Modelling topological and magnetic materials for charge and spin-based devices

Abstract

The imminent halt of Moore’s law and discontinuation of scaling of transistors based on three-dimensional materials, e.g., silicon, has prompted researchers to look for different ma- terials and device systems apart from the conventional ones to form the backbone of the electronics industry of the future. Topological insulators (TIs) open a vast avenue to realize devices with high ON current and low power consumption. TIs are a class of materials with topologically protected edge states which are spin-polarized and robust against impurity scattering. The possibility of spin-polarization in TIs and efficient transfer of spin-current in soft-layered magnets opens another avenue of research for realizing fast memory devices. In this dissertation, first, we model carrier transport through imperfect two-dimensional (2D) TI ribbons. In particular, we investigate the impact of vacancy defects on the carrier trans- port of 2D TIs. We show that carrier transport through the topologically protected edge states is robust against a high percentage of defects (up to 2%), whereas the carrier trans- port through the bulk state is strongly suppressed due to backscattering. We show that the suppression of bulk transport in 2D TIs can be used to design devices using 2D TI ribbons. Next, we develop a computational method to model the magnetic interactions in layered magnetic materials and calculate their critical temperature from the first principles, taking into account both the magnetic anisotropy as well as the out-of-plane interactions. We ap- ply our method on Cr-compounds: CrI3, CrBr3, and CrGeTe3, and FeCl2, and show that our predictions match well with experimental values. Using the same model we next inves- tigate the magnetic order in two-dimensional (2D) transition-metal-dichalcogenide (TMD) monolayers: MoS2, MoSe2, MoTe2, WSe2 , and WS2 substitutionally doped with period-four transition-metals (Ti, V, Cr, Mn, Fe, Co, Ni). We show that five distinct magnetically or- dered states can exist among the 35 distinct TMD-dopant combinations including the non- magnetic (NM), the ferromagnetic (FM) with out-of-plane spin polarization (Z FM), the out-of-plane polarized clustered FMs (clustered Z FM), the in-plane polarized FMs (X–Y FM), and the anti-ferromagnetic (AFM) state. Most remarkably, we find from our study that V-doped MoSe2 and WSe2, and Mn-doped MoS2, are the most suitable candidates for realizing a room-temperature FM at a 16–18% atomic substitution. We then compare three first-principles methods (the MC, the Green’s function, and the RNSW) of calculating the Curie temperature in 2D FMs in the presence of exchange anisotropy, modeled using the Heisenberg model. We find that the Curie temperature obtained from the Green’s function in high-anisotropy regimes is higher compared to MC, whereas the Curie temperature cal- culated using the renormalized spin-waves (RNSW) is lower compared to the MC and the Green’s function for all anisotropies. Finally, we present a theoretical model to simulate spin- dynamics and spin-induced switching in a semiconductor-ferromagnet heterostructure. Our theoretical model combines the non-equilibrium Green’s function method for spin-dependent electron transport and time-quantified Monte-Carlo for simulating magnetization dynamics. We use the adiabatic approximation for combining the electron dynamics and the magne- tization dynamics. We study spin-induced switching in a 2D TI-FM interface. Finally, we show that for a certain range of magnetic exchange parameters (or certain materials), it is possible to change magnetic domains in a 2D FM using spin-torque from TIs, which can be used for designing high-speed memories.