Sensitivity Study of Simulation Parameters Controlling CO<Subscript>2</Subscript> Trapping Mechanisms in Saline Formations

The primary purpose of this study is to understand quantitative characteristics of mobile, residual, and dissolved CO₂ trapping mechanisms within ranges of systematic variations in different geologic and hydrologic parameters. For this purpose, we conducted an extensive suite of numerical simulations to evaluate the sensitivities included in these parameters. We generated two-dimensional numerical models representing subsurface porous media with various permutations of vertical and horizontal permeability (k _v and k _h), porosity (f{\phi}), maximum residual CO₂ saturation (S_gr^max{S_{\rm gr}^{\max}}), and brine density (ρ _br). Simulation results indicate that residual CO₂ trapping increases proportionally to k_v, k_h, S_gr^max{k_{\rm v}, k_{\rm h}, S_{\rm gr}^{\max}} and ρ _br but is inversely proportional to f.{\phi.} In addition, the amount of dissolution-trapped CO₂ increases with k _v and k _h, but does not vary with f{\phi } , and decreases with S_gr^max{S_{\rm gr}^{\max}} and ρ _br. Additionally, the distance of buoyancy-driven CO₂ migration increases proportionally to k _v and ρ _br only and is inversely proportional to k_h, f{k_{\rm h}, \phi } , and S_gr^max{S_{\rm gr}^{\max}} . These complex behaviors occur because the chosen sensitivity parameters perturb the distances of vertical and horizontal CO₂ plume migration, pore volume size, and fraction of trapped CO₂ in both pores and formation fluids. Finally, in an effort to characterize complex relationships among residual CO₂ trapping and buoyancy-driven CO₂ migration, we quantified three characteristic zones. Zone I, expressing the variations of S_gr^max{S_{\rm gr}^{\max}} and k _h, represents the optimized conditions for geologic CO₂ sequestration. Zone II, showing the variation of f{\phi} , would be preferred for secure CO₂ sequestration since CO₂ has less potential to escape from the target formation. In zone III, both residual CO₂ trapping and buoyancy-driven migration distance increase with k _v and ρ _br.     

STREAMLINE-BASED MATHEMATICAL MODEL FOR CO2 MISCIBLE FLOODING

According to the research theory of improved black oil simulator, a practical mathematical model for C02 miscible flooding was presented. In the model, the miscible process simulation was realized by adjusting oil/gas relative permeability and effective viscosity under the condition of miscible flow. In order to predict the production performance fast, streamline method is employed to solve this model as an alternative to traditional finite difference methods. Based on streamline distribution of steady-state flow through porous media with complex boundary confirmed with the boundary element method (BEM), an explicit total variation diminishing (TVD) method is used to solve the one-dimensional flow problem. At the same time, influences of development scheme, solvent slug size, and injection periods on CO2 drive recovery are discussed. The model has the advantages of less information need, fast calculation, and adaptation to calculate CO2 drive performance of all kinds of patterns in a random shaped porous media with assembly boundary. It can be an effective tool for early stage screening andmiscible oil field.reservoir dynamic management of the CO2 miscible oil field.     

Modeling of Rocks and Cement Alteration due to CO<Subscript>2</Subscript> Injection in an Exploited Gas Reservoir

The injection of CO₂ in exploited natural gas reservoirs as a means to reduce greenhouse gas (GHG) emissions is highly attractive as it takes place in well-known geological structures of proven integrity with respect to gas leakage. The injection of a reactive gas such as CO₂ puts emphasis on the possible alteration of reservoir and caprock formations and especially of the wells’ cement sheaths induced by the modification of chemical equilibria. Such studies are important for injectivity assurance, wellbore integrity, and risk assessment required for CO₂ sequestration site qualification. Within a R&D project funded by Eni, we set up a numerical model to investigate the rock–cement alterations driven by the injection of CO₂ into a depleted sweet natural gas pool. The simulations are performed with the TOUGHREACT simulator (Xu et al. in Comput Geosci 32:145–165, 2006) coupled to the TMGAS EOS module (Battistelli and Marcolini in Int J Greenh Gas Control 3:481–493, 2009) developed for the TOUGH2 family of reservoir simulators (Pruess et al. in TOUGH2 User’s Guide, Version 2.0, 1999). On the basis of field data, the system is considered in isothermal (50°C) and isobaric (128.5 bar) conditions. The effects of the evolving reservoir gas composition are taken into account before, during, and after CO₂ injection. Fully water-saturated conditions were assumed for the cement sheath and caprock domains. The gas phase does not flow by advection from the reservoir into the interacting domains so that molecular diffusion in the aqueous phase is the most important process controlling the mass transport occurring in the system under study.     

