sign in sign up
<<
Home , Celadonite

American Mineralogist, Volume 7 1, pages I I 09-1 I I 7, I 986

Toward a practical plagioclase-muscovitethermometer

N,q,rHA.NL. GnnnN. SrnvnN I. Uso.q.Nsxv* Department of Geology, University of Alabama, Tuscaloosa,Alabama 35487, U.S.A.

Ansrnlcr The Na-K exchangereaction between plagioclaseand has been formulated as a geothermometerutilizing a ternary-feldspar subregular solution model, a modified binary white solution model, and available thermochemical data that account for high and low structural statesof feldspar. The plagioclase-muscovitegeothermometer has been applied to pelitic metasedimentary rocks and peraluminous granitoids that have equilibrated at temperaturesof 49G750'C and at pressuresof 2-13 kbar. The equilibration temperaturesof pelitic rocks, predicted from compositions of coexisting plagioclaseand muscovite, are generallysimilar to those predicted by garnet- geothermometry.The plagioclase-muscovitetemperatures of peraluminous granitoids derived using the com- positions of secondarymuscovites are distinctly subsolidus, whereas temperatures esti- mated using primary muscovite compositions are near liquidus. Application of the pla- gioclase-muscovitethermometer extends the range of pelitic metasedimentaryrocks and peraluminous granitoids for which equilibration temperaturesmay be estimated.

INrnonucrroN geothermometerbased on the Na content of muscovite Muscovite and feldsparsare important constituents of coexisting with sodic plagioclases(Ano-,r). Tracy (1978) pelitic metasedimentary rocks and peraluminous gran- and Cheney and Guidotii (1979) demonstrated that the itoids. The compositions of thesecoexisting can Ca, Na, and K distributions betweenmuscovite and pla- be related by the Na-K exchangereaction: gioclaseexhibit regular variations with metamorphic grade. Cheneyand Guidotti (1979) further showedthat the com- KAlSi3O8 + NaAlrSirAlO,o(OH), positions of coexisting muscovite and plagioclasecould K-feldspar paragonite be used as a viable indicator of intensive parameter (Z : NaAlSi3Ot + KAlrSi3AlO,o(OH)r. and/orfr*) variation within a single, petrographically well- characterizedterrane. They used the paragonitedehydra- albite muscovlte tion reaction involving albite, aluminosilicate,and quartz to estimateequilibration temperaturesof lower sillimanite Thompson(1974) and Chatterjeeand Froese(1975) have through sillimanite + K-feldspar zone metapelites from provided extensivetheoretical treatments of paragonite- PuzzleMountain area,Maine. Becausethe paragonite muscovite-alkali feldsparrelations in the systemNaAlOr- the dehydration equilibrium used by Cheney and Guidotti KAIOr-AlrO3-SiOr-HrO. In particular, Thompson (1974) requires both an estimate of/"ro and the presenceof an noted the temperature dependenceofthe exchangeequi- aluminosilicate polymorph, this geothermometercannot librium and presentedfree-energy relations basedon end- be applied to rocks in which muscovite and plagioclase member thermodynamic data derived from both ion-ex- coexist in the absenceof andalusite, sillimanite, or kya- changeexperiments (Iiyama, 1964) and from muscovite nite. In order to overcome difficulties involved with sub- and paragonite dehydration experiments (Day, 1973; solidusre-equilibration of alkali feldsparduring late-stage Chattefee, 1973;Chatterjee and Johannes,1974). How- metamorphic and magmatic processes,or the absenceof ever, when theserelationships are applied to natural rock an aluminosilicate phase,the Na-K exchangerelationship systems,predicted muscovite-alkali feldsparequilibration betweenmuscovite and plagioclasehas beeninvestigated. temperaturesbased on endmember thermodynamic data This paper presentsthe formulation of a geothermometer are generally less than 100'C, suggestingloss of albite basedsolely on the compositions of coexistingmuscovite component from K-feldspar during subsolidusre-equili- and plagioclase. bration (Labotkaet al., l98l). Kotov et al. (1969) proposed a geothermometerbased on the distribution of Na between coexising muscovite TxnnlronyNAMrc BACKGRoUND and plagioclase.Talantsev (1971) presented a graphical Ternary-feldsparsolution model and excessterrns The albite and orthoclaseactivities of plagioclasefeld- I Presentaddress: Department of Geosciences,Murray State spars can be formulated in terms of a ternary model as University, Murray, Kentucky 42071, U.S.A. follows (Wohl, 1953;Saxena, 197 3): 0003-004x/86/09I 0- r l 09$02.00 I 109 lll0 GREENAND USDANSKY:PLAGIOCLASE-MUSCOVITE THERMOMETER

