396 BURSTALL AND WELLS: COLUMNS OF CELLULOSE AND ORGANIC [Vol. 76 Inorganic Chromatography on Cellulose The Use of Columns of Cellulose in Combination with Organic Solvent Extraction for the Separation of Uranium from Other Metals BY F. H. BURSTALL AND R. A. WELLS (Presented at the meeting of the Society o:a Wednesday, February 7th, 1951) A new method is described for the separation of uranium from other metals; it is based on the extraction of uranyl nitrate with ether containing 6 per cent. v/v of nitric acid in the presence of cellulose. The method has been applied to the determination of uranium in minerals and ores with results that compare favourably in simplicity, rapidity and accuracy with other methods. The behaviour of a number of other anions and cations in the process has been investigated and methods have been devised for overcoming difficulties caused by the presence of some metals and acid materials on the extraction process.IT has been known for many years that uranyl nitrate is soluble in ether and other organic solvents, and this fact has been widely used as a basis for the separation of uranium for the determination of the element in minerals, ores and other products. The method, however, does not always give a pure extract of uranyl nitrate and further treatment is necessary. In early work with paper strips and organic solvents containing nitric acid, Arden, Burstall and Linsteadl found that very good separations of uranium from a large number of other metals could be achieved. The use of paper strips limits the quantity of material that can be used, but in later work by Burstall, Davies and Wells2 it was found possible to separate much larger quantities of metallic products by using columns of cellulose pulp packed in organic solvents and contained in glass tubes, and the application to uranium was briefly indicated. A detailed account is now given of the use of ethyl ether containing 5 per cent. v/v of nitric acid as solvent and cellulose pulp asl adsorbent, for the quantitative extraction of uranium from a variety of materials.The method consists in preparing a nitrate solution of the sample for analysis in dilute nitric acid and transferring this mixture to the top of a column packed with cellulose in the presence of the ethyl ether - nitric acid solvent. The solvent is allowed to percolate through the column; uranyl nitrate is dissolved and passes quantitatively into the liquid eluent, whereas a larger number of other metals remain stationary or move only slowly in comparison with uranium.The uranyl nitrate is readily recovered from the eluent after it has been diluted with water and the solvent removed by distillation ; the uranium can be determined by gravimetric, volumetric, colorimetric, polaro- graphic or fluorimetric techniques. The method has proved widely applicable in the separation of uraniumL from minerals, ores and other products. The procedure is simple and rapid and has given results comparable in accuracy with those of other methods. The mechanism of the separation process is complicated and is dependent chiefly on the following factors- (a) Selective extraction of uranyl nitrate by ,!he solvent-Nitrates of metals other than uranium (e.g., ceric, ferric, mercuric and thorium nitrates) also dissolve in ether, but under the conditions of the extraction these metal salts are retained by the cellulose.Ceric nitrate is reduced to the cerous condition and is then retained, mercuric nitrate is also retained by the cellulose, and movement of ferric and thorium nitrates depends largely on the amount of water present in the system. The solubilities of metal nitrates are substantially dependent on the concentration of nitric acid in the solvent, an increase in nitric acid concentrationJuly, 19511 SOLVENT EXTRACTION FOR SEPARATING URANIUM 397 causing an increase in solubility.The use of 5 per cent. v/v of nitric acid in ethyl ether for the extraction of uranium has been found most satisfactory for analytical work; it provided a suitably pure uranyl nitrate with a small amount of solvent. (b) Partition between the nitrates dissolved in the organic solvent and water present in the celldose-This is also important ; indeed, water is a key factor in chromatography with organic solvents and solid adsorbents. (c) Adsorption of metals on the cellulose-This also plays a part in the separation. Uranyl nitrate in ethyl ether containing 5 per cent. v/v of nitric acid is not adsorbed although many other metals are strongly retained on the cellulose. This chemical adsorption is due to the presence of reactive groups in the cellulose, and can be increased by pre-treatment such as boiling the cellulose with dilute nitric acid.The foregoing factors are still being investigated in order to gain further details of the mechanism of chromatographic separation, but lack of information on this aspect of the process does not affect the value of the experimental technique in the analysis of uranium. EXPERIMENTAL SOLVENT PREPARATION AND RECOVERY- The solvent was freshly prepared each day by mixing ethyl ether, free from peroxide, with concentrated nitric acid, sp.gr. 1.42, in the proportion of 5 ml of acid to 100 ml of ether. The ether was recovered by addition of water followed by distillation and purified for re-use by neutralisation with caustic soda, treatment with alkaline permanganate, distillation, drying over caustic soda and a final fractionation. Each batch of ether was tested for peroxide with potassium iodide solution before use.Estimations by the Fischer method indicated that the water content was less than 0.1 per cent. PREPARATION OF CELLULOSE- The cellulose pulp was prepared by boiling 450 g of cellulose (Whatman Ashless Tablets) with 3 litres of 5 per cent. v/v nitric acid for 2 minutes. Other forms of cellulose can be used, for example, Whatman No. 1 Waste Paper Clippings, which require boiling for 20 minutes, however. The pulp is filtered and washed free from nitric acid with water and then washed with 2 litres of ethyl alcohol and finally with about 2 litres of ether. After draining at a filter-pump, the pulp is stored in a closed container and is ready for use.Experiments were carried out to ascertain the amount of inorganic impurities extractable from a column of cellulose pulp, 2 cm in diameter and 25 cm long, by the mixed ether - nitric acid solvent. As shown in Table I, the pulp made from No. 1 clippings is sufficiently pure for most purposes provided that the column is first washed through with 250 ml of solvent. TABLE I INORGANIC MATERIAL EXTRACTABLE FROM CELLULOSE Volume of ether - nitric acid mixture passed Weight of ignited Source of pulp through 25-cm column, residue from eluent, Dl1 g Ashless tablets . . .. .. 