Gas hydrate characterization in sediments via x-ray microcomputed tomography

microcomputed tomography ( μ CT) has gained recent attention in producing high-resolution three-dimensional images of the pore structure and habits of hydrate-bearing sediments and providing the spatial distribution and morphology of gas hydrate. Further, μ CT offers the direct visualization of the hydrate structure and growth habits at a high resolution ranging from the macro-to micro-metric scale; therefore, it is extensively used in NGH characterization. This review summarizes the theoretical basis of μ CT imaging spanning the setup of the experimental apparatus and visualization techniques. The applications of μ CT in NGH reservoir characterization, such as hydrate types and their constituents, physical and chemical properties, occurrence, and accumulation, are presented. Hydrate characterization using μ CT imaging is explicitly discussed, including a general understanding of hydrate pore-habit prediction, saturation and percolation behavior, seepage and permeability, and the influence of hydrate saturation on the mechanical properties of hydrate-bearing sediments. Last, conclusions and recommendations for future research are provided. This review offers a reference for understanding the application of μ CT to evaluate gas hydrates, which contributes to exploiting these energy resources.


Introduction
The world is moving towards the clean energy and advanced methods to meet the net zero emission targets (Al-Anssari et al., 2021;Ali, 2018;Ali et al., 2022b,a;Iglauer et al., 2022;Mahesar et al., 2020;Pal et al., 2022).Natural gas hydrates (NGHs) are the most efficient natural gas resources and have remained stable since their discovery in the 1960s.These ice-like solid compounds comprising hydrocarbons are available in marine and permafrost (permanently frozen ground) conditions.In addition, NGHs are found in permafrost regions at depths of 0 to 900 m and marine areas at depths of 300 to 500 m (Chibura et al., 2022).Gas hydrate resources have a geographically heterogeneous distribution and exist in a far greater amount than both conventional and unconventional hydrocarbon resources, which is significant for energy provision worldwide.
Numerous aspects of gas hydrates, including hydrate formation, recovery, gas production, depressurization, reserve evaluation, and thermal gas generation, have recently been the focus of laboratory and field research (Ripmeester, 2000).Therefore, a comprehensive understanding of the physical and chemical changes during hydrate formation and decomposition is vital for refining existing gas production models from gas hydrate accumulation and for efficient recovery planning (Swinkels and Drenth, 2000).Heat capacity and thermal conductivity are critical physical properties that govern the thermal gas generation potential, whereas relative permeability characteristics depict the associated fluid flow dynamics in hydrate-bearing sediments (HBSs).
The occurrence of gas hydrate in porous media significantly depends on specific environmental conditions (temperature and pressure; (Abbasi et al., 2021a;Abbasi et al., 2021b;Chong et al., 2016;Yin et al., 2016).Further, NGHs are formed under low-temperature and highpressure conditions where they are thermodynamically stable.This fact implies that hydrates become more stable with decreasing temperature and increasing pressure, as illustrated in Fig. 1.However, researchers have argued that gas hydrates (formed in a laboratory for experimental investigation) can form at moderate or ambient pressure conditions (Abbasi et al., 2021a;Abbasi et al., 2021b;Tariq et al., 2022).Furthermore, experiments replicating natural NGH conditions to investigate hydrate behaviors are conducted in high-pressure vessels (Aksnes and Kimtys, 2004;Chaouachi et al., 2015;Chen and Hartman, 2018;Haligva et al., 2010;Li et al., 2017a;Li et al., 2017b).
Several experimental techniques have been employed to investigate the physical behavior of HBSs.For example, ultrasonic technology has been employed to elucidate primary information on hydrate formation and study the influence of ultrasonic waves on hydrate dissociation (Biot, 1955;Bu et al., 2022a;Miura et al., 2006;Winters et al., 2007).Raman spectroscopy has routinely been used to examine the structure of gas hydrates.Raman spectroscopy provides information on the structure type and can identify the phase transformation, hydration number, growth process, and other aspects (Sasaki et al., 2003;Sum et al., 1997;Uchida et al., 1995).Furthermore, the micromodels, which serve as a lab-on-a-chip (Al-Shajalee et al., 2022;Arif et al., 2019), have also been used to characterize gas hydrates.In particular, gas hydrate morphology, occurrence, and growth have been visualized using highresolution microfluidic imaging (Almenningen et al., 2018;Mahabadi et al., 2016a;Mahabadi et al., 2016b).Microscopic and macroscopic investigations can also be conducted using nuclear magnetic resonance (NMR), a powerful tool for providing such information (i.e., morphology, occurrence, and growth).Since the 1960s, NMR has been employed in hydrate compound studies.Moreover, the capability of NMR in instant image acquisition facilitates the recording of the dynamic process of HBSs (Cha et al., 2015;Gilson and McDowell, 1961;Melnikov et al., 2012).However, the three-dimensional (3D) spatial distribution and morphology of gas hydrate cannot be represented by any of the discussed methods.Nevertheless, it is crucial to examine the hydrate structure and growth.Thus, x-ray microcomputed tomography (μCT) has been extensively applied to produce high-resolution images of the pore structure and habits of HBSs (Gupta et al., 2009;Kneafsey et al., 2010;Seol and Kneafsey, 2009).The advancement in imaging technology is evident from the ten-fold increase in μCT acquisition speed every five years (Arif et al., 2021), indicating the promising potential of its applications.Furthermore, μCT can be combined with other techniques, such as acoustic, electrical, and mechanical measurements (e.g., scanning electron microscopy [SEM] and low-field NMR) to promote μCT technology.For instance, (Bu et al., 2022b) investigated the influence of microscopic distributions on acoustic characteristics by evaluating the elastic wave velocity and hydrate saturation during hydrate decomposition using μCT imaging.Moreover, (Bu et al., 2022a) integrated porescale visualization using μCT and an ultrasonic test system.These integrated systems also offer innovation to existing μCT applications.In addition, researchers have also integrated μCT with other techniques to analyze gas hydrate physical properties.For instance, (Zhao et al., 2022) combined resistivity measurements with μCT to derive the 3D spatial structure and resistivity of HBS.(Li et al., 2019b)   testing integrated with μCT that can record the cementation failure behavior of hydrate saturation, hydrate dissociation characteristics in HBS, and the effective stress and strain rates.(Zhang et al., 2021b) combined μCT and NMR to evaluate pore-scale characteristics.Previous reviews have focused on the feasibility of using gas hydrates as an efficient energy source (Chong et al., 2016;Liu et al., 2019;Zhu et al., 2021), the accumulation behavior of NGHs (Xie et al., 2020), and the formation and dissociation kinetics of gas hydrates (Dong et al., 2021a;Hassanpouryouzband et al., 2020;Malagar et al., 2019).In comparison, others (Ruan et al., 2021) have highlighted using numerical methods to characterize gas hydrates.
Thus, the present review focuses on the applications of μCT in gas hydrate reservoir characterization.This study presents the theoretical basis of μCT.Then, the application of μCT in NGH studies is discussed.In particular, the review emphasizes the published data on estimating hydrate saturation in porous media, hydrate pore-habit prediction, and percolation using μCT imaging.The review also provides insight into the results from studies in the last decade.Thus, this work presents a fundamental understanding of the implementation of μCT in gas hydrate characterization.

