Journal of Petrology Advance Access originally published online on March 24, 2008
Journal of Petrology 2008 49(5):885-909; doi:10.1093/petrology/egn010
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Petrogenesis of Ultramafic Rocks from the Ultrahigh-pressure Metamorphic Kimi Complex in Eastern Rhodope (NE Greece)
1National Technical University of Athens, Department of Mining and Metallurgical Engineering, Section of Geological Sciences, Heroon Polytechniou 9, 15780, Athens, Greece
2California Institute of Technology, Division of Geological and Planetary Sciences, Pasadena, CA 91125, USA
RECEIVED JANUARY 25, 2007; ACCEPTED FEBRUARY 13, 2008
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSAMPLE CHARACTERIZATIONDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Widespread bodies of garnet–spinel metaperidotites with pyroxenitic layers occur in the ultrahigh-pressure metamorphic Kimi Complex. In this study we address the origin of such peridotite–pyroxenite associations in the context of polybaric melting regimes. We conduct a detailed geochemical investigation of major and trace element relations and compare them with a range of major element modelling scenarios. With increasing bulk-rock MgO content, the garnet–spinel metaperidotites exhibit decreasing CaO, Al2O3, TiO2, and Na2O along with increasing Ni and a gradually increasing Zr/Zr* anomaly, consistent with an origin as residues after variable degrees of melt extraction. The major element modelling further suggests a polybaric adiabatic decompression melting regime beginning at high to ultrahigh pressure, with an intermediate character between pure batch and fractional melting and a mean extent of melting of 9–11%. The pyroxenites exhibit major element compositions that cannot be reproduced by experimental or calculated melts of peridotite. Moreover, the Kimi pyroxenites have highly variable Ni and Sc contents and a wide range of Mg-number (0· 76–0· 89), inconsistent with an origin as frozen melts or the products of melt–peridotite interaction. However, both the major element systematics and the observed rare earth element patterns, with both convex and concave shapes, can be explained by an origin as clinopyroxene-rich, high-pressure cumulates involving garnet and/or Cr-spinel.
KEY WORDS: peridotite; pyroxenite; partial melting; UHP metamorphism; cumulate
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSAMPLE CHARACTERIZATIONDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The current consensus among observational, experimental, and theoretical geochemists and petrologists is that upper mantle processes are dominated by adiabatic, polybaric, decompression-induced partial melting (McKenzie, 1984
; Klein & Langmuir, 1987; Asimow, 2002), leading to the formation of new crust from the aggregated liquids (McKenzie & Bickle, 1988; Kinzler & Grove, 1992; Langmuir et al., 1992; Asimow et al., 2001). Data regarding upper mantle mineralogy, element partitioning, partial melting processes and the generation of primary basaltic magmas are interpreted within this broad paradigm. However, a range of important variations on the theme of polybaric melting exist, including equilibrium vs fractional melt production (Niu, 1997; Johnson et al., 1990) and melt migration (Spiegelman & Kenyon, 1992; Iwamori, 1993), the pressure range of melt production (Salters & Hart, 1989; Hellebrand et al., 2002; Presnall et al., 2002; Weyer et al., 2003; Brunelli et al., 2005), the importance of deep (i.e. subcrustal) differentiation (e.g. Grove et al., 1992), and the significance of lithological heterogeneity in the mantle (Hirschmann & Stolper, 1996; Sobolev et al., 2005).
The origin of peridotites and related rocks and variations in their chemical composition reflect processes such as partial melting and melt–rock interaction; the petrogenetic history of such rocks is, therefore, invaluable for understanding upper mantle processes. Pyroxenites are commonly associated with peridotites; they have been the focus of several studies leading to various hypotheses for their origin and geochemical significance, including frozen melts (e.g. Pearson et al., 1993), melt–rock interaction (e.g. Kelemen et al., 1998; Litasov et al., 2000; Hermann et al., 2006) and/or cumulate processes (e.g. Hirschmann & Stolper, 1996; Kopylova et al., 1999; Xu, 2002; Dantas et al., 2007).
Here we study the geochemistry of garnet–spinel metaperidotites and associated spinel–garnet clinopyroxenites tectonically emplaced into the crustal rocks of the ultrahigh-pressure (UHP) metamorphic Kimi Complex in eastern Rhodope, Greece. We compare whole-rock compositions of the peridotites with residues generated in polybaric melting models (Asimow, 1999) to define the detailed melting history of the suite. We infer constraints about the nature of the mantle source and melt extraction process that affected the metaperidotites and the processes controlling the formation of the spinel–garnet clinopyroxenites. In particular, we attempt to give answers to the following questions, using geochemical evidence from whole-rock major and trace element concentrations.
- Are the peridotites formed as residues from a batch, fractional or mixed polybaric melting regime? It should be noted that we use batch melting here as a proxy for physical settings in which, despite melt mobility, melts remain in equilibrium with residues throughout the melting column; the equivalence of this case to batch melting in terms of residue composition has been demonstrated by several workers (Ribe, 1985; Spiegelman & Elliott, 1993; Asimow & Stolper, 1999).
- Which processes are responsible for the formation of the clinopyroxenites?
Geological framework of the UHP Kimi Complex: general
The Rhodope high-pressure (HP) province in the easternmost part of the Hellenic Orogen is an Alpine synmetamorphic thrust and nappe complex (Burg et al., 1996; Ricou et al., 1998; Liati & Gebauer, 1999; Mposkos & Krohe, 2000; Krohe & Mposkos, 2002) that incorporates several tectonic slivers of UHP and HP metamorphic rocks (Mposkos & Krohe, 2000, 2006; Mposkos & Kostopoulos, 2001; Mposkos et al., 2004). The Rhodope HP province is subdivided into several tectonometamorphic units that are bounded by thrust and normal faults. In eastern Rhodope, the Kimi Complex, representing the structurally uppermost metamorphic unit, records an alpine UHP metamorphism followed by an HP granulite- to upper amphibolite-facies event. It was exhumed between 65 and 48 Ma (Mposkos & Wawrzenitz, 1995; Liati et al., 2002; Mposkos & Krohe, 2006).
The UHP Kimi Complex in eastern Rhodope
The Kimi Complex (Fig. 1) is a tectonic mixture of crustal and mantle-derived rocks. The crustal rocks comprise amphibolitized eclogites, orthogneisses, marbles and migmatitic pelitic gneisses. The presence of diamond inclusions in garnet and needle-like rutile exsolution in Na-bearing garnet in migmatitic pelitic gneisses documents UHP metamorphism with maximum P–T conditions of >4· 5 GPa at 
1000°C (Mposkos et al., 2004; Perraki et al., 2004, 2006; Mposkos & Krohe, 2006). The mantle rocks consist of garnet–spinel metaperidotites, spinel–garnet pyroxenites and olivine pyroxenites. The P–T conditions of the metaperidotites are >2· 3 GPa for an assumed temperature of 1000°C calculated from the association of garnet with Cr-Spinel (Cr-number = 0· 27). Independently, the reintegration and incorporation of the Ca-Ts component in Cpx and Mg-Ts in Opx requires a pressure of >2· 5–3· 0 GPa at a temperature of 1200°C. The maximum P–T conditions for the pyroxenites are calculated from the incorporation of the Ca-Ts and enstatite component into the clinopyroxene at about 1000°C and >1· 8 GPa. A first stage of decreasing pressure and temperature, recorded in the metaperidotites, shows equilibration at about 1· 9–2· 0 GPa and 1000–1100°C. Subsequently, a long-lived equilibration stage, with substantial cooling at depth, is documented both in the metaperidotites and pyroxenites. In the metaperidotites, application of the Al-in-Opx barometer, assuming that garnet was in equilibrium with that orthopyroxene, yields a pressure of 1· 06–1· 19 GPa at 700°C and 1· 55–1· 68 GPa at 800°C. The transition from the garnet peridotite to the spinel peridotite stability field is constrained at a pressure of <1· 4 GPa at a temperature of 700°C. During this stage, hydration reactions of the form Grt + Opx + H2O 
Ol + Hbl and Opx + Cpx + Spl + H2O Ol + Hbl produce hornblende at a pressure of 1· 2–1· 3 GPa and temperatures of 650–750°C. In the pyroxenites, application of garnet–clinopyroxene geothermometry yields equilibration temperatures of 750–765°C for a pressure of 1· 5 GPa, whereas the minimum pressure is constrained at 1· 1 GPa at 750°C (see Mposkos & Krohe, 2006, fig. 9, reaction curve 1).