250 o*oooo J 1st 250 0.0008 2nd 250 o*ooo 1 No. 1 clippings . . . . THE EXTRACTION TUBE- The adsorption apparatus consists of a glass tube about 2 cm in diameter and 40 crn long, the upper end being widened to form a funnel to allow easy transfer of material to the tube.The lower end of the tube is narrowed and is closed by a short length of polyvinyl chloride tubing carrying a screw clip or a tap. The inside surface of the glass extraction tube was treated with dichlorodimethyl silane, (CH,),SiCl,, which conferred strong water- repellent properties to the glass surface (see Burstall, Davies and Wells,, Part I11 of this series, p. 180). Another method of achieving water-repellent properties was to use Fluid 200 (Albright and Wilson Ltd.) in carbon tetrachloride solution; with this, the tube must be heated to 250" C to provide a stable water-repellent film. The adsorption tube was packed with cellulose pulp in the following way. The tube was first half-filled with ether -nitric398 BURSTALL AND WELLS: COLUMNS OF CELLULOSE AND ORGANIC [VOl.76 acid mixture and cellulose pulp was added in sniall quantities. Each portion of pulp was gently pushed down with a glass rod flattened at one end to form a plunger of diameter slightly less than that of the tube; a brisk up and down movement of the plunger then served to b r e a up any aggregated pieces of pulp. A column properly packed in this way allowed ether solution to pass freely through the cellulose at a rate of approximately 100ml in 20 minutes with the end of the tube completely open. Cellulose columns from 5 to 8cm in length were ultimately used, but initial experiments were made with 25-cm columns. PREPARATION AND TRANSFER OF SAMPLE- A solution of the sample was prepared in aqueous nitric acid.Use of a high concentration of nitric acid favoured the rapid extraction of uranium in a narrow band. The movement of many impurities, however, decreased with the concentration of the nitric acid. A suitable compromise was found to be a solution containing 25 per cent. v/v of nitric acid. The method of preparing a nitric acid solution of the sample varied with the type of mineral, and is described later separately for each type of ore, together with any special treatment of the solution found necessary. The acid solution of the sample could be transferred directly to the top of the column, but the following method was usually adopted. Sufficient cellulose pulp was added to the sample solution to ensure complete adsorption; the amount required was 2 g of pulp for 10 ml of 25 per cent. v/v nitric acid.This wad of cellulose was then transferred to the top of the tube, gently beaten with a plunger and prelssed down to fonn a continuous part of the column. THE EXTRACTION OF URANIUM- After transfer of the sample, successive small volumes (10 ml) of the ether containing 5 per cent. v/v of nitric acid were added and the solvent was allowed to flow through the tube. This procedure was continued until 150ml of solvent eluent had been collected, this quantity being adequate for an 8-cm cellulose column, although proportionately larger quantities must be used with longer columns. ‘The solvent was added in such a manner that the level of the solvent a t the top of the column fell to the top of the cellulose packing, but not below, between successive additions. The extraction column was not allowed to run dry at any stage.To the eluent from the extraction was added water in the proportion of 50 ml of water to each 100 ml of solvent and the ether was removed by distillation. The uranium was then determined by one of the following methods: (a) evaporation and ignition to U,O,; (b) precipitation with oxine, filtration amd ignition to U,O,; (c) evaporation with sulphuric acid and heating to fumes followed by dilution, reduction in a Jones reductor and titration with ceric sulphate; (d) evaporation with sulphuric acid and heating to fumes followed by a colorimetric estimation with alkaline solution and hydrogen peroxide; (e) as method (d) but with a polarographic determination in place of the colorimetric method.THE BEHAVIOUR OF ELEMENTS OTHER THAN URANIUM- A study has been made of the behaviour on *a cellulose column of other elements under the conditions used for uranium extraction and of the effect of these elements on the extraction of uranium. The following observations are based on a preliminary study. A more detailed account will be given in a further paper. The nitrates of Li, Na, K, Cs, Rb, Cu, Ag, Be, Mg, Ca, Sr, Ba, Ra, Zn, Cd, Al, Y, La, Ce, Pr, Nd, Sm, Eu, Ho, Er, Ga, In, T1, Ti, Hf, Ge, Sn, Pb, Nb, Ta, Cr, W, Te, Mn, Fe, Co and Ni all remained stationary or moved only very slightly. These metals, therefore, do not interfere with the estimation of uranium. The reimaining elments are dealt with separately. Gold-The dilute solution of gold prepared by the action of nitric acid on gold metal was partly reduced by the cellulose of the column and gave a purple tint to the absorbent.There was also a tendency for colloidal gold to pass through the column. Prior reduction of the gold by treatment of the sample solution with ferrous sulphate resulted in its complete retention at the top of the extraction column. In presence of chloride, gold was extracted readily from the column. Mercury-Mercury was found to move only in the mercuric state and then not sufficiently rapidly to affect the estimation of uranium. Mercuric salts are partly reduced by the ether - nitric acid solvent in presence of cellulose, and this aided the retention of mercury. A solutionJuly, 19511 SOLVENT EXTRACTION FOR SEPARATING URANIUM 399 of 0.5 g of mercuric nitrate in 5 ml of 25 per cent.v/v nitric acid was extracted with ether containing 5 per cent. v/v of nitric acid through a 10-cm cellulose column. After the passage of 200 ml of solvent, mercury was detected 3 cm from the bottom of the column, but not in the eluent. Selenium, arsenic, antimony and bismuth-Selenium, arsenic, antimony and bismuth in a column all moved, but not at sufficient speed to be extracted with uranium. For example, bismuth nitrate had moved only 5 cm down a column after the passage of 200 ml of ether containing 5 per cent. v/v of nitric acid. The effect of large amounts of arsenic on the extraction of uranium will be dealt with in a later paper. Cerium-Ceric nitrate