American Mineralogist, Volume 78, pages 988-1001, 1993
O isotope zoning in garnet and staurolite: Evidencefor closed-systemmineral growth during regional metamorphism M,llrrrBw J. KonN, JonN W. Yar-r-ny,DoN Er-snrunnrvrnno* Mrcruer, J. Srltcuzzt Department of Geology and Geophysics,University of Wisconsin-Madison, Madison, Wisconsin 53706, U.S.A.
Ansrn,rcr Microanalysis of O isotope ratios has been made in traversesacrosstwo garnet crystals and one staurolite sample (all - I cm in diameter) from a metapelitic schist and across one garnet sample -4.5 mm in diameter from an amphibolite from the Cordillera Darwin metamorphic complex, Tierra del Fuego, Chile. The metapelitic sample contains garnet + biotite * staurolite * plagioclase * muscovite + quarlz + ilmenite + chlorite, and P-Z conditions determined from mineral rims were approximately 575 'C and 7 kbar. O isotope analyseswere collected with a spatial resolution of 0.5-1 mm using a laser extraction system. The 72 analysesof the two garnet crystals in the metapelite indicate a gradualincreasein 6'80 from -3.5Vm(V-SMOW) in the core to -4.0Vu at the rims. The 70 analysesofthe staurolite grain show that it is generallyunzoned, with an averagevalue of 4.31 + 0.087- (1o),exceptfor localizedincreasesin d'8Otoward someparts of the rim; theseincreasesmay be the result of late chlorite alteration. The garnet amphibolite sample contains garnet + biotite + plagioclase+ quartz + chlorite + hornblende + cummingtonite, and P-Z conditions were approximately625'C and 4.5 kbar. The 25 analysesof the garnet sample indicate that it is homogeneous,with an averageisotopic value of 3.44 + 0.05Vm(lo), exceptfor three analysesthat
[email protected] analysesof the finegrained matrix near the garnet indicate that the matrix is compositionally neady homogeneous,with an averagevalue of 3.70 + 0.15V*(lo). With this P-T-t palh, the reaction history, and no infiltration by a fluid with disequilibrium d'8O, the predicted prograde zonation for the metapelitic garnets is between *0.2 and +0.75Vu (rim to core), with a best estimate of +0.57o. The zonation for staurolite is --0.05Va, and for the amphibolitic garnet it is between -0.01 and +0.06Vm.The correspondenceofthe predictions and observationssuggeststhat the rocks were not infiltrated by a fluid that is not in isotopic equilibrium during regional metamorphism. The data show that regional metamorphic garnet and staurolite can preserveprogradevariations of O isotopes in amphibolite faciesterranes. IurnonucrroN A principal goal of metamorphic petrologists is to determine the degreeto which rocks are open or closed to infiltration by a fluid that is out of equilibrium with the rock. Studies demonstrate the power of O isotope analysis for interpreting such metamorphic processes(e.g.,see reviewsby Rumble, 1982;Valley, 1986;Nabelek, 1991). Prograde O isotope zoning in metamorphic porphyroblasts(e.g.,Garlick and Epstein, 1967;Chamberlainand Conrad, 1991; Kohn, 1993) is a potentially sensitive monitor of fluid evolution becausethe changesof mineral d'8Ofor regional metamorphic systemsthat are closed to fluid infiltration are readily predictable for a given bulk composition,assemblage, and P-Z history (Kohn, 1993). With analytically and spatially precise analysis of O isotopes, a direct comparison of theory and observation is possible for rocks whose P-Z and reaction histories are well understood. * Presentaddress:Geological Survey ofJapan, l-l-3 Higashi, Tsukuba, Ibaraki, 305 Japan. 0003-004x/93l09 l 0-0988$02.00
We present data on the zoning of O isotopes from regional metamorphicgarnetand stauroliteporphyroblasts from Cordillera Darwin, Tierra del Fuego,Chile. These results are unique in that (l) no previous study of the zoning of O isotopes has been conducted on regional metamorphic rocks whose P-T-t and reaction histories are as well known, and (2) no previousstudy hasobtained such good analytical precision (99o/oinclusion-free, and reach - 5 mm in diameter. Books of biotite and chlorite are typically -0.25 mm wide and 0.05 mm thick. Plagioclase and quartz are equant and xenoblasticand average-0.05 mm in diameter; individual plagioclasegrains are zoned from Anro in their cores to Anu, at their rims (Kohn