A Simulation Study on Evaluating the Influence of Impurities on Hydrogen Production in Geological Carbon Dioxide Storage

Abstract

In this study, we examined the effect of CO₂ injection into deep saline aquifers, considering impurities present in blue hydrogen production. A fluid model was designed for reservoir conditions with impurity concentrations of 3.5 and 20%. The results showed that methane caused density decreases of 95.16 and 76.16% at 3.5 and 20%, respectively, whereas H₂S caused decreases of 99.56 and 98.77%, respectively. Viscosity decreased from 0.045 to 0.037 cp with increasing methane content up to 20%; however, H₂S did not affect the viscosity. Notably, CO₂ with H₂S impacted these properties less than methane. Our simulation model was based on the Gorae-V properties and simulated injections for 10 years, followed by 100 years of monitoring. Compared with the pure CO₂ injection, methane reached its maximum pressure after eight years and eleven months at 3.5% and eight years at 20%, whereas H₂S reached maximum pressure after nine years and two months and nine years and six months, respectively. These timings affected the amount of CO₂ injected. With methane as an impurity, injection efficiency decreased up to 73.16%, whereas with H₂S, it decreased up to 81.99% with increasing impurity concentration. The efficiency of CO₂ storage in the dissolution and residual traps was analyzed to examine the impact of impurities. The residual trap efficiency consistently decreased with methane but increased with H₂S. At 20% concentration, the methane trap exhibited higher efficiency at the end of injection; however, H₂S had a higher efficiency at the monitoring endpoint. In carbon capture and storage projects, methane impurities require removal, whereas H₂S may not necessitate desulfurization due to its minimal impact on CO₂ storage efficiency. Thus, the application of carbon capture and storage (CCS) to CO₂ emissions containing H₂S as an impurity may enable economically viable operations by reducing additional costs.

1. Introduction

The South Korean government announced the “First Master Plan for implementing the Hydrogen Economy” in November 2021, the goal of which was to become a leading country in the hydrogen economy. They intended to produce 750,000 tons of blue hydrogen by 2030 and 2 million tons by 2050, according to the Ministry of Trade, Industry, and Energy in 2021 [1]. Blue hydrogen is produced using fossil fuels to generate hydrogen, and the resulting CO₂ is captured and stored using carbon capture and storage (CCS) technology to prevent its atmospheric release. This is the most cost-effective method for producing clean hydrogen. One notable project in this field is the CarbonNet Project, which occurred in Victoria, Australia. Additionally, in 2023, Republic of Korea plans to implement CCS in the East Sea Gas Field, which is a depleted gas reservoir, and gradually expand its usage.

There are two methods of hydrogen production using fossil fuels: steam methane reforming (SMR) using natural gas and coal gasification. During hydrogen production, methane and H₂S are generated in addition to hydrogen and CO₂. The greenhouse effect of methane is approximately 20-fold that of CO₂, whereas H₂S is highly toxic; therefore, its atmospheric emissions are regulated. When conducting CCS projects, pure CO₂ is injected because impurities in CO₂ can cause an increase in the two-phase region, resulting in potential transportation issues as the critical point changes. According to a report by the International Energy Agency Greenhouse Gas R&D Programme in 2011, depending on the type of impurity, compared to pure CO₂, storage capacity can be reduced by up to 40% when injected at a depth of 895 m [2]. However, removing impurities from CO₂ requires installing desulfurization devices or using membranes, which incur additional costs.

In the LaBarge gas field operated by ExxonMobil in Wyoming, a gas mixture consisting of 70% H₂S and 30% CO₂ is injected [3]. In the Borzecin gas field in Poland, enhanced gas recovery (EGR) is the objective, and a gas mixture containing 20% H₂S and CO₂ is injected into the subaqueous zone of a gas–water reservoir [4]. Furthermore, Carbfix, a project conducted in Iceland, observed faster mineralization when a mixture of H₂S and CO₂ generated from a geothermal power plant was injected into basaltic formations than with pure CO₂ [5].

