LBNL is one of the main international research organizations addressing concerns about potential impact of deployment of CO2 geological storage on the nation’s groundwater resources. A significant body of our work has targeted the possible groundwater quality changes in response to leakage of CO2 from deep sequestration reservoirs if it were to occur. To better understand and be able to predict this risk, we have developed and applied capabilities that combine sophisticated numerical modeling with laboratory and field experiments. Several past and present research projects are listed below:
- Understanding Groundwater Quality Changes in Case of CO2 Intrusion by Numerical Modeling
- Groundwater Chemistry Changes as a Result of CO2 Injection at the ZERT Field Site in Bozeman, Montana
- Laboratory Studies of Effects of Dissolved Carbon Dioxide (CO2) on Trace Metal Release from Shallow Sediments
- Potential Impact of Carbon Dioxide on Potable Groundwater: A Controlled Release Experiment at Plant Daniel, Mississippi
- Assessment of Potentially Deleterious Effect of CCS Operations on Groundwater Quality in California
Preview the project publications list here.
Funding Period: 2007-2010
Primary Contact: Jens Birkholzer
Proper site selection and management of geological storage projects will ensure that the risks of carbon sequestration to human health and the environment are low. However, there are realistic scenarios under which CO2 could migrate from the deep storage formation(s) to shallower aquifers, and very little research has been done regarding the potential impact of such leakage. In this project, the consequences of CO2 leakage on groundwater quality were evaluated to provide sound scientific information to regulators and the public. Injection of high-pressure CO2 could impact shallow aquifers through several processes: (1) CO2 gas migrating from depth could reach an underground source of drinking water (USDW) and dissolve in the water, increasing the acidity of the groundwater, which could enhance the solubility of inorganic hazardous constituents. (2) In deep storage formations, the enhanced solvent properties of CO2 are expected to lead to the leaching of organic compounds present in the reservoir rock matrix (e.g., aromatic hydrocarbons such as benzene). Subsequent transport of the contaminated CO2 from depth and intrusion into a USDW could result in contamination by these hazardous organic compounds. (3) Contaminants such as H2S, a byproduct of coal gasification, could be co-injected with CO2. H2S would preferentially partition into the formation brine, but H2S-bearing CO2 could also leak into USDWs and adversely affect water quality. H2S could furthermore interact with organic matter in the CO2 reservoir.
The following three subtasks were conducted during this three-year project to evaluate the potential hydrochemical impact of CO2 storage projects on USDWs.
CO2-Related Dissolution of Heavy Metals and Other Constituents in USDWs. For this sub-task, we investigated by numerical modeling the water quality changes upon intrusion of CO2 into potable groundwater. The intruding CO2 lowers groundwater pH and thereby can enhance the solubility of many metals present in the aquifer sediments (including heavy metals). How and to what extent groundwater quality would be affected depends largely on the initial abundance and distribution of these constituents in the aquifers, as well as on the aquifer mineralogy and the oxidation state. Using the USGS NWIS (National Water Quality Information System) data base, we have conducted a systematic evaluation of more than 38,000 groundwater quality analyses from aquifers throughout the United States that report non-zero concentrations of selected hazardous constituents. The results of the evaluation were employed to set up an equilibrium geochemical model of the aquifer chemistry in order to estimate the distribution of each heavy metal between the aqueous phase and adsorption and ion exchange sites, and in solid solution in primary and secondary minerals. Important qualitative conclusions can be drawn immediately from this evaluation regarding the geochemical vulnerability of the groundwater. For quantitative evaluation, we used the equilibrium geochemical model as a starting point for reactive geochemical transport simulations that predict the impact of CO2 intrusion into a fresh-water aquifer and the related changes to the host rock mineralogy and water chemistry. Our simulation results suggested that CO2 ingress into a shallow aquifer can lead to significant mobilization of lead and arsenic, contaminating the groundwater near the intrusion location and further downstream. Our simulations indicated that adsorption/desorption via surface complexation is arguably the most important process controlling the fate of hazardous constituents mobilized by CO2 leakage, but many conditions and processes impact the severity and spatial distribution of groundwater contamination in response to CO2 intrusion. Details of this work can be found in Birkholzer et al. (2008, LBNL-1251E), Apps et al. (2010, Transport in Porous Media, 82:215–246) and Zheng et al. (2009, Chemical geology, 268(3-4): 281-297)
