Reactive Transport Modeling and Sensitivity Analysis of CO₂–Rock–Brine Interactions at Ebeity Reservoir, West Kazakhstan

Abstract

This study investigated the reactive transport modeling of CO₂ injection into the Kazakhstan reservoir to identify mineralogical and porosity changes due to geochemical reactions. Additionally, sensitivity analysis was performed to test the effect of the surface area and gas impurity on the CO₂ storage capability. Despite the current need to investigate carbon sequestration in Kazakhstan, a limited number of studies have been conducted in this field. The Ebeity oil reservoir sandstone formation in the Pre-Caspian Basin has been tested as a potential CO₂ storage site. The 1D PHREEQC simulation results of 10,000 years suggest that reservoirs with a higher abundance of these secondary carbonates may be better suited for long-term CO₂ sequestration. The concentration of Fe³⁺ fluctuated, influenced by magnetite and siderite dissolution, leading to ankerite precipitation at 20 and 40 m. The porosity increased from 15% to 18.2% at 1 m and 20 m, favoring a higher CO₂ storage capacity, while at 40 m, it remained stable due to minor mineral alterations. A reduced surface area significantly limits the formation of dawsonite, a crucial secondary mineral for CO₂ trapping. For instance, at λ = 0.001, dawsonite formation dropped to 6 mol/kgw compared to 24 mol/kgw at λ = 1. Overall, the results of this study can play an essential role in future geological analyses to develop CO₂ storage in Kazakhstan for nearby reservoirs with similar geological characteristics.

1. Introduction

Fossil fuels are used worldwide; they increase energy consumption and contribute to global warming [1,2] by releasing greenhouse gases (GHG). The main contributor to greenhouse gas emissions is carbon dioxide (CO₂) [3]. Carbon capture and sequestration (CCS) technology is used to address the issue of unabated GHGs released into the atmosphere [3], which has been a general topic of climate change mitigation-related research to date.

As part of the Paris Agreement, which aims to reduce global warming well below 2 °C, Kazakhstan aims to reduce its GHG emissions by 15% below the 1990 limits by 2030, which establishes a net carbon footprint limit of 331 MtCO_2e per year [4]. The final net zero emissions are planned to be achieved by 2060, when the Land Use, Land Use Change and Forestry (LULUCF), and CCS sectors will cover 77 MtCO_2e. However, according to the recent assessment report [5], Kazakhstan’s climate policies and actions are marked as insufficient, meaning that if all countries were to follow the same approach, the temperature rise could reach up to 3 °C. Additionally, recent studies indicate that, based on energy system modeling (MARKAL-TIMES) results, the annual emissions in Kazakhstan are expected to increase [6]. Thus, there is an urgent need to develop and implement CCS technologies in the country.

According to Abuov et al. [7], Kazakhstan has a high storage capacity to sequester CO₂ in its geological formations. Earlier studies [8,9] have provided geological investigations of the south-eastern (SE) Pre-Caspian Basin, West Kazakhstan. The basin is also known for many anthropogenic CO₂ sources, constituting a significant part of the country’s economy through the oil and gas, power, and chemical industries. The Ebeity Reservoir, a part of the south-eastern Pre-Caspian basin, was chosen in this study to investigate the reactive transport modeling of CO₂ injections.

Depleted oil/gas reservoirs, saline aquifers, and coal seams have been recognized as suitable geological formations for successful geologic CO₂ sequestration [2,10,11]. Solubility and mineral trapping mechanisms immobilize CO₂ through interactions with the rock and formation water [12]. CO₂ can trigger geochemical reactions after interaction with brine water, changing the ion concentrations and mineral contents and lowering the pH. Consequently, fluid–mineral reactions can lead to the precipitation of secondary minerals.

To obtain the unique environmental parameters of the storage sites, geochemical and hydrogeological modeling has been considered essential as it can provide comprehensive knowledge about long-term interactions, such as aluminosilicates’ dissolution/precipitation [13], the behavior of minerals, their distributions and abundance, the effect of the relevant geological factors, and the impurity of injected CO₂ on its storage performance. Recently, the simulation of CO₂–rock–brine interactions has been implemented in several research projects and studies [14,15,16,17,18,19]. PHREEQC v3 has been widely used for reactive transport modeling (RTM).

Ensuring the safe CO₂ injection and storage reaction mechanisms and analyzing the short- and long-term fate of the CO₂ injected into the Ebeity oil field were the primary purposes of our study. A numerical simulation of the mass transport of the injected CO₂ has been performed to estimate its transport behaviors with diverse phases using the genuine flow parameters of the specific geological formation. Equilibrium and kinetic reaction-controlled geochemical modeling have been used to investigate brine–rock–gas interaction processes, such as mineralogical-fluid composition alterations, advective-diffusive transport, and secondary phase growth. Also, they provide the simulation results for the effect of the mineral surface area and mixed gases.

This study first presents the safe CO₂ storage potentials of the west Kazakhstan geological formation (Ebeity oil field). It provides comprehensive knowledge for successfully implementing its CO₂ geological sequestration in the near future as a starter.

2. Lithology and Sample Characterization

The Pre-Caspian Basin is a sizeable Paleozoic-Cenozoic basin located southeast (SE) of the East European Platform [20]. The Ural Mountains and Turanian Platforms are neighbors with the Pre-Caspian basin on the east and south sides [21]. Regional clastic sediments grew towards the SE side at the Triassic age [8], where sandstones are well established on discontinuous Triassic beds. Floodplain sediment facies and interconnected fluvial streams characterize these sandstone reservoirs between shales and siltstones [9]. This study considers the Zharkamys West-1 region in the SE part of the basin (Figure 1) and is embodied by post-salt mini basins and intra-salt deposit sets [8]. The Ebeity sandstone reservoir is a potential candidate for carbon storage in this region. It was selected due to its thickness (~40 m), porosity-permeability (15% and 1.38 × 10⁻¹³ m², respectively), cap rocks (shale) with low permeability, and the relative homogeneity (sandstone with well-distributed silicates) of the reservoir. A 1 m long core interval retrieved from an 850–890 m thick reservoir was used to identify the mineralogy of the reservoir.

