Myron G. Best. Igneous and metamorphic 2003 Blackwell Science

Подождите немного. Документ загружается.

Evolution of Imposed Metamorphic Fabrics: Processes and Kinetics

557

Quartz grains

and their c-axes

in rock

Projection of c-axes

onto lower hemisphere and

stereographic projection

of a c-axes onto

equatorial plane

(a)

(b)

Individual

-axes

(c)

Contoured

stereographic

projection

(d)

17.39 Lattice preferred orientation in an aggregate of equant mineral grains. (a) Schematic of aggregate of quartz grains (shown as spheres for

simplicity) with preferred orientation of c-axes. (b) Orientation of the c-axis in each grain in (a) is drawn from the center of a sphere

whose equatorial plane (shaded) is oriented horizontally. From the point (solid circle) on the lower hemisphere where the c-axis

intersects it a projection line (dashed) is drawn to a point (*) on top of the sphere. The point where this projection line intersects the

equatorial plane (open circle) is called the stereographic projection of the c-axis. (c) Stereographic projection of c-axes from (a) forms a

bandlike pattern. (d) Contoured diagram with each line bounding an area of some speciﬁed minimum point density (Hobbs et al., 1976,

pp. 120–1, Appendix A).

A hypothetical rock that contains oriented c-axes is

shown in Figure 17.39. Lattice orientation can be

determined by means of a universal-stage attachment

on a polarizing petrographic microscope or by X-ray

diffraction techniques.

Lattice orientations are useful in determining the

pattern of the deformation path, or the strain history,

including shear sense asymmetry, the relative T of the

plastic strain (because different slip systems are activ-

ated at different temperatures), and the magnitude and

symmetry of the total (ﬁnite) strain in the tectonite

(Law, 1990). It is important to integrate the crystallo-

graphic data with other fabric data on the tectonite

(e.g. Special Interest Box 17.3 and Figure 17.40) and to

be aware of possible overprinting and obliteration of

earlier fabrics in polymetamorphic rocks.

Experiments and observations indicate that shape

orientation by pressure solution yields little if any lat-

tice orientation, so other mechanisms of ductile ﬂow

must be involved. Ample evidence exists for the efﬁcacy

of intracrystalline plasticity under nonhydrostatic stress

in producing lattice preferred orientations, especially

558 Igneous and Metamorphic Petrology

in quartz and calcite. Figures 17.15–17.17 and 17.19

reveal that slip, twinning, and kinking are capable of

rotating grain lattices in aggregates. Slip is especially

effective because large amounts of strain are possible,

allowing correspondingly strong preferred orientations.

Dynamic recrystallization during dislocation creep yields

new, more stable, smaller, unstrained grains or subgrains

by migration of dislocations and grain boundaries; the

new grains have more or less the same lattice orienta-

tion as the original plastically deformed ones. Static

annealing yields new grains with a complex relation to

the pre-existing strained and lattice oriented grains.

17.4.3 Cleavage, Schistosity, and Compositional Layering

A long-standing controversy in structural petrology has

been the origin of cleavage in slates and phyllites. A

related dilemma is the origin of compositional layering

that, although characterizing coarse-grained, high-

grade gneisses, also occurs in ﬁner grained, lower grade

rocks as well. Neither of these foliations imposed

onto the rock during metamorphism has been fully

accounted for. From symmetry, the imposed planar

fabric must develop from a planar pattern of strain

during deformation. Obviously, recrystallization plays

a role, as does the availability and movement of ﬂuids.

Cleavage and Schistosity. While simple progressive

mimetic recrystallization of shale yielding the meta-

morphic sequence of slate, phyllite, and schist—as

suggested at the end of Section 17.4.1—may take place

in some terranes, many foliated tectonites reveal a more

complex evolution. Overprinting effects of deformation

and recrystallization are common. Foliations range over

a wide spectrum of types and scales (Section 15.1.1),

from continuous to spaced, and including slaty, crenu-

lation, and disjunctive cleavages and counterpart schis-

tosity and crenulation schistosity in coarser tectonites.

A recent critical review of a vast number of studies by

Vernon (1998) forms the basis of the following.

However the fabric may be designated, slates, phyl-

lites, and schists derived from sedimentary protoliths

typically possess two alternating and locally anastomos-

ing domains (Figure 15.3; see also Borradaile et al.,

1982):

1. P-domains are folia of mostly oriented ﬂakes of

phyllosilicates (micas and chlorite) and more ﬁne

grained carbonaceous matter, locally graphite, and

Fe–Ti oxides. They contain only minor quartz

grains, none obviously detrital and all conspicu-

ously ﬂattened; some quartz grains appear trun-

cated by adjacent phyllosilicate ﬂakes and have

local pressure-shadow overgrowths (Figure 17.10).

