EPA/540/R-05/007 July 2004 Field Evaluation of TerraTherm In Situ Thermal Destruction (ISTD)Treatment of Hexachlorocyclopentadiene Innovative Technology Evaluation Report Diana Bless, Project Officer National Risk Management Research Laboratory Office of Research and Development U. S. Environmental Protection Agency Cincinnati, Ohio 45268 NOTICE The information in this document has been funded by the U.S. Environmental Protection Agency (EPA) under Contract No. 68-C-00-181 to Tetra Tech EM Inc. It has been subjected to the Agency’s peer and administrative reviews and has been approved for publication as an EPA document. Mention of trade names or commercial products does not constitute an endorsement or recommendation for use. FOREWORD The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the Nation’s land, air, and water resources. Under a mandate of national environmental laws, the Agency strives to formulate and implement actions leading to a compatible balance between human activities and the ability of natural systems to support and nurture life. To meet this mandate, EPA’s research program is providing data and technical support for solving environmental problems today and building a science knowledge base necessary to manage our ecological resources wisely, understand how pollutants affect our health, and prevent or reduce environmental risks in the future. The National Risk Management Research Laboratory (NRMRL) is the Agency’s center for investigation of technological and management approaches for preventing and reducing risks from pollution that threaten human health and the environment. The focus of the Laboratory’s research program is on methods and their cost-effectiveness for prevention and control of pollution to air, land, water, and subsurface resources; protection of water quality in public water systems; remediation of contaminated sites, sediments and ground water; prevention and control of indoor air pollution; and restoration of ecosystems. NRMRL collaborates with both public and private sector partners to foster technologies that reduce the cost of compliance and to anticipate emerging problems. NRMRL’s research provides solutions to environmental problems by: developing and promoting technologies that protect and improve the environment; advancing scientific and engineering information to support regulatory and policy decisions; and providing the technical support and information transfer to ensure implementation of environmental regulations and strategies at the national, state, and community levels. This publication has been produced as part of the Laboratory’s strategic long-term research plan. It is published and made available by EPA’s Office of Research and Development to assist the user community and to link researchers with their clients. Sally Guitterez, Acting Director National Risk Management Research Laboratory ABSTRACT This report summarizes the U.S. Environmental Protection Agency (EPA) Superfund Innovative Technology Evaluation (SITE) Program evaluation of the In Situ Thermal Destruction (ISTD) technology developed by others and refined by TerraTherm, Inc. The demonstration was designed to evaluate the technology's ability to treat soil-and-waste material contaminated with hexachlorocyclopentadiene (hex) and chlorinated pesticides at a former disposal pit (the Hex Pit) located at the Rocky Mountain Arsenal in Commerce City, Colorado. Operation of the system was terminated soon after initial startup and before the SITE demonstration could be completed, due to the destruction of system components from highly corrosive vapors and liquids. ISTD is a soil remediation process that applies heat and vacuum simultaneously to contaminated soils, either with surface heater blankets or with an array of vertical heater and vacuum extraction wells. The ISTD system at the Hex Pit used an array of vertical heater and combination heater and vacuum extraction wells. According to the developer, as the soil is heated, volatile contaminants are vaporized or destroyed by a number of mechanisms, including the following: (1) evaporation into the vapor stream, (2) steam distillation into the vapor stream, (3) boiling, (4) oxidation, and (5) pyrolysis (Stegemeier and Vinegar 2001). Most of the contaminants are expected to be destroyed in the soil before the vapor stream is removed by vacuum extraction. Contaminants that have not been destroyed in situ and remain in the vapor stream are destroyed by an off-gas treatment system. Evaluation of the ISTD technology as part of this SITE demonstration included extensive sampling to characterize soil-and-waste material in the Hex Pit before construction and startup of the ISTD system. In general, the Hex Pit contains layers or bands of virtually pure, tar-like waste material interlayered with soil that was used to cover the waste. Due to the early termination of the treatment process, SITE’s project objectives and post-treatment sampling were modified from the original plan. For post-treatment sampling, the revised demonstration objective was to evaluate potential contaminant destruction or removal resulting from short-term operation of the system in the near vicinity of combination heater and vacuum extraction wells. Sampling results were inconclusive regarding evidence of contaminant destruction or removal from short-term operation of the system. ISTD treatment at the Hex Pit was terminated 12 days after initial startup of the system due to the destruction of system components, likely from higher-than-anticipated production of hydrogen chloride (HCl). In addition, vapor-phase HCl condensed to the more corrosive liquid form in the system piping. Corrosion occurred in both aboveground and subsurface piping components constructed of 304 stainless steel. Destruction of the system components appeared to result from a combination of circumstances, including (1) the occurrence of layers of virtually pure, tar-like waste material that were not destroyed in situ; (2) the generation of HCl that was not adequately neutralized by in situ materials; (3) the choice of 304 stainless steel for system components, which was insufficiently resistant to corrosion; and (4) the inability of the system to maintain extracted vapors in the vapor phase for transport to the off-gas treatment system. ACKNOWLEDGMENTS This report was prepared for the U.S. Environmental Protection Agency (EPA) Superfund Innovative Technology Evaluation (SITE) program by Tetra Tech EM Inc. (Tetra Tech) under the direction and coordination of Marta Richards at the National Risk Management Research Laboratory (NRMRL) in Cincinnati, Ohio. The In Situ Thermal Destruction (ISTD) technology evaluation was a cooperative effort that involved the following personnel from EPA, the Rocky Mountain Arsenal (RMA), and TerraTherm, Inc. (TerraTherm): Marta Richards, EPA SITE Technical Project Manager Lorri Harper, RVO Project Manager Kerry Guy, EPA Region 8 Ralph Baker, TerraTherm Project Manager SECTION 1. INTRODUCTION This section provides background information about the U.S. Environmental Protection Agency (EPA) Superfund Innovative Technology Evaluation (SITE) Program, and discusses the purpose and organization of this Final Report. The technology evaluated in this report is the In Situ Thermal Destruction (ISTD) system developed by TerraTherm, Inc. (TerraTherm). The evaluation site is a former hexachlorocyclopentadiene disposal pit (the Hex Pit), located at the Rocky Mountain Arsenal (RMA) in Commerce City, Colorado. This technology evaluation has been conducted by the EPA SITE Program in cooperation with EPA Region 8, the Colorado Department of Public Health and Environment, and RMA’s Remediation Venture Office (RVO) (U.S. Army, Shell Oil Company, and U.S. Fish and Wildlife Service). Key contacts for additional information about the SITE Program, this technology, and the demonstration site are listed at the end of this section. 1.1 DESCRIPTION OF THE SITE PROGRAM AND REPORTS The Superfund Amendments and Reauthorization Act of 1986 mandates that EPA select, to the maximum extent practicable, remedial actions at Superfund sites that create permanent solutions (as opposed to land-based disposal) for contamination that affects human health and the environment. In response to this mand ate, the SITE Program was established by EPA’s Office of Solid Waste and Emergency Response and Office of Research and D evelopment (ORD). The SITE Program promotes the development, demonstration, and use of new or innovative technologies to clean up Superfund and other contaminated sites across the country. The SITE Program’s primary purpose is to maximize the use of alternatives in cleaning up hazardous waste sites by encouraging the development and demonstration of innovative treatment and monitoring technologies. It consists of the Demonstration Program, the Emerging Technology Program, the Monitoring and Measurement Technologies Program, and the Technology Transfer Program. This evaluation of TerraTherm’s ISTD technology was completed under SITE’s Demonstration Program. The objective of the SITE Demonstration Program is to develop reliable performance and cost data on innovative treatment technologies so that potential users may assess specific technologies. Technologies evaluated either are currently, or will soon be, available for remediation of Superfund sites. SITE demonstrations are conducted at hazardous waste sites under conditions that closely simulate full-scale remediation, thus assuring the usefulness and reliability of information collected. Data collected are used to assess the performance of the technology, the potential need for pre- and post-treatment processing of wastes, potential operating problems, and approximate costs. The demonstrations also allow evaluation of long-term risks and operating and maintenance costs. For this evaluation of the ISTD technology, however, no cost information was developed, because the ISTD system did not complete the demonstration. Technologies are selected for the SITE Demonstration Program through annual requests for proposals. ORD staff review the proposals, including any unsolicited proposals that may be submitted throughout the year, to determine which technologies show the most promise for use at Superfund sites. Technologies chosen must be at the pilot- or full-scale stages of development, must be innovative, and must have some advantage over existing technologies. Once EPA has accepted a proposal, cooperative agreements between EPA and the technology developer establish responsibilities for conducting the demonstration and evaluating the technology. The technology developer is responsible for demonstrating the technology at the selected site and is expected to pay any costs for transportation, operation, and removal of equipment. EPA is responsible for project planning, site preparation, sampling and analysis, quality assurance and quality control, and preparing reports and disseminating information. 1.2 PURPOSE AND ORGANIZATION OF THE FINAL REPORT The Final Report (Report) provides information on TerraTherm’s ISTD technology and includes a description of the demonstration and its results. EPA provides information regarding the applicability of each technology to specific sites and wastes; therefore, the Report includes information on site-specific characteristics. Each SITE demonstration evaluates the performance of a technology in treating a specific waste. The waste characteristics at other sites may differ from the characteristics of the treated waste; therefore, successful field demonstration of a technology at one site does not necessarily ensure that it will be applicable at other sites. Data from the field demonstration may require extrapolation for estimating the operating ranges in which the technology will perform satisfactorily. Only limited conclusions can be drawn from a single field demonstration. TerraTherm’s ISTD system did not complete the demonstration at the Hex Pit at RMA. Operation of the ISTD system was terminated soon after initial startup due to the destruction of system components from highly corrosive vapors and liquids. Consequently, this Report focuses primarily on site characteristics unique to the Hex Pit and the ISTD system design (Section 2.0); a description of the demonstration methodology and results, including a chronology of activities and events that occurred during operation of the ISTD system (Section 3.0); and a description of the component destruction and conditions that may have lead to the system’s destruction (Section 4.0). Section 5.0 lists the references used in preparing this Report. This report does not include cost information for the ISTD technology, because the demonstration was stopped during initial operation of the system. 1.3 DEMONSTRATION BACKGROUND This section describes the history of the Hex Pit at RMA and the selection of the ISTD technology for remediating contamination at the Hex Pit and for evaluation under the SITE Program. 