Lithas,25 (1990) 171-188
171
Elsevier Science Publishers B.Y., Amsterdam
Garnet peridotite and associated high-grade rocks from Sulawesi, Indonesia H. Helmers, P. Maaskant and T.H.D. Hartel' Instltute aJEarth Sciences, Free University, De Baelelaan 1085, 1081 HV Amsterdam (The Netherlands) (Revised and accepted June 28, 1990)
LITHOS
ABSTRACT Helmers, H., Maaskant, P. and Hartel, T.H.D., 1990. Garnet peridotite and associated high-grade rocks from Sulawesi, Indonesia. In: M. Okrusch (Editor), Third International Eclogite Conference. Lithos, 25: 171-188. The effects of collision between three major plates define the geological development of eastern Indonesia. Garnet peridotite and associated granulite-facies contact rocks are described from two sites within the valley of the active Palu-Koro left-lateral, strike-slip fault crossing central Sulawesi. Disrupted parts of a medium- to low-grade metamorphic complex intruded by Neogene granite occur on both sides of the fault. Thermobarometry on minerals and fluids in the garnet peridotite reveals a re-equilibration path from a depth of 60 km upward. Chemistry points to metasomatic effects - isolated trace-element enrichment - by a COz-rich liquid and fluid in a peridotite of oceanic affiliation. The granulite shows an increase in T and incipient melting at the arrival of peridotite. The sequence of fluid inclusions of an evolving CO 2 -CH 4 (-N 2 )-H 2 0-bearing fluid defines a concave decompression path suggesting rapid uplift. Trace element chemistry of granulite with basaltic to peraluminious rhyolitic composition indicates island-arc affinity. The described history may well reflect the processes beneath a mantled gneiss dome, present as a coeval metamorphic aureole around the garnet peridotite outcrops.
Introduction Garnet peridotite, stable at pressures above 15 kbar, provides information about mantle processes. Its survival at the surface depends on rapid uplift (O'Hara, 1977), as evidenced by its presence as xenoliths in kimberlites and alkalibasalts, notably related to established cratonized shield environments. Some xenoliths show changes in composition due to the formation process of the transporting magma, while others have been picked up by the magma en route to the surface without chemical change. Another type of garnet peridotite occurs sparsely in unroofed ophiolite complexes which have undergone deep subduction (Obata and MorI Present address: Umversity of Calgary, Dept. of Geology and Geophysics, Alberta, Canada.
0024-4937/90/$03.50
ten, 1987). Their history is usually demonstrated by spinel cores surrounded by garnet or by zonation of the garnet porphyroblasts, both indicating increase in pressure. The garnet peridotite of central Sulawesi (formerly Celebes) outcrops in a tectonically active area at the junction of three major plates, the Australian, the Pacific and the Asian ones (Fig. 1). At present, it belongs to the lastnamed plate. In this region, continuing tear-off and rotation of miniplates occurs with changes of direction and intensity with time (Hamilton, 1979). Locally, the fragments have been welded together again. Characterization of garnet peridotites in such areas contributes to the understanding ofthe behaviour ofthe upper mantle in plate-collision processes. The other source of direct information on the mantle is the young oceanic lithosphere, which is obducted as part of ophiolite
© 1990 - Elsevier Science Publishers B.Y.
172
H. HELMERS ET AL.
southeast from Palu, and north of Labua, 74 km south-southeast from Palu, where garnet peridotite together with various types of granulite outcrop on the valley floor (Fig. 1). No field relations between the rock-types were observed. Around the three mentioned outcrop areas, scarce exposures of amphibolite-facies rocks and of Neogene volcanics occur. Some very fresh samples come from cobbles found adjacent to the granulite outcrops; these have been included in the present study.
Geological setting
Fig. 1. Configuration of geologic units in Sulawesi (after Hamilton, 1979, Sukamto, 1975 and Silver et aI., 1983). Triangles: Neogene Island Arc; w=melange windows; narrow vertical lining = metamorphites; solid rhombs=garnet peridotite and granulite occurrences. Icelets: active volcanoes of the Sangihe Island Arc. Line with squares: Median Line. Horizontal lining = blueschist belt; points=obducted ophiolite; white = micaschists, Mesozoic and Tertiary sedimentary rocks. Double lining and arrows: Miocene rifting and transform faulting. Heavy lines with teeth and arrows: faults bordering miniplates; wide vertical lining: minicontinents derived from New Gumea. Inserted: map of Indonesian Archipelo with map-area delineated and main plate boundaries indicated.
nappes in the collision zones of the area. During his expedition in 1929, Brouwer found garnet peridotite and granulite on one of his crossings of the interior of Sulawesi. The petrography has been described by Egeler ( 1947), who concentrated on the metamorphic development ofsedimentary and volcanic rocks in the northwestern part of Central Sulawesi. Three occurrences of ultrabasic rocks were sampled, all present in the up to 1500 m deep rift valley of the active Palu-Koro transform fault (Tjia et al., 1974). The first occurrence is at Sakedi, 35 km south-southeast ofPalu, where only talc serpentinites and picotite-serpentine marbles occur. The other two are located south of Tuwa, 52 km south-
The island of Sulawesi (Fig. 1) is a type example of a paired metamorphic belt (Miyashiro, 1973; Katili, 1978). Its geology has been described by Brouwer (1934) and summarized by Van Bemmelen (1949). Sukamto (1975) constructed an excellent map. Hamilton ( 1979) and Hartono and Tjokrosapoetro (1984) provided plate tectonic models for the development of Sulawesi. The two western arms and the western central part of Sulawesi (Fig. 1) feature a mid-Miocene to Cenozoic island arc (Sasajima et al., 1981). This arc has been developed on three different crustal segments, from south to north characterized as follows: (a) On the southwestern arm of the island, a midCretaceous melange wedge with chaotic internal imbrication and containing lower-Cretaceous blueschists and eclogites (Van Leeuwen, 1979), is overlain by a sequence of cherts, graywackes and pillow lavas filling an outer arc basin. After folding, Paleogene shallow-water sediments containing minor arc volcanics (31 Ma in age) were deposited. From the mid-Miocene onwards the volcanic arc developed discordantly until the present. (b) In western central Sulawesi, a metamorphic basement is discordantly overlain by metamorphosed marine shallow-water sediments (shalesgraywackes containing metamorphic and tuffaceous detritus) of Paleogene age. The volcanic arc started to develop in the mid-Miocene and contains volcaniclastic rocks. Thick molasse-type sediments of upper Miocene to Quaternary age occur locally, notably around the lower Palu basin (Sukamto, 1975 ). (c) On the northern arm of Sulawesi the volcanic arc developed from the mid-Miocene onwards on an upper-Cretaceous to Paleogene oceanic crust
GARNET PERIDOTITE AND ASSOCIATED HIGH-GRADE ROCKS FROM SULAWESI
(Trail, 1974), containing pillow basalts, dike swarms and deep-sea sediments. Especially in the western central part and on the northern arm of Sulawesi, a large number of granitic to dioritic intrusives is exposed. The rocks composing the island arc (basalt-andesite-rhyolite) are calc-alkaline, but some Cenozoic volcanics on the west coast are alkaline. All radiometric ages of intrusives and extrusives range between 31 and 1.6 Ma, with a large majority between 17 and 5 Ma (Sukamto, 1975). The active volcanoes on the northeasternmost part of the northern arm belong to the Sangihe arc which extends north to the Philippines. On central Sulawesi, a steep west-dipping median line at 129 20' E separates the island arc in the west from the blueschist belt in the eastern and southeastern parts of the island (De Roever, 1947, 1950, 1953; Helmers et al., 1989). The latter constitutes the uplifted part of the west-dipping Paleogene subduction complex (Fig. 1). This complex comprises graphite-bearing micaschists and some mafic intrusives, overlain by radiolarian cherts and calcarenitic limestones of probable Cretaceous age, and basal epidote amphibolites of an obducted ophiolite complex further eastwards (Silver et al., 1981, 1983). Typical low- T-high-P minerals such as jadeite, lawsonite, aragonite, ferrocarpholite and blue amphibole occur (De Roever, 1947). In the west, Sulawesi is separated from Kalimantan by the Strait of Makassar, opened during a Miocene spreading event. The lower-Cretaceous subduction complex of southwestern Sulawesi is also exposed on southeast Kalimantan (Hamilton, 1979). The oceanic basins to the east and north of Sulawesi (Fig. 1) are of Cretaceous and Paleogene, and of Miocene age respectively (Lee and McCabe, 1986). They are probably parts of the Indian Ocean trapped during the collision. The Banda Sea in the east is cut by east-west transform faults between which Australian continental fragments moved westwards from New Guinea (Hamilton, 1979; Linthout et al., 1989; De Smet, 1989). To the west of the Banggai-Sula fragment (Fig. 1), an imbricate wedge has developed clnthe eastern tip ofthe northeastern arm of Sulawesi. The westward displacement of this fragment may also be responsible for the left-lateral movement along the active PaluKoro (Katili, 1978) and Matano fault systems and for the development of the connected active north Sulawesi trench in the Celebes Sea (Hamilton, 0
173
1979). This is in accordance with clockwise rotation of northern Sulawesi (Sasajima et al., 1981; Nishimura and Suparka, 1986). The eastern boundary of this new miniplate has been defined by Silver et al. (1983). The eastward growth of the Asian plate by renewal of subduction came to an end during the Miocene, when the configuration of the continental part of the Australian plate prohibited further accretion (Hamilton, 1979; Barber, 1981; De Smet, 1989). The increasing intensity of collision caused the break-up of major plate edges with the simultaneous development of small-scale arcs, subduction complexes and spreading zones between the miniplates in the area. The metamorphic complex of northwest central Sulawesi (Egeler, 1947), cut by the Palu-Koro fault, comprises a prograde, medium-P metamorphic belt ranging from the chlorite zone up to the staurolitekyanite zone. The prograde character is indicated along sections to the south and the northeast (Brouwer, 1934). Westwards, the overall presence of igneous rocks obscures early metamorphic zoning. Granulite occurs locally in the core of the metamorphic complex, on the bottom of the fault valley. A second phase oflow-P metamorphism overprints the first phase; it is related to the intrusion of granites and diorites. Staurolite is surrounded by a singlecrystal armour of cordierite, kyanite by andalusite and rarely by bundles of sillimanite. Arc volcanics are locally affected by this phase of metamorphism. A late phase of low-grade dynamometamorphism affected the intrusives and metamorphites. Although the ages of the metamorphic rocks and of the first phase of metamorphism are unknown, so far no radiometric pre-Cretaceous ages have been reported.
