Introduction

Western Yunnan, usually interpreted as the southeastern part of the Himalayan belt, belongs to one of the important branches of the eastern Tethyan tectonic belt and, therefore, bears significance of the Tethyan belt in SE Asia (e.g., Gansser 1981; Tapponier et al. 1981; Sengör 1988). Geological evolution of SE Asia from Paleozoic to Mesozoic can be regarded as an accretion history by the amalgamation of several micro-continents or continental blocks to the Eurasian supercontinent due to successive closure of Tethys oceans (e.g., Sengör 1988). In common view, these mirco-continents rifted from the northern margin of supercontinent Gondwana and drifted northwards (e.g., Metcalfe 1996a, b; Ueno 2000). The Tethyian orogen of western Yunnan is composed of several continental blocks (terranes) and suture zones, e.g., the Simao, Tengchong and Baoshan blocks and the Changning-Menglian suture zone (e.g., Zhong 1998).

According to the distribution of the Jurassic fauna, Neumayr (1885) developed the idea that during the Mesozoic an ocean traversing the Eurasian continent must have existed along the equator. Suess (1893) proposed that such an ocean which he named “Tethys” separated Laurentia in the north from Gondwana in the south. Closure of the Tethys and subsequently collision of these two continents built the Alpine-Himalayan orogenic belt. Further studies on the Tethyan belt and paleogeographical reconstruction of continents have revealed the existence of the Paleo-Tethys in the east of the Pangea supercontinent during the Carboniferous to Permian time (e.g., Sengör 1988; Sengör and Natal’in 1996). Consequently, it has been commonly accepted that at least two ocean basins, the Paleo-Tethys and the Neotethys, were closed before the final collision of Gondwana and Laurentia (e.g., Sengör 1988). Metcalfe (1996a, b, 2002) has proposed the existence of a Meso-Tethys ocean basin that was opened in late Early Permian and closed before Late Cretaceous times. Between Gondwana and Laurentia, numerous micro-continents or continental blocks, such as Cimmeria (Sengör 1988), Sibumasu, and the Lhasa block (e.g., Metcalfe 1996a, b; Jin 2002; Wang et al. 2001), were rifted from the northern margin of Gondwana and were subsequently accreted to the Eurasian continent during the Paleozoic and Mesozoic (Metcalfe 1996a, b; Ueno 2000). Besides biogeographical constraints, provenance analysis of the basements of these micro-continents is crucial for understanding the origin and tectonic processes, which led to the formation of the Tethyan orogenic belt. Basement rocks in the Yunnan Tethyan belt are generally concealed beneath thick cover sequences. This study aims to obtain indirect evidence from granites derived from the underlying basement. Here we present zircon ages and Hf isotopic composition as well as geochemical and Nd isotopic data of granites exposed in the Baoshan block to constrain the age and isotopic characteristics of this block and to discuss the pre-Tethyan evolution of Tethyan Yunnan.

Geological setting

The Tethyan belt in the western Yunnan (Fig. 1a) comprises the Simao block of South China affinity and the Baoshan and Tengchong blocks of Gondwana affinity (Wu et al. 1995; Metcalfe 1996a, b; Wopfner 1996; Zhong 1998; Ueno 2000). The last two blocks belong to the northern part of the Sibumasu microcontinent (Metcalfe 1988, 1996a, b, 1998) or the Shan-Thai block (e.g., Hutchison 1993; Shi and Archbold 1995, 1998). During the Triassic closure of the Changning-Menglian ocean, a branch of the Paleo-Tethys, these two micro-continental blocks formed different tectonic entities along the passive continental margin of Gondwanaland. Some researchers consider that they were separated by the eastern branch of the Banggong-Nujiang ocean basin that formed during the development of the Neotethys. This ocean basin was closed in early Jurassic and, subsequently, the Tengchong and Baoshan blocks were welded together (e.g., Zhong 1998). The collision led to the formation of the Gaoligongshan high-grade metamorphic zone.

