The Gross Brukkaros (Namibia) reflects a broad dome structure showing a crater-shaped depression with numerous peripheral beforsitic carbonatite dykes. These dykes frequently contain an extreme load of basement (Nama-group) xenoliths (> 60 vol.%) including shales, quartzites, granites and gneisses. While xenoliths of exposed country rocks (mainly shales) show an angular habit, a pronounced rounding of xenoliths from other lithologies proves a wide transport and strong abrasion. This consumption of xenolithic material may result in remarkable contamination of the carbonatitic magma. MicroXRF mapping and optical microscopy provides first evidence that the corrosion and alteration of crustal xenoliths is controlled primarily by the xenoliths´ mineralogy and geochemistry. While some xenoliths exhibit distinct zoning reflecting a progressive leaching, others appear to be relatively inert. This proves the Gross Brukkaros of being an ideal natural laboratory to study the influence of crustal contamination in the carbonatitic system, particularly at the subvolcanic-volcanic depth transition. On the other hand, cross-cutting carbonatite dykes generated diatremes almost completely composed of quartz. A closer proximity to the diatreme yields an increase of the Si content in the dykes. In some cases, dykes occur with an extremely high proportion of microscopic xenolith fragments (>95 vol.%) and only subordinate proportions of carbonate. This indicates evaporation of the carbonatite melt during eruption, while the inherent Si remains as a residue along with the xenolith fragments and is precipitated in the diatreme breccia. Combined with C and O isotope systematics, carbonate crystallization is suspected to have proceeded under super-cooled conditions at ~150 °C.

The Chico Sill Complex (Northeast New Mexico) is the result of magmatic episodes from ~37 Ma to 20 Ma and produced a diverse and compositionally discontinuous suite of mostly intrusive silica-saturated and silica-undersaturated rocks. The Chico Phonolite was emplaced ~ 26 to 20 Ma in dikes and large sills. Sills of different composition may be stacked 2 and 3 thick. They bear no chemical affinity to the bulk of other rocks in the complex based on normalized trace element diagrams. Candidates for parent melts are scarce. At least two distinct trends are noted in Zr-La Space. A higher-Zr trend includes dikes and three sills and may represent evolution of a primary phonolite melt. The most-evolved sill (Point of Rocks Mesa) is the last-gasp of phonolite magmatism and likely the companion immiscible silicate for a calciocarbonatite dike 10 km distant.

Calciocarbonatite is a miniscule portion of the complex (outcrop limited to a few hundreds of m²⁾. The carbonate mineral is impure calcite (Mn>Fe>Sr>Mg>>Ba) in matrix goethite. Other minerals present include barite, pyrite, and REE minerals containing Ca and Ca-Ti next to the calcite. Normalized Ba, Th, REE, Y and Sr show 100 times enrichment in the carbonatite. Mineralogy, texture, and O-C isotopes suggest that the original carbonatite melt may have been more sodic and experienced alteration similar to that of lavas at Oldoinyo Lengi. Owing to the distance between outcrops, the separation of the melts (and phonolite evolution) occurred at much greater depth.

Nephelinites are strongly SiO₂-undersaturated volcanic rocks that are often associated with phonolites and carbonatites. In the Gregory Rift in East Africa several major nephelinitic-phonolitic volcanoes occur, with some of them being associated with carbonatitic rocks (e.g., Oldoinyo Lengai, Kerimas, Mosonik, Shombole, Meru), while others lack carbonatites (e.g, Sadiman, Essimingor, Burko). We characterize the magmatic evolution of the Burko volcano and compare our results with published data from spatially associated nephelinite-phonolite±carbonatite associations in the Gregory Rift and elsewhere.

Overall, nephelinites show mineralogical differences, are variably evolved (in terms of X_Mg, LILE and HFSE), and in some cases peralkaline (Na+K/Al >1) nephelinites do occur. Besides nepheline, clinopyroxene and apatite, garnet, magnetite, perovskite and titanite are magmatic phases in most cases. However, magmatic ne-cpx-grt-ttn assemblages can be distinguished from those with ne-cpx-mag-prv. Other phases, such as wollastonite, melilite, combeite, aenigmatite, sodalite and others are restricted to some occurrences and resemble different geochemical flavors of nephelinites, different crystallization conditions, variable differentiation stages and different levels of peralkalinity. Redox- and silica activity-dependent phase equilibria allow for constraining and comparing the magmatic evolution of the different localities by combining textural with mineral chemical data.

In general, high redox conditions above FMQ and peralkalinity seem to favor the formation of carbonatites. However, in several cases that meet these conditions, no carbonatites are exposed and worldwide, carbonatites are often associated with nephelinites that are not peralkaline. We discuss the potential for nephelinites to exsolve carbonate-rich liquids based on a petrological and geochemical comparison of different occurrences.

