Gwchimico
d Cosmockimica
Acla Vol. 54, pp. 641-65
Copyright 0 1990 Pergamon Press plc.Printedin
0016-7037/90/$3.00
+ .OO
I
U.S.A
Methane in fluid inclusions from granulites: A product of hydrogen diffusion?*
DONALDL. HALL’ and ROBERTJ. BODNAR~
‘Department of Earth Sciences, University of California, Riverside, CA 92521, USA
*Department of Geological Sciences, Virginia Polytechnic Institute & State University, Blacksburg, VA 2406 1, USA
(Received April I, 1989; accepted in revisedform December 18, 1989)
Abstract-Many
fluid inclusion studies of granulite grade rocks reveal the presence of CO*-rich inclusions
that appear to have been trapped near the peak of metamorphism. Final melting temperatures of CO2
[Tm(C02)] reported for these inclusions are often below the CO* triple point of -56.6”C, and some are
below -60°C. This freezing-point lowering is usually attributed to the presence of a second volatile
component, such as CH4, and the presence of CH4 has been confirmed in some cases by Raman or mass
spectroscopy. A CO* melting temperature near -56.6”C is commonly offered as evidence that the inclusions
contain “nearly pure CO2”. However, significant amounts of CH4 may be present but cause seemingly
insignificant freezing point depressions.
C-O-H fluid speciation calculations for conditions representative of granulite facies metamorphism
indicate that CH4 may comprise a significant portion of peak metamorphic fluids when graphite is present,
but it is never a significant species in C02-rich fluids. In most cases the amount of CH4 should be virtually
undetectable with microthermometric or spectroscopic techniques. Only in aqueous fluid inclusions can
CO2 and CH4 both be significant species. Post-trapping speciation changes to fluid inclusions at constant
mass cannot account for the reported compositions unless graphite precipitates in the inclusions. Thus,
the observed compositions require post-trapping compositional changes due to loss or gain of components.
We have modeled variations in the fugacities of molecular fluid species in inclusion and matrix fluids
during uplift from 6 kb and 8OO”C, assuming hypothetical uplift pressure-temperature paths which are
concave or convex toward the temperature axis (T-concave and T-convex, respectively). Our results
suggest that for rocks buffered at JO, within one log unit of the fayalite-magnetite-quartz equilibrium
(FMQ f I), most uplift paths result in externalfH2 overpressures of bars to tens of bars at temperatures
> 400°C. The highest overpressures are generated during T-concave uplift. Compositional changes resulting
from equilibration of such gradients, via diffusive addition of hydrogen to peak metamorphic fluid inclusions and concomitant reduction of CO2 by conversion to CH4 and H20, are consistent with the
compositions of fluid inclusions reported from granulite terranes.
Previous workers have postulated that C02-rich fluid inclusions in granulites could originate from
post-trapping diffusive loss of Hz0 from H20-CO2 inclusions in response to an fHZO gradient between
the inclusion and matrix fluids. The results of the present study suggest that for fluids buffered by FMQ
+ 1 this is possible only if ( 1) uplift is T-convex and the matrix fluid composition remains nearly constant,
or (2) the matrix fluid evolves toward relatively H20-poor compositions. The latter could occur if influx
of C02-rich fluids occurs during uplift.
INTRODUCI’ION
such as the upper mantle (HARRIS et al., 1982), or from devolatilization of deeply subducted continental shelf sediments
(HODGES et al., 1982; DRURY et al., 1984), mafic intrusions
(TOURET, 1971; WELLS, 1979) or deeply buried carbonates
(GLASSLEY,1983), (3) scavenging of Hz0 by retrograde hydration reactions (LAMB et al., 1987a; VAN REENEN and
HOLLISTER, 1987), and (4) preferential loss of Hz0 by diffusion out of fluid inclusions in response to differential pressure during uplift (ROEDDER, 1981). The first two mechanisms imply that the fluid inclusions are syn-metamorphic
while the latter two hypotheses imply, respectively, that trapping of fluid inclusions post-dated metamorphism or that
post-trapping compositional changes have altered the original
inclusions. The distinction between syn- and post-metamorphic trapping of fluid inclusions is crucial to any model proposed for granulite formation. Unfortunately, most inclusions
in granulites are texturally secondary. Inclusion paragenesis
is based mainly on comparisons of inclusion compositions
and densities with pressure, temperature, and fluid compositions indicated by mineral equilibria, although textural fea-
THE DISCOVERY
OF C02-rich fluid inclusions in granulites
(TOURET, 1970, 197 1; see also ROEDDER, 1984, pp. 375
380) has contributed significantly to models attempting to
explain the origin of these rocks while at the same time presenting a conundrum. The presence of these inclusions coupled with mineral assemblages suggesting low water fugacities
has been used to infer a C02-rich fluid phase during metamorphism. The origin(s) of these carbonic fluids are the subject of much debate. Mechanisms that have been proposed
to explain the CO*-rich fluids include (1) partitioning of the
aqueous portion of the fluid into a partial melt, which is
subsequently removed from the system leaving a C02-rich
residual fluid (QUENSEL,I95 1; MCCARTHY, 1976; NESBITT,
1980), (2) influx of COz-rich fluids from an external source,
* Presented at PACROFI II, Second Biennial Pan-American Conference on Research on Fluid Inclusions, held in Blacksburg, VA,
January 4-7, 1989.
641
642
I). I.. Hall and R. J. Bodnar
tures such as cross-cutting inclusion planes may also be useful
(TOURET, 1981: ROEDDER, 1984).
A common problem encountered in these studies is that
fluid inclusion densities are too high or too low to have resulted from trapping at inferred peak metamorphic P-T conditions. This may indicate trapping during uplift, i.e., following the peak of metamorphism. Alternatively, re-equilibration
of inclusions to both higher and lower densities has been
demonstrated during laboratory-simulated burial and uplift
(STERNERand BODNAR, 1989). If the uplift path deviates
significantly from being isochoric with respect to the densities
of inclusions trapped at peak metamorphism, re-equilibration
of fluid inclusions to lower or higher densities may occur. In
these cases densities (or homogenization temperatures) could
not be employed to determine whether the inclusions were
trapped at peak metamorphic conditions.
Inconsistencies between fluid compositions implied by
mineral equilibria and observed compositions of fluid inclusions are common. Often the fugacity of water implied by
mineral equilibria, while low. is not zero and thus is at odds
with apparently pure CO> fluid inclusions. LAMBet al. (1987b)
have provided strong evidence that many CO&h
fluid inclusions within lowfHz0 and,fCOz mineral assemblages of
the Adirondacks were not trapped during granulite-facies
metamorphism. These authors favor a fluid absent model for
metamorphism with subsequent trapping of CO>-rich fluid
inclusions during uplift. However, in other terranes (e.g., S.
India; Scouri. Scotland; Bamble, S. Norway, and Buksefjorden, S.W. Greenland) the presence of COz-rich inclusions is
consistent with trapping during metamorphism (NEWTON,
1986).
In this study we evaluate the methane content of COa-rich
fluid inclusions in granulites and theJT)* conditions implied
by these inclusions. We have calculated the molecular speciation in a C-O-H fluid as a function of pressure, temperature, oxygen fugacity, and activity of carbon (P-TTf02-a,)
and have compared these results with observed fluid inclusion
compositions. Finally, we consider the possibility of postentrapment compositional changes to fluid inclusions during
uplift of granulite terranes as a mechanism to explain apparent
inconsistencies between observed and calculated compositions.
