ISSN:0003-2654    

DOI:10.1039/AN95782FX023

出版商:RSC    

年代:1957

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95782BX025

出版商:RSC    

年代:1957

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95782FP075

出版商:RSC    

年代:1957

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN95782BP081

出版商:RSC    

年代:1957

数据来源: RSC

摘要:

JUNE, 1957 THE ANALYST Vol. 82, No. 975 PROCEEDINGS OF THE SOCIETY FOR ANALYTICAL CHEMISTRY GIFT TO THE SOCIETY THE Council has the greatest pleasure in acknowledging the generous gift to the Society by Dr. G. Roche Lynch (Past President) of a long-case clock by Jonathan Loundes (Lowndes) of Pall Mall, ca. 1705. Dr. Roche Lynch has presented the clock to the Society for the Council Room, where it has now been installed. NEW MEMBERS ORDINARY MEMBERS Reginald Clifford Fawcett, M.Sc., Ph.D. (Manc.), A.R.I.C. ; Paul Goudime, M.A. (Cantab.) ; Clarence Bertram Hackett ; Alfred John Hookham ; Clifford Raynor Johnson ; Robert Julius Motz, Dipl.Ing.-Chem. (Vienna), A.R.I.C. ; Derek Robbins, Grad.R.1.C. ; Alfred John Shorter, M.Sc. (Birm.), M.S. (Illinois), A.R.I.C., M.Inst.F., F.Inst.Ceram.; Harold Vincent Street, B.Sc. (Birm.) , M.I.Biol., F.R.I.C. ; Reginald David Taylor, A.R.I.C. ; Joseph Vincent Westwood, MSc. (Lond.), F.R.I.C. ; Raymond Ernest Wilson, B.Sc. (Lond.) , A.R.I.C. ; Robert Wilson, B.Sc. (Edin.). JUNIOR MEMBERS Peter Adams, BSc. (Lond.) ; Ernest Draper; Michael Thomas Hall; Anthony John Harrison; John Coltherd Paxton, B.Sc. (Edin.). DEATHS WE record with regret the deaths of Robert Luman Barnard Arthur Harvey Leslie Herbert Lampit t Herman Lee Albert Edward Parkes John W. Skirvin. MIDLANDS SECTION A SPECIAL General Meeting of the Section was held at 6.45 p.m. on Wednesday, April loth, 1957, in the Mason Theatre, The University, Edmund Street, Birmingham, 3. The Chair was taken by the Chairman of the Section, Dr. R. Belcher, F.Inst.F., F.R.I.C., and an alteration was made t o the Section Rules.AN Ordinary Meeting of the Section was held at 7 p.m. on Wednesday, April loth, 1957, in the Mason Theatre, The University, Edmund Street, Birmingham, 3. The Chair was taken by the Chairman of the Section, Dr. R. Belcher, F.Inst.F., F.R.I.C. The following paper was presented and discussed : “The Analytical Chemistry of Beryllium,” by E. Booth. 377378 ANALYTICAL METHODS COMMITTEE : THE DETERMINATION OF SMALL AMOUNTS OF [VOl. 82 PHYSICAL METHODS GROUP THE 58th Ordinary Meeting of the Group was held at 6.30 p.m. on Tuesday, May 21st, 1957, in the Meeting Room of the Chemical Society, Burlington House, London, W.1. The Chair was taken by the Chairman of the Group, Dr. J. E. Page, F.R.I.C. The subject of the meeting was “Electrochemistry” and the following papers were presented and discussed : “Coulometric Titrations with an Integrated-current Source,” by L. E. Smythe, M.Sc., Ph.D., A.R.I.C., F.R.A.C.I. (presented on his behalf by G. W. C. Milner, MSc., A.Inst.P., F.R.I.C.); “Pulse Polarography,” by A. W. Gardner, B.Sc. A commercial prototype of the coulometric titrimeter and a laboratory-built model of the pulse polarograph were demonstrated.

ISSN:0003-2654    

DOI:10.1039/AN9578200377

出版商:RSC    

年代:1957

数据来源: RSC

摘要:

378 ANALYTICAL METHODS COMMITTEE THE DETERMINATION OF SMALL AMOUNTS OF [VOl. 82 Analytical MethLods Committee REPORT PREPARED B17 THE PESTICIDES RESIDUES IN FOODSTUFFS SIJB-COMMITTEE The Determination of Small Amounts of Total Organic Chlorine in Solvent Extracts of Vegetable Material THE Analytical Methods Committee has received the following report from its Pesticides Residues in Foodstuffs Sub-committee. The Report has been approved by the Analytical Methods Committee and its publication has been authorised by the Council. REPORT In June, 1954, the Pesticides Residues in Foodstuffs Sub-committee appointed a Working Party to carry out collaborative work on the determination of small amounts of chlorine in solvent extracts of foodstuffs for the purpose of detecting the presence of chlorinated hydrocarbons.This work had already been initiated by an Analytical Sub-committee of the Fungicide and Insecticide Research Co-ordination Service of the Agricultural Research Council that had carried out much valuable, but unpublished, experimental work. The members constituting the working party were nominated by the laboratories of Boots Pure Drug Co. Ltd., the Colonial Products Advisory Bureau, the Department of the Government Chemist and Dr. Bernard Dyer and Partners. IYTRODUCTION- Consideration of the comprehensive investigations already undertaken by the A.R.C. Sub-committee indicated that the cause of the wide variation in the results of their last recorded work on the determination of the organic-chlorine content of solvent extracts from grain impregnated with DDT and BHC could be either inefficient extraction from the grain or the method used for the determination of chlorine.The A.R.C. Sub-committee could not decide which of these two factors was the ca.use of failure; accordingly, the working party of the Pesticides Residues in Foodstuffs S ub-Committee concluded that the immediate problem for investigation must be the method for determining known small amounts of organic chlorine compounds in organic solvents. The problem of the efficient solvent extraction of vegetable materials contaminated with halogenated hydrocarbons would then be a matter for subsequent investigation. I t was further agreed, from examination of the available data, that the most promising procedure for the semi-micro determination of chlorine appeared to be that based on Stepanow's method'l i.e., reduction by metallic sodium in isopropanol; accordingly, this procedure was adopted for the collaborative tests by the working party.The preliminary work of preparing the solvent extracts and the details of the analytical procedures was undertaken by the members from the Government Laboratory. First, green cabbages known to be uncontaminated by chlorinated hydrocarbons were extracted EXPERIMENTAL WORK-June, 19571 TOTAL ORGANIC CHLORINE IN SOLVENT EXTRACTS OF VEGETABLE MATERIAL 379 with benzene to yield a heavily coloured extract. Since, however, benzene is an unsuitable solvent for the chromatographic separation included in the subsequent analytical procedure, this extract was evaporated to dryness and the residue dissolved in light petroleum.This solution was then washed with water to remove water-soluble chloride and was divided into three portions. To two of the portions known amounts of DDT were added, and the chlorine was determined in all three. Samples of the three solutions (I, I1 and 111) were distributed to the participating laboratories for chromatographic separation of the vegetable colouring matter and subsequent determination of the organic chlorine by both the mercuric oxycyanide RESULTS OF COLLABORATIVE TESTS- Results by the mercuric oxycyanide method were submitted by all four laboratories. Results by the turbidimetric method were submitted with confidence by two laboratories only. The fourth laboratory found this method unsatisfactory and did not submit any results.The collected results are shown in Table I. t 4 and the turbidimetric method. TABLE I SUMMARY OF RESULTS FOR THE ORGANIC CHLORINE CONTENT OF A LIGHT PETROLEUM EXTRACT OF GREEN CABBAGE CONTAINING KNOWN AMOUNTS OF DDT Chlorine* in plant material (less blank) r--.--------A-- , Chlorine Laboratory Method Solution I, Solution 11, Solution 111, in blank, P d g E*g/g P g l g P d g A Mercuric oxycyanide 3-7 3.4 17-1 } 0.1 { ;:; 18.25 0.1 0-86 19.6 (approx.) } O a l Turbidimetric 3.8 4.1 B Mercuric oxycyanide 3.86 17.62 None Turbidimetric 3.52 18.5 None { ::f3 3.0 16.8 None { ;:; C Mercuric oxycyanide 8.3 17.9 4.3 4.5 D Mercuric oxycyanide Chlorine added 3-65 18.3 None - * All results are the mean of two or more determinations.Each figure for Laboratory A is the mean from two different operators. CONCLUSION- From the results of the collaborative investigation, the working party came to the conclusion that the method of chromatographic separation followed by determination of the chlorine by the mercuric oxycyanide procedure was satisfactory for determining small amounts of organic chlorine derived from DDT, but it was not known how far this method would be applicable to other chlorinated hydrocarbons. Details of the recommended method are given in Appendix I. Grateful appreciation is expressed by the working party for the valuable help given by Dr. A. J. Feuell of the Colonial Products Laboratory and by Mr. G. A. Sergeant of the Government Laboratory. Appendix I DETERhlINATION OF TOTAL ORGANIC CHLORINE IN SOLVENT EXTRACTS OF VEGETABLE MATERIAL Nom-AlI reagents (including distilled water) and apparatus should be free from chloride, and care should be taken to prevcnt the access of chlorine or chlorine-containing compounds, e.g., hydrochloric acid, ammonium chloride or chloroform, from the atmosphere.CHROMATOGRAPHIC SEPARATION OF VEGETABLE COLOURING MATTER APPARATUS- narrow opening at one end. A glass twbe, of about 17 mm internal diameter and 20 cm long, drawn out to give a380 ANALYTICAL METHODS COMMITTEE: THE DETERMINATION OF SMALL AMOUNTS OF [VOl. 82 REAGENTS- whereby its absorptive efficiency is considerably reduced. dried at 200" to 300" C. Alumina-For chromatographic analysis. This may contain a small amount of water, If so, the material should be Benzene-"Crystallisable" grade.Light petroleum-Analytical-reagent grade, boiling range 40" to 60" C. This must be Sodium sulphate, anhydrous-Analytical-reagent grade. ('harge the tube with 8 g of alumina held in place by a wad of cotton-wool. free from aromatic hydrocarbons and chloride. PROCEDLJRE- Tap down the alumina to produce a level surface and then add a 1-inch layer of anhydrous sodium sulphate. Clamp the tube vertically. Pour in sufficient light petroleum to wet the column and then introduce the test solution, followed, at appropriate intervals, by four successive 10-ml washings of a mixture of four parts by volume of light petroleum and one part of benzene. Evaporate the eluate, which should be colourless, to about 10 ml on a hot-plate, assisting the operation by directing on to the surface of the liquid a gentle stream of air that has been passed through a suitable desiccant, e g ., silica gel, and then through a cotton-wool filter. The solution is then ready for reduction. STEPANOW SEMI-MICRO METHOD OF REDUCTION APPARATUS- A 50-ml round-bottomed jask attached to a small reflux condenser by a ground joint. REAGENTS- isoPropano1-Analytical-reagent grade, and a (1 + 1) aqueous dilution. Sodium-Pellets in liquid paraffin and uniformly free from chloride. If such a grade of sodium is not obtainable from the suppliers, a suitable grade for this method may be produced from that normally supplied by the procedure given in Appendix 111. Diethyl ether-Tested for freedom from water-soluble chloride.Nitric acid, diluted (1 + 2)-Analytical-reagent grade. Potassium hydroxide, 5 per cent. w/v-Rinse sticks of the analytical-reagent grade solid Hydrogen peroxide, 20-volume-Analyt ic al-reagen t grade. Phenolphthalein indicator solution-A 0.1 per cent. solution in 95 per cent. ethanol. PROCEDURE- Transfer the solution prepared by the chromatographic separation to the dry 50-ml flask and evaporate just to dryness by immersing the flask in water at 40" C and blowing in a gentle current of dry, filtered air. To the residue in the flask add 4 ml of isopropanol and approximately 0.3 g of sodium cut into several pieces. Attach the reflux condenser (dry) and boil for half an hour over a micro-burner (eg., a Bunsen burner from .which the top has been unscrewed); then, while continuing the boiling, slowly add 2 ml of diluted isopropanol(1 + 1).When all the residual sodium has disappeared, discontinue heating and rinse down the condenser with about 5ml of distilled water. Remove the flask, cool, add phenolphthalein indicator solution* and acidify with diluted nitric acid (1 + 2). Rinse the contents of the flask into a small separating funnel to a total volume of about 25 ml and add 10 ml of diethyl ether. Shake the mixture well, allow it to stand and then run the aqueous layer into a 50-ml calibrated flask. Wash the ethereal layer with 10ml of distilled water, and to the total aqueous extract in the flask add 5 drops of 20-volume hydrogen peroxide (to destroy any sulphide, see Note, p. 382), phenolphthalein indicator solution* and 5 per cent.w/v potassium hydroxide solution until alkaline. with distilled water before dissolving. Dilute the solution to 50 ml. * A modification of this procedure was used by one operator as follows: methyl red indicator was used instead of phenolphthalein when acidifying the solution after the treatment with sodium, and screened methyl red after the ether extraction.Jllne, 19571 TOTAL ORGANIC CHLORINE I N SOLVENT EXTRACTS OF VEGETABLE MATERIAL 381 MERCURIC OXYCYANIDE DETERMINATION OF CHLORIDE REAGENTS- Sulphuric acid, 0.002 N or 0.004 N.? Sulphuric acid, approximately 0.01 N. Potassium hydroxide solution, approximately 0.01 N. Screened methyl red indicator solution-Mix equal volumes of (a) 0.125 g of methyl red in 50 ml of 90 per cent. ethanol and (b) 0.083 g of methylene blue in 50 ml of 90 per cent.ethanol. Mercuric oxycyanide reagent-Dissolve 4 g of mercuric oxycyanide as completely as possible in 100 ml of distilled water, with stirring and moderate heating. Filter the solution through a close-grained filter-paper, cool, add a few drops of screened methyl red indicator solution and then add 0.01 N sulphuric acid until the solution assumes a grey tint. Standard potassium chloride soldion, 0.0002 N. PROCEDURE- Transfer a suitable aliquot (say, 25ml) of the 50ml of solution from the Stepanow reduction to a small evaporating basin and add almost sufficient 0.01 N sulphuric acid to decolourise the phenolphthalein. Evaporate the solution to dryness on a steam-bath or under an infra-red lamp, dissolve the residue in distilled water and rinse it into a 25-ml tall beaker.Add 2 drops of screened methyl red indicator solution, heat to boiling and add 0.01 N sulphuric acid until the solution is just acid. Boil to remove carbon dioxide and continue boiling to reduce the volume to about 5ml.* Titrate to a faint violet-grey tint with 0.002 N or 0.004 Nt sulphuric acid; then add 0.5 ml of mercuric oxycyanide reagent and titrate to the original tint with the 0.002 N acid from a 5-ml burette. BLANK- as described above. solution as is titrated in the test. STANDARDISATION OF THE SULPHURIC ACID WITH RESPECT TO CHLORINE- Place 0,5, 10,20,30,40 and 50-ml portions of 0.0002 N potassium chloride in evaporating basins. To each, add 1 drop of 0.01 N potassium hydroxide and evaporate to dryness, and then proceed as directed above.Titration values corrected for the blank (on the 50 ml from the blank Stepanow reduction) should, when plotted against volumes of standard chloride, show a very nearly linear relation- ship. This linear relationship does not hold so closely for the higher titration values if much less mercuric oxycyanide reagent, e.g., 0.15 ml, is added or if a more dilute acid is used. CALCULATION- If T ml of test solution require S ml of acid (after deducting T/50 of the volume of acid required by the whole of the blank) and S ml of the acid are, from the standardisation graph, equivalent to A ml of 0-0002 N potassium chloride, Cool the solution and make it just alkaline with 0.01 N potassium hydroxide. Take the whole of the final solution from the blank Stepanow reduction and proceed The final blank titration should be carried out on the same volume of 50 x A x 35.5 x 0.0002 x lo6 pg of chlorine.T x 1000 then the whole sample contains For a 25-g sample, the DDT content will be pg per g or p.p.m. 50 x A x 35.5 x 0.0002 x lo6 x 2 - - T x 1000 x 25 28.4 x A T Per g or P*P.m* - - * When evaporating the solution in the 25-ml beaker before the titration, some form of boiling-rod should be used t o promote uniform boiling. Especially useful for this purpose is a device consisting of a &inch square of &inch polytetrafluoroethylene sheet perforated and pushed on to the end of a 3-inch glass rod, the tip of which has been drawn out and finally melted to form a retaining knob. This reduces the dilution on titration, especially for the higher titration values and thereby sharpens the end-point. For the titration, 0.004 N sulphuric acid may be preferred to 0-002 N .382 ANALYTICAL METHODS COMMITTEE [Vol.82 NOTE-The addition of hydrogen peroxide may be omitted in the reduction stage, since any sulphide formed should be decomposed on boiling the acidified solution before titration in the mercuric oxycyanide determination. Appendix I1 ALTERNATIVE PROCEDURE FOR DETERMINING CHLORINE BY THE MERCURIC OXYCYANIDE METHOD REAGENTS- Sulphuric acid, approximately 3 N. Sulphuric acid, appyoximately 0.01 N. Sulphuric acid, 0*0005 N. Diethyl ether-Tested for freedom from water-soluble chlorides. Hydrogen peroxide, 20-volume-Analytical-reagent grade. Mercuric oxycyanide reagent-As described in Appendix I.Methyl red indicator solution-A 0.04 per cent. aqueous solution. PROCEDURE- Carry out the chromatographic separation and the Stepanow reduction as far as the elimination of excess of sodium. Acidify the resulting solution to methyl red with 3 N sulphuric acid and extract with diethyl ether. Run off the aqueous phase into the original flask, wash the ethereal layer with 10ml of water and add it to the main aqueous bulk. Add 5 drops of 20-volume hydrogen peroxide (see Note at end of Appendix I, above), make the solution just alkaline, boil, cool, neutralise with 0.01 AJ sulphuric acid and dilute to 50 ml. Transfer a 5-ml or 10-ml aliquot to a 50-ml narrow-necked conical flask and add 4 drops of methyl red indicator solution, followed by a slight excess of 0.0005 N sulphuric acid.Boil for 2 minutes to remove carbon dioxide and transfer the solution to a 25-ml wide-necked flask with distilled water free from carbon dioxide. Into another similar flask put the same quantity of distilled water free from carbon dioxide and of methyl red indicator solution. Adjust the blank solution with 0-0005 N sulphuric acid until slightly pink. Adjust the test solution to the same tint. Add 2 ml of the neutral 2 per cent. solution of mercuric oxycyanide to each ampoule and titrate the liberated sodium hydroxide with 0.0005 N sulphuric acid to the original tint, as shown by the blank, which should be diluted with distilled water free from carbon dioxide to the same volume as the titrated test solution.Appendix 111 PREPARATION OF CHLORIDE-FREE SODIUM Although the total chloride content of a sample of sodium may be small, the distribution of the chloride appears to be irregular, so that comparatively large amounts may be contained in the small portions used in this determinalion. By separating, as far as possible, the clean metal from the dross, a satisfactory grade of sodium can be obtained. The following procedure is recommended- Take about four pellets of sodium (9 to 12 g) and melt them in a 6-inch x l-inch Pyrex-glass test-tube under a few millilitres of liquid paraffin, B.P. Stir vigorously with a glass rod to separate dross and heat fairly strongly to help coagulate globules of sodium. Introduce into the molten sodium a heated glass tube about 1 cm in diameter, narrowed at the lower end to a jet about 1 mm in diameter and constricted at the top to take a length of rubber tubing. Draw up a quantity of clean sodium and transfer it to a clean test-tube containing liquid paraffin. To get the cleaned sodium in a convenient form, re-melt it, insert a piece of clean 5-mm bore glass tubing wetted internally with liquid paraffin and draw up about 2 inches of the molten metal, which, after cooling and solidifying, can be extruded as required with the aid of a glass rod. REFERENCES 1. 2. Viebock, I?., Heit., 1932, 65, 496. 3. 4. Klein, A. K., and Wic.imann, H. J., J . A s s . Off, Agvic. Clietn., 1946, 29, 191. Ingram, G., Analyst, 1944, 69, 365. Belcher, K., Macdonald, A. M. G., and Nutten, A. J., Miwoclzem. Acta, 1954, 104.

