You are on page 1of 16

Marine and Petroleum Geology 107 (2019) 407–422

Contents lists available at ScienceDirect

Marine and Petroleum Geology


journal homepage: www.elsevier.com/locate/marpetgeo

Research paper

Pore connectivity and water accessibility in Upper Permian transitional T


shales, southern China
Mengdi Suna,b, Linhao Zhanga, Qinhong Huc, Zhejun Pana,b,∗, Bingsong Yud, Liangwei Sune,
Liangfei Baie, Luke D. Connellb, Yifan Zhangd, Gang Chengf,∗∗
a
Key Laboratory of Tectonics and Petroleum Resources, Ministry of Education, China University of Geosciences, Wuhan 430074, China
b
CSIRO Energy, Private Bag 10, Clayton South, VIC 3169, Australia
c
Department of Earth and Environment Sciences, University of Texas at Arlington, Arlington, TX 76019, USA
d
School of Earth Sciences and Resources, China University of Geosciences, Beijing 100083, China
e
Key Laboratory of Neutron Physics and Institute of Nuclear Physics and Chemistry, China Academy of Engineering Physics (CAEP), Mianyang 621999, China
f
College of Life Science and Technology, Beijing University of Chemical Technology, Beijing 100029, China

A R T I C LE I N FO A B S T R A C T

Keywords: Pore connectivity of shale controls shale gas migration and production behavior. The pore connectivity and
Transitional shale water accessibility in clay-rich Upper Permian transitional shales remain unclear. Contrast matching small-angle
Pore connectivity neutron scattering (CM-SANS) tests were used to determine the accessibility of water to pores in the transitional
Water accessibility shales. Complementary analyses with SANS, gas (CO2 and N2) physisorption isotherm, mercury intrusion ca-
Small angle neutron scattering
pillary pressure (MICP), helium ion microscopy (HIM), as well as field emission-scanning electron microscopy
Contrast matching
(FE-SEM) were conducted to study the pore connectivity and distribution characteristics of closed pores. The
results show that closed pores (inaccessible to nitrogen molecules and mercury) are mainly distributed in pore
diameters <10 nm and associated with organic pores and interlayer spaces of illite-smectite mixed-layer mi-
neral. The pore volume values obtained from MICP and N2 adsorption underestimate the large pores (pore
diameters >100 nm) in shales. Based on deuterated water CM-SANS tests, 87–98% of the pores (2–200 nm
diameters) are water-connected in transitional samples. The low accessibility to water is at pore-sizes of 5–10 nm
and 20–30 nm. Results from in-situ gas contents show that closed pores have a certain gas bearing capacity, but
micropores (pore diameters <2 nm) control the gas occurrence in transitional samples. The connectivity of the
organic pore network and between organic pores and surrounding interparticle pores is directly supported by FE-
SEM and HIM imaging. Overall, improving pore-fracture connectivity through effective fracturing techniques is a
means of mitigating the rapid decline in shale gas production.

1. Introduction Carboniferous-Permian shales developed in northern and northwestern


China (Zhang and Fu, 2018). Upper Permian transitional shales were
Shale gas industry in China has entered the stage of commercial mainly deposited in coastal swamp environment. It resulted in rapid
exploitation since 2014 (Guo, 2015). Up until the end of 2018, the changes in lithofacies with interlayers of coal beds or fine sandstone
Fuling gas field in southern China has produced more than 20 billion (Mani et al., 2015; Pan et al., 2015). The vast geological distribution
cubic meters of shale gas from overmature marine shales. Given the area and multiple gas sources (shale, coalbed, and tight-sand gas) of
successful research and development of marine shale in United States, transitional shales provide many opportunities for shale gas exploration
Chinese industry and academics are more focused on marine shales in and development in China.
the Sichuan Basin and its neighboring areas (Hao et al., 2013; Zhang Some comparative studies have carried out between transitional
et al., 2016, 2018). China has three types of organic-rich shales (by shale and marine shale (Mao and Guo, 2018; Yang et al., 2017). The
area: marine shale 26%, lacustrine shale 18%, and marine-terrestrial transitional shales exampled by Upper Permian shales have more clay
transitional shale 56%). Marine-terrestrial transitional shales mainly content (usually>50%) than that of marine shales, and high to over
include Permian Longtan Formation shales of southern China and maturity (Jiang et al., 2017). With the pore structures characterized by


Corresponding author. Key Laboratory of Tectonics and Petroleum Resources, Ministry of Education, China University of Geosciences, Wuhan 430074, China..
∗∗
Corresponding author..
E-mail addresses: Zhejun.Pan@csiro.au (Z. Pan), chenggang@mail.buct.edu.cn (G. Cheng).

https://doi.org/10.1016/j.marpetgeo.2019.05.035
Received 7 March 2019; Received in revised form 25 April 2019; Accepted 27 May 2019
Available online 30 May 2019
0264-8172/ © 2019 Elsevier Ltd. All rights reserved.
M. Sun, et al. Marine and Petroleum Geology 107 (2019) 407–422

Fig. 1. Location map of sampling wells in southern China (modified after Sun et al., 2017a,b).

field emission-scanning electron microscopy (FE-SEM), mercury intru- of closed pores can be obtained by comparing intrusion techniques and
sion, and low-pressure nitrogen physisorption (LPNP), transitional that from the scattering techniques. The second one takes advantage of
shales have a larger average pore size and a higher heterogeneity of OM a unique property of neutrons; the contrast-matching small-angle neu-
pore development, compared with marine shales (Ju et al., 2018; Zhang tron scattering (CM-SANS) technique can distinguish the accessible and
et al., 2017). The studies on the pore structures of the transitional shales inaccessible pores for a specific fluid in a certain pore size range. The
have mainly focused on geometrical attributes and size distribution of contrast variation was implemented by mixing deuterated solvents such
pores. However, closed pore (inaccessible to fluids) characteristics and as D2O with regular solvents such as H2O. CM-SANS technology was
pore connectivity in the shale matrix is still poorly understood. initially used in the closed pore and connectivity studies of other ma-
The characteristics of closed pores in various coals have been ex- terials such as coal char, coke, and silica (Liu et al., 2017; Melnichenko
tensively studied (Radlinski and Mastalerz, 2018; Radlinski et al., 2004; et al., 2012; Sakurovs et al., 2018a). More recently, deuterated water
Zhang et al., 2015b). In most coals, the closed porosity, for a given pore and methane were used in CM-SANS experiments on shale for the de-
size of 12 nm, decreases linearly with the total number of pores termination of closed porosity (Bahadur et al., 2018; Gu et al., 2016;
(Sakurovs et al., 2018b). The existence of closed pores in shale has been Ruppert et al., 2013). Meanwhile, CM-SANS method can avoid the er-
reported in some literature (Chen et al., 2018; Niu et al., 2017). rors that may be introduced by applying different models to calculate
Nevertheless, the controlling factors and gas-bearing properties of pore size distribution measured by various techniques. Additional
closed pores in shales are poorly studied. Through FIB-SEM (focused ion background information of CM-SANS method can be found in Chapter 8
beam-scanning electron microscopy) and nano-CT (nano-computed to- and references in Melnichenko (2015).
mography) characterization, closed pores not only refer to the isolated A quantitative characterization of pore connectivity in shale is still a
pores but also include the pore network systems that enclosed by the challenge in reservoir evaluation. For conventional reservoirs, the
surrounding minerals (Gu et al., 2015; Tang et al., 2016; Zhou et al., hysteresis of MICP is often used to calculate the mercury withdrawal
2016). Therefore, if the closed pores in organic-rich shale have a certain efficiency and pore body/throat ratio to indicate pore connectivity
gas-bearing capacity, it is essential to evaluate the geological and re- (Bary, 2006; Katz and Thompson, 1986; Li et al., 2015). On the mi-
coverable reserves of shale gas. crometer scale, the slopes of spontaneous imbibition curves and nano-
With the application of SANS/USANS (small-angle neutron scat- sized tracer diffusion coefficient and distance were used to assess the
tering/ultra small-angle neutron scattering) techniques, two methods connectivity of shale by some researchers (Gao and Hu, 2018; Hu et al.,
can be used to characterize the closed pores. One method is based on 2015b; Sun et al., 2017a). On the nanometer scale, the FIB-SEM and
the differences between different characterization techniques. SANS/ nano-CT were utilized to rebuild 3D visual models of limited areas by
USANS can provide information on the total porosity (open and closed numerical simulation (Starnoni et al., 2017; Zhao et al., 2018). On the
pores) and pore size distribution within a sample (Sun et al., 2017b, other hand, CM-SANS obtains the averaged data on the macroscopic
2018; Zhao et al., 2017). Mercury injection capillary pressure (MICP) scale (>mm3) and quantitatively characterizes pore connectivity at the
and low-pressure gas physisorption isotherm techniques allow for the nanometer scale.
characterization of accessible, or open, pores (Clarkson et al., 2013; Although waterless fracturing is being tried, hydraulic fracturing is
Davudov et al., 2018; Mastalerz et al., 2012). Therefore, the parameters still a key technology for shale gas development (Westwood et al.,

408
M. Sun, et al. Marine and Petroleum Geology 107 (2019) 407–422

Fig. 2. Schematic of experimental setup for in-situ gas content measurement and desorption curve.

