ISSN:0003-2654    

DOI:10.1039/AN95277FX021

出版商:RSC    

年代:1952

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95277BX023

出版商:RSC    

年代:1952

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95277FP057

出版商:RSC    

年代:1952

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95277BP065

出版商:RSC    

年代:1952

数据来源: RSC

摘要:

JUNE, 1952 THE ANALYST Vol. 77, No. 91 5 PROCEEDINGS OF THE SOCIETY OF PUBLIC ANALYSTS AND OTHER ANALYTICAL CHEMISTS AN Ordinary Meeting of the Society was held at 7 p.m. on Wednesday, April 2nd, 1952, in the Meeting Room of the Chemical Society, Burlington House, London, W.l. The Chair was taken by the President, Dr. J. R. Nicholls, C.B.E., F.R.I.C. The following papers were presented and discussed: “The Determination of Traces of Arsenic in Germanium Dioxide and Tetrachloride,” by S. T. Payne ; “Inorganic Chromato. graphy on Cellulose. Part IX: The Determination of Thorium by Chromatography on Alumina and Cellulose Adsorbents and the Simultaneous Determination of Thorium and Uranium in Minerals and Ores,” by A. F. Williams, BSc., F.R.I.C. ; “Inorganic Chromato- graphy on Cellulose.Part X : The Spectrographic Determination of Micro Quantities of Thorium Separated by Chromatography from Minerals and Ores,” by G. W. J , Kingsbury and R. B. F. Temple, D.Phil. NEW MEMBERS David James Brown, A.R.I.C. ; Harold Lovell Haigh, A.R.I.C. ; James John Hayes, A.R.I.C. ; John Robert Hudson, B.Sc. (Lond.), A.R.I.C. ; Robert John Magee, M.Sc. (Q.U.B.), Ph.D. (Edin.) ; Richard Leslie Overin, B.Sc. (Lond.), F.R.I.C. ; Philip Robins; Harold Bernard Salt, M.Sc. (Bkm.), F.R.I.C. ; Derek Howard Shrimpton, B.A. (Cantab.) ; Lionel Hewgill Smith, F.R.I.C. ; Stephen Charles James Christopher Snell, B.Sc. (Lond.), A.R.I.C. DEATH John Francis Hutchins Gilbard WE regret to record the death of PHYSICAL METHODS GROUP THE Thirty-sixth Ordinary Meeting of the Group was held at 6.30 p.m.on Tuesday, April 8th, 1952, in the Meeting Room of the Chemical Society, Burlington House, London, W.l. The meeting was organised by the Polarographic Discussion Panel, and the Chair was taken by the Chairman of the Panel, Dr. A. J. Lindsey. The following papers were presented and discussed : “The Measurement of Diffusion Current with Special Reference to the Tinsley Pen-recording Polarograph,” by W. Furness, B.Sc., F.R.1 .C. ; “Hypochlorites and the Dropping-Mercury Electrode,’’ by E. N. Jenkins, MSc., A.R.I.C. ; “The Polarographic Determination of Tellurium in Selenium,” by G. H. Osborn, F.R.I.C., A.M.Inst.M.M., and J. G. C. Cobb. BIOLOGICAL METHODS GROUP A MEETING of the Group was held at 6.30 p.m. on Friday, March 14th, 1952, in the Meeting Room of the Chemical Society, Burlington House, London, W.l. Dr. H. 0. J. Collier was in the Chair. The following papers were presented and discussed : “The Measurement of Prothrombin in the Control of Anticoagulant Therapy with Tromexan,” by Dr. Rosemary Biggs; “The Assay of Heparin, Protamine and Russell Viper Venom,” by V. J , Birkinshaw and K. L. Smith. 277

ISSN:0003-2654    

DOI:10.1039/AN9527700277

出版商:RSC    

年代:1952

数据来源: RSC

摘要:

278 P.4YNE : THE DETERMINATION OF TRACES OF ARSENIC [Vol. 77 The Determination of Traces of Arsenic in Germanium Dioxide and Tetrachloride BY S. T. PAYNE (Yresenied nt the meeting of the Society on‘ Wednesday, April Znd, 1952) The determination of arsenic by a direct Gutzeit test on materials con- taining germanium is shown to be unsatisfactory. Reproducible results have been obtained by extracting the arsenic with a chloroform solution of diethyl- ammonium diethyldithiocarbamate after forming an oxalate complex with the germanlum, and completing the analysis by a modified Gutzeit test in which the hydrogen is generated electrolytically from a zinc-plated cathode. With a 5-g sample of germanium dioxide or 5 ml of germanium tetrachloride it is possible to determine as little as 0.10 p.p.m.of arsenic in the germanium material. INCREASING use is being found for germanium metal in the electrical industry for the manu- facture of crystal rectifiers in high-frequency circuits. For this purpose the metal must be extremely pure and, in particular, almost completely free from arsenic, the presence of which in more than the most minute traces seriously reduces the “turnover voltage’’ of the rectifiers. All the raw material from which germanium is prepared contains considerable amounts of arsenic ; for example, the mineral germanite, which is the richest source of germanium, generally contains more than 5 per cent. of arsenic, and most flue dusts, which are the largest source of germanium in this country, contain more arsenic than germanium.The process of preparing germanium from any of these raw materials involves, at some stage, distillation of germanium tetrachloride from diluted hydrochloric acid (1 + 1) ; in this process a con- siderable proportion of the arsenic content of the raw material accompanies the germanium in the distillate as arsenic trichloride. In subsequent rectifications of the crude germanium tetrachloride distillate, the arsenic is eliminated partly by taking advantage of the lower boiling-point of germanium tetrachloride compared with that of arsenic trichloride, and partly by taking advantage of the greater reactivity of arsenic trichloride compared with that of germanium tetrachloride. For use in rectifiers, the amount of arsenic present in the germanium tetrachloride must be well below 0.5 p.p.m.and the determination of arsenic in such a low concentration presents considerable difficulties, as germanium interferes in all the standard methods of determining arsenic when the latter is present at low concentrations. Probably the most useful test for arsenic in these circumstances is the Gutzeit test. If this test is carried out in the usual way by generating hydrogen by dissolving zinc in the solution containing germanium and arsenic in sulphuric acid, some of the germanium is reduced to germanous chloride and some to the metal, whilst a small part is apparently evolved as germane. Tests with radioactive arsenic have shown that the germanium precipitate produced in the solution retains an appreciable amount of arsenic. Although the germane evolved does not colour the mercuric chloride paper used for detecting the arsine, it does interfere if an attempt is made to intensify the arsenic stain by treatment with silver nitrate, as the silver nitrate blackens the paper more or less uniformly if the evolved gases contain germane.Attempts to overcome the interference of germanium in the Gutzeit test by forming a complex of it with oxalic acid failed, as oxalic acid does not prevent production ofsome germane under the conditions of the test. Obviously, therefore, some preliminary separation must be carried out. EXPERIMEXTAL Strafford, Wyatt and Kershawl found that, in hydrochloric acid solutions, tervalent arsenic forms a complex with diethylammonium diethyldithiocarbamate that can be extracted with chloroform.Germanium, however, also forms a complex with this reagent that is soluble in chloroform; the author has found that formation of this complex can be preventedJune, 19521 IN GERMANIUM DIOXIDE AND TETRACHLORIDE 279 by first converting the germanium into ammonium germanioxalate, and that from solutions containing 20 per cent. by volume of concentrated hydrochloric acid and an excess of oxalic acid, minute traces of arsenic can readily be separated from several grams of germanium, provided that the carbamate reagent is prepared from carefully purified diethylammonium diethyldithiocarbamate. When this reagent is prepared by the method described by Straff ord et al. it generally contains impurities that form complexes with germanium, which pass into the chloroform extract and interfere with the subsequent arsenic determination.By re- crystallising the reagent several times from ether, a pure white product can be obtained that forms a colourless solution in chloroform and does not react with germanioxalate. The chloroform extract of the arsenic is heated with a mixture of nitric, sulphuric and perchloric acids to destroy the organic matter, and then evaporated until fumes of sulphuric acid appear and the other two acids are completely expelled. The arsenic is then in the quinquivalent condition, and Strafford et al. finished the determination of the arsenic colori- metrically after converting it to arsenomolybdate and by reducing this to molybdenum blue. Attempts to use this method for the determination of arsenic in germanium gave erratic results and attention was therefore turned to the Gutzeit test and particularly to the modification described by in which the arsine is collected on a piece of string impregnated with mercuric chloride and the stain is developed with ammoniacal silver nitrate solution.How generated the hydrogen required for the evolution of the arsine by means of a special zinc alloy containing small amounts of tin, lead and iron. The preparation of this alloy with a sufficiently low arsenic content to give a reasonably small blank was found to be extremely difficult. At this stage of the investigation details of an apparatus for the generation of arsine electrolytically were ~ b t a i n e d . ~ An apparatus of this type was constructed and found to require some modifications before satisfactory results could be obtained.In the original apparatus a platinum wire cathode is coated with zinc by electrolysis of a dilute zinc sulphate solution, and after allowing sufficient time for any arsenic in the electrolyte to be evolved as arsine, the solution of the sample is added and the electrolysis continued until no further arsine is generated. In the apparatus first used by the author a vertical cathode was fitted, but considerable “treeing” of the zinc deposit occurred; frequently small portions of the cathode were not completely plated with zinc, and on these bare portions elemental arsenic was deposited so that the results obtained were erratic. When the apparatus was modified by incorporating a horizontal platinum disc to completely cover the bottom of the electrolysis tube, satisfactory zinc deposits were obtained without any “treeing” and a smooth and complete evolution of arsine occurred when the sample solution was added.A combination of this apparatus and How’s technique for collecting and measuring the arsine has given satisfactory results over a period of several months. Occasionally, batches of zinc sulphate used for the plating operation gave widely different results and, although it was possible to standardise each batch, it was naturally desirable to obtain comparable results with the different batches. I t was noticed that when a smooth, pale grey deposit of zinc was obtained, the stain lengths for a given amount of arsenic were shorter than when the deposit was highly crystalline and dark in appearance.Very careful purification of the zinc sulphate gave no better results, but it was finally discovered that addition of ammonium sulphate to the electrolyte produced a highly crystalline deposit, which resulted, as expected, in excellent sensitivity and reproducibility. It is, however, essential that the zinc and ammonium sulphates used in the plating bath should be free from foreign metals. After every determination the zinc must be removed from the platinum by dissolving it in dilute hydrochloric acid and the platinum cleaned with hydrochloric acid saturated with bromine to remove the small amount of platinum black that remains on the cathode owing, presumably, to a zinc - platinum alloy layer being formed during the plating.If this platinum black is not removed the subsequent zinc plate will be unsatisfactory. The surface of the platinum is more severely etched when cleaned with dilute aqua regia, and in time this etched surface ceases to be plated satisfactorily with zinc. To ensure that the platinum is working satisfactorily, frequent checks of the action of the cell are made with solutions containing known amounts of arsenic. Even with the bromine - hydrochloric acid treatment the platinum surface eventually becomes etched, and when this occurs the old platinum disc should be replaced with a new one, but with careful work, a disc should operate satisfactorily for many determinations.280 PAYNE: THE DETERMINATION OF TRACES OF ARSENIC [Vol. 77 , J-outlet (Fig.2) Fig. 1. Diagram of apparatusJune, 19521 IN GERMANIUM DIOXIDE AND TETRACHLORIDE 281 Blank determinations, made with the reagents only, usually give a stain that represents between 0.05 and 0-1 p.p.m. of arsenic (based on a 5-g sample of germanium dioxide). This stain can be kept to a minimum by careful selection and preparation of the reagents. For a given batch of reagents (including the zinc plating solution and strings), the blank remains almost constant, but it is advisable to make a blank determination with each batch of analyses. APPARATUS- The complete apparatus is shown in Fig. 1. The cathode, F, is a cap of 1 inch internal diameter and Q inch deep, made of 0.01-inch platinum foil. This cap fits on the end of the generator, E, and is held in position by a rubber tubing bound with wire.The lower end of the rubber tubing is sealed with a rubber stopper carrying a short length of glass tubing. A spacer, G, placed between the platinum cap and the rubber stopper, consists of a ring of glass tubing 1 inch in diameter and Q inch long. Electrical contact with the platinum cap is made by soft-soldering the wire to its centre and by passing this wire through the glass tube in the stopper to the current supply. During the analysis the temperature of the generator must be maintained at 5” C above that of the absorption unit, which is kept at room temperature; consequently the generator is immersed in the water-bath, J, which is cooled by passing a slow current of tap water through it. In this manner sufficient of the heat produced by the electrolysis in the cell is removed to keep the electrolyte at the desired temperature.To avoid variation in mains water pressure the cooling water to the water- bath is supplied by the constant-head device shown in Fig. 2. The lower end of the constant- head tube is placed about 12 inches above the top of the water-bath and, to avoid local overheating of the cell, the bath is stirred continuously by bubbling a stream of air through it. The absorption apparatus shown in Fig. 1 is similar to that recommended by How; A is the lead acetate scrubber; B, the water trap; and C, the capillary tube containing the sensitised string, D. As exposure to daylight causes some fading of the stain during the time necessary for complete evolution of the arsine, the capillary, C, is covered with a black-coated tube.Preparation of strings-The string used for the test is Strutt’s Knitting Cotton No. 4. This is wound lightly, but as closely as possible, on to a glass tube of about 1Q inches in diameter, immersed in a 0.25 per cent. solution of mercuric chloride in redistilled industrial spirit for about 16 to 24 hours in the dark and then wound tightly on to a wooden or metal frame. As the string is uncoiled from the glass tube and withdrawn from the solution it should be passed lightly through a large tuft of cotton wool to remove surplus solution. The frame on to which the string is wound is first covered with “Sellotape” to avoid con- tamination. When the string has dried, it is cut into 8-inch lengths and stored in a large stoppered test tube covered with black paper.Portions of the string that have been in contact with the frame are discarded and in all the above operations the string should be kept away from bright daylight and handled with the fingers as little as possible. As the strings become less sensitive with age they should be checked at least once a fortnight if prepared in large batches, in order to avoid the possibility of error. The string is inserted into the wide end of the capillary by applying suction to the other end. It is then pulled through until the clean-cut tail end is about 1 cm inside the capillary from where it joins the wider tube. The excess is cut off leaving about half an inch protruding to allow the string to be withdrawn when the test is completed.Scrubber wit-This consists of 12 inches of 1-inch roller bandage rolled to form a plug and moistened with 16 drops of saturated lead acetate solution containing just sufficient acetic acid to keep it clear. Not more than 16 drops must be used or the tube will be blocked. Water tva@-Two drops of water are added just before starting the test. Zinc $Zating of the cathode-The cathode is coated with zinc by the electrolysis of 40 ml of zinc sulphate plating solution (see under Reagents) at a potential of 110 v. for 30 minutes and then rinsed with distilled water. The cathode is cleaned after each determination by dissolving’the zinc in hydrochloric acid and then treated for half to one minute with hydro- chloric acid saturated with bromine. It is advisable to rinse the cell with a little ammonium hydroxide to remove any traces of acid that may be trapped between the platinum and the glass, and then with distilled water.The cathode can then be re-coated with zinc when required.283 PAYNE: THE DETERMINATION OF TRACES OF ARSENIC [Vol. 77 METHOD REAGENTS- All materials should be of recognised analytical purity. Oxalic acid-Recrystallise the commercial crystals once from 12 per cent. v/v hydrochloric Sulphuric acid-Sp.gr. 1.84. Ammonium hydroxide solution--Sp.gr. 0.88. Hydrochloric acid--Sp.gr. 1.18. Potassium iodide solution-Dissolve 20g of potassium iodide in water and dilute to 100 ml. Extract with carbamate reagent. Potassium metabisulphite solution-Dissolve 5 g of potass'iurn metabisulphite in water and dilute to 100ml.acid and twice from water. Fig. 2. Constant head device Diethylammonium diethyldithiocarbamate-Recrystallise the commercial substance three times from ether or until a completely colourless solution is obtained in chloroform. On the day of use, dissolve 0.25 g of the reagent in 100 ml of chloroform. Perchloric acid mixture-Mix 25 ml of nitric acid with 25 ml of 60 per cent. AnalaR perchloric acid and slowly add 50 ml of sulphuric acid. Stir throughout the addition of the sulphuric acid. AnalaR perchloric acid has a much lower arsenic content than other brands, which were all unsatisfactory. Sodium hydroxide-Pellets. Zinc sulphate plating solution-Dissolve good quality zinc grain in a deficiency of 20 per cent. sulphuric acid and boil for several hours.Filter off the excess of zinc and add sufficient hydrogen peroxide solution to oxidise all the iron present. Add an excess of zinc oxide, followed by a moderate quantity of filter-paper pulp, and keep the solution on the hot-plate until the precipitate settles. Filter from the insoluble matter and evaporate the solution to incipient crystallisation. Add about half its volume of distilled water, cool, and add an equal volume of redistilled industrial spirit while stirring vigorously. Cool thoroughly, filter the crystals on a Buchner funnel, and dry them by drawing air through the funnel. The dried crystals are ZnS0,.7H20. Prepare the ammonium sulphate by neutralising sulphuric acid with ammonium hydroxide solution (both suitably diluted) and recrystallise twice from water.June, 19521 I N GERMANIUM DIOXIDE AND TETRACHLORIDE 283 Make the plating solution by dissolving 95 g of ZnS0,.7H20 and 5 g of (NH,),SO, in water, add 5ml of sulphuric acid and dilute to 1 litre. Silver nitrate soldion-Dissolve 2 g of silver nitrate in 100 ml of 10 per cent.ammonium hydroxide solution. Standard arsenic solution-Dissolve 1.3208 g of arsenious oxide in 25 ml of 20 per cent. sodium hydroxide solution, saturate the solution with carbon dioxide, and dilute to 1 litre. Dilute 5 ml of this solution to 15m1, acidify with 1 drop of sulphuric acid, add 0.5g of potassium metabisulphite, heat to 80" to 85" C for 30 minutes and then boil for 2 to 3 minutes or until no trace of sulphur dioxide can be detected. This solution contains 10 p g of arsenic per ml.Immediately before use, dilute 10 ml of the solution to 100 ml, so that 1 ml contains 1 pg of arsenic. Finally dilute to 500 ml. PROCEDURE- Weigh 17 g of oxalic acid into a 500-ml conical wide-mouthed flask, cover with 200 ml of redistilled water and add 7 ml of ammonium hydroxide solution. Shake until the odour of ammonia disappears and then add 5 g of the germanium dioxide to be tested. Boil the mixture for 2 minutes, rinse the sides of the flask, and keep the solution hot for 30 minutes or until it is perfectly clear; crush any small lumps of oxide with a glass rod to facilitate their solution. For the analysis of germanium tetrachloride, shake 5 ml with the ammonium oxalate solution prepared as described above until