The Alnö alkaline and carbonatitic complex, east central Sweden – a petrogenetic study

Jaana Hode Vuorinen

Department of Geology and Geochemistry, Stockholm University, SE-106 91 Stockholm, Sweden

The doctoral thesis consists of a summary and the following papers:

Paper IHode Vuorinen J., Hålenius U., Whitehouse M.J., Mansfeld J. and Skelton A.D.L. (2005). Compositional variations (major and trace elements) of clinopyroxene and Ti-andradite from pyroxenite, ijolite and nepheline syenite, Alnö Island, Sweden. Lithos 81, 55-77. Reprinted with permission.

Paper IIHode Vuorinen J. and Hålenius U. (2005). Nb-, Zr- and LREE-rich titanite from the Alnö alkaline complex: crystal chemistry and its importance as a petrogenetic indicator. Lithos, in press. Reprinted with permission.

Paper IIIHode Vuorinen J. and Skelton A.D.L. (2004). Origin of silicate minerals in carbonatites from Alnö Island, Sweden – Magmatic crystallisation or wall rock assimilation? Terra Nova 16, 210-215. Reprinted with permission.

Paper IVSkelton. A., Hode Vuorinen, J., Arghe F., and Fallick A. Chromatographic modeling of fluid-rock interaction across a carbonatite-gneiss transition zone, Alnö, Sweden. Re-submitted to Contributions to Mineralogy and Petrology.

Paper VHode Vuorinen J., Skelton, A., Andersen T., Taylor, P. and Claesson S. Geochemical (major and trace elements, Nd-, Sr- and Pb-isotopes) characteristics of the Alnö alkaline complex, Sweden – Petrogenetic implications. In manuscript.

Field work, sampling and petrographic studies in all papers were carried out by the author, except in Paper IV where the petrographic work was made by Alasdair Skelton and Fredrik Arghe (Department of Geology and Geochemistry, Stockholm University). The author performed the electron microprobe analyses in Papers I, II and III with assistance from Hans Harrysson (Uppsala University) and Paula McDade and Peter G. Hill (University of Edinburgh, Scotland). SIMS analyses of REE in minerals for Paper I were performed by the author with assistance from Martin Whitehouse (SMNH), whereas Joakim Mansfeld (Department of Geology and Geochemistry, Stockholm University) is responsible for the LA-ICP-MS analyses in Paper I. Stable isotope (C and O) analyses on calcite samples in Paper IV were done by Klara Hajnal, Department of Geology and Geochemistry, Stockholm University. O-istope analyses on alkali feldspars in Paper IV were performed by A. Fallick, SUERC, East Kilbride, Scotland. Sample selection and preparation for analyses in Paper IV was done by the author. Whole rock major and trace element analyses for Paper V were carried out by ACME Labs, Canada. Laboratory work and analyses of Sr and Nd whole rock isotope composition was done by the author with assistance by Marina Fischerström and Hans Schöberg (SMNH). The author is responsible for all interpretations in Papers I, II and III in collaboration with co-authors. In Paper V, the Pb-isotope data interpretations are made by T. Andersen. Major and trace element modelling in Paper V was made in collaboration with A. Skelton. All interpretations in Paper V are by the author and co-workers. In Paper IV, the author answers for the mineralogical and petrological descriptions and has provided input to the discussion and interpretations.

3

IntroductionIn 1960-61 the dispute concerning the existence and definition of carbonatite magmas was settled as an eruption of sodium-carbonatite lava flows was observed at the volcano Oldoinyo Lengai in Tanzania. Years prior to this, the concept of carbonatite magmas had been proposed (Eckermann, 1948), but was not fully accepted by the scientific community. At the same time as the eruptions at Oldoinyo Lengai, the first successful experimental studies were also carried out on carbonatitic melts, seemingly confirming the possibility of their existence. The events that occurred in 1960 were first described by Dawson (1962a and 1962b) after which Oldoinyo Lengai has taken on a special significance in volcanological and petrological research in general, and more specifically in the study of carbonatites. Consequently, many studies and models of carbonatite genesis constructed from observations of the lavas at Oldoinyo Lengai have been argued to be generally applicable to carbonatite genesis. This is despite the fact that the composition of the lavas erupted at Oldoinyo Lengai is unique with respect to the >400 carbonatites recognized worldwide. For example, the carbonatite lavas of Oldoinyo Lengai are alkali-rich in contrast to most other known carbonatites which are either Ca- and/or Mg-rich. This discrepancy remains an issue of controversy and the nature of carbonatite formation remains uncertain, other than that they are mantle-derived and spatially associated with alkaline silicate rocks. Carbonatite petrologists remain undecided as to whether (1) carbonatites are derivative melts formed by unmixing or fractionation from a mantle-generated alkaline silicate magma or (2) carbonatite are mantle-derived magmas unrelated to an associated silicate melt since the time of extraction from the mantle. The study presented in this thesis aims to further our understanding of the genesis of alkaline silicate rocks and carbonatites in general by studying a particular complex in detail (i.e. the Alnö complex, east central Sweden). This is important to petrologists, because carbonatites provide a chemical “window” through which we can examine the composition of the mantle. For example, due to the unique geochemical characteristics of carbonatites (and associated silicate rocks), e.g. high Sr-

and Nd-concentrations, they are believed to buffer effectively against crustal contamination processes and might thus preserve the isotope composition of their mantle source(s) (Bell & Blenkinsop, 1987). Since most carbonatite occurrences are limited to extensional continental tectonic settings, e.g. East African Rift System, the Rhinegraben in Germany and the Gardar rift in south Greenland, they may also provide chemical information relating to continental break-up. In addition, the intrusion ages for carbonatites spread all the way back to Achaean times (Bell & Tilton, 2002). This makes it possible not only to characterize the recent composition of the sub-continental lithospheric (and astenospheric) mantle, but it also presents a unique opportunity to monitor the evolution of the mantle all the way back to the Achaean. This is a time perspective that is unmatched by other, mantle derived rocks such as MORB (Mid-Ocean Ridge Basalts).

Petrological and mineralogical definitionsAlkaline silicate rocks have, despite being very rare in the Earths crust, given rise to a bewildering range of names in the igneous nomenclature, reflecting the geochemical and mineralogical complexity of this group of rocks. In older literature, rock names have often been given with reference to type locality as for example alnöite, alvikite, beforsite (Alnö and surroundings), fenite, melteigite, hollaite, sövite (Fen, Norway), ijolite (Iivaara, Finland) and jacupirangite (Jacupiranga, Brazil). Many of the older names are now obsolete, and a general consensus regarding the essential rock names has been reached and is presented by Le Maitre et al (2002). Variable use of the term alkaline has also flourished in the literature since the beginning of the study of such rocks and is excellently summarized by Sörensen (1974). The term alkaline should only be applied to rocks with Na and K in excess of that needed to form alkali feldspars (Si-saturated to -oversaturated systems) or feldspathoids (Si-undersaturated systems). This leads to mafic minerals, i.e. clinopyroxenes and amphiboles, often with appreciable Na- and/or K-contents. Rocks with Na2O+K2O/Al2O3>1 should be termed peralkaline, a term which should not be confused with, or used synonymous with

4

agpaitic. The term agpaitic should only be applied to peralkaline nepheline syenites with rare complex Zr and Ti silicates such as

eudialyte Na15Ca6(Fe2+,Mn2+)3Zr3Si25O73(O,OH,H2O)3(OH,Cl)2

and mosandrite (Ca,Na)3(Ca,Ce)4(Ti,Nb,Al,Zr)(Si2O7)(O,F)4] Of the more common alkaline silicate rocks occurring world-wide, the ijolite series rocks and nepheline syenites are most frequently encountered. The ijolite series (melteigite-ijolite-urtite) is defined as being essentially clinopyroxene-nepheline rocks. Following the classification of Le Maitre (2002) a melteigite contains less than 30% nepheline; an ijolite between 30 and 70 % nepheline and an urtite is composed of more than 70% nepheline. Ijolite and urtite may sometimes carry small amounts of alkali feldspars and as these increase in quantity, the rocks pass into nepheline syenite (via malignite = mela-nepheline syentie). Olivine is sometimes found as an important constituent in melteigite (Dunworth and Bell, 2001; Verhulst et al, 2000). Accessory phases in all rocks of the ijolite series and in nepheline syenite include Ti-rich andradite, which locally may become a major constituent, titanite, apatite, perovskite, phlogopite/biotite, iron-titanium oxides and Fe-sulfides. Interstitial calcite is usually present and may become a major phase. According to recommendations by the IUGS (Le Maitre, 2002), ijolite series rocks in which calcite increase to more than 10 vol% (but is less than 50 vol%) should be called calcite- or carbonatitic ijolites. Late-stage and secondary phases in ijolite series rocks and nepheline syenite include cancrinite, natrolite and sodalite. Alnöite and kimberlitic alnöite are potassic ultramafic dyke rocks with clinopyroxene (diopside-augite), olivine and phlogopite as chief phenocryst constituents. Calcite and sometimes melilite are abundant in the groundmass along with the already mentioned minerals. Garnet and perovskite may also be found. Melilitolites are rare on a worldwide scale but are important components of many alkaline-carbonatite complexes and play an important role in the understanding of

the genesis of SiO2-poor magmas. Their modal mineralogies are highly variable with modal melilite varying from 10 vol% (melilitolite) to more than 65 vol% (ultramelilitolite). Some melilitolites are essentially oxide-melilite rocks whereas others contain up to 50 vol% of primary carbonates (Dunworth & Bell, 1998). For a comprehensive review of melilitolite mineralogy and classification schemes, the reader is referred to the work of Dunworth & Bell (1998). Carbonatites are defined as igneous carbonate rocks in which the modal percentage of carbonates is more than 50 % (Le Maitre, 2002). If the dominating carbonate is calcite the rock is a calcite carbonatite (calciocarbonatite), which is called sövite if it is coarse grained and alvikite (Eckermann, 1948) if fine grained. However, although both alvikite and sövite are calciocarbonatites, they are distinctly different chemically with respect to trace elements (Le Bas, 1999). Mg-rich carbonatites are dolomite-bearing and sometimes referred to as beforsite although this term is becoming somewhat obsolete. The term dolomite cabonatite or magnesiocarbonatite should be applied to those rocks. Fe-rich carbonatites are a bit more difficult to recognize as Fe in carbonatites does not only reside within carbonates but is also contained within silicates (which may be derived from wall-rock or may have crystallized in the carbonatite magma) or magnetite and/or hematite which is an important cumulus phase in some carbonatites. This has lead to some ambiguities as to how to correctly classify this rather rare type of carbonatite. Gittins & Harmer (1997) recognized the difficulties and recommended that the term ferrocarbonatite should be used only when the modal mineralogy of the rocks is not known (i.e. the term should be used solely in a chemical sense), whereas terms as siderite carbonatite, magnetite-calcite carbonatite etc. should be used when the modal mineralogy is known. Natrocarbonatite is a fine-grained carbonatite lava composed chiefly of nyerereite [(Na,K)2Ca(CO3)2] and gregoryite [(Na,K)2CO3]. This is an extremely rare type of carbonatite and its occurrence is at present restricted to the volcano Oldoinyo Lengai in Tanzania.

5

Carbonatite petrogenesis – a literature review

A majority of all known carbonatites worldwide are associated with alkaline silicate and/or ultramafic rocks, i.e. phonolites, nephelinites, melilitites and kimberlites, or their plutonic equivalents, but carbonatites apparently occurring without any associated silicate rocks, are also known. The almost ubiquitous association however, indicates that it is difficult to separate the origin of carbonatites from the origin of the associated silicate rocks. Still the relationships between the different rock types are not completely understood and a number of theories exist regarding their origin. Bell et al (1998) note that the diverse associations that exist (i.e. carbonatite–nephelinite/ijolite, carbonatite–phonolite/nepheline syenite, carbonatite–syenite, carbonatite–kimberlite, carbonatite–lamprophyre, carbonatite–melilitite/melilitolite, carbonatite–pyroxenite and dolomitic carbonatite–olivine-rich ultrabasites), argues against any simple model of magma generation. This raises the question as to whether the associated rocks are generated from a common parental magma, and if so by what processes, or if they are derived independently of one another. Essentially, three models for

carbonatite genesis are discussed, two of which argue that carbonatite magmas are derived from a carbonated silicate parent:

(a) carbonatite magma is produced through immiscible separation from a carbonated silicate magma which can be nephelinitic or phonolitic (with or without preceding crystal fractionation) (Kjarsgaard & Hamilton, 1988, 1989a; Kjarsgaard & Peterson, 1991; Kjarsgaard et al, 1995; Church & Jones, 1995 among others)

(b) carbonatite magma is primary and derived through melting of a carbonated peridotite (Wyllie & Huang, 1975, 1976a; Bailey, 1993; Lee & Wyllie, 1997a; Wyllie & Lee, 1998; Harmer, 1999; Harmer & Gittins, 1998)

(c) carbonatite magma is produced through fractional crystallization of a carbonated silicate magma (Cooper & Reid, 1998; Lee & Wyllie, 1994; Otto & Wyllie, 1993)

The three models are not thought to be mutually exclusive and the genesis of carbonatite may reflect the combined effects of one or more of these processes.

