Shale pore size classification: An NMR fluid typing method

Highlights

• Novel method of full-scale pore size distribution classification for shale.

• Determine T_2C1 and T_2C2 by NMR experiments on cyclic centrifugal and heat shales.

• T_2C2 ranges from 0.09 ms to 0.36 ms, and T_2C1 ranges from 0.45 ms to 2.98 ms

• Surface relaxivities of shales range from 0.00426 μm/ms to 0.02822 μm/ms.

• Unrecoverable-, capillary bound- and movable-pores in shales were identified.

Abstract

Shale reservoir is characterized by complex pore networks, within which there are various pore fluids including unrecoverable fluid, capillary bound fluid and movable fluid. Although considerable literature has investigated the pore structure of gas shale using different laboratory testing techniques, few papers have provided a quantified model to distinguish different types of fluids (corresponding to different pore types) in shale. In this study, seven shale core plugs from the Sichuan Basin were measured in a series of NMR experiments under full brine-saturated, centrifugal and heat-treated conditions to analyze the pore structure information and pore fluid transport during the processes of centrifuging and heating. For a typical T₂ spectrum of 100% brine-saturated shale, the movable fluid T₂ cutoff (T_2C1) and unrecoverable fluid cutoff (T_2C2) were derived from NMR centrifugal and heat-treated experiments to distinguish the unrecoverable fluid (T₂ < T_2C2), capillary bound fluid (T_2C2 <T₂< T_2C1) and movable fluid (T₂ > T_2C1). Our results show that for the investigated shales, the T_2C2 ranges from 0.09 ms to 0.36 ms, and T_2C1 has a wide range from 0.45 ms to 2.98 ms. The surface relaxivities range from 0.00426 μm/ms to 0.02822 μm/ms, and the shales having high silicate mineral contents commonly have low surface relaxivities. A conceptional model based on the dual T₂ cutoff method was constructed to illustrate the full-scale pore size distribution: unrecoverable fluid pores, capillary bound fluid pores and movable fluid pores. This study provides a new method of pore fluid typing and full-scale pore size distribution classification for shales.

Introduction

The production of shale gas has been highly successful in North America, followed by China, where production reached 7.9 billion m³ in 2016 in the marine shale of the Sichuan basin (Liu et al., 2017). The success in shale gas industry raises great interests in fundamental investigations on shale petrophysical properties, such as pore size distribution (PSD), effective porosity and clay-bound water, which can be used to estimate hydrocarbon storage capacity (Furmann et al., 2016; Testamanti and Rezaee, 2017). For example, clay-bound water, residing in nanometer-scaled pore structure, is the most key petrophysical parameter and commonly used to calibrate log porosity for shale formation evaluation (Coates et al., 1999).

Unlike conventional natural gas reservoirs, gas shale is particularly complex in its mineral composition (Curtis et al., 2012), and it is characterized by low porosity, extremely low permeability, and a complicated pore structure (Curtis, 2002; Sondergeld et al., 2010; Hinai et al., 2014). In past decades, various analysis methods, including image analysis, intrusive fluid detection methods and nonintrusive fluid detection methods, have been used to investigate the pore structure features of shale (Ambrose et al., 2010; Clarkson et al., 2013; Kuila and Prasad, 2013; Hinai et al., 2014; Testamanti and Rezaee, 2017). Among these methods, each method has advantages and limitations (Hinai et al., 2014; Clarkson et al., 2013). Typical image analysis methods include scanning transmission electron microscopy (TEM), atomic force microscopy (AFM) and focused ion beam scanning electron microscopy (FIB-SEM). Based on these methods, Reed et al. (2007) and Bustin et al. (2008) defined the nanopores inside the organic matter of gas shale, and Loucks et al. (2012) classified shale pores as organic pores, intra-pores and inter-pores. Intrusive fluid detection methods, including mercury injection capillary pressure (MICP), low pressure gas (N₂ and CO₂) adsorption, and nuclear magnetic resonance (NMR) methods, can be used to characterize pore characteristics, such as the PSD, pore throat and pore body sizes, and specific surface area (Fleury, 2007; Ambrose et al., 2010; Yao et al., 2010a, 2010b; Clarkson et al., 2013; Kuila and Prasad, 2013; Hinai et al., 2014; Zolfaghari and Dehghanpour, 2015; Chen et al., 2015). Nonintrusive fluid detection methods, such as X-ray computerized tomography (X-CT), can be used to simulate the 3D distribution of the PSD, fracture morphology, and fluid transport (Desbois et al., 2011; Yao and Liu, 2012; Blunt et al., 2013; Korost et al., 2013; Vega et al., 2014). However, the laboratory methods mentioned above have certain limitations. For example, MICP can only detect certain parts of pores but not all nanopores because the original pore structure may be destroyed by applying an extremely high intrusion pressure (Yao and Liu, 2012; Clarkson et al., 2013). The X-CT scanning technique is also not applicable for the analysis of nanopores, and low-pressure gas adsorption technology is not effective for pores of >300 nm (Clarkson et al., 2013). The limitations of these methods make it difficult to characterize the full-scale PSD of shale using an individual method.