Modeling of CO<Subscript>2</Subscript> Leakage up Through an Abandoned Well from Deep Saline Aquifer to Shallow Fresh Groundwaters

This article presents a numerical modeling application using the code TOUGHREACT of a leakage scenario occurring during a CO₂ geological storage performed in the Jurassic Dogger formation in the Paris Basin. This geological formation has been intensively used for geothermal purposes and is now under consideration as a site for the French national program of reducing greenhouse gas emissions and CO₂ geological storage. Albian sandstone, situated above the Dogger limestone is a major strategic potable water aquifer; the impacts of leaking CO₂ due to potential integrity failure have, therefore, to be investigated. The present case–study illustrates both the capacity and the limitations of numerical tools to address such a critical issue. The physical and chemical processes simulated in this study have been restricted to: (i) supercritical CO₂ injection and storage within the Dogger reservoir aquifer, (ii) CO₂ upwards migration through the leakage zone represented as a 1D vertical porous medium to simulate the cement–rock formation interface in the abandoned well, and (iii) impacts on the Albian aquifer water quality in terms of chemical composition and the mineral phases representative of the porous rock by estimating fluid–rock interactions in both aquifers. Because of CPU time and memory constraints, approximation and simplification regarding the geometry of the geological structure, the mineralogical assemblages and the injection period (up to 5 years) have been applied to the system, resulting in limited analysis of the estimated impacts. The CO₂ migration rate and the quantity of CO₂ arriving as free gas and dissolving, firstly in the storage water and secondly in the water of the overlying aquifer, are calculated. CO₂ dissolution into the Dogger aquifer induces a pH drop from about 7.3 to 4.9 limited by calcite dissolution buffering. Glauconite present in the Albian aquifer also dissolves, causing an increase of the silicon and aluminum in solution and triggering the precipitation of kaolinite and quartz around the intrusion point. A sensitivity analysis of the leakage rate according to the location of the leaky well and the variability of the petro-physical properties of the reservoir, the leaky well zone and the Albian aquifers is also provided.     

A collisional-radiative code for computing air plasma in high enthalpy flow

A time-dependent collisional-radiative model for air plasma has been developed to study the effects of nonequilibrium atomic and molecular processes on population densities in a weakly ionized high enthalpy flow. This model consists of 15 species: e^-,N, N⁺,N²⁺,O, O⁺,O²⁺,O^-,N₂,N₂⁺,NO, NO⁺,O₂,O₂⁺{{\rm e}^{-},{\rm N, N}^{+},{\rm N}^{2+},{\rm O, O}^{+},{\rm O}^{2+},{\rm O}^{-},{\rm N}_{2},{{\rm N}_{2}}^{+},{\rm NO, NO}^{+},{\rm O}_{2},{{\rm O}_{2}}^{+}}, and O₂^-{{{\rm O}_{2}}^{-}} with their major electronic excited states. Many elementary processes are considered in the number density range of 10¹²/cm³ ≤ N ≤ 10¹⁹/cm³ and the temperature range of 300 K ≤ T ≤ 40,000 K. We then compare our results with an existing collisional-radiative code to validate our model. Additionally, the unsteady nature of pulsively heated air plasma is investigated. When the ionization relaxation time is of the same order as the time scale of a heating pulse, the effects of unsteady ionization are important for estimating air plasma states. From parametric computations, we determine the appropriate conditions for the collisional-radiative steady state, local thermodynamic equilibrium, and corona equilibrium models in that density and temperature range.     

