Early Permian mantle–crust interaction in the south-central Altaids: High-temperature metamorphism, crustal partial melting, and mantle-derived magmatism

    Early Permian mantle–crust interaction in the south-central Altaids: Hightemperature metamorphism, crustal partial melting, and mantle-derived magmatism T.N. Yang, J.Y. Li, M.J. Liang, Yu Wang PII: DOI: Reference:

S1342-937X(14)00180-4 doi: 10.1016/j.gr.2014.05.003 GR 1264

To appear in:

Gondwana Research

Received date: Revised date: Accepted date:

31 August 2013 4 May 2014 5 May 2014

Please cite this article as: Yang, T.N., Li, J.Y., Liang, M.J., Wang, Yu, Early Permian mantle–crust interaction in the south-central Altaids: High-temperature metamorphism, crustal partial melting, and mantle-derived magmatism, Gondwana Research (2014), doi: 10.1016/j.gr.2014.05.003

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

SC R

IP

T

Manuscript (reversion 3) prepared for Gondwana Research

Early Permian mantle–crust interaction in the south-central Altaids: high-temperature metamorphism, crustal partial melting, and mantle-derived

TE

D

MA

NU

magmatism

CE P

T. N. Yang, J. Y. Li, M. J. Liang

Institute of Geology, Chinese Academy of Geological Sciences, Beijing 100037, China

AC

Yu Wang

Geological Laboratory Center and Department of Geology, China University of Geosciences, Beijing, 100083, China

Corresponding author: Tian-Nan Yang E-mail address: [email protected] Telephone number: *86-10-68999722 1

ACCEPTED MANUSCRIPT ABSTRACT Early Permian mafic magmatism, the partial melting of the crust, and

IP

T

high-temperature metamorphism in the Chinese Altai, south-central Altaids, provide

SC R

an excellent case study of exchange of energy and mass between the mantle and crust. Field and petrographic observations, together with microprobe mineral analyses, have, for the first time, allowed us to identify scapolite-bearing calc-silicate granulites along

NU

the southern margin of the Chinese Altai. The conditions of metamorphism were 680

MA

to 800°C and 6–7 Kbar, based on mineral phase relationship and compositions, as well as the results of previous studies. New LA–ICP–MS zircon U–Pb geochronology

D

demonstrates that the high–T metamorphism in the Chinese Altai was accompanied by

TE

the emplacement of leucogranites at approximately 295 Ma. SHRIMP II dating results,

CE P

combined with previously published data, reveal a mafic magmatic event at around 275 Ma in a large region of northern Xinjiang, NW China. This mafic magma was

AC

derived from N-MORB-like depleted mantle, as deduced from our new bulk geochemical and zircon Hf isotopic data. Synthesizing available geochemical (including isotopic) and geochronological data, we propose a two-stage model of mantle–crust interaction to explain the early Permian geology of northern Xinjiang. The early stage of interaction involved high-temperature metamorphism and the coeval partial melting of crustal rocks, indicating solely a heat exchange between the mantle and crust; in contrast, the later stage involved the exchange of both mass and energy between the mantle and crust. Keywords: South-central Altaids, high-temperature metamorphism, partial melting, 2

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

Early Permian, mantle–crust interaction.

3

ACCEPTED MANUSCRIPT 1. Introduction The Altaids in Central Asia (inset of Fig. 1, Sengör et al. 1993) is the world’s largest

IP

T

accretionary orogen (Xiao et al. 2010, 2013; Wilhem et al. 2012; Xiao and Santosh

SC R

2014) and were formed during the Paleozoic by the accretion and amalgamation of different types of allochthonous fragments such as island arcs, ophiolites, accretionary prisms, seamounts, oceanic plateaus, and continental blocks (Sengör et al. 1993; Xiao

NU

et al. 2004a, b, 2009, 2010; Windley et al. 2007). Igneous activity in the Altaids

MA

spanned the entire Phanerozoic, including the Mesozoic, and was characterized by the production of abundant granitoids and volcanic rocks showing significant admixture

D

of the mantle component (Jahn et al. 2000; Kröner et al. 2014). Numerous new

TE

geochronological data derived from modern techniques (some are marked on Fig. 1)

CE P

have revealed abundant latest late Carboniferous to early Permian bimodal intrusion or extrusion of magmatic rocks throughout the south-central Altaids (inset of Fig. 1)

AC

which have been identified within all tectonic units. It would appear that early Permian basalt developed in all basins, whereas the coeval bimodal plutons only occur in the basin-bounding ranges where the pre-Permian subduction-accretion related rocks are exposed. Given this context, the south-central Altaids in Chinese Xinjiang provide a natural laboratory for the study of mantle–crust interactions. In general, there are two aspects to mantle–crust interactions: exchanges in mass, and exchanges in energy (heat) between the mantle and crust. Most previous studies of the south-central Altaids have focused on the geochronology and geochemistry of the magmatic rocks (e.g., Han et al. 1997, 2004, 2006; Geng et al. 2009; Wang et al. 4

ACCEPTED MANUSCRIPT 2006, 2009a; Zhao et al. 2009), thereby providing fundamental data for constraining of mass exchanges. However, except for the high-pressure metamorphic rocks in the

IP

T

western Tianshan (e.g., Gao and Klemd 2000), little attention has been paid to the

SC R

metamorphic rocks mainly because of the generally low grade of metamorphism of the south-central Altaids. However, some high-temperature metamorphic rocks are exposed in the Chinese Altai (Wei et al. 2007; Wang 2009b), and these may provide

NU

constraints on the possible exchange of heat between the mantle and crust. Zhuang

MA

(1994) identified several metamorphic domains in the Chinese Altai, which consist of Barrow-type metamorphic belts with migmatitic granitic gneisses at their cores. Yang

D

et al. (2011a) pointed out that the metamorphism in the Chinese Altai is highly

with

leucogranites.

Recently

published

geochronological

and

CE P

relationship

TE

heterogeneous in space, and that the metamorphic domains have an obvious spatial

thermochronological data (e.g., Laurent–Charvet et al. 2003; Briggs et al. 2007; Wang

AC

et al. 2009b) indicate that the regional metamorphism in the Chinese Altai occurred exclusively during the late Paleozoic, regardless of the fact that the magmatism in this region spanned from the Cambrian (Windley et al. 2002; Yang et al. 2011a) to the Permian (e.g., Han et al. 1997, 2006). Thus, the Chinese Altai is an ideal region for studying

these

possibly

complex

relationships

between

magmatism

and

metamorphism. In this paper, we document a newly identified scapolite-bearing granulite in the Altai region, northernmost Xinjiang, NW China (Fig. 1). We determined the metamorphic age of this granulite using in situ LA–ICP–MS zircon U–Pb dating. In 5

ACCEPTED MANUSCRIPT addition, we report new LA–ICP–MS zircon U–Pb data for two garnet-bearing leucogranites, and sensitive high-resolution ion microprobe (SHRIMP II) zircon U–Pb

IP

T

data for a gabbroic pluton; for the gabbro, we also provide bulk geochemistry and

SC R

zircon Lu/Hf isotopic data. We then synthesize these data with previously published geochronological and geochemical data, on which basis we discuss the spatial and temporal relationships between magmatism and metamorphism in northern Xinjiang.

MA

2. Geological Setting and Sampling

NU

This case study helps to shed light on mantle–crust interactions in general.

D

Our study area is located in the southeastern segment of the Chinese Altai, which

TE

along with the Siberian Altai comprise the Altai–Mongolia microcontinent (Wilhem et

CE P

al. 2012) in the south-central Altaids (Fig. 1). The Ertix (Irtysh–Zaysan) suture zone separates the Chinese Altai in the north from a series of Devonian to Carboniferous

AC

island arc belts (Zhang et al. 2008a; Chen et al. 2010, and references therein) in the south (Fig. 2). The Chinese Altai is regarded to have been a passive continental margin during the Cambrian to Middle Ordovician (Windley et al. 2003; Wilhem et al. 2012), consisting of fine-grained turbidite clastics and bimodal volcanic rocks (Windley et al. 2003, 2007; Yang et al. 2011a). From the Late Ordovician (Wilhem et al. 2012, and references therein), a continental arc started to develop along the passive continental margin, giving rise to widespread lower to middle Paleozoic arc-like volcanics and volcaniclastics and associated plutons (Xu et al. 2002; Wang et al. 2006; Yuan et al. 2007; Sun et al. 2008a; Yang et al. 2011a).

6

ACCEPTED MANUSCRIPT Meanwhile, one or more early–middle Paleozoic island arcs developed in north Junger, south of the Ertix suture (e.g., Zhang et al. 2008a; Chen et al. 2010). Thus, the

IP

T

Ertix suture represents a subduction zone along which an ocean slab subducted

SC R

northwards under the Altai–Mongolia microcontinent, giving rise to a continental-arc belt. This suture was subsequently modified by a collision between the continental-arc belt and the northern Junger island arc belts (e.g., Briggs et al., 2007), and likely

NU

contains juxtaposed crust slices with different geochemical (including isotopic)

MA

features (e.g., Wang et al. 2009a). Numerous early Permian granitic and gabbroic plutons intruded the pre-Permian rocks along both sides of the Ertix suture (see Fig. 2)

D

(e.g., Han et al. 2004; Wang et al. 2005; Tong et al. 2006a and b; Zhou et al. 2007,

TE

2009; Zhu et al. 2009), which suggests that the collision was likely completed before

CE P

the early Permian. At present time, however, only rare structural data are available to constrain this collision and its relationship with the Ertix shear zone.

AC

The early to middle Paleozoic volcanics and volcaniclastics and coeval plutons of the Chinese Altai, north of the Ertix suture, are metamorphosed (Zhuang 1994; Wei et al. 2007; Wang et al., 2009b), whereas the Devonian to Carboniferous island-arc volcanics and volcaniclastics in northern Junger, south of the Ertix suture, are generally not metamorphosed. Two slices can be identified in term of metamorphic degree for the Chinese Altai despite the fact that both have the same protolith assemblage as mentioned above (Fig. 2): (1) the relatively-low-grade Altai–Qinghe slice, and (2) the high-grade Fuyun slice. Between these two slices is the high-angle, northward-dipping Altai–Qinghe normal fault. 7

ACCEPTED MANUSCRIPT The metamorphism of the Altai–Qinghe slice is highly heterogeneous. This slice can be broadly subdivided into two segments separated by the dextral Fuyun

IP

T

fault (Fig. 2). Its eastern segment, in the Qinghe area, has a relatively higher degree of

SC R

metamorphism than that of its western equivalent in the Altai-Fuyun area. Several metamorphic domains with strongly migmatized cores can also be identified in the eastern segment, and they display the following metamorphic mineral zones as

NU

indicated by the first appearance of an index mineral (from the core outwards):

MA

sillimanitekyanitegarnetbiotite; biotite and garnet are common in the inner two zones. The results of geological mapping (Xinjiang Bureau of Geology and Mineral

D

Resource (XBGMR) 1989) have revealed much more abundant early Permian

TE

leucogranitic plutons within the eastern segment than within the western segment (Fig.

CE P

2). The metamorphic grade of the western segment is evidently much lower than that of the eastern segment. Most Cambrian clastics in the western segment have been

AC

metamorphosed to schist or paragneiss consisting of quartz, biotite, muscovite, chlorite, K-feldspar, and garnet suggesting greenschist facies metamorphism. The metamorphic degree of the Early Devonian volcanic or volcaniclastic rocks in the western segment is broadly very low (i.e., greenschist to sub-greenschist facies), whereas the Middle to Late Devonian clastics or bioclastics in the region of Altai City are generally unmetamorphosed (The Team One of Geology of Xinjiang Province 1976) (Fig. 3). The Fuyun slice consists of high-grade metamorphosed clastic and volcanic rocks and mylonitized granodiorites, in which pelitic granulites have been identified (Wang 8

ACCEPTED MANUSCRIPT et al. 2009b). The clastics and volcaniclastic rocks of this slice have been metamorphosed to garnet- (and locally) orthopyroxene-bearing gneiss (Fig. 4a) with

IP

T

widespread migmatite veins. These veins contain numerous euhedral garnet and

SC R

tourmaline crystals. Locally, the migmatite veins enlarge and gradually become part of a leucogranite pluton that also contains garnet and tourmaline. Abundant early Permian leucogranites (Tong et al. 2006; Han et al. 2006; Zhou et al. 2007; this study)

NU

or gabbros (Han et al. 2004; Zhang et al. 2008b; this study) are exposed throughout

MA

this slice (Figs. 2 and 3).

