thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes...

20
American Mineralogist, Volume 67, pages 203-222, 1982 Thermodynamiccalibration of geobarometers based on the assemblages garnet-plagioclase-orthopyroxene (clinopyroxene)-quartz' R. C. NewroN AND D. PemrNs III2 Department of the Geophysical Sciences University of Chicago, Chicago, Illinois 60637 Abstract Two mineralogic geobarometers based on the assemblage garnet-plagioclase-pyroxene- qvartz, common in granulite-gradequartzofeldspathic and basic lithologies, have been calibrated almost entirely from measured thermodynamicquantities,especiallyenthalpy of solution and heat capacity measurements. For the continuous reaction: CaAl2Si2O6 + Mg2Si2O6: l/3Ca3Al2Si3O12 * 2/3Mg3Al2Si3Or2 + SiO2 (A) (plagioclase) (opx) the resulting geobarometric expressionis: (quartz) Peo* (bars) : 3944+ 13.0702(K) + 3.5038ZlnKa (l) (garnet) (ag^' azM)l wn€r€ r(4 : ,45(oRf") The pressureuncertainty is + 15fi) bars, mainly from the calorimetric uncertainty in AGi. For the reaction CaAl2Si2O6 + CaMgSi2O6:2l3Ca3Al2Si3O12 + 1/3Mg3Al2Si3Or2 + SiO2 (B) (plagioclase) (cpx) (garnet) the expression is: Pcp* : 675 + I7.l79T + 3.5962TlnKs (43" ' arraJc' whef€ Kn = _:--------=--. (48 (43fit") (quartz) The uncertainty is -t1600 bars. The activity expressionsare simple formulae based on calorimetry, except for those of the pyroxene components,which are basedon the "ideal two-site" model. A literature survey of analyzed occurrencesof the above assemblages shows that they can be classified into (l) thermal aureoles, (2) transitional granulite terranes, (3) massif granulite terranes, (4) tectonically uplifted deep crustal granulites and (5) deep crustal granulite exotics in explosive igneouspipes. Characteristicpressurerangesfor each class were found with the present geobarometers. Thermal aureoles in the Nain (Labrador) province and the Scottish Dalradians show l-4 kbar; massif granulite terranes show a notable cluster in the range 8.911.5 kbar, transitional granulite terranes fall at lower pressures, as well as lower temperatures, than the massif granulites, and the highest pressuresare registeredby the deep crustal uplifts, such as the Ronda (Spain) peridotite aureole and the Doubtful Sound, New Zealand, mafic granulites. Pressures of up to 12.5 kbar for the Doubtful Sound granulites confirms their deep-crustal character. The consistent8 kbar pressures of many granulite terranescould correspondto the base I This report was first presentedat a symposium on "Modeling of High-Temperature Petrologic Processes" held at Iowa State University in Ames, Iowa in connectionwith the North-Central G.S.A. meetings April 30-May I, l9El. 2 Presentaddress:Geology Department, University of North Dakota, Grand Forks, North Dakota 5E202. 0003-004)v82l0304-0203$02.00 203 (2\

Upload: others

Post on 17-May-2020

16 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

American Mineralogist, Volume 67, pages 203-222, 1982

Thermodynamic calibration of geobarometers based on the assemblagesgarnet-plagioclase-orthopyroxene (clinopyroxene)-quartz'

R. C. NewroN AND D. PemrNs III2

Department of the Geophysical SciencesUniversity of Chicago, Chicago, Illinois 60637

Abstract

Two mineralogic geobarometers based on the assemblage garnet-plagioclase-pyroxene-qvartz, common in granulite-grade quartzofeldspathic and basic lithologies, have beencalibrated almost entirely from measured thermodynamic quantities, especially enthalpy ofsolution and heat capacity measurements. For the continuous reaction:

CaAl2Si2O6 + Mg2Si2O6: l/3Ca3Al2Si3O12 * 2/3Mg3Al2Si3Or2 + SiO2 (A)

(plagioclase) (opx)

the resulting geobarometric expression is:

(quartz)

Peo* (bars) : 3944 + 13.0702(K) + 3.5038ZlnKa (l)

(garnet)

(ag^ ' azM)lwn€r€ r(4 :

,45(oRf")

The pressure uncertainty is + 15fi) bars, mainly from the calorimetric uncertainty in AGi.For the reaction

CaAl2Si2O6 + CaMgSi2O6:2l3Ca3Al2Si3O12 + 1/3Mg3Al2Si3Or2 + SiO2 (B)

(plagioclase) (cpx) (garnet)

the expression is:Pcp* : 675 + I7.l79T + 3.5962TlnKs

(43" ' arraJc'whef€ Kn = _:--------=--.

(48 (43fit")

(quartz)

The uncertainty is -t1600 bars. The activity expressions are simple formulae based oncalorimetry, except for those of the pyroxene components, which are based on the "idealtwo-site" model.

A literature survey of analyzed occurrences of the above assemblages shows that theycan be classified into (l) thermal aureoles, (2) transitional granulite terranes, (3) massifgranulite terranes, (4) tectonically uplifted deep crustal granulites and (5) deep crustalgranulite exotics in explosive igneous pipes. Characteristic pressure ranges for each classwere found with the present geobarometers. Thermal aureoles in the Nain (Labrador)province and the Scottish Dalradians show l-4 kbar; massif granulite terranes show anotable cluster in the range 8.911.5 kbar, transitional granulite terranes fall at lowerpressures, as well as lower temperatures, than the massif granulites, and the highestpressures are registered by the deep crustal uplifts, such as the Ronda (Spain) peridotiteaureole and the Doubtful Sound, New Zealand, mafic granulites. Pressures of up to 12.5kbar for the Doubtful Sound granulites confirms their deep-crustal character.

The consistent 8 kbar pressures of many granulite terranes could correspond to the base

I This report was first presented at a symposium on "Modeling of High-Temperature PetrologicProcesses" held at Iowa State University in Ames, Iowa in connection with the North-Central G.S.A.meetings April 30-May I, l9El.

2 Present address: Geology Department, University of North Dakota, Grand Forks, North Dakota5E202.

0003-004)v82l0304-0203$02.00 203

(2\

Page 2: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES

of a 30-40 km-thick continent, which provides evidence that continental-scale overthrust-ing, such as that postulated to be currently in progress under the Tibetan Plateau, could bethe principal tectonic mechanism of ancient granulite metamorphism.

IntroductionGeobarometry of garnetiferous crustal rocks

In recent years several mineralogic geobarom-eters for metamorphic rocks have been proposedwhich make use of the common coexistence ofgarnet with less dense aluminous minerals such ascordierite, plagioclase, or aluminous pyroxenes.Simple reactions relating garnet to the less denseminerals form the basis of geobarometric equations.The reactions are of high variance in natural compo-sitions, so that the corresponding thermodynamicequations are sliding-scale, or continuous, pressureindicators. A simple example is the reaction ofMg3Al2Si3Orz (pyrope) to Mg2Si2O6 (enstatite) andMgAl2SiO6 (MgTs), first calibrated from experimen-tal phase equilibrium studies as a geobarometer byWood and Banno (1973). The garnet+rthopyrox-ene barometer was originally intended for garnetperidotites of subcrustal origin, but has also beenapplied numerous times to high-grade crustal rocks(Weaver et al. 1978; Wells, 1979; Hormann et aI.,1980). This barometer has some severe limitationsin the crustal application, mainly arising from inade-quate experimental calibration for the crustalranges of temperature, pressure, and mineral com-positions. The association garnet-plagioclase-AlzSiOyquartz in high-grade pelitic rocks is the basisof another widely-used geobarometer. The calibra-tion, based on experimental phase equilibrium work(Goldsmith, 1980) and measured thermodynamicproperties of garnet and plagioclase (Newton andHaselton, 1981), is fairly precise, and the volumechange of the reaction is large, which is a prerequi-site of an accurate geobarometer. The main disad-vantage is comparative rarity of pelitic composi-tions yielding the four-phase assemblage in somehigh-grade terranes. A third widely-used geobarom-eter is based on the reaction of Mg-cordierite topyrope, sillimanite and quartz. The volume changeof reaction is very large, and the requisite four-phase assemblage has been reported from manyterranes. Experimental calibrations based on at-tempted direct equilibration of garnet and cordieriteare in some conflict (Hensen and Green, 1971;Currie, l97l), as are semiempirical theoretical cali-brations (Thompson, 1976; Perchuk et al., l98l).One problem stems from the poorly-known effect ofmolecular HzO in the cordierite structure (Newton

and Wood, 1979). Another problem is the limitedpressure range over which garnet and cordieritecoexist, which is probably only 3-4 kbar in manyrock compositions.

Quartzo-feldspathic lithologies are the dominantcomponents of many granulite-facies terranes. Py-roxene-bearing gneisses, or charnockites, common-ly contain garnet, including the type charnockitefound near Madras (Pichamuthu, 1970). Basic gran-ulites are ubiquitous and abundant in all gneissterranes as lenses, enclaves, and, sometimes, thickbodies of supracrustal or plutonic origin. Clinopyr-oxene or orthopyroxene or both are often accompa-nied by garnet and quartz. The mineralogy of char-nockites and basic granulites suggests thepossibility of geobarometers based on the reactions:

CaAl2Si2Os+ Mg2Si2O6 :ll3CatAlzSi3O12

anorthite enstatite grossular

in in inplagioclase orthopyroxene garnet

+ 2l3MgtAl2Si3O12 + SiO2 (A)

pyrope quartztn

garnet

CaAl2Si2O6* CaMgSi2O6 :2l3CatNzSi3O12

anorthite diopsidein in

plagioclase clinopyroxene

grossularin

garnet

+ 1/3Mg3Al2Si3O12 * SiOz G)

pyrope qtartztn

garnet

These reactions have large volume changes and arethus suitable in principal for accurate geobarom-etry. They are of high variance in crustal rocks,however. and therefore are less amenable to directexperimental calibrations.

Thermodynamic formulation of the barometers

The equilibrium condition for any reaction is thatthe Gibbs energy difference, AG, vanish at a giventemperature, T, and pressure, P:

Page 3: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES

-Arib asB agl .@ED, qLVB: u r -

R? R -"

agl*. aX" RT

aG(T, P, Xfl : 0 (1)

X; represents the mole fractions of the components.i, in the participating phases, B. Equation (l) maybe expanded for reactions (A) and (B):

-Alli +

A,Si _ ,,, (o$r,Jt . o8l

+ pLV^

()\RT R 48R. .4 i ' " RT

\ - '

Archaean charnockites and metagabbros. He evalu-ated AGi from phase equilibrium data of Kushiroand Yoder (1966), which required a long tempera-ture extrapolation of the data, obtained mostlyabove 1200"C. Moreover, compositions of thephases in their reactions were not accurately deter-mined. Activities of garnet components were takenfrom phase equilibrium reductions of Hensen et a/.(1975a) and plagioclase activities were based on thephase equilibrium dara of Orville (1972). Woodfound maximum pressures of 11-13 kbar at plausi-ble granulite temperatures of 700-900'C. Thesepressures seem rather high considering that theSouth Harris charnockites do not actually containgarnet. Wells (1979) deduced pressures of 9.1-11.4kbar (mean of 10.5 kbar) for basic granulites of theBuksefiorden region, SW Greenland, from the samegeobarometer and assuming a temperature of800'C. He used more recent phase equilibrium data(Hensen, 1976) to evaluate AGo, and nearly thesame activity assumptions used by Wood (1975).The Hensen (1976) equilibrium curves also requirelong extrapolations to crustal conditions, and thecompositions of phases along them were not well-determined. Perkins (1979) calibrated geothermom-eters based on reactions (A) and (B) for rocks of atransitional amphibolite-granulite terrain in the Ot-ter Lake, S. Quebec, region. He used experimentalphase equilibrium work on plagioclase, pyroxenes,garnet and quartz of Kushiro (1969) and Kushiroand Yoder (1966) to evaluate the AGp's, and sum_maized available measurements, both calorimetricand experimental, to define the activity-composi-tion relations of Ca-Mg-Fe garnets. Indicated pres-sures of the Otter Lake rocks were zero to negativekilobars. The inconsistencies were traced to thedubious nature of the AG's derived from phaseequilibrium data.