Obliquely oriented relict bedding contacts are dis-

placed along mica ﬁlms.

2. Q-domains consist of quartz and lesser feldspar

plus possible carbonate minerals and minor, less

well aligned phyllosilicate grains. Grains are less

deformed and relict detrital grains are preserved in

low-grade rocks.

These two domains are commonly interpreted (e.g.,

Williams, 1990) to represent the effects of strain parti-

tioning and differential pressure solution. P-domains

Special Interest Box 17.3 Integrating

anisotropic lattice fabric into the evolution of a

tectonite terrane: a case history

An example of how lattice preferred orientation of

quartz can be related to other aspects of tectonite

fabric is provided by lineated metasedimentary

quartz–feldspar schists in the Raft River Mountains

metamorphic core complex, northwestern Utah

(Section 18.2.1 and Figure 18.11; see also Compton,

1980). The uniform east–west lineations consist of

streaked mineral aggregates and hinges of local re-

cumbent isoclinal folds. The subhorizontal folia-

tion deﬁnes a small amplitude, elongate east–west

structural dome—the core complex. Quartz and

minor feldspar grains in the schists are relict sand

grains of fairly uniform size that have been de-

formed into triaxial ellipsoids with shortest axes

perpendicular to foliation and longest axes parallel

to lineation. Local relict conglomeratic beds in the

metasedimentary sequence have remarkable triaxial

cobbles of similar shape and orientation (Figure

14.13). Ribbonlike lenticular quartz grains with un-

dulose optical extinction, elongate subgrains, and

basal-plane deformation lamellae are consistent

with translation slip. A few relict quartz sand grains

that are undeformed have c-axes oriented perpen-

dicular to the foliation, an orientation unfavorable

to either basal or prism slip under a compression

acting normal to the foliation. Deformed quartz

grains have volumes similar to those of less-

deformed feldspar grains, supporting a constant

volume process of plastic deformation acting on

initially uniform size sand grains. Although some

samples have lattice fabrics that are monoclinic or

triclinic c-axis orientation patterns, orthorhombic

symmetry patterns are more common (Figure 17.40).

Experimental and computer-simulated deformation

of quartzite by coaxial strain via dislocation glide

and climb has yielded similar orthorhombic

patterns (Tullis, 1977; Lister and Hobbs, 1980).

Compton (1980) concluded that the L–S tectonites

of the core complex were produced by east–west

extension and subvertical ﬂattening during essen-

tially coaxial strain, modiﬁed locally by simple

shear, resulting from upward doming of the heated

buoyant core of the metamorphic complex.

Evolution of Imposed Metamorphic Fabrics: Processes and Kinetics

559

17.40 Stereographic projections of lattice orientation of grains in tectonites subjected to coaxial pure shear. (a) Quartz c-axes in a sample of

quartz (89%)  feldspar (8%)  muscovite (3%) schist from the Raft River Mountains metamorphic core complex show an or-

thorhombic symmetry. Solid line is the trace of the foliation, square is the lineation, and dashed arc is the trace of the horizontal plane

with geographic north (N) and east (E). Numbers indicate average relative lengths of axes of triaxial ellipsoidal quartz grains, assuming

an original sphere of radius 1. (b) Normals to (001) in muscovite from same sample as in (a). (c) Quartz c-axes in quartzite deformed

experimentally in plane strain at 1.5 GPa, 800°C and d/dt  10

6

/s. Arrows indicate direction of compression and extension. Note

orthorhombic symmetry and resemblance to (a). Redrawn from Compton (1980) and Tullis (1977).

(a) Quartz c-axes

Triaxial

strain ellipsoid

0.33 : 1.0 : 2.7

(b) Muscovite

poles to (001)

Triaxial

strain ellipsoid

0.5 : 1.0 : 2.0

Extension

Compression

that are limbs of crenulations of a pre-existing foliation

have obviously experienced more strain. P-domains

are less soluble residues left after pressure solution

has preferentially extracted more soluble felsic and

carbonate minerals. The volume decrease permitted

the relatively inert phyllosilicate ﬂakes to rotate into a

preferred orientation. Once selective dissolution has

been initiated to produce phyllosilicate-rich folia,

subsequent prograde reactions will liberate more

water from them, facilitating additional corrosion of

remaining soluble grains and bodily rotation of inert

ﬂakes. This metamorphic differentiation can thus con-

tinue during prograde metamorphism.