FLUID INCLUSIONS FROM GRANULITES
A survey of the literature reveals that final melting temperatures of CO2 [Tm(CO,)] in fluid inclusions from granulites are often depressed below the triple point value for the
p,ure CO* system (-56.6”Q indicating that other constituents
are present. A cumulative histogram of Tm(C02) from 7
studies (Fig. 1) reveals values for Tm(C02) ranging from
-55.5”C to < -62.5”C with a mean near -57.5”C. We do
not claim that these studies are entirely representative of
Tm(C02) from granulites; however, they do encompass ranges
reported in other studies in which actual data is not tabulated
(e.g., SCHREURS, 1984; KREULEN and SCHUILING, 1982;
HANSENet al., 1984). Furthermore, the presence of CHI has
been mentioned in additional studies (e.g., PERCHUKet al.,
1985; CHACKOet al., 1987; TOMILENKOand CHUPIN, 1983).
Some of the data included in Fig. 1 are from inclusions which
>r
100
0
e
80
5
60
g
t
40
20
-63
-61
-59
Tm(CO,)
-57
-55
(“C)
FIG. 1. Cumulative histogram showing the melting point of CO2
in CO&h
inclusions from seven studies of granulite and related
rocks. The vertical dashed line shows the triple point of pure CO2
(-56.6”C). Data from each study was normalized to 100 total inclusions. The number of data points in the original data sets varied from
33 to 1400. Data from HOLLISTER
and BURRUSS
(1976), CO1inclusions of KONNERUP-MADSEN
(1977), high-gradeand pegmatite samples of KREULEN(1980), SWANENBERG
(1980), RUDNICKet al.
(l984), SANTOSH(1986), and scattered monophase inclusions of
SANTOSH
( 1987).
have been interpreted to be secondary based on textural relationships and/or densities. However, as outlined earlier,
these criteria are rarely unequivocal. Some might argue that
the magnitude of the freezing-point depression shown on Fig,
1, which ranges from a few tenths of a degree to a few degrees
Celsius, is not significant but, rather, reflects the accuracy of
the measurement. In some cases this may be true. However,
the reported accuracy and precision of the measurements is
generally kO.1’ to -+0.5”C and we accept these values as
being correct.
Of the possible components that can cause depression of
Tm(CO& methane and nitrogen are the most common.
Methane has been detected during analysis of inclusions from
granulites using Raman spectroscopy (RUDNICKet al., 1984;
SCHREURS, 1984), mass spectrometry (TOMILENKO and
CHUPIN, 1983), and gas chromatography (SWANENBERG,
1980; KREULEN and SCHUILING, 1982); however, in many
studies such analyses have not been conducted and it is unresolved what additional component(s) are causing the
depressions in Tm(CO& although CH4 is a reasonable possibility.
Figure 2 shows the CH4 content of a binary C02-CH4 inclusion as a function of Tm(C02). Because CH4 is partitioned
into the vapor phase, and Tm(C0,) is a function only of the
amount of CH., in the coexisting liquid phase (see BURRUSS,
198 I), actual bulk inclusion compositions in the presence of
a significant volume of vapor are more CH.,-rich than indicated by Tm(C02). For instance, for a Tm(CO*) of -57.5”C
a bulk composition of approximately 6 mol% CH4 is indicated
if a negligible ~01% vapor is present, while 9.5 mol% CH, is
indicated if 50 ~01% vapor coexists with liquid. Thus, even
minor depressions of Tm(C02) below -56.6 indicate that
significant quantities of CH4 may be present. A minimum
of about one mol% CH, in a fluid inclusion cannot be detected
microthermometrically
due to the resolution limits of the
643
Presence of methane in fluid inclusions from granulites
saturation, the system is completely constrained. We have
assumed that the only species present are HZ, 02, H20, COZ,
CO, and CH4. These are related through the linearly independent reactions
c + 02 = co*
co +
10 -
%02
=
(1)
co2
(2)
HZ + ‘/202= HZ0
-0 ~01%vapor
(3)
CH4 + 202 = COZ + 2HzO
(4)
and the constraint that
Tm(CO,)
(“Cl
FIG. 2. Final melting temperature of solid CO, [Tm(C02)] and
corresponding bulk mol% methane in CO&H,, fluid inclusions for
negligible (-0 ~01%)and 50 vol% vapor present at Tm(CO*).
Tm(CO*) is insensitive to the amount of vapor present but the bulk
mol% methane for a given Tm(COJ will vary from the limiting value
at -0 ~01%vapor to some higher value as the ~01%vapor increases
because CH4is partitioned into the vapor phase. Data from VAN DEN
KERKHOF( 1988).
fluid inclusion stage (ideally kO.1 “C for the U.S.G.S.-type
stage). For a typical value of Tm(C02) of -575°C (as indicated by our compilation of inclusion data from granulites,
Fig. 1) the mole fraction of methane may range from 6 to 10
mol% depending on the ~01% vapor present (550 ~01%) at
Tm(C02). Minimum Tm(C02) values near -6 1“C (Fig. 1)
imply 2 l-27 mol% CH., if the inclusions are compositionally
represented by the system C02-CH4 (Fig. 2).
In the preceding discussion, the effect of an aqueous component on the CH4 content indicated by Tm(C02) has been
ignored. Due to its wetting characteristics, the aqueous phase
forms a rim on the walls of the inclusions and up to lo-20
mol% HZ0 may be undetectable by optical microscopy.
Likewise, Raman and infrared spectroscopy may fail to detect
similar quantities of HZ0 in fluid inclusions (VRY et al., 1987).
If HZ0 is present, clathrate hydrate should form upon cooling
and CO* or CH4 may be differentially partitioned into the
clathrate (SEITZ and PASTERIS, 1989). In the absence of differential partitioning, HZ0 merely acts as a diluent and the
COJCH., ratio indicated by Tm(C02) will be unaffected.
C-O-H FLUID SPECIATION
The speciation of C-O-H fluids at elevated P-T conditions
has been calculated using the “equilibrium constants - mass
balance” technique (FRENCH, 1966; EUGSTERand SKIPPEN,
1967; OHMOTO and KERRICK, 1977; FERRY and BAUMGARTNER, 1987; HOLLOWAY,1977). For a homogeneous CO-H fluid in the system C-O-H the phase rule dictates that
four variables can be independently and arbitrarily varied
(within limits as indicated by the phase diagrams) without
causing a phase change. Hence, four variables must be specified for the system to be invariant. In the case of graphite
saturation two phases are present (graphite + C-O-H fluid)
and the variance is reduced to 3. By specifying total fluid
pressure (PO&, temperature (T), oxygen fugacity (f02), and
the activity of carbon (a& which is defined to be 1 at graphite
ZXi=
1
(5)
where Xi is the mole fraction of component i. The mole fmctions of species are related to their fugacities through the
equation
(f;huid,T
=
rrXiYiPfluid
(6)
where fi‘, Pi and yi are the fugacity, activity coefficient and
fugacity coefficient, respectively, of component i. If idea1
mixing is assumed, Eqn. (6) reduces to
(f;huad,T
=
XiYiPIluld.
