Mantle Input to the Crust in Southern Gangdese, Tibet, during the Cenozoic: Zircon Hf Isotopic Evidence

INTRODUCTION

The Gangdese batholith in the southern margin of the Lhasa terrane, which is one of the largest intrusive belts on earth, extending over 2 000 km east-west along the India-Yarlung Zangbo suture zone, is generally accepted as the product of the Tethyan oceanic subduction and subsequent India-Asia continental collision. The magmatism started as early as the Late Triassic, and ceased at ~10 Ma, recording long-lasting crust-mantle interaction processes and therefore providing an important opportunity for exploring the timing and style of the crustal growth of the Asian continent (Ji et al., 2009; Mo et al., 2008, 2007, 2005a, b; Wen et al., 2008; Chu et al., 2006; Chung et al., 2005, 2003; Dong et al., 2005; Hou et al., 2004).

Based on our previous detailed field geological mapping and SHRIMP zircon U-Pb dating on part of the Quxu granitoid complex close to Lhasa, abundant mafic micro-enclaves (MME) and gabbros are recognized within the host granitoid and are involved as the mafic end-member during the magma mixing events that peaked at ~50 Ma (Dong et al., 2005; Mo et al., 2005a). Such peak magmatism at ~50 Ma was recently confirmed by 25 SHRIMP zircon U-Pb ages sampled along ~800 km length of the Gangdese batholiths (including gabbro, gabbroic enclave, diorite, granodiorite, and granite) from Dajia Co in the west to Bomi in the east (Wen et al., 2008). These authors referred to this peak magmatism as the magmatic "flare-up" event and attributed it to the slab breakoff of the subducted Neo-Tethyan oceanic lithosphere. Unfortunately, no isotopic data are available in Wen et al. (2008) that could further address the mantle origin of the complex. A more recent work by Ji et al. (2009) revealed a systematic dataset of LA ICP-MS zircon U-Pb ages and Hf isotopic compositions on the Gangdese batholiths in southern Tibet, revealing that the magmatism occurred from 205 to 13 Ma, and can be compared with the Kohistan-Ladakh batholiths in the west and the Chayu-Burma batholiths in the east. They also pointed out that the third stage of 65–41 Ma, which had the most juvenile nature among the four divisions, is the most prominent period of granitic magmatism in the Gangdese batholith. These new zircon Hf isotopic data substantially support the previous recognition that the Gangdese batholiths are distinctly characterized by positive ε_Nd(t) values (Jiang et al., 1999; Harris et al., 1988). Such prominent granitoid magmatism, in combination with the contemporaneous Linzizong volcanic successions (65–41 Ma, Lee et al., 2007; Zhou et al., 2004) that are also characterized by mantle-like positive ε_Nd(t) values, suggests a significant mantle contribution to the crust, which therefore played a significant role in creating the thickened crust beneath the Tibetan plateau (Mo et al., 2008, 2007).

In this article, based on our previously published SHRIMP zircon U-Pb ages of the granitoids and related mafic rocks in the Quxu batholith, we present our newly obtained zircon Hf isotopic data on the same samples. We, (1) discovered an extensive mantle input when magmatism peaked at ~50 Ma, (2) discuss the contribution of this huge flux from the mantle to the crust, and (3) consider it as an important mechanism in thickening the crust during the early stage of the India-Asia continental collision.

GEOLOGICAL SETTING AND SAMPLES

The Quxu pluton is one of the typical plutons among the Gangdese batholiths that have attracted many studies both in dating and geochemistry since the Sino-France and Sino-British collaborations in the 1980s. A detailed field mapping on this pluton (Dong et al., 2005; Mo et al., 2005a, b) revealed that the granitoids and related mafic bodies and MME were emplaced almost synchronously, ranging from 45 to 55 Ma, with a peak emplacement age of ~50 Ma. The elemental and Sr-Nd-Pb isotopic investigation on rocks, from gabbros to granites, in the Quxu complex led to the recognition of a magma mixing event and an associated magma underplating process (Dong et al., 2008, 2006, 2005; Mo et al., 2005a).

