Laser-Raman Analysis of Individual Fluid Inclusions and Hydrocarbon-Fluid Evolution in the Marine–Terrestrial Transitional Longtan Formation Shale, Southern Sichuan Basin

Abstract

This study integrates microthermometry and laser-Raman spectroscopy of individual fluid inclusions with basin modelling to reconstruct the hydrocarbon-fluid evolution and multistage re-mobilisation of the Permian Longtan Formation transitional marine–terrestrial shale in the YJ-LJ area, southern Sichuan Basin. Systematic analysis of aqueous two-phase, methane-rich, and associated bitumen inclusions hosted in fracture-fill veins and sandy partings identifies four fluid episodes, enabling subdivision of petroleum evolution into five stages. Results show trapping pressure increasing with temperature to a maximum of 107.77 MPa, equivalent vitrinite reflectance (EqVR) between 1.49% and 2.49%, and formation-water salinity that first rises then falls, remaining within the high-salinity continental-leaching brine field. Coupled with thermal-history modelling, shale-oil/gas evolution is divided into: (1) low-pressure slow burial, (2) over-pressured rapid burial, (3) sustained over-pressured deep burial, (4) high-pressure uplift adjustment, and (5) late uplift adjustment. The study demonstrates that over-pressure has been partially preserved, providing critical palaeo-fluid and pressure evidence for exploring transitional marine-terrestrial shale gas.

1. Introduction

Fluid inclusions are micron-scale domains of ancient fluid that were trapped in structural defects or growth cavities of host minerals during crystallisation. They preserve critical palaeo-fluid information such as temperature, pressure and composition at the time of entrapment [1,2]. Investigation of fluid inclusions is therefore a direct and robust approach [3,4] and has been widely applied in conventional hydrocarbon-accumulation studies [5,6,7]. In unconventional reservoirs, however, the fine-grained size and low crystallinity of shale mean that usable inclusions seldom form during palaeo-fluid flow [8]. Fortunately, mineralised veins that fill shale fractures and sandy partings within mudstone commonly contain abundant workable inclusions and can provide key evidence for shale-gas-reservoir evaluation [9,10].

Raman spectroscopy is a powerful tool for fluid-inclusion analysis [11,12,13]. Because the laser beam can be focused to <3 µm, both entire inclusions and individual phases within them can be interrogated in situ [14]. The recent development of sealed capillary standards with fixed compositions [15,16,17] is driving the technique from qualitative toward semi-quantitative or even quantitative analysis, enabling rapid quantification of organic-matter maturity, trapping pressure and fluid density [18,19] that can be applied directly to shale-gas exploration.

In the southern Sichuan Basin, intensive shale-gas drilling has prompted inclusion-based investigations of fracture-filling veins in the Silurian Longmaxi marine shales, revealing palaeo-pressure and accumulation histories [3,20,21,22]. By contrast, the Permian Longtan Formation (LTF)—a marine–terrestrial transitional mudstone succession in the same region—has received little attention. Expanding research on the LTF within the region will significantly benefit the exploration of unconventional hydrocarbon resources in new frontiers of China.

Here, we collected vein material from black LTF shales and sandy interbeds from the YJ-LJ area, southern Sichuan Basin, and prepared doubly polished thin sections. Individual two-phase (aqueous vapour–liquid) brine inclusions and associated solid-bitumen inclusions were subjected to non-destructive petrography, microthermometry and laser-Raman spectroscopy. Trapping pressures, thermal maturity and formation-water compositions were reconstructed to unravel the hydrocarbon-fluid evolution of the LTF and to discuss its geological implications.

2. Geological Setting and Samples

The Sichuan Basin, situated in the western part of the Yangtze Platform, is one of China’s most important petroliferous basins. Within it, the Permian Longtan Formation (LTF), deposited in a marine–terrestrial transitional setting, has emerged as a new ‘hot spot’ for shale-gas exploration and development [23,24,25]. The Middle–Late Permian Dongwu orogeny was accompanied by large-scale volcanism in the southwestern basin—the Emeishan basalt eruptions—linked to the ascent of a mantle plume and constituting a major large igneous province [26,27]. Consequently, the Late Permian Sichuan Basin shows pronounced sedimentary differentiation: from the southwestern margin landward, successively, continental, transitional and marine environments are recognised [28,29,30] (Figure 1a). Longtan Formation strata are best preserved in the central and eastern basin. They were laid down in transitional coastal-plain to lagoonal settings and consist dominantly of fine-grained rocks—mudstone, shale, siltstone, and coal seams—intercalated with thin limestones and siltstone lenses [31,32,33,34] (Figure 1b,c). Rapid alternation of multiple lithologies produces an overall coal.

