THE UNIVERSITY OF MICHIGAN INDUSTRY PROGRAM OF THE COLLEGE OF ENGINEERING RADIOCHEMICAL SEPARATIONS: II. SEPARATION OF RADIOACTIVE SILVER BY ISOTOPIC EXCHANGE Duane N. Sunderman W. Waynel Meinke November, 1955 IP-141

ACKNOWLEDGEMENT We would like to express our appreciation to the authors for permission to give this prepaper limited distribution under the Industry Program of the College of Engineering.

RADIOCHEMICAL SEPARATIONS: II. SEPARATION OF RADIOACTIVE SILVER BY ISOTOPIC EXCHANGE Duane N. Sunderman and W. Wayne Meinke Department of Chemistry, University of Michigan Ann Arbor, Michigan ABSTRACT The principle of isotopic exchange has been adapted to the separation and determination of radioactive silver from a mixture of radioactive species, and a quantitative procedure developed. The separation of radioactive silver by isotopic exchange with a silver chloride surface was found to be quantitative in fifteen minutes or less, even in the presence of high concentrations of many inorganic salts and organic solvents. The method is quite specific for silver, with decontamination factors between 103 and 107 for eighteen representative radioactive species. The method was also applied to the separation of silver from fission products and bombardment samples and found satisfactory in both cases. 1

INTRODUCTION The exchange of silver ion between a precipitate of silver chloride and a solution of silver nitrate reaches isotopic equilibrium very rapidly. Because of the low solubility of silver chloride, a very favorable ratio exists at equilibrium between the silver atoms in the precipitate and the silver atoms in the solution. For this reason if silver chloride is added to a solution containing only trace amounts of radioactive silver, a high percentage of this silver will have exchanged with the silver in the precipitate by the time equilibrium is attained. Use has been made of this fact to develop a rapid, high decontamination, single-step method for the separation of traces of radioactive silver from a solution containing other radioactive species. By use of the technique of isotopic exchange (6, 7) it is possible to shorten existing procedures for the separation of silver without sacrificing specificity or yield. Isotopic exchange involving silver halides and silver salts in solution has been studied by Langer (2, 3) and in one case was applied to the determination of macro amounts of silver (4). This method has also been applied to the preparation of beta-ray sources (5), and the separation of 1-131 (9). A method for the separation of iodine on a base material of AgCl recently reported in the literature (1) is based upon elemental substitution of chloride for iodide by the silver chloride rather than isotopic exchange. 2

APPARATUS AND REAGENTS Apparatus International Clinical Centrifuge, Scintillation well counter, Nuclear Instrument and Chemical Corporation, Model DS-3 with2in. additional lead shield. Scaler, Nuclear Instrument and Chemical Corporation, Model 162, scale of 128, modified to count with preset time and preset count and to reset automatically. Battery Eliminator, Heathkit Model BE-4, Heath Company, Benton Harbor, Michigan. Magnetic Stirrer, manufactured b- G. F. Smith Chemical Company, Columbus, Ohio. Platinum Electrodes, 10 mm diameter, 30 mm high, 52 mesh, manufactured by American Platinum Works, Newark, New Jersey. Reagents Sodium Cyanide, (Mallinckrodt Reagent No. 7616), 3 M. Sodium Hydroxide, (Mallinckrodt reagent No. 7680), 5 M. Hydrochloric Acid, concentrated, (B and A reagent No. 1090). Nitric Acid, concentrated, (B and A reagent No. 1120). Sulfuric Acid, concentrated, (B and A reagent No. 1180). Hydrobromic Acid, 48%, (B and A reagent No. 1085). Hydriodic Acid, sp gr 1.5, (Merck reagent No. 0246). Hydrofluoric Acid, 48%, (B and A reagent No. 1100). Acetic Acid, glacial, (B and A reagent No. 1019). Aluminum nitrate, (Mallinckrodt reagent No. 3172). 3

Bismuth nitrate, (B and A reagent No. 1456). Cupric nitrate, (B and A reagent No. 1641). Ferric nitrate, (B and A reagent No. 1739). Sodium acetate, (B and A reagent No. 2191). Zinc nitrate, (B and A reagent No. 2445). Radioactive Tracers: Obtained from the Isotope Control Division of Oak Ridge National Laboratory and used without added carrier. Nuclear and chemical characteristics of these tracers are tabulated in the first paper of this series (8). 4