(7)
Fluid speciation was calculated using functions describing
the equilibrium constants for reactions (l)-(4) presented in
OHMOTOand KERRICK(1977). The procedure was modified
to allow for nonideal mixing of fluid species and for carbon
activities less than 1. For idea1 mixing, fugacity coefficients
for HzO, COZ, and H2 below 3 kb were taken from BURNHAM
et al. (1969), BURNHAMand WALL (unpubl. data), and SHAW
and WONES ( 1964), respectively. Ideal mixing fugacity coefficients for CO, CH4, and HZ above 3 kb were calculated
from the reduced variable equations of RYZHENKO and
VOLKOV ( 197 1). For nonideal mixing, fugacities were calculated from the equation
In (fi) = In [V/( V - b,)] + bi/( V - b,)
- [~ZQZ,/(R~‘~~~,)]
+ (a,bi/bZ,RT”5){1n
In [(I/ + b,)/v]
[(V + b,)/v]
- &/(b,
+ I?> - ln (FIRT)
(8)
(FLOWERSand HELGESON,1983), where V is the mean molar
volume of the fluid mixture, a and b are empirical coefficients
of the Redlich-Kwong equation of state, subscript m refers
to the fluid mixture, subscripts i and j refer to individual fluid
components, subscript z’jrefers to interaction between unlike
molecules i and j, and R is the gas constant. The terms a and
b were calculated using the subroutine MRKMIX (HOLLOWAY, 198 1). The general Redlich-Kwong equation was then
solved explicitly for mean molar volume. In order to calculate
the mixing terms the mole fractions of the species must be
known. These were obtained by iteration: Ideal mixing was
initially assumed to calculate the nonideal mixing fugacity
coefficients. The resulting mole fractions then became the
new mole fractions to re-calculate the nonideal mixing fugacity coefficients. Using this procedure convergence was
usually achieved in about six iterations. The criterion for
634
I). I.. Hall and R. J. Bodnar
convergence is that the relative variation in mole fraction
between two successive iterations is less than IO ’ for all species. by the relation
(9).
where k is the iteration number.
Mineral thermobarometry
generally indicates pressures of
6-9 kb and temperatures
of 700- 1000°C for granulite facies
metamorphism (BOHLEN, 1987: VALLEY, 1987). Although
granulites show a wide range of oxidation states from graphitebearing varieties to varieties containing hematite (NEWTON,
1986), most oxygen fugacities indicated by coexisting mag-
netite-ilmenite solid solutions or orthopyroxene-bearing assemblages suggest JO2 within 1 to 2 log units of the fayalitemagnetite-quartz equilibrium (FMQ + 1-2) (e.g., ROLLINSON, 1980; PERKINS et al., 1982; HANSEN et al., 1984; LAMB
and VALLEY, 1984; SCHREURS and WESTRA, 1986; WEBER
and BARBEY, 1986). We have calculated C-O-H fluid speciation over the P-T range l-8 kb and 400-800°C at three,fO,
conditions: FMQ + 1, FMQ and FMQ - 1. We have contoured isopleths of constant mol% CO* and CH4 in P-T space
at these three .fO, conditions (Fig. 3, a-f) because we are
concerned with speciation among the carbonic species. At
temperatures
below the graphite maximum (graphite stable
region), a, = 1. At temperatures above the graphite maximum,
a, is defined as 0.6 for,fO, = FMQ and 0.06 for,f’O:! = FMQ
+ I. These values of a, were chosen in order to generate low
water fugacities at P-T conditions of granulite-facies
metamorphism. Over the P-T-j”02 conditions considered here the
fluid can be adequately modeled as a mixture of the three
species COz, CH4, and HzO, all other species combined generally comprising less than 1 mol% of the fluid. Thus, the
balance of the fluid not contoured
in Fig. 3 is composed
essentially of HzO. It is evident from Fig. 3 that at granulite
pressures and temperatures
the fluid phase should be dominantly a binary CO*-Hz0 fluid, except at low ,JO, where
CH., may comprise several mol% of the fluid (Fig. 3f). The
presence of primary graphite in some granulites (HOL.LISTER
and BURRUSS, 1976; RUDNICK et al., 1984; CHACKO et al.,
1987) is consistent with metamorphism
at,fO, conditions at
or below FMQ. Above the graphite maximum, the fluid phase
may vary from essentially pure H20 to pure CO*, depending
on the activity of carbon. One trend which is clear from Fig.
3 is that methane should never be a significant species in
HzO-poor fluids at any P-TTf’02 conditions ofgranulite facies
metamorphism.
In general, COz and CH, can both be significant species only in fluids which contain appreciable H20,
and then only at relatively low T-P-j’02 (e.g., Fig. 3 e-f).
Methane contents at reasonable granulite P-T-j”02 conditions
should not exceed -0.2 mol% in C02-rich fluids and should
reach this value only when graphite is present. Above the
graphite maximum
methane contents are uniformly low
(5.0 I mol%). The smalk quantity of CH, predicted at granulite-facies metamorphic
conditions would be undetectable
in fluid inclusions with current microthermometric
and
spectroscopic techniques.
These results are consistent with those of other workers
(e.g., LAMB and VALLEY 1985; LAMB et al., 1987b) who have
used similar calculations to show that COz-rich inclusions
containing appreciable CH4 are incompatible with estimated
P-T-/‘02 conditions of granulite facies metamorphism.
It has
been suggested in these studies that leakage, re-equilibration
or late-stage trapping. even for those aberrant inclusions that
“appear to be primary” (LAMB and VALLEY, 1985), may explain the discrepancies.
However, the details of these mechanisms, or direct evidence for a particular one. have not been
presented.
Results described above present us with a dilemma. Fluid
calculations predict that COz-rich inclusions from granulites
should contain only trace quantities of CH.,, while fluid inclusions reported from some granulites commonly contain
greater than 6 mol% CH4. Assuming that our calculations
adequately model fluid speciation at conditions of granulitefacies metamorphism,
then either the inclusions were not
trapped at peak metamorphism
or they have undergone posttrapping compositional changes. Our calculations suggest that
the observed fluid compositions
require trapping at low P
and,/O, and fairly high T (e.g.. (2 kb, FMQ-1 and 700°C;
Fig. 3e-f) during uplift. Such inclusions would have low
densities and their isochores would not pass through peak
metamorphic
conditions.
POST-TRAPPING
CHANGES
IN FLUID COMPOSITION
C’loscYlSJZV
tcm
One possible explanation for the discrepancy between calculated and observed fluid compositions is that compositions
recorded in fluid inclusions at room temperature are not representative of fluid compositions
originally trapped, owing
to post-trapping
re-equilibration
among fluid species during
cooling. We have calculated speciation changes in mode1 fluid
inclusions
trapped at various P-T7f02 conditions
using
methods
outlined by HOLLOWAY (1981) and DUBESSY
( 1984).