Figure 1. Comparison of miscibility gaps (curves at 1.0 and 2.5 GPa), liquid compositions from carbonated lherzolite (dark grey field), natural rock compositions (AB, EN, NA and SV) and possible magmatic processes (paths 1-4) in silicate-carbonate systems at mantle to crustal conditions. The miscibility gaps are at 1200°C and show the effect of pressure (a) and composition (b). NA = natro-carbonatite, SV = average sövite, AB = alkali basalts and EN = evolved nephelinites. Arrows are: (1) progressive melt-ing path for carbonated peridotite, (2) cooling path of carbonated, undersaturated mantle silicate melt through the crystallization of amphibole (pargasite), (3) fractional crystallization path of mantle carbonatite liquid to NA and (4) metasomatic trend for the formation of weh-rlite and average sövite by successive reaction of lherzolite with more magnesian carbonatite liquid. From Lee & Wyllie (1997a).

ENLs

Lc

Ls+Lc

2.5 GPa1.0 GPa

(1)(2)

(3)

Na2O + K2O

NA

ABSiO2 + Al2O3 + TiO2

CaO + MgO+ FeO*

wt%

CO2-saturated

AB

ENSV cc

dol

ol

1.0 GPa

(1)

(4)

(3)

Na2O + K2O

MgO + FeO*

CaOSiO2 + Al2O3 + TiO2

(b)(a)

NA

6

Liquid immiscibility has become the most favoured model to explain carbonatite genesis (Simonetti & Bell, 1994; MacDonald et al, 1993; Morogan & Lindblom, 1995; Peterson, 1990; Ruberti et al, 2002; Beccaluva et al, 1992 among many others) and has been shown by experimental studies to be a possible process for carbonatite genesis (Lee & Wyllie, 1998a, 1998b, 1997b; Kjarsgaard & Hamilton, 1988; among others). Interestingly, studies of liquid immiscibility in nature have claimed carbonatite unmixing from a variety of silicate rocks (i.e. pyroxenite [Morogan & Lindblom, 1995], peralkaline wollastonite nephelinite [Church & Jones, 1995; Peterson, 1990], melilitite [Stoppa & Cundari, 1995] and nepheline syenite [Beccaluva et al, 1992]). Additionally, carbonatites of widely differing compositions (i.e. Na-rich to Mg-rich) have been suggested to form through unmixing from a carbonated parent. Liquid immiscibility has even been claimed to occur under mantle conditions (e.g. Kogarko et al., 1995; Pyle & Haggerty, 1994 among others). Lee & Wyllie (1996) showed that pure calciocarbonatites (> 80 % CaCO3) cannot form by immiscibility and must be cumulates. Thus most naturally occurring carbonatites plot in the so-called “forbidden volume” in CaO+MgO-FeO* - Na2O+K2O – SiO2+Al2O3+TiO2 –space, and in their current state (chemical composition) cannot represent the carbonatite fraction that was separated from a carbonated silicate parent. Lee & Wyllie (1998a, 1997a) further showed that no proposed paths

of progressive partial melting of carbonated peridotite, or of fractional crystallization of carbonated silicate liquids intersects the miscibility gap under mantle conditions (paths 1-4, Fig. 1a), and concluded that it is unlikely that silicate-carbonate immiscibility occurs during magmatic processes in the mantle. However, Lee & Wyllie (1997a) note that as the MgO-content of evolving nephelinites decrease en route to crustal depths, their compositions are driven close to the silicate volume of the miscibility gap (AB-EN in Fig. 1b) indicating that immiscible carbonate-rich liquids may be generated at crustal pressures. Furthermore, it has been shown that if the parental silicate liquid reaches the miscibility gap during fractionation, more alkali-rich carbonatite liquids can form if the parental melt has a higher Na2O+K2O/CaO-ratio. The experimental results (cf. Lee & Wyllie, 1998a; 1998b; 1997a among others) have also shown that natrocarbonatites cannot be parental to calciocarbonatites as has been proposed (Le Bas, 1989). Finally, experimental studies have also shown that the compositions of dolomite-carbonatites are far removed from the miscibility gap and that such magmas are probably not derived through unmixing, but must represent primary magmas (Lee et al., 2000; Lee & Wyllie, 1998a; 1997a). The case for primary mantle carbonatite melts have mostly been argued by Harmer (1999), Harmer & Gittins (1998; 1997) and Bailey (1993). The idea stems, in part, from the observation that carbonatite sometimes occur without any

Figure 2. Two hypothetical paths for late-stage evolution in ultramafic alkaline complexes. The star shows the position of evolved melt and the arrows show it’s progressive evolution during fractionation. In (a) melilite is stable in the crystallizing sequence and the liquid reaches the miscibility gap. In (b) melilite is not stable and the liquid evolves toward the calcite liquidus field (ne+cpx+phl crystallization). From Veksler et al. (1998a).

Two-liquid field

cpx mel

phl mel ccne

cpx cc

ne

SiO2+Al2O3

TiO2

CaO+MgOFeO

Wt%

Projected from CO2

Na2O+K2Oa)

cpx

phl ccne

cc

ne

SiO2+Al2O3

TiO2

CaO+MgOFeO

Wt%

Projected from CO2

Na2O+K2Ob)

Two-liquid field

cpx+ne+mel

7

association to alkaline silicate rocks. There is experimental evidence that partial melting of mantle peridotites containing CO2 could yield carbonatite liquid with a composition (major elements) corresponding to dolomite carbonatite (Dalton & Presnall, 1995 and 1996). However, experimental studies have also shown that not all carbonatite compositions are likely to represent primary melts, and it has been argued that primary carbonatite magmas from the mantel should be dolomitic (Wyllie & Lee, 1998), Consequently, many proposed primary mantle-derived carbonatite are not actually primary (cf. Lee & Wyllie, 1998a). However, more Ca-rich compositions can be produced if the primary dolomite carbonatite is protected from mantle lherzolite by metasomatic wehrlite. Through the reaction [opx+melt→cpx+ol+CO2 (free fluid)], a dolomitic melt can react progressively and reach calciocarbonatite compositions (Dalton & Wood, 1993: Lee & Wyllie, 2000a). One further limitation to the occurrence of primary mantle-derived carbonatites is that they should occur in isolation, much like kimberlites (Eggler, 1989; Lee & Wyllie, 1997a, 1998a). Petrological evidence for the formation

of carbonatite magma through (extreme) fractionation of a carbonated silicate parent (normally “carbonated nephelinite”) has been argued by some authors (Cooper & Reid, 1998; Andersen, 1988; Veksler et al, 1998a; Church & Jones, 1995). Some of these studies invoke unmixing of a silicate and an (alkali-)-carbonate fraction at an evolved stage following prolonged fractionation of carbonated silicate parent magma (i.e. Andersen, 1988; Church & Jones, 1995), while other evidence point to a straightforward fractionation without late-stage unmixing (i.e. Cooper & Reid, 1998; Veksler et al, 1998a; see Fig. 2). In the study by Veksler et al (1998a), it is apparent that the fractionation of melilite plays a key role in the evolution of the silicate parent magma towards the miscibility gap (cf. Fig. 2) and Cooper & Reid (1998) showed that nepheline sövites from Dicker Willem in Namibia have compositions which allow co-precipitation of silicates and calcite, i.e. they plot at the silicate-carbonate liquidus field boundary (cf. Fig. 3 after Lee & Wyllie, 1998a). The fractional crystallization model appears in the literature to be least favoured, and compared to the liquid immiscibility and

Figure 3. Silicate-carbonatite phase relations at 1.0 GPa showing the three major liquidus volumes for the miscibility gap, silicate liquidus field and carbonate liquidus field, and the liquidus surfaces between them. Contours for 10 wt% MgO+FeO* in-tervals are shown. Nepheline sövites from Dicker Willem (Cooper & Reid, 1998) would plot at the silicate-car-bonate liquidus field boundary indi-cating the possibility of co-precipita-tion of calcite and silicate minerals.

MgO + FeO*

SiO2 + Al2O3 + TiO2

Na2O+K2O

CaO

wt%

Projected from

1.0 GPa

LsLc

CO2

MgO + FeO*

SiO2 + Al2O3 + TiO2

Na2O+K2O

CaO

Na2O+K2O

Projected fromCO2

wt%

Carbonate liquidusSilicate liquidus

MgO + FeO*

SiO2 + Al2O3 + TiO2

CaO

1.0 GPa

8

primary mantle carbonatite models it has not been given much attention.

Each of the above models has advantages and short-comings, which have lead some authors to speculate (and even conclude) that there are “carbonatites and carbonatites”. For example, Harmer (1999) pointed out a number of serious short-comings of the fractional crystallization model, which would appear to make it an unlikely process for carbonatite genesis other than in a few cases. A rather serious short-coming is that for carbonatites to form as residual liquids, Ca (and Mg) would have to behave incompatibly in the crystallizing sequence. This does not seem entirely likely as many of the associated alkaline silicate rocks (i.e. melilitolite, pyroxenite, ijolite) are dominated by (Mg-,) Ca-rich mineral assemblages and these elements thus appear to behave compatibly. However, CO2 is likely to behave incompatibly and may therefore become enriched in residual liquids as shown by the occurrence of interstitial calcite in many of the associated silicate rocks. The advantage of this model compared to both the liquid immiscibility and the primary mantle-derived carbonatite magma hypotheses is that it could be expected to agree with the intrusive sequences observed in carbonatite complexes, e.g. carbonatites are generally intruded after associated silicate rocks. The liquid immiscibility model, fails to delineate

a process by which an immiscibly separated carbonatite magma could be selectively retained at depth, while the silicate magma continues its course upwards through the crust. This is crucial, especially considering that a carbonatite liquid, due to its low viscosity, its lower density relative to the silicate fraction and high volatile content, should be expected to percolate more easily upwards through the mantle and crust (cf. Hunter and McKenzie, 1989; McKenzie, 1985). Similarly, considering the incompatibility of CO2 in the mantle (cf. Olafsson & Eggler, 1983) and presuming that the silicate and carbonatite melts form in the same source area, advocates of the primary mantle carbonatite hypothesis need either to explain why silicate melts would form prior to carbonatite melts in the mantle source, or why carbonatite melts are selectively ponded at some level in the crust, while being ”overtaken” by the silicate melt. In contrast, arguing in favour of primary mantle-derivation of carbonatite, Harmer (1999), suggests that because carbonatite melts are extremely reactive, they are not able to move until channelways lined by unreactive products have been created by preceeding alkaline silicate melts.

Plume-lithosphere interactionSr, Nd and Pb isotopes are extensively used when trying to detect the origin of and relationships between different rock series. Carbonatites have been proven to be of particular use also

Figure 4. Location of EACL (East African Carbonatite Line). Plotted is data from Oldoinyo Lengai (Ne=nephelinites, Ph=phonolites, P.B.=plutonic blocks), Ugandan carbonatites (U.C.; Bell & Belnkinsop, 1987; Nelson et al., 1988), Canary Islands (C.I.; Hoernle & Tilton, 1991) and mafic granulites (M.G.; Cohen et al., 1984). HIMU, DMM and EMI are mantle end-members (Hart, 1988) defining the mixing line. From Bell & Simonetti, 1996.

Figure 5. Location of KCL (Kola Carbonatite Line) after Dun-worth & Bell, 2001. Other data from Mahotkin et al. (2000), Kramm & Kogarko (1994), Zaitsev & Bell (1995) and Beard et al. (1996).

9

when trying to characterize the composition of the upper mantle under continents as carbonatites, due to their high Nd- and Sr-contents effectively buffer against any effects of crustal contamination (Bell & Blenkinsop, 1987). Pb-isotope ratios on the other hand can be used to monitor crustal contamination (because the amount of Pb in mantle derived melts generally is very low: Andersen, 1987) and to characterize the mantle source in cases where crustal contamination is not an issue. One of the most fundamental questions of carbonatite research today is whether the parental melts of carbonatites are derived from the astenosphere or the lithosphere or if they represent mixtures of these components. Many recent studies (Bell & Simonetti, 1996; Bell & Tilton, 2001; Simonetti et al, 1998; Veena at al, 1998 among others) have shown that carbonatites are indeed mantle derived and that a component of astenospheric material is present in many carbonatite and alkaline silicate rocks. Radiogenic isotope data from young carbonatites have shown that they are also very similar to OIB (Ocean Island Basalts), leading to the hypothesis that young carbonatites from East Africa are mixtures between two principal mantle components (Bell and Dawson, 1995), i.e. HIMU and EM1 (HIMU=high 238U/204Pb, EM1=Enriched Mantle 1, high Rb/Sr and low U/Pb and Sm/Nd). These mantle components were originally defined on the basis of Nd, Sr and Pb isotope ratios in OIBs. The young carbonatites of East Africa show such a consistency in isotopic data that the line they represent plotting between the two mantle

end-member has been referred to as the East African Carbonatite Line (EACL, Fig. 4) (Bell & Blenkinsop, 1987). A similar mantle mixing line has also been identified for carbonatites and associated alkaline silicate rocks from the Kola Peninsula (Fig. 5) and is called the Kola Carbonatite Line (KCL; Kramm, 1993; Dunworth & Bell, 2001). However, deviations from this common feature also exist and carbonatites from China for example appear to be derived from a completely different and enriched lithospheric mantle source (Fig. 6; cf. Ying et al, 2004). The associated alkaline silicate rocks in the East African Rift System or elsewhere however, do not show data that are so straightforward and a number of alternative models have been suggested also for the generation of these rocks, including small degrees of partial melting of metasomatized, heterogeneous mantle, crustal contamination and rheomorphism of fenites at crustal levels (Kramm, 1994). Bell & Simonetti (1996) pointed out that of all silicate rocks associated with carbonatites in Africa, phonolites show the greatest scatter in the εNd–εSr diagram. This requires a genesis model involving a third component such as lower crustal granulites, in addition to HIMU and EM1. The preferred model for the generation of the natrocarbonatites and associated silicate rocks of Oldoinyo Lengai is that both the are produced by mixing of two mantle components (HIMU and EM1), with the involvement of a third component to support the generation of the silicate rocks (Bell & Simonetti, 1996). The HIMU component is attributed to a plume or “thermal cell” interacting with the base of the sub-continental lithosphere and causing metasomatism. This provides a mechanism for generating nephelinitic and carbonatitic melts through low-degree partial melting. This type of two-stage model for the generation of alkaline and carbonatitic magmas has also been proposed for the Deccan alkaline province (Simonetti et al, 1998) and for the Sung Valley carbonatite complex, India (Veena et al, 1998). The HIMU components in these two cases are attributed to the Réunion and Kerguelen-Heard plume respectively, and in the case of Sung Valley the model involves mixing with an EM2-type mantle (=Enriched Mantle 2, higher Rb/Sr than EM1 plus high

Figure 6. MORB and OIB field compared to EACL (Bell & Blenkinsop, 1987), Jacupiranga carbonatites (Huang et al., 1995), China carbonatites (Ying et al., 2004) and Italian car-bonatites (Stoppa & Wolley, 1997). From Ying et al. (2004).