ISSN:0003-2654    

DOI:10.1039/AN9578200378

出版商:RSC    

年代:1957

数据来源: RSC

摘要:

June, 19571 EASTERBROOK : SODIUM CARBONATE AS A VOLUMETRIC STANDARD 383 Sodium Carbonate as a Volumetric Standard* BY W. C. EASTERBROOK Sodium carbonate has frequently been proposed as a working standard for acidimetry, but objections have been made t o its use on the grounds of alleged lack of purity and stability. In spite of this, sodium carbonate prepared by heating sodium sesquicarbonate is used by Imperial Chemical Industries Ltd. Recent experiments carried out on a thermobalance have confirmed that sodium carbonate prepared by heating sodium sesquicarbonate a t 270" C is perfectly stable over long periods a t this temperature. Although the preparation of sodium carbonate and the method of referring it to pure silver used as the ultimate volumetric standard have been described, no account has ever been given of the work that first led to the adoption of sodium carbonate as a working standard for Imperial Chemical Industries Ltd.or of the preparation and assay of subsequent batches of the material. This work is now described and includes the assay of materials containing 100.023 100.013 and 100-008 per cent. of sodium carbonate. It establishes beyond doubt the purity and stability, and hence the suitability as an acidimetric standard, of properly prepared sodium carbonate. A recent spectrographic analysis of the ultimate standard silver, to which the sodium carbonate was referred, has given a value of rather better than 99.998 per cent. of silver. SINCE early in the nineteenth century, sodium carbonate has been recommended as a volu- metric standard and, in spite of alleged lack of purity and stability when prepared by different methods, it is still widely approvedl for this purpose, The numerous papers published on the subject contain statements that appear to be contradictory, and these have been responsible for much of the uncertainty existing on the reliability of sodium carbonate as a reference standard in acidimetry.Amongst the chief difficulties reported in the literature have been the removal of residual water,2 y 3 the apparent instability on heating,4,5 the high solubility in water, which defeats easy purification, and the hygroscopic nature of the anhydrous ~ a r b o n a t e . ~ , ~ I t would seem that these objections may be associated with too severe or ill-defined experimental conditions, or both.Sodium hydrogen carbonate of high purity has been largely adopted as the source of anhydrous carbonate, because it is simply converted at relatively low temperatures. This procedure was recommended by Gay-Lussac and also by L ~ n g e , ~ , ~ , l O who showed that sodium carbonate obtained in this way was stable for long periods when heated in the region of 300" C, and his work is confirmed by the experimental work reported in this paper. There are, however, widely differing views in the literature, not only on the best temperature of conversion to the normal carbonate, but also on the composition of the final product, which is alleged to contain oxide or hydroxide.8,g,10,11,12,13,14 The temperature recommended for the conversion of sodium hydrogen carbonate to sodium carbonate \.aries over a wide range, namely from 102" to 450" C.14 At temperatures below 200' C some workers claim that the decomposition is incomplete, whereas at temperatures of 380" C or over it ~70uld appear that decomposition proceeds beyond the stage of normal carbonate, to carbon dioxide, water and sodium oxide or hydroxide.On these points there is some evidence to show that at lower temperatures complete conversion from the sesquicarbonat e or hydrogen carbonate to the normal carbonate is not so much a function of the temperature as of the duration of the heating period; for instance, both salts are transformed to the normal carbonate by exposure in a boiling-water oven for a period of about 50 hours. It would be difficult to reconcile all the conflicting views on this subject; however, the experimental work summarised below should help to dispel some of the many doubts that still exist in the minds of those who by tradition rely on sodium carbonate as a reference standard. * Presented a t the XVth International Congress on Pure and Applied Chemistry- (Analytical Chemistry), Lisbon, September 8th to 16th.1956.384 EASTERBROOK: SODIUM CARBONATE AS A VOLUMETRIC STANDARD [VOl. 82 THERMAL STABILITY AND HYGROSCOPICITY OF SODIUM CARBONATE DERIVED FROM SODIUM SESQUICARBONATE Sodium hydrogen carbonate is readily converted to sodium sesq~icarbonate,~ 915 which has obvious advantages as a starting material in preparing sodium carbonate. First, it is possible in this conversion to remove mechanical and other impurities and, secondly, the sodium sesquicarbonate so obtained is easier to handle than the finely divided sodium hydrogen carbonate from which it is derived. The loss in weight of sodium sesquicarbonate on sustained heating is not a reliable criterion on which to judge purity, because of the ease with which the material dissociates at relatively low temperatures.The chief factor of concern in the present instance was the stability of the sodium carbonate derived as a residue on heating sodium sesquicarbonate under defined and controlled temperature conditions. The constancy in weight of the sodium carbonate was obviously one of the simplest means of measuring this stability. Simple experiments were carried out by heating the sodium sesquicarbonate in platinum or silver capsules in an electrically heated air-oven at 270" C.The decomposition is indicated by the equation- 2Na2C0,.NaHC03.2H,0 = 3Na2C03 + 5H,O + CO,. The progressive change in weight and percentage loss in terms of time that sodium sesqui- carbonate and sodium hydrogen carbonate undergo when heated to 270" C are given in Table I. STABILITY OF SODIUM CARBONATE OBTAINED BY HEATING SODIUM SESQUICARBONATE AND SODIUM HYDROGEN CARBONATE AT 270°C Time of hours 2 6 10 14 30 46 heating, j Sodium sesquicarbonate Weight Loss Weight Losrj of in of in residue weight residue weig'ht Ln test 1, in test 1, in test 2, in test 2, 10-5158 29.89 10.5151 29.90 10.5169 29.88 10.5143 29.91 10.5173 29.88 10.5149 29.90 10,5157 29.89 10.5145 29.90 10.5152 29.89 10.5147 29-90 10.5161 29.89 10.5133 29-91 7 A 6 Yo g % Sodium hydrogen carbonate Weight Loss Weight Loss of in of in residue weight residue weight in test 1, in test 1, in test 2, in test 2, 9.4745 36.85 9.4729 36-86 9.4750 36.85 9.4736 36-86 0.4765 36.84 9-4752 36-84 9.4750 36.85 9.4741 36.85 9.4742 36.85 9.4737 36.85 9.4743 36.85 9.4732 36.86 7 A -I g Y O f3 Y O Weight of sample 14.9983 Theory Theory 15.0001 29.66 15.0036 15.0030 36.91 The losses in weight suffered by sodium hydrogen carbonate and sodium sesquicarbonate, respectively, remain constant during the heating time of from 2 to 46 hours.Although reasonable constancy in weight was found after 2 hours, small fluctuations that would appear to be of little practical importance persisted throughout the tests.It seemed reasonable, however, to carry out tests of a more critical character with the aid of a thermal balance, so that a continuous instead of an intermittent record of the weight could be obtained. The results obtained with sodium sesquicarbonate are shown in Fig. 1, which shows that dissociation with loss of carbon dioxide and water was virtually complete at a temperature of 210" C and within 2 hours of the commencement of heating. Thereafter, the chart readings indicate no detectable change. This is shown in Figs. 2 and 3 of the continuous record, that is, after exposure at 270" C for 6 to 8 hours and for 19 hours. The operating temperature of the heating chamber was then increased to 340" C. In the inter- vening periods no appreciable change in weight was apparent.The stability test was repeated with sodium sesquicarbonate that had already been converted to normal carbonate by exposure in an electrically heated air-oven at 270" C for 18 hours. By this means it was possible to increase the size of the sample under test from 2 to 5 g. The results obtained are shown in Fig. 4. Figs. 4 and 5 represent 6 to 8 hours and 18 to 20 hours, respectively, of additional heating, and again they show no significant change in weight. The temperature of the heating chamber was finally raised to 340" C. At that temperature an increase of 1 mg in the weight was recorded; this is attributed to the change in buoyancy with increase of temperature. A384 EASTERBROOK: SODIUM CARBONATE AS A VOLUMETRIC STANDARD [VOl.82 THERMAL STABILITY AND HYGROSCOPICITY OF SODIUM CARBONATE DERIVED FROM SODIUM SESQUICARBONATE Sodium hydrogen carbonate is readily converted to sodium sesq~icarbonate,~ 915 which has obvious advantages as a starting material in preparing sodium carbonate. First, it is possible in this conversion to remove mechanical and other impurities and, secondly, the sodium sesquicarbonate so obtained is easier to handle than the finely divided sodium hydrogen carbonate from which it is derived. The loss in weight of sodium sesquicarbonate on sustained heating is not a reliable criterion on which to judge purity, because of the ease with which the material dissociates at relatively low temperatures. The chief factor of concern in the present instance was the stability of the sodium carbonate derived as a residue on heating sodium sesquicarbonate under defined and controlled temperature conditions.The constancy in weight of the sodium carbonate was obviously one of the simplest means of measuring this stability. Simple experiments were carried out by heating the sodium sesquicarbonate in platinum or silver capsules in an electrically heated air-oven at 270" C. The decomposition is indicated by the equation- 2Na2C0,.NaHC03.2H,0 = 3Na2C03 + 5H,O + CO,. The progressive change in weight and percentage loss in terms of time that sodium sesqui- carbonate and sodium hydrogen carbonate undergo when heated to 270" C are given in Table I. STABILITY OF SODIUM CARBONATE OBTAINED BY HEATING SODIUM SESQUICARBONATE AND SODIUM HYDROGEN CARBONATE AT 270°C Time of hours 2 6 10 14 30 46 heating, j Sodium sesquicarbonate Weight Loss Weight Losrj of in of in residue weight residue weig'ht Ln test 1, in test 1, in test 2, in test 2, 10-5158 29.89 10.5151 29.90 10.5169 29.88 10.5143 29.91 10.5173 29.88 10.5149 29.90 10,5157 29.89 10.5145 29.90 10.5152 29.89 10.5147 29-90 10.5161 29.89 10.5133 29-91 7 A 6 Yo g % Sodium hydrogen carbonate Weight Loss Weight Loss of in of in residue weight residue weight in test 1, in test 1, in test 2, in test 2, 9.4745 36.85 9.4729 36-86 9.4750 36.85 9.4736 36-86 0.4765 36.84 9-4752 36-84 9.4750 36.85 9.4741 36.85 9.4742 36.85 9.4737 36.85 9.4743 36.85 9.4732 36.86 7 A -I g Y O f3 Y O Weight of sample 14.9983 Theory Theory 15.0001 29.66 15.0036 15.0030 36.91 The losses in weight suffered by sodium hydrogen carbonate and sodium sesquicarbonate, respectively, remain constant during the heating time of from 2 to 46 hours.Although reasonable constancy in weight was found after 2 hours, small fluctuations that would appear to be of little practical importance persisted throughout the tests. It seemed reasonable, however, to carry out tests of a more critical character with the aid of a thermal balance, so that a continuous instead of an intermittent record of the weight could be obtained. The results obtained with sodium sesquicarbonate are shown in Fig. 1, which shows that dissociation with loss of carbon dioxide and water was virtually complete at a temperature of 210" C and within 2 hours of the commencement of heating. Thereafter, the chart readings indicate no detectable change.This is shown in Figs. 2 and 3 of the continuous record, that is, after exposure at 270" C for 6 to 8 hours and for 19 hours. The operating temperature of the heating chamber was then increased to 340" C. In the inter- vening periods no appreciable change in weight was apparent. The stability test was repeated with sodium sesquicarbonate that had already been converted to normal carbonate by exposure in an electrically heated air-oven at 270" C for 18 hours. By this means it was possible to increase the size of the sample under test from 2 to 5 g. The results obtained are shown in Fig. 4. Figs. 4 and 5 represent 6 to 8 hours and 18 to 20 hours, respectively, of additional heating, and again they show no significant change in weight.The temperature of the heating chamber was finally raised to 340" C. At that temperature an increase of 1 mg in the weight was recorded; this is attributed to the change in buoyancy with increase of temperature. A384 EASTERBROOK: SODIUM CARBONATE AS A VOLUMETRIC STANDARD [VOl. 82 THERMAL STABILITY AND HYGROSCOPICITY OF SODIUM CARBONATE DERIVED FROM SODIUM SESQUICARBONATE Sodium hydrogen carbonate is readily converted to sodium sesq~icarbonate,~ 915 which has obvious advantages as a starting material in preparing sodium carbonate. First, it is possible in this conversion to remove mechanical and other impurities and, secondly, the sodium sesquicarbonate so obtained is easier to handle than the finely divided sodium hydrogen carbonate from which it is derived.The loss in weight of sodium sesquicarbonate on sustained heating is not a reliable criterion on which to judge purity, because of the ease with which the material dissociates at relatively low temperatures. The chief factor of concern in the present instance was the stability of the sodium carbonate derived as a residue on heating sodium sesquicarbonate under defined and controlled temperature conditions. The constancy in weight of the sodium carbonate was obviously one of the simplest means of measuring this stability. Simple experiments were carried out by heating the sodium sesquicarbonate in platinum or silver capsules in an electrically heated air-oven at 270" C. The decomposition is indicated by the equation- 2Na2C0,.NaHC03.2H,0 = 3Na2C03 + 5H,O + CO,.The progressive change in weight and percentage loss in terms of time that sodium sesqui- carbonate and sodium hydrogen carbonate undergo when heated to 270" C are given in Table I. STABILITY OF SODIUM CARBONATE OBTAINED BY HEATING SODIUM SESQUICARBONATE AND SODIUM HYDROGEN CARBONATE AT 270°C Time of hours 2 6 10 14 30 46 heating, j Sodium sesquicarbonate Weight Loss Weight Losrj of in of in residue weight residue weig'ht Ln test 1, in test 1, in test 2, in test 2, 10-5158 29.89 10.5151 29.90 10.5169 29.88 10.5143 29.91 10.5173 29.88 10.5149 29.90 10,5157 29.89 10.5145 29.90 10.5152 29.89 10.5147 29-90 10.5161 29.89 10.5133 29-91 7 A 6 Yo g % Sodium hydrogen carbonate Weight Loss Weight Loss of in of in residue weight residue weight in test 1, in test 1, in test 2, in test 2, 9.4745 36.85 9.4729 36-86 9.4750 36.85 9.4736 36-86 0.4765 36.84 9-4752 36-84 9.4750 36.85 9.4741 36.85 9.4742 36.85 9.4737 36.85 9.4743 36.85 9.4732 36.86 7 A -I g Y O f3 Y O Weight of sample 14.9983 Theory Theory 15.0001 29.66 15.0036 15.0030 36.91 The losses in weight suffered by sodium hydrogen carbonate and sodium sesquicarbonate, respectively, remain constant during the heating time of from 2 to 46 hours. Although reasonable constancy in weight was found after 2 hours, small fluctuations that would appear to be of little practical importance persisted throughout the tests.It seemed reasonable, however, to carry out tests of a more critical character with the aid of a thermal balance, so that a continuous instead of an intermittent record of the weight could be obtained.The results obtained with sodium sesquicarbonate are shown in Fig. 1, which shows that dissociation with loss of carbon dioxide and water was virtually complete at a temperature of 210" C and within 2 hours of the commencement of heating. Thereafter, the chart readings indicate no detectable change. This is shown in Figs. 2 and 3 of the continuous record, that is, after exposure at 270" C for 6 to 8 hours and for 19 hours. The operating temperature of the heating chamber was then increased to 340" C. In the inter- vening periods no appreciable change in weight was apparent. The stability test was repeated with sodium sesquicarbonate that had already been converted to normal carbonate by exposure in an electrically heated air-oven at 270" C for 18 hours.By this means it was possible to increase the size of the sample under test from 2 to 5 g. The results obtained are shown in Fig. 4. Figs. 4 and 5 represent 6 to 8 hours and 18 to 20 hours, respectively, of additional heating, and again they show no significant change in weight. The temperature of the heating chamber was finally raised to 340" C. At that temperature an increase of 1 mg in the weight was recorded; this is attributed to the change in buoyancy with increase of temperature. A384 EASTERBROOK: SODIUM CARBONATE AS A VOLUMETRIC STANDARD [VOl. 82 THERMAL STABILITY AND HYGROSCOPICITY OF SODIUM CARBONATE DERIVED FROM SODIUM SESQUICARBONATE Sodium hydrogen carbonate is readily converted to sodium sesq~icarbonate,~ 915 which has obvious advantages as a starting material in preparing sodium carbonate.First, it is possible in this conversion to remove mechanical and other impurities and, secondly, the sodium sesquicarbonate so obtained is easier to handle than the finely divided sodium hydrogen carbonate from which it is derived. The loss in weight of sodium sesquicarbonate on sustained heating is not a reliable criterion on which to judge purity, because of the ease with which the material dissociates at relatively low temperatures. The chief factor of concern in the present instance was the stability of the sodium carbonate derived as a residue on heating sodium sesquicarbonate under defined and controlled temperature conditions. The constancy in weight of the sodium carbonate was obviously one of the simplest means of measuring this stability.Simple experiments were carried out by heating the sodium sesquicarbonate in platinum or silver capsules in an electrically heated air-oven at 270" C. The decomposition is indicated by the equation- 2Na2C0,.NaHC03.2H,0 = 3Na2C03 + 5H,O + CO,. The progressive change in weight and percentage loss in terms of time that sodium sesqui- carbonate and sodium hydrogen carbonate undergo when heated to 270" C are given in Table I. STABILITY OF SODIUM CARBONATE OBTAINED BY HEATING SODIUM SESQUICARBONATE AND SODIUM HYDROGEN CARBONATE AT 270°C Time of hours 2 6 10 14 30 46 heating, j Sodium sesquicarbonate Weight Loss Weight Losrj of in of in residue weight residue weig'ht Ln test 1, in test 1, in test 2, in test 2, 10-5158 29.89 10.5151 29.90 10.5169 29.88 10.5143 29.91 10.5173 29.88 10.5149 29.90 10,5157 29.89 10.5145 29.90 10.5152 29.89 10.5147 29-90 10.5161 29.89 10.5133 29-91 7 A 6 Yo g % Sodium hydrogen carbonate Weight Loss Weight Loss of in of in residue weight residue weight in test 1, in test 1, in test 2, in test 2, 9.4745 36.85 9.4729 36-86 9.4750 36.85 9.4736 36-86 0.4765 36.84 9-4752 36-84 9.4750 36.85 9.4741 36.85 9.4742 36.85 9.4737 36.85 9.4743 36.85 9.4732 36.86 7 A -I g Y O f3 Y O Weight of sample 14.9983 Theory Theory 15.0001 29.66 15.0036 15.0030 36.91 The losses in weight suffered by sodium hydrogen carbonate and sodium sesquicarbonate, respectively, remain constant during the heating time of from 2 to 46 hours.Although reasonable constancy in weight was found after 2 hours, small fluctuations that would appear to be of little practical importance persisted throughout the tests.It seemed reasonable, however, to carry out tests of a more critical character with the aid of a thermal balance, so that a continuous instead of an intermittent record of the weight could be obtained. The results obtained with sodium sesquicarbonate are shown in Fig. 1, which shows that dissociation with loss of carbon dioxide and water was virtually complete at a temperature of 210" C and within 2 hours of the commencement of heating. Thereafter, the chart readings indicate no detectable change. This is shown in Figs. 2 and 3 of the continuous record, that is, after exposure at 270" C for 6 to 8 hours and for 19 hours. The operating temperature of the heating chamber was then