An Integrated Numerical Study of CO₂ Storage in Carbonate Gas Reservoirs with Chemical Interaction between CO₂, Brine, and Carbonate Rock Matrix

Abstract

In light of the burgeoning interest in mitigating anthropogenic CO₂ emissions, carbonate reservoirs have emerged as promising sequestration sites due to their substantial storage potentials. However, the complexity of CO₂ storage in carbonate reservoirs exceeds that in conventional sandstone reservoirs, predominantly due to the rapid interactions occurring between the injected CO₂, brine, and carbonate rock matrix. In this study, a numerical model integrated with the chemical CO₂–brine–rock matrix interaction was developed to analyze the carbonate rock dissolution process and the physical property variations of different carbonate gas reservoirs during the CO₂ injection and sequestration process. More specifically, a total of twenty scenarios were simulated to examine the effects of lithology, pore size, pore–throat structures, and CO₂ injection rate on carbonate rock matrix dissolution and reservoir property variation. Calcite is significantly easier and quicker to react with CO₂-solvated brine than dolomite; as a result, limestones exhibit an expedited rock dissolution and pore volume increase, along with a slower pressure buildup in comparison to dolomites. Also, the carbonate reservoir with a smaller pore size has a higher rock dissolution rate than one with a larger pore size. Furthermore, the simulation results show injected CO₂ can modify the pore-dominant carbonate reservoir to a more pronounced extent than the fracture-dominant carbonate reservoir. Lastly, the carbonate rock dissolution is more obvious at a lower CO₂ injection rate. The insights derived from this research aid evaluations related to CO₂ injectivity, storage capacity, and reservoir integrity, thereby paving the way for environmentally and structurally sound carbon sequestration strategies.

1. Introduction

Carbon dioxide storage and sequestration in underground geological formations such as saline aquifers, depleted oil and gas reservoirs, and coal seams is a critical strategy for mitigating anthropogenic CO₂ emission and reducing atmospheric greenhouse gas levels [1,2]. Carbonate reservoirs are considered attractive CO₂ storage sites as they hold the majority of the world’s oil and gas reserves, making them a primary storage target when CO₂ based enhanced oil recovery (CO₂-EOR) is incorporated [3,4]. Furthermore, it can be assumed that CO₂ can be stored in a carbonate reservoir for an indefinite period since the hydrocarbons have been held in these reservoirs for millions of years [5]. Compared to conventional sandstone reservoirs, CO₂ storage in carbonate reservoirs is much more difficult and challenging for the following reasons. First, carbonate rocks are known to be structurally heterogeneous, with dual or triple pore–throat systems, and exhibit a range of wetting behaviors, all of which strongly impact CO₂ storage performance [3,6]. Second, carbonate rocks are very reactive with injected CO₂, thus undergoing a series of physical and chemical changes that may affect the reservoir’s integrity, injectivity, and storage capacity [7,8].

The CO₂–brine–carbonate rock interactions can cause mineral precipitation or rock dissolution in the carbonate reservoir, depending on the chemical species present in the formation brine [1,9]. Mineral precipitation, which solidifies the injected CO₂ as carbonate minerals, is one of the major CO₂ trapping mechanisms [10]. It happens when the brine pH is high and alkaline elements are present in the formation brine [11]. The precipitated carbonate minerals can occupy the reservoir pore spaces and block the flow paths leading to a reduction in porosity and permeability [11,12]. Another pronounced phenomenon is rock dissolution when injected CO₂ reduces the pH in formation water, converting ${\mathrm{CO}}_3^{2-}$ to ${\mathrm{HCO}}_3^{-}$ [13]. On the plus side for CO₂ storage and sequestration, dissolution of carbonate rocks facilitates injectivity, increases storage capacity, and enhances the performance of carbonate hosted CO₂ storage reservoirs [14]. Sometimes, the dissolution of carbonate minerals may result in the formation of highly porous and conductive channels or wormholes [5,15]. On the down side, dissolution creates concern for the long-term storage security of CO₂ in carbonate reservoirs [15,16]. The injected CO₂, being a less dense fluid than the formation fluids in subsurface pore spaces, will move laterally or migrate vertically after injection until it is restricted by a caprock or seal [3]. The dissolution and reactivity of carbonate reservoirs exposed to CO₂ may negatively affect their storage integrity, thus jeopardizing the long-term security of CO₂ sequestrated in the carbonate reservoirs [3,17].

