Abstract

Graphene is a fascinating 2D material, known for its unique charge transport and optical properties due to its low dimensionality and unique band structure. Although photocurrents in graphene have been heavily studied, there is little consensus on the photogeneration mechanisms contributing to photocurrents in applied graphene devices. There are two primary contributors to photocurrent that we investigate herein: the photovoltaic effect and the photothermoelectric effect. For short-circuit measurement configurations, the photobolometric effect is negligible due to requiring a non-zero bias voltage. Understanding the role each mechanism plays can aid in the design and operation of graphene-based photodetectors. If the mechanisms’ contributions are tunable, we can design a photodetector such that the most photoresponsive mechanism dominates, resulting in greater photosensitivity.

We report simultaneous photocurrent and micro-Raman measurements in mono- and bi-layer graphene rectangular transistors of at least 2 μm width and 4 μm length, on Si substrate with 300 nm SiO2 layer at room temperature and ambient pressure in source-drain configuration, seeking to disambiguate the contribution between photovoltaic and photothermoelectric
effects from spatial and power dependencies. Devices presented demonstrated photoresponsivities of up to (229.4±2.5) μA W−1 for bilayer and (159±2) μA W−1 for monolayer when comparable in size and shape. To ensure consistent and comparable results throughout our experiments, we measure Raman spectra during the photoresponse measurements, as well as gate sweeps before each set of measurements. We found that in order to yield reproducible results without a vacuum, we must employ a laser-annealing technique to reduce the influence of surface moisture. Adsorbed water molecules dope the graphene with holes, affecting transport, but through laser annealing we can ensure that the sample is in the same initial state before each experiment. We measured the photocurrents while observing their dependence on laser spot position on the sample and on laser power intensity. The photocurrent exponents extracted ranged from 0.6 to 1.1, indicating that likely both photovoltaic and photothermoelectric effects contribute to the total photocurrent in varying amounts, with photovoltaic being the dominant effect close to the contacts, likely due to limitations of the electron mean free path of around 7-200 nm and laser spot FWHM between 0.3 μm and 0.4 μm. We also noted that some sample properties evolved over the course of hours or days, suggesting perhaps a shift in charge density during the course of longer experiments experiment.