Graphene Transistors Do Triple Duty in Wireless Communications

Triple transistor: Single graphene transistors like this one can be made to operate in three modes and perform functions that usually require multiple transistors in a circuit.
Credit: Alexander Balandin

Communications

Graphene Transistors Do Triple Duty in Wireless Communications

A single graphene transistor that does the job of many conventional ones could lead to compact chips for cell phones.

  • Friday, October 22, 2010
  • By Katherine Bourzac

Graphene's potential was recognized earlier this month when those who first studied it in the lab won the 2010 Nobel Prize in Physics. But researchers are just beginning to figure out how to take advantage of the novel carbon material in electronic devices.

Researchers have already made blisteringly fast graphene transistors. Now they've used graphene to make a transistor that can be switched between three different modes of operation, which in conventional circuits must be performed by three separate transistors. These configurable transistors could lead to more compact chips for sending and receiving wireless signals.

Chips that use fewer transistors while maintaining all the same functions could be less expensive, use less energy, and free up room inside portable electronics like smart phones, where space is tight. The new graphene transistor is an analog device, of the type that's used for wireless communications in Bluetooth headsets and radio-frequency identification (RFID) tags.

Graphene's perfect structure at the atomic level provides smooth sailing for electrons, and the material conducts electrons better than any other materials do at room temperature. So far, it's been used to make transistors that switch at about 100 gigahertz, or 100 billion times per second, 10 times faster than the best silicon transistors; it's predicted the material could be made into transistors that are even 1,000 times faster than this. And because graphene is smooth and flat, it should be compatible with the chip-making equipment at semiconductor fabs.

But graphene offers other properties besides just being a great conductor of electrons, says Kartik Mohanram, professor of electrical and computer engineering at Rice University. It's also possible to change the behavior of a graphene transistor on the fly, something that can't be done with conventional silicon transistors. The transistors that make up conventional silicon logic circuits can only behave in one of two ways, called "n" for negative or "p" for positive--they either control the flow of electrons or the flow of "holes," or positive charges. Whether a conventional transistor is p-type or n-type is determined during fabrication. But graphene is ambipolar: it can conduct both positive and negative charges.

Mohanram has designed a transistor that can be changed, and has made and tested it with Alexander Balandin, professor of materials science and engineering at the University of California, Riverside. By changing the voltage applied to a sheet of graphene using three electrical gates, they could switch the graphene between three different modes: n-type, p-type, and a mode where it conducted positive and negative charge equally. This triple-mode transistor acts as an amplifier and can be used to encode a data stream by changing the frequency and the phase of a signal. Changes in phase and frequency are used to encode data in telecommunications devices such as Bluetooth headsets and RFID tags.

Mohanram and Balandin's device is the first that can do this level of signal processing in a single transistor. Usually such signaling requires multiple transistors. Their transistor is a proof-of-concept device, but Mohanram says it demonstrates what might be possible with graphene.

Other groups have demonstrated multimode transistors using graphene, carbon nanotubes, and organic molecules. The researchers say that the new graphene triple-mode circuit can be controlled better than those devices.

Control is critical when designing transistors that are ambipolar, says Subhasish Mitra, professor of electrical engineering and computer science at Stanford University. "People used to consider ambipolarity a bad thing" because it's typically difficult to control how an ambipolar transistor will behave, which makes it difficult to use them at all, he says.

Mitra notes that the benefits shown at the single-transistor level must now be demonstrated in systems. The electrical gates needed to control the behavior of arrays of ambipolar transistors might end up making circuits much harder to design and fabricate. "Now that they have shown that they can do this, we need to see what benefit it brings at a system level," he says.

Balandin and Mohanram are now working on graphene circuits to test the benefits of ambipolarity at a higher level. They're also changing the design of the transistors themselves to make them more efficient.

No one has yet published any articles on the creation of integrated circuits made of graphene transistors, but Balandin says researchers are now on the verge of putting it all together. As materials scientists and device fabricators work on overcoming the challenges of working with graphene, says Mohanram, circuit designers should keep pace with them and think creatively about ambipolarity and other possibilities opened up by graphene and other nanomaterials. "New designs and new ways of thinking can lag behind the development of new materials," he says.

Making Graphene Nanomachines Practical

Machine making: A transparent sheet of graphene is stretched over the surface of this silicon wafer. The graphene can oscillate over holes in the silicon beneath, acting as a nanomechanical device called a resonator.
Credit: ACS/Nano Letters

Computing

Making Graphene Nanomachines Practical

Graphene devices could make extremely sensitive sensors and superfast electronic switches for consumer electronics.

  • Wednesday, December 1, 2010
  • By Katherine Bourzac

Many of today's consumer electronics rely on microscopic machines. These tiny devices are found in smart-phone motion sensors, inkjet printheads, and the switches that activate some display pixels, to name just a few components.

