2003 Progress Report: In-Situ Containment and Treatment of Contaminated Sediments: Engineering Cap Integrity and ReactivityEPA Grant Number: R828773C002
Subproject: this is subproject number 002 , established and managed by the Center Director under grant R828773
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
Center: HSRC (2001) - South and Southwest HSRC
Center Director: Reible, Danny D.
Title: In-Situ Containment and Treatment of Contaminated Sediments: Engineering Cap Integrity and Reactivity
Investigators: Wiesner, Mark R. , Hughes, Joseph B , Willson, Clinton S. , Valsaraj, Kalliat T. , Edge, Billy
Institution: Rice University , Texas A & M University , Louisiana State University - Baton Rouge , Georgia Institute of Technology
Current Institution: Rice University , Georgia Institute of Technology , Louisiana State University - Baton Rouge , Texas A & M University
EPA Project Officer: Lasat, Mitch
Project Period: October 1, 2001 through September 30, 2006 (Extended to September 30, 2007)
Project Period Covered by this Report: October 1, 2002 through September 30, 2003
Project Amount: Refer to main center abstract for funding details.
RFA: Hazardous Substance Research Centers - HSRC (2001) Recipients Lists
Research Category: Hazardous Waste/Remediation , Land and Waste Management
The overall objective of this research project is to develop "second-generation" sediment caps with specific chemical transport characteristics and built-in sustainable reactivity. The specific objectives of this research project during 2002-2003 are to: (1) determine how postplacement processes such as consolidation, deposition, and colloidal transport impact cap function, and how to control cap stability and permeability (Weisner); (2) develop innovative approaches to be used during cap placement to achieve desired reactivity and transport characteristics (Weisner); (3) determine the long-term variability of caps subject to the dynamic processes identified above because of strong episodic storm events (Edge); and (4) develop a mathematical model that incorporates known mechanisms to design a cap (Valsaraj).
The specific objectives for additional proposed work during 2004 are to: (1) identify conditions leading to mobilization or entrapment of nonaqueous phase liquid (NAPL) by a cap (Willson); (2) identify gas generation resulting from microbial processes and subsequent effects on cap integrity and contaminant migration (Valsaraj and Hughes); and (3) evaluate the effectiveness of a conventional cap as a permeable reactive barrier (Hughes).
Because of the cost of dredging and concerns in the management of dredge spoils, it is imperative that in situ solutions be developed for the management and treatment of contaminated sediments. Capping, the placement of a clean layer of sediment over the contaminated bed, is an attractive solution in many cases, as it is effective in rapidly isolating contaminants from benthic organisms and can improve sediment stability. However, there are concerns regarding the long-term effectiveness of caps as a result of storm events and other phenomena (gas and NAPL migration) that may compromise the cap integrity. Successful implementation of a cap will require that depositional processes used in cap placement be tailored such that the desired cap characteristics are achieved. Work also is necessary to understand and predict the stability of the cap under various conditions. Our studies also focus on the development of sediment caps, where the contaminant destruction is achieved during the transport through the cap. The net effect will be the ability to reduce contaminant flux to the water column and allow for natural recovery of the sediment. To accomplish these objectives, it must become possible to engineer the cap such as for specific chemical transport characteristics, and to incorporate sustainable reactive processes (biotic and abiotic). These studies will form the basis for larger scale second-generation caps.
The work in the past 2 years (2002 and 2003) has accomplished much in regard to Objectives 1 and 3, with a few more items to be investigated next year. Objective 2 has received limited attention during the first 2 years of work. Objective 4 has been accomplished, and the modeling effort is complete. Only experimental verification is needed and is continuing, but no funds are requested. In terms of cap integrity, we are now at the point where we need to know more about the failure mechanisms for in situ capping at various sediment sites. In situ capping is not a strong candidate as a remediation option when contaminant transport is facilitated by the following: (1) migration of a NAPL enriched in the contaminant; (2) significant gas generation and migration; and (3) migration of groundwater by active seeps. The assumption that these conditions are inappropriate for capping, however, is largely because of the inability, given current state of the science, to adequately assess fate and transport processes under these conditions. The future works (see Objectives 5 through 7) for 2004 are aimed at resolving these issues.
This is a multidisciplinary, multi-institutional project. There already has been a large degree of integration among the groups. For example, the cells used by the Rice University group was previously designed at Louisiana State University (LSU) and modified for Objectives 1 and 2. The consolidation work of the Rice University group (Objective 1) formed the basis for the modeling work by the LSU group (see Objective 4). The same cells are now being used by the LSU group in conducting experiments to calibrate aspects of the model that pertains to contaminant transport. The Texas A&M work on episodic events now is progressing in a direction that incorporates the work conducted by the LSU group. Periodic meetings between the groups were conducted throughout the year to discuss these collaborative efforts.
Various investigators with complimentary strengths (Drs. Hughes, Willson, and Valsaraj) have been assembled to ascertain the issues relevant to Objectives 5 through 7. Issues relevant to Objectives 2 and 3 also are important in addressing the newer objectives, and the work of Drs. Wiesner and Edge will be integrated into the wider objectives within 5 through 7 as described in the section on proposed work. Dr. Hughes' work at Georgia Tech will form the basis for Objectives 5 through 7. The gas generation rates from representative sediments (e.g., Anacostia) will be ascertained in microcosms at Georgia Tech. Gas will be injected at these representative rates into bubble columns at LSU by the Valsaraj group to understand the facilitated transport of contaminants and colloids by gas bubbles. The Willson group at LSU will use the same rates in the sediment columns to understand the effects of bubble transport on sediment cap structure and migration of NAPL through x-ray tomography. The Edge group at Texas A&M will employ the same gas generation rates to study the effects of gas bubbles on scour rates of typical cap materials designed by the Weisner group at Rice University.
