Dec., 19501 OSBORN 67 1 The Use Of Polarographic Methods for the Analysis of Fine Chemicals BY G. H. OSBORN (Read at the meeting of the Physical Methods Group 092 Friday, October 'ith, 1949) SYNoPsIs-'rhe function of polarographic methods in a laboratory dealing with the analysis of fine chemicals is discussed at length and followed by outlines of the methods at present in me. Detailed procedures are given for new methods. THE analyst in charge of a laboratory controlling the production of thousands of different fine chemicals is faced with what appears to be an almost insuperable task. The vast variety of the samples to be analysed, together with all the various possible impurities in them and the ever-growing demand by customers for very pure chemicals of known analysis, sets the analyst a very difficult problem.It is only by the use of every possible modern tool that he is able to attempt to cope with the problem, and the polarograph is one of the most essential of these tools. It must first be said that the application of the polarograph to any problem in such a laboratory is only justified if there exists no simpler or more rapid method by other techniques. Each particular application of the polarograph calls for very careful investigation before the method can be put into routine use, and it will be obvious, therefore, that the amount of research required would render its wholesale application to the analysis of thousands of different chemicals out of the question. The polarograph is, therefore, only used in our laboratory where it is justified on the grounds of accuracy, speed or simplicity, or where no other method is known to exist.There are many hundreds of polarographic methods recorded in literature, but this paper deals mainly with those found to be useful in one particular fine chemical laboratory. We find the polarograph to be of most use for the estimation of trace amounts of elements or groupings, either organic or inorganic; it is seldom used for the determination of the major constituents in a chemical. In the determination of trace elements or groupings in fine chemicals, it may be stated as a general rule that, if the half-wave potential of the trace to be determined precedes the half-wave potential of the compound under test, then the determination of the trace is usually a simple matter.Solution of the compound in water, in which the compound itself acts as a supporting electrolyte, followed by taking a polarogram should theoretically give the desired result. If, however, the half-wave potential of the compound under test precedes that of the trace being determined, then interfering elements or groupings must either be complexed or removed by some method such as electrolysis. We find the most expedient method of measurement to be that using internal standards.l This greatly decreases the time required for a determination and avoids the use of thermostats. We find it advisable to make two separate additions of known amounts of the substance being determined and to take a polarogram after each addition to ensure that the increases in wave heights are linear.INORGANIC TRACES ZINC- The polarograph is especially valuable for the estimation of zinc owing to the difficulty of estimating traces of this element by other methods. Zincgives very well-defined waves both in acid and in alkaline solutions and can readily be estimated, e.g., in ferrous sulphate,l cadmium salts: aluminium salts3 and copper salts. In the determination of zinc in ferrous sulphate it is not necessary to remove the iron, and this transforms what was a very difficult problem into a very simple one. For the determination of zinc in cadmium, it is necessary to remove the cadmium by electrolysis and we have found that in doing so there is no loss of zinc. This electrolysis must be from a solution made just acid with either sulphuric or672 OSBORN: THE USE OF POLAROGRAPHIC METHODS [Vol.76 perchloric acid, and the current should be made 3 to 4 amps., copper-plated platinum electrodes4 being used. Similarly, in copper salts, the copper is removed by electrolysis prior to the polarographic estimation. The determination of zinc in organic compounds falls into two classes, (a) those compounds, such as insulin and thiouracil, which must first be subjected to wet oxidation because they yield interfering waves, and (b) those substances that do not give interfering waves. Zinc may be determined directly in phenol after treatment with excess of caustic soda. If an organic sample is wet oxidised with sulphuric and nitric acids before the polarographic estimation of zinc, care must be taken to remove all traces of residual nitrate by treatment with ammonium oxalate because the nitrate group has a half-wave potential in alkaline solutions close to that of zinc