What it does, and what it does not
A particle filter removes particles: dust, pollen, pet dander, mold spores, smoke particulate, and other things with mass that a fibrous medium can catch. That is a real and worthwhile job, and on a bad air day a well-sized unit noticeably lowers what is suspended in a closed room.
What it does not do is generate air. It does not add oxygen, remove carbon dioxide, or substitute for ventilation — the CO2 you exhale in a closed bedroom accumulates whether the purifier runs or not, and stuffiness is a ventilation problem that a filter cannot touch. It does not remove gases unless it carries a sorbent for the purpose, and even then activated carbon has a finite capacity that fills silently. It does very little about anything that has settled: a filter cannot pull dust out of a rug, which is why cleaning and a filter are complementary rather than alternatives.
The useful framing is that a purifier reduces concentration in a specific room, continuously, as long as the doors are closed and the filter has life in it. It is not a whole-house device, it does not make outdoor air safe to bring in, and it does nothing about a source you have not removed. Take the source out first — the damp material, the litter box, the smoker on the balcony, the gas range used without the hood — and the filter has less to fight.
CADR, the number that lets you compare
Clean Air Delivery Rate is the output of a standardized industry test: a unit runs in a test chamber against a known challenge and the result is expressed as an equivalent volume of perfectly clean air delivered per unit time, usually reported separately for smoke, dust and pollen. Smoke is the useful one to look at for fine particulate, because it uses the smallest challenge particles of the three.
Its virtue is comparability. Manufacturers can describe filter media however they like, but a CADR measured to a common method is the same measurement everyone else took. Two cautions go with it. First, CADR is measured at the highest fan setting, and the highest fan setting on most units is loud enough that nobody actually runs it there — so the number in the catalogue reflects a mode you will use rarely. Second, a coverage area printed on a box is a marketing translation of that number using assumptions about ceiling height and how many times per hour the air should be turned over, and those assumptions vary between manufacturers.
| Spec on the box | What it really tells you |
|---|---|
| CADR, smoke / dust / pollen | Tested delivery of clean air at maximum fan. The comparable number, and the smoke figure is the one to weigh for fine particulate. |
| Coverage area in square feet | A derived claim, dependent on assumed ceiling height and assumed air changes per hour. Useful only if you know which assumptions were used. |
| Air changes per hour | How many times the unit could theoretically process the room's volume in an hour. Higher is better for a room with a continuing source, like smoke coming in from outdoors. Work out your room volume with the air changes calculator. |
| "True HEPA" versus "HEPA-type" or "HEPA-like" | The qualifiers are the tell. A filter described as HEPA-type is claiming a resemblance, not a rating. |
| Rated wattage | What it costs to leave running. Cross-check with the appliance running cost calculator. |
| Noise in decibels, per fan speed | The number that decides whether it stays on. Look for the lowest-speed figure, not just the maximum. |
Sizing it to the room
The mistake is buying one unit for a house. Air does not circulate freely between rooms with doors and hallways in the way that would require, and a purifier in the living room does very little for a closed bedroom down the hall. Size to the room you will actually run it in, with the door shut.
Buy larger than the minimum, for one specific reason: a unit that is oversized for the room can hit the same clean-air delivery on a low fan speed that an undersized unit only reaches at maximum. Low speed is quiet, so it runs. The most common failure in practice is not a badly chosen filter, it is a machine turned off or held on the lowest setting because it is annoying, and buying extra capacity is how you avoid that.
For a bedroom, prioritize the low-speed noise figure and keep the unit off the floor beside the bed where its intake collects everything and its output blows on you. For a living space with ongoing sources, prioritize the air-change figure. Give it a foot or so of clearance on the intake and exhaust sides — pushed against a wall or into a corner it recirculates its own output and treats a small pocket of the room. Ceiling fans help by mixing the room so the unit is not just cleaning one corner.
Filter types, and one to avoid
| Technology | What it does | Honest note |
|---|---|---|
| True HEPA mechanical filter | Captures fine particulate at high efficiency to a defined standard | The workhorse. This is what you are buying. It is a consumable. |
| Pre-filter | Catches hair, lint and large dust before the main filter | Usually washable. Cleaning it regularly is what makes the expensive filter last. |
| Activated carbon or other sorbent | Adsorbs some gases and odors | Capacity is finite and fills with no visible sign. A thin carbon-coated mat does less than a thick bed. It will not solve an ongoing source. |
| Ionizer | Charges particles so they clump and settle out or stick to surfaces | Particles end up on your walls and floors rather than in a filter. Some ionizers produce ozone as a byproduct — check whether the unit has a switch to turn it off, and use it. |
| Ozone generator, sometimes sold as an activated-oxygen or ozone purifier | Deliberately produces ozone | Ozone is a lung irritant. Public health and air-quality agencies have been consistent that ozone-generating devices should not be used in occupied spaces, and that at concentrations low enough to be tolerable they do not effectively clean air anyway. This one is worth saying plainly: do not buy it for a room people are in. |
| UV-C lamp inside the unit | Marketed for microorganisms | Effectiveness depends heavily on lamp intensity and how long air is exposed as it passes, which is brief in a fast-moving unit. Some designs also produce ozone. Treat it as a bonus feature rather than a reason to choose a unit. |
| Photocatalytic or plasma technologies | Various proprietary claims | Vendor-specific and difficult to compare. Some produce byproducts. If a unit's case rests on a technology with no comparable published test, weigh it accordingly. |
The short version: a fan pulling air through a genuine mechanical filter is a well-understood, testable, boring approach that works. Everything sold on the promise of destroying rather than capturing deserves harder questions, and anything that intentionally makes ozone belongs outside an occupied room.
