Sept., 19511 THE ACIDITY OF MILK 509 An Improved Volumetric Method for the Determination of Hydrogen Sulphide and Soluble Sulphides BY J. A. KITCHENER, A. LIBERMAN AND D. A. SPRATT (Presented at the meeting of the Society on Wednesday, April 4th, 1951) The usual iodimetric method of determining hydrogen sulphide or sulphur in iron and steel is not satisfactory when accurate results are required. The use of alkaline sodium hypochlorite as combined absorbing and oxidising reagent has been investigated and found to offer many advantages. The reagent is stable to boiling and gives quantitative oxidation of sulphide to sulphate. The recommended procedure has been shown to give results correct to within 0.3 per cent. of the sulphur content. IN the course of researches into the thermodynamical activity of sulphur in liquid iron,l it was necessary to make a large number of determinations of (a) hydrogen sulphide in mixtures of hydrogen sulphide and hydrogen and (b) sulphur in iron.The quantity of sulphur in the samples was geflerally between 1 and 10 mg, and it was required to determine it to within IfIl per cent. A method of analysis was therefore needed that would be (i) reliable, both in reproducibility and absolute accuracy, (ii) sensitive to 0.01 mg and (iii) reasonably rapid. A review of the literature suggested that, of the many and various methods that have been described for the determination of hydrogen sulphide and sulphur in iron, the commonly used volumetric meth0d~9~ would be the most suitable for the purpose. In this method, the hydrogen sulphide in a gas mixture, or resulting from the dissolution of a sample of iron in acid, is absorbed and then titrated by iodimetry.There are many modifications for the absorbing medium and method of titration, several of which are in common use in steelworks. Absorbing solutions include ammoniacal or buffered cadmium chloride or acetate or zinc sulphate, while in the determination of hydrogen sulphide in gases, sodium hydroxide is often used. The resulting sulphide suspension or solution may be acidified and titrated immediately with iodine or iodide - iodate mixture, but the most common procedure is to add the sulphide to an excess of acidified iodine and then titrate the residual iodine. The basic reaction is- The usual evolution method3 is generally considered satisfactory for routine analysis of plain carbon steels, but when it was applied to a number of the hydrogen sulphide - hydrogen and pure iron - sulphur samples arising from the present work the results were not S” + I, -+ s J.+ 21’.510 KITCHENER, LIBERMAN AND SPRATT IMPROVED VOLUMETRIC METHOD FOR [VOl. 76 sufficiently concordant, although several of the usual absorbents were tried. Closer examina- tion of the ordinary volumetric method shows that it is really semi-empirical rather than accurately stoicheiometric. It is reliable only for a carefully standardised set of conditions and any variation of the conditions may lead to unexpected errors from a variety of causes. Examples of errors up to 17 per cent. arising merely from variation in the size of the sample used in the analysis have been quoted by Lundell, Hoffman and Bright.2 In the present work it was often necessary to use very small specimens of iron because of the abnormally large proportion of sulphur in many of them (e.g., about 1 per cent.of sulphur instead of the usual range of 0.01 to 0.1 per cent.). In these circiimstances it would clearly be unsatisfactory to take the ordinary volumetric, method on trust when the research required an absolute accuracy of 1 per cent., as the results might easily be subject to large unknown errors. SOURCES OF ERRORS IN THE ORDl NARY VOLUMETRIC METHOD Consideration was given to the possibility of developing a more reliable volumetric method, and the potential sources of error in the existing methods were first examined. The following errors have been recognised- Incomplete evolution of the sulphur as hydrogen sulphide during dissolution of the iron- This is a well known difficulty with certain cast irons and alloy steels, but since it does not apply to pure iron - sulphur samples or hydrogen sulphide - hydrogen mixtures it need not be discussed further here.Incomplete absorption of the hydrogen sulphide--Rapid dissolution of the sample is usually recommended as this is found empirically to give better results.2 Theoretically it should make no difference; that it is found to do so is another indication of the existence of weakness in the method. If the gas contains much hydrogen sulphide some of it may escape absorption and two absorption bulbs in series must then be Loss of hydrogen sulphide was noticed with some of the high-sulphur samples when tested in conjunction with the usual absorbents.Alkali was found to be the most effective absorbent. Errors in the titration procedures-(a) If cadmium solutions are -used, the cadmium sulphide must be protected from sunlight since it is photo-~ensitive.