Boiling concentrates fluoride rather than removing it. What actually takes it out is a metal oxide surface

A family in a fluoride-affected district does everything they have been told. They boil the water. Some of it evaporates. The fluoride does not — it stays behind in what remains, slightly more concentrated than before. The boiling made the water microbiologically safer and chemically marginally worse.

This is not an unusual failure. It is the standard one, and it happens because almost every water treatment most households know about — and most of what a municipal treatment plant does — was designed for a completely different problem. Boiling, chlorination, candle filters and cloth straining all target things that are suspended or alive: particles, bacteria, protozoa, viruses. Fluoride and arsenic are neither. They are ions in true solution, individually far smaller than any filter pore, chemically indifferent to chlorine, and thermally stable well past the temperature at which the water leaves as steam.

Removing them means removing something dissolved, which is a fundamentally harder job than removing something suspended — and the technologies that do it are quite different from the ones in most Indian kitchens.

The contamination comes from the rock, not from a factory

The first thing that makes these two contaminants confusing is that there is usually nobody to blame. They are geogenic: they dissolve out of the aquifer itself.

Fluoride occurs in fluorite, apatite and several micas, common in granitic and volcanic terrain. Groundwater sitting in long contact with those rocks — especially in arid regions where recharge is slow and evaporation concentrates what is there — picks up fluoride steadily. This is why the affected belts across parts of Rajasthan, Gujarat, Telangana, Andhra Pradesh, Karnataka, Madhya Pradesh and Bihar map onto geology and aridity rather than onto industry.

Arsenic in the Ganga–Brahmaputra alluvial plains has a more interesting and more troubling origin. The arsenic is bound to iron oxyhydroxide coatings on sediment grains, where it is harmless. In deep, oxygen-poor layers rich in organic matter, bacteria respire by reducing that iron from Fe(III) to Fe(II) — and as the iron mineral dissolves, the arsenic it was holding is released into the water. The contamination is therefore produced by ordinary subsurface microbiology, not by a spill.

There is a bitter irony in the history. Millions of tubewells were installed across Bengal, Bihar and Assam precisely to move people off microbially contaminated surface water, and that programme succeeded — it saved an enormous number of lives from diarrhoeal disease. It also moved people onto water from exactly the reducing aquifers where arsenic is mobilised. Solving one problem revealed another that nobody was testing for.

Both contaminants are slow. Fluoride above the limit causes dental fluorosis and, over years, skeletal fluorosis — joint pain, stiffness, and bone deformity that is essentially irreversible. Arsenic causes skin lesions and pigmentation changes, and over decades raises the risk of cancers of the skin, bladder and lung. Neither makes water taste, look or smell wrong, and neither makes anyone ill this week. That combination — invisible, tasteless, and slow — is why the problem persists in places where it has been known about for thirty years.

Indian drinking water standards set an acceptable fluoride limit of 1 mg/L, relaxable to 1.5 mg/L only where no alternative source exists, and an acceptable arsenic limit of 0.01 mg/L, matching the WHO guideline value.

Why arsenic has to be oxidised before anything will hold it

Here is the single most important operational fact in arsenic treatment, and the one most often missed.

Arsenic exists in groundwater in two forms. Arsenate, As(V), is what you find in oxygenated water; at normal pH it carries a negative charge, so it binds strongly to a positively charged metal oxide surface. Arsenite, As(III), is what dominates in the reducing aquifers where the arsenic problem actually occurs — and at neutral pH it is a neutral, uncharged molecule. It has almost nothing for an adsorbent to grab.

The consequence is blunt: an adsorption unit that performs beautifully in a laboratory test with arsenate can remove very little from real reducing groundwater full of arsenite. Any working arsenic removal system therefore begins with an oxidation step — aeration, chlorine, permanganate, or a media that oxidises and adsorbs together — to convert As(III) to As(V) before the adsorbent ever sees it.

A unit sold without an oxidation stage, or installed by someone who does not know why it is there, will underperform in exactly the water it was bought for. Fluoride has no equivalent problem: it is a simple anion in one form, and its difficulty lies elsewhere.