no oily drops of undecomposed chloride are visible; the solution must not be heated above 50" C.Adjust the temperature of the solution to 50" C and add 50 ml of hydrochloric acid, followed by 2 ml of potassium iodide solution to reduce any quinquivalent arsenic. Maintain at 50" C for 15 minutes, add 0-5 ml of potassium metabisulphite solution and transfer without cooling to a 500-ml separating funnel. Introduce about 2 ml of chloroform into the funnel, shake well, release the pressure, and add 5 ml of the carbamate reagent. Shake the mixture for 45 seconds, allow it to settle, and remove the lower layer into a 50-rnl conical flask. Repeat the extraction with another 5 ml of the carbamate reagent and finally with 5 ml of chloroform. Add 4 ml of the perchloric acid mixture to the flask containing the chloroform extracts, evaporate on the hot-plate until fumes of perchloric acid appear, and then heat strongly until fumes of sulphuric acid appear; heat the sides of the flask so that no remaining perchloric acid will condense and run back into the solution.Cool, dilute to 40 ml, add 1 drop of phenol- phthalein solution and sodium hydroxide pellets until just alkaline (about 32 pellets). Acidify again by adding concentrated sulphuric acid dropwise and add a further 1 ml of the same acid. Add 06ml of potassium metabisulphite solution whilst the sample solution is still hot and leave it overnight. Carefully boil out the remaining sulphur dioxide by rotating the flask over a bare flame. Then cool and transfer the solution to the prepared Gutzeit apparatus, place in position in the water-bath, and electrolyse for 18 hours at 1.0 amp. (65 v.approximately). Develop the stain by immersing the string for a few seconds in the silver nitrate solution and, after rinsing it, measure the length of the stain and read the arsenic content from a calibration graph produced by plotting stain lengths against the arsenic content of standards that have been previously determined in the recommended manner. The standards are prepared by adding the arsenic solution, in quantities containing up to 5 pg of arsenic, to the oxalate solution and by continuing the analysis as described above. Each new batch of strings or plating solution must be re-standardised. It is important that the greatest care be exercised to avoid contamination in the tests as the amounts of arsenic involved are very small.Strict attention must also be paid to temperature control to obtain consistent results. A blank determination must be made with all the reagents used in the analysis. RESULTS In order to verify the completeness of the extraction a series of tests was carried out to compare the stains produced by direct additions of arsenic to the generator cell with those obtained by adding arsenic to the original oxalate solution, without germanium, and by subsequent extraction. The results of this series are shown in graphical form in Fig. 3. Curve A represents the additions made directly to the cell and curve B those made to the original solution and extracted. It will be seen that, under the conditions laid down, a recovery of about 85 per cent.can be expected and this is regarded as being satisfactory.284 PAYNE: THE DETERMINATION OF TRACES OF ARSENIC Wol. 77 A higher recovery value can be attained by increasing the number of extractions, but no practical advantage is gained by so doing. Further tests were then carried out to check the extraction procedure in the presence of germanium, and for this eight 5-ml portions were taken from a batch of germanium tetra- chloride of low arsenic content. To seven of these were added various amounts of arsenic, the quantities being unknown to the operator. All eight portions were then analysed with the results shown in Table I. Fig. 3. Results of recovery experiments. Curve A , additions of arsenic made directly to cell; Curve B, additions of arsenic made to original solution and extracted It d l be seen that the recovery under these conditions is satisfactory in view of the minute mounts of arsenic involved. TABLE I RECOVERY OF ARSENIC IN THE PRESENCE OF GERMANIUM Arsenic added, p.p.m.nil 0.10 0.11 0.16 0.20 0.25 0.29 0.32 Arsenic recovered, p.p.m. nil 0.09 0.13 0.16 0.19 0.24 0.25 0-30 Error, p.p.m. - 0.01 + 0.02 - 0.01 -0.01 - 0.04 - 0-02 - - The original germanium tetrachloride used for this series contained 0.05 parts per million of arsenic and this has been allowed for in Table I. A sample of germanium dioxide, found to contain less than 0-05 p.p.rn. of arsenic by the above method, was analysed at A.E.R.E., Harwell, by an improved version of the radio- activation method of Smales and Brown4 and was reported to contain 0.035 p.p.m.of arsenic. The author wishes to thank Mr. R. H. A. Crawley of British Thompson Houston CO. Ltd. for help in testing the method, Mr. €3. J. Cluley of G.E.C. Research Laboratories for many helpful suggestions and for advice concerning the development of the generator cell, and the Directors of Johnson, Matthey & CO. Ltd., for permission to publish this paper.June, 19521 IN GERMANIUM DIOXIDE AND TETRACHLORIDE REFERENCES 1. 2. 3. 4. RESEARCH LABORATORIES Strafford, N., Wyatt, P . F., and Kershaw, F. G., Analyst, 1945, 70, 232. How, A. E., Ind. Eng. Chem., Anal. Ed., 1938, 10, 226. Private communication from the G.P.O. Research Station, Dollis Hill. Smales, A. A., and Brown, L. O., Chem. and Ind., 1960, No.23, 441. JOHNSON, MATTHEY & Co. LTD. WEMBLEY, MIDDLESEX 285 DISCUSSION MR. A. A. SMALES asked what would be the method of dissolving metallic germanium and what was the normal blank level of the method, in practice. MR. PAYNE replied that metallic germanium could be dissolved in the normal oxalic acid mixture, as described in the method, by the addition of 100-volume hydrogen peroxide solution as necessary. The rate of dissolution was slow, but 1 g of the crushed material could be dissolved in about one day. Blanks were normally of the order of 0.05 parts of arsenic per million, but might be as high as 0.10 p.p.m. For a given batch of reagents, however, the blank was quite constant and, therefore, not inconvenient provided it did not get inordinately high.MR. T. MCLACHLAN expressed surprise that the author had found it necessary to allow the reduction with potassium metabisulphite ,to continue for so long, oiz. , overnight ; this was generally considered unnecessary. He also asked if Mr. Payne had compared the stains given on string with those obtained by passing the gas through mercuric chloride paper as described in the British Pharmacopoeia. MR. PAYNE was emphatic that it was essential to allow the reduction to proceed overnight or low arsenic figures would result. No other interpretation could be given to this than that the reduction was incomplete, as quinquivalent arsenic appeared not to be reduced in the Gutzeit cell. He had not had practical experience with the B.P. method and was unable to make a comparison.However, in his view, greater accuracy would result from measuring the length of a stain than from visually comparing the intensity of a stain with a standard. DR. K. A. WILLIAMS said that he had found that if either thread or a sufficiently long paper strip was used none of the arsine missed it and that quite satisfactory and reproducible stains were produced. The efficacy of the two methods was about the same. MR. G. G. ELKINGTON enquired if the author had tried mercuric bromide in place of the chloride, as he himself had found that it gave much darker undeveloped stains than did the chloride. MR. PAYNE said that he had not tried mercuric bromide and, although i t might well give darker undevel- oped stains, the main advantage of the method described lay in the sharp demarcation of the stain when developed.MR. SMALES said that when the development method (with silver nitrate) was used, the separation of germanium must be absolutely complete, but this might not be so with thenormal Gutzeit technique with only the mercuric bromide, as there would then be no coloration from any germane formed. He thought that a few micrograms might, therefore, be tolerated, and this would make the extraction procedure much easier. MR. PAYNE agreed that evolution of germane did not cause interference with the undeveloped stain, but germanium present in the Gutzeit cell was largely reduced to metal and this caused adsorption of arsenic and, in consequence, low results. If germanium was present in insufficient quantity to colour the string when developed, it might be assumed that there was insufficient present to cause low results.To develop the string was useful, therefore, to check the com- pleteness of the extraction. DR. J. HASLAM asked what contribution to the blank value of 0.05 parts of arsenic per million was made by the reagents and, in particular, by the sodium hydroxide. MR. PAYNE replied that, as previously stated, most of the blank was due to the hydrochloric acid. Only about 10 per cent. was due to the sodium hydroxide, although different makes vary considerably in this respect. MR. R. C. CHIRNSIDE said that he thought the real value of this paper would be in drawing attention to the possibilities of determining arsenic, in the absence of germanium, in smaller quantities than had prev- iously been thought possible or satisfactory.Its value to those in the electrical industry who were con- cerned with the use of germanium was now less than it had been and its main application was as a control method by the germanium suppliers. In the electrical industry they were now more concerned with impurities of an entirely different order; he had a sample in which the arsenic content might very well lie between 0.01 and 0.001 p.p.m. and he thought it would be agreed that no conventional chemical method was at all likely to deal with these quantities. In the early days of the work on germanium he had said to his physicist colleagues that they were looking for the impossible in expecting to get a material having arsenic It was due, in large measure, to the hydrochloric acid.It must, therefore, be removed as completely as possible. The brand used by the author was particularly good.286 PAYNE: THE DETERMINATION OF TRACES OF ARSENIC [Vol. 77 in amounts so much smaller than those legally allowed in many foods and, even if it were possible to get it, it would seem impossible to maintain it at such a standard of purity. He would admit that this view had been incorrect and that the order of impurity, which they obtained and which they maintained, was in the region he had mentioned. The methods that now had to be used were essentially electrical in nature, e.g., a measurement of resistivity gave a fairly good idea of the nature and amount of impurities present in the germanium. The other method, which he tended to regard as an absolute one, was that developed by Mr.Smales a t Harwell based on radiochemical methods. MR. N. L. ALLPORT said, with reference to the determination of traces of arsenic, that theanalyst concerned with examining pharmaceuticals must conform to the procedure described in the British Pharma- copoeia and that this involved comparing the intensity of the yellowish brown stains produced on discs or mercuric chloride paper. The method described by Mr. Payne, in which the criterion was the length of stains developed by silver nitrate and not the intensity of the colour, seemed to be a most interesting and useful development of the time-honoured Gutzeit test, and if subsequent work confirmed the accuracy of the new method, i t might be hoped that the present official specification would be modified in a future edition of the Pharmacopoeia. MR.SMALES said that, with regard to the activation method that had been developed a t Harwell, the sensitivity was now certainly better than 0.01 p.p.m., but he was still waiting for samples of germanium containing amounts of arsenic small enough to apply a really stringent test to the method. The best sample of oxide he had so far met contained 0.005 parts of arsenic per million. DR. J. H. HAMENCE said that he thought a warning should be given a t this stage in connection with the use of a paper strip or string in the modified Gutzeit test. He had used the Sanger and Black methods in his laboratory for many years and had found that in this process i t was absolutely essential to ensure a uniform rate of hydrogen generation.The length of the stain that was produced in the test, in his exper- ience, depended not only on the amount of arsenic present in the solution, but also on its rate of liberation, and unless a uniform liberation of hydrogen was ensured by adding the acid slowly, as in the Marsh test, the mere length of stain after development was of little value, since the more rapid the evolution of the hydrogen, the longer became the stain. In his view, the excellence of the stains that had been produced by the author was in no small measure due to the electrolytical equipment that had been devised, which quite clearly ensured a uniform evolution of hydrogen and, consequently, of arsine. MR. PAYNE agreed that the rate of evolution of the hydrogen had a bearing on the result.With the electrolytic method this could be very accurately controlled. MR. H. J. CLULEY added that How had investigated numerous variables in the Gutzeit procedure, even to such details as controlling independently the temperature of the generating solution and that of the arsine absorption unit. Valuable as all this work had been, the weakest point in How’s procedure was the method of generating the arsine, which involved the use of a complex zinc alloy that was most difficult to prepare. Mr. Payne’s use of an electrolytic method of generating the arsine gave much closer control over this part of the process and constituted a valuable improvement to How’s procedure. In this respect much credit was also due to Dr.Speight and Mr. Carasso of the Post Office Research Station for the development of the electrolytic cell used, with some modification, by Mr. Payne. There was some reason to believe that the arsenic content of distilled water might be a limiting factor in thepurity of germanium dioxide prepared by hydrolysis of the tetrachloride. Would the author say what part of the blank could be attributed to the distilled water. MR. PAYNE said that i t was not possible to state exactly what proportion of the blank was due to the distilled water, but i t must have been of a very low order. I t was, of course, compensated for in the determination. Generally the arsenic figure with a given sample of germanium dioxide was slightly lower than in the chloride from which i t had been prepared.There was evidence, therefore, that part of the arsenic remained in the liquor when hydrolysis took place. The distilled water used by the author originated from condensed steam as supplied to the laboratory. This was always redistilled in an all-glass Pyrex still. THE PRESIDENT thanked the author for presenting such a stimulating paper. Many chemists who had used the strip method had experienced difficulty in measuring the length of the stain. He said that he was impressed by the very sharp demarcation given when the stain was developed. DR. E. A. SPEIGHT and MR. J. I. CARASSO, in a written contribution, said that they were privileged to receive details of Mr. Payne’s method of extraction prior to its publication and had used it successfully, in conjunction with their own (Post Office Research Branch) design of electrolytic Gutzeit cell, for the estima- tion of arsenic in germanium dioxide.Occasionally, however, abnormal results were obtained, the causes of which were still obscure, and they felt that there was scope for further study of the process. The most disturbing abnormality was the apparently fortuitous occurrence on the sensitive strip of stains that differed markedly in length or intensity, or both, from those expected. They had a considerable amount of evidence suggesting that these phenomena were critically related to the particular batches of In the author’s method the reagent used in greatest quantity was distilled water.June, 19521 IN GERMANIUM DIOXIDE AND TETRACHLORIDE 287 reagents used, especially the ammonium hydroxide, but they had been unable so far to discover the nature of the essential factors determining whether or not a particular batch of a reagent would be “suitable” for the determination. Usually, but not invariably, the use of an unsuitable batch of a reagent led to abnormal results both in the “blank” experiment and in the parallel determination. Occasionally, even, anomalous. results had been obtained with reagents that had previously appeared to be “suitable.” Another abnormality was that some types of arsenic compound appeared to be complexed by the oxalate solution and to escape detection. Frequently the total arsenic found was much less than the amount of arsenic known to be present in the ammonium hydroxide alone. In spite of these difficulties they usually obtained a satisfactory recovery of a few tenths of one part per million of arsenic added to germanium dioxide when working on half the scale used by Mr. Payne. Although they hoped shortly to publish a full description of their design of cell and to report critically on the electrolytic Gutzeit method generally, they wished to clarify one point mentioned by Mr. Payne, The cathode of their cell had the form of a single-turn flat spiral and not a helix. This was chosen to avoid the formation of zinc “trees” during the plating stage, but they were aware that satisfactory results had been obtained with a similar design of cell in which the formation of “trees” was deliberately emmu-aged.