The Cambrian to Devonian rocks of the Fuyun slice were intensively sheared to

D

mylonite, and define the Ertix shear zone. In contrast, the early Permian leucogranite

TE

in the Fuyun slice can be divided into two types in term of deformation (Fig. 2): (1)

CE P

highly to moderately sheared leucogranite; and (2) leucogranite lacking deformation. The deformed leucogranites are always enriched with both garnet and tourmaline, and

AC

show elongate shapes in map view with long axes parallel to the strike of the slice. The undeformed leucogranites are biotite bearing but lack peraluminous mafic mineral, and show rounded shapes in map view (Fig. 2). Available geochronological data indicate that the undeformed leucogranites have ages < 285 Ma (e.g., Wang et al. 2005; Tong et al. 2006a; Han et al. 2006), whereas the deformed leucogranites are commonly older with zircon U–Pb age > 286 Ma (e.g., Zhou et al. 2007; Briggs et al. 2009; this study). These data suggest that the main phase of shearing along the Ertix shear zone occurred during the earliest early Permian. It is worth pointing out that all Permian granites in northern Junger, south of the Ertix suture, are undeformed and 9

ACCEPTED MANUSCRIPT that they display rounded shapes in map view (Fig. 2). In the Fuyun slice, several relatively large mafic to ultramafic plutons roughly

IP

T

defining a belt are located in the regions between the cities of Altai and Fuyun (Fig. 2).

SC R

In addition, many small gabbroic stocks (with diameter < 50 m, not shown on Fig. 2) occur extensively throughout the entire Fuyun slice. These mafic bodies have a rounded shape in map view and have well-preserved intrusive contacts. One small

NU

gabbro body was selected for petrological, geochemical, and geochronological study.

MA

Numerous metamorphosed calc-silicate rocks, with individual exposing areas of 0.1 m  1 m to 8 m  40 m, are limited to a small region of the Fuyun slice (Fig. 3).

D

Some xenoliths of the metamorphosed calc-silicate rocks are present in slightly

TE

deformed early Permian leucogranite, whereas others exhibit folded layers that are

CE P

intercalated with garnet- and diopside-bearing paragneiss and minor marble (Figs. 4a, b, and c). Most garnet of the xenoliths has been pseudomophosed by epidote. A

AC

leucogranite with xenoliths and a paragneiss were sampled for zircon geochronology (Fig. 3).

The high-grade layers were subjected to two stages of folding, resulting in interference fold patterns (Figs. 4a and b). The granulite facies minerals (see below sections 3.3 for details) lack preferred orientations, which suggests that the first stage of folding developed before the granulite facies metamorphism. However, the late-stage folding led to the development of some spaced cleavages that cut across the granulite facies minerals. Further detailed structural study is required to reveal the relationship between the deformation and the high-temperature metamorphism. 10

ACCEPTED MANUSCRIPT 3. Petrology

T

3.1. Early Permian gabbros

IP

The early Permian gabbros are dark, massive, and quite homogeneous rocks, without

SC R

any cumulate structures, consisting of equally coarse (ca. 2–6 mm) euhedral to sub-euhedral pyroxene and plagioclase. Accessory minerals include Fe–Ti oxides and

NU

minor zircon. Plagioclase is slightly more euhedral than pyroxene. It is worth pointing out that no textural variations have been identified from core to margin of the gabbro

3.2. Early Permian leucogranites

MA

pluton, regardless of the sharp contact with the country rocks.

TE

D

The leucogranites are made up of K-feldspar, quartz, plagioclase, biotite, muscovite, and minor garnet; accessory minerals include epidote and zircon. Tourmaline is

CE P

common but distributed heterogeneously; locally, the coarse grain-size is similar to tourmaline present in pegmatite veins. Biotite and muscovite and slightly elongated

AC

quartz are weakly preferred-oriented, defining gneiss foliation; meanwhile undulating extinction is common in the elongate quartz grains. 3.3. Calc-silicate granulite Field observations and examination of thin-sections demonstrate that the granulite facies rocks have simple mineral assemblages consisting of one to four of the minerals quartz, garnet, plagioclase, diopside, wollastonite, scapolite, and calcite; accessory minerals include apatite, zircon, zoisite, and titanite. All the granulite rocks consist of thin (ca. 10 cm) metamorphic calc-silicate layers that are characterized by sharp variations in mineral modal contents (Fig. 4c). Plagioclase-rich monomineralic layers 11

ACCEPTED MANUSCRIPT are common, most of which have irregular and gradational boundaries, and some of which have clear cut-crossing relationship with the host metamorphic rock (Fig. 4d).

IP

T

Mineral compositions and texture feature described in below sections would indicate

SC R

that there was rare mass exchange between adjacent layers during metamorphism; thus we suggest that these monomineralic layers represent metamorphosed liquefied calc-sandstone dykes containing mudstone breccias.

NU

On the basis of variations in modal contents of the seven principal minerals listed

MA

above, the granulite facies rocks can roughly be subdivided into seven lithologic types: marble, quartzite, garnetite, clinopyroxene-rich rocks, wollastonite-rich rocks,

D

scapolite-rich rocks and plagioclase-rich rocks, all of which form distinct layer-like

TE

structures. Under the microscope, the boundaries between different layers are

CE P

distinctive, though often with narrow gradational zones. (1) The marbles consist of coarse calcite grains that enclose numerous

AC

fine-grained euhedral diopsides. Minor interstitial scapolite occurs, which displays intergrowth with epidote, calcite, and minor diopside. (2) In the quartzites, the quartz forms a groundmass to assemblages of either (1) garnet + diopside + plagioclase, (2) garnet + diopside, or (3) garnet + wollastonite + diopside (Fig. 5). The quartz content decreases from quartzites with assemblage (1) to quartzites with assemblage (3). The mineral phases in all three kinds of assemblage have highly irregular, even vermicular, boundaries. Most diopside is set in poikiloblastic garnets as isolated, roundish grains, whereas the wollastonite, if present, is commonly euhedral or sub-euhedral occurring as porphyroblasts (Fig. 5). 12

ACCEPTED MANUSCRIPT (3) The garnetites are made up mainly of garnet that displays a poikiloblastic texture where the enclosed minerals are mainly diopside, less commonly titanite (Fig.

IP

T

6), or sometimes calcite, scapolite, and quartz (Fig. 7a, b). In the scapolite-bearing

SC R

garnetite, symplectites of vermicular garnet, quartz, and calcite, set in a groundmass of plagioclase, are developed in some domains (Figs. 7a, c). These symplectites may have resulted from a reaction involving scapolite and wollastonite.

NU

(4) The wollastonite-rich rock layers are intercalated with garnetite layers, and

MA

consist of nearly pure wollastonite with minor diopside (Fig. 6). (5) Most of the clinopyroxene-rich rocks are predominantly made up of diopside,

D

along with plagioclase and garnet (Fig. 8a). The garnet is vermicular and set in the

TE

plagioclase forming symplectites on the microscale (Fig. 8b). The symplectites of

CE P

plagioclase + garnet + minor quartz form the groundmass to large diopsides with highly irregular grain boundaries (Fig. 8a), and these, in turn, form symplectites on

AC

the thin section scale.

(6) In the scapolite-rich rocks (Figs. 9a and b), the scapolite with irregular grain boundaries seem to connect with each other, displaying a matrix to irregularly-shaped garnets and diopsides. (7) The plagioclase-rich layers consist mainly of pale plagioclase and sporadically distributed blue scapolite, and the layers enclose numerous garnetite blocks that stand out in vivid contrast (Fig. 10a). The plagioclase grains are commonly euhedral or sub-euhedral, and the interstices are filled by cuspate or film-like scapolite (Figs. 10a and b). The garnetite blocks are generally quartz-bearing, 13

ACCEPTED MANUSCRIPT and they contain some large diopside porphyroblasts up to 1.0 mm in size (Fig. 10a). Some tiny garnets are enclosed in plagioclase. Large garnetite blocks commonly have

IP

T

boundaries with numerous embayments (Fig. 10a), whereas small ones are generally

SC R

round (Figs. 10a, b).

In summary, the following mineral assemblages are present in the granulite rocks from the Chinese Altai (not including accessory minerals):

NU

(1) Quartz + garnet + diopside ± wollastonite (e.g., Fig.5)

MA

(2) Quartz + garnet + diopside ± plagioclase (e.g., Fig. 8b) (3) Garnet + calcite + scapolite (e.g., Fig. 7a)

D

(4) Scapolite + garnet + diopside + minor quartz (e.g., Figs. 9a, b)

TE

(5) Calcite + diopside + minor scapolite (e.g., marble)

CE P

(6) Plagioclase + scapolite (plagioclase-rich layer) (e.g., Fig. 10b) Coexisting plagioclase and wollastonite or scapolite and wollastonite have not

AC

been identified in any of the assemblages. Except for (6), these assemblages are very common in calc-silicate granulites worldwide, as for example in Mary Kathleen, Australia (Ramsay and Davidson 1970), Rayagada, India (Shaw and Arima 1996; Satish–Kuma and Harley 1998), Eastern Antarctica (Harley and Buick 1992; Buick et al. 1993), and Maligawila, Sri Lanka (Mathavan and Fernando, 2001).

4. Analytical methodologies

In order to study the metamorphic evolution of the granulites, and to reveal the possible relationships between the metamorphism and magmatism, more than fifty

14

ACCEPTED MANUSCRIPT samples were taken for petrography, mineral chemistry, geochemistry, and geochronology. The samples for geochronology include, respectively, one

IP

T

leucogranite from the Altai-Qinghe slice (X0522-7) and another leucogranite from the

SC R

Fuyun slice (X0504-13); one gabbro (X0478-8) and one paragneiss (X0509-11) samples from the Fuyun slice. Their locations are labeled on Figs. 2, 3, and 4. Five fresh gabbro samples were collected for geochemical analyses, while other samples

NU

are mainly of granulite for petrological observations followed by microprobe analyses,

and data-processing procedures.

MA

the locations of which are indicated on Fig. 4. We now briefly describe our analytical

D

Mineral compositions were determined with a JEOL JXA 8100 SUPERPROBE

TE

at the Key Laboratory of Nuclear Resources and Environment, East China Institute of

CE P

Technology, Nanchang, China. Operating conditions were a 15kV accelerating voltage, a 20 nA specimen current, and a 2µm bean diameter. Natural minerals and synthetic

AC

oxides were used as internal standards during the analytical procedure. Most mineral formulae were calculated using the MINPET software (version 2.0, Richard 1994). Ferric iron in clinopyroxene was estimated by assuming a charge-balance based on a total of six oxygen atoms per formula unit. Ferric iron in garnet was calculated assuming stoichiometry, as described by Droop (1987). Analytical results are presented in Supplementary Tables 1 to 5, and some of the results are plotted in diagrams (Fig. 11). Major element analyses of the five gabbro samples were undertaken at the National Research Center of Geo-analysis, Chinese Academy of Geological Sciences 15

ACCEPTED MANUSCRIPT (CAGS), Beijing, using a wavelength X-ray fluorescence spectrometer (Rigaku, 3080E). Analytical errors were less than 2%. Trace element analyses were undertaken

IP

T

at the same center using an inductively coupled plasma-mass spectrometer (ICP-MS,

SC R

X-series). The analytical errors were less than 5% for elements in concentrations of >10 ppm, less than 8% for concentrations of <10 ppm, and about 10% for the transition metals.

NU

Zircons of the four samples were obtained using the standard techniques of

MA

crushing, heavy liquids, and magnetic separation. We handpicked 150 to 200 grains from the >25 m non-magnetic fraction of each sample, and these grains were then

D

mounted and cast in an epoxy resin disc, together with the standard TEM (206Pb/238U

TE

age of 417 Ma), for photographing under reflected and transmitted light and obtaining

CE P

cathodoluminescence (CL) images. CL images of the zircons were obtained using a HITACH S-3000N scanning microscope fitted with a Gatan Chroma at Beijing

AC

SHRIMP center.

The U–Pb analyses of zircons from gabbro sample X0478-8 were performed on a SHRIMP-II instrument at Curtin University of Technology, Western Australia. The instrument was controlled, and data acquired, from a remote control center in the Beijing SHRIMP Centre, Institute of Geology, CAGS, Beijing. This was achieved by using the SHRIMP Remote Operation System (SROS), developed jointly by the Beijing SHRIMP Centre, Chinese National Institute of Meteorology, and Jilin University. SROS allows the remote operator to control the instrument, choose sites for analysis, and print out data in real time through the internet. The details of the 16

ACCEPTED MANUSCRIPT analytical procedures used with the SHRIMP are similar to those described in Song et al. (2002) and Williams and Claesson (1987). The analytical data are shown in

IP

T

Supplementary Table 7.

