Defluoridation refers to methods of water treatment that reduce the concentration of fluoride in the water, normally, in order to make it safe for human consumption. Some water treatments that have the capacity of reducing the fluoride concentration along with most other anions, or anions and cations, in the water, are not considered as defluoridation methods. Thus general demineralising methods like distillation, reverse osmosis, electrodialysis and resin de-anionisation, which are able to remove fluoride fully or partly from the water, are not considered as defluoridation methods. On the other hand methods that only remove fluoride without any addition or reduction of other parameters are not yet discovered. That's why the expression “fluoride removal” lacks precision. Defluoridation is used to characterise methods that reduce the fluoride ion specifically, without major other changes to the quality of the treated water.


Defluoridation of water differs from normal piped water treatments:

The objective of defluoridation is to prevent or to mitigate endemic fluorosis.










Defluoridation, cf. text box, is only required in small scale, mostly in rural areas in developing countries. As such, many defluoridation approaches have been launched, proclaiming success, without unbiased field proof and thorough optimisation. Further confusion comes with the fact that the process or the technical set-up that may work in one context of socio-economic and environmental conditions may fail in another. Local availability and acceptability of the required materials, fluoride contamination level and water quality are major factors to be considered when selecting the process and the design that minimises the capital and running costs.

Defluoridation technology has to be simple, affordable, reliable and operational at least at three different levels







Defluoridation is carried out in one of three different processes:



A sorption process is normally designed in a plug flow filter column, including a medium that has a certain capacity of absorption, adsorption or ion exchange of the fluoride. This process requires recharge or regeneration of the medium upon saturation. Numerous media are known to have defluoridation properties, cf. Table.


Some natural media capable to sorb fluoride






Acidic clay

Fuller’s earth





Alkalkine soil

Kaolinitic clay

Diatomaceous earth





Aiken soil

China clay

Ando soil


However, due to capacity, availability and subsequent water quality limitations only bone char and activated alumina are worth mentioning. The plants shown in attached figures are based on bone char sorption, which is the process of choice if acceptable by the community.

Alternatively, similar set-ups can be utilised in the activated alumina process. This is often preferred e.g. in strictly vegetarian societies that consider the use of (cow) bone char as unethical.




Text Box:    The EnDeCo Domestic Defluoridator is an example of a field-tested and optimised unit. It can treat 9000 L of water containing 5 mgF/L, suitable for one family and 1 yearly recharge.    Text Box:    The EnDeCo Institutional Defluoridator. It can be made at any desired capacity, normally in the range >20 m3 for a raw water fluoride of 5 mg/L.


Sorption design

The sorption process may be designed on basis of daily water demand, capacity of the medium, the raw water contamination level, proposed length of operation period and the bulk density of the medium. The required column treatment capacity and the amount of medium required in a filter column are calculated, eq. 1 & 2. From the bulk volume of the medium, eq.3, the dimensions of the column are derived, eq. 4 & 5:










D is daily water demand, L/day.

C is capacity of the medium, g/kg.

F r is raw water contamination level, mgF/L.

F t is treated water concentration, (0.5) mgF/L.

P is proposed length of operation period, days

k is required column capacity, L.

M is amount of medium required in a filter column, kg.

V m is bulk volume of the medium, L.

d is inner diameter of the filter column, cm.

h is height of medium in the column, cm.

H is height of the column, cm.


Unfortunately the capacity of the medium, and its purity, is subject to batch variation. Further, it depends on the fluoride concentration in the raw water. In practice however, the capacities of 4, 1 and 0.03 g/kg for good qualities of bone char, activated alumina and activated clay respectively.




Text Box:    The Nalgonda process as optimised for use in the Rift Valley.
A co-precipitation process is often carried out in batch containers, where the precipitating chemicals are totally mixed with the raw water. This process requires filling of water and chemicals, mixing and settling and subsequent withdrawal of the treated water and the produced sludge. In the so-called Nalgonda technique high dosage of alum and lime are used in a coagulation/sedimentation process, where the precipitation of aluminium hydroxide instantly binds a part of the water fluoride.


Nalgonda process design

 In the Nalgonda and similar processes the container used has to have a volume well over the daily usage. This will allow for comfortable mixing in the container as well as include provision for the sludge water loss. A plastic bucket of 20 L is often sufficient for the daily use of a family. About 18 L can be treated at a time, out of which about 16 can be used.

The Nalgonda process design is complicated and must be confirmed empirically. This is mainly because of the co-precipitation that lacks stoichiometry, the great variation in media quality and solubility and the efficiency being dependent on the raw water quality, in particular the fluoride concentration, pH and the alkalinity. The best possible estimation is based on the Freundlich equation:



m is amount of alum required for a daily treatment of a batch, in g.

F r is fluoride concentration in the raw water, in mg/L.

F t is residual fluoride concentration in the treated water, in mg/L.

V b is volume of water to be treated in batch, L.

a is sorption capacity constant, L (1-1/ b ) mg 2/ b g -1.

b is sorption intensity constant, -.

Any resulting pH between 6.2 and 7.6 is adequate. For pH = 6.7 and required residual fluoride between 1 and 1.5 mg/L, a = 6 and b = 1.33. The amount of lime required may be 20-50 % of the alum dosage.

Alum and lime are added simultaneously to the raw water bucket where it is dissolved/suspended by stirring. The operator should stir fast while counting to 60, i.e. about 1 minute, and then slowly while counting to 300, i.e. about 5 minutes. The mixture is left for settling for about one hour. The treated water is then tapped through the cloth into the treated water bucket from where it is stored for daily drinking and cooking. This separation of the water from the sludge is essential in order to avoid escape of the toxic aluminium to the treated water and in order to avoid the detachment of removed fluoride from the aluminium hydroxide flocs during storage.


Contact Precipitation

Text Box:    A Contact Precipitation Unit sat up as a demonstration plant and run for a couple of years. Since it has been operating as a simple Bone Char Community Plant

Contact-precipitatio n is a process where the water is mixed with the precipitating chemicals, being calcium and phosphate ions, flowing into a catalytic filter column. So far this process is only known to operate in a fluoride-saturated bone char, being the catalyst, dosed with calcium chloride and sodium di-hydrogen phosphate. Contact precipitation is considered to be the process of choice in the future, when the local availability of bone/bone char reaches its limit.

Testing or metering

Text Box:      A safe and user-friendly set-up of defluoridation based on sorption, in casu activated alumin. The system, containing 8 kg AA, is commercially available in Scandinavia and declared to be able to treat 2000 L of water containg 5 mg/L.

The co-precipitation technique, even though arbitrary in nature, is reliable in the sense that the same removal is obtained if the procedure followed exactly.

On the other hand the process results in poor removal efficiency and it demands burdensome daily preparation.

The sorption process is less burdensome. It is also able to provide high removal efficiencies, even at high fluoride levels in the raw water. However, the operator has to watch for the breakthrough the column at saturation. As fluoride is oganoleptically neutral, this cannot be done without either regular testing of the treated water fluoride or checking the accumulated water flow.

Testing the fluoride concentration in the treated water used to be impossible without expensive laboratory equipment. Now simple, reliable and cheap test kits are available.

The Figure attached shows a user-friendly set-up for household defluoridation. The secondary cartridge filter ensures that no toxic aluminium can escape with the treated and the water meter ensures that accumulated water flow can be compared with declared capacity.


Written by Eli Dahi, March 2009