ISSN:0003-2654    

DOI:10.1039/AN9527700278

出版商:RSC    

年代:1952

数据来源: RSC

摘要:

June, 19521 IN GERMANIUM DIOXIDE AND TETRACHLORIDE 287 The Separation and Analysis of Metallic Carbides from Steel BY R. PEMBERTON (Preseated at the meeting of the Society on Wednesday, February 6th, 1952) A technique for isolating the carbide phase from solutions of highly alloyed steels is described and methods for determining the elemental com- position of the separated residue are detailed. The steel specimen is anodically decomposed in a citric acid - potassium iodide electrolyte to leave a residue of carbides, which is collected and analysed by semi-micro analytical methods. Procedures are given for the determination of carbon, manganese, chromium, vanadium, molybdenum, tungsten, titanium, niobium and iron in the residue. THE resistance t o “creep” of highly alloyed steels at high temperatures is believed to be largely dependent on the composition and distribution of carbides that have been precipitated out of solid solution.With the advent of the gas turbine the need for efficient creep-resisting steels assumed major importance, and further information was required about the nature of the carbides present in these steels in order to improve their performance. Microscopic examination will reveal the dimensions and distribution of the carbide particles ; if there are several types of carbide, some attempt can be made to distinguish between them and to estimate approximately the amount of each. But more exact distinction was necessary; the quantity of carbide is usually too small to permit of its identification by X-ray crystallo- graphy and recourse was had t o chemical analysis.First, some method of isolating the “out of solution’’ phase or “carbide residue” from the remainder of the steel was sought. An attempt to dissolve the steel directly in acid was not successful owing, presumably, to acid attack on the carbides. Attempts to separate residues with solvents such as copper potassium chloride and ammonium iodide - iodine - ammonium citrate were abortive because the time required for complete solution of sufficient steel millings (10 g) was too great. By dissolving the metal phase anodically in an electrolyte that attacked steels but little in the absence of an applied potential, the yield of residue was quantitative. The technique and design of this, the most satisfactory method, were simple. After various solutions had been tried, an electrolyte containing citric acid and potassium iodide in dilute hydrochloric acid solution was found to give effective and selective decomposition of highly alloyed steels, with complete solution of the metal portion and no evidence of solvent action on the separated residual carbides.288 PEMBERTON : THE SEPARATION AND ANALYSIS OF METHOD APPARATUS- (Vol.77 A diagram of the apparatus assembly is shown in Fig. 1, and a view of the apparatus in Fig. 2. The cathode consists of a piece of platinum foil welded to platinum wire and enclosed in a porous pot to prevent contamination of the carbide residue by cathodically deposited iron, which tends to flake off. The steel specimen itself forms the anode and can conveniently take the form of a smoothly ground cylindrical bar, 6 inches long and 8 inch in diameter.The bar should have n Fig. 1. Apparatus for electrolytic separation of carbides from steel. C, solid carbon dioxide: B, bubbler for carbon dioxide: E, electrolyte: W, water cooling spiral: K, platinum cathode: P, porous pot; G, split cover-glass ; R, variable resistance; A, ammeter; S, test specimen a smooth emery finish and be free from scale and other surface contamination. The circuit includes a 110-volt D.C. supply, an ammeter and avariable resistance for regulating the current to the cell. Considerable heat is generated in the cathode compartment during electrolysis and a water-cooled glass spiral is required to maintain the temperature at 25" C, the optimum.A slow stream of carbon dioxide is bubbled through the electrolyte to maintain an inert atmosphere immediately above the surface of the electrolyte, where there may be some danger of atmospheric decomposition of iron carbide on the surface of the anode. The carbon- dioxide bubbler produces also a steady agitation of the solution, which helps to maintain homogeneity in the electrolyte throughout the electrolysis. REAGENTS- of hydrochloric acid, sp.gr. 1.16, and make the solution up to 1 litre with water. the solution through a sintered-glass filter of No. 3 porosity. EZectroZyte solutiort-Dissolve 450 g of citric acid and 300 g of potassium iodide in 60 ml Filter Residue wash solution-Dilute 5ml of the electrolyte solution to 1 litre with water. PROCEDURE- Set up the apparatus as described, weigh the anode, suspend it in the electrolyte and electrolyse until sufficient of the anode metal has been dissolved to give a carbide residue large enough for the analysis.Experience has shown that about 1 g is dissolved per ampere- hour and analyses can be based on the residue from about l o g of steel. When sufficient steel has been dissolved, switch off the current, rinse the spiral, bubbler and cathode pot and withdraw them from the solution.June, 19521 METALLIC CARBIDES FROM STEEL 289 The steel anode bar then carries a lightly adherent surface residue of carbides, which can be completely detached from the bar by brushing gently with a nylon toothbrush; collect. the residue in the electrolyte. Dry and re-weigh the clean steel bar; the loss in weight represents the amount of steel corresponding to the carbide residue obtained. Filter the electrolyte and washings through a dried and weighed sintered-glass filter of No.4 porosity, wash three or four times with the wash solution, then three times with water and finally once with industrial spirit and once with ether. Draw air through the filter for 10 to 15 minutes, and place the filter in an air oven at about 150" F for a few minutes to ensure complete dryness. Allow the filter to cool and weigh it. Calculate the carbide residue as a percentage w/w of the steel dissolved from the anode. The percentage of total mixed carbides isolated from the steel having been determined, semi-micro analytical methods are applied to the various elements present and from the results an assessment is made of the probable constitution of the residue.For this purpose sufficient of the residue to meet the analytical requirements is detached from the filter and reserved in a glass sample tube. A GENERAL SCHEME OF CARBIDE RESIDUE ANALYSIS- The mean yield of residue from the highly alloyed rustless and heat-resistant steels is of the order of 1 per cent. From a 10-g sample of steel the amount of residue available is some- what less than 0.1 g. An outline of the methods of analysis of these residues as normally practised, follows. Carbon-Carbon is determined, in 5 to 10-mg samples, with a specially designed micro- combustion apparatus (Fig. 3). Oxygen from a 10-cubic-foot cylinder is passed through a l-litre vessel, which serves as a reservoir and maintains a positiveflow without further manual regulation during the period of combustion. The oxygen then passes through a preheater tube (1200" C) packed with platinised asbestos, to decompose traces of hydrocarbon ; the resultant carbon dioxide is absorbed in the soda-asbestos tube.Magnesium perchlorate extracts traces of water and the dry oxygen passes via the side-arm into the combustion tube, also maintained at 1200" C. The end of the combustion tube is closed by means of a ground-glass joint, which can be removed to allow the small combustion boat, containing the sample, to be pushed into the tube. The exit gases resulting from the combustion pass through a solution of chromic acid, to remove sulphurous gases, and then through magnesium perchlorate; finally the carbon dioxide is collected in the small absorption tube packed with soda asbestos.The absorption tube is weighed on a micro balance before and after combustion, and from the increase in weight of the tube, after suitable correction for a blank run, the carbon content is computed. Decomposition of the residue-For the remainder of the analysis it is most emmnient to prepare a stock solution from a known weight of residue, from which suitable aliquots can be taken for the determination of the various elements. To effect decomposition, the residue is digested in a nitric - sulphuric acid mixture and evaporated until dense fumes d sulphur trioxide appear; further nitric acid is added dropwise until charring ceases and decomposition is complete.The sulphated concentrate is extracted in aqueous medinm and, if basic sulphates are present, the solution is filtered and the residue recovered by h i o n with potassium bisulphate, the extract being returned to the main solution. A suitable concentration for the stock solution is 1 mg of carbide residue per millilitre. Manganese-Manganese is determined by the potassium periodate reaction in sulphuric - phosphoric acid solution. The resulting colour intensity is measured on an absorptiometer with Calorex and Ilford 604 spectrum green filters. The working range is 0 to 0126 mg of manganese in a final volume of 10ml. Chromium-Chromium is determined volumetrically with standard 842 N ferrous sulphate and dichromate solutions, with barium diphenylamine sulphonate as indicator.The oxidation is merely a scaled-down replica of the normal ammonium persulphate method for the determination of chromium. An amount of residue calculated to provide the equivalent of 2 to 3 mg of chromium is used. If vanadium is present in the residue, the chromium titre is inclusive of this element and correction must be made based on a separate determination of vanadium (see p. 290)- If the available material is insufficient for this volumetric method with persulphate, the determination can be made absorptiometrically with the diphenylcarbazide reaction afterFig. 2 . Apparatus for separating carbides from steel Fig. 3. Micro-combustion apparatus for the determination of carbon290 PEMBERTON: THE SEPARATION AND -4NALYSIS OF [Vol.77 oxidation with persulphate and selective reduction of permanganic acid with nitrite in the presence of urea. The much higher sensitivity of this colour reaction permits the use of samples of residue weighing about 1.0 mg. The range is 0 to 0.2mg of chromium in a final volume of 100ml. Vanadiztm-Vanadium is determined volumetrically with standard 0.02 N solutions of ferrous sulphate and dichromate, with barium diyhenylamine sulphonate as indicator. The element is oxidised t o the quinquivalent condition with permanganate solution in the cold, and excess of permanganate is reduced with nitrite in the presence of sulphamic acid. The titre value of the vanadium determination must be deducted from the chromium titre. Molybdenum and tungsten-Molybdenum and tungsten are determined in turn on a 4-mg sample of residue by selective application of their reactions with toluene-3 : 4-dithiol.The molybdenum - dithiol complex is developed first at a carefully regulated acidity so that there is no simultaneous formation of the corresponding tungsten complex. The tungsten complex is developed in a solution of suitably modified acidity after extracting the molybdenum complex. Hence, molybdenum is extracted from a dilute hydrochloric acid solution, sp.gr. 1-06, with 1 per cent. dithiol solution in amyl acetate and made up to standard volume with amyl acetate, and the colour is measured with an absorptiometer with Calorex and Ilford 601 spectrum violet filters. A working range equivalent to 0 to 6 per cent. of molybendum (0 to 0.24 mg of molyb- denum) on a 4-mg residue sample can be obtained with a suitable choice of 1.0, 0.5 or 0-25 cm micro-cells.The acid layer from the molybdenum extraction is evaporated until fumes of sulphur trioxide appear, to remove residual dithiol and amyl acetate. The sulphated concentrate is then treated with hydrochloric acid, sp.gr. 1.16, and the tungsten - dithiol complex is developed by a further treatment with the reagent in the concentrated acid extract. This extract is made up to 10-0 ml with amyl acetate and its colour intensity measured on an absorptiometer with Calorex and Ilford 607 spectrum orange filters. The range of tungsten concentration covered is 0 to 4 per cent. of tungsten with a 0-5 cm cell. The 3-cm cell calibration is normally only required for tungsten determination on steels.Iron-Iron is determined by an adaptation of the thiocyanate reaction. The thio- cyanate complex is developed in a solution containing 3 per cent. of free sulphuric acid. The colour intensity is measured on an absorptionmeter with Calorex and Ilford 604 spectrum green filters. Titanium-The hydrogen peroxide reaction is not of sufficient sensitivity to permit of adequate measurement of the small amounts of titanium present in the type of residues under review. The procedure used *is based on the reaction with 1 :e-dihydroxybenzene- 3: 5-disulphonic acid. This reagent gives an intense yellow colour with minute quantities of titanium and, in solutions of controlled acidity, e.g., at approximately pH 4.5, the colour shows a linear relationship with the titanium concentration.Chromium and iron interfere with the reaction. Chromium is removed as chromyl chloride by means of controlled additions of hydrochloric acid to a fuming perchloric acid concent rate. The deep purple iron complex simultaneously formed can be selectively destroyed with 0.1 g of sodium dithionite (N+,S,O,) to leave the clear yellow of the titanium - dihydroxy- benzene complex. The colour is measured on the absorptiometer with Calorex and Ilford 601 spectrum violet filters. The range is 0 to 0.4 mg of iron in a final volume of 10.0 ml. , The range is 0 to 0.03 mg of titanium in a final volume of 10.0 ml. Niobium-For the determination of niobium a large initial sample (0.1 g) is required; this is better taken separately from the separated residue as a direct test rather than as an aliquot from the stock solution. Niobium together with titanium is precipitated with cupferron from a reduced (ferrous) solution so that only small amounts of iron are simul- taneously precipitated. Copper, molybdenum and so on are removed as sulphides from an extract of the ignited residue and iron is removed as ferrous sulphide from ammoniacal tartrate solution.A further precipitation with cupferron yields all the niobium along with the titanium ; the titanium is determined absorptiometrically by a hydrogen peroxide test on the final weighed residue. After correction the difference is taken as niobium pentoxide.June, 19521 METALLIC CARBIDES FROM STEEL 291 OTHER DETERMINATIONS- Other determinations occasionally made include : silicon, which is determined by the usual hydrofluoric acid procedure on the insolubles obtained from the sulphated extract of the initial bulk decomposition; nickel, which is sometimes determined in an aliquot of the stock solution by means of the oxidised dimethylglyoxime colour reaction ; nitrogen, which is determined by micro steam-distillation in a modified Kjeldhal apparatus on a sepa.rate sample of residue; and oxygen, which is determined by the vacuum fusion procedure1 on similar lines to the determination of the element in steel.RESULTS Five typical examples of residue compositions are shown in Table I. The residues vary widely, depending on the composition and heat-treatment of the steel. The results shown are included to convey merely a general impression of the kind of figures from which it is possible to make a fair estimate of the probable constitution of the residues.TABLE I TYPICAL ANALYSES OF RESIDUES FROM FIVE STEELS Sample 1 Carbon, yo .. .. 0.37 Silicon, Yo .. .. 2.67 Manganese, % .. .. 0.63 Nickel, Yo * . .. 9.39 Chromium, yo . . , . 45.94 Molybdenum, yo . . .. Niobium, yo . . .. Vanadium, yo . . .. Titanium, yo . . .. Nitrogen, Yo . . .. Oxygen, yo . . .. - - - - Iron, yo . . .. . . 39.48 - - Sample 2 7.54 0.89 3.00 24.93 5-80 3-28 46-00 0-24 5-62 I - - Sample 3 6.20 4.20 0.03 0.50 4.10 63-95 1.20 4.76 2.45 - - I Sample 4 4.76 6-14 0.03 1.20 13.40 40.60 0.80 17-32 1-53 - - - Sample 5 7.11 5-93 0.18 0.37 67.50 1-51 trace 2-47 - L - - Total . . .... 98-48 97.30 87-39t 85.78t 85.07t Residue in steel, yo . . 6.00* 4-50 1-56 2.20 1-22 * This residue contained a large proportion of sigma phase, an iron - chromium - nickel intermetallic t Oxygen was not determined in these three residues; this probably accounts for the low totals. compound. The author is indebted to Mr. B. Bagshawe for his advice and interest in this work, and to Dr. C. Sykes, F.R.S., for permission to publish the paper. REFERENCE 1. Newell, W. C., “The Eighth Report of the Heterogeneity of Steel Ingots Committee,” The Iron and Steel Institute, London, Special Report No. 25, 1939, p. 97. THE BROWN - FIRTH RESEARCH LABORATORIES PRINCESS STREET SHEFFIELD, 4 DISCUSSION MR. E. C. MILLS asked what range of anodic current densities was used in view of possible polarisation effects, and if it was necessary to adjust the current density according to the type of steel and its heat treatment. MR.PEMBERTON replied that he had used current densities over a range of 0.3 to 1.0 ampere per square inch without any polarisation. Polarisation effects were often marked in the experimental solutions initially tried, but were eliminated in the citric acid - potassium iodide electrolyte. The efficiency of carbide isolation was not affected by the nature or composition of the steel, nor was it influenced by the heat-treatment condition. MR. D. G. HIGGS enquired about the stability of the blank weighings and the precautions taken to guard against electrostatic effects. He asked if the author would explain the nature of the combustion tube and the mode of its attachment to the glass-stoppered tube.292 PEMBERTON [Vol.77 He remarked that although the steels used were quite complex they were not as complex as he had expected them to be. He asked if the author could give an idea of the effect of high concentrations of molybdenum and say if the molybdenum complex went into solution easily. Could the technique be applied to the determination of carbides in molybdenum metals ? Blank values were always steady and reproducible between 0.06 and 0.10 mg. The combustion tube was of the micro pattern, 14 inches long by # inch internal diameter, tapered a t one end and with side arm delivery. The glass tube, carrying the male ground-glass joint, was fastened to the combustion tube by means of a rubber sleeve, which was waxed on to the tube.The method had been applied to high molybdenum steels, i.e., containing up to 10 per cent. of molybde- num, in the solution-treated condition. There was complete extraction of molybdenum in the electrolyte and no contamination of the residue with molybdenum owing to incomplete insolubility in the citrate medium. It would appear likely that the use of such an electrolyte could be extended to molybdenum- base and high-molybdenum materials. MR. C. WHALLEY asked if there was any tendency for the carbides to explode or to oxidise during the determination of carbon, since i t appeared that the authors took no precautions to prevent their exposure to the atmosphere during drying and storing. No mention had been made of the incidence of copper except at one point in the methods, where it was stated to have been removed.He thought that residual copper present in the steel might come down with the carbides and be found in the residues. He said that the author had not referred to the determination of silica, to the form in which silica would be found in the residues or to the method of attack used to remove it before proceeding with the determination of the other elements. MR. PEMBERTON replied that no difficulty had been experienced in controlling the combustion of carbide residues, as there had been little or no tendency for spontaneous or explosive combustion. It should be noted that sample weights were only of about 5 to 10 milligrams. There was no evidence to support the contention that partial decomposition of the carbides might occur during drying or storing. Copper determinations were not normally made on carbide residues, but the element was likely to separate and contaminate the residue in the form of cuprous iodide. This could be removed by washing the residue with ammonium hydroxide. This treatment would also extract any elemental sponge copper deposited along with the carbides. Silica was found in the residue mainly in the form of silicates present in the steel as slaggy inclusions. It was determined on the bulk sample of separated residue by sulphuric acid dehydration and the conventional treatment with hydrofluoric acid. The main solution without the silica was then used for the determination of the various carbide elements in the usual way. MR. PEMBERTON replied that he had met no difficulty attributable to electrostatic effects.

