Every cleanroom manager has lived through the same scene: a particle counter trips above its operational limit, the alarm hits the shift supervisor’s phone, and the next two hours are spent answering three questions in order — is this a real excursion, where is it coming from, and what do we do about it before the next batch ships. The difference between a one-hour recovery and a two-day investigation is almost always the play you wrote down before the event happened.
This article lays out a practical excursion response playbook for ISO 5 to ISO 7 cleanrooms in semiconductor and pharmaceutical operations. It assumes you have a continuous airborne particle counter on the wall — such as the GCC-MST-5100XPro — and a documented trend, not just a snapshot. Where the playbook touches on environmental monitoring broadly, we will cross-reference our ISO 14644-1 classification guide and the particle counter selection guide so the framework stays grounded in the underlying standards.
What Counts as an Excursion (and What Does Not)
The first failure mode in any excursion response is misclassifying a normal event as an excursion — or worse, treating a real excursion as noise. The definition has to be written down, signed off, and reviewed at least annually.
For an ISO 5 in-operation class, the per-cubic-meter limits under ISO 14644-1 are 3,520 particles at ≥ 0.5 µm and 29 particles at ≥ 5.0 µm. A single reading at or below those values is not an excursion. An excursion is one of the following:
- Sustained exceedance. Two or more consecutive samples above the operational limit during a defined monitoring window (typically 5 to 15 minutes depending on your sampling flow rate and the room’s clean-air design).
- Trending drift. A 30-day rolling average that climbs by more than a defined percentage (commonly 25 to 50%) without a corresponding process change.
- Alert-level breach. A single sample above the “alert” level (typically set at 50 to 70% of the action level) that persists across two counter locations in the same zone.
- Operator report. A qualified operator reports a visible event — condensation, smoke, mist, or a particle plume — even if the counter has not yet tripped.
Defining these four triggers up front lets you avoid the “is this real?” debate in the moment. If the playbook says “two consecutive samples over 3,520 particles/m³ at 0.5 µm = excursion,” then two consecutive samples over the limit is an excursion, full stop. The argument moves from interpretation to action.
The First 15 Minutes: Stabilize, Do Not Investigate
Most cleanroom managers try to find the root cause before they stabilize the room. This is the wrong order. The first 15 minutes are about preventing the excursion from spreading, not about diagnosing it.
Your first-15-minutes checklist should look like this:
- Stop the inflow. If the suspected source is a personnel transfer, close the air shower cycle and lock the entry door. Most excursions linked to people clear in 8 to 12 minutes if the air shower and cleanroom pressurization are functioning correctly (see our air shower design guide for the engineering background).
- Stabilize the pressure cascade. Confirm the room is at positive pressure relative to the corridor. A drop of 1 to 2 Pa is enough to allow infiltration of unfiltered air.
- Log the alarm. Capture the counter ID, the time stamp, the channel that tripped, the value, and the sample volume. The counter itself should store this in a 21 CFR Part 11 compliant log. If your counter does not log automatically, write it on the excursion form before doing anything else.
- Notify the next person in the chain. Production supervisor, QA on call, facility engineer — whoever your playbook names. Do not wait until you have a root cause hypothesis.
- Do not open any door you do not have to. A door opened mid-excursion can both extend the event and contaminate your investigation.
After 15 minutes, you should have a stable room (or you should have escalated to the facility team for HVAC intervention) and a written log entry. Now you can start the investigation.
The Root Cause Hypothesis Tree
Excursions in well-run cleanrooms fall into one of six categories. Going through them in order keeps you from chasing the wrong cause first.
- Personnel transfer. Look at the access log for the room in the 30 minutes before the alarm. Was anyone entering or exiting? Did the air shower complete a full cycle? Was the door held open during the transfer? The vast majority of ISO 5 excursions in fab and aseptic operations trace back to a personnel event.
- Process event. Was a tool started, stopped, or serviced? Was a chemical dispensed? Was a wafer or container moved through the room? In semiconductor fabs, a process tool exhaust or a vacuum pump can spike the local counter without a corresponding room-wide event.
- Facility event. Has the AHU, FFU, or exhaust fan changed state? Was a HEPA filter scan recently performed? Is the cleanroom pressure cascade still within spec? A facility-side event often shows up across multiple counters, which is the easiest tell.