Given the absence of a fully cored well through the Longtan Formation across the basin, six samples were selected from the LTZ in two key wells in the YJ-LJ area of southern Sichuan Basin: YJ-1 (coring the upper-middle LTZ) and LJ1 (coring the lower-middle LTZ), together providing good coverage of the formation. The suite comprises two vein samples from LJ-1 (Figure 2a,b), three vein samples from YJ-1 (Figure 2c–e) and one sandy parting sample extracted from mudstone in YJ-1 (Figure 2f). Owing to discontinuous coring, samples from well LJ-1 are concentrated in the Tan1 member, whereas those from well YJ-1 are mainly from the Tan3 member. Given the essentially uniform sedimentary and structural setting, we infer that the fluid-inclusion assemblages are comparable between the two intervals. Consequently, all samples were sealed immediately after retrieval and transported promptly to the laboratory for thin-section preparation and subsequent analyses.

3. Experiment

3.1. Preparation of Double-Polished Wafers

Three preparation techniques exist for fluid-inclusion analysis: cleavage chips obtained by splitting along mineral parting planes with a razor blade, quick mounts—ground but unpolished wafers whose surface relief is locally flattened with immersion oil matched to the host-mineral refractive index—and double-polished wafers. For this work, we selected double-polished wafers prepared in five sequential steps after McNeil and Morris [35] (1992): impregnation, sectioning, grinding, polishing, and mounting.

3.2. Microthermometry

A UK-manufactured Linkam THMSG600 (Linkam Scientific Instruments Ltd., Tadworth, UK) heating–freezing stage was used to determine homogenisation (T_h) and final ice-melting (T_m) temperatures with an uncertainty of ±0.1 °C. During T_h acquisition, samples were first heated at 5–10 °C min⁻¹; once the vapour bubble began to shrink rapidly the rate was reduced to 0.5 °C min⁻¹, the temperature of complete homogenisation was logged, heating was continued 5–10 °C beyond this point to check for phase changes, the section was then cooled until the bubble reappeared and the cycle was repeated to verify reproducibility. For two-phase aqueous inclusions, cooling was initiated at 15–20 °C min⁻¹ to nucleate ice; warming was then performed at 1 °C min⁻¹ to −10 °C and finally at 0.5 °C min⁻¹, with the last-melting ice-crystal temperature recorded as T_m. Methane-rich inclusions were quenched with liquid nitrogen to the phase-transition point, cooled further to monitor bubble behaviour, and subsequently warmed at 2 °C min⁻¹ until homogenisation was achieved; the corresponding Tₕ(CH₄) was recorded. In this study, microthermometric measurements were performed on as many inclusions as possible, provided their size and morphology met the required criteria (typically >3 μm and regular shape).

3.3. Laser-Raman Spectroscopy

Methane-rich, aqueous two-phase and bitumen inclusions in the present samples were non-destructively characterised using a LabRAM HR800 Raman spectrometer (HORIBA Scientific, Kyoto, Japan) equipped with a 532 nm YAG solid-state laser; individual inclusions > 1 μm were targeted. Gas, liquid and solid phases were identified from their characteristic Raman shifts. v₀ denotes the methane Raman peak position at near-zero pressure; for this study, the calibration performed in the laser-Raman laboratory gives v₀ = 2917.52 cm⁻¹.