EXPERIMENTAL PROCEDURE A silver chloride electrode of large surface area is made from an electrode of fifty-two mesh platinum gauze, ten mm in diameter and thirty mm high. Ten mg of silver metal is deposited on the gauze by electrolysis for ten minutes at four volts from ten ml of a solution three M in NaCN and 0.5 M in NaOH. A simple battery eliminator was used as a source of voltage for these experiments. The whitish surface of silver metal is rinsed with water, inserted in ten ml of 0.05 M HC1 and electrolyzed as the anode for five minutes at one volt, completely converting the silver metal to silver chloride. The electrode is washed with eight M HN03 and is ready for use. To separate silver by isotopic exchange, the silver chloride electrode is contacted with ten ml of solution containing a very low concentration of radioactive silver. The solution is stirred magnetically. The gauze is then removed from the solution and washed with eight M HN03 for one minute. It is rinsed with acetone, placed in a lusteroid cone and counted in the scintillation well counter. If quantitative counting and recovery of the silver is desired, the silver chloride may be dissolved in ammonium hydroxide, acidified, and the silver chloride precipitate separated by centrifugation or filtration and mounted for counting. The standard method referred to hereafter as the "quantitative procedure" for separating trace silver requires the use of ten mg of silver as silver chloride on the electrode and fifteen-minute contact with the tracer solution one M in HN03 at room temperature. The silver chloride is then dissolved, transferred to a culture tube, reprecipitated, centrifuged, the supernate removed and the precipitate counted in the scintillation well counter. 5

RESULTS AND DISCUSSION Yields of Silver Recovery of silver tracer in the quantitative procedure using ten mg, four mg and one mg of silver as AgCl on the gauze are given in Table I. For each determination 20,000-40,000 d/m of Ag-ll0 tracer was used. All samples were counted to a probable error of 0.4%. Errors quoted in the table are standard deviations. Table I Yield of Radioactive Silver by Exchange with Silver Chloride Dete- Weight of Silver on Gauze Deter-, matin 10 mg 4 mg 1 mg mination ( 1 99.5 99.0 93.0 2 98.5 99.1 94.3 3 99.8 98.8 94.7 4 99.8 99.0 94.9 5 99.8 98.8 95.8 6 99.2 99.0 7 100.0 98.7 Std. De- 99.5+0.5 98.9+0.2 94.5+1.0 viation The silver chloride surface prepared in the manner described above is not ideal from the point of view of mechanical stability. During the course of the procedure a rather small amount of silver chloride is lost, due apparently to chipping or chalking of the electrode surface in the contacting and washing periods. Experiments indicate that this 6

chalking loss is of the order of 0.05 mg of silver chloride. Thus if only one mg of the chloride precipitate is used originally, this loss is an appreciable portion of the total while it is negligible in the ten mg samples used routinely. As noted in Table I, however, reproducible results can be obtained with four mg and even one mg amounts of silver. Thus it is possible to use such small amounts if, because of self-absorption or other considerations, it is necessary to obtain a high specific activity. Fifteen minutes contact time under ideal conditions is more than sufficient for complete removal of the radioactive silver from solution, as shown in a later section. Actually at 25~C the 97% removal point occurs at five minutes and over 99% is obtained in six to seven minutes. This range of five to seven minutes, however, is a very sensitive one as far as interfering ions and nonideal conditions are concerned. Thus the fifteen minute time was chosen to minimize the effect of small changes in acidity, stirring, and the character of the silver chloride deposit. The one-minute wash of the silver chloride gauze after exchange is necessary where decontamination from other activities by a factor of 106 is desired. Washing for a longer period of time had no effect. Eight M HN03 proved effective as the wash solution for all the contaminating activities used in this work. If other conditions warrant their use, even organic wash solutions could be applied. It was found essential to dissolve the silver chloride surface and reprecipitate it for the quantitative recovery of radioactive silver and its reduction to a reproducible form for counting. This was necessary because it was impossible to reproduce counts of the gauze to better than 2-5%. 7

