Rates of CO2 Mineralization in Geological Carbon ... - ACS Publications

Article pubs.acs.org/accounts

Rates of CO2 Mineralization in Geological Carbon Storage Published as part of the Accounts of Chemical Research special issue “Chemistry of Geologic Carbon Storage”. Shuo Zhang†,‡ and Donald J. DePaolo*,†,§ †

Department of Earth and Planetary Science, UC Berkeley, 307 McCone Hall Berkeley, California 94720-4767, United States Aramco Research Centers, Houston 16300 Park Row Drive Houston, Texas 77094, United States § Energy Geosciences Division, Lawrence Berkeley National Laboratory 1 Cyclotron Road, MS74R316C Berkeley, California 94720, United States ‡

CONSPECTUS: Geologic carbon storage (GCS) involves capture and purification of CO2 at industrial emission sources, compression into a supercritical state, and subsequent injection into geologic formations. This process reverses the flow of carbon to the atmosphere with the intention of returning the carbon to long-term geologic storage. Models suggest that most of the injected CO2 will be “trapped” in the subsurface by physical means, but the most risk-free and permanent form of carbon storage is as carbonate minerals (Ca,Mg,Fe)CO3. The transformation of CO2 to carbonate minerals requires supply of the necessary divalent cations by dissolution of silicate minerals. Available data suggest that rates of transformation are highly uncertain and difficult to predict by standard approaches. Here we show that the chemical kinetic observations and experimental results, when they can be reduced to a single cation-release time scale that describes the fractional rate at which cations are released to solution by mineral dissolution, show sufficiently systematic behavior as a function of pH, fluid flow rate, and time that the rates of mineralization can be estimated with reasonable certainty. The rate of mineralization depends on both the abundance (determined by the reservoir rock mineralogy) and the rate at which cations are released from silicate minerals by dissolution into pore fluid that has been acidified with dissolved CO2. Laboratorymeasured rates and field observations give values spanning 8 to 10 orders of magnitude, but when they are evaluated in the context of a reservoir-scale reactive transport simulation, this range becomes much smaller. The reservoir scale simulations provide limits on the applicable conditions under which silicate mineral dissolution and subsequent carbonate mineral precipitation are likely to occur (pH 4.5 to 6, fluid flow velocity less than 5 m/year, and 50−100 years or more after the start of injection). These constraints lead to estimates of 200 to 2000 years for conversion of 60−90% of injected CO2 when the reservoir rock has a sufficient volume fraction of divalent cation-bearing silicate minerals and confirms that when reservoir rock mineralogy is not favorable the fraction of CO2 converted to carbonate minerals is minimal over 104 years. A sufficient amount of reactive minerals is typically about 20% by volume. Our approach may allow for rapid evaluation of mineralization potential of subsurface storage reservoirs and illustrates how reservoir scale modeling can be integrated with other observations to address key issues relating to engineering of geologic systems.

1. INTRODUCTION Understanding the geochemical behavior of anthropogenic carbon dioxide stored in geological reservoirs is crucial for quantifying the risk of leakage and the evolution of the form of the CO2 through the life of an individual storage site.1 Mineral trapping is considered to be the most secure form of CO2 storage but is also thought to be slow to develop because it depends on the release of cations like Fe, Mg, and Ca by dissolution of silicate and oxide minerals in the rocks that constitute the underground storage reservoirs. Dissolution of CO2 into brine and subsequent formation of carbonic acid is the first step in the mineral trapping process. Carbon dioxide dissolves in water at the level of a few percent by volume to form carbonic acid with pH ≈ 3, which in turn can slowly dissolve silicate minerals. Under appropriate conditions, divalent cations released by silicate mineral © 2017 American Chemical Society

dissolution can combine with dissolved carbonate ions to form stable carbonate minerals. This mechanism requires host rocks with a high “acid neutralization potential”.2 Rocks rich in calcium, magnesium, and iron silicate minerals can neutralize acids by providing Ca2+, Mg2+, and Fe2+, which can later form stable carbonate phases.3 The most likely candidate storage formations for CO2 are sedimentary rocks with relatively high porosity and permeability, such as sandstones, that are overlain in stratigraphic succession by low-permeability rock formations such as shale or mudstone. High porosity and permeability allow rapid injection of CO2, and the overlying low-porosity caprocks (or “seals”) can physically trap the buoyant supercritical CO2 phase so that Received: July 5, 2017 Published: August 28, 2017 2075

