Introduction

CO2 is one of the most common inorganic gas species in natural gas fields with significant economic implications1. It serves as a valuable tracer for understanding the formation and evolution of natural gas fields due to its close association with hydrocarbon gases2,3. Studies have shown that CO2 circulation in reservoirs impacts preservation conditions, fluid compositions, and physical properties, influencing the recovery and economic value of hydrocarbons4,5,6,7. High CO2 concentrations have been observed in various regions, such as the Changqing gas field in the Ordos Basin and the Tazhong area of the Tarim Basin, China. CO2 dissolution of carbonate minerals like calcite increases reservoir porosity but may also lead to acidification, altering reservoir permeability and impacting oil and gas recovery rates8. In the Norwegian North Sea Basin, CO2-induced dissolution forms secondary pores and fractures, enhancing reservoir storage and permeability, though precipitation of new minerals can complicate migration and enrichment processes9. Identifying the source of CO2 is critical for understanding natural gas genesis, migration, and enrichment mechanisms. It also aids in optimizing extraction and processing technologies while predicting long-term reservoir stability and natural gas economic value10,11,12.

Authigenic minerals such as calcite, quartz, and dolomite can trap CO2 during precipitation, with geochemical and isotopic analyses providing insights into the genesis of CO2 and its links with hydrocarbons. Carbon isotopic (δ13C) and strontium isotopic compositions are used to discern hydrothermal fluid sources13, while oxygen isotopic (δ18O) compositions estimate fluid temperatures during mineral formation14. Isotopic data combined with rare earth element (REE) models help delineate fluid pathways, diagenetic alterations, and water‒rock interactions15.

CO2 and hydrocarbons often coexist in sedimentary basins and are trapped during mineral recrystallization16. Fluid inclusions (FIs) provide valuable information about multi-phase fluid occurrences and evolutionary histories. Techniques such as laser Raman spectroscopy (LRM) enable non-destructive, high-precision analysis of inclusion composition, density, and chemical properties17,18. Coupled with microthermometry and thermodynamic modeling, LRM reconstructs trapping pressure and temperature conditions, offering a comprehensive view of reservoir evolution. The Huangqiao field in the Lower Yangtze region contains vertically layered CO2 and hydrocarbon reservoirs within sandstone and carbonate formations from the Carboniferous to Cretaceous. Analyses of fluid inclusions through petrography, Raman spectroscopy, and fluorescence spectrum studies revealed detailed inclusion distributions, compositions, and hosting mineral characteristics. Quantitative data on CH4 and CO2 phases, including density, pressure, and temperature, were derived through thermodynamic simulations. Elemental and isotopic indicators (REEs, δ13C, δ18O, strontium isotopes) provided insights into diagenetic environments, while U–Pb dating of calcite and fluid inclusion homogenization temperatures calibrated the thermal evolution history of the field. These findings elucidate the basin’s abnormal pore pressure recovery and the interplay between organic and inorganic fluids, offering guidelines for quantitative fluid studies and enhancing the understanding of reservoir evolution processes.

Geological setting

The Huangqiao gas field, located in the Lower Yangtze region, is a significant CO2-dominated gas field with the largest proven CO2 reserves in China19. Geographically, it is located at the northern junction of the Jiangsu sag and the southern boundary of the Jiangsu uplift within the Nanjing depression of the Lower Yangtze basin20 (Fig. 1A). This region is characterized by well-preserved marine and Paleozoic strata, while the overlying continental Cenozoic strata possess substantial sedimentary thickness, making it conducive to the exploration of oil and gas. The eastern boundary of the Huangqiao area is marked by the Qidong salient, and to the north, it is bordered by the Qintong. Taizhou is located in the western portion of this depression, while Gushan lies to the south21,22. The exploration area of the Huangqiao oil and gas field covers approximately 1896.7 km2. To the north, the lower Yangtze region’s boundary adjoins the North China plate along the Lianyungang-Yellow Sea and Tanlu faults, while the southern boundary separates it from the South China plate by the Jiangshan‒Shaoxing fault belt. Four stages of tectonic movements have affected the Huangqiao area: (1) a stage of steady uplift before the Indochina movement; (2) during the late Indo-Chinese to middle Yanshan compressional nappe stage, the strata underwent significant tectonic deformation; (3) a late Yanshan-early Himalayan tensile rifting stage; and (4) a late Himalayan uplift and denudation stage23.

Fig. 1
Fig. 1
Full size image

(A) Location of the Huangqiao gas field in the Subei basin of the Lower Yangtze Platform; (B) Division of the Huangqiao gas field and well location; (C) Stratigraphic column of the Paleozoic to the Cenozoic in Well S174.

The upper assemblage in the Huangqiao gas field comprises the Devonian, Carboniferous, Permian, and Triassic formations (Fig. 1C). The stratigraphy includes the Gaolishan Formation (C1g), Huanglong Formation (C3h), Chuanshan Formation (C3c), Qixia Formation (P1q), Gufeng Formation (P1g), Longtan Formation (P2l), Dalong Formation (P2d), Qinglong Formation (T1q), Pukou Formation (K2p), Yancheng Formation (Ny), and Dongtai Formation (Qd). Notably, P1q represents a marine carbonate deposit that is primarily composed of bioclastic limestone and serves as the main CO2 production layer in the Huangqiao region; P1q is widely distributed in the working area. P1g and P1q are in disconformable contact, with the lower portion being dominated by yellow shale and the upper portion characterized by black siliceous shale. P2l, deposited in a marine-continental transition zone with coal-bearing formations, comprises a lower section of black mudstone with a coal seam, a middle section of black mudstone with gray quartz sandstone, and an upper section predominantly consisting of lenticular limestone with a thickness exceeding 200 m, serving as a significant tight sandstone reservoir. The lithology of P2d is primarily gray‒black siliceous rock and purple mudstone with thicknesses ranging from 35 to 55 m. T1q is extensively distributed across the central and southern areas of the Huangqiao region and is characterized by black mudstone and argillaceous shale in a deep-water sedimentary environment influenced by denudation during the Indosinian period. K2p is extensively developed, with a thickness varying from 700 to 3500 m24,25.

