Journal of Petrology | Volume 39 | Number 11-12 | Pages 1943-1951 | 1998
© Oxford University Press 1998
The Transition from Carbonate to Silicate Melts in the CaO—MgO—SiO2—CO2 System
Department of Earth Sciences, Wills Memorial Building, University of Bristol Queen's Road,Bristol BS8 1RJ, UK
Received September 30, 1997; Revised typescript accepted June 23, 1998
 |
ABSTRACT |
|---|
 TOP ABSTRACT IntroductionExperimental ProcedureResultsDiscussionConclusionsReferences |
|---|
The compositions of melts in equilibrium with a lherzolite mineral assemblage were determined in the analogue system CaO–MgO–SiO2–CO2 at 3 GPa. Carbonate liquids coexist with olivine and two pyroxenes between the solidus fo carbonated lherzolite at 1250°C and 1450°C. The Ca/(Ca + Mg) ratio of these melts is 0.64 and the main effects of rising temperature are increasing SiO2 content (from <4.3 to 7.5 wt %) and decreasing CO2 content. Between 1475°C and
1525°C the SiO2 conten of th liquid increases dramatically from 10 to 30% and, thereafter, the CO2 content decreases rapidly as the CO2-absent invariant point (at >1700°C is approached. The progression from carbonate to silicate liquids is, therefore abrupt and the field of transitional compositions (10–30% SiO2) is restricted to very narrow temperature intervals at pressures greater than the solidu ledge. All liquids appear to be miscible. In the context of upwelling magma, our result provide possible insight into the origins of complexes that are considered to contain primary carbonatites. The solidus ‘ledge’ between 2.5 and 3 GPa acts a a filter for both carbonatites and transitional melt compositions. Carbonatites, which have a wide stability field at 3 GPa, may rise through the mantle if they are isolated from lherzolite by wallrock reaction and production of wehrlite. Transitional carbonate–silicate melts must also, however, react with the mantle at low pressures. This fact, combined with the small range of physical conditions over which they are generated and higher (than carbonatite) viscosity, means that they rarely reach crustal levels. Low-CO2 silicate melts, in contrast, are not required to react extensively en route to the surface and are abundant. We suggest that the binary nature of some carbonatite complexes may be controlled by the compositions of primary mantle melts produced at pressures greater tha the solidus ledge. KEY WORDS: carbonatites; primary liquids; CO2 saturation; metasomatism
|
Introduction |
|---|
|
TOPABSTRACTIntroductionExperimental ProcedureResultsDiscussionConclusionsReferences |
|---|
The common association of carbonatites with silica-undersaturated alkaline igneous rocks has led to many speculative petrogenetic schemes for their generation. These includ limestone assimilation (Shand, 1945
) to produce silica-undersaturated rocks crystal fractionation to produce carbonatite from a primary CO2-rich silicate melt (Watkinson & Wyllie, 1971) and liquid immiscibility to produc both rock types from an intermediate parent (Freestone & Hamilton, 1980). It has become apparent from isotopic studies (e.g. Powell et al., 1966; Bell & Blenkinsop, 1987a, 1987b) that the primitive magma is of mantle origin. An understanding of the genesis of associated carbonate and silicate rocks therefore depends on determining the compositions of melts that are produced from peridotitic compositions under upper-mantle conditions. This study is aimed at the elucidation of melting behaviour in mantle analogue systems.
