UV & Ozone Air Purifiers

These are two different products and conflating them is the first mistake

“UV purifier” and “ozone purifier” get used interchangeably in listings, and they are not the same thing at all. Keeping them apart is the whole of the decision:

• An ozone generator is built to produce ozone deliberately and sell that as the cleaning mechanism. This is the category the EPA tells you to avoid.
• A UV-C lamp is an extra stage inside an otherwise ordinary filter unit, intended to kill microorganisms as air passes. It is not sold as an ozone producer - but EPA notes that air cleaners containing “electrostatic precipitators, ionizers, UV lights without adequate lamp coatings” “may have the potential to emit ozone.”
• Photocatalytic oxidation (PCO) pairs a UV lamp with a catalyst and aims at gases rather than microbes. It is the stage most often sold under a proprietary name.

The ionizing stage is a fourth thing again and it has its own page - air purifier vs ionizer. What all four share is that none of them is the part with a measurement standard behind it.

Ozone generators: what the EPA actually publishes

This is the clearest guidance in the whole category, and it is worth reading in the agency’s own words rather than a summary of them. On effectiveness, EPA states that at concentrations within public health standards ozone “has little potential to remove indoor air contaminants,” and that it “is not effective at removing many odor-causing chemicals.” It also does not remove airborne particles such as dust and pollen.

On health, EPA states that “When inhaled, ozone can damage the lungs,” and that “Relatively low amounts can cause chest pain, coughing, shortness of breath and throat irritation.” The concentration figures it reports are the part that makes this concrete:

And there is a second-order problem that gets no attention in listings at all. EPA warns that when ozone reacts with other chemicals in the air, “the reaction can form a variety of harmful or irritating by-products.” In one carpet-chemistry experiment it reports that total airborne organic chemicals “increased rather than decreased” after ozone was introduced. So the device marketed as removing a new-carpet smell can add to the chemical load in the room.

UV-C: the contact-time problem, in one sentence

A germicidal lamp needs a certain dose to inactivate an organism, and dose is intensity multiplied by exposure time. In a portable purifier the air is past the lamp in a fraction of a second, which is the structural reason this stage underdelivers in a domestic unit compared with the sustained exposures UV is used for elsewhere.

Our position on the evidence has not changed and we are not going to overstate it in either direction: we have not tested any of these units, and the EPA guide does not make a claim about UVGI’s effectiveness against particles or microbes in homes one way or the other. What the guide does say is that research has not shown UV light, plasma or photocatalytic oxidation to remove gases effectively in portable residential air cleaners, and it does not recommend any of them.

Which leaves a simple buying rule. The filtration in a purifier has a published, comparable number behind it. The lamp does not. Pay for the number.

What to look for instead, in order

Every one of these is a published figure you can check before you pay, which is the whole difference between a specification and a feature:

1. An AHAM-verified CADR for smoke, dust and pollen, against the exact model number in the listing. If a unit publishes no CADR at all, you cannot size it and you cannot compare it.
2. Coverage stated at a useful air-change rate, not at one air change per hour. The two-thirds rule is on CADR explained.
3. A real carbon stage with its own part number if odors or gases are your problem - a coating on the HEPA wrap is not a carbon bed. See activated carbon explained.
4. Published power draw, ideally per fan speed, because a purifier is a machine you leave running.
5. CARB certification, read with the UV caveat above.
6. An off switch on any extra stage. Where an ionizing or UV stage is optional, you can simply switch it off and rely on the filters, which is what we would do.

The three units below, and why one of them has a lamp

We have deliberately included a unit with a UV-C lamp rather than pretending the category does not exist, because the honest position is not “never buy a purifier that has one.” It is “never pay for the lamp.”

• Coway Airmega AP-1512HH Mighty - filtration plus a bipolar stage we would rather do without and would check can be disabled. It publishes power draw per fan speed, which almost nobody does: 3.7 W on speed one against 75.3 W on speed three. That is the kind of figure a lamp will never give you.
• Levoit Core 400S - 231 CFM CADR with coverage published at both one and 4.8 air changes per hour. No UV, no ozone stage, and 38 W for the air it moves.
• GermGuardian AC4825E - this one does have a UV-C lamp, and we rank it anyway because it publishes all three AHAM CADR figures against its printed model number at the cheapest price we could verify that in. Our registry entry says the lamp “is a marketing feature at this air speed and we do not count it in the ranking,” and that is exactly how we would have you buy it: for the 99 CFM smoke CADR, with the lamp as an irrelevance you are not paying a premium for.

Honest cons, including on the units we recommend

• Coway: 75.3 W on speed three is the highest draw of the three, so how its auto mode behaves decides your bill. The bipolar device is a stage we would rather not have at all.
• Levoit Core 400S: we could not find a published filter replacement interval in Levoit’s specification, so we do not quote one - and the 1,718 sq ft headline is a single air change per hour, not a useful rate.
• GermGuardian: the advertised 743 sq ft per hour works out to about 149 sq ft at AHAM’s two-thirds smoke rule, 55 dBA on high is audible in a quiet room, the carbon is a thin coating rather than a bed, and filters run on a 6-8 month cadence that erodes the price advantage over three years.
• All three: none of them will remove a gas your carbon stage is too small for. If your problem is VOCs rather than particles, the honest answer is more carbon by weight, which is its own page.

What we would actually do

Treat UV as a feature you neither seek out nor refuse: if the unit you want on CADR, carbon and watts happens to have a lamp, buy it for the first three and ignore the fourth. Treat an ozone generator as a product to walk away from, because the agency that studies this has told you to in plain words, and the devices it measured were producing concentrations it describes as several times the public health limits. Then spend the money you did not spend on a lamp on more carbon, which is the stage that actually does something about the smell you bought the machine for.