Engineering documentation for yield teams.
Detailed PDF notes on the methodologies underlying Wafercadence — written for process engineers who want to understand how the algorithms work, not just that they work.
Technical note library
Each note provides methodology-level detail — the kind of documentation you'd need to present findings to a process committee or integrate the approach into an internal analysis workflow.
Defect Classification Methodology: CNN Architecture and Training Protocol
Describes the convolutional neural network architecture used for inline defect classification, per-process-node training protocol, confidence scoring system, and validation methodology against KLA SURFmonitor and AMAT ComPlus inspection outputs.
Spatial Pattern Detection: Edge Ring, Arc, Scratch, and Center Spot Classification
Covers the spatial clustering algorithm used to detect systematic defect patterns at low D0 counts where standard SPC rules are statistically blind. Includes classification accuracy benchmarks at 14nm and 7nm nodes.
KLARF Ingest Specification: Format Handling and Data Normalization
Defines the complete KLARF (KLA Results File) parsing specification used by the Wafercadence ingest layer — including KLARF 1.0/1.1/1.2 format variants, field normalization, coordinate system alignment, and handling of mixed scan modes.
Yield Correlation Approach: Layer-to-Device Kill Ratio and WAT Integration
Explains the statistical method used to compute layer-to-device kill ratio from in-line inspection data correlated against WAT probe data and sort yield. Covers Pearson correlation thresholds, lag window selection, and multi-layer attribution disambiguation.
Technical notes are available to yield engineers and process engineers under NDA. Contact [email protected] or use the request form and we'll arrange access within 1 business day.
Need a specific technical question answered?
Talk directly to our yield engineering team. We can walk you through any of these methodologies in a 30-minute session.
Request Technical Evaluation