Qualitative interpretation of the behaviour of carbonated peridotite at solidus and supersolidus conditions can be derived from the model system CaO–MgO–SiO2–CO2 (CMS–CO2) (Fig. 1). Decarbonation reactions (Wyllie & Huang, 1975a, 1975b, 1976; Eggler, 1978) define the stability fields of carbonate in lherzolitic mantle. The following reaction separates regions in the mantle where CO2 exists as a free fluid phase and where it resides in mineral phases (dolomite):
 |
 |
; Eggler, 1978; White & Wyllie, 1992). The melt at the solidus in this region approximates a calcic dolomite carbonatite. The position of the back-bend is, in the simple system, constrained by the intersection of the carbonation reaction (1) with the solidus at 2.8 GPa. Liquid compositions may be represented by projection from CO2 onto the CaO–MgO–SiO2 side of the compositional tetrahedron, as shown in Fig. 2. At 2 GPa, the first melt of carbonated lherzolite (at peritectic T) is approximately basaltic, containing less than 5 wt % CO2 (Wyllie & Huang, 1976). Carbonate-rich liquids, which exist near the solidus as small fraction melts, cannot be in equilibrium with lherzolite at this pressure, and instead coexist with olivine and clinopyroxene (Fig. 2a). The association of carbonate liquid with the phases olivine and clinopyroxene has been used to link the frequent occurrence of wehrlite veins in peridotite to low-pressure metasomatism by ascending carbonate melts (Dalton & Wood, 1993; Rudnick et al., 1993). Above 2.5 GPa, the peritectic T moves rapidly in the approximate direction of the vapour-saturated eutectic E (Wyllie & Huang, 1976; Eggler, 1978), merging to generate two new peritectics A and B at 3 GPa (Fig. 2b). In the context of carbonatites and associated rocks, the important point is that carbonate melts produced at 3 GPa can be in equilibrium with harzburgite and lherzolite mineral assemblages in the upper mantle. Thus carbonatite is a viable primary melt of peridotite at a range of depths that may extend to >100 km.
|
|
Given that carbonate melt and carbonated silicate melts can be produced at pressures near 3 GPa, the relationships between them and the compositional paths taken en route t the surface remain to be established. Melts in equilibrium with orthopyroxene-bearing assemblages are dolomitic and have a maximum Ca/(Ca + Mg) at 3 GPa of
0.7 (Dalton & Wood, 1993). Carbonatite rocks have Ca/(Ca + Mg) averages of 0.5 and 0.86 for dolomitic carbonatites and calcio-carbonatites, respectively [based on the data of Woolley & Kempe (1989)], with the latter being more common. Current experimental data indicate that calcio-carbonatites cannot equilibrate with harzburgite or lherzolite and must evolve from a primary mantle melt by fractional crystallization or wallrock reaction. At low pressure the latter is due to the instability of orthopyroxene in carbonate melt (Fig. 2a) so that ascending carbonate melts react with this phase, enriching the melt in the CaCO3 component and evolving CO2:
 |
extend to Ca/(Ca + Mg) ratios as high as 0.87, approximating those observed in many carbonatites.
At pressures greater than those of the solidus ‘ledge’ of Fig. 1 and at temperatures considerably higher than the solidus at 3 GPa, harzburgite and lherzolite mineral assemblages are in equilibrium with CO2-rich silicate melts. Although the relationships between these melts and eruptive rocks are uncertain, it has been demonstrated that some carbonated silicate rocks are plausible primary mantle melts. For example, Brey & Green (1977) and Brey (1978) have found that olivine nephelinite and olivine melilitite may be generated from lherzolitic mantle at pressures between 2.7 and 3.5 GPa. The petrogenetic links between the different rock types of the carbonate–alkali silicate association remain unclear, however. In this study we have attempted to elucidate them by investigating the nature of the transition from carbonate to silicate melts at 3 GPa in the CaO–MgO–SiO2–CO2 system, the simplest model for the range of primary melts generated from carbonated lherzolite.
|
Experimental Procedure |
|---|
|
TOPABSTRACTIntroductionExperimental ProcedureResultsDiscussionConclusionsReferences |
|---|
Three starting compositions were used (Table 1): (a) MgSiO368.5–CaMg(CO3)231.5 (wt %); (b) MgSiO3 44.5–Mg2SiO413.7–CaMgSi2O610.3–CaMg(CO3)231.5; (c) MgSiO317–Mg2SiO431–CaMgSi2O642–CaMg(CO3)210.These were prepared by mixing and grinding analytical grade oxides and carbonates under ethanol to produce silicate and carbonate powders. These were packed into 2 mm o.d. platinum capsules in layers with carbonate at the base of the capsule so that it did not dissociate with loss of CO2 during welding. Proportions of silicate and carbonate were chosen to be comparable with previous studies and to produce large enough melt pools (normally >50µm diameter) to be analysed by electron microprobe. Packed capsules were stored open-ended overnight in a drying oven to ensure that no atmospheric water absorbed by the powders during packing was retained. The weight of capsules was noted before and after arc-welding to confirm that there had been insignificant decarbonation of the starting mixture during the welding process.
