Air filtration, MERV ratings, and indoor air quality field guide for HVAC

Air filtration, and the tradeoff you are always making

Air filtration is the part of an air system that pulls particulate out of the airstream before it reaches the coil, the fan, and the people downstream. Every filter does two jobs at once. It protects the equipment from a fouled coil and a dirty blower, and it cleans the air the building's occupants breathe. Those two jobs do not always want the same filter.

Here is the tradeoff that runs through every decision in this guide. A higher-efficiency filter catches smaller particles, which is better for the coil and far better for the air. It also resists airflow harder, which costs fan energy and, past a point, starves the system of the air it was designed to move. Clean air on one side of the scale, fan static and energy on the other. You are always picking a point between them.

The filter that protects the equipment can be a modest one. The filter that protects a person's lungs from fine aerosol is a different animal, and it pulls harder on the fan. Most field arguments about filters are really arguments about where on that scale the system can actually live, given the fan it has and the static it can hold.

What is a MERV rating?

A MERV rating is the Minimum Efficiency Reporting Value, the number from 1 to 16 that ASHRAE Standard 52.2 assigns to a filter based on how well it captures particles across a range of sizes. Higher means it catches more, and catches smaller. The word that earns its place in the name is minimum. The rating reflects the filter's worst-case capture during the test, not its best day or its average.

The test runs particles across twelve size ranges from 0.3 to 10 microns and sorts them into three groups. E1 is the fine end, 0.3 to 1.0 micron. E2 is the middle, 1.0 to 3.0 micron. E3 is the coarse end, 3.0 to 10 micron. A filter's MERV is set by its weakest performance across those groups, which is why a filter that does well on coarse dust but poorly on fine particles lands at a lower number than its marketing suggests.

That worst-case logic is the whole point. A filter that catches 95 percent of big particles and almost nothing fine is not protecting anyone from smoke or aerosol, and MERV is built to keep that filter from claiming a high number. When someone quotes a MERV, they are quoting the floor, not the ceiling.

What does each MERV step actually capture?

The jump that matters is not even across the scale. MERV 8 to 11 looks like a big step on paper and often is not much in the fine range, while MERV 11 to 13 is a real jump in capture of the small particles that carry smoke and aerosol.

MERV 8 captures roughly 70 percent or more of the 3 to 10 micron range and little of the fine end. It keeps a coil clean and stops visible dust, which is why it is the default in a lot of commercial equipment. MERV 11 adds meaningful capture in the 1 to 3 micron range, better for allergens and finer dust. MERV 13 is where the filter starts catching the 0.3 to 1 micron particles in quantity, on the order of half of them, which is the range that matters for infectious aerosol and combustion smoke. MERV 16 pushes that fine capture up toward 95 percent.

The exact threshold for each step shifted with the 2017 edition of 52.2, so confirm the numbers against the edition the filter was actually tested to before you put them in a submittal. The shape of the curve holds regardless. The fine-particle capture is what separates a high MERV from a middling one.

MERV-A and what happens when the filter loads

MERV alone has a blind spot, and Appendix J of Standard 52.2 was written to close it. Some filter media carry an electrostatic charge, natural or added in manufacturing, that boosts capture when the filter is clean. That charge fades as the filter loads with dust in service, and the real-world efficiency drops below the clean MERV the lab measured. A filter can test MERV 13 on day one and behave like a MERV 11 a month in.

Appendix J adds a conditioning step. It exposes the filter to fine potassium chloride particles to discharge that electrostatic effect, then re-rates capture. The result is reported as MERV-A, the efficiency you can expect after the charge is gone and the filter is doing the work mechanically. A charged synthetic filter often shows a clean MERV well above its MERV-A. A mechanical filter, the kind that catches by fiber structure rather than charge, holds its number, so its MERV and MERV-A sit close together.

For anything where the filter has to actually perform over its life, IAQ and infectious control especially, read the MERV-A, not just the MERV. The MERV-A is the honest number. A filter sold on a high MERV with a much lower MERV-A is selling you the first week, not the service interval.