P1g, P2l, and P2d mudstones primarily serve as superior hydrocarbon source rocks in the Huangqiao region, with thicknesses ranging from 120 to 514 m and burial depths ranging from 1600 to 3500 m. Notably, P2l has a high organic content, with the total organic carbon contents (TOC) ranging between 2 and 3%26. The Huangqiao area primarily features two reservoir systems: T1q and P1q carbonate reservoirs (marked by diminished porosity and permeability) and P2l clastic reservoirs (also displaying low porosity and low permeability). Argillaceous rocks from the P2l-P2d, K2p, and Ny Formations serve as the main regional cap rocks26,27.

The Huangqiao CO2 gas field can be vertically segmented into a deep marine gas field and a shallow continental gas field. The deep marine gas field is buried at depths ranging from 1800 to 2300 m, with over 90% CO2 content, accompanied by a small amount of hydrocarbon gas and condensate oil28. The CO2 gas-bearing layers include the C3c, C3h, P1q, and K2q Formations. In contrast, the shallow continental gas field primarily resides in the Neogene Yancheng Formation, with the gas reservoir buried at depths ranging from 371 to 378 m. This gas is predominantly composed of N2 (more than 56%), along with varying proportions of CO2, He gas, and small quantities of hydrocarbons29,30. Shallow gas reservoirs are derivatives of deep gas reservoirs. The δ13CCO2 values in the Huangqiao area closely resemble those of magmatic origins, indicating that the CO2 in the Huangqiao area originated from deep mantle. The intense tectonic activity during the Himalayan period gave rise to deep and extensive lithospheric faults along the Yangtze River, which serve as the primary conduits for the upwelling of mantle material. The Huangqiao region is in close proximity to these fault zones31,32. However, previous research on the migration-accumulation patterns and source discrimination of CO2 and hydrocarbon gas in the Huangqiao gas field has not been conducted systematically.

Samples and methods

Forty core samples, encompassing both carbonate and clastic rocks, were systematically collected from various boreholes, including S174, X3, X2, X1, N6, N9, and N13 (Fig. 1B). These boreholes span multiple formations, namely, C3c, P1q, P2l, P2d, and T1q (Fig. 2).

Fig. 2
Fig. 2
Full size image

Continuous well profile with well logging interpretation and sampling point labeling.

Petrological examinations were conducted utilizing an Olympus BX60 microscope. The homogenization temperature (Th) and ice melting temperature (Tm) of aqueous fluid inclusions were precisely determined using a Linkam (TH–600) heating–cooling stage, with a reproducibility of ± 1 °C for Th and ± 0.1 °C for Tm. Microthermometric analyses were carried out on saline aqueous inclusions coentrapped with gas or oil inclusions. The salinities of the inclusion fluids were calculated from the ice final melting temperatures (Tm-ice) following the methodology described by Bodnar (1993)33. Cathodoluminescence (CL) analysis was employed to discriminate between quartz and calcite and identify distinct cement generations. Polished thin sections were analyzed utilizing a CL8200 MK5 microscope with a beam voltage set at 17 kV and a beam current of 500 μA. Imaging procedures were standardized using a Leica DM2500 microscope, with uniform exposure time and gain set at 70 ms and 1.005 s, respectively. All these analyses were conducted at the Petroleum Geochemistry Laboratory of Yangtze University.

The instrument used for rare earth element (REE) analysis of calcite was a laser ablation inductively coupled plasma mass spectrometer (LA‒ICP‒MS) with an ARF excimer laser ablation device (operating wavelength: 193 nm, maximum energy: 200 mJ) and a MicroLas optical apparatus. An Agilent 7700e ICP‒MS instrument was used to measure the ion signal strength, and helium served as the carrier gas. The laser beam spot diameter used was 90 μm. The synthetic silicate glass standard sample BIR-1G was used for calibration during mineral trace element content processing. The sample ablation method adopts single-point ablation, which includes 20–30 s of background collection and 50 s of data collection. The software ICPMSDataCal (a developed Excel-based software tool) was used to complete the offline processing of the analysis data, including background subtraction, time offset correction, and quantitative standardization of the REE analysis34. We adopted the method standardized by the North American shale standard (NASC). The experiment was completed in Wuhan Shangpu Analytical Technology Co., LTD.

Carbon and oxygen isotopic analyses were performed at the China University of Geosciences (Wuhan). Sample collection consisted of micro-drilling followed by pulverization to 200-mesh size and subsequent chemical treatment with 100% phosphoric acid at 25 °C. Upon completion of the reaction, the carbon and oxygen isotopes of the liberated CO2 were analyzed using a MAT253 gas stable isotope mass spectrometer. Carbon isotope results and oxygen isotope results were normalized according to VPDB standards and SMOW standards. The test accuracies are δ13C < 0.01‰ and δ18O < 0.02‰.

Strontium isotopic analysis was conducted utilizing a Nu Instruments multi-collector inductively coupled plasma mass spectrometer and was completed at Wuhan Shangpu Analysis Technology Co., Ltd. Sample collection involved the use of microdrills, followed by crushing to 200 mesh. Subsequently, centrifugation, purification, and separation procedures were applied. Calibration of the Sr isotope test data was performed with reference to the 86Sr/88Sr ratio of 0.1194. The data quality was monitored using the international standard sample SRM-987.

The analysis of oil inclusion maturity was conducted through the application of ultraviolet (UV) light irradiation, employing fluorescence spectroscopy facilitated by a Nikon 80I microscope outfitted with a high-pressure mercury lamp, spanning an emission spectrum of 300–380 nm and was completed at China University of Geosciences (Wuhan). Spectroscopic examination of hydrocarbon inclusions revealed the λmax value, which is indicative of the wavelength at peak intensity, along with the QF535 metric, which quantifies the integral area ratio spanning 535–750 nm to 430–535 nm (Mounds, 2001). The selection of characteristic hydrocarbon inclusions for microscale thermovolumetric evaluation was executed utilizing a confocal laser scanning microscope, adhering to proven techniques for compositional and pressure‒temperature (P‒T) entrapment condition reconstruction35,36. The determination of the C7+ molar proportion within the oil inclusions was performed through the assessment of parameters such as Th and the degree of bubble occupancy (Fv). Furthermore, quantitative associations between the C7+ molar fraction, Thoil, and Fv were employed to infer the minimal entrapment pressure.

