American Mineralogist, Volume 77, pages 631442, 1992

An empirical phasediagram for the clinozoisite-zoisitetransformation in the system CarAlrSirOrr(OH){arAlrFe3 +Si3Or r(OH)

Gnnnlno FuNz FachgebietPetrologie, Technische Universitiit Berlin, StraBe des 17Juni 135,D-1000 Berlin 12,Germany JaNr Srr,vERSToNE Departmentof Earthand PlanetarySciences, Harvard University, 20 OxfordStreet, Cambridge, Massachusetts 02138, U.S.A.

Ansrnlcr The effects of pressure,temperature, and Fe3* + Al exchangeon the transformation from orthorhombic zoisite to clinozoisite are examined using thermodynamic data and analysesofnaturally coexisting mineral pairs. Theoretical considerationsindicate a steep transformation loop in the T-X sections, with the Al end-member reaction occurring at low temperature and the Fe end-member reaction at high temperature;the orthorhombic phaseis the high-temperatureform. The position of the two-phasezoisite-clinozoisite area in a Z-Xsection varies with pressure.Equilibrium mineral pairs from low- to intermediate- pressureregimes indicate a smaller two-phase area (betweenapproximately 8-15 molo/o AlrFe in zoisite and 25-40 molo/oin clinozoisite) than those from high-pressureregimes (samevariation in zoisite, 35-55 molo/oin clinozoisite). The width of the two-phaseregion also dependson minor elements,such as the rare earth elements.The effect on the phase diagram of a possible low-temperature miscibility gap in the monoclinic series is also discussed.Disequilibrium textures, such as complex zoning patterns and multiple gener- ations of grains, are common features in naturally coexisting clinozoisite-zoisite pairs. Owing to the sluggishreaction behavior of the epidote minerals, these can serve as im- portant petrogeneticindicators of prepeakor relict metamorphic conditions in a wide range of bulk compositions.

INrnolucrroN ing temperature,both the monoclinic and the orthorhom- forms become enriched in Fe3*' However' Acker- The epidote mineral group, Iike the pyfoxenesand feld- bic a1d Raase (1973) observed that clinozoisite and spars, exhibits borh chemical and structural ;;i;;r. ry1d prograde core to rim zoning toward Fe- The principal isomorphous substitution Al * FJil;;- zoisite showed poor Prunier and Hewitt (1985) presented tinuous over a wide range of compositions unA op"*t"t compositions' evidence that at high temperaturesboth inbothorthorhombicandmonocliniccrystals.Th;il;: experimental forms are-relatively Fe rich and thus verified Enami and tural transformation from the orthorhombi;i; ,h" Banno's hypothesis' Comparing the loop proposed by monoclinic form thus potentially depends o" pr"rru.L, Prunier and Hewitt (1985) with the data of Enami and temperature, and chemical composition. This ;;i;;- (1980) still leaves considerableuncertaintv about mation is petrologically significant in a variery ;i;;;;- Banno the exact position in P-?"spaceof the two-phaseloop and morphic environments: zoisite is part or tn" rtigi-p."r- its form in T-X and P-x projections' The aim of this sure assemblagethat replaceslawsonite una u"o.tiltJu"i is 10 present new data on coexisting zoisite and is common in many eclogites;epidote -i"".ufr'u." study "U+ clinozoisite mineral pairs in order to construct an emplr- uitous in mafic rocks throughout the blueschisi, fi";- ical phase diagram that stressesthe important influence schist, and epidote amphibolite facies; and ,"iri;;;; ofpressure on the phasetransformation' clinozoisite both occur in calc-silicate rocks rro* u "u- riety of P-Z environments. Despite its petrologic impor- tance. however. the effects of structural and chemical IHEORETTCALCONSTDERATTONS variationson reactionsinvolving epidotegroup minerals This investigation is restricted to the system are not well understood. CarAlrSirO,r(OH)-CarAlrFe3'Si3Orr(OH);the influenceof The controversy concerning the zoisite-clinozoisite other potentially important substitutions involving Mn, phasediagram can be summarized as follows. Enami and Cr, and rare earth elements is only briefly discussed.In Banno (1980) analyzed coexisting zoisite-clinozoisite addition, we restrict it to the Al-rich part of the system, mineral pairs from rocks of different metamorphic grade. where only up to one-third of the total Al is replacedby They arranged the analysesby increasing metamorphic p"rt (and we assumethat Fe2*is not important). Higher grade (i.e., temperature)and suggestedthat, with increas- Fe contents are only rarely reported in the literature (e.g., ooo3404x/92l05064631$02.00 631 632 FRANZ AND SELVERSTONE: CLINOZOISITE-ZOISITE TRANSFORMATION

Al2Al rather low critical temperature or both. Their possible 0 influence is discussedin a subsequentsection. The transformation from monoclinic to orthorhombic forms can be described in terms of two end-member re- actrons:

mcl CarAl.SijO,r(OH)

: orh CarAlrSi3O,r(OH) (l)

mcl CarAlrFeSi3Orr(OH)

: orh CarAlrFeSirO,r(OH). (2)