Table 1
Mineralogical composition and organic petrographic characteristics of shale samplesa.
Sample Well Depth (m) Quantitative analysis of whole-rock minerals (wt.%) Relative content of clay TOC Ro (%) OM elemental analyses (wt.%)
minerals (wt.%) (wt.%)

Quartz Plagioclase Calcite Dolomite Pyrite Clays C I I/S %Sm N C H O

(X)Y1-16 (X)Y1 467.7 23 0 7 0 29 41 0 30 70 25 4.21 2.77 0.67 39.21 1.87 5.34


(X)Y1-2 (X)Y1 542.1 23 4 7 0 25 41 0 22 78 15 15.4 2.82 0.64 42.60 2.01 3.90
(X)Y1-1 (X)Y1 549 12 3 0 16 11 58 0 24 76 15 8.53 2.82 0.75 61.78 2.59 10.25
FY1-6 FY1 937 31 3 0 0 6 60 1 10 89 15 34.1 2.52 0.78 61.43 2.62 17.86
FY1-8 FY1 940 58 0 0 0 4 38 18 0 82 40 3.20 2.57 0.68 60.01 2.59 15.99
FY1-13 FY1 962 32 4 0 7 10 47 24 0 76 35 2.01 2.56 0.48 44.63 1.82 15.35

a
C = chlorite; I = illite; I/S = illite-smectite mixed-layer mineral; %Sm = percentage of smectite in mixed-layer mineral; TOC = total organic carbon;
Ro = vitrinite reflectance; OM = organic matter; A part of data for (X)Y1-1 were previous published by Sun et al. (2017a,b).

2017). A large volume of water (2–6 million gallons) was pumped into hydraulic fractures by reducing the relative permeability of gas (Gupta
shale formations to create hydraulic fractures for economic production et al., 2018). Therefore, it is significant to study the water accessibility
(Singh, 2016). Across all shale plays in US, only 6–10% of the pumped in shale which is conveniently investigated here by CM-SANS that uses
water is recovered (Singh, 2016). The abundant retention of injected H2O/D2O as the solvents.
water in shale matrix keeps the hydrocarbons from migrating into In this work, six core samples of two wells from Upper Permian

409
M. Sun, et al. Marine and Petroleum Geology 107 (2019) 407–422

Table 2 to 70° for quantitative mineralogy analyses. Total organic carbon (TOC)
Parameters used in calculating Scattering Length Density (SLD)a. and vitrinite reflectance (Ro) of samples were determined by a Leco
Mineral Density, (g/ Elemental composition SLD ( × 1010 cm−2) carbon analyzer and an MPV-3 microphotometer 806 apparatus, re-
cm3) spectively. The organic matter extraction and elemental (CHON) ana-
lyses of samples were conducted following the standard methods of GB/
Quartz 2.65 SiO2 4.18
T19143-2003 and GB/T19144-2010.
Calcite 2.71 CaCO3 4.69
Dolomite 2.86 CaMg(CO3)2 5.44
Plagioclase 2.61 NaAlSi3O8 3.97
2.2. SANS experiments
Pyrite 5.01 FeS2 3.81
Illite 2.70 K(Al3Mg)(Si7Al)O20(OH)4 3.80
Smectite 2.35 Na0.7(Al3.3,Mg0.7) 3.26 SANS measurements were carried out on the General Purpose SANS
Si8O20(OH)4H2O (GP-SANS) instrument at Oak Ridge National Laboratory (ORNL), USA.
Chlorite 3.00 (Mg9Al3)(Al3Si5)O20(OH)16 3.75
A wide Q range (0.001 Å−1 ≤Q ≤ ∼0.5 Å−1) was obtained by collecting
Kerogen 1.50 C90H47O15N 3.39
SANS data at three sample-to-detector distances of 19.2 m, 10.7 m, and
a
Online calculators are available at https://www.ncnr.nist.gov/resources/ 1.7 m using neutron wavelengths of λ = 4.57 Å (10.7 m, 1.7 m) and
activation/. λ = 12 Å (19.2 m), where Q = 4πλ−1 sin (θ) and 2θ is the scattering
angle. Core samples were pulverized and sieved into specific size ranges
Longtan formation transitional shales were selected. Although these (35–80 mesh), such that more than 99.6% neutron transmission is
limited samples cannot adequately represent the transitional shales, the achieved (Sun et al., 2018). Samples were then oven-dried at 60 °C for
methodology is applicable for other organic-rich transitional shales in 48 h before loaded into 1 mm path length quartz glass cells for mea-
evaluating their pore connectivity and water accessibility. With a surement, yielding an effective thickness of ∼0.5 mm (Bahadur et al.,
combination of SANS, CM-SANS, MICP, low-pressure gas physisorption, 2015).
FE-SEM, HIM, the main objectives of this work are as follows: (1) CM-SANS experiments were performed on the Suanni SANS spec-
characterize the closed pore (e.g., pore volume, pore size distribution, trometer at 20 MW China Mianyang Research Reactor (CMRR). The Q
and relationship with gas content); (2) assess the pore connectivity; (3) range of this study covered from 0.003 Å−1 to 0.3 Å−1 with neutron
determine the water accessibility in clay-rich transitional shales. This wavelengths of λ = 5.3 Å (4 m, 1 m) and λ = 8 Å (10 m). The mixtures
work is beneficial to understand the pore connectivity and water ac- of H2O/D2O (water/d-water) with different SLDs were added into the
cessibility of transitional shales, and provides guidelines for future ac- 1 mm path length Hellma cells (Hellma Analytics, 404-1-46). The shale
tive development. samples were slowly poured into the cells approximately 8 h before
SANS analyses to ensure that samples were fully in contact with H2O/
D2O mixtures.
2. Shale samples and experiments The raw 2D scattering patterns were corrected by subtraction of
scattering data of empty-cell and acid-washed quartz sand with the
2.1. Sampling and analyses same particle size (35–80 mesh). Corrected scattering from acid-washed
quartz sand is to eliminate the influence of pore space between sample
A total of six fresh core samples were selected from the Upper particles. The original SANS data was reduced to 1D absolute scale via
Permian marine-terrestrial transitional Longtan Formation in Well (X) IGOR Pro software. The data were analyzed using IRENA macros
Y1 and Well FY1 (Fig. 1). The powdered whole-rock samples (less than (https://usaxs.xray.aps.anl.gov/software/irena) and PRINSAS software
200 mesh) and clay minerals separated using gravity method were (http://smallangle.org/content/software). Prior to the analysis, the
carried out on a Bruker X-ray diffractometer with a step of 0.02° from 2° background scattering intensity, as calculated from the slope of a plot of

Fig. 3. SANS profiles of (a) samples and (b) corrected for background.

410
M. Sun, et al. Marine and Petroleum Geology 107 (2019) 407–422

Table 3
Values of porosity and other parameters obtained from SANS analysesa.
Sample ID SANS SLD H2O/D2O PI porosity PDSP porosity PDSM porosity PDSM PDSM NDP PDSM porosity PDSM Pore volume
( × 1010 cm−2) volume (2–400 nm) (%) (2–400 nm) (%) (2–400 nm) (%) SSA (m2/ ( × 1017cm−3) (3–400 nm) (%) (2–100 nm)
ratio g) (10−2 cm3/g)

(X)Y1-16 3.86 1:1.75 10.8 13.6 13.46 24.81 27.5 12.00 2.37
(X)Y1-2 3.71 – 3.5 4.12 3.65 6.13 6.3 3.34 0.70
(X)Y1-1 3.93 1:1.83 2.9 3.78 3.38 11.20 9.5 2.88 0.93
FY1-6 3.63 1:1.52 3.9 4.79 4.73 5.98 3.8 4.54 0.97
FY1-8 3.90 – 6.3 7.83 7.33 13.11 7.4 7.00 2.68
FY1-13 3.92 1:1.81 4.4 5.82 5.42 5.95 2.8 5.28 0.97

a
PI= Porod invariant; PDSP= Polydisperse Spheres; PDSM= Polydisperse sphere model; SSA= Specific surface area; NDP= Number density of pores.