1.3.1 Site History RMA is located in Commerce City, Colorado, 10 miles northeast of downtown Denver. The U.S. Arm y originally developed the 27-square-mile facility in 1942, primarily for manufacturing chemical weapons. After World War II, parts of the facility were leased to private industry for pesticide manufacturing. The Hex Pit is an unlined, earthen-disposal pit located near the northern edge of the South Plants Manufacturing Complex (South Plants) at RM A (Figure 1-1). The pit was used to dispose of distillation bottoms and other residues from the production of hexachlorocyclopentadiene (referred to as “hex” throughout this report), a manufacturing interm ediary used in the production of pesticides. Hex was produced in South Plants by Julius Hyman and Company from 1947 to 1951, and by Shell Chemical Company from 1951 to 1955. The black, tar-like distillation bottoms and residues, in drums and in bulk, were buried in the pit from mid-1951 to mid-1952. The waste material was periodically covered with soil backfill. Although the exact quantity of waste material disposed of in the Hex Pit was not recorded, it has been estimated that 833 cubic yards (cy) of waste was disposed of and that the pit contains a total of 2,005 cy of waste materials interlayered with soil backfill (TerraTherm 2001). By the end of 1952, the Hex Pit was completely covered with a soil cap. By 1954, it appeared as an unvegetated rectangular ground scar on aerial photographs. In 1976, waste materials from the H ex Pit were uncovered during construction of the foundation for Building 571B. Building 571B was constructed over the southern end of the pit. Building 571B was later demolished, and most of the foundation was removed (Tetra Tech EM Inc. [Tetra Tech] 2001). 1.3.2 Technology Selection Innovative thermal treatment was specified for remediation of the Hex Pit in the Record of Decision (ROD) (Foster Wheeler Environmental Corporation [FW ENC] 1996). Through the process identified in the ROD Dispute Resolution Agreement (DRA) (Program Manager Rocky M ountain Arsenal 1996) for this area at RMA., regulatory agencies overseeing environmental activities at RMA selected ISTD as the specific innovative thermal treatment to be used at the Hex Pit. RMA Remediation Goal 1 outlined in the ROD and DRA involves the destruction of contaminants to levels that met human health exceedance (HHE) criteria for the six site contaminants of concern (COCs). The six site CO Cs consisted of hex and the pesticides aldrin, dieldrin, endrin, isodrin, and chlordane. RMA Remediation Goal 2 involved the destruction of the six COCs to levels that met preliminary remediation goals (PRG). T able 1-1 summarizes the HHE criteria and PRGs for the six COCs. The standard that ISTD was to achieve, as expected by RMA, was 90 percent destruction and removal efficiency (DRE) for hex, dieldrin, and chlordane. Endrin, isodrin, and aldrin were reportedly below detection limits in pre-characterization sampling results, and therefore, RMA did not include them in the post-treatment DRE standard (TerraTherm 2001). The primary objective of the SITE demonstration of TerraTherm’s ISTD technology was to determine the ability of the technology to meet the HHE criteria for the six COC s. Additional discussion of the SIT E Program’s originally-planned primary and secondary objectives for this evaluation is included in Section 3.1.1 1.4 GENERAL TECHNOLOGY DESCRIPTION ISTD is a soil remediation process that applies heat and vacuum simultaneously to contaminated soils, either with surface heater blankets or with an array of vertical heater and vacuum extraction wells. Surface heater blankets are used for the removal of surficial contamination down to about 2 feet, while vertical well arrays are used to treat deeper contamination in subsurface soils. Heaters are operated at 1,450 to 1,650 degrees Fahrenheit (°F). According to the developer, as the soil is heated, volatile contaminants are vaporized or destroyed by a number of mechanisms, including the following: (1) evaporation into the vapor stream, (2) steam distillation into the vapor stream, (3) boiling, (4) oxidation, and (5) pyrolysis (Stegemeier and Vinegar 2001). The vaporized water, contaminants, and natural organic compounds are drawn in a direction counter-current to the heat flow to the vacuum source in the blankets or wells. Because the soil in the proximity of the heater-vacuum wells is heated to high temperatures (above 900 °F) for many days, the technology developer claimed that contaminants in the heated soil can be almost completely removed (Stegemeier and Vinegar 2001). Most of the contaminants are expected to be destroyed in the soil before the vapor stream is removed by vacuum extraction. For the Hex Pit site, the technology developer claims that this expectation was borne out in the results of the Hex Pit Treatability Study, in which the DREs for the site COCs within the treatability study samples exceeded 99 percent (ENSR Corporation [ENSR] 2000, see also Section 3.2.1). Contaminants that have not been destroyed in situ that remain in the vapor stream are destroyed by the off-gas treatment system. The technology developer claims that both thermal blankets and thermal wells have been highly effective in removing a variety of contaminants, including polychlorinated biphenyls (PCB), pesticides, chlorinated solvents, and heavy and light hydrocarbons (Stegemeier and Vinegar 2001). 1.5 KEY CONTACTS Additional information on the SITE Program, TerraTherm’s ISTD technology, and the demonstration site can be obtained from the following sources: The SITE Program Marta K. Richards and Scott Jacobs U.S. Environmental Protection Agency Office of Research and Development 26 West Martin Luther King Drive Cincinnati, Ohio 45268 Telephone: (513) 569-7692 and (513) 569-7635 Fax: (513) 569-7676 and (513) 569-7585 Email: richards.marta@epa.gov Email: jacobs.scott@epa.gov TerraTherm’s ISTD Technology Ralph BakerTerraTherm, Inc. 356 Broad Street Fitchburg, Massachusetts 01420 Telephone: (978) 343-0300 Fax: (978) 343-2727 Email: rbaker@terratherm.com RMA’s Hex Pit Site Lorri Harper Remediation Venture Office U.S. Fish and Wildlife Rocky Mountain Arsenal Building 111 Commerce City, Colorado 80022-1748 Telephone: (303) 289-0411 Fax: (303) 289-0485 Email: Harper@FWS.gov SECTION 2 TECHNOLOGY APPLICATION ANALYSIS This section describes the general applicability of TerraTherm’s ISTD technology to contaminated waste sites. Previous ISTD applications are described; and, since the technology treatment was not completed at this site, this section focuses primarily on descriptions of the Hex Pit site characteristics and the IST D system specifically designed for the site. 2.1 PREVIOUS APPLICATIONS OF IN SITU THERMAL DESTRUCTION The technology developer currently describes case studies of six completed thermal treatment projects using the ISTD technology on its internet website (www.terratherm.com). These case studies include four sites contaminated with PCB s, one chlorinated solvent site, and one petroleum hydrocarbon site. Of the four PCB sites, three included vertical wells installed to depths of 12 to 15 feet below ground surface (bgs), similar to the thermal treatment approach at the Hex Pit. Two of the sites used therm al blankets, in addition to the thermal wells, to treat near-surface contamination or stockpiled soil, one used only thermal wells, and one site used thermal blankets in a batch-treatment process on stockpiled soil. PCB concentrations ranged up to highs of 20,000 milligrams per kilogram (mg/kg) in soil treated in situ using vertical wells and were greater than 10,000 mg/kg in stockpiled soil treated using thermal blankets. One site was also contaminated with dioxins and furans up to a toxicity equivalent (TEQ) concentration of 3.2 parts per billion (ppb). TerraTherm reports that treatment at all four PCB sites achieved cleanup goals ranging from less than 1 mg/kg to 10 mg/kg PCBs. Dioxin and furan contamination at the one site was reduced to the TEQ cleanup goal of less than 1 ppb. The technology developer claims that soil contaminated with chlorinated solvents, including trichloroethene (TCE), tetrachloroethene (PCE), and 1,1-dichloroethene, were reportedly successfully remediated at one site using the ISTD technology. The site included two vertical well fields; one consisting of 15 wells installed to a depth of 12 feet bgs and the other consisting of 130 wells installed to depths of up to 19 feet bgs. For PCE, the pre-treatment concentrations were as high as 3,500 mg/kg, while those for TCE were as high as 79 mg/kg. For PCE, the post-treatment concentrations in all samples were less than 0.5 mg/kg, while concentrations of TCE were less than 0.02 mg/kg. ISTD technology was applied at one site contaminated with petroleum hydrocarbons, including gasoline, diesel-range organics, and benzene. Reportedly, approximately 200,000 pounds of hydrocarbons, including immiscible product, were successfully removed and treated during the 120day heating cycle. The TerraTherm web site also includes the description of a thermal treatment project at a former wood treatment site that is apparently ongoing. Soil at the site is described as contaminated with polyaromatic hydrocarbons, pentachlorophenol, and dioxins and furans. Thermal treatment will be conducted using vertical wells. 2.2 HEX PIT SITE CHARACTERISTICS This section describes the geologic and hydrogeologic setting and previous investigations completed at the Hex Pit site. Information describing the characteristics of the pit’s contents is then summarized. Previous investigations at the site include those completed by Morrison Knudson (MK) (MK 1989), ENSR (ENSR 1999), and EPA (Tetra Tech 2001). Descriptions of the characteristics of the waste material contained in the pit are summarized from these previous investigations, a bench-scale treatab ility study of the ISTD technology (ENSR 2000), and from the pre-treatment sampling and analysis completed as part of this technology demonstration. 2.2.1 Geologic and Hydrogeologic Settings The Hex Pit was excavated in alluvial material, predominantly silty sand. This alluvial material is approximately 25 feet thick in the immediately vicinity of the Hex Pit and appears to thicken to the north. The alluvium is underlain by Denver Formation bedrock. The Denver Formation consists of weathered clayey sandstone and sandy shale. The top of the Denver Formation in the area forms an apparent shallow paleochannel that generally trends northward. The local bedrock topography controls the northward thickening of the alluvium and influences the pattern of groundwater flow (MK 1989). Recently, the water-table surface has been about 13 to 14 feet bgs in the immediate vicinity of the Hex Pit (Tetra Tech 2001, measured during pre-treatment sampling). The depth to the water-table surface reportedly varies seasonally by about 3 feet and is at its lowest during the winter and highest in late spring (TerraTherm 2001). Regional groundwater flow is to the north-northeast at a gradient of about 0.008 feet per foot, or about 42 feet per mile (MK 1989). 