Rock descriptions
Medium-grained garnet peridotite (samples 15326, 15-328, 16-266, 16-271, 16-337 and 16-340) generally shows an equigranular foliated texture, a moderate degree of serpentinization and a strong kelyphitization of garnet. All minerals have an anhedral habitus and recrystallization features are uncommon. Olivine is usually quite fresh, although serpentinization cracks are commonly present. The olivine-garnet pair is apparently not stable as reac-
174
tion rims of orthopyroxene mantle olivines adjacent to garnets. Olivine inclusions within garnet, however, are observed without any reaction features. Orthopyroxene displays a more pronounced cleavage than clinopyroxene and sometimes shows exsolution lamellae, flakelets and rods of clinopyroxene, ilmenite and spinel. Garnets, often up to several mm in size, constitute above 5 vol. % of the rock and are invariably surrounded by kelyphitic rims, up to 2 mm across; kelyphitic alteration is also present along cracks in the garnets. A clear distinction between inner and outer kelyphitic zones is difficult. The host mineral phase orthopyroxene is crowded with vermicular, brownish green spinel and stringers and patches of light brownish amphibole together with minor clinopyroxene. Outwards from the garnet contact the vermicular spinel turns into worm-like spinel and further away into spinel blebs. Colourless clinoamphibole and chlorite replace clinopyroxene and garnet, respectively; talc is also occasionally present. Three types ofgranulites have been distinguished by Egeler (1947); (a) mafic clinopyroxene-garnet granulites, (b) felsic plagioclase-clinopyroxene granulites and (c) felsic feldspar-quartz-garnet granulites with locally kyanite and sillimanite. P-T estimates have only been obtained from types (a) (samples 16-336 and 16-338) and (c) (samples 16274 and 16-332). One sample of each of the three groups (15-322, 16-338 and 16-269, respectively) was analyzed chemically. The clinopyroxene-garnet granulite samples 16336 and 16-338 are medium-grained rocks showing marked cataclastic features which are most evident as a mortar texture of quartz and, to a lesser extent, plagioclase. Plagioclase, quartz, garnet and clinopyroxene constitute, in decreasing order, the main phases. Brown and secondary green biotite, brown hornblende, rutile, apatite, olive-green spinel, epidote, zircon and magnetite occur in minor amounts. Plagioclase (An20-30) forms large anhedral grains showing abundant antiperthite exsolution. Xenoblastic garnet is frequently surrounded by brown hornblende and magnetite. Light green clinopyroxene is present as discrete anhedral grains in the plagioclase-quartz matrix. Both garnet and clinopyroxene appear to be in microstructural equilibrium. The felsic granulite sample 16-274 is a mediumgrained, mylonitic rock in which garnet and quartz are the predominant minerals. The rock shows a
H. HELMERS ET AL.
complex set of coexisting but partly reacting minerals. A primary assemblage is represented by garnet-kyanite-quartz-rutile ± opaque oxides. Xenoblastic garnet is usually bordered by a multizoned corona: an inner rim of cordierite and plagioclase with threads of orthopyroxene is surrounded by a continuous rim of slightly pleochroic orthopyroxene, with an outer zone of plagioclase and magnetite. Figure 2 shows the idioblastic growth of cordierite over plagioclase. Sillimanite is present both as individual grains and as sheaf-like bundles; the mineral is often rimmed by an intergrowth of vermicular olive-green spinel and plagioclase. This intergrowth may have been derived from the reaction: 3CaM 2AbSi 30
12
+2AbSiOs=3CaAbSi20 g
+2MAb04 +2M 2Si20 6 +Si0 2
(1)
[garnet + silimanite (or relic kyanite) = plagioclase + spinel + orthopyroxene + quartz] in which M= (Mg,Fe2+). Kyanite occurs in minor amounts and is also rimmed by spinel. Biotite is usually found in close association with ilmenite and magnetite. Green spinel is also present as small grains enclosed in garnet. Rutile and pyrrhotite occur in large amounts. Graphite flakes occur near the garnet reaction zones.
Fig. 2. Backscattered electron image (somewhat distorted) of the kelyphite-forming reaction in sample 16-274. cor=cordierite, g{=graphite, gf=garnet, kel=kelyphite of cordierite and orthopyroxene, op = orthopyroxene, ox= magnetite, pi = plagioclase, q = quartz. Scale bar= 100,um.
GARNET PERIDOTITE AND ASSOCIATED HIGH-GRADE ROCKS FROM SULAWESI
In a first stage, plagioclase and orthopyroxene were formed by the reaction: CaM2A12Si3012 + Si02= CaAhSi20 s + M2Si20 6 (2 ), (garnet + quartz = plagioclase + orthopyroxene)
in which M= (Mg,Fe2+) Subsequent anhydrous reactions were: AhSiOs,kyanite=AhSiO s , sillimanite,
(3),
the GRISP reaction, calibrated by Bohlen and Liotta (1986): Ca3AhSi3012 +2Fe3A12Si3012 +6Ti0 2 = 3CaA12Si 2Os + 6FeTi0 3 + 3Si02
(4 ),
(garnet + rutile = plagioclase + ilmenite + quartz) and the symplectite reaction (1 ). The rising peridotite body released water by the reaction CO2 + CH 4 = 2H 20 + 2 C, causing the growth of graphite flakes (see below). With the influx of water, cordierite was formed: 6 (Ca,M hAhSi 30 12 + 7Si0 2+ H 20 = 6M 2Si20 6 + 4CaAhSi20 s + M2A14SisOls (H 20) (5) (garnet + quartz + water = orthopyroxene + plagioclase + cordierite ) Garnet porphyroblasts in the felsic granulite sample 16-332 are crowded with oriented opaque minerals and lesser amounts of sillimanite, kyanite, biotite, quartz and spinel. Sillimanite occurs in bundles, often around garnet, and frequently forms pseudomorphs after kyanite. Abundant dark brown biotite confers a distinct gneissose texture to the rock. Antiperthitic andesine occurs in the granulated matrix as well as in larger grains. Kyanite occurs in minor amounts; green spinel is present in symplectitic intergrowth with plagioclase. Pinite-like sericite patches have probably been derived from cordierite.
Geothermobarometry Electron-probe microanalyses were performed with a Cambridge Instruments Microscan 9 with two automated spectrometers and an on-line ZAF correction program. Accelerating voltage was 15 kV and
175
probe current 40 nA. Natural and synthetic minerals were used as standards. For each mineral phase, at least five grains were selected for analysis and attention was paid to the rim compositions of the garnets in the garnet peridotites. Representative analyses are given in Tables 1, 2, 3, and 4. Several geothermobarometers can be applied to garnet peridotites. The most consistent data were obtained by the methods of Ellis and Green (1979) and Raheim and Green (1974) for the clinopyroxene-garnet pair, the methods of Harley (1984), Wood (1974) and Sen and Bhattacharya (1984) for the garnet-orthopyroxene pair and the method of Wood and Banno (1973) for the orthopyroxeneclinopyroxene pair. The pressure-independent first three geothermometers of Mori and Green (1978) have also been incorporated. The five garnet peridotite assemblages, in which garnet is still a major phase and present as large grains, contrast with the one sample (16-337) in which only one small garnet remnant is present. They give the P-T estimates listed in Tables 5 and 6. Table 5 presents the mean values for the various methods, while Table 6 shows the mean values for the five samples. A P-T box of 1050-1100°C and 15-20 kbar is considered a reasonable P-T estimate for the garnet peridotites (Fig. 3 ). These P-T estimates were calculated under assumption that all Fe is present in the ferrous state. This assumption generally leads to maximum temperatures, except when applying the methods of Harley (1984), Wood (1974) and Wood and Banno (1973). The calculation of ferric iron has its greatest impact on pyroxene compositions, notably those of low-Fe clinopyroxenes. The maximum amounts of ferric iron are obtained by calculating on a basis of 6 oxygens and 4 sum cations, while intermediate values are given by the method of Papike et al. ( 1974). Propagation of analytical errors is the main reason for the authors to prefer Papike's et al. method to calculate maximum Fe3+. Compared with the all-Fe2+ P-T estimates, the use of the values of Papike et al. (1974) generally leads to T and P lower by about 40°C and 3 kbar, respectively. A small increase of Fe and Mn was usually n'oted at the garnet rims, but this did not lead to a lowering of the T estimates by more than 50 °C. The chemical compositions of the mineral phases included in garnet are commonly within the compositional range of the individual mineral phases. Alteration
