Journal of Petrology Advance Access originally published online on March 14, 2006
Journal of Petrology 2006 47(7):1261-1315; doi:10.1093/petrology/egl008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Genesis of Ultramafic Lamprophyres and Carbonatites at Aillik Bay, Labrador: a Consequence of Incipient Lithospheric Thinning beneath the North Atlantic Craton
1 MAX-PLANCK-INSTITUT FÜR CHEMIE POSTFACH 3060, 55020 MAINZ, GERMANY
2 INSTITUT FÜR GEOWISSENSCHAFTEN, UNIVERSITÄT MAINZ BECHERWEG 21, 55099 MAINZ, GERMANY
3 DEPARTMENT OF EARTH SCIENCES, MEMORIAL UNIVERSITY ST. JOHN'S, NEWFOUNDLAND, CANADA A1B 3X5
4 DEPARTMENT OF EARTH AND ATMOSPHERIC SCIENCES, UNIVERSITY OF ALBERTA EDMONTON, ALBERTA, CANADA T6G 2E3
5 GEOLOGICAL SURVEY OF CANADA OTTAWA, ONTARIO, CANADA K1A 0E8
6 GEOFORSCHUNGSZENTRUM POTSDAM TELEGRAFENBERG, 14473 POTSDAM, GERMANY
7 GEOCHEMISCHES INSTITUT, UNIVERSITÄT GÖTTINGEN GOLDSCHMIDTSTRASSE 1, 37077 GÖTTINGEN, GERMANY
RECEIVED JULY 16, 2004; ACCEPTED FEBRUARY 2, 2006
 |
ABSTRACT |
|---|
TOP
ABSTRACT
INTRODUCTIONNumerous dykes of ultramafic lamprophyre (aillikite, mela-aillikite, damtjernite) and subordinate dolomite-bearing carbonatite with U–Pb perovskite emplacement ages of
590–555 Ma occur in the vicinity of Aillik Bay, coastal Labrador. The ultramafic lamprophyres principally consist of olivine and phlogopite phenocrysts in a carbonate- or clinopyroxene-dominated groundmass. Ti-rich primary garnet (kimzeyite and Ti-andradite) typically occurs at the aillikite type locality and is considered diagnostic for ultramafic lamprophyre–carbonatite suites. Titanian aluminous phlogopite and clinopyroxene, as well as comparatively Al-enriched but Cr–Mg-poor spinel (Cr-number < 0.85), are compositionally distinct from analogous minerals in kimberlites, orangeites and olivine lamproites, indicating different magma geneses. The Aillik Bay ultramafic lamprophyres and carbonatites have variable but overlapping 87Sr/86Sri ratios (0·70369–0·70662) and show a narrow range in initial 
Nd (+0·1 to +1·9) implying that they are related to a common type of parental magma with variable isotopic characteristics. Aillikite is closest to this primary magma composition in terms of MgO (15–20 wt %) and Ni (200–574 ppm) content; the abundant groundmass carbonate has 
13CPDB between –5·7 and –5
, similar to primary mantle-derived carbonates, and 18OSMOW from 9·4 to 11·6. Extensive melting of a garnet peridotite source region containing carbonate- and phlogopite-rich veins at 4–7 GPa triggered by enhanced lithospheric extension can account for the volatile-bearing, potassic, incompatible element enriched and MgO-rich nature of the proto-aillikite magma. It is argued that low-degree potassic silicate to carbonatitic melts from upwelling asthenosphere infiltrated the cold base of the stretched lithosphere and solidified as veins, thereby crystallizing calcite and phlogopite that were not in equilibrium with peridotite. Continued Late Neoproterozoic lithospheric thinning, with progressive upwelling of the asthenosphere beneath a developing rift branch in this part of the North Atlantic craton, caused further veining and successive remelting of veins plus volatile-fluxed melting of the host fertile garnet peridotite, giving rise to long-lasting hybrid ultramafic lamprophyre magma production in conjunction with the break-up of the Rodinia supercontinent. Proto-aillikite magma reached the surface only after coating the uppermost mantle conduits with glimmeritic material, which caused minor alkali loss. At intrusion level, carbonate separation from this aillikite magma resulted in fractionated dolomite-bearing carbonatites (13CPDB –3·7 to –2·7) and carbonate-poor mela-aillikite residues. Damtjernites may be explained by liquid exsolution from alkali-rich proto-aillikite magma batches that moved through previously reaction-lined conduits at uppermost mantle depths. KEY WORDS: liquid immiscibility; mantle-derived magmas; metasomatism, Sr–Nd isotopes; U–Pb geochronology
|
INTRODUCTION |
|---|
The ultramafic lamprophyres (UML; Rock, 1986
) are a widely recognized group of alkaline igneous rocks associated with continental extension; however, their origin is poorly understood. Although they are volumetrically minor components of continental magmatism, they are of fundamental significance for our understanding of deep melting events during the initial stages in the development of continental rifts. UML typically occur as dyke swarms and in central complexes (Rock, 1991), but their genesis has commonly been discussed as though they are varieties of kimberlite (Dawson, 1971; Griffin & Taylor, 1975; Raeside & Helmstaedt, 1982; Alibert & Albarède, 1988; Dalton & Presnall, 1998) mainly as a result of a similar macroscopic appearance and often problematic identification within existing classification schemes (Tappe et al., 2005a). However, compositional differences and the lack of spatial coexistence between contemporaneous UML and kimberlites (Rock, 1991; Mitchell, 1995) suggest that they are derived from distinct magma types. The occurrence of UML is largely confined to regions of lithospheric extension and they are commonly associated with carbonatite magmatism. Additionally, UML magmatism forms the earliest igneous activity in some flood basalt provinces (Queen et al., 1996; Leat et al., 2000; Riley et al., 2003) and also occurs on oceanic plateaux (Nixon et al., 1980; Neal & Davidson, 1989). In contrast, kimberlites occur exclusively within areas of stable Archean cratons or in the surrounding Proterozoic mobile belts (Mitchell, 1986; Janse & Sheahan, 1995). As for kimberlites, UML magmas may contain diamonds (Hamilton, 1992; Mitchell et al., 1999; Digonnet et al., 2000; Birkett et al., 2004), indicating that the depth of melting can be in excess of 150 km (>5 GPa), and, thus, may not be the crucial petrogenetic difference.
