| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology Volume 42 Number 3 Pages 529-554 2001
© Oxford University Press 2001
Metamorphic Evolution of Garnet–Epidote–Biotite Gneiss from the Moine Supergroup, Scotland, and Geotectonic Implications
1MINERALOGISCHES INSTITUT DER UNIVERSITÄT WÜRZBURG, AM HUBLAND, D-97074 WÜRZBURG, GERMANY
2BRITISH ANTARCTIC SURVEY, c/o NERC ISOTOPE GEOSCIENCES LABORATORY, KINGSLEY DUNHAM CENTRE, KEYWORTH, NOTTINGHAM NG12 5GG, UK
Received October 26, 1999; Revised typescript accepted June 16, 2000
 |
ABSTRACT |
|---|
TOP
ABSTRACT
INTRODUCTIONMetapelitic gneisses from the Glenfinnan Group of the Moine Supergroup, Scotland, contain sparse large and numerous small garnets, associated with complex zoned epidote and plagioclase in a biotite matrix. The large garnets show four zones (AI–AIV), whereas the small garnets show three or fewer zones, indicating successive garnet nucleation with increasing nucleation densities. Garnet zones AI and AIV grew under static conditions, whereas the formation of AII and AIII was accompanied by deformation. Garnet zones AI and AII were formed in the assemblage (all + biotite + epidote + plagioclase + quartz + fluid + apatite) garnet + chlorite + muscovite ± ilmenite ± sphene ± magnetite; zone AIII in the assemblage garnet + muscovite + sphene ± magnetite; and zone AIV in the assemblage garnet + sphene ± ilmenite. The chemical zonation and microstructures of garnet A indicate two important discontinuities; one at the transition between garnet zones AI and AII, and a second between zones AII and AIII, which correlate with complex zonation shown by epidote and plagioclase. These discontinuities may result from polymetamorphic garnet growth during different orogenic cycles affecting the Moine Supergroup. Geothermobarometric calculations and Gibbs method modelling provide evidence that garnet zone AI grew rapidly during heating from about 550 to 560°C at pressures of about 4–6 kbar. In contrast, the formation of zone AII was accompanied by nearly isothermal compression from 6 to 8·5 kbar (560 575°C), indicating crustal stacking. After a certain period of cooling, garnet zone AIII grew during renewed heating at P–T conditions of about 640°C and pressures between 5 and 9 kbar. Growth of garnet AIV was accompanied by further temperature rise, reaching maximum conditions of about 670°C at 5 kbar.
KEY WORDS: epidote; garnet; Gibbs method; Moine Supergroup; P–T path
|
INTRODUCTION |
|---|
Although numerous models exist for the structural evolution of the Moine Supergroup, less is known about its metamorphic history. One reason is the scarcity within the Moine Supergroup of aluminosilicate-bearing metapelites, which can be used to unravel the P–T history in detail. Garnet-bearing schists and gneisses, which are designated in the Moine as metapelites, mostly contain garnet, biotite, white mica, quartz and plagioclase. Locally, the Moine metasediments contain lenses of coarse garnet–biotite–epidote–plagioclase gneiss, which may also contain hornblende. Such rocks can also be very useful monitors of the metamorphic P–T history, as outlined by Menard & Spear (1993)
, because of the preservation of zonation in garnet, plagioclase and epidote, and the presence of associated minerals enclosed in different zones of these minerals. In addition, the textures of these rocks, in combination with petrological data, provide valuable information about their deformation–crystallization history. Successful reconstruction of the P–T path from these zoned minerals is possible only when the reaction history of the rock is interpreted correctly (e.g. Spear & Selverstone, 1983; Selverstone & Spear, 1985; Menard & Spear, 1993).
In this paper we present a comprehensive petrological study of a garnet–epidote–biotite gneiss from the Glenfinnan Group of the Moine Supergroup, and describe mineral composition and zonation patterns and their relationship to the complex deformation–crystallization history. P–T conditions are calculated using conventional thermobarometry and internally consistent datasets (Thermocalc: Powell & Holland, 1988; Holland & Powell, 1990). Additional P–T path constraints are obtained from compositional zoning of garnet, plagioclase and epidote employing the program GIBBS (Spear & Menard, 1989; Spear et al., 1991). Finally, garnet growth modelling is carried out to explain the complex garnet zonation patterns found in the Moine samples. The data presented in this paper set new petrological constraints for the interpretation of the tectono-metamorphic history of the Moine Supergroup.
|
GEOLOGICAL SETTING |
|---|
The Moine Supergroup is situated between the Great Glen Fault in the SE and the Moine Thrust in the NW (Fig. 1), and has been subdivided into three lithostratigraphic units: from west to east, the Morar, Glenfinnan and Loch Eil groups (Johnstone et al., 1969
; Roberts et al., 1987; Holdsworth et al., 1994; Soper et al., 1998). The dominantly psammitic Loch Eil Group stratigraphically overlies interbanded pelites and psammites of the Glenfinnan Group. The Glenfinnan and Loch Eil groups structurally overlie the dominantly psammitic Morar Group, from which they are separated by the ductile Sgurr Beag thrust, assumed to be formed during the Caledonian orogeny (Tanner et al., 1970). Together, the Glenfinnan and Loch Eil groups form the Sgurr Beag nappe. The boundary between the Glenfinnan and the Loch Eil groups is broadly coincident with a rapid change in the level of strain, known as the Loch Quoich Line (Clifford, 1957; Roberts & Harris, 1983). To the east of the Loch Quoich Line, in the ‘flat belt’ (Fig. 1), the dominant Loch Eil Group psammites are characterized by flat-lying foliation and recumbent isoclinal folds, formed during the D2 event (Roberts & Harris, 1983). To the west, in the ‘steep belt’, Glenfinnan Group rocks and Loch Eil Group outliers are deformed by tight, upright, steeply plunging folds during D3. All of the earlier D2 structures preserved in the ‘flat belt’ can also be identified in the ‘steep belt’, indicating a common tectono-metamorphic history. According to Roberts & Harris (1983), D2 in both belts is pre-dated by rare isoclinal folds, and an early migmatitic foliation (D1). These D1 structures are restricted to sparse intrafolial F1 isoclines, and a bedding-parallel S1 fabric.
|
A discontinuous series of highly deformed and metamorphosed granite bodies, the West Highland Granite Gneiss of Johnstone (1975), lies close to the boundary between the Glenfinnan and Loch Eil groups. The protoliths of these orthogneisses are thought to have been emplaced during the D1 event affecting the Moine country rocks. They are typical S-type granites, and were derived, at least in part, by partial melting of Moine metasediments (Barr et al., 1985). U–Pb dating of zircons from the Ardgour granite gneiss indicates emplacement of the gneiss protoliths at 873 ± 7 Ma (Friend et al., 1997). In places, the Glenfinnan and Loch Eil group metasediments contain deformed metabasic bodies, which are in part spatially associated with the West Highland Granite Gneiss. The emplacement of early, gabbroic metabasites has been dated at 873 ± 6 Ma (Millar, 1999). A subsequent suite of metabasite dykes has a tholeiitic composition and shows a mid-ocean ridge basalt (MORB) signature. They are interpreted to be the result of a continental rifting event at 
870 Ma, and may have contributed to the heat source for the generation of the West Highland S-type granites (Millar, 1999). This model conflicts with the interpretation of Barr et al. (1985) and Friend et al. (1997), in which the granite gneiss was emplaced during compressional deformation. The metabasites were isoclinally folded together with the Ardgour gneiss, during D2. The timing of this major deformation is not well constrained. However, recent P–T–t data of Vance et al. (1998) from the Morar Group point to crustal stacking between 820 and 790 Ma and, therefore, indicate plate convergence during the Neoproterozoic. In marked contrast to this view, some workers believe that extensional tectonics were active in Moine and Dalradian rocks throughout the Neoproterozoic (e.g. Soper et al., 1998). At the moment, because of the lack of P–T and geochronological data from the Loch Eil and Glenfinnan groups, it is not clear if these groups underwent a similar metamorphic evolution to the Morar Group. The petrological data presented here will help to solve some of the unresolved questions.
|
SAMPLE LOCATIONS AND FIELD RELATIONSHIPS |
|---|
Layers of unusual garnetiferous metapelitic gneisses and schists occur within the Moine metasediments close to the contact of the Glenfinnan and Loch Eil groups (Fig. 1). These are particularly abundant to the south of Glen Doe (Peacock, 1977
; Millar, 1990; Peacock et al., 1992), the area from which our samples come (NH 210 127). The garnetiferous gneisses here form layers of 0·3 m thickness within metapsammites, and can be traced over several hundred metres. The flat-lying metapelitic and metapsammitic layers seem to be flanks of decametre-scale isoclinal D2 folds. Such structures can be seen affecting deformed metabasite dykes within the Ardgour gneiss 500 m east of the sample location (Millar, 1999). Frequently, isoclinal folds of centimetre scale can be observed in the metapelitic and metapsammitic gneiss bands. Large garnet porphyroblasts are restricted on metapelitic gneiss layers. They are up to 5 cm in diameter and frequently show a euhedral shape. |
PETROGRAPHY AND MINERAL CHEMISTRY |
|---|
Microprobe analysis
Electron microprobe analysis of relevant minerals was performed at the University of Würzburg using a CAMECA SX-50 instrument. Operating conditions were 15 kV acceleration voltage, 15 nA beam current, 1 µm beam size, and an element-dependent integration time of 15–20 s. White mica and plagioclase were analysed with a defocused beam of 5 µm size. Pure oxides, and natural and synthetic silicates were used as standards. Corrections for atomic number, fluorescence and absorption were carried out by means of the PAP program supplied by CAMECA. The procedures used for the calculation of mineral formulae and of end-member activities are described in Appendix A, and representative analysis are presented in Appendix B.