increased to 340" C.In the inter- vening periods no appreciable change in weight was apparent. The stability test was repeated with sodium sesquicarbonate that had already been converted to normal carbonate by exposure in an electrically heated air-oven at 270" C for 18 hours. By this means it was possible to increase the size of the sample under test from 2 to 5 g. The results obtained are shown in Fig. 4. Figs. 4 and 5 represent 6 to 8 hours and 18 to 20 hours, respectively, of additional heating, and again they show no significant change in weight. The temperature of the heating chamber was finally raised to 340" C. At that temperature an increase of 1 mg in the weight was recorded; this is attributed to the change in buoyancy with increase of temperature.AJune, 19571 385 blank test made under the same conditions showed the same response. The maximum variation of the blank was estimated to be about & 0.25 mg and is attributed chiefly to inherent variability of the instrument. It is concluded from these tests that sodium carbonate obtained from sodium sesqui- carbonate by heating under controlled conditions at 270" C for a period of 2 to 20 hours is completely stable at that temperature. The sodium carbonate prepared in this way is very slightly hygroscopic at ordinary temperatures and under ordinary conditions. The magnitude of any error that might arise from this cause was examined in the following way- Weigh the quantity of working standard into a silver or platinum crucible, and heat it to 270" 10" C in an electrically heated oven until it is constant in weight. Transfer the crucible to a stoppered weighing bottle of suitable size, close the bottle and place it in a desiccator to cool thoroughly.Transfer the bottle to the balance case, momentarily release the stopper of the weighing bottle and, after an interval of 30 minutes, weigh it accurately. Empty the contents of the crucible into a dry 500-ml conical flask, placing the crucible well inside the neck of the flask in order to avoid loss. Replace the crucible immediately in the bottle, transfer the whole to the balance case and weigh it after an interval as before. When the weight of the weighing bottle with contents has been found to be constant, the time necessary to transfer the sodium carbonate to the conical flask was estimated at about 30 seconds.The increase in weight sustained by the weighing bottle with capsule and sodium carbonate was in the interval about 30 pg, so that any uptake of moisture by the empty capsule and weighing bottle would be a fraction of this weight. In an experiment 4.5 g of working-standard sodium sesquicarbonate were heated over- night in a silver capsule at 270" C. The capsule was then cooled in a weighing bottle as described above. The results are shown in Table 11. For the first hour the increase in weight is fairly constant at about 0-00003 g per 30 seconds; after this the rate of increase falls off. The procedure adopted in weighing the sodium carbonate precludes any increase in weight that could significantly affect the analytical result.EASTERBROOK : SODIUM CARBONATE AS -4 VOLUMETRIC STANDARD TABLE I1 HYGROSCOPICITY OF FRESHLY HEATED WORKING-STANDARD SODIUM CAKBONATE PREPARED BY HEATING SODIUM SESQUICAKBONATE Test No. Details of test Weight, g Remarks 1 Weighing bottle with capsule and carbonate 54.961 12 Weighing a t ,&minute intervals 2 were weighed 54.961 18 3 54.961 17 Initial weighing after- 4 The stopper of the wcighing bottle was 54.96117 4 minute 5 removed and the wcighings were con- 54.96120 1 minute ti tinuetl 54.96123 1; minutes 7 54-96 1 2 8 2 minutes 8 54.96 132 2; minutes 9 54.963 137 3 minutes 1 0 54.96 166 10 minutes 11 54.96197 15 minutes 12 54.9643 1 hour I3 54.9657 2 hours 14 54.9658 3 hours 15 54.9664 4$ hours SCHEME OF STANDAKL)ISATIOPJ OF VOLUMETRIC SOLI-TIONS SODIUM CAKBONATE AS A REFERENCE STANDARD IN ACIDIMETRY- Unfortunately, there is insufficient systematic investigational work in the literature to justify reasonably the basis of a general scheme of standardisation with reference and ultimate standards, although such a scheme was recommended by Wagner at the Fifth International Congress in 1903.He suggested that reference and ultimate standards should be examined in two or more different ways, for example, acidimetrically and oxidimetrically. The success of such a plan would necessarily depend on the ultimate standard being of high386 EASTERBROOK: SODIUM CARBONATE AS A VOLUMETRIC STANDARD [vol. 82 purity. This principle was in effect put into practice in 1912 by Dr.E. G. Beckett in the Research Laboratories of Nobel’s Explosives Cio. Ltd., when the initial part of the Imperial Chemical Industries Ltd. scheme of standardisation was first evolved. It was then decided that the substance that best fulfils the exacting and critical properties of an ultimate standard issilver. However, it is admitted that silver has certain disadvantages, namely it cannot be used directly for the standardisation of acids, except hydrochloric acid, and even in this instance the standardisation would only be correct if the acid were entirely free from its salts and from other acids. Further, it cannot be used directly for the standardisation of oxidising or reducing substances. To overcome these disadvantages, it was decided to employ “working standards” to be used for the ,standardisation of laboratory solutions direct.As a working standard for acids and alkalis, pure sodium carbonate was selected and its strength was determined by titration with pure hydrochloric acid, the strength of which had been ascertained by comparison with the ultimate standard silver. SELECTION OF SODIUM CARBONATE AS WORKING STANDARD- carbonate and two sources were available for its preparation, namely, At this stage it had been decided to use sodium sesquicarbonate as the source of anhydrous ( a ) from pure sodium hydrogen carbonate (“sodium carbonate l”), (b) from commercially pure anhydrous sodium carbonate (“sodium carbonate 2”). Preparation of sodium carbonate l-The purest grade of sodium hydrogen carbonate available was added in small quantities to water at 86” C in a resistance-glass flask until no more would dissolve.The solution was then rapidly filtered and cooled thoroughly and the mother-liquor was poured off. The crystals of sodium sesquicarbonate were ground in a mortar and sucked dry in a Hirsch funnel. ‘The moist powder was dried in a porcelain basin and heated on a water bath until the powder had a perfectly dry appearance. It was powdered again, well mixed and sealed in resistance-glass ampoules. Preparation of sodium carbonate 2-An alternative method for the preparation of sodium carbonate from pure anhydrous carbonate was also examined. After recrystallisation, the salt was dissolved in water and purified carbon dioxide gas was passed into the solution at 0” C until no further gas was absorbed.The sodium hydrogen carbonate was isolated and it was then converted to the sodium sesquicarbonate by the method referred to above. The purity of the two sodium sesquicarbolnates and the hydrogen carbonate and an- hydrous carbonate from which they were derived, after heating to 270” C, was determined by reference to ultimate standard silver. The true weights of sodium carbonate and silver were obtained by using a high-precision balance with a sensitivity of 0.01 mg. Weighings were made by the method of substitution with weights recently calibrated to class “A” accuracy on a mass basis at the National Physical Laboratory. The results are given in Table 111. TABLE I11 PURITY OF SODIUM CARBONATES OBTAINED THROUGH SODIUM SESQUICARBONATE FROM SODIUM HYDROGEN CARBONATE AND SODIUM CARBONATE BY EVALUATION WITH ULTIMATE STANDARD SILVER Purity of sodium carbonate after heating to 270” C until constant in 99-988 100.003 99-996 Description of sample weight, yo Sodium hydrogen carbonate used in the preparation of sodium sesquicarbonate Sodium carbonate 1 derived from sodium sesquicarbonate .. . . .. 100*000 Sodium carbonate (anhydrous) used in preparation of sodium sesquicarbonate odium carbonate 2- ( a ) derived from sodium sesquicarbonate . . . . .. .. .. 99.976 ( b ) derived from sodium sesquicarbonate after recrystallisation . . .. 99.974 The results in Table I11 show that- (a) sodium carbonate 1 (sesquicarbonate) and the sodium hydrogen carbonate from which it was derived gave carbonates of high purity, and (c) derived from sodium sesquicarbonate mother-liquor .. . . . . 99.972June, 19571 EASTERBROOK SODIUM CARBONATE AS A VOLUMETRIC STANDARD 387 sodium carbonate 2 (sesquicarbonate) prepared from pure anhydrous sodium carbonate was slightly less pure than the original anhydrous carbonate from which it was derived. On account of this slight fall in purity, the sodium carbonate (sesquicarbonate) was recrystallised and both crystals and residue from the mother- liquor were tested as described above, but the results failed to show an improvement. This may be explained by small amounts of silica dissolved from the glass vessels used in recrystallisation. The experiments, therefore, favoured the method in which high-grade sodium hydrogen carbonate was further purified and transformed to sodium sesquicarbonate before conversion to normal carbonate by heating to 270" C, and this method was finally adopted in the preparation of working-st andard sodium carbonate.During the preparation of a new lot of sodium sesquicarbonate the individual batches, the blended batches and the bottled material were exhaustively tested by weight titrations with N hydrochloric acid to ensure homogeneity. In order to assess the purity of the sodium carbonate derived from the heat treatment of sodium sesquicarbonate at 270" C, pure silver was used as the ultimate standard. This procedure has been described in detail in the 1iterature.l Briefly, it involves the weight titration of working-standard sodium carbonate with N hydrochloric acid previously standardised gravimetrically against the ultimate standard silver.This method has been employed for the evaluation of several lots of sodium carbonate prepared from different batches of sodium sesquicarbonate. The results of the evaluations are shown in Tables IV and V. TABLE IV PURITY OF WORKING-STANDARD SODIUM CARBONATE : SERIES 1 (a) Determination of the ratio of silver to N hydrochloric acid by using ultimate standard silvev and chemically pure silver, precipitants being 0.1 N- True weight of Weight of silver Expt. silver required True weight of equal to 100 g No. by N HC1, N HCl, of N HC1, Ultimate standard 5 9.80898 92.2569 10- 6322 6 9.72351 91-4705 10-6302 Ultimate standard 7 8.10283 76.2103 10-6322 Chemically pure. Guaranteed Silver used g g g Mean = 10.6315 minimum- purity 99-99 per cent.(b) Determin.ation of the ratio of sodiuiii carbonate to N hydrochloric acid and purity of sodium cavbon&- Equivalent weight of Expt. NO. 1 2 3 4 8 9 10 True weight of I Na,CO,, g 3.63071 3.68262 3.69363 3.64518 5.25585 5.24120 5-25910 True weight of N HC1 -equired by Na,CO,, 6 69.5332 70.5230 70- 7 376 69.8081 100.6529 100.3753 100.7 198 Weight of Na,CO, equal to 100 g of N HCI, g 5.22154 5.22187 5.22159 5.22171 5.22155 5-22 170 5.22152 Mean weight of silver equal to 10og of N HC1, g 10.6315 10.6315 10.6315 10.6315 10.6315 10.6315 10.6315 Weight of Na,CO, equivalent to 107.88 g of silver, g 5 2.9 84 ( 0) 5 2.9 87 (4) 52.984( 6) 5%985( 8) 52.984( 1) 52.985(7) 52- 9 83 ( 8) NCCO, from Inter- national Atomic Weights (1925) 52.997 52.997 52.997 52-997 52.997 52.997 52.997 Mean = Purity of 100.025 100.018 100.024 100.02 1 100.024 100.021 100.025 100.023 From time to time during the interval covering the period of the experiments involving the determination of the actual sodium carbonate present, the standard hydrochloric acid was checked against the silver standard and the results given in the Tables show that no change occurred. The experiments described in Table IV were carried out with 0.1 N reagents, whereas those in Table V were obtained with reagents of normal strength.This was done in order to look for the possibility of adsorption of silver nitrate by silver chloride during the standardisation of the hydrochloric acid. The procedures recommended by388 EASTERBROOK: SODIUM CARBONATE AS ,4 VOLUMETRIC STANDARD [VOl.82 Richards and WellslG for dilution, order of addition of precipitants and so on were followed throughout this work. Comparison of the results shown in Tables IV and V indicates that there is no evidence that the results were affect,ed by adsorption of silver nitrate. Table V also shows that there is no evidence of difference in the strength of hydrochloric acid resulting from storage in wax-lined containers or resistance-glass bottles. TABLE V PURITY OF WORKING-STANDARD SODIUM CARBONATE : SERIES 2 (a) Determination of the ratio of silver to N hydrclchloric acid by 24sing chemically pure silver, precipitants being N- I<xpt True weight of silver Weight of silver equal so. 11 7.93567 7352884 10-7926 12 8.07762 74.83967 10.7932 13 8.23127 7 13.2 6 7 80 10.7926 14 8.28278 7 6.7 463 7 10.7924 Mean = 10.7927 required by N HC1, True weight of AV HCl, g g g t o 100 g of N HCI, (b) Determination of the ratio of sodium carbona,te to N hydrochloric acid and purity of sodium carbonate- Equivalent weight of Na,CO, weight of Na,CO, wleight of Na,CO, national True N HC1 equal t o silver equal equivalent Atomic Purity True Weight of Mean Weight of from Inter- Expt.weight of required by 100 g of to 100 g t o 107.88 g Weights of No. Na,CO,, Na,CO,, N HC1, of N HCl, of silver, (1925) Na:CO,, g g g g g /o 15 5.23072 98.67292 6.30107 10.7927 52*987(6) 52.997 100.018 I6 5.23038 98.6591 1 5.30148 10.7927 52.991(7) 52.997 100.0 10 17 5.22948 98.64379 5.30138 10.7927 52*990(7) 52.997 100.0 12 18 5.22374 98.72326 8.30143 10-7927 52.991(2) 52.997 100*011 Mean = 100.013 In experiments No.11, 12, 15 and 16 the acid was stored in wax-lined bottles. I n experiments No. 13, 14, 17 and 18 the acid was stored in resistance-glass bottles. TABLE VI Expt. No. 22 23 24 25 PURITY OF WORKING-STANDARD SODIUM CARBONATE : SERIES 3 Determination of the ratio of silver to N hydrochloric acid by w i n g ultiwaate standard silver- Expt. True weight of silver VE7eight of silver equal No. required by N HC1, True weight of N HC1, g g 6 t o 100 g of N HC1, 19 9.85422 92.0528 10.7060 20 9.83063 9 1.8282 10.7 055 21 9.94792 92.9289 10.7049 Mean = 10-7051 Determination of the ratio of sodium carbonate to X hydiochloric m i d and purity of sodiunz cagtbonate- Equivalent weight of Na,CO, weight of Na,CO, weight of Ka,CO, national True N HC1 equal to silver equal equivalent Atomic Purity weight of required by 100 g of to 100 g t o 107.88 g Weights of Na,CO,, Na,CO,, N HCI, of iV HCl, of silver, (1925) Na,CO,, b" g g 6 g Yo 3.64248 67.3714 5.25814 10-7061 52.988(0) 52-997 100*01 G 3.49 7 02 66.5003 5.25865 10.7051 52*993(7) 52.997 100.006 True Weight of Mean Weight of from Inter- 3.53805 67.2790 5.25879 10.7051 52.996(2) 52-997 100.004 3.56779 67.8450 5.25874 10.7051 52-994(6) 52.997 100.005 Mean = 100.008June, 19571 EASTERBROOK SODIUM CARBONATE AS A VOLUMETRIC STANDARD 389 The mean values of the purity of the sodium carbonate from Tables IV and V are 100.023 & 0.003 and 100.013 This leads to the conclusion that some oxide or hydroxide is formed during the decomp~sitionl~ of the sodium sesqui- carbonate, but the amount is not sufficiently large to make any significant difference when the material is used as a standard.As Table I and the charts on thermal stability show that prolonged heating of sodium carbonate at 270" C does not result in any appreciable decom- position, it seems reasonable to conclude that the slight decomposition that occurs (accounting for the slightly high results, see Tables IV (a) and ( b ) ) takes place during the decomposition of the sesquicarbonate (or hydrogen carbonate), oxide or hydroxide only being produced while moisture is being lost. The dry sodium carbonate is then completely stable to heat at 270" C. The results of tests for the purity of a further lot of sodium carbonate are shown in Table VI.For effective comparison, the results in Table VI have been calculated with the same atomic weights (1925) as those of the earlier figures. If the atomic weight of carbon, however, is taken as 12.01, i.e., the 1954 value, the level of the purity of sodium carbonate would be raised by 0.009 per cent., and the mapitude of this increase is of the order of precision of the results, namely 5 0.007 per cent. if the three lots are treated as one population in the statistical sense. The results, in general, show that sodium carbonate obtained by heating sodium sesquicarbonate to 270" C is consistently of high purity.ls~lg Samples of sodium sesquicarbonate representing the material from Series 3 were sub- mitted to three independent laboratories for examination.The detailed methods employed were the same in each instance. The results returned by each of the laboratories are given in Table VII. 0.005 per cent., respectively. PURITY OF WORKING-STANDARD SODIUM CARBONATE DETERMINED BY THREE INDEPENDENT LABORATORIES EVALUATION WITH ULTIMATE STANDARD SILVER Expt. No. 1 2 3 4 5 6 Purity of sodium carbonate from sodium sesquicarbonate, yo Evaluation by laboratory 100.03 99-96 99.95 99-98 100.03 100.03 Mean = 100.00 These results confirm the values in Table VI. ULTIMATE STANDARD SILVER The pure silver prepared by electrolytic deposition from chemically pure silver has a It was assumed to have, a purity of 100.00 per cent. in the evaluation of the various Four methods were used to check the purity, namely- (a) direct gravimetric comparison with chemically pure silver of guaranteed minimum purity of 99.99 per cent.of silver, through N hydrochloric acid; the purity of ultimate standard silver was deduced to be 0.01 per cent. higher than that of the silver from which it was prepared, (b) chemical examination for impurities before and after purification ; this showed that the original silver contained 0.0007 per cent. of copper and some carbon, but these two impurities could not be detected in the purified silver, determination of the equivalent of working-standard iodine used for oxidimetry, which gave a value of 126.93 compared with the International Atomic Weight of 126.92, and (d) recent spectrographic examination, which gave the total impurities as not more than 0.002 per cent. high degree of purity, probably 99.995 per cent. lots of sodium sesquicarbonate. (c)390 CABELL AND SMALES: THE DETERMINATION OF RUBIDIUM AND CAESIUM I N [VOl. 82 The thermal balance was lent by the courtesy of Messrs. Stanton Instruments Limited. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 16. 16. 17. 18. 19. REFERENCES Analytical Chemists’ Committee of I.C.I. Ltd.; Analyst, 1950, 75, 577; see also Strouts, C. R. N., Gilfillan, J. H., and Wilson, H. N., “Analytical Chemistry, The Working Tools,” Clarendon Press, Oxford, 1955, Volume 1, Chapter 12. Carmody, W. R., I n d . Eng. Chem., Anal. Ed., 1945, 17, 578. Richards, T. W., and Hoover, C. R., J . Amer. Chem. SOC., 1915, 37, 97. Smith, G. F., and Hardy, V. R., J . Chem. Educ., 1933, 10, 507. Smith, G. F., and Croad, G. F., I n d . Eng. Chem., Anal. Ed., 1937, 9, 42. Sebelien, J., J. Chem. Ztg., 1905, 29, 638. Kolthoff, I. M., Pharm. Weekblad, 1926, 63, 37. Lunge, G., 2. angew. Chem., 1897, 10, 522. Kunz-Krause, H., and Reichter, R., Arch. Pha:mz., 1917, 255, 540. Linder, J., and Figala, N., 2. anal. Chem., 1933, 91, 105. Waldbauer, L., McCann, D. C., and Tuleen, L. F., I n d . Eng. Chem., AnaE. Ed.: 1934, 6, 336. Reinitzer, B., 2. angew. Chem., 1894, 7, 551. Richards, T. W., and Wells, R. G., J . Amer. Ciiem. SOC., 1905, 27, 459. Schmitt, K. O., 2. anal. Chem., 1927, 70, 321. “P.V.S. Reagents for Volumetric Standardisation,” Hopkin & Williams Ltd., Chadwell Heath, Strouts, C. R. N., Chem. 6 Ind., 1956, 346. -, Ibid., 1904, 17, 231. -, Ibid., 1905, 18, 1520. -- , , J . Chem. S O ~ . , 1919, 116, 423. Essex, 1955, p. 7. RESEARCH DEPARTMENT IMPERIAL CHEMICAL INDUSTRIES LTD. NOBEL DIVISION, STEVENSTON, AYRSHIRE October loth, 1966

