FEBRUARY, 1968 Voi. 93, No. I103 THE ANALYST Separation of Silicon and Phosphorus with Ammonium Molybdate and their Successive Determination BY A. M. G. MACDONALD AND F. H. VAN DER VOORT (Departwent of Chemistry, The University, Birmingham 16) A method for the successive determinations of silicon and phosphorus is described. Phosphate is separated cleanly from silicate as ammonium molybdophosphate; phosphate in the precipitate and silicate in the filtrate are determined by conventional quinoline procedures. The method is suitable for the separation and determination of 0.2 to 2 mg of each of the ions. The separation method can be applied to organic materials, after decomposition with peroxide. A study of organic bases for the precipitation of molybdo- silicate showed that none is preferable to quinoline.IN a recent paper,f a study of various organic bases for the quantitative precipitation of the molybdophosphate ion was described; a method based on 2-methylpyridine was found to be less sensitive to interferences from silicate, arsenate and citrate than the well known quinoline method, although most of the favourable characteristics of the quinoline method were retained. The investigation described below was undertaken in an attempt to find an organic base that would prove analogously less sensitive to interfering effects than quinoline for the determination of silicate. This aim was not accomplished but a satisfactory method was developed for the separation and successive determination of phosphate and silicate, which readily proved applicable to the analysis of organic materials.EXAMINATION OF ORGANIC BASES FOR PRECIPITATION OF MOLYBDOSXLICATE QUALITATIVE STUDY- Sensitivity tests were made in 1 N, 1.5 N and 2 N hydrochloric acid medium for both known and possible precipitants for molybdosilicate; for these tests, precipitation was made at 70" to 80" C by the addition of ammonium molybdate solution and the organic base solution. The sensitivities quoted in Table I are the least amounts of silicon (in a final volume of 6 ml) that could be detected when compared with a blank solution prepared simultaneously; at acidities below 1 N, precipitation of the molybdosilicate could scarceIy be distinguished from that of the base molybdate itself. Better sensitivities for some of the reagents could probably be achieved with a change in the conditions, for example, King and Watson2 recommend a sulphuric acid medium for pyramidon, but the hydrochloric acid medium used seemed satisfactory for comparison purposes.As expected, the sensitivities deteriorated rapidly as the acidity of the medium increased. Hexamine,s pyridine? 1-naphthylamine, piperidine, 2,4-dimethylpiperidine, cyclohexylamine, dibutylamine, 2-aminopyridine, benzotriazole, 2,2'-pyridoin and methyl and ethyl nicotinate proved of no value for molybdosilicate precipitation. Where their behaviour was unknown, the bases listed in Table I were tested for their sensitivity to the molybdophosphate ion. All showed greater sensitivity to molybdophosphate than to molybdosilicate, in the same way as quinoline and 8-hydroxyquinoline. QUANTITATIVE STUDY- The quantitative precipitation of molybdosilicate with the bases listed was examined, the procedure used being analogous to the gravimetric quinoline method of Miller and Chalmers.6 Generally, the precipitates were formed in about 1 N hydrochloric acid medium at 70" to 75' C, cooled in ice - water, washed with ice-cold water, and dried at 120" to 140" C for 1 hour; they were then weighed without ignition, their hygroscopicity being generally 66; 0 SAC and the authos.66 MACDONALD AND VAN DER VOORT: SEPARATION OF [APzdySt, VOl.93 negligible with a reasonable amount of care. The methods were tested for about 1.5 and 3.5mg of silicon as silicate; the results for the former amount are listed in Table I. All results were calculated on the basis of the formation of 4 (base).12Mo03.Si02.2H,0.TABLE I QUALITATIVE AND QUANTITATIVE RESULTS ON PRECIPITATION OF SILICATE Sensitivity as mg of silicon per 6ml* of Base 8-Hydroxyquinoline . . 2,4Dimethylquinoline 2-Methylquinoline . PMethylquinoline . . 2,4-Dimethylpyridine PAminopyridine . . 2-Benzoyipyridine . . Tributylamine . . N,N-Diethylaniline . . 4-Picolylamine . . 2-Aminoquinoline . . 3-Aminoquinoline . . 8-Aminoquinoline - . Quinoline . . .. Pyramidon .. .. Reference .. 