Inspection systems use multiple sensing modes because each produces a different form of wafer evidence. Optical imaging, laser scattering, and electron-beam methods can examine surfaces or features, while the resulting data help locate particles, scratches, pattern defects, and dimensional irregularities. Using these modalities broadens detection across distinct defect types and process variations.
Reference comparison turns raw measurements into actionable defect information. The system compares wafer images or measurements with reference patterns, then records where observed features depart from the expected pattern. Engineers can classify findings by size and type, separating isolated particles or scratches from pattern defects and dimensional irregularities. This organization supports consistent review and process monitoring.
Inspection results become especially useful when viewed across manufacturing steps rather than as isolated observations. Data from lithography, deposition, etching, and cleaning can reveal where process drift begins and whether a recurring defect follows a particular operation. Statistical process control uses these measurements to monitor variation, allowing engineers to identify emerging problems before they substantially affect yield or reliability.
A typical inspection workflow scans the wafer surface, collects optical, laser-scattering, or electron-beam observations, and compares the resulting measurements or images with reference patterns. The system then locates and classifies detected features by size and type. Engineers review this output to connect defects with upstream processing and decide where intervention or closer monitoring is needed.
Engineers apply inspection data after or around lithography, deposition, etching, and cleaning to track how each operation affects wafer quality. Examining results at these points can expose process drift and identify the step associated with a defect pattern. This timing supports early intervention instead of waiting for reduced integrated-circuit yield, reliability, or device performance to appear.
Rather than treating every defect as an isolated failure, engineers use inspection measurements to observe variation over production and relate changes to process operations. Defect classifications by size and type provide a structured basis for recognizing drift. In engineering practice, that feedback supports targeted corrective action, improved manufacturing efficiency, and more consistent production quality.