The Reaction Albite = Jadeite * Quartz Determined Experimentally in The

American Mineralogkt, Volume 65, pages 129-134, 1980

The reaction albite = jadeite * quartz determinedexperimentally in the range600-1200.c

TrMorHy J. B. Hou-eNo Departmentof the GeophysicalSciences, University of Chicago Chicago,Illinois 60637

Abstract

The reaction

NaAlSirOu+ SiO2 :NaAlSirOs jadeite quaftz high albite has been determinedby direct experimentin the temperaturerange 6fi)-1200"C, and the P- T line can be describedby the equation P: 0.35+ 0.0265?("C) + 0.50kbar The brackets,when comparedwith the gas-apparatusdeterminations of this reactionby Hays and Bell (1973)and by Birch and LeComte (1960),show clearly that pressurecorrections are vanishingly small when the NaCl pressurecell is used with the piston-+ylinder device if piston-out methodsare used.The slope(dP/df : 26.5bar oC-t), io conjunction with tabulated entropiesfor quartz andjadeite, leadsto a value for the entropy ofsynthetic high albite. The disorderingentropy of albite is (14.2+2.4J K-' mol-'), and the enthalpy is (12950=3320J mol-t). The disorderingenthalpy is in excellentagreement with recentcalorimetric measurements on Amelia albite and its heat-treatedmodification. The best valuesfrom phase equi- libria for the 298 K, I bar, thermochemicaldata for synthetichigh albite are AJI3I".rt*: -3922170+ 4300J mol-' S&r., o- :221.6 + 2.5J K-t mol-' The lack of curvatureof the reaction in the P-T plane imFlies that high albite doesnot undergo ordering down to 600'C, at least within the duration time of the experiments.This is also conflrmed by the X-ray patternsof albites from runs at 600"C.

Introduction which high_temperature calorimetry (Holm and The stability of the end-member feldspars is of Kleppa, 1968) predicts a slope of 27.6 bar deg-', as fundamental importance to petrology, yet consid- shown later. erable uncertainty exists as to the P-Tlocation of the The importance of this reaction to petrology can- albite to jadeite * quartz curve at high temperatures. not be overstated, as it forms the basis for a number In the temperature range 500-600"C the experi- of geobarometric inferences on the stability of mental location is precise (Newton and Smith, 1967; plagioclase-bearing assemblagesat depth (Boettcher, Johannes et al., l97l; Hays and Bell, 1973), but at l97l\. higher temperatures the slope and position of the This study defines closely the P-Z stabitty of high curve are only loosely constrained by the reversal in- albite relative to jadeite and quartz, using the precise tervals determined by Birch and LeComte (1960) and pressure control aforded by the low-friction NaCl Bell and Roseboom (1969). Boettcher and Wyllie pressure cell with the piston-cylinder apparatus. The (1968) published brackets at 600oC and 800"C which precision made possible by this technique allows ac- demand a lower dP/d'T slope (19.5 bar deg-') than curate determination of the standard thermodynamic expected for the breakdown ofdisordered albite, for properties ofhigh albite. 0003-004X/80/0102-0129$02.00 r29 130 HOLI-AND: REACTION ALBITE : TADEITE + QUARTZ