The structuresof the orthorhombic and monoclinic forms are similar. They consist of SiOoand SirO, groups linked Fig. l. Nomenclatureof the epidotegroup minerals in the by octahedral chains occupied by Al and Fe3*.The two system CarAl3Si30r r(OH)-CarAlrFeSirO,r(OH).Orthorhombic forms are related to each other by a glide-twin operation, zoisiteincludes the Fe-poorvariety zoisitewith orientationof B but the types of octahedral chains in the two forms are the optic axial planeparallel to (010)and the Fe-richvariety a zoisitewith the optic axial planeparallel to (100).Monoclinic essentiallydifferent: in zoisite only one type ofchain oc- zoisiteincludes clinozoisite proper, which is opticallypositive, curs, whereasin clinozoisite there are two types of chains epidoteproper (optically negative), and the hypotheticalcom- (Dollase, 1968). The transformation can also be de- positionpistacite (not shown),CarFerSirO,r(OH). Because of the scribedas cation shuffieor synchroshear(Ray et al., 1986) highoptic axialangle dispersion ofthe epidoreminerals and the with a rearrangementof (Al,Fe) and Ca on the M3 site. strongeffect of minor elementson opticalproperties, the exact Such a transformation is probably of first order with a positionof the curvesvaries (dashed lines, diagram modified two-phase region in a T-X or P-X section. It cannot be afterTrdger, 1982; Maaskant, 1985). describedas a miscibility gap,which would be a function only of cation ordering. In the epidotes, ordering of oc- tahedrally coordinated Al and Fe3*is almost complete in Bird et al., 1988)and appearto be confinedto low-tem- zoisite(Fe3+ in the A13 site;Ghose and Tsang,l97l) and perature environments; thesemonoclinic high-Fe epidote lessperfect in clinozoisite (Fe3*in the M3 site, with small sampleshave not been reported coexistingwith an ortho- amounts in the Ml site; Dollase, l97l). rhombic member of the epidote seriesand henceare not Both Enami and Banno(1980) and Prunierand Hewitt relevant to the problem addressedhere. (1985)argued that pressurehas no influenceon the trans- We use the following nomenclature, which is basedon formation loop, owing to the very small differencein mo- optical properties and summarized in Figure l. Zoisite is lar volumes of the orthorhombic and monoclinic forms the orthorhombic form. The optical varieties with differ- (V"",-- 135.9cm3lmol, V^": 136.2cm3lmol, Helgeson ent optic axial orientations and a changefrom positive to et al., 1978; V,"t : 135.58, V"t": 136.73,Holland and negative character (see Myer, 1966; Vogel and Bahezre, Powell,1990; V*,: 135.88,V*:136.76, Berman,1988). 1965;Maaskant, 1985),which is stronglydependent on However, the P-T slope of a reaction depends on both Fe content, are consideredas one phase,since there is no AV and,AS, and since the entropy values for the epidote indication that thesechemical (and optical) variations are minerals are also similar (S,", : 296.1 J/K.mol, S.o : discontinuous or that they are related to structural dis- 295.1 J/K.mol, Helgesonet al., 1978;S,.1 : 295, S.,, : continuities. There is also no clear indication of structur- 291, Holland and Powell, 1990; S,", : 297.58,5.,": ally different types in the monoclinic series(Deer et al., 287.08,Berman, 1988; &' :296.58,,S"k: 294.06,Gott- 1986).A continuousrange of compositionsfrom low to schalk, 1990), it is possible that pressureexerts a signifi- high Fe contentsexists at certain P-?"conditions, and the cant effect on the position and shape of the two-phase monoclinic seriescan also be representedby one phase. region. The reported molar-volume value with the small- We use the term "clinozoisite" throughout this paper to est error in determination of the lattice constantsfor syn- refer to monoclinic forms and reserve "epidote" for the thetic orthozoisitewas given by I-angerand Lattard (1980) whole mineral group; "pistacite," which is also a term as 135.77cm3/mol, which is in good agreementwith the usedfor Fe-rich epidote, may be usedfor the hypothetical data from the above mentioned data sets. No reliable end-memberCarFerSirO,r(OH). The Fe-freemonoclinic data for synthetic Fe-free clinozoisite are known. Pisto- end-member is known from experimental studies (e.g., rius (1961) reportedidentical values for zoisite and cli- Jenkinset al., 1985).Several studies on natural samples nozoisite(within the limits of error) of 136.4+ 0.7 cm3/ suggestthat miscibility gapsoccur within the monoclinic mol. Extrapolation of data for Fe-bearing clinozoisites series (Hietanen, 1974; Raith, 1976; Selverstoneand presentedby Myer (1966),Bird and Helgeson(1980), and Spear, 1985).These proposed miscibility gapsare located Seki (1959) yields a value of 136.2 cm3/mol.Linearly at relatively high Fe contents of the system or have a extrapolateddata from H0rmann and Raith (1971) in- FRANZ AND SELVERSTONE: CLINOZOISITE-ZOISITE TRANSFORMATION 633 dicate a slightly higher value of 136.5,but they suggest that a linear extrapolation is probably not correct, and their givenline would extrapolatetoward a valueof 137.5 cm3/mol. Reaction I has a negative slope in P-Zspace, although there is a large range in the magnitude of the slope ob- tained with different data sets (-124 bars/"C, Berman, 1988;- 13.3bars/"C, Helgeson et al., 1978;-34.8 bars/ oC, Holland and Powell, 1990). If an average value of -50 bars/'C is accepted,a pressure increase of 2 GPa would shift the equilibrium temperature for Reaction I by 400 'C. No thermodynamic data for Reaction 2 are Xotf" available. If we assumethat the changein entropy arising X n,f" from the exchangeof Fe3* for Al is essentiallythe same regardlessof whether the exchangeoccurs in an ortho- Fig.2. SchematicT-X phasediagram for the transformation (A) possibilitiesof eithera rhombic or a monoclinic crystal, the P-T slope of Reac- from clinozoisiteto zoisite. The two positive(+) or a negative(-) P-7 slopefor theAlrFe end-mem- tion 2 dependsonly on the molar volumes and the entro- ber Reaction2 for two isobaricsections at low pressurePl and py from phenomena.Bird and contribution order-disorder highpressure P2. (B) An enlargedpart of A, showingthe possible Helgeson(1980) compiled data on lattice constantsand zoning trendsof coexistingminerals (squares : zoisite,dia- calculatedmolar volumes. In their Figure 2 it can be seen monds: clinozoisite)during prograde and retrogradedevelop- that, with increasing Fe3+contents, the difference in the ment (pathsa-d, seetext). The rangeof compositionsof the molar volumes of the orthorhombic and monoclinic forms orthorhombicand the monoclinictypes during such a cyclemay may get smaller. overlap,and Fe-poor clinozoisite may therefore coexist with rel- Ifthere is significant order-disorder ofAl and Fe in the ativelyFe-rich zoisite in complexlyzoned minerals. epidotes, it is likely that the degreeofdisorder in clino- zoisiteis higherthan in zoisite(Dollase, 197 l; Ghoseand existing phases.With the assumption that the model with Tsang, 197l), which resultsin a decreasein AS for Re- a positive slope for Reaction 2 is correct, coexisting zois- action 2. Bird and Helgeson (1980) give an entropy dif- ite and clinozoisite would exhibit different types of zon- ferencefor ordered-disorderedepidote of 0.3 J/K.mol. ing, dependingupon the P-f path they had followed dur- Changing the value of AS and a decreasingA Z will have ing metamorphism: a simple prograde path with a large effect on the slope of the reaction, but unfortu- simultaneous burial and heating would result in increas- nately uncertainties in the available entropy and volume ing Fe contents from core to rim ofboth phases(path a data and lack ofdata concerning their sensitivity to sub- in Fig. 2B), whereas isothermal uplift (path b) and iso- stitution and order-disorder phenomenado not allow us baric cooling (path c) would result in decreasingFe con- to distinguish with confidence between a positive and tents; heating during uplift (path d) would also cause a negative slope for Reaction 2. rimward decreasein Fe content of both minerals. If Re- For the construction of a schematicP-T-X diagram, it action 2 has instead a negative P-Z slope, the two-phase is also necessaryto have some information about the region is shifted with increasing pressure toward lower temperature range in which Reactions I and 2 are locat- temperatures. Qualitatively, the change in composition ed. Jenkinset al. (1985) determinedReaction I experi- of coexisting pairs would be the same for the P-I paths mentally and located it at temperaturesbelow 350'C; discussedfor Figure 2B, but the changewould be much calculationswith the data of Berman(1988) yield a tem- more pronounced for both minerals, in contrast to the peraturenear 200'C at I bar. Despitethe uncertainties small changesin zoisite composition illustrated in Figure in the absolute temperatures,there is no doubt that the 2B' orthorhombic form is the high-temperature form. Reac- tion 2 is probably metastablebecause it must lie at tem- NlTrlRAll.y coExrsrrNc zolsrrE{LrNozorsrrE PArRs peratureswell above 600'C. This was demonstratedex- In order to constrain the position ofthe two-phase re- perimentally by Holdaway (1972) and empirically by the gion, we analyzedcoexisting minerals from a variety of common occurrenceof monoclinic epidote near its upper P-Tregimes. This approachhas some inherent problems. stability limit in amphibolite facies rocks. The first problem is that epidotes, especiallyFe-rich cli- Figure 2 shows I-X sections for Reactions I and 2, nozoisite, can preserve extremely complicated zoning with the assumption of both positive and negative slopes patterns, which show up clearly in backscatteredelectron for Reaction 2 and extreme temperature differencesbe- images and indicate slow diffusion rates at metamorphic tween the two reactions.ln a T-X diagram the limbs of temperatures. In theory, slow Fe3*-Al diffusion should the two-phase region are very steep, and their position make the epidote minerals suitable for an empirical study changeswith pressureand the changingslope ofReaction of the two-phase region becausetheir sluggish reaction 2.If anegative slope for Reaction I is correct, it is clear behavior inhibits retrograde reequilibration. In practice, that pressurehas a large effect on the composition ofco- however, epidotes can form over a wide range of P-T 634 FRANZ AND SELVERSTONE: CLINOZOISITE-ZOISITE TRANSFORMATION

