Journal of Petrology | Volume 39 | Number 4 | Pages 663-688 | 1998
© Oxford University Press 1998
Contact Metamorphism in Pelitic Rocks on the Island of Kos (Greece, Eastern Aegean Sea): a Test for the Na-in-Cordierite Thermometer
Mineralogisches Institut, University of Heidelberg IM Neuenheimer Feld 236, D-69120 Heidelberg, Germany
Received April 10, 1996; Revised typescript accepted November 17, 1997
 |
ABSTRACT |
|---|
 TOP ABSTRACT IntroductionGeological Setting and SamplesAnalytical MethodsPetrography and Mineral...Calculation of PressuresCalculation of TemperaturesFluid Composition During Contact...DiscussionSummary and ConclusionsReferences |
|---|
Most of the applicable thermometers involving cordierite require additional Fe–Mg phases and knowledge of pressure and fluid phase composition. These conditions are not necessarily met with natural cordierite-bearing metamorphic rocks. A potential pressure-independent geothermometer for cordierite- and plagioclase-bearing rocks is the experimentally determined inverse linear relationship between temperature and Na contents of Mg-cordierite coexisting with NaOH or albite and NaOH. In this paper, the Na-in-cordierite thermometer is empirically tested on cordierite-bearing metapelites from the island of Kos, contact metamorphosed by a quartz monzonite intrusion. The variety of phase assemblages in the metapelites allows for tight constraints to be placed on temperatures (508–762°C) and pressure (0.2 GPa) during contact metamorphism by calculated and experimentally determined phase equilibria and by Al-in-hornblende barometry. Fluid compositions are crudely estimated from the presence and absence of graphite and calcite and from 2Vx in cordierite. Applying the Na-in-cordierite thermometer to 19 rock samples taken at increasing distances from the intrusion yields realistic peak temperatures. The Na-in-cordierite temperatures are consistent with constraints from petrogenetic grids, with experimentally determined phase equilibria and with temperatures based on Fe–Mg exchange equilibria (garnet–cordierite and garnet–biotite), although the latter tend to give lower temperatures. The Na-in-cordierite temperatures show no systematic variations with Mg values and Al contents of cordierites, Na contents of coexisting plagioclase (in the range An10–An40) and biotite, and H2O/CO2 ratios in coexisting fluids, though the last are only poorly known. Cordierites have either high Na and significant Li contents or display low levels of both elements. Li is probably introduced into cordierite by a coupled substitution involving tetrahedral and octahedral sites, whereas Na may not be charge-balanced by lattice oxygen. The results obtained suggest that the experimentally established relationship between Na contents of cordierite and temperature is a useful thermometer for cordierite- and plagioclase-bearing metapelitic rocks. The nature of Na incorporation in cordierite and its relation to fluid composition, however, must be constrained by further experiments.
KEY WORDS: Contact metamorphism; Kos; metapelites; Na-in-cordierite thermometer
|
Introduction |
|---|
|
TOPABSTRACTIntroductionGeological Setting and SamplesAnalytical MethodsPetrography and Mineral...Calculation of PressuresCalculation of TemperaturesFluid Composition During Contact...DiscussionSummary and ConclusionsReferences |
|---|
The calculation of pressures and temperatures for metamorphic rocks is one of the keys to understanding orogenic processes. Metamorphism at high temperatures and low to medium pressures is widespread in many orogenic belts. The precursors to these high-temperature metamorphic rocks are commonly sediments of psammitic to pelitic composition that very often form cordierite during metamorphism.
A number of common phase equilibria involving cordierite have been used as geothermometers, e.g. the reactions bt + sil + qtz = crd + kfs + v, grt + sil + qtz = crd, and spl + qtz = crd (Richardson, 1968
; Hensen, 1971; Weisbrod, 1973; Holdaway & Lee, 1977; Lonker, 1981; Martignole & Sisi, 1981; Vielzeuf, 1983; Bhattacharya et al., 1988; Mukhopadhyay & Holdaway, 1994). Although these phase equilibria are widely used to constrain temperatures, they bear several handicaps that restrict their application:
- All three reactions are strongly pressure dependent, and pressures can often not be calculated independently because of the lack of suitable phase assemblages.
- The respective experiments have been carried out under conditions of PH2O = Ptotal. Although equilibration pressures and temperatures for equilibria involvingcordierite may be calculated for other values of aH2O(Mukhopadhyay & Holdaway, 1994; Carey, 1995), the problem remains that for many natural rocks no constraints can be placed on the existence or composition of a fluid phase during metamorphism.
- Fe–Mg exchange equilibria such as garnet–cordierite, biotite–cordierite, and garnet–biotite are sensitive to retrograde diffusion, as Fe–Mg diffusion is known to be fast compared with most other main elements at least in garnet (Elphick et al., 1985) and to impede the calculation of peak metamorphic temperatures in many cases (Jiang & Lasaga, 1990; Spear, 1991).
- Garnet is not always present and its common zoning may also hamper the calculation of peak temperatures (Florence & Spear, 1991).
- Biotite is very readily transformed to chlorite during hydrothermal overprint, resulting in a change of Fe/Mg ratios.
Cordierite is a tectosilicate with channel-like cavities parallel to its crystallographic c-axis (Schreyer, 1965; Gibbs, 1966; Armbruster & Bloss, 1981). Apart from volatile species such as H2O, CO2, and minor N2 and Ar (Armbruster & Bloss, 1980, 1982; Johannes & Schreyer, 1981; Armbruster, 1985; Le Breton, 1989; Vry et al., 1990; Santosh et al., 1993), these channels can preferentially accommodate alkali ions (Schreyer, 1965; Gibbs, 1966; Goldman et al., 1977). Measurable amounts of K have also been found (Schreyer et al., 1990). Li has often not been analysed but can reach significant amounts (Armbruster et al., 1982; Gordillo et al., 1985; Cerny et al., 1997). Significant but varying Na contents are very often encountered (Armbruster & Irouschek, 1983; Armbruster, 1986; `, 1986). The Na content of Mg-cordierite coexisting with albite and NaOH has been shown to be a pressure-independent function of temperature in the range between 0.2 and 0.8 GPa at 650–850°C (Mirwald, 1986; Fig. 1). Hence, this thermometer offers the possibility to calculate pressure-independent temperatures from a variety of cordierite-bearing rocks, the precondition of Na availability being fulfilled in most high-temperature metapelitic and metapsammitic rocks by the presence of sodic plagioclase. Although the Na-in-cordierite thermometer has been found to yield values consistent with independently constrained temperatures for several metamorphic rocks from diverse tectonic settings (Mirwald, 1986), there has been no systematic empirical evaluation of the thermometer up to now.
|
The latter is the purpose of the present investigation. The applicability of the Na-in-cordierite thermometer is tested by using it on natural rocks from a contact aureole where independent temperature constraints are available from calculated petrogenetic grids, experimentally determined phase equilibria, and Fe–Mg exchange thermometry. To test whether the experimental relationship found in a simple system may be extended to complex natural systems, the following issues are addressed. (1) As the thermometer was calibrated with Mg-cordierite but most of the natural cordierites contain significant amounts of Fe2+, the possible influence of Fe/Mg ratios on Na incorporation must be checked. (2) A possible correlation between the Na contents in cordierite and the Na contents of coexisting plagioclase and biotite must be tested as the thermometer is calibrated with pure albite and NaOH. (3) Testing whether Na incorporation is also controlled by coexisting fluid phases is another important point, as natural metamorphic rocks have often equilibrated either under conditions with PH2O < Ptotal or possibly under fluid-absent conditions, whereas the thermometer is calibrated at conditions of PH2O
= Ptotal.
