Surface and bulk characterization of reservoir and cap-rocks: Past, present, and future

While zero net carbon emission initiatives may offset carbon emissions in the future, the current level of CO₂ in the atmosphere has already surpassed 400 parts per million, a point we have not reached in a million years.¹ The environmental impacts of increased CO₂ emissions, such as global warming and climate change, necessitate large scale capture and storage of CO₂. The process of capturing CO₂ produced by coal plants and other sources and injecting CO₂ into the subsurface formations for permanent storage is known as carbon capture and storage (CCS). CO₂ injection into the subsurface itself has been employed for the past 40 years to increase oil recovery. While a number of CCS projects have been conducted around the world, one of the biggest concerns of the current and future projects is the possible leakage of CO₂ through insufficient trapping, damaged seals, faults, and fractures.

Conventionally, storage and propagation of CO₂ in the subsurface can be detected using time-lapse seismic data or forward modeled using rock physics/fluid substitution computations. Seismic data lack the resolution to capture information at the nano-scale, and the rock physics models fail to accurately consider the different CO₂ trapping mechanisms. A challenge in accounting for the various trapping mechanisms is caused by a lack of reasonable representation of governing surface interactions between CO₂ and nanoporous structures. The stored CO₂ may interact with nanoporous shale rocks (cap-rocks/seals) and possibly create conduits for flow. While shale layers are conventionally used as cap-rocks, a large amount of CO₂ can also be stored in the micropores of shale by surface interactions such as adsorption.² A special report by the Intergovernmental Panel on Climate Change recommends laboratory studies of reservoir and cap-rocks with native pore fluids and CO₂ to understand the effects of interfacial reactions in complex mineralogies and pore structures.³ 

This review paper discusses the types of reservoir rocks targeted for CO₂ storage, outlines the trapping mechanisms by which CO₂ is trapped in the subsurface, and then summarizes the various methods currently used to characterize bulk mineral, porous, and surface properties of rocks. The overview of widely used rock characterization techniques shows the information that can be gained with the current methods and reveals gap in current characterization tools. A review of x-ray photoelectron spectroscopy (XPS) studies to determine the elemental surface composition of rocks and its correlations to surface wettability are highlighted. Finally, the perspective toward investigating the dynamic surface and interface properties of porous rocks when interacting with CO₂ and other in situ fluids using ambient pressure XPS (AP-XPS) is presented as the future direction of this work.

Subsurface CO₂ storage is recognized to occur through four trapping mechanisms as shown in Fig. 1. The four main mechanisms are (i) structural and stratigraphic trapping where CO₂ flow is thwarted by fault, fracture, or sealing shale layer above the reservoir rocks; (ii) residual CO₂ trapping where CO₂ is trapped by capillary forces in the micropores; (iii) solubility trapping where CO₂ is stored in the dissolved state in brine; and (iv) mineral trapping where CO₂ reacts with surface minerals and is precipitated as carbonate minerals.

Commonly, CO₂ storage estimation and monitoring are conducted based solely on structural or stratigraphic trapping along pathways of sealing faults, fractures, and seals. Seismic monitoring lacks the resolution sensitivity to detect the density contrast between brine and CO₂ dissolved brine. Most models and seismic monitoring technology are unable to consider the residual trapping occurring at nanoporous capillaries and mineral trapping happening at the pore-scale. Interestingly, the permanence of CO₂ trapping is increasingly governed by mineral trapping over time and depends on the reactivity of CO₂ with the existing rock minerals. Note that the time scale shown in Fig. 1 for mineral trapping lies between 1 and 10k years. It has been since recognized that the time scale shown in Fig. 1 for mineral trapping might be too long for some lithologies, such as carbonates and basalts.⁵ 

Understanding the composition of sedimentary rocks (e.g., sandstones, shales, carbonates, and evaporites) is of great importance because the pore volume (or porosity) and the connectivity of pore network (or permeability) in the sedimentary rocks provide a location for CO₂ storage and conduits for CO₂ flow, respectively. Sedimentary rocks are formed either as detrital or clastic sediments (e.g., sand, silt, pebble, etc.) that are transported and deposited by different environments, or as chemical deposits of silicate and carbonate minerals, or as accumulation of organic debris. Subsequent diagenesis processes compact and modify the sediments into sedimentary rocks.⁶ Sandstones are mostly known as conventional reservoir rocks due to their high porosity and permeability, and durability of minerals both chemically and physically.⁷ Sandstones are primarily made of detrital quartz and are considered clean when they contain more than 90% quartz. Other common minerals found in sandstones are feldspar minerals, particularly K-feldspar. Sandstones have varying grain sizes from silt to gravel, with particles ranging from 0.004 to 64 mm. Examples of CO₂ storage in saline aquifers composed of sandstones are the Utsira formation of Sleipner field in the North Sea and Tubåen formation of Snøhvit in the Barents Sea.

Carbonate rocks are formed from skeletal remains of living marine organisms and chemical precipitation.⁷ Carbonates are primarily composed of calcite, dolomite, aragonite, and small amounts of quartz and feldspar minerals. Porosity within carbonate minerals is formed during deposition by an arrangement of carbonate grains and postdeposition by mineral alteration, dissolution, and precipitation. Carbonate rocks are most reactive in a CO₂-rich environment and goes through dissolution in a CO₂-dissolved low pH brine. Therefore, study of chemistry and physical structure of carbonate formation rocks is required for potential CO₂ sequestration activity in carbonate-rich formations.⁸ 

Evaporites are chemical sediments that are precipitates of supersaturated water body.⁷ Examples of evaporites are gypsum, halite, and anhydrite. The extreme low porosity of evaporite sediments makes them a less desirable site for CO₂ storage.

