In this paper the Late Pleistocene carbonatite-melilitite association (CMA) of Italy is evaluated in terms of the Plio-Quaternary extensional structures in the crust and mantle lithosphere. In particular, we briefly discuss the geometry and stress field of the host geological structures, the tectonic control on the eruptive style, the thickness and composition of the lithosphere. Some implications for a passive rift geodynamic environment are also considered.

We present petrographic and mineralogical data for 21 mantle xenoliths (12 lherzolites, 8 wehrlites and 1 composite) selected from a suite of more than 70 samples collected from the Monticchio Formation, Mt. Vulture volcano, southern Italy. The xenoliths are rounded, coarse- to porphyroclastic-textured, and very fresh, with the following equilibrated mineral assemblages; olivine (Fo90–92), orthopyroxene (∼En89, Wo2.0), clinopyroxene (Mg90–92, 3–6% Al2O3, 1–1.5% Cr2O3), and chrome-spinel (14–20% MgO, ∼30–40% Cr2O3). Many xenoliths contain partial melt glasses and accessory sulphide (pentlandite) Some contain primary mica (phlogopite with ∼4% FeO, 1.8% Cr2O3, 1.4–2.8% TiO2) with slightly zoned rims (Fe-, Ti-, Al-enriched). One contains relics of garnet (pyrope; Mg84). Secondary veins in several xenoliths contain carbonate with significant Sr levels (∼0.5–1.0% SrO), occasional apatite and scarce melanite, all typical of carbonatites and presumably related to the host magma (melilitite/carbonatite). Although amphibole is a common megacryst in the same volcanic units, no primary amphibole was found in the xenoliths themselves. Calculated pressures and temperatures using a range of geothermometers/barometers give values of 14–22 kbar and 1050–1150°C. In particular, the En-Sp and Di-Sp thermo/barometers (Mercier, 1980) show a good positive correlation between P and T. The Monticchio xenoliths lie on the high-T side of an ‘oceanic’ geotherm. The xenolith geotherm is hotter than general heat flow values in this region at the current day (50 mWm−2) but it compares well with the high-pressure end of a typical alkaline continental rift.

Intrusive calcite-carbonatite ejecta (sovite) in the lowermost tephra layers of the Mt. Vulture alkaline mafic-ultramafic volcano (Upper Pleistocene), is the first intrusive carbonatite sample from southern Europe. The sovite is of coarse granularity and shows some textural and mineralogical layering. It is mainly formed of calcite (up to 3.5 wt.% MgO, and 0.53 wt.% SrO), less dolomite (average 18.2 wt.% MgO, and up to 2.1 wt.% SrO), spinel (60 wt.% Al2O3, 26.5 wt.% MgO, 10.7 wt.% FeO) and olivine (Fo99). Perovskite and apatite have been found only as microlites. Intergranular vugs are scattered throughout the rocks and small composite inclusions occur in calcite.

The mineral chemistry, high temperature melt inclusions, overall isotopic compositions, and the REE distribution are consistent with a primary igneous origin. Compared with world average sovite compositions, the Vulture sovite has lower LILE and HFSE but Rb, Sr, Th and U are high. The REE abundance is typical of carbonatites, having an LREE/HREE value of ∼100. The δ13C (−4.8% SMOW) is in the range for mantle-derived carbonatites. The 143Nd/144Nd (0.512648±15) and 87Sr/86Sr (0.705978±10) ratios show close similarity between the sovite and the Vulture alkaline mafic rocks. The sovite ejecta are interpreted as a crystallization product of carbonate ultramafic liquid. In common with many shallow-level carbonatites from other localities, the recrystallization of rather pure Mg-calcite, the presence of dissolution vugs and the depletion of some HFSE and the relatively high δ18O values, suggest that a secondary process, such as hydrothermal leaching, affected the rock.

New Sr-Nd isotope data were obtained from Late Pleistocene carbonatite-kamafugite associations from the Umbria-Latium Ultra-Alkaline District of Italy (ULUD) with the aim of constraining their origin and possible mantle source(s). This is relevant to the origin and evolution of ultrapotassic (K/Na ≫2) and associated rocks generally, notably the occurrences from Ugandan kamafugites,Western Australian lamproites and South African orangeites. The selected ULUD samples yielded 87Sr/86Sr and 143Nd/144Nd ranging from 0.7100 to 0.7112 and from 0.5119 to 0.5121 respectively, similar to cratonic potassic volcanic rocks with higher Rb/Sr and lower Sm/Nd ratios than Bulk Earth. Silicate and carbonate fractions separated from melilitite are in isotopic equilibrium, supporting the view that they are cogenetic. The ULUD carbonatites yielded the highest radiogenic Sr so far reported for carbonatites. In contrast, sedimentary limestones from ULUD basement formations are lower in radiogenic Sr, i.e. 87Sr/86Sr = 0.70745–0.70735. The variation trend of ULUD isotopic compositions is similar to that reported for Ugandan kamafugites and Western Australian lamproites and overlaps the values for South African orangeites in the εSr-εNd diagram. A poor correlation between Sr/Nd and 87Sr/86Sr ratios in ULUD rocks is inconsistent with a mantle source generated by subduction-driven processes, while the negligible Sr and LREE in sedimentary limestones from the ULUD region fail to account for a hypothetical limestone assimilation process. The Nd model ages of 1.5–1.9 Ga have been inferred for a possible metasomatic event, allowing further radiogenic evolution of the source, a process which may have occurred in isolation until eruption time. While the origin of this component remains speculative, the Sr-Nd isotope trend is consistent with a simple mixing process involving an OIB-type mantle and a component with low εNd and high εSr.

