Air Quality Gas Sensor NO2 O3 Outdoor: A Selection Guide

Last spring, a systems integrator named Priya stood beside a new ozone monitoring node she had just commissioned for a coastal city. The unit was reporting 120 parts per billion of O3 on a quiet Sunday. The reference station nine blocks away read 41 ppb. She had bought the sensor on a clean-looking datasheet and deployed it without a partner gas cell, and now a city council member was asking why the numbers did not match.

Her problem wasn’t a faulty sensor. It was a chemistry problem, and it’s the single most common mistake in air quality gas sensors for outdoor NO2 and O3 monitoring: an ozone cell also responds to nitrogen dioxide. That gap between a datasheet promise and field reality is what this guide is about.

Particulate matter gets most of the headlines, but the pollutants that breach limit values in dense cities are gases. Choosing air quality gas sensors for outdoor NO2 and O3 monitoring means reading a ppb specification critically, understanding cross-sensitivity, planning for a cell that ages, and calibrating on a schedule.

This guide is for station OEMs, systems integrators, and municipal air quality engineers. By the end, you’ll match the right sensing technology to each pollutant and build a data path your regulators will accept. For the full station picture, start with our outdoor IoT air quality monitoring station guide, then widen the lens with the environmental monitoring station smart city deployment guide.

Why Gases Need Different Sensors Than Particulate Matter

PM and gas monitoring look similar on a station diagram. They share nothing underneath.

A PM sensor is optical. It counts particles by the light they scatter.

A gas sensor is chemical. It converts a specific molecule into an electrical signal through a reaction, an absorption, or an ionization event. You can’t measure NO2 the way you measure dust, and mixing up the two families is the first procurement mistake.

An air quality gas sensor is a device that detects the concentration of a gaseous pollutant in ambient air, usually reporting parts per billion (ppb) or parts per million (ppm). The five targets in an outdoor station are NO2, O3, SO2, and CO, which are inorganic and measured individually, plus VOC, which is a family of compounds rather than a single molecule.

That last distinction matters. VOC monitoring isn’t a measurement; it’s an index. Treat it as a trend, not a compliance number.

For the particulate half of the station, our guide to PM2.5 sensor accuracy for outdoor IoT covers the optical side in depth.

The Four Air Quality Gas Sensor Technologies, Compared

Four technologies do nearly all the work in outdoor stations. Each wins on a different axis.

Electrochemical Sensors

An electrochemical cell holds a working electrode, a counter electrode, and often a third reference electrode behind a capillary diffusion barrier. Target gas molecules diffuse in and react at the working electrode, producing a small current proportional to concentration.

These cells are the workhorse of low-cost outdoor gas monitoring. A 2025 review of 65 outdoor NO2 studies found electrochemical sensors in 95% of them, against just 17% for metal-oxide alternatives. They are ppb-capable, low power, and inexpensive. Their weaknesses are drift, a limited life, and cross-sensitivity.

Metal-Oxide Semiconductor (MOS) Sensors

A MOS sensor heats a metal-oxide surface, often tin dioxide or tungsten trioxide, to 300 to 500°C. Target gas changes the surface resistance. These cells are tiny, cheap, and long-lived, and they respond to a wide range of gases.

The trade-offs are real. MOS sensors are sensitive to humidity and temperature, they need significant power to stay heated, and they are better at trends than absolute values. Most MOS parts now output a VOC or NOx index rather than a concentration, and that’s the honest way to use them.

NDIR Sensors

Non-dispersive infrared sensors measure how much infrared light a gas absorbs, following the Beer-Lambert law. They are precise, drift slowly, and can last 10 to 15 years, which makes them excellent for carbon monoxide and carbon dioxide. They can’t see homonuclear molecules like O2 or N2, so NDIR isn’t a universal answer.

Photoionization Detectors (PID)

A PID uses ultraviolet light to ionize gas molecules, producing a fast response with a very low detection limit, sometimes as low as 100 parts per trillion. For speciated VOC and BTEX work, it’s the tool of choice. It isn’t selective among VOCs, so it needs response factors and a reference gas. It’s also more expensive and power-hungry than cells.