|
A garnet–diopside-whole rock Sm–Nd age for a spinel–garnet pyroxenite of c. 119 ± 3· 5 Ma (Wawrzenitz & Mposkos, 1997
) probably reflects the age of a metamorphic event in the lithospheric mantle (at 1· 5 GPa, 750°C), assuming that equilibration in the Sm–Nd system between garnet and clinopyroxene stopped at the same temperature (750°C) as the Grt–Cpx Fe–Mg exchange reaction (Mezger et al., 1992). Muscovite pegmatites with a crystallization age of c. 65–63 Ma intrude all the lithological successions (Mposkos & Wawrzenitz, 1995; Liati et al., 2002). |
SAMPLE CHARACTERIZATION |
|---|
|
TOPABSTRACTINTRODUCTIONSAMPLE CHARACTERIZATIONDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Petrography and mineral chemistry
Garnet–spinel metaperidotites
Between Organi and Kimi villages (Fig. 1) several bodies of garnet–spinel metaperidotite occur. The overall mineral assemblage of the metaperidotites is Ol–Cpx–Opx–Spl–Grt–Hbl. Representative mineral assemblages with their mineral modes are listed in Table 1. The mineral abbreviations used throughout this paper are in accordance with those proposed by the International Mineralogical Association (IMA; Martin, 1998
). The mineral chemical data referred to herein are drawn from Mposkos & Krohe (2006).
|
Several pre-, syn- and post-deformation generations of Spl, Opx and Cpx occur. Macroscopically they are characterized by a distinct foliation and lineation defined by flattening (stretching) of large, older Cpx and Opx grains and elongated recrystallized smaller Cpx and Opx grain aggregates. The primary textural features are preserved in all of the studied samples; even in strongly serpentinized samples pseudomorphs of serpentine after olivine occur (Fig. 2). Post-deformational mineral assemblages document substantial cooling within the stability field of spinel peridotite but still at high pressures. This is indicated by the garnet exsolution in deformed clinopyroxene in the associated clinopyroxenites (Mposkos, 2002
; Mposkos & Krohe, 2006).
|
Clino- and orthopyroxenes. Large Cpx-1 and Opx-1 grains contain exsolution lamellae of Spl as well as of Opx and Cpx, respectively, that formed during cooling (Fig. 2a and b). A second Opx-2 and Cpx-2 generation is exsolution-free and probably formed by static recrystallization of Px-1. The Al2O3 content of both exsolution-bearing and exsolution-free Opx and Cpx is in the range of 1· 6–2· 6 wt % and 0· 11–2· 1 wt % respectively.
Spinel. Three successive spinel generations are present. (1) Large spinel crystals (Spl-1) with Cr/(Cr + Al) values (Cr-number) ranging from 0· 27 to 0· 48 in their cores formed near the maximum P–T conditions. (2) During decompression and cooling, light brown spinel (Spl-2; Cr-number 0· 12–0· 19) grew around Spl-1 (Fig. 2c) in the matrix interstitially between exsolution-free Px-2 and as exsolution lamellae in Cpx-1 and Opx-1. (3) Spl-3 (Cr-number 0· 03–0· 05) formed symplectites with diopside and enstatite or enstatite and hornblende from the reaction between garnet and olivine (Fig. 2d).
Garnet. Rare garnet grains (Grs13–17Alm22–24Prp58–62Sps1–2Uvr1· 6–2· 3) commonly with resorbed edges, occur only as small inclusions in olivine, spinel (Spl-2), orthopyroxene (Opx-2) and hornblende (Fig. 2e).
Hornblende. Chromian hornblende (Si = 6· 5–6· 9 atoms per formula unit, Cr2O3 0· 8–1· 1 wt %, Al2O3 8–11 wt %) coexists with olivine (Fig. 2f) or with symplectitic Spl-3, indicating hydration of the peridotites.
Spinel–garnet clinopyroxenites and olivine clinopyroxenites
Clinopyroxenites occur as layers (millimetres to centimetres scale, locally >1 m wide) within the metaperidotites. They commonly show sharp contacts with the host peridotites. Analogous sequences of layered pyroxenites alternating with peridotites occur worldwide (i.e. Pearson et al., 1993; Hirschmann & Stolper, 1996; Kelemen et al., 2000; Schiano et al., 2000; Berly et al., 2006; Montanini et al., 2006). Based on the modal mineral assemblages, the Kimi pyroxenites have been subdivided into spinel–garnet clinopyroxenites (Cpx + Grt + Spl + Hbl ± Ol; type-I) and olivine clinopyroxenites (Cpx + Ol ± Hbl ± Spl; type-II). The mineral assemblages along with the mineral proportions are listed in Table 1. Petrographic features, such as garnet inclusions in olivine and clinopyroxene (Fig. 3a), suggest that the clinopyroxenites may represent HP mantle rocks formed above the stability field of plagioclase.
|
Garnet. Garnet forms elongated polycrystalline aggregates characteristically showing homogeneous grain-size distribution, straight grain boundaries and 120° triple junctions, indicating static annealing that followed crystal-plastic deformation (Fig. 3b). Garnet composition ranges between Grs21–27Prp45–51Alm24–31Sps0· 5–1· 2 (in olivine-free layers) and Grs16–18Prp51–52Alm28–31Sps1–1· 5 (in olivine-bearing layers).
Clinopyroxene. Two clinopyroxene generations can be distinguished. Cpx-1 shows compositional zoning with Al2O3 content decreasing from 3· 5 wt % in the cores to 2· 5 wt % at the rims. In the olivine clinopyroxenites, the Al2O3 content of the clinopyroxene ranges from 2· 7 wt % in the cores to 1· 7 wt % at the rims. During cooling, garnet was exsolved in the Cpx-1 cores (Fig. 3c and d). Cpx-2 occurs in polycrystalline ribbons stretched into layers. Such Cpx ribbons are homogeneous in grain size; grains are strain free and show 120° triple junctions, indicating that static annealing and grain growth followed deformation. The exsolution-free rims of Cpx-1 crystals and exsolution-free Cpx-2 in the ribbons are similar in composition. Moreover, garnet exsolved from Cpx-1 and matrix garnet (Grt-2) are similar in composition.
Spinel. Green spinel occurs either as elongated porphyroblasts up to 1 mm in length or as smaller interstitial grains between Grt-2 and Cpx-2 grains. In olivine-bearing garnet clinopyroxenites, spinel forms symplectites (Spl-2) with enstatite and diopside as reaction products from olivine and garnet.
Hornblende. Tschermakitic to pargasitic hornblende forms elongated and oriented grains up to 800 µm in length or smaller grains in textural equilibrium with Cpx-2.
Whole-rock major and trace element analysis
Methods
Major and trace element compositions of 29 samples, including garnet–spinel metaperidotites (14) and spinel–garnet clinopyroxenites (15), were determined by inductively coupled plasma-emission spectroscopy (ICP-ES) and X-ray fluorescence spectroscopy (XRF). For major and trace element analysis, structural water was removed from sample powders by heating at 1000°C for 1 h. Loss on ignition (LOI) was determined from the total weight change. Major and trace element analyses were performed on solutions after LiBO2 fusion and nitric acid digestion of rock powder for ICP-ES analysis and on prepared beads after mixing with di-lithium tetraborate and fusion for XRF analysis. Rare earth element (REE) analyses were determined by inductively coupled plasma-mass spectroscopy (ICP-MS) after LiBO2 fusion and nitric acid digestion. The detection limits for the REE (in ppm) are <0· 5 for La and Ce, <0· 02 for Pr, <0· 4 for Nd, <0· 1 for Sm, <0· 05 for Eu, Gd, Dy, Ho, Er, Tm and Yb, and <0· 01 for Tb and Lu. The whole-rock analyses were carried out at Acme Analytical laboratories in Canada. The chemical compositions of the analysed samples are presented in Table 2.
|
Major elements and compatible trace elements
Garnet–spinel metaperidotites. The garnet–spinel metaperidotites have SiO2 contents in the range of 43· 5–47· 9 wt %, Al2O3 1· 4–4· 6 wt %, CaO 0· 1–3· 5 wt % and TiO2 0· 02–0· 2 wt % (Table 2). FeO* content varies from 7· 1 to 11· 3 wt % and Na2O content from 0· 01 to 0· 2 wt %; these two oxides show poor correlation with the other oxides. Cr, Ni and Co abundances are in the range of 1644–3530 ppm, 1283–3000 ppm and 85–137· 5 ppm, respectively, whereas V and Sc contents are 25–132 ppm and 2–28 ppm, respectively.
Variations of selected major and trace elements vs MgO are shown in Fig. 4. SiO2, TiO2, Al2O3, FeO*, CaO, Na2O and V are negatively correlated with MgO whereas Ni is positively correlated. Cr and Co show no correlation with MgO. Sample I33 represents the most fertile composition (Table 2) and displays slightly lower MgO, Al2O3 and CaO contents compared with the primitive mantle composition (Sun & McDonough, 1989), but it is strongly depleted in Na2O (0· 05 wt %) relative to estimates of primitive mantle (0· 35 wt %) or depleted MORB source mantle (e.g. 0· 28 wt %, Workman & Hart, 2005).
|
Spinel–garnet clinopyroxenites and olivine clinopyroxenites. The clinopyroxenites have SiO2 contents in the range of 40· 6–50· 7 wt %, Al2O3 4· 1–18· 2 wt %, FeO* 4· 3–11· 3 wt %, CaO 6· 3–19· 8 wt % and TiO2 0· 1–0· 5 wt %. Based on the Al2O3 contents, two types of clinopyroxenites can be distinguished: type-I (spinel–garnet clinopyroxenites), with Al2O3 ranging from 9· 5 to 18· 2 wt %, and type-II (olivine clinopyroxenites), with Al2O3 contents of 4· 1–4· 8 wt %. The discrimination of the two clinopyroxenite types based on the Al2O3 content agrees with the petrographic observations based on the olivine (serpentine) abundance.
