RENDICONTl DELLA SOCIETA ITALlANA Dl MINERALOGIA E PETROLOGIA, 1988. VoL ~l. pp. 6~·82

Geothermobarometry and fluid inclusionsin a rock from the Doddabetta charnockite complex, Southwest India

}ACQUES L.R. TOURETInslitme of Earth Scie~, Fttt Univeuit}", de Boddaan 108:i, 1081 HV Amsterdam, 1be Nelherlands

THOR H. HANSTEENMineralogisk-Geologisk. Museum, SITS gate I, N.Q:i62 Ch!o :i, Norway

ABsnAcr. - A de:formed, garl'lC't and. OI'thopyroxenc.bearing c::harnockite from Oodda Bella near Ooty, NiIgiriHills India, ronlains many COl Auid inclusions in 3major lTIinB'als fsilmC'l, plagiodue and quutz). Iktailcdstudy of the mincra1 phases pl«'CS PT peak lI'lC"Iamorphicrondilions al 8 ± 0,7 kb and 800 ± l:i°C, respectively.1be rock is conspicuously deformed .nd recrystalliud(quartz more than p1l11ioclue), whereas plnet hasresisled to the deformation. 1be densities of the COinclusions in each mineral are relatively uniform, ~they increase in order of increasing deformation from0,96 - 1,02 s/cmJ in garnet to 1,12 s/cm) in quartz.Primary indusioru in garTlC't COfrespooo to a pressureabout 1 kbat lower thal peak metamorphic conditions.Higher densities in p1agiodase and quartz indicatc a postmctamorphic PoT-trajectOry concave towards T axis(.. isobaric cooling_). Geodynamic implications and fluidregime after metamorphi$ltl (important quantity of freeCOl during mineral growth) are discussed.

Introduction

The Archean craton of Southern Indiaconsist of a southern granulite facies and anorthern greenschist-amphibolite fadesterrane, connected by a narrow transitionalzone from Mangalore in the west, throughMysore and Bangalore to Madras in the east(DRURY and Holt 1980> (Fig. 1). The granulitefades metamorphism is essentially apostkinematic transformation of a .3400 to2900 m.y. migmatite-gneiss terrain(BECKlNSDALE et al., 1980). CRAWFORD (1969)found a metamorphic age of 2600 m.y.Incipient proterozoic retrogression of thegranulite fades rocks occurs mainly alongseveral extensive E-Wand NE-SW trending.

ductile mega - shear zones (DRURY and HOLT,1980).

The Doddabetta (or Dodda Bena)Charnockite Complex is located in thenorthwestern part of the Indian «CharnockiteUplands» (Fig. 1). (In the subsequent text.«Doddabeua charnockite. refers to the rockused in this study). Both geographically andstructurally, it is marginally connected to theregional shear zones. The Doddabenacharnockite is an example of a rocks wellsuited for diverse geochemical investigations.It contains a relict mineral assemblage frompeak granulite conditions, including severalpotential geothermometers and-barometers.The rock contains as much as three majorminerals with a surprising quantity of fluidinclusions, much more abundant than in mostsimilar rocks in other terranes.

The present study tries to establish thepressure and temperature conditions of highgrade metamorphism in the Doddabeuacharnockite by a combination of mineralogicaland fluid inclusion methods. The maintools are petrographic studies, electronmicroprobe analyses and microthermometry.Geothermobarometrical data are compared toindependent results obtained from a differentmethod (fluid inclusions),

From this, semiquantitative P-T paths areinferred (see e.g, CRAWFORD and HOLUSTER,1986; TOURET, 1985; NEWTON, 1985, etc.).

This publication is an illustration of the

value of combining independent geochemicalmethods at the scale of the handspecimen, inorder to estimate the conditions of formationfor high grade rocks. However, we wish toemphasize that we by no means claim that thisstudy, based on only onc sample, isrepresentative for the area in generaL Allconclusions have to be confirmed byinvestigations on other specimens .

66

,,'1b ...'"

H·R. mURET, UI. HAN$TEEN

in the collection of che Free University)classifies as a tonalite in the nomenclature ofSTRECKEISEN (1976) see (Table 1), A properrock name would be «cnderbitic-granolitc»(WINKLER 1974), indicating the strongdominance of plagioclase over alkali feldspar,which occurs only as antiperthitic lamellae inthe plagioclase grains. However, the«granolite» nomenclature has never been

Fig. 1. - Geological setting of Doddabt':ttala. (lnsert) Sketch of Somh India, (after GoPAlAKRISJlNA et aI., 1986). The metamorphic grade increases on aregional scale from the Dharwar craton in the north to the high grade terrane in the south. (Northern part ofthe "Southern High grade terrane. Charnockite Upland»).lb. Doddabetta is situated. in the Nilgiris hill, near Doty in the northern part of the Indian "Charnockite Uplands ».The high grade terrane is cut by regional, mainly E-W {tending ductile mega shear zones, inducing an incipientretrogression of the granulite facies rock (DRURY and BOlT, 1980).1: Location of the investigated. sample (6074, VU collection); 2: AmphiboJite/granulite transition (orthopyroxeneisograd); 3: Ductile mega shear wnes; 4: Supracrustal rocks of the DARWAR and SARGUIl group; 5: Proterowicgranites; 6: Closepet plutonic belt; 7: Migmlltite geniss complex; 8: Charnockitc/knondalite complex; 9:Charnockitc/khondalite complex affectcr! by retrogression. A., Amphibolite, G ~ Granulite facies.

Earlier results from the charnockites nearDoddabetta are given by JANARDHAN et al.(1982), HARRIS et al. (1982) and NEWTON(1985). For a more thorough description ofthe Indian granulites, the reader is referrede.g. to HoWlE (1955), WEAVER (1980), RArrHet al. (1983), and SANTOSH (986).

Petrography

The investigated sample (sample no. 6047

widely accepted and the name « Dodda Bettacharnockite», taken in a very broad sense, willbe sufficient for the purpose of the presentpaper. Qualitatively, its mineral compositionresembles other Dodda Betta charnockites,e.g. samples 11.1 and 11.1B, Table 1 inJANARDHAN et aL (1982). However, fromcomparison with sample 11-1 (obtained bycourtesy from R.c. Newton), it is obvious thatsample 6047 is more deformed and contains

GEOTHERMOBAROMETRY AND FWID INCWSIONS IN A ROCK FROM THE DODDABETTA ETC. 67

TABLE 1

Af0d41 analyses and whole rock chemistry ofSIlmple 6047 and other Doddabetta chamockites

OANARDHAN et al. 1982)

~ ..._U) _1·-_..... 4'1)

C_I. toll)

'" '" '"PlocI.

