Skip to content
SemiAI

Defect Model Simulator

Train, manage and simulate overlay correction recipes in one workspace Trains overlay models stage-by-stage on the selected wafer data, then applies recipes from a shared library to real measurements — previewing next-lot corrections before they reach production.

What it does

  • 01

    Physically meaningful multi-stage model training

    Trains a registered model on the selected wafer measurements to isolate systematic components across three physical stages — Inter-field (global), Intra-field (in-shot) and CPE (per-exposure residual).

  • 02

    Recipe-driven correction simulation

    Applies a recipe from the library — or a freshly trained model recipe — to real measurements, previewing the correction result before it reaches the next lot.

  • 03

    Recipe library management

    Registers, edits and removes recipes — model structure with stage and coefficient combinations — in a shared library, then reuses the same recipes across training and simulation runs.

    Recipe Library
    3 recipes
    Recipe A
    Stage 1
    Stage 2
    Stage 3
    Recipe B
    Stage 1
    Stage 2
    Recipe C
    Stage 1
    Stage 2
    +
    Add new recipe

Tell us the challenge — we'll scope it with you.

You don't need detailed data ready. We reply within two business days with a tailored deployment scenario.