Journal of Petrology Advance Access originally published online on April 25, 2008
Journal of Petrology 2008 49(6):1097-1131; doi:10.1093/petrology/egn019
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The Alkaline–Peralkaline Tamazeght Complex, High Atlas Mountains, Morocco: Mineral Chemistry and Petrological Constraints for Derivation from a Compositionally Heterogeneous Mantle Source
1Institut Für Geowissenschaften, AB Mineralogie UND Geodynamik, Eberhard-Karls-Universität, Wilhelmstrasse 56, D-72074 Tübingen, Germany
2Solid Earth Studies Laboratory (SESL), Department of Geology, University of Regina, Regina, Saskatchewan, S4S 0A2, Canada
RECEIVED SEPTEMBER 4, 2007; ACCEPTED MARCH 26, 2008
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONGEOLOGICAL SETTING AND PREVIOUS...FIELD RELATIONSPETROGRAPHYMINERAL COMPOSITIONSDISCUSSIONSUMMARY AND CONCLUSIONSAPPENDIX 1: CALCULATION OF...REFERENCES |
|---|
The Eocene Tamazeght complex, High Atlas Mountains, Morocco is a multiphase alkaline to peralkaline intrusive complex. A large variety of rock types (including pyroxenites, glimmerites, gabbroic to monzonitic rocks, feldspathoidal syenites, carbonatites and various dyke rocks) documents a progression from ultramafic to felsic magmatism. This study focuses on the silicate plutonic members and the genetic relationships between the various lithologies. Based on detailed petrographic and mineral chemical data we show that the various units crystallized under markedly different oxygen fugacity and silica activity conditions and demonstrate how these parameters influence both the phase assemblage and the detailed chemical evolution of the fractionating phases. Nepheline, olivine–clinopyroxene and hornblende–plagioclase thermometry indicate equilibration temperatures
800°C for all major rock types. Highly oxidized conditions (close to the hematite–magnetite buffer) are characteristic of the garnet-rich pyroxenites, ultrapotassic glimmerites and associated olivine-shonkinites. The parental magmas to these rocks evolved from low initial aSiO2 values of 0·1 to values of 0·5–0·8 during nepheline and alkali feldspar saturation. In contrast, the monzonitic rocks evolved from initially high aSiO2 values (up to 0·75) down to about 0·1 at intermediate values of oxygen fugacity (
FMQ = +2–5 to –1, where FMQ is the fayalite–magnetite–quartz buffer). For nepheline syenites and malignites, more reduced conditions (FMQ = –2) and intermediate aSiO2 values (between 0·25 and 0·5) dominate. We conclude that fractional crystallization is not a likely mechanism to explain the large variety of lithologies present in the Tamazeght complex. It is more probable that successive melting of a compositionally heterogeneous mantle source region gave rise to several melt batches with distinct chemical and physico-chemical characteristics. Low-degree melts from a K-phase-bearing mantle domain resulted in the formation of ultrapotassic glimmerites, whereas garnet-rich pyroxenites and olivine-shonkinites may have originated from hybrid melts and partly from a pyroxene-dominated source. Less alkaline lithologies such as monzonites potentially reflect larger degrees of melting and the increased importance of a basaltic component, whereas nepheline syenites and malignites may be explained by lower degrees of melting and a more alkaline character for the parental melt of these rocks. KEY WORDS: Tamazeght; Morocco; alkaline magmatism; source heterogeneity; Ti-bearing andradite
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTING AND PREVIOUS...FIELD RELATIONSPETROGRAPHYMINERAL COMPOSITIONSDISCUSSIONSUMMARY AND CONCLUSIONSAPPENDIX 1: CALCULATION OF...REFERENCES |
|---|
Alkaline to peralkaline igneous rocks represent a volumetrically small, but mineralogically highly variable, group typically located within intracontinental extensional settings. Chemically, these rocks are characterized by high contents of alkalis and incompatible elements, particularly the high field strength elements (HFSE; such as Ti, Zr, Hf and Nb). The residual fluids of such rock associations are known to give rise to a number of exotic mineral associations in pegmatites and hydrothermal veins (e.g. Salvi & Williams-Jones, 1990
; Khomyakov, 1995; Chakhmouradian & Mitchell, 2002), some of them being of economic interest (Sørensen, 1992).
The exceptional geochemical character of alkaline to peralkaline igneous rocks is reflected by an unusual phase assemblage and by the chemical composition of these phases; otherwise less-common minerals can appear as major constituents. For example, Ti-bearing andradite is commonly found in ultramafic alkaline lithologies (e.g. Coulson et al., 1999; Vuorinen et al., 2005) and eudialyte-group minerals (Na–Ca-zircono- and titanosilicates) are typical of highly evolved agpaitic nepheline syenites (e.g. Sørensen, 1997; Mitchell & Liferovich, 2006). It has been shown that the evolution of intensive parameters (e.g. fO2, aSiO2) during the crystallization of such rock types significantly influences the chemical composition of the phases present (e.g. Jones & Peckett, 1980; Coulson, 2003; Marks & Markl, 2003; Mann et al., 2006).
The association of ultramafic pyroxenites, leucocratic ijolites, and highly evolved nepheline syenites ± carbonatites is a common feature of alkaline plutonic complexes world-wide (e.g. Harmer, 1999; Dunworth & Bell, 2001; Vuorinen et al., 2005). Petrological and geochemical studies have revealed two principal genetic relationships in such complexes: either (1) closed-system fractionation of a common parental magma produces the various lithologies (e.g. Beccaluva et al., 1992; Markl et al., 2001; Marks et al., 2004; Halama et al., 2005) or (2) the various lithologies represent crystallization of magmas derived from different sources, or are related to each other by combined assimilation–fractionation–mixing processes (e.g. Kramm & Kogarko, 1994; Morikiyo et al., 2000; Arzamastsev et al., 2006).
The Tamazeght complex, which is the focus of this study, comprises numerous intrusive phases that document a progression from ultramafic to felsic alkaline to peralkaline rock types. A wide range of lithologies is present, including pyroxenites, glimmerites, gabbroic to monzonitic rocks, and predominating feldspathoidal syenites. Additionally, several carbonatitic diatremes and dyke rocks of lamprophyric, carbonatitic, phonolitic and foiditic composition occur throughout the complex and its sedimentary cover (Agchmi, 1984; Bouabdli et al., 1988; Mourtada et al., 1997; Neukirchen & Markl, in preparation). These are not, however, the focus of this work.
The large variety of rock types present in the Tamazeght complex questions the possibility that these rocks were derived from one parental magma by fractional crystallization alone; thus important questions concerning the origin and the genetic relationships between the different lithologies remain to be answered. Until now, there has been no systematic study of the mineral chemical and petrological evolution of the Tamazeght rocks. In this study we investigate in detail the chemical evolution of the fractionating phases to derive crystallization conditions for the various rock types in terms of oxygen fugacity (fO2) and silica activity (aSiO2). We further show how these parameters influence the chemical evolution of the mineral phases present and how such investigations are useful in deciphering the role of chemically different source components for such multiphase intrusive complexes.
|
GEOLOGICAL SETTING AND PREVIOUS WORK |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTING AND PREVIOUS...FIELD RELATIONSPETROGRAPHYMINERAL COMPOSITIONSDISCUSSIONSUMMARY AND CONCLUSIONSAPPENDIX 1: CALCULATION OF...REFERENCES |
|---|
The Tamazeght complex (also known as Tamazert complex) is located in the Moroccan High Atlas Mountains, about 20 km south of the city of Midelt (Fig. 1). Here, in the northern range of the High Atlas, NE–SW-striking dome and trough structures are the dominant structural features. Jurassic to Cretaceous marine sediments were deposited in intra-continental pull-apart basins that are related to the opening of the Atlantic Ocean (Laville, 1981
; Laville & Harmand, 1982). The Tamazeght complex is the largest of several alkaline intrusions associated with these graben structures. It intrudes Liassic marbles and crops out as an elongated body (16 km NE–SW and 5 km NW–SE) following the trend of the graben. Along the medial axis of the complex, Mesozoic marbles form the roof of the intrusion (Agard, 1960).
|
Numerous intrusive phases in the Tamazeght complex document a progression from ultramafic to felsic magmatism. Kchit (1990
) proposed the following chronology based on structural features, cross-cutting relationships and enclaves within the various units: (1) ultramafic rocks (pyroxenites & glimmerites); (2) shonkinites; (3) monzogabbroic rocks; (4) foid-monzosyenites; (5) malignites and associated pegmatites; (6) a range of textural varieties of nepheline syenites; (7) monzonitic rocks; (8) several carbonatite diatremes associated with carbonatitic and phonolitic to foiditic dyke rocks; (9) a lamprophyric dyke swarm.
All the intrusive units show vertical or near-vertical planar internal structures. Close to their margins these tend to be oriented parallel to the contacts, which are also sub-vertical. This led Kchit (1990) to conclude that the Tamazeght intrusive units represent irregular pipe-shaped bodies, which cross-cut each other. These magmatic ‘pipes’ were interpreted to represent magma in-fills of crustal fractures created by the same SW–NE sinistral shearing that characterizes the post-Cretaceous Atlas folding (Laville & Harmand, 1982). The presence of roof pendants, numerous pegmatites and contact metamorphic minerals within the surrounding marbles suggests that these magmatic bodies intruded to shallow depths of <3 km (Salvi et al., 2000).
Radiometric ages of 44 ± 4 Ma (Rb/Sr) and 42 ± 3 Ma (K/Ar) (Tisserant et al., 1976) have been determined for some of the monzonites. Nephelinitic dyke rocks, however, have an age of 35 Ma (Klein & Harmand, 1985). This relatively large time gap led Khadem Allah et al. (1998) to question the genetic relationship between the various intrusive phases. Nevertheless, based on geochemical data, Bouabdli et al. (1988) and Kchit (1990) assumed that all the rock units originated by fractional crystallization of a common parental magma of nephelinitic or monchiquitic composition. This parental magma was considered to have originated by low-degree partial melting of a carbonated amphibole-lherzolite mantle source. The carbonatites were thought to have formed through liquid immiscibility (Bouabdli et al., 1988).
