Y
+
1.003 ('t0.031) X 9.9 ( 5 1.9) p.p.m.
where Y is the concentration of oxygen found and X is the concentration of oxygen added to the samples. The value 9.9 f 1.9 p.p.m. indicates the oxygen content of the unspiked samples-Le., the oxygen content of the cesium metal itself.
Since much of the interference from the oxygen content of the containers and the surrounding atmosphere has been eliminated, one may now also take advantage of the higher fast-neutron fluxes available from 2 or 5 curies per sq. inch tritium targets to obtain better counting statistics for samples containing only a few parts per million of oxygen.
CONCLUSION
ACKNOWLEDGMENT
With the new vials and the method described it is now possible to determine trace amounts of oxygen even in very reactive and hard to handle samples.
The authors are indebted to R. Moolenaar for providing and handling of the cesium metal as well as preparing the spiked samples.
=
LITERATURE CITED
(1) Anders, 0. U., Briden, D. W., ANAL. CHEM.36, 287 (1964). (2) Gray, P. R., Phillips Petroleum Co., Bartlesville, Okla., Sixth Eastern Analytical Symposium, New York, Paper 4, NOV.11, 1964. (3) Stallwood, R. A., Mott, W. E., Fanale, D. T., ANAL.CHEM.35, 7 (1963). (4) Steele. E. L.. Meinke. W. W.. Zbid.. 34, 185 '( 1962).' ( 5 ) Veal, D. J., Cook, C. F., Zbid., 34, 178 (1962). RECEIVEDfor review May 25, 1964. Accepted January 21, 1965. Work supported in part by the Air Force Materials Laboratory, Project No. 7360, Task No. 736005, Wright Patterson Air Force Base, Ohio.
Trace Impurity Analysis of Thorium-Uranium and PI uto nium-T ho rium-U r a nium Alloys by Anion Exc ha nge-Partitio n Chromatog ra p hy EDMUND A. HUFF Argonne National laboratory, Argonne, 111.
b An analytical method was developed for the analysis of trace impurities in high purity thoriurnuranium and plutonium-thorium-uranium alloys. The procedure involved the separation of metal contaminants from the alloy matrix by column elution in an 8N nitric acid medium. The two-phase chromatographic column incorporated a strong base anion exchange resin for the absorption of plutonium(1V) and thorium(lV), and a TBP coated trifluorochloroethylene polymer for the retention of uranium (VI). A one-pass elution effectively removed the matrix constituents, while most impurities of interest remained in the effluent, and were determined spectrographically, using the copper spark method. High concentrational sensitivity was attained without the inherent standardization and toxicity problems associated with the carrier distillation technique.
-
I
SUPPORT of metallurgy research activities a t the Argonne National Laboratory, an analytical method was needed for the determination of trace metal impurities in plutonium-thorium, thorium-uranium, and plutoniumthorium-uranium alloys. Emission spectrographic analysis appeared to be the most promising technique for the determination of the large number of impurities sought a t the expected low concentrations. The widely used
N
carrier distillation procedure (12) for trace impurities in refractory materials was not attractive because, as Kofoed (9) has shown, samples and standards should be of nearly identical composition for most reliable results. The wide variation in matrix ratios among alloys actually received for analysis would have presented a formidable standardization task. Preconcentration techniques prior to spectrographic analysis have often been used as a means of producing more immediate results. The anion exchange separation of impurities from plutonium (2, I I ) , thorium (S), and uranium ( 2 ) gave excellent concentrational sensitivity and good reproducibility. However, the data by Kraus and Nelson (IO) and Faris and Buchanan (4) showed that a practical one-pass anion exchange elution system could not be conceived for alloys containing all three elements as matrix components. Recently, Hamlin, et al. (6, 7) reported on a successful separation of uranium from a number of matrices by a reversed-phase partition chromatographic technique. I n their work, a tri-n-butyl phosphate (TBP)-Kel-F column was used to adsorb uranium (VI), while the impurities were being eluted with 5.5N nitric acid. Even at an 8N liitric acid concentration, the column showed satisfactory adsorption for uranium(V1). Thus, a combined anion exchangereversed-phase partition chromato-
graphic separation in a nitric acid medium appeared applicable to the analysis of binary thorium-uranium and ternary plutonium-thoriumuranium alloys. A two-phase chromatographic column, containing both strong base anion exchange resin and TBPcoated Kel-F polymer, effectively retained the matrix components and allowed the trace impurities to be eluted with 8'47 nitric acid. High concentrational sensitivity could be attained by subsequent spectrographic analysis of the effluent solutions. The absence of inherent standardization and toxicity problems rendered the procedure preferable to the carrier distillation technique. EXPERIMENTAL