DOI: 10.1021/acs.accounts.7b00334 Acc. Chem. Res. 2017, 50, 2075−2084

Article

Accounts of Chemical Research Table 1. Data Used in Calculating Release Rate of Cations from Rocks data type and reference

cation concentration (mol/kg)

rock/fluid ratio (kg/L)

closed batch experiment, Nagaoka29 flow through batch experiment, Nagaoka17 core flood experiment, Nagaoka18 field evidence, Nagaoka29 field evidence, Cranfield19 core flood experiment, Sleipner20 field evidence, CarbFix6

5.06 2.04 2.04 5.06 1.01 0.67 4.5

0.01 0.000095 0.12 13.00 0.05 1.38 10.80

ΔCa (mmol/L) 0.5 0.0026 0.475 7.2

ΔMg (mmol/L)

ΔFe (mmol/L)

5.13 0.0014 0.357 0.538 1.43 1.5

820

Table 2. Fraction of Available Cations Released from Dissolution of Reactive Minerals for CO2 Mineralization Compiled from Laboratory and Field Experiments and Natural Analogs data type and reference 29

closed batch experiment, Nagaoka flow through batch experiment, Nagaoka17 core flood experiment, Nagaoka18 field evidence, Nagaoka29 field evidence, Cranfield19 core flood experiment, Sleipner20 field evidence, CarbFix6 core flood experiment, the Pretty Hill Formation31 natural analog, the Pretty Hill Formation32

reaction rate (year−1)

pH

48.7 0.258 0.252 9.49 × 10−5 0.27 0.059 8.44 × 10−3 8.03 1.70 × 10−4

3.07 3.07 3.07 6.35 5.2 3.18 3.85 3.28

flow rate (m/year) 41.5 10.7 4.05 113.9 456.25 376 0.005

temp (°C)

pressure (MPa)

time scale

50 50 50 48 120 70 20−33 95

10 10 10 10.8 32.4 10

20 h 700 h 1200 h 633 days 37 days 10 days 2 years 157 h 4500 years

10

these divalent cations be held in minerals that dissolve relatively rapidly. Prediction of the long-term fate of the carbon dioxide stored in volcanogenic sandstones requires determination of the relevant gas−fluid−mineral reactions and their chemical kinetics. Modeling the progress of such reactions is often frustrated by uncertainties in the absolute mineral surface reaction rates and the significance of other rate-limiting steps in natural systems. Observed or inferred silicate dissolution rates in the natural environment are typically 2 to 5 orders of magnitude slower than would be inferred from far-fromequilibrium, laboratory-derived rates at similar pH and temperature conditions,10 and the laboratory rates are typically the only rates available for simulations.11−13 This study uses the measured release rates of divalent cations in volcanogenic sandstones and related rock reservoirs from experimental studies, field evidence, and natural analogs to evaluate the potential and rate of CO2 mineralization in a largescale subsurface injection. To compare the release rates of cations determined from different studies, we normalize the rates to the fraction of available cations (Ca, Mg, Fe) released to aqueous solution from mineral dissolution per unit time. This approach essentially lumps the dissolution rate constants, surface areas, and volume fractions of minerals into one value. The objective is to determine the order of magnitude of CO2 mineralization rate in common volcanogenic sandstone reservoirs and to evaluate the extent to which we can use seemingly divergent observations to make useful predictions.