7,s . . 5, 6, 12 .. 5 .. 5 .. 6 . * 2, 9, 10 * . - - .. - .. - . . - .. - .. - .. I .. - .. r ~~ 1 1.5 N 2N hydrochloric hydrochloric acid acid 0.0 1 0.01 0.02 0-05 0.05 0.50 0.02 0.05 0.02 0.05 0.50 0.50 0.1 0-50 0.05 0.05 0.02 0.10 0.05 0.50 0.02 0.10 0.06 0.10 0.02 0.02 0.02 0.05 0.02 0.10 Silicon present, mg 1.464 1.470 1-470 1.464 1.635 1.470 1.452 1.470 1.470 1.635 1.635 1.580 1.665 1-665 - Silicon foundt , mg 1.442 1.464 1.467 1.450 1.636 1-366 1.202 1.438 1-457 1.610 1.323 1.578 1.651 1.658 - Error t - 0.02 1 - 0.006 - 0.003 -0.014 + 0.001 -0.104 - 0.250 - 0.032 -0.013 - 0.025 -0.312 - 0.002 - 0.014 - 0.007 - * Qualitative test (see text).Average of two results. $ Blanks generally lay in the range 5 to 10 pg of silicon and have been deducted. The behaviour of these bases as quantitative precipitants for molybdosilicate ions falls into three groups. I. Quinoline,6 2,4-dimethylq~inoline,~ 4-rnethylq~inoline~ and 8-aminoquinoline all gave satisfactory results ; the 2-aminoquinoline precipitate tended to creep and was less satisfactory.11. 8-Hydro~yquinoline,~ s 8 2-methylq~inoline,~ 3-aminoquinoline and tributylamine gave slightly low results. These would have been improved by calculation on the basis of a monohydrated complex, but none of the previous evidence on base molybdosilicates seems to favour this formula; a proper thermogravimetric study would be of value. The low results obtained with 2-benzoylpyridine and N,N-diethylaniline can be explained by the slight solubility of the precipitates and a tendency to slight decomposition on drying. 111. 2,4-Dimethylpyridine, 4-aminopyridine and 4-picolylamine gave poor results, because of the solubility of the precipitates. A formula containing only three molecules of base has been favoured for pyramidon rnolybdosili~ate,~~~~~~~~ but none of the three above bases yielded sufficiently reproducible results for an analogous formula to seem justified. Heslop and Pearsonll have shown that 2,4-dimethylpyridine has a much better sensitivity for molybdophosphate than for molybdoarsenate, similarly to 2-methylpyridinel ; the molybdosilicates of the two bases appear to be more soluble than the molybdoarsenates.None of the bases examined had any marked advantage over quinoline itself for pre- cipitation of molybdosilioate. For the compounds in group 11, modification of experimental conditions would probably have improved the results but none of these bases showed any favourable characteristics in comparison with quinoline, hence this did not seem worthwhile. Accordingly, other procedures were sought for the separation of silicate and phosphate. As the method was ultimately intended for samples containing silicon in the range 1 to lOrng, the titximetric6 and gravimetric5 methods of completing the determination of silicate were compared ; the gravimetric method has been preferred by most later Under the conditions described in the Experimental it was found that silicon in the range 0.2 to 36mg could be determined gravimetrically with excellent accuracy and precision, whereas for the titrimetric method, the precision was less, an empirical factor was essential (as mentioned by Wilson6) and the range was only 0.2 to 1 nig, above which the end-points were masked by the formation of molybdenum blue, even when potassium chloratel5 was added.February, 19681 SILICON AND PHOSPHORUS WITH AMMONIUM MOLYBDATE 67 However, for the determination of phosphate, the titrimetric procedure of Wilson16 pr7 was preferred to the gravimetric method.18 No trouble arises from molybdenum blue in the phosphate titration whereas, for the gravimetric method, the molybdophosphate is more finely divided and requires more care in filtering on sintered crucibles than quinoline molybdosilicate.SEPARATION AND DETERMINATION OF PHOSPHATE AND SILICATE- Ion exchange and solvent extraction have often been used for the separation of phosphate and silicate. Ion exchange was not considered in the present work because of its general tediousness. Recently, extraction with isobutyl acetate at pH 1 has proved useful for the separation of phosphate from silicate,lg and methods for the selective formation of molybdophosphate and molybdosilicate in mixtures have been described.20 In both methods colorimetric procedures were used, which were considered