Photocurrent modeling and device characterization of a graphene photodetector integrated on a silicon nitride waveguide

Abstract

Graphene is a promising material for novel photodetector technology. Nevertheless, the responsivity of the graphene photodetectors is not at the level of its competitors. The responsivity is the key figure of merit for several photodetector applications. Thus, this thesis aims to increase the understanding of the factors affecting responsivity in unbiased and gated metal-graphene-metal photodetectors. These factors are investigated by modeling the current generating mechanisms, simulating different device configurations with the model and fitting the model to the experimental results. In an unbiased graphene detector the current generating mechanisms are photovoltaic and photothermoelectric effect. The model implies that the dominating effect is photothermoelectric. By inducing hot spots and optimizing the waveguide location in the graphene channel, yielding a steep temperature gradient over an optimal distance, the model predicts significant improvement to the photothermoelectric current and to the internal responsivity, consequently. By fitting the model to the experimental results, it was found that the model complies the trends in the photocurrent, implying that the photocurrent is, indeed, dominated by the photothermoelectric effect. Moreover, the current maximum is obtained at a specific channel doping level. The main parameter affecting the required doping with respect to the Dirac point is the graphene quality in the channel. Additionally, due to the contact effects, the asymmetric structure is likely to cause less shift in the power distribution than was expected, reducing the responsivity. Understanding the parameters affecting the responsivity paves the way for developing power-efficient and compact integrated photonic circuits.