A CO₂ monitor measures how much infrared light at one specific wavelength (about 4.26 micrometers) is absorbed by the air in a small chamber. CO₂ molecules absorb that wavelength and almost nothing else in room air does, so the amount absorbed is a direct measure of the CO₂ concentration. This method is called NDIR. Devices that report “eCO₂” do not measure CO₂ at all; they estimate it from other gases.
Infrared absorption
Every gas molecule vibrates at frequencies set by its structure, and it absorbs infrared light at those frequencies. CO₂ has a strong absorption band at 4.26 µm, in a part of the spectrum where water vapor and the other common indoor gases are nearly transparent. That coincidence is what makes CO₂ easy to measure optically.
How an NDIR sensor is built
- An infrared source (a small lamp or LED) shines across a chamber that room air diffuses into.
- An optical filter in front of the detector passes only the 4.26 µm band.
- A detector measures how much of that band arrives.
- More CO₂ in the chamber means more absorption and less light at the detector. The electronics convert the drop into ppm using a calibration curve.
Better sensors add a reference channel at a wavelength CO₂ does not absorb, so lamp aging and dust affect both channels equally and cancel out. Cheaper single-channel sensors rely on periodic recalibration instead.
Photoacoustic sensors
A newer design, used in the Sensirion SCD4x modules found in several popular monitors, pulses the infrared source. Each pulse heats the CO₂ molecules slightly, causing a tiny pressure wave that a microphone detects. More CO₂, louder pulse. The physics is the same absorption; the detection is acoustic, which allows a much smaller package. Manufacturers describe these as photoacoustic NDIR, and for practical purposes they belong in the same trustworthy category.
Why “eCO₂” is not a measurement
Metal-oxide (MOX) gas sensors change resistance when volatile organic compounds land on a heated surface. They are cheap and respond to hundreds of compounds. Some products run an algorithm that converts that VOC signal into an “equivalent CO₂” figure, on the assumption that people are the main VOC source and therefore VOC tracks CO₂. Hand sanitizer, cooking, a new sofa or a glass of wine breaks that assumption completely, and the eCO₂ number jumps to 2,000 ppm while real CO₂ is unchanged. If a product does not say NDIR or photoacoustic, assume it is estimating. See NDIR vs metal-oxide sensors.
Compensation and calibration
Absorption depends slightly on pressure and temperature, so good monitors compensate using built-in sensors. Over months the optical path drifts, which is why monitors have automatic baseline correction (assuming the room reaches outdoor levels at some point each week) or a manual fresh-air calibration. See how to calibrate a CO₂ monitor.
What limits accuracy
Consumer NDIR monitors typically state ±(30 to 50 ppm) plus 3 to 5 percent of the reading. The main sources of error in practice are placement (breath within a meter), warm-up time, and a drifted baseline rather than the sensor itself. See CO₂ monitor accuracy explained.
Choosing one
Any of the NDIR or photoacoustic monitors in the best CO₂ monitors guide measures the gas directly. The differences between them are battery, display, connectivity and price, not the measurement principle.
Frequently Asked Questions
What does NDIR stand for?
Nondispersive infrared. "Nondispersive" means the sensor uses a filter to select one wavelength rather than a prism or grating to spread the spectrum.
Are photoacoustic sensors NDIR?
They use the same infrared absorption principle but detect it by the pressure pulse the absorbed energy creates rather than by measuring transmitted light. Manufacturers such as Sensirion describe them as photoacoustic NDIR.
How long does a CO₂ monitor take to respond?
Most consumer NDIR monitors report a new reading every few seconds to a few minutes and take one to three minutes to fully respond to a step change in concentration.