It is well known that the anaerobic biodegradation rates for a single chlorinated compound can vary over four orders of magnitude. When a new chemical compound was developed that could potentially increase the contaminant biodegradation rates it was decided that a study of mechanisms of the chemical reactions involved would be useful to help understand the key features of the process. A chemical model has been developed that can be applied to microcosm and larger scale simulations in the laboratory, as well as results from the field.
As with any sequence of chemical reactions, the model has to be specific for the study of the biodegradation of each chlorinated compound. The new chemical compound that will assist anaerobic biodegradation is a polylactate ester of sorbitol, xylitol or glycerol. These polylactate esters slowly release lactic acid into groundwater over a period of many months. The lactic acid is biotransformed into pyruvic acid and subsequently into acetic acid releasing hydrogen in both steps. The ultimate fate of the acetic acid is probably to methane and carbon dioxide by methanogenesis. The acids can also be used as as carbon source for bacterial cell growth.
It would appear that the new compounds could be used wherever enhanced anaerobic biodegradation is desired. To study at least one of these cases in detail we selected the biodegradation of 1,1,2-trichloroethylene (TCE). The work is being extended to tetrachloroethylene (PCE) and pentachlorophenol (PCP). The TCE study is somewhat simpler to analyze and document since there fewer chemical steps involved.
MATERIALS AND METHODS
The experimental data on TCE degradation is obtained by injection of known amounts of TCE in water into soil samples that are suspected to contain bacteria that could dechlorinate TCE. The lactic acid releasing polymer was added to the soil and water in various amounts. Generally TCE and its degradation products dichloroethylene (DCE) and vinyl chloride (VC) were measured by extracting an aqueous phase sample and mixing it with an internal standard of toluene dissolved in water. The CAHs and the toluene are then measured by headspace analysis using gas chromatography and both a photoionization detector (PID) and finally by a flame ionization detector (FID). Detection limits of the PID are 5 m g/L based on the concentration in the original liquid sample being analyzed.
Lactic acid, acetic acid and pyruvic acid are measured by high performance liquid chromatography using a UV detector. Citric acid is used as an internal standard and the limits of detection are 1 mg/L in the original sample. Bacterial counts were made through use of standard pour plate techniques. Total Plate Counts (Colony Forming Units per ml) were measured both aerobically and anaerobically.
A positive control was made by using liquid from a bioreactor that had been continuously degrading TCE at the rate of 5 mg/(L-day). A series of two ten-fold dilutions were made and the bacterial counts of the starting liquid was also made. The diluted solutions were placed in a system with sterile soil and known amounts of TCE and the lactic acid releasing ester.
The main test system is 200 ml test tubes containing 10 grams of soil and 160 ml of liquid at concentrations of 25 mg/L and 10 mg/L of TCE. Concentrations of TCE as high as 100 mg/L have been used. An alternative pilot systems uses 33 L of soil in a tube 1.83 m long and 15.25 cm in diameter. In this system the lactate releasing material in injected into the soil and the contaminated water flows through the tube at rates similar to groundwater flow rates. Sample ports in the tube occur every 15 cm.
Numerical analysis of the coupled differential equations was carried out by the Runga-Kutta-Gill algorithm. The model was programmed in Visual Basic for Applications as a series of macros in an Excel spreadsheet.
RESULTS AND DISCUSSION
The system of equations to be described here can be generalized to any number of species limited only by the programmer’s patience and the availability of sufficient data to allow meaningful interpretation of the parameters. The large number of parameters involved leads to the accusation that it is always possible to make the model fit the data if the parameters are selected correctly. The value of the model, however, is not in fitting the data.
The model provides a rationale for selection of various mechanisms. A good model will work over a wide range of experimental parameters. Changes in the behavior of the system will be easily understood in terms of the model equations.
