American Mineralogist, Volume 71, pages 946-954, 1986

Pyroxene exsolution in granulites from Fyfe Hills, Enderby Lando Antarctica:Evidence for 1000t metamorphic temperatures in Archean continental crust

Mrcn,l'nr, SlNorrom" Rocnn Pownr,r-Department of Geology, University of Melbourne, Parkville, Victoria 3052, Australia

Arsrucr

Orthopyroxene (Opx) and augite (Aug) in metasedimentary and meta-igneous gneissesin a supracrustal succession at Fyfe Hills, Enderby Land, Antarctica, exhibit intimateexsolution intergrowths. In magnesian assemblages, exsolution lamellae occur only on 100as a result of exsolution of Opx from Aug and Aug from Opx. In more Fe-rich assemblages,exsolution lamellae occur on both 00 I and 100 in Opx. The 00 I lamellae indicate exsolutioninvolving pigeonite (Pgt) subsequently "inverted" to Opx and Aug. In intermediate assem-blages, Opx-Aug intergrowths are blebby rather than lamellar, a feature attributed to thedecomposition of primary metamorphic Pgt in the three-phase field of Opx-Pgt-Aug. Thetextural relations and compositional data suggest the following phase relations at peakconditions: Opx-Aug (for XB. < 0.43), Opx-Aug-Pgt (for XPl.:0.5-0.43), Opx-Pgt (forXPj- > 0.5), Aug-Pgt (for Xfy > 0.48), where Xr" : Fe2+/(Fez+ + Mg). Application of Lind-sley's (1983, American Mineralogist, 68,477-493) phase diagram at the relevant pressure(9 kbar) implies a peak metamorphic temperature of at least 1000t.

INrnonucrroN

The Napier Complex granulites in Enderby Land, Ant-arctica, contain a number of exceptionally rare assem-blages indicative of exceedingly high temperature meta-morphism, including sapphirine + quartz * osumilite(Ellis et al., 1980; Grew, 1982); sapphirine * orthopy-roxene * mesoperthite + quartz (Sandiford, 1985a); andpigeonite (now inverted) + magnetite + qlrurtz (Grew,1982). These assemblages imply that the Napier Complexrepresents one of the highest-temperature granulite ter-ranes known, along with the granulites of I-abwor Hills,Uganda (Nixon et al., 1973); Wilson Lake, Canada (Morseand Talley, 1971); and Scourie, Scotland (Barnicoat andO'Hara, 1979). Establishing the temperatures at whichsuch granulite terranes formed is particularly importantin understanding granulite-facies metamorphism and itsrole in crustal growth and differentiation. In this contri-bution we determine the peak metamorphic temperaturesrecorded by the Napier Complex granulites using con-straints imposed by compositional and textural relationsin complexly exsolved metamorphic pyroxenes from theFyfe Hills region of the Napier Complex.The texturallybased approach to thermometry used herein circumventssome of the problems common in conventional thermom-etry that are caused by chemical re-equilibration duringslow cooling of rocks from very high temperatures (Es-sene, 1982; Powell, 1985).

Gnor,ocrc sETTTNG or rnr FyrnHrr.r.s cRANULTTES

The Fyfe Hills occur near the western boundary of anArchean gneissic terrane termed the Napier Complex

0003404x/86/0708-0946$02.00 946

(Sheraton et al., 1980). Two distinct lithologic associationsare found in the Napier Complex (Sheraton and Black,1983). The oldest association consists of pelitic, psam-mitic, and femrginous metasedimentary rocks intimatelyinterlayered with meta-igneous rocks of presumed vol-canic origin. The younger association is an intrusive meta-igneous sequence that is believed to have been emplacedimmediately prior to or during peak metamorphism. Al-though the meta-igneous rocks from both lithologic as-sociations are similar petrographically, they are readilyseparated by their distinctive field occurrence and trace-element geochemistry (Sheraton and Black, 1983; San-diford and Wilson, 1986). All pyroxene-bearing rocks dis-cussed in this paper apart from sample R25384B belongto the older supracrustal association. Sample R25384B isinterpreted as a dike, intruded and isoclinally folded priorto the peak metamorphism (Sandiford and Wilson, 1984).

Sapphirine-quartz assemblages in the metapelites arewidespread in the central and southwest part ofthe NapierComplex (Sheraton et al., 1980) and indicate that tem-peratures throughout this region exceeded 800-850'C(Newton, 1972; Newton et al., 1974). On the basis of thedistribution of osumilite and equivalent parageneses inbulk compositions approaching KFMAS, Sandiford(1985a) argued that metamorphic temperatures were ap-proximately equivalent affoss the region in which sap-phirine-quartz occurs, an area ofabout 100-km diameter(see also Grew, 1982). These assemblages do, however,indicate a systematic increase in pressure from northeastto southwest across the Napier Complex, suggesting thatthe current exposure of the Napier Complex representsan oblique profile through the lowermost part of an Ar-chean continental crust (Sandiford, 1985a).

SANDIFORD AND POWELL: PYROXENE MICROSTRUCTURES IN GRANULITES 947

The quantification of the temperature of formation ofthis lower-crustal, nearly isothermal section has proveddifrcult. Estimates based on pyroxene thermometry afterWells (1977) are in the range 78G980"C (Ellis, 1980; Grew,1980; Black et al., 1983; Sandiford, 1985a). The lowerend of this range is defined by the estimates of Grew (840-890'C) and Sandiford (780-920"C), neither ofwhom at-tempted to reintegrate the composition of the ubiquitousexsolution lamellae found in the highest-grade pyroxeneassemblages. Neither Ellis (1980) nor Black et al. (1983)indicated whether their temperature estimates were basedon reinteglated pyroxene compositions. Given that ex-perimentalists claim that they achieve equilibrium in thissystem at temperatures as low as 1000'C, calculated tem-peratures ofthis order in slowly cooled rocks are unlikelyto reflect peak metamorphic temperatures. Thus, the sig-nificance of published pyroxene thermometry in NapierComplex gneisses must be questioned, although the es-timates of Grew (1980) and Sandiford (1985a) may pro-vide absolute minima.