10

U/Pb and low Sm/Nd). Isotope data from intrusive alkaline rocks also record complex evolution histories (like the silicate rocks at Oldoinyo Lengai) and these data must therefore be interpreted cautiously. The fact that many silicate rocks of a complex show significantly more scatter with respect to isotope data than the carbonatites also means that establishment of genetic relationships between the two, such as liquid immiscibility, may become ambiguous and should be evaluated carefully.

The Alnö Complex

Specific aims of the projectA detailed petrogenetic study of the Alnö complex on the central Swedish east coast was initiated by docent Viorica Morogan, then at the Department of Geology and Geochemistry at Stockholm University, in 1999. The key questions addressed in, and the specific aims of, this thesis are:

a) petrogenesis of the alkaline silicate rocks at Alnö (Papers I, II and V)

b) petrogenesis of carbonatites at Alnö (Papers III and V)

c) evaluation of element and fluid transport within and around an alkaline-carbonatite intrusion (Paper IV)

d) evaluation of mantle source characteristics and parental magma(s) and numerical modelling of crystal fractionation and crustal contamination in the genesis of the Alnö rocks (Paper V)

Understanding the genesis of carbonatite magmas requires also that we understand the cause for the enormous compositional variations displayed by their associated silicate rocks. Failure to do so, may lead to oversimplifications and misinterpretation of the relationships. Several studies of carbonatite genesis have short-comings, the most serious being overlooking the associated silicate rocks. The carbonatites of a complex are often thoroughly investigated while associated alkaline silicate rocks in some cases are only given brief mention. In order to fully characterize the voluminously most important silicate rocks of the Alnö complex with respect

to mineralogy and mineral chemistry, a detailed petrographic and major- and trace element study of clinopyroxene and Ti-andradite from pyroxenite, ijolite and nepheline syenite was undertaken at the beginning of the project (Paper I). The overall aim of this study was to evaluate the genetic relationships of these rocks by processes such as fractional crystallisation, possible contamination and magma mixing. During this study, important accessory phases were also chemically and petrographically characterized. In doing so, it was recognized that accessory titanite sometimes contained substantial amounts of, from a petrogenetic point of view, important trace element such as Nb, Zr and LREE. This led to a separate, thorough investigation of titanite with the aim of detecting any compositional differences of titanite between different rocks (e.g. pyroxenite, ijolite and nepheline syenite) and to evaluate its petrogenetic significance (Paper II). One of the most fundamental issues for the understanding of carbonatite genesis is first to constrain what constituents observed in a carbonatite have actually crystallized within the carbonatite magma and which are potentially derived from surrounding wall-rock. This has important implications since assimilated phases may substantially alter the chemical composition of carbonatite magma. It also has implications for the interpretation of liquid immiscibility relationships between carbonatites and associated silicate rocks based only on petrography and mineral chemistry (identical liquidus mineralogies in the immiscibly separated liquids, cf. Treiman and Essene, 1985). For this reason, a detailed petrographic and mineral chemistry study was undertaken on three selected carbonatite dykes from Alnö in order to evaluate the mineralogical contributions from surrounding wall-rock and to discuss the implications of this (Paper III). The importance of fluid-rock interactions around alkaline-carbonatite complexes is evidenced in the extensive fenite aureoles surrounding almost all reported occurrences of this association. Such extensive interaction transports elements from the intruding magmas into adjacent wall rock and processes involved have been thoroughly studied in many places (Morogan 1994 and 1989; Morogan & Woolley, 1988; Morogan & Martin, 1985;

11

Kresten, 1988; Kresten & Morogan, 1986; Sindern & Kramm, 2000; Kramm & Sindern, 1998; Kramm et al, 1997; Kramm, 1994; Martin et al, 1978). However, few studies have aimed at quantifying the amount of fluid required to bring about the mineralogical and geochemical changes associated with fenitisation, and fewer studies still have been concerned with the possible transport of elements from surrounding bedrock and into the crystallising magma. This scenario is not only probable but also important from a contamination point of view since elements, which are important tracers in petrogenetic studies, may be mobile under such conditions (cf. Morogan, 1989; Martin et al, 1978; Kramm et al, 1997; Kramm, 1994). Due to this, a study of stable isotopes (C and O) in calcite from carbonatite and in alkali feldspar from fenitized to unfenitized wall rock was undertaken in order to try to quantify the amount of fluid necessary to produce the fenitisation aureole at Alnö (Paper IV). The aim was also to evaluate the amount of element transport, from the source carbonatite and outward, and also from the fenitized wall-rock and inward. All of the above results were later incorpo-rated in the final synthesis of a model for

the generation of the Alnö rocks (Paper V). This model discusses: (1) the genetic relationship between the silicate rocks in more detail and includes a discussion concerning possible contributions from crustal contamination via bulk mixing or fluid-rock interaction; (2) the genetic relationships between the carbonatites and the silicate rocks and possible liquid immiscibility processes in the generation of carbonatite magma on Alnö and (3) mantle sources for the Alnö magmas. Numerical modelling of crystal fractionation of the associated silicate rocks on Alnö was performed.

Geological backgroundThe intrusionsThe Alnö complex consists of five separate small intrusions (Kresten, 1979 and 1990) outlined in Fig. 8. All but one of these are accessible on land, the fifth occurs beneath the Bothnian Bay and is inferred from geophysical data (Kresten, 1976) and boulders found on the small skerries north of the main island. The main intrusion, found on Alnö Island (Fig. 7), is composed of ijolite series rocks (melteigite-ijolite-urtite: modal nepheline increasing in that order) and nepheline syenite, minor pyroxenite and concentric Ca-carbonatite

Figure 7. Geographical overview of the different intrusions of the Alnö Complex (dark grey shaded areas). Location of Alnö Island is shown in the inset. After Kresten (1990).

N

5 km

0o 10o 20o 30o

70o

65o

60o

55oBåräng

Sälskär

Långharsholmen

The main intrusion

Alnö

Fig. 9

Söråker

12

dykes. Later alnöitic, phonolitic and syenitic dykes also occur. Fenitisation is recorded in the surrounding country rock. On the northern shore of Alnö Island and on the islands immediately north of it (Hörningsholmen and Långharsholmen, Fig. 7), the Långharsholmen ring intrusion is composed of pyroxenite, Ca-carbonatite, ijolite, nepheline syenite and pieces of metasomatic wall-rock (i.e. fenite). Morogan & Lindblom (1995) performed a fluid inclusion study on pyroxene, apatite and nepheline from this ring intrusion and suggested the existence of a liquid immiscibility relationship between the pyroxenite and carbonatite in this area. The ijolite and nepheline syenite, found on

Långharsholmen, are probably related to the main intrusion on Alnö Island (Kresten, 1979 and 1990). Centred west of the main intrusion is the smaller Båräng intrusion (Fig. 7) consisting solely of Ca-carbonatite dykes and fenitised wallrock. On the mainland north of Alnö Island, the Söråker intrusion (Fig. 7) consists of subordinate Ca-carbonatite and melilite-bearing rocks (melilitolite). Due to poor exposure, no detailed investigation of the intrusion has been possible. Boulders of carbonatite breccia has been found on the skerries north of the main island, which are believed to have been derived from the bottom of the sea west of the skerries (Kresten, 1979). Geophysical data indicate a

Figure 8. Detailed geology of the main intrusion, and the Båräng and Långharsholmen satellite intrusions of the Alnö complex modified after Kresten (1990). Dashed lines and question marks around nepheline syenite and ijolite bodies in western and southern part of the main intrusion indicate uncertainty regarding the extension of the igneous bodies. The western nepheline syenites may to a large extent be high-grade syenitic fenites. Names of localities used in the text are indicated on the map.

No outcrops

Gulf of Bothnia

0 500 1000m

N

Långharsholmen

Båräng

?

?

?

? ?

??

?

?

??

Smedsgården Ås

Hörningssholm

Näset

Migmatite, unaltered

Fenite, qz-bearing

Fenite, qz-free

Pyroxenite

Ijolite

Nepheline syenite

Carbonatite

Alkaline igneousrocks

Metasomaticrocks

?

?

13

small circular body, which has been interpreted as an explosive carbonatite breccia vent (Kresten, 1979 and 1990). Kresten (1979) also described the existence of a possible carbonatitic centre at Åvike Bay (northeast of map area in Fig. 7). Numerous dykes of carbonatitic and alnöitic character occur around this bay but the actual centre has not been found.

AgeThe Alnö complex was intruded into Precambrian gneissic basement (ortho- and paragneisses and migmatites) of Svecokarelian age. Brueckner & Rex (1980) reported a preferred emplacement age of the complex of 553±6 Ma (Rb-Sr whole rock isochron) while Kresten et al (1977) published K-Ar ages of 546±16 Ma and 552±16 Ma on phlogopites from carbonatite and alnöite respectively. Andersen (1996) interpreted a whole rock Pb-Pb isochron giving an age of 584±13 Ma as the best estimate of the crystallization age of the complex and Walderhaug et al. (2003) reported an age of 590 Ma for the complex.

Alkaline silicate rocks of the complexMelilitolite and other melilite-bearing rocksMelilite-bearing rocks occur both on the mainland north of Alnö Island and within the main complex on the island. On the island, the melilite-bearing rocks occur as dykes and are alnöites and kimberlitic alnöites, which intruded after the main intrusive phase of ijolite, nepheline syenite and carbonatite. The type locality (Näset, Fig. 8) from which alnöite was originally described and given the name melilitebasalt by Törnebohm (1882), shows a black, dense rock with phenocrysts of gold-brown phlogopite and apatite. Kimberlitic varieties of alnöites have been described both from the mainland north of Alnö Island (Kresten, 1990) and from the Hartung area southwest of Hörningsholm (Fig. 8) on the island (Kresten, 1990). These are slightly more primitive than alnöite proper as shown by a high content of olivine and the occurrence of rounded enclaves of mantle-derived rocks composed of olivine, diopside, chromite and phlogopite. Another enclave-rich variety of alnöite is found west of the main complex at Hovid and is either a dyke or a small diatreme. Here, the enclaves are mainly derived from upper and lower crust

although eclogites have also been reported (Kresten, 1990). The position of the alnöititc dykes rocks in the petrogenetic scheme of Alnö is somewhat uncertain but their relevance to our understanding of the nature of the mantle source is unequivocal. At Söråker on the mainland (Fig. 7) a melilite-bearing plutonic rock occurs in a few outcrops and numerous boulders. The rock is extremely coarse-grained and individual melilite crystals occur in sizes of up to 3 cm. This rock is the only known occurrence of melilitolite in the Alnö area and the volume is probably small. However, given its ubiquitous occurrence in other complexes very similar to Alnö (i.e. the Turiy and Kovdor massifs on the Kola Peninsula [Dunworth & Bell, 2001; Verhulst et al, 2000]), and the fact that it is the rock that is most likely to have a composition close to that of the parental magma(s) at Alnö, it is an important piece of the petrogenetic puzzle.

Pyroxenite, ijolite and nepheline syeniteFour series of alkaline silicate rocks and carbonatites have been recognized in the complex on Alnö Island (Kresten, 1990; Morogan, 1988; this study). The following order of intrusion has been established based on field relations: pyroxenite/ijolite → nepheline syenite → carbonatite. The relationship between the pyroxenite and the ijolite (the name is here used collectively for clinopyroxene-nepheline rocks) is somewhat ambiguous although a majority of the pyroxenites appear to be cumulates formed in the same magma that

Figure 9. Field relations of ijolite series rocks from the northern part of the intrusion (Hörningsholm). Light grey areas are urtite and darker grey are ijolite and melteigite.