Contact metamorphism in pelitic rocks on the island of Kos (Greece, Eastern Aegean Sea), caused by the intrusion of a quartz monzonite, was considered appropriate for such an investigation, as the petrological requirements detailed above are all met. Contact metamorphic mineral reactions involving cordierite can be observed over a wide temperature range of 550–800°C (Altherr et al., 1976) at constant load pressures. The chemical diversity of the metapelites has produced a variety of phase assemblages that allow temperatures and pressure to be constrained from a number of calculated and experimentally established exchange equilibria. The chemical diversity also accounts for a wide range in bulk Fe/Mg ratios (and thus in cordierite) as well as for a considerable range in the An contents of coexisting plagioclase. The partly intimate interlayering of metapelites with marbles and calcic metapelites has yielded variable H2O–CO2 fluid compositions during contact metamorphism.
|
Geological Setting and Samples |
|---|
|
TOPABSTRACTIntroductionGeological Setting and SamplesAnalytical MethodsPetrography and Mineral...Calculation of PressuresCalculation of TemperaturesFluid Composition During Contact...DiscussionSummary and ConclusionsReferences |
|---|
Geology of Kos
The island of Kos off the Turkish coast in the Eastern Aegean Sea (Fig. 2) forms part of the Hellenic-Tauric system, a complex mountain belt extending from southeastern Europe to Turkey. The Hellenic arc results from the collision of the Apulian microcontinent with the crystalline complexes of northeastern Greece in Mesozoic to Cenozoic times. It displays a pronounced southwest vergent nappe character. Apart from minor Palaeozoic units, the nappes consist of Mesozoic to Cenozoic sediments and magmatic rocks, including ophiolites. Some of these nappes have been subject to different degrees of metamorphism during collision [see compilation by Jacobshagen, (1986)
]. Metamorphism and thrusting took place during several events between Jurassic and Miocene times (Altherr et al., 1994, and references cited therein; Seidel et al., 1977, 1982).
|
The geological units on the island of Kos form part of the Central Hellenic nappes (Fig. 2, Jacobshagen, 1986
). The oldest units form the Dicheo Massif in the central part of the island towards the south coast (Fig. 2b and c). They consist of Permocarboniferous marls, impure limestones and sandstones, phyllites and rare mafic intercalations (Altherr et al., 1976; Gralla, 1982) that have been subject to post-Permocarboniferous regional metamorphism at very low grades (qtz + ms ± chl ± cc). At 12 Ma, this series was partly contact metamorphosed by the intrusion of a large quartz monzonite (Altherr et al., 1982; Henjes-Kunst et al., 1988). Subsequent to intrusion and exhumation Permocarboniferous metasediments and Miocene quartz monzonite were tectonically overridden by Cretaceous to Eocene limestones that crop out mainly north and west of the pluton (Fig. 2c). After tectonic emplacement of these limestones, the area west of the Dicheo Massif was covered by Pliocene and Pleistocene sediments and mostly Quaternary volcanics (Keller, 1969; Dürr & Jacobshagen, 1986). To the east, the Dicheo Massif is tectonically bordered by a flysch sequence of late Cretaceous to Tertiary age (Dürr & Jacobshagen, 1986).
The Permocarboniferous sedimentary unit contains a basal marble–metapelite sequence (1000 m), followed by 200 m of massive marble, 100 m of calc-silicate rocks and a top layer of dolomite marble (Gralla, 1982). The unit is mainly exposed to the east of the quartz monzonite intrusion (Fig. 2c). Within the basal marble–metapelite sequence and the calc-silicate rocks, contact metamorphic mineral assemblages and mineral isograds have been described and mapped (Altherr et al., 1976; Fig. 2). From mineral parageneses and experimentally determined phase relations, temperatures of 
800°C at the contact declining to 500°C over a distance of 4 km at pressures of 0.3–0.6 GPa were deduced (Altherr et al., 1976).
Sample selection
The rationale behind the sample selection was (1) to cover the entire temperature range of contact metamorphism through which cordierite is stable, to test the applicability of the Na-in-cordierite thermometer, and (2) to study different bulk compositions at a given temperature, to check for possible influences of chemical parameters on the Na contents of cordierite. For the first purpose, samples were selected on the basis of their cartographic distance from the intrusion (Fig. 2). However, precise fixing of the actual distance from the intrusion was not always possible, because of morphological effects and the complex shape of the intrusion and numerous dykes associated with it. For the second purpose, samples with different phase assemblages were chosen from the same locality wherever possible. Only very fresh samples were selected, with no indication of pinitization of cordierite. Furthermore, only those samples where petrographic evidence for equilibrium between cordierite and plagioclase was obvious were included in the investigation. Thus, a total of 19 cordierite-bearing samples were selected (Table 1).
|
|
Analytical Methods |
|---|
|
TOPABSTRACTIntroductionGeological Setting and SamplesAnalytical MethodsPetrography and Mineral...Calculation of PressuresCalculation of TemperaturesFluid Composition During Contact...DiscussionSummary and ConclusionsReferences |
|---|
Mineral analyses were performed using a CAMECA SX 51 microprobe equipped with five wavelength-dispersive spectrometers. Operating parameters were 20 nA beam current, 15 kV accelerating voltage, 10 s counting time for all elements except Ti in spinel (30 s), Mg, Ca, Al in spinel (20 s) and Zn in spinel, staurolite, and cordierite (40 s). PAP correction was applied to the data. Natural and synthetic silicate and oxide standards were used for calibration.
The detection limit for Na in natural oxide and silicate standards is 0.049 wt % employing operating parameters as detailed above. Under these conditions, the amount and precision of Na2O measurements (n = 20) is 0.10 ± 0.01 wt % (0.007 ± 0.001 Na p.f.u.) on a diopside standard, 1.32 ± 0.04 wt % (0.094 ± 0.003 Na p.f.u.) on an augite standard, 0.85 ± 0.003 wt % (0.060 ± 0.002 Na p.f.u.) on a Cr-augite standard, and 0.12 ± 0.02 wt % (0.018 ± 0.002 Na p.f.u.) on a garnet standard. Analyses using a defocused beam or longer counting times at lower beam currents did not change these values significantly. In all measurements, Na and K were analysed first, to avoid volatilization.
Analysis of Li and Be in the cordierites was performed using a CAMECA ims3f ion-probe. Depending on the grain size of the cordierites a primary beam of 14.5 keV O– ions with a current ranging from 30 pA to 2 nA was used, resulting in a beam diameter of 5–30 µm. Secondary ions with an energy range of 0–10 eV were analysed with a mass resolution of 2000. As reference material to determine the relative ion yields of Li and Be the NIST SRM-610 glass with 428 ppm Li and 421 ppm Be (Rocholl et al., 1997) and cordierite TS25 [described by Schreyer et al., (1993) and analysed by Cerny et al., (1997) for Li and Be by inductively coupled plasma mass spectrometry (ICP MS) and ion-probe] with 1626 ppm Li and 7.2 ppm Be were used for daily routine calibration. For our analytical conditions cordierite showed no significant matrix effect for Li and Be compared with SRM-610. The accuracy of our analyses is limited by the accuracy of the Li and Be data for TS25 and NIST SRM-610 and is therefore not better than 15–20%. Ten analyses per point were typically performed, resulting in a standard deviation <5% (1
). Five to 15 points in cordierites were run per sample.
|
Petrography and Mineral Chemistry |
|---|
|
TOPABSTRACTIntroductionGeological Setting and SamplesAnalytical MethodsPetrography and Mineral...Calculation of PressuresCalculation of TemperaturesFluid Composition During Contact...DiscussionSummary and ConclusionsReferences |
|---|
The rocks display various assemblages comprising the phases cordierite + plagioclase + biotite + ilmenite + quartz ± garnet ± Al2SiO5 ± K-feldspar ± orthoamphibole ± staurolite ± chloritoid ± hercynite ± accessory phases (Table 1, Figs 3 and 4). All of the samples (except for K-346, Fig. 3c) are characterized by heterogranular textures, the largest grains in all cases being garnet (where present), cordierite, and andalusite (where present). Most of the textures are poikiloblastic, whereas nematoblastic textures are only rarely observed (K-345, K-267, K-151, Fig. 3d). Grain sizes range from very fine grained to medium grained, and the rocks may display either isotropic textures or foliation to variable degrees. Most of the large cordierite blasts show sector trilling and complex twinning as described by Venkatesh, (1954)
, Zeck, (1972) and Maresch et al., (1991).
|
|
Mineral phases in all samples were tested for inter-grain and intra-grain chemical variations by microprobe analyses. Cordierite compositions are given in Table 2 and Table 3; the range of Na contents for each sample is shown in Fig. 5. Representative compositions of other main phases from key samples are listed in Table 4. Additionally, the XMg ranges of cordierite, garnet (where present), biotite and ilmenite, as well as the An contents of plagioclase for all samples are given in Table 5. Significant inter-grain variations are only displayed by biotites, namely, in Ti contents and XMg values (Table 5) with no clear relation to textural sites. Single grains of all phases were generally found to be unzoned except for the outermost rims, which may display variable changes in element concentrations as a result of retrogression. Large garnets may record extremely weak zoning with slightly higher XMg values and Mn contents in the cores (Table 5). In cordierite grains, intra-grain scatter in Na and Li contents was observed. For samples K-339 and K-357, intra-grain scatter of all element contents in the extremely poikiloblastic cordierites was found. This is probably due to formation temperatures only slightly above the minimum required for cordierite (see section on accuracy and precision of the Na-in-cordierite temperatures), resulting in only sluggish achievement of equilibrium. Moreover, because of the extremely fine-grained nature of these two samples, it may be possible that spot analyses partly also hit tiny inclusions. This applies particularly to ion-probe analyses (beam diameter 5 µm) but may also affect microprobe analyses (beam diameter 1 µm).
|
|
|
|
|
|
Calculation of Pressures |
|---|
|
TOPABSTRACTIntroductionGeological Setting and SamplesAnalytical MethodsPetrography and Mineral...Calculation of PressuresCalculation of TemperaturesFluid Composition During Contact...DiscussionSummary and ConclusionsReferences |
|---|
Pressures calculated from Al contents of amphiboles from the intrusion
Hammarstrom & Zen, (1986)
and Hollister et al., (1987) showed empirically and theoretically that in near-solidus calc-alkalic plutons containing quartz, plagioclase, potassium feldspar, biotite, amphibole, sphene, and Fe–Ti oxides and melt in equilibrium, the aluminium content of amphibole rims is controlled by total pressure. The Al-in-hornblende barometer was experimentally confirmed and calibrated by Johnson & Rutherford, (1989) with different natural starting materials between 0.2 and 0.8 GPa at temperatures of 720–780°C and H2O/CO2 fluid ratios between 50/50 and 25/75. The experimental data yielded a different equation for the regression line, resulting in lower pressures for any given Al content when compared with the older empirical calibrations. Experimental results that match the empirical relations more closely were obtained by Schmidt, (1992) on a natural tonalite under water-saturated conditions between 0.25 and 1.3 GPa at temperatures of 655–700°C.