et al., 1993). Hornblende and cummingtonite are typically -0.5 mm long and -0.075 mm in diarneterand are compositionally homogeneous. ANal,yrrc,l.r, TECHNTeUES Sample preparation and data collection Sampleswere prepared using the thin sawblade technique (Elsenheimer and Valley, 1993). Rocks were first slabbed,and garnet and staurolite crystals that were particularly inclusion-freewere identified. Standardthin-section chips containing the crystalsand surrounding matrix material were then cut, and the chips were polished and mounted on a glassslide with acetone-soluble,quick-setting cement. A wafer 500-750 pm thick of garnet, staurolite, and matrix material was cut from each chip, the cut surfaceswere lightly polished, and glass coverslips were affixed with the same cement. Each crystal was cut into strips 500 pm wide using a slow-speedultrathin sawblade (kerfs 100-200 pm thick). Becauseoffractures, the strips would commonly break apart after the cement was dissolved, but reconstruction was readily possible with photo documentation. Prior to analysis, each strip was broken into piecesbetween 300 and 1000 s.m in length and inspectedfor inclusions and alteration. Each sample piecewas then loaded onto a solid block of Ni with holes 2 mm in diameter drilled into it, along with our laboratory standard Gore Mountain garnet. Crowe et al. (1990)and Elsenheimerand Yalley (1992, 1993)have describedin detail the laser extraction system and proceduresused at the University of Wisconsin for O isotope analysis. They are similar to the system and procedure describedby Sharp (1990, 1992). After preparation, sampleswere pretreatedovernight in the sample chamberwith 1000 pmol of BrF, (-70 torr) to remove HrO and any tracesof organic compounds. Although this pretreatment can leach O from altered feldspars (Elsenheimer and Valley, 1992, 1993); our additional studies indicate that it doesnot affect the analysisofgarnet, staurolite, quartz, biotite, muscovite, unaltered plagioclase, magnetite, ilmenite, diopside, wollastonite, hornblende, corundum, or zircon. Most sampleswere heated with a defocusedCO, laser beam -500 pm in diameter in the presenceof - 1000 pmol of BrF, from an initial power density of 5 x 106W,/m2to a final power density of 4.5
x 107Wm2. Three analysesof quartz were made at a singlepower density of 4.5 x l0? Wm'z. Samplescompletely reacted within l-2 min. Sample sizes were typically in the range5-20 pmol of CO, (0.5-2 mg). An intralaboratory standard (UW Gore Mountain garnet) was analyzed each day prior to analyzing unknowns to correct for any O fractionation within the extraction system.The 0'8O value for our standard was determined relative to the international standardsNBS-28 quartz and NBS-30 biotite, as describedbelow. A memory effect occurred in the extraction system prior to September l, 1992;suchthat analyseswerebiasedby the 6'80 value of the preceding analysis. The amount of gas required to overcomethe memory of a samplethat differed by - l4Vm was found to be - 10-20 pmol. Becausethe differences between samplesand the garnet standard were 0.5 and 0.2V* would be readily recognized in garnetsfrom the metapelite and amphibolite, respectively.Iffluid-rock ratios are consideredon the basis of a simple zero-dimensional model of fluid flow, and if the fluid is assumedtobe lV* out of equilibrium, then a fluid-rock molar ratio of 1.0 producesa shift of 0.5Vm. If the fluid is l}Vm out of equilibrium then the sameshift is producedwith a fluid-rock molar ratio of only 0.1. If the model of Dipple and Ferry (1992) is used,and a length scaleof I km for the flow path is assumed(basedon the distance to the nearest major lithologic contact), then fluxeson the order of 5 x 103moles of HrO per square centimeter of rock would produce shifts of 0.5Vmor greater and would be detectable. The close correspondencebetween the models and measurementsgives us confidence that fluid-rock interactions during amphibolite facies metamorphism can be quantitatively evaluated in the future becausedeviations ofisotopic zoning from closed-systempredictions will be identifiable and interpretable. This study demonstrates the utility of the microanalytical isotopic approach for interpretingmetamorphicprocesses.