Raman spectroscopic analysis was performed using a Horiba LABRAM HR800 apparatus equipped with a green laser at a wavelength of 532 nm, delivering output powers between 400 and 500 mW.  The instrument is stationed at the China University of Geosciences (Wuhan). For calibration purposes, a polished silicon standard exhibiting a Raman shift near 520.7 cm−1 was utilized. The characterization of the gaseous inclusions within the geological samples involved the use of a 200-µm confocal aperture coupled with a 300-g·mm−1 diffraction grating centered at 2650 cm−1. To enhance the signal clarity relative to background noise, the collection duration was finely tuned, ranging from 0 to 500 s, ensuring optimal detection conditions. The accurate ascertainment of Raman shifts for discrete gaseous inclusions was achieved through the utilization of an 1800 gr·mm−1 grating alongside a 50-μm aperture, ensuring refined spectral resolution. Calibration corrections facilitated by neon illumination sources were systematically applied to validate the precision of Raman spectral positions, in line with protocols established in the literature37,38,39,40. The methane (CH4) inclusion density was predicated on the Raman shift observed in the C–H symmetric stretch vibration (v1) of CH4 under ambient conditions (25 °C, ensuring phase homogeneity). For pure carbon dioxide (CO2) inclusions, density calculations were predicated on the analysis of the separation between the Fermi diad peaks associated with CO2. The quantitative determination of CO2 inclusion densities incorporated the application of multiple regression analyses, as delineated in extant research40. The relative molar concentrations of the constituent gases were discerned by comparative analysis of the integrated peak intensities for CO2 (~ 1280 cm−1 and 1380 cm−1) and CH4 (~ 2917 cm−1), employing the method proposed by Placzek et al. (1934)41, which considers the Raman cross-section, integrated peak intensity, and instrument-specific detection efficiency.

Results

Petrographic characteristics

P1q and T1q are the primary carbonate reservoirs, and our sampling primarily targeted pore/hole/fracture filling or partial filling within gas or oil layers (Fig. 3A). These veins are predominantly composed of calcite and formed as a result of secondary diagenetic fluid activity. Based on their occurrence characteristics, we classified the core observations of joint-hole veins into two main types:

Fig. 3
Fig. 3
Full size image

Typical photos of carbonate and sandstone core samples in this study. (A) Multitypes veins in Carbonate fracture; (B,C) Oilbearing sandstones; (D) Carbonate rock with Corrosional hole; (E) Carbonate rock with structural calcite veins; (F) Carbonate rock with karst and structural fractures.

Filling of karst, sutured, or interbedded fractures (Fig. 3D,F)

These veins exhibit smooth-face fracture orientations ranging from subhorizontal and bedding parallel to oblique 30°. In addition, visible small wrinkles and folds can be observed, and these veins may exhibit mixed characteristics with the surrounding rock. Vein widths in this category range from 7 to 9 mm.

Filling in structural fractures with straight edges occurs in a coplanar manner (Fig. 3E): These veins are characterized by vertical or near-vertical orientations and do not show evident parallels to the fractures. The vertical persistence length of these fracture veins ranges from 2 to 6 cm. Incomplete filling can be observed in some of the larger veins.

In addition, the other two samples in Fig. 3 are sandstone samples in the P2l reservoir. Oil-bearing sandstone (Fig. 3 B and C): B. The core is coarse sandstone with a particle diameter of 1–0.5 mm. The rock surface is yellow–brown. There are cracks with an inclination of 30° to 60° on the surface. The length is 5–7 cm, but there is no mineral filling in the cracks. Obvious black oil stains can be seen on the surface of the sandstone. C. The core is fine sandstone with a particle diameter of 0.25–0.1 mm. Obvious black oil stains can be seen on the surface.

Typical petrographic studies revealed that relatively wide calcite veins (with widths of approximately 3 cm, referred to as Cal-vein 1, Fig. 4) exhibit a bright red cathodoluminescence color. These veins are characterized by fibrous crystallization and suitable transparency. Fluorescence images of these veins show the presence of fluid inclusions without excitation fluorescence. However, a significant number of gas inclusions are also present, as indicated by the Raman characteristic peak, which suggests the presence of gas inclusions containing CO2 and CH4. Two conjugate intersecting veins connected to a wide vein (with a width of approximately 0.02 cm, referred to as Cal-vein 2, Fig. 4) exhibit a dark red cathodoluminescence color. These veins are characterized by acinose crystallization and poor transparency. Fluorescent photographs of these veins reveal the presence of many excited blue fluorescence fluid inclusions and the absence of gas-rich inclusions. These typical calcite veins, which formed during different periods, trapped different types of fluids, indicating an accumulation relationship between the two types of fluids.

Fig. 4
Fig. 4
Full size image

Microphotographs contrast diagram of typical calcite veins. (A) Transmitted light panoramic diagram, yellow and red points are laser location; (B) Panoramic diagram of cathode luminescence; (C) Fluorescent panoramic photograph; (D,E) Lithographic photographs of fluid inclusions occurring in different types of veins.

The P2l sandstone serves as the primary tight sandstone reservoir and is primarily composed of quartz, feldspar, and debris. Cementation in this formation is mainly due to carbonate and argillaceous minerals. Our research focuses on fluid inclusions occurring in quartz microfractures, overgrowths, and cementation within the Longtan Formation sandstones.

Rare earth elements

The fractionation of light rare earth elements (LREEs) to heavy rare earth elements (HREEs), as well as Ce and Eu anomalies and La/Ho and Y/Ho mass ratios, are important indicators of fluid composition and paleo-hydrological conditions. P1q veins with typical fluid inclusion characteristics were selected for REE analysis via laser ablation-inductively coupled plasma‒mass spectrometry (LA‒ICP‒MS) at the micron scale.