Pressures calculated from the garnet-aluminosilicate-plagioclase-quartz barometer (Newton and Haselton, I 9 8 I )and the garnet-orthopyroxene-plagioclase-quartz barom-eter (Newton and Perkins, 1982) suggest that the Fyfe Hillsgranulites equilibrated at 9 kbar (Sandiford, 1985a).

PynoxrNn ASSEMBLAGES AT FyFE Hrr-r,s

Because of the extremely anhydrous nature of the Na-pier Complex granulites, pyroxene forms a common com-ponent of both meta-igneous and metasedimentary rocktypes. In the supracrustal sequence at Fyfe Hills, ortho-pyroxene and clinopyroxene occur together in femrginousmetasedimentary rocks (typically composed of subequalproportions of pyroxene, magnetite, and quartz) and inbasic and intermediate meta-igneous gneisses presumablyderived from volcanic rocks (Sandiford and Wilson, 1986).Representative whole-rock analyses of the assemblagesdescribed herein are presented in Table l. Owing to thisdiversity of bulk composition, two-pyroxene assemblagesexhibit a wide range of X." [:Fe,+/Mg + Fe2*)] with Xr"rangrng from 0.14 to 0.70. All samples with coexistingorthopyroxene (Opx) and augite (Aug) show evidence ofexsolution involving Opx and Aug and in some casespigeonite (Pgt). The extent ofexsolution and the texturalcomplexity ofthe exsolution intergrowths are, in a generalfashion, dependent on the X." of the participant pyrox-enes, suggesting that the highest-temperature Fe-rich py-roxenes showed considerable solid solution into the in-terior of the pyroxene quadrilateral, with the more Fe-rich bulk compositions "seeing" the stability field of pi-geonite.

In the following discussion we attempt to reconstructthe peak metamorphic phase relations for the pyroxenesin these rocks through the interpretation of exsolutiontextures. This reconstruction provides a useful basis forevaluating the peak temperature of metamorphism of theFyfe Hills granulites, a basis that is, to a large extent,independent ofthe pronounced Fe-Mg exchange that con-

Table l. Representative xnr whole'rock analyses ofpyroxene-bearing rocks described in this paper; all Fe as

Fe,O.

R 2 5 4 0 2 R 3 1 0 3 1

s i 0 2

T 1 0 2

A 1 2 0 3

F " 2 o 3

M n 0

M o n

C a O

N a 2 0

K 2 op n' 2 " 5

L O I

T o t a l

5 1 . 2 3

0 . 5 3

L 4 . t 7

1 0 . 3 0

o . t 7

8 . 0 6

I I . 3 2

z . t 2

0 . 0 6

- 0 . 2 2

9 9 . 3 5

4 2 . 4 3

U . J )

0 . 6 1

i o I q

0 , 0 7

7 . 9 4

L . 2 4

0 . 1 5

0 . 0 1

0 , 0 4

3 . 9 3

o o o o

tinued to much lower temperatures during cooling andsubsequent re-equilibration.

The samples described herein were collected from theFyfe Hills and the adjacent islands of Khmara Bay overan area approximately 200 km'zin size. This region formsa coherent tectonic unit (Sandiford and Wilson, 1984),and the distribution of assemblages in metapelites (whichinclude sapphirine + garnet + qtrartz and sillimanite +hypersthene + quartz) suggests that temperatures at thepeak of metamorphism did notvary significantly through-out this region (Sandiford, 1985a).

AN,lr,yrrcAl, TECHNTeUES

Electron-microprobe analyses for orthopyroxene and clino-pyroxene from two-pyroxene assemblages are presented in Tables2 and 3, respectively. Analyses were performed on the MelbourneUniversity JXA-5A wavelength-dispersive instrument operatingat 15 kV and with a beam width of l0 pm. The raw data werereduced according to the method ofFerguson and Sewell (1980).The analyses presented in Tables 2 and 3 represent individualspot analyses of pyroxenes. Except where otherwise indicated,the analyses represent the composition ofthe host pyroxene. Noattempt has been made to reintegrate the composition of lamellae$iith hosts in these analyses. Structural formulae for pyroxeneshave been recalculatedaccording to the method oflindsley(1983).

Pnrnocru.prrv oF pyRoxENE-ExsoLUTroNINTERGROWTHS

Three distinct types of pyroxene-exsolution intergrowthoccur in the Fyfe Hills gneisses (Figs. l-3). Each type isrestricted to a distinct compositional range as indicatedby the X.. ofthe coexisting pyroxenes.

Pyroxene-exsolution textures in Mg-rich rocks

In the most magnesian rocksthat containprimary coarse-grained Opx and Aug, the Opx has fine exsolution lamellae

948 SANDIFORD AND POWELL: PYROXENE MICROSTRUCTURES IN GRANULITES

Table 2. Electron-microprobe analyses oforthopyroxenes, recalculated according to Lindsley (1983)