According to the classification of Streckeisen, (1976), the intrusive body on Kos is mainly a quartz monzonite, with compositions ranging to monzogranites and quartz monzodiorites (Altherr et al., 1982, 1988). The quartz monzonitic massif contains dioritic inclusions and is crosscut by numerous aplitic and lamprophyric dykes. The modal mineralogy of the quartz monzonite (Altherr et al., 1982) is quartz (2–16 vol. %), plagioclase (20–50 vol. %), orthoclase (15–35 vol. %), biotite (5–10 vol. %), hornblende (2–10 vol. %), sphene, magnetite, and orthite (0–1 vol. %), and in places clinopyroxene (0–5 vol. %). According to the nomenclature of Leake, (1978), the amphiboles are edenites, ferro-edenites, edenitic hornblendes, and ferro-edenitic hornblendes (Table 6).
|
The pressures calculated from mean rim compositions of these amphiboles with each of the calibrations given above range from 0.313 to 0.506 GPa (Table 7). These values are to be taken as maximum pressures for several reasons. (1) Not all of the amphibole rims directly coexist with quartz, meaning that they may not have been silica-buffered during growth. This results in increased incorporation of Al and thus in an overestimation of pressure. (2) The comparatively high modal abundance of biotite and amphibole possibly indicates slightly higher solidus temperatures than those used for calibration of the barometer. The assumption of a solidus temperature that is too low would also result in an overestimation of pressure. As discussed below, we now think the conditions were closer to 0.2 GPa.
|
Pressure constraints from petrogenetic grids for metapelitic rocks
Pressures can also be estimated from experimentally determined and calculated phase equilibria for metapelitic rocks. A very crude pressure estimate may be derived from the fact that in all cordierite-bearing samples andalusite is the stable aluminosilicate over a temperature range of 510–750°C. Sillimanite is only present as fibrolite or small needles at one locality (K-262, K-262a) and kyanite was never found. Taking into account the various versions of the Al2SiO5 triple point (Berman, 1988
; Holland & Powell, 1990; Hemingway et al., 1991; Pattison, 1992; Holdaway & Mukhopadhyay, 1993a) pressure is limited to below 0.4 GPa for contact metamorphism on Kos. This estimate reasonably coincides with that derived from Al-in-hbl barometry (see above). Samples K-352 and K-354 allow for estimation of the pressure prevailing during contact metamorphism, as they give textural evidence of phase equilibria that are well constrained in P–T space either from experiments or petrogenetic grids. The bulk compositions of these samples are best described by the FASH system or the KFASH system as they are very Fe rich, a fact that emerges from the low XMg values of cordierites (0.26–0.35, Table 5), from the presence of chloritoid, and from the absence of chlorite.
Both rocks contain cordierite, biotite, andalusite, ilmenite, staurolite, quartz, plagioclase, hercynite, and chloritoid in order of decreasing abundance. Sample K-354 additionally bears garnet. Remnants of the reaction ctd = hc + crd (Fig. 3e) can be clearly identified in both rocks, with chloritoid relics being only preserved in sample K-352. Staurolite coexists with these hercynite–chloritoid aggregates or may have started to grow a little later, as indicated by rare tiny hercynite inclusions in staurolite. Both staurolite and the hercynite–chloritoid aggregates are included in large cordierite porphyroblasts and never coexist with quartz or andalusite. No relics of the staurolite-forming reaction were found, but from most of the common petrogenetic grids (Spear & Cheney, 1989; Powell & Holland, 1990; Xu et al., 1994) it is clear that staurolite must have formed at the expense of chloritoid in some way. Comparing the described textures with a grid for metapelitic rocks (KFASH system,Xu et al., 1994) and assuming an isobaric temperature increase, as is typical for contact metamorphism, the following sequence of reactions is indicated (Fig. 4). At the lowest temperature, the reaction ctd + qtz = crd takes place. As the bulk system is comparatively poor in SiO2, chloritoid is left over and is further consumed at increasing temperature by the reaction ctd = hc + crd. Chloritoid grains in contact with andalusite take part in the reaction ctd + als 7equals; crd + st. This sequence explains the formation of hercynite, cordierite, and staurolite within a narrow temperature range and the lack of coexisting quartz or andalusite.
Pressure constraints on these reactions are tight (Fig. 4). The maximum pressure for the reaction ctd = hc + crd is 0.27 GPa at 575°C according to experiments by Grieve & Fawcett, (1974) on the stability of Fe-chloritoid below 1.0 GPa (PH2O = Ptotal, Ni–NiO buffer). The same reaction is terminated in the petrogenetic grid of Xu et al., (1994) at 0.15 GPa and 525°C by an invariant point, where the reaction ctd + als = crd + st continues to higher pressures. The reactions ctd + als = crd + st and ctd + qtz = crd have a maximum pressure of 0.23 GPa at 532°C (Xu et al., 1994). The latter also requires a minimum pressure of 0.15 GPa. The P–T field for these reactions is further limited by the minimum pressure of 0.15–0.25 GPa required for staurolite stability in the (K)FASH system (Dutrow & Holdaway, 1989; Holdaway & Mukhopadhyay, 1993b; Xu et al., 1994; Fig. 4) and a maximum pressure of 0.17–0.23 GPa given by cordierite stability (Spear & Cheney, 1989; Powell & Holland, 1990; Bucher & Frey, 1994; Xu et al., 1994; Fig. 4).
Therefore, depending on whether the Grieve & Fawcett, (1974) data or the KFASH grid of Xu et al., (1994) are used for the reaction ctd = crd + hc, pressures between 0.15 and 0.25 GPa must have prevailed in these rocks during intrusion of the pluton. An intermediate pressure of 0.20 GPa is thus selected for temperature calculation.
|
Calculation of Temperatures |
|---|
|
TOPABSTRACTIntroductionGeological Setting and SamplesAnalytical MethodsPetrography and Mineral...Calculation of PressuresCalculation of TemperaturesFluid Composition During Contact...DiscussionSummary and ConclusionsReferences |
|---|
Na-in-cordierite temperatures
The Na-in-cordierite thermometer of Mirwald, (1986)
is based on preliminary experimental results in the systems Mg-cordierite-1N NaOH and Mg-cordierite-1N NaOH-albite, carried out at temperatures of 650–850°C and pressures of 0.1–0.8 GPa. The results show a linear inverse relationship between Na contents of Mg-cordierite and temperature and suggest an insignificant influence of pressure (Fig. 1). Errors on the thermometric relationship are estimated to be ±30°C (Mirwald, 1986).
In Fig. 5, the analytical results for Na in cordierites of our study are presented in the form of histograms which show a normal distribution of measured Na contents for all samples. The considerable scatter within each sample is almost exclusively due to intra-grain variations of the Na contents (see section on petrography and mineral chemistry). Thus, the mean values of Na contents of cordierites of each sample were used for temperature calculations. Calculated temperatures are given in Table 8. For each sample, 1 errors were calculated; these are directly translatable to temperatures (0.001 cations per formula unit [c.p.f.u.] = 3°C) and accessible from Fig. 1. Calculated temperatures range from 762 to 508°C and are shown in Fig. 6 for each sample in relation to its distance from the pluton.
|
|
Garnet-cordierite temperatures and garnet-biotite temperatures
The Fe–Mg partitioning between garnet and cordierite and between garnet and biotite form the basis of several thermobarometric formulations. The thermometers of Perchuk & Lavrent'eva, (1983)
have been chosen for this study as they are based on experiments with garnet, cordierite and biotite compositions close to those in natural greenschist- to amphibolite-facies metapelites and inherently take Ca and Mn in garnet and Ti and Al in biotite into account.
