Ozone Safe: Only 0.006ppm, 1/8th the Legal Limit

It produces a trace amount, far below the legal limit. Intertek measured 0.002 to 0.006 ppm depending on fan speed in 2025 under UL 867 Section 40 (report 106174008CRT-002), against a limit of 0.050 ppm set by UL 867 and the California Air Resources Board.

No. An ozone generator deliberately fills a room with ozone as its cleaning method, and the room must be unoccupied while it runs. WellisAir cleans the air with hydroxyl radicals and is designed to run continuously in occupied rooms. Its measured ozone output is 0.006 ppm, about one eighth of the legal limit.

The concern behind this question is real and worth taking seriously. When ozone meets terpenes — the aromatic compounds in citrus oils, pine cleaners and many air fresheners — the reaction can produce formaldehyde and ultrafine particles. It is why general advice about air cleaners often warns against anything that generates ozone, and our citrus cartridge contains d-limonene from orange peel, so it is a fair thing to ask about.

We measured it rather than argue about it.

CRESCA, at the Polytechnic University of Catalonia, burned incense in a chamber of about 8,000 cubic feet to fill the air with VOCs, ran the unit, and measured against a control running at the same time with no device in it. Thirty minutes in:

With the citrus (d-limonene) cartridge — formaldehyde at 33 µg/m³, against the control’s 500. Down 93.4%.
With the unscented hydrogen-peroxide cartridge — 25 µg/m³. Down 95.0%.

Formaldehyde fell with both cartridges. It did not rise. Korea Conformity Laboratories measured the same direction separately: formaldehyde down 65% over four hours.

Two things make that unsurprising. The amount of ozone present sets how much of this chemistry can happen at all, and ours is 0.002–0.006 ppm — two to six parts per billion. And formaldehyde is itself one of the compounds hydroxyls act on, which is why both laboratories measured it falling.

The other half of the concern is ultrafine particles, and the same study measured those too. With the citrus cartridge, PM2.5 read 6 µg/m³ at thirty minutes against the control’s 113, and the finest fraction the meter reports, PM1.0, read 4 against the control’s 85. Readings never rose above 6 at any point in the two hours, with either cartridge. Particles did not increase.

That is a narrower statement than it may look, and we mean it narrowly. It is not a claim that WellisAir removes particles — it does not, and a HEPA filter is the right tool for dust, pollen and dander. What the readings show is that running the unit in a room full of smoke did not generate particles, which is what the objection predicts it would. The measurements come from a consumer-grade meter photographed on site, from a single run, and the control clears on its own within an hour, so thirty minutes is the only point where the comparison carries weight.

Those two results are worth reading together. Formaldehyde and ultrafine particles are produced by the same reaction between ozone and terpenes, so if that chemistry were running at any rate worth worrying about, both should have climbed. Neither did.

The unscented cartridge is our default for this reason among others. It contains hydrogen peroxide and no d-limonene, so the terpene reaction has nothing to work with at all.

What we do not have is a study designed specifically to look for byproducts — measuring acetaldehyde and ultrafine particles during operation alongside formaldehyde. We have asked for one. The test reports are here.

It is designed for continuous use in occupied rooms, with ozone output about one eighth of the legal limit and CARB certification. As with any air treatment device, if anyone in the home has a respiratory condition or a known sensitivity to fragrance or chemical emissions, discuss it with their doctor first — and note that the citrus cartridge contains d-limonene, a natural orange-peel extract that some people are sensitive to. Unscented is our default cartridge and contains no d-limonene.

Yes. WellisAir is a CARB certified air cleaning device, listed by the California Air Resources Board. California operates the strictest ozone regime for air cleaners in the United States, and devices sold there must be certified.

They are different technologies. Ozone generators work by producing ozone at concentrations intended to oxidize contaminants, which is why they must be used in empty rooms and why the EPA warns against using them in occupied spaces. Hydroxyl-based devices such as WellisAir work by releasing hydroxyl radicals, which react with contaminants and leave water vapor. WellisAir’s measured ozone output is 0.006 ppm.

The EPA cautions that ozone generators sold as air cleaners can produce ozone at levels that may be harmful, and advises against using them in occupied spaces. That guidance is the reason we publish our UL 867 test results: it is the only way to tell one kind of device from another. Our 2025 result is 0.006 ppm against a 0.050 ppm limit.

A HEPA purifier is a fan and a filter, so it produces no ozone at all; its limitation is that it only removes what passes through the filter and does nothing about odors or gases. WellisAir produces 0.006 ppm, a small fraction of the legal limit, and works on odors, smoke gases and VOCs in the air and on surfaces. The two are complementary, and many customers run both.

No. It is designed to run 24 hours a day with people and pets present. That is the practical difference between a hydroxyl device and an ozone generator.