Fig. 1
figure 1

a Geological sketch of the Tethyan belt in western Yunnan (Zhong 1998). b Geological map of the study area (Yunnan Geological Survey 1981) showing the sample locality. LLRF the Lushui-Luxi-Ruili fault

Full size image

In its present position, the Tengchong and Baoshan blocks are separated by the Lushui-Luxi-Ruili fault (LLRF). The basement is largely covered by medium-grade to low-grade metamorphic clastic Paleozoic to early Mesozoic sedimentary rocks. These rocks are intruded by numerous granitoids mainly Mesozoic to Cenozoic in age (Yunnan Geological Survey 1981; Zhong 1998). The Mesozoic granitoids are probably related to subduction of the Neotethys before the Indo-Eurasian collision, whereas the Cenozoic granitoids post-date the Indo-Eurasian collision (e.g., Rowley 1996). In the Tengchong block, rift-related volcanism was active over the last five million years. Volcanic rocks are basaltic to andesitic in composition and were derived from a metasomatized mantle source that probably resulted from prior subduction beneath Asia (e.g., Zhu et al. 1983; Chen et al. 2002a). The basement of the Tengchong block, also interpreted as the southeastern part of the Himalayan fold zone or part of the Sundaland plate (e.g., Fan 1978; Powell and Johnson 1980), is mainly composed of gneiss, migmatite, migmatitic granite, and small lenses of mafic rocks. The Gaoligongshan metamorphic zone comprises diverse rock types of different ages, including migmatitic granites, probably being Cenozoic in age, low-grade metasediments, supposed to be early Paleozoic in age, and amphibolite-facies orthogneisses which have Mesoproterozoic to Neoproterozoic protolith ages (Yunnan Geological Survey 1981). Orthogneisses are probably part of the basement of the Gondwana-affinity blocks. The basement of the Baoshan block is largely hidden by young sedimentary cover rocks. It is commonly accepted that the Pake Group of the Ximeng rock complex represents part of the Baoshan basement. Rocks of this group which underwent greenschist-facies metamorphism are Neoproterozoic in age (e.g., Zhong 1998).

A sedimentary rock unit, known as the Gongyanghe Group of late Precambrian (?) to early Cambrian age is exposed in the study area (Fig. 1b). It was formerly considered as part of the basement of the Baoshan block in this area. This group is mainly composed of low-grade metamorphic arkose, sandstone and silt slate in the lower part and contains silicalite layers in the upper part. The thickness and number of the silicalite layers increases upwards. The Gongyanghe Group is bounded to the high-grade metamorphic rocks of the Tengchong block in the northwest by the Lushui-Luxi-Ruili fault. Paleozoic and late Mesozoic to early Cenozoic granites intruded into this group that is partly covered by late Cambrian to Cenozoic sedimentary rocks. Six analyzed samples whose localities are shown in Fig. 1b were collected from the early Paleozoic Pingdajie granite (samples 03BS06, 03BS08 and 03BS09) and from the Huataolin granite of late Mesozoic to early Cenozoic age (samples 03BS07, 03BS10 and 03BS11a). In Chinese literatures, the Huataolin granite belongs to the late Yanshanian (Late Cretaceous to Early Tertiary) magmatic cycle.

Analytical methods

Whole-rock powder was obtained by crushing and splitting of about 15 kg of rock material. Zircons were isolated from crushed rocks by standard mineral separation techniques and were finally handpicked for analysis under a binocular microscope. For cathodoluminescence (CL) investigation zircon grains were mounted in epoxy resin and polished down to expose the grain centres. CL images were obtained on a microprobe CAMECA SX51 at the Institute of Geology and Geophysics, Chinese Academy of Sciences (IGG CAS). Zircon U–Pb analysis was performed on a Finnigan MAT-262 at Tübingen University, Germany, and on a GV IsoProbe-T mass spectrometer at the Laboratory for Radiogenic Isotope Geochemistry (LRIG), IGG CAS. Sm–Nd and Rb–Sr isotopic composition was measured using a Finnigan MAT-262 mass spectrometer at the LRIG. Major element contents were analyzed by the X-ray fluorescence spectrometry (XRF) with a Philips PW1400 spectrometer at the IGG CAS. Trace element contents were measured by an inductively coupled plasma mass spectrometry (ICP-MS; Finnigan MAT Element) at the IGG CAS. Analytical uncertainties are generally better than 5%.

