Third-author research / Published 2026
Numerical Investigation of Short-Channel Effects and RF Performance in Top-Gate In2O3 Thin-Film Transistors
A third-author contribution to a two-dimensional TCAD study of short-channel and RF scaling in top-gate In2O3 thin-film transistors.
Reported transition region
Numerical long-channel to short-channel transition in gate length.
Simulated fT at 20 nm
Extracted from the study's simplified small-signal framework.
Simulated fmax at 20 nm
A numerical result, not a fabricated-device measurement.
Problem
Ultrathin In2O3 is promising for high-current and high-frequency electronics, but aggressive lateral scaling makes short-channel behavior increasingly important. A systematic view was needed of how gate length changes threshold behavior, transconductance, and RF figures of merit in a top-gate device architecture.
Approach
The paper uses two-dimensional Silvaco Atlas simulation for a top-gate In2O3 TFT with a 1.5 nm channel and 7 nm HfO2 dielectric across gate lengths from 20 to 700 nm. Effective material and interface parameters are calibrated against a previously reported back-gate experimental device, then transferred to the independently constructed top-gate simulation platform.
The numerical analysis connects DC characteristics, DIBL, transconductance, lateral electric field, field-dependent mobility, and a simplified small-signal RF equivalent circuit.
My contribution
I am the third author. My contributions, as recorded in the published CRediT statement, were validation, data curation, and writing-review and editing. I do not present the simulation setup, primary analysis, or reported device performance as work completed independently by me.
Reported results
The study identifies a numerical transition near a 100 nm gate length. For gate lengths at or below 100 nm, it reports a negative threshold-voltage shift and DIBL rising to approximately 130 mV/V, together with non-classical transconductance scaling linked to a crossover between field-assisted transport and gate-controlled channel modulation.
At 20 nm gate length, the simulation reports fT = 124.32 GHz and fmax = 157.64 GHz. The paper also reports a less distinct transition in fT scaling than classical long-channel expectations and multiple fmax scaling regimes shaped by capacitance, transport enhancement, and short-channel non-idealities.
Evidence boundary
Every device and RF value highlighted here is a numerical simulation result, not a measurement from a fabricated top-gate In2O3 transistor. Experimental data from prior literature are used for calibration context, but the target device sweep itself remains a drift-diffusion TCAD study with a simplified intrinsic RF model. Tunneling, self-heating, substrate dependence, and additional extrinsic parasitics remain future extensions.