Nogueira and Mamora [6] experimentally reported that impurities, such as nitrogen, oxygen, water, sulfur dioxide, nitrous oxide, and carbon monoxide, are present at approximately 1% when storing CO₂ in depleted gas fields and storing the same amount of pure CO₂ is possible. The impure CO₂ injection does not significantly affect storage efficiency. However, this study has the limitation of lower impurity concentrations than those encountered under actual field conditions [6]. Li et al. (2017) confirmed that nitrogen is included as an impurity, and compared to that of pure CO₂, its density decreases, resulting in rapid vertical movement of the gas plume, reaching the top of the reservoir, followed by a relatively large horizontal dispersion [7]. Gaus, Azaroual, and Czernichowski-Lauriol evaluated the impact of CO₂ diffusion in the cap rock of the Sleipner field using a one-dimensional model and found a decrease in porosity owing to geochemical reactions [8]. Cho et al. [9] developed a simulation model for water alternating gas (WAG) using CO₂ and methane. The simulation results showed that methane inclusion reduced miscibility, leading to improved oil recovery, minimized changes in permeability and porosity, and reduced moisture content in the reservoir [9]. Due to the recent increase in data availability, the application of AI in the CCS field is expanding and aiding in faster decision-making [10,11]. Kim et al. [10] presented an artificial neural network (ANN) model to predict CO₂ storage efficiency in deep saline aquifers, validated against the Gorae-V structure—a potential saline aquifer for CO₂ storage in Republic of Korea [10]. Sedaghat and Rouhinakhsh used the least square support vector machine (LSSVM) to study the vapor–liquid equilibrium (VLE) and phase behavior of impure CO_2, such as nitrogen, hydrogen, oxygen, and argon [11]. Schöneich, May, and Vosteen observed that the density and storage capacity decreases when CO₂ contains argon, oxygen, nitrogen, carbon monoxide, and hydrogen impurities. However, this study did not consider reservoir flow through simulation but used thermodynamic calculators, which have limitations in assessing the impact of H₂S and methane [12].

The literature demonstrates the possibility of injecting impure CO₂ based on the type and concentration of impurities and the characteristics of the storage reservoir. However, research involving impurities in CCS has primarily focused on enhanced oil recovery (EOR) and EGR, and limitations have been observed concerning lower impurity concentrations than those encountered under actual field conditions. Herein, we examined CO₂ injection into deep saline aquifers, considering impurities present in blue hydrogen production. The properties of the Gorae-V structure were obtained, and a 3D model was created to investigate the injection of impure CO₂ using the reservoir simulation software GEM 2023.10 developed by CMG. Impurities that can be generated during hydrogen production, such as methane and H₂S, which are regulated by atmospheric emissions, were selected. The concentration of impurities included in CO₂ was set to a realistic level for field applications. The changes in properties that occur when these impurities are present in CO₂ were calculated, and the pressure changes in the reservoir and variations in CO₂ storage capacity were researched according to the type and concentration of impurities.

2. Theoretical Background

First, understanding hydrogen production methods that use fossil fuels and the types of impurities is necessary to evaluate the impact of the impurities generated during hydrogen production in CCS, which can be involved in the CO₂ capture process. Hydrogen production from fossil fuels can be divided into natural gas and coal gasification methods. The typical method for producing hydrogen using natural gas is SMR, which involves reacting methane with water to produce hydrogen and carbon monoxide, as shown in Equation (1). Equation (2) describes the water–gas shift (WGS) reaction that converts carbon monoxide into hydrogen and CO₂. To prevent corrosion in an SMR reactor, natural gas should be pretreated to capture H₂S using desulfurization or capture devices. In the CO₂ capture process, methane that does not react with water is captured along with CO₂. Understanding the effects of impurities entering the storage layer is crucial [13].

$${\mathrm{CH}}_4+{\mathrm{H}}_2\mathrm{O}\leftrightarrow 3{\mathrm{H}}_2+\mathrm{CO}$$

(1)

$$\mathrm{CO}+{\mathrm{H}}_2\mathrm{O}\leftrightarrow {\mathrm{H}}_2+{\mathrm{CO}}_2$$

(2)

Coal gasification involves the partial oxidation of pulverized coal using a gasifier to produce synthesis gas, which consists primarily of carbon monoxide and hydrogen. The composition of syngas varies depending on coal composition, and methane and H₂S can be generated during the gasification process. Coal gasification is primarily performed in the form of an integrated gasification combined cycle, and the major chemical reactions involved are shown in Equations (3)–(5). The carbon monoxide emitted from the gasifier is converted to hydrogen through a WGS reaction, similar to Equation (2) [14].

$$2\mathrm{C}+{\mathrm{O}}_2\leftrightarrow 2\mathrm{CO}$$

(3)

$$\mathrm{C}+{\mathrm{H}}_2\mathrm{O}\leftrightarrow {\mathrm{H}}_2+\mathrm{CO}$$

(4)

$$\mathrm{C}+2{\mathrm{H}}_2\leftrightarrow {\mathrm{CH}}_4$$

(5)

Table 1. Possible impurities during the CO₂ capture process [18].

