ISSN:0003-2654    

DOI:10.1039/AN95782FX027

出版商:RSC    

年代:1957

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95782BX029

出版商:RSC    

年代:1957

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95782FP087

出版商:RSC    

年代:1957

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95782BP095

出版商:RSC    

年代:1957

数据来源: RSC

摘要:

JULY, 1957 THE ANALYST VOI. 82. NO. 976 PROCEEDINGS OF THE SOCIETY FOR ANALYTICAL CHEMISTRY DEATHS The Rt. Hon. Lord Clinton (Honorary Member) Gerald Roche Lynch. NORTH OF ENGLAND SECTION THE Twentieth Summer Meeting of the Section was held at Llandudno from Friday, May 17th, to Monday, May 20th, 1957. The Chairman of the Section, Mr. A. N. Leather, BSc., F.R.I.C., presided over an Ordinary Meeting at 10.15 a.m. on May 18th, 1957, at which V. L. S. Charley, BSc., Ph.D., gave a talk entitled “Some Chemical Features of the Composition of Fruit Juices of Interest to Analysts and their Ladies.” This provoked an animated and prolonged discussion. On the Sunday afternoon the party was taken on a coach tour of the coast and hinterland between Llandudno and Caernarvon, and at the latter place tea was taken and the Castle was visited.MIDLANDS SECTION AND MICROCHEMISTRY GROUP A JOINT Meeting of the Midlands Section and the Microchemistry Group was held at 7 p.m. on Friday, May loth, 1957, in the Mason Theatre, The University, Edmund Street, Birming- ham, 3. The Chair was taken by the Chairman of the Midlands Section, Dr. R. Belcher, F.Inst.F., F.R.I.C., who introduced the Chairman of the Microchemistry Group, Mr. D. F. Phillips, F.R.I.C. A discussion on “The Micro-determination of Functional Groups” was opened as follows : “Some Developments in the Analysis of Functional Groups,” by W. I. Stephen, Ph.D., A.R.I.C.; “The Determination of N-Methyl Groups,” by M. K. Bhatty, M.Sc., A.R.I.C.; “The Determination of Equivalents,” by T. S. West, Ph.D., A.R.I.C.; “Titrations in Non- aqueous Media on the Sub-micro Scale,” by T.S. West, Ph.D., A.R.I.C. (see summaries below). Before the meeting, at 2.30 p.m., a visit was made to Little Bromwich Hospital, Yardley Green Road, Birmingham, 9, to see some of the laboratory work being carried out in the Metabolic Ward, the Regional Virus Laboratory and the Pathological Laboratory. WE record with deep regret the deaths of SOME DEVELOPMENTS IN THE ANALYSIS OF FUNCTIONAL GROUPS DR. W. I. STEPHEN gave a survey of recent literature on the determination of organic functional groups. A great deal of work had been reported in this branch of quantitative organic analysis, but only those developments that had improved the methods for the more important organic groups were discussed.One of the most important determinations was that of the alkoxyl group, and recent work had been aimed at improving the apparatus, increasing the accuracy and extending the scope of the method. The main difficulty in the past had been the wash liquid used to free the alkyl iodide liberated by the classical Zeisel procedure from iodine, hydriodic acid and possibly hydrogen sulphide. Cadmium sulphate or antimony potassium tartrate solutions appeared to be the best wash liquids. Numerous modifica- tions to apparatus and procedure had been described. Noteworthy were the contributions 457PR0CEE:DINGS [Vol. 82 458 of Kirsten and Erlich-Rogozinsky (Mikrochim. Acta, 1955, 786) and of Gran (Svensk Papperstidning, 1952, 55, 255, 287; 1953, 56, 179, 202; 1954, 57, 702).The latter had perfected a suitable procedure for the simultaneous determination of methoxyl and ethoxyl groups. Franzen and co-workers had made an extensive study of the various factors involved in this deter- mination and recommended simplified apparatus (Mikrochim. Acta, 1955, 845). Acetyl and C-methyl groups were del ermined by hydrolysis and oxidation, respec- tively, by using Wiesenberger's apparatus, recently standardised in B.S. 1428 : Part C1 : 1954. Wiesenberger had further modified his apparatus to deal with volatile acidic substances formed during the hydrolysis of certain acetylated carbohydrates of high molecular weight. The Zerewitinoff method as used by Soltys was widely applied in the determination of active hydrogen. Recent work had tended to favour the use of lithium aluminium hydride in place of the conventional Grignard reagent.Recently, Stevens (Anal. Chem., 1956,28, 1184) had shown that a solution of methyl magnesium chloride in tetraethylene- glycol dimethyl ether functioned satisfactorily as a general reagent for active-hydrogen determinations. The work of Kainz (Mikrochim. Acta, 1953, 349) on the determination of the primary amino group had resulted in a great simplification of the established van Slyke procedure. The new apparatus was simple to use and gave accurate results. Finally, the methoxymercuric acetate methods of hiarquardt and Luce and of Martin for the determination of unsaturation in organic compounds had been improved by Das (Afial. Chem., 1954, 26, 1086). E,xcess of mercuric acetate was determined by non-aqueous titration.The method was by no means universally applicable, but a number of important ethylenic substances, including esters such as vinyl and ally1 acetates, could be determined with accuracy. THE DETERMINATION OF N-METHYL GROUPS MR. M. K. BHATTY said that the Ilerzig-Meyer method for methylimino and methylimino and methoxyl groups had been adapted to the 50-pg scale of analysis. An apparatus consisting of a double-distillsition arrangement for hydriodic acid had been modified for the ultra-micro determinations. The substance was allowed to react with hydriodic acid, the excess of which was then distilled over and kept boiling in a second flask under a reflux condenser. The residual quaternary ammonium iodide was decom- posed at 300" to 360" C in presence of ammonium iodide and gold chloride.The decom- position of the salt was best achieved by filling the flask initially with glass beads. The resulting methyl iodide was absorbed in a solution of bromine in sodium acetate - glacial acetic acid mixture. The iodine in the solution was determined by titration with 0.01 N sodium thiosulphate. Three distillations and decompositiorls for the substance under test and two for a blank gave a net titre representing the arnount of N-methyl group. When methylimino and methoxyl groups were present together, the methoxyl was determined first by digestion with hydriodic acid for 18 hours. The methylimino group was then determined, as outlined above, after the distillation of the acid and decom- position of the salt.Emphasis had been laid on precautions for minimising the blanks. High blanks due to aerosol formation had been eliminated by passing the hot gas stream through the boiling acid in the second flask. Methyl iodide vapours were purified in an efficient condensation - scrubbing system. Heating and sweeping of the apparatus had been regulated. Determination of tertiary nitrogen-An ion-exchange method had been devised on the ultra-micro scale for the determination of tertiary nitrogen. About 50 pg of the substance were treated with methyl iodide, the excess of which was afterwards removed by evapora- tion. The methiodide salt was passed down an anion-exchange column of Amberlite IRA-400(0H) resin, which produced the corresponding quaternary ammonium hydroxide in the effluent.The base was titrated acidmetrically, 0.01 per cent. alcoholic solutions of methyl red and methylene blue being used as indicators. The limitations of the method were enumerated. Closely related to the alkoxyl group was the alkylimino group.July, 19571 PROCEEDINGS 459 THE DETERMINATIOS OF EQUIVALENTS DR. T. S. WEST said that ultra-micro samples were weighed by difference on an ultra-micro balance and dissolved in hemispherical-ended borosilicate-glass tubes of 1 cm diameter. Each sample was dissolved in 200 pl of ethanol and treated with an excess of standard aqueous sodium hydroxide. The latter was titrated with standard aqueous benzoic acid to a phenolphthalein end-point by using a micrometer-syringe burette. Standard illumination from a “daylight” lamp was used throughout the titrations.Identical tubes were placed on either side of the titration tube and were filled with the same volume of indicator, alcohol and water to act as comparison standards for the detection of the end-point. The titrations were carried out with protection from atmospheric carbon dioxide. Results by this method were as accurate as those obtained on the micro scale, provided the volume of ethanol present was controlled within certain limits. TITRATIONS I N NON-AQUEOUS MEDIA ON THE SUB-MICRO SCALE DR. T. S. WEST said that many substances that were too weakly acidic or basic to be titrated in aqueous solution could be satisfactorily determined in anhydrous organic solvents such as benzene, butylamine and glacial acetic acid.Generally speaking, protogenic solvents enhanced basicity and protophyllic solvents encouraged acidity. Ionisation was not, however, a necessary process, and many aprotic solvents had much to offer because of their excellent solvent powers on organic molecules and because they minimised the solvolysis of reaction products and so on. Ultra-micro titrations of many substances could be carried out in non-aqueous solvents. For example, the alkali-metal salts of carboxylic acids could be titrated as strong bases in acetic acid by dissolving the sample in an excess of standard perchloric acid in glacial acetic acid and titrating the excess with standard sodium acetate in glacial acetic acid, with crystal violet as indicator and with a comparison system (4 1 p1 of titrant contained in each tube) as described previously. Quaternary ammonium halides could be determined similarly by application of the mercuric acetate reaction.Visual end-points in non-aqueous solvents were somewhat unsatisfactory and more precise results could be obtained by application of potentiometric or radio-frequency end-point techniques. BIOLOGICAL METHODS GROUP THE Summer Meeting of the Group was held on Thursday, May 23rd, 1957, when, by kind permission of the Director, 30 members paid a visit to the Wellcome Research Laboratories, Langley Court, Beckenham, Kent. In the morning, visits were made to the Pharmacological and Chemical Research Laboratories, after which the Group was entertained to lunch at the Eden Park Hotel. In the afternoon demonstrations were given by the Bacteriological and Immunological Depart- ments and visits were made to the Stables and to the Serum Concentration Laboratory. At the close of the meeting the Chairman of the Group, Dr. S. K. Kon, F.R.I.C., expressed thanks on behalf of the visitors. WESTERN SECTION AN Ordinary Meeting of the Section was held on Saturday, November loth, 1956, taking the form of a visit to the factory of Messrs. H. W. Carter and Co. Ltd., Coleford, Glos. During the morning, the party was conducted round the factory and laboratories, and in the afternoon the following papers were presented and discussed : “The Application of the Polarised Platinum Electrode to the Determination of Ascorbic Acid in Fruit Products,” by R. C. Curtis, B.Sc. ; “The Effect of Ripening on Certain Constituents of the Blackcurrant,” by Miss A. D. Ayres, BSc., R. C. Curtis, BSc., M. J. Egerton, BSc., A.R.I.C., and H. Fore, BSc., Ph.D., F.R.I.C.

ISSN:0003-2654    

DOI:10.1039/AN9578200457

出版商:RSC    

年代:1957

数据来源: RSC

摘要:

460 OBITUARY [Vol. 82 0 bituary LESLIE HERBERT LAMPITT died on June 3rd at the age of 69; by his death chemistry, particularly that branch dealing with food, lost one of its outstanding men and personalities. This was clearly demonstrated by the large number of people who attended the funeral at the church of St. Mary, Harrow-on-the-Hill. There were representatives from all the main chemical societies in this country, and Professor Stoll, President of the International Union of Pure and Applied Chemistry, came to England especially for the service, as did others, including Professor R. Delaby. All wished to pay their last respects to an outstanding man. Lampitt was educated at the old Birmingham Technical School, and then entered Birmingham University as a Priestley Scholar in 1906.He had a brilliant career as a student and obtained first-class honours in chemistry and biochemistry. One of his first positions (1911) was that of chemist to the large Belgian flour mill, the Meunerie Bruxelloise. He fought in the 1914-18 war, during which time he first met Mr. S. M. Gluckstein of J. Lyons & Co. Ltd. His life’s work proper began when he started and organised, after the 1914-18 war, the now world-famous laboratories of the firm of J. Lyons & Co. Ltd. There are many first-class chemists, but alas, only few with the vision of Lampitt and with the enthusiasm and the organising and business ability to make the vision come true. The best monument to him must still remain the magnificent, well equipped and efficiently run laboratories at Cadby Hall.It was no simple task to do this, but L,%mpitt was as able a business man as he was a chemist. He had uncanny judgment and collected around him able lieutenants whom he could inspire. It is not surprising, therefore, that before long he was elevated to the Board. Lampitt realised the importance of organised chemistry and, as far as this country was concerned, although he helped all and sundry, his main devotion was to the Society of Chemical Industry. He was President in 194648, Honorary Foreign Secretary in 1935-36 and 1948-57, Honorary Treasurer in 193646, Chairman (and Founder) of the Food Group, 1931-36, and Jubilee Memorial Lecturer for 1933-34. He became the Society’s Medalist in 1943. However, he was active in many other chemical and scientific organisations. He was a Member of Council of the Royal Institute of Chemistry in 1928-31, 1935-38, 193942, 1949-50 and a Vice-president in 194245.In our own Society he was a Member of Council in 1925-26, 1933-34 and 193940, as well as being a Vice-president in 1935-36 and 194142. He turned his attention to international chemistry and was one of the leading lights and a driving force of the International Union of Pure and Applied Chemistry, of which he was the Honorary Treasurer. He was often the Chairman of Executive Committees for organising International Congresses, and this was a job he could do so much better than anyone else. He watched carefully to see that the interests, dignity and importance of the United Kingdom were looked after. He never lost close contact with his main interest, the chemistry of food, and did as much as any man to ensure high standards, with particular attention to hygiene.He was able to impress upon food manufacturers that they undertook great responsibilities in providing food and had to maintain high standards, only possible by proper attention to the use of scientific methods. In his own laboratory he was an absolute autocrat, but a benevolent one. The time he gave to outside activities, only some of which have been mentioned, was enormous. Many came to him for advice and he had time for all, and great understanding. Like all great personalities he had his critics, but all were agreed as to his ability and honesty of purpose. He knew good food and wine as do few men, and indeed was a great diner-out. One side of him not always realised was thai he was a devout Christian, worked for the Church and was always doing quietly some kind and thoughtful act that might have surprised those who only knew him as a public man. He lived at Harrow-on-the-Hill, and his hobby was his wonderful garden. He was married in 1915 to Edith Potter Potts, who survives him, and who is so well known and respected in all chemical circles, while his son is also a frequenter of chemical social functions. We shall all miss Leslie Herbert Lampitt (ind it is hard to see how he can be replaced- certainly no one man can do it. LESLIE HERBERT LAMPITT It is difficult to write of him briefly. D. W. KEKT- JONES