the slower processes of dissolution into brine and reaction with the minerals can proceed without leakage of the CO2 back to the surface. Rock types with the highest porosity and permeability usually are rich in quartz (SiO2) with some feldspar, (Na,Ca)(Si,Al)AlSi2O8, and do not contain an abundance of readily soluble silicate minerals that contain divalent cations that can combine with CO2 to form carbonate minerals (CaCO3, MgCO3, FeCO3). For storage in cation-poor rocks, modeling studies show that only a few percent of injected CO2 is trapped in minerals even after 1000 to 10000 years has elapsed from the time of injection.4 In contrast, there is an increasing literature advancing the possibility of using certain silicate rock types that are known to react rapidly with acidic aqueous solutions as a way to sequester CO2.5−7 Basalt lava flows, for example, contain acid-neutralizing minerals as well as, in some cases, volcanic glass that can react rapidly with acidic solutions. Among common silicate minerals, olivine, (Mg,Fe)2SiO4, is the most unstable in contact with acidic solutions and reacts quite fast even at room temperature. Other minerals that are common and somewhat less reactive but still potentially useful for neutralizing acids are orthorhombic pyroxene [(Mg,Fe)2Si2O6], monoclinic pyroxene [Ca(Mg,Fe)Si 2 O 6 ], and Ca-rich plagioclase feldspar [(CaxNa1−x)Al1+xSi2−xO8]. These minerals and olivine are commonly found in basalt lava flows as well as in so-called ultramafic rocks such as peridotite. Mineralization of injected CO2 can be significant if the reservoir sandstones have a sufficient percentage of volcanic rock fragments (VRF) or the divalent cation-bearing minerals listed above. Such “volcanogenic” sandstones8,9 could be a promising target for CO2 mineralization because they are porous and permeable enough to allow injection at acceptable rates. In most volcanogenic sandstones, the divalent cationbearing minerals are chlorite, zeolites, smectite, amphibole, and devitrified lithic fragments that may still contain some of the original igneous minerals. To mineralize injected CO2, the requirement is that there be sufficient Ca, Mg, and Fe available to combine with CO2 to form carbonate minerals, and also that

2. FORMULATION OF DISSOLUTION TIME SCALES Available data on mineral dissolution rates in potential reservoir rocks for carbon storage come from laboratory experiments, field experiments, and observations of natural analogue systems (Tables 1 and 2). Laboratory experiments include batch reactor experiments where crushed or disaggregated rock material is mixed with CO2-saturated fluid and allowed to react for periods of weeks to months. Dissolution is monitored by measuring the increase with time in dissolved ions in the fluid phase. Another 2076


Article

Accounts of Chemical Research

Figure 1. Experimental and field data showing relationships between observed fractional dissolution rate (eq 5 of text) and (a) fluid flow velocity and (b) pH. Expected ranges applicable to large-scale geologic carbon sequestration are also shown. Data and relevant sources are listed in Table 2.

approach is to flow CO2-saturated water or brine through a cylindrical core of intact porous rock and to monitor the change in ion concentrations between the inflow and outflow.30 Field experiments involve injection of supercritical CO2 or CO2-charged water into the subsurface through a well and measurement of fluids extracted from another well positioned 10s to 100s of meters from the injection well. Field experiments are analogous to flow-through laboratory experiments but at a larger scale and involving additional uncertainties. Natural analogue observations are based on measurements of natural fluid compositions or rock alteration phenomena in circumstances where the history of interaction between fluid and minerals can be inferred. There is extensive literature on dissolution rates of individual common rock-forming minerals when exposed to aqueous solutions of differing pH and temperature.8,14 Although the results of such experiments are reasonably consistent, there is ambiguity about how to apply the data to natural rock−fluid systems.15,16 The representation of mineral dissolution rate is generally done with an equation of the form: R mineral(mol/year) = Srk(1 − Q /K )α

R total(mol/(kg rock· year)) =

⎝

i

Qi⎞ ⎟ Ki ⎠

αi

(3)

Application of this equation in natural fluid−rock systems requires information on mineralogy, mineral-by-mineral reactive surface area, kinetic rate constants, and departures from equilibrium. Significant progress has been made recently on how to determine these parameters in sandstones.17,18 Nevertheless, estimating the rates for actual rocks appropriately remains difficult. To more easily evaluate the data available in the literature, we simplify by dividing eq 3 by the total amount (in moles) of available cations in the rock: R fractional(year ‐1) =

(

∑i m i niSriki 1 −

Qi Ki

αi

)

∑i m i ni /Mi

(4)

Where Mi is the molecular mass of each mineral (kg/mol). With this formulation, we can evaluate and compare rates from a variety of rock types and fluid compositions. To simplify further and focus on the aspects of dissolution that are most germane to carbon mineralization, we restrict consideration to noncarbonate minerals that contain the divalent cations Ca, Mg, and Fe, which are the key components necessary to mineralize dissolved CO2. Hence, the rate that we discuss here can also be written as