What actually removes them

Adsorption onto metal oxide surfaces is the workhorse for both contaminants, and it is the same physical chemistry in each case: a hydrated oxide surface carries a pH-dependent charge, and the target ion exchanges onto it, displacing hydroxide.

For fluoride, the classic medium is activated alumina. It works well, it is cheap, and it has two demanding conditions. Its capacity depends sharply on pH, peaking in mildly acidic water around 5.5 to 6.5 and falling off in the alkaline groundwater that fluoride-affected regions frequently have. And it is competitive — bicarbonate, sulphate, phosphate and silicate all contend for the same sites, so the capacity measured in clean laboratory water is an optimistic upper bound on what a real borewell will deliver. Regeneration is possible with sodium hydroxide followed by acid neutralisation, which restores most but not all of the capacity and produces a concentrated fluoride waste stream that then has to go somewhere.

For arsenic, iron-based media — granular ferric hydroxide and related iron oxyhydroxides — have a very strong affinity for arsenate and dominate the field. They tolerate a wider pH range than alumina does, which matters in practice. They are usually run to exhaustion and replaced rather than regenerated, and the spent media is arsenic-bearing waste requiring proper disposal, not landfill by the roadside.

Coagulation and precipitation is the other established route. The Nalgonda technique, developed in India, doses alum and lime so that the aluminium hydroxide floc which forms carries fluoride down with it as it settles. It is inexpensive and needs no exotic material. Its weaknesses are equally real: it generates substantial sludge, leaves residual aluminium in the treated water, requires accurate dosing that varies with the raw water, and depends on an operator who turns up every day. Many community plants built on this principle failed for the last reason rather than the first four.

Membrane processes remove both contaminants reliably. Reverse osmosis rejects nearly all dissolved ions including fluoride and arsenic, and nanofiltration does much of it at lower pressure. This is why RO has become the default household answer in affected areas, and it works. It also costs: electricity, a reject stream that is commonly one to three litres discarded for every litre produced, the removal of calcium and magnesium along with everything else, and a maintenance requirement that households frequently ignore. And in an arsenic-affected village, pouring concentrated reject water onto the ground puts the arsenic straight back into the aquifer everyone is drawing from.

Ion exchange works for both, on the same competitive-ion caveat as alumina, and is more common in institutional than household settings.

Where nanomaterials genuinely help, and where they are a sticker

Now the part worth being careful about, because "nano" appears on a great many water products and means something real in only some of them.

The genuine argument is straightforward and it is the standard nanoscale one: adsorption happens at a surface, and dividing a given mass of adsorbent into much smaller particles multiplies the available surface enormously. Nanostructured iron oxyhydroxides, nano-alumina, mixed and doped metal oxides, and various engineered composites do show substantially higher adsorption capacity per gram than conventional granular media in laboratory tests, and a higher capacity per gram means a smaller column, less frequent replacement and less waste.

The reasons that advantage shrinks on the way to a working filter are worth knowing, because they are where most of these materials stop.

• A fine powder cannot be put in a flow-through column. It packs, the pressure drop becomes unworkable, and it washes downstream into the treated water. Real use requires granulating the material or growing it on a support — and both steps bury a large fraction of the surface area that justified the material in the first place.
• Beaker capacity is not column performance. The number quoted in most papers is an equilibrium capacity from a batch test with a single contaminant in clean water. What a field unit is judged on is the breakthrough curve: how many litres pass before the outlet exceeds the limit, in real water containing competing ions, at a real flow rate. The two numbers are related only loosely, and far more of the literature reports the first than the second.
• Regeneration and disposal decide the economics. A medium that adsorbs superbly and cannot be regenerated is a consumable, and its cost per litre treated includes the hazardous-waste disposal of the loaded material.
• Cost per gram is the wrong metric. Cost per thousand litres brought below the limit is the metric, and a cheap conventional medium replaced often frequently beats an expensive advanced one.

Then there are the claims that are simply not about these contaminants at all. Nano-silver in a purifier is an antibacterial agent; it does nothing whatsoever to fluoride or arsenic. Activated carbon blocks, with or without a nano label, adsorb organics, chlorine, taste and odour — not these ions. Graphene oxide membranes and carbon nanotube filters are a genuinely active research area with real promise and are not what is inside a retail purifier today.