inappropriate for the present purpose, i.e.,occasional organic analysis.Another colorimetric procedure for the simultaneous determination of silicon and phosphorus, which has been applied to organic compounds after fusion with sodium peroxide, involves the molybdosilicate blue and molybdovanadophosphate procedures.21 Brabson, Duncan and Murphyl3 have used quinoline for the simultaneous precipitation of molybdophosphates and silicates, the phosphate being determined on a separate aliquot as quinoline molybdopho~phate.1~~~7 So far as is known, no methods are available for the consecutive determination of the two ions on a single solution based on titrimetric or gravimetric techniques involving the heteropolyanions. The work done in these laboratories on the precipitation of molybdophosphates and silicates with organic bases indicated the impossibility of obtaining clean separations of the two ions by such means; the masking of one ion by, for example, citric or oxalic acid to allow precipitation of the other is valid only for small amounts of the unwanted ion, for there is always a slight effect on the required ion also if much of the masking agent is added.Accord- ingly, attention was re-directed to the ammonium heteropolymolybdates. There is little definite information in the literature about the precipitation of ammonium inolybdophosphate in the presence of silicate.Mellor and Thompson22 quoted many early workers in support of the statement that silicate can be co-precipitated as molybdosilicate, and mentioned that it is usually separated by a preliminary dehydration, although it may be removed by washing the ammonium molybdophosphate with ice-cold water. Hillebrand and Lunde1123 also state that silicate retards the precipitation and may co-precipitate as ammonium molybdosilicate, but most modern textbooks simply mention the interference of silicate; several imply that the gravimetric procedure for phosphate is affected by the formation of hydrated silica rather than by molybdosilicate. In contrast, Feig12* recommends a spot test in which molybdosilicate is detected in the filtrate after the bulk of the phosphate has been precipitated as ammonium molybdophosphate. Although Feigl’s separation was, of course, only for qualitative purposes, it was considered worthwhile to examine its quanti- tative possibilities. If ammonium molybdophosphate could be precipitated in a pure form, it should be possible to determine molybdosilicate in the filtrate by the quinoline method. The ammonium molybdophosphate precipitate could be dissolved and determined analogously by a quinoline rnethod,l6s17 for no advantage would accrue from the use of 2-methylpyridine in the absence of silicate.Measuring the ammonium molybdophosphate as such was un- attractive, because the conditions for the classical micro-gravimetric procedure are critical, requiring an empirical factor, and the micro-titrimetric procedure was shown to be inaccurate because of poor end-points despite several variations (cf.Stockdale25 and WilsonlSJ7). The procedure for the precipitation of ammonium molybdophosphate is based on that of Thistlethwaite26 and on the recommendations of Archer, Heslop and Kirby.27 REAGENTS- Standard silicate solution, 0-5 m g of silicon per mLWeigh about 200 mg of pure precipi- tated silica (British Drug Houses Ltd.) into a weighed platinum crucible, ignite to constant weight and fuse with 3 g of analytical-reagent grade sodium carbonate for 1Q hours. After cooling, transfer the crucible and contents to a beaker and dissolve the melt in distilled water ; then dilute to 100ml. Standard Phosphate solution, 0-5 mg of phosphorus $er md-Prepare from analytical- reagent grade potassium dihydrogen phosphate.EXPERIMENTAL68 MACDONALD AND VAN DER VOORT: SEPARATION OF [Auzdyst, Vol. 93 Ammonium molybdate reagefit-Dissolve 35 g of analytical-reagent grade ammonium molybdate in 50 ml of water and add 50 ml of analytical-reagent grade ammonia (sp. gr. 0.88). Dilute 168 ml of nitric acid to 360 ml with water, cool, and add the molybdate solution. Leave for 24 hours, dilute to 900 ml with water and filter. Sodium moly bdate solzttiort-Dissolve 43.2 g of analytical-reagent grade molybdenum trioxide and 24.0 g of analytical-reagent grade sodium hydroxide in water, add 58-0 ml of 6 N hydrochloric acid, dilute to 500 ml with