Journal of Petrology Advance Access originally published online on April 22, 2005
Journal of Petrology 2005 46(8):1725-1746; doi:10.1093/petrology/egi034
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
A New Interpretation of Centimetre-scale Variations in the Progress of Infiltration-driven Metamorphic Reactions: Case Study of Carbonated Metaperidotite, Val d'Efra, Central Alps, Switzerland
1 DEPARTMENT OF EARTH AND PLANETARY SCIENCES, JOHNS HOPKINS UNIVERSITY, BALTIMORE, MD 21218, USA
2 GEOPHYSICAL LABORATORY, CARNEGIE INSTITUTION OF WASHINGTON, 5251 BROAD BRANCH ROAD, NW, WASHINGTON, DC 20015, USA
3 EARTH SYSTEM SCIENCE INTERDISCIPLINARY CENTER, UNIVERSITY OF MARYLAND, COLLEGE PARK, MD 20742, USA
RECEIVED JULY 28, 2004; ACCEPTED MARCH 7, 2005
 |
ABSTRACT |
|---|
TOP
ABSTRACT
INTRODUCTIONProgress (
) of the infiltration-driven reaction, 4olivine + 5CO2 + H2O = talc + 5magnesite, that occurred during Barrovian regional metamorphism, varies at the cm-scale by a factor of 3·5 within an 
3 m3 volume of rock. Mineral and stable isotope compositions record that XCO2, 
18Ofluid, and 13Cfluid were uniform within error of measurement in the same rock volume. The conventional interpretation of small-scale variations in in terms of channelized fluid flow cannot explain the uniformity in fluid composition. Small-scale variations in resulted instead because (a) reactant olivine was a solid solution, (b) initially there were small-scale variations in the amount and composition of olivine, and (c) fluid composition was completely homogenized over the same scale by diffusion–dispersion during infiltration and subsequent reaction. Assuming isochemical reaction, spatial variations in image variations in the (Mg + Fe)/Si of the parent rock rather than the geometry of metamorphic fluid flow. If infiltration-driven reactions involve minerals fixed in composition, on the other hand, spatial variations in do directly image fluid flow paths. The geometry of fluid flow can never be determined from geochemical tracers over a distance smaller than the one over which fluid composition is completely homogenized by diffusion–dispersion. KEY WORDS: Alpine Barrovian metamorphism; diffusion; metamorphic fluid composition; metamorphic fluid flow; reaction progress
|
INTRODUCTION |
|---|
Carbonation and decarbonation reactions during metamorphism in the crust typically are driven by infiltration of rocks by chemically reactive fluids (e.g. Ferry & Gerdes, 1998
; Ferry et al., 2002). Significant differences in the progress () of the infiltration-driven reactions commonly occur between contrasting lithologic layers within individual outcrops (Ferry, 1994; Ferry & Rumble, 1997; Ferry et al., 1998, 2001) and, in some cases, large differences occur between adjacent layers only 1 cm thick (Ferry, 1987). The variations in conventionally are interpreted in terms of channelized, layer-parallel fluid flow, with elevated flow in the high- layers and reduced flow in low- layers (Ferry, 1987, 1994). The interpretation is correct only if there is no significant chemical communication during metamorphism between adjacent high- and low- layers, either by advection, diffusion or mechanical dispersion (the combination of the latter two is referred to as ‘diffusion–dispersion’ in the rest of the paper). The conventional interpretation once appeared reasonable because limited cross-layer chemical communication during metamorphism seemed to be documented by significant layer-by-layer differences in fluid composition (e.g. Rumble, 1978; Ferry, 1979; Kohn & Valley, 1994), some at the cm-scale (e.g. Rumble & Spear, 1983). More recent field (e.g. Bickle et al., 1997; Evans et al., 2002; Ague, 2003), theoretical (e.g. Ague, 2000, 2002) and experimental studies (e.g. Wark & Watson 2004), however, indicate efficient homogenization of fluid composition over several metres across lithologic layers during regional metamorphism caused by exchange of CO2, H2O and other fluid species by diffusion–dispersion. If correct, at least some layer-by-layer variations in the progress of infiltration-driven reactions demand another explanation. We propose the alternative explanation that cm- to m-scale variations in may result when adjacent layers initially contain different amounts and/or compositions of reactant mineral solid solutions, and fluid composition is homogenized across layering by diffusion–dispersion at all times during subsequent infiltration and reaction. The importance of cross-layer diffusion–dispersion in driving metamorphic devolatilization reactions has been recognized by others as well (e.g. Hewitt, 1973; Ague & Rye, 1999).
Modal, mineral chemical and stable isotope data for the carbonated metaperidotite body in Val d'Efra, Central Alps, Switzerland (Evans & Trommsdorff, 1974), were used to test whether the conventional or new interpretation better explains cm-scale variations in the progress of an infiltration-driven reaction during one instance of Barrovian regional metamorphism. The metaperidotite is nearly ideal for the investigation because of several reasons. First, most samples experienced a single, simple mineral–fluid reaction at or near the peak of Barrovian metamorphism,
 | (1) |
1, are up to a factor of 2·6 over a distance of <1 m. Second, rocks contain numerous proxies for metamorphic fluid composition (mole fraction of the forsterite component in olivine, Xfo,Ol, for XCO2; 18OMgs and 18OOl for 18Ofluid; and 13CMgs for 13Cfluid, where subscripts Ol and Mgs refer to olivine and magnesite, respectively). The proxies allow accurate determination of the scale of homogenization of fluid composition relative to the scale of variations in 1 without explicit consideration of T. Third, the minerals are close to binary Fe–Mg solid solutions and reaction (1) involves only one reactant mineral. The relationship between 1 and the amount and composition of mineral reactants therefore can be completely and quantitatively represented on a single two-dimensional diagram. Fourth, the study benefits from three decades of excellent mineralogical and petrologic work, both on the metaperidotite body (Evans & Trommsdorff, 1974) and on associated rocks in the region (summaries by Pfiffner & Trommsdorff, 1998; Pfiffner, 1999; Nimis & Trommsdorff, 2001). |
GEOLOGIC SETTING |
|---|
The metaperidotite body at Guglia, Val d'Efra (Fig. 1), is one of numerous boudins composed of metamorphosed ultramafic and mafic rocks and rodingite, ranging from several metres to several hundred metres in size, in the Cima Lunga unit of the Penninic nappe system (Pfiffner & Trommsdorff, 1998
; Nimis & Trommsdorff, 2001). The best known are at Alpe Arami and Cima di Gagnone. In Val d'Efra, the boudins are set in a matrix of felsic gneisses, pelitic schists and metacarbonate rocks (Evans & Trommsdorff, 1974). Boudins and their host rocks are considered to have been part of an ocean basin near a continental margin and exhumed oceanic mantle lithosphere that were subducted, metamorphosed and uplifted during the Alpine orogeny. The ultramafic boudins represent oceanic mantle lithosphere. Some of the metamorphosed mafic and ultramafic rocks retain a mineralogical record of Eocene (35–43 Ma) ultra-high pressure (UHP) metamorphism. Mineral equilibria in prograde metamorphosed garnet lherzolite at Cima di Gagnone, 1 km SW of the metaperidotite body in Val d'Efra, for example, record P 30 kbar and T 740°C (Nimis & Trommsdorff, 2001). Where fluids gained access to ultramafic rocks, as in Val d'Efra, however, all mineralogical evidence for UHP metamorphism was obliterated by later Alpine Barrovian regional metamorphism at P = 6–8 kbar and T = 600–660°C (Grond et al., 1995).
|
The metaperidotite body at Guglia, Val d'Efra, is exposed over a 400–500 m2 area (Fig. 1) and vertically over 10–15 m on a vertical exposure that bounds its western margin. The contact between metaperidotite and surrounding rock is buried by vegetation and alluvium. Metaperidotite is primarily composed of two mappable lithologies (Fig. 1). The schlieren facies is schist composed of olivine (Ol), talc (Tlc), magnesite (Mgs) and chlorite (Chl) with and without enstatite (En). [These and other abbreviations for minerals follow Kretz (1983)
]. Schlieren are defined by wispy lighter-colored regions, richer in Tlc, set in a darker matrix richer in Ol [Fig. 2a of this study and plate 1A of Evans & Trommsdorff (1974)]. All samples of schlieren and matrix collected for this study contain Ol, Tlc, Mgs and Chl. Some schlieren contain small amounts of En in addition (0·5–3·8 modal %); the matrix to the schlieren contains no En. The matrix grades into schlieren over 1 cm; boundaries between schlieren and matrix cut foliation at a low angle. Evans & Trommsdorff (1974) concluded that the schlieren developed by replacement of the matrix. The schlieren, however, differ from adjacent matrix in bulk composition [lower (Fe + Mg)/Si] rather than simply in greater progress of reaction (1) (e.g. compare modes of matrix sample 16B with schlieren samples 16H and 16 M, Table 1). We interpret the schlieren as features that (a) developed prior to Barrovian metamorphism, either by a primary magmatic process during formation of the igneous parent rock, by deformation in the mantle, or by Si-metasomatism of the parent rock during serpentinization, and (b) were then ductilely deformed during regional metamorphism.
|
|
Rocks of the prismatic enstatite facies are composed of randomly oriented, prismatic En crystals, up to several centimetres long, set in a finer-grained foliated matrix of Ol, Tlc, Mgs and Chl [Fig. 2b of this study and Plate 3 of Evans & Trommsdorff (1974)
]. They differ from rocks of the schlieren facies, both in their larger grain size and mineralogy (significantly more En in samples collected for this study, 11–42 modal %). In three dimensions (3D), the prismatic enstatite facies appears to form a thin shell, 1–2 m thick, around the margin of the metaperidotite body. Because of its much greater volume, this study focused on the schlieren facies.
