How Do You Interface a Gas or CO2 Sensor to a Microcontroller?

Detailed Explanation

"Gas sensor" and "CO2 sensor" are used loosely across consumer and industrial products to describe three genuinely different measurement technologies, each with its own accuracy characteristics, calibration burden, and power profile. Choosing the wrong one for the application — most commonly, using a VOC sensor's estimated CO2 figure where an actual measured CO2 value is required — is a common and avoidable design mistake.

NDIR (Non-Dispersive Infrared) CO2 Sensors

An NDIR sensor shines infrared light through a small gas-sample chamber onto a detector. CO2 absorbs infrared light strongly at a specific wavelength (around 4.26 µm); the detector measures how much light at that wavelength is absorbed, and the sensor's internal signal processing converts that absorption into a CO2 concentration. Because the measurement is based on a specific physical property of the CO2 molecule itself, NDIR is the only one of the three technologies covered here that measures CO2 directly rather than inferring it.

Integrated NDIR modules such as the Sensirion SCD40/SCD41 or Sensair S8 combine the optical bench, detector, and signal processing into a single package that reports a calibrated ppm value directly over I2C (SCD4x) or UART (S8) — no analog front-end design is required, in contrast to older discrete NDIR modules. The main design considerations are optical: the sensor typically requires an open path or vent to ambient air (a fully sealed enclosure prevents the gas sample chamber from equilibrating with the surrounding air), and the automatic self-calibration behaviour covered in the FAQ above needs to be understood for the specific deployment environment.

MOx (Metal-Oxide Semiconductor) VOC/Air-Quality Sensors

A MOx sensor uses a heated metal-oxide sensing element (commonly tin dioxide) whose electrical resistance changes when reducing gases and volatile organic compounds adsorb onto its surface at operating temperature. This makes MOx sensors broadly sensitive to a wide range of VOCs, combustible gases, and general "air freshness" — but not specifically selective to any one gas, and largely insensitive to CO2 itself.

Because MOx sensors don't measure CO2 directly, integrated parts like the Bosch BME680 and Sensirion SGP30/SGP41 report either a relative index (IAQ — Indoor Air Quality index) or a modelled "equivalent CO2" (eCO2) estimate derived from the raw VOC signal by an onboard or vendor-supplied algorithm — see the FAQ above for why this estimate can diverge from true CO2 concentration. MOx sensors also draw meaningfully more power than NDIR or electrochemical alternatives, because the sensing element's heater typically needs tens to over 100 mW continuously (or during each duty-cycled measurement) to reach and hold its operating temperature — a significant factor for battery-powered designs.

Electrochemical Gas Sensors

An electrochemical sensor uses a chemical reaction at a sensing electrode, specific to a target gas, to produce a small current proportional to that gas's concentration. This makes electrochemical sensors highly selective — a CO sensor responds to carbon monoxide specifically, not to a broad mix of VOCs — which is why this technology dominates safety-critical single-gas detection: CO detectors, O2 depletion monitors, and toxic industrial gas detectors (H2S, NH3, Cl2, and others, each requiring its own sensor type).

Electrochemical sensors have two design considerations the other two technologies don't: a finite operating lifetime (commonly one to a few years, driven by consumption of the internal electrolyte and electrodes, so safety-critical products must plan for sensor replacement or an end-of-life self-test) and, on many parts, a small output current requiring a dedicated transimpedance amplifier stage rather than a ready-made digital interface — see transimpedance amplifier design for the same feedback-resistor sizing principle applied to a different sensor type. Some newer electrochemical modules integrate this analog front end and expose a digital (I2C) output directly, following the same integration trend seen in MEMS barometric pressure sensors.

Comparison

Practical Examples

A demand-controlled ventilation system that opens a damper when a meeting room's CO2 rises above a set threshold needs an NDIR sensor — an MOx-derived eCO2 estimate is not a defensible basis for a ventilation control loop, since it can diverge from true CO2 concentration depending on what else is in the air. Reading the SCD41 over I2C at a periodic interval (commonly every few seconds to match the sensor's own internal measurement cycle) and comparing against a configured threshold is a straightforward integration.

A consumer indoor air-quality display, by contrast, is a good fit for an MOx-based sensor — the product's goal is a relative "how stuffy is this room" trend indicator, not a certified CO2 measurement, and the lower cost and smaller footprint of an integrated MOx part like the SGP41 typically outweighs the accuracy limitation for that use case.

A portable gas detector for confined-space entry work needs an electrochemical O2 sensor (to detect oxygen depletion) potentially alongside an electrochemical CO or H2S sensor, each independently selected and calibrated for its specific target gas — this class of product typically has regulatory certification requirements (for example, relevant workplace gas-detection standards) that should be confirmed against the specific application and jurisdiction before a sensor and product design are finalised.

Design Considerations

• Provide an air path to the sensor. All three technologies need the gas sample chamber or sensing element to actually exchange air with the environment being measured. A sensor mounted inside a fully sealed enclosure, or behind a dense, unvented mesh, will read whatever gas mixture is trapped inside the enclosure rather than the ambient environment.
• Budget power for the specific technology's duty cycle. An NDIR sensor's periodic infrared source pulse and a MOx sensor's continuously (or near-continuously) heated element are both meaningful current draws in a battery-powered design — check the specific part's datasheet for both active measurement current and any lower-power duty-cycled mode it supports, and plan the measurement interval around the product's battery life target.
• Design in a path to recalibration or replacement. NDIR sensors depend on periodic exposure to a known reference (see the FAQ above); MOx sensors' baseline drifts over the sensor's life and some parts benefit from periodic recalibration against a reference; electrochemical sensors have a hard end-of-life. A product with a multi-year expected lifetime should account for whichever of these applies to the chosen technology, rather than assuming a "set and forget" sensor.
• Don't conflate air-quality index with a gas-specific measurement in product claims or firmware logic. A control loop, alarm threshold, or user-facing claim that assumes a MOx-derived eCO2 or IAQ figure is a true, absolute CO2 or VOC concentration is building on an estimate, not a measurement — state the actual technology and its limitations clearly where the distinction matters to the product's function or claims.

Zeus Design's engineering team selects and integrates gas, air-quality, and environmental sensing — including analog front-end design for sensors without an integrated digital interface — as part of complete product development.

Common Mistakes

• Choosing a MOx/VOC sensor for an application that actually needs measured CO2. As covered above, this is the single most common gas-sensor selection mistake — the eCO2 figure a MOx sensor reports is a modelled estimate, not a direct CO2 measurement, and can diverge from reality in ways that matter for ventilation control or compliance claims.
• Sealing the sensor inside an enclosure with no air path. A gas sensor measures the air immediately around its sensing element — if that air can't exchange with the environment being monitored, the reading reflects a trapped microclimate, not the actual space.
• Reading the sensor before its warm-up period completes. As covered in the FAQ above, all three technologies specify a warm-up time after power-on; a reading taken before that time elapses is unreliable, and for MOx sensors specifically, the very first power-on after installation requires a much longer initial burn-in than every subsequent power-on.
• Ignoring electrochemical sensor end-of-life in a safety-critical product. Unlike NDIR and MOx sensing elements, an electrochemical cell is a consumable with a finite lifetime — a product relying on one for safety-critical gas detection needs a defined replacement interval or an active end-of-life self-test, not an assumption that the sensor will keep working indefinitely.