R 2 5 4 0 2 R 3 1 0 9 6 R 2 5 8 1 2 R 2 5 8 1 2 p . 2 5 5 4 7

h o s t h o s t h o s E l a n h o s t

R 2 5 3 8 4 R 2 5 3 9 0 R 2 5 3 9 3 R 3 r 0 3 1

h o s E l a n 1 a n I a D

R 2 5 3 8 4 R 2 5 6 9 9 R 2 5 5 9 3

l a o l a n h o s t

s t o ^T l o ^

A I . O ^

F e ^ o ^

F e O

M n 0

C a O

N a 2 O

T o t a I

T i

A 1- 3 +

, +

M n

M g

N a

4 8 . 4 1

0 . 1 r

0 . 8 9

o . 7 6

0 . 5 9

1 0 . 8 2

0 . 6 5

0 . 0 0

9 9 . 6 4

5 1 . 2 0 5 0 . 4 5

0 . 0 8 0 . 0 5

t . 3 2 1 . 3 9

0 . 5 5 3 . 0 3

2 6 . 4 8 2 5 . O 1

0 . 4 4 0 . 5 5

1 8 . 8 8 1 8 . 6 9

0 . 5 6 0 . 8 8

0 . 0 0 0 . 0 8

9 9 , 5 2 1 0 0 , r 2

5 1 . 3 1 5 0 . 8 1

0 , 1 2 0 . 0 0

0 . 9 8 0 . 5 5

2 . 7 4 0 . 6 7

2 6 . 6 t 3 1 . 0 6

o . 2 4 0 . 3 1

r 9 , 0 3 1 5 . 0 8

0 . 5 4 0 . 4 6

0 , 0 0 0 . 0 3

r 0 r . 5 7 r 0 0 . 0 7

4 9 . 0 9 4 8 . 9 8 4 9 . 3 5

0 . 0 1 0 . I 0 0 . 0 1

r . 0 0 0 . 7 6 0 . 5 6

r . 7 9 1 . 9 I 2 . 7 6

3 2 . 4 4 3 2 . 5 5 3 3 . s 3

o . 4 5 0 . 3 4 0 . 3 I

1 3 , 9 7 1 3 . 9 5 r 3 . 7 1

0 . 6 1 0 . 6 8 0 . 5 8

0 . 0 3 0 , 0 0 0 . 0 0

9 9 . 3 9 9 9 . 2 9 r 0 0 . 8 r

4 8 . 8 8 4 9 . 6 3

0 . 0 0 0 . 0 3

0 . 7 3 0 . 2 5

1 . 8 5 0 . 0 0

3 4 . 5 2 3 7 . 8 4

o . 2 4 0 . 2 3

1 2 . 8 1 1 1 . 1 3

0 . 6 7 0 , 9 0

0 . 0 0 0 . 0 6

9 9 . 7 0 1 0 0 . 0 7

r . 9 5 5 0 1 . 9 9 5 9

0 . 0 0 0 0 0 . 0 0 0 9

0 . o 3 4 4 0 . 0 1 1 9

0 . 0 5 5 6 0 . 0 0 0 0

1 . 1 5 4 6 t . 2 7 2 7

0 . 0 0 8 1 0 , 0 0 7 8

o . 7 6 3 6 0 . 6 6 7 I

o . o 2 a 7 0 . 0 3 8 8

0 . 0 0 0 0 0 , 0 0 4 7

4 8 , 8 9 4 8 . 3 5

0 . 0 7 0 . 0 7

0 . 0 7 0 . 0 5

0 . 5 6 1 . 2 3

3 9 . 0 8 3 9 . 8 4

0 . 0 3 0 . r 4

1 0 . 4 9 9 . 4 5

0 . 5 6 0 , 5 8

0 , 0 0 0 . 0 4

9 9 . 7 5 9 9 . 8 4

r . 9 6 0 0 | . 9 2 6 8 1 . 9 3 5 9 | , 9 1 6 4 r . 9 5 0 6 r . 9 5 0 6 1 . 9 4 5 8

0 . 0 0 2 3 0 . 0 0 1 4 0 . 0 0 3 4 0 . 0 0 0 0 0 . 0 0 0 3 0 . 0 0 3 0 0 . 0 0 0 3

0 . 0 5 9 5 0 . 0 6 2 6 0 . 0 4 3 6 0 . 0 2 9 8 0 . 0 4 6 8 0 . 0 3 5 7 0 . 0 2 6 0

0 . 0 1 5 8 0 . 0 8 7 0 0 . 0 7 7 8 0 . 0 1 9 6 0 . 0 5 3 7 0 . O 5 7 2 0 . 0 8 1 8

0 . 8 4 7 9 0 . 7 9 8 7 0 . 8 3 9 8 1 , 0 1 0 3 r . 0 7 7 9 1 . 0 8 4 5 1 . 1 0 s 6

0 . 0 1 4 3 0 . 0 1 7 8 0 . 0 0 7 7 0 . 0 1 0 2 0 . 0 1 5 1 0 . 0 I I 5 0 . 0 1 0 4

L 0 1 7 t r . 0 6 3 8 r . 0 7 0 0 0 . 9 3 2 2 0 . a 2 7 3 0 . 8 2 8 5 0 . 8 0 5 6

0 . 0 2 3 0 0 . 0 3 6 0 0 . 0 2 1 8 0 . 0 1 9 2 0 . 0 2 6 0 0 , 0 2 9 0 0 . 0 2 4 5

0 . 0 0 0 0 0 . 0 0 5 9 0 . 0 0 0 0 0 . 0 0 2 3 0 . 0 0 2 3 0 . 0 0 0 0 0 . 0 0 0 0

1 . 9 5 3 8 1 . 9 8 7 1 1 . 9 7 9 3

O . O O 3 4 0 . 0 0 2 I o . O o 2 2

o . 0 4 2 5 0 . 0 0 3 4 o . O o 2 4

0 . 0 2 3 3 0 . 0 I 7 0 0 . 0 3 7 8

L 2 6 5 3 1 , 3 2 8 8 1 . 3 6 3 9

o . o 2 o 2 0 . 0 0 1 0 0 . 0 0 4 9

0 . 6 5 3 3 0 , 5 3 5 6 0 . s 7 6 6

o . 0 2 8 2 o . 0 2 4 4 0 . 0 2 9 8

0 . 0 0 0 0 0 . 0 0 0 0 0 . 0 0 3 2

0 . 4 4 0 5 0 . 4 2 8 8 0 . 4 3 9 7 0 . 5 2 0 1 0 . 5 6 5 8 0 . 5 6 6 9 0 . 5 7 8 5 o . 6 0 1 9 0 . 6 5 6 1 0 . 5 5 9 5 0 . 6 7 6 4 0 . 7 0 2 9

o . o i , z o 2 0 . 0 1 9 8 3 o . o l t S o o . o o 9 9 o o . o t 3 8 2 o . o L 5 4 2 o . o r 3 2 2 o . o 1 5 r 8 0 . 0 1 9 6 4 o . o r 4 7 4 0 . 0 1 2 3 9 0 . 0 1 5 4 3

Note: Rocks R25402, R31096, and R25812 contain pyroxenes having only 100 exsolution lamellae; rocks R25547 and R25384contain Opx with blebby Aug; and the remaining rocks contain pyroxenes with coarse 001 lamellae of Aug and Opx (after Pgt).

