IR gas detection is now the dominant fixed-instrument technology for hydrocarbon and CO2 monitoring on offshore platforms, LNG terminals, and semiconductor cleanrooms, with open-path units routinely specified to 100+ m beam length and point units operating from -40 °C to +65 °C [S1][S5].
The 2026 fixed-gas-detection market indexes 22 manufacturers offering roughly 165 IR-detector models, with target-gas coverage concentrated on CH4 (11 vendors), CO2 (11), H2 (9), C2H6 and propane (4), and refrigerant gases (1) [S3]. CO2-specific catalogues list 34 manufacturers and 66 models, of which 24 use IR, 15 electrochemical, 3 catalytic bead, 3 laser/TDL, and 2 photoacoustic [S4].
Two Physical Architectures: Point vs Open-Path IR
Point IR detectors (sometimes called NDIR or "extractive" when sampling is pumped) co-locate emitter, dual reference/measure detector, and optical bench inside a single Ex d or Ex e housing, and are specified for cell volumes typically 50-250 mm long with mirror bounces folding the optical path to 1-3 m effective length [S6]. The MSA Safety IR5500 is the open-path counter-example: a through-beam design monitoring both LEL-m and ppm-m over line-of-sight distances well beyond what a point cell can resolve [S1]. Open-path instruments answer the question "is there a flammable cloud drifting between these two posts?" rather than "what is the concentration at one point" — a categorical difference in plant siting philosophy.
Within point instruments, the dominant non-dispersive IR (NDIR) configuration pairs a single broadband incandescent source with two bandpass filters at the detector — one at the gas absorption wavelength (measure) and one at a non-absorbing reference — then ratio-tracks the two channels to cancel source ageing, window contamination, and ambient drift [S6]. A dual-wavelength variant alternates the two beams through the same physical path, which improves common-mode rejection further. TDL (tunable diode laser) instruments add a narrow-line semiconductor laser tuned to a single rotational-vibrational line of the target molecule, giving sub-ppm selectivity and zero cross-sensitivity to other hydrocarbons — useful for ppm-methane leak detection in coal-mine drainage [S3].
Detection Principle: Beer-Lambert Absorption on Hydrocarbon Bands
All IR gas detection is, at heart, a measurement of how much mid-infrared radiation at a specific wavelength (roughly 3.0-3.5 µm for C-H stretches of methane, 4.26 µm for CO2, 4.6 µm for CO) survives passage through the gas cell [S6]. Each target gas has a characteristic absorption fingerprint; the detector subtracts the non-absorbing reference channel and converts the ratio into a concentration via a 3rd- or 4th-order polynomial calibration. Because the absorption is selective to the molecular bond, IR detectors do not respond to every combustible gas the way a catalytic-bead or pellistor sensor does — which is both a strength (no poisoning by H2S or silicones) and a limitation (you must know which gas you are looking for) [S2].
The basic optical train in every IR gas sensor contains four canonical elements: an IR source (either a tungsten-filament incandescent or a MEMS-style micro-emitter, operating at 600-1100 K), a measurement chamber with gas-diffusion or pumped inlet, a narrow-band optical filter selecting the absorption band, and a thermopile or pyroelectric / photodiode receiver [S6].
Target-Gas Coverage Across 2026 Catalogs
Counting entries in the DirectIndustry IR-detector manufacturer index, the gas-types breakdown is: CH4 and methane-class flammable (11 vendors), CO2 (11), flammable gas generally (9), H2 (9), O2 (9), toxic gas class (9), VOC (9), CO (8), N2 (7), benzene and Cl2 (6 each), formaldehyde (6), SO2 and sulfur-bearing (6), ammonia (5), ozone (5), propane (4), LPG (3), butane (2), ethanol (1), hexane (1), pentane (1), and refrigerant gases (1) [S3]. H2 detection is a notable outlier — IR absorption by H2 is intrinsically weak, so 9 IR-vendor entries for hydrogen mostly means IR-confirmation roles around an electrochemical H2 cell rather than pure IR H2 measurement.