RT ln ao,: R?" ln ao, tropy (l/"), and volume (Wr) terms, and the mole frac- paragoniteare given as + {x bII'I4N^+ 2X',(WE"K- 'yy^)l tions of muscovite and + x + 2xo,(w?K- wc^11 X'" : (X")(XX)' "lw" : + &"xAb[0.5(w3""+ w[c^ a 14t5N' X"" (X""XXX)' * w\,r - w?"" where - Wf;^c^1 &: tK/(Na+K+Ca+Ba)1, + xo,(wK - w5c" X"" : [Na/(Na * K + Ca + Ba)], * Wf;^x- WY^) and + (x"" - Xoo) 4r: [AlvI/(Fe + Mg + Mn + Ti + Al"I)] '(w7*" - ttl5'91), defined on an atom per formula basis for the mica. Be- RT ln aoo: RZ ln aoo causemuscovite and paragonite components of a white + Ixz,lwE^K+ 2x^b(wN^- tt?*)l mica appear on opposite sides of the plagioclase-mica + x7"1w1"."+ 2x^b(wg^Na- w}aca)f Na-K exchangeequilibrium, the octahedralAl terms (i.e., + Xo.XA"[0.5(WE^K+ W$N^ a 1YaN. -XX) of these mole-fraction expressionscancel. Thus, the to be independentof - - exchangeequilibrium would appear * Wf;'a W7K W5c^7 Fe-Mg (celadonite)substitution in the octahedral sites of + X^b(Wl'- Wf;"r the white mica. Pigageand Greenwood (1982), however, + W?"" - Wf;"c^1 have demonstratedthat muscovite-paragoniteactivity re- + (x* - x"J lations are strongly affectedby the ubiquitous presenceof '(lv5c" - W*\IL celadonite and/or phengite component. They considered a quaternary mixing model involving paragonite, mus- where the interaction (Margules) parameters, VIto,which covite, and two celadonite endmemberswritten in terms describemixing relations within the Ab-Or, An-Ab, and ofpure K and pure Na, respectively,and excessV[/Gpa' An-Or binary systems,can be resolved into enthalpy, en- rameters that were assumedidentical for the muscovite- tropy, and volume terms; X-, Xor,and Xtu are mole frac- paragoniteand celadonitebinary joins. On the other hand, tion of orthoclase,albite, and anorthite; ar;.da terms rep- severallines of evidence suggestonly minor substitution resent the ideal contribution to the activity of albite of the Na endmember in white mica. Crystallographic component(Price, 1985), studies indicate that becauseof the sizes of the cations involved, addition of Mg, Fe2*,or Fe3*to the octahedral - - a"o: (&uX2 Xot X",)(X", * Xo), sites would causerotation and comrgation of the tetra- and orthoclasecomponent, hedral sheet, such that the interlayer site enlarges and becomesunfavorable for holding the small Na* in place - - ao,: (X)(2 Xoo X".X&, + X".). (Bailey, 1984; Guidotti, 1984).Moreover, muscovite is Green and Usdansky (1986) have shown that theseactiv- always enriched in Fe and Mg relative to coexisting parag- ity-composition expressions,combined with the An-Ab onite (Henley,1970; Hock, 1974;Hoffer, 1978;Katagas Margules parametersof Newton et al. (1980), the Ab-Or and Baltatzis, 1980; Grambling, 1984; Guidotti, 1984). Margules parametersof Haselton et al. (1983), and An- These relationships imply an antipathy between Fe-Mg Or mixing parametersderived from Seck's(197 la, 197lb) substitution in the octahedral sites and Na occupancyof experimentalfeldspar compositions, can be usedto model neighboring interlayer sitesof white . It is therefore ternary-feldsparequilibria. possiblethat most Na substitution in natural occursas a paragonitecomponent. This suggestionis sup- ported by the generallylow Na contents(0.02-0.04 atoms) Muscovite activity relations of natural celadonites(Wise and Eugster,1964; Boles and Binary solid solution between muscovite and parago- Coombs, 1975; Odom, 1984). Accordingly, the binary nite is nonideal,and an asymmetricbinary solution model white mica solution model has been modified to reflect generally has been used to expressactivity-composition the possibility that Na enters the lattice of natural mus- relations of muscovite (Eugsteret al., 1972; Thompson, covites only as the paragonite endmember, whereas K 1974; Chatteiee and Froese,1 97 5 ; Cheneyand Guidotti, would be present in both the muscovite and celadonite 1979), components.The muscovite mole-fraction terms (X'") in the muscovite and paragoniteactivity expressions,given RIln ar" : R?"ln XM"+ X2p^lW"+ 2XM"(Wd - W")\ previously, have been modified by reducing X'" empiri- and paragonite, cally as RTlnar^:RIln Xp.+ XM"lWi + 2Xp"(W - W)1, X,":[K/(Na+K+Ca+Ba)] where lQ'and Wo^arebinary excessfree energy ofmixing .(Al"' - Na)/(Fe + Mg + Mn + Ti parametersthat can be expressedin enthalpy (W"), en' + Alu')l GREEN AND USDANSKY: PLAGIOCLASE-MUSCOVITE THERMOMETER 1l1l on an atoms per formula basis, where by virtue of the action, however,is complicatedby differencesin the struc- (Alu' - Na) term, only the K associatedwith the mus- tural statesof the low albite-microcline and high albite- covite component is considered. The paragonite mole- sanidine solid solutions. The free-energyrelationships of fraction terms remain the reaction have therefore been calculated from the in- ternally consistentthermodynamic data of Helgesonet al. X". : + K + Ca + Ba)], [Na/(Na (1978), which take into account the structural state of where all Na is present as paragonite endmember. This plagioclaseand alkali feldspars(Table l). The variation asymmetricformulation ofXr. and X"" is compatible with of RZ ln KD with temperature is illustrated in Figure l. Grambling's (1984) suggestionthat increasedceladonite The equilibrium condition that the Gibbs free-energy substitution shifts the position of the muscovite limb of change(in joules per mole) is zero at a given temperature, the white mica solvus markedly toward higher K/(Na * T, and pressure,P, results in the following expression: K) ratios, but has limited efect on the compositions of AGo: -7.58052 - 2087.6587- 0.043rP. coexistingparagonite (i.e., paragonite becomes only slight- ly less K-rich). This free-energyrelation, combined with the ternary-feld- In addition, muscovite and paragoniteendmembers have spar and modified binary white mica activity models, leads beenassumed to interact with all other componentsin the to the following plagioclase-muscoviteNa-K geother- same manner as they interact with each other. In the mometric expression,with f in kelvins and P in bars: absenceof appropriate Wo parametersfor celadonite or phengitecomponents, the activity expressionshave there- T : 09 456A + 122308+ 27 320C+ l8 8lOD fore been taken as + 84738 + 28226F - 65407.0G - R?" ln cr" : RZ ln XM"+ (l - X*")t + 65305.4H 2087.6s87) .IW" +2XM"(W;- W)l + p(-0.0431 - 0.456A+ 0.66538+ 0.364C and +0.364D+ 2.tt2tc + 0.9699m/ (7.5805- 8.3147ln K" - 1.6544A R?nln a"" : RZ ln Xr" + (l - Xr^), .lW€+2X,^(W["-W.)] - 0.71048+ 10.3c+ 10.3D - lr4.r040c+ 12.5365rn. with (XM"+ Xp) <1. parameters Muscovite-paragoniteMargules have been where determined experimentally by Eugster et al. (1972) and Blencoe and Luth (1973). Chatterjee and Froese (1975) A:l-4XM"+5Xh"-2xil" adjusted the pressure dependenceof the Eugster et al. - 2X,^ + 4X?^ - 2Xl^, Margulesparameters based on a review of published data on the molar volumes of muscovite-paragonitesolid so- B : 2X-" - 4Xh" + 2I/3M"+ 4XP^ (injoules, lutions kelvins, and bars): - 5X7" + 2X7" - r, : - vlft" t9 456.0+ 1.65437 0.456tP C : 2XooX[, + 0.5Xo,X^, - Y"o + 2Xo,P^b wo^ : 12230.3+ 0.71047 + 0.6653P. - 0.5X^^X^b+ zXo,XA"X^b, Pigageand Greenwood (1982) subsequentlymodified the D: X6, - 2XooX6,+ 0.5Xo.XA"- zXo,XAb Eugsteret al. interaction parametersto take into account excessvolumes of mixing determined experimentally by - O.sxA^x^b- 2xo,x^nx^b, (1977). Blencoe Although the various Wrval.ues lead to E:2XooX'z[^ + O.sxo.xA, + Xo,x^"X^b significantly different equilibrium solvus compositions of - - muscovite-paragonite solid solutions at high pressures + O.sxA.xAb x^"x^b(xAn xno), (Blencoe,1977),Ihere is very little differencein calculated F: )fo^- 2X^bX1^+ 0.5Xo,X^^- X^nXobXo, equilibrium curves using the various W[" and Wo" pa- + O.sxA"xAb+ x^^x^b(x^" - xoo)' rameters(Cheney and Guidotii,1979; Pigageand Green- - wood, 1982).The Margules parametersof Chatterjeeand G: -0.5Xo,X^" - XA,Xo,(Xo, Xo^) Froese (1975) have been adopted here, but the revised - 2Xo,X2^,- o.sxo^X^b- Xo,X^nX^b, constants based on Blencoe's (1977) experimental data : -}.SXo,Xo^ - yield similar results when used to describe plagioclase- H * X^"Xo,(Xo, Xo,) muscovite Na-K exchangeequilibria. - x7^+ 2X"J&" - 0.5X",&o + xo,x^nx^b, PucrocusE-MuscovrrE THERMOMETER ln Ko : ln X'. + ln[X"o(2 - Xoo- Xo) The ternary albite and orthoclase solution models, white '(&o + X*)l mica activity relations,and high-temperatureendmember - ln Xr^ - lnlX",(2 - Xoo- Xo) thermodynamic data can be combined to formulate apla- gioclase-muscovitethermometer. The Na-K exchangere- (x^b + xo)], ll12 GREEN AND USDANSKY: PLAGIOCLASE-MUSCOVITE THERMOMETER

