Article pubs.acs.org/est
Regional Assessment of CO2−Solubility Trapping Potential: A Case Study of the Coastal and Offshore Texas Miocene Interval Changbing Yang,* Ramón H. Treviño, Tongwei Zhang, Katherine D. Romanak, Kerstan Wallace, Jiemin Lu, Patrick J. Mickler, and Susan D. Hovorka Bureau of Economic Geology, The University of Texas at Austin, 10100 Burnet Road, Austin, Texas 78758, United States S Supporting Information *
ABSTRACT: This study presents a regional assessment of CO2-solubility trapping potential (CSTP) in the Texas coastal and offshore Miocene interval, comprising lower, middle, and upper Miocene sandstone. Duan’s solubility model [Duan et al. Mar. Chem. 2006, 98, 131−139] was applied to estimate carbon content in brine saturated with CO2 at reservoir conditions. Three approaches (simple, coarse, and fine) were used to calculate the CSTP. The estimate of CSTP in the study area varies from 30 Gt to 167 Gt. Sensitivity analysis indicated that the CSTP in the study area is most sensitive to storage efficiency, porosity, and thickness and is least sensitive to background carbon content in brine. Comparison of CSTP in our study area with CSTP values for seven other saline aquifers reported in the literature showed that the theoretical estimate of CO2-solubility trapping potential (TECSTP) has a linear relationship with brine volume, regardless of brine salinity, temperature, and pressure. Although more validation is needed, this linear relationship may provide a quick estimate of CSTP in a saline aquifer. Results of laboratory experiments of brine-rock−CO2 interactions and the geochemical model suggest that, in the study area, enhancement of CSTP caused by interactions between brine and rocks is minor and the storage capacity of mineral trapping owing to mineral precipitation is relatively trivial.
■
INTRODUCTION Carbon capture and storage (CCS) has been proposed as a viable technology to reduce the emission of greenhouse gases generated through combustion of fossil fuels.1−7 Although various deep geological media, such as depleted oil/gas reservoirs and unmineable coal seams, have been identified as possible geological storage targets, deep saline aquifers are of particular interest because of their wide distribution and huge storage potential.4,8 A saline aquifer for CO2 storage is generally specified as a porous and permeable body of rock capped by one or more regionally extensive, low-permeability rock formations (seals).9 Storage of CO2 in saline aquifers is optimal below 800 m, depths at which pressure and temperature maintain the CO2 in a high-density fluid or supercritical state.2 The long-term safety and performance of geological storage relies on various physical and chemical trapping mechanisms by which CO2 is sequestered. Physical trapping of CO2 occurs at a relatively early stage during and after CO2 injection, when CO2 is immobilized as a free gas or, preferably, a supercritical fluid.10 Chemical trapping occurs at a later stage, when CO2 dissolves into brine (solubility trapping), interacts with sedimentary rocks, and potentially is absorbed onto mineral surfaces.1,3,10 Because of the complexity of trapping mechanisms, which can occur spatially and temporally interdependently, assessment of CO2-storage potential (or capacity) in a saline aquifer is © 2014 American Chemical Society
particularly difficult. Published estimates of CO2-storage potential in deep saline aquifers focus mainly on physical trapping mechanisms.2,8,10,11 Much less attention has been given to regional assessment of CO2-solubility trapping potential (CSTP), although some limited studies, such as those of the Viking aquifer12 and Winnipegosis aquifer13 in the western Canada sedimentary basin, the Utsira Formation at Sleipner,14 the Oriskany Formation,15 and the Lower Yancheng and Upper Sanduo Formations in the Subei basin, East China,16 have been published. In these studies, CSTP values in deep saline aquifers were assessed using a volumetric method, assuming that brine in the pore space can be completely accessed by the injected CO2. We also used the volumetric method in this study to assess regional CSTP in the coastal and offshore Texas Miocene interval, one of the largest saline aquifers for CO2 storage in the United States. As a potential saline CO2 storage site, the Miocene interval has various advantages, including (1) high porosity and permeability in the sandstones; (2) thick regional shale intervals that provide potential seals; and (3) numerous depleted oil and gas fields Received: Revised: Accepted: Published: 8275
May 1, 2014 June 23, 2014 June 24, 2014 June 24, 2014 dx.doi.org/10.1021/es502152y | Environ. Sci. Technol. 2014, 48, 8275−8282
Environmental Science & Technology
Article
that may serve as initial storage or enhanced recovery targets.17 The aims of this study are to (1) estimate CSTP; (2) compare CSTP values in this study with those in other locations reported in the literature; (3) validate the CO2-solubility model for estimation of CSTP using laboratory experiments; and (4) assess impacts of interactions of brine, CO2, and rocks on CSTP using an approach that integrates laboratory experiments and numerical modeling.
