American Mineralogist, Volume 79, pages 951-959, 1994

O-isotope variations in a porphyroclastic meta-anorthosite: Diffusioneffects and false isotherms

Z. D. SrHnpInstitut de Min6ralogie et P6trographie, Universit6 de I-ausanne, CH-1015 Lausanne, Switzerland

D. P. MoncHERDepartment of Geological Sciences, University of Kentucky, Lexinglon, Kentucky 40506-0053, U.S.A.

AssrRAcr

A detailed O-isotope study of annealed porphyroclastic meta-anorthosite was performedto assess the degree of postpeak-metamorphic and postshearing reequilibration. Stableisotope and petrologic evidence indicate that the Whitestone anorthosite, Ontario, wasinfiltrated by an 'sQ- and COr-rich fluid subsequent to peak Grenville granulite faciesmetamorphism. Ductile shearing occurred during and following peak metamorphism. Rel-ict igneous plagioclase phenocrysts (An.o) occur as coarse-grained porphyroclasts in a fine-grained recrystallized matrix of plagioclase (Anoo) + hornblende + garnet * scapolite *epidote + ilmenite. The 6'80 values of mineral fragments as small as 200 pm were ana-lyzed to assess constituents developed at different stages of the rock's history. The 6r8O.ro"nvalues (7oo) are as follows: porphyroclastic plagioclase : 8.1-8.5 (av. 8.38 ! 0.14,2o);matrix plagioclase :8.2-9.2 (av. 8.65 ! 0.32); garnet: 6.87 + 0.02, n: 2;hornblende: 6.44 + 0.00, n: 2; ilmenite : 1.93 + 0.18, n : 2. In-situ analyses agree within errorbut are far more scattered (av. 8.2 + 0.5).The plagioclase porphyroclasts are isotopicallyhomogeneous, a result of rapid difitsion of 180/160 at hlgh Z and slow cooling rates.

The theoretical d'8O values of all minerals were calculated on the basis of O-volumediffi.rsional considerations using diffusion rate data for both hydrous and anhydrous con-ditions. The Whitestone data set is in far better agreement with the diftrsion data obtainedin hydrothermal experiments, indicating that cooling probably occurred in the presence ofa hydrous fluid. Similar calculations made for other granulite terranes are generally con-sistent with cooling under anhydrous conditions. The observed d'8O values of all phasesfrom the Whitestone anorthosite yield an enoneous equilibration temperature of 470 L30'C that can be explained by an infinite reservoir model involving postpeak-metamor-phic O-isotope exchange with an infinite reservoir of plagioclase. Comparable false iso-therms are characteristic of other granulite terranes as well.

INrnooucrtox

The use of an isotherm plot for O-isotope thermometry(Javoy et al., 1970) has become a standard tool for eval-uating equilibria among coexisting mineral pairs and forestimating equilibration temperatures (e.g., J av oy, | 9 7 7).The isotherm plot, a graphical representation ofthe d'8Ovalues of minerals vs. the temperature coefficient of frac-tionation, is an extension ofthe concept that equilibriumamong three minerals is demonstrated if two indepen-dent, isotopic, temperature estimates agree within exper-imental error (Clayton and Epstein, l96l). Isotherm plotshave also been used to estimate fractionation factors be-tween coexisting phases for which no experimental dataare available (Bottinga and Javoy, 1973, 1975). The re-liability of this approach has been questioned by Deines(1977), who suggested that most terrestrial samples donot preserve isotopic equilibrium. Retrograde diffusionalexchange may explain the partial resetting that can beobserved in slowly cooled rocks (Giletti, 1986). If dis-

0003-o04x/94l09 1 M95 I $02.00

equilibrium among mineral triplets is common, then theapplication of the isotherm plot to thermometry and ther-mometric calibrations is not valid.

In the present study, we investigate the O-isotope vari-ations in coexisting minerals in a high-grade, ductilelyrecrystallized, scapolite-bearing meta-anorthosite forwhich the history of fluid-rock interaction is well knownfrom a petrological basis. The anorthosite has been infil-trated by an r8O- and COr-rich fluid at near-peak meta-morphic conditions (Moecher et al., 1992).In this study,6'80 values of plagioclase porphyroclasts and coexistingphases were determined on a scale of millimeters to sub-millimeters to evaluate the scale of isotopic equilibrationthat accompanied fluid infiltration and continued duringcooling. The validity of isotherm plots is assessed in termsof isotopic variations (or lack of variations) within singlecrystals and the degree of isotopic equilibrium amongcoexisting phases. The D18O values ofgarnet, hornblende,ilmenite, and two generations of plagioclase yield an ap-

95 r

952 SHARP AND MOECHER: O ISOTOPES IN META-ANORTHOSITE

Rlm to Rlmfiayerce ol PlagloclasePotphytoclast lc

5 0.69FE 0.sq)

o 0.4

.5 0.6I6Llz 0.5q)

€ 0.4a

: Grain Boundaries :',^------ --------.--,--""..i -E

i F - - i . 8-ffi*lhq-m jer4.qffi d

l +Mattix Plag

Fig. l. Photomicrograph of the annealed porphyroclastictexture in meta-anorthosite [(A) plane light, (B) crossed polars;field of view is 6 mml. A plagioclase porphyroclast (Pg) exhib-iting internal subgrain formation is surrounded by a matrix con-sisting mainly of plagioclase (clear in plane light) with lessergarnet (Gt), hornblende (Hbl), scapolite (Sc), epidote (Ep), il-menite (Ilm), and biotite (Bt).

parent equilibrium temperature that has no known geo-logical significance but can be explained in terms of ret-rograde reequilibration by O-diffirsional exchange.

