Garnet peridotites provide insight into geochemical processes within the upper mantle, which can be recorded in the mineral zircon. We separated zircon from a garnet peridotite of the Erzgebirge (Zöblitz), Bohemian Massif, and dated it by means of U-Pb with SHRIMP, determined its trace and REE geochemistry and investigated its inclusions. Two types of zircon were distinguished: (a) Inclusion-free zircon with weak cathodoluminescence (CL) emission. For some of these zircons oscillatory zoning or, partly, sector zoning could be observed indicating precipitation from a melt or fluid. These zircons are characterised by high HREE contents and a strong negative Eu anomaly. The latter feature indicates that they formed within the stability field of plagioclase. SHRIMP results reveal a 206Pb/238U age of 332.1 ± 4.8 Ma (error at 95 % c.l.). This age is younger than the one ascribed to the metamorphic peak by previous studies (ca. 341 Ma) and similar to ages suggested for exhumation stages (ca. 332–330 Ma). (b) Inclusion-bearing, very U-rich zircons with no, or very weak, CL emission. The inclusions (a few μm large) are quartz, albite, K-feldspar and felsic melt. These zircons have high trace- and REE contents and are especially rich in LREE. Their chondrite-normalized REE pattern is flat. They formed later in the exhumation path than the inclusion-free zircons. Zircons of this type did not yield a reliable SHRIMP age because of their very high U contents. Garnet shows a REE pattern typical for this mineral. MREE-HREE partitioning between zircon and garnet reveals that the two minerals were far from equilibrium. This applies for both zircon types. The whole-rock chemistry displays absolute REE abundances slightly higher than the chondritic values (except for La and Ce) and enrichment in LREE.
Our results show that zircon in the garnet peridotite formed during exhumation, within the stability field of plagioclase (presence of albite inclusions or negative Eu anomaly) during two distinct stages by infiltration of two types of fluid/melt with different chemistry (very different trace and REE composition of the two zircon types), far from equilibrium with the (U)HP mineral garnet (DMREE-HREEzrc/grt ≫ 1). These fluids/melts probably originated from the immediately adjacent felsic gneisses. It is likely that two pulses of influx of a highly differentiated melt (very U-rich zircons, presence of albite) occurred and led to precipitation of zircon: the first at higher-T (no mineral inclusions in zircon) and the second at lower T, closer to the solidus (felsic inclusions in zircon).