For conventional U–Pb isotopic dilution analyses, single zircons or populations consisting of a few morphological identical grains (up to four grains) were shortly washed in warm 7 N HNO3 and warm 6 N HCl prior to dissolution to remove surface contamination after air abrasion (Krogh 1982). A mixed 205Pb–235U-tracer solution was added to the grain(s). Dissolution was performed in PTFE vessels in a Parr acid digestion bomb (Parrish 1987) using the vapor digestion method. The bomb was placed in an oven at 210°C for 1 week in 22N HF acid and for 1 day in 6 N HCl acid to dissolve fluorides into chloride salts and avoid U–Pb fractionation. Separation of U and Pb was carried out by ion exchange chromography using Teflon columns filled with 40-μl AG1-X8 (100–200 mesh) anion exchange resin. The separation procedure is described in Poller et al. (1997). Intensity of 204Pb was measured on the ion counter when using the MAT262 mass spectrometer or the Daly detector of the IsoProbe-T mass spectrometer, while other Pb isotopes were measured with Faraday cups. Total procedural blanks were <10 pg for Pb and U. A factor of 1‰ per atomic mass unit for instrumental mass fractionation was applied to all Pb analyses, using NBS 981 as reference material. Common Pb contribution remaining after correction for tracer and blank was corrected using values of Stacey and Kramers (1975). U–Pb analytical data were evaluated using the Pbdat program (Ludwig 1988). More details on analytical techniques are given in Chen et al. (2000, 2002c). Results of in house measurements on natural zircons from the Phalaborwa Igneous Complex/South Africa and from Kuehl Lake/Canada (Zircon 91500; Wiedenbeck et al. 1995) are summarized in Chen et al. (2002b). Using the U–Pb method, we obtained a 207Pb/206Pb age of 2053.5 ± 1.2 Ma for Phalaborwa zircon, similar to the age of 2051.8 ± 0.4 Ma obtained by the evaporation method (Kröner et al. 2001). Analysis of zircon 91500 gave nearly concordant U–Pb ages of 1065.6 ± 2.2 Ma, consistent with the reported U–Pb age of 1065.4 ± 0.3 Ma obtained in different laboratories (Wiedenbeck et al. 1995). Pb evaporation ages are expressed as a weighted average of 207Pb/206Pb ratios and the errors refer to the 95% confidence level, calculated using the Isoplot program (Ludwig 2001).

For Nd–Sr isotope analyses, Rb–Sr and light rare-earth elements were isolated on quartz columns by conventional ion exchange chromatography with a 5-ml resin bed of AG 50W-X12 (200–400 mesh). Nd and Sm were separated from other rare-earth elements on quartz columns using 1.7-ml Teflon powder coated with HDEHP, di(2-ethylhexyl)orthophosphoric acid, as cation exchange medium. Sr was loaded with a Ta–HF activator on pre-conditioned W filaments and was measured in single-filament mode. Nd was loaded as phosphate on pre-conditioned Re filaments and measurements were performed in a Re double filament configuration. The 87Sr/86Sr and 143Nd/144Nd ratios are normalized to 86Sr/88Sr = 0.1194 and 146Nd/144Nd = 0.7219, respectively. In the LRIG, repeated measurements of Ames metal and the NBS987 Sr standard during the 2004/2005 period gave mean values of 0.512149 ± 0.000022 (n = 98) for the 143Nd/144Nd ratio and 0.710244 ± 0.000033 (n = 100) for the 87Sr/86Sr ratio. Results of repeated Rb–Sr and Sm–Nd analyses on the standard material BCR-1 (basalt powder) are given in Table 2. The external precision refers to the 2σ uncertainty based on replicate measurements on these standard solutions over 1 year. Total procedural blanks were <300 pg for Sr and <50 pg for Nd.

