Peter C. Lichtner

Analysis and performance of oil well cement with 30 years of CO2 exposure from the SACROC Unit, West Texas, USA

Abstract A core sample including casing, cement, and shale caprock was obtained from a 30-year old CO2-flooding operation at the SACROC Unit, located in West Texas. The core was investigated as part of a program to evaluate the integrity of Portland-cement based wellbore systems in CO2-sequestration environments. The recovered cement had air permeabilities in the tenth of a milliDarcy range and thus retained its capacity to prevent significant flow of CO2. There was evidence, however, for CO2 migration along both the casing–cement and cement–shale interfaces. A 0.1–0.3 cm thick carbonate precipitate occurs adjacent to the casing. The CO2 producing this deposit may have traveled up the casing wall or may have infiltrated through the casing threads or points of corrosion. The cement in contact with the shale (0.1–1 cm thick) was heavily carbonated to an assemblage of calcite, aragonite, vaterite, and amorphous alumino-silica residue and was transformed to a distinctive orange color. The CO2 causing this reaction originated by migration along the cement–shale interface where the presence of shale fragments (filter cake) may have provided a fluid pathway. The integrity of the casing–cement and cement–shale interfaces appears to be the most important issue in the performance of wellbore systems in a CO2 sequestration reservoir.

Diffusion and reaction in rock matrix bordering a hyperalkaline fluid-filled fracture

Abstract Multicomponent diffusive and advective-dominant transport-reaction calculations are used to analyze water-rock alteration in rock matrix adjacent to a hyperalkaline fluid-filled fracture. The calculations indicate that rock alteration resulting from diffusive transport may be fundamentally different from that observed in the case of advective-dominant transport. Since advective-dominant and diffusive transport can result in differing reaction products as a function of time and space, the two transport processes may modify the chemical and physical properties of the rock at different rates. We apply reactive transport calculations to an analogue of the Cretaceous and Tertiary marls proposed as host rocks for the Swiss low-level nuclear waste repository. Diffusion-reaction calculations predict that the rock matrix bordering high-pH fluid-filled fractures could be completely cemented within 10 to 500 years. The bulk of the porosity reduction occurs because of the precipitation of calcite resulting from the interdiffusion of Ca 2+ and CO 2− 3 . In contrast, advective-dominant transport results in the precipitation of calcite only as a replacement of dolomite. Both advective-dominated and diffusive transport result in a porosity increase within millimeters or less of the fracture wall, an effect which could widen the fracture and thus increase the rate of radionuclide transport along the fracture. Because solute diffusion is coupled to porosity and tortuosity change in the rock matrix, cementation causes the fracture to become physically and chemically isolated from the rock matrix if no expansion of the rock occurs. As a consequence, reaction-induced porosity reduction may potentially decrease the buffering and sorbing capacity of a fractured host rock, thus reducing the physical and chemical retardation of contaminants migrating along fractures. These effects may occur within the time required for radionuclides in the repository to decay to environmentally safe levels.

On the upscaling of reaction-transport processes in porous media with fast or finite kinetics

Abstract We show that for reaction-transport processes with fast kinetics (in the limit of thermodynamic equilibrium), conventional volume averaging for determining effective kinetic parameters applies only when the macroscopic variable approaches its equilibrium value. Even under such conditions, computing the effective mass transfer coefficient requires solving an eigenvalue problem, which couples the local microstructure with the global. Two examples, one involving a simple advection–dissolution problem and another a drying problem in a pore network, illustrate the theoretical predictions. Similar considerations apply for the case of finite kinetics, when the macroscale concentration approaches an equilibrium value. In that case, the effective kinetic parameter is not equal to the local, as typically assumed, but it becomes a function of the local Thiele modulus.