A fluid inclusion is assumed to contain a fluid of known
composition and molar volume at some initial P-T condition
(e.g., of trapping or observation). By specifying an arbitrary
inclusion volume at this P-T and assuming the system remains
isovolumetric,
or by correcting for thermal expansion and
compressibility of the host, the new composition, molar volume, and internal pressure of the fluid inclusion at some
final temperature
may be calculated. This is accomplished
via the functions describing equilibrium constants for reactions (2)-(4) (OHMOTO and KERRICK, 1977). Eqn. (5), and
the additional mass balance constraints
Z:n,C = Zn,C
(10)
Zn,O = Bn,-0
(1 I)
&z,H = Bn/H
(12)
where n, and n/are the initial and final moles, respectively,
of carbon, oxygen, or hydrogen in the inclusion. Fugacity
coefficients for Hz, 02, HzO, CO2, CO, and C& are calculated
using HOLLOWAY’S (198 1) program MRKMIX. The new PV-X properties of the inclusion are obtained at the new temperature by simultaneously solving six equations (3 describing
fluid speciation and 3 describing mass balance) using New-
645
Presence of methane in fluid inclusions from granulites
Mole % CO,
50
450
550
850
750
850
Mole % CH,
350
450
550
850
750
850
‘i>z/l&y;
T=?-&f~
.d’
I 50 .‘.n”.n’
450
550
850
750
850
850
450
550
650
750
950
750
850
b01.
450
550
650
750
L....,.,.I450
550
650
I.
I.
1.
I
850
.I
850
Temperature (OC)
Ptc. 3. P-T projections contoured for isopleths of mol% CO2 (a, c, and e) and CH4 (b, d, and f), atf0, defined by
fayalite-magnetitequartz f I log unit (FMQ f 1). The activity of carbon is defined as I below (low T side) the graphite
maximum (gr) and is arbitrarily defined to be 0.6 and 0.06 above the graphite maximum for FMQ and PMQ + 1,
respectively.
ton’s method for non-linear equations (BURDENand FAIRE&
1985). The criterion for convergence is defined by Eqn. (9).
A temperature of 400°C was chosen as the minimum temperature for our calculations for three reasons. Firstly, this is
the lower temperature limit of Holloway’s equation for
aqueous fluids. Secondly, it is questionable if chemical equilibrium is attained among carbonic fluid species at temperatures much below 400°C (SACKETT and CHUNG, 1979;
HARTING and MAASS, 1980; ZIEGENBEINand JOHANNES,
1980). Thirdly, limiting our calculations to >4OO”C allowed
us to neglect the effects of fluid immiscibility over the P-T
conditions considered here.
Figure 4 shows the speciation changes generated upon
cooling a fluid inclusion, which was trapped at 6 kb, 800°C
andfOz = FMQ, to 400°C. At the time of entrapment, the
inclusion contains 83.6 mol% CO*, 15.4 mol% H20, 0.006
mol% CH.,, 0.9 mol% CO, and 0.05 mol% Hz. At 400°C the
inclusion composition has changed to 84.7 mol% C02, 15.1
mol% H20, 0.2 mol% CH4, 0.01 mol% CO, and 0.006 mol%
Hz. While the amounts of the major components COz and
Hz0 remain nearly constant, the concentrations of minor
components including CH4 change significantly. Although
the mole fraction of methane has increased over one order
of magnitude, the final concentration is still below the detection limit of microthermometry and probably other spectroscopic techniques as well. Our calculations for other trap-
I). L. Hall and R. J. Bodnar
646
"0
Temperature (“C)
200
400
600
800
Temperature (OC)
FIG. 4. Closed system speciation changes during cooling a fluid
inclusion at constant volume. The original fluid was trapped at 6 kb,
800°C. FMQ, and a, = 0.4.
suggest that post-trapping,
closed system
speciation changes may produce maximum CH4 contents of
perhaps 1 mol%, which is insufficient to account for phase
relations described from natural fluid inclusions from granulites.
The preceding calculations assumed that graphite does not
precipitate in the inclusions during cooling. If graphite were
to precipitate, significant quantities of CH4 may be produced
at low temperatures and pressures (Fig. 3; see also DUBESSY,
1984). Graphite-bearing inclusions are not uncommon in
granulites (TOURET, 1985; VRY and BROWN,1987) and often
contain detectable to significant quantities of CH4. Graphitebearing inclusions sometimes contain more CH, than apparently coeval graphite-absent inclusions, and it is likely
that the methane contents of graphite-bearing inclusions (and
the impliedfOz conditions) are not representative of the values at trapping. In the absence of graphite precipitation,
however, the presence of detectable methane is inconsistent
with predicted fluid compositions. We are left with the hypothesis that the inclusions have changed bulk composition
due to preferential loss or gain of fluid components and not
by closed system changes in speciation during cooling.
FIG. 5. Proposed uplift paths for granulite terranes. Sources as
follows: (1) PERCHUK
et al., 1985; (2) OLSEN, 1987; (3) HOLLISTER
et al., 1979; (4) SWANENBERG,
1980; (5) SCHREURS,
1984;(6) SANTOSH, 1987; (7) HOLLISTER
et al., 1979, and (8) BOHLEN.1987.
ping conditions
during uplift assuming the granulite host is continuously buffered by FMQ + 1, FMQ or FMQ - 1. During uplift the
host rock fluid composition will be a function of the uplift
path, and several different P-T-time paths have been proposed
for granulites (Fig. 5). Two endmember types of uplift paths
are noted, one which is concave toward the temperature axis
(hereafter referred to as T-concave uplift) and one which is
convex toward the temperature axis (hereafter referred to as
T-convex uplift). From these we constructed two hypothetical
paths from peak metamorphic conditions of 6 kb and 800°C
(Fig. 6) and then calculated matrix fluid compositions along
these paths assuming that theSO* of the fluid is buffered by
FMQ+ l,FMQorFMQ1.
The variation infHz during T-concave and T-convex uplift
is shown in Fig. 7. Also shown are them2 variations in a
model fluid inclusion trapped at 6 kb and 800°C during cooling at constant volume. The graphite maximum is shown
schematically by vertical dashed lines; due to the pressure
dependence of the graphite maximum it intersects the two
uplift paths at different temperatures. The activity of carbon
Open system
Recent studies (e.g., PASTERISand WANAMAKER, 1988;
STERNERet al., 1988; HALL et al., 1989a, b) have shown that
fluid inclusions can change composition at elevated temperatures and pressures during laboratory experiments. The
mechanisms of exchange, the conditions under which these
changes occur, and the actual species which are diffusing have
not been fully investigated, but it is clear that fugacity gradients between inclusion fluids and matrix fluids play a key
role in the process. In order to investigate the possibility that
high methane contents in granulite fluid inclusions result from
inward hydrogen diffusion and conversion of CO* to CH, +
Hz0 via the reaction
CO2 + 4H2 = CH., + 2H20
(13)
we have modeled the variations in fluid composition in the
host rock (e.g., grain boundary fluids) and fluid inclusions
200
Te
400
600
800
ure (CC)
FIG. 6. Hypothetical temperature-concave (T-concave) and temperature-convex (T-convex) uplift paths used to model matrix fluid
composition during uplift of granulite terranes. Paths were chosen
to be representative of paths shown in Fig. 5. The 25”, 35” and 45”C/
km geotherms are shown for reference.
647
Presence of methane in fluid inclusions from granulites
550
450
550
650
750
850
3050
450
550
650
750
850
*
r
r
650
750
’
200
.FMO-1
50
450
T-concaw
550
C.
850
Temperature (OC)
FIG. 7. Variation infH2 in matrix (host) fluids and a model fluid
inclusion as a function of temperature at FMQ + 1 (a), FMQ (b) and
FMQ - 1 (c). Pressures along the T-concave and T-convex host uplift
paths are constrained from Fig. 6. The internal pressure of the fluid
inclusion is dictated by the P-V-T-X properties of the trapped fluid
and the assumption that the inclusion maintains a constant volume.