Large areas of Labrador, adjacent northeastern Quebec and western Greenland consist of Archean blocks surrounded by Paleoproterozoic mobile belts (Fig. 1) stabilized at 1900–1700 Ma (Wardle & Hall, 2002). Continental extension affected this cratonic area repeatedly during Mesoproterozoic (1350–1140 Ma; Romer et al., 1995; Upton et al., 2003), Neoproterozoic (620–550 Ma; Gower et al., 1986; Kamo et al., 1989; Larsen & Rex, 1992; Murthy et al., 1992) and Mesozoic times (Hansen, 1980; Keen et al., 1994; Chian et al., 1995; Larsen et al., 1999; Srivastava & Roest, 1999), eventually causing the break-up of the North Atlantic craton and opening of the Labrador Sea at 60 Ma (Chalmers & Laursen, 1995; Chalmers & Pulvertaft, 2001).
|
All of these extensional episodes were accompanied by volatile-rich alkaline igneous activity (Larsen & Rex, 1992
), but the most productive, in terms of UML magma generation, was the Late Neoproterozoic episode, which was distally associated with initiation of the Iapetus Ocean (Tappe et al., 2004). Neoproterozoic UML dykes related to this lithospheric stretching occur in the Sisimiut–Sarfartoq–Maniitsoq areas of western Greenland (Scott, 1981; Thy et al., 1987; Larsen & Rex, 1992; Mitchell et al., 1999; Heaman, 2005), the Torngat Mountains in northern Quebec and Labrador (Digonnet et al., 2000; Tappe et al., 2004), the Otish Mountains region in central–north Quebec (Heaman et al., 2004), single occurrences along the northern Labrador coast at Hebron, Saglek, Eclipse Harbour, Iselin Harbour and Killinek Island (Tappe et al., 2005b), and Aillik Bay in central-east Labrador (Malpas et al., 1986; Foley, 1989a). Here, we report the results of a petrological and geochemical study, combined with U–Pb perovskite and 40Ar/39Ar phlogopite age determinations, on a diverse suite of Neoproterozoic UML and associated carbonatites from the Aillik Bay area on the Labrador Sea coast. We discuss whether the large compositional diversity reflects mantle source heterogeneity or variability in the melting process, or relates to modification of a common parental UML magma by low-pressure processes, such as liquid immiscibility and devolatilization. Additionally, we specify and emphasize fundamental differences in the characteristics and genesis of UML and other ultramafic magma types, such as kimberlites. Late Neoproterozoic UML magma production occurred throughout the North Atlantic region attendant with widespread lithospheric stretching, thinning and the eventual break-up of the supercontinent Rodinia.
|
GEOLOGY OF THE AILLIK BAY AREA |
|---|
The southern North Atlantic craton margin
The Aillik Bay area is situated within the Paleoproterozoic Makkovik orogen at the southern edge of the Archean North Atlantic craton (NAC; Fig. 1). Reworked Archean orthogneisses (protolith ages 3260–2800 Ma) equivalent to the adjacent NAC are exposed along the western margin of the Makkovik Province (Fig. 2), whereas a juvenile high-grade magmatic arc crust dominates the central (Aillik Group) and eastern part close to the Grenville deformation front (Culshaw et al., 2000a
, 2000b; Sinclair et al., 2002). These supracrustal units formed during a sequence of subduction and accretion events between 1900 and 1700 Ma (Makkovikian Orogeny) and were later detached from the basement and thrust onto the edge of the NAC (Ketchum et al., 2002; Wardle & Hall, 2002). Seismic data and the unradiogenic initial Nd isotope composition of widespread post-orogenic granites (1720–1650 Ma; e.g. Strawberry granite in Fig. 2) clearly indicate that the western part of the Makkovik orogen, including the Aillik Bay area (Fig. 2), is underlain by the Archean crust of the NAC (Kerr & Fryer, 1994; Hall et al., 1995; Kerr & Wardle, 1997; Kerr et al., 1997).
|
Widespread lithospheric thinning occurred throughout eastern North America along the former Laurentian margin during the Late Neoproterozoic (Bond et al., 1984
; Kamo et al., 1995; Torsvik et al., 1996; Cawood et al., 2001; Puffer, 2002), resulting in continental break-up and subsequent opening of the Iapetus Ocean at 600 Ma. In central Labrador, this episode of continental stretching is recorded by remnant graben structures forming the eastward continuation of the prominent St. Lawrence valley rift system (Gower et al., 1986; Murthy et al., 1992).