Major and trace element vs MgO variations are shown in Fig. 4. CaO, Al2O3 and Na2O more or less exhibit negative correlations with MgO whereas Cr and Ni are positively correlated. SiO2 and FeO* remain nearly constant at variable MgO content. TiO2 displays a weak but still negative correlation with MgO content. Cr, Ni and Co abundances are in the range of 407· 1–1770· 6 ppm, 87· 3–976· 8 ppm and 35· 8–75· 1 ppm respectively, whereas V and Sc contents range from 51· 5 to 264· 2 ppm and from 16· 6 to 59· 9 ppm respectively.
Incompatible trace elements
Garnet–spinel metaperidotites. Primitive mantle-normalized trace element patterns (Fig. 5a) are characterized by positive Ba anomalies, positive to negative Sr anomalies, and relatively flat high field strength elements (HFSE), with the exception of Zr, which shows negative anomalies with Zr/Zr* [(Zr/Zr* = ZrN/(NdN· SmN)0· 5] roughly inversely correlated with MgO content.
|
The REE patterns normalized to chondrite (Fig. 5b) are characterized by nearly horizontal light REE (LREE), slightly negative to positive Eu anomalies (Eu/Eu* = 0· 87–1· 71) and relatively flat heavy REE (HREE) patterns. They fall within the range of 0· 5–3 times chondrite for both LREE and HREE.
Spinel–garnet clinopyroxenites and olivine clinopyroxenites. Primitive mantle-normalized trace element patterns (Fig. 5c) are characterized by variable enrichment of incompatible elements (e.g. Ba) and negative to positive Sr anomalies. They also display relatively flat HREE patterns with enrichment from Dy to Lu. The relative abundances of HFSE and HREE are higher than those of the metaperidotites.
The chondrite-normalized REE patterns from the type-I and type-II pyroxenites are illustrated in Fig. 5d. Type-I clinopyroxenites are characterized by relative HREE enrichments, depletion of LREE [(La/Yb)N <1· 0] and slight negative to positive Eu anomalies (Eu/Eu* = 0· 93–1· 14). The pattern within the LREE is often concave-upwards, yielding a spoon-shaped overall REE pattern. Type-II clinopyroxenites are characterized by smaller LREE depletion than the Type-I samples, flat HREE and a slight negative Eu anomaly (Eu/Eu* = 0· 73–0· 89). The REE pattern of the sample 5K29A1 shows a convex-upward segment for the middle REE (MREE)–HREE pattern, similar to that reported for spinel pyroxenites from Qilin Island, South China (Xu et al., 2002).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSAMPLE CHARACTERIZATIONDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Alteration and element mobilization
Numerous studies have revealed that mantle rocks are often metasomatized through chemical interaction with fluids and/or melts (Downes, 2001
; Szabo et al., 2004). Low- to high-temperature hydrous alteration may have induced modification to the mostly incompatible elements such as Cs, Ba, Pb, Rb and Sr. Volatile components are reported as loss on ignition (LOI) in Table 2 and are considered to represent the serpentinite component. The studied samples are affected by variable serpentinization and by the formation of small modal amounts of retrograde amphibole and chlorite (see Table 1). The serpentine modal content is in the range of 28–60%, hornblende 4–10% and chlorite 2–10%. A plot of molar Si/(Mg + Fe) vs LOI can be used to test for the effects of major element mobility with increasing degree of serpentinization (Fig. 6a). In peridotites, the molar Si/(Mg + Fe) ratio ranges from 0· 57 to 0· 74, without a clearly visible trend with LOI variation, indicating that hydration processes occurred without major changes of Si/(Mg + Fe) ratio. Pyroxenites are more variable in major element composition and overall show low LOI values, presenting no clear evidence of major element mobility associated with alteration. Moreover, by plotting MgO vs MgO/CaO (in wt %), we are able to define the degree of mobilization of Ca. In the pyroxenites, MgO and MgO/CaO are well correlated; in the peridotites, most of the samples define a positive correlation with the exception of two samples that have MgO/CaO ratios higher than 100 (Fig. 6b). CaO displays only a weak negative correlation with LOI (not shown). The data suggest that minor Ca was removed from the peridotites during serpentinization.
|
The primitive mantle-normalized trace element diagrams are also good indicators of the element mobilization. As noted above, the mostly incompatible elements (e.g. Cs, Ba, Pb, Rb and Sr) and the LREE are more mobile than the others. In Fig. 5, we note the variable Ba and Ce contents together with nearly constant Rb and La. Also, given the evidence for Ca removal we suspect that the range of Sr anomalies from negative to positive may be an alteration effect. As a result, the behaviour of the most incompatible elements is not steady state, probably because of the addition or removal of fluids or melts. For the above reasons, this study does not focus on the highly incompatible elements, but instead we evaluate the petrogenetic processes by interpreting the major elements and the moderately incompatible to compatible trace elements.
Do the peridotites represent residues after variable degrees of partial melting?
Partial melting is an important process for generation of heterogeneity among mantle rocks; the range of chemical variation is most simply interpreted in terms of residues after variable degrees of melt extraction (Dick et al., 1984; Frey et al., 1985; Niu, 1997; Burnham et al., 1998; Paulick et al., 1996).
The studied peridotites exhibit variations in their major element composition (Table 2). Observed variations of major elements vs MgO, qualitatively similar to those in the Kimi suite, are widely ascribed to extraction of partial melts from a fertile protolith subjected to different degrees of partial melting, where the peridotites represent the residual material (Frey et al., 1985). The gradual increase of the Zr/Zr* anomaly observed in the metaperidotite samples is consistent with variable degrees of melt extraction (Yu et al., 2006). Although Zr/Zr* measured in clinopyroxene is thought to be a good indicator of the extent of melting, the whole-rock values reported here are more probably a function of variations in the modal abundance of Cpx, and therefore not such a strong independent indicator of depletion.
Geochemical modelling
To provide constraints on the melting regime of the Kimi peridotites we try to simulate the major and trace element variations. We test two different end-member melting models, batch and fractional melting. A candidate fertile peridotite source could be sample I33; however, as noted above, its TiO2 and Na2O contents are too low for a fertile source. Therefore, a hypothetical source composition was adjusted to fit the observed data. This procedure allows us to examine what process might explain the variation in peridotite composition within the suite, but it prevents us from directly evaluating how the original source composition might have been derived from typical primitive or depleted mantle reservoirs. Major and selected trace element concentrations of this model source peridotite are given in Table 3.
|
Batch and fractional melting were modelled for the hypothetical fertile peridotite BHS (Table 3) using the program Adiabat_1ph, a simple text-menu driver for subroutine versions of the algorithms MELTS, pMELTS and pHMELTS with added capability for tracking the behaviour of trace elements (Ghiorso & Sack, 1995
; Ghiorso et al., 2002; Asimow et al., 2004; Smith & Asimow, 2005). The MELTS algorithm is a computational thermodynamic package that incorporates different solution models of minerals and liquids to calculate phase equilibria of various systems. MELTS is suitable for variable bulk compositions and pressures <3· 0 GPa, whereas pMELTS is optimized for peridotite compositions in the pressure range of 1· 0–4· 0 GPa. The processes of melting and crystallization in a batch, fractional or continuous manner can be simulated more easily than using the original MELTS interface. We used pMELTS to model batch and fractional melting and to calculate the compositions of the resultant liquids and residues. The two melting regimes were modelled under isentropic conditions and at the same potential temperatures, following the method described by Asimow et al. (2001). In both cases, melting continues up to a pressure equivalent to the base of the crust to produce aggregate liquids that are subsequently used as the possible primary magma to explain the cumulate character of the clinopyroxenites. It should be noted that these are the first published calculations to use pMELTS to model complex metamorphic systems such as the Kimi ultramafic rocks. Asimow (1999) used the original MELTS calibration. Batch melting. In Fig. 7a–j selected major and trace elements are plotted vs MgO along with pMELTS calculations of residues for polybaric batch melting under isentropic conditions. We used two mantle states in our calculations, a ‘hot’ and a ‘cold’ case, but only the ‘hot’ case is plotted (‘hot’ mantle at Po = 3· 5 GPa, Tp = 1456· 1°C; ‘cold’ mantle at Po = 1· 5 GPa, Tp = 1341· 4°C).
|
As mentioned above, despite some scatter in SiO2 and FeO*, all the major oxides in the peridotite bulk compositions exhibit a general negative trend with MgO content. The calculated residues reproduce the overall negative trend of SiO2, CaO, Al2O3 and TiO2. The Na2O model curves provide a reasonable fit to the Na2O data. However, the Na2O data are too scattered to assign much meaning to the model fitted to this oxide. The FeO* model curve does not provide a good fit to the observed data.