ISSN:0003-2654    

DOI:10.1039/AN9527700287

出版商:RSC    

年代:1952

数据来源: RSC

摘要:

June, 19521 RYAN AND WILLIAMS 293 Inorganic Chromatography on Cellulose Part VIII The Use of a “Compound” Column of Alumina and Cellulose for the Determination of Uranium in Minerals and Ores Containing Arsenic and Molybdenum BY W. RYAN AND A. F. WILLIAMS A new technique in inorganic chromatography, with alumina and cellulose adsorbents in the same extraction column, is described for the separation of uranium from minerals and ores. The purpose of the alumina is to retain arsenic and molybdenum, which are not readily retained by cellulose alone when ether containing 5 per cent. v/v of concentrated nitric acid, sp.gr. 1.42, is used as the extraction solvent. A METHOD for the chromatographic separation and determination of uranium and its applica- tion to minerals and ores has been described by Burstall and We1ls.l Ethyl ether containing 5 per cent. v/v of concentrated nitric acid, sp.gr.1.42, is used as the solvent and cellulose pulp as the adsorbent. Although the method is widely applicable, difficulties are encountered with materials containing molybdenum and arsenic, both of which may be extracted to varying degrees and so interfere with the final estimation of uranium. It was found that extraction of molybdenum and arsenic by ether solvent containing 5 per cent. v/v of nitric acid, sp.gr. 1.42, could be largely prevented by using a solvent containing only 1 per cent. of nitric acid. At such low acidity, however, it was found difficult to extract uranium completely from ores rich in material other than uranium unless a very large volume of solvent was used.For this reason a search was made for an adsorbent that would retain both molybdenum and arsenic. It was found that a column of activated alumina (chromatographic grade) was very efficient in retaining these two metal ions, so that for ores containing these two elements, uranium separations can be made in a compound column consisting of alumina placed above cellulose. The cellulose retains small quantities of impurities such as iron, aluminium and vanadium, which are not so efficiently retained by alumina. The presence of small amounts of molybdenum and arsenic are objectionable in the standard volumetric procedure for the determination of uranium in the eluate. On the other hand, traces of these elements extracted with the uranium from a low-grade material do not interfere in the spectrophotometric method based on the development of colour in the presence of hydrogen peroxide and excess of alkali., EXPERIMENTAL EXTRACTION AND DETERMINATION OF URANIUM IN SAMPLES CONTAINING MOLYBDENUM BY Molybdenum interferes with the volumetric estimation of uranium and causes low results by its catalytic effect in promoting oxidation of uranium. As high-grade samples of uranium- bearing ores, such as pitchblende ore, often contain molybdenum, it is important to ensure the absence of this metal from a uranium solution before a volumetric determination, as this is based on the oxidation of UO, to UO, by an oxidising agent of known equivalence.I t was disclosed in early work on pitchblende that molybdenum trioxide, usually present in high-grade pitchblende ore in small amounts of about 0-25 per cent., was partly extracted by ether containing nitric acid in a column containing cellulose as used for the determination of uranium in low-grade siliceous ores.1 Moreover, the extraction of molybdenum was found to be conditioned to a certain extent by several factors, such as the concentration of uranium in the ore and the effect of light.The movement of molybdenum with uranyl nitrate has been confirmed by experiments in which samples of pitchblende in nitric acid were extracted USE OF A CELLULOSE ADSORBENT-294 RYAN AND WILLIAMS : INORGANIC CHROMATOGRAPHY [Vol. 77 with an ether - nitric acid mixture in a column of cellulose, 10 cm long, sheltered from direct sunlight.Some results on pitchblende sample :No. 1, containing 0.255 per cent. of MOO,, extracted with 125ml of solvent, were as follows- Weight of ore taken, g . . .. .. 0-2158 0.2126 MOO, in ore taken, pg . . .. . . 550 540 MOO, found in eluate, pg. . .. .. 154 182 Spectrographic analyses of a number of samples treated under broadly similar conditions indicated quantities of molybdenum up to 200 pg in the extracted fractions. Determinations of molybdenum were carried out by a colorimetric method involving a preliminary separation with a-benzoin oxime followed by solvent extraction with fi-butyl acetate of the coloured complex of molybdenum with potassium thiocyanate, thioglycollic acid being used as reducing agent. THE RETENTION OF MOLYBDENUM BY ACTIVATED ALUMINA AND ITS APPLICATION TO THE I n the course of the study on the adsorption of metals in ether - nitric acid medium by various solid materials, it was found that activated alumina had good retentive properties for molybdenum without adsorbing uranyl nitrate.Activated alumina, however, proved less satisfactory for retaining other elements such as vanadium ; accordingly, a combination of activated alumina and activated cellulose in the form of a column has been used for purification of uranium from molybdenum and other elements. I n experiments with two samples containing 0.255 per cent. and 2.55 per cent. of MOO,, the molybdenum found in the eluate was 0 and 8.3 pg, respectively. The procedure, described in detail on p. 295, in which an adsorbent containing alumina and cellulose is used, has been tested with a number of pitchblende ore samples and pitchblendes containing added amounts of impurities.The results shown in Table I indicate good reproducibility and satisfactory agreement with determinations of uranium made by standard chemical procedures. Determinations made without the use of an alumina trap are shown in the third column of the table for comparison. ANALYSIS OF URANIUM I N PITCHBLENDE ORE- TABLE I COMPARISON OF Sample 5* 6* 7* DETERMINATIONS OF URANIUM BY PROPOSED AND STANDARD METHODS U 0 found by aluAiia - cellulose adsorbent method, Y O 71.35, 71.61, 71-51, 71-50, 71.40 48.49, 48.36 80.98, 81.03 56.82, 56.52, 56.63, 56.70, 56.74, 56.80, 56.70, 56.61, 56.68 59.15, 59.03 59.47, 59.47, 59-50 40-20, 40.34 U,O, found by cellulose adsorbent method, 69.45, 69-50, 70.55, 70.08, 68.0 48.49, 48.56 80.64, 79.91 Y O 54.00 U,O, found by standard chemical procedure, 71.20, 71.30 Y O 48.28, 40.32 80.94, 80.84 56.53, 56.59 59-01] * These samples were pitchblendes containing the following added impurities: 2 per cent.each of molybdenum and vanadium, and 0.5 per cent. each of bismuth, arsenic, boron, lead, cadmium and mercury. Two further points should be noted in connection with the analysis of pitchblende. The first concerns the volume of ether - nitric acid solvent for completely removing uranyl nitrate in the alumina - cellulose adsorbent method. It has been found that 125 ml is sufficient, but 150 ml is recommended as providing a reasonable excess. With sample No. 1 (Table I), 71.47 per cent.of U,o, was extracted by 100ml of solvent; a further 25ml of solvent extracted 0.14 per cent. of U,O,, whilst the next 25 ml did not remove any more. Under these conditions extraction of molybdenum is considerably increased,3 even in the presence of an alumina trap, owing to reduction of the molybdenum to a mobile blue form of oxide. An experiment showed that 7 pg of MOO, were extracted from an adsorption column shielded from sunlight, The second point is the effect of direct sunlight on the extraction.June, 19521 ON CELLULOSE. PART VIII 295 whereas 45pg of MOO, were removed from the column when it was exposed to sunlight. The analysis of uranium by this method should therefore be carried out in diffused light. THE DETERMINATION OF URANIUM IN MIXTURES CONTAINING ARSENIC BY USE OF ALUMINA Arsenic is known to interfere with the volumetric procedure for determining uranium by causing high results.It has been shown that 1 g of pitchblende ore with 0.2 g of As06 added yielded most of the arsenic with the uranium on extraction under staqdard conditions with ether containing 5 per cent. v/v of nitric acid, sp.gr. 1.42, and a column of cellulose adsorbent material. It has now been found that by using similar conditions, but with a column containing alumina in addition to cellulose, according to the method described for the determination of uranium in ores containing molybdenum, the extracted uranium con- tained less than 5 p.p.m. of arsenic. A method is described below for the extraction and TABLE I1 TO RETAIN THE ARSENIC- DETERMINATION OF URANIUM IN PRESENCE OF ADDED ARSENIC PENTOXIDE pitchblende taken, As,O, added, U80, present, U,O, found, Amount of g g g g 0.2392 0.3243 0.1707 0.1702 0-2398 0.1025 0.1711 0.1706 0.3842 0.3776 0.1860 0.1862 TABLE I11 DETERMINATION OF URANIUM IN SAMPLES OF CONCENTRATES CONTAINING ARSENIC (10 TO 20 PER CENT.OF AS205) U,O, by standard U,O, by alumina - cellulose Sample chemical method, chromatography, % % 1 11.55 11.59, 11.65 2 12.15 12.17, 12-20, 12.36 determination of uranium from its mixtures with arsenic pentoxide and the results are shown in Tables I1 and 111. The extracted uranyl nitrate was free from arsenic. The method described for ores containing arsenic is essentially the same as that for those containing molybdenum, except that a larger volume of ether solvent containing 5 per cent.v/v of nitric acid, sp.gr. 1-42, must be used to ensure complete extraction of uranium. Hence, if an ore is known to contain appreciable quantities of arsenic and molybdenum, the amount of solvent used for extraction of uranium would be the same as that used if arsenic alone were present. METHOD FOR SAMPLES CONTAINING MOLYBDENUM OR ARSENIC OR BOTH Solvent-Add 5ml of concentrated nitric acid, sp.gr. 1.42, to each 95ml of ether. PREPARATION OF ALUMINA - CELLULOSE COLUMN- The adsorption tube for the preparation of the column is a glass tube about 25 cm long and 1-8 cm in diameter. The upper end is flared to a diameter of about 8 cm to form a funnel that permits easy transfer of the sample. The lower end terminates in a short length of narrow tubing and is closed by a short length of polyvinyl chloride tubing carrying a screw clip.The inside surface of the glass tube is coated with a silicone in the manner described by Burstall and We1ls.l Weigh 5 or 6 g of cellulose pulp* into a stoppered flask and cover it with ether - nitric acid solvent. Pour the suspension into the glass tube, agitate gently and then gently press down the cellulose to form a homogeneous column. Wash the column with about 100 ml of ether - nitric acid solvent and finally allow the level of the solvent to fall to the top of the cellulose. Next pour about 15 g of activated aluminium oxidet on top of the cellulose, * Whatman’s Standard Grade cellulose powder is suitable. Type H, Chromatographic Alumina, 200 mesh.Supplied by Peter Spence Ltd.296 RYAN AND WILLIAMS [Vol. 77 pour on 30 ml of either - nitric acid solvent and vigorously agitate the alumina with a glass plunger. Allow the packing to settle. Allow the level of the ether to fall to the surface of the alumina and the column is ready for use. PREPARATION OF SAMPLE SOLUTION FROM MINERAL OR ORE- Weigh into a platinum dish sufficient of the sample to give 100 to 150mg of U,O,, which is a convenient amount for a volumetric determination. Decompose the sample by treatment with nitric and hydrofluoric acids in the manner described by Burstall and Wells.1 Finally remove hydrofluoric acid by repeated evaporations with nitric acid and take the sample to dryness. If the addition of dilute nitric acid indicates the presence of undecomposed material, filter the insoluble residue on to a filter-paper and ignite and fuse it in a nickel crucible with a few pellets of potassium hydroxide.Then add the melt to the filtrate and take the whole to dryness. Add 4ml of diluted nitric acid, 25 per cent. v/v, to the dry residue, gently warm to dissolve the mixture and then cool the solution, which is then ready for chromatography. EXTRACTION OF URANIUM- Transfer the sample on a wad of dry cellulose pulp to the top of the prepared alumina - cellulose column1 and extract the uranium with 200 to 250 ml of ether - nitric acid solvent if arsenic or molybdenum and arsenic is present in the original sample. If molybdenum alone is present, the amount of solvent can be reduced to 150 ml. Screen the column from direct sunlight. After removal of ether fr6m the eluate, determine the uranium volumetrically.lP2 The authors thank Mr. K. C. Overton for assistance in the experimental work. The investigations were carried out on behalf of the Ministry of Supply. published by permission of the Director of the Chemical Research Laboratory. This paper is REFERENCES 1. 2. 3. Burstall, F. H., and Wells, R. A,, Analyst, 1951, 76, 396. Burstall, F. H., and Williams, A. ,,F., “Handbook of Chemical Methods for the Determination of Arden, T. V., and Wells, R. A., Unpublished work. Uranium in Minerals and Ores, H.M. Stationery Office, London, 1950. RADIOCHEMICAL GROUP CHEMICAL RESEARCH LABORATORY TEDDINGTON, MIDDLESEX January, 1962

