Contributions to Mineralogy and Petrology

Contrib Mineral Petrol (1988) 98:28 32

9 Springer-Verlag 1988

High integrated fluid/rock ratios during metamorphism at Naxos: evidence from carbon isotopes of calcite in schists and fluid inclusions Rob Kreulen Institute of Earth Sciences, University of Utrecht, P.O. Box 80.021, NL-3508 TA Utrecht, The Netherlands

Abstract. Calcite in schists o f the m e t a m o r p h i c complex at N a x o s is depleted both in 13C and in 180 with respect to massive marbles. This effect is attributed to isotope exchange with circulating CO2-rich fluids, which had an Xco2 > 0.5 according to fluid inclusions. The carbon isotopic composition o f the calcites is close to equilibrium with fluid inclusion CO2 at m e t a m o r p h i c temperatures. Mass balance calculations assuming initial fi13C values o f 0 for calcite and - 5 for the fluid, give integrated fluid/rock volume ratios between 0.1 and 2.0. Such high fluid/rock ratios are supported by observations on the distribution of C O 2 / H 2 0 ratios o f fluid inclusions, carbon isotopic compositions o f fluid inclusion CO2 and oxygen isotope systematics o f silicates.

sites in the A d i r o n d a c k s N.Y., where an area o f 3000 k m 2 o f anorthosites was isotopically altered from initial values o f 6-7.5%0 to values of 10.5_+1%o. F e r r y (1986) reviewed the use o f petrologic d a t a to m o n i t o r reaction progress in fluid-rock interaction. He concluded that petrologic evidence for chemical interaction o f m e t a m o r p h i c rocks with at least one to five rock volumes is common. A detailed isotope study o f N a x o s by Rye et al. (1976) showed evidence for large scale oxygen isotope exchange with a deepseated reservoir. The present study is an attempt to estimate the a m o u n t of fluid involved, independently of the oxygen isotopic evidence.

Rocks and fluids

Introduction It is now well established that high integrated fluid/rock ratios (total mass o f fluid integrated over time) occurred a r o u n d m a n y shallow intrusions (Taylor 1977). The hot intrusions caused convective circulation of groundwater in the surrounding country rocks, and many ore deposits related to shallow intrusions p r o b a b l y formed from such circulating meteoric groundwaters (Taylor 1974). M e t a m o r p h i c rocks were formed deeper in the Earth's crust where thermal gradients are usually less steep, where fluid pressure equals lithostatic pressure so that the large density contrast between fluid and rock inhibits d o w n w a r d fluid flow (Walther and Orville 1982), and where the availability o f a large mass o f fluid is less obvious. One would, therefore, predict that high integrated fluid/rock ratios in m e t a m o r p h i c rocks were rare. According to oxygen isotope studies, however, this prediction is wrong. Some m e t a m o r phic rocks behaved as closed systems. F r e y et al. (1976), for instance, concluded that no significant external oxygen was a d d e d during m e t a m o r p h i s m o f the M o n t e R o s a granite and that the system merely "stewed in its own juices". M o s t high grade m e t a m o r p h i c terrains, however, show a tendency for fi 180 to a p p r o a c h igneous values. Such isotope systematics were first observed by Garlick and Epstein (1966) in Dutchess County, N.Y. and are generally attributed to isotope exchange with a large reservoir o f fluids from a deep-seated source. One o f the most convincing examples o f f i l s o buffering by pervasive fluids was described by Taylor (1964, 1969) for m e t a m o r p h i c anortho-

Naxos is the largest island of the Greek Cyclads and consists of a central migmatite core surrounded by a series of schists and marbles. Marbles constitute about 50% of the rocks (Fig. 1). The metasediments are more or less concentrically arranged around the migmatite core. Temperatures of metamorphism, deduced from mineral assemblages in various rock types, decrease rapidly from 700 ~ C in the central part of the migmatite core to slightly below 400 ~ C in the SE part of the island. This drop in temperature occurs over a horizontal distance of not more than 15 km. Pressures were about 4-5 kb. A detailed description of the area and PIT estimates were given by Jansen and Schuiling (1976). Figure 1 shows the following six isograds: the appearance of corundum in meta-bauxites, the appearance of biotite in pelitic schists, disappearance of chloritoid in iron-rich pelitic schists and meta-bauxites, the appearance of sillimanite in pelitic schists, disappearance of kyanite, and first melting. In the SE part of the island (farthest away from the migmatite core) the mineralogy is still dominated by an earlier high-pressure-type of metamorphism (Mr). Towards the migmatite core, however, overprinting becomes stronger and from the + biotite isograd onwards the mineralogy of the rocks is essentially M2. Andriessen (1978) dated the M1 metamorphism at 40-50 Ma and the M2 metamorphism around 23 Ma. The composition of the fluid phase was studied in fluid inclusions (Kreulen 1980). Most of this work was done on conformable quartz lenses, as this material is commonly very rich in fluid inclusions and allows the extraction of sufficient fluid for isotope analyses. The quartz lenses are thought to contain representative samples of metamorphic fluids because similar inclusion assemblages (albeit less abundant) occur in matrix quartz in the schists and in minerals such as kyanite, andalusite, garnet, feldspar, dolomite, calcite, tremolite, corundum that were formed during metamorphism. Most fluid inclusions (both primary-looking and secondary) are CO2rich. Optical estimates on 1000 inclusions in fifty samples from all over the area show that about 70% of the inclusions have a composition 60-90 mole% CO2 (the balance being water). The