The impacts of mineral dissolution in carbonate reservoirs due to CO₂ injection have been extensively investigated. It is found by Hosseini et al. that the interfacial tension between calcite and gas decreases after the introduction of supercritical CO₂ (scCO₂) into the carbonate pore spaces, and the decrease level is highly related to the reservoir pressure, temperature, and formation water salinity [18]. Menke et al. experimentally investigated dissolution in two different carbonate rocks with different levels of heterogeneity. They observed that both rocks exhibited channel widening at the millimeter scale, and the dissolution was more focused and progressed along the flow direction [16]. Similarly, it is claimed by Cui et al. that the formation of high permeability pathways due to rock dissolution is only observed in pre-existing open fractures where most injected scCO₂ flowed through [19]. Seyyed et al., recognized that the impact of calcite dissolution may be limited to the area near the scCO₂ injection point, while at distances far from the injection point, only minimal or no change was observed in the pore structure, pore roughness, and pore populations [20]. The interaction between the injected CO₂ and carbonate rock can also promote the diffusion of CO₂ in the carbonate reservoir. It has been found that the effective diffusion coefficient of CO₂ can increase by approximately one order of magnitude when the interaction of the CO₂–brine–carbonate rock matrix is considered in comparison to the pure CO₂ diffusion in the carbonate reservoir [21].

Many studies have been conducted to optimize the dissolution effects on CO₂ storage in carbonate reservoirs. Valle et al. found that scCO₂ can substitute acids and be used as safe, non-aqueous fracturing fluid in reservoir fracturing [22]. They claimed that the use of scCO₂ enables a greater development of fractures, and the permeability of such fractured rocks is three orders of magnitude higher than the ones created by conventional hydraulic fracturing [23]. Zhang et al., using a combined CT scanning and discrete element method (DEM), found that the carbonate dissolution will be much stronger when scCO₂ is injected after live brine flooding [24]. A series of carbonate water injection tests revealed that flow rate can notably influence the natural fracture profiles in carbonate rocks. The fracture widening is more pronounced when CO₂ is injected at a higher rate [24]. In addition, the carbonate dissolution rate can be changed by adding some external material during CO₂ injection, such as microbials and nanoparticles. For example, in Kolawole’s study, samples from deep carbonate formations are treated with microbial media and scCO₂. The results indicate that carbonate dissolution is slower with appropriate post-injection remediation [3]. In another study, Tan et al. found that CO₂–brine–rock interactions can be more intense and occur in small pores with the addition of SiO₂ nanoparticles [25].

In summary, it is known that the injection of large amounts of CO₂ into geological reservoirs may lead to a series of alterations due to chemical and physical interactions between minerals and fluids in carbonate- or carbonate-rich reservoirs. However, some research gaps need to be filled. First, so far, most of the CO₂ capture, utilization, and sequestration (CCUS) studies are focused on siliciclastic formations rather than highly chemically reactive carbonate formations [26]. Also, even though few studies regarding CO₂ storage in carbonate reservoirs have been conducted, researchers prioritized scCO₂ injection in deep saline aquifers rather than shallow hydrocarbon reservoirs, whose CO₂ storage advantages, such as low-cost drilling, injection, and transportation, abundance of available reservoir data and knowledge, and minimal uncertainties, should not be ignored. In this study, the carbonate rocks are selected as CO₂ storage sites, and different carbonate geological models are constructed to numerically investigate the effects of chemical CO₂–brine–carbonate rock interactions on CO₂ storage and sequestration. Specifically, a sensitivity analysis of four factors, i.e., lithologies, pore–throat structures, pore sizes, and rate, is presented, and the physical property variations before and after CO₂ injection in different carbonate reservoirs are compared. This study, to some extent, fills the knowledge gap in rock–pore fluid interaction in CO₂ storage in shallow carbonate reservoirs and helps researchers and engineers in aspects of reservoir selection and operation optimization in the projects of CO₂ storage in carbonate reservoirs.