Table 3. Electron-microprobe analyses ofclinopyroxenes, recalculated according to Lindsley (1983); see note, Table 2

R 2 5 4 0 2 R 3 1 0 9 6 R 2 5 8 r 2

h o s t h o s t h o s t

R 2 5 5 4 7

b 1 e b

R 2 5 3 8 4 R 2 5 3 9 0

b l e b h o s t

R 2 5 3 9 3 R 3 1 0 3 1 R 2 5 3 8 4 R 2 5 6 9 9 R 2 5 6 9 3

h o s t h o s t h o 6 t h o s t l a n

s i o ^

T l o ^

A t ^ 0 -

F e O

U n O

M g o

C a O

N a 2 0

T o t a l

T t

A I1 +

F e -

, +

M n

C a

N a

5 1 . 1 0 4 9 . 9 5

o . 2 5 0 . 2 9

2 . 1 9 3 . 2 0

2 . O 5 2 . 8 5

8 . 9 7 7 . 8 8

0 . 1 9 0 . 2 4

1 2 . 6 2 1 r . 9 9

2 r . 3 5 2 1 . 6 7

0 . 5 0 0 . 5 9

9 9 . 2 3 9 8 . 6 6

r . 9 3 3 2 r . 9 0 0 9

0 . 0 0 7 I 0 . 0 0 8 3

o . o 9 7 7 0 . 1 4 3 6

0 . 0 5 8 5 0 . 0 8 1 6

0 . 2 8 3 9 0 . 2 5 0 7

0 . 0 0 6 1 0 . 0 0 7 7

0 . 7 r 1 5 0 . 6 8 0 0

0 . 8 6 5 4 0 . 8 8 3 6

0 . 0 3 6 7 0 . 0 4 3 5

5 0 . 8 3 5 1 . 0 4

0 . 1 5 0 . 1 0

2 . 4 6 r . 7 9

2 . 5 5 2 . 3 9

1 0 . 2 9 1 2 . t 4

0 . 1 2 0 . 2 1

1 1 . 4 1 t O . 2 9

2 t . 3 4 2 r . 4 7

0 . 6 1 0 . 6 3

9 9 . 7 6 r 0 0 . 0 6

5 0 . 0 1 5 0 . 5 6

o . 2 9 0 . 1 2

r . 9 3 r . 7 4

2 . t 2 3 . 1 0

1 2 . 5 5 1 2 . 4 7

o . r 7 0 . r 29 . 9 3 1 O . 2 9

2 0 . 9 6 2 I . 6 8

0 . 6 0 0 . 4 0

9 8 . 5 6 r 0 0 . 4 8

4 9 , 7 0 s 1 . 3 3

0 . 0 8 0 , 0 1

1 . 8 1 0 . 4 4

1 . 4 2 2 . 3 1

r 4 . 4 7 1 5 . 1 9

0 . 0 9 0 . 1 0

9 . 4 7 8 . 9 8

2 1 . O 7 2 1 . 0 3

o . 2 3 0 , 5 5

9 8 . 3 4 1 0 0 . 5 4

L . 9 7 6 7

0 . 0 0 0 3

0 . 0 2 0 0

0 . 0 6 7 1

0 . 5 0 8 4

0 . 0 0 3 3

0 . 5 r 5 4

0 . 8 6 7 8

0 . 0 4 r 1

4 9 . 3 2 5 0 . 9 4 5 0 . 8 9

o . 2 4 0 . 0 0 0 . 0 0

2 . 0 6 0 , 1 8 0 . 0 5

2 . 3 5 2 . 2 9 1 . 5 4

1 5 . 5 1 1 5 . 3 5 1 7 . 1 9

o . 2 7 0 . 1 4 0 . 1 5

8 . 1 5 8 . 8 7 8 . 3 9

2 0 . 4 4 2 2 . O 3 2 L . 6 4

0 . 5 8 0 . 3 0 0 . r 8

9 8 . 9 3 1 0 0 . 1 0 1 0 0 . 0 2

r . 9 3 2 7 r . 9 7 3 7 1 . 9 8 3 1

0 . 0 0 7 r 0 . 0 0 0 0 0 . 0 0 0 0

0 . 0 9 5 2 0 . 0 0 8 2 0 . 0 0 2 3

0 . 0 6 9 3 0 . 0 6 6 9 0 , 0 4 5 1

0 . 5 0 8 4 0 . 4 9 7 2 0 . 5 6 0 1

0 . 0 0 9 0 0 . 0 0 4 6 0 , 0 0 5 0

o , 4 7 6 0 0 , 5 1 2 2 0 . 4 8 7 3

0 . 8 5 8 3 0 . 9 1 4 6 0 . 9 0 3 6

0 . 0 4 4 1 o . 0 2 2 5 0 . 0 r 3 6

0 . 5 r 6 5 0 . 4 9 2 6 0 . 5 3 4 8

o . 4 2 4 6 0 . 4 4 6 4 0 . 4 4 2 8

r . 9 2 6 9 1 , . 9 4 5 9 1 . 9 3 9 1

0 . 0 0 4 3 0 . 0 0 2 9 0 . 0 0 8 5

0 . 1 0 9 9 0 . 0 8 0 5 0 . 0 8 8 2

o , o 7 2 6 0 . 0 6 8 5 0 . 0 6 1 8

o . 3 2 6 2 0 . 3 8 7 1 0 . 4 0 7 r

0 . 0 0 3 9 0 . 0 0 6 8 0 . 0 0 5 6

o . 6 4 4 6 0 . 5 8 4 7 0 . 5 7 3 8

0 . 8 6 6 8 0 . 8 7 7 1 0 . 8 7 0 8

0 . 0 4 4 8 0 . 0 4 6 6 0 . 0 4 5 r

r . 9 2 7 8 r , 9 4 3 7

0 . 0 0 3 4 0 . 0 0 2 4

o . o 7 8 2 0 . 0 8 3 5

0 . 0 8 8 9 0 . 0 4 I 9

0 . 3 9 7 8 0 . 4 7 3 3

0 . 0 0 3 9 0 . 0 0 3 0

0 . 5 8 4 7 0 . 5 5 2 0

0 . 8 8 5 7 0 . 8 8 2 9

o . 0 2 9 6 0 . 0 1 7 4

o . 2 8 5 2 0 . 2 6 9 4 0 . 3 3 6 0

o . 4 2 4 7 0 . 4 2 2 3 0 . 4 2 3 5

0 . 3 9 8 4 0 . 4 1 5 0 0 . 4 0 4 8

o . 4 3 7 7 0 . 4 3 6 0 0 . 4 2 5 0

0 . 4 6 r 6 0 . 4 9 6 6

o . 4 2 4 4 0 . 4 4 1 4

Fig. 1. Exsolution texture in R25547 showing blebby exso-lution of Aug (stippled) in Opx host, with later augite lamellaeon 100. Note the elongation of Aug blebs parallel to 001 in B.The pyroxene grain has the bulk composition of, and is believedto be derived from, primary Pgt. The bar scale represents I mm.

ofAug on (100) and the coexisting Aug has fine exsolutionlamellae of Opx on (100). The proportion of exsolutionlamellae to host never exceeds 5olo in the primary Aug and2o/o in the primary Opx. In contrast to the more Fe-rich

949

assemblages, there is no textural evidence that Pgt wasstable in the primary metamorphic assemblage or that itdeveloped during the exsolution history of these assem-blages. These intergrowths are derived from primary co-existing Opx and Aug and are restricted to rocks in whichthe pyroxenes (host pyroxene without lamellae) have thecomposition X83- < 0.44 and Xf"* < 0.34.

Pyroxene-exsolution textures in rocks ofintermediate composition