Generalities
Large euhedral and small, euhedral and xenomorphic garnets occur within a layered matrix comprising alternating bands of biotite–epidote–quartz and plagioclase–quartz (Fig. 2), which locally contain centimetre-scale isoclinal rootless folds. Euhedral garnet porphyroblasts up to 3 cm in size show symmetric and asymmetric, locally sigma-shaped pressure shadows, formed of plagioclase and quartz (Fig. 2). The latter minerals show 120° triple point junctions and often straight grain boundaries, indicating static relaxation of the fabric above the plagioclase recrystallization temperature (500°C, Kruhl & Voll, 1976). Along sparse cracks the rocks show a greenschist-facies overprint, with the formation of retrograde chlorite and rare calcite.
|
Garnet
Four types of garnet (A–D) can be distinguished in the investigated gneiss sample, on the basis of their size, internal textures, mineral inclusions and zonation patterns (Figs 2–7). To avoid misinterpretations resulting from cutting effects, only the largest garnets, which were well characterized before thin-section preparation and in thin section, were analysed. To prove that individual garnets of similar macro- and microscopic features belong to the same garnet type, qualitative line scans over two garnets of type A and three of type B, C and D, respectively, were performed. Garnet type A forms euhedral crystals up to 3 cm in size, which show complex internal textures related to static and syndeformational garnet growth. Four zones, designated as garnet zones AI–AIV, can be distinguished (Figs 2b, 3a and b, 4, 6a and 7a). Garnet type B reaches only 0·7 cm in diameter. It is hypidiomorphic, and has an inclusion-rich core and inclusion-poor rim. Garnet type B shows only three zones (BI, BII and BIII), which correlate with zones II, III and IV of garnet type A (Figs 6a and 7). Garnet type C has nearly the same size as garnet type B (0·5 cm), but is xenomorphic and contains only two zones (CI and CII), which are similar to those of garnet AIII and AIV. Typically, some garnets of type C show s-shaped inclusion trails (Fig. 2c). Garnet type D is only 0·1 cm in size and contains one zone, equivalent to garnet AIV. In places, garnet type D forms more or less completely idiomorphic atoll garnets.
|
|
|
|
|
|
Garnet AI
The texture of the garnets, their compositional zoning and their mineral inclusions provide evidence for a very complex history of crystallization and deformation, which starts with the growth of the euhedral garnet core (Fig. 6b). The formation of the cores must have taken place under static conditions, as indicated by their idiomorphic shape (Figs 2b, 4 and 6a), and the random orientation of numerous apatite, sphene and quartz inclusions (Fig. 3b). Furthermore, garnet A contains inclusions of white mica, chlorite, epidote, rare ilmenite, allanite and plagioclase (Table 1).
|
The composition of garnet zone AI, especially the high spessartine content (Xsps), clearly indicates that it formed before the growth of garnet B, C or D (Fig. 7). As there are relatively few garnets of type A in the rock, the nucleation density during the first growth of garnet must have been very low (10 garnets/dm3). The large size of garnet AI (2 cm in diameter) and the low nucleation density require material transport over a relatively long distance, and may point to a high permeability of the rock during this early stage of metamorphism.
Garnets AII and BI
Textures, mineral inclusions and compositional zoning indicate at least three domains in garnet AII (AIIa, b and c; Figs 2b, 3a, 4, 6a and 7a). Domain AIIa forms a pressure shadow and occurs on two opposite sides of garnet A (Figs 2b, 4 and 6a). It is characterized by numerous inclusions of quartz, sphene, apatite and rare epidote (Table 1). These are randomly oriented at the direct contact with garnet zone AI, but form trails oblique to the garnet AI surface toward the rim of garnet AIIa (Fig. 6a). This clearly indicates increasing strain toward the rim of domain AIIa. In contrast to garnet domain AIIa, inclusions in domains AIIb and c are parallel to the surface of garnet AI (Figs 2b, 3a and 6b). Domains AIIb and AIIc are successively formed and can be distinguished by their different mineral contents. Domain AIIb forms the inner zone in direct contact with the euhedral garnet AI and contains ilmenite, apatite, sphene, epidote, biotite, muscovite and rare chlorite, whereas domain AIIc grew later and is ilmenite free (Fig. 3a). It is notable that garnet AIIa is directly overgrown by zone AIIc but never by zone AIIb (Fig. 6b).
Generally, the compositional zoning of garnet AII is characterized by an increase of Xalm, Xpy and Xgrs, and a decrease of Xsps (Fig. 7a). In detail, however, the zonation is much more complicated, as can be seen from the variation in Xsps (Figs 4, 6a and 7a). In the pressure shadow domain AIIa, Xsps decreases rapidly from 13 to 10–8 mol % and then forms a plateau, whereas Xsps decreases continuously from 13 to 4 mol % throughout domains AIIb and c (Figs 6a and 7a). The zonation pattern indicates that garnet domain AIIb initially grew faster than AIIa (Figs 4, 6b and 7a). The absence of ilmenite in domain AIIa might result from differential stress during simultaneous growth of both garnet zones, leading to the formation of different assemblages in distinct domains. However, it could also result from the fact that garnet AIIa growth started later than garnet AIIb. Interestingly, at Xsps = 10–8 mol % domain AIIa suddenly grew much faster than AIIb, and finally zones AIIa and AIIb were overgrowth by garnet zone AIIc (Fig. 6a and b).
Taken together, the chemical composition and the textural observations clearly indicate different growth rates in different domains of garnet AII, which seems to correlate with different magnitudes of deformation. Fast initial growth of garnet AIIb was obviously accompanied by high strain, and slow garnet AIIa formation with low strain. These features indicate that at least the first period of garnet AII growth was syntectonic. The symmetric array of the individual garnet AII domains points to deformation under pure shear conditions (Fig. 6b). However, it should be emphasized that some metres away from sample NH 210 127 in the same gneiss layer, type A garnets with typical snowball textures can be observed (see Millar, 1990). These rather indicate simple shear deformation during garnet AII growth.
An additional deformation event after garnet AII growth is evident from different orientations between the pressure shadows represented by garnet AIIa (PS1 in Fig. 6a), and the sigma-shaped plagioclase–quartz pressure shadows around garnet A (PS2 in Fig. 6a). Furthermore, some garnets of type A from sample NH 210 127 were sheared either before or during garnet zone AIII overgrowth, at the latest (Figs 2a and 6b).
Finally, it should be noted that at least parts of garnet AII zonation correlates with that of garnet type B core (garnet BI, Figs 6a and 7). That indicates new garnet formation in the matrix at the same time as garnet AII growth. In garnet BI, a foliation is weakly traced by inclusions of quartz and rare epidote (Fig. 6a).
Garnets AIII, BII and CI
As is clearly shown in Fig. 5, the boundary between garnet zones AII and AIII is characterized by numerous complex embayments, which indicate garnet resorption before garnet AIII overgrowth. A gap in the garnet growth history at the AII–AIII transition is also documented by a slight increase of Xsps (4·0–4·5 mol %) and Fe/(Fe + Mg) (0·962–0·971), as shown in Fig. 7a (point F). Increase of Fe/(Fe + Mg) points to a period of cooling before garnet AIII overgrowth, whereas the increase in Xsps results from manganese refractionation as a result of garnet resorption. Following this transition zone, garnet AIII growth is characterized by a relatively abrupt increase of Xgrs from about 31 to 39 mol % and decreasing Xalm (from 63 to 57 mol %) and Xsps (from 4·4 to 2 mol %, Fig. 7a). Toward garnet zone AIV Xgrs decreases (from 39 to 32 mol %) and Xalm (from 57 to 62 mol %) increases, whereas Xsps (1–2 mol %) is nearly constant. In places the garnet AIII zonations grade continuously into AIV (Fig. 7b), but mostly change abruptly. Typical garnet AIII zonations can also be observed in garnet type B (zone BII) and in the core of garnet type C (garnet CI; Figs 6a and 7). This clearly indicates formation of new garnet C simultaneously with garnet AIII and BII overgrowth. Some garnets of type CI show s-shaped inclusion trails, which indicate syntectonic garnet growth (Fig. 2c). All garnet types contain inclusions of quartz, biotite, epidote, apatite, sphene, plagioclase and very rare magnetite (Table 1).