Gas hydrates
Gas hydrates are crystalline solids consisting of gas molecules surrounded by a water molecule cage.The hydrogen bonds connect water molecules.The gas molecules are trapped in the cage and are not bonded.The van der Waals force holds the gas molecules.Hydrogenbonded water molecules form gas hydrates.However, gas hydrates possess different crystallographic forms stabilized by gas molecules (Dillon, 2002).Gas hydrates are usually formed by methane (CH 4 ).However, other hydrocarbons, such as carbon dioxide (CO 2 ) and hydrogen sulfide (H 2 S), can also form a gas hydrate.

Physical properties of gas hydrates
Physical properties are a crucial aspect in the design of processes involving gas hydrates (Table 1).Estimating the physical properties of gas hydrates is complex because they depend upon multiple factors: the (1) hydrate type, (2) saturation degree, and (3) encaged guest molecule in the gas hydrate.The mechanical and electrical properties and heat capacity of gas hydrates are similar to those of ice.However, thermal conductivity significantly differs from ice (Handa and Cook, 1987).
(a) The molar mass (molecular weight) depends on the saturation degree and crystal structure.The molar mass also relies on pressure and temperature because the saturation degree is a function of these parameters.The molar mass of gas hydrates can be obtained using Eq. ( 1): where N W denotes the number of water molecules per unit cell, M W indicates the water molar mass, the fractional occupancy of cavities of type I by component j is represented by Y ij , vi denotes the number of type i cavities, n reveals the number of cavities of type i, and the number of components is represented by c.
(b) Density is the mass of molecules divided by the unit volume, as presented in Eq. ( 2).The hydrate density is similar to that of ice; however, hydrocarbon-gas hydrates are significantly denser: where N A represents Avogadro's number, and the volume of the unit cell is V cell .
Thermal conductivity is the most crucial property of hydrates.The hydrates are significantly less conductive than ice.The thermal conductivity of hydrocarbons ranges from 0.50 ± 0.01 to 2.2 W/m K for ice.Thermal conductivity is widely considered to thaw/melt/dissociate the hydrates.Due to their low thermal conductivity value, hydrates require a longer duration to thaw.
(c) Researchers rarely study the heat capacity of hydrates.The heat capacity of hydrates is assumed constant because there is a narrow range of hydrate temperatures.The heat capacity values of hydrates are provided in the following table, where the heat capacity of ice is also included for comparison.
Other physical properties of gas hydrates include the volume of the gas hydrates and enthalpy of fusion (heat of formation).