Garnet–cordierite and garnet–biotite temperatures were calculated on the basis of mean values (Table 8). The outermost rim compositions (10–40 µm) of all phases were excluded as they may show slightly different and variable compositions. In addition, maximum temperatures were calculated using the garnet cores, which are slightly higher in XMg than the average compositions, and the minimum XMg values of biotites and cordierites (Table 5 and 8). All temperatures were calculated assuming a pressure of 0.20 GPa as deduced from constraints from petrogenetic grids (see section oncalculation of pressures and Table 7). Garnet–cordierite mean temperatures range between 574 and 636°C (maximum temperatures 593–686°C) and garnet–biotite mean temperatures between 604 and 670°C (maximum temperatures 618–705°C). No errors on the temperature values are given as the individual uncertainties on XMg, XFe, and XMn mean values are not independent and hence hamper a realistic calculation of error propagation. These X values were calculated from cordierite, garnet, and biotite formulae on the basis of a fixed oxygen number with all iron assumed as divalent. Calculating formulae with fixed oxygen and cation numbers and Fe2+/Fe3+ calculations based on charge balance has neither systematic nor significant influence on calculated temperatures and is moreover problematic as cation totals in cordierites are not constant (Schreyer, 1965). Both thermometers are slightly pressure dependent. In both cases, higher assumed pressures result in higher temperatures.
Temperature constraints from experiments and petrogenetic grids
Temperature estimations have been made using grids in the KFASH or FASH systems. As has already been argued in the section on calculation of pressures, this seems justified as in the Mg-bearing systems chlorite always takes part in the lower-temperature reactions but has never been observed in the studied rocks.
Concerning the expected temperatures, the samples may be divided into three groups, although this subdivision is a little arbitrary. Samples of the first group come from the immediate vicinity of the contact. There is only one cordierite-bearing sample representing this group (K-252). Samples of the second group (K-345 to K-151, Table 1) have distances from the intrusion ranging between 167 and 592 m. This region is characterized by numerous granitic, aplitic and lamprophyric dykes and monzonitic apophyses. In addition, as has been mentioned in the section on sample selection, morphology, the irregular shape of the pluton, and the dip of the metapelite sequences do not always allow for a clear correlation between distance and temperatures. Therefore, the temperatures in this region are expected to vary only little, and if so, probably not in any systematic relation to distance. The samples of the third group (K-269 to K-339) span a range in distances between 697 and 1258 m. Within this region, only a few dykes are encountered.
For sample K-252, representing group 1, a minimum temperature is given by the fact that samples at the same distance bear orthopyroxene (unfortunately no cordierite). The orthopyroxene-in temperature at 0.2 GPa is about 740°C (Bucher & Frey, 1994). A maximum temperature of 770°C is given by the absence of spl + qtz assemblages (Vielzeuf, 1983; Bucher & Frey, 1994). For samples of group 2, minimum temperatures of 640°C at 0.2 GPa are given by the fact that no ms + qtz or ms + crd assemblages are stable in these rocks (Xu et al., 1994). The maximum temperature is the opx-in temperature (about 740°C,Bucher & Frey, 1994), as orthopyroxene is absent both in the studied cordierite-bearing rocks and in any other samples at these distances. For samples K-262 and K-262a, the presence of sillimanite requires minimum temperatures of about 675°C (Bucher & Frey, 1994; Xu et al., 1994).
For samples K-352 and K-354 of group 3, there is a comparatively tight temperature bracket resulting from evidence for the reactions ctd + qtz equals; crd, ctd + als = crd + st, and ctd = crd + hc. As has been argued in the section on calculation of pressures, these reactions occur in a temperature range of 525–575°C at 0.2 GPa. In samples K-357 and K-339 of group 3, ms + qtz + crd assemblages are stable. This implies a maximum temperature of 600-640°C at 0.2 GPa and a minimum temperature of 520°C (Bucher & Frey, 1994; Xu et al., 1994) required by the presence of cordierite. The presence of significant amounts of CO2 and CH4 in possible metamorphic fluids would reduce all of the temperatures except that of the reaction crd – spl + qtz by about 20–40°C.
|
Fluid Composition During Contact Metamorphism |
|---|
|
TOPABSTRACTIntroductionGeological Setting and SamplesAnalytical MethodsPetrography and Mineral...Calculation of PressuresCalculation of TemperaturesFluid Composition During Contact...DiscussionSummary and ConclusionsReferences |
|---|
Fluid pressure and composition can have an influence on the position of mineral equilibria in P–T space. The thermometers and some of the grids used in this study refer to fluid-saturated systems with XH2O = 1. Incipient fluid saturation or deviation of fluid compositions from XH2O = 1 will have the effect of destabilizing hydrous phases at lower temperatures and pressures when compared with fluid-saturated systems with XH2O – 1 (Ohmoto & Kerrick, 1977
; Connolly & Cesare, 1993). The effect on Na contents of cordierite and thus on Na-in-cordierite temperatures has not been studied yet.
For the cordierite-bearing metapelites on Kos, lack of hydrothermal alteration features in the metapelites and the pluton, excellent preservation of cordierite and confinement of calc-silicate mineralogies to pelite-marble interfaces render a significant fluid input from the intruding pluton unlikely and argue in favour of lithology-controlled fluid production. The presence of H2O- and CO2-bearing fluids during contact metamorphism is likely from the formation of phyllosilicates and amphiboles in the metapelites and from the presence of calcite in adjacent metacarbonate rocks. Because of the intimate interlayering of both rock types, fluid composition during contact metamorphism varied on a small scale. This is consistent with varying contents of accessory graphite, calcite, pyrite, and margarite in the metapelites (Table 1). Comparatively high haematite contents in some ilmenites (Table 3) point to the presence of significant amounts of CO2 in some rocks (Ohmoto & Kerrick, 1977; Connolly & Cesare, 1993).
The composition of metamorphic fluids may be constrained by the amount of channel H2O and CO2 of cordierites (Johannes & Schreyer, 1981; Armbruster & Bloss, 1982). Experimentally derived and natural total channel gas contents (H2O + CO2) of cordierite show a positive correlation with pressure and a negative correlation with temperature and with XCO2 of the coexisting fluid phase (Schreyer & Yoder, 1964; Mirwald & Schreyer, 1977; Mirwald et al., 1979; Johannes &Schreyer, 1981; Armbruster & Bloss, 1982; Vry et al., 1990). XCO2 in cordierite channels essentially monitors XCO2 in the coexisting fluid phase (Johannes & Schreyer, 1981). Within the contact aureole on Kos, the extremely poikiloblastic nature of the cordierites impedes the determination of H2O and CO2 in cordierite channels by IR spectroscopy (Le Breton, 1989; Vry et al., 1990) and also by ion-probe, as extremely small solid and fluid inclusions are not visible and can thus not be avoided. Optical data provide the only source of information on cordierite volatile components (Armbruster & Bloss, 1980, 1982), although the poikiloblastic character of the cordierites makes determinations of 2Vx (optic angle bisected by the x-axis) extremely difficult. For Mg-cordierites devoid of channel H2O and CO2, 2Vx is approximately 87°. The angle decreases with increasing H2O contents and increases with increasing CO2 contents. With mixed H2O–CO2 fluids, these contrary effects impede calculation of fluid compositions from the optical data. Moreover, Na has the same effect on 2Vx of cordierite as H2O (Selkregg & Bloss, 1980; Armbruster & Irouschek, 1983; Armbruster, 1986).
At pressures of 0.15–0.25 GPa as deduced for contact metamorphism on Kos (see section on calculation of pressures) cordierite channels can potentially accommodate 1.5–1.9 wt % H2O (T = 600°C) if Pfl = PH2O and 0.7–1.2 wt % CO2 (T = 600°C) if Pfl = PCO2, corresponding to 2Vx = 59–53° and 2Vx = 92–97°, respectively (Armbruster & Bloss, 1982). 2Vx values in cordierites from Kos range from 73° to 94° (Table 9), suggesting that CO2 and H2O and/or Na are important channel filling species. If 2Vx values lower than 87° in Kos cordierites were overall due to Na incorporation, there should be a positive correlation between Na contents and 2Vx values. Figure 7d clearly shows that this is not the case. Therefore, although Na may have a minor influence, the comparatively large range in 2Vx values must be caused by varying XCO2 in the cordierites and thus in the metamorphic fluid(s). This conclusion is consistent with evidence from accessories and geological arguments (see above). Whether or not contact metamorphism occurred under fluid-saturated conditions cannot be deduced.
|
|
|
Discussion |
|---|
|
TOPABSTRACTIntroductionGeological Setting and SamplesAnalytical MethodsPetrography and Mineral...Calculation of PressuresCalculation of TemperaturesFluid Composition During Contact...DiscussionSummary and ConclusionsReferences |
|---|
Accuracy and precision of the Na-in-cordierite temperatures
In general, the accuracy of calculated temperatures for natural rocks is difficult to test as there is virtually no independent way to determine absolute and correct temperatures for metamorphism or plutonism. Temperature brackets emerging from experimental results and petrogenetic grids should be more reliable in this respect than thermometric formulations that consider only distinct exchange reactions with distinct closure temperatures.
