Monitoring autoclave cycles: the sensors and loggers that show what a run really did

What the printout shows

Many autoclaves end a cycle with a printout or a screen of numbers: the program, a cycle number, and time, temperature and pressure through the run. CDC's sterilization guidance counts reading that record, along with the pressure gauge, as the mechanical part of monitoring a steam sterilizer. WHO's reprocessing manual has the operator review the printout before unloading, confirm the right cycle ran with the right parameters, check that the cycle number matches the load's label, and initial it.

That record is worth keeping, but it is the machine describing itself. Unless the unit times its hold from a probe in the load, the temperature on the printout comes from a sensor in the chamber, and it says nothing about the inside of a 2-liter bottle of medium. It also cannot catch its own drift. If the machine's sensor reads two degrees high, so does the printout.

Nor does it say what was in the load, where the big bottles sat or which media batches rode along. Those are lab records, and they are what turn a cycle number into something you can trace back from a contaminated jar. For units with no printer, the WHO manual gives a simple log: date, autoclave number, load number, cycle start, start and end of the sterilization stage, cycle end and a signature.

Chamber temperature is not load temperature

Liquids lag. A biosafety guide published by University College Cork shows thermocouple readings from a 2-liter flask of broth. The chamber reached 121 °C about three minutes into the cycle, the lower part of the liquid took almost 20 minutes, and the top of the flask heated faster than the bottom. In the guide's own example, most of the liquid never reached 121 °C during a 15-minute cycle.

There are three common ways to see the load itself. A load probe is a flexible sensor on a lead, fitted to some autoclaves so the hold starts only when the load reaches temperature. A reference bottle is a spare container of the same size and fill as the largest in the load, carrying that probe or a logger's probe, set where heating is slowest. Thermocouples, fine wire pairs whose junction gives a temperature-dependent voltage, are the long-standing validation tool, led into the chamber through a sealed port and placed in bottles across the load.

Each has its catches. NHS Scotland's guidance on sterilizer test equipment calls for thin, single-strand thermocouple wire, because thicker cable can let steam track along its outside and spoil the reading, and for junctions to be inspected for corrosion and remade as needed. It also notes that platinum resistance sensors drift less in the short term than thermocouples. Whichever you use, put the tip low in the largest bottle.

A logger reading is evidence about one spot in one load. Put it where the medium heats slowest, or it is evidence about the wrong spot.

Loggers that ride in the load

A standalone logger removes the wiring. The same NHS Scotland guidance describes small, self-contained temperature and pressure loggers, with one channel or several, that run on their own power, sample at a programmed rate for a programmed time, and either transmit wirelessly or are downloaded afterwards. It notes they are most useful where feeding sensor leads into the equipment is impractical.

The class is defined by what it has to survive. The housing has to take full cycle temperature and chamber pressure, including any vacuum stage, with seals and a battery to match. The guidance also warns that a logger must measure the load, not just its own body, which in practice means a probe on a lead that sits inside a bottle while the logger rests beside it. Where two or more loggers run together, their clocks should be synchronized so every trace lines up on one chart.

Its numbers, written for hospital sterilizer testing rather than plant labs, make a useful yardstick: a sampling interval of no more than one second, enough readings to have at least five inside the hold, temperature recorded in steps of 0.1 °C or finer, and an uncertainty for the whole system, sensor included, of no more than ±0.5 °C. A logger that samples once a minute can miss the moment a load dipped below temperature.

For buyers

Before you buy a logger

Check the temperature rating against your hottest cycle with some margin, and the housing's pressure rating against the chamber's working pressure, including any vacuum stage. Ask for the accuracy of the whole system at 121 °C, sensor included, and the shortest sampling interval it supports. Confirm the probe can sit inside a bottle on a lead rather than measuring only the logger's body.

Ask how often it needs calibration, who can calibrate it, and whether the report shows readings before and after adjustment. Ask how long the battery and seals last at cycle temperature. Then ask how data comes off: a dock or cable, wireless transfer, a file format you can open without the maker's software, or an API.

Pressure as a second witness

Pressure is the other half of mechanical monitoring in CDC's guidance, and a useful cross-check. In saturated steam, temperature and pressure rise and fall together, so a run that records both lets you test one against the other. If the load reads cooler than the chamber pressure implies, the usual suspect is air rather than steam in that part of the load. On a unit with a vacuum stage, the pressure trace also shows whether each pulse still pulls as deep as it did when the cycle was set up.

