American Mineralogist, Volume 67, pages M3452, 1982 Review of the behavior of plagioclase under metamorphic conditions Julrex R. Gor-osurru Departmentof the GeophysicalSciences The Universityof Chicago Chicago,Illinois 60637 Abstract The behavior of plagioclasefeldspars under metamorphic conditions is reviewed. The literature on plagioclasecomposition as related to reactions involving zoisite (epidote) and other calcic phases,the peristerite gap, and the coexistenceofintermediate plagioclasesof various compositions is presented. The various ideas on the relative effects of structural statesand of the role of chemicalreactions on subsolidusrelations are considered.Two views, not mutually exclusive, relative to plagioclasecompositions, are apparent: a "homogeneous"view, emphasizingcontrol by crystal-structuralfactors, and a "heteroge- neous" view, dominatedby sequentialmineralogical reactions. Mineralogy and petrology are supplementaryand dence, attempted to explicitly define the nature of complimentary disciplines. Much of our under- the relationship. Ramberg proposed an equilibrium standingof the petrology and ultimately the geology between epidote and "anorthite-molecule" in the of any portion of the earth is dependent on our plagioclase: knowledge of the properties and characteristics of 2CaAl2Si2O8+ llzHzO = Ca2Al3Orz(OH) minerals and of mineral interactions. The petrolo- anorthite "epidote" gist examines the minerals to help unravel the history of a rock; the mineralogistgains insight on + l/2Al2sio5 + 1/2sio2 mineral behavior from accumulatedinformation on kyanite quartz the rocks. Someminerals with limited composition- He then wrote a reaction in which calcic plagioclase al and structural variation have little to reveal. plus water breaks down to a more albitic plagioclase Others, such as the feldspars,may contain a great plus epidote,kyanite, and quartz and constructeda deal of information, but can be complex and stub- theoretical curve for the equilibrium in which, with born in revealingtheir secrets.A voluminouslitera- decreasingtemperature, an increasingly sodic pla- ture on plagioclasehas not yet clearly characterized gioclase is stable with epidote. Ramberg was fully the phaserelations. The equilibriumpicture is over- cognizantofthe role ofpressureand drew schemat- lain by confusion compoundedof and dominated by ic diagramsof both isobaric and isothermal surfaces kineticfactors. Furthermore,heterogeneous miner- for the reaction. He considered this reaction rela- alogical reactions can be affected by the varying tion to be a continuous one all the way from the energeticsof differentstructural states. This review greenschistto the granulitefacies. briefly covers and is restrictedto the highlightsof Ramberg(1949) extended his ideas on the plagio- papers on plagioclasefeldspars in metamorphic clase-epidote equilibrium to potassium-rich rocks rocks, and is followed by a report on an experimen- with orthoclaseand muscovite.A secondcontinu- tal investigation that bears on plagioclase stability ous reactionoccurs: at_elevated temperatures and water pressures. Petrologistshave long been aware of an increase 4CaAlzSizO8+ KAlSi3Os + H2O in anorthite content in some metamorphic rocks anorthite orthoclase accompanying an increase in metamorphic grade ? KAhSigOro(OH)z+ 2Ca2Al3SirOrz(OH)+ 2SiO2 (Becke, 1913).Although earlier workers had con- muscovite epidote quartz sidered a relationship between plagioclaseand epi- dote or other Ca-rich minerals(see Christie, 1959), Figure I is a reproduction of his (2-dimensional) Ramberg (1943), largely on the basis of field evi- diagram.The dashedcurve, representingthe equi- 0003-004x/82/0708-0643$02.00 il3 644 GOLDSMITH: REVIEW OF PLAGIOCLASE data clearly indicated a gradual I curve, and the increasein the anorthitecontent of the plagioclases U F HOMOGENEOUSFIELD with increasinggrade of metamorphism.He noted of :+ significantdeviations, however, and indicatedthat z PLAGIOCLASE there are problems, as recognizedby Ramberg,in its useas an index of intensityof metamorphism.In 1t addition, Lyons noted, ". a great paucity of F ar%\ metamorphic plagioclase in the range of Anrr An2o." a = -^\ Ramberg's (1943, 1949, 1952) deduced curves 1l = tl EPIDOTE lt were continuous in nature, implying an equilibrium A that variescontinuously with temperature(or meta- prlllocusr U z 4 morphicgrade). A numberof field studiesin region- (Kyonite, tl E9 ally metamorphosedterranes had given some indi- - Muscovite) Yo' i r9 cation of a rapid conversionof albite to oligoclase U t- or sodic andesineover a short distance, but it is o 1l likely that the first clear-cutevidence for a discon- tl t tinuous increase in plagioclasecomposition with increasinggrade was describedfrom greywackesof 0 20 40 60 80 t00 (Ambrose, Co,Al No,Al the Missi series, Manitoba 1936; see Silicotes Silicotes Brown, 1962).Ambrose commentedon abundant Fig. l. ReproductionofRamberg's (1949)subsolidus diagram epidote, and assumed that original intermediate of the plagioclaseP epidote equilibrium, deduced in part from plagioclaseof volcanic rocks has reacted, ". field observations. according to the familiar equation: plagioclase + water * iron : albite + epidote * alumina * librium of plagioclaseand epidotein the presenceof silica." He noted that water-clear albite persists muscoviteand potashfeldspar is drawn somewhat with little changethrough the biotite zone, but with abovethe K-free equilibriumbecause muscovite in the appearanceof garnet, the anorthite content of associationwith K-feldspar and quartz has some- the plagioclaserises abruptly from An64 to An2s-32 what lower Al-activity than kyanite. Ramberg and that epidote is greatly diminished in amount pointed out that increasedH2O pressuredisplaces above the garnet isograd, concomitant with the the equilibrium toward epidote, as does increasing appearanceof oligoclase. the Fe3+/Alratio becauseof the ability of epidoteto Sincethe earlier years of the 20th century, when incorporateFe3* in its structure. Rambergconsid- the nature of the isomorphousreplacement of Na ered the reactions of Figure I to span a temperature and Si for Ca and Al in the plagioclasefeldspars intervalof approximately500" C, and facies,repre- becameknown, petrologistshad long considered sentingtemperature, are ploted on the vertical axis. the plagioclasesto be an ideal solid solution series. He felt that the epidote-plagioclaseequilibrium in The dogmaof one rock, one plagioclase(excluding most quartzo-feldspathicrocks cannotregister P-Z zoning, or other non-equilibriumphenomena) pre- conditions higher than those that would develop vailed. Chao and Taylor (1940)were the first to Ansg-aoplus epidote,due to the low CalNa ratio in observe superstructurereflections in X-ray diffrac- these rocks. Ramberg(1952) further extendedthe tion patterns of intermediateplagioclases, indicat- plagioclase-epidote relations to a more complex ing structuralcomplexities. Chao and Taylor con- systemcontaining CaCO3 and indicatedthat rocks sideredthis as evidence against a complete solid with the sameplagioclase compositions may have solution series, although a strong hint had earlier formed at differenttemperatures, depending on the appearedwhen Taylor et al. (1934)found the c-axis partial pressureof water. of anorthite to be twice that of albite. Kohler Lyons (1955) plotted plagioclasecompositions (1942a,b) and Larsson (1941) were the first to againstmetamorphic grade in the Hanover quadran- explicitly describeoptical diferences betweenpla- gle (New Hampshire-Vermont)as a check against gioclasesofplutonic and volcanicrocks, and thus to Ramberg's(1943) diagram. He observedthat in a differentiate high-temperature from low-tempera- general way the separation of the plagioclase-epi- ture plagioclases.Cole et al. (1951)showed that dote and plagioclase fields followed Ramberg's low-temperatureplagioclase from An3qto AnT2had GOLDSMITH: REVIEW OF PLAGIOCLASE 645 complex superstructures.Laves (1951, 1954) re- free energy raises and flattens the plagioclase- ported that there are plagioclasesin the composi- epidote equilibrium curve in the An35-An5sr&ng€, tional range of approximately An5-Ar17, &rlorg reducing the temperature interval in which calcic them peristerites,that show two phasesin "single- andesineis stable. crystal" X-ray photographs.Laves estimatedthe Brown (1962)presented an excellentreview and compositionsof the exsolvedphases to be approxi- clarification of the subject of peristerite unmixing in mately Ane and An3s. Gay and Smith (1955)con- the plagioclasesand related metamorphic and facies firmed this observation,and refined the estimation series.His summaryof the field observationsindi- of composition of the two phases as An3*2 and cated that the rate of increasein the composition of Anzt=z.Thisregion of compositionwas termedthe the most basicplagioclase with increasingmetamor- "peristeritegap". Additional crystallographicwork phic grade is continuous or nearly so in rocks of the and refinement was done by Ribbe (1960, 1962), andalusite-sillimanite type and perhaps in the low- Brown (1960),and Fleet and Ribbe (1965). pressureintermediate type. A jump occursbetween Christie(1959) pointed out the significanceof the about An7 and about An26in the kyanite-sillimanite peristeritegap in low-temperatureplagioclases, and type andthe high-pressureintermediate type, and in that as a consequenceRamberg's smooth unbroken the jadeite-glaucophane type only albite seems to curve needed modification. He also assumedan exist. He mentionedthat the relation