The shale mineral compositions are presented in Table 1. The XRD results reflect differences in the mineral compositions of the over-matured Niutitang Formation shale which is mainly composed of quartz and clay minerals. Based on the content of siliceous minerals, clay minerals, and carbonate minerals, Niutitang Formation shale can be divided into two lithofacies: siliceous shale and argillaceous shale (Fig. 2; Niu et al., 2018; Wang et al., 2017). The quartz content ranges from 25 wt.% to 56 wt.% averaging at 34 wt.% (Fig. 3b), whereas clay minerals content varies from 31 wt.% to 64 wt.% averaging at 52 wt.%. Clay minerals are composed of chlorite (with an average content of 10.7 wt.%) (Fig. 3c) and illite (with an average content of 41.3 wt.%) (Fig. 3e). Besides, there are albite (the average content is 12.8 wt.%) (Fig. 3e) and a small amount of pyrite (Fig. 3j) and siderite. The TOC content ranges from 0.90 wt.% to 1.33 wt.% with an average of 1.09 wt.%.
Sample TOC Q Ab Ill Chl Py Sd G-01 1.29 56 12 28 3 0 1 G-02 1.33 45 11 35 6 3 0 G-03 1.15 25 17 40 17 1 0 G-04 0.98 27 11 48 12 0 2 G-05 0.90 36 14 47 1 1 1 G-06 1.00 26 9 45 19 1 0 G-07 1.10 34 15 47 4 0 0 G-08 1.04 27 14 40 18 1 0 G-09 1.01 28 12 42 17 1 0 TOC. Total organic carbon; Q. quartz; Ab. albite; Ill. illite; Chl. chlorite; Py. pyrite; Sd. siderite.
Table 1. Mineral compositions (wt.%) and TOC (wt.%) content of Niutitang Formation shale
Figure 3. Four types of pores observed in the Lower Cambrian Niutitang Formation. (a) Organic matter pores near the pyrite; (b) organic matter strip; (c) interparticle pores between albite and chlorite and between albite particles; (d) interparticle pores between quartz and albite and intraparticle pores inside quartz; (e) intraparticle pores inside albite and illite; interparticle pores between albite and illite; (f) interparticle pores between orthoclase and albite; (g) interparticle pores between quartz; (h) and (i) intraparticle pores inside chlorite; (j) intraparticle pores inside pyrite framboid; (k) micro-fractures along the albite; (l) micro- fractures inside clay mineral. Ab. Albite; Chl. chlorite; Ill. illite; Kf. orthoclase; Py. pyrite.
Shale pore space is essential for hydrocarbon migration and reservoir space analysis, and numerous studies on pore morphology have been done using the FE-SEM method (Li et al., 2019; Sun et al., 2019; Xu et al., 2019; Zhang et al., 2019). According to FE-SEM observation, the pores of Niutitang Formation shale are at nano-scale to micron scale on the whole. Based on pores morphology, four types of pores are observed: organic matter (OM) pores, interparticle (InterP) pores, intraparticle (IntraP) pores and micro-fractures (Loucks et al., 2012). Among them, InterP pores are most developed, followed by IntraP pores. The OM pores are seldom developed and isolated in the studied shale samples, not while as commonly seen in Niutitang Formation shale in other area (Sun et al., 2019; Yang F et al., 2016). They usually show the bubble-like shape and develop near the pyrite (Fig. 3a), and there is also an elongated organic strip where OM pores are seldom developed (Fig. 3b). InterP pores are often observed between different mineral particles and their shapes in the study area are mainly angular. The width of InterP pores ranges from dozens of nanometers up to several micrometers depending on the size of the neighboring mineral grains. There are triangle pores between albite particles, plate-like pores between albite particles, or between albite and chlorite (Fig. 3c). Also, stylolite-like pores between albite and quartz (Fig. 3d), slit-like pores between albite and illite (Fig. 3e), triangle pores between quartz and orthoclase, albite and orthoclase (Fig. 3f) and irregular pores between quartz particles (Fig. 3g) can be observed. IntraP pores are pores developed inside mineral particles, and intercrystalline pore is included. Clay minerals make up about 50% in studied shale samples, and abundant pores inside clay minerals are observed. Linear pores often locate along cleavage sheet of illite (Fig. 3e), and sometimes they are bent (Fig. 3g). Besides, there are pores in quartz (Fig. 3d) and albite (Fig. 3e). Intercrystalline pores can often be observed inside pyrite (Fig. 3j), and their shapes are more regular. The pyrite framboid is surrounded by clay minerals, and its diameter is less than 10 μm. As for micro-fractures, they can extend almost 60 μm and present along rigid minerals such as albite (Fig. 3k) and cut through the clay minerals (Fig. 3l).
