INORGANIC ANALYSIS 291 INORGANIC ANALYSIS. Separation and Estimation of Aluminium in Presence of Iron by the Action of Acetyl Chloride in Acetone. H. D. Minnig. (Amer. J. Sci., 39, 197 ; through Chern. News, 1915, 111, 172-173.) - The sparing solubility of certain chlorides--e.g., barium chloride and aluminium chloride in a saturated solution of hydrochloric acid in presence of ether or acetone-forms the basis of methods of separation by precipitation of the chlorides. For the separation of aluminium chloride from ferric chloride, acetyl chloride serves as the source of hydrochlorio acid, and the reagent consists of a mixture of four parts of acetone to one part of acetyl chloride.The solution containing ferric and aluminium chloride is evaporated on the water-bath to the smallest possible volume ; the concentrated solution is cooled, and the acetone-acetyl chloride mixture is added drop by drop while stirring, 15 to 20 C.C.being required. The precipitate is collected in a perforated crucible, washed with the precipitating mixture, dried at some distance above a Bunsen flame for ten to fifteen minutes, and then converted into the oxide by ignition. With small quantities of aluminium chloride (about 0.02 grm.of aluminium oxide) the results are perfectly satisfactory, but the subsequent estimation of the iron in the filtrate is affected by the presence of phosphoric acid introduced as an im- purity in tho commercial acetyl chloride. With larger quantities of aluminium (0.05 to 0.18 grm. of the oxide), the precipitated aluminium chloride also tends to contaia phosphate, owing to the larger quantity of precipitating agent required.Pure acetyl chloride may be prepared by passing a rapid stream of hydrogen chloride through a mixture of glacial acetic acid and phosphorus pentoxide, or, preferably, through acetic anhydride at 100' C., which has been purified by distillation over sodium acetate.J. F. B. Estimation of Carbon Dioxide in Minerals. L. H. Borgestrom. (Zeitsch. anal. Chenz., 1914, 53, 685-687.)-1n the estimation of oarbon dioxide in minerals, such as scapolite, the addition of a small quantity of hydrofluoric acid to the hydro- chloric acid used for the decomposition of the carbonates accelerates the rate at which the mineral is dissolved, and causes the whole of the carbon dioxide to be evolved in a comparatively short time (about fifteen minutes).The results obtrtingd are higher and more concordant than when hydrochloric acid ie used alone. The292 ABSTRACTS OF CHEMTCAL PAPERS hydrofluoric acid attacks the flask to some extent, but the same flask may be used for a considerable number of estimations. w. P. s. Estimation of Carbon and Phosphorus in Cerium and Cerium Alloys. H.Arnold. (Zeitsch. and. Chem., 1914, 53, 678-682.)-The total quantity of carbon is estimated by oxidation with a mixture of chromic acid and sulphuric acid. An explosion is, however, liable to occur when cerium is brought into contact with the oxidising mixture unless the following procedure is adopted : Five grms. of the metal are placed in the evolution flask and 60 C.C.of a 20 per cent. copper sulphate solution are added. The metal is disintegrated gradually and coated with copper. The small quantity of hydrocarbon gases liberated at this stage is burnt to carbon dioxide by passing it over heated platinised asbestos. After about one hour 30 C.C. of chromic acid solution (720 grms. of chromic acid in 700 C.C.of water) are added, followed by the gradual addition of a mixture of 100 C.C. of sulphuric acid and 235 C.C. of water. The whole mixture is then boiled for two to three hours, and the process carried out as in the estimation of carbon in iron or steel. Free carbon is estimated by dissolving 10 grms. of the metal in 100 C.C. of 30 per cent. copper chloride solution, adding 100 C.C.of ferric chloride solution (125 grms. of ferric chloride dissolved in 150 C.C. of water and 100 C.C. of hydrochloric acid), and collecting the insoluble residue containing the carbon on an asbestos filter. The filter and its contents are washed and the carbon then estimated by oxidation with chromic acid and sulphuric acid. The phosphorus is estimated by dissolving 10 grms.of the metal in nitric acid, neutralising the solution, and precipitating the rare earths as oxaIates. The filtrate from the latter is evaporated, the excess of oxalic acid decomposed by heating with nitric acid, and the phosphoric acid precipitated with molybdic acid reagent. Four samples of cerium examiued con- tained-total carbon, 0.058 to 0.150 per cent. ; free carbon, 0-037 to 0.074 per cent.; carbide, 0.031 to 0.076 per cent.; phosphorus, a trace to 0.059 per cent. w. P. s. Sensitive Reaction of Chromates. P. N. van Eck. (Chem. WeekbZud, 1915, 12, 6-8.)