|
Experiments were performed in a piston-cylinder apparatus using barium carbonate and crushable alumina pressure media. The capsule was surrounded by a tightly packed mixtur of haematite and Pyrex to inhibit ingress of hydrogen into the capsule. Temperature was monitored and controlled directly above the capsule by the use of a calibrated W–Re thermocouple. Using the piston-in technique, with a pressure correction of –10%, experiments were run at 3 GPa and temperatures between 1300 and 1500°C for 3 h. Experiments above 1500°C were run for 2 h and those at 1600°C for 40 min (Table 1).
Quenching was performed by turning off the power supply to the graphite furnace. The resulting charge was mounted in epoxy resin and prepared for analysis. Charges were ground and impregnated with resin at frequent intervals to minimize plucking of carbonate and quenched silicate phases, which can be severe in such experiments. They were then polished by hand, using diamond pastes and oiled-based lubricants on polishing pads uncontaminated with water, to produce a flat surface for electron microprobe analysis.
Analysis was carried out using the JEOL 8600 electron microprobe in the Geology Department at the University of Bristol. Compositions of silicate phases were obtained using spot analyses and a beam current of 15 nA. Melts, with their high carbonate content, devolatilize easily under the electron beam and were analysed using a beam current of 5 nA and a diffuse spot (beam rastered over an area up to 20 µm2). Composition of liquids that quenched to dendritic crystals were obtained by averaging a large sampl number of melt analyses (Table 2). Analyses were carried out at 15 kV usin olivine (SJIO), quartz and wollastonite as standards, with results processed using ZAF procedures. To check for errors introduced by not analysing for CO2 we performed a number of analyses of carbonate standards and of our quenched melts in which neighbouring points were processed by (a) ignoring CO2 completely and (b) adding oxygen to the ZAF correction until the analysis totalled 100%. Method (a) is the one that is normally adopted, but method (b) should be more accurate if one is certain that the surface is absolutely flat and only one phase is being analysed. The results show that method (a) gives a linear increase in the error in Ca/(Ca + Mg) with CO2 content from 0.0 at 0.0% CO2 to +0.028 at 50% CO2. Thus, for essentially pure dolomitic melts Ca/(Ca + Mg) appears to be overestimated by 0.028. Data in Table 2 have not been corrected for this effect.
|
|
Results |
|---|
|
TOPABSTRACTIntroductionExperimental ProcedureResultsDiscussionConclusionsReferences |
|---|
Enstatite–dolomite join
An experiment run at 1250°C using starting composition (a), MgSiO3 68.5 (top layer)–CaMg(CO3)2 31.5 (base layer) wt %, produced liquid in such small amounts that it could not be reliably analysed and is therefore not included in the tables. From this we inferred that the solidus for composition (a) lies very close to 1250°C, in agreement with previous work (Eggler, 1978
, 1987; Wyllie, 1987). Near-solidus meltin in this composition produced enstatite and a dolomitic melt of homogeneous appearance and a Ca/(Ca + Mg) ratio of 0.51 (Table 2, Fig. 3). The texture of the quenched melt remains homogeneous up to 1400°C, at which temperature hopper crystals of olivine appear and orthopyroxenes have a corroded appearance Although all melt compositions contain some silica, this is the first textural evidence observed that silicate phases are participating in the melting reaction. At 1450°C equant crystals of forsteritic olivine and diopsidic clinopyroxene join orthopyroxene in the silicate phase assemblage and the dolomitic melt has a Ca/(Ca + Mg) ratio of 0.55 (Fig. 3). Clinopyroxene reacts out at 1475°C as Ca/C + Mg) of the liquid reaches 0.58. Between 1475 and 1500°C extensive melting of silicate phases dramatically changes the nature of the liquid from being dominantly carbonate to a silicate composition, which quenches to a heterogeneous mixture of crystals rather than to a glass. In all cases gas was produced on rupture of the capsule, indicating that either excess CO2 was present and/or CO2 was evolved during quenching.
|
Melt composition (Fig. 3) varies little over the first 150–200°C above the solidus, with SiO2 content increasing from 5 to 8 wt %. Then, in the temperature interval between 1475°C and 1500°C the SiO2 content more than triples as the liquid changes from carbonate to silicate, with concomitant lowering of the Ca/(Ca + Mg) ratio from 0.57 to 0.28. At temperatures higher than 1500°C olivine reacts out and orthopyroxene is left as the liquidus phase.