ISSN:0003-2654    

DOI:10.1039/AN9578200383

出版商:RSC    

年代:1957

数据来源: RSC

摘要:

390 CABELL AND SMALES: THE DETERMINATION OF RUBIDIUM AND CAESIUM IN [Vd. 82 The Determination of Rubidium and Caesium in Rocks, Minerals and Meteorites by Neutron-activation Analysis BY M. J. CABELL AND A. A. SMALES A method is described for the determination of rubidium and caesium in rocks, minerals and meteorites by neutron-activation analysis. Except that simple precautions must be taken .to eliminate the possibility of self- shielding of samples and standards during neutron irradiation, the results are not subject to interference by other constituents of the sample. After irradiation radiochemical separations with carriers present are based on ferric hydroxide scavenges, cobaltinitrite precipitations and cation-exchange chromatography. The purified rubidium and caesium fractions are finally recovered for counting as their chloroplatinates. The samples examined have varied from 3.4 per cent.to 4 x 10-6 per cent. in rubidium content and from 7.5 per cent. to 1 x per cent. in caesium content. They include an international i nter-comparison suite of lepidolites, the two international standard rocks W1 and G1, samples from the Skaergaard intrusion of East Greenland and several meteorites. Except at the lower end duplicate or tiriplicate determinations of rubidium contents usually agree within 5 per cent. and often within 2 per cent. Agree- ment between the results for caesium is rather less good. When it has been possible to compare results with those obtained by other methods, agreement has been excellent. Each determination requires 15 to 300 mg of sample.UNTIL quite recently accurate determinations of the concentrations of the alkali metals rubidium and caesium in many geological samples, especially when these concentrations are very small, have not been possible, because of the lack of suitable analytical methods. This limitation is particularly regrettable with rubidium, since the reliability of the rubidium - strontium method of geological-age determination depends upon determining rubidium and strontium with the highest possible accuracy.June, 19571 ROCKS, MINERALS AND METEORITES BY NEUTRON-ACTIVATION ANALYSIS 391 In general, two methods have been employed for determining small amounts of the elements. The first, emission-spectrum analysis, which has been used by Ahrens1v2 and others, suffers from the disadvantage that it depends on accurately known standards of composition very similar to the sample being available.The second, isotope-dilution analysis with enriched isotopes and the mass spectrometer, is undoubtedly superior from the points of view of both sensitivity and accuracy-a coefficient of variation of the results of 1 to 2 per cent. is attainable. I t has been applied recently to rubidium determinations by Davis and Aldri~h,~ Tomlinson and Das Gupta? Herzog and Pin~on,~ Pinson, Herzog, Backus and Cormier,6 Schumacher,’ Wetherill, Tilton, Davis and Aldrichs and by Webster and Smales.@ Its application to the determination of caesium is complicated, but by no means eliminated, by the necessity for using a radioactive caesium isotope, since naturally occurring caesium is monoisotopic.So far, however, no geochemical applications for caesium determinations have been reported. Neutron-activation analysis is another technique that is applicable to the determination of traces of rubidium and caesium. It has been used for their determination in sea-water, coal and sea-weed by Smales and Salmon,lo and Cabell and Thomasll have used it for their determination in sodium - potassium liquid-metal alloys, but no general application in geo- chemistry has been reported so far. The advantages of the method, however, great sensitivity, specificity and freedom from contamination troubles, make it very suitable for the deter- mination of traces of a large number of elements, and it has been chosen for this work.An outline of the application of the technique to geological specimens has been given by Vincent and Smales12 and need not be expanded further. FEASIBILITY OF THE NEUTRON-ACTIVATION METHOD- subjected to neutron bombardment are given in Table I. The nuclear characteristics of the isotopes involved when rubidium and caesium are TABLE I NUCLEAR DATA FOR RUBIDIUM AND CAESIUM Abundance Isotopic of nuclide activation Product of Target in natural cross-section, neutron nuclide element, barns irradiation % 8SRb 72.2 0.85 s6Rb 87Rb 27-8 0.14 s8Rb 3 134mCs 133CS 100.0 26 134cs Radiation and energy Half-life of product of product nuclide, nuclide /I- 1.8 19.5 days P- 5 18 minutes i- :a13 3- 1 hours 8- 0.66 2-3 years y 0.60, 0.79 MeV y 1.1 For practical purposes the short half-lives of 88Rb and the metastable state of l34Cs make them unsuitable for use when fairly lengthy chemical procedures are necessary to ensure radiochemical purity of the product. But B6Rb and 2.3-year lacs are quite suitable under these conditions and have been used exclusively in this work.It can be calculated that, by using these two radioactive nuclides and an irradiation time of 4 weeks in a flux of 10l2 thermal neutrons per sq. cm per second, i.e., that of the Harwell Pile, the limit of detection under normal counting conditions is about 6 x 10-log for both caesium and rubidium. EXPERIMENTAL SAMPLING, STANDARDS AND IRRADIATION- The samples for analysis were ground to less than 200-mesh B.S. sieve whenever possible. I t proved difficult to grind lepidolite samples as fine as this, but small thin flakes proved to be quite suitable.Suitable quantities (10 to 250 mg) were weighed into small dry silica ampoules, which had been manufactured from silica tubing having an internal diameter of 4 or 6 mm, as described by Smales and Loveridge,13 and the ampoules were immediately sealed. Weighed quantities (100 to 200 mg) of standard solutions were treated similarly, and samples and standards were packed side by side and irradiated in the “self-serve” positions of the Harwell Pile (see Cabell14) for periods of up to 4 weeks, or in the “rabbit” position for short periods.392 CABELL AND SMXLES: THE DETERMINATION OF RUBIDIUM AND CAESIUM IN [Vd. 82 Preliminary investigations showed that it was necessary to use solutions as standards.The chlorides are the only commercially available salts of caesium and rubidium that are sufficiently pure and thermally stable to serve as solid standards for the elements, but the chlorine nucleus has an average absorption cross-section of 32 barns, which is large enough for the solid chlorides to exhibit the phenomenon of self-shielding, i e . , reduction of the effective neutron flux progressively through them, so that there is unequal irradiation of standard and sample. Of even greater importance with caesium is the fact that the caesium nucleus has several large neutron resonance absorption peaks, including one with a maximum of 2500 barns for neutrons of 6 eV, which is at least partly within the spectrum of the neutrons used for irradiations.The effect of this self-shielding is shown in Table 11, in which the specific activity of different weights of solid caesium chloride is compared with that of a solution that was irradiated under identical conditions and for the same time. The bottom of the phials used in irradiations is conical rather than flat and it can be seen that the specific activity of the solid decreases as the phials are filled up until it reaches a constant value when the addition of more solid merely increases the height of solid in the phial without affecting the average diameter of the solid presented to the neutron beam. TABLE I1 EFFECT OF SELF-SHIELDING ON THE SPECIFIC ACTIVITY OF CAESIUM CHLORIDE Physical state of sample Liquid Solid Solid Solid Solid Solid Weight of caesium chloride in sample, mg 0.153 5.3 14.5 "4.4 42.9 45-9 Concentration of Specific activity of caesium, solution, counts per minute mg Per Per mg 1.000 34,870 27,000 -.26,800 24,700 _. 24,900 24,800 - By contrast, Table I11 shows that when the caesiuni chloride has been diluted one- hundredfold with water, further dilution, even another hundredfold, does not produce any significant increase in specific activity. Rubidium chloride in the solid state and in solution exhibits the same phenomena, but to a less mxked extent. TABLE I11 EFFECT OF DILUTION ON THE SPECIFIC AC'IIVITY OF CAESIUM CHLORIDE SOLUTIONS Concentration of caesium chloride solution, mg per g 0.09994 0.3998 0.9994 0.9994 3.976 10.17 Specific adivity of caesium, counts per minute per mg 35,500 35,000 35,600 35,700 34,800 35,100 hverage 35,300 Coefficient of variation 1.0 per cent.The same two solutions were used as standards throughout this work. Specpure caesium and rubidium chlorides were dried to constant weight in platinum dishes at 170" C and a solution of each containing approximately 10 mg per g was made up accurately. The rubidium content of the caesium chloride and the caesium content of the rubidium chloride were determined by neutron activation of the standard solutions. Two deter- minations of each showed that there were 45 pg of rubidium per g of caesium chloride and 35 pg of caesium per g of rubidium chloride. As rubidium and caesium attain approximately the same specific activity during irradiations of 1 day to 4 weeks, it was not necessary in the determinations either to separate the rubidium contained in the caesium chloride standards (or vice versa) or applv a correction for it, as its concentration was far too low to affect the result.June, 19571 ROCKS, MINERALS AND METEORITES BY NEUTRON-ACTIVATION ANALYSIS 393 DISSOLUTION AND SEPARATION OF THE ALKALI METALS- After irradiation it is necessary to get the sample into solution with the carrier and in the same chemical form.The sodium peroxide sinter method of Seelye and Rafter15 was used for this, but heating for a longer time and at a slightly lower temperature than those authors used was found necessary to obtain a good sinter and complete dissolution of beryls and lepidolites. Carriers were added to the samples directly after opening the irradiation tubes and before sintering.No special step was necessary to ensure chemical equilibrium between carrier and tracer, since the alkali metals can only exist as univalent cations in aqueous solutions. The rubidium and caesium were separated from the other elements in the sample and from each other in two stages. In the first stage the alkali metals were collected together and freed from other constituents of the sample that had been made radioactive, and most of the sodium, by ferric hydroxide scavenges of the solution and precipitation of the alkali-metal cobaltinitrites. In the second stage the rubidium and caesium were separated from each other, and from potassium and residual amounts of the other alkali metals, by cation-exchange chromatography.The separation of the alkali metals by cation-exchange chromatography has been demonstrated before, notably by Kayas,lG but a more rapid procedure than any of those that have been published so far was required. The resin chosen was Zeo-Karb 315, which was known to be particularly favourable for alkali-metal separations. Two separate batches of the resin were used. For the preparation of batch B sufficient “Normal” grade Zeo-Karb 315 was taken to form a column 1 metre long and 3-8 cm in diameter, and this was washed over a period of 2 days with 24 litres of 3 M hydrochloric acid to convert it to the hydrogen form and then with 5 litres of water. The resin was dried by exposure to the air, ground in a coffee grinder and sieved. The fraction passing through a sieve of aperture 0.179mm and caught on a sieve of aperture 0.125 mm was collected and freed from “fines” by repeated decantation with water.Excess of water was poured off and the resin was washed twice with industrial methylated spirit and then with ether. Finally it was dried in a current of air and stored in a tightly stoppered bottle. Batch A was treated in a similar manner, but was ground in a porcelain ball-mill. This method of grinding takes longer and possesses no advantage for this application over the use of a coffee grinder. Although batches A and B were nominally of the same material, they had in fact been supplied by the makers at different times, and investigation showed that they possessed different properties. Weighed portions were dried to constant weight in vaczto at 80” C and were found to contain different proportions of water, but, even allowing for this, potentio- metric titration in an atmosphere of nitrogen showed that their capacities were appreciably different.Hence 1 g of vacuum-dried resin from batch A required 2.25 milli-equivalents of sodium hydroxide for titration to pH 7.0, whereas 1 g of batch B required 2-51 milli-equiva- lents. Likewise their behaviour in ion-exchange chromatography differed and, although equal quantities from the same batch behaved reproducibly, the same quantity from the other batch required different volumes of eluent for a complete separation. The separation of the alkali metals was studied by using synthetic mixtures of radio- active sodium, potassium, rubidium and caesium chlorides in proportions that approximated to those arising at the ion-exchange stage from genuine sample solutions.They were pre- treated as if they were genuine sample solutions and their elution from an ion-exchange column under different conditions was studied by monitoring the eluate for gamma radiation, It was found that sodium was completely removed by the pre-treatment and that the separa- tion could be effected in one working day if 0.1 M hydrochloric acid was used to elute the potassium, and then 0-5M hydrochloric acid for the rubidium fraction and 1.0 M hydrochloric acid for the caesium fraction. Fig. 1 shows the successive elution of the alkali metals from a column made by packing 25.0 g of resin A in a tube of approximately 1 cm diameter. Fig.2 shows a similar separation on a column made from 21.6 g of resin B in a tube of the same diameter-this occupied almost exactly the same volume as 25-0 g of resin A. Based on experiments such as these, it was possible t o set conditions under which the alkali metals of sample solutions could be separated Their swelling properties on addition of water were also different.394 CABELL AND SMALES: THE DETERMINATION OF RUBIDIUM AND CAESIUM I N [VOl. 82 into pure fractions, even though the separati'on could not be followed by monitoring the eluate as described above, since the radioactivity of sample solutions is too small for this to be practicable. The actual conditions arrived at are set out in Table IV. 100- e, .- c =I x 75- U .- -", .$50- m .- +I ; 25- W 0.5 M HCI Time, hours I I I I I 1 1 Feed 0.5 M HCI I I t b Elution with 0.1 M HCI - - - _ I, I.0M HCI-June, 19571 ROCKS, MINERALS AND METEORITES BY NEUTRON-ACTIVATION ANALYSIS 395 cobaltinitrites are unsuitable for the gravimetric determination of the elements, whereas the chloroplatinates are easily obtained as stoicheiometric compounds.The cobaltinitrites were therefore dissolved in nitric acid, and ethanol and then a solution of hexachloroplatinic acid were added. The precipitate was washed with ethanol, transferred to a weighed counting tray, dried under an infra-red lamp and weighed. TABLE IV CONDITIONS USED FOR THE SEPARATION OF THE ALKALI METALS ON COLUMNS OF ZEO-KARB 315 Resin A Resin B Weight of resin taken, g . . .. .. .. .. .. 25-0 Feed solution, ml .. .. .. .. .. . . . . 5, plus a 5-ml Flow rate for applying feed, ml per hour . . .. . . .. 30 Flow rate used throughout elution, ml per hour . . .. . . <150 Volume of rejected potassium fraction, ml of 0.1 M hydrochloric acid 700 100 Volume of rejected mixed rubidium and caesium fraction, ml of Volume of pure caesium fraction, ml of 1.0 M hydrochloric acid . . wash with water Volume of pure rubidium fraction, ml of 0.5 M hydrochloric acid . . 0.5 M hydrochloric acid . . .. .. . . . . . . 100 150 21.6 5, plus a 5-ml wash with water 30 < 150 900 100 120 200 This procedure has other advantages. First, any sodium from the added sodium hydroxide that had been co-precipitated with the cobaltinitrite is dissolved in the ethanol and removed under these conditions and, secondly, the alkali-metal chloroplatinates are dried more easily.Whereas Duval17 has shown that caesium hexachloroplatinate that has been precipitated from aqueous solution does not lose all its water until heated to 200" C, we found that precipitation from ethanolic solution and washing with ethanol gave a product that was stoicheiometric and anhydrous on being heated to 80" C and did not lose weight thereafter, at least up to 370" C. Rubidium and caesium chloroplatinates are easily slurried evenly into aluminium counting trays and form compact adhering cakes that are most suitable for counting. The chemical yields obtained in practice varied between 50 and 70 per cent. for samples and between 80 and 90 per cent. for standards. MEASUREMENT OF RADIOACTIVITY- The alkali-metal chloroplatinates were counted by using an EHM2 thin-window Geiger tube and a conventional Geiger beta - gamma-counting assembly.At least 104 counts were recorded whenever practicable. The sources used were 5 sq. cm in area and it was shown that no self-absorption or back-scattering correction was necessary for rubidium chloro- platinate sources that were less than 10 mg per sq. cm in thickness; greater thicknesses could probably be tolerated without a correction being necessary, but these were never used. No self-absorption or back-scattering corrections were necessary with caesium chloroplatinate sources either, provided that their thickness was within the range 4 to 10 mg per sq. cm, as it always was in practice. The combined effects of self-absorption and back-scattering on the counting-rate of caesium chloroplatinate are seen in Table V.The apparent specific activity, as measured by Geiger counting, of sources of varying thicknesses prepared from a common stock is recorded. The difference in apparent specific activity between sources of thickness 4.4 mg per sq. cm and those of thickness 9-7 mg per sq. cm is less than 3 per cent. Sources from samples and standards used in analyses seldom differed by as much as 2 mg per sq. cm in thickness, so the error due to these factors in using the two counting rates as a direct measure of their alkali-metal contents would nearly always be less than 1 per cent., i.e., less than the statistical error involved in counting either of them. TABLE V COMBINED EFFECTS OF SELF-ABSORPTION AND BACK-SCATTERING ON CAESIUM CHLOROPLATINATE SOURCES Thickness of source, mg per sq.cm . . .. . . 1.29 1.44 2-15 2.61 2.89 4.40 5.48 5.63 6.62 9.67 Specific activity of source, counts per minute per mg 74.5 75.6 77.1 78-1 78-6 80.6 79.8 79.3 78.3 78-4 Another correction that must be considered when rubidium chloroplatinate sources of very low beta - gamma activity are being counted is the radioactivity of the 87Rb, which396 CABELL AND SMALES: THE DETERMINATION OF RUBIDIUM AND CAESIUM IN [VOl. 82 comprises 27.8 per cent. of the naturally occurriing rubidium used as the carrier for the traces of radioactive rubidium in the sample. A source of known weight of rubidium chloroplatinate prepared from natural rubidium was kept for this purpose and this was counted directly after samples for which a correction was thought to be necessary.With this source placed within 1 mm of the Geiger tube, the count recorded was only 2-2 counts per minute per mg of rubidium. Under the most unfavourable counting conditions used in practice, i.e., with the sources within 1 cm of the Geiger tube, the correction never amounted to as much as double the counter background correction ; usually it was quite insignificant. CONFIRMATION O F RADIOCHEMICAL PURITY- The preliminary experiments on the separation of the alkali metals by ion exchange showed that pure alkali-metal fractions could he obtained without difficulty. In applying the information obtained from these experiments to the sample solutions, large safety factors were allowed in deciding which fractions of the eluate to accept and which to reject, so that any small variations in column behaviour would be allowed- for and not affeci the of the product.Nevertheless an essential part of the neutron-activation method confirmation of the radiochemical purity of the products. purity is the Fig. 3. Absorption curves for caesium: curve A, pure caesium standard; curve B, caesium separated from a sample; curve C, caesium and added rubidium (ratio of beta activity of rubidium t o beta activity of caesium is 0.005) ; curve D, caesium and added rubidium (ratio of beta activity of rubidium to beta activity of caesium is 0.014) Decay curves are possible for rubidium and have been used by Smales and Salmon,lo but ideally they require four half-lives, i.e., 80 days, for certainty of radiochemical purity.They are not, of course, practicable for caesium. Beta-absorption curves are suitable for the confirmation of caesium purity and have been used in this work. Typical examples are displayed in Fig. 3 and show clearly that the beta particles of maximum energy produced by caesium just fail to penetrate an absorber having a thickness of 240mg of aluminium per sq. cm as expected for a beta particle of energy 0-66MeV. The Figure also shows the curves produced when small amounts of rubidium activity are deliberately added to pure caesium sources. It is clear that if rubidium impurity were present in caesium products it could be detected with considerable sensitivity in this way, but none has ever been found in practice.June, 19571 ROCKS, MINERALS AND METEORITES BY NEUTRON-ACTIVATION ANALYSIS 397 Beta-absorption curves have also been used to confirm rubidium radiochemical purity.The most energetic beta particles produced by rubidium just fail to penetrate an absorber having a thickness 880 mg of aluminium per sq. cm and this is clearly shown in the examples displayed in Fig. 4, but the presence of caesium impurity in even quite large amounts could escape detection very easily, as is demonstrated by another curve in the Figure, which was obtained by deliberately adding radioactive caesium to a pure rubidium source. For this reason a close similarity of the beta-absorption curves obtained from sample and standard is not a very reliable test for rubidium radiochemiml purity.0 Fig. 4. Absorption curves for rubidium: curve A, pure rubidium standard; curve B, rubidium separated from a sample; curve C, rubidium and added caesium (ratio of beta activity of caesium to beta activity of rubidium is 0.15) B cs r- D I l l 1 0.33 38 I I , Gamma energy, MeV Fig. 6. Gamma-ray spectra for rubidium: curve A, pure rubidium standard; curve B, rubidium and added caesium (ratio of beta activity of caesium to beta activity of rubidium is 0.006); curve C, rubidium and added caesium (ratio of beta activity of caesium to beta activity of rubidium is 0.011); curve D, rubidium and added caesium (ratio of beta activity of caesium to beta activity of rubidium is 0.057)398 CABELL AND SMALES: THE DETERMINATION OF RUBIDIUM AND CAESIUM I N [VOl.82 In contrast to this, the presence of a small amount of caesium impurity in a rubidium source is readily shown by the gamma spectrum of the mixture. The reason for this is that for each beta particle produced by the disintegration of a 134Cs nucleus approximately two gamma photons are also produced, whereas the pr'oduction of a beta particle by a disintegrating ssRb nucleus is accompanied by the production of a gamma photon only once in about five times. Moreover, each gamma photon produced. by a disintegrating lSCs nucleus is of lower energy than a corresponding gamma photon from a s6Rb nucleus and is recorded with approximately twice the sensitivity by the gamma spectrometer (where the height of the gamma peak is taken as the criterion of sensitivity).These two effects reinforce one another so that when sources of lacs and ssRb of equal1 beta activity are examined in the gamma spectrometer the caesium source produces a gamma peak many times the height of that given by the rubidium source. Fig. 6. Gamma-ray spectra for rubidium : curve A, rubidium sample separated from a lepidolite; curve B, pure rubidium standard Gamma energy, MeV Fig. 7. Gamma-ray spectra for caesium: curve A, caesium sample separated from a lepidolite; curve B, pure caesiurn standard This is seen in Fig. 5, where the gamma spectrum obtained from a pure rubidium source is compared with those obtained from similar rubidium sources to which caesium impurity has been added in known amounts. It can be seen that, when the beta activity of the caesium is only 0-6 per cent.of the rubidium beta activity (both being measured by a Geiger counter before mixing), the gamma spectrum still shows this small amount very clearly. Fig. ti shows the gamma spectra obtained from some rubidium separated from an irradiated lepidoliteJune, 19571 ROCKS, MINERALS AND METEORITES BY NEUTRON-ACTIVATION ANALYSIS 399 and that due to a standard rubidium source; no evidence of caesium impurity can be detected. For comparison purposes the gamma spectra of a pure caesium source and a caesium sample separated from a lepidolite are displayed in Fig. 7. METHOD PROCEDURE- Grind the sample for analysis to less than 200-mesh B.S. sieve or as finely as practicable. Weigh a suitable quantity (10 to 250 mg) into a dry silica ampoule and, if necessary (see later), add a diluent, seal the tube and mix the contents thoroughly.Prepare some clean dry transfer pipettes