Low-temperature N2 adsorption is mainly used to analyze the pore volumes and surface area distributions of mesopores when the relative pressure (P/P0) was between 0.01 and 1. At relatively low pressure range (P/P0 < 0.45), the adsorbed quantity increases slowly, representing monolayer adsorption dominated by van der Waals' force. When relative pressure is between 0.45 and 0.8, the adsorbed quantity increases significantly, indicating multilayer adsorption. The adsorption curve and desorption curve depart at the relative pressure of 0.45, and the hysteresis loop begins to develop, which is related to capillary condensation taking place in mesopores (Gregg and Sing, 1982). At higher relative pressure (0.8 < P/P0 < 1.0), the adsorbed quantity increases sharply, but there's a limiting uptake over a range of high P/P0 (Fig. 4). For the studied samples, the isotherm shapes belong to type Ⅳ according to IUPAC classification (Sing et al., 1985), manifesting a continuous pore system from dominating micropore and mesopore to a few macropores.
Hysteresis appearing in the multilayer range of physisorption isotherms is usually associated with capillary condensation in mesopore structure. The shapes of hysteresis loops have been identified with different pore structures (Sing et al., 1985). For the studied Niutitang Formation shale, the hysteresis loops belong to type H3, indicating the aggregates of plate-like particles giving rise to slit-shaped pores (Fig. 4) (Sing et al., 1985), which is reflected in Figs. 3e, 3h and 3i.
Pore structure parameters calculated from the N2 adsorption and desorption isotherms are presented in Table 2. The BET specific surface area ranges from 4.15 to 18.09 m2/g with an average of 11.67 m2/g. The adsorption BJH volume which was calculated at the relative pressure of 0.9 ranges from 0.016 to 0.032 cm3/g with an average of 0.022 cm3/g. The average pore diameter varies from 4.72 to 10.83 nm with an average of 6.41 nm, which indicates mesopores are the dominating pore type. In addition, the maximum of adsorbed quantity is between 10.266 to 21.987 cm3/g with an average of 15.740 cm3/g.
Samples Average pore
G-01 10.83 4.15 0.016 10.266 G-02 6.72 10.84 0.020 16.943 G-03 4.72 18.09 0.025 18.056 G-04 6.20 16.30 0.032 21.987 G-05 5.51 11.91 0.021 14.809 G-06 6.83 8.22 0.018 12.728 G-07 5.74 16.53 0.031 21.016 G-08 5.57 10.87 0.019 13.518 G-09 5.61 8.15 0.018 12.335
Table 2. Pore structure parameters of the Niutitang Formation shale samples obtained from N2 adsorption-desorption isotherms
The pore size distribution can be described according to the distribution of pore volume for pore size distribution, including differential, incremental and cumulative pore volume distribution curves (Liu et al., 2015; Clarkson, 2013) where we gain the information about pore size range, dominating pore size and pore size distribution peaks. In the case of the tensile strength effect (Gregg and Sing, 1982), the adsorption branch data is more suitable to gain the pore size distribution (Bertier et al., 2016). Based on BJH model, the plots of dV/dlogW versus D (V is the pore volume, and W is the pore diameter) illustrate the pore size distribution of the Niutitang Formation shale (Fig. 5). It can be seen that the distribution of shale samples presents bimodal or multimodal (except for sample G-03) with the major peak between 1.9 and 2.5, 3 and 4, 90 and 110 nm. As a whole, it's observed that the pore size distribution has a wide range between 1.9 and 160 nm.