When very small amounts of inactive silver are desired on the gauze, the lower limit appears to be dependent upon the manipulatory techniques in making the silver plate on the gauze. This means first that the more uniform the deposit, the more area is made available for exchange for a given small amount of silver. Secondly, silver in sub-milligram amounts is perhaps dissolved or in any case lost from the electrode in the process of removing it from the cyanide solution. This is shown by the graph in Fig. 1 in which the time of electrolysis is plotted against the fraction of metallic silver deposited on the platinum gauze from cyanide solution using the standard four-volt cell and varying the amount of silver in solution. Silver-llO was used as the tracer to follow the deposition in these experiments. Silver iodide surfaces are also quite suitable for the exchange separation of tracer silver. The rate of exchange is not significantly different than that with silver chloride in that 97% is removed in six minutes and 99% in nine minutes with ten mg silver as AgI. In some cases, the iodide may be considerably better for exchange purposes due to its lower solubility, but further separations are often complicated by the presence of the active silver in the form of the difficulty soluble iodide. One method used in this laboratory to by-pass these difficulties was to electrolize the "AgI gauze" containing the tracer silver as cathode in a neutral conducting solution such as sodium acetate for one minute at a voltage sufficient to decompose the silver iodide. The iodide ion goes into solution and free silver remains as a black deposit on the electrode. This free silver may then be dissolved in nitric acid. 8

100 2.5,5,10,12.5 MG _ 80 C/) - E Xol0. MG_ IL60- - - r, /U_40- n 4 8 12 I o20 /'" — IME OF DE POS 2 1ON(MIN) TIME OF DEPOSITION (MIN) Fig. 1. Influence of Total Amount of Silver Present on the Yield of Silver Deposited Electrolytically from a Cyanide Solution. 9

Contamination The quantitative procedure was checked with eighteen different representative tracers to determine the levels of contamination which might be expected. Long-lived daughter activities of the tracers were also considered since no carriers were present and no attempt was made to separate parents from daughters. In these experiments the only radioactivity present was that of the element being studied. The contamination of the AgCl gauze by each element is given in Table II. All results are an average of at least two determinations. Table II Contamination of Silver by Other Activities in Isotopic Exchange Method Percentage Carried on Gauze Ba-140, La-140 0.00005 Ce-144, Pr-144 0.0001 Cs-134 0.0001 Co-60 0.001-0.0001 Cr-51 0.01 I-131 0.01-0.001 Ir-192 0.1-0.01 Ru-106, Rh-106 0.1 Sb-124 0.01 Se-75 0.1 Sn-113, In-113 0.1-0.01 Sr-90, Y-90 0.0001 Ta-182 0.01 Zr-95, Nb-95 0.01-0.001 10

To obtain more information concerning the mechanism of contamination of the AgCl gauze by foreign activities it is advantageous to arrange the data in the form of Table III illustrating the similarity of those activities giving the largest contamination. Table III Contamination Summary Percentage Carried Elements Group A 0.1-0.001 Cr, I, Ir, Ru-Rh, Sb, Se, Sn-In, Ta, Zr-Nb Group B 0.00.-O.00001 Ba-La, Ce-Pr, Co, Cs, Sr-Y The elements in Group A which show some tendency to contaminate are those elements which in general, form very strong halide complexes with chloride and fluoride. (The iodide case is different in that about 10% of the tracer iodide is carried on a silver chloride gauze after a fifteen-minute contact. The 0.01% figure of Table II for iodine is obtained after treatment of the I-151 solution with HN03 and KMn04 to oxidize the iodide to iodate. The iodate is more soluble than the iodide and the chloride by factors of 104 and 102 respectively so it would not be expected to replace either of them on the gauze surface.) The contamination of the gauze by the complex formers may be by adsorption of chloride complexes of the individual elements on the silver chloride surface. Perhaps these complexes actually replace the silver chloride with insoluble compounds of the type Ag2IrCl6, similar to the slightly soluble Cs2PtCl6 or even Ag4Fe(CN)6. 11

