Coal Facies and Its Effects on Pore Characteristics of the Late Permian Longtan Coal, Western Guizhou, China

,


Introduction
The implementation of China's "carbon emission peak and carbon neutrality" goals has resulted in unconventional natural gas being an important bridge for the transition from fossil fuels to new energy sources [1]. Coalbed methane, along with shale gas and tight gas, are the main components of unconventional natural gas known today [2]. However, unlike other unconventional natural gas, the development of coalbed methane not only provides clean energy but also safe and efficient mining of coal resources [3,4].
In 2020, China's coalbed methane production reaches 6.7 billion cubic meters [5], strengthening our confidence in the development of the coalbed methane industry. The evaluation results of national coalbed methane resources by the Strategic Research Center of the Ministry of Land and Resources show that the amount of coalbed methane geological resources in China is as high as 36 trillion cubic meters, of which 3.15 trillion cubic meters is located in Guizhou Province (accounting for about 10% of the country). Guizhou is thus the province of the most abundant coalbed methane resources in South China, and it is also the prospective province for the coalbed methane industry in China [6,7].
Coal is a highly heterogeneous combustible organic rock composed of different macerals and minerals, and the macerals compositions, types, and distribution are controlled by the initial stage of coal formation, such as peat formation, diagenesis, and coalification [8][9][10][11][12][13]. The type and environment of bogs in the peatification stage are the main controlling factors for the composition and distribution of coal macerals [12,[14][15]. According to macerals compositions, coal facies can be subdivided into five types, and they are shallow-water covered forest peat moor, wetland herbaceous moor, shallow-water covered wetland herbaceous moor, moist forest peat moor, and low peat moor coal facies [13]. The composition and spatial distribution of coal macerals during peatification stage are important factors leading to interbedded, intrabedded, and planar heterogeneous of coal reservoirs, and coal facies has significant controls to reservoir pore structure [11,[16][17][18][19][20][21][22]. However, the effect of coal facies on pore characteristics of coal reservoirs and its mechanism are still unclear. The coal reservoirs of the Upper Permian Longtan Formation in Guizhou Province have a complete range of coal ranks, and the effects of the thermal evolution on pore structure were analyzed [7,23]. However, the research of coal facies for these coal reservoirs is inadequate.
In this study, we take the medium-rank coal and highrank coal of the Longtan Formation in western Guizhou as research samples. Using the analyzing methods including macerals compositions, minerals compositions, vitrinite reflectance, and mercury intrusion porosimetry experiment, we divide the coal facies types, and then discussed the effect of coal facies on pore characteristics.  [31,40]