Newton (1978) recalibrated the reaction (A) ba-rometer using recent heat of solution data for AIfand entropy measurements for AS". Activity rela-tions were taken largely from deductions fromnatural parageneses of Ganguly and Kennedy(1974), together with experimental phase equilibri-um measurements of Ca-Mg garnet mixing of Hen-sen et al. (1975b), based on the anorthite-garnet-Al2SiO5-qu artz reaction. Recomputed pressures ofthe South Harris terrane were 3-4 kbar lower thanfound by Wood (1975) at the same assumed tem-peratures. The mixing properties of garnets used byNewton (1978) have been superceded by recentphase equilibrium and calorimetric measurements.

where AIf and ASo denote, respectively, the stan-dard enthalpy and entropy changes ofthe reactions,Af the partial molal volume differences, a theactivities at one bar, and R is the gas constant. Theapproximation sign is used because of neglect ofdifferences in compressibilities. The subscripts Mg,Ca, En, Di and An denote, respectively, the compo-nents MgAl273SiOa, CaAl273SiOa, Mg2Si2O6, CaMg-Si2O6, and CaAl2Si2Os, and the superscripts Gt,Opx, Cpx, and Pl denote, respectively, the phasesgarnet, orthopyroxene, clinopyroxene and plagio-clase.

Equations (2) define P in terms of ? and thecompositions of the phases provided that the activi-ties and partial molar volumes are known as func-tions of composition and Z and that the standardGibbs energy changes, which are the left-hand sidesof equations (2), are known as functions of ?. Theactivities may be determined by calorimetric mea-surements on solid solutions and by modelling ofphase equilibrium data. The partial molar volumesare determined by X-ray diffraction data for thesolid solutions. The standard Gibbs energies forreactions (A) and (B) may not be deduced directlyand accurately from phase equilibrium data becausethe reactions are not univariant and because rele-vant experimental phase equilibrium data (Kushiroand Yoder, 1966; Hensen, 1976) do not defineclosely enough the compositions of coexisting gar-nets and pyroxenes. However, enthalpy of solutionand heat capacity measurements on the end-mem-ber substances allow purely thermodynamic deter-mination of AGi and AGfi.

Previous calibrations

Wood (1975) first used reaction (A) in geobaro-metric calculations of the South Harris (Scotland)

Page 4: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES

Scope of the present work

In this paper we calibrate the two garnet-pyrox-ene-plagioclase'4uartz geobarometers based on re-actions (A) and (B) using recently determined ther-modynamic data and discuss the uncertainties inpressure indication. Various parageneses of thefour-phase assemblages in charnockites and basicgranulites are tabulated from the literature, andapparent pressures of crystallization are calculated.The survey, although not exhaustive, shows pat-terns which suggest new interpretations about vari-ous kinds of granulites, as discussed in subsequentsections.

Thermodynamics of the barometric reactions

Enthalpy of reaction

Table I lists enthalpy of formation data at970Kdetermined by oxide melt calorimetry for thephases anorthite (Newton et al, 1980), diopside andgrossular (Charlu et al., 1978) and enstatite andpyrope (Charlu et al., 1975). The data of Table 1lead to the following enthalpies of reaction at970K,taken as the differences in the enthalpies of forma-tion of the reacting phases:

AIli : 2237-1570 cal (3)

AHb : 373+650 cal

The error limits are taken as the square roots of thesums of the squares of the uncertainties in theenthalpies of formation in reactions (A) and (B) andare certainly overestimates, in that the uncertaintiesin the actual enthalpy of solution measurementshave entered twice (Anderson, 1977) in the errorassessment. High temperature heat capacity data

(Holland, 1981) are given in Table I to correct AIffrom 970 K to any temperature in the crustal range.The corrections for 900 K < T < 1150 K are verysmall and may be neglected.

Entropies of reaction

Table 1 lists entropy data at 1000 K for the sixsubstances of reactions (A) and (B). The mostimportant recent data are those of Haselton (1979)for pyrope and grossular which include both high-accuracy, low-temperature adiabatic heat capacitymeasurements and heat content measurements onthe same samples at high temperatures. Krupka etal. (1979a) measured the low-temperature heat ca-pacity of synthetic Mg2Si2O6 enstatite and heatcapacities to 1000 K by differential scanning cal-orimetry @SC). Haselton (1979) made drop cal-orimetry measurements on synthetic enstatite at973,ll73 and 1273 K and fitted a high-temperatureheat capacity equation to his and Krupka et al.'s(1979a) data. Robie et al. (1978) made precise low-temperature heat capacity measurements on syn-thetic anorthite, and Krupka et al. (1979b) addedDSC measurements to supplement earlier drop cal-orimetry. According to Haselton (1979) and Perkinset al. (1980), a configurational entropy ofabout 1.0caVK due to partial Al,Si disordering of anorthite isconsistent with phase equilibrium data of anorthiteand other minerals with well-measured parameters.X-ray diffraction data also require a small amount oftetrahedral disorder (Smith, 1974). For these rea-sons, a configurational entropy of 1.0 caVK is addedto the Third Law entropy of anorthite in Table l.The uncertainty of this term, probably no more than+0.5 caVK, is one of the largest sources of uncer-

Table l. Heats of formation, entropies, volumes, and heat capacities

anorthi teenstat i tedi ops i depyropegrossul arquartz

AHi, r oooGc-af6oTI

-24. 06r . 3l-17.62t ,34-34.991.41-20. 2 l r .38-77 .91t.67

0

sioooG-Vo:-moTT

127.7893 .5196.01

I 8s .83r87 . t 027.62

bxl 0-33.001-.0743.451

4 .9435. 695I . 451

cxI o6-.7257-.3387-.3349

- t . 9 9 t 2-2.2006

.0082

dxl 03- .6174-.6843-.6021-.5457-.4781- .2?6

-23.73 cm3-23.12 cm3

AHi^Hi

( looo) = 2.237 t .57 kcal(1000) = 0.373 t .65 kcal

asi =^si =

a

I 00.93 93.5562.66 82.5866.08 78.44

l l 3 . l 3 1 3 0 . 2 5125.23 130.2622.69 25.179

-7,417 ! 0 .64 cal / (-9.493 t 0 .54 cal /K

Heat CaDacitv Coeff icients\ a a , / a r - " " r , ,

avi =Avi =

Sources shten in teat.

Page 5: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES

tainty in AGP. The standard entropies of reaction at1000 K are:

ASi : -7.417-10.64 caUK

A,Si : -9.493-10.64 callK G)

These values may be corrected to any temperatureby the heat capacity data of Table 1; however, theymay be taken as virtual constants in the tempera-ture range of 850-1150 K.

Partial molar volumes

Haselton and Newton (1980) showed that thenonsymmetric partial molar volume function ofgrossular has a significant effect on its activitycoefficient in Ca-Mg-Fe garnets in the range 0.05 <Xcu - 0.20, and that the partial molar volumeversus X6, function could be represented by arelatively simple numerical expression which isalmost independent of Mg/Fe ratio. Unfortunately,this cannot be done for the partial molar volume ofpyrope, because the molar volumes of F+Mg gar-nets have not been carefully measured yet. There-fore, the AV terms in equations (2) have beenapproximated by AIu", the 298 K, one bar volumechange. These are -23.73 cm3 for reaction (A) and-23.12 cm3 for reaction (B). Trial replacement ofVq. for Vq. in a few calculated pressures shows thatthe substitution results in very small pressurechanges. Thermal expansion effects on the AV's arenegligible.

Activities of garnet components

The activity coefficient, 7, which is multiplied bythe mole fraction of a garnet component to give theactivity, can be expressed over a limited range ofcomposition by the formula:

u.r:Y!- xfRT

(5)

where X is the mole fraction of the component in abinary solution and W is a constant. Most solidsolutions whose properties have been measured donot obey this "symmetrical solution" formulationover the complete range of compositions. The factthat pyrope and grossular components of typicalgranulite garnets are usually restricted to the range0.05 < X < 0.40 allows the application of (5).

Enthalpy of solution measurements on severalsynthetic gaxnets of the pyrope-grossular series(Newton et al., 1977), together with heat capacitymeasurements on synthetic pyrope, grossular, and

Py.oGr.+ (Haselton and Westrum, 1980), define theone bar activities in the join by (5) and:

Wcilnre: 3300 - 1.5f(K), calories/mole (6)

which applies to all except very Ca-rich composi-tions (Newton and Haselton, 1981). The coefficientof the temperature term is the excess entropyparameter, lVs, the existence of which was predict-ed by the high temperature, high pressure phaseequilibrium data of Hensen et al. (1975b). Haseltonand Newton (1980) showed that expression (6),based on calorimetry, is in good agreement with theHensen et al. (1975b) data when the peculiar shapeof the grossular-pyrope molar volume curve, and itseffect on activities at high pressures, are taken intoaccount.

Cressey et al. (1978) derived the Gibbs energiesof mixing of synthetic Ca-Fe garnets from phaseequilibrium measurements based on the anorthitebreakdown reaction in the system FeO-CaO-Al2O3-SiO2. The mixing energy is temperature-dependent and depends on composition in a com-plex way. However, the excess Gibbs energy ofmixing is very small for Xcu - 0.30, at least attemperatures down to 900oC, and, within the accu-racy of their measurements, may be taken as zero.Deductions based on the compositions of coexistingnatural phases yield Wgur.. values of - 1000 to 1000caVmole for small grossular contents (Ganguly andKennedy, 1974; Ghent, 1976). The value of zero istaken here as the simplest assumption compatiblewith the experimental evidence, for the composi-tions of most of the garnets considered here.

The value of V/p"y, is still subject to debate.Earlier deductions from natural parageneses result-ed in a value of 2900 caVmole (Ganguly and Kenne-dy, 1974) but later workers using the same methodhave found much smaller values (Kawasaki andMatsui, 1977; Dahl, 1980). The latter referencegives the value of 1390-11160 caVmole. O'Neill andWood (1979) determined that Fe,Mg mixing ingarnet is more ideal at 1000'C and above by severalhundred caVmole than is olivine at the same condi-tions from their experimental Fe,Mg distributionisotherms at high pressures. Wood and Kleppa(1981) found, by enthalpy of solution measurementson synthetic olivines, a non-ideality which wouldcorrespond to a W of about 1200 callmole in the Fe-rich range. This implies aWpey.eof garnet of about600 callmole. Sack (1980) concluded, after a consid-erable review of available experimental work, thatW for olivine is probably less than about 800 cal/

Page 6: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES

mole, which suggests that Wp"y* for garnet isnegligible. A value of zero is taken in this work,which is probably a lower limit of values compatiblewith currently-available evidence.

There is little experimental work available todetermine the mixing properties of spessartine withother garnet components. Hsii (1980) experimental-ly found a stable join spessartine-grossular down toat least 400"C at 2 kbar, and spessartine-grossulargarnets are characteristic of low-temperature blue-schist terranes (e.g. Hashimoto, 1968), which sug-gests a very small Ca-Mn interaction energy. Themixing of Mn with Fe2+ and, especially, with Mgmay be substantially non-ideal, however. This un-certainty rules out use of Mn-rich garnets with thegeobarometers considered here. This is unfortu-nate, since some high grade terranes have abundantmanganiferous quartzites with barometrically diag-nostic assemblages (e.9. Janardhan and Srikan-tappa, 1975). In the present study we exclude anyassemblages from consideration where the garnetshave Mn > Mgl3, which insures that Ca interactionswith Mg and Fe far outnumber interactions withMn.