2.4. FE-SEM and HIM imaging

Before the imaging with FE-SEM (Zeiss Merlin) equipped with an


energy-dispersive spectrometer observation, samples were prepared by
cutting as 5 mm × 5 mm × 2 mm slices which parallel to bedding and
ion milled perpendicular to the bedding plane by Ilion + II (Model 697,
Gatan) for a smooth surface. Before imaging, the samples were coated
with carbon at a thickness of 5 nm on the surface to enhance the
electrical conductivity. Then, FE-SEM images were collected at a
2–4 mm working distance with an acceleration voltage of 1 kV. HIM
(Zeiss Orion NanoFab) has a higher resolution than FE-SEM and can
produce ultra-high resolution images of nonconductive surfaces, espe-
cially useful for observing sub-10 nm pores in shale. The instrument
was operated at a working distance of 7 mm, a beam current set at
0.9–1 pA, and a scan dwell time of 0.5 μs.

2.5. Gas content test

The in-situ gas content in this study can be divided into lost gas
content, desorbed gas content, and residual gas content (Ma et al.,
2015; Tang et al., 2017). The AD-LJ-150-3 in-situ gas desorption ap-
paratus manufactured by China University of Geosciences was used to
Fig. 4. SANS-generated pore size distribution (from PDSM analysis) for tested carry out the desorbed gas and residual gas content (Fig. 2a and b). The
samples. rate of minimum response flow is 0.1 ml/h, while the precision of
pressure and temperature measurements are 0.01 MPa and 0.1 °C, re-
spectively. During the first 3 h, the instrument recorded one datum
Q4I(Q) vs. Q4, was subtracted.
every 10 s. The initial gas desorption rate was measured at reservoir
temperatures by placing at least 1 kg core sample into the desorption
canister. The lost gas content was estimated by the initial gas deso-
2.3. Mercury injection capillary pressure (MICP) and low-pressure gas
rption rate and the time of samples coring from underground to can-
physisorption isotherm analyses
ister. Reservoir temperature setting was based on sample depth and
geothermal gradient in this area. Then, the samples were heated to
MICP tests were performed using a Micromeritics AutoPore V 9620
90 °C until the released gas cannot be detected. Thus, the desorbed gas
with an initial filling pressure of 5 psi (0.034 MPa) to a maximum
content includes the amount of gas desorbed at reservoir temperature
pressure of 60000 psi (413 MPa). This pressure range corresponds to a
and 90 °C. Finally, the core samples were crushed into powder in the
pore-throat range of 36 μm to 3 nm based on Washburn equation using
canister to obtain the residual gas content. A schematic diagram of an
physical constants: surface tension = 485 mN/m, and contact
in-situ gas content curve is shown in Fig. 2c. The measured in-situ gas
angle = 130° (Katz and Thompson, 1986). Prior to the test, 1 cm-sized
content was converted into the gas content under standard temperature
cuboid samples were oven-dried at 60 °C for 48 h to remove moisture
and pressure conditions (STP, 0 °C, and 101.325 kPa).
and then cooled down to room temperature (∼23 °C) in a desiccator.
Three bulb penetrometer with 0.096 mL stem volume is most suitable
for shale samples with low-porosity (less than 8%). The mercury com- 3. Results
pressibility and blank penetrometer errors in high pressure have been
corrected. 3.1. Geochemical characteristics
N2 and CO2 adsorption measurements were conducted at 77 K and
273 K by a Micromeritics ASPS 2020 apparatus. The samples were The results of mineral composition, TOC content, vitrinite re-
crushed into 35–80 mesh as same in the SANS test and degassed at flectance index and organic matter elemental value for samples are
110 °C under vacuum for about 5 h prior to the adsorption experiments. listed in Table 1. All shale samples in this study with a vitrinite re-
The pore volume and pore size distribution of pores with diameters flectance greater than 2% are over-mature and within the dry gas
between 0.35 nm and hundreds of nanometers were characterized window. TOC contents of marine-terrestrial transitional shale vary
based on density functional theory (DFT) (Sang et al., 2018; Sun et al., greatly (2.01%∼34.1%) due to the change of microfacies. Clay and
2016). quartz are the dominant minerals, and the illite-smectite mixed-layer
mineral accounts for 70% or more of the clays in these shale samples. In

411
M. Sun, et al. Marine and Petroleum Geology 107 (2019) 407–422

Fig. 5. Combined SANS raw data and water contrast matching data for samples showing pores in all samples being penetrated by deuterated water.

general, clay content is higher than quartz content in transitional shale convergence when data span large Q ranges and intensity (Ilavsky and
samples, which is consistent with the results in literature studies (Jiang Jemian, 2009). The polydisperse size-distribution model (PDSM) was
et al., 2017; Mao and Guo, 2018). Brittle minerals such as pyrite and used to analyze to data in the Irena macros software with the SLDs
carbonate mineral are ubiquitous in these shale samples. Though CHON shown in Table 3. The assumption of the PDSM was that pore network
analysis, the average elemental composition of kerogen was calculated. conforms to polydisperse spherical pore network in shale and ignores
The scattering length density (SLD) of each components, calculated the contribution from the structure factor term. In this case, the scat-
based on above analyses, are listed in Table 2. tering intensity in this PDSM is defined as (King et al., 2015; Zhang
et al., 2019b):
3.2. Pore characterization I (Q) = (ρ1 − ρ2 )2 ∑ P (Q, r )2F (r ) V (r )2NΔr (1)
3.2.1. SANS and CM-SANS analyses where ρ1 and ρ2 are the SLDs of shale matrix and pores, r is the radius of
Fig. 3 shows the raw SANS profiles and the profiles corrected for sphere, P (Q, r ) is the form factor of spheres with radius r, F (r ) is the
background scattering. An approximate relation between the pore ra- pore size distribution function, V (r ) is the volume for spheres with
dius (r) and scattering vectors (Q), (r ≈ 2.5/Q) has been used in poly- radius r, N is the number of sphere. The calculated porosity, specific
disperse spherical pores in shale samples (Radlinski et al., 2000). A surface area (SSA), and number density of pores (NDP) for samples are
comparative analysis of data was carried out for Q range between listed in Table 3. A plot of pore size distribution (PSD) of samples for
0.25 Å−1 and 0.00125 Å−1 (equivalent to pore diameter between 2 nm comparison is given in Fig. 4.
and 400 nm). The formula that used for the SLD calculation of shale In addition, the Porod invariant (PI) method has been utilized to
samples was the same as in previous studies (Bahadur et al., 2015; Sun calculate the porosity when shale samples are treated as a two-phase
et al., 2018). system. The porosity can be calculated by Eq. (2) using a scattering
A total non-negative least-squares (TNNLS) method was used in the contrast ( Δρ ) (Porod et al., 1982):
Irena macros to obtain porosity, specific surface area (SSA), number

density of pores (NDP; the number of pores per unit volume (in units of
cm−3)), and generate pore size distribution (PSD) within a certain size
∫ Q2I (Q) dQ = 2π 2 (Δρ)2∅ (1 − ∅)
0 (2)
range. An optional ‘Q-weighting’ of intensity is implemented to improve

412
M. Sun, et al. Marine and Petroleum Geology 107 (2019) 407–422

Fig. 6. Pore size distributions derived from neutron scattering data for dry and water-contrast-matching samples to show water-inaccessible pores.

Table 4
Pore parameters of samples from MICP and gas adsorptiona.
Sample ID MICP Bulk MICP MICP porosity Mercury Body-throat N2 Pore volume N2 SSA N2 APD CO2 Pore volume CO2 SSA
density (g/ porosity (%) (3–400 nm) (%) Withdrawal ratio (2–100 nm) (m2/g) (nm) (10−2 cm3/g) (m2/g)
cm3) Efficiency (%) (10−2 cm3/g)

(X)Y1-16 2.94 2.59 2.46 28.9 32.5–8.2 1.00 5.72 8.35 0.26 8.66
(X)Y1-2 2.89 3.21 2.00 41.4 3.5–1.4 0.28 0.51 30.00 1.41 40.38
(X)Y1-1 2.07 3.70 2.80 86.1 1.3–1.1 0.76 7.58 4.71 2.12 62.21
FY1-6 1.94 5.79 3.60 47.8 1.6–1.1 0.25 0.44 28.10 2.29 62.33
FY1-8 2.55 6.31 5.80 35.6 13.3–1.8 2.18 7.58 13.02 0.46 15.24
FY1-13 2.56 3.27 3.10 33.9 11.4–4.5 0.47 1.74 11.95 0.072 2.32

a
APD = Average pore size distribution.