Shrinking these electromechanical machines down to the nanoscale would enable new devices, such as extremely sensitive chemical sensors, incredibly precise accelerometers, and super-fast integrated circuit switches. In an important step toward this goal, researchers at Cornell University have made large arrays of nanoscale resonators using graphene.

An atom-thin form of carbon called graphene is among the most promising materials for making nanoelectromechanical systems (NEMS). Graphene is the strongest known material, and the most electrically conductive. Graphene's atom-thin size means it is also incredibly lightweight and can move very fast. Cornell physics professor Paul McEuen says graphene can be used to build large numbers of nanodevices with equipment developed for etching silicon chips on flat wafers. But building mechanical nanomachines from graphene is challenging, and most of the devices created so far have been one-offs.

McEuen and fellow Cornell professor Harold Craighead have now shown that they can make graphene nanodevices called resonators on the surface of a silicon wafer. Each resonator is made of a film of graphene that oscillates back and forth, like a trampoline moving up and down, in response to a mechanical force applied to its surface or to an electrical field.

The Cornell group first etched trenches into the surface of a silicon wafer. They then topped the wafer with a film of graphene grown on top of copper. The graphene sticks to the surface of the silicon wafer like plastic cling wrap would. The researchers finally add electrical contacts to the graphene to complete the resonators. The work is described online in the journal Nano Letters.

"We're making large numbers of identical resonators, which demonstrates a transition from a lab experiment to a technology," says McEuen. Previous nanoresonators made at this scale were either much thicker and less sensitive, or they had to be made one at a time. "The two major obstacles in implementing nanodevices are scale-up and reproducibility in performance," says Alex Zettl, professor of physics at the University of California, Berkeley. Zettl has made similar devices from carbon nanotubes, including a radio made from a single carbon nanotube. "Using single-layer graphene allows many devices to be made in one shot, with similar performance," Zettl says.

Graphene nanoresonators could make very sensitive chemical detectors or accelerometers. The suspended graphene films respond dramatically when any weight is added—even just a molecule or an atom. "It couples very strongly to the outside world," which makes for a good sensor, says McEuen.

Rod Ruoff, professor of mechanical engineering at the University of Texas at Austin, who pioneered the graphene growth-and-transfer technique used by the Cornell group, says this work demonstrates that this type of graphene performs well in nanomechanical systems. But Ruoff says he sees room for improvement in the performance of the resonators.

The Cornell researchers are now working to push the graphene resonators to their ultimate performance limits. The crystalline structure of graphene, which determines its strength and electrical conductivity, is not perfect in the Cornell devices made so far.

The researchers also hope to take advantage of quantum effects that occur at the nanoscale. This could improve their sensitivity, McEuen says.

Writing Circuits on Graphene

Hot wire: An AFM tip heated to over 150 °C can etch an insulating graphene oxide surface to create thin conductive nanoscale wires.
Credit: Debin Wang, Georgia Tech

Computing

Writing Circuits on Graphene

A heated AFM tip can draw nanometers-wide conductive lines on graphene oxide.

  • Tuesday, June 15, 2010
  • By Prachi Patel

Using a heated atomic force microscope tip, researchers have drawn nanoscale conductive patterns on insulating graphene oxide. This simple trick to control graphene oxide's conductivity could pave the way for etching electronic circuits into the carbon material, an important advance toward high-speed, low-power, and potentially cheaper computer processors.

Graphene, an atom-thick carbon sheet, is a promising replacement for silicon in electronic circuits, since it transports electrons much faster. IBM researchers have already madetransistors, the building blocks of electronic circuits, with graphene that work 10 times faster than their silicon counterparts. But to make these transistors, researchers first have to alter the graphene's electronic properties by cutting it into thin ribbons, which are then incorporated into devices. Researchers have made these nanoribbons with lithography, with chemical solution-based processes, or by unzipping carbon nanotubes.

In the new Science paper, researchers at the Georgia Institute of Technology and the U.S. Naval Research Laboratory instead "write" such nanoribbons on a surface rather than cutting graphene. The researchers start with a graphene oxide sheet, which doesn't conduct electric current. When they pull an AFM tip heated to between 150 °C and 1060 °C across the sheet, oxygen atoms are shed at the spots that the tip touches. This leaves behind lines of almost-pure graphene that are 10,000 times more conductive than the surrounding graphene oxide.

"It's a fast, reproducible technique, it's one-step, it's simple," says Paul Sheehan, who led the work at the Naval Research Laboratory. "Instead of putting down resist and trying to cut graphene in different ways, you can use local heat and write the lines exactly where you want them." Sheehan says that an array of thousands of AFM tips could sketch circuits on graphene oxide at the same time.

Lithographic methods to make nanoribbons are cumbersome and expensive, says Jing Guo, an electrical and computer engineering professor at the University of Florida in Gainesville. These methods can also create ribbons with rough edges, which affect graphene's electronic properties and result in low-quality transistors. "This is a new way to [make nanoribbons] that's very simple and reliable and potentially scalable to large scale," he says. "You basically have a paper and take a pencil to scratch it, and you have a very narrow line."