10 o5 roz MnO.;' o I Fe2O3 Ror o rl' o 'r.'". o c 02 PO ; ;e v .a Eo t# o F., "'J N7 o - E N13-3 03 O6 O RM71 Cr2Og..:" .c . NR125 + RWl03 5 r FT1A o2 = tr DTBD ^ 85.1 a 5

01234567 ol 02 irirrn roi.ilj nwl %in% in zoisite wt % in zoisite Fig. 4. Distributionof the minor componentsTiO', MnO, Fig.3. Distributionof FerO,between clinozoisite and zoisite MgO,and CrrO, betweenclinozoisite and zoisite(same data set (onlyanalyses ofgrains in contactwith oneanother are plotted); asFig. 3; overlappinganalyses not shown).The l:l slopeline is clinozoisiteis stronglyenriched in FerOr. givenas a referenceline; for TiO, a referenceline with a slope of 2.5:I is alsoshown. conditions between the lower greenschistfacies and the In most samples,other optical features,such as anoma- eclogite facies, which at times makes it difficult to deter- lous interferencecolors, crystal shapeand size,and inclu- mine if coexistingzoisite and clinozoisite grew during the sion patterns, were also helpful in distinguishing between same event. We selectedgrains that appearedto belong zoisite and clinozoisite. Additional transmission electron to the same fabric generation and worked with backscat- microscopy was done on two samples;a detailed report tered electron images to analyze only grains in mutual on this work is in preparation (Smelik and Franz). contact and with no evidence for late-stagegrowth or reequilibration. Compositionsof coexisting phases Uncertainties in the microprobe analysesgive rise to The FerOr, TiOr, MnO, MgO, and CrrO, contentsof an error on the order of + 1 molo/oAlrFe. Assuming an all coexistingclinozoisite-zoisite pairs from all rock types error of 0. I absolute wto/ofor an Fe-bearingzoisite with are given in Figures 3 and 4. All analysesare compiled 1.75 wto/oFerO, resultsin +0.6 molo/oAlrFe. An error of in Table l.r Figure 3 shows that FerO. is preferentially 1.25 relative wt0/oin SiO, has no influence on the calcu- incorporated into the monoclinic form. TiO, (Fig. 4) is lated AlrFe component, whereasan error of 1.25 relative also preferentially incorporated in clinozoisite, up to a wto/oin AlrO. results in +0.3 molo/oAlrFe. The presence maximum value of 0.4 wto/o.MnO and MgO are also of minor elements such as Cr or Mn results in only a enriched in clinozoisite (up to 0.2 wto/o),whereas CrrO, slight overestimation of the AlrFe component. In most of partitioning behavior is unclear (Fig. a). Preferential in- the analysesthe sum ofminor octahedralcations per for- corporation of the minor elements into the monoclinic mula unit is below 0.05, which has little influnce on the crystalswill contribute to the uncertainty in position and molo/oAlrFe,calculatedas100[Fe,.,/(-2 + Al ., + Fe,",)]. slope of the two-phase loop in the Al-Fe system,but be- All analyseswere carried out with automated Camebax causethe absolute amounts are small, they are probably electron microprobes at Harvard University and the of minor importance. This is not the case, however, if Technische Universitiit Berlin, with simultaneous ener- these elements are present in large amounts, as is dem- gy-dispersive analysis to check for minor elements, and onstrated for the rare earth elements. Sample BP I 19 with backscatteredelectron images to avoid mixed anal- (metagranite from the Granatspitzkern Zentralgneis, yses on adjacent grains or on different zones within the Tauern Window, Austria; Fig. 5) contains zoisite blasts grains. Despite different standards and correction pro- with small (approximately 20- 1rm)clinozoisite inclusions, grams (Harvard-natural minerals, Bence-Albeecoffec- which have variable amounts of REE. The zoisite has a tion; TUB-natural minerals and synthetic metals, PAP constant composition with 6-8 molo/oAlrFe, but the Fe correction), analyseson the same materials are identical in the range mentioned above. ' Distinction Table I may be ordered as Document AM-92-496 from the between orthorhombic and monoclinic Business Office, Mineralogical Society of America, 1130 Sev- crystalswas made in thin section by determination of the enteenthStreet NW, Suite 330, Washington,DC 20036,U.S.A. extinction angle in grain sectionsshowing the a-c plane. Pleaseremit $5.00 in advance for the microfiche. ctrnoa lamellae n zor core

sec clhoz al am ol zol o===-_-:9

rcl(?) alleralion ol zor sec clinoz h cracks c E@o@OO@OO o o oo ooc€ o@

10 12 t4 t6 26 30 3436

sr1 ftl(?) alle.alion sec clnoz (:--+ mrT]Drr @ @

6142024 68r01214t6t820

l------

r0 12 !4 16 30 34

mol7o AlzFe

Fig. 5. Analytical results for the investigatedsamples of cli- AlrFe in the other. The compositional gap between 15 and 24 nozoisite-zoisite pairs; squares : orthorhombic; diamonds : molo/oin one crystal does not indicate a miscibility gap, since monoclinic; rotated squares : structural state uncertain; trian- microprobe analysesof other crystals fall into this range. Sam- gles: rare earth-bearing clinozoisite. Lines connect coexisting ples 85-l and 85-2: Pegmatoid vein in eclogite with formation minerals; analyseswere made a few micrometers away from the ofclinozoisite in core ofzoisite, secondaryclinozoisite near and grain boundary. Sample BP I 19: REE-bearingand REE-free in- at the rim ofzoisite, secondaryclinozoisite in cracks in zoisite. clusionsofclinozoisite in zoisite; the width ofthe two-phasearea The samples show both equilibrium and disequilibrium pairs widens with increasing REE content. N l3-3: Matrix and vein and also the alteration ofprimary zoisite into clino(?)zoisitewith assemblagein a low-grade amphibolite; filled symbols indicate increasing Fe contents. Newly formed crystals with < I molo/o core compositions. Sample NR 125: Tourmaline-epidote segre- Al,Fe are too small for structural determination; an orthorhom- gation from contact metamorphic environment; filled symbols bic structure is consistentwith the proposedphase diagram, but indicate core compositions. Sample RW 103: Calc-silicatenod- not proved. Sample St l: Quartz-zoisite vein from eclogites, ule from a polymetamorphic granulite to amphibolite faciester- showing the same type of alteration of primary zoisite as in rane. Solid symbol indicatesFe-poor granulite facieszoisite with samples85-1 and 85-2. Core compositions of the zoisite (filled overgrowth of Fe-richer zoisite during amphibolite faciesmeta- symbols) is Fe rich, rims are Fe poor, but near to alteration morphism. Sample 1266:.Law-grade calc-silicate rock. Closed products Fe content increases.Samples DT 8 and FT I are from and open squaresrepresent core and rim analyses,respectively, eclogites,filled symbols indicate core compositions. Both cli- of grains determined optically to be orthorhombic; closed and nozoisite and zoisite show a wider range of compositions com- open diamonds are core and rim analysesfrom clinozoisite, with pared to the low-P rocks. Dashed line connectscore composi- a continuous range of intermediate compositions (no analysis tions of two adjacent crystals. Sample RM 7l is an epidote + points shown). Closed diamonds (lower two rows) represent garnet + albite vein in an eclogite.The high Fe rim ofclinozois- monoclinic intergrowths (two crystals determined with TEM) ite is only 5-10 pm wide (open diamonds), inside of this rim that contain 8-32 molo/oAl"Fe in one case ard 22-46 molo/o (filled diamonds) the clinozoisite is homogeneous. 636 FRANZ AND SELVERSTONE: CLINOZOISITE-ZOISITE TRANSFORMATION