2.2.2 Previous Investigations Previous field investigations have been completed at the and three existing monitoring wells located location of the Hex Pit. In 1989, MK completed an investigation to evaluate whether the Hex Pit was an active primary source of groundwater contamination in the South Plants area (MK 1989). In 1998, MK completed a preliminary investigation to evaluate the boundaries of the Hex Pit and to characterize its contents (MK 1998). In 1999, ENSR completed a more extensive evaluation of the boundaries of the Hex Pit and the characteristics of the contained waste material (ENSR 1999). The 1999 ENSR investigation also involved collection of samples of material disposed of in the Hex Pit were used for a bench-scale treatability study of the ISTD technology (ENSR 2000). On behalf of EPA, Tetra Tech completed a screening investigation in 2000 to further evaluate the boundaries of the Hex Pit, focusing primarily on the south end of the pit that was previously covered by the concrete foundation slab of Building 571B (Tetra Tech 2001). The screening investigation also involved collection of soil samples from just outside the boundaries of the Hex Pit to evaluate the potential migration of contaminants from the Hex Pit to native soils, and installation of piezometers to measure the water table elevation in the immediate vicinity of the Hex Pit. Finally, samples were collected and analyzed as part of this technology demonstration in 2001, further characterizing the contents of Hex Pit and contaminant concentrations in soil covering, adjacent to, and immediately below the pit before the ISTD system was constructed and operated (pre-treatment sampling and analysis). This section describes the objectives of, and activities completed as part of, these previous investigations and the pre-treatment sampling and analysis. Section 2.2.3 summarizes the characteristics of the Hex Pit based on the results of these previous investigations. Groundwater Impact Study (MK 1989) MK completed the following activities to evaluate whether the Hex Pit was an active primary source of groundwater contamination in the South Plants area (MK 1989): * Aerial photographs from 1948 to 1978 and a blueline sketch dated November 19, 1967 were examined to delineate the approximate boundaries of the site. * Five new groundwater monitoring wells were installed, one hydraulically upgradient and four downgradient of the Hex Pit. The nearest downgradient monitoring well was located approximately 60 feet from the Hex Pit. * Groundwater samples were colleted and analyzed from the five new monitoring wells and three existing monitoring wells located in the general area. Water-level elevation measurements were also obtained. The MK study concluded that two compounds associated with waste material in the Hex Pit, hexachlorobenzene and hexachlorobutadiene, may be migrating at relatively low concentrations from the Hex Pit into the alluvial groundwater. However, the study also concluded that the risk to human and non-human biotic receptors from groundwater emanating from the Hex Pit area was insignificant and that no long-term benefit would be gained by conducting an interim response action at the site. The study also established the direction of groundwater flow in the area of the Hex Pit (north-northeast). Preliminary Investigation (MK 1998) The preliminary investigation of the boundaries of the Hex Pit and characteristics of its contents included the following activites (MK 1998): * Geophysical surveys of the area, including an electromagnetic-conductivity survey to evaluate the dimensions of the pit, a metal-detector survey to evaluate the presence of metal objects, and direct-current measurements to evaluate the character of the waste material. * Drilling three paired soil borings (six total borings) to evaluate the boundaries of the pit. * Drilling three soil borings to collect waste samples from the pit for chemical, odor, and physical analyses. The results of the geophysical surveys and observations from drilling the three paired borings provided a preliminary indication of the dimensions of the Hex Pit. Metal objects, presumably buried drums, were detected within the boundary of the pit. The waste samples were found to contain 33 to 38 percent volatile material, 5 to 27 percent carbon, and 14 to 23 percent chlorine. Concentrations of hex ranged from 1.3 to 16 percent. Although the odor from the Hex Pit was judged offensive, it was determined to be unlikely to present any off-post odor problems. Characterization Study (ENSR 1999) The objectives of the ENSR Hex Pit characterization study were as follows (ENSR 1999): • Delineate the vertical and lateral extent of the planned ISTD treatment zone. • Characterize the chemical and physical nature of the material in the pit. • Collect samples of the material in the pit for use in a bench-scale treatability study. • Collect samples outside and beneath the pit to establish background levels of contaminants and physical properties of soil. • Locate and examine buried utilities in the vicinity of the pit. • Confirm the depth to groundwater at the pit. In addition, several former site workers were interviewed as part of the EN SR investigation regarding their recollection of activities at the Hex Pit. As part of the ENSR investigation, 51 soil borings were drilled within and around the perimeter of the Hex Pit to visually identify its lateral and vertical boundaries. Samples collected to characterize the contents of the Hex Pit included three composite samples obtained from the north, middle, and south portions of the pit, and one sample collected from beneath the concrete foundation that remained from Building 571B. Two composite samples were also collected for a bench-scale treatability study. The SITE Program witnessed the process of opening the collected soil cores and compositing the sub-samples into the Master and Waste Composite samples that were tested during the bench-scale treatability study. “Background” soil samples were collected from beneath the pit and at four locations just outside the boundaries of the pit. Treatability Study (ENSR 2000) A bench-scale treatability study of the ISTD technology was conducted on contaminated samples collected from the Hex Pit as part of the characterization study (ENSR 1999). Two composite samples were tested during the treatability study, including the “Mast Composite,” which was representative of the entire contents of the pit, and the “Waste Composite, which was representative of only visibly contaminated soil-and-waste material. The purpose of the treatability study was to evaluate whether the ISTD technology could achieve a 90 percent DRE for each of the site COCs. Additional objectives of the study included comparing post-treatment concentrations of the COCs to the site-specific clean-up goals established in the site ROD, and evaluating the off-gas stream produced for use in designing an emission control system. Results from analyses of the Master- and Waste-Composite samples before treatment are included in the summary of Hex Pit characteristics (Section 2.2.3). The results of the treatability study are summarized in Section 3.2.1. Screening Investigation (TetraTech 2001) EPA’s screening investigation included drilling 57 soil borings to evaluate the boundaries of the Hex Pit and to collect samples of native soil surrounding the pit to evaluate the potential lateral migration of contaminants. In addition, four piezometers were installed near the sides of the pit to measure he local water-table elevation. The screening investigation was completed between September and October 2000, immediately after most of the foundation of Building 571B was demolished and removed. During demolition of the concrete foundation slab of Building 571B, it was discovered that foundation structures under the slab were more extensive than had been previously estimated. These foundation structures included concrete footers and columns that extended to depths exceeding 16 feet bgs near the northwestern corner of the slab. In addition, deteriorating drums and other waste material were discovered beneath the northern half of the slab and extending an unknown distance to the west. Because of these observations, the screening investigation was modified from the outset to focus primarily on evaluating the lateral boundaries and vertical depth of waste material beneath the foundation of Building 571B. Technology Demonstration Pre-treatment Sampling an Analysis Samples were collected as part of this SITE demonstration to establish conditions existing at the Hex Pit before construction and operation of the ISTD treatment system. The “pre-treatment” samples were collected and analyzed as described in the SITE project quality assurance project plan (QAPP) (EPA 2001) in July 2001. Pre-treatment samples included composites of the materials disposed of in the Hex Pit (Hex Pit soil-and-waste material); soil above, below, and laterally contiguous to the disposal pit (contiguous soil); and groundwater from the four piezometers previously installed as part of the screening investigation. The pre-treatment sampling is further described in Section 3.1.2, and all pre-treatment sampling results are included in Section 3.2.3. 2.2.3 Summary of Hex Pit Characteristics The characteristics of the Hex Pit can be summarized based on the results of previous investigations and the pre-treatment sampling and analysis completed as part of this technology demonstration. Figure 2-1 shows the lateral boundaries of the Hex Pit. The main part of the Hex Pit measures approximately 94 feet long, 45 feet wide, and varies from 8 to 10 feet deep. A narrow trench extends west near the south end of the pit. A ramp is also evident at the south end of the Hex Pit where, presumably, a bulldozer or other heavy equipment entered the pit when it was originally excavated. The north end of the Hex Pit is also sloping, while the east and west sides and the sides of the trench extending west are nearly vertical. The total volume of material in the Hex Pit is estimated to be 2,005 cubic yards (TerraTherm 2001). Figure 2-2 shows a generalized stratigraphic column through the Hex Pit. As shown in Figure 2-2, materials logged in borings completed as part of the previous investigations can be divided into the following general categories: * Cover material * Soil-and-waste material * Mixed fill-and-waste material from removal of the foundation of Building 571B * Native soil The Hex Pit cover material is primarily composed of mixed sand, gravel, and silt that were placed as a cap over the entire area. The soil-and-waste material is composed of the material that was originally disposed of in the pit. It consists of soil (primarily silty sand) that is often stained dark brown, rust orange, or black, and may be mixed with granules or globules of hex. Black, tar-like relatively pure hex residue occurs in layers or bands usually less than 1 foot thick. Other substances include rusted metal fragments (probably drum remains), black to orange and occasionally white crystalline substances, layers of light bluish-gray paste-like material that is probably lime, and wood fragments. The layered nature of the soil-and-waste-material unit reflects the historical disposal practices; that is, hex disposed of in drums (that ruptured when dumped or later corroded) or in bulk that was then covered with soil backfill. It is also apparent that lime was occasionally dumped into the pit. The mixed fill-and-waste material from the removal of the foundation of Building 571B is from the demolition and removal of the building’s concrete foundation in September 2000. Foundation structures, including concrete footers and columns, were found to extend below the concrete slab, and attempts were made to excavate and remove these structures. Clean fill was used to cover the excavation at the end of each day to control odors from the Hex Pit waste material. The next morning, this fill material was dug out of the excavation so demolition and removal of the foundation structures could continue. Moving this material in and out of the excavation each day resulted in a mix of clean fill-and-waste material. The mixed fill and waste generally consists of silty sand with occasional gravel or concrete rubble fragments; streaks of granules of black, tar-like hex waste material; and trace amounts of rusted metal fragments. This material is restricted to the southern end of the Hex Pit beneath the location of the former building foundation Native soil beneath and adjacent to the pit consists of sand, silty sand, and silt, usually yellow-brown in color. The native soil may be stained rust orange to depths of several feet below the Hex Pit waste material. Occasionally, streaks of black hex staining also occur in native soil immediately beneath the pit. Samples of Hex Pit soil-and-waste material were analyzed as part of the characterization study (ENSR 1999) and the SITE pre-treatment sampling effort. The characterization study samples included three composite samples obtained from the northern, middle, and southern portions of the pit and one sample collected beneath the concrete foundation slab of Building 571B. Two composite samples were collected for the treatability study (ENSR 2000), including the “Master Composite, which was representative of the entire content of the pit, and the “Waste Composite, which was representative of only visibly contaminated soil-and-waste material. These samples were analyzed for volatile organic compounds (VOCs), total chlorine, and the Hex Pit COCs. The Master Composite sample was also analyzed for dioxins and furans. The SITE pre-treatment sampling effort included the collection of six composite samples analyzed for the site COCs, semivolatile organic compounds (SVOCs), and dioxins and furans. In addition, the SITE pre-treatment sampling included the collection of eight grab samples that were analyzed for VOCs. These samples were collected from depths of approximately 5 feet bgs, without regard to whether the material was primarily waste or soil backfill. Table 2-1 summarizes the concentrations of selected chemical constituents detected in these samples of soil-and-waste material disposed of in the Hex Pit. Samples of native soil (referred to as “contiguous soil”) were collected from beneath and adjacent to the Hex Pit as part of the characterization study (ENSR 1999), the screening investigation (Tetra Tech 2001), and the pre-treatment sampling effort. Many of the native soil samples collected beneath or very near the sides of the Hex Pit were visibly stained rust-orange or with streaks of black hex. Visibly contaminated contiguous soil samples often contained concentrations of the site’s COCs similar to the soil-and-waste material composite samples. Contamination did not appear to migrate more than a few feet laterally into contiguous soil as evidenced by the lack of hex detected in contiguous soil samples collected approximately 8.5 feet from the sides of the Hex Pit as part of the pre-treatment sampling effort (see also Section 3.2.3). Groundwater samples were analyzed as part of the screening investigation (Tetra Tech 2001) and pre-treatment sampling effort. Several VOCs, including chloroform, carbon tetrachloride, benzene, TCE, and PCE, were detected in these groundwater samples (Tetra Tech 2001), which are typical of a regional groundwater contaminant plume in the area (MK 1989). However, hex was not detected in these groundwater samples collected as near as approximately 13 feet downgradient of the Hex Pit boundaries. 