Alkaline magmatism and previous age constraints
Late Neoproterozoic UML and carbonatite dykes occur in an area at least 30 km by 30 km around Aillik Bay (Fig. 2; Appendix A). These dykes are narrow (up to 3 m wide) and dominantly steeply dipping; subordinate flat-lying sheets also occur. Recognized UML types are aillikite, mela-aillikite and damtjernite, following the scheme devised by Tappe et al. (2005a). The subvertical dykes are roughly north–south-oriented and appear to converge towards a focus in the Labrador Sea (Fig. 2). Flow banding, back-veining and internal chill-bands are often seen, whereas fluidized globular ‘autolithic’ segregations are rare. Individual members of this dyke swarm cross-cut each other in a rather arbitrary manner. Some aillikite sheets or dykes grade laterally into carbonatite and/or mela-aillikite (Fig. 3). A weighted K–Ar mica age of 570 Ma, obtained for a poorly described ultramafic dyke rock from Aillik Bay (Leech et al., 1962), provided the only age constraint for the carbonate-rich magmatism when this study was initiated.
|
UML magmatism was preceded by Mesoproterozoic ultrapotassic magma production (
1374 Ma; Tappe et al., in preparation), represented by subvertical, 0·2–2 m wide, fine- to medium-grained olivine lamproite dykes within the same area. The youngest record of alkaline igneous activity around Aillik Bay is a Mesozoic suite of melilitite, nephelinite and basanite dykes (142 Ma; Tappe et al., in preparation), which appears to be related to the poorly exposed Ford's Bight alkaline intrusion (King & McMillan, 1975). |
GEOCHRONOLOGY |
|---|
U–Pb dating of ultramafic lamprophyres
Aillikite, mela-aillikite and damtjernite dykes were selected from all parts of the Aillik Bay area for U–Pb perovskite dating (Fig. 2). Analytical details can be found in Appendix B. Results are reported in Table 1 and displayed in concordia diagrams in Fig. 4. Two perovskite fractions of aillikite dyke ST123 from the east shore of Kaipokok Bay yielded similar 206Pb/238U dates of 560·7 ± 2·4 and 564·5 ± 3·0 Ma, respectively. Hence, a weighted average 206Pb/238U date of 562·2 ± 1·9 Ma is considered the best age estimate for emplacement of dyke ST123. A similar 206Pb/238U age of 569·2 ± 1·8 Ma was obtained from mela-aillikite dyke ST114A exposed on the west shore of Aillik Bay. The emplacement age of aillikite dyke ST228 from the southern shore of Makkovik Bay was determined to be 576·4 ± 6·5 Ma (weighted average 206Pb/238U date of two perovskite fractions 574·4 ± 1·8 Ma and 578·6 ± 2·0 Ma). The emplacement age of aillikite dyke ST220II from West Turnavik Island is 589·6 ± 1·3 Ma (weighted average of 589·4 ± 1·4 and 590·5 ± 2·8 Ma).
|
|
The youngest perovskite ages obtained from damtjernite dykes are 555·0 ± 1·8 Ma (ST211A; Main Turnavik Island) and 563·9 ± 2·5 Ma (ST256; east shore of Makkovik Bay). Perovskites from damtjernite dyke ST174 (Pigeon Island) yielded a 206Pb/238U date of 574·6 ± 1·6 Ma. Strikingly similar weighted average ages of 581·9 ± 2·3 Ma (582·5 ± 4·8 and 581·9 ± 2·6 Ma) and 582·5 ± 2·1 Ma (582·8 ± 3 and 582·1 ± 2·2 Ma) were obtained from the damtjernites ST140A (east shore of Aillik Bay) and ST188A (Red Island), respectively.
Taken together, the high-precision 206Pb/238U perovskite dates for four individual aillikite/mela-aillikite dykes and five damtjernite dykes cover a similar age range between 562–590 Ma and 555–583 Ma, respectively (Fig. 4). Hence, aillikite and damtjernite magmatism can be considered coeval over 30–35 Myr during Late Neoproterozoic extension at a craton margin. Although no carbonatites have been dated, their close association with the various dated UML types implies contemporaneous emplacement.