4. Geological Applications and Results

4.1. Reconstructing Palaeo-Pressure During Geological Time

For fluid inclusions trapped in diagenetic minerals, coeval oil-bearing and aqueous inclusions share identical temperature and pressure conditions at the time of entrapment; consequently, the intersection of their respective isochores defines the trapping temperature (T_t) and trapping pressure (P_t) [36,37]. Pironon [38] (2004) discriminated three inclusion scenarios in hydrocarbon-bearing environments: (i) both aqueous and petroleum phases are undersaturated in gas at entrapment—hence, T_t and P_t correspond to the intersection of the liquid-phase isochores, whereas the homogenisation temperature (T_h) and homogenisation pressure (P_h) are lower than T_t and P_t (Figure 3a); (ii) both phases are gas-saturated, and the isochore intersection coincides with the bubble-point curve, hence T_h and P_h are equal to T_t and P_t (Figure 3b); (iii) the aqueous phase is gas-saturated whereas the petroleum phase is gas-undersaturated—hence, the isochore intersection lies on the bubble-point curve of the aqueous inclusion only (Figure 3c). Petrographic inspection and geological context indicate that the studied assemblage corresponds predominantly to scenario (ii); accurate palaeo-pressure reconstruction therefore hinges on precise determination of T_h and P_h [39].

Accurate T_h values were acquired via microthermometry on a heating–freezing stage, whereas P_h was derived by laser micro-Raman spectroscopy. Since Fabre et al. [40] first documented in 1979 that the C–H stretching band of gaseous CH₄ shifts systematically with density or pressure, numerous subsequent studies have confirmed that the symmetric C–H stretching vibration moves to lower wavenumbers as density or pressure increases [16,41,42,43]. Xi et al. [44] (2020) constructed an experimental setup following the protocol of Chou et al. [17] (2008) and, using the calibration procedure of Qiu et al. [22] (2020), established a CH₄ peak-position versus pressure calibration curve (Figure 4); a corresponding density relationship was not provided. Lu et al. [42] compiled multi-laboratory data to generate a universal Raman-shift versus pressure equation for methane (Equation (1)); although minor deviations occur at elevated pressures, their companion equation linking CH₄ peak position to density (Equation (2)) has found wide application.

P(MPa) = −0.0148 × (ν_p − ν₀)⁵ − 0.1791 × (ν_p − ν₀)⁴ − 0.8479 × (ν_p − ν₀)³ − 1.765(ν_p − ν₀)² × 5.876 × (ν_p − ν₀)

(1)

ρ(g/cm³) = −5.17331 × 10⁻⁵(ν_p − ν₀)³ + 5.53081 × 10⁻⁴(ν_p − ν₀)² − 3.51387 × 10⁻²(ν_p − ν₀)

(2)

where ν_p–ν₀ denotes the Raman shift difference between the measured CH₄ peak position and that of nearly zero-density methane.

Fluid inclusions archive the temperature and pressure of the ambient fluid at the time of entrapment and thus constitute a robust tool for reconstructing palaeo-fluid P–T conditions. Previous studies have demonstrated that, within an immiscible fluid-inclusion assemblage, the lowest homogenisation temperature (T_h) of the aqueous two-phase inclusions corresponds to the trapping temperature. The trapping pressure of the sample inclusions is calculated with the equation of state for super-critical methane systems developed by Duan [45,46] (1992a, 1992b) (Equations (3)–(12)). By coupling the derived methane-inclusion density range with the above equation, the trapping pressure of each inclusion was obtained.

$$\mathrm{Z}={\displaystyle \frac{\mathrm{PV}}{\mathrm{RT}}}={\displaystyle \frac{{\mathrm{P}}_{\mathrm{r}}{\mathrm{V}}_{\mathrm{r}}}{{\mathrm{T}}_{\mathrm{r}}}}=1+{\displaystyle \frac{\mathrm{B}}{{\mathrm{V}}_{\mathrm{r}}}}+{\displaystyle \frac{\mathrm{C}}{{\mathrm{V}}_{\mathrm{r}}^2}}+{\displaystyle \frac{\mathrm{D}}{{\mathrm{V}}_{\mathrm{r}}^4}}+{\displaystyle \frac{\mathrm{E}}{{\mathrm{V}}_{\mathrm{r}}^5}}+{\displaystyle \frac{\mathrm{F}}{{\mathrm{V}}_{\mathrm{r}}^2}}\left(\beta +{\displaystyle \frac{\gamma }{{\mathrm{V}}_{\mathrm{r}}^2}}\right)\exp \left({\displaystyle \frac{\gamma }{{\mathrm{V}}_{\mathrm{r}}^2}}\right)$$

