RETRACTED: Equilibrium vs. disequilibrium melting relations in the Higher Himalayan Crystalline (HHC) pelitic migmatites, Sikkim Himalaya

GEOSCIENCE FRONTIERS 2(4) (2011) 539e549

available at www.sciencedirect.com

China University of Geosciences (Beijing)

GEOSCIENCE FRONTIERS journal homepage: www.elsevier.com/locate/gsf

ORIGINAL ARTICLE

Bhaskar Kundu*

TE D

Equilibrium vs. disequilibrium melting relations in the Higher Himalayan Crystalline (HHC) pelitic migmatites, Sikkim Himalaya

National Geophysical Research Institute, Council of Scientific and Industrial Research, Uppal Road, Hyderabad 500007, India

Abstract Higher Himalayan Crystalline (HHC) complex of the Sikkim Himalaya predominantly consists of high-grade pelitic migmatites. In this study, reaction textures, mineral/bulk rare earth elements (REE), trace element partition coefficients and trace element zoning profiles in garnet are used to demonstrate a complex petrogenetic process during crustal anatexis. With the help of equilibrium REE and trace element partitioning model, it is shown that strong enrichment of Effective Bulk Composition (EBC) is responsible for the zoning in garnet in these rocks. The data strongly support disequilibrium element partitioning and suggest that the anatectic melts associated with mafic selvedges are likely produced by disequilibrium melting because of fast melt segregation process. ª 2011, China University of Geosciences (Beijing) and Peking University. Production and hosting by Elsevier B.V. All rights reserved.

R

ET

Pelitic migmatites; Crustal anatexis; Mafic selvedge; Disequilibrium melting

R

KEYWORDS

AC

Received 8 March 2011; accepted 14 June 2011 Available online 13 September 2011

* Tel.: þ91 9490077905. E-mail address: [email protected]. 1674-9871 ª 2011, China University of Geosciences (Beijing) and Peking University. Production and hosting by Elsevier B.V. All rights reserved. Peer-review under responsibility of China University of Geosciences (Beijing). doi:10.1016/j.gsf.2011.06.005

Production and hosting by Elsevier

1. Introduction Migmatites are of great interest to petrologists because these rocks are considered to represent the source regions for granite melts. Careful study of migmatites can potentially yield important information about melt-forming reactions, melt-extraction processes, and scales of chemical equilibrium in granite source regions. Residueemelt equilibrium is a fundamental assumption in trace elements models for granite petrogenesis (Hanson, 1978). Previous studies of peraluminous migmatites have demonstrated that high field-strength elements (HFSE), such as Zr, Th, trivalent rare earth elements (REE), may not reach equilibrium concentrations in partial melts because their concentrations are controlled by accessory minerals that may be refractory or remain as inclusions in residual major minerals during melting (Sawyer, 1987; Watt and Harley, 1993; Johannes et al., 1995; Nabelek and Glascock, 1995; Bea, 1996; Watt and Burns, 1996). Studies on

B. Kundu / Geoscience Frontiers 2(4) (2011) 539e549 eliminated from the assemblages of pelitic rocks up to its northern contact with the Tethyan sedimentary sequence (Fig. 1). Reaction textures, mineral/bulk REE and trace element partitioning in migmatite and trace element zoning profiles in a garnet crystal are used to demonstrate a complex petrogenetic process during crustal anatexis (Hickmott et al., 1987; Hickmott and Shimizu, 1990). In the Sikkim region, the lithounits display a horse-shoe shaped regional fold pattern (Fig. 1), reflecting a culmination structure often referred to in the literature as the “Tista Dome” (Neogi et al., 1998). A group of low-grade metapelites, psammites and interbanded metabasic rocks, belonging to the Daling Group of possibly Proterozoic age, occur in the central part and constitute the Lesser Himalayan belt. Medium to high-grade crystallines (pelitic schists, pelitic migmatites, quartzites, calcsilicate rocks, conformable bands of metabasites and granites) constituting the Higher Himalayan Crystallines (Cenozoic) occupy outer part of the dome. A prominent ductile-brittle shear zone, the Main Central thrust (MCT), separates the two belts. The MCT is defined here as the southernmost shear zone of a number of northward dipping ductile shear zones within the HHC. Gondwana (CarboniferousePermian) and molasse-type Siwalik (MioceneePliocene) sedimentary rocks of Sub-Himalayan Zone occur in the southern part of the region. In the far north, a thick pile of Cambrian to Eocene fossiliferous sediments of the Tethyan Zone overlie the HHC on the hanging wall side of a series of north-dipping normal faults constituting the South Tibetan Detachment System (STDS). The HHC predominantly consists of high-grade pelitic migmatites with subordinate calc-silicate rocks, metabasites and granites. HHC pelitic migmatites are pervasively heterogeneous on a macroscopic scale, having a dark schistose selvedges (i.e., melanosome) that is intimately associated with light colored, coarse-grained, cmscale layers, veins, or pods of poorly schistose material (i.e., leucosome). The pelitic migmatites are stromatic, with layer-parallel granitic leucosomes and biotite-rich melanosomes. Patchy leucosomes and discordant veins are also present. Banded, finely foliated, and augen gneisses show transitions from stretched leucosomes to composite crystal augens with porphyroblast of K-feldspar. The augen gneisses display pervasive myolintic microfabric suggesting that augen development reflects strain heterogeneities.