S*rpr,n DEscRrPTroN

The samples are from a Grenville meta-anorthosite ofthe Central Gneiss belt in southern Ontario, Canada.Ductile deformation and mylonitization occurred duringand following the peak of granulite facies metamorphismbetween l160 and ll20 Ma (van Breemen et al., 1986;Tucillo el al., 1992; Mezger et al., 1993). The sample isa porphyroclastic gabbroic meta-anorthosite, with centi-meter-scale, relict, igneous plagioclase phenocrysts in amatrix of fine-grained (

agate pestle to produce fragments as small as 200 pm ona side (O'Neil et al., 1992). Individual fragments wereremoved from the tape, and their spatial positions wererecorded accurately with respect to the original, unbrokenplate (Fig. 3). Grains of matrix plagioclase, garnet, horn-blende, and ilmenite were also sampled with this tech-nique. Individual grains were reacted with BrF' by heat-ing with a CO, laser (Sharp, 1990,1992). O was convertedto CO, and passed directly into the mass spectrometer.

The plagioclase porphyroclasts are isotopically homo-geneous at the submillimeter scale, with a 6180 weightedaverage value (given errors of 0. lToo analytical uncertain-ty) of 8.38 + 0.14 (950/o confidence) (Table l, Fig. 3a, 3b).The more sodic matrix plagioclase is, on average, slightlyenriched in lEO relative to porphyroclasts (6'EO : 8.65 +

0.32). Garnet averages 6.87 + 0.02, hornblende averages6.44 + 0.00, and ilmenite averages 1.94 t 0.13V*.

In-situ rneasurements were made by heating a plate,200 pm thick, with a focused laser. Short, 100-ms pulseswere made repeatedly until a hole -200-300 pm in di-ameter was bored through the sample. Reactions weremade in a BrF, atmosphere at 0. I bars with a laser powerof 6 W. The in-situ data are in good agreement with themineral-separate data, but the scatter is significantly larg-er (av. : 8.3 + 0.5, 950/o confidence given analytical errorof 0.4Vm; Table 1). The higher degree of scatter in the in-situ data is most likely due to partial reaction and isotopicfractionation in the laser pit and with the surroundingmaterial. The in-situ method has been used successfullyfor some minerals (quartz: Conrad and Chamberlain,1992; Kirschner et al., 1993; garnet Chamberlain andConrad, 1991), but, apparently, feldspars are too reactivewith F for the in-situ technique (see also Elsenheimer andValley, 1992). The advantages of the microsamplingtechnique over the in-situ technique for this kind ofstudyinclude (l) complete reaction of mineral phases with BrF ;(2) no interaction with less reactive material near the laserbeam (edge eflects); (3) experimental yields may be mea-sured; and (4) careful assessment of mineral purity bypetrographic inspection can be made prior to analysis.Physical mineral separation is far more time-consumingthan in-situ analyses, but it must be considered the meth-od of choice until a more accurate method for in-situanalyses of feldspars is developed.

DrscussroN

Infiltration by a COr-rich fluid during metamorphismperturbed the isotopic equilibrium of the anorthosite(Moecher et al., 1992). The response to this perturbationwas evaluated at two scales: first, by determining varia-tions in the D'8O values ofsingle feldspar porphyroclasts,and second, by determining the isotopic fractionation thatexists between coexisting phases. Isotopic variationswithin a single crystal is controlled only by O-volumedifusion, whereas isotopic fractionations between coex-isting phases in a single hand specimen are controlledboth by grain boundary and by volume diffirsion.

953

Fig. 3. (a) Illustration of a thick section and O-isotopic com-position of samples extracted from the porphyroclastic meta-anorthosite (all values parts per mil, relative to SMOW). Thedark areas consist offine-grained matrix plagioclase with lesserhornblende, garnet, ilmenite, scapolite, and epidote. @) Sketchof an individual porphyroclast, with sample locations and iso-topic values. Starred value is a mixed garnet + feldspar inclusionin the porphyroclast.

