Fluid inclusion experiments were conducted in Research Institute of Petroleum Exploration Development. Linkam THMSG600 heating–freezing stage was prepared under a petrographic microscope and used for microscopic temperature measurement of fluid inclusions. The temperature measurement error was ±0.2 °C at below 50 °C and ±2 °C at above 100 °C. The freezing temperature and inclusion homogenization temperature were measured at a heating rate of 0.1 °C/min. Salinity of the NaCl-H₂O inclusion was calculated according to the freezing temperature of two-phase inclusions. The composition of the selected fluid inclusions was identified using an RM-2000 laser Raman microprobe containing an argon ion laser with a 514 nm laser source. The spectral scanning range was 100–4,300 cm^-1, the accumulation time of each scan was 60 s, and the spectral resolution was ∼0.14 cm⁻¹. After temperature measurement, the fluid inclusions were observed under a back scattering electron microscope (BSEM) and scanning electron microscopy-cathode luminescence (SEM-CL). The BSEM and SEM-CL images of crack cement inclusions were recorded by a Quanta FEG 650 SEM device equipped with an energy dispersion spectrometer in Peking University. The accelerating voltage was 15 kV, and scanning time was 45 s.

C-O isotopes were tested on a Thermo-Finnigan MAT 253 isotope mass spectrometer in Guangzhou Institute of Geochemistry, Chinese Academy of Sciences. At each time, a sample (about 0.1 mg) was placed into a single reaction vessel and reacted with 1.92 g/cm³ H₃PO₄ at 90 °C ± 1 °C. Then the generated CO₂ was collected, purified, and transferred to a mass spectrometer. The international carbonate standard sample NBS-19 and the China national standard sample GBW04405 were repeatedly measured for comparison. The precision of δ¹³C and δ¹⁸O was ±0.06‰(1σ) and ±0.08‰(1σ) respectively.

As for Sm-Nd isotopic dating, pure calcite was manually sorted under binocular microscopy, and crushed in an agate mortar to 200 meshes. The isolated calcite was sent to detection of rare earth element (REE) concentrations on an ELAN 6000 inductively-coupled plasma mass spectrometer (ICP-MS) in Guangzhou Institute of Geochemistry, CAS. The Sm and Nd concentrations and isotope composition of the calcite were analyzed. Specifically, each sample was dissolved in a mixture of HF and perchloric acid (10: 1) in a polytetrafluoroethylene container (PTFE vessel), which was maintained below 150 °C for at least 12 h until the whole sample was decomposed. Two portions of the same sample (each about 150 mg) were dissolved separately, and one portion was used for Sm and Nd concentration spiking and measurement, and the other was for detection of ¹⁴³Nd/¹⁴⁴Nd ratio. When Sm and Nd were separated through reversed phase extraction, one key step was to coat di (6-methylheptyl) phosphate on polytetrafluoroethylene powder. The isotope ratio was measured on an IsoProbe T thermoionization mass spectrograph in Tianjin Geology Survey Center. The Nd ratio was normalized to the ¹⁴⁶Nd/¹⁴⁴Nd ratio, which was about 0.7219. The reproducibility of isotope ratio measurement was higher than 0.005% (2σ), and the precision of Sm and Nd concentration detection was smaller than the cited value of 0.5% (2σ). Experimental measurements showed the BCR-1 concentrations of Sm and Nd were 6.57 and 28.75 ppm respectively, and the ¹⁴³Nd/¹⁴⁴Nd ratio was 0.512644 ± 5 (2σ, n = 6). Repetitive analysis with the Johnson and Mattey® standard sample (JMC) showed the ¹⁴³Nd/¹⁴⁴Nd ratio was 0.511132 ± 5 (2σ, n = 6). The blank sample weights of Sm and Nd were 0.03 and 0.05 ng respectively. The attenuation constant (λ¹⁴⁷Sm) for age calculation was 6.54 × 10⁻¹²/year. The Sm-Nd isochrone ages were computed on ISOPLOT 2.9.

As for Nd isotope, the powder samples were first labeled using a mixed isotopic tracer, then dissolved in HF and HNO₃, and finally Nd and REEs were isolated. Isotopes were detected on an thermal ionization mass spectrogram in Tianjin Geology Survey Center. After measurements and according to BCR-1 and La Jolla Nd isotopic standards, the ¹⁴³Nd/¹⁴⁴Nd ratios were 0.512663 ± 9 (2σ) and 0.511862 ± 7 (2σ) respectively.