Shales are fine-grained sedimentary rocks with high surface area, large sorption capacity, and low permeability. Grain sizes of shale are in the scale of nano- to micrometers.⁹ While sandstones are mostly composed of quartz, shales have complex mineral composition made of clay minerals, quartz, organic matter (OM), and carbonate minerals. The heterogeneous mineral composition also leads to the complex physicochemical properties of shales. It is important to gain a comprehensive understanding of shales, especially for CO₂ storage and trapping purposes, because the large specific surface area of shale potentially provides a large storage space for CO₂ and its low permeability provides long term trapping of CO₂.

Rocks are complex and do not always look alike. Figure 2 shows examples of the electron microscopy images of sandstone, shale, carbonate, and evaporite that demonstrate possible structure, composition, and mineral colocation of different types of sedimentary rocks.^10–13 Shale rock is known to be most heterogeneous in its composition and structure. Therefore, this paper provides examples of two shale samples, Niobrara (N1) and Agardhfjellet (A1), to show the extent of their complex surface chemistry and the distinct characterization techniques required to gain full understanding of shale rocks. N1 is a carbonate-rich Niobrara shale rock with negligible organic content. The concentration of its mineral composition toward carbonate minerals makes it a relatively homogeneous shale sample. A1 is predominantly clay rich followed by silicate minerals, along with the presence of organic matter. The mineral heterogeneity of A1 compared to N1 can be observed through the multitechnique measurements presented in Sec. IV.

In general, all sedimentary rocks can be described as materials composed of minerals making up the matrix and fluids filling the pore space. Figure 3 schematically illustrates the heterogeneous composition of a rock composed of minerals making up sandstone, carbonate, and shale. Sandstones are composed of a mixture of quartz and K-feldspar, carbonate can be a mixture of calcite, dolomite, and siderite minerals, and shale can be made of nanoporous clay minerals (smectite, illite, mica, kaolinite, etc.) with traces of quartz. Organic matter commonly exists as an additional material in shale rocks. In Fig. 3, we classify grains composed of quartz as sandstone, grains composed of clay minerals as shale, and grains composed of carbonate minerals as carbonate rocks. One way to distinguish between the minerals is by their chemical compositions. For instance, SiO₄ is indicative of the quartz mineral. Table I shows chemical compositions of rock minerals identified in Niobrara (N1) and Agardhfjellet (A1) samples. Larger porosity is formed by the pore space between mineral grains. Pores are termed as interparticle when they exist between minerals and intraparticle when they exist within a mineral. Intraparticle-porosity may be external and accessible, or internal and inaccessible by fluids. The schematic in Fig. 3 succinctly shows that a complete characterization of CO₂ host formations requires an understanding of the bulk minerals that make up the composite rock, their respective elemental content, and pore structure. The collective information is necessary to understand how fluids fill up pore space and react with mineral surfaces.

Mineral precipitation and dissolution could alter both rock and fluid chemistry. CO₂ is soluble in water and forms carbonic acid that reacts with the host rock causing mineral dissolution and carbonate precipitation, a process known as mineral carbonation where CO₂ can be permanently trapped. Studies show both carbonate and clay minerals dissolve in acidic brine, while a simultaneous precipitation takes place at a slower rate of kinetics.¹⁵ Several elements in the subsurface can be carbonated, but alkaline earth metals such as calcium, magnesium, and iron have been investigated the most to date.¹⁶ Shown below are reaction steps for the precipitation of calcium carbonate from the dissolution of plagioclase (CaAl₂Si₂O₈).^17,18 The process starts with the (1) dissolution of CO₂ in brine, (2) formation of carbonic acid, (3) dissociation of carbonic acid, (4) dissolution of plagioclase, and (5) subsequent precipitation of calcium carbonate.

A carbonation process takes years to reach reaction equilibrium in the subsurface. For laboratory studies, several carbonation steps can be used to speed up the reaction: direct gas-mineral carbonation in the presence and absence of aqueous environments and additives, and multistep indirect carbonation with the aid of acetic acid, low pH, and other additives.¹⁶ The multistep process separates the reaction of mineral carbonation and hence, reduces the total time for each reaction studied individually. CO₂ trapping by the multistep reactions provides motivation to conduct studies at the surface and interfaces of rock minerals.

The heterogeneous nature of rocks with features at various scales (from macro- to micro- to nano-) motivates fundamental studies that incorporate the following characterization of rocks: (1) bulk mineral characterization that extracts information on the grains and minerals making up the matrix, (2) pore characterization that illustrates the fluid spread or access to the mineral surface, and (3) surface characterization that governs rock-fluid interaction at the interface of matrix and pore space down to the elemental scale. Importantly, any surface characterization of the external and internal surfaces of the rocks should be accompanied by mineral grains and pore size distribution analyses. The following sections include data from various literature sources as well as unpublished work by co-authors. Samples A1 (Agardhfjellet) and N1 (Niobrara) are used to describe the common experimental techniques.

The techniques discussed in this section are x-ray diffraction, thermogravimetric analysis, and Rock-Eval pyrolysis. These methods both qualitatively and quantitatively describe the materials/minerals making up rocks.

Each rock mineral has a characteristic crystal structure which can be determined using x-ray diffraction analysis (XRD). In XRD, the diffracted x-rays are detected, and the count is recorded in terms of diffraction angles, which later converts to the distance between adjacent planes of atoms (d-spacing).¹⁹ Minerals are identified from their unique set of d-spacing. Table II shows the mineral composition of Niobrara and Agardhfjellet shale samples in weight percent determined using XRD. Note that N1 is predominantly composed of calcite, whereas A1 is a heterogeneous mixture of several clay minerals and quartz.

