The Grader layered intrusion (Havre-Saint-Pierre Anorthosite, Quebec) and genesis of nelsonite and other Fe–Ti–P ores

The Grader layered intrusion is part of the Havre-Saint-Pierre anorthosite in the Grenville Province (Quebec, Canada). This intrusion has a basin-like morphology and contains significant resources of Fe–Ti–P in ilmenite and apatite. Outcropping lithologies are massive oxide alternating with anorthosite layers, banded ilmenite–apatite–plagioclase rocks and layered oxide apatite (gabbro-)norites. Drill cores provide evidence for stratigraphic variations of mineral and whole rock compositions controlled by fractional crystallization with the successive appearance of liquidus phases: plagioclase and ilmenite followed by apatite, then orthopyroxene together with magnetite, and finally clinopyroxene. This atypical sequence of crystallization resulted in the formation of plagioclase–ilmenite–apatite cumulates or “nelsonites” in plagioclase-free layers. Fine-grained ferrodiorites that cross-cut the cumulates are shown to be in equilibrium with the noritic rocks. The high TiO₂ and P₂O₅ contents of these assumed liquids explains the early saturation of ilmenite and apatite before Fe–Mg silicates, thus the nelsonites represent cumulates rather than crystallized Fe–Ti–P-rich immiscible melts. The location of the most evolved mineral and whole rock compositions several tens of meters below the top of the intrusion, forming a sandwich horizon, is consistent with crystallization both from the base and top of the intrusion. The concentrations of V and Cr in ilmenite display a single fractionation path for the different cumulus assemblages and define the cotectic proportion of ilmenite to 21 wt.%. This corresponds to bulk cotectic cumulates with ca. 8 wt.% TiO₂, which is significantly lower than what is commonly observed in the explored portion of the Grader intrusion. The proposed mechanism of ilmenite-enrichment is the lateral removal of plagioclase due to its relative buoyancy in the dense ferrodiorite melt. This plagioclase has probably accumulated in other portions of the intrusion or has not been distinguished from the host anorthosite.     

Origin of the giant Allard Lake ilmenite ore deposit (Canada) by fractional crystallization,multiple magma pulses and mixing

The late-Proterozoic Allard Lake ilmenite deposit is located in the Havre-Saint-Pierre anorthosite complex, part of the allochtonous polycyclic belt of the Grenville Province. Presently the world's largest Fe–Ti oxide deposit, it had a pre-mining amount in excess of 200 Mt at grades over 60 wt.% hemo-ilmenite. The main ore body is a funnel-shaped intrusion, measuring 1.03 × 1.10 km and 100–300 m-thick. Two smaller bodies are separated by faults and anorthosite. The ore is an ilmenite-rich norite (or ilmenitite) made up of hemo-ilmenite (Hem_22.6–29.4, 66.2 wt.% on average), andesine plagioclase (An_45–50), aluminous spinel and locally orthopyroxene. Whole-rock chemical compositions are controlled by the proportions of ilmenite and plagioclase ± orthopyroxene which supports the cumulate origin of the deposit. Ore-forming processes are further constrained by normal and reverse fractionation trends of Cr concentration in cumulus ilmenite that reveal multiple magma emplacements and alternating periods of fractional crystallization and magma mixing. Mixing of magmas produced hybrids located in the stability field of ilmenite resulted in periodic crystallization of ilmenite alone. The unsystematic differentiation trends in the Allard Lake deposit, arising from a succession of magma pulses, hybridisation, and the fractionation of hemo-ilmenite alone or together with plagioclase suggest that the deposit formed within a magma conduit. This dynamic emplacement mechanism associated with continuous gravity driven accumulation of Fe–Ti oxides and possibly plagioclase buoyancy in a fractionating ferrobasalt explains the huge concentration of hemo-ilmenite. The occurrence of sapphirine associated with aluminous spinel and high-alumina orthopyroxene (7.6–9.1 wt.% Al₂O₃) lacking exsolved plagioclase supports the involvement of a metamorphic overprint during the synchronous Ottawan orogeny, which is also responsible for strong textural equilibration and external granule of exsolved aluminous spinel due to slow cooling.     

The Corossol structure: A possible impact crater on the seafloor of the northwestern Gulf of St. Lawrence,Eastern Canada

We report on a 4.1 (±0.2) km diameter and 185 m deep circular submarine structure exposed on the seabed in >40 m water depths in the northwestern Gulf of St. Lawrence (Eastern Canada) from the analysis of high‐resolution multibeam bathymetric and seismic data. The presence of a circular form characterized by a central uplift and concentric rings resembles the morphology and geometry of complex meteorite impact structures. Also, other origins, such as kimberlites, intrusions, karsts, or diapirs, can be eliminated on geological criteria. A single 4 cm long breccia fragment recovered from the central uplift has numerous glassy droplets of fluorapatite composition, assumed to be impact melts, and a single quartz grain with planar intersection features thought to be shock‐induced planar deformation features (PDFs). The absolute age of this possible impact structure is unknown, but its geological setting indicates that it was formed long after the Mid‐Ordovician and before regional pre‐Quaternary sea‐level lowstands. Present results outline the need for further examination to confirm an impact origin and to precisely date the formation of the structure.     