R

AC

HFSE distribution have led to inferences of rapid melt extraction or rapid cooling of granite source regions (e.g., Sawyer, 1987; Watt and Burns, 1996). In contrast to HFSE, the behavior of Eu (þII), Rb, Cs, Sr, and Ba in migmatites is different in many cases. Although in some leucosomes, Sr and Ba occur in relatively low concentrations and REE patterns have negative Eu anomalies, in many other leucosomes these elements are highly enriched in comparison to spatially related leucosomes (e.g., Brown and D’Lemos, 1991; Power, 1993). Such enrichments in the divalent cations have previously been attributed to processes such as fractional crystallization of feldspar-containing layers (Cunery and Barbey, 1982; Sawyer, 1987; Johannes et al., 1995). However, Fourcade et al. (1992) argued that trace element distributions between selvedges and leucosomes in St. Malo, France, the migmatites approach mineralemineral equilibrium rather than residueemelt equilibrium. Using published diffusion rates of trace elements in relevant minerals, Fourcade et al. (1992) inferred the cooling rate of the St. Malo migmatite terrene. In contrast, Bea et al. (1994) assumed that the distribution of alkali, alkali earth, transition metals, and some non-high field-strength elements between selvedge minerals and leucosomes in Pena Negra, Spain, migmatites represent mineralemelt equilibrium. One consequence of this assumption is that some of their inferred mineral/melt distribution coefficients are drastically different from those predicted by experimental studies and observed in rhyolite system (Mahood and Hildreth, 1983; Nash and Crecraft, 1985; Icenhower and London, 1995). This raises the question of whether the distribution coefficients of Bea et al. (1994) reflect frozen residueemelt equilibrium rather than mineralemelt equilibrium. This study aims to evaluate equilibrium vs. disequilibrium melting relation in the Higher Himalayan Crystalline (HHC) pelitic migmatites of Sikkim Himalaya by using the distribution of trace elements controlled by rock-forming minerals in selvedges and leucosomes of migmatites as an analog for equilibrium between residue and melt in granite source regions.

TE D

540

ET

2. Regional geologic setting

R

With the advent of modern concept of plate tectonics, the “Himalaya-Tibetan orogenic system” is considered as a classic continental collision zone and has attracted the attention of geoscientists world-wide in the context of the study of tectonothermal evolution of continent collision zones (e.g., Aitchison et al., 2011 and references therein). Himalayan metamorphism has, therefore, largely been considered within the framework of the collision of the Indian and Eurasian plates at some period during the Eocene (w50 Ma), resulting in the subduction of the Indian plate below Tibet and attendant intense crustal shortening and deformation, taken up by crustal stacking along a system of intracontinental thrusts and internal deformation of Indian plate (Hodges, 2000). An enigmatic feature of Himalayan metamorphism is the presence of a sequence of progressively higher grade rocks occurring at higher structural levels, first reported by Gansser (1964) from Darjeeling Sikkim region, which is referred as inverted metamorphism and the sequence as inverted metamorphic sequence (IMS). This study focuses on the geochemical aspects of the HHC pelitic migmatite in Sikkim, eastern Himalaya. The study region is the upper part of Higher Himalayan slab in Sikkim (ChungthangeLacheneYumthang), beginning at the structural levels from which primary muscovite has been

3. Petrology and geochemistry 3.1. Petrography The rocks of the Higher Himalayan Crystallines around Chungthang in north Sikkim comprise predominantly two types of gneissic rocks (Table 1, and Table 2). The first one is the pelitic gneiss comprising biotite, K-feldspar, plagioclase, sillimanite, cordierite, garnet, ilmenite, quartz, and spinel as the dominant mineralogical constituents. The second type is the quartzofeldspathic gneiss that does not contain sillimanite, cordierite and spinel. The following possible mineral reactions can be inferred from the studied rock samples of HHC pelitic migmatites: (1) Absence of primary muscovite in the pelitic assemblages suggests metamorphic conditions beyond the vapor absent incongruent melting reaction and the leucosomes may represent the liquid phase of the melting reaction (Nabelek, 1999):

541

R

AC

TE D

B. Kundu / Geoscience Frontiers 2(4) (2011) 539e549

ET

Figure 1 Schematic geological map of the Sikkim Himalayas (Neogi et al., 1998), showing the studied area (marked by black polygon) and the location of the mineral (red square) and whole-rock samples (black circle). HHC consists of migmatitic paragneisses (dark yellow) and minor bands of calcareous rocks (purple). Inset thin section of HHC pelitic migmatite (sample No. 7A/93/1) shows segregation of leucosomes from selvedge portion under cross polarized light (XPL). G: garnet; Bt: biotite; MCT: Main Central Thrust; STDS: South Tibetan Detachment System.

Table 1 Mineral assemblages (south to north oriented, Fig. 1) in the studied pelitic migmatites. Data type

Mineral assemblage

15-95-M L1 M-1-1 M-3-4 M-A-1 M-5-8 26/93 7-93-1 7A-93-6 7A-93-9 7A-93-12

EPMA EPMA EPMA EPMA EPMA EPMA EPMA EPMA EPMA EPMA EPMA

G þ Bt þ Pl þ Apt þ Opq Bt þ Pl þ Apt þ Mus Pl þ Bt þ Mus þ Kfs þ Zr þ Apt Pl þ Kfs þ G þ Pt Pl þ Apt þ Zr þ Bt G þ Bt þ Apt þ Pt þ Kfs Bt þ Apt þ Zr þ Pt þ G þ Mon þ Pl Bt þ Pl þ Mus þ Kfs G þ Pl þ Apt þ Anti-Pt þ Bt G þ Bt þ Apt þ Kfs þ Mon Pl þ Apt þ Zr þ Bt þ Kfs

R

Sample number

G: garnet; Bt: biotite; Pl: plagioclase; Apt: apatite; Mus: muscovite; Kfs: K-feldspar; Mon: monazite; Zr: zircon; Pt: perthite; Anti-Pt: antiperthite; Opq: opaques.

muscovite þ plagioclase þ quartzZbiotite þ sillimanite þ K  feldspar þ liquid (2) The melting reaction that can be constrained from textural criteria, e.g., garnet mantling plagioclase against biotite, involves biotite. The possible biotite dehydration melting reaction (Nabelek, 1999) is: biotite þ sillimanite þ plagioclase þ quartzZgarnet þ K  feldspar þ liquid In case of both the above mineral reactions, the role of K-feldspar remains uncertain. The phase present in the rock could have been produced through crystallization of the granitic liquid as well. (3) Formation of spinel and quartz along the grain boundaries of porphyroblastic garnet in the presence of sillimanite suggests the following reaction (Nabelek and Glascock, 1995; Nabelek, 1999):

99.04 97.08 99.76 100.73 99.29 99.42 100.78 99.67 100.22 99.97 100.20 100.06 100.74 100.30 98.90 100.89 99.66

garnet þ sillimaniteZspinel þ quartz

0.13 e 0.11 0.08 0.04 0.55 0.16 0.30 0.29 e 0.07 0.07 e 0.15 0.08 0.08 0.10

This reaction moves to the right in response to decompression along the retrograde trajectory of metamorphic evolution.