Decomposition of the rocks as a function of temperature can provide complimentary information on the mineralogy of the rocks. This can be accomplished using thermogravimetric analysis (TGA) which measures the change in mass, commonly a decrease in mass, as a function of time and increasing temperature. When the temperature is increased, minerals within a rock are decomposed at a temperature range specific to each mineral. Identification of decomposed minerals provide mineral composition in rocks. Other phenomena that occur during thermal analysis are desorption of adsorbed species, such as water molecules and phase transition. Gips²⁰ and Gips et al.²¹ have used thermogravimetric analysis to identify these phenomena and recorded the typical temperature range for mineral degradation in rocks.

Figure 4 shows the decomposition profiles of the Niobrara and Agardhfjellet shale samples, from Table II. For Niobrara shale, the absence of peaks below 200 °C indicates the lack of free water and clay bound water in the sample. The major loss of mass occurs at 700 °C, corresponding to decomposition of calcite minerals. Although mass loss with increasing temperature is not as significant in the Agardhfjellet sample as in the Niobrara shale, several events can be identified with increasing temperature: free and clay bound water are lost at temperatures below 100 °C, illite degrades close to 500–550 °C, smectite continues to degrade after 600 °C, and pyrolysis of kerogen, which is organic-matter, and bitumen, which is an early stage hydrocarbon generated from kerogen, starts as early as 400 °C. TGA provides qualitative assessment of minerals and adsorbed species in rock samples and insights on the effect of temperature on different rock minerals. This information proves useful during surface studies to differentiate the adsorbed species from native minerals.

While XRD and TGA provide qualitative and quantitative mineral composition, Rock-Eval pyrolysis provides additional information to the total composition by determining the total organic carbon (TOC). As powdered samples are exposed to increasing temperatures during pyrolysis, three major processes occur, each accompanied by the release of hydrocarbon species. The release of free hydrocarbons (peak S1) is followed by the generation and release of hydrocarbon from kerogen (peak S2), release of CO₂ due to kerogen cracking (peak S3) and finally residual carbon which is not converted (S4).²² The TOC is determined by combining S1, S2, and S4 peaks, while a combination of TOC with S1, S2, and S3 gives hydrogen index (HI), oxygen index (OI), and production index (PI) as shown by the widely used equations below.

HI indicates the amount of hydrogen and OI indicates the amount of oxygen relative to TOC of a sample. HI and OI are used to evaluate the type of kerogen. The temperature at which maximum hydrocarbon is generated during the pyrolysis (T_max) denotes organic maturity, with higher T_max corresponding to higher organic maturity. Maturity is an essential information as matured organic matter have aromatic carbon for which CO₂ has higher affinity compared to the aliphatic carbon on the surface of immature kerogen.²³ Table III shows the Rock-Eval analysis for the Niobrara shale and the Agardhfjellet shale samples. Note that Niobrara has low organic carbon, maturity, and hydrocarbon generation. Therefore, Niobrara has lower storage capacity for CO₂ compared to Agardhfjellet because (1) Niobrara lacks the amount of kerogen pores available for storage, (2) CO₂ has lower affinity to aliphatic carbon on immature shales, and (3) the lower hydrocarbon generation from Niobrara shows only a limited amount of pores become available for subsequent storage after hydrocarbon production.

This section focuses on the common pore characterization methods found in the literature for geological samples: water, mercury, and helium porosimetry, nuclear magnetic resonance, and gas adsorption. Measurements on N1 and A1 are only shown for the adsorption method.

The immersion/saturation method measures connected porosity by calculating the difference between dry and fully saturated samples with a liquid of known density. Prior to each measurement, the samples are oven dried with a user-specific temperature to remove the free and adsorbed species such as the volatile hydrocarbon, contaminants, residual, and sometimes clay-bound water in the pore space without altering the solid matrix. As described below, the choice of saturating liquid depends on several properties especially in its ability to penetrate pore spaces and low reactivity with the mineral composition.

The high penetration coefficient of water allows it to access nano-pores and capillaries that are abundant in shales. After drying, the sample is weighed dry and saturated with water, and the Archimedes principle is used to calculate bulk density (g/cc), grain density (g/cc), and porosity of sample (p.u.). Kuila et al. provide a detailed experimental and numerical procedure for shale samples.²⁴ Due to limited access to organic matter and the tendency of clay minerals to swell with water, the saturating fluid can be changed to kerosene.²⁴ Topόr et al. present how to utilize both water and kerosene using dual liquid porosimetry for oil- and gas-bearing shales.²⁵ 

Mercury intrusion porosimetry (MIP) is used to measure the pore throat distribution in rocks. Pore throat is the entrance diameter to the inner porous space, and empirical relations can be used to calculate pore size distribution from pore throat distribution. In MIP, mercury, a nonwetting liquid, is injected into the sample with incremental pressure step that allows the mercury to access different pore sizes. Pore throat distribution can be calculated using the Washburn equation.^26,27 Refer to Olson and Grigg for MIP conducted on shale samples.²⁸ Studies also show the limitation of MIP to access pores smaller than 3.6 nm in diameter due to the excessive pressure needed to access these pores.²⁹ 

Helium porosimeters (HPs) compute porosity based on gas expansion as formulated by Boyle's law. As helium gas at known volume and pressure is allowed to expand into a sample, the gas pressure decreases until equilibrium expansion is achieved in the entire system.^30–32 The grain volume can then be calculated from the new pressure and volume. If the bulk volume is known, porosity can be calculated. Refer to Oliveira et al. for a detailed example of helium porosimetry for sandstone and carbonate rocks.³² 