is appreciably soluble in ether - nitric acid mixtures; in addition, its coefficient of partition between ether - nitric acid mixtures and water is high.Since ceric nitrate is absorbed by cellulose from ethereal solutions to a small extent only, steps must be taken to ensure that any cerium is reduced to the cerous state, in which form it is insoluble in most organic solvents. Reduction of ceric nitrate takes place in ether solution in the dark and in the absence of added acid, but the process is slow. For the extraction of uranium from monazite sands, ferrous sulphate was first added to reduce cerium, but a better method was to boil the original solution in dilute nitric acid with hydrogen peroxide. Thoriztm-Thorium was extracted in small amounts with ether containing 5 per cent.v/v of nitric acid, but the extraction was very sensitive to the concentration of acid in the solvent. The use of ether containing 3 per cent. v/v of nitric acid permitted uranium to be extracted completely before thorium was detected in the eluent, as shown in Table 11. TABLE I1 EXTRACTION OF THORIUM Volume of test Weight of Tho, Nitric acid in Length of Weight of Tho, solution (in extracted with solvent, column, in test solution, 40% HNO,), 400 ml of solvent, % cm g ml g 3 15 0.450 5 nil 5 25 1.027 10 0.170 When uranium is present as phosphate, as in monazite sands, special conditions must be observed for its extraction; this is referred to later. Zirconium-Zirconium was also extracted by ether containing nitric acid. Table I11 shows the results of an extraction of a solution of zirconyl nitrate containing the equivalent of 0.19 g of ZrO, dissolved in 3 ml of water and 2 ml of nitric acid, sp.gr.1.42, when ether containing 5 per cent. v/v of nitric acid was used as solvent. The extraction of zirconium was inhibited by the presence of a number of anions, e.g., phosphate, sulphate, oxalate and tartrate. The use of tartrate for retention of zirconium in the analysis for uranium in zircon- bearing minerals is mentioned later in this paper. TABLE I11 EXTRACTION OF ZIRCONIUM Fraction of eluent r A 3 1st 2nd 3rd Total Proportion 200 ml 200 ml 200 ml 600 ml extracted, % Weight of ZrO, extracted . . 0.8 mg 26.0 mg 30.4 mg 57.2 mg 31 Scandium-The behaviour of scandium was similar to that of thorium.The movement of scandium can be inhibited by the addition of tartrate. Tin-In nitric acid tin is precipitated as insoluble meta-stannic acid, which does not move in the column. But in order to avoid the risk of occlusion of uranium when dealing with minerals containing large amounts of tin, it is best 'to remove the tin by volatilisation as the iodide. In presence of chloride, tin was extracted very readily from a cellulose column. Vanadiztm-Vanadium was immobile in a cellulose column provided that the ethereal solvent used is free from peroxides. In the presence of ether peroxides, a pink peroxy- vanadium compound was formed and moved rapidly down the column. The presence of400 BURSTALL AND WELLS: COLUMNS OF CELLULOSE AND ORGANIC [Vol. 76 reducing agents such as ferrous sulphate in the column converted the pink compound to a non-moving vanadium salt.Phosphorus-Phosphoric acid was readily extracted from cellulose by ether - nitric acid mixtures. In combination with metallic radicals, however, phosphoric acid was much more strongly retained; ferric nitrate has been found a particularly useful complexing agent for this purpose. A small amount of the acid, presumably resulting from the dissociation of ferric phosphate, was, however, still extracted in trace amounts. The presence of a trap of freshly prepared meta-stannic acid in the column reduced the amount of extracted phosphoric acid further, but trace amounts were still detectable in the eluent. The presence of phosphoric acid inhibited the extraction of uranium, but addition of ferric nitrate overcame this effect.Molybdenum-The behaviour of molybdenum in cellulose columns appears to vary with a number of factors. When added as a dilute nitric acid solution of ammonium molybdate to a cellulose column and extracted with an ether - nitric acid solvent, the bulk of the molybdenum moved only slightly, but a low concentration of molybdenum was detect- able in the eluent. In sunlight there was a strong tendency for molybdenum to be reduced t o molybdenum blue and this reaction was catalysed by the presence of uranium. There are indications that there are two forms of the blue complex, one of which was almost immobile and another that was readily extracted from the column. There is also some evidence to show that in concentrated solutions a molybdenum - uranium complex is formed.This behaviour results in a small but definite quantity of molybdenum being present in uranium oxide samples extracted from molybdenum-bearing ores. The proportion of molybdenum in twenty such samples of oxide varied between 7 and 200 p.p.m. Although small, these quantities of molybdenum interfered with the vo'lumetric estimation of uranium by catalysing the aerial re-oxidation of U"" to UO," ions. Decrease in the concentration of nitric acid in the solvent or in the aqueous test solution showed little effect. Similarly, carrying out the extraction in the absence of sunlight gave no improvement. It was found that molybdenum could be reduced to an insoluble trioxide by standing it overnight with an excess of ferrous sulphate.The large excess of ferrous sulphate required, however, made this method in- convenient, and other methods, which have been more successful, will be described in a further paper. The platinum metals-Of the six platinum metals, osmium was neglected, since once it is in solution it is removed by evaporation with nitric acid. Iridium and rhodium were not extracted and hence gave no trouble. Platinum, palladium and ruthenium behave differently. Ruthenium was fused with potassium hydroxide and a solution of the melt was acidified with nitric acid. The resulting suspension of ruthenium hydroxide was absorbed on cellulose and treated with ether - nitric acid solvent. 