Early Silurian Slab Break-Off and Crustal Reworking in the Southern Central Asian Orogenic Belt: Insights from Liuyuan A-Type Granites

Abstract

The southern Central Asian Orogenic Belt (CAOB) underwent a major Early Paleozoic tectonic transition, yet its timing and mechanisms remain unclear. We present zircon U-Pb-Hf, whole-rock geochemical, and Sr–Nd isotopic data for newly identified Early Silurian (ca. 439–431 Ma) granitoids from the Liuyuan area of the southern Beishan Orogenic Belt. These high-silica, high-K calc-alkaline intrusions not only show arc-like trace-element patterns but also display elevated Ga/Al ratios and enriched Sr–Nd isotopic compositions ((⁸⁷Sr/⁸⁶Sr)_i = 0.7158–0.7189; ε_Nd(t) = −4.6 to −3.9), consistent with aluminous A₂-type granites derived mainly from ancient crust. Their heterogeneous zircon ε_Hf(t) values (−6.3 to +3.7) suggest a minor, localized input from mantle-derived mafic magmas superimposed on the dominant crustal signature. Integrating regional metamorphic constraints, we interpret this magmatism to have formed during the transition from oceanic subduction to incipient collision/continent involvement and subsequent post-subduction extension, plausibly triggered by slab break-off at the slab root (ocean–continent transition). Slab-window-related asthenospheric inflow and localized thermal perturbation could have promoted high-temperature crustal melting and facilitated Early Silurian crustal reworking in the southern CAOB.

1. Introduction

The CAOB is the largest Phanerozoic accretionary orogen on Earth, formed by long-lived subduction and final closure of the Paleo-Asian Ocean (PAO), and consisting of amalgamated arcs, oceanic plateaus, and micro-continents [1,2,3,4]. Within the southern CAOB, the Beishan Orogenic Belt (BOB) is a key segment for reconstructing Paleozoic accretionary processes, yet the timing and geodynamic nature of PAO closure remain debated. While many studies argue for a Late Permian–Early Triassic terminal collision [5,6,7], others suggest a Mid-Paleozoic closure followed by intracontinental rifting [7,8]. This debate is most prominent in the Liuyuan area, where interpretations of the Permian Liuyuan Complex diverge sharply between an ophiolitic suture and a rift-related sequence [9,10]. A clearer understanding of the earlier Paleozoic framework is therefore required to resolve these contrasting tectonic models.

The Ordovician–Silurian evolution of the southern Beishan belt is particularly contentious, with competing models proposing opposite subduction polarities and distinct interactions with the Dunhuang Block. One model invokes southward subduction of the Beishan Ocean beneath the Dunhuang Block, producing an active continental margin and Silurian collision [11,12,13]. Alternatively, a northward subduction model suggests progressive arc accretion involving the Huaniushan and Gongpoquan arcs onto the Beishan margin [7,14,15]. The presence of high-pressure (HP) granulites and eclogites (ca. 460–430 Ma) in the nearby Gubaoquan area further complicates interpretations: these rocks may record deep continental subduction and Beishan–Dunhuang collision [16,17] or reflect subduction-channel processes within a long-lived accretionary wedge [18]. Constraining the nature and timing of Silurian magmatism in Liuyuan is therefore crucial for distinguishing between these conflicting tectonic scenarios.

Beyond the geodynamic models, uncertainties also persist regarding crustal growth processes and basement composition in the southern BOB during the Early Paleozoic. The CAOB is widely regarded as a major locus of juvenile crust formation [1,19,20]; however, geochemical and isotopic constraints on Silurian granitoids from the Liuyuan area remain scarce, making it unclear whether they reflect mantle–oceanic components or reworked ancient continental materials. Determining their petrogenesis is essential for reconstructing the Early Paleozoic crustal architecture prior to extensive Permian magmatism.

To address the aforementioned gaps, this paper presents a systematic investigation of the newly identified granites in the Liuyuan area, combining zircon U-Pb geochronology, zircon Hf isotopes, whole-rock geochemistry, and Sr–Nd isotopic data. Our results reveal a distinct Early Silurian (ca. 431–439 Ma) magmatic event, providing a crucial but previously missing snapshot of the Early Paleozoic evolution of the region. Based on these data, we constrain the petrogenesis and magma sources of these rocks. Furthermore, we discuss the implications of these findings for the subduction polarity of the Paleo-Asian Ocean and the tectonic evolution of southern Beishan.

2. Geological Background

2.1. Regional Tectonic Framework

The CAOB, situated between the Siberian Craton to the north and the Tarim–North China cratons to the south, represents one of the largest Phanerozoic accretionary orogens on Earth (Figure 1a) [7]. The BOB, located in the southern limb of the CAOB, connects the Tianshan Orogen to the west and the Alxa Block to the east. It was assembled through the Paleozoic amalgamation of island arcs, ophiolitic mélanges, and micro-continental fragments during the subduction and closure of the PAO [6,10,11].

In this study, we adopt the widely accepted tectonic scheme of Xiao et al. (2010) (Figure 1b) [7], which delineates the orogen using major E–W trending sutures: the Hongshishan, Shibanjing–Xiaohuangshan, Hongliuhe–Niujuanzi–Xichangjing, and Liuyuan ophiolitic mélange belts. These sutures separate a collage of discrete micro-continental blocks and arcs, including the Queershan Arc, Heiyingshan–Hanshan Arc, Mazongshan Arc, Shuangyingshan–Huaniushan Arc, and Shibanshan Arc. The study area is situated within the southernmost tectonic unit, the Huaniushan–Liuyuan belt, bounded to the south by the Liuyuan–Daqishan Fault, a major crustal discontinuity separating the orogenic collage from the Precambrian Dunhuang Block [12,13,17].