2. Simulation

2.1. Scenarios

In this study, twenty scenarios were numerically simulated to analyze the effects of lithology, pore size, pore-structure type, and CO₂ injection rate on the process of carbonate rock dissolution. These four parameters are studied because the lithology, pore-structure type, and pore size are the most important geophysical features, and the injection rate is one of the most important operating parameters. The numerically constructed geological models used in the scenarios have the same dimensions of 10 m in length, 10 m in width, and 1 m in height. The porosity of the geological model used in each scenario is also the same and set at 10%. The lithology, pore space type, pore size, and CO₂ injection rate in each scenario are listed in

Table 1. The lithology, pore space type, pore size, and CO₂ injection rate in each scenario.

Scenario No.	Lithology	Pore Space Type	Pore Size (µm)	Injection Rate (mol/min)
1	Limestone	Pore-dominant	5	1.0
2	Limestone	Pore-dominant	5	0.5
3	Limestone	Pore-dominant	5	0.1
4	Limestone	Pore-dominant	10	0.1
5	Limestone	Pore-dominant	50	0.1
6	Limestone	Fracture-dominant	5	1.0
7	Limestone	Fracture-dominant	5	0.5
8	Limestone	Fracture-dominant	5	0.1
9	Limestone	Fracture-dominant	10	0.1
10	Limestone	Fracture-dominant	50	0.1
11	Dolostone	Pore-dominant	5	1.0
12	Dolostone	Pore-dominant	5	0.5
13	Dolostone	Pore-dominant	5	0.1
14	Dolostone	Pore-dominant	10	0.1
15	Dolostone	Pore-dominant	50	0.1
16	Dolostone	Fracture-dominant	5	1.0
17	Dolostone	Fracture-dominant	5	0.5
18	Dolostone	Fracture-dominant	5	0.1
19	Dolostone	Fracture-dominant	10	0.1
20	Dolostone	Fracture-dominant	50	0.1

. The lithologies in Scenarios #1–10 and #11–20 are limestone and dolostone, respectively. Scenarios #1–3, #6–8, #11–13, and #16–18 are simulated to investigate the CO₂ injection rate effect on carbonate matrix dissolution in different reservoirs with different lithologies and pore space types. Scenarios #1–3 have limestone lithology, pore-dominant space type, 5 µm pore size, and varying CO₂ injection rates of 0.1, 0.5, and 1.0 mol/min to examine carbonate matrix dissolution in limestone pore-dominant reservoirs. Similarly, scenarios #6–8, #11–13, and #16–18 are simulated to understand the effects of CO₂ injection rate on carbonate matrix dissolutions in limestone fracture-dominant, dolomite pore-dominant, and dolomite fracture-dominant reservoirs, respectively.

Scenarios #3–5, #8–10, #13–15, and #18–20 are simulated to investigate the effect of pore/fracture size on the carbonate matrix dissolution process in different carbonate reservoirs with different lithologies and pore space types. For example, the pore size effect on limestone pore-dominant reservoirs can be found in Scenarios #3–5. Similarly, the other three groups of scenarios, i.e., Scenarios #8–10, #13–15, and #18–20, are simulated to investigate the pore space size effect on limestone fracture-dominant reservoirs, dolomite pore-dominant reservoirs, and dolomite fracture-dominant reservoirs, respectively. It is worth noting that the residual water saturations in the reservoirs with different pore sizes are different. The residual water saturations are 76.5%, 55.0%, and 18.4% for pore-dominant reservoirs with pore sizes of 5 µm, 10 µm, and 50 µm, respectively. For the fracture-dominant reservoirs, the residual water saturations are 68.2%, 44.1%, and 8.7% for the reservoirs with fracture diameters of 5 µm, 10 µm, and 50 µm, respectively. The calculation of the residual water saturation will be presented in Section 2.3.