'The ruthenium appeared to be completely retained at the top of the column, but traces were found in the ethereal eluent.Platinum behaved in a similar manner to ruthenium, the main bulk of the platinum being retained at the top of the column, although small quantities were extracted. Palladium in nitric acid solution was readily extracted from a cellulose column. Reduction of the platinum and palladium solutions with ferrous sulphate before extraction was only partly successful in overcoming this difficulty. The bulk of the paliadium was retained, but traces of platinum and palladium were still found in the eluent. SuZphate-Small quantities of sulphuric acid do not appear to have any appreciable effect on the extraction of uranium, but this question has been the subject of a fuller investigation that will be described in a later paper. Free dphuric acid under normal conditions was retained at the top of the cellulose column.Halides-Halide ions must be absent from samples used for the estimation of uranium since extraction of other elements is greatly increased. Under normal conditions hydrochloric acid is retained in the column. Both hydrobrornic acid and free bromine move slowly down the extraction column. Hydriodic acid and iodine behave similarly. RESULTS- Initial experiments were made with nitrate solutions prepared from weighed quantities of pure U,O, with and without added impurities. With a 25-cm column of cellulose,. 250 ml of ether - nitric acid solvent were necessary for the complete extraction of uranium, in the absence of phosphate, as shown in Table IV. The presence of phosphate slowed the rateJuly, 19511 SOLVENT EXTRACTION FOR SEPARATING URANIUM 401 of extraction of uranium considerably, but the addition of ferric nitrate complexed phosphate sufficiently to allow uranium extraction to proceed normally.TABLE IV EXTRACTION OF URANIUM FROM SYNTHETIC MIXTURES Weight of U,08 g Elements present taken, Uranium . . .. .. .. .. 0.0975 Uranium . . .. .. .. 0.1721 Uranium . . .. .. Uranium + 0.2 g of each of Fe, Zn, Mn, .. .. .. 0.9992 * Uranium + 0.3 g of Fe(NO,), . . .. 0.0808 Cr, V and Cu nitrates . . .. .. 1-0053 Weight of U,O, found, U,08 found, f3-0970 99-5 0.9960 99.7 0.1721 100.0 100.0 0.0808 g % 1-0043 99.9 APPLICATION TO ANALYSIS OF SILICEOUS MATERIALS The method was then applied to a number of low-grade siliceous materials. These were all ores that could be dissolved completely by the action of nitric and hydrofluoric acids, Fluoride was removed from the test solution by repeated evaporations with con- centrated nitric acid and the residue was finally dissolved in 10 ml of 25 per cent.v/v nitric acid. The weight of sample taken for analysis varied between 0.5 and 5 g; in Table V results are compared with those found by standard methods of chemical analysis. TABLE V EXTRACTION O F URANIUM FROM SILICEOUS ORES ON 25-CM COLUMNS Sample 1 2 3 4 5 6 7 8 9 10 U,O, by cellulose column 0.36 0.29 0.20 0.20 0.36 0.17 0.38 1.39 0.77 5-03 % U,08 by standard chemical methods3s4 (mean value), 0.35 0.28 0.19 0.22 0.36 0.16 0.39 1.41 0.73 4.98 % The effect of decreasing the column length was then investigated and results with a It will be observed that for some estimations 5-cm column of cellulose are shown in Table VI.TABLE VI EXTRACTION OF URANIUM Sample 10 10 10 10 11 11 11 11 Weight of sample, g 0.5670 0.5019 2.5024 0.5180 2.7869 2.5933 2-4818 2.2242 FROM SILICEOUS ORES ON 5-CM COLUMNS U,08 by standard Volume of U,O, by cellulose chemical methods**P ether, column, (mean value), ml % % 250 4-92 165 100 100 100 2.13 100 100 2.15 4-98 ::;; } 75 4-96 the amount of solvent used was only 75 ml. Further experiments (Table VII) in which known weights of uranium oxide were added to a standard ore indicated that results were slightly low when 70ml of extracting solvent were used; 100ml were used, therefore, in further experiments. Although the results shown in Table VI were considered satisfactory, the ores used in these experiments contained neither vanadium or molybdenum.In view of the tendency for small amounts of molybdenum to be extracted and because of the possible formation402 BURSTALL AND WELLS: COLUMNS OF CELLULOSE AND ORGANIC [Vol. 76 of a solvent-soluble peroxy-compound of vanadium with traces of ether peroxides, further analyses were carried out in which small quantities of vanadium and molybdenum were added. The results, as shown in Table VIII, were not unsatisfactory, but during these estimations it was noted that vanadium moved rapidly as a pink band and had usually TABLE VII EXTRACTION OF URANIUM ADDED TO SILICEOUS ORES Weight of U,O, Weight of Volume of ether Weight of U,O, added, vanadium added, used, found, mg mg ml mg 20 10 70 19-85 20 3 80 19.99 20 10 80 19.91 reached the bottom of the column before 100ml of ether had passed through the column.Movement of molybdenum under the same conditions appeared to be slight. Addition of ferrous sulphate to the dilute nitric acid solution of the ore maintained the vanadium in a reduced immobile form and did not interfere with the extraction of uranium (Table IX). TABLE VIII THE EFFECT OF VANADIUM AND MOLYBDENUM ON THE EXTRACTION OF URANIUM Sample 10 11 11 11 12 12 13 Weight of sample, g 0.9665 2.5000 2.4494 2.5012 3.0526 2.4996 2.4994 Volume of ether used, ml 80 100 100 100 80 100 100 Molybdenum added, mg nil 25 nil 25 nil 25 25 Vanadium added, mg 5 25 25 nil 5 25 25 U,O, found by cellulose column, Yo 4-95 2.08 2.17 2-07 0.63 0.66 0.54 U30, found by standard chemical (mean value), % 4.98 2.13 2.13 2-13 0.66 0.66 0.54 TABLE IX EXTRACTION OF URANIUM AFTER ADDITION OF FERROUS SULPHATE U308 found by standard chemical Sample added, added, cellulose column, (mean value), Molybdenum Vanadium U,O, found by methods3$* mg mg YO Y O 11 25 25 2.13 .2.13 11 nil 25 2.13 2.13 11 25 nil 2-15 2.13 If the strength of the nitric acid in the test solution was allowed to fall there was a tendency for molybdenum to be reduced to a mobile molybdenum blue by ferrous sulphate. In view of this and because of the undesirability of adding solid material to the original solution, further experiments were made in which the column length was increased to 7-5 cm. Columns of this length allowed the uranium extraction to be completed before the pink peroxy-vanadium band reached the bottom of the column.The formation of peroxy- vanadium compounds was later avoided by the iise of fresh peroxide-free ether. RECOMMENDED METHOD FOR SILICEOUS ORES Cellulose PuZP-This should be prepared as (described on p. 397. Ether - nitric acid solvent-A 5 per cent. v/v solution of nitric acid in ether prepared Nitric acid-Concentrated, sp.gr. 1.42. REAGENTS- by mixing 5 ml of nitric acid, sp.gr. 1-42, with 100 ml of dry peroxide-free ethyl ether.July, 19511 SOLVENT EXTRACTION FOR SEPARATING URANIUM 403 PROCEDURE- Add 2 ml of concentrated nitric acid, sp.gr. 1.42, and 5 ml of hydrofluoric acid and evaporate the mixture just to dryness on a hot-plate. Wash the sample with a minimum of water into a 100-ml beaker.After evaporation