The useful question about any water product is never "does it contain an advanced material?" but "which contaminant does it remove, by what mechanism, verified by whose test, and after how many litres does it stop?" Every honest supplier can answer all four.

The step that has to come first

Because every one of these technologies is contaminant-specific, the most valuable thing a household or a panchayat can do costs a few hundred rupees and is skipped almost universally: test the water.

Fluoride and arsenic need different media. Iron and hardness need different treatment again, and high iron will foul an arsenic unit that was not designed for it. Nitrate needs ion exchange or RO. A purifier chosen without a test is a guess, and an expensive guess that removes the wrong thing offers a family the most dangerous outcome available — confidence without protection.

State public health engineering departments, district laboratories, agricultural universities and accredited private laboratories all test water, and the parameters worth asking for in an affected area are fluoride, arsenic, iron, nitrate, total dissolved solids and hardness. Test the source that is actually drunk from, and retest after any change to the well, because groundwater chemistry varies over short distances and with depth — two borewells a hundred metres apart can be genuinely different water.

Why it matters for students and researchers

Water treatment adsorbents are one of the most-published and least-deployed areas in applied materials science, and the gap between those two facts is itself the lesson. Thousands of papers report a novel material with a high equilibrium adsorption capacity for fluoride or arsenic. Very few report a packed column run on real groundwater to breakthrough, with competing ions present, over regeneration cycles, with a disposal route for the spent medium and a cost per cubic metre treated. The second kind of paper is much harder to write and is the kind the field actually needs.

For anyone choosing a project, that gap is an opportunity rather than a discouragement. Column studies on real water, competitive-ion behaviour at realistic concentrations, regeneration over many cycles, granulation methods that preserve accessible surface area, and honest techno-economic analysis are all publishable, all useful, and all under-supplied.

The sociotechnical half deserves equal weight. India's history with community defluoridation plants is largely a history of good chemistry defeated by maintenance, spare parts, dosing discipline and sludge disposal. A treatment that works in a laboratory and fails in a village has not solved the problem, and understanding why is as much a part of the discipline as the isotherm.

Frequently asked questions

Does boiling remove fluoride or arsenic?

No — it does the opposite. Boiling kills pathogens, which is valuable, but it removes water as vapour and leaves dissolved salts behind, so the concentration in what remains goes slightly up. The same is true of any process that evaporates water without capturing the vapour. Only distillation, which condenses and collects the steam, separates them this way, and that is a different appliance entirely.

Will my RO purifier remove them?

A properly functioning reverse osmosis unit rejects the great majority of both, and in an affected area that is a reasonable choice. The conditions are that the membrane is intact and not past its life, that the unit is actually maintained, and that any pre-treatment it needs is present. It is worth knowing that RO also removes calcium and magnesium, wastes a substantial fraction of the input water, and produces a concentrated reject that should not be returned to the ground in an arsenic-affected area.

Is a "nano silver" or "activated carbon" filter any use against these?

Not for fluoride or arsenic. Silver is an antibacterial agent and activated carbon adsorbs organic compounds, chlorine, taste and odour. Both are useful for what they do and neither addresses a dissolved inorganic ion of this kind. If a product claims fluoride or arsenic removal, ask which medium does it and ask for test results showing the outlet concentration after a stated volume.

Why does a treatment unit stop working after a few months?

Because adsorbents fill up. Every gram of medium has a finite number of sites, and once they are occupied the water passes through unchanged — this is breakthrough, and it happens without any visible sign. This is why a unit needs a stated capacity in litres for your water quality, a replacement or regeneration schedule, and periodic outlet testing. A filter that is never tested and never changed reverts to being a piece of plumbing.

If the water looks and tastes fine, is it safe?

Not on this question. Fluoride and arsenic at harmful concentrations are colourless, odourless and tasteless, and their health effects take years to appear, which is precisely why affected communities often go on drinking the water long after the hazard is documented. Clarity and taste report on turbidity and on some dissolved organics; they say nothing at all about these two ions. Only a laboratory test does.