Six samples previously dated by SHRIMP zircon U-Pb dating (Mo et al., 2005a) were selected for zircon Hf isotopic analysis. These six samples were collected from two localities: the first set (samples SZ0345-A, B, D) from Mianjiang Village (29°21.7'N, 90°42.5'E) in the southern Quxu pluton and the second set (samples SQ0343-A, B, C) from Niedang Village (29°29.9'N, 90°56.3'E) in the northern Quxu pluton. The age and lithology of these six samples are listed in Table 1. The ages range from 47 to 51 Ma, with a peak age of ~50 Ma. The dating results (Mo et al., 2005a) and geochemical modeling (Dong et al., 2006) suggest that the rocks (including gabbro, MME, and granodiorite) are all mantle origin and/or mantle-input-related, and members of magma mixing processes.

Table  1.  Summary of zircon Hf isotopic results of the gabbro-granite rock in the Gangdese botholiths, southern Tibet

 | Show Table

DownLoad: CSV

ANALYTICAL METHODS

Zircon Hf isotope analysis was done on the same dated spots using LA ICP-MS with a beam size of 60 μm and a laser pulse frequency of 8 Hz at the Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing. The details of the instrumental conditions and data acquisition are given in Wu et al. (2006) and Ji et al. (2009). During analysis, the ¹⁷⁶Hf/¹⁷⁷Hf and ¹⁷⁶Lu/¹⁷⁷Hf ratios of the standard zircon (91500) were 0.282 322±22 (2σ_n, n=28) and 0.000 319, consistent with the values (0.282 307±31, 2σ_n, n=44) obtained previously in this laboratory (Wu et al., 2006).

ZIRCON Lu-Hf ISOTOPIC RESULTS

A total of 90 sets of ¹⁷⁶Hf/¹⁷⁷Hf isotopic data on the zircons from the 6 samples are listed in Table 2. A summary of the results is given in Table 1. The results are also plotted in ε_Hf(t) against their ages in Fig. 1. The ε_Hf(t) values (the parts in 10⁴ deviation of the initial Hf isotope ratios between the zircon sample and the chondritic reservoir) and T_DM^C (the zircon Hf isotope crustal model ages based on a depleted-mantle source and an assumption that the protolith of the zircon's host magma has an average continental crustal ¹⁷⁶Lu/¹⁷⁷Hf ratio of 0.015) were calculated following Griffin et al. (2002) using the ¹⁷⁶Lu decay constant adopted in Blichert-Toft and Albarède (1997).

Table  2.  Zircon Hf isotopic results of the gabbro-granite rock in the Gangdese batholiths, southern Tibet

 | Show Table

DownLoad: CSV

Figure  1.  Plots of ε_Hf(t) vs. ages (a) and whole-rock SiO₂ content (b) of the Quxu gabbro-granitoid complex in southern Tibet. Sources of age data are the same as in Tables 1 and 2.

DownLoad: Full-Size Img PowerPoint

In the three samples (SZ0345-A, B, D) of the first set that was collected in Mianjiang Village, the U-Pb ages range from 48.9 to 49.9 Ma (Mo et al., 2005a); the ¹⁷⁶Hf/¹⁷⁷Hf ratio ranges from 0.283 045 to 0.283 118 (ε_Hf(t)=10.6–13.2) in the gabbroid MME sample (SQ0345A), from 0.282 928 to 0.283 070 (ε_Hf(t)=6.3–11.7) in the gabbro sample (SQ0345B), and from 0.283 058 to 0.283 123 (ε_Hf(t)=11.2–13.4) in the diorite sample (SQ0345C) (Tables 1 and 2, Fig. 1a). Their ε_Hf(t) values decrease slightly with decreasing SiO₂ content (Fig. 1b). The positive ε_Hf(t) values in these samples are consistent with a mantle origin of the gabbros and mantle-input-related diorite. The zircon Hf isotope model ages based on a depleted-mantle source (T_DM) are in the range of 230–338 Ma.

In the three samples of the second set that was collected in Niedang Village (47–51 Ma, Mo et al., 2005a), the ¹⁷⁶Hf/¹⁷⁷Hf ratios and ε_Hf(t) values of the amphibole gabbro (SZ0343B, 0.283 012–0.283 163, ε_Hf(t)=9.3–14.7) are slightly greater than those of the granodiorite (SZ0343A, 0.283 005–0.283 057, ε_Hf(t)= 9.4–11.2) and the gabbroidic MME (SZ0343D, 0.282 990–0.283 054, ε_Hf(t)=8.7–11.0). The T_DM values of the three samples range from 293 to 335 Ma. These samples also show depleted-mantle Hf isotopic features. As a whole, the gabbros show a decreasing trend in ε_Hf(t) value with decreasing SiO₂ content (Fig. 1b).