Leaching of Organic Compounds. Organic compounds that could potentially be mobilized by supercritical CO2 were evaluated in terms of probability of occurrence and abundance in deep storage formations, as well as potential for environmental hazard. Benzene, PAHs, and phenols were identified as good candidates for further evaluation. Of these, benzene was selected for further investigation by numerical modeling. A literature search was conducted to obtain reasonable initial geochemical conditions in the modeled reservoir, and define appropriate ranges for key model input parameters. Two-dimensional (2-D) models were developed for a hypothetical scenario in which benzene is mobilized by supercritical CO2 injection into a deep saline reservoir, and then transported vertically along a preferential pathway into an overlying shallow aquifer. Model results indicate that benzene may co-migrate with CO2 into overlying aquifers if a leakage pathway were present. Because the aqueous solubility of benzene in contact with CO2 is lower than the aqueous solubility of CO2, benzene becomes enriched in the CO2 phase at the front of the CO2 plume as it advances. Details of this work can be found in Zheng et al. (2013, International Journal of Greenhouse Gas Control 14(0): 113-127) and Zheng et al. (2010, LBNL-4339E).
Impact of Co-Injection of H2S. Because the separation of some gas impurities from power plant effluents is quite costly, and their release to the atmosphere is environmentally harmful, it may be advantageous to consider co-injecting these gases with CO2. If the co-injection of H2S is deployed, the impact of CO2+H2S leakage from deep saline reservoirs to overlying shallow aquifers is also a question that the scientific community must answer for the public. Simulations were conducted for the case of H2S co-injection with CO2. Three different 2-D models were set up to evaluate the potential for groundwater contamination by H2S. The first model comprised the storage formation and a sealing cap rock. This model was established to assess the fate of H2S within the storage formation. A second model similar to that used for the benzene transport case was developed to investigate how H2S might migrate along a leakage pathway. Modeling results show that H2S arrival in the shallower aquifer is delayed in comparison with that of CO2 due to the preferential dissolution of H2S into the aqueous phase. A third model was used to evaluate the potential consequences of CO2+H2S leaking into a hypothetical shallow groundwater aquifer. Model results showed that the potentially adverse impacts of leakage on shallow groundwater quality may be exacerbated for cases of leaking CO2+H2S, compared to intrusion of pure CO2, possibly leading to the mobilization of thiophilic elements such as arsenic. Geochemical reactions included in the simulations involve adsorption/desorption, reductive dissolution of goethite, precipitation of pyrite, siderite, and arsenic sulfide phases. The models presented are generic in nature, exploring important processes regarding organic compounds and co-injected H2S, and calling attention to the need for more site-specific studies taking into account the variability and uncertainty of key hydrogeologic and geochemical parameters. Details of this work can be found in Zheng et al. (2013, International Journal of Greenhouse Gas Control 14(0): 113-127) and Zheng et al. (2010, LBNL-4339E).
Groundwater Chemistry Changes as a Result of CO2 Injection at the ZERT Field Site in Bozeman, Montana
Funding Agency: Funded primarily by the Electric Power Research Institute, EPRI; additional funds were also obtained from EPA, DOE-NETL, LBNL and USGS
Funding Period: 2008-2010
Primary Contact: Jens Birkholzer
In this project, we studied the geochemical and hydrological changes in a shallow CO2-injection test performed by the Zero Emission Research and Technology Center (ZERT) during July and August 2008 at a field site located on Montana State University (MSU) property near Bozeman, Montana. From July 9 to August 7, 2008, approximately 300 kg/day of food-grade CO2 was injected through a perforated pipe placed approximately 1.1-2.5 m below ground surface, and about 0.5–1 m below the water table. The objective of this project was to investigate, using a small-scale field experiment, whether sequestered CO2 released from a hypothetical geologic storage reservoir would have an adverse impact on USDW. The objective was accomplished by collecting approximately 80 shallow groundwater samples before, during and after CO2 gas injection and analyzing these samples for their concentrations in major, minor and trace inorganic and organic compounds and by interpreting these observations with geochemical modeling. The study was undertaken as a collaborative project between the Earth Sciences Division at LBNL and the U.S. Geological Survey (USGS), in part funded by the Electronic Power Research Institute (EPRI).