The trace elements Ni and Co in the Kimi peridotites display a roughly positive correlation with MgO content, whereas V is clearly negatively correlated, indicating the increased role of olivine in the modal assemblage. Cr behaviour is more complicated as it increases and then apparently decreases with increasing MgO content. The batch residues match the observed trends well, except they do not reproduce the possible reversal in the Cr trend at high MgO. The selected partition coefficients (Table 3) for the trace element modelling are the standard set compiled in the source coding of Adiabat_1ph [references given by Smith & Asimow (2005)]. The exception is V, where we have used different values to fit the peridotite data.
The degree of melting is negatively correlated with pressure; it ranges from zero to 26· 8% as pressure decreases from 3· 5 to 1 GPa, with a mean value of 8· 2%.
Fractional melting. The calculated fractional melting residues (Fig. 7a–j) reproduce the major element compositions for CaO and Al2O3 in a similar manner to those of polybaric batch melting, but fail to match the negative correlation in SiO2. Without re-equilibration with liquids at low pressure, which favours opx dissolution and olivine precipitation (Kelemen et al., 1997), residues remain roughly constant or even increase in SiO2 during polybaric fractional melting. The model fractional residues lie at lower values of TiO2 and Na2O compared with the batch model, further from the average trend through the data. The calculated residues reproduce the general negative trend of FeO but are unable to reach the lowest FeO values. The Ni, Co and Cr fractional melting trends are similar to those of polybaric batch melting. Only the V trend differs noticeably at high MgO contents and the difference is smaller than the scatter in the Kimi data. The calculated MgO contents fail to reach the most refractory peridotites even at the ‘hot’ mantle state (Tp = 1456· 1°C), if melting stops at 1 GPa. The degree of melting at 1 GPa is somewhat lower than in the batch melting case, 23· 3%. The mean extent of melting FB [as defined by Plank et al. (1995); see also Asimow et al. (2001) and Asimow & Langmuir (2003)] is 11· 0% and the mean pressure of melt extraction PB is 1· 87 GPa, for the ‘hot’ mantle case. The ‘cold’ mantle case yields, surprisingly, a higher mean extent of melting FB (12· 3%) at lower mean pressure of melt extraction PB (0· 93 GPa). Although the maximum extent of melting in the cold model is lower, the mean value ends up higher because the melting column is so much shorter; for a passive-flow type integrated melting region at high potential temperature in this model, a very large volume of mantle melts to low degree only before exiting the ‘wings’ of the melting regime, which pulls down the value of FB. From the perspective of residues, however, the cold fractional case yields less depleted residues than the hot fractional case, as expected.
Batch melting, fractional melting, or a combination? The two melting models, batch and fractional melting, yield more or less similar results in most simple major–trace element variation diagrams as in Fig. 7. Both melting processes can broadly explain the natural peridotite data from the Kimi Complex. However, to elucidate which process is predominant, we use the CaO/Al2O3 and TiO2/Al2O3 ratios, which are sensitive to melting process (Fig. 8a and b). These ratios can also reflect source heterogeneities; however, they were introduced by Asimow (1999) in analysis of abyssal peridotites to distinguish reactive melt migration from simple olivine addition as proposed by Niu (1997).
|
The calculated CaO/Al2O3 ratios (Fig. 8a) remain more or less constant at a value of
0· 8, for both batch and fractional melting regimes. However, the peridotite data do not follow the calculated CaO/Al2O3 trend and show considerable scatter in CaO/Al2O3. The model TiO2/Al2O3 ratios are higher for batch melting than for fractional melting (Fig. 8b). As outlined by Asimow (1999), for abyssal peridotites neither batch nor fractional melting alone can explain the variations of these ratios. Asimow suggested a combination of fractional melting followed by batch melting or vice versa, as an approximation to the range of behaviours expected in two-porosity type melting regimes in which both reactive flow and fractional melt extraction are occurring together, to explain the bulk composition of abyssal peridotites. In addition, variation in whole-rock compositional data (from abyssal peridotites or from the Kimi suite), if they are not a consequence of analytical uncertainties, could be also the result of source heterogeneity or of minor trapping of migrating liquids.
In the case of the Kimi peridotites, a polybaric melting model starting at high to ultrahigh pressures and continuing adiabatically at lower pressures is preferred to explain the peridotite compositions. We advocate a mixed melting regime, similar to that of Asimow (1999), with a potential role for the fractional melting component, to explain the range of TiO2/Al2O3 ratios and the MgO content of the most refractory peridotites.
The incompatible trace element and LREE variations within the peridotite suite are substantially smaller than would be expected in a suite of residues of fractional melting. On the other hand, the HREE variations (about a factor of nine from the least to most depleted sample) are consistent with 25% melt extraction, especially if a significant fraction of the melting is in the garnet peridotite field. We conclude that a substantial component of batch melting and a significant history of late modification of incompatible trace elements prevents us from observing the large variation expected for LREE in residue suites.
Processes responsible for the formation of pyroxenites
Frozen melts
To test the possibility that the pyroxenites represent frozen melts derived from partial melting of host peridotite, experimental and calculated melts are compared with the natural data. First we compare the studied pyroxenites with experimental melts produced by partial melting of dry peridotites under various pressure conditions and different experimental datasets (Hirose & Kushiro, 1993; Falloon et al., 1999, 2001; Bulatov et al., 2002; Wasylenki et al., 2003). Subsequently, we consider the calculated melts produced by partial melting of peridotite (see previous section) and whether or not there is a possible genetic relationship between the host peridotites and the associated pyroxenites.
The Kimi pyroxenites exhibit higher MgO and lower TiO2, SiO2 and Na2O contents than any experimental melt of dry peridotite (Fig. 9a–f). As MgO increases in Hirose & Kushiro's (1993) experimental melts with increasing degree of melting, the SiO2 and FeO* remain roughly constant. The CaO content of the experimental melts increases during lherzolite melting and then decreases after cpx exhaustion and Al2O3, TiO2 and Na2O contents uniformly decrease (Fig. 9). In the Kimi pyroxenites, both CaO and Al2O3 contents uniformly decrease with increasing MgO content, whereas FeO* contents, despite some scatter, show a general negative trend extending from the experimental trend towards higher MgO contents. For the experimental melts of Bulatov et al. (2002) and Wasylenki et al. (2003), the CaO trend is parallel to that in the studied pyroxenites but displaced toward lower MgO contents. Moreover, the MgO content is positively correlated with pressure and degree of melting, indicating that if the pyroxenites represent liquid compositions they should have been formed under very high pressure (>2 GPa) and high degrees of melting (>>30%). However, even at such pressures and extensive degrees of melting, the other major elements fail to fit the natural data.
|
We have advocated a mixed melting model to explain the Kimi peridotites as residues after liquid extraction during adiabatic upwelling of the mantle. The range of liquids generated by such melting models does not fit the major element compositions of the Kimi pyroxenites (Fig. 10a–f). In particular, the calculated melts display lower SiO2, Al2O3 and CaO at a given MgO content (Fig. 10). The TiO2, FeO* and Na2O contents of the calculated melts display rather linear positive trends vs MgO content (Fig. 10). In contrast, the Kimi pyroxenites display very weak correlations between TiO2, FeO* and Na2O vs MgO.
|
In summary, the pyroxenite data fail to fit the melt compositions from isobaric experimental batch melting or from pMELTS calculations of polybaric batch, fractional or mixed melting under upper mantle pressure conditions. Therefore we conclude that such mantle pyroxenites, in agreement with previous work on other areas (e.g. Takazawa et al., 1999
, and references therein), do not represent frozen melts produced by peridotite partial melting.
Melt–rock interaction
As an alternative hypothesis for the origin of the Kimi pyroxenites, it is possible that they may represent the products of melt–rock interaction between a silicic melt and the host peridotite. In this case, the concentration of compatible elements in the melt is buffered by the peridotite for all reasonable melt–rock ratios in an increasing to near-constant melt-mass system. As a result, the melt–rock reaction products should have high Ni contents and uniform Mg-number at around 0· 85–0· 90, assuming an olivine–melt Fe–Mg KD value of 0· 3 (Roeder & Emslie, 1970; Ulmer, 1989).