ISSN:0003-2654    

DOI:10.1039/AN9527700293

出版商:RSC    

年代:1952

数据来源: RSC

摘要:

June, 19521 WILLIAMS 297 Inorganic Chromatography on Cellulose Part IX The Use of Alumina and Cellulose Adsorbents for the Determination of Thorium and the Simultaneous Determination of Thorium and Uranium in Minerals and Ores BY A. F. WILLIAMS (Presented at the meeting of the Society on Wednesday, April 2nd, 1952) A method is described for the chromatographic separation of thorium from minerals and ores. Its application to the quantitative determination of this element in macro and micro amounts is described. The method can be applied to simple minerals such as monazite and to complex materials such as samarskite and pyrochlore, which are very difficult to analyse by the classical analytical procedures. Although thorium nitrate cannot be extracted from a cellulose adsorbent in the presence of phosphate ion by ether containing nitric acid, an extraction can be made from an alumina adsorbent.By the use of cellulose beneath the alumina, thorium nitrate can be extracted free from other ions. The method has been extended to the chromatographic determination of uranium and thorium in the same sample of material. Alumina is used as the principal adsorbent. THE problem of determining thorium in minerals and ores such as pyrochlore, euxenite and samarskite may present many difficulties when undertaken by the classical methods. Usually the difficulties increase progressively as the concentration of thorium in the sample decreases. Carriers are often added to ensure complete precipitation of the thorium and many analytical steps may be involved; relatively large amounts of sample may have to be taken.With the increased importance of low-grade sources of uranium ores, an accurate deter- mination of the thorium content is often desirable to assist in interpretation of radiometric results. Efforts were therefore made to find a method that was (a) reasonably simple and speedy in operation, (b) entirely selective and applicable at any level of thorium concentration in the sample, (c) applicable to relatively small quantities of sample and (d) generally applicable to all types of material. It was considered that a selective chromatographic method would be ideal for the separation and determination of thorium and might overcome all the difficulties inherent in precipitation procedures , particularly with low-grade samples.Uranium is readily determined after separation by chromatography on cellulose with ether - 5 per cent. nitric acid so1vent.l Kember2 extended the chromatographic separation method to the determination of thorium by using ether containing 12.5 per cent. of nitric acid and analysed the same sample of uranothoranite for both uranium and thorium by extraction with ether solvents containing 3 and 12.5 per cent. of nitric acid, respectively. This procedure is not directly applicable to ores containing phosphate ions, as preliminary chemical separation for the removal of such ions is necessary when analysing phosphatic ores. Examination of other adsorbents under similar conditions showed that thorium nitrate could be quantitatively extracted by ether - nitric acid solvent from alumina adsorbent when phosphate and ferric ions are present.This represented an important difference between alumina and cellulose. With this knowledge it was possible to devise methods for the deter- mination of thorium in any complex ore at all levels of concentration.298 WILLIAMS INORGANIC CHROMATOGRAPHY [Vol. 77 The general procedure for determination of thorium (and uranium, if required) involves fusion of the sample with potassium hydroxide t o effect decomposition. After acidification with dilute nitric acid, the hydroxides are precipitated* by adding ammonium hydroxide t o the hot solution. After filtration, the precipitate of hydroxides, which may also contain phosphate ions, is dissolved in nitric acid, the solution is evaporated to dryness and the residue is treated with a standard amount of nitric acid.Hydrogen peroxide is added to reduce all the cerium to the cerous state in order to prevent its extraction in ether - nitric acid solvent. The only ion other than uranium arid thorium that shows any marked tendency to be extracted from an alumina adsorbent is zirconium,2 but its movement is suppressed by adding phosphate ion, which is in turn made into a complex by adding ferric nitrate. The sample solution is adsorbed on alumina and transferred to an alumina - cellulose column. Cellulose, placed below the alumina, retains any small amounts of aluminium and iron that are not completely retained by the alumina. The “compound” column furnishes a novel example of the use of two adsorbents in inorganic chromatography. If thorium only is to be determined, extraction is carried out with ether containing 12.5 per cent.v/v of con- centrated nitric acid, but if both uranium and thorium are to be determined, this extract is re-extracted in a second column, first with ether containing 1 per cent. v/v of nitric acid and then with the solvent of higher acidity. With some ores, uranium and thorium can be extracted in turn direct from the same sample solution, only one column being used. The presence in some other ores of certain impurities enhances the extraction of thorium SO that it appears in the uranium extract, thereby necessitating the double extractions. Normally, thorium is determined as oxide after ignition of the oxalate, but, for routine purposes, provided that the uranium has been removed by the solvent containing 1 per cent.of acid, the thorium can be determined by ignition of the hydroxide. The lower limit of determination of the thorium depends solely on the method finally used for its measurement. The macro procedure is limited to about 0.1 to 0.2 per cent. of thoria when determined on a 1-g sample, and for this reason the micro method was developed. The micro procedure involves the spectrographic determination of the thorium - lanthanum ratio of the extracted thorium after co-precipitation in the presence of an excess of lanthanum (see part X of this series3). Experience with the chromatographic procedure for thorium has shown that its extreme simplicity leads to excellent reproducibility for repeated analyses, which would be difficult to attain with complex materials by classical chemical procedures.This advantage becomes more marked as the thorium content of the material decreases. The method generally appears to give results that are low by about 1 per cent. of the true thorium content of the sample and this has been confirmed by tracer techniques with uranium-XI. Results of analyses on very low-grade materials containing 0.001 to 0.2 per cent. of thorium show an over-all accuracy of about +lo per cent., which is considered excellent at such low levels. EXPERIMENTAL In the preliminary experiments made in order to find an adsorbent that would permit extraction of thorium in the presence of phosphate ions, alumina (chromatographic grade) showed great promise.Hence comparison between alumina and cellulose was made in a number of ways, as described below. EXTRACTION OF THORIUM NITRATE FROM ACTIVATED ALUMINA AND CELLULOSE- Phosphate ions absent.-Under similar conditions of extraction, with thorium nitrate equivalent to 0-3110 g of Tho2 in a volume of 1 ml of 25 per cent. v/v nitric acid solution and ether solvent containing 12.5 per cent. v/v of nitric acid, a comparison was made between alumina and cellulose. Results shown in Table I, relating to the rate and degree of extraction of thorium, show that thorium nitrate is extracted more quickly from alumina than from cellulose. Phosbhate ions $,resent-Under the same conditions as those described above but with the addition of 1 g 6f di-sodium hydrogen phosp‘hate to the sample wad, 20 per cent.of the thorium was extracted from alumina in the first 50 ml of solvent, and progressively smaller amounts were extracted in following fractions; there was no extraction from cellulose after passing 150 ml of solvent. * Recent experiments have shown that this operation is not essential,June, 19521 ON CELLULOSE. PART IX 299 Phosphate complexed with iron-A comparison of alumina and cellulose was made with iron added as complexing agent for phosphate. A larger volume (5 ml) of dilute nitric acid was used to dissolve the thorium nitrate and added salts; this resulted in a slower extraction TABLE I COMPARISON OF RATE OF EXTRACTION OF THORIUM FROM ALUMINA AND CELLULOSE ADSORBENTS IN ABSENCE OF PHOSPHATE Thorium, weighed as Tho,, extracted from 7 A 3 25-ml fraction alumina adsorbent, cellulose adsorbent, g g First 0-2483 0.1048 Second 0.0572 0.1826 Third 0.0028 0.0108 Fourth 0.0003 0.0028 of the thorium nitrate.Thorium nitrate equivalent to 0.3110 g of Tho, was taken together with 1 g of di-sodium hydrogen phosphate and 3 g of ferric nitrate, Fe(NO,),.SH,O, in a volume of 5ml of 25 per cent. v/v nitric acid solution, and extractions were made with ether containing 12-5 per cent. v/v of nitric acid; 150-ml fractions were taken. The results in Table I1 show an extraction of 99 per cent. from alumina compared with only about 30 per cent. from cellulose. The efficiency of extraction of thorium from the alumina was no less when a column of cellulose was placed below the column of alumina.TABLE I1 COMPARISON OF RATE OF EXTRACTION OF THORIUM FROM ALUMINA AND No phosphate ions were detected in the eluates. CELLULOSE ADSORBENTS I N PRESENCE O F PHOSPHATE AND IRON Thorium, weighed as Tho,, extracted from - A \ 150-ml fraction alumina adsorbent, cellulose adsorbent, First Second g 0.261 0.046 g 0.030 0.037 Third 0.001 0.049 APPLICATION OF ALUMINA ADSORBENT TO EXTRACTION OF THORIUM FROM MIXTURES WITH Kember2 showed that in the extraction of thorium from other ions by the use of cellulose adsorbent, the most troublesome elements were zirconium and, to a lesser extent, scandium, both elements being extracted with the thorium. He suppressed the movement of these ions with tartaric acid, but the acid was itself partly extracted. When alumina is used as adsorbent, zirconium is similarly extracted, but its movement is suppressed by the presence of phosphate ions.Movement of zirconium in the presence of phosphate ions-Columns of alumina adsorbent supported on cellulose were prepared. Solutions of zirconyl nitrate were prepared in 5 ml of diluted nitric acid (1 + 3) and different amounts of di-sodium hydrogen phosphate were added to find conditions under which zirconium was not extracted. Ferric nitrate, in the ratio of 3 g of the hydrated salt, Fe (N0,),.9H20, to 1 g of Na,HPO,, was added to complex the phosphate ions. The volume of ether - 12.5 per cent. nitric acid solvent collected was 450 ml. I t was shown that up to 0-5 g of zirconia (25-0,) was made into a complex by 1.4 g of di-sodium hydrogen phosphate, and no zirconium was extracted.With l-g samples of ores it is doubtful whether more than 0.5 g of zirconia would be present in practice, so that the addition of 1.4 g of di-sodium hydrogen phosphate should be sufficient to complex the whole of the zirconia; this amount of phosphate was used in all further experiments. On the assumption that ores might generally contain up to the equivalent of 0-6g of Na,HPO, as phosphate, about 8 g of hydrated ferric nitrate was used as complexing agent for the total of 2 g of Na,HPO,. With materials such as “pure” monazites and phosphatic rocks it would usually be unnecessary t o add any phosphate. Movement of zirconium and rare earth salts im presence of phosphate and ferric ions-The work described above was extended to mixtures of zirconium with rare earths.The movement OTHER IONS-300 WILLIAMS INORGANIC CHROMATOGRAPHY [Vol. 77 of the more common metal ions was not closely studied because it was known that these ions are normally retained by cellulose2 (see p. 302 for synthetic and complex ores). With a sample consisting of the following mixture of nitrates: Zr, 0-5g; Ce, 0.25g; Sm, 0.05 g ; Gd, 0-03 g ; Nd, 0.05 g; Y, 0-05 g ; La, 0.02 g ; Er, 0.003 g ; Tb, 0-007 g ; Ho, 0.009 g ; Sc, 0.002 g; extractions were made under the cond.itions described above. Cerium was first reduced with hydrogen peroxide and extraction carried out with 450ml of ether solvent. After removal of ether and precipitation of the aqueous extract with ammonium hydroxide a total of 3-9 mg of oxides was found; with oxalic acid, a precipitate that was only just visible was obtained.Extraction of small amounts of zirconium, iron and aluminium therefore accounted for this hydroxide precipitate. The double extraction procedure (see p. 303) decreases these impurities still further. GENERAL METHOD FOR THE DETERMINATION OF THORIUM IN MINERALS AND ORES AND ITS EXTEKSION TO THE DETERMINATION OF URA.NIUM AND THORIUM- I n the method described for the determination of thorium the size of the column, the diameter of which was 2.7 cm, was larger than that used by Burstall and Wells1 for the extraction and determination of uranium, in order to overcome mechanical difficulties involved in handling the comparatively large volume of sample solution (20 ml of dilute nitric acid) when alumina is used as adsorbent. With quantities of thorium ranging from 0.0038 to 0.196 g of Tho, recoveries of thorium were excellent.I t was found that with monazites and a number of other ores, uranium could be extracted before thorium by using ether solvent containing 1 per cent. of nitric acid and, as will be seen later, a general method was developed whereby uranium and thorium could be deter- mined on the same sample. A synthetic sample containing 0.050 g of U,O, and 0.0785 g of Tho, was prepared in dilute nitric acid solution containing 8 g of ferric nitrate nonahydrate and 1.4g of di-sodium hydrogen phosphate and two successive extractions were made, the first with 400 ml of ether solvent containing 1 per cent. of nitric acid and the second with 400 ml of solvent containing 12.5 per cent. of acid.The recoveries were quantitative. The methods given below describe the application of this technique for uranium and thorium determinations and also the simpler technique involving single extraction with ether containing 12.5 per cent, of acid when the determination of thorium alone is required. METHOD REAGENTS- All reagents should conform to recognised analytical standards. Potassium hydroxide-Pellets. Ammonium hy droxid e-Sp. gr . 0.880. Nitric acid, concentrated-Spgr. 1.42. Nitric acid, diluted (1 + 3). Hydrochloric acid, concentrated-Spgr. 1.18. Hydrojuoric acid, dihte-Add 1 ml of 40 per cent. w/v hydrofluoric acid to 100 ml of Oxalic acid-Crystals. Ferric nitrate-Solid Fe(NO,),.SH,O. Di-sodiztm hydrogen phosphate-Solid Na,HPO,.Hydrogen peroxide-A 20-volume solution. Ether-Redistilled; the water content must be less than 0.1 per cent. water. SOLVENTS- sp.gr. 1.42, to each 100ml of ether. acid, sp.gr. 1.42, to each 87.5 ml of ether. Ether containing 1 per cent. v/v of nitric acid-Add 1 ml of concentrated nitric acid, Ether containing 12.5 per cent. V/V of nitric acid-Add 12.5ml of concentrated nitric ADSORBENTS- Activated alumina-Type H, 100 to 200 mesh, as supplied by Peter Spence and Co., Ltd Activated celldose-Whatman cellulose powder for chromatography.June, 19521 ON CELLULOSE. PART IX 301 PURIFICATION OF FERRIC NITRATE FOR MICRO-DETERMINATION- Dissolve 250 g of ferric nitrate, Fe(NO,),.SH,O, in a final volume of 250 ml containing 25 ml of concentrated nitric acid.Transfer this solution to a 500-ml separating funnel and extract three times with 100-ml portions of ether. Warm the solution to remove ether, add 180ml of concentrated nitric acid and dilute to 750 ml. Use 20 ml of this solution, which contains approximately 25 per cent. of nitric acid and about 7 g of ferric nitrate, at the appropriate stage of the chromatographic extractions when very small amounts of uranium are to be determined. PURIFICATION OF ALUMINA FOR MICRO-DETERMINATION- Gradually pack about 1 kg of alumina into a glass column, 4.5 cm in diameter and 50cm long, by pouring small quantities into the column together with ether containing 12.5 per cent. v/v of nitric acid, until a homogeneous column of the alumina is formed. Wash the column with a litre of ether containing 12.5 per cent.of nitric acid and then with 1 litre of ether alone. Transfer the alumina to a large dish, dry it under infra-red lamps and store it in a dry bottle. PROCEDURE FOR PREPARING THE SAMPLE FOR CHROMATOGRAPHY- Weigh 10 g of potassium hydroxide pellets into a nickel crucible, 5 cm in diameter and 5 cm deep, and heat gently for about 10 minutes to remove water. Cool the melt and gently brush the accurately weighed sample, which should be about 1 g (or less if the thorium concentration exceeds 26 per cent. of Tho,), on to the surface of the cold melt. Cover the crucible and slowly heat the mixture to redness; continue heating for about 1 hour (monazites only need about 20 minutes, but more refractory ores such as pyrochlore or euxenite need longer).Allow the crucible to cool and leach the melt by immersing the crucible and lid in about 150 ml of dilute nitric acid in a 400-ml beaker covered with a clock glass. Wash the crucible with distilled water and then boil the contents of the beaker for 5 to 10 minutes. At this stage, the addition of 1 drop of dilute hydrofluoric acid sometimes helps to clear the solution,l but some ores, such as those rich in zirconia, do not respond to this treatment. Allow the contents of the beaker to cool slightly and cautiously add ammonium hydroxide until the solution smells distinctly ammoniacal. A heavy hydroxide precipitate is usually formed and this contains uranium and thorium together with some phosphate ions. Filter the precipitate on Whatman No. 541 filter-paper and wash it twice with hot water containing a few drops of ammonium hydroxide.Only slight washing is required since most of the neutral salts will be in the original filtrate. Open out the paper over the original beaker and wash the precipitate into the beaker with a stream of 50 to 100 ml of hot water.