Haplolherzolite–dolomite
Composition (b) was chosen as a simple analogue of carbonated lherzolite to generate near-solidus liquids in equilibrium with a lherzolitic residuum. The melt at 1300°C does indeed coexist with olivine, orthopyroxene and abundant clinopyroxene, and has Ca/(Ca + Mg) of 0.64 (Table 2, Fig. 3). Clinopyroxene disappears at 1400°C, and the melt, coexisting with olivine and orthopyroxene, begins to precipitate large silicate quench crystals during quenching from 1450°C. At 1500°C the carbonated silicate melt coexists with orthopyroxene only.
It is clear from our results on compositions (a) and (b) that, despite the evolution of gas from the capsules on puncturing, the liquids generated in this study were not all CO2 saturated. This is required by the fact that, isobarically, we have generated phase assemblages containing four condensed phases (ol–opx–cpx–L) over a temperature interval of 1300–1450°C (Fig. 4). We have therefore determined part of the four-phase cotectic at 30 kbar, that part in equilibrium with carbonate melts and hence close to CO2 saturation. To characterize this cotectic more fully we performed experiments on a CO2–poor composition (c), MgSiO3 17–Mg2SiO4 31–CaMgSi2O6 42–CaMg(CO3)2 10 (wt %). At 1600°C this produces a carbonated silicate melt containing 37% SiO2 (Table 2, Fig. 4), with Ca/;(Ca + Mg) of 0.36 (Fig. 3), which coexists with olivine, orthopyroxene and clinopyroxene.
|
|
Discussion |
|---|
|
TOPABSTRACTIntroductionExperimental ProcedureResultsDiscussionConclusionsReferences |
|---|
Figures 4 and 5 imply that all liquid compositions between carbonate and CO2-free silicate can be produced isobarically in equilibrium with lherzolite (in the simple system CMS–CO2) by partial melting at varying temperatures. For starting compositions with fixed carbonate content, melt compositions lie on the four-phase cotectic projected in Fig. 5 and the CO2 content of the liquid decreases with rising temperature. The end-point of the cotectic, corresponding to the volatile-free system, was taken from Eggler (1978)
, after Kushiro (1968). The rapid change in composition at 1475 ± 25°C is the result of melting generated in a system with few phases and there is no evidence that carbonate and CO2-free silicate liquids are immiscible liquids.
|
) at 1300, 1450 and 1600°C were used to construct the ol–cpx–opx–L cotectic (dashed line). Melt compositions in equilibrium with harzburgite are also shown (
). The end-point of the cotectic, corresponding to the volatile-free system, was taken from Eggler (1978)Comparison of the trace of the CO2-undersaturated cotectic at 3 GPa with the CO2-saturated peritectic (ol + opx + cpx + L) at 2 GPa in Fig. 2a (Wyllie & Huang, 1976
) shows that the latter has similar Ca/(Ca + Mg) to the low PCO2 end of the 3 GPa cotectic. This demonstrates that the progression from carbonate to silicate melt in equilibrium with lherzolite can occur either by raising temperature isobarically or by holding the peridotite just above its solidus and lowering pressure at CO2 saturation. In the latter case, the initial liquid follows the trace of the peritectic from CO2-rich (at high pressure) to CO2-poor melts at low pressure. Olafsson & Eggler (1983) noted a similar progression over the pressure interval 22–17.5 kbar in the presence of both H2O and CO2.
As can be seen from Figs 2 and 5, the progression from carbonate to silicate melts both iso- and polybarically is abrupt and the field of transitional compositions in equilibrium with lherzolite is small. At 3 GPa, for example, melts are carbonatitic over the temperature range 1250–1475°C and SiO2 rich from about 1525 to 1800°C. The field of transitional melt compositions, that is, those liquids having SiO2 content between 10 and 30 wt %, in equilibrium with lherzolite is restricted to a narrow temperature interval of 50°C. Similarly, near-solidus melts shift from carbonate to silicate over the very small pressure interval 2.8–2.5 GPa as shown in Fig. 1. Although the transitional region must be broadened by additional components such as Na2O, H2O, P2O5 and F (Jago & Gittins, 1991) preliminary experiments on Na- and P-bearing compositions in our laboratory suggest tha the field of transitional melts remains narrow even in bulk compositions more closely approximating the natural system.