-Sohtions of chromates give an intense blue coloration with a-naphthylamine in the presence of a little tartaric, citric, or oxalic acid. Tbe reagent is prepared by grinding 0.5 grm. of a-naphthylamine with 50 grms.of tartaric acid in a mortar and dissolving the mixture in 100 C.C. of water. I t will detect 0.001 mgrm. of chromium in the form of chromate. I t may also be used for the colorimetric estimation of chromium, the standard soIution used for the comparison consisting of potassium chromate (0.373 grm.tper litre ; 1 C.C. = 0-1 mgrm. of chromium). Chromic oxide may be fused with sodium carbonate and potassium chlorate, and the coloration obtained in the test matched with that of a solution of chromate prepared in the same way from pure chromic oxide.C. A. M. Volumetric Estimation of Cobalt in the Presence of Nickel. A. Metzl. (Zeitsch. anal. Chem., 1914, 53, 537-541.)- The cobalt and nickel solution, which should not measure more than 100 c.c., is treated with 15 C.C.of 10 per cent.INORGANIC ANALYSIS 293 hydrogen peroxide, 30 C.C. of 5 per cent. sodium bicarbonate solution, and 30 C.C. of sodium hydroxide solution. The mixture is shaken, heated for twenty minutes to destroy the excess of hydrogen peroxide, the precipitate then dissolved by the addition of potassium iodide and sulphuric acid, and the liberated iodine titrated with thiosulphate solution.One molecule of iodine is equivalent to 1 molecule of cobalt. An alternative method depends on the different behaviour of cobalt and nickel towards a mixture of ammonia, ammonium chloride, and an oxidising substance. The solution containing the two metals is treated with 10 C.C. of 16.6 per cent. ammonium chloride solution, 10 C.C. of ammonia, and 20 C.C.of 10 per cent. hydrogen peroxide, then heated for ten minutes ; 50 C.C. of 25 per cent. potassium hydroxide solution are next added, and the solution boiled for about forty-five minutes, or until all the ammonia has been expelled, water being added from time to time to compensate for loss by evaporation. The quantities of reagents mentioned are required for each 0.1 grm.of cobalt present. After cooling, the solution is treated with potassium iodide, acidified, and the liberated iodine titrated. The cobalt precipitate dissolves slowly, and the mixture should be shaken for about fifteen minutes after the addition of the potassium iodide and acid. Results are recorded which show that the method is accurate. w. P. s. Etching Reagents and their Application.0. F. Hudson. (J. Institute ofMetaZs, 1915, advance proof, p. 22.)-This paper was prepared at the suggestion of the Publioation Committee of the Institute, and in its preparation the author received assistance from a number of well-known metallographists. The merits of the various etching reagents are discussed in such a manner that the reader can have no difficulty in selecting a reagent suitable for any particular metallographic purpose, and will as a rule select the best.The method capable of giving the best results in practical hands, however, is not always the best where experience is lacking, and the paper, as its title implies, deals very fully with the application of etching reagents, on which success so much depends, as well as with their com- position.It also deals generally with the preparation of the specimen in order to bring it into a condition suitable for etching, and with the developments of heat- tinting by Stead. Chief prominence is given to the ebching of non-ferrous metals and alloys, but the most important of the reagents used for iron and steel are dealt with briefly. Some space is devoted to methods of polish attack, which the author believes to be deserving of more attention than is generally bestowed on them.No ordinary abstract of this paper, with its carefully balanced statement of conflicting views of eminent experts, is possible. G. C. J. Separation of Gold and Platinum from Other Metals. A. Christensen. (Zeitsch. ana2. Chem., 191’5, 54, 158, 159.)