ISSN:0003-2654    

DOI:10.1039/AN9578200460

出版商:RSC    

年代:1957

数据来源: RSC

摘要:

July, 19571 DREW AND KING 46 1 The Determination of Atmospheric Mercury Trapped in Permanganate Solutions : A Modified Method BY R. G. DREW AND E. KING The use of hydrogen peroxide as a reducing agent for acid potassium permanganate solutions before the determination of trapped mercury is described. The degree of accuracy of a modified extractive titration method under various interfering conditions is indicated. THE problems involved in measuring atmospheric mercury fall into two main categories, associated with the method of sampling and the method of analysing the samples so obtained. The principle of sampling suggested by Kuziatina,l involving the use of bubblers containing acid potassium permanganate as the air-scrubbing solution, has been adopted. Since dithizone, the reagent used in this work for the determination of the mercury, is sensitive to oxidation, it is necessary to reduce the permanganate before using this reagent.This paper reports the use of hydrogen peroxide as a reducing agent and the suitability of the chemical determination under conditions of contamination and interference likely to be encountered under actual sampling conditions. PRIXCIPLE OF THE METHOD- As in the method described by Kuziatina, a known volume of the atmosphere to be tested is drawn through a bubbler containing acid permanganate solution. The perman- ganate is decolorised by reduction with aqueous hydrogen peroxide. The mercury in the sample is then determined by extractive titration with dithizone solution. In the method described below, the chief innovation is the use of hydrogen peroxide as a reducing agent to prepare the contents of the bubbler for determination.This reagent gives more reproducible results under varying conditions than either oxalic acid or hydroxyl- amine, which have been used hitherto. Our whole experience during this work was that dithizone is remarkably stable towards hydrogen peroxide at the dilutions used and under the conditions of the method. Extractive-titration methods have been employed in many instances for the separation and determination of heavy metals with dithizone solution. Eckert2 and Pvlaren3 used the method for the determination of mercury. Strafford, Wyatt and Kershaw4 recommended the extractive titration of zinc when present in large amounts. Milton and Hoskins5 revived the method for the preferential extraction of mercury in the presence of copper, and BuckelP adopted the method for the determination of mercury in the atmosphere.METHOD APPARATUS- The bubblers used in this work are of the glass-bead type, as described by Milton and Duffield,' but without the attached filter-paper head. Before use, soak each bubbler in chromic acid cleaning solution. Rinse the bubbler, first with tap water and then with water redistilled in all-glass apparatus. Carry out a blank by filling the bubbler completely with 0.05 N potassium permanganate solution and setting it aside for 1 hour, after which the contents should be titrated for mercury as described below. Not more than 0.5 ml of dithizone solution should be required.In this work some 15 to 20 ml of potassium permanganate solution are necessary to fill the bubblers adequately and allow a reasonable sampling rate without undue splashing. In view of this, a fixed amount (20 ml) of potassium permanganate solution was taken for most of the interference experiments. All glass apparatus, including the bubblers, should be made from Pyrex glass. All apparatus should be thoroughly cleaned in chromic acid solution. REAGENTS- metal still, all water should be redistilled in Pyrex-glass apparatus. All reagents should be of recognised analytical grade. After normal distillation in a462 DREW AND KING: THE DETERMIhATION OF ATMOSPHERIC MERCURY [VOl. 82 Potassium fiermanganate solution, 0.05 &-Prepare this solution in 5 per cent. v/v It should be discarded sulphuric acid (1 litre contains 1.58 g of potassium permanganate).if drastic ageing reduces the permanganate concentration owing to decomposition. Hydrogen peroxide, 20-vohme-This should be stored in the dark at about 4' C. Chloroform, redistilled-Redistil analytical-reagent grade chloroform in all-Pyrex-glass apparatus. Dithizone stock solution-Dissolve 0.05 g of diphenylthiocarbazone (dithizone) in 100 ml of redistilled chloroform. This solution should be stored in the dark at about 4' C, and it should not be used if it shows any signs of deterioration, i.e., an increase in the volume of the solution equivalent to a known amount of mercury. The solid dithizone used should be up to a reliable standard of purity and, once decided upon, the source of supply should not be changed.Dithizone extraction solution-Dilute the $stock solution 1 in 100 with redistilled chloro- form. Prepare this solution freshly each day and, when it is not in use, keep it in the dark. Standard mercury solution A-Dissolve 0.06767 g of pure mercuric chloride in water and dilute to 100ml. Standard mercury solution B-Dilute standard mercury solution A to 1 in 100 with water. This solution should be freshly prepared as required. 1 ml = 5 pg of mercury. PROCEDURE- The method is essentially an extractive titration in which the technique has been rigorously standardised. Wash the contents of the bubbler into a Eeaker with redistilled water until the washings are only faintly pink, and then rinse the bubbler with 10 ml of hydrogen peroxide and finally with redistilled water.The hydrogen peroxide removes any brown deposit in the bubbler that has not been affected by washing with water. If much dilution with water has occurred, complete reduction of the permanganate may be slow, but the reaction can be accelerated by adding a few millilitres of 50 per cent. v/v sulphuric acid. After complete reduction, quantitatively transfer the solution or an aliquot to a separating funnel and dilute to a standard volume with distilled water. As the manner and time of shaking and settling out in the separating funnel must be carefully standardised, the contents of the separating funnel must be diluted to an approximately constant volume before the titration; we dilute to about 250ml in a 500-ml separating funnel.At this point check the acidity of the diluted sample to ensure that its pH is 1 or less, adding more 50 per cent. v/v sulphuric acid if necessary. The use of universal indicator paper is sufficiently accurate to check the acidity. Saturate the aqueous layer with chloroform by vigorously shaking it with about 5 ml of redistilled chloroform. Then add measured portions of the dithizone extraction solution from a burette. After each addition, mix the two phases intimately by shaking the separating funnel 2 or 3 times per second for 20 seconds. After each shaking, set the funnel aside for 30 seconds to allow settling out of the lower chloroform layer. lin the presence of mercury this layer will be orange, owing to the conversion of the blue-green dithizone to the orange-red mercury dithizonate, which is also soluble in chloroform.Run off the bulk of the chloroform and discard it, and swirl and shake the contents of the funnel to dislodge drops of chloroform suspended from the top aqueous layer. Run off this residue also. Add the next portion of dithizone solution and repeat the shaking procedure as before. By this means mercury is gradually extracted from the aqueous layer. The end-point of the titration is reached when a final addition of 0.5 ml of dithizone solution, after shaking, does not turn orange-red, but has a grey-green appearance. The judgment and selection of this end-point is a matter of experience, but it has been found that it is better, in the presence of other metals, not to titrate to a full green, but to a green- tinted grey colour.The end-point shade of green decided on by the operator must be adhered to rigorously. In order to make the procedure quantitative, it is necessary to reduce the volumes of dithizone solution added as the end-point is approached. To do this, an approximate estimate of the mercury content of the sample is first made by titrating a small aliquot In practice, further addition of acid is rarely necessary. After settling out, run off the lower chloroform layer and discard it.July, 19571 TRAPPED IX PERMAKGASATE SOLUTIONS : A MODIFIED METHOD 463 (say one-tenth) of the sample, 2 to 3-ml portions of the dithizone solution being added. From this preliminary titration, an approximate expected titration value for a larger aliquot can be determined.This must be done by multiplying the approximate mercury value determined from the calibration graph by the factor for the larger aliquot to be taken. The expected mercury content for the larger aliquot is then converted, from the graph, to an expected approximate titration volume. In carrying out the accurate titration, single additions of dithizone solution should not exceed 5 ml, and a few millilitres before the expected end-point the additions should be in 1-ml or 06-ml amounts, according to the degree of accuracy required. PROCEDURE FOR PREPARING A CALIBRATION GRAPH- Construct the calibration graph by making up a series of samples, each containing approximately 20 ml of permanganate solution plus various amounts of mercury, added as standard mercury solution B, through a range of 0 to 60 pg of mercury.Decolorise the samples by adding 10ml of hydrogen peroxide and titrate. Plot the volume of dithizone solution in millilitres against the amount of mercury in micrograms. The graph is slightly curved, as shown by Figs. 1 and 2. Each time the calibration curve is used it should be checked at one or two points, as any variation in conditions, technique or solutions is likely to be shown up by such a procedure. Amount of mercury, pg Fig. 1. Calibration graph prepared by using mercuric chloride solution x Metallic mercury dissolved in dilute nitric acid I 1 0 10 20 30 40 50 60 Amount of mercury,pg Fig. 2. Comparison of standard solutions for the preparation of a calibration graph NOTES ON THE METHOD- We have found that mercury dithizonate is sensitive to natural light, especially when in contact with reduced permanganate solutions (see Barness and Irving, Risdon and Andrewg), and variations in the intensity of natural light have caused difficulties in this laboratory.We therefore consider it essential to carry out the titrations in a “dark-room’’464 DREW AND KING: THE DETERMINATION OF ATMOSPHERIC MERCURY [VOl. 82 illuminated by electric-light bulbs. Fluorescent-tube lighting is not so satisfactory for assessing the end-point colours. In view of evidence of losses of mercury from reduced permanganate solutions on standing, it is recommended that all samples should be stored in the permanganate form and not reduced until immediately before titration (see p.466). The temperature at which samples are stored and analysed should be as close as possible to 20” C, as higher temperatures sometimes cause loss of mercury. CHECKS ON THE METHOD Owing to the coarseness of the titration technique, the accuracy of the method has a definite limit. However, results were obtained on a simple solution of mercuric chloride in reduced potassium permanganate solution to a precision of & 0.5 ml of dithizone solution, which is approximately equivalent to 1 pg of mercury. In order to apply this analysis technique to samples taken from the atmosphere in factories, etc., it was necessary to find out how the chemical accuracy varied with other factors. The main procedure was always as follows. Twenty millilitres of permanganate solution were taken and a known amount of standard mercury solution was added.Contaminating material was then added or a possibly interfering factor was allowed to operate, and then the hydrogen peroxide was added. The sample was titrated as described under “Procedure.” Small variations in this procedure are mentioned as they occur in the separate seci.ions that follow. These checks were carried out by five different workers, two of whom had not participated in the original development work on the method. INTERFERENCE FROM OTHER METALS- Sandelllo has stated that dithizone reacts preferentially with copper, mercury, silver and palladium when these metals are in aqueous solution