Startingmaterials and experimentaltechnique pressure,ensures initiation of the run under conditions of retreating piston. Occasionallyit was neces- High albite was synthesized hydrothermally at sary to bleed off a little oil pressureduring the final 7 50"C, 2 kbaL from a mixture of natural vein quartz stagesof heating to prevent pressureovershoot. and NaAlSi,O. glass(prepared by P. A. M. Ander- Chromel-alumel thermocouples, in AtSiMag son) in a l: I molar ratio. The X-ray pattern is that of sleeves,with the junctions protected by a layer of high albite,with A2d(CuKa) : 1.90ofor (l3l)-(l3l) alundum cement, were used to measure tem- (Goldsmith and Laves, 1954).Natural white jadeite peratures.No pressurecorrection was applied to the from a vein in serpentinite,donated to R. C. Newton thermocouple readings. The reversal mixture was by R. G. Coleman,was used;it containsas principal made up of the crystallinestarting materialshigh al- impurities 0.13 weight percentCaO, 0.12weight per- bite, jadeite, and quartz in reactingproportions. The cent MgO, and 0.45weight percentFerO, (Coleman, mixture was ground repeatedly under acetone to 1961). thoroughly homogenizethe reactants,and was then An end-loaded piston-cylinder apparatuswith a baked at 350oC overnight to burn off any organic 0.75" diameter cylindrical pressurechamber and residue from the acetone. Following Boettcher and NaCl pressuremedium was used for all runs. The Wyllie (1968)in the successfuluse of interstitial melt chamber contained a smooth,but not polished,steel as a flux, the chargewas moistenedby breathing on liner and was lubricated by dry MoS, powder. Pis- the powder before sealingin platinum capsules.The ton-out type runs were made,g5ing pistonswith bev- small amount of interstitial water-undersaturated elled tops cappedwith pyrophylli{s ssalinggaskets as melt was found sufficient to produce 100 percent re- describedin Holland (1979a),by the following proce- action in many runs. Direction of reaction was de- dure. The cold assembly was brought up to a pre- tectedby changesof more than 15 percentin relative determinedpressure of 3.5 kbar or more below the fi- peak intensitieson the diffractometertrace. Runs at nal desiredrun pressureand was then heatedto the 600"C (from Holland,1979a) were conductedunder required temp€rature for the run. The large thermal HrO-excessconditions. expansion of the NaCl assembly during heating, a and final run pre-calibrated function of temperature Results jadeite Table 1. Experimental results on the reaction jadeite + quartz : Kinetics of the albite to * quartz reaction high albite in the presenceof small amounts of melt are such that complete reaction occurred in less than 3 hours 1200oand,24 hours at 800"C. Runs at and above -o^ at Ti-me Resul t* (hours) ll00"C cannot be consideredreversed, as the albite breaks down to jadeite + quartz during the approach (Table l). However, 4 600 16.O 24.O A to the final P-T conditions 6 600 L6.5 8.O JQ growth of albite is extremely rapid (Newton, in Jo- 24 800 2l .o t3.25 resultscertaidy represent 23 800 22.O 23.75 JQ hanneset al.,l97l) and the equilibrium. Table 1 and Figure I display the experi- 28 900 23.5 23.5 A straight-line fit (by eye) 25 900 24.25 22.75 NR mental results. The best 27 900 25.O 2L.50 JQ through the brackets can be described by P (kbar) : 18 1000 26.o 23.25 A 22 1000 27.5 zz, I ) JQ 0.0265TC+0.35+0.50.

l3 1100 28-O 3.25 A Discussion L4 I 100 29.O 5.2 A to I 100 30.0 o.2 JQ Pressurecalibration of piston cylinder apparatus 11 I 100 30.o -** JQ The use of the NaCl pressure medium in recent L5 1200 31.O A L6 1200 32.O 5.4 NR years has significantly improved the pressuredeter- L7 1200 33.O 4.25 JQ mination in piston-cylinder experiments (Johannes et al., l97l; Mirwald et al., 1975;Johannes, 1978; A - albite growth; JQ - jadeite + quartz growthi Danckwerth and Newton, 19781,Holland, 1979a).It NR no reaction. is instructive to comparethe presentresults at 600oC ** Run was quenched as soon as desired P,T conditions with the gas-apparatusdetermination by Hays and had been reached. Bell (1973)and in the range 800-1000'C with the de- HOLI,AND: REACTION ALBITE : IADEITE + QUARTZ 13r