14

subsequently crystallized ijolite. Layering with gradational relationships between these rocks can be observed at several localities at Alnö, but pyroxenite is also found as fragments included in and sometimes cut by ijolite. All these varieties of pyroxenite are undoubtedly igneous. However, some pyroxenites occur within the fenite aureole and are thought to have formed through metamorphic differentiation during the main fenitisation event (Kresten, 1990). The ijolite series rocks areally dominate the intrusion and modal proportions between clinopyroxene, Ti-andradite and nepheline are highly variable. The rocks of the ijolite series display such variation of textures and structures indicating that they might be derived from several batches of magma (Fig. 9). Melteigites are found in the southern part of the intrusion at Smedsgården and in the northern part at Hörningsholm (Fig. 8). Ijolite is found west of Pottäng, at Smedsgården, at Ås and along the road-cuts east of Hörningsholm towards Stornäset (Fig. 8). Urtite is rare and found mostly as diffuse layers or coarse-grained pockets within ijolite and melteigite. Interstitial K-feldspar occurs sporadically in some ijolites and urtites, which indicates a possible genetic link with nepheline syenites. The nepheline syenites occur as plugs with diameters up to several hundred meters, as well as small bodies, sheets and dykes clearly cross-cutting the ijolite series rocks. Texturally, the nepheline syenites are also variable and three types can be distinguished on the basis of textural and field appearance. These are normal nepheline syenite, pegmatitic nepheline syenite and micro nepheline syenite. Normal nepheline syenite is texturally homogenous and intrudes prior to the micro- and pegmatitic varieties. It is also the most common of the three types; micro nepheline syenite has been observed at three localities and the pegmatitic variety occurs sporadically in the southern part of the main intrusion (Ås and Smedsgården, see Fig. 8). Outside the main intrusion, ijolite series rocks and nepheline syenite have only been observed at Långharsholmen (Fig. 8). Based on field relationships it is thought that the occurrence of ijolite may be related to the main intrusive phase of the main intrusion (Kresten, 1990). Whether the nepheline syenite at Långharsholmen is related to the main intrusive phase at the main intrusion or to the intrusion of

the Långharsholmen rocks is more difficult to establish. There is an extensive occurrence of syenitic fenites on Långharholmen which could render such interpretation ambiguous, as it has been shown that many syenitic rocks in alkaline complexes are in fact rheomorphic fenites (Kramm, 1994; Kramm & Sindern, 1998). This could also be the case for the Långharsholmen nepheline syenite and for some of the areas mapped as nepheline syenite within the main intrusion (Fig. 8 western and southern areas).

CarbonatitesThe carbonatites within the main intrusion on Alnö Island are mainly Ca-carbonatites with only minor Mg-carbonatite. They are intruded as dykes with widths from several tens of meters to only a few centimeters and cross-cut all major igneous rocks of the main intrusion, i.e. pyroxenite, ijolite series rocks and nepheline syenite (Fig. 8). Within the main intrusion the carbonatites predate the alnöites (Kresten, 1990). It has not been possible to establish if there are different generations of Ca-carbonatites within the Alnö complex. Most Ca-carbonatite appears to have been intruded at more or less the same time as no cross-cutting relationships have been observed. However, it has been observed that Mg-rich carbonatites, which only occur as smaller dykes (< 50 cm in width), in general cross-cut the Ca-carbonatite dykes (Kresten, 1990 and 1979). Silicate minerals are ubiquitous in Ca-carbonatite. These are primarily clinopyroxene (diopside → aegirine-augite), phlogopite and titanite, but olivine, alkali feldspars, wollastonite and quartz have also been observed. It should be noted however, that some of the silicate minerals present in the Ca-carbonatites are derived from surrounding wall-rock and thus did not crystallize within the carbonatite magma.

Metasomatic alteration of wall-rockThe fenites surrounding the magmatic rocks of the Alnö complex have been extensively studied by Morogan (1989) and Morogan & Woolley (1988) and are fairly widespread as in most cases of alkaline-carbonatitic magmatism. Originally, all the alkaline silicate rocks at Alnö were considered by Eckermann (1948) to be in situ or rheomorphic fenites, formed by fenitizing fluids emanating from the carbonatite intrusion. The

15

extension of fenitisation at Alnö has since been reduced (Morogan & Woolley, 1988; Kresten, 1979) and many of the rocks previously classified as fenites have been reclassified as truly igneous ijolite and nepheline syenite. The true fenites at Alnö have been divided into low-, medium-, high grade and contact fenites (Morogan & Woolley, 1988) on the basis of mineralogical assemblages and changes in fabric. The nature of fenitisation is in general progressive (Morogan & Woolley, 1988) although this pattern may locally be broken. By using the composition of amphiboles and pyroxenes, Morogan & Woolley (1988) were also able to show that fenites from different parts of the Alnö intrusion equilibrated with compositionally different fluids, i.e. one originating from an ijolite magma and a subsequent fluid emanating from a carbonatitic magma. Neither ijolitic nor carbonatitic fluids are believed to be able to induce nepheline formation in fenite and instead ijolitic fenitisation followed by re-equilibration with carbonatitic fluids is inferred (Morogan & Woolley, 1988). However, this has been contradicted by studies from Iivaara (Kramm, 1994) where it has been shown that ijolite-derived fluids were able to induce nepheline-formation in surrounding bedrock.

Methods and samplingFieldwork (mapping and sampling) has been performed on exposed intrusions, i.e. Långharsholmen, Båräng, Söråker and the main intrusion. From the collected samples about 200 polished thin sections have been prepared and characterized in order to determine crystallization sequences, equilibrium-disequilibrium textures, zoning in minerals which can give important information about possible magma mixing and/or assimilation and also to identify late- and post-magmatic alterations. Samples were then chosen for electron microprobe analyses and whole rock major and trace element analyses as well as for a detailed isotope study. For some purposes, both altered and unaltered samples were chosen to monitor potential influences from later intruding carbonatite dykes on earlier formed alkaline silicate rocks.

Major element analyses of clinopyroxene, Ti-andradite and titanite (Paper I, II and III) were obtained with a Cameca Camebax electron microprobe at the University of Edinburgh,

Scotland and additional electron microprobe work was performed at Uppsala University, Sweden, using a Cameca SX50 WDS. Standards used and operating conditions are described in Papers I, II and III.

Analyses of REE’s in Ti-andradite and clinopyroxene (Paper I) were performed using a Cameca IMS1270 ion microprobe at the Swedish Museum of Natural History. Analytical routines were adapted from those previously described for zircon and garnet by Whitehouse and Platt (2003). Operating conditions are described in Paper I. For REE analyses of clinopyroxene and garnet (Paper I), a reconnaissance laser ablation ICP-MS study was also done. Analyses were performed on a Finnigan MAT Element high resolution sector field ICP-MS and a Finnigan MAT UV laser probe at NGU, Trondheim, Norway. Analytical techniques are described in Paper I.

In order to quantify Fe-valency distribution in titanite, 57Mössbauer spectra were collected at the Swedish Museum of Natural History. Powder X-ray diffractometry using Ni-filtered CuKα-radiation was utilised to obtain cell parameters on titanite. Unpolarized FTIR spectra were collected at room temperature for (010)-sections of titanite single crystals in order to determine the H2O-content. Experimental techniques for all methods are described in Paper II.

δ18O and δ13C of calcite separates were measured with an on-line Kiel device equipped to an IRMS (Finnigan Mat 252) at the Department of Geology and Geochemistry, Stockholm University. δ18O of feldspar separates was measured at the Scottish Universities Environmental Research Centre. Analytical techniques are described in Paper IV.

Major elements were analyzed at ACME Labs, Canada. Sample solutions were analyzed with an ICP emission spectro-graph (Jarrel Ash Atom Comp Model 975) for the determination of major oxides and Cr, Ba, Ni, Sr, Sc, Y, Zr, Co, Cu, Nb, Ta and Zn. Additional trace elements (Ba, Bi, Co, Cs, Ga, Hf, Mo, As, Cd, Pb, Nb, Rb, Sb, Sn, Sr, Ta, Zn, Cu, Ni, Sc, Th, Tl, U, V, W, Y, Zr, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb and Lu) were analyzed with an ICP mass spectrometer (Perkin-Elmer Elan 6000). Loss on ignition (LOI) is determined by igniting a 1 g sample split at 950°C for 90

16

minutes and then measuring the weight loss. The Nd and Sr isotope ratios of whole

rock samples (Paper V) were measured by thermal ionization mass spectrometry on a Finnigan MAT 261 mass spectrometer at the Laboratory for Isotope Geology, Swedish Museum of Natural History, Stockholm. Sample preparation and the constraints for calculating initial isotope ratios are described in Paper V. The additional Sr-Nd-data of T. Andersen were determined in the isotope laboratory of the Mineralogical-Geological Museum, Oslo. Pb isotopes in those samples were analysed at the Department of Earth Sciences, University of Oxford, UK.

ResultsPetrogenesis of the alkaline silicate rocks

(Paper I, II, IV and V)The radiogenic isotope compositions of the silicate rocks from Alnö are shown in Fig. 10. Some samples show evidence of disturbed isotope systematics, in particular with respect to Sr-isotope compositions but otherwise most rocks plot within the depleted quadrant. The samples which have a more radiogenic isotope composition with respect to Sr are collected from the outer margins of the main intrusion. Similar trends of radiogenic Sr-isotope composition, with retention of juvenile Nd, have been reported for silicate rocks from Grönnedal-Ika, Igaliko and Qassiarsuk on Greenland (Fig. 11), from Turiy on the Kola Peninsula (Dunworth & Bell, 2001) and from Italy (Stoppa & Woolley, 1997). Bulk mixing calculations between a parental magma composition (=melilitolite from Söråker) and local crust, predict between 20-30% assimilation in order to explain the more radiogenic Sr-isotope signatures (cf. Paper V). This is generally considered unlikely in terms of AFC(LI) processes (cf. Dunworth & Bell, 2001; Ray, 1998; Ray et al, 2000), and it is suggested that element exchange may have occurred as a result of fenitisation and hydrothermal leaching of the surrounding bedrock at late- to post-magmatic stages of crystallization. A similar model has also been suggested to explain the most shifted Sr-isotope ratios observed at Turiy (Dunworth & Bell, 2001). Results, presented in Paper IV, show that fluid coupled with chemical transport during fenitisation is not only directed outward from the intruding magma body, but

also inwards. Mobility of certain trace elements has been shown to occur during fenitisation (Morogan, 1989; Kramm & Sindern, 1998; Kramm et al, 1997; Kramm, 1994; Martin et al, 1978) and it is likely that some transport of trace elements into the intruding magma should also occur. Thus, this could represent a viable mechanism to selectively alter certain isotope signatures, but would require that Sr (and Rb) is more mobile than Nd (and Sm) during hydrothermal activity. This is probable since titanite, which has been observed to form during fenitisation (Morogan & Woolley, 1988), could easily accommodate Nd (and Sm) and thus selectively retain Nd the within the metasomatized wall-rock. Studies of alkaline complexes have shown that nephelinitic, basanitic and alkali basaltic liquids can crystallize either under closed system conditions (no exchange with surrounding bedrock) or as open systems (allowing exchange to occur) to yield a variety of rock types from gabbros, through diorites, pyroxenites and ijolite series rocks to nepheline syenites and syenites (Pati et al., 2000; Beccaluva et al, 1992). Some complexes even show the presence of ultramafic cumulates such as peridotites and dunite (Verhulst et al, 2000; Dunworth & Bell, 2001). The Alnö complex is more restricted in the compositional range of exposed rock types, albeit still diverse. Any proposed petrogenetic model for the formation of the exposed rocks should take into account all different types of silicate rocks in order to fully constrain and characterize the magmatic evolution of the complex. Co-genesis of some rocks (ijolite series and clinopyroxenite) is indicated by field relations, i.e. gradational contacts between the different members of the ijolite series rocks and clinopyroxenite are common, whereas other relations are more obscure (i.e. ijolite rocks in relation to nepheline syenite, melilitolite and alnöite). In volume, the ijolite rocks (including clinopyroxenite) and nepheline syenite are most important within the main intrusion. Major element composition of clinopyroxene and Ti-andradite from these rocks (Paper I) show fractionation trends consistent with clinopyroxenite, ijolite and nepheline syenite being derived through differentiation of a common parental magma (Fig. 12). However, closed conditions do not