The most recent studies of the Tamazeght complex focused on the fenitizing effects of carbonatitic fluids (Bouabdli & Liotard, 1999; Neukirchen & Markl, in preparation), on the influence of sedimentary carbonate rocks on the evolution of the peralkaline to agpaitic pegmatites of the complex (Khadem Allah et al., 1998), on the hydrothermal mobilization of HFSE within some of the nepheline syenites (Salvi et al., 2000, 2001) and the compositional variation of clinopyroxene in some of the nepheline syenites (Khadem Allah et al., 1996).
|
FIELD RELATIONS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTING AND PREVIOUS...FIELD RELATIONSPETROGRAPHYMINERAL COMPOSITIONSDISCUSSIONSUMMARY AND CONCLUSIONSAPPENDIX 1: CALCULATION OF...REFERENCES |
|---|
Field relations between the various rock units were described in great detail by Kchit (1990
) and Al-Haderi et al. (1998), constraining the chronological order of emplacement noted above. However, our own field work in the Tamazeght complex has shown that the distinction between the rock types in the field is not as obvious as indicated on the geological map of Kchit (1990) and the labeled rock types are not exclusively restricted to the areas indicated. For example, according to Kchit (1990) monzogabbro occurs as only one distinct body in the northeastern part of the complex (Fig. 1). However, the detailed petrographic investigation of our own samples (>150 specimens) has revealed that, for example, monzogabbroic rocks also occur within the foid monzosyenitic unit and vice versa. Obviously, the ratio between plagioclase and alkali feldpar is highly variable within single rock units on a small scale and clear intrusive contacts between them are only rarely visible in the field. If present, these are in many cases gradual without sharp contacts, and thus provide evidence for only a rather short time gap between the emplacement of the various rock units. Thus, here we treat monzogabbros, monzonites and foid monzosyenites as a single group; namely, the monzonitic group. All of these rock types are characterized by the occurrence of both feldspar types and we classify them based on their plagioclase:alkali feldspar ratio, irrespective of where they have been collected in the field. Similarly, the nepheline syenitic group shows heterogeneities in terms of grain size, mineral assemblage and macroscopic textures. In all, this indicates that most of these rocks were probably emplaced within a rather short time interval as a crystal mush, possibly also allowing for mixing and mingling among them. |
PETROGRAPHY |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTING AND PREVIOUS...FIELD RELATIONSPETROGRAPHYMINERAL COMPOSITIONSDISCUSSIONSUMMARY AND CONCLUSIONSAPPENDIX 1: CALCULATION OF...REFERENCES |
|---|
In this section, we describe the phase assemblages observed in the various lithologies and the micro-textural characteristics of the investigated samples. Figure 2 gives an overview of the mineral assemblages present.
|
The ultramafic group
Two types of ultramafic rocks can be distinguished: (1) pyroxenites consisting of variable amounts of clinopyroxene, nepheline and garnet; (2) glimmerites, which are dominated by biotite.
Pyroxenites (TMZ23b, 23c and 25) are dominated by euhedral clinopyroxene (Cpx), nepheline and euhedral to subhedral garnet (Fig. 3a and b). Minor phases are apatite, calcite, mica, magnetite, titanite and pyrite (occasionally with inclusions of pyrrhotite). Magnetite is locally transformed to hematite (Fig. 3c). Kchit (1990) also reported the rare occurrence of olivine. Amphibole and feldspar are absent in these rocks. With increasing nepheline content, some of the rocks are classified as mafic foidolites (melteigites and ijolites). Cumulus minerals are colourless to pale green clinopyroxene (showing discontinuous zonation patterns; Fig. 3d), nepheline, oscillatory zoned garnet and magnetite. Nepheline and garnet have a prolonged crystallization interval and are also present as intercumulus phases (Fig. 3b). Along the rims, nepheline is in places altered to cancrinite and/or sodalite. In one sample (TMZ23b) a zone several centimetres wide consisting of euhedral calcite, analcime and a late clinopyroxene generation is observed to cross-cut garnet-rich pyroxenite. Pyroxene in this vein is bright green and very fine-grained (generally <100 µm) and occurs as radiating clusters.
|
In glimmerites (TMZ20 and 22), poikilitic biotite is the dominant phase, at up to 65 modal %. Early minerals are clinopyroxene-I, garnet and minor perovskite, magnetite (occasionally with cores of chromite) and apatite. Feldspar, nepheline and amphibole are lacking. Compared with the pyroxenites, garnet is not euhedral but occurs in subhedral to anhedral granular aggregates. Clinopyroxene-I is locally rimmed or even replaced by a mixture of fine-grained mica and magnetite (Fig. 3e); early perovskite is always overgrown by titanite. Compositionally zoned ocelli-like textures, which occur throughout the rocks, consist of granular and colourless clinopyroxene-II in the outer parts and of interstitial calcite and/or magnetite in their cores (Fig. 3f). Commonly, magnetite is replaced by pyrite and hematite. Other opaque minerals include sphalerite and chalcopyrite. In places, a third pyroxene generation (Cpx-III) crystallized interstitially with respect to clinopyroxene-II.
The monzonitic group
This group of rocks is characterized by the occurrence of both plagioclase and alkali feldspar in addition to foid minerals (nepheline and minor sodalite and cancrinite). Based on their relative modal abundance, it is subdivided into foid (-bearing) monzogabbros, (foid-bearing) monzonites, (foid-bearing) syenites and foid-monzosyenites.
Monzogabbros are generally rich in euhedral to subhedral grey to pale green clinopyroxene and magnetite with minor amounts of ilmenite. However, both amphibole-rich and biotite-rich varieties exist. In amphibole-rich varieties (TMZ159), minor pyroxene is commonly overgrown by reddish brown amphibole, and in places, small rounded relics of clinopyroxene can be seen within euhedral amphibole (Fig. 4a). Both minerals contain subhedral inclusions of magnetite, and titanite occurs as subhedral crystals and as narrow (< 200 µm wide) rims overgrowing earlier ilmenite (Fig. 4b). In biotite-rich varieties (TMZ320), subhedral biotite is commonly associated with pyroxene, but is never seen to overgrow or to resorb it, unlike amphibole in amphibole-rich varieties. Also, biotite hosts inclusions of subhedral magnetite and needles of apatite. In these varieties, titanite is much more abundant and occurs exclusively as large (up to 2 mm) subhedral to euhedral crystals, occasionally with rounded inclusions of ilmenite (Fig. 4c).
|
Monzonites are porphyritic with euhedral phenocrysts of plagioclase and alkali feldspar set in a medium-grained matrix of clinopyroxene, magnetite, ilmenite, amphibole, biotite, titanite and feldspars. The grain size ranges from several centimetres to <1 mm. Accessory minerals are apatite and zircon. Subhedral to euhedral pyroxene co-crystallized with Fe–Ti oxides and titanite. In common with the monzogabbros, these rocks initially crystallized magnetite and ilmenite, the latter of which almost exclusively occurs as partly resorbed inclusions within titanite. Biotite is commonly corroded and shows a rim of fine-grained magnetite (Fig. 4d). Subordinate amphibole is subhedral and some of the amphibole cores host tiny patches of exsolved Fe–Ti oxides.
In foid-monzosyenites the relative modal amounts of pyroxene and amphibole are highly variable. Generally, euhedral pyroxene is pale grey to green in the core and has distinct bright green to yellow–green outer parts that show patchy heterogeneities. Euhedral titanite and subhedral magnetite (now coarsely exsolved to ilmenite and magnetite) appear to have co-crystallized with clinopyroxene (Fig. 4e) and both occur as inclusions in amphibole. No primary ilmenite was found in these rocks. Additionally, most amphibole shows the above-mentioned exsolution textures and in some samples (TMZ157 and TMZ312), a late pyroxene population overgrows earlier amphibole (Fig. 4f). Biotite-rich varieties typically are amphibole-free (TMZ318), but in samples with both phases (TMZ313), biotite predominates, occurring as rounded inclusions in amphibole or as a complex intergrowth.
The foid syenitic group
Following the IUGS nomenclature, these rocks are subdivided into shonkinites, nepheline syenites and malignites, based on the proportion of mafic minerals (Le Maitre, 2002). Mafic minerals include clinopyroxene, amphibole, perovskite–titanite, apatite, ± olivine, ± biotite, ± magnetite, ± eudialyte, ± zircon, ± garnet, ± calcite, ± fluorite. Primary felsic minerals include alkali feldspar, nepheline and sodalite. Pure albite occurs as a late magmatic phase.
Shonkinites have a colour index >60 (Le Maitre, 2002) and are subdivided into olivine-bearing and amphibole-rich varieties. Olivine-shonkinites (TMZ12 and 130) are characterized by large (1–5 mm) phenocrysts of olivine, pale grey clinopyroxene-I and magnetite-I set in a fine-grained groundmass of greenish clinopyroxene-II, garnet, apatite, alkali feldspar and nepheline. The last mineral is in most cases strongly altered to calcite, analcime, cancrinite and sodalite. Olivine phenocrysts have rounded grain boundaries and are partly rimmed by a fine-grained mixture of magnetite-II, amphibole and biotite (Fig. 5a); clinopyroxene-I is commonly overgrown or invaded by clinopyroxene-II (Fig. 5b). If present, perovskite is rimmed by titanite (Fig. 5c). However, titanite also occurs as rims around magnetite-I and, occasionally, also as euhedral grains. Garnet is present as a minor phase in one sample (TMZ12).
|
Amphibole-shonkinites (TMZ68, 139 and 308) are coarse-grained and do not show any noticeable phenocrysts. Here, euhedral amphibole (up to 5 mm) strongly dominates over clinopyroxene and is commonly associated with subhedral biotite (Fig. 5d). Minor euhedral clinopyroxene is finer grained (generally <2 mm) than amphibole and grey to pale green in colour. Occasionally, it shows irregular and patchy heterogeneities, where the outer regions of the crystals are more greenish and the inner regions are more greyish in colour. Olivine and perovskite are absent and titanite is always euhedral. Magnetite and minor ilmenite occur either as subhedral grains associated with clinopyroxene, amphibole and titanite (TMZ139 and 308) or as subhedral to anhedral rounded grains as inclusions in these three minerals (TMZ68). Petrographically this group shows similarities to some of the monzonitic rocks.
Nepheline syenites have a colour index <30 (Le Maitre, 2002) and, based on their general texture, a number of varieties can be distinguished.
Foyaitic nepheline syenites (TMZ165, 221 and 223) are generally coarse-grained and consist of a framework of large alkali feldspar laths (up to 5 mm) associated with euhedral to subhedral nepheline and minor sodalite, the latter of which is strongly altered. Locally, interstitial albite also occurs. Euhedral clinopyroxene with grey cores and distinct yellow–greenish rims forms larger aggregates and is commonly associated with euhedral titanite, apatite, magnetite and biotite (Fig. 6a). In samples, which were collected in the vicinity of carbonatite dykes, biotite is commonly intergrown with clinopyroxene and appears to replace it (Fig. 6b). Primary subhedral amphibole in these rocks is rare. If present, it shows fine-grained exsolution textures in the core region (TMZ221).
|
In granular nepheline syenites (TMZ74, 94, 95 and 126) subhedral clinopyroxene shows grey cores with distinct greenish coloured rims (Fig. 6c) and is associated with euhedral titanite and magnetite. Euhedral to subhedral amphibole is brown to dark green in colour and shows similar exsolution textures to those in the foyaitic varieties. Locally, it is overgrown by green fine-grained clinopyroxene (Fig. 6d). Clinopyroxene, amphibole and titanite contain tiny needle-shaped inclusions of apatite.
Porphyritic nepheline syenites (TMZ311) consist of centimetre-sized euhedral alkali feldspar phenocrysts with euhedral to subhedral nepheline, clinopyroxene, amphibole, titanite and rounded magnetite filling the space between them. Clinopyroxene is generally pale green, showing no distinct greenish rim but a patchy inhomogeneity. As in the foyaitic and granular varieties, amphibole is subhedral to euhedral and also shows characteristic exsolution in the core regions (Fig. 6e).
