The fluorine compounds liberated during the acidulation of phosphate rock in the manufacture of phosphoric acid and fertilizers are now rightly regarded as a menace, and the industry is now obliged to suppress emissions of fluorine-containing vapours to within very low limits in most parts of the world.

As with any pollution control operation, it is highly desirable for the operator of the fluorine scrubbing operation to find a use or market for the recovered fluorine to help defray at least partially the cost of the operation.

This article reviews the chemical and technical principles of gaseous fluorine compound removal, the principal types of practical fluorine recovery processes that have been developed and their limitations, and possible methods of utilizing the fluosilicic acid solution which these processes generate.

Most phosphate rocks mined today contain an average of 3-4% fluorine. When they are processed to phosphoric acid (the basic material from which a variety of fertilizers are manufactured) fluorine compounds appear at various process stages. For the purposes of this review the volatile fluorine compounds HF and SiF4 are of prime interest, as they can be separated relatively easily from the reaction vapours during the acidulation or concentration by scrubbing with water or dilute fluosilicic acid. Many authors have dealt with the processes for and problems of fluorine recovery from wet-phosphoric acid in the last decade. Besides a few review articles, (67-75) the publications refer to separation techniques for fluorine compounds, e.g. precipitation, (1-16) solvent extraction, (27-35) ion exchange (35-40) and volatilization. (41-66).

In the past, little attention was paid to the emission of gaseous fluorine compounds in the fertilizer industry. But today fluorine recovery is increasingly necessary because of stringent environmental restrictions which demand drastic reductions in the quantities of volatile and toxic fluorine compounds emitted into the waste gases. These compounds now have to be recovered and converted into harmless by-products for disposal or, more desirably, into marketable products. At the same time, the expected depletion of natural fluorspar reserves, the main source of fluorine compounds, within the next 2-3 decades increases the importance of fluorine recovery from phosphate rock. As phosphate rock reserves are guaranteed until the end of the next century (78) silicon tetrafluoride or fluosilicic acid might well become the most important source of fluorine for the chemical industry.

Only part of the fluorine contained in phosphate rock is economically recoverable with today’s technology. In the course of wet-process phosphoric acid production by sulphuric acid attack (dihydrate and hemihydrate processes) 45-60% of the fluorine is released in gaseous, recoverable compounds, 30-45% of the fluorine precipitates in the gypsum in solid compounds while 5-10% remains as an impurity in the acid. During single or triple superphosphate production, the portion of volatile compounds diminishes to about 10-25%.

Fluosilicic acid recovered by scrubbing these volatile compounds could in future become the primary raw material for chemicals such as aluminum fluoride and cryolite – auxiliaries indispensable in Hall-process aluminum smelting – or hydrofluoric acid and others which, until now, are normally produced from natural fluorspar. Even synthetic fluorspar can be obtained for use as flux in steel making.

(NOTE FROM FAN: It should be noted while reading this article that the phosphate industry has tried, but has so far been unsuccessful, in trying to convert fluosilicic acid into the main raw materials for industrial fluorine chemicals, e.g. hydrofluoric acid and synthetic fluorspar. According to the Tampa Tribune: “Even though 600,000 tons of fluorine are contained in the 20 million tons of phosphate rock mined in Florida, the fluorine market has been inaccessible because the fluorine is tied up with silica, a hard, glassy material.” A chemical engineer in the phosphate industry, who members of FAN spoke with in the summer of 2001, confirmed that this problem still exists, as the industry hasn’t yet been able to separate out, in a commercially viable way, the silica from the fluoride.)

Fluorine recovery

During the production of phosphoric acid from fluorapatite (3Ca3(PO4)2CaF2) and a strong mineral acid, the calcium fluoride present in the rock is converted, by reaction with the silica also present, into fluosilicic acid according to the following equations:

CaF2 + 2H+ (H2SO4, HNO3, H2PO4, HCl) >
(1) 2HF + Ca++
(2) 4HF + SiO2 > SiF4 + 2H20
(3) 3SiF4 + 2H2O > 2H2SiF6 + SiO2

The hydrogen fluoride and silicon tetrafluoride are partly evolved directly as vapours and partly form fluosilicic acid which, under the influence of heat, decomposes again into volatile SiF4 and HF, leaving the reaction vessel together with the water vapour.