176
H. HELMERS ET AL.
TABLE I Representative electron microprobe analyses of orthopyroxenes and clinopyroxenes Sample
15-326
15-328
SI02 Ah 0 3 TI02 Cr203 FeO MnO MgO NiO CaO Nap
54.8 3.7 0.11 0.35 6.4 0.13 32.8 0.09 0.61 0.06
55.7 3.5 0.14 0.39 6.3 0.12 32.8 0.12 0.65 0.10
Sum
99.05
99.82
16-271 55.8 3.4 0.13 0.46 6.3 0.15 32.8 0.11 0.80 0.05 100.0
16-340
16-337
16-266
16-2741
15-326
15-328
16-271
16-340
16-337
16-336
16-338
55.5 4.0 0.12 0.36 6.3 0.13 32.4 0.08 0.70 0.10
55.6 3.6 0.12 0.37 6.5 0.15 32.7 0.09 0.60 0.05
54.9 3.1 0.14 0.36 6.3 0.15 32.9 0.09 0.65 0.05
49.8 6.5 0.09
52.8 5.6 0.59 0.98 2.8 0.08 14.8 0.07 19.7 2.1
52.1 5.7 0.56
51.3 6.2 0.53 1.0 3.1 0.07 14.6 0.06 20.5 1.7
49.2 5.4 0.78
2.9 0.08 15.3 0.D3 19.9 1.7
52.5 5.6 0.52 0.84 3.0 0.07 15.2 0.02 19.9 1.7
49.9 4.9 0.60 11.7 0.46 11.5
11.8 0.23 11.1
0.15 0.02
52.0 5.2 0.42 0.89 2.7 0.06 15.4 0.08 20.4 1.7
18.0 1.6
18.7 1.4
99.69
99.78
98.64
99.70
98.85
99.52
99.47
99.35
99.06
98.66
98.61
1.86 0.29 0.003
1.92 0.24 0.016 0.028 0.085 0.002 0.80 0.002 0.77 0.15
1.90 0.25 0.Q15 0.034 0.087 0.002 0.83 0.001 0.78 0.12
1.91 0.24 0.014 0.024 0.092 0.002 0.83 0.001 0.78 0.12
1.88 0.27 0.015 0.03 0.095 0.002 0.80 0.002 0.80 0.12
1.90 0.22 0.017
1.88 0.24 0.021
0.37 0.013 0.65
0.38 0.007 0.63
0.73 0.12
0.76 0.10
4.013
4.019
4.013
4.032
4.02
4.018
23.0 0.24 19.9
1.2
Number of ions on the basis of 6 oxygens: SI Al TI Cr Fe Mn Mg Ni Ca Na
1.91 0.15 0.003 0.010 0.19 0.004 1.71 0.003 0.023 0.004
1.93 0.14 0.004 0.011 0.18 0.004 1.69 0.003 0.024 0.007
1.93 0.14 0.003 0.013 0.18 0.004 1.69 0.003 0.030 0.003
1.92 0.16 0.003 0.010 0.18 0.004 1.67 0.002 0.026 0.007
1.93 0.15 0.003 0.010 0.19 0.004 1.69 0.003 0.022 0.003
1.93 0.13 0.004 0.010 0.19 0.004 1.72 0.003 0.024 0.003
0.006 0.001
1.91 0.23 0.012 0.026 0.082 0.002 0.84 0.002 0.80 0.12
Sum
4.007
3.993
3.996
3.982
4.005
4.018
3.998
4.024
16-266
16-336
16-338
16-274
0.72 0.008 1.11
TABLE 2 Representative electron microprobe analyses of garnets Sample
15-326
Si0 2 A1 2 0 3 Ti0 2 Cr203 FeO MnO MgO CaO Na20 Sum
16-332b
16-332c
38.9 22.1
37.8 21.4
37.5 20.6
18.1 0.41 10.0 8.5
26.2 1.1 8.6 3.1
31.4 2.3 5.9 1.1
30.4 3.9 4.2 2.4
99.87
99.89
100.0
99.9
99.0
3.01 1.94 0.01
3.01 1.93 0.012
3.01 2.00 0.Q1
3.00 2.01
2.99 2.00
3.02 1.96
1.47 0.11 1.01 0.45
1.51 0.063 0.95 0.54
1.14 0.026 1.12 0.68
1.69 0.072 0.99 0.26
2.08 0.15 0.70 0.09
2.05 0.27 0.50 0.21
8.00
8.015
7.986
8.022
8.01
8.01
15-328
16-271
16-340
16-337
42.1 22.4 0.30 1.3 8.2 0.35 20.0 4.7 0.04
42.4 22.2 0.44 1.6 7.8 0.27 20.6 4.7 0.06
42.5 22.0 0.32 1.9 8.1 0.33 20.0 5.1 0.03
42.5 22.5 0.32 1.4 8.3 0.31 20.0 4.8 0.04
41.9 21.6 0.46 1.6 10.5 0.48 18.6 4.9 0.06
42.1 21.9 0.45 1.7 7.4 0.26 20.8 4.7 0.06
39.3 21.5 0.18
39.1 21.3 0.21
40.1 22.6 0.18
22.9 1.7 8.9 5.5
23.5 0.96 8.3 6.5
99.37
100.07
100.28
100.17
100.10
99.37
99.98
3.02 1.84 0.017 0.106 0.48 0.02 2.11 0.39 0.004
3.02 1.88 0.017 0.076 0.49 0.019 2.12 0.36 0.006
3.01 1.83 0.025 0.093 0.63 0.029 2.00 0.38 0.008
3.01 1.85 0.024 0.094 0.44 0.016 2.21 0.36 0.008
7.987
7.988
8.005
8.012
16-332"
Number of ions on the basis of 12 oxygens: Si Al Ti Cr Fe Mn Mg Ca Na
3.01 1.89 0.016 0.075 0.49 0.021 2.13 0.36 0.006
3.01 1.85 0.023 0.089 0.46 0.016 2.18 0.36 0.008
Sum
7.998
7.996
a
=large grains, b =small grain, = next to green biotite. C
177
GARNET PERIDOTITE AND ASSOCIATED HIGH-GRADE ROCKS FROM SULAWESI TABLE 3 Representative electron microprobe analyses of amphiboles and blOtites Sample
15-326
15-328
16-271
16-340
16-337
16-3381
15-328
16-336
16-340
16-274
16-274
16-332
SiOz Al z0 3 TiO z Cr203 FeO MnO MgO NIO CaO Na20 K 20
42.7 16.2 1.4 1.2 3.8 0.08 17.0 0.13 11.0 3.0
43.5 15.0 2.1 1.0 3.8 0.08 17.4
43.2 15.4 1.9 1.2 4.0 0.09 17.0
39.3 16.9 3.5 0.68 4.8 0.04 21.9 0.19
37.2 14.4 5.3
37.7 13.9 4.7
37.0 15.6 4.9
34.7 17.5 4.5
36.1 16.8 0.51
11.8 0.13 16.4
10.1 0.06 17.8
13.0 0.Q3 14.2
19.2 0.08 9.8
17.2 0.03 14.0
11.3 3.6 0.21
10.9 3.2 0.66
42.6 14.7 1.9 0.69 4.2 0.05 17.1 0.06 119 3.0 0.35
42.5 10.9 2.3
1.1
41.8 15.1 2.5 0.75 4.5 0.13 16.3 0.09 11.6 2.2 1.2
11.3 1.9 1.8
1.2 8.3
0.23 9.8
0.18 9.9
0.04 0.16 9.8
0.11 9.2
0.15 8.9
Sum
97.61
96.17
97.99
97.55
96.55
96.86
96.81
95.26
94.34
94.73
95.09
93.69
5.48 2.78 0.37 0.075 0.56 0.005 4.55 0.021
5.50 2.51 0.59
5.58 2.43 0.52
5.52 2.75 0.55
5.31 3.16 0.52
5.53 3.03 0.058
1.46 0.016 3.61
1.25 0.008 3.94
1.63 0.003 3.16
2.46 O.oJ 2.24
2.20 0.004 3.19
0.052 1.86
0.007 0.046 1.86
0.033 1.80
0.043 1.74
15.526
15.533
15.795
12.2 0.16 13.8
Number of ions on the basis of23 and 22 oxygens, respectively SI Al Tl Cr Fe Mn Mg Ni Ca Na K
6.08 2.72 0.15 0.14 0.45 0.01 3.61 O.oJ5 1.68 0.83 0.20
6.07 2.58 0.27 0.086 0.55 0.016 3.53 0.011 1.80 0.62 0.22
6.15 2.50 0.22 0.11 0.45 0.01 3.66
6.14 2.58 0.20 0.14 0.48 0.011 3.60
1.71 0.99 0.Q38
Sum
15.885
15.753
15.838
6.36 1.93 0.26
1.66 0.88 0.12
6.13 2.49 0.21 0.078 0.51 0.006 3.67 0.007 1.84 0.84 0.064
1.81 0.56 0.35
0.33 1.48
0.066 1.85
15.811
15.845
15.891
15.651
15.602
of clinopyroxene in sample 16-266 apparently does not affect garnet-orthopyroxene P-T estimates. The P-T conditions of spinel-orthopyroxene pairs in the kelyphitic rims around garnet have been constrained by the method of Gasparik and Newton (1984). Samples 16-271 and 16-337 give equilibration temperatures of 834°C and 865°C, respectively. The garnet-hornblende geothermometer of Graham and Powell (1984) has been applied and gives T estimates of about 850°C for three samples and 950°C for sample 16-328. These estimates were obtained using several solid-solution models for garnet, i.e. ideal solution and the models by Perkins and Chipera (1982), Ghent and Stout (1981) and Ganguly and Saxena (1984). It should be noted that according to Graham and Powell (1984), the thermometer is applicable at temperatures below about 850°C; its significance is therefore questionable in these garnet peridotites, although they show a reasonable fit with the spinel-orthopyroxene kelyphite equilibration temperatures given above.
1.53 0.021 3.07
15.64
Using the clinopyroxene-garnet calibration methods of Ellis and Green (1979), Rilheim and Green (1974), Saxena (1979) and Mori and Green (1978), the two clinopyroxene-garnet granulite samples 16-33.6 and 16-338 give T estimates highly depending on the assumed valency state of Fe. Differences of over 100°C were obtained for calculations with all Fe = Fe 2 or with maximum amounts of Fe3+. Applying the Papike et al. (1974) Fe 3 +_ calculation method to the thermometers of Ellis and Green (1979), Raheim and Green (1974), Saxena (1979), and Mori and Green (1978), equilibration temperatures of 926°, 927°, 983°, 953° were obtained for sample 16-336, and 927 0, 906 °, 901° and 920°C for sample 16-338, both calculated at 8 kbar pressure (T increases in all methods with less than 10 °C per kbar). These values fit well with P-T estimates in the felsic granulites (see below). Application of the Graham and Powell (1984) garnethornblende geothermometer yields temperatures of about 77 5°C for both samples. For the felsic-granulite sample 16-274, various