The activity coefficient of grossular in ternarygarnets may be approximated by the followingexpression, due to Prigogine and DeFay (1954) andused by Ganguly and Kennedy (1974):

RTlnygu = Wg^y",*y. 4 W cuufiue+ (lVc'r'.-Wr.rvrg*Vll6"1ar)Xr$ue Q)

A corresponding expression exists for the pyropecomponent. The activity coefficients of grossularand pyrope are, with the parameters adoptedabove:

RTlnygu : (3300-1.51)1Xlo*+X-;Xr.'1 (g)

RTlnla, : (3300- l.5Tl(Pc^+XcJy.)

Plagioclas e activitie s

Enthalpy of solution measurements at 1000 K onmany synthetic high-structural state plagioclases,together with the theoretical entropy of mixinggiven by the "Al-avoidance?' model of Kerrick andDarken (1975),lead to the following expressions forthe activity coefficient of anorthite (Newton et a/.,1980):

RZlnTa, : *oalw e-]- 2X 6n(W 66- lyaJlWl.: 2025 calW.a,a: 6746 cal

The activity of anorthite is then determined by:

xA,(l+xAJ2(Al-avoidance model)aen : ^ f t n

(10)

It is not yet established that plagioclase crystallizesin granulite metamorphism in the high structuralstate. Ordering of Si and Al could have some effectin changing the anorthite activity from expressions(9) and (10). It is likely, however, that Al'avoidance(i.e., avoidance of adjacent Al-filled tetrahedra inthe framework structure) gives an adequate aqcountofthe solid solution energetics at least above 700"Cand for X^n = 0.25.

Enstatite and diopside activities

Ferrous iron is the main substituent in enstatiteand diopside. There are currently no thermochemi-cal data for these solid solutions; however, a num-ber ofauthors (e.9. Saxena,1973; Sack, 1980) haveconcluded that FezSizO6 and Mg2Si20.6are virtuallyideal orthopyroxene substituents'at temperaturesabove 700oC, and Bishop (1980) and Oka and Ma-tsumoto (1974) concluded that CaMgSi2O6 and CaFe-Si2O6 solid solution is attended by a W of, at most,2.5 kcal. Wood and Banno (1973) and Wood (1979.)have shown, by phase equilibrium deductions, thatthe aluminous pyroxene molecules MgAl2SiO6 andCaAlzSiOo, in enstatite and diopside respectively,are virtually ideal substituents, and Ganguly Q973)found, also from phase equilibrium work, that NaAlSizOo 6adeite) forms a nearly ideal solid solutionwith diopside. These considerations lead to, andjustify, the "ideal two-site" pyroxene mixing model(Wood and Banno, 1973; Wells, 1977) expressed as:

an : xM? 'xMl in orthoPYroxene(1 1)

aoi: XW'xS! in clinoPYroxene

where M2 denotes the larger structural site contain-ing divalent and monovalent cations and Ml thesmaller site. These formulations are only approxi-mations; intercationic substitutions are generallyattended by some non-ideality (excess Gibbs energyof solid solution). However, it is fortunate that themole fractions of enstatite and diopside in ortho-pyroxene and clinopyroxene, respectively, are usu-ally high, so that effects of non-ideal mixing areminimized. The largest departure from the systemCaGMgO-Al2O3-SiO2 in granulite pyroxenes is

(e)

Page 7: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES

P(bars):3944+ 13.0707

+ 3'5038?t- (48)(49'e)'z

lm oil"' a$ff*

for opx-garnet-plagioclase-quartzand

P : 675 + 17.r7gT + 3.5962Ttn@Ej'z(@--- aR''' a$i.

Fe2Si2O6 substitution in orthopyroxene, which hasbeen shown to be nearly ideal. Thus, the ideal two-site model should give good results in many applica-tions. The cation-site assignments for clinopyrox-ene followed here are: Ca, Na, Mn and Fe2+ in M2and Mg, Fe3+, Ti, Al and the remaining Fe2* in Ml.Mg and Fe2* are equipartitioned between Ml andM2 in orthopyroxene (Wood and Banno, 1973).

Geobarometric expressions

The numerical geobarometric expressions basedon equations 2) and with the foregoing input dataa n d T i n K a r e :

Selection of analytic data

Criteria for selection

The requirements for a valid pressure calculationare textural evidence of four-phase equilibrium,reliable analyses of the minerals, and an indepen-dent temperature estimate. The first criterion mayconsist of contact or close local association ofunzoned minerals, or compositional "plateaus" incores of coarse mineral grains. Many granulitesshow evidence of polymetamorphism in the form ofcorona assemblages around garnet rims, containingpyroxenes and plagioclases of sharply differentcompositions from those of the broad plateaus, andthe garnets may be strongly zoned with rim compo-sitions approximating local equilibrium with thecorona assemblages, or, there may be a discretesecond generation of garnet. In such cases twodistinct pressure indications, one for the primary orcore compositions and one for the corona and rimcompositions may be obtained, which may afford aninterpretation about uplift from deepseated condi-tions or later regional retrogression. The commonpattern is more calcic plagioclase, less aluminouspyroxene (frequently orthopyroxene rather thanclinopyroxene) and less pyropic garnet rims in thecorona assemblage than in the primary assemblage.Wet chemical analyses of garnets may fail to distin-guish garnet and plagioclase cores and rims thatmicroprobe analyses can resolve. It is very impor-tant in the temperature estimates to distinguishprimary from secondary assemblages. For instance,retrograde biotite replacing garnet may not yield avalid temperature estimate with garnet cores. It isfrequently hard to decide if biotite and hornblendeare primary or retrogressive.

Compilation of data

The literature of petrology abounds in descrip-tions of the assemblage pyroxene-garnet-plagio-clase-quartz. A disappointingly small fraction ofthe reports satisfy the above criteria, however.Electron microprobe analyses have become abun-dant since about 1965, but some authors havepresented only representative analyses of mineralsfrom different rocks. Frequently, plagioclase analy-ses are not presented, and plagioclase is character-ized only as "oligoclase," for instance, or a broadrange of An contents may be given. In the olderliterature, some minerals from a rock mav have

(12)

for cpx-garnet-plagioclase-quartz.

The activities at a temperature are evaluated fromcompositions of the phases as outlined in the pre-ceeding section. Although the choice of tempera-ture depends on mineralogic geothermometry, withits inherent uncertainties, it turns out that thebarometric indicators are quite insensitive to tem-perature uncertainties: a temperature error of+100"C introduces pressure errors of only a few toseveral hundred bars. The major pressure uncer-tainty lies in AG' of the reactions. The combineduncertainties in thermochemistry lead to uncertain-ties of +1500 bars for reaction (A) and -r1700 barsfor reaction (B) in evaluating the errors in AG" asthe square roots of the sums of the squares of theindividual uncertainties in the A/f and ?A^S'termsat 1000 K. Uncertainties in the volume terms intro-duce much smaller errors. The combined uncertain-ties resulting from errors in the activity expressionsare difficult to assess. However, calculation ofpressures from many different terranes, to be pre-sented, affords considerable insight into the accura-cy problem, particularly where comparison of thetwo barometers in the same terrane or, in manycases, the same rock, is possible.

Page 8: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

2IO NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES

been analyzed by wet chemistry, while composi-tions for the others may only have been estimatedfrom optical properties. In some descriptions, it isnot clear whether qtartz is part of the equilibriumassemblage. Table 2 lists the suitable describedparageneses which were found after a considerablesearch. Some well-described parageneses have un-doubtedly been overlooked. The occurrences fallnaturally into a few categories: low-pressure ther-mal aureoles (pyroxene hornfels facies), transitionalgranulite terranes, regional granulite terranes, gran-ulite exposures of very deep-seated origin in mod-ern uplift areas, and deep-crustal granulite exoticsfrom explosive volcanic pipes. The temperatureestimates for each locality are generally those pre-ferred by the authors of the mineral analyses,except where noted in the following discussions.Temperatures are based on quantitative mineralogicgeothermometry.

Results of calculations

Table 2 gives the calculated pressures of theassemblages using the given compositional data andthe temperature estimates. It is seen that the calcu-lated pressures span almost the entire crustal rangeof pressure, from about 2 kbar for andalusite-sillimanite bearing migmatite terranes of the Scot-tish Dalradians, to about 12.5 kbar for the deepestcrustal samples, the mafic granulites from DoubtfulSound, New Zealand. The clinopyroxene barome-ter, based on reaction (B), consistently yields pres-sures which average two kbar lower than thosefrom the orthopyroxene, or reaction (A), barome-ter. Although this discrepancy is within the com-bined uncertainties in the AGP values of reactions(A) and (B), other possible interpretations will bediscussed. The average spread of the calculationsfor a given terrane is about 1850 bars for eachbarometer. Details of the results for a large numberofindividual terranes, given in subsequent sections,afford insight into the problems of the precision,accuracy, and validity ofthe pressure calculations.

Thermal aureoles

Results for the granulite aureole of the Nain,Labrador, anorthosite complex are given in Table 2.Calculations for two orthopyroxene-bearing granu-lites are complicated somewhat by zoned garnetsand variable plagioclase compositions, probablyresulting from reequilibration during uplift of theanorthosite mass. The garnet cores and most sodicplagioclases in each rock yield an average of 3700

bars, which is considerably lower than Berg's(1977) assessment based on the garnet--cordierite-hypersthene-quartz barometer (Hensen and Green,1973\ and earlier calibrations of the orthoferrosilite-fayalite-quartz barometer (Smith, 1972), but is ingood agreement with the value of 3.2 kbar calculat-ed for the Nain aureole by Bohlen and Boettcher(1981) based on their accurate experimental workon the compositions of coexisting orthopyroxeneand olivine in the presence of quartz. A combina-tion of the most calcic plagioclase and garnet rim inone rock (NAR-2893 R) yields only 1100 bars,which may represent a late, very shallow reequili-bration.

The garnet<rthopyroxene assemblage is devel-oped in the Huntley-Portsoy area of the Buchanzone of the Dalradians, and inthe Lochnagar areaof the Central Dalradians, Scottish Highlands (Ash-worth and Chinner, 1978), as aluminous-basic res-tites of anatectic migmatites developed in the aur-eoles of the Newer Intrusives. Both occurrencesare in metamorphic terranes containing andalusiteas the regional Al2SiO5 polymorph, and both havesillimanite in the thermal aureole. If, as seemsprobable, the aureoles reflect nearly isobaric heat-ing, the pressures should be at least as low as theAl2SiO5 triple-point. The calculated average pres-sures of 3880 bars for the Huntley-Portsoy area and2700 bars for the Lochnagar aureole satisfy thiscondition for the Holdaway (1971) triple-point with-in the experimental uncertainty. Pressures calculat-ed for the aureole localities using the Thompson(197 6) garnet-cordierite-sillimanite-quartz barome-ter are about two kbar higher.

Precambrian granulite mas sifs

Published mineral analyses suitable for geobaro-metric calculations using reactions (A) and (B) werefound for twenty Precambrian localities in the pre-sent survey. Of these, ten were from large granulitetracts. most of which show the definitive features ofuniformly high grade mineralogy and pronounceddepletion of large-ion lithophile elements (Newtonand Hansen, 1981).

Six suitable metabasites and charnockites fromthe Adirondack Highlands, northern New York,contain both pyroxenes in addition to plagioclase,garnet and quartz. Garnet and clinopyroxene analy-ses are from Bohlen and Essene (1980), orthopyrox-ene analyses from Bohlen and Essene (1979), andplagioclases and confirmation of quartz were pro-vided by S. R. Bohlen (written communication).