The maximum stability of graphite along a particular f02 buffer is
shown schematically by vertical lines in a and b, where the dashed
line corresponds to T-concave uplift and the dotted line corresponds
to T-convex uplift. Above the graphite maximum the activity of carbon is defined to be 0.04 in a and 0.4 in b. Graphite is stable over
the entire range of P-T in c.
above the graphite maximum is defined as 0.04 in Fig. 7a
and 0.4 in Fig. 7b in order to produce a COz-rich fluid at
conditions of peak metamorphism. Graphite is stable over
all P-T considered in Fig. 7c and the activity of carbon is
defined as 1.
During both T-concave and T-convex uplift significantfH2
overpressures may be generated by matrix fluids. Maximum
jI-Iz gradients between inclusion and matrix range from 5 to
25 bars for T-convex uplift, depending onf02, and 10 to 75
bars for T-concave uplift, depending onf02 (Fig. 7). The
higher jI-Iz attained in the matrix fluid during T-concave
uplift is due to the higher total fluid pressure (Fig. 6). Given
the high diffusion rates of hydrogen through quartz at elevated
temperatures (KATS, 1962; KATS et al., 1962; KRONENBERG
et al., 1986) it is unlikely thatfH2 gradients of the magnitudes
indicated in Fig, 7 could be maintained for geologic time,
and it is probable that at least partial re-equilibration would
occur via hydrogen transport into fluid inclusions. For instance, the diffusion coefficient for hydrogen in P-quartz at
800°C is 2 X 10-r’ m*/sec (KRONENBERG
et al., 1986). This
corresponds to about 0.07 mm’/hr. The average path length
for diffusion of hydrogen from grain boundaries into fluid
inclusions is on the order of 1 mm. Thus, within a few hours
inclusions can be in hydrogen communication with matrix
fluids, and over geologic time significant amounts of hydrogen
could enter the fluid inclusions.
Hydrogen diffusion into CO*-Hz0 fluid inclusions and
concomitant reduction of CO2 to CH4 (SHzO) via reaction
(13) has been verified experimentally (HALL et al., 1989b).
Fluid inclusions containing 50 mol% CO2 were synthesized
in Brazilian quartz at 825°C and 3 kb. The inclusions were
then re-equilibrated for 15 days at 650-825°C and 1.5 kb at
fHz conditions defined by the graphite-methane
buffer
(_ 100-250 bars). Preliminary results indicate that inclusions
with initial Tm(C02) = -56.6 + 0.1 “C now have Tm(C02)
as low as -65.4”C. Raman spectroscopic analysis indicated
only the presence of CO* (+HzO) in the original inclusions,
whereas large quantities of CH., and detectable Hz were present, in addition to COZ and H20, in re-equilibrated inclusions.
Perhaps the most significant aspect of these results is that
hydrogen diffusion into fluid inclusions occurred in spite of
the 0.825- 1.5 kb internal overpressures during re-equilibration. This observation indicates that it is the fugacity of hydrogen and not the total pressure that is driving diffusion.
We have modeled the effects of hydrogen diffusion on the
P-V-X properties of a C-O-H fluid inclusion trapped at 6 kb,
800°C and FMQ. The procedure used to calculate closed
system speciation changes was modified to allow for changes
in total hydrogen content. Addition of an arbitrary amount
of hydrogen to a fluid inclusion is assumed to occur at a
given temperature and the equilibrium speciation is recalculated based on the modified bulk composition of the inclusion. Figure 8 shows the effect of addition of hydrogen on
thejH2 inside the fluid inclusion assuming addition occurs
at 450°C or 550°C. The percentage of hydrogen which is
added to the inclusion (Fig. 8) is relative to the amount of
total hydrogen originally present in the inclusion (i.e., ZH
= 2nHz0 + 2nH2 + 4nCH4), so that 100% diffusion corresponds to adding an amount of hydrogen equal to the total
amount originally present. Because the original inclusion was
relatively hydrogen-poor (i.e., 0.3 moles H per mole of fluid,
compared to 1.84 moles 0 per mole of fluid) a small amount
of hydrogen diffusion contributes significantly to the total
hydrogen content. It should be emphasized that these calculations do not incorporate any diffusion coefficients and
that instantaneous additions of the given amounts of hydrogen are assumed.
AS expected, fH, inside the inclusion increases as progressively more hydrogen
is added to the inclusion.
At 550°C
D. L. Hall and R. J. Bodnar
648
-0
100
%
200
300
400
creases the molar volume (hence bulk density) of the inclusions. Ideally, one might expect to observe a positive correlation between bulk density (as inferred from Th) or H@
content, and CH4 content (as inferred from Tm(C02)). To
our knowledge these trends have not been observed. However.
the increase in water content resulting from reaction ( 13) is
fairly subtle considering
the difficulties and inaccuracies
involved in quantifying XHzO in water-poor C’02-Hz0 inclusions. Furthermore,
inclusions do not remain constant
volume systems when subjected to modest overpressures or
underpressures
(STERNERand BODNAR, 1989). thus complicating any interpretation
of density variations.
Hydrogen Diffused
DIFFUSION
FIG. 8. Variation in/H2 in a model fluid inclusion trapped at 6
kb, 800°C, FMQ, and a, = 0.4 and subsequently cooled to 550°C
or 450°C at which point hydrogen is added to the inclusion. The
percent of hydrogen added is referenced to the total amount of hydrogen originally contained in the inclusion (i.e., contributed from
the species H20, CH4, and Hl). At each point, representing instantaneous addition of the given amount of hydrogen, the C-O-H fluid
is assumed to be in equilibrium.
the initialfH2 inside the inclusion is 4 bars while that of the
matrix, assuming T-convex uplift, is 13 bars (Fig. 7b). If hydrogen diffusion were allowed to progress to completion (i.e.,
to the point where the &
inside the inclusion equals 13
bars), 175% of the total hydrogen originally present in the
inclusion would have to be added (Fig. 8). If uplift were Tconcave to 550°C them,
in the matrix would be 25 bars
and 275% hydrogen must be added to equilibrate fH2 (Fig.
7b; Fig. 8).
Speciation changes resulting from hydrogen diffusion are
significant. Figure 9 shows the variation in mol% CO*, H20,
and CH4 during progressive hydrogen diffusion at 550°C.
Addition of hydrogen causes reduction of CO* to CH4 via
reaction (13). The mole fraction ofCOz decreases while those
of CH4 and H20 increase (Fig. 9). After 175% hydrogen addition, the amount required to equilibrate inclusion and matrixfHz during T-convex uplift to 550°C the fluid composition in the inclusion has been modified to 68.4 mol% COa,
25.3 mol% HzO, and 6.2 mol% CH,. After 275% addition,
the amount required to equilibrate inclusion and matrixfH*
during T-concave uplift to 550°C the fluid composition
in
the inclusion has been mod;fied to 60.9 mol% COz, 30 mol%
HzO, and 9 mol% CH4. Thus, significant quantities of CH4
can be produced as a result of hydrogen diffusion into carbonic fluid inclusions. Methane concentrations
on the order
of 5- 10 mol% may be produced in inclusions which contained
negligible CH., at the time of entrapment.
