January 15, 1934
INDUSTRIAL AND ENGINEERING CHEMISTRY
Gold and tellurium may be determined in the presence of each other by first precipitating the gold and tellurium together and then separating the gold by reduction with nitrous acid. ACKNOWLEDQMENT The subject for this investigation was suggested in 1917 by the late Victor Lenher. The experimental work was started in Professor Lenher’s laboratory a t the University of Wisconsin in 1917, was continued with Professor Lenher’s permission by D. C. Knowles in 1926-27, and part of this paper is based upon the thesis presented by Mr. Knowles in partial fulfilment of requirements for the degree of Bachelor
45
of Science a t the Polytechnic Institute of Brooklyn, in June, 1927. The authors are pleased to acknowledge their indebtedness to Douglas Fronmuller, who has made several series of determinations. Publication in the present form is with the permission of Samuel Lenher.
LITERATURE CITED (1) Fisher, Pogo. Ann., 77, 480 (1849). (2) Hutchins, Univ. of Wis., BUZZ. 119 (1905). (3) Lenher and Homberger, J. Am. Chem. Soc., 30, 387 (1908).
RECEIVED ApriI 7, 1933. Contribution from the Laboratory of Inorganic Chemistry of the University of Wisconsin and Contribution 20 from the Department of Chemistry of the Polytechnic Institute of Brooklyn.
Determination of Borate Ion in Ores of Borax H. L. PAYNE, 552 Figueroa St., Los Angeles, Calif.
I
N 1932, Scott (8) published the results of research on the determination of boric acid in the ores and mill products so abundantly produced from the mines in Kern County, Calif., which at the present time are the only source of borax ore in the United States. He concluded with a formal method based upon the results of these researches. The death of Doctor Scott prevented further research along the original lines, but as one of the participants in these investigations, the present author has continued the work and now offers a completed analytical method differing from Scott’s directions in some essential particulars. In general it is the method originally proposed by Thomson in 1893 and presented in revised form in 1908 by Wherry (6). In so far as it relates to titrating boric acid it has been officially accepted by the Association of Official Agricultural Chemists and by the American Ceramic Society, and for routine work has been approved by Hillebrand and Lundell, of the U. S. Bureau of Standards. With more or less incomplete detail it has been copied by standard books on chemical analysis, and is merely elaborated here. It is offered as a procedure applicable to all borate ores, to partially refined products, and to completely crystallined borax. By a more drastic decomposition treatment it may be applied to the analysis of the borosilicate minerals and glasses (4). ores and mill products-need Borate minerals-borate not be ground finer than 30 mesh. Finer grinding of many of the calcined products results in a loss of values by dusting. All milled material is hygroscopic, cannot be dried to any permanently constant weight, and samples should be kept in tightly stoppered containers. Close checks by different laboratories are practically impossible. The method is based upon the precipitation of the bases of the iron-alumina group from an acid borate solution and titration of the liberated boric acid with sodium hydroxide. Wherry (6) used calcium carbonate in neutral or alkaline solution as the precipitant, reprecipitating two or three times and washing carefully. Sullivan and Taylor ( 4 ) did the same, but separation under these alkaline conditions was never quite satisfactory and Wherry says that the method is not to be recommended. Wherry also tried an acid separation but did not follow it up. Scott’s researches included acid precipitation using phenolphthalein indicator as control, demonstrating in this way that no aluminate passed into the filtrate to interfere with the boric titration, but he did not include this plan in his final method. The author follows Scott’s sodium hydroxide-acid separation but with a modified procedure which involves only one precipitation and avoids all error by occlusion.