The metaperidotite is cut by three sets of veins. The commonest, the ‘composite veins’ of Evans & Trommsdorff (1974), are vertical, with NE strike, 1 mm wide, composed of Tlc, Mgs and anthophyllite (Ath), and bounded by a selvage of Ol-free, Tlc–Mgs–Chl rock [Plates 1B and 5 of Evans & Trommsdorff (1974)]. Composite veins are well exposed in the schlieren facies, with typical spacings of 20–40 cm (minimum 0; maximum 130 cm); selvages have fairly uniform thickness, with a half-width of 1–2 cm. As recognized by Evans & Trommsdorff (1974), the composite veins also record infiltration of metaperidotite by reactive CO2-rich, CO2–H2O fluid. Composite veins and their selvages cut across both foliation and schlieren. Selvages of composite veins and adjacent host rocks of the schlieren facies have significantly different Tlc and Mgs contents (e.g. compare samples 7H and 7S, Table 1) and significantly different O- and C-isotope compositions, separated by steep gradients in both modes and isotopic composition. Fluids that produced the composite veins therefore were different from those that drove reaction (1) in the schlieren facies, and they infiltrated the metaperidotite body along fractures after the mineralogy of the schlieren facies developed. The composite veins are not directly relevant to the study's focus on mineral reactions in the schlieren facies. Nevertheless, the composite veins cannot be ignored because their later formation disturbed the stable isotope composition of adjacent samples of the host schlieren facies. In addition to the composite veins, there is a 7-m-long, folded actinolite (Act)–Chl vein and a set of millimetre-wide Ath veins without selvages. Because they are minor constituents, the Act–Chl and Ath veins are not considered further.
|
METHODS OF INVESTIGATION |
|---|
Internal contacts between lithologies within the metaperidotite body were mapped in 3D to decimetre accuracy, using a laser rangefinder and a digital fluxgate compass (Fig. 1). The locations of samples within areas of <10 m2 were recorded with compass and metal measuring tape.
Twenty-four samples of the schlieren facies, three of the prismatic enstatite facies and three of the composite veins and their selvages were collected for modal, mineral and stable isotope analysis (Fig. 3). All samples of the prismatic enstatite facies and one of the schlieren facies were obtained in place. Because of the difficulty in sampling glacially polished surfaces of the schlieren facies and to avoid defacing the beautiful exposures of the metaperidotite body, the remaining samples of the schlieren facies and the composite veins were obtained from rectangular blocks, 1–5 m in long dimension, that have fallen from the vertical exposure that bounds the western margin of the boudin. The rectangular shapes of the blocks result from their breaking along parallel composite veins that often define two faces of the blocks. Thirteen samples of schlieren facies were obtained from a single block over a 110-cm-long traverse oriented perpendicular to foliation (location 16, Fig. 3). These were supplemented by four other samples from the same block, offset from the line of traverse parallel to foliation. Together, the 17 samples are referred to in the text, figures and tables as from the ‘m-scale traverse’ and are designated samples 16A–16M (a numerical suffix indicates the four samples collected offset from the line of traverse). One of the 17 samples contains a composite vein (16M); no other composite veins occur within or between the other samples. An additional six samples of the schlieren facies (designated 2 and 11–15) were collected from other blocks, and they are representative of the range in colour (and hence in proportions of Ol, Tlc and Mgs) of rocks from the schlieren facies exposed in situ. The three samples of composite veins (designated 7, 17 and 18) include the complete selvage as well as host schlieren facies rock outside the selvage on both sides of the vein. A sample of pelitic schist was collected from each of two outcrops located 10–35 m from the metaperidotite body for mineral thermometry and barometry (locations 6 and 19, Fig. 3).
|
Mineral assemblages were determined in thin section with optical petrography and backscattered electron (BSE) imaging using the JEOL JXA-8600 electron microprobe at Johns Hopkins University. Compositions of minerals in all samples of the schlieren and prismatic enstatite facies, two samples of selvages to the composite veins and the two samples of pelitic schist were determined by electron microprobe using wavelength-dispersive spectrometry with natural and synthetic mineral standards and a ZAF correction scheme (Armstrong, 1988
). X-ray maps were made of garnets in the pelitic schists, and regions near the rim with the lowest Mn contents were then analysed for mineral thermometry and barometry. Modes of all metaperidotite and two vein-selvage samples were measured by counting 
2000 points in thin section using BSE imaging. Any uncertainty in the identification of a particular point was resolved by obtaining an energy-dispersive X-ray spectrum.
Magnesite in all samples of the schlieren facies and the three samples of vein selvages was analysed for O- and C-isotope composition, following procedures described by Rumble et al. (1991). Approximately 6–30 mg finely powdered rock was obtained with a 2 mm diamond-tipped drill from a polished rock slab. Magnesite was dissolved overnight in phosphoric acid (McCrea, 1950) in evacuated reaction vessels at 100°C. Evolved CO2 was analysed with the Finnigan MAT 252 mass spectrometer at the Geophysical Laboratory. The acid fractionation factor was taken from Sharma et al. (2002). Results were normalized to the composition of calcite standard NBS-19 (18O = 28·65
, VSMOW; 13C = 1·95, VPDB, Coplen, 1988, 1996). Analyses of NBS-19 and a working calcite standard indicate that analytical precision for both oxygen and carbon isotopes is approximately ±0·1 (1
). All 18O analyses of samples weighing >25 mg are suspect because a significant decrease in T occurred during the relatively long time it took to remove the reaction vessel from the Al-metal heating block and mix the larger powdered samples with phosphoric acid. Values of 18O for samples >25 mg therefore are not reported. Because variations in T during reaction do not affect measurements of C-isotope composition, all measured values of 13C are reported.
The O-isotope composition of Ol in 11 samples of the schlieren facies was measured following procedures of Yui et al. (1995). Olivine separates were obtained by gently crushing samples and hand picking grains under a binocular microscope, followed by ultrasonic cleaning in distilled H2O. Oxygen was extracted from 2 mg of mineral separate in an atmosphere of BrF5 using a CO2 laser fluorination system similar to that of Sharp (1990). The O2 gas was collected, purified and directly analysed with the Finnigan MAT 252 mass spectrometer at the Geophysical Laboratory. Duplicates of all but one sample were measured. Results were normalized to garnet standard UWG-2 (18O = 5·8; Valley et al., 1995), whose composition was measured at the beginning and end of each analytical session. Based on multiple analyses of UWG-2 and of Ol pairs, the precision for 18OOl is considered ±0·1 (1).
