Late Cretaceous adakitic intrusive rocks in the Laimailang area, Gangdese batholith: Implications for the Neo-Tethyan Ocean subduction

3.1 In situ zircon U–Pb LA-inductively coupled plasma mass spectrometry (ICP-MS) analyses

Zircons were separated from sample D2053-18-N1 by standard density and magnetic separation and then fixed with a standard zircon with epoxy resin. The surface of the sample target was polished until approaching the near-center section of the zircon crystal, and then a cathodoluminescence (CL) photograph was taken. In situ zircon U–Pb dating was conducted using an LA-ICP-MS at the Beijing GeoAnalysis using the analytical conditions defined by Yuan et al. [32]. Offline processing was performed using ICPMS Data Cal 11.5 software [33], and normal lead correction was performed on the test data [34]. Harmonic plot mapping and weighted mean age calculation were performed using Isoplot 3.0 [35].

3.2 Major and trace elements

These were trimmed to remove weathered surfaces, cleaned with deionized water, crushed, and powdered in an agate mill. The samples for geochemical analysis were ground to pass through a 200-mesh sieve and then further ground and homogenized in an agate mortar under alcohol. Major elements were analyzed by X-ray fluorescence (XRF), and trace and rare earth elements were analyzed by ICP-MS at the ALS Mineral-ALS Chemex Laboratory (Guangzhou, China). Major elements were assayed using a RIX-2100 XRF spectrometer, with an analytical precision of ±5%. Trace element concentrations (rare earth elements included) were assayed using an Agilent 7500a ICP-MS. The precision of the ICP-MS analyses is ±10% for all elements, but some are ±5%.

3.3 In situ zircon Hf isotope analysis

In situ Hf isotope measurements were taken on zircon grains at the same position as previously dated, using a Thermo Scientific Neptune (Plus) multi-collector-ICP-MS, coupled with a New Wave 193 nm solid-state laser ablation system at Beijing GeoAnalysis (Beijing, China). After removing ages with anomalies (≤95% confidence level), Hf isotopic compositions were determined for the remaining samples. Lu-Hf isotopic measurements on zircon were performed by LA-MC-ICP-MS with a beam size of 38 μm, an energy density of 10–11 J/cm², and a laser pulse frequency of 10 Hz.

4.1 Petrography

We focused on the Laimailang intrusive rocks that contain a number of regional structural features and ductile shear zones that are oriented approximately E–W (Figure 1c). Weathered surfaces are khaki in color, whereas fresh surfaces are gray and show a massive structure, and five samples were collected from rocks without visible alteration (Figure 2). They had a hypidiomorphic granular structure, with a mineral composition of mostly plagioclase (40%), potassium feldspar (25%), quartz (25%), amphibole (5%), and biotite (<5%), with apatite and zircon as an accessory mineral (Figure 2). Plagioclase is semi-idiomorphic plate, most of those have zoned crystal structure, and a small amount of the crystals have weak clay alteration (Figure 2c). Potassium feldspar is microcline in the shape of other granular, and part of the crystal contains plagioclase, hornblende, and biotite, which constitute the intergrowth textures (Figure 2d–f). Quartz crystals are similar in size to feldspar and fill between feldspar crystals (Figure 2e and f). Amphibole and biotite are rare, and chlorite alteration was found in some edges (Figure 2c). The granitic structure of the sample is complete and identified as monzogranite in QAP classification diagram (Figure 3a).

Figure 2

Representative petrographical images of the Laimailang monzogranite (a, b) field occurrences; (c–f) representative microstructures of granitoid: Qtz, quartz; Am, amphibole; Bt, biotite; Pl, plagioclase; Kfs, potassium feldspar.

Figure 3

(a) QAP classification and (b) TAS diagram [40] for the Laimailang monzogranite.

4.2 Whole-rock geochemistry

The original whole-rock major and trace element data are shown in Table 1 as standardized values. The loss on ignition (LOI) values of the five samples were all <1%, indicating only weak effects from weathering and erosion. The studied samples have acidic composition, showing SiO₂ ranging ∼65–68 wt%, in association with high Na₂O (3.41–4.07 wt%), Al₂O₃ (14.95–16.60 wt%), and K₂O (mostly 2.92–3.78 wt%, except D2053-DH1 6.50 wt%), low MgO (1.38–1.83 wt%), and medium Mg^#(46–48). On the total alkalis versus silica (TAS) diagram [40], samples mostly plot in the dacite fields, with only sample D2053-DH1 plotting in the trachyte field (Figure 3b). The samples are dominantly high-K calc-alkaline (Figure 4a), and they have metaluminous affinity on the plot of A/CNK versus A/NK (Figure 4b).