Chemical properties and typical hydrate types and formers
Gas composition determines the structure, composition, and thermodynamic properties of gas hydrates (Table 2 and 3).The gas composition also influences the kinetics and hydrate formation.Complex mixed hydrates and simple methane hydrates have been recovered from natural sediments (Sloan et al., 2007).Complex mixed hydrates have more than one component, in which cages of the same kind are occupied by two or more molecule types, with the restriction of, at most, one molecule per cage.Simple methane hydrates have only one guest species.
There are two common hydrate types (type I and II), also called structure I (sI) and II (sII), as mentioned.Another hydrate type infrequently encountered (type H) is usually referred to as structure H (sH).
The simplest hydrate structure is sI, comprising two cavities.There are 46 water molecules in sI, represented as 46H₂O + 2S + 6 L, where S is a small molecule, and L is a large molecule.Methane (CH₄), carbon dioxide (CO₂), ethane, and hydrogen sulfide (H₂S) are common sI formers.In the ethane hydrate, the ethane molecule engages in large cavities, whereas CO₂, CH₄, and H₂S can engage in both large and small cavities.

Hydrate type
Heat capacity (J/g  Comparatively, sII is slightly more complex than sI, and two cavities also form it.There are 136 water molecules in sII, represented as 136H₂O + 16S + 8 L. The most common sII formers are propane, nitrogen, isobutene, and tetrahydrofuran (THF).Propane and isobutene engage in large cavities, whereas nitrogen can engage in both large and small cavities.
An uncommon hydrate type is sH.A small molecule, such as CH₄, is required, along with an sH former, to form this hydrate type.There are three cavities and 34 water molecules in sH.To form sH, a small molecule, such as CH₄, engages in small, medium, and large cavities, whereas larger molecules (sH formers) engage in a large cavity.Theoretically, it is represented as 34H₂O + 3S + 2 M + 8 L. Both sI and sII require only one former; however, sH requires two hydrate formers (e.g., CH₄, as a small molecule, and larger molecules, such as sH formers).Moreover, sH comprises the following hydrocarbons: 2-methybutane, 2,2-dimethylbutane, 2,3-dimethybutane, 2,2,3-trimethylbutane, and 2,2-dimethylpentane.

Table 3
Basic properties and hydrate formation of common natural gas components, based on (Carroll, 2020)

Hydrate resource occurrence
The occurrence of gas hydrates depends on low-temperature and high-pressure conditions, where the presence of gas is sufficient to initiate and stabilize the hydrate.At <2000 m subsurface depth, these conditions are met in certain regions: at (1) cold surface temperatures (˂0 • C) and continental areas and (2) high pressures (˃3 MPa) and cold bottom water (~0 • C) and submarine continental slopes.The worldwide occurrence of gas hydrates is presented in Fig. 3.

Theory of microcomputed tomography imaging
Microcomputed tomography (μCT) is a nondestructive imaging technique used to acquire 3D porous-structure imaging data and quantitatively investigate opaque systems (Cnudde and Boone, 2013;Maire and Withers, 2014).Moreover, μCT technology is based on the same principles as those applied in medical computerized axial tomography.As the radiation of the x-ray beam passes through the sample, it experiences continuous attenuation due to the absorption of the constituent atoms (Li et al., 2019a).A relationship between the x-ray transmission intensity, incident radiation, and linear absorption coefficient is provided by Lambert-Bee's equation: where I o represents the x-ray transmission intensity, I l denotes the incident radiation, μ is the linear absorption coefficient, and d is the thickness of the sample (Maser, 2001).The x-ray transmission intensities can be measured using a complementary metal-oxide semiconductor and charge-coupled devices (Chung et al., 2019).The CT number N is reported by the CT image given by Eq. ( 4): where μ w represents the linear absorption of water, and c denotes a scaling constant (Li et al., 2017a;Li et al., 2017b).The gray value contrast between various phases produces the different attenuation used in CT.Different N values describe the gray level of various matters in CT images, such as − 1000 for air and 0 for pure water.

Microcomputed tomography experimental setup
A synchrotron-based μCT and a cone-beam or fan-beam lab-based μCT are the most common types of μCT setup.The cone-beam μCT uses geometrical magnification by placing the sample between the detector and x-ray source for examination.The highest attainable resolution (<1 μm) can be achieved with microfocus x-ray tubes (Cnudde and Boone, 2013; Maire and Withers, 2014;Mees et al., 2003).In contrast, synchrotron-based μCT generates a parallel x-ray beam that transits via the monochromators using a bending magnet.A scintillator screen detects the beam after transmitting the sample that transforms the x-rays to light.Visualizing optics are employed to magnify the image on a detector, such as a complementary metal-oxide semiconductor or chargecoupled device camera.A high resolution can be attained with a voxel size of ~10 nm (Chao et al., 2005;Dierolf et al., 2010;Godard et al., 2011) using coherent diffractive imaging.The experimental lab-based μCT setup schematic diagram is presented in Fig. 4.

