Microfiltration (MF): What It Removes and When You Need It

Who this is for: anyone dealing with biological contamination — bacteria, protozoa, cysts — who wants a membrane-based barrier rather than a cartridge filter. MF sits between a sediment filter and ultrafiltration on the size spectrum: coarser than UF, finer than a standard depth filter. It stops everything with a physical size above roughly 0.1 micron. Viruses are smaller than that, and dissolved chemistry has no size to catch. If viruses or chemistry are your concern, MF alone is not the answer.

The Short Answer

Microfiltration is a membrane process — water is pushed through a membrane with pores in the 0.1 to 10 micron range. Anything larger than the pores does not get through. That range covers bacteria, protozoan cysts (Giardia and Cryptosporidium), most microplastics, and all suspended particles. It does not cover viruses, which are smaller, and it does not touch anything dissolved — lead, nitrate, PFAS, hardness, and chlorine all pass through a microfiltration membrane without any trouble.

The thing MF does that a sediment depth filter cannot is hold a consistent pore size across the entire membrane. A depth filter is statistical — most particles above its rating are caught, but the bed has a distribution of gap sizes and a few things sneak through. A membrane is more absolute: if the pore is smaller than the organism, the organism does not get through. That is why MF is a legitimate microbiological barrier in a way that a cartridge sediment filter is not, and why it is used in water systems where bacteria and cysts are actual concerns rather than theoretical ones.

How It Actually Works

The membrane is a sheet of polymeric material — typically polypropylene, polyethersulfone (PES), or polyvinylidene fluoride (PVDF) — with a controlled pore structure. Water flows through under pressure; particles larger than the pores are retained on the feed side. In residential systems this usually happens in a hollow-fiber configuration: thousands of hair-thin tubes bundled together, with water flowing either inside-out or outside-in through the tube walls. The hollow-fiber format packs enormous surface area into a small housing, which keeps pressure requirements reasonable and flow rates usable.

MF membranes need periodic cleaning or backwashing to clear the accumulated layer of rejected material — called the cake layer — that builds on the membrane surface and reduces flow. In a well-designed system this is automated. In a residential cartridge format it is simply a replacement schedule, the same as a depth filter, just on a different timeline because the membrane holds its structure longer than a fibrous cartridge.

The size cutoff at the fine end of MF — around 0.1 micron — is why bacteria are reliably caught. E. coli runs 0.5–2 microns; most pathogenic bacteria fall in similar ranges. Giardia cysts are 8–14 microns and Cryptosporidium oocysts 4–6 microns — both well above the MF cutoff, and both reliably removed. Viruses, at 0.02–0.3 microns, straddle or fall below the bottom of the MF range, which is why the standard position is that MF cannot be relied on for virus removal without a separate disinfection step.

Who Actually Needs It

MF earns its place in systems where biological contamination is a real concern, not a theoretical one. The primary candidates are:

• Private wells with documented or suspected bacterial contamination. A positive coliform result, a well near livestock or a septic system, a shallow well in agricultural land — any of these makes a physical bacterial barrier worth having. UV disinfection inactivates bacteria but does not remove them; MF removes them. Depending on your situation you may want both.
• Surface water intakes. Streams, ponds, lakes, and rainwater catchment all carry cysts and bacteria at levels that make MF a standard first membrane stage. Sediment prefiltration ahead of MF is not optional here — turbid water loads and blinds membranes quickly.
• Emergency or off-grid water treatment. MF membranes are the working element in most gravity-fed field filters and pump-style backcountry filters. The same physics that makes them useful at a treatment plant makes them useful on a camping trip.
• Immunocompromised households. The CDC and EPA recommend that people with compromised immune systems — HIV/AIDS, chemotherapy, organ transplant recipients — take extra precautions against Cryptosporidium specifically, which is chlorine-resistant and which MF reliably removes.

You probably do not need MF if you are on a municipal supply that already delivers biologically safe water, your concern is dissolved chemistry (lead, PFAS, nitrate, chlorine), or your well tests consistently negative for bacteria and you are not in a high-risk situation. A sediment filter ahead of your carbon block or RO unit is the right call there — adding MF to a system that does not have a biological problem is a cost without a benefit.

What It Cannot Do

This is where the marketing around MF most often misleads people.

Viruses. Norovirus runs 0.02–0.04 microns. Hepatitis A is similar. Standard MF pores at 0.1–0.2 microns let most viruses through. In a municipal system this is handled by the chlorination step, which inactivates viruses reliably. On a private well or surface water, if viruses are a concern — which they genuinely are in some rural settings, around livestock, and in developing-world contexts — MF alone is not enough. You need UV disinfection, a UF membrane (tighter pores), or boiling alongside it.