~ (b) With the cadmium or zinc back-titration method, the end-point is not altogether satisfactory, the colour changing from a reddish-brown to a deep blue colour. Attempts to increase the sensitivity of the titration by working with more dilute solutions gave dis- couraging results. (c) If the sulphide solution is first acidified and then titrated directly with iodine, some hydrogen sulphide may be lost before the titration is completed.(d) If the back-titration procedure is used there is a possibility of occlusion of iodine by the colloidal sulphur that is formed.6 (e) If sodium hydroxide is used as the absorbing solution, atmospheric oxygen is apt to cause rapid oxidation of the alkali sulphide in solution. (f) If alkali absorbent is mixed with excess of acidified iodine there is a possibility of volatilisation of iodine owing to the rise of temperat~re.~ To avoid errors (d) and (f) the solution is usually diluted considerably,8 but, with alkali sulphides, this emphasises the danger of (e). The above list gives the chief potential sources of error in the volumetric methods; certain others have been noted by Etheridge.g Although they do not necessarily vitiate the determinations, they indicate the importance of very careful control of conditions.Shawlo and Treadwell and Hall11 have pointed out that differences of technique, dilution of the solution, rate of addition, and amount of potassium iodide present may produce large variations in the results. There is clearly a need for a better method, giving greater reliability and sensitivity. In developing such a method the following points deserve attention- (i) The reactions employed in the titration should be accurately stoicheiometric (no side-reactions) . (ii) Absorption should preferably be by alkali hydroxide, as this minimises the danger of loss of hydrogen sulphide. (iii) If alkali is to be used, there must be complete exclusion of air from the apparatus and the solutions. Further, transference or dilution of the sodium hydroxide - sodiumSept., 19511 51 1 sulphide solution or addition of large volumes of reagent solutions containing dissolved air should be avoided.(iv) The neutralisation of large amounts of alkali should be avoided to minimise loss of iodine. (v) If an oxidising agent could be found that would oxidise sulphide to sulphate instead of to colloidal sulphur, it would improve the sensitivity fourfold, avoid occlusion of iodine, and improve the end-point. (vi) For accurate work an all-glass apparatus should be used, so avoiding rubber, which absorbs hydrogen sulphide. DETERMINATION OF HYDROGEK SULPHIDE AND SOLUBLE SULPHIDES USE OF ALK.4LINE HYPOCHLORITE FOR ABSORPTION AND OXIDATION All the above desiderata were satisfied by the introduction of an alkaline solution of Kolthoff and Sandell12 state that sulphides are quantitatively oxidised to sulphates sodium hypochlorite as a combined absorbing and oxidising reagent.by calcium hypochlorite in alkaline media according to the reaction- but they do not specify the conditions necessary to secure complete oxidation. The use of hypobromites and hypochlorites for the determination of sulphides was proposed by Willard and Cake.13 Hypochlorites are the more stable, but it was found that at least 4 N sodium hydroxide was needed to effect oxidation at room temperature, whereas 2.5 N sodium hydroxide was sufficient with sodium hypobromite. Willard and Cake therefore proposed absorbing hydrogen sulphide in 2.5 N sodium hydroxide and then washing this into 0.3 N sodium hypobromite.The method to be described below avoids transferring alkali sulphide solutions (cf. point (iii) above) by using a combined absorbing and oxidising reagent. The more stable hypochlorite is used, at a concentration of 0.1 N instead of 0.3 N to give greater sensitivity. The alkali is 0.4 N instead of 2 N to reduce the heat of neutralisation (cf. point (iv) above). The use of dilute calcium hypochlorite solutions for the approximate deter- mination of microgram quantities of sulphides has been described recently by P e p k 0 ~ i t z . l ~ Solutions of pure hypochlorites are remarkably stable if kept free from organic matter (dust). They can be boiled without decomposition.16 They are more stable still in the presence of free alkali,16J7 and such solutions can be kept for many months without appreciable change of titre, provided they are stored in dark bottles to avoid photochemical decomposition by sunlight.When attempts were made to absorb hydrogen sulphide directly into alkaline hypo- chlorite it was found that some colloidal sulphur was always formed, even with fairly con- centrated sodium hydroxide and a considerable excess of sodium hypochlorite. Experiments were therefore made to find a means of securing quantitative oxidation to sulphate. It was not considered desirable to use very concentrated alkali for reasons already given. The use of higher temperatures was therefore considered. Willard and Cake reported more nearly complete oxidation with hypochlorite at 45” C. I n the present work it was found that any colloidal sulphur formed during absorption at room temperature could be removed by subsequently heating the solution to about 70” C.How- ever, warming just sufficiently to clear the solution of sulphur did not lead to concordant results. There must certainly be intermediates between sulphide and sulphate, and apparently some stages of the reactions are sluggish, This procedure led to concordant results and stoicheiometric relations corresponding to formation of sulphate. As, so far as is known, the boiling of alkaline hypochlorite solutions is a novel step in volumetric analysis the following evidence is presented to demonstrate its reliability. The validity of the procedure is also supported by the accuracy of the results of the new method as recorded below.Solutions approximately 0.4 N with respect to sodium hydroxide and 0-1 N with respect to sodium hypochlorite were made by passing chlorine from a cylinder into the alkali until the required titre was reached. Provided there is a considerable excess of alkali and the solutions are cool and dilute, the product obtained is almost entirely hypochlorite, no appreciable amount of chlorate being formed.18319 It is then satisfactory to standardise the solution by the iodimetric method (see Kolthoff and Sandell,12 pp. 587 and 640). The s” + 4 OC1’ 3 SO,’’ + 4 Cl’, Pepkowitz recommended 80” to 90” C. Finally, it was found satisfactory to boil the solution.512 KITCHENER, LIBERMAN AND SPRATT: IMPROVED VOLUMETRIC METHOD FOR [VOl. 76 solution is just acidified with hydrochloric acid, excess of potassium iodide is added, and the iodine liberated is titrated with sodium thiosulpha.te, using starch at the end-point. Aliquot 25-ml portions of such solutions were transferred by pipette into conical flasks and boiled briskly for different times.The flasks .were then cooled under the tap, and the contents titrated as described above. Table I shows the results (in millilitres of 0.1 N sodium thiosulphate) of independent tests by two of the authors on three solutions prepared at different times. TABLE I EFFECT OF BOILING ON SOLUTIONS OF SODIUM HYPOCHLORITE Titre of solution A Titre of solution B ml ml ml Titre of solution C Duration of boiling (A.L.), (A.L.), (D.A.S.), 0 (cold) 1 min. 2 min. 5 min. 10 min.25.45 25.25 25-20 25-19 25-20 30.00 29.85 29-80 29.80 29-85 47.05 46.84 46.86 46-85 46.81 It is seen that during the first 2 minutes’ boiling there was a small loss of titre amounting to about 0.7 per cent. in each experiment. Thereafter the titre remained constant within the accuracy of titration (about kO.1 per cent.) during a further 8 minutes’ boiling. The initial loss is believed to be due to reaction with traces of dust. It is quite clear that alkaline hypochlorite alone is perfectly stable to boiling, and that the titre reached after, say, 5 minutes’ boiling is a highly reliable figure.