NDIR Gas Analyzers — Non-Dispersive Infrared Technology

NDIR stands for non-dispersive infrared. Unlike scanning FTIR, which resolves the full spectrum, NDIR uses a broadband IR source and one or more bandpass filters that isolate a pre-selected absorption feature. This is why NDIR is cheaper, simpler, and mechanically more stable — and also why its selectivity is lower than TDLAS (single laser line) or FTIR (full spectral deconvolution). It is the right tool for mature, high-volume gas targets like CO and CO₂ where the absorption feature is strong and the matrix is well understood.

Three reasons. First, CO₂ has an unusually strong, well-isolated mid-IR absorption at 4.26 μm that NDIR exploits efficiently. Second, the target concentration range for most applications — from 400 ppm ambient up to 20 %vol flue gas — is comfortably within NDIR dynamic range. Third, commoditization: NDIR CO₂ modules have been volume-manufactured for decades, so capex and OEM integration are far cheaper than a TDLAS laser head. TDLAS only wins on CO₂ when ppb-trace or single-laser multi-gas requirements dominate the project.

Water has broad continuum absorption across the mid-IR that overlaps many target bands — especially CO near 4.6 μm and N₂O near 4.5 μm. The three engineering options are: (1) dry the sample upstream via a Peltier cooler or Nafion tube (simplest, but can strip water-soluble target species); (2) use dual-beam optics with a reference filter at an adjacent wavelength to ratio out the continuum; or (3) use GFC architecture with a sealed target-gas reference cell that self-references against the wet matrix at the target feature. GFC is how most compliance-grade wet-stack N₂O and CO analyzers handle this.

Usually no. NDIR detection limits are set by broadband filter bandwidth, source-intensity fluctuation, and Beer-Lambert shot noise, and realistic low-range duty is around 0.5–5 ppm on clean matrices. Anything ppb-level — fenceline toxics, fuel-cell-grade hydrogen impurity audits, semiconductor ppb HCl / HF work — belongs to TDLAS, CRDS, or lab-grade chromatography, not NDIR. Buyers who were quoted a “ppb NDIR” should re-read the datasheet for the matrix and path-length assumptions.

GFC — Gas Filter Correlation — is a variant of the NDIR optical family, not a separate technology. The difference is that the reference path runs through a sealed cell filled with the target gas itself. When the rotating filter wheel passes the target-filled cell in front of the beam, all of the target absorption is pre-absorbed, so the detector sees “everything except target.” When the non-filled cell is in place, the detector sees “everything including target.” The ratio of the two states cancels matrix interferences at the target molecule’s own spectral fingerprint. GFC costs more than plain NDIR and adds a rotating mechanism, but it is what turns NDIR into a CEMS-grade instrument on difficult matrices.

In the US, Performance Specifications 3 (O₂ / CO₂) and 4 / 4A (CO) under 40 CFR Part 60 Appendix B set the ongoing accuracy / calibration drift targets, with Reference Methods 3A and 10 for instrumental measurement. In the EU, MCERTS (UK Environment Agency) and EN 15267 / EN 14181 QAL1 / QAL2 / AST cover type-approval and plant-level calibration. Greenhouse-gas MRV adds 40 CFR Part 98 (US) and Regulation 2018/2066 (EU ETS). Procurement should pin the target certifications in the specification, not rely on vendor marketing badges.

An NDIR gas analyzer works by passing infrared light through the sample and measuring how much the target gas absorbs at its characteristic wavelength, in four steps. First, a broadband IR source emits across the mid-IR; the beam is chopped or modulated so the electronics can separate it from background heat. Second, the beam crosses the sample cell, where the target gas (CO, CO₂, CH₄, N₂O) absorbs at its band — for example CO₂ at 4.26 μm. Third, a bandpass filter in front of the detector isolates that band and the detector measures the surviving intensity; the fractional absorption follows the Beer-Lambert law, so it scales with gas concentration and path length. Fourth, a single-beam design auto-zeros on clean gas, while a dual-beam design ratios the sample against a sealed reference path to cancel source drift. The firmware then converts absorption to ppm or %vol on a 4-20 mA / Modbus output.

Single-gas NDIR is enough when you measure one or two well-defined IR-active gases; you need FTIR when the job demands many species at once or a multi-component reference method. Pick NDIR when the target list is short and stable — CO, CO₂, CH₄, N₂O on a known matrix — because one bandpass filter per gas is cheap, rugged and mature, and a single-gas NDIR easily covers ppm-to-%vol process and CEMS duty. Move to FTIR when you need roughly four or more species simultaneously, when speciation across HCl / HF / NH₃ / N₂O matters (waste-incineration EU BAT, semiconductor), or when a permit names a multi-component method such as EPA Method 320 — duties where stacking four NDIR channels is costlier and clumsier than one spectrometer. The crossover is component count and method, not raw accuracy. For how FTIR works as a method and how it compares to TDLAS and NDIR, see the FTIR vs NDIR method guide; for GESHINE multi-gas hardware the production route is configured multi-pass TDLAS. This NDIR page remains the single-gas IR home.

A closed-path (extractive) NDIR analyzer draws a gas sample into a sealed internal cell of fixed path length, while an open-path NDIR measures across an open distance in the process or ambient air with no sample extraction. In a closed-path design a probe and heated line pull stack or process gas through conditioning (filter, cooler) into the analyzer’s own cell, where temperature, pressure and path length are controlled — which is why extractive NDIR dominates CEMS and process duty where a stable, repeatable reading matters. Open-path NDIR sends the IR beam across a duct, a stack diameter or a fenceline span to a retroreflector or separate receiver, giving a path-averaged concentration with no sample handling — useful for area and fenceline monitoring but exposed to alignment, dust and weather. The trade-off is control versus coverage: closed-path for a precise point measurement, open-path for a spatially averaged one.

An NDIR gas cell is calibrated with a two-point zero and span: flow a zero gas (clean N₂ or scrubbed air) to set the no-absorption baseline, then a certified span gas near the top of the range to set the slope, with the firmware fitting the Beer-Lambert response between. Frequency depends on duty: process indicators may hold for weeks, but regulated CEMS typically run an automatic daily zero and span check against reference cylinders, logged for audit, with a manual multi-point verification on a longer interval as part of the QA routine. The drift sources are source aging, detector response and window fouling — which is exactly what the zero/span cycle catches before it biases the reading. Dual-beam and GFC designs self-reference between calibrations to stretch the interval, but they do not remove the need for traceable span gas. Match the cadence to the matrix and the regulator, not to a fixed calendar.

No — a portable NDIR analyzer is built for spot checks and short surveys, not for continuous fixed CEMS duty. A portable NDIR is optimised for battery life, fast warm-up and one-hand operation for boiler tune-ups, leak surveys and commissioning, and it auto-zeros between measurements rather than holding a traceable baseline for months. A fixed CEMS runs unattended around the clock with conditioned sampling, automatic daily zero/span, a logged data record and the type-approval (US Performance Specifications, EN 15267 / EN 14181) that regulated reporting requires — none of which a portable carries. Using a portable as a stand-in for a CEMS fails on uptime, traceability and certification, and an auditor will reject the data. The right pattern is a fixed CEMS at the regulated point and a portable for surveys and cross-checks around it.