INORGANIC ANALYSIS 217 INORGANIC ANALYSIS. Reduction of Arsenic to the Arsenious State by Cuprous Chloride and Estimation of Arsenic by Distillation as Trichloride. R. C. Roark and C. C, MacDonnell. (J. Ind. and Eng. Chem., 1916, 8, 327-331.)-1n the analysis of arsenical insecticides, the authors obtained very low results when attempting to estimate the arsenic by reduction with ferrous salts followed by distillation a s trichloride.Yet this method has often been recommended for the estimation of arsenic, the recommendation being supported by excellent test numbers. 9 review of the literature, however, discloses the fact that in all these cases of the successful use of the method one or more of the following conditions obtained : (a) the arsenic was present in very small amount, ( b ) it was present as element or as arsenite and not as arsenate, (c) copper or a copper salt was present.Cuprous chloride in hydrochloric acid solution effectively reduces arsenic com- pounds to arsenious chloride, which can then be readily distilled in a current of hydro- gen chloride. The sample, containing not more than 0.4 grm. of arsenic, is washed into a distilling flask with 100 C.C.of hydrochloric acid (sp. gr. 1*19), 5 grms. cuprous chloride are added, and the mixture distilled until only about 40 C.C. remain in the fiask? when an additional 50 C.C. of acid are added through a tap funnel and distillation continued, this process being repeated until 200 C.C. have been distilled. The distillate is collected in a series of wash-bottles.The first should not contain more than 40 C.C. of water, otherwise a compound of arsenic may separate which cannot readily be redissolved. The adapter from the condenser penetrates the cork of this wash-bottle, but is not sealed in the liquid. A safety tube about twelve inches long penetrates the cork and is sealed in the water. The outlet tube connects to a bend which dips into 100 C.C.of water contained in the second wash-bottle. Both these bottles should be surrounded by ice and water. A third wash-bottle is added as a precaution, but it is unusual for any arsenic to get bsyond the second. The distillate and washings from the flasks are nearly neutralised with sodium hydroxide, using phenolpthalein as indicator, and keeping the solution cool.Sodium bicarbonate is then added in excess, and the solution titrated with iodine solution in the usual manner. G. C. J.218 ABSTRACTS OF CHEMICAL PAPERS Estimation of Carbon in Steels and Irons by Direct Combustion in Oxygen at High Temperatures. J. R. Gain and H. E. Cleaves. (J. I d . and Eng. Chem., 1916, 8, 321-324.)-The general consensus of opinion seems to be that higher results are obtained with higher combustion temperatures, but this conclusion is rendered doubtful by the fact that the recorded differences are mostly no greater than the unavoidable error of a blank determination, and may quite possibly be due to other causes than temperature, such as variation in size of drillings.Moreover, the published work does not indicate that investigators have always assured them- selves that the fluxes and material used to support the drillings were completely free from carbon. The authors have re-investigated the point and to this end set up rather elaborate apparatus permitting the maintenance of temperatures up to 1500" C. As a result they conclude that the certificates of carbon content sent out with samples by the Bureau of Standards', based as they are on simpler methods of analysis, are not in error by more than 0-015 per cent.(generally minus), and that in most cases the error is much less. I n view of this confirmation of the accuracy of simpler methods of direct combus- tion, the authors do not recommend their elaborate apparatus and somewhat incon- ' venient method, but these are fully described in the paper, as they may possibly prove useful in the analysis of some special alloys.G. C. J. Estimation of Carbon in Steel by Eggertz Method. H. Le Chatelier and F. Bogitch. (Comptes rend., 1916, 162, 709-714.)