August 15, 2026 Steven Eaton

Gas Sensor Poisoning: What Kills an LEL Sensor and How to Catch It

What poisons a gas detector sensor?

Short answer: Poisoning is the permanent loss of sensitivity in a catalytic bead combustible sensor after exposure to certain compounds — silicones above all, along with sulfur compounds, phosphorus compounds and lead compounds. The danger is the direction of failure: a poisoned LEL sensor reads low. It does not blank, it does not fault, and it does not warn you. It reports a reassuring number in an atmosphere that is not reassuring.

This is the failure mode that a daily function check exists to catch, and the reason "the instrument was on and reading zero" is not by itself evidence that an atmosphere was safe. This guide covers what poisons a sensor, the difference between poisoning and inhibition, why the failure is silent, and how the problem is detected and prevented. Where a decision depends on sensor technology rather than on sensor health, the trade-offs between catalytic and infrared detection are treated separately.

It is written for confined-space crews, gas-detection program owners and anyone maintaining a fleet of instruments. If your immediate question is how to interpret the numbers rather than whether to trust them, start with how to read a 4-gas monitor.

Why this matters.
Almost every other instrument fault announces itself. A dead battery, a failed pump, a faulted cell — all of them stop the instrument or throw an error. Sensor poisoning does neither. The instrument powers up, passes a visual check, displays a clean baseline and holds it, and under-reports a flammable atmosphere. OSHA's safety and health information bulletin on calibrating and testing direct-reading portable gas monitors is explicit that a bump test verifies sensor response and alarm function — and it is the only routine check standing between a poisoned sensor and an entry decision made on its readings.

How a catalytic bead sensor works, and how it dies

A catalytic bead sensor — often called a pellistor — contains a small ceramic bead coated with a catalyst and heated by an internal coil. When combustible gas reaches the bead it oxidises on the catalytic surface. That reaction releases heat, the bead's temperature rises, its electrical resistance changes, and the instrument converts that change into a %LEL reading.

Everything depends on the catalytic surface staying active. Poisoning is what happens when a compound reaches that surface and permanently degrades it — typically by depositing a non-reactive layer over the catalyst or by chemically altering it. The heater still works. The circuit still reads. The bead simply no longer converts gas into heat as efficiently as it did, so the same atmosphere produces a smaller temperature rise and therefore a lower reading.

Why oxygen matters to this sensor too

Because the mechanism is combustion, the reaction needs oxygen. In an oxygen-deficient atmosphere a catalytic sensor can under-report for reasons that have nothing to do with poisoning — which is one of the reasons oxygen is read before the combustible channel during a confined-space test. Two different mechanisms produce the same misleadingly low number, and neither is visible on the display.

The poisons, and the inhibitors

The distinction between a poison and an inhibitor is the distinction between permanent and temporary damage, and it decides whether a sensor can be recovered.

Compound class	Effect	Where crews meet it
Silicones and siloxanes	Poison — permanent. The most notorious class; widely reported to degrade catalytic beads at very low concentrations.	Sealants and caulks, mould release agents, some lubricants and greases, polishes, certain personal-care products.
Sulfur compounds	Poison — generally treated as permanent at meaningful exposures.	Sour gas, wastewater and digester atmospheres, refining, some fuels.
Phosphorus compounds	Poison — permanent.	Some fire retardants, agricultural chemicals, certain industrial additives.
Lead compounds	Poison — permanent.	Legacy leaded fuels and some specialist industrial processes.
Halogenated hydrocarbons	Inhibitor — often temporary. Suppresses response while present; sensitivity may recover in clean air, but recovery must be verified rather than assumed.	Refrigerants, chlorinated solvents, some degreasers and cleaning agents.

Sensitivity to each class, and the concentrations at which damage occurs, vary by bead formulation and manufacturer. Poison-resistant catalytic beads exist and are specified for exactly these environments. Your instrument's documentation is the authority on which class its sensor tolerates and to what degree — a general table like the one above tells you what to be careful about, not what your specific sensor will survive.

Why the failure is silent, and what that implies

A poisoned sensor produces a reading that looks exactly like a clean atmosphere: a stable zero. There is no error code and no visual indication. The bead is still heated, the electronics still function, and the instrument has no way to distinguish "no gas present" from "gas present but not being converted."

Two consequences follow. The first is that the only routine way to detect poisoning is to present the sensor with a known concentration of gas and confirm it responds — a bump test. The second is that partial poisoning is the harder case: a sensor might still respond, but respond weakly, so it passes a casual visual check while reading substantially low. That is why a bump test is judged on whether the reading reaches the expected value for the test gas, not merely on whether the number moved.

