Myron G. Best. Igneous and metamorphic 2003 Blackwell Science

Подождите немного. Документ загружается.

between initial and ﬁnal conditions, internally buffered

reactions may not yield much change in modal com-

position of a rock.

Inﬁltration Metamorphism in Open Systems. Suppose

the rock is a highly permeable, open system and invas-

ive, perfectly mobile ﬂuids from an external source

can penetrate along fractures and/or grain boundaries.

The inﬁltrating ﬂuid signiﬁcantly dilutes or even ﬂushes

away that created by any local reaction; the mass of the

mobile ﬂuid can be greater than that of the reacting

solid phases, the ﬂuid/rock ratio is relatively high. Cal-

culations based on the reaction progress variable indic-

ate time-integrated ﬂuid/rock ratios over the entire

course of the reaction are commonly 1 to 4 and locally

greater (Spear, 1993, p. 680). It has been argued by

some petrologists that the typical Barrovian metamor-

phism of carbonate-bearing metasedimentary rocks

requires signiﬁcant inﬁltration of aqueous ﬂuids, as

otherwise very high temperatures are needed. Systems

involving inﬁltration metamorphism in permeable

rocks include “ﬂooding” of a local decarbonating sys-

tem by water liberated from nearby pelitic rocks experi-

encing dehydration reactions or by aqueous solutions

expelled from adjacent magmatic intrusions. The con-

tact metamorphic aureole is a particularly favorable

environment for inﬁltration metamorphism because of

the heat source that can drive advective ﬂow. Two sce-

narios can be considered, again using the same isobaric

diopside-producing reaction as in Figure 16.19a.

In one case (I in Figure 16.19b), as before, a rock

composed of tremolite  calcite  quartz is heated

above 400°C but is simultaneously ﬂooded by a mobile

ﬂuid from a uniform external source whose composi-

tion is X  0.5. At T  525°C (or above as required

by necessary overstepping), diopside begins to be pro-

duced from the reactant phases, while the ﬂuid com-

position remains the same. Under these conditions the

perfectly mobile ﬂuids CO

and H

O are not counted

as components and consequently not counted as phases

either. Hence, the isobaric system is invariant because

it consists of three components (CaO, MgO, and

SiO

), four (solid) phases, and with T still an indepen-

dent variable, F  1  C 1  3  4  0. Any

additional heat input into the invariant system is con-

sumed by the endothermic devolatilization reaction

while T remains ﬁxed at 525°C; more diopside is pro-

duced until one of the reactant phases is consumed.

Once this happens, the heating system can continue to

get hotter, leaving its invariant state on the reaction line

and rising along a vertical trajectory. This open-system

process involving a ﬁxed ﬂuid composition is basically

a discontinuous reaction that occurs at some speciﬁc T.

In the second scenario (II in Figure 16.19b), sup-

pose a rock at 500°C composed of tremolite  calcite

 quartz is suddenly invaded by essentially pure water

of the same T, perhaps by hydraulic fracturing in a

contact aureole or by pervasive advective ﬂow along

permeable relict bedding planes. As the ambient ﬂuid,

with a composition X  0.5, mixes with the greater

amount of inﬁltrating aqueous ﬂuid, the rock system

follows a horizontal trajectory in Figure 16.19b that

eventually intersects the univariant curve where the

ﬂuid is about X  0.3, producing diopside. In this

isobaric–isothermal situation, the changing composition

of the ﬂuid drives the devolatilization reaction. As more

aqueous ﬂuid invades the rock the composition of the

Metamorphic Mineral Reactions and Equilibria

497

V  0.105 moles  44 cm

/mole

 4.62 cm

per 100 cm

of rock

V 0.015 moles  23 cm

/mole

0.345 cm

per 100 cm

of rock

The remainder of the volumetric modal changes are

found in Problem 16.8. The changes in modal pro-

portions of solid phases are so small as to be barely

noticeable in thin section, despite the reaction pro-

gressing over 225°C with substantial changes in

ﬂuid composition.

Consider next an open system subjected to

inﬁltration metamorphism at 525°C and 5 kbar

where the composition of the equilibrium ﬂuid is

X  0.52. In this system, the amount of inﬁltrat-

ing ﬂuid, n , is obtained from the equation

The inﬁltrating ﬂuid is assumed to be X  0.2

and the amount of tremolite produced in the reac-

tion is assumed to be 20 modal percent. From the

ﬁrst basic equation above, n

 n



(mole/272.92 cm

) 20 cm

 0.07328 mole of

tremolite in 100 cm

of rock. Substituting this value

in the equation above

 0.07328

 0.07328 = 0.888 mole

Since the molar volume of the inﬁltrating ﬂuid is

approximately 25 cm

/mole (Spear, 1993, Figure

7.21) the total volume of the inﬁltrating ﬂuid is

0.888 mole  25 cm

/mole  22.2 cm

per 100 cm

of rock. In other words, the ﬂuid/rock ratio is

about 0.22.

For a more CO

-rich inﬁltrating ﬂuid the ratio

would be greater, whereas for a more water-rich

ﬂuid the ratio would be smaller.