and can cause serious interference.Kolthoffb states that the nitrate group gives no interference in alkaline solution, but this is not our experience. Another interesting procedure recently developed is for zinc in thorium salts, when the thorium is complexed with sulphosalicylate at pH 8-5 in the presence of a small amount of gelatin. The method is excellent for the range of 20 p.p.m. to 1.0 per cent. of zinc in thorium.6 It is better to use sulphuric and perchloric acids whenever possible. NICKEL- There are a number of elegant methods for the determination of nickel in the presence of many elements, but most of these are unsuitable for the determination of small quantities of nickel in the presence of large amounts of cobalt.Traces of nickel can be determined in cobalt salts by the simple expedient of neutralising and then complexing the cobalt with ammonium thiocyanate.' When dealing with complex amino-cobalt compounds or organic salts of cobalt it is preferable to convert them to cobalt sulphate by wet oxidation with sulphuric acid before neutralisation. The colorimetric determination of nickel in the presence of large amounts of cobalt was extremely tedious, even by those methods that were accurate, whereas by aid of the polaro- graph it is simple and rapid. This effectively separates the cobalt and nickel waves. COBALT- The problem of determinating small traces of cobalt in nickel has been for a long time unsolved, although polarographic methods were known by which it was possible to separate large traces of cobalt from nickel, e.g., those involving the use of ammonium oxalate as a supporting electrolyte.8 None of the methods proposed, however, was found to be satis- factory for the determination of small traces of cobalt in nickel.A method has been proposed lately involving the use of Trilon B (ethylene diamine tetra-acetic acid) as a base solution.9 We have tried the method suggested, but quite apart from the fact that the sensitivity claimed by the authors is poor, not indicating less than 0.4 per cent. of cobalt in nickel, we have not succeeded in obtaining good waves for low concentrations of cobalt.We feel that the application of Trilon B to this particular problem is far more likely to be solved on a photometric basis. We are, however, proceeding with further polarographic experiments with this reagent. LEAD- Lead is not normally determined polarographically in our laboratories as other methods exist, but in certain cases where it is tedious to estimate by these other methods, such as in ferrous sulphate,l it is rapidly determined polarographically, using the salt as the ground solution. Lead can be determined directly in nickel salts using the nickel salt as supporting electrolyte.l* Another useful application is the determina- tion of traces of lead in zinc salts simply by dissolving the salt in dilute hydrochloric acid and taking a polarogram.ll Lead is also determined in cadmium salts by using a cyanide supporting electrolyte when the lead wave appears about -0.4 volt before the cadmium wave.We have also found that we can determine lead directly in phenol by adding excess of sodium hydroxide and then taking the polarogram. Lead may be determined in the presence of large amounts of tin, aluminium and iron since at pH 6-5 to 7-0 these elements are not reduced, although lead gives a wave at about -0.5 volt. Nitrates and free hydro- chloric acid must be eliminated as they give drawn-out waves.12 The supporting electrolyte Reliable results are obtained.Dec., 19501 FOR THE AN-4LYSIS OF FINE CHEMICALS 673 is a 30 per cent. solution of calcium chloride. Lead can be determined in aluminium salts in a sodium carbonate ground solution.Interference from tin is prevented by oxidation, from iron by reduction in alkaline solution and from copper by precipitation with potassium thiocyanate. l3 COPPER- Trace amounts of copper are not normally determined on the polarograph since the colorimetric method using sodium diethyldithiocarbamate is so rapid and simple. The polarographic method is, therefore, only used where it is difficult to apply other methods. For example, with ferrous salts we have found it advantageous to use the polarographic method which does not involve the removal of the ir0n.l As in the case of lead, copper can be determined directly in nickel salts by using the nickel salt as supporting electrolyte.lo Copper may also be determined polarographically in the presence of cadmium, nickel, zinc and manganese in lead and its salts.14 Davies and Key15 have described a method in which copper is determined in the presence of iron with N potassium fluoride or sodium potassium tartrate as supporting electrolyte.CADMIUM- The polarographic estimation of this element is most usefully applied in our laboratories to the determination