Two groups of scenarios, i.e., #3 and #6, as well as #13 and #16, are compared to analyze the CO₂–brine–rock interactions in pore-dominant and fracture-dominant reservoirs. Scenarios #3 and #6 are simulated to compare the dissolution process in limestones. Scenarios #3 and #6 have the same lithology, pore/throat size, and CO₂ injection rate, which are limestone, 5 µm, and 10 mol/min, respectively. The difference between these two scenarios is that the geological model used in Scenario #3 is pore-dominant type, whereas the other is fracture-dominant in Scenario #6. Scenarios #13 and #16 are similar to Scenarios #3 and #6. The difference between these two series is that scenarios #13 and #16 are used to compare the dissolution processes in different pore–throat types in dolomite reservoirs instead of limestone reservoirs.

2.2. Simulation of CO₂ Injection

In this study, the process of continuous CO₂ injection and sequestration in a carbonate-depleted gas reservoir is simulated. The code for simulating the process of continuous CO₂ injection was written by the author and compiled using MATLAB software (Mathworks, MATLAB, R2023a). The simulator involves four major parts: CO₂ injection, CO₂ dissolution in residual water, carbonate rock dissolution, and reservoir property variation.

CO₂ is injected and stored in the carbonate-depleted gas reservoir in a gas phase. Thus, first, there is no phase change during the entire process, and second, the geological model only takes in injected CO₂, but no CO₂ flows out. In each simulation step, CO₂ is injected into the depleted gas reservoir at a constant injection rate (Table 1). With CO₂ injection, the reservoir pressure will build up. For the reservoir dimension (10 m × 10 m × 1 m), it is assumed that the injected CO₂ can instantaneously and evenly spread into and fill the pore space. As a result, the reservoir pressure can instantaneously reach equilibrium once CO₂ is injected. The reservoir pressure is calculated using the Peng–Robinson equation of state (P-R EOS) [27].

2.3. CO₂ Dissolution in Residual Brine

After CO₂ is injected into the depleted carbonate gas reservoir, the injected CO₂ will start to dissolve in the residual brine, which will lead to a decreased amount of CO₂ in the pore space and an increased amount of dissolved CO₂ in the residual brine. This will result in a decrease in CO₂ partial pressure in pore space and an increase in CO₂ partial pressure in brine. An equilibrium will be reached once the CO₂ partial pressure in the pore space is equal to that in the residual brine.

The CO₂ partial pressure in the pore space is calculated using the P-R EOS. Meanwhile, the CO₂ partial pressure in the residual brine is calculated using Henry’s law:

P = k/C

(1)

where P is the CO₂ partial pressure in brine, Pa; k is Henry’s law constant, mol/L/Pa; and C is the CO₂ concentration in brine, mol/L.

The methods to calculate the residual brine volumes for pore-dominant reservoir and fracture-dominant reservoir are as follows. For both the pore-dominant and fracture-dominant reservoirs, it is assumed that residual brine exists in the pore or fracture spaces in the form of brine film attached to the surface of pore or fracture wall. The thickness of the brine film is set at 0.2 µm [28,29]. The pores and fractures are assumed to be spheric and cylindrical shapes, respectively. The volume of the residual brine is calculated by using the deduction of total pore or fracture volume to void space volume. As for the pore-dominant reservoir, the equation to calculate residual brine volume is the following:

V_w = V_Por ∗ [r³ − (r − 2)³]/r³

(2)

And for the fracture-dominant reservoir, the equation is:

V_w = V_por ∗ [r² − (r − 2)²]/r²

(3)

where V_w is the residual brine volume, m³; V_por is the pore volume, m³; and r is the pore or fracture radius, m. The residual brine saturation is calculated using the residual brine volume divided by the pore volume, which is:

S_w = V_w/V_Por

(4)

2.4. CO₂–Brine–Carbonate Interaction

The kinetics of the CO₂–brine–carbonate interaction are illustrated by the following kinetic rate law, which is given by TOUGHREACT [10,29].

