y 'b^/)-t^r'^JA]pnc-'^/(^ MDDC - 910 UNITED STATES ATOMIC ENERGY COMMISSION A COMPARISON OF ANALYTICAL METHODS FOR THE DETERMINATION OF URANIUM by D. A. Maclnnes L. G. Longs worth Rockefeller Institute for Medical Research This document consists of 5 pages. Date of Manuscript; November 24, 1942 Date Declassified: March 4, 1947 This document is for official use. Its issuance does not constitute authority for declassification of classified copies of the same or similar content and title and by the same author (s). Technical Information Division, Oak Ridge Directed Operations Oak Ridge, Tennessee Digitized by tlie Internet Archive in 2011 with funding from University of Florida, George A. Smathers Libraries with support from LYRASIS and the Sloan Foundation http://www.archive.org/details/comparisonofanalOOrock A COMPARISON OF ANALYTICAL METHODS FOR THE DETERMINATION OF URANIUM By D. A. Maclnnes and L. G. Longsworth In the analysis for uranium, conversion to the oxide U30g is frequently used. The conditions under which this oxide exists are, however, still in some doubt. We have attempted to develop a precise in- dependent analytical method by which the composition of the UjO, oxide might be checked after varying ignition procedures. The most promising independent method is the standard oxidimetric titration in which a uranyl solution is reduced somewhat below the quadrivalent stage in the Jones reductor and then reoxidized with a standard oxidyzing agent, the titer between the two end-points, U'''^-U''"* and U'*''* -U"*"' representing the uranium present. Successful results have been reported' for the adaptation of the method of differential potentiometric titration to this system. Unfortunately, as will be discussed later in this report, the technique fails when the dilute solutions and small titrating increments required for accurate results are used. For our comparisons, we were obliged to use the standard volumetric procedure,^ in which the over-reduced solution is oxidized to the first end-point by air, before the subsequent titration with potassium permanganate or eerie sulfate. Samples of a preparation of hydrated UO3 were analyzed in this way; other portions were ignited in air to UjO, for prolonged periods at several temperatures and finally reduced with hydrogen to UOj. EXPERIMENTAL Ignition to UjO^ Four 0.4 g samples of UOj-xHjO (prepared from UQ,) were weighed into small platinum crucibles (0.9 g) on the microbalance, and transferred to an electrically heated furnace where the temperature was gradually raised to 700° and held at that point for 7 hours. After removal from the furnace, they were cooled in a desiccator for several hours and reweighed. They were then reheated three times at the same temperature. Subsequent series of heatings and weighings were made at 820° and 930° and finally again at 700°. The average of the weights of the four portions of resulting lower oxide calculated as the per cent of the initial weight of UOj-xH^O, and the average deviation of the four from the mean are recorded with the corresponding temperature and length of heating in the first 11 lines of Table 1. Following these ignitions, the samples were stored in the atmosphere above a saturated solution of Mg(N03)2-6HjO (which has a vapor pressure of water corresponding to an average humidity of about 50%) to determine whether the uptake of water was sufficiently rapid to have taken place to a weighable extent in the few minutes that the samples were exposed to the laboratory air during weighing. The absorption of water under these conditions was evidently slight as is shown by a comparison of the weights in lines 11, 12 and 13 of Table 1. If the weights recorded in Table 1 are plotted against the corresponding cumulative hours of heating and the seemingly anomalous value of line 6 is ignored, it is seen that the oxide tends to approach a con- stant weight at a given temperature. Furthermore, the loss or uptake of oxygen is apparently reversible, as is shown by the asymptotic approach to constant weight at 700° from both sides. The true constant weight is presumably somewhere between the value of line 4 and of line 11. In the results reported, the value of line 4 is assumed to be correct and cannot differ by more than 0.013% from the true value. For the purposes of comparison, this weight is arbitrarily assumed to represent stoichiometric U^O, , and the per cent UO3 in the original sample of UOs'xHjO calculated, using this assumption, is: MDDC - 910 [1 2] MDDC - 910 93.314 X 11^ = 95.09% UO3 Reduction to UO2 The four crucibles, containing their portions of oxide, were transferred to a quartz tube heated by a combustion furnace. Dry, oxygen-free hydrogen was passed in and the temperature raised to 900° and held at that point for 1 hour. The flow of hydrogen was maintained during the subsequent cooling. When room temperature was reached, the samples were removed individually and weighed, the final weight on the microbalance being taken 2 to 3 minutes after a sample was first exposed to