Figure 1. (a) Simplified tectonic framework of the Eurasian continent showing the position of the Beishan orogenic collage within the CAOB (modified after Sengör et al., 1993, Xiao et al., 2013) [2,21]. TC: Tarim Craton; NCC: North China Craton, SCC: South China Craton; (b) geological map of the Beishan orogenic collage showing the major lithostratigraphic framework and major thrust faults, with Figure 2 outlined, modified after Song et al. (2016) [22], The age labels shown in Figure 1b are compiled from published sources: the eclogite age (467 Ma) is from Qu et al. (2011) [18] and Saktura et al. (2017) [16], whereas the other ages are from Song et al. (2016) [22].

Figure 1. (a) Simplified tectonic framework of the Eurasian continent showing the position of the Beishan orogenic collage within the CAOB (modified after Sengör et al., 1993, Xiao et al., 2013) [2,21]. TC: Tarim Craton; NCC: North China Craton, SCC: South China Craton; (b) geological map of the Beishan orogenic collage showing the major lithostratigraphic framework and major thrust faults, with Figure 2 outlined, modified after Song et al. (2016) [22], The age labels shown in Figure 1b are compiled from published sources: the eclogite age (467 Ma) is from Qu et al. (2011) [18] and Saktura et al. (2017) [16], whereas the other ages are from Song et al. (2016) [22].

Figure 2. Geology of the Liuyuan area in the Huaniushan arc, southern Beishan region (modified after Zhu et al., 2016) [23].

Figure 2. Geology of the Liuyuan area in the Huaniushan arc, southern Beishan region (modified after Zhu et al., 2016) [23].

2.2. Geology of the Liuyuan Area

The Liuyuan area is located at the southern margin of the BOB, a key tectonic segment of the southern CAOB [1,7]. The region is characterized by the widespread development of the Liuyuan ophiolitic mélange belt, which preserves a record of subduction-related accretion and the terminal closure of the Southern Beishan Ocean [10,15].

The mélange consists of an Ordovician to Early Silurian volcano-sedimentary matrix (commonly referred to as the Liuyuan Group) [11] and a variety of tectonically entrained blocks representing remnants of oceanic lithosphere and related magmatic suites [10,12]. The block assemblage includes mantle peridotites, cumulate gabbros, diabase swarms, and basaltic lavas with MORB- to OIB-like geochemical affinities [6,10]. Structurally, the area is dominated by the Liuyuan fault zone, an E–W-trending shear system that facilitated tectonic assembly of the mélange and exerted primary control on the distribution and emplacement of the studied granitoid bodies [7,10]. The Early Silurian granitoids investigated in this study intruded the mélange belt during the late stage of orogenesis and locally show intrusive/thermal interactions with adjacent mafic–ultramafic bodies (Figure 2 and Figure 3) [12,23]. The locations of all granitoid samples analyzed in this study are shown in Figure 2.

2.3. Petrography

Seven representative granite samples (21LY-51, -53, -57, -63, -81, -83 and -86) were collected from the Liuyuan area for detailed analysis. The sampling locations are marked in Figure 2, with coordinates ranging from 41°09′08′′ N to 41°11′14′′ N and from 95°36′08′′ E to 95°39′50′′ E (see Supplementary Table S2 for details).

Field investigations reveal significant intrusive relationships between the granitic plutons and regional mafic–ultramafic rocks. Sharp, irregular intrusive contacts are observed between the light-colored granites and dark-colored pyroxenites (Figure 3a), with granitic veins occasionally cutting through the mafic host. Similarly, clear contact boundaries exist between the granites and gabbros (Figure 3b). These field relationships imply that the felsic magmatism was spatially and temporally associated with mantle-derived mafic intrusions. The granite samples collected are typically grey to light flesh-red in color and exhibit a massive structure (Figure 3c).

Microscopically, the samples display a medium- to coarse-grained allotriomorphic granular texture (Figure 3d). The modal composition is dominated by alkali feldspar (35–40 vol.%), quartz (30–35 vol.%), and plagioclase (25–30 vol.%), with minor amounts of biotite (3–5 vol.%) and accessory minerals (zircon, apatite). Alkali feldspar is commonly perthitic (Figure 3e), consistent with the high-K affinity of the rocks. Quartz occurs as anhedral interstitial grains exhibiting distinct undulatory extinction (Figure 3f). Plagioclase forms subhedral prismatic crystals characterized by polysynthetic twinning and local sericitization.

3. Analytical Methods

The various types of analyses were all undertaken at the Guangxi Key Laboratory of Hidden Metallic Ore Deposits Exploration, Guilin University of Technology, Guangxi, China.

3.1. Zircon U-Pb Dating and Trace Element Analyses

Zircon crystals were extracted by conventional crushing, heavy liquid, and magnetic separation techniques, and then handpicked. Prior to isotopic analysis, cathodoluminescence (CL) images were obtained to assess internal structures and select optimal sites for U-Pb dating. The CL imaging was conducted at Langfang Yuhong Rock and Mineral Technology Co., Ltd., using a TESCAN MIRA LMS high-vacuum scanning electron microscope (SEM) equipped with a Sunny CL detector (Jinjing Technology, Beijing, China). The imaging was performed with an accelerating voltage of 10–15 keV, a filament emission current of 80 μA, a beam current of 2–3 nA, and a working distance of 20 mm.