Description	Possible Impurities
CO2 captured from natural gas sweetening	CH4, amines, H2O
CO2 captured from heavy oil production and upgrading	H2S, N2, O2, CO, H2O, COS, Ar, SOx, NOx
CO2 captured from power plants using post-combustion capture	N2, amines, H2O, O2, NH3, SOx, NOx
CO2 captured from power plants using oxy-combustion capture	N2, O2, SO2, H2S, Ar
CO2 captured from power plants using pre-combustion capture	H2, CO, N2, H2S, CH4

displays possible impurities that may occur during the CO₂ capture process using various methods. The most common method for capturing CO₂ is absorption using amine-based solvents such as monoethanolamine (MEA). Amine solvents have excellent absorption capabilities for CO₂, as well as acidic gases such as H₂S. The CO₂ absorption process consists of absorber and regenerator towers, as shown in Figure 1 [15]. In the absorber tower, the synthesis gas containing CO₂ and the lean amine solution are injected into the upper and lower sections, respectively, where the amine solution absorbs CO₂. The amine solution, now rich in CO₂ (rich amine), is discharged to the lower part of the absorber tower and preheated in a heat exchanger before being injected into the upper part of the regenerator tower. In the high-temperature regenerator tower, the CO₂ is released in the upper section through a reverse reaction, whereas the amine solution (lean amine) is returned to the absorber tower through a heat exchanger and cooler [15,16]. As shown in Figure 2, the absorbent can capture CO₂ and also acidic substances, such as H₂S or sulfur dioxide. Depending on the process, impurities such as methane or hydrogen may be included [17].

The CO₂ injected into the reservoir should not leak. To ensure this, the target storage formation should have good porosity, and an impermeable cap rock should be present at the top, such as shale. The primary target formations for CCS include depleted oil and gas fields, saline aquifers, and deep coal seams. Saline aquifers are more widely distributed than other CO₂ storage candidates and are globally recognized as potentially storing the greatest amount of CO₂, ranging from a minimum of 1000 Gt to a maximum of 10,000 Gt, as shown in

Table 2. CO₂ potential storage capacity of the CCS site [19].

CCS Site	Capacity (Approximately)
Oil and gas fields	675–1200 Gt
Deep saline aquifers	1000–10,000 Gt
Unmineable coal seams	100–200 Gt

[19].

The injected CO₂ in the reservoir is trapped through four mechanisms: structural trapping, residual trapping, dissolution trapping, and mineral trapping. However, storing impure CO₂ underground can affect CO₂ storage due to the characteristics of the impurities and changes in properties caused by mixing. Compared to pure CO₂, when methane is mixed with CO₂ for EOR and used in a WAG process, the deposition of asphaltenes decreases, the hydrophilicity of the reservoir rock is maintained, and the relative permeability increases. Consequently, compared to pure CO₂, after the WAG process, the amount of CO₂, including methane, captured in the residual trap decreases by 30% [9].

From a geochemical perspective, CO₂ and impurities affect the acidity generation, dissolution, and precipitation of minerals in reservoir rock. According to Aminu’s research [20], when pure CO₂ was injected into a saline aquifer, the pH decreased to 5.97, and carbonate minerals such as calcite in the rock dissolved, resulting in changes in permeability. When nitrogen dioxide was included as an impurity, strong acidity was formed, and the pH decreased from 5.97 to 5.23, whereas in the case of sulfur dioxide, the pH decreased to 5.86. However, when H₂S was included, the pH increased from 5.97 to 6.55. When pure CO₂ was injected, the permeability of the sample increased from 240 to 254 md. However, nitrogen dioxide decreased the permeability to 140 md, and sulfur dioxide decreased it to 225 md, whereas H₂S showed a similar permeability to pure CO₂ at 255 md. The decrease in permeability in the presence of acidic nitrogen dioxide and sulfur dioxide was due to an increase in the precipitation of carbonate and sulfate minerals [19]. Additionally, when sulfur dioxide is injected along with CO₂ into an undersaturated quartz state in a saline aquifer, it causes precipitation and reduces the permeability of the reservoir [19]. Ionization occurs when H₂S and CO₂ are injected into a saline aquifer. Equations (6)–(8) represent the reactions that occur when these substances are injected into a saline aquifer [21].

$${\mathrm{CO}}_2+{\mathrm{H}}_2\mathrm{O}\leftrightarrow {\mathrm{H}}_2{\mathrm{CO}}_3$$

(6)

$${\mathrm{CO}}_2+{\mathrm{H}}_2\mathrm{O}\leftrightarrow {\mathrm{H}}^{+}+{\mathrm{HCO}}_3^{-}$$

(7)

$${\mathrm{H}}_2\mathrm{S}+{\mathrm{H}}_2\mathrm{O}\leftrightarrow {\mathrm{HS}}^{-}+{\mathrm{H}}_3{\mathrm{O}}^{+}$$

(8)

Furthermore, impurities can lead to a decrease in gas density, resulting in an increased buoyancy-driven increase in the velocity of the gas plume and a decrease in the contact time between the brine and gas, leading to a reduction in dissolution efficiency. A decrease in density reduces the amount of CO₂ that can be retained within the pore space. When impurities such as nitrogen are present, they can also cause a decrease in viscosity, similar to the case of reduced density, resulting in an increased plume flow velocity and decreased dissolution trapping and residual trapping efficiencies [7].