at a pH of 1 or less, i.e., in 0.1 to 0.5 N acid. Milton and Hoskins5 have shown that, if the pH of the solution is kept within the limits 0.075 to 0.15 N with respect to hydrochloric acid, then mercury will be extracted by a chloroform solution of dithizone in preference to copper.In general, all other heavy metals do not normally react with dithizone under these conditions. In view of these facts, it has been specified in this method that the sample shall have a pH of 1 or less, to reduce interference from other metals. Variations of acidity towards the more acid side in practice did not cause great interference. TABLE. I EFFECT OF CONTAMINATING METALS Range of amounts of contaminant 20 to 35 mg of Cu 16 to 30 mg of FeII and FeIII 32 to 200 mg of Zn 20 to 50 mg of Pb 8 to 64 mg of CrvI 8 to 64 mg of A1 8 to 64 mg of Cd 50 mg of Mg 50 mg of Ca Xumber of recovery experiments 7 16 13 I 4 4 7 3 3 Range of absolute errors for mercury contents of 5 to 6Cl pg, Pg 0 to 4- 2 0 to 4- 7 0 to 4- 3 0 to 4- 2 0 to 4- 1 (20 pg of Hg only) 0 (20 pg of Flg only) 0 to $- 1 0 t o $ - 2 O t o $ - l Recovery for 5 p g of mercury (single value or average), % 140 130 140 130 - 120 140 120 Recovery for contents within the range 10 to 60 pg of mercury (average), 103 113 103 101 101 100 101 102 100 % In spite of the precautions described above, it was nevertheless thought desirable for two reasons to check the effect of other metals.The chemical system involves the use of dithizone in a solution containing an excess of hydrogen peroxide. This chemical environ- ment might have altered the reactivity of other metals towards dithizone. Also, the amount of contaminating metals such as copper and iron encountered in some factory atmospheresJuly, 19571 TRAPPED IS PERMANCANATE SOLUTIONS : A MODIFIED METHOD 465 is likely to be very large, e.g., of the order of milligrams in the presence of micrograms of mercury.Standard solutions of various metal salts were made up in water or acid and known amounts were added to samples containing known amounts of mercury. The results are shown in Table I, and they indicate that when a small amount of mercury, i.e., 5 pg or below, is being determined, there is considerable interference from other metals in terms of per- centage recoveries, although the absolute deviations are rarely more than a few micrograms. When mercury is being determined in the 10 to 60-pg range, the percentage recoveries are considered acceptable, except when iron is present.The iron probably causes gradual oxidation of the dithizone, as it is well known that in basic solutions this oxidation easily occurs.11 INTERFERENCE FROM MIXTURES OF METALS- first series had known mercury contents and the results were as follows- Mixtures of standard solutions were made up in a series of increasing complexity. The Sumber of determinations . . . . = 9 Range of amounts of metal added . . = 10 t o 20 mg of Fe; 5 to 50 mg of Cu; 30 to 50 mg of Zn; 30 to 50 mg of Pb; 50 mg of Cd ; 50 mgof Ca; 25 mg of CrvI Range of mercury recoveries as a per- centage of the mercury added . . = 100 to 115 per cent. Average mercury recovery , , . . = 111 per cent. Range of absolute errors .. . . = 0 to $ 3 p g of mercury The second series again contained mixtures of metals, but the mercury content was Consequently, preliminary aliquots had to be taken. The results unknown to the operator. were as follows- Number of determinations . . . . = 9 Contaminating materials in each sample = 20 mg of Fe; 30 mg of Cu; 50 mg of each Al, Zn, Mg, Cd, CrvI and Ca Range of mercury content . . . . = 7 to 57 pg Range of recoveries as a percentage of Average mercury recovery . . . . = 110percent. Range of absolute errors . . . . = - 1 pg to + 4 p g of mercury mercury added . . . . . . = 96 to 127 per cent. There were three anomalous values not included in the above results and these gave recoveries of 178, 172 and 136 per cent. They were found to be due to a worker titrating past the end-point.I t is essential, in the presence of much contaminating material, to titrate to the first perceptible shade of green that appears. This was done with a second group of unknown samples by the same operator, with the following results- Number of determinations .. .. .. .. .. = 3 Contaminating materials as in the previous group of experiments Range of mercury content .. . . * . . . . * = 22 to 48 pg Range of recoveries as a percentage of the mercury added Average mercury recovery * . . . . . * . . . = 108per cent. Range of absolute errors . . .. . . . . . . . . = + 1 pg to $ 4 p g of mercury It is realised that some of the metals tested here do not combine with dithizone. Never- theless, in view of their possible occurrence in samples from factory atmospheres, it did not seem advisable to assume that they could not interfere in any way and so this was checked.A mixture containing the following solid oxides in unknown proportions was prepared : ferric oxide, lead dioxide, lead monoxide, stannic oxide, manganese dioxide, chromic oxide and cupric oxide. Approximately 0.5 g of this mixture was added to each sample of mercury in permanganate and the determination procedure was carried out normally, except that each sample, after reduction, was extracted several times with pure chloroform, in order to clear away (at the chloroform - water interface) most of the undissolved solid. The results were as follows- . . = 103 to 118 per cent. INTERFEREKCE FROM SOLID OXIDES- Number of determinations ... . . . .. .. = 6 Range of mercury content .. . . . . .. .. = 5 to 6 0 p g Range of recoveries as a percentage of the mercury added Average mercury recovery .. .. .. .. . . = 106per cent. Range of absolute errors . . . . . . .. . I . . - - - 1 p g to $ 2 pg of mercury . . = 95 to 120 per cent.466 [Vol. 82 INTERFERENCE FROM MAKGANESE DIOXIDE- If the atmosphere being sampled contains small amounts of reducing substances, a partial decomposition of the permanganate trapping solution sometimes occurs to give a deposit of brown solid on the inside of the bubbler. This deposit probably contains manganese dioxide and therefore the effect of adding solid manganese dioxide to samples was determined. Each sample was made by adding about 1 6 g of commercial manganese dioxide to 20 ml of permanganate solution, known amounts of mercury then being added.The samples were decolorised in the cold, either with hydrogen peroxide, as in the standard method, or with hydroxylamine or oxalic acid, as used in earlier methods. With hydroxylamine sulphate 10ml of a 20 per cent. w/v solution (stripped of interfering metals by extraction with dithizone at about pH 1) were used per sample, .and with oxalic acid, 14 ml of a 10 per cent. w/v solution were used per sample. DREW AND KING: THE DETERMIXATION OF ATMOSPHERIC MERCCRY The results were as follows- Reducing agent Hydrogen peroxide Hydroxylamine Oxalic acid Mercury added, pg . . . . .. --z 20 ---z- 20 20 35 Difference in dithizone titration value from standard sample, ml , . . . - 0.6 0 - 3 -5 -3.6 -4.5 Although these experiments were carried out with more manganese dioxide than can be produced from one charge of trapping solution, it is apparent that the use of hydrogen peroxide removes any significant variation in the results.EFFECT OF ADDITIONAL TRAPPING SOLUTION AND HYDROGEN PEROXIDE- When bubblers are used for an extended period in dry atmospheric conditions, evapora- tion sometimes necessitates addition of a further charge of trapping solution. Although, therefore, the normal volume of solution used is about 20 ml, checks were carried out for possible interference from larger volumes. The results were as follows- Amount of mercury added. . . . . . .. . . . . = 20 pg in all tests Range of volumes of permanganate solutions added . . .. = 20 to 200 ml Mercury recovered . . . . .. .. . . .. . . = 20 pg in all tests In a single experiment in which 100 ml of 20-volume hydrogen peroxide were added to a Sumber of determinations . . . . .. . . . . = 6 mercury sample the recovery was 100 per cent. EFFECT O F STORING SAMPLES- In view of the fact that atmospheric sampling is sometimes carried out in places distant from the analytical laboratory, when it is impossible immediately to analyse the samples, it was considered advisable to check the keeping qualities of a typical sample. There are two possible situations, one when. the samples are kept in the original form in the bubbler, and the other when the samples are reduced and kept in the colourless form. W i t h samples kept as the permafiganate form- Range of added mercury contents .. . . . . Range of times left standing .. . . .. Air temperature . . . . . . . . .. Range of mercury recoveries . . . . . . Range of absolute errors . . . . .. . . Range of added mercury contents . . . . . . Range of times left standing . . .. .. .4ir temperature . . . . . . . . . . Range of mercury recoveries . . . . . . Range of absolute errors . . . . . . .. (there was also one value of 80 per cent., repres out of 2.5 pg of mercury) W i t h samples kept in the reduced form- - .. - - .. - - .. - - .. - ienting - . . - - .. - - . . - - .. - - . . - - . . - 5 to 60 pg 50 minutes to 24 hours 200 to 220 c 100 to 103 per cent. a loss of about 0.5 pg - 4 pg to + 1 p g of mercury 20 to 60 pg 60 to 85 minutes 18' to 21' C 90 to 98 per cent.-2 pg to -2.5 pg of mercury The losses occurring on setting aside a sample with additional hydrogen peroxide Recoveries dropped to a minimum of 75 per cent. after the sample had been set aside for 2 hours. (60 ml of 20-volume solution) are higher. SAMPLING CHECK UNDER WORKSHOP CONDITIONS- In order to check the performance of this method under actual sampling conditions a series of atmosphere samples was taken in an M.R.C. workshop (metal work and woodJuly, 19571 467 work). With most samples a low artificial mercury content was introduced into the atmosphere by exposing dishes of liquid mercury to the air. An aliquot from each sample was analysed for trapped mercury and then a known amount of mercury (standard solution) was added to an identically sized aliquot and the determination was repeated.Hence recoveries under practical conditions could be assessed. Of the first group of twelve determinations carried out, there were three anomalous recoveries of 280, 220 and 150 per cent. The nine remaining results may be summarised as follows- TRAPPED IN PERMANGANATE SOLUTIONS : A MODIFIED METHOD Range of recoveries . . . . . . . . . . . . = 90 to 130 per cent. Average recovery . . . . . . .. . . . . = 120 per cent. Range of absolute errors . . . . . , . . . . = - 1 to + 3 pg of mercury Two of the anomalous values mentioned above were associated with the operation of “melting down” lead piping. Further atmosphere samples were taken during a similar operation and this time recoveries were 100 to 110 per cent. I t is thought that the first anomalies may have been due to silver derived from laboratory plumbing, although this could not be verified. We record here our appreciation of the technical assistance received in the above work from Miss A. Murdon, Mr. B. Biles, Mr. T. R. Emerson and Mr. E. Palmer. Acknowledgment is also made to Mr. H. Hardwick (Engineer-Maintenance) for co-operation in carrying out the sampling check in the M.R.C. Workshop. We finally thank Dr. P. L. Bidstrup for her continuing interest and assistance in the preparation of this paper. REFERENCE s 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. Kuziatina, S. S., Zavod. Lab., 1939, 8, 174. Eckert, H. W., I d . Eng. Chem., A ? d . Ed., 1943, 15, 40G. hlaren, T. H., J . Lab. Clin. Med., 1943, 28, 1511. Strafford, iY., Wyatt, P. F., and Kershaw, F. G., Analyst, 1945, 70, 232. Milton, R. F., and Hoskins, J. L., Ibid., 1947, 72, 6. Buckell, M., Brit. J. Ind. Med., 1951, 8, 181. Milton, R. F., and Duffield, W. D., Analyst, 1947, 72, 11. Barnes, H., Ibid., 1947, 72, 469. Irving, H., Risdon, E. J,, and Andrew, G., J. Chenz. SOC., 1949, 541. Sandell, E. B., “Colorimetric Determination of Traces of Metals,” Interscience Publishers Inc., - , op. cit., pp. 88 and 89. 1LR.C. LABORATORIES, HOLLY HILL New York, 1944, Volume 3, p. 80. DEPARTMENT FOR RESEARCH IN INDUSTRIAL MEDICINE First submitted, June 3rd, 1955 Amended, July 25tR, 1956 HAMPSTEAD, LONDON, N.W.3

