USEPA - Technology Innovation and Field Services Division (TIFSD)

Single field mobilization completes site investigation and removal actions

The Kentucky Research Consortium for Energy and the Environment (KRCEE) worked with the U.S. Department of Energy (DOE) last summer to introduce new approaches and technologies to the site cleanup program operating at the Paducah Gaseous Diffusion Plant (PGDP) in western Kentucky. A Triad demonstration was conducted at PGDP's Area of Concern (AOC) 492 that integrated several real-time characterization tools into dynamic work strategies (DWSs) for expediting characterization and remediation within one field deployment. Field work focused on methods to address uranium and polychlorinated biphenyls (PCBs), two of the most common soil and sediment contaminants at PGDP. Those methods included laser-based gamma 'walkover' surveys (GWS), x-ray fluorescence (XRF), in situ gamma spectroscopy (ISGS), PCB immunoassay kits, multi-increment sampling (MIS), and adaptive compositing (AC).

The PGDP has operated for more than 50 years and is now the only active uranium enrichment facility in the United States. Past operations released contaminants to ditches that emptied into neighboring streams, contaminating sediments. Maintenance activities within the streams and ditches resulted in contaminated sediments being placed on streams banks and adjacent upland areas including AOC 492. Limited historical information (three surface samples) identified uranium and PCB concentrations of 1,150 and 44 ppm, respectively, in AOC 492 surface soil.

In the fall of 2007, a group of technical representatives from EPA Region 4, the Commonwealth of Kentucky, DOE, KRCEE, and Argonne National Laboratory (ANL) collaborated in a systematic planning session to devise the initial CSM and form the project's DWSs. The session yielded default risk-based, wide-area averaged and hot spot project criteria of 11 and 99 ppm, respectively, for uranium and 3.6 and 33 ppm for PCBs. It also identified three exposure units—based on a teenage recreational user exposure scenario—within AOC 492. Based on an understanding of uranium processing and past experience, PCBs were expected to be collocated with uranium contamination. The field strategy included three activities deployed in one field effort: characterizing the level and extent of soil contamination, excavating soil with concentrations exceeding project criteria, and demonstrating post-excavation that project criteria were attained for each of the exposure units.

The laser-based GWS for determining the presence of elevated gamma radiation in near-surface soil was the first tool to be deployed. Approximately 24,000 GWS data points were collected over the course of three days, providing spatially dense information (an average of four measurements per square meter) about the location of uranium contamination in near-surface soil. Based on the GWS data, twenty locations were further sampled and analyzed by XRF and ISGS. The selected locations spanned the range of GWS results; the XRF/ISGS data were used to develop a relationship between GWS results and surface soil uranium concentrations. Once that relationship was established, uranium hot spots could be identified based on GWS data (Figure 2), and the layout of exposure units modified to reflect the spatial distribution of contamination. Approximately 13 m³ of soil were subsequently removed from the uranium hot spots and placed in an offsite disposal facility. GWS and in situ XRF readings verified that the excavation surface complied with the project hot spot criteria.

MIS and AC sampling techniques were used jointly with the PCB immunoassay kits to verify that PCB hot spots did not exist for the two exposure units considered most likely impacted by contamination. MIS uses soil collected from multiple locations spread systematically across a decision unit to form a sample that is more representative of the average conditions across an area than any one individual sample location would be. AC combines samples from adjacent decision units into composites before analysis; during the compositing process, the contributing samples are split, with one half of the splits archived for possible later analysis and the other half used to make the composites. Field investigation levels are defined as a function of the number of samples contributing to the composite and the original project criteria. If the analytical results of a composite exceed the field investigation level, each of the archived sample splits contributing to the composite is analyzed. Use of MIS and AC minimized the number of sample analyses needed; only 23 were necessary to verify compliance with hot spot and area-averaged cleanup criteria for the entire study area. More than 300 analyses would have been required if each of the soil increments had been analyzed individually.

Rigorous data quality control (e.g., use of calibration standards and control charts) was developed and implemented for the project to ensure technically defensible datasets were obtained from field analytics. Data quality for the XRF uranium measurements was comparable to standard laboratory alpha spectroscopy data quality, with a correlation coefficient of 0.99 and detection limits at background levels. Each sample analyzed by immunoassay also was analyzed at an offsite laboratory for confirmation. In general, the PCB immunoassay kits compared favorably with laboratory total PCB data, with a correlation coefficient greater than 0.9.

The field work generated more than 20,000 individual GWS data points, several hundred XRF measurements, and almost 400 surface soil increments. Characterization of AOC 492, contaminated soil removal, and verification that project criteria had been achieved were completed in just 10 days of field work. The use of MIS and AC significantly reduced the number of soil samples requiring analysis, while the use of field analytics reduced the cost of each sample analysis. The availability of real-time data (e.g., XRF, ISGS, GWS, and PCB immunoassay kit results) allowed a seamless integration of characterization, remediation, and verification sampling. Demonstration results indicated that use of Triad and a suite of investigative tools effectively characterized the soil, identified locations requiring remediation, guided soil removal, and verified that cleanup criteria were achieved.

Contributed by Rich Bonczek, DOE-Portsmouth/Paducah Project Office, Steve Hampson, KRCEE, and Robert Johnson, ANL.

Customer comments

No comments were found for Single field mobilization completes site investigation and removal actions. Be the first to comment!