Geological Background
The western Guizhou area is located at the southern edge of the Upper Yangtze block, mainly including the eastern part of the southwestern Guizhou Depression, eastern Yunnancentral Guizhou uplift, Qiannan Depression and southeastern Panjiang Depression, and this area was subjected to the joint action of the Tethys tectonic domain and the Pacific Coastal tectonic domain [24]. First, the deep-water sedimentary zone in the south was further deepened, expanded, and developed with volcanic clastic turbidites followed by the appearance of the terrestrial, maritime, and land interactions and marine phases from west to east ( Figure 1) [25]. The marine intrusion during the Changxing stage of the Leping Period continued until the Middle Triassic. With the expansion of the marine intrusion, the terrestrial clastic deposits were gradually replaced by carbonate deposits [26]. The coal seam of the Longtan Formation in the western Guizhou region has experienced two stages of sedimentation and three phases of hydrocarbon generation, but there are obvious differences between the southwestern Panguan syncline and northeastern Shuigonghe syncline [23].
In the Panguan syncline, the first stage of sedimentary burial occurred in the Early-Middle Triassic, which was in a stable terrace stage and formed a giant thick shallow terrace-phase carbonate deposit. The second stage of sedimentary burial occurred in the Early-Middle Jurassic and formed a large terrestrial lake basin deposit, which reached 1200 m in thickness in this area (Figure 2(a)) [23]. The first phase of hydrocarbon generation occurred in the Early-Middle Triassic, with the paleo-temperature of the Longtan Formation reaching about 90°C at the end of the Middle Tri-assic. The reflectivity (R o ) of the mirror group was about 0.6%, corresponding to the long-flame coal stage, and the primary biogenic gas and early thermogenic gas were dominant in this stage. The second phase of hydrocarbon generation occurred in the Middle Jurassic, with the paleotemperature of the Longtan Formation reaching about 110°C at the end of the Middle Jurassic, exceeding the highest temperature of the previous stage, and the coal seam again. The third phase of hydrocarbon generation occurred in the Late Jurassic-Early Cretaceous, and under the influence of tectonic-thermal events, the paleo-geothermal gradient reached up to 5.5°C/100 m. The paleo-geothermal temperature of the Longtan Formation in the Panguan syncline rose sharply to about 140°C, with R o value in 0.9%-1.2%, corresponding to the stage of fatty coal. During this phase, condensate began to cracking into methane, generating a large amount of thermogenic methane.
In the Shuigonghe syncline, the first stage of sedimentary burial occurred in the Early-Middle Triassic, depositing about 5200 m of the Middle and Lower Triassic (Figure 2 3 Geofluids of thermogenic methane gas. The third phase of hydrocarbon generation occurred in the Middle Yanshanian period. Owing to the tectonic-thermal events in the Early Cretaceous, the paleo-temperature of the Longtan Formation rise sharply to about 220°C with R o values in 2.6%-3.5%, reaching the anthracite stage and producing a large amount of pyrogenic methane.

Sample Collection and Preparation.
A total of 20 fresh coal samples were collected from two wells for the analyses conducted in this study. Eleven of them were from well PQ-1 in the Shuigonghe syncline (Figure 3(a)), and the rest were from well YV-1 in the Panguan syncline ( Figure 3(b)). These fresh coal samples were sealed using plastic wrap immediately, and then transported to laboratory, and then were processed to powdered, thin slice, cylinder samples in the Laboratory of Coal Geology Bureau of Guizhou Province.

Analytical Methods.
In this study, the macerals compositions, minerals compositions, and vitrinite reflectance (R o ) of 20 coal samples were analyzed to determine their petrological characteristics. Based on these analyzing results, coal facies parameters were calculated to divide coal facies of these coal samples. Mercury intrusion porosimetry 4 Geofluids experiment was conducted to determine the pore features of these coal samples, including porosity, pore size distribution, and pore connectivity. All these experiments were conducted in the Laboratory of Coal Geology Bureau of Guizhou Province. The mercury intrusion porosimetry experiment was conducted using a Poremaster 60GT automatic mercury porosimeter following the China Petroleum and Natural Gas Industry Standard (SY/T 5346-2005). The equilibration time was set as 10 s, and evacuation time was 5 mins during experiments. Compared with the gas adsorption method, a more comprehensive range of pore sizes could be measured with mercury intrusion porosimetry, including pore characteristics of mesopore and macropore that cannot be measured by the gas adsorption method. During the MIP experiment, pressure of mercury injection increases with decreasing pore size. Pore radiuses were calculated using the Washburn equation [27] as follows: where P T is mercury injection pressure, MPa; σ is surface tension, set to be 0.48 N/m; θ is the contact angle between mercury and coal, set to be 141 degree; r max is the maximum capillary radius, μm.
In addition to the experiments above, gas content, reservoir pressure, and permeability of well testing were collected from the Guizhou Panjiang Coalbed Methane Development and Utilization Company Limited.  Figure 4: GI-TPI cross-plot (a) (modified from [28]) and GWI-VI cross-plot (b) (modified from [30]) for identifying coal facies.