17

seem to have prevailed and a straightforward closed-system fractional crystallization model is contradicted by clinopyroxene major element data which show the existence of two separate fractionation trends. Oscillatory zoning of clinopyroxenes in ijolite and nepheline syenite, and the occurrence of resorbed clinopyroxene (diopside) cores in nepheline syenite, indicate that different magma batches may have been introduced into the magma chamber during the course of crystallization (Paper I). It is possible that mixing of earlier residuals (from trend [1]; c.f. Paper I) and a new batch of more primitive magma, lead to slightly different crystallization conditions (as evidenced by the lack of Ti-

andradite in the rocks with clinopyroxenes plotting along trend [2]; c.f. Paper I) and the formation of the more Na-rich melteigites and nepheline syenites of trend [2]. None of the exposed silicate rocks on Alnö Island are likely to represent parental magma composition. However, based on initial isotope ratios (Sr and Nd) the melilitolite, which crops out at Söråker, can be considered the closest available approximation. Numerical modeling (AFC) assuming the parental magma composition to be olivine-melilititic, indicates that fractionation of olivine, melilite and diopsidic clinopyroxene with minor apatite, magnetite and perovskite could generated ijolite-like residual liquids (cf. Paper V). The degree of fractionation required is 35 % (19 vol% olivine and 16 vol% melilite + diopside + apatite + magnetite + perovskite). However, the model only gives a significant fit if 2.4 % carbonatite is added to the parental magma. Numerical modeling also indicates that the same parental magma could produce phonolitic residuals after prolonged fractionation of the same cumulus phases, nepheline and Ti-andradite although at larger degree of fractionation (86 vol%) and if 1.5 % carbonatite is extracted (via immiscibility?) during the course of crystallisation. The model also suggest that production of ijolite+nepheline syenite residual magma require fractionation of 17 vol% pyroxenite + melilitolite cumulates and 22 vol% olivine, without addition or removal of a carbonatite fraction. Ultramafic cumulates such as peridotite and olivine melteigites have not been found at Alnö, but are frequently reported from similar complexes elsewhere (Dunworth & Bell, 2001; Verhulst et al, 2000; Beccaluva et al, 1992). This indicates that olivine is an important fractionating phase in alkaline complexes. The validity of the major element model is supported by trace element models, with reasonably good results (Fig. 13), considering uncertainties regarding the partition coefficients used (which are poorly constrained for phases such as melilite and Ti-andradite). Furthermore, major and trace element plots of silicate rocks from Alnö Island (Fig. 14 and 15) show trends, albeit scattered, which are consistent with fractionation of the phases used in the model calculations. The apparent differences between pyroxenite, ijolite and melilitolite in the trace element plots with respect to elements such as Nb, La, Ce, Zr

�Nd

(590

Ma)

�Sr (590 Ma)-15 -5 5 15 20

CHUR

B.E

.

0-10 10

4

2

0

5

3

1

-4

-2

-5

-3

-1 melilitolitepyroxeniteijolitenepheline syeniteapatite-rich rockcarbonatite

gr I

gr II

Figure 10. εSr-εNd (at 590 Ma) for Alnö rocks. Symbols are given in the figure. CHUR (CHondritic Uniform Reservoir) and B.E. (Bulk silicate Earth) calculated at 590 Ma. Coloured fields relate to rock groups. Carbonatites have been divided into group I and II. Grey square is data from the Fen alkaline complex in Norway (Andersen, 1987). See text for explanation.

Figure 11. Isotope data for carbonatites and associated syen-ites from Grönnedal_Ika on Greenland (Halama et al., 2003). Included is also data from the Igaliko dyke swarm (Pearce & Leng, 1996) and the Qassiarsuk carbonatite (Andersen, 1997).

18

Na

(a.p

.f.u.

)

-0.800 -0.600 -0.400 -0.200 0.000 0.200 0.400 0.600Na-Mg (a.p.f.u.)

Ca

(a.p

.f.u.

)

1.100

0.900

0.800

0.700

0.600

0.500

0.400

1.000b)

pyroxenitecpx field [1]cpx field [2]

-0.800 -0.600 -0.400 -0.200 0.000 0.200 0.400 0.600Na-Mg (a.p.f.u.)

pyroxenitecpx field [1]cpx field [2]

0.600

0.800

0.400

0.200

1.000

0.000

FeTO

T (a

.p.f.

u.)

c)

pyroxenitecpx field [1]cpx field [2]

0.600

0.500

0.400

0.300

0.200

0.100

0.000-0.800 -0.600 -0.400 -0.200 0.000 0.200 0.400 0.600

Na-Mg (a.p.f.u.)

a)

and Hf, particularly highlights the importance of titanite (Paper II), apatite and perovskite in controlling trace element distribution. One interesting feature that has emerged from the geochemical study of the Alnö complex is that many rocks previously mapped as nepheline syenites may in fact be (rheomorphic) fenites. These rocks display magmatic textures but have geochemical signatures intermediate between that of magmatic nepheline syenite and fenite. Petrogenesis of the carbonatites (Papers III, IV and V)The petrogenesis of carbonatite magma is, as was described previously, complex and may involve several processes. One complication in the interpretation of carbonatite petrogenesis, by using whole rock major and trace element data, arises from the fact that carbonatite magma composition may be extensively modified through interaction with wall rock on the way to the surface. Many carbonatites exposed at crustal levels, in particular Ca-carbonatites, contain a variety of silicate minerals, some of which are obviously not crystallized within the carbonatite magma (e.g. quartz and alkali feldspar) and some of which the origin is more difficult to constrain (i.e. clinopyroxene and phlogopite). When studying the Alnö

carbonatites, it was noticed that in order to more completely comprehend the genesis of the carbonatites and their relation to the associated silicate rocks, distinction between primary silicates and assimilated/secondary phases was necessary (Paper III). This issue has been rather poorly examined in carbonatite literature and few investigations have been aimed at trying to constrain which silicates can actually crystallize within carbonatite magmas (Barker, 2001; Chakmouradian & Mitchell, 2003). Essentially, four possible origins of silicate minerals in carbonatites on Alnö Island can be discussed: (1) magmatic phases crystallized within the carbonatite magma, (2) assimilated phases from wall-rocks (this includes quartz and alkali feldspar), (3) metasomatic phases formed through interaction of assimilated phases with carbonatite liquid at late- to post-magmatic stages, and (4) pseudo-primary phases (Paper III). Experimental studies have shown that carbonatite liquid is capable of percolation through wall-rock by coupled corrosion and low dihedral angle flow (cf. Hunter and McKenzie, 1989; Hammouda and Laporte, 2000). Interestingly, Hammouda & Laporte (2000) also noted that corrosion and dissolution of mantle dunite resulted in precipitation of pseudo-primary olivine in the bulk carbonatite liquid.

Figure 12. Compositional evolution of Alnö clinopyroxenes in terms of differentiation index (Na-Mg a.p.f.u.) against Na, Ca and FeTOT (a.p.f.u.). Clinopyroxenes from the southernmost part of the main intrusion (Smedsgården, field [2]) consist-ently plot at higher Na and lower Ca and FeTOT compared to clinopyroxenes of field [1]. See text for discussion.

19

60.0

0.0

10.0

20.0

30.0

40.0

50.0

60.050.040.030.020.010.00.0

CaO

(wt%

)

SiO2 (wt%)60.050.040.030.020.010.00.0

SiO2 (wt%)

10.0

0.0

2.0

4.0

6.0

8.0

MgO

(wt%

)

60.050.040.030.020.010.00.0SiO2 (wt%)

25.0

0.0

5.0

10.0

15.0

20.0

Al 2O

3 (w

t%)

60.050.040.030.020.010.00.0SiO2 (wt%)

25.0

0.0

5.0

10.0

15.0

20.0

Fe2O

3 (w

t%)

60.050.040.030.020.010.00.0SiO2 (wt%)

Na 2O

(wt%

)

10.0

0.0

2.0

4.0

6.0

8.0

12.0

60.050.040.030.020.010.00.0SiO2 (wt%)

K2O

(wt%

)

10.0

0.0

2.0

4.0

6.0

8.0

60.050.040.030.020.010.00.0SiO2 (wt%)

TiO

2 (w

t%)

2.5

0.0

0.5

1.0

1.5

2.0

3.0

3.5

60.050.040.030.020.010.00.0SiO2 (wt%)

0.0

2.0

4.0

6.0

P 2O5 (

wt%

)

a) b)

c) d)

e) f)

g) h)

Figure 14. Major el-ement variations in Alnö rocks. Symbols: open circle = melili-tolite, filled cicle = al-nöite, black diamond = pyroxenite, blue diamond = melteigite, pink square = melt-eigite-ijolite, white square = ijolite, green triangles = transition-al ijolite-nepheline syenite, black trian-gle = nepheline syen-ite, plus = carbonatite and red diamond with cross = fenite.

20

Figure 15. Incompatible trace element distribution patterns for a) melilitolite, alnöite and pyroxenite-melteigite, b) ijolite, c) nepheline syenite and d) carbonatite. Colours are: a) grey=pyroxenite, green=melilitolite, purple=melteigite, red=jacupirangite, black lines=alnöites and dashed lines=Turiy melilitolite. b) grey field=ijolite, green=melteigite-ijolite, blue=garnet-poor ijo-lite and red=transitional ijolite-nepheline syenite. c) red=transitional ijolite-nepheline syenite, blue=micro nepheline syenite, green=pegmatitic nepheline syenite, black lines=normal nepheline syenite and dashed line=fenite. d) grey field=carbonatite, black=dolomite-bearing carbonatite, red=cpx-carbonatite with zirconolite, dark blue=calcite carbonatite, light blue=pyrochlore-bearing carbonatite, green=magnetite-rich carbonatite and dashed line=olivine-bearing apatite-rich carbonatite.

This indicates that a carbonatite may be capable of extensively modifying its silica-content to such a degree as to allow for crystallization of seemingly primary phases, a feature also noted by Chakmouradian & Mitchell (2003) in the Prairie Lake carbonatite in Canada. Studies of fluid inclusions in apatite and clinopyroxene in carbonatites from Alnö Island (Morogan & Lindblom, 1995) have shown high contents of salt (up to 65 %). Many other studies of carbonatite fluid inclusions have shown them to coexist with alkali rich H2O-CO2 fluids which contain appreciable amount of SO2 and Cl (Rankin, 1977; Rankin & LeBas, 1974; Samson et al, 1995). Experimental studies on carbonatites have also shown that the solubility of H2O in carbonatite magma is extremely high when compared to most silicate melts (Keppler, 2003), and Nielsen & Veksler (2002) have argued that the natrocarbonatites of Oldoinyo Lengai are not lavas but expelled cognate, alkaline and CO2-

rich fluid condensates. The water/carbonatite ratio in turn has been shown to greatly affect DNa

(fluid/melt) in carbonatite systems (Veksler & Keppler, 2000). Deep crustal fluids also contain large amounts of water (cf. Schmulovich et al. 2001) in addition to carbon dioxide, methane and chloride salts which have implication for element distribution, dissolution, transport and precipitation of silicate material during crustal processes. Consequently, understanding the effects of salt-rich solutions on mineral stabilities under lower and upper crustal conditions in particular may be important in the study of carbonatite petrogenesis. Schmulovich et al. (2001) showed that the solubilities of quartz and albite decreased with increasing salt concentration in the fluid (salt-out effect) whereas the opposite was observed for diopside (salt-in effect). Studies of melt inclusions in carbonatites from the Gardiner and Kovdor complexes (Veksler et al. 1998; Nielsen et al.

10000

1000

100

10

1

0.1

0.01

sam

ple/

prim

itive

man

tle

Rb Th Nb K Ce Pr P Zr Sm Ti Y LuCs Ba U Ta La Pb Sr Hf Eu Dy YbNd

transitional (37AA, 2AB)micro nepheline syenite (06AG, 06AF)pegmatitic nepheline syenite (17SGA)normal nepheline syenitefenite

Rb Th Nb K Ce Pr P Zr Sm Ti Y LuCs Ba U Ta La Pb Sr Hf Eu Dy YbNd

Rb Th Nb K Ce Pr P Zr Sm Ti Y LuCs Ba U Ta La Pb Sr Hf Eu Dy YbNd

Rb Th Nb K Ce Pr P Zr Sm Ti Y LuCs Ba U Ta La Pb Sr Hf Eu Dy YbNd

10000

1000

100

10

1

0.1

0.01

sam

ple/

prim

itive

man

tle

jacupirangite (27AA, 27AB)melilitolite (SÖR05B)melteigite (79MSR, 03SG, 18AC)alnöite (05H, 15H)pyroxenite

transitional ijolite-nepheline syenite (37AA, 07HB)grn-poor ijolite (17SGB)melteigite-ijolite (36AA, 34AA, 03AC)ijolite

cpx-carbonatite (with zirconolite, 20SGD)cc-carbonatite (20SGC)mt-carbonatite (24SG)carbonatite

Mg-carbonatite (L2a)

BA0101SG

Turiy melilitolite

a) b)

d)c)