Rocks ofintermediate Mg content contain both primarycoarse-grained Opx with a small percentage of thin 100Aug lamellae, as described for Mg-rich rocks, and a py-roxene intergrowth consisting of blebs ofAug up to I mmacross in an Opx host (Fig. l). The Aug blebs are typicallyblocky, or uue elongated parallel to (100) or parallel to aplane at a high angle to (100), presumably (001). Fine 100Aug lamellae occur in the Opx host, whereas very fine100 Opx lamellae occur in the Aug blebs. The proportionof Aug bleb to Opx host in individual aggregates is ap-proximately 1:3.7 (the average of estimates from fivegrains), indicating qualitatively that the bulk compositionofthe aggregates correspond to Pgt. In the two rocks foundwith the blebby exsolution texture, the Opx hosts havethe compositions X83- : 0,57, and the Aug blebs have thecomposition XfJ':0.41. There is no difference in com-position between Aug blebs and Aug lamellae.

SANDIFORD AND POWELL: PYROXENE MICROSTRUCTURES IN GRANULITES

\oor

Fig. 2. Exsolution texture in pyroxene aggregate in R25699. The exsolution history ofthese grains involved the formation ofbroad 001 lamellae of Pgt in the primary subcalcic Aug. Subsequently, the Pgt lameliae decomposed to Opx and Aug. Finer 100lamellae of Opx in Aug and Aug in Opx were then developed. Note the absence of exsolution-free grains along the boundaries ofthe large exsolved grains, indicating that granule exsolution was not a significant process. Photomicrograph is 3 cm wide.

950 SANDIFORD AND POWELL: PYROXENE MICROSTRUCTURES IN GRANULITES

Fig. 3. sBvr back-scattered electron image of exsolution texture in a pyroxene grain in R3 I 03 1. Aug (dark colored) and Opx (lightcolored) form broad lamellae subparallel to 001 and have subsequently exsolved finer 100 lamellae. Very fine "001" lamellae inthe Aug have been identified as Pgt. Primary pyroxene was a subcalcic Aug.

Pyroxene-exsolution textures in Fe-rich rocks

In the most Fe-rich rocks, pyroxene-exsolution inter-gowths contain broad lamellae up to 0.5 mm across ona plane close to (001) (the "00l" lamellae of Robinson etal., l97l1, Robinson, 1980; Figs. 2-3). The lamellae mayconsist either of Opx in an Aug host, or vice versa, andin both cases both the host and "001" lamellae containfiner (<20 pm wide) 100 lamellae of the coexisting py-roxene. In intergrowths in which Aug is the host, the ratioof Aug to Opx varies from 3:l to l:l;in intergrowths inwhich Opx is the host, the ratio of Aug to Opx is typicallyabout 1:4. Because of bulk-composition arguments andthe fact that "001" exsolution between Opx and Aug isnot allowed (Robinson et al., l97l; Robinson, 1980), allexsolution intergrowths involving "001" lamellae musthave involved Pgt. In pyroxene aggregates with "001"lamellae, the Aug component has compositions withXfy > 0.46, and the Opx component has compositionswith Xfl3- > 0.58.

INlnnpnnuuoN oF THE ExsoLUTroN TExruREs

The exsolution textures in the Fyfe Hills pyroxene in-dicate that at peak conditions the stable assemblages in-cluded Opx-Aug in Mg-rich assemblages and Pgt-Aug andOpx-Pgt in intermediate and Fe-rich assemblages.

The exsolution features in primary Opx-Aug assem-blages can be attributed to the widening ofthe pyroxenesolvus with decreasing temperature without any changein stability relations among quadrilateral pyroxene phas-es. This is consistent with the primary Opx'Aug assem-blages occurring in the most Mg-rich assemblages (withXFj- < 0.44). The exsolution features in the most Fe-richrocks can be attributed to the widening of the two-phaseAug-Pgt field with decreasing temperature, with exsolu-tion taking place prior to the decomposition of Pgt to Augand Opx.

The blebby intergrowths in the rocks of intermediateMg content are regarded as the product of decompositionof Pgt to Opx and Aug as the three-phase field Pgt-Opx-

SANDIFORD AND POWELL: PYROXENE MICROSTRUCTURES IN GRANULITES 9 5 1

Fig. 4. Schematic phase relations among pyroxenes in CaO-MgO-FeO-SiO, in terms of Wo : Ca/(Ca + Mg + Fe) and X :Fe/(Fe + Mg), in the form of isothermal sections for a series oftemperatures (in degrees Celsius) constructed for 9 kbar (afterLindsley,1983).

