256 JEWSBURY: A RAPID PHOTOMETRIC [Vol. 75 A Rapid Photometric Determination of Silicof luoride in Hydrof luoric Acid, Ammonium Fluoride, Sodium Fluoride and Soluble Coloured Fluorides BY ALAN JEWSBURY SYNoPsrs-The colorimetric silicomolybdic: acid method for the determination of silica has been adapted to the determination of small amounts of silica in hydrofluoric acid and certain fluorides of analytical reagent grade. The interference of fluoride ion is prevented by the addition of boric acid, and the colour is developed by the additic'n of nitric acid and ammonium molybdate. The influence of boric acid, concentration of fluoride ion and the effect of time on the development of the colour, to all of which the method is sensitive, have been studied. Detailed procedures are specified for deter- mining silica in ammonium fluoride over the range 0.02 to 0.8 per cent.; hydrofluoric acid (40 per cent.) 0.01 to 0-056 per cent.: hydrofluoric acid (anhydrous) 0.03 to 1.4 per cent.and sodium fluoride 0-01 to 0.7 per cent. The application of the method to coloured and insoluble fluorides is also described. A determination of silica in hydrofluoric acid and soluble fluorides can be completed in from 15 to 20 minutes. THE methods hitherto available for the determination of traces of silicofluorides in fluorides are not wholly satisfactory. There has been a long-felt need in these laboratories of a quick and accurate method for the determination of silicofluoride in fluorides of analytical reagent grades, e.g., hydrofluoric acid, ammonium fluoride and sodium fluoride.One method that has been proposedl y 2 , 3 , 4 depends on neutralising any free hydrofluoric acid under conditions that prevent hydrolysis of fluosilicic acid, subsequently hydrolysing the fluosilicic acid and titrating the liberated hydrofluoric acid. The first end-point is -9ot very distinct and for small amounts of fluosilicic acid large samples must be taken. Cade5 has proposed a method in which the hydrofluoric acid is evaporated in the presence of sodium chloride. Boric acid is added to convert any remaining sodium hydrogen fluoride to fluoboric acid and then the fluosilicic acid is converted tcl silicomolybdic acid and determined colori- metrically. Substantially the same method was proposed by Dobkina.6 Brabson et a1.' have described two methods for the determination of silica in the presence of fluoride.In the first method, which is applied more particularly to high purity hydrofluoric acid, the fluoride, after careful neutralisation, is converted by boric acid to fluoboric acid and the silica then estimated by development of the molybdenum blue colour and photometric measurement. The second method is intended for larger quantities of silica, 5 to 25mg. of SiO,, and consists in weighing the silica as the oxine salt of silicomolybdic acid. The trace methods of both Cade and Brabson are, however, subject to interference by phosphate and a determination takes about an hour.* EXPERIMENTAL Cade's method was first examined. One gram of AnalaR hydrofluoric acid was evaporated to dryness in a platinum dish with 10ml.of 2 per cent. sodium chloride solution. The residue was dissolved in 10 ml. of water, 10 ml. of 3 per cent. boric acid solution were added, followed by 1 ml. of 5 N nitric acid and 5 ml. of 10 per cent. ammonium molybdate solution. After 10 minutes, the colour was measured on the Spekker photo-electric absorptiometer with Ilford violet 601 filters in 1-cm. cells. A calibration graph had previously been prepared * During preparation of this paper, a communication by Lacroix and Labalade* has described a method for the determination of silica present in various materials including fluorides. In the presence of fluorides an excess of boric acid is added after solution of the material. The solution is then neutrahed with sodium hydroxide until added aluminium is just precipitated ,as hydroxide-this fixes the pH a t 4.6.A known quantity of sulphuric acid is then added to adjust the pH to the optimum value for development of the silicomolybdic acid colour. This method takes longer than the one proposed in this paper and possesses the disadvantage, for trace determinations, of introducing a considerable blank due to the silica in the caustic soda.May, 19501 DETERMINATION OF SILICOFLUORIDE 257 from standard silicate solutions containing sodium chloride and boric acid. The