Table 1. Thermodynamicdata (Helgesonet al., 1978)

HO a | 146,t) v{48,t) w (cm3/ (kJ/mol) (J/mol.K) cx105 mol) Muscovite -5972.275 287.859 408.191 't10.374 -106.440" 140.71 Albite -3931.621 207j50 258.153 58.158 -62.802b 100.25 342.586 14.870 -209.844. Paragonite -s928.573 277.818 407 647 102.508 -110.625" 132.53 Alkali feldspar -3971.403 213.928 320.566 18.037 -125.29e 108.87 Nofej Upper temperature limits for the heat-capacitypowerJunction coefficientsare (a) 1000'K, (b)479'K, and (c) 140tr K.

X*, Xoo,and XA, are mole fractions of orthoclase,albite, The Penfold Creek occurrence represents complexly and anorthite in plagioclase,and XM. and Xp" are mole folded garnet, kyanite-staurolite, and sillimanite zone rocks fractions of muscovite and paragonitein white mica. The that outcrop within the Omineca geanticline and at the application ofthis equation requires an independent es- northern extremity of the Shuswap melamorphic com- timate or a geobarometric determination of P, because plex, British Columbia (Fletcher and Greenwood, 1979). the pressure dependenceof the calculated temperature The pelites are characterizedby some variation in com- varies from ll to 21" per kilobar depending upon the position between plagioclase grains within individual compositions of the coexisingminerals. specimens,both primary and secondarymuscovite, and two generationsof garnet that may representtwo pulses TrsrrNc oF GEoTHERMoMETER of the same metamorphism. The various garnet-biotite There are various ways in which the plagioclase-mus- thermometersyield temperaturesthat range from 383 to covite geothermometercould give spurious results. Sys- 7 l2C at an estimated pressureof 7000 bars; some sam- tematic errors might easily result from erroneous as- ples are characterizedby a bimodal distribution of cal- sumptionsconcerning the mixing relations in feldsparand culated temperatures,399 to 538"C and 599 to 695"C. mica solid solutions. Moreover, it is likely that the geo- The plagioclase-muscovitethermometer indicates equil- thermometer results will be extremely sensitive to ana- ibration temperaturesbetween 466 and 674{. The tem- lytical uncertaintiesin the Or composition of plagioclase peraturesofPenfold Creek rocks are in generalagreement feldspars.In view of these potential difficulties, it is im- with those predicted using the Hodges-Spear,Holdaway- portant to show that temperatures calculated with the ke, and Ganguly-Saxenaformulations of the garnet-bio- geothermometerare geologically reasonable.As the log- tite thermometers, but there is no apparent correlation ical way of testing the plagioclase-muscovitegeother- between the plagioclase-muscovitetemperatures and those mometer, temperatures have been calculated utilizing derived from any specificgarnet-biotite thermometer (Ta- compositionsof coexistingfeldspar-mica pairs from pelit- ble 2). ic metasedimentaryrocks and peraluminous granitoids, Pigage(1982) described metamorphic assem- for which independent temperature determinations and blagesin the Azure Lake area of the Shuswapmetamor- complete mineral analysesare available. phic complex, British Columbia, that are characteristicof the kyanite, kyanite-sillimanite, and sillimanite zones of Pelitic metasedimentaryrocks the Barrovian series.Pigage and Greenwood (1982) con- Plagioclaseand muscovite commonly coexist with gar- cluded that observed mineral assemblagesrepresent a close net and biotite in thesepelitic assemblages;thus, the pla- approachto equilibrium and used the intersection ofsev- gioclase-muscovitetemperatures may be compared di- eral mineral equilibria with the kyanite-sillimanite reac- rectly to those calculated with garnet-biotite thermometry. tion curve to estimate metamorphic pressuresof 5133 to Results of application of the plagioclase-muscovitether- 6731 bars. Their revised Ferry and Spear(1978) garnet- mometer to five metapelite suites are presentedin Table biotite geothermometeryields temperaturesup to 100 de- 2. Unfortunately, severalmodifications ofthe experimen- greeshigher than those calculatedwith the other calibra- tally calibrated garnet-biotite thermometer of Ferry and tions (579-603'C). The calculated plagioclase-muscovite Spear(1978) have been proposedto model nonideal mix- temperaturesof the Azure Lake rocks rangebetween 500 ing of garnet solid solutions (Holdaway andl;ue, 1977; and 601"C and, typically, are generally intermediate be- Hodges and Spear, 1982; Pigageand Greenwood, 1982; tween those predicted by the Holdaway-ke and Pigage- Ganguly and Saxena,1984). In some rocks, the different Greenwood garnet-biotite thermometers. garnet-biotite calibrations yield somewhat varied tem- Hodgesand Spear(1982) used pelitic schistsof the Mt. peratures,so that it is difficult to determine,a priori, which Moosilauke region, New Hampshire, to evaluate the in- calibration is appropriate. As a result, temperaturespre- ternal consistency of a variety of published geother- dicted by all these calibrations are presentedin Table 2 mometers and geobarometerswith the AlrSiO, invariant for comparison. point of Holdaway (1971). The rocks studied by Hodges GREEN AND USDANSKY: PLAGIOCLASE-MUSCOVITE THERMOMETER l1l3