REEs were analyzed by ICP-MS, which can be used to detect sample solutions digested by HF + HClO₄ in closed containers. Quality control was ensured by using GBPG-1 and GSR-1. The relative standard deviations of trace element concentrations (including REEs) were smaller than 10%. The sample preparation, instrumental operation and calibration procedure proposed by Qi and Grégoire (2000) were adopted. The testing results of REEs were listed in appendix C, in which no valid results were obtained from samples TF-72 and TF-157 in group B and sample TF-101 from group C.

Totally 11 rock cores at different depths were selected from wells J32, J36, J174 and J176 for acoustic emission experiments. Uniaxial loading was induced by an RMT-159C rock mechanical testing system developed by Institute of Rock and Soil Mechanics, CAS, and acoustic emission was tested on the SAEU2S full-waveform track acoustic emission detector. This detector can acquire and record the value of acoustic emission, energy, number of rings, rise time, and usable voltage, and automatically generate interparameter relationship charts. All tests were conducted using an SR150Msyntony acoustic emission probe, and its displacement loading rate was 2 × 10⁻³ mm/s, center frequency was 60–400 kHz, sampling frequency was 5,000 Hz, parameter interval was 50, filter frequency was 20–100 kHz, waveform threshold was 40 dB, and gain of preamplifier loaded on two sides of a sample was 40 dB. Tests were started only when the parameters of the acoustic emission system were correctly calibrated. Each test was conducted after two repeated loadings. The sample was continually reloaded until it was completely destroyed.

It is critical to characterize chemistry of fracture fillings in order to determine fluid activity episodes and corresponding paleoenvironments (Kerrich et al., 1984). When fluids enter fractures induced by tectonic stresses, a certain amount of crystalline minerals will precipitate and deposit on the fracture wall (Wendler et al., 2012; Griffiths et al., 2016). Therefore, analysis of rock fracture history can promote effective identification of fluid activity episodes. In addition, geochemical and isotopic characteristics of fracture fillings in the core can further constrain division accuracy of fluid activity episodes.

Using the imaging logging data, we first obtain strike and inclination information of fractures in each well, and then plot corresponding rose diagrams showing the strikes of plane tectonic fractures (Figure 3a) and those of all the fractures (Figure 3b). There are mainly three sets of fractures in the study area, with NEE (45°–63°), NNW (317°–345°) and near NS (0°–27°) strikes, respectively (Figure 3c). Their distributions conform to certain rules. Specifically, the NEE fractures are mainly in the central and southern parts, the near NS ones are primarily in the northern and eastern parts, and the NNW ones are extensively distributed throughout the sag, though they are of low degree of development. According to their morphologies and crosscutting relationships, near NS fractures predate NEE ones, which further predate NNW ones. Dominant dip direction are in the range of 270°–360°.

When subjected to paleo-tectonic stress, corresponding memory will be recorded in the tight, brittle rocks (Segreto et al., 2015). In general, most of the underground rock formations are subjected to tectonic stress fields during Earth’s history, accompanied by development of microscopic or concealed micro-fractures (Li et al., 2025; Luo et al., 2025). When the applied stress reaches the paleo-stress field strength, micro-fractures will begin to expand, resulting in sharply increasing quantity and intensity of acoustic emission. Accordingly, the stress applied at this point is supposed to be equivalent to the ancient stress field strength. Therefore, the number of Kaiser effect points appearing on the acoustic emission curve can be used to judge the minimum stress episodes and stress field strengths that the rock layer has experienced. Kaiser effect points in the acoustic emission curve correspond to those positions with stepwise sudden increases. However, the occurrence order of those steps does not necessarily indicate the sequential order of Kaiser effect points. All the other sub-figures in Figure 4 have four Kaiser effect points, which indicate at least four fracturing episodes. All the four Kaiser effect points in each sub-figure have a certain range of stress field strength variation, which demonstrates the corresponding fractures to be products of tectonic activities. However, we only observe three sets of fractures in the core, which might be attributable to the fact that two episodes of tectonic events result in the fractures with the same occurrence features.

Calcite fillings in the near NS tensile fractures have an average C-O isotope value of −5.43 and 10.23, an average ɛNd(t) value of 8.79 and an average homogenization temperature of 165–195 °C. In comparison, those in the NEE shear fractures have an average C-O isotope value of −4.60 and 16.07, an average ɛNd(t) value of 3.91 and an average homogenization temperature of 125–150 °C. What’s more, those in the NNW shear fractures are featured by an average C-O isotope value of −0.93 and 21.12, an average ɛNd(t) value of −2.14 and an average homogenization temperature of 75–105 °C (Figure 5; Table 1). Large differences in geochemical characteristics are seen among calcite fillings in three set of fractures, while homogeneity is observed within a certain set of fractures. Therefore, we suggest that there are mainly three independent episodes of fluid activities, given the distinct differences between the sets, no clear evidence for fluid mixing is observed. The Sm-Nd isochron ages of NW tensile fractures, NEE shear fracture and NNW shear fracture are 272.4 Ma, 189.5 Ma and 124.3 Ma, respectively (Figure 6; Table 2).