ISSN:0003-2654    

DOI:10.1039/AN9578200461

出版商:RSC    

年代:1957

数据来源: RSC

摘要:

July, 19571 TRAPPED IN PERMANGANATE SOLUTIONS : A MODIFIED METHOD 467 Comparison of Flame-photometric and Chemical Methods for Determining Sodium and Potassium in Soil, Plant Material, Water and Serum BY M. PUFFELES AND N. E. NESSIM Flame-photometric and chemical methods of determining sodium and potassium in soil, plant material, water and serum are compared. The paper shows how the correction of the results by the chemical methods for the recovery error and of the results by flame-photometric methods for interference by calcium and phosphate considerably reduces the otherwise large discrepancies. TIME-CONSUMING and laborious determinations of sodium and potassium by chemical methods are rapidly being replaced by more convenient and economical flame-photometric methods. Procedures have been described for such determinations in industrial, agricultural and biological materia1.1~2~3~4~5~6~7~a~g Whereas the sources of error and their elimination have been investigated by many workers who used American flame photometers based on internal- standard techniques,lOV11 relatively little attention has been devoted to the simpler, single-cell468 PUFFELES AND NESSIM : COMPARISON OF FLAME-PHOTOMETRIC AND [VOl.82 flame photometers, such as the British-made, E.E.L. instrument (Evans Electroselenium Ltd.), which has lately come into use in many laboratories here and elsewhere.12 The work described is a study of sodium and potassium determinations in soil, plant extracts, natural waters and blood serum carried out with the E.E.L. flame photometer and by chemical methods, with a view to preparing flame-photometric correction curves for the two main interfering ions, calcium arid phosphate, and thereby to achieve better agreement of results.Physical sources of error such as (a) fluctuation of the spraying rate, (b) changes of air and gas pressures, (c) variation of flame temperature, (d) impurities affecting the viscosity and surface tension of the test solution, and (e) insufficiency of the filters, were not taken into consideration. These errors are discussed, among others, by and Robinson and 0ven~ton.l~ Halstead and Chaiken15 state that calcium interferes with the flame-photometric deter- mination of sodium and potassium. SpectorI6 reports a positive error due to calcium for 30 p.p.m. of sodium that increases as the concentration of calcium ions increases and amounts to 40 per cent. for 500 p.p.m.of calcium. The corresponding error effect on 50 p.p.m. of potassium is about 12 per cent. Both errors are appreciably high compared with the inter- ference of potassium in the determination of 10 p.p.m. of sodium (3 per cent. for 1000 p.p.m. of potassium) and of sodium on the determination of 10 p.p.m. of potassium (0 per cent. for 1000 p.p.m. of sodium). Of all the cations dealt with, calcium seems to be the main source of interference in the flame-photometric determination of both sodium and potassium. Collins and Polkinhorne,12 however, call attention to the large differences in interference caused by N solutions of various anions. Phosphate causes a 34 per cent.depression in the reading for 10 p.p.m. of sodium and 78 per cent. for 10 p.p.m. of potassium, these figures representing the greatest effect caused by anj7 anion investigated. EXPERIMENTAL The E.E.L. flame photometer used in this work has been described in detail by Collins and Polkinhorne.12 The main parts were illustrated and its operation was explained. Essentially it is a single-cell direct-reading instrument. A butane - air flame is used, primarily because it is expected to cause lower mutual interference effects between sodium and potassium themselves than would be experienced with, for example, an oxygen - acetylene flame. Amount of calcium, D.D.m. 0 2 4 6 8 10 Amount of metal, p.p.rn. , , Fig. 2. Effect of calcium on the determination of 10p.p.m.of sodium Fig. 1. Calibration curves: curve A, sodium; curve B, potassium AnalaR reagents and demineralised wateir (containing not more than 0.5 p.p.m. of total All determinations were carried salts expressed as sodium chloride) were used throughout.July, 19571 CHEMICAL METHODS FOR DETERMINING SODIUM AND POTASSIUM 469 out in duplicate. To ascertain the reproducibility of the results, flame-photometric readings were first made serially and then repeated, the scale adjustment being re-checked against standard solutions during the repeat tests. Fig. 1 shows the calibration curves for sodium and potassium prepared by spraying standard solutions of their chlorides. EFFECT OF CALCIUM ON SODIUM- TABLE I ERROR I N DETERMINING 10 p.p.m. OF SODIUM IN THE PRESENCE OF CALCIUM BY FLAME PHOTOMETER Calcium present, p.p.m.50 100 200 300 400 500 Apparent sodium exhibited by sprayed solution of Sodium found, Error, sodium-free calcium, p.p.m. % p.p.m. 11.4 + 14 0.6 12.2 T 22 1.3 13.1 +31 2.6 13.9 + 39 3.6 14.6 46 4.3 15.2 + 52 4.9 The E.E.L. flame photometer is adjusted to produce a scale reading of 0 for demineralised water and 100 (full-scale deflection) for 20 p.p.m. of sodium. A series of solutions containing 10 p.p.m. of sodium and various amounts of calcium as calcium chloride are then sprayed and the reading for sodium is recorded. Solutions of calcium chloride containing the same amounts of calcium as before but free from sodium are then sprayed and the apparent sodium reading is noted. I t can be seen from Table I that the error caused by the insufficiency of the sodium filter to eliminate the transmittance of the neighbouring calcium bands does not appear to be quite additive.The presence of calcium together with sodium seems to enhance the emission of the latter. Brealey and Rossx7 reported a constant negative error of 2.8 per cent. caused by 100 and 1000 p.p.m. of calcium when working with the first of their two flame photometers on 30 p.p.m. of sodium. This effect of calcium on sodium was never observed with our E.E.L. instrument, nor with the Langele or Beckman model DUls instruments. EFFECT OF CALCIUM ON POTASSIUM- With 400p.p.m. a positive interference error of only 0.4 per cent. was found for 10p.p.m. of potassium. This is significantly lower than that reported by Spector,lG who used the Lange single-cell flame photometer. This improved reduction of the calcium effect may be due partly to the higher selectivity of the E.E.L.potassium filter and to the lower temperature butane - air flame used, as acetylene, chosen by Spector, burns at a much higher temperature, thereby increasing the mutual interference effect. This was found to be negligibly small below 300 p.p.m. of calcium (see Table 11). TABLE I1 ERROR IN DETERMINING 10 p.p.m. OF POTASSIUM IK THE PRESENCE OF CALCIUM BY FLAME PHOTOMETER Calcium present, p.p.m. 100 200 300 400 500 600 Error, % + O . l l +0.21 +0*31 -+ 0.40 + 0.49 + 0.57 Calcium present, p.p.m. 700 1000 2000 4000 7000 10,000 Error, % t-0.65 t-0.90 + 1.65 +2.75 +3.55 + 4.20 EFFECTS OF PHOSPHATE ON SODIUM AND POTASSIUM- The effects were investigated by preparing two series of standard solutions containing 10 p.p.m.of sodium and 10 p.p.m. of potassium, respectively, and various amounts of AnalaR ammonium dihydrogen phosphate. The solutions were sprayed into the E.E.L. flame photo- meter, the instrument having been adjusted to give full-scale deflections with both 10 p.p.m.470 PUFFELES AND NESSIM: COMPARISON O F FLAME-PHOTOMETRIC AND [VOl. 82 of phosphate-free sodium and with potassium. The results are shown in Fig. 3 and confirm Collins and Polkinhorne's reportla that phosphate ions cause negative interference errors in the determination of both sodium and potassium, the effect on potassium being by far the greater. 0 2000 4000 6000 8000 10,000 Amount of phosphate, p.p.m.Fig. 3. Effect of phosphate ions on the determination of sodium and potassium: cLrve A, 10p.p.m. of potas- sium; curve B, 10p.p.m. of sodium DETERMIXATION OF SODIUM AND POTASSIUM IN SOIL EXTRACTS- Six samples of typical local soils were extracted with neutralised AT ammonium acetate solution. Sodium and potassium were then determined in the extracts (a) with the E.E.L. flame photometer and (b) by chemical methods, i.e., sodium gravimetrically as sodium magnesium uranyl acetatelS and potassium volumetrically as cobaltinitrite.20 The results are given in Tables I11 and IV. TABLE I11 COMPARISON OF THE RESULTS OF THE DETERMINATION OF SOOIUM IN SOIL SAMPLES BY CHEMICAL AND FLAME-PHOTOMETRIC METHODS Calcium in Type sprayed of soil solution, p.p.m. Light .. 41 Medium.. 155 Medium.. 56 Heavy . . 57 Heavy .. 150 Loess . , 60 Cor- rection for calcium, -11.0 - 27.5 - 15.0 -15.5 -27.0 - 14.0 % Phos- phate c in sprayed solution,* p.p.m. 30.2 17.6 15.4 11.1 8.5 2.0 Sodium found by :hemica1 method- +- uncor- cor- rected, rected, p.p.m. p.p.m. 2.46 2.52 7.03 7.20 4.59 4.70 17.36 17.74 11.98 12.25 28.80 29.45 Sodium found by flame-photometric method- - uncor- cor- rected, rected, p.p.m. p.p.m. 2.80 2.49 9.70 7.02 5.50 4.67 21.98 18.60 18.80 12.72 33.63 28.95 Deviation Deviation of uncor- of cor- rected rected results, results, Yo % +17.5 -1.2 $37.8 -2.5 +19*8 -1.5 1-26'6 +3.7 $12.9 -1.7 +17*2 $3.8 * No correction for phosphate was applied, as any interference caused by the concentrations found was negligible. DETERMINATION O F SODIUM AND POTASSIUM IN PLANT-ASH EXTRACTS- For the purpose of this investigation, ten samples of leaves from four types of local plants were chosen. About 5 g of each dried sample were ashed in silica crucibles at 550" to 600" C, and the ash was extracted with 2.5 rn.1 of concentrated hydrochloric acid and hot water.The extracts were then filtered, cooled and made up to 250 ml with demineralised water.21 Sodium and potassium were determined in the extracts both chemically and flame photometrically. The results are expressed as parts per million in dry matter and are given in Tables V and VI.July, 19571 CHEMICAL METHODS FOR DETERMINING SODIUM AND POTASSIUM TABLE IV 47 1 COMPARISON OF THE RESULTS OF THE DETERMINATION OF POTASSIUM IN SOIL SAMPLES BY CHEMICAL AND FLAME-PHOTOMETRIC METHODS Potassium found by Phos- Cor- Potassium found by flame-photometric Calcium phate rection chemical method- method- Deviation Deviation in in for <-*-, ,-*-, of uncor- of cor- Type sprayed sprayed phos- uncor- cor- uncor- cor- rected rected of soil solution,* solution, phate rected, rected, rected, rected, results, results, Light .. 41 30.2 $0.36 8.22 8.47 8.52 8.55 $3.6 +0*9 Medium.. 138 17.6 +0.21 12.50 12.89 12.99 13.02 i 3 . 9 +1.0 Medium.. 56 15.4 $0.18 18.40 18.95 18.73 18'76 $1.8 -1.0 Heavy .. 57 11.1 $0.13 40'61 41'90 41.25 41.30 +2.1 -1.4 Heavy . . 135 8.5 i-O.10 35.05 36.15 35.44 35.80 q 1 . 0 -1.8 Loess . . 