The total content of rare earth elements (ΣREE) in Cal vein 1 varies from 3.799–23.575 ppm (mean value: 10.081 ppm), while the ΣREE of wall rock 1 varies from 5.144–24.993 ppm (mean value: 14.217 ppm). The calcite veins in this stage exhibit a positive Eu anomaly (δEu = 6.439–15.548) and a weak negative Ce anomaly (δCe = 0.679–0.724) (Fig. 5A,D). The standardized La/Yb (NASC) value of Cal-vein 1 is 29.887, whereas that of the wall rock is 1.474. The rare earth element partitioning pattern is marked by the predominance of light rare earth elements (LREEs). The ΣREE of Cal-vein 2 varies from 47.531–50.872 ppm (mean value: 48.583 ppm), which is higher than that of wall rock 2, with a ΣREE ranging from 1.525–6.311 ppm (mean value: 3.526 ppm). The calcite vein in this stage exhibits a Ce-positive anomaly (δCe = 3.302–21.281) (Fig. 5B,D). The La/Yb (NASC) mean value of Cal-vein 2 is 6.752, and that of the wall rock is 1.392. The abundance of light rare earth elements (LREEs) also characterizes the rare earth partition pattern (Table 1).

Fig. 5
Fig. 5
Full size image

(A) Distribution pattern curve of rare earth elements in Cal-vein 1 and wall rock; (B) distribution pattern curve of rare earth elements in Cal-vein 2 and wall rock; (C) diagram of La/Ho versus Y/Ho; (D) crossplots of δEu versus δCe.

Table 1 Rare earth element (REE) concentrations of the calcite veins and wall rock in Well S174 in the Huangqiao oil and gas field.
Full size table

The La/Ho values of Cal-vein 1 (mean value: 350.568) and Cal-vein 2 (mean value: 122.993) are different. However, the Y/Ho values of Cal-vein 1 (mean value: 30.839) are essentially consistent with those of Cal-vein 2 (mean value: 25.395). The Y/Ho values of the wall rocks are the same as those of the veins, but the late diagenetic processes leave traceable markers for the differentiation and similarity of rare earth elements (REEs) (Fig. 5C). Among these, LREE/HREE ratios, Ce and Eu anomalies, and La/Ho and Y/Ho ratios are crucial indicators of the mineralized fluid composition and paleo-hydrological conditions.

Carbon, oxygen, and strontium isotopic compositions

The δ13CVPDB of the calcite veins in T1q vary from − 2.01‰ to − 1.53‰, which is lower than that of the wall rock (− 1.30‰ to − 1.14‰); the δ13CVPDB of the calcite veins in P1q vary from 2.55‰ to 3.17‰, which is essentially consistent with wall rock (2.60‰ to 2.98‰); and the δ13CVPDB of the calcite veins in C3c vary from 2.59‰ to 2.82‰, which is lower than that of the wall rock (4.23‰) (Table 2).

Table 2 Carbon, oxygen and strontium isotopic compositions of calcite veins and wall rocks in Well S174.
Full size table

The δ18OSMOW values of the calcite veins in T1q range from − 13.45‰ to − 12.57‰, which are lower than those of the wall rock samples (− 6.75‰ to − 6.65‰); the δ18OSMOW values of the calcite veins in P1q range from − 12.57‰ to − 8.11‰, which is lower than that of the wall rock (− 8.21‰ to − 6.17‰); and the δ18OSMOW values of the calcite veins in C3c range from − 11.60‰ to − 8.79‰, which is lower than that of the wall rock (− 7.04‰). Keith and Weber (1964)42 proposed an empirical formula (the result was calculated as the Z value) for determining the diagenetic environment of carbonate rocks based on paleo-salinity, and all the formula values were obtained from the SMOW standard. When Z > 120, it originated from marine limestone, and when Z < 120, it originated from freshwater limestone. According to the calculation results, only Triassic fractured calcite veins have Z values of 117 (less than 120), while the Permian and Carboniferous fractured calcite veins have Z values ranging from 126 to 130 (greater than 120) (Table 2).

The 87Sr/86Sr ratios of the two calcite veins in T1q are 0.7084 and 0.7088, which are higher than those of the surrounding rock (0.7081 and 0.7083); the 87Sr/86Sr ratios of the calcite veins in P1q range from 0.7086 to 0.7096, which are slightly greater than those of the surrounding rock (0.7073 and 0.7084); and the 87Sr/86Sr ratios of the two calcite veins in C3c are 0.7094 and 0.7098, which are greater than those of the surrounding rock (0.7086) (Table 2).

Analysis of fluid inclusions

Petrographic characteristics

During diagenetic processes, both quartz and calcite can effectively trap various types of fluid inclusions. These fluid inclusions can be categorized into four distinct types based on petrological examinations at 25 °C and the Raman peak identification:

Inclusion Type I: These inclusions consist of one-phase CO2 gaseous inclusions (V) primarily located in the interior of grains. They are often found as isolated inclusions and typically suggest a primary origin. In some instances, these inclusions are observed along healed microfractures or within microfractures (ranging from 5 to 15 μm in size) that cut across enlargements (Fig. 6A) or occur as clusters within enlargements, indicating a secondary origin. In terms of shape, these inclusions can be irregular or elliptical. Pure CO2 inclusions are distributed in linear or isolated square or rod-like shapes, further suggesting both secondary and primary origins (Fig. 6B). The diameters of these inclusions typically range from 5 to 10 μm. Generally, pure CO2 gas inclusions occur alongside CO2 aqueous two-phase inclusions.