In a classic study of slates in the Delaware Water

Gap area of New Jersey and Pennsylvania, Maxwell

(1962) envisaged that, prior to consolidation, a thick

sequence of muds, silts, and sands was dewatered as

it was subjected to tectonic forces, bodily rotating

phyllosilicate ﬂakes into a preferred planar orientation

parallel to the plane of ﬂattening and to axial planes of

folded beds. Grains of sand were entrained into the

expelled water that migrated along incipient cleavage

surfaces, forming cleavage-parallel sheets of sandstone

that penetrate into the slate from adjacent sandy beds.

Subsequent workers on these and other slates around

the world have concluded that mechanical rotation of

detrital grains can occur in the early stages of cleavage

development but that pressure solution and mimetic

recrystallization are dominant processes during later

progressive metamorphism.

After one and a half centuries, controversy continues

as to the amount of volume lost during evolution of slaty

cleavage and foliation in tectonites in general. Estim-

ates range upwards to 60% of the protolith volume,

as Sorby (1853) ﬁrst claimed. Estimates are based on,

for example, a comparison between the amount of per-

ceived ﬂattening perpendicular to the foliation and the

extension parallel to it. Debate persists regarding how

much of the lost volume was ﬂuids, principally water,

and how much was nonvolatile components, and the

degree to which the rock system remain closed, except

for the expelled water, and on what scale—domainal or

outcrop or larger. Local overgrowths in strain shadows

indicate grain-scale precipitation of transferred solute.

On the other hand, low-grade metasedimentary terranes

are characterized by widespread veins of predomin-

antly quartz that probably represent the sinks for the

substantial silica dissolved during foliation evolution,

as well as dehydration mineral reactions. Many geo-

chemical studies (see references in Vernon, 1998) have

attempted to cast light on the degree to which meta-

morphosing systems are closed by comparing ratios

of mobile to immobile elements, such as Al, Ti, and

Zr (Section 16.8.4). However, this approach suffers

from the apparent absence of truly immobile elements.

Rather than making complementary studies, geochemists

and structural petrologists have tended to work on dif-

ferent terranes. An exception is the work of Erslev (1998),

who concluded that nonvolatile losses in three of four

North American Paleozoic slate terranes are minimal.

Differentiated Compositional Layering. Layers of con-

trasting composition are evident at a range of scales in

metamorphic rocks of a wide range of grade.

Large-scale layers in recognizable, generally low-

grade metasedimentary and metavolcanic rocks are

commonly relict depositional beds and ﬂows, especially

where there are contrasts in composition and thickness

in a sequence. Smaller scale compositional layers, on

the order of centimeters, can be more difﬁcult to inter-

pret accurately. Bedding features, such as graded grain

size and cross-bedding, can be helpful, but other

purely metamorphic fabrics can locally resemble these

features, which are usually evident only in weakly

metamorphosed, low-grade rocks. Even if it can be

unequivocally demonstrated that a rock protolith was

sedimentary, there is no assurance that the layering is

relict bedding because other origins are possible. For

example, some layers and especially lenses of very small

aspect ratio (thickness divided by length) result from

transposition processes (Figures 14.19 and 17.33).

Other layers originate in other ways, as the following

paragraphs suggest.

Differentiated layering is a product of organization

or reorganization of matter during metamorphism of

a protolith lacking such layering (Williams, 1990). Its

true origin can be veriﬁed where the layering over-

prints an earlier foliation, which might or might not

be a relict bedding in the protolith, or where it can be

seen to have evolved from essentially isotropic rock by

metamorphic processes (e.g. Figure 14.17). As in the

growth of large porphyroblasts during metamorphism

(e.g. soccer-ball-size garnet, Figure 14.8), segregation

of chemical elements by diffusive transport is likely to

be involved, except creating a planar fabric element

rather than a subspherical one.

This layer-generating metamorphic differentiation

raises several questions. Can the coarser-grained and

thicker compositional layering typifying gneisses (see,

for example, Figure 19.24a) somehow evolve by further

progressive metamorphism of the differentiated P- and

Q-domain layering in ﬁner grained tectonites described

above? How does compositional layering develop in

orthogneisses derived from isotropic magmatic pro-

toliths? How much compositional layering in high-

grade migmatitic gneisses (Figures 15.12–15.14) is of

anatectic origin? Field observations and experimental

data on phase equilibria indicate that at least some

migmatites are a product of partial melting and that

concurrent deformation can play a signiﬁcant role

in segregating the partial melt from the crystalline

residue and concentrating it into planar domains

(Section 11.6.4 and, for example, Brown and Rushmer,

1997). But other examples of migmatization might be

formed by a strictly subsolidus process and, if so, how

does the differentiation into contrasting leucosome

and melanosome take place (Figure 15.15)? These

and other questions that could be posed highlight the

dilemma of how segregation occurs to create the com-

positional layering in gneisses.