178
H. HELMERS ET AL.
TABLE 4 Representative electron microprobe analyses of ohvines, spinels, plaglOciases and cordierite Sample
15-326
15-328
Si0 2
40.8
41.1
16-271
16-340
16-337
41.3
41.2
41.1
16-2661
16-337
16-271
16-332
59.0
56.9
56.0 0.05
62.2
38.7 0.29 3.7
29.8 0.17 29.8
40.7
Ah0 3 TI0 2
Cr203 FeO MnO MgO NiO
10.0 0.10 48.6 0.38
9.6 0.13 48.6 0.43
9.9 0.13 48.6 0.42
10.0 0.13 48.3 0.43
9.6 0.16 49.0 0.41
9.4 0.11 49.3 0.34
7.6 12.0 0.15 19.8 0.41
16-274
16-3321
16-274
46.2 33.4
61.6 23.3
49.4 33.4 0.Q3
16-2741
11.8
11.2 0.07 19.3 0.30
0.57
CaO Na20 K 20 Sum
17.7
1.2 0.Q3
99.88
99.86
100.35
100.06
100.27
99.85
4.7 0.05 10.0 5.4 8.3 0.6
0.15 0.06 97.79
98.96
99.57
98.74
100.07
99.1
99.2
1.81
1.75
1.92 0.001
2.00
2.14 1.83
2.76 1.23
0.16 0.26 0.003 0.77 0.009
0.24 0.25 0.002 0.75 0.006
0.94 0.007 0.16
0.68 0.004 0.32
Number of ions on the basis of 4, 4, 8 and 18 oxygens, respectively
Si
1.00
1.01
1.01
1.01
1.01
1.00
Al Ti
Cr Fe Mn Mg Ni
0.21 0.002 1.78 0.008
0.20 0.003 1.78 0.008
0.20 0.003 1.77 0.008
0.21 0.003 1.76 0.008
0.20 0.003 1.78 0.008
0.19 0.002 1.80 0.007
0.022
Ca Na K Sum
3.00
3.001
2.991
2.991
3.001
2.999
3.012
2.998
3.028
3.004
5.03 4.00 0.002 0.40 0.004 1.52
0.88 0.11 0.002
0.25 0.72 0.Q3
0.Q3 0.008
4.984
4.99
10.994
TABLE 5
TABLE 6
p- T estimates for garnet peridotites (several P-T meters, all samples, except 16-266) (Tin °C)
P- T estimates for garnet peridotites (several samples, all P-T meters)
15 kbar
20 kbar
(J
Ellis and Green (1979) Raheim and Green (1974) Harley (1984) Wood (1974) Sen and Bhattacharya (1984) Wood and Banno (1973) Mori and Green (1978) Mon and Green (1978) Mori and Green (1978)
1068 1056 985 1005 1120 1041 1160 1121 1074
1088 1101 1017 1119 1166 1041 1160 1121 1074
30 26 22 41 34 19 43 52 59
mean value
1070
1099
geothermobarometers may be postulated. The authors are aware that the significance of P-T estimates from coexisting minerals in corona structures, as occurring in this sample, may be questioned. However, the outermost rim of most coronas is composed of polygonally shaped crystals like those forming the common rock fabric outside
(Till °C)
15-326 15-328 16-271 16-340 16-266
15 kbar
20kbar
(J
1025 1095 1075 1084 1078
1055 1124 1106
45 70 50 61
1113 1145
the coronas. Inside the coronas, vermicular crystals are present with increasingly finer dimensions inwards. This indicates that at least at the start of the corona-forming reaction equilibrium crystallization occurred. Moreover, the general agreement with P-T estimates from the other samples warrants their use for establishing the P-T path. The well-known reactions: 3CaAh Si 20 8 =Ca3A12Si3 0
12
+ 2Al2SiO s + Si02 (6)
(plagioclase = garnet + AI-silicate + quartz)
GARNET PERIDOTITE AND ASSOCIATED HIGH-GRADE ROCKS FROM SULAWESI 20
p (kb)
GNT+PLAG+KY GNT+OPX+PLAG
D
10
500
Fig. 3. P-T1oop for garnet peridotite, mafic and felsic granulite constructed from mineral equilibria and trapped fluids using various thermobarometers and fluid isochores. Explanation in the text. "Gran" are solid inclusions of granitic composition probably representing former melt. TABLE 7
P-T estimates for the high-pressure granulite assemblage T=750°C, P in kbar
N-P
P-C
G-S
Newton and Perkins ( 1982) 11.7 Perkins and Chipera (Mg-basis) (1985) 10.7 Perkins and Chipera (Fe-basis) (1985) 7.9 Ghent (1976) Newton and Haselton (1981) Koziol and Newton (1988)
12.2 Il.l 8.6 10.9 I\.9 12.3
12.8 1\.8
T=800°C, Pm kbar Newton and Perkins (1982) 12.0 Perkins and Chipera (Mg-basls) (1985) 10.9 Perkins and Chipera (Fe-basis) (1985) 8.5 Ghent (1976) Newton and Haselton (1981) Koziol and Newton (1988)
12.5 11.4 9.0 I\.9 13.0 13.2
13.0 12.0
11.7 12.8 13.3
12.6 13.7 13.7
Gh-S
11.4 12.4 12.8
12.3 13.4 14.1
Garnet solution models: N-P (Newton and Perkins, 1982), P-C (Perkms and Chipera, 1985), G-S (Ganguly and Saxena, 1984), Gh-S (Ghent and Stout, 1981 ).
has been calibrated by Ghent (1976) , Newton and Haselton (1982), Koziol and Newton (1988), and critically discussed by Bohlen and Lindsley (1988). Reaction (2) has been calibrated by Newton and Perkins (1982) and Perkins and Chipera (1985).
179
The use of various garnet solid-solution models produces considerable spread in the P-T estimates for both reactions, see Table 7. The difference between the P estimates based on the Mg- or Fe-end members of garnet is quite evident. As estimates from the Mg-end member are similar to those from reaction (6) and as orthopyroxene is richer in Mg than in Fe, it seems warranted to prefer the values obtained from the Mg-based reaction. The Ganguly and Saxena (1984) solid-solution model includes thermodynamical, experimental and empirical data on the mixing of garnet components. It also treats the mixing of almandine and pyrope as an asymmetrical solution with the greatest non-ideality for almandine-rich compositions. If these reasons are taken into consideration, a P- T box of 750-800 °C and 11.5-13 kbar is a fairly good estimate for the high-P assemblage (Fig. 3). In sample 16-274, sillimanite and kyanite contain 0.38 and 0.21 wt.% of Fe203, respectively; X FeTI 0 3 in ilmenite is 0.94 andXTio2 in rutile is 0.99. Sillimanite in sample 16-332 contains 0.94 wt.% of Fe203' Use fo the GRISP-reaction (4) leads to pressures of8.5 and 9.3 kbar at temperatures of 800 ° and 900°C, respectively. Garnet-cordierite temperatures, calculated according to Holdaway and Lee (1977), Ellis (1986), Perchuk and La~rent'eva (1983), Wells (1979), AranovlchetaL (1986) and Bhattacharya et al. (1988), all yield values between 900 and 950°C, showing minor vari.ations with P ( ± 10° C per kbar). Garnet-cordierite-sillimanite pressures may be evaluated on a Mg- or a Fe-end member basis. Calibrations for the Fe-system have been given by Thompson (1976), Holdaway and Lee (1977), Wells and Richardson (1980), Newton and Wood (1980), Bhattacharya (1986) and by Aranovich and Podlesskiy (1983), while for the Mg-system only the methods of Aranovich and Podlesskiy (1983), Aranovich et al. (1986) and Perchuk (1986) are available. The fluid composition of the cordierite (H 20-rich, COrrich, or dry) is crucial in the determinations: anhydrous cordierites yield P estimates which are more than 2 kbar:below those for hydrated ones. The P-T dependent X H20 calculation method of Aranovich and Podlesskiy ( 1983 ) has been chosen here, since it gi~es more consistent P-T estimates for the Mg- and Fe-based values than the XH20 values obtained by Lonker's (1981) method. With a calculated X H20 =0.7, the Aranov-
180
ich-Podlesskiy method gives pressures of 7.9 and 8.8 kbar at T=900°C for the Mg- and Fe-based reactions, respectively. Applying the cordierite-spinel-quartz and cordierite-garnet-orthopyroxenequartz reactions as calibrated by Bhattacharya (1986), pressures of 8.4 and 8.6 kbar were obtained at T = 900 °C (Fig. 3). When geothermometry is applied to garnet-late biotite pairs, only the methods of Ganguly and Saxena (1984), Perchuk and Aranovich (1986) and Perchuk and Lavrent'eva (1983) yield reasonable values of 832 0, 816 ° and 855°C, respectively. Other methods (e.g., both models A and B of Indares and Martignole (1985), Ferry and Spear (1978), Hodges and Spear (1982), Bhattacharya and Raith (1987) give Testimates of 1094°,1176°,1208°,1379° and 1722°C, respectively. These values nicely illustrate the statements of Essene (1982) that in high-grade metamorphic rocks "garnet-biotite thermometry is increasingly erratic" and that "in general, the garnet-biotite thermometer does not appear to work well in highgrade rocks". Similar considerations apply to the felsic granulite sample 16-332: the garnet-biotite temperatures are highly erratic and only Perchuk and Aranovich's (1986) and Perchuk and Lavrent'eva's (1983) approaches provide reasonable estimates of 936° and 989°C, respectively. Small garnets in combination with green biotite indicate retrogressive temperatures of 630°C. In this sample, the distinct replacement of kyanite by sillimanite marks the crossing of the kyanite-sillimanite boundary curve. The garnet-plagioclase-Al-silicate-quartz barometer also indicates P-Tconditions of750-800°C and 11.5-13 kbar.
Fluid and solid inclusions The numerous fluid and solid inclusions in the samples of garnet peridotite and granulite are, unfortunately, mainly very small in size. The phase behaviour of fluid inclusions was studied in the T range between _110° and + 35°C with a Chaixmeca heating-freezing stage cooled by liquid nitrogen (Poty et al., 1976). All temperatures were measured on heating after cooling to avoid metastability of the phases. Twenty-three fluid inclusions were analyzed by Raman microspectrometry, using a Microdil-28 with a multichannel detector (Burke
