Semiconducting graphene: Fabrication on patterned silicon carbide produces bandgap to advance graphene electronics

Nov 18, 2012
This is a perspective atomic force microscope (AFM) view of graphitized trenches that are 18 nanometers deep. Credit: Georgia Institute of Technology

By fabricating graphene structures atop nanometer-scale "steps" etched into silicon carbide, researchers have for the first time created a substantial electronic bandgap in the material suitable for room-temperature electronics. Use of nanoscale topography to control the properties of graphene could facilitate fabrication of transistors and other devices, potentially opening the door for developing all-carbon integrated circuits.

Researchers have measured a bandgap of approximately 0.5 electron-volts in 1.4-nanometer bent sections of graphene nanoribbons. The development could provide new direction to the field of graphene electronics, which has struggled with the challenge of creating bandgap necessary for operation of electronic devices.

"This is a new way of thinking about how to make high-speed graphene electronics," said Edward Conrad, a professor in the School of Physics at the Georgia Institute of Technology. "We can now look seriously at making fast transistors from graphene. And because our process is scalable, if we can make one transistor, we can potentially make millions of them."

The findings were scheduled to be reported November 18 in the journal . The research, done at the Georgia Institute of Technology in Atlanta and at SOLEIL, the French national synchrotron facility, has been supported by the National Science Foundation' Materials Research Science and Engineering Center (MRSEC) at Georgia Tech, the W.M. Keck Foundation and the Partner University Fund from the Embassy of France.

Researchers don't yet understand why graphene nanoribbons become semiconducting as they bend to enter tiny steps – about 20 deep – that are cut into the silicon carbide wafers. But the researchers believe that strain induced as the carbon lattice bends, along with the confinement of electrons, may be factors creating the bandgap. The nanoribbons are composed of two layers of graphene.

Production of the semiconducting graphene structures begins with the use of e-beams to cut trenches into silicon carbide wafers, which are normally polished to create a flat surface for the growth of epitaxial graphene. Using a high-temperature furnace, tens of thousands of graphene ribbons are then grown across the steps, using photolithography.

During the growth, the sharp edges of "trenches" cut into the silicon carbide become smoother as the material attempts to regain its flat surface. The growth time must therefore be carefully controlled to prevent the narrow features from melting too much.

This is an atomic force microscope (AFM) top view of side wall graphene nanoribbons showing their long range order. Credit: Georgia Institute of Technology

The graphene fabrication also must be controlled along a specific direction so that the carbon atom lattice grows into the steps along the material's "armchair" direction. "It's like trying to bend a length of chain-link fence," Conrad explained. "It only wants to bend one way."

The new technique permits not only the creation of a bandgap in the material, but potentially also the fabrication of entire from graphene without the need for interfaces that introduce resistance. On either side of the semiconducting section of the graphene, the nanoribbons retain their metallic properties.

"We can make thousands of these trenches, and we can make them anywhere we want on the wafer," said Conrad. "This is more than just semiconducting graphene. The material at the bends is semiconducting, and it's attached to graphene continuously on both sides. It's basically a Shottky barrier junction."

By growing the graphene down one edge of the trench and then up the other side, the researchers could in theory produce two connected Shottky barriers – a fundamental component of semiconductor devices. Conrad and his colleagues are now working to fabricate transistors based on their discovery.

Confirmation of the bandgap came from angle-resolved photoemission spectroscopy measurements made at the Synchrotron CNRS in France. There, the researchers fired powerful photon beams into arrays of the graphene nanoribbons and measured the electrons emitted.

"You can measure the energy of the electrons that come out, and you can measure the direction from which they come out," said Conrad. "From that information, you can work backward to get information about the electronic structure of the nanoribbons."

Theorists had predicted that bending graphene would create a bandgap in the material. But the bandgap measured by the research team was larger than what had been predicted.

