Get Started
Home
Topics
Search
Library
Research questionHow can surrogate models accelerate TCAD exploration of large 3D FinFET designs without losing physics across geometries?High-fidelity drift-diffusion TCAD becomes prohibitively slow as three-dimensional mesh complexity and fin count increase. Scalar surrogates can reduce runtime but omit spatial fields and may transfer poorly when device geometry changes.
AI
Machine Learning
Research Paper
Technology
Latest papersRecent research connected to this question, newest first.Mesh-Native Physics-Informed Graph Surrogates for TCAD-in-the-Loop Design Space ExplorationThe source studies a physics-informed graph surrogate operating on tetrahedral TCAD meshes for multi-fin tri-gate FinFETs. It predicts node-level electrostatic potential and electron and hole quasi-Fermi levels, uses finite-volume current-continuity residuals during training, and applies ensemble uncertainty to select designs for full simulation. Evidence comes from comparisons with Sentaurus Device, including tests of larger multi-fin arrays and Pareto-front exploration; reported performance is limited to the studied device configurations.research paper · Sep 2, 2026
Related questions
How can compact device surrogates stay physically consistent and accurate across unseen process or geometry splits with scarce data?How can physics-informed neural networks handle stiff, multiscale ion-electronic transport while supporting inverse estimation?How can parameter sweeps of spatial FitzHugh–Nagumo dynamics be accelerated without losing excitable or oscillatory behavior?How can aerodynamic surrogate models predict vehicle pressure and wall shear stress without losing global flow or local geometric effects?