Modal abundances of minerals were converted to molar abundances using mineral compositions and molar volumes of mineral components from Holland & Powell (1998). All calculations of mineral equilibria used Holland & Powell's (1998) thermodynamic database and THERMOCALC (version 3.1, 2001). Except for the anorthite component of plagioclase, activities of components in mineral solid solutions were computed from measured mineral compositions and Holland & Powell's AX program. The activity coefficient of the anorthite component in plagioclase was calculated from the experimental data of Goldsmith (1982) at 650°C and 9 kbar using thermodynamic data from Holland & Powell (1998) and THERMOCALC, v. 3·1, following methods described by Carpenter & Ferry (1984).
|
MINERALOGY AND MINERAL CHEMISTRY |
|---|
Modes and mineral compositions in selected samples of metaperidotite from the schlieren and prismatic enstatite facies and from the selvages of the composite veins are listed in Tables 1 and 2. Compositions of minerals in the two samples of pelitic schist used for mineral thermometry and barometry are given in Table 3.
|
|
All samples from the schlieren facies contain Ol, Tlc, Mgs and Chl, with and without En, along with accessory chromite (Chr), pyrrhotite (Po) and pentlandite (Pn). Retrograde serpentine (Srp) and magnetite (Mag) are ubiquitous (Table 1). Enstatite occurs in small amounts (0·5–3·8%) in
20% of the samples. The modal amount of Mgs varies by a factor of 6 (4·7–28·0%). Olivine, Tlc, Mgs and Chl are close to binary Fe–Mg solid solutions (Table 2). The principal divalent cations other than Fe and Mg are Ca, Mn and Ni, and they occur in relatively small concentrations. In all analysed minerals, Ca/(Mg + Fe) and Ni/(Mg + Fe) are both <0·004 and Mn/(Mg + Fe) is <0·005. Minerals have remarkably uniform Mg/(Fe + Mg) and Ol, in particular, displays no growth zoning. Chlorite contains significant but fairly constant amounts of Cr, 0·20–0·26 atoms per formula unit. Analysed samples from the prismatic enstatite facies have the same mineral assemblage as those from the schlieren facies, except that En is always present in substantial amounts (11–42%). There is a complete overlap in measured mineral compositions between the prismatic enstatite and schlieren facies (Table 2). Reconstructed from measured modes and mineral compositions, the range in bulk Fe/(Fe + Mg) of silicates and carbonate in analysed samples from the prismatic enstatite facies (0·081–0·093) overlaps with that of En-free samples from the schlieren facies (0·069–0·093). Likewise, the range in bulk (Mg + Fe)/Si of silicates and carbonate in analysed samples of the prismatic enstatite facies (1·48–1·72) overlaps with that of En-free samples from the schlieren facies (1·47–1·89). The greater amounts of En in rocks of the prismatic enstatite facies compared with those of the schlieren facies cannot be explained in any simple way by differences either in mineral chemistry or in bulk-rock composition.
Selvages to the composite veins are composed of Mgs, Tlc and Chl with accessory Chr, Po and Pn. Selvages are devoid of Ol, except minute quantities (0·1%) that occur as isolated inclusions in Mgs. The selvages are also devoid of retrograde Srp and Mag (e.g. sample 7S, Table 1). A sharp interface, 1 mm wide, separates vein selvages with no Ol (except as inclusions in Mgs) from adjacent rock of the schlieren facies with normal Ol contents (cf. samples 7H and 7S, Table 1). The veins themselves contain the same assemblage as in the vein selvages with the addition of 0·2–1·9% Ath. Compositions of Mgs, Tlc and Chl in the vein selvages are similar to those in the schlieren and prismatic enstatite facies but have systematically slightly higher Fe/(Fe + Mg), the result of reaction (1) having gone to completion in the selvages.
Analysed pelitic schists contain garnet, muscovite, biotite, kyanite, staurolite, plagioclase and quartz, with accessory ilmenite, rutile, monazite and Po, all with unexceptional compositions (Table 3).
|
STABLE-ISOTOPE GEOCHEMISTRY |
|---|
Measured O- and C-isotope compositions of Mgs and Ol from the schlieren facies and of Mgs from the selvages to the composite veins are listed in Table 4.
|
The O-isotope composition of Mgs depends on its occurrence. Magnesite in the schlieren facies >10 cm from a composite vein has fairly uniform
18OMgs = 9·2–9·7 (VSMOW); 18OMgs in composite veins and their selvages is significantly higher, at 10·9–11·9 (Table 4). Magnesite in the schlieren facies <10 cm from the vein selvages has intermediate 18OMgs = 9·8–11·2. Values of 18OMgs measured along a traverse from the composite vein in sample 7 through the vein selvage into adjacent schlieren facies (Fig. 4) indicate that a narrow 18O-enrichment halo, 10 cm wide, exists in the schlieren facies adjacent to the composite vein selvages.
|
Magnesite in the schlieren facies likewise has fairly uniform
13CMgs = –6·4 to –7·4 (VPDB). There is also 13C-enrichment within and adjacent to the composite veins. Measured 13CMgs for veins and their selvages is –5·3 to –6·4. The halo of 13C-enrichment around the composite veins, however, appears to be restricted to the vein selvage and does not extend into adjacent host rocks of the schlieren facies (Fig. 5).
|
Analysed Ol in the schlieren facies >10 cm from composite veins has remarkably uniform
18OOl = 4·4–4·7 (Table 4). The single sample of analysed Ol collected <10 cm from a composite vein (16L1) has slightly higher 18OOl = 4·8—a value, however, the same as the others within error of measurement. |
PRESSURE, TEMPERATURE AND FLUID COMPOSITION |
|---|
Published estimates of P recorded by mineral assemblages developed during Barrovian regional metamorphism in the area are: 6–7 kbar (Heinrich, 1982
), 6–8 kbar (Grond et al., 1995) and 6·1 kbar (Todd & Engi, 1997, fig. 7). Additional P estimates were calculated from the ‘average PT’ routine of THERMOCALC, using mineral compositions in the two analysed samples of pelitic schist (Table 3) and the mineral components an, ab, mu, pa, phl, ann, east, alm, py, gr, ilm, ru, fst, ky and q [abbreviations from Holland & Powell (1998)], with XH2O = 0·8 (as explained below). Results are P = 7·5 ± 1·4 (±2) kbar (sample 6) and 7·4 ± 1·6 kbar (sample 19). The preferred value of P, based on all four sets of estimates, was taken as 7 ± 1 kbar.
Published estimates of T recorded by mineral assemblages developed during Barrovian regional metamorphism in the area are: 600–650°C (Heinrich, 1982), 600–660°C (Grond et al., 1995) and 645°C (Todd & Engi, 1997). Published estimates are consistent with those computed from the ‘average PT’ routine—628 ± 24°C (±2) for sample 6 and 629 ± 28°C for sample 19. An additional T of metamorphism is recorded independently by the equilibrium between coexisting Ol, Mgs, Tlc and En in the metaperidotite and CO2–H2O fluid. Using representative reduced activities for the Mg-components in the minerals and THERMOCALC, calculated T = 643°C at 7 kbar (Fig. 6). The range in measured mineral compositions and the uncertainty in P of ±1 kbar introduce uncertainties of ±3 and ±7°C, respectively. The T of equilibrium among Ol, Mgs, Tlc, En and CO2–H2O fluid, computed from the data of Berman (1988, updated 1991), using ideal ionic mixing models to calculate reduced activities of Mg-components in minerals, is nearly the same—645°C at 7 kbar. Mineral assemblages in the metaperidotite evidently equilibrated at the same conditions, as did mineral assemblages in other lithologies in the region during Barrovian regional metamorphism. The preferred T of equilibration, based on all five sets of estimates, was taken as 645 ± 10°C.
|
The composition of CO2–H2O fluid in equilibrium with metaperidotite of the schlieren and prismatic enstatite facies at the preferred P–T conditions of mineral equilibration during Barrovian metamorphism was XCO2 = 0·20 ± 0·01 (Fig. 6), explaining the value of XH2O used in the ‘average PT’ calculations. The uncertainty in XCO2 is based on the range in measured mineral compositions.
|
CARBONATION OF METAPERIDOTITE |
|---|
Carbonation reaction
In principle, carbonation of the metaperidotite body could have occurred either during amphibolite facies Barrovian regional metamorphism by reaction (1), sometime earlier at lower grades of Barrovian metamorphism, or even prior to Barrovian metamorphism. If the precursor mineral assemblage subject to carbonation was not Ol + Tlc + Chl, progressive metamorphism of metaperidotite in the Alps indicates other plausible possibilities (Trommsdorff & Evans, 1974
). At progressively lower grades of metamorphism, rocks with compositions equivalent to those in the metaperidotite body in Val d'Efra are composed of antigorite (Atg) + Ol + Chl; brucite (Brc) + Atg + Chl or Atg + Tlc + Chl, depending on whole-rock (Mg + Fe)/Si; and chrysotile/lizardite (Ctl/Lz) + Brc + Chl or Ctl/Lz + Tlc + Chl, depending on (Mg + Fe)/Si. The Ctl/Lz + Brc + Chl and Ctl/Lz + Tlc + Chl assemblages also correspond to the mineralogy of any serpentinite precursor that could have been carbonated prior to Barrovian metamorphism.
There are several arguments that carbonation of the metaperidotite body in Val d'Efra occurred by reaction (1) during Barrovian metamorphism, and that the observed mineral assemblages do not simply represent metamorphism of ultramafic rock carbonated at an earlier time. First, Mgs is the dominant carbonate mineral in Alpine metaperidotites from the amphibolite facies but not in metaperidotites from lower grades (Trommsdorff & Evans, 1974). Metaperidotite elsewhere in the Central Alps at a grade equivalent to that in Val d'Efra is typically composed of Ol + Tlc + Chl. The regional distributions of minerals imply that Mgs formed at conditions of the amphibolite facies by reaction (1).