to dryness, add 10 ml of concentrated nitric acid and again evaporate the solution just to dryness, Redissolve the residue in 8ml of water containing 2ml of concentrated nitric acid. Prepare a cellulose column, as described on p. 397, 7.5 cm in length, and wash it through with 100 ml of ether - nitric acid solvent. Adjust the solvent level in the column until it coincides with the top of the cellulose. Add sufficient cellulose pulp to the solution of the sample to absorb completely the aqueous nitric acid solution. Transfer the wad containing the absorbed sample to the top of the extraction column with the aid of the glass rod. Wash final traces into the column with not more than 10 ml of ether - nitric acid solvent from a wash bottle; care must be taken to avoid “ether creep.” Break up the pulp containing the sample with a glass plunger and gently press it down t o form a continuation of the original column of cellulose.Remove the clip and tubing from the bottom of the column and allow the ether to run out into a 250-ml Kjeldahl flask until the level of ether solution in the extraction tube reaches the top of the cellulose column. Add a further 10 ml of ether - nitric acid solvent to the top of the extraction tube and repeat the procedure with successive 10-ml portions of the solvent mixture until 100 ml of eluent have been collected. Use each 10 ml of ether - nitric acid mixture to wash out the sample beaker. Add 50ml of water to the eluent ether solution and remove the organic solvent by distillation on a steam-bath.Add 5 ml of sulphuric acid and 5 ml of perchloric acid to the aqueous solution and evaporate to fuming. Complete the estimation by any suitable r n e t h ~ d . ~ ~ ~ RESULTS- 100 ml when using a 7.5-cm column. 2.5-g portions of ore are shown in Table X. strated the suitability of the method for the analysis of low-grade siliceous materials. Weigh about 2.5 g of sample into a platinum dish. Stir the mixture thoroughly with a glass rod. It was not found necessary t o increase the volume of ether required for extraction beyond Results obtained on columns of this length with It was considered that these results demon- TABLE X EXTRACTION OF URANIUM FROM SILICEOUS ORES BY RECOMMENDED METHOD Sample 11 11 14 13 13 13 15 16 17 Molybdenum added, mg nil 25 25 25 nil 25 25 25 25 Vanadium added, mg 25 nil 25 25 nil 25 25 25 25 U,O* by Yo cellulose column, 2.14 2-17 0.06 0.57 0.55 0.54 2.09 2.88 3-35 u30, by standard chemical methods394 (mean value), Yo 2-13 3-13 0-06 0.54 0-64 0.54 2-16 2.89 3-16 APPLICATION TO ANALYSIS OF MONAZITE SANDS AND OTHER REFRACTORY ORES The application of the cellulose column technique to the analysis of monazite sands presented several difficulties.A fusion was necessary in order to obtain a nitrate solution of the mineral, and this gave rise to the presence of large amounts of neutral salts in the test solution. Most of the ores contained considerable quantities of phosphate, which inhibited the extraction of uranium; others contained large amounts of zirconium, which was partly extracted. The addition of ferric nitrate to the test solution overcame the retaining effect of phosphate and addition of tartaric acid prevented the extraction of zirconium.A large amount of cellulose was needed to absorb the test solution, the volume of which was increased to cater for the high concentration of potassium salts, but this large wad did not impair theBURSTALL AND WELLS: COLUMNS OF CELLULOSE AND ORGANIC [Vol. 76 404 extraction of uranium. A number of different types of mineral have been studied, including “pure monazites,” consisting mainly of rare earth and thorium phosphates ; “crude monazite,” consisting of a small quantity of monazite with a large quantity of other refractory ores; and zircons, which were mainly zirconium silicate.The development of the method is described immediately below. SOLUTION OF THE SAMPLE- In order to obtain a nitric acid solution of monazite, advantage was taken of the observa- tion that treatment of the sample with hydrofluoric acid followed by fusion with potassium hydroxide gave a melt that was soluble in nitric acid. This procedure was improved by fusing with potassium hydroxide, dissolving in nitric acid and then adding dilute hydrofluoric acid to the solution. Except for a trace of gelatinous silica, complete solution was obtained. Excess hydrofluoric acid was avoided, as it precipitated thorium and the rare-earth fluorides. With a pure monazite, 2 or 3 drops of a 2 per cent. solution of hydrofluoric acid were found sufficient. The same effect was obtained by the addition of potassium bifluoride to the potassium hydroxide fusion, but addition of hydrofluoric acid, as described, was preferred, because potassium bifluoride rapidly attacked the nickel crucible.The amount of potassium hydroxide required for fusion varied with each sample, but for pure monazite a 5 to 1 ratio of potassium hydroxide to sample, heated for 30 minutes at red heat, was found sufficient to ensure complete breakdown, although for crude monazite the ratio was increased to 8 to 1 and the heating time to 1 hour. For the present work, 2.5-g portions of sample were used for each estimation. EXTRACTION OF URANIUM- Initial experiments with a 25-cm column of cellulose showed that uranium could be completely extracted by 300ml of ether containing 5 per cent.v/v of nitric acid. The nitric acid solution of the sample was evaporated t o about 25 ml, 4 ml of water were added and the resultant solution was absorbed on a wad of cellulose. During these experiments it was noticed that a dark orange band, identified as cerium in the ceric state, moved rapidly down the column and coloured the effluent. This did not affect the result if a titrimetric finish was used, but interfered with a colorimetric determination. Since cerous nitrate showed little movement in the column, further experiments were made with a ferrous sulphate trap to reduce cerium. This method was found to hold back cerium satisfactorily and results are shown in Table XI for extractions both with and without the ferrous sulphate trap. TABLE XI EXTRACTION OF URANIUM FROM PURE MONAZITE SANDS USING A 25-CM COLUMN u3°8 by standard chemical Volume of U30, found meth0ds~9~ Sample ether, U30, found, (mean value), (mean value), ml % % % Without ferrous sztlphate trap- 1 500 1 300 1 300 1 300 1 300 With ferrous sulphate &up-- 2 300 2 300 2 300 2 300 2 300 300 300 0.37 0-37 0.37 0.35 0.40 J 0.36 7 0.39 I 0.38 J 0.37 0.40 1 0.38 0*37(5) 0.37 0.37 } 0‘37 0.37 300 0.35 0.35 