68 }L.R lOURET, HI. HANSTEEN

Fig. 2. _ Auid and solid inclusions in the Doddabeua Chamocldte. I: Photognph of a polished slab of theinvestigated sample. Numeroos globular glll"ncts (high rdicO occur in I strongly ddormcd matrix of qUartl,antiperthitk pIagiodasc and cnhopyroXale. Note me postkinemadc horUontal foliation n.ning around the gunet.s.2 to :i: Fluid and solid inclusions in garnet· ($ • Solid inclusion, mainly quarlz). r_high density COl inclusion).In ) and '. only fluid inclusions. All fluid inclusions an: randomly dispersed in the con:: of the gamel (primary

GEQTIIERMOBAROMETRY AND FWID INCWSIONS IN A ROCK FROM THE OODDABETIA ElC. 69

or two daughter minerals. From the highbirefringence, one of the phases can betentatively identified as a carbonate, probablysimilar to the ones described in garnet (Fig. 3).

x__live: lOOs Pruet.: lOOs R~lnjn9:Rea 1: 105s 5% Oead

0,

1,

< .800FS= 11(I'£J'Il: - ,.

10.2'lO >11 CTS

Fig. J. - Electron microscope imagery of lln opened fluid inclusion in garnet.Left: SEM image Oeol JSM 840 scanning dectron microscope, lnstitutt for Geologi, 0510). In the open cavity,which has a dearly visible negative crystal shape, an aggregate of platy (white) daughter millerlls remains. Someof them (the smallest) have 1»'ltly left the cavity; they could correspond to an original bigger crystal which hasbeen broken during the fracturing of the garnet.Right: EDS-Spens strinly in the fOCU$ing plane are visible (but changing the fotUS makes now inclusions apPeal"in lhe whole mass of the game!). 6: Auid inclusions If) in pbgiodase (negation c:rystal shape; approximatdy cubicshape). 1be complex pattern illustrated in the figure ronesponds to lln early $ubvcrtical trail (N·S) later transposedalong 5C'COOdary small trai1s paralld 10 the main mvaae (NE·SW). 7 and 8: Trail bound inclusions in quam { • traceof late, nearly opm traih. 1be sudden diuwc-rance of nearly all indwions to the right in Fig. 7 corresponds10 a post kinellUltic reaystalliution (annealing) of the quanz grain. Lenght of the bar (photO 2): Photo I .8 mm,all Olhers: 10,lUll.

70 J-LR. roURET, T H. HANSTEEN

stable, primary paragenesis. No fluidinclusions have been found in this mineral,but it contains numerous primary solidinclusions of carbonate.

Mica

Dark mica (biotite/phlogopite, see below)occurs both as primary and secondary grains,mostly in dusters with other ferromagnesianminerals. The distinction between «primary~versus «secondary It is basically made on thesha~ (primary = subrectanguJar sections withwell-defined boundaries, homogeneouscoloration and strong pleochroism) and therelation with neghbouring minerals(secondary .. transionaI gradation and partial

pseudomorphism of other minerals, e.g. garnet

GEOTHERMOBAROMETRY AND FLUID INCLUSIONS IN A ROCK FROM THE DODDABElTA ETC. 71

inclusion trails end abruptly at the grainboundaries; the fluids have evidently escapedduring the annealing process. In other cases,the inclusion trails crosscut boundariesbetween the grains constituting thegroundmass of the mortar fabric.

This puts time constraints on the fluidinclusion formation in quanz and feldspar.They are formed after host mineral growth,but at least partially before the annealingprocess connected to the latest deformationalphase.

Fluid inclusions in garnet occur in discreteclusters, well removed from grain boundaries.They are definitely related to the growth ofthe host mineral ~nd are hence regarded asprimary. As quartz, feldspar and garnet arepan of the same mineral paragenesis, this putsan additional time constraint on the inclusionchronology: fluid inclusions in garnet areformed previous to fluid inclusions in quartzand plagioclase.

1be abundance of inclusions decreases fromcore to rim of the garnet grains. By opticalestimate, the volume proportion of the pureCOl inclusions in some of the garnet cores.are ound to be 10 to 15 volume percent. Thedensity of the garnet is roughly 4 g/cm).With a CO

2density of 1 g/cm). this means

that these parts of the garnets contain about3 to 5 weight percent CO2!

Mineral analyses

Microprobe analyses were carried out witha Cambridge Microscan-9 microprobe at theFree University of Amsterdam, and with aCameca Camebax MICROBEAM microprobeat the Geological Museum in Oslo. Thechemical composition of coexisting mineralphases were obtained from three different thinsections. denoted A2. B2 and C2. Analyticalresults are given in Table 2. Garnet (Table2a), orthopyroxene (fable 2b) and feldspar(Table 2c) analyses from the three differentsections show rather small chemical variations,supponing the view that they show a weakretrograde zonation. The composition of themicas is highly variable. but as in general MgOcontent is significantly higher that FeO. thename ~phlogopite-will be used in the rest of

the study. Strong variations in FeO. MgO andespecially Ti02 indicate the possibility of acontinuous evolution and a probablesecondary origin for some micas, confirmingthe petrographical observations. Low TiO)'weakly pleochroic micas are indeed clearlysecondary. On the basis of their occurrencein the rocks. 3 groups of phlogopite have bttndefined and analyzed:

I) Primary phlogopite entirely isolated inquartz (2d. Table 2).

lI) Primary phlogopite enclosed in solidinclusions in garnet (no reaction rim) (2e,Table 2).

UI) Clearly secondary phlogopite, most ofthem transitional with the outer limit ofthe garnet (2f. Table 2).

Geothermobarometry

Four geothermometers and twOgeobarometers are applied to the Doddabenacharnockite in order to establish the P-T·conditions during metamorphism. Thethermometers used are:

Orthopyroxene·garnet (SEN andBHATTACHARYA. 1984);Alkali feldspar-plagioclase (STORMER,1975);Garnet-biotite (TIIOMPSON. 1976; FERRYand SPEAR, 1978).