Objectives 1 and 2
An important task in developing new cap designs is the ability to form versatile structures that can be tailored to provide a range of permeabilities, suitable matrices for reactivity, and appropriate physical integrity. We have succeeded in creating such structures using bentonite/cement composites as a new material for the isolation of contaminated sediments. Having only limited control over the depositional stage of the cap formation process, we create various cap microstructures by controlling postdepositional chemical processes in a cap. The technological appeal of this choice is because of the richness of design choices stemming from the complexity of the microstructure and strength development in cementitious materials. By judiciously choosing cement/bentonite and liquid/solid ratios of the composite, desired transport and mechanical properties of the reactive barrier may be attained. Controllable permeability and the capacity of cement materials to bind heavy metals (Conner, 1989) make such barriers especially attractive.
Using chlorine as a tracer compound, diffusion tests were performed with a modified LSU diffusion cell (Wang, et al., 1991). Modifications allowed for a higher closed-circuit cross flows, pump pulse dumping section at the entrance, and sampling ports to monitor contaminant depletion in the contaminated sediment. Cap materials were mixed at a given proportion and at given water content for 15 minutes, and were placed over the model sediment to fill the remaining 16 mm of the bottom portion of the T-shaped diffusion cell. The upper compartment of the cell was purged with deionized water for 10 minutes to remove residuals of the tracer compound, which could be introduced during the cap placement stage. Aliquots (~ 10 mL) of the solution circulated above the caps were taken periodically to determine the concentration of chlorine in the upper ("water") compartment of the diffusion cell. At the end of each experiment, chlorine concentration profiles in the model sediments were recorded by sampling through septa ports in the diffusion cell.
To calculate diffusion coefficients, a numerical model was developed. Nonuniform discretization was employed to solve the problem more efficiently, while accurately representing the steepest concentration gradients present in the system. The higher the nonuniformity of the grid, the less accurate these approximations become: first derivative and second derivative approximations become less than second order and less than first order accurate, respectively. To represent the stepwise change in the diffusion coefficient at the sediment-cap and cap-water interfaces, a central difference scheme incorporating two diffusion coefficients was derived and used to solve the diffusion equation. From the value of flux at the cap-water interface calculated using Fick's first law,
the accumulated concentration was computed for the upper (“water”) compartment of the cell and used to determine the concentration at the cap-water interface. Evaporation and dilution resulting from periodic sampling also were accounted for in the model. Diffusion coefficients were calculated by minimizing the residual in the least-squares problem.
The objectives of this component are as follows:
1. We will provide a complete, operational, three-dimensional (3-D) hydrodynamic model that can determine risk of sediment caps being scoured in a hurricane or other rare event. The density field driving the baroclinic force is determined from the salinity, temperature, and suspended sediment.
2. We will provide an improvement to the standard resuspension relationships to incorporate the effects of gas bubbles emanating within the sediments through laboratory studies.
3. We will employ the 3-D model to extend the Thibodeaux-Valsaraj "hockey stick" model of desorption in a real environmental setting. This would be in combination with an extreme event that would resuspend contaminated sediments that contain contaminants that exhibit a "hockey stick" type of desorption.
4. We will modify basic resuspension concepts as defined below to incorporate the buoyancy changes as gases are generated in the sediments.
Contaminant fate and transport modeling studies were conducted at LSU. This part of the study focused on the modeling of contaminant transport from the sediment through a layer of a composite reactive cap to the bulk water phase. Both pre- and postconsolidation stages were considered in the model. The reactive cap constitutes degradation agents such as microbial populations or other inorganic oxidation agents. The model considers the extent of oxygen diffusion into the cap layer from water column and the consequent aerobic degradation rates in the reactive cap layer. The modeling efforts are being validated using laboratory-scale experimental data that are being obtained in this study. Issues relevant to any cap design are being obtained in this study.
Objectives 1 and 2. We have succeeded in creating a range of sediment cap structures that can be reproduced in the field following a simple procedure and low-cost materials. By changing the procedure or "recipe" caps with different permeabilities, diffusional transport characteristics and mechanical strengths can be created. We are quantifying these characteristics by measuring diffusivity and mechanical strength, as well as by performing quantitative image analysis on the cap structures.
From a diffusion test with a 1-mm glass bead, molecular diffusivity of chlorine was determined to be 1.43·10-5 cm2/s, which compares well with values computed from empirical correlations. This value was used in the model to compute apparent diffusion coefficients for all other cap materials studied in this work. For example, the sand cap value resulted in a remarkable match between porosity measured experimentally and porosity computed as a result of the least square fit of model concentration profile to the measured one. Apparent diffusion coefficients for different composite caps were determined. Missing values correspond to measurements in progress. Note that these results show that caps with significantly different permeabilities and diffusive transport characteristics can be formed via this procedure.