and Spear contain andalusite andJor sillimanite, but do not contain kyanite. Sillimanite occurseither as subhedral prismatic grainsor as fibrolite intergrown with muscovite; o where present,andalusite occurs as subhedralto anhedral Y porphyroblasts. c The described textures suggestthat all F samples equilibrated at temperatures above the alumi- (r Albite-K Feldspar I -7.58057 nosilicate point (501'C) AG= triple and that the two samples -2 08 7.658 7 containingboth andalusiteand fibrolite (1468, 146D) may have equilibrated near the intersection of the muscovite 600 800 1000 1200 breakdown and andalusite-sillimanite reaction curves Temperature(oK) (545-590"C). In this case,only the calibration ofthe gar- net-biotite thermometerby Pigageand Greenwood(1982) Fig. l. Variationof RZln KDwith temperaturefor the Na-K producescomparable temperatures(561 and 621"C); all exchangereaction between plagioclase (all structuralstates) and muscovite(heavy line) based on thermodynamicdata of Helge- other calibrations generallyyield temperaturesof the alu- sonet al. (1978).Reaction curves for the exchangereaction cal- minosilicate triple point or lower. Application of the pla- culatedusing thermodynamic data of sanidine-highalbite and gioclase-garnet-AlrSiO,geobarometer (Ghent, I 976; Ghent microcline-lowalbite solid solutions are also shown. and Stout, l98l) produced preferred pressureestimates that rangefrom 2677 to 2918 bars.Two Mt. Moosilauke rocks have plagioclase-muscovitetemperatures (422 and, timates are basedupon phase-equilibriumconsiderations, 441"C) that suggestsubsolidus re-equilibration, but four such as biotite- oxide stability relations. Five peralu- samplesyield temperaturesabove the triple point (5 lzl- minous granitoid suites,chosen to test the thermometer, 588t) that are consistentwith the observedaluminosil- are discussedbelow, and estimatedeqirilibration temper- icate minerals. The compositions of plagioclaseand mus- atures are presentedin Table 3. covite in Mt. Moosilauke specimen 90A define a tem- Green and Usdansky (1984) described the petrology, peratureof 801"C,which reflectsthe anomaloustyhigh Or mineralogy, and crystallization conditions of two-mica content ofplagioclase porphyroblasts in this rock. epidote-bearinggranitoids of the Alabama tin belt. Gar- Zen ( I 98 I ) describedmineral assemblagesfrom slightly net-biotite and apatite-biotite F-OH geothermometrysug- calcic, chloritoid, garnet, and kyanite-staurolite zone pe- gested temperatures of 630 to 740C, whereas ilmenite litic rocks from the Taconic Rangein southwesternMas- and biotite stability relations suggestedtemperatures up sachusettsand adjacent areas of Connecticut and New to 770"C. Calculatedplagioclase-muscovite temperatures York. Comparison of the mineral assemblageswith hy- range from distinctly subsolidus(414C) to near liquidus drothermal phase-equilibrium data suggeststhat the ap- (740"C). proximate range of metamorphic conditions was 400- Kistler et al. (1981) describedtwo-mica granitesin the 600{ at pressuresprobably about 4 kbar. Garnet-biotite Ruby Mountains, Nevada, and reported garnet-biotite thermometry yields a similar rangeof temperatures(416- equilibration temperaturesbetween 365 and 505qCbased 607"C). Plagioclase-muscovitethermometry yields tem- on the Goldman and Albee (1977) calibration. Other cal- peraturesof 438-567t, which are in generalagreement ibrations of the garnet-biotitethermometer, however,give with garnet-biotite temperatures. temperatures of 527-725'C. The calculated plagioclase- Hodges and Royden (1984) described kyanite-grade muscovite temperaturesare higher than those predicted pelitic schistsfrom the Efiord area,northern Norway, that by all but the Pigage-Greenwoodgarnet-biotite thermom- exhibit abundant textural evidenceof retrogradere-equil- eter (Table 3) and reflect the coexistenceof K-rich mus- ibration under nonuniform pressure-temperaturecondi- covite with K-rich plagioclasein thesegranites. tions. The presenceof second-stagekyanite in all retro- Czamanskeet al. (1981) detailed the chemistry of rock- gradesamples indicates that the P-f path oftectonic uplift forming minerals of the Cretaceous-Paleocenebatholith occurred entirely within the kyanite stability field. The in southwesternJapan. They noted significant differences plagioclase-garnet-AlrSiO,geobarometer (Ghent, 1976; in the chemistry of primary and secondarymuscovites, Ghent and Stout, l98l), coupled with garnet-biotite ther- in agreementwith the characteristicsoutlined by Miller mometry, yields estimatedpressures that rangefrom 3 to et al. (1981). Temperaturesbased on two-feldspargeo- 14 kbar at temperatures of 471 to 853'C. Plagioclase- thermometry indicated that none of the suites faithfully muscovite temperatures (423-8i3C) exhibit the same retains solidus feldspar relations; perthitic alkali feldspar variation with pressureas that shown by the garnet-biotite in all sampleshas lost Ab. Ilmenite and biotite stability temperatures(Hodges and Royden, 1984; Fig. 8). relations in for-T space, however, suggesteqrilibration temperaturesof 68O-8l0'C. Plagioclase-muscovitetem- Peraluminousgranitoids peratures calculated using primary and secondary mica Plagioclaseand muscovite are also important constit- compositions are given in Table 3. uents ofthese granitic rocks, but only locally coexist with Bradfish (1979) investigated conditions of muscovite mineral pairs that have compositions applicable to geo- crystallization in the Teacup Granodiorite, Arizona, and thermometry. As a result, most available temperaturees- concluded that the mica in all facies of the intrusion was ll14 GREENAND USDANSKY:PLAGIOCLASE-MUSCOVITE THERMOMETER