These concentrations are consistent with Tm(C02) of granulite fluid inclusions
of -57” to -58°C (Fig. 1). Therefore, compositions produced
by hydrogen diffusion described above are similar to those
reported from granulites, at least with respect to the CO*/
CH, ratio. The amount of Hz0 after 275% diffusion at 550°C
(30 mol% H20) may be sufficiently large to be detected optically, microthermometrically
(e.g., by formation of clathrate), or spectroscopically.
In a constant volume system addition of hydrogen and
production of CH., + Hz0 via reaction (13) significantly in-
01; &O
Preferential diffusive loss of Hz0 from C02-Hz0 inclusions
has been postulated to explain the occurrence of nearly pure
COz inclusions in granulites (ROEDDER, 198 1). The driving
force for such diffusion is an.lHzO gradient between inclusion
and matrix. This mechanism has been shown to be experimentally viable at granulite temperatures for saline synthetic
fluid inclusions in Brazilian quartz (STERNER et al., 1988;
HALL et al., l989a), and thus might be expected in similar
geologic environments.
We have plotted the variation in,fH*O
of matrix fluids during T-convex and T-concave uplift, and
thefHzO
inside our model fluid inclusion trapped at 6 kb,
800”C,J02 = FMQ, and a, = 0.4 to test this possibility (Fig.
10). Pressure-temperature
conditions in the host rock during
uplift are the same as in Fig. 7. Compositions
more HzOrich can be generated by specifying a, < 0.4, in which case
thefHz0
at trapping would be greater than that shown in
Fig. 10 and the loci of,fHzO in the inclusion during uplift
would be translated to higher values.
Over much of the uplift history fH20 is higher in the matrix
than in the inclusion, particularly during T-concave uplift.
Thus, HI0 would be expected to diffuse into, and not out of,
fluid inclusions during uplift. Exceptions occur if the uplift
is strongly T-convex and the fluid composition
remains relatively constant. such as would occur if uplift followed one
of the isopleths of CO* shown in Fig. 3. In this case the ratios
of various species are similar inside and outside the inclusion
but the total pressure, henceJH20.
inside the inclusion is
200
‘%
H;:egen
300
400
DlWa
FIG. 9. Variation in mole/o CO1, H20, and CH4 during hydrogen
diffusion at 550°C under the same conditions described in Fig. 8.
649
Presence of methane in fluid inclusions from granulites
I.
850
.il
8.1”
450
550
650 '750
850
Temperature (“C)
FIG. 10. Variation in_/&0 in matrix (host) fluids and a model
inclusion as a function of temperature at FMQ. Pressures for the
host fluids and the model inclusion, and other symbolsare as described
in Fig. 7.
higher than outside. Diffusion of water out of fluid inclusions
is also feasible if matrix fluids evolve toward more H20-poor
compositions during uplift. This may occur if fluids undersaturated with respect to graphite become saturated, perhaps
due to influx of CO&ch fluids from an external source. In
graphite-saturated fluids near the graphite maximum jH,O
is low because the fluid is essentially pure CO2 (Fig. 3). Secondary graphite is sometimes found in granulite terranes,
suggesting that graphite saturation was reached in matrix
fluids during uplift of these areas. Thus, CO*-rich fluids may
be present in the matrix after peak metamorphism, consistent
with the proposed paragenesis for these inclusions in some
terranes (LAMB et al., 1987b).
CONCLUSIONS
Calculations of fluid speciation for P-T-fOz-a, conditions
consistent with granulite facies metamorphism predict that
methane contents should be uniformly low in COZ-rich fluids.
Despite this result, many fluid inclusions which have apparently been trapped during or shortly after granulite-facies
metamorphism contain on the order of 6 to 10 mol% methane
relative to CO2 + CH4. Post-entrapment, closed system speciation changes in fluid inclusions do not produce detectable
amounts of CH4 unless graphite precipitates within the inclusions, and even then the amounts produced may be small.
Modeling of uplift paths which are concave (T-concave)
and convex (T-convex) toward the temperature axis suggests
that decompression and cooling from conditions of granulitefacies metamorphism may result infH2 gradients between
matrix fluids and peak metamorphic fluid inclusions that
would promote diffusive addition of hydrogen to these inclusions. Addition of hydrogen causes conversion of CO2 to
CH4 + HzO, producing fluid inclusions with methane contents and CHJCOz ratios similar to those reported in the
literature.
Formation of pure CO2 or C02-rich fluid inclusions by
diffusive loss of Hz0 from C02-H20 fluid inclusions in response tofH20 gradients between inclusion and matrix appears to be feasible only if the uplift is strongly T-convex and
the matrix fluid composition remains relatively constant, or
the matrix fluid composition evolves toward more H20-poor
compositions. The latter would occur if the activity of carbon
increases after peak metamorphism, for instance by influx
of CO*-rich fluids from an external source. The matrix fluid
may also become more C02-rich by selective removal of Hz0
from the fluid and its incorporation into hydrous phases during retrograde hydration. The influx of more reduced fluids
may cause hydrogen diffusion into fluid inclusions as well.
This latter mechanism has resulted in hydrogen diffusion into
primary fluid inclusions in pyroxene at the Ducktown, Tennessee, massive sulfide deposits after metamorphism to middle-amphibolite grade (HALL et al., 1989~).
Acknowledgments-Some
of the computer programs used in this
study were developed with the assistance of Bob Downs, John Doyle,
and Mike Sterner, and additional discussions with Cheryl Knight,
Matt Nyman, and Mike Sterner were invaluable. This manuscript
benefited immeasurably from reviews by Andreas Kronenberg, Will
Lamb, and Ed Roedder. This study was supported by grants from
the Earth Sciences section of the National Science Foundation (NSF
Grant EAR-8607356), the American Chemical Society Petroleum
Research Fund (Grant # 1886 1-G2) and by funds from the Department of the Interior’s Mining and Mineral Resources Research Institute program administered by the Bureau of Mines under allotment
grantG1194151.
Editorial handling: R. A. Schmitt
REFERENCES
BOHLENS. R. (I 987) Pressure-temperature-time paths and a tectonic
model for the evolution of granuhtes. J. Geol. 95, 6 17-632.
BURDENR. L. and FAIRESJ. D. (1985) NumericalAnalysis, 3rd edn.
Prindle, Weber and Schmidt, Boston.
BURNHAMC. W., HOLLOWAYJ. R., and DAVISN. F. (1969)Thermodynamic properties of water to 1,OOO”Cand 10,000 bats. Geol.
Sot. Amer. Spec. Paper 132.
BURRUSSR. C. (1981) Analysis of phase equilibria in C-O-H-S fluid
inclusions. In Fluid Inclusions: Applications to Petrology (eds.
L. S. HOLLISTER
and M. L. CRAWFORD);Short Course Handbook.
Mineral. Assoc. Canada 6, pp. 39-74.
CHACKOT., KUMAR G. R. R., and NEWTONR. C. (1987) Metamorphic P-T conditions of the Kerala (south India) Khondahte
Belt, A granulite facies supracrustal terrain. J. Geol. 95, 343-358.
DRURYS. A., HARRISN. B. W., HOLT R. W., REEVES-SMITH
G. J.,
and WIGHTMANR. T. (1984)Precambrian tectonics and crustal
evolution in south India. J. deal. 92 l-20.
DUBESSYJ. (1984) Simulation des equihbres chimiques dans le systeme C-O-H. Consequences methodologiques pour les inclusions
fluides. Bull. Mineral. 107, 155-168.