Mechanical stability is an important characteristic of the sediment cap. To evaluate the dynamics of strength development in composite caps, setting times were determined using a standard method that involves Vicat apparatus. Initial setting time and final (when applicable) setting time were determined. Setting time may be used as a characteristic time in the expression for the time-dependent porosity of such composites. The setting time depended both on the cement fraction and liquid/solid ratio. As anticipated, strength developed quicker in caps with higher cement content and lower liquid/solid ratio. It was observed that for the final setting to occur, cement fractions higher than 33 percent were necessary.
Various techniques have been used to place the cap on top of the contaminated sediment. All of them, however, involve settling of the concentrated slurry of capping materials through the water column. To measure flowability of cement-bentonite composites, the standard test (American Society for Testing and Materials C 230-90) was performed for various liquid/solid ratios and cement contents. Flowability increased with an increase in liquid/solid ratio, as expected. The dependence of flowability on the cement fraction, however, was not linear, which is attributable to concentration-dependent interactions between clay and cement particles.
Hydraulic permeability of the cap will be an important factor in allowing caps to remain in place in the presence of groundwater seepage. These measurements are in progress. To correlate observed macroscopic transport characteristics to the cap microstructure, information on the cap porosity has to be obtained. Also, reactivity of the cap in terms of metal adsorption will be determined largely by the composition of the cap at the microscale.
Morphology and elemental content of cap components were characterized using scanning electron microscopy (SEM), energy dispersive x-ray (EDX), and particle sizing. Bentonite particle and particle aggregates were smaller than 10 microns with the following elemental composition: Na2O 2.75 percent, MgO 3.48 percent, Al2O3 23.82 percent, SiO2 68.81 percent, CaO 0.2 percent, and Fe2O3 0.94 percent. Cement particles were smaller than about 60 microns and, in terms of elemental composition, were of two types. Type I contained Al2O3 5.61 percent, SiO2 46.63 percent, CaO 43.49 percent, Fe2O3 0.72 percent, and SO3 3.76 percent. Type II (sulfur rich) contained Al2O3 3.71 percent, SiO2 6.78 percent, CaO 34.79 percent, Fe2O3 0.31 percent, and SO2 54.41 percent. Thus, based on the EDX spectra of individual grains, it is possible to discriminate different cap components. In addition, the re was a morphological difference between bentonite and cement particles, the former having more flaky appearance.
To investigate the effects of hydration on the morphology of cement grains, SEM images of cement samples hydrated at three different water/cement ratios (L/S = 0.3, 0.4, 0.5) were recorded for different stages of hydration. A significant decrease in porosity was observed after 14 days of hydration, more so samples with lower L/S ratio. Both the type and amount of various hydration products, such as ettringite and syngenite, varied as functions of time and L/S fraction.
Objective 3. This study was performed using the 3-D advanced circulation (ADCIRC) hydrodynamic model. The mathematical model of hydrodynamic circulation in coastal water is based on the 3-D Navier-Stokes equations, as summarized by Pandoe and Edge (2003) and in Luettich and Westerink (2002).
The newly developed of extended 3D-ADCIRC include the transport of sediment, salinity, and temperature, which is summarized below:
For suspended sediment transport with erosion and deposition terms, the surface and bottom boundary conditions of equation (3) are:
where C is a sediment concentration (g/L or kg/m3), salinity (psu), or temperature (°C); u and v are depth-dependent horizontal velocity, Dh and Dv are horizontal and vertical dispersion coefficients for sediment (m2 s-1); is a vertical velocity in -coordinate [m/s]; s is settling velocity of sediment in -coordinate (m/s); E is an erosion flux (kg/m2/s); and D is a deposition flux (kg/m2/s).
Cohesive Sediment (Clay [1µm] – Silts [50µm])
Whitehouse, et al. (2000) assumed that the flocs of cohesive sediment could be treated as low-density grains, where aggregation of flocs, break-up of flocs, and water-flow within flocs are neglected. Thus, the formula of settling velocity ws of any volume concentration C of cohesive sediment is given as:
|is a dimensionless floc diameter||(6)|
and de is the effective diameter of a floc that increases with the volume concentration C of the suspension, and is formulated as:
where: m is a coefficient; m =1.06; l is a length scale given by:
where: s is mineral density of the grains; k = multiplicative coefficient with k = 0.00043; e is effective density of the floc; Cin is the internal volume concentration of grains inside a floc with Cin = 0.032 as a default value; Cf is the volume concentration of flocs in water (nondimension); C is the volume of grains in the suspension (nondimension); and Ck is the mass concentration of the suspension (mass/volume).
Depositional flux (Whitehouse, et al., 2000; Partheniades, 1990) is:
|P = probability of deposition =||(13)|
ws = sediment settling velocity (m/s) given in equation (3) and cd is a critical bed shear stress and is estimated from laboratory tests to be in between 0.06 and 0.10 N/m2.
The analysis to estimate the erosion rate as a function of shear stress is given in Partheniades (1992) as:
where Eo and are experimental coefficients. Eo is between 0.38 x 10-4 and 1.53 x 10-4 g/cm2/minute, and a is between 5.6 and 12.1. ce is the critical bed shear stress for erosion given in Whitehouse (2000) as:
with typical ce around 0.1-0.2 N/m2 (not exceed 1.0 N/m2), with b as the bulk density of the bed over the density range 1,000 kg/m3 < b < 2000 kg/m3.
