An Enhanced Carbon Capture and Storage Process (e-CCS) Applied to Shallow Reservoirs Using Nanofluids Based on Nitrogen-Rich Carbon Nanospheres

The implementation of carbon capture and storage process (CCS) has been unsuccessful to date, mainly due to the technical issues and high costs associated with two main stages: (1) CO2 separation from flue gas and (2) CO2 injection in deep geological deposits, more than 300 m, where CO2 is in supercritical conditions. This study proposes, for the first time, an enhanced CCS process (e-CCS), in which the stage of CO2 separation is removed and the flue gas is injected directly in shallow reservoirs located at less than 300 m, where the adsorptive phenomena control CO2 storage. Nitrogen-rich carbon nanospheres were used as modifying agents of the reservoir porous texture to improve both the CO2 adsorption capacity and selectivity. For this purpose, sandstone was impregnated with a nanofluid and CO2 adsorption was evaluated at different pressures (atmospheric pressure and from 3 × 10−3 MPa to 3.0 MPa) and temperatures (0, 25, and 50 °C). As a main result, a mass fraction of only 20% of nanomaterials increased both the surface area and the molecular interactions, so that the increase of adsorption capacity at shallow reservoir conditions (50 °C and 3.0 MPa) was more than 677 times (from 0.00125 to 0.9 mmol g−1).


Introduction
In recent decades, climate changes have generated negative consequences such as loss of sea ice, accelerated rise of sea level, extinction of some species, drought and population displacements, among others [1][2][3][4][5]. These changes are caused by human activities, mainly through the release of greenhouse gases [4][5][6][7]. The anthropogenic emissions of carbon dioxide, CO 2 , from fossil fuel combustion and industrial processes contributed to about 78% of the total greenhouse gases emissions [8][9][10], and CO 2 emissions have increased by 46% since pre-industrial times [11,12]. The huge CO 2 daily emissions and their growth increase its responsibility on global climate change [12][13][14][15][16]. Raw material industries such as chemical, petrochemical, iron, steel, cement, and others, contribute with the most to CO 2 emissions worldwide [17,18]. The advantage of CO 2 industrial emissions is that they are fixed sources so that they can be controlled in-situ. A typical flue gas from coal-fired boilers may contain 12-14 vol% CO 2 , 8-10 vol% H 2 O, 3-5 vol% O 2 , and 72-77% N 2 [19][20][21]. The method most frequently used at industrial levels for CO 2 capture is absorption. However, this method has significant limitations related to solvent regeneration, composition of flue gas, CO 2 concentration, corrosion, high-energy consumption, etc. [15,16,18,22]. Other methods are cryogenic distillation, membranes, and adsorption by porous solids. Hence, the current methods are not enough, and CO 2 emissions continue to increase [12,17,23]. For this reason, it is necessary to broaden the portfolio of methods to reduce CO 2 emissions significantly in the years to come.
The Intergovernmental Panel on Climate Change (IPCC) promotes the carbon capture and storage process (CCS), which allows the geological storage of CO 2 emitted by the industry over the long term [10,18,24]. The CCS process might decrease CO 2 emissions by approximately 22% in 2035 [25,26]. The CCS process has three main stages: 1) CO 2 capture and separation from flue gas, 2) CO 2 transport to the storage site, and 3) CO 2 injection into deep geological storage sites, 300 to 4000 m, and 800 m on average [22]. In this case, CO 2 capture and storage are mainly due to the inter-particle volume filling. Figure 1a presents a scheme of the CCS process. [25,26]. The CCS process has three main stages: 1) CO2 capture and separation from flue gas, 2) CO2 transport to the storage site, and 3) CO2 injection into deep geological storage sites, 300 to 4000 m, and 800 m on average [22]. In this case, CO2 capture and storage are mainly due to the inter-particle volume filling. Figure 1a presents a scheme of the CCS process.
It is estimated that 50% of the CCS projects use sandstone deposits for storage because they have technical advantages and high availability [24]. Oil and gas deposits are indeed mostly composed of sandstone, which provides a natural storage structure, with porous texture and upper and lateral seals that allow long-term storage. Carbon capture and storage have been coupled with other processes such as enhanced oil recovery (EOR) or enhanced coal bed methane production (ECBM) [24].
However, its in situ industrial implementation has been unsuccessful mainly due to the technical and economic costs associated with the separation and storage stages. In general, the cost of CO2 capture is 70-80% of the total CCS costs [26][27][28]. For this reason, the current research focuses mainly on increasing the CO2/flue gas separation capacity in order to decrease energy consumption, or on the use of waste energy, among others. This study proposes an alternative to minimize the technical and economic cost for the viability of the CCS process. Figure 1b presents the configuration of an enhanced CCS process (e-CCS), in which the stage of CO2 capture/separation is removed, and the flue gas is injected directly into shallow deposits, at depths less than 300 m. In this case, CO2 remains in a gaseous state, and the adsorption process controls the capture and storage. The density of CO2 is very different in gaseous or supercritical conditions, which affects the amount of CO2 stored in the e-CCS process. For this, it is necessary to add a modifying agent to the surface of the porous medium in order to improve the adsorption capacity and the CO2 selectivity, since the adsorption (capture/storage) process is done underground. In addition, the modifying agent should not affect the naturally porous structure of the deposit to avoid operational problems.  It is estimated that 50% of the CCS projects use sandstone deposits for storage because they have technical advantages and high availability [24]. Oil and gas deposits are indeed mostly composed of sandstone, which provides a natural storage structure, with porous texture and upper and lateral seals that allow long-term storage. Carbon capture and storage have been coupled with other processes such as enhanced oil recovery (EOR) or enhanced coal bed methane production (ECBM) [24].
However, its in situ industrial implementation has been unsuccessful mainly due to the technical and economic costs associated with the separation and storage stages. In general, the cost of CO 2 capture is 70-80% of the total CCS costs [26][27][28]. For this reason, the current research focuses mainly on increasing the CO 2 /flue gas separation capacity in order to decrease energy consumption, or on the use of waste energy, among others. This study proposes an alternative to minimize the technical and economic cost for the viability of the CCS process. Figure 1b presents the configuration of an enhanced CCS process (e-CCS), in which the stage of CO 2 capture/separation is removed, and the flue gas is injected directly into shallow deposits, at depths less than 300 m. In this case, CO 2 remains in a gaseous state, and the adsorption process controls the capture and storage. The density of CO 2 is very different in gaseous or supercritical conditions, which affects the amount of CO 2 stored in the e-CCS process. For this, it is necessary to add a modifying agent to the surface of the porous medium in order to improve the adsorption capacity and the CO 2 selectivity, since the adsorption (capture/storage) (2) CN.MEL: Carbon nanostructures obtained from a solvothermal method, using carbon tetrachloride as carbon precursor and melamine as nitrogen precursor.
Both CN.LYS and CN.MEL were characterized in order to select the best material, considering the nanometer size, adsorption capacity, lower technical and economic cost, and method of synthesis. Ottawa sandstone was used as porous medium and was impregnated with the best nanomaterial at different percentages. The performances of the materials were evaluated by CO 2 adsorption at atmospheric pressure and at 0 • C, and by varying pressure and temperature conditions. The detailed procedures are presented below.
For cleaning, drying, and carbonization, the following chemicals were used, all from Sigma-Adrich again except N 2 : acetone (99.9%), ethanol (99.5%), hydrochloric acid (37%,), tert-butanol (≥99.5%), and N 2 (high purity, grade 5.0). Clean Ottawa sandstone and sandstone from a real reservoir were used as porous media. The real sandstone was obtained from a Colombian oil field, which allows evaluating the real rock that might be used to implement the e-CCS process in depleted oil fields.