Three-dimensional visualization
A reconstruction approach can be used to compute the x-ray attenuation inside a stack of 2D slices after radiographic projections of the sample have been obtained.Some software packages, such as 3DMA, Fiji, and Avizo, can reconstruct the 3D sample (Ketcham, 2005;Lindquist et al., 2000;Schindelin et al., 2012).Artifacts and noise in μCT imaging can occur during data acquisition (Blunt et al., 2013;Ketcham and Carlson, 2001;Mathews et al., 2017;Saxena et al., 2017;Van Geet et al., 2000).Usually, some practices are conducted to reduce or diminish them before the segmentation process (Blunt et al., 2013;Mathews et al., 2017;Otsu, 1979).Moreover, before presenting the raw data for the reconstruction process, numerous mathematical filtering methods enhance the signal-to-noise ratio of the data (Cnudde and Boone, 2013; Mathews et al., 2017).Moreover, μCT is a powerful tool for quantitative investigation because it enables 3D data visualization, such as grain analysis and pore-system characterization (Cnudde and Boone, 2013;Van Geet et al., 2000).Further, μCT imaging quantifies the internal rock structure and high-resolution computations using a quantitative investigation to determine the pore geometry, mineral assemblages, pore-size distribution, fracture-aperture distribution, and porosity (Dong et al., 2021b;Isah et al., 2022).

Gas hydrate characterization using microcomputed tomography imaging
For a better understanding of gas hydrate formation and the dissociation mechanism, it is essential to analyze the pore-habit alteration, saturation and percolation behavior, and permeability and seepage characteristics that cannot be observed using traditional methods.In addition, μCT can provide direct visualization of the hydrate structure; hence, it is extensively used in hydrate studies.The following sections discuss and critically review the current literature on gas hydrates using μCT.

Gas hydrate pore-habit predictions using microcomputed tomography
This section reviews the literature available in the past decade on using μCT on hydrate pore habit, growth, and occurrence (Table 4).The appearance of NGH is affected by several factors, such as the pore-throat diameter, local capillary force, wettability, and pore-surface roughness.Currently, there are limited studies of the hydrate growth pattern in sediments.Researchers have often assumed different growth patterns and conducted experiments covering possible hydrate pore habits.
The μCT technique can identify the NGH phase spatial distribution in the pores.(Zhao et al., 2016) executed methane hydrate formation experiments and used three sizes of sand particles (2.85 to 4.8 mm, 1.18 to 2.85 mm, and 0.85 to 1.18 mm).The authors observed hydrate formation in the pore space and along the surface of the sand particles.The hydrate growth pattern is random in the pore space during hydrate formation; however, hydrates are primarily present in the pore space without contacting the pore surface; thus, hydrate pore filling can advance in all directions.This mode of hydrate filling in porous media is called a floating model (Sadeq et al., 2018;Yang et al., 2018).The surrounding medium is compacted due to hydrate formation because hydrate formation extracts water from sediments.This effect is due to the temperature decline in the pressure-temperature curve of the hydrate phase diagram (see Fig. 1).
Moreover, (Lei et al., 2019) published comprehensive research on methane hydrate formation over 13 days.However, they declared that hydrate grows over the surface of sand particles and usually forms a network in the 3D volume.The NGH usually forms an inhomogeneous distribution in the pore space.Several researchers have reported an inhomogeneous hydrate distribution in partially and fully brinesaturated porous media (Schindler et al., 2015;Yang et al., 2016a;Yang et al., 2016b;Yang et al., 2016c).(Schindler et al., 2015) reported partial hydrate contact with grain at a higher hydrate saturation.Fig. 5 indicates that hydrate formation occurred in the pore space and seemed connected between the grains.This process is attributed to the initial growth at the grain surface, connecting by slowly growing in the pore space.
Furthermore, different shapes of hydrate growth are found in porous media, and spike shapes appear in gas pockets in several spots.This shape has been termed soda straw stalactites and ice spikes by various researchers (Libbrecht and Lui, 2004).An essential factor for hydrate formation is the growth rate.In addition, μCT can calculate the growth rate in a high-pressure vessel.The decline in hydrate growth rate is observed initially and rises dramatically with overpressure and undercooling.
Hydrate nucleation (the initial process of hydrate formation involving the arrangement of the constituent particles into hydrate crystal) is promoted by the presence of clay minerals.The induction time with clay is 180 min less than without clay (Lei et al., 2019;Lei and Seol, 2020).The presence of sodium dodecyl sulfate (SDS) in an appropriate amount (300 ppm) also promotes the induction time of hydrate formation.A low concentration of SDS decreases the bubble size, expanding the effective mass transfer area; thus, the process of hydrate formation advances.At a high SDS concentration, the interfacial adsorption of SDS has a dominant effect and impedes the gas molecule diffusion from the gas to the liquid phase (Lv et al., 2020).Thus, the promoting efficacy of SDS on hydrate formation is restricted.