Pressure estimates are obtained by usingthe two different garnet-orthopyroxene·plagioclase·quartz barometers given byPERKINS and CHIPERA (1985) and NEWTONand PERKlNS (982) (Fig. 4).

Thermometry

The opx-gnt thermometer (SEN andBHATTACHARYA. 1984) is based on thepartitioning of Fe2· and Mg betweencoexisting opx and gamet. and the Ca contentof the garnet. It is only sligthly pressuredependent, but a pressure corr:ection isincluded in the general formula. given by

• 211}. 0.022 p.}}Ol)~.m oc:?"".~ "'"Il T • . .·L9812 In 1'4.0.787. 1.,~

where T is the equilibrium temperature of theopx-gnt assemblage, given in degrees C and

72

9

7

3

}.L.R. TOURET. ni. IIANSTEEN

200 400 600 800 '·CFig. 4. - Geothermometric and barometric datl. A summary. Geothcrmomctry bubvcnical curves in the P.T.diagram); Loose hatching; Two feldspar thermomeler (STORMEJt, 197'); I, 1I.nd Ill: garnc{.bimilc (I: PrimaryphJogopilc cnrirdy isolated in quartz, U; Primary phlogopitc in garnet, tu: Sel;ondary phlogopitc:l. NOle thecoincidence betw«n III and the feldspar thermometry. Small m~; &llI'nCt-orthopyroxcne equilibri. (SEN andBItArrAOIAuA. 198'). Gcobuometry (dell5e hatching): gamct-orthopyt'oxcne • pl*liodase . quam: equilibri.(NEWTON .nd PnluNs, 1982; PunNs and ClIlPUA, 1985). X _ Preferred peak mctt.tllOtpb.ic P.T. conditions(Gm.Op" and Gnt.Opx.Plag.Qtz. curves).

P the pressure in bars. The distributioncoefficient for Fe1· and Mg betweenorthopyroxene and garnet is given by

X"OPX

2) Ko-~X"GNT

x,

When the mineral analyses are recalculatedto give a sum of 1.00 for the cation contentof each of the minerals, and the K iscalculated, the general formula can be appteddirectly. In Table 3, the tem~ratu~ are gi~nalong with corresponding, selected pressures.(The precisions used in this table, are thosestated by the authors in the originalpublications). Due to the ~ry constant mineralcompositions, the temperature show ~ry littlespread around an average of about 800°C.~garnet u.s«I in these calculations, are the sameas those used for gnt-bim thermometry.

The garnet show rather constantcomposition, while the phlogopites show ~ryvariable Fe2 °jMg ratios (see Tables 2d, e andn, giving a broad range of gnt-biottemperatures. Primary gm-phlog pairs (analysesare given in Table 2e) give temperatures of 425to 535°C at 6 kb. Due to a weak zonation,the use of core compositions of the garnetsgive temperatures roughly 30 to 40°C abovethose obtained from analyses near the graincontacts. The primary phlogopites, occurringas inclusions in quartz, give the much highergnt-biot temperatures of 639 to 783°C. Thesecondary, and at least partly synkinematicphlogopites, give temperatures of 461 to 630°Cat 6 kb (L 11 and Ill, Fig. 4).

The twO feldspar thermometer of SroRMER(1975) give temperatures of 480 and 505°C(Table 3). WHrrNEY and SWR1I.{ER'S (1977) twOfeldspar thermometer gives ~u1ts about 50°Chigher. But according to the reasoning givenby BROWN and PARSONS (1981), the WHITNEYand SWRMER (1977) thermometer should not

GEOTHERMOBAROMETRY AND FWJD INCWSIONS IN A ROCK FROM THE OODOABEITA ETC. 7}

TABLE 3

Thennometry (garnet-orthopyroxene, two feldspar, garnet-biotite)(5en and Bh '" SEN and BHATIACHARYA, 1984; STORN/ER, 1985; F. & 5. '" FERRY and SPEAR,

1978; Thompson '" THOMPSON, 1976). Kd = FejMgGnt-biot Gnt-b1ot Gnt-biotPria. gnt-phlg pain Pri.ol. phlg in qU, + gnt 6041-82 S8condary phlg-,. " , " ThOlllpaon -,. ," " ThOlllpoon -,. ," " ThOlllpaon8 Kbar ". "nHO 0.101 U9HO ". 187±~0 0.28" 696'~ " "68.50 O.llB "68150"" "96:t.5O 0.131 49HSO ". 102tSO 0.2H 639:050 " 637:t.50 0.l03 S94HO", S3h~0 0.1"9 S18050 m 712:lS0 0.2"9 6

74 j,l,R. TOURET, T.H. IIANSTEEN

...J611tl5001691dSOO

""511500ea. 8100

6)

a} a'Fe-reaction: In K2

= In FS A1'.j,acR • a AlM

In order to apply the barometer, theactivities of the hyJXlthetical end members areneeded. The PERKINS and CHIPERA (1985)paper contains a set of formulas for calculatingthe garnet md member activities, which arenot repeated here, but which calls for II specialattention of the reader, as s~raJ formulasmust be corrected according to the publishederrata ((ontr. Min. Petr., 1985. 90:410).Particularly important is the expression ofplagioclase activity (NEWIUN, 1983):

7)

, X . XAN (l-XANFAN· AN

4

There are three prerequisites for thecalculation of Opx end member activities:

A) The first one is the site assignement:All Ca, Mn, Na should be assigned [0 the M2site, and all Tit Fe'· , AI (tetrahedral) shouldbe assigned to the Ml site (WooD and BANNO,1973).

B) Secondary, the ~. to Mg ratioshould ~ equal in the two sites. This can beexpressed by the expressions:

X Ml X M1~ =-.!!!..:..