–3.88

0.32

(0.52{–1}  0.48{7})

(0.52  0.2)

total





 X



)

 X

)

total

mixed ﬂuid should change so that X  0.3; how-

ever, this cannot occur until one of the reactant phases

is consumed. Once this happens, still at 500°C, the sys-

tem can move off the univariant reaction curve and the

ﬂuid can track towards pure water, X  0.0. En

route, the aqueous ﬂuid might encounter a rock com-

posed of calcite and quartz, inducing a decarbonation

reaction that produces wollastonite. This conversion at

500°C contrasts with a possible 710°C production of

wollastonite in the open system whose ﬂuid composi-

tion is X  0.75 described above.

Clearly, the actual progress of devolatilization reac-

tions in mixed carbonated and hydrated solid-phase

assemblages can be very complex in space and time.

Isograds, metamorphic grades, and mineral reactions

in terranes of interlayered carbonate and silicate rocks

and impure carbonate rocks are governed by the com-

position of mixed C–O–H ﬂuids as much as, or more

than, they are by P and T. Some reactions can even

occur isothermally and isobarically where aqueous ﬂuids

inﬁltrate the rock; the ﬂuid is an active, rather than

passive, agent in the metamorphism and participates

both thermodynamically and kinetically. The possible

reaction paths in terms of P, T, and X

ﬂuid

are endless

but their elucidation provides rich insights into

ﬂuid–rock interactions during metamorphism. Broader

petrologic implications of ﬂuid movement during

metamorphism are explored in the next section.

16.7 FLUID FLOW DURING

METAMORPHISM OF THE

CONTINENTAL CRUST

It is obvious from the preceding discussions that the

accessibility of ﬂuids to reacting mineral systems can be

crucially important in their outcome. Accessibility, in

turn, can be dependent on the ﬂuid source, ﬂow prop-

erties of the rock (how the ﬂuid moves), how much

ﬂuid advects, and so on. Flow rates and ﬂuid volumes

relative to the affected rock volume are important in

determining whether a reacting system is closed and

internally buffered or open to large amounts of inﬁltrat-

ing solutions.

Evidence for large-scale ﬂuid ﬂow in continental

metamorphic terranes came about mainly through the

investigations of J. M. Ferry in the 1980s on ﬂuid–rock

interactions in thinly bedded metasedimentary rocks in

New England. For a lucid summary see Ferry (1994).

Ironically, this realization lagged behind the discovery

of huge advecting systems in far less accessible oceanic

spreading ridges. Although large ﬂuxes of hydrous

ﬂuid in many metamorphic terranes are widely ac-

cepted, some experienced metamorphic petrologists

remain skeptical and urge caution. For example, Yard-

ley (1997, p. 111) concludes, “it is likely that for most

of the time metamorphosing rocks are of very low per-

meability and contain small amounts of ﬂuid at near-

lithostatic pressures....The case for large pervasive

ﬂuxes of reactive ﬂuid through metamorphic sequences

is ﬂimsy, and based on selective arguments.”

16.7.1 Evidence for Fluid Flow

The concept of large ﬂuid ﬂuxes during metamorphism

of continental rocks is supported by the following lines

of petrologic evidence.

Discharge from Geothermal Wells. As much as 10

of saline water per hour at temperatures of 200–400°C

have been recorded, providing proof that hydrother-

mal ﬂuids not only exist but can readily ﬂow through

rock into drill holes kilometers in depth. Authigenetic

calcite and “sericite” (illite, phengite, muscovite) ce-

mented sandstones in the Salton Sea geothermal ﬁeld

near the California–Mexico border have experienced

decarbonation reactions in alkali chloride brines (X

0.05) yielding, at increasing T, the minerals wairakite,

epidote, prehnite, and clinopyroxene (see Special

Interest Box 18.1). Mass-balance calculations based on

oxygen isotope compositions of ﬂuids and rocks indic-

ate a time-integrated ﬂuid/rock ratio of about 3.

Veins in Metamorphic Terranes. Veins offer strong evid-

ence for channeled ﬂow of ﬂuids though fractured rock.

They are most commonly of quartz (Figure 15.23), but

locally include other minerals, such as feldspar, epi-

dote, and carbonates. Mineral assemblages in some

veins are similar to assemblages in the host wall rock,

suggesting formation during peak metamorphic condi-

tions and possibly local interchange of material via a

ﬂuid into an opening ﬁssure. In this situation, indi-

vidual grains locally extend continuously from the wall-

rock into the vein interior. In other veins, the mineral

assemblage is inconsistent with the immediate wall

rock and transport of dissolved constituents from a dis-

tant external source is implied. This would apply, for

example, to veins of pure quartz in host marble. Wide-

spread ﬂuid inclusions (Special Interest Box 16.3) and,

where present, minerals coexisting with quartz can be

used to infer P–T conditions of vein emplacement.

A number of petrologists have considered the for-

mation of quartz veins and all have come to the same

conclusion that large amounts of ﬂuid, principally

aqueous solutions, are required, despite the fact that

quartz is one of the most soluble minerals under meta-

morphic conditions. Its solubility decreases signiﬁc-

antly as both P and T decrease (Figure 16.20). Hence,

quartz generally precipitates as silica-laden ﬂuids rise

into shallower crust. For example, Spear (1993) calcu-

lates that to precipitate a 1 cm-thick vein for every

1 cm

area on opposite walls (1 cm

volume) there

must be 3.3  10

of aqueous solution ascending

to lower P and T from 5 kbar and 500°C.