29 25~50'

Table 1. Calcite in schists: weight percentages, isotopic composi-

25~ 3

4

5 KM

tions and calculated fluid/rock ratios Sample no

Isograd temp. ~

Weight % 5 lSCcc calcite %o

618Occ %o

Fluid/rock vol. ratio

61 62 83-6 65 155-1 156-1 67 72 73-6 170 169 87 77 78 140-2 80 162-2 1 143-2 19 30 93 89 98 150-1 102

380 380 380 385 385 385 390 390 390 390 420 425 435 435 460 500 500 520 520 530 575 605 610 630 650 700

2 7 17 32 8 16 11 7 14 3 16 6 43 30 61 16 0.5 4 36 11 1 14 1 1I 12 1

20.9 15.9 20.l 19.7 24.2 24.1 18.8 16.0 25.6 17.6 14.4 23.7 23.1 24.3 24.9 25.1 24.8 18.4 24.1 20.3 20.1 17.1 17.5 16.1 9.0 :11.2

0.098 0.34 0.12 0.95 0.70 2.0 0.26 0.31 0.094 0.084 1.8 0.051 0.49 8.5 1.1 0.29 2.1 0.13 0.16 0.15 0.057 0.10 0.43 0.0053

--5.6 -5.6 -2.2 -4.8 --6,3 -6.6 -4.4 + 1.7 -4.3 -4.9 - 1.8 --7.1 -0.5 - 3.7 -6,7 --6.0 --8.7 -6.1 -5.8 -3.1 -6.8 --2.9 --5.8 --2.7 -5.1 - 1.8

TECT~N[C ~NTACI

Fig. 1. Geological map of Naxos showing rock types and isograds 6180

%0 25

fluid composition does not change with metamorphic grade, and in most cases it does not change with lithology either. This suggests that the schists generally behaved as an open system where most of the fluid was pervasive. Some exceptions were found where the (chemical and/or isotopic) compositions of the fluid inclusions are related to the local mineralogy. The carbon isotopic composition of fluid inclusion CO2 was studied by Kreulen (1980). A wide range of 613C values was found from - 1 6 to + 5%0. The positive values, however, occur only in siliceous dolomites and values below -5%0 are restricted to a narrow graphite-rich zone. If these samples are omitted, the resulting range is - 5 to -i%o, again independent of metamorphic grade. This range of --5 to -1%o is characteristic for more than 90% of the schists on Naxos. For a more detailed description of the fluid inclusion work the reader is referred to Kreulen (1980).

84

O

massive

0 0 0 O0 0

o

o 20

marbles

9

o

o

o o

o o

o ~o

o 15

10

-8

-6

-4

-2

0

2 6~3C

%0

Fig. 2. Isotopic compositions of calcite in schists compared with massive marbles from Naxos

Isotopic composition of calcite in schists Some o f the m e t a m o r p h i c schists at N a x o s contain calcite. The isotopic compositions o f the calcites are shown in Table 1 and Figs. 2, 3 and 4. Figure 2 compares calcite in schists with the data on massive marbles from Naxos obtained by Rye et al. (1976). Rye et al. found that the marbles are often depleted in tsO and 13C near the marble/schist contacts and attributed this to isotope exchange with m e t a m o r phic fluids that m o v e d through the schist areas. Figure 2 shows that the 513C of calcite in schists is significantly lower than in massive marbles. 5 ~so shows a similar behaviour although there is some overlap between the schists and the marbles. Thus it seems that the mechanism o f

isotope exchange causing the ~sO and 13C depletions observed by Rye et al. (1976) also affected calcite dispersed in the schists. If calcite in the schists derived (part of) its isotopic composition by isotope exchange with m e t a m o r p h i c fluids, it seems pertinent to compare the isotopic compositions o f calcite with those of the fluid inclusions. Fluid inclusion CO2 in more than 90% o f the schists has 513C values between - 5 and - 1 (see above). Calcite in equilibrium with this CO2 will be ab o u t 2.5%~ lighter (Bottinga 1969) and have 513 C values between - 7 . 5 and - 3 . 5 . Figure 3 shows

30

S13c %0

calcite 9 fluid i n c l u s i o n s

2

o

0

-2

9

. . . . . . . . . . . . . . . . . Q.... -4

.....