Materials. Dowex 1 X 8, 100to 20O-meshJ strong base anion exchange resin was supplied by Bio-Rad Laboratories, Berkeley, Calif., in the nitrate form. Alternate washings with purified nitric acid and distilled, deionized water ensured the removal of trace metal impurities. T h e resin was stored as a water slurry in polyethylene bottles. Commercial grade T B P was purified by washing with a 5% sodium carbonate solution, distilled water, and 8N nitric acid. A trifluorochloroethylene polymer, Kel-F, grade 300 (Minnesota Mining and Manufacturing Co.) or Haloport-K ( F & 11 Scientific Corp.), was mesh graded and used without further purification. Kel-F and HaloVOL. 37, NO. 4, APRIL 1965
* 533
port-K were equivalent in performance, but production of these has been discontinued. Plaskon, type C T F E 2300 (Allied Chemical Corp.), performed satisfactorily and will be used in future work. This trifluorochloroethylene polymer showed good absorption for T B P (1 gram of T B P per gram of resin), was inert toward all mineral acids, and a specific gravity of greater than two facilitated column preparation of uniform bed density. Polyethylene chromatographic columns and inert plastic laboratory ware were used exclusively throughout the investigation. Their use ensured minimum blank contamination by common impurities and permitted elution with hydrofluoric acid. Column Preparation. A schematic drawing of a chromatographic column as used in this investigation is shown in Figure 1. The lower portion contained a 2-cm. high bed of settled Dowex 1 X 8 anion exchange resin. T h e next phase was prepared by stirring 2 ml. of purified T B P with a 2-gram portion of Kel-F polymer until a uniform mixture was obtained. This preparation was then transferred quantitatively with water to the column, agitated to remove air pockets, and allowed to settle by gravity. The slurry was compacted by repeated drainage of excess liquid, tapping, and finally by a slight external pressure. Care was taken not to let the aqueous phase drop below the solid support at any time, since otherwise excessive channeling was observed, rendering the preparation useless for an efficient column separation. Finally, a 10-cm. high bed of anion exchange resin was added to complete the chromatographic column. Polypropylene wool, obtained from the American Felt Go., was used to separate the individual phases. The two resin portions bracketing the Kel-F-TBP preparation were
;1_ 1.5 cm-
lOWEX I X 6 RESIN
2 0 cm
-POLYPROPYLENE WOOL U
Figure 1. Schematic drawing of a chromatographic column
quite effective in keeping the latter from drying out during intermittent periods of flow stoppage. This arrangement effectively eliminated the use of elaborate flow controls and rendered the setup very satisfactory for routine glove-box operations. Column Separation. T h e chromatographic column was conditioned with 8 N nitric acid before use. A 0.20- t o 0.50-gram alloy sample, containing generally less than 0.10
gram of uranium, was dissolved in 10 ml. of 8iL’ nitric-0.005N hydrofluoric acid mixture in a Teflon ( D u Pont) beaker. Heating on a hot plate was necessary to effect complete dissolution. After cooling, the sample solution was adsorbed on the column as the nitrate anion complex, and the unadsorbed cation impurities were eluted with 8.V nitric acid a t an approximate flow rate of 1 ml. per minute. Recovery experiments indicated that a 50- to 80-ml. fraction was usually sufficient to quantitatively separate all impurities of interest. Visual observation of the green plutonium(1V) band on the anion resin and the yellow uranium(V1) band on the Kel-F-TBP preparation generally provided a good measure for the column separation efficiency. After each impurity elution, the matrix was stripped from the column with 0.1S nitric acid for binary thorium-uranium samples, and a mixture of 0.4s nitric0.011V hydrofluoric acids for ternary plutonium-t horium-uranium alloys, A 100-ml. fraction was usually sufficient for complete elution. The chromatographic columns could be used repeatedly for as many as 10 separations before a broadening of the uranium band indicated resin exhaustion. The column effluents were evaporated to dryness under infrared heat lamps to remove free nitric acid. The impurity residue was redissolved in 1 ml. or more of 62; hydrochloric acid or a mixture of hydrochloric-hydrofluoric acids. Cobalt was added a t a concentration of 20 pg. per ml. to serve as an internal standard in the subsequent spectrographic analysis. Analysis. T h e impurity concentrations were determined by copper spark excitation ( 5 ) and visual evaluation (2, 3 ) . The estimated accuracy per determination was judged to be within a factor of two of the amount present, which was considered to be
Table 1.
Recovery of Added Impurities from Thorium-Uranium Alloys (Results in p.p.m.) Sample A (80y0Th-207, V) Sample B (90% Th-lO%
Initial impurities
Impurities added
RP
15