SC R

The U–Pb analyses of zircons from leucogranite samples X0504-13 and X0522-7, and from the paragneiss sample X0509-11, were undertaken on a Finnigan Neptune multiple collector ICP-MS equipped with a New Wave UP 213 laser-ablation system

NU

installed in the MRL Key Laboratory of Metallogeny and Mineral Assessment,

MA

Institute of Mineral Resources, CAGS, Beijing. These analyses were carried out with a beam diameter of 25 m, a repetition rate of 10 Hz, and an energy of 2.5 J/cm2.

D

Zircons GJ-1 and Plešovice were used as internal standards during the analyses. The

TE

analytical procedures followed those described in Hou et al. (2009). The analytical

CE P

data are shown in Supplementary Table 8. Measurements of Hf isotopes and rare earth elements (REE) in the zircons from

AC

the gabbro sample X0478-8 were carried out using a multiple collector-inductively coupled plasma-mass spectrometer (MC-ICP-MS, Nu Plasma HR) connected to an ArF excimer laser-ablation system (GeoLas 2005, 193nm wavelength) at the State Key Laboratory of Continental Dynamics, Northwest University, China. Standards 91500 and Mon-1 were used during the analyses. All analyses were carried out with a beam diameter of 44 m, a repetition rate of 8 Hz, and energy of 2.4 J/cm2. The analytical procedures were similar to those described by Yuan et al. (2008). Off-line selection and integration of background and analytical signals, and time-drift corrections and quantitative calibrations for trace element analyses and 17

ACCEPTED MANUSCRIPT U–Pb dating, were performed using the software ICPMSDataCal (Liu et al. 2008). The concordia diagrams and weighted mean ages were made and calculated using

IP

T

software Isoplot/EX, ver. 3.00 (Ludwig 2003). Individual analyses for REE

The measured

176

SC R

concentrations are presented with 1σ errors in Supplementary Table 9. Lu/177Hf ratios (Supplementary Table 10) and the

176

Lu decay

constant of 1.865  10–11 year–1 reported by Scherer et al. (2001) were used to Lu/177Hf ratios. The chondritic values of

176

Hf/177Hf = 0.0332 and

Lu/177Hf = 0.282772, reported by Blichert–Toft and Albare`de (1997), were adopted

MA

176

176

NU

calculate initial

for the calculation of Hf(t) values (the parts in 104 deviation of initial Hf isotope

D

ratios between the zircon sample and the chondritic reservoir). The depleted mantle

TE

Hf model ages (TDM) were calculated using the measured 176Lu/177Hf ratios of zircon,

growth from

176

Hf/177Hf = 0.279718 at 4.55 Ga to 0.283250 at present, with

Lu/177Hf = 0.0384 (Griffin et al. 2000).

AC

176

CE P

based on the assumption that the depleted mantle reservoir has a linear isotopic

5. Analytical results 5.1. Mineral chemistry Garnet from all the granulite facies samples is essentially a grossular–andradite solid-solution (ss), with minor almandine, pyrope, and spessartine components (maximum up to 10 mol%, Fig. 11a). Garnets from different samples have highly variable grossular contents that range from 35 to 90 mol%. In contrast, the variation in the grossular content among the different grains of one sample is generally less than 18

ACCEPTED MANUSCRIPT 15 mol % (Fig. 11a). The variations would appear to reflect variations in original bulk chemical compositions of the meta-clastic rocks. For example, garnet from quartzite

IP

T

sample X0502-10 contains 49–54 mol% grossular, but garnet from quartzite sample

SC R

X0503-4 contains ca. 66 mol% grossular. The same is true for garnet in the garnetite samples. Garnet in scapolite-bearing assemblages has a relatively high grossular content of up to 86 mol%.

NU

Clinopyroxene from all metamorphic rocks is a diopside–hedenbergite solid

MA

solution (Supplementary Table 2) with a very minor jadeite component (0.5–1.9 mol%). For the whole dataset, the variation in Mg# [Mg/(Mg + Fe2+)] is large, ranging

D

from 80 to 64, but for a single sample, the range of Mg# values is very narrow (< 5;

TE

Fig. 11b). This observation likely reflects the variation in bulk chemistry of the rocks.

CE P

In all cases, wollastonite is an almost pure phase (Supplementary Table 2) with just minor amounts of MnO (up to 0.5 wt%). The anorthite content of the plagioclase in

AC

the plagioclase-rich layer is in the range 95–100 mol%, but plagioclase in other samples is slightly less rich in anorthite, with 91–97 mol%; the amount of the orthoclase molecule present is insignificant (<0.2 mol%) (Supplementary Table 3). The scapolite has a homogeneous composition with the meionite component (100Ca/(Ca + Na + K), after Deer et al. 1996) varying from 75 to 83% (Supplementary Table 4). Several analyses show that the accessory minerals such as titanite and zoisite have constant compositions (Supplementary Table 5).

19

ACCEPTED MANUSCRIPT 5.2. Mineral reactions and metamorphic P–T conditions of the Altai calc-silicate granulites

IP

T

On the basis of field and petrological observations, supported by mineral chemistry,

SC R

the metamorphosed calc-silicate rocks of the Chinese Altai have the following characteristics:

NU

(1) It is evident that the diversity of assemblages in the calc-silicate rocks is determined by the marked inhomogeneity of the precursor carbonate-bearing rocks.

MA

For example, wollastonite does not occur with plagioclase and scapolite, and the coexistence of scapolite + plagioclase is rare, except in the plagioclase-rich layers.

D

The wollastonite- or scapolite-bearing rocks contain relatively low amounts of quartz.

TE

These observations indicate that the protoliths of the Altai granulites were layered

CE P

calc-silicate clastic rocks of varying bulk chemical composition, and suggest that mass exchange between adjacent layers was very limited during metamorphism.

AC

(2) In the granulite-facies rocks, one mineral phase forms a matrix that encloses the other minerals to form an interpenetrative mosaic. The dihedral angles between adjacent phases are commonly greater than 160° or less than 30°. Equilibrium textures with 120° triple-point junctions are extremely rare. In fact, except for the marbles and the plagioclase-rich rocks, the minerals of the granulite-facies rocks exhibit symplectitic textures at the thin-section scale. These observations indicate that textural equilibrium was not attained. (3) The garnets from different layers have very different compositions, but within a single thin layer the composition of the garnet is quite constant. A similar 20

ACCEPTED MANUSCRIPT pattern is observed for the diopside. In contrast, the plagioclase and scapolite in all samples have similar compositions, despite variable mineral modal compositions

IP

T

indicating that chemical equilibrium was attained within each metamorphosed

SC R

calc-silicate layer.

A reasonable interpretation of the textural and compositional data described above may include the following:

NU

(1) The Altai scapolite-bearing granulites are the result of rapid metamorphic

MA

reactions due to intensive heating over a short time period, possibly causing rapid nucleation.

D

(2) Subsequently, the rapid reactions slowed, and immature textures were locked

TE

in because of the absence of hydrous fluids in the metamorphosed calc-silicate clastic

CE P

rocks. Except for minor epidote, very few hydrous minerals have been identified in the granulites. As a result, the chemical heterogeneity of the rocks has been preserved

AC

owing to limited material diffusion. The garnets from the Altai granulites contain a substantial amount of andradite component, from 6 to 57 mol%. Furthermore, the garnets from the domains where scapolite or wollastonite is present usually have lower andradite contents (<25 mol%); otherwise, the andradite component is generally >30 mol%. No Fe-Tschermak component was found in the clinopyroxenes, which helps to rule out the possibility that the clinopyroxene is a source of Fe3+ for the garnet; oxidation reactions caused by the activity of pervasive O2-bearing fluids (e.g., Shivaprakash 1981) cannot explain either the layer-controlled variations in the andradite contents of garnet or the values 21

ACCEPTED MANUSCRIPT of Mg# in the clinopyroxene. Thus, the high modal wollastonite and scapolite contents of these rocks indicate high calcite contents in their protoliths. We also note here that

IP

T

the layers with higher calcite content contain less ferric content.

SC R

Based on the above discussion, two mineral reactions may be deduced with respect to the metamorphism: (1) quartz + calcite = wollastonite + CO2, and as either quartz or calcite was totally consumed during the reaction, the univariant quartz +

NU

calcite + wollastonite assemblage is not found; and (2) plagioclasess + calcite =

MA

scapolitess, and similarly, the univariant plagioclase + calcite + scapolite assemblage is not found, suggesting that one phase of either calcite or plagioclase was totally

D

consumed during the reaction. The absence of coexisting wollastonite and anorthite

TE

suggests the following garnet-forming reaction: (3) wollastonite + anorthite =

CE P

grossular + quartz. Similarly, the absence of coexisting scapolite and wollastonite suggests another garnet-forming reaction: (4) wollastonite + scapolite = grossular +

AC

calcite + quartz.

These vapor-absent reactions (2), (3), and (4) have been commonly used to constrain the P–T conditions of calc-silicate granulites (Shaw and Arima 1996; Mathavan and Fernando 2001; Dasgupta and Pal 2005). In previous studies, numerous calculations have been made using the internally consistent dataset of Holland and Powell (1998) and the activity-composition models of Baker and Newton (1995), Engi and Wersin (1987), and Newton (1983), and all have yielded similar positions for the vapor-absent equilibria in a P–T diagram (Fig. 12). Thus, we can use the results of these earlier works to constrain the P–T conditions for the Altai granulite. 22

ACCEPTED MANUSCRIPT Our microprobe data reveal that the meionite component of scapolite is 75–83 mol% (Supplementary Table 5), and that the associated plagioclase contains 91–100

IP

T

mol% anorthite (Supplementary Table 3), whereas the composition of garnet is highly

SC R

variable from sample to sample with a range of grossular content from 35 to 82 mol% (Fig. 11a). The association plagioclase + scapolite displayed in Fig. 10 has the same phase compositions as those of association IA of Dasgupta and Pal, 2005, and those

NU

authors’ calculations led them to conclude that the temperature of the

MA

pressure-insensitive reaction (2) is ~680 °C, providing a minimum value of temperature for the Altai calc-silicate granulites (see Fig. 12).

D

The absence of coexisting wollastonite and plagioclase indicates that

TE

equilibrium (3) (Fig. 12) cannot constrain the maximum value of temperature for the

CE P

Altai calc-silicate granulites. Also the garnet + calcite + quartz + scapolite assemblage is developed locally (e.g., Fig. 7). Taking account of the fact that no coexisting

AC

scapolite and wollastonite have been identified, the highest temperature may be constrained by equilibrium (4), that is, wollastonite + scapolite = grossular + calcite + quartz. In this assemblage (Fig. 7), the grossular content of garnet is within the range 75–87 mol%, whereas the coexisting scapolite contains a meionite component of 75–82 mol%. Dasgupta and Pal (2005) made calculations for equilibrium (4) using compositions of garnet (XGrs = 0.83) and scapolite (Xme = 0.75) much the same as those in the Altai granulites. This equilibrium indicates that the temperature of the Altai granulite was lower than 780–800°C at a pressure of 6 to 7 kbar (Fig. 12) (Wang et al. 2009b). 23

ACCEPTED MANUSCRIPT 5.3. Geochemistry of the Early Permian gabbro

T

The results of the bulk geochemical analyses of five gabbro samples are listed in

IP

Supplementary Table 6. All show a low loss-on-ignition (LOI) that varied from 0.3 to

SC R

1.96 wt%. The studied gabbros have a narrow compositional range: 47.9–51.4 wt% SiO2, 14.6–17.6 wt% Al2O3, 11.7–9.0 wt% CaO, and 7.9–9.7 wt% MgO, which

NU

corresponds to a Mg# of 62–72. Their K2O + Na2O contents range from 3.1 to 4.5 wt%, the K2O/Na2O ratios are low at less than 1, and the A/CNK ratios are 0.61 to

MA

0.73.

The homogeneity in composition of the gabbros is also displayed in their

TE

D

chondrite-normalized (values after McDonough and Sun, 1995) rare earth element (REE) patterns (Fig. 13a) and primitive mantle (values after Hofmann, 1988)

CE P

normalized trace element patterns (Fig. 13b). The REE patterns of four samples are nearly the same as that of N-MORB (data from Workman and Hart, 2005), whereas

AC

the other sample is broadly similar to E-MORB. In the primitive mantle normalized trace element patterns (Fig. 13b), all elements that have a higher compatibility than La (the order of increasing compatibility of elements follows that recommended by Albarède, 2003) broadly plot in upwards curves, the same as N-MORB, while other elements show irregular patterns, especially the large-ion lithophile elements.