Original plagioclase phenocrysts should have a D'EOvalue of 6-7V*. Infiltration of the CO, fluid from thesurrounding marbles has increased the average 0'8O valueof the plagioclase porphyroclasts up 1o l}Vm (Moecher etal., 1992). In the absence of deformation and recrystal-lization, the plagioclase crystals would exchange O withthe infiltrating fluid by volume diftrsion. The isotopicdifiirsion profrles for a plagioclase porphyroclast 4 mm indiameter are calculated as a function of time (Fig. a). Odiffirsion rates in feldspars are strongly dependent on thecomposition of the exchange fluid. Diffirsion rates deter-mined from exchange experiments with anhydrous fluids(O, or COr) are orders of magnitude slower than those

SHARP AND MOECHER: O ISOTOPES IN META-ANORTHOSITE

954

TaBLE 1. O-isotope results

SHARP AND MOECHER: O ISOTOPES IN META-ANORTHOSITE

Sample no.6180

(7@ vs. SMOW)Weight(ms)

CO,re@vercd

(pm) Comments

Plagioclase porphyroclast89

1 01 41 51 8

2730a30bAv.

23252628313233Av.

8.488.348.208.558.228.07

8.488.518.s38.38 + 0.14

9 . 1 78.698.408.968.338.238.748.65 + 0.32

6.856.896.446.442.061.81

d'60(9o vs. SMOW)

n.d.n.d.12.18.8

n.d.1 1 . 9

28.440.417.5

Plagioclase groundmassn.d.n.d.33.024.1n.d.n.d.n.o.

Other phases19 .828.617.415 .416 .4n.d.

1.O21 . 1 0o.770.591.560.79

1 .762.441 . 1 0

1 .041 . 1 61 .921 .471 .751 . 1 31 .64

porphyroclast, near edgeporphyroclast at edge@ntains very small garnet inclusionoenter of porphyroclastcenter of porphyroclasteratic trace on mass spec.

contaminated gas sample?center of porphyroclast; very cleanporphyroclastfrom same fragment as 30a

minor garnet and hornblende inclusionsvein, cleanmultiple small garnet and hornblende inclusionsvein, minor dusty inclusionsvetnvetnvein, minor garnet and hornblende inclusions

GarnetGarnetHornblendeHornblendellmenitellmenite

Expt.

1 .502.O71 .261 . 2 11 .672.63

In-situ analyses of plagioclase pophyroclast

Voltage.(mass 44) Comments

s101-2s101-3s101-4s101-5s101-6s101-7s101-8s101-9s101-10s l 0 1 - 1 1s101-12s101-13s101-14s101-15s101-16s101-17s101-18s101-19Av.

8.906.368.844.588.587.907.778.258.507.688.737.927.396.558.656.126.609.198.24 + 0.4

5.73.54.5n.d.4.24.23.94.3n.o.3.02.93.83.63.34.22.84.24.2

center of porphyroclast, clean hole73 distance to porphyroclast edge+5 distance to edgeat edge, contaminated by groundmass phases?at center, next to 5101-2% distance to edge, next to S101-3% distance to edge, next to S101-4at edge next to 5101-5, clean holey3 distance to edge, other directiony2 distance to edge, next after 5101-10y4 distance to edge, next after 5101-11at edge, next after 510112, clean holecentral portion of crystalnext to 5101-15, bad trace on mass spec.73 distance to edge% distance to edgeat edgeoutside of porphyroclast-groundmass plag.

Nofe: locations of samples correspond to Fig. 3a and 3b; n.d. : not determined.. One volt - 0.1 mol COa.

obtained from exchange with hydrous fluids (Giletti etal.,1978; Elphick et al., 1988). The plagioclase porphy-roclasts would completely exchange with a hydrous fluidin several thousands ofyears (Fig. 4), whereas the originald18O values would be retained in the cores of crystals fortens of millions of years if exchange occurred with a com-pletely anhydrous fluid (Fig. 4). There is no observablevariation in the dlEO value of the plagioclase from coreto rim (Fig. 3b), indicating either that the infiltrating fluidwas hydrous or that deformation accompanied fluid in-filtration. If a hydrous fluid is present, the preservationoffeldspar zoning can only be expected as a result oflow-

temperature hydrothermal events (Elsenheimer and Val-l ey ,1993 ) .

Among multiple phases, the traditional approach forevaluating isotopic equilibrium is to check for equilibri-um isotopic fractionations among at least three phases(Clayton and Epstein, 1958; Bottinga and Javoy, 1975;Deines, .l 977). Temperatures of isotopic equilibration be-tween any two phases i and, j are related by the equation

1000 ln d,i - b,i: aii '106/72 (l)

where 1000 ln au = A'8O(i - .;1 : D'EO, - d"Or, and aija\d bij are the temperature coefficients of fractionation

200

150

100

15

50

$ t144.y.\s _5f-

center

for the mineral pair I and j. A plot of 1000 ln a,, - b,, vs.a,,. 106 for all phases (where phase i : reference phase)defines a straight line with a slope proportional to l/72if all phases are in equilibrium (Javoy et al., 1970). Thepresent data define a straight line (r'z : 0.98) with anequilibrium temperature of 470'C (Fig. 5). The differ-ences between the DltO values of the feldspar porphyro-clasts (8.387-) and groundmass (8.657-) can be explainedin terms of the different Al-Si ratio. Albite is enriched in18O relative to anorthite (O'Neil and Taylor, 1967). Theequilibrium isotopic fractionation between Anoo and Anuois 0.2 at 700 'C and 0.4 at 470 .C. Althoueh chemicalequilibrium has not been attained between the plagio-clase porphyroclasts and groundmass, O-isotope equilib-rium has been attained.