The depositional flux for noncohesive sediment is:
where: D2 is a depositional flux of noncohesive sediment (kg/m2s); C2 is a concentration of suspended noncohesive sediment; ws,2 is sediment settling velocity; and P2 is a probability of deposition for the noncohesive sediments.
Settling velocity of natural sand particle is evaluated using the formulation given by Cheng (1997) as follows:
where d* is a dimensionless particle parameter defined as:
Above the threshold of motion, sand/sediment in the bed is lifted off into suspension, where it is carried by the current. The bottom friction governs the entrainment of sediment from the bed. In a sediment suspension, the settling of sediments towards the bed is counterbalanced by diffusion of sand upward near the bed. Rousby (1997) assumed if the eddy diffusivity for sediments Dvc varies parabolically with height in the lower half and is constant with height in the upper half of water column, then the corresponding concentration profile, defined by van Rijn's profile, is:
|for za < z < h/2||(18)|
|for h/2 < z < h|
where b' is a modified form of Rouse number or suspension parameter; z is a height above sea bed (m); za is a reference height near sea bed (m); Cb(z) is a sediment concentration at height z because of bed load (mass/volume); and Ca is a sediment reference concentration at height za (mass/volume)
The solution strategy of the computed variables is determined from the nonconservative form of the momentum equation, which involves the finite element method for spatial and finite difference for temporal.
Objective 4. The coupled differential equations for transport were solved to obtain the concentrations in the sediment and cap with depth for various scenarios. The differential equations for this model are derived from Choy and Reible (1997). The solutions to the equations were accomplished using FEMLAB modeling tools, and the graduate student, Mr. Andre Marquette, developed a user-friendly version over the last year. This model is now available free of charge to end users on the Hazardous Substance Research Center (HSRC) Web Site at: http://capping.hsrc.lsu.edu Exit . The above model is useful for both preliminary field design of a cap as well as laboratory model calibrations. This model considers the pre- and postconsolidation scenarios within the cap. The simultaneous transport of oxygen into the cap and reactivity of pollutants both within the cap and the underlying sediment were incorporated in the model. The input to this model comprised of estimated physical parameters such as cap and sediment porosity, bulk density, cap thickness, fractional organic carbon, sediment concentration, effective diffusivity, retardation factor, initial porewater concentration, and cap-water interfacial mass transfer coefficient. Modeling tasks in this area have been completed, and only experimental efforts to calibrate the model remain.
Experimental microcosms were set up to evaluate the simultaneous transport of contaminants from the sediment through a cap and the transport of oxygen from overlying water into the cap. Native sediments (University Lake, Baton Rouge, LA; Anacostia River, Washington, DC) contaminated with polycyclic aromatic hydrocarbons (PAHs) were used to retain the microbial activity within sediments and cap materials. The cap material used in the initial evaluations was sand. LSU diffusion cells that we have used previously for several capping experiments were modified to obtain both redox and oxygen concentration profiles in porewater along with contaminant concentrations in the efflux water. Replicate cells were prepared. At the end of the experiment, the sediment and cap were cored into thin sections to obtain concentration profiles and penetration of contaminant and oxygen into the cap. These profiles, under controlled laboratory conditions, are to be compared with model predictions. The experimental work is currently the research topic for another graduate student, Mr. Yuan Qingzhong.
Summary of Results to Date
Objectives 1 and 2
• Designed sedimentation/diffusion column and modified existing diffusion cell to study transport properties of in situ deposited sediment caps.
• Developed numerical model of tracer diffusion in the above geometries.
• Conducted nonreactive tracer tests to determine permeability of sediment caps of different compositions (clay loam, sand, clay/cement composite).
• Determined diffusion coefficients of for tracer and dichlorophenol in sediment caps of different compositions (clay loam, sand, clay/cement composite).
• Quantified mechanical properties of the caps.
• Demonstrated applicability of environmental SEM for the study of in situ hydration of cements and cement-clay composites. Presence of clay minerals was found to significantly alter the processes of cement hydration and strength development.
• Innovated the methodology for the resin impregnation of 100 percent water-saturated sediment caps for the subsequent thin-sectioning and morphological study.
• Developed parallel (message passing interface) implementation of lattice gas automata diffusion model and demonstrated the model's almost perfect scalability.
• Up to last year, we developed a 3-D flow of salinity, temperature, and density in the stratified fluid. The inclusion of baroclinic term governs important flow pattern, where the layer stratifications are well defined. Up to this stage, the model development has developed successfully the 3-D transport of sediments, tracers, salinity, and temperature. The inclusion of erosion and deposition for sediment transport both for cohesive and noncohesive sediments is the most advanced result. The conservation of mass among eroded, suspended, and deposited sediments has been investigated carefully with positive results.
• Developed numerical solutions to pre- and postconsolidation scenarios for the cap using FEMLAB. Incorporated different rate constants for the cap and the sediment. The model was made available on the HSRC Web Site.
• Developed numerical algorithm for coupled solution of the oxygen transport and contaminant transport equations.
• Experiments were set up in microcosms in the laboratory to calibrate the model with experimental data.
Adams DD, et al. In: Baudo R, Geisy J, Muntau H, eds. Sediments: Chemistry and Toxicity of In-Place Pollutants. Boca Raton, FL: Lewis Publishers, 1997, pp. 215-236.
Al-Raoush RI, Thompson K, Willson CS. Soil Science Society of American Journal (in press, 2003).
Al-Raoush RI, Willson CS. Journal of Hydrology (accepted, 2003).