Transport of Ca+2 and $${{\rm SO}_{4}^{-2}}$$ Ions in Porous Media of Clay Product and Its Association with the Efflorescence Phenomenon

This study investigates the displacement of SO₄^-2{{\rm SO}_{4}^{-2}} and Ca⁺² ions in a red-clay ceramic, simulating the process of efflorescence. Ceramic bodies were molded (70 × 27 × 9mm³) by vacuum extrusion formulated with different contents of CaSO₄· 2H₂O (0, 2, 4, 8, and 16% in weight) and burnt at different temperatures (800, 850, 900, and 950°C) for 12 h. Ceramic bodies were characterized in terms of water absorption, apparent porosity and pore size distribution. Efflorescence was evaluated according to the norms of ASTM C67/2003 and by testing the solubilization of SO₄^-2{{\rm SO}_{4}^{-2}} and Ca⁺² ions after 1 h with the ceramic bodies immersed in hot water as well as after 7, 14, and 28 consecutive days with the ceramic bodies immersed in cold water. In the quantification of efflorescence, a new image analysis methodology was developed by using the graphic software Image Tools 3.0. The results allowed in establishing a relationship between the efflorescence of the investigated ions, physical properties (water absorption and apparent porosity), pore size distribution, and solubilization.     

Transport of Ca+2 and $${{\rm SO}_{4}^{-2}}$$ ions in Porous Media of Clay Product and its Association with the Efflorescence Phenomenon

This study investigates the displacement of SO₄^-2{{\rm SO}_{4}^{-2}} and Ca⁺² ions in a red-clay ceramic, simulating the process of efflorescence. Ceramic bodies were molded (70 × 27 × 9 mm³) by vacuum extrusion formulated with different contents of CaSO₄ · 2H₂O (0, 2, 4, 8, and 16% in weight) and burnt at different temperatures (800, 850, 900, and 950°C) for 12 h. Ceramic bodies were characterized in terms of water absorption, apparent porosity, and pore size distribution. Efflorescence was evaluated according to the norms of ASTM C67/2003 and by testing the solubilization of SO₄^-2{{\rm SO}_{4}^{-2}} and Ca⁺² ions after 1 h with the ceramic bodies immersed in hot water as well as after 7, 14, and 28 consecutive days with the ceramic bodies immersed in cold water. In the quantification of efflorescence, a new image analysis methodology was developed by using the graphic software Image Tools 3.0. The results allowed in establishing a relationship between the efflorescence of the investigated ions, physical properties (water absorption and apparent porosity), pore size distribution, and solubilization.     

Quantification of Density-Driven Natural Convection for Dissolution Mechanism in CO<Subscript>2</Subscript> Sequestration

Dissolution of CO₂ into brine causes the density of the mixture to increase. The density gradient induces natural convection in the liquid phase, which is a favorable process of practical interest for CO₂ storage. Correct estimation of the dissolution rate is important because the time scale for dissolution corresponds to the time scale over which free phase CO₂ has a chance to leak out. However, for this estimation, the challenging simulation on the basis of convection–diffusion equation must be done. In this study, pseudo-diffusion coefficient is introduced which accounts for the rate of mass transferring by both convection and diffusion mechanisms. Experimental tests in fluid continuum and porous media were performed to measure the real rate of dissolution of CO₂ into water during the time. The pseudo diffusion coefficient of CO₂ into water was evaluated by the theory of pressure decay and this coefficient is used as a key parameter to quantify the natural convection and its effect on mass transfer of CO₂. For each experiment, fraction of ultimate dissolution is calculated from measured pressure data and the results are compared with predicted values from analytical solution. Measured CO₂ mass transfer rate from experiments are in reasonable agreement with values calculated from diffusion equation performed on the basis of pseudo-diffusion coefficient. It is suggested that solving diffusion equation with pseudo diffusion coefficient herein could be used as a simple and rapid tool to calculate the rate of mass transfer of CO₂ in CCS projects.     

Pressure Buildup During CO<Subscript>2</Subscript> Injection into a Closed Brine Aquifer

CO₂ injected into porous formations is accommodated by reduction in the volume of the formation fluid and enlargement of the pore space, through compression of the formation fluids and rock material, respectively. A critical issue is how the resulting pressure buildup will affect the mechanical integrity of the host formation and caprock. Building on an existing approximate solution for formations of infinite radial extent, this article presents an explicit approximate solution for estimating pressure buildup due to injection of CO₂ into closed brine aquifers of finite radial extent. The analysis is also applicable for injection into a formation containing multiple wells, in which each well acts as if it were in a quasi-circular closed region. The approximate solution is validated by comparison with vertically averaged results obtained using TOUGH2 with ECO2N (where many of the simplifying assumptions are relaxed), and is shown to be very accurate over wide ranges of the relevant parameter space. The resulting equations for the pressure distribution are explicit, and can be easily implemented within spreadsheet software for estimating CO₂ injection capacity.     