Redox chemistry of iron and manganese minerals in river-recharged aquifers: A model interpretation of a column experiment

Abstract The transport of Fe and Mn observed in a laboratory column filled with sections of riverbed sediments and fine grained aquifer materials is modeled for steady-state conditions. The simulated processes include oxidation of organic matter by O 2 and NO - 3 in the riverbed sediment section, reductive dissolution of Fe and Mn (hydr) oxides, and precipitation of dissolved Fe and Mn, mainly as amorphous FeS. Local equilibrium for aqueous species is assumed. Kinetic reactions are used in the calculations for the oxidation of organic matter and for dissolution and precipitation of minerals. The observed concentration profiles of H + , NO - 3 , SO 2- 4 , and dissolved Fe and Mn in the laboratory column are reasonably explained by the model. The most sensitive model parameters are the rate constant for decomposition of organic matter and those for the dissolution of Fe and Mn (hydr) oxides. The results of a sensitivity analysis suggest that the decomposition rate of organic matter is responsible for the observed seasonal variations in concentrations of dissolved Fe and Mn in river-recharged aquifers, such as at Glattfelden in Switzerland. The observed reactions are mediated by bacteria.

Evaluation of trapping mechanisms in geologic CO2 sequestration: Case study of SACROC northern platform, a 35-year CO2 injection site

CO2 trapping mechanisms in geologic sequestration are the specific processes that hold CO2 underground in porous formations after it is injected. The main trapping mechanisms of interest include (1) fundamental confinement of mobile CO2 phase under low-permeability caprocks, or stratigraphic trapping, (2) conversion of CO2 to mineral precipitates, or mineral trapping, (3) dissolution in in situ fluid, or solubility trapping, and (4) trapping by surface tension (capillary force) and, correspondingly, remaining in porous media as an immobile CO2 phase, or residual CO2 trapping. The purpose of this work is to evaluate and quantify the competing roles of these different trapping mechanisms, including the relative amounts of storage by each. For the sake of providing a realistic appraisal, we conducted our analyses on a case study site, the SACROC Unit in the Permian basin of western Texas. CO2 has been injected in the subsurface at the SACROC Unit for more than 35 years for the purpose of enhanced oil recovery. Our analysis of the SACROC production and injection history data suggests that about 93 million metric tons of CO2 were injected and about 38 million metric tons were produced from 1972 to 2005. As a result, a simple mass-balance suggests that the SACROC Unit has accumulated approximately 55 million metric tons of CO2. Our study specifically focuses on the northern platform area of the SACROC Unit where about 7 million metric tons of CO2 is stored. In the model describing the SACROC northern platform, porosity distributions were defined from extensive analyses of both 3-D seismic surveys and calibrated well logging data from 368 locations. Permeability distributions were estimated from determined porosity fields using a rock-fabric classification approach. The developed 3-D geocellular model representing the SACROC northern platform consists of over 9.4 million elements that characterize detailed 3-D heterogeneous reservoir geology. To facilitate simulation using conventional personal computers, we upscaled the 9.4 million elements model using a “renormalization” technique to reduce it to 15,470 elements. Analysis of groundwater chemistry from both the oil production formations (Cisco and Canyon Groups) and the formation above the sealing caprock suggests that the Wolfcamp Shale Formation performs well as a caprock at the SACROC Unit. However, results of geochemical mixing models also suggest that a small amount of shallow groundwater may be contaminated by reservoir brine possibly due to: (1) downward recharge of recycled reservoir brine from brine pits at the surface, or (2) upward leakage of CO2-saturated reservoir brine through the Wolfcamp Shale Formation. Using the upscaled 3-D geocellular model with detailed fluid injection/production history data and a vast amount of field data, we developed two separate models to evaluate competing CO2 trapping mechanisms at the SACROC northern platform. The first model simulated CO2 trapping mechanisms in a reservoir saturated with brine only. The second model simulated CO2 trapping mechanisms in a reservoir saturated with both brine and oil. CO2 trapping mechanisms in the brine-only model show distinctive stages accompanying injection and post-injection periods. In the 30-year injection period from 1972 to 2002, the amount of mobile CO2 increased to 5.0 million metric tons without increasing immobile CO2, and the mass of solubility-trapped CO2 sharply rose to 1.7 million metric tons. After CO2 injection ceased, the amount of mobile CO2 dramatically decreased and the amount of immobile CO2 increased. Relatively small amounts of mineral precipitation (less than 0.2 million metric tons of CO2 equivalent) occurred after 200 years. In the brine-plus-oil model, dissolution of CO2 in oil (oil-solubility trapping) and mobile CO2 dominated during the entire simulation period. While supercritical-phase CO2 is mobile near the injection wells due to the high CO2 saturation, it behaves like residually trapped CO2 because of the small density contrast between oil and CO2. In summary, the brine-only model reflected dominance by residual CO2 trapping over the long term, while CO2 in the brine-plus-oil model was dominated by oil-solubility trapping.