* Add about 40 ml of concentrated nitric acid to the beaker and then evaporate the solution until the residue is just moist. Add 20 ml of dilute nitric acid to the beaker, cover it with a clock glass, and heat the beaker on the edge of a hot-plate or below an infra-red lamp for 5 minutes. Then add 2 ml of hydrogen peroxide, replace the cover and heat the beaker for 10 minutes to reduce cerium to the cerous state. Add 8 g of ferric nitrate and heat the beaker again for 15 minutes.Finally add 1.4g of di-sodium hydrogen phosphate and heat for a further 15 minutes. The ferric nitrate must be added before the sodium phosphate, as otherwise the results will be low. With materials such as monazites, the addition of phosphate can be omitted or reduced to about 0.3 g, as such samples are usually low in zircon content; the quantity of ferric nitrate to be added is accordingly reduced to about 4 g. After thorough cooling the sample is ready for chromatography. PROCEDURE FOR THE CHROMATOGRAPHIC EXTRACTION OF URANIUM AND THORIUM TOGETHER Preparation of the column-The column consists of a tube of diameter 2-7 cm and about The bottom of the tube terminates in a short length It is closed by a short length A number of indentations should be made The whole column should preferably AND DETERMINATION OF THORIUM- 30 cm long, with a funnel at its top.of narrow-bore tubing to facilitate collection of the eluate. of polyvinyl chloride tubing with a screw clip. near the base of the tube to support the packing. be water-jacketed in a similar manner to a condenser. of alumina is sufficient to form a granular wad in the presence of neutral salts. * Recent work has shown that the ammonium hydroxide precipitation step can be omitted, as 50 g302 WILLIAMS : INORGANIC CHROMATOGRAPHY [Vol. 77 Make the column water-repellent by treating with siliconing fluid.l Place a small piece of Whatman’s ashless block at the bottom of the column, add activated cellulose and then ether containing 12.5 per cent.v/v of concentrated nitric acid. Add more pulp, if necessary, so that the cellulose reaches to a height of about 5 cm after “beating” with a glass p1unger.l Next pour activated alumina into the column so that after “beating” and allowing to settle, the alumina reaches to a height of about 6cm. Cover the completed column with solvent and it is then ready for use. Trans fer of sample to column and extraction of uraniwm and thorium-Add approximately 50g of alumina to the sample solution and thoroughly stir to give a homogeneous powder. Transfer the mixture to the prepared column with the aid of a stirring rod. When trans- ference is complete, immerse the alumina mixture in ether - 12.5 per cent. nitric acid solvent, the final level of which should be about 1 cm above the top of the sample wad.Beat this wad gently with a glass plunger to ensure that the column is homogeneous; during this operation care should be taken not to penetrate far into the main column. Rinse the beaker with successive small amounts of ether solvent and pour the rinsings on to the column. Continue to rinse until 500 ml of eluate” have collected in a 1-litre conical flask; the level of the solvent in the column must not be allowed to fall below the top of the wad. Remove the conical flask and add 20 ml of water followed by 50 ml of ammonium hydroxide to neutralise most of the nitric acid; cautiously swirl the flask during the additions. Add two glass beads and remove the ether by heating the flask on a steam-bath. Transfer the aqueous residue to a 250-ml beaker, cover it and boil for 5 minutes.Determination of thorium-Precipitate the hydroxides of thorium and uranium with ammonium hydroxide, filter, dissolve the precipitate in hot diluted nitric acid (1 + 3) con- taining 35 to 40 ml of concentrated nitric acid and collect the solution in the original beaker. Dilute the contents of the beaker to 250 ml, add ammonium hydroxide until the solution is just alkaline and then acidify with 10ml of concentrated hydrochloric acid. Heat the solution to boiling, add 10 g of oxalic acid and continue boiling for 2 to 3 minutes. Set the beaker aside for at least 4 hours and then filter the precipitated oxalates, wash and ignite them to Tho, in the usual way and weigh the oxide. For amounts of thorium giving a final precipitate of less than about 0.07 g of thoria it is advisable to precipitate the oxalate in a smaller volume, in which event the quantities of all reagents should be proportionally decreased.For less than 0.01 g of thoria the pre- cipitation should be made in a volume of about 12.5 ml instead of 250 ml. PROCEDURE FOR THE CHROMATOGRAPHIC EXTRACTION OF URANIUM AND THORIUM SEPARATELY Prepare the chromatographic column as described above (p. 301), except that ether containing 1 per cent. of concentrated nitric acid is used as solvent. Transfer the sample to the column as described above and elute the uranium and thorium in two extractions, first with ether containing 1 per cent. v/v of nitric acid, 400 ml of eluate being collected, to extract uranium, then with ether containing 12.5 per cent.v/v of nitric acid (the wad should be beaten up once on changing the solvent), the 400 ml of eluate collected containing the thorium. Determine the uranium by a standard procedure, either colorimetrically or volumetrically, depending on the level of uranium pre~ent.~ Determine the thorium as described above. AND THEIR DETERMINATION I N CERTAIN SAMPLES- RESULTS Table I11 shows results obtained for widely different materials by the procedures described above. The uranium and thorium contents of the samples were known from many other analyses made by the standard procedure. Some of the low-grade samples had known amounts of thorium added as a standard thoriuni nitrate solution immediately after fusing the sample with potassium hydroxide. The results in Table I11 show that thorium cart be determined in a wide range of materials by using a direct extraction with ether containing 12.5 per cent.v/v of nitric acid. The synthetic material of sample 11 was composed of 0.125 g of siliceous ore, 0.075 g of stannic * This larger volume of eluate was found preferable for complex ore samples.June, 19521 ON CELLULOSE, PART I X 303 oxide, 0.025 g of titanium dioxide, 0-25 g of zirconia, 0.3 g of a mixture of cerium and rare earths isolated from a monazite, and 0.05 g of each of the elements Sr, Ba, Zn, Pb, Mg, Mo, Sb, Bi, V and Cu. This mixture was very complex but the extracted thorium was found to be of high purity. With this sample and a number of complex ores it was not possible to determine uranium and thorium in turn owing to extraction with the ether - 1 per cent.nitric acid solvent of thorium nitrate, which contaminated the uranium fraction. However, TABLE I11 TYPICAL RESULTS OF DIRECT EXTRACTIONS OF MINERALS AND ORES Uranium (U,O,) Thorium (Thol) Sample Description of sample of sample, present, found, present, found, Weight (-fi-, 1 2 3 4 6 6 7 8 9 10 11 12 13 14 Extraction for uranium and thorium in tur% with ether - 1 per cent. nitric acid and ether - 12.5 per cent. nitric acid sohents- Blank on column and reagents (full “Pure” monazite . . .. .. Highly refractory ore ; containing more than 50 per cent. of zircon, some monazite, ilmenite, magnetite, cassi- terite, thorite, euxenite and silica . . ditto . . .. . . .. Refractory ore, mainly monazite . .Uranothorianite . . .. . . .. Thorianite . . .. .. .. Refractory ore, largely siliceous . . procedure) . . .. .. .. Extraction for thorium with ether - 12.5 per cent. nitric acid solvent- Pure monazite . . .. . . .. Pyrochlore . . .. .. .. Synthetic material . . . . .. Low grade monazite . . .. .. A.217 .. .. .. .. .. Monazite .. .. .. .. nil nil nil nil nil 1.0 0.0035 0.0035 0-0980 0.0960 1.0 0.0018 0.0018 0-0070 0.0069 1.0 0.0568 0-0570 0.0855 0.0860 1.0 0.0046 0.0046 0-0930 0.0930 0-5 0.1920 0.1932 0-2685 0.2665 0.25 0.0164 0.0165 - 0.2202 1-0 Analysis not completed owing to extraction of thorium with uranium fraction (see Table IV) 1.0 0.0980 0.0970 1.0 - 0.0250 - 0.0795 0-0797 - nil 1.0 - 0.0009 - 0.0272 1.0 0.0070 0.0080 1.0 0.0980 0.0960 the procedure was successful with other samples, particularly pure monazites, so that in a laboratory dealing with standard types of ore, the procedure could probably be applied successfully.The explanation of the occurrence of this phenomenon with certain types of ore is not clear, but as will be seen later the difficulty has been overcome by carrying out these successive extractions on the eluate from an initial extraction with ether - 12.5 per cent. nitric acid solvent, which will contain only uranium and thorium. The following procedure appears to be applicable to all samples. METHOD FOR DETERMINING URANIUM AND THORIUM IN ANY SAMPLE The simultaneous determination of uranium and thorium can be applied to aI1 ores by A second extraction after a modification of the method described above (pp.301-302). isolation of the two nitrates together is involved. PROCEDURE- Prepare the sample and carry out the chromatographic extraction of uranium and thorium together as described above (pp. 301-302). After removing the ether from the eluate from the extraction with ether - 12-5 per cent. nitric acid, add to the hot solution 0.25 g of ferric nitrate, 0.05 g of di-sodium hydrogen phosphate and sufficient ammonium hydroxide to precipitate the hydroxides (or phosphates) of iron, uranium and thorium. The iron and phosphate have the dual purpose of acting as carriers should the amounts of uranium and thorium be very small and also as complexing agents, in the second extraction column, for traces of zirconium that may have been extracted in the first chromatographic separation.Filter the precipitate, dissolve it in 50 ml of hot diluted nitric acid (I + 3) and then evaporate this solution until the residue is just moist. To the residue add 5 ml of diluted nitric acid (1 + 3), followed by 1 ml of hydrogen peroxide; then gently heat the solution for 10 minutes304 WILLIAMS INORGANIC CHROMATOGRAPHY [Vol. 77 to reduce ceric compounds, traces of which may be present from the first column, Adsorb this solution on 12-5 g of alumina and transfer this wad to an alumina - cellulose column prepared in ether - 1 per cent. nitric acid solvent:; the heights of alumina and cellulose in this column should be about 5 and 4 cm, respectively. Extract the uranium with 250 ml of ether - 1 per cent. nitric acid solvent and the thorium with 350 ml of ether - 12-5 per cent.nitric acid solvent. The amounts of adsorbents and solvents are less than those used in the direct extractions as less phosphate and aqueous solution are present in the wad. If the blank for reagents is allowed for, thorium can be determined by hydroxide precipitation for all except low-grade ores. Finally determine uranium and thorium by standard methods. RESULTS The results shown in Table IV are given for a very wide range of materials. Some of the lower-grade samples had known amounts of uranium and thorium added. Determina- tions on samples 1 and 2 were made on the refractory ore (Table 111, sample 8) that could not be analysed by direct extractions for both uranium and thorium. TABLE IV TYPICAL RESULTS BY THE PROCEDURE APPLICABLE TO ANY SAMPLE Thorium (Tho,) v - 7 Weight Uranium (U,O,) (oxalate (hydroxide of 7- precipita- precipita- found found Sample Description of sample sample, present, found, present, tion), tion), 1 2 3 4 6 6 7 8 9 10 11 12 13 14 16 16 17 18 19 20 21 Refractory ore, largely siliceous Refractory ore, largely siliceous Refractory ore (containing Highly refractory, >50 per cent.Monazite (“pure”) . . .. Monazite (“pure”) . . .. monazite) . . .. .. Refractory (thorianite) . . * . of ZrO, .. .. .. Standard sample . . .. Standard sample . . .. Standard sample . . .. Standard sample . . .. Monazite .. * . .. Refractory, siliceous . . .. Synthetic .. .. .. Synthetic .. .. .. Synthetic .. .. .. Synthetic .. * . .. Synthetic .. .. .. Synthetic .. .. .. Monazite .... .. Monazite .. .. .. g 1.000 1.000 1.000 1.000 1.000 0-500 1.000 1.000 1.000 1.000 1.000 1.000 1.000 - - I - - - 1.000 1,000 g 0.0006 0.0506 0.0035 0.0035 0.0012 0.0047 0-05 18 0.0004 0.0004 0.0004 0.00004 0.0035 < 0~00006 0~0001 0.0501 0.0001 0*0001 0~0001 0*0001 0.0067 - g 0.0006 0.0500 0.0034 0.0036 0~0010 0.0045 0.0500 0.00034 0.00043 . 0.00034 0*00006 0.0035 <0.00005 0*0002 0.0498 0-0002 - - - 0.0017 0.0067 g 0.0820 not known 0.0980 0.0980 0.0144 0.2520 0.0070 0.0112 0.0112 0.01 12 0.0012 0.0940 <0*0010 0.0010 0.00 10 0*0040 0.0124 0.0645 0.1980 0.1390 - g 0.0780 0.0040 0-0960 0.0975 0-0146 0.2510 0.0074 0-0104 0.0119 0.0012 0.0950 N.D. 0-0020 0.0013 0.0037 0.0120 0.0660 0.1961 0-0640 0.1393 - g 0.07 80 0.0980 0.0984 0.0158 0.2519 - - 0-0120 - - 0.0004 0.0038 0.0024 The composition of the synthetic samples 14 to 19 was as follows: 0.125 g of siliceous refractory ore, 0.075 g of stannic oxide, 0.025 g of titanium dioxide, 0.25 g of zirconia, 0.3 g of a mixture of cerium and rare earths from monazite (0.012 g of Sc, 0.05 g of Sm, 0.03 g of Gd, 0.05 g of Nd, 0.05 g of Y, 0.02 g of La, 0.003 g of Er, 0.007 g of Tb, 0.009 g of Ho and 0458 g of Ce, all as oxides) and 0-05 g of each of Slr, Ba, Zn, Pb, Mg, Mo, Sb, Bi, V and Cu.I n addition, sample 19 contained 0.1 g of Nb,O, ‘and 0.1 g of Ta,O,. Thoria from a large number of the oxalate precipitations was examined spectrographically and found to be pure. The thoria precipitate was free from scandium, which showed that by this procedure there is unlikely to be extraction of scandium from any ore that would be met in practice.Scandium was one of the few elements forming hydroxides that showed any tendency to move.2 The results in Table IV show that the method is reasonably accurate for both uranium and thorium, even at low concentration. The results for thorium by hydroxide precipitationJune, 19521 ON CELLULOSE. PART IX 305 are slightly higher than those by oxalate precipitation to an extent that approximates to the blank (cj. result on sample 13). EXTENSION OF CHROMATOGRAPHIC PROCEDURE TO THE DETERMINATION OF MICRO AMOUNTS OF THORIUM With 1-g samples, concentrations of less than 0.2 per cent. of thoria are too low for application of the usual gravimetric techniques. Further work has shown that the chromato- graphic method described can be extended to the separation of traces of thorium in high purity, so enabling a standard spectrographic procedure to be applied in the final estimation.After chromatographic separation of the thorium it is co-precipitated with lanthanum as oxalate and the lanthanum is used as an internal standard for the spectrographic procedure described in part X of this series3 After initial extraction of both uranium and thorium with ether containing 12.5 per cent. of nitric acid, second extractions are made with ether containing 1 per cent. and 12-5 per cent. of nitric acid, respectively, as described for the macro method. It was shown that thorium is not extracted by the 1 per cent. solvent even in micro amounts from mixtures of pure uranium and thorium.Thorium nitrate equivalent to 4 mg of thoria (Tho,) was extracted with ether containing 1 per cent. of nitric acid under the standard conditions described and the analysis of the extract for thorium showed less than 10 pg of thoria. As with the macro method, uranium can be determined on the same sample. CHROMATOGRAPHIC SEPARATION OF THORIUM AND ITS DETERMINATION IN MINERALS AND The procedure for the separation of thorium from the material under analysis is almost identical with that described in the macro method (p. 301). If the amount of uranium present is greater than about 1 per cent. the full procedure (p. 303) is used and similarly if both uranium and thorium are to be determined. If a determination of thorium alone is wanted and if uranium is less than about 1 per cent.of the sample, the procedure with only one extraction with ether- 12.5 per cent. nitric acid solvent (p. 301) may be used. After chromatographic separation, the thorium is precipitated as oxalate, lanthanum being used as carrier, and the mixed thorium - lanthanum oxalates are ignited to the corre- sponding oxides prior to spectrographic determination of the thorium according to the procedure described by Kingsbury and Temple.3 Experiments have shown that the precipitation of lanthanum oxalate is not quantitative in the presence of large amounts of neutral salts. Ether is therefore removed from the eluate containing 12.5 per cent. of nitric acid without adding ammonium hydroxide (see macro procedure). It was found that, provided a volume of not less than 200 ml of water is added, ether can be removed by distillation without danger of peroxidation.ORES- PREPARATION OF MIXTURE OF THORIUM AND LANTHANUM FOR SPECTROGRAPHIC ANALYSIS- After removing ether from the ether extract containing thorium (no ammonia added), evaporate the aqueous solution to a few millilitres over a bunsen burner and wash the contents of the flask into a 250-ml beaker. Add 2 ml of a stock lanthanum solution (approximately 1 g of “Specpure” lanthanum oxide (La,O,) dissolved in a few millilitres of nitric acid and diluted to 100 ml) and evaporate the contents of the beaker just to dryness. Add 2 ml of concentrated nitric acid followed by 20 ml of water, make the solution just alkaline with ammonium hydroxide and then acidify with 0.75 ml of concentrated hydrochloric acid.Boil the solution, add 0.5 g of oxalic acid and continue boiling for approximately 2 minutes. Set the beaker aside for at least 8 hours. Filter the oxalates through a Whatman No. 42 filter-paper con- taining a small amount of added paper pulp, wash with a solution containing 2 per cent. of oxalic acid and 0.1 per cent. v/v of concentrated hydrochloric acid, and ignite. Weigh the ignited oxides exactly (there should be about 20 mg), determine the percentage of thorium spectrographically and calculate the thorium content of the original sample. The blank is usually less than 10 pg of thoria, so that with a 1-g sample of material, the lower limit for the deter- mination of thorium is 0.001 per cent. Results for two different types of refractory ore and for reagents to which known amounts of thorium had been added are given in the table Perform a blank determination on the reagents and full procedure.306 WILLIAMS [ V O L 77 (p. 311) that follows the spectrographic procedure described by Kingsbury and Terrrpk+ in the following paper (part X of this series). They show that the over-all accuracy of the procedure is of the order of *lo per cent. for amounts of thoria down to 0.001 per cent. Good results were obtained for samples of low-grade shales. DETERMINATION OF URANIUM- If the determination of uranium is also required, it can be made on the appropriate solvent fraction as described for the macro procedure (p. 302). If the uranium content of the sample is less than about 200 pg, the alumina should be washed with ether containing 12.5 per cent. v/v of nitric acid before use; the ferric nitrate should be treated similarly (see p. 301). Uranium can be determined by a standard procedure, usually colorirnetrically or fluorime trically . The author thanks Messrs. E. C. Hunt and G. H. Smith for assistance in the experimental The work was carried out on behalf of the Ministry of Supply. This paper is published work. by permission of the Director of the Chemical Research Laboratory. REFERENCE: s 1. 2. 3. 4. NOTE-References 1, 2 and 3 are to parts V, VII and X of this series. CHEMICAL RESEARCH LABORATORY TEDDINGTON, MIDDLESEX Burstall, F. H., and Wells, R. A., Analyst, 1951, 76, 396. Kember, N. F., Ibid., 1952, 77, 78. Kingsbury, G. W. I., and Temple, R. B. F., Ibid., 1952, 77, 307. Burstall, F. H., and Williams, A. F., “Handbook of Chemical Methods for the Determination of Uranium in Minerals and Ores,” H.M. Stationery Office, London, 1950. RADIOCHEMICAL GROUP

