An infrared open path detector works on Beer-Lambert absorption: a transmitter projects a narrow IR beam across the monitored zone, and the receiver measures how much energy arrives at one or more specific wavelengths chosen to match the absorption lines of the target gas [S4][S3]. The deeper the gas cloud and the higher its concentration, the larger the fractional loss at the measurement wavelength, while a parallel reference wavelength stays nearly untouched and is used to cancel out fog, rain, snow and partial obscuration [S4][S2].
Typical path lengths run from a few metres to a few hundred metres, with a fixed full-scale of 5 LFL.m and alarm thresholds at 1 LFL.m (low) and 3 LFL.m (high) on North Sea offshore fire and gas systems, a configuration that reflects the instrument's natural bias toward total gas volume rather than local concentration [S4].
Beam Geometry: Transmitter, Receiver, and the Three Optical Topologies
The most common layout uses a dedicated IR transmitter at one end and a separate receiver at the other, each mounted on stable, vibration-isolated structures, because any drift in beam alignment produces a steady loss of signal and eventually a fault [S2][S4]. A second topology folds the path back on itself: the source and detector sit in the same housing, and the beam is returned by a passive retroreflector at the far end, which simplifies alignment and reduces the number of powered field devices [S4]. A third, used mainly for portable instruments, replaces the retroreflector with the natural albedo of nearby walls, terrain, or equipment, at the cost of much weaker and less stable return signal [S4]. The longer the path, the more sensitive the device becomes to a distributed cloud but the more vulnerable it becomes to partial blockage, which is why rain/sun hoods over both optics are standard practice even on weather-tolerant designs [S2].
Path length is also the dominant variable in coverage planning: a single open path beam can replace a string of point detectors along a pipeline corridor or tank-farm perimeter, but it cannot tell the operator whether the gas is one large diffuse cloud or one tight high-concentration puff at an unknown point along the line [S2][S4]. For practical leak detection this is usually acceptable, because the alarm is raised on total hydrocarbon mass in the beam, which is the quantity that drives explosive-cloud growth, not on the local ppm at a single sniff point [S4].
The Measurement Principle: Two Wavelengths, One Ratio
Open path detectors are not simple energy meters. They select one or more measurement wavelengths at peaks of the target gas's IR absorption spectrum and pair each with a reference wavelength that the same gas does not absorb, then form the ratio of the two received signals [S4]. This ratiometric architecture is what makes the reading stable in fog, rain, and snow: common-mode attenuation hits both wavelengths almost equally and drops out of the ratio, while genuine gas absorption depresses only the measurement wavelength and shows up as a clean differential signal [S4]. Spectrex's Quasar 900 series, for example, packages this as a "spectral fingerprint" Differential Optical Absorption Spectroscopy (DOAS) analysis of the path atmosphere [S2].
Because every hydrocarbon family has a slightly different fingerprint, the choice of measurement wavelengths, and therefore the firmware, is what defines which gases an open path unit will see. A detector aimed at C1-C5 alkanes will respond to methane, ethane, propane, butane and pentane with weighting set by their absorption cross-sections, while a unit tuned for heavier fuels will weight C6+ fractions more heavily [S3][S7]. This selectivity is also why the same beam geometry can be marketed as "hydrocarbon" or "toxic H2S" but the latter has, per the technical literature, so far failed to reach widespread adoption because ppm-level detection demands much longer integration and tighter wavelength control than percent-level LFL work [S4].
Path-Integral Units: Why the Reading Is LFL.m, Not ppm
The quantity the instrument actually outputs is the integral of gas concentration along the beam, which is why the natural unit is ppm.metres for toxic work and LFL.metres for flammable work, not the ppm or %LFL that point sensors report [S4]. A 1 LFL.m reading on a 50 m beam corresponds to an average of 0.02 LFL (2% of the lower flammable limit) along the path, while the same 1 LFL.m on a 10 m beam corresponds to 0.1 LFL average, so the same alarm threshold means very different things on different path lengths [S4]. Offshore installations typically standardise on a 5 LFL.m full-scale with low and high alarms at 1 LFL.m and 3 LFL.m, which gives roughly a 1.5:1 signal-to-alarm margin and a 5:1 headroom to full scale [S4].
This unit system is the single biggest source of confusion when comparing open path detectors to point detectors. A point infrared gas detector reports local concentration, while the open path unit reports the product of concentration and path length, so direct head-to-head alarm-setting is meaningless without converting both to the same physical quantity [S4]. For a process engineer sizing a fire and gas layout, the practical consequence is that open path is preferred when the leak's drift direction is unknown and the alarm should fire on total released mass, while point detection is preferred when the leak source is well known and the operator wants the earliest possible local warning [S4][S2].