Gas hydrate saturation and percolation behavior using microcomputed tomography
This section discusses the use of μCT to investigate the saturation and percolation behavior of HBSs Table 5.The movement of a fluid through a fully interconnected network is called percolation.Hence, percolation is the flow behavior of HBSs.Percolation behavior is affected by several factors, such as permeability, porosity, hydrate saturation, wettability, pore size, and capillarity (Daigle, 2016;Wang et al., 2020).It is generally agreed that hydrates grow (thus, increasing hydrate saturation) and percolate within porous media.However, different modes of saturation changes and the percolation behavior of gas hydrates have been reported by various researchers based on their experimental observations and numerical analyses (Fig. 6).
Microcomputed tomography scanning is a nondestructive method for obtaining detailed observations of hydrate formation and fluid migration and insight into the changes in hydrate saturation within HBSs.(Wang et al., 2018b) studied methane hydrate formation in sand samples.Their results demonstrated that hydrate grows in graincementing mode except in some cases, where pore-filling and floating models were observed.In another study, (Wang et al., 2015a) conducted experiments on methane hydrate formations in a sand sample (Fig. 7).They observed a thin layer of free water between the sand grains and hydrate and stated that the hydrate does not adhere to the pore walls.They noted that NGH grows in pores in the pore-filling model.
Moreover, (Seol and Kneafsey, 2011) conducted an experimental study and numerical simulations of methane hydrate formation in an unsaturated sand sample.The authors suggested that hydrate behaves as a pore body-filling hydrate accumulation.This pore-filling behavior is attributed to increased hydrate saturation, which decreases the average pore and throat radii (Wang et al., 2015a).However, (Sadeq et al., 2018) performed experiments on CO 2 hydrate in Bentheimer sandstone.The reported CO 2 hydrate predominantly grows in pores at the gas-water interface without contacting the grain surface.They also measured the P-wave velocities, revealing an increase of 7% to 8% after hydrate formation, verifying the hydrate-free-floating behavior.Furthermore, (Zhao et al., 2014) observed a water layer coating the grains.They found that gas hydrates formed randomly in the pore space instead of contacting the grain surface, which is consistent with the findings by (Sadeq et al., 2018).

Gas hydrate saturation and mechanical behavior
Hydrate saturation, the proportion of hydrate in the pore space volume, is discussed by several researchers of HBSs (Table 6).However, researchers have rarely explored the interplay between the hydrate saturation and mechanical behavior of HBSs using μCT; therefore, we highlighted the available literature in recent years on various hydrate saturation phases and the corresponding mechanical characteristics of the HBS using different modern instruments.Moreover, modern instruments and techniques depict multiple benefits of analyzing the Note: CO 2 = carbon dioxide; NaCl = sodium chloride; NaBr = sodium bromide.
G.R. Abbasi et al. formation mechanisms, microstructural properties, and thermal physical properties of gas hydrates.Various techniques have been successfully incorporated into gas hydrate-related research with the advancement of specialized experimental tools.Fig. 8 presents four hydrate saturation phases: hydrate-free sediment, low hydrate saturation, intermediate hydrate saturation, and high hydrate saturation.The pore-scale behavior of the sand particles during the shearing process for hydrate-free sediments is primarily governed by the energy principle, and the sand particles could reorganize to control the increased shear stress and friction by slippage, rotation, and other means (Wu et al., 2020).However, crushing and breakage may occur when the shear stress surpasses the strength of the sand particle, resulting in a significant decline in the strength of the sediment.
Moreover, dilation occurs in dense sediments (fine particles), and compaction occurs in loose sediments.In low hydrate saturation, hydrate saturation has less effect on the strength due to the low hydrate volume.The hydrate formation is on the surface of the sand particle without being in contact with other particles.Hence, the hydrate cementation strength is lower than the shear stress, and the influence of the hydrate on the sediment strength is finite (Kajiyama et al., 2017).The volumetric dilatancy can be escalated considerably because the crushed and broken hydrate particles occupy the surrounding pore space to increase the sediment density.In the case of high hydrate saturation, the sand particle rearrangement is hindered by pore-filling and cementing hydrates.
Because of the formation of the skeleton, the hydrate cementation strength could be significantly improved in the initial state once the generated hydrate particles attach (Wu et al., 2020).Typically, the formation of a large hydrate cementer cluster occurs, making rearrangement and rotation challenging during the shearing process.When the hydrate-sand bonding strength is greater than the hydrate cementation strength in these clusters, a shear plane forms through the hydrate mass, and the mechanical properties of the hydrate dominate the HBS strength evolution (Wu et al., 2020).This outcome is primarily anticipated in sediments with low specific surface areas and rough surfaces (Wu et al., 2020).Thus, when the cementation strength exceeds the hydrate-sand bonding strength, failure emerges along the hydrate-sand interface, and several hydrate particles appear along the shear plane due to shearing, resulting in more noticeable strain dilation and softening.This result is more anticipated in HBS with high specific surface areas or smooth surfaces than in others (Yun et al., 2007).