Pore spaces of most reservoir rocks are filled with water and hydrocarbon fluids, both consisting of a hydrogen atom in their molecular formula. Nuclear magnetic resonance (NMR) can directly measure the mass of hydrogen nuclei in the reservoir fluids due to its ability to read the distinctive angular momentum of an atom, caused by spinning of the nucleus.³³ Since the density of hydrogen atom is known, the information obtained from NMR can be converted to volume filled by the hydrogen containing molecules, which then can be correlated to the effective porosity of the rock assuming that it was fully saturated. T₂ transverse relaxation time, indicating how fast the tipped proton from hydrogen atoms in the fluid relaxes transversely relative to the axis of static magnetic field, is sufficient to invert pore size distribution and total porosity of a core sample as shown by Eq. (10).³³ The relaxation rate is a product of surface relaxivity $(ρ)$ and surface to volume ratio of the pore.³⁴ With known surface relaxivity, the pore size distribution can be approximately obtained from NMR T₂ relaxations,

Figure 5 shows the T₂ relaxation time obtained for a set of nine sandstone cores saturated with 2000 ppm KCl brine. The smaller pores of grains tend to have a large surface to volume ratio which causes quicker relaxation time of the hydrogen nuclei due to increased electron interaction with the large surface area.³⁵ Thus, shorter and longer relaxation times can be approximately associated to smaller and larger pores, respectively. The relaxation times in Fig. 5 give qualitative visualization toward the pore size distribution.

Gas adsorption measurements allow access to the following pore sizes: ultra-micropores (<0.7 nm), micropores (<2 nm), mesopores (2–50 nm), and macropores (>50 nm), as defined by the International Union of Pure and Applied Chemistry (IUPAC).^36,37 Adsorption is the enrichment of fluids in an interfacial layer, where the free fluid, sample surface, and adsorbed fluid are denoted as adsorptive, adsorbent, and adsorbate, respectively, to be consistent with the terminology accepted by IUPAC.³⁷ Adsorption can be utilized for understanding porous material using a low-pressure or subcritical adsorption experimental setup. The experimental method is based on static manometric adsorption, where a known volume of adsorptive is dosed into a precalibrated volume of a sample cell containing the adsorbent at equilibrium pressure points. Gas adsorption phenomenon can be illustrated through an isotherm as shown in Fig. 6. The hysteresis between the adsorption and desorption curves depends on the pore geometry of the adsorbent.^36,37

Interpretation of isotherms provides information on physical properties such as specific surface area of pores, pore size distribution, and cumulative pore volume of a sample. The phenomena captured by an isotherm can be divided into four sections as shown in Fig. 6: (i) micropore filling where stronger attraction between walls of the narrow pores attracts gas to begin filling micropores first, yielding information on microporosity;^36,39,40 (ii) monolayer coverage where the adsorbate tends to form a monolayer on pore surfaces due to stronger adhesion to adsorbent than adsorptive molecules, yielding specific surface area of pores;⁴¹ (iii) capillary condensation due to the successive layers of adsorptive forming with increasing relative pressure until pore condensation takes place in the mesopores, leading to the hysteresis in the isotherm;⁴² and (iv) presence of macropores is indicated by the steep increase in the adsorption volume near P/Po ≈ 1.³⁶ 

Specific surface area of the pores can be inverted from the isotherm using the Brunauer–Emmett–Teller (BET) model.^37,43 Several models exist for inverting pore size distribution and porosity from an isotherm. Figure 7 shows the pore size distribution obtained for the Niobrara (N1) and Agardhfjellet (A1) shales using density functional theory (DFT). DFT is a statistical method that constructs the configuration of an adsorbed layer at a molecular level, to quantify pores at the micropore level. The pore sizes displayed in Fig. 7 vary with the adsorptive used. CO₂ is limited to access pore widths ranging from 0.3 to 1.5 nm, whereas N₂ accesses pores between 0.7 and 30 nm.

This section transitions from morphological analysis to compositional analysis, starting from atomic force microscopy, electron microscopy to x-ray photoelectron spectroscopy. Measurements on N1 and A1 are only shown using the electron microscopy technique.

Surface topography and roughness of rocks can be evaluated by atomic force microscopy (AFM), which operates based on the cantilever deflection in response to material surface forces and surface topology. The monitored friction, topography, and small feature images as shown in Fig. 8 can be used to investigate surface properties in terms of composition, roughness of specific area, hardness, and surface forces.^44,45 Using the tip diameter as small as nanometers, AFM is able to assess shale composition at the nano-scale, especially that of organic matter.⁴⁵ AFM has been used on geological materials with various goals: (1) evaluating surface wettability by deriving force versus distance relation between fluid and mineral surface; (2) tracking mineral reactions such as dissolution, precipitation, and mineral growth like that of carbonate minerals in salt solution; and (3) measuring elastic properties and stiffness of clay and organic matter for geophysical interpretations.^46–51 

Electron microscopy (EM) is a common technique for the characterization of rock mineral morphology and their respective elemental composition. In SEM, a focused beam of electrons is accelerated by an electron gun toward the sample surface. As electrons interact with sample, they produce backscattered electrons, secondary electrons, and x-ray energy from an electron rearrangement in the atoms. Energy dispersive x-ray spectroscopy (EDS) detects the x-ray energy that is characteristic of the element from which it was emitted. FE-SEM operates on a similar principle but, because of the emitter type, a cold field emission gun, provides high topographic contrast with a resolution of 1 nm as compared to the resolution of SEM (15–20 nm).^52,53 The Environmental-SEM (E-SEM) allows the user to image samples at low vacuum conditions. Figure 9 shows a comparison between FE-SEM and E-SEM for Agardhfjellet shale at an accelerating voltage of 10 kV and a magnification of 100×. The FESEM shows higher contrast on the sample morphology compared to E-SEM.