The best-fit temperature of 470 oC from the stable-isotope data has no known geological significance. Thereis no evidence for a postpeak-metamorphic event at -500

"C. The temperature of the last metamorphism is 700'C,and cooling below that point appears to have been grad-ual and continuous. Cooling rate estimates for the Gren-ville meta-anorthosite from thermochronological dala arel-5 'Clrn.y. over the temperature range 700-

956 SHARP AND MOECHER: O ISOTOPES IN META-ANORTHOSITE

TABLE 2. Diffusion data used in calculations

Hydrous/Mineral anhydrous

Preexponentialfactor Do (cm,/s)

Activationenergy

(O) (kJ/mol) Referenc€

a quartzP quartz

Anorthite

Potassium feldspar

Pyroxene

Homblende

Magnetite

Hematite

llmeniteGarnetGarnet

hydroushydrousanhydroushydrousanhydroushydrousanhydroushydrousanhydroushydrousanhydroushydrousanhydroushydrousanhydrous

hydrousanhydrous

1904 x 1 9 t2.1 x 1g-a1 .4 x 10 -71 x 10 -s4.5 x 10- '1 x 1 9 s1 . 5 x 1 0 u6.31.0 x 10-?

assumed equal to hydrous value27 x 1g -a4.3 x 10-76.3 x 10

630assumed equal to hematite

2 x 10 -6assumed equal to hydrous value

2U't421591 1 0236107236226405171

203211405405

280

1I23435678

91 01 11 2

1 3

References: (1) Giletti and Yund, 1984; (2) Sharp et al., 1991; (3) Giletti et al., 1978; (4) Elphick et al., 1988; (5) assumed equal to anorthite; (6)Farver, 1989; (7) Connolly and Meuhlenbachs, 1988; (8) Farver and Giletti, tg85; (9) Giletti and Hess, 1988; (10) Sharp, 1991; (11) equal to theanhydrous values of Reddy and Cooper (1983) with Do : Dq*r x 100; (12) Reddy and Gooper, 1983; (13) Fortier and Giletti, 1989.

calculated to be in equilibrium with garnet at 700 "C.Both methods give identical values of 6r8o*hor" **, indi-cating that the rock behaved as a closed system followingpeak metamorphism.

The final 6'tO values of each phase were calculatedfollowing the procedures of Gilerti (1986) and of JenkineI al. (1994).In the Giletti model, the closure temperatureis based on two assumptions: (l) that there is an infiniteO reservoir with which all phases can exchange and (2)that cooling begins from above the closure temperatureof all minerals. The model of Jenkin et al. (1994) incor-porates the effects ofa finite reservoir and the conditionthat peak temperature may equal the closure temperatureof any phase. The calculations for the Giletti model weremade \ilith the computer program Cool (Jenkin et al.,l99l), and the more complex modeling following Jenkinet al. (1994) was carried out by G.R.T. Jenkin at theScottish Universities Research and Reactor Centre, Scot-land. The D'8O value of each phase calculated by eachmethod, agrees to within 0.17o.

The degree ofretrograde diffusional exchange was cal-culated with both anhydrous and hydrous diffi,rsion-ratedata. Of the phases considered in this study, anhydrousand hydrous experimental diffusion data have been ob-tained only for anorthite (Table 2). For the phases qtaftzand magnetite, the preexponential factor for O diftrsionis two orders of magnitude larger under hydrous condi-tions (Table 2), and it was assumed that a similar effectexists for hematite (isostructural with ilmenite). Castleand Surman (1969) and Schmalzried and Wagner (1962)presented the experimental and theoretical basis for thisassumption. The diffusion rate of O in hornblende wasassumed to be independent of fluid composition. The clo-sure temperature for garnet is above 700 oC, regardless ofthe composition of the fluid. The d'8O values calculated

with the hydrous diffusion data are in excellent agreementwith the measured values (Table 3). Only the plagioclasematrix value differs from the measured value. The cal-culated matrix value is 9.15V*, compared with the aver-age measured value of 8.657m. Several D'8O values of thematrix plagioclase are high (no. 23,9.2Vu;r.o.28,9Vn\,and these may be more representative of the true matrixvalues; the lower measurements could be a m6lange ofporphyroclasts and matrix. The agreement between thecalculated and measured 6'80 values for the dry systemis not nearly as good (Fig. 5), suggesting that cooling oc-curred in the presenca ofa hydrous fluid. This conclusionis supported by the lack ofisotopic zoning in the feldsparporphyroclasts. Ifthe diffusion rate ofO in feldspar wasas slow as predicted by dry diffusion experiments (El-phick et al., 1988), then isotopic zoning acquired duringretrogression should be preserved in the feldspar porphy-roclasts. The agreement between the diftision calcula-tions using hydrous diftrsion data and the measured val-ues supports the conclusion that a mixed COr-HrO fluidphaso was present at peak metamorphic conditions(Moecher and Essene, 199 1) and persisted down to lowertemperatures.