A group of carbonate-rich tuffs are described from the Murumuli crater, Katwe-Kikorongo volcanic field, SW Uganda which contain abundant carbonatite pelletal lapilli, together with melilitite lapilli and a range of xenocrysts and lithic fragments including clinopyroxenites considered to be of mantle origin. The carbonatite lapilli consist essentially of Sr-bearing calcite and Mg-calcite which form quench-textured laths. The lapilli contain microphenocrysts of Ti-magnetite, perovskite, apatite, clinopyroxene, sanidine and altered prisms of melilite. A 7 cm long dolomite carbonatite bomb is described which displays a form typically assumed by lava clots erupted in a molten state. Chemical analyses of a tuff, the bomb and a range of minerals are presented. Carbonatite clearly played an important role in the Katwe-Kikorongo magmatism and it is suggested that carbonatite magma evolved from carbonate-bearing melilitite.

We have performed an experimental study of the influence of varying size and charge on cation partitioning between wollastonite and silicate-carbonate melt in the system CaCO3-SiO2. The experimental conditions (3 GPa, 1420°C) lie close to the wollastonite II tc/I tc phase boundary. Results for 1+, 2+, 3+ and 4+ partitioning show parabolic dependence of partition coefficients on ionic radius, which can be fitted to the elastic strain model of Blundy and Wood (1994), wherein partitioning is described using three parameters: site radius (r0), site elasticity (apparent Young's Modulus) and the ‘strain-free’ partition coefficient (D0) for an element with radius r0. The apparent Young's Modulus of the Ca site in wollastonite, obtained from modelling the 2+ partitioning data, is 99±3 GPa, similar to the bulk-crystal value for the polymorph wollastonite I tc. r0 decreases with increasing charge on the substituent cation, while D0 also shows an approximately parabolic dependence on charge, with a maximum for 2+ cations. Partition coefficients for divalent cations Zn, Co, Fe, Cd, Mn and Pb are lower than would be predicted from their ionic radii alone, indicating a preference for the melt. This may be a consequence either of cation-carbonate complexation in the melt, or of the more distorted nature of cation co-ordination environments in melts.

The Tertiary (22 Ma) carbonatite centre at Buru hill is located towards the eastern end of the Nyanza rift, western Kenya. As revealed by geochemistry and petrology of drill core samples, the Buru centre ranges in composition from surficial lateritized pyroclastic ferrocarbonatite to recrystallized ferruginous calciocarbonatite at depth.

Mineralogically the lateritic cover of the Buru carbonatite centre is characterized by goethite, hematite, psilomelane, baryte, fluorite with minor calcite, bastnäsite and significant amounts of monazite. A transitional middle zone consists of ferrocarbonatite grading into ferruginous calciocarbonatite containing siderite, baryte and fluorite, a greater proportion of bastnäsite replacing calcite, but lesser proportions of monazite. The lowermost zone consists of ferruginous calciocarbonatite containing intimate mixtures of siderite and calcite, with parisite and synchysite existing as major replacive components of calcite down to depths of 200 m. Synchysite appears to be the more stable form of lanthanide fluorcarbonate at depths below 150 m.

Oxygen isotope measurements on separated calcite concentrates from three of the Buru drill cores reveal significant deviations from primary igneous carbonatite values suggesting isotopic re-equilibration with meteoric water. Accepting that the African plate may have been stationery during the past 30 Ma, it is possible to use a modern sample of water from the Kenya rift to estimate the equilibration temperature at which Buru carbonatite in western Kenya has recrystallized. Temperatures of ∼60 to 90°C suggest that the hot-spring activity encouraged the formation of replacive lanthanide fluorcarbonate in ferruginous calciocarbonatite. Lanthanide enrichment also occurred closer to the surface within the pyroclastic ferrocarbonatite at temperatures of ∼40 to 60°C corresponding to waning stages of hot-spring activity. Further lanthanide enrichment redistribution and oxide formation occurred at surface supergene temperatures of ∼20 to 30°C forming the lateritized top of the carbonatite.