Coupling SEM/FE-SEM/E-SEM with EDS allows us to map the distribution of elements and to trace the minerals forming the rock matrix on the surface. Table IV shows the distribution of elements found in Niobrara (N1) and Agardhfjellet (A1) samples. For 88.9% calcite-rich Niobrara, every atom of calcium in calcite (CaCO₃) should have an equal amount of carbon and three times the amount of oxygen. Carbon element is a lighter element that cannot be accurately captured by E-SEM analysis. Note that A1 is mineralogically more heterogeneous than N1, as indicated by prior XRD studies. This is also shown by the additional elements present in A1 that are not found in N1.

Therefore, heterogeneity of samples can be assessed, and morphological features can be correlated to composition. Table V shows the elemental heterogeneity of Agardhfjellet shale in average atomic percent, conducted with the goal to differentiate the prominent bright features observed among the darker background in an E-SEM image of Agardhfjellet (Fig. 10).⁵⁴ The darker features have higher carbon, aluminum, silicon, potassium, and iron elements, metal cations characteristic of clay and organic rich minerals. Both bright and dark features have relatively high silicon and oxygen, indicative of SiO₄ of quartz. This region is sand- or silt-rich. Note that despite the slight differences, both dark and bright spots are composed of a mixture of various elements, evident to the heterogeneous mineral composition of shales at any location. The major limitation of SEM for analysis of rocks is its lack of sensitivity toward carbon, which is an important element that signifies the presence of organic matter in shale rocks in the context of their applications for CO₂ storage.

Other electron microscopy techniques are also increasingly being used for geomaterials. Transmission electron microscopy (TEM) is a non-destructive method that studies interfaces at the nanoscale.^55,56 SEM can also be combined with focused ion beam (FIB-SEM) to extract images of microstructures in shale in 3D.⁵⁷ 

XPS provides relative quantification of elemental and chemical composition at an information depth of 5–10 nm from the surface. Sensitivity of XPS allows to determine functionalities and oxidation states of elements and to detect small differences between samples. Such ability of XPS provides opportunity to understand surface chemistry of rocks in more detail. Figure 11 shows an example of a survey spectrum obtained from a reservoir rock demonstrating elemental composition.

XPS has the ability to detect carbon, unlike other surface instruments conventionally used for characterizing rock surfaces. This advantage has advanced the use of XPS to characterize minerals of rocks, especially kerogen/bitumen composed of organic carbon,^59–66 and carbonates such as siderite, calcite, and dolomite composed of inorganic carbon.^67–69 

XPS characterizes organic matter such as kerogen or bitumen by quantifying the amount and the functionality of carbon in these minerals. XPS has also been utilized to estimate heteroatoms such as nitrogen, sulfur, and oxygen in kerogen. The functional groups of carbon provide essential information for petroleum generation such as thermal maturity and kerogen type. Figure 12 shows the XPS high resolution C1s spectra of a shale rock with deconvolutions of aromatic carbon (sp²), aliphatic carbon (sp³), oxidized carbon (C—O, C=O, O—C=O), and inorganic carbon $(CO3=)$⁠.⁶¹ Note that the higher coverage of aromatic carbon compared to aliphatic carbon in the C1s spectra indicates relatively mature rocks. Maturity and kerogen type can also be identified based on the atomic ratio of oxygen to carbon as shown by the Van Krevelen diagram (refer to Fig. 13).⁷⁰ The inorganic carbon $(CO3=)$ in Fig. 12 is associated with carbonate minerals that coexist with the organic matter in the rock matrix. The binding energy associated with the inorganic carbon typically ranges from 289.4 to 290.1 eV.⁶⁹ 

Characterization of nitrogen and sulfur in kerogen is equally important as the presence of these elements determines the quality of petroleum. The presence of nitrogen and sulfur makes the crude oil hazardous, and they need to be eliminated using costly processes. XPS determines the relative amount of these elements in the rocks. Additional knowledge of the functional groups of nitrogen and sulfur provides independent confirmation about the thermal maturity of rocks based on carbon 1s. Figures 14(a) and 14(b) show the common functional groups of nitrogen and sulfur, respectively.⁶⁶ Kerogen contains five major forms of organic nitrogen: pyridinic, amine, pyrrolic, quaternary, and nitrogen oxide, and five major forms of organic sulfur: aliphatic, aromatic, sulfoxide, sulfone, and a small amount of inorganic pyrite sulfur.⁶⁶ Some observations on nitrogen and sulfur behavior with regard to kerogen maturity and type are (1) aromatic sulfur to total organic sulfur increases with aromatic carbon and maturity;^64,65 (2) amine nitrogen is transformed into pyridinic nitrogen by cyclization with increasing aromatic carbon and maturity;⁶⁶ (3) type I kerogen has a higher ratio of aliphatic to aromatic sulfur than type II kerogen;⁶⁵ and (4) pyrrolic and pyridinic nitrogen account for majority of the nitrogen functional groups regardless of maturity and kerogen type.⁶⁵ 

Existing characterization of rocks, especially organic matter, using XPS provides the basis for correlating XPS-derived chemical composition with fine scale rock surface wettability, which is discussed in more detail in Sec. IV D.^58,71–75

Wettability of rocks can be defined as the preference of one fluid to wet the rock surface compared to another fluid. The pore surfaces of rocks consist of multiple minerals with varying preferential affinities for pore fluids such as brine, hydrocarbon, and CO₂. The interfacial chemistry of composite minerals determines the effective wetting behavior of rocks. Wettability is a surface property that governs rock-fluid interactions and several CO₂ trappings, particularly by adsorption.