XM1 XM,

C) The third and probably most essentialprerequisite (which is not clearly stated in thepaper), is that the cation content in the opxshould ~ recalculated, so that each of the M1and M2 sites sum up to one:M 1 cations = M1cations(Note that recalculated opx analyses are givenin Table 2b). Once these requirements areobserved, the activities of the opx endmembers are calculated by simple expressions:

8)

9)

aM,lSilO6 .. X~ . ~

Two gerl(:ra.l formulas are gIVen for thepressure calculations:

10)

P~ = 6.1349 • 0.3471 InK1 + 0.02136 T

0.001140 T InK I11)

Pft-=0.0630 - 0.3482 1nK~+0.0143 T -0.000997 T InK1, where P is given inkilobars and T in degrees C.

Results art: given in Table 4.NEWJON'S and PERKlN'S (1982) gnt-opx-plag-

qtz geobarometer is also applied to the rocks.It is based on reaction 3 given above (the Mg-reaction), but a differt:nt calibration than thePERKINS and Clltl'ERA (1985) Mg-reactionbarometer.

Calculation procedures for the enstatite endmember activities are exactly the same as thosementioned above. The garnet end memberactivities are based on a different set ofsolution models given by GANGUlY andKENNEDY (1974). Calculations of plagioclaseend member activities are based on the «Al-avoidance» model given by NEWTON et al.(1980). The general barometric formula isgiven by:

12)

P = 3,994 + 13.070 T + 3.5039 T InK"whith P in bar and T in K. K", is given by:

(llc. . a1~(;NTK,.~=~=-

(a~) {a~

Results from this geobarometer arepresented in two versions in Table 4. Row no.1 in this table features results obtainedfollowing NEWTON and PERKINS (1982) in allcalculation steps. Row no. 2 is obtained usingthe more up to date plagioclase activity modelgiven by NEWTON (1983).

The NEWTON and PERKINS (1982) andNEWI'ON and PERKINS (1982)/NEWTON (1983)barometers show results in the range 7.6 to 8.7kbar with an average value of 8.2 Kbar, and

T,uILl.. -IBaromelry: GN1·-OPX.PLAG-Q7Z T .. 800 C

11 .. r I ""dins" c::trlpoora 19U11 19l1l P Ifs r r•

7~19tUOO 701011000 JJllO:tlooo"'J5d500 798Hlooo 498Ulooo.UOti500 1~ilOOO .5l0UOOOea. 1000 ea. 1600 ea. 1400

GEOTHERMOBAROMETRY AND FLUID INCLUSIONS IN A ROCK FROM nlE OODDABETTA ETC. 75

7.4 to 8.5 kbar with an average of 8.0 kbar,respectively. This small differences in averagepressure is of course due to the differentplagiodase activity models used, the onefeaturing the newest model (NEWION, 1983)giving the lower pressures. This last one gi~sresults which are comparable to the graphicalbarometer of BOHLEN at al. (1983).

TIle PEluaNs and CHlPERA (1985) barometerpair gives results of about 7.6 kbar for theMg-reaction and 8.4 kbar for the fe-reaction.The averaged value of 8.0 kbar is identicalto the pressure average of the pressuresobtained by the other barometers. A schematicsummary of thermobarometric dala are givenin Figure 4.

NEWI'ON and PERKINS (982) foundpressures of roughly 7.5 kbar for other rocksin the Doddabetta area. TIlls is consistent withour results obtained with the same barometer.In conclusion, the best pressure estimate forour Ooddabetta charnockite lies in the range7 to 8 kbar for a given temperature of 800degrees C. Further elaboration concerningthese barometric results is beyond the scopeof this publkation.

Both two feldspar thermometry and gnt-opx-plag-qtz barometry involve the use ofplagioclase end member activities. Resultsfrom two feldspar-thermometry sh()O,l,l that thefeldspar have reequilibrated aCt-er peakmetamorphic conditions, and thereby the endmember activities have changed. This, ofcourse, influences the barometrical results. Inthe case of the two feldspar thermometer, asmall change in the Ab content of the alkalifeldspar gives a relatively large change in theresulting temperature. This relation is exactlyopposite for the plagioclase, because the Abcontent of the alkali feldspar is here moreimportant than the Ab content of theplagioclase. But due to the high Ab contentin plagioclase, and the high pla.g,lkfsp ratio, theinfluences of the feldspar reequilibration onthe barometrical results are not necessarilylarge.

Both the Mg and Fe·reactions (equation3 and 4) involve a cakium exchange betweenanorthite and grossularite. A change in theplagioclase end member activities wouldalso change the grossularite activity

correspondingly, and again minimize the effecton the barometric results.

Mic:rothennometry/F1uid inc:lusions

Many inclusions in garnet contain anaggregate of daughter minerals (Fig. 3). ~ra1of these could remain in an opened inclusion(freshly and gently broken surfa~) and theyhave been analyzed using a leol ISM 840scanning electron microscope fined with aLink AN 10000 energy dispersive analyser(EOS) unit. The EOS spectrum in Figure 3,is interpreted as a Fe, Mg dominated carbonate.As shown in the Figure, the aggregatescommonly consist of carbonate platelets. Theextremely small Ca content is striking.

At room temperature, each of the inclusionsare filled with a monophase liquid. Thismeans, of course, that their finalhomogenization temperalure are less thatabout 20°C. All inclusions homogenize intothe liquid phase (denoted by Th(L)).

Contaminants in the form of other gaseousspecies (notably N2, CH,.) are commonlyfound in inclusions from other high gradeterranes (see e.g. TOURET, 1985 andSWANENBm:i. 1980). This may seriously affectthermobarometric interpretations (BURRUSS,1981; GUILHAUMOU et al., 1981).

All measured inclusions show meltingtemperature of -56.6°C, which is the triplepoint of CO2, This observation is of majorimportance, as it indicates the purity of theCO

2, The composition of the inclusion

contents was also checked with a Laser-Ramanmicroprobe with a multichannel detectorsystem (OILOR MICRODIL 28). Thisequipment facilitates fast and reliable in situanalyses of molecular species inside singleinclusion cavities. (The resolution lies in theonc micron range). The analytical resultsconfirmed the purity of the CO2 (Fig. 5).(For a description of this system and itsadvantages, the reader is referred to BURKEand LUSfENHOUWER (1987). Several analyseshave been done on inclusions in quartz,plagioclase and garnet, respectively, which allconcluded with the absence of any other gasthan CO2 (no CH., N2, H1, H 2S)_ For any ofthese gases, the deteoion limit is below 1%.