Both types of pyroxenites (type-I and type-II) have highly variable Ni and Sc contents, lower than those of the coexisting peridotites, and a wide range of Mg-number (0· 76–0· 89). The compatible elements Ni and Sc are positively correlated with Mg-number. The non-uniformity of Ni and Sc contents and the wide range of Mg-number appear to preclude buffering of a silicic melt by the host peridotite. As mentioned above, if the clinopyroxenites were to represent liquids frozen after such a reaction process, then the compatible element content should be equilibrated with the peridotites; however, this is not the case for the Kimi pyroxenites.
High-pressure cumulates
Another important process to evaluate to explain the origin of the Kimi pyroxenites is crystal segregation at high pressures (e.g. Irving, 1974; Hirschmann & Stolper 1996; Xu, 2002). We test this hypothesis through major and trace element relationships and geochemical modelling.
Type-I (Al-rich) pyroxenites are enriched in the incompatible elements Ba, Sr, Na and K, with variable LREE and HREE enrichment. Two REE patterns are observed, convex-upwards and concave-upwards. These features are diagnostic of a cumulate origin for the clinopyroxenites. The distinct REE patterns indicate garnet- and spinel-rich cumulates. The notably HREE-enriched pattern of clinopyroxenite sample 5K26A (Fig. 5) demonstrates the presence of garnet as a cumulus phase, and indeed substantial garnet is observed in the mode of this sample. Type-II clinopyroxenites (e.g. sample C34) have relatively high LREE contents, convex-upwards REE patterns and variable Ni contents, typical of clinopyroxene-rich cumulates derived from a magma of basaltic composition (Liu et al., 2005). The nearly constant La in all samples, whatever the concentrations of the other LREE, could be attributed to a LREE-enriched primary melt composition, to a former trapped basaltic melt crystallized in the mantle, or to overestimation of La in the geochemical analysis. The ‘trapped basaltic melt’ hypothesis becomes rather weak as soon as some criteria are not satisfied; the high Mg-number (0· 87–0· 93) of the pyroxenes and olivine is consistent with values for such minerals from primitive mafic cumulates and is far from the range of known melts. None of the clinopyroxenites shows any distinct positive Eu anomaly, consistent with the absence of plagioclase in the observed mode and indicating that plagioclase was also not a cumulus phase at any earlier stage of evolution.
For geochemical modelling, primary liquid compositions (Table 3) obtained from the polybaric, batch and fractional, pMELTS melting models presented above were used as possible parent liquids for derivation of cumulates that could be compared with the type-I pyroxenites. The selected primary liquids used for isobaric crystallization modelling, picked from four pressures (1· 0, 1· 2, 1· 5 and 2· 0 GPa), correspond to polybaric fractional melting liquids aggregated from the solidus up to the final pressures chosen. We examine and evaluate only the batch crystallization of these liquids, assuming local equilibrium with the surrounding melt. The calculated fractionation paths for the 1· 5 GPa liquid yield the fractionation sequence Spl + Cpx, Spl + Cpx + Ol and Spl + Cpx + Ol + Grt, whereas the lower pressure (1 GPa) liquid yields the crystallization sequence Ol + Spl + Cpx, Spl + Cpx, Cpx + Pl and Cpx + Pl + Grt. The appearance of garnet in the fractionating assemblage, near 20% MgO, coincides with a change from increasing to decreasing concentration of the HREE in many of the samples. Figure 11a–f shows the results of crystallization modelling at various pressures, in which TiO2 and Na2O in the cumulates increase monotonically with decreasing MgO during progressive crystallization; Al2O3 and CaO increase until the appearance of garnet in the assemblage and then become constant or, at 2· 0 GPa, decrease somewhat. SiO2 decreases with decreasing MgO, whereas FeO* initially decreases and then increases after garnet enters the assemblage. Although the pyroxenite compositions exhibit more scatter than is described by the range of models considered from a single parent liquid, the average trend is rather similar to the calculated crystallization products (Fig. 11a–f) at a pressure interval 1· 0–1· 5 GPa. In terms of our calculated crystallization paths, early crystallization of spinel could be attributed to the manner of handling of Cr2O3 in solid phases in pMELTS, wherein the stability field of spinel is overestimated and spinel saturates earlier in the fractionating sequence (Asimow et al., 1995). The appropriate melt to produce the Kimi pyroxenites has a high-Mg basaltic composition (Table 3). The considerable scatter about the model fractionation trends could be attributed to modal variations as is commonly seen in cumulate rocks.
|
In summary, major and trace element relationships and geochemical modelling suggest that the Kimi pyroxenites are neither frozen melts nor melt–peridotite interaction products but they represent clinopyroxene-rich cumulates formed at pressures >1· 0 GPa. The P–T evolution of the pyroxenites from published thermobarometric data (Mposkos & Krohe, 2006
) distinguishes two equilibration stages: a first stage near the peak P–T conditions calculated from the Ca-Ts and enstatite component of a clinopyroxene at about 1000°C and >1· 8 GPa, and a second dominant stage characterized by decompression and cooling at depth. For the second stage, garnet–clinopyroxene geothermometry yields equilibration temperatures of 750–765°C at a pressure of 1· 5 GPa. The minimum pressure is constrained at 1· 1 GPa at 750°C (Mposkos & Krohe, 2006, fig. 9, reaction curve 1). The pMELTS modelling results suggest that the pyroxenites originally crystallized during fractionation at about 1–1· 5 GPa. Comparing the first-stage thermobarometric data, requiring equilibration above 1· 8 GPa, with the results from our pMELTS modelling, there is a small misfit in pressure. This is most probably within the uncertainties and shortcomings of the model and need not imply a stage of compression after crystallization.
Metasomatic signatures of the Kimi ultramafic rocks
Various geochemical features of a subset of the Kimi ultramafic rocks, the type-I pyroxenites and peridotite sample C4, suggest variable degrees of metasomatic enrichment. Other samples lack evidence for metasomatism. The type-I pyroxenites are enriched in the incompatible elements Ba, Sr, Na and K compared with the type-II clinopyroxenites and additionally show a depletion of Ce–Pr in the LREE pattern. Peridotite sample C4 shows similar REE patterns at lower absolute abundances. REE patterns similar to those of the Kimi pyroxenites have been observed from massif-type peridotites (e.g. Ronda; Garrido & Bodinier, 1999) and pyroxenite xenoliths (e.g. Hannuoba; Xu, 2002) and previously attributed to mantle metasomatism.
The enrichment is commonly attributed to a chromatographic effect during melt percolation (Navon & Stolper, 1987). Two speculative scenarios can be proposed to explain the sources for the metasomatic signature observed in one peridotite and the type-I pyroxenites. First, in the mantle wedge above a subduction zone, slab-derived melts or fluids unrelated to the origin and previous history of the Kimi rocks could have infiltrated the complex (Scambelluri et al., 2006). Second, at the site of crystallization of the pyroxenite cumulates, residual melts directly related to the origin of the Kimi rocks might have reached highly evolved compositions and then interacted with the pyroxenites and peridotites (Xu et al., 2002). This process might be akin to the formation of pegmatites, granophyres, and related evolved products late in the evolution of crustal magmatic systems. In the crust these late fluids form distinct veins, whereas in the mantle lithosphere environment, the evolved melts might instead penetratively react with their hosts. The enrichment patterns observed do not show any particular subduction affinity (e.g. HFSE depletion), so we prefer the second scenario as a mechanism to explain restricted local enrichment of selected Kimi ultramafic rocks. That only type-I pyroxenites and peridotite sample C4 show metasomatic features argues against a widespread regional infiltration event.
Geodynamic significance
In this section, we use our geochemical results compared with previously published petrological, thermobarometric, textural and geochronological data to constrain a suitable geodynamic framework for the formation of the Kimi ultramafic rocks.
The P–T thermal regime of the Kimi ultramafic rocks has been constrained by Mposkos & Krohe (2006). They suggested a high-pressure equilibration stage at P > 3· 0 GPa, within the garnet peridotite stability field, followed by annealing at lower crustal levels and high-temperature deformation. Isobaric cooling is indicated by garnet exsolution lamellae in Cpx-1 in the pyroxenites. Deformation-free mineral assemblages and an influx of water characterize this stage. Available geochronological data, including a garnet–whole-rock Sm–Nd age for a spinel–garnet pyroxenite of c. 119 ± 3· 5 Ma (Wawrzenitz & Mposkos, 1997) and an essentially synchronous U–Pb sensitive high-resolution ion microprobe (SHRIMP) age (c. 117· 4 ± 1· 9 Ma) from zircon from an eclogite (Liati et al., 2002), are interpreted to define the timing of the static-annealing stage inferred from the textural evidence (Mposkos & Krohe, 2006).
We suggest that the Kimi peridotites represent residues formed after adiabatic decompression melting involving a combination of near-fractional and batch melt extraction processes, whereas the associated pyroxenites are cumulates, crystallized at pressures exceeding 1 GPa. Two possible tectonic settings can be broadly considered: an oceanic setting and a continental one.