Dissolved contaminants. Lead, copper, nitrate, PFAS, arsenic, fluoride, chlorine, hardness — none of these are particles. None have a size to catch. MF does not touch them. A membrane that markets itself as removing “99% of contaminants” while showing a bacteria icon in the fine print is telling you something true and implying something much broader. The broad implication is wrong.

Chemical taste and odor. Chlorine, hydrogen sulfide, and organic compounds that make water taste or smell off are all dissolved. They sail straight through an MF membrane. If taste and odor are your issue, carbon adsorption is the mechanism you want, not a membrane.

The Honest Tradeoffs

MF offers a genuine, consistent microbiological barrier — that is real and worth having when you need it. The costs are pressure drop, fouling management, and the need for prefiltration.

Membranes require more driving pressure than a depth filter, and as the cake layer builds between cleanings, the pressure drop across the membrane increases and flow falls. In a residential system sized correctly this is a slow process — weeks to months between backwashes — but it is non-zero. Turbid, heavily particle-laden water accelerates fouling dramatically, which is why a sediment pre-filter is not optional when feed water quality is poor.

Membrane integrity is also worth naming. A crack or breach in the membrane creates a bypass — water that never sees the pores. In municipal systems, membrane integrity is continuously monitored with pressure-decay tests. In a residential cartridge, you are trusting the housing seal and the cartridge itself. Buy from manufacturers with third-party certification and handle cartridges carefully during replacement — a cartridge you ripped putting it in is not doing what the box says.

Can You DIY This?

Yes, with a reasonable level of plumbing competence. Residential MF housings install the same way as sediment housings — in-line, with inlet and outlet connections, usually with a pressure gauge before and after so you can watch the pressure drop develop over time. The main considerations are sizing the membrane correctly for your flow rate, and adding a sediment pre-filter if your water has any turbidity at all. Undersizing a membrane means fast fouling and frequent replacement; oversizing is mostly a cost question.

For off-grid and surface water use, gravity-fed MF systems require no plumbing at all — they run on head pressure from a raised feed reservoir. The trade is slow throughput. Know your daily volume needs before you choose a system.

How to Buy Without Getting Fooled

• Look for NSF/ANSI 58 or NSF/ANSI 53 certification for the specific claims being made. For cyst removal specifically, NSF/ANSI 53 cyst reduction certification is the mark to look for. A filter that says “removes Cryptosporidium” without a certification number behind that claim is asking you to trust the label.
• Check nominal versus absolute pore size. Unlike depth filters where “nominal” and “absolute” are mostly marketing vocabulary, in membrane filtration the distinction has some meaning. An absolute-rated membrane holds its pore size tightly across the entire surface; a nominal rating allows more variation. For biological applications — cyst removal — look for absolute-rated membranes with certification.
• Prefiltration is not optional on dirty water. If a vendor sells you an MF membrane for a turbid surface water source without mentioning a sediment pre-filter, they are either uninformed or selling you a fast replacement cycle. Fouled membranes are not failed membranes — they clean up — but a membrane that cannot be backwashed in a residential cartridge format just becomes trash sooner.
• Virus claims require specifics. If a product claims virus removal, look for the mechanism (UF membrane with tighter pore specification, or a disinfection stage) and the certification (NSF/ANSI P231 for portable water purifiers covers virus reduction). An MF membrane without those things is not removing viruses, regardless of what the marketing says.

The Deep End

The reason membrane filtration works the way it does — reliably and reproducibly — is that it is not a statistical process the way depth filtration is. A depth filter holds a distribution of gap sizes, and its efficiency is an average across that distribution. A membrane is a defined structure: pores of a specific size, made by a controlled manufacturing process, covering the entire active area of the membrane. If the pore is 0.2 microns and the organism is 0.5 microns, the organism does not get through. This is also why membrane integrity matters so much — one crack is not a slightly worse membrane; it is a bypass that the rest of the membrane cannot compensate for.

The MF/UF boundary is a practical designation, not a sharp physical line. Manufacturers and standards set it roughly at 0.1 microns — below that is UF, above is MF. In practice, a tight MF membrane and a loose UF membrane may perform similarly for many organisms, and some products deliberately straddle the line. What matters for purchasing is the actual pore specification and the independent test data behind it, not which category the marketing puts it in.

One thing the particle-size framing consistently obscures: the relationship between membrane filtration and disinfection. They are complementary, not substitutes. MF removes organisms physically — they accumulate on the membrane, and when you backwash or replace the cartridge, they leave the system. UV disinfection inactivates organisms without removing them — they stay in the water, dead, and pass through. A system with both removes and inactivates: more robust against both membrane bypass and UV shadowing by particles. Neither one alone is as reliable as both together on genuinely risky source water.

MF is a real microbiological barrier — for the organisms in its range. Test your water first and find out whether a biological barrier is actually what you need before you buy one.