💧 Water guide

What an RO Purifier Removes — and the Things It Does Not

8 min readPublished 10 April 2026Updated 1 August 2026

The complete answer table

Contaminant	Removed by RO?	Typical rejection	Notes
Total dissolved solids	Yes	90-98%	The defining capability. Nothing else domestic does this
Hardness (Ca, Mg)	Yes	95-98%	Removes scale-formers, and the useful minerals with them
Sodium / salinity	Yes	90-98%	Why RO works on coastal and brackish supply
Arsenic (As-V)	Yes	95-99%	As-III is removed less well; pre-oxidation improves it
Fluoride	Yes	85-95%	The main domestic option for the fluoride belt
Nitrate	Yes	85-95%	Critical where farming has contaminated groundwater
Lead	Yes	95-99%	Usually a plumbing problem rather than a supply one
Iron (dissolved)	Partly	Variable	High iron should be pre-treated; it fouls membranes
Bacteria	Yes	>99.99%	Physically excluded, not killed
Viruses	Yes	>99.9%	Still far larger than the membrane's pores
Cysts (Giardia, Crypto)	Yes	>99.99%	Orders of magnitude too large to pass
Pesticide residues	Mostly	85-95%	Larger molecules are removed well; carbon helps too
Chlorine	No	—	Removed by the pre-carbon block, which protects the membrane
Dissolved gases (CO₂, H₂S)	No	—	Pass through; H₂S needs separate treatment
Some low-MW VOCs	Partly	Variable	Carbon stages do most of this work

Why it removes microbes without being a disinfection device

An RO membrane's effective pore size is around 0.0001 microns. The smallest viruses are roughly 0.02 microns — two hundred times larger. Bacteria are thousands of times larger. Exclusion is therefore near-total, and it is physical rather than chemical, so nothing has to be dosed and nothing survives in an inactive form.

The reason RO purifiers still carry a UV lamp is not that the membrane is inadequate. It is that a membrane can fail — a pinhole, a compromised seal, a badly seated cartridge — and that contamination can enter the system downstream of the membrane, in the storage tank or tubing. UV is the backstop for both.

The contaminants where the detail matters

Arsenic — RO removes arsenate (As-V) at 95-99%, but arsenite (As-III), the form more common in reducing groundwater, is removed less reliably. Where arsenic is a known problem, use a purifier with a pre-oxidation stage and confirm removal with a lab test on the output, not just on the input.
Fluoride — rejection is typically 85-95%, which brings most Indian fluoride-belt water inside the 1.0 mg/L limit but may not be enough at very high source concentrations. Test the output if your area is severely affected.
Iron — dissolved iron passes partly and, worse, oxidises and fouls the membrane. Water with visible iron staining should be pre-treated with an iron removal filter ahead of the RO, not simply pushed through it.
Hydrogen sulphide — the rotten-egg smell is a dissolved gas that passes straight through. It needs aeration or oxidising media, not RO.
Chlorine — not only unremoved by the membrane but actively destructive to it. The pre-carbon block exists to strip chlorine before it reaches the membrane, which is why skipping that cartridge is the most expensive economy in the whole machine.

What it removes that you would rather keep

RO is indiscriminate. Calcium and magnesium go out with the arsenic, and so do the bicarbonates that buffer pH. Output at 20-30 ppm is mineral-poor, mildly acidic and tastes flat.

The remedy is not exotic. A TDS controller blends a measured amount of pre-membrane water back in, landing the output at 60-100 ppm — enough to taste like water and to retain some mineral content. Mineraliser and alkaline cartridges achieve something similar by dosing minerals in after the membrane.

Worth keeping in proportion: drinking water is a minor contributor to dietary calcium and magnesium compared with food. The taste argument for a TDS controller is stronger than the nutritional one, and both point the same way.

How to know your membrane is still doing its job

A failing RO membrane looks exactly like a working one. There is no noise, no leak and no indicator. The only reliable signal is the output TDS.

Measure the output the week the purifier is installed and write the number down. Measure again every three months. A reading that has climbed substantially — say from 45 ppm to 140 ppm on unchanged input — means the membrane is passing what it used to reject, and it is time to replace it regardless of what the service schedule says.

Tip

Record input and output TDS on every service visit. A technician who will not write both numbers on the slip is one you cannot audit, and this is the single easiest way to tell whether you are getting the service you paid for.

Frequently asked questions

Does RO remove bacteria and viruses?

Yes, very effectively. The membrane's pores are around 0.0001 microns while the smallest viruses are around 0.02 microns, so both bacteria and viruses are physically excluded rather than killed. The UV lamp fitted alongside is a backstop against membrane damage and downstream tank contamination, not a sign the membrane is inadequate.

Does RO remove arsenic and fluoride?

Yes — typically 95-99% for arsenate and 85-95% for fluoride, which is why RO is the recommended domestic technology in India's arsenic and fluoride belts. Arsenite, the form common in reducing groundwater, is removed less reliably, so in a badly affected area have the purifier's output lab-tested rather than assuming.

Does an RO purifier remove hardness?

Yes, at 95-98% — it removes calcium and magnesium along with everything else dissolved. That said, for whole-house scale problems affecting geysers, taps and washing machines, a water softener at the inlet is the right tool. RO treats only the drinking-water line.

Does RO remove chlorine?

The membrane does not, and chlorine actively damages it. Chlorine is removed by the pre-carbon block placed ahead of the membrane, which exists largely for that purpose. Letting that cartridge run past its interval is the most common cause of premature membrane failure.

Does RO remove essential minerals from water?

Yes — calcium and magnesium are removed along with the contaminants, which is why RO output tastes flat. A TDS controller blends some untreated water back in to land the output at 60-100 ppm, and mineraliser cartridges dose minerals back after the membrane. Diet contributes far more of these minerals than water does, so the stronger argument is taste rather than nutrition.