Malignites have a colour index of 30–60 (Le Maitre, 2002) and are generally amphibole- and biotite-free. Occasionally, more leucocratic varieties exist. However, to distinguish this rock type from the other foid syenites, we call them malignites throughout this work. Based on the presence of eudialyte or låvenite [simplified formula (Na,Ca)2(Mn2+,Fe2+)(Zr,Ti,Nb)Si2O7(O,OH,F)], they are subdivided into miaskitic and agpaitic varieties. Euhedral apatite and titanite occur as inclusions in clinopyroxene, nepheline or alkali feldspar. Mostly euhedral green to yellow–green clinopyroxene (up to 5 mm in size) shows irregular heterogeneities throughout most samples (Fig. 6f). Nepheline and alkali feldspar are both subhedral in habit, and sodalite occurs as an interstitial phase. In some spatially restricted areas eudialyte (Fig. 6g) or låvenite were formed during the late-magmatic stage, accompanied by felty clinopyroxene-II and small albite laths (Fig. 6h). Late-stage hydrothermal processes are documented by the formation of symplectitic cancrinite–sodalite seams around precursor nepheline. Within the malignites, a number of pegmatites and hydrothermal veins are recognized. The pegmatites have been intensively studied by Khadem Allah et al. (1998) and, thus, we investigated only one aegirine-rich pegmatite sample (TMZ247) for this study. It consists of euhedral centimetre- to decimetre-sized yellow–green sector-zoned clinopyroxene. Locally, hematite occurs interstitially between the pyroxene and as rounded inclusions within pyroxene crystals. Minor minerals are alkali feldspar and eudialyte. Hydrothermal veins (TMZ177, 229, 231 and 234) are generally several centimetres wide. They consist of alkali feldspar laths several centimetres in size and yellow–green pyroxene needles of similar length. Accessory minerals are zircon or eudialyte, catapleite, magnetite and Nb- and Mn-rich ilmenite.
|
MINERAL COMPOSITIONS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTING AND PREVIOUS...FIELD RELATIONSPETROGRAPHYMINERAL COMPOSITIONSDISCUSSIONSUMMARY AND CONCLUSIONSAPPENDIX 1: CALCULATION OF...REFERENCES |
|---|
Analytical techniques
The major and minor element compositions of the constituent minerals were determined using a JEOL 8900 electron microprobe in wavelength-dispersion mode at the Institut für Geowissenschaften, Universität Tübingen (Germany). For silicate minerals, we used a beam current of 15 nA and an acceleration voltage of 15 kV; for Fe–Ti oxides, we used 20 nA and 20 kV. The peak counting time was 16 s for major elements and 30–60 s for minor elements. Background counting times were half of the peak counting times. The peak overlap between the Fe Lβ and F K
was corrected for. To avoid Na migration under the electron beam, analyses of feldspar, nepheline and sodalite were performed with a defocused beam of 10 µm diameter. In cases where Fe–Ti oxides showed fine-grained exsolution textures they were analysed with a defocused beam of 20–40 µm diameter. For calibration, both natural minerals and synthetic phases were used as standards. Processing of the raw data was carried out with the internal 
Z correction method of JEOL (Armstrong, 1991). Analytical uncertainties are below 1% relative for major elements and around 15–20% for minor elements. The bulk composition of coarsely exsolved Fe–Ti oxide grains was reconstructed by combining image processing (NIH Image software) of back-scattered electron (BSE) images of the exsolved mineral grains with point analyses of exsolved ilmenite and magnetite. The bulk composition was then recalculated using the area proportions of both exsolved phases and using molar volumes of 44·52 and 31·70 cm3/mol for magnetite and ilmenite, respectively. Generally, this procedure was applied to 3–5 grains in each investigated sample.
Olivine
Within two samples of olivine-shonkinite, the compositional variation of olivine is small (Fo90–87 in TMZ12 and Fo88–75 in TMZ130; Table 1). Most of this variation is related to normal growth zonation with decreasing XMg [Mg/(Mg + Fe2+)] from core to rim but essentially unzoned olivine is also present (Fig. 7). However, the olivine from the two samples differs significantly in terms of its minor element composition. In TMZ12, the olivine is relatively high in NiO (up to 0·4 wt %) but low in CaO (< 0·3 wt %) and MnO (< 0·25 wt %), whereas in sample TMZ130, the opposite is the case (NiO <0·16 wt %; CaO <0·4 wt %, MnO <1·3 wt %).
|
|
Clinopyroxene
As is typical for alkaline intrusive complexes, clinopyroxene shows a wide range of compositions. For a detailed chemical classification, 10 end-members were computed, assuming stoichiometry (six oxygen atoms and four cations). Details of the applied calculation scheme are given in the Appendix. In addition to the Quad-components (enstatite [En], ferrosilite [Fs], diopside [Di] and hedenbergite [Hed]), we include the Na-bearing components aegirine [Aeg], (Ti, Zr)-aegirine [Ti-Aeg], and jadeite [Jd]. In most analyses, the calculated Fe3+ content exceeds the Na content, implying the presence of a ferri-Tschermak component [Fe-Ts]. The AlIV-bearing components Ca-Tschermak [Ca-Ts] and Ti-Tschermak [Ti-Ts] are also considered. The variation of these components in the various lithologies is illustrated in Figs 8 and 9 and representative clinopyroxene analyses are given in Table 2.
|
|
|
The end-members diopside, hedenbergite and aegirine are the most important ones to describe the compositional evolution of clinopyroxene from the Tamazeght suite. In all lithologies, similar Di-rich pyroxene compositions are found and these evolve towards more Hed- and Aeg-rich compositions with progressive differentiation. However, the various rock types show variable relative amounts of Hed enrichment while evolving towards Aeg-rich compositions, resulting in rather flat evolutionary trends for rocks of the monzonitic and nepheline syenitic group and comparatively steep trends for shonkinites. Ultramafic rocks show an intermediate trend. Also, the overall variation of clinopyroxene compositions observed in one rock type is highly variable, with clinopyroxene from nepheline syenites showing by far the largest chemical variation. Within one sample, the whole trend from Di-rich via intermediate towards Aeg-rich compositions can be traced (Fig. 8). These differences are of major importance and will be discussed in detail below.
Intermediate (aegirine–augite) pyroxene compositions in shonkinites and in some malignites do not follow well-defined trends as is found for most other rock types. These broad compositional fields can be correlated with irregular heterogeneities as is evident from BSE images (Fig. 6f).
The minor components Fe-Ts, Ca-Ts and Ti-Ts are generally <10 mol %. Typically, they are lower in ultramafic rocks than in the other rock types. However, within the ultramafic rocks, Fe-Ts is relatively enriched in the two inner zones of discontinuously zoned clinopyroxene from pyroxenites (Figs 3d and 9). In glimmerites, partly resorbed phenocrysts of clinopyroxene-I (Fig. 3e) are relatively rich in all three Tschermak components (Fig. 9). Significant amounts of the Ti-Aeg end-member are generally restricted to Aeg-rich clinopyroxene compositions, where it may be up to 
30–40 mol%.
Zoning profiles for clinopyroxene are different in the various rock groups. In pyroxenites, clinopyroxene shows discontinuous zonation, displaying three distinct pyroxene compositions (Figs 3d and 9). Pyroxene in monzonitic rock types shows continuous zoning patterns, starting with Di-rich compositions in the cores and evolving towards more Hed- and Aeg-bearing compositions at the rims, coinciding with an increase in Tschermak components. In granular syenite, clinopyroxene is discontinuously zoned (Fig. 6c). Within the Di-rich core a sudden increase and decrease of Fe-Ts and Ti-Ts is observed. The rim evolves towards high amounts of the Aeg component but is low again in Tschermak components. This evolutionary trend is accompanied by a continuous decrease in Ca-Ts from core to rim. According to their heterogeneous appearance (Fig. 6f), zoning profiles for clinopyroxene from malignites reveal that, despite a rough trend of increasing Aeg-component from core to rim, the evolution in XMg and Tschermak components is not strictly systematic.
Fe–Ti oxides
Fe–Ti oxides show considerable variation in terms of phase assemblage, composition and exsolution textures among the various rock types. Figure 10 illustrates this variation and Table 3 provides representative analyses.
|
|
In pyroxenites, primary, almost Ti-free magnetite (Mag99–100Usp0–1) is commonly replaced by hematite (Ilm0–2Hem98–100Pyr0–1). In glimmerites, euhedral opaque phases in the mica-rich matrix are mostly Ti-poor magnetite (Mag79–93Usp4–8Spl1–13) and in rare cases, these show distinct Cr-rich cores (Mag14–17Usp3–5Spl78–81), with an XCr value [Cr/(Cr + Al)] of
0·75 and an XMg value of 0·2. The composition of magnetite from ocelli-like textures overlaps with the range observed in matrix magnetite (Mag90–95Usp5–10Spl0–1). In monzogabbros and monzonites, homogeneous and Ti-poor magnetite (Mag97–100Usp0–2Spl0–1) and ilmenite (Ilm80–83Hem10–12Pyr7–10) were observed; the latter is much less abundant and occurs almost exclusively as rounded inclusions in titanite or is overgrown by the latter. In foid-monzosyenites, Ti-bearing magnetite (Mag67–83Usp15–27Spl2–5) shows coarse sandwich-type exsolution textures, but primary ilmenite is absent.
In olivine-shonkinites, ilmenite is absent but the magnetite composition is relatively variable (Mag55–90Usp6–40Spl4–7). In amphibole-shonkinites both Ti-bearing magnetite (Mag88–98Usp1–11Spl0–4) and ilmenite (Ilm84–87Hem8–11Pyr4–7) are present. In nepheline syenites, the composition of magnetite (Mag76–99Usp1–23Spl0–2) shows no systematic variation between the textural varieties. In miaskitic malignites, magnetite (Mag83–99Usp1–17Spl0–1) is present; this is, however, lacking in the agpaitic varieties. Late-stage hydrothermal veins contain either Mn-rich ilmenite (Ilm46–49Hem1–2Pyr50–53; TMZ234) or Ti-poor magnetite (Mag86–98Usp2–13Spl0–1; TMZ229).
Generally, V2O3 contents are higher in ilmenite (up to about 3 wt %) than in magnetite (<0·6 wt %) but no systematic differences between the various rock types were observed. The highest Cr2O3 contents were found in the cores of spinel grains from glimmerites (up to 38·6 wt %); however, the vast majority of the magnetite in the ultramafic rocks contains <3 wt % Cr2O3. In shonkinites, magnetite contains up to 1· 5 wt % Cr2O3, but in all other rock types, Cr2O3 contents in magnetite and ilmenite are much lower (<0·4 and <0·1 wt %, respectively). ZnO contents are generally below <1 wt %, with no obvious difference between magnetite, hematite or ilmenite nor with any systematic evolution within the complex. Only in the Cr-rich spinel of the glimmerites were increased ZnO contents (up to 3·2 wt %) detected. For all analyses, ZrO2 contents are below detection limit, as are Nb2O5 contents, except for the Mn-rich ilmenite from hydrothermal vein TMZ234, where up to 1· 6 wt % of Nb2O5 was detected.
Garnet
Garnet in the Tamazeght rocks is generally rich in Ca, Fe3+ and Ti, covering the compositional range between Ti-bearing andradite and schorlomite (Ca3Ti4+2[Si3–x(Fe3+,Al,Fe2+)x]O12). Representative compositions are reported in Table 4. The nomenclature concerning schorlomite is somewhat controversial [see Chakhmouradian & McCammon (2005) for a recent discussion]. In the absence of crystallographic and spectroscopic data, we do not attempt to constrain the distribution of Ti, Fe and Al between the cation sites.
|
In the ultramafic rocks, TiO2 contents range between 3·99 and 14·29 wt %. Si shows significant deviation from the ideal stoichiometry (2·96–2·49 p.f.u.) and Fe3+ varies considerably (1·12–1·57 p.f.u.). In olivine-shonkinites, TiO2 (1· 96 and 18·84 wt %), Fe3+ (0·99–1· 66 p.f.u.) and Si (2·12–3·0 p.f.u.) display strong variance. In nepheline syenites, TiO2 and MgO contents are comparatively lower (3·13–4·38 wt % and 0·22–0·28 wt %, respectively), whereas Al2O3 (3·14–4·12 wt %), MnO (0·89–1· 54 wt %) and ZrO2 (up to 1· 2 wt %) are significantly higher than in other lithologies (Table 4). This may reflect the more evolved character of these rocks.
