Contact-Governed Macroscopic Signatures of Strain-Induced Valley Sorting in MoS$_2$ Field-Effect Transistors
Suhas Suresh Bharadwaj
arXiv:2610.07203·cond-mat.mes-hall·Published 2026-10-05
Strain engineering is a leading route to control valley-selective transport in monolayer transition metal dichalcogenides. Whether this quantum effect survives inside a fully contacted device remains an open question. Quantum transport models capture the underlying valley deflection but are limited to isolated nanoscale fragments, leaving it unclear if a full device registers strain-induced valley sorting in the bulk, at the contacts, or at all. Here, the quantum valley-deflection tensor of strained monolayer MoS$_2$ is embedded directly into a macroscopic Poisson-drift-diffusion model, extending this physics to a complete, contact-inclusive device. Applying this framework to a 300~nm monolayer MoS$_2$ transistor with a 3~nm HfO$_2$ gate dielectric at 77~K, under a localized 6.14~T pseudo-magnetic strain field, shows the resulting 63.6\% suppression of the ON-state current is not a bulk property. It is set almost entirely by the source and drain metal-semiconductor interfaces. The full suppression is reproduced when only the 15~nm contact regions are strained and nearly vanishes when the channel interior alone is strained, identifying the metal contact as the dominant site of strain-induced valley physics. Observing it requires tunneling-transparent, degenerately-doped contacts, since classical Ohmic barriers overwhelm the topological signal. A transmission-line analysis reduces this behavior to a single valley-topological contact resistance ($R_{c,V}$) and a predictive inverse-length scaling law. This framework applies to any strain-engineered 2D-material heterostructure and can be tested with standard multi-length contact structures. Ultimately, these results identify the metal-semiconductor interface as the decisive location for engineering and detecting valley-selective transport in real 2D-material devices.
TopicsGenerative Design & Molecule Optimization
Tags2d-materials diffusion-model tmd
arXiv categoriescond-mat.mes-hall, cond-mat.mtrl-sci, physics.app-ph
arXiv abstract pagePDF