Alexandra Amann-Hildenbrand

On the use and abuse of N2 physisorption for the characterization of the pore structure of shales

PIETER BERTIER , KEVIN SCHWEINAR, HELGE STANJEK, AMIN GHANIZADEH, CHRISTOPHER R. CLARKSON, ANDREAS BUSCH, NIKO KAMPMAN, DIRK PRINZ, ALEXANDRA AMANN-HILDENBRAND, BERNHARD M. KROOSS, and VITALIY PIPICH Clay & Interface Mineralogy, RWTH-Aachen University, Bunsenstr. 8, D-52072 Aachen, Germany Department of Geoscience, University of Calgary, Calgary, Canada Shell Global Solutions International, Kessler Park 1, 2288 GS Rijswijk, The Netherlands Dynchem, Saarstrasse 98, D-52062 Aachen, Germany Institute for Petroleum & Coal, RWTH-Aachen University, Lochnerstr. 2, D-52062 Aachen, Germany Jülich Centre for Neutron Science JCNS, Forschungszentrum Jülich GmbH, Outstation at MLZ, Lichtenbergstrasse 1 85747 Garching, Germany e-mail: Pieter.Bertier@emr.rwth-aachen.de

Determining the porosity of mudrocks using methodological pluralism

Abstract Porosity of shales is an important parameter that impacts rock strength for seal or wellbore integrity, gas-in-place calculations for unconventional resources or the diffusional solute and gas transport in these microporous materials. From a well section obtained from the Mont Terri Underground Rock Laboratory in St Ursanne, Switzerland, we determined porosity, pore size distribution and specific surface areas on a set of 13 Opalinus Clay samples. The porosity methods employed are helium-pycnometry, water and mercury injection porosimetry, liquid saturation and immersion, and low pressure N2 sorption, as well as small-angle to ultra-small-angle neutron scattering (SANS–USANS). These were used in addition to mineralogical and geochemical methods for sample analysis that comprise X-ray diffraction, X-ray fluorescence, total organic carbon content and cation exchange capacity. We find large variations in total porosity, ranging from approximately 23% for the neutron-scattering method to approximately 10% for mercury injection porosimetry. These differences can partly be related to differences in pore accessibility, while no or negligible inaccessible porosity was found. Pore volume distributions between neutron scattering and low-pressure sorption compare very well but differ significantly from those obtained from mercury porosimetry: this is realistic since the latter provides information on pore throats only, and the two former methods on pore throats and pore bodies. Finally, we find that specific surface areas determined using low-pressure sorption and neutron scattering match well.

Stress-dependence of porosity and permeability of the Upper Jurassic Bossier shale: an experimental study

Abstract In order to characterize the stress-dependence of porosity and permeability of Bossier shale, a series of measurements was conducted on three dry, horizontally orientated samples using various gases under controlled stress conditions. The Klinkenberg-corrected permeability and gas slippage factors varied by more than two orders of magnitude (0.21–86 µD) and by one order of magnitude (0.09–0.89 MPa), respectively. Porosity values measured under in situ stress conditions were lower by up to 30% than those measured at ambient conditions. Therefore, disregarding the stress-dependence of porosity may lead to a substantial overestimation of the free gas storage capacity. The stress sensitivity of Klinkenberg-corrected permeability coefficients (−0.012–−0.063MPa−1) is much larger than the stress sensitivity of porosity (−0.0014–−0.0033 MPa−1). Particularly for pore systems dominated by microfractures or slit-shaped pores, the permeability is highly sensitive to effective stress changes. While conventional pore models use porosity stress-sensitivity exponents (m) ranging between 3 and 5, we measured values of up to 27. Strongly stress-sensitive permeability behaviour is a result of effective stress preferentially reducing the volume and effective cross-section of transport pathways. In contrast, stress-dependent permeability of a less stress-sensitive sample is instead controlled by the redistribution of flow.