Technology	Principle	Best Gases	Accuracy	Power	Lifespan	Cost Tier
Electrochemical	Redox at electrodes	NO2, O3, SO2, CO	ppb-capable	Very low	~2 years	Budget to mid
MOS	Resistance change when heated	VOC/NOx index, trends	Semi-quantitative	High (heated)	10+ years	Budget
NDIR	IR absorption	CO, CO2	High precision	Moderate	10-15 years	Mid to premium
PID	UV ionization	VOC, BTEX	ppt-capable	Higher	Lamp-dependent	Premium

One caveat before you compare cells to a reference station: the instruments regulators trust don’t use any of these low-cost parts. Reference-grade O3 is measured by UV photometry, SO2 by UV fluorescence, NOx by chemiluminescence, and CO by NDIR. Those are the methods your cells are being validated against, and knowing the gap keeps expectations honest.

Matching the Sensor to the Pollutant

Not every pollutant needs the same sensor, and not every cheap part can do the job. Here is the map.

Pollutant	Typical Ambient Range	Resolution Needed	Reference Method	Key Interference
NO2	0-2,000 ppb	~1-10 ppb	Chemiluminescence (EN 14211)	O3, H2S
O3	0-500 ppb	~1-10 ppb	UV photometry (EN 14625)	NO2
SO2	Low ambient ppb	~1-10 ppb	UV fluorescence (EN 14212)	NO, Cl2, H2S
CO	0-50 ppm	~0.1 ppm	NDIR (EN 14626)	H2, VOC (cross-response)
VOC	Variable	Trend only	None	Everything (non-selective)

The most important column is resolution, and most buyers skip it.

The Resolution Gate

A sensor that resolves only 50 to 100 ppb can’t tell you anything useful about ambient NO2, where limit values sit far lower. This isn’t a theoretical concern. Public sensor reviews flag the Mics 2714 as unable to detect NO2 below about 50 ppb, which corresponds to an Air Quality Index of 50. Two other common parts, the Honeywell 4NO2-20 at roughly 100 ppb resolution and the DD GS+4NO2 at plus or minus 200 ppb, are likewise unsuitable for ambient NO2.

Before you buy, convert the limit values you must report into a required resolution, then reject every part that can’t reach it. A cheaper cell that can’t resolve the pollutant isn’t a saving. It’s a data gap with a sensor attached.

Why NO2 and O3 Must Be Measured Together

Ozone is a significant interferent for NO2 sensors. Among the studies in that 2025 review, 79% of NO2 projects also measured O3, precisely because the two gases interact in both the atmosphere and the sensor. If you plan to report both pollutants, plan to measure them together with a dedicated cell for each.

NO2 and O3 Cross-Sensitivity in Outdoor Gas Sensors

Why does an ozone sensor need a nitrogen dioxide correction? Because an oxidising electrochemical cell responds to both gases. The standard fix is to measure total oxidants, called Ox, with one cell and NO2 with a second, then compute O3 as Ox minus NO2.

This is the number one source of bad gas data outdoors, and it is worth understanding the failure modes.

Documented Interferences

Cross-sensitivity is quantifiable, and vendors publish the numbers in their datasheets. The Alphasense OX-B431 cell, for example, responds to hydrogen sulphide at up to 170% of its signal. It also reacts to chlorine at up to 90% and NO at under 5%, while sulphur dioxide runs slightly negative.

A SO2 cell can show a strong response to NO, and a CO cell can react to hydrogen and some VOCs. None of these cells is truly blind to its neighbours.

Correction Approaches

There are three ways to handle interference, and they stack.

• Paired cells. Measure Ox and NO2 separately, then subtract. This is how most ozone networks correct their readings.
• Selective filters. Chemical filters remove specific interferents before the sample reaches the cell, but they add maintenance and can themselves drift.
• Multi-sensor compensation. Regression and machine-learning models trained on co-located data untangle multiple gases at once. Machine learning was the most common calibration approach in the 2025 NO2 review, ahead of simple linear regression.

One warning worth repeating: naive interference correction can make results worse. In a year-long field study, correcting an ozone cell for NO2 interference increased the error through propagation, because the NO2 cell’s own error carried into the subtraction. Correction only helps when the partner cell is good enough to trust.

Newer networks push this compensation onto the device itself. Edge models can correct a reading before it leaves the pole, which is where edge computing for real-time environmental monitoring earns its place.

What Really Limits Accuracy Outdoors

Cross-sensitivity gets the attention, but temperature, humidity, drift, and ageing quietly do more damage over a multi-year contract.

Temperature and Humidity

Electrochemical gas sensors are calibrated at a comfortable reference point, usually around 20°C and 60% relative humidity. Move away from that, and errors grow.

