Date of Award
Doctor of Philosophy (PhD)
Environmental Engineering and Earth Science
Murdoch, Lawrence C
Falta , Ronald W
Freedman , David L
Moysey , Stephen M
Chlorinated volatile organic compounds (CVOCs) in fractured, low permeability material can be a long-term source of groundwater contaminants and they are among the most difficult remediation challenges. Thermal methods are a possible remediation option, where heat is used to boil water and the volatile compounds are stripped out by partitioning into the resulting water vapor. Theoretical analysis indicates that only a small fraction of pore water removal may lead to a significant geometric reduction in the dissolved concentration [(Udell(, 1998]. Previous modeling studies [(Falta and Murdoch(, 2011; (Pruess(, 1983] suggest that a lower permeability limit exists for effective mass transfer during boiling. The simulations [(Pruess(, 1983] predict that in low permeability materials (k < 10-17 m2) boiling is restricted to the fractures, which would not contribute to contaminant removal from the matrix. However, limited experimental studies are available describing mass transfer under boiling conditions, so the theoretical results have not been validated with data. This is important because mass transfer under boiling conditions involves several strongly coupled processes, which may be incompletely represented in the theoretical analyses.
A recent experiment heated a contaminated core of Berea sandstone with permeability of 10-14 m2 to boiling temperature and then depressurized the top end of the core and collected the condensate [(Chen et al.(, 2010]. The CVOC contaminant was completely recovered after ~35% of the pore water was removed. These results demonstrate the potential for thermal methods to remediate materials of moderate permeability, but they were limited to sandstone with of (k( of~10-14 m2.
The objective of this research is to characterize contaminant mass transfer in low permeability clays at boiling temperatures. The research is motivated by the need to advance understanding of the thermal remediation of tight sediments containing CVOCs.
The approach is to extend the experimental technique developed for sandstone cores in (Chen et al.( (2010) so it can be used with remolded clay. Preliminary tests indicated that excessively long times would be required to contaminate cores of low permeability rock (mudstone), so tight rock was abandoned as impractical. Clay was adopted because it could be hydrated with contaminated water and packed in a mold, thereby creating a contaminated, low permeability material in a cylindrical form.
The experimental design considers a cylinder normal to the plane of a fracture. The upper end of the cylinder represents the wall of the fracture, so transport along the axis of the clay cylinder is used to characterize transport from the matrix to a fracture--the process assumed to limit the rate of remediation [(Chen et al.(, 2010]. 1,2-dichloroethane (DCA) was used as the CVOC because it is a common and persistent groundwater contaminant [(Henderson(, 2008]. It has a relatively large value of water solubility and a small Henry's constant, so it is a conservative choice for stripping into water vapor--other common contaminants, such as trichloroethene (TCE), will be more readily stripped than DCA. DCA was dissolved in degassed water containing bromide and this solution was mixed with dry kaolin powder to create the clay used in the tests.
A series of tests were conducted where different amounts of pore water removed from clay samples in a rigid-wall cell. Clay was heated to the boiling temperature and a valve at the upper end cap was opened, dropping the pressure and allowing fluid to flow to the condenser. During tests where the temperature was ~130 °C when the valve was opened, the flow rate from the condenser increased rapidly to 200-300 ml/h for several tens of minutes, and then it decreased to ~20 ml/h for the next 6 to 10 h. DCA concentrations were 4 to 12 times greater in the condensate than in the pore water concentration during the high rate of outflow and bromide concentrations were nearly zero. Essentially all of the DCA, but less than 5% of the non-volatile bromide was removed in the first 30% of the recovered pore water. Removal of DCA and water occurred uniformly along the clay length.