These results demonstrate, therefore, that the association of carbonatites with nephelinites–melilitites could be a primary feature of mantle melting and that transitional compositions should in this case, as observed, be rare. While recognizing that carbonate–silicate liquid immiscibility is frequently observed in nature (e.g. Dawson, 1966; Le Bas, 1977; Kjarsgaard & Peterson, 1991; Macdonald et al., 1993; Dawson et al., 1994), the field observations do not demonstrate that it is the major petrogenetic process. The silicate suite of Shombole, for example, includes nephelinites that contain irregular calcite-rich bodies interpreted as immiscible liquid (Peterson, 1989). This is clearly, however, a late-stage immiscibility, which is not necessarily related to the generation of the main carbonate and silicate end-members.
Analyses of carbonites and associated rocks from a number of carbonite complexes are plotted, in projection, in Fig. 6 together with the experimentally produced liquid compositions in equilibrium with lherzolite at 3 GPa. As noted by Bailey (1989), primary carbonate melts formed experimentally from carbonated lherzolite at high pressure closely approximate natural dolomitic carbonatites. As previously shown by Brey & Green (1975), observed olivine melilitite compositions also fall within the range of liquids that can coexist with the model lherzolite mineral assemblage at 3 GPa. The nephelinite–phonolite association could, in this simplified projection, be interpreted in terms of fractional crystallization (dominantly olivine) from a CO2–bearing silicate parent, whereas calciocarbonatites would be produced from primary dolomitic carbonatites partly by wallrock reaction (Dalton & Wood, 1993) and partly by fractional crystallization (Harmer & Gittins, 1997).
|
Bailey (1989
, 1993) has provided extensive evidence, based on xenocrysts, that many effusive calcio- and magnesio-carbonatites are of mantle origin unrelated, as suggested above, to liquid immiscibility. As such carbonate melts are the first liquids to form above the mantle solidus at high pressure their low viscosities should ensure that they are the earliest components of high-level complexes. Carbonatites with accompanying silicate volcanism are, however, frequently observed to be late rathe than early in the intrusive sequence. Advocates of a primary origin for carbonatites in complexes cite ‘thermal death’ as the delaying factor in their emplacement (Bailey, 1987). That is, near-solidus liquids formed at pressures greater than the solidus ledge (Fig. 7) ascend through the mantle only until they encounter the ledge. An ascending carbonate melt that maintains equilibrium with surrounding lherzolite will crystallize and solidify by the reaction of orthopyroxene with the liquid to produce olivine, clinopyroxene and carbon dioxide vapour, effectively metasomatizing the mantle (Wyllie, 1980; Schneider & Eggler, 1986; Wallace & Green, 1988; Haggerty, 1989; Meen et al., 1989; Dalton & Wood, 1993) at the depth of reaction. Successive melts meeting the ledge and freezing extend the altered zone until a reaction-lined conduit is formed, which allows ascending carbonatite melts to rise to the point at which evolution of CO2 accelerates the melt sufficiently for it to ‘take off’ to the surface (Bailey, 1985). Thus, the delaying factor is the time interval between the production of the first small fraction melt that freezes at the ledge and the first small fraction melt that ascends to crustal levels. Wyllie, (1980) envisioned released vapour as enhancing the prospects for crack propagation through overlying lithosphere in tension to produce an initial channel to the surface. In this scenario, the delaying factor incorporates the time taken for vapour pressure to increase to levels that will exert a sufficient force on the overlying mantle. At higher degrees of melting, silicate melts ascend and by-pass the ledge where previously formed carbonatite melts may have ponded, metasomatizing the surrounding mantle as they solidify. The silicates that by-pass the ledge ascend to form the large pile of volcanics that make up the vast majority of complexes.