Quantitative modeling of Sr,Ca, Rb,and K in the Bjerkrem-Sogndal layered lopolith (S.W. Norway)

A systematic approach with graphic techniques is used to establish a quantitative model of fractional crystallization process in igneous layered complexes. Modeling of the evolution of Sr-Ca in plagioclase and K-Rb in plagioclase and whole rock coming from the Bjerkrem-Sogndal layered lopolith (Rogaland-S.W. Norway) is taken as an example. The relationships in logarithmic coordinates can be decomposed in a succession of segments. This permits identifying the Rayleigh law as controlling the process. A step by step solution is used to determine the parameters of the model which lead to the adjustment of the calculated evolution to the observed trend. Evidence in favour of an open system crystallization of the cumulate rocks permits determining the equilibrium partition coefficients between the various minerals and the liquid. The mean cumulate corresponding to a phase of crystallization of the intrusion is determined by averaging the mineral composition of the rocks belonging to that phase. The concentrations of the major elements Ca and K are used instead of activities. The adopted plagioclase-magma partition coefficients are close to those measured in anorthositic rocks for the same range of plagioclase composition between megacrysts and liquid. For an anorthite content of respectively 50, 43 and 31, D _Sr ^plag is equal to 2.0, 2.3 and 3.9, D _K ^plag varies between 0.40 and 0.25, D _Rb ^plag is either constant (ca. 0.10) or increases from 0.12 to 0.25, D _Ca ^plag is supposed to remain at an approximately constant value of 1.48. The fraction of residual liquid in the intrusion is 0.47 at the end of the anorthositic-leuconoritic phase, and 0.21 at the end of the cpxnoritic phase.     

Magma influx and mixing in the Bjerkreim-Sokndal layered intrusion, South Norway: evidence from the boundary between two megacyclic units at Storeknuten

The Bjerkreim-Sokndal layered intrusion belongs to the Proterozoic anorthositic province in the Rogaland area of southern Norway. The northwestern part of the intrusion comprises a ca. 6 km-thick Layered Series made up of megacyclic units (MCU) arranged in a syncline; each megacyclic unit reflects the influx of fresh magma into the chamber. The boundary between megacyclic units III and IV has been studied in detail at Storeknuten on the southern flank of the syncline. The megacyclic units can be subdivided into a series of cumulate stratigraphic zones; the interval from the top of zone IIIe to the base of zone IVd is exposed in the Storeknuten area. Modally layered plagioclase-hypersthene-ilmenite-magnetite-augite-apatite cumulates belonging to zone IIIe are overlain by 30 m of massive plagioclase-rich rocks (commonly containing ilmenite and/or hypersthene) constituting zone IVa. The entry of cumulus olivine defines the base of zone IVb (dominantly plagioclase-olivine-ilmenite cumulates) which is about 100 m thick. Many of the olivines are partly or completely replaced by Ca-poor pyroxene/Fe---Ti oxide symplectites. This massive leucotroctolitic zone is overlain by modally layered, laminated plagioclase-hypersthene-ilmenite cumulates of zone IVc. The successive entry of magnetite, apatite (accompanied by Ca-rich pyroxene) and inverted pigeonite defines zones IVd, e and f respectively. The entry of K-feldspar (accompanied by Fe-rich olivine) defines the base of a jotunitic transition zone which passes upwards into mangerites and quartz mangerites.

There is a compositional regression through zone IVa. The upper part of zone IIIe has Ca-poor pyroxene with about En68, plagioclase with An44–48 and a Sr-isotope ratio of about 0.7062, while the base of zone IVb has olivine with Fo75 together with En78, An53 and 0.7050 respectively. Similar reversals are shown by the minor element compositions of plagioclase and Fe---Ti oxides. Sr-isotope ratios increase systematically up through zone IVb (reaching 0.7058 in zone IVd) while An% and Sr in plagioclase and Ni and Cr in Fe---Ti oxides decrease. Olivine compositions vary unsystematically and are believed to have changed their Fe:Mg ratios as a result of trapped liquid shift.