What is the difference between MERV and HEPA?

MERV and HEPA are two different scales, and HEPA sits above where MERV stops. The MERV scale runs to 16. HEPA, High Efficiency Particulate Air, is a separate classification that begins beyond it. A true HEPA filter captures 99.97 percent of particles at 0.3 micron, the most penetrating particle size, the size hardest for any filter to catch. ULPA goes further, to 99.999 percent at about 0.12 micron.

The 0.3 micron figure trips people up. It is not the smallest particle a HEPA catches. It is the worst case. Particles both larger and smaller than 0.3 micron are caught more easily, because of how filtration physics works across that size, so the rating is taken at the hardest point. A HEPA does even better above and below it.

That performance costs air. HEPA media is dense, and ULPA denser still, passing 20 to 50 percent less air than HEPA for the same face area. You do not drop a HEPA into a standard rooftop unit and expect the fan to move design airflow. HEPA belongs where the application demands it: cleanrooms, isolation rooms, critical exhaust, biosafety cabinets. There it is usually the final stage behind lower-MERV prefilters that take the coarse load so the expensive HEPA lasts. For general commercial comfort and IAQ, MERV 13 does the job a HEPA would only do at a fan and energy penalty the building was never built to pay.

ISO 16890 and the PM-based standard

ISO 16890 is the newer global filter standard, and it rates filters by the particulate matter sizes that public health actually tracks rather than by a single index number. It replaced the old EN 779 in Europe and shows up more on international specs and on filters sold by global manufacturers. Where MERV gives one number, ISO 16890 reports capture against the PM fractions: ePM1, ePM2.5, and ePM10.

The labels map to the particle sizes that matter for health. ePM1 is the fine fraction, 0.3 to 1 micron, the same range as MERV's E1 and the same range that carries smoke and aerosol. ePM2.5 covers up to 2.5 micron, the fraction air-quality regulators report. ePM10 covers up to 10 micron. A filter earns a class only if it captures at least 50 percent of that fraction, and the rating is the rounded average efficiency in that band.

MERV and ISO 16890 do not convert cleanly, because one is a worst-case minimum across ranges and the other is an average against PM fractions. A rough field correspondence puts ISO ePM1 filters in MERV 13 and up territory, but treat any conversion chart as approximate. When a spec calls out one standard, match a filter rated to that standard rather than translating from the other and hoping it lines up.

The filter types and where each fits

MERV tells you what a filter catches. The physical type tells you how it fits, how long it lasts, and what it does to your static. Two filters at the same MERV can pull very differently on the fan depending on how much media is packed into the frame.

Depth is the lever most techs underuse. A 1 in pleated filter at MERV 13 can pull harder than a 4 in pleated at the same MERV, because the deeper filter has far more media area, so each square inch sees less air and less resistance. When a high MERV is the goal and the fan is the constraint, going deeper is often the answer instead of going down in MERV. The rack has to accept the depth, which is the catch.

Carbon is the odd one out. Activated carbon does not have a MERV, because it does not work by trapping particles. It adsorbs gases and odors, VOCs, and the gaseous contaminants a particle filter passes straight through. Where odor or off-gassing is the complaint, no MERV solves it. You need a gas-phase stage.

Does a higher MERV filter hurt my system?

A higher MERV filter can hurt a system that was not designed for it, and the mechanism is static pressure. Every filter resists airflow, and that resistance is part of the total external static the fan has to overcome. The filter typically accounts for 20 to 50 percent of system pressure drop depending on the configuration and how loaded it is. Push the MERV up in a thin filter and you push that share up with it.

What happens next depends on the fan. A PSC motor, the older constant-torque kind, simply moves less air as static climbs, so the building quietly loses airflow and comfort while nobody changes a setting. An ECM motor does the opposite and arguably worse. It holds the target CFM by working harder and drawing more power, and if the static climbs far enough it runs hot and shortens its own life trying to hit a number the ductwork will no longer allow. Either way the higher MERV cost you, in airflow or in energy and motor life.