results were subsequently proved to be high and agreement between individual experiments was wanting. The results varied from 0.24 to 0.39 per cent. of 50,. By the method recom- mended below the silica figure was shown to be 0.02 per cent.Attempts were then made to remove the fluoride ion from hydrofluoric acid by adsorption on the synthetic anion exchange resin Amberlite 1R4B. This experiment was carried out in a 100-ml. silver beaker. About 0.5 g. of 40 per cent. AnalaR hydrofluoric acid was diluted to 30 ml. with water and stirred with about 15 g. of the activated resin for 10 minutes by means of a rubber-covered glass stirrer. The resin was filtered off on a Whatman No. 541 filter-paper, washed with about 50ml. of water and the filtrate made up to 100ml. This treatment not only removed the fluoride but led to loss of added silica, present in the form of silicofluoride, by adsorption on the resin. But, as silica in the form of silicate was not adsorbed by the resin, the difficulty was surmounted by removing the hydrofluoric acid and hydrolysing the silicofluoride by means of evaporation in the presence of sodium chloride.The residue from the evaporation was then dissolved in water, treated with the resin and filtered. Boric acid was finally added to form a complex with any remaining traces of fluoride and the silicomolybdic acid colour developed. Reasonable recoveries of added silica were obtained. The results are shown in Table I. TABLE I RECOVERY OF ADDED SILICA AFTER EVAPORATION AND TREATMEKT BY RESIN SiO, added, mg. . . . . 0.7 1.4 1.4 2.8 5.6 9 9 recovered, mg. . . 0.7 1.7 1.7 2-7 5.2 This method took just over an hour for a single determination, and, as a speedier one was desirable, further work was not carried out either to investigate it in greater detail or to discover the cause of interference in Cade's method that necessitated the resin treatment. The method of converting the free fluoride ion into fluoboric acid was next investigated.It was found that it was unnecessary to neutralise the free hydrofluoric acid, as Brabson7 had done, before forming the complex with boric acid. It was necessary for satisfactory development of the yellow colour due to silicomolybdic acid to control carefully the fluoride and boric acid concentrations. After investigating these conditions, methods were devised for determining silicofluoride in hydrofluoric acid, sodium fluoride and ammonium fluoride. The time for a single determination was 15 to 20 minutes. These methods are described in detail below.A method was also worked out for soluble fluorides that give coloured solutions, e g . , nickel fluoride. Time was also an important factor. THE CALIBRATION GRAPH Boric acid does not interfere with the development of the yellow coloured silicomolybdic acid. AnalaR boric acid gives a very small blank, however, and as a 4 per cent. solution proved to be the most satisfactory concentration for forming the complex with fluoride, the colour was developed in boric acid of this concentration. A standard silicate solution was prepared by diluting a weighed amount of sodium silicate of known 50, content so that 1 ml. of the diluted solution was equivalent to 0.7 mg. of 50,. The calibration graph is prepared as follows. To each of seven 50-ml. measuring cylinders add 2 g.of AnalaR boric acid and about 40 ml. of water a t 25" to 35" C. Reserve one solution for a blank and add known volumes (0.2, 0.5, 1.0, 2-0, 3.0, 4.0ml.) of the standard silicate solution to the others and dilute to 50 ml. At 25" to 35" C. add 2 ml. of 5 N nitric acid, stir, and then add 5 ml. of 10 per cent. ammonium molybdate. Plot milligrams of SiO, against drum reading difference on the Spekker absorptiometer using Ilford violet 601 filters and 4-cm. cells, after setting the instrument against water a t drum reading 1.30. In these solutions the colour developed to its full intensity in a minute or two and then remained constant until, and after, the expiration of the 10-minute period allowed before measurement. The influence of temperature within the range of normal room tem- perature is negligible.As the solutions were warmed slightly to affect reasonably rapid solution of the boric acid, the colour was always developed and measured at 25" to 35" C. Stir until dissolved. Stir and allow to stand for 10 minutes.268 JEWSBURY: A RAPID PHOTOMETRIC Fol. 75 DETERMINATION OF SILICA IN