Gas Detector Cross-Sensitivity: Why Sensors Read the Wrong Gas
What is cross-sensitivity in a gas detector?
Short answer: Cross-sensitivity is a sensor responding to a gas it was not designed to measure. Most portable toxic channels use electrochemical cells, which detect a target gas by reacting it at an electrode — and other substances that react in a similar way produce a signal too. The result is a reading, sometimes an alarm, attributed to the wrong gas. Hydrogen showing up on a carbon monoxide channel is the most commonly encountered example.
Cross-sensitivity is not a fault and it is not a sign of a defective instrument. It is an inherent characteristic of how the sensing chemistry works, and manufacturers document it because it cannot be engineered away entirely. What it means in practice is that a gas detector reading is a statement about sensor response, not an unambiguous identification of a substance. This guide explains the mechanism, sets out the categories of interference that show up in the field, and — importantly — explains why the only cross-sensitivity numbers you should act on are the ones published for your specific sensor and model.
It is written for confined-space crews, wastewater and utility staff, and safety managers investigating an alarm that does not match the known hazards of the space. If you are working through an unexplained reading right now, how to read a 4-gas monitor covers the other explanations that should be ruled out alongside this one.
Why this matters.
Cross-sensitivity cuts both ways, and the dangerous direction is the one people forget. A false positive erodes trust in the instrument and trains crews to silence alarms. But interference can also cause a sensor to read low or even negative in the presence of certain gases — meaning a real hazard is under-reported at the moment it matters. Under 29 CFR 1910.146 entry decisions rest on those readings being correct. Treat every alarm as real, and resolve cross-sensitivity as a diagnosis made afterwards from clean air, never as a reason to stay in a space.
Part 1 — Why electrochemical sensors respond to the wrong gas
An electrochemical cell contains electrodes in an electrolyte. The target gas diffuses in through a membrane, reacts at the sensing electrode — giving up or taking on electrons — and the resulting current is proportional to concentration. That current is what the display converts into ppm.
The cell has no way of confirming which molecule produced the current. Any substance that reaches the electrode and undergoes a similar reaction at a similar potential generates a similar signal. Selectivity is improved by chemistry and construction — choosing an electrode material and operating potential that favour the target reaction, and fitting a chemical filter that absorbs known interferents before they reach the cell — but no cell is perfectly selective, and none of those measures is permanent.
The filter is a consumable
The chemical scrubber that keeps an interferent away from the electrode has a finite capacity. As it saturates, interference that the sensor previously rejected begins to get through. This is one reason an aging cell can start producing readings it never used to — the cell itself may be fine while its filter is exhausted. It is also why cross-sensitivity behaviour is specified for a new sensor and can drift over the service life.
Part 2 — The categories of interference you actually meet
Rather than a numeric table — see Part 4 for why this page does not publish one — these are the interference patterns that recur across manufacturers and cell types.
| Pattern | What happens | Where it shows up |
|---|---|---|
| Hydrogen on a CO cell | Hydrogen oxidises readily at the same electrode and is the classic positive interferent on carbon monoxide channels. | Battery charging rooms, forklift charging bays, some digestion and process gases. |
| Reducing gases across toxic cells | Substances that give up electrons easily can add signal on cells intended for other reducing gases, inflating a reading. | Wastewater, digesters, refining and any mixed-contaminant atmosphere. |
| Oxidising gases producing negative readings | Strong oxidisers can drive current the opposite way, suppressing a reading or pushing it below zero — the under-reporting direction. | Water treatment, bleaching, chemical dosing areas. |
| Catalytic LEL sensors are non-selective by design | A catalytic bead responds to essentially any combustible, scaled to its calibration gas. Not a defect — the channel is meant to catch flammables generally. | Any space with solvents, fuels or vapours other than the calibration gas. |
| PID lamps respond by ionisation energy | A photoionisation detector reports everything ionisable by its lamp as one total figure — it does not separate compounds. | Solvent work, tank cleaning, spill response — see the VOC detectors shelf. |
Non-selectivity is not the same as cross-sensitivity
The last two rows deserve separating from the first three. A catalytic bead sensor responding to propane when it was calibrated on methane is the sensor doing its job — it is a general combustible detector, and the %LEL number just needs the correct correlation factor applied, as covered in ppm vs %LEL vs %volume. A photoionisation detector reporting a total VOC figure is likewise behaving as specified. Cross-sensitivity in the strict sense is a cell intended to be selective proving not to be.