The lithological map shows the rock stratum of the formation, including the caprock, reservoir, and bedrock (Figure 2). The caprock, comprising compacted and massive clay, overlies the fine- to medium-grained sandstone reservoir. The sandstone is porous, grey, and poorly cemented with carbonate [22]. The physical properties and mineralogy of the reservoir are typical of quartz-dominated, carbonate-cemented sand.

The mineralogical composition of the sandstone was analyzed through X-ray diffraction (XRD). Core samples from oil well EB-1 (48°08′32″ N and 55°22′01″ E) were collected and gently crushed for characterization. Four samples from injection zone Eb-1 were analyzed (Table S1). The core samples were crushed using a ball mill to avoid the effect of heterogeneity and packed to protect them from oxidation. The X-ray diffraction was performed using a Rigaku diffractometer (SmartLab), and semi-quantitative analysis was performed using the Profex5.2.2 software based on the Rietveld method. The XRD scan showed that the core samples comprise quartz, albite, kaolinite, and a small amount of calcite (

Table 1. Mineral composition in weight percentage and calculated concentrations retrieved by XRD analysis.

Mineral	Chemical Formula	Weight Percent [%]	Concentration [mol/kgw]
Primary
Quartz	SiO2	62.2	204.002
Albite	NaAlSi3O8	21.6	16.183
Calcite	CaCO3	3.4	6.694
Kaolinite	Al2Si2O5(OH)4	10.1	7.712
Magnesite	MgCO3	0.30	0.701
Siderite	FeCO3	0.70	1.191
Muscovite	KAl3Si3O10(OH)2	1.0	0.494
Magnetite	Fe3O4	0.30	0.255
Secondary
Dawsonite	NaAlCO3(OH)2	0.0	0.0
Ankerite	Ca (Mg, Fe) (CO3)2	0.0	0.0

).

Table 2. The aqueous phase chemical composition of brine aquifer in the reservoir.

Parameter	Value	Elements	Concentration [mg/kgw]
Density at 20.0 °C (kg/L)	1.1596	Cations
Resistance at 20.0 °C (Ω)	0.0547	Na	100,513
Temperature (°C)	40	Ca	1830
pH	5.97	K	1400
pe	−4.00	Mg	876
		Sr	61
		Fe	34
		Anions
		Cl−	165,211
		SO42−	0.1

includes information on the brine composition, such as the elements concentration and physical parameters. The provided water composition is balanced for the simulation using the PHREEQC. Al–Si-containing minerals cannot be equilibrated without primary elements (Al, Si); therefore, in-situ formation water was reconstructed and discussed in the following sections.

3. Geochemical Modeling

CO₂–fluid–rock interactions were simulated using geochemical software (PHREEQC v3) to predict the long-term effects of CO₂ injection on the variations in the mineralogical composition and brine chemistry of the geological formation. A zero-dimensional equilibrium simulation and one-dimensional RTM were applied to evaluate the influence of CO₂ injection under reservoir conditions. Equilibrium modeling provides information on the final composition of the aqueous phase and mineral assemblage, which are affected by interactions with CO₂ once a thermodynamic equilibrium is reached within the system. A set of nonlinear mass balance equations for each component and mass action equations among species, minerals, and gases is solved iteratively using the Newton-Raphson method [23,24]. The equilibrium model computes the species distribution among phases in a batch reactor and applies the best for sufficiently fast reactions relative to hydrologic processes. As the geochemical reactions in CO₂ sequestration are slow, kinetic simulations allow for characterizing the processes at different time intervals during storage [25,26]. Reactive transport models extend the system boundaries to several cells, rather than a single representative volume, to evaluate the variation in the properties across the system. PHREEQC considers the movement of a single aqueous phase between adjacent cells as the primary process in transport simulations [24]. The transport of dissolved species in groundwater is modeled using the Advection–Dispersion–Reaction (ADR) equation, which describes the concentration changes along the brine flow direction [27].

Additionally, RTM incorporates the reaction rates into the transport laws. CO₂ initiates fluid–rock interactions by forming an acidic environment in the brine, thus dissolving the host rock’s primary minerals. Subsequently, being transported through molecular diffusion, carbonic acid reacts with brine elements and forms stable carbonate minerals to permanently fix CO₂ [28,29]. Mineral dissolution/precipitation processes proceed in five steps: (1) diffusion of reactants from the bulk fluid to the mineral surface; (2) adsorption on reactive sites; (3) chemical reaction; (4) desorption from the surface; (5) diffusion of the products to the bulk fluid [23]. Adsorption/desorption is typically fast, while transport or surface reaction controls the kinetic rate [30].

Among the available PHREEQC databases, carbfix.dat was selected to calculate the fluid–rock interactions [31], considering the geochemical properties at elevated temperatures and pressures. Unlike other database files, such as phreeqc.dat, llnl.dat, and pitzer.dat [32,33], carbfix.dat was developed based on core10.dat with modified thermodynamic properties for ankerite and additional aqueous Al-, Fe-, Cl-, S-bearing species, and carbonate ions [31]. Due to the lack of available analytical data for aluminum and silica in brine, the formation water composition was first equilibrated with the mineral composition retrieved from the XRD analysis to exclude other reactions that could disturb the process after the injection of CO₂. It was also assumed that the initial brine and mineral contents of the reservoir with a porosity of 15% were uniform over the entire section considered in the RTM model. The temperature of 40 °C and pressure of 88.8 bar were used in the simulations. Isothermal and isobaric conditions were maintained throughout all of the simulated processes.