of traces in zinc salts. Since the cadmium wave precedes the zinc wave by about 0.4 volt it is only necessary to dissolve the zinc salt in water and take a polaro- gram ; the salt itself serves as supporting electrolyte.ll IRON- Iron is not normally determined polarographically as so many other methods exist, but we find it useful to determine this element in cobalt salts by dissolving the salt in alkaline tartrate solution when, if all the iron is present in the ferric state, its wave appears well before the cobalt wave.We have not found this method much use below 0.1 per cent. of iron, although at and above this figure good waves are obtained. This procedure has also been recommended for the determination of iron in manganese salts.lB A method has recently been proposed whereby iron may be determined in an acidic oxalate supporting electrolyte at pH 5 provided the iron is first reduced to the ferrous state. With the exception of copper, which must be removed, no other ion soluble in dilute acid solution containing sulphur dioxide gives an interfering wave. It is claimed that for small amounts of iron this procedure is more accurate than either L-olumetric or gravimetric procedures.lC VANADIUM- oxidation,la although it is not usually determined polarographically in fine chemicals. Vanadium is sometimes determined polarographically in organic materials after wet TELLURIUM- We find it convenient to determine small amounts of tellurite in the presence of a large amount of selenitels by addition of a slight excess of ammonia, de-oxygenation with sodium sulphite and polarography over the range -0.5 to -1.0 volt. It is advisable to add a little gelatin to improve the wave shape. ?. I IN- A useful method for the determination of tin in phenol has been published,20 but normally we do not determine this element polarographically. ANTIMOPU’Y AND BISMUTH- Page and Robinson21 have described methods for the micro-estimation of antimony and bismuth in inorganic and organic compounds.Tervalent antimony may be directly determined; quinquevalent antimony must first be reduced. In N sulphuric acid the half- wave potential of antimony and bismuth are respectively -0-34 and -0.02 volt and in N nitric acid they are -0-17 and -0.03 volt. Thus it is possible to estimate traces of both in sulphuric674 OSBORN : THE USE OF PO1,AROGRAPHIC METHODS [Vol. 75 acid; the wave, however, coalesces in hydrochloric or nitric acid. Bismuth can be determined in copper salts by polarography in tartrate and citrate media.% In acid tartrate of pH 4.5, bismuth can be determined in the presence of large amounts of lead and cadmium. ,iLKALINE EARTHS- The alkaline earths are reduced at large negative potentials.Thus their trace estimation in the presence of most other elements is impossible without some preliminary chemical separation, so that a polarographic method has no advantage over the standard technique, Zlotowski and Kolthoff,s however, claim that barium, strontium and calcium can be deter- mined in the presence of each other in alcohol - water media. When present in approximately the same amounts, the barium wave appears first followed by the strontium wave and finally the calcium wave. Magnesium interferes with the calcium determination. In our laboratories, however, we have found the flame photometer to be the ideal instrument for determination of trace amounts of alkalies and alkaline earths, and we do not, therefore, see any reason for application of the polarograph.NITRATE- The determination of nitrate in sodium nitrite can be readily accomplished by means of the polarograph. Haslam and Cross% have described a method in which the nitrite is first decomposed with sodium azide in hydrochloric acid solution and, after concentration, the solution is polarographed in a lanthanum base solution over the voltage range -1.2 to -2.1 volts. This method is satisfactory over the range 0.04 to 1 per cent. of sodium nitrate. NITRITE IN NITRATE- In the presence of uranyl ions and dilute hydrochloric acid, nitrite and nitrate ions are both reduced at approximately -0.9 volt (versus the saturated calomel electrode). The diffusion current is proportional to the nitrite concentration when the ratio of uranyl ion to nitrite is greater than unity.Under these conditions, the reduction of nitrite involves three electrons, indicating a reduction to nitrogen. Analysis of the wave shows that the reduction is irreversible. A solution can be analysed for both nitrate and nitrite ions in two polarographic experiments. The diffusion current due to the two constituents in the original solution is first measured. The nitrite in a second sample is oxidised to nitrate by hydrogen peroxide in acid solution, the excess of peroxide is destroyed catalytically by manganese dioxide in alkaline solution, and the diffusion current is measured.