AMMONIUM FLUORIDE In order to establish the correct conditions for the determination of silica in ammonium fluoride, a series of experiments was carried out. In the first experiments, 1-g. portions of AnalaR ammonium fluoride were dissolved in water, the equivalent of 243 mg. of SiO, was added to each, together with varying amounts of AnalaR boric acid and the volume made up to 100 ml. at 25" to 36" C. Fifty ml. of this TABLE I1 DETERMINATION OF SILICA IN AMMONIUM FLUORIDE Effect of boric acid with 1 g.of sample HsBOS, SiO, found,* on 2.9 mg., 1 0.5 17.2 2 2.1 72.4 3 2.6 86.2 4 2.4 82-8 6 2.4 82-8 6 2.4 82.8 8 0.8 27.6 Recovery, calculated g- mg. % * Not corrected for the original silica content of the ammonium fluoride which contained 0.1 mg. SiO, per g. solution were taken and the silicomolybdic acid colour developed and measured 5 minutes after mixing, as described in the preparation of the calibration graph. The reasons for taking the reading after 5 minutes are discussed below. The original solution of the fluoride and boric acid was carried out in a 100-mi. silver beaker. Once the fluoborate had been formed the solution could be transferred to glass vessels without risk of the solution attacking the glass.TABLE I11 The results are shown in Table 11. (Platinum or polythene beakers are equally suitable.) DETERMINATION OF SILICA IN AMMONIUM FLUORIDE Effect of boric acid with 0.5 g. of sample Recovery calculated on 2-85 mg., % 96-5 94.7 98.2 * Not corrected for SiO, in the a.mmonium fluoride. With a 1 per cent. solution of ammonium fluoride a recovery greater than 86.2 per cent. of the silica added was not obtained even though relatively considerable amounts of boric acid were used. The recovery increased as the boric acid concentration increased, reached a maximum at about 3 per cent., and then decreased with higher concentrations of boric acid. The experiment was repeated with 0-5 g. of ammonium fluoride and 2.8 mg. of SO,. Varying amounts of boric acid were added before making up to 100ml.The results are shown in Table 111. TABLE IV DETERMINATION OF SILICA IN AMMONIUM FLUORIDE Ratio of sample to boric acid 1 : 8 Total SiO, NHP, SiO, found, present, g. mg.. mg. 0.6 2.8. 2.86 0.8 2.75 2-88 0.9 2.4: 2.89 1.0 0.8 2.90 Recovery, 98.2 95.6 83.0 27.6 %May, 19601 DETERMINATION OF SILICOFLUORIDE 259 These results were more satisfactory; the best being that in which the ratio by weight of ammonium fluoride to boric acid was as 1 : 8. But, as shown by the results in Table IV, in which this ratio was kept constant at 1 : 8, this was not the best ratio over a wide range of concentrations. The results in Tables I1 to IV show that an excess of boric acid alone is insufficient to give satisfactory conditions for the development of the silicomolybdic acid colour.It is evident that there is a maximum concentration of both boric acid and ammonium fluoride for satisfactory silica determinations. In view of this, the concentration of boric acid was fixed at 4 per cent. and experiments were carried out to find the maximum amount of ammonium fluoride that could be taken to obtain good recoveries of added silica. The results of these experiments, in which the amount of ammonium fluoride in 100 ml. of 4 per cent. boric acid containing 2.8 mg. SiO, was varied from 0.5 to 1 g., are shown in Table V. These results show that in 4 per cent. boric acid the ammonium fluoride concentra- tion must not exceed 0.7 per cent., which is the equivalent of 0.36 per cent. of fluoride ion. TABLE V To each solution 2-8 mg.silica were again added. DETERMINATION OF SILICA IN AMMONIUM FLUORIDE Optimum concentration of ammonium fluoride SO, found (corrected NH,F, for 50, in NH,F), Recovery, g. mg- % 1 2-3 82.1 0-8 2-54 90-7 0.7 2-73 97.5 0.6 2-78 99.3 0.5 2.79 99.6 E'ect of time-It had been found during the early experiments that the intensity of the yellow colour due to silicomolybdic acid varied with time in a different way from that resulting when the colour was developed in a solution containing no ammonium fluoride. The colour in the presence of ammonium fluoride developed rapidly at first, reached a maximum for a solution containing much silica after about 4 minutes, remained constant for about 3 minutes and then slowly faded. An example of this is given in Table VI.It shows the rate of colour development of a solution of 0.6 g. of ammonium fluoride in 100 ml. of 4 per cent. boric acid, to which was added 5.6 mg. of SiO,. For solutions containing less silica the time taken to reach maximum intensity of colour was less than 4 minutes, but no fading occurred within 9 minutes. All measurements on the absorptiometer were, therefore, made 5 minutes after mixing. TABLE VI DETERMINATION OF SILICA IN AMMONIUM FLUORIDE Effect of time on the colour Time, min . 