50 2.0 $0.00 14.22 14.64 14.52 14'52 $2.1 -0.8 p.p.m. p.p.m, yo p.p,m. p.p.m. p.p.m. p.p.m. Y/, % * No correction for calcium was applied, as any interference caused by the concentrations found was negligible. TABLE V COMPARISOX OF THE RESULTS OF THE DETERMINATION OF SODIUM IN PLANT LEAVES BY CHEMICAL AKD FLAME-PHOTOMETRIC METHODS Sodium found by Phos- Sodium found by flame-photometric Calcium Cor- phate chemical method- method- Deviation Deviation in rection in (-*-, {--*-, of uncor- of cor- Type sprayed for sprayed uncor- cor- uncor- cor- rected rected of leaves solution, calcium, solution,* rected, rected, rected, rected, results, results, p.p.m.yo p.p.m. p.p.m. p.p.m. p.p.m. p.p.m. yo % Green Yellow eucalyptus 50 -14.0 17.7 2900 2970 3500 3010 +20.6 $1.3 eucalyptus 16 -4.5 5.2 9020 9240 9460 9050 $4.9 -2.1 Dried tobacco 330 -41.0 56.6 607 621 1020 602 +68*1 -3.1 Dried tobacco 451 -49.0 41.3 581 595 1140 582 +91.2 -2.1 Desert weed 3 - 0.8 1.6 43,500 44,500 33,700 33,400 -0.2 -2.0 Green citrus 135 -26.0 74.4 63 1 646 560 636 +36,2 -1.5 Dried tobacco 325 -40.5 52-4 735 752 1110 660 +51*0 $132 Green citrus 135 -26.0 74.4 659 675 940 695 +58.0 $3.0 Yellow citrus 143 -26.5 13.6 2062 2118 2905 2142 $40.8 4-1.0 Yellow citrus 150 -27.0 13.6 2050 2100 2835 2070 $27.7 -1.4 was negligible.* No correction for phosphate was applied, as any interference caused by the concentrations found TABLE VI COMPARISON OF THE RESULTS OF THE DETERMINATION OF POTASSIUM I N PLAST LEAVES BY CHEMICAL AND FLAME-PHOTOMETRIC METHODS Potassium found by Phos- Cor- Potassium found by flame-photometric Calcium phate rection chemical method- method- Deviation Deviation in in for (-A-, r-A----, of uncor- of cor- Types sprayed sprayed phos- uncor- cor- uncor- cor- rected rected of leaves solution,' solution, phate, rected, rected, rected, rected, results, results, p.p,m.p,p,m, yo p.p,m, p.p.m, p.p.m. p.p.m. O' /O Y O Green Yellow eucalyptus 50 17.7 $0.21 5820 GOO1 6020 6033 1 3 . 5 +0.5 eucalyptus 125 41.3 $0.50 2220 2320 2280 2291 $2.6 -0.1 Dried tobacco 42 7.1 $0.09 20,420 21,080 20,900 20,919 +2.3 -0.8 Dried tobacco 108 17.5 $0.21 10,010 10,334 10,300 10,321 $2.8 -0.1 Dried tobacco 150 13.5 +0.17 6650 6850 6940 6952 $4.4 $1.5 Desert weed 7 3.3 +0.04 22,740 23,444 23,200 23,209 $2'1 -1.0 Green citrus 14 7.4 $0.09 13,850 14,290 14,200 14,213 $2'5 -0.5 Green citrus 20 10.6 $0.13 8630 8900 9040 9052 t-5'0 $1.7 Yellow citrus 143 13.6 $0.17 6797 7048 7000 7012 t 2 .9 -0.6 Yellow citrus 150 13.6 +0.17 7150 7372 7430 7443 $3.9 $0.4 was negligible. * No correction for calcium was applied, as any interference caused by the concentrations found472 PUFFELES AND NESSIM : COMPARISON OF FLAME-PHOTOMETRIC AND [VOl. 82 DETERMINATION OF SODIUM AND POTASSIUM IN NATURAL WATERS- As before, sodium and potassium were determined in aliquots of the concentrated samples both by chemical methods and with the E.E.L. flame photometer. The results are shown in Tables VII and VIII. Six samples of deep-well and river waters were concentrated by evaporation. TABLE VII COMPARISON OF THE RESULTS OF THE DETERMINATION OF SODIUM IN NATURAL-WATER SAMPLES BY CHEMICAL AND FLAME-PHOTOMETRIC METHODS Sodium found by Phos- Sodium found by flame-photometric Calcium Cor- phate chemical method- method- Deviation Deviation in rection in r-J----, r-A-, of uncor- of cor- Source sprayed for sprayed uncor- cor- uncor- cor- rected rected of water water, calcium, water,* rected, rected, rected, rected, results, results, p.p.m.Yo p . p m p.p.m. p.p.m. p.p.m. p.p.m. Yo YO Bat-Yam . . 19.7 -4.3 0.11 21.1 22.3 23.1 22.1 $6.0 -1.1 Holon . . 13.7 -3.0 0.10 50.0 51-2 54.4 52.8 $8.8 $1.3 Yarkon . . 8.9 -2.0 0.14 123.8 126.8 128.8 126.2 $4.1 -0.6 Tel-Aviv . . 6.1 -1.3 0.02 230.0 235.5 236.0 232.9 $2.6 -1.1 Kfar-Ono . . 11.3 -2.5 0.13 26.2 26-8 27.6 26.9 1-5.3 +0.4 Tel-Gibborim 15.5 -3.4 0.10 46.0 47.1 48.0 46.4 +4*3 -1.5 * No correction for phosphate was applied, as any interference caused by the concentrations found was negligible.TABLE 'VIII COMPARISON OF THE RESULTS OF THE DETERMINATIOK OF POTASSIUM I N NATURAL-WATER SAMPLES BY CHEMICAL AND FLAME-PHOTOMETRIC METHODS Potassium found by Potassium Deviation Deviation Source in sprayed in sprayed r p A - - - , photometric uncorrected corrected Calcium Phosphate chemical method- found by flame- of of p.p.m. p.p.m. p.p.m. p.p.m. p.pm. Y O of water water,* water,* uncorrected, corrected, method, results, results, 1 4 . 5 Bat-Yam . . 59 0.33 1.50 1.55 1.62 + 8.0 Holon , . 55 0.40 2.58 2.66 2.62 +2.7 - 1.6 Yarkon .. 62 0.80 6.80 7.02 7.12 +4*7 + 1.4 Kfar-Ono . . 34 0.40 1.17 1.21 1.20 + 1.8 -0.8 Tel-Gibborim 62 0.40 0.92 0.95 0.95 + 3.3 0.0 Y O Tel-Aviv . . 122 0.50 6.15 6.35 6.27 + 2.0 - 1.3 * No correction has been applied to the flame-photometric results, as the interference due to calcium and phosphate was negligible.DETERMIXATION OF SODIUM AND POTASSIUM IN SERFM- Sodium was determined in six fresh samples of human blood sera both chemically by the potassium pyroantimonate method22 and flame photometrically by preparing and spraying a (1 + 499) dilution in demineralised water. Potassium was determined in the same hemolysis-free sera both chemically by the cobaltinitrite methodz2 and flame photometrically by preparing a (1 + 49) dilution in de- mineralised water. The results are shown in 'Table IX. TABLE I X COMPARISON OF THE RESULTS OF THE DETERMINATION OF SODIUM AND POTASSIUM IN SAMPLES OF BLOOD SERA BY CHEMICAL .AND FLAME-PHOTOMETRIC METHODS Calcium Phosphate Sodium Potassium in sprayed in sprayed Sodium found by Potassium found by diluted diluted found by flame- found by flame- (1 + 499) (1 + 49) chemical photometric chemical photometric serum,* serum,* method,T method, Deviation, method,? method, Deviation, 0.22 0.28 3160 3250 + 2.8 193 200 + 3.6 0.19 0.26 3120 3200 + 2.6 220 225 + 2.3 0.21 0.22 3200 3170 - 0.9 202 195 - 3.5 0.24 0.20 3400 3350 - 1.5 220 210 -4.5 0.17 0.23 3350 3450 + 3.0 170 165 - 2.9 0.20 0.28 3280 3220 - 1.9 172 180 + 4.6 p.p.m.p.p.m. p.p.m. p.p.m. /O p.p.m. p.p.m. % O f * No interference caused by the concentrations present. t Corrected for error of recovery.July, 19571 CHEMICAL METHODS FOR DETERMIKING SODIUM AND POTASSIUM 473 DISCUSSION OF RESULTS Very large discrepancies are observed when the results of chemical determinations are compared with those of flame-photometric determinations, especially for sodium, amounting in one extreme case to 91.2 per cent.(see Table V). The flame-photometric results are almost always higher than those by chemical methods. For sodium this seems to be due chiefly to two factors (a) the low recovery by chemical determinations [a series of ten recovery tests on standard solutions containing 2 to 5 mg of sodium (the usually encountered range) was carried out by the same chemical method and under similar conditions; the average for the well agreed results obtained was 97.6 I 0.3 per cent.] and (b) a relatively largepositive interference effect due to calcium, which appears to be caused by the insufficiency of the sodium filter to prevent completely the transmission of calcium bands (6030, 6240 and 6480 A), which are quite close to the sodium doublet (5890 and 5 8 9 6 ~ ) , and the enhancement of the sodium emission in presence of calcium. The negative error effect of phosphate ions is too weak to counteract the calcium effect, as the concentration of phosphate ions in all the soil, plant-material, water and serum samples investigated was found to be either below the limit required to cause interference or, at the most, negligibly small.The relatively high flame-photometric results for potassium are mainly due to the low recoveries in the chemical determinations. The average obtained for a series of ten recovery tests on standard solutions containing 1 to 5 mg of potassium was 96.9 i.0.4 per cent. The calcium effect on 10 p.p.m. of potassium is negligible up to 300 p.p.m., as their bands are widely separated, the potassium doublet being at 7660 and 7690 A. However, none of the samples analysed contained such a high concentration of calcium. The negative error effect of phosphate ions, although stronger for potassium than sodium, was negligibly small with the concentrations of phosphate ions encountered. To obtain better agreement between the results of the chemical and flame-photometric determinations, it was found necessary to apply a constant positive correction for the per- centage error of the recovery of sodium and potassium by the chemical methods, and, a negative correction for calcium and a positive correction for phosphate ion-concentrations of the sprayed solutions on which sodium and potassium were determined flame photo- metrically.For this purpose calcium and phosphate concentrations were determined and the corrections applied were based on Figs. 2 and 3. The phosphate-ion concentrations of the samples of soil, plant material, water and serum were below the level required to affect the determination of sodium. This is also true of the calcium effect on the determination of potassium in these samples and in all such cases no correction was applied. With blood serum neither the calcium nor phosphate-ion content warranted any significant interference correction for either sodium or potassium. The success of the applied corrections can be seen from the percentage deviation of the flame-photometric results from the chemical ones (see Tables I11 to IX).Most of the large discrepancies have been considerably reduced and the deviations kept within acceptable limits, considering the small quantities measured. It was found that, in general, the corrected results of the determinations of sodium and potassium were lower by the flame-photometric method than by the chemical methods. This is probably because of a negative error effect due to the presence of anions other than phosphate. The exceptions may indicate the presence of some interfering agents that have a positive error effect and that have not been considered. Small corrections for reagent blanks were applied in the calculation of the results of the chemical determinations. REFERENCES 1 .2. 3. 4. 5. 6. 7. 8. 9. 10. Brodericlr, E. J., and Zack, P. G., Anal. Chem., 1951, 23, 1455. Dickens, P., Angew. Chem., 1951, 63, 494. Edgcombe, L. J., and Hewett, D. R., Analyst, 1954, 79, 765. Cavell, A. J., Ibid., 1952, 77, 637. Richarde, L. A,, “Diagnosis and Improvement of Saline and Alkali Soils,” Agricultural Handbook KO. 60, U S . Dept. of Agriculture, Washington, 1954. Hald, P. M., J . Biol. Chew., 1947, 167, 499. Keirs, R. J , , and Speck, S. J., J . Dairy Sci., 1950, 33, 413. Overman, R. R., and Davis, S. K., J . BioZ. Chem., 1947, 168, 641. Stone, D., and Shapiro, S., Science, 1948, 108, 503. Eggertson, F. T., Wyld, G., and Lykken, L., “Symposium on Flame Photometry,” A S T M S@eciaZ Technical Publication No. 116, American Society for Testing Materials, Philadelphia, Pa., 1951, pp. 52 to 56.474 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. ALLSOPP: THE DETERMINATION OF EXCESS OF ZINC IK ZISC OXIDE [VOl. 82 Elliott, F. H., Canad. J . Technol., 1951, 29, 111. Collins, G. C., and Polkinhorne, H., Analyst, 1952, 77, 430. Fox, C. L., jun., Anal. Chem., 1951, 23, 137. Robinson, A. &I., and Ovenston, T. C. J., Anal:vst, 1951, 76, 416. Halstead, W. J., and Chaiken, B., Public Roads, 1950, 26, No. 5, 199, Spector, J., Anal. Chew., 1955, 27, 1452. Brealey, L., and Ross, R. E., Analyst, 1951, 76, 334. West, P. W., Folse, P., and Montgomery, D., Anal. Chem., 1950, 22, 667. Piper, C. S., “Soil and Plant Analysis,” Interscience Publishers Inc., Sew York, 1944, p. 1i6. “Official Methods of Analysis,” Seventh Edition, The Association of Official .4gricultural Chemists, U7ashington, D.C., 1950, p. 41. Leyton, L., Analyst, 1951, 76, 723. Rappaport, F., “Rapid Microchemical RIethods of Blood and CSF Examinations,” Grune & Stratton, Sew York, 1949, p. 105. LABORATORY FOR CHEMICAL a h A L Y S I S RfINISTRY OF AGRICULTURE TEL-AVIV, ISRAEL Soventbev 21st, 1956