* It is essential, of course, to protect the solutions from bright sunlight during treatment. Experiments in which duplicate samples of sodium sulphide solutions were added to the alkaline hypochlorite and boiled for 5 minutes gave similarly reproducible results.The rate of addition of the sodium sulphide was found to be immaterial, whereas the amount of sulphur initially precipitated and the titre in the cold were very dependent on the mixing conditions. Since sulphate could be detected after the reaction, and since all possible oxidation reactions had evidently finished after 5 minutes’ boiling, it was concluded that the treatment effected complete oxidation of sulphide to sulphate irrespective of the intermediates formed. Since also the starting and final hypochlorite solutions were stable it was evident that the process could form the basis of a reliable and stoicheiometric method of determining hydrogen sulphide or any sulphides from which hydrogen sulphide could be completely evolved.Proof of the absolute accuracy of the hypochlorite titration is not easy, as there is no convenient substance that can be taken as a primary standard for hydrogen sulphide. Even sulphides such as those of zinc and cadmium, which are sometimes used in the gravimetric determinations of those metals, cannot be prepared completely pure and of stoicheiometric composition.22 It has therefore been necessary to resort to indirect checks, which are described later in this paper. APPARATUS- The complete apparatus for the determination of soluble sulphides or sulphur in iron and steel is shown in Fig. 1. The chief parts, preferably made of Pyrex glass, are the evolution flask, A, which has a wide ground-glass joint carrying a small dropping funnel, B, and a spray trap and condenser (conveniently combined, C), and the absorption flask, E.Connection is made by the tube, D, which has ground cones at each end. If desired, the ground joint to flask E may be replaced by a rubber stopper. If the evolution apparatus is likely to be used in bright daylight, the absorption flask must be screened by a box of wood or cardboard. The dimensions of the apparatus are not critical. METHOI) REAGENTS- Sodium hypochlorite, 0.1 N , in sodium hydroxide, 0.4 N-Prepare by passing chlorine from a cylinder into a 0-5 N solution of sodium hydroxide until the required concentration * It is worth recording that “Chloramine T,” which is a very convenient substitute for hypochlorite for titrations in the cold,20J1 does not behave in the same way to boiling.Instead, there is a progressive loss of titre, no doubt due to reaction with the organic part of the compound.Sept., 19511 DETERMIXATION OF Hl7DR0GEK SULPHIDE AKD SOLUBLE SULPHIDES 513 is reached (see above). For 2 litres of solution, this takes between 5 and 7 minutes at a moderate rate of bubbling. Sodium thiosulphate, 0-05 N-Prepare by dissolving 25 g of sodium thiosulphate and 7.6 g of borax (as preservative) in 2 litres of water. It is most conveniently standardised against a solution of potassium iodate (1.7 g of potassium iodate dried at 180" C for 2 hours). Hydrochloric acid, diluted (1 + 1)-Prepare an air-free solution by diluting concentrated hydrochloric acid with its own volume of boiled, cooled water.Store in a Winchester bottle fitted with a siphon and protect the contents from oxidation by means of a bubbler containing alkaline pyrogallol. Potassium iodide-A freshly made approximately 10 per cent. solution. Hydrochloric acid, approximately 2 N . Starch solzttion-A freshly made approximately 0.5 per cent, solution. [L L Fig. 1. Details of apparatus L E f J PROCEDURE- To determine, for example, sulphur in iron, first wash the whole apparatus with water, rinse it with alcohol and dry with warm air. Weigh the sample to k0.5 mg and place it in the evolution flask. Assemble the apparatus as shown in Fig. 1, the ground joints being very lightly greased, Transfer 25 ml of standard alkaline hypochlorite solution by means of a pipette into the absorption flask and add 25 ml of water.The absorption flask must be screened from bright light. Pass a steady stream of hydrogen from a cylinder through the apparatus for 10 minutes to sweep out all air. Place a small piece of lead-acetate paper and a small piece of starch - iodide paper in the exit tube of the absorption flask to test the completeness of absorption. Admit about 20 ml of diluted hydrochloric acid (1 + 1) into the evolution flask through the tap-funnel, keeping the amount of air introduced with it to a minimum. Place a small flame or, better, an electric hot-plate under the evolution flask and maintain a continuous, slow stream of hydrogen through the apparatus.. With high-sulphur irons the rate of dissolution must not be too great.514 KITCHENER, LIBERMAN AND SPRATT: IMPROVED VOLUMETRIC