The zircon U-Pb dating and mineral trace element analyses were undertaken at the Guangxi Key Laboratory of Hidden Metallic Ore Deposits Exploration, Guilin University of Technology (GLUT), China. These analyses were performed using a 193 nm Coherent GEOLAS HD Excimer ArF laser ablation (LA) system (Coherent, Göttingen, Germany) coupled with an Agilent 7900 inductively coupled plasma mass spectrometer (ICP–MS) (Agilent Technologies, Santa Clara, CA, USA). A laser beam diameter of 44 μm was used for zircon U-Pb analyses (the same spot size as that used for zircon Hf isotope analyses; see Figure 4). Zircon 91,500 and NIST610 glass were used as external standards for U-Pb dating and trace element analyses, respectively. The zircon standard TEMORA (417 Ma) [24] was used as a secondary standard to assess the data quality. The ICPMSDataCal (version 12.2) software was used to process the data and obtain the trace element contents and U-Pb isotope ratios. Concordia ages and diagrams were obtained with the Isoplot/Ex (3.0) software [25]. A common Pb correction was applied following Andersen (2002) [26].

3.2. Zircon Lu-Hf Isotope Analyses

Representative igneous zircons from the samples with concordant U-Pb isotope data (≥95%) were selected for LA-multiple-collector (MC)–ICP–MS Hf isotope analysis. These analyses were conducted with a Neptune MC–ICP–MS coupled to a Geolas-193 nm LA system. The laser beam diameter was 44 µm, and the repetition rate was 10 Hz. Helium was used as a carrier gas. The analytical sites were the same as or adjacent to the U-Pb dating locations. Zircons 91,500 and GJ-1 were used as external standards. Calculation of ε_Hf(t) values was undertaken using the present-day chondrite values of ¹⁷⁶Hf/¹⁷⁷Hf = 0.282772 and ¹⁷⁶Lu/¹⁷⁷Hf = 0.0332 [28] and the ¹⁷⁶Lu decay constant of 1.867 × 10⁻¹¹ yr⁻¹ [29]. Calculation of single-stage model ages (T_DM1) used the present-day depleted mantle values of ¹⁷⁶Hf/¹⁷⁷Hf = 0.28325 and ¹⁷⁶Lu/¹⁷⁷Hf = 0.0384 [30]. The present-day continental crust ¹⁷⁶Lu/¹⁷⁷Hf ratio of 0.015 [30] was used for calculation of the two-stage model ages (T_DM2).

3.3. Whole-Rock Major and Trace Element Analyses

Fresh samples were collected and crushed, and then the chips were soaked in 4 N HCl for 30 min to remove any altered material. A total of seven representative granitoid samples were collected for this study (Figure 2). To ensure the integrity of the primary magmatic compositions, care was taken during field sampling to select fresh granitoids from the pluton interior, strictly avoiding the immediate contact margins and any visibly associated mafic–ultramafic inclusions/enclaves observed in several localities. The rock chips were then powdered in an alumina ceramic shatter box. Prior to major element analyses, loss-on-ignition (LOI) values were measured by heating in a muffle furnace at 1000 °C. The ignited powders were fused into glass beads using Na₂B₄O₇·10H₂O as a flux at ~1150 °C. Major-element concentrations were determined by X-ray fluorescence (XRF) using a ZSX Primus II spectrometer.

Trace element analysis was undertaken with an Agilent 7900cx ICP–MS, following the procedures described by Liu et al. (2020) [31] and Zhang et al. (2021) [32]. Approximately 50 mg of sample powder was dissolved in a bomb with 0.5 mL of purified HNO₃ and 1.0 mL of HF, placed in a high-pressure bomb, and held at 190 °C for 48 h. An internal standard solution containing Rh and Re was used to monitor the signal drift. Data were standardized by reference to analyses of the United States Geological Survey (USGS) standards BHVO, AGV, W-2, and G-2, and Chinese national rock standards GSR-1, -2, and -3. The precisions of the major and trace element contents were better than ±2% to ±5%.

3.4. Whole-Rock Sr–Nd Isotope Analyses

Whole-rock Sr–Nd isotopic compositions were determined with a Neptune Plus MC–ICP–MS. Strontium was separated with SR-B50-A (100–150 µm) resin. The rare earth elements (REEs) were separated using cation exchange resin (AG50-X8). Neodymium was purified on HDEHP resin. Mass fractionation corrections for the Sr and Nd isotope ratios were based on ⁸⁸Sr/⁸⁶Sr = 8.375209 and ¹⁴⁶Nd/¹⁴⁴Nd = 0.7219, respectively. The USGS reference material BCR-2 was used to monitor the data accuracy and yielded a ⁸⁷Sr/⁸⁶Sr ratio of 0.704900 ± 0.000012 (n = 10) and a ¹⁴³Nd/¹⁴⁴Nd ratio of 0.512633 ± 0.000013 (n = 10) [33].

4. Results

Analytical data are presented in

Table 1. Whole-rock major and trace element analysis of Liuyuan granites from the Beishan region.