The honest move is to check before you upsize the MERV. Read the static across the filter and across the system with a manometer, compare it to the fan's available external static, and confirm the system has the room. A blower CFM and static check, the kind the blowercfm tool runs, tells you whether the fan can carry a higher MERV before you commit to it. If the system is already tight, go to a deeper filter at the same MERV rather than a thin high-MERV that starves the coil and the building. This is the same static the air balancer is fighting on the whole system, which is why the filter belongs in the balancing conversation, not off to the side.

MERV 13, IAQ, and ASHRAE 241

MERV 13 became the IAQ reference point because it is the first step on the scale that captures fine aerosol in real quantity while still running at a static most commercial systems can carry. It catches on the order of half the 0.3 to 1 micron particles, the range that holds combustion smoke and the aerosol that carries airborne illness, and roughly 85 percent of the 1 to 3 micron range. That is why it shows up as the recommendation when the conversation turns to airborne disease and indoor air.

ASHRAE Standard 241, the standard on controlling infectious aerosols, sets its prescriptive filtration minimum at MERV-A 11 (effective January 1, 2025) and treats MERV 13 as a common higher target for occupied spaces during infection-risk operation. The standard frames the goal as equivalent clean airflow, a combination of ventilation, filtration, and air cleaning that adds up to enough clean air per person. Filtration is one of the levers that gets you there, and MERV 13 is the entry point because it gives meaningful fine-particle capture at a lower resistance than a 14 or higher, so the fan can move more total air for less energy.

The mistake during the rush to upgrade was dropping a MERV 13 into a system sized for a MERV 8 and walking away. The filter fit the slot and the airflow fell off a cliff, so the building got less total clean air, not more. The upgrade only delivers if the fan can carry it. Check the static first, go deeper in media if you have to, and confirm the airflow held after the swap.

What MERV filter should I use?

The right MERV is the one that meets the air-quality need the application actually has without asking the fan for static it does not have. There is no universal answer, and the spec or the equipment data controls it, but the decision sorts into a few clear cases.

For equipment protection alone, where the only job is keeping the coil and blower clean, a MERV 8 does it without loading the fan. For general commercial IAQ with people in the space, MERV 11 to 13 covers allergens and fine dust at a static most systems handle. For infectious-aerosol control, MERV 13 is the floor. For healthcare, surgical suites, smoke, or anywhere fine aerosol and bacteria are the target, you move to 14 through 16, often with HEPA as a final stage. Cleanrooms and isolation are HEPA territory by definition.

The trap is over-MERVing a system that cannot take it. A high MERV in a system with no static headroom does not give you cleaner air. It gives you less air, a starved coil, and a fan working itself to death, and the building ends up worse than it was at MERV 8. Match the filter to the application and to the fan. When the application demands more than the fan can carry in a thin filter, the answer is a deeper filter or a different air-cleaning strategy, not a thin filter the system will choke on.

The fit and the bypass

A filter only cleans the air that goes through it. Air that slips around the edges is unfiltered, and it does not matter how high the MERV is on a filter the air is dodging. Bypass is the quiet killer of filtration performance, and it is almost always a fit problem, not a filter problem.

Air takes the path of least resistance. A high-MERV filter resists airflow hard, which means the air is actively looking for a gap, and a loose filter in an oversized rack, a missing or hardened gasket, or a frame that does not seal against the track gives it one. A 1/4 in gap around a MERV 13 can pass enough unfiltered air to drop the whole system's effective efficiency to something far below the rating on the box. The higher the MERV, the worse the bypass hurts, because the pressure difference driving air through the gap is larger.

Seal it. The filter has to fit the rack snugly, the gasket has to compress against the frame, and the clips or holding mechanism have to hold it there. Check the gasket on every change, because it hardens and cracks with time and temperature and stops sealing long before it looks failed. A filter sitting loose with daylight around it is a filter the air is mostly going around, and you have paid for efficiency the system is not getting.

When should I change a filter?

Change a filter on its final pressure drop, not on the calendar. The right time to pull a filter is when it has loaded enough to reach the manufacturer's recommended final resistance, the static at which it has held its rated dust and is now costing the system more airflow than it is worth. That point comes sooner in a dusty building and later in a clean one, which is exactly why a calendar interval is a guess and a pressure reading is a measurement.