Aug tracks to more Fe-rich compositions during cooling.Thus, the blebby intergrowth is a reaction texture ratherthan an exsolution texture. Ishii and Takeda (1974) pro-posed a similar explanation for blebby intergrowths in"inverted" pigeonites from Kintokisan in Japan. The ab-sence of lamellar exsolution prior to decomposition of thePgt, suggests that these assemblages crystallized close tothe three-phase field Opx-Aug-Pgt at peak temperatures.As textural evidence for the coexistence ofthe three pri-mary phases has not been found, this interpretation mustremain somewhat conjectural. However, it is supportedby the general compositional considerations provided byother exsolution textures; the less Fe-rich compositionshave only primary Opx and Aug whereas the more Fe-rich compositions have either primary Opx and Pgt orAug and,/or Pgt. Moreover, the nearly identical compo-sitions of the pyroxenes in the two assemblages is consis-tent with this explanation.

The compositional dependence of the exsolution inter-growths is readily explained with the aid of the experi-mentally determined phase relations for quadrilateral-sys-tem pyroxenes oflindsley (1983), Figure 4. Note that thethree-phase assemblage Opx-Aug-Pgt moves to more Fe-rich bulk compositions with decreasing temperature. Thus,matching the inferred compositions of the pyroxenes inthe primary assembages, particularly those involving Pgt,with the isothermal sections in Figure 4 allows the tem-perature corresponding to peak metamorphic conditionsto be estimated. In fact each ofthe three textures describedabove and their inferred primary assemblages provide aconstraint on the peak metamorphic temperature.

Assemblages derived from primary Opx-Cpx in the moreMg-richrocks provide an absolute maximum temperaturefor metamorphism because they provide a compositionalbound on the position of the three-phase field Opx-Aug-Pgt in the quadrilateral at peak conditions (Fig. 5). Theposition of this compositional bound is provided by themostFe-rich sampleswith primary Opx-Aug, i.e., R25812.In this rock there is a considerable difference in X." be-tween host Opx and Opx lamellae in Aug. Thus, Fe-Mgexchange during cooling had ceased by the time the Opxlamellae had achieved their present form. The tie line

wo

XFig. 5. Phase relations for pyroxenes in the more Mg-rich

rocks exemplified by R258 12. Thinner lines are positions of Opx-Aug boundary of the Opx-Aug-Pgt tie triangle for indicated tem-peratures. The analyzed hosts and lamellae are shown. See textfor discussion.

between the Aug host and the Opx lamellae is much shal-lower than the high-temperature tie lines determined ex-perimentally by Lindsley (1983), indicating substantialhostJamellae Fe-Mg equilibration during cooling. Giventhe small proportion of Opx lamellae in Aug and thesmaller proportion of Aug lamellae in Opx, integratedpyroxenes will not be displaced far along the plotted tielines. The anticlockwise rotation of the tie lines causedby the higher-temperature intercrystalline Fe-Mg re-equil-ibration can be undone qualitatively if the bulk compo-sition about which the tie lines rotated is known. As thiscomposition is unknown, although Aug predominatesmodally, this rock provides an upper temperature limitof about I100'C if the bulk composition approached thatof the Aug or a limit of 1060"C if it approached that ofthe Opx.

In the rocks of intermediate composition, the blebbyintergrowths are believed to be derived from the break-down of Pgt that formed part of the three-phase assem-blage, or formed at compositions only slightly more Fe-rich than the three-phase assemblage, in the Pgt-Aug orPgt-Opx field. Consequently, the composition of the Pgtprovides an excellent constraint on the position of thethree-phase assemblage at peak conditions. The compo-sition ofthe bulk blebby intergrowth that has a proportionof Aug to Opx of l;3.7 should corespond to the com-position ofthe primary Pgt. Graphical reconstruction yieldsa Pgt composition -$g' : 0.53 for both R25547 andR25348B (Fig. 6). At peak conditions, this Pgt could havebeen stable with Opx (XP:.: 0.50) or Aug (XSJ': 0.48)or both. For the three-phase assemblage Pgt-Opx-Aug, apeak temperature of 1020'C is implied. Higher temper-atures are implied if the Pgt was originally part of eitherof the two-phase assemblages Opx-Pgt or Pgt-Aug. Sup-port for the interpretation that the blebby intergrowthsoriginally formed part of the three-phase assemblage Pgt-Opx-Aug is provided by the observation that in the twosamples in which it occurs, the pyroxene compositionsare nearly identical.

R258t2@ @ t N

iloo

9s2 SANDIFORD AND POWELL: PYROXENE MICROSTRUCTURES IN GRANULITES

X

Fig. 6. Phase relations for pyroxenes in intermediate com-position rocks. The dashed tie triangle, for the limiting case ofPgt bulk composition (open circle) in these rocks, gives a mini-mum temperature of 1020.C for the formation of the pyroxenesin these rocks. See text for discussion.

In the more Fe-rich rocks the primary pyroxene assem-blages involved crystallization in the Aug-Pgt or Opx-Pgtfields, or as a single subcalcic Aug. The pyroxenes in theserocks provide a lower temperature limit because they pro-vide a compositional bound on the position of the three-phase field Opx-Pgt-Aug, in analogous fashion to the Mg-rich assemblages. Moreover, the inferred subcalcic natureof some of the primary Aug provides some independentevidence of extremely high temperatures. Here the lowertemperature limit is provided by the most Fe-rich pyrox-enes. Granule exsolution is a problem (Bohlen and Essene,1978); R25699 is preferred to R25390 for providing thislimit because there is good textural control for this rock(Fig. 2), and granule exsolution is not suggested. The pri-mary pyroxene in R25699 was a subcalcic Aug, nowcoarsely exsolved, making analytical reintegration difr-cult. In this case, reintegration was achieved by combiningthe compositions of the analyzed Aug and Opx lamellaein proportions appropriate to their areas. This somewhathazardous procedure gives an approximate compositionWoo.rFsorrEno.r,. This composition provides a minimumtemperature of 980"C, because the primary three-phasefield Aug-Pgt-Opx must have existed at more Mg-richcompositions (Fig. 7). Higher temperatures, well in excessof 1000"C, 31s suggested by the subcalcic nature of thereintegrated Aug composition, although the actual tem-perature cannot be estimated because it is strongly de-pendent on the Wo content of the primary pyroxene, whichis sensitive to small errors in reintegration.