and rub the tips with a small piece of cotton-wool that has been dipped in a dilute solution of dimethyldichlorosilane in carbon tetrachloride. Wipe the tips clean and use the pipettes to transfer 100 to 200-mg amounts of standard solutions of Specpure rubidium and caesium chloride of accurately known concentration (less than 1 per cent.) to weighed ampoules. Mark the ampoules for identification, pack the samples together with duplicate standards for each alkali metal, and irradiate them in the Harwell Pile for suitable periods of up to 4 weeks. After irradiation allow the ampoules to “cool” sufficiently before handling. Weigh accurately approximately 40 mg of Specpure caesium chloride and 40 mg of Specpure rubidium chloride into a platinum crucible and cover with a thin layer of AnalaR sodium peroxide.Take a sample ampoule, tap it until the contents fall to the bottom, scratch the constriction with a diamond and open the ampoule. Invert the tube and tap it until the contents fall into the crucible and then add more sodium peroxide and mix the contents of the crucible thoroughly with a piece of flattened platinum wire. Cover with a thin layer of sodium peroxide (of which the total amount used need not exceed 4 to 5 g). Cool the crucible by dipping the outside in water, invert it and tap the bottom so that the sinter falls into a 250-ml conical beaker. Fill the crucible with water to dissolve any of the sinter remaining and pour the solution into the beaker.Repeat until the sinter is covered with water. Boil until the sinter is dissolved and no more bubbles are evolved. Cool, add 2 ml of a solution of ferric chloride in hydrochloric acid (10 mg of iron per ml) , mix and cool again. Now rapidly add 10-ml amounts of concentrated hydrochloric acid at a time, swirling the contents and cooling the beaker after each addition, until a clear yellow solution is obtained. Spin the ampoules containing the standard solutions in a centrifuge until the liquid is collected at the bottom of each tube. Open them in the same way as the samples, but observe caution. Considerable pressure is released as the ampoules are broken , although this does not result in any loss. Without loss of time transfer the contents to a 100-ml calibrated flask with a transfer pipette and with the same pipette thoroughly wash both parts of the ampoule with dilute hydrochloric acid, transferring the washings to the flask. Add two drops of concentrated hydrochloric acid and make up to the mark. Take suitable aliquots for direct use or dilution.Weigh again and seal. Heat the crucible in the furnace for 1 hour at, but not exceeding, 470” C. Cover with a watch-glass and allow to cool. PREPARATION OF THE SAMPLE FOR THE COLUMN- Treat the sample solution in portions of about 40 ml in separate centrifuge tubes. To each add 5 drops of phenolphthalein and then 6 M sodium hydroxide dropwise until the solution is just alkaline. Acidify the filtrate with glacial acetic acid, add 1 ml of ferric chloride solution (10 mg of iron per nil) and again make alkaline with 6 M sodium hydroxide.Acidify the filtrate with acetic acid, add 10 ml of a freshly prepared 10 per cent. solution of sodium cobaltinitrite, stir and set aside for at least 5 minutes in an ice bath. Spin in a centrifuge and wash three times with water. Collect all the cobaltinitrite precipitates from the same sample together, remove any supernatant liquid and add 2 ml of diluted nitric acid (1 + 1). Heat with shaking over a bunsen flame until the precipitate has dissolved. Transfer the solution to a small beaker and remove all the water and acid by evaporation under an infra-red lamp until a blue tarry residue is left. Dissolve in 5 ml of water and filter if the solution is not clear.The solution is now ready for feeding to the column. Spin in a centrifuge and filter. Spin in a centrifuge and filter.400 CABELL AND SMALES: THE DETERMINATION O F RUBIDIUM AND CAESIUM I N [VOl. 82 THE ION-EXCHANGE SEPARATION- I t has been emphasised already that each batch of resin must be treated on its merits. Prepare a batch of Zeo-Karb 315 (if this is not available a resin with similar cation-exchange properties may be used) of particle size such that it passes through a No. 85 B.S. sieve and is caught on a No. 120 B.S. sieve. Determine the weight of resin required such that its volume is approximately 36 ml when prepared for use, and transfer this amount to a glass tube 1 cm in diameter that has been shaped to form a column of a conventional kind. Dissolve .40 mg of rubidium chloride, 40 nng of caesium chloride, 190 mg of potassium chloride and 60 mg of cobalt as chloride in 5 ml of water and feed this solution to the column at a flow-rate not exceeding 0.5 ml per minute. Allow the liquid level to fall to the top of the column and then add 5 ml of washings, and again allow the level to sink to the top of the column.Using a reservoir of 0.1 M hydrochloric acid supported above the column and feeding it by a siphon flow, elute at a flow-rate as near to, but not exceeding, 2.5 ml per minute as possible. A total pressure of about 17'0 cm of liquid above the bottom of the column will be required. Determine the change in concentration of potassium in the eluate by a suitable method and continue elution until all the potassium has been removed.Now elute with 0.5 M hydrochloric acid at a flow-rate not exceeding 2.5 ml per minute and determine the rubidium elution curve. Finally use 1.0M hydrochloric acid to recover the caesium. From the results obtained in this way devise a scheme for obtaining pure rubidium and caesium fractions from the sample solutions, .making allowance for the rejection of cross- contaminated fractions. Prepare similar colum.ns for these separations. After each caesium fraction has been recovered regenerate the column for the next sample solution by passing 1500 ml of 3 M hydrochloric acid through it over a period of 8 hours, followed by water until the pH of the eluate is 4-0 or greater. FINAL PURIFICATION AND RECOVERY- Evaporate to dryness on the hot-plate and dissolve the residue in 20 ml of hot water.Transfer to a centrifuge tube and cool. Add 2 drops of phenolphthalein, make alkaline with 6 M sodium hydroxide, spin in a centrifuge and filter (the caesium fraction will contain cobaltic hydroxide). Make the filtrate acid with acetic acid and add 10 nil of a freshly prepared 10 per cent. solution of sodium cobaltinitrite. Set the solution aside for at least 5 minutes in an ice bath while the precipitate forms, then spin in a centrifuge and wash the precipitate three times with distilled water. Add 2 ml of diluted nitric acid (1 + 1) and heat over a bunsen flame until the precipitate has dissolved. Cool. Add another 2 ml of diluted nitric acid and then 15 ml of anhydrous industrial methylated spirit and 1 ml of a 10 per cent.solution of hexa- chloroplatinic acid. Stir and set the solution in an ice bath for 10 minutes while the pre- cipitate forrns, then spin in a centrifuge and wash the precipitate three times with 74 O.P. industrial methylated spirit. Make a slurry of the precipitate with a few drops of industrial methylated spirit and transfer it to a weighed aluminium tray, dry under an infra-red lamp, cool and re-weigh in order to determine the cjhemical yield. Treat both the rubidium and the caesium fractions as follows. STANDARDS- Add carriers to the standards and treat them in exactly the same way as the sample solutions throughout. If, however, they have been analysed and it is known that the rubidium chloride does not contain more than traces of caesium and vice versa, the ion-exchange step and the second cobaltinitrite precipitation maty be omitted. Arrange that the standards and samples for counting are of approximately the same weight and between 5 and 10 mg per sq.cm in thickness. COUNTING- Count the precipitate of caesium chloropla,tinate or rubidium chloroplatinate for beta - gamma activity, using an end-window Geiger tube of the GM4 or EHM2 type, and count the standards directly afterwards under identical conditions. Obtain a total count of at least lo4 counts whenever practicable and make use of lower geometry and absorbers so that the counting rate does not exceed 5000 counts per minute. Correct the measured activities for background, counter dead-time and, with rubidium, the natural activity of the rubidium carrier.June, 19571 ROCKS, MINERALS AND METEORITES BY NEUTRON-ACTIVATION ANALYSIS 401 Calculate the amount of alkali metal in the sample x, by means of the relationship- Activity of x from sample corrected for 100 per cent.recovery Activity of x from standard corrected for 100 per cent. recovery' - - Mass of x in sample Mass of x in standard using the conversion factors- 10 mg of Rb = 14.14 mg of RbCl 10 mg of Cs z 12.68 mg of CsCl 10 mg of Rb == 33.86 mg of Rb,PtCl, 10 mg of Cs EZ 26-35 mg of Cs,PtCl, Note that when the sample or standard contains a weight of alkali metal significant in comparison with that used for the carrier this must be taken into account when calculating the chemical yields. Finally confirm the radiochemical purity of the products by means of half-life deter- mination, beta-absorption curves and gamma-spectra determinations as appropriate. MINERALS- Table VI records the results obtained for the alkali-metal contents of an International Intercomparison Suite of Lepidolites that had been prepared at the Massachusetts Institute of Technology and supplied by Dr.L. H. Ahrens, Department of Geology and Mineralogy, University of Oxford. For comparison purposes the rubidium figures obtained by two groups of workers by isotope-dilution analysis are also recorded. The results for two other lepidolites that were collected by Prof. H. Bassett in Tanganyika are also given. RE s ULT s TABLE VI RUBIDIUM AND CAESIUM CONTENTS OF AN INTERNATIONAL SUITE OF LEPIDOLITES, TWO OTHER LEPIDOLITES AND PRECIPITATED SILICA Sample Sample and origin weight, mg International intercomparison suite- L-N45, Bikita Quarry, 24.95 S.Rhodesia 24.84 GAlO6, Bob Ingersoll Mine, S. Dakota L-,4107, Pala Mine, S. California L-A109, along Winnepeg River, S.E. Manitoba LA110 ( b ) , Varutrask, Sweden 21.89 28.40 29.44 28.48 15.20 23.47 19.54 20.97 26-70 23.36 20.81 21-62 Rubidium found, Yo 3.40 3.42 1-21 1-29 1.34 - 1-74 1.77 2-23 2.24 2.26 - 1-58 1.62 Caesium found, % 0.290 0.302 0.146 0.137 0.139 0.154 0.248 0.259 0.170 0.164 0.163 0.176 0.173 0.179 Average rubidium found, % 3.41 1.28 1.76 2-24 1.60 Average caesium found, % 0.296 0.144 0.253 0.168 0.176 Rubidium found by Pinson et a1.,18 % 3.29 1.26 1.39 2.08 1.85 Rubidium found by Aldrich et C Z Z . , ~ ~ % 3.3 1 1.21 1.23 1.16 1.76 2.20 2.05 Other lepidoldes- B1, Hombola, 20.59 1-15 0.293 1-16 0.290 - - Tangan yika, 14.22 1.16 0.288 collected 9/11/51 Tanganyika, 18.63 2.38 0.839 B2, Zati, near Hombola, 22.87 2.35 0.831 2.37 0.835 - - collected 24/7/52 Silica- Specpure precipitated 296.6 7 x lW7 A X 5 x 6 x - - silica, No.8141 321.6 5 x 5 x 297.0 4 x lo-' 6 x With minerals such as lepidolites it was thought necessary to take precautions against Lepidolites may contain up to 5 per cent. by weight of lithium self-shielding in the samples.402 CABELL AKD SMALES: THE DETERMINATION OF RUBIDIUM AND CAESIUM I N [VOl. 82 and, since the absorption cross-section of the GLi nucleus is 950 barns and ‘jLi constitutes 7-5 per cent. of natural lithium, this concentration is sufficient to increase the mean absorption cross-section for each atom in a lepidolite by a factor fifty to two hundred times that of an atom in an igneous rock. The samples that are given in Tables VI and VII were therefore diluted (after weighing) at least ten times by weight (and many more times by volume, which is the important thing) with Specpure precipitated silica.The silica was analysed separately by neutron-activation analysis with the results also given in Table VI. Its content of caesium and rubidium was too low to make a correction of the results for the minerals necessary. Table VII records the results obtained for a series of minerals supplied by Dr. J. M. Axelrod of the United States Department of the Interior, Geological Survey. Dr. Axelrod has analysed the same samples by X-ray spectroscopy and by flame photometry.His results are also recorded in Table VII for comparison. For X-ray spectroscopy he claims a sensitivity of 0.02 per cent. of rubidium oxide or 0.1 per cent. of caesium oxide and an accuracy of 10 per cent. of the amount present or the sensitivity, whichever is larger, TABLE VII RUBIDIUM AND CAESIUM COXTENTS OF SOME I1lIP;EIIALS Rubidium Caesiuni 7-----h----CT ---J--------, oxide found by-- oxide found 1))- -- Sample Sample and origin weight, mg 143024, lepidolite, from 16.91 Stewart’s Mine, Pala, 23.05 California 143025, lepidolite, from 42.00 San Diego Mine, Mesa 18.94 Grande 143029, lepidolite, froni 19-98 Foote Mineral Co., 20.35 144206, lepidolite, froin 22.58 Stewart’s Mine, Pala, 20.10 California 143950, white beryl, 17.62 from Le Grange, 14.06 Georgia 143062, morganite, from 17.51 Madagascar 143006, microcline from 16-06 Tourmaline Queen 19.25 Mine, Pala, California 143948, beryl, from 29.72 Bumpus Quarry, 51.79 Albany, Maine 5 1 A-32-2 .. Iiubi- Cac- neutron neutron dium siuiii activa- X-ray flame activa- X-ray oxide oxide tion spectro- photo- tion suectro- found, found, (average), <copy,* % 1.50 1-49 1.12 1.12 3.13 3.23 1-34 1-47 0.021 1 0.0214 0-0202 0.065 0.065 0.0066 0-0066 % 0.81 0.81 0.362 0.364 0.288 0.29‘3 0.082 0.084 0.152 0.149 7.52 0.0084 0.0080 0.054 0.055 % 1.51 1.12 3.20 1.47 0.02 0.02 - 0.01 % 1.58 1-21 3.4 1.40 < 0.01 <on01 0.10 <0-01 Getry,* (average), Scopy,* flame met r y , * 0.8!) photo- 70 0.38 0.37 0.06 0.12 7.6 0.00 < 0.05 * Figures supplied by Dr.J. M. Axelrod, U.S. Geological Survey. ROCKS- The results of analyses of the typical basic and acidic rocks Wl and G1, which have been suggested as international standards for the determination of both major and minor con- stituents of igneous rocks, are given in Table VIII. The rubidium results are compared with those of workers at A.E.R.E., Harwell (Webster and Smalesg), and the Massachusetts Institute of Technology (Herzog and Pinson5) both of whom used isotope-dilution analysis. The new values for rubidium are the result of further work since the Note by Smales20 and are preferred to those quoted there, which were, for G1, 221, 254 and 243 p.p.m. and, for W1, 27, 20 and 26 p.p.m.June, 19571 ROCKS, MINERALS AND METEORITES BY NEUTRON-ACTIVATION .4N.41dYSIS 403 TABLE VIII RUBIDIUM AND CAESIUM CONTENTS OF THE INTERNATIONAL STANDARD ROCKS W1 and G1 Sample Sample and origin weight, mg G1, granite from Westerly, 199.7 Rhode Island 194.0 246.1 W l , diabase from Center- 200.5 ville, Virginia 196.9 404.5 Rubidium CaesiuIn found, found, p.p.m.p.p.m. 217 1-51 219 1-48 222 1.54 20.6 1-03 2l*tj 1.1 3 21.8 1.07 Rubidium found by- A , Caesium isotope- isotope- found by neutron dilution dilution neutron activation method a t method a t activation I average), A. E. R.E ., M. I .T., (average), p.p.m. p.p.m. p.p,ni. p.p.m. 219 “14 216 1.5 r- A number of samples taken from various parts of the Skaergaard intrusion of East They were supplied by Prof. L. R. Wager and Dr. E. A. Vincent The results are given in Greenland were also analysed.of the Department of Geology and Mineralogy, University of Oxford. Table IX, but no comparison figures by other workers are available. TABLE IX RUBIDIUM AND CAESIUM CONTENTS O F SAMPLES FROM THE SIMERGAARD INTRUSION OF EAST GREENLAND Sample and origin EG3058, transgressive granophyre sill EG4489, transgressive granophyre sill EG4332, hedcnbergite granophyre EG4330, fayalite ferrogabbro . . EG4328, upper ferrogabbro . . EG5181, lower ferrogabbrn . . EGi427, middle gabbro . . . . EG6086, hyperstliene-oli\,~ac-gabbro EC1851, perpendicuIar feldspar rock EC4526, gabbro picrite . . . . EG4507, chilled marginal gahbro . . METEORITES- . . # . . . . . . . . . . . . . . . . . . . Sample weight, mg 171-4 170.5 196.2 169.4 161.0 205.9 102.7 101.3 105.1) 1043 223.0 521.7 201.2 103.4 104.9 104.8 102.2 102.1 104.0 247-7 1 !I 2.9 172.1 180.7 220.5 195.3 101.3 107.x Caesium content, p.p.m. 0.28 0.40 0.35 0.59 0.6 1 0.20 0.34 0.18 0.20 0.4 5 0-5 1 0.50 0.1 ,0.1 -: 0.1 . - : O n 1 <0.1 - _ _ 0-155 0.159 0.143 0-3135 0.363 0.09 0.1.’ - Average rubidium content, p. p . In. 96 47.2 42 21.9 12.5 2-4 1 .o -> .<> .). ’3 I . 3.f1 5.8 --lveragt: caesiuni content, 11.p.m. 0.34 0.60 0.32 0.1‘3 0.4!) 0.1 .--o. 1 0.1 0.1.5 0.3; 0.10 Several stony meteorites have also been analysed for their rubidium and caesiurn contents. The sample of “Homestead” meteorite had had 10-7 per cent. by weight of metal extracted, but otherwise drillings (about 1 g) of whole meteorite were taken and mixed before sampling, without any attempt being made to separate the dispersed metallic phase.The results are given in Table X.404 CABELL AND SMALES: THE DETERMINATION OF RUBIDIUM AND CAESIUM I N TABLE X RUBIDIUM AND CAESIUM CONTENTS OF SOME STONY METEORITES Sample and origin XMNH100,* from Forest City, Winne- AMNH1018, from Ness Co., Kansas . . bago Co., Iowa OUM2,t from Faha, Co. Limerick, Rep. Homestead,: from Homestead, Iowa Ireland AMNH2454, from Modoc, Scott Co., Kansas AMNH2399, from Long Island, Phillips Co., Kansas AMNH1043, from Bluff, Fayette Co., Texas AMNH2497, from Johnstown, Weld Co., Colorado Sample weight, g; 0.1:10 1 0.161 1 0.13L75 0.1 <C96 0.1903 0.1 ::82 0.1445 0.155 5 0.1 1.40 0.1032 0.2 6 3 7 0.1 ti77 0-14.85 0.24 39 0.1140 0.1239 0.1174 0.22 15 0.1165 0.1123 0.1306 0-1069 0.1750 0-1564 0.3392 0.2281 Rubidium content found, p.p.m.3.01 3.06 2.88 2.93 2.19 2.37 2.96 3.00 3.29 3.34 2.91 2.98 3.01 2.05 2.09 2.15 2-15 1.00 1.00 1-03 1.03 0.04 0.04 0.05 0.05 - Caesium content found, p.p.m. 0.101 0.096 0.036 0.041 0.099 0.099 0.09 0.06 0.06 0.09 0.08 0.07 0.07 0.02 0.01 0.01 0.01 0.0 1 0.01 - - 0.005 0.007 0.008 I Average rubidium content, p.p.m. 3-04 2.90 2.28 3.15 2.97 2.11 1.01 0.04 [Vol. 82 Average caesium content, p.p.m. 0.098 0.039 0.099 0.07 0.08 0.01 0-01 0.007 * Kindly supplied by Dr. Brian H. Mason, American Museum of Natural History, New York, pu’.y. t Kindly supplied by Dr. L. H. Ahrens, Department of Geology and Mineralogy, University of Oxford. $ Kindly supplied by Drs. W. H. Pinson and L.F. ~ e r z o g Department of Geology, Massachusetts Institute of Technology. INTERFERING ELEMENTS In neutron-activation analysis the possibility that radioactive nuclides of the elements being determined may arise by other nuclear ~xocesses than the one under consideration, must always be borne in mind. Of these (n,$), (%,a) and (n,f) reactions are the most likely. Hence, apart from the n,y reaction with 85Rb, “Rb may be produced by any of the following reactions- 86 38sr + + in -+ ;:Rb and 2izU + -+ fission products, including i;Rb (%f >- Reactions of a similar nature, which give rise to lacs and other radioactive caesium isotopes, are possible when barium, lanthanum and uranium are irradiated with neutrons. I f therefore any of these elements, Or strontium Or yttrium, are present in samples and such reactions do take place to any appreciable extent, they will, of course, increase the apparent n,y induced radioactivity due to alkali metals jn the samples and lead to results that are too high.TO determine whether such interferences are appreciable, known weights of Specpure barium and strontium carbonates, and lanthanum, yttrium and uranium oxides, were irradiated, together with rubidium and caesium standards, and analysed for their “apparent” alkali-metal contents by the method described above. The results are given in TablexI.June, 19571 ROCKS! MINERALS AND METEORITES BY NEUTRON-ACTIVATION ANALYSIS 405 TABLE XI YSEUDO-RUBIDIUM AND PSEUDO-CAESIUM CONTENTS OF ELEMENTS THAT GIVE RISE TO RADIOACTIVE NUCLIDES OF RUBIDIUM AND CAESIUM ON NEUTRON BOMBARDMENT, BY REACTIONS OTHER THAN n,y “Apparent” rubidium “Apparent” caesium Element irradiated content, Average, content, Average, p.p.m.p.p.m. p.p.n. p.p.m. Strontium . . .. 0.70 (0.1 - Yttrium . . .. Uranium . . . . .. Barium . . . . .. Lanthanum .. . . 0.12 i o - 1 - < 0.01 - 0-10 0.10 0.10 } The figures quoted include, of course, the real amounts of rubidium and caesium impurity in the samples so they only represent an upper limit on the “apparent” amounts of alkali metals that have been produced. With these figures and the abundances of the elements under consideration in rocks and stony meteorites, the extent of possible errors in alkali-metal determinations in this type of sample that can arise from unwanted nuclear reactions, can be determined.The figures are given in Table XI1 and show the errors to be quite negligible for all samples that have been examined. TABLE XI1 MAXIMUM ERRORS POSSIBLE IN THE DETERMINATION OF CAESIUM AND RUBIDIUM IN ROCKS AND STONY METEORITES DUE TO UNWANTED NUCLEAR REACTIONS Igneous rocks I E r r o r p o s s i b l e i n Average concentration determination of Element of element,* alkali-metal content, p.p.m. p.p.m. Strontium . . 300 2 x of Rb Yttrium . . 28.1 3 x of Rb Uranium . . 4 6 x lo-’ of Rb 9 x 10-4ofcs 13arium . . 250 2.5 x 10-3ofcs Lanthanum . . 18.3 2 x 10-5of c s Stony meteorites Average Error possible in concentration determination of of element,* alkali-metal content, p.p.m. p.p.m. 20 1 x of Rb 4.7 6 x lo-’ of Rb 0.36 6 x 10-80f Rb I A \ 8 x of Cs 1.58 1 x 10-6 of c s 6.9 7 x 10-5ofcs * From Rankama and Sahama.21 CONCLUSIONS The method described in this paper has been applied to samples containing as little as 6 x g of caesium, which, allowing for chemical yields and decay of the rubidium before counting, approaches the limit of detection of the method. At the other end of the scale samples containing as much as 1 rng of rubidium and 1.3 mg of caesium have also been examined.For amounts greater than 2 pg of either alkali metal in the sample, the coefficient of variation is 2 per cent. or less for the rubidium results and less than 3 per cent. for caesium results. Below an estimated 2 pg of alkali metal the precision becomes progressively worse. All of the many products that have been examined have been found to be radiochemically pure.Making allowance for possible errors in standards, differences in weight between standards and samples for counting, etc., it seems likely that the results quoted for samples containing 10 p.p.m. or more of either alkali metal are within 5 per cent. of the absolute value, and possibly nearer. Certainly the rubidium results that can be reliably compared with those obtained by other workers using other methods (see Tables VII and VIII) tend to confirm this statement. A possible explanation for larger differences, which can be seen in Table VI, for example, for lepidolite A110, may be inhomogeneity of the samples. g of rubidium and 1 x406 LANE A 250-MEGACYCLE HIGH-FREQUENCY TITRIMETER [Vol. 82 The present method supplies an alternative to the method of isotope-dilution analysis for the determination of traces of rubidium and an alternative to isotope dilution and other methods for the determination of rubidium in rubidium minerals.For caesium deter- minations no method of equal sensitivity is available. The method is a refinement of the procedure described by Cabell and Thomasll and refines and extends the scope of that described by Smales and Salmonlo to cover most materials of geochemical interest. We thank those who, as mentioned above, have supplied us with samples and the results of their own analyses. We are also grateful to Mr. ,4. M. Thomas of this establishment, for his assistance in analysing samples. REFERENCES 1. 3. 4. 5. 6. 3 -. I . x. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. Ahrens, L. H., Pinson, M’. H., and Kearns, M. M., Geochim. Cosmocham. .+’rta, 19ci2, 2, 229. Pinson, W. H., Ahrens, L. H., and Franck, M. L., Ibid., 1953, 4, 251. Davis, G. L., and Aldrich, L. T., Bull. Geol. Soc. Amer., 1953, 64, 379. Tomlinson, R. H., and Das Gupta, A. K., Caizad. J . Chem., 1953, 31, 909. IIerzog, L. F., and Pinson, W. H., Geochim. Cosmochirn. Acta, 1955, 8, 295. Pinson, W. H., Herzog, L. F., Backus, M. M., and Cormier, R., “Variations in Isotopic Abundances of Strontium, Calcium and Argon and Related Topics,” Annual Progress Report for 1955-56, Department of Geology and Geophysics, Massachusetts Institute of Technology, Report NYO 3936, 1956. Schumacher, E., Helv. Chim. Acta, 1956, 39, 538. Wetherill, G . W., Tilton, G. R., Davis, G. L., and Aldrich, L. T., Georhim. Cosmochim. A c f a , W’ebster, R. K., and Smales, A. A., unpublished results. Smales, A. A., and Salmon, L., Analyst, 1955, 80, 37. Cabell, M. J., and Thomas, A., Atomic Energy Research Establishment Report C/R 1725, HAL Vincent, E. A., and Smales, A. A,, Geochim. Cosmochim. Acta, 1956, 9, 154. Smales, A. A., and Loveridge, B. A., Anal. Chim. Acta, 1955, 13, 506. Cabell, M. J., Ind. Chem. Mfr, 1957, 33, in the press. Seelye, F. T., and Rafter, T. A., Nature, 1950, 165, 317. Kayas, G., J . Chim. Phys., 1950,47, 408. Duval, C., “Inorganic Thermogravimetric Analysis,” Elsevier Publishing Co., Amsterdam, 1953. Pinson, W. H., Herzog, L. F., Backus, M. M., and Cormier, R., 09. cit., p. 3. Aldrich, L. T., Wetherill, G. W., and Davis, <;. L., Geochim. Cosmochim. Acta, 1956, 10, 238. Smales, A. A., Ibid., 1955, 8, 300. Rankama, K., and Sahama, T. G., “Geochernistry,” University of Chicago Press, Chicago 35, 1956, 9, 292. Stationery Office, London, 1955. 1950, p. 38. ANALYTICAL CHEMISTRY GROUP ATOMIC ENERGY RESEARCH ESTABLISHMENT HARWELL, NR. DIDCOT, BERKS. October 4th, 1956