Foidites and melilitites are strongly SiO₂-undersaturated rocks that form by extremely low degrees of partial melting of the metasomatically overprinted lithospheric mantle. In Central Europe, they occur in volcanic fields, dike swarms or as isolated stocks and diatremes.

Our detailed study on foidites from SW Germany indicates two distinct age groups with marked differences in mineralogy and mineral chemistry: Based on in-situ U‑Pb age data (apatite, perovskite, zircon) a Miocene cohort (~ 9–19 Ma) of predominantly olivine melilitites and melilite-bearing nephelinites can be distinguished from a much older Upper Cretaceous to Lower Eocene group (~ 48–68 Ma) of melilite-free nephelinites and nepheline basanites. This contrasts with previous K-Ar whole-rock and mineral data suggesting continuous magmatism between 90 and 6 Ma.

The older group is characterized by the frequent occurrence of green core pyroxenes, hydroxyapatite, and minor feldspar, whereas the younger group contains melilite, late magmatic fluorapatite, Ba- and F-rich mica and occasionally perovskite, but no feldspar. It crops out in the Freiburger Bucht and the Bonndorf Graben, the Vosges (France), the Odenwald and Kraichgau region, in the Taunus and the Lower Main Plain, whereas the younger group is represented by occurrences in the Hegau, the Urach region and the Central Upper Rhine Graben including the Kaiserstuhl.

As part of the Central European Volcanic Province, the spatial distribution and age of these rocks reflect regional tectonic events, while the petrologic contrasts between the two age groups indicate heterogeneous crystallization conditions and/or magma source variations such as different formation depths.

The Gardiner (E-Greenland) and Kovdor (Russia) alkaline complexes display a similar succession of rock types comprising dunites-pyroxenites, ijolite series rocks, melilitolites and carbonatites. Although similar melanephelinitic parental magmas are suggested for both complexes, they display enrichment in HFSE at strikingly different evolutionary stages: At Kovdor, melilitolites are barren but carbonatites are mineralized with HFSE. In contrast, melilitolites at Gardiner contain ore-grade accumulates of perovskite having wt.%-level contents of Nb, Ta and REE, while associated carbonatites are barren. Previous studies suggested that HFSE-poor carbonatites at Gardiner were formed by liquid immiscibility while Kovdor carbonatites result from fractional crystallization and retained high HFSE contents. These two evolutionary trends were explained by a different CO₂-dependent stability of melilite vs. clinopyroxene+nepheline+calcite during the ijolite stage [1]. However, it is poorly investigated how the HFSE budget is affected by the crystallization of Ti-phases during different stages of the magmatic evolution, which are stabilized depending on magma composition (i.e. aTiO₂, aSiO₂) but also on intensive parameters such as P, T, and fO₂ [2]. Preliminary results suggest that, in contrast to Kovdor, magmas at Gardiner had physiochemical conditions which favoured abundant crystallization of Ti-phases along with co-precipitation of HFSE earlier in the sequence. This is supported by (1) pyroxenites with abundant Ti-magnetite and ilmenite, (2) titanite-rich ijolites and (3) perovskite-rich melilitolites. Possibly, Ti-rich melts reflect a distinct mantle regime beneath E-Greenland, which also produced anomalously Ti-enriched flood basalts ~6-10 Ma before.

[1] Veksler et al. (1998) J.Pet. 39, 2015-2031; [2] Marks et al. (2008) CG 257, 153-172

Halogen (F, Cl and Br), S and δ³⁷Cl variations within grains of Cl-rich minerals sodalite and eudialyte from peralkaline rocks of the Ilímaussaq intrusion were determined using Secondary Ion Mass Spectrometry (SIMS). Samples show either sodalite and eudialyte in direct contact, or sodalite/eudialyte embedded in Cl-free minerals (nepheline, feldspar). Comparing samples allows deciphering potential halogen and S exchange between these minerals during rock cooling.

Results suggest that sodalite (7 wt%) and eudialyte (1.2 wt%) have remarkably constant Cl concentrations. In samples with adjacent sodalite and eudialyte F increases at sodalite boundaries and decreases at eudialyte boundaries. In sodalite not in contact with eudialyte F concentrates at the edges, something obscured by F-rich inclusions. In eudialyte not in contact with sodalite F is constant with no variations at the edges. Br is also constant in eudialyte, but in sodalite its concentration decreases towards the edges. S also is constant in eudialyte, and concentrates significantly at the edges of sodalite, especially strongly if sodalite contacts eudialyte. In hydrothermal eudialyte Cl is low at the edge with higher concentrations away from the edge. Br and S correlate with Cl while little F variation is observed.

δ³⁷Cl in eudialyte is higher than in adjacent sodalite. Within individual grains δ³⁷Cl is higher at the edges than in the centre. Between sodalite grains δ³⁷Cl can vary a few tenths of a permille, while between eudialyte grains, variations can even be higher. In hydrothermal eudialyte δ³⁷Cl increases significantly at the grain boundary with values up to +3.5‰.