DNAPL is an acronym for "Dense Non-Aqueous Phase Liquid." "Dense" means the density of the liquid is greater than that of water. It weighs more than water. "Non-Aqueous Phase" means that when the liquid is added to water, it remains separated from the water and does not readily dissolve. Your favorite pancake syrup is probably a viscous DNAPL. However, toxic low-viscosity DNAPLs that flow more easily through porous soils are of concern in the contamination of groundwater in aquifers. Large quantities of low-viscosity DNAPLs are used for commercial, industrial, and military purposes. The most common are solvents used as dry cleaning fluids, degreasers, and other processes; and they have been spilled and released to the subsurface at many sites. They are known by acronyms for complex chemical names, such as PCE and TCE. Their behavior in aquifers is complex.
UTCHEM (University of Texas Chemical flooding simulator) is a computer program that is capable of simulating the complex behavior of DNAPLs in aquifers. So far as I know, the original program that was distributed by the University of Texas is the most comprehensive model applicable to DNAPLs that exists. It is a finite-difference model, which means that it divides the groundwater system into a grid of blocks and calculates changes within each block and the simultaneous movement of water and DNAPL from block to block through time. Consequently, it is capable of dealing with aquifers with properties, such as permeability, that vary from place to place in the aquifer. It can simulate the chemical interaction between DNAPL and water, and perform other complex groundwater contamination modeling. It can simulate the behavior of LNAPLs such as gasoline and oil that float on the groundwater.
UTCHEM DNAPL CONTAMINATION MODEL
The UTCHEM simulation discussed in this non-technical webpage and the corresponding technical webpage (DNAPL Contamination (T)) is a model of PCE behavior that is simplified by using a vertical slice through the aquifer oriented in the direction of groundwater flow. The length of the slice is 162 feet. The depth is 39 feet. The width is 49 feet. The simulation is for a PCE spill that is released at the center of the top of the model at a rate of 33.3 cubic feet per day (about 1.3 pints per minute) for 30 days, resulting in a total spill of 1000 cubic feet (7,481 gallons). The model simulates the downward percolation of the PCE that is caused by its density being greater than the water in the aquifer. The weight of the PCE causes it to move downward. This downward percolation is slowed because the PCE is surrounded by the water and must squeeze through the small interconnections between the pores of the aquifer. Figure 1 shows the variation of the permeability in the aquifer and depicts the rate of downward percolation of the PCE. The higher permeability is surrounded by the red contours, and the lower permeability is surrounded by the blue contours. The rate of downward percolation is depicted by showing the approximate extent of PCE intrusion at successive times since the beginning of the spill of 5, 15, and 30 days using purple contours. The contours in this figure were produced from the UTCHEM data files by a computer program called CONTOUR that was developed by the U.S. Geological Survey for contouring such gridded data.
Figure 1. Permeability contours and extent of liquid PCE at 5, 15, and 30 days.
The downward percolation of the PCE is affected by several properties that were input to the model. These properties included (1) the "dispersivity" of the PCE, which is a measure of its tendency to spread as it moves through the porous subsurface material, (2) weight of the PCE relative to the water (it is 60 percent heavier than water), (3) the solubility of PCE in water, (4) the viscosity of the PCE (about 10 percent less than water), (5) data relating the pressure in the PCE, pressure in the water, and the amount of PCE in a block ("capillary pressure curve"), (6) data relating the permeability to PCE relative to the permeability to water as the amount of PCE in a block changes (the less PCE, the lower the permeability to PCE), and (7) the irreducible water saturation (a measure of the amount of water in a block that cannot be displaced by PCE).
As the liquid PCE moves in the aquifer, some of it dissolves into the water until a maximum (equilibrium) concentration is reached in the water. Dissolved PCE is carried away from the liquid PCE by the flowing groundwater and disperses to form a plume of dissolved PCE. Figure 2 characterizes the distribution of dissolved PCE at the end of the 30-day spill. It also shows the model grid blocks. The red block is the spill location. The brown contour encloses the part of the aquifer where the concentration is close to the maximum because it is near chemical equilibrium with the liquid PCE. The green contour encloses the part of the aquifer where the PCE concentration is above the maximum contaminant level (MCL) set by the U.S. Environmental Protection Agency for public water systems (5 parts per billion). The groundwater in this model is flowing from left to right, so Figure 2 shows that the dissolved PCE is at its maximum equilibrium concentration near the liquid PCE, and the plume where dissolved PCE is greater than the MCL extends a short distance in the direction of groundwater flow. At 30 days, only 0.18 percent of the PCE has dissolved in the water.
Figure 2. Concentration of PCE dissolved in the groundwater at 30 days (the end of the spill).
Figure 3 depicts the movement of the liquid PCE downward after the end of the spill. The blue contour encompasses the liquid PCE at 5 days after the spill; the green contours represent 15 days after the spill; and the red contours encompass locations of liquid PCE 30 days after the spill. The liquid PCE is shown to have moved downward and then spread along the bottom of the aquifer, leaving behind isolated pockets where restrictions that slowed downward flow were encountered. The spread is greater in the direction of groundwater flow (rightward) than in the upstream direction.
Figure 3. Location of liquid PCE 35, 45, and 60 days from beginning of spill (5, 15, and 30 days after cessation of the spill).
The blue contours in Figure 4 show the location of dissolved PCE at 30 days after the cessation of the spill. The plume has followed the liquid PCE downward and laterally, and spread in the direction of groundwater flow.
Figure 4. Location of dissolved PCE above MCL at 30 days after the end of the spill.
Figure 5 compares the distribution of liquid PCE at 60 and 300 days after the beginning of the spill (30 and 270 days after the cessation of the spill). Most of the liquid PCE has descended to the base of the aquifer and continued to spread (more in the direction of groundwater flow than against the flow).
Figure 5. Contours enclosing most of the liquid PCE at 60 and 300 days (30 and 270 days after the end of the spill).
Before 270 days after the cessation of the spill, the plume of dissolved PCE greater than the MCL extended past the right boundary of the model, which is in the direction of groundwater flow. At this time, 990.24 cubic feet of the 1000 cubic feet spilled is still liquid PCE. Less than 1 percent has dissolved into the groundwater.
These results demonstrate problems associated with PCE aquifer contamination. PCE will tend to migrate downward rapidly to depths that are impractical to excavate. Then it will tend to spread slowly, providing a source for dissolved PCE, which is deemed toxic at very low concentrations. The dissolved PCE can move rapidly with the groundwater flow, creating an extenesive plume of contaminated groundwater.
MONITORED NATURAL ATTENUATION OF DNAPL CONTAMINATION
These results also demonstrate why monitored natural attenuation (MNA) is unlikely to be an acceptable method for dealing with DNAPL contamination at a site.