231Pa/235U Measurements in Volcanic Rocks by Thermal Ionization Mass Spectrometry

Bernard Bourdon Laboratoire de Géochimie et Cosmochimie-CNRS, IPGP T14-24, 4 Place Jussieu,

75252 Paris Cedex 05, France

bourdon@ipgp.jussieu.fr

Claude J. Allègre Laboratoire de Géochimie et Cosmochimie-CNRS, IPGP T14-24, 4 Place Jussieu,

75252 Paris Cedex 05, France

Jean-Louis Joron Laboratoire Pierre Süe, CEA-CNRS, CE/Saclay, BP n° 2, 91190 Gif-sur-Yvette, France

The presence of large 231Pa-235U (Goldstein et al., 1993; Goldstein et al., 1994) disequilibria in young volcanic rocks compared to 230Th-238U warrants further investigation of this isotopic system as a technique for understanding trace element fractionation during magma genesis. Among other applications this system can be used to examine carbonate dating, sedimentation rates and particle scavenging in the oceans. We present here a new high resolution technique for analyzing 231Pa/235U ratios in rock matrices by thermal ionization mass spectrometry.

Compared with 230Th, the abundances are ×50 times lower but the ionization efficiency attainable with Pa is greater. There are, however, some difficulties in the chemical separation due to, the hydrolysis and polymerization of Pa when it is not complexed with fluoride. In order to solve the problem of isotopic homogeneization for an element that is difficult to extract with a 100% yield, we have used neutron irradiation to produce the spike (233Pa, T1/2=27 days) in-situ by neutron activation of 232Th contained in the rock powder. With this technique, isotope homogeneization is already achieved prior to sample dissolution. About 0.5-1 g of rock powder was irradiated for 30 minutes in the Orphée reactor at Saclay under a 2x1013 n/cm2/s neutron flux to obtain a 231Pa/233Pa ratio of ×1. Standard rock GSN was used to monitor the flux across the irradiation vial and no systematic lateral gradient was found within the measured uncertainties (×2-3%). The powder was then dissolved in HF-HNO3 and the protactinium is kept in a fluoride or sulfate complex form during column chemistry to avoid losses by hydrolysis or polymerization. The yields measured by g-counting of 233Pa are ×80-90% and, unlike other published technique, we do not use solvent extraction for the Pa separation (Anderson and Fleer, 1982; Pickett et al., 1994). The high field strength elements which may potentially interfere during mass spectrometry are also efficiently separated, and they cannot be detected during analysis.

Pa is loaded as a fluoride on a tungsten filament and run as a double oxide on a Finnigan MAT 262 mass spectrometer with an ionization efficiency of 2-4”. Preliminary results show the feasibility of this new analytical scheme for a few hundred femtograms of Pa. An Icelandic dacite with a known 231Pa/235U ratio (analyzed by D. Pickett at Los Alamos) is used as a flux monitor and gives 233Pa/231Pa that are consistent with the nominal neutron flux to within 4%. Replicate analyses for an alkali basalt from the Comores give reproducible a (231Pa/235U) ratio=2.53±0.08. These and other results will be discussed at the meeting.

References

Anderson, R.F. & Fleer, A.P., Anal. Chem. 54, 1142-1147 (1982).

Goldstein, S.J., Murrell, M.T. & Williams, R.W., Earth Planet. Sci. Lett. 115, 151-159 (1993).

Goldstein, S.J., Perfit, M.R., Batiza, R., Fornari, D.J. & Murrell, M.T., Nature 367, 157-159 (1994).

Pickett, D.A., Murrell, M.T. & Williams, R.W., Anal. Chem. 66, 1044-1049 (1994).