In summary, the pyroxene-exsolution textures andcompositional data combined with Lindsley's (1983) ex-perimental data suggest that metamorphic temperaturesat Fyfe Hills were in the vicinity of 1000'C. A preferredtemperature of 1020'C is indicated by our interpretationof the origin of blebby intergrowths in rocks of interme-diate composition and is supported by the upper and lowerbound constraints imposed by the textural and compo-sitional data in both more Mg- and more Fe-rich assem-blages.

DrscussroN

The exsolution textures in supracrustal pyroxene gran-ulites from Enderby Land are indicative ofexceptionallyhigh metamorphic temperatures. Indeed, the tempera-tures suggested by the pyroxene granulites are significantlyhigher than temperatures deduced from more conven-tional geothermometry. Support for these exceedingly hightemperatures is provided by the occuffence of variousunusual mineral assemblages in metapelites from the Na-pier Complex including Mg-rich sapphirine (Spr) + quartz(1P8. : 0. I 9-0.25), Ca-rich mesoperthite (orrrAbuoAn,r),and Mg-rich osumulite (Osu) * quartz (XP;" : 0.03-0.21)(Ellis et al., 1980; Grew, 1980, 1982; Sandiford, 1985a).The disparity between temperatures deduced from thepyroxene-exsolution textures (> 1000"C) and the conven-tional two-pyroxene thermometers provides additionalevidence that extreme caution must be employed whenusing conventional thermometry to deduce peak temper-atures in high-temperature, slowly cooled metamorphicrocks (see also Essene, 1982).

Although the Enderby Land granulites represent an ex-ceptionally high temperature metamorphism, they are notunique. Similar pyroxene textures have been recorded fromgranulites in the Scourian of Scotland (Barnicoat and O'-Hara, l9i9), and similar metapelitic assemblages havebeen recorded from a number ofterranes including WilsonLake, Labrador (Morse and Talley, l97l), and LabworHills, Uganda (Nixon eI al., 1973). The similarity in as-semblages suggests that all these terranes may have formedat temperatures close to or in excess of l000qC. Such hightemperatures are significant because they represent verysubstantial departures from the equilibrium crustal geo-therm. Steady-state lower-crustal temperatures in modernshields are generally estimated to be in the range 400-500qC (Turcotte and Schubert, 1982). That such pertur-bations resulting in lower-crustal metamorphism of su-pracrustal rocks at temperatures of 1000.C may be rathermore common than indicated by their current represen-tation at the Earth's surface is suggested by the post-

980

R25699

900

o o Lo 303 04 ou

x 06 o.7

Fig.7. Phase relations for the more Fe-rich rocks exemplifiedby R25699. The estimate of the original bulk composition andthe compositions of Aug and Opx lamellae are shown. DashedOpx-Aug-Pgt triangJe is for the temperature at which Pgt breaksdown for this bulk composition. See text for discussion.

o o

SANDIFORD AND POWELL: PYROXENE MICROSTRUCTURES IN GRANULITES 953

metamorphic peak pressure-temperature-time paths of theexceptionally high temperature terranes in Enderby Land,Labwor Hills, and Wilson Lake, which show isobaric cool-ing (Ellis, 1980; Sandiford, 1985a). This implies that theseterranes continued to reside near the bottom ofthe con-tinental crust for long periods after metamorphism. Thesubsequent exposure ofthese terranes at the surface oftheEarth is due to unrelated tectonic thickening of the crustoccurring much later than the thermal perturbation. Inthe case of the Enderby Land granulites, near-surface em-placement occurred in late Proterozoic time some 1.0-1.5Ga after granulite metamorphism (Sandiford, 1985b).Thus, many terranes metamorphosed at exceedingly hrghtemperatures may still reside near the base of modern-day continental crust.

How are such extreme perturbations of continentalgeotherms generated? Because of the inferred formationof the high-temperature terranes near the base ofthe crust,abnormal heat flow from the mantle is implied. Whetherthe abnormal heat flow was produced by accretion of man-tle-derived magmas, or was related to a dramatic thinningofthe continental lithosphere, is unknown, although thereis some evidence that metamorphism at Fyfe Hills wasbroadly contemporaneous with the emplacement of sill-like basic igneous bodies (Sandiford and Wilson, 1986).An observation that does not discriminate between mech-anisms of heat supply but that is important with respectto the consequences of the high temperatures is that thereappears to be no significant temperature gradient acrossthe highest-temperature part of the Napier Complex de-spite pressure variations in excess of 2 kbar. The impli-cation is that the dT/dP was very small. One explanationfor this is that the temperature was buffered via the en-thalpy of melting and, particularly, the loss of graniticmelts upward. In Enderby Land this is supported by thedry residual character of the granulites (Sandiford andWilson, 1986).

In light of this possibility it is interesting to note thatthe metamorphic temperatures in Enderby Land closelyapproach those required for dry melting of quartz + alkalifeldspar (1050-1100'C, Huang and Wyllie, 1975). Wewould propose that this melting reaction provides an im-portant constraint on the maximum temperature for thepreservation of residual supracrustal granulites duringcrustal metamorphism. It suggests that the peak meta-morphic temperatures at Fyfe Hills were not governed bythe heat supply, but rather by an internal thermal bufferingprocess. Whatever the nature of heat supply from themantle, it must have been efficient and must have pro-vided a considerable amount of heat to the lower crust.

AcxNowr,nrcMENTS

The Antarctic Division of the Department of Science is thankedfor the opportunity to collect these rocks during the 1979-1980Australian Antarctic Research Expedition. D. H. Lindsley andS. Bohlen are thanked for critically reviewing an earlier versionof the manuscript.

RnrenBNcnsBarnicoat, A.C., and O'Hara, M.J. (1979) High temperature py-

roxenes from ironstone at Scourie, Sutherland. MineralogicalMagazine, 43,371-375.