termination by Birch and Lecomte (1960),also performed with gas pressure medium. At 600'C the pressuresobtained in this study, 16.0-16.5kbar, agree well with the Hays and Bell value (16.4+0.5 JADEITE kbar). The bracketsin the range 800-1000"C all tie E. +QUARTZ <.? within the reversal interval of Birch and LeComte ao- and can be regardedas a refinement :< of their results. I Although Birch and LeComteused natural low albite l,! ALBITE in their starting material, the product albite which t/) t/) was nucleated and/or grown in their runs was with- t! out doubt disorderedhigh albite. On the other hand o-1E the lessstable low albite in their runs probably broke down to jadeite + qtrurtz at slightly lower pressures than thoserequired for the breakdownof high albite. 500 700 900 1100 1300 The equilibrium curve for high albite : jadeite + TEMPERATURE- DEGt qtartz should therefore be expected to lie within or Fig. l. P,7 brackets on the jadeite + quartz to albite reaction. at marginally higher pressures than the Birch and Filled squares denote growth of jadeite + qtJaftz, open squares LeComte reactionbrackets. The fact that the present denote growth of albite, and partially-fiIled squares denote no resultsall lie within the Birch and LeComte intervals apparent reaction. indicates that the pressuresdetermined with the NaCl cell, as used here, are not overestimates, and extrapolatingthe c-quartz heat contentdata in Robie that subtractive friction corrections yxli5hingly ur" et al. (1978),the one-barenthalpy for the reactionja- small. Indeed, any correction due to friction would deite + quartz: albite, involving a-qtartz and high be much less at these higher temperatures (800- albite, becomes11405+2670 I at 964 K. With mean 1200'C) than at 600oC, where it was shown to be thermal expansionsand compressibilities from Table negligible (this work; Hays Bell, and 1973;Johannes 2 and the equilibrium pressureof 18.7kbar, the slope et al., l97l). Johannes(1978) reported pressure a of mav be calculatedfrom: lessthan 16 kbar for a piston-out reversalat 600oC, in contrastto the 16.25=0.25kbar reportedhere. Al- dP/dr: ,..+ Wv- r^@ryd4fr^v though both studiesused the NaCl pressuremedium, [or, 1," the methodsof arrival at the final pressureand tem- The result,27.6!1.5 bar "C-', is in good agreement peraturediffered. It is apparentthat with this method with the measuredslope, 26.5+1.5 bar oC-'. no pressurecorrection is necessary,and this redeter- The entropy of synthetichigh albite may be calcu- mination is sufficiently preciseto be used as a pres- lated from the experimentalbrackets, using the measure calibration for alternativepressure cells such as sured volumes and heat capacities(in Robie et al., Pyrex, talc, or soft glass.Results in this laboratory suggestthat the 7-10 percent subtractive correction jadeite, commonly applied to Pyrex-talc assembliesabove Table2. Standardttr"-odffifproperties for albite ll00"C with piston-in runs is valid.

s3o* A,Hloo dvoxlo5 pvo*ro6 Thermodyn amic cons ide r atio ns .t *-lLJr-t .l roitr J b".-l K-l J bar-2

The slope for the reaction between albite and ja- Jadeite 60 4 133,47 -3O2427O I +r.25 +4180 deite + quartz can be readily calculatedfrom high- to High albiEe IOO 43 221 6 -3922170 L5.94 temperatureoxide melt solution calorimetry.For low {O. 09 +2.5 +4300

albite, the data of Hlabse and Kleppa (1968) yield Low albite 100 07 2O7.4O -3935120 15.94 LHn** : -2385+2090 J, for the reaction from ja- J{ 13 r{ 40 +3415 deite + quartz.Recent measurements (R. C. Newton, Quartz 22.688 4L46 - 910700 5.18 +o o01 +o 20 +1000 personalcommu nication) on the enthalpy di-fference betweenlow albite and albite heat-treatedat 20 kbar, Thermal exFension data taken from Skinner (1966), conpressibil- : ity deEe Eaken frm Birch (1966) Al1 lhermodynamic daEa are frm 1200'C yield Aflu,*,o.. 14580+1660J mol-' at 970 Robie et al (1978) excepr for high albite and jadeiEe dara which K. Taking account of the a-B quafiz transition by are from this study, 132 HOLLAND: REACTION ALBITE: JADEITE + QUARTZ