The researchers wrote lines as narrow as 12 nanometers across and at speeds of up to 0.1 millimeters per second. The writing speed increased with temperature. "It is exciting to see that this conversion can be done and controlled at the nanoscale," says Yu-Ming Lin, a nanoscale science and technology group researcher at IBM's Watson Research Center in Yorktown Heights, NY. "This is an important step for graphene-based [electronics]."

Starting with graphene oxide sheets rather than graphene is easier and cheaper, says Elisa Riedo, a physics professor at Georgia Tech who led the work with Sheehan. Pristine graphene sheets are typically obtained by mechanically separating flakes from graphite or by growing graphene on two-inch silicon carbide wafers. "Graphene oxide was cheaper to produce in large areas compared to graphene," Riedo says. "It's a different path to arrive to graphene."

The researchers plan to make transistors using their technique, but they might need additional processing first, says Yanwu Zhu, a graphene researcher at the University of Texas at Austin. For one thing, they will have to find a way to remove graphene oxide remnants from the conductive ribbons.

Making Graphene More Practical

Bigger, better graphene: A new way to coat large areas of silicon with single-layer sheets of graphene makes it easy to fabricate an array of field-effect transistors by depositing gold electrodes on top of the graphene. The SEM image (bottom) shows graphene spanning the seven-micrometer gap between the source and drain electrodes.
Credit: Yang Yang Laboratory, UCLA

Computing

Making Graphene More Practical

A novel process yields big pieces of single-ply graphene for smaller, faster electronics.

  • Tuesday, November 18, 2008
  • By Prachi Patel

Researchers at the University of California, Los Angeles, have found a simple way to make large pieces of the carbon material graphene. Graphene, a flat, one-atom-thick sheet of carbon, can transport electrons at very high speeds, making it an attractive material for electronic devices. But producing sufficient quantities of large, uniform single-layer sheets of graphene has been a challenge. So far, processes to make graphene create small quantities of graphene flakes or films made of overlapping pieces.

Using the new method, presented online in Nature Nanotechnology, the researchers report making single-layer 0.6-nanometer-thick pieces that are tens of micrometers wide. Materials-science and engineering professor Yang Yang and his colleagues deposit the sheets on silicon wafers to make prototype field-effect transistors.

Testing at least 50 such transistors, the researchers found that the devices had an output current of a few milliamperes. That is 1,000 times higher than the output current of the devices that others have recently reported while using similar techniques to make graphene. "We believe this is a game-changing approach which will significantly improve graphene electronics in the future," Yang says.

Electrons flow through graphene sheets tens of times faster than they flow in silicon. The material could lead to electronic devices that are smaller, faster, and less power hungry than are those made of silicon. Thin and transparent, graphene is also a promising replacement for the indium tin oxide electrodes and the silicon thin-film transistors used in flat-panel displays.

The easiest way to make single sheets of graphene is by using adhesive tape to peel graphene flakes off of pieces of graphite, which is a stack of multiple graphene layers. This process results in a very small amount of tiny flakes of graphene. The pieces would have to be much larger for any practical use. "If you can coat an entire silicon wafer with a single sheet of graphene, then you can do lithography or patterning and have little devices," says James Tour, a chemistry professor at Rice University.

About two years ago, researchers came up with a chemical method that yields larger graphene pieces. They oxidize graphite to make graphite oxide and dissolve it in water. The oxygen atoms pry apart the individual graphene sheets, which get dispersed in the solution. After the researchers deposit the sheets on a substrate, the oxygen is removed using another chemical or by heating.

Manish Chhowalla, a materials-science and engineering professor at Rutgers University, has made one-to-two-nanometer-thick films with this method. He uses vapors of a chemical called hydrazine to remove the oxygen groups from the deposited film. The films, made of slightly overlapping graphene pieces, are a few centimeters wide.

Yang points out that the quality of the sheets made so far has not been very good. Because the graphene sheets are deposited on a substrate first, many oxygen groups get trapped between the sheets and the substrate underneath and are not removed. "These are detrimental to electrical properties," he says.

Yang and his colleagues have simplified the method. They dissolve graphite oxide pieces in pure hydrazine. This splits apart the individual graphene sheets and gets rid of nearly all the oxygen groups in a single step. The researchers then deposit the pieces on a silicon wafer. They could also deposit the flakes on flexible surfaces.

"The main contribution is that they've figured out a better way of [removing oxygen groups]," Chhowalla says.

The researchers uniformly cover large areas of silicon wafers about 1.5 centimeters in length and width with graphene sheets. Then they deposit gold electrodes on top of the flakes to make field-effect transistors.

The researchers are working to further improve the quality of the graphene sheets. Pure, flat graphene sheets have a thickness of 0.34 nanometers. The 0.6-nanometer thickness of the sheets that the researchers make implies that a few oxygen groups remain stuck to the graphene. "So it still might not be as good as the graphene you want, but it's getting close," Tour says. "It's certainly good enough for lots of devices."