560 Igneous and Metamorphic Petrology

One possible approach applicable to layering in or-

thogneisses is to “smear” or transpose coarse protolith

grains into thin, highly attenuated lenses (Williams

et al., 2000). But such smears typically consist of ﬁne-

grained lenticular aggregates segregated on the scale

of millimeters, rather than the coarser grained and

continuous layers several centimeters thick typical of

gneisses. Subsequent coarsening by elimination of ﬁner

grains and growth of larger ones might take place in

such segregations as well as in the differentiated layer-

ing of micaceous schists.

Robin (1979) proposed that silica ought to migrate

from P- to Q-domains because the more mica-rich ones

are weaker, allowing more strain to be partitioned

into them. Even subtle modal variations between

neighboring domains may become ampliﬁed in non-

hydrostatic stress ﬁelds by means of solution transfer.

A similar model for generation of compositional layer-

ing in alpine peridotites has been advocated by Dick

and Sinton (1979). Several workers have noticed that

where two grains of different minerals, such as a phyl-

losilicate and quartz, are in tight contact dissolution

occurs but dissolution does not occur at contacts of

grains of the same mineral. Intergranular pathways

are more extensive in polymineralic aggregates than

in monomineralic. Thus, mica–quartz rocks can

manifest greater solution activity than pure quartzite.

Dewers and Ortoleva (1990) explain this in terms of an

extra free energy induced in grain aggregates of differ-

ent minerals resulting from mismatched and distorted

lattices at their mutual boundary. They argue that tex-

tural and modal compositional factors interact with

mineral reaction/dissolution and diffusive transfer kine-

tics to create a feedback phenomenon that ampliﬁes

textural contrasts in a rock. Their theoretical models of

mechano-chemical coupling predict that small-scale

layering can be signiﬁcantly coarsened in realistic geo-

logic time scales. Further studies are needed.

SUMMARY

Metamorphic fabrics overprinted onto the relict pro-

tolith fabric evolve through a complex interplay of

thermally, chemically, and mechanically induced de-

structive and constructive subsolidus processes. Rela-

tive to melt-bearing magmatic systems, the evolution

of metamorphic fabrics is kinetically more sluggish

but more susceptible to grain-boundary textural equi-

libration. Fabric evolution is impacted by “catalyzing”

ﬂuids and more strongly inﬂuenced by broadly con-

current deformation during recrystallization in wide-

spread tectonites in orogenic belts.

Constructional, usually thermally activated, solid-state

growth processes adjust shapes and sizes of existing

grains of stable phases and create new, more stable

phases from decomposing or dissolving grains of

metastable phases. Considerable overstepping of equi-

librium conditions may be necessary to form new min-

eral grains by heterogeneous nucleation and to drive

diffusive transport of nutrient chemical components

from reactant phases along intergranular paths to

product grain surfaces. In the absence of deformation,

grain size is governed by that of the protolith, the

relative rates of nucleation and growth, and the kinetic

conditions controlling coarsening, or Ostwald ripen-

ing, whereby smaller grains are eliminated by grain-

boundary migration at the expense of more stable

larger grains. “Second” phases can restrict grain-

boundary migration and inhibit Ostwald ripening.

Minimization of the surface free energy in grain

aggregates drives equilibration of grain shapes as well

as size. Monomineralic aggregates of grains possessing

low surface-energy anisotropy, such as quartz and car-

bonate minerals, approach the ideal predicted 120°-

angle geometry of triple grain boundary junctions.

Departures from this ideal result from disequilibrium

or the need to create compromising low-energy bound-

aries between inherently inequant grains possessing

high surface-energy anisotropy, such as phyllosilicates

and amphiboles.

The kinetics of relatively rapid growth of porphy-

roblasts is governed by rates of intergranular diffusion

or incorporation of nutrient ions onto growing sur-

faces. Despite the higher surface energy in poikiloblasts

resulting from the incorporation of inclusions of resid-

ual matrix phases during growth, they can persist

indeﬁnitely. Inclusions and chemical zoning provide

opportunities for interpreting the history of mineral

reactions, diffusive transfer, and deformation during

growth of poikiloblasts.

In most metamorphic rocks, recrystallization is

intimately intertwined with broadly concurrent defor-

mation of the solid mass and leads to the rich variety

of anisotropic fabrics that characterize tectonites in

orogens. Ductile deformation resulting from non-

hydrostatic stress under metamorphic conditions

exerts a strong control on fabric by reducing grain

size and storing strain energy, mineral reaction rates

can be enhanced, rock can be locally heated, and

avenues for advective ﬂow of ﬂuids can be created.

Cohesive, penetrative, grain-scale ductile deforma-

tion under nonhydrostatic stress involves thermally

activated processes that take place at elevated P and

T and slow strain rates in metamorphic systems. In-

tracrystalline plasticity and diffusive transfer are the

two basic processes of ductile ﬂow that are involved in

changing the shape of individual mineral grains, strain-

ing the whole rock body, and creating foliation.