The PI porosities of samples generated from the PRINSAS software model (Radlinski, 1999; Sun et al., 2017b), the two algorithms are
are listed in Table 3. In the Porod invariant method, the SANS intensity consistent with each other.
is considered to follow the power law scattering and independent of the rmax

shape of the scattering object (Porod et al., 1982). Nevertheless, the I (Q) = Δρ2

∫ V 2F (r ) P (Q) dr
SANS intensity shows a deviation in the higher-Q region due to non- rmin (3)
fractal scattering of small pores (Bahadur et al., 2018). The porosity ∞
calculated from PDSM model is not restricted to fractal pore structure where V̄ = ∫ V 2f(r)dr is the average pore volume and
and general higher than the PI porosity. 0
2
sin(Qr) − (Qr)cos(Qr)
The porosity and PSD of shale also can be calculated from the P(Q) = 9 ⎡ ⎤ is a form factor for spheres. The PDSP
⎣ (Qr)3 ⎦
polydisperse spheres (PDSP) model in the PRINSAS software. The PDSP porosities of specimens are very close to PDSM porosities. The slight
and the PDSM models both assume that the pores are in spherical shape discrepancies may be due to the differences in data fitting.
and have a random size distribution. According to Eq. (3) of PDSP In order to investigate the water accessibility of shale, four samples

413
M. Sun, et al. Marine and Petroleum Geology 107 (2019) 407–422

Fig. 7. Intrusion/extrusion curves (a) and pore volume distribution (b) from MICP and gas physisorption isotherms (low-pressure N2 ad/de-sorption isotherms (c)
and low-pressure CO2 adsorption isotherms (d)) for all samples.

were tested on CMRR for CM-SANS following their SANS measurements to pore throat size during MICP tests and the results are shown in
at ORNL. Based on the SLD difference between water (SLD: Fig. 7b. Fig. 7b shows most pore volumes are connected by pore-throat
−0.56 × 1010 cm−2) and heavy water (SLD: 6.39 × 1010 cm−2), the with diameters smaller than 50 nm. Meanwhile, the efficiency of mer-
contrast matched liquid mixture of H2O/D2O corresponding to the cury withdrawal and pore-body-to-throat aspect ratio (mercury sa-
average SLD of shale was mixed with the samples about 8 h before the turation more than 80%) can be obtained from the hysteresis loop be-
CM-SANS test. The ratio of H2O/D2O at the contrast match point for tween intrusion and extrusion curves (Fig. 8 and Table 4).
each shale samples is given in Table 3. Given the different conditions of Nitrogen adsorption/desorption and carbon dioxide adsorption
the two SANS instruments, the absolute intensity of neutron scattering isotherms of shale samples are shown in Fig. 7c and d, respectively. The
was calibrated by the same standard sample of the Al-4. Fig. 5 shows pore volume, surface area and average pore diameter of the samples
the decrease of scattering intensity of contrast-matched shale samples within a specific pore diameter (based on Kelvin Equation) can be
due to accessible pores filled with water. The residual scattering in- calculated by nitrogen adsorption and the results are shown in Table 4.
tensity can be used to reveal information about inaccessible pores to BET specific surface area of N2 adsorption does not include the con-
water (Bahadur et al., 2018). The PSDs and porosities of original tribution of micropores; therefore the BET values cannot represent the
samples and contrast-matched samples estimated using Eq. (1), are real surface areas of shales. Nevertheless, the reproducible BET values
shown in Fig. 6. It is evident from Fig. 6 that water accessibility in shale under carefully controlled conditions have been used to evaluate the
samples varies with pore sizes. SANS data within the same Q range were “apparent” surface area of shale for decades (Gregg and Sing, 1982;
used for the analysis. Mastalerz et al., 2012; Sun et al., 2018). Based on the CO2 adsorption
data, the micropore surface area and DFT micropore volume are de-
3.2.2. MICP and gas physisorption rived and also shown in Table 4. The difference between SSA derived
MICP tests give information about pore volume of mercury acces- from N2 and CO2 adsorption reflects the contribution of different sizes
sible pores and pore-throat size distribution. Pore volume and porosity and types of pores to surface area.
of samples obtained from the mercury intrusion curves were listed in
Table 4. Fig. 7a is used to calculate the connective pore volume related

414
M. Sun, et al. Marine and Petroleum Geology 107 (2019) 407–422

Fig. 8. Plots of pore-throat diameter and pore body-to-throat aspect ratio vs. mercury saturation for all samples.

3.2.3. Observation of FE-SEM and HIM development of OM pores. Fig. 9d clearly shows that the dissolution
According to the classification proposed by some researchers, pores leads to further expansion of Inter-pores and facilitates the connection
in shale are divided into interparticle pores (Inter-pores), intraparticle of OM pore network. In such a well-connected pore network, hydro-
pores (Intra-pores) and organic matter (OM) pores (Loucks et al., 2012; carbon will migrate out from the OM pore system along the conductive
Milliken et al., 2013). The type and connectivity of pores affect the pathway. In addition, some organic particles are mixed with clay mi-
permeability and wettability of shale reservoirs (Gao and Hu, 2016). nerals to form organo-clay complexes (Fig. 9e), which developed nu-
FE-SEM images (Figs. 9–10) show that various types of pores are de- merous pores at the edge of contact between minerals and OM (Fig. 9f).
veloped in the Upper Permian marine-terrestrial transitional shales. Pyrite mainly exists in the form of framboidal grains and associates
Meanwhile, the OM pores and Inter-pores are mainly developed in with OM in Longtan shale samples (Fig. 10a). Dissolution causes the
Longtan Formation shales (Fig. 9a). Abundant OM pores within nano- deformation of crystals in pyrite framboids (Fig. 10b), while the crystals
and micron-meter scales are observed in FE-SEM and HIM images are generally of euhedral shape in marine shales. A complete dissolu-
(Fig. 9b and c), and these pores with bubble-like, rounded or elliptical tion of some crystals forms crystal-mold pores in pyrite framboids
shapes are unevenly distributed within OM. Considering that the ma- (Fig. 10a). Also, the interparticle space between crystals in some pyrite
turity of Longtan shale range from 2.52% to 2.82% within the dry gas framboids are filled with OM (Fig. 10c). The HIM image shows that the
generation window, the OM pores may mostly come from the genera- OM within pyrite framboids contains OM pores with a similar pore size
tion and expulsion of hydrocarbons (Bernard and Horsfield, 2014). less than 20 nm (Fig. 10d). Sheet-like pores are usually occurred in clay
Fig. 9c shows a HIM image of a higher magnification of the OM pore particles in samples (Fig. 10e). From the observation of FE-SEM
network. Numerous pores less than 10 nm in diameter can be seen in (Fig. 9a, d, 10f), the dissolution mostly occurs along the interparticle
the image. The internal structure observed in the HIM image reveals pores between brittle and plastic minerals. Fig. 10f demonstrates that
that the OM pore network is internally connected. Meanwhile, a few the quartz cement produced by dissolution contains some tiny pores.
large OM pores made up of several small interconnected OM pores
(Fig. 9c). The OM pores in Fig. 9d illustrate the heterogeneity of OM 3.3. In-situ gas content
pore development in organic matter. The two closed organic particles
shown in Fig. 9d with likely the same maturity have very different Bases on the volume of gas released over time, the measured

415
M. Sun, et al. Marine and Petroleum Geology 107 (2019) 407–422

Fig. 9. FE-SEM and HIM photomicrographs for Longtan transitional shale samples. (a) Pore types overview of sample (X)Y1-16; (b) OM pores in a single OM grain of
sample (X)Y1-1; (c) An enlargement of organic pores by HIM in the rectangle area marked in Fig. 9 (a); (d) Heterogeneously distributed OM pores within OM grains
and dissolution of inorganic minerals of sample (X)Y1-16; (e) Organo-clay complex of sample FY1-6; (f) An enlargement of abundant pores in organo-clay complex in
the rectangle area marked in Fig. 9 (e).

