Wednesday, January 6, 2010

Technology Forward

Computers keep getting more powerful because silicon transistors keep getting smaller. But that miniaturization can't continue much further without a change to the transistors' design, which has remained more or less the same for 40 years.

One potential successor to today's silicon transistors is silicon nanowires, tiny filaments of silicon suspended like the strings of a guitar between electrically conducting pads. But while silicon nanowires are certainly small enough to keep the miniaturization of computer circuitry on track, there's been doubt about whether they can pass enough electrical current for high-speed computing. At 2008's International Electron Device Meeting, researchers at MIT's Microsystems Technology Laboratories demonstrated silicon nanowires with twice the electron mobility — which indicates how easily current can be induced — of their predecessors. Now, the same group has shown that they can build chips in which up to five high-performance nanowires are stacked on top of each other. That would allow nanowire transistors to pass up to five times as much current without taking up any more area on the surface on the chip, a crucial step toward establishing the viability of silicon-nanowire transistors.

A transistor is basically a switch: when it's on, it passes an electrical current, and when it's off, it doesn't. Flipping the switch requires charging a part of the transistor called the "gate." In today's design, the gate sits on top of the transistor. But if the transistor gets small enough, electricity will leak across it whether the gate is charged or not. Turning the switch off becomes impossible.

Because silicon nanowires are suspended in air, the gate can be wrapped all the way around them, like insulation around an electrical wire, which improves control of the switch. But the narrowness of the nanowires limits the amount of current they can pass.

Electrical-engineering professor Judy Hoyt
and her graduate students Pouya Hashemi
and Leonardo Gomez
improved the performance of silicon-nanowire transistors by, basically, prying the atoms of the silicon slightly farther apart than they would be naturally, which allows electrons to flow through the wires more freely. Such "strained silicon" has been a standard way to improve the performance of conventional transistors since 2003. But Hoyt was one of the early researchers in the field.

"Starting in the early 1990s, she's really played a pioneering role in strained-silicon technology," says Tahir Ghani, director of transistor technology and integration for Intel's Technology and Manufacturing Group. "She did a lot of this pioneering work that for the first time demonstrated that you can have significant performance gains by implementing strain into silicon technology." Hoyt and her group's work on strained-silicon nanowires, Ghani says, "combines the two key elements of transistors" — performance and space efficiency — "both of which are very key to scaling in the future. And so from that standpoint, it makes it very relevant for industry."

Handling stress

To build their stacked nanowire transistors, the MIT researchers begin with a normal silicon wafer, on which they deposit a silicon-germanium composite. Because germanium atoms are bigger than silicon atoms, the distances between atoms in the silicon-germanium layer are greater than they would be in a layer of pure silicon. When the researchers deposit another layer of silicon on top of the composite, the silicon atoms try to align themselves with the atoms beneath them, so they, too, end up spaced slightly farther apart.

This layer of strained silicon is bound to a second silicon wafer, and the other layers are removed, leaving the second wafer covered with a base layer of strained silicon. The researchers then stack alternating layers of silicon-germanium and silicon on top of the base layer, passing its strain on to each successive layer of silicon. Using a technique called electron-beam lithography, the researchers pattern fine lines onto the stacks and then etch away the material between the lines. Finally, they etch away the remaining silicon-germanium, and they're left with several layers of suspended silicon nanowires. Hoyt and her students have manufactured nanowires with a diameter of only eight nanometers, which they described in a 2009 paper in the Institute of Electrical and Electronics Engineers journal Electron Device Letters; by contrast, the smallest elements of today's computer chips are 45 nanometers across.

Hoyt says that her group can create silicon with two times the strain seen in chips built by commercial vendors. "We increase the germanium fraction of the initial layer, so we therefore build more stress into the silicon," Hoyt says. Moreover, says Hashemi, "we are the only group in the world that has showed that we can maintain this strain after suspension" — that is, once the underlying layers have been cut away.

So far, Hoyt's group has built nanowire transistors in which charge is carried by moving electrons. But to maximize computational efficiency, a standard computer chip in fact uses two types of transistors. In the other type, charge is carried by so-called holes. A hole is simply the absence of an electron in a crystal of semiconducting material. When an electron slides over to fill the hole, it vacates its own spot in the crystal; another electron slides over to fill that spot; and so on. In this way, the hole in effect moves along the length of the crystal.

Increasing the mobility of holes in such transistors requires a different type of strain: the atoms of the crystal actually have to be jammed closer together than is comfortable. So Hoyt's group is now working to build nanowires from a silicon-germanium composite, where intervening layers of pure silicon cause compression rather than tension.

Tuesday, September 15, 2009

OPEN ACCESS TO SCINETFIC RESEARCH

September 14, 2009 - Five of premier institutions of higher learning - Cornell, Dartmouth, Harvard, the Massachusetts Institute of Technology, and the University of California at Berkeley - today announced their joint commitment to a compact for open-access publication.

