Search

Graphical abstract Highlights • Link between low–grade metamorphism and deformation on the ore genesis. • Carbonaceous phyllites and carbonate rocks as sources for ore–forming fluids. • P–T–t evolution from metamorphic peak to ore deposition based on fluid and mineral data. • Contrasting f O 2 – f S 2 as a chemical control for ore precipitation by fluid–rock interaction. • The Buracão area in the context of orogenic metasediment–hosted gold deposits. Abstract The orogenic Au (As, Bi, Ag) mineralization (up to 31 g/t) of the Buracão area is hosted in the metasedimentary Paleo–Mesoproterozoic Traíras Formation, Araí Group, Brasília Fold Belt, Brazil. In order to discuss the P–T–X conditions that generated this mineralization, we present mineralogical, geochemical and fluid inclusion data of veins, gold and wall rocks. Coupled with previous structural data these additional data allow us to link metamorphism and deformation with fluid migration and ore genesis. The integration of fluid inclusion microthermometry with mineral geothermobarometers and structures show that rocks of the study area evolved from: (i) syn–D 1a metamorphic peak (400–560 °C, <6 kbar), through to the (ii) development of syn–D 1b type I barren veins from low salinity H 2 O–NaCl fluids during syn– to late–metamorphic peak; and to the (iii) development of syn–D 2a type II veins with the pyritic ore precipitation from low to high salinity H 2 O–CO 2 –CH 4 –NaCl ± HCO 3 fluids during post–metamorphic peak (345–285 °C, 2.7–1.6 kbar). The high fineness gold nuggets accompanied by As, Bi and Ag within pyrite–arsenopyrite–bismuthinite association in the pyritic ore allows the Buracão area to be classified as "a gold-only orogenic ore deposit". Fluid inclusion compounds and the enrichment of As, Bi, Pb, Cu, Te, Se and Ag in pyrite from the pyritic ore point out to metamorphism of carbonaceous phyllites and carbonate rocks of the Traíras Formation as a source for ore–forming fluids. We suggest that sulfur and gold in these fluids were derived by metamorphism of sedimentary rocks and conversion of pyrite into pyrrhotite in a metamorphic–hydrothermal system. The Au(HS) 2 −–bearing fluids percolated along intrabasinal D 2a regional shear zones through the Traíras Formation towards the Buracão area, where they precipitated type II ore–bearing veins within local structures. The ore precipitation was induced by contrasting f O 2 – f S 2 related to the interaction between the sulfur-bearing fluid and host magnetite–graphite phyllites. The lithological control is represented by a series of reactions such as 2FeO (in silicates/magnetite) + 4Au(HS) 2 − (aq) + 4H+ (aq) ↔ 2FeS 2(s) + 4Au (s) + 4H 2 S (aq) + 2H 2 O (l). The deposits of the Buracão area and other orogenic gold deposits hosted in metasedimentary rocks in the Brasília Fold Belt and worldwide confirm the potential of ore genesis in zones that do not show any relation with (meta)igneous rocks, thus, expanding the exploration possibilities in metasedimentary successions. [ABSTRACT FROM AUTHOR]

Abstract: European Tertiary sedimentary basins as the Forez Graben, France, are potential records of continental paleoclimates. The Forez Basin hosts deposited and precipitated sediments of Oligocene to Miocene age. Geochemical data of carbonates indicate strictly continental origin starting at Eocene–Oligocene with tropical to temperate climate conditions, then during the Middle Miocene a temperate continental climate prevails. Combining volume of calcite deposits and their geochemical data, volumes of large lakes and evaporation/inflow ratios were reconstructed. The Late Miocene in the Forez Graben has been affected by dissolution and secondary precipitation of calcite, barite, which is the result of wetter and colder climate conditions. These lake volume calculations represent the first estimation of large lakes volumes in Western Europe during the Miocene. [Copyright &y& Elsevier]