Techniques commonly used to determine pore surface wettability are Amott–Harvey core flooding, centrifuge capillary flooding, relative permeability tests, and contact angle. The Amott–Harvey test is laborious and takes several months for samples with low permeability.^72,76 For centrifuge flooding, the reliability can be limited if it does not correspond to equilibrium conditions.⁷¹ Relative permeability tests allow inferring wettability from the end-point data, but the end-points tend to have skewed data for strongly oil wet samples.⁵⁸ The three-phase, flat-surface contact angle may not be applicable for matrix pores of tight reservoirs.⁷⁷ The common limitation of these techniques is lack of fine-scale wettability evaluation, which can be supplemented with the use of XPS.

One major advantage of XPS is its ability to detect light element such as carbon, which has been widely utilized to predict oil wettability of reservoir rocks. As discussed in Sec. IV C, XPS can distinguish between organic carbon, inorganic carbon, and adsorbed air-borne carbon; this distinction is important as wettability properties are affected by the type of carbon. For example, crude oil tends to preferentially wet organic molecules, therefore the carbon content that determines oil wettability is the organic type. Organic carbon is typically present in solid organic matter (source of hydrocarbon), asphaltene (solid component of crude oil), and other polar fractions in crude oil. A literature survey presented in Table VI summarizes the objective, methods, and sample preparation used by various groups to study wettability of rocks with XPS.^58,71–75

Organic carbon can be determined by estimating amount of aromatic and aliphatic carbon or the C—H species through curve-fitting of the C1s spectra. To study the effectiveness of XPS in detecting organic carbon, Huang and Holm aged the rocks with the asphaltene-containing toluene solvent to alter the composition of rock surfaces (refer to Table VI).⁷¹ Aging is a term used when a sample is saturated with certain liquid for a period of time until the sample has reacted with the liquid. After aging with asphaltene, Huang and Holm observed an increase in the organic carbon content of the samples.⁷¹ 

The ability of XPS to qualitatively estimate wettability is reinforced by its correlation to the wettability index obtained from the Amott–Harvey method, a core flooding method widely used to characterize the wettability of rocks (refer to Fig. 15). The Amott–Harvey scale provides wettability indices for a drainage process ranging from −1 to 1 where −1 refers to highly oil wet and +1 indicates highly water wet. Mitchell et al. Quet et al., and Toledo et al. show the positive correlation between atomic percentage of organic carbon and oil wettability (Amott–Harvey): oil wet samples show higher carbon content than water wet samples.^58,72,73 To show a general correlation, the data from Mitchell et al., Quet et al., and Toledo et al. were combined, as shown in Fig. 15. Note that the lines drawn serve to indicate potential trends across reported data.

The collective trend matches the correlation shown by each paper, where oil wet indices (orange dots) show increasing carbon content with stronger oil wetting tendency, and water wet indices (blue dots) show decreasing carbon content with stronger water wetting tendency. Note that for intermediate wet rocks (shown by dashed box in Fig. 15), the carbon content covers a large range of carbon atomic percent. The intermediate region can also be denoted as weakly oil wet and weakly water wet rocks, or mixed wettability rocks. Further studies of elemental composition and speciation of samples falling into this region are required to derive correlations at the intermediate region.

Information about the presence and nature of other elements and their speciation is also studied to assess wettability of rocks. For example, oil wettability is linked to heteroatoms (O, N, and S) present in the structure of solid kerogen and of asphaltene. Asphaltene is a dissolved solid component of crude oil.⁷⁸ Additionally, a sample with higher Si—CH as compared to Si—O is considered more hydrophobic because the bond of silicon from quartz and phyllosilicates to carbon from organic matter makes it less water-wet.⁷³ Hydrophilicity (attraction to water), on the other hand, is observed by the elements associated to clay minerals such as K, Mg, Na, Al, Si, and O.^58,71

Despite the usefulness of elemental characterization using XPS, Mitchell et al.⁵⁸ and Durand and Beccat⁷⁴ showed that error in the elemental quantification of geological samples ranges from 5% to 20%, suggesting that XPS studies be combined with other analysis methods such as electron microscopy coupled with EDS to evaluate the elemental prediction errors. One possible factor affecting the accuracy of surface information can be surface contamination. Air-borne carbon likely creates a coating on the surface affecting the surface composition obtained from XPS. Air-borne carbon content is often significant enough, between 6% and 20% of atomic percentage, to mask the original surface elements.^58,72 Using a comprehensive study of carbon contamination on clay minerals, Durand and Beccat showed that while homogeneous carbon coating on the surface does not alter the ratios of Si/Al, K/Al, and O/Al (elements of kaolinite and illite), it does reduce their total signal due to attenuations.⁷⁴ To illustrate the effect of surface contamination, Cánneva et al. show the C 1s XPS spectra of shale rock before and after ion sputtering with Ar⁺ (refer to Fig. 16 where i refers to before sputtering and i_s refers to after sputtering).⁶³ Before sputtering, the majority of signal is at 285 eV; these species are due to surface contamination with aliphatic and aromatic carbon. After sputtering, the main signal around 285.0 eV reduced after eliminating surface carbon contamination and another signal indicating inorganic carbonate centered around 289.0 eV appears.