$$R_{\mathrm{m}}=-S_{\mathrm{m}}{k}_m{e}^{-\frac{E_{\mathrm{m}}}{R}\left(\frac{1}{T}-\frac{1}{298.15}\right)}\left(1-\frac{Q_{\mathrm{m}}}{K_m^{\mathrm{eq}}}\right)$$

(5)

where R_m is the mineral reaction rate, S_m is the specific surface area for the mineral, T is the absolute temperature, R is the gas constant, $K_m^{eq}$ is the equilibrium constant of mineral reaction, and Q_m is the corresponding activity product. k_m and E_m are the rate constant at 298.15 K and activation energy, respectively, which are listed in

Table 2. The mineral kinetic reaction constant for calcite and dolomite used in this study [10,13,30].

Lithology	Reaction Constant
	logk	Em (kJ/mol)
Calcite	−5.81	23.5
Dolomite	−7.53	52.2

[10,13,30].

2.5. Simulation Workflow

Figure 1 illustrates the workflow employed to simulate the dissolution of carbonate rocks during the CO₂ injection and sequestration process. This simulation begins with the assignment of initial property values to the geological model, encompassing parameters such as initial reservoir pressure (P_ini, MPa), initial reservoir porosity (Φ_ini, %), initial pore volume (V_por, in m³), time interval (Δt, s), residual brine saturation (S_w, m³/m³), pore or fracture radius (r, m), and CO₂ injection rate (q_inj, mol/min). Specifically, P_ini, Φ_ini, and V_por are standardized across all simulation scenarios at 1 MPa, 10%, and 10 m³, respectively, with the assumption that the residual gas in the carbonate reservoir is methane. Distinct values for pore radius (r_Por, m), fracture radius (r_f, m), and CO₂ injection rates for each simulation scenario are provided in Table 1. Residual brine saturation is determined through equations [2,3,4] for pore-dominant and fracture-dominant reservoirs.

Upon each CO₂ injection into the reservoir, the simulation updates the quantity of CO₂ in the pore space ($N_{{\mathrm{CO}}_{2\_\mathrm{gas}}}$). Subsequently, the injected CO₂ starts to dissolve into the residual brine, and the dissolved CO₂ in the brine accounts for the brine volume (V_w) change in response to CO₂ concentration change, which is characterized by using the swelling correlation. The pressures of CO₂ in both gas and liquid phases are updated employing the Peng-Robinson equation of state (P-R EOS) and Henry’s law, respectively. The equilibrium between these phases is assessed by comparing their pressure differences. If the difference falls below a threshold of σ = 0.001 bar, the system is deemed at equilibrium, and then the system is prompted to calculate the CO₂–brine–rock matrix interaction rates. If not, 0.1% of CO₂ from the gas phase is dissolved into the liquid phase for another comparison, continuing until equilibrium criteria are met.

The acidic brine, resulting from dissolved CO₂, reacts with the carbonate rock matrix, where the reaction rate is calculated using Toughreact’s model as referenced in the previous section. This reaction updates the concentrations of dissolved Ca²⁺ (N_Ca²⁺, mol) and CO₂ (N_{CO2_water}) and calculates the enlargement of pore space due to CO₂–brine–rock matrix interaction. This procedure is repeated at an incrementing time Δt of 10 min at each iteration until a continuous injection period of 72 h is concluded. Upon reaching this end condition, the simulation compiles outcomes, including changes in reservoir pressure, space volume, CO₂ concentration, and porosity.