water and filter. QztGzoZine soZu,tkon-Prepare a 2 per cent.solution of quinoline in 0.26 M hydrochloric acid. Ammonizm nitrate wash solzction-Dissolve 16 g of analytical-reagent grade ammonium nitrate in 2 litres of water. Store all reagent solutions in polythene containers. Water from an all-glass distillation apparatus was used throughout. SEPARATION OF PHOSPHATE- To a 10-ml sample of the mixed phosphate - silicate solution, add 2 ml of 1 per cent. nitric acid and heat to 50" to 55' C on a water-bath. Heat the ammonium molybdate reagent to the same temperature and add 12 ml of it to the sample solution, while stirring. Digest at 50" to 55" C for 30 minutes with occasional stirring, and then cool for 30 minutes. Filter through a 1.5-cm layer of paper pulp supported on a Witt plate, transferring and washing the precipitate with 30ml of ammonium nitrate solution in four portions.Suck the paper pad dry and use the filtrate (volume 55 to 60 ml) for the determination of silicon. A suction "bell-jar" apparatus, similar to that described by Hillebrand and L~nde11,~~ was used. DETERMINATION OF PHOSPHORUS- Add 10.0 ml of 0-5 M sodium hydroxide solution to the original precipitation vessel and use this to dissolve the ammonium molybdophosphate off the pulp pad in small portions, collecting the filtrate in a 250-nil ground-glass stoppered conical flask. Wash the vessel, funnel and pad with five 4-ml portions of water and suck the filter dry. Neutralise the filtrate, the volume of which is about 30 ml, with 6 M hydrochloric acid in the presence of methyl red indicator, acidify with 8 ml of the acid and add 200 mg of citric acid followed by 6 ml of sodium molybdate reagent.Heat to boiling, and add 6 ml of quinoline solution. Re-heat to boiling and cool in ice - water to about 15" C. Filter on a pulp pad, washing the flask and filter with water until they are free from acid. Transfer the pad and precipitate back to the flask, add an excess of 0.1 N sodium hydroxide solution, shake to dissolve the precipitate and back-titrate the excess of alkali with 0.05 M hydro- chloric acid in the presence of phenolphthalein indicator. DETERMINATION OF SILICON- To the filtrate of the amnionium molybdophosphate separation, add 1.8 ml of 6 M hydro- chloric acid and 6 ml of sodium molybdate solution and allow to stand for 5 minutes. Heat in a water-bath to 65" to 70"C, but not higher, and add 6*2ml of 6~ hydrochloric acid and 6 ml of quinoline solution.Digest at 70" C for 10 minutes and then cool to about 15" C in ice - water. Filter on a weighed No. 4 sintered-glass crucible, transferring the precipitate and washing it with small portions of cold water (20ml in all). Dry the crucible and pre- cipitate for 1 hour at 120" to 140"C, cool in a desiccator over phosphorus pentoxide for 30 minutes and weigh immediately. In every determination, establish blank values by taking 0.5 mg of the relevant ion through the entire procedure. DECOMPOSITION OF ORGANIC MATERILU- Place some M.A.R. sodium peroxide in a weighed nickel "fluorine" bomb (C. W. Cook Ltd., Birmingham) and add the weighed (5 to 8 mg) organic sample, followed by more sodium peroxide to give 200mg in all.Seal the bomb with a copper gasket (not asbestos-filled) and heat in a muffle furnace €or 15 minutes at 500" C. Swirl the contents of the bomb after 10 minutes and again when removing the bomb from the furnace. Allow the bomb to cool and open it, rinsing any of the melt on the lid with water into a platinum dish. Rinse the outside of the bomb, wipe with a tissue and half-cover the bomb in the dish with water. Cover, heat gently to boiling to dissolve the melt and cool. PROCEDURESFebruary, 19681 SILICON AND PHOSPHORUS WITH AMMONIUM MOLYBDATE 69 Place the amount of 4 M nitric acid needed to neutralise the melt (1-80 ml for 200 mg of sodium peroxide) in a polythene beaker and rinse the watch-glass, bomb and the contents of the platinum dish into the beaker with water.Add 300ml of 1 M sodium hydroxide solution, allow to cool and finally adjust the pH to 7 with 2 M nitric acid (pH meter). Transfer the resulting 70 to 80 nil of solution to a glass beaker and use it for the separation and deter- mination of silicon and phosphorus as described above. RESULTS AND DISCUSSION DETERMINATION OF PHOSPHORUS- For the titrimetric determination of phosphate as