Gas Sensors

Infrared (NDIR) LEL & CO₂ Sensor

Measures flammable gas (and CO₂) by how it absorbs infrared light. Immune to poisoning and works with no oxygen — but has one blind spot that matters enormously: hydrogen.

OPS flammability survey where fitted TECH choosing IR vs. cat-bead, cal-gas selection, inert-atmosphere work

How it works

NDIR stands for Non-Dispersive InfraRed. Inside is an infrared light source, a sample chamber the air flows through, and a detector. Many gas molecules absorb infrared light at specific wavelengths — a hydrocarbon has a characteristic "IR fingerprint." The sensor shines IR through the sample and measures how much light is absorbed at the wavelength its filter is tuned to. More gas → more absorption → less light reaching the detector → higher reading. A second reference wavelength (where the gas doesn't absorb) compensates for dust, drift, and optics fouling.

Crucially, nothing is burned and nothing chemically reacts with the sensor — the gas just passes through a light beam. That single fact drives most of NDIR's advantages over the catalytic bead.

Gas absorbs IR at its characteristic wavelength; the drop in transmitted light is the signal. No combustion, no chemical reaction.

What it's good for

Immune to poisoning. Silicones, sulfur, lead, and halogens that wreck a catalytic bead don't touch an optical sensor. This is the reason to choose NDIR where poisons are likely (firefighting-foam residues, sewers, industrial degreasers).
Works in inert / oxygen-deficient atmospheres. No combustion needed, so NDIR reads correctly with little or no oxygen — purged tanks, nitrogen-blanketed vessels, CO₂-flooded spaces. This is where the cat bead fails.
No burnout at high concentration. Drive it to 100% LEL and beyond and it doesn't damage itself or roll over the way an old cat bead can; it fails high/over-range rather than deceptively low.
Stable and low-drift with a long sensor life.
CO₂ measurement. CO₂ has a strong IR signature (and no combustion signature), so NDIR is the standard way to measure carbon dioxide — useful for confined spaces, fermentation, dry-ice, fire suppression discharge.

What it CANNOT do / limitations

⚠ Warning — NDIR is BLIND to hydrogen

Infrared detection relies on the gas absorbing IR light. Hydrogen (H₂) is a symmetric diatomic molecule — it does not absorb infrared, so an NDIR sensor cannot see it at all. It will read a flat zero in a hydrogen-rich, fully explosive atmosphere. This is a life-threatening gap around battery/UPS rooms, hydrogen fuel systems, charging areas, and lithium-ion battery fires (which vent hydrogen and other flammables). If hydrogen is possible, you need a catalytic bead or a dedicated H₂ sensor — NDIR alone will get you hurt.

Other limitations:

Symmetric diatomic gases are invisible to IR generally — besides H₂, that includes O₂ and N₂ (not combustibles, but worth knowing why IR "can't see everything").
Doesn't identify — it's a %LEL/CO₂ number, not a chemical name.
Response is strongly gas-dependent (see next section) — worse than the cat bead in this respect.
Optics can be fouled by condensation, heavy dust, or oil film, blocking the beam.

Response varies strongly by hydrocarbon — cal gas selection

Because every hydrocarbon absorbs IR a bit differently, an NDIR LEL sensor's response varies even more by compound than a catalytic bead's. A sensor calibrated to methane may respond quite differently to propane, hexane, or gasoline vapor. You must:

Calibrate to the reference gas your SOP/manual specifies, and know what it is.
Apply the manufacturer's correction factor for a known target gas.
On an unknown, treat %LEL as a relative trend with generous margin — the true value could be substantially higher for a poorly-responding compound.

Field Tip

NDIR and catalytic beads have different response curves for the same fuel. If your five-gas has both, disagreement between them is a useful flag — and if one channel reads zero while you have combustion products or an H₂ source, suspect the blind spot (IR = no hydrogen; cat bead = low O₂/poisoned).

Degradation & failure modes

Condensation / fogging on the optics — going from cold to warm, or high humidity, can fog the mirrors/windows and block the beam, giving false or drifting readings until it clears.
Dust, dirt, oil film on the optical surfaces attenuates the beam; the reference channel compensates somewhat but heavy fouling still degrades accuracy.
Source/detector aging — slow, and largely handled by the reference channel, but real over years.
No poisoning, no burnout — the two big cat-bead killers simply don't apply.

Catalytic bead vs. NDIR — which is the right choice?

Choosing between combustible-sensor technologies. Many 5-gas monitors can carry both.
Factor	Catalytic bead (pellistor)	Infrared (NDIR)
Needs oxygen?	Yes — unreliable/low below ~10% O₂	No — works in inert atmospheres
Detects hydrogen?	Yes	No — blind to H₂
Detects acetylene?	Yes	Weak/poor for some — verify
Poisoning (silicone/sulfur/lead)	Vulnerable — reads low when poisoned	Immune
High-concentration behavior	Can burn out / roll over past 100% LEL	Fails over-range (high), no damage
Failure direction	Often low (dangerous)	Usually zero-for-H₂ or over-range (more obvious)
Optics fouling / condensation	Not applicable	Susceptible
Best for	General combustibles in normal air; H₂ present	Poison-heavy or O₂-deficient/inert spaces; CO₂

✓ Remember — they cover each other's blind spots

Cat bead fails in low oxygen and when poisoned; IR fails on hydrogen. That's exactly why serious flammability work uses both, or at least chooses deliberately based on the scenario. Battery room? You need the cat bead (or a hydrogen sensor). Purged tank full of solvent vapor and foam residue? You need the IR.

Calibration & bump test schedule

Bump test before each day's use (ISEA position) — even though NDIR can't be poisoned, a bump verifies the optics, electronics, pump, and alarm still work and catches a fogged/fouled sensor.
Full calibration per the manual — commonly monthly, extendable with a documented bump program.
Calibrate after a failed bump, optics cleaning, a big temperature/humidity swing that may have fogged the chamber, or a suspected over-range event.
Calibrate to the specified reference gas and apply correction factors for known targets.

Field care & storage

Let the instrument acclimate between temperature extremes to avoid condensation on the optics before you rely on it.
Keep the sample path clean — use inlet filters and keep out oil mist, heavy dust, and water.
Store dry; watch for fogging when moving from an air-conditioned rig into hot, humid air.

Common rookie mistakes

⚑ Common Rookie Mistakes

Deploying an IR-only flammability meter in a battery room or on a lithium-battery fire and trusting its zero — it can't see the hydrogen.
Assuming "immune to poisoning" means "never needs bumping." It still fogs, drifts, and can lose its pump/alarm.
Forgetting the strong compound-dependent response and taking an uncorrected %LEL as absolute.
Being fooled by condensation on the optics after a cold-to-hot transition.

Representative instruments

In a RAE fleet, NDIR appears mainly as the CO₂ sensor option on the MultiRAE family (and on AreaRAE area monitors). NDIR LEL channels are options on various multi-gas platforms across manufacturers (e.g., Dräger X-am, Industrial Scientific Ventis/Radius, MSA Altair) — check which combustible-sensor type your specific units carry, because the hydrogen blind spot only applies to the IR version. Some units carry both an IR LEL and a catalytic LEL specifically to close each other's gaps. Brands are illustrative; your model and manual govern.

Next: the broadband VOC survey tool — Photoionization Detectors (PID) →