0.34 e 0.74 0.36 2.42 2.01 1.40 0.57 0.97 e 0.30 0.62 e 0.40 0.54 0.52 0.78

(4) Formation of cordierite rims on garnet can be explained by the following reaction (Nabelek and Glascock, 1995):

0.00 e 0.30 0.05 0.01 0.12 0.03 0.19 0.32 e 0.00 0.00 e 0.02 0.09 0.00 0.00

garnet þ sillimanite þ quartzZcordierite

0.34 0.88 0.44 0.31 2.41 1.89 1.37 0.38 0.65 0.98 0.30 0.62 0.56 0.38 0.45 0.52 0.78

This reaction also moves to the right due to decompression, with or without hydration/carbonation. The latter cannot be evaluated in the absence of data on possible presence of fluid in cordierite.

3.14 2.03 3.80 0.80 0.01 0.56 4.19 2.51 1.22 1.76 3.25 2.83 5.71 2.23 2.93 3.45 3.15

0.35 3.05 3.06 2.19 6.23 4.36 0.89 0.67 3.14 4.05 3.35 3.00 1.27 4.38 5.79 6.14 3.60

TE D

0.02 0.13 0.02 0.00 0.15 0.10 0.13 0 0.02 0.12 0.01 0.11 0.04 0.06 0.08 0.02 0.06

(5) Presence of intergrowths of biotite-quartz and of biotitesillimanite on the marginal part or along the fractures of garnet indicates meltesolid interaction possibly during retrogression according to the reaction (Nabelek and Glascock, 1995; Nabelek, 1999):

1.34 1.74 2.88 0.61 0.27 0.29 2.03 2.52 0.67 1.82 2.83 2.73 5.24 1.64 1.28 1.45 1.68

garnet þ meltZbiotite þ sillimanite þ quartz

0.00 0.03 0.01 0 0.08 0.03 0.01 0 0.01 0.15 0.02 0.07 0.01 0.00 0.02 0.01 0.02

0.18 1.44 0.33 0.09 7.27 2.14 0.33 0.15 0.43 3.16 0.21 1.03 0.11 0.14 0.25 0.19 0.70

It is likely that this reaction occurred in response to cooling subsequent to decompression. Therefore, petrographic study shows that the rocks evolved through prograde dehydration melting reactions, followed by decompression and cooling during retrogression.

0.04 0.48 0.07 0 0.58 0.72 0.12 0.04 0.08 0.86 0.16 0.60 0.04 0.04 0.14 0.09 0.40

R

8.82 13.49 15.14 5.12 14.13 20.27 14.68 11.86 7.09 15.84 14.65 13.68 21.15 12.40 14.04 15.56 13.06

0.31 2.14 0.52 0.09 6.53 6.08 0.89 0.43 0.53 7.40 1.16 4.15 0.19 0.16 1.07 0.44 2.65

0.00 e 0.06 0.01 0.14 0.69 0.06 0.10 0.07 e 0.20 0.59 e 0.03 0.23 0.08 0.48

AC

0.30 e 0.52 0.07 5.75 4.85 0.75 0.30 0.54 e 0.86 3.20 e 0.17 0.76 0.32 1.95

3.2. Mineral chemistry

Note: all of oxides are measured in % unit.

XRF XRF XRF XRF XRF XRF XRF, XRF, XRF, XRF, XRF, XRF, XRF, XRF, XRF, XRF, XRF, L1/LSM M-4/1 M-4/5/LSM M4/6/LSM M4/2/SEL 29/99 29/99/LSM1 29/99/LSM2 M-A/1/LSM M-A/1 26/93/LSM 26/93 7A/93/1 7A/93/18/LSM 29/93/1/LSM 29/93/2/LSM 29/93/2

INAA INAA INAA INAA INAA INAA INAA INAA INAA INAA INAA

Data type Sample No.

Table 2

R

Leucosome Gneiss þ selvedge þ leucosome Leucosome Leucosome (pool type) Selvedge (restite) in melt zone Sillimanite grade metapelite Leucosome Leucosome Leucosome Gneissic part Leucosome Gneiss, mus-out Leucosome Leucosome Leucosome Leucosome High-grade gneiss

Sample description

ET

84.34 71.67 73.03 91.4 62.17 62.85 75.93 80.65 85.68 63.83 74.08 71.25 66.42 78.57 72.60 72.83 73.46

TiO2 Al2O3 Fe2O3(T) Fe2O3 FeO MnO MgO CaO Na2O K2O P2O5 H2Oþ H2O- H2O(T) CO2 Total

B. Kundu / Geoscience Frontiers 2(4) (2011) 539e549

SiO2

Whole-rock samples (south to north oriented, Fig. 1) of pelitic migmatites in different structural levels.