EUGSTERH. P. and SKIPPENG. B. (1967) Igneous and metamorphic
reactions involving gas equilibria. In Researches in Geochemistry,
II (ed. P. H. ABELSON).John Wiley and Sons.
FERRYJ. M. and BAUMGARTNER
L. (1987) Thermodynamic models
of molecular fluids at the elevated pressures and temperatures of
crustal metamorphism. In Thermodynamic Modeling of Geological
Materials: Minerals, Fluids and Melts (eds. I. S. E. CARMICHAEL
and H. P. EUGSTER);Mineral. Sot. Amer. 17, pp. 323-366.
FLOWERSG. C. and HELGESONH. C. (1983) Equilibrium and mass
transfer during progressive metamorphism of siliceous dolomites.
Amer. J. Sci. 283, 230-286.
FRENCHB. M. (1966) Some geological implications of equilibrium
between graphite and a C-H-O gas phase at high temperatures and
pressures. Rev. Geophys. 4, 223-253.
GLA~SLEYW. E. (1983) Deep crustal carbonates as CO2 fluid sources.
Evidence from metasomatic reaction zones. Contrib. Mineral. Petrol. 84, 15-35.
HALL D. L., STERNERS. M., and BODNARR. J. (1989a) Fluid in-
650
I). 1.. Hall and R. J. Bodnar
elusions in regional metamorphic rocks: experimental evidence
for post-entrapment volumetric and compositional modifications
(abstr.). 10th Biennial Symposium on European Current Rcseun,h
on Fluid Inclusions, Royal School of .Minc~~.Impriul ~‘oll~~~c.
London, Program and Abstracts, 44.
HALLD. L., STERNERS. M., and BODNARR. J. (1989b) Experimental
evidence for hydrogen diffusion into fluid inclusions in quartz
(abstr.). Geol. Sot. Amer. Abstr. Prog. 21, A-358.
HALLD. L., BODNARR. J., and CRAIGJ. R. (1989~) Fluid evolution
during metamorphism and uplift of the massive sulfide deposits
at Ducktown, Tennessee, U.S.A. II. Evidence for compositional
changes to fluid inclusions via hydrogen diffusion. Amer Mineral.
(in review).
HANSENE. C., NEWTONR. C., and JANARDHANA. S. (1984) Fluid
inclusions in rocks from the amphibolite-facies gneiss to charnockite progression in southern Karnataka, India: direct evidence
concerning the fluids ofgranulite metamorphism. J. Me~arn. Geol
2, 249-264.
HARRISN. B. W., HOLT R. W., and DRURY S. A. (1982) Geobarometry, geothermometry and late Archean Geotherms from the
granulite facies terrain of south India. J. G‘rol. 90, 509-527.
HARTINGP. and MAASS1. (1980) Neue Ergebnisse zum KohlenstoffIsotopenaustausch im system CH4-COZ. Milreilungen ZIU z. Arbeitsragung ‘%sotopein der Natur” 26, 13-24.
HODGESK. V., BARTLEYJ. M., and BURCHFIELB. C. (I 982) Structural evolution of an A-type subduction zone Lofoten-Rombak
area, Northern Scandinavian Caledonides Tectonics 1, 441-462.
HOLLISTERL. S. and BURRUSSR. C. (I 976) Phase equilibria in fluid
inclusions from the Khtada Lake metamorphic complex. Geochim.
C’osmochim. Acta 40, 163-l 75.
HOLLISTERL. S., BURRUSSR. C.. HENRYD. L., and HENDELE. M.
(1979) Physical conditions during uplift of metamorphic terranes.
as recorded by fluid inclusions. Bull. Sot. France ,I-lincra/. (‘risrallogr. 102, 555-561.
HOLLOWAYJ. R. (1977) Fugacity and activity of molecular species
in supercritical fluids. In Thermodynamics in Geology (ed. D. G.
FRASER).pp. 161- I8 1. D. Reidel Publishing Company.
HOLLOWAYJ. R. (198 I) Compositions and volumes of supercritical
fluids in the Earth’s crust. In Fluid Inclusions: Applicatiom to Petrology (eds. L. S. HOLLISTER
and M. L. CRAWFORD): Short Course
IIundbook. Mineral. Assoc. Canada 6. pp. 13-38.
KATS A. (1962) Hydrogen in alpha-quartz. Philips Rrs. Repor 17,
1-31. 133-279.
KATS A., HARENY., and STEVELSJ. M. (1962) Hydroxyl groups in
o-quartz. Phys. Chem. Glasses 3, 69-75.
VAN DEN KERKHOFA. M. (1988) The System CO,-CH,-h; in Fluid
Inclusions: Theoretical Modelling and GeologicalApplications. Free
University Press, Amsterdam.
KONNERUP-MADSEN
J. (1977) Composition and microthermometry
of fluid inclusions in the Kleivan granite South Norway. .4mer. .I.
Scl. 277, 673-696.
KRELJLENR. (1980) CO*-rich fluids during regional metamorphism
on Naxos (Greece). Amer. J. Sci. 280.745-77 I.
KREULENR. Bnd S&WILING R. D. ( 1982) N2CH4-CO2 fluids during
formation ofthe Dome de I’Agout, France. Geochim. Cosmochim.
Acta 46, 193-203.
KRONENBERGA. K., KIRBY S. H., AINES R. D., and ROSSMAN
G. R. (1986) Solubility and diffisional uptake of hydrogen in quartz
at high water pressures: Implications for hydrolytic weakening. J.
GeophJjs. Res. 91, 12723-12744.
LAMBW. M. and VALLEYJ. W. (1984) Metamorphism of reduced
granulites in low-CO2 vapour-free environments. Nature 312, 5658.
LAMBW. M. and VALLEYJ. W. (1985) C-O-H fluid calculations and
granulite genesis. In The Deep Proterozoic Crust in the North Atlantic Provinces (eds. A. C. TONI and J. L. R. TOURET):NATO
.4SI Series 158, pp. I I9- I3 I.
LAMBW. M., BROWNP. E., and VALLEYJ. W. (1987a) Post-metamorphic COz-rich fluid inclusions in granulites, evidence from the
Adirondack Mountains, New York (abstr.). Amer. Curr. Res. on
Fluid Inclusions, Jan. .F-7, 1987, Socorro, N. M., Prog. and Abstrac1.s
(unpaginated).
LAMBW. M., VALLEYJ. W., and BROWNP. E. (1987b) Post-meta-
morphic C02-rich fluid inclusions in granulites. Cbnrrrh. Mineral.
Petrol. 96, 485-495.
MCCARTHYT. S. ( 1976)Chemical interrelationships in a low-pressure
granulite terrain in Namaqualand, South Africa, and their bearing
on granite genesis and the composition ofthe lower crust. Geoc,hirn
C‘osmochirn.Acta 40, 1057-1068.
NESBIIT H. W. (1980) Genesis of the New Quebec and Adirondack
granulites: evidence for their production by partial melting. C’ontrih
Mineral. Petrol. 72, 303-3 IO.
NEWTONR. C. (1986) Late-Archean/early Proterozoic CO2 streaming
through the lower crust and geochemical segregation (abstr.). Eos
67, 387.
OHMOTOH. and KERRICKD. (1977) Devolatilization equilibria in
graphitic systems. Amer. J. Sci. 277, 1013-1044.
OLSEN S. N. (1987) The composition and role of the fluid in migmatites: a fluid inclusion study of the Front Range rocks. Conlrih.