3.1. Estimation of Geochemical Equilibrium

The brine composition and mineral content determined through XRD were used to calibrate the initial parameters by scaling down to 15% porosity, assuming the complete filling of the pore space by 1 L of brine. The PHREEQC input required a unit conversion from the weight to the mole percentage of the mineral composition (Equation (1)).

$$c_m=\frac{10\times wt{\%}_m\times {\rho }_{rock}\times vol{\%}_{rock}}{M_m\times vol{\%}_{water}{\times \rho }_{water}}$$

(1)

where $c_m$ is the molality of the mineral (mol/kgw); $wt{\%}_m$ is the mass fraction of the minerals in the XRD analysis; ${\rho }_{rock}$ is the average density of the minerals comprising a rock formation (g/cm³); ${\rho }_{water}$ is the average formation water density (g/cm³); $vol{\%}_{rock}$ and $vol{\%}_{water}$ are the volumetric proportions of rock and brine, respectively, based on the 15% porosity; and $M_m$ is the molar mass of the mineral (g/mol). No additional secondary phases were introduced in this model as the simulation was used solely to establish the initial state of the geological formation and for the subsequent calculations. The missing concentrations of Al and Si were estimated by the equilibration of the aluminum silicate (muscovite) and quartz in the host rock [34]. The resulting compositions are listed in

Table 3. The initial condition for equilibrium and reactive transport modeling in PHREEQC.

Parameter	Unit
Brine composition	mol/kgw	Al	1.16 × 10−8
		C	8.39 × 10−3
		Ca	4.56 × 10−2
		Cl	5.23
		Fe	7.46 × 10−4
		K	4.02 × 10−2
		Mg	7.48 × 10−2
		Na	4.87
		S	1.85 × 10−3
		Si	1.84 × 10−4
	- kg/m3	pH	6.552
		Density	1.16 × 103
Mineral composition	mol/kgw	Albite	16.18
		Quartz	204
		Kaolinite	7.71
		Calcite	6.69
		Siderite	1.19
		Magnesite	0.7
		Magnetite	0.26
		Muscovite	0.49
		Ankerite	0
		Anhydrite	0
		Dawsonite	0
Temperature	°C		40
Pressure	bar		88.8

. The updated brine chemistry and rock composition were input parameters for the aqueous solution content and primary mineral composition in the CO₂–fluid–rock interaction simulations. They were further equilibrated with CO₂ to estimate the final reservoir conditions when all variations of the CO₂ injection were stabilized. The selection of the secondary minerals could affect the results of the subsequent simulations; therefore, the most relevant minerals were selected based on their reaction mechanisms and previous studies. Dawsonite, anhydrite, and ankerite were included as secondary minerals and were allowed to precipitate to evaluate the possible mineral inclusion of CO₂ in the deposit [35,36,37]. Anhydrite is an anhydrous calcium sulfate that can reduce the amount of CO₂ trapped in the solid phase and can negatively affect rock permeability [38]. The result of the equilibrium model provided a general understanding of the final state of the simulated system. It was also used as a reference for estimating more informative time-dependent reactive transport.

3.2. Reactive Transport Modeling

The RTM estimating the physical transport of the injected CO₂ and its geochemical reactions with reservoir environments allows for a more detailed characterization of geochemical species in different time frames and locations. The one-dimensional RTM simulation was performed in 40 cells (1 m long each), with the first cell representing the gas injection point. An abundant amount of CO₂ was introduced to the first cell from the constant-pressure source as a gas component, dissolved in brine under a constant flux condition, and moved forward at a rate of 10 years per cell with a dispersivity of 0.1 m²/s in each cell [39]. An advective-diffusion transport model was applied to simulate the flow of acidified brine along the 1D column, where the brine and rock components were distributed uniformly. A partial equilibrium (or semi-kinetic) model was applied to simulate a series of geochemical reactions triggered by the injection of CO₂ and fluid transport between adjacent cells. The local equilibrium assumption was used for calcite and the secondary minerals, i.e., dawsonite and anhydrite. These elements can follow fast reaction kinetics; thus, short time scales were required to reach their equilibrium conditions [40]. Additionally, the geochemical interaction kinetics of CO₂–brine–rock minerals was estimated over a 1000 year period. Calcite, dawsonite, and anhydrite were assumed to reach equilibrium at each time step of the RTM simulations. Other mineral phases were defined as kinetic reactants with a rate expression adopted from [30]:

$$r_{+.-}=k_{+}S\prod _i{a}_i^{v+}\left(1-exp\left(\frac{∆G_R}{\sigma RT}\right)\right)$$

(2)

where $r$ is the reaction rate (mol/s); $S$ is the reactive surface area of the mineral (m²); $k_{+}$ is the forward reaction rate coefficient (mol/m²s); $a$ is the activity of aqueous solute species $i$ that affect the rates; $v$ is the reaction order; $∆G_R$ is the Gibbs free energy of the reaction; $R$ is the gas constant; $T$ is the absolute temperature; $\sigma$ is the Temkin’s coefficient. The parameters used to estimate the kinetic rates of the minerals are provided in

Table 4. Kinetic rate parameters are used in the reactive transport model [41]. Siderite data are taken from [42].

Mineral	Neutral Mechanism		Acid Mechanism			Reactive Surface Area
	logknu25	Enu [kJ mol−1]	logknu25	Enu [kJ mol−1]	nH	S [m2/g mineral]	Ref.
Quartz	−13.99	87.7	-	-	-	0.111	[43]
Albite	−11.84	69.8	−10.16	65.0	0.457	0.164	[43]
Calcite	Equilibrium
Kaolinite	−13.18	22.2	−11.31	65.9	0.777	3.17	[43]
Magnetite	−10.78	18.6	−8.59	18.6	0.279	0.1	[40]
Siderite	−13.6	56.0		56.0	0.6	0.175	[44]
Muscovite	−13.55	22.0	−11.85	22.0	0.370	0.68	[40]
Ankerite		48.0	−7.5	48.0	0.94	0.18	[45]
Anhydrite	Equilibrium
Magnesite	−9.34	23.5	−6.38	14.4	1.00	0.8	[46]
Dawsonite	Equilibrium

.