76 J.L.R. mURET, T.Il. HANSTEEN

This, combined with the melting data, willallow us to interpret all presem Th data interms of pure C01,:. This does not imply thatno OIher gases will ever been found in theDodda Betta area, but a more completeinvestigation is far beyond the scope of thepresent paper.

(It Fig. 6), contains two separate peaks, withmaxima at -5 and _18°(, corresfK)nding toCOl densities of 0.96 and 1.02 gJcmJ,respectively. (Density data are adapted fromANGUS et al., 1973 and HOlLlSfER, 1981). Thecomplete separation of the peaks may be dueto variations in CO2 ~ressure during garnet

I'R£[ l.tllYOlSllYl'f!STDtllAll ltl.l

IHST. EIlfITH SCl.

DllOR~ICRllll[L-2e

Of'(RAtOR TllRfT fXC IT .L1NE1",,1 514.531 9 Sl"ttl.:;t IT VlDlHlca-ll 6.3SDATE 31 l1'Wl: 19$7 u&R .1.VI 6e3 DOECTCfUHbr- of dlodul 512SAI'I'U 152.7 81 FOll£HOtlO.I",-1I lee47 FILTOl toE

IIHlOI 87/$ Sl'[CTRtl.lo-1I 16247 lllTEGMTlON 1I1'£lsl :3

ItlIIE tU..T101f'0'ft].. SLIT IIIOTlll",,1 m IOIlER or IUlIUATlOO5 seRE.'II'lRK , C0'2 IN lHCl.USIOtl VlTIlIlllGNESll'E

e-r;