On the Domain of Validity of Brinkman’s Equation

An increasing number of articles are adopting Brinkman’s equation in place of Darcy’s law for describing flow in porous media. That poses the question of the respective domains of validity of both laws, as well as the question of the value of the effective viscosity μ ^e which is present in Brinkman’s equation. These two topics are addressed in this article, mainly by a priori estimates and by recalling existing analyses. Three main classes of porous media can be distinguished: “classical” porous media with a connected solid structure where the pore surface S _p is a function of the characteristic pore size l _p (such as for cylindrical pores), swarms of low concentration fixed particles where the pore surface is a function of the characteristic particle size l _s , and fiber-made porous media at low solid concentration where the pore surface is a function of the fiber diameter. If Brinkman’s 3D flow equation is valid to describe the flow of a Newtonian fluid through a swarm of fixed particles or fibrous media at low concentration under very precise conditions (Lévy 1983), then we show that it cannot apply to the flow of such a fluid through classical porous media.     

A Discussion of the Effect of Tortuosity on the Capillary Imbibition in Porous Media

In the past decades, there was considerable controversy over the Lucas–Washburn (LW) equation widely applied in capillary imbibition kinetics. Many experimental results showed that the time exponent of the LW equation is less than 0.5. Based on the tortuous capillary model and fractal geometry, the effect of tortuosity on the capillary imbibition in wetting porous media is discussed in this article. The average height growth of wetting liquid in porous media driven by capillary force following the [`(L)] _s(t) ~ t^1/2D_T{\overline L _{\rm {s}}(t)\sim t^{1/{2D_{\rm {T}}}}} law is obtained (here D _T is the fractal dimension for tortuosity, which represents the heterogeneity of flow in porous media). The LW law turns out to be the special case when the straight capillary tube (D _T = 1) is assumed. The predictions by the present model for the time exponent for capillary imbibition in porous media are compared with available experimental data, and the present model can reproduce approximately the global trend of variation of the time exponent with porosity changing.     

Anomalous Reactive Transport in the Framework of the Theory of Chromatography

The anomalous reactive transport considered here is the migration of contaminants through strongly sorbing permeable media without significant retardation. It has been observed in the case of heavy metals, organic compounds, and radionuclides, and it has critical implications on the spreading of contaminant plumes and on the design of remediation strategies. Even in the absence of the well-known fast migration pathways, associated with fractures and colloids, anomalous reactive transport arises in numerical simulations of reactive flow. It is due to the presence of highly pH-dependent adsorption and the broadening of the concentration front by hydrodynamic dispersion. This leads to the emergence of an isolated pulse or wave of a contaminant traveling at the average flow velocity ahead of the retarded main contamination front. This wave is considered anomalous because it is not predicted by the classical theory of chromatography, unlike the retardation of the main contamination front. In this study, we use the theory of chromatography to study a simple pH-dependent surface complexation model to derive the mathematical framework for the anomalous transport. We analyze the particular case of strontium (Sr²⁺) transport and define the conditions under which the anomalous transport arises. We model incompressible one-dimensional (1D) flow through a reactive porous medium for a fluid containing four aqueous species: H⁺, Sr²⁺, Na⁺, and Cl⁻. The mathematical problem reduces to a strictly hyperbolic 2 × 2 system of conservation laws for effective anions and Sr²⁺, coupled through a competitive Langmuir isotherm. One characteristic field is linearly degenerate while the other is not genuinely nonlinear due to an inflection point in the pH-dependent isotherm. We present the complete set of analytical solutions to the Riemann problem, consisting of only three combinations of a slow wave comprising either a rarefaction, a shock, or a shock–rarefaction with fast wave comprising only a contact discontinuity. Highly resolved numerical solutions at large Péclet numbers show excellent agreement with the analytic solutions in the hyperbolic limit. In the Riemann problem, the anomalous wave forms only if: hydrodynamic dispersion is present, the slow wave crosses the inflection locus, and the effective anion concentration increases along the fast path.     