The model can cover both reaction mechanisms in neutral and acidic conditions. Several assumptions were applied for the given model: (1) isothermal and isobaric conditions were kept along the simulated distance and time; (2) uniform distribution of brine species and minerals in the cell; (3) no movement of the gas phase along the simulated column; (4) constant fluid flow rate along the simulated distance without the influence of porosity and permeability variations.

The above-mentioned assumptions could have limitations that influence the simulation results, i.e.: (1) Isothermal and Isobaric conditions may not fully capture the dynamic variations found in real geological formations; (2) uniform distribution of brine species and minerals within the cell streamlines calculations may not represent the natural heterogeneity often encountered in geological systems; (3) the omission of gas phase migration should be considered when applying the model to scenarios where gas mobility plays a significant role; (4) the model may not accurately represent the complexities of fluid transport in such systems due to the constant flow rate assumption. Thus, the model’s predictions should be understood in the context of these simplifications.

The evolution of the dissolution/precipitation of minerals, changes in the brine parameters, and the amount of CO₂ bound by carbonates are characterized at any location and time from the start of CO₂ injection using the RTM estimations. In addition to the geochemical species redistribution between the aqueous and solid phases, the injection of CO₂ also led to changes in the poromechanical properties of the reservoir rock [47]. The porosity and permeability are the governing parameters determining the host rock’s physical structure and brine flow during CO₂ injection. They were influenced by the extent of the mineral dissolution and precipitation processes and determined the available space for CO₂ storage in the geological formation [48]. Although the precipitation of carbonates is favorable for the mineralization of CO₂, a decrease in porosity due to the higher solids content delays the fluid migration and limits the CO₂ storage capacity. Deviations from the initial porosity were calculated based on the difference in the total volume of the initial mineralogical assemblage (more on this can be found in Section S1 of the Supplementary Materials). The results of this reactive transport simulation and modeling parameters are labelled as the base case.

3.3. Sensitivity Analysis

3.3.1. Reactive Surface Area

The reactive surface area values of each mineral were obtained from the literature and are listed in Table 4. The reactive surface area affects the model’s accuracy as the dissolution/precipitation rates depend highly on the available surface area of rock minerals [11,49,50]. It is typically one to three orders of magnitude smaller than the geometric surface area or that measured by BET [51]. The relationship between the reactive surface area and specific surface area can be represented by the reactive fraction, as follows:

$$A=\lambda nM\beta$$

(3)

where $n$ is the number of mol/kgw; $M$ is the molar mass (g/mol); $\beta$ is the specific surface area measured by BET (m²/g); $\lambda$ is the reactive fraction of the total surface area.

The highest reaction rate is achieved when the surface of the mineral is available at a full extent for brine species and $\lambda =1$. However, the uncertainty of the reactive surface area fraction could significantly affect the results. The reactivity of minerals is affected by the irregularity and discontinuity of the surface and particle shapes [43]. They are not identical in the natural geological formations. Thus, models tend to overestimate the reactive fraction.

Additional RTM simulations were performed at $\lambda$ values of 0.1, 0.01, and 0.001 to visualize the influence of the reactive surface fraction. The other parameters were kept constant at 40 °C and 88 bar, as in the base model. Changes in the carbonate minerals, anhydrite molar contents, and porosity in the middle of the simulated range were compared to the base case results for evaluating the storage capacity with reduced rock reactivity.

3.3.2. Impurities in the Injected CO₂

The RTM model assumed the injection of pure carbon dioxide, which is an ideal case due to the possible impurities in the gas [52]. The impurity of the CO₂-rich mixture varies depending on the industrial source and capture technology [53]. The injected gas could have significant variations in its composition and concentration, ultimately affecting the fate of the injected gas. The Pre-Caspian basin is a source of sour natural gas and could produce acid gas, such as a mixture of CO₂ and H₂S alongside hydrocarbons. H₂S and CO₂ are usually removed from natural gas refineries, e.g., acid gas removal units by amine solvents [54]. It is not economically sustainable to separate the gases from each other for further treatment. In this study, the injections of several different mixture contents—i.e., 0–3% equimolar mixture of H₂S and O₂ impurities—to the CO₂ stream were simulated, analyzed, and compared to the base case. O₂ gas was added to the injected gas mixture to allow the formation of SO₂ from the hydrogen sulfide. Subsequent formation of sulfuric acid from SO₂ would decrease the pH of the reservoir water, promoting the dissolution of the primary minerals.

In a set of kinetic batch simulations, equal amounts of H₂S and O₂ were injected with CO₂. In each of the 10,000-year batch kinetic simulations, the total gas pressure of the CO₂–H₂S–O₂ mixtures was kept constant at the same level as the reservoir conditions. During the simulation, the gas phase dissolved and reached equilibrium with the formation water, passing the aqueous species to the subsequent cells. The effect of injecting the gas mixture on the dissolution of the primary minerals, the formation of carbonates, the change in porosity, and brine chemical species were investigated to evaluate the capacity of CO₂ sequestration at different gas compositions.

4. Results and Discussion

4.1. Estimation of the Equilibrium of CO₂–Rock–Brine Interaction

The results of the equilibrium modeling revealed that the rock mineral phases and chemical species of the brine solution were affected by the CO₂ injected into the reservoir (Figure 3). The increase in the Ca²⁺ concentration was attributed to the calcite dissolution, and the decrease in Na⁺ was attributed to the dawsonite precipitation. Moreover, the aqueous Fe²⁺ concentration increased ten times, indicating the dissolution of Fe²⁺-containing minerals. A small amount of muscovite dissolution provided K⁺ to the solution.