Possible oceanic setting
Various controlling factors influence the final depth of melting and onset of crystallization in a mid-ocean ridge environment, but the local spreading rate is believed to be the most important. In purely passive models, the upwelling rate is proportional to the spreading rate (Shen & Forsyth, 1995). Spreading rate controls mantle heat loss as a result of a larger time interval for conductive cooling at slow upwelling rates (Reid & Jackson, 1981; White et al., 2001). Our polybaric melting model suggests melting beginning in the garnet stability field and continuing adiabatically to a peak degree of melting of 23% at about 1 GPa, followed by the onset of crystallization of the resultant melts. In such a scenario the high-pressure crystallization of pyroxenites is more likely to occur beneath a slow- to ultraslow-spreading ridge, because melting at faster spreading rates should continue to near the base of the crust at <0· 3 GPa (Michael & Cornell, 1998; White et al., 2001).
In any case, an oceanic setting is not preferable, as it requires, after the isobaric cooling stage, a P–T path that would be characterized by increasing pressure and temperature accompanying the obduction of the Kimi ultramafic rocks during incorporation into the crust. However, as discussed above, the P–T history of the Kimi ultramafic rocks is characterized first by decompression and then by isobaric cooling at great depth. There is no evidence of a compression phase, which makes it unlikely that the assemblage could have been transferred from oceanic to continental lithosphere.
Possible continental setting
In an extending continental lithosphere, adiabatic decompression melting models (e.g. McKenzie & Bickle, 1988) suggest that the degree of partial melting is related to the thickness of the lithosphere as well as the asthenospheric potential temperature. In particular, if lithosphere is stretched to the same final thickness, then greater quantities of melt will be produced from the extension of an initially thicker lithosphere, because the stretching factor (β) is greater. In addition, with increasing lithosphere thickness, as in the case of a slow-spreading oceanic ridge, conductive cooling is increased, thereby enhancing the crystallization of melts coming from peridotite melting at greater depth.
The formation of the Kimi ultramafic rocks is compatible with a model of upwelling mantle in a continental environment. Following Medaris (1999), the Kimi peridotites are designated as mantle-derived peridotites of ‘low-P/T’ type. A more likely tectonic scenario for the ascent of the deep-seated Kimi mantle rocks is upwelling of hot asthenospheric mantle associated with break-off of the subducted oceanic lithosphere following continental collision, lithospheric delamination or diapirism (Medaris, 1999; Brueckner & Medaris, 2000). The metaperidotites and the associated pyroxenites of the UHP Kimi Complex probably represent a segment of upwelling mantle from the wedge above the subducted Vardar oceanic slab below the European continent. Mantle upwelling and emplacement of the hot peridotite at the base of the (thickened) continental crust occurred in Jurassic to Early Cretaceous times and may have underlain a magmatic arc or a back-arc extensional province. We venture to suggest that the stage of cooling at high pressure recorded in the P–T path may indicate renewed subduction of the oceanic slab and consequent refrigeration of the overlying material, probably combined with tectonic erosion of the overriding plate.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONSAMPLE CHARACTERIZATIONDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The ultramafic rocks of the UHP metamorphic Kimi Complex consist of peridotites with pyroxenite layers. The metaperidotites most probably represent residual compositions after variable degree of melt extraction in a melting regime that shares some characteristics with the mid-ocean ridge melting environment; hence the compositional trends are similar to those in abyssal peridotites. Melt modelling suggests a process intermediate between the simple end-members of equilibrium porous flow (i.e. batch melting) and perfectly fractional melt extraction; a mixed or two-porosity polybaric melting model beginning in the garnet stability field above 3 GPa and continuing adiabatically down to about 1 GPa satisfies most of the constraints. The differences between the Kimi suite and abyssal peridotites can be attributed to the relatively high pressure at which melting ceased in the probably sub-continental setting of the Kimi suite. There is scatter in the major element compositions that is not explained by this model and requires some combination of analytical uncertainty, initial source heterogeneity, and later modifications. The Kimi pyroxenites are neither frozen melts nor the products of melt–peridotite interaction but are likely to correspond to products of crystal segregation from a crystallizing melt; melts that could have been extracted from the host peridotites constitute plausible parents for such cumulates. The pyroxenites represent clinopyroxene-rich cumulate rocks crystallized at pressures >1 GPa. Various geochemical features of a subset of the Kimi ultramafic rocks suggest variable degrees of metasomatic enrichment, perhaps caused by highly evolved residual liquids that formed at the site of crystallization of the pyroxenites and infiltrated through the more evolved pyroxenites and some peridotites.
|
ACKNOWLEDGEMENTS |
|---|
I.B. and E.M. were financially supported by the Project ‘Pythagoras I’, co-funded by the European Social Fund (75%) and National Resources (25%), and by the National Technical University of Athens for the Special Research Project ‘Protagoras’. P.D.A. was financially supported by the US National Science Foundation through grant EAR-0239513. Critical and constructive reviews by O. Müntener, E. Hellebrand and an anonymous reviewer helped us to improve the manuscript and are gratefully acknowledged. We want to express our sincere thanks to M. Wilson for her helpful remarks and extraordinarily patient editorial handling.
*Corresponding author. Telephone: + 1-626-395-4133. Fax: + 1-626-568-0935. E-mail: asimow{at}gps.caltech.edu
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSAMPLE CHARACTERIZATIONDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Asimow PD. A model that reconciles major- and trace-element data from abyssal peridotites. Earth and Planetary Science Letters (1999) 169:303–319.[CrossRef][Web of Science]
Asimow PD. Steady-state mantle–melt interactions in one dimension: 2. Thermal interactions and irreversible terms. Journal of Petrology (2002) 43:1707–1724.
Asimow PD, Langmuir CH. The importance of water to oceanic mantle melting regimes. Nature (2003) 421:815–820.[CrossRef]
Asimow PD, Stolper EM. Steady-state mantle–melt interactions in one dimension: 1. Equilibrium transport and melt focusing. Journal of Petrology (1999) 40:475–494.[CrossRef][Web of Science]
Asimow PD, Hirschmann MM, Ghiorso MS, O’Hara MJ, Stolper EM. The effect of pressure-induced solid–solid phase-transitions on decompression melting of the mantle. Geochimica et Cosmochimica Acta (1995) 59:4489–4506.[CrossRef][Web of Science]
Asimow PD, Hirschmann MM, Stolper EM. Calculation of peridotite partial melting from thermodynamic models of minerals and melts. IV. Adiabatic decompression and the composition and mean properties of mid-ocean ridge basalts. Journal of Petrology (2001) 42:963–998.
Asimow PD, Dixon JE, Langmuir CH. A hydrous melting and fractionation model for mid-ocean ridge basalts: Application to the Mid-Atlantic Ridge near the Azores. Geochemistry, Geophysics, Geosystems (2004) 5. paper number Q01E16, doi:10.1029/2003GC000568.
Berly TJ, Hermann J, Arculus RJ, Lapierre H. Supra-subduction zone pyroxenites from San Jorge and Santa Isabel (Solomon Islands). Journal of Petrology (2006) 47:1531–1555.
Brueckner HK, Medaris LG. A general model for the intrusion and evolution of ‘mantle’ garnet peridotites in high-pressure and ultra-high-pressure metamorphic terranes. Journal of Metamorphic Geology (2000) 18:123–133.[CrossRef][Web of Science]
Brunelli D, Seyler M, Cipriani A, Ottolini L, Bonatti E. Discontinuous melt extraction and weak refertilization of mantle peridotites at the Vema lithospheric section (Mid-Atlantic Ridge). Journal of Petrology (2005) 47:745–771.[CrossRef][Web of Science]
Bulatov VK, Girnis AV, Brey GP. Experimental melting of a modally heterogeneous mantle. Mineralogy and Petrology (2002) 75:131–152.[CrossRef][Web of Science]
Burg JP, Ricou LE, Ivanov Z, Godfriaux I, Dimov D, Klain L. Syn-metamorphic nappe complex in the Rhodope Massif: structure and kinematics. Terra Nova (1996) 8:6–15.[Web of Science]
Burnham OM, Rogers NW, van Calsteren PW, Hawkesworth CJ. The petrogenesis of the eastern Pyrenean peridotites: An integrated study of their whole-rock geochemistry and Re–Os isotope composition. Geochimica et Cosmochimica Acta (1998) 62:2293–2310.[CrossRef][Web of Science]
Dantas C, Ceuleneer G, Gregoire M, Python M, Freydier R, Warren J, Dick JB. Pyroxenites from the Southwest Indian Ridge, 9–16°E: Cumulates from incremental melt fractions produced at the top of a cold melting regime. Journal of Petrology (2007) 48:647–660.
Dick HJB, Fisher RL, Bryan WB. Mineralogic variability of the uppermost mantle along mid-ocean ridges. Earth and Planetary Science Letters (1984) 69:88–106.[CrossRef][Web of Science]
Downes H. Formation and modification of the shallow sub-continental lithospheric mantle: a review of geochemical evidence from ultramafic xenolith suites and tectonically emplaced massifs of western and central Europe. Journal of Petrology (2001) 42:233–250.