Many workers advocate a simple homovalent substitution of Ti 
Si, based on the negative correlation between these two elements, to account for the apparent deficit on the Z-site. However, for the Tamazeght garnets the strong negative correlation between Ti and Si does not exactly fall on the ideal 1 : 1 correlation and even when Si is ideal (at 3 a.p.f.u.) 0·2 Ti p.f.u. is present (Fig. 11). This indicates that Ti and/or Si are involved in other substitutions. In an attempt to further identify important substitutions occurring within the Tamazeght garnets, we employed principal component analysis (PCA), a statistical method that has proven useful in petrological studies (Jiménez-Millán et al., 1994; Ragland et al., 1997) to our data, using XLSTAT 2007.6 (Addinsoft). The method extracts a set of principal components, which allows us to explain the observed variability in compositions. We deduce that the most substitution schemes are Ti4+Fe3+Fe3+–1Si–1, Ti4+MgFe3+–2 and/or Ti4+Fe2+Fe3+–2, which are responsible for about 80% of the observed variability of the data.
|
Amphibole
The majority of amphibole analyses show a good 1 : 1 correlation between Ti and (Mg, Fe, Mn) (Fig. 12) indicating that the substitution Ti4+O2–2Mg–1(OH–)–2 (e.g. Oberti et al., 1992
) plays an important role in the incorporation of Ti in the Tamazeght amphiboles; consequently, the amphibole formula unit was calculated on the basis of (23 + Ti) oxygen atoms and 16 cations. Table 5 gives some representative amphibole analyses. Following the nomenclature scheme of Leake et al. (1997), all analyses are calcic amphiboles of hastingsitic (AlVI 
Fe3+) and kaersutitic (Ti 0·5 p.f.u.) composition. The latter is restricted to monzogabbros, some monzonites and to shonkinites (the least evolved rock types of the respective lithological groups). Figure 13 illustrates the variation in XK [K/(Na + K)], AlVI p.f.u., F p.f.u., XMg, XFe3+ [Fe3+/(Fe2+ + Fe3+)], and Ti p.f.u. observed throughout the complex. The last three variables show a systematic evolution within the two lithological rock groups, each parameter decreasing with progressive evolution. The variation in octahedrally coordinated aluminium (AlVI) and XK appears to be unsystematic. However, within the monzonitic group the absolute range of XK seems to decrease from monzogabbros via monzonites towards foid-monzosyenites, whereas the maximum XK value in question increases slightly. In terms of halogens, chlorine content is always low (<0·04 p.f.u.) and fluorine contents are variable, with foid-monzosyenites and foyaitic nepheline syenites showing comparatively high F contents of <0·53 and <0·72 p.f.u., respectively.
|
|
|
Biotite
Biotite occurs in all samples of the ultramafic and the monzonitic groups. Within the foid syenitic group, only shonkinites and foyaitic nepheline syenites contain biotite. Similar to the amphiboles, the Tamazeght biotites are characterized by low Si contents and 8 – Si + Al deficits of up to 0·34 p.f.u., which indicates the presence of tetrahedrally coordinated Fe3+ or Ti4+ (e.g. Dunworth & Wilson, 1998
; Mann et al., 2006). Figure 14 illustrates the variation in XMg, Ti and F p.f.u. observed throughout the complex and Table 6 gives some representative analyses.
|
|
XMg values are highest in biotites from the ultramafic rocks (up to 0·96) and decrease towards the more evolved rock types, reaching their lowest values (<0·3) in some of the foyaitic nepheline syenites. In olivine-shonkinites, however, two types of biotite occur, groundmass biotite and biotite growing at the expense of olivine, with the latter having exceptionally high XMg values (around 0·9), reflecting the XMg value of the precursor olivine.
Ti contents show considerable variation, being lowest in ultramafic rocks (<0·35 Ti p.f.u.), in biotite replacing pyroxene in foyaitic nepheline syenites (Fig. 6b) and in the biotite from the olivine-shonkinites that overgrows olivine (<0·23 Ti p.f.u.). All other biotites have elevated Ti contents, with the highest Ti contents found in monzogabbros (up to 1·16 Ti p.f.u.) and shonkinites (up to 0·78 Ti p.f.u.). It should be noted that these two rock types also contain the most Ti-rich amphiboles. Such high Ti contents could potentially explain the 8 – (Si + Al) deficits on the tetrahedral site. The positive Ti–Al correlation and the negative Ti–Si and Ti–(Mg,Fe,Mn) correlations (Fig. 15) imply the importance of the coupled substitution MgSi2Ti4+–1Al–2, which was proposed by Wagner et al. (1987) and Mann et al. (2006) for biotite from alkaline rocks of the Katzenbuckel volcano, Germany. While replacing divalent cations by Ti4+ on octahedral sites, charge-balance might also be reached by the substitution mechanism MgK2Ti4+–1Al–2, which creates vacancies on the X site (Deer et al., 1992). According to the applied formula calculation (normalization to 22 oxygens), up to 15% of the X site may be vacant. In Fig. 15a and b, low-Ti biotites from glimmerites and from foyaitic nepheline syenite TMZ223 deviate from the trend shown by biotites from all other samples. In Fig. 15c, only the latter plot off the trend. This feature also coincides with elevated F contents in these samples and may indicate that their chemistry is governed by other substitution mechanisms, potentially implying a different origin for these micas.
|
In most biotites, chlorine contents are <0·03 p.f.u, except for monzogabbros and monzonites, where slightly higher Cl contents of up to 0·07 p.f.u were found. Fluorine contents are highly variable (from <0·01 to >1 p.f.u.; Fig. 14). Generally, F is negatively correlated with Ti content and reaches high values in ultramafic rocks, in biotite around olivine from shonkinites, and in biotite in evolved foyaitic nepheline syenites. F contents in biotite from the monzonites do not fit this relationship, but this might be explained by simple alteration of mica in these rocks (see above and Fig. 4d).
Feldspar
Ca-bearing plagioclase is restricted to rocks of the monzonitic group where individual grains are strongly zoned with decreasing mol % anorthite (An) and increasing mol % albite (Ab) from core to rim; orthoclase (Or) is generally low. Overall, plagioclase composition varies between An68Ab31Or1 and An22Ab74Or4 (Fig. 16; Table 7). The most anorthite-rich compositions are found in samples of monzogabbro, whereas the most anorthite-rich plagioclase in monzonites and foid-monzosyenites is very similar (An52 and An44, respectively).
|
|
Alkali feldspar is in most cases exsolved into pure albite and orthoclase. These textures are partly rather coarse and/or heterogeneous, and this feature makes it difficult to reconstruct a primary magmatic composition. However, in many samples (except for foyaitic nepheline syenites) some alkali feldspar grains (or at least parts of them) show no signs of exsolution. Unexsolved alkali feldspar in rocks of the monzonitic group varies in composition between Ab48Or48An4 and Ab20Or78An2 (Fig. 16; Table 7), and exhibits no systematic evolution from monzogabbros to monzonites to foid-monzosyenites. In porphyritic and granular nepheline syenites, alkali feldpar composition varies between Ab70Or26An4 and Ab20Or80An0, and in malignites, as well as in hydrothermal veins, An-free and relatively Or-rich (Ab29Or71–Ab12Or88) alkali feldspar is found. Interstitial albite in foyaitic nepheline syenites as well as late-stage albite laths in some of the malignites are An-free and contain generally <2 mol% orthoclase.
Foid minerals
Nepheline
The variation of nepheline composition is illustrated in Fig. 17; representative nepheline compositions are given in Table 8. In pyroxenites, nepheline composition varies between Ne60Ks25Qtz15 and Ne72Ks25Qtz4. The relatively Qtz-rich and Ne-poor compositions are typically found in the cores of euhedral nepheline grains, which occur as inclusions in garnet, whereas the Qtz-poor and Ne-rich compositions are typical of interstitial nepheline grains.
|
, the Morozewicz nepheline composition; 
, the Buerger nepheline composition.
|
In rocks of the monzonitic and nepheline syenitic group, nepheline varies in composition between Ne67Ks12Qtz21 and Ne72Ks18Qtz10 with no systematic differences between the various rock types. The evolution from Qtz-rich and Ks-poor to relatively Qtz-poor and Ks-rich compositions is in contrast to the compositional evolution of nepheline from the pyroxenites and has been described as being typical of post-magmatic re-equilibration (Powell, 1978
). A similar compositional variation is observed within miaskitic and agpaitic malignites (Ne70Ks12Qtz18–Ne75Ks22Qtz3), with a tendency for Ks-rich compositions to be more frequent in agpaitic malignites. Nepheline compositions in a hydrothermal vein overlap with the Qtz-poor compositions of the malignites.
CaO and Fe2O3 contents may be up to 1· 5 wt % and 1· 6 wt %, respectively. The highest Fe contents are present in pyroxenites and malignites, and the lowest contents were observed in monzonites and nepheline syenites.
Sodalite
Sodalite-group minerals occur as euhedral and interstitial phases in most samples. Sodalite is not found in either the ultramafic or shonkinitic lithologies. Compositional differences between primary sodalite and sodalite associated with cancrinite in reaction textures are not obvious. The chemical composition of both types is close to end-member sodalite, with Cl between 1· 55 and 1· 89 a.p.f.u. and SO3 ranging from 0·01 to 0·4 a.p.f.u. (Table 9). A weak negative correlation between S and Cl is observed. Minor elements include Fe (< 0·5 wt % Fe2O3), K (< 0·1 wt % K2O) and Ca (< 0·4 wt % CaO).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTING AND PREVIOUS...FIELD RELATIONSPETROGRAPHYMINERAL COMPOSITIONSDISCUSSIONSUMMARY AND CONCLUSIONSAPPENDIX 1: CALCULATION OF...REFERENCES |
|---|
Evidence from the mineral chemical variations for a heterogeneous magma source
Clinopyroxene occurs in all lithologies and is therefore most suited to track the physico-chemical evolution of the Tamazeght magmas. The evolution from diopside-rich pyroxene compositions towards end-member aegirine is typical of alkaline complexes worldwide. The major difference between various complexes is the amount of Fe2+ enrichment relative to Na and Fe3+ enrichment during their evolution (Fig. 18). In that sense, two extreme evolutionary paths have been documented: from diopside to aegirine without significant Fe2+ enrichment [e.g. Murun, Siberia (Mitchell & Vladykin, 1996
) and Katzenbuckel, SW Germany (Mann et al., 2006)] and from diopside via hedenbergite, and thus strong Fe2+ enrichment, prior to evolution towards aegirine-rich compositions [from the Ilímaussaq Complex, South Greenland (Larsen, 1976; Marks & Markl, 2001; Markl et al., 2001)]. In addition to these two extremes, intermediate paths have been documented in many studies (e.g. Mitchell & Platt, 1982; Korobeinikov & Laajoki, 1994; Coulson, 2003; Vuorinen et al., 2005). The most obvious factor influencing the extent of Fe2+ enrichment in clinopyroxene during differentiation is the oxidation state of the magma (e.g. Larsen, 1976). However, the presence of coexisting mafic minerals (e.g. amphibole, biotite, garnet) has also been shown to play a role (e.g. Chakhmouradian & Mitchell, 2002; Vuorinen et al., 2005) and it seems likely that the Na/Ca ratio of the melt or fluid from which the pyroxene crystallizes also influences the evolutionary path. For the two extreme trends, quantitative data on oxygen fugacitiy (fO2) are available for the Katzenbuckel (SW Germany) and the Ilímaussaq suite (South Greenland), and indeed, for these two suites relatively oxidized (FMQ = +1 to +2, where FMQ is the fayalite–magnetite–quartz buffer) and extremely reduced crystallization conditions (FMQ = –2 to –4), respectively, were determined (Marks & Markl, 2001; Markl et al., 2001; Mann et al., 2006). For intermediate suites (e.g. North Qôroq, South Greenland; Alnö, Sweden; Fig. 18), no quantitative estimates have been reported, but it seems likely that, in terms of fO2, these suites formed under conditions somewhere around the FMQ buffer.
|
Although in all Tamazeght units similar diopside-rich compositions are observed, the amounts of Na and Fe3+ increase during Fe2+ enrichment (compare the slopes of the clinopyroxene evolutionary path) and the lengths of the evolutionary paths differ between the various rock types (Fig. 8). By far the steepest slope is observed in the evolution trend for early clinopyroxene in shonkinites whereas a comparatively flat path is tracked by clinopyroxene from the monzonitic group and from nepheline syenites. Qualitatively, these differences might indicate differences in the oxidation state of the parental magma, with shonkinites probably crystallizing under more oxidized conditions compared with both rocks of the monzonitic group and nepheline syenites. This observation is in accordance with the composition of coexisting Fe–Ti oxides in the respective rocks. Some of the shonkinites contain Ti-poor magnetite, whereas nepheline syenites and monzonites contain either Ti-enriched magnetite or magnetite and ilmenite (Fig. 10). The trend shown by clinopyroxene from the ultramafic rocks is intermediate, although in these rocks Ti-free magnetite and hematite occur.