21

1997) suggest that many plutonic carbonatites that now have a calcium-dominated composition may have contained substantial amounts of alkalies. All these studies imply that carbonatite magma/fluid have the chemical criteria which may allow it to act as a solvent on assimilated wallrock. This concurs with the observations that diopside (and aegirine-augite) in Alnö carbonatites are partially or wholly replaced by calcite (i.e. “dissolved”). Interestingly, Schmulovich et al (2001) observed that olivine, talc, serpentine and wollastonite precipitate on diopside surfaces after incongruent dissolution. All these phases have been observed and are relatively common in plutonic carbonatites. The extent of the dissolution effect in carbonatite is unfortunately poorly known and should be investigated further. The main concern, however, that arises from these results is the validity of some of the criteria (i.e. identical liquidus mineralogies and evolution of the liquidus phases in carbonatite and conjugate magma, cf. Treiman & Essene, 1989) that previously have been used to establish a liquid immiscibility relationship between carbonatite and associated silicate rocks (cf. Andersen, 1996; Cooper & Reid, 1998). This study particularly highlights the difficulty in using such criteria when looking at plutonic carbonatites as these may have had plenty of time to interact with surrounding wall rock and thus may have been extensively modified relative to their original chemical composition. Considering the Sr- and Nd-isotope composition of carbonatites, the assimilation of substantial amounts of wallrock may not be a significant problem due to the high buffering capacity of most carbonatites, but it is definitely a considerable problem in whole rock major and trace element analyses of carbonatites as the buffering capacity with respect to elements such as Si, Al, K, Na and Pb can be expected to be low or non-existent. This also presents opportunities in the study of carbonatites. Pb-isotopes are for instance a useful tool in the study of contamination processes in carbonatites (due to very low concentrations in carbonatite magmas), contrary to Sr and Nd which are not. Major element data show elevated Al, Si, K and Na in carbonatites with high amount of silicate material, which based on results presented in Paper III, can be attributed to crustal contamination. Depending

on the trace element composition of assimilated phases, some effects on other trace elements, besides Pb, can also be expected, i.e. Zr and Ti may be affected by assimilated titanite and Ti-andradite and K may be affected by assimilated K-feldspar and/or biotite/phlogopite. For the Alnö carbonatites, the effect of assimilated phases on carbonatite major and trace element composition needs to be further and more systematically explored. Carbonatite formation by liquid immiscibility from a carbonated silicate parent magma has been discussed in the literature [Bell, 1989; Bell & Keller, 1995; Simonetti & Bell, 1994; MacDonald et al, 1993 among others]. The composition of the inferred conjugate silicate liquid is most often nephelinitic/ijolitic or phonolitic (Church & Jones, 1995), but syenitic and pyroxenitic (cf. Morogan & Lindblom, 1995) compositions have also been invoked. Given the results from the modeling and the similarities in radiogenic isotope signatures between the carbonatites and silicate rocks on Alnö Island, it could be argued that the carbonatites of the main intrusion were derived from associated silicate rocks by liquid immiscibility. Recent experimental studies (Veksler et al, 1998; Jones et al, 1995) on the partitioning of trace elements between conjugate silicate and carbonatite liquids have shown that most Nb, Ta, Zr, Hf and REE (except La and Ce) preferentially partition into the silicate liquid whereas elements such as Ba, Sr, Pb, Th and U should all prefer the carbonatite fraction. However, if the carbonatites (and silicate rocks) are affected by fractionation processes (c.f. Cooper & Reid, 1998) the trace element ratios will not reflect the initial magma concentrations in the two conjugate liquids. The carbonatites on Alnö Island are affected by fractionation processes and have variable modal amounts of primarily apatite and pyrochlore as seen in the widely scattered behaviour of P, LREE and also U and Th. Due to the highly fractionated trace element patterns, with respect to Th and U, in the Alnö carbonatites (cf. Fig. 15d), these element cannot be expected to reflect initial concentrations at the time of unmixing. It is apparent that the trace element data for Alnö rocks at this stage are inconclusive regarding a possible liquid immiscibility relationship between carbonatites and associated silicate

22

rocks and that such a model can neither be confirmed nor discarded.

Mantle source(s)Trace element data together with radiogenic isotope ratios for the Alnö silicate rocks and carbonatites can be used to constrain the composition of the mantle source. The most depleted samples are represented by the melilitolites and the most depleted carbonatites (group I, cf. Fig. 10), which have εSr ≤ -9 and εNd ≥ 2.6. All but three samples from Alnö Island plot within the depleted quadrant showing that all rock were derived from (a) mantle source(s) that had experienced a time-integrated depletion in LIL elements (i.e. Rb and LREE), much like the case for many other alkaline silicate rocks and carbonatites (Verhulst et al, 2000; Dunworth & Bell, 2001; Simonetti et al, 1998; Riley et al, 1999; Harmer et al, 1999 among many others). As in the case for the Alnö samples, isotope data for carbonatite complexes tend to sometimes show slight variations in initial ratios but still plot within the depleted mantle quadrant, which in some case is argued to reflect mantle heterogeneity. For the Alnö samples, this aspect needs to be considered as the two carbonatite groups in Fig. 10 may reflect derivation of magmas from a heterogenous (or different?) mantle source. If this is true, then the more radiogenic carbonatites appear to have been generated within the same or similar source as the bulk of the silicate rocks (i.e. ijolite, pyroxenite, apatite-rich rocks and nepheline syenite of the main intrusion), whereas the carbonatites of group I appear to be more closely related to the ultramafic rocks and possible also some ijolites. Generation from a different or heterogenous source could be likely as the most depleted carbonatite samples, the ultramafic rocks and one of the depleted ijolite samples are collected from the Söråker and Långharsholmen ring intrusions which pre-date the main intrusive phase of the Alnö complex.

Another important feature for the most “primitive” rocks from Alnö (as for many other alkaline-carbonatitic complexes) is that although having depleted signatures with respect to isotope compositions, they are enriched with respect to trace elements and as such, similar to OIB (Ocean Island Basalts). A comparison between some of the mafic cumulates from

Alnö Island and lavas derived from well-known modern mantle reservoirs (Honolulu, Hawaii; Yang et al., 2003), suggests that the trace element composition of the isotopically most depleted melilitolite and pyroxenites are closely similar to modern nephelinites and basanites whose petrogenesis arguably involve a plume component.

The decoupling of the (often) depleted Sr- and Nd-isotope signatures and highly enriched trace element character of alkaline silicate rocks and carbonatites is a feature which is often attributed to mantle metasomatism at a stage when the mantle had already experienced earlier melt extraction (Andersen, 1987). The need for a metasomatising event preceding the partial melting to form carbonated alkaline silicate melt is supported by experimental studies (e.g. Olafson & Eggler, 1983), which have shown that carbonates are low-melting phases in mantle lherzolite. Thus, even if carbonates were present in the mantle prior to the earlier depletion events, CO2 would need to be introduced to the mantle source again before melting can occur to produce the carbonatite and alkaline silicate melts considered parental to most melilitolite-ijolite-carbonatite complexes. The agent inducing metasomatism in earlier depleted mantle is often considered to be fluids expelled from a rising astenospheric plume (Veena et al, 1998; Dunworth & Bell, 2001; Verhulst et al, 2000). Involvement of a plume component in the generation of the carbonatites of the Kola Peninsula is supported by the He-isotope signature of these rocks (Marty et al, 1998). However, as shown for many of the complexes for which a plume component has been invoked, this would lead to mixing between two components of differing isotope composition (a plume component and an enriched sub-continental lithospheric mantle) so that the magmas generated from such sources would fall along a mixing line in the εSr-εNd diagram (Bell & Tilton, 2002 and 2001; Simonetti & Bell, 1995; Dunworth & Bell, 2001; Kramm, 1993 among many others). Such a trend is not observed in the current set Alnö samples and metasomatism induced by a plume component is thus not considered likely in the Alnö case. The same was observed for the Fen complex by Andersen (1987) who suggested that percolation of a CO2-H2O phase through

23

the upper mantle may selectively remobilize the LREE on a local scale and redeposit them in metasomatic amphibole-phlogopite-carbonate parageneses, which subsequently would melt and form an initial carbonated nephelinitic melt. Further LREE enrichment was suggested to occur through volatile transfer into the magma after its formation (Andersen, 1987). Such a model is considered to be likely also for the Alnö complex.

ConclusionsThe melilitolite and ijolite series rocks exposed within the main intrusion, the Långharsholmen ring intrusion and at Söråker are co-genetic and related to the same parental magma(s) through fractional crystallization of olivine, clinopyroxene and melilite. The formation of the ijolite series rocks occur under essentially closed system and crustal contamination of the magmas probably did not occur until late- to post-magmatic stages. The nepheline syenites within the main intrusion can also be related to the ijolite series rocks through fractionation of additional clinopyroxene, nepheline and Ti-andradite. The formation of the nepheline syenites however cannot have occurred at completely closed conditions as shown by clinopyroxene data and petrography. Feldspar-bearing ijolites and urtites represent transitions from ijolite series to nepheline syenite. Some nepheline syenites/syenites exposed at the western and southern margins of the main intrusion and the nepheline syenites/syenites at Långharsholmen have magmatic textures but display major and trace element chemistries that are transitional between magmatic nepheline syenite and fenite. Thus, these rocks are suggested to be rheomorphic fenite. The extension and volume of these rheomorphic fenites is unknown and should be investigated further. Accessory phases like titanite and apatite have been shown to be important fractionating phases in the sequence of ijolite series rocks, due to their ability to incorporate substantial amounts of trace elements such as Nb, Zr and LREE. The importance of them and other accessories should be evaluated further. Carbonatites on Alnö are mainly Ca-carbonatites which have been variably contaminated by wall-rock material. Consequently, most silicates (either dispersed

grains or wall-rock fragments) found in the carbonatites on Alnö Island are either (1) assimilated from wall-rock, (2) produced by reaction between assimilated material and carbonatite liquid, (3) pseudo-primary as a result of increased aSiO2 (and aAl2O3, aK2O, aNa2O) due to dissolution of assimilated silicate phases or (4) magmatic. This shows that contamination of carbonatite magma is real and that significant modification of the major and trace element composition of the carbonatite magma may occur as a result of it, which will complicate interpretations of genetic relationships with associated silicate rocks. However, support for a liquid immiscibility relationship may be argued from the similarities in initial Nd- and Sr-isotope rations between some carbonatites and ijolites and also from the major and trace element modeling results. The parental magma (i.e. olivine-melilitite) at Alnö was derived from a LILE-depleted mantle source which hade experienced subsequent enrichment in trace elements, probably through small-scale percolation of CO2-H2O fluids in the mantle. Further trace element enrichment of the magma(s) may have occurred through volatile transfer during partial melting and ascent. No evidence for a plume component, which is often inferred from isotope data ok alkaline rocks and carbonatites, is seen in the isotope data from Alnö.

Suggestions for future workAlthough a lot of work to date has been performed at Alnö and a large set of data from various rock types and minerals have been compiled, there is still a number of issues regarding the petrogenesis that need attention. The first and most important objectives should be a more detailed isotopic study of the nepheline syenites since some of the rocks previously mapped as nepheline syenites have been shown in this study to have geochemical signatures intermediate between magmatic nepheline syenite and fenite (particularly in the western and southern part of the complex). The possibility that some nepheline syenites/syenites may be rheomorphic fenites (cf. Kramm, 1994; Kramm et al, 1997) is thus indicated and needs further attention. The results presented in Paper II show that the accessory phases in alkaline silicate rocks and carbonatites have important

24

petrogenetic implications. It is obvious that detailed knowledge of both major and trace element chemistry of these phases is crucial to the understanding of genetic relationships between the associated rocks types. Of the most important phases it is suggested that apatite, perovskite and pyrochlore should be given special attention in such a study, but rarer phases such as zirconolite, wöhlerite, fersmanite and reported baddeleyite should also be considered.

Stable isotopes were not systematically analyzed in carbonatites or silicate rocks. The only investigated carbonatite dyke (at Smedsgården, cf. Paper IV) has a mantle signature (δ18O of about 8 ‰ and δ13C about -4 ‰). It would be of interest to further investigate different dykes at Alnö and also to analyze interstitial calcite and alkali feldspar in the silicate rocks in order to monitor fluid-rock interaction (late- to post-hydrothermal alteration by meteoric water) and possible crustal influence in the genesis of some of the silicate rocks. This would help to further refine the petrogenetic model of the Alnö complex.

Since carbonatite petrogenesis still is heavily debated in the scientific community any new information regarding the character and nature of carbonatite magmas is essential. Experimental studies of liquid immiscibility and generation of primary carbonatite magma from the mantle are numerous and have provided a rigid framework for interpretation of carbonatites and associated silicate rocks in general. Despite this, further experimental work is needed on carbonatites, in particular studies dealing with crystallization of silicate minerals in carbonatites as this has important implications on the composition of crustal carbonatites. Ideally, such experiments would be carried out using a variety of starting compositions for both carbonatite and wall rock at a set of different physical conditions. A very important aspect of these experimental studies should also be to quantify the chemical modification that will occur in a carbonatite magma during its ascent through and interaction with the crust, since interaction with siliceous and aluminous crustal material will undoubtedly change its chemical composition significantly (cf. Paper III; Chakmouradian & Mitchell, 2003; Hammouda & Laporte, 2000; Schmulvich et al., 2001).