H. HELMERS ET AL.
and Lustenhouwer, 1987). The accuracy of such analyses depends on the amounts present and on bubble parameters such as size, depth below surface and fluid density, but is in the order of 10 to 20% relative. Isochores of the mixtures were constructed using the modified Redlich-Kwong equation ofstate (Holloway, 1977). The major-element composition of about 40 solid inclusions was obtained by the use of a scanning electron microscope with connected EDS-system giving an accuracy of about 1 wt.% for most major elements. The selected fluid inclusions are the largest ones present in the garnet porphyroblasts and quartz crystals of the felsic granulites. In the quartz crystals and quartz aggregates, three chronologically different groups of secondary fluid inclusions were distinguished. Their relative ages were determined by inclusion mapping of some domains (Touret, 1981) using trail intersections and transpositions. Frequently, zones with trails of fluid inclusions contain some tiny graphite flakes. Inclusions of the oldest recognized group contain up to 30 vol.% of H 20. Their vapour bubble consists of about 85 mole% of CO 2 and 15 mole% of CH 4 , their densities ranging from 0.84 to 0.77. The last-grown green biotite flakes cut the inclusion trails in quartz aggregates. Consequently, they had been formed during re-equilibration ofgarnet and biotite above 630 °C (see above). Inclusions of the next group are low in H 20 and have vapour phases consisting of up to 90 mole% of CH 4 besides CO 2 • Their densities are 0.30 to 0.35. They have been trapped after the growth of the last green biotite. The third and youngest group of inclusions is present in irregular H 2 0-trails with low amounts of CO 2 and CH 4 • Their densities range from 0.75 to 0.89. Most probably, this last group of high density inclusions has been trapped at temperatures below 400°C. The inclusions of the second group have been formed between 3 and 1.5 kbar in a T range between 400° and 630°C, trapping the low- T derivate of an evolving fluid in which the amount of CO 2 , H 20, CH 4 in the presence ofgraphite vary with T (Eugster and Skippen, 1967). Their compositions point to a T slightly above 400 °C. The first rather COr rich group, trapped at between 900 ° and 570°C and 3.5 to 5 kbar, shows a composition pointing to a T above 700 °C. The P- T estimates for the three groups of inclusions define a concave cooling path (Fig. 3) compatible with uplift asso-
181
GARNET PERIDOTITE AND ASSOCIATED HIGH·GRADE ROCKS FROM SULAWESI
ciated with delayed cooling. This is commonly explained by rapid uplift (Thompson and England, 1984) which can be expected also in the described geological setting. In the felsic granulite 16-274, garnet porphyroblasts contain mixed solid-fluid inclusions up to 10 .urn in size with a negative crystal shape showing a vapour bubble and outer films of H 20. The solid is not uniform in composition and its chemistry generally resembles that of a peraluminous granite (Hartel, in prep. ). The accompanying vapour bubbles are up to 3 .urn in size and contain CO 2 , CH 4 and N 2 in mole fractions of 60-70%,20-30% and 10%, respectively, indicating a low 102 (Green et al., 1987; Taylor and Green, 1987). During heating, the fluid bubbles decrepitate at about 925°C. They are too small to allow observations on homogenization. These composite inclusions may give evidence of incipient anatectic melting in the presence ofan immiscible fluid. A possible relation between them and a tiny granitic vein in the felsic granulite 16-338 needs to be studied in detail (Hartel, in prep. ). Tiny rutile needles accompany these solid inclusions. Solid inclusions are common in garnets and olivines of the garnet peridotite. Rhomb-shaped sections in combination with high-order interference colours indicate the frequent presence of carbonates. Ca, Mg and Fe are also present; locally a trace of S has been found. Three groups of carbonates have been distinguished: (a) A few calcite-dolomite-ankerite solid solutions have Ca-rich compositions and are above the 700°C immiscibility gap (Anovitz and Essene, 1987) but well within the 950°C solvus intersection (Fig. 4). (b) Magnesite-rich solid solutions, which contain up to 28 mole% of siderite and up to 12 mole% of calcite, are also within the 950°C solvus intersection. (c) Dolomite-magnesite intermediate compositions low in ankerite are present in a confined group (Fig. 4). Unmixing has not been observed. The dolomite-magnesite compositions represent the top oftheir solvus at 1400-1450 °C (Goldsmith and Heard, 1961; Goldsmith, 1980; Byrnes and Wyllie, 1981 ), which appears unlikely in relation to the equilibration T of the host minerals. According to Wyllie (1987b), at 1050-1 100°C magnesite is only stable in peridotite at a minimum pressure of 40 kbar. This also seems improbable when seen against
ANKERITE
MAGNESITE
SIDERITE
Fig. 4. Composition of carbonate inclusions. 700°C solvus intersection from and 950°C solvus constructed from data of Anovitz and Essene (1987).
the background of the complete re-equilibration of the silicates at 20-16 kbar. Unfortunately, the exact meaning of "magnesite" and "dolomite" is not well defined in these systems of extended solid solution. "Dolomite" is a stable solid phase in peridotite above about 18 kbar at 1050-l100°C (Wyllie, 1987b ). Crystallization of droplets of carbonatitic melt rather than re-equilibration of solids along a solvus is inferred from the absence of unmixing. For this reason, the meaning of the carbonate compositions can be considered in phase diagrams of Mg-rich carbonatitic melts. According to Irving and Wyllie (1975), a field of liquid and magnesite s.s. develops above 15 kbar between a magnesite s.s.-dolomite s.s. field and a periclase-liquid-vapour field (Fig. 5). The T interval of this field and the amount of solid solution ofdolomite in magnesite increase with rising P; at 25 kbar the magnesite-liquid field exists between 1350° and l550°C. This configuration explains the compositions of group (b) carbonates with falling temperature but not those ofgroup (c). The presence of an extra phase (H 2 0, see below) in the trapped droplets of Mg-carbonatitic melt is thought to be responsible for the compositions of the group (c) carbonates. The reason is that the presence ofH 20 generally shifts the boundary curves to lower temperatures and causes extended solid so-
182
H. HELMERS ET AL.
L
Per
+
L+ V
Fig. 5. Schematic phase diagram of the system dolomite (Dol)-magnesite (Mgs) above 1000 C and IS kbar (after Irvine and Wyllie, 1975). Per = peric1ase, L = liquid, V= vapour. Arrow indicates possible changes due to increasing H2 0-content. 0
lution. The latter has been argued by Lippmann ( 1960) for 10w-P carbonates. It is assumed that the carbonatitic compound persisted until re-equilibration of the garnet peridotite. Then, the untrapped part of the carbonatitic melt dissociated to COrrich fluid and Mg, Ca and Fe which are added to the silicates (Wyllie, 1987b) by a reaction such as: orthopyroxene +magnesite =olivine + CO2. The silicate and oxide inclusions of the garnets comprise: (1) olivine with F0 89 ; (2) orthopyroxene with En86-89 and a slight enrichment in Ca and Mg in comparison to the common orthopyroxene (Table 1); (3) spinel with Sp7sHc2oChrs, less Cr-rich than the coronitic variety, which is Sp76Hc14ChrIO (Table 4); (4 ) rutile; (5) amphibole, with an average composition of magnesiohornblende to tschermakite Ko.D3 N aO.38 Ca 1.62Mg3.83 Feo.34Tio.o8 Cro.Q7Alo.93 [Si6.23 AI!.??] ... , i.e. different from the average pargasite in the garnet coronas, which is: Ko.13Nao.83Ca1.74Mg3.61Feo.49Tio.21CrO.llAlo 68 [Si6.11 All 89]'" (Table 3). These compositional differences are explained by the participation of the CrTi-rich clinopyroxene (Table 1) in the late coronaforming reactions. The earlier growth of spinel, orthopyroxene and hornblende took place without major involvement of clinopyroxene at the overstepping of the garnet +olivine +H 20 stability at about 15 kbar. The remaining fluid phase became relatively enriched in H 20 after incorporation of CO 2 into the carbonatitic melt. With X H20 above 0.2 (Wyllie, 1987b), hornblende became stable because the Na20/K20 ratio of the peridotite was high enough. Below 25 kbar, the position of the peridotite solidus
is dependent on the stability of hornblende, implying an increase in T at constant P in respect to the wet solidus (Wyllie, 1987b). Because spinel was found as a stable phase, the formation and re-equilibration of solid inclusions must have occurred till after the equilibration of the peridotite at 20-16 kbar and 1075°C. Observations on the distribution of the solid phases suggest that recrystallization starting along fluid-enriched, subsequently healed cracks, is probable. Inspection of phase diagrams (Wyllie, 1987b) indicates that re-equilibration at a lower P may be a better explanation of the presence of the solid inclusions than the existence of relics of an enigmatic, incipient silicate melt produced by fast decompression. Since the CO2 dissociating from the carbonatitic melt (see above) is again added to the fluid phase ofthe peridotite, and part of the H 20 is incorporated in a hydrous silicate phase, the composition of the fluid will be COrrich at the end.
Major and trace elements In five samples (Table 8), major elements and Cr were analyzed on fused glass beads using standard XRF techniques. Ni, Cu, Zn, Ga, Rb, Sr, Y, Zr, Nb, Ba and Pb were analyzed on pressed pellets by XRF using the method of Vie Ie-Sage et al. (1979). Sc, Cs, Co, La, Ce, Nd, Sm, Eu, Tb, Vb, Lu, Hf, Ta, Th and U were analysed by standard INAA-techniques using USGS MAG-1 as standard (Potts et al., 1981). The composition of the three granulites indicates a common igneous origin, although palimpsest structures were not observed. Clinopyroxene-garnet granulite represents original basalt, felsic plagioclase-clinopyroxene granulite represents dacite to andesite, and quartz-mesoperthite-garnet granulite appears to indicate a peraluminous rhyolite. The non-analysed biotite-rich variety (16-274) was probably a metapelite. In the trace-element spidergram (Fig. 6), the low chondrite-normalized values ofNb, Ta, Zr, Hf, Yand Ti are conspicuous. When normalized to MORB (Pearce, 1983), the patterns agree well with an island-arc origin of the parent rocks of the granulites. The normative major element composition of the garnet peridotite (Fig. 7) is identical to that of an oceanic peridotite and not to the average garnet lherzolite in the Mg-enriched lithospheric mantle beneath continents (Jordan, 1978; Pollack, 1986). The spidergram (Fig. 6) of
183
GARNET PERIDOTITE AND ASSOCIATED HIGH-GRADE ROCKS FROM SULAWESI TABLE 8
Representative XRFS and INAA analyses of garnet peridotites, clinopyroxene-plagIOclase-garnet granulites and a quartz-kyanite-plagioclase-garnet granulite. Trace elements in ppm (d.l. = detection limit) Sample
16-266
16-337
15-322
16-338
SI02 Ti0 2 AI 20 3 Fe203" MnO MgO CaO Na20 K 20 P 2O,
42.54 0.15 4.06 8.84 0.14 36.87 2.92 0.23 0.09
44.85 0.14 3.44 9.11 0.14 39.11 3.04 0.32 0.09
O.oJ
O.oJ
50.43 0.57 16.16 6.42 0.11 9.92 13. 76 2.55 0.13 0.05
57.49 1.16 19.17 7.09 0.18 2.40 5.95 5.13 1.20 0.41
77.33 0.10 12.30 0.75 0.03 0.09 0.51 3.68 4.35 0.01
Sum
95.85+
100.25
100.11
100.18
99.15
6.96
38.02 0.62 25.69 3\.07 2.50
Peridolite- or CIPW norm Gnt 14.51 13.37 Di"" Hy 16.96 01 55.85 II 0.26
1l.81 14.02 14.61 59.02 0.24
Sum mg
99.70 0.89
Q C Or Ab An Hy 01 II Mt* Ap
0.78 21.54 3\.83 28.62 3.52 9.15 \.09 2.78 0.12
7.12 43.39 25.66 0.97 8.54 2.20 4.11 0.90
0.20 0.44 0.02
Sum
99.43
99.85
99.02
Di 100.95 0.89
16-269
"Total Fe as Fe203; ""incl. jd perc.; *FeO /Fe203 ace. Le Maitre;
0.46
16-266
16-337
15-322
16-338
16-269
Sc Cr Co NI Cu Zn Ga Rb Sr Y Zr Nb Cs
16.5 2662 116 1800 18.5 61 2.5 7 11.5 6 5.5 d.l. 0.7
14 2410 128 1960 26.5 49 d.l. 5.5 12.5 6.3 10.5 d.l. 0.8
43 301 41 120 8 35 10 13 115 17.5 40.5 d.l. 0.8
21 7 13.3 5 8.5 83 22 36.5 495 45 451 4 d.l.