While bulk mineral, pore, and surface characterizations including wettability of rocks are abundant in the literature, we identify a gap in knowledge in qualitative and quantitative characterization of the chemical and physical interactions that occur in the interface and alter pore surface and fluids, particularly CO₂ during CO₂ sequestration. Such studies are not possible using conventional XPS instruments that require ultrahigh vacuum conditions (10⁻⁹ Torr) to maintain an operational environment for the x-ray source.⁷⁹ 

The ambient pressure XPS (AP-XPS) can be operated at much less stringent vacuum conditions and allow us to measure solid-vapor and solid-liquid-vapor interactions at near ambient pressure (greater than 2500 Pa) and elevated temperatures (<800 °C).⁸⁰ Note that AP-XPS allows pressure to be as high as near ambient pressure and is not representative of reservoir pressure, but in the absence of strict vacuum condition, AP-XPS still gives fundamental knowledge of in situ reactions at the rock-fluid interface in the presence of various gases and moisture, filling the gap that exists in the literature. To investigate the effect of CO₂ injection on rock surface and interface composition, it is of high interest to conduct AP-XPS while dosing relevant gases such as CO₂, moisture, and hydrocarbon gas. While there are no AP-XPS data reported for investigation of rocks, these types of studies have been conducted for other materials and applications. Investigations of catalysts for various catalytic reactions are just one such example.^81,82

The idea of AP-XPS was first developed in the late 1960s by Siegbahn and his co-workers by testing low volatility liquids in XPS environment with the aid of differential pumping.⁸³ Since its early development, the use of AP-XPS has been expanding to study solid-gas interfaces and solid-liquid interfaces of catalysts and other materials. Generally, UHV-XPS is not suitable for samples that outgas significantly such as rocks. The AP-XPS also inherently deals with charge compensation when analyzing insulating materials by ionizing the free gas molecules in the XPS environment.⁸⁴ A series of tests have been published in the Surface Science Spectra on gases, liquids, biological products, and synthetic polymers.^84–90 An example of natural/biological calcite sample investigated by AP-XPS without the presence of external gas is shown in Fig. 17.⁹¹ The split peaks in Ca 2p is a spin–orbit doublet with Ca 2p1/2 at 351.1 eV and Ca 2p3/2 at 347.5 eV, corresponding to calcium carbonate. The double peak in the C1s spectrum contains peaks at 285.1 eV and 289.8 eV, indicating the presence of hydrocarbon (C—C/C—H) carbonate. Natural calcite is a common mineral existing in rocks. Similar principles, therefore, can be applied to other naturally occurring minerals and geological samples.

The increasing interest in CO₂ sequestration motivates both, further investigations with XPS and new studies with AP-XPS to better understand chemical alteration of reservoir and cap-rocks due to interactions along CO₂ pathways. First, the conventional XPS can be used to study samples before and after CO₂ treatment at high pressure and temperature conditions that expediate mineral carbonation process. Second, AP-XPS can be used to study the in situ interactions to gain insights on the carbonation mechanism itself. Due to the polymineralic nature of rocks, it will be useful to conduct XPS analysis on individual minerals before analyzing a composite shale sample. The minerals of interest are minerals commonly found in reservoir and cap-rocks as shown in Table II. Gas adsorption narrows the list down to minerals that can contribute to CO₂ storage and Hellevang et al. list the divalent cation minerals capable of forming carbonates with CO₂ reaction.⁹² 

Sample preparation and experimental condition are determined based on information gained from preliminary measurements. Mineral composition and organic carbon content information furnished by XRD and Rock-Eval, respectively, provide a measure of carbon content that exists in the sample. Information from TGA analysis suggests the optimal temperature to be used to outgas adsorbed species without degrading the minerals of interest. Pore characterization shows the specific surface area of the sample, providing an estimate for outgassing time depending on the sample surface area exposed for carbon contamination and adsorption of moisture. Depending on the carbon content, several outgassing steps with different periods of time may need to be tested with several samples to determine the best practice. Furthermore, electron microscopy surface analysis narrows the choice of elements to capture in XPS studies. After loading the sample into a glovebox that is directly connected to the instrument, it is recommended to first cleave the sample to expose a fresh surface and then heat-treat the sample at a suitable temperature.

Once the sample is loaded into the instrument through the glovebox, fluids such as water, methane, and CO₂ can be dosed individually to assess their direct reaction with the rock minerals. Water and methane are naturally occurring fluids in the reservoir, whereas CO₂ can be naturally occurring or introduced for storage purposes. For controlled moisture condition in XPS analysis, the sample can be outgassed with heat treatment in a glovebox to eliminate the naturally adsorbed water, according to the information provided by prior TGA analysis. Then, water vapor can be introduced at a specified rate, temperature, and pressure in the XPS chamber to accurately quantify the effect of amount of water vapor and experimental conditions on the sample.

Sequentially, water can be injected followed by methane, and finally, CO₂. Water is of interest because it is naturally present in rocks at an irreducible saturation. Following water, methane gas can be dosed to introduce the hydrocarbon molecules that are often trapped in some pores of the rocks in formation. Finally, CO₂ can be introduced to study the reaction of CO₂ in the presence of moisture and hydrocarbon with the minerals. For oil-wet rocks, however, the hydrocarbon molecules will be adsorbed to the rock surface. Thus, the dosing can be changed to methane followed by water vapor and CO₂ in a sequential manner.

The fluids can also be dosed simultaneously to study preferential adsorption of fluids at different areas on the sample. This necessitates spatial mapping of the chemical composition on the rocks. For simultaneous dosing, any two gases from water vapor, methane, and CO₂ can be dosed together. For the final step, the three gases can be dosed together.