quinoline molybdophosphate, it was shown that the presence of the additional molybdate, ammonium ion and the sodium chloride arising from the neutralisation had no effect on the accuracy of the final results. When phosphate was separated first as the ammonium complex, some low results were obtained when sintered-glass filters were used; there was no escape of the ammonium salt with pulp-pad filters.Because of the general doubt surrounding the stoicheiometry of ammonium molybdo- phosphate, it was felt necessary to check that all the phosphorus was present in the precipitate. This was done by precipitating the molybdate in the precipitate after dissolution, as lead molybdate ; ThistlethwaiteZB has shown that the molybdenum - phosphorus ratio in ammonium molybdophosphate is always constant whatever the total composition is. Accordingly, the excellent results obtained for molybdate in the above tests indicate that none of the original phosphate escaped precipitation. Ten determinations of standard phosphate solution by the two successive steps showed an average recovery (after blank deduction) of 9994 per cent.for 1 mg of phosphorus, with an average error of 0.6 per cent.; the slightly low recovery can certainly be ascribed to solubility losses. DETERMINATION OF SILICON- For the gravimetric determination of silicon as quinoline molybdosilicate, it was shown in the same way as for the phosphorus determination that the various ions present in the filtrate of the ammonium molybdophosphate precipitation did not affect the accuracy. Good results were obtained whether pure silicate solutions or those containing the additional ions were analysed. For example, six determinations of standard silicate solutions containing 1 mg of silicon with all the extraneous ions added showed an average recovery of 99.8 per cent. (average error 0.3 per cent.); when the same amount of silica was taken through the entire separation procedure, the average recovery was 99.65 per cent.and the average error was 0.6 per cent. for six results. Blank values, when determined by straightforward blank determinations, varied beyond the permissible limits, a common feature of heteropolymolybdate procedures, and were best determined by using a standard solution of the relevant ion. TABLE 11 CONSECUTIVE DETERMINATION OF PHOSPHATE AND SILICATE Phosphorus Phosphorus taken, found, mg mg 1.992 2.01 1 1.992 2.010 0.996 0.992 0.906 0.986 0.996 1.005 0.996 0.999 0.747 0.746 0.498 0.602 0,249 0.246 0.249 0.246 0.100 0.090 Error, mg + 0.019 + 0.018 - 0.004 - 0.010 + 0.009 + 0.003 -0.001 + 0.004 - 0.004 - 0.004 - 0.010 Silicon taken, mg 0.990 1.980 0.099 0.248 0.495 0-743 0.990 0.990 0.990 0.248 0.990 Silicon found, m g 0.979 1.961 0.101 0.243 0.497 0.736 0.981 0-987 0.981 0.237 0.978 Error, m g - 0.01 1 -0.019 + 0.002 - 0.006 + 0.002 - 0.008 - 0.009 - 0.003 - 0.009 -0~011 -0.012 The range over which the separation method could be used was about 0-2 to 2 mg of each of the ions concerned. Typical results are shown in Table 11.With amounts of silicon70 MACDONALD AND VAN DER VOORT: SEPARATION OF [Analyst, VOl. 93 larger than 2 mg there was a slight co-precipitation with the ammonium molybdophosphate, so that the method could not be used; at the 2-mg level, the relative error does not exceed 1 per cent. ANALYSIS OF ORGANIC COMPOUNDS- An attempt was made to use the oxygen-flask combustion for the decomposition of silicon compounds.Reverchon and Legrand2* have described a method in which a nickel flask is used, sodium peroxide is added to the sample and the sample holder is boiled for 45 minutes in the 1 M sodium hydroxide absorbent solution to recover all the silica from it. In the present work, thick-walled, 500 or 1000-ml polypropylene flasks were used and the products were absorbed in 0.1 M alkali or acid. Under no conditions could quantitative recoveries of silicon be achieved, despite the addition of sodium peroxide, sodium carbonate, sucrose, fully fluorinated or fully chlorinated compounds to the sample before the combustion. The various combustion aids were added both singly and in various mixtures; sodium peroxide was certainly the most effective additive, but the recoveries of silicon were still 5 to 10 per cent.low. Because of the plastic flasks, it was not possible to boil the absorbent solution. Christopher, Fennell and