How to choose an air purifier: a step-by-step decision

How to choose an air purifier comes down to five checks made in order: what you are trying to remove, how big the room is, whether the filter can do that job, how loud the unit is at the speed you will really use, and whether it makes ozone. After that, count what it takes to keep one running. The best air purifier for you is the one that passes all five for one specific room, which is why no list of top models can settle it. Start with the problem, because a machine that is excellent at particles can do nothing for a chemical smell. Then compare published numbers instead of adjectives. And check the last section first if you have not yet dealt with the source of the problem, because for some homes the right decision is not to buy one.

Step 1: Name the problem

Different pollutants need different hardware, and EPA notes that most filters are designed to filter either particles or gases.

If your concern is in the first two rows, the answer is particles, which keeps the rest of the choice simple. Be honest about the goal too. If it is a health condition, take the decision to a clinician; the evidence on what these machines do is more modest than the advertising.

Step 2: Size it to the room

Measure the room you will use it in, and multiply length by width. Then use smoke CADR, which EPA calls the most conservative estimate for small particles. AHAM’s rule of thumb, repeated by EPA, is that the CADR should be at least two-thirds of the room’s area in square feet.

A worked example: a 12 by 15 foot bedroom is 180 square feet, and two-thirds of 180 is 120, so look for a smoke CADR of at least 120 cfm. EPA says the chart is for estimation purposes and is calculated on an 8-foot ceiling. For a higher ceiling, both EPA and AHAM say to use a unit sized for a larger area. If you want a figure, scale by volume: a 10-foot ceiling on that same room is the air of a 225 square foot room with a normal ceiling, which works out at 150 cfm. That scaling is simple arithmetic and not a published rule.

For wildfire smoke the two sources disagree. AHAM recommends a smoke CADR equal to the size of the room in square feet, which is 180 for that bedroom. EPA’s wildfire guidance uses two-thirds, so 120. Size to the higher figure if smoke is the main reason you are buying. More on the sums, ceilings and open-plan rooms is in the room size guide, and the explanation of CADR ratings covers the three numbers on the label and where the sources differ on them.

Step 3: Check the filter

For particles

Look for a filter described as HEPA, and read that word carefully. EPA, using the definition set by the Department of Energy, says a HEPA filter can theoretically remove at least 99.97% of particles of 0.3 microns, but its Technical Summary says there is no widely accepted definition of HEPA performance in consumer products. Words such as “HEPA-type” or “HEPA-style” mean even less. EPA’s earlier edition called pleated filters given such names similar in appearance to true HEPA but made with less efficient media. So the label is only a starting point, and the CADR test result is the better evidence, because it measures the whole machine, including any air that leaks around the filter. The guide to HEPA filters goes further.

For gases

If gases are on your list, the amount of sorbent is what counts. EPA says activated carbon filters can be effective provided there is a large amount of material, “the thicker the better”, and it adds that many consumer-grade systems with small amounts of carbon have unknown effectiveness. There is no rating to compare: EPA says there are no widely used performance rating systems for portable air cleaners designed to remove gases, and AHAM agrees that no universal test method exists yet. A thin carbon-dusted sheet will not do the job for long, and EPA says activated carbon is not especially effective against formaldehyde. Ask the seller for the weight of carbon, and be skeptical of a unit that cannot say.

Step 4: Check the noise at the speed you will use

EPA says performance ratings are determined at maximum airflow and therefore typically maximum noise, and that as of 2017 quantified noise is not part of the US performance ratings. A box can therefore show a noise figure from one setting beside a CADR from another, and those two numbers do not describe the same moment.

Before buying, look for a noise figure and an airflow figure for the same setting. For scale, EPA says around 50 dB is roughly the noise of a modern refrigerator. In the 2024 trial by Lu and colleagues, the unit on its quieter setting measured 59.6 +/- 11.0 dB at 1 meter, with an advertised 330 cfm at top speed and a measured 263 cfm on that setting. In a 2022 home test by Chen and colleagues, a unit ran at about 31.4 dB on low and about 56.5 dB on high.

This matters because of how people behave. EPA reports that units were used less over time in some studies, and that noise was a stated reason for turning them off at night. A practical answer is to buy more CADR than the room strictly needs and run it at a middle speed. If the published sheet gives no figures for lower speeds, assume the delivered air is meaningfully lower: EPA says it could be considerably lower than the highest advertised.

Step 5: Rule out ozone

EPA says to avoid portable air cleaners that intentionally produce ozone, and it calls ozone a lung irritant. CARB and EPA both warn that some marketing hides the word, using terms such as “activated oxygen” or “energized oxygen”. The guide to ozone generators explains why they do not work as claimed.

The most useful public check is California’s. Section 94802(a) of the state’s regulation says no person shall sell an indoor air cleaning device in California “unless the device is certified by CARB to produce an ozone emission concentration not exceeding 0.050 ppm”. CARB publishes a list of certified devices, and it is worth knowing what a listing means. CARB’s page says certification “does not reflect air cleaner effectiveness or ‘health safety’”, and that certification does not imply a specific air cleaner is “safe to use”. A listing tells you about ozone emissions and electrical safety. It tells you nothing about CADR. Units using mechanical filtration alone are exempt from the ozone test, per section 94804(b).

EPA also says that some electrostatic precipitators, ionizers, uncoated UV lamps and plasma units may emit ozone. For most homes the simplest route is a filter-only machine, or one where the electronic feature can be switched off, and EPA says mechanical filters do not produce ozone.

Step 6: Count the running commitment

• Filters. EPA says all filters need regular replacement. Before buying, confirm that the replacement is sold separately, how often the maker says to change it, and what happens to the machine if the part goes out of production. The guide to replacing filters has the details.
• Electricity. Federal rules set minimum efficiency levels for conventional room air cleaners, and the higher tier applies to units manufactured on or after December 31, 2025. An ENERGY STAR label indicates a unit met that program’s criteria. To estimate running cost, the Department of Energy’s formula is watts divided by 1,000, times hours, times your rate per kilowatt-hour; the guide to electricity use works through it.
• Size and handling. A heavy machine stays in one room, so if you plan to move it between bedroom and living room, check for a handle or wheels.

Reading a spec sheet and the marketing

Also, a logo or a registration number is weaker than it looks. EPA says the presence of an EPA number on a product’s packaging does not imply that it endorses the product or has found it safe or effective.

When not to buy one

Skip the purchase, at least for now, if any of these describes your situation.

• The smell or problem comes from something you can remove, seal off or vent to the outdoors. EPA says the most effective steps are to reduce or remove sources and to ventilate with clean outdoor air.
• The problem is damp or mold. Fix the moisture first.
• You are worried about radon or carbon monoxide. EPA does not recommend air cleaning for radon, and carbon monoxide is not readily captured by sorbents.
• Someone smokes indoors and you hope the machine will make that acceptable. EPA says filtration can reduce exposure but will not eliminate it.
• You cannot keep it running in the room you care about, with a filter you will actually replace.

If you get past that list, the steps above will narrow the field to a small number of machines, and the choice between them can reasonably come down to noise and the replacement filter.