You are here:
CHLORINATED SOLVENT IMPACT AND REMEDIATION STRATEGIES FOR THE DRY CLEANING INDUSTRY
Impact/Purpose:
Groundwater contaminated with chlorinated solvents from dry cleaner sources has been the focus of much research in recent years. The most common dry cleaning solvent is perchloroethene (PCE), which is a suspected carcinogen. Studies in Texas and California have shown that due to inadequate separation equipment and handling practices, PCE may be released directly into the ground via migration through degraded or cracked PVC or ABS sewer lines. The pump-and-treat remediation method often used at dry cleaner sites is only capable of removing the mobile, dissolved plume, and has been ineffective in many cases.
Certain microbes can metabolize PCE by using it as an electron acceptor. Lab studies have confirmed that in an anaerobic setting with a suitable substrate, PCE will be dechlorinated to trichloroethene (TCE), then to 1,2-dichloroethene (DCE), then to vinyl chloride, and finally to ethene, a non hazardous end product. Electron donors, such as hydrogen, are required to maintain the reaction. Providing hydrogen or a compound that releases hydrogen may be the key to driving the reaction to completion.
Two methods of hydrogen delivery are currently under study. One method involves the injection of a hydrogen releasing compound (HRC) into the aquifer, while the other method directly injects hydrogen gas. The main objective of this research is to study the effectiveness of these hydrogen delivery systems.
Description:
Additional funding has been received from the Advanced Technologies Program of the State of Texas by collaborators on this project, Dr. Herb Ward and Dr. Joseph Hughes. This follow-up study will provide a further in depth analysis of the benefits of bioremediation for the treatment of chlorinated solvents beyond the laboratory-scale, in the Experimental Controlled Release System tanks.
The earlier work that was performed in these pilot-scale tanks has provided us the insight and the background that has led to additional near-field scale studies funded by both GCHSRC and the American Petroleum Institute. This new work is investigating the environmental implications associated with the use of ethanol additive fuels as oxygenates in gasoline.
The findings from our research show that enhanced bioremediation and bioaugmentation may be viable options for decreasing source longevity and potentially overall exposure to chlorinated solvents with shorter plume lengths. This treatment option is also low maintenance and requires minimal equipment when a long-term electron donor such as Hydrogen Release Compound is used in sufficient quantities. Furthermore, our study reassures the public that huge quantities of methane will not be produced nor will there be the need to have the direct injection of hydrogen gas through these techniques and thus preventing potential explosive hazards.
Tracer Test: The main objective of this research was to study the effectiveness of a hydrogen delivery system that will provide an electron donor for effective stimulation of biological reductive dechlorination of chlorinated solvents. A hydraulic characterization of the system was first step towards reaching this goal and was essential to making informed decisions for later injections of PCE and microbes. Therefore, a tracer test was performed to evaluate the variation in hydraulic conductivity in the Experimental Controlled Release System (ECRS) tank located at Rice University. The test began with the injection of a six hour pulse input of 1000 mg/L bromide solution into the inlet of the ECRS tank. Approximately 400 samples of 50 to 100 mL were taken from 12 sample lines down the 18 ft length of the tank over the 107 hour length of the experiment. The results indicate that the ECRS tank was not packed evenly, with what appears to be areas of hard pack through the center that retard the mass of bromide flowing through. Despite the heterogeneity, the bromide mass was conserved and some general trends were seen. There was an uneven bromide front, with flow reaching the lines located ~ 2ft from the side walls appearing first, showing that the gravel pack did not fill up before it began leaving the area and the bromide solution didn’t then move as a uniform wave as was predicted. Sampling locations 2 feet off the bottom vertically at any location saw only trace amounts of bromide, on the order of 10 times less than their counterpart 1.5 ft below at the same x,y coordinate, demonstrating that there was almost no vertical mixing. Both of main sets of results lead us to believe that there is very low dispersion through the sand and the ECRS is very similar to field conditions marked with much heterogeneity.
Tank Preparation: The tracer test proved to be very beneficial to the entire group and served as the foundation for all further experimental decisions. The understanding of the hydraulic characteristics dictated the placement the of PCE DNAPL injection into the ECRS tank. PCE was injected through manually syringing approximately 0.33 L into 3 sampling lines, which made up the source zone. Each of these injection points are located 2 feet from the base of the tank and roughly two-thirds down the length of the tank. This location was chosen so that dechlorination products could be seen in the effluent in a relatively short period, 3-5 days, and to allow for sufficient distance upstream for fermentation to occur. Based on bench scale work in the laboratory, this amount was calculated so that pooling of DNAPL on the bottom on the tank would not occur. The PCE was allowed to distribute under recycle flow conditions for two weeks to establish a residual high aqueous concentration of 10 to 30 mg/L. Following the PCE addition anaerobic conditions were established by the addition of acetate. Once the ECRS was made anaerobic, 30 gallons of microbe solution (10 mg/L biomass) was added. Shortly afterwards, lactate was added as an electron donor to help establish the microbe population. Finally, 21 gallons of hydrogen release compound (HRC) was added to act as the electron donor for the remainder of the experiment.
