92 ABSTRACTS OF CHEMICAL PAPERS INORGANIC ANALYSIS. Colorimetric Method for the Estimation of the Carbon Dioxide in Air. H. L. Higgins and W. M. Marriott. fJ. Amer. CRem. SOC., 1917, 39, 68-71.)-The method is based on the principle that when a current of air con- taining carbon dioxide is passed through a solution of sodium bicarbonate until the latter is saturated with carbon dioxide, the reaction of the saturated solution (hy- drogen-ion concentration) will depend on the relative amounts of sodium bicarbonate and carbon dioxide present, and this will depend on the partigl pressure of the carbon dioxide in the air-mixture used, regardless of the volume of air passed through, provided the saturation-point has been attained.The reaction of the solution is determined by adding to it a suitable indicator, such as phenolsulphonephthalein, on the principle of Sorensen’s method, and comparing the result colorimetrically with solutions of known hydrogen-ion concentmtion.Standard solutions are pre- pared by mixing M/15 solutions of acid potassium phosphate (KqPO,) and &sodium hydrogen phosphate, both carefully purified by crystallisation, and the latter dehy- drated to a stable condition corresponding to N%HPO,*ZH,O by exposure to the air for several days.These solutions are each made up with 200 C.C. of 0.01 per cent. solution of phenolsulphonephthalem and diluted to 1 litre. A series of standard mixtures of these two solutions is prepared in sealed test-tubes, the proportions being determined empirically to correspond with definite percentages of carbon dioxide in air analysed by a chemical method.For applying the method, two standard bicarbonate solu- tions may conveniently be employed : A is ~+c, snd is prepared by taking 10 C.C. of .f6 sodium hydroxide solution, adding 200 C.C. of phenolsulphonephthaleln solution, and making up .to 1 litre. B is for higher percentages, and is prepared similarly from 85 C.C.of 6 sodium hydroxide. They are converted into bicarbonate by passing carbon dioxide through them from a cylinder. For the analysis, 2 to 3 C.C. of one of the bicarbonate solutions are placed in a test-tube of the same diameter as the Bta;nda,rd tubes, but 100 to 150 mm. long. Air from the room is then blown in by means of an atomiser bulb and a glass tube drawn out to a fine point until the solution is saturated, when no further change in colour is observed.The test solu- tion is then compared with the scale of standards, which are devised to show directly the percentages of carbon dioxide, the same set of standards used with B showing ten times the values indicated with A. A colorimetric box for comparing three tubes a t a time may be conveniently employed.J. F. B.INORGANIC ANALYSIS 93 Estimation of Fluorine in Soluble Fluorides. J. G. Dinwiddie. (Amer. J . Sci., 1916, 42, 464-468; through J. SOC. Chem. I d . , 1917, 36, 30.)-A neutral solution of the fluoride is heated to boiling, and powdered chlcium sulphate is added; after standing for one hour, with frequent stirring, the precipitate, con- sisting of calcium sulphate and calcium fluoride, is washed several times by decanta- tion and then collected on a filter.The latter consists of a disc of filter-paper fitted into the bottom of a perforated platinum crucible. The precipitate is now washed (the wash-water used should be saturated previously with calcium sulphate and calcium fluoride), then rinsed into an ordinary platinum crucible, and the water evaporated; the disc of filter-paper is, meanwhile, ignited on the crucible lid and the ash introduced into the crucible.The dry contents of the crucible are then heated at 300" C. for one hour, or until constant in weight, then sulphated, again heated a t 300" C., and weighed. The increase in weight after sulphating is due to the replace- ment of two atoms of fluorine by the sulphuric acid radicle, and a simple calculation gives the quantity of fluorine present.The error of the method is about 0.1 per cent. Reactions for Distinguishing between Perehlorates, Periodates, Per- sulphates, Perearbonates, and Perborates. A. Monnier. (Ann. Chim. Analyt., 1916, 21, 237-240 ; through J . sbc. Chew. Id., 1917, 36, 30.)-The following reactions are given for the detection and identification of per-salts in aqueous solution, but the latter must be free from heavy metals, vanadates, tungstates, molyb- dates, bichromates, and ferricyanides.Perborates and Percarbonates,-These salts give the general reactions for boric acid, carbon dioxide, and hydrogen peroxide. To detect their presence when mixed with borates and carbonates, a portion of the solution is treated with dilute potassium bichromate solution acidified with sulphuric acid; a blue coloration indicates that a percarbonate or perborate is present.To identify a perborste, a solution of chrome alum is treated with an excess of the solution to be tested, ether and dilute sulphuric acid are added, and the mixture is shaken gently; the ethereal layer is coloured blue.Percarbonates and hydrogen peroxide do not give this reaction. Iodates and Periodates.