CN.LYS Synthesis
The process was adapted from Yong-Rong et al. [48], changing the lysine concentration, the reaction time and the resorcinol/water molar ratio. A solution (S1) of resorcinol/formaldehyde in a 1:2 molar ratio, and deionized water was stirred at 25 • C for 1 h. In parallel, a solution (S2) of L-lysine and deionized water was stirred at 60 • C for 1 h. The molar ratio of resorcinol/L-lysine was 1:0.16. Subsequently, the solutions S1 and S2 were mixed at 60 • C for 1 h to obtain the solution S3. The latter was maintained at 25 • C for 20 h to benefit from the natural precipitation of the nanomaterial. Finally, the obtained polymer was dried at 120 • C for 12 h and carbonized under N 2 flowing at 60 mL min −1 , using a tubular furnace. The temperature was increased up to 800 • C at a rate of 1 • C min −1 , and the final temperature was held for 5 h. The employed molar ratios of resorcinol/water (for S1) were 1:2778 (without dilution), 1:5556 (dilution 1), and 1:11112 (dilution 2) for obtaining the CN.LYS1, CN.LYS2, and CN.LYS3 materials, respectively. Different molar ratios of resorcinol to water were used to reduce particle size.

CN.MEL Synthesis
The CN.MEL synthesis was adapted from Bai et al. [49] by dissolving 2 g of melamine in 120 mL of carbon tetrachloride [49]. This solution was put in a stainless-steel autoclave (Parr Instrument, Illinois, USA) with a capacity of 200 mL and introduced into an oven (Thermo Fisher Scientific, Massachusetts, USA) at 250 • C for 24 h. The synthesis was carried out under auto-generated pressure. Subsequently, the carbonaceous material that formed was separated from the solution, and was cleaned with acetone, ethanol, and finally 0.1 mol L −1 HCl. A mixture of nanospheres and fibers (formed by aggregation of nanospheres) was obtained (Gel.MEL1).
To obtain N-rich carbon spheres, the Gel.MEL1 was coated with a mixture of resorcinol and formaldehyde in a 1:2 molar ratio. Initially, the gel was stirred with sodium dodecylbenzene sulfonate (SDBS) at 25 • C for 18 h to promote the subsequent interaction with resorcinol/formaldehyde. After 18 h, the gel/SDBS was mixed with resorcinol/formaldehyde (in proportions of 0.37 g of nanoparticles per 1 g of resorcinol) and put in a stainless-steel autoclave at 130 • C for 24 h. In order to maintain its porous texture, the gel was dried by freeze-drying. First, the material was impregnated with tert-butanol for three days. Then, the impregnated material was frozen at −5 • C and lyophilized for two days until the tert-butanol was entirely removed (Gel.MEL2). Finally, the lyophilized gel was carbonized under N 2 flow (60 mL min −1 ) in a tubular furnace (Thermo Fisher Scientific, Massachusetts, USA). The temperature was increased at a heating rate of 1 • C min −1 up to 700 • C, and the latter temperature was held for 6 h.

Impregnation of Sandstones
The sandstone was impregnated to decorate the rock surface with the nanoparticles and improve the surface area and molecular interactions. Ottawa sandstone (SS) was impregnated with CN.LYS2 at mass fractions of 0.01, 0.1, 1, 5, 10, and 20% by immersion and soaking [50]. Initially, a nanofluid composed of nanoparticles and deionized water was sonicated at 40 • C for 4 h. Subsequently, the SS was introduced in the nanofluid at 60 • C for either 6 h at 600 rpm or for 24 h without stirring.
The latter method better mimics the reservoir conditions in which the porous medium might be impregnated. Finally, the impregnated material was dried at 110 • C for 12 h. The same procedure was followed for impregnating the sandstone from a real reservoir (RS) but using only a mass fraction of 10 and 20% of CN.LYS2 to RS and 24 h of soaking.

Characterization of the Nanomaterials
The following procedures allowed characterizing the physicochemical characteristics of the materials, essential for a good understanding of the results. For e-CCS application, nanoparticles must have a nanometer size, a spherical shape, a high surface area, and a high nitrogen content.

Size and Structure of Nanomaterials
Different techniques were used to evaluate the particle size distribution: scanning electron microscopy (SEM) was used to obtain the dry particle size, size distribution, and morphology of CN.MEL, whereas transmission electron microscopy (TEM) was used to analyze the dry particle size, size distribution, and surface characteristics of CN.LYS. The CN.LYS showed a different porous structure after dilutions, which was observed by N 2 adsorption (at −196 • C). Further, TEM was used to characterize the structure after dilutions to analyze the causes of pore structure modification. Scanning electron microscopy analysis was also carried out to analyze the distribution of carbon nanoparticles on sandstone after impregnation. The observations were carried out by means of a JSM-7100 emission electron microscope (JEOL, Nieuw-Vennep, The Netherlands), a GEMINI-LEO1530 VP FE-SEM emission electron microscope (Carl Zeiss, Cambridge, UK), and a Tecnai F20 Super Twin TMP transmission electron microscope (FEI, Hillsboro, USA).
Dynamic light scattering (DLS) was carried out by means of a NanoPlus-3 zeta/nanoparticle analyzer (Micromeritics, Norcross, USA) at 25 • C in a glass cell (capacity of 0.9 mL), which was used to obtain the mean particle size of nanoparticles dispersed in a fluid, which hydrate and interact with each other. The mean particle size was calculated from the diffusional properties of the particle indicating the size of the hydrated and solvated particle. For this purpose, a nanoparticle solution, 10 mg L −1 , was dispersed in water or ethanol and sonicated for 6 h before analysis. Particles suspended in a liquid have a Brownian motion due to the random collisions with solvent molecules. This motion causes the particles to diffuse through the medium. The diffusion coefficient, D, is inversely proportional to the particle size or hydrodynamic diameter, d, according to the Stokes-Einstein equation: where, k B is Boltzmann's constant, T is the absolute temperature, and η is the viscosity. All materials were characterized by N 2 and CO 2 adsorption at −196 • C and 0 • C, respectively, using a 3-Flex manometric adsorption equipment (Micromeritics, Norcross, USA). The total adsorbed volume (V 0.95 ) was taken as the physiosorbed volume of N 2 at a relative pressure P/P 0 = 0.95. The Brunauer-Emmett-Teller (BET) model was applied to obtain the BET area (A BET ). Micropore volume (V mic ), average pore size (L 0 ), and CO 2 adsorption energy (E ads -CO2 ) were obtained by application of the Dubinin-Radushkevich equation. The mesopore volume (V meso ) was obtained through the Barrett-Joyner-Halenda (BJH) model.

Chemical Composition of Nanomaterials and Sandstone
The chemical characterization was carried out by carbon, hydrogen, oxygen and nitrogen (CHON) analysis for nanomaterials, and by Fourier transform infrared spectroscopy (FTIR) for sandstone. An IRAffinity-1S FTIR spectrometer (Shimadzu, Columbia, USA)was operated at room temperature using potassium bromide in a KBr-to-material ratio of 30:1 (% w/w). The impregnation percentages of nanoparticles on sandstone were corroborated by thermogravimetric analysis (TGA) (TA Instruments, New Castle, USA). For this, the sample was burned under an air atmosphere at 10 • C min −1 up to 800 • C.