2.3 IN SITU THERMAL DESTRUCTION SYSTEM DESIGN AT THE HEX PIT TerraTherm’s ISTD configuration at the Hex Pit was described in the Hex Pit Remediation Final (100%) Design Package (TerraTherm 2001). At this site, ISTD was designed to heat the soil above the boiling points of the COCs using a network of heater wells. The ISTD remediation design for the Hex Pit assumed that contamination extended 10 feet bgs. To attempt to ensure adequate heating and treatment of the contaminated soils within the delineated boundaries of the Hex Pit, the ISTD remediation design included heating the soil 5 feet laterally and 2 feet vertically beyond the delineated boundaries of the Hex Pit. This area encompassed a target treatment soil volume of 3,198 cy, extending from 0 to 12 feet bgs and 5 feet laterally beyond the boundaries of the Hex Pit. The ISTD heating duration was designed to be 85 days. Approximately one-quarter of the heater wells were configured as combined heater-and-vacuum extraction wells (HV wells) to allow collection of the volatilized vapors. The well-field layout consisted of a triangular grid of thermal wells spaced on 6-foot centers with a 3.75:1 ratio of heater-only to heater-vacuum wells. The grid resulted in a total of 266 wells, of which 210 were heater-only wells and 56 were HV wells. All well casings (and screens for the HV wells) were constructed of Type 304 stainless steel. Figure 2-3 shows the well-field layout for the ISTD system. According to the developer’s design, electrical heating elements placed in the wells were designed to reach temperatures of 1,400 to 1,600 °F, resulting in an extremely hot zone surrounding each heater well. The thermal well field was designed to achieve a minimum temperature of 617 °F between wells within the delineated boundary of the Hex Pit. A thermal heat front was to advance radially outward from the heater wells through thermal conduction. As contaminants were drawn through the extremely hot zone that surrounds the heater wells, the technology developer expected the majority of the contaminant mass to be destroyed by oxidation or pyrolysis. Thus, the majority of contaminant mass destruction was expected to occur in situ. Steam stripping of contaminants was also expected to occur as the soil pore water was boiled off during the initial heating phase. Soil along the boundaries of the treatment area were maintained under negative pressure to attempt to ensure that steam and volatilized contaminant vapors were captured and directed to the off-gas treatment system. A small vacuum (approximately 20 inches of water column) was expected to provide adequate capture of the vapors released during heating. Vapors extracted from the subsurface were treated aboveground. The aboveground piping network designed to transport vapors to the treatment system was constructed of Type 304 stainless steel, except for high-temperature reinforced flexible hose connecting vapor tees at the HV wellheads to the piping network. The off-gas treatment system was designed to treat the incoming process vapor stream from the ISTD wellfield to reduce concentrations of organic and inorganic contaminants, including acid gases. The off-gas treatment system consisted of a cyclone separator, flameless thermal oxidizer (FTO), heat exchanger, knock-out pot, two acid gas dry scrubbers, two carbon bed adsorbers, and two main process blowers. The main process blowers were induced draft fans. The fans were designed to supply the motive force (vacuum) needed to draw the vapors from the well field and through the off-gas treatment system. Figure 2-4 is a process flow diagram of the ISTD system. The cyclone separator was designed to remove particulates from the incoming vapor stream to prevent damage to, or clogging of, downstream off-gas treatment system equipment. The technology developer expected the quantity of particulates to be low at all times, but to increase with time as the soil dried out. The FTO was designed to convert organic constituents in the process stream to carbon dioxide and water vapor. Because a significant quantity of chlorinated organics was expected in the waste stream, hydrogen chloride (HCl) was expected to be produced during the oxidation process. Generation of the acid gas required a separate neutralization step before discharge to the atmosphere. The FTO was expected to operate at temperatures in the range of 1,500 to 1,900°F. A heat exchanger was incorporated to decrease the temperature of the hot process gases that exited the FTO before it entered the scrubber and carbon adsorbers. The high-efficiency air-to-air heat exchanger was designed to cool the hot process stream from 1,600 to 200°F with a residence time of less than 0.3 second. Following the heat exchanger, the knock-out pot was used to separate the liquid from the vapor. The vapor passed into a dry scrubber used to neutralize acid gases in the vapor stream. The vapor stream flowed through two packed beds of granular scrubbing media, which were expected to neutralize hydrochloric acid vapor. Two vapor-phase carbon adsorbers were installed downstream of the scrubber beds as a final polishing step to remove any remaining organic contaminants from the vapor stream. Contaminant mass loading on the adsorber was expected to be low because the technology developer expected that most of the contamination would be destroyed upstream of the carbon adsorbers. As a precaution, an emergency generator was provided and connected so that in the event of a loss of grid power, an automatic transfer switch would cause the generator to start within 30 seconds and continue to power the blowers and air quality control equipment throughout such an outage. SECTION 3. TREATMENT EFFECTIVENESS The following sections describe the methods by which the ISTD treatment technology was evaluated and the results of the evaluations. 3.1 DEMONSTRATION METHODOLOGY The following sections describe the SITE demonstration objectives, including the original demonstration objectives and how the objectives were modified after failure of the ISTD system, the SITE pre- and post-treatment sampling that was completed, and the data quality assessment of the analytical results. 3.1.1 SITE Demonstration Objectives Similar to other SITE demonstration projects, the ISTD demonstration at the Hex Pit included primary and secondary objectives designed to evaluate the ability of the technology to achieve specific clean-up criteria and to assess the cost and overall effectiveness of the treatment system. The primary objective planned for the demonstration, as described in the SITE project QAPP (EPA 2001), was as follows: * P1 To determine the ability of the TerraTherm ISTD remediation technology to meet RMA HHE cleanup criteria for COCs in soil-and-waste material within the Hex Pit boundaries. The COCs are hexachlorocyclopentadiene (hex), aldrin, dieldrin, endrin, isodrin, and chlordane. The HHE clean-up criteria are included in Table 1-1 in Section 1.3.2. Secondary objectives planned for the ISTD demonstration were the following: * S1 Determine the cost of treatment for contaminated soil-and-waste material in the RM A Hex Pit. * S2 Evaluate the effluent gas-phase emissions from the TerraTherm treatment process. * S3 Evaluate the DREs of the Hex Pit COCs and dioxins and furans by in situ thermal treatment and the off-gas treatment system (FTO, heat exchanger, dry scrubber, and carbon bed). * S4 Compare contaminants remaining in the site soil after treatment to the contaminants present before treatment. * S5 Evaluate changes in concentrations of hex in soil and groundwater outside the boundary of the treatment area. * S6 Determine the ability of TerraTherm’s ISTD technology to meet PRG clean-up criteria (shown in Table 1-1 in Section 1.3.2). These objectives formed the basis for the sampling design described in the SITE project QAPP (EPA 2001) to evaluate the ISTD treatment process. The SITE objectives were to be achieved by collecting and analyzing soil-and-waste samples in the northern portion of the Hex Pit beore and after the ISTD demonstration. Pre-treatment sampling was completed as described in the QAPP and is summarized in Section 3.1.2. However, as described in Section 4.0, the ISTD demonstration was terminated prematurely due to unexpected material failures. The average concentration of contaminants in post-treatment samples was considered unlikely to be much different from the average concentration of contaminants found in the pre-treatment samples. Consequently, the sampling strategy to achieve the demonstration objectives was no longer considered viable and was re-evaluated in the SITE post-treatment sampling and analysis plan (SAP) (EPA 2002). Consistent with TerraTherm’s Operations and Maintenance Manual, the heater-only wells were energized in stages. On the fifth day of heating, all heater-only wells in the southern third of the well field were energized; however, the heater-only wells in the northern two-thirds of the well field, which were scheduled to be energized around the time of the failure of the piping, were never turned on. Thus, heating within the northern portion was limited to the HV wells. The SITE post-treatment SAP considered that all HV wells were active for 12 days before system shutdown and may have produced discernable changes in contaminant concentrations in soil-and-waste material immediately adjacent to the wells. Thus, the objective of the post-treatment sampling was to characterize contaminant concentrations in soil-and-waste material in close proximity to the HV wells (approximately 0.5 feet) for comparison to pre-treatment soil-and-waste material contaminant concentrations. Section 3.1.3 summarizes the post-treatment sampling. 3.1.2 SITE Pre-treatment Sampling SITE pre-treatment sampling was completed as described in the SITE project QAPP (EPA 2001) to establish baseline conditions at the Hex Pit before construction and operation of the ISTD system. Sampling was confined to the northern half of the Hex Pit and was completed in July 2001. Sampling was confined to the northern half of the Hex Pit because the southern portion of the Hex Pit had been disturbed during the demolition and removal of the foundation of Building 571B, including the mixing of clean fill with material originally disposed of in the Hex Pit. Sampled materials included the soil-and-waste material originally disposed of in the pit; contiguous soil above, below, and laterally adjacent to the pit; and groundwater from piezometers flanking the pit. Table 3-1 summarizes the pre-treatment sampling, and Figures 3-1 and 32 show the sampling locations. The pre-treatment sampling results are summarized in Section 3.2.3. The soil-and-waste material unit was the focus of the ISTD treatment process. Six composite soil-and-waste material samples were collected for analysis. Each composite sample was created by mixing material from three soil cores collected from 2 to 10 feet bgs. Boreholes were drilled using direct-push techniques, and soil cores were obtained with dual-tube sampling equipment. Samples were composited by mixing core material in disposable aluminum pans with disposable plastic scoops. Nine grab samples were also collected for analysis of VOCs. These grab samples were collected from soil cores from 5 feet bgs before the core material was transferred to the aluminum pans for compositing. Figure 3-1 shows the cores that were combined to form the composite samples, and the cores that were used to collect the grab samples for V OC analysis. Three separate areas of contiguous soil were sampled: cover material above the Hex Pit soil-and-waste material unit (0 to 2 feet bgs); native soil below the soil-and-waste material unit (from two different depth intervals, including 10 to 12 feet bgs and 12 to 13 feet bgs); and native soil outside the perimeter of the Hex Pit. Three composite samples each were collected from the cover material and soil beneath the soil-and-waste material unit (from the two different depth intervals). Each composite sample was created by mixing material from six soil cores collected from the specified depth intervals. Nine grab samples were also collected for analysis of VOCs from a depth of 1 foot bgs in the cover material. Twelve native soil samples were collected outside the perimeter of the Hex Pit, approximately 3.5 feet beyond the boundary of the treatment area (8.5 feet beyond the edge of the Hex Pit). These soil samples were created by homogenizing core material collected from 2 to 10 feet bgs in boreholes drilled outside the Hex Pit. Figure 3-2 shows the cores that were combined to form the composite samples, the cores that were used to collect the grab samples for VOC analyses, and the outside perimeter borehole locations. Compositing and grab-sampling procedures for the contiguous soil samples were the same as procedures described for the soil-and-waste material samples. Groundwater samples were collected from four piezometers located about 28 feet from the edges of the Hex Pit in each major compass direction (north, south, east, and west). Figures 3-1 and 3-2 show the piezometer locations. 