Numerical Study on CO2-Brine-Rock Interaction of Enhanced Geothermal Systems with CO2 as Heat Transmission Fluid

Enhanced Geothermal Systems (EGS) with CO2 instead of water as heat transmission fluid is an attractive concept for both geothermal resources development and CO2 geological sequestration. Previous studies show that CO2 has lots of favorable properties as heat transmission fluid and also can offer geologic storage of CO2 as an ancillary benefit. However, after CO2 injection into geological formations, chemical reaction between brine and rock can change chemical characteristics of saline and properties of rock such as porosity and permeability. Is this advantage or disadvantage for EGS operating? To answer this question, we have performed chemically reactive transport modeling to investigate fluid-rock interactions and CO2 mineral carbonation of Enhanced Geothermal Systems (EGS) site at Desert Peak (Nevada) operated with CO2. The simulation results show that (1) injection CO2 can create a core zone fulfilled with CO2 as main working domain for EGS, and (2) CO2 storage can induced self-enhancing alteration of EGS.


Introduction
With the increasing concern of Enhanced (or engineered) Geothermal Systems (EGS) and carbon dioxide sequestration in deep formation, Donald W. Brown [1] firstly proposed a new concept of an EGS system that may be operated with supercritical CO 2 instead of water as heat transmission fluid. Such a scheme could combine recovery of geothermal energy with simultaneous geologic storage of CO 2 . In comparison with water, Brown [1] noted that CO 2 has three favorable properties: large expansivity, lower viscosity, and much more stable than water as a solvent for rock minerals. He pointed out there are still unfavorable properties, such as lower heat conductivity of CO 2 . In recent years, many studies have been undertaken to investigate the CO 2 -brine-rock reaction, heat transmission, and rock mechanics of EGS with CO 2 [2,3,4,5,6,9] . There are not only the advantages of CO 2 -EGS which include increased heat extraction rates and wellbore flow compared to water-based systems and lesser potential for unfavorable rock-fluid chemical interactions [6,7] , but also the disadvantages that includes the solubility of CO 2 in water has obvious implications for long-term carbon sequestration and water-rock interactions [2,8,10] . The solubility of water into CO 2 is very important because it affects the reaction of CO 2 with surrounding rocks and the change of porosity and permeability of rock. These effects depend on the time for removal of water and the lithology of rock. Therefore, it is curial to study geochemical processes and associated changes in permeability and porosity of a CO 2 -EGS system. A great deal of specific and detailed information is required to assess the geochemistry of such a system. In this study, data for mineralogical composition were taken from the Desert Peak (Nevada) geothermal field [11] . The study is not intended to represent complete flow system of Desert Peak geothermal field, but to understand geochemical process of CO 2 -EGS at this filed.

Geometric and Flow Conditions
In our simulation, a 10 m thick reservoir formation with an injection well was modeled. A one-dimension radial flow model (Fig.1.) was used, consisting of 60 radial blocks with logarithmically increasing radii out to a distance 10000 m. A initial homogeneous porous medium is considered in the model, with the assumption that the fluid exchange with the surrounding low permeability matrix is insignificant. The uncertainty on the permeability specification does not affect modeling results of reactive transport and porosity changes (as long as pressure buildup at the wellbore can be afforded) because a constant injection rate was specified in the present study. Conductive heat exchange with rocks of low permeability above and below simulation domain is an important process when injection temperature differs from the reservoir temperature. The confining layers are modeled as semi-infinite half spaces, and heat exchange is treated with a semi-analytical technique due to Vinsome&Westerveld (1980). Initial reservoir temperature is 210°C and initial hydrostatic pressure is 200 bar (Fig.1.). Hydrogeologic specifications of the one-dimensional radial problem are given in Table.1.