498 Igneous and Metamorphic Petrology

Stable Isotope Studies. Since pioneering investiga-

tions in the late 1960s by Taylor (see 1974), it has

been known that extensive

O–

O exchange (Section

2.6.1) occurs between advecting heated meteoric water

and country rock surrounding granitic intrusions in the

shallow crust (Figures 4.12 and 4.14). If the isotopic

composition of the unaltered protolith and the perco-

lating ﬂuid are known, the time-integrated ﬂuid/rock

ratio can be determined from mass balance calcula-

tions. For many advecting systems this ratio is 1.

Fluid-Induced Mineral Reactions in Deep Crustal Shear

Zones. In the vast granulite-facies terrane of western

Norway near Bergen (Jamtveit et al., 1990), the occur-

rence of eclogite is invariably spatially related to ex-

tensional veins and kink bands (see Figure 17.17) in

cm- to m-thick shear zones. Inﬁltration of water-rich

ﬂuids (X  0.25) in these zones transformed the anhy-

drous granulite assemblage of plagioclase  Al-rich

pyroxenes  garnet into a somewhat hydrated assem-

blage of omphacite  garnet  kyanite  clinozoisite 

phengite  amphibole  quartz. Consequently, once

eclogitization into a smaller volume mineral assemblage

was initiated, the enhanced permeability resulted in a

positive feedback loop, promoting the inﬁltration of

additional aqueous ﬂuid.

Hydration-Decarbonation Reactions in Contact Aureoles.

Hydrating and accompanying decarbonating mineral

reactions occur in heated calcareous rocks in contact

aureoles surrounding granitic intrusions; expelled

hydrothermal solutions act in concert with increased

temperatures to drive the mixed-volatile reactions.

Formation of metasomatic skarns by silicating hydro-

thermal solutions at Crestmore, California, was pre-

viously described (Section 14.2.5). Here, we describe

a more common type of open-system metamorphism

where an inﬁltrating aqueous ﬂuid permeated quartz-

bearing dolomite rocks in the Ubehebe Peak contact

aureole, California (Figure 16.21). Other than the

mobile ﬂuid, components (Si, Ca, Mg, Al, and so on)

appear to have remained essentially immobile in these

calc-silicate hornfelses, except on a grain scale during

recrystallization.

The presence of tremolite rather than the more

water-rich talc in the outer lowest T part of the aureole

might suggest lowest T metamorphism at 1–2 kbar

(based on stratigraphic reconstructions; Roselle et al.,

1997) in the presence of CO

-rich ﬂuids, rather than

aqueous. However, prograde reactions, including those

in closed systems, with such ﬂuids would yield diopside

and then forsterite (Figure 16.18). However, diopside

is almost nonexistent in the aureole. In addition, peri-

clase of the inner part of the Ubehebe aureole could

not be stabilized in ﬂuids that contain very much CO

Kinetic factors or retrograde effects might account for

the absence of talc in the outermost part of the aureole.

Stabilization of the hydrous mineral tremolite in

the lower grade hornfelses in itself indicates inﬂux of

water into the “dry” siliceous dolostone protolith but

its “blossomlike” form centered along relict cracks in

the rock offers further evidence for the inﬁltration

of water (Figure 16.22). Moreover, the sequence of

mineral zones toward the intrusion, namely tremolite,

forsterite, and periclase (retrograded to brucite), shown

in Figures 16.21 and 16.23, is consistent with inﬂux of

Metamorphic Mineral Reactions and Equilibria

499

Log moles SiO

/kg aqueous solution

1

2

3

200 400 600 800

0.5 kbar

1000

T (°C)

16.20 Solubility of quartz in aqueous solution. Data from Manning

(1994). In terms of wt.%, solubilities range from about 0.1 un-

der low-grade metamorphic conditions to about 5.0 under

high-grade conditions in the middle continental crust. Walther

and Wood (1984) note that quartz solubility is markedly re-

duced with increasing X in the solution.

Special Interest Box 16.6 Decarbonation

reactions and CO

ﬂux into the atmosphere:

paleoatmospheric inﬂuences

Kerrick and Caldeira (1993; see also Ingebritsen

and Manning, 1999) have postulated that metamor-

phism during orogeny can release sufﬁcient CO

via decarbonation reactions to signiﬁcantly alter the

atmospheric concentration and cause as much as

2°C global warming. The calculated CO

ﬂux is on

the order of 10

g/y; this may be compared with

potential diffuse degassing of the Earth (10

g/y),

ﬂux at oceanic ridge systems (3.5  10

g/y), and

from volcanic arcs (10

g/y). Continental exten-

sional regimes are also a potential site for signiﬁcant

ﬂux into the atmosphere from decarbonation reac-

tions because of deeply circulating meteoric water

in the fractured crust.

aqueous ﬂuids that ﬂushed out liberated CO

so as to

maintain a sufﬁciently high water activity during meta-

morphism. That this is so can be appreciated from Fig-

ure 16.24, which shows relevant univariant reaction

curves and equilibria appropriate for siliceous dolo-

stones (10% quartz, 90% dolomite). In an open system

at 1 kbar ﬂooded by water-rich ﬂuid (X  0.1), dis-

continuous prograde reactions produce tremolite near

415°C, forsterite at about 470°C, and periclase near

640°C. These temperatures compare favorably with

calcite–dolomite solvus thermometry in the hornfelses

determined by Roselle et al. (1997): 475°C in the upper

tremolite zone, 475–500°C in the lower to middle

forsterite zone, and 620°C at the periclase-in isograd.

