ISO 16890 Explained: Why Air Filter Labels All Changed After 2018

What does "ePM1 70%" on your air purifier actually mean? And why did every filter label change after 2018?

Where Did EN 779 Fall Short?

From 1993, Europe graded general ventilation filters using EN 779: G3/G4 (coarse), M5/M6 (medium), F7/F8/F9 (fine).

The problem? All grading was based on efficiency at a single particle size — 0.4 μm.

Real air pollution isn't just 0.4 μm. WHO and national EPAs track PM1 / PM2.5 / PM10, and 0.4 μm single-point efficiency does not translate directly to those. A designer holding an F7 filter couldn't answer "how much PM2.5 does this remove?" with a single lookup.

So in 2016, ISO 16890 was published; by 2018 it had fully replaced EN 779. The new standard maps directly to PM indices, so engineers can pick filters based on actual air-quality targets.

What Are ePM1 / ePM2.5 / ePM10?

ISO 16890 defines three particle-size bands:

• ▸ePM1: 0.3–1.0 μm — aligns with PM1 (the most dangerous fraction, reaches lung alveoli)
• ▸ePM2.5: 0.3–2.5 μm — aligns with PM2.5 (respirable, reaches lungs)
• ▸ePM10: 0.3–10 μm — aligns with PM10 (reaches upper airways)

All three start at 0.3 μm because that's the ISO 16890 test lower bound. The difference is the upper bound.

Why is PM1 the priority? Particles under 1 μm can bypass the bronchi, reach alveoli, even enter the bloodstream. WHO is moving toward treating PM1 as a distinct monitored pollutant.

Why "Average" the Test Result? The Three-Step Flow

The biggest reform over EN 779: ISO 16890 folds post-loading efficiency into the rating.

Why Test This Way?

Many medium-grade filters lean heavily on electrostatic charge for efficiency. Brand new, they show beautiful capture numbers. But after a short service life, that charge dissipates (dust build-up, humidity), and efficiency can fall from 90 % to 60 %.

EN 779 measured only the initial number — catalogs looked great, real-world performance did not match.

ISO 16890 takes the arithmetic mean of both, which tracks reality more honestly:

ePM = (initial efficiency + post-loading efficiency) / 2

If the average is below 50 %, that grade isn't designated. It's a hard threshold — filters below 50 % don't get to claim the class.

Decoding "ISO 16890 - ISO ePM1 65 %"

A real filter label:

ISO 16890 - ISO ePM1 65 %

Where does it come from? Suppose measurements give:

• ▸ePM1 average = 65 % (≥ 50 %, passes)
• ▸ePM2.5 average = 82 %
• ▸ePM10 average = 95 %

The classification rule searches from smallest band upward, picking the first passing band:

1. 1Does ePM1 ≥ 50 % ? → Yes, so the filter is classed ePM1
2. 2If ePM1 failed, drop to ePM2.5
3. 3If ePM2.5 also failed, drop to ePM10

Result: ISO ePM1 65 % — "ePM1" says it handles the finest band, "65 %" is the actual efficiency.

Which Grade Fits Your Space?

Each use case has a minimum recommended grade — both over- and under-spec are waste:

Common scenarios:

• ▸General office / residence — ISO ePM2.5 ≥ 50 % upward, captures most respirable particles at reasonable energy cost
• ▸Hospitals / schools — ISO ePM1 ≥ 50 % or higher; fine particles reach deep lung tissue
• ▸Pre-filter before HEPA — ISO ePM1 ≥ 80 %, extends HEPA service life and cuts long-run cost
• ▸Industrial / underground carpark — ISO ePM10 ≥ 50 % suffices, dominant pollutant is coarse

In heavily polluted areas (annual PM2.5 > 25 μg/m³), step up one grade. For high-static processes or spaces serving infants, default to ePM1 ≥ 50 % minimum.

Three Common Misconceptions

Misconception 1: "ePM10 Is Stronger Than ePM1"

The opposite. ePM1 handles everything ePM10 does, plus more. ePM1 is the strictest — it must handle the smallest, hardest fraction. ePM10 only commits to handling up to 10 μm coarse particles.

Misconception 2: "65 % Average Sounds Worse Than EN 779's F7"

You can't directly compare. EN 779 F7 is ~80 % at 0.4 μm *initial* — after a year the same filter might be at 60 %. ISO 16890's 65 % is the average of initial and loaded, closer to real service-life performance. Don't compare old-standard initial against new-standard average.

Misconception 3: "Higher Is Always Better — Just Pick ePM1 95 %"

Over-spec costs you three ways: higher purchase price, higher pressure drop (more electricity), and earlier clogging (shorter life). Pick based on your air-quality target, not on the highest number.

ISO 16890 Is Not Just a Label Change — It's Filters Finally Aligned With Public Health

From single-point to PM-mapped, from initial-only to loaded-average, ISO 16890 does one thing: match filter labels to the air we actually breathe.

Before picking a filter, nail down which PM index you're targeting and the concentration you need — then the grade falls out. Both over-spec and under-spec are waste.