Imaging a semiconductor sandwich

A technological mash-up between graphene and the semiconductor gallium arsenide as characterised by optical microscopy and Raman spectroscopy and other techniques could pave the way to hybrid electronics devices that bridge the gap between current silicon circuitry and future molecular electronics.

Graphene is a modified form of the all-carbon pencil lead material graphite and is being touted as the material of choice for a future generation of computer chips to augment, or even usurp, silicon. It is akin to a single layer of graphite and has some fascinated properties that have gradually unrolled since its discovery in 2004 by Andre Geim and colleagues at The University of Manchester. Graphene has already revealed much exquisite new physics, not least because of its thickness and the fat that its electrons can incredibly freely.

Gallium arsenide in contrast is a semiconductor material whose properties can be tailored and it is become the starting material of choice for the development of fast experimental electrical and opto-electronic components. One property of gallium arsenide that meshes nicely with the graphene lattice structure is that it can be produced with atomic-layer smoothness and so could act as the perfect substrate for fabricating graphene devices.

Now, scientists at the Physikalisch-Technische Bundesanstalt (PTB) in Braunschweig, Germany, the country's national metrology institute, have now developed a way to make graphene visible on gallium arsenide, something that was previously only possible on silicon/silicon dioxide.

Franz-Josef Ahlers and colleagues M. Friedemann, Klaus Pierz, and Rainer Stosch, at PTB, are experts in precision measurements, and so have plans for studying the electrical properties of their materials in great details. However, they first needed to confirm that the graphene layers deposited on the gallium arsenide are as near perfect as possible. For this they turned to Raman spectroscopy and optical microscopy.

Intriguingly, the graphene layer is less than on thousandth of the wavelength of light in thickness. So they exploited the principle of an anti-reflective layer to image the graphene optically. If one superimposes a very thin, almost transparent layer of one material on to another, i.e. graphene on gallium arsenide, then the reflectivity of the underlying layer will change visibly. To image the graphene, required an additional five wafer-thin layers of aluminium arsenide (AlAs) to enhance the interference effects that would be seen with optical microscopy exploiting the anti-reflective layer phenomenon.

Ahlers explains: "Even with optically similar materials, as AlAs and GaAs, it is possible, in a manner of speaking, to 'grow' interference effects", Ahlers PTB department head, says, "This principle is known from optical interference filters. We have simply adapted it for our purposes."

In order to obtain the information they needed, Ahlers and his colleagues calculated the optical properties of different GaAs and AlAs layers and optimized the layer sequence so that they could produced a sufficiently good detectability of graphene. Once they had the recipe in hand, they then used the PTB's molecular beam epitaxial facility to produce a corresponding GaAs/AlAs crystal atom layer. This substrate was then coated with graphene.

The results are different to those seen with silicon/silica, but as they predicted single carbon layers visible at some wavelengths of visible light using a green filter, for instance, to limit the wavelength range in optical microscopy.

"In our images, all lighter areas of the dark GaAs are covered with graphene," Ahlers says. The contrast in the image allows the researchers to determine the number of individual layers of atoms. The marked areas are 'real', that is, single-layer graphene. But next to them, there are also two, three and multiple layers of carbon atoms, which also have interesting properties. The team confirmed these findings with Raman.

"Our reflectivity data, and their agreement with the reflectivity calculations, alone would already strongly support that indeed mono-, bi-, tri-, and more layers of graphene were observed," he explains. On the other hand one can never be perfectly sure whether another processing artefact, such as glue residues from the tape employed for exfoliation, or resist residues from other lithographic preparation steps) could cause similar reflectivity changes." The Raman spectra provide fingerprint identification of the materials, he told SpectroscopyNOW. "The Raman results were important since they showed the known fingerprint-spectrum of graphene, and they also enable one to distinguish between mono-layer (i.e. 'true') graphene, and bi- or more layers: One of the graphene Raman lines is an unstructured, single-component line only for mono-layers," Ahlers adds

The main advantage of being able to image graphene on a substrate, such as gallium arsenide optically, is that the fabrication process of electrical components based on graphene surfaces should now be much easier to follow.