Rock characterization should include bulk, surface, and pore characterization as well as quantification of rock-fluid interactions at the surface and interfaces. Bulk mineral characterization of rocks is an important first step to identify all minerals present in the rock matrix, hence a combination of XRD, TGA, and Rock-Eval is recommended to identify common minerals, adsorbed species, and total organic carbon, respectively. Additionally, TGA identifies the appropriate temperature range for AP-XPS analysis. Elevated temperature is necessary to increase the kinetics of reactions between CO₂ and the host rocks,⁹³ but note that high temperature could alter the minerals, causing phase transition of organic matter and thermally degrading some minerals. Second, pore characterization detailing the pore volume available for CO₂ storage in each dominant mineral in the rock is necessary to determine minerals of interest for XPS studies. For this, the gas adsorption method can be utilized. Besides the ability to directly use CO₂, the gas adsorption method can also distinguish the nature of a sample to either permanently or temporarily store CO₂ in the pore spaces. For instance, an open hysteresis between the adsorption and desorption curves in an isotherm can indicate permanently trapped CO₂ in a sample. Third, EM/EDS is needed to assess elemental composition to indicate possible elements, and these studies can be complemented by UHV-XPS analysis. Electron microscopy is also invaluable for the identification of sample heterogeneity and provides valuable information for both sample preparation and selection of representative areas. One needs to determine how many areas should be analyzed in XPS to get statistically relevant data.

XPS has generally been used to determine surface chemical composition and correlate it to wettability of rocks. For instance, carbon content on pore surfaces is directly related to oil wetting tendency of the rocks. While rock wettability indicates what may adhere to and what may bypass the rocks, a direct quantification of the process and subsequent chemical alteration on rock surfaces will provide useful information. XPS under UHV conditions can be conducted on initial samples and after exposing the samples to conditions that rapidly alter the chemistry and morphology of the rocks. Such conditions may include CO₂ at high temperatures and high pressures.

AP-XPS offers an opportunity to probe surfaces of geological samples in the presence of gases and moisture at elevated temperatures and ambient pressure with the purpose of quantifying the mineral-fluid interactions. Fluids could include carbon dioxide, water, and methane gas at varying temperature conditions. These commonly existing in situ fluids in the reservoirs can be dosed individually, sequentially, and simultaneously, providing insights into the complex interactions and potential storage mechanisms. Evaluation of CO₂ interactions at the surface and interface of minerals is important in order to properly understand CO₂ trapping by mineral precipitation during sequestration, providing better idea when choosing potential sites and lithologies for CO₂ sequestration programs.

This material is based upon the work supported by the U.S. Department of Energy (DOE) National Energy Technology Laboratory (NETL) under Grant No. DEFE0023223 and start-up funds from the Colorado School of Mines. Partial support for graduate student was provided by the Chevron fellowship. Special thanks to Michael Dzara for his insightful discussions and ideas in regard to the theoretical and experimental applications of XPS and AP-XPS, Kurt Livo for his help on conducting the experiments on NMR, and Helge Hellevang for providing the organic-rich Agardhfjellet shale from University Centre in Svalbard. The authors also acknowledge resources and support from the Electron Microscopy user facility at the Colorado School of Mines.

 Manju Pharkavi Murugesu completed her B.S. and M.S. from the Colorado School of Mines in the Department of Petroleum Engineering under guidance of Professor Prasad and Professor Pylypenko. Murugesu's interest in Carbon Capture, Storage, and Utilization (CCUS) has led her to study the intricate surface chemistry and physics of rocks and their interactions with reservoir fluids and injected CO₂, along with the simultaneous reactive transport. Manju is a recipient of PETRONAS scholarship, Chevron fellowship, and E-days Engineer award during her times at Mines. Presently, Manju Pharkavi Murugesu is pursuing her Doctoral degree in the Department of Energy Resources Engineering in Stanford University, where she is a 2019 distinguished Knight-Hennessy scholar.

 Manika Prasad is a Professor in the Geophysics Department at the Colorado School of Mines. She has been at Colorado School of Mines (CSM) for the past 14 years, and was previously at the Stanford University and University of Hawaii. She received her B.S. from Bombay University and her M.S. and Ph.D. from Kiel University in Germany. Prasad's main interests lie in understanding microstructural controls on geophysical data. She is the recipient of the Virgil Kaufmman Gold Medal in 2017, the Outstanding Educator Award (2015), and the AAPG-SEG Distinguished Lecturer Award (2012). Known as the “mud queen” among her peers and students, she pioneered integral research in source rich rock and fluid properties using tools and techniques from the geosciences and engineering domains. In addition to her teaching and research duties at CSM, Prasad serves as the 1st Vice President of SEG.

 Svitlana Pylypenko is an Assistant Professor in the Chemistry Department at the Colorado School of Mines. She is also involved in the interdisciplinary Materials Science Program at Mines and holds a joint appointment at the National Renewable Energy Laboratory. Prior to joining the Chemistry Department, Svitlana was an Assistant Research faculty in the Department of Materials and Metallurgical Engineering. Svitlana received her B.S. and M.S. in Chemistry and Chemical Engineering from the National Technical University of Ukraine and Ph.D. in Chemistry from the University of New Mexico. Svitlana's group at Mines investigates surfaces and interfaces of applied materials with the emphasis on building relationships between surface composition and structure, material properties, and their performance with the eventual goal to design next generation of materials based on earth abundant elements which provide high efficiency at the fraction of the cost. Research focuses on multi-technique, multiscale analysis, and in situ and in operando studies bridging surface analysis, surface science, and catalysis. Svitlana served as the board member of the AVS Applied Surface Science Division, Rocky Mountain Chapter of AVS, ECS Physical and Analytical Electrochemistry Division. She was a co-chair of 2014 Surface Analysis Symposium and is a co-chair of the upcoming Surface Analysis Symposium, which will be held at the Colorado School of Mines. Svitlana serves as the chair of AVS Educational Materials and Outreach Committee, is involved in the AVS Science Educators Workshop and is a faculty adviser of the Colorado School of Mines AVS student chapter.