Second, comparison of the mineral assemblage in the selvages of the composite veins with that in adjacent host rock of the schlieren facies unequivocally demonstrates that Mgs + Tlc in the selvages formed from Ol by reaction (1). The veins and their selvages are undeformed and therefore could not have formed prior to amphibolite facies Barrovian regional metamorphism (Pfiffner, 1999). The selvages of the composite veins are proof that at least some parts of the metaperidotite body were carbonated by reaction (1) during Barrovian metamorphism.
Third, Mgs in the metaperidotite body commonly contains inclusions of Ol but not of Tlc or other minerals. The Ol inclusions are often in optical continuity with each other (Evans & Trommsdorff, 1974) and sometimes with Ol in the matrix. Carbonation of Ol-free equivalents composed of Atg + Brc + Chl, Atg + Tl + Chl, Ctl/Lz + Brc + Chl or Ctl/Lz + Tlc + Chl, either at conditions of lower-grade Barrovian metamorphism or prior to metamorphism, appears to be ruled out. In principle, Mgs with Ol inclusions could develop from either an Ol–Atg–Chl or an Ol–Tlc–Chl precursor. The Ol–Atg–Chl precursor is less likely for two reasons. When modes of analysed samples of metaperidotite are recast as an isochemical equivalent combination of Ol, Atg and Chl, the equivalent Ol–Atg–Chl rocks have modal Atg/(Ol + Atg) = 0·19–1·00. More than half the equivalent Ol–Atg–Chl rocks contain too little Ol to form amounts of Mgs now observed in the metaperidotite body in Val d'Efra by the reaction
 | (2) |
T–XCO2 conditions of reaction
Replacement of Ol with Tlc and Mgs by reaction (1) occurs at T between the Atg–Ol–Tlc–Mgs and En–Ol–Tlc–Mgs isobaric invariant points, 575–645°C at 7 kbar (Fig. 6). The corresponding range in XCO2 is 0·03–0·20 at 7 kbar. Carbonation could have occurred at a single T or over any range of T between 575 and 645°C.
Source of carbon
Values of 13CMgs = –5·3 to –7·4 provide the only constraints on the origin of C involved in the carbonation reaction. The C probably was derived from a combination of marine carbonate and reduced organic material, although a source in the mantle cannot be ruled out (Kyser, 1986).
|
SPATIAL DISTRIBUTION OF REACTION PROGRESS |
|---|
Progress of reaction (1) was computed for samples from the schlieren facies and from the selvages to the composite veins as (moles Mgs)/5, referenced to 1 l of Ol–Tlc–Chl schist prior to reaction. Measured modes therefore were corrected for the increase of rock volume caused by reaction (1) and by the retrograde reaction that produced Srp. Because Srp replaces both Tlc and Ol in Ol-bearing rocks, but is absent from the Ol-free selvages to the composite veins, the Srp-producing reaction probably was
 | (3) |
The increase in rock volume caused by reactions (1) and (3) is the sum of (
Vs) for each reaction, where Vs is the solid molar volume of reaction. All corrections for the formation of Srp were small—0·2–5·1% of 1. Five samples from the schlieren facies contain small amounts of En that required an additional correction. The pair Tlc + Mgs is stable at or at a lower T than, and En is stable at or at a higher T than the Ol–Tlc–Mg–En isobaric invariant point in Fig. 6. Following Evans & Trommsdorff (1974), En is considered to have formed after reaction (1) by an increase in T and reaction at the P–T–XCO2 conditions of the isobaric invariant point. Under these conditions, the reaction is
 | (4) |
The measured amounts of En in the five samples were corrected for by running reaction (4) backwards to 4 = 0 and adjusting the measured amounts of Tlc, Ol and Mgs accordingly. All corrections for the formation of En were very small, 0·1–0·9% of 1. Calculated values of 1 are listed in Table 4 and illustrated in Fig. 7.
|
Along the line of the m-scale traverse at location 16,
1 = 0·72–1·88 mol/l—variation of a factor of 2·6 over 1 m (Fig. 7). Reaction (1) has occurred but not gone to completion in every sample along the traverse. Significant differences in 1 occur over distances of several centimetres. Considering samples collected from positions offset from the line of traverse as well, 1 varies by a factor of 3·5 within a volume of rock 3 m3. Values of 1 are not necessarily higher in Tlc-rich schlieren than in the Ol-rich matrix (cf. samples 16B, 16H and 16 M, Tables 1 and 4). The schlieren therefore did not develop simply from greater 1 than in surrounding rock, but are regions where Tlc/(Ol + Tlc) was elevated in the Ol–Tlc–Chl precursor prior to carbonation because of lower whole-rock (Mg + Fe)/Si. The range in measured 1 for the m-scale traverse is similar to the range for samples of the schlieren facies collected from other parts of the metaperidotite body (open and filled diamonds, Fig. 7). Metamorphic processes that controlled 1 along the m-scale traverse therefore are representative of those that affected the body as a whole. Values of 1 adjacent to the selvages of the composite veins (filled diamonds, Fig. 7) are not greater than those measured for samples far from the veins (open diamonds). In terms of reaction progress, the effects of vein formation do not extend beyond the vein selvages. Reaction (1), however, has gone to completion in the vein selvages themselves; measured values of 1 for the selvages (open squares, Fig. 7) indicate that the maximum value of 1 4 mol/l. Measured values of 1 in samples from the schlieren facies therefore correspond to 9–63% reaction. |
SPATIAL DISTRIBUTION OF FLUID COMPOSITION |
|---|
Rocks of the schlieren facies contain four proxies for metamorphic fluid composition: Xfo,Ol,
18OMgs, 18OOl and 13CMgs. The proxies are considered, rather than the corresponding fluid compositional variables themselves, because they are directly measured quantities not subject to uncertainties introduced by estimates of P and T and by activity–composition relations. If P and T were uniform across the metaperidotite body at all times during Barrovian regional metamorphism, spatial variations in fluid composition can simply be tracked by variations in the proxies.
Activities of components in Ol, Tlc, Mgs and fluid are related through the equilibrium constant for reaction (1),
 | (5) |
 | (6) |
 | (7) |
) demonstrates that all values of Xfo,Ol along the traverse are consistent within error of measurement with a single value whose best estimate is the weighted mean, 0·888 ± 0·001 (±95% confidence interval for the standard error). Correspondingly, Xfo,Ol records a single value of XCO2 within error of measurement. Measured values of Xfo,Ol along the m-scale traverse are similar to those measured in samples of the schlieren facies from other parts of the metaperidotite body (right-hand panel of Fig. 8). In addition, there are no large differences in Xfo,Ol and, hence, XCO2 between samples of the schlieren facies <10 cm from composite veins and those farther away. Some of the small differences in Xfo,Ol between samples along the m-scale traverse and samples from other parts of the body, however, are statistically significant. Whereas measurable differences in XCO2 did not occur during metamorphism over distances of 1 m or less, small differences in XCO2 < 0·01 did develop over the scale of the entire metaperidotite body (30 m or less).
|
The
18O of Mgs and Ol are proxies for 18O of fluid. With two exceptions, measured values of 18OMgs along the m-scale traverse are consistent with a single value (MSWD = 1·90) whose best estimate is 9·37 ± 0·06 (Fig. 9). The exceptions, samples 16L1 and 16 M, occur <10 cm from a composite vein and, like other samples of the schlieren facies collected near composite veins, experienced 18O-enrichment associated with formation of the vein (cf. Fig. 4 and right-hand panel of Fig. 9). Measured values of 18OOl along the m-scale traverse are consistent with a single value (MSWD = 1·50) whose best estimate is 4·62 ± 0·09 (Fig. 10). The O-isotope compositions of Mgs and Ol record uniform 18Ofluid along the m-scale traverse during Barrovian metamorphism within error of measurement. Although 18OMgs along the m-scale traverse is similar to that measured for samples of the schlieren facies from other parts of the metaperidotite body (Fig. 9), some of the small differences are statistically significant. Like XCO2, 18Ofluid therefore was uniform over distances comparable to the m-scale traverse but not over distances on the scale of the entire body.
|
|
The
13C of Mgs is a proxy for 13C of fluid. With the exception of three samples at the ends (A, L1, M), measured values of 13CMgs from the m-scale traverse are consistent with a single value (MSWD = 1·66), whose best estimate is –6·81 ± 0·06 (Fig. 11). Although values of 13CMgs along the m-scale traverse overlap with those of samples of the schlieren facies from other parts of the metaperidotite body, small but statistically significant differences occur. Like XCO2 and 18Ofluid, 13Cfluid was uniform over distances comparable to the m-scale traverse but not over distances on the scale of the entire body.