3 5 8.5 3.8 4 7.5 I\.9 175 18 6 45.4 d.l. 3.3
Ba La Ce Nd Sm Eu Tb Yb Lu Hf Ta Pb Th U
68 1.4 4.2 d.l. 0.3 0.2 d.l. 0.6 0.1 0.3 d.l. 4.6 d.l. d.l.
99 1.2 3.4 d.l. 0.3 0.2 d.l. 0.4 0.1 d.l. d.l. 4.9 d.l. d.l.
18 6.3 14.8 8.8 2.3 0.8 0.3 1.1 0.2 1.3 d.l. 9.5 d.l. d.l.
790 5\.9 105.2 63.5 I\.4 2.9 1.5 4.1 0.7 13.7 0.7 10.1 4.6 3
104 17 25.5 d.l. I 0.2 0.1 0.6 0.1 2.9 0.3 18.5 12 2
+some serpentinisation.
two peridotite samples shows a typical ellfichment of LIL and LREE elements in comparison to normal HFS-element abundances. In Fig. 8, the normalized REE contents of the peridotites are compared to several suites of peridotites (Nixon, 1987; Sen, 1987). The patterns do not fit any of these because, in the present case, LREEN/HREE N is 2 to 2.5 with HREEN above 1. Assuming an independent process of LREE enrichment, the other REE fit with the trend of Alpine peridotites in ophiolite complexes (Morten, 1987) and with that of Aleutian nodules (Swanson et al., 1987) but both have lower normalized LREE values. Harte (1987) distinguished three types of mantle metasomatism caused either by infiltrating fluids or by passing melts. Incorporation of a melt fraction leads to enrichment in AI, Fe and Ti which has not been observed in the present case (the Mg-values
are 89-89.5). There are no changes in the modal composition of the rocks (O'Reilly and Griffin, 1988) and no chemically zoned crystals have been found. This indicates the absence of the two types of diffusion metasomatism distinguished by Harte (1987). The only exception is sample 15-326 where nearly 1 vol. % of phlogopite is present. However, the chemistry fits well with Harte's third type, his "isolated trace-element enrichment". The LIL enrichment must be explained by the presence of H 2 0 and the LREE enrichment by the presence of CO 2 in the percolating fluid phase (Menzies et al., 1985). The solubility ofLREE in carbonatitic melts is well established at low P (Jones and Wyllie, 1986) but not yet at high P. The temporary existence of such a melt is indicated by our data (see above). This suggests that the concentration of LREE may have commenced during the temporary existence of the carbonatitic melt. Subsolidus conditions and the
184
H, HELMERS ET AL. 10
!:
i" .. _4
I
I
".1 \\
t
'\
AP+AX
1",
\
\-\-_!
100
~\-~~~~~~~~--l100
!r, \/\ ,r, \/ \ ,/ \" "
"~
/ !.'"
II
I'
"
\
10
ill.: /
\\
Ii
\\
,! 1-,'.\ VIi\----/1
'M' /\, \ 'I
\
I I
'v/ ''l.,!L
I
\1"
:.::
'"
()
\ / '"
~ \/ 1\ i
l
\
I
.\ "
,'\
V X.-ot-l f
~
~
II
:
!'.\''
\\
\
.\\
~
0,2
, I: ,
0,1
I
" II
\II
l1l ;;
~ ~
'"
10
C")
M
;;
~ ~ ~ ~ 0
II)
10
Pj [€
~ ~ ~
co
£3 N
~co
\ II , 'I : i:1 ~
~
0 N
CsPbRbBaU Th K NbTa LaCe SrNd P SrnZrHfTl
~\ \~ \\\
N
gj
r
La Ce
~
-,---1\ ,
01
05
'--:----j10
!
2,
z
I
~l t t-....l
\
\ \ \
2,
G 1,
,
\, I/ I /. 'I
W
f-
a: o o
'i'
•
I" ,
~
\
5
5
, '1
! \ .... -'1 16-269/ \1 l'
"","",~~~~~~~~~~~~~~----'10
!
----
Y YbScCrNI
01
Fig. 6. Spidergram offelsic granulite 16-269, two mafic granulite samples 16-338 (plagioclase-rich) and 15-322 and two garnet peridotite samples 16-337 and 16-266. The chondritenormalizing values from various sources are indicated.
. 0,5
..),
'"
02
Sm Eu
Yb Lu 01
Fig. 8. Chrondrite-normalized REE pattern of garnet peridotite samples compared to various sources (from Nixon (Editor), Mantle Xenoliths, 1987). AP and AX = Alpine peridotites and Aleutian xenoliths, HSLS=Hawaiian spinel lherzolite suite, GDLT = granular depleted low-temperature - and SFHT = sheared fertile high-temperature garnet lherzolite of South Africa (explanation in the text).
absence of partial melting in the peridotite are implied, since fluids will easily dissolve in any melt at pressures below 20 kbar (Wyllie, 1987a; Ellis and Wyllie, 1980) and be largely removed, thus preventing the formation of the observed enrichments.
Conclusions OL + OPX
20
30
CPX
20
30
GNT
Fig. 7. Normative composition ofgarnet peridotite compared to average continental garnet lherzolite (ACGL) of stable continents and to the oceanic mantle peridotite composition (OMPC) (Jordan, 1978).
The observations and data obtained by the various methods of investigation are discussed at the end of the preceding paragraphs. Only the most important information is repeated here and integrated in the conclusions below: ( 1) The garnet peridotite of Sulawesi occurs in a zone of continental accretion below which fragmented and probably degassing relics of a former subduction zone are present. (2) The geothermobarometry of the garnet peridotite indicates a simple uplift pattern. The concave upper part of the P- Tcurve and the presence of onecrystal armours around the unstable minerals in the surrounding metamorphites indicate fast uplift. (3) The garnet peridotite is not part of a former subduction complex because no increase in P can be demonstrated due to the absence of significant
GARNET PERIDOTITE AND ASSOCIATED HIGH-GRADE ROCKS FROM SULAWESI
garnet zonation. No subduction eclogites are present. (4) Major-element chemistry indicates that the garnet peridotite is an oceanic peridotite which does not belong to an evolved continental keel. This is in agreement with a young age of the overlying continental crust. (5) Chemical variation of the carbonate inclusions in the garnet peridotite suggests crystallization from droplets of carbonatitic melt rather than the presence of a re-equilibrating solid carbonate phase in the peridotite. (6) The solid and fluid inclusions indicate that the garnet peridotite and the granulites contained an evolving COz-HzO-bearing fluid. Differences of f02 and the crystallization of graphite and hornblende describe the exact composition at entrapment. (7) The trace-element chemistry indicates that the garnet peridotite was enriched by HzO-COrbearing infiltrating fluids. Our data are insufficient to propose enrichment of the LIL elements by fluids in the lower-crustal granulites. (8) The evolution of the surrounding metamorphites from intermediate-P to low-P conditions and the occurrence of large amounts of dioritic to granitic melts may indicate the presence of a thermal dome. (9) The scarcity of features pointing to anatexis in the granulites and the obvious absence of LIL-depletion (see however, paragraph 7 above) indicate that the lower crust had hardly been involved in the formation of the Miocene granite and diorite batholiths in the area. A mid-crustal origin is more feasible. (10) The deep-seated Palu-Koro fault allowed the uplift and surface exposure of garnet peridotite at the boundary of a newly created miniplate. The synoptic model is: A garnet peridotite with a fluid phase enriched in COz and HzO was emplaced into the lower crust in the near-absence of mantle-derived magmas. Small quantities of peridotite and granulites were transported to the surface along a fault. Studies of fluid inclusions indicate fluid transfer into the lowercrustal granulites with the possible removal of preexisting HzO to higher crustal levels. During this process, the temperatures rose in the granulites and a thermal dome was formed. The fluid in the rising peridotite may have been derived from the under-
185
lying subducted slab or, as commonly assumed (Wyllie, 1987a), it may have been released from a mantle reservoir below the asthenosphere. The f02 must have been below the QFM-buffer to allow the formation of CH 4 in the fluid phase after a re-equilibration event at 20 to 15 kbar (Green et aI., 1987; Saxena, 1989). Further isotope studies are needed to confirm the source of the fluids. We present this study as a case history of mantle degassing during the rapid rise ofa garnet peridotite to the lower crust. This caused the formation of a mantled gneiss dome in the overlying crustal segment.
Acknowledgements The authors are grateful for facilities to use the electron microprobe, the laser Raman microprobe and heating and freezing equipment, provided by the Free University, Amsterdam and N.W.O. (Netherlands Organization of Pure Scientific Research). The use of the scanning electron microscope at the University of Siena, Italy by THDH and his grant from the ERASMUS organization are also acknowledged. The aid ofE.A.J. Burke in providing Raman analyses is highly appreciated. F.F. Beunk and Th.G. van Meerten are sincerely thanked for providing the XRF and INAA major- and trace-element chemical analyses. J. Werner of the Geological Museum of the University of Amsterdam provided the samples from the Sulawesi collection of Brouwer and thin sections from the Ph.D. collection ofEgeler. His consent is highly appreciated. The manuscript was improved considerably by critical reviews of R. Klemd and N. Pearson and profited from readings by E.A.J. Burke, A.M. van den Kerkhof and H.J. Kisch.