Open-access scholarly journals have arisen as an alternative to traditional publications that are founded on subscription and/or licensing fees. Open-access journals make their articles available freely to anyone, while providing the same services common to all scholarly journals, such as management of the peer-review process, filtering, production, and distribution.

According to Thomas C. Leonard, University Librarian at UC/Berkeley, "Publishers and researchers know that it has never been easier to share the best work they produce with the world. But they also know that their traditional business model is creating new walls around discoveries. Universities can really help take down these walls and the open-access compact is a highly significant tool for the job."

The economic downturn underscores the significance of open-access publications. With library resources strained by budget cuts, subscription and licensing fees for journals have come under increasing scrutiny, and alternative means for providing access to vital intellectual content are identified. Open-access journals provide a natural alternative.

As Dartmouth Provost Barry P. Scherr sees it, "Supporting open-access publishing is an important step in increasing readership of Dartmouth research and, ultimately, the impact of our research on the world."

Since open-access journals do not charge subscription or other access fees, they must cover their operating expenses through other sources, including subventions, in-kind support, or, in a sizable minority of cases, processing fees paid by or on behalf of authors for submission to or publication in the journal. While academic research institutions support traditional journals by paying their subscription fees, no analogous means of support has existed to underwrite the growing roster of fee-based open-access journals.

Stuart Shieber, Harvard's James O. Welch, Jr. and Virginia B. Welch Professor of Computer Science and Director of the University's Office for Scholarly Communication, is the author of the five-member compact. According to Shieber, "Universities and funding agencies ought to provide equitable support for open-access publishing by subsidizing the processing fees that faculty incur when contributing to open-access publications. Right now, these fees are relatively rare. But if the research community supports open-access publishing and it gains in importance as we believe that it will, those fees could aggregate substantially over time. The compact ensures that support is available to eliminate these processing fees as a disincentive to open-access publishing."

The compact supports equity of the business models by committing each university to the timely establishment of durable mechanisms for underwriting reasonable publication fees for open-access journal articles written by its faculty for which other institutions would not be expected to provide funds.

Additional universities are encouraged to visit the Compact web site and sign on.

Cornell Provost Kent Fuchs offers his perspective on participating in the Compact. "As part of its social commitment as a research university," Fuchs says, "Cornell strives to ensure that scholarly research results are as widely available as possible. The Compact for Open-Access Publishing Equity could increase access to scholarly literature while at the same time ensuring that the valuable services that publishers provide are supported."

A full account of the motivation for the compact can be found in the article "Equity for Open-Access Journal Publishing" published in the open-access journal Public Library of Science Biology.

"Supporting OA journals is an investment in a superior system of scholarly communication," states Peter Suber of the Scholarly Publishing and Academic Resources Coalition (SPARC) in Washington, DC, and a fellow of Harvard Law School's Berkman Center and Harvard University's Office for Scholarly Communication. "Before this compact, a number of funding agencies and universities were willing to pay OA journal processing fees on behalf of their grantees and facult. It's significant that five major universities recognize the need to join the effort, extend fee subsidies to a wider range of publishing scholars, enlist other institutions, and start to catch up with their long practice of supporting traditional -- or non-OA -- journals."

Summing up the Compact, MIT Provost L. Rafael Reif observes, "The dissemination of research findings to the public is not merely the right of research universities: it is their obligation. Open-access publishing promises to put more research in more hands and in more places around the world. This is a good enough reason for universities to embrace the guiding principles of this compact."


Source : http://web.mit.edu/newsoffice/2009/open-access-0914.html


Wednesday, January 16, 2008

Pictures from Dubai, in a good raining day

 

 

 

 
Posted by Picasa

Two Cities and Two Civilizations






الفرق بين نيويورك و دبي، هو كالفرق بين "خادشات السماء" و"ناطحات السحاب"، ولا أدري من هو بالتحديد ذلك الذي أطلق على المباني العالية في نيويورك مصطلح sky scrapers الذي ترجمته الحرفية من اللغة الإنجليزية إلى اللغة العربية هي "خادشات السماء"، ولا أدري كذلك من هو الذي وضع المصطلح العربي "ناطحات السحاب" اسماً للمباني العالية. لكنني أعرف كلاً من نيويورك ودبي ، وأعرف المباني العالية في كل منهما ، واوافق تماماً على تسمية المباني العالية في نيوورك "خادشات السماء" والمباني العالية في دبي "ناطحات السحاب" وذلك لا علاقة له بالإرتفاع كما قد يظن بعض غير ذوي الهوى العربي، ولكن له علاقة بالروح والمحتوى الإنساني، والفرق بين الرأسمالية المتوحشة التي تستوطن نيويورك والرأسمالية الأليفة التي تستوطن دبي، نعم ذلك هو الفرق الحضاري بين هاتين المدينتين، وذلك أيضاً هو الفرق الحضاري بين هاتين الحضارتين!
Posted by Picasa


Technology Forward

Computers keep getting more powerful because silicon transistors keep getting smaller. But that miniaturization can't continue much furt...