No single sensing technology covers every flammable or toxic gas threat in a refinery, battery cell plant or confined-space entry job, so engineers match the detector to the gas, the environment and the certification envelope.
The four workhorse technologies in 2026 spec sheets are catalytic bead, non-dispersive infrared (NDIR), electrochemical cell and metal-oxide semiconductor; semiconductor and PID sit alongside for lower-cost or VOC applications. Choosing among them means balancing lower explosive limit (LEL) response, ppm-level toxic sensitivity, poisoning resistance, warm-up time and intrinsic-safety power budget against the ATEX 2014/34/EU or IECEx zone they must live in.
Catalytic Bead (Pellistor) Detectors: Cheap, Proven, Poisonable
Catalytic bead sensors oxidise flammable gas on a heated catalyst-coated element and compare the resulting temperature rise against an inert reference bead, producing a Wheatstone-bridge output proportional to gas concentration in the 0–100 % LEL range [S2].
Advantages: lowest unit cost in the combustible-gas detector category, linear response from roughly 0 to 100 % LEL for most hydrocarbons, no need for a separate IR source, and decades of field-proven calibration procedures — a portable gas detector with a pellistor head is still the default kit for utility contractors entering gas-distribution manholes.
Disadvantages: the catalyst is permanently damaged by silicones, leaded petrol vapours, hydrogen sulphide above a few hundred ppm and phosphate esters, causing loss of sensitivity that calibration cannot recover; oxygen is required for the oxidation reaction, so readings in inerted vessels collapse toward zero; and the element must run at 400–500 °C, which draws more current than an IR source and complicates IECEx Ex ia certification for multi-gas detectors running four sensors at once.
Non-Dispersive Infrared (NDIR) Detectors: Selective, Poison-Proof, Range-Limited
NDIR detectors pass broadband IR through a gas-filled measurement cell and measure absorption at a wavelength matched to the target molecule's C–H, C=O or N–H stretch, using a reference wavelength to compensate for lamp ageing, window fouling and optical-path drift. [S1]
Advantages: immune to the silicone and H2S poisoning that kills pellistors, fail-safe in inert atmospheres where the sensor simply reads zero instead of giving a dangerous under-read, and highly selective — a properly designed methane NDIR cell rejects propane interference and vice versa; this selectivity is why NDIR is the workhorse for fixed gas detector heads on LNG terminals and battery dry-room solvent monitoring.
Disadvantages: NDIR cannot detect homonuclear diatomic gases (H2, N2, O2) because they have no IR fingerprint, so a toxic gas detector for hydrogen must be electrochem or palladium-MOS; pressure and humidity swings shift the absorption line, requiring on-board compensation; and the IR source plus detector pair costs 2–4× a pellistor at the BOM level, pushing the four-gas cart up by a noticeable margin.
Electrochemical Cells: ppm-Level Toxic Sensitivity, Limited Life
Electrochemical sensors hold a target gas across a gas-permeable membrane to a working electrode where it is oxidised or reduced, generating a current in the nanoamp-to-microamp range proportional to concentration, typically over 0–10 ppm, 0–50 ppm or 0–500 ppm ranges depending on the electrolyte chemistry. [S2]
Advantages: detection limits down to single-digit ppm for H2S, CO, NO, NO2, SO2, NH3 and Cl2, intrinsic safety is straightforward because cell current is intrinsically limited, and the cell draws microamps — letting a portable gas detector run four toxic sensors plus a pellistor and an O2 cell for a 12-hour shift on one battery charge; cross-reference the cell chemistries against the gas detector selection guide before locking the bill of materials.
Disadvantages: electrolyte dries out at high temperature and freezes below roughly −20 °C without a heated enclosure; many cells suffer from interference gases — a CO cell responds to H2, an H2S cell is fouled by high concentrations of its own target gas; and every cell has a finite 18–36 month operating life regardless of use, so the spares budget must include a scheduled replacement cycle, not just a calibration cycle.