The wider implications from these carbonatite studies are the recognition that lanthanide enrichment can be preserved in volcanic-subvolcanic centres, and that lanthanide fluorcarbonates can replace calcite in response to circulating hot-spring waters. From a classification point of view, ferrocarbonatites should be seen as products of meteoric hydrothermal systems unrelated to carbonatite magmatic processes

Italian carbonatites form part of a suite with melilitites, normally an association characteristic of continental interiors; the perfect analogue of the Italian suite being the kamafugites (from the type area in SW Uganda, where the western branch of the East African Rift Zone cuts across the craton). The latter are commonly attributed to plume generation, whereas the Italian carbonatites, strung along the Appennine front, are usually linked to subduction. Evidently these two mechanisms are not essential, since neither can apply in both provinces. This conclusion is re-inforced by the related magmatism registered in both provinces in the Cretaceous. Phlogopite is ubiquitous in the mantle debris, and compositions from the two provinces overlap. Xenolithic phlogopites are distinct from cognate micas in the lavas, and from the carrier melt compositions, with similar distribution patterns in both suites. Kamafugitic magmas must be products of exceptional conditions, and added to the many near-identical magmatic features, the Italian and Ugandan volcanoes have sampled similar mantle conditions. Although the large scale geodynamic regimes are in total contrast, as are the deep mantle tomographic structures, the crucial common factor at the igneous province level is extensional tectonics. Extension, promoting release of volatiles (esp. CO2), is the vital trigger for this small volume, primary magmatism.

Pyrochlore group minerals have the general formula A16–xB16O48Z8–y·nH2O, with A mainly Na, Ca, Sr, REE, Th, U; B mainly Nb, Ta, Ti and Z being O, F, OH. In this study, pyrochlore specimens were examined from carbonatites at Argor, Carillon Dam, Chilwa Island, Fen, Lueshe, Oka, Mbeya, Meech Lake and Verity. Primary features include a background with little compositional variation, from core to rim, upon which are commonly superimposed narrow oscillatory zones, parallel to {111}. These zones are usually characterized by high Ta, in many cases coupled with U (here Argor, Meech Lake and Verity specimens), but Chilwa Island and Fen pyrochlores have little Ta and zonation is mainly by enrichment in Ce and Nb. Primary zonation may persist through high-temperature metamorphism (Meech Lake and Verity) and metamictization (Meech Lake). Oscillatory zones were generated by a disequilibrium system that cooled under tranquil conditions, signalling absence of magma turbulence and, in many cases, the end of crystal growth. Some fresh crystals (Oka, Fen) have no oscillatory zones, possibly the product of magma turbulence in space or time. Low-temperature effects may mimic those of primary high temperature and are especially characterized by replacement rims, pyrochlore-inpyrochlore veinlets and low A-ion totals (Carillon Dam, Lueshe, Myeba).

Physical and chemical properties of lustrous black mica from greisen veins at the Bom Futuro tin mine indicate that it is a relatively Fe-F-rich, Li-poor zinnwaldite with a composition near the midpoint of the polylithionite-siderophyllite join. On the basis of 22 oxygens, the mica contains 1.28–1.73 atoms of Li with an average octahedral occupancy of 5.61 and a filled interlayer site. Compositional variation reflects the net effects of alteration, coupled tetrahedral-octahedral substitutions, and petrogenetic factors related to vein fluid chemistry. Geological constraints and compositional data suggest the Bom Futuro zinnwaldite crystallized at T ≈ 500°, P ≈ 1–500 bar, and fO2 ≈ 10−23. High Mn-Zn concentrations in the Bom Futuro mica appear to be a feature of Fe-Li micas from evolved anorogenic granitic systems not shared by orogenic counterparts.

Cordierite+orthoamphibole (Crd+Oam)-bearing gneisses in the Cormacks Lake complex are regionally associated with metapelites containing prismatic sillimanite and K-feldspar, metabasites that locally contain metamorphic orthopyroxene, and other high-grade rocks in the Central Gneiss (Dashwoods) subzone, in the southwestern Newfoundland Appalachians. Retrograde features formed at the expense of the granulite-facies assemblages are ubiquitous. For example, in some migmatitic rocks, garnet is resorbed by Crd+Oam, and in metapelites, cordierite separates corroded garnet and sillimanite. Mineral thermobarometry suggests that, following granulite-facies metamorphism (T<785°C, P<7.5 kbar), retrogression occurred as the Cormacks Lake gneisses cooled through Mg-Fe diffusional blocking temperatures as they decompressed to a pressure of ∼3–4 kbar. Given the absence of Barrovian (or higher pressure) mineral assemblages in the metapelites, regional tectonic reconstructions involving the thrusting of a neighbouring terrane (Notre Dame subzone) over the Central Gneiss subzone appear to be supported only by the moderate pressure determined for the granulite facies event. Although scarcely discernible given re-equilibration effects and the imprecision of thermobarometers, subsequent decompression nonetheless had a marked impact on the mineralogy of the gneisses.