-The method depends on the reduction of gold and platinum by means of hydrazine hydrochloride; gold is reduced in the cold, whilst platinum is reduced only when the mixture is heated.The solution containing gold and platinum and other metals is treated with an excess of hydra- zine hydrochloride and heated for one hour on a water-bath. The precipitated gold294 ABSTRACTS OF CHEMICAL PAPERS and platinum are collected, washed with hot nitric acid to remove mercury and traces of copper, then dissolved in aqua regia, and identified by the usual tests.Metals other than gold and platinum, even arsenic, antimony, and tin, are not precipitated by the reagent. The filtrate from the gold and platinum precipitate may be used for the detection of the presence of other metals; if Petersen's method (ANALYST, 1910, 35, 369) is used for this purpose,'the excess of hydrazine hydro- chloride must be removed previously by evaporating the solution and igniting the residue.w. P. s. Corrosion of Non-Ferrous Alloys. C. H. Deseh. (J. SOC. Chem. Id., 1915, 34, 258-261,)-The defects of laboratory tests, and especially of accelerated tests, are discussed, and a new method is described which is rapid, applicable to small Epecimens, and suitable for the investigation of adherent films.I n the case of alloys composed of two or more micrographic constituents, it was also thought desirable to observe the relative rate of corrosion of those constituents. For this purpose it was deoided to use specimens of,the size usually adopted for metallographic examination, having a polished surface suitable for direct observation by means of the microscope. Further, it was thought advisable to assist and regulate the corrosion by the applica- tion of an external electromotive force.Objections have been urged against such a procedure, on the ground that chemical and electrolytic corrosion are different in character, but experimental evidence is adduced by the author which shows clearly that the mechanism of the two processes is identical, whilst the conditions of a, rapid laboratory test, cannot be fulfilled without the aid of an applied electro-motive force.The vertical brass rod of a stand is divided into two parts, insulated from one another by a short section of ebonite. The lower rod carries a brass plate which may be clamp3d at any height, and a binding Bcrew for connection with the positive pole of some source of electrical energy, usually a storage battery, provided with a distributing board and measuring instruments for electrolytic analysis.To the brass plate is attaqhed a brass block, carrying a pair of spring clips, between which the specimen, 12.5 mm. square, is held with the polished horizontal surface upwards. The cathode is a piece of fine platinum gauze, 10 mrn.square, attached to a vertical wire suspended from the outer end of a brass arm attached to the upper half of the vertical brass rod of the stand by means of a sliding plate whicb can be raised or lowered by a rack and pinion, a scale allowing the height to be adjusted accurately.The cathode is lowered until in cQntact with the specimen, and then raised 5 mm. A wall of plasticine is built up to contain the electrolyte, which is usually 1 or 2 C.C. of a 5 per cent. solution of sodium chloride. Corrosion is allowed to proceed for five to sixty minutes, according to circumstances. The wirea are disconnected and the electrolyte is rinsed out of the plasticine cell into a beaker, using a wash-bottle With a fine jet. A loose, flocculent precipitate is usually obtained.Any solid deposit adhering to the corroded Burface so loosely that, it is detached by light rubbing with the finger-tip may usually be added to the bulk : but a firmly adherent deposit, requiring the use of a wooden chisel-edge or a knife-blade to detach it, should be collected, if present, for separate A special apparatus is figured in the paper.INORGANIC ANALYSIS 295 analysis The surface of the corroded metal is examined under the microscope both before and after the removal of the adherent layer.The results of experiments with various brasses are given in the paper, together with the conclusions that may be drawn from them, among them the conclusion that the process of corrosion by sea and other natural waters is of essentially the same character as that of electrolytically stimulated corrosion under the conditions described above.G. C. J. Comparison of Methods for the Analysis of the Higher Lead Oxides. J. Milbauer and B. Pivnicka. (Zeitsch. anal. Chem., 1914,53, 569-580.)