3. Results and Discussion

3.1. Effects of Lithology

The intricate CO₂–brine–rock matrix interaction in different lithological carbonate reservoirs, particularly limestone and dolostone reservoirs, showcases markedly different physicochemical responses under equivalent operational conditions. Figure 2a,b demonstrates these disparities, focusing on the respective variations in pore volume augmentation and pressure escalation within the limestone and dolostone reservoirs during CO₂ injection at a fixed rate of 0.1 mol/min. With an underlying assumption that the limestone is exclusively composed of calcite and the dolostone solely of dolomite, the depicted outcomes reveal a more pronounced pore volume enlargement observed in the limestone reservoir in comparison with that in the dolostone reservoir. Specifically, the introduction of 4200 mol of CO₂ into the limestone reservoir catalyzed an 8.5% increment in pore volume. However, this value is reduced to a paltry 0.16% in the dolostone reservoir, one-fifth of the limestone value. This pronounced disparity can be attributed to the inherent reactivity characteristics of the reservoir materials, where calcite is characterized by a significantly higher propensity to react with carbonate brine compared to dolomite. This is further evidenced by the superior reaction rate between calcite and carbonate brine. Moreover, the reactivity variance precipitates a more rapid pressure accumulation in the limestone reservoir in comparison to that in the dolomite reservoir, as delineated in Figure 2b. This observation is primarily driven by two key factors. Firstly, the rapid interaction between the limestone’s rock matrix and carbonate brine facilitates an accelerated CO₂ consumption, which consequently results in a smaller amount of CO₂ in the pore space in the limestone reservoir in comparison to that in the dolostone reservoir. As a result, the reservoir pressure increases at a slower rate in the limestone reservoir than its dolostone counterpart. Second, the above-mentioned faster increase in the pore volume within the limestone reservoir, relative to the dolostone reservoir, plays a crucial role. According to the real gas law expressed as P = ZnRT/V, a larger pore volume can lead to a lower reservoir pressure. There, an expedited volumetric expansion effectively slows down the rate of reservoir pressure build-up in the limestone reservoir.

These findings underscore the paramount importance of mineralogical composition in the practice of CO₂ storage and sequestration in carbonate reservoirs. For example, dolostone reservoirs commonly have superior initial injectivity due to their more developed natural factures compared to limestone reservoirs. However, the limestone reservoirs may hold greater potential for CO₂ storage and less operational difficulties in CO₂ injection in the long run because of their higher reactivity with carbonate brine [31].

3.2. Effects of Pore Size

Figure 3 and Figure 4 present the physical variances, i.e., the reservoir pressure and pore volume, in the carbonate reservoirs during the CO₂ injection process, specifically targeting the limestone and dolostone formations with different pore radii of 5, 10, and 50 µm. The constant CO₂ injection rate of 0.1 mol/min serves as a controlled parameter to elucidate the effects of pore size on reservoir behavior as a result of the introduced CO₂. Analysis of these two figures reveals a consistent trend across both types of carbonate reservoirs, whereby a diminution in pore radius correlates with an increase in reservoir pressure and a more pronounced augmentation in reservoir pore volume. The underlying mechanisms driving this observed phenomenon can be attributed to the increased surface area afforded by smaller pore sizes, since for a fixed pore volume, a smaller pore radius results in a larger surface area. The larger surface area in the carbonate reservoir with a smaller pore radius significantly enhances the interaction between the CO₂-saturated carbonate brine and the carbonate rock matrix, which accelerates the dissolution process of the rock matrix and leads to a faster pore volume increment and slower pressure buildup. Moreover, the residual brine saturations are higher in the reservoir with smaller pore radii due to the hypothesis in this study that no free brine exists in the reservoir and all the residual brine adheres to the pore walls. This is exemplified by the calculation showing a dramatic difference in residual water saturation levels—75% in reservoirs with a pore radius of 5 µm versus only 8.5% in those with 50 µm pores. The higher residual brine saturation levels in smaller pores facilitate a larger CO₂ dissolution capacity in the brine, leading to a more pronounced and rapid increase in pore volume. A comparative analysis between limestone and dolostone reservoirs reveals that with identical pore sizes and no matter what the pore radius is, limestone reservoirs exhibit lower reservoir pressures and more significant increases in pore volumes compared to dolostone reservoirs. This distinction is attributed to limestone’s inherently greater reactivity with CO₂-saturated carbonate brine compared to dolostone, which has been stated in Section 3.1.