The infinite reservoir model (Giletti, 1986) is particu-larly well suited to the present rock type if the rapid wetdiftrsion data for feldspar are applicable. Plagioclasemakes up 70 modalo/o of the rock and has a very rapid Odiffirsivity under HrO-rich conditions (Giletti et al., 1978).For a cooling rate of l-10 "C/m.y. and time scales of 0.1-l0 m.y., the scale of O diftrsion at 700 "C is on the orderof tens of centimeters, easily allowing the plagioclase tohomogenize isotopically. The assumption that plagro-clase served as an infinite reservoir is therefore appro-priate for this sample. The more rigorous treatment ofJenkin et al. (1994) involves none of the assumptions

0.0 l.o 2.o 3.0 4.0Temperature coefficient of fractionation (a)

Fig. 5. Isotherm plot for the Grenville meta-anorthosite.Measured and predicted [A'EOour_-*a ) - b] values vs. the tem-perature coefrcient of fractionation , a. The relationship is givenby Eq. L The line defined by the locus ofpoints should be pro-portional to ?F-, (Javoy et d., 1970). The best-fit line to themeasured data corresponds to 470 oC, far lower than the peaktemperatures of metamorphism. The low isotherm temperaturecan be explained by retrogade diffusional exchange. The pre-dicted (A'tO - D) values for the slowly cooled rock (3 "C/m.y.)were calculated using the method outlined by Giletti (1986) basedon the closure temperature concept ofDodson (1973). Calculat-ed 6'EO values were determined with both the hydrous (wet) andanhydrous (dry) difrrsion coefrcients (Table 2). The measuredand hydrous calculated 6ttO values agree well except for ilmen-ite. Values for the a and b coefrcients for all calculations arefrom Bottinga and Javoy (1975).

inherent in the closure temperature concept (Dodson,1973) as used by Giletti (1986). Even so, the calculateddltO values by Jenkin are within 0.17- of the values ob-tained with the Giletti model, illustrating that the far sim-pler Giletti model may be valid as long as appropriaterock types are chosen. Rock types in which the modallydominant phase has a very rapid diftrsion rate, such asfeldspar (this study) or calcite (Sharp and Jenkin, 1994),can be modeled accurately with the simple approach ofGiletri (1986).

Although the isotherm derived from a best fit of all thedata has no geologic meaning, peak temperatures ofmetamorphism can still be extracted from this data setwhen diftrsional effects are considered. The calculatedchange in the 6'80 value of the modally abundant feldsparporphyroclasts is only 0.l7oo during cooling, and the D'8Ovalue ofthe garnet did not change during retrogression.Therefore, this mineral pair should yield nearly peak-metamorphic temperatures. The AltOo*@ru"._r"r valueof l.7Vu corresponds to a temperature of 660 "C, indis-tinguishable from peak temperatures within error. Theconclusion that most mineral triplets have no meaning(Deines, 1977) must be reviewed in light of known dif-fusional effects-when appropriate mineral pairs are cho-sen, peak temperatures can be extracted. When measuredstable isotope data are evaluated in terms of expected,retrograde, difusional-exchange effects, apparent dis-equilibrium effects can be explained and information such

957

TaBLE 3. Calculated and measured D18O values from high-grademetamorphic terranes

Graindiam.

Mode (mm)

Calculated final 6180value/closure

temp€rature ("C)

Hydrous Anhydrous

Grenville mela-anorthosite (present study)Pl porph An@ 0.5 4.0 8.4 265 8.2 695 8.4Pl gm An4 0.2 0.3 9.1 175 8.8 560 9.1-Garnet 0.05 0.2 6.9 >700 6.9 >700 6.9Hornblende 0.2 0.3 6.0 47O 6.5 470 6.1llmenite 0.05 0.3 3.4 595 4.1 665 1.9

Wind River iron formation-T2-h (Sharp et al., 198E)Quartz 0.26 0.5 11.2 455 9.9 500 9.9Garnet 0.03 0.5 6.8 775 6.6 775 6.7Opx 0.60 0.5 6.8 660 6.8 >800 6.9Magnetite 0.11 0.5 0.7 455 2.3 585 1.9

Mafic glanulite-EO1 1 (McNaughton and Wilson, 1980);diameters approx.

0.01 0.8 8.9 465 8.5 525 8.40.35 0.8 5.5 215 5.4 620 5.4

Cox 0.10 0.8 4.3 690 4.3 . >730 4.1Oox 0.18 0.8 4.3 690 4.3 >730 4.4Hornblende 0.34 0.8 3.8 535 4.0 535 4.2

Orlhogneiss, Sri Lanka-BSL-88-46-4 (Hotfbauer et al., 1993)

SHARP AND MOECHER: O ISOTOPES IN META-ANORTHOSITE

6'60vatuemea-sureo

,a| 5.0

O

T r obo

-

I r o

OuartzPl Ane

QuartzKsparPl AnsGarnet

o.24 0.4 12.9 445 13.0 525 12.80.33 0.4 11.4 200 11.1 585 10.30,10 0.2 8.9 170 10.5 550 n.d.0.07 0.4 9.0 760 9.0 760 9.0

Hornblende 0.27 0.4 7.8 500 8.4 500 8.0llmenite 0.001 0.2 5.1 590 6.0 660 4.9

,Vote.' abbreviations are defined as tollows: porph : porphyroclasts;gm : groundmass; Cpx : clinopyroxene; Kspar: potassium feldspar;Opx : orthopyroxene; Pl : plagioclase; n.d. : not determined.