temperature and pressure conditions at which density of CO2 is greater than 0.42 g/cm3 (supercritical state), generally at depths greater than 1006 m in the study area. The lower depth limit of the capacity interval was determined by the depth at which the pore water pressure is greater than the hydrostatic pressure, ranging from 1767 to 3688 m.17 The total thickness of the sandstone units within the Miocene capacity interval at each well location was obtained by adding the thicknesses of each individual sand unit, as shown in Figure s2b. On the basis of wireline log sandstone-thickness interpretations, we estimate that 40% of measured sand units are less than 2 m thick, comprising about 14% of total sand volume.17 Approximately 50% of total sand volume is attributed to units less than 8 m thick, suggesting that a significant portion of sand volume is found in relatively thin sand beds. Brine Compositions. Brine composition data in the study area was selected from the USGS produced-water database, a subset of a database originally provided to the USGS by GeoINFORMATION at the University of Oklahoma.20 The USGS produced-water database contains 58,706 records of brine compositions from 102 fields. Most samples were collected and analyzed from the 1950s through 1980s. Table s1 presents 66 brine composition records and statistics from samples in the study area. Brine salinity in the Miocene interval ranges from 4070 mg/L to 274,000 mg/L, with an average value of 119,106 mg/L. Na+ and Cl− are the two dominant ions in the brines (Figure s3a). Brine salinity shows a strong correlation with Na+ (or Cl−) concentration (Figure s3b). Brine pH ranges from 5.1 to 8.2, with an average value of 6.7. Background alkalinity of brine varies from 24 mg/L to 1110 mg/L, with an average value of 244 mg/L, showing no correlation with depth (Figure s3d). Other ions, such as Ca, Mg, and K, have much smaller concentrations than do Na and Cl (Table s1).
■
COASTAL AND OFFSHORE TEXAS MIOCENE INTERVAL Geological Setting. As summarized by Wallace et al.,17 the early Miocene in the study area (Figure s1) comprises two, thick, prograding, clastic wedges (lower Miocene 1 (LM1) and lower Miocene 2 (LM2)) separated by a significant shale tongue containing Marginulina ascensionensis (Marg. A) fauna. The middle Miocene (MM) is a progradational, clastic section formed over a relatively brief period of deposition (ca. 3 m.y.). A transgressive shale containing Amphistegina chipolensis (Amph. B) fauna separates LM2 from MM. Overlain on a transgressive shale containing either Textularia stapperi fauna or Textularia W fauna, the upper Miocene (UM) deposits from the late middle to early late Miocene record extensive margin offlap over a period of 7 million years. Generally, the Miocene interval in this area is capped by a regional flooding event associated with the Robulus “E” biostratigraphic marker. The major tectonic elements of the study area are shown in Figure s2a. During Miocene (LM1 and LM2) deposition, the strike-parallel Clemente-Tomas fault system actively displaced strata by up to 1200 m, an expansion of approximately 200% and formed as a result of salt evacuation and sediment loading. Subparallel to the Clemente-Tomas system, the Corsair fault system and the contemporaneous Wanda fault systems roughly 19.3 km basinward to it developed in the middle to late Miocene as a result of salt evacuation and progradation of the South Brazos Delta in the case of the former and salt evacuation in the case of the latter.18 In addition, 2-fold expansion of the upper Miocene section occurs across the Wanda fault system. Sandstone Units. The offshore Miocene stratigraphy typically exhibits thick sand intervals with high porosity and permeability, effective trapping mechanisms, and regional seals. The primary targets for CO2 storage in the study area include the fluvio-deltaic sandstones contained in the LM1, LM2, MM, and UM clastic wedges. The primary sealing intervals are the regional transgressive mudrock units (e.g., Marg. A and Amph. B.).17 The data set for determining formation tops and bases, thickness, volume, and