The Na-in-cordierite temperatures for the contact metamorphic rocks from Kos were derived from the mean values of each sample (Fig. 5) by application of the thermometer calibration shown in Fig. 1. The obtained temperatures are very consistent with the constraints from experiments and petrogenetic grids. For sample K-252 of group 1, the Na-in-cordierite temperature is 762°C (Table 8), which is slightly above the minimum temperature required by the presence of orthopyroxene in adjacent samples (740°C). The Na-in-cordierite temperatures of group 2 samples range from 662 to 731°C (Table 8) and thus fall within the temperature bracket of 640–740°C deduced from petrogenetic grids. The Na-in-cordierite temperatures of 508–603°C (Table 8) calculated for group 3 samples are also consistent with those estimated from petrogenetic grids and experiments (520–630°C, see above). The only exception is sample K-269, with an Na-in-cordierite temperature of 719°C. Consequently, there is also a fairly good correlation between Na-in-cordierite temperatures and distance from the intrusion (Fig. 6). As has already been pointed out in previous sections, a perfect correlation between distance and temperatures is not to be expected.
Samples from the same localities but with different bulk chemistries generally give the same Na-in-cordierite temperatures within 1 error limits (Table 8, Fig. 2). There are four samples that record slightly higher temperatures than adjacent samples. Sample K-254 at 258 m distance from the intrusion gives higher temperatures (731 ± 18°C) than samples K-255 (690 ± 24°C) and K-256 (677 ± 24°C). The first and the last temperature values do not overlap within 1 error ranges. Sample K-346 at 500 m (722 ± 30°C) shows higher temperatures than sample K-143 (685 ± 27°C) at the same distance, but apart from the fact that these values coincide within 1 error limits the samples do not come from the same locality (Fig. 2) and hence the temperature difference may be `real'. The Na-in-cordierite temperatures for sample K-142 at 592 m (717 ± 27°C) are also slightly higher than for sample K-151 from the same locality (680 ± 27°C), but as in the latter case there is coincidence within 1 error limits. Whether or not the temperature of 719 ± 27°C for sample K-269 at a distance of 697 m from the intrusion is realistic cannot be clarified, as there is no other sample from this locality. Although a lower temperature is to be expected on the basis of distance, there may very well be a dyke or an apophysis beneath the locality that has induced higher temperatures. Therefore, only one sample (K-254) remains with Na-in-cordierite temperatures that are obviously too high.
It is possible that the Na-in-cordierite thermometer systematically overestimates temperatures as a result of the experimental calibration with NaOH. Na activity is probably lower in natural fluids and hence the Na contents of coexisting cordierites would also be lower at a given temperature. However, this effect only applies to fluid-present systems and must be small as the temperatures for group 3 on Kos cannot be much lower, the Na-in-cordierite temperatures for two of the samples being already at the lower stability limit of cordierite when using the NaOH-albite calibration.
The 1 errors on the Na-in-cordierite temperatures are fairly small. This reflects the fact that errors on Na-in-cordierite temperatures, expressed as standard mean deviations, reflect only the analytical error of Na determination in cordierite by microprobe (see section on analytical methods) and the heterogeneity of cordierites within one sample (Fig. 5). Although this and the lack of pressure dependence may be advantages of the Na-in-cordierite thermometer when compared with Fe–Mg exchange thermometers, it must be stressed that numerous cordierite analyses for one sample are necessary to obtain a representative mean value. The errors on Na-in-cordierite temperatures increase with decreasing temperature (Table 8, Figs 5 and 6). This is not due to enlarged analytical errors, as Na contents become higher with decreasing temperature (Fig. 1). The large errors reflect heterogeneities in the cordierite chemistry within the samples, probably because of only sluggish attainment of equilibrium at temperatures just above the minimum required for cordierite to be stable (520°C; see section on calculation of temperatures).
Comparison of calculated temperatures
For eight samples, garnet-cordierite and garnet-biotite temperatures (Perchuk & Lavrent'eva, 1983) could be calculated (Table 8). Garnet–cordierite mean temperatures range between 574 and 636°C and garnet–biotite mean temperatures between 604 and 670°C. These mean temperatures may include effects of partial retrograde equilibration. Hence, the calculated maximum temperatures (garnet–cordierite: 593–686°C; garnet–biotite: 618–705°C) are discussed further. Assuming an average error of ±50°C, these temperatures are mostly consistent with the temperature brackets deduced from petrogenetic grids and with Na-in-cordierite temperatures, although tending to lower values. For sample K-252 of group 1, garnet-cordierite and garnet-biotite temperatures (686°C and 691°C, respectively) lie just at the lower limit of the minimum temperature (740°C), given by the presence of orthopyroxene in adjacent samples. For samples of group 2, garnet-cordierite and garnet–biotite temperatures of 620–705°C are consistent with the temperature brackets emerging from petrogenetic grids and experiments (640–740°C). For sample K-354 of group 3 there is a good coincidence of all three calculated temperatures around 600°C within 1 error limits of the Na-in-cordierite temperatures (Table 8).
The garnet–cordierite and garnet–biotite temperatures may trend to lower values as a result of incipient equilibration, retrograde diffusion or lower blocking temperatures for the Fe–Mg exchange compared with the Na-in-cordierite thermometer. Retrograde equilibration effects should play no major role, as rim compositions were excluded and maximum temperatures were calculated. Incipient equilibration is not a common phenomenon in contact metamorphism. However, garnet-forming reactions in Fe-rich metapelitic rocks generally occur at fairly low temperatures during progressive contact metamorphism. The first garnet-in reaction can take place just above 500°C, which may be slightly lower than the first cordierite-in reaction (Bucher & Frey, 1994; Xu et al., 1994). Thus, garnets may have grown earlier than cordierites and failed to equilibrate with them later on. The very limited degree of zoning in large garnets and the inter-grain variation in biotite compositions may support this hypothesis. Low blocking temperatures for the Fe–Mg exchange between garnet, cordierite and biotite may also be a reason for the observed temperature discrepancies, because of the comparatively high diffusion velocities of Mg and Fe (see Introduction).
Relationship between Na contents of cordierite and chemical parameters
Reasons for the deviations in Na-in-cordierite temperatures between samples from equal distances may lie in the possible dependence of Na incorporation on chemical parameters. As set out in the Introduction, the thermometer has been calibrated with pure Mg-cordierite and pure albite (Fig. 1). In natural rocks, additional components such as Fe in cordierite and Ca in plagioclase may thus cause deviations from the experimentally determined thermometric relationship.
From Fig. 7a it is evident that there is no systematic relationship between XMg and Na contents of cordierites. These findings are consistent with the high-temperature crystal chemistry of Mg- and Fe-rich cordierites (Hochella, 1979). The incorporation of Fe in octahedral positions leads only to a flattening along the crystallographic c-axis but does not influence the channel size and thus Na incorporation. The results of structural refinements of Mg- and Fe-rich cordierites (Hochella, 1979) show that thermal expansion in both Fe- and Mg-cordierites does not affect the size of the channels. This, in turn, is in line with experimental results (Boberski & Schreyer, 1990) which indicate the amount of water that can be accommodated in the channels to be equal for Fe- and Mg-cordierites. Probably the same holds true for even smaller channel fillings such as Na or K.
Figure 7b shows the Na contents of cordierite related to the An contents of coexisting plagioclase. It is evident that there is no systematic relationship between the two parameters. Nevertheless, there may be an influence of An contents in plagioclase on Na contents in cordierite. Out of the four samples with higher Na-in-cordierite temperatures when compared with results from adjacent samples or expected temperatures (see section on calculation of temperatures), two samples (K-346 and K-269) display An contents of plagioclases around 40 mol %, which is distinctly higher than any value measured in plagioclases of other cordierite-bearing samples from Kos (Table 5). Therefore, Na incorporation in cordierite seems to be a function of temperature as long as aNa is buffered at or above some minimum level, expressed by Na contents in plagioclase. Below this critical level, there may be a different relationship between Na in cordierite, Na in plagioclase, and temperature.
In Fig. 7c, the Na contents of cordierite related to the Na contents of coexisting biotite are shown. A systematic relationship between Na contents of cordierite and biotite is not evident.
The influence of fluids and fluid composition on the Na contents of cordierite cannot be clearly constrained from our data (Fig. 7d), as neither the level of fluid saturation nor the exact fluid compositions are known. Crystal chemical arguments (Armbruster, 1986) and the results of hydration experiments (Jochum, 1986) indicate that Na channel contents of cordierite control the amount of water accommodated in a certain orientation in the channels (type II,Goldman et al., 1977; Vry et al., 1990), which may amount to two-thirds of the total water content in Na-saturated cordierites. Hence, a reverse dependence seems possible under water-saturated conditions. However, the occurrence of natural cordierites containing Na but lacking significant amounts of water argues against Na contents in cordierite being controlled by H2O under conditions of PH2O < Ptotal. CO2 plugs the cordierite channels as the large molecules arrange with their long axis perpendicular to the channels (Farrell & Newnham, 1967; Armbruster & Bloss, 1982) and impede dehydration (Jochum, 1986). Thus, incorporation of large amounts of CO2 into cordierite channels should also hinder Na entry. This phenomenon will, however, only play a major role with very CO2-rich fluids, as with composite fluids H2O is preferentially partitioned into cordierite (Johannes & Schreyer, 1981). Therefore, a major influence of fluid composition on Na incorporation in cordierite does not clearly emerge from the available experimental data. This is consistent with our observation that varying fluid compositions during contact metamorphism on Kos obviously do not affect Na incorporation in any systematic way. However, no constraints can be placed on the effect of fluid saturation.