desorbed and residual gas contents and the estimated lost gas content inaccessible pores for mercury and nitrogen molecules range from 2.8%
were listed in Table 5. All the samples have some residual gas, ap- to 79.5% and from 18.3% to 74.2%, respectively (Table 5). The fraction
proximately average 29% of the total in-situ gas content, which cannot of closed-pores of nitrogen is generally higher than that of mercury,
be desorbed. Moreover, methane accounted for 89.1% of the total except for sample (X)Y1-16. In order to characterize the closed-pore
content by analyzing the composition of the desorbed gas. distribution, the pore volume distributions derived from SANS, gas
adsorption (N2 and CO2), and MICP results were plotted in Fig. 11.
The pore volume distribution curves of each sample derived from
4. Discussion
MICP and SANS cross each other within the pore-size range between 30
and 70 nm (Fig. 11). The pore volume distribution from SANS is higher
4.1. Closed (inaccessible) pores analyses
in magnitude than that from MICP after these cross points; similar re-
sults were found in U.S. shale oil samples (Zhao et al., 2017) and other
Pore parameters determined from SANS, MICP, and nitrogen ad-
gas-bearing shale samples (King et al., 2015). One reason is that SANS
sorption offer important insight into the structure of closed pores in
measured total porosity of the pores while MICP measures pores in-
shale samples (Tables 3–4). The parameters of MICP and nitrogen ad-
truded by mercury. That is to say, there are closed pores in this size
sorption in the pore-throat range of 3–400 nm and in the pore size range
range where MICP cannot access. Another reason is related to MICP
of 2–100 nm were used to calculate the fraction of closed pores by
itself. As shown in Figs. 9b and 10b, some OM or inorganic pore net-
comparing with SANS results in the same size range, respectively
works containing macropores (>50 nm) connected with surrounding
(Table 5). The SANS porosity and pore volume, especially obtained
space through smaller pore throats. Other studies also have reported
from PDSM model, are much higher than that from MICP and nitrogen
that MICP would underestimate the volume of some macropores in fine-
adsorption in each sample (Tables 3–4). The total fractions of

416
M. Sun, et al. Marine and Petroleum Geology 107 (2019) 407–422

Fig. 10. FE-SEM and HIM photomicrographs for Longtan transitional shale samples. (a) Pyrite framboids distributed in organic matter of sample FY1-8; (b)
Dissolution causing the deformation of crystals in pyrite framboid of sample FY1-8; (c) Organic matter within the space between pyrite crystals of sample (X)Y1-2; (d)
An enlargement of organic pores within a pyrite framboid by HIM in the rectangle area marked in Fig. 10 (c); (e) Sheet-like pores in clay particles of sample FY1-13;
(f) Quartz cement produced by dissolution containing some tiny pores of sample FY1-13.

Table 5
In-situ gas content and multiple-scale fraction of closed-pores of studied samplesa.
Sample ID In-situ gas content (cm3/g,STP) Fraction of closed-pores Fraction of closed-pores from LPNP and SANS (%) Fraction of water
from MICP and SANS (%) inaccessible Pores (%)
Lost gas Desorbed gas Residual gas 2–5 nm 5–10 nm 10–50 nm 50–100 nm Total
content content content

(X)Y1-16 0.88 1.69 2.29 79.5 85.7 72.2 3.2 75.4 57.8 4.31
(X)Y1-2 0.57 2.12 0.94 40.1 81.8 57.7 9.5 73.7 60.0 –
(X)Y1-1 0.51 5.71 0.47 2.8 42.9 0.1 16.4 73.0 18.3 5.28
FY1-6 0.89 4.15 1.22 20.7 99.9 79.0 44.7 85.2 74.2 2.04
FY1-8 0.52 1.23 0.92 17.1 59.5 41.3 0.1 28.5 18.7 –
FY1-13 0.67 1.39 1.43 41.3 67.6 65.5 20.6 76.3 51.5 12.63

Fraction of closed-pores from LPNP and SANS =(V PDSM Pore volume (2–100 nm)-V N2 Pore volume (2–100 nm))/V PDSM Pore volume (2–100 nm).
Fraction of water inaccessible Pores = ∅contrast matched sample/∅ original sample (data in Fig. 6).
a
Fraction of closed-pores from MICP and SANS =(∅PDSM porosity (3–400 nm)- ∅MICP porosity (3–400 nm))/∅PDSM porosity (3–400 nm).

417
M. Sun, et al. Marine and Petroleum Geology 107 (2019) 407–422

Fig. 11. Comparison of pore volume distribution obtained from SANS, MICP, and gas (CO2 and N2) adsorption for the studied samples.

grained rocks due to complicated pore shapes (Hildenbrand and Urai, Since the 3 nm pore-throat is the limit of MICP, the data in this area
2003; Wang et al., 2016). Meanwhile, the underestimation is usually show some fluctuations (Fig. 11). In some shale samples ((X)Y1-2, FY1-
related to the pore body-to-throat aspect ratio (Abell et al., 1999; 6, FY1-8, FY1-13), the pore volumes of MICP were higher than the
Dewhurst et al., 1998). results of SANS in the mesopore range (2–50 nm) (Fig. 11). Since MICP

418
M. Sun, et al. Marine and Petroleum Geology 107 (2019) 407–422

be detected by gas adsorption (Sang et al., 2018). In the study of marine


Niutitang gas shale, the closed pores are mainly related to organic pores
(Sun et al., 2018). Therefore, closed pores, inaccessible to nitrogen and
mercury, are mainly distributed in the pore-size range of less than
10 nm and associated with OM, clay minerals and organo-clay com-
plexes in transitional shales.

4.2. Shale connectivity and water accessibility

The connectivity of pores in the shale matrix is a significant con-


trolling factor for hydrocarbons migration, enrichment, and production
behavior (Hu et al., 2015a; Zhang et al., 2015a). Over the past decades,
experimental permeability measurements are often used to evaluate
pore connectivity in rock (Cui et al., 2018; Davudov and Moghanloo,
2018). However, for shale with extremely low matrix permeability, the
determination of absolute permeability is challenging, and the aniso-
tropy is extreme. Therefore, the pore connectivity can be reflected by
the fractions of water, mercury, and nitrogen molecule inaccessible
pores in shale (King et al., 2015; Zhao et al., 2017; Bahadur et al.,
2018).
In this work, CM-SANS tests were used to evaluate the pore con-
nectivity of Longtan transitional shales. Shown in Fig. 12, under am-
bient pressure, accessibility of pores to water varies with pore size and
Fig. 12. Fraction of accessible pores to water in four Longtan transitional shale show similar trends in four samples. The lowest accessibility to water is
samples by H2O/D2O contrast matching.
associated with pore sizes of 5–10 nm, which also exhibited a peak at
the CM-SANS volume distribution in Fig. 6. Approximately 87–98% of
represents the distribution of pore throats, the pore throats connect the the pores (2–200 nm diameters) in these four shale samples are water-
pores with large pore sizes, resulting in an overestimate of pore volume connected (Table 5). In the CM-SANS tests, the water accessibility of the
in this range. Therefore, MICP would overestimate the pore volume of shale samples in the pore-size range of 2–5 nm are considerably dif-
mesopores due to the complex pore morphology. However, as in sample ferent from each other (Fig. 6). Though the qualitative observation by
(X)Y1-1, if the body-throat ratio is small and mercury withdrawal ef- FE-SEM and HIM (Figs. 9 and 10), OM pores and clay-associated pores
ficiency is high with good pore connectivity, the overestimation will be are generally developed in shale samples. In general, OM pores are oil-
significantly reduced. Meanwhile, as in sample (X)Y1-16, if the fraction wet, and clay-associated pores are water-wet (Gao and Hu, 2016). Some
of closed-pores is high in this pore-size range, it may conceal this nitrogen-inaccessible pores with diameters of 2–5 nm are penetrable by
phenomenon. water and this may due to the wettability. Water also appears to access
The fraction of nitrogen-inaccessible pores with different pore dia- more of the pores with pore size diameters less than 30 nm than me-
meters were obtained by comparing SANS pore volume with that de- thane in Barnett Shale in a previous study (Ruppert et al., 2013). An-
rived from the nitrogen adsorption experiments (Table 5). In order to other reason is that the strong capillary pressure of small pores makes it
study the relation between fractions of closed-pores and pore size, the more accessible for water by spontaneous imbibition (Gao and Hu,
fraction of nitrogen-inaccessible pores were divided into four pore- 2018; Sun et al., 2017a). The results of water accessibility in transi-
diameter intervals, as SANS and N2 adsorption both measure pore tional shales are higher than previous studies on Marcellus Shale and
bodies. For pores with diameters larger than 50 nm, the fractions of Barnett Shale (Bahadur et al., 2018; Gu et al., 2016; Ruppert et al.,
nitrogen-inaccessible pores may be overestimated. Based on the Kelvin 2013), which may be indicative of different pore connectivity. The
equation, the pressure P/P0 of nitrogen adsorption is higher than 0.981 Permian transitional shale samples in this work are very clay-rich (clay
when determining the pore diameter larger than 50 nm (Gregg and content 38–60%) with a high concentration of illite-smectite mixed-
Sing, 1982). When the shale pore size distribution is extensive, the layer mineral. In contrast, the clay mineral contents in marine shales
Ⅳ-type isotherm can become the Ⅱ-type isotherm, and it is difficult to are usually less than 30% which are mainly composed of illite. Some
measure the accurate saturation adsorption (Kuila and Prasad, 2013; internal pores contained in illite-smectite mixed-layer mineral (in par-
Tian et al., 2015). Therefore, both nitrogen adsorption and MICP un- ticular of smectite in the I/S group) are inaccessible to nitrogen mole-
derestimated the pore volume of macropores. The partial OM pores and cules but accessible to water due to different wettability (Sang et al.,
inorganic pores with pore-size diameters larger than 50 nm are not 2018).
reflected in FE-SEM images (Figs. 9b, 10b and e). Within the pore The water-inaccessible pores are mainly distributed in the pore size
diameter range of 10–50 nm, the proportion of closed pores is relatively range of 5–10 nm and 20–30 nm in all samples (Fig. 12). Interestingly,
low. This is consistent with the overlaped curves of the pore volume Table 5 shows that the sample (X)Y1-1 has the lowest fraction of ni-
distribution of the SANS and nitrogen adsorption analysis shown in trogen-inaccessible pores (0.1%) in pore-size range of 5–10 nm. Its
Fig. 11. The closed pores ratio increased within the pore diameter range fraction of mercury-inaccessible pores is also very low, 2.8%. One ex-
of 5–10 nm, except for sample (X)Y1-1. planation is that water film confined inside hydrophobic organic na-
According to Table 5, the percentage of nitrogen-inaccessible pores nopores prevent the pores from being filled by water in ambient con-
with diameters less than 5 nm is the highest, consistent with the results dition (Li et al., 2017; Zhang et al., 2019a). Bahadur et al. (2018)
of U.S. shale oil research and Longmaxi Shale (Sun et al., 2017b; Zhao suggested that the water-inaccessible pores in 5–50 nm pore size range
et al., 2017). The organic pores in U.S. shale oil samples with maturity are associated with OM and not with a specific mineral in Marcellus
ranging from 0.67% to 1.05% are inchoately developed by FE-SEM Shale.
observation (Grathoff et al., 2016). Zhao et al. (2017) indicated that Fig. 13a exhibits a positive relationship between residual gas con-
clay-associated pores with pore diameters less than 10 nm are in- tent and total closed pore volume. The latter was obtained from com-
accessible to nitrogen in oil shale samples. The internal space of mon- paring SANS with N2 adsorption results (V PDSM Pore volume (2–100 nm)-V
tmorillonite is inaccessible to nitrogen molecules in shale which cannot nitrogen pore volume (2–100 nm)) (Tables 3 and 4). Here the closed pores