porosity in the study area comprises well logs and paleontological data for 3300 wells; these are a subset of a larger Petra software database containing 89,566 wells.17 Well logs used include spontaneous potential (SP), resistivity, gamma ray, caliper, neutron density, bulk density, and sonic (DT).17 Estimation of sandstone units’ porosity was primarily based on digital well logs (Log ASCII standard [LAS]). Thicknesses of Miocene sandstone units at well locations were determined by identifying significant negative deflections in the SP curve of wireline logs.19 The SP curves were used for sand content determination primarily not only because they were the most abundant curves in the data set but also because they highlighted permeable zones.17 Note that only those sandstones within the upper and lower depth limits of the Miocene capacity interval at each well location were selected. The upper depth limit was determined by the minimum
■
METHODS CO2-Solubility Trapping Potential Estimation. At the basin- or regional-scale, CSTP in brine can be estimated according to the following equation12,13,21 MCO2t =
∭ ⌀E(ρsCO2 XsCO2 − ρb X0CO2)dxdydz
(1)
where E is the storage efficiency [-]; ϕ is porosity of the saline formation [-]; ρb and XCO2 are, respectively, background brine 0 density [g/cm3] and mass fraction of dissolved carbon [-]; and ρs and XCO2 are, respectively, brine density [g/cm3] and mass s fraction of dissolved CO2 [-] after the brine is saturated with the injected CO2 [-]. Background mass fraction of dissolved carbon in brine, XCO2 0 , was estimated from the alkalinity measurements (Table s1). Note that if E is equal to 1, MCO2t is a theoretical estimate of CO2-solubility trapping potential (TECSTP), equivalent to the ultimate CO2-sequestration capacity in solution proposed by Bachu and Adams.12 Calculation of ρb, ρs, and XCO2 suggested by Bachu and s Adams,12 are provided in eqs s1−s6 in the Supporting Information. Note that CSTP in eq 1 is different than that with the US-DOE method because the later focuses on physical occupation in the pores of a saline reservoir, while eq 1 is based on CO2 dissolution in brine which is initially present in a saline reservoir.2,22 Three approaches were proposed for the calculation of CSTP, eq 1: simple, coarse, and fine. The simple approach simplifies the integration in eq 1 as 8276
dx.doi.org/10.1021/es502152y | Environ. Sci. Technol. 2014, 48, 8275−8282
Environmental Science & Technology MCO2t = EAT Th⌀̅(ρsCO2 XsCO2 − ρb X 0CO2)
Article
(2)
SCO2 =
where AT is total area, Th is average thickness of the sandstone formation, and ⌀̅ is average porosity. Parameters for calculating CSTP using the simple approach are listed in Table s2. The coarse and fine approaches (see Section 4 in the Supporting Information) discretize the integration of eq 1 into 14,456 square prisms, each having the same square base with an area of one square mile (2.6 km2, shown in Figure s3a). Thickness of the sandstone within each square prism was interpolated from the thickness of sandstones at the well locations extracted from the Petra software database. Sensitivity Analysis. It is of particular interest to compare quantitatively how the CSTP estimate is sensitive to the parameters in eq 1: thickness of the sandstone, porosity, storage efficiency, midpoint depth, brine salinity, and background carbon content. Temperature and pressure in a deep saline aquifer can be estimated from depth (see the Supporting Information) and thus are not included in the sensitivity analysis. Relative sensitivity of the ith parameter is calculated to compare sensitivity of the parameters according to the following equation RSi =
ΔMCO2t /MCO2t δi /Pi ̅
MCO2v + MCO2l MH 2O + MCO2v + MCO2l
(4)
where MCO2v is the mass of CO2 in the head space of the degassing apparatus, MCO2l is the mass of CO2 that was dissolved in brine, and MH2O is the mass of water. Calculation of MCO2v, MCO2l, and MH2O can be found in the Supporting Information. Batch Experiments of Brine-Rock-CO2 Interactions. Dissolution of carbonates and feldspar minerals present in rocks can convert CO2 gas to HCO3− and thus increase CO2solubility trapping potential, according to the following reactions: CaxMg1 − xCO3 + CO2 + H 2O ⇒ xCa 2 + + (1 − x)Mg 2 + + 2HCO3− 2KAlSi3O8 + 2CO2 + 11H 2O ⇒ Al 2Si 2O5(OH)4 + 2K+ + 2HCO3− + 4H4SiO4