3.3. Effects of Pore–Throat Structures

Figure 5 and Figure 6 illustrate the impact of pore–throat structures on the space volume increase and reservoir pressure dynamics on the respective limestone and dolostone reservoirs during the CO₂ injection process, respectively. The space volume in this subsection means pore volume for the pore-dominant reservoir and fracture volume for the fracture-dominant reservoir. Figure 5a,b presents a comparative analysis of space volume changes and pressure buildup in both pore-dominant and fracture-dominant limestone reservoirs during the CO₂ injection process. Similarly, Figure 6a,b details the comparison of space volume change and pressure buildup in the dolostone reservoir under the same conditions. It is worth noting that both the pore or fracture radius and the CO₂ injection rate were held constant across all scenarios, set at 5 µm and 0.1 mol/min, respectively.

The analysis shows a notably faster increase in the space volume within the pore-dominant reservoirs compared to their fracture-dominant counterparts. For instance, a limestone reservoir exhibited an 8.5% increase in space volume upon the injection of 4200 mol of CO₂ in a pore-dominant structure, as opposed to a 6.8% increase in a fracture-dominant structure under identical injection conditions. This discrepancy is attributed to the differences in the reaction surface area provided by the geometrical shapes of the reservoir structures: cylindrical for fractures and spherical for pores. The spherical shape offers a larger surface–volume ratio, hence providing a larger reaction surface area in pore-dominant reservoirs when the space volume of the two types of reservoirs is the same. It has been stated in Section 3.2 that the larger reaction surface area not only accelerates rock matrix dissolution but also facilitates higher CO₂ dissolution in brine, leading to a more significant increase in space volume. Furthermore, the disparity in space volume increase between pore-dominant and fracture-dominant reservoirs is more pronounced in limestone than in dolostone. This is due to the faster CO₂–brine–rock matrix interaction in limestone, which amplifies the effects of pore–throat structures on space volume increment. The reservoir pressure change with the injected CO₂ amount follows an opposite trend as the space volume increase vs. injected CO₂ amount data. Fracture-dominant reservoirs display larger and faster reservoir pressure increases than pore-dominant reservoirs, either in limestone or dolostone. This phenomenon is directly associated with the fact that fracture-dominant reservoirs possess smaller and slower space volume enlargement. Since the pressure is inversely related to the space volume based on the real gas EOS stated in Section 3.1, the reservoir pressures in the fracture-dominant reservoirs are larger and increase faster than those in the pore-dominant reservoirs. Similarly, the larger pressure difference between the pore-dominant and fracture-dominant reservoirs observed in limestone in comparison to that in dolostone can be explained by the larger discrepancy in space volume increase between the pore-dominant and fracture-dominant reservoirs in limestone.

These findings underscore the dynamic interplay between pore–throat structures and CO₂ injection in carbonate reservoirs. While fracture-dominant reservoirs may initially seem advantageous for CO₂ sequestration due to easier injection, pore-dominant reservoirs exhibit a greater long-term CO₂ storage capacity [32]. This is primarily due to their enhanced and more rapid increase in pore space, highlighting the critical role of reservoir structure, lithology, and mineralogy in optimizing CO₂ storage and sequestration processes.

3.4. Effects of CO₂ Injection Rate

The dynamic responses of pore volume and reservoir pressure to varying CO₂ injection rates in limestone and dolostone reservoirs are demonstrated in Figure 7 and Figure 8. Specifically, Figure 7 delineates the increment in pore volume and the corresponding reservoir pressure build-up in the limestone reservoir during CO₂ injection at injection rates of 0.1, 0.5, and 1 mol/min. Figure 8 illustrates the same results, but in the dolostone reservoir.