air. The weights before and after reduction are given in lines 13 and 14 of Table 1. Table 1. Ignition of UOg-xHjO to V^Og and the reduction to UOj. Line Treatment Oxide Hours Total Percentage of Avg deviation treat- hours original wt (4 samples) ment 1 Ignition at "U3O3" 7 7 93.345 .003 2 7 Go- '* 21 28 93.329 .005 3 to (( 46 74 93.314 .004 4 49 123 93.314 .002 5 Ignition at "U3O," 18 18 93.301 .003 6 820° (( 59 77 93.248 (?) .005 7 to if 48 125 93.267 .002 8 Ignition at "U3O3" 19 19 93.245 .003 9 930° re 44 63 93.238 .003 10 Ignition at "U3O8" 52 52 93.284 .005 11 700° t( 92 144 93.288 .004 12 Atm of Mg(N03)j-6HjO "U30," •xHjO 5 5 93.293 .005 13 te 15 20 93.294 .005 14 Reduction to UOj 1 1 89.798 .006 15 Atm of Mg(N03),.6H20 UOjxHjO 16 16 89.814 .002 As with the U3OJ, the samples of UO2 were stored in the Mg(NCl^).6I^O atmosphere to determine the rate of water absorption. Comparison of lines 14 and 15 shows that the uptake of water was slight. The reduction of uranium oxides to UOj has been criticized as aa analytical method.^ However, Blitz and Muller' found that UO2 was produced by the reduction of U3O8 with hydrogen very nearly in the theoretical ratio. In the reduction described, we have followed their procedure. If line 14 is assumed to represent pure UOj the per cent UO3 in the original sample of UOj-xHsO is: 89.798 X 286.07 270.07 = 95.12% UO. Titration with eerie sulfate Ceric sulfate solution was standardized against well-dried Bureau of Standards sodium oxalate, by adding an excess of the oxidizing agent from a weight buret to a sulfuric acid solution of the oxalate, heating, cooling, and titrating the excess ceric sulfate with O.OIN ferrous ammonium sulfate to the MDDC - 310 [3 ferrous-o-phenanthroline end-point. Four determinations gave a weight normality (equivalents per 1000 g solution in air) of 0.10506 with an average deviation of 0.013% 0.5 g samples of the same UOj-xHjO taken for the ignitions were dissolved in 50 ml dilute H^SO^ (5 : 95) and reduced in the Jones reductor. Air was bubbled through the resulting olive-green solution until the color changed to clear blue-green. Five drops of 0.005N ferrous-o-phenanthroline indicator were added, followed by the eerie sulfate in slight excess from a weight buret. The excess was titrated to the pink end- point with O.OIN ferrous ammonium sulfate. Four blank determinations (indicator correction plus reagent correction) were made, averaging 0.022 g of eerie sulfate, or, about 0.07% of the net titer. The average of uve uranium titers was 63.328 g eerie sulfate per gram UOj-xHjO with an average deviation of 0.017%. The per cent UO, in the UOjXHjO is then: 63.328 X 0.10506 x 0.14304 x 100 = 95.17% UO3 A comparison of the final results of the ignitions, reduction, and titration are recorded in Table 2. Table 2. Per cent UO, in UOj-xHjO by three independent methods. Ignition U3O,, const, wt 700° 95.09% Reduction to UOj 95.12% Titration U"^" -U'^^ 95.17% It will be seen, therefore, that the oxide assigned to formula 1130^ has that composition within the experimental error of the analytical methods when brought to constant weight at 700°.* This is shown by an independent titration method and by direct reduction of the oxide to V<\ . Application of the titration to small amounts of uranium If the titration procedure described above is adapted to 5 mg of UO3 the results generally agree closely but fail to represent the exact quantity of uranium. In one experiment, portions of a standard uranyl sulfate solution, each equivalent to 5 mg UO3, were dispensed from a weight buret into a micro Jones reductor, and the resulting 20 ml of over-reduced so- lution aerated briefly to the usual color change. Standard .OIN permanganate was added in slight excess from a weight buret, and the excess determined by differential potentiometric titration with .OOIN ferrous sulfate. Blank corrections were made. Four determinations yielded the following values for the % UO3 in the original U03-xHjO: 93.93, 93.85, 93.93, and 93.41(?)%. The value obtained by ignition to U3O, was 95.1%. Locating the U"*^^ -. U"*"* end-point by differential titration A number of investigators have demonstrated by direct potentiometric titration the sudden increase in potential occurring when U*' is completely oxidized to U''"*. In attempting to determine the precise location of this end-point by the method of differential titration, we have come upon unexpected diffi- culties. If a O.IN solution of oxidizing reagent is used with a considerable quantity of over-reduced uranium the end-point may be found. However, when the titrating increments are decreased by using a more dilute reagent in order to obtain greater accuracy the drifts in potential characteristic of the U+3.u+< system become more and more dominant until they completely obscure the end-point. If an over-reduced solution is taken in the usual apparatus for differential titration,* using platinum electrodes and an inert gas, the following typical behavior is noted: *The usual method of igniting at the higher temperature of a Meeker burner, but for the much shorter .period of 5 minutes, gave 95.06% UO3 as the composition of the hydrated oxide. 