When All PM2.5 Isn’t the Same: Why Particle Size Distribution Matters for Optical PM Sensors

Not all particulate matter is created equal. Two environments can show the same PM_2.5 value on a monitor, yet be composed of completely different types of particles. One might be mostly fine soot, the other made up of coarse dust. To a typical user, these differences are invisible, but to the sensor inside your monitor, they matter a great deal.

At AirGradient, we believe understanding these nuances is essential, especially when it comes to evaluating how accurate or meaningful an air quality monitor truly is.

Incense Smoke vs Ambient Air

To investigate this, we compared the particle size distributions of two pollution sources:

• Incense smoke: used for controlled sensor testing and calibration due to its reproducibility.

• Ambient air: representing the complex mix of pollutants we encounter daily outdoors.

The results, measured with a reference-grade instrument, showed a clear difference in the mass distribution by particle size:

• Incense emissions were heavily weighted toward very fine particles, mostly between 0.1 to 0.3 µm.

• Ambient air, on the other hand, showed a broader and more variable size distribution across the full 0.1–2.5 µm range.

Even though both scenarios could register the same PM_2.5 concentration, the actual particle profiles are very different - and probably also the associated health risks.

Why This Matters for Calibration and Accuracy

High-end optical FEM monitors use single‑particle counting: each particle passing through the laser beam produces a scattering pulse, and the intensity of that pulse is used, after calibration, to assign the particle to a size bin. At high loads they correct for “coincidence” (two or more particles in the beam at once) by using design features and correction algorithms to minimize the effect.

In contrast, low‑cost optical PM sensors convert a single total‑scattering signal into mass using an assumed size distribution. This means they cannot distinguish whether the light came from one large particle or many smaller ones. Their internal algorithms make assumptions about particle size distributions, so the reported PM_2.5 or PM₁₀ values can be biased if the true distribution is different. This mismatch between assumed and actual distributions introduces bias, particularly during incense, fresh wildfire smoke, or dust‑heavy events.

This creates a core challenge:

A low‑cost optical PM sensor calibrated using incense smoke might perform well in a lab. But in real-world ambient air, with its different particle mix, it may systematically over- or underestimate pollution levels.

This disconnect complicates any effort to compare monitors, interpret readings across regions, or evaluate accuracy based solely on lab testing.

Practical Implications

• If your calibration method uses ultrafine aerosols like incense, be aware it might work well in that setting, but could bias readings outdoors.

• Conversely, ambient air calibration may underperform in indoor or wildfire-dominated environments.

This is why it’s strongly recommended, where feasible, to calibrate sensors locally, in the same environment and conditions where they’ll be deployed.

Why We’re Sharing This: We’re Still Learning

Most low-cost air quality monitor manufacturers gloss over this issue. We think it’s too important to ignore.

AirGradient is built on a foundation of openness and transparency. We share these insights not to overwhelm, but to inform and to invite conversation. If you’re comparing monitors, conducting your own testing, or just trying to make sense of the numbers, understanding particle size differences is key.

This is a complex topic, and we’re still learning too. If you have questions, thoughts, or ideas on better calibration methods or testing protocols, we’d love to hear them.

Reach out anytime or join the discussion in our community forum.

The goal isn’t perfection but clarity. And it starts with understanding what our monitors are really measuring.