Falloon TJ, Green DH, Danyushevsky LV, Faul UH. Peridotite melting at 1· 0 and 1· 5 GPa: an experimental evaluation of techniques using diamond aggregates and mineral mixes for determination of near-solidus melts. Journal of Petrology (1999) 40:1343–1375.[CrossRef][Web of Science]
Falloon TJ, Danyushevsky LV, Green DH. Peridotite melting at 1 GPa: Reversal experiments on partial melt compositions produced by peridotite–basalt sandwich experiments. Journal of Petrology (2001) 42:2363–2390.
Frey FA, Suen CJ, Stockman HW. The Ronda high temperature peridotite: Geochemistry and petrogenesis. Geochimica et Cosmochimica Acta (1985) 49:2469–2491.[CrossRef][Web of Science]
Garrido CJ, Bodinier JL. Diversity of mafic rocks in the Ronda peridotite: evidence for pervasive melt–rock reaction during heating of subcontinental lithosphere by upwelling asthenosphere. Journal of Petrology (1999) 40:729–754.[CrossRef][Web of Science]
Ghiorso MS, Sack RO. Chemical mass transfer in magmatic systems processes IV. A revised and internally consistent thermodynamic model for the interpolation and extrapolation of liquid–solid equilibria in magmatic systems at elevated temperatures and pressures. Contributions to Mineralogy and Petrology (1995) 119:197–212.[Web of Science]
Ghiorso MS, Hirschmann MM, Reiners PW, Kress VC. The pMELTS: a revision of MELTS for improved calculation of phase relations and major element partitioning related to partial melting of the mantle to 3 GPa. Geochemistry, Geophysics, Geosystems (2002) 3. doi:10.1029/2001GC000217.
Grove TL, Kinzler RJ, Bryan WB. Fractionation of mid-ocean ridge basalt (MORB). In: Mantle Flow and Melt Generation at Mid-Ocean Ridges. Geophysical Monograph, American Geophysical Union—Phipps Morgan J, Blackman DK, Sinton JM, eds. (1992) 71:281–310.
Hellebrand E, Snow JE, Hoppe P, Hofman AW. Garnet-field melting and late-stage refertilization in ‘residual’ abyssal peridotites from the Central Indian Ridge. Journal of Petrology (2002) 43:2305–2338.
Hermann J, Spandler C, Hack A, Korsakov AV. Aqueous fluids and hydrous melts in high-pressure and ultra-high pressure rocks: implications for element transfer in subduction zones. Lithos (2006) 92:399–417.[CrossRef][Web of Science]
Hirose K, Kushiro I. Partial melting of dry peridotites at high pressures: determination of compositions of melts segregated from peridotite using aggregates of diamond. Earth and Planetary Science Letters (1993) 114:477–489.[CrossRef][Web of Science]
Hirschmann MM, Stolper EM. A possible role for garnet pyroxenite in the origin of the ‘garnet signature’ in MORB. Contributions to Mineralogy and Petrology (1996) 124:185–208.[CrossRef][Web of Science]
Irving AJ. Geochemical and high pressure experimental studies of garnet pyroxenite and pyroxene granulites xenoliths from the Delegate Basaltic Pipes, Australia. Journal of Petrology (1974) 15:1–40.
Iwamori H. Dynamic disequilibrium melting model with porous flow and diffusion-controlled chemical equilibration. Earth and Planetary Science Letters (1993) 114:301–313.[CrossRef][Web of Science]
Johnson TM, Dick HJB, Shimizu N. Melting in the oceanic upper mantle: an ion microprobe study of diopside in abyssal peridotites. Journal of Geophysical Research (1990) 95:2661–2678.
Kelemen PB, Hirth G, Shimizu N, Spiegelman M, Dick HJB. A review of melt migration processes in the adiabatically upwelling mantle beneath oceanic spreading ridges. Philosophical Transactions of the Royal Society of London, Series A (1997) 355:283–318.[CrossRef]
Kelemen PB, Hart SR, Bernstein S. Silica enrichment in the continental upper mantle via melt/rock reaction. Earth and Planetary Science Letters (1998) 164:387–406.[CrossRef][Web of Science]
Kelemen PB, Braun M, Hirth G. Spatial distribution of melt conduits in the mantle beneath oceanic spreading ridges: Observations from the Ingalls and Oman ophiolites. Geochemistry, Geophysics, Geosystems (2000) 1. doi:10.1029/1999GC000012.
Kinzler RJ, Grove TL. Primary magmas of mid-ocean ridge basalts 2. Applications. Journal of Geophysical Research (1992) 97(B5):6907–6926.
Klein EM, Langmuir CH. Global correlations of ocean ridge basalt chemistry with axial depth and crustal thickness. Journal of Geophysical Research (1987) 92:8089–8115.
Kopylova MG, Russell JK, Cookenboo H. Petrology of peridotite and pyroxenite xenoliths from the Jericho kimberlite: implications for the thermal state of the mantle beneath the Slave craton, northern Canada. Journal of Petrology (1999) 40:79–104.[CrossRef][Web of Science]
Krohe A, Mposkos E. Multiple generations of extensional detachments in the Rhodope Mountains (N. Greece): evidence of episodic exhumation of high-P rocks. In: The Timing and Location of Major Ore Deposits in an Evolving Orogen. Geological Society, London, Special Publications—Blundell DJ, Neubauer G, Von Quant A, eds. (2002) 204:151–178.
Langmuir CH, Klein EM, Plank T. Petrological systematics of mid-ocean ridge basalts: Constraints on melt generation beneath ocean ridges. In: Mantle Flow and Melt Generation at Mid-Ocean Ridges—Phipps Morgan J, Blackman DK, Sinton JM, eds. (1992) 183–280.
Liati A, Gebauer D. Constraining the prograde and retrograde P–T–t path of Eocene HP rocks by SHRIMP dating of different zircon domains: inferred rates of heating, burial, cooling and exhumation for central Rhodope, northern Greece. Contributions to Mineralogy and Petrology (1999) 135:340–354.[CrossRef][Web of Science]
Liati A, Gebauer D, Wysoczanski R. U–Pb SHRIMP-dating of zircon domains from UHP mafic rocks in the Rhodope zone (N. Greece); evidence for Early Cretaceous crystallization and Late Cretaceous metamorphism. Chemical Geology (2002) 184:281–300.[CrossRef][Web of Science]
Litasov KD, Foley SF, Litasov YD. Magmatic modification and metasomatism of the subcontinental mantle beneath the Vitim volcanic field East Siberia: evidence from trace element data on pyroxenite and peridotite xenoliths from Miocene picrobasalt. Lithos (2000) 54:83–114.[CrossRef][Web of Science]
Liu Y, Gao S, Aeolus Lee C.-T, Hu S, Liu X, Yuan H. Melt–peridotite interactions: Links between garnet pyroxenite and high-Mg# signature of continental crust. Earth and Planetary Science Letters (2005) 234:39–57.[CrossRef][Web of Science]
Martin RF. Symbols of the rock-forming minerals. The nomenclature of minerals, in. In: The Nomenclature of Minerals: A Compilation of IMA Reports (1998) Québec: Mineralogical Association of Canada. 148–149.
McKenzie D. The generation and compaction of partial melts. Journal of Petrology (1984) 25:713–765.
McKenzie D, Bickle MJ. The volume and composition of melt generated by extension of the lithosphere. Journal of Petrology (1988) 29:625–679.
Medaris LG. Garnet peridotite in Eurasian HP and UHP terranes: a diversity of origins and thermal histories. International Geology Reviews (1999) 41:799–815.
Mezger K, Essene EJ, Hallidy AN. Closure temperature of the Sm–Nd system in metamorphic garnets. Earth and Planetary Science Letters (1992) 8:397–409.
Michael PJ, Cornell WC. Influence of spreading rate and magma supply on crystallization and assimilation beneath mid-ocean ridges: evidence from chlorine and major element chemistry of mid-ocean ridge basalts. Journal of Geophysical Research (1998) 103:18325–18356.[CrossRef]
Montanini A, Tribuzio R, Anczkiewicz R. Exhumation history of a garnet pyroxenite-bearing mantle section from a continent–ocean transition (Northern Apennine ophiolites, Italy). Journal of Petrology (2006) 47:1943–1971.
Mposkos E. Petrology of the ultra-high pressure metamorphic Kimi complex in Rhodope (N.E. Greece): a new insight into the Alpine geodynamic evolution of the Rhodope. Bulletin of Geological Society of Greece (2002) XXXIV:2169–2188.