In Fig. 7e the Na and Li contents of 15 cordierites are shown. They display a largely positive correlation in the sense that the four samples with high Na contents (K-339, K-357, K-354 and K-352) also have high Li contents (except K-357) whereas the rest of the samples have fairly low Na and Li contents. However, in neither of those groups can a clear positive correlation be observed, reflected by the fact that cordierites with approximately the same Na contents may show variable Li contents. For K-339 and K-357, Li values must be taken with caution because of the extremely fine-grained nature of the rocks (see section on petrography and mineral chemistry). Figure 7e suggests that Li incorporation in cordierites may also be temperature dependent. To eliminate temperature effects when discussing the relationship of Na contents to other chemical parameters, samples of equal temperatures (equal distances from the intrusion) need to be compared. These are samples K-254, K-255 and K-256 at 258 m as well as K-352 and K-354 at 712 m. In the latter case there is a positive correlation, in the former not, leaving the problem unsolved. The possible relation between Li and Na incorporation in the cordierites of Kos will be further discussed in the next section.
Possible mechanisms of Na and Li incorporation in cordierite
Possible mechanisms of Na and Li incorporation in cordierites have been proposed and discussed by various workers. Wolfsdorff & Schreyer, (1992) proposed and confirmed experimentally the following substitutions in Li-free cordierites (ch, channel site; VI, octahedral site; IV, tetrahedral site): (1) Na[ch] + Al[IV] = Si[IV], (2) Na[ch] + Mg[VI] = Al[IV], (3) 2Na[ch] = Mg[VI] and (4) Na[ch] + Si [IV] = Mg[VI] + Al[IV]. Kirchner et al., (1984) showed the substitutions (5) Li[ch] + Li[VI] – Mg, (6) Na[ch] + Li[VI] = Mg and (7) Li[VI] + Si[IV] = Mg[VI] + Al[IV] experimentally. Substitutions (6) and (7) were also confirmed on natural cordierites [Gordillo et al., (1985) and Cerny et al., (1997) and references therein], where also Fe, Mn and traces of Ti, Cr and Zn can occupy octahedral sites. In addition to substitutions (1)–(7), Be on tetrahedral positions normally occupied by Al can enhance Na incorporation in the channels [Gordillo et al., (1985) and references therein].
Table 3 clearly reveals that Be is only present in insignificant amounts in the cordierites from Kos and thus plays no role in Na and Li incorporation. Substitutions (1), (2), (3), (5) and (6) are rendered unlikely by the fact that plots of the respective substitution parameters (e.g. Fig. 8a–c) show no clear correlations. Additionally, Fig. 7e reveals that substitution (6)-if active at all-cannot be the only mechanism of Na incorporation as there is always more Na than necessary to charge-balance Li. Similar observations have also been made by Schreyer et al., (1979), Armbruster & Irouschek, (1983), Gordillo et al., (1985), Grew et al., (1988) and Cerny et al., (1997). Substitutions (2), (3), (5) and (6) do not involve Si. The consistently high Si contents (>5.000 c.p.f.u.) and low Al contents (<4.000 c.p.f.u.), however, point to a substitution mechanism involving Si (Table 2). Substitutions (4) and (7) involve Si and show a rather good negative correlation and a very vague tendency of low-temperature samples towards the substitution of Na, Li and Si for Mg (Fe, Mn) and Al (Fig. 8d and e). Fig. 8d reveals that the cordierites contain more Na than needed to balance substitution (4) in the Li-bearing cordierites. The best correlation is clearly displayed in Fig. 8e; this result suggests that Li is introduced into the cordierites by substitution (7) [Li[VI] + Si[IV] = (Mg, Fe, Mn)[VI] + Al[IV]]. This suggestion is supported by the almost perfect cation totals when calculating the cordierite formula without Na (Table 2). As a consequence, Na should not be charge-balanced by lattice oxygens but be fixed in the channels by either volatiles or by very weak bonds to channel oxygens.
|
It must be clearly stated, however, that the poikiloblastic and partly fine-grained nature (especially of the crucial low-temperature samples) makes the samples unsuited for a systematic study of substitution mechanisms in cordierites. Moreover, the intra-grain variations of Li and Na (see section on petrography and mineral chemistry) suggest that charges are balanced on a very small scale. In this study, however, mean values of more than 100 Na analyses (for the sake of accuracy of Na-in-cordierite temperatures) but significantly fewer Li analyses were used to calculate average formulae. Further studies measuring Li and Na simultaneously on the same spot have to be conducted to clarify this point.
|
Summary and Conclusions |
|---|
|
TOPABSTRACTIntroductionGeological Setting and SamplesAnalytical MethodsPetrography and Mineral...Calculation of PressuresCalculation of TemperaturesFluid Composition During Contact...DiscussionSummary and ConclusionsReferences |
|---|
Petrologic studies on contact metamorphosed cordierite-bearing pelitic rocks on the island of Kos (Greece, Eastern Aegean Sea) allow for several conclusions on P–T conditions and fluid compositions during contact metamorphism and on the applicability of the Na-in-cordierite thermometer of Mirwald, (1986)
:- The intrusion of the Dicheo quartz monzonite on the island of Kos occurred at a depth corresponding to 0.15–0.25 GPa. This is deduced from calculated and experimentally determined phase equilibria for metapelites. Al contents of amphiboles from the intrusion give maximum pressures of 0.31–0.51 GPa.
- Temperatures at the time of intrusion declined from 762°C at 46 m from the contact to 508°C at a distance of 1258 m. These temperatures are calculated with the Na-in-cordierite thermometer.
- Na-in-cordierite temperatures agree with temperature constraints from petrogenetic grids for metapelitic rocks and thus represent peak metamorphic conditions. Na-in-cordierite temperatures of samples with equal distances from the pluton coincide within error limits except for one sample pair. Errors for the Na-in-cordierite temperatures in the high-temperature range are between ±18 and ±30°C, those in the low-temperature range between ±36 and ±72°C, as a result of insufficient equilibration.
- The Na incorporation in cordierite shows no systematic variations with the XMg values of cordierite, the An contents of coexisting plagioclase (for An below 40 mol %), the Na contents of coexisting biotite, and fluid composition. The thermometric relationship should thus also be valid in systems containing additional components such as Fe, Ca, and CO2.
- The garnet–cordierite and the garnet-biotite thermometers of Perchuk & Lavrent'eva, (1983) give temperatures that are mostly consistent with the Na-in-cordierite temperatures and with constraints from petrogenetic grids but tend to underestimate temperatures above 700°C.
- The cordierites from Kos have either high Na and significant Li contents or display low levels of both elements. Li is probably introduced into cordierite by a coupled substitution involving tetrahedral and octahedral sites, whereas Na may not be charge-balanced by lattice oxygen.
- In the temperature range of 508–762°C, the experimentally determined pressure-independent relation between Na content of cordierite and temperature is a potentially useful geothermometer for cordierite- and plagioclase-bearing metapelitic rocks as long as aNa is buffered at a minimum level by sodic plagioclase. Further experiments have to clarify the mechanism as well as the temperature and fluid dependence of Na and Li incorporation in cordierite. Simultaneous in situ analysis of Na and Li must reveal the extent of intra-grain charge balance.
|
Acknowledgements |
|---|
The authors would like to thank Udo Geilenkirchen and Ilona Salzmann for preparing the polished thin sections, Mario Koch for help with the microprobe work, and Hans-Peter Meyer for maintaining microprobe facilities and providing formula calculation programs. M. Holdaway, W. Schreyer and D. Carrington are thanked for critical and constructive reviews that helped to significantly improve the manuscript. It is stated that the conclusions drawn in this paper are those of the authors and not necessarily those of the reviewers. We thank W. Schreyer for providing cordierite TS25 as reference material for Li and Be analyses. Financial support within the special research programme of the Deutsche Forschungsgemeinschaft ‘Orogene Prozesse-ihre Quantifizierung und Simulation am Beispiel der Varisziden’ is gratefully acknowledged.
* Corresponding author. Fax: 0049-6221-544805. e-mail: akalt{at}classic.min.uni-heidelberg.de
|
References |
|---|
|
TOPABSTRACTIntroductionGeological Setting and SamplesAnalytical MethodsPetrography and Mineral...Calculation of PressuresCalculation of TemperaturesFluid Composition During Contact...DiscussionSummary and ConclusionsReferences |
|---|
Altherr R., Keller J., Kott K., et al. Der jungtertiäre Monzonit von Kos und sein Kontakthof (Ägäis, Griechenland). Bulletin de la Société Géologique de France (1976) 7(XVIII):403–412.