DISCUSSION

Extensive Mantle Input during India-Asia Collision

The zircon Hf isotopic data of the six samples of the Quxu pluton from Mianjiang and Niedang reported in this study further support this point. The average ε_Hf(t) values of the samples, including gabbros, gabbroic MMEs, and granodiorites, range from 10 to 12.2 (Table 2). The variation within one sample can be as high as 5.4 units of ε_Hf(t) (such as that in two gabbros, SQ0345B and SZ0343B). They have a young Hf model age (230–340 Ma). These depleted mantlelike high and positive ε_Hf(t) values suggest that the Quxu complex, one of the typical Gangdese batholiths, is generated from either the mantle wedge or a pre-existing mafic lower crust beneath the collision zone instead of the general granites that originated from the recycling of a sedimentary lower continental crust. The majority of a widely sampled research also shows this positive and wide range variation in ε_Hf(t) values, with the ε_Hf(t) values changing from 4 to 15, corresponding to the ages between 47 and 70 Ma, and the T_DM(Hf) from 510 to 730 Ma (Ji et al., 2009). These zircon Hf isotopic signatures are consistent with the Nd isotopic compositions (ε_Nd(t)=2–8.5, DePaolo et al., 2008; Dong et al., 2008; Jiang et al., 1999; Harris et al., 1988).

The Paleogene Linzizong volcanic succession that was studied in detail in the Linzhou basin (LVS, 40–65 Ma) temporally overlaps with the Gangdese batholiths. It is similarly regarded as a product of Tethyan subduction-related magmatism and also shows mantle-like Nd-Sr features (Mo et al., 2008, 2007, 2005a, b, 2003; Zhou et al., 2004). The zircon Hf isotope of LVS reported by Lee et al. (2007) (Fig. 1) is also the same as that of the Gangdese batholiths.

The Hf isotope results obtained in this study and others (Ji et al., 2009; Lee et al., 2007) in the central Gangdese batholith and related volcanic rocks further support that the mantle materials were added to the crust via the partial melting of the subducted remaining part of the Tethyan Ocean crust (Mo et al., 2008) or the recycling of the pre-existing oceanic arc terrane (with positive ε_Nd(t) values) (Ji et al., 2009). This is an important issue in figuring out whether the collisional zone is also the place of crustal growth from below (Mo et al., 2008).

To the west syntaxis, the western extension of the Gangdese batholith in Ladakh and Karakoram, the same situation was also found by Ravikant et al. (2009). The granite and diorite from the Ladakh batholith (50–68 Ma) have ε_Hf(t) values ranging from 7.4 to 10.3, corresponding to a young model age (T_DM(Hf)=510–730 Ma). Therefore, such a significant feature (high and positive ε_Hf(t) values and a young Hf model age) in a large spatial frame from Ladakh to the Gangdese batholith suggests a giant mantle-to-crust process through the magmatism during the early collision between India and Asia (Ji et al., 2009).

Constraints in Magma Mixing

Magma mixing processes were employed in describing the petrogenesis of the Quxu pluton, and the portions of basic and acid end-members involved in the mixing event were quantitatively estimated using major elemental compositions (Dong et al., 2008, 2006, 2005; Mo et al., 2007, 2005a, b). The remaining issue is that the two proposed end-members cannot be recognized by Nd-Sr-Pb isotopes in terms of the most mafic portion having the highest Nd and the least Sr isotopic ratios. For instance, the ε_Nd(t) values changed from 2.7 to 8.5, unrelated to their SiO₂ contents (47 wt.%–56 wt.%) (Dong et al., 2008). This is explained as one of the geochemical features for supporting a magma mixing origin of granite and related rocks, indicating that the magma has been chemically homogeneous during the mixing processes (Dong et al., 2008, 2006; Mo et al. 2005a, b).

In the Hf results of this work (Fig. 1b), there are no correlations between the ε_Hf(t) values and their SiO₂ contents. For example, the most basic sample (SZ0343D) has the lowest mean ε_Hf(t) values (8.8) among all the 6 samples, even lower than that of the granodiorite (9.1 in sample SZ0343A). This is similar to the above-mentioned Nd-Sr isotopes. In this point, the zircon Hf isotopic results are consistent with the Nd-Sr data. This was also discovered by Ji et al. (2009), who pointed out that no significant Hf isotopic difference existed between the mantle-derived gabbro and the crustal-derived granite, further indicating that these zircons were crystallized from different kinds of host magmas.