The irregularity of the periclase isograd and layer-

speciﬁc metamorphism in the Ubehebe aureole suggest

heterogeneous inﬁltration of ﬂuids. O- and C-isotope

ratios strongly support nonuniform inﬁltration of mag-

matic ﬂuids and conﬁrm their water-rich composition,

X  0.3 (Roselle et al., 1999).

16.7.2 Mechanics of Fluid Flow

In order for ﬂuid to move through solid rock there must

be both openings (porosity) that are interconnected so

500 Igneous and Metamorphic Petrology

Ubehebe

Peak

0 km 0.8

0 miles 0.5

Intrusion

36°

42′

30″

117°35′

Tremolite zone

Alluvium

Periclase

Zone

16.21 Ubehebe Peak contact metamorphic aureole in siliceous dolostones, Death Valley National Monument, California. In contrast to the

broad tremolite and forsterite zones, the periclase zone (darkest shade) is thin and impersistent adjacent to the granitic intrusion.

Quaternary alluvium (Qa) enclosed by dotted line. Inset in upper right shows location of Death Valley National Park in the state of

California. Pz, Paleozoic rocks. Redrawn from Roselle et al. (1999).

16.22 “Blossoms” of tremolite grown below a quartzite layer into

dolomarble, Ubehebe Peak contact aureole, Death Valley

National Monument, California (Baumgartner et al., 1997).

Tremolite growth was positioned along cracks where advect-

ing aqueous ﬂuids inﬁltrated into the rock and is believed to

have formed by the reaction (see Figure 16.24) 5 dolomite

 8 quartz  H

O  tremolite  3 calcite  7CO

. Coin is

24 mm in diameter. Photograph courtesy of Gregory

T. Roselle.

as to allow transfer of the ﬂuid (permeability), and a

driving force to propel the ﬂuid.

Porosity and Permeability. Other than widespread frac-

tures that were mostly produced during uplift and

stress relief, metamorphic rocks exposed at the surface

of the Earth possess only a trivial porosity, generally a

few tenths of 1% by volume. Because most openings in

the rock existing during its metamorphism deep in the

crust have vanished, careful observations of rock fabric

combined with laboratory investigations simulating

metamorphic conditions are necessary to determine the

origin and nature of porosity and permeability during

metamorphism (Brenan, 1991).

Openings can be inherited from the protolith—such

as pore spaces between clasts and vesicles in volcanic

rocks—but these generally become closed by increas-

ing P during burial and by ﬁlling with mineral precipi-

tates (e.g. amygdules).

Experiments on rocks at elevated P and T indicate

the character of very small volumes of ﬂuid that can ex-

ist along grain boundaries in metamorphic rocks. The

basic principle underlying this transient intergranular

porosity and possible permeability involves the rela-

tions between surface energies of crystalline grains and

the ﬂuid as formulated in the dihedral angle,  (Section

11.2.2 and Figure 11.7). At equilibrium, no natural

ﬂuid–solid pair exists for which 0°, ensuring that

no ﬂuid ﬁlm can form a continuous grain-boundary

ﬁlm covering the entire grain. Under some conditions,

interconnected threadlike volumes of ﬂuid may exist

along three-grain tubular edges for 0° 60° (Hol-

ness, 1997). In pure calcite aggregates containing water,

 60° at any P and T but with addition of NaCl or in

nearly equimolar H

O–CO

ﬂuids  60° but only at

low P and probably high T. Hence, relatively pure car-

bonate rocks are generally impermeable except where

special ﬂuids exist or where the rock is being deformed

by ductile ﬂow or by fracturing. In silicate rocks, dis-

solved NaCl and feldspar components can reduce  to

60°. In most metamorphic rocks, ﬂuids will generally

occur in isolated “pools” at multiple grain corners

where  60°. Of course, these spatial relations do not

apply where the ﬂuid/rock ratio may be large during

the course of an active devolatilization reaction.

An important means of creating openings in rock

undergoing metamorphism is by reaction-enhanced

permeability (Rumble et al., 1982). This phenomenon

is a consequence of the fact that the volume of solid

products is generally less than the volume of solid

reactants undergoing devolatilization. For dehydration

reactions involving silicates, V may not be very great

Metamorphic Mineral Reactions and Equilibria

501

Tremolite

Brucite

Calcite

Dolomite Forsterite Calcite Dolomite

0 0.5mm 0 1mm 0

(a) TREMOLITE ZONE

5 dolomite  8 quartz  H

O →

tremolite  3 calcite  7CO

(b) FORSTERITE ZONE

11 dolomite  tremolite →

8 forsterite  13 calcite

 9CO

 H

dolomite → periclase 

calcite  CO

(Retrograde:

periclase  H

O → brucite)

0.5mm

16.23 Contact metamorphosed siliceous dolostones from the Ubehebe Peak contact aureole, Death Valley National Monument, California.