Fig. 6
Fig. 6
Full size image

(A) CO2-rich gas inclusions coeval with aqueous inclusions cut across overgrowths and grains of quartz; (B) CO2 gaseous inclusions are distributed along microfractures in calcite veins; (C) CH4 gaseous inclusions are linearly distributed in healed microfractures coeval in quartz; (D) CH4 gaseous inclusions are distributed in microfractures of calcite veins; (E) CO2 and CH4 inclusions are present in quartz overgrowths; (F) Hybrid CO2 and CH4 inclusions are arranged in calcite veins; (G) Pale-yellow oil inclusions are arranged in quartz overgrowths; (H) Pale-blue oil inclusions cut across multiple quartz grains; (I) Pale-blue oil inclusions are dispersed and distributed as clusters in calcite cement. All the components were identified by micro–Raman spectroscopy.

Inclusion Type II: These inclusions consist of CH4 single-vapor inclusions (V) and are the most abundant within calcite cements or microfractures in quartz. They often form linear arrays, indicating secondary generation (Fig. 6C,D). Their diameters vary from 5 to 15 μm. Inclusions of this type are typically tubular and aligned parallel to calcite axes, exhibiting rectangular or irregular shapes. Gas-rich inclusions frequently coexist with water-rich inclusions.

Inclusion Type III: These inclusions consist of hybrid CO2 and CH4 gas phases. They are typically found within enlargements (Fig. 6E) and occur as clusters within calcite veins (Fig. 6F), suggesting secondary generation. In terms of shape, these inclusions are commonly irregular and elliptical, with sizes ranging from 9 to 17 μm.

Inclusion Type IV: These inclusions are two-phase oil inclusions (L and V) exhibiting blue and yellow fluorescence, respectively. They are distributed along the edges of enlargements or span multiple quartz grains (Fig. 6G–I), implying a secondary origin. The diameters of these oil inclusions typically range from approximately 13 μm. They exhibit irregular, elliptical, and rod-like shapes. These inclusions typically exhibit irregular and elliptical shapes, with sizes ranging from 9 to 17 μm.

Microthermometry

Microthermometric analyses were carried out on saline aqueous inclusions coentrapped with gas or oil inclusions. The microthermometric data and associated parameters for the secondary-generation fluid inclusions are visually presented in Fig. 7A–C.

Fig. 7
Fig. 7
Full size image

(A) Homogenization temperature (Th) statistics of aqueous inclusions cotrapped with secondary CO2 fluid inclusions (FIs) and secondary hybrid CO2 and CH4 inclusions (HIs); (B) Homogenization temperature statistics of aqueous inclusions cotrapped with secondary CH4 gaseous inclusions and oil fluid inclusions (PIs); (C) Crossplot of Th and salinity of aqueous inclusions coeval with secondary pure different types of inclusions; (D) Density versus Th of CO2 gaseous inclusions and CH4 gaseous inclusions.

For the aqueous inclusions coexisting with type I inclusions, their homogenization temperatures (Th) exhibited two distinct unimodal distributions: one ranged from 140.0 to 150.0 °C, corresponding to salinity levels spanning from 6.02 to 11.55 wt.% NaCl equivalent, and the other ranged from 190.0 to 200.0 °C, with salinity values ranging from 4.65 to 27.52 wt.% NaCl equivalent.

Aqueous inclusions trapped with type II fluid inclusions displayed a bimodal Th distribution, with peaks at 80.0–90.0 °C (salinity interval: 2.01–5.88 wt.% NaCl eq.) and 140.0 °C to 150.0 °C (salinity interval: 0.22–14.82 wt.% NaCl eq.).

In contrast, aqueous inclusions cotrapped with type III inclusions exhibited a more constrained Th range, varying only from 120.0 to 140.0 °C, with corresponding salinities ranging from 2.81 to 9.75 wt.% NaCl equivalent and displaying a well-defined maximum distribution.

The measured Th for aqueous inclusions coentrapped with yellow fluorescence inclusions ranged from 80.0 to 90.0 °C, corresponding to salinity levels between 0.12 and 6.16 wt.% NaCl eq. On the other hand, the Th values for aqueous inclusions coexisting with blue fluorescence inclusions ranged from 80.0 to 90.0 °C (salinity interval: 0.12–6.16 wt.% NaCl eq.) and from 120.0 to 130.0 °C (salinity interval: 0.92–4.05 wt.% NaCl eq.).

Fluorescence spectroscopy

The hydrocarbon fluorescence spectra exhibit a blue-shift phenomenon with increasing maturity, and there is a negative correlation between the thermal maturity of hydrocarbons and the λmax and QF535 values43,44. In this study, oil inclusions of type IV from the Huangqiao oil and gas field were subjected to fluorescence spectrometry analysis. The results reveal that oil-bearing inclusions can be categorized into two distinct assemblages, as shown in Fig. 8. One assemblage exhibited λmax values ranging between 535 and 555 nm and QF535 values ranging from 1.1 to 1.4 (designated P2l in wells S174, X1, and X2). The other assemblage displays λmax values between 480 and 520 nm and QF535 values ranging from 0.7 to 1.1 (referred to as P1q in Well S174; X1, N9, and N13; and P2l in Well X1). Inclusions with relatively high maturity levels (0.85–1.26 Ro%) exhibit pale yellow‒blue fluorescence, while those with relatively lower maturity levels (0.75–0.82 Ro%) exhibit luminous yellow fluorescence.

Fig. 8
Fig. 8
Full size image

Microfluorescence spectra of oil inclusions (A) and λmax (wavelength of the maximum intensity) versus QF535 (ratio of the 535–750 nm flux to the 430–535 flux) diagram (B).

Quantitative Raman analysis of fluid inclusions

Quantitative Raman analysis enables accurate assessment of the density, composition, and pressure of gaseous fluid inclusions within calcite veins and quartz45. Classic Raman spectroscopy data for fluid inclusions containing CH4 and CO2 are presented in Fig. 9. Variations in the density of pure CO2 and CH4 fluid inclusions in relation to the homogenization temperatures of Co-existed aqueous inclusions are depicted in Fig. 7D.