Summary: A re-appraisal of interactions between basinal waters and hydrothermal fluids within the silici-clastic series of the intracontinental Cenozoic Forez Basin is proposed using major element and isotope geochemistry together with thermodynamic models. Recharge processes and shallow sedimentary units in the Forez Basin were compared using major elements and isotopic data (87Sr/86Sr, δ18O, δ2H, δ13C, Tritium and 14C) for waters, sedimentary carbonates and Hercynian basement rocks. Waters derived from Late Pleistocene to Early Holocene surface recharge followed by mixing and chemical evolution through percolation into deep groundwater horizons are investigated. Isotopic ratios and chemical speciation computed for deeper waters from 0 to 200m depth indicate that mineralised waters were produced by initial surficial dissolution of calcite in calcareous-mudstone providing its Sr isotopic signature. At about 200m depth, Na–H3O+ cation-exchange with clayey-sand rich levels provides additional Ca, Na, and OH contents, for carbonate-saturated water. Na–H3O+ cation exchange can not contribute large Na–HCO3 concentrations. Geochemical changes such as increasing δ13CDIC values observed from 200 to 500m depth were modelled by addition of thermal water and geogenic CO2 in increasing proportions with depth (500mbs). Thermodynamic calculations of mineral proportions indicate water–rock interactions with primary minerals in the metamorphic basement and secondary minerals observed in the basin. Water temperature and its meteoric origin suggest a basin-wide recharge system with water–rock interaction driven by heat flow rather than through deep crustal fractures. [Copyright &y& Elsevier]

Abstract: In north-eastern Algeria, autochthonous limestone units at Djebel Debbagh varying in age from Aptian to Cenomanian underwent burial diagenesis followed by deformation during the Alpine orogeny. Karst windows formed during Oligocene to Miocene times, and are now filled with clay minerals, which currently help to sustain a local mineral industry. The largest karst window at Djebel Debbagh contains substantial volumes of 7 Å halloysite ore. Mineralogical investigations on the 7 Å halloysite indicate a non-expandable material, with measured water contents varying from 14.1 to 15.5 wt.% H2O. Non-expandable 7 Å halloysite showed limited hydrogen-exchange capacity during a 24-month laboratory experiment, suggesting that hydrogen and oxygen isotopes had not exchanged since equilibrium at the time of halloysite formation. Therefore, stable-isotope data were used to calculate reliable temperature profiles and to determine fluid composition. It is suggested on the basis of geochemical and mineralogical data that 7 Å halloysite formed between 25 °C and 40 °C in equilibrium with meteoric water during Eocene to Miocene times. Estimates of supergene temperatures and calculated stable-isotope values for rainfall correspond to continental climate conditions during Miocene times, coeval with the closure of the Liguro oceanic basin. [Copyright &y& Elsevier]