In situ zircon Hf isotopic composition was measured on a Finnigan Neptune Multi-Collector ICP-MS at the MC-ICP-MS laboratory, IGG CAS. The laser system delivers a beam of 193 nm UV light from a Geolas excimer laser ablation system. Sampling spots have beam diameters of 63 or 32 μm for small grains. Ar and He carrier gases were used to transport the ions derived from the ablated material in the laser-ablation cell via a mixing chamber to the ICP-MS torch. For details on measurement procedures and interference correction see Xu et al. (2004). Isotope analyses were performed on those zircons monitored by CL studies, and the CL images provided guidance in the choice of the beam spot site. Isobaric interference of 176Lu on 176Hf was corrected using the intensity of the interference-free 175Lu isotope and a recommended 176Lu/175Lu ratio of 0.02655 (Machado and Simoetti 2001), while isobaric interference of 176Yb on 176Hf isotope was corrected using the interference-free 172Yb isotope and a recommended 176Yb/172Yb ratio of 0.5886 (Chu et al. 2002). Zircon 91500 was used as the reference standard during routine analyses, with a recommended 176Hf/177Hf ratio of 0.282293 ± 28 from laser analyses (Woodhead et al. 2004). ɛHf(t) values were calculated using chondritic Hf data (Blichert-Toft and Albarede 1997). Parameters used for calculation of depleted mantle model ages (T DM(Hf)) are from Griffin et al. (2000).

Chemical and Nd–Sr isotopic composition

Three analysed samples collected from the early Paleozoic Pingdajie granite are fine-grained biotite granites, containing quartz, feldspar and biotite as major mineral constituents with minor secondary muscovite, apatite, zircon and opaque phases. Three samples of the late Yanshanian Huataolin granite are medium-grained granites containing mainly quartz and feldspar, and minor biotite and muscovite. Accessory minerals are zircon, apatite and opaque phases. Sample 03BS11a is a light colored tourmaline-bearing rock.

Analytical data for major and trace elements of six whole-rock samples are given in Table 1. SiO2 and Al2O3 contents range from 71.2 to 77.0 and 12.6 to 16.7 wt%, respectively. A/CNK values (molar Al2O3/(CaO + Na2O + K2O ) ratio) vary from 1.14 to 1.49 (Fig. 2a), indicating peraluminous composition (Fig. 2b). The samples are characterized by high Rb/Sr ratios and plot in or close to the field of syn-collisional granites in the (Y+Nb) versus Rb diagram (Fig. 3). The three samples from the Pingdajie granite exhibit similar chondrite-normalised rare earth element (REE) patterns with moderate light over heavy REE enrichment (normalised La/Yb ratios 9.5–10.3; normalised Tb/Yb ratio 1.44–2.93) and negative Eu-anomalies (Eu/Eu* 0.42–0.45) (Fig. 4a). The three samples from the Huataolin granite have low and variable REE contents (ΣREE 131.1–7.94 ppm) and distinct negative Eu-anomalies (Eu/Eu* 0.08–0.43). The flat REE pattern of two of these samples (Fig. 4b) is likely the result of fractionation of LREE-rich accessory phases. A stronger degree of differentiation is also indicated by the high Rb/Sr ratios of these samples. Normalised La/Yb and Tb/Yb ratios range from 1.6 to 33.6 and from 0.89 to 2.78, respectively. Both the Pingdajie and Huataolin granites have similar mantle-normalised multi-element patterns (Fig. 4c, d), that show negative Nb, Sr and Ti and positive Pb anomalies characteristic of the continental crust.