Page 9: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANT]LITES 2II

4

6o

ts

N OF Oo oF F

O @ FF r O F

c l r 6 6@ t s @

@os

ooioN

6 0o oO F

o cO N6 e@ @

ts

F

ooNo

N @ €o F < io O Ft s F O

o t s oo t s oN N O6 F C O

o@CO(o

@€Fo

@

o

co@@@

GOGOq

!zx r3 Ez = . / ,> q d\ o <f u E= E- c

<t@@o

@ oo c ot s o

@ NF c )€ FO N

oGO6

@ < it s 66 C O

€ o oO N Ft s o oo o o

(ocoF C O

F 6

o 6o o€ No @

tsGOo6

@@<t@

goF@

o o6 N

O N

<t <t

o €o o

F O Oo o €N + OO O t r

6€@e

6NGO

@GoF

@oF6

tsts@@

O UU E

< t sE < cH E O

F Un 4q =

UF

o6F

o oo 6@ F

o oo o@ F

oFts

e o6 Nt s N

o o oo 6 0F t s t s

oNF

o@

o o€ c oN @

o oN b@ @

o6o

o@

oGO

otsts

o oo oc o @

o o oo o oo @ c o

o o oo o o@ o o

oNGO

6tsco

otsCO

ooF

ooF

Gr6==e

ooGOta ro mo 6o 6O N

t lY

6 Oo c oO N

@

oJ

I

F -w =

I

o co N o< F ZA Y '

No

I

6I=

F

F <r €

P =

(.t

N Tt o

o lo

G

6NN

ooN

o

N

O N

I N6 6= oo l

d T -@ 9 -N N O

l l

aj ":-O N t s

c o N q tt t l

r U C )= -

I

F

€C'l

6coo@o

o@N

o

o

tsNo

F

a

LoUaE=o

FFo

oLo

coFqi

E FC vt L

s o9 gt cO +t s

t o

ogo

u o 6@ t s

E 0 6C - P

sc,

oo

€No

O Fc

s !E P GO + dr E 6q q F

D . .> 5. . =

u D o, E

a

o @o

g FG v

cEL6g6

D

cFo

G

PoEox=

G

@o

coc

oe

Ecoor

coC

oe

o@o

Eoooo

oc ^O r

o @gr! rt vo

0!E@

3N

ococti

! vE6 E

o> t AO C

F O

o oE q

o

G E@

€ 6O F

g

I

L(.t

=oFd

4o6ga

Io

oooL O> PG +

o oP S

F l= oc co 6I

o oo >o oo 5L I= p6 C

6 Lo o: 6P =

Er-

EEo

o Oo !f 5? aJ Go oo 6t #L OG E

EI

! ^c h6 v

^ oc oV G

o o+ E+ O= 6c o6L GO E

oO L, o9 0@

o

xooEIo<o

q,

c'oGPo-

og5o6oGpq,=

of

5Eosoo

rF

E

oo o

o! 9O +E -L !O E E

P E dE 6 t

o

Eg 6

qrp

=c6l-oo

ao

o

Lo

O FP OF ! tF €

5 9c od EL(,

c,P

G!€pE

F

dooJ

xo

F Le oE 1 t9 GU L

oE G

6

€E

o €o o+ , 9b = -

4 E -r @ !

>TF G9 ? o

€ 9 Ec 6 E: o- o

J

Eo

o-

J! Lo oG >E-o ou =E

xo

goqt

r oE FC E3 C

c-G .p

- aE

r E! H

o .

=

- 6E

! Eo ot oG

J

xooE=( J t

GO F

l ot 9o q9

oEo- o

F J

- LL

d ooao

6

xo+ oO E

O GN

€ E= 6S Ftr

xo

F O

I , co sa o

X Uc .6 =

a

x{,€F C

+ oo c0 € ,

0,q Lo c 5

L .o -

co+ EG E

F 6

O FO E6 Oo o

< ( Lo

6

F -

J

ssv'I3aloirJnP IPuuaqJ

aJnssaJd-xn1saueJraf ell [nuPJg JlssEN

()

lsbo

' r

$l

I

o6

O

(')6lq)

F

Page 10: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

oLt

oL

NN

N G@ t sN 6o o

N N@ o

o 6

+ c o o@ F b@ o o! i 6 q t

@Fsls

@Fho

€@

6

b F( D 66 0@ 6

N

o@

ooo

@ @O Ft s oF N

t s FN Oo oC O e

N@<i@

oo

sooo

oFo

o o< t €oc tN O

o

.iF

loo@

@@@ts

oL

oL

c oco <to o@ @

@ @€ 6

<toto@

ooo@

o @ Fo N 6N € 6e N @

hF€ts

6oots

c@NF

NqFts

@IoF

coFtsts

coForF

F

@ts

o<tFo

tssN

oF@

€<tct

oI

oIo

U€

o ot s F

o oo ots ts

a ot s F@ r o

6 0 0t s t s t s@ ( o @

o 6 6 6F t s F F@ r o @ @

ooco

ooc0

oq

oots

oorF

o o6 0tsts

o o& ots ts

6sF

6*ts

oqF

ooF

o ob oF F

o oo 6t s t s

oo@

eoco

oots

oots

o o .6 6F F

oco€

oo@

o6@

IG'

3

E U@ oc l bN N

t l

@ N< oN N

I tI

I+F

I J< c !

c t 6 |c t o < (

l l F lY A l FE A < =

NosF

F

Ih

IN

NF

I

€it

IG=

O pt l

N Oo o

t lN N

oN

I

Fo

Iv=

NN

Io=E

@@rt(oo

6 . Oq l €@ @

o oo @a €io ro

o

@oN

I

A

@o

Ie

ot

I

t l

= -N

IFJ

6N

oNN

€€6E

ooIL3oa

@tsor

EIo

EgoP.A

otso

E

!Lo

d

oF6

o

6

aro@

c oB F6 v

E30

6

9 G Oo t s

o

6o=

€PoC Fo c oo 6= F

oe

o@o

sI

e@o

ss

tso

Lc,

o

6ho

Eo

a

t 6o !

c .G !- =

o. gc t =

U

o@6

E(to!-

ap ^o o

FE O '

I

L<9

orts

! oC F6 v

O E! ?o l Fo sg +O LL ( 9=u

o+oLI

oo

!c 66 0

O F+ o+ 5o te o .E 6o

I

6 0 ,6 €

E Fc =G C

o! l+ oo

o

coIo

I

G@

c,p

tgGLo

o

o6

0,P

IocLG

0

of

5c

IoI

G6

a,P

os oO G

L c ,I E

oL

oP

|UE Og 6

t gor

o

5F EG €

+ Lo otL ( t

o cO Eoo

. !6 Eo 6P

F r ^ Or v o ( tt v G6 > r! L d 4O 6 E E

E 0 c t+ L O o9 A O 6g

oE +o o> oo sP E+ EP 6tC E6 ?!

(5

6 06 0G O

6

c o Go o E

O+o o o- c LL O Oo o=

6 ( ,

5 FL O tO ' F E

9 6o G !o E ooo

q,+t

=E Eg 66 !

o

6 0 ^€ N Oo N 65 v NL Vo !

o o ?o @ 9O Gc r E

G

oJ

9 EG g

! 6E -o =L O

E

6oL

9o o

J OG q ,

qLa .Io

dO ES E

< 6g co o

E OI Lo(. '

o =

z o5

6

GoL< t t+ 6

L C: 6 tE T

L .6 =c

E6I6

@

o oC r-E6

E @

E

O G

= ! tO E! Eo

q

-

- o0 , EO G +c c =o @E # >

E F5 O -

6 = O= -&

O Po 6 ?= C e6 6 G

E 9 9C O

-agt= o

d

EE E= go 6

6F O= N

IP B! o= =oo

xo

d cE +O E9 e

o€ .E qo

E

6

. EJ !6 0o s

e F

o oE Ooo

GoL

E @c -= Lo oc =op =

LlZ

oGIc , .F O

l #

; - or ! E

o o 95 € BoJ

oI

o o xE iEs o

F E =+ @

o E =sF

ssP lc saueJJal al! lnueJg lEuoIllsuPJI salllnuPJg lqsnJS deaoseold .

ut 5a(lucJ loAI I nucJo

NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES2t2

a*G

s\Jsts

Eo$G,

so

\)

!)

s\Js:(tFtJ

sB

oI

e

P

sst*F6.s

t)

Id

I

o

$&

s:$RXs

s

o

$

o

H$

d$

a$

qsr\):

\ID

oooL

oE

o

oEoxoL

oEPLo

EopGE

U

{.)

o

6l

o

F

Page 11: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES 213

Temperatures for pressure calculations were takenfrom the regional isotherm map of Bollen et al.(1980). Calculated pressures with the orthopyrox-ene barometer average 9.2 kbar, and, with theclinopyroxene barometer, 7.0 kbar. Some of thescatter in the calculated pressures may be due toreal variations over the Adirondack terrane. Theorthopyroxene mean pressure is somewhat higherthan the assessment of 8.0 {- 1.0 kbar for theAdirondack Highlands of Bohlen er a/. (1980) basedmainly on a few occurences of orthoferrosilite-fayalite-quartz associated with the Adirondack an-orthosite.

The average discrepancy between the orthopy-roxene and clinopyroxene pressure indications inthe same rock for the 6 Adirondack Highlands rocksis 2.3 kbar, close to the average discrepancy of 2.2kbar for all rocks in Table 2 with both pyroxenes.This figure may be used in an empirical readjust-ment of the geobarometers, to be discussed below.

The South Indian Highlands granulite massifareas yield pressures of 7-10 kbar. These localitiesrepresent the culmination of progressive metamor-phism across 200 km in southern Karnataka (Picha-muthu, 1965). Of particular interest are the Sittam-pundi garnet granulites (Subramaniam, 1956),which show a primary association (92P) with clino-pyroxene and a secondary, or corona assemblage(92S) with a distinctly different plagioclase andclinopyroxene. The primary assemblage crystal-lized at 8.0 kbar, according to the barometer ofreaction (B) and the retrogressive assemblage at3.4kbar. The secondary assemblage represents a dis-tinct and probably much younger regional metamor-phism in the contact zone of the Tiruchengodogranite (Naidu, 1963).

The Archaean high-grade terranes of the LabworHills, Uganda (Nixon et al. 1974) and, EnderbyLand, Antarctica (Grew, 1980) are higher tempera-ture terranes than many granulite areas, as evi-denced by regional development of sapphirine-qtartz, and in the latter area, osumilite. The veryhigh temperatures account for the presence of cor-dierite even at the elevated pressures. Garnet meta-basites from Enderby Land were not described byGrew (1980). Newton and Haselton (1981) calculat=ed a pressure of 7.0-t-0.3 kbar for Grew's rocksbased on the garnet-plagioclase-sillimanite-quartzassemblage. .

The Furua, Tanzania, Complex (Coolen, 1980)and the Qianxi Complex, northeastern China (R.Zhang and B. Cong, unpublished data) are high-

500 600 700 800 900 1000

Tempero lu re (oC )Fig. 1. Calculated pressures for various granulites. Triangles:

thermal aureole granulites. Circles: Precambrian shieldgranulites. Inverted triangles: uplifted deep-crustal sections.Squares: exotic granulites in explosive pipes. Open symbols:orthopyroxene (reaction A) barometer. Filled symbols;clinopyroxene (reaction B) barometer. Al2SiO5 (kyanite,sillimanite, andalusite) stability relations from Holdaway (1971).Localities: LN : Lochnagar, HP = Huntley-Portsoy, NA :

Nain, RS : Ruby Range (S), AL = Adirondack Lowlands,OL : Otter Lake, SL = Sierra Leone, BU = Buksefiorden, AH: Adirondack Highlands, SI : Sittampundi, MA : Madras, AK: Akilia, IN : Inarijervi, NH : Nilgiri Hills, SC : Scourie, IC: Isortog, TH : The Thumb, LT : Letseng-le-Terae, CP :

Cone Peak, FC : Furua, QC : Qianxi, DS = Doubtful Sound,RP : Ronda, KN : Kristiansund.

grade massifs where kyanite is the regional Al2SiO5polymorph, rather than the more usual sillimanite.The calculated pressures, at the estimated tempera-tures of -800'C in both regions, are compatiblewith the experimental Al2SiO5 diagram (Holdaway,l97l), as shown in Figure 1. The Qianxi Complex,dated at 3675 m.y., is the oldest very high gradearea known, and the 10+ kbar pressure shows thatthe early Archaean crust was, locally.at least, verythick. The early Archaean Akilia metabasites (Grif-fin et al., 1980) yield the same information (Table2).