Fig. 9
Fig. 9
Full size image

Classic Raman spectra of CO2 and CH4 fluid inclusions. (A) Spectra of CO2 gas inclusions in quartz collected using a 300 gr mm−1 grating. Fermi diad peaks of CO2 were collected using the 1800 gr mm−1 grating; (B) Spectra of CH4 gas inclusions in quartz collected using the 300 gr mm−1 grating. Symmetric stretching band (v1) of methane was collected using the 1800 gr mm−1 grating; (C,D) Spectra of hybrid CO2 and CH4 components were collected using the 300 gr mm−1 grating in quartz; (E) Spectra of CO2 fluid inclusions collected using the 300 gr mm−1 grating obtained in a homogeneous state in calcite veins. (F) Spectra of CH4 fluid inclusions collected using the 300 gr mm−1 grating obtained in a homogeneous state in calcite veins.

Distinct density categories are observed for pure CO2 and CH4 fluid inclusions. For CO2, a relatively high-density range of 0.524 to 0.656 g/cm3 (Fermi doublet distance: 103.50–103.81 cm−1) is associated with homogenization temperatures (Th) of aqueous inclusions between 185.7 °C and 216.4 °C. In contrast, a relatively low density range of 0.372 to 0.443 g/cm3 (Fermi doublet distance: 103.09–103.29 cm−1) corresponds to the Th of aqueous inclusions between 141.2 and 180.2 °C (Table 3). Similarly, for CH4, a high density range of 0.175 to 0.192 g/cm3 (ν1 band shift: 2912.68–2913.04 cm−1) is linked with the Th of aqueous inclusions ranging from 140.0 to 158.1 °C, while a lower density range of 0.106 to 0.114 g/cm3 (ν1 band shift: 2914.50–2914.71 cm−1) corresponds to the Th of aqueous inclusions ranging from 86.8 to 110.2 °C (Table 4). The homogenization temperature shows a positive correlation with the inclusion density.

Table 3 Summary of the Raman quantitative analysis of pure CO2 inclusions.
Full size table
Table 4 Summary of the Raman quantitative analysis of pure CH4 inclusions.
Full size table

Additionally, Table 5 provides a quantitative compilation of Raman data for hybrid CO2 and CH4 inclusions collected under homogeneous conditions (Fig. 9C,D). The fitting peak area ratios (PAR) of the hybrid inclusions are calculated based on the molar proportions of CO2 and CH4, with most data falling within the range of 41 to 48 mol% CH4, while other estimated values for the inclusions indicate an average composition of 13 mol% CH4 and 87 mol% CO2.

Table 5 Raman quantitative analysis summary of hybrid CO2 and CH4 inclusions.
Full size table

Calcite laser in situ U–Pb isotopic dating results

From the relationship between the density of CO2 gas inclusions and Th (Fig. 7D), it can be seen that there are two stages of CO2 fluid activity. The homogenization temperature of the aqueous inclusions associated with the CO2 inclusions in the first stage (Th interval: 185.7–216.4 °C) was used for laser in situ U–Pb dating of calcite, and the dating results showed that the U–Pb isotope age of the calcite veins with CO2 inclusions was 59.51 ± 8.59 Ma. The homogenization temperature of the aqueous inclusions associated with the CO2 inclusions in the second stage (Th interval: 141.2 °C to 180.2 °C) is included in the burial history analysis, and the results show that the age of the CO2 inclusions in the calcite veins is about 23–4 Ma. We also performed laser in situ U–Pb of calcite dating on the calcite vein containing primary luminous yellow oil inclusions, and the dating result was 198.7±3.9 Ma (Fig. 10).

Fig. 10
Fig. 10
Full size image

Laser in situ U–Pb isotope ages of vein calcite from the T1q in the Huangqiao area of the Subei Basin.

Discussion

Recognition of hydrothermal venting in natural gas reservoirs

The distribution pattern of rare earth elements (REEs) in calcite veins is useful for understanding the source of fluids46. The results reveal distinctive features for Cal-vein 1, showcasing a pronounced enrichment of light rare earth elements (LREEs), a positive europium (Eu) anomaly compared to the wall rock, and a weakly negative anomaly of cerium (Ce). Sorption and complexation are key mechanisms leading to the fractionation of light rare earth elements (LREEs) and heavy rare earth elements (HREEs)47. Acidic and reducing environments are conducive to the enrichment of light rare earth elements (LREEs) in fluids48. Due to their larger ionic radii and relatively weaker sorption strengths, LREEs are more easily mobilized49. The REE content decreases from LREEs to heavy rare earth elements (HREEs) with undulating distribution patterns (Fig. 5A). Following the intrusion of hydrothermal fluids into the fractured reservoir, the migration of Eu2⁺ and LREEs to fluids suggests robust hydrothermal alteration during vein formation50,51. The presence of CO2-bearing fluid inclusions (Th > 190 °C) in this vein further supports hydrothermal alteration. In contrast, Cal-vein 2 exhibits LREE enrichment, a positive Ce anomaly, and a higher total REE content than the wall rock. These characteristics align with a diagenetic fluid-hydrocarbon fluid REE model31,47. A positive Ce anomaly indicates a reducing diagenetic environment, suggesting a relatively closed system with prolonged interactions with wall rocks. Organic matter enhances rare earth element dissolution, facilitating the continuous migration of REEs within rocks. The Yttrium/Holmium (Y/Ho) ratio, which is usually consistent in most geological environments, changes due to fluid‒rock interactions3. The Y/Ho values of Cal-vein 1 (Y/Ho = 25.933–40.268) suggest that the vein-forming fluid originated from formation water evolving under strong water‒rock reaction conditions. Conversely, the Y/Ho values of Cal-vein 2 (Y/Ho = 21.312–34.606) indicate weaker fluid‒rock interactions, suggesting earlier formation of these veins.

In comparison with wallrocks, calcite veins exhibit an increase in 87Sr/86Sr ratios and a decrease in carbon and oxygen isotopic ratios. Both wall rock and calcite vein data points are plotted within the sedimentary carbonate region (Figs. 11, 12). The similarity in δ13C values between calcite veins and wall rock implies that the carbon in fractured calcite veins primarily originated from the wall rock. The 87Sr/86Sr ratios in calcite veins are consistently higher than those in wall rocks and the normal values of contemporaneous seawater, indicating that precipitation was influenced by the input of an external fluid. Moreover, the increasing 87Sr/86Sr values with depth suggest significant reservoir reconstruction in deep reservoirs and good sealing properties in shallow reservoirs.