40Ar/39Ar dating of cognate inclusion ST162I
A clinopyroxene–phlogopite inclusion recovered from aillikite dyke ST162 on the west shore of Aillik Bay (Fig. 2) yielded a phlogopite 40Ar/39Ar plateau age of 573·3 ± 3·3 Ma (Table 2), which falls within the U–Pb perovskite age range of the four dated aillikite dykes.
|
The plateau was calculated over 10 consecutive steps, which contained 98% of the total released 39Ar. The gas release spectrum is shown in Fig. B1 (Appendix B). The 3·5%, 3·9% and 4·2% laser power steps significantly overlap the plateau but have slightly older apparent ages, indicating a potential presence of excess argon. However, on an inverse isochron diagram, a regression through these three data points and the seven others included in the plateau age calculation passes through a 40Ar/36Ar value of 433·1 ± 257·7, which is within error of the atmospheric value (295·5). As the three apparently older steps have no significant effect on the overall age, they have been included in the plateau age calculation.
|
PETROGRAPHY |
|---|
Rock types of the Neoproterozoic Aillik Bay UML suite include carbonatite, aillikite, mela-aillikite and damtjernite, listed in order of decreasing carbonate content. Additionally, aillikite dykes host a wide variety of micaceous cognate inclusions (Fig. 5). Modal mineral abundances are listed in Table 3 and mineral compositional data are given in Tables 4–11 (extended tables can be downloaded from the Journal of Petrology website at http://www.petrology.oupjournals.org as Electronic Appendix 1). It should be noted that descriptions of the following accessory minerals are solely provided as Electronic Appendix 2: ilmenite, rutile, perovskite, titanite, apatite, alkali feldspar, feldspathoids, pectolite.
|
|
|
|
|
|
|
|
|
|
Carbonatite dykes
Two distinct types of carbonatite can be distinguished: (1) a dolomite carbonatite devoid of any mafic silicates; (2) a mixed dolomite–calcite carbonatite containing minor amounts of clinopyroxene, phlogopite and olivine crystals. The dolomite carbonatite mainly consists of a mosaic of equigranular Fe-rich dolomite crystals (100–300 µm). Hydroxy-fluorapatite forms abundant euhedral microphenocrysts (50–150 µm). Interstices may be filled by barite, quartz, alkali feldspar (orthoclase and albite) and/or tiny rare earth element (REE)-carbonate crystals. Large rutile grains commonly occur (50–100 µm), whereas opaque phases including magnetite are comparatively rare.
The dolomite–calcite carbonatites exhibit a granular to interlocking texture dominated by calcite grains and laths (150–300 µm). Calcite coexists with subordinate laths of Fe-rich dolomite (Fig. 6a). Zoned phlogopite plates (up to 0·5 mm) and olivine grains (up to 1·0 mm; replaced by carbonates) are observed, suggesting gradation into aillikites. However, the presence of diopside-rich clinopyroxene phenocrysts (up to 1·2 mm) contrasts with the aillikites. Fresh rutile grains and apatite prisms (up to 0·4 and 1·0 mm, respectively) are abundant, as are opaque oxides. Secondary interstitial barite and/or fluorite may occur.
|
Aillikite dykes
Aillikites are texturally heterogeneous (e.g. Fig. 5a and b); some exhibit an inequigranular texture with olivine and phlogopite macrocrysts up to 7 mm in diameter, whereas the majority are weakly inequigranular and have a porphyritic texture (Fig. 5b). Porphyritic aillikites are characterized by phenocrysts of euhedral to subhedral olivine (0·6–1·3 mm), phlogopite (0·25–0·5 mm), apatite and magnetite (0·2–0·4 mm) in a carbonate matrix. Mela-aillikites are distinguished from aillikites in containing more mafic silicate phases (>70 vol. %) and less carbonate (<10 vol. %). The change in the modal mineral proportions is gradational from aillikite to mela-aillikite. End-member mela-aillikite contains abundant clinopyroxene prisms in the groundmass (100–300 µm; Fig. 5c), which are rare in aillikite. Both rock types carry microphenocrysts of olivine (0·25–0·5 mm), phlogopite (<0·25 mm), apatite, opaque oxides (dominantly titanomagnetite and Mg-rich ilmenite) and perovskite or rutile (50–200 µm). Primary kimzeyitic garnet typically occurs (<100 µm).
A rare textural variety occurs locally in otherwise uniformly textured aillikite dykes and consists of globular aillikite segregations ‘cemented’ by primary calcite laths. The globular segregations are made up of olivine and/or glimmerite kernels surrounded by concentrically arranged fine-grained aillikite matrix (Fig. 5a). Their strong resemblance to nucleated autoliths suggests an origin by fluidization of partly solidified early magma fractions as a result of local near-surface devolatilization.
Damtjernite dykes
Damtjernites are medium- to fine-grained, porphyritic to intergranular rocks (Fig. 5d) containing rare macrocrysts (up to 2·0 cm) of virtually Cr-free diopside-rich clinopyroxene and/or Cr-free titanian aluminous phlogopite. Modal layering, internal chill zones, bounded felsic segregations, flow-alignment and rotation structures are common macroscopic features of these rocks, which were called ‘sannaites’ by Foley (1984) and Malpas et al. (1986), but have been renamed here following Tappe et al. (2005a).
