Meaning
Quality control phenomena describe the non-random, localized grouping of physical imperfections or errors across the surface of a manufactured component, such as a silicon wafer or a flat panel display. This uneven distribution, referred to as spatial defect clustering, indicates that the underlying manufacturing issue is not a random occurrence but is tied to a specific localized process variation. It is used to analyze yield losses in high-precision industries.
This analysis stops when the defects are distributed uniformly across the production batch.
Failure Mechanism
Process variations often arise from localized temperature changes, chemical contamination, or mechanical stress during the production cycle. When engineers analyze spatial defect clustering, they can trace the root cause to a specific malfunctioning nozzle, heating element, or handling arm. This tracing helps to prevent widespread batch failures.
It allows for targeted maintenance before the entire run is ruined.
Process Diagnosis
Yield management software uses statistical algorithms to identify these clusters in real time. When a high density of defects is detected in one area of a wafer, the system flags the issue for immediate correction. This automation minimizes waste.
Contractual Claim
Commercial agreements often include clauses that allow buyers to reject batches where the defects are clustered, even if the total defect count is below the allowable limit. This rejection right protects the buyer from receiving components that are structurally compromised.