- Material event. A new lot of wipes, gloves, or garments can outgas and spike the AMC (airborne molecular contamination) reading — and on some counters, the particle reading. Cross-reference with the materials receiving log and our material outgassing guide.
- Counter event. A counter that is itself drifting, dirty, or losing sample flow will produce a false excursion. Run a zero-count verification on a filtered exhaust — most cleanroom managers discover their “excursions” are actually sensor drift during the third or fourth false alarm in a month.
- Unknown. If you cannot categorize the event in 60 minutes, document it as such and escalate to the next review level. Do not let a string of “unknown” excursions pile up — that is the early signal of a systemic problem.
Containment vs. Recovery: Knowing the Difference
Containment stops the excursion from contaminating product. Recovery returns the room to its classified state. The two are not the same activity, and your playbook should treat them as separate decision points.
For a semiconductor fab, containment usually means:
- Quarantining the lot or wafer batch in the affected tool or zone
- Not releasing product to the next process step until the room is back in spec and the exposure window is documented
- Notifying the customer if a dispatch has already been made and the exposure window includes the dispatch time
For a sterile pharmaceutical operation, containment is heavier because of the patient safety line. The Annex 1 contamination control strategy requires that any Grade A event — even a single counter exceedance during an aseptic operation — is investigated and that the batch disposition considers the event. A documented trend of low-level excursions across many batches can become a PAI inspection finding; an isolated event is typically handled inside the batch record.
Recovery has its own checklist, and most of it is mechanical:
- Confirm pressure cascade is in spec
- Confirm FFU / AHU fans at design speed
- Run a recovery test (the 100:1 rule per ISO 14644-3 — covered in our qualification guide)
- Re-zero all counters in the affected zone and confirm they read the local background
- Resume operations only when the room is at “in-operation” classification for a continuous 15-minute window
Documentation and Trend Analysis
The single most valuable thing you do during an excursion is write it down. Most cleanroom managers document the trigger, the time, and the recovery. The strongest programs document everything else:
- Counter ID, channel, value, sample volume, time stamp
- Personnel on shift (names, not just headcount)
- Tools running and their state at the time of the event
- Door events (air shower cycles, pass box openings, door held open events)
- Recovery time to return to in-operation classification
- Root cause hypothesis and the evidence that supports or refutes it
- Corrective action taken (and the date it is closed)
Once a quarter, plot the excursion data as a Pareto. In most fabs and aseptic operations, three to four causes will explain 80% of events. Fix those, and the trend drops sharply. If your top cause is “unknown,” the fix is to upgrade the documentation, not the cleanroom.
Instrumentation: Why Excursion Response Depends on Your Counter
The playbook above assumes a counter that:
- Logs every sample to a 21 CFR Part 11 compliant data store
- Alarms at both alert and action levels on every channel
- Reports a sample volume failure if the inlet is blocked or the pump is weak
- Integrates with the access control and BMS systems so a personnel event and a particle event can be time-correlated
The GCC-MST-5100XPro covers the first three by design (six-channel resolution at 100 L/min, integrated data logging, and a low-flow alarm). The fourth is a building integration decision — your counter has to be able to send an event to the BMS in a format the BMS understands. Plan that integration during the spec phase; it is much harder to retrofit after the cleanroom is running.
Training the Shift Team
A playbook nobody has read does not work. The minimum training pattern is:
- All operators complete the playbook training during onboarding
- Shift supervisors complete a tabletop exercise annually using a real anonymized excursion from the past 12 months
- The QA team reviews the playbook annually and updates it for any new tools, materials, or counter models
- One drill per year — the supervisor walks a new operator through a fake excursion and times the first-15-minutes checklist
Closing: Treat the Excursion as a Data Source
The cleanest mental shift a cleanroom team can make is to treat excursions as a stream of useful data, not as failures to be hidden. The first hundred excursions you document will tell you where the room is genuinely weak — a door that is held open too long, a material that outgasses more than the spec, a fan that drops speed under load, a counter that drifts. Most of those are cheap fixes once you can see them. The cost is the discipline of writing the playbook, training the team, and reviewing the data.
If you are evaluating your current excursion response — or building one from scratch — reach out to us with a sketch of your current cleanroom layout, your counter models, and your typical event frequency. We will respond with a sample playbook template and a recommended instrumentation set, typically within two business days.