The short version: the housing, the electronics and the signal processing are shared. The laser, the detector, the spectral line and everything downstream of the line choice are not. Ten gases is not one product family. It is ten.

This page shows where the boundary sits, so you can ask a supplier the question that separates a built gas from a listed one.

The wavelength map decides the hardware

Every gas absorbs at its own wavelengths. The wavelength decides the laser, and the laser decides the detector.

Diode lasers used for gas detection cover roughly 1260 to 2400 nm. That band carries H₂O, CO₂, CH₄, NH₃, HF, H₂S and CO. Within it, an InP-based DFB laser pairs with an InGaAs detector, and a module built for one of these gases shares its optical stack with the others.

Oxygen does not sit in that band. It absorbs near 760 nm, which needs a GaAs laser and a silicon detector. Nothing in the optical stack carries over. A supplier who lists oxygen alongside methane is either running two device stacks or has not built one of them.

O₂ 760 nm GaAs laser Si detector H₂O · CO₂ · CH₄ · NH₃ · HF · H₂S · CO 1260–2400 nm InP DFB laser InGaAs detector CH₄ 1650 nm longer wavelengths above 2400 nm QCL or ICL third device stack 700 1000 1500 2000 2500 3000 Wavelength, nm Three device stacks. Nothing optical carries between them.
Oxygen sits in a different band from the other gases, and shares none of their optical stack. A gas list that spans these bands spans more than one product line.

This is also where TDLAS earns its selectivity. A DFB laser has a linewidth on the order of 10 MHz or below. A traditional non-dispersive infrared source has a linewidth of hundreds of GHz. That difference of several orders of magnitude is why one technique can sit on a single absorption line and the other cannot. It is a property of the light source, not a claim about a product.

What carries across gases, and what does not

Shared across the platform Mechanical envelope and housing Power input and communication Modulation and demodulation Signal-processing architecture 2f/1f normalisation framework Gas path fittings, in most cases Built once per gas Laser, for that absorption band Detector, matched to the laser Spectral line selection Calibration, per gas and per range Cross-interference verification Gas path material and heating A gas list counts the right column, not the left.
The housing and the electronics carry across gases. Everything from the laser down is built again.

The last item is the one most often left out of a datasheet. Ammonia adsorbs onto surfaces. A module reading ammonia accurately depends on the wetted material and on whether the line is heated, which means part of the accuracy specification lives outside the module. A supplier quoting an ammonia figure without stating the gas path conditions is quoting half a specification.

Line selection is where the background gas matters

Choosing an absorption line is not just picking the strongest one.

The line has to be strong enough at your concentration range, and it has to be clear of the absorption features of everything else in your gas. Change the background and the answer changes. A line that works cleanly in nitrogen may sit under a water or CO₂ feature in a real process stream.

In nitrogen target line clear on both sides In a stream with CO₂ and H₂O target line background feature the same line now sits under another absorber Wavelength → A cross-interference figure without its background gas describes a different stream from yours.
A line that is clear in nitrogen can sit under another gas’s feature in a real process stream. This is why the background composition has to travel with the figure.

This is why cross-interference figures are meaningless without the background composition they were measured against. It is also why “we support that gas” and “we have run that gas against your background” are two different statements.

Ask for

Which wavelength and laser type, which detector, whether line selection was done against your background gases, what calibration ranges exist, and what the cross-interference check covered. Those five answers tell you whether a gas on the list has been built.

What we carry as standard, and what we do not

Our Box-level Extractive TDLAS Gas Measurement Module carries four standard gases: CH₄, CO₂, CO and NH₃. All four sit in the 1.5 to 2.7 µm band and share one device stack.

Oxygen is not in that set. It is available as a project adaptation, and we say so rather than adding it to a list, because it needs the 760 nm stack described above. The same applies to H₂O, H₂S, HCl, C₂H₂ and C₂H₄: possible as project work, not standard product.

Narrowing the standard set is a choice about what we can hold to a published specification. A shorter list that each entry can back is more useful to you than a longer one you have to interrogate.

Accuracy is given per gas and per range, in two segments

For CH₄ on the 0–5000 ppm range: ±10 ppm below 500 ppm, ±2 % of reading above it. On the 0–100 %vol range: ±0.05 %vol below 1.00 %vol, ±3 % of reading above it. Every other gas and range has its own pair. We do not publish a single ±X % of full scale across four gases and eight ranges, because that specification cannot hold across a range whose full scale changes by four orders of magnitude.

Figures apply at a sample flow of 1 L/min at 20 °C and 101.3 kPa, module at 25 °C, ambient pressure, nitrogen balance, after warm-up. T90 is 30 seconds at that same 1 L/min.

What we do not publish

Detection limits, cross-interference figures and gas path material are confirmed at configuration, against your background composition and your range. We do not put a platform-wide number on any of the three, because a number that has not been checked against your background is not a number you can design with.

Where this module does not go

It runs on 12 V at up to 3 W, which places it outside any intrinsic safety envelope. For a coal mine or any Ex-classified installation, this is the wrong module and we will say so early. → Pipeline Insertion TDLAS Methane Measurement Module