Ozone follows different rules depending on whether anyone is in the room

Why does one gas get two completely different rules?

Because nobody drew a line around ozone. They drew it around whether somebody is standing in the room.

Picture two buildings on one street. The first has a restoration crew who taped the doors shut, ran a generator hard for hours and went home, and nobody walks back in until it’s been aired out.

The second has a box humming beside a sofa while a family watches television.

Same gas. Sometimes the same machine. One is a job with a protocol behind it. The other is a Tuesday night nobody planned.

EPA’s own wording carries both halves in one breath. High amounts of it, with nobody present, are sometimes used to clean out an empty space, and conditions should be controlled well enough that no person or pet gets exposed.

Read what that permits before you read what it warns about. Treating an empty building is a recognized approach, and a trained operator who does it properly breaks no federal rule.

California went harder and ended up in the same place. Its air cleaner rule reaches anything used in a space people can occupy, from portable units to ionizers to corona discharge generators. Then it carves out industrial ozone gear on one condition. It may run only “when no people are present.”

Two of them draw the same line, and we think that line is the whole subject.

What happens if you run one in a closed bedroom?

Somebody smells something in a bedroom. They put a generator in there, turn it up, and close the door to keep it in.

That closed door is the whole problem. It is also the move nearly everybody makes, and we keep coming back to it.

EPA researchers ran a set of these machines inside a real house in 1995, and on high with interior doors shut they produced 0.20 to 0.30 ppm. Same house, same machines, doors open and units off high, and ozone stayed under public health limits.

So the difference between safe and unsafe came down to a door, and that’s most of our advice on the subject.

Something else makes that bedroom impossible to judge from inside it. Ozone smells sharp and clean, a smell people describe after a thunderstorm. Some notice it at once. Others barely register it. And whoever you are, your nose gives up fast while you stand in it.

Smell fades. Ozone doesn’t. Somebody in that room decides it worked and goes to bed.

We’d rather hand you that scene than a warning, because the scene is the mechanism.

What happens if you use a commercial ozone machine at home?

Shaughnessy and Oatman ran the test everybody pictures. Researchers in 1991 took a large generator, one its maker rated for spaces “up to 3,000 square feet,” and switched it on in a room of 350 square feet.

Ozone climbed to 0.50 to 0.80 ppm, five to ten times what public health limits allow.

Nothing broke. That is a machine sized for a warehouse, running in a bedroom.

Britigan, Alshawa and Nizkorodov carried purifiers into offices, bathrooms, bedrooms and cars for a 2006 paper, including the wearable kind people hang around a neck. Ozone ran well past public and industrial safety limits in most of those runs. They also showed us something buyers never expect. Run two of them in one room and their output adds up.

Fifteen years on, Richard Shaughnessy was at University of Tulsa, writing a technical report for Consumer Product Safety Commission, and anybody expecting a crackdown put it down disappointed. He found an existing 50 ppb ceiling “is considered to be adequate to protect consumers,” and evidence to move it either way was “not compelling.”

He didn’t ask for a tighter rule. He asked for the existing one to stand, and the ordering matters to us.

What he was barred from weighing is worth knowing, because it surprised us. His report notes that byproduct chemistry “was beyond the scope of this project,” and that the authors “were requested by CPSC to specifically limit the references to these issues.”

Look where that rule points. Filed at 21 CFR 801.415, it bites once ozone builds past 0.05 ppm “in the atmosphere of enclosed space intended to be occupied by people for extended periods of time, e.g., houses, apartments, hospitals, and offices.”

Houses. Apartments. Rooms keeping people for hours. A room again, and not a gas.

Does ozone remove dust, mold or smells?

Not much, for three of the four reasons people buy one.

Start with dust and pollen. Federal assessment of these devices is flat about it. Ozone “does not remove particles (e.g., dust and pollen) from the air, including the particles that cause most allergies.” Those are solid bits drifting around a room, and getting them out takes something that physically catches them.

Mold and bacteria arrive back at that same door. Dyas and colleagues in 1983, then Foarde and colleagues in 1997, are where this comes from. Ozone has to run 5 to 10 times higher than public health standards allow before it stops those things regrowing. That is a trained job in an empty building, not a bedroom on a weeknight. Foarde’s group added a catch. Anything buried in porous material, duct lining or ceiling tiles, may go untouched even then.

