Controlling Defects in CVD Grown Graphene : Device Application Perspective

Abstract

Necessity is the mother of all inventions. With Si hitting the speed bottleneck, newer materials to replace Si are being sought out. The ex-foliation based experiments on graphene by Geim and Novoselov at this point was perfect as many of its physical properties were fascinating from an electronics standpoint and hence it was very soon projected as a Si replacement for logic applications. In addition, graphene is also an attractive alternative to applications such as radio frequency devices, ultra-sensitive mass/chemical sensing, high-speed optoelectronics and transparent conductors for photo-voltaic applications. While the widespread success and utility of Si can be attributed to easy availability of source material and the ability to synthesize large areas of ultra high quality material, chemical vapor deposition (CVD) is the only available method to controllably produce large area monolayer graphene. CVD graphene is however polycrystalline and therefore defective. Hence, in order to promote graphene towards large-scale commercialization, it is necessary to be able to grow spatially homogeneous graphene with tailored defect densities. Transfer of atomic layers of graphene from the substrate on which it is grown, a Cu foil typically, on to an insulating substrate for electrical measurements is typically a major defect inducing step. Hence, a direct transfer-free fabrication of suspended device using graphene grown on thin films of electro-deposited Cu was attempted and successfully reported for the first time. Though it was shown that the fabrication process itself did not introduce any additional defects, the maximum obtained mobility on such fabricated structures was 5200 cm2/V·s. This value is lower than reported values in literature and thus improvements for electronic applications warranted further optimization. However, limitations on ability of electro-deposited Cu films (melting point of 1083 ◦C) to withstand high temperatures, 1000 ◦C, impeded further optimizations. Hence, growth on Cu foils was taken up. On Cu foil, we were able to identify the roles of the growth kinetics and system thermodynamics on the final quality of graphene. Specifically, by carefully altering the conditions during appropriate growth phases, we were able to obtain graphene films of tunable defect densities with motilities ranging from 200 - 20000 cm2/V·s. Using a host of characterization Techniques like electrical transport, Raman spectroscopic measurements, TEM imaging and water permeation studies, we find that the defect densities in graphene are largely concentrated at the boundaries, while the bulk of the graphene grain remains pristine. Further investigations revealed a thermodynamic correlation between the growth conditions and quality of the grain boundary in terms of defect density and structure. In addition to the influence of defects in graphene on charge mobility as seen before, their impact on the device contact resistance and charge transport hysteresis in graphene field effect transistors were also investigated. With a careful control on the film defect density, we were able to demonstrate devices with low contact resistance (1000 Ωµm ) and tunable hysteresis behavior. Finally, alternate substrates for graphene and its impact on the carrier densities were explored. Non-polar substrate SiO2 and polar substrates such AlN and AlGaN were chosen. On AlN, we obtained higher carrier mobility due to reduced phonon-electron scattering and a higher ’P’ doping behavior due to piezo-electric effects. Hence, to leverage the previous observation, novel FET device architecture with a HEMT based substrate using AlGaN was demonstrated.