Article pubs.acs.org/est
Metal Release from Sandstones under Experimentally and Numerically Simulated CO2 Leakage Conditions Katie Kirsch,†,‡ Alexis K. Navarre-Sitchler,*,†,‡ Assaf Wunsch,†,§ and John E. McCray†,§ †
Hydrologic Science and Engineering Program, Colorado School of Mines, Golden, Colorado 80401, United States Department of Geology and Geological Engineering, Colorado School of Mines, Golden, Colorado 80401, United States § Department of Civil and Environmental Engineering, Colorado School of Mines, Golden, Colorado 80401, United States ‡
S Supporting Information *
ABSTRACT: Leakage of CO2 from a deep storage formation into an overlying potable aquifer may adversely impact water quality and human health. Understanding CO2-water-rock interactions is therefore an important step toward the safe implementation of geologic carbon sequestration. This study targeted the geochemical response of siliclastic rock, specifically three sandstones of the Mesaverde Group in northwestern Colorado. To test the hypothesis that carbonate minerals, even when present in very low levels, would be the primary source of metals released into a CO2-impacted aquifer, two batch experiments were conducted. Samples were reacted for 27 days with water and CO2 at partial pressures of 0.01 and 1 bar, representing natural background levels and levels expected in an aquifer impacted by a small leakage, respectively. Concentrations of major (e.g., Ca, Mg) and trace (e.g., As, Ba, Cd, Fe, Mn, Pb, Sr, U) elements increased rapidly after CO2 was introduced into the system, but did not exceed primary Maximum Contaminant Levels set by the U.S. Environmental Protection Agency. Results of sequential extraction suggest that carbonate minerals, although volumetrically insignificant in the sandstone samples, are the dominant source of mobile metals. This interpretation is supported by a simple geochemical model, which could simulate observed changes in fluid composition through CO2-induced calcite and dolomite dissolution.
■
INTRODUCTION A proposed strategy to reduce anthropogenic carbon dioxide (CO2) emissions and mitigate climate change is to capture CO2 from the flue gas at power plants, compress it, and pump it underground into the pore space of rock.1 At the pressure and temperature of targeted injection formations, CO2 would exist as a supercritical fluid, but still be less dense than the groundwater. As a result, it would migrate upward toward the surface, and if faults, fractures, or faulty wellbores punctured the caprock, it could leak into an overlying, potable aquifer. The primary concern for such a leak is the dissolution of CO2 into groundwater, which acidifies the water and may mobilize naturally occurring toxic metals through mineral dissolution or desorption from mineral surfaces.2−5 CO2-induced metal mobilization has been well documented in laboratory experiments,4,6−10 modeling simulations,2,11−15 and field tests,3,16−18 however few studies have focused explicitly on identifying which mineral phases control the aqueous metal concentrations. Most numerical models simulate sulfide dissolution as the primary source (e.g., As in Arsenoypyrite, FeAsS; Pb in Galena, PbS; Cd in Greenockite, CdS),2,13,15,19,20 even though carbonate minerals would likely contain impurities (e.g., Co, Pb, and Cd),21,22 and may be more reactive than sulfides in the acidic, low-oxygen conditions expected in a CO2-impacted, confined aquifer. Using a reactive transport model, Navarre-Sitchler et al.12 found concentrations of Pb were higher in simulations where calcite was the source © 2014 American Chemical Society
than in simulations where galena was the source. Batch reaction experiments also suggest that rapid dissolution of impure carbonate minerals is the primary mechanism controlling the release of metals in carbonate aquifers,9,10 as well as in sandstone aquifers.7,23 This study further investigates the potential for metal release from carbonate mineral dissolution in sandstone aquifers impacted by CO2 leakage with the purpose of testing the hypothesis that carbonate minerals, even when present in very low levels, would be the largest source. We perform batch reaction experiments on sandstones, but at low oxygen conditions typical of confining aquifers, and in parallel with a detailed characterization of the minerals. If carbonate dissolution is found to consistently be the primary contributor to water quality degradation, knowing the distribution and chemical composition of these fast-reacting minerals will aid in assessing risk to fresh water resources overlying potential CO2 sequestration sites.
■
MATERIALS AND METHODS Three sandstone samples were collected from Mesaverde outcrop in northwest Colorado: two from the Iles Formation Received: Revised: Accepted: Published: 1436
July 16, 2013 December 20, 2013 January 14, 2014 January 14, 2014 dx.doi.org/10.1021/es403077b | Environ. Sci. Technol. 2014, 48, 1436−1442
Environmental Science & Technology
Article
(Campbell Scientific) (SI Figure S2). A stream of premixed CO2−N2 gas or pure CO2 gas (General Air Service) was piped into the headspace of each vessel, whose pressure was maintained at 1 bar by a downstream regulator (Equilibar, model QPV). The vessels were continuously agitated on a slowrotating shaker table to maximize exposure of the rock sample to the water. Water samples (4 mL) were extracted at 0, 1, 5, 12, 24, 48, 96, 144, 192, 240, 288, 336, 408, 480, 552, and 648 h, which caused a decrease in the water−rock ratio of less than 10%. They were filtered to 0.45 μm (Whatman syringe filters), acidified to a pH