In shonkinites and in some malignites, intermediate pyroxene compositions (aegirine–augite) were found, which do not follow a well-defined path, but are remarkably variable in composition and plot along a broad band within the central part of the Di–Hed–Aeg triangle (Fig. 8). This is in contrast to the well-defined clinopyroxene trend observed in foyaitic nepheline syenites, which similarly evolve via intermediate to aegirine-rich compositions but follow a tight path. Such tightly defined evolutionary paths most closely resemble the chemical evolution of clinopyroxene during its primary crystallization history, which is directly linked to the physico-chemical evolution in the crystallizing melt. In contrast, the poorly defined compositional fields of late clinopyroxene from shonkinites and from miaskitic malignites (Fig. 8) and their heterogeneous micro-textural appearance (Fig. 6f) imply that these compositions reflect different extents of diffusional re-equilibration with a fluid phase during sub-solidus conditions.
Detailed clinopyroxene zoning profiles reveal that the relative proportions of Al-Ts, Fe-Ts and Ti-Ts change systematically during evolution and that this systematic evolution is different in rocks of the monzonitic group compared with that in the nepheline syenites. In monzonitic rocks, all three Tschermak components increase from core to the rim of crystals, whereas the opposite is the case in nepheline syenites. The schematic equilibrium
|
| (1) |
). The observed variations in Ti-Ts between the rock units (Fig. 9) show that, generally, clinopyroxene in the ultramafic rocks is significantly lower in Ti-Ts than in all other rock types (neglecting some of the most evolved nepheline syenites). The most obvious difference between the ultramafic and the other rocks is the presence or absence of Ti-bearing andradite. Although no simple schematic equilibrium between Ti-bearing andradite and the Ti-Ts molecule can be expressed, it seems likely that Ti-bearing andradite acts as a sink for Ti, and thus coexisting clinopyroxene (Fig. 9) and biotite (Fig. 14) are comparatively starved of Ti. The exceptionally high Ti contents in early clinopyroxene-I phenocrysts in glimmerites do not contradict this observation, as micro-textures show that clinopyroxene-I is not in equilibrium with garnet and mica (Fig. 3e); they may simply have crystallized before garnet appeared on the liquidus.
The composition of clinopyroxene-I in glimmerites is distinct from that in most other rocks from the complex. In addition to being diopside-dominated (which does not, however, make clinopyroxene unique for the Tamazeght suite), clinopyroxene-I in glimmerites shows comparatively high proportions of all three Tschermak components and is extremely low in Na (Figs 8 and 9). We interpret these data as evidence that clinopyroxene-I from glimmerites crystallized from a melt source chemically distinct from the parental melt of the other rocks. Furthermore, given the high abundance of biotite in these rocks, this parental magma must have been exceptionally rich in potassium. Bouabdli & Liotard (1992) reported major and trace element data for the Tamazeght glimmerites and suggested that a kimberlitic magma was a likely parent to these rocks. However, Tamazeght glimmerites differ from typical kimberlites in the lack of Mg-rich ilmenite (instead, a Cr-bearing but relatively Mg-poor spinel phase is present), the atypical Na-poor and Ti-rich clinopyroxene compositions (see above) and the occurrence of Ti-rich and Cr-poor garnet. In any case, the presence of calcite and large amounts of phlogopite implies a potassium-rich and carbonated mantle source; such a source rock has already been proposed for the lamprophyre dyke swarm that cross-cuts the Tamazeght complex rocks (Bouabdli et al., 1988). In all, it seems likely that a carbonated amphibole-lherzolite was the source rock for the generation of the lamprophyric dykes, carbonatites and the Tamazeght glimmerites.
In the remaining rock units, very similar diopside-rich core compositions are observed. However, the various rock types document dissimilar evolutionary paths resulting in different phase assemblages and different phase compositions—a fact that is hard to reconcile with the assumption of a homogeneous parental melt source for all rock types. We thus argue that the various rock units in the Tamazeght complex possibly resulted from successive (or progressive) melting of a chemically and mineralogically heterogeneous mantle source. The generated melt batches were very similar in their XMg value, but in terms of their physico-chemical characteristics they were obviously distinct from each other. In turn, these differences in intensive parameters (fO2, aSiO2, aH2O) resulted in the stabilization of different phase assemblages (e.g. ilmenite or magnetite, Ti-andradite or titanite, amphibole or pyroxene, presence or absence of plagioclase) and these influenced the continuing chemical evolution of the melts from which they crystallized. The influence of plagioclase fractionation on the chemical evolution of the remaining melt and the composition of later crystallizing phases can be seen in the monzonitic rocks. Plagioclase crystallization (K/Na ratio << 1) increases the K/Na ratio of the melt. Amphibole in these rocks shows increasing K/(Na + K) ratios from core to rim and the minimum K/(Na + K) ratio of amphibole in the various monzonitic members increases with evolution from monzogabbros via monzonites to foid-monzosyenites. However, such a systematic evolution is not seen in the plagioclase-free rocks (Fig. 13).
Olivines from two samples of olivine-shonkinite have similar high XMg values of around 0·9 in their cores. Together with their relatively high Ni contents, this indicates that olivine in these rocks crystallized from a near-primary mantle melt. However, Ni, Ca and Mn contents are very different in these two samples (see above; Fig. 7). The concentration of such elements in olivine of a fixed XMg value is relatively independent of parameters such as, for example, oxygen fugacity, being mainly dependent on the composition of the crystallizing melt (Snyder & Carmichael, 1992). Thus, the observed heterogeneities in olivine from this rock type show that melt source heterogeneities may occur on a relatively small scale and these may later be documented not only in different lithologies but also in slight chemical variations of phases within a single rock type. However, other possibilities, such as mixing of different magma batches having distinct trace element compositions, cannot be excluded.
Quantitative constraints on the evolution of intrinsic parameters
At a given depth of intrusion, the parameters mainly governing the evolution of the Tamazeght magmas are T, fO2, aSiO2 and aH2O. The Al-in-hornblende barometer (e.g. Schmidt, 1992) commonly provides the only means of constraining the emplacement depth of plutonic complexes, such as the monzonitic group of the Tamazeght complex. However, Anderson & Smith (1995) showed that this barometer is significantly affected by T and fO2. Given this, the known restrictions of the application (Schmidt, 1992), the unusual Ti-rich composition of amphiboles and the strong zonation of plagioclase in the monzonitic rocks, the results need to be treated with extreme caution. However, a combination of the Al-in-hornblende barometer and amphibole–plagioclase thermometry (Blundy & Holland, 1990; Holland & Blundy, 1994) yields pressure estimates between 0·1 and 2·3 kbar (uncertainty of ± 0·6 kbar; Anderson & Smith, 1995) and equilibration temperatures between 790 and 860°C (uncertainty of ± 40°C; Holland & Blundy, 1994). These estimates seem reasonable, as they are in accordance with estimated conditions in upper crustal alkaline magma chambers elsewhere (e.g. Larsen & Sørensen, 1987; Potter et al., 2004). The presence of numerous pegmatites, roof pendants and contact-metamorphosed sediments (Salvi et al., 2000) indicates a shallow depth of intrusion. Consequently, we apply a pressure of 1 kbar in subsequent calculations.
In addition to amphibole–plagioclase thermometry, further constraints on minimum liquidus temperatures can be made by plotting the feldspar compositions on the temperature-dependent feldspar solvus of Fuhrman & Lindsley (1988) and by nepheline thermometry (after Hamilton, 1961). The results of the latter represent minimum liquidus temperatures, as a result of the known late- to postmagmatic equilibration of nepheline resulting in Si loss and hence lower estimates of temperature (Powell, 1978). Near-solidus temperatures for olivine-shonkinites can be calculated with the QUILF program (Frost & Lindsley, 1992; Lindsley & Frost, 1992; Andersen et al., 1993) from the assemblage olivine–clinopyroxene based on the Fe–Mg-exchange equilibrium between these two phases. For these calculations, average olivine core compositions and the most Fe-rich pyroxene core compositions were used, to minimize the possible effects of later diffusive re-equilibration, during which pyroxene tends to become enriched in Mg (Markl et al., 1998; Marks & Markl, 2001).
In addition to reaction (1) above, various phase equilibria allow us to constrain the T–fO2–aSiO2 evolution of the different rock types:
|
| (2) |
|
| (3) |
|
| (4) |
|
| (5) |
|
| (6) |
|
| (7) |
Phase diagrams were calculated using the GEOCALC software of Berman et al. (1987) and Liebermann & Petrakakis (1990) with the database of Berman (1988). Thermodynamic data for titanite and perovskite were taken from Robie & Hemingway (1995). End-member component activities were calculated using the solution model of Fuhrman & Lindsley (1988) for feldspar, the models of Wood (1979) and Green et al. (2007) for clinopyroxene, and a mixing-on-site model for nepheline. The activity of andradite was calculated after Cosca et al. (1986), and for Fe–Ti oxides either unit activities or, if necessary, the solution models implemented in QUILF were used. Titanite, perovskite, baddeleyite and zircon were treated as pure phases. Unit activity of SiO2 was referred to the standard state of the relevant pure SiO2 phase at P and T.
Estimation of equilibration temperatures
Two-feldspar thermometry using the temperature-dependent feldspar solvus of Fuhrman & Lindsley (1988) was applied to the monzonitic rocks and resulted in minimum liquidus temperatures between 750 and 900°C (Fig. 16).
Applying nepheline thermometry, maximum temperatures for the pyroxenites reach 1000°C. For monzonitic rocks and nepheline syenites, slightly lower but still high temperatures well above 800°C are indicated, as is the case for malignites. Nepheline from one of the hydrothermal veins yields temperatures of about 400–500°C. It is interesting to note that the evolution of nepheline compositions is different in pyroxenites compared with the other rock types (Fig. 17). In pyroxenites, the Ne content increases with decreasing SiO2 component, whereas in the other rock types, Ne content decreases, which was considered to indicate sub-solidus re-equilibration by Powell (1978). The nepheline trend observed in pyroxenites may be interpreted to reflect the primary evolution trend of nepheline in these rocks, evolving towards Ne-rich compositions during differentiation.