3. Methods

3.1. Fluid Model

To simulate the behavior of CO₂ in a deep saline aquifer, the pressure–volume–temperature and density of mixtures were calculated in CMG-WinProp using the Peng–Robinson equation (PR-EOS). The PR EOS is suitable for predicting the phase behavior of gas–liquid systems under high pressure and is commonly used for CCS processes involving mixtures, such as gas processing, oil refining, and petrochemical processes [22]. The PR-EOS applied to a mixture of components is given by Equations (9)–(16), as follows:

$$P=\frac{RT}{V_m-b}-\frac{a\left(T\right)}{V_m^2+2V_m-b^2}$$

(9)

$$a\left(T\right)=0.45724\frac{R^2{T}_c^2}{P_c}\alpha \left(T\right)$$

(10)

$$\sqrt{\alpha }=1+k\left(1-\sqrt{T/T_c}\right)$$

(11)

$$k=0.37464+1.54226\omega -0.26992{\omega }^2 \mathrm{for} \mathrm{lighter} \mathrm{components} \left({\mathrm{CH}}_4, {\mathrm{H}}_2\mathrm{S}\right)$$

(12)

$$k=0.3796+1.485\omega -0.1644{\omega }^2+0.01667{\omega }^2 \mathrm{for} \mathrm{heavier} \mathrm{components} \left({\mathrm{CO}}_2\right)$$

(13)

$$a_{mix}={\displaystyle \sum }_{i=1}^{N_c}{\displaystyle \sum }_{j=1}^{N_c}{x}_i{x}_j{a}_{ij}$$

(14)

$$b_{mix}={\displaystyle \sum }_{j=1}^{N_c}{x}_i{b}_i$$

(15)

$$a_{ij}=\sqrt{a_{ii}{a}_{jj}}\left(1-k_{ij}\right)$$

(16)

where $P$ is the pressure (atm), $T$ is the temperature (K), $V_m$ is the volume of the mixture, $R$ is the universal gas constant, a is the coefficient for volume calculation by the attraction effect, $b$ is the coefficient for volume calculation by the repulsion effect, $T_c$ is the critical temperature, $P_c$ is the critical pressure, $\omega$ is the acentric factor, $x_i$ is the mole fraction (single component $i$ in mixtures), and $k_{ij}$ is the binary parameter [23].

The viscosity of the gas was estimated using the Jossi–Stiel–Thodos correlation. The process of dissolving CO₂ in the formation of water is described by Equation (17).

$${\mathrm{CO}}_{2\left(\mathrm{g}\right)}\leftrightarrow {\mathrm{CO}}_{2\left(\mathrm{aq}\right)}$$

(17)

where ${\mathrm{CO}}_{2\left(\mathrm{g}\right)}$ is gas phase of ${\mathrm{CO}}_2$, ${\mathrm{CO}}_{2\left(\mathrm{aq}\right)}$ is aqueous phase of ${\mathrm{CO}}_2$ [24].

CO₂ rapidly dissolves in formation water, and from a thermodynamic perspective, it can be assumed to be in an equilibrium state. Equations (18) and (19) represent the thermodynamic equilibrium state using the fugacity [25].

$$f_{ig}-f_{iw}=0, i=1, 2, 3, \cdots n_g$$

(18)

$$f_i=x_{iw}\times H_i$$

(19)

where $f_{ig}$ is the fugacity of the $i$ component of the gas phase, calculated by the PR-EOS; $f_{iw}$ is the fugacity of the $i$ component of the aqueous phase, calculated by Henry’s law using components of fugacity and concentration; $x_{iw}$ is the mole fraction of component $i$ in the liquid phase; and $H_i$ is Henry’s constant, which is a function of the system pressure, temperature, and aqueous phase in saline water [25].

The molecular diffusion of gas in a liquid phase is driven by concentration gradients and is modeled based on Fick’s second law, as described by Equation (20).

$$\frac{\partial c}{\partial t}=D{\nabla }^{2}c$$

(20)

where c is the concentration of a component, and $D$ is Fick’s diffusion coefficient [25].

