Direct titrimetric determination of fluoride in natural waters.

The fluorine content of drinking waters has become a matter of great interest to those concerned with the control and diminution of dental caries. In the U.S.A. many are of the opinion that the epidemiological studies of Dean (1938) and his co-workers (Dean, Arnold & Elvove, 1942) provide convincing evidence for the belief that, within certain limits, there is an inverse relationship between the fluorine content of drinking water and the incidence of dental caries in children. Mackenzie (1952) has given a good general account, including pertinent references, ofthe evidence put forward in support of this thesis. The result of such convictions has been the addition ofsodium fluoride to the water supplies ofa number ofAmerican communities. The relevant authorities of this country have this matter under consideration. It is essential in all these investigations to have reliable estimations of the fluorine content of the water, whether the fluorine is naturally present or artificially added. The existing methods for the determination of small amounts of fluoride fall into two classes: (1) colorimetric methods, which use the bleaching action of the fluoride ion on a zirconium alizarin lake and require a set of fluoride standards for comparison, (2) titrimetric methods in which the fluoride is directly titrated with thorium nitrate in the presence of alizarin or a solochrome dye as indicator. The theoretical bases for these methods are discussed by Talvitie (1943); both types are sensitive to many substances found in natural waters. The method of estimation commonly used is that of Lamar (1945), but it has the disadvantage that the alkalinity of the water needs first to be determined, and that sulphate, phosphate, etc., interfere with the colorimetry. The method applies to waters containing less than 500 parts per million (p.p.m.) S042and 1000 p.p.m. CF, and the suggestion is made that the errors introduced by these two ions cancel one another out. Until recently we used a colorimetric method. Our technique resembled that of Elvove (1933), in that the bleaching was carried out in the presence of 0-4N-hydrochloric acid, which has considerable advantages: we have found that at this high acidity alumninium does not cause interference. Elvove advised a complete analysis ofeach water so that the fluoride standards required for comparison could be made 'similar in composition to the unknown'. His reason for this was the interference caused by PO4-, SO4'and other anions. We overcame this difficulty by the addition of barium chloride. This removed most of the interfering ions and did not itself interfere with the subsequent bleaching action of fluoride on the lake. Fluoride was not precipitated or adsorbed on to the precipitate under these conditions. Although we have had considerable success with this type of method it has the following disadvantages: it is tedious and time-consuming because ofthe number ofstandards required, it is susceptible to temperature and it takes some 6-18 hr. for the bleaching action to be complete. Moreover, the estimation may have to be repeated if the amount of fluoride present falls outside the expected range. It is not proposed here to give a full critical consideration of the numerous titrimetric methods used by others in the past, but to describe our present method for the direct determination in natural waters.We have had long experience in the methods of estimating fluoride both in natural waters and in the distillates from natural waters, rocks, bones and teeth, and have gradually modified and simplified our technique.