Simulating Subsurface Flow and Transport on Ultrascale Computers using PFLOTRAN

We describe PFLOTRAN, a recently developed code for modeling multi-phase, multi-component subsurface flow and reactive transport using massively parallel computers. PFLOTRAN is built on top of PETSc, the Portable, Extensible Toolkit for Scientific Computation. Leveraging PETSc has allowed us to develop—with a relatively modest investment in development effort—a code that exhibits excellent performance on the largest-scale supercomputers. Very significant enhancements to the code are planned during our SciDAC-2 project. Here we describe the current state of the code, present an example of its use on Jaguar, the Cray XT3/4 system at Oak Ridge National Laboratory consisting of 11706 dual-core Opteron processor nodes, and briefly outline our future plans for the code.

Time‐space continuum description of fluid/rock interaction in permeable media

The feasibility of integrating multicomponent mass transport equations over geologic time spans is demonstrated for the case of pure advection in a homogeneous porous medium. The mathematical formulation of the problem is based on the quasi-stationary state approximation, or multiple reaction path description, in which the time evolution of a geochemical system is represented by a sequence of stationary states or reaction paths. The method is implemented in the computer code MPATH which solves the transport equations in a single spatial dimension taking into account irreversible mineral precipitation/dissolution reactions and local equilibrium of aqueous complexing reactions. An adaptive grid enables the positions of reaction zones, with widths which vary over many orders of magnitude and which move with greatly differing velocities, to be tracked simultaneously over geologic time spans. There appears to be virtually no limitation to the number of chemical species that can be included in the code without rendering the computational effort beyond the bounds of a high-performance workstation. The numerical accuracy of the solution can be verified through global mass conservation equations and by comparing the asymptotic kinetic solution with the corresponding solution to algebraic equations representing local equilibrium conditions for pure advective transport, if such solutions exist. The code MPATH is applied to several examples including migration of redox fronts, weathering and hydrothermal alteration in a spatially varying temperature field. These examples demonstrate the absolute necessity of solving the governing transport equations for sufficiently long time spans in order to fully characterize the behavior of the system.

Role of Competitive Cation Exchange on Chromatographic Displacement of Cesium in the Vadose Zone beneath the Hanford S/SX Tank Farm