KOHN ET AL.: O ISOTOPE ZONING IN GARNET AND STAUROLITE
l00l
Kohn, M.J., Spear,F S.,and Dalziel,I W.D (1991)Rapid coolingfollowing exhumation in the Cordillera Darwin metamorphic complex, Tierra J. Eiler and L. Baumgartnerare thanked for valuable discussions,and del Fuego, Chile. Geological Society of America Abstracts with ProB. Dutrow, J. Grambling, R. Gregory, and T. Labotka are thanked for grams2 , 3,A134. their reviews. This work was funded by NSF grant EAR-91-05709 to -(1993) Metamorphic P-7 paths from Cordillera Darwin, a core J W V. and an NSF postdoctoralfellowship to M.J.K. complex in Tierra del Fuego,Chile. Journal ofPetrology, 34, in press. Matthews,A., Goldsmith,J R, and Clayton,R.N (1983)Oxygenisotope RnrnnnNcrs cITED fractionation betweenzoisite and water. Geochimica et Cosmochimica Acta,47, 645-654. Bottinga, Y., and Javoy, M. (1975) Oxygen isotope partitioning among Nabelek, P.L. (1991) Stableisotope monitors. In Mineralogical Societyof the minerals in igneousand metamorphic rocks. Reviews in GeophysAmerica Reviewsin Mineralogy,26,395-435. ics and SpacePhysics,13, 401-418 Nelson,E.P.,Dalziel,I W.D., and Milnes,A.G. (1980)Structuralgeology Chamberlain,C.P, and Conrad, M.E. (1991)Oxygenisotopezoning in ofthe Cordillera Darwin: Collisional-style orogenesisin the southerngarnet. Science,254, 403-406 most Chilean Andes. EclogaeGeologicaeHelvetiae, 73, 727-7 51. Clayton, R.N, O'Neil, J.R., and Mayeda, T.K. (1972) Oxygenisotope O'Neil, J R, and Taylor, H.P., Jr. (1967)The oxygenisotopeand cation exchangebetweenquartz and water. Journal of GeophysicalResearch, exchangechemistry of feldspars American Mineralogist, 52, 141477,3057-3067. t437. Conrad,M, and Chamberlain,C.P. (1992)Iaser-based,in situ measure(1969) Oxygenisotope equilibrium betweenmuscovite and water. mentsof fine-scalevariations in the 6'80 valuesof hydrothermalquartz. Joumal of GeophysicalResearch,74, 6012-6022. Geology,20,812-816. Richter, R., and Hoemes, S (1988) The application of the incrernent Crowe,D.E., Valley,J.W., and Baker,K (1990)Microanalysisof sulfurmethod in comparison with experimentally derived and calculated isotope ratios and zonation by laser microprobe. Geochimica et CosO-isotopefractionations.Chemieder Erde,48, l-18. mochimica Acta. 54. 2O75-2092 (1982) Stable isotope fractionation during metamorphic Dalziel, I.W.D., and Brown, R. (1989)Andean core complex evolution Rumble, D, III devolatilization reactions. In Mineralogical Society of America Rerelatedto margtnalbasin collapse:Implications for Cordillera tectonics. views in Mineralogy,lO,327-353. Geology, 1'l, 699-703 Sharp, Z. (1990) A laser-basedmicroanalytical method for Ihe in situ Dipple, G.M., and Ferry, J.M (1992) Fluid flow and stable isotopic aldetermination of oxygen isotope ratios of silicates and oxides. Geoteration in rocks at elevated temperatureswith applications to metachimicaet CosmochimicaActa. 54. 1353-1357 morphism.Geochimicaet CosmochimicaActa, 56, 3539-3550. (1992) In situ laser microprobe techniquesfor stableisotope analEiler, J.M, Baumgartner,LP, and Valley, J.W. (1992) Intercrystalline ysis.ChemicalGeology,l0l, 3-19. stableisotope diffusion: A fast grain boundary model Contributions to Shieh,Y.