•s•~m-

f--~

~~~

~

~ \ m~,~

,.I. ".. 14'5. .... 'N ,.. ~".. ".. OHFig. 5. - Raman spectrum of COl (Microdil 28, Free Univetsily, Amsterdam, Analyst: E. Burke).

Fluid inclusions of sufficient size andabundance for microthermometric studies arcfound in thl'e(: minerals; garnet, quartz andpbgioclase. The measurements were performedon eight selected grains: 3 garnets, 3 quartzcsand 2 plagioclases. The results are presentedin Fig. 6. Note that in plagioclase,measurements were done on inclusion trailsboth parallel to and inclined to the bestcleavage plane (00l).

Fluid inclusion measurements from each ofthe mineral species were used to constructfrequency histograms, presentinghomogenization temperatures reJati~ to thenumber of measurements (Fig. 6). Somedifferences between the thrtt histograms areimmed.iawy evident. The one with garnet data

growth, possibly reflectmg changes in the fluidregime during the early stages of peakmetamorphism. However, it has not beenpossible to observe a clear spatial distributionbetween inclusions with different densitiesand both values have been used for thebracketting of the garnet field (Fig. 6).

The plagioclase histogram (2. Fig. 6) seemsto consists of two partly superimposed peaks.They occur at -27 and -35 degrees C,corresponding to CO2 densities of 1.06 and1.09 g/cm l , respectively. Homogenizationtemperatures from the quartz inclusions O.Fig. 6) show a remarkable uniformity. Theyare concentrated in a single, narrow peak witha maximum at _39°C, corresponding to aCO2 density of 1.12 g/cm

l .

GEOTlIEMtOBAROMETRY AND FLUID INCLUSIONS IN A ROCK FROM THE DODDABETIA ETC. 77

Another histogram, labelled «(combi» (4,Fig. 6), combines all Th measurements. Themost conspicuous peak is represented by thequartz inclusions. To sum up, CO

2densities

are very uniform for each separate mineral,compared to many other examples (e.g. inSouthern Norway, TouRET, 1985; Thcommonly varies between -20 and + 3QoCin a single mineral grain). Moreover, thedensity tends to increase in order of increasingdeformation and recrystallization (increasinglyhigher densities from garnet, throughplagioclase to quartz).

consistent with ToURET and 8crrnNGA, 1979).Here they are shown together with P-T-estimates from solid phase thermobarometry.Because these isochores represent frequencyhistograms, they are oot drawn as single lines,but rather as «isochore bands», covering thedensity ranges indicated in the figure.

Fluid inclusion petrography has shown thatthe garnet inclusions are primary with respectto peak metamorphic conditions. Plagioclaseand quartz inclusions are trailbounded andsecondary with respect to host mineralgrowth. This makes them, at least partially,

COMBIN~ 143Max {23J -40/-38

QUARTZN~ 59Max (22J:-40_-36

0.96 1.12

1.02 00) GarnQtN~34

Max[6J:-6/-4

-~ -~

1.06 1.12

PlagioclasQ ®Nz 50Max [8J:-28/-26

Fig. 6. - Homogenization temperature (Th) histograms for CO2 inclusions in garnet (1), plagioclase (2) and quartz(3). (All temperatures denote liquid homogenization). The above three histograms are combined in «combi" (4).For each histogram; The first line (N) in the local number of measurements. The second line is the maximumfrequency (in number of the total number of measurements) in a given temperature interval. For instance, inhistrogram No. 4 (oombil, 23 of the measurements (maximum frequency) occur between -38 and ~40°C. Numbersabove each peack: CO2 density in gJcmJ.

Evidence for a «isobaric cooling» PT pathafter the peak of metamorphism

Geothermobarometric calculations showpeak metamorphic conditions of 795 to 815°Cand 7.4 to 8.7 khar. This is illustrated by the«P_T box» in Figure 4. These P-T-estimatesare regarded as the starting point for the p-T-story indicated by the present study.

The maximum frequency values from thehistograms (Fig. 6) are used to constructisochores, which are drawn in Fig. 7(Thermodynamic data are taken fromHOLLOWAY, 1977; 1981 and they are

later. Knowing this fluid inclusion chronology,the frequency histograms can be used in thefurther explanations: The average CO2density increases gradually from garnet,through plagioc1ase to quartz.

Isochores bracketting the «low» densityprimary inclusions in garnet pass somewhatbelow the Doddabetta peak metamorphicconditions. Results from other granuliteterranes suggest that some H 20 was presentduring prograde metamorphism (COOLEN,1982). In our case, this is proven by thepresence of primary phlogopite. In cases

78 j.L.R. TOURET, T.ll. HANSTEEN

similar to this, a water pressure of I to 2 kbarin addition to the measured CO

2pressure

has been indicated (TOURET, 1974). Thiswould make the primary garnet inclusionscompatible with the peak metamorphicconditions. The water in the garnet inclusionscould have removed through reactionsforming daughter minerals, but also at leastup to 10 mole% H20 may remain unnoticed

have lower densities than the quartz ones. (Itis also noteworhy that they show much greaterinternal variations, the ones along the beStcleavage plane being less dense). This indicatesthat most plagioclase inclusions areintermediate in time those from garnet andquartz. All inclusions in quartz haveremarkably similar homogenizationtemperatures, which can only reflect a well

800

,

8

'0

8

6

4

200 400 600

~milXH P-Tcondj ~

•

/..-M:' .......-::'Milln"

gr;llnUllt~ trvnd

Tourvt 1981

T·e

Fig. 7. - Mood of post-metamorphic poT plIth in the O

GEarJ1ERMOIlA.ROME'RY AND FWID INCWSIONS IN A ROCK FROM THE OODDABETTA ETC. 79

are strongly deformed. Some of the secondaryphlogopite is thus pre- to syn-kinematic. Thegnt-biot thermometers (FERRY and SPEAR,1978; THOMPSON, 1976) give a large spreadin temperatures for secondary micas. Thispoints toward the development of very localequilibria after peak metamorphic conditions.The temperatures obtained by STORMER'S(1975) two feldspar thermometer, are roughlyin the same range as the lowest temperatureobtained by using the two gnt-biotthermometers on the group 1 primary gnt-phlgpairs, i.e. 500 + /_800 C. If the fluidinclusions in quartz are trapped during thedeformation phase, and the subsolvus feldsparre-equilibration is contemporaneous with thelatest phlogopite re-equilibtation, asemiquantitative area on the PoT-path can bedefined. This is the area where the quartzisochores intersect the gnt-biot and twofeldspar thermometric curves (Fig. 7). Evenif the gm-biot temperatures are erroneous, thisarea is well defined by the intersection of thequartz isochores and the two feldsparthermometric curve.