9

o

8

8

o

9

g

0 8

0

o

9

. . . . . . . . . . .

o

~176

-6

9

0

~

o

9

o

9

.....................................

8

9

oO

o~

..................

-8

400

500

600

temperature

of

700

metamorphism

Fig. 3. 13C of calcite in schists and fluid inclusion CO/ versus metamorphic grade. The fluid inclusion data (taken from Kreulen 1980) have been recalculated to calcite in equilibrium with CO2 under metamorphic conditions

CO2/calcite ratios depend on: 1) the amount of fluid moving through the particular rock site (related to local variations in permeability), 2) the fluid composition (CO2/H20 ratio), 3) the concentration of calcite in the schist. Variations in permeability (1) may well have played a role. In general one would expect to find the lower permeabilities (for pervasive flow) in the lower grade rocks because these contain more (relatively impermeable) marbles and also because temperature increases the rock permeability. Fluid flow in these lower grade rocks may be more channeled along fractures, bedding planes, etc. and consequently have less effect on the 6~3C of dispersed calcite. Variations in fluid composition (2) were probably of minor importance, considering the results on fluid inclusions. The effect of the concentration o f calcite in the schist (3) is tested in Figure 4, where the 6 ~ac of calcite is plotted against its concentration in the schist. Indeed, Fig. 4 suggests a tendency for 613 C to decrease with decreasing calcite concentration and to approach values of - 5 to -8%o at low calcite contents. These values are exactly the ones to be expected if calcite equilibrated with a fluid of the isotopic composition found in the fluid inclusions.

S13C %0 0

o

I o

-2

-4

~

Integrated fluid rock ratios

0,1 0.3

o o

0.5

o

1.0 2.0

-6

-8

10

20

30

40

weight

% calcite

Fig. 4. 13C versus weight % calcite in schists. Curves for fluid/rock volume ratios of 0.1, 0.3, 0.5, 1 and 2 were calculated under assumptions explained in the text

both the fi*3C of calcite in schists and the fluid inclusion data plotted versus metamorphic grade. The fluid inclusion data are from Kreulen (1980); samples from the graphiterich zone, siliceous dolomites and pegmatites are omitted. Carbon isotope fractionation between CO2 and calcite (Bottinga 1969) at metamorphic temperatures has been subtracted from the fluid inclusion data so that the resulting values represent calcite in equilibrium with the trapped fluids. The values deduced from fluid inclusions agree well with the actually measured calcite data. This suggests that the CO2-rich fluid and the calcite were in good isotopic communication with each other and that the amounts of fluid were sufficient to modify the 613C of the calcite. Calcites with a higher 61ac probably retained part of their original (sedimentary) composition. This may be either because the isotope exchange was too slow to approach equilibrium (the highest fi~3C values occur in the lowest grade rocks) or because in those cases the CO2/calcite ratios were lower.

Calculations based on 613 C of calcite

The amount of fluid that must have equilibrated with calcite in order to produce the observed depletions in 13C can be obtained by a simple mass balance calculation, provided that the initial 613C of the calcite and the fluid are known. Calculations based on initial values of 0 for calcite and - 5 for the fluid give the results in Table 1. An initial value of 0 was chosen because this is a common value for marine carbonates. Actually, most of the massive marbles at Naxos have a slightly higher 613C (between 0 and + 3%o, Fig. 2). An initial calcite value of + 1 or + 2 would give somewhat higher integrated fluid/rock ratios. Much more critical (especially at high fluid/rock ratios) is the choice of the initial 613C of the fluid. The value of --5%0 corresponds to the lowest values for fluid inclusion CO2 in the schists (excluding samples from the graphite-rich zone where the isotopic composition of C O / i s clearly affected by the presence of graphite). Such CO2 of - 5%o is considered to have suffered least from interaction with carbonates in schists or in marbles and be a reasonable estimate of the isotopic composition before CO2-calcite interaction. Lower initial COz values will result in lower integrated fluid/rock ratios; higher initial CO2 values will increase the calculated fluid/ rock ratios until, at values higher than - 4 , several of the calcites have too low a 613C to have equilibrated with the fluid. This also limits the choice of the initial 6~3C value of CO2. The integrated fluid/rock volume ratios in Table i were calculated assuming a fluid composition of 50 mole% CO2 and fluid densities corresponding to the metamorphic conditions. The following expressions were combined: (1) The volume of metamorphic fluid containing 1 mole