-Concentrated nitric acid eBects more rapid solution of the carbon than does weaker acid, but tends to retard or even prevent complete solution of the iron. The authors, therefore, treat 1 grm.of the borings with 20 C.C. of acid of sp. gr. 1-16, heat rapidly to boiling during one minute, maintain in ebullition another minute, then add 30 C.C. of acid of sp. gr. 1.33 and continue to boil for three minutes longer. After cooling rapidly for one minute, comparison is made with standards in the usual manner, The intensity of the final coloration being dependent on the temperature of digestion, as well as other condi- tions, the authorsprefer to make this digestion at the boiling-point, which is fixed by the acid concentration, rather than to make u ~ e of a bath at SO" C.or other tempera- ture, which it is less easy to keep uniform. Experiments are described showing the influence on the final colour intensity of the temperature of digestion, duration of heating, concentration of acid and period of exposure to diffused daylight before the comparison is made, also the necessity for preserving standgrds from unnecessary exposure to light and for their frequent renewal.The disturbing effect of chlorides in the nitric acid is also demonstrated and the fact that sulphuric acid does not interfere. G.C. J. New Method of Estimating Fluorine. F. Pisani. (Comptes rend., 1916, 162, 791-793.)-0n adding thorium nitrate to the solution of an alkali fluoride, slightly acidified with acetic or nitric acid, a gelatinous precipitate of thorium fluoride, ThF4+4H,O, is obtained. When dried at 100" C. this loses 1 molecule of water, but for estimating fluorine it is preferable to ignite the precipitate and weigh theINORGANIC ANALYSIS 219 residue of thorium oxide, Tho,.Too large an excess of the reagent must be avoided, since the precipitate is somewhat soluble in strong thorium nitrate solution. Subject to this precaution the test is capable of detecting 0.01 per cent. of fluorine. In the case of certain insoluble fluorides preliminary fusion with 5 parts of sodium car- bonate is necessary, the melt being then taken up with water, and the solution acidified with acetic acid.Hydrofluosilicic acid precipitates thorium in the cold from solutions rendered slightly acid with hydrochloric acid, and is thus a sensitive reagent for thorium even in the presence of cerium, lanthanum, and didymium. Thorium nit rate precipitates h ydrofluosilicic acid quantitatively from soluble hydro- fluosilicates, while insoluble hydxofl uosilicates may first be decomposed by boiling or fusion with sodium carbonate, and separation of the silica by treatment with ammonium carbonate.Hydrofluotantalates and hydrofluoborates &re boiled with sodium carbonate solution, the tantalum filtered off, and the filtrate treated with acetic acid and thorium nitrate as described.C. A. M. Rapid Approximate Assay for Lead. G. Torossian. (J. Ind. and Eng. Chem.,. 1916,8,33l.)-An aluminium plate, 5 inches by 2 inches by 0.03 inch, is required. It serves for about 30 estimations. Near each end of the plate a cup-shaped depression is made by laying the plate on the thumb-hole of a crucible tongs and striking with a pestle. The diameter of the cup need not exceed 1 inch, nor its depth i$ inch.Before use, the plate is rubbed with emery-cloth to remove the protective coating of oxide. From 0.15 to 0.2 grm. of the finely-powdered sample is placed in one of the cups, moistened with 1 or 2 drops of dilute (1 : 3) hydrochloric acid, and more acid is added, drop by drop, until the action of the acid on the aluminium is well started.The sample is subjected to this treatment for several seconds, or until the colour of the sample (if coloured) disappears or is notably changed. By this time the sample becomes spongy, may be turned over by st pointed glass rod, or gently stirred and more acid added if necessary. After a minute or two the contents of the cup are stirred with a pointed glass rod for about five minutes, or until all indications point to the completion of the reaction, such indications being the disappearance of the colour of coloured compounds, or the disappearance of the smell OF hydrogen sulphide in the case of sulphides.The spongy metallic lead is washed by decantation four or five times until free from acid, and is then pressed together with a glass rod to make it solid and compact, after which a piece of filter-paper is pressed on it to dry it as far as is possible in this way.Finally, 1 drop of water is placed near the cup on the plate, which is then warmed above a small gas flame or lighted match until the water is evaporated. When the water is evaporated, the lead may be assumed to be substantially dry. The dried lead is then detached from the plate and weighed.The test results with lead oxides, carbonate, sulphate, and acetate are all within I per cent. of theory, but there are no test results on naturally occurring minerals, although the author recommends the method as an approximate one for mining engineers and prospectors. G. C. J. Estimation of Silica. V. Lenher and E. Truog. (J. Amer. Chem.SOC., 1916,. 