290 THE ANALYST. INORGANIC ANALYSIS. Estimation of Aeids in Hydrogen Peroxide. H. Endemann. (Zed. angew. Chem., 1909, 22, 673-674.)-Hydrogen peroxide nearly always contains appreciable quantities of acid, added for the purpose of increasing its stability. The quantity of acid present in pharmaceutical preparations should be as small as possible, and is generally estimated by titration with sodium hydroxide in presence of phenol-THE ANALYST, 291 phthalein. The author has observed, however, that decomposed samples of hydrogen peroxide apparently contain far more acid than the same samples in the fresh condition. If the estimation be effected by adding excess of alkali and titrating back, the same result is obtained as by direct titration. This value only corresponds to one-half of the true amount of free acid present.I n order to determine the whole of the acid, it is necessary to add excess of standard sodium hydroxide, heat the liquid in a platinum dish until the hydrogen peroxide is completely decomposed, and titrate back with acid in presence of phenolphthalein. Glass vessels are not suitable unless a piece of platinum wire be present to eifect the catalytic decomposition of the peroxide. For technical purposes it is sufficient to multiply the acidity, found by direct titration in the cold, by 2. The explanation of the anomaly observed by the author in the titration of the acids in hydrogen peroxide is found in researches of Tafel. When sodium hydroxide is added to hydrogen peroxide, sodium peroxide is formed, which immediately splits up into a perhydroxide and ordinary hydroxide, thus : Na202 + H,O = NaOOEI + NaOH.The latter, of course, has an alkaline reaction towards phenolphthalein, but the perhydroxide is neutral. The latter reacts with concentrated hydrochloric acid in alcoholic solution, giving a sodium oxychloride, NaOCl, which is not identical with sodium hypochlorite. J. F. B. Direet Volumetric Method for the Estimation of Alumina. W. H. Seamon. (Western Chemist and Metall., February, 1909 ; Chem. Engineer, 1909, 9, 116.)-The precipitate of alumina, and iron oxide, obtained as usual, is dissolved in the smallest possible quantity of concentrated hydrochloric acid. To this solution 2 to 5 grams of sodium peroxide are added, so as to yield it strongly alkaline solution containing aluminium as aluminate.The solution is filtered from ferric hydroxide, and exactly neutralised with hydrochloric or sulphuric acid, after the addition of phenolphthalein. The liquid is then heated slmost to boiling, methyl orange is added, and then standard sulphuric acid, until the aluminium hydroxide previously precipitated is entirely redissolved, and a permanent acid reaction is obtained. The quantity of standard acid used corresponds to the alumina present. The method is inapplicable in the presence of lead or zinc. Sodium peroxide is used to make the solution alkaline, instead of sodium hydroxide, on account of its greater purity. A. G. L. Separation of Antimony from Tin. G. Panajotow. (Ber. deut. Chem. Ges., 1909, 42, 1296-1299.)-The quantitative separation of antimony and tin is one of the most difficult in analytical practice.Of the methods hitherto proposed, that of Vortniann and Metzl (ANALYST, 1905, 30, 281) is the simplest, and gives useful results. The author, however, has found that a ready means for the separation of the sulphides of these two metals may be based on their different solubilities in hydrochloric acid of different concentrations. Antimony sulphide, though soluble in hot concentrated hydrochloric wid, is insoluble in the 15 per cent. acid, whilst tin292 THE ANALYST. sulphide is completely soluble. When both metals are present in an acid solution of such concentration, saturation with hydrogen sulphide at the ordinary tempera- ture throws out a little of the tin sulphide in association with the antimony, but this may be avoided by performing the precipitation at a temperature of 50" to 60" C.A mixture of antimony and stannic salts is treated with strong hydrochloric acid sufficient to give a concentration of 15 per cent. of the acid. The beaker containing the solution is then placed in a water-bath, and the temperature is maintained at 50° to 60' C.; a strong current of hydrogen sulphide is then