Run a filter past its final pressure drop and the static keeps climbing. The fan loses airflow or burns energy fighting it, the coil starves, and on a deeply loaded filter the media can fail or the dust can blow through. Change it too soon and you are throwing away filter life and money for no air-quality gain. The filter holds its rated efficiency across its loading life, so pulling it early buys nothing.

The practical method is to set a change setpoint. Record the clean initial pressure drop when the filter goes in, take the manufacturer's recommended final pressure drop, and change the filter when the measured drop reaches that final number. A loaded filter actually filters better than a clean one, up to a point, because the captured dust adds capture area. The skill is changing it where rising static outweighs that benefit, and that point is a number on a gauge, not a date.

The Magnehelic and the DP gauge

The gauge that tells you when to change a filter is a differential pressure gauge across the filter bank, and on most commercial systems that means a Magnehelic or a digital equivalent. One port reads upstream of the filter, the other downstream, and the gauge shows the difference, the pressure drop the filter is causing right now, in inches of water column.

That reading is the whole decision. A new filter reads its clean initial drop. As it loads, the needle climbs, and when it reaches the final drop you change it. No guessing, no calendar, no pulling a filter that still has life or running one that is choking the system. On systems without a permanent gauge, a handheld manometer with two probes gives you the same number on a maintenance round.

The gauge only helps if someone reads it and acts on it. A Magnehelic pegged at the top of its range for months means the building has been running on a choked filter while the gauge sat there telling the story to nobody. Make the reading part of the round, log it, and change on the number.

Protecting the coil

The first job of a filter, before anyone talks about IAQ, is keeping the coil clean, and a fouled coil is one of the most expensive quiet failures in HVAC. Dust that gets past the filter lands on the coil fins, builds an insulating layer, and chokes the airflow through the coil. Heat transfer drops, the system works harder for less cooling or heating, and energy use climbs while capacity falls.

A dirty coil is also hard to clean and easy to damage. Once dust packs into the fins, especially with any moisture on a cooling coil, it turns into a mat that a simple cleaning will not move, and aggressive cleaning bends the fins and makes the airflow worse. The filter is far cheaper than the coil cleaning, and the coil cleaning is far cheaper than the lost efficiency over a season of running fouled.

This is the argument for never running a system without a filter, or with a filter so bypassed it might as well not be there. A construction crew that pulls the filter to move more air during the dusty phase of a build, or runs the unit with no filter at all, hands the owner a coil that fouled before the building was even occupied. Protect the coil from day one, because the coil does not recover the way a filter swap does.

Dust-holding capacity and filter life

Dust-holding capacity is how much dust a filter can hold before it reaches its final pressure drop, and it sets the filter's service life as directly as the MERV sets its efficiency. Two filters at the same MERV can have very different lives, because the one with more media area and more depth holds more dust before it chokes.

This is where filter type pays back. A deep pleated cartridge or a bag filter holds far more dust than a thin panel at the same MERV, so it runs longer between changes and spends more of its life at a lower static. On a system with high dust load, paying for the higher dust-holding filter cuts change-outs and keeps the average static lower, which is a fan-energy saving across the year, not just a labor saving.

Dust-holding is also why a prefilter makes sense ahead of an expensive final filter. A cheap lower-MERV prefilter takes the coarse load and protects the costly high-MERV or HEPA behind it, so the expensive filter loads slowly and lasts. In a built-up air handler the prefilter is the filter you change often and cheaply, and the final filter is the one the prefilter is there to protect.

The filter in the system static

The filter's pressure drop is not a standalone number. It is one term in the total external static the fan has to overcome, and it shares that budget with the ductwork, the coil, the dampers, and every other resistance in the path. When a balancer sets a fan to deliver design airflow, the filter's drop is baked into the static the fan was selected for, clean filter and all.