Garnets AIV, BIII, CII and D
The formation of garnet zones AIV, BIII and CII took place simultaneously with the growth of new garnet D under static conditions (Figs 6a and 7). Garnet zone AIV can be distinguished from zone AIII by much lower Xgrs and higher Xalm contents (Fig. 7d and e). The rim of garnet zone AIV shows the lowest Fe/(Fe + Mg) ratio. Notably, garnet D locally forms idiomorphic atoll garnets in the matrix, enclosing plagioclase, quartz, biotite, epidote, sphene and rare ilmenite (Table 1).
Epidote
Epidote forms euhedral crystals in the matrix and occurs in garnets A, B, C and D. It invariably contains cores of allanite (Fig. 3c), which served as crystallization seeds. Twinning of allanite frequently continues into the surrounding epidotes. Quartz, apatite and biotite are common inclusions, whereas ilmenite occurs rarely. Most matrix epidotes are completely mantled by biotite (Fig. 3c), although in some cases they occur in direct contact with plagioclase and quartz. In such domains, epidote frequently forms symplectites with quartz, rarely accompanied by magnetite (Fig. 3d). Such symplectites were also found enclosed in garnet zones AIII, AIV, CII and D, indicating that the symplectites are either pre- or syn-genetic to AIII (see below). The symplectites were probably formed as a result of the inverse net-reaction [(R3), see below].
Most epidotes show three zones (EI–EIII), with more or less abrupt transitions between them (Fig. 8). Epidote zone EI occurs at the direct contact to allanite inclusions and shows the lowest pistacite content of Xps = 16 mol % [Xps = Fe3+/(Fe3+ + Al)], whereas Xps within zone EII varies between 20 and 23 mol %. Frequently, Xps decreases in zone EII from 23 to 20 mol % toward the rim (see Fig. 8: Ep1). In epidote zone EIII Xps increases to 29·7 mol % (Fig. 8: Ep2, Ep3), which is a nearly pure pistacite end-member composition. Symplectitic textures are restricted to epidote zones EII and EIII.
|
Numerous epidote inclusions allow a correlation between the growth history of garnet and epidote. Epidote with Xps = 23 mol % (= EII), observed at the rim of garnet AI indicates that epidote EI formation was already finished before garnet AII growth started (Fig. 8). We assume that epidote EI formation took place before or during garnet AI growth. The composition of epidotes enclosed in garnet AII is nearly constant (Xps = 20–23 mol %), whereas Xps increases rapidly to 29·5 mol % in garnet AIII (Ep3: Fig. 8). As epidotes with all three zones (EI–EIII) are overgrown by garnet zone AIII, epidote zone EIII must be formed either before or with garnet zone AIII, at the latest (see below).
Biotite
Biotite is widespread in the matrix and occurs as inclusion in garnets AII, AIII, AIV, B and C. Biotite also occurs as inclusions in epidote zones EII and EIII, and rarely is enclosed in allanite surrounded by epidote EI. Because all allanite and epidote (with Xps < 20 mol %) were already formed before garnet AII growth started (see above), biotite must also have been present during garnet AI growth.
Biotite in all textural domains has extremely high contents of titanium, at 0·39–0·43 p.f.u., and of AlVI, at 0·31–0·42 p.f.u. (see Appendices A and B). The silica contents of biotite are relatively low, ranging between 5·36 and 5·43 p.f.u. The Fe/(Fe + Mg) ratio of matrix biotites, of large biotite inclusions in garnet D and of small biotites included in epidote (EI and EII) and allanite ranges between 0·76 and 0·79. In contrast, biotites enclosed in garnet AIII + IV and CI + II show large variations of Fe/(Fe + Mg) between 0·77 and 0·84. The highest Fe/(Fe + Mg) ratio (0·85) was noted for biotites in contact with retrograde chlorite. A local increase of Fe/(Fe + Mg) of garnet in contact with biotite indicates retrograde, diffusional Fe–Mg exchange.
Chlorite
Numerous chlorite inclusions were observed in garnet AI, and possibly formed during the prograde evolution. All other chlorite occurrences, especially those formed along garnet cracks and in the matrix, result from retrograde overprint. According to the nomenclature of Hey (1954) all chlorites are daphnites, with a calculated amount of Fe3+ between 0·00 and 0·37 p.f.u. (Appendices A and B). Chlorite in garnet AI shows Fe/(Fe + Mg) ratios between 0·960 and 0·963, and Mn contents between 0·351 and 0·327 p.f.u., whereas chlorites formed along cracks have a Fe/(Fe + Mg) value between 0·821 and 0·872, and Mn contents between 0·161 and 0·07 p.f.u.
Muscovite
Muscovite was frequently observed in garnet zones AI and AII, rarely in AIII, but never in AIV. The Si content of muscovite in garnet zone AI ranges from 6·07 to 6·22 p.f.u., and in garnet zones AII and III from 6·02 to 6·17 p.f.u. Notably, all muscovites have very low aluminium contents (AI muscovites: 5·14–5·30 p.f.u.; AII and AIII muscovites: 5·11–5·53 p.f.u.) and high iron contents, up to 5·0 wt % Fe2O3tot. As the Si content of muscovite is low, iron cannot be explained as ferrous iron, entering the octahedral position by the Tschermak’s exchange: Si1Fe2+1AlIV-1AlVI-1. It seems more likely that ferric iron substituted for octahedral aluminium via the exchange AlVI-1Fe3+ VI1. The calculated amount of ferric iron (Appendix A) ranges between 0·25 and 0·55 p.f.u. in AI muscovites and between 0·14 and 0·30 p.f.u. in AII muscovites. The paragonite component of AI muscovites varies between 1·7 and 6·0 mol %, that of AII and AIII muscovites varies between 8·4 and 14·4 mol %.
Plagioclase
Plagioclase occurs frequently in the matrix and forms rare inclusions in all garnet zones. Matrix plagioclase shows highly variable zonation patterns. Four typical plagioclase types (Pl1–Pl4) are presented in Fig. 9. Plagioclase Pl1 first shows an increase of Xan from 36 to 46 mol % followed by a decrease to 26 mol % towards the rim. In contrast, plagioclase Pl2 is characterized by a decrease of Xan from 32 to 22 mol %, followed by a rapid increase to 36 mol %. The same increase can be observed in plagioclase Pl3 which, however, shows an unzoned core with Xan = 22–23 mol % and a rim with Xan = 30–36 mol %. Plagioclase Pl4 has an almost unzoned core with Xan = 28 mol % and shows increasing anorthite contents up to 38 mol % (not shown) towards the rim.
|
Zoned and unzoned plagioclase inclusions in garnet indicate a decrease of Xan from garnet zone AI rim (42–35 mol %) to garnet zone AII rim (22 mol %), followed by an increase of Xan in garnet zone AIII to 28–32 mol %. The maximum anorthite content found in garnet zone AIV was Xan =32 mol %. The change of Xan of the plagioclase inclusions between garnet AI rim and AIV correlates well with that found for matrix plagioclases Pl2 (Fig. 9).