Rheological Analysis of CN.LYS Synthesis Solutions
The stability of the synthesis solutions was evaluated at different concentrations of CN.LYS (CN-LYS1, CN.LYS2, and CN.LYS3). For this purpose, a Kinexus Ultra+ rheometer (Malvern Panalytical, Malvern, UK) utilizing a concentric cylinder sensor equipped with a Peltier cell for temperature control was used. The tests were first carried out by varying the shear rate from 1 to 250 s −1 in order to define the adequate shear rate (50 s −1 ). The test conditions were carried out to mimic the real synthesis conditions in terms of temperature and stirring. The process started at 60 • C for 1 h, after which the temperature was controlled to simulate the natural cooling process at 57, 45, 37, and 34 • C until the viscosity reached a constant value. Only 30 mL were needed for the test, while the current reaction was carried out in 1.8 L.

Dispersion of Nanoparticles in Solution
The electrophoretic light scattering (ELS) technique was used to evaluate the surface charge of the particles and their dispersion stability at 25 • C in a NanoPlus-3 zeta/nanoparticle analyzer (Micromeritics, Norcross, USA). In this test, several nanoparticle suspensions were prepared at 10 mg L −1 , with a pH adjusted between 2 to 12 by adding solutions of 0.1 mol L −1 HCl or 0.01 mol L −1 NaOH, and then subjected to analysis. The zeta potential was calculated using the Smoluchowski equation, derived from the calculation of the Doppler effect.
where ζ is the zeta potential, η is the viscosity of the fluid (water), U is the electrophoretic mobility, ε is the permittivity, V represents the speed of movement of the particles, E is the electric field, ∆ν is the Doppler effect, n is the index of refraction, θ is the angle of detection, and λ is the wavelength of the incident light.

Adsorption Tests at High Pressure
The adsorption tests carried out below atmospheric pressure were described in sub-Section 2.3.2. At high pressure, the CO 2 adsorption capacity was evaluated in two different conditions: (i) under pure CO 2 in a manometric device (up to 3.0 MPa) and (ii) under a CO 2 /N 2 flow in a gravimetric device (up to 2.6 MPa). The details of each protocol are presented below.
2.4.1. Adsorption at High Pressure for Pure CO 2 -Manometric Device The carbon nanospheres (CN.LYS2), sandstone, and impregnated sandstone (at mass fractions of 10 and 20%) were investigated in High Pressure Volume Analyzer, HPVAII-200 (Micromeritics, Norcross, USA) at 0 • C, 25 • C, and 50 • C and at pressures from 3 × 10 −3 up to 3.0 MPa. In order to have enough total surface area for adsorption and to minimize measurement errors, the amount of each material inside the sample holder was around 0.5 g for nanoparticles, 1.5 g for impregnated sandstone, and 14 g for sandstone. The contribution of the empty sample holder was systematically measured and subtracted to all data to improve accuracy. The isosteric heat of adsorption, Q ST , was calculated using the isosteric method with the Microactive software (from Micromeritics, Norcross, USA) from three adsorption isotherms at 0, 25, and 50 • C, based on the Clausius-Clapeyron equation [51]: where R is the universal gas constant (8.314 J mol −1 K −1 ), P is the absolute pressure, and T is the temperature. The excess adsorbed CO 2 amount (N exc , g CO2 .g adsorbent −1 ) was equal to the absolute adsorbed CO 2 amount (N ads, g CO2 .g adsorbent −1 ) minus the product of gas density in the bulk phase by the volume of the adsorbed phase. The values provided by the HPVA device were obtained on an excess basis, and therefore, the absolute amounts had to be determined as follows [52]: where M CO2 is the molecular weight of CO 2 (44.013 g mol −1 ), Z is the compressibility factor at the considered pressure and temperature, and ρ liq is the density of liquid CO 2 (1032 × 10 3 g m −3 ). The isotherms were fitted with the Sips and Toth models, which take into account multilayer adsorption. Table 1 presents the equations for each model [53][54][55]. K S and K T represent adsorption equilibrium constants for the Sips and Toth models, respectively, and the n and t parameters indicate the heterogeneity of the system for the Sips and Toth models, respectively. The heterogeneity may originate from the solid structure, from the solid energy properties, or from the adsorbate [54]. The n or t parameters are usually greater than unity, and when they are the unit, the models assume the Langmuir equation [54].

Sips (7)
N ads (mmol g −1 ) is the adsorbed amount, N m (mmol g −1 ) is the adsorption capacity at equilibrium, P (kPa) is the equilibrium pressure, and K S and n are the Sips adsorption equilibrium constants, related to the affinity and the heterogeneity of the system, respectively.

Toth (8)
N ads , N m , and P have the same meaning as above, and K T and t are the Toth adsorption equilibrium constants, related to the affinity and the heterogeneity of the system, respectively.

Adsorption at High Pressure for CO 2 and N 2 -Gravimetric Device
The CN.LYS2, sandstone, and impregnated sandstone CO 2 isotherms (with a mass fraction of 20% of nanoparticles) were investigated using a HP TGA 750 thermogravimetric analyzer (TA Instruments, New Castle, USA) at 50 • C and high pressure from 0.03 to 3.0 MPa for CO 2 and N 2 . This device was equipped with a magnetic levitation top-loading balance, which made it possible to achieve high accuracy and reduce the volume of the system. The amount of each material put inside the sample holder was around 15 mg for nanoparticles, 40 mg for sandstone, and 40 mg for impregnated sandstone, to have enough total surface area for adsorption. The contribution of the buoyancy effect was manually subtracted from the data using blank tests carried out in the same conditions but with an empty sample holder. From the adsorption results of each N 2 and CO 2 isotherm, it was possible to predict the selectivity by applying the ideal adsorbed solution theory (IAST), which allows estimating the competitive adsorption of the compounds in a mixture of gases as the flue gas. Based on the literature, a model flue gas comprising 80% N 2 and 20% CO 2 was selected. The IAST was implemented in a Python routine (package-pyIAST from Simon et al. [56]).

Results and Discussion
The results are divided into two main sections: (a) materials characteristics (nanoparticles and sandstone) and (b) study of the interaction between CO 2 /nanoparticles/sandstone by adsorption isotherms under different operation conditions (T, P).

Materials Characteristics
The morphology of the carbon materials obtained from melamine and carbon tetrachloride was very heterogeneous. Figure 2a,b presents SEM images of two different zones wherein more or less agglomerated nanospheres can be observed (Figure 2b). When the images are observed at lower magnification, it can be noticed that some areas have fiber and block morphologies (Figure 2c), while other areas have nanospheres/microspheres morphologies ( Figure 2d).
After coating with resorcinol/formaldehyde, the hydrodynamic diameter of the CN.MEL particles was higher than the limit of detection of the equipment (10 µm). For the e-CCS process, this material might, thus, induce technical problems, due to the possible obstruction of the naturally porous structure of the rock. The ultimate analysis (Table 2) shows that this material had a nitrogen content close to 50% before coating and carbonization (Gel.MEL1), but of only 9.2% after coating with resorcinol/formaldehyde (Gel.MEL2) and 2.2% after carbonization (CN.MEL). Therefore, CN.LYS and CN.MEL materials exhibited similar nitrogen contents. Some N-rich carbon materials, reported in the literature, have nitrogen content close to those obtained in this work [16]. The oxygen content was measured independently from carbon, hydrogen and nitrogen content.