If this explanation were correct, a less soluble surface such as AgBr should reduce contamination and a AgI surface should reduce it even further. The degree of Ru-Rh contamination of AgBr and AgI surfaces was studied to determine whether this were the case. These results along with the corresponding AgCl data are shown in Table IV and indicate that the availability of silver ions or halide ions or both does indeed influence the contamination by this particular activity. Table IV Contamination of Silver Halide Surfaces by Ru-Rh Tracer Surface Percentage of Ru-Rh Material Activity Carried AgCl 0.1 AgBr 0.04 AgI 0.003 In the cases of zirconium-niobium and tantalum activities, the contamination is probably due to hydrolytic effects such as radio-colloid formation or adsorption of basic salts. Three more general determinations of decontamination were made with this procedure. First, the silver chloride gauze was contacted with a solution containing four-year-old fission products in unseparated form. After a fifteen-minute contact, 0.001% of the fission product activity was found on the gauze. In the second determination, cadmium metal was bombarded with deuterons in the University of Michigan cyclotron, the target was dissolved in nitric and hydrofluoric acids and the silver produced by the (d, a) reaction removed by isotopic exchange. Due to the rapidity 12

of the separation it was possible to detect a two-minute silver isotope which was never seen when precipitation separations were used. The decay was followed in the scintillation well counter as well as a four-pi beta proportional counter and the decay curves showed as clean a separation from cadmium and indium activities as had a series of precipitation separations requiring fourty-five minutes. For the third determination a Pd109 irradiated unit, containing about 300 me of thirteen hr Pd activity and 2-3 me of Ag-lll with a 7.5 day half-life, was obtained from Oak Ridge National Laboratory. The palladium was dissolved in aqua regia, diluted to a four M HN03 solution and the silver removed by isotopic exchange. The AgCl gauze obtained weighed ten mg and contained two me of Ag-lll which gave a reading of three "r/hr" at six inches when monitored. Its decay was followed in the scintillation well counter and the GM end-window counter. The semilog plot of the decay was linear over more than five half-lives and no other activity was detected. Interferences A study was made of the effects of inactive species present in varying concentrations and forms in the solution from which tracer silver is to be removed. About 40,000 d/m of Ag-llO were added to solutions containing the given concentration of interfering specie, the quantitative procedure used to remove the silver activity, and the yield determined. The species chosen were those which were considered to be of most interest, with the greatest likelihood of appearance in process solutions. The values for the silver tracer yield are given in Table V. This study of the effect of inactive species in high concentrations revealed few specific interferences. The only substance which 13

Table V Effect of Interfering Substances upon Exchange of Silver with Silver Chloride'...................Percentage Yield Specie Concentration Percentage Yield of Ag-llO HNOs 1 M 99 4- 99 8 93 12 91 16 85 HC1 0.05 M 98 0.5 99 1.0 98 3 89 6 41 H2S04 1.5 M 97 3 90 6 79 HF 1.1 M 98 3.3 98 6.6 83 13.2 57 22 18 HC2H302 99% 42 Al (N3)3 2 M 86 1 95 0.5 99 Cu(NO3)2 2 M 95 1 97 0.5 98 Zn(NO3)2 2 M 95 1 97 Fe(N03)3 2 M 94 1 93 0.5 95 Bi(N03)3 2 M 93 1 97 0.5 100 14

Table V (cont.) Percentage Yield Specie Concentration Pe Age el of Ag-110 NaC2H302 2 M 95 1 97 0.5 98 C2H50H 99% 36 50 95 25 98 (CH3)2CO 99% 70 50 95 25 98 showed adverse effect on the yield of tracer silver was Fe(N03)3. Yield values remained between 93 and 95% for this compound. For other salts and acids included in the study, the effect seems to be one of salt concentration rather than a specific effect of any individual ion or salt. In Table VI the anionic molarity of each salt is listed opposite the concentration at which that salt first resulted in less than 97% yield of the tracer silver. In this work the AgCl gauzes were counted with an accompanying error of 2-3% as explained earlier. Table VI Interference Summary Anionic Molarity at __Specie Critical Concentration HN03 4-8 M HC1 1-3 H2S04 1.5-3 HF 3-6 Al(NO3)3 3 Bi(N03)3 3-6 Cu(N03)2 4 Zn(N03)2 4 NaC2H302 2 15