Sample	21LY-51	21LY-53	21LY-57	21LY-63	21LY-81	21LY-83	21LY-86
Major elements(wt.%)
Al2O3	14.47	14.40	14.76	12.38	12.01	12.94	13.51
CaO	1.67	1.19	2.43	0.54	1.45	1.67	1.43
Fe2O3T	1.55	2.80	2.85	0.29	1.43	1.26	1.63
K2O	4.74	4.12	4.13	5.78	4.46	3.94	5.31
MgO	0.51	0.83	0.88	0.34	0.62	0.53	0.45
MnO	0.02	0.04	0.05	0.01	0.02	0.02	0.02
Na2O	3.25	3.29	2.93	2.74	2.55	2.97	1.88
P2O5	0.06	0.07	0.07	0.02	0.05	0.03	0.04
SiO2	72.45	70.77	70.35	76.51	75.42	75.04	74.64
TiO2	0.22	0.33	0.35	0.04	0.20	0.16	0.18
LOI	0.49	1.75	0.94	0.43	1.20	0.95	0.53
Total	99.42	99.58	99.75	99.08	99.39	99.52	99.62
A/NK	1.38	1.46	1.59	1.15	1.33	1.41	1.53
A/CNK	1.07	1.20	1.08	1.05	1.03	1.06	1.18
Na2O + K2O	8.08	7.57	7.14	8.63	7.13	7.02	7.25
trace element (ppm)
V	9.58	25.5	29.2	50.1	15.7	14.9	19.8
Co	23.2	30.9	24.8	19.3	2.48	30.6	25.3
Ni	10.2	12.0	10.2	11.7	3.25	12.2	9.14
Cu	70.4	1.64	13.9	5.03	3.37	3.21	3.70
Zn	8.41	29.0	29.5	41.5	13.0	11.7	17.5
Ga	46.8	42.6	61.2	42.3	39.2	37.2	66.7
Rb	148	157	138	235	124	115	155
Sr	192	103	182	305	151	160	143
Y	22.0	24.6	24.3	28.6	5.34	5.82	9.11
Zr	219	196	195	246	119	114	149
Nb	11.9	14.2	15.1	27.5	4.32	7.02	8.01
Mo	2.22	0.303	0.380	0.621	0.102	0.102	0.189
Cd	0.132	0.127	0.124	0.176	0.0827	0.0664	0.0912
Cs	3.57	3.65	3.67	5.66	1.58	1.71	1.61
Ba	897	670	949	639	692	655	1125
La	31.0	33.7	36.1	47.8	36.3	22.5	48.4
Ce	51.3	62.5	69.7	108	70.4	42.7	89.6
Pr	5.81	6.66	7.44	12.4	7.16	4.24	8.97
Nd	20.1	23.2	25.8	44.7	23.5	14.2	29.2
Sm	3.64	4.41	4.83	8.41	3.53	2.69	4.66
Eu	1.26	1.13	1.37	1.64	0.925	0.771	1.05
Gd	3.86	4.72	5.04	7.87	3.35	2.56	4.47
Tb	0.571	0.678	0.712	1.01	0.316	0.291	0.473
Dy	3.62	4.07	4.17	5.20	1.26	1.29	2.07
Ho	0.757	0.842	0.844	0.983	0.210	0.223	0.362
Er	2.30	2.56	2.54	2.84	0.634	0.644	0.987
Tm	0.364	0.388	0.394	0.408	0.0803	0.0880	0.135
Yb	2.38	2.55	2.55	2.63	0.546	0.603	0.897
Lu	0.374	0.395	0.393	0.391	0.094	0.098	0.153
Hf	5.40	5.17	5.16	7.14	3.60	4.02	4.42
Ta	0.919	1.14	1.09	1.79	0.370	0.466	0.651
W	235	316	224	141	0.593	321	248
Tl	1.00	0.863	0.745	1.82	0.765	0.756	0.929
Pb	9.00	5.43	12.8	16.3	8.53	11.4	16.8
Th	22.6	19.2	20.9	35.1	33.5	33.4	39.0
U	3.72	3.15	2.55	4.57	3.24	3.23	3.80
Y + Nb	33.9	38.8	39.4	56.1	9.66	12.8	17.1
10,000 × Ga/Al	6.05	5.47	7.73	6.36	6.06	5.35	9.23
(La/Yb)N	9.32	9.49	10.2	13.1	47.7	26.8	38.7
Eu/Eu *	1.03	0.76	0.85	0.62	0.82	0.90	0.70

Note: LOI = loss on ignition. Eu/Eu * = Eu_N/ (Sm_N × Gd_N)^1/2. A/NK (Alumina-to-Alkalis Ratio): Defined as the molar ratio of Al₂O₃/(Na₂O + K₂O). A/CNK (Aluminum Saturation Index, ASI): Defined as the molar ratio of Al₂O₃/(CaO + Na₂O + K₂O).

and

Table 2. Whole-rock Sr–Nd isotopic compositions and concentrations for representative Liuyuan granites from the Beishan region.