% IODATES IN IODIDE- We have found it very convenient to determine iodates in iodides when both are water- soluble, e.g., the potassium salts. For potassium iodide, the solution is made slightly alkaline with sodium hydroxide, de-oxygenated with hydrogen and a polarogram is taken, when the iodate wave appears at about -1.2 volts.BROMATE IN BROMIDE- Bromate in bromide may easily be determined in the same way as iodate in iodide. It is, unfortunately, not possible to determine chlorate in chloride in a similar manner owing to the fact that the reduction potential of the chlorate is higher than that of the supporting electrolyte . ORGANIC APPLICATIONS In the field of fine organic chemical analysis, the use of the polarograph has been less well developed than in the inorganic field. Quantitative organic polarography in general has from the beginning lagged behind inorganic developments. However, among other applications of polarographic methods to organic analysis the following are valuable methods in the analysis of fine chemicals.ALDEHYDES I N ALCOHOL- The rapid and accurate determination of acetaldehyde in ethyl alcohol has become possible by means of the following procedure, which is an application of the method by Adkins and Cox26 for the polarographic measurement of aldehyde.Dec., 19501 FOR THE ANALYSIS OF FINE CHEMICALS 675 The alcohol is mixed in equal proportions with a I M solution of ammonium chloride and, after de-oxygenation of the solution, a polarogram is taken, when the acetaldehyde wave occurs in the range - 1.0 to - 3.0 volts. Internal standards are then added and measure- ment made as usual. We have found this method to have great advantages over the usual method with Schiff's reagent.FURFURALDEHYDE IN FORMALDEHYDE- Reduction of furfuraldehyde occurs in acid, neutral and alkaline media giving a single wave in acid and alkaline solutions and two waves in solutions of pH 4.0 to 7.0. Within this range, with increasing acidity the first wave gets smaller and the second larger, but the total wave-height remains unchanged. With pH 7, the reduction potential is 0.25 to 0.30 volt more positive than that of hydrogen, and under these conditions formaldehyde does not interfere.27 NITROBENZENE IN ANILINE- Small quantities of nitrobenzene in aniline may very conveniently be determined by the method of Haslam and Cross.28 To a known amount of the sample is added a little con- centrated hydrochloric acid containing nigrosine and a polarogram is taken.From 0.01 to 0.05 per cent. of nitrobenzene in aniline can be determined with an accuracy closer than 4 per cent. GAMMA ISOMER IN GAMMEXANE (BENZENE HEXACHLORIDE)- A useful method of assay as distinct from the trace determinations so far discussed is the determination of the gamma isomer of benzene hexachloride as described by Dragt.29 The gamma isomer is the only one of the five isomers reduced at the dropping mercury electrode under the conditions described. The method consists of taking a solution of the isomers in acetone, alcohol and water, buffered with a potassium chloride - sodium acetate buffer and, after de-oxygenating, taking a polarogram through the range -0.5 to -2.0 volts. Other methods exist for the determination of this isomer, but this is by far the most rapid with the possible exception of the infra-red method.MALEIC - FUMARIC ACID MIXTURES- It is claimed that there exists no satisfactory method for the estimation of these acids in the presence of each other, except by a recent polarographic method.aO This method enables the analyst to solve the problem in a simple manner by polarography in an ammonium hydroxide - ammonium chloride buffer solution of pH 8-2 as supporting electrolyte. The maleate wave precedes the fumarate wave. If interfering substances are present, maleic and fumaric acids must be precipitated as their barium salts in alcoholic solution. The precipitate is soluble in the base solution. We find that the method works quite well when maleic and fumaric acids are present in roughly equal quantities.It fails entirely when there is an overwhelming proportion of one or other constituent. Thus we were quite unable to detect 6 per cent. of fumaric acid added to maleic acid. The authors of this method refer throughout their work to roughly equal quantities of both and make no reference to unequal proportions. We find that 25 per cent. of fumaric acid had to be present before we could detect the slightest sign of a second wave. Our experience with this method has not been particularly fortunate. PEROXIDES AND ALDEHYDES IN ETHER- The method normally employed in our laboratories for the determination of peroxides in ether is based on the oxidation of ferrous thiocyanate and subsequent measurement of colour. Recently, however, we have been trying out the polarographic method of G ~ s m a n .