The Paleoproterozoic Thelon Basin, located on the border between Nunavut and the Northwest Territories of Canada, is a contemporaneous analog of the uranium-rich Paleoproterozoic Athabasca Basin in Canada. Early diagenesis resulted in precipitation of extensive hematite on the surfaces of detrital quartz grains throughout the Thelon Formation and minor hydroxy-phosphate in veins locally. Continued diagenesis then resulted in syntaxial quartz cementation of detrital quartz at 130°C from fluids having ca. 17 wt.% equivalent NaCl, similar to the Athabasca Basin. Cementation of this type is most pronounced in fine-grained sequences in the Thelon Basin. A period of extensive desilicification during continued burial was followed by formation, at ca. 200°C, of peak-diagenetic illite having Ar–Ar ages of ca. 1400–1690 Ma in the Thelon Formation. This illite was associated with fluids with δ[sup 18] O and δD values of ca. 6‰ and –50‰, respectively, similar to those during peak diagenesis of the Athabasca Basin. Although the timing, salinity, and isotopic composition of the peak-diagenetic fluids in the Thelon and Athabasca Basins are similar, the peak-diagenetic mineral assemblage in the Athabasca Formation is dickite and illite, with minor dravite and goyasite rather than simply illite. Consequently, the fluids at peak diagenesis, which in the Athabasca Basin are synchronous with formation of world-class unconformity-type uranium deposits, had different compositions in each basin. Post-peak diagenesis in the Thelon Basin was quite distinct from that in the Athabasca Basin in that illite was replaced in the central portion of the basin by K-feldspar and then sudoite, which crystallized from saline brines at ca. 1000 Ma and 100°C. Evidence for later infiltration of these brines is absent in the Athabasca Basin, although uranium mobilization at ca. 900 Ma from fluids having the same characteristics as those at peak diagenesis was pronounced in the Athabasca Basin. Recent incursion of meteoric waters along reactivated structures into the Athabasca Basin has variably affected hydrous and uraniferous minerals, but evidence for this is lacking in the Thelon Basin. The Thelon Basin reflects less intensive fluid–rock interaction in its early history than that recorded in the basal units of the Athabasca Basin.Le bassin de Thelon (Paléoprotérozoïque), situé sur la frontière entre le Nunavut et les Territoires du Nord-Ouest du Canada, est l'analogue contemporain du bassin riche en uranium d'Athabasca (Paléoprotérozoïque), au Canada. Une diagenèse précoce a fait précipiter une grande quantité d'hématite sur les surfaces des grains de quartz détritique à travers toute la Formation de Thelon et, localement, une faible quantité d'hydroxy-phosphate dans les veines. La poursuite de la diagenèse a ensuite résulté en la cémentation quartzique par surcroissance du quartz détritique à 130 ºC à partir de fluides qui avaient environ 17 % (poids) d'équivalent NaCl, semblable au bassin d'Athabasca. La cémentation de ce type est le plus prononcé dans les séquences à grains fins du bassin de Thelon. Une période de désilicification extensive, au cours d'un enfouissement continu, a été suivie de la formation d'illite au sommet de la diagenèse, à environ 200 ºC; les âges Ar–Ar de l'illite sont d'environ 1400–1690 Ma dans la Formation de Thelon. Cette illite était associée à des fluides ayant des valeurs de δ[sup 18] O et δD d'environ 6‰ et – 50 ‰ respectivement, semblables a celles présentes durant le sommet de la diagenèse du bassin d'Athabasca. Bien que la séquence, la salinité et la composition isotopique des fluides au sommet de la diagenèse dans les bassins de Thelon et d'Athabasca soient semblables, l'assemblage minéralogique du sommet de la diagenèse dans la Formation de l'Athabasca consiste en dickite et en illite avec de la dravite et de la goyasite plutôt que simplement de l'illite. Par conséquent, les fluides au sommet de la diagenèse, lesquels dans le bassin d'Athabasca sont synchrones à la formation de dépôts d'uranium de classe mondiale et de type discordance, ont des compositions différentes dans chaque bassin. Dans le bassin de Thelon, la diagenèse après le sommet a été assez différente de celle dans le bassin d'Athabasca en ce que l'illite dans la partie centrale du bassin a été remplacée par un feldspath-K puis par de la sudoïte, qui a cristallisé à partir de saumures vers 1000 Ma et à 100 ºC. Il n'existe aucune preuve d'une infiltration postérieure de ces saumures dans le bassin de l'Athabasca, bien que la mobilisation de l'uranium, vers 900 Ma et à partir de fluides ayant les mêmes caractéristiques, ait été forte dans le bassin de l'Athabasca. Une récente incursion d'eaux météoriques dans le bassin de l'Athabasca, le long de structures réactivées, a affecté les minéraux hydratés et uranifères à divers degrés, mais il n'y a pas d'évidence de cela dans le bassin de Thelon. Le bassin de Thelon reflète une interaction fluide–roche moins intense au début de son histoire que celle enregistrée dans les unités à la base du bassin de l'Athabasca.[Traduit par la Rédaction] [ABSTRACT FROM AUTHOR]