Migration of radionuclides under the SX-tank farm at the Hanford nuclear waste complex involves interaction of variably water saturated sediments with concentrated NaOH–NaNO3–NaNO2 solutions that have leaked from the tanks. Constant K d models for describing radionuclide retardation are not valid under these conditions because of strong competition for sorption sites by abundant Na+ ions, and because of dramatically changing solution compositions with time as the highly concentrated tank fluid becomes diluted as it mixes with infiltrating rainwater. A mechanistic multicomponent sorption model is required that can account for effects of competition and spatially and temporally variable solution compositions. To investigate the influence of the high ionic strength tank fluids on Cs+ migration, numerical calculations are performed using the multiphase-multicomponent reactive transport code FLOTRAN. The computer model describes reactive transport in nonisothermal, variably saturated porous media including both liquid and gas phases. Pitzer activity coefficient corrections are used to describe the high ionic strength solutions. The calculations take into account multicomponent cation exchange based on measured selectivity coefficients specific to the Hanford sediments. Solution composition data obtained from Well 299-W23-19, documenting a moderately concentrated leak from the SX-115 tank, are used to calibrate the model. In addition to exchange of cations Na+, K+, Ca2+, and Cs+, aqueous complexing and a kinetic description of precipitation and dissolution of calcite are also included in the calculations. The fitted infiltration rate of 0.08 m yr−1, and fitted cation exchange capacity of 0.05 mol kg−1 are consistent with measured values for the Hanford sediments. A sensitivity analysis is performed for Na+ concentrations ranging from 5 to 20 m to investigate the mobility of Cs+ interacting with a highly concentrated background electrolyte solution believed to have been released from the SX-108/SX-109 tanks. The calculations indicate that during the initial period of the tank leak when Cs+ is associated with high Na+ concentrations, there is little retardation of the Cs+ plume. However, as time increases the Na+ and Cs+ profiles become chromatographically separated due to differences in their selectivity coefficients and dilution of the tank leak plume with infiltrating rainwater. Eventually the two species become separated spatially, and Cs+ becomes highly retarded and remains essentially fixed in the sediments by cation exchange. For the 20 m Na+ simulated tank leak, the sorbed Cs+ profile is in close agreement with data obtained from the slant borehole and consistent with the estimated tank supernatant concentration. The simulations suggest that natural attenuation processes should result in strong fixation of Cs+ in the vadose zone in spite of the release of high Na+ concentrations during a tank leak event.

Mechanism of Plagioclase Albitization

ABSTRACT Observations of closely spaced porous and cemented sandstones from the San Joaquin basin, the Los Angeles basin and the Gulf Coast reveal that albitization of plagioclase occurs during fracturing of the detrital plagioclase grains. Albitization is found to be a chemically selective hydrous reaction in which An-rich plagioclase is preferentially albitized first relative to An-poor plagioclase. The observed non-positive volume change during albitization indicates that aluminum mobility takes place on a scale at least as large as a thin section. These findings are supported by thermodynamic stability calculations for plagioclase with different compositions and degrees of aluminum-silicon ordering at 25°C in a representative sea water and at 100°C in an oil field water. The calculations predict that 1) An-rich plagioclase is less stable than An-poor plagioclase, and 2) plagioclase with a random Al/Si distribution is less stable than plagioclase with the same composition but an ordered Al/Si distribution. Any composition and structural state of plagioclase, including albite, is unstable in sea water at 25°C. The width of the albitization zone may be controlled by the degree of aluminum-silicon ordering. This agrees well with the observed smaller temperature interval of albitization in sandstones containing only volcanically derived plagioclase.

Fluid Flow, Heat Transfer, and Solute Transport at Nuclear Waste Storage Tanks in the Hanford Vadose Zone