passed through the liquid for thirty minutes, and the antimony sulphide separates readily as a scarlet precipitate. The liquid is then cooled below 30" C., and a moderate stream of hydrogen sulphide is passed for ten minutes. The clear liquid is decanted off through a Gooch crucible, previously dried at 110' C. and tared. The precipitate is rapidly collected, washed with 50 C.C.of 15 per cent. hydrochloric acid saturated with hydrogen sulphide, then with water and hydrogen sulphide, and then in succession with alcohol, a mixture of alcohol and carbon bisulphide, carbon bi- sulphide, again with alcohol, and finally with ether. I t is dried at l l O o C. and weighed. The tin remains in the filtrate, which is partially neutralised with ammonia, diluted with water, warmed and saturated with hydrogen sulphide. The results are exact, and are independent of the relative quantities of the two metals. The method is applicable in all cases where antimony is present in the form of a solution of the trioxide. J. F. B. Volumetric Estimation of Small Quantities of Arsenic. L. W. Andrews and H. V. Farr. (Zeit.anorg. Chem., 1909, 62, 122-128.)-The method proposed is an application of Bettendorff's reaction (Zeit. aaal. Chem., 1869,9,105), and consists in precipitating the arsenic in the metallic state by means of stannous chloride; the arsenic is then titrated with iodine and thiosulphate solution. The process is suitable for estimating quantities of arsenic ranging from 0.1 to 100 mgm. The method of operating is as follows : The arsenical solution under examination is neutralised, evrllporated to a volume of about 15 c.c., and transferred to a flask of about 100 C.C. capacity. To the solution is added 2.5 times its volume of stannous chloride solution, prepared by dissolving 20 grams of crystallised stannous chloride and 40 grams of tartaric acid in 1 litre of concentrated hydrochloric acid (containing 40 per cent. HCI) ; the flask is then closed securely, and its contents are kept at a temperature of about 40° C.until the precipitated arsenic has settled down, leaving the supernatant liquid quite clear. This usually takes about three hours. The precipitate is then brought on to an asbestos filter by the aid of small quantities of concentrated hydrochloric acid (free from chlorine), and the flask, precipitate, and filter are next washed completely with water. Air should be excluded as far as possible from the precipitate during the filtration. An excess of about 100 per cent. of the quantity of & iodine solution required for combination with the arsenic, as shown by the equation : As + 51 + 7NaHC0, = Na,HAs04 + 5NaI + 7C0, + 3H,O, is added to the flask, and the filter and precipitate are then introduced.Sufficient 5 per cent, sodium hydrogen carbonate solution is added to keep the mixture neutralTHE ANALYST. 293 during the reaction, but an undue excess is to be avoided. The mixture is thoroughly shaken, and the excess of iodine is then titrated back with + thiosulphate solution, using starch solution as indicator. For quantities of arsenic of less than 0-5 mgm., & iodine and thiosulphate solutions may be used, but a correction should in this case be made for the amount of the iodine solution required to give a reaction with the starch solution. iodine solution is equivalent to 0.15 mgm. of arsenic. Results of estimations, in which quantities of arsenic varying from 0.375 to 75.0 mgm.were taken for the experiment,lshow that the accuracy of the method lies Each C.C. of within limits of kO.8 per cent. w. P. s. An Acidimetric Method of Estimating Alkali Iodides. E. Rupp and F. Pfenning. (Arch. Pharm., 1909, 247, 108-110.)