Table2. Plagioclase-muscovitetemperatures of peliticmetamorphic rocks

Sample Xo^ Xoo Xo, xru Xu" F-S G-S P (bars) Penfold Creek (Fletcherand Greenwood, 1979) 5 0.359 0.635 0.005 0.146 0.786 596 560 668 617 599 674 7 000 o 0.292 0.705 0.003 0.238 0.700 s34 596 553 549 523 565 7 000 7 0.302 0.694 0.003 0.167 0.766 648 676 663 626 577 561 7 000 8 0.260 0.738 0.002 0.238 0.695 572 628 600 576 545 488 7 000 0.290 0.708 0.002 0.241 0.693 521 577 538 536 510 505 7 000 10 0.310 0.685 0.003 0.202 0.721 529 594 551 546 514 574 7 000 11 0.290 0.706 0.003 0.186 0.742 539 599 573 553 525 558 7 000 12 0.202 0.794 0.003 0.222 0.714 602 630 617 596 568 508 7 000 13 0.206 0.789 0.005 0.236 0.704 625 677 640 613 582 585 7 000 14 0.282 0.712 0.005 0.163 0.762 631 700 659 617 619 625 7 000 15 0.129 0.868 0.003 0.202 0.726 640 674 650 621 623 466 7 000 AzureLake (Pigage, 1982) 373 0.277 0.718 0.004 0.170 0.763 545 602 573 557 545 565 5 678 121 0.314 0.681 0.004 0.160 0.772 535 651 603 578 535 581 5 546 367 0.394 0.602 0.003 0.140 0.791 533 608 571 548 535 580 5 783 82 0.278 0.716 0.004 0.160 0.772 525 626 549 542 524 552 5 133 398 0.207 0.788 0.005 0.209 0.734 541 652 557 553 544 562 5 528 492 0.227 0.766 0.003 0.209 0.734 543 618 579 569 534 505 5 979 223 0.337 0.658 0.004 0.150 0.782 533 621 570 556 529 601 6 035 2-376 0.277 0.716 0.005 0.150 0.782 555 641 585 564 547 592 5 828 2-13 0.309 0.685 0.005 0.140 0.791 523 656 552 542 535 600 s 396 74 0.358 0.638 0.003 0.120 0.792 581 658 612 582 578 569 6 731 5V 0.256 0.738 0.005 0.150 0.782 557 682 581 565 549 579 5 828 40 0.240 0.758 0.003 0.179 0.754 540 580 ss8 5s3 530 500 5 264 Mt. Moosilauke(Hodges and Spear, 1982) 788 0.116 0.876 0.006 0.245 0.713 468 533 477 499 488 514 3 918 80D 0.280 0.714 0.006 0.172 0.776 489 565 512 518 513 588 3 444 90A 0.141 0.831 0.026 0.237 0.721 486 568 497 513 504 801 3 801 92D 0.108 0.888 0.003 0.252 0.701 491 557 501 518 508 422 3 493 145E 0.089 0.90s 0.004 0.231 0.727 498 567 507 524 524 441 3 167 1468 o.24't 0.753 0.005 0.190 0.767 519 621 538 540 527 531 2722 146D 0.243 0.750 0.005 0.190 0.762 464 561 480 498 486 532 2677 TaconicRange (zen, 1981) 463-1 0.190 0.804 0.005 0.280 0.687 509 568 524 530 544 538 4 000 339-1 0.266 0.729 0.004 0.237 0.721 416 500 447 461 458 547 4 000 356 0.138 0.853 0.008 0.225 0.727 469 508 492 501 479 567 4 000 103-1 0.120 0.875 0.005 0.187 0.733 460 515 475 495 5't7 484 4 000 103-2 0.300 0.694 0.004 0.170 0.761 510 607 545 531 561 548 4 000 234-1 0.368 0.630 0.002 0.133 0.815 452 495 473 489 479 472 4 000 14-1 0.247 0.750 0.002 0.200 0.763 443 474 460 482 469 438 4 000 1052-2 0.059 0.930 0.010 0.187 0.780 505 548 520 527 516 535 4 000 Efjord(Hodges and Royden, 1984) 8B 0.541 0.454 0.004 0.120 0.785 712 853 775 673 653 873 13000 8C 0.143 0.853 0.003 0.152 0.737 491 527 507 516 466 423 3 600 12E 0.207 0.788 0.004 0.101 0.776 485 528 503 512 475 451 3 300 18A 0.29s 0.699 0.004 0.101 0.751 476 573 523 507 492 546 6 000 21A 0.302 0.694 0.003 0.166 0.703 471 634 499 503 489 524 4 900 21C 0.312 0.680 0.008 0.075 0.780 554 640 597 565 533 675 I 000 21F 0.252 0.744 0.004 0.11 1 0.781 557 615 584 565 526 546 7 000 Notes: Garnet-biotitethermometer calibrations as follows: F-S, Ferry and Spear (1978); P-G, Pigage and Greenwood (1982); H-S' Hodges and Spear (1982); H-L, Holdaway and Lee (1977); and G-S, Ganguly and Saxena (1984). f = plagioclase-muscovitegeothermometer.