The magma residing in the chamber when the influx at the base of megacyclic unit IV took place was compositionally zoned, and assimilation of gneissic country rock at the roof had resulted in the Sr-isotope ratio increasing up through the magma column. The new magma had a Sr-isotope ratio of about 0.7050 while the resident magma had a ratio of 0.7062 at the floor, increasing upwards. The new magma mixed with the basal layer(s) of the compositionally zoned resident magma and crystallization of this hybrid magma during influx and mixing produced the compositional regression in zone IVa. When magma influx ceased, olivine-bearing rocks began to crystallize at the base of zone IVb. The leucotroctolites at the base of this zone are the most primitive rocks in the entire intrusion. The systematic increase in Sr-isotope ratios up through zone IVb resulted from progressive mixing between new and resident magma. This mixing either took place during magma influx or by the progressive mixing of overlying resident magma layers during crystallization.

Calculations based on geochemical modelling, the thickness of cumulate stratigraphy repeated and Sr-isotope ratios indicate that the new magma influx had a thickness of 350–500 m in the Storeknuten section and that the leucotroctolites of zone IVb represent about 20–30% crystallization of this influx.     

Post-collisional melting of crustal sources: constraints from geochronology,petrology and Sr,Nd isotope geochemistry of the Variscan Sichevita and Poniasca granitoid plutons (South Carpathians,Romania)

The Sichevita and Poniasca plutons belong to an alignment of granites cutting across the metamorphic basement of the Getic Nappe in the South Carpathians. The present work provides SHRIMP age data for the zircon population from a Poniasca biotite diorite and geochemical analyses (major and trace elements, Sr–Nd isotopes) of representative rock types from the two intrusions grading from biotite diorite to biotite K-feldspar porphyritic monzogranite. U–Pb zircon data yielded 311 ± 2 Ma for the intrusion of the biotite diorite. Granites are mostly high-K leucogranites, and biotite diorites are magnesian, and calcic to calc-alkaline. Sr, and Nd isotope and trace element data (REE, Th, Ta, Cr, Ba and Rb) permit distinguishing five different groups of rocks corresponding to several magma batches: the Poniasca biotite diorite (P₁) shows a clear crustal character while the Poniasca granite (P₂) is more juvenile. Conversely, Sichevita biotite diorite (S₁), and a granite (S₂*) are more juvenile than the other Sichevita granites (S₂). Geochemical modelling of major elements and REE suggests that fractional crystallization can account for variations within P₁ and S₁ groups. Dehydration melting of a number of protoliths may be the source of these magma batches. The Variscan basement, a subduction accretion wedge, could correspond to such a heterogeneous source. The intrusion of the Sichevita–Poniasca plutons took place in the final stages of the Variscan orogeny, as is the case for a series of European granites around 310 Ma ago, especially in Bulgaria and in Iberia, no Alleghenian granitoids (late Carboniferous—early Permian times) being known in the Getic nappe. The geodynamical environment of Sichevita–Poniasca was typically post-collisional of the Variscan orogenic phase.     

Shoshonitic liquid line of descent from diorite to granite: the Late Precambrian post-collisional Tismana pluton (South Carpathians, Romania)

The post-collision late-kinematic Tismana pluton belongs to the shoshonitic series. It is part of a Late Precambrian basement within the Alpine Danubian nappes of the South Carpathians (Romania). This pluton displays an exceptionally complete range of compositions from ultramafic to felsic rocks (granites). Widespread mingling/mixing relationships at all scales give rise to a variety of facies. A liquid line of descent from the diorites to the granites is reconstructed by considering the variation in major and trace elements (REE, Sr, Rb, Ba, Nb, Zr, Hf, Zn, V, Co, Cr, U, Th, Ga, Pb) from 33 selected samples as well as mineral/melt equilibrium relationships. The first step of fractional crystallization is the separation from a monzodioritic parent magma of a peridotitic cumulate similar to the ultramafic rock found in the massif. A possible contamination by lower crustal mafic component takes place at this stage. The second step marks the appearance of apatite and Fe–Ti oxide minerals as liquidus phases, and the third step, saturation of zircon. Mixing by hybridisation of magmas produced at different stages of the evolution along the liquid line of descent is also operating (endo-hybridisation). As depicted by Nd and Sr isotopes, fractional crystallization was combined to an important early contamination by a mafic lower crust in a deep-seated magma chamber and to a later and mild contamination by a felsic medium crust in an intermediate chamber. The mingling essentially occurred during the final emplacement in the high-level magma chamber. The monzodioritic parent magma, identified by major and trace element modelling, is shown by Sr and Nd isotopes to have its source in the lithospheric mantle or in a juvenile mafic lower crust derived from it. The necessarily recent enrichment in K₂O and associated elements of the lithospheric mantle is likely to be related to the preceding Pan-African subduction period. The partial melting of this newly formed deep source has to be linked to a major change in the thermal state of the plate.