See also Figure 16.21. Drawn from annotated photomicrographs supplied through the courtesy of Gregory T. Roselle. Possible reactions

progressively consuming (from left to right) dolomite and yielding calcite are indicated below drawings. (a) Lowest grade tremolite

( calcite) zone believed to have formed by reaction between dolomite and minor quartz that was entirely consumed but leaving excess

dolomite. Exact reaction is uncertain as there is no lower grade talc zone as expected where aqueous ﬂuids drove the mixed-volatile ﬂuid

reactions. (b) Forsterite zone. Large robust forsterite prisms and surrounding calcite grains were developed in the outer part of the

aureole whereas much smaller, equant anhedral forsterite grains were produced closer to the intrusion (see Figure 17.3). Note diameter

of ﬁeld is twice that of other two views. (c) Brucite–calcite rock from the periclase zone immediately adjacent to the intrusion. Periclase

in contact metamorphosed carbonate rocks is almost invariably hydrated during cooling to form platy aggregates of brucite. Note sparse

excess dolomite.

—about 3% for the reaction muscovite  quartz →

sillimanite  K-feldspar  water. On the other hand,

V for decarbonation reactions can be much larger—

about 33% for calcite  quartz → wollastonite  CO

Once again, such porosities are only transient during

the course of reactions and compaction processes are

likely to expel ﬂuids quite rapidly.

Fracturing is an obvious means of inducing open-

ings in otherwise porosity-free brittle rocks. These

cracks can either be interconnected or connect to other

types of openings, resulting in rock permeability. Obvi-

ously, channeled ﬂow in fractures provides the most

efﬁcient and rapid means for ﬂuid transport. Fractures

are produced in at least three ways:

502 Igneous and Metamorphic Petrology

T (°C)

500

600

500

(c)

400

0.0 0.2 0.4 0.6 0.8 1.0

SiO

Qtz

(b)

Tlc

Per

MgODol

CaO

Cal

T (°C)

450

350

650

550

0.0 0.2 0.4 0.6 0.8 1.0

1.0 0.20.40.60.8 0.0

(a)

2 kbar

ﬂuid

 5 kbar

ﬂuid

 1 kbar

1 kbar



l



700

16.24 Mixed-volatile ﬂuid equilibria in the system CaO–MgO–SiO

–CO

–H

O depicting possible reactions in metamorphism of siliceous

dolostones. (a) Petrogenetic grid showing univariant reaction curves at P

ﬂuid

 1 kbar. Dashed line segment on right is the metastable ex-

tension into the tremolite stability ﬁeld of the univariant reaction line for dolomite  2 quartz  diopside  2CO

that is too short where

stable to the right of invariant point B to allow labeling. Univariant reaction lines converging into the three invariant points, such as A at

420°C and X  0.72, that are thinner and unlabeled are unimportant for the purposes of this diagram (for these other

reactions see Slaughter et al., 1975). Triangles in each ﬁeld indicate stably coexisting phases keyed to larger labeled triangle in (b) where

circle indicates quartz-bearing dolostone protolith. (c) Lower portion of grid from (a) but P

ﬂuid

 5 kbar to show leftward shift of

invariant point B and especially A with respect to increasing P, greatly expanding the range of ﬂuid compositions over which tremolite

can be created as the initial product of metamorphism rather than talc. Note the invariant points have also shifted to higher T. Redrawn

from Slaughter et al. (1975) and Ferry (1991).

1. Intercrystalline mismatches during heating as

anisotropic grains, especially quartz, expand differ-

entially in different crystallographic directions (Section

8.3.1). In effect, this is “thermal decompaction.”

2. Nonhydrostatic stresses during tectonic (orogenic)

activity that exceed the brittle strength of the rock

(Section 8.2.1).

3. Fluid overpressures surrounding magmatic intru-

sions (Section 9.1.2) and in rocks undergoing

devolatilization reactions. Metamorphic rocks typic-

ally have insufﬁcient permeability to accommod-

ate the rate of ﬂuid production in devolatilization

reactions. Consequently ﬂuid pressure will build

up, exceeding the tensile brittle strength of the

rock, creating hydraulic fractures (Section 8.2.1

and Figure 8.2). In pioneering experiments, Rayleigh

and Paterson (1965) found that the strength of ser-

pentinite dropped by almost an order of magnitude

with increasing T as it partially dehydrated at

500–600°C to olivine  talc, releasing several wt.%

water. Rather than becoming more ductile, as occurs

during heating of volatile-free rocks, the weakened

serpentinite was embrittled and lost cohesion as the

released water collected into hydraulically created

openings.

On a grain scale, fractures are preserved as healed

microcracks, or planar secondary ﬂuid inclusions, that

appear as either single entities or as trains of many in-

line parallel inclusions that were ﬁlled with ﬂuid before

being pinched off by recrystallization (Figure 16.25).

Larger scale fractures ﬁlled with minerals precipitated

from ﬂowing ﬂuids are manifest by veins whose thick-

nesses are measured in centimeters or meters and

lengths in meters to hundreds of meters (Figure 15.23).

Crack-seal veins reveal multiple episodes of ﬂuid

ﬁlling, precipitation of dissolved solids, renewed crack-

ing, ﬁlling, and so on.

