Ultrafiltration (UF): Tighter Than MF, Looser Than RO

Who this is for: people who need a biological barrier that reaches into the virus range — beyond what microfiltration can reliably claim. UF pores run 0.01 to 0.1 micron, which puts most viruses within reach and makes UF the tightest membrane you can run without also stripping out dissolved minerals. It does not remove dissolved ions — lead, nitrate, fluoride, hardness — so if your concern is chemistry, UF is not the answer. If your concern is biological and you want more than MF can reliably guarantee, UF is the mechanism to look at.

The Short Answer

Ultrafiltration sits between microfiltration and nanofiltration on the membrane tightness spectrum. Its pore range — roughly 0.01 to 0.1 micron — is tight enough to stop bacteria and protozoa with the same reliability as MF, and tight enough to catch most viruses as well. It is not tight enough to stop dissolved salts, hardness ions, or most small dissolved contaminants — those pass through along with the water.

That combination — reliable biological removal, mineral passthrough — is what makes UF the right membrane for applications where you want biologically safe water without stripping the mineral content that affects taste and long-term palatability. RO removes everything, including the minerals. UF removes organisms and particles, and leaves the water chemistry largely intact.

How It Actually Works

UF membranes are typically hollow-fiber, the same configuration as MF: bundles of thin tubes, water pressed through the tube walls under pressure, rejected material accumulating on the feed side. The manufacturing process for UF membranes is controlled more tightly to achieve the smaller pore size, and the operating pressures needed are somewhat higher than MF — though still far lower than RO, which forces water through a dense membrane that is effectively non-porous.

The key measurement for UF membranes is molecular weight cutoff (MWCO), expressed in Daltons. This is the molecular weight above which a molecule is largely rejected by the membrane — conventionally defined as the molecular weight at which 90% rejection occurs. A UF membrane with a 100,000 Dalton (100 kDa) MWCO passes water and small molecules freely but rejects larger ones. Viruses, depending on type, run from roughly 1,000 to 1,000,000 Daltons — a range that explains why some viruses pass tighter UF membranes and some do not, and why virus removal claims need to specify which viruses were tested and at what log reduction.

Fouling management is the operational core of any UF system. The rejected material — cells, particles, macromolecules — accumulates as a cake layer on the membrane and as internal pore fouling. Backwashing reverses flow to dislodge the cake; chemical cleaning (typically caustic or acid, depending on the foulant) addresses internal fouling. Residential UF systems automate backwashing on a timer or differential pressure trigger; cartridge formats simply get replaced.

Who Actually Needs It

UF’s practical niche is situations where biological risk includes viruses and where you cannot or do not want to strip dissolved minerals:

• Private wells in areas with documented viral contamination risk. Septic system proximity, heavy agricultural runoff, or a community with documented norovirus or hepatitis A transmission through water are all situations where MF’s partial virus coverage is not enough. UF with appropriate virus-removal certification closes that gap.
• Surface water treatment. Streams, ponds, and lake water carry the full biological spectrum. UF is a standard primary treatment membrane in many small community water systems for exactly this reason — it handles the biology so that downstream treatment stages can focus on chemistry.
• Households where immunocompromised individuals need virus protection. HIV/AIDS, active chemotherapy, organ transplant — if someone in the household is in this category, the standard municipal treatment plus a carbon block is often not enough assurance. UF adds a physical virus barrier independent of chemical disinfection.
• Applications where you want biological safety without demineralization. RO would also solve a virus problem, but it removes everything — minerals that affect taste, trace elements, the water chemistry your household is used to. UF removes organisms and leaves dissolved chemistry essentially untouched. For households that have tried RO water and find it flat or off-tasting, UF is worth knowing about.

You probably do not need UF if your water is municipally treated and your concern is dissolved chemistry — lead, chlorine, PFAS, hardness. A carbon block or RO handles those. If your well tests negative for biological contamination, the cost of a UF system is buying protection against a risk you have not documented. Test first.

What It Cannot Do

Dissolved contaminants. This is the same limit as MF, and it is the one most often glossed over. UF pores at 0.01–0.1 micron are far too large to intercept dissolved ions or small molecules. Lead, copper, arsenic, nitrate, fluoride, PFAS, chlorine, hardness — all of these are in solution and pass through UF freely. If these are your contaminants, the mechanism you need is either adsorption (carbon, activated alumina), ion exchange, or RO.

All viruses with certainty. UF catches most viruses under most conditions, but the word “most” is doing real work there. The smallest viruses, at the low end of the 20-nanometer range, approach UF pore sizes. Adsorption to the membrane plays a role alongside size exclusion, which means performance depends on water chemistry, pH, and membrane condition in ways that size-based removal does not. For the highest-risk applications — immunocompromised individuals, documented viral contamination — UF should be paired with UV disinfection rather than relied on alone.