Fig. 6. Photomicrographof equilibriumintergrowth of cli- Fig. 7. Photomicrographof zoisitewith 10-12 mol0/oAlrFe nozoisite(34-36 molo/oAlrFe) in thecore of a zoisite(12 molo/o with alterationzones parallel to (100)(dark because ofanoma- AlrFe),together with (sericitized)albite (sample85-1, crossed lous interferencecolors) with 15 molo/oAlrFe, intergrown with polars). albite(ab; sample 85-1, crossed polars). content in the clinozoisite varies from 25 molo/0,where ty, Granatspitz Zentalgneiskern, Tauern Window, Aus- the REE are below the detectionlimit (0.5 wto/o),to 5l tria). Although these rocks subsequentlyunderwent Ter- molo/0,where the REE total is I I wto/o(estimated from tiary Alpine metamorphism, the Paleozoic contact Ca deficiency and deviation of ideal anhydrous total of metamorphism is well preservedand is characterizedby 98.5 wto/o;semiquantitative energy dispersive analysis). the mineral assemblagecalcite + wollastonite + quaftz The REE end-member of the epidote group, allanite, con- * grossular+ diopside.Temperatures near 550 "C at0.2 tains Fe2*in large proportions (Dollase, I 97 I ), and there- GPa are assumed to represent the equilibration condi- fore these high Fe contents do not reflect the true AlrFe tions. There is no significant zoning in the epidote min- mol0/0,and clinozoisite of this type cannot be used to erals, and tourmaline inclusions in the epidote cores are constrain the two-phase loop. similar in composition as the matrix tourmaline crystals. Zoisite contains l0-13 molo/oAlrFe, clinozoisite32-37 Petrographical description molo/0. Owing to the problems mentioned above (complex Samples85-l and 85-2 (Figs. 5, 6, 7) are qtrartz-al- zoning, growth at diferent P-Z conditions) each sample bite(An,o)-zoisite-phengitesegregations in the Weissen- is different and must therefore be describedseparately. stein eclogite, Miinchberg Massif (Germany). The eclo- SampleN l3-3 (Fig. 5) comesfrom an ophiolite com- gite itself formed at high pressuresof at least 1.9 GPa at plex in the Gebel Rahib area, Darfur, northwestern Su- temperatures near 620 "C (Klemd, 1989). The segrega- dan (Abdel Rahman et al., 1990). Z-"- and P-.. are rough- tions and associatedpegmatoids probably formed during ly estimatedfor this areaas 400 + 50'C and <0.5 GPa, uplift when the P-Ipath just crossedthe minimum melt- respectively, from the mineral assemblageamphibole + ing curve (Franz et al., 1986) at pressuresof approxi- chlorite * epidote * plagioclase+ quartz + calcite. Ret- mately 0.9 GPa. Crystallization of a granitic melt at these rograde shearingin the rocks is recorded by transforma- conditions yields zoisite and potassium white mica in- tion of amphibole with 6 wto/oAlrO, into actinolite with steadof potassium feldsparand anorthite component (Jo- 3 wIo/oAlrOr. The sample shows coexisting zoisite and hannes,1985). Within the centimeter-sizedzoisite crys- clinozoisite in the matrix and in a small, deformed vein- tals are 0. l-mm-wide lamellae of clinozoisite (Fig. 6) let. In the matrix, both minerals have the samegrain size together with pure albite; the lamellae have straight grain and morphological featuresand show only minor varia- boundaries,typically occur in the coresof the zoisite crys- tion in chemical composition from core to rim. In the tals, and are oriented parallel to (100) ofzoisite. They are veinlet, zoisite is larger in grain size than clinozoisite and interpreted as equilibrium intergrowths that formed dur- shows sector zoning (variation from 5 to 9 molo/oAlrFe). ing cooling of the segregations.The zoisite crystals are In the veinlet, coexistingmineral pairs are zoisite with 7- zoned with rims slightly lower in Fe content (10 molo/o 9, clinozoisitewith20-25 molo/oAlrFe, compared to ma- AlrFe) compared to the cores (12-13 molo/o).Coexisting trix zoisite with 9-14 molo/oand clinozoisite with 27-28 mineral pairs give consistentvalues of I 2- l3 molo/oAlrFe molo/oAlrFe. and 35-36 molo/0,with one exception(14 and 4l molo/o). Sample NR 125 (Fig. 5) is a tourmaline-epidote seg- Texturally similar clinozoisite located at the rim of the regation. It comes from the contact of a granitoid with large zoisite crystals shows a larger variation in AlrFe metasediments,the so-called Hiillserie ("Seettirl" locali- content from 35 to 42 molo/o. FRANZ AND SELVERSTONE: CLINOZOISITE-ZOISITE TRANSFORMATION 637

A secondtype of clinozoisite (with 3l-53 molo/oAlrFe) occursalong irregular cracksin zoisite (seeFig. 9 in Ham- merschmidt and Franz, 1992) and also parallel to (100) and (010) planesof the host zoisite (Fig. 7), where it is accompaniedby a slight changein interferencecolors of - the adjacentzoisite. The clinozoisite formation and mod- *20 ification ofthe zoisite took place at a later stage,probably below 350 oC at low pressure.A late (postmetamorphic) hydrothermal stagethat might have been responsiblefor the features shown in Figure 7 has been documented by Hammerschmidt and Franz (1992) for the Miinchberg Massif. The modified zoisite crystals vary between 12 and26 molo/o(analyses at points where interferencecolors 20