The initial temperature and pressure of the saline aquifer, along with the critical properties and acentric factors for each component used in the calculation of the fluid properties, are presented in

Table 3. Parameter for fluid modeling, Reproduced with permission from Reid et al. [26], The Properties of Gases and Liquids; published by AIP Publishing in 1959.

Parameter	Value
Reservoir temperature (K)	376.2 K
Reservoir initial pressure (atm)	244.8
Components	CO2	CH4	H2S
Critical temperature (K)	30.1	−82.8	100.1
Critical pressure (atm)	11.4	4.6	8.9
Acentric factor	0.228	0.011	0.031
Impurity contents in the CO2 stream	3.5 mol.%, 20 mol.%

.

3.2. Reservoir Model

Simulations of CO₂ storage in the deep saline aquifer were conducted using the compositional simulation software GEM developed by CMG. The simulations were run on Intel i9-12900K CPU, 32 GB RAM, and NVIDIA GeForce RTX 3070 GPU. The average elapsed time for simulations was 5 min. The deep saline aquifer of the Gorae-V structure, which includes the Donghae gas field, was used as the target reservoir for storage and is located 60 km offshore, east of Ulsan, as shown in Figure 3 [27]. As shown in

Table 4. Model parameters for CO₂ injection simulation [28].

Parameters	Value
Grid type	Block type
Number of cells	33 × 33 × 10 (10,890)
Grid size	100 × 100 × 6 m
Depth	2450 m
Initial pressure	24.8 MPa
Initial temperature	103
Horizontal permeability	50 md
Vertical permeability	5 md
Porosity	16%
Thickness	60 m
Initial water saturation	100%

, the depth of the saline aquifer is 2450 m, and the initial pressure and temperature are 24.8 MPa and 103 °C, respectively. The porosity of the aquifer is 16%, and it exhibits horizontal and vertical permeabilities of 50 and 5 mD, respectively [28]. The reservoir model was constructed as a rectangular shape with a grid size of 100 m in the horizontal direction and a height of 6 m. The grid array was set to 33 × 33 × 10, totaling 10,890 grids. The reservoir characteristics in the simulation model were assumed to be homogeneous and isotropic; the relative permeability is shown in Figure 4 [29].

3.3. CO₂ Injection Scenarios

Table 5. Operating parameters for CO₂ injection simulation.

Parameters	Value
Maximum bottom-hole pressure	38 MPa
Gas injection rate	146.4 ton/d
Injection period	10 years
Monitoring period	100 years
Impurities	CH4, H2S
Impurity contents in the CO2 stream	3.5%, 20%

displays operating parameters for CO₂ injection simulation in this study. CO₂ was injected through the three centered bottom cells of the model to enable the gas to rise and implement residual trapping owing to buoyancy differences. The injection rate was controlled by adjusting the bottom-hole pressure (BHP) of the injection well. The maximum BHP was set to 38 MPa, which corresponds to 80% of the rock fracturing pressure, known to be 48 MPa for the Gorae-V structure deep saline aquifer [30]. The CO₂ injection rate was set to 146.4 tons per day. After the injection was completed, monitoring was conducted for the next 100 years to observe changes in trapping owing to CO₂ plume migration. This model only included structural trapping, residual trapping, and dissolution trapping without considering the influence of mineral dissolution and precipitation on the deep saline aquifer. The selected impurities for injection were methane and H₂S, which can be present in CO₂ from precombustion processes. The impurity concentration was set to 3.5%, representing the maximum H₂S content for amine-based absorbents [15] and 20%, as reported in a project conducted in Borzecin, Poland [4].

4. Analysis of Simulation Results

4.1. Fluid Model Results

The properties of CO₂ and impurities are presented in

Table 6. Properties of gas at the target formation.

Parameter	Pure CO2	CO2-CH4		CO2-H2S
Impurity Content (mol.%)	-	3.5	20	3.5	20
Impurity Volume (vol.%)	-	5.13	27.16	2.24	13.66
Molar mass (g/mol)	44.01	43.03	38.42	43.66	42.02
Molar volume (m3/kmol)	0.0802	0.0824	0.0921	0.0800	0.0776
Density (kg/m3)	548.50	521.97	417.25	546.07	541.78
Viscosity (cp)	0.045	0.043	0.037	0.045	0.046