As the heat of the reaction evolved in the attack stage is much less than that required for evaporation, the major portion of the volatile fluorine compounds is obtained during subsequent concentration of the phosphoric acid.

In the production of single and triple superphosphate or weak (28-32%) phosphoric acid, silicon tetrafluoride is preferentially volatilized because under the conditions prevailing its vapour pressure is higher than that of hydrogen fluoride. As the phosphoric acid is concentrated up to 54% P2O5, more and more hydrogen fluoride escapes. The molar ratio HF:SiF4 in the vapours increases sharply with the concentration of the phosphoric acid and surpasses 2 when the acid concentration is 50% P2O5 or more. At molar ratios below 2, reaction (3) will take place when the vapours are scrubbed, and surplus silica will be precipitated in the scrubber liquor, an effect that has to be considered when designing equipment for fluorine recovery.

There are two distinct basic types of process in use:

– fluorine recovery under atmospheric pressure (as used in single and triple superphosphate and weak phosphoric acid production)

– fluorine recovery under vacuum (used in the concentration of phosphoric acid from 30%-50% P2O5 and in evaporative cooling of reaction slurry during phosphoric acid production)

Typical descriptions of the two process types are given below. The second is of greater importance, as it represents the larger recoverable fluorine source.

Fluorine recovery at atmospheric pressure

The gases (mainly silicon tetrafluoride) extracted from the reaction vessel are fed to a venturi scrubber in which the silicon tetrafluoride is absorbed, forming fluosilicic acid and silica (Fig. 1). The scrubbing liquid is dilute, circulating fluosilicic acid. To increase scrubbing efficiency (up to 99%) two or more units are placed in line. Dust can be eliminated first, if necessary, in a special scrubber. Precipitated silica must be removed from the product, for example by filtration. The concentration of the formed fluosilicic acid depends on the use to which it is to be put; normally it is maintained at between 18 and 25%. The higher the concentration of the acid, the lower the washing efficiency.

Fluorine recovery under vacuum

The superheated vapours from the flash vessel of the phosphoric acid concentration plant first pass through a high-efficiency entrainment separator. This is essential to reduce the P2O5 contamination of the vapours, and thus the product, to a minimum; this is particularly important if the product fluosilicic acid is to meet the purity specifications demanded for certain of its uses. The collected mixture of dilute phosphoric and fluosilicic acid is sent back to the concentration unit and thus does not represent a loss of either fluorine or P2O5. The cleaned vapours are then fed to a fluorine scrubber, where the silicon tetrafluoride and hydrogen fluoride they contain are absorbed using circulating fluosilicic acid as the scrubbing liquor. Fluosilicic acid (18%-25%) is withdrawn continuously under density control and the corresponding amount of water is introduced into the system. (Fig 2) For economic reasons, it is desirable to achieve the required fluorine recovery with one scrubber stage only. However, this depends on various factors which need to be carefully investigated before the final decision is made.

Whereas the attainable fluorine recovery largely depends on the fluorine content of the incoming vapour as well as the concentration and the temperature of the fluosilicic acid produced, the P2O5 content of the fluosilicic acid is mainly dependent on the P2O5:F ratio in the vapours from the flash vessel and on the efficiency of the P2O5 separator. Figure 3 shows the fluorine recovery efficiency versus the fluorine content of the vapours for a single-stage scrubbing unit for different concentrations of circulated fluosilicic acid. From this it is quite clear that a high fluorine recovery cannot be achieved with a single-stage unit when a high fluosilicic acid concentration is required and at the same time the fluorine content of the vapours is low. In that case, a second scrubbed stage would be necessary.

Figure 4 shows the P2O5 contamination of the fluosilicic acid in relation to the P2O5 content of the vapour, expressed as the P2O5:F ratio for different fluosilicic acid concentrations, based on a constant fluorine level in the vapours and a given efficiency of the P2O5 separator of 98%. The P2O5 impurities of a 25% fluosilicic acid in this case are almost twice as high as for an 18% fluosilicic acid.

