January, 19581 USES OF ION-EXCHANGE RESINS. A REVIEW 9 The Determination of Uranium by Solvent Extraction Part I. The Separation of Uranium-233 from Irradiated Thorium as the Diethyldithiocarbamate Complex BY W. H. HARDWICK AND M. MQRETON-SMITH (Chenaical Engineering Divisioiz, A .E.R.E., Harwell, n y . Didcot, Bevks.) The analytical separation of trace amounts of uranium by solvent extraction has been investigated, with particular reference to the determina- tion of uranium-233 in irradiated thorium. The methods that have been evolved are applicable to the separation of uranium from many other elements under appropriate conditions. The use of isobutyl methyl ketone in con- junction with sodium diethyldithiocarbamate is described. URAXIUM-233 is produced by neutron irradiation of thorium as follows- n)Y P B 232Th -+ 2zTh 233Pa ---+ 233U 23.3 nz 27.4 d Uranium233 itself decays by alpha emission with a half-life of 1.62 x lo5 years.One microgram of the isotope gives 2.09 x lo4 alpha disintegrations per minute and hence it10 HARDWICK AND MORETON-SMITH : THE DETERMINATION OF [Vol. 83 can be determined in trace amounts by alpha counting. The thermal-neutron capture cross- section for thorium-232 is 7 barns, so that the arnount converted to uranium-233 in neutron fluxes of the order of 10l2 neutrons per sq. cm per second is small. Irradidation for 1 year and then “cooling” for long enough to allow most of the protactinium-233 formed to decay to uranium-233 produces only 0.22 g of uraniuni-233 per kg of thorium. Fission products are also formed by the fission of a very small amount of the uranium-233.Hence the analytical control of a process in which these srnall amounts of uranium-233 are separated from irradiated thorium1 will involve the determination of from 0.01 to 100 pg of uranium-233 per ml in approximately 0.7 M solutions of thorium nitrate. The solutions are also 0.05 M in fluoride ions to catalyse the dissolution of thorium metal or oxide. The accuracy required is from & 10 to Before uranium-233 can be determined by alpha counting, it must first be separated from thorium, which would reduce the alpha count by absorption, and from the decay products of thorium, which would contribute to the alpha count. The relative amounts of the daughter products of the decay of thorium will depend upon the interval between purification of the thorium before irradiation and the analysis.The total equilibrium alpha activity of the chain is about 1.8 x lo4 disintegrations per minute per mg of thorium, or about 3 x los disintegrations per minute in 0.7 M solutions of thorium. This compares with 2-09 x lo2 disintegrations per minute for 0.01 pg of uranium-233. An excellent review of methods of analysis for uranium has been given by Rodden.2 For the present requirements of rapid separation of traces of uranium-233, it seemed that a solvent-extraction procedure would be most suita.ble. Provided that it were simple enough, it would have the advantage of confining the uranium-233 and fission products to small volumes of liquids and would permit aliquots of the solvent phase containing the purified uranium-233 to be evaporated directly for alpha counting. The equipment required is very simple and this reduces the risk of radioactive cross-contamination.2 per cent. at the 100 pg per In1 level. EXPERIME NTAL A solvent-extraction procedure that could separate trace amounts of uranium from large amounts of thorium in a single extraction was already available, based on the work of Betts and Cahn.3 This was the extraction by isobutyl methyl ketone (hexone) of the uranium - diethyldithiocarbamate complex from a thorium nitrate solution made 2 M in ammonium nitrate at pH 2.5 to 3.0. Unfortunakely, the thorium daughters lead-212 (half- life 10.64 hours) and bismuth-212 (half-life 60.5 minutes) are also extracted. The alpha activity that results from the decay of the extracted bismuth-212 and from that bismuth-212 formed in the solvent by decay of lead-212 does not permit the determination of uranium- 233 in the extract by counting until this unwanted activity has decreased to a relatively insignificant level.This requires at least 4 days, and much longer for very small amounts of uranium-233, and is therefore impractical for plant control. Neither lead nor bismuth is extracted by hexone, whereas uranium, and to a lesser extent thorium, is extractedby this solvent. A method was developed in which uranium is separated from lead and bismuth by extraction with hexone. The thorium that is extracted with the uranium is then separated by washing, under the conditions given above, after the addition of sodium diethyldithio- carbamate.A description of the experiments carried out to establish the four basic assumptions upon which the method depends is given below. EXTRACTION OF URANIUM BY HEXONE FROM THORIUM NITRATE SOLUTIONS- To achieve the highest extraction of uranium-233 it was proposed to saturate the aqueous phase with ammonium nitrate. In the