-*-r-r- changed or immediately adjacent to the secondary cli- \ nozoisite).We were unable to determine by optical meth- ods whether thesecrystals are monoclinic or orthorhom- bic, but preliminary transmission electron microscopy results show deviations from orthorhombic symmetry in theseareas (a detailed description ofthese results will be 21.-26/'' - presentedelsewhere). We were also unable to determine the symmetry of small crystals approximately 5 pm in diameter that are located in the albite matrix parallel to /> (100) of zoisite(Fig. 7) and in alteredpegmatitic plagio- I clasenext to the pegmatitic zoisite These crystals. small n+ crystalshave AlrFe contentsnear I mol0/0.Assuming that the latter are orthorhombic and the zoisite modifications are monoclinic, a compositional gap between I and 12 molo/oexists. The rocks from this locality provide an ex- cellent example of the variety of epidote minerals that form during cooling. Only the primary intergrowths were used in determination of the transformation loop, al- though the secondaryclinozoisite crystalsgive some con- 50 rlm straints about the position ofthe loop at low P and L Sample 1266 (Figs. 5, 8) is also from the Miinchberg Fig. 8. Zoning in clinozoisite(drawing after backscattered area,but from the border zone adjacentto very low-grade electronimage, sample 1266)in a calcite + phengitematrix sediments(locality Schwingen,near Schwarzenbach/Saa- (cleavageindicated) showing central compositions with low Fe le, Germany). This so-calledprasinite-phyllite seriesdid content(6 mol0/0,as indicatedin the figure)increasing toward not experiencetemperatures in excessof 350 'C (the sta- the rim in irregularzones (separated in the figureby dottedand molo/o;all in- ble assemblageis stilpnomelane * actinolite + chlorite solid linesaccording to their sharpness)up to 3l termediatecompositions were found. Some of the Fe-richzones + calcite + albite; pressruesare not determined,but there haveREE presentin smallamounts. Small crystals are zoisite, is no indication high pressures). of The specimen is a rangingfrom 4 to 12 molo/oAlFe from coreto rim, and subor- calcite-phengiterock with minor amounts of epidote, ti- dinatetitanite (stippled). tanite, and tourmaline, from a gteenschist-metapelite contact.Zoisite occursas needle-shapedcrystals 5-20 pm, with 4-12 molo/oAlrFe mostly restricted to the calcite- conditions, low to intermediate AlrFe clinozoisites exist, rich part of the sample, whereaslarge clinozoisite grains indicating that there is no obvious miscibility gap, and (300 pm) wfih 7-46 molo/oAlrFe occur mainly in the that there can be an overlap in compositions of zoned phengite-rich part of the sample.They were not found in clino- and orthozoisite that grew during a progradeburial clear grain contact but were within a few micrometers of and heating event. each other. Closer inspection of thesecrystals with back- Sample St I (Fig. 5) is another example of modifica- scattered electron images revealed zoning from core to tions to primary zoisite during retrograde development. rim toward Fe-rich compositions in both the monoclinic This sample comes from the eclogitezone (Knappenhaus and the orthorhombic grains, although zoning in the cli- locality, Frosnitztal, Tauern Window, Austria), from a nozoisite was lessregular (Fig. 8). The monoclinic nature zoisite-quartz segregationin a banded eclogite (Thomas of those parts of the crystals with very low Fe content (8 and Franz, 1989). Following eclogite facies metamor- molo/0,determined by analytical electron microscopy)was phism, these rocks underwent subsequentalteration in verified by selectedarea diffraction analysis of two crys- the P-T rangefrom the lower amphibolite faciesdown to tals. The analysesclearly demonstrate that, at low P-T temperaturesbelow 350 "C (Spearand Franz, 1986).The 638 FRANZ AND SELVERSTONE: CLINOZOISITE-ZOISITE TRANSFORMATION

Fig 9. Backscatteredelectron image ofclinozoisite and zois- ite from eclogitesequilibrated at 2 GPa(sample DT-S), together with omphacite(O) and amphibole (black, not labeled).The om- phacitecrystal in the lowerleft growsacross the grainboundary betweenzoisite and clinozoisiteand also acrossthe core (Fe poor, 37 molo/oAlrFe) and the rim (Ferich, 49 molo/oAlrFe).

Fig. 10. (A) Backscatteredelectron image of zoisiteand cli- nozoisite,together with albite(black, unlabeled), (B) 20-1rmin- centimeter-sizedzoisite crystalsof the segregationare re- clusionsofclinozoisite in zoisite(close-up ofthe left centralpart placed along their rims by clinozoisite, paragonite,albite, of the zoisitecrystal in A). Matrix and inclusionclinozoisite and calcite, and adjacentto thesenewly formed minerals, showthe same type of zoning(sample RMTl). zoisite changesits composition from 6-10 molo/oAlrFe at the rim (coresare I l- l5 molo/o)to 16-21 mol0/0.The secondaryclinozoisite rangesfrom 39 ro 47 molo/oAlrFe. lesspronounced. The zoisitecomposition ranges from l0 SamplesFT-l and DT-8 (Figs.5, 9) are from the same to 15 molo/oAlrFe (one exceptionwith 18 molo/o),the areaas sample St l. They are both from banded eclogites clinozoisite from 34 to 52 mol0/0,with core compositions (localitiesFrosnitztal, southern margin of the eclogitezone, between 29 and 38 mol0/0.There are two averagecom- and Dorfer Tal, western end of the Gastacher Wiinde). positions of clinozoisite,one near 34, the other near 45 The eclogitesexperienced pressures near 2.0 GPa, at tem- molo/0.Figure 9 shows that omphacite inclusions are in peraturesnear 590 "C (Holland, 1979;Frank et al., 1987; textural equilibrium with both of thesecompositions. We Spearand Franz, 1986)with some subsequentreequili- use this observation to argue that the clinozoisite grains bration to lower greenschistfacies conditions. The FT with the high AlrFe component are not late-stageover- sample shows no signs of this reequilibration, whereas growths but rather formed during prograde metamor- the DT sample is slightly retrograded(as indicated by the phism. beginning of symplectite formation of some omphacite Sample RM 7l (Figs. 5, l0) is a garnet amphibolite and the scatteringof Ko values for garnet clinopyroxene). with strong indications that it is a retrograded eclogite The mineral assemblagein both samples is epidote + (symplectite of amphibole and plagioclase,corroded gar- garnet + omphacite * amphibole + phengite + dolomite net; the main mineral assemblageis amphibole + plagio- + magnesite + rutile + quartz. Clinozoisite is in most clase * chlorite + epidote + titanite). It comes from the cases strongly zoned with increasing Fe contents from basement complex of the Tauern Window, Maurer Tal. core to rim; in coexisting zoisite the trend is similar but Eclogiteswithin this basementrecord lower pressures(1.2 FRANZ AND SELVERSTONE: CLINOZOISITE-ZOISITE TRANSFORMATION 639