on the Columbia River plateau, a semiarid region in south-central Washington (Fig. 1), the Hanford Site At the Hanford Site, highly radioactive and chemically aggressive served as a plutonium production facility for nuclear waste fluids have leaked from underground storage tanks into the vadose zone. This paper addresses hydrogeological issues at the weapons from 1944 to the end of the Cold War era. A 241-SX tank farm, especially focusing on Tank SX-108, which is one total of 177 large underground tanks were constructed of the highest heat load, supernate density and ionic strength tanks in the Hanford vadose zone to store waste fluid streams at Hanford and a known leaker. The behavior of contaminants in from the plutonium extraction facilities. Many of the the unsaturated zone near SX-108 is determined by an interplay of older single-shell tanks have leaked radioactive waste multiphase fluid flow and heat transfer processes with reactive chemifluids, posing a contamination hazard for the underlying cal transport in a complex geological setting. Numerical simulation aquifer and ultimately the Columbia River. Our analysis studies were performed to obtain a better understanding of mass and specifically addresses the 241-SX tank farm (Fig. 2), energy transport in the unique hydrogeologic system created by the SX tank farm. Problem parameters are patterned after conditions where highly radioactive and chemically aggressive at Tank SX-108, and measured data were used whenever possible. aqueous fluids of high ionic strength have leaked into Borrowing from techniques developed in geothermal and petroleum the vadose zone. Of interest is the nature and extent of reservoir engineering, our simulations feature a comprehensive desubsurface contamination; the past, present, and future scription of multiphase processes, including boiling and condensation migration of contaminants; and hydrogeologic issues phenomena, and precipitation and dissolution of solids. We find that posed by possible future remedial actions. the thermal perturbation from the tank causes large-scale redistribuConstruction and operation of the SX storage tanks tion of moisture and alters water seepage patterns. During periods since the mid 1950s involved massive perturbations of of high heat load, fluid and heat flow near the tank are dominated by vapor–liquid counterflow (heat pipe), which provides a much more the natural hydrogeologic system. At the SX tank farm, efficient mechanism than heat conduction for dissipating tank heat. the ground was excavated down to 15.4 m depth, and The heat pipe mechanism is also very effective in concentrating dis15 large cylindrically shaped storage tanks with approxisolved solids near the heat source, where salts may precipitate even mate dimensions of 13.6-m height and 11.8-m radius if they were only present in small concentrations in ambient fluids. were emplaced in a regular pattern with 30.4-m spacing Tank leaks that released aqueous fluids of high ionic strength into between tank centers (Fig. 2, 3). The excavated material the vadose zone were also modeled. The heat load causes formation was then backfilled, and a gravel layer 1.8 m thick was dry-out beneath the tank, which is accompanied by precipitation of placed on top. The latter led to a substantial reduction solutes. These may become remobilized at a later time when tank temperatures decline and previously dried out regions are rewetted. in evapotranspiration and increase in net infiltration. Simulated temperature and moisture distributions compare well with Water migration in the unsaturated zone was profoundly borehole measurements performed in 2000. The temperature maxialtered by the umbrella effect of the tanks that diverts mum observed beneath Tank SX-108 can be explained from past water around the tank perimeters and by altered hydrothermal history of the tank; it is not necessary to invoke heat generageologic properties in the backfilled region. Additional tion from leaked radioactive contaminants. A novel composite meeffects arose from the heat generated by the radioactive dium model is used to explore effects of moisture tension–dependent wastes in the tanks. Temperatures in several tanks rose anisotropy, which is shown to have important impacts on fluid flow to well above the nominal boiling point of 100 C, in one and solute transport in the Hanford sediments. case up to 160 C, for extended time periods (Fig. 4). This caused elevated formation temperatures with vaporization–condensation effects and associated redistriA unusual hydrogeologic system, namely, the bution of moisture and solutes. Tank leaks introduced large complex of underground storage tanks at the into the subsurface hot and highly saline aqueous fluids, Hanford reservation is the subject of this study. Located whose thermophysical properties and flow behavior K. Pruess, Earth Sciences Division, Lawrence Berkeley National Labmay be quite different from pure water. Further changes oratory, University of California, Berkeley, CA 94720; Steve Yabuin flow behavior could result from chemical alteration saki, Pacific Northwest National Laboratory, Richland, WA 99352; of the sediments because of reactions with the fluids. Carl Steefel, Energy and Environment Sciences Directorate, LawThe physical and chemical processes affecting condirence Livermore National Laboratory, Livermore, CA 94551; Peter Lichtner, Earth and Environmental Sciences Division, Los Alamos tions at and around the tanks are being played out in National Laboratory, Los Alamos, NM 87545. Received 3 Dec. 2001. a complex natural setting that is subject to human alter*Corresponding author (K_Pruess@lbl.gov). ations and perturbations that are only imperfectly known. Modeling of the impacts of heat-generating tanks and Published in Vadose Zone Journal 1:68–88 (2002).