(3)

$$\mathrm{B}={\mathrm{a}}_1+{\displaystyle \frac{{\mathrm{a}}_2}{{\mathrm{T}}_{\mathrm{r}}^2}}+{\displaystyle \frac{{\mathrm{a}}_3}{{\mathrm{T}}_{\mathrm{r}}^3}}$$

(4)

$$\mathrm{C}={\mathrm{a}}_4+{\displaystyle \frac{{\mathrm{a}}_5}{{\mathrm{T}}_{\mathrm{r}}^2}}+{\displaystyle \frac{{\mathrm{a}}_6}{{\mathrm{T}}_{\mathrm{r}}^3}}$$

(5)

$$\mathrm{D}={\mathrm{a}}_7+{\displaystyle \frac{{\mathrm{a}}_8}{{\mathrm{T}}_{\mathrm{r}}^2}}+{\displaystyle \frac{{\mathrm{a}}_9}{{\mathrm{T}}_{\mathrm{r}}^3}}$$

(6)

$$\mathrm{E}={\mathrm{a}}_{10}+{\displaystyle \frac{{\mathrm{a}}_{11}}{{\mathrm{T}}_{\mathrm{r}}^2}}+{\displaystyle \frac{{\mathrm{a}}_{12}}{{\mathrm{T}}_{\mathrm{r}}^3}}$$

(7)

$$\mathrm{F}={\displaystyle \frac{\alpha }{{\mathrm{T}}_{\mathrm{r}}^3}}$$

(8)

$${\mathrm{P}}_{\mathrm{r}}={\displaystyle \frac{\mathrm{P}}{{\mathrm{P}}_{\mathrm{C}}}}$$

(9)

$${\mathrm{T}}_{\mathrm{r}}={\displaystyle \frac{\mathrm{T}}{{\mathrm{T}}_{\mathrm{C}}}}$$

(10)

$${\mathrm{V}}_{\mathrm{r}}={\displaystyle \frac{\mathrm{V}}{{\mathrm{V}}_{\mathrm{C}}}}$$

(11)

$${\mathrm{V}}_{\mathrm{C}}={\displaystyle \frac{{\mathrm{RT}}_{\mathrm{C}}}{{\mathrm{P}}_{\mathrm{C}}}}$$

(12)

In the equation, P is pressure (bar), T is temperature (K), R is the gas constant (0.08314467 bar·dm³·K⁻¹·mol⁻¹), V is the molar volume (dm³/mol) derived from the density (ρ) and molar mass of methane, and Z is the compressibility factor. P_r and T_r denote reduced pressure and reduced temperature, respectively, both dimensionless; P_c and T_c are the critical pressure (46 bar) and critical temperature (190.4 K), consistent in units with P and T. The coefficients are: a₁ = 0.0872553928, a₂ = −0.752599476, a₃ = 0.375419887, a₄ = 0.0107291342, a₅ = 0.0054962636, a₆ = −0.0184772802, a₇ = 0.000318993183, a₈ = 0.000211079375, a₉ = 0.0000201682801, a₁₀ = −0.0000165606189, a₁₁ = 0.000119614546, a₁₂ = −0.000108087289, α = 0.0448262295, β = 0.75397, γ = 0.077167.

Microthermometric data for aqueous two-phase inclusions yield T_h values between 112.4 and 201.2 °C (Figure 5a,b), clustering in four intervals at 110–120 °C, 140–150 °C, 160–170 °C and 180–190 °C. A limited number of outliers are attributed to metastable behaviour of the system [47] and are excluded from further discussion. Accordingly, the dataset was subdivided into four groups: low-T, low-to-moderate-T, moderate-to-high-T, and high-T. Contemporaneous methane-bearing inclusions from each group were selected for laser-Raman analysis (Figure 6b–g); inclusion IDs are listed in

Table 1. Summary of fluid-inclusion sample information.