GPa) and temperatures(500 + 50 "C; Zimmermann and Toc Franz, 1989) than those within the eclogitezone. The coexistingzoisite-clinozoisite minerals occur in a small layer in the sample, coexisting with albite and garnet (+ paragonite),with good indications for textural equilibri- um. The zoisite crystalsare embeddedin clinozoisite and albite but also have inclusionsof clinozoisite.These in- /tw-interm e P clusions show the same type of zoning as the matrix cli- nozoisite, with very narrow rims of Fe-rich composition (42-54 mol0/oAlrFe), central parts with 3 I -34 molo/o(Fig. l0b). This is again taken as an argument for prograde zoning of the minerals not modified by retrogression.The zoisite rangesfrom 8 to 16 molo/oAlrFe, with core to rim Fig.rr. )-* ur^*')0"..0 -trr,"", .",",";r;"; zoning toward Fe-richcompositions. and text. The averagevalues for ":several analyses are shownby Sample RW 103 (Fig. 5) is a calc-silicatenodule in an the symbol;the horizontalline throughthe symbolgives the amphibolite. It comesfrom the Nubian Desert basement, rangeof composition(compare with Fig. 5 for numberof anal- approximately 30 km west of Delgo (20"N, Nile River) ysispairs). Symbols with doublelines refer to eclogites,and the in the northern province of Sudan.This basementis char- doubleline labeled"high P" indicatesthe two-phaseloop for acteized by a main stage of mineral formation in the high pressure.The solid symbolsrefer to the corecompositions upper amphibolite facies (P near 0.5 GPa, I near 650 ofthe highpressure clinozoisite. 'C), but contains relicts of granulite faciesmetamorphism (P : 0.8 GPa, temperaturesnear 800 "C; Bernau et al., that minor elements can influence the width of the gap. 1987;Huth et a1.,1984). The calc-silicatenodule consists This is evident for the REE (Fig. 5) and has also been ofgrossular + zoisite + diopside + anorthite + qtarlz, shown for Sr by Maaskant(1985). separatedby a small reaction zone ofclinozoisite + zois- In Figure I I we have arranged the data on coexisting ite + diopside + anorthite from the amphibolite (mainly clinozoisite-zoisite pairs (outer rim analyses)in a T-X plagioclase+ amphibole). Zoisite is strongly zoned from diagram, using averagevalues (mean of the analyses),but 2 to 12 molo/oAlrFe, with Fe-rich rims, but zonesof all also showing the range of analyses,together with an es- compositionscoexist with clinozoisite.The variationsin timate of the temperature of formation of the epidote clinozoisite composition (36-42 molo/oAlrFe) occur in minerals. The rocks were split into two groups, one with sharp zones parallel to the b axis. It is assumedthat the low to intermediate pressureof formation (< I GPa), the zoisite formed during granulite facies conditions above other with high pressure(> I GPa). These estimatesare the upper stability limit of clinozoisite,which is near650 rather uncertain: even when rigorous thermobarometric 'C (Holdaway,1972). During the amphibolite faciesevent, calculations are made, the timing of equilibration of the epidote component was formed, changing the zoisite epidote minerals is still unknown. Error boxesin the T-X composition and also forming separateclinozoisite crys- diagram would be large, but several key featuresare ev- tals. Although both minerals are in grain contact and ident despitethese uncertainties: (l) The overall chemical present in equal grain sizes, the large range in zoisite variation in zoisite is much smaller than in clinozoisite, rim compositions coexisting with clinozoisite indicates but both show an increasing AlrFe component with in- disequilibrium. creasingtemperature. This supportsthe previous findings of Enami and Banno (1980),Maaskant (1985), and Pru- DrscussroN nier and Hewitt (1985). (2) The largestcompositional A primary observation in nearly all of the samplesthat variations, in both zoisite and clinozoisite, were consis- we have studied is that both zoisite and clinozoisite com- tently observed in rocks of high-pressureorigin. High- monly show disequilibrium features, such as complex pressurerocks necessarilyhave a complex metamorphic zoning patterns (especiallyin rocks of high-pressureori- history with more possiblestages of mineral growth than, gin), formation of texturally and chemicallydifferent types, for example, during simple contact metamorphism. (3) and modifications ofearlier minerals.Within a singlegrain Clinozoisite samplesfrom the high-pressureterranes are the composition can vary widely and often irregularly, significantly higher (by approximately l0 molo/o)in AlrFe i.e., not parallel to the outer crystal faces. These varia- component than those from low-pressure terranes, al- tions in composition are controlled by mineral reactions though their estimated formation temperaturesare sim- producing epidote component in the rock and by crystal ilar. The same trend can be observedin Enami and Ban- growth phenomena such as nucleation and growth rates. no's (1980) data: Their FU and OM samples(their Fig. Intra- and intercrystalline exchangereactions seem to be 4) come from low-pressure terranes, whereas their TO, of minor importance, and we conclude that volume dif- NI, and IR samplesare from eclogites;the compositional fusion in the epidote mineral group is slow. Another im- gap for the high-pressurerocks is significantly wider. portant factor influencing the compositional variations These observationsprovide evidencethat pressureex- for coexisting minerals shown in Figure 5 may be the fact erts an important influence on the position of the clino- 640 FRANZ AND SELVERSTONE: CLINOZOISITE-ZOISITE TRANSFORMATION