Table 1 Major and trace element concentrations of granites from the Baoshan block
Full size table
Fig. 2
figure 2

SiO2-content versus A/CNK ratio diagram (a) and A/CNK ratio versus A/NK ratio diagram (b) after Maniar and Piccoli (1989)

Full size image
Fig. 3
figure 3

Rb versus Y+Nb diagram after Pearce et al. (1984) and Pearce (1996). ORG ocean-ridge granitoid; GPG within-plate ganitoid; VAG volcanic granitoid; COLG collisional granitoid

Full size image
Fig. 4
figure 4

Normalized REE and trace element content of the early Paleozoic Pingdajie granites (a, c) and the late Yanshanian Huataolin granites (b, d). Normalizing values for chondrite and primordial mantle are referred from Sun (1982) and Taylor and McLennan (1985)

Full size image

Analytical results of whole-rock Rb–Sr and Sm–Nd isotopic composition of six samples are given in Table 2. Due to their chemical differentiation, the analysed samples display high 87Rb/86Sr ratios and, as a consequence, calculated initial 87Sr/86Sr ratios show a large variation. When calculated back to 470 Ma, initial ɛNd values of the early Paleozoic Pingdajie granites range from −9.1 to −11.5, depleted-mantle modal ages, using a two-stage model (Liew and Hofmann 1988), range from 1.9 to 2.16 Ga. Irrespective of the different REE pattern, the late Yanshanian Huataolin granite has similar ɛNd (470 Ma) values (−8.57 to −9.03) and similar two-stage modal ages (1.93–1.96 Ga) as the Pingdajie granite (Fig. 5). With respect to these data, both granites might have been derived from similar magma sources.

Table 2 Analytical data of whole-rock Rb–Sr and Sm–Nd isotopic composition
Full size table
Fig. 5
figure 5

a Initial ɛNd values (calculated for an age of 470 Ma) and b depleted-mantle model ages T DM for the early Paleozoic Pingdajie and the late Yanshanian Huataolin granites

Full size image

Zircon ages and Hf isotopic composition

Internal structures of typical zircon populations were studied using the cathodoluminescence (CL) technique prior to U–Pb and Hf isotopic analyses. Zircons of the Pingdajie and Huataolin granites have magmatic habitus (Fig. 6). Most grains of samples 03BS06, 03BS09 and 03BS10 show oscillatory zoning indicative of igneous crystallization (e.g., Hanchar and Miller 1993; Pidgeon et al. 1998). The grains do not show solid-state recrystallization textures or metamorphic-related overgrowth, which would indicate post-crystallization processes. Some zircon grains of samples 03BS06 and 03BS09 contain inherited cores with weak oscillatory CL zoning. Sample 03BS10 largely contains zircon grains having distinguishable cores. Zircon grains of sample 03BS11a exhibit a multitude of CL features. Most grains from this sample do not show typical oscillatory zoning but show sponge-like CL textures.

Fig. 6
figure 6

Cathodoluminescence (CL) images of typical zircon populations for the early Paleozoic Pingdajie granite (03BS06, 03BS09) and the late Yanshanian Huataolin granite (03BS10, 03BS11a)

Full size image

Two samples of the Pingdajie granite (03BS06, 03BS09) and two samples of the Huataolin granite (03BS10, 03BS11a) were analyzed by the single zircon U–Pb isotopic dilution method. Anaytical data for 43 zircon fractions are given in Table 3 and shown in concordia diagrams in Fig. 7. Eight out of thirteen zircon fractions of sample 03BS06 plot on or close to the concordia around 470 Ma, giving a mean 206Pb/238U age of 472 ± 5 Ma. One fraction yields younger discordant U–Pb ages of about 400 Ma, indicating partial Pb-loss. The other four zircon grains contain inherited cores probably as old as Archean in age (upper intercept at 2950 ± 300 Ma; Fig. 7a). Similarly, seven zircon grains of sample 03BS09 yield near concordant U–Pb ages around 470 Ma, with a mean 206Pb/238U age of 466 ± 11 Ma. This sample contains old inherited zircon components of Archean to Mesoproterozoic age (upper intercepts at 3580 ± 380 Ma and 1350 ± 130 Ma; Fig. 7b). Two zircon fractions gave nearly concordant U–Pb ages around 520 Ma, likely representing the contribution of younger crustal material to the magma source. Most zircon grains of sample 03BS10 gave discordant U–Pb ages (Fig. 7c), consistent with the internal zircon structure shown in the CL images (Fig. 6). Only one out of eight grains yielded nearly concordant U–Pb ages of about 80 Ma. Ages of inherited cores vary from late Proterozoic to early Paleozoic (upper intercepts at 828 ± 71 and 495 ± 150 Ma; Fig. 7c). Despite the lack of typical magmatic zoning and sponge-like CL textures, all the analyzed zircon grains from sample 03BS11a gave concordant or nearly concordant U–Pb ages, yielding a mean 206Pb/238U age of 60.9 ± 1.4 Ma (Fig. 7d).