For CO2 specifically, the parallel catalog of 34 manufacturers and 66 models splits: 24 IR-based, 15 electrochemical, 6 "other", 3 catalytic, 3 laser/TDL, 3 semiconductor, 2 photoacoustic, 2 thermal, 1 LED-based, 1 TDL, 1 digital, 1 laser-diode, 1 photodiode, 1 pellistor, 17 not-specified [S4]. That puts IR at roughly 36 % of the CO2-specific product count and laser/TDL at about 5 % — NDIR remains the default CO2 transducer, with TDL reserved for purity, medical, and low-ppm specialty applications.
Selection Criteria That Actually Drive the Spec
Five parameters drive 90 % of IR-gas-detector selection. (1) Target gas and its detection range — CH4 is calibrated in %LEL for safety and in ppm-m for environmental compliance; CO2 is normally 0-5 %VOL or 0-10,000 ppm for IAQ, but 0-100 %VOL for process and CO2 fire suppression supervision. (2) Response time T90, normally 5-30 s for point NDIR and under 3 s for TDL. (3) Path length for open-path (5-200 m typical) and the resulting ppm-m sensitivity floor — at 100 m the MSA IR5500-class instrument reaches ppm-m low-end resolution. (4) Hazardous-area certification: ATEX II 2G Ex d IIC T6, IECEx equivalents, and SIL 2 hardware (Honeywell Searchpoint Optima Plus holds SIL 2 / SIL 3 with optional HART) [S5]. (5) Output protocol — 4-20 mA sink/source, HART 7 over the same loop, Foundation Fieldbus, PROFIBUS PA, or digital-only Modbus / Ethernet-APL for greenfield builds [S5].
Environmental limits matter for outdoor and process applications. The Honeywell Searchpoint Optima Plus is rated for -40 °C to +65 °C ambient with stainless-steel Ex d housing and operates in 0-95 %RH non-condensing [S5]. Wet dirty sites, paint-booth environments, and offshore salt-spray zones all push the specifier toward open-path with heated optics, or to a remote-gland-mounted point detector with a sintered flame arrestor and weather shield.
Where IR Wins, Where It Loses: Comparison Against Catalytic, Electrochemical, and TDL
Against the three alternatives, IR's profile is: failsafe in oxygen-deficient atmospheres (catalytic bead needs O2 to burn the gas — IR does not), immune to H2S and silicone poisoning that kills pellistors, no consumable electrolyte to replace, and a 5-15 year calibration interval [S5][S6]. Against electrochemical cells, IR wins on lifetime and against the 9-vendor H2 / NH3 / O2 cross-sensitivity list, but loses on ppm-level cost — an IR NH3 or HCl instrument is materially more expensive than its electrochemical equivalent.
For ppm-methane leak survey, fixed TDL hits detection floors around 0.1 ppm-m, well below anything NDIR can resolve; the trade-off is single-gas lock-in and 5-10× price. For the midstream LNG and wellhead-leak-detection space, open-path IR is the workhorse because ppm-m sensitivity at 30-100 m path detects a 0.5 m gas cloud that a point detector simply cannot see [S1]. A useful contrast in the broader fixed-gas-detector landscape is that catalytic-bead and electrochemical sensors are point-and-leader-cable architectures, whereas IR pivots between point and open-path. The same logic drives multi-gas detector and combustible-gas detector families, where IR modules typically replace catalytic beads for CH4/LEL and replace electrochemical cells for CO2/CO in the higher-end cart configurations.
Application Map by Industry
Three application clusters dominate 2026 IR-detector sales. Offshore and onshore oil & gas: open-path IR for perimeter and turbine-enclosure monitoring, point IR for LEL-m and ppm-m at wellheads and compressor stations, all with ATEX/IECEx/SIL 2/3 certification and HART or Foundation Fieldbus output [S1][S5]. LNG, LPG, and propane storage: point IR in %LEL at loading arms and pump skids, plus open-path across bund walls [S3]. Semiconductor and pharmaceutical cleanrooms: wall-mounted suction-type multi-gas detectors using IR for CO2, CH4, and VOC in the same housing as electrochemical O2 and toxic sensors — the New Cosmos SH-4002-WAD is representative, listing 17 gas types from NH3, NO, H2, H2S, O2, O3, benzene, Cl2, CH4, CO, SO2, HCl, HCHO, CO2 through VOC in one chassis with a built-in pump, replaceable sensor modules, and data logging [S2].