Intracrystalline plastic slip and twinning are well

documented processes of cohesive strain that occur on

particular crystallographic planes in a crystal. Slip is

Evolution of Imposed Metamorphic Fabrics: Processes and Kinetics

561

made possible by the presence of dislocations which

allow parts of the crystal to translate past one another

at much lower applied stress than for a theoretically

perfect crystal. In the presence of water, silicates can be

hydrolytically weakened and made more susceptible to

plastic slip. If thermally activated at high T, disloca-

tions can climb away from obstacles in their slip planes,

preventing work hardening of the strained grain and

allowing recovery to a state where more slip can occur.

Recovery can create more stable, less strained sub-

grains of slightly different lattice orientations bounded

by low-angle dislocation walls. Recovery can also

involve grain-boundary migration in the process of

dynamic (syntectonic) recrystallization, resulting in

knobby, or sutured, grain margins on strained grains

and ultimately creating aggregates of small, polygonal,

essentially strain-free grains that replace the less stable

strained grains. Simultaneous, continuous intracrys-

talline slip, recovery, and dynamic recrystallization can

accomplish large strains during steady-state ﬂow in

tectonites. Because this dislocation creep is a constant-

volume process, it is insensitive to P but depends on T

and strain rate. A power law equation incorporates

these environmental factors with applied stress and the

activation energy of ﬂow, permitting extrapolation of

experimental data to real geologic systems, and show-

ing that large ductile strains can occur at only small

stress differences over long periods of time. Deep

crustal and mantle rock deform by ductile ﬂow at slow

strain rates as if they were low viscosity liquids with

almost no shear strength. Slip and dislocation creep

can create a strong preferred crystallographic orienta-

tion of grains in deformed rocks.

Diffusive transfer and creep occurs by intergranular

diffusion of atoms; is especially effective in the pres-

ence of a grain-boundary ﬂuid, in which case it is called

pressure solution because more highly stressed parts of

grains are preferrentially dissolved. Grain boundary

sliding can accompany diffusive transfer but develop-

mant of a preferred crystallographic orientation does

not. Transfer is strongly T dependent because of the

diffusion rate dependence and generally occurs at

slower strain rates than intracrystalline plastic ﬂow.

In simple orogenic cycles, prograde metamorphic

reactions that accompany increasing depth of burial

release volatiles that facilitate deformation. After

peak metamorphism, rocks may be relatively strong

and resist further deformation unless there is sub-

sequent inﬂux of ﬂuid that will render them more

susceptible to ductile or brittle deformation. Although

deformation and recrystallization may be broadly con-

current, in most orogens each is typically episodic and

repetitive. Unraveling the timing of multiple overprints

of recrystallization and deformation in polymetamor-

phic tectonites requires careful attention to fabric fea-

tures such as overgrowths and cross-cutting relations.

Inclusion trails in poikiloblasts and their geometric

relation to the surrounding matrix foliation can pro-

vide important clues to the nature of deformation

whose effects have been erased in the matrix by later

deformation events or mineral modiﬁcations. Kinem-

atic indicators in tectonite fabrics record details of

the deformation path, which is a complex function of

the state(s) of stress, rates of strain, and rheological

properties of the rock, including the character of its

inhomogeneities.

Imposed metamorphic foliations reﬂect ductile ﬂow

and grain-boundary adjustments under nonhydrostatic

stress. Typically, a single foliation is parallel to the

plane of ﬂattening in the ﬁnite (bulk) strain ellipsoid.

Intracrystalline plastic twinning and especially slip are

effective processes of creating preferred shape orienta-

tions of mineral grains in tectonites. Plastic strain also

orients crystal lattices that can be preserved by recov-

ery processes. Diffusive transfer accompanied by grain

boundary sliding yields shape orientations but not

lattice, while bodily rotations of inherently inequant

grains in a more ductile matrix create both. Foliation

in ﬁne-grained metasedimentary tectonites probably

evolves by bodily rotation of grains, perhaps accom-

panying volume loss by dewatering, followed by later

pressure solution and mimetic recrystallization. These

tectonites possess phyllosilicate-rich and quartz-rich

domains that can evolve during further metamorphism

into more strongly differentiated layering by continued

deformation, strain partitioning, and differential pres-

sure solution. Coarser compositional layering typifying

gneisses can originate by anatectic migmatization and

possibly other little understood processes involving

mechano-chemical self-organization and ampliﬁcation

of subtle layering.