Mineralogical Composition and Reaction Kinetics
The initial mineralogical composition used in the modeling is summarized in Table 2. The composition specified was based on the mineralogy of the crystalline rock of pre-Tertiary unit 2 (pT2) from well DP 23-1 of DesertPeak (Nevada) geothermal field, as previously used by Xu et al. (2009). Plagioclase was modeled using 50% low-albite and 50% anorthite. Other trace minerals including epidote, pyrite, and biotite were not considered in the model, because their reactions with the injected fluid are slow and not important for the chemical purpose. Details of the primary mineral composition can be found in Xu et al. (2009).Dolomite, siderite, ankerite, dawsonite, smectite, and kaolinite could be formed after CO 2 injection and are specified as secondary minerals in the simulations. Parameters for the kinetics of mineral reaction used in the mode were taken from Xu et al. (2009 and 2010). The initial waterchemistry is in equilibrium with the initial mineralogy at areservoir temperature of 210°C.

Simulation Tool
The present simulations employed the nonisothermal reactive transport code TOUGHREACT (Xu et al., 2006). This code introduces reactive chemistry into the multiphase fluid and heat flow code TOUGH2 (Pruess, 2004). A new fluid property module, ECO2H, based on Spycher and Pruess (2010) work was used. This module provides an accurate description of the thermophysical properties of mixtures of water and CO 2 under conditions typically encountered in CO 2 -EGS systems of interest (12°Cd T d300°C; 1 bar < P d 600 bar). Further details on the process capabilities of the TOUGHREACT code are given in Xu et al. (2006).  Fig.2 shows CO 2 saturation as a function of radial distance at different times. Three different zones are formed: (1) a central zone or "core" in which all aqueous phase has been removed by dissolution into the flowing CO 2 stream, so that the reservoir fluid is a single supercritical CO 2 phase; (2) a surrounding intermediate zone, in which the reservoir fluid consists of a two-phase brine-CO 2 mixture; and (3) an outer zone, in which the reservoir fluid is a single aqueous phase with dissolved CO 2 .In Zone 1, due to only containing supercritical CO 2 , the chemical interactions are weak and the change porosity and permeability are very small. It is the major zone for heat exchange with rock around.

Results and Discussion
In Zone 2, pH decrease to lower than 5due to dissolved CO 2 . The lower pH causes albite, anorthite, K-feldspar, quartz and chlorite dissolve, and calcite, kaolinite, illite, Na-smectite, Ca-smectite, dolomite, ankerite, and siderite precipitate (Fig.3.). Consequently, the porosity and permeability decrease significantly near the left boundary of Zone 2 (Fig.4.). This is due to (1) CO 2 -rich water adds solute mass to the rock matrix and (2) molar volumes of secondary clays are larger than those of the primary minerals. The extent of porosity change is consistent with the mineral alteration mentioned above. The dissolution fronts of primary minerals move over time and occur near the CO 2 source and decreases away from it. The precipitation fronts of secondary minerals also move over time. Precipitation of dolomite, ankerite, and siderite sequesters a significant amount of CO 2 . A maximum of about 36kg/m 3 of medium were trapped by carbonate precipitation. This zone is the major area that CO 2 is captured by minerals.
Zone 3 only contains water with dissolved CO 2 . In this area, the brine-rock interactions are strong due to lower pH, including albite, anorthite, chlorite, K-feldspar and quartz precipitate, and calcite, kaolinite, illite, Na-smectite, Ca-smectite, dolomite, ankerite and siderite dissolve. As a feedback effect, the porosity and permeability increase.

Conclusions
Major findings from our simulation are: In Zone 1, rock-fluid interactions for EGS with CO 2 are weak, which may be more favorable than for water-based EGS but little information is available about chemical interactions between supercritical CO 2 (containing water vapor) and rock minerals.
In Zones 2 and 3, dissolution of primary minerals and precipitation of secondary carbonate and clay minerals are very significant. For Desert Peak mineralogy, the principal minerals undergoing dissolution are albite, anorthite, K-feldspar, quartz and chlorite. Major precipitating secondary minerals are calcite, kaolinite, illite, Na-smectite, Ca-smectite, dolomite, and siderite. Significant CO 2 could be fixed through precipitation of carbonate minerals, which can offer geologic storage of carbon as an ancillary benefit.
Mineral alteration due to dissolved CO 2 generally decreases porosity and permeability which could form a seal that would impact reservoir growth and longevity.