Materials Characteristics
The morphology of the carbon materials obtained from melamine and carbon tetrachloride was very heterogeneous. Figure 2a,b presents SEM images of two different zones wherein more or less agglomerated nanospheres can be observed (Figure 2b). When the images are observed at lower magnification, it can be noticed that some areas have fiber and block morphologies (Figure 2c After coating with resorcinol/formaldehyde, the hydrodynamic diameter of the CN.MEL particles was higher than the limit of detection of the equipment (10 μm). For the e-CCS process, this material might, thus, induce technical problems, due to the possible obstruction of the naturally porous structure of the rock. The ultimate analysis (Table 2) shows that this material had a nitrogen content close to 50% before coating and carbonization (Gel.MEL1), but of only 9.2% after coating with 500 nm 500 nm 5 μm 5 μm

C (Mass Fraction %) H (Mass Fraction %) N (Mass Fraction %) O (Mass Fraction %)
Gel The Gel.MEL1 (49.1% of nitrogen) was submitted to CO 2 and N 2 adsorption (at 0 • C and −196 • C, respectively) but it was impossible to obtain the corresponding isotherms, possibly because of its too low surface area. It is well known that the pyrolysis step induces a significant increase of narrow porosity by elimination of volatile species. However, when the Gel.MEL1 is directly pyrolyzed, most of the material undergoes decompositions and the yield is very low (< 5%).
The adsorption and desorption isotherms (N 2 at −196 • C and CO 2 at 0 • C) for nanomaterials synthesized with melamine (CN.MEL) and L-lysine (CN.LYS) are presented in Figure 3, and the textural parameters obtained from adsorption isotherms are presented in Table 3.
The micropore and the mesopore fractions of the CN.MEL and CN.LYS2 materials were similar (67-64% and 33-36%, respectively) although their total pore volumes (at P/P 0 = 0.95) were rather different. Both CN.LYS1 and CN.LYS3 exhibited higher micropore fractions (72.7 and 92.3%, respectively), but had lower values of A BET and total pore volumes than those of CN.MEL and CN.LYS2 materials ( Figure 3). The mesoporous volumen of CN.MEL is evidenced by the hysteresis loop in the range of relative pressures between 0.46 and 0.66. In the CN.LYS series, only CN.LYS2 was also mesoporous, but with a different structure as deduced from the different shape of the hysteresis loop, occurring at higher relative pressure. The CN.LYS1 texture was moderately mesoporous (28.6%). After the first dilution, the CN.LYS2 texture was a little more mesoporous (36.6%), but after the third dilution, CN.LYS3 had a predominantly microporous texture as evidenced by the type Ia of its N 2 isotherm. The adsorption capacity of CO 2 at 0 • C was as expected according to the range reported in the literature, but the synthesis process reported in this study was easier than many others reported so far [16,43,57]. The experimental development considered important parameters for a possible industrial application, such as operation at atmospheric pressure, relatively low temperature (60 • C), and relatively low time (1 h) before carbonization. The carbon nanospheres are usually synthesized by methods that demand greater energy and time [29]. by the type Ia of its N2 isotherm. The adsorption capacity of CO2 at 0 °C was as expected according to the range reported in the literature, but the synthesis process reported in this study was easier than many others reported so far [16,43,57]. The experimental development considered important parameters for a possible industrial application, such as operation at atmospheric pressure, relatively low temperature (60 °C), and relatively low time (1 hour) before carbonization. The carbon nanospheres are usually synthesized by methods that demand greater energy and time [29].
In order to provide explanations of the aforementioned trends, Figure 4 presents TEM pictures of Gel.LYS (materials before carbonization) and CN.LYS (materials after carbonization).  Figure 4a,b present CN.LYS1 before and after carbonization, respectively. Here a mixture of larger and smaller nanospheres can be seen. Figure 4c,d present smaller particles (approximately 50 nm) for CN.LYS2 before and after carbonization. These particles are more transparent than those of the CN.LYS1 and CN.LYS3 materials, due to the more mesoporous texture of the CN.LYS2 material. Figure 4e,f present CN.LYS3 before and after carbonization. It can be seen that a gel was formed around the spheres. The L-lysine acts as a catalyst in the reaction, and therefore, if its amount is limited for the third dilution process, it might be the reason for the formation of a gel coating the spheres instead of producing more nanospheres, which obstructs the porous structure of CN.LYS3. This would affect the results presented in Figure 3 and Table 3 for CN.LYS3. The reproducibility for CN.LYS is significant, the variation of size and CO2 capacity is less than 1%.    In order to provide explanations of the aforementioned trends, Figure 4 presents TEM pictures of Gel.LYS (materials before carbonization) and CN.LYS (materials after carbonization). Figure 4a,b present CN.LYS1 before and after carbonization, respectively. Here a mixture of larger and smaller nanospheres can be seen. Figure 4c,d present smaller particles (approximately 50 nm) for CN.LYS2 before and after carbonization. These particles are more transparent than those of the CN.LYS1 and CN.LYS3 materials, due to the more mesoporous texture of the CN.LYS2 material. Figure 4e,f present CN.LYS3 before and after carbonization. It can be seen that a gel was formed around the spheres. The L-lysine acts as a catalyst in the reaction, and therefore, if its amount is limited for the third dilution process, it might be the reason for the formation of a gel coating the spheres instead of producing more nanospheres, which obstructs the porous structure of CN.LYS3. This would affect the results presented in Figure 3 and Table 3 for CN.LYS3. The reproducibility for CN.LYS is significant, the variation of size and CO 2 capacity is less than 1%. Figure 4a,b present CN.LYS1 before and after carbonization, respectively. Here a mixture of larger and smaller nanospheres can be seen. Figure 4c,d present smaller particles (approximately 50 nm) for CN.LYS2 before and after carbonization. These particles are more transparent than those of the CN.LYS1 and CN.LYS3 materials, due to the more mesoporous texture of the CN.LYS2 material. Figure 4e,f present CN.LYS3 before and after carbonization. It can be seen that a gel was formed around the spheres. The L-lysine acts as a catalyst in the reaction, and therefore, if its amount is limited for the third dilution process, it might be the reason for the formation of a gel coating the spheres instead of producing more nanospheres, which obstructs the porous structure of CN.LYS3. This would affect the results presented in Figure 3 and Table 3 for CN.LYS3. The reproducibility for CN.LYS is significant, the variation of size and CO2 capacity is less than 1%.  Figure 5 presents the rheological behavior of the synthesis solution/colloid/suspension (Sol.LYS1 without dilution, Sol.LYS2 for dilution 1, and Sol.LYS3 for dilution 2). The changes in temperature and the stirring were chosen to reproduce the conditions of synthesis (60 °C for 1 h and after cooling at room temperature). The cooling stage was controlled in the rheometer. The viscosity has changed during synthesis because the reactive system underwent polymerization and condensation-related changes to form the nanospheres, so that the system passed from the solution to the colloid and the suspension. The differences in viscosity were related to the concentration of reagents in the solutions. The third solution, Sol.LYS3, presented fewer changes during the synthesis process because the system had a lower concentration of reagents as presented in Figure 4.