Minerul. Pcfrol. 96, 104- 120.
PASTERISJ. D. and WANAMAKERB. J. (1988) Laser raman microprobe analysis of experimentally re-equilibrated fluid inclusions
in olivine: Some implications for mantle fluids. .4mpr. Mineral.
73, 1074-108X.
PERC‘HLIKI.. I... AKANOVICI~ L. YA.. PODLESSKIIK. K., LAVRANT’EVAI. V.. GERASIMOVV. Y~J, FED’KIN V. V.. KITSUL
V. I.. KARSAKOVL. P., and BERDNIKOV
N. V. (1985) Precambrian
granulites of the Aladan shield. eastern Siberia. USSR. J. Metam.
Geol. 3, 265-3 10.
PERKINSD. 111,ESSENEE. J.. and MARCOTTYL. A. (1982) Thermometry and barometry of some amphibolite-granulite facies rocks
from the Otter Lake area. southern Quebec. Canadian .I. Earth
Sri. 19, 1759- 1774.
QUENSEI. P. (I 95 1) The charnockite series of the Varberg district on
the southwest coast of Sweden. Arkiv. Mineral. Geol. 1. 229-332.
VAN REENEN, D. D. and HOLLISTERL. S. (1987) Fluid inclusion
study across a hydration isograd in granulite facies rocks, southern
marginal zone. Limpopo Belt, South Africa (abstr.). Amer. C‘zrrr.
Rc,.c on f+id Inclusions. .Ian. .5- 7. 1987. Socorro N M Pro,q.and
.,Ibstracr.s(unpaginated).
ROEDDERE. (198 1) Origin of fluid inclusions and changes that occur
after trapping. In Fluid 1rwlusion.s:Applications to Petrology (ed.
L. S. HOLLISTER
and M. L. CRAWFORD);Short Course flandbook.
Mineral. Asso~. of(hnada 6. pp. 103-l 37.
ROEDDERE. (1984) Fluid inclusion.%Mineral. Sot :lmer. 12, 644
R&INSON
H. R. (1980) Iron-titanium oxides as an indicator of the
role of the fluid phase during
cooling of granites
metamorphosed
to granulite grade. Mineral. Mug. 43, 623-63 1.
RUDNICKR. L.. ASHWALL. D.. and HENRYD. J. (1984) Fluid inelusions in high-grade gneisses of the Kapuskasing structure zone,
Ontario: metamorphic fluids and uplift/erosion path. Contrib.
Mine&. Petrol. 87, 399-406.
RYZHENKOB. N. and VOLKOVV. P. (197 I) Fugacity coefficients of
some gases in a broad range of temperatures
&em. Intl.. 468-48
SACKETT W. M. and
and pressures.
Gco-
1.
CHUNG
H. M. (1979) Experimental confirmation of the lack of carbon isotope exchange between methane
and carbon oxides at high temperatures. Geochim. Cismochim.
.Icra 43, 273-276.
SANTOSHM. ( 1986) Fluids in deep crustal metamorphism: Data from
the Precambrian
granulites of Kerala Region. Indian Mineral. 27,
84-94.
SANTOSHM. (1987) Cordierite gneisses of southern Kerala, India:
petrology,
fluid inclusions and implications
for crustal uplift history.
Contrib. Minerul. Petrol. 96, 343-356.
SCHREURSJ. (1984) The amphibolite-granulite
West Uusimaa, S.W. Finland.
Geol. 2, 327-34 1.
A fluid inclusion
facies transition in
study. J. Metam
SCHREURSJ. and WESTRAL. (1986) The thermotectonic evolution
of a Proterozoic, low pressure, granulite dome. West Uusimaa,
SW Finland. Conrrib. Mineral. Petrol. 93, 236-250.
SEITZJ. C. and PASTERIS
J. D. (1989) Theoretical and practical aspects
of differential partitioning of gases by clathrate hydrates in fluid
inclusions (abstr.). Second Biennial Pan-American Coeference on
Rtwurch on Fluid Inclusions. Program and Abstracts, 58.
Presenceof methanein fluidinclusionsfrom granulites
SHAW
H. R. and WONBD. R. (1964)Fugacity
coefficients
forhy
drogengasbetweenO”-1000°C
forpressures
to 3000atm.Amer.
Js Sci.l&918-929.
STERNER
S.M. andBODNAR
R. J. (1989)Synthetic
fluidinclusionsVII.Re-equilibration
of fluidinclusionsin quartz during labomtory-
simulated metamorphic burial and uplift. J. Metam. Geol. 7,243260.
STERNERS.M., HALL D. L., and BODNARR. J. (1988) Post-entrapment compositional changes in fluid inclusions: Experimental evidence for water diffusion in quartz (abstr.). Geol. Sot. Amer. Abstr.
Prog. 20, AlOO.
SWANENBERG
H. E. C. (1980) Fluid inclusions in high-grade metamorphic rocks from SW. Norway. Geologica Ultraiectina(Univ.
Utrecht)25.
TOMILENKO
A. A. and CHUPINV. P. (1983) Thermometry,Barometry
and fluid chemistry of metamorphic complexes. ‘Nauka’ Press,
Novosibirsk (in Russian).
TOURET J. (1970) Le facies granulite, metamorphisme en milieu
carbonique. C. R. Acad. Sci. Paris 271, 2228-223 1.
TOURET J. (197 1) Le facies granulite en Norvege meridionale, II.
Les inclusions fluides. Lithos 4,423-436.
TOURETJ. (198 I) Fluid inclusions in high grade metamorphic rocks.
In Fluid Inclusions:ApplicationstoPetrology(eds. L. S. HOLLISTER
651
andM.L,CRAWFORD);
ShortCourseHandbook.
Mineral.
Assoc.
Canada6, pp. 182-208.
TOURET
J. (1985)
Fluid
regime
insouthern
Norway:
Therecord
of
fluidinclusions.
In TheDeepProterozoic
CrusfintheNorth
Atlantic
Provinces (eds. A. C. TOBI and J. L. R. TOURET); NATO ASI
Series 158, pp. 5 17-549.
VRY J. K. and BROWNP. E. (1987) Fluid inclusions across the Pikwitonei granulite domain, Manitoba (abstr.). Amer. Curr. Res. on
Fluid Inclusions,Jan. j-7,1987, Socorro,N.M., Prog. andAbstracts
(unpaginated)
VRYJ., BROWNP. E., and BEACHAINE
J. (1987) Application of microFTIR spectroscopy to the study of fluid inclusions (abstr.). Eos
68, 1538.
VALLEYJ. W. (1987) Granulites and fluids in the deep crust (abstr.).
Int ‘l Unionof Geodesyand Geophys.XIX Gen. Assembly,Abstracts
1, 56.
WEBERC. and BARBEYP. (1986) The role of water, mixing processes
and metamorphic fabric in the genesis of the Baume migmatites
(Ardeche, France). Contrib. Mineral. Petrol. 92,48 l-49 1.
WELLSP. R. A. (1979) Chemical and thermal evolution of Archaean
sialic crust, southern West Greenland. J. Petrol. 20, 187-226.
ZIEGENBEIND. and JOHANNESN. (1980) Graphite in C-H-O fluids:
an unsuitable compound to buffer fluid composition at temperatures up to 700°C. N. Jb. Miner. Abh. 7, 289-305.