doi:10.1016/S0016-7037(02)01169-9 Textural and geochemical discrimination between xenotime of different origin in the Archaean Witwatersrand Basin, South Africa NATALIE KOSITCIN, 1, *NEAL J. MCNAUGHTON, 1 BRENDAN J. GRIFFIN, 2 IAN R. FLETCHER, 1 DAVID I. GROVES, 1 and BIRGER RASMUSSEN 1 1 Centre for Global Metallogeny, School of Earth and Geographic Sciences, University of Western Australia, Crawley WA 6009, Australia 2 Centre for Microscopy and Microanalysis, University of Western Australia, Crawley WA 6009, Australia (Received March 29, 2002; accepted in revised form August 27, 2002) Abstract—Xenotime (YPO 4 ) of detrital, diagenetic, and hydrothermal origin within siliciclastic rocks of the Archaean Witwatersrand Basin, South Africa, has been identified on the basis of petrography and in situ ion microprobe (SHRIMP) age data. The chemical composition of xenotime, determined by in situ electron microprobe analysis, can be correlated with its origin. This allows the origin of any xenotime grain to be assessed by a non-destructive microanalytical method prior to ion microprobe geochronology. The main chemical discriminators are MREE-HREE abundance, normalised HREE slope and Eu anomaly, and, in some cases, U and Th contents. Igneous-detrital xenotime ( 2800 Ma) is distinguished from diagenetic (2780 Ma) and hydrothermal ( 2780 Ma) xenotime in having lower Eu, Dy, and Gd concentrations and a distinctively lower Gd/Yb ratio. Hydrothermal xenotime has distinctively lower U and Th concentrations when compared to igneous-detrital and diagenetic xenotime. Three separate hydrothermal fluid events and episodes of post-diagenetic xenotime growth are recognised in the geochemical and geochronological data, which correspond in time to the extrusion of the Ventersdorp lavas at 2720 Ma, the emplacement of the Bushveld Igneous Complex at 2061 Ma, and an event of unknown affinity at 2210 Ma. Although geochemical discrimination of the xenotime types from the Witwatersrand Basin, in combination with careful petrography, appears achievable, universal application of these discriminators to xenotime in other sedimen- tary basins remains untested. Copyright © 2003 Elsevier Science Ltd 1. INTRODUCTION Xenotime (YPO 4 ) is a common accessory phase in granites (e.g., Wark and Miller, 1993) and pegmatites (Amli, 1975; Demartin et al., 1991) and is a common detrital mineral in siliciclastic sedimentary rocks (van Emden et al., 1997). It is also a common authigenic mineral that grows on detrital zircon grains in siliciclastic sedimentary rocks of all ages (Rasmussen, 1996). Diagenetic xenotime typically occurs as m-scale, py- ramidal, isostructural outgrowths from zircon and is considered to begin growth within tens of metres of the sediment-water interface. However, xenotime also forms during burial diagen- esis, hydrothermal alteration, and low-grade to high-grade metamorphism (Rasmussen et al., 2001; England et al., 2001a; Brown et al., 2002; Tallarico et al., in press). With such diverse modes of formation, it is important to develop criteria to determine the origin of xenotime grains in common siliciclastic sedimentary rocks to complement the rapidly evolving disci- pline of xenotime geochronology (e.g., McNaughton et al., 1999; Fletcher et al., 2000; Rasmussen et al., 2001). Petrogra- phy remains the main tool to determine the origin of xenotime. However, this study shows that trace-element geochemistry may also help to distinguish xenotime of differing origin. Only a few multi-element electron microprobe studies have been conducted on the composition of xenotime. It has been analysed from granitic rocks (Wark and Miller, 1993; Casillas et al., 1995; Bea, 1996; Fo ¨ rster, 1998), pegmatites (Amli, 1975; Demartin et al., 1991), placer deposits (van Emden et al., 1997), and metamorphic rocks (Suzuki and Adache, 1991; Franz et al., 1996; Bea and Montero, 1997). Systematic varia- tions in xenotime composition have been demonstrated be- tween granites and aplites (composite S-type Sweetwater Wash two-mica granite pluton; Wark and Miller, 1993) between biotite granites, two-mica granites, S-type Li-mica granites, and A-type Li-mica granites (Fo ¨rster, 1998), and between metapelites of differing metamorphic grade (Bea and Montero, 1997). Therefore, variations in the geochemistry of xenotime may provide important information on both the origin and conditions of its formation. This study presents the results of an electron microprobe investigation of xenotime from the Archaean Witwatersrand Basin, South Africa (Fig. 1). The origin of the xenotime ana- lysed has been interpreted from morphological and in situ ion microprobe (SHRIMP) age data. The aim of this study is to identify diagnostic criteria for the discrimination of xenotime genesis based on geochemical data. Geochemical discrimina- tion of xenotime origin using the non-destructive electron mi- croprobe method is at a finer scale than SHRIMP analysis, allowing detailed imaging of textural relationships using a quicker, cheaper, and more readily available analytical method. If successful in discriminating xenotime origin, geochemical analyses and imaging may become a routine method of assess- ing samples prior to SHRIMP analysis. Xenotime of known origin from various locations has been analysed to compare with both textural types of xenotime from the Witwatersrand Basin and with data from the literature. 1. GEOLOGIC FRAMEWORK The Late Archaean Witwatersrand Supergroup (Fig. 2) is an approximately 7.5 km-thick succession of siliciclastic sedimen- * Author to whom correspondence should be addressed (nkositci@geol.uwa.edu.au). Pergamon Geochimica et Cosmochimica Acta, Vol. 67, No. 4, pp. 709 –731, 2003 Copyright © 2003 Elsevier Science Ltd Printed in the USA. All rights reserved 0016-7037/03 $22.00 + .00 709