Part 3 — Cross-sensitivity in the under-reporting direction
Most discussion of interference focuses on nuisance alarms, because those are the ones that annoy people. The more serious case is suppression.
If an interferent drives current in the opposite direction to the target gas, the two signals partly cancel. A cell exposed to both simultaneously can report less than the true concentration of the hazard — and in the extreme, a channel sitting at zero or reading negative in an atmosphere that genuinely contains the gas it is there to find. A negative reading on any channel is therefore never cosmetic. It means the cell is being acted on by something, and until you know what, the instrument's other readings deserve the same suspicion.
This is the strongest practical argument for two habits: zero the instrument only in air you know is clean, and treat a channel that will not sit at zero as a failed instrument rather than a quirk. The daily function check that catches most of this is covered in bump test vs calibration for gas detectors.
Part 4 — Why this page does not publish a cross-sensitivity table
Cross-sensitivity figures are usually expressed as a percentage: how much apparent target gas a given concentration of interferent produces. Those numbers are real and useful — and they are specific to a particular sensor, in a particular instrument, from a particular manufacturer, when new.
They vary by electrode chemistry, by operating potential, by whether a filter is fitted and how depleted it is, and by temperature and humidity. Two instruments from different makers, both nominally carrying a carbon monoxide cell, can have materially different interference behaviour. Publishing a generic table would invite exactly the error this page exists to prevent: applying a number that does not describe the instrument in your hand.
So the correct source is the cross-sensitivity table in the documentation for your specific sensor and model, from the manufacturer, for the cell as configured. If you cannot find one, that is itself worth knowing before you rely on the instrument in a mixed atmosphere.
Part 5 — Worked example: diagnosing a CO reading in a battery room
A maintenance crew is working near a forklift battery charging bay. A wearable such as the Honeywell BW Clip4 4-Gas Detector alarms on carbon monoxide. There is no combustion source anywhere nearby. Charging lead-acid batteries evolve hydrogen, and hydrogen is the best-documented positive interferent on carbon monoxide cells — so cross-sensitivity is a plausible explanation. It is not, however, a conclusion you can reach inside the room.
- Leave first, diagnose second. The alarm is treated as a real carbon monoxide reading until proven otherwise. Nobody investigates a possible interference while still in the atmosphere producing it.
- Record what the instrument saw. Note the peak value and the time. If the instrument datalogs, the profile — a slow climb during charging versus a sharp spike — is diagnostic information you cannot reconstruct later.
- Check the instrument in known-clean air. Outdoors and upwind, confirm the channel returns to zero. A channel that will not re-zero points at the sensor, not at the atmosphere.
- Bump test all four channels. Confirm the cell still responds correctly to its target gas at a known concentration. A cell that reads a real interferent may also be at end of life.
- Consult the manufacturer's cross-sensitivity table for that exact sensor. Confirm whether hydrogen is listed as an interferent for the cell in your instrument, and at what magnitude. This is the step that turns a hypothesis into a documented finding.
- Verify the actual hazard independently before changing anything. If the conclusion is hydrogen interference, the room still has a hydrogen accumulation question of its own, and that needs the right sensor — the hydrogen gas detectors shelf covers dedicated H2 monitoring. Never resolve an interference finding by simply raising the alarm threshold.
The same discipline applies to the instruments on the portable gas detectors and personal gas detectors shelves, and to permanently mounted sensors from the fixed gas detection systems shelf, where a nuisance alarm is more likely to be silenced at a panel than investigated.
Part 6 — Designing around interference instead of arguing with it
Once an interference is documented for a specific instrument in a specific setting, there are only a few legitimate responses, and raising the alarm threshold is not among them.
- Measure the interfering gas directly. If hydrogen is producing apparent carbon monoxide, the hydrogen is itself worth knowing about. Adding the right sensor turns a nuisance into data — the gas detector hub groups instruments by target gas.
- Choose a sensor with a filter rated for that interferent, and treat the filter as a consumable with a replacement interval rather than a permanent feature.
