The remediation of Department of Defense sites contaminated by chlorinated solvents and source zones with dense nonaqueous phase liquids (DNAPLs) continues to be extremely difficult, often relying on in situ remediation approaches such as in situ chemical oxidation, surfactant enhanced aquifer recovery, and six-phase electrical heating. Remediation goals can be complicated by questions concerning the validity of the monitoring data used to assess a site and judge the performance of a remediation project. Standard practices for characterization and performance assessment are based on sampling subsurface soil and groundwater within and around the contaminated zone before and after remediation has occurred. Although widely employed, these standard practices do not necessarily yield accurate and unbiased estimates of contaminant levels or remediation effectiveness. Performance assessment errors can result from both DNAPL compound losses that occur during sample acquisition and handling and unaccounted-for-effects of remediation-induced changes in the partitioning behavior of any DNAPL compounds that remain in the subsurface following remediation. Understanding how sampling methods can impact the accuracy of volatile organic compound (VOC) measurements in samples of soil and subsurface porous media is often critical to sound decision making during characterization and remediation of VOC-contaminated sites.

The objectives of this project were to (1) increase the fundamental understanding of how monitoring practices can impact the validity of data collected from groundwater wells and direct-push cores as well as its use for performance assessment of remediation effectiveness and (2) identify improved practices to mitigate monitoring-related measurement errors and assessment impacts at chlorinated solvent sites, including those with DNAPL source zones.

An integrated set of tasks were carried out during a laboratory study focusing on methods for monitoring chlorinated solvents in soil and groundwater systems when the concentrations are at levels below and above the threshold for a DNAPL phase to be present. The primary chlorinated solvents studied included three pervasive contaminants of concern (COCs) with contrasting properties: tetrachloroethene (PCE), trichloroethene (TCE), and 1,1,1-trichloroethane (TCA). Several types of porous media were studied, including those with different properties potentially important to COC behavior (e.g., grain size, porosity, water content, fraction organic carbon content [f_oc]). For control purposes, a simulated groundwater matrix was used.

Two primary series of controlled laboratory experiments were conducted. One series of experiments involved sampling of cores of porous media containing known quantities of PCE, TCE, and TCA at varied concentrations with media temperatures between 5°C and 80^°C. Five methods of sample collection and preservation were selected to be representative of field methods, as determined through a survey of practitioners. The methods have different degrees of media disaggregation and atmospheric exposure. The other series of experiments investigated the impact of treatment on contaminant behavior by quantifying changes to f_oc and the nature of contaminant-organic carbon partitioning (K_d, K_oc). Zero headspace reactors were used with porous media of varied grain size and natural organic matter (NOM) content with different levels of TCE present. The remediation agents examined include two oxidants (potassium permanganate and activated sodium persulfate) and two surfactants (DowFax 8390 and Tween 80). To examine thermal remediation effects, soil samples from a field site also were examined. Changes to sorption behavior are characterized by measuring soil f_oc and the distribution coefficient (K_d) before and after treatment, then calculating the organic carbon partition coefficient (K_oc).

Experiments to quantify chlorinated solvents in porous media demonstrated that sampling method attributes can impact the accuracy of VOC measurements in porous media by causing negative bias in VOC concentration data ranging from near 0% to 90% or more. The magnitude of the negative bias is highly dependent on the attributes of the sampling method used and interactions with key contaminant properties and environmental conditions (i.e., VOC K_H, temperature, water saturation level).

Investigations on the effect of remediation of porous media properties and contaminant partitioning revealed that the f_oc of the porous media treated by oxidants or surfactants was higher or lower relative to that in the untreated media controls, depending on the treatment agent and the extent of contamination. Isotherm experiments were run using the treated and control media to experimentally measure the K_d of TCE. Values of K_oc calculated from the experimental data revealed that the K_oc values for TCE in the porous media were altered via treatment using oxidants and surfactants.

Common and widely employed practices for monitoring DNAPL compounds in the subsurface are not clearly understood and may lead to flawed conclusions regarding in situ remediation effectiveness. During performance assessment, the distribution of remedial agents in the subsurface and the effectiveness of in situ remediation may be judged as unsuccessful when it actually is successful or as successful when it really is not. The results of this project help to delineate the limitations of current standard practices and to develop improved monitoring methods. These methods will account for the effects of remedial agents on subsurface properties that affect chlorinated solvent behavior and, in turn, how those changes affect performance assessments of in situ remediation effectiveness.