|
As melts transitional between carbonate and silicate are stable only in a very small P–T region, their ascent, as shown in Fig. 7, is also likely to be influenced by the solidus ‘ledge’. Transitional melts [Ca/(Ca + Mg) of 0.55–0.45, SiO2 10–30%] should intersect the solidus ledge and participate in wallrock reaction, thus being trapped by the overlying mantle. The ‘filtering’ effect is likely to be more severe in this case than for carbonatites because of the smaller P–T field over which the transitional melts are generated and their higher viscosities. This effect, when combined with their narrow P–T field of stability, could account for the lack of recorded occurrences of rocks transitional in composition between carbonatites and nephelinites.
|
Conclusions |
|---|
|
TOPABSTRACTIntroductionExperimental ProcedureResultsDiscussionConclusionsReferences |
|---|
We have produced liquid compositions in the system CaO–MgO–SiO2–CO2 that show that the complete compositional range of melts from magnesiocarbonatite [Ca/(Ca + Mg) = 0.64] to volatile-free silicate melt [Ca/(Ca + Mg) = 0.27] is likely to be in equilibrium with the lherzolite mineral assemblage at 3 GPa. In the presence of carbonate the solidus is a
1250°C and the CO2-free invariant point (Fo–En–Di–L) is between 1700°C and 1800°C. The carbonate and silicate end-members are connected by a four-phase cotectic fieldboundary along which CO2 content (and therefore PCO2)decreases with increasing temperature.
We determined the isobaric melting paths for two bulk compositions in the system CaO–MgO–SiO2–CO2. The data show that, for a peridotite with low initial carbonate content, a small amount of carbonatitic melt [Ca/(Ca + Mg) = 0.64]) coexists with olivine + orthopyroxene + clinopyroxene from the solidus at 1250°C to 1475°C. Melt composition in this temperature interval changes steadily, with SiO2 content increasing from 4.3% at 1300°C to 7.5 wt % at 1450°C. Between 1475°C and 1525°C the melt SiO2 content increases rapidly to 28%, thereafter rising to 37% (with 15% CO2) at 1600°C. Carbonated silicate melts have decreasing CO2 contents from 1600°C to the ternary (CaO–MgO–SiO2) invariant point between 1700°C and 1800°C. There is no evidence that the rapid change in melt composition at 1475° ± 25°C is influenced by liquid immiscibility.
The data relate to one possible explanation for the binary nature of carbonatite complexes. Melts dominated by carbonate and silicate compositions are both produced over wide temperature intervals. Transitional compositions are produced over a much narrower temperature interval (50°C in the CMS–CO2 system) and must, like carbonatites themselves, react with the mantle at pressures lower than those of the solidus ‘ledge’. Thus, carbonate melts, formed at high pressure, can only reach the surface after extensive metasomatism of the overlying mantle and production of a wehrlitic ‘pathway’. Transitional compositions need also to react a pathway to reac the surface, but the narrow range of conditions under which they are produced, coupled with relatively high viscosity, compounds the tendency for them to be trapped in the mantle. In contrast, silicate melts with modest CO2 contents escape without significant wallrock reaction and are hence the most voluminous constituents of carbonatite complexes.
|
Acknowledgements |
|---|
The help of our colleagues has proved invaluable in conducting this study. We acknowledge particularly F. Wheeler and M. Dury for construction and maintenance of experimental apparatus, and S. Kearns for expertise regarding microprobe analysis. We thank P. Wyllie, D. Canil, G. Brey and B. Kjarsgaard, who provided helpful reviews of the manuscript. K. R. Moore also acknowledges the receipt of an NERC Postgraduate Studentship.
* Corresponding author. Telephone: 0117 928 9000, ext. 4780, Fax: 0117 925 3385. e-mail: K.R.Moore{at}bris.ac.uk
|
References |
|---|
|
TOPABSTRACTIntroductionExperimental ProcedureResultsDiscussionConclusionsReferences |
|---|
Bailey D. K. Fluids, melts, flowage and styles of eruption in alkaline ultramafic magmatism. Transactions of the Geological Society of South Africa (1985) 88:449–457.
Bailey D. K. Mantle metasomatism–perspective and prospect. In: Alkaline Igneous Rocks. Geological Society, London, Special Publication—Fitton J. G., Upton B. G. J., eds. (1987) 30:1–13.