- Match the instrument to the atmosphere rather than to convention. Where a space routinely contains a documented interferent, a dedicated single-gas monitor for the real hazard alongside the multi-gas unit removes the ambiguity. The trade-off is worked through in 4-gas monitor vs single-gas detector, and the dedicated options are compared in the best H2S monitor guide and the best industrial CO monitor guide.
- Record the finding in the program, not in someone's memory. An interference that is understood by one experienced operator and nobody else will be rediscovered as a mystery alarm on the next shift.
- Keep ventilation as the first control. Interference is a measurement problem; accumulation is a hazard problem. Purging a space before entry addresses both — blowers and supporting kit are on the confined space equipment shelf.
The reason threshold-raising is singled out as the wrong answer is that it is indistinguishable, at the panel, from raising the threshold on a genuine hazard. A permanently mounted sensor from the fixed gas detection systems shelf that has been desensitised to stop nuisance alarms has also been desensitised to the real thing, and nothing on the display will say so.
Frequently asked questions
What is cross-sensitivity in a gas detector?
It is a sensor responding to a gas other than the one it is meant to measure. Electrochemical cells detect a target gas by reacting it at an electrode, and other substances that react similarly generate a similar current, which the instrument reports as the target gas.
Why is hydrogen making my CO monitor alarm?
Hydrogen oxidises readily at the same electrode used to detect carbon monoxide, making it the best-documented positive interferent on CO cells. Battery charging areas are the classic setting. Confirm against the cross-sensitivity table for your specific sensor before concluding that is what happened.
Can another gas cause a false reading on my H2S sensor?
Yes. Both positive interference, which inflates the reading, and negative interference, which suppresses it, are possible depending on the substance and the cell. Which gases affect your sensor and by how much is published by the manufacturer for that specific cell.
Can cross-sensitivity make a reading too low?
Yes, and this is the more dangerous case. An oxidising interferent can drive current opposite to the target gas, partly cancelling the signal, so the channel under-reports a hazard that is genuinely present. A negative reading is a warning sign, never a cosmetic quirk.
Why does my monitor show CO when there is no CO?
Common explanations are a real source you have not found, a zero captured in air that was not clean, an interfering gas, or a cell at end of life. All of them are diagnosed from clean air after leaving, not from inside the space.
Is cross-sensitivity a defect?
No. It is an inherent property of electrochemical sensing that manufacturers characterise and publish rather than eliminate. Selectivity is improved through electrode chemistry, operating potential and chemical filters, but no cell is perfectly selective.
What is the difference between cross-sensitivity and non-selectivity?
Cross-sensitivity is a sensor meant to be selective responding to the wrong gas. Non-selectivity is a sensor designed to respond broadly — a catalytic bead reacting to any combustible, or a photoionisation detector reporting all ionisable compounds as one total.
Do chemical filters stop cross-sensitivity?
They reduce it for specific interferents, but they are consumables with finite capacity. As a filter saturates, interference it used to block starts reaching the cell — which is why an older sensor can produce readings it never used to.
Does a catalytic LEL sensor have cross-sensitivity?
It responds to essentially all combustibles by design, so the issue is not selectivity but scale: the reading is expressed relative to the calibration gas. Applying the manufacturer's correlation factor for the gas actually present is what makes that number meaningful.
Where do I find the cross-sensitivity table for my detector?
In the manufacturer's documentation for the specific sensor and instrument model — usually the sensor datasheet or the technical section of the instrument manual. Generic tables found elsewhere may describe a different cell chemistry entirely.
Can I calibrate cross-sensitivity out of a sensor?
No. Calibration sets the instrument's response to a known concentration of its target gas; it does not change which other substances the cell reacts to. An interference problem is addressed by sensor selection, filters, or measuring the interfering gas directly.
Should I ignore an alarm I think is cross-sensitivity?
No. Treat every alarm as real, leave, and diagnose from clean air. An interference hypothesis formed inside the space is a guess, and acting on a guess is how a genuine exposure gets rationalised away.
Does humidity or temperature affect cross-sensitivity?
Both affect electrochemical cell behaviour generally, and published interference figures are specified under stated conditions. Readings taken well outside those conditions deserve more caution, particularly after moving an instrument between very different environments.
Can cross-sensitivity affect the oxygen channel?
Oxygen sensing is less prone to the interference patterns that affect toxic cells, but the oxygen reading is still an indirect indicator: it falls when any gas displaces oxygen, without identifying what did the displacing. That is a different limitation, not cross-sensitivity.