Table 3 Zircon U–Pb data of granites from the Baoshan block
Full size table
Fig. 7
figure 7

Zircon U-Pb concordia diagrams for the early Paleozoic Pingdajie granite (a, b) and the late Yanshanian Huataolin granites (c, d)

Full size image

A total of ninety-four zircon grains from the four granite samples were analyzed for Hf isotopic composition using the LA-MC-ICP MS method. Analytical data are given in Table 4. All analyzed zircon grains have low 176Lu/177Hf ratios, but measured 176Hf/177Hf ratios vary widely from 0.281816 to 0.282686 and corresponding ɛHf(0) values range from −33.8 to −3.0 but most data are between −16 and −10 (Table 4). The derived depleted-mantle modal ages T DM(Hf), calculated using a mean crustal Lu/Hf ratio (176Lu/177Hf ratio of 0.015; Veevers et al. 2005), show a roughly similar distribution within the four samples. T DM(Hf) model ages range from 3,809 to 1,075 Ma (Fig. 8a–d), having a major peak at about 1,900 Ma and a minor peak of about 1,600 Ma (Fig. 8e). Combining all data, a mean T DM(Hf) model age of 1834 ± 46 Ma is obtained, which is close to whole rock TDM(Nd) model ages of the granites (Fig. 5). ɛHf values of all the analyzed grains, calcaluted back to 470 Ma, range from about −37.9 to 7.1 (Fig. 8f), but most of the data cluster around −5. About 15% of analyzed grains have positive ɛHf(470 Ma) values indicating that younger material (possibly of mantle origin) contributed to magma sources.

Table 4 Zircon Hf isotopic composition of granites from the Baoshan block
Full size table
Fig. 8
figure 8

Zircon Hf isotopic composition of granite samples. a–d histograms of T DM(Hf) values of zircon samples from the early Paleozoic Pingdajie granite (03BS06 and 03BS09) and the late Yanshanian Huataolin granite (03BS10 and 03BS11a); e histograms of T DM(Hf) values of zircon grains from the four granite samples, mostly clustering around 1,800 Ma. f histogram of initial ɛHf values (470 Ma) of zircon grains from four granite samples, mostly clustering around −4

Full size image

Discussion

Early Paleozoic magmatism

The Pingdajie granite, which intrudes the late Precambrian (?) to early Cambrian Gongyanghe Group, was previously dated at about 530–495 Ma by the Rb–Sr method and at 465–349 Ma by the K–Ar dating of mica. The granite is interpreted as a product of a Caledonian orogenic cycle (Yunnan Geological Survey 1981; Zhong 1998). Similar K–Ar mica ages of about 490 Ma have been obtained for other small granite bodies exposed in the Baoshan block (Yunnan Geological Survey 1981). Our study of single zircon U–Pb dating yields an age of about 470 Ma for the crystallization time of the Pingdajie granite, providing a significantly more precise age than pre-exisiting age constraints. Contemporaneous magmatism was found in the eastern Sibumasu block, where a lens-shaped granite body in the late Paleozoic sedimentary sequence within the Nan-Uttaradit suture zone, the boundary between the Sibumasu and Indochina blocks, has been dated at 486 ± 5 Ma by the zircon U–Pb method (Metcalfe 1998). Early Paleozoic magmatism (e.g., the about 460 Ma old Dongpu granite) has been also found in the eastern Tibet (Li et al. 2002).