(1)

where k is the absolute kinetic rate constant, Q is the ion activity product of the mineral in solution, K is the equilibrium ion activity product of the mineral in solution, α is an empirical constant, and Sr is the reactive surface area. In some cases, k is made to be an explicit function of pH, temperature, or other characteristics of the solution. In this Account, temperature is set at 75 °C for all simulations. In a rock containing a mass fraction m1 of mineral 1, which contains n1 (the stoichiometric factor) divalent cations, the release rate of cations from that mineral in the rock would be represented as R1(mol/(kg rock· year)) = m1n1Sr1k1(1 − Q 1/K1)α1

⎛

∑ miniSriki⎜1 −

R fractional(year ‐1) =

(Ca + Mg + Fe)si,released Δt ∑ (Ca + Mg + Fe)si,rock

(5)

where the numerator is the number of moles of divalent cations released by dissolution, and the denominator is the product of the amount of time elapsed in the experiment, and the total initial amount of divalent cations in the rock. Ultimately it is this rate that is most germane to CO2 mineralization, and the question is whether it is predictable and with what confidence.

(2)

3. SYSTEMATICS OF CATION RELEASE RATES Experimental and field study data are summarized in Tables 1 and 2 and Figures 1 and 2. Only certain measurements of Ca, Mg, or Fe release were suitable for interpretation in terms of eq

where now F, Sr, k, Q, and K are mineral-specific parameters. The overall release rate of cations from the rock would be written: 2077


Article


Figure 2. Relationship between observed fractional dissolution rate for rocks and soils and time. The time axis refers to the duration of the experiment, the time period over which the rate applies, or the age of soils and other geologic features. The data compiled in this work represent slightly higher temperatures than the other data shown.

Figure 3. Representative values for dissolution rate constants as a function of pH at temperature of 80 °C for several minerals that contain divalent cations. Data from summaries in Palandri and Kharaka (2004)14 and Zhang et al. (2015).8 Also shown is the expected range of pH attending mineral dissolution in a simulated sequestration system.

5 as noted in Table 1. Several of the rock compositions used are rich in cations (2 to 5 mol/kg), with the exceptions being those from the Cranfield19 and Sleipner20 sites (1.01 and 0.67 mol/ kg). The range in normalized release rate is almost 6 orders of magnitude. In general dissolution rates are higher when fluid pH is low, and when flow rates are high; high flow rates ensure that the fluid does not approach equilibrium with the dissolving mineral phases.21 The dependence of cation release rate on pH and flow rate is not simple, but for pH > 4 and flow rates 10 m/year and pH < 4, either do not occur or occur only in the region within 200 m of the injection well. Since little mineralization occurs in that region in

obtained by pre-equilibrating 1.0 mol/L NaCl saline water in a batch model for 10 years with minerals that are used later for reactive transport modeling. This simulation follows the procedures and model parameters of Zhang et al. (2013; 2015)8,9 using the software package TOUGHREACT,24 but we use the mineralogy and reactive surface area measurements reported by Beckingham et al. (2016)17 for the Nagaoka reservoir rock. The mineralogy and RSA values of the Nagaoka example, plus those used previously by Zhang et al. (2013; 2015),8,9 are provided in Table 3. The Nagaoka composition, because of the large values of reactive surface area and reactive mineralogy, presents an example of particularly rapid subsurface mineralization. The model uses a first order formulation for mineral dissolution and precipitation kinetics (effectively eq 1 with α = 1), and typical values and formulations of the relative permeabilities of brine and supercritical CO2, porosity, permeability, and capillary pressure.8 Gravitational (buoyancy) effects are not included, but tests have shown that adding them to the simulation and allowing the model to be 2-dimensional with radial symmetry produces only minor changes in the results.9,25 Selected results of the simulation using the Nagaoka mineralogy are shown in Figures 5 and 6, and additional results including the other rock compositions (from Zhang et al., 2013, 2079


Article


Figure 5. Model results showing pH and brine flow velocity as functions of distance from the injection well at different times after the start of injection. Brine flow velocities are close to zero after injection stops at 100 years. Values are shown for the simulation in this study using the Nagaoka mineralogy and surface areas.