The phenocryst assemblage of the damtjernites consists of olivine (up to 1 mm), phlogopite (up to 5 mm), rare clinopyroxene (250–800 µm) and apatite (up to 1 mm). The modal abundance of euhedral to subhedral olivine phenocrysts may vary even within dykes from 20 vol. % to only a few crystals. Phlogopite forms large plates typically enclosing clinopyroxene and apatite needles. The groundmass mainly consists of a mesh made up of clinopyroxene and apatite needles (up to 200 µm) and biotite flakes resembling the late rims on phlogopite. Sr-calcite, alkali feldspar (almost pure orthoclase and albite) and nepheline occur in variable but small modal proportions interstitial to the mica and clinopyroxene of the groundmass (Fig. 5d). Additional rare felsic phases are analcime and sodalite. Pectolite is observed as a fibrous replacement product of groundmass clinopyroxene or as primary crystals in interstices. Olivine is absent in the groundmass. Abundant titanomagnetite grains, ilmenite laths, and rutile or perovskite crystals are the principal groundmass oxide phases. Perovskite relicts may be enclosed by Zr-rich titanite crystals recording fluctuations in silica activity during magma evolution or slow cooling. Schorlomite and/or melanite garnet occurs rarely in the groundmass in association with perovskite (Fig. 6e). Felsic segregations (orthoclase, albite, nepheline, analcime, sodalite, calcite, Mg-ankerite) with fairly sharp contacts with the groundmass are a characteristic feature of the damtjernites (Foley, 1984).
Cognate inclusions
A suite of undeformed micaceous inclusions, exclusively hosted by aillikites, comprises (1) glimmerite, (2) clinopyroxene–phlogopite and (3) olivine–phlogopite nodules in order of decreasing abundance (Table 3).
Glimmerite nodules are typically oval and less than 2 cm in diameter (Fig. 5e); rare examples approach 5 cm. Most glimmerites consist of interlocking 20–100 µm phlogopite flakes (Fig. 5f); some contain larger isolated phlogopite ‘clasts’ (up to 500 µm). Fluorapatite and rare orthoclase fill interstices or form discontinuous bands (Fig. 5e and f) that may open into radiating patches. Tiny spinel and ilmenite grains (<100 µm; also composite) are scattered throughout the fine-grained matrix.
Clinopyroxene–phlogopite nodules (up to 8 cm across; Fig. 5e) are similar in shape to the glimmerites, but have more variable mineralogy and are coarser grained. Large poikilitic phlogopite plates and clinopyroxene prisms dominate (up to 2 mm), whereas lath-like to interstitial Mg-ilmenite, chromite–titanomagnetite and prismatic hydroxy-fluorapatite occur only as minor components (Fig. 5 g). Mica plates occasionally enclose carbonated subhedral olivine. A subtle grain-size layering was observed in larger nodules. Interstitial calcic amphibole occasionally replaces clinopyroxene and phlogopite (Figs 5g and 6f), presumably as a product of melt infiltration. Subhedral titanite occurs in a single nodule (ST250C).
Rare cumulate-textured olivine–phlogopite nodules contain irregular olivine grains (300–800 µm), which are typically enclosed by large phlogopite plates (0·5–1·0 mm; Fig. 5 h). Hydroxy-fluorapatite (200–400 µm), titanomagnetite, ilmenite (200–800 µm) and rare clinopyroxene (<300 µm) occur as intercumulus phases. Perovskite and zirconolite (<100 µm) are rare accessories associated with titanomagnetite surrounding olivine grains.
|
MINERAL COMPOSITIONS |
|---|
Olivine
Aillikite/mela-aillikite olivine phenocrysts or microphenocrysts exhibit a fairly large range in forsterite component (Fo91–80 mol %; Figs 6b and 7, and Table 4), NiO (0·5–0·05 wt %), CaO and MnO (maxima of 0·9 and 0·4 wt %, respectively). Contrasting Mg/Fe evolutionary trends may indicate that different olivine populations are present. Olivine phenocrysts with normal zoning have core compositions of Fo87–91, decreasing to Fo82–85 towards the rim; NiO decreases, whereas CaO and MnO typically increase (0·1–0·3 wt %). Repetition of a normal zoning pattern may occur (Fig. 6b). Reverse zoning was often observed, with core compositions of Fo82–84 (NiO 0·2–0·3 wt %) steadily increasing towards the rim (Fo87–88; NiO 0·4 wt %). A discrete Fe-enriched overgrowth (Fo82; NiO 0·1 wt %) with sharp contact to the inner phenocryst usually occurs on reverse-zoned crystals, and can be strongly CaO enriched (up to 0·9 wt %).
|
Olivine phenocrysts in damtjernites are normal zoned (Fo80–86·5) and contain 0·18–0·5 wt % NiO (Fig. 7). CaO and MnO approach 0·4 wt % at the rims. The most primitive olivine cores in damtjernites are more evolved than their most primitive counterparts from aillikite/mela-aillikite (Fo91) but have similar high NiO concentrations (Fig. 7).
Rare subhedral olivine crystals in clinopyroxene–phlogopite nodules are normally zoned with a fairly evolved composition (Fo77–86·6; <0·1 wt % NiO; <0·4 wt % CaO; Fig. 7) and a conspicuously high MnO content (0·3–0·7 wt %). The olivine compositions (Fo80–86) in olivine–phlogopite nodules overlap with the low Mg/Fe end of aillikite phenocrysts, but are richer in NiO (up to 0·4 wt %) at a given Fo content (Fig. 7).