542

The biotite is rich in first row transition elements (FRTE) but has extremely low concentrations of Y, Pb and REE (32.77 ppm). Chondrite-normalized REE patterns (Fig. 2) show almost flat profiles (LREE/HREE Z 1.72e2.17) with a very small negative Eu anomaly (Eu/Eu* Z 0.48e0.53). The garnet grains are strongly zoned, the cores of the grains have high concentration of REE (638e826 ppm) and Y (1783 ppm) in contrast to the rims. The garnet is rich in Y (954.3e1783 ppm), Ti (76.53e231.3 ppm) and Sc (148.04e265.4 ppm), but has extremely low concentrations of Pb (0.17e1.02 ppm), Co (15.82e22.97 ppm) and V (22.47e37.94 ppm). Chondrite-normalized REE patterns (Figs. 2 and 4) show an almost flat, irregular profile from LREE (LaN/SmN Z 0.26e0.89), a strong negative Eu anomaly (Eu/Eu* Z 0.02e0.32) and a strong, progressive increase from Gd to Lu (GdN/LuN Z 0.09e0.03). South to north oriented garnet samples show overall decreasing Eu anomaly and increasing Cr concentration. The plagioclase crystals are strongly zoned and have low concentration of REE (15.52e27.27 ppm), Ti (17.0e69.64 ppm), Y (1.077e2.16 ppm), Pb (19.72e28.59 ppm), Cr (12.28e54.07 ppm). Chondrite-normalized REE patterns (Fig. 2) show strong LREE/HREE fractionation (LaN/LuN Z 18.58e36.27) and a strong positive Eu anomaly (Eu/Eu* Z 7.7e29.54). Core part is less enriched in REE compare to rim part. The plagioclase shows an increase in progressive degree of fractionation of LREE/HREE, Eu anomaly becomes stronger with a decreasing concentration of Cr and Pb. Zircon is rich in REE (1966.24e4434.4 ppm), Y (1235.50e4902 ppm), Ti (493.91e432.9 ppm), V (100.24 ppm), Sc (153.5e470.75 ppm) and Th (282.45e702.8 ppm). Chondritenormalized patterns (Fig. 2) show an almost flat, irregular profile for LREE (LaN/SmN Z 1.35), a strong negative Eu anomaly (Eu/ Eu* Z 0.03e0.12) and a strong, progressive increase from Gd to Lu. Thus zircon is enriched in HREE compare to LREE (LaN/ LuN Z 0.0000755834e0.000177028) of the zircons a marked increase in Y, Th and Cr along the south-north traverse. Monazite is, by far, the richest mineral in REE (1309016, 221339.68,

543

TE D

B. Kundu / Geoscience Frontiers 2(4) (2011) 539e549

AC

Figure 2 Chondrite-normalized REE patterns of apatite, plagioclase, monazite, biotite, zircon and garnet crystals. Position of the Eu anomaly is indicated by red arrow. Concentration of elements is normalized with respect to chondrite values (Taylor and McLennan, 1985).

ET

R

2351851.76 ppm respectively for south to north sample traverse), Y (6844e32370 ppm), Th (32940e230800 ppm) and Cr (71.62e307.7 ppm). In contrast to biotite, monazite contains much higher concentration of Pb (128.2e441.9 ppm), Sc (15.91e241.7 ppm) and Cr (71.62e307.7 ppm). Chondrite-normalized REE patterns (Fig. 2) show heavy LREE/HREE fractionation (LaN/LuN Z 98.58e6391.27) together with a small negative Eu anomaly (Eu/Eu* Z 0.12e0.37). Another notable point is that all the REE patterns show almost flat profile for LREE (LaN/ SmN Z 2.52e4.36) but heavy fractionation for HREE (GdN/ LuN Z 20.36e615.09). Zr and Cr concentrations in the samples increase toward the north. Apatite also shows higher Y (1071e2977 ppm) and REE (2181.9e5942.67 ppm) content than any other major mineral. Chondrite-normalized REE patterns (Fig. 2) are almost flat (LaN/LuN Z 1.09e2.83), with a small negative Eu anomaly (Eu/Eu* Z 0.12e0.41).

3.3. Whole-rock chemistry

R

The selvedges have relatively lower SiO2, Na2O but higher TiO2, FeO, Al2O3, MgO, MnO, ASI than leucosome whereas K2O contents and (Fe/Fe þ Mg) Z Fe# values are variable (Fe# Z 0.66e0.90). Selvedges are enriched in those elements contained either in biotite (Li, Rb, Cs, Sc, V, Ni, Nb, Ta) or in monazite (LREE, Th, U, Y) and zircon (Zr, Hf). Selvedge samples display chondrite-normalized patterns (Fig. 3) with moderate LREE/HREE fractionation (LaN/ LuN Z 7e12), i.e., relatively enriched in LREE and a small negative Eu anomaly (Eu* Z 0.43e0.78). HREE have a flat profile from Dy to Lu. However, leucosomes have a composition typical of K-rich peraluminous leucogranite. There is a large systematic change in the composition of leucosomes as K2O deceases, due to the simultaneous increase of TiO2, Al2O3, FeO, MgO, CaO and Na2O. This variability in leucosome composition is even greater in the trace elements and gives important clues about the behavior of minerals during melting.

The good positive correlation of K2O with Ba and Rb (Fig. 4) reveals that the main source of potassium was the incorporation of K-feldspar into the melt. Interestingly, Pb is almost constant with changes in K2O (Fig. 4). There is also remarkable positive correlation of TiO2 with Y (Fig. 4) and Sr with Na2O (Fig. 4). This suggests that incorporation of REE-, Y-bearing minerals into the melt is controlled by the physical incorporation of biotite crystals. In different migmatite parts, correlations are good, except for a few samples. However, the slopes of the correlations are variable as the trace elements reside in different minerals. For example, the concentration of Ba in most leucosomes is related to the amount of feldspar, whereas in the melanosomes it is related to the amount of biotite. The leucosomes samples display chondrite-normalized patterns (Fig. 3) with high LREE/HREE fractionation (LaN/LuN Z 16.0e116.7) and a high positive Eu anomaly (Eu/Eu* Z 1.19e6.76). As the SREE increases, the Eu anomaly decreases, whereas LaN/LuN ratio increases, although variations are not well defined as for the Eu anomaly. This picture suggests that REE contents and patterns in leucosomes are essentially controlled by the proportions of K-feldspar, plagioclase and monazite incorporated into the segregate, the behavior of this last mineral being controlled by that of biotite. Since biotite in leucosome is always restitic, it seems that the efficiency of restite-melt separation is the key factor in controlling leucosome chemistry. From the compatibility diagram (Fig. 3), it is clear that there is an incomplete separation of trace and REE elements between selvedges and leucosomes. In general incompatible elements are more concentrated in leucosomes compared to selvedges. Particularly three selvedge samples (M-A/1, 29/93/2, 29/99) absolutely match the leucosomes compatibility pattern. Apart from that another important point to be noted is that the leucosomes show positive anomalies for Th, U, Pb, Yb and negative anomalies for Ba, Nb, Ce, Sr, Hf, Lu. However the two completely separated selvedge samples (M4/2/ SEL, M4/1) show considerable variations between them. Sample M4/ 1 shows positive anomalies for Th, U, Ce, Sm and negative anomalies