Figure 4

Laimailang monzogranite sample classification within (a) K₂O vs SiO₂ diagram and (b) A/NK vs A/CNK diagram, with other results for comparison.

All samples presented marked fractionated REE patterns in chondrite-normalized REE diagrams and primitive mantle-normalized trace element spider diagrams (Figure 5). They had low HREEs and Y but relatively high (La/Yb)N, being enriched in LREEs and depleted in HREEs (Table 1), which have strong similarity with the typical adakite (Figure 5a). Samples relatively enriched in large ion lithophile elements (LILEs: Rb, Th, K, Pb, Sr, etc.) and depleted in high-field strength elements (HFSEs: Nb, Ta, Ti, etc.) (Figure 5b).

Figure 5

Laimailang monzogranite (a) chondrite-normalized REE patterns and (b) primitive mantle-normalized trace element spider diagrams (normalization values after [59]).

4.3 Zircon U–Pb ages and Hf isotopes

LA-ICP-MS U–Pb data from zircon were determined in 35 groups of ages (Table 2). The zircon grains are mostly euhedral, with elongated habits (crystal lengths of 60–120 μm) and length-to-width ratios from 2:1 to 3:1. Most zircons are generally transparent and colorless. The CL images show oscillatory and planar magmatic growth zoning with broadly homogeneous cores (Figure 6a), which are very similar to the zircon morphology of typical granitoids [36]. Their Th and U concentrations ranged from 327 to 965 and 519 to 930 ppm, respectively, yielding Th/U ratios of 0.63–1.08 that indicated igneous zircons [37]. Weighted means of pooled ²⁰⁶Pb/²³⁸U ages were taken to represent the crystallization ages reported with 95% confidence level. Of these, eight zircons had a harmonicity of <95% and were not used further in the harmonic age calculation. The ²⁰⁶Pb/²³⁸U weighted age of the remaining 27 zircons was 81.22 ± 0.49 Ma (n = 27, MSWD = 0.4) (Figure 6b), placing the crystallization ages of the monzogranite in the Late Cretaceous.

Figure 6

(a) CL images and (b) U–Pb concordia diagrams for zircon from the Laimailang monzogranite.

A total of 27 in situ Hf isotope analyses were successfully undertaken on zircons within sample D2053-18-N1 (Table 3), and initial ¹⁷⁶Hf/¹⁷⁷Hf ratios were calculated on the basis of the ²⁰⁶Pb/²³⁸U-weighted age (81 Ma). The ¹⁷⁶Hf/¹⁷⁷Hf ratio obtained for zircon standard GJ-1 during the data acquisition was 0.282005 ± 34 (2 standard deviations, n = 6), in good agreement with the recommended Hf isotopic ratio [38]. The decay constant used for ¹⁷⁶Lu was 1.867 × 10⁻¹¹ [39]. These zircons were characterized by distinctly negative initial ε _Hf(t) values (−2.1 to −5.4) and yielded ¹⁷⁶Hf/¹⁷⁷Hf ratios of 0.282661–0.282568. Zircon D2053-18-N1-14 had the lowest initial ε _Hf(t) value.

5.1 Petrogenesis and tectonic setting

The data presented herein show that the studied Laimailang intrusive suite consists of high-silica (≥56 wt% SiO₂), Na-rich, and high-Al granitoids that are characterized by high Sr, low Y, and HREE contents, and strongly fractionated REE patterns, with no significant Eu anomaly (Figure 5a), indicating that they are consistent with the definition of adakite [2]. In addition, all samples for the studied rocks plot within the adakite area on the Sr/Y vs Y diagram (Figure 7a) and (La/Yb)_N vs Yb_N diagram (Figure 7b), indicating the adakitic feature.

Figure 7

Laimailang monzogranite samples plotted on (a) Sr/Y vs Y diagram [2] and (b) (La/Yb)_N vs Yb_N diagram. Data sources: ca. 80 Ma [14] (lower crust-derived adakitic rocks); 137 Ma [17] (subduction oceanic slab-deriver adakites).

