You feel it at the junction. You turn your head to check for a bus, and the top edge of the mask lifts off your cheekbone — a half-second of cool air arriving somewhere it shouldn't. Then you straighten up and it settles back. On the streets of Ho Chi Minh City, researchers measured what that half-second costs: the best-performing mask types removed 60–80% of traffic particles while riders were actually on motorbikes, against filter media rated well above 95% on a bench. The gap between those two numbers is not a manufacturing defect. It is the whole story of why a pollution mask for motorcycle riding has a ceiling, and why no better filter will lift it.
The short version
- Measured on riders on real motorbikes in Ho Chi Minh City traffic, N95, valved-filtering and carbon-layer masks removed 60–80% of particles, while surgical and cloth masks removed 25–60% (Environmental Science: Atmospheres, 2022).
- The study's authors attribute the shortfall to fit, not filtration: "no mask can completely remove all particles under practical conditions. It is largely due to inappropriate mask fitting."
- Only 14% of the world's cities met the WHO annual PM2.5 guideline in 2025, down from 17% the year before, across 9,446 cities in 143 countries (IQAir World Air Quality Report).
- Three teams on three continents — India in 2022, Spain in 2025, and Easi Breezi — arrived at the same answer at a 17× price spread: filter and move the air before it reaches the face.
How well does a pollution mask actually work on a motorcycle?
Between 25% and 80%, depending on the mask — and that upper figure belongs to respirator-grade masks, not the fabric ones most riders wear. Those numbers come from a study published in Environmental Science: Atmospheres in November 2022, which is unusual in a useful way: instead of testing masks on a bench or a mannequin in a lab, the researchers measured particle concentrations on riders moving through real Ho Chi Minh City traffic.
N95 respirators, reusable valved filtering masks and locally made carbon-layer masks landed in a band of 60–80% removal. Surgical masks and cloth masks landed at 25–60%. Every one of those masks carries filter media capable of far more than that in a static test.
Read that carefully, because the popular version of this finding is wrong in both directions. Masks are not useless — the authors are explicit that any mask provides some protection, and a rider in an N95 is meaningfully better off than a rider in nothing. But a mask is also not the 95%-plus device its packaging implies. It is capped. And the cap sits in a place most riders never think about.
It was never the filter. It was always the seal
The study's own explanation is one sentence long: "no mask can completely remove all particles under practical conditions. It is largely due to inappropriate mask fitting."
Here is why that sentence matters more than any filtration rating.
A mask is a pressure-driven device with an unsealed boundary. Every time you inhale, you drop the pressure inside the mask, and air rushes in to equalise. It does not politely queue at the filter. It takes the path of least resistance, and it takes all of them at once, in proportion to how easy each one is.
Filter media is deliberately high-resistance — that resistance is how it works. A one-millimetre gap at the cheekbone is close to zero resistance. So a large share of each breath arrives through the gap, completely unfiltered, no matter what the media is rated for. Upgrading the filter from 95% to 99.97% does nothing to a leak, because the leak was never going through the filter.
Now add riding. Every condition that breaks a face seal is present at once: you rotate your head at every junction, the helmet's cheek pads press the mask out of plane, the chin strap tugs it, your jaw moves, and at 60 km/h there is dynamic pressure across your whole face that simply is not there in a fit-test booth. Seals also degrade with time on the face — a prospective cohort study of emergency-department clinicians found N95s failing fit testing in 38.7% of cases after a single shift, with failure rates climbing the longer a mask was worn. A commuter's mask is on for an hour, twice a day, in heat, under a helmet, getting damp.
Where the air actually goesSCHEMATIC
Illustrative diagram of the mechanism — not measured data
That inversion is the entire engineering argument. Stop asking a leaky seal to hold dirty air back. Instead, push filtered air toward the face faster than the rider can breathe it in. The gap stops being an inlet and becomes an exhaust. Fit no longer sets the ceiling, because there is nothing for the gap to let in.
Four years, three helmets, one unsolved problem
The interesting thing about that conclusion is that several groups reached it independently, and none of them talked to each other.
In 2022, a Delhi startup launched Puros — a full-face helmet with an intake at the rear, a battery-powered fan and an H13-grade HEPA filter, claiming up to 80% reduction in pollutant exposure. It sold for Rs 4,500, roughly US$56, and was backed by India's Department of Science & Technology.