The 40Ar/39Ar ages (muscovite and biotite) and P-T estimations presented here allow us to reconstruct the Pressure–Temperature–time (P–T–t) path from the metamorphic peak to the generation of the orogenic Au (As, Bi, Ag) mineralization in the Buracão area. The application of software-based pseudosections for bulk phyllite composition allows us to better recognize three main stages in the P-T-t path. Rocks from the study area evolved from (i) syn-D 1a greenschist metamorphic peak (T = 480–550 °C and P < 4.0 kbar) attained at ca. 538–530 Ma during the W-E shortening related to local D 1 deformational stage, which is regionally reported in the external domain of the northern Brasilia Fold Belt. These younger ages, compared to those of mafic and felsic granulites from the northernmost internal domain west of the study area (ca. 610–570 Ma), indicate that the metamorphic peak in the external domain was attained after the internal domain. The (ii) syn-D 1b barren type I quartz vein formation (T = 510–260 °C and P < 4.0 kbar) at ca. 531–525 Ma; and (iii) syn-D 2a ore-bearing type II quartz vein formation and gold mineralization (T = 345–285 °C and P = 2.7–1.6 kbar) at ca. 522–515 Ma suggest that low temperature may represent cooling P-T conditions of the peripheral supracrustal cover. Moreover, the diachronicity between the local metamorphic peak and the gold mineralization (22 M.y.) is consistent with the structural relationship between gold veins and post-metamorphic peak D 2a high deformation zones, which may have acted as fluid corridors through the metasedimentary pile. These features are coherent with the time span observed in several orogenic gold deposits worldwide, especially those associated with high-angle transpressive shear zones developed in the late stages of orogeny. The combination of the Ar–Ar petrochronology with previous published fluid reconstruction and tectonic data suggests metal-rich fluid percolated along regional shear zones through the Traíras Formation under post-metamorphic peak conditions. This regional tectonic and chronology might be useful for geodynamical reconstruction and as a prospection guide in poorly studied orogenic gold of the area. [Display omitted] • 40Ar/39Ar-petrochronology to reconstruct the P–T–t path from the metamorphic peak to orogenic gold deposit. • syn-D 1a metamorphic peak (T = 480–550 °C and P < 4.0 kbar) at ca. 538–530 Ma. • syn-D 1b barren type I vein formation (T = 510–260 °C and P < 4.0 kbar) at ca. 531–525 Ma. • syn-D 2a ore-bearing type II vein formation (T = 345–285 °C and P = 2.7–1.6 kbar) at ca. 522–515 Ma. • Contribution to reconstruct the tectonic and low-temperature evolution of the northern Brasília Fold Belt. [ABSTRACT FROM AUTHOR]

The orogenic pyrite–rich Au (As, Bi, Ag) mineralization (up to 31 g/t) of the Buracão area is one of several orogenic deposits in the Brasilia Fold Belt (BFB; Central Brazil), exposed to tropical weathering that shows extensive oxidized profiles. The distribution and thickness of these gossan profiles lead to variable gold enrichments and uncertainties to gold prospecting in this region. In order to investigate physicochemical processes related to the evolution of the gossan cover and the supergene gold enrichment, we present mineralogical, textural and geochemical data from each zone of twenty gossan profiles in the Buracão area. These results show that the mineralogy and textures along the gossan profile reflect oxidation, hydrolysis, dissolution and precipitation processes controlled by different Eh–pH conditions, which are controlled by supergene conditions and hypogene ore composition. The geometry, distribution, and thickness of gossan profiles, as well as the style of gold occurrence, reflects textures of the hypogene orebodies, while the weathering has an important role in the gold chemical refining (i.e., increasing ca. 25% in gold fineness) and the distribution of gold grades along the profile. These points allowed us to identify that the leached spongy hematite–rich zone in the upper indigenous profile contains the higher fineness gold grades (i.e., the "bonanza level"), and Bi–minerals such as waylandite and bismutite that may be important pathfinders for high gold grades in these gossan profiles. Our results demonstrate that ore prospection in gossans over Au–As–Bi–Ag deposits should be done focusing on leached zones, different to that traditionally conducted in gossans related to base metal sulfide deposits that have focused on precipitation zones. The similarities between the hypogene mineralization of the Buracão area and many deposits in the BFB and worldwide encourage the application of this approach elsewhere, which improve gold prospecting in weathering profiles over orogenic deposits. • Gossan evolution over orogenic pyrite–rich Au (As, Bi, Ag) mineralization • The role of environment, climate and primary ore composition on the gossan evolution • Mineralogy and textures related to physicochemical processes controlled by contrasting Eh–pH conditions • The nature of supergene gold enrichment and its distribution along the gossan profile • The implication of mineralogy and textures as prospecting guides in weathered profiles [ABSTRACT FROM AUTHOR]