Gas Transfer Through Clay Barriers

Gas transport through clay-rocks can occur by different processes that can be basically subdivided into pressure-driven flow of a bulk gas phase and transport of dissolved gas either by molecular diffusion or advective water flow (Figure 1, Marschall et al., 2005). The relative importance of these transport mechanisms depends on the boundary conditions and the scale of the system. Pressure-driven volume flow (“Darcy flow”) of gas is the most efficient transport mechanism. It requires, however, pressure gradients that are sufficiently large to overcome capillary forces in the typically water-saturated rocks (purely gas-saturated argillaceous rocks are not considered in the present context). These pressure gradients may form as a consequence of the gravity field (buoyancy, compaction) or by gas generation processes (thermogenic, microbial, radiolytic). Dissolved gas may be transported by water flow along a hydraulic gradient. This process is not affected by capillary forces but constrained by the solubility of the gas. It has much lower transport efficiency than bulk gas phase flow. Molecular diffusion of dissolved gas, finally, is occurring essentially without constraints, ubiquitously and perpetually. Effective diffusion distances are, however, proportional to the square root of time, which limits the relevance of this transport process to the range of tens to hundreds of metres on a geological time scale (millions of years). 2 Process understanding and the quantification of the controlling parameters, like diffusion coefficients, capillary gas breakthrough pressures and effective gas permeability coefficients, is of great importance for up-scaling purposes in different research disciplines and applications. During the past decades, gas migration through fully water-saturated geological clay-rich barriers has been investigated extensively (Thomas et al., 1968, Pusch and Forsberg, 1983; Horseman et al., 1999; Galle, 2000; Hildenbrand et al., 2002; Marschall et al., 2005; Davy et al., 2009; Harrington et al., 2009, 2012a, 2014). All of these studies aimed at the analysis of experimental data determined for different materials (rocks of different lithotype, composition, compaction state) and pressure/temperature conditions. The clay-rocks investigated in these studies, ranged from unconsolidated to indurated clays and shales, all characterised by small pores (2-100 nm) and very low hydraulic conductivity (K < 10-12 m·s-1) or permeability coefficients (k < 10-19 m²). Studies concerning radioactive waste disposal include investigations of both the natural host rock formation and synthetic/engineered backfill material at a depth of a few hundred meters (IAEA, 2003, 2009). Within a geological disposal facility, hydrogen is generated by anaerobic corrosion of metals and through radiolysis of water (Rodwell et al., 1999; Yu and Weetjens, 2009). Additionally, methane and carbon dioxide are generated by microbial degradation of organic wastes (Rodwell et al., 1999; Ortiz et al., 2002; Johnson, 2006; Yu and Weetjens, 2009). The focus of carbon capture and storage (CCS) studies is on the analysis of the long-term sealing efficiency of lithologies above depleted reservoirs or saline aquifers, typically at larger depths (hundreds to thousands of meters). During the last decade, several studies were published on the sealing integrity of clay-rocks to carbon dioxide (Hildenbrand et al., 2004; Li et al., 2005; Hangx et al., 2009; Harrington et al., 2009; Skurtveit et al., 2012; Amann-Hildenbrand et al., 2013). In the context of petroleum system analysis, a significant volume of research has been undertaken regarding gas/oil expulsion mechanisms from sources rocks during burial history (Tissot & Pellet, 1971; Appold & Nunn, 2002), secondary migration (Luo et al., 2008) and the capillary sealing capacity of caprocks overlying natural gas accumulations (Berg, 1975; Schowalter, 1979; Krooss, 1992; Schlomer and Kross, 2004; Li et al., 2005; Berne et al., 2010). Recently, more attention has been paid to investigations of the transport efficiency of shales in the context of oil/gas shale production (Bustin et al., 2008; Eseme et al., 2012; Amann-Hildenbrand et al., 2012; Ghanizadeh et al., 2013, 2014). Analysis of the migration mechanisms within partly unlithified strata becomes important when explaining the 3 origin of overpressure zones, sub-seafloor gas domes and gas seepages (Hovland & Judd, 1988; Boudreau, 2012). The conduction of experiments and data evaluation/interpretation requires a profound process understanding and a high level of experience. The acquisition and preparation of adequate samples for laboratory experiments usually constitutes a major challenge and may have serious impact on the representativeness of the experimental results. Information on the success/failure rate of the sample preparation procedure should therefore be provided. Sample specimens “surviving” this procedure are subjected to various experimental protocols to derive information on their gas transport properties. The present overview first presents the theoretical background of gas diffusion and advective flow, each followed by a literature review (sections 2 and 3). Different experimental methods are described in sections 4.1 and 4.2. Details are provided on selected experiments performed at the Belgian Nuclear Research Centre (SCK-CEN, Belgium), Ecole Centrale de Lille (France), British Geological Survey (UK), and at RWTH-Aachen University (Germany) (section 4.3). Experimental data are discussed with respect to different petrophysical parameters outlined above: i) gas diffusion, ii) evolution of gas breakthrough, iii) dilation-controlled flow, and iv) effective gas permeability after breakthrough. These experiments were conducted under different pressure and temperature conditions, depending on sample type, burial depth and research focus (e.g. radioactive waste disposal, natural gas exploration, or carbon dioxide storage). The interpretation of the experimental results can be difficult and sometimes a clear discrimination between different mechanisms (and the controlling parameters) is not possible. This holds, for instance, for gas breakthrough experiments where the observed transport can be interpreted as intermittent, continuous, capillary- or dilation-controlled flow. Also, low gas flow rates through samples on the length-scale of centimetres can be equally explained by effective two-phase flow or diffusion of dissolved gas.