Gas hydrate permeability/seepage using microcomputed tomography
Seepage characteristics of HBS play an essential role in accurately estimating natural gas production and characterizing fluid behavior in exploiting gas hydrate resources (Table 7).The relative permeability of the gas and water phase and the porous media is critical among these seepage characteristics, and μCT can present a detailed understanding of hydrate morphology, which is a critical factor for calculating permeability.Furthermore, μCT scans can provide an accurate and integrated pore structure of HBSs for advanced numerical simulations for permeability prediction.The available literature in the past 10 years on seepage characteristics of HBSs using μCT is discussed in Table 7.In the following paragraphs, we discuss studies examining the seepage characteristics of HBSs employing μCT.
Permeability laboratory measurement methods are primeval and usually conducted in cylindrical press-volume-temperature (PVT) vessels.In addition, μCT experiments are employed to observe the flow behavior of a gas/water flooding test in HBSs.For instance, (Seol et al., 2006) used methane-hydrate-saturated consolidated sand packs (fine) to estimate gas permeability.They applied the gas flow through the consolidated fine sand packs and measured the pressure variation at both ends of the PVT vessel.A water flooding test was conducted to investigate the water-flow behavior in HBSs.The variation in saturation was observed at fixed spots using μCT.The results indicate that the water-flow behavior deviates from higher porosity to higher hydrate saturation.(Kneafsey et al., 2010) reported a more complex water-flow behavior.They stated that the water-flow path improves at higher hydrate saturations with increased local capillary suction on some occasions.In many cases, water flow is obstructed, and the relative permeability of gas is low at higher hydrate saturations because gas hydrates engage most of the pore space.In addition, μCT scanning was used to detect the displacement front of injected water in a core holder, which was correlated with the adjustment in the differential pressure of each layer in the core holder.A precise prediction of the effective permeability of the water/gas phase was obtained by removing the capillary end effect with this method.
In core-flooding experiments, the precision and reliability of the measured permeability values are limited due to the nature of the flow law, location of the μCT observation area, core size, and experimental conditions.Many advanced simulation models have been employed in studies to predict the permeability of HBSs to resolve these challenges.The pore network model is one of the numerical models capable of evaluating the 3D percolation properties of gas, water, sand, and hydrate.The advancement in μCT techniques has benefited the pore network model.Thus, the permeability progress during hydrate formation/dissociation is better simulated, and more accurate permeability measurements for the water/gas phase are acquired.(Wang et al., 2018a;Wang et al., 2015a;Wang et al., 2015b;Wang et al., 2016a) conducted extensive research on the seepage properties of water/gas flow in HBSs using μCT imaging in conjunction with the pore network model.In a high-pressure vessel containing consolidated quartz  sand packs, methane hydrate was fully formed.Dong's maximal ball method was used to extract pore structure information from HBSs preserved by μCT images, which was derived from (Silin et al., 2003) and the maximal ball concept (Zhang et al., 2020).
The extraction of the pore network model from the original μCT images is depicted in Fig. 9.The simulation of the pore network model revealed that the capillary pressure increases with methane hydrate saturation.Conversely, the relative and absolute permeability of the water phase were inversely proportional to the hydrate saturation.(Wang et al., 2016a) further studied the wettability influence and particle size effect on the permeability of HBSs.They reported that waterphase relative permeability was negatively correlated with the hydrophilicity of the porous media under the same conditions.(Yang et al., 2018) investigated the inhomogeneity of the porous media using three compositions of minerals (quartz, dolomite, and feldspar).The anisotropy of pore structures causes the difference in permeability.Higher absolute permeability was achieved in smaller pore/throat ratios.This outcome demonstrates that the absolute permeability is low due to the complex pore structures of HBSs.