Altherr R., Kreuzer H., Wendt I., Lenz H., Wagner G. A., Keller J., Harre W., Höhndorf A., et al. A late Oligocene/early Miocene high temperature belt in the Attic–Cycladic Crystalline Complex (SE Pelagonian, Greece). Neues Jahrbuch für Geologie (1982) E23:97–164.
Altherr R., Henjes-Kunst F., Matthews A., Friedrichsen H., Hansen B. T., et al. O–Sr isotopic variations in Miocene granitoids from the Aegean: evidence for an origin by combined assimilation and fractional crystallisation. Contributions to Mineralogy and Petrology (1988) 100:528–541.[Web of Science]
Altherr R., Kreuzer H., Lenz H., Wendt I., Harre W., Dürr S., et al. Further evidence for a late Cretaceous low-pressure/high-temperature terrane in the Cyclades, Greece. Chemie der Erde (1994) 54:319–328.[Web of Science]
Armbruster T. Ar, N2 and CO2 in the structural cavities of cordierite, an optical and X-ray single crystal study. Physics and Chemistry of Minerals (1985) 12:243–245.
Armbruster T. Role of Na in the structure of low-cordierite: a single-crystal X-ray study. American Mineralogist (1986) 71:746–757.[Abstract]
Armbruster T., Bloss F. D. Channel CO2 in cordierites. Nature (1980) 286:140–141.
Armbruster T., Bloss F. D. Mg-cordierite: Si/Al ordering, optical properties, and distortion. Contributions to Mineralogy and Petrology (1981) 77:332–336.[Web of Science]
Armbruster T., Bloss F. D. Orientation and effects of channel H2O and CO2 in cordierite. American Mineralogist (1982) 67:284–291.[Abstract]
Armbruster T., Irouschek A. Cordierites from the Lepontine Alps: Na + Be – Al. Substitution, gas content, cell parameters and optics. Contributions to Mineralogy and Petrology (1983) 82:389–396.[Web of Science]
Armbruster T., Schreyer W., Hoefs J., et al. Very high CO2 cordierite from Norwegian Lapland: mineralogy, petrology, and carbon isotopes. Contributions to Mineralogy and Petrology (1982) 81:262–267.[Web of Science]
Berman R. G. Internally-consistent thermodynamic data for minerals in the system Na2O–K2O–CaO–MgO–FeO–Fe2O3–Al2O3–SiO2–TiO2–H2O–CO2. Journal of Petrology (1988) 29:445–522.
Bhattacharya A., Mazumdar A. C., Sen S. K., et al. Fe-Mg mixing in cordierite: constraints from natural data and implications for cordierite–garnet geothermometry in granulites. American Mineralogist (1988) 73:338–344.[Abstract]
Boberski C., Schreyer W. Synthesis and water contents of Fe2+-bearing cordierites. European Journal of Mineralogy (1990) 2:565–584.
Bucher K., Frey M. Petrogenesis of Metamorphic Rocks (1994) Heidelberg: Springer.
Carey J. W. A thermodynamic formulation of hydrous cordierite. Contributions to Mineralogy and Petrology (1995) 119:155–165.[Web of Science]
Cerny P., Chapman R., Schreyer W., Ottolini L., Bottazzi P., McCammon C. A., et al. Lithium in sekaninaite from the type locality, Dolni Bory, Czech Republic. Canadian Mineralogist (1997) 35:167–173.[Web of Science]
Connolly J. A. D., Cesare B. C–O–H–S fluid composition and oxygen fugacity in graphitic metapelites. Journal of Metamorphic Geology (1993) 11:379–388.[Web of Science]
Dürr St, Jacobshagen V. Ostägäische Inseln. In: Geologie von Griechenland (1986) Berlin: Gebrüder Borntraeger. 169–187.
Dutrow B. L., Holdaway M. J. Experimental determination of the upper thermal stability of Fe-staurolite + quartz at medium pressures. Journal of Petrology (1989) 30:229–248.
Elphick S. C., Ganguly J., Loomis T. P., et al. Experimental determination of cation diffusivities in aluminosilicate garnets. I. Experimental methods and interdiffusion data. Contributions to Mineralogy and Petrology (1985) 90:36–44.[Web of Science]
Farrell F., Newnham R. E. Electronic and vibrational absorption spectra in cordierite. American Mineralogist (1967) 52:380–388.[Web of Science]
Florence F. P., Spear F. S. Effects of diffusional modification of garnet growth zoning on P–T path calculations. Contributions to Mineralogy and Petrology (1991) 107:487–500.[Web of Science]
Gibbs G. V. The polymorphism of cordierite I: the crystal structure of low cordierite. American Mineralogist (1966) 51:1068–1087.[Web of Science]
Goldman D. S., Rossman G. R., Dollase W. A., et al. Channel constituents in cordierite. American Mineralogist (1977) 62:1144–1157.[Abstract]
Gordillo C. E., Schreyer W., Werding G., Abraham K., et al. Lithium in NaBe-cordierites from El Penón, Sierra de Córdoba, Argentina. Contributions to Mineralogy and Petrology (1985) 90:93–101.[Web of Science]
Gralla P. Das Präneogen der Insel Kos (Dodekanes, Griechenland). (1982) University of Braunschweig. Ph.D. Thesis.
Grew E. S., Chernosky J. V., Werding G., Abraham K., Marquez N., Hinthorne J. R., et al. Chemistry of kornerupine and associated minerals, a wet chemical, ion microprobe, and X-ray study emphasizing Li, Be, B and F contents. Journal of Petrology (1988) 31:1025–1070.
Grieve R. A. F., Fawcett J. J. The stability of chloritoid below 10 kb PH2O. Journal of Petrology (1974) 15:113–139.
Hammarstrom J. M., Zen E. Aluminium in hornblende: an empirical igneous geobarometer. American Mineralogist (1986) 71:1297–1313.[Abstract]
Hemingway B. S., Robie R. A., Evans H. T. Jr, Kerrick D. M., et al. Heat capacities and entropies of sillimanite, fibrolite, andalusite, kyanite, and quartz and the Al2SiO5 phase diagram. American Mineralogist (1991) 76:1597–1613.[Abstract]
Henjes-Kunst F., Altherr R., Kreuzer H., Hansen B. T., et al. Disturbed U–Th–Pb systematics of young zircons and uranothorites: the case of the Miocene Aegean granitoids (Greece). Chemical Geology (1988) 73:125–145.[Web of Science]
Hensen B. J. Theoretical phase relations involving cordierite and garnet in the system MgO–FeO–Al2O3–SiO2. Contributions to Mineralogy and Petrology (1971) 33:191–214.[Web of Science]
Hochella M. F. Jr. High-temperature crystal chemistry of hydrous Mg- and Fe-cordierites. American Mineralogist (1979) 64:337–351.[Abstract]
Holdaway M. J., Lee S. M. Fe–Mg cordierite stability in high-grade pelitic rocks based on experimental, theoretical, and natural observations. Contributions to Mineralogy and Petrology (1977) 63:175–198.[Web of Science]
Holdaway M. J., Mukhopadhyay B. A reevaluation of the stability relations of andalusite: thermochemical data and phase diagram for the aluminum silicates. American Mineralogist (1993a) 78:298–315.[Abstract]
Holdaway M. J., Mukhopadhyay B. Geothermobarometry in pelitic schists: a rapidly evolving field. American Mineralogist (1993b) 78:681–693.[Abstract]
Holland T. J. B., Powell R. An enlarged and updated internally consistent thermodynamic data set with uncertainties and correlations: the system K2O–Na2O–CaO–MgO–MnO–FeO–Fe2O3–Al2O3–TiO2–SiO2–C–H2–O2. Journal of Metamorphic Geology (1990) 8:89–124.[Web of Science]
Hollister L. S., Grissom G. C., Peters E. K., Stowell H. H., Sisson V. B., et al. Confirmation of the empirical correlation of Al in hornblende with pressure of solidification of calc-alkaline plutons. American Mineralogist (1987) 72:231–239.[Abstract]
Jacobshagen V. Introduction. In: Geologie von Griechenland—Jacobshagen V., ed. (1986) Berlin: Gebrüder Borntraeger. 1–9.
Jiang J., Lasaga A. C. The effect of post-growth thermal events on growth-zoned garnets: implications for metamorphic P–T history calculations. Contributions to Mineralogy and Petrology (1990) 105:454–459.[Web of Science]
Jochum C. Experimentelle Untersuchung zum Wassereinbau und zur Kinetik der Hydratation und Dehydratation von synthetischen und natürlichen Cordieriten. (1986) University of Bochum. 235. Ph.D. Thesis.
Johannes W., Schreyer W. Experimental introduction of CO2 and H2O into Mg-cordierite. American Journal of Science (1981) 281:299–317.