under the initial temperature and pressure conditions in the target formation. Pure CO₂ had a molar mass of 44.01 g/mol, molar volume of 0.0802 m³/kmol, density of 548.5 kg/m³, and viscosity of 0.45 cp. When methane was present at 3.5%, the volume occupied by the gas mixture increased by 5.13%, and at a concentration of 20%, it occupied 27.16%, exceeding its proportion in the mixture. In contrast, when H₂S was present at 3.5%, it occupied a lower percentage of the volume (2.24%), and at a concentration of 20%, it occupied 13.66%, which was lower than its concentration in the injected H₂S. In both cases, the molar mass decreased when methane or H₂S was present as an impurity. With 3.5% methane, the molar mass decreased by 97.78% to 43.03 g/mol, and with 20% methane, it decreased by 87.3% to 43.42 g/mol. With 3.5% H₂S, the molar mass decreased by 99.2% to 43.66 g/mol, and with 20% H₂S, it decreased by 95.5% to 42.02 g/mol. The molar volume increased in the presence of methane. With 3.5% methane, it increased to 0.0824 m³/kmol (a 2.74% increase), whereas with 20% methane, it reached 0.0921 m³/kmol (a 14.84% increase). However, H₂S caused a slight decrease in molar volume. With 3.5% H₂S, it decreased to 0.0800 m³/kmol (a 0.25% decrease), and with 20% H₂S, it decreased to 0.0776 m³/kmol (a 3.24% decrease). Differences in the molar volume resulted in a significant decrease in the gas density in the presence of methane. With 3.5% methane, density decreased to 521.97 kg/m³ (a 95.16% decrease), and with 20% methane, it decreased to 417.25 kg/m³ (a 75.16% decrease). In contrast, H₂S caused a smaller decrease in density. With 3.5% H₂S, the density decreased to 546.07 kg/m³ (a 99.56% decrease), and with 20% H₂S, it decreased to 521.78 kg/m³ (a 98.77% decrease), showing a smaller reduction than that of methane. The viscosity also decreased when methane was present. With 3.5% methane, the viscosity was 0.043 cp (a 95.56% decrease), and with 20% methane, it was 0.037 cp (an 82% decrease). H₂S did not affect viscosity when present at 3.5%; however, with 20% H₂S, the viscosity increased to 0.045 cp (a 2.22% increase). When non-condensable gases such as methane are included as impurities, an increased volume ratio inside the gas results in reservoir conditions. However, H₂S is compressed, leading to a decrease in its volume. Owing to these characteristics, when methane is included as an impurity, the molar volume increases compared to that when only pure CO₂ is present, whereas the presence of H₂S causes a decrease in the molar volume. The density difference between pure CO₂ and the impurities varied. Methane showed a larger reduction in density as its concentration increased; however, when H₂S was present, the relative difference in density was smaller.

4.2. Dynamic Simulation Model Results

Figure 5 shows the BHP (solid line) and cumulative gas mass (dotted line) over the 10-year period of CO₂ injection. When CO₂ is injected into the saline formation, the injection fluid is momentarily compressed near the injection well to displace the formation water, causing a rapid increase in the BHP. The increased pressure then decreases as CO₂ flows and subsequently increases linearly with increasing injection rate. The BHP exhibited a constant trend after reaching a maximum pressure of 38,000 kPa, with the timing varying depending on the presence and concentration of impurities. The sequence of BHP reaching the maximum pressure was methane, pure CO₂, and H₂S. The cumulative gas mass increased linearly until BHP reached the maximum pressure, and the injected gas mass followed the order of pure CO₂–H₂S–methane. After the BHP reached its maximum pressure, the cumulative gas mass decreased gradually. Additionally, comparing the overall changes in BHP and the trends in cumulative gas mass between cases with impurity concentrations of 3.5% (Figure 5a) and 20% (Figure 5b), the difference caused by the impurities became more pronounced as concentration increased.

The BHP and cumulative gas mass during the last 3 years of injection are shown in Figure 6. For the pure CO₂ injection, the BHP reached its maximum pressure in January 2029. With 3.5% methane, it took 8 years and 11 months from the start of injection to reach the maximum pressure, and when the methane concentration was 20%, it took 8 years to reach the maximum pressure from the start of injection. When the concentration of H₂S was 3.5%, it took nine years and two months from the start of injection to reach the maximum pressure, and with a concentration of 20%, it took nine years and six months. These observations reveal that, compared to pure CO₂, methane accelerates the reaching of the maximum pressure, whereas the presence of H₂S delays the time required to reach the maximum pressure relative to pure CO₂. This is because methane causes a significant decrease in density and an increase in the volume it occupies within the gas mixture, resulting in a rapid increase in pressure. In contrast, the volume fraction of H₂S decreases, and the pressure increase within the aquifer is relatively slow.