ISSN:0003-2654    

DOI:10.1039/AN9578200390

出版商:RSC    

年代:1957

数据来源: RSC

摘要:

406 A LANE A 250-MEGACYCLE HIGH-FREQUENCY TITRIMETER 250-megacycle High-frequency Ti trimeter* [Vol. 82 BY E. S. LANE The design and construction of a sta'ble high-frequency titrimeter operat- ing at 250 Mc/s are described. Examples are given of its use in conventional titrimetric procedures and indications are given of its potentialities in com- plexometric titrations and in a number (of new analytical procedures, e.g., a direct titrimetric determination of potassium as the tetraphenylboron salt and a non-aqueous titrimetric determination of mercury as copper propylene- diamine mercuric tetraiodide. ONE of the chief limitations to the more extensive use of high-frequency titrimetric methods in analytical chemistry lies in the inability of most of the instruments described to operate with solutions containing electrolyte at even moderate concentrations. Theoretical con- siderationsl have shown that the sensitivity of high-frequency titrimeters falls off rapidly when the electrolyte concentration in the titration vessel rises above a certain value and that the maximum tolerable electrolyte concentration is directly proportional to the frequency at which the circuit is worked.Whereas an instrument working at 30 Mc/s can be used to titrate solutions of approximately 0.06 N sodium chloride (or its equivalent), a 350 Mc/s * Presented a t the XVth International Congress on Pure and Applied Chemistry (Analytical Chemistry), Lisbon, September 8th t o 16th, 1956.June, 19571 LANE : A 250-MEGACYCLE HIGH-FREQUENCY TITRIMETER 407 circuit is required to deal with solutions of 0.7 N sodium chloride or its equivalent. To date the number of instruments described operating at these higher frequencies is few.Blaedel and Malmstadt used a 350-Mc/s quarter-wavelength concentric-line oscillator, but stated that extreme care in layout of parts and mechanical construction of their instrument was essential in order to reproduce the performance.2 Johnson and Timnick3 have described a more simply constructed half-wavelength coaxial-line oscillator operating at 130 Mc/s, but, as both these units are used in conjunction with expensive ancilliary equipment, it was thought that a less expensive instrument in this higher frequency range would be of use in furthering the wider application of high-frequency titrimetry.Accordingly, the design and construction of a simple and stable titrimeter working at 200 to 250 Mc/s, reasonably free from critical dimensions and made from readily available inexpensive components, is presented. No difficulty has been experienced in building a number of these units. The titrimeter is developed from a11 oscillator designed at Oak Ridge for the measurement of concentrations of alkali in aqueous solution and is based on a parallel-transmission-line oscillator, in which a type-955 acorn triode with constant current control of the oscillator plate current is used. The development of this oscillator has been fully described by Stelzner and Kelley.4 Ex PE RIME x TAL T)ESCKIPTION OF APPARATUS-- Fig. 1 1 2 @ j 9 b' * 1.: 0 3 C ,*. CI . u n - Circuit for power unit (for values of components, see Appendix, p.414) _ . I Fig. 2. Circuit for 240-ILIc,'s oscillator (for values of com- ponents, see Appendix, p. 414)408 LANE A 250-MEGACYCLE HIGH-FREQUENCY TITRIMETER [Vol. 82 The circuits of the power unit and of the instrument are shown in Figs. 1 and 2, respec- tively, and the general appearance of the app;xatus is seen from Figs. 3 and 4. Oscillator-The parallel-transmission-line oscillator makes use of an RCA 955 acorn triode. Two parallel 8-inch lengths of &inch copper tubing, spaced $ inch apart, are soldered to the plate and grid pins of the valve holder. These lines run vertically in the housing between plastic formers, while rigidly mounted between them is placed the titration vessel. A shorting condenser is soldered in place along the lines so as to establish a suitable oscillating frequency (200 to 250 Mc/s).The cathode and filaments of the valve are isolated from radio-frequency ground by 2 radio-frequency chokes, which are essential to the operation of the oscillator at this frequency because they ensure that the inter-electrode capacities of the valve are part of the tank circuit. Another radio-frequency filter array keeps stray voltages from the power supply. A portion of the grid voltage developed by the oscillator is filtered and fed through a zero-set circuit to a galvanometer. The zero-set circuit consists of a variable potentiometer for coarse control, and a 10-turn helically wound potentiometer for fine adjustment, in series with a large-capacity bell-type dry cell.Power supply-A conventional power supply with full-wave rectification and capacitor- input filtering is used. The design is due to Stelzner and Kelley, modified for British voltages and components. Housing-The power supply is housed in a standard rack unit. The oscillator, titration vessel and zero-set circuit, including the dry cell, *are mounted in a Winch x 64-inch x 6i-inch aluminium box, which is rigidly bolted to the rack unit. The titration vessel is shielded and connected to earth to avoid stray hand capacitance effects. Titration vessel-A Pyrex-glass tube, 2 inch1 diameter x 8 inches long, flared out at the upper end for a distance of 1+ inches to a diameter of 2 inches is used. This has a working capacity of about 60ml and the level of the liquid is maintained about 1 inch above the transmission lines, so that changes in liquid level during titration do not affect the response of the instrument.Slightly shorter tubes (6 inches long, and having a capacity of 40 ml) have been used in the more sensitive concentration ranges to achieve more rapid mixing. The cell is rigidly fixed to the chassis and is emptied by means of a suction tube. An elec- trically driven stirrer, comprising a 6-inch x $-inch x &-inch glass strip, twisted longi- tudinally through 360" and fixed to a glass rod, is required to mix the solution during titration. Stability-The designers of this oscillator showed it to have a remarkable stability, which develops after a warm-up period of about half an hour. When the titration vessel is filled with a conducting liquid, the absorption of radio-frequency energy results in a rise in the temperature of the solution and, since the conductivity of the solution is temperature dependent, this absorption of energy leads to some instability.When the cell was filled with water (50 ml), the temperature rise amounted to 3.5" C per hour from a room temperature of 20" C. Since a-titration requires only 10 minutes at most for completion, this effect is negligible for most practical purposes. A further example of the temperature dependence of conductivity interfering with titration occurred when concentrated solutions of acids and bases (2 N ) were being titrated during the preliminary investigation of the performance of the instrument.The heat of reaction was sufficient to cause a departure from linearity of the titration curves. However, this concentration of titrants is rarely encountered and dilution before titration minimises this complication. RESPONSE CURVES- Some idea of the performance of the instrument can be obtained from Fig. 5, in which the response of the instrument to progressively increasing concentrations of sodium chloride is plotted. Although Fig. 5 indicates that the instrument can still detect concentration changes at a concentration level of 60 to 70 1; of sodium chloride per litre, it should be remembered that this is a null instrument arid that the movement of the galvanometer needle from zero and its return by means of the zero-set circuit are highly damped at these high ionic concentrations.Here, a change in concentration causes a relatively slight diver- gence of the galvanometer needle from zero and its return by means of the zero-set circuit requires a relatively large movement of the potentiometer compared with the restoration of the same divergence at a lower ionic concentration. With high concentrations of elec- trolyte, therefore, the accuracy with which the potentiometer can be read is markedly depen- dent on the sensitivity of the galvanometer.408 LANE A 250-MEGACYCLE HIGH-FREQUENCY TITRIMETER [Vol. 82 The circuits of the power unit and of the instrument are shown in Figs. 1 and 2, respec- tively, and the general appearance of the app;xatus is seen from Figs. 3 and 4. Oscillator-The parallel-transmission-line oscillator makes use of an RCA 955 acorn triode.Two parallel 8-inch lengths of &inch copper tubing, spaced $ inch apart, are soldered to the plate and grid pins of the valve holder. These lines run vertically in the housing between plastic formers, while rigidly mounted between them is placed the titration vessel. A shorting condenser is soldered in place along the lines so as to establish a suitable oscillating frequency (200 to 250 Mc/s). The cathode and filaments of the valve are isolated from radio-frequency ground by 2 radio-frequency chokes, which are essential to the operation of the oscillator at this frequency because they ensure that the inter-electrode capacities of the valve are part of the tank circuit. Another radio-frequency filter array keeps stray voltages from the power supply.A portion of the grid voltage developed by the oscillator is filtered and fed through a zero-set circuit to a galvanometer. The zero-set circuit consists of a variable potentiometer for coarse control, and a 10-turn helically wound potentiometer for fine adjustment, in series with a large-capacity bell-type dry cell. Power supply-A conventional power supply with full-wave rectification and capacitor- input filtering is used. The design is due to Stelzner and Kelley, modified for British voltages and components. Housing-The power supply is housed in a standard rack unit. The oscillator, titration vessel and zero-set circuit, including the dry cell, *are mounted in a Winch x 64-inch x 6i-inch aluminium box, which is rigidly bolted to the rack unit.The titration vessel is shielded and connected to earth to avoid stray hand capacitance effects. Titration vessel-A Pyrex-glass tube, 2 inch1 diameter x 8 inches long, flared out at the upper end for a distance of 1+ inches to a diameter of 2 inches is used. This has a working capacity of about 60ml and the level of the liquid is maintained about 1 inch above the transmission lines, so that changes in liquid level during titration do not affect the response of the instrument. Slightly shorter tubes (6 inches long, and having a capacity of 40 ml) have been used in the more sensitive concentration ranges to achieve more rapid mixing. The cell is rigidly fixed to the chassis and is emptied by means of a suction tube. An elec- trically driven stirrer, comprising a 6-inch x $-inch x &-inch glass strip, twisted longi- tudinally through 360" and fixed to a glass rod, is required to mix the solution during titration. Stability-The designers of this oscillator showed it to have a remarkable stability, which develops after a warm-up period of about half an hour.When the titration vessel is filled with a conducting liquid, the absorption of radio-frequency energy results in a rise in the temperature of the solution and, since the conductivity of the solution is temperature dependent, this absorption of energy leads to some instability. When the cell was filled with water (50 ml), the temperature rise amounted to 3.5" C per hour from a room temperature of 20" C. Since a-titration requires only 10 minutes at most for completion, this effect is negligible for most practical purposes.A further example of the temperature dependence of conductivity interfering with titration occurred when concentrated solutions of acids and bases (2 N ) were being titrated during the preliminary investigation of the performance of the instrument. The heat of reaction was sufficient to cause a departure from linearity of the titration curves. However, this concentration of titrants is rarely encountered and dilution before titration minimises this complication. RESPONSE CURVES- Some idea of the performance of the instrument can be obtained from Fig. 5, in which the response of the instrument to progressively increasing concentrations of sodium chloride is plotted. Although Fig. 5 indicates that the instrument can still detect concentration changes at a concentration level of 60 to 70 1; of sodium chloride per litre, it should be remembered that this is a null instrument arid that the movement of the galvanometer needle from zero and its return by means of the zero-set circuit are highly damped at these high ionic concentrations. Here, a change in concentration causes a relatively slight diver- gence of the galvanometer needle from zero and its return by means of the zero-set circuit requires a relatively large movement of the potentiometer compared with the restoration of the same divergence at a lower ionic concentration. With high concentrations of elec- trolyte, therefore, the accuracy with which the potentiometer can be read is markedly depen- dent on the sensitivity of the galvanometer.408 LANE A 250-MEGACYCLE HIGH-FREQUENCY TITRIMETER [Vol. 82 The circuits of the power unit and of the instrument are shown in Figs.1 and 2, respec- tively, and the general appearance of the app;xatus is seen from Figs. 3 and 4. Oscillator-The parallel-transmission-line oscillator makes use of an RCA 955 acorn triode. Two parallel 8-inch lengths of &inch copper tubing, spaced $ inch apart, are soldered to the plate and grid pins of the valve holder. These lines run vertically in the housing between plastic formers, while rigidly mounted between them is placed the titration vessel. A shorting condenser is soldered in place along the lines so as to establish a suitable oscillating frequency (200 to 250 Mc/s).The cathode and filaments of the valve are isolated from radio-frequency ground by 2 radio-frequency chokes, which are essential to the operation of the oscillator at this frequency because they ensure that the inter-electrode capacities of the valve are part of the tank circuit. Another radio-frequency filter array keeps stray voltages from the power supply. A portion of the grid voltage developed by the oscillator is filtered and fed through a zero-set circuit to a galvanometer. The zero-set circuit consists of a variable potentiometer for coarse control, and a 10-turn helically wound potentiometer for fine adjustment, in series with a large-capacity bell-type dry cell. Power supply-A conventional power supply with full-wave rectification and capacitor- input filtering is used.The design is due to Stelzner and Kelley, modified for British voltages and components. Housing-The power supply is housed in a standard rack unit. The oscillator, titration vessel and zero-set circuit, including the dry cell, *are mounted in a Winch x 64-inch x 6i-inch aluminium box, which is rigidly bolted to the rack unit. The titration vessel is shielded and connected to earth to avoid stray hand capacitance effects. Titration vessel-A Pyrex-glass tube, 2 inch1 diameter x 8 inches long, flared out at the upper end for a distance of 1+ inches to a diameter of 2 inches is used. This has a working capacity of about 60ml and the level of the liquid is maintained about 1 inch above the transmission lines, so that changes in liquid level during titration do not affect the response of the instrument. Slightly shorter tubes (6 inches long, and having a capacity of 40 ml) have been used in the more sensitive concentration ranges to achieve more rapid mixing. The cell is rigidly fixed to the chassis and is emptied by means of a suction tube.An elec- trically driven stirrer, comprising a 6-inch x $-inch x &-inch glass strip, twisted longi- tudinally through 360" and fixed to a glass rod, is required to mix the solution during titration. Stability-The designers of this oscillator showed it to have a remarkable stability, which develops after a warm-up period of about half an hour. When the titration vessel is filled with a conducting liquid, the absorption of radio-frequency energy results in a rise in the temperature of the solution and, since the conductivity of the solution is temperature dependent, this absorption of energy leads to some instability.When the cell was filled with water (50 ml), the temperature rise amounted to 3.5" C per hour from a room temperature of 20" C. Since a-titration requires only 10 minutes at most for completion, this effect is negligible for most practical purposes. A further example of the temperature dependence of conductivity interfering with titration occurred when concentrated solutions of acids and bases (2 N ) were being titrated during the preliminary investigation of the performance of the