5.4. Zircon U/Pb results

Zircon grains from gabbro sample X0478-8 all have similar morphologies and internal structures, and all are transparent, colorless, euhedral to sub-euhedral, and up to 150 24

ACCEPTED MANUSCRIPT m in size, with a length/width ratio of 1.5 to 2.5. CL images (Fig. 14a) commonly reveal sector zoning (Watson and Liang 1995). Ten analyses were made on 10 grains

IP

T

using the SHRIMP II technique. All spots were located in zoned parts, resulting in U

SC R

and Th contents of 141–595 ppm and 68–1002 ppm, respectively, and 206Pb/238U ages of 266–278 Ma. The ages form a single and tight cluster on the concordia plot (Fig. 14b), and yield a weighted mean

206

Pb/238U age of 272.3 ± 4.1 Ma (mean square

NU

weighted deviation (MSWD) = 0.45). Except for one analysis, the U/Th ratios are less

MA

than 1.0, ranging between 0.99 and 0.50, and similar to those in magmatic zircons. We interpret the U–Pb dates as the emplacement age of the gabbro pluton.

D

Most zircons from the garnet-bearing leucogranite sample (X0522-7) of the

TE

Altai-Qinghe slice (Fig. 2) are clear, long prismatic crystals with a length/width

CE P

ratio >3.0, whereas just a few are rounded in shape and contain cores (Fig. 15a). All prismatic zircon grains have clear oscillatory zoning. Twenty-five analyses were

AC

conducted on 25 zircon grains from leucogranite sample X0522-7 (Fig. 15a), and 24 results were acceptable (the rejected one had a low concordance of 79%). Four analyses were from zircon cores, and yielded

206

Pb/238U ages of 398 to 456 Ma. The

other 20 analyses on oscillatory zoning portions gave

206

Pb/238U ages of 286 to 297

Ma, forming a single and tight cluster on the concordia plot with a weighted mean 206

Pb/238U age of 290.5 ± 1.4 Ma (MSWD = 0.34) (Fig. 15b). We interpret the 290.5 ±

1.4 Ma age as the time of emplacement of the leucogranite. Zircons from another garnet-bearing leucogranite sample (X0504-13) from the Fuyun slice (Figs. 3 and 4) have the same grain shapes as those of sample X0522-7, 25

ACCEPTED MANUSCRIPT but the CL images reveal different internal textures, and most long prismatic zircon grains from X0504-13 do not display oscillatory zoning (Fig. 16a). We made 39

IP

T

analyses on 38 zircon grains from leucogranite sample X0504-13, using the

SC R

LA-ICP-MS technique. Two analyses were of zircon cores, yielding 206Pb/238U ages of 850 and 524 Ma. Another two results are less concordant (concordance less than 95%). The remaining 35 analyses gave

206

Pb/238U ages of 269 to 300 Ma, forming two

NU

distinctive tight clusters: (1) twenty two have a weighted mean

206

Pb/238U age of

cluster, yielding a weighted mean

MA

291.3 ± 1.0 Ma (MSWD = 1.2) (Fig. 16b), and (2) thirteen form another single, tight 206

Pb/238U age of 282.0 ± 1.6 Ma (MSWD = 2.2)

D

(Fig. 16c). Field observations revealed that the leucogranite pluton (X0504-13)

TE

contains numerous calc-silicate granulite xenoliths, suggesting a younger age than the

CE P

granulite (ca. 293 Ma, Wang et al. 2009b; 295 Ma, this study). Thus we interpret the smaller weighted mean age of ca. 282 Ma as the emplacement age of the leucogranite;

AC

whereas the larger weighted mean age of ca. 291 Ma may indicate commencement of the partial melting of the crustal rocks. Most of the zircons from paragneiss sample X0509-11 of the Fuyun slice (see Fig. 3 or 4 for its location) have a core–mantle structure (Fig. 17, left). The core portions of most grains have obvious oscillatory zoning, and the cores are commonly surrounded by a rim with a thickness of up to 50 µm. Some of these rims are wide enough for LA-ICP-MS U/Pb analysis with a laser-bean diameter of 25µm. CL images have not revealed any zoning in the rims. We made 45 analyses on 45 grains. Nine analyses were located in the rim parts, and they yielded 206Pb/238U ages of 277 to 26

ACCEPTED MANUSCRIPT 295 Ma and a mean age of 284.5 ± 5.5 Ma (with large MSWD = 13). Three analyses gave mixed values due to the location of the spots in boundary positions. Two spots in Pb/206Pb ages (2606 Ma and 1817

T

207

IP

cores that lacked any zoning gave Precambrian

SC R

Ma, respectively). The other 31 analyses were located in cores that had clear zoning, and 20 of the results form a tight cluster on the concordia plot (Fig. 17, right), and yield a weighted mean 206Pb/238U age of 404.6 ± 1.8 Ma (MSWD = 1.8). Six analyses

Pb/238U age of 422.5 ± 3.0 Ma (MSWD = 1.1) These ages indicate that the protolith

MA

206

NU

form another tight cluster on the concordia plot (Fig. 17) with a weighted mean

of the paragneiss was derived from the widespread Late Silurian to Early Devonian

D

volcanics of the Chinese Altai (e.g., Wilhem et al. 2012, and references therein). Our

TE

new data reported here suggest that the Paleozoic volcanism was likely pulsed. The

CE P

meanings of the analyses with Late Paleozoic ages will be discussed further in the sub-chapter 6.1.

AC

5.5. Zircon Hf isotopes and REEs in the Early Permian gabbros

Twenty zircon grains from gabbro sample X0478-8 were analyzed for Hf isotopes and REE abundances, and 19 acceptable results were obtained. The rejected analysis shows irregular normalized REE patterns, and we suspect that mineral inclusions with high contents of LREEs, such as allanite or monazite, were responsible for interference at the spot site. The analyzed zircons have homogenous internal texture (Fig. 14a) and

206

Pb/238U ages (Fig. 14b); thus, the SHRIMP II weighted mean age

(272 Ma) of the sample was adopted to calculate zircon Hf (t) values.

27

ACCEPTED MANUSCRIPT Different zircon grains have slightly different Hf isotopic compositions, but their

176

Lu/177Hf values are always low, within the range 0.0006–0.0029. Because of

IP

T

these low Lu/Hf ratios, there are very few radiogenic ingrowths of Hf. The 176Hf/177Hf

SC R

ratios in zircons from the gabbros fall in a wide range from 0.282608 to 0.283038, and this corresponds to a range of –5.8 to +9.4 in present-day εHf values (εHf(0)), a range of 15 units. Except for one analysis with a negative value (–0.22), the initial εHf (εHf(t))

NU

values of these zircons are consistently positive, ranging from +0.05 to +15.18. The

MA

zircon Hf (t) values cluster between the CHUR and DM evolution lines, which suggests mixing of crustal material with mantle-derived magma (Fig. 18a).

D

The 19 analyzed zircon grains show relatively low REE concentrations (total

TE

REE abundances range from 316 to 2,548 ppm). Their chondrite-normalized patterns

CE P

(Fig. 18b) are characterized by steeply rising slopes from LREE to HREE, with positive Ce anomalies and negative Eu anomalies.

AC

6. Discussion

6.1. The timings of granulite facies metamorphism and emplacement of leucogranite and gabbro: A Permian geochronological framework

Geological mapping results (Xinjiang Bureau of Geology and Mineral Resources 1989) have revealed two types of early Permian pluton in the regions on both sides of the Ertix suture: gabbro and leucogranite. Numerous geochronological data have been published to constrain the emplacement times of these plutons, which, along with our new data, provide a robust basis on which to construct a Permian geochronological 28

ACCEPTED MANUSCRIPT framework. Available in situ zircon U–Pb data suggest that: (1) the magmatic stage that

IP

T

produced the leucogranite seems to have had a prolonged history, from ca. 291 Ma

SC R

(this study) via 285 Ma (Tong et al. 2006a; Briggs et al. 2007; Zhou et al. 2007; this study) to 272 Ma (Wang et al. 2005; Tong et al. 2006b; Briggs et al. 2007); and (2) the emplacement of gabbros took place during a relatively short period from ca. 270 to

NU

280 Ma (Han et al. 2004; Yang et al. 2004; Chen and Han 2006; Zhang et al. 2008b;

MA

Zhu et al. 2009; Cai et al. 2010; this study).

The dated paragneiss (ca. 10 cm thick, sample X0509-11) is conformably

D

intercalated with two layers of scapolite-bearing granulite, and both the paragneiss

TE

and granulite show the same deformation including folding (Fig. 4a and b). This

CE P

paragneiss consists of plagioclase, K-feldspar, biotite, minor diopside, and abundant garnet porphyroblasts. Migmatite veins are pervasive; some of them surround garnet

AC

porphyroblasts. Such a mineral assemblage and its field relationship with the scapolite-bearing rocks suggest that both the paragneiss and granulite are metamorphosed carbonate-bearing rocks of varying bulk chemical composition. Thus, the U–Pb analyses of the metamorphic rims of the zircons derived from the paragneiss may constrain the time of metamorphism of the scapolite-bearing granulite. Nine out of 45 spots were located in the metamorphic rims of zircons, but the results do not form a tight cluster (Fig. 17b) and the mean age (284.5 ± 5.5 Ma with large MSWD = 13) does not reflect the time of metamorphism. Taking into account that (1) the dated granulite-facies rocks are located within a slightly deformed leucogranite (Fig. 3), (2) 29

ACCEPTED MANUSCRIPT all nine analytical spots are located in the thin rims of zircons, and (3) the leucogranite pluton has a weighted mean zircon U–Pb age of 282 Ma, we suggest that the scatter in

IP

T

the results is likely to have resulted from intrazircon diffusion in a high-temperature

SC R

environment. Therefore, it is much more likely that the older ages of up to 295 Ma represent the time of metamorphism. Wang et al. (2009b) reported SHRIMP zircon U–Pb results (mean age: 292.8 ± 2.3 Ma, MSWD = 1.4) as the timing of

NU

metamorphism of a pelitic granulite 10 km northwest of our samples. Together, these

MA

data suggest that the high-temperature metamorphism took place during the earliest stage (295 to 290 Ma) of the leucogranitic magmatism, which occurred from 295 to

TE

D

275 Ma, and which was then followed by mafic magmatism at around 275 Ma.

CE P

6.2. Spatial and temporal variations in geochemistry of Permian granites

The early Permian mafic plutons on both sides of the Ertix suture seem to have a similar geochemistry. New geochemical data reported in the present study show that

AC

the early Permian mafic plutons are chemically N-MORB-like, with Mg# > 60 and positive zircon εHf (t) values, suggesting that these plutons may have crystallized from magmas derived from N-MORB-like depleted mantle. The early Permian gabbros in the western Chinese Altai, north of the Ertix suture (Cai et al. 2010), have the same geochemical features, including positive εNd(t) values, as found in the early Permian gabbros in the northern Junger, south of the Ertix suture (Yang et al. 2004; Chen and Han 2006; Zhou et al. 2009). In contrast to the mafic plutons, the granitic rocks exhibit highly variable

30

ACCEPTED MANUSCRIPT geochemical features (e.g., Wang et al. 2009a). Available data suggest that the variations in geochemistry are related to (1) their tectonic position, and (2) the times

IP

T

of their emplacement. Within the Chinese Altai, north of the Ertix suture, the

SC R

leucogranites with ages of >280 Ma are strongly peraluminous, and show high SiO2 (>70%) and Al2O3 contents, with A/CNK ≥ 1.1. These granites contain peraluminous mafic minerals (including garnet, cordierite, and epidote) and other Al-enriched

NU

minerals (including muscovite and sillimanite) that are diagnostic of this type of

MA

granite (Zhou et al. 2007; this study). The initial ɛNd values of these leucogranites are commonly negative (e.g., −0.33 to −3.53, 283 Ma; Chen and Jahn 2004; Zhou et al.