The full procedure and the distinction between a bump test, a calibration check and a full calibration is covered in bump test vs calibration for gas detectors. Cylinders, regulators and tubing are on the gas detector calibration gas and accessories shelf.

The failure chain, and where it gets caught

Poisoning rarely appears as an isolated event. It usually surfaces as a link in a chain, and recognising the chain is what turns a confusing morning into a diagnosis.

Exposure. The instrument is used where a poisoning compound is present — often without anyone identifying it as a hazard to the sensor, because it is not a hazard to people at that concentration.
Silent degradation. Response falls. Readings still look normal because the atmosphere is genuinely clean most of the time.
A weak or failed bump test. The first observable symptom. The channel reads low against the test gas, or does not reach the expected value.
A failed or drifting calibration. The instrument cannot be adjusted back to a correct span, or it can but does not hold.
Sensor replacement. A poisoned bead is not restored by calibration. Calibration adjusts the instrument's interpretation of the signal; it cannot rebuild a degraded catalytic surface.

That fourth step is worth dwelling on, because it is where a real mistake gets made. An instrument that will not hold calibration can often be forced through the procedure by repeating it. Doing so on a poisoned sensor produces an instrument that passes calibration and still under-reports in the field, because the adjustment has been made against a response that is no longer linear or stable.

Prevention

Poisoning cannot be undone, so the whole strategy is avoidance and early detection.

Know your atmosphere. Identify poisoning compounds in the hazard assessment alongside the toxic and flammable hazards. Silicone sealant work in a confined space is a sensor hazard even when it is not a health hazard.
Specify poison-resistant sensors where a known poison is routinely present, and accept the trade-offs the manufacturer states for them.
Bump test before each day's use. The single most effective control, and the one that converts a silent failure into a visible one.
Watch the trend, not just the pass. A channel that needs progressively more adjustment at each calibration is degrading, whether or not it has failed yet. A fleet record makes this visible; a pass/fail sticker does not.
Segregate instruments by environment where practical, so an instrument used around known poisons is not the one relied on elsewhere.
Consider a detection principle that cannot be poisoned where the environment justifies it. Infrared sensing does not rely on a catalytic surface and is therefore immune to this failure mode, though it carries its own limitations — notably that it cannot detect hydrogen.

Worked example: a silicone exposure in a valve chamber

A crew spends a shift applying silicone sealant inside a below-grade chamber, carrying a portable instrument such as the Forensics 4 Gas Meter throughout. The atmosphere is clean, nothing alarms, and the instrument goes back on the shelf. Here is how that becomes a finding rather than an incident:

Flag the exposure at the time, not later. The moment silicone is used around a catalytic sensor, the instrument is a suspect. Note it against that unit's record while the crew still remembers.
Bump test before the next use and read the value, not just the response. Apply the test gas and confirm the combustible channel reaches the expected reading for that concentration. A channel that moves but falls short is a partially poisoned channel.
Compare against the unit's own history. A response that is weaker than this instrument's previous checks is meaningful even if it is still inside tolerance. The trend is the early warning.
Attempt a calibration and watch whether it holds. A sensor that needs unusual adjustment, or that drifts again shortly after, is telling you the bead is degraded rather than merely out of adjustment.
Replace the sensor rather than re-calibrating around it. Calibration cannot restore a catalytic surface. An instrument forced through calibration on a poisoned bead will pass on the bench and under-report in the space.
Close the loop on the cause. Record the exposure so the next silicone job is planned with a different instrument, a poison-resistant sensor, or an infrared combustible channel. Otherwise the same sensor is bought twice.

The same reasoning applies across the portable gas detectors and personal gas detectors shelves, and to permanently mounted units from the fixed gas detection systems shelf — a fixed catalytic sensor in a plant room can be poisoned by maintenance work nobody connected to the gas detection system. Combustible instruments generally are on the combustible gas detectors shelf, and the wider category overview is in the gas detection complete guide.

What poisoning means for how you buy instruments

Sensor poisoning is one of the few failure modes that should influence procurement rather than just maintenance, because the cheapest instrument to buy is not always the cheapest to keep running in a hostile atmosphere.