Moreover, SiO_2(aq) ions were mainly consumed to form quartz. The reactions increased the acidity of the brine solution and, ultimately, the pH was equilibrated at 4.14. The geochemical reactions among the rock–brine–CO₂ resulting from the CO₂ injection caused the dissolution of calcite, albite, magnesite, muscovite, kaolinite, and siderite, and the precipitation of quartz and secondary carbonates, such as ankerite and dawsonite (Figure 3). It is suggested that successful CO₂ sequestration via in-situ carbonation and storage could be determined by the abundance of relevant minerals, such as ankerite and dawsonite, precipitations of which can increase the CO₂ storage capacity of the reservoir. The change in the mineralogical composition induced by the CO₂ was closely related to the change in the reservoir porosity, showing a 4% decrease in porosity. The higher content of precipitated minerals led to a decrease in the porosity.

The evaluation of the mineral dissolution and precipitation states in the previous studies illustrated that dissolved calcite and aluminosilicates led to the growth of silica-containing minerals and secondary carbonates [39]. Albite and muscovite dissolution led to the precipitation of kaolinite by releasing silica elements [55]. Ankerite formation was caused by the carbonate minerals dissolution; Fe²⁺ and Mg²⁺ elements have been proven to precipitate. Pham et al. [55] stated that the released Mg²⁺ and Fe²⁺ initiated the formation of ankerite in the simulation.

4.2. Reactive Transport Modeling of Injected CO₂

The variations in the mineral type and brine chemical species concentrations are represented in Figure 4 and Figure 5 for 10,000 years. Three different distances from the injection point were taken to investigate the variations. The figures (Figure 4 and Figure 5) demonstrate the changes in the aqueous chemical species and minerals at distances of 1 m (a and b), 20 m (c and d), and 40 m (e and f) from the injection point.

The acidity of the solution was increased by the reaction of the injected CO₂ with the reservoir environment, and the decrease in the pH over time was observed at all different distances (Figure 4). At the CO₂ injection point, the pH decreased to ~3.4 during the first 4000 years, followed by a plateau value (Figure 4a). At 20 m, it reduced to 4.6, showing that the activity of H⁺ decreased in that region. Moreover, an increase in pH was observed at a 40 m distance, which peaked at 7.7 at about 3000 years (Figure 4e). The pH stabilization could mainly be attributed to the dissolution of the carbonate and bicarbonate species after the interaction with CO₂, showing a buffering effect in the solution. The brine water formed carbonic acid once the CO₂ was dissolved, and the weak acid further dissociated into H⁺ and HCO₃⁻. The dissolution of carbonates (Figure 5a–f) could be monitored after their reactions with H⁺, releasing Ca²⁺ and Mg²⁺ into the brine (Figure 4a–c). Generally, the reaction starts from the interaction between the dissolved CO₂ species and the host sandstone and carbonate minerals [56]. The production of carbonic acid (H₂CO_3aq) and the subsequent release of elements can lead to the dissolution of carbonate cement, neutralizing the acid and changing the porosity and permeability of the rock units [57].

The minerals were categorized into two parts; i.e., the minerals on the left have high molarity (Figure 5a,c,d), and those on the right have low concentrations (Figure 5b,e,f). The primary phase alterations could deliver the essential elements required to form secondary minerals and aqueous chemical species. The reactions in the first cell were reflected in other cells, with a delay due to the lower concentration of CO₂. The CO₂ entered the system and started to dissolve in the brine water; therefore, after 300 years, the amount of initial CO₂ was lowered. After reaching saturation with brine, the dissolution rate is minor; this is due to the limited dissolution capacity. As the pressure value was kept constant, the modeling was performed under continuous gas injection by providing enough CO₂ during the injection period. The interaction of the primary minerals with the brine containing dissolved CO₂ resulted in secondary minerals—i.e., dawsonite and ankerite—serving as the primary mineral storage of CO₂. The rapid dissolution of calcite (Figure 5a), magnesite, and magnetite (Figure 5b) occurred at the injection point. Prompt kinetic reactions occurred during the first 1000 years, when the mineral dissolution rates of the minerals were remarkable. Carbonates were dissolved simultaneously under acidic conditions (Equation (4)).

$$CaC{O}_{3\left(s\right)}\leftrightarrow {Ca}_{\left(aq\right)}^{2+}+{C{O}_3}_{\left(aq\right)}^{2-}$$

(4)

Carbon sequestration in solid mineral phases is among the most stable and long-term carbon storage strategies [13]. Carbon sequestration in the form of carbonate minerals can increase the reservoir’s retention capacity and improve the emplacement rock’s sealing ability by the reducing porosity and permeability via injection, given the preferable host rock and mineral chemistry [56]. Calcite, dawsonite (NaAlCO₃(OH)₂; found in zones of Na⁺-rich feldspar dissolution), siderite (FeCO₃), and ankerite (Ca(Fe,Mg,Mn)(CO₃)₂) are minerals frequently precipitated in reactive CO₂ transport models [58]. Primary minerals determine the nature and extent of carbonate precipitation in the deposit rocks. A continuous carbonate dissolution process could be limited as the calcite and magnesite were exhausted/reached equilibrium quickly. Therefore, the total carbonate dissolution (Figure 5a) resulted in the precipitation of secondary minerals, such as ankerite (Figure 5b).

The concentration of Fe³⁺ immediately increased and then decreased (Figure 4b,c). Magnetite is the primary source of Fe²⁺ and Fe³⁺; therefore, iron species could be dissolved throughout the simulation (Figure 5b,d,f). Once the magnetite started to dissolve, slight siderite precipitation occurred (Figure 5b), and magnetite dissolution with the release of Fe³⁺ was continuously observed. During this period, one could observe the dominant precipitation of ankerite-consuming chemical species (Figure 5d,f). Ankerite precipitation was observed at 20 m and 40 m distances, reaching their concentration at 2.3 and 0.9 mol/L, respectively. The degradation of the silicates resulted in ankerite precipitation, which required Ca²⁺, Mg²⁺, and Fe²⁺ for its formation. Albite, an alumino-silicate mineral, dissolved completely at the injection point after 4000 years (Figure 5a). At the following two distance points, the dissolution of albite was delayed, started to dissolve substantially after approximately 3000 years (Figure 5c), and reached a steady-state condition at 40 m, with a slight decrease at the end (Figure 5e). The albite was dissolved after the reaction with H⁺ (Equation (5)); thus, the dissolution was fostered within the low pH area.

