Life after silicon: Using exotic materials to help microchips keep improving

Dec 08, 2009 by Larry Hardesty
A test chip used to evaluate the performance of indium gallium arsenide in logic circuits. Image courtesy of the del Alamo group

(PhysOrg.com) -- The huge increases in the power and capacity of computers, cell phones and communications networks in the last 40 years have been the result of ever-shrinking silicon transistors. But silicon transistors are now getting so small that they’re running up against fundamental physical limits: soon, it will be impossible to squeeze any better performance out of them.

Researchers in MIT’s Microsystems Technology Laboratories, led by professor of Jesús del Alamo, have been investigating whether made from more exotic materials can keep the processing power coming. At the International Electron Devices Meeting this week in Baltimore — the premier conference on microelectronics — they are presenting four separate papers that offer cause for hope.

Del Alamo’s group works with compound semiconductors, so called because, unlike , they’re compounds of several other materials. In particular, the group works with materials that combine elements from columns III and V of the periodic table and have names like gallium arsenide and indium gallium arsenide.

Electrons travel through these so-called III-V materials much more rapidly than they do through silicon, and III-V semiconductors have been used for years in high-speed electronics, such as the devices that process data in fiber-optic networks. But according to del Alamo, the transistors in III-V optical components are “larger by several orders of magnitude” than the transistors in computer chips. Whether they can maintain their performance advantages at dramatically smaller scales is the question that del Alamo’s group is tackling.

In a computer chip, transistors serve as on-off switches that help execute logic operations — comparing two values, for instance, or performing arithmetic functions. But transistors can also be used to amplify electrical signals — as they do in transistor radios. Last year, del Alamo’s group built a III-V transistor that set a world record for high-frequency operation, meaning that it was able to amplify higher-frequency signals than any previous transistor. While that gives some sense of the transistor’s capacities, two of the four papers being presented in Baltimore assess properties of the transistor that better predict its performance as a logic element.

To measure those properties, del Alamo says, the group built chips with multiple transistors that were identical except for the length of their most critical element, called the gate. If a transistor is a switch, the gate is what throws it. When the gate is electrically charged, it exerts an on a semiconductor layer beneath it; that force is what determines whether the semiconductor can conduct electricity or not.

By comparing the performance of transistors with different gate lengths at different frequencies, del Alamo’s group was able to extract precise measurements of both the velocity of the electrons passing through the transistor and the electrostatic force that the gate exerted on the semiconductor layer.

Electron velocity, del Alamo says, is the “key velocity that is going to set the performance of a future logic switch based on these kinds of materials, and we have obtained velocities that are easily two and a half times higher than the best silicon transistors made today.” While the electrostatic force exerted by the gate was lower than the researchers had hoped, measuring it so precisely allowed del Alamo’s group to develop better physical models of III-V transistors’ behavior. On the basis of those models, del Alamo says, he believes that the gate’s performance is a “manageable problem.”

“Before the industry is willing to switch from one technology to the next technology, people in the industry would want to make sure that they were making the right bet, the right investment,” says Robert Chau, a senior fellow at chip giant Intel who has worked closely with del Alamo’s group. “So obviously, people will need to thoroughly understand the physics behind this proposed new technology. I think that’s what Professor del Alamo’s group has been doing — getting to the science of the device operation.”

The third paper is in a similar vein, a collaboration with researchers at Purdue University who have developed simulators to model the performance of III-V transistors that are even smaller than the MIT prototypes. Del Alamo’s group used its precise measurements of the prototypes’ performance to help the Purdue team calibrate its simulator. The fourth paper, however, addresses a different topic: it proposes a new design for III-V transistors that, del Alamo says, will work better at smaller scales, because it permits a thinner layer of material to separate the gate and the semiconductor material beneath it.

For all their speed, III-V semiconductors have a disadvantage as a next-generation chip material: they’re rarer, and therefore more expensive, than silicon. But while the MIT prototypes were built entirely with III-V materials, del Alamo envisions “a silicon-like technology where just under the gate … you take silicon out and stick in indium gallium arsenide. It’s a minute amount of material that is required in every transistor to make this happen.” Indeed, “III-V is not really in competition with silicon,” Chau agrees. “It still will be a silicon transistor. It’s just that you’re using this non-silicon element to make it even better.”

Provided by Massachusetts Institute of Technology (news : web)

Explore further: Renesas announces SRAM using leading-edge 16 nm FinFET for automotive information systems

add to favorites email to friend print save as pdf

Related Stories

Intel and QinetiQ Collaborate On Transistor Research

Feb 09, 2005

The results of a two-year joint research programme by Intel Corporation and QinetiQ into new transistor technology that could become a promising candidate for making microprocessors in the middle of the next decade was made ...

Intel: New Material to Help Chips Run Cooler, Use Less Energy

Dec 07, 2005

Intel Corporation today announced development of a new, ultra-fast, yet very low power prototype transistor using new materials that could form the basis of its microprocessors and other logic products beginning in the second ...

Scientists Develop World's Fastest Graphene Transistor

Dec 19, 2008

(PhysOrg.com) -- IBM Researchers today announced that they demonstrated the operation of graphene field-effect transistors at GHz frequencies, and achieved the highest frequencies reported so far using this ...

Intel Researchers Improve Tri-Gate Transistor

Jun 13, 2006

Intel Corporation researchers today disclosed they have developed new technology designed to enable next era in energy-efficient performance. Intel's research and development involving new types of transistors ...

Recommended for you

Cyclist's helmet, Volvo car to communicate for safety

6 hours ago

Volvo calls it "a wearable life-saving wearable cycling tech concept." The car maker is referring to a connected car and helmet prototype that enables two-way communication between Volvo drivers and cyclists ...

California puzzles over safety of driverless cars

7 hours ago

California's Department of Motor Vehicles will miss a year-end deadline to adopt new rules for cars of the future because regulators first have to figure out how they'll know whether "driverless" vehicles ...

Britain's UKIP issues online rules after gaffes

7 hours ago

UK Independence Party (UKIP), the British anti-European Union party, has ordered a crackdown on the use of social media by supporters and members following a series of controversies.

Sony saga blends foreign intrigue, star wattage

7 hours ago

The hackers who hit Sony Pictures Entertainment days before Thanksgiving crippled the network, stole gigabytes of data and spilled into public view unreleased films and reams of private and sometimes embarrassing ...

User comments : 1

Adjust slider to filter visible comments by rank

Display comments: newest first

dirk_bruere
not rated yet Dec 08, 2009
I've been hearing this same story for years. It ain't gonna happen. The future is (probably) graphene.

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

Click here to reset your password.
Sign in to get notified via email when new comments are made.