Photodetectors – devices that detect light by converting optical signals into electrical current – are routinely employed in applications such as communications, sensing and imaging. Most light detectors are made of III-V semiconductors like gallium arsenide and they work by absorbing photons to produce electron-hole pairs that then separate and generate an electrical current.

Graphene – a sheet of carbon just one atom thick – has a number of unique physical and mechanical properties that make it ideal for detecting light. One important advantage is that electrons move much faster through graphene than through other materials. They behave, in fact, as if they had no mass and travel at 1/300 the speed of light. These particles are called massless Dirac fermions and their behaviour could be exploited in a host of applications, including transistors that are faster than any that exist today.

Graphene is also very good at absorbing light over a very wide range of wavelengths, ranging from the visible to the infrared. III-V semiconductors do not work over such a wide range.

Until now, researchers believed that graphene absorbed light via at least five different mechanisms – through photovoltaic, thermoelectric or bolometric effects, and by photodesorption of oxygen or phototransistor amplification. A team led by Phaedon Avouris at IBM’s TJ Watson Research Center in Yorktown Heights, New York, has now looked at all of these effects in detail in photoconductivity experiments on graphene field-effect transistors.

The IBM researchers obtained their results by illuminating the FETs with focused infrared laser light and then measuring the resulting photocurrent using a lock-in technique. The experiments were performed on homogenous graphene, rather than on graphene p-n junctions, as in many previous such experiments, which allowed the team to measure the intrinsic photoresponse of the carbon material.

Hot electrons and cooler lattice

When graphene absorbs light, electron-hole pairs are excited that then rapidly interact with other electrons and holes. These interactions increase the overall temperature of the electrons, explains Freitag, but the electrons remain “hot” because they couple poorly to the carbon lattice and thus transfer their heat to it only very slowly.

“It is these hot carriers that produce the photovoltaic current in graphene,” he told “When the lattice temperature does increase, however, this changes the electron mobility and produces a bolometric current in the opposite direction. At low charge densities, the photovoltaic effect thus appears to dominate while at high electron doping levels, it is the bolometric effect,” he said. “We found that we could switch between these two photocurrent response mechanisms by changing the electron density in the graphene FET using a back gate voltage.”

Knowing exactly how photocurrent is produced in graphene will be crucial for improving the efficiency of photodetectors made of this material. For example, changing the dielectric on which the graphene transistor is placed will alter electron-phonon coupling in the device and so change which effect, bolometric or photovoltaic, dominates.

The team is now busy working on improving graphene’s light absorption in the mid-infrared range by exploiting intrinsic graphene plasmons (collective electron oscillations in the material lattice). “We believe that graphene has a clear advantage in this energy range thanks to its Dirac fermions,” said Freitag.

The current work is detailed in Nature Photonics.