Abstract The mineralization of the Lavras do Sul Mining District and Seival Mine are controlled by brittle structures associated with the emplacement of the volcanogenic rocks and granitoids located in Lavras do Sul region, southernmost Brazil. This magmatism is associated with the evolution of the Camaquã Basin, an Ediacaran basin generated during the post–collisional stage of development of the Dom Feliciano Belt. The regional tectonic process has led to the activation of a large–scale high–angle shear zones and intense fracturing in the brittle to brittle–ductile regime. The relative chronology of structures and stress–field variations were recognized through remote sensing and structural data. The ore–deposits are controlled by extensional fractures and normal fault systems with a N70°W/70–85°NE direction in Lavras do Sul and a N40–60°E/70–88°NW direction in Seival and are disseminated or occur associated with quartz and calcite veins, respectively. We made a comparison of structures going from regional scale to smaller mineralized areas. The structural analysis of remote sensing and geophysical images (aeromagnetometry and aerogamma–spectrometry) integrated with structural data allowed us to reconstruct the paleostress tensor that generated the main fault systems and associated fractures of the host rock, which enabled the hydrothermal fluid circulation and the ore deposition. The integration of structural data such as fault slip planes, slickenlines, shear fractures systems, extensional fractures and kinematic indicators with the stratigraphic data allowed us to reconstruct the main directions of the tectonic efforts associated with the emplacement of the mineralizing magmas. The mineralization is associated with the sinistral N70°W/80–88°NE oriented Palma–Jacques Fault in the Lavras do Sul Mining District as well as the sinistral N50°E/80–88°SE oriented Cabritos–Perau Fault and the sinistral N20°E/80–88°SW oriented Segredo Fault in Seival Mine region. Their maximum compressional stresses are subhorizontal and NW–SE oriented, while a late subhorizontal and NE–SW oriented compressional stress affected only the rocks of the Seival Mine. These two main compressive stresses are associated with the development of a transcurrent (escape) tectonics during the final evolution of the Dom Feliciano Belt, constrained by U Pb zircon geochronology between 600 and 580 Ma. Transcurrent and normal NE–SW oriented faults are more expressive in volcanogenic rocks of the Seival Mine area, while extensional vein–type NW–SE oriented fractures are the most important structures in the Lavras do Sul area. We suggested that N70°W/70–88°NE oriented monzonitic dikes in Lavras do Sul and N50°E/70–88°NW oriented andesitic to trachyandesitic dikes in Seival area are probably the source of the mineralization. Highlights • Brittle structural control in Au Cu Ag mineralizations. • Regional stress associated to faults and fractures ore–bearing. • Compressive stress related to mineralized areas in Camaquã Basin. • Neoproterozoic faults associated to ore deposition. [ABSTRACT FROM AUTHOR]

The Trindade Island is located in the South Atlantic Ocean, 1170 km from the Brazilian coast and represents the eastern end of the submarine E–W Vitória–Trindade Chain. This is interpreted as the Trindade plume track beneath the South American plate during the Cenozoic. Almeida (1961) recognized five volcanogenic successions at Trindade in decreasing age: the Trindade Complex (TC) and the Desejado (DF), Morro Vermelho (MV), Valado (VF) and Paredão (PF) formations, composed of nephelinitic to phonolitic effusive-pyroclastic deposits, dykes and necks. Here, we present new 40 Ar/ 39 Ar ages and re-evaluation of available 40 K/ 40 Ar* data that, coupled with previous petrological information, allowed us to reconstruct the volcanic history of the island (i.e., 3.9–0.25 Ma) distinguishing near synchronous volcanic episodes and solving several stratigraphic uncertainties reported in the literature. The geochronological dataset show that the nephelinitic–phonolitic volcanism at Trindade was discontinuous throught time, being marked by periods of high volcanic activity and periods of quiescence, which suggests variable melt production and eruption rates. The peak of the volcanic activity occurred between 3.9 and 2.5 Ma (i.e., Lower Pliocene to Lower Pleistocene) that is represented by the TC, which correspond to the largest volume of volcanic deposits preserved in the island. The volcanic activity slows down progressively from 2.5 Ma to cease at ca. 1.6 Ma, period that is represented by the DF. This volcanism was followed by a quiescent period that lasted until the ephemeral nephelinitic volcanism of the MV (no age), VA (no age) and PF (ca. 0.25 Ma). Thus, the volcanic activity in Trindade was ceased completely at ca. 0.25 Ma, event registered in the uppermost volcanic deposits of the PF and that represents the last volcanic activity in the Brazilian territory. [ABSTRACT FROM AUTHOR]