QUILF calculations for olivine-shonkinites resulted in equilibrium temperatures between 950 and 980°C, which represent near-solidus conditions.
Variations of aSiO2
The presence or absence of perovskite provides an important constraint on silica activity [equilibrium (2)]. Perovskite occurs only in some of the ultramafic rocks and the olivine-shonkinites, where it always exhibits rounded grain boundaries and rims of titanite (Fig. 5c). Titanite itself occurs (although rarely) as euhedral grains in equilibrium with andradite. In all other rock types, perovskite is absent. This implies that silica activity in these rocks was initially significantly lower than in the other rock types and the preserved textures indicate an increase of aSiO2 during the evolution of these rocks. For high temperatures above 800°C (as indicated by nepheline thermometry), the transformation of perovskite to titanite takes place at aSiO2 values of 0·1. An upper limit of aSiO2 is given by the absence of alkali feldspar according to equilibrium (5), which results in aSiO2 values between 0·5 and 0·75 (Fig. 19). A very similar evolution is observed in olivine-shonkinites (Fig. 19) and an upper limit of aSiO2 of about 0·8 is estimated for these rocks.
|
In monzonitic rocks, silica activity is constrained by equilibria (1), (2) and (4). Decreasing activity of anorthite (in plagioclase) and increasing activity of the Tschermak component (in clinopyroxene) displaces reaction (1) to lower values of aSiO2 (Fig. 19). Using the most An-rich plagioclase composition of the monzonitic group and core compositions of clinopyroxene, the calculated initial aSiO2 ranges between
0·5 for foid-monzosyenites and 0·75 for monzogabbros (at temperatures of 790–860°C as calculated above); similar aSiO2 values between 0·4 and 0·7 are calculated based on equilibrium (5). Additionally, a lower limit of aSiO2 of 0·25 for the early crystallization stage is given by the occurrence of zircon. Higher Tschermak components in the rims of clinopyroxene and decreasing An contents in plagioclase imply that aSiO2 dropped significantly during differentiation, which is in contrast to the evolutionary trend determined for the ultramafic rocks. Combined with the presence of titanite, a lower limit of aSiO2 of about 0·1 can be determined. The even lower aSiO2 indicated by the most An-poor plagioclase compositions can be explained by the fact that these compositions were no longer in equilibrium with clinopyroxene.
For nepheline syenites and malignites, equilibrium (5) was used to constrain aSiO2 (Fig. 19). Calculated initial aSiO2 ranges between 0·25 and 0·5 and was, therefore, initially lower than in the monzonitic rocks (assuming T = 800–900°C as indicated by nepheline thermometry). This is in accordance with the absence of plagioclase in the nepheline syenites, but similar amounts of Tschermak's component in clinopyroxene.
Constraints on the crystallization conditions of the hydrothermal veins are given by equilibrium (5) and the presence of zircon in some of these veins [equilibrium (3)]. At temperatures around 500°C (as indicated by nepheline thermometry), aSiO2 was around 0·1–0·2 (Fig. 19). In many of the investigated rocks, late-stage to hydrothermal alteration features are documented. Nepheline and sodalite are altered to cancrinite, whereas in malignites late-stage pure albite is observed. Based on mineral textures, the formation of agpaitic rocks, which crystallize eudialyte-group minerals, catapleite, låvenite and other Na–Zr-silicates, is also seen to occur at late-magmatic stages (Schilling et al., 2007). However, a detailed account of the late-stage to hydrothermal processes observed in the Tamazeght rocks is not the subject of this study and will be discussed in detail elsewhere.
Significance of clinopyroxene–garnet–Fe–Ti oxide–titanite textures: further constraints on fO2 and aSiO2 evolution
In the ultramafic rocks, oxygen fugacity is buffered by reaction (6) and was initially 2–5 log units above the FMQ buffer (Fig. 20). The replacement of magnetite by hematite indicates that fO2 rose subsequently to values around the hematite–magnetite buffer. The intersection of reaction (5) with the hematite–magnetite buffer (indicated by a grey dot in Fig. 20) is temperature-dependent and takes place at aSiO2 values of between 0·55 (at 600°C) and 0·8 (at 1000°C). Although no estimate on the temperature of this transformation is possible, it is implied that during this oxidation process aSiO2 simultaneously increased to higher values. The relative scarcity of magnetite and titanite in garnet-rich pyroxenites compared with all other rock types may indicate that Ti-rich garnet influences the stability of these phases. However, a detailed quantitative treatment of the relevant phase relations is not possible, as thermodynamic data for Ti-rich garnet are lacking.
|
In olivine-shonkinites oxygen fugacity is buffered by reactions (4) and (6) and similarly oxidized conditions (
FMQ = +2·5 to +4) to the ultramafic rocks are indicated (Fig. 20). When comparing pyroxenites with olivine-shonkinites, it seems surprising that despite the similarly oxidized crystallization conditions, their clinopyroxene evolutionary paths are distinct from each other (Fig. 8). In fact, clinopyroxene-I from garnet-poor olivine-shonkinite has higher Fe3+/(Fe2+ + Fe3+) ratios of 0·85–0·93 than clinopyroxene from garnet-rich pyroxenites (0·4–0·7). Additionally, Fe3+/(Fe2+ + Fe3+) values in the latter increase from the inner to the outer zone; Figs 3d and 8; Table 2). The observed irregular zonation patterns in clinopyroxene from pyroxenites (Fig. 3d) do not exclude the possibility of redistribution of several cations as a result of secondary re-equilibration and, therefore, the achievement of equilibrium cannot ultimately be assumed. In contrast, garnet in both rock types has similar Fe3+/(Fe2+ + Fe3+) ratios of 0·8–0·95, which is the same range as found for clinopyroxene from olivine-shonkinites. In olivine-shonkinites, the Fe3+/(Fe2+ + Fe3+) ratios for both minerals are similarly high and very similar to the Fe3+/(Fe2+ + Fe3+) ratio in garnet from pyroxenites, and it therefore seems likely that in olivine-shonkinites garnet and clinopyroxene co-crystallized or at least reflect the same evolutionary stage of the melt they crystallized from.
In the remaining rock types, titanite textures show some interesting variation. In monzogabbros, monzonites and amphibole-shonkinites, either most titanite occurs as rims around ilmenite, or subhedral titanite contains abundant inclusions of rounded relics of Fe–Ti oxides (Fig. 4b and c). Despite the fact that these rocks contain much more magnetite than ilmenite, these relics are in almost all cases ilmenite and the above-mentioned titanite rims almost exclusively occur around ilmenite. Primary magnetite grains do not seem to be affected by this reaction. In the more evolved rock types, which do not contain primary ilmenite (foid-monzosyenites and nepheline syenites), titanite is always euhedral and seems to have co-precipitated with clinopyroxene and magnetite (Fig. 4e). Additionally, it occurs much more commonly as euhedral inclusions in clinopyroxene and amphibole in these rocks. These textures imply that both oxygen fugacity and silica activity in these rocks were buffered by the schematic equilibrium (7).
Wones (1989), Xirouchakis et al. (2001a, 2001b) and Ryabchikov & Kogarko (2006) demonstrated that the stability of titanite is controlled by T, fO2, aSiO2 and the composition of the coexisting oxides and Fe–Mg silicates. If reaction (7) is calculated for rocks of the monzonitic group, which contain both magnetite and ilmenite (monzogabbros and monzonites), comparatively less oxidized conditions between 0·5 and 2·5 log units above the FMQ buffer are indicated (Fig. 20). Foid-monzosyenites, however, lack ilmenite, and the magnetite has a higher ulvöspinel content. Nevertheless, we calculated equilibrium (7) for these rocks, using the full range of magnetite and ilmenite compositions observed in the monzonitic group. In this case, the range of estimated fO2 expands towards relatively reduced conditions up to 1 log unit below the FMQ buffer. Using a similar approach for nepheline syenites and malignites, and taking into account the observed compositional variations of the phases (including pure ilmenite as a possible lower limit for oxygen fugacity), fO2 values around and significantly below the FMQ buffer (FMQ = +1 to –2) are estimated (Fig. 20).
|
SUMMARY AND CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTING AND PREVIOUS...FIELD RELATIONSPETROGRAPHYMINERAL COMPOSITIONSDISCUSSIONSUMMARY AND CONCLUSIONSAPPENDIX 1: CALCULATION OF...REFERENCES |
|---|
Our work on the various lithologies of the Tamazeght complex demonstrates that the combination of detailed petrographic studies with careful interpretation of mineral chemical variations not only reveals details of their petrological evolution, but also can be used to constrain the role of compositionally distinct mantle domains in their origin.
If the relevant phase assemblages indicative for intensive parameters are identified, quantification of the important phase equilibria is generally straightforward, if reliable thermodynamic data for the phases of interest are available. For all rocks, high temperatures between 750 and 1000°C for initial crystallization conditions are demonstrated. However, in terms of aSiO2 and fO2, the principal rock groups crystallized and evolved under markedly different conditions.
The most oxidized conditions were determined for the ultramafic rocks (FMQ up to +5) and olivine-shonkinites (FMQ = +2·5 to + 4). Both groups evolved from low initial aSiO2 values (possibly as low as 0·1) to higher values, reaching nepheline saturation in the ultramafic rocks (aSiO2 around 0·5) and alkali feldspar saturation in the olivine-shonkinites (aSiO2 = 0·5–0·8). In terms of their crystallization conditions and phase assemblages, the olivine-shonkinites share some similarities with pyroxenites, although the modal abundances for the phases (garnet, clinopyroxene, olivine, nepheline, feldspar) are very different. We conclude that these two lithologies might have a similar parental magma and could be linked to each by crystal–liquid differentiation processes.
For amphibole-shonkinites and monzonitic rocks, intermediate fO2 conditions are calculated (FMQ = +2·5 to –1). Their evolution with respect to aSiO2 is in the opposite sense to that indicated for the ultramafic rocks and olivine-shonkinites. The fractionation of plagioclase and clinopyroxene resulted in a decrease in aSiO2 from around 0·75 in the early stages to about 0·1, still during magmatic conditions.
For nepheline syenites and malignites, relatively low aSiO2 values of between 0·25 and 0·5 were calculated. Although the values for fO2 (FMQ = –2) have to be taken as rough estimates, they indicate rather reduced conditions of formation. The formation of hydrothermal veins occurred at temperatures around 500°C and low aSiO2 values between 0·1 and 0·2.
This study shows that the conditions of crystallization in alkaline plutonic rocks influence both the crystallizing mineral assemblage and the detailed chemical evolution of the phases during differentiation and cooling. Both in terms of fO2 and aSiO2, we found very different crystallization conditions for the various lithologies. In a general sense, high fO2 favours the crystallization of garnet. At intermediate fO2 titanite and magnetite are the preferred phases, whereas relatively low fO2 will lead to an enhanced stability of clinopyroxene and ilmenite. The evolution of aSiO2 during magmatic differentiation also shows contrasting trends. In the most primitive lithologies, low initial aSiO2 prevents the crystallization of alkali feldspar and plagioclase. In these rocks, aSiO2 increases during differentiation. In turn, in plagioclase- and alkali feldspar-bearing rocks, aSiO2 is buffered by the co-crystallization of Al-Tschermak-bearing clinopyroxene and nepheline, respectively, and indicates decreasing aSiO2 with progressive differentiation for some of the lithologies.