That creates a problem nobody sees coming. A system balanced with clean filters has its airflow set at the clean static. As the filters load, the static climbs, the airflow drifts down on a PSC fan or the energy climbs on an ECM, and the balance the report certified slowly stops being true. The balance is a snapshot at one filter condition. The filter that loads moves the operating point away from it.

This is why filter selection and the air balance belong in the same conversation. A higher MERV than the balance assumed shifts the whole operating point and can put the system below design airflow before the filter is even dirty. If the filter spec changes after the balance, the balance has to be revisited. The filter DP, the coil, the duct, and the fan curve add up to one operating point, and changing the filter changes it. See the air balancing report procedure guide for how that total static and the fan setting get measured and documented.

Outdoor-air filtration and the economizer

Outdoor air needs filtration too, and on a system with an economizer it needs more of it at exactly the moments the economizer is working hardest. An air-side economizer opens the outdoor-air damper wide to pull free cooling, which means it is pulling large volumes of unfiltered outdoor air, with its pollen, traffic soot, and whatever the local air quality is doing that day, across the same filter bank.

That changes the filter's load. A system running heavy economizer hours loads its filters faster and with a different mix of particulate than a system on mostly recirculated air, and on a smoky-air day the outdoor-air fraction is exactly the wrong thing to maximize without good filtration behind it. The filter that was sized for a recirculation load can run short on a building that economizes aggressively.

The two systems interact in the static budget as well. The economizer's damper position changes the airflow path and the resistance, and the filter drop rides on top of that. When you commission an economizer for minimum outside air and free-cooling changeover, the filtration behind the outdoor-air intake is part of the same airflow picture, not a separate concern. The economizer and demand-control ventilation guide covers setting those damper positions and the outside-air fraction the filter then has to clean.

Data center and gaseous filtration

Data centers filter for a different enemy. In a white space the concern is not occupant health but the electronics, and the threats are both fine particulate that settles on boards and, increasingly, gaseous contaminants that corrode them. Particle filtration handles the dust. It does nothing for the corrosive gases.

ASHRAE's TC 9.9 committee and the ANSI/ISA 71.04 classification frame the gaseous problem. ISA 71.04 rates the environment by how fast it corrodes a copper or silver coupon, with G1, the mild level, as the target where corrosion is not a factor in equipment reliability. Contaminants like hydrogen sulfide, sulfur dioxide, and the oxides of nitrogen drive the corrosion rate up, and where outdoor air carries them, particle filters pass them straight through.

The answer in those facilities is gas-phase filtration, activated carbon or chemical media, added to the particulate filtration rather than instead of it. The white space gets clean particle filtration for the dust and a gas-phase stage to hold the environment at the target corrosion severity. Confirm the particulate and gaseous targets against the current ASHRAE data center guidelines and the equipment manufacturer's requirements before you spec either, because the thresholds are specific and they move between editions.

Filtration is one leg of IAQ

Filtration is one leg of indoor air quality, and a building that treats it as the whole answer still has air problems. The three legs are ventilation, source control, and filtration, and they cover different things. Filtration removes particulate already in the air. It does not bring in fresh air, and except for a gas-phase stage it does not remove gases, odors, or the off-gassing from a new build.

Ventilation is the leg filtration cannot replace. Bringing in outdoor air dilutes the carbon dioxide people exhale and the contaminants that come off furniture, finishes, and processes, and ASHRAE Standard 62.1 sets the ventilation rates for that. You can filter recirculated air all day and still have a stuffy, high-CO2 space if the outdoor-air rate is wrong. Source control is the cheapest leg. Stop the contaminant at the source, the off-gassing material, the unsealed process, the moisture that feeds mold, and you have less for the other two legs to handle.

The practical version is that a complaint about air quality is not automatically a filter problem. Odor points at gases and source control. Stuffiness and high CO2 point at ventilation. Visible dust and allergen symptoms point at filtration. Diagnose which leg is short before you sell a higher MERV that will not fix a ventilation or a source problem.

Field checklist

What to document

A filter program nobody documents is a filter program nobody can run when the tech who knew it leaves. The record is what tells the next person which filter goes in which unit, when it was last changed, and what static it should be reading.