It is observed that the pore volume expansion within both limestone and dolostone reservoirs inversely correlates with the CO₂ injection rate. For instance, the pore volumes increase by 8.5% and 0.16% in the respective limestone and dolostone reservoirs at a CO₂ injection rate of 0.1 mol/min. In contrast, the increases in pore volume in the limestone and dolostone reservoirs are merely 0.85% and 0.016%, respectively, at a higher injection rate of 1 mol/min, which are ten times less than those at the lower injection rate of 0.1 mol/min. The mechanism for this inverse relationship is attributed to the slower injection rates, which provide a longer residence time for CO₂ to physically dissolve and chemically interact with the carbonate rock matrix, thereby facilitating a greater increase in pore volume. It is worth noting that the reduction in pore volume increase is not linearly proportional to the increase in CO₂ injection rate. For instance, in the limestone reservoir, a five-fold increase in the CO₂ injection rate from 0.1 mol/min to 0.5 mol/min can result in an 8.5-fold reduction in pore volume increase. However, doubling the injection rate further to 1 mol/min only led to a 1.6-fold reduction in the pore volume increase. This phenomenon illustrates the non-linear and more pronounced impact of CO₂ injection rates at lower ranges on pore volume expansion. Conversely, reservoir pressure exhibits a directly proportional relationship with the CO₂ injection rates, which is opposed to the trend observed in pore volume changes. This observation is consistent with the established understanding that an inverse relationship exists between pore volume and reservoir pressure; a faster rate of pore volume increase leads to a comparatively slower rate of reservoir pressure buildup. The findings suggest the existence of an optimal CO₂ injection rate for carbonate reservoirs. While a rapid injection rate may expedite reaching CO₂ storage targets and reduce costs, it could potentially compromise the long-term maximum CO₂ storage capacity of the reservoirs, indicating a trade-off between immediate efficiency and future storage potential.

Overall, the findings challenge traditional perceptions regarding carbonate reservoir suitability for CO₂ storage, particularly highlighting the favorable characteristics of pore-dominant limestone reservoirs with small pore sizes, which is very much against the traditional opinion of seeking high-permeable fracture-dominant dolostone reservoirs with large pore sizes. This underscores the importance of constantly reevaluating reservoir characteristics because of CO₂ injection-induced modifications. It is also worth noting that this study aims to investigate the effects of CO₂–brine–carbonate rock interactions from a mechanistic perspective. Thus, the geological model was ideally constructed, such as the geological features of homogeneity, pore space of perfect spherical and cylindrical shape, uniform pore size, etc. It is evident that the actual carbonate reservoirs are far more complex than the geological model constructed in this paper. Therefore, further research is necessary to quantitatively apply the results of this research to real-world practices.

4. Conclusions

In this paper, the carbonate rock matrix dissolution process is comprehensively studied, and its implications on the physical variation in different carbonate reservoirs during CO₂ injection and sequestration are simulated and demonstrated. The following conclusions can be drawn from this study.

Firstly, mineralogy plays an important role in the practice of CO₂ sequestration in carbonate reservoirs. Limestone reservoirs exhibit a faster reaction with injected CO₂ and residual brine compared to dolostone reservoirs due to their higher reactivity with carbonate. This leads to a slower reservoir pressure buildup in limestone reservoirs due to the more rapid dissolution of pore space.

Secondly, the type of reservoir, whether pore-dominant or fracture-dominant, significantly influences the dissolution process. Pore-dominant reservoirs exhibit a larger reaction surface area, resulting in faster dissolution and slower pressure buildup compared to fracture-dominant reservoirs.

Thirdly, carbonate mineral dissolution behavior is also related to pore size. Smaller pore sizes lead to larger reaction surface areas and greater residual brine saturation, resulting in faster pore volume expansion but slower pressure buildup. Conversely, carbonate reservoirs with larger pore sizes exhibit faster pressure buildup and slower pore volume expansion.

Furthermore, the CO₂ injection rate also has an impact on dissolution dynamics. Slower injection rates allow for more time for CO₂ to dissolve into residual brine and react with the rock matrix, leading to increased pore volume expansion and reduced pressure buildup compared to faster injection rates.

Importantly, these effects, i.e., pore space type, pore size, and injection rate, were found to be more pronounced in limestone reservoirs than in dolostone reservoirs, which indicates that the dolostone reservoir is more stable and less affected during the CO₂ injection process.

Lastly, the above-mentioned conclusions hint at the necessity of continually reassessing the characteristics of the carbonate reservoir due to variations induced by injected CO₂ and challenge the traditional perspective in the selection of high porosity and permeability zones for CO₂ storage in carbonate reservoirs. Also, the idealistic geological model can be used to mechanistically and qualitatively study carbonate rock dissolution during CO₂ injection. However, the quantitative application is limited.