Accessories
Apatite, sphene, allanite and ilmenite are common accessories, whereas magnetite is relatively rare. Apatite is widespread in the matrix and is enclosed in all garnet generations, whereas allanite forms cores of epidote or independent crystals within garnet, but is never found in the matrix. Ilmenite occurs commonly in the matrix and in garnet zone AIIb (Fig. 3a), and is rarely enclosed in garnet AI and in epidote. Ilmenites enclosed in garnet zone AIIb show a systematic decrease of the haematite and pyrophanite components from 6 to 4 mol %, and from 3 to 2 mol %, respectively, whereas the ilmenite component increases from 92 to 96 mol % toward the garnet rim. In contrast, rare ilmenite inclusions in garnet AI show pyrophanite contents of 16–17 mol % and ilmenite contents of 79–81 mol %. Matrix ilmenites have the same compositional variations as ilmenites enclosed in garnet zone AIIb. Magnetite, which has nearly stoichiometric composition, forms idiomorphic grains in some matrix domains and very rarely in garnet BI–II, and occurs along retrograde cracks within garnet. Sphene from all domains shows slightly elevated contents of aluminium (0·17–0·24 p.f.u.) and iron (0·05–0·08 p.f.u.). Zircon was observed, enclosed in allanite, in garnet, and in matrix quartz and biotite.
|
MINERAL ASSEMBLAGES AND REACTION HISTORY |
|---|
Textural relationships and mineral inclusions clearly indicate that the growth of garnet AI took place in the assemblage
 |
 |
 |
 |
 |
 |
A reaction explaining the formation of garnet AI and AII is
 |
The transition between garnet zones AII and AIII forms an important discontinuity. This is evident from resorption textures before garnet AIII growth (Fig. 5), from the abrupt increase of Xgrs with the start of garnet zone AIII formation (Fig. 7a), and from the sudden appearance of numerous epidote, biotite and plagioclase inclusions in garnet AIII. Inclusions found in garnet type AIII indicate that this garnet zone grew in the chlorite-free assemblage
 |
 |
Zoned plagioclase (Xan = 23 32) and epidote inclusions (Xps = 20 30) in garnet AIII provide evidence that they were formed either before or simultaneously with garnet AIII. Formation of both minerals before garnet AIII growth seems likely, because some of the epidote and plagioclase inclusions are resorbed by garnet. Nevertheless, it may also be possible that new epidote and plagioclase were formed together with garnet AIII, whereas older epidote and plagioclase, no longer in chemical equilibrium with AIII, were consumed. Formation of plagioclase and epidote before garnet AIII growth can be explained by the inverse reaction (R1), if chlorite was present. Otherwise the inverse chlorite-free reaction
 |
 |
All three inverse reaction (R1)–(R3) are in agreement with garnet resorption before garnet AIII growth (Figs 5 and 6b). They support the interpretation that epidote symplectites (Fig. 3d) were formed before garnet AIII (see below). Following garnet AII resorption, growth of garnet AIII took place, via either reaction (R2) or reaction (R3). Subsequently, garnet AIV was formed in the muscovite-free assemblage
 |
 |
Assemblage (P3) requires that muscovite reacted out with the start of garnet AIV growth, at the latest. Resorption of epidotes in the matrix and enclosed in garnet AIV indicates that epidote was probably consumed during garnet AIV growth. Finally, as a result of the subsequent retrograde overprint new epidote, plagioclase, chlorite, magnetite and rare calcite were formed along thin cracks, which cut through all garnet zones (Fig. 2b).
|
P–T PATH RECONSTRUCTION |
|---|
Geothermobarometry
P–T calculations were carried out using compositions of coexisting zones of garnet, epidote and plagioclase, as well as ilmenite, muscovite and chlorite inclusions, assumed to represent equilibrium conditions during different stages of garnet formation. Diffusive smoothing of the initial growth zonation of garnet, plagioclase and epidote can be excluded, because all these minerals still show steep compositional gradients between individual zones (e.g. epidote EI–EII–EIII; garnet AII–AIII; Pl2; see Figs 7–9). The continuous decrease of Fe/(Fe + Mg) toward garnet AIV rim (Fig. 7b) also indicates that a retrograde Fe–Mg exchange of garnet with matrix biotite is insignificant. Only at the direct contact (
5–10 µm) with some biotite and chlorite inclusions does the Fe/(Fe + Mg) in garnet increase slightly. Furthermore, there is no evidence for any chemical change of plagioclase, epidote and ilmenite compositions after their entrapment by garnet, because the garnet composition at the contact with these inclusions is unmodified.
P–T calculations were carried out with the program Thermocalc V2.4, which is based on the internally consistent thermodynamic dataset of Powell & Holland (1988) and Holland & Powell (1990). The ln K values, slopes and errors of the end-member reaction (Fig. 10)
 |
. Mineral end-members and activity models used for calculations are listed in Table 2 and Appendix A, respectively. The garnet–biotite geothermometer of Kleemann & Reinhardt (1994, see below) was also employed to calculate temperatures for garnet zone AII-rim and AIV-rim, assuming that biotite inclusions in epidote zone EII coexists with garnet AIIc, and matrix biotites coexist with garnet AIV.
|
|
The results of the P–T calculations, which are summarized in Table 2, indicate different P–T conditions during formation of different zones of garnet A (Fig. 10). From minerals enclosed in garnet AI P–T conditions of 495 ± 13°C at 3·5 ± 0·7 kbar were estimated, whereas the rims of garnet AIIb and AIIc give higher metamorphic conditions of 634 ± 45°C at 9·4 ± 1·0 kbar and 657 ± 59°C at 10·0 ± 1·3 kbar, respectively. The temperatures calculated with Thermocalc for AIIc are higher than that calculated with the garnet–biotite geothermometer of Kleemann & Reinhardt (1994), ranging between 560 and 590°C (at 9 kbar). Coexisting garnet AIII (Xgrs = 0·39), epidote EIII (Xps = 29·7) and plagioclase (Xan = 0·29) yield a pressure of 4·8 ± 0·8 kbar, at a temperature of 640°C. Garnet AIV matrix–biotite pairs calculated with the garnet–biotite geothermometer of Kleemann & Reinhardt (1994) yielded temperatures between 653 and 680°C at an assumed pressure of 5·5 kbar.
At the moment, because of the lack of any geochronological data, it is not clear whether the individual P–T points represent increments during a single, clockwise P–T evolution, or if they result from a polymetamorphic evolution (see below). Furthermore, it should be noted that the P–T calculations given above should be regarded as rough approximations. This especially concerns the temperature estimates obtained with the ‘average pressure–temperature calculations’ mode of Thermocalc, which is strongly dependent on the composition of the phyllosilicates chlorite, biotite and muscovite. As mentioned above, it is possible that chlorite re-equilibrated during the retrograde evolution or that the composition of biotite, although enclosed in epidote EII, changed during the prograde or retrograde evolution. In the case of retrograde chlorite re-equilibration, the temperature of garnet AI-rim is too low; in the case of prograde biotite re-equilibration the temperatures of garnets AIIb and c are too high.
Another important point is that the biotites and muscovites in our sample are complex solid solutions. Biotite, for instance, contains high amounts of Ti, Mn and Alvi and low silica contents, whereas muscovite contains important amounts of Fe3+ instead of Alvi (see above). Furthermore, we have no control over the amounts of Fe3+ of biotite. The activity models used in Thermocalc calculations for the phyllosilicates (Appendix A) do not consider the complex effects resulting from excess energies of mixing. In contrast, effects of Mn, Ti and Alvi in biotite and the complex garnet composition are taken into account in the garnet–biotite geothermometer of Kleemann & Reinhardt (1994) which, therefore, should constrain the temperatures much better. A further uncertainty is the water activity, which is assumed here to be 1·0. At a water activity <1·0 all P–T points estimated with Thermocalc will shift to lower temperatures and pressures, as shown in Table 2.
Pressure estimates seem to be less problematic, because they dominantly result from reaction (R4), using end-members of plagioclase, epidote and garnet. The compositions of these minerals are unaffected by prograde and retrograde diffusive alterations, as clearly indicated by the preserved zonation patterns (see above). Activity models are available that take into account excess energies of mixing for plagioclase and garnet (Appendix A). Nevertheless, water activities below 1·0 shift the position of reaction (R4) in Fig. 10 for a given ln K to lower temperatures, so that higher pressures at the same temperatures would be obtained. Furthermore, it should be noted that the pressure of 4·8 ± 0·8 kbar estimated for garnet zone AIII is significant only when the rims of epidote and plagioclase inclusions are really in equilibrium with AIII. For the case that they represent metastable relics formed before garnet AIII growth, the estimated pressure would be meaningless (see below).
Gibbs method modelling
To set additional P–T path constraints that result from the zonation of garnet, plagioclase and epidote, the Gibbs method of differential thermodynamics is employed (Rumble, 1974; Spear et al., 1982; Spear, 1988, 1989a, 1989b). In this method, the intensive variables pressure (P) and temperature (T), and extensive variables, such as phase composition (X) and phase abundance (M), are related by a set of differential thermodynamic, stoichiometric and mass balance equations. Dependent upon the number of phases and system components chosen, a certain number of variables is independent [monitor variables according to Menard & Spear (1993)] and all remain dependent (see Table 3 and 4). Changes of independent variables, e.g. Xalm, Xgrs, Xsps of garnet, result in a simultaneous change of all other variables, e.g. P, T, Xan, etc. The advantage of the Gibbs method is that depending upon the purpose of the study, independent variables can be chosen in different ways.
|
|
Three types of Gibbs method calculations were performed here, using the Apple Macintosh program GIBBS (Spear & Menard, 1989; Spear et al., 1991), with activity and mixing models supplied with the program (Table 3). For magnetite, thermodynamic data were added to the program using the dataset of Berman (1988). First, P–T paths were computed using the technique of Spear & Selverstone (1983). Second, P–T space was contoured with mineral composition isopleths to determine relations between P–T–X–M in different assemblages. Third, garnet zoning models were constructed along estimated P–T paths to evaluate effects resulting from different garnet nucleation densities during the metamorphic history.