Chatzis I, et al. Society of Petroleum Engineering Journal 1983;23:311.
Conner JR. Chemical Fixation and Solidification of Hazardous Wastes. New York, NY: Van Nostrand Reinhold, 1989, p. 784.
Dalton SK. Transport and lipid solubility of hydrophobic organic compounds using semipermeable membranes: influence of dissolved organic matter and solution chemistry. M.S. Thesis. Department of Environmental Science and Engineering, Rice University, Houston, TX, 2002.
Fendinger NJ, Glotfelty DJ. Environmental Science and Technology 1988;22(11):1289-1293.
Hull JH, Jersak JM, Kasper CA. In situ capping of contaminated sediments: comparing the relative effectiveness of sand versus bentonite-mineral based sediment caps. In: Proceedings of the 1999 Conference on Hazardous Waste Research, 1999, pp. 286-311.
Hydroqual Inc. Development and application of a modeling framework to evaluate hurricane impacts on surficial mercury concentrations in Lavaca Bay. Project Report, Aluminium Company of America, Point Comfort, TX, 1998.
Luettich RL, Westerink JJ. Formulation and numerical implementation of the 3D ADCIRC finite element model version 36.01, NC, 2002.
Illangasekare TH, et al. Journal of Contaminant Hydrology 1995;20(1-2):1-25.
Johnson BD, et al. Marine Geology 2002;187(3-4):347-363.
Kochetkov A, et al. Environmental Toxicology and Chemistry 2001;20:2184-2188.
National Research Council. Adaptive site management. National Academy Press, 2003.
Neale CN, Hughes JB, Ward CH. Ground Water 2000;38(5):784-794.
NeuBauer CM, Jennings HM. Scanning 1996;18:515-521.
Pandoe WW, Edge BL. Three-dimensional hydrodynamic model, study cases for quarter annular and idealized ship channel problems. Journal of Ocean Engineering 2003;30:1117-1135.
Partheniades E. Estuarine sediment dynamics and shoaling processes. In: Herbich JB, ed. Handbook of Coastal and Ocean Engineering. Gulf Publishing Co., 1992.
Poindexter-Rollins. Technical Report D-90-2, US Army Engineer Waterways Expt. St., 1990.
Raja S, et al. Journal of Chemical and Engineering Data 2002;47:1213-1219.
Reible D, Constant WD. Comparative validation of innovative capping technologies. Active capping, AWTA update, 2002.
Robertson R, Padman L, Levine MD. A correction of the baroclinic pressure gradient term in the Princeton Ocean Model. Journal of Atmospheric and Oceanic Technology 2001;18(6):1068-1075.
Smith JS, Valsaraj KT. Industrial and Engineering Chemistry Research 1997;36:903-914.
Sojitra I, et al. Colloids and Surfaces 1996;110:141-157.
Soulsby R. Dynamics of Marine Sands. United Kingdom: Thomas Telford Services Ltd., 1997.
VanRijn LR. Sand transport by currents and waves: general approximation formulae. In: Proceedings of the Conference on Coastal Sediments 2003, Clearwater Beach, FL, May 18-23, 2003.
Whitehouse R, Soulsby R, Roberts W, Mitchener H. In: Wallingford HR, ed. Dynamics of Estuarine Muds. United Kingdom: 2000.
Wang XQ, Thibodeaux LJ, Valsaraj KT, Reible DD. Environmental Science and Technology 1991;25:1578-1584.
The proposed research project will respond to each of the key areas of uncertainty associated with facilitated transport processes in sediment caps. The work proposed for the following year continues to address the two key elements of this project, engineering cap reactivity and integrity, through the following five tasks:
• Quantify integrity as a function of structure through an evaluation of cap hydraulic permeability, gas permeability, strength, and diffusivity as a function of structure (Lead Principal Investigator [PI]: Weisner).
• Advance the episodic event modeling work on three fronts to make it more readily applicable to real field conditions and to the realistic determination of fate of contaminates in episodic or hurricane-type conditions (Lead PI: Edge).
• Identify conditions that lead to mobilization or entrapment of NAPL by a cap (PIs: Willson and Valsaraj).
• Identify the rate of gas generation resulting from microbial processes and the subsequent effects on cap integrity and contaminant migration (PIs: Hughes and Valsaraj).
• Evaluate the effectiveness of a conventional cap as a permeable reactive barrier for treatment of groundwater contaminated by chlorinated solvents (Lead PI: Hughes).
In the face of the uncertain influence of facilitated transport on capping effectiveness and lack of confidence in fate processes accelerated by the interactions of physicochemical and biological factors in the high-gradient conditions in the cap, many managers choose to avoid capping for removal approaches that are potentially more expensive, less effective, and may exhibit greater environmental impacts. The primary objective of the proposed research is to develop the processes of understanding, and to develop quantitative criteria necessary to make informed decisions as to the applicability or nonapplicability of capping under these complicating conditions and, where possible, means of designing a cap to enhance these attenuation processes.