3.1. Shale Mineralogical Analysis and TOC
3.2. Pore Morphology from FE-SEM Images
3.3. Low-Temperature N2 Adsorption Isotherms
3.4. Pore Structure Parameters and Pore Size Distribution
Owing to the intricate shale pore structure, fractal dimension is an effective proxy to reflect the complexity of pore surface roughness and spatial structure (Shao et al., 2017; Hu et al., 2016; Li et al., 2016; Liu et al., 2015). There are several methods to calculate fractal dimensions of porous material, among which Frenkel-Halsey-Hill (FHH) model proves to be the most effective and is widely used because it's only based on nitrogen adsorption data (Yao et al., 2008).
FHH model can be described as follows.
where P is the equilibrium pressure; P0, the saturation pressure of nitrogen; V, the adsorption volume at P. C is a constant. If the pores possess fractal features, a plot of lnV versus ln(ln(P0/P)) will show a linear relationship, and K is the slope of the straight- line determined by adsorption mechanism.
D is the fractal dimension. D can be calculated by
when van der Waals' force dominates (at the early stage of the multilayer), the former equation fits better. When capillary condensation dominates, the latter equation fits better. Based on the N2 adsorption isotherm data, nine plots of lnV versus ln(ln(P0/P)) were obtained, and two linear segments occurred at a relative pressure of 0–0.45 and 0.45–1, showing two different adsorption mechanisms in two regions (Fig. 6). At a relative low pressure stage, the adsorption is mainly affected by monolayer coverage, and multilayer coverage begins to arise. As the adsorption proceeds, multilayer adsorption strengthens and capillary condensation takes control. The fractal dimensions D1 of the lower relative pressure and the fractal dimensions D2 of the higher relative pressure were both calculated by Eq. (2) and Eq. (3). The calculation results are shown in Table 3. Since some data of D1 and D2 calculated by Eq. (2) is less than 2, which is inconsistent with the theory of fractal dimension on pore structure, Eq. (3) seems more suitable. D1 ranges from 2.558 0 to 2.710 2 averaging at 2.651 6. D2 ranges from 2.541 5 to 2.765 2 averaging at 2.674 2. All fractal dimensions are between 2 and 3, demonstrating that the shale pores possess fractal characteristics. According to Pyun and Rhee (2004), surface fractal dimension represents the pore surface irregularity: the greater the value of the surface fractal dimension is, the more irregular and rougher the pore surface is. Pore structure fractal dimension characterizes the complexity of pore distribution and connectivity in space: when the value of the pore structure fractal dimension increases, the pore structure is more complicated (Li et al., 2016). For this paper, both D1 and D2 are adopted to describe the fractal characteristics. D1 is calculated based on the first linear segment and represents the pore surface fractal dimension. D2 stands for pore structure fractal dimension based on the second segment.
Figure 6. Plots of lnV versus ln(ln(P0/P)) reconstructed from the adsorption branch of the N2 adsorption-desorption isotherms of the Niutitang Formation shales samples.
Samples P/ P0=0–0.45 P/P0=0.45–1 D1 D2 Slope (K) R2 3K+3 K+3 Slope (K) R2 3K+3 K+3 G-01 -0.442 0 0.988 5 1.674 0 2.558 0 -0.458 5 0.995 2 1.624 5 2.541 5 G-02 -0.351 8 0.983 1 1.944 6 2.648 2 -0.389 9 0.992 9 1.830 3 2.610 1 G-03 -0.321 8 0.993 0 2.034 6 2.678 2 -0.234 8 0.998 4 2.295 6 2.765 2 G-04 -0.382 4 0.994 3 1.852 8 2.617 6 -0.308 3 0.994 7 2.075 1 2.691 7 G-05 -0.328 5 0.989 1 2.014 5 2.671 5 -0.281 3 0.998 1 2.156 1 2.718 7 G-06 -0.339 8 0.986 8 1.980 6 2.660 2 -0.348 0 0.992 3 1.956 0 2.652 0 G-07 -0.354 0 0.991 9 1.938 0 2.646 0 -0.289 7 0.998 0 2.130 9 2.710 3 G-08 -0.325 2 0.984 8 2.024 4 2.674 8 -0.291 9 0.998 3 2.124 3 2.708 1 G-09 -0.289 8 0.977 8 2.130 6 2.710 2 -0.330 0 0.989 7 2.010 0 2.670 0
Table 3. Calculation results of fractal dimensions
The relations among pore structure parameters (average pore size, BJH pore volume, and BET specific surface area) of Niutitang Formation shale are illustrated in Fig. 7. There is a negative correlation between BET specific surface area and average pore diameter (R2=0.44) while the relation between BJH pore volume and average pore diameter is not clear (Fig. 7a), which is consistent with Wei et al. (2013) and Zeng et al. (2014) and manifests that with the increasing average pore diameter, BET specific surface area gets smaller and the influence on pore surface area is greater. BJH pore volume increases with increasing BET surface area (R2=0.73 in Fig. 7b). The relations were previously found in the study of Barnett, Haynesville, Marcellus, Woodford, and Doig Phosphate shales in North America (Chalmers et al., 2012) and Wufeng Formation shale (Liang et al., 2015).
