Definition
Metrology quantifies specified properties, while inspection searches for anomalies or defects; together they provide the evidence used to control processes, disposition wafers, and learn yield.
Process position
Inputs
- Wafer, reticle, die, or package
- Sampling plan
- Measurement recipe
- Reference and calibration data
Outputs
- Measurements and defect maps
- Statistical signals
- Disposition and root-cause evidence
How it works
- 01Define the measurand or defect class
- 02Calibrate the system
- 03Select sites and sampling frequency
- 04Acquire data
- 05Classify and correlate signals
- 06Feed results to process and yield control
Measurement answers a defined question
A useful control plan states what property matters, where and how often it is sampled, how uncertainty compares with the process window, and what action follows an excursion. More data is not automatically better control.
Semiconductor process control relies on inspection, metrology, and data analysis across reticle, wafer, and packaging operations.
Inspection and review form a learning loop
High-throughput inspection finds candidate anomalies; review and classification determine which signals represent meaningful defect mechanisms. Correlation with process history and electrical yield turns those observations into root-cause evidence.
Inspection and metrology data can be combined with analysis to identify defect sources and support yield improvement.
Boundary: A supplier annual report describes the intended benefit of its own inspection products; it does not independently quantify yield gains.
Sources
Citations support the tagged claims above. Access dates record when Maha Strategies last checked the public source.
- [1]Annual Report: Process Control and Yield Management · KLA · 2024 · accessed 2026-08-13
- [2]Annual Report: Inspection, Metrology, and Yield Analysis · KLA · 2019 · accessed 2026-08-13