Sources and timing of pyroxenite formation in the sub-arc mantle: Case study of the Cabo Ortegal Complex, Spain

•We use isotope geochemistry, geochronology and modelling of REE diffusion timescale.•Cabo Ortegal pyroxenites record Cadomian arc magmatism off a pre-Gondwanan block.•Melt–rock interaction occurred during the reworking of a rifted continental margin.•Subduction metasomatism and metamorphism obscure the age of orogenic mantle rocks.•2nd-stage Nd model ages may provide robust time constraints on pyroxenite formation. Pyroxenites exposed in ophiolites and orogenic peridotite massifs may record petrogenetic processes occurring in mantle domains generated and/or transferred in supra-subduction environments. However, the timing of their formation and the geochemical characteristics of their source region commonly are obscured by metamorphic and metasomatic overprints. This is especially critical in arc-related environments, where pyroxenites may be formed during the differentiation of primitive magmas. Our approach combines Sr- and Nd-isotope geochemistry and geochronology, and modelling of REE diffusion, to further constrain the origin of a well-characterized set of pyroxenites from the arc-related Cabo Ortegal Complex, Spain. In the light of petrological constraints, Sr- and Nd-isotope systematics consistently indicate that cpx and amphibole have acquired disequilibrium during two main episodes: (1) a magmatic/metasomatic episode that led to the formation of the pyroxenites, coeval with that of Cabo Ortegal granulites and corresponding to the incipient stage of a potential Cadomian arc (459–762 Ma; isochron and second-stage Nd model ages); (2) an episode of metamorphic amphibolitization upon the percolation of relatively unradiogenic and LREE-enriched hydrous fluids, subsequent to the delamination of the pyroxenites from their arc-root settings during Devonian subduction. Calculations of diffusional timescale for the re-equilibration of REE are consistent with this scenario but provide only poor additional constraints due to the sensitivity of this method to grain size and sub-solidus temperature. We thus emphasize the necessity to combine isochron ages and Nd model ages corrected for radiogenic ingrowth to put time constraints on the formation of subduction- and/or collision-related pyroxenites, along with petrological and geochemical constraints. Homogeneous age-corrected 143Nd/144Nd of 0.5121–0.5125 (εNd between 0 and +7.5) and 87Sr/86Sr of 0.7037–0.7048 provide information on the sources of the metasomatic agents involved (and potentially the parental mel