Sample Number	Depth (m)	Lithology	Occurrence	Type of Inclusions	Temperature Category	V(CH4) #G	Gas–Liquid Ratio (%) #L	Tf #L (°C)	Th #L (°C)
LJ-6-G (LJ-6-CH4)	3170.75	Shale	Quartz vein	Secondary	high-T	2912.72	8	−5.5	187.2
LJ-1-G (LJ-1-CH4)	3181.18	Shale	Quartz vein		high-T	2912.72	7	−5.5	185.3
YJ-1-2-ZG (YJ-1-2-CH4)	3019.82	Shale	Quartz vein		moderate- to-high-T	2911.92	7	−6.2	168.8
YJ-2-3-ZG (YJ-2-3-CH4)	3019.60	Shale	Quartz vein		moderate- to-high-T	2911.61	6	−6.0	161.4
YJ-2-1-ZD (YJ-2-1-CH4)	3019.27	Shale	Calcite vein		low-to- moderate-T	2911.58	7	−5.2	140.4
YJ-8-D (Y-8-CH4)	3055.59	Sandstone	Micro-fractures confined to quartz grains		low-T	2911.10	5	−5.1	112.4

Sample Number: aqueous two-phase inclusion (methane-bearing inclusion); ^#G: methane-inclusion data; ^#L: aqueous two-phase inclusion data; T_f: Full width at half maximum; T_h: Homogenization temperature.

.

Laser-Raman spectra of methane inclusions coeval with aqueous two-phase inclusions (Figure 6a) display intense host-mineral bands together with CH₄ peaks of variable relative intensity. Overall, trapping pressure increases with trapping temperature (Figure 7), ranging from 29.93 to 107.77 MPa, whereas inclusion densities cluster between 0.19 and 0.26 g/cm³. An anomalous data point (elevated density and pressure) occurs within the low-to-moderate-T group. Integrating the discussion in Section 4.2, this inclusion most likely formed during a late entrapment event.

4.2. Determine the Maturity of Thermal Evolution

Within a single stratigraphic unit of comparable organic-matter abundance, the degree of thermal maturation dictates the maximum hydrocarbon yield. For assessing the maturity of palaeo-organic matter, the Raman Maturity Method (RaMM-Ro) is regarded as an accurate and rapid diagnostic tool [18,48]. Mixed or discrete Raman spectra of vitrinite and inertinite are acquired; multivariate linear regression of selected G- and D-band parameters [49] (Figure 8) yields the equivalent vitrinite reflectance (EqVR). A key advantage of RaMM is its applicability to dispersed organic matter where vitrinite is unidentifiable [50,51], making it well-suited to bitumen trapped within fluid inclusions. Using the maturity models of Lu et al. [52] or Wilkins et al. [53] (Equations (13)–(16)), Raman parameters of associated bitumen inclusions from each temperature group were processed to derive EqVR values.

νG–νD denotes the peak-to-peak separation (cm⁻¹) between the G and D bands; b is the slope of the spectral baseline; Saddle index is the ratio of the G-band height to the saddle minimum between the overlapping G and D peaks; FWHM-G and FWHM-D are the full widths at half maximum of the G and D bands, respectively; I-G/I-D is the height ratio of the G to D band (Figure 8). For the protocol established by Wilkins et al. [53], a reliable EqVR calculation further requires simultaneous acquisition of the background baseline. Equation (3) is valid only for equivalent vitrinite reflectance (EqVR) between 0.4% and 1.0%, whereas Equation (4) applies exclusively to EqVR > 1.25%. When EqVR falls in the transitional window of 1.0–1.25%, both equations are solved, and the arithmetic mean of the two results is adopted.

_RmcRo% = 0.0537 d(νG-νD) − 11.21

(13)

_RmcRo% = 1.1659 (I-D/I-G) + 2.7588

(14)

EqVR% = −6.384 + 5.439 log(νG − νD) + 0.863 log(b) + 0.832 log(saddle index) − 2.677 log(FWHM-D)
− 0.661 log(FWHM-D)

(15)

EqVR% =−52.3363 + 13.1992 log(νG − νD) − 0.7964 log(b) + 8.1121 log(FWHM-D) + 15.3255 log(I-G/I-D)

(16)