-Bunsen’s methoa, which depends on the reaction PbO, + 4HC1= PbCI, + 2H,O + GI2, yields results which are at the most 0.1 per cent.too low. The modification of LUX’S method described by Chwala and Colle (ANALYST, 1911, 36, 366) tends to give results which are slightly too low, whilst the figures obtained by Finzi and Rapuzzi’s method (ANALYST, 1913, 38, 343) are too high by about 1.4 per cent. The authors have endeavoured to estimate lead dioxide gasometrically by a process based on the reaction YbO, + 2NB,OH + 2KOH = Pb(OK), + 4H,O + N, ; the reaction proceeds in the cold, but the results obtained are invariably too low, owing to the retention of nitrogen by the reacting solution.w. P. s. Analysis of Litharge. P. Beek. (Zeitsch. anal. Chenz, 1915, 54, 137-147.)- A complete analysis of litharge is seldom required when the material is to be used for technical purposes, but the estimation of certain impurities contained in most specimens of litharge is of importance. For the estimation of copper, 100 grms.of the sample are dissolved in nitric acid, the lead is precipitated as sulphate and separated, and the solution is then diluted to 1 litre. According to the amount of copper present, a quantity of this solution, varying from 10 to 100 c.c., is treated with an excess of ammonia, aluminium and ferric hydroxides are separated by filtration, the filtrate is acidified with dilute sulphuric acid, and the copper deposited electro- lytically.If the colour of the deposited copper indicates the presence of other metals, the metal is dissolved in nitric acid, the solution treated with ammonia and ammonium carbonate, any precipitate (bismuth hydroxide) which forms is separated, and the copper again deposited electrolytically.Litharge for use in pottery and glass manufacture should not contain more than 0.004 per cent. of copper, or more than 0*006 per cent. of iron. These two impurities may be estimated in one portion of the sample as follows : One hundred grms. of the litharge are dissolved in nitric acid, the lead is precipitated as sulphate, washed with dilute nitric acid, and the filtrate and washings are evaporated and heated until fumes of sulphuric acid are given off; after dilution, the solution is filtered to remove any lead sulphate, and the copper precipitated as sulphide in the filtrate.The copper sulphide is collected, dissolved in nitric acid, evaporated with sulphuric acid, any traces of lead sulphate are separated, bismuth is removed by treatment with ammonia and ammonium carbonate, and the copper is again precipitated as sulphide, ignited, and weighed as oxide.The iron, contained in the filtrate from the first copper sulphide precipitate, is estimated gravimetrically in the usual way.296 ABSTRACTS OF CHEMICAL PAPERS Litharge intended for use in the manufacture of accumulators should be free from nitrites and nitrates, and should contain not more than traces of iron, copper, or other metals precipitated by hydrogen sulphide; the chlorine content should not exceed 0.05 per cent.The quantity of metallic lead in litharge may be estimated by mixing 100 grms.of the sample with 700 C.C. of water, adding 70 C.C. of nitric acid (sp. gr. 1.4) in small quantities at a time 80 that the solution does not become heated, collecting the insoluble metallic lead on a filter, then dissolving it in hot dilute nitric acid, and estimating it electrolytically or gravimetrically. Insoluble substances- e.g., lead sulphate, silica, etc.-are best estimated by heating 100 grms.of the litharge with 750 C.C. of water and 80 C.C. of nitric acid (sp. gr. 1.4) until lead oxide, metallic lead, etc., have dissolved, and collecting the insoluble portion on a filter ; the filter and its contents are washed six times with hot water, dried, ignited in a porcelain crucible, and the residue weighed. w. P. s. Estimation of Manganese in Soils.B. von Horvath. (Zeitsch. anal. Chem., 1914,53, 581-593.)