probably secondary and resulted from subsolidus reac- assemblagesconsisting of muscovite, calcite, quartz, ep- tions involving a coexistingaqueous phase. Predicted pla- idote, plagioclase,and alkali feldspar that record temper- gioclase-muscovitetemperatures (478 and 530'C) are aturesof4l5 to 435t at 3500 bars.The measuredfrac- compatible with such an interpretation and with temper- tionation ofAb componentbetween coexisting plagioclase atures predicted by some but not all garnet-biotite ther- and microcline yields temperaturesof 41 5 to 454qCat the mometers (Table 3). same pressure,although Ferry (1978, 1979) did not pre- Ferry (1978, 1979)discussed the origin ofhydrother- sent data that would p€rmit evaluation of the other two mally altered feldspars and mica in samples of granitic feldspar endmember equilibria in order to determine rock from south-central Maine. He reported alteration whether stableequilibrium has been achievedbetween all GREENAND USDANSKY:PLAGIOCLASE-MUSCOVITE THERMOMETER lll5

Table3. Plagioclase-muscovitetemperatures ot peraluminousgranitoids

Sample xh xe Xo, X"" Xt" F-S P-G P (bars) Alabamatin belt(Green and Usdansky,1984) ATB.67 0.337 0.644 0.008 0.070 0.753 63s 863 689 618 740 n.a. 698 8000 ATB-58 0.354 0.635 0.010 0.070 0.753 n.a. n.a. n.a. n.a. n.a. 760" 740 8000 RF-20s 0.017 0.969 0.006 0.060 0.780 n.a. n.a. n.a. n.a. n.a. 463b 414 6000 RF-3s o.224 0.754 0.008 0.060 0.785 n.a. n.a. n.a. n.a. n.a. 584b 536 6000 RubyMountains (Kistler et al.,1981) 11-66A 0.008 0.866 0.125 0.068 0.828 538 725 548 550 630 n.a. 718 5000 31-66D 0.010 0.905 0.08s 0.064 0.836 514 607 527 533 609 n.a. 67'l 5000 TeaCup Granodiorite (Bradfish, 1979) LN.1 0.126 0.863 0.012 0.065 0.776 501 774 512 s25 615 n.a. 530 5000 SN-19 0.146 0.844 0.008 0.055 0.770 590 916 607 586 720 n.a. 478 5000 Japanesebatholiths (Czamanske et al., 1981) 6510s 0.245 0.739 0.014 0.020 0.769 n.a. n.a. n.a. n.a. n.a. n.a. 485 s000 T25s 0.017 0.969 0.006 0.060 0.780 n.a. n.a. n.a. n.a. n.a. n.a. 556 5000 T173bc 0.224 0.754 0.008 0.060 0.785 n.a. n.a. n.a. n.a, n.a. n.a. 669 5000 T173Br 0.354 0.635 0.010 0.070 0.753 n.a. n.a. n.a. n.a. n.a. n.a. 651 5000 T76 0.017 0.969 0.006 0.060 0.780 n.a. n.a. n.a. n.a. n.a. n.a. 669 5000 T133r 0.224 0.754 0.008 0.060 0.785 n.a. n.a. n.a. n,a, n.a. n.a. 678 5000 0.354 0.635 0.010 0.070 0.753 n.a. n.a. n.a. n.a. n.a. n.a. 616 s000 Augusta,Maine (Ferry, 1978, 1979) 694s 0.122 0.871 0.005 0.040 0.U2 n.a. n.a. n.a. n.a. n.a. 403" 376 3500 698s 0.137 0.8s8 0.004 0.040 0.816 n.a. n.a. n.a. n.a. n.a. 447" 364 3500 821s 0.170 0.817 0.012 0.040 0.791 n.a. n.a. n.a. n.a. n.a. 415d 474 3500 838s 0.i73 0.816 0.009 0.040 0.816 n.a. n.a. n.a. n.a. n.a. 423d 446 3500 276s 0.213 0.780 0.006 0.030 0.798 n.a. n.a. n.a. n.a. n.a. 435d 401 3500 787s 0.096 0.900 0.003 0.030 0.798 n.a. n.a. n.a. n.a. n.a. 442. 31s 3s00 Notesj Garnet-biotitethermometer calibrations as follows: F-S, Ferry and Spear (1978); P-G, Pigage and Greenwood (1982); H-S, Hodges and Spear (1982); H-L, Holdaway and Lee (1977); and G-S, Ganguly and Saxena (1984). Other temperature estimates (7-) based on (a) biotite-apatite F-OHexchange equilibria, (b) biotite-ilmenitestability relations,(c) twGfeldspar albite geothermometer,and (d) muscovite-4alcite-epidote-plagioclase- alkali leldspar equilibria; n.a., not applicable. Where indicated, analyses represent (c) core or (r) rim of primary (p) muscovite, or secondary (s) muscovite. I plagioclase-muscovitegeothermometer.

feldspar endmembers. The calculated plagioclase-mus- covite geothennometer extends the range of pelitic and covite temperaturesare compatible with those predicted peraluminous rocks for which equilibration temperatures by other geothemometric techniques(Table 3). may be estimated.