B. Kundu / Geoscience Frontiers 2(4) (2011) 539e549

ET

R

AC

TE D

544

R

Figure 3 Chondrite-normalized REE patterns of bulk-rock samples (selvedge and leucosome). Eu anomaly is indicated by red arrow. Incomplete segregation of trace and REE between selvedges and leucosomes is noticed in compatibility profile. In compatibility profile solid and dashed lines indicate leucosomes and selvedges samples respectively.

for Nb, La, Pb, Sr, Hf; but sample M4/2/SEL shows positive anomalies for Nb, Ce, Nd, Hf and negative anomalies for Ba, U, Pb, Sr, Eu.

4. Trace element zoning in garnet crystal

Chemical zoning in garnet porphyroblasts (Table 3) is frequently observed in the HHC pelitic gneiss. Chemical zoning (Chernoff and Carison, 1999) may result from change in p, T, or fluid condition. This work focuses on trace element zoning in garnet because of two fundamental reasons: (1) trace elements are more sensitive to monitors of disequilibrium element partitioning during geological process than major elements; and (2) at sufficient high T, major element growth zoning may be significantly modified by intercrystalline

diffusion, but on the other hand trace element zoning is less susceptible to diffusional modification (Hickmott and Spear, 1992). Trace elements in garnet (Hickmott et al., 1987) are divided into two groups on the basis of site occupancies: cubic-site elements (Sc, Co, Y and REE) and octahedral-site elements (Ti, V, Cr, Zr). The cubic-site cations all show fairly smooth zoning profile (Fig. 5). Sc shows roughly bell-shaped zoning profile, whereas for Co, profile is almost flat with slightly low concentration in core. Y and REE show periodic humps, interestingly Yb, Y, Ho, Lu, Er and Eu profiles are all symmetric with higher concentration in core and adjacent rim parts. On the other hand octahedral-site trace elements exhibit more complicated behavior. All profiles are irregular and asymmetric. From this observation

Figure 4

545

TE D

B. Kundu / Geoscience Frontiers 2(4) (2011) 539e549

Scatter plots of some representative trace elements against K2O, TiO2 and Na2O of leucosome samples.

concentrations or ion intensity ratios. An element with a bulk distribution coefficient identical to that of Sc, yields a horizontal array of data on Fig. 6, an infinite K produces an array that approaches upward vertical, and a small K yields an inclined array approaching 45 . Any array on Fig. 6 oriented between downward vertical and inclined at 45 cannot be realized by simple Rayleigh fractionation and requires either a change in the ratio K(Sc)/K(El), disequilibrium partitioning, or open-system behavior (this region will be referred to non-Rayleigh region (NRR)). Unfortunately, none of the trace elements in the studied sample can be easily explained by a single-stage Rayleigh fractionation models, with possible exception of Co (Fig. 6). More complex mechanisms such as multiple-stage Rayleigh growth, open-system behavior or disequilibrium partitioning are required. An abrupt change in distribution coefficients for an element leads to a kink in Log(Sc/El) vs. Log(Sc) plot of Fig. 6, rather than a gradual change that leads to a smooth curve. In the expression (1) we assume constant K, but multiple episodes of garnet growth with differing K can be treated as multiple distinct Rayleigh events.

AC

some fundamental questions can be raised: (1) Can the humps in zoning profile be explained by equilibrium crystal growth? (2) If disequilibrium is required, what is the nature of the process controlling it? (3) If the process can be identified, in what geologic environments is the process important? In order to address these questions, it is important to establish two partitioning models relevant to metamorphic minor element zoning.

R

4.1. Equilibrium partitioning models (Rayleigh fractionation)

R

ET

Elemental concentration ratios between garnet and another mineral (i.e., distribution coefficients) are >1 for compatible elements (Sc, Y, Mn and HREE), w1 for intermediate elements (Co and Zr), and <1 for incompatible elements (Cr and V). The Sc profile in sample 7A/93/9 has been modeled successfully (Fig. 5) using equilibrium partitioning models such as a single-stage Rayleigh fractionation model (Hollister, 1966) simply because Sc zoning profile is smooth bell-shaped and the most importantly, it is a compatible element in garnet. This model has three constraints: (1) mass balance within a closed system, (2) constant partitioning during crystal growth, and (3) removal of the fractionating element from the reacting system by mineral growth. The second requirement generally best applies to the trace elements because distribution coefficients for major elements are often functions of composition. The simple Rayleigh fractionation equation for Sc is: LogðScðgrtÞ=ElðgrtÞÞZ½ðKðScÞ  KðElÞÞ=ðKðScÞ  1Þ ðLogðScðgrtÞÞÞ þ Z

ð1Þ

where, Sc(grt) is the concentration of Sc in garnet crystals, K(Sc) is the equilibrium bulk distribution coefficient of Sc between garnet and matrix, and Z is a constant. Thus, if K’s are constant, any trace element that follows an exponential fractionation law during garnet growth will yield a straight line on a Log(Sc/El) vs. Log(Sc) plot (Fig. 6). The slope of the line is proportional to [(K(Sc)eK(El))/ (K(Sc)e1)], provided the scale of equilibrium for two elements is similar. This relation for slopes on Fig. 6 holds for element