In the test, field measurements and laboratory analyses showed rapid and systematic changes in pH, alkalinity, and conductance, as well as increases in the aqueous concentrations of naturally occurring major and trace element species. These changes were simulated using a multicomponent reaction path model to assess possible geochemical processes responsible for the observed increases in the concentrations of dissolved constituents. Reasonable agreement between observed and modeled data suggests that (1) calcite dissolution was the primary pH buffer, yielding increased Ca+2 concentrations in the groundwater, (2) increases in the concentrations of most major and trace metal cations except Fe could be a result of Ca+2-driven exchange reactions, (3) the release of anions from adsorption sites due to competitive adsorption of carbonate could explain the observed trends of most anions, and (4) the dissolution of reactive Fe minerals (presumed ferrihydrite and fougerite-CO3) could explain increases in total Fe concentration. Details of this work can be found in Ambats et al. (2009, LBNL-2931E), Kharaka et al. (2010, Env. Earth. Sci., 60 (2), 273–284) and Zheng et al. (2012, International Journal of Greenhouse Gas Control 7(0): 202-217).
Laboratory Studies of Initial Effects of Dissolved Carbon Dioxide (CO2) on Metal Release from a Shallow Aquifer
Funding Period: 2009-2012
Primary Contact: Peter Nico
In this study, results from laboratory experiments were combined with extensive mineralogical characterization, and data from a concurrent field test to identify metals that could be mobilized by CO2 intrusion into a shallow, sandy groundwater formation located at a depth of ~ 50 m. Sequential leaching experiments were conducted under in situ redox conditions, exposing heterogeneous sediments from the field site to synthetic groundwater solutions that were either saturated with CO2 (pH ~ 5) or N2 (pH ~ 8.5) at formation pressures using a relatively low-cost, scalable setup. Comparable leaching experiments were carried out using CO2-free, pH-amended solutions to identify the mechanisms underlying metal release, and to particularly distinguish between two effects of CO2 dissolution – i.e. a decrease in solution pH, and the relesae of carbonate-ligands into solution. We have demonstrated that a systematic categorization of released metals is possible based on the kinetics of metal release reactions, even if sediments are geochemically and mineralogically different. We have also shown that lab tests such as the pH amendments in conjunction with routine sediment characterizations can be used as cost-effective screening protocols to predict metal release when field tests are not possible.
Our results indicate that some metals can be released from sediments due to the addition of dissolved CO2, by primarily pH-driven processes. The results show that a pH decrease can lead to an initial fast mobilization of primarily alkaline and alkaline earth metals (Ca, Mg, Ba, Sr, Na, Li, Rb), and a few other elements (Co, Fe, Ge, Mn, Ni, Si, Zn), with release concentrations significantly decreasing over time. In addition, carbonate ligands appear to either enhance (U, Ba, As, Mo, Sr, Mn, Co, Ge, Mg) or suppress (weak trends observed for Fe, and Li) the release of metals. Of the metals that have a tendency to be released in low pH conditions, those that exhibit fast kinetics have the highest potential for immediate mobilization under elevated CO2 levels. A fast initial release is likely caused by the exchange of ions adsorbed onto mineral surfaces, and is potentially driven by Ca (or Mg) ions that can be released into solution by the dissolution of carbonates at lower pHs. Consequently, the slower release kinetics that follows the initial pulse may be affected by the presence of carbonate ligands, which can either enhance or suppress metal release under low pH conditions. In all cases, mineral dissolution reactions may also occur under low pH conditions; the metal release from dissolution will exhibit a slow steady increase that may be overshadowed by the initial pulse and may not become apparent until later on. Slower reactions at low pH would be especially relevant in the cases of aquifers with low buffering capacities and high groundwater residence times (slow groundwater velocities), where the pH may not rebound quickly to background conditions leading to sustained mineral dissolution over long time frames.