PROCEDURE To 2.5 grams of the suitably prepared sample in a 100-ml. tall plain covered beaker add 5 or 10 ml. of water to wet the mineral and then 10 ml., pipet measure, of 6 N hydrochloric acid (1 to 1). Digest for at least 30 minutes on top of the water bath, with occasional agitation but without removing the cover. All borate ores are thus readily decomposed without loss of boron or the use of an excessive amount of acid (2). Remove from the bath, wash down the cover and beaker with plenty of cold water, and filter into a 250-ml. graduated flask, washing the insoluble matter with hot water only. (Ignite and weigh the insoluble mineral matter if that item is desired.) It may be tested qualitatively for boron; but the author has never found any, and such insoluble borate would not be considered available boric acid. To the filtrate, which will amount to about 200 ml., add carefully, preferably from a buret, a strong sodium hydroxide. solution until the liquid is nearly neutralized, using methyl red as indicator. This condition is necessary for the removal of carbon dioxide by boiling. A precipitate of iron and aluminum hydroxides will appear just before neutrality is reached and if the liquid is still acid to methyl red a proper condition obtains for the removal of carbonate by boiling. Adapt an air-condenser tube to the flask, such as the inner tube of a 24-inch (60-cm.) Liebig condenser, and heat to gentle boiling for at least 15 minutes, allowing the steam to rise not higher than half way up the condenser. All carbon dioxide will be removed without loss of boron and the ironalumina hydroxide precipitate will occlude no boric acid. Under these conditions methyl red in the presence of free boric acid shows an acid reaction a t pH of 6, and both iron and aluminum hydroxides are precipitated. Borates do not form until a pH of about 11; hence the precipitates carry no boron. Remove the flask from the heat, wash down the condenser with cold water, and cool the flask as rapidly as possible without undue exposure to the air. As soon as cool, add at once sufficient calcium hydroxide solution to fill the flask to the mark. This will make the solution alkaline without introducing any soluble carbonate and any residual iron or alumina will precipitate. Mix thoroughly and filter on a dry paper and into a dry beaker. Do not wash, but as quickly as possible pipet off two aliquots for the final titration.
TITRATION Titrations should be conducted as quickly as possible after the aliquots are taken, using a standard hydroxide practicslly-
Vol. 6, No. 1
ANALYTICAL EDITION
46
free from carbonate. To each aliquot in a flask add methyl red and a drop of hydrochloric acid just sufficient to turn it red. Then from the titration buret add hydroxide until the methyl red just turns a full lemon-yellow. Do not stop a t the so-called neutral tint, The solution will then be alkaline to hydrochloric acid, a t a pH of about 10, and will be practically neutral to phenolphthalein and boric acid a t about the same pH. The differences are well within the color indicator errors of titration. (Allen and Zies ( 1 ) working on borosilicate glass found a difference of 0.3 mg. Bz03 between the end point with p-nitrophenol and the beginning with phenolphthalein. The author has not been able to confirm this error using methyl red, although theoretically some alkali should be consumed.) To this neutralized solution add an excess of mannite and titrate to a permanent red with phenolphthalein, a color which is not bleached by further addition of mannite. If sufficient mannite is added a t the beginning, the solution will be nearly colorless until the final red. I n absence of excess mannite a pink color will maintain during much of
the titration and a true end point will be reached only when a single drop of the alkali gives a red that does not fade. Using 0.2 N alkali, titrations will habitually check to 0.1 cc. equivalent to 0.7 mg. of BzOs or less than 0.1 per cent on 1 gram titrated. Check analyses in this laboratory recently ran as follows: 38.80, 33.07, 29.67, 29.55, 44.61, 44.10, 41.01, 42.19, 42.07, 38.81, 33.22, 29.77, 29.62, 44.69, 44.25, 41.15, 42.19, 42.14, routine work. Two drops of 0.2 N alkali will always indicate an end point and the quantitative reaction between sodium hydroxide and boric acid in the presence of a polyhydric alcohol has never been brought into dispute, LITERATURECITED (1) Allen, E. T., and Zies, E. G., J . Am. Ceram. Xoc., 1,739-86 (1918). (2) Dodd, Analyst, 54,715 (1929). (3) Scott, W.W., IND.EXG.CHEM.,Anal. Ed., 4,306 (1932). (4) Sullivan and Taylor, IND.ENQ.CEIEM.,6, 897 (1914). (5) Wherry, J . Am. Chem. Soc., 30, 1684 (1908). RECEIVED July 24, 1933.