2 3 4 5 7 8 10 17 Drum difference reading (1-30 -drum reading) 0.850 0-875 0-883 0.882 0.882 0.878 0.875 0-865 SiO, found, mg. 5.2 5.4 5.55 6-66 5.55 5.5 5.4 5.3 PROCEDURE- In the light of the above results, the procedure recommended for a silica determination in ammonium fluoride is as follows.Dissolve not more than 0.7 g. of ammonium fluoride in about 60 ml. of water in a silver beaker, add 4 g. of AnalaR boric acid, warming slightly, if necessary, to dissolve it. Make up to 100 ml. in a volumetric flask at 25" to 35" C . Take 60 ml., add 2 ml. of 5 N nitric acid, stir, then add 5 ml. of 10 per cent. ammonium molybdate, again stir and after 5 minutes measure the resulting colour on the Spekker absorptiometer.260 JEWSBURY: A RAPID PHOTOMETRIC [I'ol. 75 By this procedure, with 0-7 g. of ammonium fluo.ride, silica can be determined over a range of 0.02 to 0.8 per cent. If the ammonium fluoride contains more silica than 0.8 per cent., less than 0.7 g. should be taken; the remainder of the procedure being as described.Examples of the recoveries obtained by this procedure when silica was added to 0-5-g. quantities of AnalaR ammonium fluoride are shown in Table VII. TABLE TI1 DETERMINATION OF SILICA IN AMMONIUM FLUORIDE By the recommended procedure SiO, added, Yo SiO, found, * Error, Y O Y O 0.056 0.060 + 0-004 0.140 0.140 nil 0.280 0.268 - 0.012 0.560 0.540 - 0.020 0.840 0.830 - 0.010 1.120 1.120 nil * Corrected for silica in the ammonium fluoride (0.05 mg. in 0.5 g. of sample). DETERMINATION OF SILICA IN HYDROFLUORIC ACID It was found to be unnecessary to neutralise the free hydrofluoric acid if the boric acid was added before developing the silicomolybdic acid colour. The fluoborate formed did not influence the intensity of the colour if the conditions were carefully controlled, and the calibration curve already described could be used.Solutions were prepared by diluting a weighed amount of -the hydrofluoric acid solution contained in a silver beaker to about 60ml. with water. Once the added boric acid was dissolved, it was found that the solution could be transferred safely to glass vessels without contamination by silica. The solutions were made up to 100 rnl. and 50 ml. taken for development of the colour, as already described. Preliminary experiments were carried out in which about 0-5-g. quantities of 40 per cent. hydrofluoric acid were taken with 2.8 mg. of SiO, and varying amounts of boric acid. The colour was developed at 25" to 35" C . and measured after 10 minutes. A similar series was carried out with about l 6 g . of 40 per cent.hydrofluoric acid. The results are shown in Table VIII. For these experiments 40 per cent. of w/w AnalaR hydrofluoric acid was used. TABLE 'VIII DETERMINATION OF SILICA IN HYDROFLUORIC ACID Preliminary experiments 40% HF, g. 0.6 0.5 0.8 0.45 0.45 1.5 1.3 1-4 1-5 1-5 . SiO, found," mg. 2.54 2.78 2-84 2.88 2.90 2.2 2.5 2.8 2.86 2-86 * Corrected for the original silica content of the hydrofluoric acid, of 50,. Recovery, YO 90.7 99.3 101.4 102.9 103.6 78.6 89.3 100.0 102.1 102.1 which contained 0.01 per cent. I t appeared that 4 per cent. of boric acid was again a suitable concentration, at least with 2.8 mg. of SiO, present per 100 ml. of solution. So, fixing the concentration of boric acid at 4 per cent., experiments were carried out to find the maximum concentration of hydrofluoric acid that could be used, having regard to the important factor of time, to achieve a satisfactory silica determination. Efect of tiwe and hydrofluoric acid concentration-It had been noticed that time was again a very important factor and this was investigated.The results are summarised in Table I XMay, 19501 DETERMINATION OF SILICOFLUORIDE 261 and show that the time for complete development of colour depends on both the hydrofluoric acid concentration and the silica concentration. If the concentration of silica did not exceed 2-8 mg. per 100 ml., up to 1.4 g. of 40 per cent. hydrofluoric acid could be present for a satis- factory