An Experimental Study on the Influence of Sub-Core Scale Heterogeneities on CO<Subscript>2</Subscript> Distribution in Reservoir Rocks

This article presents the results of CO₂/brine two-phase flow experiments in rocks at reservoir conditions. X-ray CT scanning is used to determine CO₂ saturation at a fine scale with a resolution of a few pore volumes and provide 3D porosity and saturation maps that can be use to correlate CO₂ saturations and rock properties. The study highlights the strong influence of sub-core scale heterogeneities on the spatial distribution of CO₂ at steady state and provides useful relative permeability data on a sample originated from an actual storage site (CO2CRC-Otway project, Victoria, South-West Australia). Two different samples tested, although different in nature, present strong heterogeneities, but differ in the detail of the connectivity of high porosity layers. In both samples, the results of the investigations show that sub-core scale heterogeneities control the sweep efficiency and may cause channeling through the porous medium. In one of the samples, CO₂ saturation appears uncorrelated to porosity close to the outlet end of the core. This observation is understood as a result of the position and the orientation of high porosity layers with respect to the inlet face of the core. Finally, in the operating conditions of the two experiments, the saturation maps demonstrate that gravity does not play a major role since no detectable buoyancy driven flow is observed.     

CO2 Injection in Geological Formations: Determining Macroscale Coefficients from Pore Scale Processes

Carbon dioxide (CO₂) injections in geological formations are usually performed for enhanced hydrocarbon recovery in oil and gas reservoirs and storage and sequestration in saline aquifers. Once CO₂ is injected into the formation, it propagates in the porous rock by dispersion and convection. Chemical reactions between brine ions and CO₂ molecules and consequent reactions with mineral grains are also important processes. The dynamics of CO₂ molecules in random porous media are modeled with a set of differential equations corresponding to pore scale and continuum macroscale. On the pore scale, convective–dispersive equation is solved considering reactions on the inner boundaries in a unit cell. A unit cell is the smallest portion of a porous media that can reproduce the porous media by repetition. Inner boundaries in a unit cell are the surfaces of the mineral grains. Dispersion process at the pore scale is transformed into continuum macroscale by adopting periodic boundary conditions for contiguous unit cells and applying Taylor-Aris dispersion theory known as macrotransport theory. Using this theory, the discrete porous system changes into a continuum system within which the propagation and interaction of CO₂ molecules with fluid and solid matrix of the porous media are characterized by three position-independent macroscopic coefficients: the mean velocity vector , dispersivity dyadic , and mean volumetric CO₂ depletion coefficient .     

Effect of Vertical Heterogeneity on Long-Term Migration of CO<Subscript>2</Subscript> in Saline Formations

For deep injection of CO₂ in thick saline formations, the movements of both the free gas phase and dissolved CO₂ are sensitive to variations in vertical permeability. A simple model for vertical heterogeneity was studied, consisting of a random distribution of horizontal impermeable barriers with a given overall volume fraction and distribution of lengths. Analytical results were obtained for the distribution of values for the permeability, and compared to numerical simulations of deep CO₂ injection and convection in heterogeneous formations, using multiple realizations for the permeability distribution. It is shown that for a formation of thickness H, the breakthrough times in two dimensions for deep injection scale as H ² for moderate injection rates. In comparison to heterogeneous shale distributions, a homogeneous medium with equivalent effective vertical permeability has a longer breakthrough time for deep injection, and a longer onset time for convection.     

A study of loop heat pipe with biporous wicks

The purpose of this study is to investigate the effects of various bimodal pore size distributions of biporous wicks for a loop heat pipe (LHP). The study was conducted following a statistical method using a two-level factorial plan involving three variables (particle size of pore former:74–88 and 125–149 μm Na₂CO₃, pore former content:20% by volume and 25% by volume, sintering temperature:700 and 750°C). Finally, the heat transport capability of the LHP between monoporous wicks and biporous wicks has been investigated. Experimental results show that, at the sink temperature of 10°C and the allowable evaporator temperature of 80°C, the heat transfer capacity of the better biporous wick achieved 200 W and the total thermal resistance was 0.31°C/W. The performance is enhanced about 60%, compared to a monoporous wick for 125 W and 0.53°C/W. Therefore, LHPs with biporous wicks are very attractive for high heat flux applications in the future.     