3.1.3 SITE Post-Treatment Sampling As described in Section 3.1.1, the ISTD demonstration was terminated prematurely due to unforseen material failures. Consequently, the post-treatment sampling strategy to achieve the demonstration objectives originally described in the SITE project QAPP (EPA 2001) was no longer considered viable. The SITE post-treatment sampling objectives and procedures were re-evaluated in the SITE post-treatment SAP (EPA 2002). The post-treatment sampling consisted of collecting six samples from the soil-and-waste material unit from close proximity (approximately 0.5 foot) to six ISTD HV wells. Table 3-2 summarizes the SITE post-treatment sampling, and Figure 3-3 shows the sampling locations. The post-treatment sampling results are summarized in Section 3.2.4. Following the failure of the ISTD system, the site was buried under approximately 3 feet of imported fill material. Since the southern portion of the site was lost to physical disturbance and was unavailable for sampling, the SITE post-treatment sampling was completed by first marking the presumed locations of buried HV wells in the northern half of the Hex Pit. Hand digging through the fill material was then conducted to find the tops of the HV wells. Once the tops were verified, offsets were measured to locate where an angled borehole would be started to collect cores from the soil-and-waste material unit adjacent to the HV well casing. The boreholes were angled to avoid steel plates welded to the well casings and to position the borehole approximately 0.5 foot from the HV well at depths of 2 to 10 feet below the original surface of the Hex Pit cover material. Figure 3-4 diagrams this approach to drilling the SITE post-treatment sampling boreholes. Similar to the SITE pre-treatment sampling effort, the boreholes were drilled by direct-push techniques, and core samples were collected using dual-tube sampling equipment. The samples were created by homogenizing core material collected from 2 to 10 feet below the original top of the Hex Pit cover material. Six grab samples were also collected for analysis of VOCs from a depth of 5 feet below the top of the soil-and-waste material unit. 3.1.4 SITE Data Quality SITE pre- and post-treatment laboratory analytical data were validated to confirm that the results were satisfactory for use in addressing the project objectives. Appendix A includes the validation reports for all SITE pre- and post-treatment laboratory analytical data generated for this project. The validation reports discuss the performance of the internal quality control (QC) checks conducted by the laboratory during the sample analyses, such as results for matrix spike/matrix spike duplicate (M S/M SD) samples and surrogate spikes. In addition to the internal QC checks, field replicate samples were collected during the treatment demonstration as external (field) QC samples. These co-located samples included one triplicate sample of contiguous soil and another of soil-and-waste material collected during the pre-treatment sampling, and one duplicate soil-and-waste material sample collected during the post-treatment sampling. Overall, the findings of the QC checks and data validation indicated that the sample analyses were acceptable as qualified; no results were considered unusable. All validation qualifiers are listed with the analytical results summarized in the validation reports in Appendix A. As described in the validation reports, the analyses rendered an expected level of data quality, given the nature of the analytical methods and the samples. The analytical methods were designed to identify and quantitate low concentrations of organic compounds in relatively uncontaminated soil matrices. However, many of the samples contained relatively high concentrations of many organic compounds. This complexity produced many failures of QC measures, such as matrix interferences manifested in irregular MS/M SD results, surrogate recoveries, and internal standard results. In other instances, QC data were lost entirely due to the high dilutions required for many samples prior to analysis. Required dilutions produced very high quantitation limits for many analytes and samples. In general, the high concentrations and complex sample matrices increase the potential for false positives in the data sets and give the quantitative results an "estimated" character. Although the complex nature of the samples remained consistent between the SITE pre- and post-treatment samples, the comparability of the two data sets is limited by the different sampling approaches applied for the pre- and post-treatment events. The utility of the data sets for assessing the effects of the ISTD treatment process is further limited by the inherent heterogeneity of the soil-and-waste material in the treatment zone. The comparability of the two sampling events is further discussed in Section 3.2.5. The SITE pre- and post-treatment analytical data were compared to precision, accuracy, representativeness, completeness, and comparability (PARCC) objectives outlined in the project QAPP (EPA 2001). The following sections summarize the evaluation of the PARCC objectives. Precision Precision is a measure of the reproducibility of an experimental value without regard to a true or referenced value. The primary indicators of precision were the relative percent difference (RPD) results for the MS/MSD analyses, the RPD between the field duplicate pair collected during the post-treatment sampling, and the percent relative standard deviation (%RSD) between the three replicate field samples collected during the pre-treatment sampling. The RPD and %RSD values for the duplicate and replicate samples are shown in Table 3-3. The inherent heterogeneity of soil samples often result in high RPD and %RSD values in duplicate and replicate analyses. This heterogeneity is apparent in some of the field replicate results shown in Table 3-3, particularly in the high RPDs calculated for the VOCs in the post-treatment duplicate. Due the high concentration of analytes, the MS/MSD spiking amounts were diluted out for many samples and could not be used to evaluate the level of precision. Overall, however, acceptable precision was found for the pre- and post-treatment analytical results for the field, given the high analyte concentrations and complex matrices in the samples analyzed. Accuracy Accuracy assesses the proximity of an experimental value to a true or referenced value. The primary indicators of accuracy are compound recoveries in surrogate, MS, and laboratory control sample (LCS) analyses. Accuracy is expressed as percent recovery. Due to the high concentration of analytes in the samples, the MS spiking amount was often diluted out and could not be used to evaluate accuracy. Having only partial data to evaluate the overall accuracy, leads to an inconclusive judgement. Though the surrogate and LCS recoveries were adequate, the overall accuracy of these data could not be determined. Representativeness Representativeness refers to the ability of data to reflect true environmental conditions. Results were evaluated for representativeness by examining items that were related to the collection of the samples, such as the chain-of-custody documentation, which included accurate sample labeling, recording correct sample collection dates, and confirming the condition of the samples when they were received at the laboratory. Laboratory procedures were also examined, including anomalies reported by the laboratory either when the samples were received or during the analytical process, including evaluating sample holding times, appropriate calibration of laboratory instruments, adherence to analytical methods, appropriate quantitation limits, and the completeness of the data package documentation. Items not meeting the criteria are documented in the validation reports. Overall, acceptable representativeness was found for the pre- and post-treatment analytical results. Completeness Completeness is defined as the percentage of measurements that are considered valid. The validity of the analytical results is assessed through the data validation process. All results that are rejected and any missing values are considered incomplete. Data that are qualified as estimated or nondetected are considered valid. Completeness is measured by comparing the total number of samples planned in the QAPP to the total number of samples collected, and the total number valid results compared to the total number of analytical results. Analytical completeness is measured by dividing the total number of valid results by the total number of results and multiplying by 100. Each analyte from each method is multiplied by the number of samples analyzed to calculate the total number of results. As no data were rejected and all data were collected and analyzed as specified in the SITE project QAPP (EPA 2001) and post-treatment SAP (EPA 2002), completeness for this investigation was 100 percent. Comparability Comparability is a qualitative parameter that expresses the confidence with which one data set may be com pared to another. Comparability of data is achieved by the use of uniform sampling procedures, standard methods of analysis, standard quantitation limits, and standardized data validation procedures. The use of approved laboratories, specified and well-documented analyses, and standard processes of data review and validation give the pre- and post-treatment data sets a high degree of analytical comparability. However, as discussed in Section 3.5.2, the need to modify the post-treatment sample collection and preparation procedures relative to those procedures used to obtain the pre-treatment samples renders accurate comparability of the data sets somewhat questionable. 3.2 DEMONSTRATION RESULTS The following sections summarize evaluations of the ISTD system at the RMA Hex Pit. Pre-construction evaluations are summarized that were not completed by EPA’s SITE Program, but by the technology developer, to estimate the performance of the system to assist in the design process. A brief chronology of system operations at the Hex Pit is presented as well as SIT E’s pre- and post-treatment sampling results. Finally, a comparison of the SITE pre- and post-treatment sampling results is presented. 3.2.1 Pre-construction Evaluations Pre-construction evaluations completed by the technology developer included a treatability study of the effectiveness of thermal treatment on representative contaminated soil-and-waste samples from the Hex Pit, computer simulation modeling to optimize the subsurface thermal and vapor flow operating parameters, and field testing of the IST D well design at a separate test site. The results of these evaluations are summarized below. Treatability Study A bench-scale treatability study of the ISTD technology was conducted on contaminated samples collected from the H ex Pit (ENSR 2000). The treatability study samples were collected during the characterization study (ENSR 1999) and included the Master Composite and the W aste Composite. Table 2-1 includes a summary of contaminant concentrations detected in the Master and Waste Composite samples before treatment. The purpose of the treatability study was to evaluate whether the ISTD technology could achieve a 90 percent DRE for each of the site COCs. Additional objectives of the study included comparing post-treatment concentrations of the site COCs to the site-specific clean-up goals, and evaluating the off-gas stream produced for use in designing an emission control system. In the treatability study, the test samples were thermally treated at a target temperature range of approximately 1,000 to 1,900 °F under controlled vapor flow conditions to simulate treatment of the Hex Pit material by the ISTD process. After treatment, the test samples were recovered and analyzed for residual contaminant concentrations. The post-treatment sampling results indicated that DREs of 99 percent were achieved for the site COCs and that the site cleanup goals could be met. Dioxin and furan concentrations were reduced by more than 90 percent, and test results indicated that dioxins and furans were not created by the thermal treatment process. Evaluation of off-gas emissions from