-The colorimetric method described by Marshall (ANALYST, 1901, 26,195) and depending on the oxidation of manganese salts to permanganate by ammonium persulphate in the presence of silver nitrate (see also ANALYST, 1914, 39, 274) is recommended as being the most suitable process for the estimation of small quantities of manganese in soils.The hydrochloric acid extract of the soil is twice evaporated to dryness with nitric acid to expel chlorine and destroy organic substances, the residue is then heated with sulphuric acid to remove nitric acid, and the estimation proceeded with as described. Gravimetric and volumetric methods were found to be untrustworthy for the purpose. w. P.s. Estimation of Potassium by the Perchlorate Method. R. G. Thin and A. C. Cumming. (J. Chem. Soc., 1915, 107, 361-366.)-The authors find that this method can be depended on to give results of very satisfaetory rtcouracy. The error need never exceed 0.5 mgrm. on the weight of the potassium perchlorate. A neces- sary condition is that the precipitate shall be washed only with alcohol saturated with potassium perchlorate.This wash liquor wag recommended by Davis in a paper (ANALYST, 1913, 38, 47) in which he strongly recommended the method, but most of his test results given in that paper were obtained before he substituted this wash liquor for a weak solution of perchloric acid in strong alcohol. Consequently Davis’s test numbers, though good, are not exact, and give the impression that the method, though the best available, leaves something to be desired, whereas the authors now show that it is fully entitled to the description of an exact as well as a convenient, rapid, and cheap method.Attention is called to the fact that Mercb’s so-called ‘‘ pure ” perchloric acid (20 per cent.) is liable to contain appreciable traces of potassium perchlorate.G. C. J. Radiometric Measurements of the Ionisation Constants of Indicators. E. J. Shaeffer, M. G. Paulus, and €I. C. Jones. (J. Amer. Chem. Soc., 1915,37, 776-807.)-A satisfactory radiomicrometer, having a half-period of ten seconds and a sensibility of 5 per sq. mm. of exposed vane (candle and scale being at a metre’sINORGANIC ANALYSIS 297 distance), was constructed.Making use of the radiomicrometer and a grating spectroscope, a radiometric method was worked out for the determination of the ionisation constants of indicators. This method, which serves as well for a twc- coloured indicator as for a one-coloured indicator, is freer from objections and limitations than any method previously used. For the proof of the above claim, the original paper must be consulted. Briefly, the experimental procedure is as follows : The radiomicromoter is a thermo-electric junction attached to a loop of non-magnetic wire, the whole system being suspended by a quartz fibre in a glass tube.A strong magnetic field surrounds that portion of the tube enclosing the loop of wire. Radiant energy falling upon ,the blackened junction is converted into electrical energy, and the suspended system rotates in the magnetic field, the loop tending to set itself at right angles to the lines of force.The deflection is given by means of a mirror attached to the suspended system, Parallel light from a Nernst lamp is made to pass through the solution of the indicator, and is then deflected by a right-angle prism and focussed on the slit of the spectroscope, sl.Directly in the rear of this slit is another right-angle prism, which reflects the light through a 4-inch lens to the grating. The dispersed light is reflected back through this same lens and focussed on the slit of the spectroscope, s2, and the light emerging through this slit is again brought to a focus on the junction of the radiomicrometer.By turning the drumhead of the spectroscope it is then possible to determine quantitatively the light transmitted by any solution for all wave-lengths of light between A= 0.4 p and X = 2.0 p. Very small concentrations of coloured components were determined, and it is shown that minute concentrations of hydrogen and hydroxyl ions can be quickly and accurately estimated by means of radiometric measurements.Satisfactory constants were obtained for the ionisation of methyl orange as a base. The value found was 2.1 x 10 -11. The ionisation and hydrolysis constants for phenolphthaleln considered as a monobasic acid are far from being satisfactory. From the known ionisation constant of methyl orange and from radiometric measurements, the ionisation constant of a very weak base and the hydrolysis constant of one of its salts have been roughly determined.The method can likewise be applied for the determination of the ionisation constants of very weak acids, and the hydrolysis constants of the salts formed by these acids. G. C. J. Estimation of the Reducing Power of Natural Waters. L. W. Winkler. (Zeitsch. anal. Chem., 1914, 53, 561-564.)