* Assumed equal to the upper range of measured values.

as the presence or absence ofa fluid can be deduced (e.9.,Eiler et al.,1993; Farquhar et al., 1993).

The apparent temperatur€ of 470 "C obtained with thegraphical method (Javoy et al., 1970) can be explained interms of retrograde diffirsional exchange. To determinewhether the near-linear locus of points is unique to theWhitestone anorthosite or is expected for most slowlycooled granulite terranes, calculated and measured 0'tOvalues were compared from other granulite-facies ter-ranes. With these data, we were able to evaluate (l)whether the false isotherm generated with the present dataset is a common feature in other rock types and (2) whetherthe calculated and measured 6rto values are best fit byhydrous or anhydrous diftrsion data. Three additionalgranulite-facies samples were evaluated-a banded ironformation, a mafic granulite, and an orthogneiss (Table3, Fig. 6A-6C). The iron formation cooled under anhy-drous conditions (Sharp et al., 1988). The measured 6180values for this sample agre€ very well with the dry dif-fusion data, and all phases plotted on an apparent or falseisotherm of -600 oC, which is not geologically signifi-cant. The calculated d'8O values for the mafic granuliteare very similar for both the hydrous and anhydrous cal-culations. The similarity can be explained in terms of thebulk composition and relative insensitivity to fluid com-position for the phases pyroxene and hornblende. Theanhydrous calculations most closely match the measuredd18O values. The granulite probably cooled under anhy-

0.0..0.0

958 SHARP AND MOECHER: O ISOTOPES IN META-ANORTHOSITE

A. Iron Formation

1.0 2n 3.0 4.0 5.0Temperature coefficient of fractionation (a)

C. Orthogneiss

0.0 1.0 2-o 3.0 4J 5.0Temperature coefficient of fractionation (a)

Frg. 6. Isotherm plot illustrating measured and predicted d'EOvalues for a granulite-facies banded iron formation (A) (Sharpet al., 1988), a granulite-facies mafic rock (B) (McNaughton andWilson, 1980), and a granulite-facies orthogneiss (C) (Hoffbaueret al., 1993). Calculated 6'80 values for anhydrous and hydrousdiftrsion coefrcients for a cooling rate of 5 "C/m.y. The best-fitline to the measured data is given, along with the correspondingtemperatures. Although the locus of points may closely define aline, the temperature derived from the slope does not have anygeologic significance.

drous conditions, or the pyroxenes would have retro-gressed to amphiboles. Similar to the meta-anorthositeand iron formation, the data define a false isotherm-onethat has no geological significance. The orthogneiss sam-

ple is the most ambiguous. The calculated hydrous sam-ples fit the measured d18O values except for the potassiumfeldspar. The discrepancy can be explained if the ana-lyzed feldspars were a mixture of potassium feldspar andplagioclase (Fig. 6C). This sample may have undergoneopen-system retrograde exchange (Hoffbauer et al., 1993),explaining the poor fit between calculated and measuredvalues for feldspar. Except for the feldspar data, the mea-sured and predicted 6'80 values define a false isotherm.

O isotope temperature estimates are often below thepeak temperatures of metamorphism. The isothermmethod based on three or more minerals is equivocal forestablishing equilibrium. As illustrated in Figures 5 and6, apparent isotherms can develop from diffusional ex-change and are not related to a recogtized geological event.Only in cases where the mineral samples are extremelyrefractory and coarse-grained, or have cooled very rap-idly, will isotherms provide peak temperatures. Carefullyselected mineral pairs, such as plagioclase * garnet in thepresent study, may provide peak temperatures, and theoverall isotopic compositions of a mineral suite may beused to constrain the fluid composition of a rock duringcooling.

AcxNowr,BocMENTS

We thank G.R.T. Jenkin (S.U.R.R.C.) for his computational assistanceand advice. The reviews of B. Giletti, R. Joesten, and especially D. Henryimproved the quality of this manuscript. J.C. Hunziker is thanked for hiscontinued advice and support at the University of Iausanne Stable Iso-tope Laboratory

Rprnnnxcrs crrEDBottinga, Y., and Javoy, M. (1973) Comments on oxygen isotopic geo-

thermometry. Earth and Planetary Science Irtters, 20,250-265.- (1975) Oxygen isotope partitioning among the minerals in igneous

and metamorphic rocks. Reviews ofGeophysics and Space Physics, I 3,40 l-4 I 8.

Castle, J.E., and Surman, P.L. (1969) The self-difrrsion of oxygen in mag-netite. The effect of anion vacancy concentrations and cation-distri-bution. Journal of Physical Chemistry, 7 3, 632-634.