Figure 7. Relations between (a) BET specific surface area, BJH volume and average pore diameter, (b) BJH volume and BET specific surface area.
In order to illustrate the influence of shale mineral compositions on pore structure, the relations between pore structure parameters and mineral contents are given in Fig. 8. According to XRD results, Niutitang Formation shale is mainly composed of quartz and clay minerals. Both of quartz and clay minerals content have unclear relations with BET pore specific surface area and BJH pore volume (Figs. 8b–8c, 8e–8f). While there is a positive relation between average pore diameter and quartz content (Fig. 8a, R2=0.63) but a slightly negative correlation between average pore diameter and clay mineral content (Fig. 8d, R2=0.44). However, the relationships between TOC content and average pore diameter, BET surface area and BJH pore volume are ambiguous (Fig. 9). Since the TOC content of our samples is much lower than that of the same formation in another area (Wang et al., 2016; Yang F et al., 2016), it's deduced that organic pores that have greater pore diameter, pore surface area, and pore volume are more developed with higher TOC. Hence, TOC content isn't the dominating factor on pore structure for our samples.
Figure 8. Relationships between quartz content and (a) average pore diameter, (b) BET specific surface area, (c) BJH pore volume, and between clay minerals content and (d) average pore diameter, (e) BET specific surface area, (f) BJH pore volume.
The relationships between fractal dimensions and pore structure parameters are presented in Fig. 10. Both of D1 and D2 have negative relations with average pore diameter (R2=0.70 and R2=0.81 in Fig. 10a), indicating that fractal dimensions increase with decreasing pore diameter (Shao et al., 2017). D2 has a positive relationship with BET surface area (R2=0.60) while D1 has no apparent relationship with BET surface area (Fig. 10b). There are unclear relationships between D1, D2, and BJH pore volume (Fig. 10c). That's to say; fractal dimensions get larger with increasing pore surface area but with decreasing average pore diameter. Overall, both D1 and D2 are controlled by average pore diameter, and D2 is more correlated with BET specific area. Taking the FE-SEM observation and the low-temperature N2 adsorption-desorption isotherms into consideration, pores of Niutitang Formation shale vary differently in morphology, size and structure complexity. As a result, fractal dimensions can be used to quantify the characteristics of Niutitang Formation shale pores.
The relationships between fractal dimensions and mineral content are given in Fig. 11. Both of D1 and D2 decrease when quartz content increases (R2=0.45 and R2=0.60 in Fig. 11a). As a result of smooth surface, poor development of secondary pores and primary pores inside quartz particle, just some interparticle pores between quartz (Fig. 3g) which are easily filled with clay or OM, the increasing of quartz content lowers the heterogeneity of pore surface, pore development degree and the complexity of pore structure to some extent, so D1 and D2 come to be smaller. D2 increases when illite content increases (R2=0.45 in Fig. 11b). As for the relationship between clay minerals and fractal dimensions, the laminated structure of clay minerals is easily changed because of activities in post diagenetic stage and various pores and fractures are produced (Figs. 3e, 3h) to increase pore surface area and pore volume and form pore-fracture network, thus D2 slightly increase with increasing illite content. There is no apparent relationship between fractal dimensions D1 and D2 and TOC content (Fig. 11c), and similar discussions can be seen in the study of Loucks et al. (2012) and Liu et al. (2015). It's caused by the poor development of organic pores due to low TOC value and over maturity. This phenomenon testifies that organic pores have little impact on our studied shale pore structure and TOC is not the dominating factor influencing the fractal dimensions.