419
M. Sun, et al. Marine and Petroleum Geology 107 (2019) 407–422

Fig. 13. Relationships between (a) Closed pore (inaccessible to N2) volume vs. residual gas content; (b) Closed pore (inaccessible to water) porosity vs. residual gas
content; (c) CO2 pore volume vs. lost and desorbed gas content; (d) Total organic carbon vs. CO2 pore volume (Black: Longtan Shale in this work; Green: Longmaxi
Shale from Sun et al., 2017a,b; Red: Niutitang Shale from Sun et al., 2016). (For interpretation of the references to colour in this figure legend, the reader is referred
to the Web version of this article.)

correspond to those which are not penetrated by nitrogen. This corre- (Fig. 13c), suggesting that the shale gas in transitional shale samples
lation indicates that the closed-pore system of transitional shale sam- mainly occurs in micropores, similar to the occurrence of coalbed me-
ples has certain gas-bearing properties. The porosity of water-in- thane (Cai et al., 2013; Mastalerz et al., 2018). Moreover, the pore
accessible pores shows no apparent positive correlation with residual volume determined by MICP and N2 adsorption does not exhibit a linear
gas content (Fig. 13b). This means that part of gas-bearing space can be relationship with lost and desorbed gas content. The average pore size
occupied by water. This process will affect the shale gas diffusion and of transitional shale samples determined by N2 adsorption (Table 4) is
lead to abundant retention of fracturing fluid (Phan et al., 2018; Singh, larger than that of marine shale samples (Mao and Guo, 2018; Yang
2016). A better understanding of the behavior of water in shale pores et al., 2017). Therefore, mesopores in transitional shale samples con-
will improve fracturing fluid management and ultimately enhance re- tribute less specific surface area than marine shale samples. The dis-
covery. crepancy of pore structure in shales is the fundamental factor that leads
In addition, the lost and desorbed gas content increase as CO2- to the difference of shale gas occurrence state.
measured pore volume increases in transitional shale samples Although micropores (pore size<2 nm) are difficult to be observed