D

2007; Shen et al. 2011; Yang et al. 2011b), and the data suggest that the leucogranites

TE

likely resulted from the partial melting of supracrustal rocks. In contrast, we note that

CE P

most granites with younger ages (mostly <285 Ma) have positive initial ɛNd values, and are strongly peraluminous and/or high-K calc-alkaline I-type granites (e.g., ɛNd (t)

AC

= +4.4, Chen and Jahn 2004; 276 Ma, ɛNd (t) = +1.3 to +2.8, Wang et al. 2005; 281 Ma, ɛNd (t) = +3.49, Hu et al. 2006; 275 Ma, ɛNd (t) = +6.1, 260 Ma, ɛNd (t) = +2.7, Wang et al. 2009a), whereas a few granites have negative initial ɛNd values (for example, 267 Ma, ɛNd (t) = −2.7 to −4.4, Wang et al. 2009a). These data suggest that the Altai Permian granites, north of the Ertix suture, should be subdivided into two age groups and types: (1) At the very start of the early Permian, numerous leucogranites with negative ɛ Nd (t) values were emplaced; and (2) high-K calc-alkaline I-type granites and minor leucogranites, both with generally positive ɛNd (t) values, were emplaced from the late early Permian to the early middle Permian. 31

ACCEPTED MANUSCRIPT Conversely, all Permian granites in northern Junger, south of the Ertix suture, have positive initial Nd values in spite of their being high-K calc-alkaline I-types (e.g., ca.

IP

T

300 Ma, ɛNd (t) = +5.1 to +6.7 and ca. 270 Ma, ɛNd (t) = +5.5 to +5.9, Han et al. 1997;

SC R

286 Ma, ɛNd (t) = +6.2 to +7.2, Tong et al. 2006b; Wang et al. 2009a; 275 Ma, ɛNd (t) = +4.6, Wang et al. 2009a) or leucogranite (289.5 ± 3.6 Ma, ɛNd(t) = +6.5, Zhou et al. 2009). It is worth noting that all the granites with positive ɛNd (t) values do not

NU

contain any peraluminous mafic minerals, but biotite or muscovite are common (e.g.,

MA

Han et al. 1997, 2006; Chen and Jahn 2004; Hu et al. 2006; Tong et al. 2006b; Wang et al. 2009a).

TE

D

6.3. Petrogenesis of the early Permian granites

CE P

It is commonly argued that the peraluminous-mafic-mineral-bearing leucogranite were formed by the partial melting of supracrustal rocks (S-type granite, e.g.,

AC

Chappell et al. 1987; Bonin 2007, and references therein; Zhou et al., 2007; Wang et al. 2009a), and that the gabbroic plutons were derived from a depleted mantle source according to our new and previously published data (e.g., Han et al. 2004; Yang et al. 2004; Chen and Han 2006; Zhang et al. 2008b; Zhu et al. 2009; Cai et al. 2010). However, major controversy exists concerning the petrogenesis of granitic rocks with positive ɛNd (t) values. Two models have been proposed to explain the positive ɛNd(t) values of the Permian high-K calc-alkaline granites: (1) the partial melting of juvenile crustal rocks, such as Early Devonian arc volcanics (Chen and Jahn 2004; Han et al. 1997, 2006);

32

ACCEPTED MANUSCRIPT and (2) the mixing of mantle-derived basaltic magma with crust-sourced magma (Tong et al. 2006b). The presence of both the widespread early Permian gabbroic

IP

T

plutons and the leucogranites implies the possibility of magma mixing, an idea

SC R

supported by the observations of Zhou et al. (2009), who provided clear field and geochemical–geochronological evidence for the mixing of basaltic and felsic magmas. However, the magma-mixing model alone cannot explain the spatial and temporal

NU

differences in geochemistry between the Permian granitic rocks to the north of the

MA

Ertix suture and those to the south. Wang et al. (2009a) have pointed out that different tectonic units in northern Xinjiang have distinctive characteristics. For example, the

D

basement of the Altai–Mongolia microcontinent consists predominantly of

TE

Proterozoic continental materials and is probably a peri-Gondwana terrane (Yang et al.

CE P

2011a). In contrast, northern Junger is composed of juvenile Phanerozoic materials of the Devonian to Carboniferous intraoceanic island-arc system (Zhang et al. 2008a;

AC

Geng et al. 2009; Xu et al. 2012). Thus, the partial melting of the crust of the Altai–Mongolia microcontinent may have yielded granitic magma with negative ɛNd(t) values and Proterozoic Nd model ages, whereas partial melting of northern Junger juvenile crustal material may have produced granitic magma with positive ɛNd(t) values and younger Nd model ages..

6.4. Early Permian mantle–crust interaction in south-central Altaids: a two-stage model

Our new geochronological and geochemical data, together with previously published studies cited in this paper, allow us to build the following two-stage mantle–crust 33

ACCEPTED MANUSCRIPT interaction model to explain the temporal and spatial relationships between mantle-derived magmatism, the partial melting of the crust, and high-temperature

IP

T

metamorphism (Fig. 19). This new model provides a more reasonable interpretation of

SC R

the spatial and temporal variations in the geochemistry of Permian granites by taking into account both the partial melting of crust and magma-mixing. (1) In the first stage, just after the start of the Permian (ca. 295 Ma; Fig. 19a),

NU

asthenospheric upwelling beneath the northern Xinjiang regions, consisting of the

MA

Chinese Altai and northern Junger, gave rise to a mantle plume that resulted in basaltic magma underplating the crust of the region (e.g., Zhao et al. 2009). This

D

underplating involved the conduction of mantle-sourced heat into the lower crust,

TE

causing partial melting and high-temperature metamorphism of the crustal rocks. The

CE P

partial melting of the juvenile northern Junger crustal rocks gave rise to leucogranites with positive ɛ Nd(t) value. Meanwhile, the partial melting of the Chinese Altai

AC

continental crust resulted in leucogranites with negative ɛ Nd(t) value. Thus, the leucogranite north or south of the Ertix suture displays distinctive isotopic geochemistry.

(2) In the second stage, a growing accumulation of basaltic magma increased pressures in the magma chambers, causing diapirs of the underplated basaltic magma to rise during the late early Permian (ca. 280 to 270 Ma). Meanwhile, the partial melting of the crust continued. The diapirism resulted in numerous gabbroic plutons that are exposed in the regions on both sides of the Ertix suture; the plutons north of the suture intruded high-grade metamorphic rocks of the Altai-Mongolia 34

ACCEPTED MANUSCRIPT microcontinent, while the plutons south of the suture were emplaced in unmetamorphosed sediments of the northern Junger arc system. During the diapirism,

IP

T

mixing of the mantle-derived and crust-sourced magmas occurred, giving rise to

SC R

granitic rocks with generally positive ɛ Nd(t) values (Fig. 19b). These rocks are exposed in regions on both sides of the Ertix suture. Our model predicts a southward increase in ɛNd(t) value of the Permian granite in northern Xinjiang regions, as has

NU

been revealed by isotopic mapping results of Wang et al. (2009a).

MA

This model also implies that early Permian high-temperature metamorphism was widespread in the middle to low crust, at least in northern Xinjiang. However,

D

high-temperature metamorphic rocks are exposed exclusively in the southern margin

TE

of the Chinese Altai, north of the Ertix suture, whereas Permian gabbro and granite

entire

CE P

are exposed in the regions on both sides of the suture (see Fig. 2), even including the south-central

Altaids

(Fig.

1).

Furthermore,

the

high-temperature

AC

metamorphosed rocks of the Chinese Altai (e.g., Wei et al. 2007; Wang et al. 2009b; this study) are located in the high-grade Fuyun slice; this slice was also intensively sheared, giving rise to the huge Ertix shear zone. Giving this spatial relationship, we suggest that the exhumation of these high-grade metamorphic rocks likely resulted from shearing along the Ertix shear zone. The temporal relationship between the mylonite and the high-temperature metamorphic rocks revealed in this paper also suggests a possible close connection between the deformation and metamorphism. These findings point to the urgent need for a detailed structural study of the Ertix shear zone. 35

ACCEPTED MANUSCRIPT Another issue concerns the role of fluid activity in heat and mass exchange between the mantle and crust. High-temperature metamorphism and coeval partial

IP

T

melting of crustal rocks likely have generated some aqueous fluids due to such as

SC R

mica dehydration melting. The presences of aqueous fluid in the crust-sourced magma may have accelerated the exchange of mass and energy between the mantle and crust during the mantle-crust interaction, resulting in widespread middle to late Permian

NU

granitic rocks with positive ɛNd(t) values. Again, further detail study is required to

MA

address this issue.

D

7. Conclusions

TE

The new data and ideas presented in this paper have significant implications for our

CE P

understanding of mantle–crust interactions in general. Our conclusions can be summarized as follows:

AC

(1) High-temperature (680–800 °C at 6–7 kbar) metamorphic rocks have been discovered in the south-central Altaids in the huge Ertix ductile shear zone along the southern margin of the Altai–Mongolia microcontinent, north of the Ertix suture. Zircon geochronology of intrusive leucogranite and of metamorphic rock suggests that high-temperature metamorphism and the partial melting of crust occurred around 295 Ma. This partial melting continued to ca. 280 Ma, and resulted in leucogranites with peraluminous mafic minerals diagnostic of this type of granite. (2) The results of bulk geochemical and zircon Hf isotope analyses, combined with zircon U–Pb SHRIMP II dates, reveal the presence of a mantle-derived

36

ACCEPTED MANUSCRIPT magmatic event at ca. 272 Ma, approximately 20 Ma later than the partial melting and the high-temperature metamorphism of the crustal rocks that resulted in the extensive

IP

T

emplacement of gabbroic plutons in regions on both sides (north and south) of the

SC R

Ertix suture zone.

(3) Available geochronological and geochemical (including isotopic) data reveal that the geochemistry of the early Permian granites in northern Xinjiang is strongly

NU

influenced by their tectonic position and evolution. In the Chinese Altai, all granites

MA

older than ca. 285 Ma have negative ɛNd(t) values, whereas those younger than 285 Ma have mostly positive ɛNd(t) values; in contrast, the Permian granites in northern

D

Junger have generally positive ɛNd(t) values.

TE

(4) A two-stage mantle–crust interaction model is proposed on the basis of

CE P

previously published data and our new data. The early stage involves mainly heat exchange between the mantle and the crust: mantle-derived magma was probably

AC

underplated at the bottom of the crust, thus conducting extra heat into the crust, and causing coeval high-temperature metamorphism and the partial melting of the crustal rocks. The variability in characteristics of the melted crust rocks led to leucogranite with different ɛNd(t) values. During the later stages, mixing of the mantle-derived and crust-sourced magmas occurred, with the result that mantle-derived juvenile material, along with the additional heat, was added to the crust.

ACKNOWLEDGMENTS This study was supported financially by the State Key Research Development 37

ACCEPTED MANUSCRIPT Programs of China (973. No. 2011CB808901 and No.2007CB411306). Mrs. Hui Zhou helped to make the CL images, and Mr. Min Gao and Dr. Hang–Qiang Xie are

IP

T

acknowledged for their assistance with the zircon SHRIMP U–Pb analyses. Dr. X. D.

SC R

Liu and G. H. Guo are thanked for their help in conducting the mineral microprobe analyses. Dr. David Shelley corrected English. Professor Chunjing Wei from Peking University and three anonymous reviewers provided constructive comments and

NU

suggestions on this paper.

MA

REFERENCES

Albarède, F., 2003. Geochemistry: An Introduction. Cambridge University Press,

D

Cambridge, UK.

TE

Ao, S.J., Xiao, W.J., Han, C.M., Mao, Q.G., Zhang, J.E., 2010. Geochronology and

CE P

geochemistry of Early Permian mafic–ultramafic complexes in the Beishan area, Xinjiang, NW China: Implications for late Paleozoic tectonic evolution of the

AC

southern Altaids. Gondwana Research 18, 466–478. Baker, J., Newton, R. C., 1995. Experimentally determined activity–composition relations for Ca-rich scapolite in the system CaAl2Si2O8–NaAlSi3O8–CaCO3 at 7 kbar. American Mineralogist 80, 744–751. Blichert-Toft, J., Albarede, F., 1997. The Lu–Hf isotope geochemistry of chondrites and the evolution of the mantle–crust system. Earth and Planetary Sciences Letter 148, 243–258. Bonin, B. 2007. A-type granites and related rocks: Evolution of a concept, problems and prospects. Lithos 97, 1–29. 38

ACCEPTED MANUSCRIPT Briggs, S., Yin, A., Manning, C.E., Chen, Z-L., Wang, X-F., Grove, M., 2007. Late Paleozoic tectonic history of the Ertix Fault in the Chinese Altai and its

IP

T

implications for the development of the Central Asian Orogenic System.

SC R

Geological Society of American Bulletin 119, 944–960.