Because prograde devolatilization reactions occur at

distinct P–T conditions during metamorphism, related

porosity and permeability production and enhance-

ment will be episodic and transient. Devolatilization

can enhance ﬂuid ﬂow and ﬂuid inﬁltration can induce

devolatilization reactions. Thus, these can be mutually

reinforcing or interactive feedback processes. Once

the rocks have passed through their peak thermal con-

dition and have experienced irreversible ﬂuid loss,

further hydraulically induced fracturing will commonly

not occur, effectively sealing the rocks to any sub-

sequent ﬂuid inﬁltration. Consequently, the peak meta-

morphic imprint is preserved as T falls because of the

lack of ﬂuids to accomplish retrograde volatilization re-

actions.

Porosity and permeability production are also likely

to be spatially heterogeneous because of the hetero-

geneity of rocks in metamorphic terranes and the

episodic nature of reaction-enhanced permeability in

both space and time.

Driving Force for Fluid Flow. The ﬂow of any liquid

through a solid material possessing interconnected

openings can be quantiﬁed by Darcy’s Law. For ﬂow in

one dimension (the direction x) this is

16.26 V

 

where  is the viscosity, about 10

4

Pa s for H

O–CO

ﬂuids in most metamorphic systems, according to

Walther and Orville (1982). V

is the volumetric ﬂux,

or volume of ﬂow per unit time per unit area of mater-

ial traversed; it has units of m

s  m/s and is,

therefore, justiﬁably called the Darcy velocity. The

actual velocity of the liquid through small grain-scale

openings in rocks is much greater. The permeability,

k, has units of m

and ranges over many orders of

magnitude in rocks, but at depths of more than about

12 km is apparently close to 10

18

(Ingebritsen and

Manning, 1999) where no devolatilization reactions are

actively going on.

The dP/dx term in Darcy’s Law is the pressure gra-

dient that drives the ﬂow. In many metamorphic sys-

tems, this gradient arises from a thermal gradient that

causes uneven expansion of a ﬂuid, imparting buoy-

ancy. For geothermal gradients 20°C/km the density

of water decreases with depth (Figure 4.3) so it be-

comes more buoyant, driving water upwards through

permeable rock in the crust. However, ﬂuid ﬂow in the



≤¢

≤

Metamorphic Mineral Reactions and Equilibria

503

16.25 Healed microcracks (secondary ﬂuid inclusions) in quartz,

Proterozoic Wolf River Batholith, Wisconsin. At least two sub-

parallel sets of cracks are evident at an angle of about 60° to

one another. Photomicrograph courtesy of Bart J. Kowallis.

continental crust is not everywhere vertically upward

because a large component of geothermal water and

advecting water around shallow magmatic intrusions is

meteoric. Moreover, the pattern of ﬂuid inﬁltration in

some contact metamorphic aureoles indicates subhor-

izontal channelized ﬂow in highly permeable layers.

Fluid ﬂow can be focused in fault zones where a

different sort of pressure gradient is produced by dila-

tancy pumping. As stress builds up immediately prior

to faulting, rocks experience a perceptible increase in

volume (dilatancy) as a result of pervasive microcrack-

ing. The opening cracks are local sites of reduced pres-

sure and consequently suck any nearby ﬂuid into them.

But once failure occurs, stress is released, microcracks

collapse, and ﬂuid is driven out. This pumping action

and other possible mechanisms may cause ﬂuid ﬂow in

ductile shear zones (marked by mylonite) as well as

more brittle faults (Ord and Oliver, 1997).

It is entirely possible that ﬂuid ﬂow would be

limited in metamorphic rocks without the effects of

reaction- and deformation-enhanced permeability.

16.8 METASOMATISM

Most metamorphic reactions involve movement of

ions on the scale of consuming reactant and growing

product grains. However, the foregoing pages have

demonstrated that large-scale ﬂuid movement occurs

in metamorphic terranes. In most instances, these

inﬁltrating ﬂuids are likely to react to some degree with

the rocks through which they ﬂow. The term meta-

somatism is conventionally applied where the distance

between the source and sink of the transported mater-

ial is more than the scale of individual grains and there

is a signiﬁcant change in the bulk chemical and modal

composition of the protolith. (Compare cryptic meta-

somatism in mantle rocks, Section 11.2.2.) Metasomat-

ism has been placed in a thermodynamic context since

the middle twentieth century by J. B. Brady, G. W.

Fisher, D. S. Korzhinskii, J. B. Thompson, and others.

For reviews see Barton et al. (1991) and Winter (2001,

Chapter 30).

Metasomatism can be driven by differences in

chemical potential between adjacent compositionally

contrasting rock volumes and accomplished by diffu-

sion at relatively high T over prolonged periods of

time. An example is the so-called bimetasomatism at

contacts between silicate and carbonate-bearing rock

where, to a ﬁrst approximation, Si has moved from the

former into the latter and Ca in the opposite direction.

Most metasomatism, however, is related to inﬁltration

of ﬂuids into highly permeable rocks. This is especially

common in, but by no means restricted to, contact

metamorphic aureoles and oceanic ridges.

Interpretations of metasomatic rocks generally must

deal with three unknowns: the character of the protolith,

the gains and losses of elements (what chemical

species were mobile versus immobile, and which were

conserved), and the metasomatic process (whether

constant volume or not). Without some independent

information, the composition of the metasomatic rock

alone cannot provide answers to all of these unknowns.