Structure- and tectonic-related gas migration into Ordovician sandstone reservoirs and its impact on diagenesis history were reconstructed in two gas fields in the Sbaa Basin, in SW Algeria. This was accomplished by petrographical observations, fluid inclusion microthermometry and stable isotope geochemistry on quartz, dickite and carbonate cements and veins. Two successive phases of quartz cementation ( CQ1 and CQ2) occurred in the reservoirs. Two phase aqueous inclusions show an increase in temperatures and salinities from the first CQ1 diagenetic phase toward CQ2 in both fields. Microthermometric data on gas inclusions in quartz veins reveal the presence of an average of 92 ± 5 mole% of CH4 considering a CH4- CO2 system, which is similar to the present-day gas composition in the reservoirs. The presence of primary methane inclusions in early quartz overgrowths and in quartz and calcite veins suggests that hydrocarbon migration into the reservoir occurred synchronically with early quartz cementation in the sandstones located near the contact with the Silurian gas source rock at 100-140°C during the Late Carboniferous period and the late Hercynian episode fracturing at temperatures between 117 and 185°C, which increased in the NW-direction of the basin. During the fracture filling, three main types of fluids were identified with different salinities and formation temperatures. A supplementary phase of higher fluid temperature (up to 226°C) recorded in late quartz, and calcite veins is related to a Jurassic thermal event. The occurrence of dickite cements close to the Silurian base near the main fault areas in both fields is mainly correlated with the sandstones where the early gas was charged. It implies that dickite precipitation is related to acidic influx. Late carbonate cements and veins (calcite - siderite - ankerite and strontianite) occurred at the same depths resulting from the same groundwater precipitation. The absence of methane inclusions in calcite cements result from methane flushing by saline waters. [ABSTRACT FROM AUTHOR]

Summary: A detailed picture of the hydrogeochemistry of an alpine thermal area (La Léchère, French Alps) was drawn up from new and complete analyses of shallow and deep groundwater, including major and trace elements, gases, and stable (δ18O, δD, δ13C, δ34S) and radiogenic isotopes (14C), combined with data on the trace and stable isotope (δ18O, δ34S) contents of the aquifer rocks. Three different groundwater facies were recognized: (1) the shallow groundwater in the carbonate and crystalline rocks presents a poorly mineralized Ca–HCO3 facies; (2) the shallow groundwater from the evaporitic fault zone presents a mineralized Ca–Mg–SO4 facies; (3) and the deep thermal water presents a highly-mineralized Ca–Na–SO4 facies. The evolution from facies 2 to facies 3 is controlled by three phenomena: (i) the presence or absence of evaporitic deposits along the fault, which governs whether or not the water can interact with carbonates and dolomites, (ii) interactions with micaschists and Permo-Carboniferous rocks, which lead to Ca–Na exchanges and enrichment in some trace elements and gases (CO2, CH4 and H2S), and (iii) the long residence time (several thousand years) and high temperature (90°C) of the thermal water. [Copyright &y& Elsevier]

Some mantle-derived Kerguelen harzburgite and dunite xenoliths have bulk-rock and mineral trace element compositions that provide evidence of carbonatitic metasomatism similar to that described in some continental and other oceanic settings. Rare xenoliths contain carbonates that are highly enriched in rare earth elements (REE), interpreted to be quenched, evolved carbonatitic melts. One amphibole-bearing dunite mantle wall-rock containing carbonates in small interstitial pockets (100–500 μm across) has been studied in detail. Mg-bearing calcite (MgO: <1.4 wt.%, XCa 0.96) with unusually high REE abundances and strong light REE (LREE) enrichment occurs in the pockets and is sometimes associated with euhedral carbonates (dolomite and Mg-free calcite), mafic silicate glass (low in alkalis) and with fine grains of spinel, sulfides and magnesio–wüstite concentrated near the boundaries of the carbonate pockets. The unusual metasomatic mineral assemblage, together with the microstructural features and chemical composition of carbonates (with trace element contents similar to those of common carbonatite magmas), suggests that the pockets of Mg-bearing calcite represent quenched carbonate melts rather than crystal cumulates from carbonate-rich melts. The associated mafic silicate glass could represent the immiscible silicate fraction of an evolved fluid produced by the dissolution–percolation of the original carbonate melt in the dunitic matrix and subsequent unmixing as the xenoliths ascended to the surface. Clinopyroxene formed during the percolation event and is therefore inferred to be in chemical equilibrium with the carbonate melt. This allowed calculation of clinopyroxene/carbonate melt partition coefficients for a large set of trace elements at relatively low pressure (1 GPa). As a result, a significant pressure control on REE partitioning between carbonate melt and silicate minerals was observed. This study provides further evidence for the occurrence of carbonate melts and demonstrates that these melts can be preserved in hot oceanic uppermost mantle. [Copyright &y& Elsevier]