Sample	21LY-51	21LY-53	21LY-83
87Rb/86Sr	2.161915	4.251223	2.011544
87Sr/86Sr	0.729239	0.742517	0.731357
±2σ	0.000013	0.000011	0.000016
(87Sr/86Sr)i	0.715843	0.716176	0.718894
147Sm/144Nd	0.111404	0.116863	0.116473
143Nd/144Nd	0.512196	0.512192	0.512174
±2σ	0.000004	0.000004	0.000005
143Nd/144Nd(t)	0.511878	0.511859	0.511842
εNd(t)	−3.9	−4.3	−4.6
TDM1(Ga)	1.43	1.51	1.54
TDM2(Ga)	1.50	1.53	1.55

and Supplementary Tables S1–S3 (S1: zircon U-Pb; S2: sampling coordinates; S3: zircon Lu-Hf).

4.1. Zircon U-Pb Geochronology

Five granite samples (21LY-51, -53, -63, -83, and -86) were selected for zircon U-Pb dating analyses. The results of the zircon U-Pb geochronology and trace element compositions are presented in Supplementary Table S1, and representative zircon CL images are presented in Figure 4. Zircon grains from all samples are euhedral, and long and prismatic in shape with length: width ratios of 2:1–4:1. They generally had varying but high Th (122–10,079 ppm) and U (418–27,349 ppm) contents and high Th/U ratios (0.07–1.07). Their chondrite-normalized REE distribution patterns showed relative depletion of light rare earth elements (LREEs), enrichment in heavy rare earth elements (HREEs), positive Ce anomalies, and negative Eu anomalies (Figure 4f), as is typical for zircon. All above-mentioned characteristics indicate a magmatic origin, and the zircon U-Pb ages represent the formation ages of the Liuyuan granites.

The zircon ages of samples 21LY-51 and 21LY-86 yielded weighted mean ²⁰⁶Pb/²³⁸U ages of 432 ± 2 Ma (MSWD = 0.053) and 432 ± 2 Ma (MSWD = 0.048), respectively. The three samples (21LY-53, 21LY-63, and 21LY-83) yielded weighted mean ²⁰⁶Pb/²³⁸U ages of 437 ± 2 Ma (1σ, MSWD = 0.056), 431 ± 3 Ma (MSWD = 0.37), and 439 ± 2 Ma (MSWD = 0.02), respectively (Figure 4a–e), representing Early Silurian crystallization.

4.2. Whole-Rock Geochemistry

The whole-rock major and trace element compositions of the Liuyuan granites are listed in Table 1.

4.2.1. Major Elements

The Liuyuan granites exhibit a coherent geochemical affinity characterized by high silica contents, ranging from 71.20 to 77.55 wt.% (mean = 74.66 wt.%). They are rich in total alkalis (Na₂O + K₂O = 7.02–8.63 wt.%; mean = 7.55 wt.%). In the Total Alkali–Silica (TAS) classification diagram (Figure 5a), all samples plot tightly within the granite field.

Regarding aluminum saturation, the samples show A/CNK values ranging from 1.03 to 1.20 (mean = 1.10). On the A/NK vs. A/CNK diagram (Figure 5b), most samples plot in the peraluminous field, with a few straddling the boundary with the metaluminous field. In terms of magmatic series, the rocks display high K₂O contents (4.00–5.86 wt.%; mean = 4.71 wt.%) and K₂O/Na₂O ratios of 1.25–2.83. On the SiO₂ vs. K₂O diagram (Figure 5c), they fall exclusively into the high-K calc-alkaline series. In the Harker variation diagram (Figure 5d), MgO contents (0.35–0.89 wt.%; mean = 0.60 wt.%) display a robust negative correlation with increasing SiO₂. This linear trend is consistent with the fractionation of ferromagnesian minerals, particularly biotite, during magma evolution. A similar decreasing trend is observed for other compatible elements (e.g., TiO₂, Fe₂O₃T, CaO, and P₂O₅; see Table 1), which generally reflects the concurrent separation of plagioclase and accessory phases (e.g., Fe–Ti oxides and apatite).

The Loss on Ignition (LOI) values are uniformly low, ranging from 0.49 to 1.75 wt.% (mean = 0.90 wt.%). These values (≤1.75 wt.%) indicate that the samples are petrographically fresh and have undergone minimal post-magmatic alteration or weathering, ensuring the reliability of the geochemical data.

4.2.2. Trace Elements

Chondrite-normalized rare earth element (REE) patterns (Figure 6b) display significant fractionation between light and heavy REEs. The samples are enriched in LREEs and exhibit relatively flat HREE profiles, with (La/Yb) _N ratios ranging from 9.32 to 47.7 (mean = 22.2). A prominent feature is a moderate negative to near-absent Eu anomaly (Eu/Eu * = 0.62–1.03; mean = 0.81). This range is consistent with variable degrees of plagioclase fractionation and/or residual plagioclase in the source.

In the primitive mantle-normalized spider diagram (Figure 6a), the granites show enrichment in large ion lithophile elements (LILEs; e.g., Rb, Th, U) relative to high field strength elements (HFSEs). Notably, all samples exhibit pronounced negative anomalies of Nb, Ta and Ti. The depletion of Nb and Ta is characteristic of subduction-related magmatism, while the negative Ti anomalies are consistent with the fractionation of apatite and Fe–Ti oxides, respectively, as implied by the major element trends.