Difference in fluid origins is speculated to be responsible for the difference in geochemical characteristics of calcite fillings. As shown in the burial history chart (Figure 7), the Permian Lucaogou Formation is at the deepest burial depth and has the highest burial temperature at this moment. Moreover, fluid inclusion homogenization temperatures of calcite fillings in near NS and NEE fractures are both higher than the current burial temperature, which implies a higher fluid temperature of the precipitated calcite than that of the surrounding rocks, and this is a typical feature of deep fluids (Machel and Lonnee, 2002; Davis and Smith, 2006).

In the high temperature environment, Eu³⁺ is reduced to Eu²⁺, resulting in an increasing Eu^2+/Eu³⁺ ratio, while an equilibrium is reached at 250 °C (Sverjensky, 1984; Bau and Möller, 1992; Cai et al., 2008; Zhu et al., 2013). Since the ionic radius of Eu²⁺ (0.117 nm) is larger than that of Eu³⁺ (0.095 nm), it is more difficult for Eu²⁺ to enter the rock-forming minerals than Eu³⁺ (Cai et al., 2008). Therefore, Eu²⁺ is the main form of Eu in the fluids at higher temperatures. Similarly, in the process of condensation and crystallization of hydrothermal fluids, Eu²⁺ is converted to Eu³⁺, which further replace Ca²⁺ into carbonate minerals due to its close ionic radius to that of Ca²⁺ (0.1 nm) (Zhu et al., 2013). In this case, the calcite fillings present a positive anomaly of Eu. This interpretation is based on the premise that the differentiation of Eu valence states is predominantly controlled by high-temperature reducing conditions. Although factors such as fluid–rock interaction may also influence the characteristics of the Eu anomaly, they do not alter the above assessment. Therefore, fluids corresponding to the calcite fillings in near NS and NEE fractures should be deep hydrothermal fluids according to their positive Eu anomalies (Figure 8; Table 3), which are in line with the high fluid inclusion homogenization temperatures.

As shown in Figure 9 and Table 4, calcite fillings in the near NS tensile fractures fall in the zone representing original mantle carbonates, which indicate relation of their origin to mantle fluids. In comparison, those in the NEE shear fractures present a transition from the original mantle carbonate zone to the sedimentary carbonate zone, reflecting the trend of fluid temperature reduction. Comparably, carbonates in the shear zone of the southern India show typical features of original mantle carbonates (δ¹⁸O(SMOW) = 7.3‰ to 8.5‰, δ¹³C(PDB) = −5.0‰ to −4.7‰; Wichman et al., 1994). Meanwhile, calcite veins in the Dongying Sag, eastern China are speculated to be mantle fluids (δ¹⁸O(SMOW) = 10.7‰ to 13.1‰, δ¹³C(PDB) = −7.9‰ to −7.3‰). Sharing similar C-O isotope features with above two examples, calcite fillings in the NEE shear fractures in the Jimsar Sag could also be related to mantle fluids. In addition, the narrow ranges of εNd(t) and C-O isotope values of the near NS tensile fractures and NEE shear fractures suggest small effects of assimilation of surrounding rocks and formation water.

Compared with calcite fillings in the NEE and NNW shear fractures, those in the near NS tensile fractures have the most depleted Nd isotopic compositions (εNd(t) = +6.44∼+10.19) (Figure 10). Moreover, recent studies suggest that the Junggar Basin and the surrounding Junggar Terrane share a common tectonic evolution from Carboniferous subduction–accretion process to Permian amalgamation, further confirming the relevant magmatism events in the north part of Xinjiang (Li et al., 2015). These magmatism events promote invasion of massive basalt magnesia-ultra-magnesium complexes in the upper crust. The Nd isotopic compositions of Permian mantle sources in the northern Xinjiang are represented by these Permian mafic rocks (Xiao et al., 2008; Gao and Zhou, 2013). These mantle sources can also account for the extremely depleted Nd isotopic compositions of the calcite fillings in the near NS tensile fractures in the Jimsar Sag based on the following evidence. (1) The Permian magnesia-ultra-magnesium complexes in the northern Xinjiang are generally featured by high εNd(t) values ranging from +6.2 to +8.5. (2) The εNd(t) values of mafic dykes of the Permian Beishan-Tianshan Region are +5.5 to +7.5 (Xue et al., 2016). (3) Olivine from the Tianshan massive unit also has high εNd(t) values, ranging from +6.0 to +7.5 (Gao and Zhou, 2013). (4) Basalts from the Junggar Basin have high εNd(t) values, ranging from ∼+6.4 to +8.0 (Li et al., 2015). All these suggest that the high εNd(t) values (from 6.44 to 10.19) of calcite fillings in the near NS tensile fractures are ascribed to thermal fluid activity induced by the mantle magmatism. This is further supported by the flatter partitioning of rare earth elements.