Based on the particle size analysis, porous texture, nitrogen content, and adsorbed amount of CO2 (at 0 °C and up to 1 bar), the CN.LYS material was selected to perform the adsorption tests in conditions closer to those of the reservoir. Despite the high CO2 adsorption capacities exhibited by the CN.MEL material, its size and shape would not allow its application in real reservoir conditions. Besides, its synthesis process was more complex, with a higher number of steps and a higher consumption of energy and time, which were not compensated by significantly better physicochemical properties regarding the CN.LYS.   Figure 5 presents the rheological behavior of the synthesis solution/colloid/suspension (Sol.LYS1 without dilution, Sol.LYS2 for dilution 1, and Sol.LYS3 for dilution 2). The changes in temperature and the stirring were chosen to reproduce the conditions of synthesis (60 • C for 1 h and after cooling at room temperature). The cooling stage was controlled in the rheometer. The viscosity has changed during synthesis because the reactive system underwent polymerization and condensation-related changes to form the nanospheres, so that the system passed from the solution to the colloid and the suspension. The differences in viscosity were related to the concentration of reagents in the solutions. The third solution, Sol.LYS3, presented fewer changes during the synthesis process because the system had a lower concentration of reagents as presented in Figure 4.
Based on the particle size analysis, porous texture, nitrogen content, and adsorbed amount of CO 2 (at 0 • C and up to 1 bar), the CN.LYS material was selected to perform the adsorption tests in conditions closer to those of the reservoir. Despite the high CO 2 adsorption capacities exhibited by the CN.MEL material, its size and shape would not allow its application in real reservoir conditions. Besides, its synthesis process was more complex, with a higher number of steps and a higher consumption of energy and time, which were not compensated by significantly better physicochemical properties regarding the CN.LYS.
The third solution, Sol.LYS3, presented fewer changes during the synthesis process because the system had a lower concentration of reagents as presented in Figure 4.
Based on the particle size analysis, porous texture, nitrogen content, and adsorbed amount of CO2 (at 0 °C and up to 1 bar), the CN.LYS material was selected to perform the adsorption tests in conditions closer to those of the reservoir. Despite the high CO2 adsorption capacities exhibited by the CN.MEL material, its size and shape would not allow its application in real reservoir conditions. Besides, its synthesis process was more complex, with a higher number of steps and a higher consumption of energy and time, which were not compensated by significantly better physicochemical properties regarding the CN.LYS. To analyze the behavior of the nanoparticles in aqueous medium, Table 4 presents the mean particle size of nanomaterials (CN.LYS) and Figure 6 presents their zeta potentials. According to the Stokes-Einstein equation, the diffusion coefficient is inversely proportional to particle size or hydrodynamic diameter; therefore, it is possible to analyze whether the nanoparticles could interact to form aggregates. Precipitation is only possible if the aggregates are big enough. Another important concept is zeta potential; if zeta potential is high (negative or positive), the particles are stable due to the high electrostatic repulsion between them. On the contrary, a low zeta potential (approaching zero) increases the probability of particles colliding, and thus forming aggregates. The hydrodynamic diameter was calculated for nanoparticles in water (at pH 5.8) and ethanol (at pH 7). Aggregate size was less in ethanol (Table 4) because the pH affects the behavior in solution ( Figure 6). However, the results for nanoparticles suspended in water was close those obtained for nanoparticles suspended in ethanol. For an industrial application and injection into the porous medium, the most economical way is suspension in water. For CN.LYS1 and CN.LYS3, a pH higher than 7 was better for rocks impregnation because the zeta potential was farther from zero. For CN.LYS2, a pH higher than 7 or lower than 4.7 was better for rocks impregnation. Gel.LYS2 presented the highest values of zeta potential at pH below 4, increasing the natural precipitation time of nanoparticles after the synthesis process. Figure 7 presents SEM images of the sandstone surface before (Figure 7a) and after (Figure 7b,c) the impregnation step. Impregnation was achieved in water because of its lower cost and non-hazardous nature, making it ideal for industrial applications. The distribution of CN.LYS2 particles was homogeneously distributed on the surface (Figure 7b). The size of the aggregates was between 100 and 200 nm (Figure 7b). After one year, the sandstone continued to be impregnated, without showing any disintegration of nanoparticles from the sandstone surface. By thermogravimetric analysis, variations of less than 5% of the percentage of impregnation were obtained. This can also be related to the impregnation method without stirring, which might produce zones of lower nanoparticle concentration. For CN.LYS1 and CN.LYS3, a pH higher than 7 was better for rocks impregnation because the zeta potential was farther from zero. For CN.LYS2, a pH higher than 7 or lower than 4.7 was better for rocks impregnation. Gel.LYS2 presented the highest values of zeta potential at pH below 4, increasing the natural precipitation time of nanoparticles after the synthesis process.   showing any disintegration of nanoparticles from the sandstone surface. By thermogravimetric analysis, variations of less than 5% of the percentage of impregnation were obtained. This can also be related to the impregnation method without stirring, which might produce zones of lower nanoparticle concentration. The sandstone presented an ABET of 0.4 m 2 g -1 , and its CO2 adsorption capacity could not be measured using conventional methods (< 0.0013 mmol g -1 at 0 °C and atmospheric pressure). Sandstone is mainly composed of silica, which has an acidic character as the CO2 molecule. Consequently, if the specific area of the sandstone is low, its CO2 adsorption capacity is even lower than that which might be expected for this specific area.
The sandstone was impregnated at a low nanoparticle concentration to evaluate its economic feasibility at the industrial level. The sandstone impregnated with mass fractions of 0.1 and 0.01% did not show a significant increase in its surface properties, unlike samples with higher mass fractions as shown in Table 5. The textural parameters of the impregnated sandstones were indeed improved as the percentage of nanoparticles on their surface increased. At a mass fraction of 20%, ABET and V0,95 increased by factors as high as 225 and 670, respectively.