Throughout the burial history, formation temperature exhibits a quasi-quadratic relationship with time within a specific interval; consequently, maturity analysis was focused on the three inclusion generations (LJ1-6-L, YJ1-2-3-L and YJ1-2-4-L) entrapped during this thermally symmetric period. Considering the validity ranges of the respective maturity algorithms, Equations (13) and (16) were selected for the present study. When Equation (13) was applied, the G–D peak separations [νG – νD] of the analysed bitumen inclusions LJ1-6-L, YJ1-2-3-L and YJ1-2-4-L are 266.14 cm⁻¹, 268.05 cm⁻¹ and 263.52 cm⁻¹, yielding RmcRo values of 3.08%, 3.18% and 2.94%, respectively. Application of Equation (16) required spectral deconvolution [57,58,59] (Figure 9) using Gaussian, Lorentzian and pseudo-Voigt functions [60] to extract the requisite parameters (

Table 2. Peak-by-peak integration parameters for each sample.

Sample Umber	Band Type	Fitted Peak Area	FWHM	Band Height	Wp (cm−1)	SP (%)
YJ-1-2-ZG (YJ-1-2-L)	-	16,986.50	19.87	803.04	169.00	2.50
	Calcite	55,678.11	32.31	1618.90	285.94	8.19
		10,384.73	27.88	349.98	723.66	1.53
		59,575.12	10.08	5550.91	1094.33	8.77
	D	460,827.98	399.82	1082.78	1339.60	67.80
	G	76,213.40	50.62	1414.33	1605.74	11.21
YJ-2-1-ZD (YJ-2-1-L)	Quartz	1078.19	8.69	325.06	130.14	0.77
		7190.89	22.42	301.88	208.59	5.12
		12,741.36	11.73	1024.70	467.30	9.08
	D	81,546.86	173.53	441.47	1336.15	58.08
	G	37,044.46	45.99	756.70	1604.21	26.39
	-	795.25	5.87	981.62	2913.71	0.57
LJ-6-G (LJ-6-L)	Quartz	0.00	0.28	2903.36	127.29	0.00
		21,188.78	32.18	618.44	207.27	6.24
		40,730.76	12.01	2011.65	466.98	11.99
	D	175,414.85	241.74	681.69	1341.99	51.63
	G	45,111.87	37.00	1145.50	1605.51	13.28
	-	57,274.66	8.80	3344.07	2913.42	16.86

FWHM: Full width at half maximum; W_p: Weighted mean centre of peak; S_P: Percentage of fitted peak area relative to total integrated area.

). The resulting equivalent vitrinite reflectance (EqVR) values are 2.49% (Figure 9a,b), 1.49% (Figure 9c,d) and 2.01% (Figure 9e,f). With reference to Zhang et al. [61], who reported a maximum vitrinite reflectance (Romax) of 3.05% for the target succession, EqVR values derived from the Wilkins [53] model are considered more representative of the studied horizon. Consequently, the maturity sequence from lowest to highest is as follows: moderate-to-high-T group (1.49%), high-T group (2.01%), and low-to-moderate-T group (2.49%).

4.3. Analyse the Properties of Formation Water During the Burial Process

The liquid phase of aqueous two-phase inclusions also preserves abundant information that provides critical insights into the nature of pore fluids during burial diagenesis [62,63]. Hall et al. [64] derived a salinity calculation based on experimental data and Raoult’s law. Knight et al. [65] determined the critical properties of the NaCl–H₂O system, thereby expanding the database for critical-condition calculations. Subsequent compilation by Bodnar [66] has made this approach the most widely used method for determining fluid-inclusion salinity (Equation (17)).

Salinity(%) = 0.00 + 1.78T_f − 0.04421.78T_f² + 0.000557T_f³

(17)

where T_f is the final ice-melting temperature.

Based on T_f data obtained with a heating–freezing stage (Table 1), salinity values for each temperature group are displayed in Figure 10. Salinity first increases and then decreases with rising thermal maturity, the inflexion preceding the corresponding shift in trapping pressure and density. All values are high (salinity > 5 wt%) and exhibit a narrow total range. Burial depth is considerable (>3000 m), and the shale matrix has inherently low permeability; hence, external fluid ingress was minimal during progressive burial. It is therefore inferred that internally generated water during hydrocarbon generation and variable degrees of water–rock interaction were the primary controls on the salinity evolution of pore fluids. The slight decrease in salinity recorded during uplift is interpreted to reflect, at least in part, the influence of externally derived fluids [67].