Buick, I. S., Harley, S. L., Cartwright, I. C., 1993. Granulite facies metasomatism: zoned calc-silicate boudins from the Rauer Group, East Antarctica. Contribution

NU

to Mineralogy and Petrology 113, 557–571.

MA

Cai, K-D., Sun, M., Yuan, C., Zhao, G.C., Xiao, W.J., Long, X.P., Wu, F.Y., 2010. Geochronological and geochemical study of mafic dykes from the northwest

TE

Research 18, 638–652.

D

Chinese Altai: Implications for petrogenesis and tectonic evolution. Gondwana

CE P

Chappell, B.W., White, A.J.R., Wyborn, D., 1987. The importance of residual source material (restite) in granite petrogenesis. Journal of Petrology 28, 1111–1138.

AC

Chen, B. and Jahn, B.M., 2004. Genesis of post-collisional granitoids and basement nature of the Junggar Terrane, NW China: Nd–Sr isotope and trace element evidence. Journal of Asian Earth Sciences 23, 691–703. Chen, B.H., Luo, Z.H., Jia, B.H., Liu, W., Wei, Y., Han, B.F., 2007. SHRIMP U-Pb zircon geochronology of igneous rocks from southern margin of the Alataw mountains, Xinjiang, China. Acta Petrologica Sinica 23, 1756–1764 (In Chinese with English abstract). Chen, H.L., Han, B. F., 2006. Geochronology, geochemistry and Sr–Nd–Pb isotopic composition of mafic intrusive rocks in Wuqiagou area, north Xinjiang: 39

ACCEPTED MANUSCRIPT constraints for mantle sources and deep process. Acta Petrologica Sinica 22, 1201–1214 (In Chinese with English abstract).

IP

T

Chen, J.F., Han, B.F., Ji, J.Q., Zhang, L., Xu, Z., He, G.Q., Wang, T., 2010. Zircon

SC R

U–Pb ages and tectonic implications of Paleozoic plutons in northern West Junggar, North Xinjiang, China. Lithos 115, 137–152.

Dasgupta, S., Pal, S., 2005. Origin of Grandite Garnet in Calc-Silicate Granulites:

NU

Mineral–Fluid Equilibria and Petrogenetic Grids. Journal of Petrology 46,

MA

1045–1076.

Deer, W.A., Howie, R.A., Zussman, J., 1996. An introduction to the rock-forming

D

minerals (2nd edition). Addison Wesley Longman limited, Harlow.

TE

Dong, Y.P., Zhang, G..W., Neubauer, F., Liu, X.M., Hauzenberger, C., Zhou, D.W., Li,

CE P

W., 2011. Syn- and post-collisional granitoids in the Central Tianshan orogen: Geochemistry, geochronology and implications for tectonic evolution.

AC

Gondwana Research 20, 568–581. Droop, G.T.R., 1987. A general equation for estimating Fe3+ concentrations in ferromagnesian silicates and oxides from microprobe analyses using stoichiometric criteria. Mineralogical Magazine 51, 431–435. Engi, M., Wersin, P., 1987. Derivation and application of a solution model for calcic garnet. Schweizerische Mineralogische und Petrographische Mitteilunger 67, 53–73. Gao, J., Klemd, R., 2000. Eclogite Occurrences in the Southern Tianshan High-Pressure Belt, Xinjiang, Western China. Gondwana Research 3, 33–38. 40

ACCEPTED MANUSCRIPT Geng, H.Y., Sun, M., Yuan, C., Xiao, W.J., Xian, W.S., Zhao, G.C., Zhang, L.F., Wong, K.N., Wu, F.Y., 2009. Geochemical, Sr–Nd and zircon U–Pb–Hf isotopic studies

IP

T

of Late Carboniferous magmatism in the West Junggar, Xinjiang: Implications

SC R

for ridge subduction? Chemical Geology 266, 364–389.

Gong, H.L., Chen, Z.L., Hu, Y.Q., Li, L., Lai, X.R. Hu, B., Zhang, W.G., 2007. Geochemical characteristics of acidic dike swarm from the eastern segment of

NU

the Ertix tectonic belt, Altai orogeny and its geological implications. Acta

MA

Petrologica Sinica 23, 889–899 (In Chinese with English abstract). Griffin, W.L., Pearson, N.J., Belousova, E., Jackson, S.E., O'Reilly, S.Y., van

D

Achterberg, E., Shee, S.R., 2000. The Hf-isotope composition of cratonic mantle:

TE

LAM–MC–ICPMS analysis of zircon megacrysts in kimberlites. Geochemica

CE P

Cosmochimica Acta 64, 133–147. Han, B.F., Wang, S.G., Jahn, B.M., Hong, D.W., Kagami, H., Sun, Y.L., 1997.

AC

Depleted-mantle source for the Ulungur River A-type granites from North Xinjiang, China: geochemistry and Nd-Sr isotopic evidence, and implications for Phanerozoic crustal growth. Chemical Geology 138, 135–159. Han, B.F., Ji, J.Q., Song, B., Chen, L.H., Li, Z.H., 2004. SHRIMP zircon U/Pb ages of Kalatongke and Huangshan Cu–Ni bearing gabbros, Xinjiang, and their geological implications. China Science Bulletin 49, 2324–2328. Han, B.F., Ji, J.Q., Song, B., Chen, L.H., Zhang, L., 2006. Late Paleozoic vertical growth of continental crust around the Junggar Basin, Xinjiang, China (Part I): timing of post-collisional plutonism. Acta Petrologica Sinica 22, 1077–1086. 41

ACCEPTED MANUSCRIPT Han, C.M., Xiao, W.J., Zhao, G.C., Ao, S.J., Zhang, J.E., Qu, W.J., Du, A.D., 2010. In-situ U–Pb, Hf and Re–Os isotopic analyses of the Xiangshan Ni–Cu–Co

IP

T

deposit in Eastern Tianshan (Xinjiang), Central Asia Orogenic Belt: Constraints

SC R

on the timing and genesis of the mineralization. Lithos 120, 547–562. Harley, S.L., Buick, I.S., 1992. Wollastonite–scapolite assemblages as indicators of

Journal of Petrology 33, 693–728.

NU

granulite pressure–temperature-fluid histories: the Rauer Group, East Antarctica.

MA

Hofmann, A.W., 1988. Chemical differentiation of the Earth: the relationship between mantle, continental crust, and oceanic crust. Earth and Planetary Science Letters

D

90, 297–314.

TE

Holland, T.J.B., Powell, R., 1998. An internally consistent thermodynamic data set for

CE P

phases of petrological interest. Journal of Metamorphic Geology 16, 309–343. Hou, K.J, Li, Y.H., Tian, Y.Y., 2009. In situ U–Pb zircon dating using laser

AC

ablation-multi ion counting-ICP-MS. Mineral Deposit 28, 481–492 (in Chinese with English abstract). Hu, A.Q., Wei, G.J., Deng, W.F., Chen, L.L., 2006. SHRIMP zircon U-Pb dating and its significance for gneiss from southwest area to Qinghe County in the Altau, China. Acta Petrologica Sinica 22(1), 1–10 (in Chinese with English abstract). Konopelko, D., Biske, G., Seltmann, R., Eklund, O., Belyatsky, B., 2007. Hercynian post-collisional A-type granites of the Kokshaal Range, Southern Tien Shan, Kyrgyzstan. Lithos 97, 140–160. Kröner, A., Kovach, V., Belousova, E., Hegner, E., Armstrong, R., Dolgopolova, A., 42

ACCEPTED MANUSCRIPT Seltmann, R., Alexeiev, D.V., Hoffmann, J.E., Wong, J., Sun, M., Cai, K., Wang, T., Tong, Y., Wilde, S.A., Degtyarev, K.E., Rytsk, E., 2014. Reassessment of

IP

T

continental growth during the accretionary history of the Central Asian Orogenic

SC R

Belt. Gondwana Research 25, 103–125.

Jahn, B.M., Wu, F. Y., Chen, B., 2000. Granitoids of the Central Asian Orogenic Belt and continental growth in the Phanerozoic. Transactions of Royal Society,

NU

Edinburg, Earth Sciences 91, 181–193.

MA

Laurent-Charvet, S, Charvet, J, Monie, P, Shu, L., 2003. Late Paleozoic strike-slip shear zones in eastern central Asia (NW China): New structural and

D

geochronological data. Tectonics 22, 4-1–4-24.

TE

Liu, Y. S., Hu, Z. C., Gao, S., Günther, D., Xu, J., Gao, C. G., Chen, H. H., 2008. In

CE P

situ analysis of major and trace elements of anhydrous minerals by LA-ICP-MS without applying an internal standard. Chemical Geology 257, 34–43.

AC

Ludwig, KR., 2003. Isoplot/EX, rev. 3.00, A Geochronological Toolkit for Microsoft Excel. Berkeley Geochronology Center Special Publication 4, 71. Luo, J.H., Che, Z.C., Gao, Y.Z., Zhang, J.Y., 2008. Geochemical and geochronological characteristics and its tectonic significance of Early Permian acid volcanic rocks of Xiaotiekelike Formation in the southern margin of south Tianshan orogen, NW China. Acta Petrologica Sinica 24, 2281–2288. Mao, Q.G., Xiao, W.J., Han, C.M., Yuan, C., Sun, M., 2008. Late Paleozoic southward accretionary polarity of the eastern Junggar orogenic belt: insight from the Dajiashan and other A-type granite. Acta Petrologica Sinica 24, 733–742. 43

ACCEPTED MANUSCRIPT Mathavan, V., Fernando, G.W.A.R., 2001. Reactions and textures in grossular– wollastonite–scapolite calc–silicate granulites from Maligawila, Sri Lanka:

IP

T

evidence for high-temperature isobaric cooling in the meta-sediments of the

SC R

Highland Complex. Lithos 59, 217–232.

McDonough, W.F, Sun, S-S., 1995. The composition of the Earth. Chemical Geology 120, 223–253.

NU

Ramsay, C.R., Davidson, L. R., 1970. The Origin of Scapolite in the Regionally

MA

Metamorphosed Rocks of Mary Kathleen, Queensland, Australia. Contributions to Mineralogy Petrology 25, 41–51.

TE

version 2.0.

D

Richard, L. R., 1994. Minpet: a mineralogical and petrological data processing system,

CE P

Satish-Kuma, M., Harley, S.L., 1998. Reaction textures in scapolite– wollastonite– grossular calc-silicate rock from the Kerala Khondalite Belt, Southern India:

AC

evidence for high-temperature metamorphism and initial cooling. Lithos 44, 83–99.

Scherer, E., Munker, C., Mezger, K., 2001. Calibration of the Lutetium–Hafnium clock. Science 293, 683–687. Sengör, A.C., Natal’in, B.A., Burtman, V.S., 1993. Evolution of the Altaid tectonic collage and Palaeozoic crustal growth in Eurasia. Nature 364, 299–306. Shaw, R. K., Arima, M., 1996. High temperature metamorphic imprint from calc-granulites of Raygada, Eastern Ghats, India: implication of isobaric cooling path. Contribution to Mineralogy and Petrology 126, 169–180. 44

ACCEPTED MANUSCRIPT Shivaprakash, C. 1981. Petrology of calc-silicate rocks from Koduru, Andhra Pradesh, India. Contribution to Mineralogy and Petrology 77, 121–128.

IP

T

Shen, X.M., Zhang, H.X., Wang, Q., Wymanc, D.A., Yang, Y.H., 2011. Late

SC R

Devonian–Early Permian A-type granites in the southern Altay Range, Northwest China: Petrogenesis and implications for tectonic setting of ‘‘A2-type’’granites. Journal of Asian Earth Sciences 42, 986–1007.

NU

Song, B., Zhang, Y-H., Wang, Y-S., 2002. Mount making and procedure of the

MA

SHRIMP dating. Geological Review 48, 26–30 (in Chinese with English abstract).

D

Su, B.X., Qin, K.Z., Sakyi, P.A., Liu, P.P., Tang, D.M., Malaviarachchi, H., Xiao,

TE

Q.H., Sun, H., Dai, Y.C., Hu, Y., 2011. Geochemistry and geochronology of

CE P

acidic rocks in the Beishan region, NW China: Petrogenesis and tectonic implications. Journal of Asian Earth Sciences 41, 31–43.