In general, olivine phenocryst compositions in UML and associated cognate inclusions from the Aillik Bay area are less primitive (<Fo91) than those found in kimberlites and lamproites, which typically approach Fo93 (Mitchell, 1986; Mitchell & Bergman, 1991; Fedortchouk & Canil, 2004; Prelevic et al., 2005).
Phlogopite
Phlogopite phenocrysts from aillikite and dolomite–calcite carbonatite typically have (1) a resorbed core with 15–16 wt % Al2O3 and up to 5 wt % TiO2, (2) a broad inner rim with elevated Al2O3 (up to 18 wt %) and lower TiO2 (about 2 wt %) and (3) a narrow outer rim with Al2O3 and TiO2 falling below 10 and 1 wt %, respectively, at constantly high MgO. This trend may culminate in virtually Al- and Ti-free tetraferriphlogopite rims in the most carbonate-rich samples (Figs 6c and 8a; Table 5). Whereas tetraferriphlogopite is uncommon in kimberlites, the rims reported here are similar to those from orangeites but distinct from titanian tetraferriphlogopites in lamproites. Phlogopite from type aillikite/mela-aillikite and carbonatite is generally Ba poor with core compositions typically below 1 wt % BaO. Rim compositions only rarely approach 3 wt % BaO. This is in contrast to kimberlite phlogopites, which commonly show a strong Al and Ba enrichment and Ti depletion toward the rim (Fig. 8a). The inner zones of less extremely zoned phlogopite plates (0·5–1·0 mm) from mela-aillikite contain 13–15 wt % Al2O3 and 3–5 wt % TiO2, but a high-Al inner rim composition such as in aillikite is absent. As in aillikites, Al2O3 depletion (8–13 wt %) toward the rim is common, but TiO2 increases rimwards (up to 6 wt %; Fig. 8a), which contrasts with the decreasing TiO2 trend observed in aillikites. Furthermore, micas from mela-aillikite follow a different Mg/Fe evolutionary trend than aillikite micas, with a strong increase in Fe at the expense of Mg leading to discrete dark brown biotite rims.
|
Damtjernite phlogopite plates compositionally resemble the less extremely zoned Ba-poor phlogopite plates from mela-aillikites with inner zones containing 13–15 wt % Al2O3 and 3–5 wt % TiO2 (Fig. 8b). A few samples were found to contain micas approaching 8 wt % TiO2 in the core. In general, these micas lack the high-Al inner rim composition of aillikite micas, but show Al depletion toward the rim (8–13 wt % Al2O3; Fig. 8b). We noted both rimward TiO2 increase (up to 8·5 wt % as in mela-aillikite) and decrease (down to 1 wt % as in aillikite). Micas in damtjernite show a strong Fe increase at the expense of Mg. This culminates in broad dark brown biotite overgrowths (Fig. 5d).
The mica compositional range in the clinopyroxene–phlogopite and olivine–phlogopite nodules (Fig. 8c) is the same as in phenocrysts from aillikites/mela-aillikites and damtjernites with the characteristically high Al2O3 (13–15 wt %) and TiO2 (1–8 wt %), but low BaO concentrations (<1·0 wt %). Fluorine concentrations are as low as in UML micas (<1·3 wt %), but much lower than in glimmerite phlogopite. The phlogopite plates are distinct from primitive mica compositions reported for MARID nodules (typically <12 wt % Al2O3, Dawson & Smith, 1977; Smith et al., 1978) and do not show evolution toward either tetraferriphlogopite or biotite.
Glimmerite phlogopites are compositionally unlike any of the phlogopite phenocrysts, plates or groundmass flakes described above (Table 5). They are highly magnesian (Mg-number 70–90), Al2O3 and TiO2 poor (5–12 and 0·3–2·0 wt %, respectively; Fig. 8c), BaO depleted (<0·2 wt %) but enriched in F (1–3 wt %).
Clinopyroxene and amphibole
Phenocrystic clinopyroxene in dolomite–calcite carbonatite and damtjernite, as well as groundmass prisms in damtjernites and mela-aillikites, are diopside-rich, showing Al2O3 and TiO2 enrichment towards the rim (up to 10 and 6 wt %, respectively). However, the average atomic Al/Ti ratio of carbonatite clinopyroxene is 3, distinctively higher than in clinopyroxene from associated mela-aillikite and damtjernite (2; Fig. 9a; Table 6). Cr2O3 concentrations in all these diopsides are below 0·1 wt %. Aegirine-rich overgrowths (up to 46 mol %) occur around diopside-rich phenocrysts in damtjernite. Some of these phenocrysts contain rare resorbed green Fe-rich salitic clinopyroxene cores that are rich in Al2O3 (up to 9 wt %).