Determination of Iron Adaptation of the Mercaptoacetic Acid Colorimetric Method to Milk and Blood GLADYSLEAVELL AND N. R. ELLIS Nutrition Laboratory, Animal Husbandry Division, Bureau of Animal Industry, U, S. Department of Agriculture, Washington, D. C. HE mercaptoacetic acid colorimetric method of Lyons
T
(9) for the determination of small quantities of iron offers advantages over the thiocyanate colorimetric method of Elvehjem (3) as adapted from Kennedy ( 6 ) , in that smaller samples may be used, thus making wet-ashing practicable and avoiding the losses and contaminat,ions encountered in the dry-ashing procedure, and in that the color produced is more stable. The authors’ experiments confirmed observations of Andreasch ( I ) , Ginsburg and Bondzynski (4), and Claesson (2) on the formation of an intensely blue or purple compound by the reaction of mercaptoacetic acid (thioglycollic acid) with ferric chloride in ammoniacal solution, the tendency of the color to fade, and the ready restoration of the color when the solution is shaken with air. Preliminary experiments with a standard iron solution (0.005 mg. of iron per cc.) showed that for quantities of iron between 0.0025 and 0.025 mg., the colored solutions could be conveniently compared in tubes of uniform diameter, such as Nessler tubes, and that it was not necessary to dilute to volume, as the addition of distilled water to the solutions in the tubes caused no noticeable change in color intensities of the columns when observed through their entire depths. For quantities of iron from 0.015 to 0.1 mg., dilution to volume and the use of the colorimeter are recommended. Larger quantities of iron necessitate dilution and the production of color in convenient aliquots. Comparison of color-development by the use of 2 drops of undiluted mercaptoacetic acid, 1 cc. of a 1 to 15 aqueous solution of the acid, 1 cc. of a 1 t o 15 alcoholic solution, and 1 cc. of a 1 to 15 aqueous solution rendered alkaline with ammonium hydroxide (the reagent described in the next section) favored the last as it combines the advantages of nearly
eliminating the unpleasant odor of the mercaptoacetic acid with good control of the amount of reagent added. Experiments on wet-ashing of milk with sulfuric acid, sulfuric and nitric acids (8, IO), and sulfuric and perchloric acids (5, 7) were performed. The ashing with sulfuric acid alone was too slow, and with nitric acid resulted in muddiness and other irregularities in the colors developed, whereas the colors with perchloric acid were like those developed when standard iron solutions were used. METHOD REAGENTS.Iron-free concentrated sulfuric acid, 60 per cent perchloric acid, concentrated ammonium hydroxide, standard iron solutions, and the mercaptoacetic acid reagent are necessary. The standard iron solutions are prepared by dissolving 1 gram of pure iron wire in dilute sulfuric acid and oxidizing with concentrated nitric acid. The oxides of nitrogen and excess nitric acid are expelled by boiling. The solution is diluted to 1 liter, with further dilution as needed. I n all dilutions it is advisable to add extra sulfuric acid to prevent hydrolysis of the dissolved iron salts to insoluble basic compounds. The mercaptoacetic acid reagent is prepared by the addition of 4 cc. of mercaptoacetic acid to a solution of 8 cc. of concentrated ammonium hydroxide in 50 cc. of water. DETERMINATION OF IRON IN MILK. Samples* of 5 cc. of milk (or 0.5 gram dried milk) are digested with 3 cc. of con1 With samples of this size, there wa8 rarely any turbidity from precipitation of alkaline earth phosphates in the final ammoniaoal solution. Filtration of occasional turbid samples through small iron-free filter paper gave clear solutions with colom matohing those of the duplicates vhioh were not turbid. Filtration of very turbid solutions from larger samples pave entirely satisfactory results.