Ultrafiltration is the workhorse of point-of-use water treatment — a 0.1-micron membrane that physically strips out bacteria and protozoa with no power and no chemicals. But UF alone is rarely the whole answer, and the “standard train” you read about is a trap. The right build depends on the water and the community. There is no one-size-fits-all.
What ultrafiltration does — and what it doesn’t
An ultrafiltration membrane is a bundle of hollow fibers riddled with pores about 0.1 microns across — small enough that bacteria and protozoa physically cannot pass. Water is pushed or pulled through the fiber wall; the pathogens stay behind and are periodically flushed out. No electricity, no dosing, no consumable chemicals in the simplest configuration.
Read the bottom of that box carefully. A single cartridge rated for 100,000+ gallons, costing a few dollars, works out to pennies per person per year — which is exactly why UF is our default filter for dispersed households. But read the middle of the box too: UF is a physical sieve. It does not reliably remove viruses, and it does nothing for dissolved problems — salts, taste and odor, hydrocarbons, or heavy metals pass straight through. Everything else in this brief is about closing those specific gaps, only when the water and the community actually require it.
It scales further than people expect
The same membrane chemistry spans an enormous range, and the scale is the first design decision:
Household. A single gravity unit that makes non-potable municipal or catchment water safe to drink — the size we deploy in dispersed, off-grid settings.
Community. A pressurized kiosk or school system serving hundreds, with an operator and a maintenance routine.
Industrial. Skid-mounted membrane banks serving tens or hundreds of thousands — the same physics, engineered for continuous municipal-scale flow.
The combinations — and when each one earns its keep
Four add-ons come up again and again. Two protect the membrane; two extend what the water is treated for. None is universal.
| Add to UF | What it adds | When it earns its keep | Cost / complexity | What it does NOT solve |
|---|---|---|---|---|
| Prefilter | A cheap, washable barrier that stops silt before it reaches the membrane. | Turbid or seasonal surface water; anywhere fouling drives downtime. | Low — a sponge or screen you rinse. | Doesn’t treat the water; it protects the filter. |
| Chlorine cleaning | Periodic chlorine backwash that clears biofilm and restores flow. | Warm, biofilm-prone water; community and industrial scale. | Low — small chlorine dose plus a routine. | Not continuous disinfection on its own. |
| UV disinfection | A second barrier that also inactivates viruses and covers a membrane breakthrough. | Communities used to treated water (see below) with reliable power. | Higher — power, lamp replacement, upkeep. | No help with chemicals; adds failure modes. |
| Activated carbon | Adsorbs taste & odor, hydrocarbons and organic chemicals, and some heavy metals. | Chemical (not just microbial) source problems. | Moderate — media that saturates and is replaced. | Not a disinfectant; finite capacity. |
Protecting the membrane: prefilter and chlorine cleaning
The UF membrane is the most expensive part of the system, so the cheapest engineering win is to protect it. A prefilter — something washable that costs a fraction of the membrane — catches silt and grit before they blind the fibers, which cuts cleaning frequency, downtime, and the risk of an early membrane replacement. Chlorine cleaning follows the same logic against a different enemy: a periodic chlorine backwash breaks down the biofilm that warm water grows on the membrane, restoring flow and extending life. Neither is treating the water for the drinker — both are protecting the asset so the real filter keeps working.
The second barrier: when UV is worth it — and when it isn’t
UV is where most designs go wrong, because the honest answer depends on the people, not just the pipes. Adding UV after UV gives dual protection: if the membrane is ever compromised, UV catches what slips through, and it covers viruses that UF misses.
That second barrier is genuinely worth its cost for a community already accustomed to treated water — a piped town in the Philippines, for instance. Their immune systems are not adapted to a sudden pathogen exposure, so a rare membrane breakthrough is a real health event, not a bad week. These are also, painfully, the communities that struggle to pay for treatment and food at the same time — but here the second barrier is the responsible engineering call.
For a community coming off untreated water — rain, a river, an open source — the calculus flips. UF alone is already a massive step up, and their exposure baseline means a rare breakthrough is not a step down. Bolting on UV buys little and costs a lot: consumable lamps, electrical complexity, more maintenance, and new ways for the system to fail silently when a bulb burns out. Here the added cost and failure modes outweigh the benefit, and UF alone is the better system.
The premium build isn’t the best build. The best build is the one that fits the water, the immune baseline, and the budget in front of you.
Treating chemistry: activated carbon
When the problem isn’t just microbes — a fuel taste, an industrial solvent, a heavy-metal load — UF won’t touch it, and neither will UV. Activated carbon adsorbs taste and odor, hydrocarbons and many organic chemicals, and some heavy metals, making it the right partner where the source has a chemical signature. It saturates over time and has to be replaced, so it is specified when the water demands it, not by default.
Start with the source water. Turbid? Add a prefilter. Biofilm-prone at scale? Plan chlorine cleaning. Chemical taste, solvents, or metals? Add activated carbon. Microbially unsafe but chemically clean? UF may be all the treatment you need.
Then weigh the community. Coming from treated water with reliable power — add UV for dual protection. Coming from untreated water — keep it to UF and put the saved money into reach and upkeep.
Then size it — household, community, or industrial — to the number of people and the flow they need.
Clean Water Help in the field
This isn’t abstract — our two drinking-water programs land on opposite sides of the UV decision, on purpose. In the DR Congo, displaced families are moving off contaminated open sources, so we deploy gravity UF alone: no power, no consumables, pennies per person. In the Philippines villages, communities are used to treated municipal water, so those systems pair UF with UV for the second barrier their situation warrants. Same organization, same core filter, two deliberately different builds — because the water and the people are different.
Sources & notes
- UF performance and cost figures: Clean Water Help field assessments, cross-checked with WHO, US EPA, and CAWST guidance; summarized in the Technical solution tables.
- Field examples: DR Congo gravity-UF program and Philippines village UF+UV systems.
- Figures are approximate and vary with source water, scale, and local prices; we publish verified per-project costs on each project page.
Common questions
Bacteria and protozoa — physically, with no power or chemicals. It does not reliably remove viruses, dissolved chemicals, salts, taste and odor, or heavy metals; those need an added barrier.
No. UF+UV is worth it for communities used to treated water, whose immune systems aren’t adapted to a breakthrough. For communities coming from untreated water, UV’s consumables and failure modes usually outweigh the benefit — UF alone is better.
Both protect the expensive membrane rather than treating the water — a washable prefilter stops silt, and periodic chlorine cleaning clears biofilm. They cut maintenance and extend membrane life.
Related
A village filter can cost as little as a few hundred dollars.
We design each system to the water and the community — no more, no less — and every dollar goes to the field.