the test indicated that a significant quantity of HCl vapor or chlorine gas was emitted during thermal treatment. However, it was postulated by the developer that actual field emission rates would be lower because of the buffering capacity of the soils in the Hex Pit. Simulation Modeling The developer conducted simulation modeling as part of the ISTD system design effort to evaluate optimal subsurface thermal and vapor flow operating parameters. The simulation modeling report was included as Appendix I to the Final 100 Percent Design Package (TerraTherm 2001). Simulations were conducted using a three-dimensional, multiphase flow, multicomponent, non-isothermal model to evaluate the following: * The optimal placement of H V and heater-only wells and the required electrical load per heater. * The expected time-course and duration of heating to achieve the target temperature throughout the treatment zone. * The extraction vacuum and flow rate required to accommodate the predicted water vapor and emissions generation rates. * The length of time required after heating for soil to cool to ambient temperatures. The simulation results indicated that a ratio of 3 to 1 heater-only to HV wells set at a 6-foot inter-well spacing was optimal to achieve the site clean-up goals in a relatively short period of time. An ed ge-well 3:1 triangular well placement pattern best ensured the capture of volatilized contaminants. Simulation results also indicated that soil temperatures in portions of the treated area may remain in the range of 450 to 500°F for up to 120 days after heating, and may remain as hot as 300°F for up to 180 days after heating ceases. Field Testing Field testing of ISTD wells at a location in Houston, Texas was completed as part of the 95 percent design effort. The field testing report was included as Appendix J to the Final 100 Percent Design Package (TerraTherm 2001). The purpose of the field test was to evaluate a new generation of HV and heater-only wells for use at the H ex Pit site. ISTD wells used during previous applications were relatively complex in design and expensive to construct. Field testing identified a new well design that could result in substantial cost savings for the Hex Pit project by using materials that were readily available and that could be routinely fabricated. Problem-free performance over the course of a 63-day field trial resulted in the new well design being incorporated into the ISTD system at the Hex Pit. 3.2.2 Chronology of System Operation at the Hex Pit The following is a summary of the chronology for the ISTD system operation at the Hex Pit (adapted from TerraTherm 2002 and FWENC 2002): * October 4, 2001 - The technology developer (TerraTherm) mobilizes to the Hex Pit site. * October 9, 2001 through February 18, 2002 - Construction of the ISTD system at the Hex Pit. Activities include site preparation, installation of wells, placement of the surface cover and aboveground piping network, installation of the electrical system, and assembly of the off-gas treatment system. In addition, RVO installed three horizontal wells under the Hex Pit as a contingency for dewatering should the water table surface rise to a level that would be detrimental to operation of the ISTD system. * February 19 through March 2, 2002 - System shakedown testing and checking, and preheating of the piping network and FTO. * March 3, 2002 - Start of ISTD heating operation. All 56 HV wells were energized and vapors were drawn from the wellfield. * March 5, 2002 - 84 heater-only wells were energized in the southern third of the wellfield. * March 11, 2002 - Liquid observed collecting in flexible hoses connecting the HV wells to the aboveground piping network. * March 11, 2002 - Sagging noticed in aboveground piping at the southern end of the well field and a faint odor noticed from the wellfield. * March 14, 2002 - Two manifold pipe taps in the aboveground piping network observed to be leaning, closer inspection concluded that tap welds had corroded. During investigation of the damage, a seal on an HV well was damaged and steam leaked out at the base plate. * March 15, 2002 - Steam and strong odors emitted from an HV well. Loss of vacuum pressure noticed in southern end of wellfield. Several heaters experience electrical shorting, including an insertion heater in the aboveground piping network and a down-hole heater in an HV well. Power to the wellfield heaters was shut down. The piping network insertion heaters and off-gas treatment system continue to operate. * March 17, 2002 - All wellfield manifold valves were closed and the off-gas treatment system and insertion heaters were shut down. Soil temperatures were variable in the northern portion of the Hex Pit (location of EPA SITE’s pre- and post-treatment sampling efforts) during the 12-day heating period. By heating day 5, thermocouples located 1 foot from HV well HVD16, located in the row immediately north of the southern third of the well field, reached temperatures of approximately 70°F, 120 to 170°F, and 250°F at near the ground surface, the 4- to 7-foot-deep, and the 10-foot-deep locations, respectively. By heating day 12, temperatures were 120°F near the ground surface, just over 200°F at 4 to 7 feet, and 416°F at 10 feet. Farther north in the wellfield, temperatures within 1 foot of HVP8 at heating day 5 were 200 to 220°F, except at a depth of 4 feet, where the temperature was approximately 125°F. By heating day 12, temperatures at that location reached a maximum of 237, 237, 398, and 458°F at depths of 1, 4, 7, and 10 feet, respectively. Soil temperatures measured by thermocouples installed in the far northern end of the pit were still below 100°F after 12 days of heating. Following shutdown of the wellfield heaters, soil temperatures in the vicinity of the operating HV wells in the northern half of the pit generally dropped 50 to 100°F or more within 1 week of shutdown. 3.2.3 SITE Pre-Treatment Sampling Results As described in Section 3.1.2, SITE pre-treatment samples were collected of soil-and-waste material originally disposed of in the pit; contiguous soil above, below, and laterally adjacent to the pit; and groundwater from piezometers flanking the pit. Table 3-1 summarizes the SITE pre-treatment sampling completed, and Figures 3-1 and 3-2 show the sampling locations. All SITE pre-treatment sample analytical results are included in the validation summary reports in Appendix A. All results are included in the validation summary reports, even though only analytical results from the soil-and-waste material samples are necessary to address the project objectives that were modified after failure of the ISTD system. For completeness, Appendix B includes all borehole logs completed as part of the pre-treatment sampling event. As expected from previous investigations, the soil-and-waste material unit consisted primarily of soil (primarily silty sand) layered with waste material. The soil was often stained dark brown, rust orange, or black, and often contained granules of probable hex. Tar-like, relatively pure hex waste material often occurred as bands or layers, usually less than 1 foot thick. Other substances observed in the soil-and-waste material unit included rusted metal fragments (probably from corroded drums), black to orange and occasionally white crystalline substances, layers of a light bluish-gray paste-like material that was probably lime, and wood fragments. The SITE pretreatment soil-and-waste material samples were composited from core samples collected from 2 to 10 feet bgs. In general, most of the tar-like hex waste material occurred between depths of 4 to 7 feet bgs. Soil from 7 to 10 feet bgs was often stained with small amounts of contamination. In general, a distinct contact between soil-and-waste material disposed of in the pit and native soil was difficult to determine. Table 2-1 in Section 2.2.3 includes selected analytical results from SITE pretreatment sampling of the soil-and-waste-material unit. Contiguous soil above the soil-and-waste material unit generally consisted of a surficial cover, often about 1 foot thick, consisting primarily of silty sand and gravel. SITE pretreatment samples were collected from 0 to 2 feet bgs and often the lower half of this interval included the silty sand material characteristic of the soil-and-waste material unit, often containing minor amounts of probable hex granules. Contiguous soil beneath the soil-and-waste material unit was collected from two intervals: 10 to 12 feet bgs and 12 to 13 feet bgs. Although the base of the Hex Pit was often difficult to accurately determine, it appeared that soil below 10 feet bgs was probably in-place native soil. Minor contaminant staining, including streaks of black hex, was occasionally observed in the native soil beneath the Hex Pit. Contiguous soil was also sampled adjacent to the Hex Pit. These soil samples all appeared as uncontaminated native soil. Samples of the laterally contiguous soil were only analyzed for hex concentrations, and no hex was detected in these samples. 3.2.4 SITE Post-Treatment Sampling Results As described in Section 3.1.3, the SITE post-treatment sampling boreholes were drilled through a soil cover that was placed over the site following failure of the ISTD system. Core samples were examined to determine when the borehole had reached the surface of the soil-and-waste material unit. Once into the soil-and-waste material unit, core samples were collected and prepared for laboratory analysis. The SITE post-treatment samples were created by homogenizing core material from single boreholes drilled through the soil-and-waste material unit. The SITE post-treatment sampling procedure was different from the SITE pre-treatment sampling procedure, which composited core material from three separate boreholes for each soil-and-waste material sample. In general, the post-treatment core samples from the soil-and-waste material unit appeared similar to the pre-treatment cores. That is, the unit did not appear to have undergone a significant change in physical characteristics as a result of the relatively short-term operation of the HV wells. All SITE post-treatment sample analytical results are included in the validation summary reports in Appendix A. Appendix B includes all borehole logs completed as part of the post-treatment sampling event. 3.2.5 Comparison of SITE Pre-and Post-Treatment Sampling Results The objective of collecting the SITE post-treatment samples was to evaluate if contaminant concentrations in the soil-and-waste material in close proximity to the HV wells were appreciably different from concentrations detected in the SITE pre-treatment samples. Table 3-4 lists the concentrations of selected compounds detected in SITE pre- and post-treatment samples collected from the soil-and-waste material unit. The selected compounds shown in Table 3-4 were consistently detected in historical and SITE pre-treatment samples and include the site COCs hex, aldrin, and dieldrin; VOCs carbon tetrachloride, chloroform, and PCE; and total TEQs calculated for dioxins and furans. The comparison between contaminant concentrations detected in the SITE pre- and post-treatment samples is intended to evaluate whether any contaminant destruction or removal took place during the brief operation of the ISTD system. Two different evaluations are presented, including a qualitative comparison and a statistical comparison conducted according to procedures specified in the SITE posttreatment SAP (EPA 2002). Qualitative Comparison of SITE Pre- and Post-Treatment Sampling Results The following sections describe a qualitative comparison of SITE pre- and post-treatment sampling data for the site COCs, VOCs, and dioxin and furan TEQs. Various plots were generated to evaluate the data including frequency plots, normal probability plots, box-and-whisker plots, and scatter plots (Figures 3-5 through 3-9). The frequency plots are similar to histograms and show the number of observations (y-axis) per concentration grouping (x-axis) for the pre-treatment and post-treatment samples. The scatter plot simply shows the concentration (y-axis) of the chemical in each sample (x-axis). The box and whisker plots show the median concentration (50th percentile) as the small square, the interquartile range (25th to 75th percentile) as the larger rectangular box, and the whiskers extending out to the minimum and maximum concentrations. The symmetry (or lack thereof) of the box and whiskers around the median reflects the data distribution (that is, normal or skewed). Finally, the normal probability plots show the concentration of each chemical in each sample in a manner that also shows how well the data set fits a normal distribution. Specifically, a probability plot is a graph of values, ordered from lowest to highest and plotted against a standard normal distribution function. The horizontal axis is scaled in units of concentration and the vertical axis is scaled in units of the normal distribution function (normal