recovery after 10 minutes. If the amount of hydrofluoric acid exceeded this amount, a longer time was required for full colour development.On the other hand, when the silica concentration was 5.6 mg. per 100 ml., then although the amount of hydrofluoric acid present had not exceeded 1.4 g. the full colour did not develop even after 40 minutes. If the colour was to be measured 10 minutes after mixing, then to use the full range of the calibration graph to estimate 5-6 mg. of SO,, not more than 1 g. of 40 per cent. w/w hydrofluoric acid or 0.4g. of the anhydrous acid must be used in 100ml. of 4 per cent. boric acid. This is equivalent t o 0.38 g. of fluoride ion. TABLE IX DETERMINATION OF SILICA IN HYDROFLUORIC ACID Effect of time and acid concentration SiO, found after, minutes* A SiO, r 7 40% HF, H,I30,, added, 3 5 7 9 10 15 25 30 40 g. g. mg. 1-4 4 2.8 - 2.6 2-66 2.7 2.76 2.76 2.76 - - 2.75 - 1.6 4 2.8 0.9 4 5.6 5.46 5.62 - - 5.62 5462 - - - 1.0 4 5 6 4.4 5.1 5.3 5.4 5.5 5.5 5.5 - - 1.4 4 5.6 1.3 1.8 2.5 - 3.2 3.5 - - 4.5 - 2.5 - - - - - ' * Corrected for silica in the hydrofluoric acid. PROCEDURE- On the basis of the above experiments, the method recommended for a silica determina- tion in hydrofluoric acid is as follows.Take not more than the equivalent of 0.4 g. of the anhydrous acid in a silver beaker and dilute to about 60 ml. with water. Add 4 g. of AnalaR boric acid, warming slightly if necessary to dissolve it. Make up to 100 ml. in a volumetric flask at 25" to 35" C. Take 50 ml., add 2 ml. of 5 N nitric acid, stir, then add 5 ml. of 10 per cent. ammonium molybdate, again stir and after 10 minutes measure the colour developed on the Spekker absorptiometer.By this method, using 1 g. of aqueous 40 per cent. w/w hydrofluoric acid, silica can be determined over the range 0.01 to 0.056 per cent. or, calculated on the anhydrous acid, over the range 0.03 to 1.4 per cent. Examples of the recoveries obtained by this method when silica was added to 40 per cent. w/w AnalaR hydrofluoric acid are shown in Table X. TABLE X DETERMINATION OF SILICA I N HYDROFLUORIC ACID Recovery of added silica 40% HF taken, g* 0.6 0-55 0.55 0.6 0.6 1.0 0.9 SiO, added, 0.047 0.127 0.254 0.350 0.467 0-560 0.622 Y O SiO, found,* 0.050 0.127 0.236 0.347 0.473 0.550 0.624 Yo Error, + 0.003 nil % - 0.018 - 0.003 + 0-006 - 0.010 + 0.002 *Corrected for silica in the hydrofluoric acid (0.01%) DETERMINATION OF SILICA IN SODIUM FLUORIDE Four per cent.boric acid was again found to be satisfactory for silica determination in sodium fluoride. Preliminary experiments again showed time to be an important factor. The colour was found to develop much more rapidly than it did in the hydrofluoric acid experiments. I t reached a maximum very quickly and stayed constant for at least 5 minutes, but in less than 10 minutes it started to fade very slowly. In this determination, therefore,262 JEWSBURY: A RAPID PHOTOMETRIC [Vol. 78 all measurements were made 5 minutes after mixing. An example of the effect of time is seen in Table XI. TABLE :XI DETERMINATION OF SILICA IN SODIUM FLUORIDE Effect of time SiO, found after, minutes NaF taken, SiO, present, - A \ 40 lo 4-6 0.8 5.0 4-95 4.95 4.95 4.95 4-86 g.mg. 3 4 6 7 For these experiments a sample of sodium fluoride containing 0.1 per cent. of silica was used. The experiments shown in Table XI1 were carried out to show the maximum amount of sodium fluoride that could be taken in 100 ml. of 4 peg cent. boric acid to obtain satisfactory recovery of added silica. TABLE XI1 DETERMINATION OF SILICA IN SODIUM FLUORIDE Effect of sodium fluoride concentration Total SiO, NaF added, SiO, added, present, 50, found, Recovery, g. mg. mg. mg. % 0.6 2-8 3-4 3.3 97- 1 0.8 2.8 3.6 3.54 98.3 0.9 2.8 3.7 3.50 94-6 1.0 2- 8 3.8 2.88 75-8 1% cam be seen that 04g. of sodium fluoride (anhydrous) was the maximum amount that cauld be taken to obtain a satisfactory recotvery of silica. This is equivalent to 0.36 g.of fluoride ion. PROCEDURE- Dissolve not more than 0.8 g. of sodium fluoride in about 60 ml. of water in a silver beaker, add 4 g. of AnalaR boric acid, warming slightly to help solution. Make up to 100 ml. in a volumetric flask at 25" to 35" C. Take 50 ml., add 2 ml. of 5 N nitric acid, stir, add 5 ml. of 10 per cent. ammonium molybdate, again stir and after 5 minutes measure the colour developed on the Spekker. By this procedure silica can be determined