Remediation of fractured geologic media contaminated with chlorinated volatile organic compounds that diffuse into the matrix is challenging using isothermal methods. Due to the low permeability of the matrix material and uncertainty of fracture networks, it is difficult to flush the system with any type of fluid or deliver remediation agents into the matrix. However, thermal methods have some promises. When the matrix is heated above water boiling temperature, depressurization in the fractures may trigger water boiling in the matrix, which, as a result, generates a large volumetric steam flow toward adjacent fractures, stripping the adsorbed or dissolved volatile contaminants from the matrix. This process has not been demonstrated in the laboratory or field, and the parameters that control the contaminant mass removal from the fractured geologic media at the scale of a single fracture and field application were not well understood. The objective of this study is to understand the contaminant mass transfer due to boiling in fractured geologic media.
Local-scale matrix boiling and contaminant mass transfer during the process were demonstrated using a sandstone core, where the unfractured core represents the matrix and an end the fracture. The core was contaminated by pumping water dissolved with 1,2-DCA and NaBr. When boiling occurred in the matrix, a temperature gradient toward the fracture was observed, indicating, under saturated vapor condition, a pressure gradient pushing the steam (water and vapor mixture) towards the fracture. When boiling occurred, 1,2-DCA was removed at a rate 6 times higher than before heating. The nonvolatile bromide concentration, as an indicator of the steam quality, showed a reverse correlation with the volatile 1,2-DCA concentration, indicating that the majority of 1,2-DCA was removed by partitioning to the vapor phase.
To accurately predict the partitioning of contaminant between aqueous and gaseous phase at high temperature, Henry¡¯s law constants were measured for 12 chlorinated solvents (tetrachloroethylene, trichloroethylene, chloroform, 1,2-dichloroethane, 1,1,1-trichloroethane, 1,1-dichloroethane, dichloromethane, carbon tetrachloride, cis-1,2-dichloroethylene, chloromethane, chloroethane, and vinyl chloride) over temperatures from 8 to 93 ¢ªC. The measured results show that Henry¡¯s law constant is strongly dependent on temperature, increasing by factors from 3 (chloromethane) to 30 fold (1,2-DCA) as temperature increases from 8 to 93 ¢ªC. The temperature dependency of Henry¡¯s law constants was modeled with the Van¡¯t Hoff equation. Better fitting is obtained by assuming the enthalpy of dissolution is a function of temperature, suggesting the inappropriateness of extrapolating the Henry¡¯s law constant from measurements at low temperature using a linear function. Using measured data for solubility, a vapor pressure-solubility model gives a reasonable prediction of the Henry¡¯s law constants. With improved data on Henry¡¯s law constants at high temperatures for the 12 common CVOCs measured in this study, it will be possible to more accurately model subsurface remediation processes that operate near the boiling point of water.
The laboratory experiment was simulated with a 2-D multiphase numerical model using the TMVOC code. The simulated results are in reasonable agreement with the experiment, provideing confidence for the TMVOC code in predicting the CVOC removal from fractured geologic media by boiling. A 1-D numerical model was further used to demonstrate the contaminant mass transfer from a core without end effects. Similar to the experiment results, a temperature gradient was observed while pore water in the matrix was boiled. After opening to the atmosphere, a high proportion of steam vapor was produced, corresponding to a high removal rate of 1,2-DCA. All of the 1,2-DCA mass was removed after boiling out around 50 mL pore water. The contaminant removal from fractured geologic media at field scale was simulated using the MINC method. After about 35 days of treatment, 27.8% of the pore water (including both steam vapor and liquid water) was extracted, and essentially all the 1,2-DCA mass (more than 99%) was removed from the fractured site. The simulation shows boiling is an effective mechanism for CVOC removal from fractured geologic media.
Effects of chemicals, fracture spacing, fracture aperture, diameter of heating pattern, matrix permeability, extent of vacuum, and different operational strategies on the performance of thermal treatment of a fractured site were investigated. The simulation results show that, under the same boiling conditions, the contaminant removal rates vary for different chemicals, depending on their Henry¡¯s law constants. The higher the Henry¡¯s law constant for a chemical, the higher the contaminant removal rate. Variations of fracture properties (aperture and spacing), size of heating pattern, and the extent of extraction vacuum have larger effects on the system temperature than the matrix permeability due to the different percentages of heat extracted from the system. The contaminant removal is more sensitive to the matrix permeability. A higher percentages of contaminant are removed in the case of higher matrix permeability, where boiling occurs in both the matrix and fracture, while less is removed in the lower permeability case (e.g. 1¡¿10-17 m2), where boiling occurs primarily in the fracture. Further simulation based on a 3-D model that includes the effects of cold water shows that the cooling effect caused by the influx of cold ground water is significant. When the fracture network is permeable, the extraction wells pull a large amount of cold water into the treatment zone, which holds the majority of the treatment zone below the water boiling temperature, minimizing the performance of thermal remediation. Compared to the effect of water flow pulled by extraction, ground water flow caused by the regional hydraulic gradient is insignificant.