Capillary effects in the pore structure of the sediment control the mobility of a NAPL at the sediment-water interface. This retarding force is offset by the destabilizing influences of groundwater seepage flow because of either regional groundwater gradients or tidal fluctuations, physical disruption associated with cap placement and consolidation, and gas generation and migration. Capping also introduces unique influences on NAPL stability. The contrast in physical properties of the sediment and capping layer provides an opportunity for the development of either a capillary barrier or a preferential migration zone similar to that described by Illangasekare, et al. (1995). During cap placement, the force of the cap settling to the sediment surface will potentially destabilize the NAPL. Generally, cap placement by gravity settling is preferred to ensure uniformity of the completed cap. Subsequent consolidation of the underlying sediment after cap placement also provides a destabilizing factor for NAPL migration. The same processes describing dewatering and consolidation of sediments (Poindexter-Rollins, 1990) can, in principle, be applied to NAPL migration and subsequent sediment consolidation; however, this approach is unproven.
Another destabilizing factor in the DNAPL plume is the gas generated by organic degradation processes in the sediment. Mineralization of sedimentary organic matter by bacteria generates gases such as CH4, N2, CO2, and other trace gases. Denitrifying bacteria produce N2, methanogenesis results in CH4, and CO2 results from fermentation and sulfate-reduction processes in sediments. These gases tend to migrate out of sediments into overlying water and are vented to the atmosphere. The preferential pathway generated by the gas migration may provide a means for the migration of separate-phase material as well as contaminant to the sediment-water interface. This facilitated transport process is of concern at several other sites as well. The same forces govern bubble entrapment and mobilization in a porous media as control NAPL migration. Buoyancy-driven migration of the gas opens channels through a cap, or, if contained by an impermeable layer, may accumulate potentially, causing greater damage when ultimately released. Gas bubbles are inherently hydrophobic and tend to accumulate both hydrophobic organic contaminants and colloids from porewater. Their migration can have significant impact on the transport of contaminants through the cap. A key parameter in describing contaminant transport by bubbles is the bubble-water partition coefficient, Kbw, which tends to significantly exceed the conventional air-water Henry's constant because of the accumulation of hydrophobics at the interface. This interface is highly conducive to the adsorption and uptake of dissolved hydrophobic contaminants (for example, PAHs) (Raja, et al., 2002; Smith and Valsaraj, 1997; Sojitra, et al., 1996).
As indicated previously, capping is not normally considered to be an aid in controlling a contaminated groundwater plume that is entering a water body. A cap, however, provides a means to control oxygen conditions within the groundwater plume. Thus, the application of a sediment cap can provide a relatively simple means of engineering a reactive permeable barrier, as has been developed for subsurface treatment of groundwater (National Research Council, 2003). Specifically, a cap can behave like a two-step reactive treatment barrier in that anaerobic conditions are normally maintained at depth in a cap, but aerobic conditions can be maintained near the surface by either diffusion or bioturbation from the water body. The cap potentially can provide the residence time necessary to achieve the degradation of halogenated compounds, such as chlorinated solvents, that are common groundwater contaminants.
A combination of physical and mathematical modeling exercises are proposed to assess and evaluate each of the facilitated transport processes in sediment caps. These laboratory experiments and theoretical tools will provide the foundation for understanding and correlation of the importance of these processes as a function of physical, chemical, and biological conditions in the sediments. The overall procedures to assess each of the facilitated transport processes are discussed separately below. Although field testing and evaluation of these concepts is not explicitly supported by this research project, it is expected that the results will be incorporated into ongoing projects at field sites including the Anacostia River, Calcasieu Estuary, and Thea Foss Waterway. In addition, sediments collected at these sites will be employed in the laboratory experimentation to assess the importance and mechanisms of facilitated transport processes.
Continuing Task 1
To Further Quantify Integrity as a Function of Structure Through an Evaluation of Cap Hydraulic Permeability, Gas Permeability, Strength, and Diffusivity as a Function of Structure (Weisner). Methodology for these experiments will be identical to that used to evaluate cap integrity and transport across caps using diffusion cells. We also will examine cap reactivity with three possible scenarios: heavy metal reactivity with the cement/clay composite, reactivity of organic contaminants with additional adsorbent materials to cap formulation, and decomposition of organic contaminants across caps formulated to include placement of zero-valent iron.
Continuing Task 2
Continue the Task of Advancing the Episodic Event Modeling Work on Three Fronts To Make It More Readily Applicable to Real Field Conditions and to Realistic Determination of Fate of Contaminates in Episodic or Hurricane-Type Conditions (Edge). Three major tasks are proposed for the coming year.
• The first effort will be to document the 3-D baroclinic sediment transport code so it can be used as an open-source code. It will be fitted to the parallel version of ADCIRC so that it will perform expediently on a supercomputer, which will be needed for real project assessments. We also will incorporate several sediment resuspension mechanisms for selection to specific sites. This will be developed with an example application to Matagorda Bay in Texas, where a large area of mercury contamination exists.
• The model will be further augmented to include desorption kinetics as described by Thibodeaux, Valsaraj, and Hughes. This "hockey stick" or two-parameter desorption model provides a realistic method to look at the fate of materials that are resuspended, either from naturally capped materials, or the "designer" reactive caps during storm events. The two-parameter model is described as:
where F is the mass of desorbed compound from the suspended sediments. The first term represents the material that is rapidly desorbed, and the second is that which is desorbed at a much lower rate. The material that quickly desorbs will be transported beyond where the actual particles may land because of deposition unless it is reabsorbed. On the other hand, the material that is more slowly desorbed may be deposited back to the bottom through deposition before it is released. This has implications not only to determine risk and fate of materials in an episodic event, but in dredging operations as well. We will work closely with LSU in developing the conceptual framework to incorporate into the fate model developed in the first 2 years.
