Do Infrared Gas Sensors Get Poisoned? NDIR vs Pellistors
Short answer: no. An NDIR sensor measures how much infrared light a gas absorbs, so nothing in the gas stream touches a catalyst and nothing is consumed. A catalytic bead sensor, by contrast, burns the gas it measures on a catalyst surface - and that catalyst can be poisoned, inhibited or burned out. The difference is not a marketing nuance: a poisoned pellistor usually keeps looking healthy while it stops detecting gas.
How a catalytic bead sensor actually measures gas
A pellistor contains two matched beads. One is coated with a catalyst, the other with an inert material, and both are heated to several hundred degrees. When a combustible gas reaches the pair, it oxidises on the catalytic bead and releases heat; the inert bead does not react. The temperature difference changes the resistance of the two elements, and that difference is read as %LEL.
Two consequences follow directly from the physics. First, the measurement requires oxygen: combustion cannot happen without it, and most pellistors need roughly 10% vol oxygen or more to read correctly. Second, the catalyst is part of the measuring chain - anything that degrades the catalyst degrades the reading, and the sensor has no way of telling you that it happened.
The chemistry of poisoning: what actually kills the catalyst
Poisoning is not a single mechanism. It is a family of surface reactions that all end the same way - the active sites of the catalyst become unavailable to the target gas.
Silicones are the classic case. Volatile siloxanes from sealants, lubricants, greases, personal care products and some cleaning agents decompose on the hot bead and leave behind silicon dioxide. That oxide layer is thermally stable, mechanically hard and permanent: it does not evaporate when the silicone source is removed. The bead keeps its resistance, keeps drawing power and keeps producing a plausible zero - with a fraction of its former sensitivity.
Sulphur compounds work differently. Hydrogen sulphide and other sulphur species adsorb strongly on noble-metal catalysts and form surface sulphides; at pellistor temperatures part of that damage is irreversible. Lead compounds and halogenated hydrocarbons - refrigerants, chlorinated solvents, some fire suppressants - act as inhibitors: they suppress the reaction while present and often leave the bead permanently desensitised afterwards.
Then there is saturation, which is not poisoning but is often confused with it. Exposure well above the calibrated range drives the bead far past its design temperature and can burn the catalyst out in a single event. That is the failure mode after a real gas release - exactly when the instrument is needed next.
Silicones are the classic case. Volatile siloxanes from sealants, lubricants, greases, personal care products and some cleaning agents decompose on the hot bead and leave behind silicon dioxide. That oxide layer is thermally stable, mechanically hard and permanent: it does not evaporate when the silicone source is removed. The bead keeps its resistance, keeps drawing power and keeps producing a plausible zero - with a fraction of its former sensitivity.
Sulphur compounds work differently. Hydrogen sulphide and other sulphur species adsorb strongly on noble-metal catalysts and form surface sulphides; at pellistor temperatures part of that damage is irreversible. Lead compounds and halogenated hydrocarbons - refrigerants, chlorinated solvents, some fire suppressants - act as inhibitors: they suppress the reaction while present and often leave the bead permanently desensitised afterwards.
Then there is saturation, which is not poisoning but is often confused with it. Exposure well above the calibrated range drives the bead far past its design temperature and can burn the catalyst out in a single event. That is the failure mode after a real gas release - exactly when the instrument is needed next.
Why a poisoned sensor is more dangerous than a dead one
A failed electronic component announces itself. A poisoned pellistor does not. Zero stays at zero, the display looks normal, self-test passes, and the only symptom is a reduced response to gas that nobody sees until a bump test - or until an incident.
This is why poisoning is written into calibration policy rather than into fault handling: the industry answer to invisible degradation is frequent verification. Bump testing before each shift and full calibration up to four times a year is not paranoia; it is the compensation for a sensor that cannot self-diagnose its own loss of sensitivity.
A second, quieter problem sits alongside it. A pellistor calibrated on methane reads other combustible gases through a correction factor. In a mixed atmosphere - hydrogen, propane, solvent vapours - those factors diverge, and the instrument reports a number that is confidently wrong.
This is why poisoning is written into calibration policy rather than into fault handling: the industry answer to invisible degradation is frequent verification. Bump testing before each shift and full calibration up to four times a year is not paranoia; it is the compensation for a sensor that cannot self-diagnose its own loss of sensitivity.
A second, quieter problem sits alongside it. A pellistor calibrated on methane reads other combustible gases through a correction factor. In a mixed atmosphere - hydrogen, propane, solvent vapours - those factors diverge, and the instrument reports a number that is confidently wrong.
What NDIR does instead, and where its own limits are
An NDIR sensor passes infrared light through the gas and measures absorption at a wavelength the target molecule absorbs. Nothing burns, nothing is consumed, no catalyst exists to be poisoned. Silicones, sulphur compounds and halogenated vapours pass through the optical path without changing the measurement; over-range exposure produces a high reading, not permanent damage.
Modern designs go further and remove the drift that optical instruments were once criticised for. A dual-beam, dual-wavelength scheme sends a reference wavelength - one the target gas does not absorb - down the same optical path as the measuring wavelength. Ageing of the emitter, dust on the windows, water vapour and general contamination change both signals almost equally, so they cancel in the ratio instead of appearing as gas. That is the reason an optical sensor can hold calibration for years rather than months.
Being honest about the limits matters more than listing advantages. NDIR is selective by physics, which means it detects only molecules with an infrared absorption band at the chosen wavelength: hydrogen acetylene is visible for NDIR methane sensor are invisible to it. High humidity can produce transient artefacts in poorly compensated designs. And an optical sensor is a more complex component than two beads on a bridge - which historically meant more power, until LED-based emitters changed that arithmetic.