X-ray diffraction analysis of rock fragments collected from the Miocene Interval at a depth of 2805 m below surface in the study area suggests that the sandstone rock fragments contain up to 50% quartz, about 14 to 20% calcite, 20% feldspars, and a small amount of clay minerals.28 In order to understand enhancement of CSTP caused by mineral reactions, three batch experiments were conducted. For each batch experiment, a sample of approximately 8 g of the rock fragments was placed into the autoclave apparatus (Figure s5a).28 About 150 mL of synthetic brine (created by adding NaCl to distilled water) with salinity of 110000 ppm was placed into the autoclave apparatus. Supercritical CO2 was pumped into the apparatus, and temperature and total pressure in the reactor were maintained at 100 °C and 200 bar in Batch A, 70 °C and 200 bar in Batch B, and 100 °C and 300 bar in Batch C. N2 was used before CO2 for several days in order to obtain equilibrium between rock and brine. Reacted fluid solution in the reactor was sampled (∼0.5 mL) through the sampling port. Geochemical Modeling of Enhancement of CSTP Caused by Mineral Reactions. A geochemical model was used to simulate the concentration measurements in the three batch experiments. The geochemical model accounted for potential geochemical reactions (aqueous complexation and mineral dissolution/precipitation) that could occur in the batch experiments. Model tool, PHREEQC, together with the LLNL database, was employed.29 The geochemical model was calibrated to fit the concentrations of major ions observed in the three batch experiments and then further used to predict the CO2 mass that dissolved in each batch experiment. Enhancement of CO2 solubility in brine caused by CO2brine-rock interactions was further quantified with a relative increase in CO2 mass dissolved
(3)
where δi is the standard deviation of the ith parameter, P̅i is the mean of the ith parameter, and ΔMCO2t is the change in CSTP obtained by changing the ith parameter at the value of a standard deviation. Laboratory Experiments of CO2 Dissolution in Brine. Various models have been used to calculate CO2 solubility in brine.12,14,23−27 In this study we used Duan’s model26,27 to estimate CO2 solubility in brine (details of Duan’s model are given in the Supporting Information). A set of laboratory experiments for the range of pressure and temperature in the Miocene off-shore repository were conducted to provide new data set for testing Duan’s model. The experimental setup included two main parts: an autoclave system (Figure s4a) for conducting dissolution of supercritical CO2 into brine under various pressure, temperature, and salinity conditions and a degassing apparatus for quantifying the amount of dissolved gas in brine (Figure s4b). The autoclave system had a 250-mL stirred stainless steel reactor vessel, which was instrumented with thermal couples that could heat the vessel to a desired temperature (≤150 °C) and was monitored with thermistors. Liquid CO2 was pumped from a CO2 tank and flowed through a chiller bath and a heat exchanger that generated an input condition of 830 psi and 5 °C required for supercritical-CO2 pump performance. Brine and other solvents were pumped directly from a solvent reservoir to the reactor vessel. Pressure in the reactor vessel (≤40 MPa) was regulated through a backpressure regulator (model: TharSFC ABPR) and monitored with a pressure transducer. Equilibrium between supercritical CO2 and brine in the reactor vessel was obtained when both desired pressure and temperature reached a steady state. A sample of about 10 mL of CO2-saturated brine was collected from the reactor vessel and transferred with a stainless steel syringe to the large glass degassing apparatus, which was placed into a water bath to maintain a temperature of 5 °C in order to minimize vapor pressure in the head space. CO2 solubility in brine (mass fraction) was calculated according to the following equation
