A Search for Isotopic Anomalies in Uranium

The 23 aU/23 $ U ratios for nine bulk chondritic meteorites and a terrestrial basalt were measured. The total range in 235U/235U determined for both total meteorites and for acid leaches was from 137.2 to 138.3 and to within the error limits indistinguishable from normal terrestrial U. The typical errors in a single determination are +6•oo (2aM) for a 2 ng U sample from a chondrite. Taking the extreme values of $ 235 U for each measurement the maximum amount of excess 235 U that can be allowed to be present ranges from 2 x 108 to 2 x 10 9 atoms per gram of bulk meteorite. These results do not support the claims of variations in 238U/235 U at the percentage l vels or number of excess 235 U atoms in some of the same meteorites as reported by several other previous workers. Introduction. The purpose of this investigation is to measure the relative abundances of 238 U and 23 s U in various classes of chondrites. Until recently isotopic variations were not expected and only limited high precision work has been done in search of them. In addition, as 23 s U is relatively rare in meteoritic material, there has been very little effort directed toward determining the 238U/23 s U ratio in these materials with high precision and reliability. Numerous measurements on terrestrial materials show no large variations except in special circumstances [Naudet and Renson, 1975]. The "absolute" ratio of terrestrial U is (23aU/235U)(•=137.88 [Shields, 1960; Cowan and Adler, 1976], possibly being uncertain to +1 •oo. Measurements onlunar materials how a 238 U/23 s U ratio indistinguishable from normal U [Rosholt and Tatsumoto, 1971' Fields et al., 1972;Barnes et al., 1972] .Arden [1977] reported analyses of five chondrites which showed that some samples contained an excess of 23 s U. In particular, excesses of 23 s U (290•oo) were reported for some bulk chondrites and certain fractions of dissolution residues gave xcesses upto 2400%0. The results reported by Arden were astounding both in terms of the magnitude of the isotopic shifts and the number of excess nuclei required. These effects would require 4x10 9 to 7x10 •ø excess 2ssU atoms per gram of bulk chondrite and would.imply a high yield (x40 to 300) at Z = 92 for the injected material compared to the value of 10 -3 104 estimated for intermediate Z assuming solar abundances. [See Wasserburg, Papanastassiou and Lee, 1979]. Arden's report stimulated work on U isotopes in meteorites. Chen and Tilton [1979] and Unruh, Hutchison and Tatsumoto [!979a,b] did not observe large effects, but did report 2 $ 8U/23SU ratios a few percent lower than the terrestrial value in some chondrites and Allende inclusions. However, Tatsumoto and Shimamura [1979] reported 238U/235U ratios from 128 to 145 for several meteorites. To clarify this matter, we carried out analyses on nine meteorites and a terrestrial rock. Prior to carrying out the measurements, we considered that the major technical problems should first be addressed. Our approach was (1)to improve the ionization efficiency of U' (2)to firmly establish the limits of detection and reliability by measurement of enriched standards; and (3)to eliminate sources of contamination from "exotic" U isotopes of anthropogenic origin. Experimental Procedures. Because of our concern over contamination, our first step was to clean the laboratory area, establish cleaner chemical procedures, use only new chemical processing equipment, and to stop any handling of 23 s U during the period of the work. The usual procedure for measuring U is by isotope dilution using pure 23 s U as a tracer and results in the presence ofpure 23 s U in the laboratory. The contamination levels necessary to produce aone percent shift in 23 s U abundance for a gram of chondrite are 2 x 10 9 atoms of 23 s U (equivalent to a particle of 2000A). The possibility of contamination from such sources must be a cause of great concern since the vogue of isotopic variations provokes a ki•d of isotopic cred,ulity which inhibits some of us from critically Copyright 1980 by the American Geophysical Union. Paper number 80L0118. 0094-8276/80/0080L-0118501.00 275 examining our measurements. An example of anthropogenic ontamination was reported by Piepgras, Wasserburg, andDasch [1979] for • 48 Nd at the level of 10 6 atoms/gm of seawater due to the unrecognized presence of this isotope in the laboratory from previous work. A 233 U.236U double spike prepared from high purity 233 U and 236 U was used to monitor mass dependent isotope fractionation. The absolute ratio of 233U/236 U and the concentration i the mixed spike were determined using a solution of mixed spike and the NBS U-500 standard [99.94% U3Os, 235U/238U =0.9997+0.001, Garner, Machlan, and Shields, 1971] to correct for fractionation. The ratio in the double spike, 233 U/236 U = 1.0066 is known to within + 1 •oo excluding the errors of 1•oo reported for NBS U-500. Following Arden [1977], we used graphite as an emitter. The procedure used by us was to load a suspension of graphite in water onto a previously outgassed Re filament. This was evaporated yielding a thin uniform layer of graphite. The U samples in 0.1 N HNO3 were then loaded on the filament and dried. Another load of graphite suspension was then deposited and dried making a sandwich with U between the graphite layers. This technique routinely yields "•1% ionization efficiency and low loading blank (<0.5 pg). The ion beam intensity (U +) for 2 ng U was 5 x 10•3 A between 1700ø-1750øC for 2-3 hrs. At each mass the zeroes and peaks were measured with an integration time of 1 sec and a settling time of 3 sec prior to signal integration. To determine instrumental mass fractionation, we measured the ion beams (I) at masses 233 and 236 in conjunction with 235 and 238. The mass jumping sequences were cycled in the order 238U-236U-233U-238U and 236U-235U-235U-233U-236U and data taken in sets of 10 ratios. The ratios (relative to 236U) in a set were averaged and corrected for discrimination (<2%0/mu) and contributions of 238U and 235U from the double spike which were less than 4 x 10-2 •oo and 7•oo respectively. In general, three sets of 10 ratios (238 U:236 U:233 U) were first measured, the illament current then raised such that I236 was similar to that of I238 in the preceding sets and three sets of 236U'-235U:233U were measured. The 238U/235U ratio was calculated from the normalized 238U/236U and 235U/236U ratios in the different sets and the error given as the vector sum of the errors of the latter two ratios in 2o M (mean). To test linearity of the detection system the 238 U/235 U ratio was determined as a function of ion beam intensity. On the Lunatic III mass-spectrometer system we observed a decrease in the measured 238U/236U ratios with increasing signal. All experiments showed the same behavior which was probably caused by saturation of the detector. The error from this apparent deficiency in