The tectonomagmatic evolution of the northern Gondwana margin is largely reflected in the presence of early Paleozoic orogenic and rift-related magmatic rocks (e.g., Matte 1986, 1991; Franke 1989; Pin 1990). These magmatic associations can be traced in different tectono-metamorphic units of the European Variscides and the basement zones of the Alps (e.g., Oliver et al. 1993; Kröner and Hegner 1998; Von Quadt 1997; Dörr et al. 1998; Von Raumer 1998; Von Raumer et al. 2002 and references therein). According to palaeogeographic reconstructions the continental terranes of eastern and southeastern Asia rifted from India-NW Australian Gondwana margin (e.g., Metcalfe 1998, 2000, 2006). Early Paleozoic magmatism in the Baoshan block and in the neighboring regions implies a comparable history during the early Paleozoic between SE Asia and the western Tethyan region, e.g., the basement outcrops in the European Variscan belt, which are also considered as continental blocks having rifted from Gondwana prior to or simultaneously with those of SE Asia.

Late Yanshanian (Late Cretaceous to Early Tertiary) magmatism

The single zircon U–Pb dating results of sample 03BS10, collected from the Huataolin granite, demonstrate that inherited cores of late Proterozoic to early Paleozoic age are commonly present within these zircon grains. Low intercept model ages of 78 ± 4 and 76 ± 8 Ma obtained from two discordia lines provide age constraints for the formation of this granite. Zircon grains of sample 03BS11a give near concordant U–Pb ages around 61 Ma. Therefore, it is reasonable to suggest that the Huataolin granite formed between 78 and 61 Ma. Many contemporaneous granites can be found in the adjacent Tengchong block (Yunnan Geological Survey 1981). These magmatic rocks can be related to the closure of the Neotethys ocean. Mafic granulites found in the Yingjiang island-arc magmatic rocks on the border of Burma and in western Yunnan, west of the Tengchong block, were dated at about 74 Ma by the Ar–Ar laser-ablation technique and are interpreted as evidence for subduction of the Neotethys oceanic crust in southwestern Yunnan (Zhong et al. 1999).

Age and geochemical feature of the basement

The Tethyan belt in western Yunnan, or the Sanjiang area in Chinese literature, comprises several geological terranes of different origins. Information about the formation age and geochemical features of these terranes is essential for understanding the provenance and evolution of the Tethyan terranes in southwestern China. Both the early Paleozoic Pingdajie and the late Yanshanian Huataolin granites are peraluminous in composition and have low initial ɛNd values at 470 Ma (−8.6 to −11.5), suggesting an origin by reworking of old crustal rocks that were situated beneath the Paleozoic sedimentary cover. Contribution of mantle material was probably of minor importance during the genesis of the early Paleozoic and late Yanshanian granitoids. Hence, inherited zircon ages, zircon Hf isotopic composition and whole rock Nd isotopic characteristics (initial ɛNd values and Nd model ages) reflect charateristics of the basement rocks of the Baoshan block.