-Mercuric cyanide is readily decomposed by potassium iodide with the formation of a compound iodo-cyanide, in which the alkali cyanide component may be titrated with acid. The following reactions take place : 2Eg(CN), + 2KI = Hg(CN),.Hg12.2KCN. Hg(CN),.Hg12.2KCN + 2HCl= Hg(CN), + HgI, + 2KC1+ SHCN. In making an estimation, the precipitation of the mercuric iodide prevents the recognition of the end-point, and it is therefore necessary to add an excess of acid and to titrate back an aliquot portion of the filtrate with normal alkali solution.Leaving out of consideration the mercuric cyanide in the triple salt, the reaction taking place may be represented by the formula : Hg(CN), + 2KI + 2HC1= HgI, + SHCN + 2KC1. From 1 to 2 grams of mercuric cyanide are dissolved in 50 C.C. of water, and 10 C.C. of a 10 per cent. potassium iodide solution added. Any crystals of the triple salt that separate are dissolved by the addition of more water, and the liquid mixed with 20 C.C. of Fifty C.C. of the filtrate are titrated with The results thus obtained with a solution of known strength ranged from 99.8 to 100 per cent. of the theoretical amount. The presence of chlorides does not interfere with the results ; on the contrary, a small addition of potassium chloride accelerates the filtration.Bromides, however, have a disturbing influence, and must be removed beforehand. C . A. M. hydrochloric acid, made up to 100 c.c., and filtered. alkali solution with methyl orange as indicator. Volumetric Estimation of Mercuric Salts. L. W. Andrews. (Zeit. anorg. Chem., 1909, 62, 171-172.)-1t is pointed out that the process described by Morawitz (ANALYST, 1909, 73) consists essentially in titrating the acidity of the solution with potassium cyanide solution, whilst in the process described by the author (ibid., 1903, 28, 323) sodium hydroxide is employed for the titration. The author sees no advantage in the modification proposed by Morawitz, as the older process is known to give trustworthy results. The end-point of the reaction is sharply defined, and it is difficult to underetand how the addition of 10 drope of TT hydrochloric acid to the mercuric chloride solution, as recommended bg294 THE ANALYST.Morawitz, can have any favourable effect on the reaction or the titration, seeing that hydrochloric acid is liberated in considerable quantity in the solution as the result of the reaction. w. P. s. Comparison of the Electrolytic, Brunck, and Grassmann Methods for the Estimation of Nickel in Steel. Prettner. (Chem. Zeit., 1909, 33, 396 and 411-412.)-According to the author, the electrolytic estimation of nickel is $00 tedious for use in steel analysis. He removes iron by extraction with ether before eleotrolysis, electrolyses in ammoniacal solution for twenty-four hours, filters off the manganese hydrate, again electrolyses the filtrate with a clean cathode to recover the last traces of nickel, and examines the deposits for the small quantities (0.1 to 0.2 per cent.) of manganese they contain.Estimation by means of dicyandiamidine or dimethylglyoxime is much simpler. With the first reagent, to which a little hydrazine sulphate should be added to reduce manganese, precipitation is absolutely complete only after forty-eight hours’ standing ; the second reagent gives com- mercially useful results after one hour’s standing, and exact results after twenty-four hours’ standing. Since the volumetric potassium cyanide method yields accurate results much more quickly, these methods will be used in Ateel and ore analysis chiefly il cobalt is also present. A. G. L. The Estimation of Nitrogen in Nitrates by Means of Stannous Chloride and Iron Filings.A. Kleiber. (Chem. Zeit., 1909, 33, 479.) -Ten grams of substance are dissolved in water and made up to 150 C.C. To 7.5 C.C. of the solution in a 700 to 1,000 C.C. distilling-flask are added 5 grams solid commercial stannous chloride, 15 C.C. concentrated hydrochloric acid, and 4 to 5 grams iron filings; the mixture is heated for fifteen minutes on a water-bat%, or on a wire gauze over a small flame. 90 to 100 C.C. water, a piece of paraffin wax, if necessary, as large as a pea, and about 40 C.C. of concentrated caustic soda solution, are added, and distillation effected with a large flame from the beginning, so that it is complete in half an hour. Twenty C.C. of seminormal sulphuric acid are placed in the receiver. The ammonia left behind in the distilling-flask is found to be constant in amount, and may be allowed for by deducting 0.2 from the number of cubic centimetres of baryta solution used in titrating back.The results are satisfactory. 0. E. M. The Reduction and Estimation of Perchlorates. V. Rothmund. (Zeit. anorg. Chem., 1909,62,108-113.)