METHOD FOR [VOl.76 When the sample has finally dissolved (as shown for iron by means of a magnet placed near any residual particles of carbon, etc.), boil the solution in the evolution flask, A, for 5 minutes to ensure that the hydrogen sulphide has been expelled. The trap and condenser prevent hydrochloric acid from passing over into the absorption flask, E. Then disconnect the absorption flask and boil its contents steadily for 5 minutes fl minute. The white colloidal sulphur that appears during the absorption is completely oxidised to sulphate and the solution becomes quite clear. Remove the tubes from the absorption flask:, E, and rinse them, collecting the rinsings in the flask. Immediately cool the flask thoroughly, e.g., by covering the neck with a small inverted beaker and standing the flask under a irunning tap.The solutions are then ready for titration. Add 10 ml of 10 per cent. potassium iodide solution and then 20 ml of 2 N hydrochloric acid. Titrate the liberated iodine against 0.05 N sodium thiosulphate, and add 5 ml of starch solution as indicator when the end-point is nearly reached. Determine the thiosulphate equivalent of the hypochlorite reagent in duplicate by boiling 25 ml and titrating in the same way as in a sulpliur determination. The difference between this figure and that given by a sulphur determination is the volume of thiosulphate solution that is equivalent to the sulphur in the sample- 1 ml of 0.05 N sodium thiosulphate = 0.200 mg of sulphur. CONFIRMATION OF THE ABSOLUTE ACCURACY OF THE METHOD In the absence of any satisfactory primary standard for hydrogen sulphide or soluble sulphides, it was necessary to resort to indirect t’ests.The sulphur contents of a number of materials were therefore determined by the hypochlorite method described above and by a standard method, and the results compared. CoMPARISON WITH DIRECT WEIGHING METHOD FOR HYDROGEN SULPHIDE- Sherman, Elvander and Chipman% have recently shown that hydrogen sulphide in the gases evolved from the dissolution of iron in acids can be accurately determined by first washing the gases to remove acid vapours, drying with “Anhydrone” and then absorbing and weighing the hydrogen sulphide directly in a bulb of “Ascarite.” This method was used to determine the sulphur content of a sample of manganese sulphide prepared according to the method described by Biltz and Vl’iechrr~ann.*~* The samples used for the direct weighing method weighed 0-25 g, and were dissolved in the evolution flask in the usual way.Samples of the same specimen of manganese sulphide weighing 0.05 g were similarly dissolved and the hydrogen sulphide determined by the volumetric method as described in this paper. The results, expressed as percentage of sulphur- in the specimen of manganese sulphide, were as follows- Weight of sample Sulphur, Mean % By direct weighing of hydrogen sulphide . . 0.26 g 34.4, 34.4 34.4 By hypochlorite titration . . .. . . 0.05 g 34.1, 34-7, 34.2 34.3 From these figures it is concluded that the hypochlorite titration method is an accurate It gives results that are probably correct to within means of determining hydrogen sulphide.0.3 per cent. of the true value. COMPARISONS WITH THE STANDARD GRAVIMETRIC (BARIUM SULPHATE) METHOD- Borings from a special iron - sulphur ingot, prepared in the Chemistry Laboratory of the British Iron and Steel Research Association, were sampled by mixing and quartering. Twelve determinations by the new volumetric method gave a mean of 0-384(5) per cent. of sulphur. Independent umpire analyses on 5-g samples carried out by Miss M. Dalziel in * It is noteworthy that Biltz and Wiechmann claimed that manganese sulphide so prepared is stoicheio- metric. The evidence, however, was only that the manganese content was correct for manganese sulphide; the sulphide content of their sample may well have been low, as was that of the manganese sulphide used for the present test.The above test was not, of course, dependent upon having a stoicheiometric sulphur content. Possibly some S (= 32) is replaced by (OH), (= 34).Sept., 19511 DETERMINATION OF HYDROGEN SULPHIDE AND SOLUBLE SULPHIDES 515 the Advanced Analytical Laboratories of this Department, using the standard A.S.T.M. gravimetric method, gave 0.387 per cent. Similar comparisons have also been made with mixtures of zinc sulphide and zinc (volumetric 1-14 per cent.; gravimetric 1.15 per cent.) and of ferrous sulphide