|
Clinopyroxene–phlogopite nodules also contain diopside-rich clinopyroxene enriched in Al2O3 and TiO2 (up to 8 and 4 wt %, respectively), with atomic Al/Ti of
2·5 similar to the UML clinopyroxenes (Fig. 9b). Slightly FeO- and Na2O-enriched salitic core compositions (up to 9 and 3 wt %, respectively) may occur with Cr2O3 below 0·3 wt %. Olivine–phlogopite nodules carry rare diopside (4 and 2 wt % Al2O3 and TiO2, respectively), which is the most Cr2O3-rich composition (0·1–0·6 wt %) of all the clinopyroxenes from the Aillik Bay UML suite. In general, the strong Al and Ti enrichment in clinopyroxene from the Aillik Bay UML and their cognate inclusions contrasts with diopsidic compositions typical for groundmass clinopyroxene in orangeites, lamproites and associated MARID-type inclusions (Mitchell & Bergman, 1991; Mitchell, 1995). The intercumulus calcic amphibole found in the clinopyroxene–phlogopite nodules (Fig. 6f) is generally MgO and TiO2 rich (Mg-number 73–90 and 1·9–5·0 wt % TiO2) and ranges from magnesiohastingsite through pargasite to rare magnesiokatophorite. Fluorine is <0·5 wt % and K2O does not exceed 1·9 wt %.
Spinel group
Spinel group minerals from Aillik Bay UMLs and related micaceous inclusions generally follow a ‘titanomagnetite trend’ (trend 2 of Mitchell, 1986) which is characterized by FeT2+/(FeT2+ + Mg) >0·7, increasing Fe and Ti but decreasing Mg, Al and Cr (Fig. 10). The most Mg-rich spinels were found in aillikites but do not exceed 13·5 wt % MgO (Table 7). By comparison, kimberlite spinels are more magnesian (12–20 wt % MgO) and follow a trend of increasing Ti at buffered Fe/Mg of 0·5 (Fig. 10b; trend 1 of Mitchell, 1986). Aillik Bay UML spinels have Cr/(Cr + Al) ratios <0·85, which is in marked distinction to Cr-rich spinels in lamproites and orangeites (Cr-number >0·85), which follow the ‘titanomagnetite trend’ (Mitchell & Bergman, 1991; Mitchell, 1995).
|
Early-stage spinels in aillikites are typically composed of chromite–spinel solid solutions (up to 43 wt % Cr2O3, 13 wt % MgO, 12 wt % Al2O3). Rims of zoned spinel microphenocrysts and individual grains are of ulvöspinel–magnesian ulvöspinel–magnetite composition (up to 11 wt % MgO). The titanomagnetites are enriched in Al2O3 (up to 11 wt %) with low atomic Cr/(Cr + Al) ratios (<0·3). Spinels from mela-aillikites may contain cores of titanian magnesiochromite–chromite solid solution (up to 12 wt % TiO2, 9 wt % MgO, 25 wt % Cr2O3) and of chromite–spinel solid solution, similar to their aillikite analogues. Individual titanomagnetite microphenocrysts or rims around zoned chromite grains contain less MgO and Al2O3 (<5 wt %) than in the aillikites (Fig. 10a).
Spinels in damtjernites are dominantly titanomagnetite, which rarely exhibits cores of titanian magnesiochromite–chromite solid solution (up to 14 wt % TiO2, 9 wt % MgO, 22 wt % Cr2O3). Titanomagnetite in damtjernites has the lowest MgO concentration (typically <1 wt %) of all the Aillik Bay UML spinels (Fig. 10a), and contains similar levels of Al2O3 to mela-aillikites, but significantly less Al2O3 than the aillikites. Individual grains have rims approaching magnetite end-member composition.
Rare spinels in glimmerites are similar to the most evolved aillikite spinels with MgO and Al2O3 typically below 5 wt % following a ‘titanomagnetite trend’ (Fig. 10b). Composite spinel grains in clinopyroxene–phlogopite nodules may contain cores of chromite (up to 43 wt % Cr2O3) and/or Cr-spinel (up to 20 wt % Cr2O3) typically mantled by titanomagnetite. Titanomagnetite–magnetite resembles late-stage spinels from pyroxene-rich mela-aillikites in being very close to magnetite end-member composition. They are much more depleted in MgO and Al2O3 (typically <2·0 and 3·0 wt %, respectively) than their analogues from pyroxene-free glimmerites and aillikites (Fig. 10b). Titanomagnetite in olivine–phlogopite nodules resembles evolved aillikite spinels. They contain more MgO than spinels from clinopyroxene-rich nodules and mela-aillikite dykes (up to 7 wt %; Fig. 10b). Cr-spinel with up to 25 wt % Cr2O3 rarely occurs as inclusion in olivine.
Carbonate and Ti-rich primary garnet
Fe-rich dolomite crystals in dolomite carbonatite contain between 2 and 9 wt % FeO; only rarely approaching 12 wt % towards the rim. MnO is elevated (0·2–1·2 wt %), whereas SrO and BaO are conspicuously low (<0·2 wt %). Rare interstitial REE-carbonate is probably bastnäsite and contains up to 56 wt % LREE2O3 (where LREE is light rare earth element). Calcite in mixed dolomite–calcite carbonatite coexists with subordinate laths of Fe-rich dolomite (Fig. 6a) which resembles its counterpart from the dolomite carbonatites (2–12 wt % FeO and 0·2–1·2 wt % MnO). It differs in that both calcite and dolomite contain up to 1·5 wt % SrO. The groundmass of aillikites is dominated by a mosaic of Sr-calcite (up to 2 wt % SrO), whereas dolomite containing up to 10 wt % FeO is rare. Fe-rich dolomite seems to dominate over calcite (<1 wt % SrO) in the generally carbonate-poor groundmass of mela-aillikites. Carbonate in the damtjernite groundmass is Sr-calcite with up to 4·8 wt % SrO.