38, 1050-1063.)-Experiments are described which demonstrate the fact that the220 ABSTRACTS OF CHEMICAL PAPERS failure to recover all the silica in one evaporation is not so much due to insufficient dehydration as to the opportunity afforded by the large mass of silica for its direct solution during the acid treatment. The authors, therefore, do not attempt to carry the dehydration of the first and principal separation of silica anywhere near comple- tion, a course which has the further advantage that the bulk of the silica is less contaminated by bases.Moreover, the first evaporation is carried out as expeditiously as possible to minimise the time during which the acid acts on the bulk of the silica. The residue from the evaporation of the filtrate is more tboroughly dehydrated, but not above 110' C., nor for more than two hours, longer heating showing no advantage, whilst higher temperatures lead to the contamination of the silica by bases, notably by magnesia.The acidified solution of the fusion is evaporated in the usual manner, but evaporation is stopped when the residue begins to powder. The residue is moistened with hydrochloric acid of sp.gr. 1.1, and not with water or concentrated acid, either of which appears to exercise a greater solvent effect. After digestion on the water-bath for ten minutes, 10 C.C. of water are added and filtration proceeded with immediately, the silica, being washed with hot 5 per cent. hydrochloric acid. The filtrate is evaporated, the residue dehydrated at 110" C.for two hours, taken up with a little dilute (1 : 1) hydrochloric acid, diluted to 50 C.C. after five minutes digestion on the water-bath and filtered. The silica separated this time is washed with cold 1 per cent. hydrochloric acid. In silicate fusions, no more sodium carbonate than is necessary should be used, as silica is appreciably soluble in sodium chloride solutions, but many silicates cannot be decomposed with less than 4 parts of sodium carbonate.G. C. J. Accurate End-Point in Volumetric Estimation of Sulphur in Steel. H. Zschiegner. (J. Ind. and Eng. Chem., 1916, 8, 324.)-For the titration of hydrogen sulphide with iodine and starch a special titration vessel is used, consisting of a white porcelain beaker, 7 crns. in diameter and 16 crns.deep, to the bottom of which is cemented a thin black disc of hard rubber or celluloid about 2 cms. in diameter. A substitute for this vessel may be made by painting the black spot on the outside of a glass beaker, and then painting the whole of the outside of the beaker with white enamel. The estimation is carried out as usual until the titration is commenced, then the black spot is carefully observed, and, as the solution darkens, the iodine is added more slowly with agitation.Finally, a point is reached when one drop of iodine renders the spot invisible on allowing the solution to come to rest. This is taken as the end-point, and is said to be more easy to work to than the first appearance of a distinct blue colour. The iodine solution is, of course, standardised in a similar manner, using a steel of known sulphur content, an equal volume of hydrochloric acid to decompose the cadmium sulphide, and an equal volume of starch solution.The author prefers a starch solution, prepared by intro- ducing 12 grms. of wheat starch, made into a cream with water, into 2,000 C.C. of boiling water containing 3 grms. of sodium hydroxide.G. C. J. Estimation of Hardness of Natural Waters, and the Use of Methyl Red as an Indieator. S. A. Kay and S. H. Newlands. (J. Xoc. Chem. Ind., 1916,INORGANIC ANALYSIS 221 35, 445-447.)