absence of thorium nitrate, partition coefficients for uranium-233 (ratio of concentration in solvent phase to concentration in aqueous phase = I<) for extraction from saturated ammonium nitrate solution into hexone at acidities of 0.25 and 0.5 N were measured by alpha counting. The values found for K , were 14.5 and 12-5. The effect of thorium and nitric acid on K , is indicated in Table I.I t is apparent that three successive extractions with equal volumes of solvent are necessary to extract more than 98 per cent. of the uranium present in the solution initially 3 N in nitric acid, but measurements of extraction by nitric acid under these conditions showed KHNO3 values of 1 and 2 for two successive extractions, and hence, after each extraction, the values of K , will be improved owing to reduced acid concentrations. These vailues of K , were high enough to meet the requirements for the proposed method.January, 19581 URANIUM BY SOLVENT EXTRACTION. PART I 11 TABLE I EXTRACTION OF URANIUM AND THORIUM FROM SATURATED SOLUTIONS OF AMMONIUM NITRATE BY HEXONE Concentration of thorium,* mg per ml 176 176 176 196 176 180 150 75 50 Concentration of nitric acid,* N 3.0 2.0 1.5 1.0 0.75 0.5 1.0 1.0 0.6 K U KTh 3.9 0.28 - 0.25 6.8 - 7.3 0.24 8.4 - - 0.21 8-5 - 7-6 - 8-1 - * The concentrations are for the aqueous phase before saturation with ammonium nitrate, which The extraction of thorium was measured in this and other experiments; typical results are included in Table I and show a value for K , of between 0-21 and 0.28.Three half- volume extractions would therefore remove about one-fifth of the thorium. For solutions of thorium containing 25 to 100 pg of uranium-233 per ml, 0.25-ml portions or less are adequate for an analysis, but for the lower range of concentrations of uranium-233, 5-ml por- tions are necessary to give enough uranium-233 alpha activity. The small portions can easily be diluted with saturated ammonium nitrate solution and nitric acid to give suitable volumes for extract ion.NON-EXTRACTION OF THORIUM DAUGHTERS BY HEXONE- When thorium nitrate solutions containing daughter activities are equilibrated with hexone, some alpha activity is extracted. This activity does not show the decay character- istics of lead-212 or bismuth-212, but shows growth characteristics associated with thorium-232 and thorium-228. As shown later, this activity is removed at pH 2.8 by washing with an aqueous phase containing sodium diethyldithiocarbamate. doubles the aqueous volume and halves these concentrations. RETENTION OF URANIUM - DIETHYLDITHIOCARBAMATE COMPLEX I N HEXONE- Control ofpH-For the second stage of the analysis, uranium must be held in the hexone phase as the diethyldithiocarbamate complex while thorium is washed into the aqueous phase.Betts and Cahn3 found the optimum pH range was 2.5 to 3.0, since above pH 3.5 thorium begins to be precipitated and, under conditions of low pH, the uranium - diethyl- dithiocarbamate complex becomes unstable. An easy way to control the pH during washing was obviously desirable, and screened methyl orange was a suitable indicator. The hexone containing extracted uranium-233 and acid was stirred with 2 M ammonium nitrate solution and 1 drop of indicator was added. Concentrated ammonia solution was added dropwise until the indicator turned green. Then 0.5 ml of a freshly prepared and filtered solution of sodium diethyldithiocarbamate was added and 3 N nitric acid was added dropwise until the indicator assumed its intermediate colour.As stirring was continued, the pH sometimes rose and the indicator turned green. More acid was added to restore the intermediate colour. Removal o f alpha activity due to the presence of thorium-A 5-ml portion of a solution of thorium nitrate was saturated with ammonium nitrate and stirred with successive 5-ml portions of hexone. The hexone extracts were combined and washed with 2 M ammonium nitrate solution containing sodium diethyldithiocarbamate and the pH was controlled at 2-8 as described above. The solvent was then evaporated and the residue was destroyed with nitric acid, transferred to a tray, prepared for counting and counted. The mean alpha count of this solvent and of others from four similar experiments was equivalent to less than the 0.001 pg of uranium-233 per ml in the original sample.This was acceptable as the limit of alpha interference from thorium. EVAPORATION OF HEXONE CONTAINING URANIUM-233 FOR ALPHA COUNTING- When determining the higher concentrations of uranium-233, it is convenient to evaporate The trays are aliquots of the purified hexone extract directly on flat stainless-steel trays.