arXiv subjects

Tian-Li Wu

Publications and source records attributed to Tian-Li Wu.

5 recordsLinked to original sources

arXiv2026

A Unified Physics-Aware Quantum Machine Learning Framework across Power GaN HEMTs and Logic Nanowire FETs: Predicting Unseen Process Splits and Held-Out Geometry Combinations with Lower Error and Tighter Split-to-Split Variability

We present a unified reinforcement-learning (RL) framework that discovers compact parametrized quantum circuits (PQCs) for data-scarce device modeling. A graph neural network (GNN) policy optimized by proximal policy optimization (PPO) searches circuit architectures using leave-one-group-out cross-validation (LOGOCV) error on held-out process or geometry groups as the reward. The framework achieves the lowest mean absolute error (MAE) on all 11 targets versus six classical baselines, with 59% lower error (Ioff) and 81% tighter fold variability (VTH) for HEMTs and 84% lower error (VTH, SS, Ioff) and 82% tighter fold variability (Ioff) for NWFETs. These results demonstrate the potential of RL-selected, classically simulated PQCs as compact surrogates with low OOD error and improved physical consistency, despite imposing no explicit physical constraints, penalty terms, or device-specific equations, on the two evaluated device datasets.

cs.AI↗

arXiv2026

Hybrid Classical-Quantum Neural Networks for Multi-Characteristic Co-Optimization of Recessed-Gate AlGaN/GaN MIS-HEMTs

Optimizing recessed-gate AlGaN/GaN MIS-HEMTs requires accurate multi-characteristic models, but experimental semiconductor datasets remain costly and encode process-induced variability that simulations cannot faithfully reproduce. This work proposes a hybrid classical-quantum neural network (HQNN) for joint optimization of six electrical targets from a 24-dimensional fabrication/process vector. We systematically screen quantum-circuit templates to extract circuit-design guidance, then select a final HQNN and compare it directly with classical baselines. On 468 experimental fabricated devices spanning 17 process splits, the selected HQNN, Circuit (13, 5) at L = 2, reduces overall normalized root mean square error (nRMSE) by 24.4% relative to ANN. Target-wise, the HQNN lowers Vth,lin RMSE from 0.297 V to 0.270 V, Vth,rev RMSE from 0.278 V to 0.263 V, DeltaVth RMSE from 0.049 V to 0.045 V, SS RMSE from 22.22 mV/dec to 19.87 mV/dec, and Id RMSE from 5.75 x 10^-8 A to 4.35 x 10^-8 A, while Ion RMSE remains competitive (0.053 A vs. 0.056 A). Controlled ansatz ablations further show that performance depends strongly on architecture: parameter count, depth, and two-qubit gate count correlate positively with accuracy, expressibility (DKL) correlates negatively, and controlled-rotation entanglers outperform static controlled-NOT (CNOT)-based circuits in aggregate. A depolarizing-noise study on a representative 4-qubit circuit further suggests that comparable HQNNs may be trainable or deployable on near-term quantum hardware.

quant-ph↗

arXiv2025

A Novel Methodology of Visualizing Orthorhombic Phase Uniformity in Ferroelectric Hf0.5Zr0.5O2 Devices Using Piezoresponse Force Microscopy

Ferroelectric Hf0.5Zr0.5O2 (HZO) thin films are promising for next-generation memory and logic devices due to their CMOS compatibility and scalability. The spatial uniformity of the orthorhombic (O) phase is crucial for optimizing ferroelectric properties like remnant polarization. This work introduces a novel piezoresponse force microscopy (PFM) approach for 2D mapping of O-phase uniformity in HZO films (5 nm, 9 nm, and 20 nm), further quantifing O-phase distribution by distinguishing polarized O-phase regions from non-polarized tetragonal/monoclinic (T/M) phases. Our results reveal that the 9 nm film exhibits the most uniform O-phase and highest remnant polarization. This PFM-based method enables comprehensive phase characterization without requiring complicated facilities, broadening access to phase analysis and advancing ferroelectric thin-film research for memory and logic applications.

cond-mat.mtrl-sci↗

arXiv2025

Demonstration of TFTs 3D Monolithically Integrated on GaN HEMTs using Cascode Configuration with High Breakdown Voltage (>1900V)

This study demonstrates 3D monolithic integration of amorphous indium-gallium-zinc oxide (a-IGZO) thin-film transistors (TFTs) on Gallium Nitride (GaN) high electron mobility transistors (HEMTs) in a cascode configuration, achieving high breakdown voltage capabilities exceeding 1900 V. Two device configurations, differing in a-IGZO channel thickness (30 nm / 10 nm), are fabricated and evaluated. Sample B, with a 10 nm a-IGZO channel, demonstrates superior electrical performance, including a high ON/OFF current ratio (~10^7), low subthreshold swing (SS), and a high breakdown voltage exceeding 1900 V comparable to standalone GaN power HEMTs. The results highlight the feasibility and potential of 3D integrated TFT on GaN power HEMTs, paving the way for new opportunities for the TFTs for high voltage applications.

physics.app-ph↗

arXiv2019

Design and analysis of high electron mobility transistor (HEMT) inspired III-V electro-optic modulator topologies

III-V heterostructure based high electron mobility transistors (HEMTs) offer superior performance as compared to CMOS silicon transistors owing to the high mobility in the 2D electron gas (2DEG) channel at the heterostructure interface. Gallium nitride (GaN) based HEMTs are also suitable for high power and high temperature applications. GaN has a rich offering of material properties spanning domains of nonlinear optics, piezoelectric micro-electro-mechanical systems (MEMS), and monolithic microwave integrated circuits (MMICs). In this paper, we propose HEMT inspired III-V electro-optic modulator topologies that could potentially outperform silicon photonic modulators. We analyze the electroabsorption and electrorefraction on account of the 2DEG interaction with light and present a design framework to selectively leverage the desired mechanism of modulation. Our analysis suggests that modulation index of electrorefractive modulation in a HEMT-like structure is comparable to silicon photonic modulators, albeit with much higher electron mobility and thereby much higher modulation rates.

physics.app-ph↗