Leveraging the high spatial resolution of laser-Raman spectroscopy, the liquid phase of aqueous two-phase inclusion Y-8-D from the low-T group was analysed (Figure 11a). Host-mineral bands (quartz: 208.10 and 467.64 cm⁻¹) are identical for coeval methane-bearing (vapour) and aqueous (liquid) inclusions, whereas the baseline is slightly displaced under the liquid-phase signal. Two prominent liquid-phase peaks occur at 977.369 and 1366.67 cm⁻¹. Following the ion-specific band assignments of Rosasco [57], Burke [68] and Frezzotti [69], these features are assigned to SO₄²⁻ and HCO₃⁻, respectively. After integrated intensity transformation (Figure 11b), the SO₄²⁻/HCO₃⁻ ratio is ~2:1. Monatomic cations (Na⁺, K⁺, etc.) are Raman-inactive and could not be detected. Within the Late Permian depositional context [26], the palaeo-formation water was classified using the Sulin scheme [70]; charge-balance calculations (Figure 12) indicate a continental-leaching water type.

5. The Process of Reservoir Oil and Gas and Its Geological Significance

Integrating the above results, the formation and evolution of shale gas in the study area can be rigorously constrained. In the YJ block, uplift initiated at approximately 70 Ma during the Yanshanian–Himalayan transition [71]. Maximum palaeo-burial depth prior to exhumation reached 5.6–5.8 km [61], palaeo-temperatures were 180–210 °C, and Longtan Formation shale maturity (Ro) was 2.4–2.6% [28,72]. Under the constraints of regional thermal events, we referenced the thermal history of the adjacent YS2 well [73], and used the drilled lithology log and stratigraphic picks from well YJ-1 (whose lithological column is broadly similar to that reported in [28,31]) to simulate the burial–uplift thermal evolution of YJ-1 using BasinMod (Figure 13). Microthermometric and Raman data were superimposed on the modelled thermal curve, combined with Raman spectral deconvolution parameters (Figure 14). Peak height (I-CH₄), peak area (S-CH₄), height ratio (I-CH₄/quartz) and area ratio (S-CH₄/quartz) were employed to quantify the relative intensity of successive hydrocarbon-fluid re-mobilisation events. This approach enabled the identification of four distinct fluid episodes preserved within the inclusions and provided a framework for analysing the associated petroleum evolution history.

Fluid re-mobilisation documented by the low-T inclusion assemblage represents the earliest episode. Approximately 172–177 Ma ago, formation pressure was the lowest of the four stages, organic maturity was 0.90%, formation-water salinity was 8.0%, and the water is classified as a continental-leaching mixed brine; hydrocarbon-fluid re-mobilisation intensity was extremely weak. The moderate-to-high-T assemblage records the second, slightly younger episode. Around 140–146 Ma ago, formation pressure was still relatively low, organic maturity reached 1.49%, salinity rose to 9.3%, and Raman spectral deconvolution parameters consistently indicate a significant increase in hydrocarbon-fluid re-mobilisation intensity. The high-T assemblage represents the third episode, approximately 69 Ma ago, when formation pressure peaked, organic maturity was 2.01%, salinity remained at 8.6%, and this stage records the most intense hydrocarbon-fluid re-mobilisation event. The low-to-moderate-T assemblage records the latest pulse, approximately 69 Ma ago, characterised by the highest organic maturity (2.49%), lower formation-water salinity (8.1%), and a pronounced drop in trapping pressure. Hydrocarbon-fluid re-mobilisation intensity was weak, suggesting that the inclusion population captures the waning stage of this final episode.

Integrating the four chronological episodes recorded by fluid inclusions with thermal-history modelling, the petroleum evolution of the LTF can be subdivided into five stages.