Where Open Path Wins, Where It Loses
Open path excels in three scenarios: outdoor process areas where wind direction and plume evolution are unpredictable, linear assets such as pipelines and tank-farm perimeters where one beam covers what would otherwise need a string of points, and harsh or remote locations where weatherproofing a point detector's local sensor head is more expensive than mounting two beam-end enclosures [S4][S2]. Response time is also typically faster than a shielded point detector, because the beam sees gas the moment it enters the line of sight rather than waiting for diffusion through a weather hood [S3][S4].
It loses in three other scenarios: indoor enclosures where the beam length is shorter than the room, applications where the leak source is a single well-known fitting and a point detector can sit on top of it, and budget-constrained single-point loops where one open path unit costs more than one point infrared gas detector [S4]. Misalignment sensitivity is the other real-world failure mode: vibration, thermal growth of the mounting structure, or accidental impact can slowly walk the beam off the receiver and drive the unit into fault long before a real gas event occurs, which is why the OEM guidance is to mount both ends on rigid, low-vibration structures and to add sun/rain shields even though the optics are weather-tolerant [S2]. Modern designs mitigate this with continuous alignment scoring and automatic gain, but they do not eliminate the underlying mechanical requirement [S2].
Field Behaviour in Rain, Fog, and Partial Obscuration
Because fog, rain and snow attenuate both measurement and reference wavelengths almost equally, the ratiometric read is largely immune to weather, which is why open path detectors are routinely deployed on offshore platforms, LNG terminals, and desert gas-gathering stations where point detectors would false-alarm in every storm [S2][S4]. The remaining weather effects are drift in the absolute signal level (handled by automatic gain control) and slow contamination of the optics windows (handled by periodic cleaning and by self-checking routines that flag a blocked-lens fault rather than a gas alarm) [S2][S3]. The same ratiometric logic, incidentally, is why the technology is closely related to point-type NDIR sensors: the same Beer-Lambert absorption physics applies, only the optical path is folded into a few centimetres inside a point cell instead of stretched across tens of metres of open air, and a fixed gas detector built on NDIR is essentially a short-path version of the same idea.
Operators should still treat the reference channel as a health indicator: if the reference signal drops more than the measurement signal, the loss is almost certainly obscuration, not gas, and the controller should suppress the alarm. If the measurement signal drops more than the reference, the controller should treat it as a genuine gas event. Most modern transmitters implement this check in firmware and expose the ratio directly, which is the cleanest way to integrate the device into a Safety Instrumented Function without nuisance trips [S4].
Selection Checklist for Specifying an Open Path Detector
Five parameters drive the spec. Path length: pick the receiver/transmitter pair whose certified range brackets the actual installation distance, leaving margin for thermal expansion and minor misalignment. Target gas: confirm the measurement wavelengths are tuned to the hydrocarbon mix or toxic species of interest, since a generic "hydrocarbon" unit will not give equal weight to methane and to hexane vapours. Output unit: specify LFL.m (or ppm.m for toxic work) in the cause-and-effect matrix, not %LFL, so the alarm thresholds remain physically meaningful regardless of beam length. Environmental rating: confirm ATEX/IECEx zone classification, operating temperature range, and IP rating of both end units, since they are usually mounted in classified outdoor locations. Alignment and diagnostics: require continuous beam-strength indication, automatic gain reporting, and a Modbus or HART diagnostic register so the maintenance system can trend alignment drift before it becomes a fault [S2][S4].
For toxic gas service, the honest position in 2026 is still that the technology is proven in principle but rarely deployed in practice, because the path-integrated ppm.m signal at the relevant part-per-million concentrations is much smaller and harder to distinguish from atmospheric variability than the percent-level LFL.m signal used for hydrocarbon work [S4]. Most toxic-gas loops therefore remain point-catalytic or point-electrochemical, with open path reserved for the perimeter and the hydrocarbon inventories. A useful engineering reference for related optical measurement thinking is the optical comparator shaft-projection method comparison, which, while mechanical rather than gas-detection, follows the same beam-vs-target contrast principle used in open path optics.
Trackable signals to watch over the next two quarters: any IECEx or ATEX certification update listing extended certified path lengths beyond the current ~200 m ceiling, and any OEM move to push ppm-level toxic open path into a commercial product line, which would mark the first real challenge to point-electrochemical dominance in H2S and NH3 loops [S4].
Detailed specification references: open channel flowmeter.