AC

Sun, M., Yuan, C., Xiao, W-J., Long, X-P., Xia , X-P., Zhao, G-C., Lin, S-F., Wu, F-Y., Kröner, A., 2008a. Zircon U–Pb and Hf isotopic study of gneissic rocks from the Chinese Altai: Progressive accretionary history in the early to middle Paleozoic. Chemical Geology 247, 352–383. Sun, L.H., Wang, Y.J., Fan, W.M., Zi, J.W., 2008b. Post-collisional potassic magmatism in the Southern Awulale Mountain, western Tianshan Orogen: Petrogenetic and tectonic implications. Gondwana Research 14, 383–394. Tang, G.J., Wang, Q., Wyman, D.A., Sun, M., Li, Z.X., Zhao, Z.H., Sun, W.D., Jia, X.H., Jiang, Z.Q., 2010. Geochronology and geochemistry of Late Paleozoic 45

ACCEPTED MANUSCRIPT magmatic rocks in the Lamasu–Dabate area, northwestern Tianshan (west China): Evidence for a tectonic transition from arc to post-collisional setting.

IP

T

Lithos 119, 393–411.

SC R

Tang, H.F., Qu, W.J., Su, Y.P. Hou, G.S., Du, A.D., Cong, F., 2007a. Genetic connection of Sareshike tin deposit with the alkaline A-Type granites of Sabei body in Xinjiang: constraint from isotope ages. Acta Petrologica Sinica 23,

NU

1989–1997 (in Chinese with English abstract).

MA

Tang, J.H., Gu, L.X., Zhang, Z.Z, Wu, C.Z., Wang, C.S., Li, S.H., Li, G.R., 2008. Peraluminous granite in Huangshan-Jingerquan area of eastern Tianshan:

TE

921–946.

D

Geochemistry, Mineralogy and geochronology. Acta Petrologica Sinica 24,

CE P

Tang, L.G., Zhou, T.F., Yuan, F., Fan, Y., Yue, S.C., 2007b. 40Ar/39Ar age of the Kalagang Formation volcanic rocks in western Junggar, Xinjiang. Chinese

AC

Journal of Geology 42, 579–586 (in Chinese with English abstract). The Team One of Geological Survey of Xinjiang, 1979. Four geological maps of Altai, with geological report, scale 1/200,000. Tong, Y., Hong, D.W., Wang, T., Wang, S.G., Han, B.F., 2006a. TIMS U–Pb zircon ages of Fuyun post-orogenic linear granite pluton on the southern margin of Altay orogenic belt and their implications. Acta Petrologica Et Mineralogica 25, 85–89 (in Chinese with English abstract). Tong, Y., Wang, T., Kovach, V.P., Hong, D.W., Han, B. F. 2006b. Age and origin of the Takkeshiken postorogenic alkali-riched intrusive rocks in southern Altai, 46

ACCEPTED MANUSCRIPT near the Mongolia border in China and its implications for continental growth. Acta Petrologica Sinica 22, 1267–1278 (in Chinese with English abstract).

IP

T

Wang, B., Su, L.S., Cluzel, D., Faure, M., Charvet, J., 2007. Geochronological and

SC R

geochemical studies on the Borohoro plutons, north of Yili, NW Tianshan and their tectonic implications. Acta Petrologica Sinica 23, 1885–1900. Wang, T., Hong, D-W., Tong, Y., Han, B-F., Shi, Y-R., 2005. Zircon U/Pb SHRIMP

NU

age and origin of post-orogenic Lamazhao granitic pluton from Altai orogen: its

MA

implications for vertical continental growth. Acta Petrologica Sinica 21, 640–650 (in Chinese with English abstract).

D

Wang, T., Hong, D.W., Jahn, B-M., Tong, Y., Wang, Y.B., Han, B.F., Wang, X.X.,

TE

2006. Timing, Petrogenesis, and Setting of Paleozoic Synorogenic Intrusions

CE P

from the Altai Mountains, Northwest China: Implications for the Tectonic Evolution of an Accretionary Orogen. Journal of Geology 114, 735–751.

AC

Wang, T., Jahn, B-M., Kovach, V. P., Tong, Y., Hong, D-W., Han, B-F., 2009a. Nd–Sr isotopic mapping of the Chinese Altai and implications for continental growth in the Central Asian Orogenic Belt. Lithos 110, 359–372. Wang, W., Wei, C-J., Wang, T., Lou, Y-X., Chu, H. 2009b. Confirmation of pelitic granulite in the Altai orogen and its geological significance. Chinese Science Bulletin 54, 918–923. Watson, E.B, Liang, Y., 1995. A simple model for sector zoning in slowly grown crystals: Implications for growth rate and lattice diffusion, with emphasis on accessory minerals in crustal rocks. American Mineralogist 80, 1179–1187. 47

ACCEPTED MANUSCRIPT Wei, C.J., Clarke, G., Tian, W., Qiu, L., 2007. Transition of metamorphic series from the Kyanite- to andalusite-types in the Altai orogen, Xinjiang, China: Evidence

IP

T

from petrography and calculated KMnFMASH and KFMASH phase relations.

SC R

Lithos 96, 353–374.

Wilhem, C., Windley, B. F., Stampfli, G. M., 2012. The Altaids of Central Asia: A tectonic and evolutionary innovative review. Earth Science Reviews 113,

NU

303–341.

MA

Williams, I.S., Claesson, S., 1987. Isotope evidence for the Precambrian province and Caledonian metamorphism of high-grade paragneiss from the Seve Nappes,

D

Scandinavian Caledonides, II. Ion microprobe zircon U–Th–Pb. Contribution to

TE

Mineralogy and Petrology 97, 205–217.

CE P

Windley, B.F., Kroner, A., Guo, J.H., Qu, G.S., Li, Y.Y., Zhang, C., 2002. Neoproterozoic to Paleozoic Geology of the Altai Orogen, NW China: New

AC

Zircon Age Data and Tectonic Evolution. Journal of Geology 110, 719–737. Windley, B. F, Alexeiev, D. V, Xiao, W-J, Kröner, A., Badarch, G., 2007. Tectonic models for accretion of the Central Asian Orogenic Belt. Journal of Geological Society, London 164, 31–47. Workman, R.K., Hart, S.R., 2005. Major and trace element composition of the depleted MORB mantle DMM). Earth and Planetary Science Letter 231, 53–72. Xiao, W.J., Windley, B.F., Badarch, G., Sun, S., Li, J., Qin, K.Z., Wang, Z.H. 2004a. Palaeozoic accretionary and convergent tectonics of the southern Altaids: implications for the lateral growth of Central Asia. Journal of Geological Society, 48

ACCEPTED MANUSCRIPT London 161, 339–342. Xiao, W.J., Zhang, L.-C., Qin, K.-Z., Sun, S., and Li, J.-L. 2004b. Paleozoic

IP

T

accretionary and collisional tectonics of the Eastern Tianshan (China):

SC R

Implications for the continental growth of central Asia. American Journal of Science 304, 370–395.

Xiao, W.J, Windley B.F., Yuan, C., Sun, M., Han, C.M., Lin, S.F., Chen, H.L., Yan,

NU

Q.R., Liu, D.Y., Qin, K.Z., Li, J.L., Sun, S., 2009. Paleozoic multiple

MA

subduction-accretion processes of the southern Altaids. American Journal of Science 309, 221–270.

D

Xiao, W.J., Huang, B.C., Han, C.M., Sun, S., Li, J.L., 2010. A review of the western

TE

part of the Altaids: A key to understanding the architecture of accretionary

CE P

orogens. Gondwana Research 18, 253–273. Xiao, W.J., Windley, B. F., Allen, M. B., Han, C. M., 2013. Paleozoic multiple

AC

accretionary and collisional tectonics of the Chinese Tianshan orogenic collage. Gondwana Research 23, 1316–1341. Xiao, W. J., Santosh, M., 2014. The western Central Asian Orogenic Belt: A window to accretionary orogenesis and continental growth. Gondwana Research 25, 1429–1444. Xiao, Y., Zhang, H.F., Shi, J.A., Su, B.X., Sakyi, P.A., Lu, X.C. Hu, Y. & Zhang, Z., 2011. Late Paleozoic magmatic record of East Junggar, NW China and its significance: Implication from zircon U–Pb dating and Hf isotope, Gondwana Research 20, 532–542. 49

ACCEPTED MANUSCRIPT Xinjiang Bureau of Geology and Mineral Resources, 1989. Regional Geology of Xinjiang Province. Geological Publishing House, Beijing (in Chinese with

IP

T

English abstract).

SC R

Xu, J.F., Castillo, P.R., Chen, F.R., Niu, H.C., Yu, X.Y., Zhen, Z.P., 2002. Geochemistry of late Paleozoic mafic igneous rocks from the Kuerti area, Xinjiang, orthwest China: implications for backarc mantle evolution. Chemical

NU

Geology 193, 137–154.

MA

Xu, Z., Han, B-F., Ren, R., Zhou, Y-Z, Zhang, L., Chen, J-F, Su, L., Li, X-H, Liu, D-Y., 2012. Ultramafic–mafic mélange, island arc and post-collisional intrusions

D

in the Mayile Mountain, West Junggar, China: Implications for Paleozoic

TE

intra-oceanic subduction–accretion process. Lithos 132–133, 141–161.

CE P

Yang, S.F. Li, Z.L., Chen, H.L., Xiao, W.J., Yu, X., Lin, X.B., Shi, X.G., 2006. Discovery of a Permian quartz syenitic porphyritic dyke from the Tarim basin

AC

and its tectonic implications. Acta Petrologica Sinica 22, 1405–1412 (in Chinese with English abstract). Yang, T.N., Li, J.Y., Zhang, J., Hou, K.J. 2011a. The Altai–Mongolia terrane in the Central Asian Orogenic Belt (CAOB): A peri-Gondwana one? Evidence from zircon U–Pb, Hf isotopes and REE abundance. Precambrian Research 187, 79–98. Yang, G. X., Li, Y.J., Wu, H.G., Zhong, X., Yang, B.K., Yan, C.X., Yan, J., Si, G. H., 2011b. Geochronological and geochemical constrains on Petrogenesis of the Huangyangshan A-type granite from the East Junggar, Xinjiang, NW China. 50

ACCEPTED MANUSCRIPT Journal of Asian Earth Science 40, 722–736. Yang, W.P., Zhou, G., Zhang, Z.C., Yan, S.H., He, L.X., He, Y.S., Chen, B.L., 2004.

IP

T

Geochemistry of mafic intrusions on the south margin of the Altay orogenic belt

SC R

and Cu–Ni sulfide ore potential of these intrusions. Geological Bulletin of China 23, 390–399 (in Chinese with English abstract).

Yuan, C., Sun, M., Xiao, W.J., Li, X.H., Chen, H.L., Lin, S.F., Xia, X.P., Long, X.P.,

NU

2007. Accretionary orogenesis of the Chinese Altai: Insights from Paleozoic

MA

granitoids. Chemical Geology 242, 22–39.

Yuan, C., Sun, M., Wilde, S., Xiao, W.J., Xu, Y.G., Long, X.P., Zhao, G.C., 2010.

D

Post-collisional plutons in the Balikun area, East Chinese Tianshan: Evolving

TE

magmatism in response to extension and slab break-off. Lithos 119, 269–288.

CE P

Yuan, H.L., Gao, S., Dai, M.N., Zong, C.L., Gunther, D., Fontaine, G.H., Liu, X.M., Diwu, C.R., 2008. Simultaneous determinations of U/Pb age, Hf isotopes and

AC

trace elements compositions of zircon by excimer laser-ablation quadrupole and multiple-collector ICP-MS. Chemical Geology 247, 100–118. Zhang, C.L., Xu, Y.G., Li, Z.X., Wang, H.YH., Ye, H.M., 2010a. Diverse Permian magmatism in the Tarim Block, NW China: Genetically linked to the Permian Tarim mantle plume? Lithos 119, 537–552. Zhang, Z-C., Mao, J-W., Cai, J-H., Zhou, G., Yan, S-H., Kusky, T.M., 2008a. Geochemistry of picrites and associated lavas of a Devonian island arc in the Northern Junggar terrane, Xinjiang (NW China): implications for petrogenesis, arc mantle sources and tectonic setting. Lithos 105, 379–395. 51

ACCEPTED MANUSCRIPT Zhang, Z.H., Mao, J.W., Du, A.D., Pirajno, F., Wang, Z.L., Chai, F.M., Zhang Z.C., Yang, J.M., 2008b. Re–Os dating of two Cu–Ni sulfide deposits in northern

IP

T

Xinjiang, NW China and its geological significance. Journal of Asian Earth

SC R

Science 32, 204–217.

Zhang, Y.Y., Dostal, J., Zhao, Z.H., Liu, C., Guo, Z.J., 2011. Geochronology, geochemistry and petrogenesis of mafic and ultramafic rocks from southern

NU

Beishan area, NW China: Implications for crust–mantle interaction. Gondwana

MA

Research 20, 816–830.