In the 1-dimensional model, the CO2 available to be converted to minerals is mostly the CO2 that is present in the pore space after injection is complete. In the 1-D model, this CO2 does not migrate upward within the reservoir as a consequence of its buoyancy and therefore decreases with time only because of dissolution into the pore fluid and conversion to silicate minerals. In two dimensions, much of the CO2 migrates upward and collects at the top reservoir layer. This process limits exposure of the CO2 to minerals and should decrease the overall amount of mineralization. However, the 1D model gives useful estimates for the mineral trapping of CO2 because there are other processes that can increase the exposure of the soluble minerals to CO2. In particular, it is thought that in many cases, especially if permeability is sufficiently high, negatively buoyant plumes of CO2-saturated brine can form from the stratigraphically trapped CO2 layer at the top of the reservoir.25−27 This process would tend to resupply acidified brine to a large volume of the reservoir formation, providing additional CO2 that could be mineralized well after injection has stopped. As noted in Zhang et al. (2015),8 vertical heterogeneity in mineralogy and permeability within the reservoir formation can also affect the rate and amount of mineralization, and makes the system behave more like the 1-D model.

the simulation, and because that region represents only about 0.25% of the total rock volume affected by the CO2 injection and subsequent mineralization, those dissolution rate values are not useful for estimating overall mineralization rates in GCS systems. The simulations show that mineralization occurs mostly between 50 and 1000 years after CO2 first encounters the reservoir rocks (Figure 7); hence the applicable values for specific dissolution rate (based on Figure 2) should typically be in the range of 10−3 to 10−4 year−1. Referring to Figure 3, since pH is typically in the range 4.5 to 6 for the dissolution process, the applicable time scale for dissolution is in the range of 102 to 104 years as discussed above, equivalent to specific dissolution rate values of 10−2 to 10−4 year−1. The amount of mineralized CO2 produced in the simulations using the Etchegoin reservoir rock composition8 is also shown in Figure 7a. The Nagaoka example is probably a best case for subsurface mineralization, and the result shows that in this type of rock, mineralization could be essentially complete within 200−300 years. Even in this example of rapid mineralization, most of the mineralization occurs after injection has stopped. For the Etchegoin example, mineralization stops at about 90%, and the time scale for 60% mineralization is 400−500 years rather than 200 years for the Nagaoka example. Figure 7b shows the results for the synthetic sandstone composition from Zhang et al. (2013).9 The curve for RSA = 10 cm2/g is the one that is comparable to the curves shown in Figure 7a. The effects of increasing or decreasing the RSA values by a factor of 5 are also shown and are substantial. 2080


Article


Figure 6. Model results showing (a) mass of CO2 mineralized and (b) porosity in the reservoir formation at different times after the start of CO2 injection (c) CO2 saturation in the reservoir pore space (fraction of pore space filled with supercritical CO2) at the end of the injection period (100 year). Values are shown for the simulation done here (Nagaoka) and for those reported in Zhang et al. (2013, 2015)8,9 that use similar model parameters.

5. MINERALIZATION RATE ESTIMATES AND IMPLICATIONS

in the pore space. This ratio is given approximately by the following formula, modified from Zhang et al. (2013):9

The potentially mineralizable CO2 is represented by the gas filling the pore space in the reservoir at the end of injection plus the CO2 dissolved in pore fluid. The mineralization potential of the reservoir is determined by the molar ratio of divalent cations in the minerals of the rock to the moles of CO2 residing

EMT =

(1 − ϕ)X rmρrm MCO2 ϕ[sg + (1 − sg)S b]ρCO M rm 2

(6)

where Xrm is the volume fraction of reactive minerals containing divalent cations, sg is the CO2 “gas” saturation at the end of 2081


Article


Figure 7. Simulated values for the amount of injected supercritical CO2, and the amount that is converted to minerals over a 1000-year time span after the start of injection. (a) Simulation results for the Nagaoka case (cf. Table 3) and for a reservoir mineralogy corresponding to the Etchegoin Formation of central California (Table 3 and Zhang et al., 20158). (b) Values of mineralized CO2 showing dependence on assumed reactive surface area of cation-bearing minerals (from Zhang et al., 20139).