Reference conditions
A prerequisite for the use of the Gibbs method is to find at least one reference point at which the phase compositions of an equilibrium assemblage at given pressures and temperatures and optionally the mineral mode are known. For well-crystallized rocks that contain continuously zoned minerals, this is simply done using the rim composition and the mode of all minerals in equilibrium, from which the start P–T conditions are obtained. For discontinuously zoned minerals, as dealt with in this study, it may be necessary to define more than one starting condition. This, however, is dependent upon the kind of the discontinuity between the different mineral zones. During a continuous P–T evolution a discontinuity in the zonation can result from
- (U1) change of the mineral assemblage;
- (U2) change of the mineral assemblage, accompanied by a temporary consumption of the zoned mineral;
- (U3) changing nucleation density in a specific mineral assemblage;
- (U4) changing nucleation density and changing mineral assemblage.
During a polymetamorphic history a discontinuity additionally results from a skip of the P–T conditions
- (U5) in a certain mineral assemblage;
- (U6) accompanied by a change of the mineral assemblage (or fluid species) in between;
- (U7) accompanied by changing nucleation density in a specific mineral assemblage;
- (U8) accompanied by a change of the nucleation density and the mineral assemblage.
Examples for discontinuities of types (U1)–(U4) have been given, for example, by Spear et al. (1990), Spear (1993) and Menard & Spear (1993). In contrast, examples for type (U5)–(U8) discontinuities are rare, and without geochronological data difficult to prove (e.g. Vernon, 1996). In case of a continuous P–T evolution only one set of starting P–T–X–M conditions is required, as shown by Spear et al. (1990) and Menard & Spear (1993). In this situation, change of the mineral assemblage is simply simulated by adding or removing minerals from an existing assemblage at a certain P–T condition, and change of the nucleation density by changing the number of garnets per volume. However, for a real polymetamorphic evolution, new starting P–T–X–M conditions are required for each zone interrupted by a discontinuity.
The petrological observations mentioned above clearly indicate important discontinuities between garnet zones AI and AII, and between zones AII and AIII, which coincide with discontinuities in zoned epidote and plagioclase. Furthermore, there is evidence that the nucleation density increases from garnet AI to AIV. Unfortunately, there are no geochronological data for the several garnet zones, which could demonstrate whether all garnet zones were formed during a single or during different orogenic events. Therefore, it is not clear whether the transition AI–AII is a (U3) or (U7) discontinuity, and whether the AII–AIII discontinuity is of type (U4) or (U8). Because of these uncertainties two models will be discussed.
Model I. As a first approach, it is assumed that the zonation of garnet AI–AIV results from a nearly continuous P–T evolution during a single orogenic event. The transition between garnet zones AI and AII represents a (U3) discontinuity, and the transition between garnet zones AII and AIII is a (U4) discontinuity.
P–T path calculations and P–T–X–M contouring for model I were started near the rim of garnet zone AIIc (point A in Fig. 7a), at which chlorite-out is inferred. Consequently, we have to change from the model assemblage
 |
 |
for minerals from garnet zone AIIc (see Fig. 10). Mineral compositions and mineral abundances used for modelling are listed in Table 3. The start compositions of biotite and chlorite were recalculated using the garnet–biotite geothermometer of Ferry & Spear (1978) and the garnet–chlorite geothermometer of Ghent et al. (1987), respectively (Tables 3 and 4). This step was necessary, because prograde chlorite was never found in zone AIIc, and the composition of natural biotite is too complex (see above). For calculations in model I pure muscovite, clinozoisite and H2O was used. For biotite and chlorite mixing on one site models and for garnet a mixing model of Berman (1990) were employed. The sets of independent reaction used by the program GIBBS for the calculations are listed in Table 5. It should be noted that for the case that the starting P–T conditions given in Table 3 are incorrect, the form of the resulting P–T paths and the P–T–X–M contours are nearly the same; however, their absolute position in the P–T diagram would be shifted.
|
Model II. For this model a polymetamorphic evolution is assumed, which led to the formation of a type (U7) discontinuity between garnet zones AII and AIII. The discontinuity between garnet zones AI and AII is considered to be the same as in model I. Furthermore, it is assumed that consumption of the garnet AII rim and formation of garnet AIII took place within the chlorite-free assemblage (P2).
Model II-1. In a first step, starting P–T conditions for garnet AII consumption are defined at 600°C and 8·5 kbar. These are slightly higher temperatures than those used in model I, but are still in agreement with the results of the conventional geothermobarometry (see Fig. 10 and Table 3). These starting conditions take into account that the outermost rim of garnet AII is already consumed. In addition to model I, magnetite is added to assemblage (P2m) and epidote is considered to be an ideal solid solution of the mineral end-members clinozoisite (Xczo) and pistacite (Xpist). These additions enable us to show constraints that result from the epidote zonation.
Model II-2. In a second step, the P–T–X conditions estimated from inclusions in garnet AIII are assumed to define new starting conditions, decoupled from the prehistory. The new starting P–T conditions are fixed at 640°C and 5·5 kbar. These are in agreement with the results obtained from inclusions in garnet AIII using reaction (R4) and the garnet–biotite geothermometer of Kleemann & Reinhardt (1994, see above). As magnetite is rarely observable in garnet BI–BII, it was not involved in the calculation, but epidote is still used as a solid solution (Table 3).
P–T path calculation
A Gibbs method P–T path was calculated only for model I, assuming a successive growth from garnet zones AI–AIII in equilibrium with all minerals in the respective model assemblages (P1m) and (P2m). The P–T path segment for zone AI–AII was derived from the change of the composition from points A to E (Figs 7a and 11), using Xgrs, Xsps and Xalm in garnet AI and AII as monitor variables (Table 4). The calculations show that garnet AI grew during heating of only 8°C (Fig. 11), with a slight pressure decrease of 0·3 kbar, whereas formation of garnet AII was accompanied by an important pressure increase of 2·5 kbar and a slight temperature rise of at most 10°C (Fig. 11, Table 4). Removal of chlorite, necessary for P–T path calculations from garnet AIII zonation, increases the variance of the system from three to four, which requires an additional independent variable (Table 4). Here we used the zonation of Xan. The assumed ‘short-lived’ consumption of garnet AII before AIII in model I is assumed to have taken place in assemblage (P2m). It leads to a slight pressure and temperature decrease to 572°C at 8·3 kbar (point F, Figs 7a and 11), followed by an important temperature and moderate pressure increase to 638°C at 9·2 kbar, represented by the initial zonation of garnet AIII (point G in Fig. 11).
|
P–T–X–M modelling
For P–T–X–M modelling, mass balance constraints are involved, which reduce the variance of the system to two, assuming closed-system behaviour. The starting conditions for models I, II-1 and II-2 are shown in Table 3.
Model I. A typical feature of all contour plots is that the slopes of the isopleths and/or the spaces between them change dramatically with chlorite-out (Fig. 12). Following the thick P–T trajectory in assemblage (P1m) (Fig. 12) the contour plots for model I predict well the zonation patterns of garnet zones AI, AII and AIII, and of plagioclase. They confirm the rapid decrease of Xsps and Xan simultaneously with the rapid increase of Xalm towards the rim of garnet AII, whereas Fe/(Fe + Mg) slightly decreases and Xgrs increases. The thick P–T trajectory in assemblage (P2m) (Fig. 12) conforms with rapid Xgrs increase and decrease of Xalm in garnet AIII, and the slight decrease of Xsps and Fe/(Fe + Mg). Furthermore, it explains the increase of Xan of plagioclase enclosed in garnet AIII (see Fig. 9).
|
Mass balance constraints of model I (Fig. 12f) indicate growth of garnet and muscovite, and release of fluid along the P–T path in assemblage (P1m), whereas chlorite, epidote, plagioclase and biotite will be consumed. This is in agreement with reaction (R1), discussed above. Furthermore, model I is consistent with the start of garnet growth at the calculated garnet-in isograd (Grt = 0 in Fig. 12f). For assemblage (P2m) the model explains garnet AIII growth accompanied by plagioclase formation, according to reaction (R2). Furthermore, the contour plots for assemblage (P2m) indicate that decrease of Xgrs, and increase of Xalm toward garnet AIII rim (point H in Fig. 7b) could result from a nearly isothermal decompression of 3–4 kbar, shown as a dotted P–T path section in Fig. 12. However, in disagreement with the natural observations, garnet AIII would be consumed along the dotted P–T path trajectory and plagioclase with Xan up to 80 mol % would be formed. Another important weakness of model I is that it cannot explain the formation of epidote before or with garnet AIII growth, as required by the thin-section observations (see above).