When all textural, thermobarometric andmicrothermometric informations arecombined, they reveal the following P-T-story;peak metamorphic conditions of 7.4 to 8.7kbar and roughly 800°C were obtained bythermobarometry, and fits relatively well withthe densities of primary inclusions found ingarnet. Secondary feldspar and quartzinclusions show a gradual increase in COldensities after peak metamorphism.Formation of secondary phlogopite possiblycovers the entire temperature range from peakmetamorphism, to less than 500°C.Deformation and sub solvus feldspar re-equilibration both occur at temperatures ofvery roughly 500°C and 5 to 6 kbar. This lastrecorded event is indicated by the intersectionof the twO feldspar and gm-bim thermometriccurves with the isochores for quartz (Fig. 7).

Comparison with other granulitefades terranes

In many granulites (TOURET in HOLLlSTERand CRAWFORD, 1981) both mineral andmicrothennometric data are compatible with

an adiabatic uplHt model (Fig. 8, I). In suchcases, the lower continental crust has beenuplifted from 6 to 8 kbar (corresponding toa depth of 20-35 km) through a cooling pathsomewhat convex towards the T axis (Stt e.g.COOLEN, 1980). Thisean be regarded as themost ..:normal» uplift path. The fluidinclusions decrease in density after peakmetamorphism. This is exactly the oppositeof what is found in Doddabeua.

The case of Southern Norway is morecomplicated.. In the western part (Rogaland),SWANENBERG (1980) found increasing COl'densitjes in post-metamorphic inclusions andhe was then the first to propose a poT pathcomparable to the Doddabetta one (

80 J,t,K. lOUKET, TJI. HANSTEEN

CD

(0) peak metamorphic conditions anddeveloping during post metamorphic ~gionaIcooling. This is (freely) illustrated in Fig. 8III as a thickening of the continental crustdue to overthrusting, fonowed or accompaniedby an isostatic readjustment which wouldbring t~ 10Wtt part of the initially above lyingcrust to the surface. At any tem~ratuI'(: alonga ~normal .. post metamorphic path (e.g. T p

Fig. 8. - Post-metamorphic poT trajectories in gnnulitcsand related rocks. I: .. Adiabatic uplift .. (A.U.). Thistrajectory. still regarded as the «normal. granulite trend(Toull.£T, 1981), corresponds to vertical movements of• continental block (hatched .. lower crusta!, granulitedomain). ll: .. Isobaric wolinglO (I.e.) .. Temperll(ured«ttase al high pre:5Sun:, relate

Conclusion

By carefully combining petrography,thermobarometry and microthermometry,severe constraints can be put both on theevolution of the fluid density and on the p.T-paths of a rapidly increasing number of highgrade rocks.

This task is never easy, but some specimensare better suited that others. In this respectthe Doddabetta is one of the best that we haveever seen, both for the quantity of inclusionsand the specific character of the inclusions ina given mineral. One sample is not enoughto drew large scale conclusions, and wewill restrain to comment here on themuch debated problem of fluid regimein the lower crust (

GEOTlIERMOBAROMETRY AND FWlD INCWSIONS IN A ROCK FROM THE OODDABETIA ETC. 81

immjscibility between carbonate and silicatemelts has ever been observed under lowercrusta! conditions. But on the other hand,direct evidence of immiscibility betw~ncarbonate and silicate melts has recently beenobserved in mantle xenoliths, in the form ofquenched droplets of carbonatic and basalticglass in lherzolites (AMUNDSEN, 1987). Thisgives way for imeresting aspects concerningthe source of CO

2, Observing the

Doddabeua chamockite, we want to st~s thepetrographic relation between primary CO2inclusions and a comparable type of carbonate(iron and magnesium rich) as found in theBamble meta intrusives (TOURET, 1985). Weare convinced that some genetic relation doesexist between them, and its preciseundestanding will greatly enhance ourknowledge of the source of CO2 in the lowercrust.

Acknowkdg~mn'lh. - The sample used in this study,is taken from the collection at the Free University in

REFERENCES

MtUNDS£N H.E.F. (987) . Evitknce /0' liiJllidimmiscibility i" ~ Ilppn malltk. Nature, v. 327,692·69'.

ANGUS S., A&:.1S'RONG B., DE REUK K.M., ALruNlNV.V., GADE'ISKII a.G., CHAPAU. G.A., RoWUNSONj.S. (1973) ·lnttmatiof/IJI tlNrmodYlltlmic tlIb~ 0/~fluid st4U: Gubondioxitk. Pcrgamon~, New York.

BECKINSI)A!.E R.D., DIIUIIY S.A., HOLT R.W. (1980)·J,J6o-My, old prriun from ~ South !nditln c,.ton.Nature, v. 283, p. 469-470.

BoHLEN S.R., W"'LL V.J., BoETI'CIlElI A.L. (1983) -Experimn'ltal in~tigation and application 0/ gametf/'tllllllitt ~ui/ibrUl. Contr. Min. Petr., v. 83, p. '2·61.

BROWN W.L., P ... IlSONS 1. (l98l) . Towards a morcpractical two feldspar goothl!mlometa. Conlr. Mineral.Petrol., v. 76, p. 369·377.

BURKE E.A.}., LUSl'ENlIOUWER W. (1987) • Theapplielltion 0/ Il multiehllnnel WsI!T' Ramlln microprobe(MICRODIL-28j to tk af/IJlysis 0/ fluid inclusions.Chem Geol. 61, p. 11-17.

BURRUSS R.C. (1981) - A1J/l/y~ 0/pbIl~~uilibriD in C·O·N-S fluid inc/Ilsiom. Mineral. fusoc. Canada ShortCourse Handbook 6, p. 138-156.

CooLEN J.j.M.M.M. (1980) . Chtmic.l pttrolof1 0/ t1xFIlIW G",nlllit~ Complec, Soutbm! T/I1Illlnu.. (Ph.D.thesis) GUA Papers of Geology, Sc:r. I, No. 13-1980.

Coou£N J.J.M.M.M. (1982) - C.rbonk jlMid inc/lluollsin V.lIulitn from TanlJ/ni.: • eom/Mrisoll 0/g«Jh.~~ basd cm /kiddmJitylUJdmi~chnniJlT'J. Chcm. GeoI., 37, p. '9-77.

AmSlerdam. It was collected by professor W.Uytenbogurl during cxcunion A 24 of the 25thImermnionaJ Geology Congreu in India in 1%4.

Half of lhe microthermometric data were measuredby M.L. Freuolti. The Raman analyses and microprobeanaly5CS which were performed in All15terdam, weredone by E. Burke .nd W. Lustenhou"'er, respectivdy;facilities fOl' dcctron· and laser RamiUl microprobe wereprovided by the Free University in Amnerdam and bythe WACOM,. working group for anaJytical chcmiSltyfor mineral and rockJ; dus group is subsidized by theNetherlands Organization for the AdvlII'ICeTllCnt of PureResearch (NWQ1.