injection (Figure 6c), ϕ is the initial porosity, Sb is the volumetric solubility of CO2 in brine (typically about 2−3%), and M is the molar weight of minerals and CO2. The larger the value of E, the more likely that a large fraction of injected CO2 can be mineralized. This ratio is higher when the proportion of reactive, cation-bearing minerals is higher and also when the porosity and average gas saturation are lower. Gas saturation values in the reservoir depend on the relative permeabilities of CO2 and brine in the rock formation, and on details of reservoir heterogeneity. Equation 6 takes account of the CO2 dissolved in brine at the time injection stops, which increases the total amount of mineralizable CO2 per unit volume of rock by about 10% relative to that accounted for by the gas phase only. The results from the reservoir scale reactive transport modeling plus the systematics of observations on mineral dissolution rates yield estimates of the time scale as well the extent of CO2 mineralization. The values of sg at the end of the injection period are in the range 0.4 to 0.25 (Figure 6c); the reservoir volume average is 0.28, and the value of Sb is 0.02. The fractional rate of mineralization can be estimated from eq 7 and the fractional cation release rates deduced from mineralogical data (Table 3) or the more empirical data in Figures 1 and 2. The result can be written:

rate(year −1) =

(1 − ϕ)X rmρrm MCO2 Δtminϕ[Sg + (1 − sg)S b]ρCO M rm 2

Ncations = ΔtminNCO2

(7)

where N represents the number of moles (of divalent cations or CO2) contained in a unit volume of rock-plus-pore space at the end of the injection period. For a system where the molar ratio of cations to pore space CO2 is 5 and the time scale of cation release is 100−104 years, the time scale for CO2 mineralization is about 20 to 2000 years. If the dissolving minerals are highly reactive, like olivine, pyroxenes, and hornblende, the mineralization time scale will be nearer the short end of this range as shown in our simulations (Figure 7). If the minerals are more slowly reacting, the time scale will be toward the longer end of the range. If the ratio of cations to CO2 is lower, for example, 1 or less, and the minerals are of the less reactive type, then the time scale for mineralization is 104 years or more and mineralization will not be important on the time scale for engineering a GCS system. Zhang et al. (2015)8 summarized occurrences of volcanogenic sandstones in sedimentary basins around the world. 2082