References Anovitz, L.M. and Essene, E.J., 1987. Phase equilibria in the system CaCOr MgCO r FeC0 3 • J. Petrol., 28: 389-414. Aranovich, L.Ya. and Podlesskiy, KK, 1983. The cordierite-gamet-sillimanite-quartz equilibrium: experiments and applications. In: S.K Saxena (Editor), Kinetics and Equilibrium in Mineral Reactions. Springer, Berlin, pp. 173-198. Aranovich, L.Ya., Podlesskiy, KK and Kosyakova, N.A., 1986. Thermodynamics ofgarnet-cordierite-orthopyroxene equilibria in the FeO-MgO-Alz0 3-SiOz system. Dokl. Akad. Nauk. SSSR, 291: 945-950. (Translated from Russian).
186 Barber, A.1. (1981). Structural interpretations of the island of Timor, Eastern Indonesia. In: A.J. Barber and S. Wiryosujono (Editors), The Geology and Tectonics of Eastern Indonesia. Geol. Res. Dev. Centre. Spec. Publ. 2. Bandung, Indonesia, pp. 183-198. Bhattacharya, A., 1986. Some geobarometers involving cordierite in the FeO-Al z0 3-SiO z (± HzO) system: refinements, thermodynamic calibration, and applicability in granulite facies rocks. Contrib. Mineral. Petrol., 94: 387394. Bhattacharya, A. and Raith, M., 1987. An updated calibration of Mg-Fe partitioning between garnet-biotite and orthopyroxene-biotite based on natural assemblages. Fortschr. Mineral., 65: 25. Bhattacharya, A., Mazumdar, A.C. and Sen, S.K., 1988. FeMg mixing in cordierite: Constraints from natural data and implications for cordierite-garnet geothermometry in granulites. Am. Mineral., 73: 338-344. Bohlen, S.R. and Lindsley, D.H., 1987. Thermometry and barometry of igneous and metamorphic rocks. Ann. Rev. Earth Planet. Sci., 15: 397-420. Bohlen, S.R. and Liotta, J.1., 1986. A barometer for garnet amphibolites and garnet granulites. J. Petrol., 27: 10251034. Brouwer, H.A., 1934. Geologische onderzoekingen op het eiland Celebes (with collaborators). Verh. Geol. Mijnbouwk. Genootsch., series X: 39-218. Burke, E.A.J. and Lustenhouwer, W.1., 1987. The application of a multichannel laser Raman microprobe (Microdil-28) to the analysis of fluid inclusions. Chem. Geol., 61: 1117. Byrnes, A.P. and Wyllie, P.1., 1981. Subsolidus and melting relations for the join CaCO r MgC0 3 at 10 kb. Geochim. Cosmochim. Acta, 45: 321-328. De Roever, W.P., 1947. Igneous and Metamorphic rocks in eastern Central Celebes. In: Geological Explorations in the Island of Celebes under leadership ofH.A. Brouwer. North Holland, Amsterdam, pp. 69-173. De Roever, W.P., 1950. Preliminary notes on glaucophanebearing and other crystalline schists from South East Celebes and on the origin of glaucophane-bearing rocks. K. Ned. Acad. Wet. Proc., 53: 1455-1465. De Roever, W.P., 1953. Tectonic conclusions from the distribution ofthe metamorphic facies in the island ofKabaena near Celebes. Pacif. Sci. Congr. 7th, New Zealand 1949, vol. 2, pp. 71-81. De Smet, M.E.M., 1989. A geometrically consistent plate tectonic model for Eastern Indonesia. In: J.E. van Hinte, Tj.e.E. van Weering and A.R. Fortuin (Editors), Geology and Geophysics of the Banda Arc and Adjacent areas. Neth. J. Sea Res., 24: 173-183. Egeler, e.G., 1947. Contribution to the Petrology of the metamorphic rocks of Western Celebes. In: Geological Explorations in the Island of Celebes under leadership of H.A. Brouwer. North Holland, Amsterdam, pp. 175-346. Ellis, D.E. and Wyllie, P.J., 1980. Phase relations and their petrological implications in the system MgO-SiOz-HzOCOz at pressures up to 100 kbar. Am. Mineral., 65: 540556. Ellis, D.J., 1986. Garnet-liquid Fe z+ -Mg equilibria and im-
H. HELMERS ET AL.
plications for the beginning of melting in the crust and subduction zones. Am. J. Sci., 286: 765-791. Ellis, D.J. and Green, D.H., 1979. An experimental study of the effects of Ca upon garnet-clinopyroxene Fe-Mg exchange equilibria. Contrib. Mineral. Petrol., 71: 13-22. Essene, E.J., 1982. Geologic thermometry and barometry. In: P.H. Ribbe (Editor), Characterization of metamorphism through mineral equilibria. Rev. Mineral., 10: 153-206. Eugster, H.P. and Skippen, G.B., 1967. Igneous and metamorphic reactions involving gas equilibria. In: L.S. Abelson (Editor), Researches in Geochemistry. 2. Wiley, Chichester, pp. 492-520. Ferry, J.M. and Spear, F.S., 1978. Experimental calibration of the partitioning of Fe and Mg between biotite and garnet. Contrib. Mineral. Petrol., 66: 113-117. Ganguly, J. and Saxena, S.K., 1984. Mixing properties of aluminosilicate garnets: constraints from natural and experimental data, and applications to geothermobarometry. Am. Mineral., 69: 88-97. Gasparik, T. and Newton, R.e., 1984. The reversed alumina contents of orthopyroxene in equilibrium with spinel and forsterite in the system MgO-AlzOrSiO z. Contrib. Mineral. Petrol., 85: 186-196. Ghent, E.D., 1976. Plagioclase-garnet-AlzSiOs-quartz: a potential geobarometer-geothermometer. Am. Mineral., 61: 710-714. Ghent, E.D. and Stout, M.Z., 1981. Geobarometry and geothermometry of plagioclase-biotite-garnet-muscovite assemblages. Contrib. Mineral. Petrol., 76: 92-97. Goldsmith, lR., 1980. Thermal stability of dolomite at high temperatures and pressures. J. Geophys. Res., 85: 69496954. Goldsmith, J.R. and Heard, H.L., 1961. Subsolidus phase relations in the system CaCO r MgC0 3 • J. Geol., 69: 45-74. Graham, e.M. and Powell, R., 1984. A garnet-hornblende geothermometer: calibration, testing, and application to the Pelona Schist, Southern California. J. Metam. Geol., 2: 13-31. Green, D.H., Falloon, T.J. and Taylor, W.R., 1987. Mantlederived magmas; roles of variable source peridotite and variable C-H-O fluid compositions. In: B.O. Mysen (Editor), Magmatic Processes: Physicochemical Principles. The Geochem. Soc. Spec. Publ., I: 139-153. Hamilton, W., 1979. Tectonics of the Indonesian Region. U.S. Geol. Surv. Prof. Pap., 1078, 345 pp. Harley, S.L., 1984. An experimental study of the partitioning of Fe and Mg between garnet and orthopyroxene. Contrib. Mineral. Petrol., 86: 359-373. Harte, B., 1987. Metasomatic events recorded in mantle xenoliths: an overview. In: P.H. Nixon (Editor), Mantle Xenoliths, Wiley, Chichester, pp. 625-640. Hartono, H.M.S. and Tjokrosapoetro, S., 1984. Preliminary Account and Reconstruction ofIndonesian Terranes. Proc. Indones. Petrol. Assoc. 13th Annu. Conv.: 185-226. Helmers, H., Sopaheluwakan, J., Surya Nila, E. and Tjokrosapoetro, S., 1989. Blueschist evolution of SE Sulawesi, Indonesia. In: J.E. van Hinte, Tj.C.E. van Weering and A.R. Fortuin (Editors), Geology and Geophysics of the Banda Arc and Adjacent areas. Neth. J. Sea Res., 24: 373381.
GARNET PERIDOTITE AND ASSOCIATED HIGH-GRADE ROCKS FROM SULAWESI
Hodges, K.V. and Spear, F.S., 1982. Geothermometry, geobarometry and the AlzSiO s triple pomt at Mt. MOOSIlauke, New Hampshire. Am. Mineral., 67: 1118-1134. Holdaway, M.J. and Lee, S.M., 1977. Fe-Mg cordierite stability in high-grade pelitic rocks based on experimental, theoretical, and natural observations. Contrib. Mineral. Petrol., 63: 175-198. Holloway, J.R., 1977. Fugacities and activity of molecular species in supercritical fluids. In: D.G. Fraser (Editor), Thermodynamics in Geology. Wiley, Chichester, pp. 161181. Indares, A. and Martignole, J., 1985. Biotite-garnet geothermometry in the granulite facies: the influence ofTi and Al in biotite. Am. Mineral., 70: 272-278. Irving, A.J. and Wyllie, P.J., 1975. Subsolidus and melting relationships for calcite, magnesite and the join CaC0 3MgC0 3 to 36 Kb. Geochim. Cosmochim. Acta, 39: 3553. Jones, A.P. and Wyllie, P.J., 1986. Solubility of rare earth elements in carbonatite magmas indicated by the liquidus surface in CaCOrCa(OHh-La(OHh at I kbar pressure. Appl. Geochem., I: 95-102. Jordan, T.N., 1978. Composition and development of the continental tectosphere. Nature, 274: 544-548. KatilL J.A., 1978. Post and present geotectonic position of Sulawesi, Indonesia. Tectonophysics, 45: 389-422. Koziol, A.M. and Newton, R.C., 1988. Redetermination of the anorthite breakdown reaction and improvement of the plagioclase-garnet-AlzSiOs-quartz geobarometer. Am. Mineral., 73: 216-223. Krogh, E.J., 1988. The garnet-clinopyroxene Fe-Mg geothermometer - a reinterpretation of existing experimental data. Contnb. Mineral. Petrol., 99: 44-48. Lee, CoS. and McCabe, R., 1986. The Banda-Celebes-Sula basin: a trapped piece of Cretaceous-Eocene oceanic crust. Nature, 322: 51-54. Linthout, K, Helmers, H., Sopaheluwakan, J. and Surya Nila, E., 1989. Aspects of metamorphism of the Northern Banda Arc. In: J.E. van Hinte, Tj.C.E. van Weering and A.R. Fortuin (Editors), Geology and Geophysics of the Banda Arc and Adjacent areas. Neth. J. Sea Res., 24: 345-356. Lippmann, F. (1960). Versuch zur AufkUirung der Bildungsbedingungen von Kalzlt und Aragomt. Fortschr. Mmeral., 38: 156-161. Lonker, S.W., 1981. The P-T-X relations of the cordieritegarnet-sillimanite-quartz eqUIlibrium. Am. J. Sci., 281: 1056-1090. Menzies. M.A., Kempton, P. and Dungan, M., 1985. Interaction of continental lIthosphere and asthenospheric melts below the Geronimo Volcanic Field, Arizona, USA. J. Petrol., 26: 663-693. Miyashiro, A., 1973. Paired and unpaired metamorphic belts. Tectonophysics, 17: 241-254. Mori, T. and Green, D.H., 1978. Laboratory duplication of phase equilibria observed in natural garnet Iherzolites. J. Geol., 86: 83-97. Morten, L., 1987. Italy: a review of xenolithic occurrences and their comparison with Alpine peridotites. In: P.H. Nixon (Editor), Mantle Xenoliths. Wiley, Chichester, pp. 135148.