CO2 : Vnuia ~- 44/0.9 + 18/1 [cm 3]

(molar weights devided by densities of CO2 and H20). (2) The volume of rock containing 1 mole CaCO3 :

31 Vrock--~100 * 100/(2.5 * wt% calcite) [cm 3] (100 = molar weight calcite; 2.5 = rock density). (3) The mass balance equation: mole CO2 * (c~13Ccalcite _}_2.5 + 5) = - mole calcite* (813Cca~cUe__0) where --5 and 0 are the initial 8'3C of fluid and calcite, and 2.5 the isotope fractionation between CO2 and calcite. It must be stressed here that the above calculation of integrated fluid/rock ratios stands up on its own and does not depend on models explaining the source of the fluids. Rye et al. (1976) and Kreulen (1980) presented evidence for an external source of the fluids, either in the crustal rocks below the migmatite or in the upper mantle. Although an external source seems most likely, it can not be completely ruled out that the fluids originated within the presently exposed metamorphic rocks by mixing of a marble source and a graphite source. The result of the fluid/rock ratio calculations will then be the same as for externally derived fluids, unless extensive recycling occurred over large distances. This is because the amount of fluid of a certain isotopic composition that is needed to balance the change in isotopic composition of the calcite is the same in both cases. The origin of the fluids was discussed in detail by Rye et al. (1976) and by Kreulen (1980). Most of the integrated fluid/rock volume ratios calculated in Table 1 are between 0.1 and 2.0. Variations in fluid/rock ratio probably reflect differences in permeability of the rocks and distribution of fractures which channeled the fluid flow. Only pervasive fluid flow on a grain size scale permits isotope exchange with calcite dispersed in the schists and leaves its traces. The fluid/rock ratios are also shown in Fig. 4 where "fluid/rock ratio curves" have been constructed. It is difficult to asses the precision of the obtained fluid/rock ratios. The results depend to a large extent on the assumption of the initial cs13C of calcite and the initial 813C of the fluid. As to the assumption of the initial 813C of the calcite, it can not be completely ruled out that the variation in 813C of the calcite is in part a primary sedimentary feature and that the low 813C calcites reflect a larger proportion of land-derived carbon than the high 61ac calcites. Similar mass balance calculations based on the 6180 of the calcite can not be used to confirm the obtained fluid/rock ratios because the calcites represent only a small part of the oxygen reservoir of the rocks. As pointed out by Rye et al. (1976), the original 81so of the pre-metamorphic rocks is not known and may in fact have varied from place to place. Due to uncertainties mainly in the initial 613C of the calcite and the initial 8 ' 3 C of the fluid, the calculated integrated fluid/rock ratios must probably be regarded as order of magnitude approximations. The concept that the carbon isotopic composition of calcite is mainly dependent on isotope exchange with COz-rich fluids and gives valid information on integrated fluid/rock ratios, is supported by the fact that other observations indicate similar high integrated fluid/rock ratios. Although probably each of these indications could also be explained by other hypotheses, taken together they make a convincing case.

Other indications.for high integrated fluid/roclc ratios I. Dehydration reactions are generally favoured at low grades and decarbonation reactions at higher grades of

metamorphism, and are obviously dependent on lithology. Such variations do, however, not show up in the fluid inclusions. The CO2/H20 ratios in fluid inclusions are largely independent of lithology and metamorphic grade, which suggests that the amount of externally derived fluid is larger than the amount of fluid produced within the metamorphic system. Walther and Orville (1982) calculated that H20 and COz produced during devolatilization of an "average pelite" will occupy about 12 volume % of the rock at 500 ~ C and 5 kb, and that their molar ratio will be about 5:2. If the amount of externally derived fluid is larger than this, this suggests that the integrated fluid/rock ratios calculated from the isotopic composition of calcite in schists are at least of the right order of magnitude. 2. The c~'3C of fluid inclusions in most schists fall in the range of - 5 to -1%o. It was shown (Kreulen 1977, 1980) that the siliceous dolomites at Naxos produced CO2 with fi13C values between + 2 and + 5%0. From the observation that the CO2 produced during extensive decarbonation of siliceous dolomites has no measurable effect on the 813C of fluid inclusions in the adjacent schists (not even within a distance of 1 m), it follows that the mass of C02 moving through the schists must have been much larger than the mass of CO2 produced by decarbonation reactions in the siliceous dolomites (i.e., fairly large). 3. Rye et al. (1976) found that 8180 values of quartz from schists and conformable quartz lenses decrease from about 19%o in the least metamorphosed rocks to about 9%0 in the migmatite. They concluded that "part of the trend may be the result of original 81sO distribution in the pre-metamorphic rocks but more likely it indicates that the rocks in the schist-rich zones exchanged enormous quantities of oxygen with a reservoir external to the metamorphic system" and "evidently large quantities of fluid from a deepseated source passed through the schist-rich zones in the high grade parts of the metamorphic complex or were recycled between the schist-rich zones and the deep-seated source". Conclusions