Article


in a saline Aquifer at the Sleipner site, North Sea. Am. J. Sci. 2007, 307, 974−1008. (5) Kelemen, P. B.; Matter, J.; Streit, E. E.; Rudge, J. F.; Curry, W. B.; Blusztajn, J. Rates and mechanisms of mineral carbonation in peridotite: natural processes and recipes for enhanced, in situ CO2 capture and storage. Annu. Rev. Earth Planet. Sci. 2011, 39, 545−576. (6) Matter, J. M.; Stute, M.; Snæbjörnsdottir, S. Ó .; Oelkers, E. H.; Gislason, S. R.; Aradottir, E. S.; Sigfusson, B.; Gunnarsson, I.; Sigurdardottir, H.; Gunnlaugsson, E.; et al. Rapid carbon mineralization for permanent disposal of anthropogenic carbon dioxide emissions. Science 2016, 352, 1312−1314. (7) McGrail, B. P.; Schaef, H. T.; Spane, F. A.; Cliff, J. B.; Qafoku, O.; Horner, J. A.; Thompson, C. J.; Owen, A. T.; Sullivan, C. E. Field validation of supercritical CO2 reactivity with basalts. Environ. Sci. Technol. Lett. 2017, 4, 6−10. (8) Zhang, S.; DePaolo, D. J.; Voltolini, M.; Kneafsey, T. CO2 mineralization in volcanogenic sandstones: geochemical characterization of the Etchegoin formation, San Joaquin Basin. Greenhouse Gases: Sci. Technol. 2015, 5, 622−644. (9) Zhang, S.; DePaolo, D. J.; Xu, T.; Zheng, L. Mineralization of carbon dioxide sequestered in volcanogenic sandstone reservoir rocks. Int. J. Greenhouse Gas Control 2013, 18, 315−328. (10) White, A. F.; Brantley, S. L. The effect of time on the weathering of silicate minerals: why do weathering rates differ in the laboratory and field? Chem. Geol. 2003, 202, 479−506. (11) Knauss, K. G.; Johnson, J. W.; Steefel, C. I. Evaluation of the impact of CO2, co-contaminant gas, aqueous fluid and reservoir rock interactions on the geologic sequestration of CO2. Chem. Geol. 2005, 217, 339−350. (12) White, A. F.; Schulz, M. S.; Vivit, D. V.; Blum, A. E.; Stonestrom, D. A.; Harden, J. W. Chemical weathering rates of a soil chronosequence on granitic alluvium: III. Hydrochemical evolution and contemporary solute fluxes and rates. Geochim. Cosmochim. Acta 2005, 69, 1975−1996. (13) Xu, T.; Apps, J. A.; Pruess, K. Numerical simulation of CO 2 disposal by mineral trapping in deep aquifers. Appl. Geochem. 2004, 19, 917−936. (14) Palandri, J. L.; Kharaka, Y. K. A compilation of rate parameters of water-mineral interaction kinetics for application to geochemical modeling, U.S. Geological Survey, 2004. (15) Maher, K.; Steefel, C. I.; DePaolo, D. J.; Viani, B. E. The mineral dissolution rate conundrum: Insights from reactive transport modeling of U isotopes and pore fluid chemistry in marine sediments. Geochim. Cosmochim. Acta 2006, 70, 337−363. (16) Steefel, C. I.: Geochemical kinetics and transport. In Kinetics of water-rock interaction; Springer, 2008; pp 545−589. (17) Beckingham, L. E.; Mitnick, E. H.; Steefel, C. I.; Zhang, S.; Voltolini, M.; Swift, A. M.; Yang, L.; Cole, D. R.; Sheets, J. M.; AjoFranklin, J. B.; et al. Evaluation of mineral reactive surface area estimates for prediction of reactivity of a multi-mineral sediment. Geochim. Cosmochim. Acta 2016, 188, 310−329. (18) Beckingham, L. E.; Steefel, C. I.; Swift, A. M.; Voltolini, M.; Yang, L.; Anovitz, L. M.; Sheets, J. M.; Cole, D. R.; Kneafsey, T. J.; Mitnick, E. H.; et al. Evaluation of accessible mineral surface areas for improved prediction of mineral reaction rates in porous media. Geochim. Cosmochim. Acta 2017, 205, 31−49. (19) Lu, J.; Kharaka, Y. K.; Thordsen, J. J.; Horita, J.; Karamalidis, A.; Griffith, C.; Hakala, J. A.; Ambats, G.; Cole, D. R.; Phelps, T. J.; et al. CO 2−rock−brine interactions in Lower Tuscaloosa Formation at Cranfield CO 2 sequestration site, Mississippi, USA. Chem. Geol. 2012, 291, 269−277. (20) Rochelle, C. A.; Bateman, K.; Pearce, J. M. Geochemical interactions between supercritical CO2 and the Utsira Formation: an experimental study. British Geological Survey, Commissioned Report CR/02/006, 2002. (21) Maher, K.; Druhan, J. Relationships between the transit time of water and the fluxes of weathered elements through the critical zone. Procedia Earth Planet. Sci. 2014, 10, 16−22.

Typical rocks of this type contain 10−20% volcanic rock fragments or other reactive minerals, about half the amount in the Nagaoka and Etchegoin rocks that we have analyzed here, so that the value of E is about 2. Given that Δtmin is typically near 1000−3000 years for these rocks, the time scale for mineralization is roughly 1000 years. Although as Matter et al. (2016)6 have shown, basalts can potentially react faster than this, the requirement of predissolving CO2 into water limits the feasibility of storing CO2 by the method of their experiment. Another possibility for storing CO2 in volcanic formations is in the submarine parts of large extinct oceanic volcanoes like those of the Hawaiian island chain. The interior flanks of these volcanoes are made up of large volumes of clastic volcanic material that retains 10−15% porosity.28 These rocks would not be attractive for storing CO2 in the near future because they are far from large sources of industrial CO2, but if direct air capture became feasible, they could be. Similarly, there are substantial volcanic aprons around most of the island arcs that circle the Pacific Ocean, and they too could potentially be highly reactive storage possibilities for CO2.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Donald J. DePaolo: 0000-0001-8880-0202 Notes

The authors declare no competing financial interest. Biographies Shuo Zhang is a research geologist in Aramco Global Research Centers, Houston. He received his Ph.D. in Earth and Planetary Science from University of California at Berkeley in 2014. His research interest has focused on geochemical aspects of CO2 sequestration, carbonate diagenesis, and coupling of fluid flow with geochemistry. Donald J. DePaolo is Chancellor’s Professor Emeritus in the Department of Earth and Planetary Science at the University of California, Berkeley, and a Senior Advisor at the Lawrence Berkeley National Laboratory. His research interests include the use of geochemistry for understanding geologic processes, and isotope geochemistry in particular for establishing the rates and time scales of geologic processes.