Vmic-N2 (cm 3 g -1 )
Vmic-CO2 (cm 3 g -1 ) The sandstone presented an A BET of 0.4 m 2 g −1 , and its CO 2 adsorption capacity could not be measured using conventional methods (<0.0013 mmol g −1 at 0 • C and atmospheric pressure). Sandstone is mainly composed of silica, which has an acidic character as the CO 2 molecule. Consequently, if the specific area of the sandstone is low, its CO 2 adsorption capacity is even lower than that which might be expected for this specific area.

Vmes (cm
The sandstone was impregnated at a low nanoparticle concentration to evaluate its economic feasibility at the industrial level. The sandstone impregnated with mass fractions of 0.1 and 0.01% did not show a significant increase in its surface properties, unlike samples with higher mass fractions as shown in Table 5. The textural parameters of the impregnated sandstones were indeed improved as the percentage of nanoparticles on their surface increased. At a mass fraction of 20%, A BET and V 0.95 increased by factors as high as 225 and 670, respectively. Figure 8 shows adsorption isotherms of CO 2 at atmospheric pressure and 0 • C, for raw sandstone and sandstone impregnated at mass fractions of 0.01, 0.1, 1, 5, 10, and 20% of CN.LYS2. In addition, it shows the slope changes related to the affinity between the adsorbent medium and the adsorbate. The materials did not have significant affinity with CO 2 at mass fractions of 0.01 and 0.1% and without CN.LYS2. The affinity and adsorption capacity increased with the percentage of nanoparticles in the system. The latter presented a different behavior above a mass fraction of 1%. Indeed, at a mass fraction of 1%, the adsorption capacity was increased by a factor 21 with respect to raw sandstone, although the value was still low, 0.03 mmol g −1 . At a mass fraction of 20%, the value was far higher, 0.63 mmol g −1 , corresponding to an increment factor of 499. Different materials have been reported in the literature [16,43] with specific surface modifications to increase the adsorption capacity of CO 2 , but the value of adsorption capacity was similar to that of sandstone by adding a mass fraction of 10 or 20% of nanoparticles. The effect of the impregnation method on the CO2 adsorption capacity at atmospheric pressure and 0°C was then evaluated. As explained before, immersion and soaking were achieved using two different sets of conditions: (i) 6 h and 600 rpm and (ii) 24 h without stirring. Increasing the soaking time by 18 h improved the adsorption capacity by more than 25% ( Table 6). The values presented in Figure 8, adsorption isotherms of CO2 at atmospheric pressure and 0 °C, thus correspond to Conditions 2. To evaluate the possible synergistic behavior between NC.LYS2 and sandstone, the theoretical and experimental values of the CO2 adsorption capacity are presented in Table 7. Theoretical values were calculated by assuming a linear relationship and taking into account the CO2 adsorption The effect of the impregnation method on the CO 2 adsorption capacity at atmospheric pressure and 0 • C was then evaluated. As explained before, immersion and soaking were achieved using two different sets of conditions: (i) 6 h and 600 rpm and (ii) 24 h without stirring. Increasing the soaking time by 18 h improved the adsorption capacity by more than 25% ( Table 6). The values presented in Figure 8, adsorption isotherms of CO 2 at atmospheric pressure and 0 • C, thus correspond to Conditions 2. To evaluate the possible synergistic behavior between NC.LYS2 and sandstone, the theoretical and experimental values of the CO 2 adsorption capacity are presented in Table 7. Theoretical values were calculated by assuming a linear relationship and taking into account the CO 2 adsorption capacities and the percentages of each solid. The difference between theoretical and experimental values ranged from 5 to 10%, which corresponds to the experimental error given the inaccuracy in the measurement of the very low CO 2 adsorption capacity of the sandstone. The differences could also be related to the segregation of nanoparticles during the impregnation process, the nanoparticles not being homogeneously distributed on the surface of the sandstone.

Pure CO 2 Adsorption at High Pressure-Manometric Measurement Method
The e-CCS process requires evaluating the behavior of the materials at high pressure (up to 3 MPa) and at the temperature of a hypothetical reservoir (50 • C). Figure 9a-c present the absolute CO 2 adsorbed amount and the excess amount for CN.LYS2 at 0, 25, and 50 • C. The difference between excess and absolute amounts appeared above 1 MPa, and represented 8.3% at 0 • C, 6.7% at 25 • C, and 5.9% at 50 • C. This difference was lower when the temperature increased. Such a trend is consistent with the fact that, at similar pressure, the density of the bulk phase decreases when the temperature increases. The high-pressure intrinsic CO 2 adsorption capacity of sandstone without impregnation had a negligible effect on the measurement. Therefore, the evaluation was only carried out for sandstone impregnated at mass fractions of 10% and 20% (Figure 9d,e). The difference between excess and absolute adsorbed amounts of impregnated sandstone was similar to that of CN.LYS2, stranded out after 1 MPa, and represented less than 10%, on average. The e-CCS process requires evaluating the behavior of the materials at high pressure (up to 3 MPa) and at the temperature of a hypothetical reservoir (50 °C). Figure 9a-c present the absolute CO2 adsorbed amount and the excess amount for CN.LYS2 at 0, 25, and 50 °C. The difference between excess and absolute amounts appeared above 1 MPa, and represented 8.3% at 0 °C, 6.7% at 25 °C, and 5.9% at 50 °C. This difference was lower when the temperature increased. Such a trend is consistent with the fact that, at similar pressure, the density of the bulk phase decreases when the temperature increases. The high-pressure intrinsic CO2 adsorption capacity of sandstone without impregnation had a negligible effect on the measurement. Therefore, the evaluation was only carried out for sandstone impregnated at mass fractions of 10% and 20% (Figure 9d,e). The difference between excess and absolute adsorbed amounts of impregnated sandstone was similar to that of CN.LYS2, stranded out after 1 MPa, and represented less than 10%, on average.
To observe the pressure effect on the adsorption capacity of CN.LYS2, the Nads at atmospheric pressure and 0 °C (3.48 mmol g -1 ) was compared to Nads at 3 MPa and 0 °C (5.80 mmol g -1 ). The increase of pressure produced an increase of Nads of 66.6%, indicating physisorption as the main adsorption mechanism. At 50 °C, as expected, the adsorption capacity decreased by 20% due to the exothermic character of adsorption. The obtained adsorption capacity is competitive compared to other results reported for nanomaterials under similar conditions [43,44,58]. The effect of pressure on the impregnated sandstone was also observed by comparing Nads at atmospheric pressure and 0 °C (0. Figure 9e presents incremental factors comparing Nads for sandstone without impregnation to Nads after impregnation with mass fractions of 10 and 20% at 3 MPa and 0, 25, and 50 °C. In the conditions of a shallow reservoir (50 °C and 3.0 MPa), the incremental factor was 677 for SS-20. For the e-CCS process, the pressure could increase up 6.0 MPa so that CO2 is still in vapor phase, which allows a higher adsorbed amount. Appendix A presents the fit of the Sips and Toth models to the CO2 adsorption isotherms at high pressure and at 0, 25, and 50°C for CN.LYS2, SS-10, and SS-20 ( Figures A-1, A-2, and A-3, respectively). As expected, the Toth and Sips models led to very good fits: R 2 > 0.99 for CN.LYS2, SS-10, and SS-20, and R 2 > 0.75 for SS-10 at 0 °C. In the latter case, the concavity of the isotherm was mainly due to the very rapid increase of the bulk density as a function of pressure with respect to the increase of the density of the adsorbed phase at pressure higher than 1.5 MPa. Figure 10a presents the isosteric heat of adsorption (Qst) of CN.LYS2 and of sandstone impregnated with mass fractions of 10 and 20%, as a function of Nads expressed in mmol per gram of total adsorbent material. Nads for SS-10 and SS-20 was small, and thus, Figure 10b presents Qst as a function of Nads in mmol per gram of carbon adsorbent material. The values of Qst varied from 25 to 33 kJ mol -1 , which indicates a strong interaction in the adsorption system (high affinity). The interactions with the carbon porous structure could be increased by nitrogen-containing groups that are present onto the carbon surface. Nitrogen groups can indeed promote interactions between CO2 and the substrate. To observe the pressure effect on the adsorption capacity of CN.LYS2, the N ads at atmospheric pressure and 0 • C (3.48 mmol g −1 ) was compared to N ads at 3 MPa and 0 • C (5.80 mmol g −1 ). The increase of pressure produced an increase of N ads of 66.6%, indicating physisorption as the main adsorption mechanism. At 50 • C, as expected, the adsorption capacity decreased by 20% due to the exothermic character of adsorption. The obtained adsorption capacity is competitive compared to other results reported for nanomaterials under similar conditions [43,44,58]. The effect of pressure on the impregnated sandstone was also observed by comparing N ads at atmospheric pressure and 0 • C (0.34 mmol g −1 for SS-10 and 0.63 mmol g −1 for SS -20) to N ads at 3 MPa and 0 • C (0.47 mmol g −1 for SS-10 and 1.04 mmol g −1 for SS-20). The corresponding increases were 38.2% (SS-10) and 66.0% (SS-20), respectively. The maximum N ads under reservoir conditions was 0.85 mmol g −1 at a mass fraction of 20% of CN.LYS2. Figure 9e presents incremental factors comparing N ads for sandstone without impregnation to N ads after impregnation with mass fractions of 10 and 20% at 3 MPa and 0, 25, and 50 • C. In the conditions of a shallow reservoir (50 • C and 3.0 MPa), the incremental factor was 677 for SS-20. For the e-CCS process, the pressure could increase up 6.0 MPa so that CO 2 is still in vapor phase, which allows a higher adsorbed amount. Appendix A presents the fit of the Sips and Toth models to the CO 2 adsorption isotherms at high pressure and at 0, 25, and 50 • C for CN.LYS2, SS-10, and SS-20 ( Figures A1-A3, respectively). As expected, the Toth and Sips models led to very good fits: R 2 > 0.99 for CN.LYS2, SS-10, and SS-20, and R 2 > 0.75 for SS-10 at 0 • C. In the latter case, the concavity of the isotherm was mainly due to the very rapid increase of the bulk density as a function of pressure with respect to the increase of the density of the adsorbed phase at pressure higher than 1.5 MPa. Figure 10a presents the isosteric heat of adsorption (Q st ) of CN.LYS2 and of sandstone impregnated with mass fractions of 10 and 20%, as a function of N ads expressed in mmol per gram of total adsorbent material. N ads for SS-10 and SS-20 was small, and thus, Figure 10b presents Q st as a function of N ads in mmol per gram of carbon adsorbent material. The values of Q st varied from 25 to 33 kJ mol −1 , which indicates a strong interaction in the adsorption system (high affinity). The interactions with the carbon porous structure could be increased by nitrogen-containing groups that are present onto the carbon surface. Nitrogen groups can indeed promote interactions between CO 2 and the substrate.  The isosteric heat of adsorption of CN.LYS2 presented two distinct behaviors, whether Nads was either (i) lower or (ii) higher than 3 mmol gcarbonadsorbentmaterial -1 . For case (i), Qst decreased with Nads due to the increasing distance between the last adsorbed layer and the carbon surface, thus decreasing the molecular interactions (Figure 10b). For case (ii), i.e., when Nads was higher than 3 mmol gcarbonadsorbentmaterial -1 , Qst increased due to the increased interactions between adsorbate molecules [61]. For SS-10 and SS-20, the value of Qst also increased with Nads. Because the interactions between sandstone (silica) and CO2 were weak and because the contribution of the surface area associated to the carbon material was lower than for CN.LYS2 alone, it can be assumed again that the interactions between the adsorbed gas layers prevailed [61].