Suprulnv AND coNcLUsroNS AcxNowr,nncMENTs Available thermochemical data have been used to cal- We are grateful to J. M. Rice, S. K. Saxena,and J. C. Stormer, ibrate the Na-K exchangereaction between plagioclase Jr., for their critical reading ofour manuscript and valuable con- and muscovite as a geothermometer.The formulation of structive suggestionsfor its improvement. Stimulating discus- the plagioclase-muscovitegeothernometer accounts for sions with R. J. Donahoe and C. M. Lesher about the thermo- dynamic modeling are greatly appreciated.We also thank S. Miller high and low structural statesof feldspar and utilizes pla- and S. Simpson who helped in the preparation of the manuscript. gioclaseactivity-composition relations expressedin tenns This work was supported by a gant from the School of Mines of a ternary subregularsolution model. Binary mica so- and Energy Development at the University of Alabama. lution modelshave beenmodified to accountfor increased K/(K + Na) in the muscovite component associatedwith RrrnnnNcns increased celadonite andlor phengite substitution. The Bailey, S.W. (1984) Crystal chemistry of true micas. In S.W. geothermometer has been applied to pelitic metasedi- Bailey, Ed. Reviews in Mineralogy, 13, 13-60. mentary rocks and peraluminous granitoids that crystal- Blencoe,J.G. (1977)Molal volumes ofsyntheticparagonite-mus- Mineralogist, 62, l20Ll2l5. lized at temperaturesof 4 | 5-7 25"C and pressuresof 2.5- covite micas. American Blencoe,J.G., andLuth, W.C. (1973)Muscovite-paragonite solvi 13.0kbar. The predicted equilibration temperaturesusing at2,4, and 8 kb pressure.Geological Society of America Ab- plagioclase-muscovite pairs generally fall within the range stractswith Programs, 5, 553-554. of equilibration temperaturesestimated using other geo- Boles,J.R., and Coombs,D.S. (1975)Mineralreactions in zeolitic thermometric techniques(e.g., garnet-biotite thermome- Triassic tuff, Hokonui Hills, New Zealand. Geological Society of America Bulletin, 86, 163-173. try). Observed discrepanciesbetween calculated temper- Bradfish, L.J. (1979) Petrogenesisofthe Tea Cup granodiorite, atures appear to be related to re-equilibration of muscovite Pinal County, Arizona. M.S. thesis, University of Arizona, and/or plagioclase.Application of the plagioclase-mus- Tucson. lll6 GREEN AND USDANSKY: PLAGIOCLASE-MUSCOVITE THERMOMETER