-Whilst perchlorates are not attacked by the usual reducing agents, such as sulphur dioxide, iron, zinc, sodium amalgam, etc., they are reduced comparatively quickly by titanous salts, and the following method, based on this fact, is described for their estimation : A solution of the perchlorate is mixed with an excess of titanous sulphate solution containing sulphuric acid, 4 C.C. of concentrated sulphuric acid are added, and the mixture is boiled under a reflux condenser for one hour, carbon dioxide being passed through the flask during the whole time.After cooling, the excess of titanous chloride is oxidised by the addition of potassium permanganate, and the chlorine is then estimated by Volhard‘s method. The reduction may be effected by the use of standardised titanous sulphate solution,THE ANALYST. 295 and the excess ti trated back with standardised ferrous ammonium sulphate solution, but in this case the greatest care must be taken to exclude sir from the apparatus during the whole operation. Vanadium and molybdenum salts also reduce prchlorates, tungsten salts act similarly but much more slowly, whilst chromium salts have no reducing action. w. P.s. The Quantitative Separation of Sulphates and Fluorides. R. Ehrenfeld and A. Indra. (Chem. Zeit., 1909, 33, 375-376.)-1n the presence of fluorides sulphates are estimated by mixing the sample with ten to twenty times its weight of zinc-dust in a Rose crucible, covering the whole with a layer of zinc-dust, and igniting over a Teclu or blast-lamp for thirty minutes in a current of hydrogen or coal-gas washed with lead acetate solution. After cooling in the reducing atmosphere, the contents of the crucible are transferred to a flask in which the sulphide formed is decomposed by dilute sulphuric acid, the evolved hydrogen sulphide being absorbed by standard iodine solution. A current of carbon dioxide is led through the flask during the whole operation. To diminish the attack on the glass by the hydrogen fluoride formed, several grams of a mixture of equal parts of precipitated silica and finely divided alumina are placed in the flask.The standard iodine solution used must be neutral; if it is alkaline, high results are obtained. The results obtained are accurate to about 0.4 per cent. on the sulphate present. A. G. L. The Electrolytic Estimation of Thallium. G. Gallo and G. Cenni. ( G a m Chim. Ital., 1909, 39, 285-296.)-The electrolytic estimation of thallium presents difficulties, owing to the fact that thallium has a tendency to separate from solutions of its salts partly in the form of the metal on the cathode, and partly as an oxide upop the anode. In the following method it is deposited completely as an oxide upon the anode.The solution of the thallium salt (sulphate), containing about 0.3 to 0.6 gram i n 100 c.c., is acidified with about 0.1 gram of oxalic acid, and electrolysed at the ordinary temperature in a Classen's dish, in which a disc of platinum is made to rotate as the negative electrode. If the speed be kept at about 800 revolutions per minute, any deposition of thallium upon the cathode is prevented. A current of from 3 to 4 volts and 0.15 to 0.20 ampere is used for the electrolysis. After about an hour the deposition of the thallium oxide begins, and it is advisable t9 continue the electrolysis overnight. The black deposit is washed by decantation with water, with alcohol, and with ether, and then dried in the dish, until constant in weight, in a well-closed air-oven maintained at a temperature of 160" C., and in the lower part of which is a uniform layer of soda-lime.From a series of experiments, the authors conclude that the deposit consists of a new oxide of thallium, T1,0,, which contains 88.48 per cent, of thallium. Applying this factor to the varying quantities of the deposit obtained as described above, the amounts of thallium found in eight test estimations in no case differed by more that 0.0005 gram from the theoretical quantities. C. A. M. Estimation of Tungsten. M. Tschilikin. (Ber. deut. Chem. Ges., 1909, 42 1302-1304.)