Zhao, Z.H., Xiong, X.L., Wang, Q., Bai, Z.H., Qiao, Y. L., 2009. Late Paleozoic

D

underplating in North Xinjiang: Evidence from Shoshonitic Series Volcanic

TE

Rocks and Adakite. Gondwana Research 16, 216–226.

CE P

Zhou, D.W., Liu, Y.Q., Xing, X.J., Hao, J.R., Dong, Y.P., Ouyang, Z.J., 2006. Formation of the Permian basalts and implications of geochemical tracing for

AC

paleo-tectonic setting and regional tectonic background in the Turpan-Hami and Santanghu basins, Xinjiang. Science in China: Series D Earth Sciences 49, 584–596.

Zhou, G., Zhang, Z.C., Luo, S.B., He, B., Wang ,X., Ying, L.J., Zhao, H., Li, A.H., He, Y. K., 2007. Confirmation of high-temperature strongly peraluminous Mayin’ebo granite in the southern margin of Altay, Xinjiang: age, geochemistry and tectonic implications. Acta Petrologica Sinica 23, 1909–1920 (in Chinese with English abstract). Zhou, G., Zhang, Z.C., Wu, G.G., Dong, L.H., He, Y. K., Dong, Y.G., He, L.X., Qin, 52

ACCEPTED MANUSCRIPT J.H., Zhao, Z. H., Liu, G.R., 2009. Postorogenic extension and continental growth of the northeastern margin of the Junger: Evidence from petrography and

SC R

331–346 (in Chinese with English abstract).

IP

T

geochemistry of the Hadansun intrusive complex. Acta Geologica Sinica 83,

Zhou, T.F., Yuan, F., Fan, Y., Zhang, D.Y., Cooke, D., Zhao G.C., 2008. Granites in the Sawuer region of the west Junggar, Xinjiang Province, China: Geochronological

NU

and geochemical characteristics and their geodynamic significance. Lithos 106,

MA

191–206.

Zhu, Y.F., Xu, X., 2009. Lithology and zircon SHRIMP geochronology of the

D

Biestuobie gabbro in Tacheng, Xinjiang. Acta Geologica Sinica 83, 1316–1326

TE

(in Chinese with English abstract).

CE P

Zhu, Z.X., Li, J.Y., Dong, L.H., Zhang, X.F., Hu, J.W., Wang, K.Z., 2008. The age determination of Late Carboniferous intrusions in Manqusi region and its

AC

constraints to the closure of oceanic basin in South Tianshan, Xinjiang. Acta Petrologica Sinica 24, 2761–2766 (in Chinese with English abstract). Zhuang, Y.X., 1994. Tectonothermal evolution in space and time and orogenic process of Altaide, China. Jilin Sciences and Technology Press, Changchun, China (in Chinese with English abstract).

53

ACCEPTED MANUSCRIPT Figure captions Figure 1. Generalized map of Xinjiang Province, NW China, showing the major

IP

T

basins and their periphery orogenic belts. The Latest Late Carboniferous to

SC R

Early Permian magmatic rocks are widespread in all tectonic units of Xinjiang Province. Inset shows the location of Fig. 1 in the Altaids. Position of Fig. 2 is marked. Abbreviations: W. Junggar: West Junggar; E. Junggar: East Junggar; C.

NU

Tianshan: Central Tianshan; E. Tianshan: East Tianshan; N. Tianshan: North

MA

Tianshan; S. Tianshan: South Tianshan. Age data from: (1) to (3) Tang et al. (2007b); (4) Zhou et al. (2008); (5) Zhang et al. (2008b); (6) Han et al. (2006);

D

(7) Geng et al. (2009); (8) & (9) Chen et al. (2007); (10) Tang et al. (2010); (11)

TE

& (12) Wang et al. (2007); (13) Chen & Jahn (2004); (14) Sun et al. (2008b);

CE P

(15) Luo et al. (2008); (16) Konopelk et al. (2007); (18) to (20), (23) & (24) Zhang et al. (2010a); (21) Yang et al. (2006); (22) Wang et al. (2007); (25) Zhu

AC

et al. (2008); (26) Dong et al. (2011); (27) to (29), (40), (42), & (44) Zhou et al. (2006); (30) Yuan et al (2010); (31) Su et al. (2011); (32) Ao et al. (2010); (33), (41) & (43) Yuan et al. (2010); (34) Chen and Han (2006); (35) Gong et al. (2007); (36) Zhang et al. (2008b); (37) Tang et al. (2007a); (38) Shen et al. (2011); (39) Mao et al. (2008); (45) Han et al. (2010); (46) Zhang et al. (2008b); (47) Tang et al. (2008); (48) Zhang et al. (2011); (49) Yang et al. (2011b); (50) Yuan et al. (2007); (51) Xiao et al. (2011). Figure 2. Geological map of part of Northern Xinjiang, NW China (modified after 4 geological maps with scale 1/200,000, The Team One of Geology of Xinjiang 54

ACCEPTED MANUSCRIPT Province 1976) showing the main tectonic units, previously published zircon

IP

(X0522-7) samples. The position of Fig. 3 is marked.

T

U–Pb data, and the locations of the gabbro (X0478-8) and leucogranite

SC R

Figure 3. Geological map of the region around Altai city, showing the distribution of the calc-silicate granulite. The numerals started by “X” and related green circles show the locations of field studies and sampling spots.

NU

Figure 4. (a) A type outcrop of calc-silicate granulite in the Chinese Altai (field study

MA

spot No. X0509, location shown on Fig. 3); the locations of most samples for petrographic study and microprobe analyses are marked. (b) Represents an

D

enlarged portion of Fig. (a) showing a refolded reversal anticline of

TE

metamorphosed calc-silicate layers. Also shown is the location of the paragneiss

CE P

sample (X0509-11) for zircon geochronology. (c) Synclinal layers of metamorphosed calc-silicate rocks; Anorthosite-vein is identified as leuco-vein

AC

in Fig. (c), the cutting-cross relationship between the anorthosite vein and its host rocks is very clear (d). Figure 5. Microphotograph (plane polarized light, PPL) that shows the microtexture and mineral assemblage of the wollastonite (Wo), garnet (Grt), and diopside (Di)-bearing quartzite. Sample X0509-4 (for location, see Fig. 4). Figure 6. Microphotograph (PPL) that shows the microtexture and mineral assemblage of the layered garnetite, which is intercalated with layers consisting mainly of wollastonite (Wo); sample X0509-8 (for location, see Fig. 4). Figure 7. (a) A mosaic of microphotographs (PPL) of a garnetite showing both 55

ACCEPTED MANUSCRIPT poikiloblastic and symplectite textures. The poikiloblastic texture is reflected by irregularly shaped inclusions of calcite (Cc), scapolite (Scap), and minor titanite

IP

T

(Tnt) in garnet (b); (c) is a BSE image displaying the detailed mineral

SC R

composition and texture of symplectitic intergrowth of garnet with plagioclase, calcite, and quartz and minor ziosite (Zi); the shorter arrows indicate tiny vermicular calcite in the plagioclase matrix. Sample X0509-13 (for location, see

NU

Fig. 4b).

MA

Figure 8. (a) A mosaic of microphotographs (PPL) of a clinopyroxene-rich calc-silicate rock consisting of porphyroblastic diopside grains with highly

D

irregular boundaries set in a matrix of anorthite + garnet + minor quartz. The

TE

composite matrix is symplectitic, involving vermicular garnet set in a

CE P

plagioclase matrix (b, BSE image). Sample X0503-10 (for location, see Fig. 3). Figure 9. A mosaic of microphotographs (PPL) (a) and a BSE image (b) of a

AC

scapolite-rich layer composed of diopside and garnet set in a scapolite matrix. Sample X0509-10, its location is marked in Fig. 4. Figure 10. (a) Mosaic of microphotograph (PPL) of a layered anorthite-rich calc-silicate rock comprising alternating garnet-rich and anorthite-rich layers. Euhedral diopside (Di) is locally present in the garnet-rich (Grt) layers as porphyroblastic crystals. (b) Microphotograph (crossed polarized light) of the anorthite (An)-rich layer showing euhedral anorthite and interstitial scapolite (Scap) and garnet (Grt). Sample X0509-14, location shown in Fig. 4. Figure 11. (a) Pyrope + Almandine + Spessartine–Grossular–Andradite triangular 56

ACCEPTED MANUSCRIPT diagram illustrating the compositional variation of all analyzed garnet in the calc-silicate granulite of the Chinese Altai; (b) Wollastonite (Wo)–Enstatite

IP

T

(En)–Ferrosilite (Fs) triangular diagram reveals that the composition of all

SC R

analyzed clinopyroxene in the calc-silicate granulite of the Chinese Altai is essentially a diopside–hedenbergite solid solution.

Figure 12. P–T diagram showing the positions of the vapor-absent equilibria adjusted

NU

for respective phase compositions in their different textural modes (modified

MA

after Dasgupta and Pal 2005). The mineral phases labeled in gray letters are absent in the calc-silicate granulite of the Chinese Altai. For the

D

pressure-insensitive equilibrium 2, the absent phase is located on the

TE

lower-temperature side of its univariant line, which constrains the minimum

CE P

temperature. For other equilibria, especially equilibrium 4, the absent phases are generally located on the higher-temperature side of their univariant lines; as

AC

such, they are used to constrain the maximum temperature. At a pressure of 6 to 7 kbar (Wang et al. 2009b), the metamorphic temperature of the calc-silicate granulite of the Chinese Altai lies in the range of 680 to 800°C. Figure 13. Chondrite-normalized REE patterns (a) and primitive mantle-normalized trace element patterns (b) for the Early Permian gabbros of the Chinese Altai. Figure 14. (a) Cathodoluminescence images of analyzed zircons from gabbro sample X0478-7, where numbers, locations, and resultant

206

Pb/238U ages of data points

are labeled. (b) U–Pb concordia diagrams for zircons from the gabbro sample X0478-8 (SHRIMP II results). Location shown in Fig. 2. 57

ACCEPTED MANUSCRIPT Figure 15. (a) Cathodoluminescence images of analyzed zircons from leucogranite sample X0522-7, where numbers, locations, and resultant 206Pb/238U ages of data

IP

T

points are labeled. (b) U–Pb concordia diagrams for zircons from the

SC R

leucogranite sample (LA-ICP-MS results). Location shown in Fig. 2. Figure 16. (a) Cathodoluminescence images of analyzed zircons from leucogranite sample X0504-13, where numbers, locations, and resultant

206

Pb/238U ages of

NU

data points are labeled. (b) U–Pb concordia diagrams for zircons from the

MA

leucogranite sample (LA-ICP-MS results). (c) Weighted mean age of the younger cluster of analyses. Location shown in Figs. 2 and 3.

D

Figure 17. Left: cathodoluminescence images of analyzed zircons from paragneiss

TE

sample X0509-11, where numbers, locations, and resultant

206

Pb/238U ages of

CE P

data points are labeled. Right: U–Pb concordia diagrams for zircons from the sample (LA-ICP-MS results), where the analyses form three distinctive clusters

AC

with different weighted mean ages. Location shown in Figs. 3 and 4. Figure 18. (a) Hf evolution diagram showing the result of single-zircon LA-ICP-MS analyses, and (b) chondrite-normalized REE patterns of zircon grains from the Early Permian gabbroic plutons of the Chinese Altai. For detailed discussion, see the text. Figure 19. A two-stage Early Permian mantle–crust interaction model for northern Xinjiang, NW China. For detailed discussion, see the text.

58

Figure 1

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

59

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

Figure 2

60

Figure 3

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

61

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

Figure 4

62

Figure 5

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

63

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

Figure 6

64

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

Figure 7

65

Figure 8

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

66

Figure 9

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

67

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

Figure 10

68

Figure 11

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

69

Figure 12

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

70

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

Figure 13

71

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

Figure 14

72

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

Figure 15

73

Figure 16

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

74

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

Figure 17

75

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

Figure 18

76

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

Figure 19

77

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

Graphical Abstract

78

ACCEPTED MANUSCRIPT Highlights

AC

CE P

TE

D

MA

NU

SC R

IP

T

 We report scapolite-bearing granulite in Chinese Altay for the first time.  Granulite facies metamorphism was accompanied by partial melting of the crust.  Earlier stage of mantle-crust Interaction involved energy exchange mainly.  Mass and energy exchange occurred at later stage of mantle-crust interaction.

79