A peer-reviewed field evaluation published in Atmospheric Measurement Techniques found that errors increase and correlation drops as temperature climbs from below 10°C to above 30°C. Accuracy also falls as humidity drops from above 75% to below 30%. The likely cause is electrolyte drying at low humidity.

The same study found the extreme case is the damaging one. Prolonged exposure above 40°C and below 10% relative humidity produced irreversible drift of more than 50% per year, more than double what the manufacturer specified. Temperature and humidity compensation, a hydrophobic barrier, and condensation control aren’t optional outdoors.

Drift and Realistic Lifespan

Vendor lifespan figures and field reality diverge sharply. The Alphasense NO2-B43F carries a two-year warranted life, but it has been reported to malfunction after as little as 200 days in the field. At the other end, the MOS Sensirion SGP41 specifies more than 10 years with on-chip drift compensation, though it reports an index rather than a concentration.

The lesson isn’t that one technology wins. It’s that warranted life isn’t expected field life, and your maintenance budget should assume the shorter of the two.

Calibration Reality

One-time calibration doesn’t hold. A year-long co-location study in Augsburg, Germany concluded that monthly recalibration was the most suitable interval for electrochemical cells, and that performance varies by season with temperature and humidity. Vendors agree on the direction even where they differ on the number: Libelium rates its cells at maximum accuracy for about six months, and Winsen advises bump tests weekly to monthly with full calibration every three to six months.

The deep co-location protocol, including how long to run it and how to validate the fit, belongs in its own guide. For the post-deployment maintenance calendar, our smart city sensor maintenance and calibration guide covers the schedule.

Gas Sensor Modules for Outdoor Stations Compared

If you are specifying hardware, this table is your shortlist. Treat every brand the same way and evaluate against your resolution requirement first.

Module / Family	Gas	Principle	Strength	Watch Out For
Alphasense B43F / B4 series	NO2, O3/OX, SO2, CO	Electrochemical, 4-electrode	Best-validated low-cost NO2 cell; ppb-capable	Needs recurring calibration; O3 needs NO2 partner
Winsen MICS / ME series	NO2, O3, SO2, CO	Electrochemical	Low cost, wide availability	Wider unit-to-unit spread; verify resolution
Honeywell AQ7 Series	NO2, O3, SO2, CO	Electrochemical	Individual temperature and cross-sensitivity compensation	Higher cost than bare cells
Sensirion SGP41	VOC, NOx index	MOS	10+ year life, on-chip drift compensation	Index output, not ppb; needs clean-air exposure
PID modules	VOC, BTEX	UV ionization	ppt-level, fast, stable baseline	Non-selective; expensive; response factors

The Alphasense family has the deepest independent validation record. In the year-long Nicosia study, its NO2 cell was the only low-cost sensor tested that qualified for European Directive indicative measurements, with relative expanded uncertainty near 25% above about 70 ppb. Its ozone cell, notably, performed badly without correction, which is exactly the point.

The Winsen cells cost less but varied more in the field, with one ozone cell under-reporting and a NO2 cell overestimating by roughly double the Alphasense error. Honeywell bundles individual compensation into the AQ7 family, which raises the price but lowers integration work. The Sensirion SGP41 suits low-power nodes that need a robust VOC and NOx trend, not a compliance number.

When a Cell Is Not Enough

Match the instrument class to the job. For supplemental or indicative monitoring, a corrected electrochemical cell is fine. For fenceline compliance or regulatory determinations, you need a Federal Reference or Equivalent Method analyser, and no correction turns a low-cost cell into one.

How to Choose Gas Sensors for an Outdoor Air Quality Station

Here is a six-step selection process you can take into a procurement meeting.

1. List the pollutants your limits or AQI reporting require. If NO2 is on the list, put O3 on it too, with a dedicated cell for each.
2. Convert limit values into a required resolution. Work in ppb, not marketing language, and set a hard minimum before you look at parts.
3. Screen candidate cells against that resolution. Reject anything that cannot resolve ambient concentrations, regardless of price.
4. Map the site’s interferents. Know whether H2S, Cl2, or NOx are present, then choose cells, filters, and paired measurements accordingly.
5. Confirm environmental protection. Insist on temperature and humidity compensation plus a hydrophobic barrier, and if you’re choosing radio, our comparison of NB-IoT vs LoRaWAN for smart city sensors covers the uplink side.
6. Set the calibration and replacement plan before deployment. Decide the interval, the co-location method, and the spare-cell budget now, not after drift appears.