From the perspective of the origin of the large lithological variation found in the Tamazeght complex, in contrast to the findings of Kchit (1990) and Bouabdli et al. (1988) we suspect that the various rock types probably originated from distinct melt batches derived from a heterogeneous mantle source (heterogeneity caused by earlier metasomatic enrichment processes) or were produced from a stratified mantle source. However, crystal fractionation and accumulation processes may also play a role for some of the rocks. Models for mantle metasomatism include cryptic and patent mantle metasomatism (e.g. Wilshire & Shervais, 1975), and the vein-plus-wall-rock mantle model of Foley (1992) and others. The main differences between the various models are the proposed metasomatizing agents (melt vs fluid phase), the spatial effects of this metasomatism (locally vs universal) and the resulting mineralogical and geochemical changes [formation of hydrous phases vs enrichment in incompatible elements without other obvious changes; see review by Wilshire (1987)]. Regardless of the process, a later melting event in such a modified mantle source region will initially produce highly alkaline melts, which are strongly enriched in incompatible elements if the degree of melting is low enough. The higher the degree of melting, the less alkaline and more basalt-like the resultant melts will be. In this sense, the various lithologies in the Tamazeght complex might be interpreted either as representing variable degrees of melting of a cryptically metasomatized mantle domain or, if the vein and wall-rock model of Foley (1992) is applied, as reflecting melts of hydrous vein material, pristine wall-rocks and hybrid mixtures between them. The data presented here will serve as a basis for further geochemical and geochronological work, which is needed to resolve the origin of the Tamazeght rocks in detail.
|
APPENDIX 1: CALCULATION OF CLINOPYROXENE END-MEMBERS IN THE 10-COMPONENT SYSTEM DI–Hed–En–Fs–Aeg–Jd–Ti-Aeg–Fe-Ts–Ti-Ts–Al-Ts |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTING AND PREVIOUS...FIELD RELATIONSPETROGRAPHYMINERAL COMPOSITIONSDISCUSSIONSUMMARY AND CONCLUSIONSAPPENDIX 1: CALCULATION OF...REFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(1)Fe3+rest = Fe3+ – Aeg; (2) R2+ = Fe2+ + Mg2+ + Mn2+; (3) (Ti + Zr)rest = [(Ti + Zr) – (0·5 x Ti-Aeg)]; (4) Carest = Ca – Fe-Ts – Ti-Ts – Al-Ts; (5) XFe = (Fe2+ + Mn2+)/(Fe2+ + Mn2+ + Mg2+); (6) R2+rest = R2+ – 0·5 x Ti-Aeg.
|
ACKNOWLEDGEMENTS |
|---|
We acknowledge the support of Francois Fontan, Pierre Monchoux and Stefano Salvi (Toulouse, France), who gave us interesting insights into their earlier work on Tamazeght and encouraged us to investigate this intrusive complex in more detail. Ali Bajja (Marrakesh, Morocco) is thanked for his co-operation, which facilitated the field trip in May 2006. We also thank the citizens of the small Berber village of Anougal for their hospitality, and Boujemaa Boudaoud (Azrou, Morocco) for being our guide in the High Atlas Mountains and for supplying important infrastructure in the field. Florian Neukirchen is thanked for his assistance in the field. We also thank Sebastian Staude for his help with reflected light microscopy and Sylvia Mettasch for her interest in this work and for careful and detailed petrographic work on some of the samples. Funding for this work by the Deutsche Forschungsgemeinschaft (grant Ma 2135/11-1) and the Natural Sciences and Engineering Research Council of Canada (IMC: Discovery grant funds) is gratefully acknowledged. The constructive comments of M. Wilson, R. Mitchell, A. Chakhmouradian and one anonymous reviewer are greatly appreciated.
*Corresponding author. E-mail: michael.marks{at}uni-tuebingen.de
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTING AND PREVIOUS...FIELD RELATIONSPETROGRAPHYMINERAL COMPOSITIONSDISCUSSIONSUMMARY AND CONCLUSIONSAPPENDIX 1: CALCULATION OF...REFERENCES |
|---|
Agard J. Les carbonatites et les roches à silicated du massif de roches alcalines du Tamazert (Haut Atlas de Midelt, Maroc) et les problémes de leurs genéses. Proceedings of the 21st International Geological Congress, Norden, Norway (1960) 13:293–303.
Agchmi M. Les carbonatites filoniennes de l’Oued Tamazeght et leurs relations avec des métasyénites (Haut Atles de Midelt). (1984) Toulouse: Université Paul Sabatier. 106. Ph.D. Thèse.
Al-Haderi M, Tayebi M, Bouabdli A, El-Hanbali M. Chronologie et conditions de mise en place des différents faciès pétrographiques du complexe alcalin de Tamazert (Haut-Atlas de Midelt, Maroc). Africa Geoscience Review (1998) 5:159–171.
Andersen DJ, Lindsley DH, Davidson PM. QUILF: a PASCAL program to assess equilibria among Fe–Mg–Mn–Ti oxides, pyroxenes, olivine, and quartz. Computers and Geosciences (1993) 19:1333–1350.[CrossRef]
Anderson JL, Smith DR. The effects of temperature and fO2 on the Al-in-hornblende barometer. American Mineralogist (1995) 80:549–559.[Abstract]
Armstrong JT. Quantitative elemental analysis of individual microparticles with electron beam instruments. In: Electron Probe Quantitation—Heinrich KFJ, Newbury DE, eds. (1991) New York: Plenum. 261–315.
Arzamastsev A, Bea F, Arzamastseva LV, Montero P. Proterozoic Gremyakha–Vyrmes polyphase massif, Kola Peninsula: An example of mixing basic and alkaline mantle melts. Petrology (2006) 14:384–414.
Beccaluva L, Barbieri M, Born H, Brotzu P, Coltorti M, Conte A, Garbarino C, Gomes CB, Macciotta G, Morbidelli L, Ruberti E, Siena F, Traversa G. Fractional crystallization and liquid immiscibility processes in the alkaline-carbonatite complex of Juquia (São Paulo, Brazil). Journal of Petrology (1992) 33:1371–1404.
Berman R. Internally consistant thermodynamic data for minerals in the system Na2O–K2O–CaO–MgO–FeO–Fe2O3–Al2O3–SiO2–TiO2–H2O–CO2. Journal of Petrology (1988) 29:445–522.
Berman RG, Brown TH, Perkins EH. Geo-Calc; a software for calculation and display of P–T–X phase diagrams. American Mineralogist (1987) 72:861–862.[Abstract]
Blundy JD, Holland TJB. Calcic amphibole equilibria and a new amphibole–plagioclase geothermometer. Contributions to Mineralogy and Petrology (1990) 104:208–224.[CrossRef][Web of Science]
Bouabdli A, Liotard J.-M. Affinité kimberlitique des lamprophyres ultrabasiques du massif carbonatitique de Tamazert (Haut Atlas marocain). Comptes Rendus de l’Académie des Sciences, Série II (1992) 314:351–357.
Bouabdli A, Liotard J.-M. Rôle des fluides carbonatés dans le contrôle de la composition des clinopyroxènes ferrisodiques: exemple du massif carbonatitique de Tamazert (Haut-Atlas de Midelt, Maroc). Africa Geoscience Review (1999) 6:291–300.
Bouabdli A, Dupuy C, Dostal J. Geochemistry of Mesozoic alkaline lamprophyres and related rocks from the Tamazert massif, High Atlas (Morocco). Lithos (1988) 22:43–58.[CrossRef][Web of Science]
Chakhmouradian AR, McCammon C. Schorlomite: a discussion of the crystal chemistry, formula, and inter-species boundaries. Physics and Chemistry of Minerals (2005) 32:277–289.[CrossRef][Web of Science]
Chakhmouradian AR, Mitchell RH. The mineralogy of Ba- and Zr-rich alkaline pegmatites from Gordon Butte, Crazy Mountains (Montana, USA): comparisons between potassic and sodic agpaitic pegmatites. Contributions to Mineralogy and Petrology (2002) 143:93–114.[Web of Science]
Cosca MA, Moecher DP, Essene EJ. Activity–composition relations for the join grossular–andradite and application to calc-silicate assemblages. Geological Society of America, Abstracts with Programs (1986) 18:572.
Coulson IM. Evolution of the North Qôroq centre nepheline syenites, South Greenland: alkali–mafic silicates and the role of metasomatism. Mineralogical Magazine (2003) 67:873–892.
Coulson IM, Russell JK, Dipple GM. Origins of the Zippa Mountain pluton: a Late Triassic, arc-derived, ultrapotassic magma from the Canadian Cordillera. Canadian Journal of Earth Sciences (1999) 36:1415–1434.
Deer WA, Howie RA, Zussman J. An Introduction to the Rock-forming Minerals (1992) 2nd. Harlow: Longman. 696.
Dunworth EA, Bell K. The Turiy massif, Kola peninsula, Russia: Isotopic and geochemical evidence for multi-source evolution. Journal of Petrology (2001) 42:377–405.
Dunworth EA, Wilson M. Olivine melilitites of the SW German Tertiary volcanic province: mineralogy and petrogenesis. Journal of Petrology (1998) 39:1805–1836.[CrossRef][Web of Science]
Foley S. Vein-plus-wall-rock melting mechanism in the lithosphere and the origin of potassic alkaline magmas. Lithos (1992) 28:435–453.[CrossRef][Web of Science]
Frost BR, Lindsley DH. Equilibria among Fe–Ti oxides, pyroxenes, olivine, and quartz: Part II. Application. American Mineralogist (1992) 77:1004–1020.[Abstract]
Fuhrman ML, Lindsley DH. Ternary-feldspar modeling and thermometry. American Mineralogist (1988) 73:201–205.[Abstract]
Green E, Holland T, Powell R. An order–disorder model for omphacitic pyroxenes in the system jadeite–diopside–hedenbergite–acmite, with application to eclogitic rocks. American Mineralogist (2007) 92:1181–1189.
Halama R, Vennemann T, Siebel W, Markl G. The Grønnedal–Ika carbonatite–syenite complex, south Greenland: Carbonatite formation by liquid immiscibility. Journal of Petrology (2005) 46:191–217.
Hamilton DL. Nephelines as crystallisation temperature indicators. Journal of Geology (1961) 69:321–329.[Web of Science]
Harmer RE. The petrogenetic association of carbonatite and alkaline magmatism: constraints from the Spitskop Complex, south Africa. Journal of Petrology (1999) 40:525–548.[CrossRef][Web of Science]
Holland T, Blundy J. Non-ideal interactions in calcic amphiboles and their bearing on amphibole–plagioclase thermometry. Contributions to Mineralogy and Petrology (1994) 116:433–447.[CrossRef][Web of Science]
Jiménez-Millán J, Jiménez-Espinosa R, Velila N, Chica-Olmo M. The use of correspondence analysis and principal component analysis as a tool for deducing metamorphic mineral crystallization in Mn-rich lithologies. Terra Nova (1994) 6:267–273.[CrossRef][Web of Science]
Jones AP, Peckett A. Zirconium-bearing aegirines from Motzfeldt, South Greenland. Contributions to Mineralogy and Petrology (1980) 75:251–255.[CrossRef][Web of Science]
Kchit A. Le plutonisme alcalin du Tamazeght (Haut Atlas de Midelt, Maroc). (1990) Toulouse: Université Paul Sabatier. 287. Ph.D. Thèse.
Khadem Allah B, Monchoux P, Fontan F, Béziat D, Kadar M. Clinopyroxènes des syénites néphéliniques du Tamazeght (Haut-Atlas de Midelt, Maroc). Comptes Rendus de l’ Académie des Sciences, Série IIA (1996) 323:841–847.