1978), thermal expansion and compressibility data in times of the order of three weeks.Note that the (Table 2). Solving the relation: low-temperature acid calorimetric results (Wald- baum and Robie, l97l; Kracekand Neuvonen',1952) yield somewhatlow enthalpiesof disorder. LH.,n".,- ZASrr..,: - dr- r- ar+ tv arl Applications [/-o.o I:"^+ 1," The improved precision in the location of the albite breakdowncurve to high temperaturesallows for by plotting the right side against absolute temper- more reliable geobarometryof sodic pyroxenesthan ature yields ASrnr,,as slope and LH2s8'as intercept. was possible before. An important application is in The derived entropy of albite is 221.6+2.5 J K-l the estimation of the stability of jadeitic pyroxenes mol-'. Comparing this value with that for low albite coexisting with quartz in natural rocks. As an ex- gives the disordering entropy as 14.2+2.4J K-' per ample we consider the metamorphic rocks of the mole of albite, which is somewhat lower than the Sezia-Lanzo zone of the Western Alps, and in partic- maximum configurationalterm -4R (0.25 ln 0.25 + ular the assemblagepyroxene I quartz, described by 0.75 ln 0.75) : 18.7 J K-' mol-'. X-ray crystallo- Reinsch (19-17).Reinsch estimated680'C as an ap- graphic site-occupancystudies show that natural low propriate temperaturefor the equilibration of the py- albite is not fully ordered and that high albite may roxene, which contained 8l percent of the jadeite not be completely disordered (Holm and Kleppa, end-member.Direct applicationof the presentresults 1968).From the observeddistribution of Al and Si yield a pressureof 18.4 kbar below which pure ja- on the tetrahedralsites in low and heat-treatedalbite, deite + qrartz would break down. However, the ef- the configurational disordering entropy is calculated fects of 19 percent diluent in the natural pyroxene as 16.0J K-' mol-' (Hotm and Kleppa, 1968).The can be calculatedand will result in a destabilization agreementwith the value, 14.2!2.4 J K-t, derived of the quartz * pyroxene assemblageby an amount from the slopeof the experinental study is encourag- = -(RT/LV")lno,a ing. AP By comparing the derived LHr"".,for the reaction where R is the gas constant, T the absolutetemper- (12800+2580J) obtained from the experimental ature,LV" the volume changefor the reactionjadeite bracketswith that for the reaction involving low al- * quartz: high albite, and crothe activity ofjadeite bite (-150+2090J) calculated from the data of in the pyroxene.The experimentalwork of Holland Hlabse and Kleppa (1968)with the help of the high- (1979b)on the activity-compositionrelationships for temperature heat contents (Robie et al., 1978),we the join jadeite-diopside at 600"C showed that the can estimatethe disorderingenthalpy of albite. This activity ofjadeite can be expressedas estimate, 12950t3320 J, can be compared with the measuredenthalpies of disorderll0l5+840 J mol-' o.,d: Ka"lta (Waldbaumand Robie, l97l),9640 J mol-'(Kracek and Neuvonen, 1952), and 14220=1660J mol-' where Xro and yrd are the mol fraction and the activ- (R.C.Newton and T. V. Charlu,unpublished data). ity coefficient of jadeite respectively.no was deter- This last value was determinedfrom the differencein mined (Holland, 1979b)as heatsof solution at 970 K of Amelia albite and high - 1,ra: €xPI(W/RZ) (l X,")1 albite crystallized from Amelia albite glass at 20 kbar, l200oC, and should representmaximum dis- where the interaction energy, W, was found to be ordering enthalpy for albite. All enthalpies, unless 24kJ. For Reinsch'spyroxene we find 1',a: l.l2 ar.d otherwise stated, are given corrected to 298 K for aro: (0.81)'z(1.12): 0.73 at 680oC,which leads to a easeof comparison.The high-temperaturecalorimet- destabilizationof AP: 1.4kbar, and an estimateof a ric measurementsof Holm and Kleppa (1968)on al- minimum pressureof 17 kbar for the assemblage. bites heat-treatedfor times of the order of one month The quantitative interpretation (Ganguly, 1973)of at 1050"C yielded a disordering enthalpy of the experimentsof Kushiro (1969) on the join diop- 14230+1050J mol-' at97l K, which is equivalentto side-albite was hamperedby lack of data for the al- 13870+1050J mol-' at298K.It may be (R. C. New- bite breakdown curve at the high temperaturesin- ton, personal communication) that albite does not volved. Application of the mixing relationsdiscussed disorder fully at 1050"C under atmosphericpressure aboveto Kushiro's data showsthat predictionsbased HOLLAND: REACTION ALBITE: JADEITE + QUARTZ 133