Model II-1. The same discrepancies with respect to epidote result if calculations are carried out using starting conditions of model II-1, which takes magnetite and an epidote solid solution into account. As shown in Fig. 13, the additional phase components lead to a change of the slopes of the isopleths Xalm, Xgrs, Xsps and Fe/(Fe + Mg) with respect to model I (Fig. 12). Nevertheless, these contours also constrain a similar isobaric temperature rise during initial garnet AIII growth (dotted arrow; Fig. 13) as obtained for model I (Fig. 11, points A–G). In contrast, formation of epidote with increasing Xpist in this assemblage can occur only when the temperature falls (e.g. black arrows in Fig. 13). Thus, the contour diagrams in Fig. 13 indicate that formation of epidote and garnet AIII cannot be cogenetic. The diagrams rather support the interpretation that epidote was formed during cooling at the expense of garnet AII before garnet AIII growth. This is also in agreement with the resorption pattern and the increase of Xsps and Fe/(Fe + Mg) observed at the rim of garnet zone AII (point F in Fig. 7a), and with mass balance constraints indicating that during cooling of 50°C some 2 vol. % garnet will be consumed (Fig. 13f). Less than 0·5 mm would be consumed from a garnet of 12 mm radius, provided that all garnets in the rock are consumed in the same manner. Depending on the form of the cooling path plagioclase is either consumed during epidote growth (Fig. 13f, black arrow) or formed together with epidote.
|
From the contour plots shown in Fig. 13, the textures observed at the transition between garnet zones AI and AII could be explained as follows. In a first step, garnet, plagioclase, muscovite and magnetite were consumed during a possible isobaric cooling, whereas epidote with increasing Xpist contents and biotite were formed (black arrow). During a subsequent pressure decrease (grey arrow in Fig. 13) plagioclase with increasing Xan (up to 35 mol %) and biotite were formed. During this stage most of the well-zoned euhedral epidotes were overgrown by biotite (Fig. 3c and d). All other epidotes underwent an intensive resorption. This evolution may have stopped when magnetite reacted out. During renewed heating (white arrow in Fig. 13) garnet AIII was formed. The rapid change of the composition at the garnet AII–AIII transition probably indicates that the advent of garnet AIII was a sudden event and not a continuous process. P–T conditions for an instant start of garnet AIII growth are marked by the black star in Fig. 13. The contours of Xgrs, Xalm, Xsps and Fe/(Fe + Mg) at these P–T conditions (625°C at 6 kbar) are in good agreement with those observed from garnet AIII (point G in Fig. 7a). Notably, the P–T conditions are near those estimated for garnet AIII (see above).
Model II-2. In this model garnet AIII growth is assumed to start abruptly at 640°C and 5·5 kbar, in an assemblage represented by the inclusions found in garnet AIII, and independent of the prehistory. In contrast to model II-1, magnetite was removed (see above). That leads to a change of the slopes of Xgrs, Xalm, Xsps, Fe/(Fe + Mg) and Xpist with respect to model II-1 (Figs 13 and 14). The black arrow shown in Fig. 14 defines a P–T vector, which roughly explains the increase of Xalm and the decrease of Xgrs and Fe/(Fe + Mg) toward the rim of garnet AIII (profile G–H in Fig. 7b). Furthermore, it is consistent with plagioclase growth with increasing Xan contents, and explains the disappearance of muscovite during AIII formation. However, in disagreement with the natural observations, garnet would be consumed along this P–T vector.
|
Garnet zoning model
Garnet modelling was performed along the P–T path shown in Fig. 11 for model I, with growth of garnet AI, AII and BI from point E to A in assemblage (P1m), and of garnet AIII, BII and CI in assemblage (P2m) from point A to G (point F was excluded). Starting phase compositions are those obtained at point E (Grt = 0) by Gibbs P–T path calculations with model I (Tables 3 and 4). Garnet fractionation was modelled for garnet zone AI and AII, but not for AIII (see below), resetting the number of moles of garnet to zero after each P–T increment. Garnet nucleation densities were defined with 10 Grt/dm3 during formation of garnet AI, 40 Grt/dm3 for AII–BI and 400 Grt/dm3 for garnet AIII–BII–CI (Fig. 15), in agreement with the textural observations made on the investigated gneiss sample. Because of the requirements of GIBBS an ideal mixing on one site model for garnet was used. Growth progress of garnets of different size was correlated for the assumption of constant area growth, which corresponds to garnet growth kinetics controlled by diffusion in the matrix (Carlson, 1989; Kretz, 1993). In this model, the same constant area increment is added to each existing garnet.
The calculated size and zonation of garnet zones AI and AII in assemblage (P1m) agree well with those measured (Fig. 15), except that the change in composition at the garnet AI–AII transition is less sharp in the computer model than in the natural garnet. In contrast to garnet A, the zonation of garnet BI is poorly constrained. This particularly concerns the size of garnet B, which is smaller than predicted by the computer simulation. There are at least three possibilities for this disagreement: (1) important amounts of garnet BI rim were resorbed, before it was overgrown by garnet BII; (2) the measured garnet BI is a piece of an initially larger garnet, detached by deformation before garnet BII overgrowth; (3) nucleation of the measured garnet BI started later than garnet AII growth. Whatever the reason, new formation of garnet type B must have started simultaneously with garnet AII growth, otherwise the zonation profile of garnet zone AII would appear as a simple continuation of the garnet AI zonation, and that clearly is not the case. The garnet growth modelling supports the assumption that the garnet AI–AII discontinuity results at least from a change in the nucleation density (U3). However, an additional skip of the P–T conditions between garnet AI and AII growth cannot be excluded (U7 discontinuity).
The change from assemblage (P1m) to (P2m), and the increase of the nucleation density to 400 Grt/dm3 along the path section A–G (Fig. 11), predicts well the formation of a narrow AIII and BII zone, and a wide core in garnet C. However, it should be kept in mind that garnet AIII–BII–CI modelling is carried out under the assumption that the P–T evolution is continuous (model I) and the AII–AIII transition is of type (U3). However, the situation in the investigated gneiss sample seems to be more complex (see above). Furthermore, it should be noted that growth modelling of garnet zones AIII–BII and CI in assemblage (P2m) was only achieved without considering fractionation. With fractionation, after a short period of growth along the P–T vector A–G in Fig. 11, garnet AIII, BI and CI would be consumed.
Taken together, the good agreement between the computed and measure zonation profiles of garnet A (Fig. 15) supports the thin-section observation that the garnet nucleation density increased during the metamorphic evolution. This independently constrains a rise in temperature during successive garnet growth, because the nucleation density correlates exponentially with temperature (Carlson, 1989).
|
DISCUSSION AND CONCLUSIONS |
|---|
The results of this study show that garnet–epidote–biotite-gneisses are valuable monitors, which allow the reconstruction of the complex pressure–temperature–deformation history of the Glenfinnan Group (Fig. 16). Within the investigated gneisses sparse large and numerous small garnets occur, indicating successive garnet formation with increasing nucleation densities during rising temperatures (Fig. 16). Large garnets show four zones (AI–AIV), which are successively formed in the assemblages
 |
 |
 |
 |
 |
|
The four zones of the large garnets are, in part, represented in smaller garnets (B, C and D), which contain three or fewer zones. Results of conventional geothermobarometry and Gibbs method modelling can be interpreted to reflect a polymetamorphic evolution, recorded by the zonation pattern of garnet, plagioclase and epidote. Gibbs modelling points to a fast garnet AI growth during heating from about 550 to 560°C, accompanied by an insignificant pressure decrease from 6·3 to 6·0 kbar. The nucleation density during the formation of garnet AI was very low (10 Grt/dm3), leading to growth of the large garnets. The internal structures (Fig. 3b) and the idiomorphic habit (Figs 2b, 4 and 6) of garnet AI indicate static growth conditions. The prograde P–T path section and the internal textures of garnet AI are consistent with contact metamorphism, perhaps associated with intrusion of the West Highland Granite Gneiss protolith during Proterozoic rifting (Fig. 16). Such a model is supported by the occurrence of the Glen Doe body of the West Highland Granite Gneiss near the sample location, which is transected by metabasalts and metagabbros with MORB signatures (Millar, 1999). Both the granite gneiss and the intercalated metagabbros have yielded identical U–Pb zircon ages of 873 ± 7 Ma (Friend et al., 1997) and 873 ± 6 Ma (Millar, 1999), respectively. Rifting is correlated with deformation D1 in the Glenfinnan Group, during which sparse migmatitic structures were locally formed (Roberts & Harris, 1983; Millar, 1999).
