Data Sources for Yield

Inline metrology, WAT data, FDC, and merging heterogeneous data

The Yield Data Landscape

Yield prediction requires integrating multiple data sources across the entire manufacturing flow:

Inline metrology: CD, overlay, film thickness, and other measurements taken during fabrication. Sparse sampling (5–20 sites per wafer, 5–10% of wafers).
FDC (equipment sensor data): Process conditions for every wafer on every tool. Complete coverage but indirect — must be correlated to yield outcomes.
Defect inspection: Defect counts, maps, and classifications from optical and e-beam inspection.
WAT (Wafer Acceptance Test): Electrical measurements on test structures after fab completion — transistor parameters (Vt, Idsat, Ioff), resistances, capacitances.
Sort/probe data: Die-level pass/fail and bin results from electrical testing.
Design data: Die layout features — pattern density, metal coverage, critical design rules.

Key Concept: The Data Integration Challenge

Each data source has different granularity (wafer-level, die-level, site-level), different sampling rates, and different schemas. Merging them into a unified dataset is often 80% of the ML project effort. Wafer ID and lot ID are the typical join keys, but handling missing data and mismatched sampling is non-trivial.

Understanding Yield

Yield Prediction Models

Knowledge Check

1 / 1

What fraction of the ML project effort is typically spent on data integration for yield prediction?

~80%~10%~50%~5%