-To the solution is added a small quantity of chloroform and then titanium trichloride solution; a white pre- cipitate is formed and iodine is liberated, giving a violet-coloured solution with the chloroform. Titanium trichloride also liberates iodine from potassium iodide in the presence of chlorates or perchlorates, but in this case a white precipitate is not pro- duced.The reaction with silver nitrate serves to distinguish periodates from iodates. If an alcoholic solution of benzidine is poured on the surface of the solution under examination, a brown zone appears in the presence of periodates, but not with iodates. Chlorates and P'erc7iZorates.-A portion of the solution is shaken with the addition of potassium bromide, chloroform, and titanium trichloride solution ; chlorates liberate bromine, and this gives a brown chloroform solution, but perchlo- rates do not give the reaction. Another portion of the solution is treated, drop by drop, with a 0.2 per cent methylene blue solution; a violet precipitate, which deto- nates when heated on a piece of platinum foil, indicates the presence of a perchlorate.In testing for chlorates and perchlorates, iodides, if present, must be removed pre-94 ABSTRACTS OF CHEMICAL PAPERS viously by treatment with silver oxide, and iodates and periodates by means of silver nitrate, the excess of the latter being then removed by the addition of potassium bromide.Detection of Bromates in the Presence of Chlorates and Iodates.-To the solution is added manganous sulphate solution acidified with sulphuric acid; in the presence of a bromate the solution is coloured violet, owing to the formation of manganic sulphate, and bromine is liberated. This reaction is not given by chlorates or iodates. PersuZphates.-In the absence of iodides and perchlorates, a rose-red precipitate is obtained when a persulphate is treated with methylene blue solution.Persulphates may be identified in the presence of perchlorates by pour- ing an alcoholic benzidine solution on the surface of the solution t o be tested; a blue zone a t the junction of the two liquids denotes the presence of it persulphate. A 2 per cent. aniline solution may be used in place of the benzidine solution. Combustion Methods for Use in the Laboratory (Estimation of Sulphur).J. Hewett. (London and Southern Distr. Jun. Gas Assoc., Jan. 5, 1917. Gas J . , 1917, 137, 72-73 ; through J . Xoc. Chem. Ind., 1917, 36, 165.)--Sulphur may be estimated in volatile liquids--e.g., crude benzols-by drawing air through the liquid and, after further diluting the air- vapour mixture, passing through a silica tube heated to redness, when complete combustion occurs.The gases then pass through two wash- bottles containing sodium hydroxide. After oxidation of the product with bromine or hydrogen peroxide, sulphur is estimated as sulphafe in the usual way. The author indi- cates the wide applicability of the estimation of sulphur by combustion with sodium per- oxide.It is carried out in a nickel crucible (see illustration) fittedwith a screwed top and electrical ignition device. A ring of filter- paper makes a tight joint. The crucible may be immersed in water when the charge Is fired. The substance under test must be thoroughly incorporated with the sodium peroxide in suitable proportions. The reaction should be complete in a few seconds.The c0ntent.s of the crucible, when cold, are extracted with distilled water, and, after acidification and filtration, sulphur is estimated as sulphate in the usual way. This method may be used for any substance which can be brought to react vigorously with sodium peroxide, either alone or with the addition of sulphur-free combustible matter-e.g., wood dust or alcohol.It has been used with tar and tar products, oils, fats, greases, coal, coke, new and spent oxide of iron. With spent oxides the reaction product is suitable for the estimation of iron and alumina as well as sulphur. The analysis of spent oxide in this way can be conducted much more speedily and conveniently than by the usual methods. Not more than 1 grm. of combustible substance should be taken, and then 20 grms.of sodium peroxide usually give a suitable melt. The substance should be approxi- mately free from moisture, as water may cause