quantile). T he straight line on the probability plots shows the normal distribution, which is a theoretical probability distribution that is symmetric and has other specific attributes (Gilbert 1987). Site COCs Evaluations for the selected site COCs assessed the range, variability, and distribution of SITE pre- and post-treatment sampling data, and compared results from the two sampling events. A review of the box-and-whisker plot in Figure 3-5 suggests that hex concentrations may have decreased from the SITE pre- to post-treatment sampling events. The same trend is evident for aldrin and dieldrin (Figures 3-6 and 3-7), although the evaluation is complicated by the number of non-detected results in the SITE post-treatment data set. Comparison of the SITE pre- and post-treatment data sets, however, must take into account differences in the way samples were collected during the two events. As described in Section 3.1.1, each pre-treatment sample was obtained by compositing soil-and-waste material from three separate boreholes. For the post-treatment samples, however, core material was not composited from multiple boreholes; instead, samples were collected from single boreholes (see Section 3.1.2). Post-treatment samples from several boreholes contained relatively low concentrations of the site COCs, including samples HVH8, HVJ6, and HVL4 (Table 3-4). A review of the borehole logs (Appendix A) indicates that layers or bands of relatively pure, tar-like hex were not observed in these borings. Relatively thick layers of probable lime material (approximately 3.5 feet thick) were observed through the sampled intervals in borings HVH8 and HVJ6. The high pH values (12) measured in these samples supports the observation of probable lime material in the borehole logs (see sample analytical results summarized in Appendix A). The relatively low concentrations of COCs in samples from these borings may or may not be representative of typical concentrations remaining in the Hex Pit. VOCs Grab samples were collected for analysis of VOC concentrations from predetermined depths during both the SITE pre- and post-treatment sampling events. These samples were collected without regard to sample matrix and may have been obtained from relatively uncontaminated soil or highly contaminated waste material. Figure 3-8 presents an evaluation of analytical results for PCE, which is representative of trends observed for VOCs frequently detected in the pre- and post-treatment soil-and-waste material samples. The box plot shown in Figure 3-8 illustrates the relatively broad range of PCE concentrations detected in both the pre- and post-treatment samples. The broad range of PCE concentrations detected probably results from the different sample matrices collected. The box-and-whisker plot shown in Figure 3-8 suggests a slight decrease in PCE concentrations from the pre- to post-treatment sampling events. However, the wide scatter in both the pre- and post-treatment data sets complicates any comparison. Presumably, VOCs should have been quickly volatilized and removed had the ISTD system reached the intended operating temperatures. As described in Section 3.2.3, the chronology of system operation, soil temperatures measured near HV wells in the northern part of the Hex Pit did not reach the minimum treatment temperatures designed for the system. Dioxin and Furan TEQs Figure 3-9 presents an evaluation of analytical results for total TEQs calculated for dioxins and furans. A review of the box-and-whisker plot in Figure 3-9 suggests that TEQ concentrations may have increased slightly from the SITE preto post-treatment sampling events. However, the wide scatter of TEQ concentrations in the post-treatment data set suggests that a meaningful comparison with the pre-treatment data set may not be possible. In addition, soil temperatures measured near HV wells did not reach minimum treatment temperatures Statistical Comparison of SITE Pre- and Post-Treatment Sampling Results The SITE post-treatment SAP specified two types of statistical tests to compare the SITE pre- and post-treatment sampling results (EPA 2002). The following sections describe these statistical tests, test assumptions (hence, applicability to the data collected), and the results of the comparison of SITE pre-and post-treatment sampling results. Three representative compounds were selected for the statistical comparison, including hex, PCE, and TEQs for dioxins and furans. Hex was selected as a representative compound because in was the site COC detected in greatest concentration in the SITE pretreatment samples. PCE, although not a site COC, was selected to evaluate whether brief operation of the thermal treatment system had any affect on a volatile compound. Dioxin and furan TEQs were evaluated to assess potential creation of these compounds from operation of the thermal treatment process. Summary statistics for these selected compounds are presented in Table 3-5. Method 1: Linearized Ratios Method 1 evaluated the SITE pre- and post-treatment means for contaminant concentrations using a linearized ratio test and a null hypothesis of a 50 percent reduction in contaminant concentrations; that is, the null hypothesis stated that a 50 percent reduction in contaminant concentrations occurred between the SITE pre- and post-treatment sampling results. The test was to be applied to data for the three representative compounds discussed in the qualitative comparison (hex, PCE, and TEQs for dioxins and furans); however, one of the fundamental that approximately equal variance – was violated. Another test assumption – that data sets be normally or log-normally distributed – could not be quantitatively evaluated, but qualitative review of the data suggests that this assumption was also violated in some cases. result of these violations, the linearized ratio test was not performed. The second statistical test described in the work plan Wilcoxon Signed Rank Test) is parametric test (that is, the test does not assume data are normally or log-normally distributed). Results from this non-parametric test (Method 2) are discussed in the following paragraphs. Method 2: Bootstrapping and the Wilcoxon Signed Rank Test A second statistical method to evaluate the data used the "bootstrap" method to provide a better estimate of the SITE pre-treatment mean concentrations for the three representative compounds. “Bootstrapping” is a tool that uses random re-sampling of the original data sets, then provides an estimate of the mean for (in this case) 1,000 samples instead of the eight or nine samples that composed the original data sets. Bootstrapping or resampling methods take the combined samples as a representation of the population from which the data came, and create 1,000 or more bootstrapped samples. The bootstrapping process was applied 10 times (10 iterations) to produce 10 different estimates of the mean for pre-treatment concentration of the three representative compounds (Tables 3-6 through 3-8). The Wilcoxon Signed Rank test (a non-parametric one-sample test) was used to compare the SITE post-treatment data to each of the 10 bootstrapped estimates of the SITE pre-treatment mean concentrations of the representative compounds (hex, PCE, and TEQs for dioxins and furans). The SITE post-treatment SAP specified a null hypothesis stating that a 50 percent reduction in contaminant concentrations was not achieved (EPA 2002). That is, the null hypothesis stated that the post-treatment concentration of a compound was greater than the threshold value. The threshold value in this case, was one-half of each of the 10 bootstrapped pre-treatment mean concentrations. To conduct the Wilcoxon Signed Rank test, the SITE post-treatment data were compared with each iteration value of the SITE pre-treatment bootstrapped mean, then the absolute values of the difference between the estimated mean and the post-treatment data were assigned a rank based on their magnitude. After the results were ranked, then the rank values were assigned the appropriate sign (negative or positive value) and the positive values of rank were summed. If the sum was greater than the critical value (from a lookup table), which is based on sample size and the specified confidence (95 percent in this case), then the null hypothesis was rejected. In all cases, there was a failure to reject the null hypothesis; thereby indicating that the post-treatment data could not be shown to indicate a 50 percent reduction in contaminant concentrations. In these tests, however, failure to reject the null hypothesis was due to extreme variability in sample concentrations and too few samples to adequately characterize post-treatment conditions. These two factors resulted in poor power of the statistical test to reject the null hypothesis. Results of the Wilcoxon Signed Rank test for the three representative compounds are summarized in Tables 3-6 through 3-8. With regard to assumptions, the Wilcoxon Signed Rank test assumes the data constitute a random sample from a symmetric continuous population. The statistical plots (Figures 3-5 through 3-7) show that the data for hex and dioxins and furans (as TEQs) are roughly symmetrical; however the data for PCE are not symmetrical, which violates this test assumption. Nonetheless, the results from the Wilcoxon Signed Rank test offer information to be evaluated in the context of other evidence. Summary of Statistical Test Results A statistical hypothesis is a statement that may be supported or rejected based on relevant data. In statistical hypothesis testing, the “burden of proof” rests on the alternative hypothesis, which is the logical opposite of the null hypothesis. When testing a statistical hypothesis, two types of errors may occur; these are termed Type I error (false rejection of the null hypothesis) and Type II error (false acceptance of the null hypothesis). The Type I error is specified by the confidence level; for example, a 95-percent confidence level means there is a 5 percent probability of making a Type I error. The probability of making a Type II error is related to the “power” of the test. Power can simply be defined as “the probability of rejecting the null hypothesis when it is indeed false.” Poor power means that the probability of correctly rejecting the null hypothesis is low. For the statistical test (Wilcoxon Signed Rank test) used on the SITE pre- and post-treatment data, a confidence level of 95 percent was specified. The null hypothesis stated that the contaminant concentrations were not reduced by 50 percent. Results of the W ilcoxon Signed Rank T est indicate that, in every case, there was a failure to reject the null hypothesis. In other words, results of the statistical test do not indicate that contaminant concentrations were reduced by 50 percent. Results for the Wilcoxon Signed Rank T est may appear to contradict what is visible in the data plots (see Figures 3-5 through 3-9), until one reviews the summary statistics. The table of summary statistics (Table 3-5) shows extremely large variability (quantified as the standard deviation and variance) in contaminant concentrations. In six out of ten cases, the standard deviation was larger than the mean. The consequence of this variability is that any statistical test will have poor power to reject the null hypothesis. The power of a statistical test can be checked to determine if an adequate number of samples were collected to achieve a specified level of confidence (here, 95 percent). When the power of the tests is examined, for all data sets, the power of the test to reject the null hypothesis using data from the seven post-treatment samples, was poor in all cases. In the case of the data examined here, poor power to resolve differences and reject the null hypothesis is a consequence of examining populations with high variance for which there are too few samples. Generally, the desired performance for a statistical test is spelled out in project data quality objectives and includes the selection of a minimum detectable difference, which is the width of the gray region on a test performance plot, the confidence level, and the power desired. The number of samples required can then be estimated using existing information on population variance. Because information on population variance was not available for this SITE demonstration, the number of samples collected was not based on existing data. As a result, the extreme variance (standard deviation approximately equal to or greater than the mean value in many cases, see Table 3-5) translated into poor power and poor performance for the statistical tests to reject the null hypothesis. Due to the extreme variance in contaminant concentrations, there are insufficient data to statistically determine whether or not contaminant concentrations were reduced by 50 percent or more of their pre-treatment concentrations during this SITE demonstration. In summary, the results of the statistical tests are inconclusive. SECTION 4. TECHNOLOGY STATUS The following sections describe the physical destruction of ISTD system components, and summarize the results of investigations conducted to determine the cause of the component destruction. 