Then decide station versus module. OEMs building a product should buy cells and integrate. Integrators and cities deploying many nodes usually move faster with a complete station that already handles signal conditioning, compensation, and enclosure design.

That’s the approach we take at Black Ant. Our outdoor environmental monitoring station combines air quality, temperature, humidity, wind speed, noise, and particulate sensing in an IP65 enclosure with NB-IoT or 4G uplink and smart city platform compatibility. Because the sensing, enclosure, and data path are designed together, environmental compensation and the sample path are handled before the unit ships. Tell our engineering team which gases you need to report and your climate, and we will confirm the right gas-sensing configuration before you commit.

Standards That Make Gas Data Defensible

When you present gas data, you will be asked how it maps to the official frameworks. Keep the map ready.

• United States. FRM and FEM instruments for SO2, CO, O3, and NO2 sit under 40 CFR Part 53, which defines the testing and comparability procedures for these pollutants. Designated methods are published on the EPA’s reference and equivalent methods list.
• European Union. Continuous analysers follow EN 14211 for NOx, EN 14212 for SO2, EN 14625 for O3, and EN 14626 for CO. Lower-cost sensors are addressed by CEN/TS 17660, and the revised Ambient Air Quality Directive is tightening limits toward WHO levels.
• World Health Organization. The WHO 2021 air quality guidelines set the benchmarks most readers compare against. NO2: 10 µg/m³ annual and 25 µg/m³ for 24 hours. O3: 60 µg/m³ peak season and 100 µg/m³ over 8 hours. SO2: 40 µg/m³ over 24 hours. CO: 4 mg/m³ over 24 hours.

Put your corrected reading next to the relevant framework and state plainly whether your network is a reference method or an indicative supplement. That single sentence prevents most disputes.

Frequently Asked Questions

How does an electrochemical gas sensor work?

Target gas diffuses through a capillary barrier into an electrolyte and reacts at a working electrode, driving a small current proportional to concentration. A counter electrode completes the circuit, and a reference electrode, when present, keeps the baseline stable.

Why does my ozone sensor read too high?

An oxidising electrochemical cell responds to both O3 and NO2, so a raw ozone reading includes the NO2 present. The standard fix is to measure total oxidants and NO2 separately and compute O3 as Ox minus NO2, using a well-calibrated partner cell.

How long do electrochemical gas sensors last outdoors?

A common warranted life is about two years, but field failures have been reported from roughly 200 days. Plan for recalibration every few months and replacement well inside the warranty, especially in hot, dry, or polluted sites.

What is the difference between electrochemical, MOS, NDIR, and PID sensors?

Electrochemical cells use a redox reaction and suit NO2, O3, SO2, and CO. MOS sensors change resistance when heated and suit VOC and NOx trends. NDIR measures infrared absorption and suits CO and CO2. PID uses UV ionization and suits speciated VOC and BTEX.

Which gas sensors do I need for an outdoor air quality station?

It depends on what you must report. Most stations need NO2, O3, SO2, and CO cells with adequate ppb resolution, always pairing NO2 and O3. Add VOC sensing only if you need a trend, and expect a higher-grade analyser for any compliance role.

Can low-cost gas sensors be used for regulatory monitoring?

Not on their own. FRM and FEM analysers are required for legal determinations. Low-cost cells serve as dense, corrected, indicative supplements around a small set of reference instruments.

How often do gas sensors need calibration?

More often than most buyers expect. Field studies point to monthly recalibration for demanding campaigns, and vendors advise bump tests weekly to monthly with full calibration every three to six months. A one-time factory calibration degrades within months outdoors.

Getting Outdoor Gas Monitoring Right

Return to Priya and her two numbers. Her ozone cell wasn’t broken. It was doing exactly what an uncorrected oxidising cell does: reporting O3 and NO2 together, without a partner sensor to tell them apart.

The fix was never a more expensive cell in isolation. It was a parameter matched to the limit value, a paired NO2 cell for the Ox correction, environmental compensation that matched the coastal climate, and a calibration interval set before deployment.

That’s the real lesson on air quality gas sensors for outdoor NO2 and O3 monitoring. Accuracy isn’t a datasheet line. It’s a property of the whole system, sensor chemistry, interference handling, environmental compensation, and a maintenance plan working together. Get the system right, and low-cost cells produce data you can defend to a council, a regulator, or a client.

When you are ready to move from modules to a deployed station, our engineering team will help you specify the right gas-sensing configuration for the pollutants you must report. Explore the Black Ant environmental monitoring station range or request the datasheet, and we will respond within one business day.