Experimental and theoretical investigation of the graphene ribbons nonlinear electrical behavior

In this study, we experimentally investigate the carrier transport nonlinear behavior in back-gated graphene transistors. The graphene ribbons are monolayer graphene formed by chemical vapor deposition process and transferred on SiO2/Si substrate. The electrical characterization with a micro prober setup of hundreds of devices has been performed under vaccuum and for a temperature ranging from 300K down to 77K. We can distinguish two categories of electrical behavior, respectively, linear and nonlinear electrical output characteristics. The linear behavior excludes the impact of the Schottky barrier between both metallic source and drain contacts and graphene. The study is focused on the impact of the technological parameters on the electrical characteristics of back-gated graphene transistors. Equivalent circuit simulations show that, first, the output characteristics slopes are strongly influenced by the doping concentration value. Second, during the fabrication process the graphene is doped by its environment, which makes the doping parameter experimentally difficult to control. The in situ Raman characterization along the ribbons area highlights interesting results for 2D peaks shift and intensity. The Raman shift variation corresponding to the position of the 2D peak between all measured spectra extends from 2697 cm-1 to 2700 cm-1 for devices exhibiting a linear behavior. While, for the devices exhibiting a nonlinear behavior, the variation range is more important, from 2686 cm-1 to 2704 cm-1. These results reveal that the doping level between two regions within the ribbon area is clearly different. Furthermore, the Raman mapping characterization shows a high variability of the 2D peak intensity. The extraction of the ribbons doping profile for both the minimum doping corresponding to the charge neutrality point and the doping issued from the gate oxide are in agreement with these results. Indeed, the electrical behavior appears strongly related to the doping concentration and distribution along the graphene layer.