• Gas in sediments is a very real phenomenon that exists in contaminated and noncontaminated sediments that are rich in organics. All scour experiments that have been conducted to determine erosion rates or resuspension rates have not included gases being released. We plan to conduct tests in the 3 m deep by 4 m wide by 50 m long flume in the Coastal Engineering Laboratory to determine the effects of gas on scour rates of various cap materials. The flume also has a 1.6 m deep sediment pit that is 8 m long, which will be used for the placement of sediments. The tests will be conducted with currents ranging from 0.2 m/s to 1 m/s. Compressed air will be added through a manifold at rates suggested from the literature and as recommended by research from Objective 6. We will work closely with Dr. Wiesner to select "designer" cap materials for tests. The results will produce a new data set on the effect of gas emission from sediments on scour rates.
Proposed New Task 3
Identification of Conditions Leading to Mobilization or Entrapment of NAPL by a Cap (Willson). Selected sediments will be collected and subjected to several different experiments. (1) We will perform synchrotron x-ray tomographic imaging of the sediment pore structure. (2) We will perform column experiments examining displacement force requirements for the NAPL phase.
Synchrotron x-ray tomographic imaging has proven useful for quantitatively describing the pore structure and the distribution and characteristics of various fluid phases in unconsolidated porous media systems (Al-Raoush, et al., 2003; Al-Raoush and Willson, 2003). Dr. Willson’s group will utilize the synchrotron x-ray tomography stations located at the LSU Center for Advanced Microstructures and Devices and the GeoSoilEnviroCARS synchrotron x-ray tomography beamline at the Advanced Photon Source at Argonne National Laboratory.
The effect of this pore distribution and heterogeneity on NAPL migration will be evaluated in a series of column experiments that will seek to correlate displacement efficiency with capillary and Bond numbers (after Chatzis, et al., 1983). The effect of various cap layers on displacement efficiency also will be assessed. The experiments will be conducted by forcing a controlled flow through undisturbed sediment cores collected from NAPL-containing sediment sites. Additional experiments will be conducted by introducing a well-characterized NAPL phase (e.g., soltrol, a mineral oil) to a sediment column, and displacement efficiency will be assessed with and without various capping layers.
The laboratory experiments will be compared to predictions of NAPL migration from numerical models of sediment consolidation and NAPL displacement. Conventional sediment consolidation models that describe dewatering (Poindexter-Rollins, 1990) will be adapted to predict consolidation in the presence of a migrating NAPL phase. The ability of such models to accurately predict such consolidation will be tested. Similarly, models of NAPL migration will be compared to observed NAPL displacement in the column experiments. Several NAPL migration models are available to the investigators, including a state-of-the-art petroleum reservoir model, STARS (Computer Modeling Group), which can simulate both bulk phase movement and contaminant transport.
Proposed New Task 4
Identification of Gas Generation Rates in Sediments and Impacts on Contaminant Migration and Cap Integrity (Hughes and Valsaraj). A series of laboratory experiments also will be directed at investigating the effects of gas bubble ebullition. The primary parameters requiring definition are the rate of gas generation and the bubble-water partition coefficient. Thus, the four main tasks in the proposed work are to: (1) determine the rates of gas migration in typical capped sediments; (2) obtain experimental values of gas bubble-porewater partition constants for typical PAH compounds in the same sediments; (3) assess the contaminant transport within gas bubbles from a contaminated sediment column at typical gas flow rates observed in the field; and (4) investigate the structural changes and secondary porosity formed during gas bubble transport through sediments.
Task 1 will again be evaluated by measurement of gas ebullition rates in undisturbed cores collected from sediments. Gas generation rates will be compared to the amount and type of organic matter found in the sediments. The Anacostia River is expected to provide excellent samples of gas-generating sediments. The partition constant for a pollutant between porewater and methane or other gas bubbles can be obtained using a variety of methods in Task 2. Examples include a bubble column apparatus as was used previously by us (Smith and Valsaraj, 1997), a wetted wall column method as used by Fendinger and Glotfelty (1988), or a dynamic vapor loop stripping method recently described by us (Kochetkov, et al., 2001). The dynamic vapor loop stripping method is especially suited to compounds with low gas bubble-porewater partition constants. The methodology is described in a recent paper by the investigators in Kochetkov, et al. (2001). In Tasks 3 and 4, column experiments will be used to evaluate contaminant transport as a result of gas movement and the effect of the gas growth and movement on the structural integrity of the sediment and/or cap. The goal of the task is to ascertain the structural (physical) changes and secondary porosity that would be observed within a sediment column as a result of bubble growth and migration as done by Boudreau and coworkers (Johnson, et al., 2002). We propose to inject small gas bubbles at typical rates through a two-dimensional (2-D) glass column with the use of a gear motor, a piston, and cylinder arrangement modeled after Johnson, et al. (2002). The 2-D rectangular column will not only allow us to observe the physical and structural changes within the sediment, but also allow us to make quantitative measurements of sediment porosity changes.