$$NaAl{Si}_3{O}_8+4H^{+}={Al}^{3+}+{Na}^{+}+3Si{O}_2+H_{2}O$$

(5)

Kaolinite dissolved throughout the simulation and reduced to 3 mol/kgw at 1 m and 40 m and 5 mol/kgw at 20 m (Figure 5a,c,e). Its dissolution rate decreased at the 20 m distance due to the increased interaction with H⁺ ions (Figure 5c). Once the kaolinite dissolved, it could release Al^+3, and the quartz could form:

$${Al}_2{Si}_2{O}_5{\left(OH\right)}_4+6H^{+}=2{Al}^{3+}+2Si{O}_2+5H_{2}O$$

(6)

Some studies refer to the effect of clay minerals on the process and mineralization of CO₂ [25,59]. Although the changes occurred with kaolinite, the clay mineral characteristics must be considered. As reported by [25], swelling of the clay minerals, caused by the formation of water molecules outside of the clay structure, negatively affects the storage of CO₂. Unlike in other minerals, this phenomenon is less observed in kaolinite due to its characteristic exchange capacity and elasticity [25]. Clay-rich reservoirs with high-pressure conditions are restrained by their storage capacity.

At 1 m, the muscovite was dissolved fully after 8000 years (Figure 5b). Unlike the 1 m and 20 m cases, the muscovite formation was observed at 40 m (Figure 5f). Precipitation occurs through the interaction of K⁺ ions with aqueous chemical species released from the kaolinite dissolution (Equation (7)).

$$K{Al}_3{Si}_3{O}_{10}{\left(OH\right)}_2+10H^{+}=K^{+}+3{Al}^{3+}+3Si{O}_2+6H_{2}O$$

(7)

A positive correlation was observed between the K⁺ concentration and the dissolution of muscovite, although its change is not uniform (Figure 4b,d,f). The dissolution of aluminosilicates estimated by the model could be caused by a breakdown of mineral surface complexes [60]. Decomposition is usually dependent on the brine solution pH, and its chemical species change; e.g., the activity of Al can significantly influence the dissolution rate of aluminosilicates. The relevant effect of the pH and Al activity was observed from the near-to-injection point (1 m), where more silicates were dissolved than at the 20 m and 40 m levels. The dissolution of aluminosilicates could be the source of quartz and dawsonite formations, and the reactions became dominant in this estimation. Al³⁺, Na⁺, and SiO₂ were released into the reservoir water (Equation (6)) through albite dissolution, enabling quartz and dawsonite formation. The formation of quartz was observed after reaching the saturation of Al³⁺, Na⁺, and SiO₂ (Figure 5a,c,e), and reduced silica activity. Thus, the albite dissolved faster, and then the dawsonite concentration increased. As the feldspar group minerals provide further precipitation, high alumino-silicates and abundant quartz lead to massive dawsonite formation [61].

The quartz concentration increased to ~260 mol/kgw (Figure 5a), while the dawsonite increased rapidly to 23 mol/kgw at 3000 years. The formation rate of dawsonite decreased significantly, and a slight increase was monitored by the end of the estimation (Figure 5a). Once the silica minerals were fully dissolved or remained unreactive, the change in the contents of silica precipitation was limited due to the small dissolution amount (Figure 5a,c,e). It can be observed from Figure 4c that the delay in albite dissolution led to the stable concentrations of related minerals, such as quartz and dawsonite. This trend was further observed in Figure 5, where a small amount of quartz could form at the end of the estimation, while dawsonite was not precipitated. The decrease in silica activity and destabilization of albite led to the formation of a significant amount of dawsonite, and a corresponding increase in the total amount of carbon stored as secondary carbonate minerals. The precipitation of dawsonite and ankerite can safely fix the CO₂ gas in the reservoir as soil minerals via in-situ carbonation processes, thus providing the underground mineral storage of CO₂. The long-term stability of these minerals prove the excellent potential of the selected reservoir for safe and successful carbon storage.

On the other hand, several studies debate the stability and formation of dawsonite experimentally at a high concentration under CO₂ storage [61,62,63]. It is reported [61] that a TST-derived kinetic rate overestimates the precipitation of dawsonite due to the expressions and approach. Siliciclastic reservoirs with a low temperature and shallow depth are questioned with the formation of the dawsonite because of the strong correlation between the model and thermodynamic properties. The amount of dawsonite rises near the injection point due to a higher concentration of total dissolved carbon and feldspar dissolution, and decreases with distance [64]. Golubev et al. [44] showed that pH buffered by albite at ~0.1 mol/kg of HCO₃ creates better conditions for dawsonite precipitation.

It has also been reported that dawsonite is more stable at low pH values (acidity is equilibrated by buffering); therefore, the range of dawsonite stability is limited given the boundaries of the albite dissolution effect. Moreover, significant care must be taken to preserve the dawsonite and ankerite formation under in-situ conditions by maintaining an increased pressure of the injected CO₂ [37]. These minerals may not provide CO₂ fixation at low partial pressure, unlike calcite. One could question the reliability of the studies of reactive transport modelling with abundant dawsonite formation as a secondary phase; however, once the thermodynamic parameters of the minerals are set and provided, there is a need for further experimental verifications and validations.

The porosity increased from 15 to 18.2% at the 1 m and 20 m distances (Figure 5a,c), which was caused by the total dissolution of the carbonates and silicates. However, it did not change at the 40 m distance and was stable over the estimation time. The minor alteration of the minerals in the reservoir could explain the last cell’s stability in terms of the porosity. The increase in porosity provided a higher CO₂ storage capacity in the reservoir—a preferable condition for in-situ carbon mineralization.