As shown in Figure 5, all samples of Laimailang pluton rocks exhibit similar characteristics, indicating that all samples could have been derived from a common magma source region. Several lines of evidence indicate that the Laimailang monzogranite are most likely derived from partial melting of thickened lower crust.

1. The fractional crystallization of amphibole can significantly increase the ratio of Sr/Y to La/Yb and reduce the content of HREEs and Y [4]. However, it has less influence on the partition coefficient of HREEs than middle rare earth elements (MREEs) [10], generating concave downwards patterns between MREEs and HREEs [4,5]. The Laimailang pluton did not show this “U-shaped” partition curve (Figure 5a). Trace element signatures for the Laimailang monzogranite are consistent with the formation of lower crust rather than subducted slab (Figure 8). Our samples showed that the remarkable HREE depletion in the Laimailang monzogranite, as attested by the elevated Sr/Y, (La/Yb)_N, and (Gd/Yb)_N (>1) values, is indicative of garnet remained in the magmatic source [6]. The absence of a pronounced Eu anomaly implies that plagioclase was not a common residual phase [2,41,42]. These arguments suggest that the Laimailang monzogranite formation likely occurred at high-pressure conditions of melting [9,41,43]. Previous studies have shown that this may have been generated from eclogites or garnet amphibolites source region [9]. The ratio of the concentration of a highly incompatible element (e.g., Th) to a moderately incompatible element (e.g., REE, such as La and Sm) can be used to identify partial melting trends [44]. In the (Th/trace-element)-Th diagrams (Figure 9), the Laimailang adakitic rocks define a linear trend, similar to typical Late Cretaceous adakitic rocks in the southern Gangdese, indicating that the magmas were generated by partial melting of the subduction oceanic lithosphere or the lower crust. However, the K₂O contents (most of 2.90–3.77 wt%) of the Laimailang monzogranite differ significantly from the subduction-related adakitic rocks, because adakitic rocks generated by partial melting of the lower crust tend to be K-rich [45] (average 2.78 wt%). Compared to the adakitic rocks in the southern Gangdese derived from partial melting of the subducting oceanic crust, the Laimailang monzogranite is relatively high silica, K rich similar to those of thickened lower crust-derived adakitic rocks (Figure 4).

2. The zircon ε _Hf(t) value (−2.1 to −5.4) of the Laimailang monzogranite is uniform and negative and old model ages (∼1.3–1.4 Ga), which likely originated from partial melting of ancient crustal materials [13]. In fact, the existence of ancient-lower crust in the Lhasa microcontinent further corroborates this suggestion [13]. It is also important to note that the very large negative ε _Hf(t) values (up to −22.0) of zircons in the central Lhasa subterrane [13], which differ significantly from the Laimailang monzogranite. Another piece of evidence is that Mamba (ca. 10 km from Laimailang) mafic enclaves (∼85 Ma) are used as the products of mantle-derived magmas, which mixed with crust-derived felsic magmas [24]. In addition, the magmatic activity during the same period of this study has obvious positive isotopic characteristics (Figure 10). All these features indicate that the Laimailang monzogranite is likely to be mainly derived from partial melting of thickened ancient-lower crust, and there is a small amount of enriched fluid-metasomatized mantle magma mixing, resulting in a small negative ε _Hf(t) value.

Figure 8

(a–h) Harker variation diagrams showing the major and trace element variations in the Laimailang monzogranite (after Wang et al., 2006). Data sources: typical adakitic rocks on research area: 137 Ma [17]; 95 Ma [16]; 90 Ma [15]; 88 Ma [51]; 85 Ma [19,24]; and ca. 80 Ma [14].

Figure 9

Laimailang monzogranite samples plotted on (a) Th/Sm vs Th diagram and (b) Th/La vs Th diagram. Data sources: 137 Ma [17] (subduction oceanic slab-deriver adakites) and ca. 80 Ma [14] (lower crust-derived adakitic rocks).

Figure 10

Laimailang monzogranite plotted on zircon ε _Hf(t) vs U–Pb age. Data sources: 137 Ma and 100–89 Ma [17,46] subduction oceanic slab-deriver adakites and 85 Ma [19,24] lower crust-derived adakitic rocks.