T. Hansteen gratefully rlnowkdges u.: constant helpand encou~cmcnt from Ihe vakgroc:p EMP during Islay at the Free University in Amsterdam February-April1986. This stay was subsidized by the NorwegianCouncil for Scientific and Technical Research (NTNF)under the programme .. Ores associated with GraniticRocks,. which is part of lhe EEC reasearch programmeon minerals. Thanks arc also due 10 T. Kirkedy forproviding a computer program for gnt.biot thermimetry,and to P. Hagdia, B.Jamtvcit and F. Theisen for helpfuldiscussions.

Several drafts of this paper were reviewed by E.R.Neumann, T. Andersen and R.C. Newton, who is alsothanked for sending a reference sample of a Doddabcttacharnockile.

C"AwFOllD A.R. (1969) - R«cmnaiSSlUlu Rb-Srdating 0/~ f'rrumbri411 rocks 0/JOUtbm! pt1finsu14r bulU. J.Gco.l. Soc. India, v. 10, p. 117·166.

C.....Wf'ORI) M.L., Hou..rstU L.S. (1986)· Mnmnorphkfluids, the cviikrlu from fluid indusiom, ill WAI.nlEl.J.J., WOOD D.j. (Ed.). Fluid rock intl!T'actio" dllringmel

82 J.L.R. TOURET, T.H. HANSTEEN

GtobD1'Qt1/ttry. vothmnommy, and IiItt Arr/xoanVOlhmm/rom lM granu/i~jlU:1d tnrrJi" 0/wulh Indi4.Ceol, Y. 90, p. ':09·527.

110UJSTD. LS. (1981) . brfOnnatiOfl inJri1tsically tlVtlilobkfrom jluiJ incu.UOnl. Mineral. Assoc. Canada ShorlCOUl"$(' Handbook 6, p. 1·12.

Hou..ow,o\Y J. R. (l977) • Fuf/Kil1-' /lCtir,ir, 0/mol«uWrspuk in 1upt~,itiCtIi fluids, p. /6/·181 If/ThnmodY1Ulmics In gt'O/ao. (D G. Fraset, ed.l. D.Reidel, Dordudll, lIon~nd.

HOL.LOWAY J.R. (l981)· CompositIons i",d uc/umn 0/supnmtitll/jJuids in ,Ix Emth's mm. Mint:ral. Anoc.Canada Shorl Course Handbook 6, p. 13·28.

HOWlE R.A. (I9.5.5) . The- geochernimy of IheCharnockilc $Cries cl Madras, India. R. Soc.Edinbut8. v. 62, p. 72:5·768.

JANAMDllAN A.S., NEWTON R.e., SMI111 J.V. (1979)-Andrtll crusw{ metamorphism III low Plfld chamockiteIOm/dtion at I0Jbbalduryjl, South India. Nature, v. 278,p.511·:i14.

)ANMWItAN AS, NEWTON R.c., HANSEN E.C. (l982). The trrms/ormation of amphibolite facin gnein tochamockitt in muthem Kamataka and northern Tami!Nadu, India. Connib. Mineral. Petrol., v. 79, p.130·149.

KERRICK D.M., DARKEN L.S. (l9n) • Statisticalthnmodynamic mOlklsfoT idal oxiu IVId si!ica(t solidwlutions with applialtion to p/agiot;~. Geochim. Acta39, p. 1431·1442.

NEWTON R.C. (1983) . G~rometry of higb-",.Jemetllmcrphic rocks. Amer.]. Sci., v. 283·A, p. 1·28.

NEWTON R.e. (1985) . T~perahJ~, prnlMTr andmetllmorp!Jk fluid rrtim~ in tM tlJ'tphibo/i!t ftlci~ togmnu/ikffICin mmsition lOtteS. In Toal A.C., TOUu:T].R.L. (Eds.), D. Reidd 198', p. 75-104.

NEWTOH R.e. (1986) . Fluw of gmnMlite fKinmetllmorphism. In WALl'HEII. J.T., WOOD B.J. (£d.).Fluid·rocks imeracrion during metamorphism,Springer Vcrlag, p. 36-'9.

NEWTON R.C., CIIAlI.LU T.V., K1.EPPA 0.]. (1980) .Thnmochnnistry ofthehig, ~tuftll JtIIk p/agiot;ltue.Geochim. Cosmochim. Acta, v. 44, p. 9)3·941.

NEWTON R.C. WINDLEY B.P., SMml J.V. (1980) .urbonic meUlmorphism, granulitn and crusUll growth.Nature 288, p. 4'-50.

NEWTON R.C., PERKINS D. (1982) - Thermodynamiccalibration ofbarometers based on the assemblage garnet·plagioc!ase·orthopyroxtlle (cJinopyroxene).quartl. Am.Mineral., v. 67, p. 203·222.

OTT'O J.W. (1984) . Melting Tr!ation in some carbonate·sifica!t sySletm: source and products ofCOrrich liquids.Ph.D. Thcsis, University of Chigaco, p. 220.

PERKINS D. 1II, CIIlPEk.\ S.j. (1985) . Garnet·ortoop)'roxene-p!agiocltJSNjU4m baroffrtry: ~finementand OIppfication to the English River suhprovinulmd tMMinnesota RiwT VlJfky. Comrib. MiOttal. Petrol., v.89, p. 69·80.

R.\ASE P., RAmi M.. ACKERM.\SD D., UL R.K. (1986). Prog,ruive metamorphism of mafte rockJ from

green~hiJt to granulite fades in the DhtJrwaT Cnton ofSouth /ndilf. Journ. Geol., v. 94, 2, p. 261-282.

RAlT11 M., R.\ME P., ACKU."V.Nll D.,l.AL R.K. (1983). Regionalg«Jthermometry in tM gmnulite j4cieJ temJneof South /ndiOl. Royal Soc. (Edinburgh) Earth Sri.Tr2nS., v. n, p. 221·244.

SA:\'TOSlI M. n98') . FIMid evolution ~hmKteriSlkJ OIWJp1ezothermk If""" of So"th lndum ~b.mot;kittJ.Geology, v. IJ, p. 361-363.

SA:\-rosll M. (1986) . GJrbonic metamorphism of~htJmocJrit~iw tlx JO"tJm:~temIndian 5hi~lJ: A fluidindusion Jlud,. Lithos. ,'. 19. p. 1·10.

SES S., BlIATfAl..llAK\.\ t\ '1'lb5j· An orlbopylWlene-gpmet thermometer IJnJ IlJ IJpp/ication to the Mad1tJJcbtJmockiteJ. Contrib. Mincral. Petrol., v. 88, p. 64·71.

STORMER ].C. JR. (I9n) . A PrlJtical twojelJJpa,geothermometer. Am. Mineral.. v. 60, p. 667·674.

STRECKISEN A. (1976) . To efUh plutonic rock its propername. Earth Sci. Rev., v. 12, p. I·jj.

SW.\NENIIERG H.E.C. (1980)· Fluid indusiom in high.g'ade metOlmorphi~ rocks from Notwoy S. W. (Ph.D.Thesis, Uttecht) Gcologia Uhraiectina No. 25.

TllOMPSON A.B. (1976) . Mineral TrfUtiom in pelilic rocks:Il. ukuwtion of Jome p. T.X(Fe.MgJ pha~ Ttwtiom.Am. Journ. Sci., v. 726, p. 42'·454.

THOMl'SON A.B. (1980) - Fluid-absmt metamorphism. J.Geol. Soc., 140, p. S33·HS.

TOUJtET J.L.R. (1974) . Fad~ gTOInulize et fluidecIJ1boniqu~. Cem. Soc. Gcd. Bclgiquc des domaine$criuallins, Liege, p. 267·287.

TOURET ].L.R. (1981)· Fluid ilfClusions in hish vmkmeUlmorphic rocks. Mineral. AsStX. Canada ShorlCOUrK Handbook 6, p. 182-208.

TOUu:T J.LR. (198') - Fb«J Pr'gime in 5cf4them NorwtIJ:The rrrord offlMid inclusionJ. In TOBI A.e., TOUJ.ET].L.R. (Eds.), The deep Protcrozoic crust in the NorlhAllantic Provinccs, D. Rddd 198'. p. SI7·H9.

TOURET J., Ol.SEN S.W. (l9S,) . fluid inclusionJ inmigptatitn. In ASlIWOIlTI-l J.R. (Ed.). Migmatitcs,Blakie Pub., p. 26'·286.

TOURET j.L.R., VAfot DEN KERJtIlOf" A_M. (1986) . Hig,density fluids in the 10IWl' Cfl/st IJnd upper mantk'.Physica 193 &: 140B, p. 843·840.

TOURET J., BornNGA]. (1979) . Equation d'Elat pou'le COi IJpp/ication IJUX incluJiom COIrOOniqueJ. Bull.Min~ra1., No. 2, ,77·,S3.

WEAVER B.L. (I9S0) . Rart:-Earth ElemCtlt geoclxmiJtryofMadras gratll/lites. Comrib. Mineral. Petrol., v. 71,p.271.279.

WHIDIEY J.A., STORMER J.C. JR. (1977) - The diJtributionof NaAl5iJO, httween coexisting microcline OIndplagioc/ase and its effect on g«Jthermometric clJkulationJ.Am. Mineral., v. 62, p. 687-691.

WII':"UR H.G.F. (1974) . PetrogeneJis of mttIInlOrphicrockJ (third edition). Springcr-.Vcrlag.

WOOD B.J., B.\NNO S. (1973) . GtJmet-{)rlbo/1)7OU7le lmdorlbopyroxme-c/inopyroxene ~/ationshi/1J in simple OIndcompkx JYJt~s. Comrib. Mineral_ PC'lrol., v. 42, p.109·124.