The cumulative gas mass tended to decrease after the rate of increase after BHP reached the maximum pressure. When the methane concentration was 3.5%, the cumulative gas injection at the maximum pressure was 97.6% compared to that of pure CO₂; however, at the end of the injection, it decreased to 96.31%. At a concentration of 20%, 87.21% of the gas was injected at the maximum pressure; at the end of the injection, it decreased to 79.82%. For H₂S at a concentration of 3.5%, the cumulative gas masses at the maximum pressure reaching point and at the end of the injections were 99.21% and 99.73%, respectively. Furthermore, at a H₂S concentration of 20%, the cumulative gas masses, compared to pure CO₂ at max pressure, were 96.60% and 97.86%, respectively. Based on these results, the cumulative injected gas mass varies depending on the time required to reach the maximum pressure, and the variation increases with impurity concentration.

The difference in CO₂ injection efficiency owing to impurities during the last 3 years before the end of the injection is described in

Table 7. Differences in CO₂ injection efficiency owing to impurity.

Parameter	1 January 2028	1 January 2029	1 January 2030
CO2 96.5% + CH4 3.5%	96.5%	96.1%	95.04%
CO2 96.5% + H2S 3.5%	96.5%	96.5%	96.91%
CO2 80% + CH4 20%	79.9%	76.2%	73.16%
CO2 80% + H2S 20%	80.8%	80.8%	81.99%

. Two years before the end of the injection, on 1 January 2028, when the impurity concentration was 3.5%, the CO₂ injection efficiency was 96.5% regardless of the type of impurity. However, when the impurity concentration was set to 20%, including methane, it resulted in an efficiency of 79.7%, and including H₂S, it resulted in an efficiency of 80.8%, showing a difference of within 1% compared with the CO₂ injection efficiency without impurities. One year before the end of the injection (1 January 2029), the BHP reached the maximum pressure when injecting pure CO₂. However, when injecting CO₂ mixed with methane, the BHP reached the maximum pressure earlier, resulting in reduced injection rates. Consequently, when the methane content was 3.5%, the CO₂ injection efficiency decreased to 96.1%, and when 20% methane was present, it decreased further to 76.2%. However, the CO₂ injection efficiency with H₂S as an impurity showed no significant change compared with that of the previous year. In 2030, when methane was present, the injection efficiency of pure CO₂ showed a gradual decline, with an efficiency of 95.4% at a 3.5% methane concentration and an efficiency of 73.16% at a 20% methane concentration. When H₂S was present as an impurity, BHP reached its maximum pressure between 2029 and 2030, after which the injection rate began to decrease. As a result, at the end of the injection period, when the H₂S concentration was 3.5%, the CO₂ injection efficiency increased to 96.91% compared with that of pure CO₂ and 81.99% when the concentration was 20%. Therefore, before reaching the maximum BHP pressure, the designed impurities were injected; however, after reaching that point, the storage efficiency varied based on the properties of the impurities.

4.3. Effects of Impurities on CO₂ Trapping Efficiency

The efficiency of CO₂ storage in the dissolution and residual traps was analyzed to examine the impact of impurities on CO₂ trapping. Figure 7 illustrates the variation in the efficiency of the residual trap. When methane concentration was 3.5%, the efficiency of the residual trap was 96.5% at the end of the injection but decreased to 95.7% after 100 years of monitoring. At a methane concentration of 20%, the efficiency of the residual trap was 83.3% at the end of the injection but decreased to 71.2% at the monitoring endpoint. However, in the case of a H₂S concentration of 3.5%, the efficiency of the residual trap was 96.7% at the end of the injection and increased to 97.1% at the monitoring endpoint. At a concentration of 20%, the efficiency was 80.8% at the end of the injection but increased to 82.3% at the monitoring endpoint. The efficiency of the residual trap consistently decreased when methane was included and showed an increasing trend with H₂S. At a concentration of 20%, the efficiency of the residual trap for methane was higher than that for H₂S at the end of the injection; however, at the monitoring endpoint, the efficiency of the H₂S trap was higher.

The efficiencies of the dissolution traps are shown in Figure 8. When methane was present at 3.5%, the dissolution trap efficiency was 94.2% at the end of the injection and remained constant at 94.2% at the monitoring endpoint. When the methane concentration was 20%, the efficiency was 75.7% at both the end of injection and the monitoring endpoint. In the case of H₂S at a concentration of 3.5%, the dissolution trap showed an efficiency of 97.8% at the end of the injection, which decreased slightly to 97.5% at the monitoring endpoint. For H₂S at a concentration of 20%, the efficiency remained the same at both the end of injection and the monitoring endpoint at 86.2%. H₂S exhibited a higher dissolution trap efficiency than methane, and unlike the residual trap, it showed minimal changes over time.