Calcite fillings in the NEE shear fractures are characterized by moderate to high positive εNd(t) values, indicating significant contribution from mantle-like components. However, εNd(t) values of calcite fillings in the NEE shear fractures are slightly different from those in the near NS tensile fractures and mafic rocks in the northern Xinjiang, which indicates different fluid origins of calcite fillings in two sets of fractures. This is further demonstrated by the difference in the C-O isotope system (Figure 9). In terms of partitioning of rare earth elements, calcite fillings in the NEE shear fractures are characterized by light rare earth element enrichment, which is a typical feature of crustal fluids. Therefore, we infer that calcite fillings in the NEE shear fractures are juvenile crustal-derived in nature, but also present mantle-like isotopic characteristics. Calcite fillings in the NEE shear fractures share similar εNd (t) values (from −0.52 to 6.27) with granites in the northern Xinjiang, which are speculated to be mainly-derived from lower continent crust (Zhang et al., 2018). The basement beneath the Junggar Basin has been interpreted as a complex, from sedimentary unit to volcanic unit (Li et al., 2015), and basalts beneath the Junggar Basin are speculated to be generated by partial melting of the metasomatized wedge mantle by hydrous fluids and melts of subducted sediments. Therefore, the relatively low εNd(t) values of calcite fillings in the shear fractures are likely products of the lower crust magmatism, which means their origin is related to magmatism of mixed basalts and sedimentary basement.

As for calcite fillings in the NNW shear fractures, they fall in the zone representing sedimentary carbonates. In conjunction with the negative εNd(t) values and low fluid inclusion homogenization temperatures, we suggest them to originate from basin fluids in the Jimsar Sag.

As direct products of regional tectonic activity, tectonic fractures provide migration channel and reservoir space for deep fluids. As demonstrated in the acoustic emission curve (Figure 4), there are mainly four episodes of tectonic activities. Meanwhile, according to different fluid inclusion homogenization temperatures of calcite fillings in the three sets of fractures and homogeneous geochemical features within one set of fractures, we infer that there are mainly three episodes of tectonic-fluid filling events, accompanied by no occurrence of fluid mixing. Three sets of fractures are filled by fluids at 272.4, 189.5, and 124.3 Ma, respectively, based on analysis of Sm-Nd isochron ages, which are consistent with crosscutting relationships among them recognized in the field outcrops.

As an important part of the Central Asian Orogenic Belt, the northern Xinjiang is characterized by development of the Paleozoic orogenic belt that dominantly consists of continental crust accretion. During the Late Carboniferous, collision among continents of the Junggar Basin and folding orogenic movement become intensified, which result in development of fault basin and intracontinental rift, accompanied by widely distributed, frequently occurring magmatic and volcanic activities. Subsequently, the post-collision mantle-derived magmatism is extensively developed in the northern Xinjiang during the Permian, which marks the entrance of the Junggar Basin into the post-collision extensional period. At the same time, the emergence of the Santai Uplift in the Jimsar Sag leads to different episodes of tectonic stresses, associated with two episodes of fracturing events (Liu et al., 2019). According to the Sm-Nd isochron ages of calcite fillings, the Jimsar Sag is under a tensile tectonic setting before 272.4 Ma, after which NS extrusion generates a large amount of NS tensile fractures. Then, mantle magmatism-related fluids enter these tensile fractures, forming calcite fillings with high εNd(t) contents, high δ¹⁸O values and high fluid inclusion homogenization temperatures. At 189.5 Ma, the compressive stress from the southwest to the northeast leads to formation of NEE shear fractures, and crustal magmatism-related fluids invade these fractures, forming calcite fillings with second highest εNd(t) contents, second highest δ¹⁸O values and second highest fluid inclusion homogenization temperatures. During the Early Cretaceous, the Bogda Piedmont Foreland Depression experiences persistent foldbacks under the tectonic stress (Shen et al., 2006; Tang et al., 2015). Later at 124.3 Ma, the Fukang Fault Belt in the southern part of the Jimsar Sag becomes intensively active, exhibiting strong overthrust features from south to north. During this period, NNW shear fractures emerge and basin fluids enter these fractures, forming calcite fillings with low εNd(t), low δ¹⁸O values and low fluid inclusion homogenization temperatures.