|
Taken together, data in Figs 7–11 firmly establish that fluid composition during Barrovian metamorphism was uniform within error of measurement in the same
3 m3 volume of rock in which progress of infiltration-driven reaction (1) varies by a factor of 3·5. |
INTERPRETATIONS OF cm-SCALE VARIATIONS IN REACTION PROGRESS |
|---|
Conventional interpretation
When a carbonation or decarbonation reaction is driven by infiltration of rock by chemically reactive fluid, reaction progress is proportional to the time-integrated fluid flux (Baumgartner & Ferry, 1991
; Ferry & Gerdes, 1998). Metre- to cm-scale variations in the progress of infiltration-driven reactions, as in Fig. 7, therefore are conventionally interpreted in terms of channelized fluid flow with elevated flow in the high- areas and reduced flow in the low- areas. The interpretation implies that rocks were physically and chemically isolated from each other at the cm–m scale over which the variations in 1 are observed. Assuming isochemical metamorphism (in the petrologic sense), Xfo,Ol decreases and XCO2 correspondingly increases with increasing 1 because of the fractionation of Fe and Mg among Ol, Tlc and Mgs (Fig. 12). If samples from the m-scale traverse were isolated systems, the different values of 1 should systematically correlate with differences in Ol composition; specifically, high- samples should contain Ol with relatively low Xfo,Ol and low- samples should contain Ol with relatively high Xfo,Ol. The predicted range in Xfo,Ol for the observed range in 1, 0·7–2·5 mol/l, is at least 0·02 (Fig. 12)—a range that would be unequivocally detected by microprobe analysis (Fig. 8). The absence of any statistically significant correlation between 1 and Xfo,Ol therefore suggests some other process accounts for the variations in 1 in Fig. 7. The process, in particular, must be consistent with spatially uniform Xfo,Ol and fluid composition.
|
and 
, and were computed for P = 7 kbar and T = 645°C from K1, assuming CO2–H2O fluid, calculated mineral compositions, and a–X relations described in text. Right-hand ends of curves correspond to New interpretation
Qualitative description and quantitative analysis
Spatial variations in the progress of an infiltration-driven reaction inevitably develop in a suite of rocks that experiences the same reaction if (a) at least one of the mineral reactants is a solid solution, (b) different rocks initially contain different amounts and/or compositions of the reactant mineral(s), and (c) fluid composition is the same at all times and in all samples during reaction (if, for example, it is homogenized by diffusion–dispersion). Samples of metaperidotite that record different values of
1 illustrate the process. Consider two rocks with Ol of the same composition prior to reaction (e.g. Xfo,Ol = 0·9215), but the first rock contains twice as much Ol (e.g. 21·2 mol/l) as the second (e.g. 10·6 mol/l). The remainder of the rocks is Chl ± Tlc but not Mgs. Fluid composition changes as reaction (1) proceeds, but at any one time it is the same in both rocks; Ol composition must be the same as well. At all times, 1 in the first rock therefore must be approximately twice 1 in the second in order to maintain the equality in Ol composition (Fig. 12). In this case, the difference in 1 between the two rocks develops because the rate of reaction (1), 
1/t, in the Ol-rich rock is approximately twice that in the Ol-poor rock on a volume basis. Alternatively, consider two rocks with the same amount of Ol prior to reaction (e.g. 21·2 mol/l) but one contains more Mg-rich Ol (e.g. Xfo,Ol = 0·9215) compared with the other (e.g. Xfo,Ol = 0·9000). If, for example, reaction (1) initiates in the first rock at 7 kbar and 645°C, fluid composition is buffered by reactants and products to XCO2 = 0·197 (Fig. 12). Fluid with XCO2 = 0·197, however, is in equilibrium with Ol with Xfo,Ol = 0·9, and reaction (1) does not proceed in the second rock. With continued infiltration and progress of reaction (1), Ol composition in the first rock eventually reaches Xfo,Ol = 0·9 (Fig. 12). With further infiltration, reaction (1) then proceeds in both rocks. If infiltration and reaction (1) cease when Xfo,Ol < 0·9 in both samples (e.g. Xfo,Ol = 0·888), 1 in the first rock will be larger than 1 in the second (Fig. 12). In the second case, differences in develop not so much from differences in reaction rate, but from differences in the duration of reaction. Specifically, all else being equal, reaction (1) proceeds longer in rocks that initially contain relatively Mg-rich Ol and for a shorter time in rocks that initially contain relatively Mg-poor Ol.
A systematic quantitative analysis of the process, specifically for progress of reaction (1) driven by infiltration of metaperidotite by CO2–H2O fluid, appears in Fig. 13. Rocks are considered composed of Ol, Tlc, Chl (0·289 mol/l) and 1 modal % other inert minerals (Chr, Po, Pn) prior to reaction. Initial amounts of Ol (
) are varied between 8 and 20 mol/l; Tlc content prior to reaction is the difference in volume between 1 l and the volumes of the other minerals. Mineral solid solutions are modelled with the formulas in Table 5. Initial compositions of Ol (
) are varied between 0·88 and 0·94 Xfo,Ol; compositions of the other minerals both prior to and at all times during reaction are specified by the composition of Ol and the Fe–Mg exchange constants in Table 5. The ranges in initial amounts and compositions of Ol are those appropriate to metamorphism of the carbonated metaperidotite body in Val d'Efra. Figure 13 illustrates the value of 1 needed to achieve a final Ol composition of Xfo,Ol = 0·888 (Fig. 8) as a function of the amount and composition of Ol prior to reaction (1). Plotting coordinates are chosen to linearize the relationship between , and 1. As expected, 1 at constant initial Ol composition is greater in rocks with greater initial amounts of Ol. At constant initial abundance of Ol (inclined solid contours), 1 is greater in rocks with Mg-richer Ol prior to reaction. No reaction occurs in any rock with 
. Significant differences in 1 can result from differences in Xofo,O1 as small as 0·01. Open and filled circles correspond to the initial amounts and compositions of Ol in all samples from the m-scale traverse, obtained by running measured values of 1, 3 and (where relevant) 4 backwards to zero and then computing the corresponding initial mineral abundances and (from mass balance of Fe and Mg) their initial compositions. The measured range in 1 for samples from the m-scale traverse (0·7–2·5 mol/l) can be completely explained by modest cm-scale variations in the amount (
) and composition (
) of Ol prior to reaction (1) and complete homogenization of fluid composition across the traverse, most plausibly by diffusion–dispersion.
|
and 
, relevant to carbonation of the metaperidotite body, and are calculated using model mineral formulas and Fe–Mg exchange constants in 
, a range in 
, and complete homogenization of fluid composition along the traverse during
|
Evidence against the conventional interpretation
There remains a remote possibility that rocks indeed were chemically isolated at the cm scale during metamorphism, that the differences in measured values of
1 (Fig. 7) represent different values of time-integrated flux independent of and 
, and that any resulting variations in Xfo,Ol after reaction are simply below the detection limits of electron microprobe analysis. The possibility was evaluated by a set of representative calculations that predict the composition of Ol that would result if rocks were chemically isolated along the m-scale traverse and if time-integrated flux, and hence 1, were completely unrelated to and . The measured value of 1 in each sample from the m-scale traverse was first assigned at random to one of the other 16 samples. Starting with the abundances and compositions of minerals in each sample prior to reaction (1), reaction (1) was then run forward by the new amount and the final Ol composition that developed was calculated. Results appear in Fig. 14. The model for chemically isolated samples predicts a range of final Xfo,Ol = 0·857–0·901—far larger than that observed (Fig. 8). Given a representative analytical precision of Xfo,Ol in the samples of metaperidotite (±0·0032, 1; the average of 1 uncertainties for samples from location 16 in Table 4), the set of calculated Ol compositions in Fig. 14 is inconsistent with a single value at a high degree of statistical significance (MSWD = 15·5). Thus, except for an unreasonably contrived set of initial conditions, any model of metamorphism of the metaperidotite based on chemical isolation of samples at a scale <1 m fails to account for both the large variations in measured 1 and the uniformity in measured Xfo,Ol (Figs 7 and 8). The new interpretation, based on spatially uniform XCO2 at the m scale during metamorphism, is further supported by evidence for spatial uniformity in 18Ofluid and 13Cfluid at the same scale as well (Figs 9–11).
|
Origin of differences in the amount and composition of Ol in Ol–Tlc–Chl schist
A full understanding of reaction progress in the metaperidotite requires an explanation of the process that controlled the initial amounts and compositions of Ol in Ol–Tlc–Chl schist prior to reaction (1). Variations in the amount and composition of Ol prior to reaction (1) were caused by variations in the bulk composition of the metaperidotite prior to Barrovian metamorphism. Many metaperidotites in the Central Alps, and the one in Val d'Efra in particular, do not have bulk compositions of normal igneous rocks (Trommsdorff & Evans, 1974
; Pfiffner, 1999). They have too low Ca contents to be typical oceanic peridotites (Dick et al., 1984) and too high Si contents to be dunites. The metaperidotite in Val d'Efra could have been either a peridotite that lost nearly all Ca or a dunite that gained Si by metasomatism, e.g. associated with serpentinization (Pfiffner, 1999). Both possibilities are modelled to explain the variations in the amount and composition of Ol in the Ol–Tlc–Chl schist prior to carbonation.