and iron (volumetric 1.01 per cent. ; gravimetric 1.01(5) per cent.). These tests confirm that the evolution - volumetric method described here gives results for sulphur in iron and other metals that are in good agreement with the gravimetric method. There can be no doubt that it is reliable within the required limits of + 1 per cent.; it is probably better than this.The data appear to suggest that the volumetric method may be giving results about 0.6 per cent. lower than the gravimetric method, but even the standard barium sulphate method is probably not of the highest absolute accuracy. The present authors do not know of any unambiguous tests of the ultimate reliability of the barium sulphate determination, whereas its liability to errors is well known.2925 TIME REQUIRED FOR A DETERMINATION AND LIMIT OF SENSITIVITY A single determination of sulphur in iron can be completed in about 1i hours, allowing 1 hour for the dissolution of the sample.For the analysis of mixtures of hydrogen sulphide and hydrogen the time required is only about half an hour. As two determinations can easily be carried out simultaneously, an experienced worker with several sets of apparatus could perform 8 to 12 analyses in a working day. Since the hypochlorite method is more than four times as sensitive as the ordinary method by virtue of its stoicheiometry and its better end-point, it is worth considering how small a quantity of sulphur as hydrogen sulphide or soluble sulphide can be detected. Some tests made with a small absorption flask and an ordinary 5-ml micro-burette showed that the end-point is sensitive to +0.005 ml of 0.062 N sodium thiosulphate. Hence, allowing this uncertainty on both the unknown and the blank, the uncertainty of the difference would be +O-Ol ml.This is a limiting sensitivity of 0-0024 mg of sulphur, which would be 0.00024 per cent. by weight in a l-g sample of iron, or 0.00017 per cent. by volume as hydrogen sulphide in 1 litre of a gas mixture at N.T.P. The above estimate gives the sensitivity; the smallest amount of sulphur that could be determined with a given accuracy is, of course, correspondingly greater. For example, 0.01 per cent. of sulphur in a l-g sample of iron could be readily determined correct to the nearest 0.001 per cent. With larger amounts of sample or higher sulphur contents the full accuracy of the method (i.e., within +O-3 to 0.6 per cent. of the true sulphur content) can be achieved.Two of us (A. L. and D. A. S.) thank the British Iron and Steel Research Association for B.I.S.R.A. Bursaries, during the tenure of which the above work was done. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. REFERENCES Kitchener, J. A., Bockris, J. O’M., and Liberman, A., Disc. Farad. SOC., 1948, No. 4, 49. Lundell, G. E. F., Hoffman‘ J. I., and Bright, H. A., “Chemical Analysis of Iron and Steel,” John Wiley and Sons Inc., New York, 1931, chap. XII. Scott, W. W., and Furman, N. H., “Standard Methods of Chemical Analysis,” Fifth Edition, D. Van Nostrand Co. Inc., New York, and the Technical Press Ltd., London, 1939, pp. 1442-1444. -- , Ibid., p. 914. Luniell, G. E. F., Hoffman, J. I., and Bright, H. A., op. cit., p. 242. Treadwell, F. P., and Hall, W. T., “Analytical Chemistry,” John Wiley and Sons Inc., New York, Hoar, T.P., and Eyles, G. E. S., Analyst, 1939, 64, 666. Brunck, O., 2. anal. Chem., 1906, 45, 541. Etheridge, A. T., in Mitchell, C. A. (Editor), “Recent Advances in Analytical Chemistry,” Shaw, J . A., I n d . Eng. Chem., Anal. Ed., 1940, 12, 668. Treadwell, F. P., and Hall, W. T., 09. cit., p. 557. Kolthoff, I. M., and Sandell, E. B., “Textbook of Quantitative Inorganic Analysis,” Macmillan Willard, H. H., and Cake, W. E., J . Amer. Chem. SOC., 1921, 43, 1610. Pepkowitz, L. P., Anal. Chem., 1948, 20, 968. Mellor, J. W., “Comprehensive Treatise of Inorganic and Theoretical Chemistry,” Longmans, 1930, p. 583. J. & A. Churchill Ltd., London, 1931, Volume 11, p. 174. and Co., Ltd., London, 1943, p.588. Green and Co., Ltd., London, 1922, Vol. 11, p. 252.516 KITCHENER, LIBERMAN AND SPRATT [Vol. 76 16. 17. Foerster, F., 2. Elektrochem., 1917, 23, 138. 18. 19. SO. 21. 