Results and Discussion
4.1. Coal Facies Identification. Coal facies could be identified using coal facies parameters, including Texture Preservation Index (TPI), Gelification Index (GI), Groundwater Index (GWI), and Vegetation Index (VI) [11-13, 19, 28, 29]. The TPI refers the degradation intensity of plant tissue and the preservation degree of plant cells of coal forming plants, and can also be used to reflect the change of pH value. The GI is the ratio of gelation components to nongelatinization components, reflecting the wettability and duration of peat bog during the process of peat accumulation. Based on GI-TPI diagram, coal facies can be divided into land, piedmont, arid forest swamp, upper delta plain, humid forest swamp, lake, and lower delta plain [28]. GWI-VI diagram is another prevalent method to evaluate coal facies [29,30]. The GWI is based on gelation and mineral input, representing the water level during the formation of peat bog. The VI is the ratio of the components retaining the cell structure to the matrix, clastic material and granular components, reflecting the type of coal forming vegetation and its preservation degree. In this The GI-TPI cross-plot of studied coal samples was shown in Figure 4 The GWI values of coals from PQ-1 well range from 0.07 to 0.45, implying that the gelation degree and minerals input content of different coal seams in this well are highly variable. Compared with PQ-1 well, the samples from YV-1 well have lower and much concentrated GWI values (0.1-0.23) (Figure 4 Combined with GI-TPI and GWI-VI, the coal facies of studied samples were divided into shallow-water covered forest peat swamp facies, moist forest peat swamp facies, dry forest peat swamp facies, low peat swamp facies, and wet land herbaceous peat swamp facies.

4.2.
Coal-Forming Environment. The location of YV-1 well was closer to the Kangdian land, which is the main provenance of the eastern Yunnan-western Guizhou area, than the location of PQ-1 well, even though both of them are distributed in tidal flat-lagoon environment during the Late Permian ( Figure 1) [26,31]. During the sedimentation of 7 Geofluids the Longtan coal-bearing formation, the location of PQ-1 well has deeper water and less terrigenous input than that of YV-1 well [25]. As a result, PQ-1 well lithology includes limestone, sandstone, siltstone, shale, and coal, while there is no limestone in YV-1 well lithology (Figure 3). These differences between YV-1 well and PQ-1 well imply that coalforming peat swamps are heterogeneous in studied areas during the sedimentation of the Longtan coal-bearing formation.
Based on coal facies parameters, five peat swamp facies were recognized among studied coal samples from the Longtan coal-bearing formation. In YV-1 well, the coals were mainly formed in moist forest peat swamp, and only one sample (YV-1-8) was formed in wet land herbaceous peat swamp ( Figure 5(b)), indicating a relatively stable coalforming environment. A possible reason is that the terrigenous input rate offsets the effect of transgression, and then maintains a relatively stable water depth and forest density at the location of YV-1 well during the sedimentation of the Longtan coal-bearing formation. During this period, however, both the water depth and forest density at the location of PQ-1 well were unstable, and the coals in this well mainly formed in shallow-water covered forest peat swamp facies, moist forest peat swamp facies, and dry forest peat swamp ( Figure 5(a)). We deduce that this unstable environment is controlled by transgression. The evidence is that both shallow-water covered forest peat swamp facies samples (PQ-1-6 and PQ-1-9) are surrounded by limestone (Figure 3(b)). In addition, coals from YV-1 well have much less difference in macerals proportion compared with those coals from PQ-1 well (Figures 5(a) and 5(b)), also implying that the location of YV-1 well had more stable environment than the location of PQ-1 well during the sedimentation of the Longtan coal-bearing formation. The paleoenvironment governs the degree and feature of pore development by affecting the coal lithotype and macerals composition [8,9,12,13,16,32], and some macerals are porous ( Figure 6). Previous studies indicate that there is close connection between coal facies parameters and different sizes pores of medium-rank coal [10,19], while there are no data to support the correlation of high-rank coal and the dual control from coal facies and coal rank to pore structure. In this study, mercury intrusion porosimetry experiment was conducted to recognize the pore features of various coal facies and coal ranks. As shown in Figure 7, moist forest peat swamp facies samples have much higher mercury volume saturation than shallow-water covered forest peat swamp facies samples and dry forest peat swamp facies samples, implying that the degree of pore development follows the following order: moist forest peat swamp facies coal > shallow-water covered forest peat swamp facies coal > dry forest peat swamp facies coal.
As shown in Figure 10, coal samples have significant heterogeneity in pore size distribution, and different kinds of pores including macropores (>1000 nm), mesopores (100-1000 nm), transition pores (10-100 nm), and micropores (<10 nm) have different corrections with various coal facies parameters. We recognized that (1) the TPI has a weak negative correlation with mesopores but a positive correlation with macropores, and no significant correlation with micropores and transition pores (Figures 10(a) (4) the VI has a weak negative correlation with transition pores and mesopores, and a weak positive correlation with macropores (Figures 10(m)-(p)). These results indicate that coal facies play an important role in controlling the pore structure of coal reservoir. By combining macroscopic coal facies with microscopic pore structure and establishing multiscale coal reservoir model, the reservoir properties can be characterized more accurately.
When comparing the average pore size distribution of moist forest peat swamp facies samples from high-rank coal (PQ-1 well) and medium-rank coal (YV-1 well), we found that there is no significant difference between them ( Figure 11). The coal facies has much stronger control on pore size distribution than coal rank for the samples in this study.