ISSN:0003-2654    

DOI:10.1039/AN9578200467

出版商:RSC    

年代:1957

数据来源: RSC

摘要:

474 ALLSOPP: THE DETERMINATION OF EXCESS OF ZINC IS ZISC OXIDE [Vol. 82 The Determination of Excess of Zinc in Zinc Oxide BY H. J. ALLSOPP This paper describes a hydrogen-evolution technique whereby excess of zinc in zinc oxide is determined. Under vacuum, the sample to be analysed is dissolved in previously distilled hydrochloric acid. After solution, the acid is evaporated to ensure that all the evolved gas is released. The gases are then analysed for hydrogen content, from the amount of which the excess of metal is calculated. The method was standardised radiochemically and, although the values obtained showed only about a 70 per cent recovery, the method is believed to be absolute. The influence of grinding on the specimen has been studied. It is estimated that the method is sensitive t o a 0.1 p.p.m.excess of zinc in zinc oxide by weight. The zinc oxide specimens analysed so far have been found to contain an excess of zinc within the range 0.2 to 18 p.p.m. by weight. A KNOWLEDGE of the amount of excess of zinc lcontained in zinc oxide was desired during diffusion and sintering investigations.l12~s~4 In the first investigation, the diffusion of zinc in zinc oxide was being determined in atmospheres of differing oxygen content and it was important to find out whether the diffusion rates observed could be correlated with departurle of the oxide from stoicheiometry. In the second investigation, the sintering behaviour of zinc oxide powder had been found to depend on the amount of oxygen in the surrounding atmosphere, and here again it was important to try to establish whether the sintering could be correlated directly with changes in the composition of the powder.It was clear at the start that the excess of zinc to be determined was very small.6 Little attention has been paid so far to the measurement of small departures from stoicheiometry, so that, apart from the above-mentioned immediate requirement, it appeared that any attempt to determine non-stoicheiometry in samples of zinc oxide would be worth while. The greatest merit of the determination of the extent of non-stoicheiometry by chemical means is the directness of the approach. Measurements on the loss of weight of zinc oxide after being heated are not helpful, since there is appreciable volatility, i.e., loss of both zinc and oxygen, under the conditions in which the presence of excess of zinc becomes evident, e.g., in terms of colour change or semi-conductivity.Optical and electrical methods of estimating non-stoicheiometry involve assumptions about the relation between the relevant defect centres and the departure from stoicheiometry. It must be emphasised that the chemical analysis of a material, the over-all composition of which is non-stoicheiometric, cannot in principle give any information about the way in which the non-stoicheiometry is present in the material. For example, when excess of zinc is determined in zinc oxide, the method of analysis cannot itself distinguish between zinc oxide crystals incorporating an excess of zinc in their lattice and a mechanical mixture of stoicheiometric zinc oxide crystals with free zinc metal.However, it is reasonable to believe The reproducibility of the method is good.July, 19571 ALLSOPP: THE DETERMINATIOK OF EXCESS OF ZINC IN ZINC OXIDE 475 that any excess of zinc found in zinc oxide that has been subjected to conditions of temperature and atmosphere in which no zinc metal would be expected to remain in the free state can be ascribed to a non-stoicheiometric lattice. Previous work on chemical determination of excess of zinc in zinc oxide has been carried out by Ehret and Greenstone,e and by von Wartenburg (reported by Mollwo and Stiickmann'). Ehret and Greenstone studied the red variety of zinc oxide first prepared by Kut~elnigg8~~ in which the excess of metal was found to be approximately 0.02 per cent.by weight by (i) vacuum distillation of metal from the oxide at 335" C, (ii) reduction of potassium dichro- mate, (iii) reduction of potassium permanganate and (iv) reaction with bromobenzene. Von Wartenburg made use of the reduction of potassium permanganate, with precautions to prevent spurious reduction of the reagent. This method was applied to determining the excess of zinc in specimens of zinc oxide that had been deposited by Mollwo and Stockmann on glass plates inserted in a coal-gas flame containing zinc vapour. By varying the amount of oxygen in the burning gas, the colour of the zinc oxide deposit could be varied from white, through yellow to black. Von Wartenburg found a 0.5 to 1 per cent. excess of zinc by weight in the yellow deposits.The above-mentioned varieties of zinc oxide may indeed consist of crystals containing all the measured excess of metal in a non-stoicheiometric lattice, although it is possible to argue against this. However, the excess of zinc in the specimens of zinc oxide used in the diffusion and sintering work already mentioned (prepared in oxygen or argon at temperatures up to 1450" C) and believed to be contained in the crystal lattice was, on the whole, certainly much less than the excess of zinc reported by Ehret and Greenstone and by von Wartenburg. Many of the present specimens are typical of those used in studies made of the semi-conduction of zinc oxide. EXPERIMENTAL EARLY WORK AND COXSIDERATIONS- Samples of zinc oxide that had been sintered in argon at 1300" C and ground in an alumina mortar were dissolved in an excess of 0.01 N potassium permanganate containing 30 per cent. w/v of sulphuric acid.The residual permanganate was then measured by adding an excess of ferrous am- monium sulphate and titrating with further permanganate. Dissohtion and titration were carried out under oxygen-free nitrogen and all solutions were prepared with doubly distilled water. The results obtained were erratic and too close to the estimated limit of detection of the technique (0.002 per cent. excess of zinc by weight) to be acceptable. An attempt to improve this approach by measuring the residual permanganate spectrophotometrically failed for the same reason. These reactions could also be criticised on the grounds that it was conceivable that some hydrogen from the reaction of the excess of metal with the acid could escape before reacting with the permanganate or other reagent used.Another method was therefore sought. Attack of the oxide by acid, followed by measurement by a gas-analysis technique of the hydrogen produced by the excess of metal was an attractive proposition. In principle, this idea was similar to the method used with success to determine excess of barium in barium 0xide,10J1~~2~~3 although with this oxide the excess of metal was greater and water or water- vapour could be used as reactant. By solution in acid, the solid zinc oxide could be completely broken down, giving the excess of metal the best possible opportunity to react. Further, if the solution could be distilled in vaczbo in the apparatus, it would be reasonably certain that all the hydrogen was liberated.In the first experiments by the hydrogen-evolution technique, sulphuric acid was used as the reactant. Under a vacuum of 10-4mm of mercury or better, the specimen to be analysed was dropped into 90 per cent. w/v sulphuric acid at room temperature (the acid having been previously degassed by being boiled under the same vacuum) and the gas evolved during the reaction was analysed for hydrogen in the way to be described later. Unfor- tunately sulphuric acid was found to be objectionable (i) because of too slow dissolution of the zinc oxide and (ii) because it attacked the wax used on some of the joints of the apparatus. However, a few experiments in which pieces of zinc metal, weighed on a microbalance, were used as specimens gave encouraging results, which are shown in Table I.(The rapid solution in this case obviated the troubles named in (i) and (ii).) At first a method similar to that of von Wartenburg was tried.476 ALLSOPP: THE DETERMINATION OF EXCESS OF ZINC IN ZINC OXIDE [VOl. 82 TABLE I RECOVERY OF METALLIC ZINC, WITH SOLUTION IN SULPHURIC ACID: ANALYTICAL VOLUME, 685 ml; TEMPERATURE, 20” C (by microbalance), Hydrogen pressure, Zinc equivalent, CLg CL of Hg CLg Metallic zinc 89 T 3 31.75 83 50 3 21.55 66 30 6 8.46 22 THE HYDROCHLORIC ACID METHOD- From the experience gained with sulphuric acid, it was decided to investigate the possi- bility of using hydrochloric acid instead. The ease of solution of zinc oxide in this acid would be helpful, and experiments soon showed that its vapours could be condensed in liquid- nitrogen traps and that dry hydrogen chloride was absorbed by soda-asbestos.Hence the wax that it was desired to use in joints in much of the apparatus could be protected and also it would be possible to use a distillation process to ensure release of all hydrogen produced. APPARATUS- The apparatus consists essentially of two parts, a reaction system and an analytical system. The reaction system consists of four bulbs, which may be cooled with liquid nitrogen. Bulb 1 is provided with a side-arm to house the sample and a sealed glass tube filled with iron powder to act as a “pusher.” Bulb 2 is provided with a side-arm through which 50 per cent, v/v hydrochloric acid may be added.The reaction vessels are connected to the main apparatus by a cone and socket joint, “A,” and then via a mercury-vapour cold-trap, 1, and tap TI to a three-stage mercury- diffusion pump, PI. This pump transfers the gases evolved from the reaction vessels to a small analytical system containing a palladium tube, platinum filament, Pirani gauge, soda- asbestos bulb, McLeod gauge and a cold-trap, 2. The palladium tube, soda-asbestos bulb, McLeod gauge and cold-trap can each be isolated from the analytical system by taps. A second mercury-diffusion pump, Pz, is used to evacuate the whole system and is backed by a conventional rotary oil pump. The reaction vessels can, when necessary, be directly evacuated by this backing pump via tap T,. AXALYTICAL PROCEDURE- With the portion of the apparatus up to tap TI and the reaction vessels filled with oxygen- free nitrogen, bulbs 2 and 4 are cooled with liquid nitrogen. Fifteen millilitres of 50 per cent.v/v hydrochloric are placed in bulb 2 and the arm is sealed. The 0.5g of crushed sample in a glass boat together with the “pusher” are placed in the side-arm of bulb 1 and this arm is sealed, and the system is then evacuated. Bulb 1 is heated with a hand torch, after which liquid nitrogen is placed around bulbs 1 arid 3 and trap 1. The liquid nitrogen at bulb 2 is removed, and the bulb is allowed to warm and the acid to distil into bulbs 1 and 3 (the liquid nitrogen at bulb 4 preventing any acid vapours that may escape from bulb 3 from reaching the waxed joints). Tap TI is closed, the sample and pusher are transferred magnetically to bulb 1 and the liquid nitrogen is removed so that the acid thaws to react with the oxide and distils back into bulb 2.Solution is assisted by magnetic stirring. The analytical system is prepared by heating the platinum filament at 600” C and the palladium tube at 350” C for 2 minutes and then cooling. Liquid nitrogen is placed at trap 2 and the soda-asbestos is exposed to the system as a precaution in the event of traces of dry hydrogen chloride findings its way past bulb 4 and trap 1. The gases are then transferred from the reaction vessels to the analytical system over a period of 5 minutes and the pressure is measured by the Pirani gauge. Any oxygen present is allowed to react with the hydrogen on the platinum filament at 600” C, and the pressure drop is measured.The remaining hydrogen is diffused to the atmosphere through the palladium tube at 350” C and the pressure drop is again measured. This procedure was recommended by Ran~1ey.l~ The total hydrogen present is found from these two pressure changes and the zinc equivalent is then calculated. A schematic diagram of the whole apparatus is shown in Fig. 1. Bulb 2 is again cooled. A powdery deposit of zinc chloride is left in bulb 1. Tap T, is closed.July, 19571 ALLSOPP: THE DETERMINATION OF EXCESS OF ZIXC IX ZINC OXIDE N a N 0 M 01 M J M m M .- n .s A? 477 Q-47 8 ALLSOPP: THE DETERMINATIOK OF EXCESS OF ZINC IN ZINC OXIDE [Vol. 82 In order to standardise the method it was desired to test known amounts of pure zinc and to compare the weights with those obtained from their hydrogen equivalents. It was decided to work with evaporated layers of radioactive zinc.The evaporation technique afforded easy deposition of microgram amounts of metal, the mass of which could be deter- mined radiochemically. This was less tedious and more accurate than the microbalance approach used in the early work with sulphuric acid. It was found to be essential to evaporate the zinc in Ztacuo straight into the reaction system. Preliminary experiments in which evaporated deposits were first made in a separate apparatus showed that the rate of oxidation of the zinc deposit when exposed to air at room temperature was sufficiently great as to render the experiments valueless, since the reaction with hydrochloric acid would give the zinc metal equivalent, whereas the radiochemical determination would give the total zinc, including any that had become oxidised.STANDARDISATION OF THE METHOD- Fig. 3. Recovery on standardisation: 0, zinc alone; A, zinc $- 0 . 5 g of zinc oxide zinc The evaporation apparatus constructed is shown in Fig. 2 ; it consists of an inverted U-tube, A, closed at one end and joined to the centre tube of a bulb, B, by means of a short length of I-mm capillary tubing, C; the envelope of bulb B was provided with a side-arm, D. Milligram portions of the radioactive zinc (zinc-65 with a specific activity of 50mC per g approxi- mately) were placed in the closed end of the U-tube before assembly and the side-arm, D, of the bulb, B, was joined to the arm of bulb 1 of the reaction system (see Fig.1). Then 50 per cent. v/v hydrochloric acid was added (under oxygen-free nitrogen) to bulb 2 , the apparatus was sealed and evacuated and a portion of the acid was distilled from bulb 2 to bulb 1, which, together with bulbs 3 and 4, was surrounded with liquid nitrogen. The bulb, B, was cooled with liquid nitrogen and the U-tube was heated to 450" C. Most of the zinc condensed in the U-tube above the capillary, but microgram portions condensed at the bottom of bulb B (see Fig. 2). At this stage the capillary was sealed with a torch, so that the small portion of zinc that had reached bulb B was incorporated in the apparatus for analysis. The reaction system, still evacuated, was tlnen isolated from the main apparatus and a portion of the acid in bulb 1 was distilled into bulb B.The liquid nitrogen was removed from bulb B, and the acid was allowed to thaw and react with the zinc and distil back into bulb 1. The evolved gases were pumped into the analytical system and analysed for hydrogen content as already described. The bulb, B, was then disconnected from the reaction system and the residue of zinc chloride was dissolved in 20 ml of 30 per cent. v//v hydrochloric acid. The gamma activity The zinc equivalent was then calculated.July, 19571 ALLSOPP: THE DETERMINATION OF EXCESS OF ZINC IN ZIKC OXIDE 479 of 10 ml of this solution was measured by means of a scintillation counter. The activity per dissolved microgram of the radioactive zinc metal used in these standardisation experi- ments had been determined previously by dissolving weighed samples (10 mg approximately) and diluting until the solution had a convenient count-rate (the same counting geometry was maintained throughout).Hence, after correction had been made for decay (the half-life of zinc-65 is 250 days), the mass of zinc deposited in bulb B in each evaporation was easily calculable. After allowance had been made for errors in the counting, the decay constant and the measurement of volumes, and errors in the calibration, the accuracy