CRITICAL THINKING QUESTIONS

17.1 Larger quartz grains in the lower left of Fig-

ure 17.5 could have been created by poikilo-

blastically encompassing magnetite grains as

inclusions. Why do you suppose this did not

happen?

17.2 Describe the textural evidence seen in thin

section for ductile deformation of quartz.

17.3 With all sorts of modern laboratory equip-

ment at your disposal, indicate two ways that

you could distinguish between diffusive and

dislocation creep as mechanisms creating a

mylonite.

17.4 Figure 17.31b shows strained plagioclase

with bent twin lamellae and undulatory

extinction under cross-polarized light has been

partially replaced by small polygonal grains

lacking undulatory extinction. What are two

562 Igneous and Metamorphic Petrology

interpretations of this texture regarding timing

of deformation and recrystallization? How

would you distinguish between these?

17.5 Diffusion is enhanced in crystals experiencing

strain. Explain why this is so.

17.6 If it is assumed that the mineral inclusions

embedded in a poikiloblast originated during

growth and incorporation of matrix grains,

how would you account for their smaller size?

Indicate two possibilities and how you might

test which one is the more likely.

17.7 Discuss reasons why strain can be partitioned

into different domains in an inhomogeneously

deforming rock body.

17.8 Describe the mechanisms that can cause

change in grain shape. To what extent do these

mechanisms also change the orientation of the

crystal lattice?

17.9 Justify the commonly invoked assumption that

states of stress at depth in the Earth are, to a

ﬁrst approximation, hydrostatic.

17.10 Would one anticipate any correlation between

metamorphic grade and the mode of deforma-

tion producing preferred grain orientation in

tectonites? Justify your answer.

17.11 Discuss the role of water in the formation of

anisotropic fabric in metamorphic tectonites.

17.12 Contrast the conditions that favor one mech-

anism of ductile ﬂow over another.

17.13 Discuss fabric features that distinguish between

pre-, syn-, and postkinematic crystallization,

indicating any possible ambiguities.

Evolution of Imposed Metamorphic Fabrics: Processes and Kinetics

563

UNDAMENTAL

UESTIONS

ONSIDERED IN

HIS

HAPTER

1. How does the tectonic and thermal evolution of

rock masses at convergent plate boundaries

through time govern the nature of regional

metamorphism?

2. Conversely, how can zones, facies, and the P–T–t

paths in ancient metamorphic terranes provide

insights into the tectonothermal evolution of the

orogens (mountain belts) in which they were

created?

3. What mineral reactions and assemblages deﬁne

metamorphic zones and facies in regional

environments?

INTRODUCTION

Metamorphism is driven by transfer of heat and mass,

including ﬂuids, in ways that are related to the tectonic

setting in which the transfer occurs. This concept was

introduced in Section 14.2.3 but is further developed

as the central theme of this chapter. Tectonothermal

factors governing the development of metamorphic

zones and facies in active orogenic belts are recorded

in the time-dependent P–T path of the rock or terrane.

Data to construct this path can be determined by

thermobarometry of mineral assemblages and zoned

porphyroblasts, mineral inclusions, petrogenetic

grids, and chronologic analyses of appropriate time

markers.

Metamorphism on Earth is almost entirely focused

near convergent and divergent plate boundaries be-

cause these are the only tectonic regimes where ﬂows

of heat and mass are of sufﬁcient magnitude to cause

changes in the states of geologic systems. Within plate

interiors, such as stable cratons, steady-state geotherms

and virtual lack of movement of mass preclude much

metamorphism; only locally do small intrusions of

highly alkaline maﬁc magmas (Section 13.12) create

negligible contact metamorphism. Strike-slip and trans-

form fault regimes likewise generally lack signiﬁcant

vertical movement of mass in the gravitational ﬁeld

but focused deformation and shear heating can create

local metamorphism, such as mylonitization. Ocean-

ridge metamorphism takes place in a relatively narrow

zone at spreading ridges but its variable effects cover

the entire ﬂoor of the oceans because of seaﬂoor

spreading.

Regional metamorphism of a variable complex char-

acter develops in broad orogens in thick continental

crust at active convergent plate junctures over time pe-

riods of tens to perhaps hundreds of millions of years.

The geometry and rates of plate convergence, plate

character, and details of tectonic and thermal processes

in the asthenosphere and lithosphere all play a part

in metamorphic development and character. The typ-

ical tectonites of convergent orogens are commonly

Metamorphism at

Convergent Plate

Margins: P–T–t

Paths, Facies,

and Zones

CHAPTER

polymetamorphic, having experienced multiple episodes

of recrystallization and pervasive solid-state ductile

deformation associated with contractional folding and

thrusting on a regional scale. Heat sources driving

metamorphism in orogens can be magmatic intrusions

and underplates, as well as adjustment of geotherms

related to crustal thickening, strain heating, advecting

ﬂuids, and decay of radioactive U, Th, and K.