TOC intergrowths of monoclinic and orthorhombic domains (Smelik and Franz, in preparation). Secondaryclinozois- ite crystals precipitated in cracks range from 30 to 54 molo/0.This indicates that at low P-T"conditions (approx- imately 0.3 GPa, 350 'C) no miscibility gap is apparent. The existenceof almost Fe-free epidote minerals (sym- metry not determined) in altered plagioclase,only a few micrometers away from the primary Fe-bearing zoisite and the secondary high-Fe clinozoisite, is typical of the disequilibrium preserved among neighboring grains of the epidote minerals. An orthorhombic symmetry of these Fe-free epidotes would be consistent with the transfor- mation temperature for Reaction I near 200 "C, as men- tioned above. There is also evidence to suggestthat miscibility gaps might occur in the monoclinic seriesat low temperatures or high pressures.Apparent miscibility gaps have been reported by several authors (Strens, 1965; Holdaway, Al2Al 2s so 75 Al2Fe L972;Hietznen, 1974; Raith, I 976; Selverstoneand Spear, 1985)on the basisof coexistingcompositions of mono- Fig. 12. Influenceof a possiblemiscibility gap at low 7, high clinic epidotes in natural rocks. Our own analysesand P, as indicatedby dataofSelverstone and Spear(1985). A mis- backscatteredelectron images ofepidotes show that dis- cibility gapat low temperaturesdemands a secondmonoclinic equilibrium textures are commonly preserved in single varietylabeled clz$. grains and even in coexistingmonoclinic and orthorhom- bic forms; it is thus possible that published reports of zoisite limb. This is supportedby observationsof Spencer miscibility gaps in the monoclinic series are merely an et al. (1988),who describedpairs ofzoisite and clinozois- artifact of complex growth histories. On the other hand, ite with l3- l4 and 40-51 molo/oAlrFe, respectively,from such effects as ordering of Al and Fe3* in Ml and M3 eclogites of the Western Gneiss Region (Norway), and sites may createstructural variations leading to miscibil- Maaskant(1985), who reporteda zoisite-clinozoisitepair ity gaps.We hope that detailed TEM studieswill resolve (16.5 and 45 molo/oAlrFe) from the high-pressurerocks this problem. Becausemost of the reported miscibility of Galicia (northwestern Spain). We cannot distinguish gaps are located at high AlrFe contents (between 35 and unambiguouslybetween the two possibilities discussedin 75 molo/o)and close at temperaturesnear 550 "C (Raith, Figure 2 becauseof the large variation in the composi- I 976)or evenlower (450'C; Selverstoneand Spear,I 985), tions, but a configurationsimilar to Figure 28 seemsquite they probably have no influence on the steeporthorhom- likely. bic to monoclinic transformation loop reported here. At temperaturesbelow approximately 400 .C the two- However, if the clinozoisite limb is significantly shifted phase region is not well constrained. Our sample 1266 toward Fe-rich composition with increasing pressure,as (see Fig. 5) shows that intermediate compositions for indicated in Figure I l, the suggestedmonoclinic misci- monoclinic crystals lying within the two-phase region at bility gaps might intersect the transformation loop. Such higher temperaturesdo exist. Therefore at low P-Z con- a configuration is shown in Figure 12 for the miscibility ditions, the limb restricting the monoclinic phase area gap proposedby Selverstoneand Spear(1985), with a extends toward very Al-rich compositions, but its exact critical temperature of 450 "C. position is still uncertain. An equilibrium temperaturefor Reaction I near 200 oCat low pressuresis likely, in agree- CoNcr-usroxs ment with calculations using the Berman (1988) data set This study supports the previous findings of Enami and and with the experimental determination of Jenkins et al. Banno (1980),Maaskant (1985), and Prunier and Hewitt (l e85). (1985) that the transformation loop between clinozoisite The secondaryepidote minerals formed in samples85- and zoisite depends strongly on temperature and that Fe3* l, 85-2, and St I show a large variation of compositions (as well as the minor elements Ti, Mq and Mg) is pref- and textures within a single thin section, reflecting pro- erentially distributed into clinozoisite. In addition, we longed cooling of the rocks. The primary zoisite crystals show that pressuremay also be a significant variable and with 10-12 molo/oAlrFe were transformed during a hy- that in high-pressurerocks clinozoisite is enriched in Fe3* drothermal event into grains with rtp to 26 molo/0,typi- compared to low-pressure rocks (at the same tempera- cally without changing the morphology and the straight ture, coexistingwith zoisite) at intermediate temperatures extinction of the crystal. The increasingFe content is vis- between400 and 600'C; zoisite-clinozoisiterelations re- ible in narrow zones with increasing birefringence. Pre- main uncertain at lower temperatures, although a nar- liminary TEM data suggestthat these zones consist of rowing of the compositional gap is likely, and at higher FRANZ AND SELVERSTONE: CLINOZOISITE.ZOISITE TRANSFORMATION 641 temperaturesthe upper stability limit of clinozoisite dis- rocks. Such studies have the possibility to yield timing turbs the simple binary phaserelations. information on the progradepaths ofhigh-pressure rocks The phasediagram that we presentfor the orthorhom- and thereby provide useful data for geodynamic studies bic to monoclinic transformation can be used as an aid of paleosubductionenvironments. in the calculationof reactionsinvolving epidoteminerals. For example, if the pressure-sensitivereaction anorthite AcxNowr,nncMENTS + HrO : zoisite-clinozoisite qvartz + kyanite + is con- The authors acknowledgea DFG grant Fr 557/!,4 (G.F.), a NATO sidered, its position in a P-T diagram depends on the travel grant (G.F. and J.S.),and an NSF grant EAR-8658145 (J.S.).Some structural state, as well as on the composition of the ep- ofthe sample material was collectedon field trips supportedby SFB 69. idotes. Calculationsinvolving such a reaction at high P-7" TEM work was carried out at Johns Hopkins University (Transmission grant conditions (eclogitefacies) should considerzoisite at bulk Electron Microscopy Facility, supported by NSF EAR-8903630). G.F. gratefully acknowledgesG. Smelik's help and introduction into TEM AlrFe compositions from 0 to l5 molo/0,coexisting zoisite work, as well as the kind permission of D. Veblen to use his laboratory. and clinozoisite in the range l5-45 molo/oAlrFe, and cli- We thank W. Heinrich, D. Hewitt, D. Jenkins, D. lattard, G. Smelik and nozoisite only at AlrFe contents greater than 45 molo/0. J.B. Thompson for discussionsand critical reviews ofthe manuscript. At low P-Z conditions (greenschistto amphibolite facies) the compositional gap is small and of only minor effect. RrrnnrNcas crrED Hence, at these conditions clinozoisite should be used in Abdel Rahman, E.M, Harms, U., Schandelmeier,H., Franz, G., Dar- the calculationsfor all bulk rock compositions that might byshire, D.P.F., Horn, P., and Mueller-Sohnius, D. (1990) A new yield more than l0-15 molo/oAlrFe; zoisitewould be the ophiolite occurrencein NW-Sudan-constraints on Late Proterozoic appropriate phaseonly for very Fe3*-poorcompositions. tectonism. Terra Nova, 2,363-376. Ackermand, D., and Raase,P. (1973) Coexistingzoisite and clinozoisite For a more temperature-sensitivereaction such as law- in biotite schistsfrom the Hohe Tauem, Austria. Contriburionsro Min- sonite : zoisite + kyanite + quafiz + H2O, the pressure eralogy and Petrology,42, 333-341. influenceis more pronounced:at low P and 7t,5-10 molo/o Berman, R.G. (1988) Intemally consistentthermodynamic data for min- AlrFe can be incorporated in zoisite, and the coexisting eralsin the systemNarO-KrO-CaO-MgO-FeO-FerOr-AlrOr-SiOr-TiOr- HrO-COr. Joumal of Petrology, 29, 445-522- clinozoisite has around 20 molo/o,whereas at low ?" and Bemau, R., Darbyshire, D.P.F., Franz, G., Harms, U., Huth, A., Man- high P the clinozoisite has around 40 molo/oAl,Fe. The sour, N., Pasteels,P., and Schandelmeier,H. (1987) Petrology, geo- likelihood of forming zoisite in addition to clinozoisite chemistry and structural development of the Bir Safsaf-Aswanuplift, increaseswith rising pressurebecause of the wider com- Southern Egypt. Journal ofAfrican Earth Sciences,6,79-90. (1980) interaction positional gap. This provides Bird, D.K., and Helgeson,H.C. Chemical of aqueous an explanation for the rel- solutions with epidote-feldsparmineral assemblagein geologicalsys- ative abundance of zoisite in high-pressure rocks, al- tems. l. Thermodynamic analysisof phaserelations in the systemCaO- though zoisite itselfis also stable at low pressure. FeO-FerOr-AlrOr-SiOr-HrO-CO,.Arnerican Journal of Science,280, Epidote minerals, owing to their stability over a wide 907-941. range of P-?" conditions, can be produced in many dif- Bird, D.K., Cho, M., Janik C.J., Liou, J.G., and Caruso,L.J. (1988)Com- positional, order/disorder, and stable isotope characteristics of Al-Fe ferent stagesduring the metamorphism of a single rock; epidote, state 2-14 dnll hole, Salton Sea geothermal system. Journal of the apparently slow diffi.rsionrates of Al e Fe3*exchange GeophysicalResearch, 93, (Bl l), 13135-13144. result in preservation of complex zoning patterns in in- Deer, W.A., Howie, R.A., and Zussman,J. (1986) Rock-forming miner- dividual grains. The epidote minerals thus have the po- als, vol. lB: Disilicatesand ring silicates,p. 2-l 5 1. Longman, london. (1968) tential provide wealth petrologic Dollase, W.A. Refinement and cornparison of the structures of to a of information rel- zoisite and clinozoisite. American Mineralogist, 53, 1882-1898. evant, for example, to the reconstruction of P-I paths, - (l 97 l) Refinementofthe crystal structuresofepidote, allanite and provided that quantitative data on P-T-X relations in the hancockite.American Mineralogist, 56, 447-464. epidote series are available. Though this is not yet the Enami, N., and Banno, S. (1980) Zoisite-clinozoisiterelations in low- to case,the empirical phasediagram presentedin Figure I I medium-grade high-pressure rocks and their implications. Mineralog- ical Magazine,43, 1005-1013. should serve as a starting point for future work. More Frank, W., Hoeck, V., and Miller, C. (1987) Metamorphic and tectonic analysesof coexistingmineral pairs from areaswith well- history of the central Tauem Window. In H.W. Fluegel and P. Faupel, known metamorphic histories are necessaryto refine this Eds., Geodynamicsofthe Fastern Alps, p. l-418. Deuticke, Vienna. diagram further. For fine-grained rocks, this will require Franz, G., Thomas, S., and Smith, D.C. (1986) High pressurephengile decomposition in the Weissenstein eclogite, Muenchberger Gneiss simultaneous determination of crystal structure and Massif, Germany. Contributions to Mineralogy and Petology,92,7l- chemical composition by transmission electron micros- 85. copy. Ghose, S., and Tsang, T. (1971) Ordering of V'*, Mn2* and Fe!* ions in Ultimately, single epidote grains might be useful can- zoisite CarAlSirOrr(OH). Science,I 7 l, 37 4-37 6. didates for radiometric agedeterminations with a micro- Gottschalk, M. (1990) Intern konsistentethermodynamische Daten im System SiOf,AlrOr-CaO-MgO-KrO-NarO-HrO-COr, 272 p. Ph.D the- beam technique,both becausethey can contain high con- sis, Universit.iit Tiibingen, Tiibingen, Germany. centrations of trace elementssuch as REE, Sr, Pb, U, and Hammerschmidt, K., and Franz, G. (1992) The influence of post-peak Th and becausethey can record different stagesofgrowth metamorphic alteration on 4O Ar/39 Ar white mica ages-a casestudy during complex metamorphic events; in this regard, they from the Miinchberger Gneismasse, Germany. Contributions to Min- press. may eventually servethe samepurpose in geochronologic eralogy and Petrology, in Helgeson,H.C., Delany, J.M., Nesbitt, H.W., and Bird, D.K. (1978) Sum- studies of intermediate-temperaturemetamorphic rocks mary and critique of the thermodynamic properties of rock forming that zircon does in higher temperature (e.g., magmatic) minerals. Arnerican Journal of Science,278 A, 229. 04z FRANZ AND SELVERSTONE: CLINOZOISITE-ZOISITE TRANSFORMATION