Chamberlain, C.P., and Conrad, M.E. (1991) Oxygen isotope zoning ingamet. Science, 254, 403-406.

Clayton, R.N., and Epstein, S. (1958) The relationship between '80/'60

ratios in coexisting quartz, carbonate, and iron oxides from vanousgeological deposits. Journal of Geology, 66, 352-37 1.

-(1961) The use of oxygen isotopes in high-temperature geologicalthermometry. Journal of Geolory, 68, 447 -452.

Connolly, C., and Muehlenbachs, K. (1988) Contrasting oxygen diftrsionin nepheline, diopside and other silicates and their relevance to isotopicsystematics in meteorites. Geochimica et Cosmochimica Acta, 52, 1585-I 5 9 1 .

Conrad, M.E., and Chamberlain, C.P. (1992) Laser-based in situ mea-surements of fine-scale variations in the d'!O values of hydrothermalquartz. Geology, 20, 812-8 16.

Cosca, M.A., Sutter, J.F., and Essene, E.J. (1991) Cooling and inferreduplifuerosion history of the Grenville Orogen, Ontario: Constraintsfrom aoArl3eAr thermochronology. Tectonics, I0, 9 59 -97 7.

Crank, J. (1975) The mathematics ofdifnrsion (2nd edition),414 p. Clar-endon, Oxford, U K.

Deines, P. (l 977) On the oxygen isotope distribution among mineral trip-lets in igneous rocks. Geochimica et Cosmochimica Acta, 41, 1709-t730.

Dodson, M.H. (1973) Closure temperature in cooling geochronologicaland petrologic systems. Contributions to Mineralogy and Petrology, 40,259-27 4.

.o

E 4 0

N

x 2.0

k

o 4.0

g 3.0

< 1.0

Eiler, J.M., Baumgartner, L.P., and Valley, J.W. (1992) Intercrystallinestable isotope diffrsion: A fast grain boundary model. Contributions toMineralogy and Petrology, ll2, 543-557.

Eiler, J.M., Valley, J.W., and Baumgartner, L.P. (1993) A new look atstable isotope thermornetry. Geochimica et Cosmochimica Acta, 57,257 L-2583.

Elphick, S.C., Graham, C.M., and Dennis, P.F. (1988) An ion microprobestudy of anhydrous oxygen diffirsion in anorthite: A comparison withhydrothermal data and some geological irnplications. Contributions toMineralogy and Petrology, 100, 490-495.

Elsenheimer, D., and Valley, J.W. (1992) In situ oxygen isotope analysisof feldspar and quartz by Nd:YAG laser microprobe. Chemical Geol-oW,101,21-42.

- (I993) Subrnillimeter scale zonation ofd'EO in quartz and feldspar,Isle of Skye, Scotland. Geochimica et Cosmochimica Acta, 57 , 3669-3676.

Farquhar, J., Chacko, T., and Frost, B.R. (1993) Strategies for high-tem-perature oxygen isotope thermometry: A worked example from theLaramie Anorthosite complex, Wyoming, USA. Earth and PlanetaryScience Irtters. ll7. 407-422.

Farver, J.R. (1989) Oxygen self-diftrsion rates in diopside with applica-tion to cooling rate determinations. Earth and Planetary Science I-et-ters, 92, 386-396.

Farver, J.R., and Giletti, B.J. (1985) Oxygen diffirsion in amphiboles.Geochimica et Cosmochimic a Actz, 49, 1 403- 1 4 1 t.

Fortier, S.M., and Giletti, B.J. (1989) An empirical model for predictingdifrrsion coeffi cients in silicate rninerals. Science. 245. l 48 l - 1 484.

Giletti, B.J. (1986) Diffirsion effecrs on oxygen isotope temperatures ofslowly cooled igneous and metamorphic rocks. Earth and PlanetarySciencc Irtters. 77. 218-228.

Giletti, B.J., and Hess, K.C. (1988) Oxygen diffirsion in magnerire. Earthand Planetary Science [.€tters, 89,115-122.

Giletti, 8.J., and Yund, R.A. (1984) Oxygen diftrsion in quartz. Journalof Geophysical Research, 89, 4039-4046.

Giletti, B.J., Semet, M.P., and Yund, R.A. (1978) Studies in diftrsion.III. Oxygen in feldspars, an ion microprobe determination. Geochimicaet Cosmochimica Acta.42. 45-57.

Hoftbauer, R., Hoernes, S., and Fiorentini, E. (l 993) Oxygen isotope ther-mometry based on a refined increment method and its application togranulite-grade rocks from Sri Lanka. Precambrian Research,66,199-) ) i

Javoy, M. (1977) Stable isotopes and geothermometry. Journal of theGeological Society oflondon, 133, 609-636.

Javoy, M., Fourcade, S., and Allegre, C.J. (1970) Graphical merhod ofexamination ofIEO/'6O fractionations in silicate rocks. Earth and Plan-etary Science ktters, 10, 12-16.