upon the location of the albite breakdown curve to- supported by NSF grant EAR 74-22851 (R. C. Newton), with gether with the activity-composition relations deter- additional support from the Minerals ResearchLaboratory, NSF. mined at 600oC are in good agreementwith the experiments of Kushiro. The agreement is particularly gratifying becauseKushiro's work was performed at References temperatures(1050-1250"C) far removedfrom those at which the mixing relations were determined. Bell, P. M. and E. H. Roseboom(1969) Melting relations of ja- The results on the albite breakdown curve can be deite and albite to 45 kilobars with commentson mghing dia- used together with the recent determination of the grams of binary systemsat high pressures.Minerat. Soc.Am anorthite breakdown curve to grossular Spec.Pap., 2, 15l-161. + kyanite * (1966) qtrafiz Birch, F. Compressibility;Elastic constants.In S. P. Clark, (Goldsmith, 1979) to predict the maximum Ed., Handbook of PhysicalConstants, p.97-173. Geol. Soc.Am. stability pressuresfor plagioclase feldspars in the Mem.97. lower crust. Reliable estfunatesfor these limits await - and P. LeComte(1960) Temperature pressure plane for al- accurate determination of the mixing relations in bite composition.Am. J. Sci.,258, 209-217. plagioclasefeldspar. Boettcher,A. L. (1971)The nature of the crust of the earth, with specialemphasis on the role of plagioclase.In J. G. Heacock, Conclusions Ed., The Structure and PhysicalProperties of the Earth's Crust, p. 261-277. Geophys. Monogr. Ser. 14, Am. Geophys. Union, The experimental determination of the jadeite * Washington,D.C. qaartz: albite reactionrepresents a considerableim- - and P. J. Wyllie (1968)Jadeite stability measuredin the provement in location of the curve as well as an ex- presenceof silicateliquids in the systemNaAlSiO.-SiO2-H2O. Geochim. Cosmochim.Acta, 32, 999-1012. tension of the measurementsto higher temperatures. Coleman, R. G. (1961)Jadeite deposits of the Clear Creek area, This study demonstratesthat the P-T lne showsno New Idria district, San Benito County, California. J. Petrol., 2, curvature, at least in the range 600-1200oC, and 209-247. hencethat albite doesnot undergo Al-Si ordering in Danckwerth, P. and R. C. Newton (1978)Experimental determi- the experinental runs. nation of the spinel peridotite to garnet peridotite reaction in the systemMgO-Al2O3-SiOr in the range 900oC-ll00oC and The location of the curve agreesprecisely with Al2O3 isopleths of enstatite in the spinel field. Contrib. Mineral. both pre-existing gas-apparatus determinations Petrol.,66,189-201. (Birch and LeComte, 1960;Hays and Bell, 1973)and Ganguly, J. (1973)Activity compositionrelation of jadeite ln om- demonstratesthe reliability of the NaCl pressureme- phacite pyroxene.Earlh Planet. Sci.Lett., 19,145-153. (1979) dium in piston-cylinder experiments.No pressure Goldsmith, J. R. Melting and breakdownreactions of anorthite at high pressuresand temperatures.EOS, 60,420. correction need be applied for such determinations - and F. Laves(1954) The microcline-sanidinestability rela- with the experinental techniqueused in this study. tions. Geochim.Cosmochim. Acta, 5, l-19. The slope of the reaction as experimentallydeter- Hays, J. F. and P. M. Bell (1973)Albite-jadeite-quartz equilib- mined is consistentwith recenthigh-temperature ca- rium: a hydrostatic determination. Carnegie Inst. lil/'ash. Year lorimetry and yields an entropy Book, 72,706-708. of disordering which Hlabse, T. and O. J. Kleppa (1968)The thermochemistryof ja- is in accord with predictions based upon site occu- deite.Am. Mineral., 53, 128l-1292. pancy studies on natural and heat-treated albites. Holland, T. J. B. (1979a)Experimental determination of the reac- The enthalpy of disordering in albite can be similarly tion paragonite : jadeite + kyanite + wat€r, and lnternally con- determined from the experimental brackets and is sistent thermodynamic data for part of the systemNa2O-Al2O3- also in agreement with SiO2-H2O,with applicationsto eclogitesand blueschists.Coz- thermochemical measure- trib. Mineral. Petrol., 68, 293-301. ments. Thermodynamic data derived for hydro- - ( I 979b) Reversedhydrothermal determinationof jadeite- thermally crystallized synthetic high albite are tied to diopsideactivities. EOS, 60,405. the entropy and enthalpy of low albite given in Hokn, J. L. and O. J. Kleppa (1968)Thermodynamics of the dis- Robie et al. (1978)and form part of a self-consistent ordering processin albite (NaAlSirO). Am. Mineral., 53, 123- set which describes 133. accuratelyall the measuredreac- Johannes,W. (1978)Pressure comparing experiments with NaCl, tions among albite, paragonite,kyanite, andalusite, AgCl, talc, and pyrophyllite assembliesin a piston cylinder ap- jadeite, q\artz, corundum, and water (Holland, paratus. N euesJahrb. Mineral. Monat sh., 84-92. 1979a). P. M. Bell, A. L. Boettcher,D. W. Chipman, J. F. Hays, H. K. Mao, R. C. Newton and F. Seifert (1971) An inter- Acknowledgments laboratory comparison of piston--cylinder pressure calibration using the albite-breakdown reaction. Contrib. Mineral. Petrol., R. C. Newton and J. R. Goldsmith have contributed greatly in 32,2+38. discussionsduring the course of this work. The project was Kracek, F. C. and K. J. Neuvoncn (1952) Thermochemistryof 134 HOLLAND: REACTION ALBITE: TADEITE + QUARTZ

