APC (Advanced Process Control)

Notes

APC is how fabs hold overlay and thickness budgets that are far tighter than any single tool's open-loop repeatability.

Overview

APC, short for advanced process control, is the family of methods a fab uses to keep its processes on target automatically, by feeding measurement and tool sensor data back into recipe settings, alarms and decisions instead of relying on fixed recipes and manual checks. It includes run-to-run control, fault detection and classification, statistical process control and virtual metrology. The aim is tighter and more consistent results, fewer scrapped wafers and faster reaction when something drifts.

How it works

Run-to-run control adjusts a recipe between lots or wafers. After a step, metrology measures the result, a model estimates how far the process has moved, and the controller changes a setting such as time, dose or position for the next run. Control can also work forward, passing measurements from an earlier step ahead so that a later step compensates. Lithography is a leading example: overlay and critical dimension data are fitted to models and sent back to the scanner as corrections, including higher-order terms, for the next lot.

Fault detection and classification watches the tool rather than the wafer, using data collected from the equipment during each run to flag abnormal behaviour and identify its likely cause, so that a problem can be stopped before many wafers are affected. Statistical process control charts track key measurements and alert engineers when a trend or limit is breached. Virtual metrology uses statistical or machine learning models to predict a wafer's result, such as a film thickness, from tool parameters and sensor data, filling the gaps between real measurements.

These pieces must share data with one another, with equipment and with other factory systems. The SEMI E133 standard defines interfaces for such process control systems and names run-to-run control, fault detection, fault classification, fault prediction and statistical process control as examples. In practice, the value of APC depends heavily on the quality and timeliness of the data it receives.

Examples

• Overlay data from each lot fitted to a correction model and sent back to the scanner for the next lot.
• A virtual metrology model predicting silicon nitride film thickness in chemical vapour deposition from tool data.
• Fault detection applied to metrology tools so that their measurements stay consistent from day to day.
• Process control software exchanging data with equipment and factory systems through interfaces defined in SEMI E133.

Why it matters

APC lets a fab hold tighter tolerances with less measurement and less manual chart watching, which matters more as process windows shrink. Process control and APC engineers build and tune the controllers, SPC engineers own the charts and rules, equipment data engineers collect the sensor data, and machine learning engineers build virtual metrology models.

In the chip lifecycle, control strategies are designed during process development and ramp, when models are first trained, and they run continuously in volume production as one of the main defences against drift and excursions.

It matters most to: Advanced Process Control Engineer, Equipment Data Acquisition Engineer, ML Engineer, Virtual Metrology, Process Control Engineer, Statistical Process Control Engineer

Common confusions

Not the same as advanced process control in the process industries

In refineries and chemical plants, APC usually means model predictive control of continuous processes. In a fab it mostly means run-to-run control and fault detection on discrete wafer lots.

Not the same as SPC

Statistical process control monitors measurements and raises alerts but does not itself change the recipe. It is one component within APC.

Related standards

• SEMI E133 (interfaces for automated process control systems)

Sources

1. SEMI E133: Specification for Automated Process Control Systems Interface (opens in a new tab), SEMI, 2026-10-22
2. How Overlay Keeps Pace With EUV Patterning (Laura Peters, 9 August 2022) (opens in a new tab), Semiconductor Engineering, 2026-10-22
3. Systematic Yield Issues Now Top Priority At Advanced Nodes (Laura Peters, 6 December 2022) (opens in a new tab), Semiconductor Engineering, 2026-10-22
4. Virtual metrology (Wikipedia) (opens in a new tab), Wikimedia Foundation, 2026-10-22
5. Advanced process control (Wikipedia) (opens in a new tab), Wikimedia Foundation, 2026-10-22

Related terms

• Run-to-run control
• FDC
• Virtual metrology
• SPC

Where it appears on the Hub

Used on 12 pages of the Hub. In each group, the pages that mention it most come first.