Garnet zone AII was formed during nearly isothermal compression from 6·0 to 8·5 kbar (from 560 to 575°C), indicating crustal stacking. The complex internal texture of garnet AII clearly indicates that growth of this zone was affected by deformation in a simple shear regime. However, as mentioned above, other garnets from the same gneiss layer and with identical composition show snowball textures (Millar, 1990; A. Zeh, unpublished data, 1999), more consistent with a non-coaxial deformation. It may be that all these garnets were formed together during isoclinal folding in the Glenfinnan Group (D2). The different garnet textures may result from garnet growth in different locations within the isoclinal folds; simple shear in the cores of the folds and pure shear at the flanks. As supported by garnet modelling the nucleation density during garnet AII growth increases to at least 40 Grt/dm3, resulting in the simultaneous formation of new garnet B. The P–T vector estimated for garnet zone AII is similar to that obtained by Vance et al. (1998) for Morar Group metasediments. Therefore, it is likely that it results from the same stacking event in Neoproterozoic times between 820 and 790 Ma, designated as the Knoydartian orogeny. If this is the case, garnet zone AII formed much later than zone AI. In this case the garnet AI–AII discontinuity would be of type (U7).
Resorption patterns and retrograde zoning at the transition between garnet zones AII and AIII provide evidence for an further gap in the garnet growth history. Gibbs modelling supports the interpretation of a cooling event before garnet AIII growth, during which new epidote and plagioclase were formed at the expense of garnet AII (model II-1). However, because of the lack of geochronological data, the duration of this cooling event is not clear. Either it was a ‘short-lived’ cooling event, perhaps caused by stacking of a relatively warm rock pile above a colder one followed by a thermal relaxation (Fig. 16), or garnet zone AIII was formed much later, perhaps during the Caledonian orogeny. In the latter case, garnet AIII growth results from a polymetamorphic P–T evolution (Fig. 16). Whatever the reason was, garnet AIII and BII growth, and new formation of garnet CI, requires higher temperatures than were necessary for garnet AII–BI formation, because the Fe/(Fe + Mg) content in these garnet zones is higher and the nucleation density increases from AII to AIII by a factor of 10. According to the results of conventional geothermobarometry, formation of garnet AIII started at conditions of about 5·5 kbar and 640°C. S-shaped inclusion trails in garnet CI indicate that zone AIII growth was accompanied by deformation. It is likely that formation of the asymmetric quartz–plagioclase pressure shadows around garnet A and shearing of previously formed garnets AI and AII (Figs 2a and 6) coincides with garnet AIII growth. The deformation that led to the formation of these structures must have taken place before the D3 event sensu Roberts & Harris (1983, see above) but after D2 mentioned above. Therefore, it will be designated here as D2a.
Finally, garnet AIV was formed at temperatures of 640–670°C at pressures <6 kbar. The idiomorphic habit of garnet AIV indicates formation under static conditions (Figs 4 and 16). As yet, there is no conclusive evidence to show if garnet AIV formation represents the end of a unique P–T history, or if it results from a metamorphic overprint, perhaps related to the Caledonian orogeny. During the final retrograde evolution the rock was locally transected by thin cracks, filled with epidote, chlorite, calcite, plagioclase and magnetite.
|
APPENDIX A: MINERAL FORMULA CALCULATIONS AND ACTIVITY MODELS |
|---|
Garnet
Garnet, X3VIIIY2VIZ3IVO12, analyses were normalized to 24 oxygens. Fe2+/Fe3+ is derived as follows: AlIV = 6 – Si, AlVI = Altotal – AlIV, Fe3+ = 4 – AlVI – Ti – Cr (assuming all Mn is divalent), Fe2+ = Fetotal – Fe3+. Activities for garnet end-members almandine (alm), pyrope (py), grossular (grs), spessartine (sps) were calculated after Berman (1990)
. Mole fractions of Xalm, Xgrs, Xsps and Xpy are Xalm = Fe2+/(Fe2+ + Mg + Ca + Mn).
Biotite
Biotite, XXIIY2VIZ4IVO10(OH)2, was normalized to 22 oxygens. Activities of biotite end-members were calculated as follows (all Fe = Fe2+):
 |
Muscovite
Muscovite, XXIIY2VIZ4IVO10(OH)2, analyses are normalized to total cations less (Na + K + Ca) = 6. Fe2+/Fe3+ estimation based on the assumption of (Fe2+,Mg,Mn),Si = (AlVI, Fe3+,Ti)AlIV substitution (Tschermak’s substitution): Fe2+ = (Si – 3) – (Mg + Mn – Ti) and Fe3+ = Fetot – Fe2+. Activities of muscovite end-members were calculated as follows:
 |
Chlorite
Chlorite, Y6VIZ4IVO10(OH)8, analyses are normalized to 10 cations and then the following assignments were made: AlIV = 4 – Si and AlVI = Altotal – AlIV.
Charge balance requires that Fe3+ = AlIV – AlVI – Cr – 2Ti; Fe2+ = Fetotal – Fe3+. Activities of chlorite end-members were calculated as follows:
 |
Epidote
Epidote, X2VIIIY3VIZ3IVO12(OH), analyses were normalized to 25 oxygens. All Fe was assumed to be Fe3+. Pistacite contents (Xps) were calculated with Xps = Fe3+/(Fe3+ + Al), activities of the epidote end-members as follows:
 |
Ilmenite
Ilmenite, R2+R4+O3, was normalized to two cations. Ferric iron was estimated as follows: Fe2+ = Ti - (Mn + Mg), Fe3+ = Fetot - Fe2+. End-members for ilmenite (ilm), haematite (hem), pyrophanite (pyph) and geikelite (gei) were calculated.
Plagioclase
Plagioclase (NaxCa1-xAl2-xSi2+xO8) analyses were normalized to five oxygens. Activities of the end-members albite (ab), anorthite (an), orthoclase (or) were calculated according to Fuhrman & Lindsley (1988).
Sphene and magnetite
Sphene, XVIIIYVIZIVO4(OH), and magnetite, R2+R3+2O4, analyses were normalized to three cations. In sphene all iron is assumed to be ferric. Activities of sphene (sph) and magnetite (mt) were asumed to be 1·0.
|
APPENDIX B: REPRESENTATIVE ELECTRON MICROPROBE ANALYSES |
|---|
|
|
ACKNOWLEDGEMENTS |
|---|
We thank M. Okrusch for help with discussions about the petrological data, P. O’Brien for garnet mapping, and U. Schüssler and K. P. Kelber for technical support. G. T. R. Droop, H. Rollinson and T. Nagel are thanked for their helpful reviews and comments on a former version of the manuscript.
|
FOOTNOTES |
|---|
*Corresponding author. Telephone: +49-931-888-405-728. E-mail: armin.zeh{at}mail.uni-wuerzburg.de
|
REFERENCES |
|---|
Barr, D., Roberts, A. M., Highton, A. J., Parson, L. M. & Harris, A. L. (1985). Structural setting and geochronological significance of the West Highland Granite Gneiss, a deformed granite with Proterozoic, Moine rocks of NW Scotland. Journal of the Geological Society, London 142, 663–675.[Web of Science]
Berman, R. G. (1988). Internally-consistent thermodynamic data for minerals in the system Na2O–K2O–CaO–MgO–FeO–Fe2O3–Al2O3–SiO2–TiO2–H2O–CO2. Journal of Petrology 29, 445–522.
Berman, R. G. (1990). Mixing properties of Ca–Mg–Fe–Mn garnets. American Mineralogist 75, 328–344.[Abstract]
Carlson, W. P. (1989). The significance of intergranular diffusion to the mechanism and kinetics of porphyroblast crystallization. Contributions to Mineralogy and Petrology 103, 1–24.
Clifford, T. N. (1957). The stratigraphy and structure of part of the Kintail district of southern Ross-shire: its relation to the Northern Highlands. Quarterly Journal of the Geological Society, London 113, 57–85.
Ferry, J. M. & Spear, F. S. (1978). Experimental calibration of the partitioning of Fe and Mg between biotite and garnet. Contributions to Mineralogy and Petrology 66, 880–899.
Friend, C. R. L., Kinny, P. D., Rogers, G., Strachan, R. A. & Paterson, B. A. (1997). U–Pb zircon geochronological evidence for Neoproterozoic events in the Glenfinnan Group (Moine Supergroup): the formation of the Ardgour granite gneiss, north-west Scotland. Contributions to Mineralogy and Petrology 128, 101–113.