4.3. Sensitivity Analysis

4.3.1. Effect of Mineral Surface Area

The formation of carbonates through CO₂ injection was investigated under different surface area conditions. Figure 6a–d showed the effect of the reactive surface area fraction (i.e., λ-factor showing the availability of surface to interact with other geochemical species such as H⁺) on the abundance of secondary minerals. The estimation results illustrated the contribution of dawsonite formation to the total CO₂ trapped in the reservoir (Figure 6a–d). Under the base case conditions (λ = 1), the molar concentration of dawsonite increased remarkably after 4000 years, reaching the final value of 24 mol/kgw (Figure 6a). Whereas, at lower λ-ratios, the growth rate of dawsonite remained stable (Figure 6b–d). A decrease in λ from 1 to 0.001 showed a four-fold reduction in dawsonite formation, resulting in 6 mol/kgw at the end of the estimation (Figure 6d). The available surface area of aluminosilicate minerals to interact with geochemical species was limited as the λ-factor decreased; therefore, the dissolution of albite and kaolinite led to decreased dawsonite precipitation compared to the base case. This resulted in the decreased mineralization of CO₂; thus, limited dawsonite and ankerite were formed. A low surface area limits the aluminosilicate mineral’s reactivity and slows down the formation of dawsonite. Calcite, a primary mineral, was removed through its dissolution following the CO₂ injection throughout the estimation (Figure 6a–d). Except for the base case, it was dissolved entirely at ~9200, ~7000, and ~8500 years, for λ = 0.1, 0.01, and 0.001, respectively (Figure 6b–d). The limited reactivity of aluminosilicate favors the suitable condition for calcite dissolution via the reaction with H⁺ ions. As calcite was assumed to be in equilibrium during the estimation, the effect of the reactive surface area at different λ-factors on its dissolution kinetics was not investigated. Siderite was also completely dissolved in four cases (Figure 6a–d). The slowest reaction kinetics was observed at λ = 0.001 for calcite and siderite (Figure 6d). The results confirmed the function of the surface area factor on the interaction kinetics. Ankerite experienced a relatively negligible effect from the decrease in surface area, and its content gradually increased to the maximum amount of 3 mol/kgw (Figure 6a–d). The highest ankerite content was observed in the base case and at λ = 0.1 (Figure 6a,b), while its content decreased by two-thirds at λ = 0.01 and 0.001 (Figure 6c,d). This may indicate the limited source of chemical species available in the brine to form ankerite. Therefore, the last two cases were unfavorable for safe gas injection because it can limit the mineralization of CO₂.

A small jump in the values of ankerite can be attributed to the sharp decrease in the calcite concentration, which is the primary source of the former mineral. Although calcite dissolution resulted in lower contents of carbonates in the deposit, it releases CO₃²⁻ into the bulk aqueous reservoir environment to form the secondary minerals and maintains their depositions. As ankerite and dawsonite contribute to CO₂ mineralization, low surface area values are not preferable where these minerals are less formed. The results demonstrated that possessing exact information about the surface area of the relevant minerals would be a critical factor in determining the contribution of mineral trapping among diverse storage mechanisms. The specific mechanism is a prerequisite for securing successful and safe geological CO₂ storage.

4.3.2. Effect of Impurities (H₂S–O₂ Gases)

We investigated the effect of small amounts of impure gases included in the injected CO₂, potentially influencing the geochemical interactions of reservoir rock minerals (carbonates and anhydrite). They contained SO_x, NO_x, O₂, and H₂S, and affected the solution pH and increased the mineral dissolution [65]. A co-injection of H₂S–O₂ and its effect at different volume fractions on the reservoir pH can be observed in Figure 7. The pH dropped at an early stage of different times and then stabilized at 0–1.5% mixed gas volume fractions, while the pH showed its peak value at the early stage and slightly decreased and stabilized at 2 and 3% (Figure 7). The distinction between 0–1% and 1.5–3% of the mixed gas effect on the pH can be characterized by the complete dissolution of carbonates (mainly calcite). At a higher pH (>4), the release of carbonic ions from the dissolved carbonates buffers the pH of the solution. Meanwhile, in acidic conditions, the total dissolution of carbonates restrains the pH stabilization. The estimated equilibrium pH reached a minimum value of 3.1 at a 3% volume fraction and a maximum of 4.2 at 0%. Regardless of the different trends of the pH variation caused by the mixed gas volume fractions, the equilibrium pH significantly decreased as the volume fraction increased, resulting in undesirable pHs, under which the maximum amount of anhydrite formed.

Figure 8a shows the effect of the mixed gases on the calcite dissolution via CO₂ injection. The calcite dissolves continuously, with a quick drop at an early stage, and is eventually stabilized by its slight and slow decrease. The calcite dissolution kinetics increased as the mixed gas volume fraction increased—e.g., even if it cannot be present at 2 and 3%. Magnesite, another carbonate, dissolved completely and released Mg²⁺ to the bulk brine solution to form secondary magnesium-containing minerals (Figure 8b). Siderite exhibited a distinct behavior under the specified geochemical conditions (Figure 8d). Its content was originally 1.1 mol/kg at a 1% mixture gas volume fraction; it gradually increased as the corresponding calcite dissolved at 1200 years, reached 1.5 mol/kg at 6000 years, and stayed constant throughout the estimation. The lowest content at equilibrium (0.3 mol/kgw) was observed in the base case (0%), and the highest content (1.7 mol/kgw) was marked at 1.5%. The siderite formation at 2 and 3% reached its peak value (1.7 and 1.9 mol/kgw) after 1100 years, and it dissolved to 1.2 and 0.7 mol/kgw, respectively, providing the bulk brine with a Fe²⁺ source to promote ankerite precipitation. Therefore, the reservoir brine was enriched by Ca²⁺, Fe²⁺, Mg²⁺, and HCO₃⁻ due to calcite, siderite, and magnesite dissolution. Together with the dissolved carbon species, the cations can promote the formation of minerals such as ankerite and dawsonite (Figure 8c,e). A further addition of CO₃²⁻ to the brine was required to increase their contents; however, their formation kinetics slowed down after 2000 and 4000 years, respectively, due to the complete dissolution of calcite and the limited availability of Ca²⁺. The dawsonite content showed a maximum value of ~20 mol/kgw at a 1% mixed gas volume fraction and lower (Figure 8c). The increase in the mixture gas volume fraction inhibited its formation by prevailing the SO₂ activity over the dissolved CO₂ species. The formation of ankerite was observed at 0–1%, and its content reached a maximum of 1.9 mol/kgw, showing an inversely proportional dependence on the mixed gas volume fraction (Figure 8e). A relatively minor fraction (i.e., 0.1 and 0.5%) of H₂S–O₂ resulted in a higher formation rate of ankerite, with the exception of the base case. Once calcite was dissolved entirely at 1% after 1200 years, the formation of ankerite was limited, and further dissolved to 0.7 mol.