In this chapter, we ﬁrst consider P–T–t paths and

how they provide insights into the tectonothermal evo-

lution of convergent-plate orogens and regional meta-

morphic terranes (Figure 14.24). Mineral assemblages

in the classic intermediate- and low-P Barrovian and

Buchan zones of these terranes are then reviewed as a

standard of comparison with other types of P–T zones

and facies. Ocean-ridge metamorphism is brieﬂy sum-

marized as a prelude to discussion of typical high P/T

metamorphic facies series in near-trench accretionary-

wedge regimes where oceanic and terrigenous rocks

are complexly juxtaposed. Lastly, recent discoveries of

very rare but intriguing ultrahigh-P rocks formed in

continent–continent collision zones are considered.

18.1 P–T–t PATHS

Before we deal with metamorphic associations of con-

vergent plate junctions, which are the major topic of

this chapter, the concept of P–T–t paths is developed

further here, building on the introduction in Section

14.2.9. Such paths depict the tectonothermal history of

an individual rock sample or a metamorphic terrane.

Because of intimate relations between thermal and tec-

tonic phenomena during orogenesis, determination of

the P–T history of a metamorphic terrane in a orogen

can provide insight into its tectonic evolution. Integra-

tion of data on changing P–T conditions in a metamor-

phic terrane in a chronologic framework of time,

t, yields a P–T–t path that can constrain the nature

of dynamic tectonic phenomena. So called “reverse”

petrologic interpretations of metamorphic rocks can

test and reﬁne “forward” tectonothermal modeling of

orogeny and lithospheric plate interaction, and vice

versa. New developments in the late 1970s and 1980s

in geophysics, computer modeling, geochronology, and

metamorphic petrology converged into this new ﬁeld

of research on the tectonothermal evolution of orogens

and the thermobarometry of metamorphic associa-

tions. Useful discourses on P–T–t paths and their im-

plications include Haugerud and Zen (1991), Hodges

(1991), and Spear (1993).

18.1.1 Thermal Considerations

The crust of the Earth is enriched in incompatible

elements relative to the mantle. Further enrichment in

radioactive K, U, and Th develops in the upper 15–20 km

of the continental crust because these incompatible

elements are taken up in partial melts that rise from

the lower to upper crust. For reasonable rates of

radioactive heat production in the crust of average

thickness, augmented by typical heat input from the

underlying mantle, the T just above the Moho is

500–600°C. However, such a “normal” geotherm is

inconsistent with peak temperatures of 600–750°C at

depths of 20–30 km recorded in typical Barrovian ter-

ranes of orogens. To overcome this thermal deﬁciency,

the geotherm must be perturbed in some way. One way

to increase the T in the lower crust is by underplating

or intruding it by hot mantle-derived basaltic magma

(Section 11.1.1).

However, it was pioneering studies by Oxburgh and

Turcotte (1974) and England and Richardson (1977)

that ﬁrst drew attention to other signiﬁcant factors

that perturb geotherms in active orogens and create

unusually high temperatures in the deep crust. They

demonstrated through forward tectonothermal model-

ing that, because rocks are such slow conductors of

heat (Section 1.1.3), tectonic movement of rock mass

on “rapid” plate tectonic time scales of cm/y in an

orogen can signiﬁcantly perturb geothermal gradients.

Indeed, young mountain belts are far removed from

thermal equilibrium. First-order factors that pro-

foundly control P–T conditions and paths during the

history of a metamorphic terrane include the dynamic

interaction between tectonic movement of rock mass

of particular dimensions and rates, time available for

conductive re-equilibration, rate of erosion or tectonic

unrooﬁng, and distribution of heat sources. Second-

order, less signiﬁcant, factors that can locally perturb a

geotherm include strain heating, advecting ﬂuids and

magma, and mineral reactions.

Tectonothermal models of orogeny published in

the last two decades of the twentieth century depend

on the speciﬁc boundary conditions just listed, but

these are never fully known for even young mountain-

metamorphic belts. Hence, only a few generalities of

the models can be accepted with any degree of con-

ﬁdence. Elevated geotherms capable of creating real-

istic temperatures for metamorphism typically invoke

combinations of a thicker crust and enhanced radioac-

tive heat production below or within the crust. Thick-

ening of continental crust by thrusting, folding, and

magmatic intrusion and extrusion during orogenesis is,

of course, a fundamental and long-standing concept.

The model in Figure 18.1 illustrates the thermal re-

sponse of instantaneously doubling the crustal thick-

ness by thrusting. Several tens of millions of years are

required to subsequently conductively re-equilibrate

the sawtooth-like geotherm that is created at the in-

stant of thrusting. During this re-equilibration, or ther-

mal relaxation, temperatures at any particular depth

progressively increase and P–T conditions and paths

appropriate to Barrovian metamorphism can develop.