RI =
MCO2rock − MCO2 × 100% MCO2
(5)
where MCO2rock is that CO2 mass dissolved in each batch experiment, and MCO2 is the CO2 mass dissolved if no rocks were added. Note that both MCO2rock and MCO2 were predicted with the geochemical model. 8277
dx.doi.org/10.1021/es502152y | Environ. Sci. Technol. 2014, 48, 8275−8282
Environmental Science & Technology
■
Article
Mt/km2 to 0.08 Mt/km2, exhibiting an increasing trend seaward in the study area. Total background CO2 content in the study area is approximately 1.06 Gt, estimated using the simple approachslightly lower than the values estimated using the coarse and fine approaches (Table 1). Estimates of CSTP in the study area using the three approaches are very similar (Table 1), suggesting that the simple method may be applicable for preliminary assessment of CSTP in a saline formation, with only sparse information needed. Possible reasons for the lack of significant difference among the estimates using the three approaches include (1) small variation coefficients in depth and thickness of the Miocene interval (Table s2) and (2) oversimplificationthickness of each sand layer in the Miocene interval was lumped together to the midpoint depth (Figure s2b). The theoretical estimate of CO2-solubility trapping potential (TECSTP), storage efficiency, E =1) is about 167 Gt (Table 1) in the study area. Spatial distribution of TECSTP in the study area, estimated using the coarse approach (Figure s9), varies from 0.38 Mt/km2 to 10.4 Mt/km2, with an average of 4.5 Mt/ km2. Background carbon content in brine in the study area appears to be negligible, only about 0.7% of the TECSTP. This finding is consistent with that reported by Bachu and Adams.12 Currently, there exists no consensus in the scientific community on how to estimate storage efficiency, E in eq 1, for assessment of CSTP. One of the reasons is that storage efficiency is known to be dependent on time and on many factors (e.g., formation layering, thickness, dip angle, permeability anisotropy, and injection activities) that affect the spread and dissolution of CO2 in the entirety of the storage formation under consideration. Some researchers have simply assumed that storage efficiency is equal to 1,12,14 and some others have used the same value of the storage coefficient derived in the US-DOE methodology.15,16 The storage efficiency derived using the US-DOE method measures the fraction of the accessible pore volume that will be occupied by the injected CO2, depending on various parameters, such as area, thickness, porosity, displacement, compressibility of the formation and the fluid, the maximum allowable pressure.2 The storage efficiency used for assessment of CSTP should be used to gauge the fraction of accessible brine volume by the injected CO2 in pore space or in the storage formations. Thus, directly using the storage efficiency derived with the US-DOE method for assessment of CSTP tends to be too conservative. SteelMacInnis et al.30 suggested that brine that may be contacted by CO2 is comprised of two portions: the volume that is left behind the CO2 plume (residual-trapped brine) and the brine outside of the free phase plume that might be contacted by CO2. Since the second portion of brine is poorly constrained and highly uncertain, the portion of residual brine volume in the pore space is used as storage efficiency which has been used
RESULTS AND DISCUSSION Laboratory Measurements of the CO2-Solubility in Brine. Experimental measurements of CO2 solubility in distilled water and synthetic brines (1.88 and 3.4 M) at temperatures from 35 to 140 °C and pressures from 9 to 40 MPa are summarized in Table s3. Results of Duan’s model26 are compared with CO2-solubility measurements in Figure 1.
Figure 1. Experimental validation of Duan’s solubility model26,27 at various temperature, pressure, and salinity values. (Experimental results of CO2 solubility listed in Table s3. (Note that an outlier of CO2 solubility shown in part a was due to a measurement error in mass of brine sample after degassing, eq s9).