Capture the unit and filter location, the size and MERV and MERV-A, the clean initial pressure drop, the final pressure drop that triggers a change, the change date and runtime, and the type and quantity per bank. With those on a sheet or in the system, the next change is a lookup, not an investigation, and the owner inherits a program instead of a mystery. Log it in the field as you go, with a tool like tradeos, so the record is built from the actual change-outs rather than reconstructed later.

Common mistakes

• Over-MERVing a system with no static headroom, starving the coil and the fan to get a number that looks good on paper.
• Dropping a high-MERV filter into a slot sized for a low one and never checking that the airflow held.
• A loose filter or a failed gasket letting unfiltered air bypass a high-MERV filter.
• Changing filters on a calendar instead of on the measured final pressure drop.
• Running a filter long past its final pressure drop until the fan and the coil pay for it.
• Reading the clean MERV on charged media and ignoring the lower MERV-A it drops to in service.
• Mixing up MERV and HEPA, or specing HEPA where the fan was never built to pull it.
• Treating a filter upgrade as a cure for an odor, CO2, or ventilation problem it cannot touch.
• Running a system with no filter during construction and handing the owner a fouled coil.

Standards and references

ASHRAE Standard 52.2 is the standard that defines MERV and the test behind it, rating filters 1 through 16 across the E1, E2, and E3 particle-size groups, with Appendix J adding the conditioned MERV-A value. It is the standard to cite when MERV is the subject, and the edition matters, because the step thresholds shifted in the 2017 edition. Confirm the values against the edition the filter was tested to.

ISO 16890 is the PM-based global standard, rating filters by ePM1, ePM2.5, and ePM10 capture, and it shows up on international and manufacturer specs in place of MERV. HEPA and ULPA classifications, with the 99.97 percent at 0.3 micron HEPA definition, sit above the MERV scale and are governed by separate standards and the IEST recommended practices. For ventilation and IAQ, ASHRAE Standard 62.1 sets outdoor-air rates, and ASHRAE Standard 241 sets a filtration minimum of MERV-A 11 for infectious-aerosol control, with MERV 13 a common higher target. Data center contamination falls under the ASHRAE TC 9.9 guidelines and the ANSI/ISA 71.04 corrosion classification.

None of these replaces the project specification or the filter manufacturer's data. The exact MERV thresholds, the final pressure drops, and the application targets come from the standard, the spec, and the manufacturer together. Cite the standard that governs the point, confirm the edition, and let the contract documents and the equipment data control the number you actually install to.

Units, terms, and conversions

Filter performance shows up in a few units and a few names that mean the same thing across a spec, a manufacturer sheet, and a controls screen.

Pressure drop, also called resistance or static, is measured in inches of water column, written in. w.c. or in. wg, and in pascals on metric sheets. Particle size is in microns, the micrometer, written µm. MERV is the ASHRAE 52.2 index, and MERV-A is the same value after conditioning. ISO 16890 uses ePM1, ePM2.5, and ePM10 for the same idea on a PM scale. Airflow is in cubic feet per minute, CFM, or cubic meters per hour on metric systems.

MERV

Minimum Efficiency Reporting Value, the ASHRAE 52.2 filter rating from 1 to 16

MERV-A

The MERV after Appendix J conditioning discharges electrostatic media, the in-service efficiency

E1 / E2 / E3

The 0.3 to 1, 1 to 3, and 3 to 10 micron particle-size groups MERV is scored across

Pressure drop

The static the filter adds to the airstream, in inches of water column, initial when clean and final when loaded

HEPA

High Efficiency Particulate Air, 99.97 percent capture at 0.3 micron, the most penetrating particle size

ISO 16890 / ePM

The PM-based global filter standard rating ePM1, ePM2.5, and ePM10 capture

Dust-holding capacity

How much dust a filter holds before reaching final pressure drop, which sets its life

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• What is air balancing?
• What is an air-side economizer?
• How do you investigate an indoor air quality problem?
• What is indoor air quality monitoring?
• Does duct cleaning really work?