■

ACKNOWLEDGMENTS This work was supported as part of the Center for Nanoscale Control of Geologic CO2 (NCGC), an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, under Award No. DEAC02-05CH11231.

■

REFERENCES

(1) Metz, B. Carbon dioxide capture and storage. Special Report of the Intergovernmental Panel on Climate Change; Cambridge University Press, 2005. (2) Baines, S. J.; Worden, R. H. The long-term fate of CO2 in the subsurface: natural analogues for CO2 storage. Geological Society, Special Publications, London 2004; Vol. 233, pp 59−85. (3) Gunter, W. D.; Perkins, E. H.; Hutcheon, I. Aquifer disposal of acid gases: modelling of water-rock reactions for trapping of acid wastes. Appl. Geochem. 2000, 15, 1085−1095. (4) Audigane, P.; Gaus, I.; Czernichowski-Lauriol, I.; Pruess, K.; Xu, T. F. Two-dimensional reactive transport modeling of CO2 injection 2083


Article

Accounts of Chemical Research (22) Fantle, M. S.; DePaolo, D. J., Sr isotopes and pore fluid chemistry in carbonate sediment of the Ontong Java Plateau: Calcite recrystallization rates and evidence for a rapid rise in seawater Mg over the last 10 million years. Geochim. Cosmochim. Acta 2006, 70, 3883− 3904. (23) Johnson, N. C.; Thomas, B.; Maher, K.; Rosenbauer, R. J.; Bird, D.; Brown, G. E. Olivine dissolution and carbonation under conditions relevant for in situ carbon storage. Chem. Geol. 2014, 373, 93−105. (24) Xu, T. TOUGHREACT user’s guide: A simulation program for non-isothermal multiphase reactive geochemical transport in variably saturated geologic media, V1. 2.1. Lawrence Berkeley National Laboratory 2008, LBNL-55460-2008. (25) Wang, K.; Xu, T.; Tian, H.; Wang, F. Impacts of mineralogical compositions on different trapping mechanisms during long-term CO2 storage in deep saline aquifers. Acta Geotech. 2016, 11, 1167− 1188. (26) Ennis-King, J.; Paterson, L. Coupling of geochemical reactions and convective mixing in the long-term geological storage of carbon dioxide. Int. J. Greenhouse Gas Control 2007, 1, 86−93. (27) Riaz, A.; Hesse, M.; Tchelepi, H.; Orr, F. Onset of convection in a gravitationally unstable diffusive boundary layer in porous media. J. Fluid Mech. 2006, 548, 87−111. (28) Stolper, E. M.; DePaolo, D. J.; Thomas, D. M. Deep drilling into a mantle plume volcano: The Hawaii Scientific Drilling Project. Sci. Drill. 2009, 7, 4−14. (29) Mito, S.; Xue, Z.; Ohsumi, T. Case study of geochemical reactions at the Nagaoka CO 2 injection site, Japan. Int. J. Greenhouse Gas Control 2008, 2, 309−318. (30) Rochelle, C. A.; Bateman, K.; Pearce, J. M. Geochemical interactions between supercritical CO2 and the Utsira Formation: an experimental study. British Geological Survey, Commissioned Report CR/02/006, 2002 (31) Luquot, L.; Andreani, M.; Gouze, P.; Camps, P. CO 2 percolation experiment through chlorite/zeolite-rich sandstone (Pretty Hill Formation−Otway Basin−Australia). Chem. Geol. 2012, 294, 75− 88. (32) Watson, M. N.; Zwingmann, N.; Lemon, N. M. The Ladbroke Grove−Katnook carbon dioxide natural laboratory: a recent CO 2 accumulation in a lithic sandstone reservoir. Energy 2004, 29, 1457− 1466.

2084


Rates of CO2 Mineralization in Geological Carbon ... - ACS Publications

Recommend Documents