420
M. Sun, et al. Marine and Petroleum Geology 107 (2019) 407–422

by FE-SEM and HIM, the positive correlation of micropores volume analysis of cement-based materials. J. Colloid Interface Sci. 211, 39–44.
with TOC (Fig. 13d) indicates that OM provides the main development Bahadur, J., Radlinski, A.P., Melnichenko, Y.B., Mastalerz, M., Schimmelmann, A., 2015.
Small-angle and ultrasmall-angle neutron scattering (SANS/USANS) study of New
space for micropores. For this, the connection between the organic pore Albany Shale: a treatise on microporosity. Energy Fuels 29, 567–576.
network system and its surroundings determines the expulsion of shale Bahadur, J., Ruppert, L.F., Pipich, V., Sakurovs, R., Melnichenko, Y.B., 2018. Porosity of
gas. The inter-particle pores formed by dissolution play an essential role the Marcellus Shale: a contrast matching small-angle neutron scattering study. Int. J.
Coal Geol. 188, 156–164.
in the connectivity of some samples (Fig. 9d). If the organic pore net- Bary, B., 2006. A polydispersed particle system representation of the porosity for non-
work system is disconnected with surrounding pore networks, over- saturated cementitious materials. Cement Concrete Res. 36, 2061–2073.
pressure will be formed when the shale gas content exceeds the max- Bernard, S., Horsfield, B., 2014. Thermal maturation of gas shale systems. Annu. Rev.
Earth Planet Sci. 42, 635–651.
imum methane adsorption capacity of the organic pores (Hao et al., Cai, Y., Liu, D., Pan, Z., Yao, Y., Li, J., Qiu, Y., 2013. Pore structure and its impact on CH4
2013). Therefore, the pressure coefficient of shale formation will also adsorption capacity and flow capability of bituminous and subbituminous coals from
reflect characteristics of pore connectivity. Northeast China. Fuel 103, 258–268.
Chen, Y., Qin, Y., Wei, C., Huang, L., Shi, Q., Wu, C., Zhang, X., 2018. Porosity changes in
Moreover, unlike sandstone reservoirs, shale sample usually does
progressively pulverized anthracite subsamples: implications for the study of closed
not have a single geometric pore body-to-throat aspect ratio (Fig. 8). To pore distribution in coals. Fuel 225, 612–622.
compare the data in Table 4, the lower pore body-to-throat aspect ratio Clarkson, C.R., Solano, N., Bustin, R.M., Bustin, A.M.M., Chalmers, G.R.L., He, L.,
of MICP is usually associated with a higher mercury withdrawal effi- Melnichenko, Y.B., Radliński, A.P., Blach, T.P., 2013. Pore structure characterization
of North American shale gas reservoirs using USANS/SANS, gas adsorption, and
ciency. For sample (X)Y1-1 with highest mercury withdrawal efficiency mercury intrusion. Fuel 103, 606–616.
and lowest pore body/throat ratio, MICP curve overlaps best with SANS Cui, G., Liu, J., Wei, M., Shi, R., Elsworth, D., 2018. Why shale permeability changes
data (Fig. 11). MICP curves have been used in some literature to reflect under variable effective stresses: new insights. Fuel 213, 55–71.
Davudov, D., Moghanloo, R.G., 2018. Impact of pore compressibility and connectivity loss
pore connectivity in shale (Davudov et al., 2018; Zuo et al., 2018). on shale permeability. Int. J. Coal Geol. 187, 98–113.
However, the compressibility or destructiveness in the high-pressure Davudov, D., Moghanloo, R.G., Lan, Y., 2018. Evaluation of accessible porosity using
segment to the shale samples needs to be clarified in further work. mercury injection capillary pressure data in shale samples. Energy Fuels 32,
4682–4694.
Dewhurst, D.N., Aplin, A.C., Sarda, J.-P., Yang, Y., 1998. Compaction-driven evolution of
5. Conclusions porosity and permeability in natural mudstones: an experimental study. J. Geophys.
Res. Solid Earth 103, 651–661.
Gao, Z., Hu, Q., 2016. Wettability of Mississippian Barnett Shale samples at different
Six Upper Permian Longtan Formation transitional shale samples depths: investigations from directional spontaneous imbibition. AAPG Bull. 100,
were examined using neutron scattering tests, MICP, gas adsorption 101–114.
experiments, and FE-SEM/HIM techniques to characterize the pore Gao, Z., Hu, Q., 2018. Pore structure and spontaneous imbibition characteristics of
marine and continental shales in China. AAPG Bull. 102, 1941–1961.
connectivity and water accessibility within the shale matrix. Based on
Grathoff, G.H., Peltz, M., Enzmann, F., Kaufhold, S., 2016. Porosity and permeability
results and discussion, the following conclusions can be drawn: determination of organic-rich Posidonia shales based on 3-D analyses by FIB-SEM
microscopy. Solid Earth 7, 1145–1156.
(1) Closed pores, inaccessible to nitrogen and mercury, are mainly Gregg, S.J., Sing, K.S.W., 1982. Adsorption, Surface Area, and Porosity. Academic Press,
New York.
distributed in the pore-size range of less than 10 nm and associated Gu, X., Cole, D.R., Rother, G., Mildner, D.F.R., Brantley, S.L., 2015. Pores in Marcellus
with organic pores and interlayer spaces of illite-smectite mixed- Shale: a neutron scattering and FIB-SEM study. Energy Fuels 29, 1295–1308.
layer mineral. Gu, X., Mildner, D.F.R., Cole, D.R., Rother, G., Slingerland, R., Brantley, S.L., 2016.
Quantification of organic porosity and water accessibility in Marcellus Shale using
(2) CM-SANS test using water shows that 87–98% of the pores neutron scattering. Energy Fuels 30, 4438–4449.
(2–200 nm diameters) are water-connected in transitional samples, Guo, T., 2015. The Fuling Shale Gas Field — a highly productive Silurian gas shale with
which are higher than that of the marine shales from the U.S.A. The high thermal maturity and complex evolution history, southeastern Sichuan Basin,
China. Interpretation 3, SJ25–SJ34.
inaccessible pores to water are mainly distributed in the pore size Gupta, N., Fathi, E., Belyadi, F., 2018. Effects of nano-pore wall confinements on rarefied
range of 5–10 nm and 20–30 nm in all samples, which may be due gas dynamics in organic rich shale reservoirs. Fuel 220, 120–129.
to the influence of water film confined inside hydrophobic organic Hao, F., Zou, H., Lu, Y., 2013. Mechanisms of shale gas storage: implications for shale gas
exploration in China. AAPG Bull. 97, 1325–1346.
nanopores.
Hildenbrand, A., Urai, J.L., 2003. Investigation of the morphology of pore space in
(3) In-situ gas behaviors indicate that closed pores have a certain gas mudstones—first results. Mar. Pet. Geol. 20, 1185–1200.
bearing capacity. Meanwhile, the occurrence of shale gas in tran- Hu, Q., Ewing, R.P., Rowe, H.D., 2015a. Low nanopore connectivity limits gas production
in Barnett formation. J. Geophys. Res. Solid Earth 120, 8073–8087.
sitional samples is more related to micropores.
Hu, Q.H., Liu, X.G., Gao, Z.Y., Liu, S.G., Zhou, W., Hu, W.X., 2015b. Pore structure and
tracer migration behavior of typical American and Chinese shales. Petrol. Sci. 12,
Acknowledgements 651–663.
Ilavsky, J., Jemian, P.R., 2009. Irena: tool suite for modeling and analysis of small-angle
scattering. J. Appl. Crystallogr. 42, 347–353.
The authors sincerely thank the National Natural Science Jiang, S., Tang, X., Cai, D., Xue, G., He, Z., Long, S., Peng, Y., Gao, B., Xu, Z., Dahdah, N.,
Foundation of China (Grant Nos. 41802146; 41830431; 41572134), 2017. Comparison of marine, transitional, and lacustrine shales: a case study from the
Fundamental Research Funds for the Central Universities (No. Sichuan Basin in China. J. Pet. Sci. Eng., 150, 2017, 334-347. 150, 334–347.
Ju, Y., Sun, Y., Tan, J., Bu, H., Han, K., Li, X., Fang, L., 2018. The composition, pore
CUG180608), Key Laboratory of Tectonics and Petroleum Resources structure characterization and deformation mechanism of coal-bearing shales from
(TPR-2017-09), and Key Laboratory of Unconventional Oil & Gas tectonically altered coalfields in eastern China. Fuel 234, 626–642.
Geology of China Geological Survey (DD20160181-YQ17W06JJ01), for Katz, A.J., Thompson, A.H., 1986. Quantitative prediction of permeability in porous rock.
Phys. Rev. B 34, 8179–8181.
their financial support. We acknowledge support by Oak Ridge National King, H.E., Eberle, A.P.R., Walters, C.C., Kliewer, C.E., Ertas, D., Huynh, C., 2015. Pore
Laboratory for instrument time on the High Flux Isotope Reactor which architecture and connectivity in gas shale. Energy Fuels 29, 1375–1390.
was sponsored by the Laboratory Directed Research and Development Kuila, U., Prasad, M., 2013. Application of nitrogen gas-adsorption technique for char-
acterization of pore structure of mudrocks. Lead. Edge 32, 1478–1485.
Program and the Scientific User Facilities Division, Office of Basic
Li, J., Li, X., Wu, K., Feng, D., Zhang, T., Zhang, Y., 2017. Thickness and stability of water
Energy Sciences, U.S. Department of Energy. We thank Dr. Lilin He of film confined inside nanoslits and nanocapillaries of shale and clay. Int. J. Coal Geol.
ORNL for his discretionary neutron beam time and his help with SANS 179, 253–268.
Li, Z., Liu, D., Cai, Y., Yao, Y., Wang, H., 2015. Pore structure and compressibility of coal
experiments at GP-SANS and Core Lab of Institute of Geology and
matrix with elevated temperatures by mercury intrusion porosimetry. Energy Explor.
Geophysics at Chinese Academy of Sciences for the help with FE-SEM Exploit. 33, 809–826.
and HIM analyses. Liu, D., Chen, J., Song, L., Lu, A., Wang, Y., Sun, G., 2017. Parameterization of silica-filled
silicone rubber morphology: a contrast variation SANS and TEM study. Polymer 120,
155–163.
References Loucks, R.G., Reed, R.M., Ruppel, S.C., Hammes, U., 2012. Spectrum of pore types and
networks in mudrocks and a descriptive classification for matrix-related mudrock
Abell, A.B., Willis, K.L., Lange, D.A., 1999. Mercury intrusion porosimetry and image pores. AAPG Bull. 96, 1071–1098.