The integrated fluid/rock volume ratios between 0.1 and 2.0 obtained from the carbon isotopic composition of calcite in schists compare well with other indications of fluid/ rock ratios. Considering the uncertainties in the initial fi 13C values of calcite and the fluid (they may well have varied from place to place), the calculated fluid/rock ratios are order of magnitude approximations. An important implication of such high integrated fluid/rock ratios was discussed by Schuiling and Kreulen (1979) who concluded that the heat transported by a flux of hot fluids from a deep source must have had a marked effect on the temperature of the rocks, and may in fact have been the sole cause of metamorphism.

Acknowledgements. I am grateful to J.A.N. Meesterburry for technical and analytical assistance. The mass spectrometer was partly financed by the Netherlands Organization for the Advancement of Pure Research (ZWO). The paper has benefited from critical remarks by Bruce Yardley. References

Andriessen PAM (1978) Isotopic age relations within the polymetamorphic complex of the island of Naxos (Cyclades, Greece). Ph D Thesis Univ Utrecht

32 BottMga Y (1969) Calculated fractionation factors for carbon and hydrogen isotope exchange in the system calcite-carbondioxidegraphite-methane-hydrogen-watervapor. Geochim Cosmochim Acta 33:49 64 Ferry JM (1986) Reaction progress: a monitor of fluid-rock interaction during metamorphic and hydrothermal events. In: Walther JV, Wood BJ (eds) Fluid-rock interactions during metamorphism. Advances in physical geochemistry vol. 5. Springer, New York Berlin Heidelberg Tokyo, pp 6(~88 Frey M, Hunzicker JC, O'Neil JR, Schwander HW (1976) Equilibrium-disequilibriumrelations in the Monte Rosa granite, Western Alps: petrological, R b - S r , and stable isotope data. Contrib Mineral Petrol 55:147-179 Garlick GD, Epstein S (1967) Oxygen isotope ratios in coexisting minerals of regionally metamorphosed rocks. Geochim Cosmochim Acta 31:181 214 Jansen JBH, Schuiling RD (1976) Metamorphism on Naxos: petrology and geothermal gradients. Am J Sci 276:1225-1253 Kreulen R (1977) CO2-rich fluids during regional metamorphism on Naxos, a study on fluid inclusions and stable isotopes. Pb D Thesis Univ Utrecht Kreulen R (1980) COz-rich fluids during regional metamorphism on Naxos (Greece): carbon isotopes and fluid inclusions. Am J Sci 280 : 745-771 Rye RO, Schuiling RD, Rye DM, Jansen JBH (1976) Carbon,

hydrogen and oxygen isotope studies of the regional metamorphic complex at Naxos, Greece. Geochim Cosmochim Acta 40:1031-1049 Schuiling RD, Kreulen R (1979) Are thermal domes heated by CO;-rich fluids from the mantle? Earth Planet Sci Lett 43 : 298-302 Taylor HP (1964) Isotopic evidence for large scale oxygen exchange during metamorphism of Adirondack igneous rocks (abstract). Geol Soc Am Spec Pap 76:163-164 Taylor HP (1969) Oxygen isotope studies of anorthosites with special reference to the origin of bodies in the Adirondack Mountains, New York. In: Origin of Anorthosites, NY State Mus Sci Serv Memo 18:111 134 Taylor HP (1974) The application of oxygen and hydrogen isotope studies to problems of hydrothermal alterations and ore deposition. Econ Geol 69 : 843-883 Taylor HP (1977) Water/rock interactions and the origin of H20 in granitic batholiths. J Geol Soc 133:509 558 Walther JV, Orville PM (1982) Volatile production and transport in regional metamorphism. Contrib Mineral Petrol 79:252-257 Received May 15, 1987 / Accepted November 9, 1987 Editorial responsibility: J. Touret