dangerous premature ignitions. The reaction is extremely violent.INORGANIC ANALYSIS 95 Removal of Phosphoric Acid by Stannic Chloride. L. Gattermann and H. Schindhelm. (Ber., 1916, 49, 2416-2422 ; through J.SOC. Chem. I d . , 1917, 36, 165.) -Working with. 1 grm. of material, the filtrate from the hydrogen sulphide group is boiled until free from hydrogen sulphide, concentrated to 100 c.c., filtered if necessary, and treated with 5 C.C. of 6 per cent. ammonium phosphate solution. It is then nearly neutralised with dilute ammonia (until only faintly acid to litmus), 3 C.C.of dilute hydrochloric acid (7 to 8 per cent.) are added, and the briskly boiling solution is treated with 3 C.C. of a freshly prepared solution contain- ing 3 grms. of crystallised stannic chloride, dissolved in the cold. Stannic phosphate is thus precipitated, and, if' necessary, more of the stannic chloride solution is added in portions of 0-5 to 1 c.c., to the boiling solution until no further precipitate is ob- tained on testing a little of the filtered liquid with ammonia molybdate.The boiling solution is then filtered, using a Biichner funnel and double filter, the precipitate washed with boiling water, the warm filtrate, which should be quite clear, is treated with dilute hydrochloric acid and hydrogen sulphide to remove excess of tin, and the metals present are detected in the usual way.When the phosphate precipitate has a greenish colour, it is tested with caustic soda and hydrogen peroxide for chromium, which, like ferric (but not ferrous) iron, is liable to be carried down with the precipitate. If too little phosphoric acid be present, the mixture containing the stannic phosphate precipitate is difficult to filter, but satisfactory conditions are insured by adding ammonium phosphate as directed.Estimation of Free Sulphuric, Nitric, and Picrie Acids in the Presence of Each Other. F. W. Richardson. (J. SOC. Chem. Id., 1917,36,13-15.)-Methods are described for the estimation of sulphuric, nitric, and picric acids in efflbents from picric acid works. The picric acid is estimated colorimetrically in a Lovibond's tintometer ; the yellow collour of the acid is eady comparable with the yellow units of the instrument, and there is no difficulty in standardising the yellow units against standard picric acid solutions.The use of a cell giving about 1 inch depth of solu- tion will be found to be the most convenient. Since the coloration of a picric acid solution varies with the temperature, the observations should be made a t some constant temperature-my, 15" or 20" C.The nitric acid is also estimated colori- metrically by means of phenoldisulphonic acid, allowance being made for the colour contributed by the picric acid, and the sulphuric acid is estimated by difference, the three acids being titrated in the presence of methyl-red, and the quantities of nitric and picric acids deducted from the result. For example, an effluent when examined in the tintometer showed a coloration equivalent to 0.28 grain of picric acid per gallon.Ten C.C. of the effluent were mixed with 5 C.C. of & sodium hy- droxide solution, evaporated to dryness, the residue was warmed for five minutes with the addition of 1 C.C. of phenoldisulphonic acid mixture, then diluted, rendered alkaline with ammonia, cooled to 15" C., diluted to 100 c.c., and observed in the tintometer.In this case the yellow units of the instrument were standardised against definite quantities of nitric acid under similar conditions ; in calculating the quantity of nitric acid alllowance was made for the colour of the picric acid at the corresponding dilution (ten times). Seventy C.C. of the effluent were then titrated96 REPORTS with & sodium hydroxide solution, using methyl-red as the indicator; after de- ducting the quantity of alkali neutralised by the nitric acid and picric acid (1 C.C. of & sodium hydroxide solution corresponds with 0-00229 grm. of picric acid), the remaining amount of sodium hydroxide solution used for the titration was equiva- lent to the quantity of free sulphuric acid present. With regard to the action of water containing small quantities of sulphuric acid and nitric acid on iron, the author finds that nitric acid may produce slightly protec- tive coatings on the metal, but that this coating does not form when sulphuric acid is also present. The concentration of the water in boilers leads to exfensive corrosion. Picric acid also acts readily on iron; as compared with sulphuric acid, the amount of iron dissolved is roughly proportional to the molecular equivalents, 49 parts of sulphuric acid producing the same effect as 229 parts of picric acid. W. P. S.