Khadem Allah B, Fontan F, Kadar M, Monchoux P, Sørensen H. Reactions between agpaitic nepheline syenitic melts and sedimentary carbonate rocks, exemplified by the Tamazeght complex, Morocco. Geochemistry International (1998) 36:569–581.
Khomyakov AP. Mineralogy of Hyperagpaitic Alkaline Rocks (1995) Oxford: Clarendon Press. 222.
Klein JL, Harmand C. Le volcanisme de la région de Zebzate: age relations avec le complex alcalin à carbonatites du Tamazert (Haut-Atlas de Midelt, Maroc). 110e Congrès National de la Société Sav. Montpellier, Sci. Fascicule VII (1985) 147–152.
Korobeinikov AN, Laajoki K. Petrological aspects of the evolution of clinopyroxene composition in the intrusive rocks of the Lovozero alkaline massif. Geochemistry International (1994) 31:69–76.
Kramm U, Kogarko LN. Nd and Sr isotope signatures of the Khibina and Lovozero agpaitic centres, Kola Province, Russia. Lithos (1994) 32:225–242.[CrossRef][Web of Science]
Larsen LM. Clinopyroxenes and coexisting mafic minerals from the alkaline Ilímaussaq intrusion, south Greenland. Journal of Petrology (1976) 17(2):258–290.
Larsen LM, Sørensen H. The Ilimaussaq intrusion—progressive crystallization and formation of layering in an agpaitic magma. In: Alkaline Igneous Rocks. Geological Society, London, Special Publications—Fitton JG, Upton BGJ, eds. (1987) 30:473–488.
Laville E. Role des décrochements dans le mécanisme de formation des bassins déffondrement du Haut-Atlas marocain au cours des temps triasique et liasiques. Bulletin de la Société Géologique de France (1981) 23:303–312.[Web of Science]
Laville E, Harmand C. Evolution magmatique et tectonique du bassin intracontinental mésozoique du Haut-Atlas (Maroc): un modèle de mise en place synsédimentaire de massifs ‘anorogeniques’ liés à des dérochements. Bulletin de la Société Géologique de France (1982) 7:221–227.
Leake BE, Woolley AR, Arps CES, et al. Nomenclature of amphiboles: Report of the Subcommittee on Amphiboles of the International Mineralogical Association, Commission on New Minerals and Mineral Names. American Mineralogist (1997) 82:1019–1037.[Abstract]
Le Maitre RW, ed. Igneous Rocks; A classification and Glossary of Terms; Recommendations of the International Union of Geological Sciences Subcommission on the Systematics of Igneous Rocks (2002) Cambridge: Cambridge University Press. 236.
Liebermann J, Petrakakis K. TWEEQU thermobarometry, analysis of uncertainties and applications to granulites from western Alaska. Canadian Mineralogist (1990) 29:857–887.
Lindsley DH, Frost BR. Equilibria among Fe–Ti oxides, pyroxenes, olivine, and quartz: Part I. Theory. American Mineralogist (1992) 77:987–1003.[Abstract]
Mann U, Marks MAW, Markl G. Influence of oxygen fugacity on mineral compositions in peralkaline melts: the Katzenbuckel volcano, Southwest Germany. Lithos (2006) 91:262–285.[CrossRef][Web of Science]
Markl G, Frost BR, Bucher K. The origin of anorthosites and related rocks from the Lofoten Islands, northern Norway: I. Field relations and estimation of intrinsic variables. Journal of Petrology (1998) 39:1425–1452.[CrossRef][Web of Science]
Markl G, Marks M, Schwinn G, Sommer H. Phase equilibrium constraints on intensive crystallization parameters of the Ilimaussaq Complex, South Greenland. Journal of Petrology (2001) 42:2231–2258.
Marks M, Markl G. Fractionation and assimilation processes in the alkaline augite syenite unit of the Ilimaussaq Intrusion, South Greenland, as deduced from phase equilibria. Journal of Petrology (2001) 42:1947–1969.
Marks M, Markl G. Ilimaussaq ‘en miniature’: closed-system fractionation in an agpaitic dyke rock from the Gardar province, south Greenland. Mineralogical Magazine (2003) 67:893–919.
Marks M, Vennemann T, Siebel W, Markl G. Nd-, O-, and H-isotopic evidence for complex, closed-system fluid evolution of the peralkaline Ilimaussaq Intrusion, South Greenland. Geochimica et Cosmochimica Acta (2004) 68:3379–3395.[CrossRef][Web of Science]
Mitchell RH, Liferovich RP. Subsolidus deuteric/hydrothermal alteration of eudialyte in lujavrite from the Pilansberg alkaline complex, south Africa. Lithos (2006) 91:353–372.
Mitchell RH, Platt RG. Mineralogy and petrology of nepheline syenites from the Coldwell alkaline complex, Ontario, Canada. Journal of Petrology (1982) 23:186–214.
Mitchell RH, Vladykin NV. Compositional variation of pyroxene and mica from the little Murun ultrapotassic complex, Aldan Shield, Russia. Mineralogical Magazine (1996) 60:907–925.[Abstract]
Morikiyo T, Takano K, Miyazaki T, Kagami H, Vladykin NV. Sr, Nd, C and O isotopic compositions of carbonatite and peralkaline silicate rocks from the Zhidoy complex, Russia: evidence for binary mixing, liquid immiscibility and a heterogeneous mantle source origin. Journal of Mineralogical and Petrological Sciences (2000) 95:162–172.[CrossRef]
Mourtada S, Le Bas MJ, Pin C. Pétrogenèse des magnésio-carbonatites du complexe de Tamazert (Haut Atlas marocain). Comptes Rendus de l’ Académie des Sciences (1997) 325:559–564.
Oberti R, Ungaretti L, Cannillo E, Hawthorne FC. The behaviour of Ti in amphiboles: I. Four- and six-coordinate Ti in richterite. European Journal of Mineralogy (1992) 4:425–439.
Potter J, Rankin AH, Treloar PJ. Abiogenic Fischer–Tropsch synthesis of hydrocarbons in alkaline igneous rocks; fluid inclusion, textural and isotopic evidence from the Lovozero complex, N.W. Russia. Lithos (2004) 75:311–330.[CrossRef][Web of Science]
Powell M. The crystallisation history of the Igdlerfigssalik nepheline syenite intrusion, Greenland. Lithos (1978) 11:99–120.[CrossRef][Web of Science]
Ragland PC, Conley JF, Parker WC, Van Orman JA. Use of principal components analysis in petrology: an example from the Martinsville igneous complex, Virginia, U.S.A. Mineralogy and Petrology (1997) 60:165–184.[CrossRef][Web of Science]
Robie RA, Hemingway BS. Thermodynamic properties of minerals and related substances at 298·15 K and 1 bar (105 Pascals) pressure and at higher temperatures. US Geological Survey Bulletin (1995) 2131:461.
Ryabchikov ID, Kogarko LN. Magnetite compositions and oxygen fugacities of the Khibina magmatic system. Lithos (2006) 91:35–45.[CrossRef][Web of Science]
Salvi S, Williams-Jones AE. The role of hydrothermal processes in the granite-hosted Zr, Y, REE deposit at Strange Lake, Quebec/Labrador: evidence from fluid inclusions. Geochimica et Cosmochimica Acta (1990) 54:2403–2418.[CrossRef][Web of Science]
Salvi S, Fontan F, Monchoux P. Hydrothermal mobilization of high field strength elements in alkaline igneous systems: Evidence from the Tamazeght Complex, (Morocco). Economic Geology (2000) 95:559–576.
Salvi S, Tagirov B, Moine BN. Hydrothermal mineralization of Zr and other ‘immobile elements’: field evidence and experimental constraints. Water–Rock Interaction 10 (2001) 745–748.
Schilling J, Marks MAW, Markl G. What governs the transition from miaskitic to agpaitic assemblages in peralkaline rocks? Geochimica et Cosmochimica Acta 71 (15S) (2007) A889.
Schmidt MW. Amphibole composition in tonalite as a function of pressure: an experimental calibration of the Al-in-hornblende barometer. Contributions to Mineralogy and Petrology (1992) 110:304–310.[CrossRef][Web of Science]
Snyder DA, Carmichael ISE. Olivine–liquid equilibria and the chemical activities of FeO, NiO, Fe2O3 and MgO in natural basic melts. Geochimica et Cosmochimica Acta (1992) 56:303–318.[CrossRef][Web of Science]
Sørensen H. Agpaitic nepheline syenites: a potential source of rare elements. Applied Geochemistry (1992) 7:417–427.[CrossRef][Web of Science]
Sørensen H. The agpaitic rocks—an overview. Mineralogical Magazine (1997) 61:485–498.[Abstract]
Tisserant D, Thuizat R, Agard J. Données géochronologiques sur le complexe de roches alcalines du Tamazeght (Haut Atlas de Midelt, Maroc). Bureau des Recherches Géologiques et Minière Bulletin, Section 2 (1976) 3:279–283.
Vuorinen JH, Hålenius U, Whitehouse MJ, Mansfeld J, Skelton ADL. Compositional variations (major and trace elements) of clinopyroxene and Ti-andradite from pyroxenite, ijolite and nepheline syenite, Alnö Island, Sweden. Lithos (2005) 81:55–77.[CrossRef][Web of Science]
Wagner C, Velde D, Mokhtari A. Sector-zoned phlogopites in igeneous rocks. Contributions to Mineralogy and Petrology (1987) 96:186–191.[CrossRef][Web of Science]
Wilshire HG. A model of mantle metasomatism. In: Mantle Metasomatism and Alkaline Magmatism. Geological Society of America, Special Papers—Morris EM, Pasteris JD, eds. (1987) 215:47–60.
Wilshire HG, Shervais JW. Al-augite and Cr-diopside ultramafic xenoliths in basaltic rocks from western United States; Structural and textural relationships. Physics and Chemistry of the Earth (1975) 9:257–272.[CrossRef]
Wones DR. Significance of the assemblage titanite + magnetite + quartz in granitic rocks. American Mineralogist (1989) 74:744–749.[Abstract]
Wood BJ. Activity–composition relationships in Ca(Mg,Fe)Si2O6–CaAl2SiO6 clinopyroxene solid solution. American Journal of Science (1979) 279:854–875.
Xirouchakis D, Lindsley DH, Andersen DJ. Assemblages with titanite (CaTiOSiO4), Ca–Mg–Fe olivine and pyroxenes, Fe–Mg–Ti oxides, and quartz: Part I. Theory. American Mineralogist (2001a) 86:247–253.
Xirouchakis D, Lindsley DH, Frost BR. Assemblages with titanite (CaTiOSiO4), Ca–Mg–Fe olivine, and pyroxenes, Fe–Mg–Ti oxides, and quartz: Part II. Application. American Mineralogist (2001b) 86:254–264.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Berger, N. Ennih, J.-C. C. Mercier, J.-P. Liegeois, and D. Demaiffe The role of fractional crystallization and late-stage peralkaline melt segregation in the mineralogical evolution of Cenozoic nephelinites/phonolites from Saghro (SE Morocco) Mineralogical Magazine, June 15, 2009; 73(1): 59 - 82. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Schilling, M. A.W. Marks, T. Wenzel, and G. Markl RECONSTRUCTION OF MAGMATIC TO SUBSOLIDUS PROCESSES IN AN AGPAITIC SYSTEM USING EUDIALYTE TEXTURES AND COMPOSITION: A CASE STUDY FROM TAMAZEGHT, MOROCCO Can Mineral, April 1, 2009; 47(2): 351 - 365. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||