It is commonly assumed that certain chemical con-

stituents remained immobile during metasomatism,

such as Al

, TiO

, and high-ﬁeld-strength trace

elements (Figure 2.20), but this assumption can be a

pitfall unless tested and evaluated in some way. Field

relations and rock fabric can provide constraints in

some cases, as in resolving the long-standing dilemma

regarding serpentinization of ultramaﬁc rocks (Section

15.2.7). In contact metamorphic aureoles, unlike re-

gional terranes, distinct rock layers can often be traced

along their strike from unaltered protoliths a kilometer

or less from the intrusion into metasomatized rocks

nearer the magmatic contact.

Some petrologists have argued for removal of as

much as 20 wt.% of the total nonvolatile mass of pelitic

rocks during Barrovian regional metamorphism. How-

ever, most believe that less than 3 wt.% of nonvolatiles

are lost and, to a ﬁrst approximation, metamorphism

is isochemical in terms of nonvolatile elements but

metasomatic in terms of perfectly mobile volatiles. In

this controversy, it is difﬁcult to accurately evaluate the

amount of mass transferred out of a regional terrane

because of the heterogeneity of the protoliths and their

uncertain initial chemical compositions (Walther et al.,

1995; Vernon, 1998). Another impediment is how to

devise a rigorous sampling plan that can accurately and

reliably test the amount of transfer.

16.8.1 Ion Exchange Reactions in Open

Metasomatic Systems

Most rock-forming silicate minerals and their chemical

components are only very slightly soluble in H

O–CO

ﬂuids; thus, very large volumes are required to trans-

port signiﬁcant masses of dissolved silicates, as dis-

cussed above for quartz. Solubilities depend on the

composition of the ﬂuid, P, and T (Labotka, 1991).

Greater Cl concentrations in aqueous ﬂuids can pro-

mote greater solubilities, especially for alkalies (Na,

K, Rb). Salts of Cl



, SO

2

, and CO



, such as KCl,

CaSO

, and Na

, have signiﬁcantly greater solubil-

ities in aqueous ﬂuids than silicates and consequently

greater amounts of these materials can be transported

by ﬂuid ﬂow. As carbonates, halides, and sulfates are

the very materials generally found in ﬂuid inclusions in

metamorphic minerals (Special Interest Box 16.3), they

are likely to have been involved in the solid–ﬂuid reac-

tions occurring during metamorphism.

Mineral components can have a source in the rocks

through which ﬂuids percolate, dissolving or leaching

out the more soluble, mobile components. Or the

504 Igneous and Metamorphic Petrology

components may have been derived from nearby mag-

matic sources, such as silica in the metasomatizing

solutions at Crestmore, California (Section 14.2.5).

In an ion-exchange reaction, ions in a ﬂuid ex-

change with other ions in neighboring solids without

destroying them. This reaction is modeled as follows:

16.27 AX  B



 BX  A



solid in ﬂuid solid in ﬂuid

A familiar example of an ion-exchange reaction is the

water-softening action of some Na-bearing zeolites in

which dissolved Ca

2

, which causes “hardness” in

water, is taken into the zeolite in exchange for two



, which pass from the zeolite into the water.

Common alkali-ion-exchange reactions in metamor-

phic rocks involve micas and feldspars as these are the

chief repositories of Na and K. Many different reac-

tions can be written but for muscovite breakdown two

possibilities include:

16.28 KAl

AlSi

(OH)

 6SiO

 2K



muscovite quartz in ﬂuid

 3KAlSi

 2H



K-feldspar in ﬂuid

16.29 2KAl

AlSi

(OH)

 2H



muscovite in ﬂuid

 3Al

SiO

 3SiO

 3H

O  2K



quartz in ﬂuid solution

For reaction 16.29 where the solid phases and water are

pure with activities  1 the equilibrium constant, K

, is

16.30 K



showing that the stabilities of micas and alkali feldspars

in aqueous solutions depend on ratios of activities at a

speciﬁed P and T (Figure 16.26). High H



concentra-

tions relative to K



and Na



stabilize an aluminosilic-

ate, whereas the opposite stabilize feldspars. Micas are

stable at intermediate alkali ion/hydrogen ion ratios.

Hydrogen metasomatism is particularly common

and widespread around granitic intrusions in the shal-

low continental crust. The high H



activity relative to

other major elements causes removal of Na, K, Ca, Mg,

and Fe in feldspars and maﬁc silicates in felsic country

rocks as well as the intrusion. In the initial sericitic

alteration stage, white micas  clay minerals are stabil-

ized at the expense of feldspars. Clay minerals, such as

kaolinite, are stabilized at the expense of K-feldspar

according to the reaction

16.31 2KAlSi

 2H



 H

 Al

(OH)

 4SiO

 2K



More aluminous chlorites and epidotes replace biotite,

amphibole, and pyroxene. In more advanced argillic

alteration, feldspars and micas are converted into

pyrophyllite, andalusite, and other aluminosilicates.

Where the activity of S is sufﬁcient, Fe from break-

down of maﬁc silicates and Fe–Ti oxides produces

widespread pyrite. At higher T, greisens may develop

(Section 15.2.8) that contain aluminosilicates such as

topaz (high F activity) and tourmaline (B). Extensive

halos of sericitic and argillic alteration around more

localized hydrothermal ore deposits provide valuable

exploration targets.