It would appear from these data that a concentration of anion of greater than 2-8 M, depending upon whether it is nitrate, chloride, sulfate, fluoride or acetate, will result in losses of silver under the procedures used. A look at the mechanism of the separation may show the reason for this concentration effect. Silver chloride, as freshly precipitated, is well known for its ability to adsorb either positively or negatively charged ions, depending upon which is in excess. Use is made of this fact by the Fajans method for determination of end point in the titration of chloride with silver nitrate, by use of dichlorofluorescein. To effect the isotopic exchange which is required for separation, high ionic mobility is required in the immediate vicinity of the silver chloride surface. A stable ionic layer adjacent to the exchanging surface might lower this mobility and cause loss of yield. Cognizance should be taken of this possibility before adoption of this method for highly salted solutions. The effect of the two aqueous-organic systems studied indicates that minor concentrations of either acetone or ethyl alcohol will not prevent quantitative removal of silver tracer since they apparently have no specific deactivating effect on the AgCl surface. "AgCl gauzes" did, however, show lower activity toward exchange if they had been rinsed with acetone and dried prior to exchange. Dependence of Yield on Time and Temperature If for any reason fifteen minutes is too long a contacting time, shorter periods can be used. Experimental determinations of the extent of exchange for a number of temperatures are plotted against time in Fig. 2. As would be expected, the exchange is more rapid at the higher temperatures. 16

100 IL-) 0 \ \ \ 20C 960C \530C 26~C 740C 380C I — L -l-l\l l l l l l l I 0 2 4 6 8 10 TIME OF CONTACT (MIN.) Fig. 2. Temperature Dependence of Rate of Exchange of Silver Ion with Silver Chloride. 17

It can be seen from this temperature dependence that considerable time economies can be realized by raising the temperature of the contacting solutions. At 96~C, for example, over 95% of the tracer can be separated with only two-minute contact, the 50% point being reached in twenty seconds. Use was made of this fact to separate silver with sufficient rapidity to measure the rate of decay of a two-minute isotope formed by the (d, a) reaction on cadmium. Aging of Silver Chloride Most of the work described above was carried out with AgCl gauzes prepared within one hour of use. Because of the light sensitivity of AgCl and the fact that isotopic exchange must necessarily occur on the surface of the electrode, time studies were made of the stability of the AgCl surface. Gauzes containing ten mg of silver as AgCl were prepared, stored in eight M HN03 for varying periods of time and then used for a silver separation. The yields of the exchange determined under quantitative conditions are shown in Table VII. The period of study extended over one week. Table VII Effect of Aging of the AgCl upon the Efficiency of the Silver Exchange Time of Aging Yield in in 8 M HNO3 15 Minute Contact 2 hrs 99.0 22 hrs 99.8 72 hrs 99.5 7 days 98.7 18

No attampt was made to keep the gauzes from direct light. However, the only light incident upon the gauzes during this period was that from two 100-watt fluorescent tubes about four feet above them. It appears that the aging of AgCl gauzes in eight M HN03 for periods up to seven days has no effect on their activity with respect to the exchange separation of silver. 19

ACKNOWLEDGEMENTS The authors wish to express their appreciation to Mr. Thomas Olcott for his help in determining the temperature dependence of the rate of exchange. This work was partially supported by the Atomic Energy Commission. 20

BIBLIOGRAPHY 1. D. G. Arnott, and J. Wells-Cole, Nature 171, 269 (1953). 2. A. Langer, J. Chem. Phys. 10, 321 (1942). 3. A. Langer, J. Chem. Phys. 11, 11 (1943). 4. A Langer, Anal. Chem. 22, 1288 (1950). 5. W. W. Meinke, and D. N. Sunderman, "Silver-ill Beta-Ray Sources", Nucleonics 13, November or December (in press). 6. D. N. Sunderman, and W. W. Meinke, Atomic Energy Commission Unclassified Document AECU-2988 (1955). 7. D. N. Sunderman, and W. W. Meinke, Science 121, 777 (1955). 8. D. N. Sunderman, and W. W. Meinke, "Radiochemical Separations: I. Barium, Strontium, and Calcium", Jour. Inorg. and Nucl. Chem. (in press). 9. D. N. Sunderman, and W. W. Meinke, "Utilization of Isotopic Exchange for Radiochemical Separations", to be presented at the Nuclear Engineering and Science Conference in Cleveland, Ohio, December, 1955. 21