As a nondestructive and relatively new method, the NMR core analysis method has been used for the petrophysical characterization of sandstones, coal and shales (Timur, 1969; Westphal et al., 2005; Yao et al., 2010b; Josh et al., 2012; Chi and Heidari, 2014; Gips et al., 2014; Gannaway, 2014; Saidian et al., 2014; Saidian and Prasad, 2015; Wang et al., 2015; Zhao et al., 2015; Minh et al., 2016; Sondergeld et al., 2016; Zolfaghari et al., 2017; Sun et al., 2018). For example, using NMR, X-ray CT scanning and SEM observation, researchers found that shale pores of 3–100 nm commonly have a typical transverse relaxation time (T₂) distribution of 0.01–10 ms (Hinai et al., 2014; Huang and Zhao, 2017), and T₂ spectra ranging from 10 to 300 ms represent micro fractures (Tinni et al., 2014). Based on different scenarios of NMR simulation experiments, Gannaway (2014) quantified the effective porosity, organic porosity and inorganic porosity in gas shale. Generally, studies have given various classifications for pore typing.

Shale is complex in its petrological composition and pore structure, and thus, fluid types within shale vary. In the normal definition used by a log analyst, pore fluid types were divided into movable fluid and irreducible fluid, and the latter was further divided into capillary bound fluid and clay bound fluid (Straley et al., 1997; Sondergeld et al., 2010; Rylander et al., 2013; Jiang et al., 2013; Mehamed and El-monier, 2016). In a generally accepted definition of sandstones or carbonate rocks, the clay bound porosity is not part of the effective porosity and is the difference between the total porosity and effective porosity. However, recent SEM studies of shales have indicated that many nanopores (gas unrecoverable) are paragenetic with organic matter but not clay minerals (Reed et al., 2007; Bustin et al., 2008; Sondergeld et al., 2010; Loucks et al., 2012; Jing et al., 2017). Considering that we cannot determine whether the so-called “clay bound fluid” occurs in the clay minerals of shales or not (Prammer et al., 1996), in this study, we defined the difference between the total porosity and effective porosity as the unrecoverable porosity. The fluid type corresponding to the unrecoverable porosity is called unrecoverable fluid. Based on a typical T₂ distribution obtained from NMR measurements, we defined a T₂ cutoff of T_2C1 to distinguish the movable fluid (T₂ > T_2C1) from the irreducible fluid (T₂ < T_2C1), and we defined another T₂ cutoff of T_2C2 to differentiate the capillary bound fluid (T_2C1>T₂ > T_2C2) and unrecoverable (conventionally called “clay bound”) fluid (T₂ < T_2C2).

The value of T_2C1 for a certain rock is commonly determined by a series of NMR centrifugal tests at different water saturation levels. Straley et al. (1997) concluded typical T_2C1 values of 33 ms and 100 ms for sandstones and carbonates, respectively. Based on NMR measurements of 12 brine-saturated shale samples, Sun et al. (2012) suggested a T_2C1 of 8.29 ms for shale. Apart from the determination of T_2C1, the determination of T_2C2 is also important for evaluating production performance in the late gas production stage. However, the determination of the value of T_2C2 for a certain rock is relatively difficult because of methodological limitations. Recently, there have been some reports of T_2C2 for sandstone (e.g., Straley et al. (1997)) but no in-depth research of the T_2C2 of shale. Based on NMR core measurements, Straley et al. (1997) provided a definition of T_2C2 = 3 ms for sandstones. For a long time, 3 ms was the most acceptable T_2C2 for both sandstones and shales (Sondergeld et al., 2010; Rylander et al., 2013). Recently, some researchers have reported different values of T_2C2 for different shales, such as T_2C2 = 0.2 ms for Swiss Molasse shales (Esteban et al., 2013) and T_2C2 = 0.25 ms for Carynginia shales (Testamanti and Rezaee, 2017). In fact, the pore structure is very different for shales deposited in different environments; thus, it is problematic to evaluate the shale PSD using a certain value of T_2C2 for different shales. It is urgent to develop a method that can be used to accurately calculate the values of T_2C1 and T_2C2 for different shales.

In this study, we collected seven shale samples from the Lower Silurian Longmaxi formation in the Sichuan basin and conducted a series of NMR experiments on full brine-saturated, cyclic centrifugal treatment, and cyclic heat treatment samples. These experiments utilized a quantitative method for the determination of T_2C1 and T_2C2 for different shales. This study also provides a pore type classification method for assessing the full-scale PSD of gas shale.