Detailed chemical analyses of different portions of laboratory-grown sector-zoned clinopyroxene been provided by Skulski et al. (1994). These data have been plotted here in PC-IR diagrams results indicate equilibrium crystallization and protosite control of element selection.

Lithium oxide contents of biotites in some Cornubian granites which have been studied by electron microprobe are estimated from their SiO2 and F contents using samples formerly analysed by XRF and wet chemistry. Better formulae which include Li show that most of these biotites are lithian siderophyllites/ lithian Al-annites although some Dartmoor samples, especially those in the basic microgranite (ME) enclaves, are closer to Al-annites. Dartmoor samples are slightly more basic and poorer in Alvi than those from the Isles of Scilly and Carnmenellis granites agreeing with the broad differences in the host rocks. Biotites in volcanic clasts from the New Red Sandstone are significantly different in having much higher Si, Ti, Mg, F and ΣY and lower Alvi contents than those of the granite biotites: these are lithian Fe-Mg biotites. These clasts are believed to be the volcanic equivalents of an evolving Dartmoor magma (Awad et al., 1996). If so, their biotites reflect the composition of the granite biotites at the time of magma emplacement. They subsequently evolved as they equilibrated with the cooling granite to give the present biotite compositions of the Dartmoor granites.

The 590±2 Ma Ben Vuirich Granite was intruded into late Proterozoic Dalradian rocks prior to the Grampian orogeny, during which it was affected by upper amphibolite facies regional metamorphism. Spotted cordierite and andalusite (chiastolite) hornfelses at the granite margin were altered to kyanite-and garnet-bearing assemblages during regional metamorphism. From the inferred mineralogy of the hornfels, together with the normative Qz–Ab–Or values of the granite and the application of a simple cooling model, we conclude that the country rocks immediately adjacent to the granite were hornfelsed at T = ∼600°C and P ≤ 2 kbar.

The hornfelsed rocks were subsequently metamorphosed during the regional D2 event to form an equilibrium mineral assemblage of muscovite + biotite + garnet + plagioclase + quartz + kyanite (after chiastolite). Garnet which grew during this event shows unusual reversed chemical zoning, with Ca increasing systematically from core to rim as Fe, Mg and Mn decrease. A study of element partitioning between coexisting phases in equilibrium (including zoned garnets), and use of an internally consistent thermodynamic dataset, suggest that isothermal (T = 577±42°C) compression (from P = 6.2±1.6 kbar to 9.0±1.9 kbar) occurred during crustal thickening. K-feldspar-plagioclase-quartz veinlets found in the hornfels close the granite contact are demonstrated to be of igneous origin.

To date, Fe-Li micas have been defined differently from other micas. The purpose of this paper is to reinterpret the actual Fe-Li mica series with new concepts of ‘essential replacement’ (the evolution direction) and ‘composition track’ (the sequence of mica varieties). Two hundred and fifty-eight analyses from the literature are used for this study in the form of eight data groups. The common compositional and substitutional characters of Fe-Li micas have been reinterpreted in light of principal component analysis and a geometric frame of ideal Fe2+-Al-Li micas in space with (Si, AlIV, AlVI, Fe2+, Li, ☐VI, K)-coordinates.

In our new interpretation, the actual Fe-Li micas are essentially neither AlIV- nor AlVI-, but ☐VI-constant. The actual Fe-Li micas are the weakest fluctuant relative to the Annite-Polylithionite-Trilithionite-Siderophyllite (APTE) plane. About 90% of variations of actual Fe-Li micas range from the trioctahedral trend described as a sequence along the segment K2Al2Fe5/22+Li☐1/2Si6Al2O20(OH,F)4+x[AlFe−22+Li3SiAl−2] (−1/3≤ x≤5/8). The substitution AlFe−22+Li3SiAl−2 (i.e. 2AlIV +4Fe2+ → 2Si +AlVI +3Li) is the main mechanism that keeps actual Fe-Li micas trioctahedral. More than 8% of variations arise from the dioctahedral trend involving created AlVI- and ☐VI-increasing replacement. The actual Fe-Li mica series comprises the composition trend from Fe2+-biotite to lepidolite. This series is not on the siderophyllite-polylithionite join, but can be expressed ideally as K2Al1+xFe4−4x2+Li1+3x(Si6+2xAl2−2x) O20(OH,F)4 (−1/3 ≤x≤1).