Figure 5 is similar to Fig 4 but it indicates the influence of the fluorine content of the vapours for a given H2SiF6 concentration.

From this it follows that, for an existing installation, neither the efficiency of the fluorine absorption unit nor the P2O5 content of the fluosilicic acid is constant. They depend rather on the type of phosphate rock processed as well as on the actual operating conditions of the phosphoric acid and concentration plant.

Direct uses of fluosilicic acid

Fluosilicic acid has only limited applications for direct use but it can be used advantageously as a raw material for the production of, for example, aluminum fluoride and cryolite; this will be described later. Its direct use is restricted because of its low concentration and the relatively high amount of impurities, as shown below for a typical acid composition:

H2SiF6 18-25%
P2O5 100 ppm
Fe2O3 70 ppm
SO4 1,000 ppm
Cl 1,000 ppm

The main characteristics of fluosilicic acid are its bactericidal and fungicidal effects, because of which there is some direct use as a sterilizing and impregnating agent in breweries and for wood protection. Today, some attempts have been made, mainly in the United States, to fluoridate drinking water with up to 1 ppm F using fluosilicic acid or its salts. (80)

Pure silicon tetrafluoride is not isolated on an industrial scale because of the great expense of doing so. Only one process is described in the literature (the Ochrate process) for direct uses of SiF4 in which dry concrete is treated with SiF4 gas to improve stability and abrasion strength. (81)

Disposal as a waste

The strong and poisonous fluosilicic acid has to be converted into inert and harmless waste products if no suitable application exists. Small plants, especially, are often confronted with the problem on economic grounds. They prefer to neutralize the acid, for example with limestone or milk of lime, to precipitate the acid as a mixture of calcium fluoride and silica.

The precipitated solids are filtered off and removed as a waste product, sometimes together with gypsum from the phosphoric acid plant. The neutralization has to be closely controlled to avoid problems in settling and filtration. However, it is difficult to achieve complete neutralization, and therefore small amounts of poisonous fluorine compounds are still found in the effluent.

Use in the production of fluorine compounds

There are various ways of using fluosilicic acid as a raw material to produce essential fluorine-containing materials on an industrial scale.

Aluminum fluoride

Aluminum fluoride and cryolite are used to reduce the melting point of alumina (forming an eutectic mixture) in electrolysis plants producing aluminum metal. Normally about 20-30 kg aluminum fluoride and about the same amount of cryolite are consumed per tonne of aluminum, depending on the specific process conditions. The P2O5 content of these flux materials should be as low as possible in order to minimize losses of electrical energy.(125)

The classical route for producing this indispensable auxiliary of the aluminum industry is from hydrogen fluoride and aluminum hydroxide; the modern processes using fluosilicic acid (82-117) are divided into the acid and the ammonia process. The acid process, especially the one developed by Chemie Linz, (118-121) is of greater significance, having been in industrial use since 1962. According to this process the required quantities of aqueous fluosilicic acid and aluminum hydroxide are mixed in a reaction vessel. At the boiling point and by careful control of distinct process conditions, the following reaction takes place:

H2SiF6 + 2Al(OH)3 + 2H2O > 2(AlF3 3H2O) + SiO2

The trihydrate crystallizes very slowly and therefore the precipitated silica is separated first from the quasi “metastable” solution. The filtrate is then discharged to a batch crystallizer, where the precipitation of the trihydrate is completed within several hours with the aid of some seed crystals. The separated trihydrate is converted into pure AlF3 (97%) by calcination at 550 C.

A variant of this process has been developed by Derivados del Fluor, (122) while Bayer (123) proposed that the reaction should be carried out at elevated temperature and pressure to form a water-depleted product, AlF3 H2O directly.

The ammonia treatment of fluosilicic acid results in a solution of ammonium fluoride in the first step, which after separation of the silica, is converted first to ammonium cryolite by addition of partly calcined aluminum hydroxide and subsequently into pure AlF3 (Mekog-Albatros process). (124)

The ammonia is recycled.


There are no significant differences between the various processes for manufacturing cryolite. (126-147) IG-Farben was the first to develop a process in its factory at Oppau in 1940. It was based on neutralization with ammonia and treatment with sodium aluminates.