over a range of 0.01 to 0.7 per cent. if 0.8 g. of sodium fluoride is taken. Examples of the recoveries obtained by this method when silica was added to sodium fluoride are shown in Table XIII. The procedure recommended is as follows. TABLE XI11 DETERMINATION OF SILICA IN SODIUM FLUORIDE Recovery of added silica Total SO2 N;LF taken, SiO, added, present, SiO, found, Error, g- mg- % % % 0.6 0.8 0.8 0.6 i 0.8 nil 0.10 0.10 - nil 0.10 0-10 - 1.4 0.2'75 0-270 - 0.005 2.8 0.587 0-550 - 0.017 2.8 0*4!iO 0.443 - 0.007 APPLICATION OF THE METHOD TO OTHER FLUORIDES The above results show that in general, by usihg 1OOml.of a 4 per cent. solution of boric acid, the maximum amount of fluoride that. can be taken without affecting the accuracy of the silica'determination is equivalent to 0.36 g. of fluoride ion. In the light of this know- ledge, and.bearing in mind the fact that some ions can influence the rate of development of full colour and also cause fading, the method may, with precautions, be applied generally.May, 19501 DETERMINATION OF SILICOFLUORIDE 263 For coloured fluorides a method was developed in which the coloured cation was first removed by a cation-exchange resin.The filtrate from the resin was colourless and contained hydrofluoric acid, together with the silica as hydrofluosilicic acid. There was no loss of silica by this resin treatment. The procedure adopted for nickel fluoride was as follows-0~5 g. of NiF2.4H20 and 4 g. of AnalaR boric acid were dissolved in about 60 ml. of water in a silver beaker. The solution was filtered, if necessary; log. of Amberlite 1R100H were then added and the mixture stirred for + hour. The resin was then filtered on a Whatman No. 541 filter-paper and washed with water. The filtrate, which was colourless, was made up to 100ml. The silica was determined in 50ml. by the general procedure already described.The same procedure was applied successfully to cobalt fluoride. It was necessary to know for certain purposes the silica content of magnesium fluoride. This compound is insoluble in water and is not easily soluble in cold dilute acids. The following procedure was found to be satisfactory. Half a gram of magnesium fluoride was fused in a silver beaker with 5 g. of AnalaR sodium hydroxide and kept in the liquid state for about 5 minutes. After cooling, the melt was warmed with about 50ml. of water to break it up. A few drops of phenolphthalein solution were added followed by 5 N sulphuric acid added dropwise until the solution was just permanently colourless. Four grams of AnalaR boric acid were then added and warmed until in solution.If necessary, the solution was filtered and after making up to 100 ml. the silica estimated as already described by taking 50 ml. of the dilution and making the measure- ment after 5 minutes. A blank experiment was carried out on the sodium hydroxide and the result was corrected accordingly. The main disadvantage of this method was that the silica content of the best available grade of sodium hydroxide was greater than that of the magnesium fluorides examined. INTERFERENCE Iron caused no interference as shown by adding quantities of up to 10 mg. of Fe per 100 ml. of final solution; this greatly exceeds the amount that is normally present in reagent grade fluorides. The method is subject to interference by phosphate. The author thanks the Directors of the British Drug Houses Limited for permission to publish these results and desires to record his appreciation of the encouragement he has received in this work from Mr. G. H. Osborn, F.R.I.C., A.M.I.M.M., Chief Analyst, B.D.H. Laboratory Chemicals Group. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. B.D.H. LABORATORY CHEMICALS GROUP Furman, N. H., “Scott’s Standard Methods of Chemical Analysis,’’ 5th Ed., Vol. 11, D. Van Kolthoff, I. M., and Furman, N. H., “Volumetric Analysis.” Vol. 11, John Wiley & Sons, New Manufacturing Chemists Association, Ind. Eng. Chem., A n d . Ed., 1944, 16, 483. Swinehart, C. F., and Hisik, H. F., Ibid., 1944, 16, 419. Cade. G. N., Ibid., 1945, 17, 372. Dobkina, B. M., Zavod. Lub., 1948, 14, 765. Brabson, J. A., Mattraw, H. C., Maxwell, G. E., Darrow, A., and Needham, H. F., A n d Ghem., Lacroix, S., and Labalade, M., A n d . Chim. Acta, 1949, 3, 383. Nostrand Co., New York, 1939, p. 2209. York, 1929, pp. 124-7. 1948, 20, 503. THE ANALYTICAL DEPARTMENT POOLE, DORSKT September, 1949