4.3. Zircon Lu-Hf Isotopes

In situ zircon Lu-Hf isotopic data are listed in Supplementary Table S3. Analyses were conducted on dated magmatic domains to ensure temporal linkage. The measured ¹⁷⁶Hf/¹⁷⁷Hf ratios range from 0.282333 to 0.282623, representing the initial magmatic compositions. Correspondingly, the ε_Hf(t) values display a restricted range from −6.3 to +3.7, with a weighted mean of −2.0. In the ε_Hf(t) versus U-Pb crystallization age diagram (Figure 7), the data form a tight cluster straddling the Chondritic Uniform Reservoir evolution line, plotting significantly below the Depleted Mantle (DM) line. This distribution is distinct from the highly positive values typical of newly accreted crust in the CAOB. The two-stage crustal model ages (T_DM2) range from 1.16 to 1.80 Ga (mean = 1.53 Ga). These isotopic signatures indicate a dominant Mesoproterozoic crustal source with minor, localized mantle-derived mafic contributions.

4.4. Whole-Rock Sr–Nd Isotopes

Whole-rock Sr–Nd isotopic data are presented in Table 2. Initial isotopic ratios were calculated based on the zircon U-Pb crystallization age of ca. 435 Ma. The Liuyuan granites display highly evolved isotopic compositions. Initial ratios (⁸⁷Sr/⁸⁶Sr)_i are significantly radiogenic, ranging from 0.7158 to 0.7189. The ε_Nd(t) values are exclusively negative, varying from −4.6 to −3.9 (mean = −4.3).

In the ε_Nd(t) versus U-Pb crystallization age diagram (Figure 8), the Liuyuan granitoids fall within the overlapping compositional fields of the Tianshan and Altai terranes. Their moderately negative ε_Nd(t) values and Mesoproterozoic T_DM2 ages (1.50–1.55 Ga) indicate involvement of evolved crustal sources comparable to both terranes. This positioning suggests that the magmas most likely originated from reworked Precambrian basement in a transitional zone between the Tianshan and Altai crustal domains.

5. Discussion

5.1. Early Silurian Magmatic Pulse

Our new zircon U-Pb ages constrain the crystallization of the Liuyuan granitoids to a tight interval of 439–431 Ma (Llandovery–Wenlock). This finding is pivotal because it fills the previously inferred “magmatic gap” in the Southern BOB that was thought to separate Ordovician arc construction from the widespread Late Paleozoic rift-related magmatism [11]. Far from being a period of tectonic quiescence, the Early Silurian represents a critical phase of thermal rejuvenation. The ca. 435 Ma magmatism in Liuyuan was not an isolated event but part of a macro-regional thermal episode along the southern CAOB margin. It correlates temporally with high-K calc-alkaline magmatism in the adjacent Dunhuang Block [40] and arc-related/adakitic intrusions in Eastern Tianshan [41], implying a coordinated geodynamic mechanism operating over a strike length exceeding 1000 km.

Crucially, this magmatic flare-up exhibits a remarkable temporal coupling with the metamorphic evolution of the adjacent Gubaoquan complex. The emplacement of these high-temperature granitoids is strictly synchronous with the retrograde cooling and rapid exhumation of high-pressure (HP) granulites and eclogites in the region (cooling ages ca. 430–428 Ma) [16,18]. Such a tight temporal link between vigorous magmatism and the rapid unroofing of deep-crustal rocks suggests a fundamental shift in lithospheric dynamics, potentially marking the transition from a compressive collisional regime to an extensional post-collisional setting.

5.2. Petrogenesis and Classification

The Liuyuan granitoids are characterized by high silica contents (SiO₂ > 71 wt.%) and elevated alkali levels, placing them firmly within the high-K calc-alkaline series (Figure 5c). Their mineral assemblages lack primary Al-rich phases such as muscovite and garnet, and their metaluminous to weakly peraluminous compositions (A/CNK < 1.20) argue against an S-type affinity [42]. Instead, several geochemical indicators point convincingly to an A-type origin. On the conventional 10,000 × Ga/Al–Zr discrimination diagram (Figure 9a) [43], all samples plot entirely within the A-type granite field. They show markedly elevated 10,000 × Ga/Al ratios relative to typical I- and S-type granites, a key characteristic of high-temperature A-type magmatism [43,44]. Collectively, their enrichment in high field strength elements (HFSEs) and high Ga/Al ratios supports the classification of the Liuyuan granitoids as A-type granites.

To refine their tectonic affinity, we further employ the Nb–Y–3 × Ga ternary discrimination scheme of Eby (1992) [45] (Figure 9b). All samples fall within the A₂-type field. According to Eby (1992) [45], A₁-type granites commonly evolve from OIB-like basaltic sources in intraplate rift or hotspot environments, whereas A₂-type granites reflect partial melting of continental crust or underplated crust previously modified by subduction-related magmatism. The A₂ classification aligns well with the “arc-like” features observed in other diagrams. For example, the samples lie within the Volcanic Arc Granite (VAG) field on the Rb vs. (Y + Nb) diagram (Figure 9c) [46] and exhibit pronounced negative Nb, Ta, and Ti anomalies—the typical “TNT” pattern—on spider diagrams (Figure 6a). These characteristics do not contradict their A-type nature; rather, they reflect the inherited subduction signatures expected from A₂-type magmas derived from arc-modified crustal sources [44,45].

Taken together, these observations indicate that the Liuyuan granitoids represent A₂-type granites emplaced in a post-collisional setting. Their formation was most likely driven by dehydration melting of a subduction-modified lower crustal source under enhanced thermal conditions [47], consistent with crustal heating associated with asthenospheric upwelling following slab break-off.