The Scourie Complex, northwestern Scotland, isone of the most intensely-studied Archaean ter-ranes. Only recently, however, have mineral analy-ses been forthcoming sufficient to characterize thetemperatures and pressures of metamorphism (Rol-

s

<l)

=aoo

Page 12: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

2t4 NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES

linson, 1980; 1981a). The pressures of7-8.5 kbar forRollinson's (1981a) basic granulites yielded by thepresent pressure method are somewhat less thanthe 10.5 kbar calculated by him by the aluminousenstatite-garnet method (Wood, 1974), and muchless than the 1513 kbar estimated by O'Hara andYarwood (1978) based on experimental phase rela-tions of garnetiferous gabbros. The latter estimatecontradicts the presence of sillimanite and absenceof kyanite throughout the Scourie terrane.

Transitional granulite terranes

Pressures calculated for those regions, all Pre-cambrian, where amphibolite facies gneiss terranesgive way to granulite facies, are given in Table 2.The Madras type charnockite locality is classifiedas transitional on the basis of variable minor ele-ment depletion and migmatitic-metasomatic aspect(Weaver, 1980). Several of the published mineral-ogical descriptions of the transition localities havebeen supplemented by personal communications tothe authors.

Pressures, as well as temperatures, are generallylower in the transitional regions than in the massifareas. Of course, the compilation does not take intoaccount transitional areas where the pressures ofcrystallization were too low to produce garnet inordinary basic and intermediate compositions, andthe pressures in Table 2 should be considered onlyas the higher pressure range of transitional granu-lites.

All of the transitional granulite terranes in Table 2lie in well-defined gradational zones adjacent toamphibolite-grade tracts except for the RubyRange, Montana, granulites (Dahl, 1980), which areisolated cordilleran "core complex" exposures.For this reason they could conceivably be classifiedwith the isolated granulite exposures in strong upliftareas described in the next subsection.

Granulites in young uplift areas

Several high pressure granulite terranes are ex-posed in areas of strong modern uplift, typically atcontinental margins. These include the PrecambrianBasal Gneisses of southwestern Norway (Kristian-sund area) (Krogh, 1980), the Doubtful Sound maficgranulites of the southwestern New Zealand Fiord-land (Oliver , 1977), the mafic granulites of the InnerAureole, Ronda Complex, southern Spain, thdCone Peak granulites of the Big Sur area, centralCalifornia (Compton, 1960), and, perhaps, the IvreaZone grarulites, northern Italy (Mehnert, 1975).

The latter are part of an isolated exposure ofprogressive amphibolite-to-granulite progression(Schmid and Wood, 1976), and perhaps should beclassified in the preceeding section.

The Doubtful Sound granulites give very highpressures of9-12.4 kbar. The granulites are garnet,two-pyroxene rocks in vein complexes in maficgarnet-clinopyroxene amphibolites. Sulfate-carbon-ate scapolite is well developed in the Fiordland area(Blattner, 1976). In the mafic character and pres-ence of scapolite these granulites greatly resemblethe lower-crustal granulite exotics of explosive vol-canic pipes in southeastern Australia (lrving, 1974)and southern Africa (Rogers and Nixon, 1975).Gravity surveys over the Fiordland show that theuppermost mantle lies only a few kilometers belowthe granulite exposures (Oliver, 1980). It seemsquite probable from these lines of evidence that anactual section of lowermost continental crust isexposed in the New Zealand Fiordland.

The Ronda Complex is believed to be a mass ofupper mantle peridotite which has been, for somereason, lifted to crustal levels during the Mid-Tertiary orogeny of the Betic Cordillera (Obata,1980). The inner thermal aureole contains maficgranulites with a primary garnet-clinopyroxene-plagioclase-quartz assemblage and a corona assem-blage with different plagioclase, clinopyroxene andgarnet (Loomis, 1977). These granulites are be-lieved to have been dragged up from the deepercrust by the peridotite mass (Loomis, 1972). Theclinopyroxene barometer gives 11.1 kbar for theprimary assemblage and 6.5 kbar for the coronaassemblage, at the temperature of about 800'C forboth assemblages suggested by Loomis (1977).These estimates correspond closely to the tempera-tures and pressures deduced by Obata (1980) fortwo equilibration stages of the peridotite mineral-ogy, presumably during a prolonged sojourn of therising mass_at two discrete levels.

Eclogite and garnetiferous peridotite masses ex-posed in the Norwegian fiord country have longbeen considered to be of subcrustal origin, incorpo-rated into the crust tectonically (e.9., O'Hara andMercy, 1963). Recent reappraisal (Krogh, 1977;Carswell and Gibb, 1980) concludes that the highpressure mafic and ultramafic rocks were metamor-phosed along with their enveloping crustal gneissesat pressures estimated by Krogh (1980) at 18r-3kbar during a mid-Proterozoic very high pressureevent, perhaps continental collision. Garnet-clino-pyroxene assemblages in acid gneisses analyzed by

Page 13: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES 215

Krogh (1980) give pressures as high as 13 kbarfromthe reaction (B) barometer. If, as is likely, current-ly-used geobarometers for peridotites and eclogitessomewhat overestimate pressures, and if the reac-tion (B) barometer somewhat underestimates pres-sures, as will be discussed in the next section, itmay well be that the eclogite pressure indicationand that of the enclosing gneisses converge, sup-porting the view that they were metamorphosedtogether, rather than mechanically juxtaposed dur-ing post-metamorphic times.

The Cone Peak, California granulites are a vein-complex in amphibolites in a coastal terrane ofprecipitous relief. The Cretaceous radiometric ageof a nearby charnockite body (Mattinson, 1978)supports the belief (Aitken, 1979) that the granulitesare the uplifted root zones of the Mesozoic Saliniablock batholiths. The pressure of8.7 kbar deducedhere (reaction (A) barometer) is in contrast with thehigh level of emplacement of most of the Californiamesozoic batholiths.

Granulites of the lvrea Zone are considered bysome (e.g. Mehnert, 1975) to represent upliftedlowermost crust, and a large positive gravity anom-aly across the zone supports this conjecture. How-ever, the pressures calculated from the assemblagegarnet-plagioclase-sillimanite-quartz by Newtonand Haselton (1981) are only 6.0-+-0.5 kbar for thehighest grade rocks, and this makes it seem doubt-ful that the Ivrea Zone is as representative of thelower continental crust as the Doubtful Sound gran-ulites. Although the garnet-pyroxene-plagioclase-quartz assemblage exists in lvreaZone mafic granu-lites, analyses satisfactory for geobarometry are notyet available.

Exotic fragments from volcanic pipes

Various granulites from explosive alkali basaltand kimberlite pipes have been reported recently. Alarge number of granulite fragments of undoubtedcrustal origin from the Lesotho kimberlites wereanalyzed by Griffin et al. (1979). Complete dataexist for one specimen (LT-2) from the Letseng-la-Terae pipe, which yielded an orthopyroxene pres-sure of 10.1 kbar at the estimated temperature of680"C. This specimen is of particular value in pro-viding the only example we have of a quenchedtwo-pyroxene granulite with which to investigatethe possibility of clinopyroxene reequilibration dur-ing slow uplift of granulite terranes, as discussedbelow.

The minette pipe, The Thumb, and many similar

tertiary occurrences in the Four Corners area,western United States, contain suites of lowercrustal granulites. Two specimens from The Thumbdescribed by Ehrenberg and Griffin (1979) yieldedclinopyroxene pressures of 7700 and 6600 bars atthe authors' preferred temperature of 650oC, con-firming the deep-crustal interpretation. Althoughone is mafic (THM 253) and the other granitic(THM 223), and although they have quite differentgarnets and plagioclases, the similarity of the pres-sures calculated for the two specimens indicatesthat they were quarried from nearly the same levelon the sub-Colorado Plateau geotherm. The garnet-clinopyroxene temperature for both of 650"C issomewhat high for a 20-35 km depth, particularlysince the garnet<linopyroxene temperature scalealmost always yields lower temperatures than thetwo-pyroxene scale where they can be compared ina single rock. In view of the difficulty of reequilibra-tion of silicate mineral compositions by diffusion indry rocks below 800"C (Ahrens and Schubert,1975), 650"C may be a temperature below whichreadjustment of mineralogic thermometers does nottake place even for long time scales.

Interpretations and conclusions

Accuracy of the barometers

Some idea of the potential accuracy of the twogeobarometers may be acquired from ancillary fieldand experimental information on the localities stud-ied here. The pressures of the Nain anorthositeaureole from reaction (A) agree with those obtainedfrom the recent revision of the ferrosilite-fayalite-qtrartz barometer (Bohlen and Boettcher, 1981).The two Scottish localities where the apparentandalusite to sillimanite transition is recorded inthermal aureoles give orthopyroxene pressures ator less than the Holdaway (197I) A12SiO5 triple-point. The Furua, Tanzania, and Qianxi, northeast-ern China, terranes contain kyanite as the regionalA12SiO5 polymorph, and the reaction (A) pressuressatisfy this condition with respect to the experimen-tal Al2SiO5 diagram (Fig. 1). All of the other largePrecambrian granulite terranes investigated havesillimanite. The Adirondack localities plot a shortdistance into the kyanite field, but theBohlen et aI.(1980) temperatures for these localities were basedon two-feldspar and Fe-Ti oxide thermometry,which methods have been criticized (Stoddard,1980) as yielding temperatures somewhat too low.

Page 14: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

Adirondock Lowlonds rs€Stoddord ( 1976)

-l zro"tro"c f-

216 NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES

^ 7

s

o r 6

a

G

700 800Temperolure ( "C)

Fig. 2. Calculated pressures of four orthopyroxene-garnet-plagioclase-quartz assemblages in basic rocks from theAdirondack Lowlands, analyzed by Stoddard (1976), asfunctions of assumed temperature. Note the slight dependenceon temperature. The pressure band width for four interbeddedmetapelites, containing ganret-sillimanite-plagioclase-quartz,arialyzed by Stoddard (1976), were calculated by the method ofHaselton and Newton (1981). The crossing area of the bandsdefined by the two barometers is consistent with the estimate of730"+30"C of Stoddard (1976) for the metamorphism.

Also, a kyanite locality in the central Adirondacks(Boone, 1978) indicates that the kyanite field wasapproached during metamorphism.

In several areas the orthopyroxene barometer ofreaction (A) can be compared with the metapelitebarometer based on garnet-plagioclase-sillimanite-qrnrtz calibrated by Newton and Haselton (1981).Figure 2 shows the d,PldT lines generated by thelatter and the reaction (A) barometers for several

rocks from the Adirondack Lowlands (Stoddard,1976; 1980). These axe all of the suitable rocksanalyzed by Stoddard, (1976) except for a fifthmetapelite, #75-C94, which plots about one kbarlower than the others, and is thought to be anoma-lous. The near temperature independence of thereaction (A) barometer is evident. The two barome-ters give consistent pressures of 5.5-6.5 kbar in thetemperature range 730o+30'C judged appropriateon the basis of eight geothermometers by Stoddard(1980). The migmatitic character of the AdirondackLowlands rocks is consistent with the experimen-tally-determined temperature of the beginning ofmelting of the assemblage biotite-K-feldspar-qaaftz in the presence of H2HO2 vapors (Wend-landt, 1981), which is about 740"C at 5-6 kbar.Table 3 gives comparisons of the average reaction(A) pressures for several localities with the New-ton-Haselton barometer. The agreement seems tobe good.