Other experimental investigations on gas hydrates
The physical and chemical properties of gas HBSs are critical parameters for producing gas hydrate.Gas hydrate production is significantly dependent on proper experimental analyses and simulations.Some properties of gas hydrates can be examined with natural samples, and the specifications of the gas hydrate characteristics can be acquired by investigating hydrate samples.Moreover, experiments provide an overview of gas hydrates.This section provides a concise discussion on gas hydrate characterization using various techniques.
Studying physical and chemical properties is crucial to understanding gas hydrate behavior.The primary focus of conventional gas hydrate evaluation techniques is pressure, temperature, and saturation (Wu et al., 2018).Modern techniques, including x-ray CT, SEM, Raman spectroscopy, high-pressure differential scanning calorimetry (DSC), xray diffraction (XRD), and NMR, have been applied to study the formation and dissociation mechanisms; structural analysis; physical, chemical, and thermal properties; and phase equilibrium of gas hydrates.Many methods have been successfully implemented with the advancements in experimental tools in studies related to gas hydrates.
Microstructures of gas hydrates have been successfully studied using NMR, XRD, and Raman spectroscopy.A basic study of gas hydrate has been conducted using Raman spectroscopy, which is most suitable to deduce hydrate existence, its molecular composition, cage occupancy, and structure type.The hydrate number and structural type can be obtained by examining the chemical bonds of water molecules (O-H bonds) and guest molecules (C -H bonds and C -C bonds).Both NMR and XRD are crucial tools for identifying the structure type, formation and dissociation kinetics, and gas storage performance (Zhang et al., 2017).The hydrate space groups, crystal structure, cavity occupancy, hydration index, and lattice parameters can be determined using XRD.For the multicomponent hydrate structure type, NMR is a powerful method and highly sensitive; therefore, it helps investigate the gas molecule dynamics and chemical behavior of different gas hydrates.Hence, NMR can be employed to study gas hydrate formation and dissociation kinetics and structure information (Fu et al., 2017;Meng et al., 2011b;Meng et al., 2015b).
Researchers in several areas of China have employed Raman spectrometry and XRD to study gas hydrates (Liu et al., 2015;Liu et al., 2017;Liu et al., 2012;Liu et al., 2010a;Meng et al., 2011a;Meng et al., 2015a).Their results revealed that most hydrates form sI and accumulate 99% of methane gas.The shift from thermodynamics to kinetic measurements is challenging for researchers (Wu et al., 2018).Magnetic resonance imaging (MRI) and CT are used to study the microscopic kinetics of hydrates in HBSs.The key parameters that control the growth and nucleation of hydrates in the pore space can be investigated microscopically using these instruments.The study of the formation and dissociation mechanism, evaluation, and exploration of gas hydrates is related to understanding the occurrence of gas hydrates.X-ray CT can be used to observe the microscopic distribution of hydrates in porous media and investigate the dissociation mechanism and internal spatial structure of HBSs.The kinetics and formation dissociation mechanisms and microscopic distribution have been investigated by researchers using xray CT (Hu et al., 2014;Jiang et al., 2005;Li et al., 2013;Pu et al., 2007;Wang et al., 2015a;Wu et al., 2006;Zhang et al., 2016).
In addition, MRI reveals a brightness difference between the solid hydrate and free water, which depicts the hydrate formation and dissociation.Proton density affects the MRI signal.Hence, MRI differentiates the phase states and produces quantifiable data for porosity, fluid distribution, pore-size distribution, and variation in hydrate saturation in the pore space.Several researchers have used MRI to investigate hydrate formation, dissociation, kinetics, saturation, fluid migration, and behavior in porous media (Liu et al., 2010b;Qingguo et al., 2012;Yao, 2010).
Furthermore, neutron scattering has also been extensively applied to investigate the gas hydrate structure (Hosseini et al., 2021).For instance, gas hydrate formation and dissociation have been quantified, which is significant in the hydrate "gun" hypothesis and unconventional fuel resource recovery.Furthermore, neutron scattering can be employed at high gas pressures, which exist in gas hydrate formations (Hosseini et al., 2021).
High-pressure and low-temperature reactors have been employed to study the phase equilibrium of gas hydrates.Phase equilibrium data with various components are acquired in abundance from gas hydrates (Du et al., 2011;Fan et al., 2001;Liang et al., 2001;Litao et al., 2009;Sun et al., 2010;Sun et al., 2002a;Sun et al., 2002b;Sun et al., 2001).Hydrate phase equilibrium data were obtained by (Ma et al., 2008) from systems comprising methane, hydrogen, ethylene, and ethane in the presence and absence of THF in water.Their results indicated that hydrogen could be successfully separated from other gases by forming a gas hydrate.(Litao et al., 2009) also observed hydrate phase equilibrium data from various solution systems.
High-pressure DSC is becoming prominent in investigating the hydrate phase equilibrium over the past few years because the contemporary PVT reactors are challenging to use, and the obtained data are usually ambiguous.The heat flux estimation during a hydrate phase change can be performed under high-pressure conditions by employing high-pressure DSC.Accordingly, specific heat and thermal dissociation equilibrium data can be acquired under the desired conditions.(Chen et al., 2012) investigated the phase equilibrium of hydrates in porous media and pore water using this method.Their results indicated that, with a pressure range from 10 to 30 MPa, the temperature of the hydrate phase equilibrium is approximately 2 K lower than that in pure water.However, this requires more study under specific conditions at sea.