In July 2025, a Spanish startup announced Zyon: a modular helmet with a patented four-layer filter in the chin guard — debris, then fine particles, then activated carbon for gases, then a protective backing — with three airflow modes and an 18-hour battery. It sold at US$935 and took more than 4,000 pre-orders.
Neither of these is news. That is the point. Across four years, at a 17× price spread, on two continents, teams working in isolation converged on the same architecture: powered, filtered, delivered. Nobody tried to build a better mask, because the people closest to the problem could see that the mask category had already found its ceiling.
What happens by 2031?
The pressure behind all of this is not easing. Only 14% of the world's cities met the WHO annual PM2.5 guideline in 2025 — down from 17% in 2024 — across a dataset of 9,446 cities in 143 countries (IQAir World Air Quality Report).
Share of world cities meeting the WHO annual PM2.5 guideline
Two measured years, and what their rate of change implies
| 2024 (measured) | 17 percent |
|---|---|
| 2025 (measured) | 14 percent |
| 2031 (extrapolated) | about 4.4 percent |
That 2031 figure is arithmetic, not prophecy, and it deserves to be treated as such. A drop from 17% to 14% is a relative decline of about 17.6% in one year. Hold that rate for six more years and you get 14% × (1 − 0.176)6 ≈ 4.4% — fewer than one city in twenty. But it rests on exactly two data points. One unusually mild wildfire season would break it, and IQAir specifically flags wildfire activity as what hit historically cleaner regions in 2025. Treat it as what the current rate implies if nothing changes, and nothing more.
The market forecasts are no more settled. Anti-pollution masks are projected to grow at 14.3% annually from 2025 to 2031 by one analyst — while published estimates elsewhere range from roughly 6% to over 20%. That spread is itself the finding. Analysts cannot agree on the growth rate of a category that is quietly being redefined underneath them. A market that is half disposable fabric and half powered filtration does not have a single number.
What actually moves the number
If you accept that the limit is mechanical, the shortlist of real options gets short.
Easi Breezi is the clip-on version of the physics Puros and Zyon build into the shell. A powered unit draws air through an H11 filter rated ≥95% on PM2.5 and delivers it into the helmet, so the rider is supplied rather than sealed. It is IP67 rated and patent pending. Because it mounts to the helmet you already own and trust, it does not ask you to re-buy certified head protection in order to get filtration — which is the trade every integrated helmet in this category asks you to make. It is US$199.
Frequently asked questions
Is a pollution mask for motorcycle riding worth wearing at all?
Yes. The Ho Chi Minh City study is explicit that any mask offers some protection, and respirator-grade masks reached 60–80% removal on real rides. If a mask is what you have, wear it — just don't assume you are getting the number printed on the box.
Why doesn't a better filter fix the problem?
Because the air that harms you mostly isn't passing through the filter. It enters through gaps around the mask edge, where resistance is near zero. Improving media from 95% to 99.97% changes the filtered fraction only, and leaves the leak untouched.
Does wearing a mask under a full-face helmet help?
It complicates the seal rather than improving it. The helmet's cheek pads press the mask out of shape, and exhaled moisture has nowhere to escape — which is why fogging is a design concern for helmet makers working in this space.
What is the difference between a mask and a powered helmet filter?
Direction of flow. A mask asks a seal to hold dirty air out while you inhale against it. A powered filter pushes clean air in faster than you breathe it, so leaks vent outward instead of letting pollution in.
How much PM2.5 are riders actually exposed to?
It varies enormously by city and time of day, and it is usually worse than the nearest monitoring station suggests. Only 14% of cities worldwide met the WHO annual guideline in 2025.
Ready to breathe cleaner on every ride?
The mask was a reasonable answer to a problem nobody had built for yet. Four years and three continents later, the answer has changed — not because filters got better, but because the people who measured the problem stopped trying to seal a face and started supplying it. If you ride in traffic, that distinction is worth understanding before you buy your next box of masks.
See how the Easi Breezi unit works →
Written by Ash — mechanical engineer and founder of Easi Breezi, building an active HEPA filtration system for motorcycle helmets (patent pending). Based between Hong Kong and Bali, riding daily in the traffic this blog writes about.