Potential Impact of Carbon Dioxide on Potable Groundwater: A Controlled Release Experiment at Plant Daniel, Mississippi, USA
Funding Agency: EPRI
Funding Period: 2009-2013
Primary Contact: Jens Birkholzer
- Charuleka Varadharajan
- Ruth M. Tinnacher
- Liange Zheng
- Nic Spycher
- Marco Bianchi
- Peter S. Nico
- Yuxin Wu
- Baptiste Dafflon
- Susan S. Hubbard
The project involved an integrated field experiment of the controlled release of dissolved CO2 into dilute groundwater in an isolated permeable sandy aquifer. The goals of the study were to investigate the potential impact that dissolved CO2 has on in situ groundwater quality, to apply and test monitoring methods, and to evaluate geochemical modeling concepts and results. For this test, groundwater was withdrawn from a pumping well, carbonated, then re-injected back through an injection well into the same aquifer (dipole, closed-loop system). Four monitoring wells weer located downstream of the injection well for water sampling and analysis. The concentration of major ions and trace elements were measured before, during and after the injection of carbonated groundwater.
The mobilization of major and minor cations in response to CO2 injection was evidenced by short-term concentration increases coinciding with the arrival of dissolved CO2. After an initial pulse-like increase, the concentrations fell back and stabilized at levels slightly higher than original values until the end of the injection period. However, concentrations started to increase again after the injection stopped, likely because the increased groundwater residence time (slower velocity) allowed for further reaction between the groundwater and the sediments. Statistically significant increases above baseline concentrations were observed for Ba, Ca, Fe, Li, Mg, Mn, K, Na, Si, Sr, HCO3– and electrical conductance. However, the concentration of many species remained below detection limits, and no obvious concentration increases from background were observed for trace solutes such as Al, Sb, As, Cd, Cu, Pb, Hg, Mo, P, Se, Ag, Ti, Zn and other species such as Br, NH3, Cl, F, NO3-, NO2-, SO42-, sulfides. Observed and modeled results captured the observed pulse response for Ca, Sr, Mg, Ba, Mn, and Fe, which suggested a fast release mechanism with depletion of metals from their source at the front of the low-pH plume (such as desorption, exchange, and/or fast dissolution of small finite amounts of minerals), possibly overlapping with a slower and more continuous release from mineral dissolution. Details of this work can be found in Trautz et al., (2012, Environmental Science & Technology 47(1): 298-305).
Funding Agency: California Energy Commission (CEC)
Funding Period: 2013-2015
Primary Contact: Peter Nico
Similar to the previous projects, in this project we expand these research activities to fresh-water aquifers in California. We propose to study potential water quality impacts at each stage of a hypothetical CO2 leak, from leaching of storage aquifer cap rock, to release of metals from shallow-aquifer sediment samples, to transport of impacted water to the surface. We will use materials from existing carbon capture and storage (CCS) research projects in California, and/or from new boreholes advanced in areas of interest. The determination of such key areas and potential synergies with ongoing CCS operations in California will be part of the proposed work. Specific tasks will be to a) assess the quantity and quality of supercritical CO2-leachable organic material from typical caprock, b) determine the impact of that organic material on metal mobilization and microbial activity within aquifers, c) assess mineral-trace metal associations in typical aquifers near areas deemed most suitable for CCS operations in California with particular emphasis on metals with known soluble carbonate complexes such as uranium (in 2008, uranium concentrations exceeded MCL in 16% of wells tested within the San Joaquin Valley and were within an order of magnitude of the MCL for 65% of the wells), d) determine the most probable mechanisms of water quality impacts as a result of CO2 leaks (through laboratory experiments leveraged with our ongoing projects), e) evaluate the extent to which these impacts reverse themselves upon depressurization during transport to the surface, and f) use this information to develop and test models that can be applied to further evaluate the mobilization, fate and transport of trace metals in groundwater by CO2 leaks from CCS operations in California. By building on existing research projects, these goals will be reached in a cost-effective manner. Also, assessment of sediment samples from shallow fresh-water aquifers in California will augment the suite of aquifer types currently being considered in efforts to develop a national risk-based assessment of CCS operations, to which we are taking part under leadership and funding by the NETL (DOE Fossil Energy) National Risk Assessment Program (NRAP).