0.37 300 0.41 0.41 0.36 As about 8 g of cellulose were used in absorbing the solution of the sample, the total In an length of both wad and column, when a 20-cm column was used, was about 50 cm.July, 19511 SOLVENT EXTRACTION FOR SEPARATING URANIUM 405 attempt to use a column of more manageable length, experiments were carried out with a 5-cm column of cellulose.Addition of hydrogen peroxide to reduce cerium was also tried, as being a more convenient method than the use of a ferrous sulphate trap. About 2 ml of 20-volume hydrogen peroxide were added to the nitric acid solution of the sample while it was being evaporated to dryness. Figures shown in Table XI1 indicate that satisfactory extraction of uranium is obtained with a 5-cm column and 100 ml of ether - nitric acid solution. TABLE XI1 EXTRACTION OF URANIUM FROM PURE MONAZITE SANDS WITH A 5-CM COLUMN Sample Length of column, cm 25 25 5 5 5 5 5 Volume of ether, ml 300 300 300 100 100 100 100 U308 found, % 0-37 0-37 0.37 0.37 0.37 0.39 0.34 uZ08 by standard chemical U,08 found methods3 94 (mean value), (mean value), Yo % ] 0.37 0.3 7 0.39 0-36 0-34 0.30 EFFECT OF ZIRCONIUM AND PHOSPHORIC ACID ON THE EXTRACTION PROCESS- The materials so far studied were all pure monazite samples. When the method was applied to crude monazites, particularly those containing a large amount of zircon, difficulty was encountered from some zirconium passing through the column and contaminating the eluent.The results of uranium determinations on samples containing zirconium were very erratic, as shown in Table XIII. TABLE XI11 THE EFFECT OF ZIRCONIUM ON THE EXTRACTION OF URANIUM FROM CRUDE MONAZITES Sample 7 8 9 10 7 7 7 8 10 10 Length of column, cm 25 25 25 25 5 5 5 5 5 5 Volume of ether, ml 300 300 300 300 100 100 100 100 100 100 U308 found, % 0.54 0-06 0.32 0.18 0.42 0-46 0.35 0.07 0.26 0.21 u.3°8 by standard chemical methods3$* (mean value), % 0-48 0.05 0-25 0.18 0.48 0.48 0-48 0.05 0.18 0.18 Accordingly, factors affecting the movement of zirconium were investigated in the following manner.A known weight of zirconium nitrate was dissolved in 10-ml portions of nitric acid of various strengths. Each 10 ml of solution was taken up on a wad of cellulose and extracted in a 5-cm column with 100ml of ether containing 5 per cent. v/v of nitric acid. The movement of zirconium down the column was measured and showed (see Table XIV) an increase with increasing nitric acid concentration. If, however, a nitric acid solution of crude monazite was evaporated to complete dryness, redissolved in 5 or 10 per cent. v/v nitric acid and extracted in a 5-cm column, zirconium was still found in the ether eluent.The increased movement of zirconium under these conditions was not due to a salting out effect of potassium nitrate, as shown in the last line of Table XIV, but in other experiments, ferric nitrate has been shown to produce such an effect. Although zirconium was still extracted from crude monazite dissolved in a 10 per cent. v/v solution of nitric acid after fusion with potassium hydroxide, the amount extracted was very much smaller than if a stronger solution of nitric acid was used. In all further estimations, therefore, the acidity was controlled by evaporating the nitric acid solution of the sample to dryness and redissolving in 20 ml of 10 per cent. v/v nitric acid. As a406 BURSTALL AND WELLS: COLUMNS OF CELLULOSE AND ORGANIC TABLE :XIV [Vol.76 BEHAVIOUR OF ZIRCONIUM WITH VARIATION OF ACID CONCENTRATION Weight of zirconium nitrate taken = 0.1 g Volume of nitric acid solution, ml 10 10 10 10 10 8 20 Strength of nitric acid solution, 10 15 20 25 30 25 10 % v/v Movement of zirconium down column, cm 2.5 2.5 3.0 4.5 > 5.0 5.0 3.0 Added potassium nitrate nil nil nil nil nil nil 20 g small amount of zirconium still passed through the column, investigations were made into means of preventing the movement by complex formation. Addition of phosphoric acid to the nitric acid solution of the sample held back zirconium but slowed down the extraction of uranium. Traps of ferrous sulphate, oxalic acid or activated carbon in the column were not successful; extraction with ether containing 5 per cent. v/v of nitric acid and 1 per cent.v/v of orthophosphoric acid still resulted in zirconium being found in the eluent. Addition of sulphuric acid to the nitric acid solution of the sample did, however, result in complete retention of zirconium and so permitted extraction of uranium. In this procedure the nitric acid solution of the sample was evaporated to dryness, redissolved in 20 ml of 10 per cent. v/v nitric acid. Sulphuric acid, 6 ml of a 1 + 1 solution, was added and the mixture was boiled for 2 or 3 minutes. Zirconium was cornpletely held back on extraction in a cellulose absorption column. The results for a series of crude monazite samples treated with sulphuric acid in the above manner are shown in Table XV. TABLE XV ADDITION OF SULPHURIC ACID AS COMPLEXING AGENT FOR ZIRCONIUM U 3 0 8 by standard chemical methods3~4 Sample Type U,08 found, (mean value), % % 11 11 } 0.05 Zircon concentrate 0.045 97 0.041 Y Y Y Y 0.048 8 Crude monazite 0.048 8 0.047 8 0.051 8 0-05 1 8 0.048 Y Y Y Y Y t 0.18 10 0.16 The amount of sulphuric acid added is large, but this quantity was found necessary to cope with samples 10 and 11, which had a high zircon content.The efficiency of extraction under these conditions was shown by addition o:f a weighed amount of U,O, to a sample of known uranium content and finding the amount: of added uranium extracted, as follows- Sample No. 8: Weight of U30, added = 25 mg. Weight of U308 recovered = 24.8 mg. The results shown in Table XV were found for ores with a low monazite content. When the method was applied to pure monazite, the extraction of uranium was very poor.This was owing to retention of uranium brought about by the large excess of uncomplexed sulphate present in the absence of a high concentration of zirconium. USE OF TARTARIC ACID AND FERRIC NITRATE AS COMPLEXING AGENTS FOR ZIRCONIUM AND PHOSPHORIC ACID- Addition of either oxalic or tartaric acids to th.e nitric acid solution of a sample of monazite before absorption on a wad was found to complex zirconium satisfactorily. The procedure used was to evaporate the nitric acid solution of monazite to dryness, redissolve in 20 mlJuly, 19511 SOLVENT EXTRACTION FOR SEPARATING URANIUM 407 of 10 per cent. v/v nitric acid, add 2 ml of 20-volume hydrogen peroxide, boil for a few minutes, add 3 g of the organic acid and then cool. The use of either acid resulted in a lower