going to Join IBM Common Platform?

Filed Under Semiconductor Industry, Video Gallery | Leave a Comment

Hitachi and IBM announced on Monday that they would begin a two-year semiconductor metrology research program for next-generation 32- and 22-nanometer CMOS devices. Engineers from both Hitachi and IBM will conduct the joint research at IBM’s Thomas J. Watson Research Center in Yorktown Heights, N.Y. and at the College of Nanoscale Science and Engineering in Albany, N.Y. It’s the first time the two companies have agreed to work together on semiconductor technology. Although the current joint research program only focus on metrology, it would well be the harbinger for Hitachi to join the IBM-led Common Platform alliance (Ref).

The IBM-led Common Platform alliance is the largest bulk CMOS technology alliance in the world. Its members consist of Chartered, Freescale, Infineon, Samsung, Sony, Toshiba, and STMicroelectronics. Toshiba is the most recent member, joining the group in December last year. The R&D cost of developing next-generation CMOS technology has soared significantly. It costed about $1.5 billion to develop a 65-nm node, but it would need nearly twice or more to develop 32nm CMOS process technology. As a consequence, many chip makers do not have the financial prowess for such development, alliance seems to be the only way to stay in the 32nm rat race for most companies, except Intel and TSMC. Byran Lewis, VP of Garnter, painted a similar picture in the DesignCon conference last week (see video below).

 

Related posts

Feb

11

The 32nm Rat race has begun…

Filed Under Semiconductor Industry, Video Gallery | Leave a Comment

Just come back from a super long Chinese New Year break. According to the Chinese Zodiac, the Year of 2008 is a Year of the Rat, which begins on February 7, 2008 and ends on January 25, 2009. A Rat Year is supposedly a good year to begin a new job, get married, launch a product or make a fresh start. Well ,on the semiconductor arena, Intel finished the 45nm race as a clear winner at the end of 2007 and the rat race for 32nm has begun in 2008…

EUV litho: The prospect of EUV litho deployed in 32nm is almost vaporized after IMEC claimed a week ago that EUV litho would not be ready for 32nm. Rather immersion and double-patterning or a combination of the two, are now the most viable lithographic method to be used at 32nm (Ref). The EUV litho was originally slated for chip production at the 45nm, but now it seems that if there is any chance, EUV litho would probably be used at the latter stages of the 22nm node in the 2011 timeframe (Ref).

Laser Spike Annealing: Laser spike annealing (LSA) has already started to use in 45nm for some logic chip makers. LSA is known to be capable of full activation with negligible diffusion as compared to conventional RTA spike anneal, thus allowing the formation of ultra-shallow junctions (USJ) and highly activated source/drain contacts, as well as reduced poly depletion. In 32nm, LSA will be the key enabling technology to meet transistor performance. Ultratech is the predominant supplier of LSA tools. Thus, analysts have upgraded Ultratech to a “buy” call due to increasing demand for its LSA products in 45 and 32nm (Ref).

HKMG:Intel began the mass production of 45nm using high-k metal gate process since the 2nd half of 2007. However, the rest of the chip makers are likely to push out the complicated high-k metal gate process to 32nm. Thanks to Intel, ASM enjoys a big chuck of the market share in high-k since Intel is apparently using ASM ALD tool for high-k deposition (Ref).

Players: Starting from 32nm, we will start to see diminishing number of players in the field due to escalating cost of developing 32nm. Several companies such as Sony, TI, STMicroelectronics, and Fujitsu (Ref) are exiting 32nm and moving towards fabless model, while others are forming alliances such as the IBM Common Platform and the recent Matsushita- Renesas alliances (Ref). The only lone wolves in 32nm will be Intel and TSMC. UMC and SMIC are in the difficult positions as they have no clear roadmap on their 32nm.

The race of 32nm has just begun and I believe 32nm will change the semiconductor arena. Below is a nice video showing IBM 32nm development.

 

Related posts

Next Page →