Applications of Coal Facies in CBM Industry for Guizhou
Province. Coal facies have significant effects on pore structure in coal reservoirs [10,11,13,19,39]. Coal facies are determined by sedimentary environment of coal reservoirs, and are the bridge between sedimentary environment and reservoirs properties [12,15]. Recognizing the heterogeneous distribution of coal facies is important to reservoirs evaluation. Especially in Guizhou Province, where the coal reservoirs are multilayer thin coal seams [7]. In addition, previous studies have proved that coal rank is another controlling factor to reservoir properties (Figures 9 and 12) [33][34][35][36][37], and coal rank reflects tectonic evolution of coal reservoir. Generally, medium rank coal has much higher   10 Geofluids  12 Geofluids proportion of micropores and lower proportion of macropores than low and high rank coals. However, the experimental results in this study show that a same coal facies (MFPS) from PQ-1 well (high rank coal samples, Figure 3(b)) and YV-1 well (medium rank coal samples, Figure 3(a)), samples from PQ-1 well have lower median pore diameter and average pore diameter, but higher total pore area than those samples from YV-1 (Table 1). Moreover, SCPS coal facies samples from PQ-1 well have lower average pore diameter and total pore area than both MFPS and DFPS coal facies samples even though they are all high rank coal samples (Table 1). These results implying that coal facies plays more important role affecting pore structure than coal ranks for the Longtan Formation coals in western Guizhou area. Combining macroscopic coal facies (related to coal reservoir sedimentary environment) and coal rank (related to coal reservoir tectonic evolution), the study of microscopic pore structure characteristics of different coal phase and coal grade of coal reservoirs will provide a new theoretical basis for coalbed methane exploration and development in Guizhou Province.

Conclusions
Twenty coal samples of the Longtan formation were collected from PQ-1 and YV-1 wells in western Guizhou, to analyze the coal facies and their effects on pore characteristics. Three conclusions have been obtained.
(1) Based on coal facies parameters including TPI, GI, GWI, and VI, the coal facies of studied samples were divided into shallow-water covered forest peat swamp facies, moist forest peat swamp facies, dry forest peat swamp facies, low peat swamp facies, and wet land herbaceous peat swamp facies. The coals samples from YV-1 well were mainly formed in moist forest peat swamp, and only one sample (YV-1-8) was formed in wet land herbaceous peat swamp. The coal samples from PQ-1 well mainly formed in shallow-water covered forest peat swamp facies, moist forest peat swamp facies, and dry forest peat swamp (2) Pore development degree follows a descending order, namely moist forest peat swamp facies coal, shallow-water covered forest peat swamp facies coal, dry forest peat swamp facies coal. In contrast to the samples of shallow-water covered forest peat swamp facies and dry forest peat swamp facies, moist forest peat swamp facies samples have higher porosity and larger pore size (3) Coal facies plays more important role affecting pore structure than coal ranks for the Longtan Formation coals in western Guizhou area

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare that they have no conflicts of interest.