The close correspondance in geochemical and isotopic composition, especially in trace elements, whole rock Nd isotopic and zircon Hf isotopic composition, imply similar sedimentary protoliths for the early Paleozoic Pingdajie and the late Yanshanian Huataolin granites. Upper U–Pb intercept model ages suggest inherited zircon cores probably as old as early Archean in age (about 3.8 Ga; Fig. 7b). Whereas the early Paleozoic Pingdajie granite contains inherited zircons predominantly of Archean to Mesoproterozoic age, the late Yanshanian Huataolin granite bears evidence for Neoproterozoic to early Paleozoic inherited zircon cores. Nd model ages (T DM) of the granites clustering around 2.0 Ga indicate an avearge late Paleoproterozoic crustal residence time for the precursor rocks. These model ages are generally older compared to those from basement rocks exposed in the European Variscides (Liew and Hofmann 1988; Hegner and Kröner 2000). A histogram of T DM(Hf) values of the four analysed samples (Fig. 8e) shows two peak values around 1.9–1.8 and 1.6–1.4 Ga. Some zircon grains give Archean T DM(Hf) values between 3.8–2.6 Ga. Geochemical data indicate S-type characteristic for both the Pingdajie and Huataolin granites. In general, great care must be taken to insure that model ages which only rely on one isotope system, e.g., Sm–Nd, are not misinterpreted with respect to crustal residence times and source ages. This holds true in particular for the evaluation of S-type anatectic magmas (Clemens 2003), which can be derived from a number of crustal sources of different age. In the present study we have combined U–Pb data from inherited zircon ages and Nd–Hf isotopic data to put constraints on the crustal residence time of the two granites from the Baoshan block. The data imply that magma sources of the granites are likely made up mainly of Paleoproterozoic to Mesoproterozoic crustal material. Some zircon grains from both the early Paleozoic Pingdajie and late Yanshanian Huataolin granites have positive initial ɛHf values at 470 Ma, i.e., the formation age of the early Paleozoic granite. This implies that the addition of juvenile (or newly mantle derived) material to the crust played some role, at least during the formation of the Pingdajie granite. However, the proportion of material sources of different ages is difficult to estimate. As mentioned above, continental terranes of E and SE Asia probably originated from the India-N/NW Australian margin of the Gondwana supercontinent (e.g., Metcalfe 1998, 2000, 2006), where Archean to Paleoproterozoic cratonic blocks are exposed (e.g., Barley 1997; Myers and Swagers 1997; van Kranendonk and Collins 1998; Wingate and Evans 2003; Yedekar et al. 1990; Jain et al. 1991; Mazumder et al. 2000; Dobmeier and Raith 2003). These cratons could have served as a major source region during the formation of the Baoshan basement. Inherited zircon ages and Nd-Hf isotopic composition of the early Paleozoic and late Yanshanian granites bear evidence for an India-N/NW Australian origin for the underlying basement of the Baoshan block. The detrital zircons of Archean age were probably derived from Archean terranes in India (e.g., the Singhbhum and Bastar cratons) or western Australia (e.g., the Yilgarn and Pilbara cratons), whereas the zircons of Paleoproterozoic age could have been derived from similar terranes in NW Australia.

Conclusions

Granites intruding the Gongyanghe Group in the Baoshan block of the Tethyan Yunnan can be devided into two groups whose emplacement is dated at about 470 Ma (Pingdajie granite) and about 78–61 Ma (late Yanshanian Huataolin granite). Geochemical and isotopic characteristics indicate that both granites mainly originated from intra-crustal reworking during two different tectonic events. The late Yanshanian magmatism is probably related to the closure of the Neotethys ocean. The early Paleozoic magmatism is more or less contemporaneous with magmatism in the neighboring areas of SE Asia (e.g., the eastern Sibumasu block and the eastern Tibet) and in the western Tethyan belt (e.g. the inner-Alpine belt and European Variscides), supporting a common derivation and separation of the different crystalline units from northern margins of Gondwana during the early Paleozoic.

Contents of trace elements in the early Paleozoic Pingdajie granite are very similar to contents found in the late Yanshanian Huataolin granite. The same holds true for the Nd–Hf isotopic record. It is concluded that both granites were derived from a similar magma sources that represent a record of the Tethyan Yunnan basement. The early Paleozoic granite contains early Archean to Mesoproterozoic inherited zircon cores, whereas the late Yanshanian granite contains late Proterozoic to early Paleozoic inherited zircons. Nd–Hf isotopic data of the granites imply Paleoproterozoic to Mesoproterozoic crustal material as the major components of the underlying basement. It is proposed that the basement was derived from the Archean and Paleoproterozoic terrains in India and NW Australia. This conclusion is consistent with a Gondwana origin of the continental terrains in SE Asia, which was previously proposed on the basis of tectono-stratigraphic, palaeontological and geophysical data (Metcalfe 1998).