As the main rocks forming continental crust, granites are closely related to tectonic environment and geodynamic conditions, and some trace elements (such as Rb, Y, Yb, Nb, and Ta) are the most effective elements for the tectonic discrimination of granites due to the sensitivity of these elements to fractional crystallization [46]. The Nb/Y and Ta/Yb diagrams of our samples show the characteristics of volcanic arc (Figure 11), indicating that the Laimailang monzogranite should be in the subduction background. This is consistent with the previous studies that the Neo-Tethyan Ocean was subducted in the south Gangdese during the Late Cretaceous [11–13].

Figure 11

Diagrams of the tectonic setting of trace elements for Laimailang monzogranite (a) Nb vs. Y and (b) Ta vs. Yb (after [46]).

5.2 Geodynamic interpretation

As mentioned above, zircon U–Pb dating results indicate that a Late Cretaceous (106–80 Ma) magmatic “flare-up” and the subsequent period of magmatic quiescence (80–65 Ma) events occurred in the southern Lhasa subterrane [20,47]. The model of normal (flab or steep angle) subduction can account for much of the Mesozoic magmatism, which generally produces continuous magmatism, but the Late Cretaceous magmatism has obviously magmatic quiescence (80–65 Ma). The existence of 90 Ma extreme high temperature (HT) condition and low H₂O activity charnockites [15,47], which have adakitic affinities, as well as the mafic magmatism (88–85 Ma) [18,24] obtained in southern Lhasa subterrane. That is, the magmatic “flare-up” cannot be explained by normal subduction, which is not coincidence with the extreme high hot. The local region in southern Lhasa terrane lacks the typical ca.85 Ma extension setting rocks, such as A-type granitoids or ocean island basalt, so more evidence is needed for the back-arc extension model [24]. The discovery of mafic rocks and adakites in this area is consistent with the Neo-Tethyan mid-ocean ridge subduction model, given that the rocks indicate HT and require mantle contributions that could be attributed to mantle upwelling through a slab window [48]. Consequently, during which high heat flow as the hot asthenosphere ascends upwards through the slab window and directly heats the subducting oceanic slab and continental crust, leading to partial melting of the oceanic or continental crust along the edges of the slab window [3,49,50,51]. At the same time, the asthenospheric upwelling induced the appearance of large-scale magmatic rock (e.g., normal calc-alkaline island-arc magmatic rocks and the adakites derived from partial melting of the subducting oceanic slab) in the Late Cretaceous (106–88 Ma) (Figure 12a) [15,16,17,18,20,21,22,52].

Figure 12

Subduction model for the Neo-Tethyan Ocean in the Late Cretaceous, showing the formation model for the Laimailang adakitic rocks studied here (a) mid-ridge subduction model; (b) thickening of the crust was accompanied by the formation of the Mailang intrusive (adapted from [16,19,29]).

However, the Laimailang monzogranite exposure more than 90 km north of the previous trench, which is now represented by the Yarlung-Tsangbo suture zone (IYZSZ, Figure 1). This distance, combined with more than 40% of the major upper crustal shortening in the Cretaceous [53,54,55], suggests that Laimailang monzogranite was located north of the IYZSZ at a distance of ∼200 km or more before Laimailang pluton emplaced (ca. 81 Ma). With such a long distance, the Neo-Tethyan mid-ocean ridge subduction may have had rarely effect on the Laimailang pluton. In the study area, the adakites formed from 85 to 81 Ma thus relate to the partial melting of thickened lower crust ([14,18,19,24]; and this study), which are consistently later than that of the adakites derived from partial melting of the subducting oceanic crust [15,16,17,18,19,20,21,22]. Considering that subduction is a continuous process, these phenomena should be combined to discuss the evolution of the tectonic setting in the local.

Late Cretaceous (106–85 Ma) magmatic “flare-up” can be well explained by an oceanic ridge subduction model, which resulted in the gradual decrease in the plate’s subduction angle with HT and positive buoyancy [56], changed to low-angle subduction or flat subduction eventually. This flat-slab stage of the Neo-Tethyan subduction led to a contractional tectonic regime in which the subducting slab may have tectonically thickened the arc crust and squeezed out the mantle wedge; thus, the arc magmatism was terminated [14,57,58]. That is, adakites derived from the thickened lower crust appeared (85–81 Ma) in the south Gangdese ([14,18,19,24] and this study). A magmatic gap or quiescent period occurred between ca. 80 and 68 Ma [14,59], caused by the mantle wedge gradually squeezing out under the flat subduction model. The formation of the Laimailang pluton represents the evolution of the tectonic setting of the Neo-Tethyan oceanic ridge subduction to flat subduction (Figure 12b).