Chemistry-driven problems. Hard water, taste and odor from chlorine or dissolved organics, iron staining — none of these are addressed by UF. The membrane is a biological tool. Layering a UF membrane into a system that does not have a biological problem adds cost and maintenance without adding benefit.

The Honest Tradeoffs

UF membranes cost more than MF cartridges and require more pressure to operate. The tighter the membrane, the higher the operating pressure and the faster the fouling rate for equivalent feed water quality. On clean source water the difference is small; on turbid or biologically active water, the sediment pre-filter ahead of a UF membrane is doing real work keeping the membrane from loading fast.

UF systems that automate backwashing add complexity and require a drain connection. Systems that do not automate put the burden on the user — which works fine if you are paying attention, and fails badly if you are not, because a partially fouled UF membrane that has not been backwashed is still filtering, just at reduced flow and with uncertain performance at the boundary of its range.

The mineral passthrough that makes UF attractive also means you get no benefit against any dissolved chemistry. People who buy UF expecting it to handle hardness, lead, or chlorine are disappointed, because the membrane was not built to do those things and does not do them.

Can You DIY This?

Yes, with the same plumbing competence that serves sediment and MF installation. The additional consideration is the backwash drain line — UF systems that automate backwashing need a place for the reject water to go, which means a drain connection that a point-of-use sediment filter does not need. Sizing matters more than with coarser filters: undersizing a UF membrane for your flow rate means fast fouling and poor performance; oversizing is mostly a cost issue.

Gravity-fed UF filters exist for off-grid use, though they are slower than pressure-fed systems and require the feed reservoir to be high enough above the filter to generate adequate head pressure. For anything beyond light daily use, a pump-fed system is more practical.

How to Buy Without Getting Fooled

• Ask for the MWCO, not just the pore size. MWCO in Daltons tells you what molecular weight is being rejected and at what efficiency. Pore size in microns is a useful rough guide but says less about performance at the boundaries where viruses live.
• Virus removal claims need log reduction data and test conditions. A claim of “removes viruses” or “99.9% virus removal” needs a test standard behind it — NSF/ANSI P231 for portable purifiers, or equivalent independent test data. The standard matters because test conditions (water chemistry, flow rate, challenge organism) dramatically affect results.
• Check what is included in the system. A UF membrane without a sediment pre-filter upstream is a membrane that will foul quickly on any real-world source water. A complete system includes prefiltration. If the product does not mention it, ask.
• Maintenance honesty. A UF system that never needs backwashing or replacement is either very lightly used or not working. Understand the maintenance schedule before you buy and make sure it fits your actual life — the membrane that sits neglected for two years is not protecting anyone.

The Deep End

The distinction between UF and MF is partly pore size and partly the mechanism by which smaller objects are retained. At MF pore sizes, retention is primarily steric — the particle is physically too big to fit through the pore. At UF pore sizes, size exclusion still operates, but adsorption to the membrane surface plays an increasing role, especially for viruses. A virus that is slightly smaller than a UF pore can still be retained if it adsorbs to the membrane material. This is why UF performance against viruses is sensitive to water chemistry — pH, ionic strength, and the presence of organic matter all affect adsorption — in ways that size-based MF retention is not.

It also explains why virus removal efficiency in a UF system can degrade over the membrane’s life in ways that particle removal does not. A fouled membrane with an established cake layer may actually show improved virus removal, because the cake itself acts as an additional barrier. A freshly backwashed or replaced membrane may temporarily show slightly lower virus removal until the cake layer re-establishes. This is well understood in municipal membrane systems where integrity testing accounts for it. In a residential cartridge, you are trusting the manufacturer’s design to account for this, which is another argument for choosing products with rigorous independent testing rather than taking marketing claims at face value.

UF also sits at an interesting position relative to RO in the broader story of home water treatment. RO is the standard recommendation for the highest-purity drinking water — it removes everything, including the dissolved contaminants that UF misses. But the “remove everything” property of RO is also its limitation for some users: the resulting water has a different taste from mineral-rich source water, requires remineralization in some applications, and wastes a fraction of the input water as reject. UF offers biological safety with mineral retention — a meaningful middle ground for households whose water chemistry is acceptable and whose concern is specifically biological. Understanding that this tradeoff exists is what makes it possible to match the mechanism to the actual problem.

UF is a real virus barrier — for the viruses in its range, under the right conditions. Test your water to find out whether biological contamination is actually what you are dealing with before you invest in the membrane to stop it.