In the case of Ca-depleted peridotite, the parent rock is considered composed of 98 modal % Ol + En + diopside (Di), 1% Al–Cr Spl that reacts with Ol and En to form 0·289 mol/l Chl, and 1% minerals that remain inert during metasomatism and regional metamorphism (Chr, Po, Pn). Olivine, En and Di in peridotite have uniform compositions. The proportions of Ol and pyroxene (Px), however, are allowed to vary in the range Px/(Ol + Px) = 0–0·8 by volume. The relative proportion of En to Di is fixed and corresponds to that in average oceanic peridotite, Di/En = 0·175 by volume (Dick et al., 1984). The variations in Px/(Ol + Px) are intended to represent dm-scale alternations of Px- and Ol-rich layers observed in oceanic peridotite (Loney et al., 1971; Dick & Sinton, 1979; Boudier & Coleman, 1981). In the first stage of the reaction history, peridotite loses all Ca (at constant Mg, Fe and Si) and is hydrated to Ol–Tlc–Chl schist using Ca(Mg,Fe)Si2O6 as for the composition of Di and formulas in Table 5 for the other minerals. The exact sequence of reactions, that in nature would have involved Ctl/Lz, Atg, Brc, and possibly UHP minerals as intermediate reaction products (Trommsdorff & Evans, 1974; Pfiffner & Trommsdorff, 1998), is inconsequential. Amounts of Ol and Tlc in the model Ol–Tlc–Chl schist depend on Px/(Ol + Px) of the parent rock; Ol composition was computed from mass balance of Fe and Mg and the Fe–Mg exchange constants in Table 5. The partitioning of Fe and Mg between Ol and Di was taken as that for Ol and En (Table 5), as appropriate for the elevated T at which the peridotite parent originally formed (Loucks, 1996). During the second stage, reaction (1) proceeds in Ol–Tlc–Chl schist until Ol composition reaches 0·888 Xfo,Ol (Fig. 8). Different values of 1 develop at the end of the second stage of reaction, depending on and prior to reaction (1), that, in turn, depend on Px/(Ol + Px) of the parent peridotite. The range in measured values of 1 along the m-scale traverse (Fig. 7) are quantitatively reproduced for Xfo,Ol = 0·923 (a representative mantle value) and Px/(Ol + Px) in the range 0·05–0·50 by volume in the original peridotite (Fig. 15). There could have been variations in Xfo,Ol as well as in Px/(Ol + Px) in the peridotite, but these are not required by the data. For Xfo,Ol = 0·923 and Px/(Ol + Px) = 0·05–0·50 in the peridotite parent, predicted ranges in 
and 
of Ol–Tlc–Chl schist nearly match those inferred for samples along the m-scale traverse from measured modes and mineral compositions (12·4–19·9, 0·901–0·920, Fig. 13).
|
In the case of a silicified dunite, the parent rock is considered composed of 98 modal % Ol, 1% Al–Cr Spl that reacts with Ol and SiO2 to form 0·289 mol/l Chl, and 1% minerals that remain inert during metasomatism and regional metamorphism (Chr, Po, Pn). Olivine has uniform composition. The amount of SiO2 added to dunite is allowed to vary in the range 0–0·7 mol SiO2/mol Ol (in excess of what is required to react with Ol and Spl to form Chl). In the first stage of the reaction history, dunite is hydrated and silicified to Ol–Tlc–Chl schist using mineral formulas in Table 5. The exact sequence of reactions, that in nature would have involved other minerals as intermediate reaction products, is inconsequential. Amounts of Ol and Tlc in the model Ol–Tlc–Chl schist depend only on the amount of SiO2 added; Ol composition was computed from mass balance of Fe and Mg and the Fe–Mg exchange constants in Table 5. During the second stage, reaction (1) proceeds in Ol–Tlc–Chl schist until Ol composition reaches 0·888 Xfo,Ol (Fig. 8). Different values of
1 develop at the end of the second stage of reaction, depending on and prior to reaction (1), that, in turn, depend on the amount of SiO2 added to dunite. The range in measured values of 1 along the m-scale traverse (Fig. 7) is quantitatively reproduced for Xfo,Ol = 0·924 in the dunite (a representative mantle value) and the amount of SiO2 added = 0·04–0·43 mol/mol Ol (Fig. 16). There could have been variations in Xfo,Ol, as well as a range in amount of SiO2 added, but these are not required by the data. For Xfo,Ol = 0·924 in the dunite parent and addition of 0·04–0·43 mol SiO2/mol Ol, predicted ranges in 
and of Ol–Tlc–Chl schist almost exactly match those inferred for samples along the m-scale traverse from measured modes and mineral compositions (12·4–19·9, 0·901–0·920, Fig. 13).
|
Regardless of whether the carbonated ultramafic rock evolved from Ca-depleted peridotite, silicified dunite, Ca-depleted and silicified peridotite, or something else, the measured variations in
1 along the m-scale traverse therefore record nothing about the geometry or distribution of fluid flow during Barrovian regional metamorphism. The variations in 1 simply image spatial variations in the bulk composition of the altered ultramafic rock, (Mg + Fe)/Si in particular, prior to metamorphism.
A variation of the new interpretation
Significant variations in 1 can also be produced by infiltration of Ol–Tlc–Chl schist by CO2–H2O fluid, even if all samples contain the identical amount and composition of Ol prior to reaction, provided there is partial rather than complete homogenization of fluid composition among the samples. Consider Ol–Tlc–Chl schist with initial mineral abundances and Ol composition as indicated in Fig. 17. Chemically reactive fluid infiltrates some reference sample, reaction (1) then proceeds, and XCO2 is buffered by reactants and products. If there is imperfect CO2–H2O exchange by diffusion–dispersion between the reference sample and some other remote sample (e.g. 1–10 m away) through which there is no fluid flow, a gradient in XCO2 develops and Xfo,Ol will differ between the rocks. In a steady state involving imperfect CO2–H2O exchange between the samples, the difference in XCO2 and Xfo,Ol will be constant, and reaction (1) will proceed in the remote sample at a reduced rate. If reaction proceeds in the reference sample until Xfo,Ol = 0·888, Fig. 17 illustrates the calculated value of 1 in the remote sample as a function of the steady-state difference in Xfo,Ol. Variations in the degree of communication result in a range of 1 in the remote sample between zero (for no communication at all) and2·76 mol/l (for perfect communication). Variations in 1 similar in magnitude to those measured along the m-scale traverse therefore can be produced by imperfect chemical communication between samples that are in every other respect identical. The good evidence for homogenization of fluid composition along the m-scale traverse (Figs 8–11) rules out imperfect chemical communication as the explanation for variations in 1 along the traverse. Imperfect communication, along with differences in and , however, may explain differences in 1 between samples from the m-scale traverse and more remote samples that record a difference in Xfo,Ol (= XCO2), 18OMgs (= 18Ofluid) and 13CMgs (= 13Cfluid).
|
|
IMPLICATIONS |
|---|
Whenever an infiltration-driven reaction involves one or more mineral reactants that are solid solutions and fluid composition is homogenized by diffusion–dispersion over a distance greater than the thickness of lithological layering, layer-by-layer differences in the progress of the reaction inevitably develop if the layers differ in the initial amount and/or composition of the reactant mineral(s). Specific consideration of the carbonated metaperidotite body in Val d'Efra demonstrates that the process can produce variations in reaction progress of a factor of
2·6 over several decimetres. Other studies that interpreted cm-scale variations in reaction progress in terms of channelized fluid flow (e.g. Ferry, 1987) failed to adequately consider the significance of solid solutions and can no longer be considered correct. The results of this study thus resolve the apparent contradiction between cm-scale variations in progress of infiltration-driven reactions and isotopic evidence for homogenization of fluid compositions over a distance of >1 m in the same outcrop of regionally metamorphosed rock (Ferry, 1987; Bickle et al., 1997). The rate of CO2–H2O interdiffusion during regional metamorphism is very rapid (Wark & Watson, 2004), and the length scale of homogenization of fluid composition by diffusion–dispersion appears to be typically 1 m (Bickle et al., 1997; Ague & Rye, 1999; Ague, 2000, 2002, 2003; Evans et al., 2002). Therefore, cm-scale variations in the progress of infiltration-driven reactions involving solid solutions are better interpreted in terms of cm-scale variations in the initial amount and compositions of mineral reactants. Variations in reaction progress at the m scale (e.g. Ferry, 1994) likewise may be controlled more by initial variations in modes and mineral chemistry than by the channelization of reactive fluid flow during regional metamorphism.