Fig. 11
Fig. 11
Full size image

Diagram of δ18OSMOW (‰) versus δ13CvPDB (‰).

Fig. 12
Fig. 12
Full size image

Diagram of the 87Sr/86Sr relationships with δ18OSMOW (‰) and δ13CVPDB (‰).

Timing of entrapment of CO2 and hydrocarbons

Aqueous inclusions and gas-rich inclusions or oil inclusions were synchronously observed, indicating typically immiscible entrapment conditions. In immiscible fluid inclusion assemblages (FIAs), the homogenization temperature and homogenization pressure reflect the entrapment state (trapping temperature and pressure) in saturated paleoenvironments52,53,54. The Tan-Lu deep fault and secondary fault, which formed in the late Yanshanian, represent deep mantle CO2 passage systems21,31. Abnormally high trapping temperatures in the CO2 fluid range cannot be used to evaluate the accumulation period. Extensive Raman quantitative results of pure CO2 inclusions indicate that CO2 accumulation can be separated into two stages (Fig. 13): high-density CO2 fluid activity (density interval: 0.524 to 0.656 g/cm3; Th interval: 185.7–216.4 °C) accumulated at approximately 59–39 Ma (early Eocene) . This is in good agreement with the U–Pb dating of CO2 inclusions (59.51 ± 8.59 Ma). And low-density CO2 fluid activity (density interval: 0.372 to 0.443 g/cm3; Th interval: 141.2–180.2 °C) accumulated at approximately 23–4 Ma (Tertiary Quaternary). CO2 accumulation times align with the time sequence of volcanism and tectonic activity (Fig. 14). The chemical composition of the alkane gas implies that it is derived from an oil-type gas mixed with primary cracking gas and secondary oil cracking gas32. The throwing-dot method can be used to constrain gas and oil charging times55,56. CH4 fluid charging into reservoirs occurred in two stages: high-density CH4 fluid activity (density interval: 0.175–0.192 g/cm3; Th interval: 140.0–158.1 °C) accumulated at approximately 61–41 Ma (early Eocene), and low-density CH4 fluid activity (density interval: 0.106–0.114 g/cm3; Th interval: 86.8–110.2 °C) accumulated at approximately 200–185 Ma (early Jurassic) (Fig. 14).

The recovered trapped pressure characteristics of the CO2 reservoirs indicate that the paleo-pressure varied from high abnormal pressure (pressure coefficient: 1.49) to low abnormal pressure (pressure coefficient: 1.16) after adjustment in the Himalayan period. Despite natural gas being released and pressure being reduced due to tectonic movement, the overpressure conditions indicate that the CO2 reservoir maintained good preservation conditions (Fig. 13). Oil fluid accumulation can be divided into two stages: luminous yellow fluorescence oil fluid accumulated at approximately 200–185 Ma (Th interval: 80.0–100.0 °C; 0.75–0.82 Ro%), U–Pb dating was also performed on the luminous yellow oil inclusions, showing an age of 198.7 ± 3.9 Ma. And pale yellow‒blue fluorescence oil fluid accumulated at approximately 60–42 Ma (Th interval: 130.0–150.0 °C; 0.85–1.26 Ro%) (Fig. 14).

Fig. 13
Fig. 13
Full size image

The relationship between the trapping pressure of CO2, CH4 and oil inclusions and the inferred paleoburial depth (Pc: pressure coefficient).

Fig. 14
Fig. 14
Full size image

The burial and thermal history of well S174 and key fluid activity event sequence and measured Ro in right figure.

Interaction between hydrothermal and hydrocarbon fluids

CO2 exhibits extraction effects against crude oil, with late entry causing the extraction of light components, predominantly CH457. The CO2 reservoirs within the T1q and P1q Formations underwent multiple compositional transformations, as indicated by fluid inclusions. Posttrap formation in the Jurassic period, the reservoirs displayed a pattern characterized by early oil and gas filling, followed by late CO2 filling. Notably, the partial top of P2l form an oil layer, while the T1q and P1q formations represent CO2 gas layers based on log interpretation (Fig. 2). The reservoir formation pattern of the P2l Formation suggests that a self-generating and self-bearing oil pool formed in the late Yanshan period. The tops of the P2l reservoirs exhibit overpressure conditions (trapped pressure: 45.25–49.31 MPa) based on paleo-pressure prediction modeling of oil inclusions (Table 6), significantly surpassing the trapped pressure of later CO2 fluids (trapped pressure: 23.92–32.36 MPa). Due to the limited pore pressure difference, lack of fracture development, and poor physical properties, the oil could not be displaced by late deep CO2 vertical migration, resulting in P2l reservoirs existing in a separate pressure system.

Table 6 Oil inclusions trapped pressure analysis summary.
Full size table

Raman quantitative analysis reveals two types of molar proportions of mixed gas inclusions. One type has an average composition in the analyzed hybrid CO2-CH4 fluid inclusions (FIs) of 13% CO2 and 87% CH4 (molar fraction), representing the component of present gas reservoirs. The majority of the other type falls within the range of 41 to 48 mol% CH4, representing inclusions of trapped hybrid CO2 and CH4 in the early Eocene. Displacement functions played a role in the accumulation progress. Displacement functions, driven by CO2 migration and pressure gradients, facilitated the accumulation process by replacing oil and water in the reservoir, altering the fluid composition and enhancing gas-phase dominance.

The vertical Ro variation in well S174 was statistically significant (Fig. 14), showing relatively high Ro segments close to ~ 2000 m and ~ 3000 m (P2l, P1q, and C3h). Hydrothermal CO2 fluid with abnormally high temperature entered the oil-bearing reservoir, accelerating the cracking of crude oil and changing the fluid composition3. This thermal event also accelerated the thermal evolution of the source rock, leading to extraction, pyrolysis, and gas displacement during the reservoir evolution process3.