CO2 and N2 Adsorption at High Pressure-Gravimetric Measurement Method
Using the PyIAST application and using the pure-components adsorption isotherms obtained by experimental tests, it is possible to characterize the behavior of each component and to predict the adsorbed amount of each component present in a mixture [56]. It is, thus, possible to obtain the adsorbed amount at a constant temperature by varying the concentration of the components (at constant pressure) or the pressure (at constant concentrations). Initially, it is necessary to use the "isothermal interpolator" to generate data points that follow a given isothermal model, avoiding the search for an appropriate analytical model and examining the quality of its fit to the data (i.e., Langmuir, Freundlich or Toth, among others) [56]. The isothermal interpolator is a tool included in the PyIAST package. After that, PyIAST takes the interpolated data for its calculations. Calculations are done using a predesigned routine [56].
In the present case, the experimental data of each pure component (CO2 and N2) were obtained by HP-TGA at 50 °C, between 0.1 and 2.5 MPa, and with a flow of CO2 or N2 (50 mL min -1 up to 1.0 MPa and 70 mL min -1 up to 2.5 MPa). The prediction was calculated at 50 °C by varying: 1) the system pressure from 0.1 to 2.5 MPa at constant CO2 concentration of 20% and 2) the concentration of CO2, from 5 to 100% at a constant pressure of 2.5 MPa. The evaluated materials were CN.LYS2 and sandstone from a real reservoir (RS) impregnated with a mass fraction of 20% of CN.LYS2. It is important to mention that it might be possible to obtain considerable adsorbed quantities for a cleaning and adsorption balance of more than 12-24 hours, because the nanomaterials (main adsorbent) are micro/mesoporous. A longer cleaning time thus allows eliminating adsorbed gases and moisture. In the same way, an adequate equilibrium time allows for greater diffusion of the gas The isosteric heat of adsorption of CN.LYS2 presented two distinct behaviors, whether N ads was either (i) lower or (ii) higher than 3 mmol g carbonadsorbentmaterial −1 . For case (i), Q st decreased with N ads due to the increasing distance between the last adsorbed layer and the carbon surface, thus decreasing the molecular interactions (Figure 10b). For case (ii), i.e., when N ads was higher than 3 mmol g carbonadsorbentmaterial −1 , Q st increased due to the increased interactions between adsorbate molecules [61]. For SS-10 and SS-20, the value of Q st also increased with N ads . Because the interactions between sandstone (silica) and CO 2 were weak and because the contribution of the surface area associated to the carbon material was lower than for CN.LYS2 alone, it can be assumed again that the interactions between the adsorbed gas layers prevailed [61].