The chemical sequence is:
TCE => DCE => VC => ethylene
There are three groups of bacteria that we must consider:
1. bacteria that metabolize TCE (B1)
2. bacteria that produce hydrogen (H2) from LA (B2)
3. bacteria that produce lipase for polylactate cleavage (B3)
These groups are not distinct. By this we mean that group B3 contains group B1 and B2 and in general will be larger than the sum of B1 and B2. Group B2 will be smaller than B3 but larger than B1. Since these bacteria catalyze the reactions we will have to keep track of their amounts during the course of the analysis. Bacteria can be modeled by first order kinetics:
The square brackets denote concentration and, in general, will be expressed in millimoles per L (mm/L). Bacteria are typically reported as numbers (CFU) per ml. In order to keep units consistent, bacterial concentration will be given as the plate count per liter and all rates will be per day. Thus, the ki have units of day-1.
If we now consider what happens in an aquifer or a test tube we will use a 'control volume' that contains soil; water and some TCE both dissolved in the water and adsorbed on the soil. A 'control volume' is a convenient way to keep track of changes that occur. In this defined volume we can keep track of everything that moves into the volume, the changes that occur in the volume and everything that moves out of the volume. This allows us to perform overall mass and species balance on the volume. The control volume used here does not allow any of the unused lactic acid (LA) or TCE to escape. In an aquifer this is essentially the volume downstream of the polylactate ester (HRC) injection site that encompasses the LA and metabolic products. The actual total amounts of material, e.g. in pounds, is calculated by multiplying the total volume by the concentration. In the aquifer case the volumes will be on the order of thousands of liters. In the test tube it is 170 ml. However, we keep the model very general by using a control volume that simply encompasses all of the aquifer of interest and is based on the concentration contained in that volume.
We place HRC into this volume and several things happen. HRC breaks down to LA. The bacterial group B3, the largest of the groups, catalyzes this. The LA can then be used to produce hydrogen. Unfortunately it can also be used to produce more bacteria that do not produce hydrogen and it can react chemically. Further LA can be adsorbed and previously adsorbed LA can be released. TCE can move into the control volume. TCE can be degraded and it can be adsorbed. TCE that was previously adsorbed can be released. In this model we will assume that TCE does not react chemically. The TCE equation is:
where RTCEI is the rate at which TCE enters the volume in mm/(L day), [S] is the number of sites associated with the soil for the adsorbtion of TCE (mm/L) and ka TCE is rate for TCE adsorbtion L/(mm day).
Note that [S] may be combined into ka TCE for all practical purposes since the combination will be measured together. The units would then be day-1.
[TCEA] is the amount of TCE adsorbed on the soil at any time (mm/L), kdTCE is the rate of desorbtion (day-1), k1 is the rate of remediation with units of l2/(mm day), a1 is the exponential catalyst (bacterial) dependence.
RTCEI is the input flow rate of TCE. The model assumes that there is some TCE initially adsorbed on the soil and some in the dissolved phase. RTCEI accounts for new TCE coming into the treatment volume. We need to account for the change in adsorbed TCE with time:
Adding the rest of the equations in similar notation we have:
The term A indicates the adsorbed species; for example, VCA is adsorbed VC.
Figure 1 shows typical results for the use of the model in explaining the reduction of TCE in a microcosm of active bacteria at around 10,000 CFUs per ml. Typical soils are less than 1,000 CFUs per ml. With bacterial levels the same reduction can take 25 days. Figure 2 is a typical result that illustrates the utilization of lactic acid when the bacterial count is very high and the polylactate ester is limited. This example also shows the model can predict the maximum lactic acid concentration.
Standard chemical kinetic formulations are useful in predicting concentrations of chemical species in active dechlorinating systems using a polylactate ester that releases lactic acid. Modeling the effect of bacterial concentration as a catalyst with the order determined by the data appears to provide a reasonable understanding of the experimental data.
FIGURE 1 Results of TCE Reduction Model vs. Experiment
FIGURE 2 Results of Lactic Acid Production Model vs. Experiment