Two practical points. An autoclave's dial typically reads gauge pressure, measured above atmosphere, while many loggers report absolute pressure, so the two differ by about one atmosphere, roughly 101 kPa at sea level. Check which a report shows before comparing. Pressure sensors also have their own errors: NHS Scotland's guidance notes that temperature swings inside a pressure-sensing element can throw off its reading, and it expects test sensors to agree with a reference within 50 mbar.

Indicators answer a different question

A logger reports conditions at one spot. Indicators report what those conditions did. CDC's guidance calls for a combination of mechanical, chemical and biological monitoring. Chemical indicators show that an item was exposed to the process. The WHO manual lists six types under ISO 11140, from process indicators that only tell processed from unprocessed to integrating and emulating indicators that react to all the critical variables. CDC adds that chemical indicators belong alongside biological indicators, not in place of them, because only resistant spores measure the killing power of a cycle.

The two kinds of evidence work best together. Use logger runs to find where the load heats slowest, then put the spore test there. For liquids, the Cork guide describes a biological indicator made for liquid loads, sealed so the medium never touches the spores and placed low in the bottle, not floating at the surface where it would heat fastest.

CDC's schedule is written for healthcare: spore tests at least weekly, daily when a sterilizer runs several loads a day, and retesting after installation, relocation, major repair or a failed cycle. A plant lab sets its own schedule in its SOP. CDC also notes that in Europe, items are often released on the physical record alone, a practice called parametric release, which requires a defined quality system and a process validated for those items.

Calibrate what does the checking

A reading is only as trustworthy as the instrument behind it. NHS Scotland's guidance puts test equipment on a planned calibration program at least once a year, with a shorter interval if the instrument drifts, using a reference standard calibrated within the previous 12 months by a laboratory accredited to ISO/IEC 17025. It asks the calibration lab to report readings before and after adjustment, so drift shows up over time.

Each instrument should carry a unique identifier and its calibration date, and a full maintenance and calibration history. Thermocouple systems are checked against an independent reference before and after each series of tests, and should read within 0.5 °C of a source known to ±0.1 °C. Self-contained loggers can skip that step only if their calibration data shows negligible drift between annual calibrations.

There are two sets of instruments in play: the autoclave's own sensors and gauge, and the logger. Each needs its own schedule and history. When either is found out of tolerance, every run it vouched for since its last good check is in question, which is why the reading before adjustment on a calibration report matters as much as the reading after.

From logger to load record

Data has to leave the logger and land somewhere it can be found again. NHS Scotland's guidance asks for recording equipment with memory that lets data be retrieved later and stored securely, and for reporting software that is validated, secure and backed up off-site.

For each load, record the date, autoclave, program and cycle number, operator, each media batch with its container size and fill volume, which logger sat where, when the load reached temperature and when the hold ended, the lowest load reading during the hold, the printout or file reference, indicator results, and who released the load. For each logger, keep its ID and serial number, probe type, calibration certificates with readings before and after adjustment, next due date, clock checks, seal and battery service, and the loads it monitored.

In xPlant Pro those records sit in two places. The autoclave can have its own entry in the equipment library. On plans that include equipment calibration, its calibration sheet holds the verification schedule and a history of calibration and maintenance events, each with an outcome such as pass, pass after adjustment, out of tolerance or fail. The history is append-only, so a correction is a new event. Uses are recorded against the autoclave too: the app records them from SOP runs and transfers, and xPlant’s API documentation shows equipment, or a small script beside it, recording a use for a media batch, with the cycle number in the use’s label and a key that makes a call fired twice record once. When a calibration comes back out of tolerance or failed, the sheet lists the uses recorded since the last pass, so the batches that need a second look are in front of you.

Logger readings arrive separately. On plans that include sensor monitoring, a device registered through the API can post temperature and other readings with a unit, a timestamp and an optional room, and the Sensors page shows each device with its status, its room and its latest value per channel, updated on the hour. Readings belong to the device and, optionally, a room, not to the autoclave’s record, and the Sensors page is a current view rather than a cycle trace, so keep the cycle number in the load record and the logger’s own exported file with the load. The plans page lists which tiers include equipment tracking, calibration, sensor monitoring and API access.