4.2. Disequilibrium partitioning models (controlled by diffusion) This model (Shimizu, 1983) was derived for binary melt glass systems. Thus, it cannot be easily applied to multicomponent, silicate systems found in metamorphic rocks. With this caveat in mind, we apply the simplified models, assuming that the HHC pelitic gneiss samples can be approximated as two-phase (garnet and “matrix”) assemblages. The model is isothermal and assumes that transport of the nutrient material to the interface is matrix diffusion-limited, that the radius of curvature of a growing grain is large enough so that the problem may be reduced to one dimension, that the matrix diffusion coefficient has a constant value which is much larger than the diffusion coefficient in the growing mineral, and that growth is at a constant rate. Garnet growth dominated by matrix diffusion-controlled disequilibrium would show: (1) parallel depletion of compatible elements (Sc, Y, HREE) and enrichment of incompatible elements (V, Co) during episodes of garnet growth, (2)

G3

G4

G5

G6

G7

G9

G10

G11

G12

G13

Note: G1 to G18 are the location across the zoned garnet porphyroblast across rim to core. Elemental concentrations are measured in ppm unit.

R

G14

G15

G16

G17

G18

258.2 271.700 415.900 110.700 426.700 305.900 183.400 113.600 114.900 411.400 146.400 412.800 0.083 e e 0.079 e 0.084 0.071 e e e e e 0.078 0.310 0.182 e 0.116 0.250 0.340 0.076 0.247 0.075 0.057 e 0.084 0.206 0.140 0.127 0.152 0.253 0.197 0.056 0.042 0.084 e 0.050 2.486 4.017 3.107 0.824 3.285 3.485 4.393 1.346 0.907 3.513 0.964 0.395 7.141 4.279 8.268 3.074 6.044 5.286 8.547 2.643 2.598 5.072 2.505 1.737 1.926 1.590 2.500 0.421 1.460 1.053 1.511 0.335 0.192 0.556 0.356 0.185 16.78 6.623 13.660 9.646 9.562 7.415 15.080 12.720 13.350 13.570 14.500 10.030 5.975 2.558 4.171 2.724 3.130 2.745 3.342 3.320 3.747 6.683 6.009 5.726 48.26 33.820 49.920 17.990 41.970 35.800 28.400 21.300 26.440 60.770 46.080 57.980 10.49 11.220 18.920 4.276 17.890 12.890 7.241 4.353 4.165 18.240 7.178 16.160 26.04 39.930 78.990 13.580 89.230 46.950 23.220 14.840 9.356 67.030 11.990 53.250 3.624 6.482 16.340 2.995 20.700 8.366 3.327 2.545 1.191 9.303 0.848 7.464 22.51 49.350 142 22.120 211.800 70.960 26.000 19.290 10.190 58.550 4.428 52.520 2.385 7.307 23.790 3.773 39.460 11.580 3.479 3.169 1.278 7.786 0.480 7.654 0.492 e e 0.434 0.190 e 0.384 e 0.575 0.742 e 0.112 e 0.140 e e 0.153 0.232 e e e e e e 101.1 141.200 167.300 265.400 160.400 157.500 132.100 120.800 101.500 99.570 81.200 62.190 230 713.100 367.100 223.400 347.800 500.100 259.900 186.300 151.700 172.300 107.500 70.790 54.09 62.110 58.470 76.440 54.750 48.470 72.090 69.250 63.720 64.500 29.060 47.830 37.16 29.040 25.160 18.220 27.510 29.070 32.800 21.170 24.740 36.210 39.880 20.330 17.22 4.985 38.330 2.626 3.274 4.131 5.560 2.082 3.253 4.675 5.987 2.118 208.5 54.890 229.600 76.020 63.240 40.510 95.430 53.110 55.970 48.370 77.690 30.400

AC

29.360 343.700 364.800 339.500 140.100 e e e e e 0.071 e 0.099 e 0.071 e 0.047 0.099 e e 0.786 0.462 1.516 0.664 0.707 2.654 2.033 2.958 1.885 2.386 0.376 0.279 0.706 0.397 0.210 11.590 10.040 11.450 12.230 11.510 3.314 4.457 5.384 5.383 4.501 13.460 48.660 59.900 55.560 29.510 1.398 15.140 17 15.670 5.336 1.833 56.310 55.940 47.500 11.860 0.184 8.640 8.420 6.331 1.495 1.223 62.480 58.370 41.730 9.117 0.112 9.471 7.854 5.481 1.513 0.859 0.327 e e e e e e e 0.179 61.610 68.690 78.820 82.200 87.960 76.260 68.320 97.380 77.340 165 41.490 46.640 26.050 52.370 68.650 38.220 21.900 36.810 19.680 21.990 6.246 2.322 2.224 3.458 15.210 90.050 72.210 59.410 29.580 62.070

G8

R

23.070 e e e 0.601 1.482 0.183 7.075 2.036 8.476 0.915 1.348 0.136 0.824 0.130 0.335 e 52.730 80.130 41.230 41.020 12.110 113.700

G2

B. Kundu / Geoscience Frontiers 2(4) (2011) 539e549

Y La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Pb Th Sc Ti V Co Zr Cr

G1

ET

Elements

EPMA data of zoning garnet porphyroblast (7A/93/9).

TE D

546

Table 3

547

TE D

B. Kundu / Geoscience Frontiers 2(4) (2011) 539e549

AC

Figure 5 Trace element ion intensity zoning profile (rim-core-rim) in garnet crystal of sample 7A/93/9 (Table 3). In thin section (under plane polarized light (PPL)), rim portion of garnet porphyroblasts is nearly free of primary inclusions, whereas primary inclusions of plagioclase, quartz and biotite are present in the core portion. Patchy secondary biotites replace the garnet porphyroblasts along the grain boundary. Chondrite-normalized REE profile of that garnet sample shows significant variation of Eu anomaly and elemental enrichment from rim to core portion. Ce anomaly is not so much prominent as compared to Eu anomaly. Two red arrows indicate the position of the Ce and Eu anomaly of the chemically zoned garnet porphyroblasts. G: garnet; Bt: biotite; Sill: sillimanite; Pl: plagioclase; Qtz: quartz.