Increase in pressure results in increase in CO2 solubility, whereas increase in temperature leads to decrease in CO2 solubility observed for the range of temperature and pressure conducted in this study. Obviously, CO2 solubility decreases with salinity (Figure 1). The modeling results are fairly comparable with the new data set of CO2 solubility measured in this study, suggesting that Duan’s model can provide reasonable estimation of CO2 solubility in brine for the range of temperature and pressure in the study area. Estimate of CO2-Solubility Trapping Potential. Spatial distributions of pressure and temperature at the midpoint depth in the study area are shown in Figure s6. In the storage formation, pressure ranges from 11 to 21 MPa (Figure s6a) and temperature ranges from 50 to 72 °C (Figure s5b). Spatial distribution of background brine density is shown in Figure s7a. After brine is saturated with the injected CO2, brine density shows a slight increase (Figure s7b). Background carbon content (=Th⌀ρbXCO2 0 ), shown in Figure s8, varies from 0.003
Table 1. CO2-Solubility Trapping Potential Estimated with the Three Approaches simplea 3
total brine volume (km ) total background CO2 content in brine TECSTP (E = 1.0)
Gt Mt/km2 Gt Mt/km2
(4.7 1.06 0.028 167 4.47
± ± ± ± ±
finec
coarseb
1.5) × 10 0.76 0.020 55.4 1.48
3
(4.7 1.36 0.036 168 4.48
± ± ± ± ±
1.5) × 10 0.83 0.022 55.6 1.49
3
(4.7 1.35 0.036 166 4.45
± ± ± ± ±
1.5) × 103 0.83 0.022 55.3 1.48
The simple approach is defined in eq 2. bThe coarse approach is defined in eq s13. cThe fine approach is defined in eq s14 in the Supporting Information. a
8278
dx.doi.org/10.1021/es502152y | Environ. Sci. Technol. 2014, 48, 8275−8282
Environmental Science & Technology
Article
interval at a regional scale. The CSTP estimated in this study should not be considered as the storage capacity at a specific site within the study area. Further site-specific information and characterization may be needed for estimating a more precise capacity at a specific site. We compared CSTP estimated in the study area with those values estimated in other saline formations reported in the literature (Table 2). The listed saline formations vary in area, depth, thickness, porosity, temperature, pressure, and brine salinity. CSTP estimated in those saline formations varies significantly, ranging from 0.1 Gt in the Leman Sandstone Formation to 198 Gt in the Viking aquifer, owing to various factors such as differences in storage coefficient. Note that Duan’s model was used to calculate CO2 solubility in brine for the saline formations except the Ultisira formation and Viking and Winnipegosis aquifers. CSTP in the Leman Sandstone Formation was estimated on the basis of residual brine volume (Table 2). The TECSTP (= SCTP divided by E listed in Table 2) shows a linear relationship with brine volume
in the study of CO2 storage capacity in the United Kingdom sector of the south North Sea basin.31 Because there is no available data for portion of residual brine in pore space after CO2 injection in this study area, we employed the method described in the report (pp 34−35) proposed by Braitsch et al.32 by converting the storage efficiency estimated using the US-DOE methodology reported by Wallace et al.17 to the storage efficiency which can be used for assessment of CSTP in a saline aquifer. With the average depth, temperature, pressure, and brine salinity in the study area (Table s2), the calculated conversion factor is 15.3 with the method by Braitsch et al.32 Thus, the storage efficiencies of 0.184 for 10 percentile (P = 10) and 0.69 for 50 percentile (P = 50). The up-scaled storage efficiency is greater than 1 for the storage efficiency derived at P = 90 which was not used in the calculation. With E = 0.69, CSTP was estimated at 114 ± 38 Gt, which is about 89% of the total CO2-storage capacity, 129 Gt, estimated using the USDOE methodology, which assumes that all pore space is occupied by the injected CO2.17 CSTP varies from 0.26 Mt/ km2 to 7.2 Mt/km2 at the upper right corner of the study area, shown in Figure 2; whereas spatial distribution of CO2-storage
TECSTP = 0.040Vb
(6)