instrument. The heat of reaction was sufficient to cause a departure from linearity of the titration curves. However, this concentration of titrants is rarely encountered and dilution before titration minimises this complication.RESPONSE CURVES- Some idea of the performance of the instrument can be obtained from Fig. 5, in which the response of the instrument to progressively increasing concentrations of sodium chloride is plotted. Although Fig. 5 indicates that the instrument can still detect concentration changes at a concentration level of 60 to 70 1; of sodium chloride per litre, it should be remembered that this is a null instrument arid that the movement of the galvanometer needle from zero and its return by means of the zero-set circuit are highly damped at these high ionic concentrations. Here, a change in concentration causes a relatively slight diver- gence of the galvanometer needle from zero and its return by means of the zero-set circuit requires a relatively large movement of the potentiometer compared with the restoration of the same divergence at a lower ionic concentration.With high concentrations of elec- trolyte, therefore, the accuracy with which the potentiometer can be read is markedly depen- dent on the sensitivity of the galvanometer.June, 19571 LANE, .4 25O-MEGACYCLE HIGH-FREQUENCY TITIUMETEK 409 Stelzner and Kelley point out that with an oscillator of this type the change in coil- ductivity of the solution in the cell contributes a greater share to the change in oscillator response than the change in dielectric constant. They found with potassium hydroxide solutions ;t cloce relationship between oscillator response and specific conductance.METHOD OF GSE- The method of operation of the instrument is similar to that used by Lane5 for the instrument designed by Dowdall, Sinkinson and Stretch.6 The solution to be titrated is placed in the titration cell and diluted to the required volume.. The stirrer is started and the galvanometer current adjusted to zero. Additions of titrant are made, the galvanometer being adjusted to zero after each addition and the scale reading recorded on the helically wound potentiometer. As the end-point is approached, there is a change of slope of the plot of potentiometer scale readings against volume of titrant added. Frequently a complete reversal of slope occurs. As is customary with many instrumental titrations, it is generally advantageous to determine the end-point approximately with large increments of titrant and then to carry out a duplicate titration with small increments in the vicinity of the end-point. APPLICATIONS OF THE HIGH-FREQUENCY TITRIMETER ACID - BASE AND ARGENTIMETRIC TITRATIOXS- Fig.6 shows the end-point obtained in the titration of aqueous hydrochloric acid with N sodium hydroxide. The same end-point is obtained when the solution is “loaded” with inert electrolyte, but then slope of the approach and retiring lines are reversed. This reversal of slope with differing ion concentrations has been noted by other workers. Fig. 7 shows the type of curve obtained when a solution of 0.1 N silver nitrate is titrated with 0.1 N potassium chloride. The end-points in these titrations agreed excellently with those obtained by well established methods.TITRATIOXS I N NON-AQUEOUS SOLVENTS- Considerable progress has been made in the application of high-frequency conductimetric methods to non-aqueous titrimetry. Wagner and Kauffman7 used the technique for the titration of bases in glacial acetic acid, and acids in dimethylformamide were titrated with sodium methoxide by Dean and Cain.* My experience5 of high-frequency titrimetry with a wide variety of compounds, including phenols, enols, quaternary ammonium salts and heterocyclic bases, suggests that the method is comparable in performance to the potentio- metric determination, but is far more convenient in operation.410 LANE : A 250-MEGACYCLE HIGH-FREQUENCY TITRIMETER [Vol. 82 The examination of several representative titrations in glacial acetic acid, ethylene- diamine and dimethylformamide with this titrimeter has indicated, however, that no advantage is to be gained by performing these titrations a.t frequencies greater than 15 Mc/s.350- 300- M .5 250- e - 2200- V I ._ 5 - C 2 150- n 1 : l I I I l i I I 2 4 6 8 Volume of N sodium hydroxide added, m l 8 9 10 I I 12 Volume of 0.IN potassium chloride added, ml Fig. 6. Acid - base titration; curve A, hydro- chloric acid titrated with N sodium hydroxide; curve B, hydrochloric acid titrated with N sodium hyroxide in presence of 5 ml of 10 per cent. w/v potassium chloride. Indicator end-point, 5.2 ml; volume a t end-point, 40 ml Fig. 7 . Argentimetric titration, 10 ml of 0.1 N silver nitrate being titrated against 0.1 AT potassium chloride HIGH-FREQUEXCY CONDUCTIMETRIC DETERMINATION OF POTASSIUM BY DIRECT TITRATION The quantitative determination of potassium is one of the less satisfactory analytical procedures and the method a t present in vogut: is based on its gravimetric determination as potassium tetraphenylboron.Various attempts have been made to devise volumetric methods based on the formation of the potassium tetraphenylboron precipitate, but apart from the conductimetric titration of Raff and B r ~ t z , ~ in which the titrant is used as the precipitant, all the methods involve the prior separation of the precipitate by filtration, and it is precisely this operation that is to be avoided in this particular determination because of the difficult physical form of the precipitate ilnd the careful washing technique required.In view of the fact that Raff and Brotz successfully titrated potassium solutions with lithium or sodium tetraphenylboron solutions and obtained sharp end-points conductimetrically, the high-frequency conductimetric method with its advantages of contactless electrodes was applied to this system and Fig. 8 shows the type of curve obtained. The sodium tetraphenyl- boron solution (0.1 N) was prepared in the usual way and standardised against potassium hydrogen phthalate solution and then used to determine the unknown potassium solution in 0.05 N acetic acid. The sodium tetraphenylboron solution may also be standardised against silver nitrate by this method. In experiments when 8.26 mg of potassium chloride were taken, the recoveries were 8.21, 8.28 and 8.27 mg.WITH SODIUM TETRAPHEKYLBOROK- NON-AQUEOES TITRIMETRIC DETERMINATION O F MERCURY (Hg2+) FROM AQUEOUS SOLUTIOKS- There are many examples in classical analytical chemistry in which a gravimetric deter- mination of a metal is avoided by allowing a precipitated metal compound (or its solution in a suitable reagent) to react with a standardised titrant in preference to the weighing ofJune, 19571 LANE : .A 250-MEGACYCLE HIGH-FREQUENCY TITRIMETER 41 1 the precipitate itself. Perhaps the most widely used is the bromimetric determination of metal oxinate precipitates. 755- 04 $750- e z745- L 01 .- $ u C !740- L 735- I 1 4341 I I I I I I 1 1 . 1 2 4 6 8 Volume of perchloric acid added, ml Fig. 9. Determination of merury (Hgzt) as copper propylenediamine mercuric iodide by titration with perchloric acid in acetic acid.direct titration with sodium tetraphenyl- Found, 0*100(4) g of Hg2+; taken, 0.1004 g of boron Hg2+ The increasing use of titrations in non-aqueous solvents greatly widens the scope of the “titrimetric” finish to the quantitative determination of metals, and a notable example in recent years of the use of this particular method has been the titration by Flaschka of potassium tetraphenylboron precipitates with perchloric acid in glacial acetic acid with a visual indicator. An example of the use of a high-frequency titrimetric finish in a conven- tional determination of a heavy metal is provided by a modification of the method of Spacu and Spacull for mercuric ions.The metal is precipitated from aqueous solution by the addition of potassium iodide and cupric propylenediamine sulphate as Cu.pn2.HgI, (pn = 1 : 2- propanediamine). This precipitate can be dissolved in glacial acetic acid and mercuric acetate and titrated with perchloric acid in glacial acetic acid, four equivalents of perchloric acid, a very favourable volumetric factor, being required for equivalence. The mercuric acetate is required for the conversion of iodide ion to acetate ion (see Pifer and Wollish12). Conventional visual indicators for this particular titration, eg., Oracet blue B and crystal violet, give indistinct end-points, but Fig. 9 shows the highly satisfactory end-point given with this titrimeter. The precipitation and washing of the mercury precipitate was carried out as described by Walton and Sniith,13 except that the precipitate, without being dried, was dissolved in glacial acetic acid and mercuric acetate and titrated with 0.1 N perchloric acid in glacial acetic acid.THE STUDY OF CHELATIOX- Cornfileximetric titratiom-Chelation may be profitably studied by means of high-fre- quency titrimetry. As ethylenediaminetetra-acetic acid (EDTA) is one of the strongest and most generally applicable chelating agents in use, some examples of the use of this technique in the study of EDTA and related compounds are presented. The analytical use of EDTA-type compounds is usually based either on the measurement of hydrogen ion released during chelation or on the direct measurement of the compleximetric reaction itself.A general equation for the reaction with a metal salt is- 2 4 6 8 Volume of sodium tetraphenylboron added, ml Determination of potassium by Fig. 8 . H, Y -+ MIIX2- -+ 2H+ + X2- -+ H2 (MIIY). The drop in pH due to liberation of the hydrogen ion may be determined by titration with alkali. Alternatively, the end-point of the compleximetric reaction may be indicated by potentiometric, amperometric, spectrophotometric or internal-indicator methods. Suit- able indicators for the last method comprise those compounds forming coloured weakly bound complexes with the ion being titrated.412 LANE : A 250-MEGACYCLE HIGH-FREQUENCY TITRIMETER [Vol. 82 Considerable impetus has been given to the use of high-frequency titrimetry in the study of these reactions by the work of West14 and his school at the Louisiana State University.The results obtained by these workers with the Sargent Oscillometer can be repeated by using the 250-Mc/s titrimeter in a much higher ionic concentration. I I I , I M I I I m EDTA added, rnl (4 :r 50 I 2 3 EDTA added, rnl (4 (4 Fig. 10. Titration of thorium with;EDTA: (a), 5 mlrof 0.04 M thorium nitrate; ( b ) , as (a) with 10 ml of 0.1 N potassium chloride added; ( c ) , as (a) with 5 ml of 0.1 N sulphuric acid added; ( d ) , as (a) with 0.25 g of ammonium sulphate ;added r-----l 7- I I cJune, 19571 LANE : A %%MEGACYCLE HIGH-FREQUENCY TITRIMETER 413 acetate ion, Fig. 11 shows a satisfactory titration of thorium by EDTA in the presence of a large excess of acetate ions. By using this technique similar (1 + 1) complexes have been observed for thorium and diethylenetriaminepenta-acetic acid and cyclohexyl-1 : 2-diamine- tetra-acetic acid.Fig. 12 shows a titration of the monosodium and disodium salts of nitrilotriacetic acid (NTA) titrated with a standard solution of copper and indicates the relatively high ionic concentrations that can be satisfactorily dealt with by this instrument. Chelating agents in non-aqueous titrimetry-Some of the limitations of the use of chelating agents for the titrimetric determination of metals are due to the following causes- ( i ) the chelating agent does not combine with the metal ion sufficiently strongly to give (ii) intermediate compounds formed between the metal ion and the chelating agent a good titration curve, and mask the titration curve E I40- e i i , V 2 100 E 2 60 0- 20- \ , \ , , I60 150- 140- 130- Volume of 0 IN cupric chloride added, ml Volume of 0,IN cupric chloride added, ml (4 ( b ) Fig.12. Titration of NTA with copper: ( u ) , 0.1 g of NTA and 10.5 ml of 0.1 N sodium hydroxide; ( b ) . 0 . 1 g of NTA and 2 1 4 ml of 0.1 A' sodium hydroxide 201 i I I I I I I 2 3 4 5 Volume of O,IN methanolic potassium mechoxide added, ml I I I I - I 2 3 4 5 Volume of 0.IN methanolic potassium methoxide added, ml Fig. 13. Uranyl nitrate - sodium salicylate Fig. 14. Uranyl nitrate - 8-hydroxyquinoline chela- chelation in methanol tion in methanol414 LANE : A %@MEGACYCLE HI(;H-FREQUENCY TITRIMETER [Vol. 82 The use of non-aqueous media for these titrations, introduced by Brummett and Holl- weg,l5 marks a considerable forward step in this field, since the choice of chelating agents is widened (many are insufficiently soluble in water for use in aqueous systems) and the formation constant may be favourably altered.The high-frequency conductimetric-titration technique has been applied t o these systems in view of my conviction that this method is simpler than the conventional potentio- metric end-point detection, particularly in inexperienced hands. This method has been applied to the determination of uranyl ion by means of sodium salicylate. An excess of solid sodium salicylate was added to an aliquot of uranyl nitrate solution (0.1 N ) and the solution was diluted to the required volume (40 ml) with methanol. Hydrogen ion is released during the chelation and this is titrated with standardised potassium methoxide in a benzene - methanol mixture.Fig. 13 shows the titration curve, which is of some theoretical importance in so far as it proves that the o-hydroxyl group of salicylic acid is displaced during the chelation process, whereas Bernstrom,16 working on the basis of the similarity between salicylic acid and o-methoxybenzoic acid in extraction experiments, concluded that only the carboxyl group was involved with the uranyl ion. Fig. 14 shows a similar titration; uranyl nitrate in methanol is treated with an excess of 8-hydroxyquinoline and the hydrogen ion released during chelation is titrated with potassium methoxide in a benzene - methanol mixture. APPENDIX LIST OF COMPONENTS USED IN THE CONSTRUCTION OF THE POWER UNIT (Fig.1) L P PL 1 PL2 Fl, F2 s, 470,000-ohm, %-watt, carbon resistance. 1200-ohm, 3-watt, wire-wound resistance. 100-ohm, 10-watt, wire-wound resistance. 12,000-ohm, 6-watt, wire-wound resistance. 2500-ohm, wire-wound variable resistance. 10-ohm, 1 0-watt, wire-wound resistance. 8-pF condenser. 0.1-pF condenser. CV 1268 valve. CT.' 686 valve. CV 216 valve. CV 449 valve. CV 848 (6 AG 5 ) valve. hfains transformer : primary winding, 250 volts; secondary windings, 250-0-250 volts, 5 volts, 6.3 volts and 6.3 volts. 10-henry 60-111.4 choke, of d.c. resistance 390 ohms. Neon indicator. Three-pin small fixed plug (Plessey). Six-pin small fixed socket (Plessey). 2-amp fuse. Double-pole switch. LIST OF COMPONENTS IJSED IN THE CONSTRUCTION OF THE OSCILLATOR (Fig.2) = 25,000-ohm resistance, with a 2 per cent. tolerance. = 10,000-ohm potent:ometer. = 1000-ohm helical wire-wound potentiometer (Colvernj. = 0.005-pF condenser, with .+ 20 per cent. tolerance. Rl R, R3 Cl, c,, c3, c4, c, c, = O.l-pF condenser, with & 25 per cent. tolerance. = 820- F condenser. = R C & ~ (acorn type). c, Vl RFCI, RFC2, RFC3, RFC4 = Radio-frequency choke I am indebted to Mr. C. Lomas, A.E.R.E., for the construction of the instrument; to Mr. T. Jaques and Mr. S. J. Burnett, A.E.R.E:., for helpful discussions on circuitry and to Dr. J. K. Aiken, Geigy Ltd., who kindly provided a supply of diethylenetriarninepenta-acetic acid and cyclohexyl-1 : 2-diaminetetra-acetic acid,June, 19571 LANE : A 250-MEGACYCLE HIGH-FREQUENCY TITRIMETER 415 REFERENCES science Publishers Inc., New York, London, 1954. 1. Reilley, C. N., in Delahay, P., “New Instrumental Methods in Electrochemistry,” Inter- 2. Blaedel. IV. 1.. and Malmstadt. H. V.. Anal. Chem.. 1950. 22. 1413. _ I , I~ 3. 4. Johnson, A. H., and Timnick, k., ha., 1956, 28, 889. Stelzner, R. \V., and Kelley, M. T., U.S. Atomic Energy Commission Report ORNL-1742, Oak Ridge National Laboratory, Tennesse, 1955, unclassified. 5 . 6. 7 . 8. 9. 10. 11. 12. 13. 14. 15. 16. Lane, E . S., Analyst, 1955, 80, 675. Dowdall, J . P., Sinkinson, D. V., and Stretch, H., Ibid., 1955, 80, 491. Wagner, W. E., and Kauffman, \V. B., Anal. Chem., 1953, 25, 538. Dean, J . A . , and Cain, C., jun., Ibid., 1955, 27, 212. Raff, P., and Brotz, JV., 2. anal. Chem., 1951, 133, 241. Flaschka, II., Chenrist-Analyst, 1955, 44, 60. Spacu, G., and Spacu, P., Z. anal. Chew., 1932, 89, 187. E’ifer, C. \V., and IVollish, E. G., Anal. Chem., 1952, 24, 300. IValton, H. F., and Smith, H. A4., Ibid., 1956, 28, 406. Hara, R., and West, P. TV., Anal. Chim. A c f a , 1954, 11, 264; 1955, 12, i 2 ; 195~5~ 12, 285; 1955, 13, 189; 1956, 14, 280. Brummet, B. D., and Hollweg, K. M., Anal. Chem., 1956, 28, 448. Hok. Bernstrom, B., Acta Chem. Scand., 1956, 10, 163. CHEMISTRY DIVISION .~TohfIC ENERGY RESEARCH ESTABLISHMENT HARWELL, NR. DIDCOT, BERKS. Octobev 12th, 1956