R

ET

R

much more significant depletions of compatible elements than enrichment of incompatible elements, (3) all trace elements should be affected unless matrix diffusion coefficients significantly vary for different elements, (4) element concentrations approach a steady state, and the most important point to be noted is that with the help of this model, not only we can calculate the growth rate of the crystal,

Figure 6 Equilibrium partitioning models (Rayleigh fractionation) of chemically zoned garnet crystals (sample No. 7A/93/9). In Log (Sc/ El) vs. Log (Sc) plots slope is proportional to partitioning coefficient. Only Co shows single-stage Rayleigh fractionation episode, however all other trace elements display multiple stages Rayleigh fractionation episodes or disequilibrium element partitioning. NRR: non-Rayleigh region.

but also we can determine the peak metamorphic period, provided there was no retrogression.

5. Discussion

Minor fluctuation in the trace and REE zoning profiles in HHC pelitic gneiss may reflect some specific petrogenetic process, but could also be due to the influence of micro-inclusions. Smooth and monotonous decreasing zoning profile in Sc, and almost flat profile in Co can be explained in terms of equilibrium closed-system partitioning models (Hollister, 1966). The elements exhibit for humps require either strong enrichment in the effective bulk composition (EBC) or a large increase in Ki over a narrow growth interval. A large change in Ki for an element is caused by a significant change in p, T, f(O2) of the melt. Presence of ilmenite and absence of magnetite suggest no systematic variation of f(O2). Large increase in p is very much unlikely, however the effect of increased p and increased T are generally in an opposite sense and may, to some extent, cancel each other. Thus the ultimate controlling factor must be an open-system behavior, i.e., the enrichment of the EBC. The two processes by which the EBC may be increased are (1) introduction of a pulse of trace element-enriched fluid into the system (metasomatism to be an explanation of anomalous trace element behavior), and (2) breakdown of a refractory mineral enriched in these trace elements. In the HHC pelitic gneiss, all the garnet samples show HREE enriched chondrite-normalized fractionated profiles. Studies on REE mobility in low-grade metamorphic, ocean-floor, and hydrothermal regimes generally show enhanced LREE rather than HREE (Wood et al., 1976), though F and CO2 appear to

B. Kundu / Geoscience Frontiers 2(4) (2011) 539e549 are separated or the melt quenches in intervals between the onset of melting and the attainment of chemical equilibrium between the melt and the whole volume of the solid. Two of the factors which determine the time required for equilibration are the grainsize of the solid (diffusion path length, D) and temperature (T ), and ln(D) is inversely proportional to T. If the melt forms and is then rapidly segregated from its residuum, the trace elements with low diffusivities in the solids will have low abundances in the segregated melt and elements with high diffusivities in melt will be concentrated in the melt even if the segregation is fast. Thus all of the observations support disequilibrium melting behavior.

Acknowledgment I am thankful to Prof. S. Dasgupta and Prof. S. Chakraborthy for providing me the EPMA, XRF and INAA data of the HHC pelitic migmatite. I am also indebted to Geological Sciences department, Jadavpur University, India, for providing me with all the laboratory facilities as per my needs. I thank Council of Scientific and Industrial Research for financial support.

References

Aitchison, J.C., Xia, X., Baxter, A.T., Ali, J.R., 2011. Detrital zircon U-Pb ages along the Yarlung-Tsangpo suture zone, Tibet: implications for oblique convergence and collision between India and Asia. Gondwana Research. doi:10.1016/j.gr.2011.04.002. Bea, F., 1996. Residence of REE, Y, Th and U in garnets and crustal protoliths: implication for the chemistry of crustal melts. Journal of Petrology 37, 521e552. Bea, F., Pereira, M.D., Stroh, A., 1994. Mineral/leucosome trace elements partitioning in a peraluminous migmatite (a laser ablation-ICP-MS study). Chemical Geology 117, 291e312. Brown, M., D’Lemos, R.S., 1991. The Caledonian granites of Mancellia, NE-Armorican Massif of France: relationship to the St. Malo migmatite belt, petrogenetic tectonic setting. Precambrian Research 51, 393e427. Chernoff, C.B., Carison, W.D., 1999. Trace element zoning as a record of chemical disequilibrium during garnet growth. Geology 27, 555e558. Cunery, M., Barbey, P., 1982. Mise an evidence de phenomenes de crystallization fractionnee danes les migmatites: Comptes-Rendus des Seances de l. Academiedes Sciences, Series 2 (295), 37e42. Fourcade, S., Martine, H., de Br€emond d’Ars, J., 1992. Chemical exchange in migmatite during cooling. Lithos 28, 43e53. Gansser, A., 1964. Alps & the Himalayas. 22nd International Geological Congress 11, 387e399. Hanson, G.H., 1978. The application of trace elements to the petrogenesis of igneous rock of granitic compositions. Earth and Planetary Science Letter 38, 26e43. Haskin, L.A., 1990. PREE conceptions pREEvent pREEcise pREEdictions. Geochimica et Cosmochimica Acta 54, 2353e2361. Hickmott, D.D., Shimizu, N., Spear, F.S., Selverstone, J., 1987. Traceelement zoning in a metamorphic garnet. Geology 15, 573e576. Hickmott, D.D., Shimizu, N., 1990. Trace element zoning in garnet from the Kwoiek area, British Colombia: disequilibrium partitioning during garnet growth? Contributions to Mineralogy and Petrology 104, 619e630. Hickmott, D.D., Spear, F., 1992. Major & trace elements zoning in garnets from calcareous pelites in NW Shelburns Falls Quadrangle, Massachusetts; garnet growth histories in retrograde rocks. Journal of Petrology 15, 965e1005. Hodges, K.V., 2000. Tectonics of Himalaya and southern Tibet from two perspectives. Geological Society of America Bulletin 100, 324e350.