Modeling Gas Transport in the Shallow Subsurface During the ZERT CO<Subscript>2</Subscript> Release Test

We used the multiphase and multicomponent TOUGH2/EOS7CA model to carry out predictive simulations of CO₂ injection into the shallow subsurface of an agricultural field in Bozeman, Montana. The purpose of the simulations was to inform the choice of CO₂ injection rate and design of monitoring and detection activities for a CO₂ release experiment. The release experiment configuration consists of a long horizontal well (70 m) installed at a depth of approximately 2.5 m into which CO₂ is injected to mimic leakage from a geologic carbon sequestration site through a linear feature such as a fault. We estimated the permeability of the soil and cobble layers present at the site by manual inversion of measurements of soil CO₂ flux from a vertical-well CO₂ release. Based on these estimated permeability values, predictive simulations for the horizontal well showed that CO₂ injection just below the water table creates an effective gas-flow pathway through the saturated zone up to the unsaturated zone. Once in the unsaturated zone, CO₂ spreads out laterally within the cobble layer, where liquid saturation is relatively low. CO₂ also migrates upward into the soil layer through the capillary barrier and seeps out at the ground surface. The simulations predicted a breakthrough time of approximately two days for the 100kg d⁻¹ injection rate, which also produced a flux within the range desired for testing detection and monitoring approaches. The seepage area produced by the model was approximately five meters wide above the horizontal well, compatible with the detection and monitoring methods tested. For a given flow rate, gas-phase diffusion of CO₂ tends to dominate over advection near the ground surface, where the CO₂ concentration gradient is large, while advection dominates deeper in the system.     

Analytical Solution Incorporating History-Dependent Processes for Quick Assessment of Capillary Trapping During CO<Subscript>2</Subscript> Geological Storage

Carbon storage in saline formations is considered as a promising option to ensure the necessary decrease of CO₂ anthropogenic emissions. Its industrial development in those formations is above all conditioned by its safety demonstration. Assessing the evolution of trapped and mobile CO₂ across time is essential in the perspective of reducing leakage risks. In this work, we focus on residual trapping phenomenon occurring during the wetting of the injected CO₂ plume. History dependent effects are of first importance when dealing with capillary trapping. We then apply the classical fractional flow theory (Buckley–Leverett type model) and include trapping and hysteresis models; we derive an analytical solution for the temporal evolution of saturation profile and of CO₂ trapped quantity when injecting water after the gas injection (“artificial imbibition”). The comparison to numerical simulations for different configurations shows satisfactory match and justifies, in the case of industrial CO₂ storage, the assumptions of incompressible flow with no consideration of capillary pressure. The obtained analytical solution allows the quick assessment of both the quantity and the location of mobile gas left during imbibition.     

An Experimental Study of CO<Subscript>2</Subscript> Exsolution and Relative Permeability Measurements During CO<Subscript>2</Subscript> Saturated Water Depressurization

Dissolution of CO₂ into brine is an important and favorable trapping mechanism for geologic storage of CO₂. There are scenarios, however, where dissolved CO₂ may migrate out of the storage reservoir. Under these conditions, CO₂ will exsolve from solution during depressurization of the brine, leading to the formation of separate phase CO₂. For example, a CO₂ sequestration system with a brine-permeable caprock may be favored to allow for pressure relief in the sequestration reservoir. In this case, CO₂-rich brine may be transported upwards along a pressure gradient caused by CO₂ injection. Here we conduct an experimental study of CO₂ exsolution to observe the behavior of exsolved gas under a wide range of depressurization. Exsolution experiments in highly permeable Berea sandstones and low permeability Mount Simon sandstones are presented. Using X-ray CT scanning, the evolution of gas phase CO₂ and its spatial distribution is observed. In addition, we measure relative permeability for exsolved CO₂ and water in sandstone rocks based on mass balances and continuous observation of the pressure drop across the core from 12.41 to 2.76 MPa. The results show that the minimum CO₂ saturation at which the exsolved CO₂ phase mobilization occurs is from 11.7 to 15.5%. Exsolved CO₂ is distributed uniformly in homogeneous rock samples with no statistical correlation between porosity and CO₂ saturation observed. No gravitational redistribution of exsolved CO₂ was observed after depressurization, even in the high permeability core. Significant differences exist between the exsolved CO₂ and water relative permeabilities, compared to relative permeabilities derived from steady-state drainage relative permeability measurements in the same cores. Specifically, very low CO₂ and water relative permeabilities are measured in the exsolution experiments, even when the CO₂ saturation is as high as 40%. The large relative permeability reduction in both the water and CO₂ phases is hypothesized to result from the presence of disconnected gas bubbles in this two-phase flow system. This feature is also thought to be favorable for storage security after CO₂ injection.