16.8.2 The Thompson Model of Metasomatic Zoning

and Local Equilibrium

For a system to be in complete thermodynamic equi-

librium, all intensive variables—T, P, and chemical

potentials of every component—must be uniform

throughout. No incompatible phases, such as quartz

and forsterite, can exist together. However, as Thomp-

son (1959, p. 430) points out, “though a large volume

of rock commonly contains mutually incompatible

phases and is thus not in internal equilibrium, it is

generally possible to regard any part of such a thermo-

dynamic system as substantially in internal equilibrium

if that part is made sufﬁciently small.” This expression

of the concept of local equilibrium has a signiﬁcant

Metamorphic Mineral Reactions and Equilibria

505

Paragonite

600°C

500°C

650°C

Muscovite

K-feldspar

Albite

Kfs

SiO

 Qtz



SiO



16.26 Phase relations in the system Na–K–Al–Si–O–H as a function

of T and the ratios of alkali and hydrogen ions at 5 kbar and

a  0.5. Ion-exchange reactions 16.28 and 16.29 are top

and bottom subhorizontal lines, respectively. Note shrinking

of muscovite stability ﬁeld with increasing T (light shaded at

500°C, medium shaded at 600°C, and dark shaded at 650°C)

and absence of paragonite ﬁeld at 600 and 650°C. Compare

Figure 14.31. Redrawn from Wintsch (1975).

bearing on the fate of juxtaposed local systems that

equilibrate via diffusion-controlled reactions to pro-

duce sharply bounded mineral layers of a small number

of phases. Juxtaposed incompatible systems include

vein solutions markedly out of equilibrium with wall-

rock mineral assemblages and bedded carbonate and

silicate rocks deposited in a sedimentary environment

that are unstable under metamorphic conditions.

To illustrate the concept of local equilibrium,

Thompson (1959) used a hypothetical example at ﬁxed

P and T in which a region of pure periclase is separated

from a region of pure quartz by three two-phase zones

in which the bulk composition varies continuously and

linearly (Figure 16.27, top). Across each of these two-

phase zones, such as periclase  forsterite, the modal

proportions of each phase varies inversely and con-

tinuously so that the variation in bulk composition is

continuous. Within each zone, as well as in the two

end regions, local equilibrium prevails. Each of the

two-phase assemblages is stable and no reaction should

occur within them. However, along zone boundary F, a

few grains of periclase in the left zone are in contact

with enstatite grains in the right zone. As this pair of

phases is less stable than forsterite, an intermediate

one-phase zone of that mineral may form, resulting in

a local reduction of the number of phases. Similarly,

a zone of pure enstatite may form in place of the

interface E. The extent to which these monomineralic

forsterite and enstatite zones replace the neighbor-

ing two-phase zones would depend upon the available

time for the diffusion of Si and Mg. Ultimately, the

diffusion process might produce the stepped chemical

conﬁguration of Figure 16.27, bottom, in which only

monomineralic forsterite and enstatite zones occur

between the periclase and quartz regions. Four areas of

local one-phase equilibrium now exist.

A natural example of local equilibrium in a diffusion-

controlled metasomatic system is the reaction at the

interface between nodules of chert (microcrystalline

quartz) and host dolostone protolith. In this case, how-

ever, the system is not binary and reaction layers need

not be monomineralic. Joesten (1991) has provided

a detailed thermodynamic analysis of the chemical

potential gradients that drive diffusional exchange

of CaO, MgO, and SiO

across the incompatible

dolomite–quartz contact. Depending on the P–T con-

ditions, buffering mineral reactions, and diffusion

coefﬁcients different types of layers and layer thick-

nesses develop (Figure 16.28).

16.8.3 Low-Variance Assemblages

in Metasomatic Rocks

Because metasomatism is accomplished by inﬁltration

of relatively large amounts of ﬂuid, at least it and one

or more dissolved components will be perfectly mobile

in the sense deﬁned in Section 16.6.3. Inasmuch as

perfectly mobile components are not counted in the

phase rule, the variance of the system is less than if the

components were not mobile. The mineralogical phase

rule derived in Section 15.3.1, C  , can be rewritten

as C  C

  for metasomatic systems, where C

the number of mobile components. Consequently, a

typical property of metasomatic mineral assemblages

is the relatively small number of solid phases, because

the number of components is small. Theoretically, all

of the components in a metasomatic system could be

mobile but that would mean that there would be no

phases (C  C

 0  ), which is impossible. So

at least one component must be immobile in a one-

phase assemblage; monomineralic assemblages tend

to be common in, for example, ﬂuid-dominated vein

systems.

16.8.4 Frames of Reference and the Isocon Diagram

If so many components are on the move in metaso-

matic systems, how is the petrologist to decide which

were actually mobile? Careful petrographic study can

help, as can examination of ﬁeld relations, especially

between the altered rock and what may be the protolith,

across isograds and other metasomatic transitions (for

example, Figure 16.28). With the proper frame of

reference, chemical changes might be ascertained. But

506 Igneous and Metamorphic Petrology

Periclase Forsterite Enstatite Quartz

Periclase

Forsterite

Enstatite

Periclase



forsterite



Enstatite



quartz

Quartz

Periclase

100

wt. %

SiO

Periclase



forsterite



Enstatite



quartz

Quartz

16.27 Hypothetical model of metasomatic zoning in system where

local equilibrium prevails. Redrawn from Thompson (1959).