rate of extraction of uranium, particularly with pure monazite, but whereas a 50 per cent.increase in the volume of ether - nitric acid solution used in extraction resulted in complete recovery of uranium when tartaric acid was used, results were still low for oxalic acid. Of the two organic acids, oxalic is the more effective complexing agent for zirconium, but tartaric acid retains zirconium sufficiently well to enable a clean extraction of uranium to be achieved. Increasing the concentration of the nitric acid in the solution of the sample before extraction did not materially affect the extraction of uranium, although some increase in the movement of zirconium was apparent. During these estimations it was noticed that, if an iron crucible was used for the fusion, results were satisfactory, but if a nickel vessel was used, the values were low.Experiments were then carried out in which fusion was carried out in a nickel crucible, but iron as ferric nitrate was added to the nitric acid solution of the sample before extraction. Under these conditions uranium was quantitatively extracted. Since a clean, easily removed melt was obtained in nickel crucibles, they were preferred and ferric nitrate was added at a later stage of the estimation. TABLE XVI USE OF TARTARIC ACID AS A COMPLEXING AGENT FOR ZIRCONIUM Sample Type With added tartaric mid- 10 Crude 10 37 1 Pure 1 99 1 99 1 93 1 39 1 19 With added oxalic acid- 10 Crude 10 99 1 Pure 1 79 1 97 Volume of ether, ml 100 100 100 100 100 150 100 150 100 150 100 150 150 Strength of nitric acid used for solution of sample, % V/V 10 10 10 10 25 25 10 25 10 10 10 25 25 Weight of ferric added, found, present, nitrate u30, U308 g % % nil nil nil nil nil nil 5 5 0.18 0.20 0.24 0.27 0.27 0.33 0.32 0.37 0.18 0.18 0.37 0.37 0.37 0.37 0.37 0.37 nil 0.13 0-18 5 0.19 0-18 5 0-32 0.37 5 0.33 0.37 5 0-34 0.37 in Table XVI.the addition of tartaric As a result of these exDeriments. summarised acid and ferric nitrate was included in the procedure recommended for the analysis of uranium in monazite sand. RECOMMENDED METHOD FOR MONAZITE SAND AND REFRACTORY ORES REAGENTS- Celldose @ul@-This should be prepared as described on p. 397. Ether - nitric acid solvent-A 5 per cent. v/v solution prepared as described on p.402. Potassium hydroxide-Solid. Nitric acid-Concentrated, sp.gr. 1.42. Hydrojuoric acid-A 2 per cent. v/v aqueous solution. PROCEDURE- Heat 12.5 to 20 g of potassium hydroxide, according to the type of mineral to be analysed, in a nickel crucible until all the water has been removed. Allow to cool slightly, add 2.5 g of sample and quickly cover the crucible with its lid. Heat the melt slowly to red heat. Continue to heat for 1 hour at bright red heat, occasionally swirling the contents of the crucible. Allow the crucible to cool and wash the contents with water into a 400-ml beaker. Make just acid with nitric acid and then add about 20 ml of concentrated acid in excess. Bring the solution to the boil with constant stirring and slowly add dropwise a 2 per cent.solution408 [Vol. 76 of hydrofluoric acid. Stop the addition of hydrofluoric acid as soon as the solution clears and evaporate to dryness on a steam-bath or under an infra-red lamp. To the residue add 20 ml of water containing 2 ml of concentrated nitric acid and 5 g of ferric nitrate. Heat with stirring until solution is complete. Add 2 :ml of 20-volume hydrogen peroxide and boil for 2 or 3 minutes to reduce cerium. Then add 3 g of tartaric acid, stir and cool rapidly. To the nearly solid mass add about 8 g of cellulose pulp and stir until a homogeneous mixture is attained. Pack an extraction tube to a depth of 5 cm with cellulose and wash the column by allowing 100 ml of ether - nitric acid solvent to flow through it. Adjust the ether level until it is about 10 cm above the top of the column. Transfer the wad containing the sample in small portions to the extraction tube.Break up each portion of pulp with a glass plunger and gently press down to form a continuous column with the original cellulose. Remove the clip and tubing from the botto:m of the extraction tube and allow the ether to run out into a 350-ml Kjeldahl flask until the level of the ether - nitric acid solution in the extraction tube reaches the top of the cellulose column. Add a further 10 ml of ether - nitric acid solution to the top of the extraction tube and repeat the procedure with successive 10-ml portions of the solvent mixture until 150 ml of eluent have been collected. Use each 10-ml portion of ethereal solvent to wash out the sample beaker.Add 75 ml of water to the eluent ether solution and remove the organic solvent by distillation on a steam-bath. Add 5 ml of sulphuric acid and 5 ml of perchloric acid to the aqueous solution and take the mixture to fuming. Complete the estimation by any suitable m e t h ~ d . ~ , ~ RESULTS- monazite minerals. BURSTALL AND WELLS: COLUMNS OF CELLULOSE AND ORGANIC Satisfactory results, as shown in Table XVII, were obtained on both pure and crude TABLE IWII EXTRACTION OF URANIUM FROM MONAZITE SANDS BY THE RECOMMENDED METHOD Sample 1 7 8 10 12 13 14 15 16 17 18 Pure Crude 1 Y Y Y Pure Crude Pure Y Y Y Y 99 Crude U308 found, % 0.37, 0.36 0.50 0.05 0.19 0:35, 0-35 0.09, 0.07 0-38, 0.42, 0.39 0.86, 0.81, 0.85 0.92, 0.96 0.26, 0.28, 0.30 0.35, 0.33, 0.33 Mean U30, by standard chemical methods334 0.37 0.48 0.05 0.18 0.35 0.07 No reliable results obtained by other methods The efficiency of the extraction, as shown in Table XVIII, was tested by addition of uranium to a standard sample of monazite of known uranium content.A determination of the total uranium was made and the “recovery” calculated by subtracting the known uranium content of the sample. TABLE XVIII RECOVERY OF URANIUM ADDED TO MONAZITE SAND Sample No. 1 I L \ Weight of U308 added, Weight of U,O, recovered, mg mg 5-0 4-96 10.0 10.07 25.0 24.94 The agreement between the two methods shown in Table XVII and the recoveries shown in Table XVIII are considered satisfactory. In spite of the special precautions necessary to prevent the movement of zirconium, the cellulose column method is preferred to normal chemical methods because of its ease and speed of operation and the accuracy and reproducibility of the results.July, 19511 SOLVENT EXTRA4CTION FOR SEPARATING URAXIUM 409 The authors wish to thank J.G. Beynon, Miss R. D. Humphreys, Mrs. P. J. Forrest and Miss P. McGlone for assistance in the experimental work. The investigations were carried out on behalf of the Ministry of Supply by whose permission this paper is published. REFERENCES 1. 