5.3. Crustal Architecture and Source

The isotopic compositions of the Liuyuan granitoids offer important insights into the deep crustal structure of the southern BOB and point to a petrogenesis dominated by reworking of ancient continental crust with a subordinate but meaningful input from mantle-derived mafic magmas. The whole-rock Sr–Nd isotopic data show a consistently evolved signature, with negative ε_Nd(t) values (−4.5 to −3.9) and Mesoproterozoic two-stage model ages (TDM2 ≈ 1.55 Ga). These features indicate that the parental magmas were derived mainly from Precambrian metamorphic basement—likely the concealed continuation of the Dunhuang or Beishan microcontinental blocks [17]—rather than from the juvenile Paleozoic accretionary crust typical of the northern Beishan [1,20].

5.3.1. Evaluation of Upper-Crustal Contamination

The potential for assimilation of the surrounding Ordovician–Silurian volcano-sedimentary host rocks during magma ascent and emplacement was evaluated. Field observations indicate generally sharp pluton–host contacts and a lack of abundant metasedimentary xenoliths within the main granitoid body. Moreover, samples used for whole-rock Sr–Nd and zircon Hf analyses were collected from the interior of the pluton, away from exposed contacts, to minimize localized wall-rock effects. The whole-rock Nd isotope compositions are relatively homogeneous (ε_Nd(t) = −4.5 to −3.9; TDM2 ≈ 1.55 Ga), which argues against substantial and variable upper-crustal assimilation. In addition, the granitoids are only weakly peraluminous (A/CNK = 1.03–1.20; mean = 1.10) and do not show strongly peraluminous/S-type characteristics expected from extensive incorporation of metasedimentary material. Significant contamination by heterogeneous host rocks would be expected to increase sample-to-sample isotopic scatter and/or produce systematic shifts in isotopic compositions and peraluminosity during differentiation. Therefore, although minor localized assimilation cannot be entirely excluded, the Sr–Nd–Hf isotopic characteristics of the Liuyuan granitoids are interpreted to primarily reflect source signatures and deep-crustal magma evolution rather than extensive upper-crustal contamination.

5.3.2. Hf–Nd Isotopic Decoupling and the MASH Model

A more complex picture emerges from the in situ zircon Hf isotopes. In contrast to the relatively uniform whole-rock Nd signatures, zircon Hf values show pronounced variability, with ε_Hf(t) spanning from strongly evolved (−6.3) to moderately radiogenic (+3.7) (Figure 7). Taken together, the whole-rock Nd isotopes define the dominant, integrated source signature of the magma, whereas the in situ zircon Hf isotopes potentially record finer-scale processes during melt generation, recharge, and crystallization [37,38]. Accordingly, the apparent mismatch between homogeneous whole-rock Nd isotopes and heterogeneous zircon Hf isotopes warrants more detailed evaluation.

The decoupling between heterogeneous zircon Hf isotopes and relatively homogeneous whole-rock Nd isotopes likely reflects the different sensitivities of these isotope systems to localized magma recharge, source heterogeneity, and isotopic homogenization during magma evolution [38]. We suggest that the zircon Hf variability may record (i) heterogeneity within the ancient crustal source and/or cryptic inheritance (if present), and (ii) minor, localized inputs of mantle-derived mafic magmas during MASH-type processes. Whereas whole-rock Nd isotopes represent a bulk, integrated signature of the magma batch that can be efficiently homogenized during storage and mixing, zircon Hf isotopes are acquired during discrete crystal-growth events and may therefore be more sensitive to small-volume mafic pulses or locally mixed melt domains [37]. Petrographically fresh textures (Figure 3) and uniformly low LOI values (≤1.75 wt.%; mean = 0.90 wt.%) make significant late-stage hydrothermal overprinting unlikely, and zircon Hf isotopes are generally robust against such alteration. Therefore, the Hf–Nd decoupling is most parsimoniously explained by a dominant ancient crustal component overprinted by spatially and temporally localized mantle-derived thermal input and minor chemical contributions.

To quantitatively test the feasibility of the MASH model, we performed a simple Nd isotope mass-balance calculation using the samples for which whole-rock Nd isotopes are available (n = 3; Supplementary Table S4). Using the most crust-dominated granite (sample 21LY-83; ε_Nd(t) = −4.6; Nd = 14.2 ppm) as the crustal reference endmember and a plausible range for contemporaneous mantle-derived basaltic/mafic melts (ε_Nd(t) = +6.0 to +8.0; Nd = 10–25 ppm) as the mantle endmember, the mixing results suggest that the Liuyuan granitoids can be produced with only a minor mantle-derived basaltic/mafic contribution. For example, reproducing the shift from 21LY-83 to the most radiogenic sample 21LY-51 (ε_Nd(t) = −3.9; Nd = 20.1 ppm) requires only ~3–9% mantle-derived input depending on the assumed mantle Nd concentration and εNd(t) (Supplementary Table S4). This modest input is consistent with a scenario in which underplated basaltic/mafic magmas provide the heat required for extensive crustal melting while introducing only a limited chemical signature into the resulting felsic melts.