-The methods described are mainly modifications of those described by Hehner (ANALYST, 1883, 8, 77) ; the use of methyl red as indicator in the titrations is recommended. Temporary hardness is most conveniently estimated by treating 100 C.C.of the water with a slight excess of & hydrochloric acid, boilingthe mixture for about one minute, and titrating the excess of acid with +% barium hydroxide solution. For the estimation of permanent hardness, the water is evaporated with the addition of $F potassium carbonate solution, the residue then extracted with 90 per cent.alcohol, and the alcoholic solution titrated, after filtration, with & hydro- chloric acid. The reasons for the adoption of this procedure are that calcium and magnesium carbonates are practically insoluble in 90 per cent. alcohol, that the soluble double magnesium-potassium salt which is liable to form is decomposed by alcohol, with reprecipitation of magnesium carbonate, and that potassium carbonate is more soluble in 90 per cent.alcohol than is sodium carbonate. Total hardness is found from the sum of the temporary hardness and the permanent hardness, or by titrating the precipitate of calcium and magnesium carbonates formed in the estimation of the permanent hardness. An alternative method consists in neutralising the water as in the estimation of temporary hardness, then evaporating it with the addition of potassium carbonate, and estimating the amount of the latter required to precipitate the calcium and magnesium salts.The electrical con- ductivity of the water, as determined by the Dionic Water Tester (see ANALYST, 1912, 37,538), affords evidence as to the degree of hardness of a water. Every 20 units of conductivity correspond, approximately, with 1 degree of hardness.w. P. s. Estimation of Calcium and Magnesium in Natural Waters. S. A. Kay and S. H. Newlands. ( J . SOC. Chem. Ind., 1916, 35, 447-449.)-The method depends essentially on the solubility of magnesium carbonate in ammonium carbonate solution, whilst calcium carbonate is insoluble. The total hardness of the water is, first of all, estimated by one of the methods described in the preceding abstract, One hundred C.C.of the water are then evaporated to dryness in a platinum basin, with the addition of 10 C.C. of ammonium carbonate solution (2 grms. of the salt and 1 C.C. of concentrated ammonia per 100 c.c.). The dry residue is treated with 15c.c. of the ammonium carbonate solution, the mixture heated for two minutes on a water- bath, the solution then decanted through a small filter, and the insoluble portion washed twice with the ammonium carbonate solution.It is next washed with 90 per cent. alcohol until free from alkali, then dissolved in a measured quantity of & hydro- chloric acid, and the excess of the latter is titrated with & barium hydroxide solution, using methyl red as the indicator.The amount of calcium, expressed in degrees of hardness, is then calculated, and the difference between the total hard- ness and that due to calcium gives the hardness due to magnesium. A simple calcu- w. P. s. lation gives mgrms. of calcium and magnesium per litre of water. Comparison of the Permanganate Methods for the Estimation of Required Oxygen [in Water Analysis].J. H. Sachs. ( J . Ind. and Eng. Chem., 1916, 8, 404-406.)-A variety of conditions have been prescribed by various222 ABSTRACTS OF CHEMICAL PAPERS authors for performing the permanganate test for oxygen absorption in water analysis. The variable factors are : Acidity or alkalinity of the mixture, quantity of permanganate, temperature and duration of the action.Comparative estimations were made with a number of organic substances under these varied conditions. I n the alkaline methods there was a tendency to the formation of green potassium manganat a, or else an unusually large precipitation of manganese dioxide, which reacts catalytically to reduce further quantities of the permanganate. The alkaline methods were therefore discarded. As regards the temperature, it was found that a t 85" and 100" C. much larger quantities of permanganate were reduced than at lower temperatures, especially in acid solution. At 37" C. the quantity consumed was somewhat larger than at the ordinary temperature. Comparative experiments. made at 85" C. for one hour and at 37' C. for three hours showed higher results for the higher temperature, but at 85" C. the consumption of oxygen increased largely with the quantity of permanganate present, whereas at 37" C. variation in the quantity of permanganate had but little effect on the consumption value. The author, therefore, expresses a preference for Thresh's method in acid solution at 37" C. The presence of chlorides interferes, but these can be readily removed by treatment with silver oxide. J. F. B. Estimation of Dissolved Oxygen in Water. J. Miller. (J. Xoc. Chern. Ind., 1916,35,457.)-Since phenosafranine, the indicator recommended for use in a method described previously by the author (ANALYST, 1914, 39, 234), is somewhat difficult to obtain, methylene biue may be used in its place. The two indicators give the same rssult. w. P. s.