6. Conclusion

R

AC

enhance the solubility of the HREE relative to the LREE (Haskin, 1990). Thus, a metasomatic origin for the trace element humps requires introduction of an unusual, perhaps F- or CO2-rich, trace element-bearing fluid flux into the system. But unfortunately we have no supporting evidence for this process. Thus, during the formation of garnet crystal, breakdown of refractory phase may control the formation of the humps. The humps actually represent event markers, which imply dissolution of refractory phases. Monazite is the source for Cr, REE, Y; apatite for REE, Y; and biotite for Ti. Presence of biotite inclusion may also be an alternative source for Ti. In fact, we know that Ti is less soluble in intergranular fluid than many other trace elements, thus primary inclusion of biotite may be potential source for Ti. Ilmenite may be the potential source for Ti, however as Ti is a structural element in ilmenite, Ti is only released when it breaks down to rutile, but we have no such petrographic evidence. Petrogenetic studies have established two types of anatectic melts: (a) equilibrium melts and (b) disequilibrium melts. According to Sawyer (1987) leucosomes derived through equilibrium melting are typically enriched in LREE, Th, and Hf and rich in Zr saturation, but are depleted in HREE relative to their source rocks. From the bulk-rock analysis it is clear that overall incomplete separation operated between selvedges and leucosomes. Both selvedges and leucosomes display enriched LREE with respect to HREE, but the difference in selvedges shows negative Eu anomaly in contrast to leucosomes showing positive Eu anomaly. Leucosomes are essentially enriched in incompatible elements compare to selvedges. Leucosome samples display enrichment of Th, U, Pb, Yb but relative depletion of Nb, Ce, Sr, Hf, Lu. Thus enrichment of LREE and Th supports equilibrium melting model but the Zr saturation is extensively low in leucosomes samples (6e55 ppm) compared to selvedges samples, which is relatively enriched in Zr (165e200 ppm) and depletion of Hf supports disequilibrium melting model.

TE D

548

ET

This study on the HHC pelitic migmatite from Sikkim Himalaya leads to the following general conclusions:

R

1. Though Zr is highly soluble in melt, Zr saturation is extensively low in the studied leucosomes, compared to selvedges. Furthermore the Hf concentration is depleted. The leucosomes in migmatite do not represent the entire melt component produced by partial melting because of the subsequent melt loss process. 2. In compatibility profile an incomplete separation of trace and REE between leucosomes and selvedges is observed. 3. The periodic humps in trace and REE zoning profiles of garnet crystal can be explained by strong enrichment in the EBC by breakdown of the refractory phases like monazite, apatite and most importantly the micro-inclusions of the accessory minerals (e.g., primary micro-inclusions of biotite in garnet).

Thus, when a melt forms during crustal anatexis, it is initially in chemical equilibrium with the surface of the grains with which it is in contact. Equilibration of the melt with the whole volume of the solid is a slow process that depends on the diffusion of the components in the solids and in the melt. However, disequilibrium melt compositions are preserved when the melt and solid

B. Kundu / Geoscience Frontiers 2(4) (2011) 539e549

TE D

Power, G.M., 1993. Geochemical differentiation between the Cadomian granites of Mancellia and the St. Malo migmatites, Armorican Massif, France. Journal of the Geological Society London 150, 465e468. Sawyer, E.W., 1987. The role of partial melting and fractional crystallization in determining the discordant migmatite leucosome compositions. Journal of Petrology 28, 445e473. Shimizu, N., 1983. Interface kinetics and trace element distributions between phenocrysts and magma. In: Augustithis, S.S. (Ed.), The Significance of Trace Elements in Solving Petrogenetic Problems and Controversies, pp. 175e195. Theophrastus, Athens, Greece. Taylor, S.R., McLennan, S.M., 1985. The Continental Crust: Its Composition and Evolution. Blackwell, Oxford, 312 pp. Watt, G.R., Burns, I.M., 1996. Chemical characteristics of migmatites. Contributions to Mineralogy and Petrology 125, 100e111. Watt, G.R., Harley, S.L., 1993. Accessory phase controls on the geochemistry of crustal melts and restites produced during water-undersaturated partial melting. Contributions to Mineralogy and Petrology 114, 550e556. Wood, D.A., Gibson, I.L., Thompson, R.N., 1976. Elemental mobility during zeolite facies metamorphism of the Tertiary basalts of eastern Iceland. Contributions to Mineralogy and Petrology 55, 241e254.

R

ET

R

AC

Hollister, L.S., 1966. Garnet zoning: an interpretation based on the Rayleigh fractionation model. Science 154, 1647e1651. Icenhower, J., London, D., 1995. An experimental study of element partitioning among biotite, muscovite, and coexisting peraluminous silicic melt at 200 Mpa (H2O). American Mineralogist 80, 1229e1251. Johannes, W., Holtz, F., Moller, P., 1995. REE distribution in some layered migmatites: constraints on their petrogenesis. Lithos 35, 139e152. Mahood, G., Hildreth, W., 1983. Large partitioning coefficients for elements in high silica rhyolites. Geochimica et Cosmochimica Acta 47, 11e30. Nabelek, P.I., 1999. Trace element distribution among rock-forming minerals in Black Hills migmatites, South Dakota: a case study for solid-state equilibrium. American Mineralogist 84, 1256e1269. Nabelek, P.I., Glascock, M.D., 1995. REE-depleted leucogranites, Black Hills, South Dakota. Journal of Petrology 36, 1055e1071. Nash, W.P., Crecraft, H.R., 1985. Partition coefficients for trace elements in silicic magmas. Geochimica et Cosmochimica Acta 49, 2309e2322. Neogi, S., Dasgupta, S., Fukuoka, M., 1998. High p-T polymetamorphism, dehydration melting and generation of migmatites and granites in Higher Himalayan Crystalline Complex, Sikkim, India. Journal of Petrology 39, 61e99.

549