Webb21 have used fusion with sodium peroxide in a nickel bomb for the decomposition of semi-micro amounts of organic silicon and phosphorus compounds, and this method was successfully scaled down for 5 to 8-mg samples. Nitric acid was used to acidify the solution because chloride tends to retard the precipitation of molybdophosphate ; the amount of sodium nitrate arising from neutralisation of the bomb leachings had no effect on the separation and determination process. The volume of solution obtained from the bomb leaching and subsequent treatment was 70 to 80m1, and it was shown on standard silicate and phosphate mixtures that the increase in volume scarcely affected the accuracy of the final results.Results obtained for mixtures of silicon and phosphorus compounds are shown in Table 111. Only one compound containing both elements was available and the results are also shown in Table 111. It can be seen that all the results obtained are within the acceptable limits for elemental organic analysis. The method is probably less convenient for continual routine use than methods based on solvent extraction and spectrophotometry. It is, however, useful for occasional analysis, as only common reagents are required. We are grateful to Prof. R. Belcher for his interest in this work, and to Mr. R. W. Fennell for providing the special sample. TABLE I11 ANALYSIS OF ORGANIC COMPOUNDS Phosphorus, per cent.Silicon, per cent. Sample weight, mg 8.260 3.863 3.468 3.972 3.960 5.761 5.942 3.604 6.640 4.500 4.769 3.876 6.664 1 Theoreti- cal 11.81 Theoreti- cal 12-97 Compound Diphen ylsilanediol Diphenylsilanediol Diphenylsilanediol Hexaphenyldisiloxane Hexaphenyldisiloxane + triphenylphosphine + triphenylphosphine + triphenylphosphine + triphenylphosphine + triphenylphosphate C,*H,*O,P*Si, Found 11.96 Error +O.lS Found 12.91 Error - 0.06 11.81 11-73 - 0.08 12.97 13.06 + 0.09 11.81 11.93 +0.12 12.97 13-15 +O*l8 11.81 12.10 + 0-29 10.60 10.77 +0.27 9.62 9.36 -0.16 10.60 10.18 -0.32 8.04 7.28 + 0.26 +Om31 + 0.28 8.18 8.13 8-06 3-0.14 + 0.09 +O.Ol 7.63 7.59 7-66 REFERENCES 1. Macdonald, A. M. G., and Rivero, A.-M., Analytica Chim. Ada, 1967, 37, 625.2. King, E. J., and Watson, J. L., Mikrochamie, 1936, 20, 49. 3. Duval, C., “Inorganic Thermogravimetric Analysis, ” Second Edition, Elsevier Publishing Co., Amsterdam, London and New York, 1963, p. 239.February, 19681 SILICON AND PHOSPHORUS WITH AMMONIUM MOLYBDATE 71 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 10. 20. 21. 22. 23. 24. 25. 26. 27. 28. Babko, A. K., J. Appl. Chem., USSR, 1937, 10, 374. Miller, C. C., and Chalmers, R. A., Analyst, 1953, 78, 24. Wilson, H. N., Ibid., 1949, 74, 243. Voluinetz, M. Y., Zav. Lab., 1936, 5, 162. Su, Y.-S., Campbell, D. E., and Williams, J. P., Analytica Chinz. Ada, 1965, 32, 559. Hecht, F. , and Donau, J ., “Anorganische Mikrogewichtsanalyyse ,’, Springer-Verlag, Wien, 1940, Furuya, M., Bunseki Kagaku, 1962, 11, 1247. Heslop, R. B., and Pearson, E. F., Analytica Chim. Ada, 1967, 37, 516. Armand, M., and Berthoux, J., Ibid., 1953, 8 , 610. Brabson, J. A., Duncan, R. D., and Murphy, I. J., Analyt. Chem., 1963, 35, 1102. Christopher, A. J., and Fennell, T. R. F. W., Talanta, 1965, 12, 1003. Fernlund, U., and Zechner, S., 2. analyt. Chem., 1955, 146, 111. Wilson, H. N., AnaZyst, 1951, 76, 65. Fennell, T. R. F. W., and Webb, J. R., Talanta, 1959, 2, 105. Paul, J., Mikrochim. Acta, 1965, 836. Chalmers, R. A., and Sinclair, A. G., Analytica Chzm. Acta, 1966, 34, 412. Christopher, A. J., Fennell, T. R. F. W , and Webb, J. R., Talanta, 1964, 11, 1323. Mellor, J. W., and Thompson, H, V., “A Treatise on Quantitative Inorganic Analysis,” Second Edition, Charles Griffin & Co., London, 1938, p. 671. Hillebrand, W. F. and Lundell, G. E. F., Revised by Lundell, G. E. F., Bright, H. A., and Hoffman, J. I.,“Applied Inorganic Analysis,” Second Edition, John Wiley & Sons Inc., New York; Chapman 6 Hall Ltd., London, 1953, pp. 100 and 701. Feigl, F., “Spot Tests. Inorganic Applications,” Elsevier Publishing Co., Amsterdam, 1968, p. 336, Stockdale, D., Analyst, 1968, 83, 24. Thistlethwaite, W. F., Ibid., 1947, 72, 531. Archer, D. W., Heslop, R. B., and Kirby, R., Analytica Chim. Ach, 1964, 30, 450. Reverchon, R., and Legrand, Y., Chim. Analyt., 1966, 47, 194. p. 244. -, Ibid., 1954, 79, 535. Received September 20th, 1967