Spatial Samplings: After the introduction of a mixed anaerobic culture into the ECRS, and the Hydrogen Release Compound (HRC) injection up gradient of the source zone to provide a slow release of electron donor, dechlorination performance was monitored with effluent samples to check for the presence of metabolites (TCE, DCE, VC, ethene). The successful bioaugmentation and subsequent detectable concentrations of dechlorination products from effluent samples in the ECRS provided a favorable system for the study of the interaction between the microbial communities and the resulting chemical products. Concentration measurements of chlorinated ethenes (PCE, TCE, DCE, VC and ethene), pH, methane, acetate, propionate and chemical oxygen demand (COD) were determined through chemical analysis. The spatial distributions of all the above constituents were evaluated through a series of comprehensive samplings of the ECRS. The chemical analyses of the dechlorination products were performed through GC head space analysis. The results from these analyses were plotted with the Groundwater Modeling System (GMS) for a graphical interpolation of the data.
Tremendous spatial variability between the zones of biological activity and in the distribution of chemical products was observed in ECRS system. The system was initially thought to be geologically homogenous. However, the ECRS system resembled the heterogeneity observed at a typical field site more closely than was anticipated. In some cases, samples from locations only a few feet apart yielded very different results. In the three sampling locations constituting the source zone, PCE and microbes were injected in equal amounts and using the same procedure. However, there existed considerable variation among the resulting chemical constituents and dechlorination activity in the source zone lines residing in the same transect.
Comparisons between the performance at the source zone locations were based on the concentrations of not only the dechlorination products themselves, but also through other biological indicators such as methane and COD. These comparisons provided an additional understanding of what mechanisms were responsible for the processes observed in the source zone. The results of this study found that the variation in electron donor delivery was primarily responsible for the differences observed in the dechlorination performance in the source zone locations. In addition, the methanogens and dechlorinators in the ECRS appeared to have a complementary relationship. Both communities thrived under high COD conditions and there seemed to be adequate electron donor. In samples taken from areas where chlorinated ethenes were present, the production of methane was concurrent with the formation of dechlorination products. These observations imply that while methanogens were competitors with dechlorinating organisms for the hydrogen substrate, the dechlorination process was not impeded by active methanogenesis in the ECRS system.
Statistical Analysis: The daily aqueous samples taken from the ECRS effluent to monitor the biological and chemical activity provided a substantial data set that up to this point had only been analyzed through qualitative means of graphical plots. In the reductive dechlorination process, assessing success (PCE going to its final product, ethene) on the field-scale is very difficult and data intensive. Therefore, by understanding which biological processes are correlated, a species or group of species may act an indicator of successful dechlorination. There were a total of 203 observations; however, some were removed from data set because they did not include all of the desired variables. The remaining data was comprised of 147 observations. This portion of the data was collected exclusively from the effluent; therefore, it is an average of the activity occurring in the ECRS tank.
Five to six predictors (methane, COD, acetate, pH, propionate, plus after
some initial tests, the previous day’s dechlorination performance) were
used to estimate the current day’s dechlorination performance through
various statistical methods. Rather than solely analyzing the concentrations
of the dechlorination products to assess the degree of biological activity,
the chlorine number was also used as an indicator of performance. The chlorine
number composites the chlorinated ethene concentrations and scales them from
0 to 4. At zero, no chlorinated species are present (all ethene or ethane remains);
at 4 all of the ethene is in the form of PCE. The optimal outcome is for the
chlorine numbers to decrease over time, demonstrating that PCE was being dehalogenated
to its daughter products. An advantage of the chlorine number analysis beyond
comparing concentrations is that it standardizes the amount of products present
regardless of concentrations, by weighing the amount of chlorinated ethenes
present. It is not known what exact initial concentrations of PCE were present
at each location after injection; however, the chlorine number compensates
for this, too. For statistical analyses, the chlorine number acts as a single
response variable for how the biological activity is affected by the chemical
environment.
Previously, r2 correlations had been calculated through Excel. These correlations
showed evidence of the relationship of some species; however, it was typically
not consistent between sampling events and proved to provide only a limited
interpretation. Therefore, this portion of the study has performed a more in-depth
statistical analysis. The statistical package Minitab was used for all calculations.
With the five to six predictor variables, the first step was to determine which
ones would best estimate the response variable, the chlorine number. This selection
was done using the Stepwise Regression function, which through repeated testing
of hypotheses was able to determine whether an adding or removing a variable
improves the statistical significance of the model. Once the significant variables
were determined, a simple Linear Regression was performed. Additional transformations
on the data were performed to optimize the prediction.
The final result was that the statistical methods of stepwise regression, linear regression and time series autocorrelation, along with a logarithmic transformation on the data set selected the lagged chlorine number variable and methane to be the best predictors of the current chlorine number, the response variable representing the cumulative concentration of the chlorinated ethenes remaining. The other two significant predictors, pH and COD, decrease as the chlorine number decreases. For pH, the increased biological activity represented by dropping chlorine number causes the production of slightly acidic conditions. While for COD, as the dechlorination process proceeds the COD, which is composed of lactate that is transformed to the microbe’s energy source hydrogen, is consumed. This result verified Capiro’s Master of Science research results that methane producing microbes (methanogens) do not impede dechlorination performance within an aquifer system, but instead they are an indictor that the process is occurring (Capiro 2002). Both microbe communities are able to utilize the hydrogen electron donor without severe issues in competition, as shown by the stepwise regression result that as chlorine number decreases and dechlorination occurs, methane increases.