Smells are where it genuinely does something, and where the result catches people out. Weschler and colleagues mixed ozone with fumes coming off new carpet in 1992. The smell improved. Total organic chemicals in that air went up rather than down, because ozone built new compounds while breaking old ones apart.

A nose says better and the air says worse. We would call that the trade to know before buying one for odor.

Has anybody been held liable for an ozone machine?

Greenville, Tennessee. Alpine Industries sold ozone machines for home use between $400 and $625 apiece, and its president William Converse stood behind what company literature said they could do.

Twelve jurors heard 13 days of evidence about Alpine Industries, then deliberated six more. All of them found, on 1 November 1999, that hundreds of those claims had no “competent and reliable scientific evidence” behind them. Relief from allergies, asthma, emphysema, lupus and migraine headaches. Removal of mold, dust, viruses, formaldehyde and named bacteria.

One claim survived. Jurors didn’t find smoke removal unsupported, and we read that as the fair detail to carry forward.

Magistrate Judge Dennis H. Inman ordered $1.49 million, worked out at $1,000 for every day the violations ran, and barred claims that those machines treat any medical condition. Sixth Circuit affirmed every bit of it in 2003.

Read what that case turned on. Not a gas. A sales pitch.

What does the marking on the box mean?

Less than shoppers think. Some brands print an EPA establishment number on their packaging. Displaying it “does not imply EPA endorsement or suggest in any way that EPA has found the product to be either safe or effective.”

It registers a factory. A building appears on a list somewhere, and that’s all it tells us. Closer to a business license than a seal of approval, and it goes on that box because a maker is entitled to put it there.

Nobody is quite lying, which is what makes it the sharpest trick in this category.

One marking does mean something. Since 2010, an air cleaner sold in or shipped to California has to be certified below 0.050 ppm, tested to a standard called ANSI/UL 867. Over 9,000 devices from more than 800 makers have made that list, and certified ones carry a line reading “Meets California ozone emissions limit: CARB certified.”

So we’d look for that line on any air cleaner we bought, wherever we lived. You are borrowing a state’s testing program for nothing, and a machine sold happily elsewhere but not there’s telling you something about itself.

California Air Resources Board keeps a second list nobody talks about. We counted 67 models on it, 66 brands, running from Air-Zone and Alpine to Zontec. PLACEHOLDERot seen UL 867 results for them. Its advice isn’t careful at all.

That same board settles on paperwork rather than harm. Walmart paid $20,100 in July 2022, and C&A Marketing, trading as Ivation, paid $150,000 in February 2024 over twelve uncertified models. A settlement is money and a promise about future sales. It carries no finding that anybody got hurt.

Is anybody regulating ozone air cleaners?

We went looking through Federal Register on 14 August 2026, and we came back with nothing. A search for “ozone-generating air cleaners” was empty. Narrowed to Consumer Product Safety Commission, “ozone generators” was empty too. A wider sweep turned up 18 documents, every one about outdoor air or appliance energy, none about a box sold for a living room.

New Jersey Department of Health says its state has no indoor air standard for ozone, and to us that’s ordinary rather than an oversight.

Meanwhile a federal page can go years without an edit. Two pages from Environmental Protection Agency disagree right now, because one prints 0.08 ppm as an outdoor ozone standard, a figure retired in 2015, while the other has carried 0.070 ppm ever since.

So the gap here is an absence rather than a fight anybody is having, and absences don’t get corrected on a schedule, because nothing forces the question. A machine sold for a bedroom falls into a category no agency claims, between a pesticide rule that covers claims, a consumer protection rule that covers advertising, and an outdoor standard that covers the air over a state.

What should you do about your own air?

Air cleaning comes last. Federal guidance ranks three approaches by how well they work and puts machines third, behind removing a source and bringing in outside air. Air cleaning on its own, it adds, is not regarded as enough.

The first two cost us nothing.

Find the source. A smell has a cause. Damp under a sink. Old carpet. Something in a garage. A gas appliance. Getting rid of it beats covering it up.

Then ventilate. Opening a window is boring and it works. Make the exits work before you try to remove anything.

Then, if you still want a machine, buy one with a filter. Look at its clean air delivery rate, which is how much clean air it puts out rather than how good its filter looks on paper.

And if you’ve a real problem, a fire, a flood, a building that smells wrong, that’s a conversation with somebody trained. About an empty building. With a plan for airing it out afterwards.

Ozone isn’t what went wrong in any story above. Who is in the room is.