The effect of microstructural heterogeneity on pore size distribution and permeability in Opalinus Clay (Mont Terri, Switzerland): insights from an integrated study of laboratory fluid flow and pore morphology from BIB-SEM images

Abstract Opalinus Clay (OPA) is considered as a potential host rock for the deep geological disposal of radioactive waste. One key parameter in long-term storage prediction is permeability. In this study we investigated microstructural controls on permeability for the different facies of OPA. Permeability and porosity were determined under controlled pressure conditions. In addition, the pore space was investigated by SEM, using high-quality surfaces prepared by broad ion beam (BIB) milling. Water permeability coefficients range from 1.6×10−21 to 5.6×10−20 m2; He-pycnometer porosities range between approximately 21 and 12%. The sample with the highest He porosity (shaly facies) is characterized by the lowest permeability, and vice versa (carbonate-rich sandy facies). This inverse behaviour deviates from the generally reported trend of increasing permeability with increasing porosity, indicating that parameters other than porosity affect permeability. Visible porosities from SEM images revealed that 67–95% of the total porosity resides within pores smaller than the SEM detection limit. Pore sizes follow a power-law distribution, with characteristic power-law exponents (D) differing greatly between the facies. The carbonate-rich sandy facies contains a network of much larger pores (D(shaly)≈2.4; D(carbonate-rich) c. 2.0), because of the presence of load-supporting sand grains that locally prevent clay compaction, and are responsible for a higher permeability.

Evolution of small-scale flow barriers in German Rotliegend siliciclastics

Abstract Many siliciclastic reservoirs contain millimetre-scale diagenetic and structural phenomena affecting fluid flow. We identified three major types of small-scale flow barriers in a clastic Rotliegend hydrocarbon reservoir: cataclastic deformation bands; dissolution seams; and bedding-parallel cementation. Deformation bands of various orientations were analysed on resistivity image logs and in core material. They are mainly conjugates, and can be used to validate seismically observable faults and infer subseismic faults. Bedding-parallel dissolution seams are related to compaction and post-date at least one set of deformation bands. Bedding-parallel cementation is accumulated in coarser-grained layers and depends on the amount of clay coatings. Apparent permeability data related to petrographical image interpretation visualizes the impact of flow barriers on reservoir heterogeneity. Transmissibility multiplier calculations indicate the small efficiency of the studied deformation bands on flow properties in the reservoir. Deformation bands reduce the host-rock permeability by a maximum of two orders of magnitude. However, host-rock anisotropies are inferred to reduce the permeability by a maximum of four orders of magnitude. The relative timing of these flow barriers, as well as the assessment of reservoir heterogeneities, are the basis for state-of-the-art reservoir prediction modelling.

Shale Porosity - What Can We Learn from Different Methods?

While the determination of porosity on sandstones is well established, porosities determined on shales are much less straightforward due to limited coring or inadequate pore preservation. Porosity in shale has an important control on many petrophysical, geomechanical and geochemical parameters of shales. Most of the porosity in shales is associated with small pore throat sizes, ranging in diameter from few up to about 100 nm. Pore throat sizes in carbonate or sandstone reservoir rocks are typically determined using mercury injection porosimetry (MIP). It is however well understood that MIP on shales underestimates porosity due to its limited accessibility. It is well known that using different methods for determining shale porosity results in different porosity values which is due to the different accessibility. Nonetheless, porosity is generally used as an absolute, intrinsic parameter without considering the method for determination. To address this issue we compare porosity, specific surface areas and pore volume distributions from fluid invasion and radiation methods on a total of 14 different Opalinus Clay samples recovered from the shaly facies at the Mont Terri underground laboratory in St. Ursanne, Switzerland.

The Dependency of Diffusion Coefficients and Geometric Factor on the Size of the Diffusing Molecule: Observations for Different Clay-Based Materials

In order to investigate in more detail the relation between the size of diffusing molecules and their diffusion coefficients (and geometric factors), diffusion experiments with gases of different size and tritiated water (HTO) have been performed on different clayey samples (Boom Clay, Eigenbilzen Sands, Opalinus Clay, Callovo-Oxfordian Clay, and bentonite with different dry densities). We observed that, for unreactive gases in clayey materials, the effective diffusion coefficient varies with the size of the diffusing molecule and this variation can be described by an exponential or a power law function. The variation of the geometric factor can also be described by an exponential function. The observed experimental relations can be used to estimate diffusion coefficients; by measuring experimentally in clay the effective diffusion coefficient of two unreactive dissolved gases with a different size, the diffusion coefficients of other dissolved gases (with a size in between the two measured gases) can be estimated by using the fitted exponential relationship.