Beyond building transistors and other devices, in future work the researchers will attempt to learn more about what creates the – and how to control it. The property may be controlled by the angle of the bend in the graphene nanoribbon, which can be controlled by altering the depth of the step.

"If you try to lay a carpet over a small imperfection in the floor, the carpet will go over it and you may not even know the imperfection is there," Conrad explained. "But if you go over a step, you can tell. There are probably a range of heights in which we can affect the bend."

He predicts that the discovery will create new activity as other researchers attempt to utilize the results.

"If you can demonstrate a fast device, a lot of people will be interested in this," Conrad said. "If this works on a large scale, it could launch a niche market for high-speed, high-powered electronic devices."

Explore further: Graphene sensor tracks down cancer biomarkers

More information: Hicks, J., A wide-bandgap metal-semiconductor-metal nanostructure made entirely from graphene, Nature Physics (2012).

Related Stories

Writing graphene circuitry with ion 'pens'

Mar 27, 2012

The unique electrical properties of graphene have enticed researchers to envision a future of fast integrated circuits made with the one-carbon-atom-thick sheets, but many challenges remain on the path to commercialization. ...

Topographical approaches to measuring graphene thickness

Sep 28, 2012

(—Graphene has long shown potential for use in electronics, but difficulties in producing the material to a high enough quality has so far prevented the commercialisation of graphene-based devices.

Recommended for you

Engineered proteins stick like glue—even in water

3 hours ago

Shellfish such as mussels and barnacles secrete very sticky proteins that help them cling to rocks or ship hulls, even underwater. Inspired by these natural adhesives, a team of MIT engineers has designed new materials that ...

Smallest possible diamonds form ultra-thin nanothreads

3 hours ago

For the first time, scientists have discovered how to produce ultra-thin "diamond nanothreads" that promise extraordinary properties, including strength and stiffness greater than that of today's strongest ...

User comments : 4

Adjust slider to filter visible comments by rank

Display comments: newest first

not rated yet Nov 18, 2012
So, evidently, we can already make transistors that are a *lot* faster than silicon ones by using germanium as a dopant, or probably even just replacing silicon.. probably not economical for entire CPUs though.. but germanium has a bandgap of 0.67 eV near room temp, and silicon has a bandgap of 1.1 eV. ..but from what I'm seeing with solar cells, the more bandgap the better, and it's also colder is better, just as with CPUs.. so I know you need *a* bandgap to have a transistor, but is bigger better?
not rated yet Nov 18, 2012
If you are trying to design electronics that require the least amount of power to process information, basically to switch the transistor on, then a lower bandgap would be better.
Congratulations to us on a fantastic breakthrough!
not rated yet Nov 19, 2012
so I know you need *a* bandgap to have a transistor, but is bigger better?

For electronics you want a small bandgap. Germanium is a good substrate for GaAs components - but unfortunately it (and the GaAs) are rather more expensive than the current Si components.

For photovoltaics the bandgap issues are a bit more complex. The photons deliver a certain amount of energy (dependent on its wavelength), and you want to tune the bandgap to have exactly that energy (which means you need several sorts of structures if you want to capture several parts of the spctrum)
not rated yet Nov 19, 2012
There are several major determinants of (reasonable) transistor switching speeds today. The first is the capacitance of the gate. This is determined by the properties of the gate, and the size of the active area. (Usually measured as effective gate length, EGL, which is then multiplied by the width of the transistor.) Higher voltages drive transistors to switch faster, but for a 10% increase in voltage, you have at least a 20% increase in waste heat to deal with. Recent switches to high-k metal gates would seem to be increasing capacitance, but they make it possible to have a thicker gate (which reduces capacitance) and still get the other properties you need.

The band gap compared to the operating voltage (Vcc) used to be a determinant, but with CMOS the only power consumed is in switching, not sitting there.

But the reason for using GaAs or graphene is electron mobility. Electrons have to move across the gate in switching. Graphene has an amazingly high electron mobility.