AcknowledgementsFirst of all I would like to express my heartfelt thank you to my three wonderful supervisors: Professors Alasdair Skelton (Department of Geology and Geochemistry, SU), Ulf Hålenius (Dpartment of Mineralogy, SMNH), and Stefan Claesson (SMNH). You have all helped me more than you know, not only scientifically, and I thank you for your patience. I especially would like to extend my heart-felt and sincere thank you to Alasdair for taking me on as a student at a time when everything was chaos and for making such a good job of it. I would also like to thank Martin Whitehouse, Marina Fischerström, Ann-Marie Kähr, Paula Allardt and Hans Schöberg (all at the Swedish Museum of Natural History) for technical support and laboratory assistance, during my work at the Swedish Museum of Natural History. I would like to take this opportunity to especially acknowledge Joakim Mansfeld and Johan Söderhielm for good company in the field (and elsewhere), nice photographs and for lots of input and interest in this project. Johan made some really nice polished thin sections for which I am grateful. Thanks also to the nice people in Fikahörnan plan 4, Curt and Dick, you have all been good company over the years and well…what can I say – you have a way to bring out the best in me when it comes to feministic ideas! Thank you to Hans Vaerneus (Stockholm University) for good cooperation, nice field trips and support and encouragement when I needed it most. Arne Lif (SU) – I thank you for keeping computer problems to a minimum although I still think you should have given me more stuff to play with! My very best friends – Lillemor, Malin and Sussie – may we share many more happy moments, laughs and adventures together! I will miss you. My other girls – Ilka and Gun – we have shared much, good and bad but the happy moments are the ones to remember. I’m so lucky to have you all in my life. Emma Mirkovic, you were an A-student during the week you helped me in the lab – I thank you for your help. My family (Vuorinen and Berggren) and my extended family (Hode and Mirkovic) – for always being interested. Most of all I would like to thank Tomas. Thank you for wanting to share your life with me (and the cats!)…and for letting me spend

25

endless hours playing FPS’s and MMORPG’s! Cosmos and Balrog were, as always, indifferent to the reults of scientific studies...

Financial support for this study was provided by Stockholm University and the Swedish Mueum of Natural History, with additional support from Arne S. Lundbergs Fond, Wallenbergstiftelsens jubileums-fond, Stiftelsen Lars Hiertas Minne and Stiftelsen Hierta- Retzius Fond.

References Andersen, T., 1997. Age and petrogenesis of the Qassiarsuk carbonatite-alkaline silicate extrusive complex, Gardar Rift, South Greenland. Mineralogical Magazine 61, 499-513. Andersen T., 1996. Sr, Nd and Pb isotopic data of the Alnö carbonatite complex. Abstract volume, 22nd Nordic Geological Winter Meeting. p 11. Andersen T., 1988. Evolution of peralkaline calcite carbonatite magma in the Fen complex, southeast Norway. Lithos 22, 99-112. Andersen T., 1987. Mantle and crustal components in a carbonatite complex, and the evolution of carbonatite magma: REE and isotopic evidence from the Fen complex, Southeast Norway. Chemical Geology 65, 147-166. Bailey D.K., 1993. Carbonate magmas. Journal of the Geological Society, London 150, 637- 651. Barker D.S., 2001. Calculated silica activities in carbonatite liquids. Contributions to Mineralogy and Petrology 141, 704-709.

Beard A.D., Downes H., Vetrin V., Kempton P.D. and Maluski H., 1996. Petrogenesis of Devonian lamprophyre and carbonatite minor intrusions, Kandalaksha Gulf (Kola Peninsula, Russia). Lithos 39, 93-119. Beccaluva L., Barbieri M., Born H., Brotzu P., Coltorti M., Conte A., Garbarino C., Gomes C.B., Macciotta G., Morbidelli L., Ruberti E., Siena F. and Traversa G., 1992. Fractional Crystallization and Liquid Immiscibility Processes in the Alkaline Carbonatite Complex of Juquia (Sao Paulo, Brazil). Journal of Petrology 33, 1371-1404. Bell, K., 1989. Carbonatites: genesis and evolution. Unwin Hyman, London. Bell K. and Tilton G. R., 2002. Probing the mantle: the story from carbonatites. EOS 83, 273- 277. Bell K. and Tilton G. R., 2001. Nd, Pb and Sr isotopic compositions of East African carbonatites: evidence for mantle mixing and plume inhomogeneity. Journal of Petrology 42, 1927-1945. Bell, K., Kjarsgaard, B. A. and Simonetti, A., 1998. Carbonatites – into the twenty-first century. Journal of Petrology 39, 1839-1845. Bell K. and Simonetti A. 1996. Carbonatite magmatism and plume activity: implications from the Nd, Pb and Sr isotope systematics of Oldoinyo Lengai. Journal of Petrol 37, 1321-1339. Bell K. and Dawson J.B., 1995. Nd and Sr isotope

systematics of the active carbonatite volcano, Oldoinyo Lengai. In: Bell K. and Keller J. (eds). Carbonatite volcanism-Oldoinyo Lengai and the petrogenesis of natrocarbonatites. IAVCEI Proceedings in volcanology 4. Berlin, Springer verlag, pp. 100-112. Bell, K. and Keller, J., 1995. Carbonatite volcanism: Oldoinyo Lengai and the petrogenesis of natrocarbonatite. IAVCEI proceedings in volcanology 4, pp 210. Bell K. and Blenkinsop J., 1987. Nd and Sr isotopic compositions of East African carbonatites: implications for mantle heterogeneity. Geology 15, 99-102. Brueckner H. K. and Rex D. C., 1980. K-Ar and Rb-Sr geochronology and Sr isotopic study of the Alnö alkaline complex, northeastern Sweden. Lithos 13, 111-119. Chakhmouradian A.R. and Mitchell R.H., 2003. The role of assimilated silica in the evolution of carbonatites. 4th Eurocarb Workshop, Canary Islands, Spain. 16th-21st September, 2003. Abstract volume, p 10-11. Church A. A. and Jones A. P., 1995. Silicate-carbonate immiscibility at Oldoinyo Lengai. Journal of Petrology 36, 869-889. Cohen R.S., O’Nions R.K. and Dawson J.B., 1984. Isotope geochemistry of xenoliths from East Africa: implications for development of mantle reservoirs and their interaction. Earth and Planetary Sciences Letters 68, 209-220. Cooper A. F. and Reid D. L., 1998. Nepheline sövites as parental magmas in carbonatite complexes: evidence from Dicker Willem, southwest Namibia. Journal of Petrology 39, 2123-2136. Dalton J.A and Presnall D.C., 1996. Melting of a model carbonated lherzolite in the system CaO-MgO-Al2O3-SiO2-CO2 at 6.0 GPa: the continuous transition between carbonatite and group 1B kimberlite magmas. EOS Transactions, American Geophysical Union 77, 817. Dalton J.A and Presnall D.C., 1995. Phase relations in the system CaO-MgO-Al2O3-SiO2-CO2 from 4.0 to 6.0 GPa with applications to the generation of kimberlites and carbonatites. EOS Transactions, American Geophysical Union 76, 697. Dalton J.A. and Wood B.J, 1993. The composition of primary carbonate melts and their evolution through wall rock reaction in the mantle. Earth and Planetary Science Letters 119, 511-525. Dawson J.B., 1962a. The geology of Oldoinyo Lengai. Bulletin of Volcanology24, 348-387. Dawson J. B., 1962b. Sodium carbonate lavas from Oldoinyo Lengai, Tanzania. Nature 195, 1075-1076. Dunworth E.A. and Bell K., 2001. The Turyi Massif, Kola Peninsula, Russia: Isotopic and Geochemical Evidence for Multi-source Evolution. Journal of Petrology 42, 377-405. Dunworth, E. A. and Bell, K., 1998. Melilitolites: A new scheme of classification. Canadian Mineralogist 36, 895-903. Eckermann H. von, 1948. The alkaline district of Alnö island. Swedish Geological Survey, ser Ca no 36,

26

pp 176. Eggler D.H, 1989. Carbonatites, primary melts and mantle dynamics. In: Bell, K. (ed). Carbonatites - genesis and evolution. London, Unwin Hyman, pp 561-579. Gittins, J. and Harmer, R.E., 1997. What is ferrocarbonatite? A revised classification. Journal of African Earth Sciences 25, 159-168. Halama R., Marks M. and Markl G., 2003. NdO isotopic and trace element constraints on the petrogenesis of the Grönnedal-Ika carbonatite-syenite complex in the Gardar Province, South Greeenland. 4th Eurocarb Workshop, Canary Islands, Spain. 16th-21st September, 2003. Abstract volume, p 63-64. Hammouda T. and Laporte D., 2000. Ultrafast mantle impregnation by carbonatite melts. Geology 28, 283-285. Harmer R.E., 1999. The petrogenetic association of carbonatite and alkaline magmatism: constraints from the Spitskop complex, South Africa. Journal of Petrology 40, 525-548. Harmer R.E. and Gittins J., 1998. The case for primary mantle-derived carbonatite magma. Journal of Petrology 39, 1895-1903. Harmer R.E. and Gittins J., 1997. The origin of dolomitic carbonatites: field and experimental constraints. Journal of African Earth Sciences 25, 5-28. Harmer R.E., Lee C.A. and Elington B.M., 1998. A deep mantle source for carbonatite magmatism: evidence from the nephelinites and carbonatites of the Buhera district, SE Zimbabwe. Earth and Planetary Science Letters 158, 131-142. Hart S.R., 1988. Heterogeous mantle domains: signatures, genesis and mixing chronologies. Earth and Planetary Sciences Letters 90, 273-296. Hoernle K.A. and Tilton G.R., 1991. Sr-Nd-Pb isotope data for Fuerteventura (Canary Islands) basal complex and subaerial volcanics: application to magma genesis and evolution. Schwezerische Mineralogische und Petrographische Mitteilungen 71, 3-18. Huang, Y.M, Hawkesworth, C.J., van Calsteren, P. and McDermott, F., 1995. Geochemical characteristics and origin og the Jacupiranga carbonatites, Brazil. Chemical Geology 119, 79-99. Hunter R.H. and McKenzie D., 1989. The equilibrium geometry of carbonate melts in rocks of mantle composition. Earth and Planetary Science Letters 92, 347-356. Jones J.H., Walker D., Pickett D.A., Murrell M.T. and Beattie P., 1995. Experimental investigation of the partitioning of Nb, Mo, Ba, Ce, Pb, Ra, Th, Pa and U between immiscible carbonate and silicate liquids. Geochimica et Cosmochimica Acta, 59, 1307-1320. Keppler, H., 2003. Water solubility in carbonatite melts. American Mineralogist 88, 1822-1824. Kjarsgaard B.A., Hamilton D.L. and Peterson T., 1995. Peralkaline carbonatite/nephelinite liquid immiscibility: comparison of phase compositions in experiments and natural lavas from Oldoinyo Lengai. In: Bell, K. and Keller, J. (eds). Carbonatite volcanism-Oldoinyo Lengai and the petrogenesis of natrocarbonatites. IAVCEI Proceedings in volcanology 4. Berlin, Springer verlag, pp. 163-190.

Kjarsgaard B.A. and Peterson T., 1991. Nephelinite-carbonatite liquid immiscibility at Shombole volcano, East Africa: petrographic and experimental evidence. Mineralogy and Petrology 43, 292-314. Kjarsgaard B.A. and Hamilton D.L., 1989a. The genesis of carbonatites by liquid immiscibility. In: Bell, K. (ed). Carbonatites - genesis and evolution. London, Unwin Hyman, pp 388-404. Kjarsgaard B.A. and Hamilton D.L., 1988. Liquid immiscibility and the origin of alkali-poor carbonatites. Mineralogical Magazine 52, 43-55. Kogarko, L.N., Henderson, C.M. and Pacheco, H., 1995. Primary Ca-rich carbonaite magma and carbonate-silicate-sulfide liquid immiscibility in the upper mantle. Contributions to Mineralogy and Petrology 121, 267-274. Kramm U, 1994. Isotope evidence for ijolite formation by fenitization: Sr-Nd data of ijolites from the type locality Iivaara, Finland. Contributions to Mineralogy and Petrology 115, 279-286. Kramm, U., 1993. Mantle components in carbonatites from the Kola alkaline province, Russia and Finland: a Nd-Sr study. European journal of Mineralogy 5, 985-989. Kramm U. and Sindern S., 1998. Nd and Sr isotope signatures of fenites from Oldoinyo Lengai, Tanzania, and the genetic relationship between nephelinites, phonolites and carbonatites. Journal of Petrology 39, 1997-2004. Kramm U., Maravic H. V. and Morteani G., 1997. Neodymium and Sr isotopic constraints on the petrogenetic relationships between carbonatites and cancrinite syenites from the Lueshe alkaline complex, east Zaire. Journal of African Earth Sciences 25, 55-76. Kramm U. and Kogarko L.N., 1994. Nd and Sr isotope signatures of the Khibina and Lovozero agpaitic centres, Kola alkaline province, Russia. Lithos 32, 225-242. Kresten P., 1990. The Alnö area (Alnöområdet) in: Lundqvist T., Gee D., Kumpulainen R., Karis L. and Kresten P. (eds), 1990. Beskrivning till berggrundskartan över Västernorrlands län. Sveriges geologiska undersökningar ser Ba nr 31, 238-278. Kresten P., 1988. The chemistry of fenitisation: examples from Fen, Norway. Chemical Geology 68, 329-349. Kresten P. 1979. The Alnö complex: discussion of the main features, bibliography and excursion guide. Nordic carbonatite symposium, pp 67. Kresten, P., 1976. A magnetometric survey of the Alnö complex. Geologiska Föreningens i Stockholm Förhandlingar 98, 364-365. Kresten P. and Morogan V., 1986. Fenitisation at the Fen complex, southern Norway. Lithos 19, 27-42. Kresten P., Printzlau I., Rex D.C., Vartiainen H. and Woolley A.R., 1977. New ages of carbonatitic and alkaline ultramafic rocks from Sweden and Finland. Geologiska Föreningens i Stockholm Förhandlingar 99, 62-65. Le Bas, M.J., 1999. Sövite and alvikite: two chemically distinct calciocarbonatites C1 and C2. South African Journal of Geology 102, 109-121. Le Bas M.J., 1989. Diversification of carbonatite.

