(Invited) An Overview of Doping Studies in MoS2

Abstract

The advent of 2D materials in recent years has brought the potential for new device applications at ever decreasing transistor design nodes. Transition metal dichalcogenides have received much of that attention, with much of the focus on MoS2 thanks to its desirable bandgap. Studies have shown the advantages of MoS2 devices in key performance metrics such as a high ION/IOFF ratio and a subthreshold swing that approaches or potentially exceeds the fundamental 60 mV/dec limit. However, the contact resistance remains prohibitively high, and questions remain regarding the ability to dope these materials. Numerous routes for doping MoS2 have been reported, including surface charge transfer via molecular adsorption, cation substitution during growth, chalcogenide substitution via immersion in solution, and plasma-assisted doping. The effectiveness of these doping methods is displayed by reductions in contact resistance, increases in drive current, and in some instances a change from n-type to p-type behavior. This talk will review these recent reports on MoS2 doping. Recent experiments have begun to explore the possibility of doping MoS2 by ion-implantation, which is a legacy doping technique that is often avoided with 2D materials due to their ultra-thin nature. These experiments utilize ultra low implant energies of 200 eV which places the species of interest (Cl+ and Ar+) completely within the 3-5 layers of MoS2 being investigated. Cl is potentially a donor in MoS2 whereas Ar was used to determine the effect of the implant damage on the electrical properties. These studies focused on room temperature implants at doses between 1 x 1013/cm2 and 1 x 1015/cm2. Preliminary structural studies utilizing angle-resolved XPS indicate a Cl peak is around 1 nm deep into the MoS2, in rough agreement with Monte Carlo TRIM calculations. STM studies have revealed a number of defects present after implantation, including sub-surface defects that may be sulfur vacancies as well as surface sulfur vacancies. The density of these defects track with the dose. Relative to the pristine surface, STS studies have shown little change in tunneling current associated with the sub-surface defects, but enhanced tunneling current associated with the sulfur vacancies. It was observed that as the dose increases the oxidation of the surface as measured by XPS after exposure to the air increases. Additionally, we are performing transmission line measurements to observe the electrical effect of the implantation. Using lithography, transmission line masks were patterned onto individual exfoliated flakes. Photoresist was used as a mask to protect the channel and the implantation preceded contact metal (Ni/Au) deposition. Thus, only the contact area was implanted. Prior to implantation, primarily non-linear IDS-VDS curves are observed for few-layer (<=10 layers) MoS2 flakes with no applied back-gate voltage. After implantation at lower doses, linear IDS-VDS curves are observed implying more ohmic contacts. This is prior to annealing implying the electrical nature of the implant damage is contributing to better contacts. However the effect is greater for Cl+ implants than Ar+ implants implying that the effect is not strictly a damage-related phenomenon. At higher doses the contacts went back to being non-ohmic and this may be associated with the aforementioned oxidation after higher dose implantation. The dose dependence of the role of implant doping on contact resistivity will be discussed as well as the results from post implant annealing to facilitate dopant activation.