How ultrafiltration works
Ultrafiltration pushes water through the walls of hollow polymer fibres perforated with pores typically between 0.01 and 0.1 micron. A cartridge packs thousands of these straw-like fibres together, giving an enormous filtering surface in a small housing. Anything physically larger than the pore is retained; anything smaller passes through with the water.
This is pure size exclusion — a mechanical barrier, not adsorption or chemistry. That distinction explains everything UF does well and everything it cannot do. It sits between microfiltration (0.1–10 micron) and nanofiltration (0.001–0.01 micron), and well above reverse osmosis, which rejects dissolved ions themselves.
Unlike RO, ultrafiltration needs no high pressure and produces no brine waste. Mains pressure or even gravity head is often sufficient, and essentially all the water you put in comes out the other side. That combination — barrier-grade microbiological performance with no wastewater and no pump — is why UF has spread rapidly through under-sink units, portable filters and rainwater systems.
What it removes — and what passes straight through
| Contaminant | Removed? | Detail |
|---|---|---|
| Bacteria (E. coli, Salmonella) | Yes — strongly | Published UF membranes achieve roughly 6–7 log reduction (up to 99.99999%) |
| Cryptosporidium, Giardia | Yes | Cysts are far larger than the pores |
| Viruses | Partly — depends on pore size | ~4 log (99.99%) typical across the range; tighter membranes (0.01–0.02 micron) do better |
| Sediment, turbidity, colloids | Yes | 98–99.5% turbidity reduction reported on challenging feed water |
| Microplastics | Yes | Orders of magnitude larger than the pores |
| Chlorine and chloramine | No | Dissolved gases pass through — needs carbon |
| PFAS, pesticides, solvents | No | Dissolved molecules smaller than the pores |
| Lead, heavy metals (dissolved) | No | Ionic — needs adsorptive media or RO |
| Fluoride, nitrates, hardness, salts | No | All dissolved ions pass freely |
Read the right-hand column as a whole and the shape is clear: ultrafiltration is outstanding against things that are particles and useless against things that are dissolved. A UF cartridge on its own leaves your water chemically identical — same chlorine taste, same hardness, same fluoride, same TDS reading.
Ultrafiltration vs reverse osmosis
The most common marketing framing for UF is “RO without the waste” or “keeps the good minerals”. That is half true and half misleading, and the difference matters.
True: UF wastes essentially no water where RO commonly discards several litres per litre produced; UF needs no pump or storage tank; UF leaves calcium and magnesium in the water; UF cartridges are usually cheaper to run.
Misleading: “Keeps the good minerals” and “removes nothing dissolved” are the same physical fact described in two different tones. A UF system keeps the calcium and the fluoride, the nitrates, the PFAS and the lead. If your reason for filtering is a dissolved contaminant, UF is not a lower-waste alternative to RO — it is not an alternative at all.
The genuinely correct use of UF is either as a microbiological barrier in its own right, or as pre-treatment ahead of RO, where it strips the particles and bacteria that would otherwise foul the RO membrane — a role it performs well enough that municipal plants use it for exactly that.
The integrity problem nobody mentions
A membrane is only as good as its worst fibre. Because UF works by physical exclusion, a single broken or cut fibre creates a bypass through which unfiltered water flows — and testing has shown that a compromised fibre measurably degrades log reduction performance. Membranes can also be damaged by freezing, by drying out between uses, and by chlorine attack depending on the polymer.
As with UV, the failure is invisible: water from a damaged UF cartridge looks, tastes and flows the same as water from an intact one. Industrial plants run routine integrity tests for this reason. Household users have no equivalent — which is why replacement intervals and physical care matter more here than with a carbon cartridge, whose decline is gradual rather than sudden.
Fouling is the other operational reality. Everything the membrane rejects accumulates on its surface, so flow drops over time. Backwashing recovers some of it; sediment pre-filtration prevents much of it. A UF system on rainwater or bore supply without pre-filtration will clog quickly.
Our verdict on the evidence
Ultrafiltration is a well-proven barrier technology with an unusually clean evidence base and an unusually mis-sold scope. The microbiological performance is real, well characterised in the literature, and achieved without power, pressure or wastewater — a genuinely attractive combination for rainwater, bore water, travel and rentals.
The category’s honesty problem is the framing. Sold as a microbiological barrier, UF delivers. Sold as an RO substitute that “keeps the good minerals”, it invites buyers to believe a filter is addressing chemical contaminants it does not touch. Our practical position: choose UF when your concern is organisms or particles, insist on a stated pore size (0.01–0.02 micron if virus reduction matters to you), pair it with carbon if you also want chlorine or taste addressed, and treat it as the wrong tool entirely for fluoride, PFAS, nitrates or hardness.
Primary sources: published characterisation of hollow-fibre UF membrane performance (6–7 log bacteria, ~4 log virus reduction across 0.01–0.1 micron; 98–99.5% turbidity reduction); NSF International evaluation of UF module virus reduction at varying flux and with cut fibres; peer-reviewed hollow-fibre ultrafilter studies on simultaneous concentration of enteric bacteria, protozoa and viruses. Reviewed August 2026. FilterOut sells nothing and takes no commissions — see who to trust for filter advice.
FAQ
Is ultrafiltration better than reverse osmosis?
Neither is better — they do different jobs. UF is a physical barrier to bacteria, cysts and particles with no wastewater or pump. RO additionally removes dissolved contaminants such as fluoride, nitrates, PFAS and salts, which UF does not touch at all.
Does ultrafiltration remove chlorine or PFAS?
No. Both are dissolved and pass straight through the membrane. UF is routinely paired with activated carbon for chlorine and taste; PFAS requires carbon rated for it or reverse osmosis.
Do ultrafiltration filters remove viruses?
Partly, and it depends on pore size. Published membranes achieve around 4 log (99.99%) virus reduction across the standard range, with tighter 0.01–0.02 micron membranes performing better. Ask for the stated pore size rather than accepting a general claim.