How do I confirm a reading is interference and not a real exposure?
Leave, re-zero in clean air, bump test the channel, check the manufacturer's cross-sensitivity table for that sensor, and where possible measure the suspected interfering gas directly with an appropriate instrument. Anything short of that is an assumption.
Does cross-sensitivity change as a sensor ages?
Yes. Published figures generally describe a new cell. Filter depletion and normal cell degradation can both change interference behaviour over the service life, which is one reason the response of an older sensor should not be assumed to match its datasheet.
Further reading on this site
- How to read a 4-gas monitor — the other explanations for an unexpected reading, and how to work through them.
- Bump test vs calibration for gas detectors — the function check that separates a bad cell from a real atmosphere.
- ppm vs %LEL vs %volume — why a combustible reading is always relative to its calibration gas.
- VOC detectors — photoionisation instruments, and why their total-VOC figure is not compound-specific.
- Hydrogen gas detectors — dedicated H2 monitoring for the battery-room case in the worked example.
- Gas detector calibration gas and accessories — single-gas cylinders for verifying a suspect channel.
- Gas detection complete guide — sensor technologies and instrument formats across the category.
Why trust this guide? WC Safety is an independent editorial desk covering industrial safety equipment. This guide was written by our editorial team, not by a sensor manufacturer or a paid third-party reviewer, and we run no laboratory and perform no testing of our own. This page deliberately publishes no numeric cross-sensitivity table: those values are specific to a cell chemistry, a model and a sensor age, and are published by the manufacturer. Presenting a generalised table would cause precisely the misapplication this guide warns against. WC Safety earns Amazon affiliate commissions on outbound clicks; that does not influence the content of this guide.
Authored by Steven Eaton, WC Safety Editorial — industrial gas detection desk · specialization: electrochemical and catalytic sensor behaviour, portable multi-gas instrument interpretation, and confined-space atmospheric testing.
Last reviewed: · Sources reviewed: OSHA 29 CFR 1910.146 (permit-required confined spaces), OSHA SHIB 09-30-2013 on calibrating and testing direct-reading portable gas monitors, NIOSH direct-reading and sensor technologies guidance, NIOSH Pocket Guide to Chemical Hazards, and manufacturer sensor datasheets and cross-sensitivity documentation.
Editorial standard: Zero sponsored listings. No manufacturer input. No paid placement on this page. No cross-sensitivity value is reproduced here that a manufacturer has not published for a specific sensor, and no generalised interference table is offered as a substitute for one.
Last reviewed: · Sources reviewed: OSHA 29 CFR 1910.146 (permit-required confined spaces), OSHA SHIB 09-30-2013 on calibrating and testing direct-reading portable gas monitors, NIOSH direct-reading and sensor technologies guidance, NIOSH Pocket Guide to Chemical Hazards, and manufacturer sensor datasheets and cross-sensitivity documentation.
Editorial standard: Zero sponsored listings. No manufacturer input. No paid placement on this page. No cross-sensitivity value is reproduced here that a manufacturer has not published for a specific sensor, and no generalised interference table is offered as a substitute for one.
How this guide was researched. Primary sources consulted directly:
- 29 CFR 1910.146 — the entry decisions that depend on a correct reading.
- OSHA SHIB 09-30-2013 — calibration, calibration check and bump test definitions for direct-reading portable gas monitors.
- NIOSH Direct Reading and Sensor Technologies — how direct-reading instruments are used to evaluate hazardous conditions.
- NIOSH Pocket Guide to Chemical Hazards — substance properties and exposure limits for the gases named here.
- OSHA annotated PEL tables — the exposure values an interfered-with reading would misrepresent.
Disclosure. WC Safety participates in the Amazon Services LLC Associates Program and earns from qualifying purchases. We accept no sponsored placements and no manufacturer payment for coverage. This guide is general reference information about sensor behaviour — it is not medical, legal or regulatory advice, and it does not replace your employer's written confined-space or exposure-monitoring program, your instrument's documentation, or the manufacturer's cross-sensitivity data for your specific sensor. For a commercial monitoring program, sensor selection and alarm response procedures should be reviewed by a Certified Industrial Hygienist.
Leave a comment