~ ~ Peroxide and aldehyde are extracted from the ether by shaking with an equal volume of 0.01 N lithium hydroxide and determined by polarography of the aqueous layer. The peroxide wave occurs at -1.3 volts and the acetaldehyde wave at -1.8 volts. It is claimed that this method shows waves where chemical methods have failed to detect any peroxide. The method has the disadvantage that not all the peroxide or aldehyde is extracted in one shaking with the lithium hydroxide. From the partition coefficients, which are 0.45 for676 OSBORN THE USE OF POLAROGRAPHIC METHODS [Vol. 75 peroxide and 0.63 for aldehyde, it is possible to calculate the amount originally present in the ether. The waves are measured by comparison with the increases in wave height obtained by the addition of dilute solutions of hydrogen peroxide and acetaldehyde.In spite of its apparent disadvantage, it appears that the method is a promising one. PEROXIDE IN DIOXAN (DIETHYLENE DIOXIDE)-- Peroxide may be determined in dioxan very easily by the following technique. The diosan is mixed with an equal volume of M lithium hydroxide solution and, after de-oxygena- + - // ' \CHLOROFORM ELECTRIC HOT PLATE Fig. 1 tion with hydrogen, a polarogram is taken over the range -0.5 to 2.0 volts. To estimate the peroxide content a known volume of a standard diluted hydrogen peroxide solution is added and polarography repeated.32 HYDROGEN PEROXIDE- The polarographic determination of hydrogen peroxide can be carried out by making use of the reduction step at -1.0 volt versus the saturated calomel electrode. The method is limited to peroxide concentrations of less than 0.15 per cent., owing to the oxidation of mercury by hydrogen peroxide a t higher concentrations. If a stationary platinum micro- electrode is used in place of the dropping mercury electrode, current - voltage curves in the range from 0 to -0.6 volt versus the Saturated calomel electrode show a reduction step whose height is proportional to the hydrogen peroxide concentration over a much widerDec., 19501 FOR THE -4NALYSIS OF FINE CHEMICALS 677 range, and determinations can be carried out simply by measuring the limiting current at -0.6 volt.The upper limit of peroxide concentrations that can be measured depends on the concentration of the supporting electrolyte, and if saturated potassium chloride solution is used, concentrations up to 0.9 per cent.can be determined.% THXOMERSALATE- Page and WallersP have recently described an interesting polarographic method for the estimation of thiomersalate. A well-defined wave at abott -0.5 volt is obtained in N hydrochloric acid. This method is very useful for the determination of this antiseptic in vaccines and pharmaceutical preparations. THE DETERMINATION OF BENZANTHRONE AND ANTHR.4QUINONE IN THE PRESENCE OF EACH OTHER- An interesting application has been suggested35 for these two chemicals since both give a good wave in 70 per cent. methyl alcohol containing 0.1 N sulphuric acid. Benzanthrone c. 5 a a 3 0 +C i 0 -0.5 -1.0 -I *s -2.0 VOLTS Fig. 2 has a half-wave potential at 0.96 volt while anthraquinone has a half-wave potential at -0-36 volt.The wave height is proportional to the concentration and this simplifies the detennina- tion of small quantities of one in the presence of large amounts of the other. DETERMIN.4TION OF PHENYL MERCURIC ACETATE IN A GELATIN BASE- Some recent interesting experimental work in the B.D.H. laboratories has led to the development of a method by L. J. J. Hillman for the determination of small quantities of phenyl mercuric acetate in a gelatine base. Page34 had reported two waves for this compound in dilute hydrochloric acid base solution and this has been confirmed by us. Our problem was complicated by the presence of the large amount of gelatin, which made the solution very viscous at room temperature, and consequently it gave an erratic polarogram owing to the mercury drop not falling at a steady rate.This difficulty was overcome by conducting the experiment at an elevated temperature in a specially constructed polarographic cell. Details of this cell are shown in Fig. 1. It consists of a graduated cell made from a burette barrel by sealing one end. The cell is heated by immersion in the refluxing vapour of an organic liquid. The solvent used to produce a constant temperature was chloroform (AnalaR) and the working temperature was found to be 58" & 0.2" C. It is essential that the temperature be kept constant. Preearation of calibration cuyve-A standard solution of phenyl mercuric acetate was made up as follows.Take 6 g . of glycerol and dissolve in it 0.15 g. of phenyl mercuric acetate, add 2.25g. of triethanolamine and make up to 100ml. with water. Take various known678 