However, it should be noted that there are large variations in zircon ε_Hf(t) values, up to 5.4-ε units within a single rock and up to 15-ε units between zircons within the six contemporaneous samples of ~50 Ma. Similar large variations in the zircon Hf isotopic compositions of the Quxu pluton are also observed in the granitoids (133–117 Ma) in eastern Tibet (Chiu et al., 2009), the Fogang batholith in SE China (Li et al., 2007), the granitoids in the Lachlan fold belt of Australia, and the separation point suite of New Zealand (Bolhar et al., 2008; Kemp et al., 2007). This heterogeneity requires an open system with more than 2 end-member mixtures of the magma source region instead of partial melting or fractional crystallization. In other words, the large variations in zircon ε_Hf(t) values observed in the Quxu pluton are indicative of magma mixing.

Then, the remaining issue is which "crust" end-member was involved during the magma mixing in the Gangdese batholiths. In Fig. 1a, there is no zircon with crustal signature (ε_Hf(t) < 0) with an age range of ~65–47 Ma, suggesting that the majority of the source region are mantle-like materials or even a small number of crustal input. But at ~47 Ma, the crustal components with negative ε_Hf(t) values appeared. Such a change was interpreted by Ji et al. (2009) as the involvement of old crustal material and was attributed to the input of Indian continental crust materials. But a just finished article by Zhu et al. (2009) reported that the Paleocene diorite (~62 Ma) south of Nanmulin contains a large amount of inherited Proterozoic zircon (466–1 632 Ma) with ε_Hf(t) values ranging from -9.9 to 4.6 (T_DM^C=1.46–2.53 Ga). The emplacement age of the diorite significantly predates by about 10 Ma the arrival of the Indian continental crust beneath the Gangdese batholith (Chung et al., 2005). Therefore, the finding would indicate that the crustal materials involved in the magma mixing of the Gangdese batholith may have come from the Lhasa terrane itself rather than the Indian continent.

Constraints in the Crust Thickening and Growth on Tibetan Plateau

In a recent study, Mo et al. (2007) divided the postcollisional process into three phases after the India-Asia collision around 65–70 Ma based on a systematic research on the associated magmatism. They discussed the contributions of the magmatism to the crustal thickening during each phase. Among the three, Phase Ⅰ, which is represented by the syn- collisional LVS volcanism (~65–40 Ma) and the emplacement of southern Gangdese batholiths (a peak age of ~50 Ma), is the key period for the formation of the lower juvenile crust via the input of mantle-derived magmas and caused the crustal thickening in southern Tibet. The authors proposed that the mantle material input contributed about 30% (about 20 km) to the total thickness of the present-day Tibetan crust. The zircon Hf isotopic compositions reported in this study, in combination with the data recently published (Ji et al., 2009; Lee et al., 2007), further confirmed this mantle input to the crust of southern Tibet. In this regard, the zircon Hf isotopic compositions provide new insights to justify the previous interpretation (Mo et al., 2007) that the magma mixing and magma underplating at about 50 Ma may have played an important role in creating the crust of the southern Tibetan plateau.

CONCLUSIONS

(1) Six samples (including gabbro, mafic microenclaves (MME), and granodiorites) used for LA ICP-MS zircon Hf isotopic analysis of the Quxu complex yielded ¹⁷⁶Hf/¹⁷⁷Hf ratios ranging from 0.282 921 to 0.283 159, corresponding to ε_Hf(t) values of 6.3 to 14.7. Their T_DM and T_DM^C are in the range of 137 to 555 Ma and 178 to 718 Ma, respectively.

(2) The mantle-like high and positive ε_Hf(t) values of the samples suggest a mantle input of the juvenile source regions.

(3) There is no relationship between the sample composition and the ¹⁷⁶Hf/¹⁷⁷Hf compositions. A large variation in ε_Hf(t) values within a single rock (up to 5.4-ε units) and within the 6 samples (up to 15-ε units) is observed. These further suggest that the magma mixing and magma underplating at about 50 Ma may have played an important role in creating the crust of the southern Tibetan plateau.