27

In: Bell, K. (Ed). Carbonatites - genesis and evolution. London, Unwin Hyman, pp 428-447. Lee W-.J. and Wyllie P.J., 2000a. The system CaO-MgO-SiO2-CO2 at 1.0 GPa, metasomatic wehrlites, and primary carbonatite magmas. Contributions to Mineralogy and Petrology 138, 214-228. Lee W-.J and Wyllie P.J., 1998a. Petrogenesis of carbonatite magmas from Mantle to Crust, Constrained by the system CaO-(MgO+FeO*)-(Na2O+K2O)-(SiO2+Al2O3+TiO2)-CO2. Journal of Petrology 39, 495-517. Lee W-.J. and Wyllie P.J., 1998b. Processes of crustal carbonatite formation by Liquid Immiscibility and Differentiation, Elucidated by Model System. Journal of Petrology 39, 2005-2013. Lee W-.J. and Wyllie P.J., 1997a. Liquid immiscibility between nephelinite and carbonatite from 2.5 to 1.0 GPa compared with mantle melt compositions. Contributions to Mineralogy and Petrology 127, 1-12. Lee W-.J. and Wyllie P.J., 1997b. Liquid immiscibility in the join NaAlSiO4-NaAlSi3O8 CaCO3 at 1 GPa: Implications for crustal carbonatites. Journal of Petrology 38, 1113-1135. Lee, W-.J. and Wyllie, P.J., 1996. Liquid immiscibility in the join NaAlSi3O8-CaCO3 to 2.5 GPa and the origin of calciocarbonatite magmas. Journal of Petrology 37, 1125-1152. Lee W-.J. and Wyllie P.J., 1994. Experimental data bearing on liquid immiscibility, crystal fractionation, and the origin of calciocarbonatites and natrocarbonatites. International Geology Review 36, 797-819. Lee, W.-J., Fanelli, M.F., Cava, N. and Wyllie, P.J., 2000. Calciocarbonatite and magnesiocarbonatite rocks and magmas represented in the system CaO-MgO-CO2-H2O at 0.2 GPa. Mineralogy and Petrology 68, 225-256. Le Maitre R. W., 2002 (Ed). Igneous rocks: a classification and glossary of terms. Cambridge University Press. 236 pp. MacDonald R., Kjarsgaard B.A., Skilling I.P., Davies G.R., Hamilton D.L. and Black S., 1993. Liquid immiscibility between trachyte and carbonate in ash flow tuffs from Kenya. Contributions to Mineralogy and Petrology 114, 267-287. Mahotkin I.L., Gibson S.A., Thompson R.N., Zhuravlev D.Z. and Zherdev P.U., 2000. Late Devonian diamondiferous kimberlite and alkaline picrite (proto-kimberlite?) magmatism in the Archangelsk region, NW Russia. Journal of Petrology 41, 201-227. Martin R.F., Whitley J.E. and Woolley A.R., 1978. An investigation of rare-earth mobility: Fenitized quartzites, Borrolan Complex, Scotland. Contributions to Mineralogy and Petrology 66, 69-73.

Marty, B., Tolstikhin, I., Kamnesky, I.L., Nivin, V., Balaganskaya, E. and Zimmermann, J.L., 1998. Plume derived rare gases in a 380 Ma carbonatite from the Kola Region (Russia) and the argon composition of the mantle. Earth and Planetary Science Letters 164, 179-192. McKenzie D., 1985. The extraction of magma from the crust and mantle. Earth and Planetary Science Letters 74, 81-91. Morogan V., 1994. Ijolite versus carbonatite as sources of fenitisation. Terra Nova 6, 166-176. Morogan V., 1989. Mass transfer and REE mobility

during fenitization at Alnö, Sweden. Contributions to Mineralogy and Petrology 103, 25-34. Morogan, V., 1988. Fenitisation - Hallmark of the ijolite-carbonatite magmatic association. Unpublished PhD-thesis, Stockholm University. Morogan M. and Lindblom S., 1995. Volatiles associated with the alkaline-carbonatite magmatism at Alnö, Sweden: a study of fluid and solid inclusions in minerals from the Långarsholmen ring complex. Contributions to Mineralogy and Petrology 122, 262-274. Morogan V. and Woolley A. R., 1988. Fenitization at the Alnö carbonatite complex, Sweden: distribution, mineralogy and genesis. Contributions to Mineralogy and Petrology 100, 169-182. Morogan V. and Martin R.F., 1985. Mineralogy and partial melting of fenitized crustal xenoliths in the Oldoinyo Lengai carbonatite volcano, Tanzania. American Mineralogist 70, 1114-1126. Nelson D.R., Chivas A.R., Chappell B.W.and McCulloch M.T., 1988. Geochemical and isotopic systematics in carbonatites and implications for the evolutions of ocean-island sources. Geochimica et Cosmochimica Acta 52, 1-17. Nielsen, T.F.D. and Veksler, I.V., 2002. Is natrocarbonatite a congate fluid condensate? Contributions to Mineralogy and Petrology 142, 425-435. Nielsen, T.D.F., Solovova, I.P. and Veksler, I.V., 1997. Partial melts of meliitolite and origin of alkaline carbonatite: evidence from crystallised melt inclusions, Gardiner complex. Contributions to Mineralogy and Petrology 126, 331-344 Olafson M. and Eggler D.H., 1983. Phase relations of amphibole, amphibole-carbonate and phlogopite-carbonate peridotite: Petrologic constraints on the astenosphere. Earth and Planetary Science Letters, 64, 305-315. Otto J.W. and Wyllie P.J., 1993. Relationships between silicate melts and carbonate precipitating melts in CaO-MgO-SiO2-CO2-H2O at 2 kbar. Mineralogy and Petrology 48, 343-365.

Pati, J.K., Arima, M. and Gupta, A.K., 2000. Experimental study of the system diopside albite-nepheline at P(H2O)=2 and 10 kbar and at P(total)=28 kbar. Canadian Mineralogist 38, 1177-1191.

Pearce, N.J.G. and Leng, M.J., 1996. The origin of carbonatites and related rocks from the Igaliko dyke swarm, Gardar Province, South Greenalnd: field, geochemical and C-O-Sr-Nd isotope evidence. LIthos 39, 21-40. Peterson T.D., 1990. Petrology and genesis of natrocarbonatite. Contributions to Mineralogy and Petrology 105, 143-155. Pyle, J.M. and Haggerty, S.E., 1994. Silicate-carbonate immiscibility in upper-mantle eclogites: implications for natrosilicic and carbonatite conjugate melts. Geochimica et Cosmochimica Acta 58, 2997-3011. Rankin, A.H., 1977. Fluid inclusion evidence for the formation conditions of apatite from the Tororo carbonatite complex of eatern Uganda. Mineralogical Magazine 41, 155-164. Rankin, A.H. and LeBas, M.J., 1974. Nahkolite

28

(NaHCO3) in inclusions in apatite from some E. African ijolites and carbonatiets. Mineralogical Magazine 39, 564-570. Ray, J.S., 1998. Trace element and isotope evolution during concurrent assimilation, fractional crystallization and liquid immiscibility of a carbonated silicate magma. Geochimica et Cosmochimica Acta, 62, 3301-3306. Ray, J.S., Ramesh, R, Pande, K., Trivedi, J.R., Shukla P.N. and Patel, P.P., 2000. Isotope and rare earth element chemistry of carbonatite-alkaline complexes of Deccan volcanic province: implications to magmatic and alteration processes. Journal of Asian Earth Sciences 18, 177-194. Riley T. R., Bailey D. K., Harmer R. E., Liebsch H., Lloyd F. E. and Palmer M. R., 1999: Isotopic and geochemical investigation of a carbonatite-syenite-phonolite diatreme, west Eifel (Germany). Mineralogical Magazine 63 (5), 615-631. Ruberti E., Castorina F., Censi P., Comin-Chiaramonti P., Gomes C.B., Antonini P. and Andrade F.R.D., 2002. The geochemistry of the Barra do Itapirapuã carbonatite (Ponta Grossa Arch, Brazil): a multiple stockwork. Journal of South American Earth Sciences 15, 215-228. Samson, I.M, Williams-Jones, A.E. and Weining Liu, 1995. The chemistry of hydrothermal fluids in carbonatites: evidence from leachate and SEM-decrepitate analysis of fluid inclusions from Oka, Quebec, Canada. Geochimica et Cosmochimica Acta 59, 1979-1989. Schmulovich, K., Graham, C. and Yardley, B., 2001. Quartz, albite and diopside solubilities in H2O-NaCl and H2O-CO2 fluids at 0.5-0.9 GPa. Contributions to Mineralogy and Petrology 141, 95-108. Simonetti A., Goldstein S. L., Schmidberger S. S. and Viladkar S. G., 1998. Geochemical and Nd, Pb and Sr isotope data from Deccan alkaline complexes: inferences for mantle sources and plume-lithosphere interaction. Journal of Petrology 39, 1847-1864. Simonetti, A. and Bell, K., 1995. Nd, Pb and Sr isotopic data from the Mount Elgon volcano, eastern Uganda-western Kenya: implications for the origin and evolution of nephelinite lavas. Lithos 36, 141-153. Simonetti A. and Bell K., 1994. Isotopic and geochemical investigation of the Chilwa Island Carbonatite Complex, Malawi: evidence for a depleted mantle source region, liquid immiscibility and open-system behaviour. Journal of Petrology 35, 1597-1621. Sindern S. and Kramm U., 2000. Volume characteristics and element transfer of fenite aureoles: a case study from the Iivaara alkaline complex, Finland. Lithos 51:75-93. Stoppa F. and Woolley A.R., 1997. The Italian carbonatites: field occurrence, petrology and regional significance. Mineralogy and Petrology 59, 43-67. Stoppa F. and Cundari A., 1995. A new Italian carbonatite occurrence at Cupaello (Rieti) and its genetic significance. Contributions to Mineralogy and Petrology 122, 275-288. Sörensen H., 1974. The alkaline rocks. John Wiley and Sons.

Treiman A.H. and Essene E.J., 1985. The Oka carbonatite complex, Quebec: Geology and evidence for silicate-carbonate liquid immiscibility. American Mineralogist 70, 1101-1113. Törnebohm A.E., 1882. Mikroskopiska bergartsstudier XVII. Melilitbasalt från Alnö. Geologiska Föreningens i Stockholm Förhandlingar 6, 240-251. Veena K., Pandey B. K., Krishnamurthy P. and Gupta J. N., 1998. Pb, Sr and Nd isotope systematics of the carbonatites of Sung Valley, Meghalaya, northeast India: implications for contemporary plume-related mantle source characteristics. Journal of Petrology 39, 1875-1884. Veksler, I.V and Keppler, H., 2000. Partitioning of Mg, Ca and Na between carbonatite melt and hydrous fluid at 0.1-0.2 GPa. Contributions to Mineralogy and Petrology 138, 27-34. Veksler I.V., Nielsen T.F.D. and Sokolov S.V., 1998a. Mineralogy of crystallized melt inclusions from Gardiner and Kovdor ultramafic alkaline complexes: implications for carbonatite genesis. Journal of Petrology 39, 2015-2031. Veksler I.V., Petibon C., Jenner G.A., Dorfman A.M and Dingwell D.B., 1998b. Trace element partitioning in immiscible silicate-carbonate liquid systems: an initial experimental study using a centrifuge autoclave. Journal of Petrology 39, 2095-2104. Verhulst A., Balaganskaya E., Kirnarsky Y. and Demaiffe D., 2000. Petrological and geochemical (trace elements and Sr-Nd isotopes) characteristics of the Paleozoic Kovdor ultramafic, alkaline and carbonatite intrusion (Kola Peninsula, NW Russia). Lithos 51, 1-25. Wyllie P.J. and Lee W-.J., 1998. Model system controls on conditions for formation of magnesiocarbonatite and calciocarbonatite magmas from the mantle. Journal of Petrology 39, 1885-1893. Wyllie P.J. and Huang W.-L., 1976a. Carbonation and melting reactions in the system CaO- MgO-SiO2-CO2 at mantle pressures with geophysical and petrological applications. Contributions to Mineralogy and Petrology 54-79-107. Wyllie P.J. and Huang W.-L., 1975. Peridotite, kimberlite and carbonatite explained in the system CaO-MgO-SiO2-CO2. Geology 3, 621-624. Yang, H.-J., Frey, F.A. and Clague, D.A., 2003. Constraints on the source components on lavas forming the Hawaiian North Arch and Honolulu volcanics. Journal of Petrology 44, 603-627. Ying J., Zhou X. and Zhang H., 2004. Geochemical and isotopic investigation of the Laiwu-Zibo carbonatites from western Shandong Province, China, and implications for their petrogenesis and enriched mantle source. Lithos 75, 413-426.

Zaitsev A. and Bell K., 1995. Sr and Nd isotope data of apatite, calcite and dolomite as indicators of source, and the relationships of phoscorites and carbonatites from the Kovdor massif, Russia. Contributions to Mineralogy and Petrology 121, 324-335.