Two more clusters are growing. Indoor-air-quality / building management: low-cost NDIR CO2 (0-2,000 / 5,000 ppm) with analog or Modbus output, often the same ZMF-100-IR-class module [S3]. Refrigerant-leak detection (HFC, HFO, CO2 in commercial refrigeration): IR and NDIR specifically because refrigerants are not combustible and not detectable by LEL sensors — only IR, semiconductor, or photoacoustic methods apply, and only 1 vendor in the index flags refrigerant gas explicitly [S3]. Mine-methane and coal-bed-methane drainage: TDL ppm-m on ventilation returns.
Failure Modes and Limitations to Spec In
Three failure modes recur in field service reports. (1) Window contamination — oil mist, salt, hydrate, or paint overspray fouling the optical window takes the reading to a low-bias state and, because IR is "failsafe on the high side" only for diagnostic faults, can mask a real leak. Mitigation: heated optics, purge gas, automatic obscuration alarm channel. (2) Condensation in the cell on rapid warm-up from cold ambient — pick a unit with explicit low-temperature-startup validation, not just a storage spec. (3) Pressure and humidity cross-sensitivity on dual-wavelength designs when the cell is operated at altitude or under rapid pressure transients — partial-pressure correction is mandatory for %-volume CO2 custody-transfer or breathing-zone IAQ. [S1]
Two hard limitations do not go away: IR cannot detect homonuclear diatomics (N2, O2, H2, Cl2 in some cases) because they have no IR-active dipole moment — so the 9-vendor H2 and 9-vendor O2 "IR" listings in the index are mostly IR-confirmation or non-IR modules carrying IR labels for multi-gas cabinet packaging, not pure IR measurement [S3]. And the infrared-thermometer and gas-detector crossover is conceptual only — the optical trains share emitter/filter/detector parts, but the spectral filter, modulation scheme, and gas-cell engineering are completely different from a pyrometer; do not cross-spec.
Standards, Certification, and Sourcing Anchors
Safety-integrity for fixed hydrocarbon IR detectors is typically demonstrated to IEC 61508 SIL 2 and, with redundancy, SIL 3 (the Searchpoint Optima Plus product line documents this with an optional HART SIL certificate) [S5]. Hazardous-area certification follows the ATEX 2014/34/EU directive (EU-Type Examination Certificate Supplement 12) for Europe and IECEx for the rest of the world; marine approvals include DNV and BV type approval for offshore deck and engine-room use [S5]. Performance testing for open-path flammable detectors falls under IEC 60079-29-4; for toxic gas it is IEC 60079-29-1. Wireless and portable IR instruments for mine and confined-space use additionally carry MSHA / SANS approvals in their target jurisdictions.
For new 2026 builds, specifiers should anchor the requirement in three documents: the project hazardous-area classification drawing, the target-gas and concentration-range matrix, and the Safety Instrumented Function (SIF) target SIL. With those, a procurement engineer can choose between an open-path IR5500-class instrument for perimeter leak detection, a point Optima-Plus-class instrument for SIL 2/3 LEL detection at a specific measurement point, and a SH-4002-WAD-class multi-gas suction detector for cleanroom toxic + combustible coverage — three architectures, three price points, three different engineering documents to back them [S1][S2][S5].
Two trackable signals to watch through 2026: (1) further migration from 4-20 mA + HART toward Ethernet-APL in greenfield chemical and refinery builds, which compresses IR-detector lead times for digital plants; (2) consolidation of low-cost NDIR CO2 modules (ZMF-100-IR class) into BMS, DCV (demand-controlled ventilation), and ASHRAE 62.1 occupant-density loops as IAQ regulations tighten in EU and APAC jurisdictions [S3][S4].
Related analysis: Magnetic Material Types and Classifications: A Spec-Driven Engineering Map.