Chamber Calibration

Chamber calibration and setup is the process-side qualification that happens after a chamber is physically mounted and docked — connecting gas lines, RF matching networks, and sensors, then running a defined sequence of baseline measurements and recipe configuration until the chamber proves it behaves like every other chamber of the same type on the fab floor. A bolted-down, leak-tight chamber is a mechanical fact; a calibrated, configured chamber is the thing a process engineer is actually allowed to run product on.

Mechanically docked
Leak check passed
Calibration and setup sequenceProcess engineering takes over
Gas/RF/sensor hookup
MFCs, matching network, pyrometer
Baseline pump-down
Base pressure and leak-up rate
Sensor calibration
OES, RF probe, thermocouple zero
Chamber matching
Compare to golden/reference chamber
Qual wafers
Run test recipe, check results
Released to production
Cleared for product wafers

Chamber matching is the step that turns an individual chamber into an interchangeable unit of fab capacity. A fab with five identical etch chambers needs any wafer lot to get the same result regardless of which chamber it happens to run in, so calibration isn't just "does this chamber work" — it's "does this chamber's etch rate, uniformity map, and critical-dimension bias match the golden reference chamber within a tight statistical band." A chamber that technically functions but drifts outside that matching window gets flagged and re-tuned before it's allowed back into the production rotation, even if nothing about it looks broken.

Chamber Calibration — Matching to the Golden Reference a new chamber isn't qualified until its results overlay the fleet's reference chamber Golden reference chamber etch rate, uniformity, CD bias — baseline New / recalibrated chamber same metrics, measured fresh statistical comparison Etch-rate uniformity map, radial profile golden reference new chamber (within matching band) a chamber outside the matching band gets re-tuned, not shipped to production as-is

Sensor calibration has to happen before the matching comparison means anything. Optical emission spectroscopy endpoint sensors, RF voltage-current probes, and thermocouples all drift slightly from their last calibration, and recipe parameters are tuned against what those sensors *report*, not the true physical state of the chamber — so a miscalibrated pyrometer reading 5°C low would get "corrected for" by a recipe change that actually just makes the real wafer temperature wrong. This is why calibration standards (certified reference gases, NIST-traceable thermocouples, calibrated RF power meters) get run through a chamber's sensors on a fixed schedule, independent of whether the chamber "seems fine."

Qualification wafers are the final gate, and they test the whole stack at once — mechanical, sensor, and recipe. A short run of monitor wafers goes through the calibrated chamber on the actual production recipe, and the resulting film thickness, etch depth, or deposition uniformity gets measured and compared against both the golden chamber's historical data and the fab's statistical process control limits. Only after those qual wafers pass does the chamber's status flip from "calibration" to "released," and even then it stays subject to periodic re-qualification as part of ongoing preventive maintenance.

Chamber mount-and-dock answers "is it physically sound," while chamber calibration-and-setup answers the harder question — "does it behave exactly like every other chamber running the same process" — and that second answer is what actually lets a fab treat a dozen identical-looking chambers as one fungible pool of capacity instead of a dozen slightly different machines.

Take chamber calibration further

Ask the copilot about this term, or have our engineers assess it against your process.