[Vol. 83 put on rings made of sections of copper tubing of diameter about 1 inch and 4 inch long, which are placed on a hot-plate controlled by a Simmerstat. The trays are also heated from above by an infra-red lamp. To avoid loss by the solvent “creeping” over the edges of the tray, a technique was devised whereby 1 drop of a 20 per cent. solution of ammonium chloride containing about 1 per cent.of a water-soluble glue (Stephen’s Stefix was used) is first placed on the tray and evaporated to dryness under the infra-red lamp. The hexone aliquot and washings from the pipette are added dropwise to the tray, which is kept hot enough to evaporate hexone without spitting. When the solvent aliquot and washings have been evaporated, the tray is heated to redness, thereby volatilising the ammonium chloride and organic matter. When this technique for (evaporating the final solvent aliquot was used, recoveries ranged from 98 to 100 per cent. 12 HARDWICK AND MORETON-SMITH : THE DETERMINATION OF RESULTS In practice, duplicate analyses were carried out, and for the range of 30 to 1OOpg of uranium-233 per ml, agreement between the duplicates was better than 2.5 per cent.In 80 per cent. of the analyses of solutions containing 0-1 pg per ml or less, agreement between duplicates was better than 5 per cent. The total time required for duplicate analysis, includ- ing counting, was about 3 hours. Although this time compares favourably with that required for other methods, such as chromatography, it was thought that the method might be improved by the introduction of ethylenediaminetetra-acetic acid to eliminate the preliminary extrac- tions with hexone. This was only partly successful, but the substitution of 8-hydroxy- quinoline for sodium diethyldithiocarbamate was satisfactory, and its use is described in Part I1 of this series. METHOD REAGENTS- Hexone. Ammonium nitrate, solid. Ammonium nitrate solution, 2 M.Sodium diethyldithiocarbamate solution-A freshly prepared and filtered 20 per cent. Ammonia solutiofi, sp.gr. 0.880. Nitric acid, concentrated and N. Screened methyL orange indicator solution. Anti-creeping soZution-A 20 per cent. soluticm of ammonium chloride containing 1 per aqueous solution. cent. of a water-soluble glue (Stephen’s Stefix was found to be suitable). PROCEDURE FOR 0.01 TO 1 pg OF URANIUM-233 PER ml- By pipette place 5ml of sample solution in a 40-ml centrifuge tube and saturate the solution with ammonium nitrate by adding the solid reagent. Then add 3 ml of hexone and stir the solution for 5 minutes. Spin in a centrifuge and then transfer the hexone layer to a clean 40-ml centrifuge tube. Repeat the extraction with two further 3-ml portions of hexone and combine the three hexone layers.To the combined layers add 5 ml of ammon- ium nitrate solution and 1 drop of screened methyl orange indicator solution and stir. Make alkaline by adding ammonia solution and then add 0.5 ml of sodium diethyldithiocarbamate solution. Add N nitric acid until the aqueous layer is mauve (not red) and stir for 4 to 5 minutes. If necessary, add more acid during this time to maintain the mauve colour. Wash the surface of the aqueous layer with 0.5-ml portions of hexone and add the washings to the hexone in the beaker. Evaporate to dryness under gentle heat from an infra-red lamp in a fume chamber and wash the sides of the beaker twice with 2-ml portions of concentrated nitric acid, evaporating to dryness each time.’Then wash the beaker twice with 4 drops of A T nitric acid each time and transfer the washings to a stainless-steel counting tray. Evaporate the washings to dryness, heat the tray to redness in the flame of a Meker burner and then cool and count. Spin in a centrifuge and then transfer the hexone layer to a 50-ml beaker. PROCEDURE FOR 1 TO 100 /.Lg OF URANIUM-233 PER ml- By pipette place a suitable aliquot of the sample solution in a 40-ml centrifuge tube Make up to approximately 1.5 ml (0-1 ml is suitable for 100 pg of uranium-233 per ml).January, 19581 URANIUM BY SOLVENT EXTRACTION. PART I 13 with N nitric acid and saturate the solution with ammonium nitrate by adding the solid reagent. Continue as for the procedure for 0.01 to 1 pg per ml up to the addition, if necessary, of more nitric acid to maintain the mauve colour. Spin in a centrifuge and then transfer the hexone layer to a 10-ml calibrated flask, taking care to ensure that none of the aqueous layer is transferred. Wash the surface of the aqueous layer with three 0.5-ml portions of hexone and add the washings to the hexone in the flask. Make the volume in the flask up to the mark by adding hexone and mix thoroughly. Put 1 drop of anti-creeping solution into each of four stainless-steel counting trays and evaporate to dryness. By pipette put 0.25 ml of the hexone solution from the flask into each tray and slowly evaporate, keeping the hexone around the spot of ammonium chloride, and then add the washings from the pipette. Heat the trays to redness in the flame of a Meker burner and then cool and count. REFERENCES 1. 2. 3. Nicholls, C. M., and Wells, I., “Progress in Nuclear Energy,’’ Series 111, Pergamon Press Ltd., Rodden, C. J., Anal. Chem., 1953, 25, 1598. Betts, R. H., and Cahn, R. P., National Research Council of Canada Report MX/219, 1946, Received izlay 22nd, 1957 London, 1956, Volume I, p. 223. classified.