Bailey D. K. Carbonate melts from the mantle in the volcanoes of south-east Zambia. Nature (1989) 338:415–418.
Bailey D. K. Petrogenetic implications of the timing of alkaline, carbonatite and kimberlite igneous activity in Africa. South African Journal of Geology (1993) 96:67–74.[Abstract]
Baker B. H. Outline of the petrology of the Kenya rift alkaline province. In: Alkaline Igneous Rocks. Geological Society, London, Special Publication—Fitton J. G., Upton B. G. J., eds. (1987) 30:293–311.
Bell K., Blenkinsop J. Archean depleted mantle: evidence from Nd and Sr initial isotopic ratios of carbonatites. Geochimica et Cosmochimica Acta (1987a) 51:291–298.[Web of Science]
Bell K., Blenkinsop J. Nd and Sr isotopic compositions of East African carbonatites: implications for mantle heterogeneities. Geology (1987b) 15:99–102.
Bergman S. C. Lamproites and other potassium-rich igneous rocks: a review of their occurrence, mineralogy and geochemistry. In: Alkaline Igneous Rocks. Geological Society, London, Special Publication—Fitton J. G., Upton B. G. J., eds. (1987) 30:103–190.
Brey G. Origin of olivine melilitites—chemical and experimental constraints. Journal of Volcanology and Geothermal Research (1978) 3:61–88.[Web of Science]
Brey G., Green D. H. The role of CO2 in the genesis of olivine melilitite. Contributions to Mineralogy and Petrology (1975) 49:93–103.[Web of Science]
Brey G., Green D. H. Systematic study of liquidus phase relations in olivine melilitite + H2O + CO2 at high pressures and petrogenesis of an olivine melilitite magma. Contributions to Mineralogy and Petrology (1977) 61:141–162.[Web of Science]
Dalton J. A., Wood B. J. The compositions of primary carbonate melts and their evolution through wallrock reaction in the mantle. Earth and Planetary Science Letters (1993) 119:511–525.[Web of Science]
Dawson J. B. Oldoinyo Lengai—an active volcano with sodium carbonatite lava flows. In: Carbonatites—Tuttle O. F., Gittins J., eds. (1966) New York: John Wiley. 155–168.
Dawson J. B., Pinkerton H., Pyle D. M., Nyamweru C. June 1993 eruption of Oldoinyo Lengai, Tanzania: exceptionally viscous and large carbonatite lava flows and evidence for coexisting silicate and carbonate magmas. Geology (1994) 22:799–802.
Eggler D. H. The effect of CO2 upon partial melting of peridotite in the system Na2O–CaO–Al2O3–MgO–SiO2–CO2 to 35 kb, with an analysis of melting in a peridotite–H2O–CO2 system. American Journal of Science (1978) 278:305–343.
Eggler D. H. Discussion of recent papers on carbonated peridotite, bearing on mantle metasomatism: an alternative. Earth and Planetary Science Letters (1987) 82:398–400.[Web of Science]
Freestone J. C., Hamilton D. L. The role of liquid immiscibility in the genesis of carbonatites–an experimental study. Contributions to Mineralogy and Petrology (1980) 73:105–117.[Web of Science]
Haggerty S. E. Mantle metasomes and the kinship between carbonatites and kimberlites. In: Carbonatites: Genesis and Evolution—Bell K., ed. (1989) London: Unwin Hyman. 546–560.
Harmer R. E., Gittins J. Dolomitic carbonatite parental magmas. In: Abstract Volume, Geological Association of Canada–Mineralogical Association of Canada Annual Meeting (1997) Ottawa: Geological Survey of Canada. A64.
Jackson E. D., Wright T. L. Xenoliths in the Honolulu volcanic series, Hawaii. Journal of Petrology (1970) 11:405–430.
Jago B. B., Gittins J. The role of fluorine in carbonatite magma evolution. Nature (1991) 349:56–58.
Kjarsgaard B. A., Peterson T. Nephelinite-carbonatite liquid immiscibility at Shombole Volcano, East Africa: petrographic and experimental evidence. Mineralogy and Petrology (1991) 43:293–314.[Web of Science]
Kushiro I. Compositions of magmas formed by partial zone melting of the earth's upper mantle. Journal of Geophysical Research (1968) 73:619–634.