Johnson M. C., Rutherford M. J. Experimental calibration of the aluminium-in-hornblende geobarometer with application to Long Valley caldera (California) volcanic rocks. Geology (1989) 17:837–841.
Keller J. Origin of rhyolites by anatectic melting of granitic crustal rocks. The example of rhyolitic pumice from the island of Kos (Aegean Sea). Bulletin Volcanologique (1969) 33:942–959.
Kirchner D., Mirwald P. W., Schreyer W., et al. Experimenteller Li-Einbau in Mg-Cordierit. Fortschritte der Mineralogie (1984) 62(Beiheft 1):119–120.[Web of Science]
Kretz R. Symbols for rock-forming minerals. American Mineralogist (1983) 68:277–279.[Abstract]
Leake B. E. Nomenclature of amphiboles. Canadian Mineralogist (1978) 16:501–520.
Le Breton N. Infrared investigation of CO2-bearing cordierites. Contributions to Mineralogy and Petrology (1989) 103:387–396.[Web of Science]
Lonker S. W. The P–T–X relations of the cordierite–garnet–sillimanite–quartz equilibrium. American Journal of Science (1981) 281:1056–1090.
Maresch W. V., Blümel P., Schreyer W., et al. A search for variations in the structural state of cordierite in contact-metamorphosed pelites. In: Equilibrium and Kinetics in Contact Metamorphism—Voll G., Töpel J., Pattison D. R. M., Seifert E., et al, eds. (1991) Berlin: Springer. 297–314.
Martignole J., Sisi J.-C. Cordierite–garnet–H2O equilibrium: a geological thermometer, barometer and water fugacity indicator. Contributions to Mineralogy and Petrology (1981) 77:38–46.[Web of Science]
Mirwald P. W. Ist Cordierit ein Geothermometer. Fortschritte der Mineralogie (1986) 64(Beiheft 1):119.
Mirwald P. W., Schreyer W. Die stabile und metastabile Abbaureaktion von Mg-Cordierit in Talk, Disthen und Quarz und ihre Abhängigkeit vom Gleichgewichtswassergehalt des Cordierits. Fortschritte der Mineralogie (1977) 55:95–97.
Mirwald P. W., Maresch W. V., Schreyer W., et al. Der Wassergehalt von Mg-Cordierit zwischen 500° und 800°C sowie 0.5 und 11 kbar. Fortschritte der Mineralogie (1979) 57(Beiheft 1):101–102.
Mukhopadhyay B., Holdaway M. J. Cordierite–garnet–sillimanite–quartz equilibrium: I. New experimental calibration in the system FeO–Al2O3–H2O and certain P–T–XH2O relations. Contributions to Mineralogy and Petrology (1994) 116:462–472.[Web of Science]
Ohmoto H., Kerrick D. M. Devolatilization equilibria in graphitic schists. American Journal of Science (1977) 277:1013–1044.
Pattison D. R. M. Stability of andalusite and sillimanite and the Al2SiO5 triple point: constraints from the Ballachulish aureole, Scotland. Journal of Geology (1992) 100:423–446.[Web of Science]
Perchuk L. L., Lavrent'eva I. V. Experimental investigation of exchange equilibria in the system cordierite–garnet–biotite. In: Kinetics and Equilibrium in Mineral Reactions—Saxena S. K., ed. (1983) New York: Springer. 199–239.
Powell R., Holland J. T. B. Calculated mineral equilibria in the pelite system KFMASH (K2O–FeO–MgO–Al2O3–SiO2-H2O). American Mineralogist (1990) 75:367–380.[Abstract]
Richardson S. W. Staurolite stability in a part of the system Fe–Al–Si–O–H. Journal of Petrology (1968) 9:467–488.
Rocholl A., Simon K., Jochum K. P., Bruhn F., Gehann R., Kramar U., Lücke W., Molzahn M., Pernicka E., Seufert M., Spettel B., Summeier J., et al. Chemical characterisation of NIST silicate glasses certified reference material SRM 610 by ICP-MS, TIMS, LIMS, SSMS, INAA, AAS and PIXE. Geostandards Newsletter (1997) 21:101–114.[Web of Science]
Santosh M., Jackson D. H., Harris N. B. W. The significance of channel and fluid-inclusion CO2 in cordierite: evidence from carbon isotopes. Journal of Petrology (1993) 34:233–258.
Schmidt M. W. Amphibole composition in tonalite as a function of pressure: an experimental calibration of the Al-in-hornblende barometer. Contributions to Mineralogy and Petrology (1992) 110:304–310.[Web of Science]
Schreyer W. Synthetische und natürliche Cordierite II. Neues Jahrbuch für Mineralogie, Abhandlungen (1965) 103:35–79.
Schreyer W., Yoder H. S. The system Mg-cordierite and related rocks. Neues Jahrbuch für Mineralogie, Abhandlungen (1964) 101:271–341.
Schreyer W., Gordillo C. E., Werding G., et al. A new sodian–beryllian cordierite from Soto, Argentina, and the relationship between distortion index, Be content and state of hydration. Contributions to Mineralogy and Petrology (1979) 70:421–428.[Web of Science]
Schreyer W., Maresch W. V., Daniels P., Wolfsdorff P., et al. Potassic cordierites: characteristic minerals for high-temperature, very low-pressure environments. Contributions to Mineralogy and Petrology (1990) 105:162–172.[Web of Science]
Schreyer W., Bernhardt H.-J., Medenbach O. Ferrogedrite, siderophyllite, septechamosite, andalusite and chloritoid as alteration products of sekaninaite (ferrocordierite) from the Dolni Bory pegmatite, Moravia. Russian Geology and Geophysics (1993) 34:125–131.
Seidel E., Schliestedt M., Kreuzer H., Harre W., et al. Metamorphic rocks of late Jurassic age as components of the ophiolitic mélange on Gavdos and Crete (Greece). Geologisches Jahrbuch (1977) B28:3–21.
Seidel E., Kreuzer H., Harre W., et al. A late Oligocene/early Miocene high pressure belt in the external Hellenides. Geologisches Jahrbuch (1982) E23:165–206.
Selkregg K. R., Bloss F. D. Cordierites: compositional controls of 
, cell parameters, and optical properties. American Mineralogist (1980) 65:522–533.[Abstract]
Spear F. S. On the interpretation of peak metamorphic temperatures in the light of garnet diffusion during cooling. Journal of Metamorphic Geology (1991) 9:379–388.[Web of Science]
Spear F. S., Cheney J. T. A petrogenetic grid for pelitic schists in the system SiO2–Al2O3–FeO–MgO–K2O–H2O. Contributions to Mineralogy and Petrology (1989) 101:149–164.[Web of Science]
Streckeisen A. To each plutonic rock its proper name. Earth-Science Reviews (1976) 12:1–34.
Venkatesh V. Twinning in cordierite. American Mineralogist (1954) 39:636–646.[Web of Science]
Vielzeuf D. The spinel and quartz associations in high-grade xenoliths from Tallante (S.E. Spain) and their potential use in geothermometry and barometry. Contributions to Mineralogy and Petrology (1983) 82:301–311.[Web of Science]
Vry J., Brown P. E., Valley J. W., et al. Cordierite volatile content and the role of CO2 in high-grade metamorphism. American Mineralogist (1990) 75:71–88.[Abstract]
Weisbrod A. Cordierite–garnet equilibrium in the system Fe–Mn–Al–Si–O–H. Carnegie Institution of Washington, Yearbook (1973) 72:515–523.
Wolfsdorff P., Schreyer W. Synthesis of sodian cordierites in the system Na2O–MgO–Al2O3–SiO2. Neues Jahrbuch für Mineralogie, Monatshefte (1992) 2:80–96.
Xu G., Will T. M., Powell R. A calculated petrogenetic grid for the system K2O–FeO–MgO–Al2O3–SiO2-H2O with particular reference to contact-metamorphosed pelites. Journal of Metamorphic Geology (1994) 12:99–119.[Web of Science]
Zeck H. P. Transformation trillings in cordierite. Journal of Petrology (1972) 13:367–380.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  D. J. J. van Hinsbergen and F. Boekhout Neogene brittle detachment faulting on Kos (E Greece): implications for a southern break-away fault of the Menderes metamorphic core complex (western Turkey) Geological Society, London, Special Publications, January 1, 2009; 311(1): 311 - 320. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Barbey, C. Marignac, J. M. Montel, J. Macaudiere, D. Gasquet, and J. Jabbori Cordierite Growth Textures and the Conditions of Genesis and Emplacement of Crustal Granitic Magmas: the Velay Granite Complex (Massif Central, France) J. Petrology, September 1, 1999; 40(9): 1425 - 1441. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Kalt, A. Berger, and P. Blumel Metamorphic Evolution of Cordierite-Bearing Migmatites from the Bayerische Wald (Variscan Belt, Germany) J. Petrology, April 1, 1999; 40(4): 601 - 627. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||