If mineral reactants and products involved in an infiltration-driven reaction are either pure substances (e.g. calcite, quartz and wollastonite) or are fixed in composition by mineral equilibria (e.g. calcite and dolomite during the dolomite–periclase–calcite reaction), amounts and compositions of reactant mineral(s) exert no control on reaction progress. In this case, the mapped distribution of the progress of the reaction at the outcrop or larger scale indeed corresponds to the spatial distribution of time-integrated fluid flux (e.g. Ferry & Rumble, 1997; Ferry et al., 1998, 2001, 2002; Lackey & Valley, 2004). Specifically, high- regions image the location, size and geometry of channels for elevated flow, and low- regions image intervening regions with reduced flow.
A corollary of this study is that the geometry of fluid flow can never be determined at a length scale smaller than the characteristic distance over which fluid composition is completely homogenized by diffusion–dispersion. Because of the homogenization of geochemical tracers in the fluid by diffusion–dispersion, it is impossible, for example, to determine over the characteristic distance whether the physical mechanism of flow was along a single thin crack, was pervasive and uniform at the grain-size scale, or was something in between. The conclusion holds, regardless of whether an infiltration-driven metamorphic reaction involves solid solutions or pure substances and regardless of the chemical tracer used to investigate flow (e.g. reaction progress, stable or radiogenic isotope compositions, trace-element concentrations). Studies that determine the characteristic length scale of diffusion–dispersion (e.g. Bickle & Baker, 1990; Bickle et al., 1997; Ferry et al., 2001; Evans et al., 2002; Ague, 2003) are essential contributions because they define the smallest scale at which the geometry of the flow system can be determined using geochemical tracers. This smallest scale may differ depending on the specific tracer considered (Bickle et al., 1997; Ague, 2003).
New methods for direct inversion of spatial patterns of mineralogical, isotopic and geochemical alteration in terms of the regional-scale, 3-D pattern of reactive fluid flow (Wing & Ferry, 2002, 2005) require representative outcrop-scale estimates of fluid composition and the progress of infiltration-driven mineral–fluid reactions. Results of this study suggest that, with some obvious counterexamples (e.g. Rumble, 1978), fluid composition may be as surprisingly uniform at the outcrop scale during regional metamorphism as predicted by Ague (2000, 2002). For example, with the exception of extensively serpentinized samples 16L and 16M, the range of measured Ol compositions for all analysed samples from the metaperidotite body is Xfo,Ol = 0·885–0·899 (Table 4). The range in Xfo,Ol, in turn, corresponds to a range in XCO2 < 0·01. If fluid composition indeed is typically so uniform during regional metamorphism at the outcrop scale, average fluid composition can be adequately determined from measurements of one or a few samples per outcrop. Reaction progress, on the other hand, is much more variable, and average values correspondingly are more difficult to measure. The obvious solution of collecting and analysing several dozen samples per outcrop would be impractical for any regional investigation. This study suggests a more efficient alternative. If a representative outcrop were intensely investigated at low grade of metamorphism where infiltration-driven reactions did not occur, variations in the amount and composition of reactant minerals could be measured and average values along with their uncertainties accurately determined. At higher grades where reaction occurred in lithologically equivalent rocks, one or a few samples per outcrop could be used to calibrate the quantitative relationship between reaction progress and the amount and composition of reactants prior to reaction. An example specifically for the carbonated metaperidotite body in Val d'Efra is illustrated in Fig. 13. An average value of reaction progress and its range for an outcrop as a whole then could be inferred from the previously determined average values and ranges for the initial amounts and compositions of mineral reactants. Thus, quantitative studies of the 3-D geometry of reactive fluid flow on the regional scale may become more tractable.
|
ACKNOWLEDGEMENTS |
|---|
We thank Volkmar Trommsdorff and Bernard Evans for advice on the logistics of fieldwork in the area; Franco Barera for arranging our stay at Capanna Efra; the Museo Cantonale di Storia Naturale, Lugano, Switzerland, for permission to conduct the fieldwork; and Jay Ague, Bernard Evans, and Craig Manning for their thoughtful reviews. Sarah Carmichael assisted with preparation of field photographs. Research supported by grant EAR-0229267 from the Division of Earth Sciences, National Science Foundation.
* Corresponding author. Telephone: 410-516-8121. Fax: 410-516-7933. E-mail: jferry{at}jhu.edu
|
REFERENCES |
|---|
Ague, J. J. (2000). Release of CO2 from carbonate rocks during regional metamorphism of lithologically heterogeneous crust. Geology 28, 1123–1126.
Ague, J. J. (2002). Gradients in fluid composition across metacarbonate layers of the Wepawaug Schist, Connecticut, USA. Contributions to Mineralogy and Petrology 143, 38–55.[Web of Science]
Ague, J. J. (2003). Fluid infiltration and transport of major, minor, and trace elements during regional metamorphism of carbonate rocks, Wepawaug Schist, Connecticut, USA. American Journal of Science 303, 753–816.
Ague, J. J. & Rye, D. M. (1999). Simple models of CO2 release from metacarbonates with implications for interpretation of directions and magnitudes of fluid flow in the deep crust. Journal of Petrology 40, 1443–1462.[CrossRef][Web of Science]
Armstrong, J. T. (1988). Quantitative analysis of silicate and oxide minerals: Comparison of Monte Carlo, ZAF and phi-rho-z procedures. In: Newbury D. E. (ed.) Microbeam Analysis—1988. San Francisco: San Francisco Press, pp. 239–246.
Baumgartner, L. P. & Ferry, J. M. (1991). A model for coupled fluid-flow and mixed-volatile mineral reactions with applications to regional metamorphism. Contributions to Mineralogy and Petrology 106, 273–285.[CrossRef][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.
Bickle, M. J. & Baker, J. (1990). Advective-diffusive transport of isotopic fronts: an example from Naxos, Greece. Earth and Planetary Science Letters 97, 78–93.[CrossRef][Web of Science]
Bickle, M. J., Chapman, H. J., Ferry, J. M., Rumble, D. & Fallick, A. E. (1997). Fluid flow and diffusion in the Waterville limestone, south–central Maine: constraints from strontium, oxygen and carbon isotope profiles. Journal of Petrology 38, 1489–1512.[CrossRef][Web of Science]
Boudier, F. & Coleman, R. G. (1981). Cross section through the peridotite in the Semail ophiolite, southeastern Oman Mountains. Journal of Geophysical Research 86, 2573–2592.
Carpenter, M. A. & Ferry, J. M. (1984). Constraints on the thermodynamic mixing properties of plagioclase feldspars. Contributions to Mineralogy and Petrology 87, 138–148.[CrossRef][Web of Science]
Chayes, F. (1956). Petrographic Modal Analysis: An Elementary Statistical Appraisal. New York: John Wiley.
Coplen, T. B. (1988). Normalization of oxygen and hydrogen isotope data. Chemical Geology 72, 293–297.[Web of Science]
Coplen, T. B. (1996). New guidelines for reporting stable hydrogen, carbon, and oxygen isotope-ratio data. Geochimica et Cosmochimica Acta 60, 3359–3360.[CrossRef]
Dick, H. J. B. & Sinton, J. M. (1979). Compositional layering in alpine peridotites: Evidence for pressure solution creep in the mantle. Journal of Geology 87, 403–416.[Web of Science]
Dick, H. J. B., Fisher, R. L. & Bryan, W. B. (1984). Mineralogic variability of the uppermost mantle along mid-ocean ridges. Earth and Planetary Science Letters 69, 88–106.[CrossRef][Web of Science]
Evans, B. W. & Trommsdorff, V. (1974). Stability of enstatite + talc, and CO2-metasomatism of metaperidotite, Val d'Efra, Lepontine Alps. American Journal of Science 274, 274–296.[Abstract]
Evans, K. A., Bickle, M. J., Skelton, A. D. L., Hall, M. & Chapman, H. (2002). Reductive deposition of graphite at lithological margins in East Central Vermont: a Sr, C, and O isotope study. Journal of Metamorphic Geology 20, 781–798.[CrossRef][Web of Science]
Ferry, J. M. (1979). A map of chemical potential differences within an outcrop. American Mineralogist 64, 966–985.[Abstract]
Ferry, J. M. (1987). Metamorphic hydrology at 13-km depth and 400–550°C. American Mineralogist 72, 39–58.[Abstract]
Ferry, J. M. (1994). Overview of the petrologic record of fluid flow during regional metamorphism in northern New England. American Journal of Science 294, 905–988.