421
M. Sun, et al. Marine and Petroleum Geology 107 (2019) 407–422

Ma, Y., Zhong, N., Li, D., Pan, Z., Cheng, L., Liu, K., 2015. Organic matter/clay mineral Sun, M., Yu, B., Hu, Q., Chen, S., Xia, W., Ye, R., 2016. Nanoscale pore characteristics of
intergranular pores in the Lower Cambrian Lujiaping Shale in the north-eastern part the lower cambrian Niutitang Formation shale: a case study from well yuke #1 in the
of the upper Yangtze area, China: a possible microscopic mechanism for gas pre- southeast of chongqing, China. Int. J. Coal Geol. 154–155, 16–29.
servation. Int. J. Coal Geol. 137, 38–54. Sun, M., Yu, B., Hu, Q., Yang, R., Zhang, Y., Li, B., 2017a. Pore connectivity and tracer
Mani, D., Patil, D.J., Dayal, A.M., Prasad, B.N., 2015. Thermal maturity, source rock migration of typical shales in south China. Fuel 203, 32–46.
potential and kinetics of hydrocarbon generation in Permian shales from the Sun, M., Yu, B., Hu, Q., Yang, R., Zhang, Y., Li, B., Melnichenko, Y.B., Cheng, G., 2018.
Damodar Valley basin, Eastern India. Mar. Pet. Geol. 66, 1056–1072. Pore structure characterization of organic-rich Niutitang shale from China: small
Mao, W., Guo, S., 2018. Comparison of factors influencing pore size distributions in angle neutron scattering (SANS) study. Int. J. Coal Geol. 186, 115–125.
marine, terrestrial, and transitional shales of similar maturity in China. Energy Fuels Sun, M., Yu, B., Hu, Q., Zhang, Y., Li, B., Yang, R., Melnichenko, Y.B., Cheng, G., 2017b.
32, 8145–8153. Pore characteristics of Longmaxi shale gas reservoir in the Northwest of Guizhou,
Mastalerz, M., He, L., Melnichenko, Y.B., Rupp, J.A., 2012. Porosity of coal and shale: China: investigations using small-angle neutron scattering (SANS), helium pycno-
insights from gas adsorption and SANS/USANS techniques. Energy Fuels 26, metry, and gas sorption isotherm. Int. J. Coal Geol. 171, 61–68.
5109–5120. Tang, X., Jiang, Z., Jiang, S., Cheng, L., Zhang, Y., 2017. Characteristics and origin of in-
Mastalerz, M., Wei, L., Drobniak, A., Schimmelmann, A., Schieber, J., 2018. Responses of situ gas desorption of the Cambrian Shuijingtuo Formation shale gas reservoir in the
specific surface area and micro- and mesopore characteristics of shale and coal to Sichuan Basin, China. Fuel 187, 285–295.
heating at elevated hydrostatic and lithostatic pressures. Int. J. Coal Geol. 197, Tang, X., Jiang, Z., Jiang, S., Li, Z., 2016. Heterogeneous nanoporosity of the Silurian
20–30. Longmaxi Formation shale gas reservoir in the Sichuan Basin using the QEMSCAN,
Melnichenko, Y.B., 2015. Small-angle Scattering from Confined and Interfacial Fluids: FIB-SEM, and nano-CT methods. Mar. Pet. Geol. 78, 99–109.
Applications to Energy Storage and Environmental Science. Springer International Tian, H., Pan, L., Zhang, T., Xiao, X., Meng, Z., Huang, B., 2015. Pore characterization of
Publishing. organic-rich lower cambrian shales in Qiannan depression of Guizhou Province,
Melnichenko, Y.B., He, L., Sakurovs, R., Kholodenko, A.L., Blach, T., Mastalerz, M., southwestern China. Mar. Pet. Geol. 62, 28–43.
Radliński, A.P., Cheng, G., Mildner, D.F.R., 2012. Accessibility of pores in coal to Wang, S., Javadpour, F., Feng, Q., 2016. Confinement correction to mercury intrusion
methane and carbon dioxide. Fuel 91, 200–208. capillary pressure of shale nanopores. Sci. Rep. 6, 20160.
Milliken, K.L., Rudnicki, M., Awwiller, D.N., Zhang, T., 2013. Organic matter-hosted pore Westwood, R.F., Toon, S.M., Cassidy, N.J., 2017. A sensitivity analysis of the effect of
system, Marcellus formation (devonian), Pennsylvania. AAPG Bull. 97, 177–200. pumping parameters on hydraulic fracture networks and local stresses during shale
Niu, Q., Pan, J., Cao, L., Ji, Z., Wang, H., Wang, K., Wang, Z., 2017. The evolution and gas operations. Fuel 203, 843–852.
formation mechanisms of closed pores in coal. Fuel 200, 555–563. Yang, C., Zhang, J., Tang, X., Ding, J., Zhao, Q., Dang, W., Chen, H., Su, Y., Li, B., Lu, D.,
Pan, L., Xiao, X., Tian, H., Zhou, Q., Chen, J., Li, T., Wei, Q., 2015. A preliminary study on 2017. Comparative study on micro-pore structure of marine, terrestrial, and transi-
the characterization and controlling factors of porosity and pore structure of the tional shales in key areas, China. Int. J. Coal Geol. 171, 76–92.
Permian shales in Lower Yangtze region, Eastern China. Int. J. Coal Geol. 146, 68–78. Zhang, J., Li, X., Wei, Q., Sun, K., Zhang, G., Wang, F., 2017. Characterization of full-sized
Phan, T.T., Paukert Vankeuren, A.N., Hakala, J.A., 2018. Role of water−rock interaction pore structure and fractal characteristics of marine–continental transitional Longtan
in the geochemical evolution of Marcellus Shale produced waters. Int. J. Coal Geol. formation shale of Sichuan Basin, south China. Energy Fuels 31, 10490–10504.
191, 95–111. Zhang, L., Li, B., Jiang, S., Xiao, D., Lu, S., Zhang, Y., Gong, C., Chen, L., 2018.
Porod, G., Glatter, O., Kratky, O., 1982. General Theory in: Small Angle X-Ray Scattering. Heterogeneity characterization of the lower Silurian Longmaxi marine shale in the
Academic, London. Pengshui area, South China. Int. J. Coal Geol. 195, 250–266.
Radlinski, A.P., 1999. Small-angle neutron scattering: a new technique to detect gener- Zhang, M., Fu, X., 2018. Study of the characteristics of marine-terrigenous facies shale
ated source rocks. AGSO Res. Newsl. 31. from the Permo-carboniferous system in the Guxian block, southwest Qinshui basin.
Radlinski, A.P., Boreham, C.J., Linder, P., Randl, O., Wignall, G.D., Hinde, A., Hope, J.M., Energy Fuels 32, 1096–1109.
2000. Small angle neutron scattering signature of oil generation in artificially and Zhang, P., Hu, L., Meegoda, J.N., Gao, S., 2015a. Micro/Nano-pore network analysis of
naturally matured hydrocarbon source rocks. Org. Geochem. 31, 1–14. gas flow in shale matrix. Sci. Rep. 5, 13501.
Radlinski, A.P., Mastalerz, M., 2018. Neutron scattering study of vitrinite: insights into Zhang, Q., Liang, F., Pang, Z., Jiang, S., Zhou, S., Zhang, J., 2019a. Lower threshold of
sub-micrometer inclusions in North American Carboniferous coals of biuminous rank. pore-throat diameter for the shale gas reservoir: experimental and molecular simu-
Int. J. Coal Geol. 186, 145–154. lation study. J. Pet. Sci. Eng. 173, 1037–1046.
Radlinski, A.P., Mastalerz, M., Hinde, A.L., Hainbuchner, M., Rauch, H., Baron, M., Lin, Zhang, Q., Liu, R., Pang, Z., Lin, W., Bai, W., Wang, H., 2016. Characterization of mi-
J.S., Fan, L., Thiyagarajan, P., 2004. Application of SAXS and SANS in evaluation of croscopic pore structures in Lower Silurian black shale(S1l), southeastern Chongqing,
porosity, pore size distribution and surface area of coal. Int. J. Coal Geol. 59, China. Mar. Pet. Geol. 71, 250–259.
245–271. Zhang, R., Liu, S., Bahadur, J., Elsworth, D., Melnichenko, Y., He, L., Wang, Y., 2015b.
Ruppert, L.F., Sakurovs, R., Blach, T.P., He, L., Melnichenko, Y.B., Mildner, D.F.R., Estimation and modeling of coal pore accessibility using small angle neutron scat-
Alcantar-Lopez, L., 2013. A USANS/SANS study of the accessibility of pores in the tering. Fuel 161, 323–332.
Barnett Shale to methane and water. Energy Fuels 27, 772–779. Zhang, Y., Hu, Q., Long, S., Zhao, J., Peng, N., Wang, H., Lin, X., Sun, M., 2019b. Mineral-
Sakurovs, R., Koval, L., Grigore, M., Sokolova, A., Campo, L.d., Rehm, C., 2018a. controlled nm-μm-scale pore structure of saline lacustrine shale in Qianjiang
Nanostructure of cokes. Int. J. Coal Geol. 188, 112–120. Depression, Jianghan Basin, China. Mar. Pet. Geol. 99, 347–354.
Sakurovs, R., Koval, L., Grigore, M., Sokolova, A., Ruppert, F., L., Melnichenko, Y.B., Zhao, J., Jin, Z., Hu, Q., Jin, Z., Barber, T.J., Zhang, Y., Bleuel, M., 2017. Integrating SANS
2018b. Nanometre-sized pores in coal: variations between coal basins and coal origin. and fluid-invasion methods to characterize pore structure of typical American shale
Int. J. Coal Geol. 186, 126–134. oil reservoirs. Sci. Rep. 7, 15413.
Sang, G., Liu, S., Zhang, R., Elsworth, D., He, L., 2018. Nanopore characterization of mine Zhao, Y., Sun, Y., Liu, S., Chen, Z., Yuan, L., 2018. Pore structure characterization of coal
roof shales by SANS, nitrogen adsorption, and mercury intrusion: impact on water by synchrotron radiation nano-CT. Fuel 215, 102–110.
adsorption/retention behavior. Int. J. Coal Geol. 200, 173–185. Zhou, S., Yan, G., Xue, H., Guo, W., Li, X., 2016. 2D and 3D nanopore characterization of
Singh, H., 2016. A critical review of water uptake by shales. J. Nat. Gas Sci. Eng. 34, gas shale in Longmaxi formation based on FIB-SEM. Mar. Pet. Geol. 73, 174–180.
751–766. Zuo, J.Y., Guo, X., Liu, Y., Pan, S., Canas, J., Mullins, O.C., 2018. Impact of capillary
Starnoni, M., Pokrajac, D., Neilson, J.E., 2017. Computation of fluid flow and pore-space pressure and nanopore confinement on phase behaviors of shale gas and oil. Energy
properties estimation on micro-CT images of rock samples. Comput. Geosci. 106, Fuels 32, 4705–4714.
118–129.

422

You might also like