Metal-Oxide Semiconductor (MOS) and Photoionisation Detector (PID): Cheap VOC Coverage, Humidity-Dependent
MOS sensors heat a tin-oxide or tungsten-oxide film to 200–400 °C and watch the resistance drop as reducing VOCs adsorb and donate electrons, while PID detectors ionise VOC molecules with a 10.0 eV or 11.7 eV UV lamp and measure the resulting ion current — both sold as low-cost alternatives for VOC and combustible gas detection at the ppm level. [S3]
Advantages: MOS is the lowest-cost sensor you can buy in volume, responds to a broad VOC envelope, and survives a brief exposure to high gas concentration that would kill an electrochem cell; PID delivers linear ppb-to-ppm response across hundreds of VOCs that no other sensor family can cover, and is the only practical detector for isocyanate and benzene leaks during tank-purging operations where an infrared gas detector cannot see the molecule.
Disadvantages: MOS is non-selective and humidity-dependent — a 50 % change in relative humidity can swing the baseline more than 100 ppm on a methane-calibrated head — and is generally not certified for life-safety combustible-gas duty in most jurisdictions; PID lamps need cleaning every few weeks in dirty environments, the 11.7 eV lamp is moisture-sensitive, and neither technology is a drop-in replacement for a pellistor on the LEL scale.
Decision Map: Matching Sensor to Application
Use pellistor when the threat is a known flammable hydrocarbon in air, the budget is tight, and the environment is free of silicones and H2S; use NDIR when poisoning gases are present, the gas is methane or a heavier hydrocarbon, and inert-atmosphere fail-safe behaviour matters; use electrochem for ppm-level toxic threats in confined spaces where battery life is a constraint; use MOS only for qualitative VOC alarms; use PID for VOC quantification or hazardous-substance leak survey. [S4]
The standard envelope that locks the choice down is ATEX 2014/34/EU for European explosive atmospheres plus IECEx for global projects, and for offshore and sour-service applications the cell housing and wetted parts must be qualified to NACE MR0175; for oxygen-deficient and oxygen-enriched monitoring the electrochem O2 cell remains the dominant technology because neither pellistor nor NDIR is reliable for O2 in a multi-gas detector form factor.
Failure Modes and Lifecycle Costs Engineers Often Miss
Sensor poisoning is the single most expensive failure mode in a fleet — a £20 silicone release from a nearby paint operation can kill a £150 pellistor in minutes, and a single 500 ppm H2S overexposure can shorten a £40 electrochem H2S cell from 24 months to 24 hours; bump-test logs and the use of catalytic bead sensors only with confirmed-clean test gas are the cheapest insurance. [S5]
Cross-sensitivity is the second silent failure — CO cells responding to H2 from a battery vent, an HCN cell responding to NO, an SO2 cell responding to NO2 — turns a perfectly calibrated instrument into a liar; the OEM cross-sensitivity matrix (typically published in ppm of interferent giving 1 ppm of signal) must be on the spec sheet before purchase, not after the incident.
Lifecycle cost is dominated by the consumable sensors, not the instrument body: a typical four-gas monitor in a refinery turnaround contract burns through 12–24 electrochem cells and 2–4 pellistor heads per year, so the total-cost-of-ownership line item that matters in the purchase evaluation is the 5-year cell-replacement cost, not the sticker price of the unit; the new SRE and IR-leak-test wave of 2026 instruments is pushing the cell-replacement interval out, but only for the larger form factors and at a BOM premium the smaller portable gas detector market has not yet absorbed.
Trackable signal to watch next: the IEC 60079-29-1 performance-class boundaries between "catalytic-only", "IR-only" and "dual-technology" certified heads are being re-evaluated as dual-IR-plus-electrochem fixed-point detectors replace single-sensor catalytic arrays on new greenfield ethylene and lithium-refinery builds; the industrial vacuum and process-gas containment auxiliary market is moving in lockstep, and the cell-pricing line on Q3 2026 OEM price lists will be the cleanest data point on whether the dual-tech premium is collapsing.