Low-pressure slow burial (Ro < 0.9%): deposition duration was short, and burial was shallow; pore waters are high-salinity, continental-leaching mixed brines that established the baseline aquatic environment for subsequent hydrocarbon generation. Brief uplift related to the Indosinian orogeny had a negligible impact on thermal evolution; organic matter remained immature, early hydrocarbon generation was incipient, and formation pressure was hydrostatic. Consequently, hydrocarbon-fluid activity was extremely weak.

Over-pressured rapid burial (0.9% < Ro < 1.5%): accelerated subsidence driven by early Yanshanian tectonics. Increasing temperature with depth carried organic matter through the diagenetic window into the oil-generation zone. Hydrocarbon generation initiated secondary over-pressure and enhanced fluid mobility. Expulsion of connate water and variable water–rock reactions reduced pore volume, elevating brine salinity.

Sustained over-pressure deep burial (1.5% < Ro < 2.0%): LTZ approached maximum burial; maturity rose rapidly, the reservoir was charged with highly evolved, dry thermogenic gas and moderate-to-strong over-pressure. Hydrocarbon-fluid activity peaked, allowing fracture-hosted veins of the high-T episode to trap abundant dense methane and aqueous two-phase inclusions that archive palaeo-P–T conditions.

High-pressure uplift adjustment (2.0% < Ro < 2.5%): late Yanshanian and early Himalayan compression drove uplift; heating and pressurisation waned. Formation pressure remained strongly over-pressured. Internal fluid activity declined, probably in response to pressure release. Pressure decline reflects both unloading during uplift and lateral gas migration along fractures into adjacent strata.

Late uplift adjustment (Ro > 2.5%): accelerated exhumation under early Himalayan compression. Organic matter reached post-mature stage and hydrocarbon generation waned [74,75]; present-day over-pressure analyses indicate weak internal fluid activity, allowing partial preservation of original over-pressure in gas-charged intervals [76].

6. Limitations

This study demonstrates the first application of in situ, non-destructive laser-Raman analysis of individual fluid inclusions in the Longtan Formation transitional marine–terrestrial reservoir rocks, Sichuan Basin, and provides a robust constraint on the petroleum evolution history of the unit. By correlating micro-Raman spectral parameters (peak area, peak position, peak height and their ratios) with stratigraphic and geologic constraints, the composition and internal pressure of single fluid inclusions can be retrieved accurately and comprehensively. This approach overcomes the limitations of conventional quantitative techniques that are sensitive to thin-section thickness, experimental conditions, inclusion composition and host-mineral properties. Nevertheless, the burial-related evolution is simultaneously governed by multiple interactive factors; the analytical workflow presented herein—particularly regarding the characterisation of host minerals to fluid inclusions—remains to be refined as a consequence of current core-acquisition limitations. In addition, quantification in this work relied primarily on ratio-based calibrations; single-sample reference standards were not prepared. Future efforts should focus on coupling this technique with complementary instrumentation to advance towards high-precision, fully quantitative analysis.

7. Conclusions

Fluid inclusions are direct archives of palaeo-fluid physico-chemistry; non-destructive, in situ analysis of single inclusions has long been a focus of petroleum-geological research. Laser-Raman-based in situ non-destructive protocols applied to individual inclusions in reservoir rocks can decipher hydrocarbon evolution histories and are widely used to reconstruct palaeo-pressure, thermal maturity and burial-stage pore-water compositions.

Integrated laser-Raman investigations of fluid inclusions in the Longtan Formation reservoir reveal four distinct fluid-flow episodes. Early (low-T) and late (low-to-moderate-T) pulses were weak, whereas mid-early (moderate-to-high-T) and mid-late (high-T) events were intense. Inclusion-derived pressure, salinity and density all increase with thermal maturity to a maximum and then decline. These four temporal anchors are used to subdivide the petroleum evolution into five stages: low-pressure slow burial, over-pressured rapid burial, sustained over-pressured deep burial, high-pressure uplift adjustment, and late uplift adjustment. The resultant evolutionary trajectory indicates that over-pressure has been well preserved in parts of the studied gas accumulations.

Given four fluid pulses of contrasting intensity, five hydrocarbon-evolution stages and the interbedded lithologies of the Longtan Formation, the potential for fluid re-mobilisation to generate vertically continuous geologic sweet spots requires further investigation.