Jenkin, G.R.T., Fallic, A.E., Farrow, C.M., and Bowes, G.E. (1991) COOL:A Fortran-77 computer program for modeling stable isotopes in coolingclosed systems. Computers and Geoscience, 17, l9l-412.

Jenkin, G.R.T., Farrow, C.M., Fallic, A.8., and Hig€:ins, D. (1994) Oxy-gen isotope exchange and closure temperatures in cooling rocks. Jour-nal of Metamorphic Petrology, 12, 221-215.

Kirschner, D.L., Sharp,2.D., and Teyssier, C. (1993) Vein growth mech-anisms and fluid sources revealed by oxygen isotope laser rnicroprobe.Geology, 21, 85-88.

lamb, W.M., and Moecher, D.P. (1992) CO:-rich fluid inclusions in theWhitestone anorthosite: Implications for the retrograde history of theParry Sound ShearZone, Grenville Province, Canada. Journal ofMeta-morphic Petrology, lO, 7 63-77 6.

959

McNaughton, N.J., and Wilson, A.F. (1980) Problems in oxygen isotopegeothermometry in mafic granulite facies rocks from near Einsaleigh,northern Queensland. Precarnbrian Research, 13, 77 -86.

Mezger, K., Essene, E.J., van der Pluijm, B.S., Halliday, A.N. (1993)U-Pb geochronology of the Grenville Orogen of Ontario and New York:Constraints on ancient crustal tectonics. Contributions to Mineralogyand Petrology, 114, 13-26.

Moecher, D.P., and Essene, E.J. (1991) Calculation of CO, activities usingscapolite equilibria: Constraints on the presence and composition ofafluid phase during high grade metamorphisrn. Contributions to Min-eralogy and Petrology, 108, 219-240.

Moecher, D.P., Anovitz, L.M., and Essene, E.J. (1988) Calculation andapplication of clinopyroxene-plagioclase-garnet-quartz geobarometers.Contributions to Mineralogy and Petrology, 100, 92-106.

Moecher, D.P., Essene, E.J., and Valley, J.W. (1992) Stable isotope andpetrological constraints on scapolitization of the Whitestone meta-an-orthosite, Grenville Province, Ontario. Journal of Metamorphic Pe-trology, 10,745-762.

O'Neil, J.R., and Taylor, H.P., Jr. (1967) The oxygen isotope and cationexchange chemistry of feldspars. American Mineralogist, 52, l4l4-1437.

O'Neil, J.R., Masuda, H., and Sharp,Z.D.(1992)Oxygen isotope analysesof microsamples of silicates and oxides. Program and Abstracts, V.M.Goldschmidt Conference May 8-10, 1992, A,-79.

Reddy, K.P.R., and Cooper, A.R. (1983) Oxygen diffirsion in MgO anda-Fe2Or. Journal ofthe American Ceramic Society, 66, 664-666.

Schmalzried, H., and Wagner, C. (1962) Fehlordnung in ternaren lonen-kristallen. Zeitschrift fiir physikalische Chemie neue Folge, 31, 198-22 t .

Sharp, Z.D. (1990) A laser based microanalytical method for the in situdetermination of oxygen isotope ratios in silicates and oxides. Geo-chimica et Cosmochimica l\ct^,54, 1353-1357.

-(1991) Determination of oxygen difiirsion rates in magnetite fromnatural isotopic variations. Geology, 19, 653-656.

- (1992) In situ laser microprobe techniques for stable isotope anal-yses. Isotope Geoscience, 101, 3-19.

Sharp, 2.D., and Jenkin, G.R.T. (1994) An empirical estimate of thediffi.rsion rate of oxygen in diopside. Joumal of Metamorphic Geology,12, 89-97.

Sharp, 2.D., O'Neil, J.R., and Essene, E.J. (1988) Oxygen isotope varia-tions in a granulite-grade iron formation: Constraints on oxygen dif-fusion and retrograde isotopic exchange. Contributions to Mineralogyand Petrology, 98, 490-501

Sharp, 2.D., Giletti, B.J., and Yoder, H.S., Jr. (1991) Oxygen diffirsionrates in quartz exchanged with COr. Earth and Planetary Science Iet-ters. 107. 339-348.

Tucillo, M.E., Mezger, K., Essene, E.J., and van der Pluijm, B.A. (1992)Thermobarometry, geochronology, and the interpretation of P-I-, datain the Britt Domain, Ontario Grenville Orogen, Canada. Journal ofPetrology, 33, 1225- 1259.

van Breemen, O., Davidson, A., Loveridge, W.D., and Sullivan, R.W.(1986) U-Pb geochronology ofGrenville tectonites, granulites and ig-neous precursors, Parry Sound, Ontario. In J.M. Moore, A. Davidson,and A.J. Baer, Eds., The Grenville Province. Geological Association ofCanada Special Paper, 31, l9l-207.

M,cNUscRrPT RrcEwED Jr:Lv 22, 1993MeNuscrupr AcCEPTED Mw 27,1994

SHARP AND MOECHER: O ISOTOPES IN META-ANORTHOSITE