Le Bas M. J. Carbonatite–Nephelinite Volcanism: an African Case History (1977) New York: Wiley–Interscience.
Lohmann L. Ein Beitrag zur Petrographie melilith-fuhrender Olivin-nephelinite aus dem Gebiet Fritzlar–Naumburg (Nordhessen). Beiträge zur Mineralogie und Petrographie (1964) 9:533–584.
Macdonald R., Kjarsgaard B. A., Skilling I. P., Davies G. R., Hamilton D. L., Black S. Liquid immiscibility between trachyte and carbonate in ash flow tuffs from Kenya. Contributions to Mineralogy and Petrology (1993) 114:276–287.[Web of Science]
Meen J. K., Ayers J. C., Fregeau E. J. A model of mantle metasomatism by carbonated alkaline melts: trace element and isotopic compositions of mantle source regions of carbonatite and other continental igneous rocks. In: Carbonatites: Genesis and Evolution—Bell K., ed. (1989) London: Unwin Hyman. 464–499.
Olafsson M., Eggler D. H. Phase relations of amphibole–carbonate, and phlogopite–carbonate peridotite: petrologic constraints on the asthenosphere. Earth and Planetary Science Letters (1983) 64:305–315.[Web of Science]
Peterson T. D. Peralkaline nephelinites I. Comparative petrology of Shombole and Oldoinyo Lengai, East Africa. Contributions to Mineralogy and Petrology (1989) 101:458–478.[Web of Science]
Powell J. L., Hurley P. M., Fairburn H. W. The strontium isotopic composition and origin of carbonatites. In: Carbonatites—Tuttle O. F., Gittins J., eds. (1966) New York: John Wiley. 365–378.
Rudnick R. L., McDonough W. F., Chappell B. W. Carbonatite metasomatism in the northern Tanzanian mantle: petrographic and geochemical characteristics. Earth and Planetary Science Letters (1993) 114:463–475.[Web of Science]
Schneider M. E., Eggler D. H. Fluids in equilibrium with peridotite minerals: implications for mantle metasomatism. Geochimica et Cosmochimica Acta (1986) 50:711–724.[Web of Science]
Shand S. J. The present status of Daly's hypothesis of the alkaline rocks. American Journal of Science (1945) 23-A:495–507.
Simonetti A., Bell K. 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 (1995) 36:141–153.[Web of Science]
Spencer A. B. Alkali igneous rocks of the Balcones Province, Texas. Journal of Petrology (1969) 10:272–307.
Wallace M. E., Green D. H. An experimental determination of primary carbonatite magma composition. Nature (1988) 335:343–348.
Watkinson D. H., Wyllie P. J. Experimental study of the join NaAlSiO4–CaCO3–H2O and the genesis of alkalic rock–carbonatite complexes. Journal of Petrology (1971) 12:357–378.
White B. S., Wyllie P. J. Solidus reactions in synthetic lherzolite–H2O–CO2 from 20–30 kbar, with applications to melting and metasomatism. Journal of Volcanology and Geothermal Research (1992) 50:117–130.[Web of Science]
Woolley A. R., Kempe D. R. C. Carbonatites: nomenclature, average chemical compositions and element distribution. In: Carbonatites—Bell K., ed. (1989) London: Unwin Hyman. 1–14.
Wyllie P. J. The origin of kimberlite. Journal of Geophysical Research (1980) 85:6902–6910.
Wyllie P. J. Discussion of recent papers on carbonated peridotite. bearing on mantle metasomatism and magmatism: response. Earth and Planetary Science Letters (1987) 82:401–402.[Web of Science]
Wyllie P. J., Huang W. L. Influence of mantle CO2 in the generation of carbonatites and kimberlites. Nature (1975a) 257:297–299.
Wyllie P. J., Huang W. L. Peridotite, kimberlite and carbonatite explained in the system CaO–MgO–SiO2–CO2. Geology (1975b) 3:621–624.
Wyllie P. J., Huang W. L. Carbonation and melting reactions in the system CaO–MgO–SiO2–CO2 at mantle pressures with geophysical and petrological applications. Contributions to Mineralogy and Petrology (1976) 54:79–107.[Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||