Black, L.P., James, P.R., and Harley, S.L. (1983) Geochronologyand structure of the early Archaean rocks at Fyfe Hills, EnderbyLand, Antarctica. Precambrian Research, 21, 197-222.

Bohlen, S.R., and Essene, E.J. (1978) Igneous pyroxenes frommetamorphic anorthosite massifs. Contributions to Mineral-ogy and Petrology, 65,433.

Ellis, D.J. ( I 980) Osumilite-sapphirine-quartz granulites fromEnderby [,and, Antarctica: P- I conditions of metamorphism,implications for garnet-cordierite equilibria and the evolutionofthe deep crust. Contributions to Mineralogy and Petrology,74,20r-2r0.

Ellis, D.J., Sheraton, J.W., England, R.N., and Dallwitz, W.B.( I 980) Osumulite-sapphirine-quartz granulites from EnderbyLand, Antarctica-Mineral assemblages and reactions. Con-tributions to Mineralogy and Petrology, 72, 123-143.

Essene, E.I. (1982) Geological thermometry and barometry. InJ.M. Ferry, Ed. Characterization of metamorphism throughmineral equilibrium. Reviews in Mineralogy, 10, 153-206.

Ferguson, A.K., and Sewell, D.K.B. (1980) A peak integrationmethod for acquiring X-ray data for online microprobe data.X-ray Spectrometry, 9, 48-5 1.

Grew, E.S. (1980) Sapphirine-quartz association from the Ar-chaean rocks of Enderby Land, Antarctica. American Miner-alogist, 65, 821-836.

- (1982) Osumulite in the sapphirine-quartz-terrane of En-derby Land, East Antarctica: Implications for osumulite pet-rogenesis in the granulite facies. American Mineralogist, 67,762-78'1.

Huang, W.L., and Wyllie, P.J. (1975) Melting relations in thesystem NaAlSirO8-KAlSi3OE-SiO, to 35 kbars, dry and withexcess water. Journal of Geology, 83, 737-7 48.

Ishii, T., and Takeda, H. (1974) Inversion, decomposition andexsolution phenomena of terrestrial and extraterrestrial pi-geonites. Memoirs of the Geological Society of Japan, 11, 19-36.

Lindsley, D.H. (1983) Pyroxene thermometry. American Min-eralogist, 68, 47 7 -493.

Morse, S.A., and Talley, J.H. (1971) Sapphirine reactions in deep-seated granulites near Wilson Lake, central Iabrador, Canada.Earth and Planetary Science Lrtters, 10,325-328.

Newton, R.C. (1972) An experimental determination of the highpressure stability limits of cordierite under wet and dry con-ditions. Journal ofGeology, 80, 398-420.

Newton, R.C., and Haselton, H.T. (1981) Thermodynamics ofgarnet-plagioclase-AlSirOr-quartz geobarometer. In R.C. New-ton, A. Nawotsky, and B.J. Wood, Eds. Thermodynamics ofminerals and melts, 3l-147. Springer-Verlag, New York.

Newton, R.C., and Perkins, D. (1982) Thermodynamic calibra-tion ofgeobarometers based on the assemblage garnet-plagio-clase<rthopyroxene (clinopyroxenelquartz. American Min-eralogist, 67, 203-222.

Newton, R.C., Charlu, T.V., and Kleppa, O.J. (1974) A calori-metric investigation of the high-pressure stability of anhydrousmagrresium cordierite with application to gxanulite facies meta-morphism. Contributions to Mineralogy and Petrology, 44,295-3rt.

Nixon, P.H., Reedman, A.J., and Burns, L.K. (1973) Sapphirine-bearing granulites from Lubwar, Ug;anda. Mineralogical Mag-azine, 39, 42U428.

Powell, R. (1985) Geothermometry and geobarometry: A dis-cussion. Journal ofthe Geological Society oflondon, 142,29-38.

Robinson, Peter. (1980) The composition space ofterrestrial py-roxenes-Internal and external limits. In C.T. Prewitt, Ed. Py-roxenes. Reviews in Mineralogy, 7, 419-494.

Robinson, Peter, Jafe, H.W., Ross, M., and Klein, C., Jr. (1971)

954 SANDIFORD AND POWELL: PYROXENE MICROSTRUCTURES IN GRANULITES

Orientation of exsolution lamellae in clinopyroxenes and cli-noamphiboles: Considerations of optimal phase boundaries.American Mineralogist, 56, 909-939.

Sandiford, M. (1985a) The metamorphic evolution of granulitesat Fyfe Hills; implications for Archaean crustal thickness inEnderby Land, Antarctica. Journal of Metamorphic Geology,3 , 155-178.

- (1985b) The origin ofretrograde shear zones in the NapierComplex: Implications for the tectonic evolution of EnderbyIand, Antarctica. Journal of Structural Geology, 7, 47 7 -488.

Sandiford, M., and Wilson, C.J.L. (1984) The structural evolu-tion of the Fyfe Hills-Khmara Bay region, Enderby Land, EastAntarctica. Aushalian Journal ofEarth Sciences, 31, 403-426.

- (1986) The origin of Archaean gneisses in the Flb Hillsregion, Enderby Land: Field occurrence, petrography and geo-chemistry. Precambrian Research, 31, 37 -68.

Sheraton, J.W., and Black, L.P. (1983) Geochemistry of Precam-brian gneisses: Relevance for the evolution of the East Ant-arctic Shield. Lithos, 16, 273-296.

Sheraton, J.W., Otre, L.A., Tingey, R.J., and Ellis, D.J. (1980)Enderby Land, Antarctica-An unusual Precambrian high gradeterrain. Geological Society of Australia Joumal, 27 , l-18.

Turcotte, D.L., and Schubert, G. (1982) Geodynamics. Wiley,New York, 450 p.

Wells, P.R.A. (1977) Pyroxene thermometry in simple and com-plex systems. Contributions to Mineralogy and Petrology, 62,r29-r39.

MnNuscnrpr REcETvED Ocrospn 23, 1984MrNuscnrpr AccEprED M.qncn 18, 1986