Fuhrman, M. L. & Lindsley, D. H. (1988). Ternary feldspar modelling and thermometry. American Mineralogist 73, 201–215.[Abstract]
Ghent, E. D., Stout, M. Z., Black, P. M. & Brothers, R. N. (1987). Chloritoid bearing rocks with blueschists and eclogites, northern New Caledonia. Journal of Metamorphic Geology 5, 239–254.[Web of Science]
Hey, M. T. (1954). A new review of the chlorites. Mineralogical Magazine 30, 278–292.
Holdsworth, R. E., Strachan, R. A. & Harris, A. L. (1994). Precambrian rocks in northern Scotland east of the Moine Thrust: the Moine Supergroup. In: Gibbons, W. & Harris, A. L. (eds) A Revised Correlation of Precambrian Rocks in the British Isles. Geological Society, London, Special Reports 22, 23–32.
Holland, T. J. B. & Powell, R. (1990). An enlarged and updated internally consistent thermodynamic dataset with uncertainties and correlations: the system K2O–Na2O–CaO–MgO–MnO–FeO–Fe2O3–Al2O3–TiO2–SiO2–C–H2–O2. Journal of Metamorphic Geology 8, 89–124.[Web of Science]
Johnstone, G. S. (1975). The Moine succession. In: Harris, A. L., Shackleton, R. M., Watson, J. V., Downie, C., Harland, W. B. & Moorbath, S. (eds) A Correlation of Precambrian Rocks in the British Isles. Geological Society, London, Special Reports 6, 30–42.
Johnstone, G. S., Smith, D. I. & Harris, A. L. (1969). The Moinian assemblage of Scotland. In: Kay, M. (ed.) North Atlantic Geology and Continental Drift. Memoirs, American Association of Petroleum Geologists 12, 159–180.
Kleemann, U. & Reinhardt, J. (1994). Garnet–biotite thermometry revisited: the effect of Alvi and Ti in biotite. European Journal of Mineralogy 6, 925–941.[Web of Science]
Kretz, R. (1993). A garnet population in Yellowknife schist, Canada. Journal of Metamorphic Geology 11, 101–120.[Web of Science]
Kruhl, J. H. & Voll, G. (1976). Fabrics and metamorphism from the Monte Rosa Root Zone to the Ivrea Zone near Finero, southern margin of the Alps. Schweizerische Mineralogische und Petrographische Mitteilungen 56, 627–633.
Menard, T. & Spear, F. S. (1993). Metamorphism of calcic pelitic schists, Stafford Dome, Vermont: compositional zoning and reaction history. Journal of Petrology 34, 977–1005.
Millar, I. L. (1990). Caledonian and pre-Caledonian events in Moine rocks of the Cluanie area, Inverness-shire. Ph.D. thesis, University of London.
Millar, I. L. (1999). Neoproterozoic extensional basic magmatism associated with the West Highland granite gneiss in the Moine Supergroup of NW Scotland. Journal of the Geological Society, London 155, 1–11.[Web of Science]
Peacock, J. D. (1977). Metagabbros in granitic gneiss, Inverness-shire, and their significance in the structural history of the Moines. Institute of Geological Sciences Report, 77/20.
Peacock, J. D., Mendum, J. R. & Fettes, D. J. (1992). Geology of the Glen Affric district. Memoir of the British Geological Survey, Sheet 72E (Scotland).
Powell, R. & Holland, T. J. B. (1988). An internally consistent dataset with uncertainties and correlations: 3. Applications to geothermobarometry, worked examples and a computer program. Journal of Metamorphic Geology 6, 173–204.[Web of Science]
Roberts, A. M. & Harris, A. L. (1983). The Quoich Line—a limit of early Palaeozoic crustal reworking in the Moine of the northern Highlands of Scotland. Journal of the Geological Society, London 140, 883–892.
Roberts, A. M., Strachan, R. A., Harris, A. L., Barr, D. & Holdsworth, R. E. (1987). The Sgurr Beag nappe: a reassessment of the stratigraphy and structure of the Northern Highland Moine. Geological Society of America Bulletin 98, 497–506.
Rumble, D. (1974). Gibbs phase rule and its application to geochemistry. Journal of the Washington Academy of Sciences 64, 199–208.
Selverstone, J. & Spear, F. S. (1985). Metamorphic P–T paths from pelitic schists and greenstones in the southwest Tauern Window, Eastern Alps. Journal of Metamorphic Geology 3, 439–465.[Web of Science]
Soper, N. J., Harris, A. L. & Strachan, R. A. (1998). Tectonostratigraphy of the Moine Supergroup: a synthesis. Journal of the Geological Society, London 155, 13–24.
Spear, F. S. (1988). The Gibbs method and Duham’s theorem: the quantitative relationships among P, T, chemical potential, phase composition, and reaction progress in igneous and metamorphic systems. Contributions to Mineralogy and Petrology 99, 249–256.
Spear, F. S. (1989a). Relative thermobarometry and metamorphic P–T paths. In: Daly, J. S., Cliff, R. A. & Yardley, B. W. D. (eds) Evolution of Metamorphic Belts. Geological Society, London, Special Publication 43, 63–81.
Spear, F. S. (1989b). Petrologic determination of metamorphic pressure–temperature–time paths. In: Spear, F. S. & Peacock, S. M. (eds) Metamorphic Pressure–Temperature–Time Paths. Short Course in Geology. Washington, DC: American Geophysical Union, pp. 1–55.
Spear, F. S. (1993). The origin and interpretation of zoned metamorphic minerals. In: Spear, F. S. (ed.) Metamorphic Phase Equilibria and Pressure–Temperature–Time Paths. Washington, DC: Mineralogical Society of America, pp. 576–639.
Spear, F. S. & Menard, T. (1989). Program GIBBS: a generalized Gibbs method algorithm. American Mineralogist 74, 942–943.[Abstract]
Spear, F. S. & Selverstone, J. (1983). Quantitative P–T paths from zoned minerals: theory and tectonic applications. Contributions to Mineralogy and Petrology 83, 348–357.
Spear, F. S., Ferry, J. M. & Rumble, D. (1982). Analytic formulation of phase equilibria: the Gibbs method. In: Ferry, J. M. (ed.) Characterization of Metamorphism through Mineral Equilibria. Mineralogical Society of America, Reviews in Mineralogy 10, 105–152.
Spear, F. S., Kohn, M. J., Florence, F. P. & Menard, T. (1990). A model for garnet growth in pelitic schists: implications for thermobarometry and P–T path determinations. Journal of Metamorphic Geology 8, 683–696.
Spear, F. S., Peacock, S. M., Kohn, M. J., Florence, F. P. & Menard, T. (1991). Computer programs for petrological P–T–t path calculations. America Mineralogist 76, 2009–2012.
Tanner, P. W. G., Johnstone, G. S., Smith, D. I. & Harris, A. L. (1970). Moinian stratigraphy and the problem of the Central Ross-shire inliers. Geological Society of America Bulletin 81, 299–306.
Vance, D., Strachan, R. A. & Jones, K. A. (1998). Extensional versus compressional setting for metamorphism: garnet chronometry and pressure–temperature–time histories in the Moine Supergroup, northwest Scotland. Geology 26, 927–930.
Vernon, H. R. (1996). Problems with inferring P–T–t-paths in low-P granulite facies rocks. Journal of Metamorphic Geology 14, 143–153.
Will, T. M. (1998). Phase equilibria in metamorphic rocks: thermodynamic background and petrological applications. Lecture Notes in Earth Science 71. Berlin: Springer, 309 pp.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. ZEH Calculation of Garnet Fractionation in Metamorphic Rocks, with Application to a Flat-Top, Y-rich Garnet Population from the Ruhla Crystalline Complex, Central Germany J. Petrology, December 1, 2006; 47(12): 2335 - 2356. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. ZEH, R. KLEMD, and J. M. BARTON JR Petrological evolution in the roof of the high-grade metamorphic Central Zone of the Limpopo Belt, South Africa Geological Magazine, May 1, 2005; 142(3): 229 - 240. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. ZEH Crystal Size Distribution (CSD) and Textural Evolution of Accessory Apatite, Titanite and Allanite during Four Stages of Metamorphism: an Example from the Moine Supergroup, Scotland J. Petrology, October 1, 2004; 45(10): 2101 - 2132. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. ZEH, I. L. MILLAR, and M. S. A. HORSTWOOD Polymetamorphism in the NE Shackleton Range, Antarctica: Constraints from Petrology and U-Pb, Sm-Nd, Rb-Sr TIMS and in situ U-Pb LA-PIMMS Dating J. Petrology, May 1, 2004; 45(5): 949 - 973. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||