2. 3. 4. “Assayer’s Guide,” A.E.C.D.-2640. NOTE-References 1 and 2 are to Parts I and I11 of this series; Part I1 is by Burstall, F. H., Davies, G. R., Linstead, R. P., and Wells, R. A., J . Chern. SOC., 1950, 516; Part IV is by Lewis, J. A., and Griffiths J.M., Atialyst, 1951, 76, 388. CHEMICAL RESEARCH LABORATORY TEDDINGTON, MIDDLESEX Arden, T. V., Burstall, F. H., and Linstead, R. P., J . Chem. SOC., 1949, S 311. Burstall, F. H., Davies, G. R., and Wells, R. A., Disc. Farad. Soc., 1949, No. 7, 179. “Handbook of Chemical Methods for the Determination of Uranium in Minerals and Ores,” H.N. Stationery Office, London, 1950. MR. C. G. DAUBNEY enquired about the of evaporation during the percolations, and the amount of water present. MR. BURSTALL replied that a normally DISCUSSION time taken for ether to run through a column, the prevention any precautions that must be taken to avoid an increase in packed column permitted a flow-rate of 100ml of ethereal solvent in 20 to 30 minutes. The eluent was collected directly in a distillation flask and no special pre- cautions were taken to prevent evaporation or to prevent an increase in the water content of the solvent.Addition of more water to the solvent tended to slow up the extraction of uranium slightly, but a t the same time the movement of other materials also was usually retarded. DR. G. E. FOSTER asked whether the presence of peroxide in the ether would affect the results. MR. BURSTALL said that peroxide in the ether should be avoided. A peroxy-vanadium compound readily soluble in the solvent was formed if vanadium was present. This substance moved down the column as a characteristic pink zone immediately following the uranium. MR. W. H. BENNETT asked, first, whether fluorides could be tolerated in the separations as described and secondly, whether the authors would comment on the use of water-repellent agents other than the toxic compound quoted in the paper.MR. WELLS said, in reply, that the presence of fluoride should be avoided because free hydrofluoric acid would remove the silicone lining from the glass tube. Hydrofluoric acid, unless suitably complexed, would also inhibit the extraction of uranium. Another, less toxic, silicone solution was now available; i t was manufactured by the Dow Corning Co. of America, and marketed in this country by Albright & Wilson under the name “Dow Corning Fluid 200.” DR. D. I. COOMBER asked whether the authors had had any experience of the use of derivatives of cellulose, for example, ethyl cellulose, in this or related problems. He thought that, with substituted celluloses, metals other than uranium would probably not be held back so much.If carboxymethyl cellulose were used the column would become an ion-exchange column rather than a chromatographic column. MR. BURSTALL said that ethyl cellulose had not, so far, been examined as an adsorbent. MR. N. STRAFFORD asked whether the authors considered the separation by cellulose to be due to partition or to adsorption chromatography. Had they tried partition chromatography on wet silica gel ? MR. BURSTALL replied that the separations possible on cellulose appeared to be due to a combination of both partition and adsorption, the predominance of either factor depending upon the metal and solvent concerned. Partition chromatography on wet silica gel has been tried for a number of inorganic separations, but without much success.DR. J. H. HAMENCE asked the authors if they would put forward any theories that they might have arrived a t in the course of their work on the mechanism of the separation. In view of the ever-increasing application of chromatography in the solution of hitherto insoluble problems, information on the mechanism of the phenomenon was always very valuable, particularly as a guide when working out conditions for a new separation. In the questioner’s experience, adsorption and solubility appeared to play the major parts in this work. The extent to which any one of these factors contributed to a separation varied considerably and difficulty was frequently experienced -in obtaining precise information as to which factor predominated. Generally, the next most important factor is the extent of partition of the material between solvent and water in the cellulose. Finally, the retention of a number of metallic salts by cellulose appeared to be far too strong to be accounted for by partition alone and for these it was possible that chemi-adsorption occurred. DR. W. STROSS asked whether the authors had found applications of this extremely elegant selective Technique (perhaps with suitable modifications) to the determinations of elements other than uranium. MR. WELLS replied that a number of factors combined to effect a separation. The solubility of the material in the solvent used was an obvious first consideration.410 OSBORN AND JOHNS: THE RAPID DETERMINATION OF SODIUM AND [VOl. 76 MR. BURSTALL said that a similar chromatographic technique had been used in a number of other separations and determinations, some of which had been published and others were to be published in the near future. These studies included the separation and estimation of nickel, cobalt, copper and iron in samples of nickel steel, the separation and determination of gold in the platinum metals, of mercury with the group IIA metals, thorium in minerals and ores, niobium and tantalum in minerals and ores and the separation of zirconium and hafnium. DR. H. LIEBMANN referred to Dr. Stross’s question and mentioned that, following very closely the methods of Burstall and his colleagues and using the polarograph for the final analysis, they had recently determined small quantities of zinc in tin - lead solders. At present they could estimate quantities of about 0.001 per cent. in a 2-g sample, but they believed that the sensitivity of the method was capable of improvement. MR. R. C. CHIRNSIDE said he would be glad to know if the work that the authors had so far carried out enabled them to give a lead as to the probable behaviour of some of the non-metals, particularly boron and arsenic, on these chromatographic columns. MR. BURSTALL replied that, although they had not studied the extraction of boron or arsenic by chromatographic means, separations that had been carried out on strips of filter-paper by the authors and others indicated that i t should be possible to develop an extraction procedure for these materials. It was hoped to publish a preliminary Note on this subject in the near future.