Diffusional Rate Limitations to Contaminant Transport in Heterogeneous Groundwater AquifersEPA Grant Number: U914736
Title: Diffusional Rate Limitations to Contaminant Transport in Heterogeneous Groundwater Aquifers
Investigators: Cunningham, Jeffrey A.
Institution: Stanford University
EPA Project Officer: Broadway, Virginia
Project Period: January 1, 1995 through January 1, 1996
Project Amount: $102,000
RFA: STAR Graduate Fellowships (1995) Recipients Lists
Research Category: Academic Fellowships , Engineering and Environmental Chemistry , Fellowship - Engineering
The main objective of this research project is to investigate the nature of the processes that govern contaminant transport in heterogeneous groundwater aquifers.
The use of groundwater as a natural resource is threatened by contamination from hazardous chemicals. Restoration of a contaminated aquifer can be impeded by contaminant sequestration into aquifer solids, or by the heterogeneous nature of the aquifer's physical and chemical properties; proper development of aquifer protection and remediation strategies must account for these impediments. One model that commonly has been used to describe contaminant uptake and release (often referred to as sorption and desorption) by aquifer solids attributes these processes to contaminant diffusion through an intra-granular or intra-aggregate pore structure. However, this model now appears insufficient to describe all of the observed results from particular laboratory experiments (reported previously elsewhere, but reviewed in this dissertation). In response to this insufficiency, I expand the sorption-desorption model to consider aquifer grains as "biporous" in nature (i.e., as containing two distinct intragranular pore regimes). This model often has been used in chemical engineering to describe engineered solids such as catalyst pellets or ion exchange zeolites, but only recently has been introduced to describe natural solids such as grains of aquifer material. I consider one of the intra-granular pore regimes as an interconnected network of "mesopores" of relatively large diameter, saturated with water under normal aquifer conditions. The second pore regime is proposed to consist of relatively straight, short, narrow "micropores," which branch off from the interconnected network of mesopores. Contaminant sorption and desorption, therefore, occur via two diffusion processes in series: first through one pore regime, then through the other.
In nearly all cases considered, model simulations compare very well with experimental results. I consider contaminant transport through the mesopore network to occur via aqueous diffusion, retarded by adsorption onto the pore walls and by pore tortuosity; all parameters affecting the rate of the mesopore diffusion process are estimated a priori. To describe the micropore diffusion process, two or three "fitting" parameters are required: one to describe the micropore capacity, and one or two to describe the micropore diffusion rate. In some instances, a single diffusion rate is sufficient to describe experimental observations; in other cases, a distribution of micropore diffusion rates is required, increasing the total number of adjustable parameters from two to three. The experimental data imply that the distributions of micropore diffusion rates might be very broad, although there is considerable uncertainty with respect to estimation of the distributions.
If an aquifer's hydraulic conductivity is homogeneous (spatially uniform) over a large domain, then contaminant plume spreading and tailing can be controlled by the sorption-desportion processes, particularly if these are slow compared to the processes of advection and dispersion. In this dissertation, I use temporal moment analysis to quantify these effects. The second temporal moment, representing the spread of plume arrival times at a control plan, depends on the mean time scale for contaminant sorption. The third temporal moment, representing skewness of plume arrival times (i.e., plume tailing), also depends on the variance of the distribution of sorption time scales. Because the distribution of sorption time scales can be quite broad, it might be reasonable to expect significant plume tailing even in cases where the mean sorption rate is relatively fast.
When an aquifer is characterized by significant heterogeneity of hydraulic conductivity, then both the contaminant sorption-desorption processes and the heterogeneous groundwater velocity field can affect plume behavior. By considering the second temporal moment for contaminant arrival at a control plane, I formulate a criterion that distinguishes between the conditions under which plume spread is controlled by slow contaminant sorption and desorption, by heterogeneity of hydraulic conductivity, or by a combination of both. Results of simulations show that, under some conditions, it might be possible to approximate contaminant sorption and desorption as occurring "instantaneously" (local equilibrium assumption); under other conditions, it might be possible to approximate the aquifer's hydraulic conductivity field as spatially homogeneous.