of the radio- chemical determination is estimated to be within &3 per cent. The results from the hydrogen evolution were plotted against those determined by the radiochemical method and are shown in Fig. 3.It will be seen that the relationship was nearly linear, but the values obtained from the hydrogen evolution were only about 70 per cent. of those obtained radiochemically. Having established a recovery for zinc metal, it was thought desirable to do experiments in which zinc metal was deposited on zinc oxide in order to check whether the large residue of zinc chloride had any catalytic or similar effect on the hydrogen produced. This was done by using the same apparatus as before with 0.5 g of zinc oxide (previously found to possess a 0.25 p.p.m. excess of zinc in zinc oxide by weight by the method described in this paper) present in bulb B. The radioactive zinc was then evaporated on to the zinc oxide powder and processed as before. The results showed a recovery of the same order as that for zinc alone (see A, Fig.3). POSSIBLE CHANGE OF COMPOSITION DURING GRINDING- It was considered essential to check whether the initial grinding of the specimen, which was necessary to obtain dissolution in a reasonable time, could alter the composition of the oxide. To investigate this, a block of zinc oxide (sintered at 1300” C in argon for 20 hours) was progressively ground and samples were taken for determination of the excess of zinc. First a portion of this block was crushed in an alumina mortar, and the particles were shaken on to a piece of millimetre graph paper and particles of about 1 mm and 0.5 mm diameter were hand-picked from there. The remaining particles were then graded into those less than 0.25 mm, much less than 0.25 mm and very much less than 0.25 mm by “tabling” on an inclined paper. Another portion of the block was ground in a mechanical agate mortar and samples were taken after grinding for 5 , 15, 20, 25, 30, 45, 60 and 120 minutes.il third portion was ground by hand in an alumina mortar for less than 1 minute in exactly the same way as in the recommended procedure and a fourth portion was ground in an alumina mortar for 120 minutes. All these samples were then analysed. RESULTS AND DISCUSSION The fact that the gas evolved, on dissolution of the oxide, was hydrogen was confirmed by the Pirani - McLeod factor. The results of the grinding experiment described in the previous section are presented first (see Fig. 4). All of the figures obtained up to 15 minutes’ grinding time are in agreement and then the figures increase sharply up to 30 minutes’ grinding time, after which, as far as can be seen, they remain roughly constant. It would seem, therefore, that particles of the oxide lose oxygen when heated during the short intervals of time during which they are under the pestle and that cooling is too rapid for the replacement of all the oxygen lost.However, the important practical point is that the usual grinding procedure does not appear to alter the composition of the oxide (see 0, Fig. 4). The result from the long grinding in an alumina mortar (see X in Fig. 4) agrees with the result of long grinding in agate, suggesting that contamination from mortar materials had not occurred in either case, since it would have been a remarkable coincidence if contamination had occurred to the same extent in both mortars.Table I1 contains the results of determination of the excess of zinc in various zinc oxide preparations. I t will be seen that the excess of zinc in the specimens was about 0.2 to 18 p.p.m. by weight, which is considerably less than that in the specimens studied by other workers.6$7 The reproducibility of the method was good and the sensitivity was estimated to be a 0.1 p.p.m. excess of zinc in zinc oxide by weight from figures obtained from blanks with480 [Vol. 82 15 ml of 50 per cent. v/v hydrochloric acid alome. The radiochemical determinations sug- gested that the results were certainly accurate ,within 30 per cent. However, it is believed that the 30 per cent. discrepancy in the standardisation experiments arose for the following reasons (i) that hydrochloric acid does not react with all of the evaporated zinc that is deposited in the upper part of the inner sleeve of bulb B and (ii) that oxidation of zinc occurs during the evaporation.It is therefore thought that the accuracy of the determinations of excess of zinc in zinc oxide is considerably better than 70 per cent. ALLSOPP: THE DETERMINATION OF EXCESS OF ZINC IN ZINC OXIDE Fig. 4. Change of composition of zinc oxide during grinding: A, ground in an agate mortar; X, ground in an alumina mortar Work is in progress on the further standardisation of the method by means of a molecular- beam technique in order to concentrate the zinc: molecules on to a desired spot in bulb B. The effective volume of the distillation apparatus will be much smaller, so minimising the danger of oxidation.It is not proposed to enter here into a full discussion of the interpretation of the various results given in Table 11, since the purpose of this paper is to describe the method of analysis, but the following comments must be made. It seems reasonable to suppose that the results on sintered material correspond to the excess of zinc present in non-stoicheiometric zinc oxide crystals, because any free zinc present in the starting powder might have been expected to oxidise or vaporise during sintering. Nevertheless, it should be emphasised that (i) it is not yet clear how uniform is the distribution of the excess of metal in the specimens analysed and (ii) the first results suggest a variation of non-stoicheiometry among specimens of different raw materials treated similarly.It is considered that (ii) is probably due to the influence of impurities on the non- stoicheiometric composition attained at high temperatures. However, it must not be over- looked that substituted impurity cations of valency higher than zinc could also cause evolution of hydrogen during solution of the oxide in hydrochloric acid, but the small amount of hydro- gen produced by some of the preparations investigated gives an upper limit to this effect of impurity and suggests that the effect is probablly small with the raw materials from which the present specimens were made. Turning to the results for powders, the relatively high content of excess of zinc in powder heated in air at 400" C (compared with that in the powder before being heated and also after being heated in air at 700" C) will be investigated further, as will powder heated under other conditions in this temperature range, which is of particular importance to studies of the sintering, catalytic and other properties of zinc oxide.The powders were made by the French process (burning zinc vapour), so that it is possible that the excess of zinc in the powder as received could be due to free metal, but this is uncertain. Similarly, the state of the excess of zinc in powder heated to 400" C is uncertain.July, 19571 ALLSOPP: THE DETERMINATIOK OF EXCESS OF ZIKC IS ZINC OXIDE 481 Finally, previous estimates of the concentration of excess of zinc in non-stoicheiometric zinc oxide by non-chemical methods should be given.From measurement of the temperature dependance of the Hall coefficient, Hahn15 calculated that the donor concentration in powder pressings sintered in air at temperatures between 895" and 1400" C (with various cooling schedules) was in the range 1.9 x 10l6 to 9.0 x lo1' per cubic centimetre; if it is assumed that the numbers of atoms with excess of zinc equals the number of donors, the excess of zinc content of the specimens of zinc oxide would have been in the range 0.4 to 18 p.p.m. by weight. In their work on the electrical conductivity of zinc oxide with additions of alu- minium oxide, Hauffe and Vierkla estimated that there is a fractional concentration of inter- stitial zinc in pure zinc oxide at 400" C of the order of 4 x Scharowsky17 estimated by dispersion theory the concentration of absorption centres in needle crystals of zinc oxide that had been heated in zinc vapour at temperatures between 920" and 1220" C (the concen- tration of zinc atoms in the vapour ranging from 7.2 x lo1* to 6.1 x 1019).The concen- tration of centres varied from 6.6 x 1017 to 6.8 x 10l8 per cubic centimetre. If it be assumed that the number of atoms in the excess of zinc equals the number of absorption centres, the non-stoicheiometry of the crystals would range from a 13 to 140 p.p.m. excess of zinc in zinc oxide by weight. The order of magnitude of the non-stoicheiometry estimated from these researches seems to be in rough accord with the present results of direct chemical analysis for excess of zinc.TABLE I1 TYPICAL RESULTS BY THE PROPOSED METHOD : ANALYTICAL VOLUME, 405 ml; TEMPERATURE 20" c Conditions of sintering A I -l Atmosphere of 1 atmos- perature, Time, Specimen phere) "C hours (pressure Tem- argon 1300 20 I Sintered bar, approximately 1.2 cm x 1.2 cm x 10 cm Sintered plate, approximate- ly 1.2 cm x 1.2 cm x 0.2 cm Bar: oxygen Cut from bar, approximately 1-2 cm x 1.2 cm x 10 cm Plate: argon (material DA) 1200 and then 1300 and then 1300 1200 1300 7 As received air 400 7 air 700 2 As received air 400 7 air 700 2 METHOD { 2:: 2: Sintered plate (as above) oxygen Cut from bar (as above) (material AF) argon Powder, particle size originally argon 0.1 to 1 p (material BH) Powder, particle size origin- ally 0.1 to 1 p (material Powder, particle size origin- ally 0.1 to 1 p (material AF) BC) PROCEDURE- Mass of sample, g 0.506 0.502 0.505 0.507 0.504 0.500 0.508 0.504 0.345 0,373 0.286 0.321 0.503 0.504 0.496 0.497 0.488 0.503 0.506 0,507 0.505 Excess of zinc in zinc oxide Hydrogen evolved, P of Hg 5.68 5.69 5.80 5.86 0.92 0.88 0.61 0.83 0.77 0.77 1.58 1.90 5.39 5.49 0.16 1.61 0.07 0.82 1.61 1.68 0.09 by weight, p.p.m.17.5 17.6 17.9 18.0 2.8 2.7 1.9 2.6 3.5 3.2 8.6 9.1 16.5 17.0 0.5 5.0 0.2 2.6 6-0 5.2 0.3 Store the specimen when received in a clean glass tube and, when required, crush it to a powder in a clean alumina mortar. Weigh 0.5 g into a small glass boat. With all taps closed, switch on all three pumps and when pumps P, and P, are working open taps T,, T,, T3, T, and T,.When a suitable vacuum is reached, as indicated by the Pirani gauge, i.e., less than 0.2 p of mercury, close T,, bleed in oxygen-free nitrogen through T, and when the pressures are equalised remove the cap from joint "A."482 ALLSOPP: THE DETERMINATION OF EXCESS OF ZINC IN ZINC OXIDE [Vol. 82 Flush the reaction system with oxygen-free nitrogen through the side-arm of bulb 1 and connect to the main apparatus with the nitrogen still flowing and leaving the vessels from the side-arm of bulb 2. Place liquid nitrogen at bulbs 2 and 4, add 15 ml of 50 per cent. v/v hydrochloric acid into bulb 2, stop the flow of nitrogen gas to bulb 1 and seal off the side-arm at bulb 2. Place the boat containing the specimen to be analysed together with a sealed glass tube containing iron powder as a pusher into the side-arm of bulb 1, and draw the side-arm down to a fine capillary with the aid of a hand-torch; stop the flow of nitrogen by closing T, and seal.Close T, and evacuate the reaction system roughly through T, and close T,. Open T, and then evacuate the vessels through the mercury-diffusion pumps by opening T,. Heat bulb 1 carefully to remove any adherent oxygen and, when a suitable vacuum is reached, again less than 0.2 p of mercury, place liquid nitrogen at bulbs 1 and 3 and trap 1. Remove the liquid nitrogen from bulb 2 and allow the acid to distil into bulbs 1 and 3, assisting dis- tillation with gentle heat. When the distillation is complete, close T,, and transfer the boat containing the oxide and the pusher to bulb 1, using a magnet for the purpose; replace the liquid nitrogen at bulb 2 and remove that at bulb 1 to allow the acid to thaw and react with the oxide and to distil back into bulb 2.Solution is assisted by mechanical magnetic stirring. Allow the solution and distillation to proceed without heat, topping up the cold- traps with liquid nitrogen meanwhile. When solution and distillation are complete, run the platinum filament at dull-red heat (600' C) and the palladium tube at 350" C for 2 minutes with T, open, and then cool, and close T,, isolating the palladium tube. Open T, to expose the soda-asbestos, place liquid nitrogen at trap 2 and close T,. Open T, and pump the gases from the reaction vessels to the analytical system over a period of 5 minutes. Close T, and take the pressure reading on the Pirani gauge.Run the platinum filament at about 600" C for 2 minutes and record the reading on the Pirani gauge. Open T,, place the small furnace, running at 350"C, over the palladium tube and diffuse out the hydrogen until the pressure is constant. Note the reading on the Pirani gauge and remove thLe small furnace from the palladium tube. Open T,, evacuate the analytical system and close T, and T,. Remove the liquid nitrogen from bulbs 2 and 3 and trap 1, and open T, to the atmosphere to let in air t o the reaction system. When the system is full of air, remove the vessels from the apparatus, replace the cap at joint ''A,'' close taps T, and T, and open tap T, to roughly evacuate the system to T,. When trap 1 has thawed completely, close T, and open T, and then T,.Pump for about 1 hour and during this period remove the liquid nitrogen from trap 2 . Close all taps, switch off the mercurydiffusion pumps, open T,, and switch oft the backing pump. Tap T,, is used only for slow rough evacuation of the system and T,, and T,, when the McLeod gauge is being used. CALCULATIOX- The pressure drops during catalysis (if any) and during the diffusion through palladium are converted to microns of mercury on an air-calibration graph and then to the pressure of hydrogen as microns of mercury, 1.60 being taken as the Pirani factor for hydrogen (determined experimentally), as follows- Pressure drop on catalysis on platinum, microns of mercury 2.10 = Hydrogen pressure, microns of mercury. Pressure drop on diffusion through palladium, microns of mercury 1.60 = Hydrogen pressure, microns of mercury.Total hydrogen, microns of mercury x 65.3'3 x analytical volume, ml 22.4 x 760 x weight of sample, g = Excess of zinc, p.p.m. I express my gratitude to Mr. J. P. Roberts for unstinted help and advice and for the radiochemical determination of zinc, to Mr. H. W. Sumner for glassblowing, to Messrs. H. C. Davis, W. N. Mair and W. Watt for advice, particularly during the early stages of thisJuly, 19571 ALLSOPP: THE DETERMINATION OF EXCESS OF ZINC IN ZINC OXIDE 433 work and to Mr. T. R. F. W. Fennel1 for weighing microgram amounts of zinc metal. Reproduced by permission of the Controller, H.M. Stationery Office. REFEREXCES Roberts, J . P., Hutchings, J., and Wheeler, C., Trans. Brit. Ceram. SOC., 1956, 55, 75. Hutchings, J., and Roberts, J. P., Paper to be published in the Proceedings of the 3rd International -, -, unpublished Ministry of Supply Technical Note, 1953. --, -, unpublished Ministry of Supply Report, 1956. Anderson, J. S., “Annual Reports on the Progress of Chemistry for 1946,” Chemical Society, Ehret, W. F., and Greenstone, A., J . Anzer. Chem. SOC., 1943, 65, 872. Mollwo, E., and Stockmann, F., Ann. Physik, 1948, [6] 3, 226. Kutzelnigg, A, 2. anorg. Chem., 1932, 208, 23. Berdennikowa, T. P., Phys. 2. Sowjet, 1932, 2, 77. Jenkins, R. O., and Newton, R. H. C., J . Sci. Instrum., 1949, 26, 172. Libowitz, G. G., J . Amer. Chem. SOC., 1954, 75, 1501. Wooten, L. A., Moore, G. E., and Guldner, W. G., J . Appl. Phys., 1955, 26, 937. Ransley, C. E., AnaZyst, 1947, 72, 504. Hahn, E. E., J . AppZ. Phys., 1951, 22, 855. Hauffe, K., and Vierk, A. L., Z. phys. Chem., 1950, 196, 160. Scharowsky, E., 2. Phys., 1953, 135, 318. 1 . 2. 3. 4. 5. 6 . 7. 8. 10. 11. 12. 13. 14. 15. 16. 17. Meeting on the Reactivity of Solids, Madrid, April 1956. London, 1947, p. 117. 9. -, Ibid., 1934, 221, 46. ROYAL AIRCRAFT ESTABLISHMENT FARNBOROUGH, HANTS. November 26th, 1956