where Vb is the volume of brine (km3). The coefficient of determination (R2) of eq 6 is close to 0.9 (Figure 3). The slope of the regression line is 0.040 Gt/km3. This finding suggests that the TECSTP in a saline formation could be reasonably estimated from the brine volume in the storage unit regardless of temperature, pressure, and salinity. Sensitivity Analysis. CSTP in the study area is most sensitive to thickness and porosity (Table s4) which are also two of three parameters (thickness, porosity, and area) that determine brine volume, suggesting that the volume of available brine in the storage aquifer is a primary control. This result is consistent with the finding shown in Figure 3. As expected, CSTP is highly sensitive to storage efficiency used in the calculation because CSTP in eq 1 is proportional to the storage efficiency. CSTP appears to be more sensitive to brine salinity than to midpoint depth of the Miocene sandstone formation in the study area (Table s4), suggesting that salinity is more important than pressure and temperature (functions of midpoint depth) for regional assessment of CSTP in this study. This pattern is consistent with the finding in the study of CSTP in the Winnipegosis aquifer.13 As expected, CSTP is almost insensitive to background carbon content in the study area (Table s4), because background carbon content is negligible compared to CSTP. Enhancement of CSTP Caused by Mineral Reactions. Concentrations of major ions (Ca, Mg, K, Si) measured in the three batch experiments over a period of about 300 h after supercritical CO2 was pumped into the autoclave apparatus are shown in Figures s11−s13 and indicate obvious increasing trends over the reaction time. The relative increase in CSTP, estimated with eq 3, varies from 2% in Batch C to 3.5% in Batch B (Figure 4a) by assuming that Dawsonite is able to precipitate in the geochemical model, suggesting that enhancement of the CSTP as a result of interactions between brine and rocks was relatively minor. In addition, relative increase in CSTP at lower pressure and temperature appears to be greater than at higher pressure and temperature (Figure 4a). As has been extensively discussed in the literature,34−39 it is possible for injected CO2 to be trapped as solid phases, such as Dawsonite, siderite, and other carbonates. Modeling results of the three batch experiments in this study suggest that Dawsonite precipitated after supercritical CO2 was pumped in
Figure 2. Spatial distribution of theoretic estimate of CO2-solubility trapping potential in study area with storage efficiency E = 0.69 for 50 percentile (P = 50). Note the contours were interpolated with the inverse-distance weighting method.
capacity estimated with the US-DOE methodology varies from 0.4 Mt/km2 to 8.4 Mt/km2, suggesting that CSTP is comparable with the storage capacity estimated using the USDOE methodology. CSTP was estimated at 30.5 ± 10 Gt for P = 10. Spatial distribution of CSTP in the study area for P = 10 is shown in Figure s10. As pointed out by Dilmore et al.,15 the two methods may represent estimation of CO2-storage potential for two extreme scenarios: one in which all CO2 injected is present as dissolved carbon aqueous species in brine that occupies all pore space, and the other in which all brine is displaced from the pore space in the storage and replaced by the injected CO2. Neither of the two scenarios could be realistic because (1) dissolution of CO2 into brine is dramatically limited by accessibility of CO2 to brine and the time scale could be up to thousands of years, depending on various factors such as storage geometry and rock properties,1,4,33 and (2) capillary force may limit complete displacement of brine from pore space in the storage unit.15 Hence, our study provides a theoretical estimate rather than attempting to provide a precise amount of CO2 that can be potentially dissolved into brine in the Miocene 8279
dx.doi.org/10.1021/es502152y | Environ. Sci. Technol. 2014, 48, 8275−8282
Environmental Science & Technology
Article
Table 2. Comparison of CO2-Solubility Trapping Potential in Brine with Those Reported in the Literature site (formation) Miocene interval, TX, USA Oriskany Formation, PA, USA The Utsira Formation, Sleipner, Norway Viking aquifer, Canada Winnipegosis aquifer, Canada Lower Yancheng Formation, Subei basin, East China The Upper Sanduo Formation, Subei basin, East China Leman Sandstone Formation, North Sea basin, UK a
area (km2)
depth (m)
thickness (m)
porosity (%)
temp (°C)
pressure (Mpa)
salinity (ppm)
brine volume (km3)
CO2-solubility trapping potential (Gt)
source
37441
1006-3100