Chatterjee,N.D. (1973)Low-temperature compatibility relations of rock-forming minerals. American Journal of Science,2734, of the assemblageparagonite + quartz and the thermodynamic 229 p. statusof the phaserectorite. Contributions to Mineralogy and Henley, K.J. (1970) Application of the muscovite-paragonite Petrology,42, 259-27l. geothermometerto a staurolite-gradeschist from Sulitjelma, Chattefee, N.D., and Froese,E. (1975) A thermodynamic study north Norway. Mineralogical Magazine, 37, 692-702. of the pseudobinaryjoin muscovite-paragonitein the system Hock, V. (1974) Coexisting phengite,paragonite and NaAlSirO8-KAlSijOE-AlrO3-SiOr-HrO.American Mineralo- in metasedimentsof the Mittlere Hohe Tauern, Austria. Con- gist,60,985-993. tributions to Mineralogy and Petrology, 43,261-273. Chatte{ee, N.D., and Johannes,W. (1974) Thermal stability and Hodges,K.V., and Royden,L. (1984)Geologicthermobarometry standard thermodynamic properties of synthetic 2M, musco- of retrograde metamorphic rocks: An indication of the uplift vite, KAlrSi.AlOro(OH)r. Contributions to Mineralogy and Pe- trajectory ofa portion ofthe northern ScandinavianCaledon- trology,48, 89-l 14. ides. Journal of GeophysicalResearch, 89, 707 7 -7 090. Cheney,J.T., and Guidotti, C.V. (1979) Muscovite-plagioclase Hodges,K.V., and Spear,F.S. (1982) Geothermometry, geoba- equilibria in sillimanite + quartz-bearingmetapelites, Pttzzle rometry and the AlrSiO5 triple point at Mt. Moosilauke, New Mountain area, Northwest Maine. American Journal of Sci- Hampshire. American Mineralogist, 67, I I l8-l 134. ence,279, 4Il-434. Hofer, E. (1978) On the "late" formation of paragoniteand its Czamanske,G.K., Ishihara,S., and Atkin, S.A.(1981) Chemistry breakdown in pelitic rocks of the Southern Damara orogen of rock-forming minerals of the Cretaceous-Paleocenebatho- (Namibia). Contributions to Mineralogy and Petrology, 67, lith in southwesternJapan and implications for magmagenesis. 209-219. Journal of GeophysicalResearch, 86, l0 43 l-l9 469. Holdaway, M.J. (197{) Stability of andalusite and the alumi- Day, H.W. (1973)The upperstability limit of muscovite+ quafiz. nosilicate phase diagram. American Journal of Science,271, American Mineralogist, 58, 255-262. 97-r3r. Eugster,H.P., Albee, A.L., Bence, A.E., Thompson, J.8., and Holdaway, M.J., and Lee, S.M. (1977) Fe-Mg cordierite stability Waldbaum, D.R. (1972)The two-phaseregion and excessmix- in high-gradepelitic rocks based on experimental,theoretical ing properties of paragonite-muscovitecrystalline solutions. and natural observations. Contributions to Mineralogy and Journal of Petrology, 13, 147-179. Petrology,63, L75-193. Ferry, J.M. (1978) Fluid interaction between granite and sedi- Iiyama, J.T. (1964) Etude des r6actions d'echanged'ions Na-K ment during metamorphism, south-centralMaine. American dansla s6riemuscovite-paragonite. Bulletin de la Soci6t6Fran- Journalof Science.278. 1025-1056. gaisede Min6ralogie et de Cristallographie,87, 532-541. -(1979) Reaction mechanisms, physical conditions, and Katagas,C., and Baltatzis, E. (1980) Coexistingceladonitic mus- masstransfer during hydrothermal alteration of mica and feld- covite and paragonitein chlorite zone metapelites.Neues Jahr- spar in granitic rocks from south-centralMaine. Contributions buch fiir Mineralogie, Abhandlungen,H.5, 206-214. to Mineralogy and Petrology, 68, 125-139. Kistler,R.W., Ghent,E.D., and O'Neil, J.R. (1981)Petrogenesis Ferry, J.M., and Spear,F.S. (1978) Experimental calibration of of garnet, two-mica granitesin the Ruby Mountains, Nevada. the partitioning of Mg and Fe2* between biotite and garnet. JournalofGeophysical Research, 86, 10591-10 606. Contributions to Mineralogy and Petrology, 66, Il3-117. Kotov, N.V., Mil'kevich, R.I., and Turchenko, S.I. (1969) Pa- Fletcher, C.J.N., and Greenwood, H.J. (1979) Metamorphism leothermometry of muscovite-bearing metamorphic rocks based and structure ofPenfold Creek area,near Quesnel,British Co- on X-ray and chemical analysis of muscovite. Doklady Aka- lumbia. Journal of Petrology, 20, 7 43-7 94. demii Nauk SSSR,184, I 180-l 182. Ganguly, J., and Saxena,S.K. (1984) Mixing properties of alu- Labotka, T.C., Papike, J.J., Vaniman, D.T., and Morey, G.B. minosilicate garnets:Constraints from natural and experimen- (1981) Petrology of contact metamorphosedargillite from the tal data, and applications to geothermometry.American Min- Rowe Formation, Gunflint Trail, Minnesota. American Min- eralogist,69,88-97. eralogist,66, 79-86. Ghent, E.D. (1976) Plagioclase-garnet-A1rSiO,-quartz:A poten- Miller, C.F., Stoddard, E.F., Bradfish, L.J., and Dollase, W.A. tial geobarometer-geothermometer.American Mineralogist, 6 1, (1981) Composition of plutonic muscovite: Genetic implica- 7r0-7 14. tions. Canadian Mineralogist, 19, 25-34. Ghent, E.D., and Stout, M.Z. (1981) Geobarometryand geo- Newton, R.C., Charlu, T.V., and Kleppa, O.J. (1980) Thermo- thermometry of plagioclase-biotite-garnet-muscoviteassem- chemistry of the high structural stateplagioclases. Geochimica blages.Contributions to Mineralogy and Petrology,76,92-97 . et CosmochemicaActa, 44,933-941. Goldman, D.S., and Albee, A.L. (1977) Correlation of Fe/Mg Odom, I.E. (1984) and celadonite minerals. In S.W. partitioning betweengarnet and biotite with OrE/Or6partition- Bailey, Ed. Reviews in Mineralogy, 13,545-572. ing between quartz and magnetite. American Journal of Sci- Pigage, L.C. (1982) Linear regressionanalysis of sillimanite- ence,277 , 7 50-767. forming reactionsat Azure Lake, British Columbia. Canadian Grambling, J.A. (1984) Coexisting paragoniteand quartz in sil- Mineralogist, 20, 349-378. limanite rocks from New Mexico. American Mineralogist, 69, Pigage,L.C., and Greenwood, H.J. (1982) Internally consistent 79-87. estimatesof pressureand temperature:The stauroliteproblem. Green, N.L., and Usdansky, S.I. (1984) Mineral chemistry and American Journal of Science,282, 943-969. crystallizationconditions of Alabama tin belt granitoids. Geo- Price, J.G. (1985) Ideal site mixing in solid solutions, with ap- logical Society of America Abstracts with Programs,16, 142. plication to two-feldspar geothermometry. American Miner- - (1986) Ternary-feldsparmixing relations and thermoba- alogist,70, 696-701. rometry.American Mineralogist,71, 1100-1108. Saxena,S.K. (l 973) Thermodynamicsof rock-forming crystalline Guidotti, C.V. (1984) Micas in metamorphic rocks. In S.W. Bai- solutions. Springer-Verlag,New York. 188 p. ley, Ed. Reviews in Mineralogy, 13,357-468. Seck, H.A. (l97la) Koexistierende Alkalifeldsplte und Plagio- Haselton, H.T., Hovis, G.L., Hemmingway, B.S., and Robie, klase im System NaAlSi3OE-KAlSi,O8-CaAIrSitOr-HrObei R.A. (1983) Calorimetric investigation of the excessentropy Temperaturen von 650t bis 900'C. Neues Jahrbuch liir Mi- ofmixing in analbite-sanidinesolid solutions:Lack ofevidence neralogie,Abhandlungen, I I 5, 3 I 5-345. for Na, K short-range order and implications for two-feldspar -(197lb) Der EinfluB des Drucks auf die Zusammenset- thermometry. American Mineralogist, 68, 398-4 I 3. zung koexistierenderAlkalifeldspite und Plagioklaseim Sys- Helgeson,H.C., Delany, J.M., Nesbitt, H.W., and Bird, D.K. tem NaAlSirO8-KAlSi3OE-CaAlrSirOs-HrO.Contributions to (1978)Summary and critique ofthe thermodynamic properties Mineralogy and Petrology, 31, 67-86. GREENAND USDANSKY:PLAGIOCLASE-MUSCOVITE THERMOMETER LttT

Talantsev,A.S. (l 97 1) The plagioclase-muscovitegeothermom- Wohl, K. (1953) Thermodynamic evaluation of binary and ter- eter.Doklady Akademii Nauk SSSR,196, I193-l 195. nary liquid systems.Chemical EngineeringProgress, 49,218- Thompson, A.B. (l 974) Calculation of muscovite-paragonite- 2r9. alkali feldspar phase relations. Contributions to Mineralogy Znn,E-an. (1981) Metamorphic mineral assemblagesof slightly and Petrology, 44, 173-194. calcic pelitic rocks in and around the Taconic allochthon, Tracy, R.J. (1978)High grademetamorphic reactionsand partial southwesternMassachusetts and adjacent Connecticut and New melting in pelitic schist,west-central Massachusetts. American York. U.S. Geological Survey ProfessionalPaper I I 13. JournalofScience, 278, 150-178. Wise, W.S., and Eugster,H.P. (1964) Celadonite:Synthesis, ther- mal stability and occurrence.American Mineralogist, 49,l03l- MeNuscnrprREcETVED JuNs 6, 1985 1083. Meuuscnrpr ACCEPTEDMav 16, 1986