Why fluoride is hard to remove
Fluoride defeats the two technologies most households already own. It is an ion, not a particle and not an organic molecule — so activated carbon, which works by adsorbing organics, does essentially nothing to it, and no mechanical filter has pores small enough to strain it. Every Australian capital fluoridates, so this is not a fringe question: our fluoride filter guide exists because the most common assumption in the category is wrong.
That leaves four approaches: reverse osmosis, distillation, and two adsorptive media — activated alumina and bone char. This review covers the media, because they are the options sold as cheaper, simpler alternatives to RO, and because their real-world performance depends on a variable most buyers never check.
Activated alumina
Activated alumina is processed aluminium oxide, typically supplied as small white or tan spheres. It removes fluoride by adsorption: fluoride ions bind to the alumina surface. The US EPA recognises it as Best Available Technology for fluoride (and arsenic), which is the strongest institutional endorsement any fluoride media holds.
Reported removal efficiency spans an uncomfortably wide range. Retailers commonly claim 98–100%. Industry sources measuring at practical system flow rates report figures closer to 65%, and note that capacity is roughly 1.5% by weight — about 6 lb of fluoride per cubic foot of media. Both can be true: adsorption is contact-time dependent, so a slow drinking-water station and a whole-house flow rate are different propositions entirely.
Bone char
Bone char is carbonised animal bone, largely hydroxyapatite with carbon remnants. Fluoride exchanges into the hydroxyapatite structure. It is cheaper than activated alumina — by as much as twentyfold — and appeals to buyers wanting a less industrial material.
The peer-reviewed picture is genuinely mixed and worth reading carefully. Laboratory work has found bone char capacity 2.8 times greater than a commercial activated alumina in batch tests, and field research in the Ethiopian Rift Valley found some bone chars outperforming activated alumina, requiring smaller doses for 90% removal. But the same literature reports laboratory column results consistently exceeding field results, real-world efficiency around 67%, and performance declining after relatively small throughput volumes.
Two practical cautions follow. Bone char manufacturing is not standardised — performance varies with bone source and peak processing temperature, and studies find meaningful differences between chars produced at 400°C and 1000°C. And bone char carries no NSF certification as a category, where activated alumina at least has EPA recognition behind it.
The pH problem — and why it matters in Australia
This is the section that should decide most Australian purchases, and it is almost never mentioned at point of sale.
Both media are strongly pH-dependent, and both work best in acidic water. Published equilibrium work on activated alumina found fluoride uptake of about 29.8 g F/kg at pH 6, falling to 6.63 g F/kg at pH 8 — roughly a fourfold loss of capacity across two pH units. Bone char is worse: laboratory work found maximum adsorption at pH 3, with capacity falling nearly twentyfold as pH rose from 3 to 12, and practical effectiveness favouring pH 6–7.
Industry assessment puts it bluntly: activated alumina capacity drops by 50% or more once feed water pH reaches 8.2 with bicarbonate alkalinity above 50 ppm — a combination said to describe the large majority of municipal supplies.
Australian supplies sit in exactly the wrong pH band. Utilities deliberately raise pH to protect pipework from corrosion. Perth’s supply zones run from pH 6.9 to 8.6, with most of the metropolitan network between 7.5 and 8.2; Canberra reports around 7.8; Melbourne, Sydney and Brisbane are similar or higher. Check your own zone in our water quality lookup before buying fluoride media — if your pH is above about 8, you are buying a medium operating at a fraction of its rated capacity.
Competing ions compound it. Chlorine and chloride occupy adsorption sites that would otherwise take fluoride, and bicarbonate alkalinity — which tracks hardness — competes directly. In a hard, alkaline zone the medium faces both problems at once.
The three options compared
| Activated alumina | Bone char | Reverse osmosis | |
|---|---|---|---|
| Typical removal | 65–98%, flow dependent | 67–90%, highly variable | 95–100%, consistent |
| pH sensitivity | High — poor above ~8.2 | Very high — favours 6–7 | Low |
| Recognition | US EPA Best Available Technology | None as a category | NSF/ANSI 58 certification available |
| Capacity | ~25× bone char | Low; declines quickly | Membrane life measured in years |
| Media cost | Up to ~20× bone char | Low | n/a — system cost |
| Wastewater | None in service | None | Several litres per litre produced |
| Failure signal | None visible | None visible | None visible |
The bottom row is the one to sit with. None of these tells you when it has stopped working. Fluoride is tasteless, odourless and invisible at treatment concentrations, so exhausted media produces water indistinguishable from treated water. With adsorptive media that is a genuine risk, because capacity is consumed faster in exactly the alkaline, high-alkalinity conditions most Australian households have.
Our verdict on the evidence
For Australian mains water, reverse osmosis or distillation is the more defensible answer, and it is not close. Adsorptive fluoride media are real technology with genuine research behind them — activated alumina is EPA-recognised Best Available Technology, and bone char has peer-reviewed results that in some conditions beat it. But both were characterised largely in the context of high-fluoride groundwater in low-resource settings, often at favourable pH, and Australian treated supply is close to the worst case for them: pH deliberately elevated, bicarbonate alkalinity present, chlorine competing for sites, and starting fluoride concentrations already within guidelines.
If you still prefer media over RO — because you cannot plumb a system, or will not accept the wastewater — the conditions to insist on are narrow: know your pH (below about 8), accept a slow flow rate at a dedicated drinking station rather than whole-house, choose activated alumina over bone char for capacity and consistency, and track throughput or test periodically, because nothing else will tell you when it has stopped. If your pH is above 8, our RO explainer and distillation review cover the two options that are not pH-dependent.
Primary sources: US EPA Best Available Technology designation for activated alumina; peer-reviewed equilibrium studies of alumina fluoride uptake versus pH (6.63, 23.78 and 29.8 g F/kg at pH 8, 4 and 6 respectively); bone char pH-dependence studies showing ~20-fold capacity loss from pH 3 to 12 and capacity 2.8× commercial activated alumina in batch tests; fixed-bed and field column comparisons of bone char and activated alumina including Ethiopian Rift Valley trials; industry assessment of capacity loss above pH 8.2 with bicarbonate alkalinity above 50 ppm. Water Corporation 2024–25 zone pH data. Reviewed August 2026. FilterOut takes no commissions and carries no paid placements — see who to trust for filter advice.
FAQ
Does activated alumina remove fluoride?
Yes, and the US EPA recognises it as Best Available Technology for fluoride. But performance is strongly pH-dependent: published work shows capacity falling roughly fourfold between pH 6 and pH 8, and industry sources report a 50%+ drop above pH 8.2 with bicarbonate alkalinity present.
Is bone char better than activated alumina for fluoride?
Sometimes in the laboratory, rarely in Australian conditions. Batch studies have found bone char capacity 2.8× a commercial activated alumina, but it is even more pH-sensitive (favouring pH 6–7), manufacturing is not standardised, real-world efficiency is around 67%, and it carries no NSF certification.
What is the most reliable way to remove fluoride?
Reverse osmosis or distillation. Both achieve 95–100% consistently and are largely unaffected by pH, unlike adsorptive media. Activated carbon does not remove fluoride at all.