Ferry, J. M. (1995). Fluid flow during contact metamorphism of ophicarbonate rocks in the Bergell aureole, Val Malenco, Italian Alps. Journal of Petrology 36, 1039–1053.
Ferry, J. M. & Gerdes, M. (1998). Chemically reactive fluid flow during metamorphism. Annual Review of Earth and Planetary Sciences 26, 255–287.[CrossRef][Web of Science]
Ferry, J. M. & Rumble, D. (1997). Formation and destruction of periclase by fluid flow in two contact aureoles. Contributions to Mineralogy and Petrology 128, 313–334.[CrossRef][Web of Science]
Ferry, J. M., Sorensen, S. S. & Rumble, D. (1998). Structurally controlled fluid flow during contact metamorphism in the Ritter Range pendant, California, USA. Contributions to Mineralogy and Petrology 130, 358–378.[CrossRef][Web of Science]
Ferry, J. M., Wing, B. A. & Rumble, D. (2001). Formation of wollastonite by chemically reactive fluid flow during contact metamorphism, Mt. Morrison pendant, Sierra Nevada, California, USA. Journal of Petrology 42, 1705–1728.
Ferry, J. M., Wing, B. A., Penniston-Dorland, S. C. & Rumble, D. (2002). The direction of fluid flow during contact metamorphism of siliceous carbonate rocks: new data for the Monzoni and Predazzo aureoles, northern Italy, and a global review. Contributions to Mineralogy and Petrology 142, 679–699.[Web of Science]
Goldsmith, J. R. (1982). Plagioclase stability at elevated temperatures and pressures. American Mineralogist 67, 653–675.[Abstract]
Grond, R., Wahl, F. & Pfiffner, M. (1995). Mehrphasige alpine deformation und metamorphose in der nördlichen Cima-Lunga-einheit, Zentralalpen (Schweiz). Schweizerische Mineralogische und Petrographische Mitteilungen 75, 371–386.[Web of Science]
Heinrich, C. A. (1982). Kyanite–eclogite to amphibolite facies evolution of hydrous mafic and pelitic rocks, Adula nappe, Central Alps. Contributions to Mineralogy and Petrology 81, 30–38.[CrossRef][Web of Science]
Hewitt, D. A. (1973). The metamorphism of micaceous limestones from south-central Connecticut. American Journal of Science 273-A, 444–467.
Holland, T. J. B. & Powell, R. (1998). An internally consistent thermodynamic data set for phases of petrological interest. Journal of Metamorphic Geology 16, 309–343.[CrossRef][Web of Science]
Kohn, M. J. & Valley, J. W. (1994). Oxygen isotope constraints on metamorphic fluid flow, Townsend Dam, Vermont, USA. Geochimica et Cosmochimica Acta 58, 5551–5566.[CrossRef][Web of Science]
Kretz, R. (1983). Symbols for rock-forming minerals. American Mineralogist 68, 277–279.[Abstract]
Kyser, T. K. (1986). Stable isotope variations in the mantle. Reviews in Mineralogy 16, 141–164.[Abstract]
Lackey, J. S. & Valley, J. W. (2004). Complicated patterns of fluid flow during wollastonite formation in calcareous sandstones at Laurel Mountain, Mt. Morrison Pendant, California. Geological Society of America Bulletin 116, 76–93.
Loney, R. A., Himmelberg, G. R. & Coleman, R. G. (1971). Structure and petrology of the alpine-type peridotite at Burro Mountain, California, USA. Journal of Petrology 12, 245–309.
Loucks, R. R. (1996). A precise olivine–augite Mg–Fe-exchange geothermometer. Contributions to Mineralogy and Petrology 125, 140–150.[CrossRef][Web of Science]
Mahon, K. I. (1996). The new ‘York’ regression: application of an improved statistical method to geochemistry. International Geology Review 38, 293–303.
McCrea, J. M. (1950). On the isotopic chemistry of carbonates and a paleotemperature scale. Journal of Chemical Physics 18, 849–857.[CrossRef]
Nimis, P. & Trommsdorff, V. (2001). Revised thermobarometry of Alpe Arami and other garnet peridotites from the Central Alps. Journal of Petrology 42, 103–115.
Pfiffner, M. A. (1999). Genese der Hochdruckmetamorphen ozeanischen Abfolge der Cima-Lunga einheit (Zentralalpen). Ph.D. thesis, ETH Zürich.
Pfiffner, M. & Trommsdorff, V. (1998). The high-pressure ultramafic–mafic–carbonate suite of Cima Lunga–Adula, Central Alps: Excursions to Cima di Gagnone and Alpe Arami. Schweizerische Mineralogische und Petrographische Mitteilungen 78, 337–354.[Web of Science]
Rumble, D. (1978). Mineralogy, petrology, and oxygen isotope geochemistry of the Clough Formation, Black Mountain, New Hampshire, USA. Journal of Petrology 19, 317–340.
Rumble, D. & Spear, F. S. (1983). Oxygen-isotope equilibration and permeability enhancement during regional metamorphism. Journal of the Geological Society, London 140, 619–628.
Rumble, D., Oliver, N. H. S., Ferry, J. M. & Hoering, T. C. (1991). Carbon and oxygen isotope geochemistry of chlorite-zone rocks of the Waterville limestone, Maine, USA. American Mineralogist 76, 857–866.[Abstract]
Sharma, S. D., Patil, D. J. & Gopalan, K. (2002). Temperature dependence of oxygen isotope fractionation of CO2 from magnesite–phosphoric acid reaction. Geochimica et Cosmochimica Acta 66, 589–593.[CrossRef][Web of Science]
Sharp, Z. D. (1990). A laser-based microanalytical method for the in situ determination of oxygen isotope ratios of silicates and oxides. Geochimica et Cosmochimica Acta 54, 1353–1357.[CrossRef][Web of Science]
Todd, C. S. & Engi, M. (1997). Metamorphic field gradients in the Central Alps. Journal of Metamorphic Geology 15, 513–530.[CrossRef][Web of Science]
Trommsdorff, V. & Evans, B. W. (1974). Alpine metamorphism of peridotitic rocks. Schweizerische Mineralogische und Petrographische Mitteilungen 54, 333–352.
Valley, J. W., Kitchen, N., Kohn, M. J., Niendorf, C. R. & Spicuzza, M. J. (1995). UWG-2, a garnet standard for oxygen isotope ratios: strategies for high precision and accuracy with laser heating. Geochimica et Cosmochimica Acta 59, 5223–5231.[CrossRef][Web of Science]
Wark, D. A. & Watson, E. B. (2004). Interdiffusion of H2O and CO2 in metamorphic fluids at 490 to 660°C and 1 GPa. Geochimica et Cosmochimica Acta 68, 2693–2698.[CrossRef][Web of Science]
Wing, B. A. & Ferry, J. M. (2002). Three dimensional geometry of metamorphic fluid flow during Barrovian regional metamorphism from an inversion of combined petrologic and stable isotopic data. Geology 30, 639–642.
Wing, B. A. & Ferry, J. M. (2005). Magnitude and geometry of reactive fluid flow from direct inversion of spatial patterns of mineralogical, geochemical, and isotopic alteration. American Journal of Science (in press).
Yui, T.-F., Rumble, D. & Lo, C.-H. (1995). Unusually low 18O ultra-high pressure metamorphic rocks from the Sulu Terrain, eastern China. Geochimica et Cosmochimica Acta 59, 2859–2864.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. M. Ferry The role of volatile transport by diffusion and dispersion in driving biotite-forming reactions during regional metamorphism of the Gile Mountain Formation, Vermont American Mineralogist, August 1, 2007; 92(8-9): 1288 - 1302. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. A. Wing and J. M. Ferry Magnitude and geometry of reactive fluid flow from direct inversion of spatial patterns of geochemical alteration Am J Sci, May 1, 2007; 307(5): 793 - 832. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. C. Penniston-Dorland and J. M. Ferry Development of spatial variations in reaction progress during regional metamorphism of micaceous carbonate rocks, Northern new England Am J Sci, September 1, 2006; 306(7): 475 - 524. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||