OSBORN THE USE OF POLAROGRAPHIC METHODS [Vol. 75 amounts of this solution and to each add 0-5 ml. of 0.01 per cent. methyl red solution, 0.4 g. of gelatin and 2 crystals of sodium sulphite (AnalaR). Make the volume of each portion to 10 ml. with distilled water. Place each solution in turn in the polarographic cell and heat in the vapour of the refluxing chloroform for about a quarter of an hour so as to allow a uniform temperature to be achieved. Then set the potentiometer at -1.05 volts, adjust PHENYL MERCURIC ACETATE, mg. Fig. 3 the zero current and condenser current to give a good deflection, and take the reading on the galvanometer scale.Disconnect the lead to the polarographic cell so that it is no longer in circuit and take another reading; the difference between these two readings gives the diffusion current from the cell at -1.05 volts, which is proportional to the phenyl mercuric acetate concentration. I t is necessary to adopt this procedure since an anodic wave due to sulphite just before the first phenyl mercuric acetate wave makes it impossible to, estimate the residual current (see Fig. 2). Plot the diffusion current obtained from the above solutions against the respective concentrations of phenyl mercuric acetate when a straight line graph (Fig. 3) is obtained. Sodium sulphite was used as it was not possible to obtain a satisfactory degree of de-oxygena- tion in a reasonable time by the use of hydrogen.The graph does not pass through the origin since the current measured by this method includes the residual plus the diffusion current at -1-05 volts. It will be found that the current in microamps represented by XY in Fig. 3 corresponds to that represented by CD in Fig. 2, and their values correspond to the residual current at -1.05 volts. Accuracy-By means of the calibration graph so constructed, a series of twenty samples containing known amounts of phenyl mercuric acetate were estimated by an operator who did not know the true contents. All the results obtained were found to be within k5.0 per cent. of the true value. Thanks are due to Mr. A. Jewsbury, BSc., A.R.I.C., and to Mr.L. J. J. Hillman of these laboratories for assistance and constructive advice in the preparation of this review, and to the Directors of the British Drug Houses, Ltd., for permission to publish. 1. 3. 4. 5. 9 d. REFERENCES Jewsbury, A., and Osborn, G. H., Analyst, 1948, 73, 506. B.D. H. Laboratories-unpublished work. Stross, W., and Osborn, G. H., Light Metals, July, 1944, 7, 323-7, and Stross, W., Metallurgin, 1947, Osborn, G. H., Metallurgia, 1949, 39, 111. Kolthoff, I. Rf., and Lingane, J. J., “Polarography,” Interscience Publishers, Tnc., New York, 36, 163-6 and 223-5. p. 314.Dec., 19501 FOR THE ANALYSIS OF FINE CHEMICALS 679 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 65. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. Patterson, J .H., and Ranks, C . V., Anal. Chern., 1948, 20, 897. Lingane, J. J., and Kerlinger, H., Ind. Eng. Chem., Anal. Ed., 1941, 13, 77. Prajzler, J., Coll. Czech. Chew. Comnz., 1931, 3, 406. Souchay, P., and Faucherre, J., .-3naZ. Chim. A d a , 1949, 3, 252. Korchunov, I. A., and Sazanova, L. N., Zavod. Lab., 1948, 14, 621. Terai, Y., Bull. Inst. Phys. Chem. Research, Tokyo, 1938. Mhlopen, 3. Ya., J . Anal. Chenz. Russ., 1947, 2, 55. Stross, W., Metallwgia, 1947, 37, 49-51. Urotschmann, C., Metallwirtshaft, 1944, 23, 343. Davies, W. C., and Key, C., Ind. Chemist, 1943, 19, 555. Verdier, E. T., Coll. Czech. Chern. Comm., 1939, 11, 340. Meites, L., Anal. Chem., 1948, 20, 895. Page, J. E., and Robinson, F. A., Analyst, 1943, 68, 269. Schwaer, L., and Sucky, K., Coll. Czech. Chem. Comm., 1935, 7, 75. Wilson, H. N., and Hutchinson, W., Analyst, 1947, 72, 149. Page, J. E., and Robinson, I;. A., J . SOC. Chem. Ind., 1942, 61, 93. Sucky, K., Coll. Czech. Chem. Coinm., 1931, 3, 354. Zlotowski, I., and Kolthoff, 1. M., J. Phys. C h e w , 1945, 49, 386. Haslam, J., and Cross, L. H., .I, SOC. Chem. Ind., 1945, 64, 259. lieilin, B., and Otvos, J . W., J . Amer. Chem. SOL, 1946, 68, 2665. Adkins, H., and Cox, F. W., Ibid., 1938, 60, 1151. Malyugina, N. I., and Korshunov, I. A., J . Anal. Chem. Rzkss., 1947, 2, 341. Haslam, J., and Cross, L. H., J . Soc. Chem. Ind., 1944, 63, 94. Dragt, G., Anal. Chenz., 1948, 20, 737. W'arskowsky, B., Eloing, P. J., and Mandel, J., Anal. Chenz., 1947, 19, 161. Gosman, I3. A., Coll. Czech. Chem. Comin., 1935, 7, 467. Unpublished work-B.D.H. Laboratories. GiguCre, P. A., and Jaillet, J . B., Canad. J . Res., 1948, B, 26, 767. Page, J. E., and W a l k , J. G., Analyst, 1949, 74, 292. Yu, I. Vainshtein, Zavod. Lab., 1949, 15, 411. ANALYTICAL DEPARTMENT THE BRITISH DRUG HOUSES, LTD. POOLE, DORSET (B.D.H. LABORATORY CHEMICALS GROUP) First submitted, Novembev, 1949 Amended, April, 1950