Fracture system as a major fluid conduit

Fractures developed in upper carbonate formations below the Indochina-Yanshan unconformity in the Lower Yangtze region. Based on core observation and microexamination, fractures are generally divided into structural joints, interbedded fractures, suture fractures, and karst fractures. These fractures are typically half-filled or completely filled, and their development stages correspond well with the tectonic evolution stages (Fig. 15).

Fig. 15
Fig. 15
Full size image

Diagram of the coupling relationship between tectonic evolution and fracture development in the Huangqiao oil and gas field.

During the early-Indosinian and middle-Yanshanian tectonic cycles, tectonic uplift led to the main epigenetic exposure period, accompanied by a shift to a drier climate58. Epigenic weathered karst fractures and interbedded fractures developed, with red‒violet mudstone generally filling these fractures. The calcite veins that formed in the early stage were stained by red‒violet mudstone (Fig. 3D). In the late Yanshanian and early Himalayan tectonic cycles, under the influence of tensile forces, numerous mini-scale sutures and interbedded fractures formed. Hydrocarbon generation and accumulation were primarily distributed in the P1q reservoirs, with oil fluid occurring in small sutures and interbedded veins.

During certain stages of the later Himalayan tectonic cycle, intense compression of the Indo-Chinese plate on the Eurasian continent transitioned to a tensile background in the region, forming a stage of uplift and compression. Large-scale structural sutures and low-angle interbedded fractures were formed, with dark gray argillaceous veins containing a small amount of calcite filling them. Under high temperature and pressure conditions, CO2 carried CH4 to generate large amounts of acidic fluid, infusing it into the reservoir. The injection of reservoir fluid led to the dissolution of early calcite veins, altering the fractures and increasing their pore space. Late CO2 fluid was trapped mainly in calcite veins cemented in large structural fractures. Petrography combined with Raman quantitative analysis also confirmed that different types of veins captured different types of fluids (Fig. 4).

Oil–gas basins (Huangqiao Sag, Lishui Sag, etc.) on the continental shelf of the East China Sea have been influenced primarily by volcanic activity from the island arc on the western side of the Pacific Ocean since the Eocene40. The deep CO2 fluid transports essential substances to sedimentary horizons and provides heat, influencing hydrocarbon generation and accumulation. Significant quantities of CO2 can extract organic components, transporting them to shallower layers, thereby expediting the pyrolysis process of reservoir crude oil. Experiments have indicated that under true formation temperatures and pressures, supercritical CO2 is highly effective in extracting a substantial portion of organic materials from source rocks. The extracted compounds are predominantly saturated hydrocarbons, with a notably high concentration of low-molecular-weight hydrocarbons57. In clastic reservoirs, CO2-rich fluid reacts with the rock to produce a large number of fibrous sheets of sodium bauxite that block secondary pores59. Upon intrusion into carbonate reservoirs, acidic environments often form, gradually corroding unstable minerals (e.g., calcite) and increasing the pore storage capacity of the reservoirs60. Compared with hydrocarbon accumulation, deep fluid accumulation mostly occurs in a late time sequence, and the CO2 in the reservoir mainly produces a displacement effect on methane in the studied area.

Conclusions

In this study, a comprehensive analysis incorporating petrographic observation, elemental and isotopic geochemistry, fluid inclusion quantitative analysis, and U–Pb dating of calcite, combined with considerations of burial history, hydrocarbon generation, and reconstruction information to track the origin of CO2 and its genetic link with hydrocarbon accumulation processes in the Subei Basin. Following conclusions are obtained from this study:

  1. (1)

    REE analysis of typical calcite veins indicates that early calcite veins trapped oil fluids, conforming to a diagenetic fluid-hydrocarbon fluid (REE) model. In contrast, late calcite veins trapped CO2-rich fluids, and the diagenetic environment was characterized by a hydrothermal alteration model. Carbon, oxygen, and strontium isotopes suggest that calcite veins in the Triassic primarily originated from freshwater limestone and were influenced by meteoric freshwater, while those in the Permian and Carboniferous originated from marine limestone facies. The addition of hydrothermal fluid modified the precipitation of calcite veins. Deep reservoirs have undergone significant reconstruction, but shallow reservoirs exhibit good sealing properties. Due to alteration by exogenous fluid, some reservoirs have experienced partial dissolution and destruction, resulting in decreased pore pressure but remaining in an overpressured environment, indicating favorable preservation conditions.

  2. (2)

    Through petrographic observation, Raman quantitative analysis, and laser in situ U–Pb dating of calcite, the fluid components were classified in detail, including CO2 fluid inclusions, CH4 fluid inclusions, Luminous yellow oil inclusions, and Pale blue-yellow oil inclusions. The infiltration of mantle-derived CO2-bearing fluid can be divided into two stages: the early Eocene (around 59–39 Ma) high-density (0.524–0.656 g/cm3) CO2 fluid activity and the Tertiary-Quaternary (23–4 Ma) low-density CO2 (0.372–0.443 g/cm3) fluid activity. The filling time of oil and CH4 fluids showed two stages: the medium-mature oil fluids and CH4 fluids accumulated in the Early Jurassic (200–185 Ma), and the high-mature oil fluids and CH4 fluids accumulated in the Early Eocene (61–41 Ma).

  3. (3)

    Component variations in fluid inclusions reflect the process of mutual displacement between CO2 and hydrocarbon fluids in the reservoir, with reservoir pore pressure identified as a major influencing factor. Hydrothermal fluid accelerates the cracking of crude oil, and abnormally high intervals of vitrinite reflectance indicate rapid thermal evolution of the source rock.

  4. (4)

    Fracture development aligns well with tectonic evolution. Karst and interbedded fractures developed during the early-Indosinian and middle-Yanshanian stages. Mini-scale sutures and interbedded fractures, accompanied by oil fluid, formed during the late Yanshanian and early Himalayan stages. Large-scale structural and interbedded fractures, accompanied by CO2 and CH4 fluids, developed during the late Himalayan stage. The injection of acidic fluid led to the dissolution of early calcite veins, altering fracture structures and enhancing the pore space of fractures.