CO 2 and N 2 Adsorption at High Pressure-Gravimetric Measurement Method
Using the PyIAST application and using the pure-components adsorption isotherms obtained by experimental tests, it is possible to characterize the behavior of each component and to predict the adsorbed amount of each component present in a mixture [56]. It is, thus, possible to obtain the adsorbed amount at a constant temperature by varying the concentration of the components (at constant pressure) or the pressure (at constant concentrations). Initially, it is necessary to use the "isothermal interpolator" to generate data points that follow a given isothermal model, avoiding the search for an appropriate analytical model and examining the quality of its fit to the data (i.e., Langmuir, Freundlich or Toth, among others) [56]. The isothermal interpolator is a tool included in the PyIAST package. After that, PyIAST takes the interpolated data for its calculations. Calculations are done using a predesigned routine [56].
In the present case, the experimental data of each pure component (CO 2 and N 2 ) were obtained by HP-TGA at 50 • C, between 0.1 and 2.5 MPa, and with a flow of CO 2 or N 2 (50 mL min −1 up to 1.0 MPa and 70 mL min −1 up to 2.5 MPa). The prediction was calculated at 50 • C by varying: (1) the system pressure from 0.1 to 2.5 MPa at constant CO 2 concentration of 20% and (2) the concentration of CO 2 , from 5 to 100% at a constant pressure of 2.5 MPa. The evaluated materials were CN.LYS2 and sandstone from a real reservoir (RS) impregnated with a mass fraction of 20% of CN.LYS2. It is important to mention that it might be possible to obtain considerable adsorbed quantities for a cleaning and adsorption balance of more than 12-24 h, because the nanomaterials (main adsorbent) are micro/mesoporous. A longer cleaning time thus allows eliminating adsorbed gases and moisture. In the same way, an adequate equilibrium time allows for greater diffusion of the gas into the porous structure and greater interactions with the material, which would allow a higher adsorbed amount. Therefore, to analyze the selectivity, a shorter time was used (cleaning and adsorption equilibrium time of 2 h at each pressure).
It was not possible to obtain the CO 2 isotherm for RS because the latter had a too low surface area, so the maximum value for RS at 50 • C and 2.5 MPa was 0.033 mmol g −1 . For RS, the adsorbed amount of N 2 was 3.19 mmol g −1 at 50 • C and 2.5 MPa. Figure 11a,b present the CO 2 and N 2 adsorption isotherms under continuous gas flow for CN.LYS2 and RS-20. The increment factor of N ads for RS-20 with respect to RS was 19 (0.66 mmol g −1 at 50 • C and 2.5 MPa). In addition, the theoretical value of N ads for RS with a mass fraction of 20% of CN.LYS was 0.78 mmol g −1 but the experimental value was 0.66 mmol g −1 (i.e., 85% of the theoretical one). into the porous structure and greater interactions with the material, which would allow a higher adsorbed amount. Therefore, to analyze the selectivity, a shorter time was used (cleaning and adsorption equilibrium time of 2 hours at each pressure). It was not possible to obtain the CO2 isotherm for RS because the latter had a too low surface area, so the maximum value for RS at 50 °C and 2.5 MPa was 0.033 mmol g -1 . For RS, the adsorbed amount of N2 was 3.19 mmol g -1 at 50 °C and 2.5 MPa. Figure 11a and b present the CO2 and N2 adsorption isotherms under continuous gas flow for CN.LYS2 and RS-20. The increment factor of Nads for RS-20 with respect to RS was 19 (0.66 mmol g -1 at 50 °C and 2.5 MPa). In addition, the theoretical value of Nads for RS with a mass fraction of 20% of CN.LYS was 0.78 mmol g -1 but the experimental value was 0.66 mmol g -1 (i.e., 85% of the theoretical one).   Figure 11c,d present the simulated adsorbed amount for a N2/CO2 mixture at constant CO2 concentration of 20% and at 50 °C by varying the pressure for RS-20 and CN.LYS2, respectively. For RS-20, the affinity for CO2 was higher (P < 1 MPa) compared to RS without impregnation, but the Nads for N2 was superior (123.5%) to that for CO2 (Figure 11c). For RS-20, the CO2 isotherm obtained a form that would be impossible to obtain without impregnation due to the lack of surface area and molecular interactions. The isotherm for CN.LYS2 showed a higher affinity for CO2 and a correspondingly higher Nads than for N2, which, above 0.5 MPa, decreased considerably (Figure 11d). Similar conditions occurred when the CO2 concentration was varied at 2.5 MPa and 50 °C ( Figure  11e,f), CN.LYS2 had more affinity for CO2 than RS-20, and CN.LYS2 had a considerably higher Nads for CO2 than for N2. Figure 11e shows the increment of Nads only for CO2 while the CO2 concentration was increased in RS-20; at low concentrations of CO2 (< 30%), the Nads for N2 was higher. For CN.LYS2 at low concentrations of CO2 (1%), the slope increment was significant, indicating a higher affinity for CO2 at low concentrations.
From Figure 11, it can be concluded that the pyIAST is a useful tool because it is possible to simulate the behavior of adsorption systems from some pure gas adsorption data.

Conclusion
To the best of our knowledge, this is the first study using nanoparticles to modify the CO2 adsorption capacities of a reservoir in a carbon capture and storage (CCS) process. Moreover, this is the first research proposing a possible new configuration of the CCS process in which the storage is performed in shallow reservoirs (less than 300 m). We called it enhanced CCS (e-CCS), for which the main advantage is that the CO2 capture/separation step is removed, and the flue gas is injected directly into shallow deposits, where the CO2 is gaseous and where the adsorption phenomena control capture and storage.
Nitrogen-rich carbon nanospheres allowed increasing the adsorption capacity by 67,700% with a mass fraction of only 20% under realistic reservoir conditions (50 °C and 3 MPa). This was possible thanks to the higher surface area and to the favorable chemical composition, which promoted the capture and storage of CO2. These N-doped carbon nanospheres, synthesized by a simple process, had competitive CO2 capture performances compared to other special materials reported in the literature. Therefore, this research opens an interesting line of research that would expand knowledge in the field of carbon nanospheres for application in the adsorption and geological storage of CO2.  Figure 11c,d present the simulated adsorbed amount for a N 2 /CO 2 mixture at constant CO 2 concentration of 20% and at 50 • C by varying the pressure for RS-20 and CN.LYS2, respectively. For RS-20, the affinity for CO 2 was higher (P < 1 MPa) compared to RS without impregnation, but the N ads for N 2 was superior (123.5%) to that for CO 2 (Figure 11c). For RS-20, the CO 2 isotherm obtained a form that would be impossible to obtain without impregnation due to the lack of surface area and molecular interactions. The isotherm for CN.LYS2 showed a higher affinity for CO 2 and a correspondingly higher N ads than for N 2 , which, above 0.5 MPa, decreased considerably (Figure 11d). Similar conditions occurred when the CO 2 concentration was varied at 2.5 MPa and 50 • C (Figure 11e,f), CN.LYS2 had more affinity for CO 2 than RS-20, and CN.LYS2 had a considerably higher N ads for CO 2 than for N 2 . Figure 11e shows the increment of N ads only for CO 2 while the CO 2 concentration was increased in RS-20; at low concentrations of CO 2 (< 30%), the N ads for N 2 was higher. For CN.LYS2 at low concentrations of CO 2 (1%), the slope increment was significant, indicating a higher affinity for CO 2 at low concentrations.
From Figure 11, it can be concluded that the pyIAST is a useful tool because it is possible to simulate the behavior of adsorption systems from some pure gas adsorption data.

Conclusions
To the best of our knowledge, this is the first study using nanoparticles to modify the CO 2 adsorption capacities of a reservoir in a carbon capture and storage (CCS) process. Moreover, this is the first research proposing a possible new configuration of the CCS process in which the storage is performed in shallow reservoirs (less than 300 m). We called it enhanced CCS (e-CCS), for which the main advantage is that the CO 2 capture/separation step is removed, and the flue gas is injected directly into shallow deposits, where the CO 2 is gaseous and where the adsorption phenomena control capture and storage.
Nitrogen-rich carbon nanospheres allowed increasing the adsorption capacity by 67,700% with a mass fraction of only 20% under realistic reservoir conditions (50 • C and 3 MPa). This was possible thanks to the higher surface area and to the favorable chemical composition, which promoted the capture and storage of CO 2 . These N-doped carbon nanospheres, synthesized by a simple process, had competitive CO 2 capture performances compared to other special materials reported in the literature. Therefore, this research opens an interesting line of research that would expand knowledge in the field of carbon nanospheres for application in the adsorption and geological storage of CO 2 .

Funding:
The authors thank COLCIENCIAS for financing the doctoral studies of Elizabeth Rodriguez Acevedo through the call 647-2014. The authors thank COLCIENCIAS, Agencia Nacional de Hidrocarburos-ANH provided by agreement 272-2017 for the support provided and Universidad Nacional de Colombia for the support provided in the agreement 272-2017. The authors thank to Spanish Ministry of Science, Innovation and Universities, FEDER funds, contract number RTI2018-099224-B-I00. The authors also thank to ERASMUS+ program (agreement F NANCY43) and ENLAZAMUNDOS-SAPIENCIA for the support of academic internships. French authors acknowledge FEDER funds, through TALiSMAN project, for the financial support.
This section presents the adsorption isotherms fitted by the Sips and Toth models, for CO2 at high pressure and 0, 25, and 50 °C. This information is related to Section 3.2.1. Pure CO2 adsorption at high pressure-Manometric measurement method.