Sealed maintenance-free units trade serviceability for simplicity: there is no sensor to replace, so a poisoning event effectively ends the instrument's useful life early. In a clean environment that is a good trade; around known poisons it is an expensive one. The wearable options are compared in the best personal gas detector guide.
Serviceable multi-gas instruments cost more up front and let you replace a single degraded channel rather than the whole unit. Across a fleet exposed to sealants or sour atmospheres, that is usually the lower total cost.
Single-gas clips for the dominant hazard can keep an expensive multi-gas instrument out of the worst atmospheres entirely — see 4-gas monitor vs single-gas detector and the hydrogen sulfide detectors shelf for sour-gas work.
Sample-draw versus diffusion changes what reaches the sensor and how much of it. Remote sampling through a probe means the instrument itself is not sitting in the contaminated atmosphere — the trade-offs are covered in diffusion vs pump gas detector.

Whichever route you take, the recurring cost that decides it is test gas rather than sensors, because the daily check is what makes any of this visible in time. Budget for the consumables on the gas detector calibration gas and accessories shelf as a running cost of the program, not an occasional purchase, and treat the wider selection question through the portable gas detectors and 4-gas monitors shelves.

Frequently asked questions

What poisons an LEL sensor?

Silicones and siloxanes are the most commonly cited class, along with sulfur compounds, phosphorus compounds and lead compounds. Halogenated hydrocarbons act as inhibitors, suppressing response while present, with recovery that must be verified rather than assumed.

Can silicone damage a gas detector?

Yes. Silicone-based sealants, lubricants, mould release agents and polishes are widely reported to degrade catalytic bead sensors, and they do so at low concentrations. The damage is permanent and produces no error indication on the instrument.

Why is a poisoned sensor dangerous rather than just broken?

Because it fails low rather than failing off. The instrument keeps working, keeps displaying a clean baseline and keeps under-reporting a flammable atmosphere. There is no fault code to notice, so the reading looks trustworthy when it is not.

What is the difference between sensor poisoning and inhibition?

Poisoning permanently degrades the catalytic surface, so sensitivity does not return. Inhibition suppresses response only while the interfering substance is present, with sensitivity potentially recovering in clean air — but recovery has to be confirmed by a bump test, not assumed.

Can a poisoned sensor be recalibrated?

No. Calibration adjusts how the instrument interprets the sensor's signal; it cannot rebuild a degraded catalytic surface. Forcing a poisoned sensor through calibration produces an instrument that passes on the bench and still under-reports in the field.

Why won't my combustible sensor respond during a bump test?

Common causes are poisoning, an exhausted sensor at end of life, expired or wrong test gas, inadequate flow, or a blocked inlet. Work through the gas supply and the flow path first, because those are the cheap explanations, then treat the sensor as suspect.

How do I know if my sensor has been poisoned?

Only by presenting it with a known concentration and confirming it reads the expected value. Partial poisoning still produces a response, so the check has to be judged on the value reached, not on whether the number moved at all.

Does poisoning affect electrochemical toxic sensors?

The catalytic poisoning mechanism described here is specific to catalytic bead combustible sensors. Electrochemical cells degrade in their own ways — electrolyte depletion, filter saturation and normal ageing — and those are separate failure modes with separate symptoms.

Are infrared sensors immune to poisoning?

Infrared detection measures absorption of light rather than relying on a catalytic surface, so the poisoning mechanism does not apply to it. That immunity comes with different limitations, including an inability to detect hydrogen, so it is a trade-off rather than a straight upgrade.

What is a poison-resistant sensor?

A catalytic bead formulated to tolerate greater exposure to known poisons before losing sensitivity. It raises the threshold rather than removing the failure mode, and manufacturers state the trade-offs, which can include response characteristics or service life.

How often should I bump test to catch poisoning?

Before each day of use is the widely applied convention and the one this site recommends throughout. Poisoning can occur in a single shift, so a check performed weekly leaves a window in which every reading taken is unverified.

Can a sensor be poisoned without anyone noticing at the time?

Yes, and that is the normal case. The compounds involved are frequently harmless to people at the concentrations that damage a bead, so nothing about the work feels hazardous and nobody records an exposure event.

Does a poisoned sensor still pass a calibration?

Sometimes, which is the trap. It may accept adjustment and then drift again, or accept adjustment while responding non-linearly across the range. An instrument that needs unusual adjustment or will not hold it should be treated as a sensor replacement.

Will an oxygen-deficient atmosphere make a combustible sensor read low?

Yes. A catalytic bead depends on combustion, so it needs oxygen to respond fully. That is a separate mechanism from poisoning but produces the same misleadingly low reading, and it is one reason oxygen is read before the combustible channel.

Should I replace all sensors after a poisoning event?

Only the affected sensor type, and only where response has actually been shown to be degraded. Verify each channel against test gas rather than replacing on suspicion — but do not return a channel to service on the assumption that it was probably fine.