In summary, the calculated pressures of theorthopyroxene barometer appear to be consistentwith most of the independent information for thelocalities where comparisons are possible.

The clinopyroxene or reaction (B) barometerconsistently yields pressures two kbar lower thanthe reaction (A) barometers for rocks that containboth pyroxenes. Some of the reaction (B) pressureswould not be in good agreement with the A12SiO5experimental diagram or other reliable barometers.For instance, amphibolite 74-C-2488B of the Adi-rondack Lowlands transitional granulite regionyields a reaction (B) pressure of 4.2 kbar, which isconsiderably lower than pressures for this localitycalculated from the other barometers. This lack ofagreement indicates that the reaction (A) barometeris of superior accuracy.

Table 3. Comparison of the reaction (A) and Newton-Haselton barometers

Loca I i ty Source Estinnted T Opx-Gar-Plaq-02oc No. of-Teli.-flfarsTGar-S i l l -P lao-02

No. o f Ca lc . P (bars )rocks avq . rocks avq .

Western Sargurs, Roll lnson et alS . r n d t a ( 1 9 8 1 )

Adi rondackLovlands, Stoddard (1976)New York

Ruba Range(Ke l ly Mlne) , Dah l (1980)Jtlontana

0tter Lake, Perkins (1929)S. Quebec

730

750

74s

675

7772

6176

7798

6844

5823

8177

5933*

z

5

* NegLecte one roek tith a ealatT,ated pressure of 10,5 kbat

Page 15: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES 217

Precision of the barometers

The orthopyroxene pressures for the six Adiron-dack Highlands rocks show a standard deviation of-t-600 bars, and those of the Otter Lake area of-f 1100 bars. These figures might indicate the preci-sion of the method were it not for the possibilitythat some of the apparent scatter is actually realvariation of the crystallization pressure. This possi-bility may be explored with samples having therequisite assemblage but many different bulk com-positions, gathered from closely-spaced localitieswhich appear to be structurally contiguous. Datathis comprehensive are not yet available from anyof the granulite localities.

Closure problem

Another possible reason for the discrepancy be-tween barometers (A) and (B) is preferential clino-pyroxene reequilibration to lower pressures enroute to the surface. Evidence on this is providedby the two-pyroxene garnet granulite LT-2 from theLetseng-la-Terae kimberlite, Lesotho (Grifin et al.,1979). This is a "quenched" sample; that is, thetravel time to the surface is negligible on a geologi-cal time scale. The reaction (A) and (B) barometersshow, nevertheless, the discrepancy characteristicof the Precambrian massif two-pyroxene garnetgranulites (Table 2). This would seem to indicatethat the discrepancy lies in the thermodynamicformulation or input parameters, rather than in realdifferences in pressure indication. Lines of evi-dence that the geobarometric minerals commonlyretain their peak metamorphic compositionsthrough the slow ascent to the surface are the usuallack of zonation of pyroxenes, plagioclase, and,often, of garnets, and the indication that whenreequilibration at lower pressures does occur, theresult is usually symplectites or coronas of pyrox-enes, plagioclase, and, sometimes, garnets of com-positions greatly different from the primary assem-blage.

Re adjus t ed g eob arome t er expre s sions

Evidence has been presented that the diferencein pressure indication of the reaction (A) and (B)barometers is not due to real pressure differences offinal equilibration before closure of individual min-erals. The possibility is strong that the discrepancyhas its source in the thermodynamics of expressions(12), either in the magnitudes of the AGr quantitiesor in the validities of the activity expressions. The

lack of correlation of the amount of pressure dis-crepancy with compositions of the minerals arguesagainst the latter possibility. The average pressurediscrepancy is thus in the AGo's. The estimateduncertainties of the two barometers due to AG"uncertainties are i1500 bars for the orthopyroxenebarometer and 11600 bars for the clinopyroxenebarometer, and these AGo uncertainities are almostequally distributed between AIf and AS'factors inthe neighborhood of 1000 K.

In order to arrive at an unbiased pressure esti-mate based on the reaction (A) and (B) barometersfor rocks containing both pyroxenes one couldsimply average the two independent calculations.This would result in a pressure that is usually within1.2 kbar of each individual calculation. Since theorthopyroxene (reaction A) pressures seem to agreewith most other independent evidence, one couldweight the average by raisin_g the reaction (B)pressure to the limit of the AGsuncertainty and bylowering the reaction (A) pressure a small amount.The two barometers would read the same pressure,on the average, for two-pyroxene rocks if 1600 barsare added to expression (12) (cpx) and 600 bars aresubtracted from expression (12) (opx). However,such adjustments are empirical and necessarily sub-jective, and are left to the reader's discretion.

Geological apPlications

Classes of granulites

The five-fold classification of granulite localitiesadopted here is supported by the more or lesscharacteristic pressure ranges found for each class'The aureole granulites are of shallow origin andperhaps should be considered in the pyroxene horn-fels facies, rather than the granulite facies. The lowpressures found for the Nain anorthosite aureoleconfirm the model of Berg (1977) and Emslie (1978),who advocated a high-level emplacement for theLabrador and Grenville magmatic complexes, fol-lowed by a high pressure metamorphic overprint forthe complexes within the Grenville belt' The granu-lite terranes classified as transitional between am-phibolite facies and granulite facies give somewhatlower pressures than the granulite massif areas,which cluster in the range 8.9-l-1.5 kbar. The re-gions of strong modern uplift, such as DoubtfulSound, New Zealand, and the aureole of the RondaComplex, Spain, give very high pressures whichcorrespond to the base of the continental crust. Thevolcanic pipe exotics display a range of high pres-

Page 16: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

218 NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES

d -

0800 900 1000

Temperolure (oC)Fig. 3. Pressures for nineteen Precambrian granulite terranes.

Plotted points are the weighted averages of all calculatedreaction (A) and reaction (B) pressures for each locality (seetext). Circles'. sillimanite-bearing terranes. Squares.. kyanite-bearing temanes. Open symbols: massif granulites. Filledsymbols'. transitional granulite terranes. Half-filled symbols:indeterminate granulite stalus. Bhavani Sagar, Doddabetta, andNorth Slopes are individual localities within the Nilgiri Hills, S.India massif. Other localities explained in Table 2. Kyanite-sillimanite relations from Holdaway (1971).

sures, the highest being possibly subcrustal. ThepressurFtemperature indications of these cannottogether be interpreted as describing a Phanerozoiccontinental geotherm since some of them are fromareas of abundant magmatism where geotherms arelikely to be perturbed, and some of the samples mayactually be magmatic cognates. Also, the cut-offtemperatures below which granulite samples maycontinue to reequilibrate may be higher than thepresent normal sub-continent geothermal curve, inwhich case temperature-pressure indications ofsome deep-crustal exotics may be fossilized fromearlier times.

Transitional g ranulite terranes

Figure 3 shows weighted average pressures ofregional Precambrian granulite terranes, with 600bars subtracted from equation (12) (opx) and 1600bars added to equation (12) (cpx). Although thisadjustment is arbitrary, it allows data for bothreactions (A) and (B) to be used in a mutuallyconsistent way. The pressures, as well as the tem-peratures, of the granulite terranes classified astransitional average lower than those of the massifs(homogeneous high grade areas). This enhances thehypothesis that the granulite terranes are related tolower grade terranes by overprinted progressivemetamorphism, rather than by initial differences,magmatic fractionation, or mobilizate and restitefractions of partial melting, as other current hypoth-

eses propose. An average Precambrian geothermmight be constructed by drawing a P-T line be-tween the centroids of the transitional granulitesand the massif granulites in Figure 3. However, therelationship between transitional granulite terranesand massifs may be something other than a simpledepth-zone arrangement. A model commonly pro-posed for metamorphism in the Adirondacks iscontinental collision and crustal doubling (e.g.Mclelland and Isachsen, 1980). For this modelstrong lateral temperature and pressure gradientsmay be expected in the vicinity of the ancient suturezone.

Granulite massifs

The origin of ancient granulite terranes has longbeen a subject of speculation and controversy. Anew aspect of the problem is provided by thenarrow spread of pressure calculations (8.9-F1.5kbar) for twelve different massif localities. Thenearly constant pressures cannot simply be theresult of progressive excavation over time by deeperosion in shield areas to expose a "granulite level"at the present time, for several reasons. One is thatthe high-grade rocks are often demonstrably meta-sedimentary and thus went through an upper crustalhistory prior to being buried to some thirty kilome-ters. Some explicit tectonic mechanism must beinvoked to explain this great burial. The second isthat the terranes considered here span the timeinterval from 3675 m.y. (Qianxi Complex) to 1000m. y. (Adirondack Highlands). There is no apparentrelation between depth of metamorphism and antiq-uity which would be implied by the progressiveexcavation theory. Instead, the strong implicationis that a specific tectonic mechanism of metamor-phism was operative in the Precambrian which stillmay be operative today but whose effects may besomewhat diferent than in the past because ofreduced geothermal gradients or reduced intensityof action. Such a mechanism is continental overrid-ing during collision, as has been proposed for theapparently double crustal thickness at the juncturebetween Asia and the Indian subcontinent (Powelland Conaghan, 1975) and for high pressure meta-morphism in the Alps (Richardson and England1979). This hypothesis explains the repeated 7-10kbar pressures as the result of effectively instanta-neous loading of a continental margin with a 30-kmthickness of another continent. Continental shelfsedimentary facies, recumbent overfolding and sub-horizontal foliations, and COz fluid inclusions (from

Adircndock,.- Osit lomDundlaisnong.D4 sgrgli^ ^

"ili:[ #i:,^.zdwsororrs >ogotv qv,,

--ffi"{ " "jcii'o'"ia"u.l. "to"'t o,ro,,oo ohiilt

Lohe. r*,i1r,"? -$QQ;";i"'ao

Adirondocka ^ Sierro Ieone- Ruby qonqe(K)Lowlonds ^ : -xuDy ronqet> I

Page 17: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES 219

destruction of shelf carbonates), all nearly ubiqui-tous characteristics of ancient granulite terranes,are explainable by this mechanism (Newton andHansen, 1981).

Nature of the lower continental crust

Speculation about the lower continental crust hasfocussed on certain exposed granulite terranes aspossible actual lower crustal sections, tectonicallyuplifted. The ancient granulite massifs have beenput forward as a model largely because of their lowintrinsic heat production, which is a necessarycondition for the low heat-flow shield terranes(Heier, 1973). However, it seems unlikely that thisis a generally valid model because of the ubiquitoussupracrustal rocks, which are found in all Precam-brian granulite terranes, sometimes as dominantcomponents, as in the Sri Lanka Highland Series(Cooray, 1962). The interbanded pelitic granulite-mafic granulite sequence of the Ivrea Zone, north-ern Italy, has also been regarded as an actual deep-crustal section (Mehnert, 1975). A strong positivegravity anomaly over the granulites has been citedas evidence of near-surface upper mantle, but theanomaly could also be explained by a large gabbrointrusion under the granulite sequence (Schmid andWood, 1976). The low pressures of 6.0+0.8 kbarcomputed by Newton and Haselton (1981) from theas se mblage garnet-plagioclas e-sillimanite-quartzin Ivrea Zone metapelites indicates that the expo-sure is not the basal section of a crust of normalthickness. A third model is the mafic granulite tractof the Doubtful Sound area, New Zealand (Oliver,1976). In addition to being petrologically similar tothe mafic granulites of explosive pipe nodule suites,commonly of plagioclase-clinopyroxen+garnet-hornblende+scapolite-r qtrartz mineralogy (e.5.Rogers and Nixon, 1975), the pressures calculatedhere for the Doubtful Sound rocks are equivalent toapproximately 40 km depth in a crust of averagedensity 2.8 gm/cm3. This fact, together with theprofound gravity high over the area, suggestsstrongly that the Doubtful Sound granulites areindeed an actual deep-crustal array. The mechanisnof modern strong uplift, exhibited in some granuliteterranes, which elevates a lower crustal section, or,in the case of the Ronda peridotite, a subcrustalsequence, remains obscure.