plagioclase and alkali feldspars. Am J. Sci., Bowen Volume, Robie, R. A., B. S. Hemingwayand J. R. Fisher (1978)Thermody- 29t-3r8. namic propertiesofminerals and relatedsubstances at 298'15K Kushiro,I. (1969)Clinopyroxene solid solutions formed by reac- and I bar (105 Pascals)pressure and at higher temperatures. tions between diopside and plagioclase at high pressures.Mrz- U. S. Geol. Sun. Bull. 1452. eral Soc.Am. Spec.Pap., 2,179-191. Skinner, B. J. (1966) Thermal expansion.In S. P. Clark, Ed., Mirwald, P. W., I. C. Getting and G. C. Kennedy (1975)Low-fric- Handbook of PhysicalConstants, p.75-96. Geol.Soc. Am. Mem. tion cell for piston-cylinder high-pressureapparatus. "L 97. Geophys.Res., 80, 1519-1525. Waldbaum, D. R. and R. A. Robie (1971)Calorimetric investiga- Newton, R. C. and J. V. Smith (1967)Investigations concerning tion of Na-K mixing and polymorphismin the alkali feldspars. the breakdown of albite at depth in the earth. J. Geol., 75, 268- Z. Kristallogr., 184, 381420. 286. Reinsch,D. (19'fD High pressurerocks from Val Chiusella(Sezia- Lanzo zone,Italian Alps). NeuesJahrb. Mineral. Abh., 130,89- Manuscript received, December 13, 1978; to2. acceptedfor publication, May 24, 1979.