Recommendations
While significant progress has been achieved in using μCT technology to evaluate hydrate pore behaviors and characteristics, we highlighted some areas that require further investigation: • A limitation of the μCT technology is that the sample size should be small enough to acquire high-resolution images, which are essential    to determine the high precision of hydrate morphology in the small pores of HBS.Hence, extensive research to address this limitation is needed.• More complex techniques (e.g., pore network modeling and artificial intelligence deep learning) are required to reproduce μCT images to optimize the restoration of HBS morphology and computational accuracy.
• The implication of numerous parameters on hydrate nucleation behavior and seepage characteristics, such as salinity, grain surface wettability, and pore structure reformation, requires extensive research.
• The presence of organic materials in reservoirs influences wettability.Thus, studies on the influence of wettability on gas hydrate formation and dissociation are crucial.Furthermore, such studies using μCT can provide critical insight into gas hydrate characterization.Hence, it is recommended that researchers investigate gas hydrate wettability using μCT imaging.
• Numerical and molecular dynamic simulation methods can have great potential for NGH characterization because experimental methods may encounter challenges due to the complexity of actual geological conditions.It can also provide a means of validating observations from μCT and other experimental techniques.
• Fossil fuels, such as oil and coal, are depleting rapidly, and reliance is emerging on renewables, such as solar, wind, geothermal, and hydrogen, as new clean energy producers to achieve the global energy demand with no greenhouse gas emissions.Hydrogen is gaining substantial attention among these renewable energy sources as a sustainable substitute.Hydrogen can be produced from nonrenewable and renewable sources and is considered the "fuel of the future."Hence, an innovative approach to hydrogen storage via gas hydrates is recommended.

Summary and conclusions
The use of μCT imaging technology to investigate gas hydrate formation and dissociation mechanisms, such as pore behaviors, permeability and seepage properties, and saturation and percolation of HBSs, is critical.The distribution and occurrence patterns of hydrates were directly studied using the μCT experimental technique.In addition, the μCT imaging technique can provide crucial data for developing more precise numerical models to calculate HBS permeability and other seepage parameters.Moreover, μCT scanning is often used to assess the quality and ability of sediments to produce gas hydrates.Therefore, the current study offers a fundamental understanding of the application of μCT to evaluate gas hydrates.Based on the reviewed data, the following conclusions are drawn: • Three hydrate types are identified: type I (sI), type II (sII), and type H (sH).The occurrence of NGHs depends on the temperature and pressure conditions.Hydrates are thermodynamically stable under low-temperature and high-pressure conditions.• The initial process of hydrate formation involves arranging the constituent particles into hydrate crystals (hydrate nucleation), promoted by the presence of clay minerals.• Nucleation of gas hydrates occurs randomly with many parameters, such as saturation, salinity, pore-size distribution, and mineralogy, affecting the nucleation process.• Estimating gas hydrate physical properties (thermal conductivity, molecular weight, heat capacity, density, and mechanical and electrical properties) is complex because these properties depend upon multiple factors, including (i) hydrate type, (ii) saturation degree, and (iii) the encaged guest molecule in the gas hydrate.Thus, robust characterization techniques and equipment, such as the μCT or recently emerging combined techniques, are needed.
• The advancement in μCT imaging technology is evident from the tenfold increase in its acquisition speed every five years, indicating the promising potential of its applications.• Hydrates grow and percolate within porous media.The percolation behavior is affected by several factors, such as permeability, porosity, hydrate saturation, wettability, pore size, and capillarity.• Different hydrate growth/filling modes include encrustation/coating (laminating), cementing, matrix-supporting, and pore-filling modes.• The mechanical strength of HBSs is significantly affected by the hydrate morphology, saturation, and wettability.• Moreover, the stress-strain behavior of HBSs depends on confining pressure, particle size, and hydrate concentration.As the hydrate saturation increases, the friction angle, strength, cohesion, and stiffness of HBSs also increase.• Last, gas hydrate formation and dissociation greatly affect the flow properties and pore morphology.As hydrate saturation increases, the average pore and throat radii decrease, reducing permeability.

Declaration of Competing Interest
The authors declare that they have no competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
G.R.Abbasi et al.

Fig. 5 .
Fig.5.Gas hydrate growth pattern in porous media using microcomputed tomography: (a) hydrate bridge, (b) hydrate growth on the surface, and (c) hydrate spike shape.Adapted from(Lv et al., 2020) with permission from Elsevier.

Fig. 9 .
Fig. 9. Schematic of pore network mode extraction using microcomputed tomography: (a) methane hydrate in porous media, (b) fluid flowing in the pore space, and (c) pore network representation, adapted from (Wang et al., 2016a) with permission from Elsevier.
developed triaxial

Table 4
Available literature on gas pore-habit predictions using microcomputed tomography.

Table 5
Available literature on gas hydrate saturation and percolation behavior using microcomputed tomography.With increased fine concentrations, gas percolation decreases, and regardless of the wetting phase, pore and throat diameters decrease, which is attributed to the dislocation in the sand • Sand-brine interfaces accumulate hydrophobic particles, and the brine-gas interface includes hydrophilic clay(Zhao et al., • Regardless of the wetting phase and grain size, hydrates nucleate as lens-shaped clusters Note: CO 2 = carbon dioxide.

Table 6
Available literature on hydrate saturation estimation using different observation methods.

Table 7
Available literature on gas hydrate permeability/seepage using microcomputed tomography.