A suite of tests were terminated after different fractions of pore fluid were recovered from the flexible-wall cells heated with a low power density and a maximum temperature of 110 °C. Profiles of water content and DCA concentration show a fairly uniform, gradual decrease along the length of clay specimens, even after a relatively small amount of water was recovered. There was a gradient in water content and DCA concentration, with decreasing water content in the direction of flow, but the slopes are gentle (roughly 0.01/cm of water content and one order of magnitude of DCA concentration over 25 cm) and fairly uniform over the length of the core for tests where the fraction of water removed was less than 40%. A larger DCA concentration gradient was observed in tests where the bottom end had temperature less than 100 °C. More than two orders of magnitudes reduction of DCA in clay was observed uniformly along the clay length when more than 50% of the pore water was removed. A zone of larger reduction in water content developed at the outflow end, but only in tests where more than 40% of the pore water was removed. This result was surprising because it was expected that both water and DCA would be removed as a front that progressed with time away from the outflow, but instead the removal occurred over the length of the cylinder.
Theoretical analyses using a multiphase flow and transport code indicate that recovery should be controlled by boiling that first occurs at the outflow end of the core. Boiling and DCA removal should progress as a front that moves into the core, away from the outflow end. The analyses also predict large pressures in clay, which suppresses boiling and substantial recovery of DCA. These predictions contradict observations during the experiments. To address these contradictions, I propose a conceptual model that involves the growth of fractures during heating, which increase the rate at which fluids are recovered. A water P-T diagram is used to show how the conceptual model can explain the heating path in different experimental conditions. This explains different behaviors of low k clay in the rigid-wall and flexible-wall cells and how the fracturing happens in the P-T space, as well as the non-fracturing behavior in moderate k sandstone material under the same testing conditions.
The major findings from this dissertation are:
(i) DCA is nearly completely removed (more than 2 orders of magnitude reduction in concentration) from clay after approximately 30% to 50% of the pore water is recovered during boiling.
(ii) Increasing temperature before depressurization can speed up the removal rate of DCA.
(iii) An outflow rich in DCA (up to 12 times that of the pore concentration) and low in bromide indicates boiling in the matrix dominates the mass removal. Measured pressure and temperatures in the clay matrix are consistent with boiling conditions.
(iv) Water and DCA were removed over the entire length of the sample instead of concentrated as a front as indicated by the theoretical analyses. A drying front does appear after approximately 40% of the water has been removed.
(v) The observations indicate that the permeability of the clay significantly increased early in the heating process. This appears to be a result of fracture growth that occurs during heating and depressurization.
(vi) A fracturing mechanism can explain the observations that appear to be anomalous in light of the conventional processes of flow and transport through non-deforming porous media. An analysis is outlined that explains pressures and temperatures associated with different experimental conditions. It can also explain why sandstone has no fracturing upon heating to the boiling temperature.
(vii) This research confirms the ability of thermal remediation to remove volatile contaminants from the matrix in fractured clays. The recognition of a new mechanism for recovery by vapor-driven fractures suggests that remediation of low permeability may be more feasible than previously recognized.
Chen, F., X. Liu, R. W. Falta, and L. C. Murdoch (2010), Experimental demonstration of contaminant removal from fractured rock by boiling, (Environ. Sci. Technol.(, 44(16), 6437-6442, doi: 10.1021/es1015923.
Falta, R. W. and L. C. Murdoch (2011), Contaminant Mass Transfer during Boiling in Fractured Geologic Media, SERDP Report ER-1553 (accessed online on 4/1/2012).
Henderson, J. K. (2008), Anaerobic biodegradation of ethylene dibromide and 1,2-dichloroethane in the presence of fuel hydrocarbons. Thesis (PhD), Clemson University, 2008.
Pruess, K. (1983), Heat transfer in fractured geothermal reservoirs with boiling, (Water Resour. Res.(, 19(1), 201-208.
Udell, K. S. (1998), Application of in situ thermal remediation technologies for DNAPL removal, Proceedings of the GQ'98 Conference, Tubingen, Germany, September 1998.
Liu, Xiaoling, "LABORATORY CHARACTERIZATION AND THEORETICAL ANALYSIS OF CONTAMINANT MASS TRANSFER" (2012). All Dissertations. 1004.