Increasing the volume fraction of the mixed gases is not preferable for successful and safe CO₂ sequestration via the mineralization of the dissolved carbon in the brine reservoir. The acidification of the brine solution through co-injected H₂S–O₂ indicates that further sulfate formation can lead to a higher dissolution of the primary minerals (Figure 8). However, the porosity of the reservoir decreased with the increased impurity volume fraction, despite the rapid dissolution of the minerals. A decrease in the porosity indicates that the precipitation of anhydrite associated with H₂S–O₂ co-injection and anhydrite formation could be greater than that of the dissolved minerals. Anhydrite was not initially present in the reservoir; it reached the highest formation of 6.7 mol/kgw at 2 and 3% volume fractions. Anhydrite (CaSO₄) is composed of Ca²⁺ and SO₄²⁻, where the former originated from the dissolution of carbonates, mostly calcite, and the latter from SO₂. Anhydrite showed stability with no variations in its contents under acidic reservoir environments throughout the estimation. The lower the volume fraction of the added mixed gases, the lower the amount of anhydrite precipitated in the reservoir—i.e., the anhydrite formation was strongly dependent on the gas impurities present. The rapid anhydrite precipitation kinetics could control the in-situ carbonation trapping mechanism, unlike the other mineralogical interaction kinetics, where H₂S–O₂ was co-injected with CO₂. Therefore, the co-injected H₂S–O₂ with CO₂ could be trapped as anhydrite in the reservoir. Anhydrite was precipitated near the wellbore, and thus plugged down into the CO₂ injectivity (Figure 9a). Therefore, the loss of porosity at high impurity levels can influence the overall reservoir performance for CO₂ storage.

Figure 9b shows the effect of the mixture gas volume fraction on the extent of the mineralization of the injected CO₂ in 10,000 years. The lowest amount was obtained at a 3% mixed gas fraction, where, based on the estimation results, 3 mol/kgw was mineralized. The base case showed the highest CO₂ fixation, reaching 19 mol/kgw. Moreover, as the volume fraction of H₂S–O₂ increased to 1.5%, the maximum possible CO₂ sequestration via mineralization decreased by two times. Thus, a significantly lower mineralization amount was observed in the presence of, and increase in, gas impurity. The purification of mixed CO₂ gas to a certain purity degree needs to be implemented for safe and successful CO₂ sequestration and storage before its injection. According to Wang et al. [65], H₂S could cause fewer issues regarding the integrity of the reservoir than other impurities. It is also stated that sulfates may decrease the porosity and permeability. The oxidation of H₂S, consequently forming sulfuric acid, leads to geochemistry alterations in the medium; therefore, CO₂ injectivity depends on the level of purification.

5. Conclusions

Investigating geochemical alterations, their causes, and the subsequent impact on the reservoir performance for carbon sequestration after CO₂ injection is essential for understanding the safe and successful operation of the carbon storage and sequestration reservoir system. This study conducted numerical modeling to identify and investigate CO₂–rock–brine interactions, the reaction mechanisms, chemical species transport, and mineralization under CO₂ injection environments, such as geological formations and reservoirs.

The CO₂–rock–brine interaction equilibrium results emphasize the complex geochemical dynamics within the system and offer specific insights that are crucial for the successful implementation of CCS strategies. The precipitation of minerals like ankerite and dawsonite represents a promising aspect for increasing the CO₂ storage capacity of reservoirs. This suggests that reservoirs with a higher abundance of these secondary carbonates may be better suited for long-term CO₂ sequestration, potentially making them more attractive candidates for CCS projects. While the formation of secondary minerals can enhance CO₂ storage, it may also impact the fluid flow properties and overall reservoir integrity. Therefore, reservoir engineers and geoscientists must consider these changes when designing and managing CCS projects to ensure both storage efficiency and long-term containment.

The RTM results revealed significant insights into the long-term behavior of the CO₂–rock–brine system. The mineral dynamics displayed a dichotomy, with rapid dissolution of minerals like calcite, magnesite, and magnetite, followed by the formation of secondary minerals like ankerite, serving as primary carbon storage reservoirs. Silicate minerals like albite, kaolinite, and muscovite played pivotal roles in quartz and dawsonite formation. The reliability of reactive transport modeling for dawsonite formation was discussed, emphasizing the importance of experimental validation. These findings are crucial for optimizing carbon capture and storage strategies, enhancing reservoir characterization, and ensuring the long-term stability of geological carbon storage solutions, particularly in the context of in-situ carbon mineralization for safe and effective carbon storage.

Finally, the sensitivity analysis results indicate that a reduced surface area significantly limits the formation of dawsonite—a crucial secondary mineral for CO₂ trapping. This underscores the importance of surface area data in optimizing CO₂ storage strategies. Additionally, co-injecting impure gases like H₂S–O₂ affects the pH and mineral dissolution. Higher impurity levels lead to porosity loss, impacting the overall reservoir performance. Ensuring CO₂ purity before injection is vital for safe and efficient CO₂ sequestration.