-The quantity of permanganate reduced in alkaline solution by natural waters (cj.ANALYST, 1902, 27, 341) may be estimated by treating 100 c.c., or less, of the water with 10 C.C. of alkaline i$a permanganate solutioo ; after standing for twenty-four hours, the mixture is acidified with sulphuric acid, potassium iodide is added, and the liberated iodine titrated with thiosulphate solution.The following procedure is recommended for waters containing nitrites, ferrous salts, and manganese salts: 100 C.C. of the water are treated with 1 C.C. of 10 per cent. sodium hydroxide solution and set aside for five minutes ; ferrous and manganous salts are thus oxidised by the dissolved oxygen in the water. The per- manganate solution is then added, and, after twenty-four hours, the mixture is298 ABSTRACTS OF CHEMICAL PAPERS 8.42 9.51 10.30 11.08 11.46 11.75 acidified with 25 per cent.phosphoric acid solution (this is preferable to sulphuric acid; as it prevents the action of the ferric salts on potassium iodide), potassium iodide is added, and the iodine titrated with thiosulphate solution. Another portion of the water is also treated with sodium hydroxide solution to oxidise the ferrous and manganous salts, permauganate solution is then added, and the excess titrated after the lapse of two minutes. The nitrites are thus oxidised whilst the organic matter is not appreciably affected; the result of this titration is a measure of the substances other than organic matters, which reduce permanganate.w. P. s.8.42 9.56 10.33 10-96 11-33 11.64 Comparison of the Relative Drying Powers of Sulphuric Acid, Calcium Chloride, and Aluminium Trioxide, when used in Ordinary Scheibler Desic- eating Jars. J. W. Marden and V. Elliott. (J. Ind. and E'ng. Chem., 1915, 7, 320-321.)-The water-vapour pressure of some materials still containing residual moisture is as low as that of the sulphuric acid over which they have been dried. By graphic interpolation it can be shown that concentrated (95 per cent.) sulphuric acid has a vapour pressure of about 0.03 mm.at 25" C. According to Thorpe (Dict. App. Chern., 1912,2,210) the speed of drying is hastened, but no more moisture is obtained by the use of vacuum. The authors found ,that with cheese, syrup, coffee, etc., using both sulphuric acid and calcium chloride, slightly more moisture was obtained with vacuum than without, in one case 0.7 per cent., in others quite small.It was found that the concentration of the sulphuric acid had little effect on its drying power for cane syrup and cheese between 95 and GO per cent. ; but with the former strength flour lost 12.48 per cent., and with the latter only 2.9 per cent., of water.Maple syrup lost as much moisture in one and a half hours when the air in the desiccator was stirred with a small motor as it lost in six hours without stirring. The following table shows the moisture lost to three different drying sub- stances kept at 25' C. The results were in the same order when vacuum desiccators were employed. Time Days. 1 2 3 5 7 13 Flour. CaC12. 8-47 9.21 10.06 10.62 10.96 11.15 14.12 18.76 23.73 29-36 29.53 29.53 H2S04 15 per Cent, 15.08 19.47 25.51 29.46 29.54 29-54 CaCI2. 15.22 16.84 21-31 28-77 29.42 29-42 Coffee. 1.31 1.98 2.66 3.30 3'54 3.66 CaClz . -- 1.17 1-02 2.38 3.11 3.32 3.46 H. F. E. R.REVIEWS 299 Estimation of Dissolved Oxygen in Water. L. W. Winkler. (Zeitsch. anal. Chem., 1914, 53, 665-672.)-The following modification of the author’s well- known method is described for the estimation of dissolved oxygen in waters containing nitrites and organic substances which interfere with the process : To the bottle containing some of ‘the water under examination, manganous chloride solu- tion and sodium hydroxide solution free from potassium iodide are added in the usual way, and a current of carbon dioxide passed in. The manganous hydroxide is thus converted into manganous carbonate and manganous hydrogen carbonate ; about ten minutes’ treatment with carbon dioxide is sufEcient. The precipitate is now collected on a filter, washed with a 2 per cent. potassium hydrogen carbonate solution, then dissolved in dilute hydrochloric acid containing potassium iodide, and the liberated iodine titrated. The manganous carbonate is not oxidised by atmos- pheric oxygen during the filtration. w. P. s.