Isotope dilution U-Th-Pb age determinations using a 205Pb/235U spike, and no chemical separation, were made on two 10-30 mg monazite grains and one xenotime grain from a sample of migmatites in the Aar Massif, Swiss Alps. The same monazite grains were also analysed by micro-XRF, which gave the ages of 275 and 330 Ma with errors of 20 to 50 Ma. Monazites from the same sample have previously been dated by an electron and ion microprobes, which gave the ages of 356 ± 64 and 378 ± 78 Ma, and 292-297 ± 2 Ma respectively.

U and Th were determined on a Nu Instruments ™ multicollector ICP-MS. Sample runs were bracketed with a mixed U-Th standard to enable correction for fractionation and the ionisation efficiency difference between U and Th. Pb was run on a single collector Ion Instruments (AVCO) TIMS using an electron multiplier in analog mode.

The monazite grains gave similar U-Th-Pb sets of ages, concordant within analytical error. In the order 208Pb/232Th, 206Pb/238U and 207Pb/235U ages, they are, grain 1: 329 ± 5 Ma, 327 ± 4 Ma, and 338 ± 17 Ma; grain 2: 314 ± 3 Ma, 313 ± 3 Ma and 324 ± 8 Ma. The results are interpreted as indicating the crystallisation ages of the monazites. The xenotime grain yielded discordant results: 222 ± 14 Ma, 341 ± 14 Ma and 2890 ± 90 Ma, probably indicating its Alpidic formation with radiogenic Pb inherited from Precambrian zircon, and incorporation of additional Th. Similarly, the previous zircon results on the same sample reveal a scatter typical of multiple discordance.

The monazite Pb runs showed a greatly suppressed Pb ionisation in TIMS, with signals between 40 and 100 x smaller than expected for the amount of Pb. As the xenotime analysis did not show this effect (and no-chemistry zircon single grain Pb analyses don’t either) the effect may be due to abundant Ce on the filament: Ce could act as an electron donor preventing the formation of Pb+. For isotope dilution analyses this merely reduces the precision and indicates a necessity for miniaturised chemistry on monazites in future. However it may imply that, for ion microprobe dating, a strongly matrix-dependent ionisation of Pb could reduce the confidence in measured Pb concentrations, and therefore in U-Pb and Th-Pb ages, and may account for the discrepancy between TIMS and SIMS ages for our sample. The large deviation of EMP ages from those by other techniques remains unexplained.