ISSN:0003-2654    

DOI:10.1039/AN9578200474

出版商:RSC    

年代:1957

数据来源: RSC

摘要:

July, 19571 ALLSOPP: THE DETERMINSTIOK OF EXCESS OF ZINC IN ZINC OXIDE 433 Determination of Gold in Sea Water by Radioactivation Analysis BY R. W. HUJIMEL Small portions of sea water were irradiated with slow neutrons in a Harwell Pile and the radioactivity from the gold-198 produced was compared with that from a standard gold solution. The gold contents found depended on the distance from the shore that the samples were taken, and varied from about 400 p g per cubic metre for English coastal water to about 15 pg per cubic metre for water from the north-w-estern limit of the Bay of Biscay. OF the many investigations of the gold content of sea waters, the first reported apparently was that of S0nstadt.l Many more followed until after the First World War, when Fritz Haber embarked on a seven-year study a t first intended to relieve the burden of Germany’s war debts, but which culminated in the desire to add to scientific knowledge by determining as closely as possible the actual gold content of sea water. This change in emphasis occurred when he discovered that the gold content was much lower than earlier studies had led him to b e l i e ~ e .~ ,3 9 4 In their very careful studies involving thousands of analysed samples from the European and North American coasts and from both the tropical and arctic Atlantic, Haber and his co-workers strove to eliminate the disturbing effects of gold introduced with the reagents and by laboratory dust, as well as gold leached out of containers by the corrosive waters and gold adsorbed on the vessel walls. Bauer has stated5 that Haber disregarded both the possibility of adsorption and the likelihood of the oxidation of colloidal gold to ionic forms by dissolved oxygen in the sea, but he was at fault in each case, as can be seen by a perusal of Haber’s paper^.^^^ The analytical method finally developed by Haber was, in brief, to precipitate lead sulphide in the sea-water sample within a few minutes of its collection by adding to it solutions of lead acetate and ammonium sulphide, all the gold being occluded and errors caused by adsorption on the walls or by leaching being minimised.After several minutes the precipitate was treated with lead formate and boric acid and the whole was converted by heating to a lead regulus embedded in a lead borate slag. Finally, by carefully controlled heating the lead nucleus was melted and scorified, leaving a gold - silver granule measurable without removing it from the slag in which it was embedded.Further careful heating near the melting-point of gold caused the silver to go into the slag. The gold granule embedded484 HUMMEL: DETERMINATION O F GOLD IN SEA WATER [Vol. 82 in the slag was measured microscopically after placing it in 1-bromonaphthalene, the refractive index of which equalled that of the slag. The mean of all the gold determinations made on samples collected during trips across the South Atlantic by the German research ship “Meteor” (1635 samples were taken for gold analysis, at 186 stations) was 4 x lO-9g of gold per kg of sea water. Samples from North Atlantic and Arctic waters led Haber to believe that the South Atlantic was a somewhat gold-poor area, and he stated3 that the sea near Iceland and east coast of Greenland exhibited a mean gold content about ten times greater than the average of the “Meteor” samples.Haber found not only that large sea areas appeared to differ in their gold contents, but also that individual samples taken almost simultaneously at the same location gave widely variable results, ,and he concluded, after many experiments, that the local variations were due to the likelihood of gold occurring not only as chloroauric ions but also as colloidal particles or adsorbed on or included in suspended inorganic and organic matter. This refutes the comment of l?utnam6 that Haber and others “have not appreciated the extreme variability of sea water with respect to biologically active trace elements.” Among the more recent investigations have been those of Caldwell,’ Stark* and Putnam.6 These re-emphasised the variability to be expected, the values found covering the range from 20 to 2000 pg per cubic metre.In general the highest values were found in coastal waters. Putnam, using a modification of Haber’s method (but with ferric hydroxide to carry down the gold, a procedure considered unsuitable by Haber not only because small amounts of the precipitate do not occlude all the gold, but also because of its awkward filtering properties) and finally determining the gold colorimetrically by an improvement of the stannous chloride test used by Sonstadt in 1872, found no detectable gold in samples taken from both the eastern and western coastal areas of the United States, from the Gulf Stream near Florida or from a point 100 miles offshore from New York.He concluded that sea water ordinarily contains less than 0.5 pg of gold per cubic metre, since the analytical method was sufficiently sensitive to detect gold added to sea water at a concentration of 0.5 pg per ton. In view of the wide variations and uncertainties in much previous work, owing almost certainly to the difficulties inherent in the classical methods when extended to their lower detection limits, it seemed desirable to apply the radioactivation method, which has a lower detection limit for gold, at a flux of 10l2 neutrons per sq. cm per second and a minimum detectable radioactivity of 20 disintegrations per minute, of 2.5 x 10-l1 g (the amount present in 2.5 ml of sea water containing 10 pg per cubic metre).EXPERIMENTAL The samples of sea water analysed for their gold contents were as follows- (a) surface samples collected by the author at various distances offshore from Portland Bill, (b) surface and sub-surface samples obtained at International Hydrographic Station El, 25 miles offshore from Plymouth, through the kind co-operation of L. H. N. Cooper and F. A. J. Armstrong of the Marine Biological Association of the United Kingdom, and (c) surface and sub-surface samples from the north-western limits of the Bay of Biscay at 46” 30’N, 08” OO’W, again collected by F. A. J. Armstrong. The results of a number of analyses of deep-sea waters stored for several years in a glass carboy or in a polythene bottle will not be discussed, since it is suspected that adsorption of gold on the vessel walls has rendered them valueless. The amount of gold found in the polythene-stored water was so low compared with that stored in the carboy that adsorption was suspected.Experiments in which gold-198 was added to sea water in polythene bottles has confirmed this view, as has also an experiment in which surface water from 1 mile off Portland Bill was collected almost simultaneously in a polythene bottle and in six silica tubes in which the water was to be irradiated. Subsequent radioactivation analyses of these samples showed that over a 3-week interv(a1 the gold content of the sea water in the polythene bottle decreased to less than a quarter of that collected directly in the irradiation tubes.New silica irradiation tubes were made for all samples collected from Portland, Plymouth and the Bay of Biscay. The silica was first rinsed with hot aqua regia, then with demineralisedJuly, 19571 BY RADIOACTIVATIOS SNALYSIS 485 water and finally flame-dried. The tubes were made with a long narrow neck and held from 8 to 10 ml. After fabrication they were flame-dried, partly evacuated on the inter- laboratory vacuum line and sealed off at the narrow end. In each experiment, four tubes contained only sea water, two contained standard gold solution (71.7 mg of gold per litre) either with or without sea water and two tubes were left empty. Both the sea-water tubes and blank tubes were made from the same two pieces of 12-mm and 3-mm bore silica tubing.The narrow ends of the standard tubes came from a second piece of 3-mm bore silica, but in this case any possible difference in the amount of gold leached from these tubes, compared with the other six, would be insignificant owing to the overwhelming preponderance of gold added in the standard solution. The surface samples from Portland were collected simply by dipping the tubes into the water and breaking off their tips. Surface samples from Plymouth and the Bay of Biscay were taken in a polythene bucket and the irradiation tubes were filled from it within a minute of collection. Sub-surface samples were taken with a Perspex-lined Nansen - Petterssen water bottle from which portions of the water were run into a 200-ml polythene beaker in which the tips of the tubes were broken off.The beaker had been previously washed with aqua regia and kept free from dust until used. The water was in contact with the Perspex for 3 or 4 minutes and with the polythene for less than 1 minute. After collection, the open ends of the tubes were closed with plastic caps tightly fitted to keep an air gap between the caps and the sea water, so minimising the possibility of adsorption of gold by the caps. On arrival at the laboratory the caps were removed and any excess of water was shaken out to give a remaining volume of from 4 to 6 ml. The outer surfaces of the tubes were then carefully rinsed with demineralised water. The tubes were weighed on a balance not used for any weighings in which milligram amounts of gold were involved.The corresponding blank tubes then had their tips removed and were washed and weighed in the same manner as the sea-water tubes. Then about 0.4 g of standard gold solution, containing 71.7 mg of gold per litre, was added to the remaining two tubes of the set. These may or may not have been filled with sea water. Control experiments revealed no appreciable difference in the results whether the gold solutions were irradiated with or without admixed sea water. The amounts of solution added were determined by weight difference. The gold used to make up the standard solution was precipitated with hydroquinone from a 2 per cent. w/v gold chloride solution. Subsequent irradiations of the standard solution, followed by chemical treatment as described below, revealed no detectable active impurity on following the decay through six to seven half-lives. For the neutron irradiations of these samples in the Pile, the apparatus designed by Pateg was used, in which the eight open narrow-necked silica tubes were held vertically in a graphite housing placed inside an aluminium can. The samples were irradiated for 6 days in a thermal column of the Harwell BEPO Pile and then left for 2 or 3 days to allow the activity to fall to a convenient working level.ANALYTICAL PROCEDURE- The analytical procedure described below was evolved with tubes containing no added gold and is based on a method described by Goldberg and Brown.lo The outside of the tube was rinsed with hot aqua regia containing a trace of inactive gold in order to remove possible active-gold contamination.Then about 2 ml of 10 per cent. hydrochloric acid were introduced and the tube was heated gently until, in those instances where the tube had gone dry during irradiation, the solids had dissolved. The tube was then inverted over a tall 250-ml Pyrex-glass beaker that had previously been cleaned with hot aqua regia containing inactive gold and the washing solution expelled by heating the closed end. (All beakers were numbered and reserved for use only with similar samples. For example, beakers 1 and 2 were used for the blanks and never contained gold activities greater than several hundred disintegrations per minute per ml. Between experi- ments all were rinsed two or three times with hot aqua regia containing added inactive gold.) The tube w;s washed once more with 10 per cent.hydrochloric acid, twice with hot aqua regia containing a trace of inactive gold and finally twice more with 10 per cent. hydrochloric acid. Five millilitres of 2 per cent. w/v gold chloride solution were added and the solution was evaporated nearly to dryness. A few drops of aqua regia were added to ensure complete chemical exchange between active and inactive gold, and the solution was again evaporated Other glassware was similarly reserved.486 HUMMEL: DETERMINATION OF GOLD I N SEA WATER [Vol. 82 almost to dryness. It was then transferred to a separating funnel by using 25 ml of 10 per cent. hydrochloric acid and then an equal volume of ethyl acetate. After extraction, the acetate layer was washed twice with equal volumes of the 10 per cent.acid and then trans- ferred back to the original beaker. This in the meantime had been thoroughly rinsed to ensure that any minute particles of silica originating from the breaking off of the tip of the tube were not carried through to contaminate the gold finally precipitated and prepared for counting. After evaporation of the ethyl acetate solution to dryness, the gold was dissolved in several drops of aqua regia. This was heated nearly to dryness to remove most of the nitric acid and then diluted to give about 25 ml of 25 per cent. hydrochloric acid solution. The solution was heated to boiling, 10 ml of 5 per cent. hydroquinone solution were added and the mixture was boiled gently for several minutes or until the gold had coagulated to leave a clear solution. The gold was washed twice with hot water and then dissolved in a few drops of aqua regia.The entire procedure from extraction with ethyl acetate to precipi- tation with hydroquinone was repeated. The final gold precipitate was washed three times with hot water (cold water was added and brought to the boil, so as to break up the larger particles and make possible a more even deposit on the counting tray) and twice with acetone. After being dried, the coarse gold powder was transferred directly to a weighed aluminium counting tray and flattened down with a numbered spatula. A measured amount of a dilute acetone solution of collodion was added to affix the gold to the tray.A few minutes' drying under an infra-red lamp was followed by weighing and counting. The contents of the tubes containing the standard solution were extracted in the same way, but were expelled into a 1-litre flask containing 50 ml of concentrated hydrochloric acid and 5ml of 2 per cent. w/v gold chloride solution. After dilution up to the mark, 10 ml were put by pipette into a 100-ml flask containing acid and gold in the same proportions as in the 1-litre flask. One millilitre of this was transferred to the appropriate 250-ml beaker for treatment as described above. The empty irradiation tubes were finally dried and weighed to determine the weights of sea water irradiated. An end-window Geiger tube was used for counting. The usual corrections were made for coincidence, background and chemical yield.The last was generally between 60 and 80 per cent. Decays of blank, sea water and standard samples were followed on several occasions over 6 to 7 half-lives and throughout that period exhibited a half-life of 2.69 days, within the statistical counting errors. The mean activity of the two blank tubes was subtracted from the activities obtained from the sea waters. The blank activities averaged slightly more than 7 per cent. of the sea-water activities for all the experiments reported in Table I. The ratio increased to 20 per cent. for the last five experiments, the blanks averaging 57 counts per minute and the sea water samples 290 counts per minute. TABLE I GOLD CONTENTS OF SEA WATERS BY RADIOACTIVATION Date of Number of Gold content, Source of water collection Depth samples pg per cubic metre 320 150 506 k 14 411 5 10 63* 15 & 9 Surface 4 96 i 30 19 & 13 89 58 Surface 4 14 9 4 29 12 Surface 4 13 2 45 metres 4 21 i 6 Portland, 1 mile.... , . Dec. 7, 1955 Surface I ; 'I: Portland, breakwater , . .. Jan. 19, 1956 Surface 3 Portland, 5 miles . . .. Jan. 19, 1956 Surface 4 185 i 115 20 metres 2 50 metres 2 { Plymouth, 50" 02' N, 04" 22' W' Feb. 21, 1956 Mar. 26, 1956 46" 30' K, 08' 00' W . . . . June 22, 1956 { 20 metres -600 metres 3 10 5 3 * Mean of three experiments with water collected in a polythene bottle.July, 19571 BY RADIOACTIVATION ANALYSIS 487 Irradiated gold from both standard solution and sea water were assumed to have the same specific activity, so that- - Weight of gold in sea water Corrected count rate of sea water' - Weight of gold in standard Corrected count rate of standard and hence the weight of gold per unit weight of sea water could be calculated.However, the results of this investigation are arbitrarily expressed in terms of the weight of gold per unit volume of sea water. The values of the gold contents given are the means of the several values obtained by comparing each sea-water sample (the numbers of which are given in the Table) with each of the two standard samples. The standard deviation of each mean is given also. The resulting error is considered insignificant. The results are given in Table I. COXLUSIONS The results are in general agreement both with those workers who have only been able to obtain quantitative results with coastal samples containing relatively large amounts of gold and with Haber and his co-workers, who, by extremely careful application of refined techniques, were able to determine the gold contents of deep-sea samples as well as of coastal samples. Somewhat smaller amounts of gold were found when samples were obtained progressively further out to sea.The means of the samples collected at Portland (1 mile), Portland (5 miles), Plymouth and Bay of Biscay are, respectively, 409, 185, 49 and 15 pg per cubic metre. The lowest values found were still well within the sensitivity of the method, taking as the criterion the relative activities of blanks and sea-water samples. As little as 3 to 5 pg per cubic metre should be distinguishable by using the method described here.If many samples were taken, to allow confident statistical treatment of the results, the lower limit would be extended still further. Reduction of the blank values seems to offer the widest scope for improvement of the method. This might best be accomplished by the use of a less corrosive reagent to remove the irradiated samples from the tubes, so reducing the amount of gold leached from the tubes themselves. The deviations given in Table I tend to confirm Haber's opinion that much of the gold content of sea water is associated with suspended particles of organic or inorganic materials. This hypothesis is supported by the decrease in gold contents as samples are obtained progressively further from shore. The results of some analyses of sea-water samples collected at the same time as those examined for their gold contents are given in Table 11.These are reproduced with the permission of the Marine Biological Association and the Government Chemist. A decrease of about 15 per cent. in the dissolved silicate contents between February 21st and March 26th, possibly associated with the spring outburst of phytoplankton, is accompanied by a decrease in the gold contents of about 60 per cent. Many more experiments would be required to establish if there is a reliable correlation. TABLE I1 ANALYSES OF SEA WATER FROM INTERNATIONAL HYDROGRAPHIC STATION E l , BY THE MARINE BIOLOGICAL ASSOCIATION (SALINITIES DETERMINED BY THE GOVERNMENT CHEMIST) Tempera- Date of ture, Salinity, collection Depth "C "a0 Surface 8.8 35.35 20 metres 8.89 39-30 45 metres 8.88 35.32 Surface 8.7 35.43 8.63 35.39 Surface 15.3 35.71 12.84 35.64 600 metres 10.30 35.58 Feb.21, 1956 Mar. 26, 1956 { 2o metres At Stutioiz 46" 30' IT, 8' 00' W in Bay of Biscuy- Phosphate, pg-atom of P per litre 0.63 0.49 0.49 0.54 0.46 Silicate, pg-atom of Si per litre 2.9 2.9 2.9 2.6 2.5 0.7 1.5 7.6 Oxygen, ml per litre 5.92 5.75 4.72488 TAYLOR: FORMOL TITRATION : AN EVALUATION OF [Vol. 82 The very low value obtained with sea water from the breakwater at Portland Bill remains inexplicable. It is a pleasure to acknowledge my indebtedness to Mr. A. A. Smales, Head of Analytical Chemistry Group, Chemistry Division, A.E.R.ES., who proposed the problem and made several valuable suggestions during the course of the work. I am also very grateful for the invaluable assistance given by members of the staff of the Marine Biological Association, without which the major portion of this work would not have been possible. REFEREXCES 1. 2. 3. 4. 5 , 6. 7. 8. 9. 10. Sonstadt, E., Chem. News, 1872, 26, 159. Haber, F., 2. angew. Chem., 1927, 40, 303. Haber, F., Z . Ges. Erdk. Berlzn, 1928, Suppl. 3, 3. Jaenicke, J., Naturwiss., 1935, 23, 57. Bauer, E., Helv. Chzm. Acta, 1942, 25, 1202. Putnam, G. L., J. Chem. Educ., 1953, 30, 576. Caldwell, W. E., Ibid., 1938, 15, 507. Stark, W., Helv. Chim. Acta, 1943, 26, 424. Pate, B. D., Atomic Energy Research Establishment Report C/R 643, H.M. Stationery Office, Goldberg, E. D., and Brown, H., Anal. Chem., 1950, 22, 308. London, 1951. ANALYTICAL CHEMISTRY GROUP ATOMIC ENERGY RESEARCH ESTABLISHMENT HARWELL, NR. DIDCOT, BERKS. February 14th, 1957

ISSN:0003-2654    

DOI:10.1039/AN9578200483

出版商:RSC    

年代:1957

数据来源: RSC