Hietanen,A. (1974) Amphibole pairs, epidote minerals,chlorite, and pla- coexisting zoisite and clinozoisite. American Mineralogist, 70, 375- gioclase in metamorphic rocks, northern Sierra Nevada, California. 378. American Mineralogist, 59, 22-40. Raith, M. (1976) The AI-Fe(III) epidote miscibility gap in a melamorphic Holdaway, M.J - (1972) Thermal stability of Al-Fe epidote as a function profile through the Penninic series of the Tauern Window, Austna. of/o, and Fe content. Contributions to Mineralogy and Petrology, 37, Contributionsto Mineralogyand Petrology,57,99-117. 307-340. Ray, N.J., Putnis, A., and Gillet, P. (1986) Polytlpic relationshipbetween Holland, T.J.B. (1979) High water activities in the generation of high clinozoisite and zoisite. Bulletin de Min6ralogie, 109, 667-685. pressurekyanite eclogitesof the Tauern window, Austria. Journal of Seki, Y. (1959) Relation betweenchemical composition and lattice con- Geology,E7, l-27. stants of epidote. American Mineralogist, 44, 72O-7 30 Holland, T.J.B., and Powell, R. (1990) An enlargedand updated inter- Selverstone,J., and Spear,F.S. (1985) Metamorphic P-T paths from pe- nally consistentthermodynamic datasetwith uncertaintiesand corre- litic schistsand greenstonesfrom the south-westTauern Window, east- lations:The systemItO-NarO-CaO-MgO-MnO-FeO-Fe,Or-Al,O,-TiOr- ern Alps. Journal of Metamorphic Geology, 3,439-465. SiO,-C-H,-O,. Journal of Metamorphic Geology, 8, 89-124. Spear,F.S., and Franz, G (1986) P-T evolution ofmetasediments from Hdrmann, P.K., and Raith, M. (1971) OptischeDaten, Gitterkonstanten, the Eclogite Zone, south-centralTauern Window, Austria. Lithos, 19, Dichte und magnetischeSuszeptibilifit von Al-Fe(III)-Epidoten. Neues 219-234. Jahrbuch fiir Mineralogie Abhandlungen, I 16, 4l-60. Spencer,J.C., Valley, J.W., and Essene,E.J. (1988) Petrology of a mar- Huth, A., Franz, G., and Schandelmeier,H. (1984) Magmatic and meta- garite-bearing meta-anorthosite from Seljeneset,Nordfiord, westem morphic rocks of NW Sudan: A reconnaissancesurvey. Bediner geo- Norway: Implications for the P-T history of the western GneissRegion wissenschaftlicheAbhandlungen (A), 50, 7-2L during Caledonian uplift. Lithos, 21, ll7-128. Jenkins, D M., Newton, R.C., and Goldsmith, J.R. (1985) Relative sta- Strens, R.G.J (1965) Stability relations of the Al-Fe epidotes. Mineral- bility of Fe free zoisite and clinozoisite. Journal of Geology, 93, 663- ogical Magazine,35, 464-47 5. 672. Thomas, S., and Franz, G. (1989) Kluftminerale und ihre Bildungsbern- Johannes,W. (1985) The significanceofexperimental studiesfor the for- gungenin Gesteinender Eklogitzone/Siidvenedigergebiet(Hohe Tau- mation of migmatites.In J.R. Ashworth, Ed., Migmatites,302 p. Blackie, ern, Osterreich). Mitteilungen der dsterreichischengeologischen Ge- Glasgow. sellschafr.81. 189-218. Klemd, R. (1989) P-T evolution and fluid inclusion characteristicsof Trdger, W.E. (1982) Optische Bestimmung der gesteinsbildendenMi- eclogites,Muenchberg Gneiss Complex, Germany. Contributions to nerale, Teil 1, p. l-188. Schweizerbart'scheVerlagsbuchhandlung, Mineralogy and Petrology, lO2, 221-229. Stuttgart Langer, K., and Lattard, D. (1980) Identification oflow-energy OH-va- Vogel,D.8., and Bahezre,C. (1965)The compositionofpartially zoned lence vibration in zoisite. American Mineralogist, 65,179-783. garnet and zoisite from Cabo Ortegal, NW Spain. NeuesJahrbuch fiir Maaskant, P. (1985) The iron content and the optic axial anglein zoisite Mineralogie Monatshefte, 140-149. from Gaficia, NW Spain. Mineralogical Magazine, 49, 97- 10O. Zimmermann, R., and Franz, G. (1989) Die Eklogite der unteren Schie- Myer, G.M. (1966) New data on zoisite and epidote. American Journal ferhiiile; Frosnitztal/Siidvenediger(Tauern, Osterreich). Mitteilungen of Science,264, 364-385. der dsterreichischengeologischen Gesellschaft, 8 1, I 67- I 88. Pistorius, C.W.F.T. (1961) Synthesisand lattice constantsof pure zoisite and clinozoisite. Journal of Geology, 69, 6O4-609. Me,m;scnrr"rREcErwD Mnv 17, l99l Prunier, A.R., and Hewitt, DA (1985) Experimental observations on Mnxr;scnrm AccEprEDJ,+ruenv 16, 1992