22. Cullen, G. E., and Hubbard, R. S., J . BioZ. C h e w , 1919, 37, 511. Gmelin’s Handbuch dev Anorg. Chem., 1927 (8), 6, 265. Fuchs, P., Bodenkunde u. P’anzenerniihr., 1942, 28, 385. Vogel, A. I., “Textbook of Quantitative Inorganic Analysis,” Longmans, Green and Co., Ltd., London, 1943, p. 459. Bendall, J. R., Mann, F. G., and Purdie, D., J . Chern. Soc., 1942, 157. See, for example, Hillebrand, W. F., and Lundell, G. E. F., “Applied Inorganic Analysis,” John Wiley and Sons Inc., New York, 1929, p. 207; :Scott, W. W., and Furman, N. H., op. cit., p. 201, footnote. Sherman, C. W., Elvander, H.I., and Chipman, J., J . Metals, 1950, 188, 334. Biltz, W., and Wiechmann, F., 2. anorg. Chem. 1936, 228, 268. Kolthoff, I. M., and Sandell, E. B., op. cit., chajp. XX and pp. 718-719. 23. 24. 25. DEPARTMENT OF INORGANIC AND PHYSICAL CHEMISTRY IMPERIAL COLLEGE LONDON, S.W.7 DISCUSSION MR. F. L. OKELL said that the authors were to be congratulated upon having removed one of the three long-standing difficulties of the volumetric evolution methods for determining hydrogen sulphide, namely, the uncertainty that attended the iodimetric titration of metallic sulphides. The authors’ opinion that the reaction was not stoicheiometric was confirmed by his own experience of the method, By applying the alkaline hypochlorite oxidation to this determination they had improved its scientific aspect, as distinct from manipulative technique, and so increased our knowledge of this attractive but hitherto disappointing method.With this improvement the method should become of service for many purposes other than those of the foundry chemist. He asked if the all-glass apparatus described by the author was indispensable to the method and mentioned that S. G. Clarke (Analyst, 1931,56, 436) had used corks and black rubber tubing in an evolution method for hydrogen sulphide on samples weighing but 0.1 g and got good results. DR. KITCHENER, in reply, thanked Mr. Okell for his remarks and said that, as an all-glass apparatus eliminated one possible source of trouble, they had not investigated the suitability of cork and rubber tubing. However, it was worth mentioning that in other work they had found polythene tubing satis- factory and had proved by direct test that, unlike rubber, it did not absorb hydrogen sulphide.Polyvinyl chloride tubing was also available and was superior to rubber for many purposes in the laboratory, e.g., in the polarograph. MR. J. HASLAM, although not necessarily implying that the test results would be affected, rather doubted the stoicheiometry of the reaction between hypochlorite, iodide and acid. That seemed to him to be taken account of in the best methods of standardisation of hypochlorite. In these methods what ultimately happened was that a slight deficiency of arsenite solution was added to a known amount of hypochlorite, and only a t this stage, when a very small amount of hypochlorite was present in excess, was iodide added, the iodine then liberated being titrated by the addition of a further small volume of standard arseni te . MR. R. F. MILTON said that in his experience titration of hypochlorites with potassium iodide and thiosulphate was unsatisfactory owing to the end-point being indefinite and variable. He had found that the arsenite and iodine titration was to be preferred. DR. KITCHENER replied that although the arsenite method was no doubt preferable in some instances for determining hypochlorite, especially when chlorates were present (e.g., in bleaching powder), they considered the direct potassium iodide reaction entirely :satisfactory for pure solutions of hypochlorites. This opinion was supported by the literature (see reference 12, above) and by the reproducibility of their results (Table I). DR. J. H. HAMENCE asked if the method could be used to estimate hydrogen sulphide in water directly, without distillation. MR. T. MCLACHLAN said that he did not think it would be possible to apply this method to the deter- mination of traces of hydrogen sulphide in waters on account of the large amount of dissolved air. More- over, when they were present in traces, sulphur dioxide was frequently there in addition to hydrogen sulphide. With regard to the reliability of the methods of Dr. Kitchener and his co-workers. it should be remembered that in the determination of available chlorine in sodium hypochlorite solution, the analyst was concerned only with the amount of hypochlorite, whereas Dr. Kitchener was indifferent to the forms in which the chlorine was present, as long as he could eventually determine the total amount after liberation with acid and iodide. The end-point in their titrations was very sharp. DR. KITCHENER concurred.