Small kimzeyitic garnets (Zr-rich andradite) are restricted to aillikites and have a fairly constant TiO2 content (9–11 wt %), whereas ZrO2 spans a wide range between 10 and 17 wt % (Table 8). Core compositions are generally richer in Zr than the rims. Schorlomite and/or melanite garnet is rare but characteristic for damtjernites, and observed zoning patterns are typically from Ti-rich core compositions to more Fe-rich rims (1·8–18 wt % TiO2; 15·7–21·6 wt % FeO). Zirconian schorlomite with up to 5 wt % ZrO2 in the core was rarely found. The presence of Ti-rich andradites and kimzeyitic garnets reflects the high Ca and Ti but low Al concentration of the UML magma and can therefore be regarded as characteristic for UML–carbonatite associations (Platt & Mitchell, 1979; Rock, 1986; Tappe et al., 2005a). These garnets do not occur in kimberlites and lamproites (Mitchell & Bergman, 1991; Mitchell, 1995).
|
PRESSURE ESTIMATES FOR COGNATE INCLUSIONS |
|---|
The clinopyroxenes and rare calcic amphibole of the clinopyroxene–phlogopite nodules allow qualitative pressure estimates. The clinopyroxene barometer of Nimis & Ulmer (1998)
requires an independent temperature estimate, which we obtained using the clinopyroxene thermometer of Kretz (1982). The uncertainty in temperature is 60 °C (1
) and results in large errors in pressure estimates (0·3 GPa, 1). Nevertheless, the crystallization pressure of clinopyroxenes from several clinopyroxene–phlogopite nodules can be bracketed between 0·8 and 1·5 GPa, corresponding to 25–45 km depth. Rare clinopyroxene from an olivine–phlogopite nodule gives a similar pressure estimate of 0·9–1·7 GPa.
Calcic amphibole in clinopyroxene–phlogopite nodules yielded the lowest crystallization pressures of 0·4–0·7 GPa (Al-in-hornblende barometer of Hammarstrom & Zen, 1986), corresponding to 10–20 km depth. This agrees with textural relations indicating late melt/fluid infiltration into the nodule material (Fig. 6f). No pressure estimate can be given for the glimmerite nodules, but the low-Ba mica compositions may be a reflection of comparably high crystallization pressures (Guo & Green, 1990).
|
MINERALOGICAL CONSTRAINTS ON CRYSTALLIZATION CONDITIONS AND THEIR IMPLICATIONS FOR MANTLE SOURCE CHARACTERISTICS |
|---|
Oxygen fugacity estimates from olivine–spinel and ilmenite–magnetite pairs
Olivine and Cr-spinel are the earliest phases crystallized from aillikite magma and may be used to constrain the oxygen fugacity conditions during early stages in UML magma evolution. We applied the FeMg–1 exchange thermometer of O'Neill & Wall (1987)
and the olivine–orthopyroxene–spinel oxybarometer of Ballhaus et al. (1991). Ferric iron in spinel was calculated assuming stoichiometry (Ballhaus et al., 1990). Because ultramafic lamprophyres are not saturated in orthopyroxene, the oxybarometer of Ballhaus et al. (1991) yields maximum fO2 values, which can be corrected for the appropriate silica activity of the melt as outlined by Fedortchouk & Canil (2004). The perovskite–titanite reaction (Nicholls et al., 1971) rather than the monticellite–diopside reaction as chosen by Fedortchouk & Canil (2004) for kimberlite was considered as the upper limit of silica activity controlling UML magma evolution at Aillik Bay. This assumption is consistent with the observation that perovskite and diopside-rich clinopyroxene frequently occur in the groundmass of these rocks. Some damtjernites contain perovskite and titanite in reaction relationship, indicating that crystallization occurred along this silica activity buffer. We assumed an equilibration pressure for olivine–spinel pairs of 1 GPa: pressure has only a minor influence on the calculation of the equilibrium olivine–spinel crystallization temperature (20 °C/GPa) and oxygen fugacity (0·03 log-bar units/GPa). The olivine–spinel equilibration temperatures for the aillikite magma range from 912 °C to 1300 °C (Fig. 11). The oxygen fugacity varies from FMQ –0·03 to FMQ +2·43 (log-bar unit deviation from fayalite–magnetite–quartz buffer) with most pairs recording fO2 slightly above the FMQ buffer. An olivine–spinel pair from a damtjernite (1253 °C; FMQ +1·84), and from an olivine–phlogopite cognate inclusion (1002 °C; FMQ +1·83) fall within the fO2–T range calculated for aillikites.
|