AcknowledgmentsThe calorimetry upon which much of the present formulations

are founded was supported by National Science Foundation

grants to the first author over a several-year period, and byadditional support from the Materials Research Laboratory,NSF. The research of R. C. Newton and D. Perkins III has beensupported by NSF grant # EAR 78-15939 (RCN). Some of theIndian granulites discussed here were gathered and analyzedwith support from an NSF grant, # 79-05723 (RCN).

The authors are grateful to the following persons who contrib-uted unpublished analyses of minerals: J. H. Berg, F. C. Bishop,S. R. Bohlen, B. Cong, P. S. Dahl, E. C. Hansen, A. S.Janardhan, C. A. Johnson, H. R. Rollinson, J. V. Smith, P. R. A.Wells and R. Zhang.

Helpful discussions and correspondence with W. L. Grifrn,N. B. Harris, F. M. Richter and H. R. Rollinson are acknowl-edged.

ReferencesAhrens, T. J. and Schubert, G. (1975) Gabbro-eclogite reaction

rate and its geophysical significance. Reviews of Geophysicsand Space Physics, 13, 3E3-400.

Aitken, B. G. (1979) Stability relations of Ca-rich scapolite.Ph.D. thesis, Stanford University, l15 pp.

Anderson, G. M. (1977) The accuracy and precision of calculatedmineral dehydration equilibria. In D. G. Fraser, Ed., Thermo-dynamics in Geology, Reidel (Dordrecht-Holland), I 15-136.

Ashworth, J. R. and Chinner, G. A. (1978) Coexisting garnet andcordierite in migmatites from the Scottish Caledonides Con-tributions to Mineralogy and Petrology, 65,379-394.

Berg, J. R. (1977) Regional geobarometry in the contact aureolesof the anorthositic Nain Complex, Labrador. Journal of Petrol-ogy, 18,399-430.

Bishop, F. C. (1980) The distribution of Fe2* and Mg betweencoexisting ilmenite and pyroxene with applications to geother-mometry. American Joumal of Science, 280, 46-77.

Blattner, P. (1976) Replacement of hornblende by garnet ingranulite facies assemblages near Milford Sound, New Zea-land. Contributions to Mineralogy and Petrology, 55, lEl-190.

Bohlen, S. R. and Boettcher, A. L. (1981) Experimental investi-gations and geological applications of orthopyroxene geobaro-metry. American Mineralogist, 66, 951-96y'..

Bohlen, S. R. and Essene, E. I. (1979) A critical evaluation oftwo-pyroxene thermometry in Adirondack granulites. Lithos,12,335-345.

Bohlen, S. R. and Essene, E. J. (19E0) Evaluation of coexistinggarnet-biotite, garnet-+linopyroxene, and other Mg-Fe ex-change thermometers in Adirondack granulites. GeologicalSociety of America Bulletin, 91, Part II, 685-719.

Bohlen, S. R., Essene, E. J. and Hotrman, K. J. (1980) Updateon feldspar and oxide thermometry in the Adirondack Moun-tains, New York. Geological Society of America Bulletin, 91,Part I , l l0-113.

Boone, G. M. (197E) Kyanite in Adirondacks Highlands silliman-ite-rich grr.eiss, and P-? estimates of metamorphism. Geologi-cal Society of America Abstracts with Program, 10, 34.

Carswell, D. A. and Gibb, F. G. F. (19E0) The equilibrationconditions and petrogenesis ofEuropean crustal gamet lherzo-lites. Lithos. 13, 19-29.

Charlu, T. V., Newton, R. C. and Kleppa, O. J. (1975) Enthal-pies of formation at 970 K of compounds in the system MgO-Al2O3-SiO2 by high temperature solution calorimetry. Geo-chimica et Cosmochimica Acta, 39, 1487-1497.

Chartu, T. V., Newton, R. C. and Kleppa, O. J. (1978) Enthalpyof formation of some lime silicates by high-temperature solu-

Page 18: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

220 NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES

tion calorimetry, with discussion of high pressure phase equi-libria. Geochimica et Cosmochimica Acta. 42.367-375.

Compton, R. R. (1%0) Charnockitic rocks of the Santa LuciaRange, California. American Journal of Science, 258, 609-639.

Coolen, J. J. M. M. M. (1980) Chemical petrology of the Furuagranulite complex, southern Tanzania. GUA (Amsterdarn)papers, 13, (Series l), 1-258.

Cooray, P. G. (1962) Charnockites and their associated gneissesin the Precambrian of Ceylon. Quarterly Journal of the Geo-logical Society of London, ll8,239-266.

Cressey, G., Schmid, R. and Wood, B. J. (1978) Thermodynamicproperties of almandine-grossular garnet solid solutions. Con-tributions to Mineralogy and Petrology, 67,397-404.

Currie, K. L. (1971) The reaction 3 cordierite : 2 garrret + 4sillimanite + 5 quartz as a geologic thermometer in theOpinicon Lake region, Ontario. Contributions to Mineralogyand Petrology, 33, 215-226.

Dahl, P. S. (1980) The thermal-compositional dependence ofFe2+-Mg2+ distributions between coexisting garnet and py-roxene: applications to geothermometry. American Mineral-ogist, 65, 852-866.

Ehrenberg, S. N. and Grifrn, W. L. (1979) Garnet granulite andassociated xenoliths in minette and serpentinite diatremes ofthe Colorado Plateau. Geology, 7,483-487.

Emslie, R. F. (197E) Anorthosite massifs, rapakivi granites, andlate Proterozoic rifting of North America. Precambrian Re-search. 7. 6l-9E.

Ganguly, J. (1973) Activity-composition relation of jadeite inomphacite pyroxene: theoretical deductions. Earth and Plane-tary Science Letters, 19, 145-153.

Ganguly, J. and Kennedy, G. C. (1974) The energetics ofnaturalgarnet solid solution: I. Mixing of the aluminosilicate end-members. Contributions to Mineralogy and Petrology, 48,137-148.

Ghent, E. D. (1975) Plagioclase-garnet-AlzSiOs-quartz: a po-tential geobarometer-geothermometer. American Mineral-og is t ,6 l , 7 lOJ l4 .

Glassley, W. E. and Sorenson, K. (1981) Constant P"-Tamphib-olite to granulite facies transition in Agto (West Greenland)metadolerites: implications and applications. Journal of Petrol-ogy , 2l , 69-1 05 .

Goldsmith, J. R. (1980) The melting and breakdown reactions ofanorthite at high pressures and temperatures. American Min-eralogist, 65, 272-284.

Grew, E. S. (1980) Sapphirine + quartz association from Arche-an rocks in Enderby Land, Antarctica. American Mineral-ogist,65, 821-E36.

Griffin, W. L., Carswell, D. A. and Nixon, P. H. (1979) Lower-crustal granulites and eclogites from Lesotho, southern Africa.In F. R. Boyd and H. O. A. Meyer, Eds., The Mantle Sample.American Geophysical Union, 59-86.

Grifrn, W. L., McGregor, V. R., Nutman, A., Taylor, P. N. andBridgwater, D. (1980) Early Archaean granulite-facies meta-morphism south of Ameralik, West Greenland. Earth andPlanetary Science Letters, 50, 59-:7 4.

Haselton, H. T. (1979) Calorimetry of synthetic pyrope-grossulargarnets and calculated stability relations. Ph.D. thesis, Depart-ment of Geophysical Sciences, University of Chicago, 98 pp.

Haselton, H. T. and Newton, R. C. (1980) Thermodynamics ofpyrope-grossular garnets and their stabilities at high tempera-tures and high pressures. Journal ofGeophysical Research, 85,6973-6982.

Haselton, H. T. and Westrum, E. F. (1980) Low-temperatureheat capacities of synthetic pyrope, grossular, and py-ropeeogrossulara6. Geochimica et Cosmochimica Acta, 44,701-:709.

Hashimoto, M. (1968) Grossular-spessartine garnet from low-grade pelitic schist of the Katsuyama District, OkayamaPrefecture. Journal of the Geological Society of Japan,74,343--345.

Heier, K. S. (1973) Geochernistry of granulite facies rocks andproblems of their origin. Philosophical Transactions of theRoyal Society of London, A.,273, 429-442.

Hensen, B. J. (1976) The stability of pyrope-grossular garnetwith excess silica. Contributions to Mineralogy and Petrology,55.279-292.

Hensen, B. J. and Green, D. H. (1971) Experimental study of thestability ofcordierite and garnet in pelitic compositions at highpressures and temperatures. L Compositions with excessalumino-silicate. Contributions to Mineralogy and Petrology,33.309-330.

Hensen, B. J. and Green, D. H. (1973) Experimental study of thestability ofcordierite and garnet in pelitic compositions at highpressures and temperatures. IL Synthesis ofexperimental dataand geologic applications. Contributions to Mineralogy andPetrology, 38, 15l-166.

Hensen, B. J., Schmid, R. and Wood, B. J. (1975a) Composition-activity relationships for pyrope-grossular garnets. NERCReprint #3 (Edinburgh-Manchester), 24-25.

Hensen, B. J., Schmid, R. and Wood, B. J. (1975b) Activity-composition relations for pyrope-grossular garnet. Contribu-tions to Mineralogy and Petrology, 51, 161-166.

Holdaway, M. J. (1971) Stability of andalusite and the aluminumsilicate phase diagram. American Journal ofScience, 271,97-l 3 l .

Holland, T. J. B. (1981) Thermodynamic analysis of simplemineral systems. In R. C. Newton, A. Navrotsky and B. J.Wood, Eds., Thermodynamics of Minerals and Melts, Spring-er. 19:34.

Hdrmann, P. K., Raith, M., Raase, P., Ackermand, D. andSeifert, F. (19E0) The granulite complex ofFinnish Lappland:petrology and metamorphic conditions in the Ivalojoki-Inari-jiirvi area. Geological Survey of Finland Bulletin, 308, l-95.

Hsu, L. C. (19E0) Hydration and phase relations of grossular-spessartine garnets at PsrO = 2 Kb. Contributions to Mineral-ogy and Petrology, 71,407415.

Irving, A. J . (1974) Geochemical and high pressure experimentalstudies of garnet pyroxenite and pyroxene granulite xenolithsfrom the Delegate basaltic pipes, Australia. Journal of Petro-logy, 15, l-40.

Janardhan, A. S. and Srikantappa, C. (1975) Geology of thenorthern parts of the Sargur schist belt between Mavinahalliand Dodkanya, Mysore District, Karnataka. The Indian Min-eralogist, 16,66-75.

Kawasaki, T. and Matsui, Y. (1977) Partitioning of Fe2+ andMg2* between olivine and garnet. Earth and Planetary ScienceLetters. 37. 159-166.

Kerrick, D. M. and Darken, L. S. (1975) Statistical thermody-namic models for ideal oxide and silicate solid solutions, withapplication to plagioclase. Geochimica et Cosmochimica Acta,39, l43t-1M2.

Krogh, E. I. (1977) Crustal and in situ origin of Norwegianeclogites. Nature, 269, 730.

Krogh, E. J. (1980) Compatible P-T conditions for eclogites and

Page 19: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANULITES 221

surrounding gneisses in the Kristiansund Area, Westem Nor-way. Contributions to Mineralogy and Petrology, 75,387-393.