The results can be interpreted based on the behavior of gases in the aquifer. Figure 9 illustrates the gas saturation within the aquifer. When methane was present as an impurity, the gas saturation near the injection well was lower than that of pure CO₂, and owing to the density difference of gases, methane moved faster to the upper part of the aquifer, reducing the contact time between the gas and formation water, leading to a decrease in the solubility effect [31]. Furthermore, when methane was included, the residual gas saturation was lower than that with pure CO₂ because the presence of methane reduced the volume fraction and viscosity of CO₂ in the gas mixture, negatively affecting the residual trapping efficiency. However, when H₂S was included, although the density decreased, the volume fraction of CO₂ within the gas increased. Because of the higher viscosity compared to that of pure CO₂, the residual trapping efficiency in the residual trap was higher.

5. Discussion

In the hydrogen production process, impurities, such as H₂S and methane, are present in the CO₂ stream. We conducted simulations that considered the properties and injection of these impurities into the CO₂ stream under aquifer conditions. When non-condensable gases such as methane were included as impurities, compared to pure CO₂, significant differences occurred, whereas the presence of H₂S did not cause significant differences. Additionally, during the aquifer injection simulations, when methane was included, the BHP reached the maximum pressure more rapidly than with pure CO₂, whereas H₂S exhibited a delayed effect. This difference increased as impurity concentration increased and affected the injection rate and total gas volume injected after reaching the maximum pressure. Moreover, when methane was included as an impurity, the density, viscosity, and volume ratio changed, leading to reduced efficiency in both residual and solubility trapping, which can permanently store CO₂. Based on these findings, for CCS projects involving CO₂ streams with methane impurities, separation techniques, such as membrane systems, may be necessary for impurity removal. However, if H₂S is present as an impurity, it may not significantly impact the storage efficiency, indicating that facilities such as desulfurization units may not be required for CO₂ streams with H₂S impurities. The application of CCS to carbon dioxide emissions containing H₂S as an impurity reduced additional costs and ensured stable carbon storage. Additionally, the presence of H₂S did not significantly affect CCS efficiency in our study. Therefore, it enables economically feasible CCS operations.

6. Conclusions

In this study, CO₂ storage efficiency was analyzed when CO₂, along with impurities such as methane and H₂S, which are captured by CO₂ in the hydrogen production process, was injected into a deep saline aquifer. Fluid modeling was conducted by mixing methane and H₂S with CO₂, followed by the simulation of injection into a 3D saline aquifer model. The following conclusions were drawn from this study:

• Impurities affected the CO₂ stream under initial reservoir conditions. In the case of methane, the density decreases compared to that of pure CO₂, and its volume ratio increases, whereas the viscosity decreases. In contrast, a smaller change in density is shown in H₂S than in methane, and its volume decreases relative to the injection concentration, whereas, compared to pure CO₂, its viscosity increases;
• When the same amount of gas is injected into the reservoir, reaching the pressure limit of the reservoir occurs more rapidly when methane is present, while that with H₂S takes longer than that with pure CO₂. The difference in the time required to reach the pressure limit increases with impurity concentration. These phenomena affect the amount of CO₂ injected into the stream;
• Once the pressure limit is reached, the injection amounts differ, leading to variations in CO₂ storage. Methane present at 3.5% resulted in 95.04% CO₂ injection, and at 20%, only 73.16% CO₂ was injected. In contrast, with H₂S at 3.5%, 96.91% CO₂ was injected, and at 20%, 81.99% CO₂ was injected. Consequently, methane stores less CO₂ relative to the mixture ratio, whereas H₂S stores more;
• Methane results in a relatively low density, reducing the contact time between the brine and CO₂ and decreasing the dissolution effect. The CO₂ volume ratio within the gas mixture also decreases, leading to a reduced amount of CO₂ that can be retained in the rock pores;
• When storing CO₂ underground, methane causes a rapid increase in BHP and decreases storage efficiency. However, if H₂S is not removed and injection occurs with CO₂, the BHP increases more slowly than with the pure CO₂ injection. Additionally, the volume ratio of CO₂ in the mixed gas decreases, leading to improved storage efficiency. For CCS projects, methane impurities in the CO₂ stream require removal. In contrast, H₂S may not require steps such as desulfurization because it does not significantly impact CO₂ storage efficiency. Employing CCS for carbon dioxide emissions containing H₂S as an impurity can enable economically viable operations by reducing additional costs.