From subduction to extension: The tectonomagmatic evolution of the Bulgarian Rhodopes

The Bulgarian Rhodopes provide an unique opportunity to study processes that take place at convergent continental margins. Ophiolite complexes incorporated in the Rhodopean nappe stack as well as the directly overlying post-collisional volcanism allow the investigation of processes from subduction, collision, and subsequent lithospheric extension triggering lithospheric mantle melting that leads to volcanism in collisional orogens. The first part of this dissertation investigates the high-pressure (HP) metamorphic history of ophiolite complexes incorporated in different levels of the Rhodopean nappe stack. The determination of the exact timing of these HP events as well as the characterization of the metamorphic protoliths is crucial for reconstructing the geodynamic evolution of the Rhodopes. In this context, the Lu-Hf isotope system has already been proven useful to date HP mineral assemblages in other Alpine units and was therefore applied to four eclogite samples from different units of the nappe stack of the Bulgarian Rhodopes. The Lu-Hf garnet dating revealed a metamorphic event during the Cretaceous (~ 126 Ma) affecting the highest nappe unit investigated (Upper Allochthon) and an Eocene event (~ 43 Ma) for the Middle Allochthon. These results provide evidence for two separate subduction events in the Rhodopes, in support of previous findings. Moreover, thrusting of the Middle over the Lower Allochthon can be narrowed down to the time span 42 - 34 Ma. The second and third part of this dissertation provides an extensive dataset on the post-collisional volcanism in the Bulgarian Rhodopes as well as for arc lavas from Santorini, which are used as a comparative suite throughout the text. The Bulgarian post-collisional volcanism is characterized by a high magnitude of incompatible trace element enrichment, which is particularly shown by its affiliation to the high-K and shoshonite series. Two petrogenetic models that were previously proposed for the generation of high-K magmas involve the melting of ancient, enriched lithospheric mantle sources (single-stage model) or melting triggered by young refertilization of subduction-related components derived from subducted sediments or oceanic crust (multi-stage model). These two models are tested for the Bulgarian K-rich rocks, based on new major, trace element and Sr-Nd-Hf-Pb isotope compositions. The single-stage model is evaluated by Sr-Nd isotope modelling assuming the presence of ancient lithospheric mantle domains whereas the multi-stage model is assessed by comparing compositions of the Bulgarian lavas with those of lavas from Santorini. Santorini Island lavas are thought to sample the current trace element and isotope inventory of the long-lived Aegean subduction-zone system. This northward facing system has been active since late Jurassic/Early Cretaceous and was potentially involved in refertilizing the mantle sources of the Bulgarian lavas. In addition to the Bulgarian lavas, we present new major, trace element and Sr-Nd-Hf-Pb isotope data for Santorini. Modelling of Sr-Nd isotope compositions of the Bulgarian lavas argues for a young (Meso- to Cenozoic) source enrichment. Therefore, single-stage models involving melting of ancient, > 1 Ga old lithospheric mantle can be confidently ruled out, in agreement with tectonic models for the region. The enriched isotope signatures found in the Bulgarian lavas, coupled with a pronounced enrichment in incompatible elements, instead indicate mantle refertilization by subduction components similar to presently subducted continent-derived sediments. Notably, the Bulgarian lavas record a predominant influx of fluid-like subduction components when compared to the Santorini lavas. Collectively, the data presented for the Bulgarian lavas are thus clearly in favour of a multi-stage model. The last part of this dissertation focuses on extended high-field-strength element (HFSE) systematics in the Bulgarian and Santorini lavas. The extended HFSE (Nb, Ta, Zr, Hf, W, Mo, and Sb) are of particular interest in magmatic rocks as their fractionations hint towards specific residual phases in their source regions, such as rutile, allanite, zircon, micas, and sulphides. Tungsten, Sb, and Mo are of particular importance in that they are mobilized in subduction zones by fluids and melts at distinct temperatures and redox conditions and might provide important insights into the conditions and magnitude of source enrichment. However, no significant fractionation of the HFSE ratios (Nb/Ta, Zr/Hf, Zr/Nb) compared to MORB were observed in the Santorini lavas and in the Bulgarian high-K rocks. An influence on the HFSE budget by residual phases like allanite, zircon, or phengite can be largely ruled out, whereas trace amounts of residual rutile in the source may account for the slightly lower Nb/Ta observed in the dataset than expected for bulk sediment addition. The W-Sb-Mo systematics of both sample suites furthermore confirm the predominance of subducted sediments on the incompatible trace element budget, which is dominated by more fluid-like components in the Bulgarian lavas and melt-like in the Santorini lavas