-N., and Schwarz,H.P. (1974)Oxygenisotopestudiesof granite Mineralogy and Petrology, | 12, 543-557. and migmatite, Grenville province of Ontario, Canada.Geochimica et Elsenheimer,D., and Valley, J W. (1992) In situ oxygen isotope analysis CosmochimicaActa,38, 2l-45. of feldspar and quartz by Nd:YAG laser microprobe. Chemical GeolY.-N., and Taylor, H P, Jr. (1969)Oxygenand hydrogenisotope Shieh, o g y ,l 0 l , 2 l - 4 2 . studies of contact metamorphism in the Santa Rosa Range, Nevada -(1993) Sub-millimeterscalezonationof D'8Oin quartz and feldand other areas Contributions of Mineralogy and Petrology, 20, 306spar, Isle of Skye, Scotland.Geochimica et CosmochimicaActa, 57, in 35 6 . press. Spear, F.S. (1989) Petrologic determination of metamorphic pressure(1989) Fortier, S.M., and Giletti, B J. An empirical model for predicting temperature-timepaths In F.S. Spear and S.M Peacock,Eds , Metadiffusioncoefficients in silicateminerals.Science,245, l48l-1484. morphic pressure-temperature-timepaths, short course in geology,p. Garlick, G.D., and Epstein,S (1967) Oxygen isotope ratios in coexisting l-55. American GeophysicalUnion, Washington, DC. minerals of regionally metamorphosedrocks. Geochimica et CosmoSpear, F.S., Kohn, M.J., Florence,F., and Menard, T. (1990)A model c h i m i c aA c r a .3 1 . l 8 l - 2 1 4 . for garnet and plagioclasegrowth in pelitic schists: Implications for Kirschner,D.L., Sharp,2.D., and Teyssier,C. (1993)Vein growth mechthermobarometry and P-f path determinations. Journal of Metamoranismsand fluid sourcesrevealedby oxygen isotope laser microprobe phic Geology,8, 683-696. G e o l o g y2, 1 , 8 5 - 8 8 . Valley, J W (1986) Stable isotope geochemistryof metamorphic rocks. (1991) Kohn, M.J. Studiesof the metamorphismin west-centralNew In Mineralogical Society of America Reviews in Mineralogy, 16,445Hampshire, U.S.A. and Tierra del Fuego, Chile. Ph.D thesis, Rens489. selaerPolytechnic Institute, Troy, New York. Wickham, S.M., and Taylor, H.P., Jr (1985)Stableisotopicevidencefor -(1993) Modeling of progrademineral 6'80 changesin metamorlarge-scaleseawaterinfiltration in a regional metamorphic terrane.The phic systems.Contributions to Mineralogy and Petrology, ll3,249Trois SeigneursMassif, Pyrenees,France Contributions to Mineralogy 261. and Petrology,9l, 122-137. Kohn, M.J., and Valley, J W (1993a)Disparatepatternsof oxygenisotope zonation in gamet: All products of closedsystemmetamorphism. EuropeanUnion ofGeosciences,Terra Abstracts, 5, 373. -(1993b) High-Z fluids in the Fall Mountain nappe, southwestern M,lNuscnryr rucEIvEDNovertlnrn 4, 1992 Mx 26, 1993 New Hampshire: Infiltration vs. anatexis(abs.).Eos, 74, 332. Mnr'r.rscnrsrAcCEPTED
AcrNowr,nocMENTS
Note added in proof More recent analysis of standardsallows a refinement ofcorrection proceduresfor laser probe analysis.A total of 371 analysesof the UW Gore Mountain garnetstandard was made by four analysts between September l, 1992, and July 15, 1993;the uncorected value of d'8Ois 6.03 + 0.l3qn (ls), virtually identical to the value reported here of 5.99 + 0.07V- for 79 analyses.The correction of these raw data to the SMOW scaleis basedon comparisonwith l7 laseranalysesof NBS28, sevenfrom
before cleaningof the extraction line in September,1992. Since September, the day-to-day variations have been typically