This fundamental process was modified in many ways, for example to improve filtration of silica (148) or to minimize the impurities in the cryolite. (149)

According to a suggestion of VEB Stickstoffwer Piesteritz, (150) ammonium fluoride, formed by the neutralization of fluosilicic acid with ammonia, can be converted into to cryolite by reaction with sodium hydroxide and then aluminum fluoride.

Chemie Linz has developed a process to neutralize fluosilicic acid in different reaction vessels with aluminum hydroxide and soda ash, forming aluminum fluoride and sodium fluoride solutions, which after separation of the precipitated silica, react to give cryolite.

Instead of soda ash, caustic soda can be used.

Other routes use fluosilicates as an intermediate product, for example, the process of Kaiser Aluminum, (151) Montedison, (152) and Onoda. (153) The Kaiser Aluminum process has been used in the United States for more than ten years. However, a major disadvantage of this process is the dilute hydrochloric acid by-product.

Hydrofluoric acid

To produce hydrofluoric acid from fluosilicic acid, a number of processes have been developed, (154-174) but none has so far been used industrially. According to their principles, five groups of processes can be distinguished.

a) Fluosilicic acid is decomposed, by the action of concentrated sulphuric acid, into the gaseous components of hydrofluoric acid and silicon tetrafluoride. Hydrofluoric acid is separated from the sulphuric acid solution by means of distillation. Processes of this kind have been developed both in the U.S.S.R. (175) and by the Tennessee Corp. (178)

b) Another suggestion (179) refers to the thermal decomposition of fluosilicic acid. Because of its higher vapour pressure, silicon tetrafluoride is evaporated preferentially and the water solution is enriched with hydrofluoric acid, which is purified afterwards by distillation.

c) Ammonium fluoride solution, prepared from fluosilicic acid and ammonia, is converted into ammonium hydrogen fluoride by means of evaporation. This component reacts with sulphuric acid forming hydrofluoric acid.

d) A quite different separation principle comprises using the better solubility of hydrofluoric acid in organic solvents (for example polyether) during the evaporation of fluosilicic acid. (181)

e) Synthetic fluorspar made from fluosilicic acid may be used in place of the natural mineral in sulphuric acid attack.


As for hydrofluoric acid, much research work has been done to develop processes for the production of synthetic fluorspar from fluosilicic acid, although no industrial-scale application has been described to date. The number of publications increased in the last years as a result of the expected shortage of natural fluorspar reserves, and the promising perspectives for the use of a mixture of calcium fluoride and silica as a fluorspar substitute in steelmaking. (182) Finally, pure synthetic fluorspar can be used as a raw material for producing hydrofluoric acid, the basic compound of the fluorine industry. Unfortunately, this process route is not yet economic.

The neutralization of fluosilicic acid with limestone or milk of lime is the main principle of fluorspar production. (182-187) Normally, the calcium fluoride and silica are precipitated together but, under certain process conditions, silica remains metastable in the solution. Alternatively, silica can be precipitated first by using the reaction between fluosilicic acid and ammonia to form ammonium fluoride, which is afterwards converted into calcium fluoride.


These components can easily be produced by treating fluosilicic acid with salts like calcium chloride and potassium chloride because of their low solubility in water.

Though their direct use is limited to some applications in disinfectants, fluosilicates can serve as raw material for the production of other fluorine compounds, as has been described.

Prospects for fluorine recovery

More than 100 million tonnes of phosphate ore are consumed annually, from which approximately 1.2 million tonnes of fluorine could be recovered and converted into essential fluorine compounds. (198) The future development of fluorine recovery can be considered optimistically because of the increasing environmental responsibility and positive perspectives in aluminum production. (199-201) Nevertheless, fluorine recovery and recycling in the aluminum industry itself have to be taken into account, which reduce the specific fluorine consumption. (202) However, as this applies mainly to the recovery of fluorine in the form of cryolite it is very likely that the specific consumption ratio of cryolite to aluminum fluoride will change in favor of aluminum fluoride.

Note: This online version of this article does not contain the lengthy list of references, nor the diagrams and all of the chemical equations that are contained in the original. To learn more about the phosphate fertilizer industry, click here.