Mposkos E, Kostopoulos D. Diamond, former coesite and supersilicic garnet in metasedimentary rocks from the Greek Rhodope: a new ultrahigh-pressure metamorphic province established. Earth and Planetary Science Letters (2001) 192:497–506.[CrossRef][Web of Science]
Mposkos E, Krohe A. Petrological and structural evolution of continental high pressure (HP) metamorphic rocks in the Alpine Rhodope domain (N. Greece). In. Panayides I, Xenophontos C, Malpas J, eds. (2000) Nicosia: Geological Survey of Cyprus. 221–232. Proceedings of the 3rd International Conference on the Geology of the Eastern Mediterranean.
Mposkos E, Krohe A. P–T–deformation paths of closely associated UHP (diamond-bearing) crustal and mantle rocks of the Kimi Complex: Implications for the tectonic history of the Rhodope Mountains, northern Greece. Canadian Journal of Earth Sciences (2006) 43:1755–1776.
Mposkos E, Wawrzenitz N. Metapegmatites and pegmatites bracketing the time of HP-metamorphism in polymetamorphic rocks of the E-Rhodope: Petrological and geochronological constraints. Geological Society of Greece, Special Publication (1995) 2(4):602–608.
Mposkos E, Baziotis I, Palikari S, Perraki M, Krohe A, Hoinkes G. Alpine UHP metamorphism in the Kimi complex of the Rhodope HP province N.E. Greece: mineralogical and textural indicators. In: 32nd International Geological Congress, Florence, Italy (2004) 1:18–28. 108.
Navon O, Stolper E. Geochemical consequence of melt percolation: the upper mantle as a chromatographic column. Journal of Geology (1987) 95:285–307.[Web of Science]
Niu Y. Mantle melting and melt extraction processes beneath ocean ridges: Evidence from abyssal peridotites. Journal of Petrology (1997) 38:1047–1074.[CrossRef][Web of Science]
Niu Y. Bulk-rock major and trace element compositions of abyssal peridotites: implications for mantle melting, melt extraction and post-melting processes beneath mid-ocean ridges. Journal of Petrology (2004) 45:2423–2458.
Paulick H, Bach W, Godard M, De Hoog JCM, Suhr G, Harvey J. Geochemistry of abyssal peridotites (Mid-Atlantic Ridge, 15°20'N, ODP Leg 209): Implications for fluid/rock interaction in slow spreading environments. Chemical Geology (2006) 234:179–210.[CrossRef][Web of Science]
Pearson DG, Davies GR, Nixon PH. Geochemical constraints on the petrogenesis of diamond facies pyroxenites from the Beni Bousera peridotite massif, north Morocco. Journal of Petrology (1993) 34:125–172.
Perraki M, Proyer A, Mposkos E, Kaindl R, Baziotis I. Micro- and nanodiamonds in garnets of metapelitic rocks from the Greek Rhodope: an in situ micro-Raman study. In. Chatzipetros AA, Pavlides SB, eds. (2004) Thessaloniki: Aristotle University of Thessaloniki. 1216–1219. Proceedings of the 5th International Symposium for Eastern Mediterranean Geology.
Perraki M, Proyer A, Mposkos E, Kaindl R, Hoinkes G. Raman micro-spectroscopy on diamond, graphite and other carbon polymorphs from the ultrahigh-pressure metamorphic Kimi Complex of the Rhodope Metamorphic Province, NE Greece. Earth and Planetary Science Letters (2006) 241:672–685.[CrossRef][Web of Science]
Plank T, Spiegelman M, Langmuir CH, Forsyth D. The meaning of ‘mean F’: Clarifying the mean extent of melting at ocean ridges. Journal of Geophysical Research (1995) 100:15045–15052.[CrossRef]
Presnall DC, Gudfinnsson GH, Walter MJ. Generation of mid-ocean ridge basalts at pressures from 1 to 7 GPa. Geochimica et Cosmochimica Acta (2002) 66:2073–2090.[CrossRef][Web of Science]
Reid I, Jackson HR. Oceanic spreading rate and crustal thickness. Marine Geophysical Research (1981) 5:165–172.
Ribe NM. The generation and composition of partial melts in the Earth's mantle. Earth and Planetary Science Letters (1985) 73:361–376.[CrossRef][Web of Science]
Ricou LE, Burg JP, Godfriaux I, Ivanov Z. Rhodope and Vardar: the metamorphic and olistostromic paired belts related to Cretaceous subduction under Europe. Geodinamica Acta (1998) 11:285–309.[CrossRef][Web of Science]
Roeder PL, Emslie RF. Olivine–liquid equilibrium. Contributions to Mineralogy and Petrology (1970) 29:275–289.[CrossRef][Web of Science]
Salters VJM, Hart SR. The hafnium paradox and the role of garnet in the source of mid-oceanic-ridge basalts. Nature (1989) 342:420–422.[CrossRef]
Scambelluri M, Hermann J, Morten L, Rampone E. Melt- versus fluid-induced metasomatism in spinel to garnet wedge peridotites (Ulten Zone, Eastern Italian Alps): clues from trace element and Li abundances. Contributions to Mineralogy and Petrology (2006) 151:372–394.[CrossRef][Web of Science]
Schiano P, Eiler JM, Hutcheon ID, Stolper EM. Primitive CaO-rich, silica-undersaturated melts in islands arcs: evidence for the involvement of clinopyroxene-rich lithologies in the petrogenesis of arc magmas. Geochemistry, Geophysics, Geosystems (2000) 1. doi:10.1029/1999GC000032.
Shen Y, Forsyth DW. Geochemical constraints on initial and final depths of melting beneath mid-ocean ridges. Journal of Geophysical Research (1995) 100:2211–2237.[CrossRef]
Smith PM, Asimow PD. Adiabat_1ph: a new public front-end to the MELTS, pMELTS, and pHMELTS models. Geochemistry, Geophysics, Geosystems (2005) 6. paper number Q02004.
Sobolev AV, Hofmann AW, Sobolev SV, Nikogosian IK. An olivine-free mantle source of Hawaiian shield basalts. Nature (2005) 434:590–597.[CrossRef][Medline]
Spiegelman M, Elliott T. Consequences of melt transport for uranium series disequilibrium in young lavas. Earth and Planetary Science Letters (1993) 118:1–20.[CrossRef][Web of Science]
Spiegelman M, Kenyon P. The requirements for chemical disequilibrium during magma migration. Earth and Planetary Science Letters (1992) 109:611–620.[CrossRef][Web of Science]
Sun SS, McDonough WF. Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and process. In: Magmatism in the Ocean Basins. Geological Society, London, Special Publications—Saunders AD, Norry JM, eds. (1989) 42:313–345.
Szabo C, Falus G, Zajacz Z, Kovacs I, Bali E. Composition and evolution of lithosphere beneath the Carpathian–Pannonian region: a review. Tectonophysics (2004) 393:119–137.[CrossRef][Web of Science]
Takazawa E, Frey FA, Shimizu N, Saal A, Obata M. Polybaric petrogenesis of mafic layers in the Horoman Peridotite Complex, Japan. Journal of Petrology (1999) 40:1827–1851.[CrossRef][Web of Science]
Ulmer P. The dependence of the Fe2+–Mg cation-partitioning between olivine and basaltic liquid on pressure, temperature and composition, an experimental study to 30 kbars. Contributions to Mineralogy and Petrology (1989) 101:261–273.[CrossRef][Web of Science]
Wasylenki LE, Baker MB, Kent AJR, Stolper EM. Near-solidus melting of the shallow upper mantle: Partial melting experiments on depleted peridotite. Journal of Petrology (2003) 44:1163–1191.
Wawrzenitz N, Mposkos E. First evidence for lower Cretaceous HP/HT-metamorphism in the Eastern Rhodope, North Aegean Region, North-East Greece. European Journal of Mineralogy (1997) 9:659–664.
Weyer S, Münker C, Mezger K. Nb/Ta, Zr/Hf and REE in the depleted mantle: implications for the differentiation history of the crust–mantle system. Earth and Planetary Science Letters (2003) 205:309–324.[CrossRef][Web of Science]
White RS, Minshull TA, Bickle MJ, Robinson CJ. Melt generation at slow-spreading oceanic ridges: Constraints from geochemical and geophysical data. Journal of Petrology (2001) 42:1171–1196.
Workman RK, Hart SR. Major and trace element composition of the depleted MORB mantle (DMM). Earth and Planetary Science Letters (2005) 231:53–72.[CrossRef][Web of Science]
Xu YG. Evidence for crustal components in the mantle and constraints on crustal recycling mechanisms: pyroxenite xenoliths from Hannuoba, North China. Chemical Geology (2002) 182:301–322.[CrossRef][Web of Science]
Xu YG, Sun M, Yan W, Liu Y, Huang XL, Chen XM. Xenolith evidence for polybaric melting and stratification of the upper mantle beneath South China. Journal of Asian Earth Sciences (2002) 20:937–954.[CrossRef][Web of Science]
Yu J-H, O’Reilly SY, Zhang M, Griffin WL, Xu X. Roles of melting and metasomatism in the formation of the lithospheric mantle beneath the Leizhou peninsula, South China. Journal of Petrology (2006) 47:355–383.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||