4] MDDC - 910 a) Erratic potential differences appear, increasing as the rate of stirring is increased. b) When all circulation is stopped, the electrodes come to the same potential. c) If one electrode is isolated and the solution about the other stirred, a large increasing potential difference appears which persists when stirring is stopped. Therefore, if a drop of titrating reagent is added to the bulk of the solution with one electrode iso- lated and both electrodes at the same potential, as after (b), the resulting natural difference in potential will be more or less covered up by that caused by necessary stirring (c). In our investigations, we have talcen precautions to exclude all traces of oxygen from the reduced solution and from the titrating reagent. Nitrogen was passed through two electric combustion furnaces containing reduced copper, heated to about 45Cf C,and into the titrating vessel through all glass connections and shown to be oxygen-free by the sensitive phosphorous streamer test. If a persistent trace of o.xygen was the disturbing impurity, its effect should have been decreased by greatly increasing the acidity of the solution. Adding acid, however, had no apparent stabilizing action. After extensive tests with several forms of apparatus, we were finally led to conclude that the difficulty was not caused by impure gas but was due to the inherent instability of the over-reduced system. It is noteworthy, that as soon as the first end-point is passed, the erratic behavior (a) suddenly completely disappears and the electrodes come together, whether or not stirring is continued. The U"""^ ion therefore appears to be the cause of the instability. Solutions of U"*"^ salts are known to liberate hydrogen readily; crystalline UH(S04)2, for example, on contact with water, produces a strong evolution of hydrogen. This tendency of an over-reduced uranium solution to evolve hydrogen is probably cat- alyzed at the electrode surface. The potential difference between the electrodes would then depend on the relative rates at which the hydrogen liberated is swept away from the electrodes. With a view to finding a metal with a favorable overvoltage under the conditions of titration, we have tested gold, silver, tungsten, and mercury as electrode materials in addition to the platinum. All these metals behaved alike, except the mercury which was rather insensitive to all effects. Observations on the reduction of uranium by silver Using a micro-silver reductor with small amounts (5 mg) of uranium, we discovered a decided dependence of the extent of reduction upon the temperature and rate at which the uranyl solution was passed over the silver. This reductor was completely jacketed so that the temperature could be con- trolled accurately. Since it contained only a small amount of silver, the solution to be reduced was passed through very slowly. At a given low rate of passage the extent of reduction was found to be a function of the temperature: at 50° reduction to the uranous stage was 55% of quantitative, at 80° it was 105 to 110%, indicating over-reduction. Blank corrections were small and uniform. The reductor gave precise results with 5 mg of iron. SUMMARY The uranium content of a preparation of hydrated UO3 was determined by three independent proce- dures: prolonged successive ignitions in air at 700°, 820°, and 930° to lower oxides of the approximate composition UjOe, further reduction of the same samples to UOj by dry hydrogen at 900°, and titration with eerie sulfate of the U(S04)2 obtained by passing a sulfuric acid solution of the UO3 through a Jones reductor and aerating the somewhat over -reduced solution. The results of these three methods showed an extreme difference of 0.08%. MDDC - 910 r g A modified titration procedure was adapted to 5 mg of UO3, in which the over-reduced uranium sulfate solution was aerated to the uranous stage and oxidized to the uranyl condition by excess per- manganate, the excess being determined by differential potentiometric titration with ferrous sulfate. The results generally agreed closely but failed to give the true quantity of uranium. Extensive application of the differential titration method to over-reduced solutions failed to locate the precise end-point of the U*^ -U'^'' change. The difficulty is believed to reside with the excessively strong reducing power of the U"^^ ion. The reductio:i of uranyl solutions with silver has been generally assumed to proceed quantitatively to the uranous stage. However, some over-reduction appears to be possible. RKFERENCES 1. Furman and Schoonover, J. Am. Chem. Soc. 53:2570 (1931). 2. Lundell and Knowles, J. Am. Chem. Soc. 47:2637 (1925). 3. Blitz and Muller, Z. anorg. Chem. 163:260 (1927). 4. Maclnnes and Cowperthwaite, J. Am. Chem. Soc. 53:555 (1931). 5. Gmelin Handbuch der anorg. Ch., 55:146 (1936). if u^.v,Hs;-„°J„[,ffi■\ll\| g^'2 08907 9601 I 1 i1