187
Newton, R.C. and Haselton, H.T., 1981. Thermodynamics of the garnet-plagioclase-AlzSiOs-quartz geobarometer. In: R.C. Newton, A. Navrotsky and B.J. Wood (Editors), Thermodynamics of minerals and melts. Advances in Physical Geochemistry, Vol. I. Springer, Berlin, pp. 131148. Newton, R.C. and Perkins, D., III, 1982. Thermodynamic calibration of geobarometers based on the assemblages garnet-plagIOclase-orthopyroxene (clinopyroxene)quartz. Am. Mineral., 67: 203-222. Newton, R.c. and Wood, B.J., 1979. Thermodynamics of water in cordlerite and some petrologic consequences of cordierite as a hydrous phase. Contrib. Mineral. Petrol., 68: 391-405. Nishimura, S. and Suparka, S., 1986. Tectonic development of East Indonesia. J. Southeast Asian Earth Sci., I: 45-57. Nixon, P.H., 1987. Kimberlitic xenoliths and their cratonic setting. In: P.H. Nixon (Editor), Mantle Xenoliths. Wiley, Chichester, pp. 215-240. Obata, M. and Morten, L., 1987. Transformation of spinel lherzolite to garnet lherzolite in ultramafic lenses of the Austridic Crystalline Complex, Northern Italy. Journ. Petrol., 28: 599-624. O'Hara, M.J., 1977. Thermal history of excavation of Archaean gneisses from the base of the continental crust. J. Geol. Soc. London, 134: 185-200. O'Reilly, S.Y. and Griffin, W.L., 1988. Mantle metasomatism beneath western Victoria, Australia. I. Metasomatic processes in Cr-diopside Iherzolites. Geochim. Cosmochim. Acta, 52: 433-447. Papike, J.J., Cameron, KL. and Baldwin, K., 1974. Characterization of other than quadrilateral components and estimates of ferric iron from microprobe data. Annu. Meet. Geol. Soc. Am. Abstr. Programs, 6: 1053-1054. Pearce, J.A., 1983. Role of the Subcontinental Lithosphere in Magma Genesis at Active Continental Margins. In: C.J. Hawkesworth and M.J. Norry (Editors), Continental Basalts and Mantle Xenoliths, Shiva Publ. U.K, pp. 230249. Perchuk, L.L., 1986. The course of metamorphism. Int. Geol. Rev., 28: 1377-1400. Perchuk, L.L. and Aranovich, L.Ya., 1986. Improvement of the biotite-garnet geothermometer: correction for the fluorine content of biotIte. Transact. U.S.S.R. Acad. Sci., E.S.S., 277: 130-133. Perchuk, L.L. and Lavrent'eva, LV., 1983. Experimental investigation of exchange equilibria in the system cordierite-garnet-biotite. In: S.K. Saxena (Editor), Kinetics and Equilibrium in Mineral Reactions. Springer, Berlin, pp. 199-239. Perkins, D., III and Chipera, S.J., 1985. Garnet-orthopyroxene-plagioclase-quartz barometry: refinement and application to the English River subprovince and the Minnesota River valley. Contrib. Mineral. Petrol., 89: 69-80. Pollack, M.N., 1986. Cratonization and thermal evolution of the mantle. Earth Planet. Sci. Lett., 80: 175-182. Potts, P.J., Thorpe, O.W. and Watson, J.S., 1981. Determination of the rare earth element abundances in 29 international rock standards by instrumental activation anal-
188 ysis: a critical appraisal of calibration errors. Chern. Geol., 34: 331-352. Poty, B., Leroy, J. and Jachimowicz, L., 1976. A new device for measuring temperatures under the microscope: the Chaixmeca microthermometry apparatus. Bull. Soc. Fr. Mineral. Cristallogr., 99: 182-186. Raheim, A. and Green, D.H., 1974. Experimental determination of the temperature and pressure dependence of the Fe-Mg partition coefficient for coexisting garnet and clinopyroxene. Contrib. Mineral. Petrol., 48: 179-203. Sasajima, S., Nishimura, S., Otofuji, Y., Hirooka, K., van Leeuwen, T. and Heluwat, F., 1981. Paleomagnetic studies combined with fission-track dating on the western arc of Sulawesi, East Indonesia. In: AJ. Barber and S. Wiryosujono (Editors), The Geology and Tectonics of Eastern Indonesia. Geol. Res. Dev. Centre. Spec. Publ. 2. Bandung, Indonesia, pp. 305-312. Saxena, S.K., 1979. Garnet-clinopyroxene geotherrnometer. Contrib. Mineral. Petrol., 70: 229-235. Saxena, S.K., 1989. Oxidation state of the mantle. Geochim. Cosmochim. Acta, 53: 89-95. Sen, G., 1987. Xenoliths associated with the Hawaiian Hot Spot. In: P.H. Nixon (Editor), Mantle Xenoliths, Wiley, Chichester, pp. 359-376. Sen, S.K. and Bhattacharya, A., 1984. An orthopyroxenegarnet thermometer and its application to the Madras charnockites. Contrib. Mineral. Petrol., 88: 64-71. Silver, E.A., Mc Caffrey, R. and Joyodiwiryo, Y., 1981. Gravity results and emplacement geometry of the Sulawesi ultramafic belt. In: A.J. Barber and S. Wiryosujono (Editors), The Geology and Tectonics of Eastern Indonesia. Geol. Res. Dev. Centre. Spec. Publ. 2. Bandung, Indonesia, pp. 313-320. Silver, E.A., Mc Caffrey, R., Joyodiwiryo, Y. and Stevens, S., 1983. Ophiolite emplacement by collision between the Sula platform and the Sulawesi island arc, Indonesia. J. Geoph. Res., 88(Bll): 9419-9435. Sukamto, R., 1975. Geologic map ofUjung Pandang sheet, scale 1.1000 000. Geo!. Surv. Indonesia. Swanson, S.E., Kay, S.M., Brearly, M., Scarfe, C.M., 1987. Arc and back-arc xenoliths in Kurile-Kamchatka and Western Alaska. In: P.H. Nixon (Editor), Mantle Xenoliths. Wiley, Chichester, pp. 303-318. Taylor, W.R. and Green, D.H., 1987. The petrogenetic role of methane: Effect on liquidus phase relations and the solubility mechanism ofreduced C-H volatiles. In: B.O. Mysen (Editor), Magmatic Processes: Physicochemical Principles. Geochem. Soc. Spec. Publ. 1: 121-138. Thompson, A.B., 1976. Mineral reactions in pelitic rocks; II. Calculation of some P-T-X(Fe-Mg) phase relations. Am. J. Sci., 276: 425-454.
H. HELMERS ET AL.
Thompson, A.B. and England, P.C., 1984. Pressure-Temperature-Time paths of regional metamorphism II. Their Inference and Interpretation using Mineral Assemblages in Metamorphic Rocks. J. Petrol., 25: 929-955. Tjia, H.D. and Zakaria, T., 1974. Palu-Koro strike-slip fault zone, Central Sulawesi Indonesia. Sains Malaysiana, 3( 1): 65-86. Touret, J .L.R., 1981. Fluid inclusions in high grade metamorphic rocks. In: L.S. Hollister and M.L. Crawford (Editors), Fluid Inclusions: Applications to Petrology. Mineral. Assoc. Can., 6: 182-208. Trail, D.S., 1974. The general geological survey of block 2, Sulawesi Utara, Indonesia. PT. Tropic Endeavour Indonesia (unpubl. rep.). Van Bemmelen, R.W., 1949. The Geology ofIndonesia. Gov. Printing Office, The Hague, 732 pp. Van Leeuwen, Th.M., 1981. The Geology of south-west Sulawesi with special reference to the Biru area. In: AJ. Barber and S. Wiryosujono (Editors), The Geology and Tectonics of Eastern Indonesia, Geol. Res. Dev. Centre. Spec. Publ. 2. Bandung, Indonesia, pp. 277-304. Vie Ie-Sage, R., Quisefit, J.P., Dejean de la Batie, R. and Faucherre, J., 1979. Utilisation du rayonnement primaire diffuse par l'echantillon pour une determination rapide et precise des elements trace dans les roches. X-ray Spectrom., 8: 121-128. Wells, P.R.A., 1979. Chemical and thermal evolution of Archaean sialic crust, southern West Greenland. J. Petrol., 20: 187-226. Wells, P.R.A. and Richardson, S.W., 1979. Thermal evolution of metamorphic rocks in the Central Highlands of Scotland. In: A.L. Harris, C.H. Holland and B.E. Leake (Editors), The Caledonides of the British Isles - reviewed. Scott. Acad. press, Edinburgh, Spec. Pap. Geol. Soc. London, 8: 339-344. Wood, B.J., 1974. The solubility ofalumina in orthopyroxene coexisting with garnet. Contrib. Mineral. Petrol., 46: 115. Wood, B.J. and Banno, S., 1973. Garnet-orthopyroxene and orthopyroxene-clinopyroxene relationships in simple and complex systems. Contrib. Mineral. Petrol., 42: 109-124. Wyllie, P.J., 1987a. Transfer of subcratonic carbon into Kimberlites and rare earth carbonatites. In: B.O. Mysen (Editor), Magmatic Processes: Physicochemical principles. Geochem. Soc. Spec. Publ., 1: 107-120. Wyllie, P.J., 1987b. Metasomatism and fluid generation in mantle xenoliths. In: P.H. Nixon (Editor), Mantle Xenoliths. Wiley, Chichester, pp. 609-622.