Modern designs go further and remove the drift that optical instruments were once criticised for. A dual-beam, dual-wavelength scheme sends a reference wavelength - one the target gas does not absorb - down the same optical path as the measuring wavelength. Ageing of the emitter, dust on the windows, water vapour and general contamination change both signals almost equally, so they cancel in the ratio instead of appearing as gas. That is the reason an optical sensor can hold calibration for years rather than months.
Being honest about the limits matters more than listing advantages. NDIR is selective by physics, which means it detects only molecules with an infrared absorption band at the chosen wavelength: hydrogen acetylene is visible for NDIR methane sensor are invisible to it. High humidity can produce transient artefacts in poorly compensated designs. And an optical sensor is a more complex component than two beads on a bridge - which historically meant more power, until LED-based emitters changed that arithmetic.
Power: the argument that used to favour pellistors
For decades the trade-off was straightforward. Pellistors were cheap and simple but fragile and thirsty; NDIR was robust but expensive to run, because an infrared source had to be powered. That is no longer the case. LED-based NDIR sensors operate at average currents measured in tens of microamps, which changes what an instrument can be.
Concretely: an average operating current below 35 µA is what makes a two-year, non-serviceable clip detector possible at all. A ppm-range optical sensor drawing under 100 µA supports twelve months of battery life in a portable leak detector. A continuously measuring sensor at under 60 µA turns a wireless fixed point from a maintenance liability into an install-and-forget device.
Field comparisons on the same instrument show the same direction: replacing a catalytic bead LEL sensor with an ultra-low-power infrared sensor has been reported to extend runtime from around 18 hours to about 32 hours on a single charge.
Concretely: an average operating current below 35 µA is what makes a two-year, non-serviceable clip detector possible at all. A ppm-range optical sensor drawing under 100 µA supports twelve months of battery life in a portable leak detector. A continuously measuring sensor at under 60 µA turns a wireless fixed point from a maintenance liability into an install-and-forget device.
Field comparisons on the same instrument show the same direction: replacing a catalytic bead LEL sensor with an ultra-low-power infrared sensor has been reported to extend runtime from around 18 hours to about 32 hours on a single charge.
Total cost of ownership over five years
Purchase price is the smallest term in the equation and the only one that favours pellistors. What follows is the structure of the comparison; put your own labour rate and fleet size into it.
Calibration labour: four calibrations a year against one, multiplied by the number of instruments and by the loaded cost of a technician, plus calibration gas and downtime. Replacement cycle: a poisoned or saturated bead is scrap, and in contaminated environments that can happen inside a year, while an optical element with no consumable chemistry lasts the life of the instrument. Battery and charging infrastructure: fewer charge cycles, smaller cells, and in the extreme case a sealed non-serviceable device with no charging infrastructure at all. Incident exposure: this is the term nobody wants to price, and it is the one that dominates. A silently poisoned sensor is a safety failure with legal and human consequences, not a maintenance line item.
The honest summary: pellistors remain the right answer for hydrogen in clean, oxygen-rich atmospheres, and they are cheap to buy. Everywhere else - silicone-bearing environments, sulphur-bearing gas streams, inert or purged spaces, battery-powered instruments, long service intervals - optical detection wins on both safety and five-year cost.
Calibration labour: four calibrations a year against one, multiplied by the number of instruments and by the loaded cost of a technician, plus calibration gas and downtime. Replacement cycle: a poisoned or saturated bead is scrap, and in contaminated environments that can happen inside a year, while an optical element with no consumable chemistry lasts the life of the instrument. Battery and charging infrastructure: fewer charge cycles, smaller cells, and in the extreme case a sealed non-serviceable device with no charging infrastructure at all. Incident exposure: this is the term nobody wants to price, and it is the one that dominates. A silently poisoned sensor is a safety failure with legal and human consequences, not a maintenance line item.
The honest summary: pellistors remain the right answer for hydrogen in clean, oxygen-rich atmospheres, and they are cheap to buy. Everywhere else - silicone-bearing environments, sulphur-bearing gas streams, inert or purged spaces, battery-powered instruments, long service intervals - optical detection wins on both safety and five-year cost.
Choosing between them: a short decision list
Is hydrogen a target gas? If yes, a pellistor or a hybrid solution is required - NDIR cannot see either.
Is oxygen guaranteed above roughly 10% vol at the sensor? If not, catalytic detection is unreliable by physics, and NDIR is the only safe choice.
Are silicones, sulphur compounds or halogenated vapours present, even occasionally? If yes, treat pellistors as consumables and budget for their replacement, or move to optical.
Does the instrument run on a battery, and for how long? Below a few days of required autonomy the question is open; above that, ultra-low-power optical sensing is what makes the specification achievable.
What is the acceptable calibration interval? If the answer is "annually, and I need to trust it", the sensor has to be one that cannot lose sensitivity invisibly.
Is oxygen guaranteed above roughly 10% vol at the sensor? If not, catalytic detection is unreliable by physics, and NDIR is the only safe choice.
Are silicones, sulphur compounds or halogenated vapours present, even occasionally? If yes, treat pellistors as consumables and budget for their replacement, or move to optical.
Does the instrument run on a battery, and for how long? Below a few days of required autonomy the question is open; above that, ultra-low-power optical sensing is what makes the specification achievable.
What is the acceptable calibration interval? If the answer is "annually, and I need to trust it", the sensor has to be one that cannot lose sensitivity invisibly.