ISSN:0003-2654    

DOI:10.1039/AN9578200406

出版商:RSC    

年代:1957

数据来源: RSC

摘要:

June, 19571 LANE : A 250-MEGACYCLE HIGH-FREQUENCY TITRIMETER 415 Polarographic Determinations of Thorium BY R. P. GRAHAM AND G. B. LARRABEE* Indirect polarographic methods for the determination of thorium are described. These depend on the precipitation of thorium by either m-nitro- benzoic acid, m-nitrophenylarsonic acid or 8-hydroxyquinoline, the dissolution of the separated and washed compound, and then the polarographic deter- mination of the regenerated organic reagent. The pH conditions required for the precipitations by the reagents and appropriate for their polarographic determination are discussed. The method with m-nitrobenzoic acid, in which thorium is precipitated as the tetra-(m-nitrobenzoate), has been shown to be applicable to the determination of thorium in magnesium alloys containing zinc and rare-earth elements.IN the last few years much study has been made of methods for the determination of thorium. New absorptiometric, titrimetric, gravimetric, spectrochemical, and X-ray fluorescence methods, amongst others, have been developed since hloeller, Schweitzer and Starrl published their review of the analytical chemistry of thorium in 1948. Very little attention, however, has been given to polarographic methods. Thorium apparently cannot be determined by a direct polarographic method, the value of its half-wave potential being too negative to permit the recording of a wave before hydrogen ions or other cations of the supporting electrolyte are reduced.2 One indirect polarographic method has been developed, by K~mBrek.~ This depends on the polarographic determination of iodate after converting, by means of sodium hydroxide, a precipitate of thorium iodate to one of hydrous thoria.Associated with this method is the difficulty of freeing precipitated thorium iodate from adsorbed iodate without the loss of significant amounts of thorium by dissolution and hydrolysis4 and, further, the difficulty of washing the desired iodate out of a precipitate of hydrous thoria. This paper describes indirect polarographic methods for thorium in which use is made of an organic reagent both as a precipitant and as a polarographically reducible substance. m-Nitrophenylarsonic acid, 8-hydroxyquinoline and m-nitrobenzoic acid have been studied, and it has been shown that each of these can be the basis of a precise and accurate polarographic method for thorium.The methods involve the quantitative precipitation of thorium, under controlled conditions, by the organic reagent, then the dissolution of the separated and washed precipitate, and finally the polarographic determination of the free organic reagent that results from the dissolution. By appropriate calibration procedures, * Present address: Canadian Westinghouse Co. Ltd., Hamilton, Ontario, Canada.416 GRAHAM 4ND LAKRABEE : POLAROGRAPHIC DETERMINATIOSS OF THORIUM ,vOl. 82 the amount of the organic reagent that is found is related to the amount of thorium in the original solution. Of the three reagents, m-nitrobenzoic acid is the most useful (vide infm). The organic reagents with which this paper is concerned have been used in other studies of the analytical chemistry of thorium.m-Nitrophenylarsonic acid has been proposed for use in an amperometric titration of thorium5; 8-hydroxyquinoline has been used for the determination of thorium by gravimetric and titrimetric methods, l s 6 and for the separation of thorium from c e r i ~ m ~ ~ ~ ~ ~ and from lanthanum7; m-nitrobenzoic acid, first proposed as a precipitant for thorium half a century ago,1° but not given much attention until recently, has lately been used or recommended for the gravimetric determination of thorium,11J2Js for the separation of thorium and zirconium11~L2~14 (by control of the pH of precipitation) and as the basis of a titrimetric method15 and a colorimetric methodl6 for thorium.METHOD -hPP.4KATUS- The polarograph was a Tinsley pen-recording instrument, type V722/13. The capillary for the dropping-mercury electrode was a piece of marine barometer tubing. A conven- tional H-type of polarographic cell1' was used, electrolytic contact with the saturated calomel reference cell being made through a potassiurn chloride - agar bridge. Measurements of pH were made with a Beckman model G pH meter, equipped with a G-1190-80 glass electrode. REAGENTS- The mercury for the dropping-mercury electrode was purified by washing it for a t least 10 hours in 10 to 15 per cent. nitric acid, then washing it several times with water and finally distilling it twice in vucuo. The 8-hydroxyquinoline was AnalaR grade, the m-nitrophenylarsonic acid was Eastman white-label grade (obtained from Distillation Products Industries, Rochester, U.S.A.), and the m-nitrobenzoic acid was prepared by the method described by Kamm and Segur.ls The magnesium was metal that had been specially purified by sublimation (obtained from Dow Chemical Co., Midland, U.S.A.), and the mischmetal (obtained from Dominion Magnesium Ltd., Haley, Canada) was stated to contain 45 1 per cent.of cerium, 97 per cent. of total rare-earth elements, 3 per cent. of iron and 0.004 per cent. of thorium. The alizarin blue was a spot-test reagent grade (obtained from British Drug Houses Ltd.) and the gelatin was purified Pigskin Gelatin (obtained from Eastman Kodak Co., Rochester, U.S.A.). A hot saturated aqueous solution of a C.P. commercial product (obtained from Amend Drug and Chemical Co., New York, U.S.A.) was prepared and allowed to cool a t room temperature; the solution was saturated with hydrogen chloride and an equal volume of ethyl ether was added; the mixture was shaken until crystallisation was complete, and then it was filtered and the product was washed several times with ether.The purified salt was dried a t room temperature in vucuo. Stock solutions of thorium chloride in 5 per cent. hydrochloric acid (about 2 mg of thorium per ml) were prepared and their thorium contents were determined by precipitation of the hydrous oxide by means of ammonia solution, followed by ignition to thorium dioxide. Thorium solutions for calibration work and other purposes were obtained, as needed, by quantitatively diluting a stock solution with water.All other reagents were of recognised analytical quality. POLAROGRAPHIC PROCEDURES- Before polarograms were recorded, each solution was de-aerated by bubbling through it for 10 minutes nitrogen that had been purified by being passed through a solution of pyrogallol in aqueous potassium hydroxide and then through water. All polarographic measurements were made at 25.0" 0.05" C. Usually five polaro- grams for each solution were recorded, and the average wave height was taken as the value. THORIUM 8-HYDROXYQUINOLATE POLAROGRAPHY- Although the behaviour of 8-hydroxyquinoline a t the dropping-mercury electrode is complicated and pronouncedly dependent on pH,19 120 this reagent has found polarographic applications in inorganic analysis. Magnesium and aluminium, for example, have been A purified sample of thorium tetrachloride was obtained as follows.June, 19571 GRAHAM AND LARRABEE : POLAROGRAPHIC DETERMINATIONS OF THORIUM 417 determined indirectly, by the measurement of one of the waves given by this reagent, a t a controlled pH, after dissolution of the precipitated metal 8-hydroxyquinolate (or by measurement of the excess of reagent after precipitation).Different media of differing pH have been recommended for the polarographic determination of 8-hydroxyquinoline : Car- ruthers21 used a phosphate buffer of pH 7.1 ; Stone and Furman,z2 an ammonia - ammonium chloride buffer of pH 10; Parks and L ~ k k e n , ~ ~ the same buffer system a t pH 9.8. We have obtained very satisfactory results in a phosphate buffer at pH 12.0 (disodium monohydrogen phosphate, 0.10 M , and sodium hydroxide, 0.10 M ) , Under these conditions, 8-hydroxy- quinoline gives a single and well formed wave in the presence of gelatin as a maximum sup- pressor.The diffusion current given by 8-hydroxyquinoline is linearly related to the amount of thorium (at least up to 10 mg) that was present in the solution from which it wasprecipitated by the organic reagent, the procedure described in detail below being used. PRECIPITATION- Whether the precipitate that thorium gives with 8-hydroxyquinoline contains 4 or 5 moles of the organic reagent for every mole of thorium depends on the temperature of pre~ipitation.~~ (Scandium, uranium, and plutonium also form 8-hydroxyquinolates with an “extra” molecule of the organic reagent.25) The structure of the “5-complex” of thorium has been the subject of recent investigation^.^^^^ We have found that homogeneous precipitation of thorium 8-hydroxyquinolate by means of urea yields a precipitate that coagulates more rapidly and that is more easily filtered off and washed than that obtained by the method of Frere.26 (Moeller and RamaniahZ4 also recommend a homogeneous precipitation of this compound.) On the basis of its colour and the temperature of its precipitation, the 8-hvdroxyquinolate that we obtained can be presumed to be the “5-complex.” PKOCEDURE- To 50.0 ml of a thorium solution containing about 10 mg of thorium there were added, drop by drop, 8 M sodium hydroxide until hydrous oxide just succeeded in re-dissolving on stirring, then 10 ml of a 2.5 per cent.solution of 8-hydroxyquinoline in 1.0 M acetic acid and finally 12 g of urea. The solution was heated and kept just below the boiling point for about 20 minutes, until the orange-red precipitate, which forms on hydrolysis of the urea, coagulated. The precipitate was collected on a M’hatman No. 42 filter-paper and washed thoroughly with water. Next, the precipitate was dissolved directly into a 100-ml calibrated flask by means of hot 6 M hydrochloric acid. Five millilitres of a 20 per cent. solution of tartaric acid were then added (to complex the thorium so that it would not be re-precipitated during the subsequent adjustment of the pH of the solution for polarography), and, by using 8 M sodium hydroxide, the pH of the solution was brought to the end-point of alizarin blue.Finally, a further 2 ml of 8 M sodium hydroxide, 35 ml of a 0-10 M disodium mono- hydrogen phosphate - 0.10 M sodium hydroxide buffer solution of pH 12.0 and 1 ml of a 0.75 per cent. aqueous gelatin dispersion were added, and the solution was diluted to volume with water and polarograms were recorded. A procedure that obviates any treatment of the thorium 8-hydroxyquinolate consists in adding an exactly known volume of the solution of 8-hydroxyquinoline and polarographi- cally determining the excess of the organic reagent in the filtrate. With concentrations of reagents slightly different from those given above, and omitting the tartrate (since there is no thorium in need of complexing), this indirect approach has been shown, like the one above, to yield a straight-line calibration curve for amounts of thorium up to a t least 10 mg.Procedures similar to this have been recommended for the polarographic determination of magnesium22 and of aluminium.23 POLAROGRAPHY- A tracing of the excellent polarographic wave that m-nitrophenylarsonic acid can give a t pH 4.0 in a solution of potassium hydrogen phthalate (with methyl red as a maximum suppressor) is shown in another paper from this laboratory.2i The diffusion current given by this arsonic acid was found to be a linear function of the amount of thorium (at least up to 20 mg) that was present in the solution from which it was precipitated (at a pH of 1.9 THORIUM YY1-SITROPHE~YLARbONATE 0.1, by the procedure given below).41 8 GRAHAM AND LARRABEE : POLAROGRAPHIC DETERMINATIONS OF THORIVM [VOl. 82 PRECIPITATION- The extent to which thorium is precipitat1:d by m-nitrophenylarsonic acid depends on the pH of the solution, as is shown in Fig.1. Each point on the curve for thorium m-nitro- phenylarsonate was obtained as follows. To 50.0 ml of a thorium tetrachloride solution, containing about 10 mg of thorium, there was added 4 M ammonia solution until the desired pH value was reached and then 50 ml of a 0.24 per cent. aqueous solution of m-nitrophenyl- arsonic acid of the same pH. The precipitate was kept at room temperature for 2 hours to coagulate it, then it was collected on a Whatrnan No. 42 filter-paper and washed four times with water. Next, the precipitate was dissolved directly into a 100-ml calibrated flask by means of a hot tartrate solution (prepared by making a 0.5 M solution of tartaric acid strongly alkaline to phenolphthalein with concentrated sc'dium hydroxide), and the pH of the solution was adjusted to the methyl red end-point with 5 M perchloric acid.Thirty-five millilitres of a saturated potassium hydrogen phthalate solution were added (as buffer and supporting electrolyte), and the solution was diluted to volunie with water and polarograms were recorded. From curve A, Fig. 1, it is clear that, from hydrochloric acid solutions, thorium gives no precipitate with m-nitrophenylarsonic acid at a pH below 0.7, and an analytically useful m-nitrophenylarsonate precipitate only in the pH range 1.4 to 2.6. At higher pH values the precipitate is a mixed one, presumably of arsonate and hydrous oxide.Similar findings have been reported for titanium m-nitr~phenylarsonate.~~ I 3-0 0 I .o 2.0 PH Fig. 1. Effectivenes of nz-nitrophenyl- arsonic acid (curve -4) and m-nitrobenzoic acid (curve B) as precipitants for thorium as a function of pH. The amount of thorium was constant throughout thc experiments represented by curve A , and likewise for curve B, although i t was somewhat different in the two instances The finding of Kolthoff and Johnson5 that m-nitrophenylarsonic acid can be used for the amperometric titration of thorium in hydrochloric acid when the pH is 2.0 to 2.5, but not when the pH is 1, is of course consistent with the results shown in Fig. 1. PROCEDURE- solution being first brought to a pH of 2.0 arsonic acid solution being of pH 2.0.The recommended procedure is that given :in the preceding section, with the thorium 0.1 (or a t most I. 0.2) and the m-nitrophenyl-June, 19573 POLAROGRAPHY- The polarographic diffusion current for a given concentration of m-nitrobenzoic acid (and also its half-wave potential) depends on the pH of the s o l u t i ~ n . ~ ~ ~ ~ ~ For analytically useful results a good buffer is therefore required, as with 8-hydroxyquinoline and m-nitro- phenylarsonic acid. We have found that potassium hydrogen phthalate serves well as a buffer (pH 4.0) and as a supporting electrolyte. Under these conditions, and with 0.005 per cent. of 1-naphthol present as a maximum suppressor, m-nitrobenzoic acid yields a very well formed and easily measurable wave, such as that shown in Fig.2. Although m-nitrobenzoic acid can yield two waves, only one appears at pH 4.0; this probably corresponds to a reduction of the nitro group to the hydroxylamine stage. GRAHAM AND LARRABEE : POLAROGRAPHIC DETERMINATIONS OF THORIUM 419 THORIUM m-NITROBENZOATE I Applied potential for the reduction of m-nitrobenzoic acid a t pH 4.0 Fig. 2. Typical polarographic wave Fig. 3. Diffusion current (wave height) of nz-nitrobenzoic acid (at pH 4.0) as a linear function of the amount of thorium present in the solution from which i t was precipitated (at pH 2.15) The linear relation between the height of the wave for m-nitrobenzoic acid and the amount of thorium present in the solution from which it was precipitated, a t pH 2.15 2 0.10, is shown in Fig.3. PRECIPITATION- As with the other organic reagents, the effectiveness of m-nitrobenzoic acid as a precipitant for thorium depends on the pH of the solution. Each point on the curve for thorium m-nitro- benzoate that is shown in Fig. 1 was obtained as follows. To 50.0ml of a thorium tetra- chloride solution, containing about 10 mg of thorium, there were added 3 to 4 ml of 12 M hydrochloric acid, 50 ml of a saturated aqueous solution of m-nitrobenzoic acid (see Note 1) and then 7.5 M ammonia solution until the desired pH was attained. The precipitate was coagulated by heating the mixture near the boiling point for about 15 minutes, and then was collected on a Whatman No. 40 filter-paper and washed with hot water.Next, the precipitate was dissolved directly into a 100-ml calibrated flask by means of hot 4 M perchloric acid (see Note 2). Five millilitres of a 20 per cent. tartaric acid solution were added (see Note 3) and the pH of the solution was adjusted to the methyl red end-point by adding 15 M ammonia solution from a burette. Thirty-five millilitres of a saturated solution of potassium hydrogen phthalate were added (as buffer and supporting electrolyte), and then 10 drops of a 1 per cent. solution of 1-naphthol in ethanol (as a maximum suppressor). Finally, the solution was diluted to volume with water and polarograms were recorded. From curve B, Fig. 1, it is evident that thorium gives no precipitate with m-nitrobenzoic acid at a pH below 1-6 and a precipitate that is useful (in the present context) only in the420 GRAHAM AND LARRABEE : POLAROGRAPHIC DETERMINATIONS OF THORIUM [VOl.82 pH range 1.8 to 2.6. At higher pH values the precipitate is a mixed one, probably of m-nitrobenzoate and hydrous oxide. NOTES- The saturated m-nitrobenzoic acid solution was prepared by dissolving 4 g of m-nitrobenzoic acid in 1 litre of hot water, allowing the solution to cool overnight, and then filtering off the acid that had crystallised out. Perchloric acid was found to be a much more satisfactory solvent for the m-nitrobenzoate than nitric, hydrochloric or sulphuric acids. The tartaric acid solution was added to complex the thorium so that the subsequent adjustment of the pH, in preparing the solution for polarogra.phy, would not cause its re-precipitation.1. 2. 3. PROCEDURE- The recommended procedure is that given in the preceding section, with the pH of precipitation controlled to 2.15 f 0.10 (rtO.20 should, however, be permissible). This procedure yields a straight-line calibration curve for amounts of thorium up to at least 12mg (see Fig. 3). Results for the determination of thorium by this method in solutions containing greater concentrations of zinc and rare-earth elements and a much greater concentration of mag- nesium are summarised in Table I, but the initial volume was 25 ml and not 50 ml. TABLE I DETERMINATION OF THORIUM IN SOLUTIONS CONTAINING MAGNESIUM, All solutions contained 230 mg of magnesium and 19 mg of zinc; the elements were present as chlorides in a volume of 25 ml Amount of rare-earth Thorium taken, Thorium found,? elements present,* mg mg mg ZINC AND RARE EARTHS 12 10.20 & 0.01 10.22 & 0.09 49 10.20 & 0.01 10.22 f 0.10 22 6.12 i: 0.01 6.15 & 0.09 74 10.20 * 0.01 10.19 & 0.03 * Added in the form of a hydrochloric acid solution of mischmetal.7 The f value represents the over-all spread in the wave-height measurements of the 5 polarograms recorded for each solution. COblPOSITION OF THORIUM W-NITROBENZOATE- The composition of thorium m-nitrobenzoate was determined by precipitating a known amount of thorium with m-nitrobenzoic acid (by the procedure given above, the pH being 2.15 i 0-lo), dissolving the washed precipitate with hot 4 M perchloric acid and determining the amount of m-nitrobenzoic acid polarographically. The linear calibration curve for these determinations was obtained from measurements with aliquot portions of a standard solution of m-nitrobenzoic acid, prepared from acid that had been recrystallised three times from hot water.Before the adjustment of the pH of the calibration solutions and the addition of the supporting electrolyte, tartaric acid and perchloric acid (in amounts equal to those used when thorium was present) were added. The results of these experiments are given in Table 11. TABLE I1 COMPOSITION OF THORIUM m-NITROBENZOATE Thorium taken, wz-Nitrobenzoic ;acid found, Molar ratio moles x lo4 moles x lo4 0.803 3.19 3.97 1.605 6*4(6) 4.03 2.410 9.5(9) 3.98 3.221 12.6(5:1 3.94 4.015 15.9(1:1 3.96 4.81 19.1 3.97 4.81 19.0 3.95 4.81 19.2 3.99 Average = 3.97June, 19571 GRAHAM AND LARRABEE : POLAROGRAPHIC DETERMINATIONS OF THORIUM 421 The precipitate is clearly the tetra-(m-nitrobenzoate).Other workers have found the same composition by other methods.10~80 If the precipitate is assumed to be the tetra compound, then the failure of the molar ratio (Table 11) to be exactly four can be attributed to incompleteness of the precipitation of the thorium. The divergence suggests that the precipitation was short of completeness by 0.7 per cent. Rather satisfactory verification of this was obtained in gravimetric experi- ments in which the nitrobenzoate precipitate obtained from a known amount of thorium was ignited to the oxide; the recovery of thorium was 99.5 per cent. DISCUSSION The conditions under which thorium gives with either m-nitrophenylarsonic acid, m-nitrobenzoic acid or 8-hydroxyquinoline an analytically useful precipitate are not unduly restrictive.Each precipitate is readily soluble in an appropriate medium, and each of the organic reagents gives an easily measurable polarographic wave. For each there is a strictly linear relationship between the diffusion current of the reagent and its concentration, at least over a substantial range of concentrations. Although the polarographic procedures for the determination of thorium with 8-hydroxy- quinoline are very convenient, they are restricted in application, owing to the unselective character of 8-hydroxyquinoline as a precipitant. The precipitation of thorium by m-nitrophenylarsonic acid is more attractive, because of the more selective precipitating ability of m-nitrophenylarsonic acid.The polarographic procedure that was developed by using this reagent is, however, rather lengthy, because of the 2-hour digestion period for the coagulation of the thorium m-nitrophenylarsonate. The use of heat to hasten coagulation is deemed unwise in view of the findings of VanDalen and Graham27 in their work with titanium m-nitrophenylarsonate. The method for thorium with m-nitrobenzoic acid combines the speed of the 8-hydroxy- quinoline procedure and the selectivity of the m-nitrophenylarsonic acid procedure. The range of allowable pH values for the precipitation of thorium by m-nitrobenzoic acid is practicable, although admittedly it is narrow. m-Nitrobenzoic acid is a very selective precipitant, being peculiarly adapted to the determination of quadrivalent elements (except titanium).Under appropriate conditions, thorium, zirconium and hafnium, and quadrivalent cerium and plutonium, can be quantita- tively precipitated by this reagent. It does not, however, precipitate tervalent cerium11 or plut~nium.~l Mercury in either state can be precipitated in a cold but not a hot solution, and stannous and stannic tin can interferelOJ1 (probably owing to hydrolysis effects). As mentioned earlier, m-nitrobenzoic acid has, with proper control of the pH of precipitation, been used to separate thorium and zirconium. Crepaz and Mar~hesini~~ reported that barium, aluminium, lead, copper and lanthanum, amongst other elements, give precipitates with m-nitrobenzoic acid (introduced into the solution as the ammonium salt) at a pH of 3.3.We have found that these elements (and also tervalent rare-earth elements obtained from mischmetal) give no precipitates that contain nitrobenzoate over the pH range 2.1 to 8.0, when the elements are present in the initial solution at concentrations up to at least 40 mg per 50 ml of solution (four times the concentration of the thorium, in which we were interested). NeishlO or Osbornll or both also found that m-nitrobenzoic acid does not precipitate these elements. Further evidence that tervalent lanthanide elements need not interfere in the determination of thorium by m-nitrobenzoic acid is furnished by the results in Table I. Osbornll reported that m-nitrobenzoic acid quantitatively precipitates thorium from a solution that is 0.02 N in nitric acid; such a solution has a pH at the lower end of the range of pH that we found appropriate (pH 1.8 to 2.6).Crepaz and Marche~ini,~~ in a study of the precipitation of thorium by m-nitrobenzoic acid in relation to the pH of the solution after precipitation, reported that no precipitate is formed when the pH is less than 1.9, that in the pH range 1.9 to 2.4 thorium is precipitated as the tetra-(m-nitrobenzoate), but not quanti- tatively, and that when the pH is above 2-4 thorium is precipitated quantitatively but not solely as the m-nitrobenzoate (the precipitate being a mixed one of indefinite composition).* * Murthy et u Z . , ~ * in a study of the use of m-nitrobenzoic acid for the separation of thorium from uranium, used a pH of 2.6 to 2.8 for the precipitation of thorium.Their method of analysis would not, however, distinguish between a thorium nitrobenzoate precipitate and hydrous thoria; their precipitate may well have been a mixed one. These workers reported that a t pH values below 2.6 the precipitation of thorium by m-nitrobenzoic acid was “incomplete.”422 GRAHAM AND LARRABEE : POLAROGRAPHIC DETERMINATIONS OF THORIUM [VOl. 82 According to these workers, then, there is no pH range over which thorium can be quantita- tively taken out of solution by m-nitrobenzoic alcid in the form of a precipitate of constant composition. Our results confirm their finding that the m-nitrobenzoate of thorium is the tetra compound; that, at a pH of about 2-4-we would say 2*&and above, the precipitate is not pure thorium tetra-(m-nitrobenzoate), the contaminant probably being hydrous thoria ; but, by contrast, we precipitated thorium tetra-(m-nitrobenzoate), almost quantitatively (99.4 per cent.), at a pH somewhat below 1.9 (as did Osbornll) and, more particularly, precipitated a constant amount of the tetra-(mmitrobenzoate) over a significant range of pH (Fig.1). Each of the curves in Fig. 1 has a straight-line portion that “breaks” at the same pH, 2.6. At this pH the hydroxyl-ion concentration of the solution has apparently become sufficiently high for hydrous thoria to begin to be precipitated. The results in Table I indicate that the polarographic method with m-nitrobenzoic acid is applicable to the determination of thorium (when present at about the 3 per cent. level) in magnesium-base alloys containing zinc (about 6 per cent.) and rare-earth elements (up to about 21 per cent.).If zirconium should be present in the alloy, a prior separation from thorium would be required. We gratefully acknowledge the suppport of this work by the Defence Research Board of Canada (D.R.B. Grant No. 7510-05; Project 1344-75-10-04) and also assistance provided by the Research Council of Ontario. We are also much indebted to Mr. E. VanDalen (now of Canadian Industries Limited, Millhaven, Canada), who carried out experiments with m-nitrophenylarsonic acid. 1 . 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.32. REFERENCES Moeller, T., Schweitzer, G. K., and Starr, D. D., Chem. Rev., 1948, 42, 63. Kolthoff, I. M., and Lingane, J. J., “Polarography,” Second Edition, Interscience Publishers Inc., Komirek, K., Chem. Listy, 1950, 44, 255. Moeller, T., and Fritz, N. D., Anal. Chem., 1948, 20, 1055. Kolthoff, I. M., and Johnson, R. A., J . E1ectroi:hem. Soc., 1951, 98, 138. Hollingshead, R. G. W., “Oxine and Its Derivatives,” Butterworths Scientific Publications, Eswaranarayana, N., and Raghava Rao, Bh. :S. V., Anal. Chim. Acta, 1954, 11, 339. Berg, R., and Becker, E., Z . anal. Chem., 1940, 119, 1 . Mannelli, G., Atti X ” Congr. Intern. Chim., 19313, 2, 718. Neish, A. C., J . Amer. Chem. Soc., 1904, 26, 780. Osborn, G. H., Analyst, 1948, 73, 381. Belcher, R., and Wilson, C. L., “New Methods in Analytical Chemistry,” Chapman and Hall Ltd., Murthy, T. K. S., Lakshmana Rao, B. R., and Raghava Rao, Bh. S. V., J . Ind. Chenz. SUC., 1950, Venkataramaniah, M., and Raghava Rao, Bh. S. V., 2. anal. Chem., 1951, 133, 248. Datt, N. K., and Chowdhury, A. K., Anal. Chim. Acta, 1955, 12, 515. Rodden, C. J., and Warf, J. C., in Rodden, C. J , , Editor, “Analytical Chemistry of the Manhattan Project,” McGraw-Hill Book Co. Inc., New ’fork, 1950, p. 178. Lingane, J . J., and Laitinen, H. A., Ind. Eng. Chem., Anal. Ed., 1939, 11, 504. Kamm, O., and Segur, J . B., Organic Syntheses,” Collective Volume I, Second Edition, John Ishibashi, M., and Fujinaga, T., Bull. Chem. Sot:. Japan, 1950, 23, 25. Hollingshead, R. G. W., op. cit., p. 18. Carruthers, C., I n d . Eng. Chem., Anal. Ed., 194:3, 15, 412. Stone, K. G., and Furman, N. H., Ibid., 1944, 16, 596. Parks, T. D., and Lykken, L., Anal. Chem., 1948, 20, 1102. Moeller, T., and Ramaniah, M. V., J . Anzer. Chem. Soc., 1953, 75, 3946. Phillips, J. P., Chem. Rev., 1956, 56, 271. Frere, F. J , , J . Anter. Chem. Soc., 1933, 55, 4362. VanDalen, E., and Graham, R. P., Anal. Chim .4cta, 1955, 12, 489. Dennis, S. F., Powell, A. S., and Astle, M. J., J . Amer. Chem. Soc., 1949, 71, 1484. Page, J. E., Smith, J. W., and Waller, J. G., 1. Phys. G. Colloid Chem., 1949, 53, 545. Crepaz, E., and Marchesini, L., Atti Ist. Venetc Sci., Classe Sci. Nut. e Nut., 1951-52, 110, 91. Harvey, B. G., Heal, H. G., Maddock, A. G., and Rowley, E. L., J . Chem. Soc., 1947, 1010. Crepaz, E., and Marchesini, L., Atti 1st. Veneto Sci., Class8 Sci. Nut. e Nut., 1951-52, 110, 33. New York, 1952, p. 446. London, 1954, Chapter 25. London, 1955, p. 2. 27, 610. Wiley and Sons Inc., New York, 1941, p. 391. BURKE CHEMICALABORATORIES HAMILTON COLLEGE, MCMASTER UNIVERSITY HAMILTOX, ONTARIO, CANADA October 22nd, 1956

ISSN:0003-2654    

DOI:10.1039/AN9578200415

出版商:RSC    

年代:1957

数据来源: RSC