PM2.5, CO₂, and VOC Sensor Technologies: How They Work and How to Compare Them

Volume I  ·  May 2026  ·  849 words

An air quality monitor is only as good as the sensors inside it. A $50 monitor and a $300 monitor may both claim to measure PM2.5 and CO₂, but the underlying sensor technologies differ in ways that determine whether the readings are trustworthy. This article explains the physics of each sensor type — what it measures, what it cannot measure, and how to identify quality from a specification sheet.

Particulate Matter: Laser Scattering vs Infrared LED

Laser Particle Counting

A laser diode (typically 650–780 nm) illuminates a stream of air drawn through the sensor by a small fan. Particles passing through the beam scatter light. A photodetector at 90° to the beam measures the scattered intensity, which is proportional to particle size (Mie scattering theory). The sensor counts individual particles and bins them by size: PM1.0, PM2.5, PM4.0, PM10.

The critical component is the laser and optics assembly. Quality sensors (Plantower PMS5003, Sensirion SPS30, Honeywell HPMA115) use a focused laser with a controlled flow path to ensure each particle is counted once. Budget sensors use an unfocused IR LED with lower signal-to-noise ratio, producing coarser bins and higher minimum detection limits (typically 10 µg/m³ vs 1 µg/m³ for laser).

All optical PM sensors share a fundamental limitation: they cannot distinguish particle composition. A 2.5 µm water droplet, dust particle, and smoke particle produce the same scattering signal. Humidity compensation (heating the inlet or applying a software correction) reduces but does not eliminate false-positive readings from water droplets.

Infrared LED Sensors

IR LED sensors use the same scattering principle but with a broader, less-intense light source. The lower signal-to-noise ratio means:

• Cannot count individual particles — measures aggregate scattering intensity and converts to a mass concentration via an algorithm calibrated against a reference instrument. The algorithm assumes a specific particle size distribution and composition; if the actual aerosol differs (e.g., wildfire smoke vs. road dust), the reading is wrong.
• Higher minimum detection limit (~10 µg/m³ vs ~1 µg/m³ for laser). Cannot reliably measure clean indoor air (< 5 µg/m³).
• Coarse size discrimination — typically only PM2.5 and PM10, not PM1.0 or PM4.0.

IR LED sensors are adequate for detecting when PM levels are "high" vs "low" but not for quantifying absolute concentrations. They are the dominant technology in sub-$100 monitors and in combination CO₂/PM monitors where PM is a secondary feature.

Carbon Dioxide: NDIR (Non-Dispersive Infrared)

NDIR is the only sensor technology that measures CO₂ directly. A broadband IR source passes through the air sample in an optical cavity. An optical filter selects the 4.26 µm wavelength where CO₂ has a strong absorption line. A photodetector measures the transmitted intensity. CO₂ concentration is calculated from the ratio of transmitted to incident intensity via the Beer-Lambert law.

NDIR sensors are self-referencing: a second detector at a non-absorbing wavelength (typically 3.9 µm) corrects for source intensity drift and contamination of the optical surfaces. Quality NDIR sensors (SenseAir S8, Sensirion SCD30/SCD40) include this dual-channel design. Budget NDIR sensors use a single-channel design that drifts over weeks to months.

eCO₂: The Estimation Trap

Some monitors report "eCO₂" — equivalent CO₂ — calculated from a VOC sensor reading, not measured by an NDIR sensor. The algorithm assumes a fixed ratio of CO₂ to VOCs based on human occupancy (people exhale both CO₂ and VOCs). This assumption fails whenever VOC sources are not human — cooking, cleaning, off-gassing furniture, outdoor pollution. An eCO₂ reading in a room with a running laser printer may spike to 2,000 ppm while the actual CO₂ is 500 ppm. eCO₂ is an estimate, not a measurement. Monitors that rely on eCO₂ should be clearly identified — and generally avoided for applications where CO₂ matters.

VOCs: MOS Sensors

Metal Oxide Semiconductor (MOS) sensors detect VOCs by measuring the change in electrical resistance of a heated metal oxide film (typically tin dioxide, SnO₂) when VOC molecules adsorb to its surface. The sensor responds to a broad range of reducing gases — alcohols, aldehydes, ketones, hydrocarbons — but cannot distinguish between them. The output is a dimensionless "TVOC index" or a concentration estimate calibrated against isobutylene.

MOS sensors have two well-documented failure modes: baseline drift (resistance changes over weeks to months as the sensor surface oxidizes) and poisoning (silicone compounds from cleaning products permanently deactivate the sensor surface). A quality monitor includes an automatic baseline correction algorithm that assumes the lowest reading over a 24-hour period represents "clean air" and adjusts the baseline accordingly. This works in environments that regularly achieve clean air; in continuously polluted environments, the baseline drifts upward and the readings become progressively more optimistic.