An assignable cause (also called a special cause) is a specific, identifiable reason for process variation that is not part of the normal, random variation inherent to the process. When an assignable cause is present, the process is out of control — its behavior differs from its established baseline in a detectable way.

Assignable Cause vs. Common Cause

• Common Cause (Random): The natural, inherent variation present even when the process is running perfectly. Due to the cumulative effect of many small, uncontrollable factors. The process mean and spread are stable and predictable.
• Assignable Cause (Special): A specific, discrete event or change that shifts the process mean, increases variability, or creates an unusual pattern. It is identifiable and correctable.

Examples in Semiconductor Manufacturing

• Chamber Leak: Air leaking into a vacuum chamber alters gas composition and etch/deposition chemistry.
• Worn Component: A degraded electrode changes plasma characteristics.
• Wrong Recipe: An incorrect version of a process recipe is loaded.
• Contaminated Chemical: A batch of contaminated photoresist or etchant gas.
• PM Error: A maintenance task performed incorrectly — misaligned hardware, wrong part installed.
• Environmental Excursion: Fab temperature spike, vibration from construction, power quality issue.
• Raw Material Change: A new lot of wafers with different surface properties or film thickness.

Identifying Assignable Causes

• SPC Charts: Control charts detect the presence of an assignable cause through OOC signals — but they don't identify what the cause is.
• Investigation: Engineers must trace the excursion to its root cause through:
• Timeline Analysis: What changed at or before the time of the excursion?
• Tool History: Recent maintenance, recipe changes, PM actions, alarms.
• Lot Genealogy: Which lots, wafers, and process steps were involved?
• Correlation Analysis: Do OOC events correlate with specific tools, operators, shifts, or materials?
• Physical Analysis: SEM, TEM, or other analytical techniques to examine defective features.

The Goal: Eliminate Assignable Causes

• Every assignable cause should be identified, understood, and eliminated (or prevented from recurring).
• Once all known assignable causes are removed, the process is in statistical control — only common cause variation remains.
• Reducing common cause variation requires process improvement (better equipment, tighter controls, improved materials) rather than troubleshooting.

Identifying and eliminating assignable causes is the primary activity of SPC-based process control — it is how fabs systematically improve yield and reduce variability.