Measurements of the pressure as a function of time will be made to obtain information on the bubble growth. To provide information on the shapes of bubbles formed, the structural changes and the formation of secondary porosity within the sediment, the entire column will be placed in an x-ray cabinet, and the sediments will be x-rayed before and after gas injection (see Task 2 above). To interpret bubble growth results, one also will need the determination of the Young's modulus for the sediment samples. As indicated by Boudreaux and coworkers (Johnson, et al., 2002), the theory of linear elastic fracture mechanics can be used to satisfactorily analyze the experimental results on the critical pressure for bubble growth. Methane or other gas bubbles will be introduced at the bottom of the contaminated sediment slurry, and the concentration of pollutants leaving with the gas stream at the top of the column will be ascertained. In addition to the testing on 2-D cores, cylindrical columns also will be injected with gas bubbles, and x-ray tomography will be used to nondestructively image these systems. Although the image resolution to particle size ratio will not be sufficient to directly measure the changes in pore body and throat sizes, variations in porosity within the system can be determined. In addition, the extreme difference in x-ray absorptivity will allow for the quantification of the bubble sizes and shapes. Because of the size of the cores (on the order of inches), imaging of these systems will most likely be performed at the Naval Research Laboratory in Stennis, MS using conventional x-ray tomography.
Proposed New Task 5
Evaluation of the Effectiveness of a Conventional Cap as a Permeable Reactive Barrier for Treatment of Groundwater Contaminated by Chlorinated Solvents (Hughes). Proposed studies will investigate the potential for the dual redox state to act as an effective barrier to chlorinated solvent migration, and to evaluate the relative participation of reductive versus oxidative transformation processes occurring. Through kinetic experiments of reductive dehalogenation followed by aerobic degradation of several chlorinated solvents of interest, the ranges of residence times necessary for this mechanism to be effective will be identified. The research also will identify conditions necessary for this mechanism to be operative and conditions that encourage or discourage the reactions. The experimental approach to be used involves hybrid column systems where sediments (identical to those used in other facets of this overall project) can be deposited, followed by a clean sand material simulating a cap. The sediment layer will be several inches thick followed by approximately 6 inches of sand. Groundwater infiltration will be initiated using a syringe pump in an upflow manner. This flow will contain (14C)trichloroethene (TCE) at concentrations of approximately 1 mg/L. A crossflow of clean water saturated with oxygen will be generated across the top of the clean sand using a peristaltic pump. Sample ports located along the length of the column will be used to monitor TCE, its reduction products (cisDCE, vinyl chloride, and ethene), and 14C-labeled oxidation products (including CO2). The Hughes laboratory is well equipped to conduct these analyses using a range of gas chromatography, high-performance liquid chromatography, and scintillation counting techniques. To probe reaction processes to a greater degree, the influent groundwater composition and flow rate will be altered periodically. For example, the addition of methane to the syringe will produce a presaturated flow; this can be used to simulate conditions with increased methanogenic activity and assess the potential for increasing methanotrophic cometabolism. By changing the flow rate, we will investigate the sensitivity of the anaerobic and aerobic processes to seepage velocity. Lastly, the concentration of dissolved oxygen in the crossflow will be increased/reduced to better understand the oxygen requirements of methanotrophic populations.
Journal Articles on this Report : 3 Displayed | Download in RIS Format
|Other subproject views:||All 20 publications||8 publications in selected types||All 8 journal articles|
|Other center views:||All 279 publications||92 publications in selected types||All 63 journal articles|
||Pandoe WW, Edge BL. Three-dimensional hydrodynamic model, study cases for quarter annular and idealized ship channel problems. Ocean Engineering 2003;30(9):1117-1135.||
||Tarabara VV, Wiesner MR. Effect of collision efficiency on the evolution of the surface of diffusion-limited deposits. Journal of Colloid and Interface Science 2001;237(1):150-151.||
||Tarabara VV, Pierrisnard F, Parron C, Bottero J-Y, Wiesner MR. Morphology of deposits formed from chemically heterogeneous suspensions: application to membrane filtration. Journal of Colloid and Interface Science 2002;256(2):367-377.||
Supplemental Keywords:enhanced natural recovery, reactive barrier, confined aquatic disposal, reaction in porous media, diffusion in porous media, gas and nonaqueous phase liquid migration in cap, waste, water, analytical chemistry, contaminated sediments, ecology and ecosystems, engineering, environmental engineering, hazardous, hazardous waste, remediation, Monte Carlo technique, computer modeling, computer models, contaminant dynamics, contaminant management, contaminant transport, contaminant transport model, contaminated sediment, contaminated soil, fate and transport, in situ remediation, kinetic models, kinetic studies, sediment caps., RFA, Scientific Discipline, Waste, Water, Contaminated Sediments, Remediation, Environmental Chemistry, Analytical Chemistry, Hazardous Waste, Ecology and Ecosystems, Engineering, Hazardous, Environmental Engineering, contaminant transport, fate and transport , in situ remediation, contaminant dynamics, sediment caps, computer models, contaminated sediment, kinetic studies, contaminated soil, computer modeling, contaminant transport model, treatment, kinetic models, contaminant management
Progress and Final Reports:Original Abstract
Main Center Abstract and Reports:R828773 HSRC (2001) - South and Southwest HSRC
Subprojects under this Center: (EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
R828773C001 Bioturbation and Bioavailability of Residual, Desorption-Resistant Contaminants
R828773C002 In-Situ Containment and Treatment of Contaminated Sediments: Engineering Cap Integrity and Reactivity
R828773C003 Phytoremediation in Wetlands and CDFs
R828773C004 Contaminant Release During Removal and Resuspension
R828773C005 HSRC Technology Transfer, Training, and Outreach