Filters, and the part that gets forgotten
A filter is a consumable, and the running cost of replacements over a few years frequently exceeds what the machine itself cost. Look up the replacement filter and its interval before you buy the unit, not after, and be wary of a cheap machine with a proprietary filter that only one seller carries.
Intervals printed in a manual assume typical use, and a unit running continuously through a smoke week or in a house with three shedding animals loads far faster. Signs a filter is finished: reduced airflow at the outlet, the fan ramping up more often to reach the same result, a musty smell in the output, or visible loading across the pleats. Carbon is the sneaky one, because a saturated carbon bed looks exactly like a fresh one and quietly stops adsorbing. Vacuum or rinse the pre-filter on a schedule you will actually keep, put the main change on a calendar reminder, and do not wash a HEPA filter unless the manufacturer says it is washable — water damages the media and leaves you a filter-shaped object that passes air freely.
Noise, sensors and the settings people leave wrong
The noise figure at the lowest speed decides whether the machine runs at night, and a unit nobody can sleep next to gets switched off, which is a hundred percent reduction in effectiveness. Check the sleep-mode decibel figure before the maximum CADR.
Auto modes work off an onboard particle sensor, and consumer sensors vary in accuracy and drift as they get dusty; some can be cleaned according to the manual. Treat the colored light as a rough indicator rather than a measurement — it reacts enthusiastically to cooking and aerosol spray and can under-react to a slow rise. On a day when you know the outdoor air is bad, override auto and run a fixed speed.
The other setting people get wrong is on the wall, not the machine. Running a purifier with the window open is filtering the outdoors. Close the room, run the unit, and reopen deliberately when conditions improve — the full protocol is in air quality and wildfire smoke.
Where a purifier is not the answer
Several common problems get a purifier bought for them and are not solved by one. Persistent musty smell is a moisture problem: the fix is finding the water and controlling humidity, which means a correctly sized dehumidifier and the work in mold removal, not a filter downstream of a growing colony. Cooking smoke and grease want a range hood that actually vents outside. Radon is a gas that filters do not remove and that needs testing and a mitigation system. And if the house is leaky enough that outdoor air pours in continuously, the filter is fighting the envelope; sealing and insulating is the upstream fix, covered in insulation and air sealing and the home energy audit.
Buy for the room, buy enough capacity that low speed is sufficient, read the replacement filter price before the unit price, and put the filter change on a calendar. That is most of it. The machine that works is the boring one that is still running on low six months from now with a clean pre-filter, not the one with the most impressive list of technologies on the box.
Questions people ask
What size air purifier do I need?
Size it to the room you will run it in with the door closed, not to the house — air does not move between rooms freely enough for one unit to serve several. Compare on CADR, which is a standardized test result, and lean larger than the minimum for a specific reason: an oversized unit reaches the same clean-air delivery on a quiet low speed that an undersized one only reaches at maximum. The most common real-world failure is a machine switched off or held on its lowest setting because it is too loud, and buying extra capacity is how you avoid that. Give it clearance from walls and corners so it does not just recirculate its own output.
What does CADR mean?
Clean Air Delivery Rate — the result of a standardized industry test that runs the unit in a chamber against a known challenge and expresses the outcome as an equivalent volume of perfectly clean air delivered per unit of time, reported separately for smoke, dust and pollen. The smoke figure is the one to weigh for fine particulate, since it uses the smallest challenge particles. Its value is that it is comparable across brands in a way that filter marketing is not. Two caveats: it is measured at maximum fan speed, which is rarely how anyone runs the machine, and the square-footage claim on the box is derived from it using assumptions about ceiling height that vary by manufacturer.
Are ozone air purifiers safe?
No, and this is one of the few product questions with a clear answer. Ozone is a lung irritant, and devices that deliberately generate it — sold variously as ozone purifiers, activated oxygen or similar — have been the subject of consistent warnings from public health and air-quality agencies against use in occupied spaces. At concentrations low enough to be tolerable to breathe, they also do not effectively clean air. Separately, some ionizers produce ozone as a byproduct rather than by design; if a unit has an ionizer function, check whether it can be switched off and switch it off. A mechanical HEPA filter with a fan is the approach that is both testable and safe indoors.
How often should I change the filter?
Start with the interval in the manual, then adjust for reality — the manual assumes typical use, and a unit running continuously through a smoke week or in a house with several shedding pets loads much faster. The signs are reduced airflow you can feel at the outlet, the fan ramping up more often to achieve the same result, a musty smell in the output, or visible loading across the pleats. Activated carbon is the deceptive one, because a saturated carbon bed looks identical to a fresh one and quietly stops working. Clean or rinse the pre-filter on a schedule, since that is what makes the expensive filter last, and do not wash a HEPA filter unless the manufacturer says it is washable.
Can an air purifier replace opening a window?
No. A purifier is a fan and a filter — it removes particles from the air already in the room and it does not generate air, remove carbon dioxide, or supply oxygen. Carbon dioxide from your own breathing accumulates in a closed bedroom regardless of whether the machine is on, and stuffiness is a ventilation problem a filter cannot touch. Without a sorbent bed it does little about gases either, and even carbon has finite capacity. During a smoke event the trade-off is deliberate: keep the house closed and filtered while the outdoor air is bad, then flush the house with a short, purposeful cross-draft when conditions improve rather than leaving a window cracked continuously.