Krupka, K. M., Kerrick, D. M. and Robie, R. A. (1979) Heatcapacities of synthetic orthoenstatite and natural anthophyllitefrom 5 to 1000 K. EOS, 60, ,()5.

Krupka, K. M., Robie, R. A. and Hemingway, B. S. (1979) Hightemperature heat capacities of corundum, periclase, anorthite,CaAl2Si2Or glass, muscovite, pyrophyllite, KAlSi3O6 glass,grossular, and NaAlSi3Os glass. American Mineralogist, 64,E6-101.

Kushiro, I. (l%9) Clinopyroxene solid solutions formed byreactions between diopside and plagioclase at high pressures.Mineralogical Society of America Special Publication #2,179-t 9 l .

Kushiro, I. and Yoder, H. S. (1966) Anorthite-forsterite andanorthite-+nstatite reactions and their bearing on the basalt-eclogite transformation. Journal of Petrology, 7, 337 -362.

Loomis, T. P. (1972) Coexisting aluminum silicate phases incontact metamorphic aureoles. American Joumal of Science,n2.933-945.

Loomis, T. P. (1977) Kinetics of a gamet granulite reaction.Contributions to Mineralogy and Petrology, 62, l-22.

Mattinson, J. M. (1978) Age, origin, and thermal histories ofsome plutonic rocks of the Salinian block, California. Contri-butions to Mineralogy and Petrology, 67,233-245.

Mclelland, J. and Isachsen, Y (1980) Structural synthesis of thesouthern and central Adirondacks: A model for the Adiron-dacks as a whole and plate-tectonics interpretations. Geologi-cal Society of America Bulletin, 91, (Part l),68-:72.

Mehnert, K. R. (1975) The Ivrea Zone. Neues Jahrbuch fiirMineralogie, Abhandlungen, 125, 156-19,

Naidu, P. R. J. (l%3) A layered complex in Sittampundi, MadrasState, India. Mineralogical Society of America Special Paper,r. 116-123.

Newton, R. C. (197E) Experimental and thermodynamic evi-dence for the operation of high pressures in Archaean meta-morphism. In B. F. Windley and S. M. Naqvi, Eds., ArchaeanGeochemistry, Elsevier, 221-240.

Newton, R. C., Charlu, T. V. and Kleppa, O.I. (1977) Thermo-chemistry of high pressure garnets and clinopyroxenes in thesystem CaO-MgO-Al2O3-SiO2. Geochimica et Cosmochi-mica Acta, 41, 369-i77.

Newton, R. C., Charlu, T. V. and Kleppa, O. J. (1980) Thermo-chemistry of the high structural state plagioclases. Geochimicaet Cosmochimica Acta, 44,933-941.

Newton, R. C. and Hansen, E. C. (l9El) The origin of Proterozo-ic and Late Archaean Charnockites---+vidence from fieldrelations and experimental petrology. Geological Society ofAmerica Bulletin (in press).

Newton, R. C. and Haselton, H. T. (1981) Thermodynamics ofthe garnet-plagioclase-Al2SiO rquiartz geobarometer. In R.C. Newton, A. Navrotsky and B. J. Wood, Eds., Thermody-namics of Minerals and Melts, Springer (New YorD, 129-145.

Newton, R. C. and Wood, B. J. (1979) Thermodynamics of waterin cordierite and some petrologic consequences of cordieriteas a hydrous phase. Contributions to Mineralogy and Petrolo-gy, 6E, 391-405.

Nixon, P. H., Reedman, A. J. and Burns, L. K. (1973) Sapphi-rine-bearing granulites from Labwor, Uganda. MineralogicalMagazine, 39,420-428.

Obata, M. (1980) The Ronda peridotite: garnet-, spinel-, andplagioclase lherzolite facies and the P-T trajectories of a high-

temperature mantle intrusion. Joumal of Petrology, 21, 533-572.

O'Hara, M. J. and Mercy, E. L. P. (l%3) Petrology andpetrogenesis of some garnetiferous peridotites. Transactionsof the Royal Society of Edinburgh, volume LXV,251:314.

O'Hara, M. J. and Yarwood, G. (1978) High pressure-tempera-ture point on an Archaean geotherm, implied magma genesisby crustal anatexis and consequences for gaf,net pyroxenethermometry and barometry. Philosophical Transactions ofthe Royal Society of London, A228, 441456.

Oka, Y. and Matsumoto, T. (1974) Study on the compositionaldependence of the apparent partition coefficient of iron andmagnesium between coexisting garnet and clinopyroxene solidsolutions. Contributions to Mineralogy and Petrology, 48,rrs-rzt.

Oliver, G. J. H. (1976) The hieh grade metamorphic rocks ofDoubtful Sound, Fiordland, New Zealand. Ph.D. thesis, OtagoUniversity (Dunedin, New Zealand).

Oliver, G. J. H. (1977) Feldspathic hornblende and garnetgranulites and associated anorthosite pegmatites from Doubt-ful Sound, Fiordland, New Zealand. Contributions to Mineral-ogy and Petrology, 65, lll-121.

Oliver, G. J. H. (1980) Geology of the ganulite and amphibolitefacies gneisses Doubtful Sound, Fiordland, New Zealand.New Zealand Journal of Geology and Goephysics,23, Tl-42.

O'Neill, H. St. C. and Wood, B. J. (1979) An experimental studyof Fe-Mg partitioning between garnet and olivine and itscalibration as a geothermometer. Contributions to Mineralogyand Petrology, 70, 59:70.

Orville, P.M. (1972) Plagioclase cation exchange equilibria withaqueous chloride solution: Results at 7C0'C and 2000 bars inthe presence of quartz. American Journal of Science, 72,234-272.

Perchuk, L. L., Podlesskii, K. K. and Aranovich, L. Ya. (1981)Calculation of thermodynamic properties of end-member min-erals from natural parageneses. In R. C. Newton, A Navrotskyand B. J. Wood, Eds., Thermodynamics of Minerals andMelts, Springer (New York), lll-129.

Perkins, D. (1979) Application of new thermodynamic data tomineral equilibria. Ph.D. thesis, University of Michigan, AnnArbor, Michigan.

Perkins, D., Westrum, E. F. and Essene, E. J. (19E0) Thethermodynamic properties and phase relations of some miner-als in the system CaO-AlzOr-SiOrHzO. Geochimica et Cos-mochimica Acta, 44, 6l-84.

Pichamuthu, C. S. (1965) Regional metamorphism and charnock-itization in Mysore State, India. The Indian Mineralogist, 6,#l and2.119-126.

Pichamuthu, C. S. (1970) On the occurrence of garnet in char-nockite. Journal of the Geological Society of India, 11,273-274.

Powell, C. M. and Conaghan, P. J. (1975) Tectonic models forthe Tibetan Plateau. Geology, 3,727:731.

Prigogine, L. and Defay, R. (1954) Chemical Thermodynamics.Longmans, G.reen and Company, 543 pp.

Richardson, S. W. and England, P. C. (1979) Metamorphicconsequences of crustal eclogite production in overthrustorogenic zones. Earth and Planetary Science Letters, 42,183-190.

Robie, R. A., Hemingway, B.S.and Wilson, W. H. (1978) Low-temperature heat capacities and entropies offeldspar glassesaird of anorthite. American Mineralogist, 63,l@-123.

Page 20: Thermodynamic calibration of geobarometers based on the ... · dard enthalpy and entropy changes ofthe reactions, Af the partial molal volume differences, a the activities at one

222 NEWTON AND PERKINS: GEOBAROMETERS FOR CHARNOCKITES AND GRANUI'ITES

Rogers, N. and Nixon, P. H. (1975) Geochemistry of lowercrustal granulite xenoliths from Lesotho kimberlites. AnnualReport of the Research Institute of African Geology, Depart-ment of Earth Sciences, University of Leeds, 38J1.

Rollinson, H. R. (19E0) Iron titanium oxides as an indicator ofthe role of the fluid phase during the cooling of granitesmetamorphosed to granulite grade. Mineralogical Magazine,43,623-631.

Rollinson, H. R. (1981a) Garnet-pyroxene thermometry andbarometry in the Scourie granulites, N.W. Scotland. Lithos,14.225-238.

Rollinson, H. R. (1981b) P-T conditions in coeval greenstonebelts and granulites from the Archaean of Sierra Leone. Earthand Planetary Science Letters (in press).

Sack, R.O. (1980) Some constraints on the thermodynamicmixing properties of Fe-Mg orthopyroxenes and olivines.Contributions to Mineralogy and Petrology, 71,237-246,

Saxena, S. K. (1973) Thermodynamics of Rock-Forming Crystal-line Solutions. Springer-Verlag, IEE pp.

Schmid, R. and Wood, B. J. (1976) Phase relationships ingranulitic metapelites from the Ivrea-Verbano Zone (northernItaly). Contributions to Mineralogy and Petrology, 54, 255-279.

Smith, D. (1972) Stability of iron-rich pyroxene in the systemCaSiQ-FiSiO3-MgSiO3. American Mineralogist, 57, l4l3-1428.

Smith, J. V. Qnq Feldspar Minerals, volume I. Crystal Struc-ture and Physical Properties. Springer-Verlag,627 pp.

Stoddard, E. F. (1976) Granulite facies metamorphism in theColton-Rainbow Falls area, northwest Adirondacks, NewYork. Ph.D. thesis, University of California at Los Angeles,Los Angeles, California.

Stoddard, E. F. (19E0) Metamorphic conditions at the northernend ofthe northwest Adirondack Lowlands. Geological Socie-ty of America Bulletin, 91, Part I, 97-100.

Subramaniam, A. P. (1956) Mineralogy and petrology of theSittampundi Complex, Salem District, Madras State, India.Geological Society of America Bulletin, 67,317423.

Thompson, A. B. (1976) Mineral reactions in pelitic rocks. II.

Calculation of some P-T-X (Fe-Mg) phase relations. Ameri-can Journal of Science, n6,425454.

Weaver, B. L. (19E0) Rare-earth element geochemistry of Ma-dras granulites. Contributions to Mineralogy and Petrology,7t,27r-279.

Weaver, B. L., Tarney, J., Windley, B. F., Sugavanam, E. B.and Venkata Rao, V. (1978) Madras granulites: Geochemistryand P-T conditions of crystallization. In B. F. Windley and S.M. Naqvi, Eds., Archaean Geochemistry, Elsevier, 177-204.

Wells, P. R. A. (1977) Pyroxene thermometry in simple andcomplex systems. Contributions to Mineralogy and Petrology,62.129-139.

Wells, P. R. A. (1979) Chemical and thermal evolution ofArchaean sialic crust, southern West Greenland. Journal ofPetrelogy, 20, 187 -226.

Wendlandt, R. F. (lg8l) Influence of CO2 on melting of modelgranulite facies assemblages: a model for the genesis ofcharnockite. American Mineralogist, 66, 1164-117 4.

Wood, B. J. (1974) The solubility of alumina in orthopyroxenecoexisting with garnet. Contributions to Mineralogy and Pe-trology, 46, 1-15.

Wood, B. J. (1975) The influence of pressure, temperature and6uk composition on the appearance of garnet in ortho-gneisses---an example from South Harris, Scotland. Earth andPlanetary Science Letters, 26, 29,-311.

Wood, B. J. (1979) Activity-composition relations in Ca(Mg, Fe)Si2OJaAl2SiO6 clinopyroxene solid solutions. AmericanJournal of Science, 279, E54-E75.

Wood, B. J. and Banno, S. (1973) Garnet-orthopyroxene andorthopyroxene-+linopyroxene relationships in simple andcomplex systems. Contributions to Mineralogy and Petrology,42. tO9-124.

Wood, B. J. and Kleppa, O. J. (19E1) Thermochemistry offorsterite-fayalite olivine solutions. Geochimica et Cosmochi-mica Acta, 45,529-534.

Manuscript received, IuIy 30, 19E1;accepted for publication, October 27, l9EI .