This combination of features is most consistent with a MASH (Melting, Assimilation, Storage, and Homogenization) process operating near the crust–mantle boundary [48]. We infer that the underplating of hot mantle-derived basaltic magmas supplied the heat necessary to partially melt the overlying refractory basement [49]. Field evidence strengthens this interpretation: the granites are locally in direct contact with contemporaneous mafic–ultramafic bodies, including ultramafic rocks (Figure 3a) and gabbroic blocks/lenses (Figure 3b), supporting close spatial association between mantle-derived mafic magmas and the crustal source region. To minimize potential effects of localized contact-level processes, all samples selected for zircon Hf isotope analyses were collected from fresh granitoid domains in the pluton interior and strictly avoided the immediate contact margins with these mafic–ultramafic bodies (Figure 2). Therefore, the zircon Hf isotopic variability discussed here is interpreted to reflect magmatic-scale processes (e.g., localized mafic recharge and MASH-type interaction) rather than shallow-level wall-rock contamination during final emplacement. These mantle-derived mafic magmas likely acted both as the thermal driver and as a minor chemical contributor, potentially providing a radiogenic Hf component locally recorded by zircon. The Liuyuan granitoids can therefore be regarded as the differentiated products of an ancient microcontinental block, generated through extensive crustal melting induced by a focused input of mantle heat during the Early Silurian.

5.4. Geodynamic Model: Slab Break-Off

Integrating the geochronological constraints, the high-temperature A₂-type magmatic characteristics, and the regional metamorphic framework, we propose that the Early Silurian magmatism in the Liuyuan area is best explained by slab break-off during the transition from subduction to incipient collision/continent involvement (Figure 10b) [49]. The regional paleogeographic framework used to contextualize and place this cross-section model is summarized in Figure 10a. Continuous subduction alone does not readily account for the abrupt appearance of high-temperature felsic magmatism nor its temporal overlap with the rapid exhumation of high-pressure (HP) rocks documented in the region [18].

During the Late Ordovician, ongoing consumption and progressive narrowing of the Beishan oceanic domain (or the HNX branch; subduction polarity debated) beneath the Dunhuang–Beishan margin led to crustal thickening and deep burial of rocks to eclogite-facies conditions (>50 km), as constrained by the regional HP metamorphic record at ca. 467–450 Ma [12,14,16,18]. The involvement of buoyant continental material (microcontinental fragments/continental margin) in the subduction system marks an incipient collisional stage in southern Beishan. In such a setting, the strong buoyancy contrast between the incoming (micro)continental lithosphere and the dense, already-subducted oceanic slab would favor slab detachment at the ocean–continent transition (OCT; slab root), rather than within an actively subducting oceanic plate.

We interpret slab break-off in the Early Silurian (ca. 439–431 Ma) as a plausible trigger for the magmatic and tectonic changes observed in the Liuyuan region. Detachment at the slab root would have opened a slab window, enabling asthenospheric inflow and localized thermal perturbation beneath the overriding lithosphere (Figure 10b). The resulting increase in basal heat flux could facilitate partial melting of the lower crust and the generation of high-temperature A₂-type felsic magmas, potentially accompanied by limited input of mantle-derived melts [38,49] (see Section 5.3 for magma evolution processes). In addition, slab break-off and related tectonic reorganization (including reduced slab pull) could promote uplift and contribute to the fast cooling and exhumation of the Gubaoquan eclogites, which record cooling ages around ca. 430–428 Ma [16,18], broadly overlapping with the Liuyuan granite emplacement. Together, these processes signal a fundamental tectonic shift in the southern CAOB during the Early Silurian, marking the transition from subduction/incipient collision to post-collisional extension [11,12,13,17].

Although back-arc extension remains a possible broader context for the Early Silurian evolution of southern Beishan [7,14,15], back-arc extension alone does not readily explain the short-lived, high-temperature A₂-type magmatic pulse in Liuyuan and its near-synchronous association with rapid HP-rock exhumation within a narrow ~20 Myr window. We therefore suggest that slab break-off at the OCT acted as a local trigger superimposed on a regional extensional regime. This staged evolution is broadly consistent with the framework proposed by Saktura et al. (2017; Ref. [16]; their Figure 10c–e) [16], in which continent involvement and subsequent slab-root break-off together account for an Early Silurian thermal pulse (Liuyuan A₂-type magmatism) and the terminal-stage exhumation/cooling of regional HP rocks.

6. Conclusions

(1)

Zircon U-Pb dating identifies a distinct Early Silurian magmatic pulse (439–431 Ma), bridging the gap between Ordovician arc construction and Late Paleozoic rift-related magmatism in southern Beishan.

(2)

The Liuyuan intrusions are aluminous A₂-type granites with high SiO₂–alkalis and elevated Ga/Al ratios. Negative Nb–Ta–Ti anomalies are consistent with melting of a previously subduction-modified crustal source, potentially inherited from arc-related subduction prior to slab break-off, and therefore do not necessarily require syn-emplacement arc-related subduction.

(3)

Whole-rock Sr–Nd isotopes indicate a dominant contribution from ancient Precambrian basement, whereas zircon ε_Hf(t) values suggest only limited juvenile input. Collectively, the isotopic data support predominantly crustal magma sources, with a subordinate mantle component at most.

(4)

The A₂-type magmatism records a transition from subduction to incipient collision/continent involvement and subsequent post-subduction extension. Slab break-off at the slab root (ocean–continent transition) is a plausible trigger, providing slab-window-related thermal input for crustal melting.

(5)

The Silurian granitoids mark an early post-subduction thermal pulse (slab break-off), whereas younger granitoids likely reflect later tectono-magmatic stages and do not necessarily require ongoing Silurian break-off; the Silurian event may have precondi-tioned the crust for subsequent re-melting.