Metamorphism at Convergent Plate Margins: P–T–t Paths, Facies, and Zones

565

In the model in Figure 18.2 that is designed to mimic

Alpine orogeny and metamorphism during continent–

continent collision beginning in the late Cretaceous,

insertion of a slice of radioactive upper crust into the

overriding crust signiﬁcantly enhances heating sufﬁ-

cient to produce metamorphism.

Basically, tectonothermal models show that the char-

acter of prograde P–T paths during metamorphism is

largely governed by the amount of thermal relaxation

and radioactive heating. The retrograde path is strongly

inﬂuenced by unrooﬁng rates, either by tectonic exten-

sional faulting and crustal thinning or by normal ero-

sion. If unrooﬁng rates are rapid and occur early in the

orogenic cycle, crustal heating is inhibited and decom-

pression is nearly isothermal; the retrograde path in

P–T space is very steep. “Tectonic erosion” can be rapid

but normal erosion accomplished by running water has

also been shown to produce rates as much as 5 km/My

in the rapidly rising Himalaya (Zeitler et al., 2001).

18.1.2 Petrologic Determination of P T t Paths

In a rock perfectly equilibrated under particular peak

metamorphic conditions, a unique set of P–T values

can be obtained by mineral thermobarometry. How-

ever, additional information is needed to determine

additional points on the trajectory of the rock or the

metamorphic terrane through P–T space as conditions

of metamorphism progressed to the peak condition

and then retrogressively declined. Fortunately, for the

geologist, not all metamorphic rocks fully equilibrate at

only one unique, usually peak, P and T; vestiges of the

prograde, peak, and retrograde parts of the P–T history

can be preserved to varying degrees in the form of

zoned minerals, reaction textures, mineral inclusions in

poikiloblasts, and ﬂuid inclusions (Special Interest Box

18.2). Thus, a relatively complete P–T history can

be determined. To make the P–T–t path completely

quantitative, points on the P–T trajectory must be dated

in an absolute sense; a certain equilibrated mineral

assemblage or a particular growth zone added to a

grain can be chronologically analyzed using a variety of

isotopic systems. This two-step procedure of thermo-

barometry and chronology is brieﬂy explored below.

Thermobarometry and P T Paths. In a sense, facies

series in a metamorphic terrane deﬁne P–T paths (Fig-

ure 14.34). Thus, the zeolite–prehnite–pumpellyite–

blueschist series has a high P/T trajectory in contrast to

the intermediate P/T path of the greenschist-amphibo-

lite-granulite facies series. Calibration of relict mineral

assemblages and frozen reaction textures using a pet-

rogenetic grid can tighten the metamorphic path in

P–T space. Mineral thermobarometry (Section 16.11.3)

can provide distinct P–T values. Regardless of the

technique, caution has to be exercised because there is

opportunity for considerable error in determinations of

566 Igneous and Metamorphic Petrology

Depth (km)

P (kbar)

Depth (km)

1008006004002000

0°C

300°

550°

0°

300°

550°

300°

550°

1000°

Mantle

(a) Normal crust Thickened crust

0 My

120

(b)

(c)

10008006004002000

T (°C)

300°

550°

0°

18.1 Conductive thermal model of doubling the thickness of a seg-

ment of the continental crust in an orogen by instantaneous

thrusting. (a) Cross-section through thickened crust. In nor-

mal crust, upper 15 km-thick layer is most radioactive and has

T  300°C at bottom. Less radioactive lower crust layer is

20 km-thick and has T  550°C at bottom. Underlying peri-

dotitic mantle is considerably less radioactive. In thickened

crust section two normal crusts are stacked on top of one

another. Note depression of isotherm in underlying mantle.

(b) Depth-T proﬁle showing sawtooth-shaped geotherm

(dashed line) at instant of thrust stacking (0 My). Full lines at

5, 10, 20, 40, 80, and 120 My show progress of geotherms as

thermal relaxation (conductive re-equilibration) takes place in

the thickened crust. Filled circles on the geotherms show

depth-T–t paths (dotted lines) of small volumes of rock as

thermal relaxation and unrooﬁng of the thickened crust takes

place beginning 20 My after thrusting. Rock volumes were at

depths of 40 and 60 km a moment after thrusting. (c) Con-

ventional P–T diagram, with P (and depth) increasing upward

instead of downward as in (a) and (b). P–T–t paths from (b)

that result from crustal thickening and thermal relaxation

followed by unrooﬁng are clockwise. Redrawn from England

and Thompson (1984), who provide further details.