-Knorre (Ber. deut. Chem. Ges., 1905, 38, 783) described a method for the296 THE ANALYST. estimation of tungsten by means of benzidine hydrochloride ; the author has found that a-naphthylamine is equally suitable for the purpose.The reaction takes place at the ordinary temperature, and gives quite satisfactory results. The benzidine salt should be recrystalliaed from water, and the a-naphthylamine from light petroleum Sodium tungstate, Na2W0, + 2H20, was employed for the tests. The composition of the benzidine compound corresponds with the formula 2(C,,H,,N2).5W0,.5H20. The combination with a-napthylamine has the composition 2(C,,H,N).5W0,.3H20~ For the precipitation of the tungsten, a solution of 25 grams of the amine and 1.5 grams molecular weight of hydrochloric acid are dissolved in water and diluted to 1 litre. To 100 C.C. of this reagent, 25 C.C. of a 2 per cent. solution of sodium tungstate are added at the ordinary temperature. The mixture is allowed to stand for three hours, the precipitate is filtered off and washed with the amine solution diluted with five times its volume of water.The moist filter is then incinerated, and the residue weighed as tungstic anhydride. J. F. B. Test for Nitrites in Potable Waters. A. Rochaix. (L’Union Pharm-, 1909, 50, 62; Pharm. J., 1909, 82, 494.)-Neutral red (toluol red, or symmetric dimethyl-diamino-toluol-phenszine hydrochloride) gives, with a minute quantity of nitrous acid, a blue colour, and may be employed for the detection of this acid in waters. A 0.02 per cent. solution of neutral red is used, 20 C.C. being added to 10 CA. of the water, together with 1 to 3 C.C. of 20 per cent. sulphuric acid. The colour changes from violet to bright blue on shaking, if nitrous acid be present.Strong sulphuric acid will itself give the blue colour with the reagent. If the water be alkaline, a yellow colour may first be seen, but this does not interfere. 0.05 mgm. of nitrous acid per litre of water may be detected by this test. A. R. T. Analytical Examination of Metallic Zirconium. E. Wedekind and S. J. Lewis. (Zeit. nngew. Chew., 1909, 22, 725-729.)-To estimate metallic zirconium in samples of commercial zirconium, 0.5 gram of the sample is placed in a boat in a combustion-tube, and first thoroughly dried, either in a stream of pure hydrogen, or else by evacuating the tube; the moisture may amount to 3.5 per cent. Without admitting even a trace of air, pure chlorine, made from hydrochloric acid and permanganate, and dried by means of sulphuric acid, is then led into the tube, the boat part of which is gradually heated to a red heat, the volatilised zirconium tetrachloride being partly collected in the cooler parts of the tube, and partly in 8, remiver containing dilute hydrochloric acid, and cooled by ice.The boat is then withdrawn from the tube, and the contents of the latter dissolved in water and filtered, together with the solution in the receiver, to free them from any traces of oxide mechanically carried out of the boat. The oxide left in the boat is also thoroughly washed on the same filter to remove traces of chloride left behind ; it is then ignited and weighed, representing the oxide originally present in the sample. The zirconium in the solution is precipitated by adding a considerable excess of strong ammonia and a little ammonium chloride ; the hydroxide obtained is ignited to oxide and weighed. It represents zirconium originally present as metal and as nitride.THE ANALYST. 297 Carbon is estimated by igniting another portion of the sample in oxygen, and weighing the carbon dioxide as usual. On account of the violence of the reaction, the combustion is commenced under highly reduced pressure, oxygen then beiug gradually admitted until the interior of the combustion-tube is again under normal pressure, when the operation is finished as usual. Hydrogen is estimated in the same manner as carbon. Nitrogen is detected as ammonia on fusing the sample with pure sodium hydroxide, or after dissolving it in boiling sulphuric acid. Nitrogen is estimated quantitatively by dissolving in boiling sulphuric acid to which a little hydrofluoric acid or persulphate is added, as solution in pure sulphuric acid requires several days ; the ammonia formed is distilled as usual. The samples examined contained 25 to 84 per cent. of metallic zirconium, besides 12 to 54 per cent. of combined zirconium, 4 to 19 per cent. of oxygen, and nil to 2 per cent. and more of nitrogen. One sample, made by a magnesium process, contained also 0.074 per cent. of carbon and 0.28 per cent. of magnesium. In two samples the hydrogen amounted to about 0-2 per cent. A. G. L.