ISSN:0003-2654    

DOI:10.1039/AN9527700297

出版商:RSC    

年代:1952

数据来源: RSC

摘要:

June, 19521 KINGSBURY AND TEMPLE 307 Inorganic Chromatography The Spectrographic Part X Determination of on Cellulose Micro Quantities of Thorium Separated by Chromatography from Minerals and Ores BY G. W. J. KINGSBURY AND R. B. F. TEMPLE (Presented at the meeting of the Society 012 Wednesday, April 2nd, 1952) A simple spectrochemical method is described for the determination of micro quantities of thorium separated by chromatographic methods from small samples of about 1 g of minerals and ores. The extracted thoria is co-precipitated with lanthanum, which forms the matrix and acts as internal standard in the spectrographic analysis. The spectrochemical procedure alone has a standard deviation of k3 per cent. and has been operated for ore concentrations between 0.001 and 0.2 per cent.of thoria, only 1 g of the original material being used. For concentra- tions above 0.2 per cent. of thoria, the amount of ore necessary is considerably reduced. SUCCESSFUL chromatographic methods making use of alumina and cellulose adsorbentsf have been described for the separation of pure thoria from very low-grade ores.12 The lower limit of estimation of the thorium content previously attained was, however, only about 0.1 per cent., with 1-g samples of ore, this limit being imposed by the use of direct weighing to determine the extracted thoria. Attention was therefore directed to developing a spectrochemical technique for the estimation of thoria with the object of lowering the limit of detection. A suitable procedure has now been evolved and has proved satisfactory in the range of 10 to 2000 p.p.m.of thoria (Tho,). The essential feature of the method is the co-precipita- tion of the extracted thoria with a lanthanum “carrier,” which enables the very small amounts of thoria to be handled and also acts as internal standard in the spectrographic procedure. A solution of known lanthanum content is added to the solution containing the purified thorium and the metals are precipitated as their oxalates, which are then ignited. The ratio of thorium to lanthanum in the final product is then determined spectrographically and the percentage of thorium in the original ore is determined from a knowledge of the exact weight of the mixed oxides resulting from the extraction process. (It is assumed that all the thoria is removed by the chromatographic procedure.) Although it is possible to separate any uranium chromatographically from the thorium, it has been shown that amounts of uranium up to about 1 per cent.of the original ore do not interfere with the spectrographic results, as the uranium is adequately separated by the final precipitation of the thorium as oxalate. It is possible that the upper limit that can be tolerated may be higher than this figure. EXPERIMENTAL AND METHOD With minor modifications, the experimental technique is similar to that described for the analysis of hafnium - zirconium mixtures,2 but for convenience the practical details are summarised below. The most important differences lie in the use of plate calibration by means of a rotating sector and in the fact that the amount of material necessary for a single determination has been reduced to 3 mg, only the lower graphite electrode being coated.The method permits routine analyses to be made without consuming large quantities of valuable accurately-known standards. These standards are used only once in order to prepare a “standard” plate.308 KINGSBURY AND TEMPLE : INORGANIC CHROMATOGRAPHY [Vol. 77 OPERATING CONDITIONS- A circuit for the source unit has already been published., A rotating auxiliary gap instead of a fixed one is now used, but this has no effect on the spectra. Operating conditions for both sample and calibration exposures are shown in Table I, and conditions that differ for sample and calibration exposures are shown in Ta.ble 11. TABLE I CONSTANT OPERATING CONDITIONS Spectrograph .. .. .. . . , . Large Hilger Quartz (Littrow) Wavelength range . . .. .. . . 2400 to 3450 A Spherical lens . . .. .. .. . . Used to focus the source on the collimating lens Distance of source from slit . . .. .. 38cm Slit width . . . . .. .. .. . . 10 p (fixed) Electrode separation . . .. .. . . 4 mm, adjusted optically Supply voltage . . . . .. .. . . 440 volts A.C. 3-phase Current . . .. .. .. .. . . 10 amp. with gaps G, and G, shorteda Pre-arcing period .. .. .. .. Nil Plate . . .. .. .. .. . . Ilford Soft Ordinary Development . . .. .. .. . . 4 minutes a t 20" C by normal dish develop- Line pair used . , .. .. .. . . Th 2837-3 A and La I1 2821.04 A Densitometer . . .. .. .. . . Hilger non-recording microphotometer inent methods in Kodak D 19 B TABLE I1 VARIABLE OPERATING CONDITIONS Sample Iron calibration Slit Zength-1.5 mm .. . . . . . . Appropriate to sector used Lower electrode-#-inch length of &-inch diameter graphite rod (Johnson, Matthey JM4B) with one end faced flat. A drop of bakelite t o the optical axis solution (25 per cent. w/v in alcohol) is placed on the flat end and 3 mg of sample immediately added through a small funnel. The sample is mixed and spread evenly over the face df the electrode, allowed to dry and then baked a t 130" C for 30 minutes. Two such electrodes are prepared for each sample Upper electrode-&inch diameter graphite rod, the end turned to a 90" truncated cone, the flat portion having a diameter of 2 mm &-inch diameter iron rod with a wedge-shaped end, the length of the wedge being parallel As lower electrode Capacitance C,-48 p F .. .. .. .. 16pF Time of exposure-30 seconds . . .. . . 90 seconds PLATE CALIBRATION- Of the various means of plate calibration that: have been suggested, for example, by Harri~on,~ Kaisel"l and Argyle and Price,6 the rotating logarithmic-step sector is far the most convenient when a wide range of concentrations is involved, and although there has been much controversy over the validity of its use, a number of workers, particularly Twymane and Webb,' have established that it can give good results. A possible objection that might be raised against it is the danger of stroboscopic effects, which might arise when a mains- operated sector motor is used in conjunction with an intermittent (impulse) arc. As a measure of the validity of this method, the standard deviation from the means of the duplicate determinations of the standards was calculated and was found to be about 2 3 per cent.; this appears to indicate that the method is self-consistlent. ChurchilP has previously described the use of a rotating sector with an intermittent source.Strictly speaking, the step sector varies exposure time and not intensity, where E = I T (E = exposure, I = intensity, T = time), and for two different exposures- El - I l T l G - V ; where the subscripts denote the corresponding quantities and where I , # I , and T , f T,. If a given density resulted from a single unique value of E regardless of I and T , 11/1, would equal T,/T,.Owing to the failure of the "reciprocity law" this is not so, and I J I , 7t Tl/T,.June, 19521 ON CELLULOSE. PART X 309 However, we may regard the exposure-time ratio as a “photographic-intensity ratio” that will differ from the true intensity ratio of the lines by an amount that is a constant for a given emulsion at a given wavelength. I t is this photographic-intensity ratio that is used in the following analyses and that is loosely termed the “intensity ratio.” (For a discussion of this problem see S t r ~ c k . ~ ) The sector is of the symmetrical double-cut type for good balancing at high speeds and it is rotated just in front of the entrance slit of the spectrograph by a small non-synchronous motor running at about 1500r.p.m. I t has 7 steps, which give exposure-time ratios of 1 : 2 : 4: 8 : 16: 32 : 64, and these are so proportioned that an eighth step with a relative exposure time of 128 can be obtained by illuminating a portion of the slit, about 1.5 mm long, with uninterrupted radiation.A “standard” plate is prepared by exposing, in duplicate, in the usual way, samples of mixtures of thorium and lanthanum in known ratios and recording on the same plate an iron spectrum photographed through the rotating sector. The exposure time for the iron spectrum is chosen so that the weakest step is just visible on the iron line chosen for calibra- tion. It is important that the same kind of source, ie., the impulse arc, be used for both analysis and calibrating exposures, but it has been found that a better ratio of line to back- ground intensity is achieved for the calibration spectrum by reducing the capacitance and increasing the exposure time.Electrodes of half-inch diameter iron rod are used for the calibration exposures to prevent incandescence, which might give rise to a continuum. A suitable iron line as near as possible in wavelength to the line used for the analysis is then chosen and the densities of the 8 steps are measured in the usual way. These densities, or some suitable function of them, are then plotted against the values of the corresponding logarithm of the relative exposure time. If densities are used directly, an S-shaped curve is obtained that is very similar to the familiar Hurter and Driffield curve, but for measurements a t low intensities, and particularly if background corrections are to be applied, this procedure leads to an inconveniently curved line.Attempts have been made to use functions of density ---or galvanometer-readings that will compensate for the curvature at low intensities and produce a straight line.4 For this purpose we have used the straightforward Seidel-function, which is defined by the expression- T = log,, (+ - 1). where A and A , are the galvanometer deflections for the line and clear plate, respectively. By always adjusting A , to a predetermined value, a table of T values can be drawn up against galvanometer readings. I n some regions of the spectrum, the ordinary Seidel-function tends to overcompensate a t low intensities and again produces a curved calibration line. It is possible to find a modified function that is exactly applicable to the wavelength in use, but this was considered unnecessary for the present work.If the method is valid, calibration curves for iron measured at various levels of oder-all densitv will produce a series of parallel calibration curves that are merely displaced sideways along the exposure-time axis by various amounts. Then, as log(Il/12), which equals log I, - log I,, is the quantity determined, the ratio obtained will be independent of the over-all density. APPLICATION TO THE DETERMINATION OF THORIUM- For a given plate, the density is plotted against the logarithm of the relative exposure time, and the logarithms of the intensity ratios of the analysis and the standard lines, i.e., log IAIIs, can then be derived for the spectra of the various standard mixtures.These are plotted against the logarithm of the corresponding thorium - lanthanum ratios. Analyses of unknown mixtures are made by photographing a sectored iron spectrum on each analysis plate, constructing the curve relating density to the logarithm of the relative exposure time as described above and measuring from it the logarithms of the intensity ratios of the lines in the spectra of the unknown mixtures. These intensity ratios are then referred back to the original “standard” plate so that the concentrations can be determined. To permit a wider choice of lines for the analysis during the exploratory work, a special cell was used to enable the arc to be struck in a stream of carbon dioxide. This reduced the general background intensity and almost entirely suppressed the dense atmospheric cyanogen bands, which largely obscure the region between 3500 and 4 3 0 0 ~ .A description of this cell has been published elsewhere.lo Apart from the advantages gained by suppressing the cyanogen bands, it has been found that use of the cell leads to a significant reduction of the310 KINGSBURY itND TEMPLE : INORGANIC CHROMATOGRAPHY [Vol. 77 standard deviation of the method. To obtain this improvement in accuracy it is not necessary to use carbon dioxide, and the analyses are normally carried out in a stream of air. If it is considered necessary, correction for the spectrum background can be made by subtracting the intensity of the background in the neighbourhood of the line (taken from the iron calibration curve for the plate) from the intensity of the line plus background. This part of the working can be simplified to some extent by the use of Gaussian subtraction logarithms.Useful tables of these and their derivatives have been given by Mitchell and Scott.11 The present determinations have been made in the range 0.1 to 10 per cent. of thoria (Tho,) in the spectrographic material (La,O,). On the ores this represents a range of about 20 to 2000 p.p.m. The upper limit could be raised, if required, either by the selection of another line pair or by the use of a step-weakener. The pair giving the best results is La I1 2821.04 A - Th 2837.3 A. It is significant that the curvatures of all the working graphs recorded indicate the presence of about 0.01 per cent. of thorium A number of line pairs have been tried in order to discover the most suitable.Fig. 1. Working graph as an impurity in the lanthanum used. An attempt at purification had no effect on this apparent residuum. Although background corrections were originally made, experience showed that greater accuracy in the concentration range studied can be attained by omitting them. Two sets of standards were prepared, as described below, with 20 and 100 mg, respectively, of the lanthanum base. As a single working graph can be drawn from these two sets of data, as shown in Fig. 1, the measured ratio of thorium to lanthanum is, between these limits, satisfactorily independent of the quantity of base used in the separation. To test the repro- ducibility of the spectrographic results, a number of exposures of a selected standard were made under identical conditions.The standard (deviation had different values depending on whether the analyses were carried out in still air or in a current of air or carbon dioxide, viz. , in still air, k4.5 per cent. ; in air flow, f 2-6 per cent. ; in carbon dioxide flow, 3.0 per cent.; these values are expressed in relation to the means, no background correction being applied. The last two values are probably not significantly different. The values are slightly higher when a background correction is made. A number of exposures, in duplicate, with standards containing between 0.1 to 15 per .cent. of thoria, made in a stream of carbon dioxide under normal conditions, gave a standard deviation from the mean of the duplicates of f3-4 per cent.(without background correction) and +4-0 per cent. (with background correction) for the line pair specified above, and these values represent the total standard deviation of the method in this range of concentration. PREPARATION OF STANDARD MIXTURES OF THORIUM AND LANTHANUM- Sufficient quantities of a standard solution to give approximately 100 mg ( ~ 2 0 per cent.) of lanthanum were measured into a series of 250-ml beakers and standard thorium solution was added to each to cover the range of 50 to 10,000 mg. About 10 concentrations were chosen to give nearly equidistant spacing 011 a logarithmic concentration scale. ToJune, 19521 ON CELLULOSE. PART X 31 1 each beaker 8 ml of concentrated nitric acid were added and the solutions diluted to approxi- mately 100ml.They were then made alkaline with ammonium hydroxide, spgr. 0.880, acidified with 4 ml of concentrated hydrochloric acid and boiled. After adding 3 g of oxalic acid, boiling was continued for about 2 minutes and the contents were then set aside for at least 8 hours. The precipitates were filtered on Whatman No. 42 filter-paper containing a small amount of added filter pulp, washed with 2 per cent. oxalic acid solution containing 0.1 per cent. of concentrated hydrochloric acid. They were then ignited in weighed platinum or silica dishes which were again weighed. The ratios of thorium to lanthanum in the ignited oxides were then determined spectro- graphically. TEST OF THE METHOD FOR THORIUM EXTRACTED CHROMATOGRAPHICALLY FROM ORES- The method was tested by adding known quantities of thoria to two sets of low-grade ores, carrying out the full chromatographic extraction12 and estimating the thoria recovered.Additional experiments were made in which thorium oxalate was precipitated in the presence of 10 mg of U,O, in addition to lanthanum carrier. There was no interference with the spectrographic method, which showed that for low-grade uranium ores a single chromato- graphic extraction would suffice (see part IX).l2 Samples numbered 1 consisted of reagents only. Samples numbered 2 were siliceous ores containing small amounts of monazite, pyrochlore, thorite and zircon. The amounts of U,O, and Tho, were 0.013 and 0.016 per cent., respectively, on the original ore.Samples numbered 3A and 3B were monazite - dunite mixtures containing 11 mg of Tho, and less than 1 mg of U,O, per g of ore. Those numbered 3B include repeat deter- minations. Typical results are shown in Table 111. TABLE I11 TYPICAL RESULTS OF DETERMINATIONS BY THE METHOD DESCRIBED Total Tho, present, Sample Tho, added, 7 Tho, found, Pg Pg % % 10 Pg 1 nil - - 1 2 3A 3B This test served as a blank and the Tho, percentages below are obtained by subtracting this value from that determined 56 56 0.0056 0.0043 150 150 0.015 0.013 0.016 0.016 nil 160 75 230 0.023 0.023 200 387 0.036 0.039 300 450 0.046 0.045 300 490 0.046 0.049 1000 1076 0.116 0.108 1125 1415 0.128 0.141 nil 10 0.001 1 0.001 7-5 20 0.0019 0.002 100 91 0.011 0.009 500 556 0.051 0.056 600 620 0.061 0.062 1000 1014 0.101 0.101 100 111 0.01 1 0.01 1 100 97 0.01 1 0.010 100 100 0.01 1 0.010 500 525 0-05 1 0.053 500 530 0.05 1 0-053 Error, % - - 26 - 13 0 0 + 8.3 - 2.2 + 6-5 - 7 + 10 - 10 + 5 - 18 + 10 + 1.5 0 0 - 10 - 10 + 4 + 4 By increasing the sample weight to 2 g, low-grade shales containing as little as 0.0005 per It will be seen that the over-all accuracy of the combined chemical and spectrographic This is considered very satisfactory in view of the cent.of thoria have been successfully analysed. procedures is at least within small samples used in the separation procedure. 10 per cent.312 KINGSBURY AND TEMPLE [Vol77 The authors thank Mr. E. C. Hunt, who carried out the chemical work. The work was This paper is published by permission of carried out on behalf of the Ministry of Supply.the Director of the Chemical Research Laboratory. RE FE REN C:ES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. Ryan, W., and Williams, A. F., Analyst, 1952, 77, 293. Kingsbury, G. W. J., and Temple, R. B., J . Appl. Chem., 1951, 1, 406. Harrison, G. R., J , Opt. Soc. Amer., 1934, 24, 5191. Kaiser, H., Spectrochim. Acta, 1941, 2, 1. Argyle, A., and Price, W. J,, J . Soc. Chem. Inci!., 1945, 67, 187. Twyman, F., and Simeon, F., Trans. Opt. SOC., 11929-30, 31, 169. Webb, J. H., J . Opt. SOC. Amer., 1933, 23, 157, 316. Churchill, J. R., Ind. Eng. Chem., Anal. Ed., 1944, 16, 664. Strock, L. W., Spectrochim. Acta, 1939, 1, 117. Kingsbury, G. W. J., and Temple, R. B., “A N,ew Electrode Holder for Running a Carbon Arc Spectrochim. Acta., 1952, 4, 473.Mitchell, R. L., and Scott, R. O., Macaulay Institute for Soil Research, Annual Reports, 1942-43 Williams, A. F., Analyst, 1952, 77, 297. in an Inert Atmosphere, and 1943-44. NOTE-References 1 and 12 are to parts VIII and IX of this series. INORGANIC GROUP CHEMICAL RESEARCH LABORATORY TEDDINGTON, MIDDLESEX DISCUSSION ON THE FOREGOING TWO PAPERS MR. A. R. POWELL said that he had had nearly 40 years’ experience of the classical methods of separating thorium and uranium from their mineral associates, but had adopted the chromatographic procedures developed by Mr. Williams and his colleagues as soon as they were available; he had found that these methods were a very great advance on the older ones both in simplicity of working and in accuracy of results. Although the author did not comment on the behaviour of scandium in the new procedure, i t would appear that it should follow the thorium, from which i t could subsequently be separated by complexing with tartrate.With regard to the separation of traces of thorium from much zirconium, e.g., in the analysis of zircon, presumably one could modify the procedure in such a way as to obtain an eluate containing a proportion of the zirconium to act as a carrier for the thorium and then re-treat this solution in the way described in the paper to eliminate the remainder of the zirconium. Cme difficulty of using ether as the eluent was its great volatility, especially in warm weather; would i t be possible to use isopropyl ether, which was a satis- factory substitute for ethyl ether in gallium extractions, and.was much less volatile and had a sufficiently low boiling-point to allow i t to be readily removed from the eluate? Decomposition of the mineral by fusion for 1 hour with potassium hydroxide seemed a lengthy procedure; he had found that fusion with sodium peroxide was simpler and much more rapid. If the mineral were very finely ground i t could usually be decomposed by sintering with sodium peroxide a t about 450” C; this could be done in a nickel or even a platinum crucible without fear of contaminating the analysis with the metal of the crucible. MR. WILLIAMS said, in reply, that scandium was not extracted in the chromatographic procedure that had been described, but it was extracted after the complete elution of thorium by the use of more solvent.As the method for thorium could be applied in the presence of 0.5 g of zirconia, there was no reason why the procedure should not be applicable to the direct extraction of traces of thorium from a zircon mineral. Solvents other than ether had been used for uranium and i t followed that a number of higher-boiling solvents might be used for thorium. One marked advantage of ethyl ether was that i t had a low boiling- point and was not easily decomposed, whereas a solvent such as ethyl acetate, which had been used in uranium determinations, was not so satisfactory. ’ Sodium peroxide had been widely used a t the Chemical Research Laboratory for decomposing minerals and ores. MR. A. A. SMALES asked whether the authors had tried using Thoronol (l-(o-arsonophenylazo)-2- naphthol-3 : 6-disulphonic acid) for the determination of small amounts of thorium. MR. WILLIAMS replied that he had not used the colorimetric method for thorium as the spectrographic procedure had been found very satisfactory, but he intended to examine the method. MR. W. H. BENNETT asked if Mr. Williams could say .whether certain samples of zircon, which might contain small amounts of thorium, could be treated by his procedure. Did borates or tungsten interfere ? MR. WILLIAMS replied that there appeared to be no rewon why the procedure for thorium should not be entirely satisfactory for zircon minerals. Tungsten should not cause any difficulty and, although boron might be extracted by the ether solvent, i t should not interfere in the final precipitation procedure for thorium. This procedure, described by Rafter, was in general use.

ISSN:0003-2654    

DOI:10.1039/AN9527700307

出版商:RSC    

年代:1952

数据来源: RSC