4.1 DESTRUCTION OF SYSTEM COMPONENTS Thermal treatment at the Hex Pit was terminated 12 days after startup of all the HV wells and 10 days after startup of heater-only wells along the southern one-third of the well field. Electrical power to the well-field heaters was shut down after corrosion that resulted in structural and containment failure of segments of the aboveground stainless steel piping network was observed and heaters began shorting, including an insertion heater in the aboveground piping and a down-hole heater in one of the HV wells. All insertion heaters and the off-gas treatment system were shut down three days later. Evaluation of damage to the ISTD system focused on several areas as described below, including the aboveground piping network and insertion heaters, the down-hole heater cans and well screens in the HV wells, and the off-gas treatment system components. This discussion is summarized from TerraTherm (2002), except where referenced otherwise. 4.1.1 Aboveground Piping Network and Insertion Heaters Initial visual observations of disassembled portions of the aboveground piping network indicated significant corrosion of the pipe interior in the immediate vicinity (within 1 to 4 inches) of corroded manifold pipe taps. (The manifold pipe taps were short pieces of vertical piping that connected flexible hoses from tee fittings at the HV wellheads to the aboveground piping network. Observations of leaning pipe taps caused by disintegration of the stainless steel were initial indications of corrosion problems with the ISTD system.) Vendor-acquired metallurgical evaluation of the corroded piping indicated that several forms of corrosion had occurred, including stress corrosion cracking and intergranular corrosion or end grain attack (Colorado M etallurgical Services [CMS] 2002). No other visual evidence of significant corrosion and only minor heat discoloration or rust-colored staining in areas was noted throughout the rest of the aboveground piping network. However, metallurgical laboratory evaluation of selected sections of piping reported that general corrosive attack was evidenced by a reduction in wall thickness from the initial 0.125 inch to 0.108 inch, considered a high rate of metal loss (CMS 2002). The flexible, high-temperature rubber hoses that connected tee fittings at the HV wellheads to the manifold pipe taps were also disassembled and evaluated. During operation of the ISTD system, these hoses trapped liquids that prevented the vacuum from pulling vapors into the off-gas treatment system (Versaw 2003). TerraTherm operators attempted to drain the hoses periodically during system operation to prevent the blockage. A majority of the tee fittings and hose end connections were observed to be encrusted with materials and in some cases were completely blocked. The deposits ranged from crystalline or fibrous to tarry, muddy, powdery, or cake-like material. Chemical analysis of these precipitates indicated that they included metallic salts and both amorphous and crystalline organic materials containing high concentrations of hex. The flexible hoses did not appear to be corroded. One of the insertion heaters near the location of a failed manifold pipe tap that experienced an electrical short was removed and evaluated. The insertion heaters were contained in sections of stainless steel pipe or “cans” designed to protect the heater elements. The heater can reportedly showed some heat discoloration and visible pitting in one area, and was substantially unaffected in other areas. The insertion heater can was pressure tested and appeared tight. The electrical failure appeared to be from the melting of a thin-gauge wire and was claimed not to be related to the corrosion observed at the failed manifold pipe tap. 4.1.2 Heater Cans and Well Screens Damage to heater cans and well screens in the HV wells was evaluated by visual inspection following removal of the heater cans, down-hole video camera inspection, and metallurgical laboratory analysis. During removal of the heater cans, several wells were corroded to the extent that the cans broke off below ground surface. Heater cans remained stuck in several other wells and at five locations, the entire units including the well screen were pulled from the ground when attempting to remove the heater cans. The well screens were observed to be severely corroded and some sections of well screen were completely corroded away. One well was completely corroded through the screen and into the heater can, and hex material was observed to have accumulated in the heater can to a depth of 6 to 7 feet bgs (approximately 5 to 6 feet of hex had accumulated in the heater can). Video camera inspection revealed that hex material could be seen on, and coming through, the screen slots in several wells. In some wells, “streamers” of hex material could be seen running down the inside of the screen interval from highly corroded areas. Metallurgical laboratory evaluation of corroded screen intervals indicated corrosion resulted from preferential corrosive attack (Rocky Mountain Engineering and Materials Technology, Inc. [EMTEC] 2002) or “molten salt corrosion” (CMS 2002). An overall assessment of the pipe corrosion in EM TE C’s 2002 report was described as “classic manifestations of chloride attack of austenitic stainless steels, from stress corrosion cracking and knifeline attack to pitting and preferential attack caused by chromium depletion.” 4.1.3 Off-Gas Treatment System Several components of the off-gas treatment system were evaluated for potential corrosion problems following shutdown of the ISTD system. Visual inspection of the interior of the cyclone separator and the base of the FTO did not reveal any significant corrosion. The knockout pot storage tank was also visually inspected. The tank had accumulated approximately 200 gallons of corrosive liquids (pH approximately 0) during operation of the off-gas treatment system. The tank was flushed and no visual evidence of corrosion was evident, except corrosion on the tank sight glass holder from contact with corrosive liquid that escaped through a small leak. However, a transfer pump and discharge line used in an initial attempt to drain liquids from the knockout pot tank were corroded and damaged (Versaw 2003). The off-gas treatment system was shut down under emergency conditions because of an operational failure (Versaw 2003). Some liquid appeared to escape the knockout pot to the acid scrubbers and some discoloration of acid scrubber media in Scrubber Bed No. 1 was observed. Samples of this discolored acid scrubber media were analyzed for remaining neutralization potential and analytical results indicated that 75 percent of the neutralization potential remained in the discolored media. However, in an attempt to dry out the scrubber bed, the heat exchanger between the FTO and the scrubber bed was bypassed. The resulting hot air caused the combustion of carbon in the final carbon bed that precipitated the emergency shutdown. 4.2 FAILURE ASSESSMENT In general, components of the ISTD system at the Hex Pit failed due to severe and rapid corrosive attack. Conditions that led to the corrosive attack appeared to include the following: * Higher than anticipated production of chloride and HCl * Lower than anticipated buffering or neutralization of HCl by other materials disposed of in the Hex Pit and in the surrounding soil * Higher than anticipated heat losses in the aboveground piping network As discussed in TerraTherm (2002), the high level of HCl production could have resulted from the occurrence of layers or lenses of highly concentrated hex residues disposed of in the Hex Pit. The tar-like waste material was disposed of in bulk or thin-walled drums, many of which probably broke when dumped or later corroded in the highly acidic environment. The waste material was periodically covered with soil or lime, eventually resulting in a mix of relatively pure waste material sandwiched between layers of soil and lime (see also descriptions of the Hex Pit contents in Section 2.2.3 and the soil borehole logs in Appendix B). With the start of thermal treatment, the tar-like waste material may have lost viscosity and flowed into the HV wells. The heat and vacuum pressure, combined with the presence of steam, may have allowed the waste material to rapidly produce HCl as it flowed into and was drawn up inside the HV wells. The waste material may have undergone very little in situ treatment (thermal destruction) and the HCl produced may not have been significantly neutralized by the soil and lime also disposed of in the pit. It appears that vaporized or steam-stripped contaminants cooled in the un-heated flexible hoses that connected the HV wells to the aboveground piping network. Cooling may have allowed precipitates to form at the tee fittings and in the hose end connectors, which restricted or completely blocked the vapor flow. The resulting loss of flow velocity in the vapor stream may have allowed the formation of corrosive liquid condensates. Conversely, cooling may have led directly to the formation of liquid condensates, which restricted or completely blocked the vapor flow. Precipitates may have formed primarily after the cessation of heating. Regardless of the mechanism of condensate formation, the resulting aqueous HCl is much more corrosive than HCl in the vapor phase, and its contact with the system components at temperatures around the boiling point of water was likely to lead to the corrosion observed. In summary, destruction of the ISTD system at the Hex Pit appears to have been primarily due to the occurrence of layers of virtually pure, tar-like waste material, which was not destroyed in situ; the generation of HCl, which was not adequately neutralized by in situ materials; the choice of 304 stainless steel for both aboveground and subsurface components, which were exposed to chloride attack during system operation; and the inability of the system to maintain the vaporized or stream-stripped contaminants in the vapor phase for transport to the off-gas treatment system. SECTION 5. REFERENCES Colorado Metallurgical Services (CMS). 2002. Report on Evaluation of Corrosion. Prepared for TerraTherm, Inc. April. Department of the Army. 2002. Memorandum from B. M. Huenefeld, Rocky Mountain Arsenal Committee Coordinator, to K. Guy, U.S. Environmental Protection Agency. Subject: Draft Hex Pits Material Failure Assessment Report. April 25. Rocky Mountain Engineering and Materials Technology, Inc. (EMTE C). 2002. Hex Pit Soil Remediation Failure Evaluation. Prepared for David Bradfield, Foster Wheeler Environmental Corporation. July 8. ENSR Corporation (ENSR). 1999. Hex Pit Site Characterization Report, Rocky M ountain Arsenal, Commerce City, Colorado. Document Number 2840-005-500. August. ENSR. 2000. Hex Pit Treatability Study Report, Part A – Treatability Test Results, Part B – Conceptual Design and Cost Estimate. February. Foster Wheeler Environmental Corporation (FWENC). 1996. Record of Decision for the On-Post Operable Unit, Final. Ver. 3.1. June. FWENC. 2002. Hex Pit Remediation Project Draft Construction Completion Report. December 10. FWENC. 2003. Amendment to the Record of Decision for the On-Post Operable Unit, Rocky Mountain Arsenal Federal Facility Site, Hex Pit Remediation. April 17. Gilbert, R. O. 1987. Statistical Methods for Environmental Pollution Monitoring. Van Nostrand Reinhold. New York, New York. MK-Environmental Services (MK). 1989. Investigation of the Hex Pit as a Possible Source of Groundwater Contamination at the RM A. August. MK. 1998. Hex Pit Design Data Collection Sampling Report. February. Program Manager for Rocky Mountain Arsenal. 1996. Rocky Mountain Arsenal On-Post Operable Unit Record of Decision Dispute Resolution Agreement (DRA). June 10. Stegemeier, G. L., and Vinegar, H .J. 2001. Thermal Conduction Heating for In-Situ Thermal Desorption of Soils. Ch. 4.6-1 in: Chang H. Oh (ed.), Hazardous and Radioactive Waste Treatment Technologies Handbook, CRC Press, Boca Raton, FL. TerraTherm, Inc. (TerraTherm). 2001. Hex Pit Remediation Final (100%) Design Package. March. TerraTherm. 2002. Hex Pit Remediation Material Failure Assessment Report. April. Tetra Tech EM Inc. (Tetra Tech). 2001. Draft Screening Investigation Report for the Hex Pit Screening Investigation. January. U.S. Environmental Protection Agency (EPA). 2000. Guidance for Data Quality Assessment: Practical Methods for Data Analysis. QA/G-9. July. EPA. 2001. Quality Assurance Project Plan, In Situ Thermal Destruction Technology Evaluation at the Hex Pit, Rocky Mountain Arsenal, Commerce City, Colorado. June. EPA. 2002. D raft Post-Demonstration Sampling and Analysis Plan, In Situ Thermal Destruction Technology Evaluation at the Hex Pit, Rocky Mountain Arsenal, Commerce City, Colorado. October. Versaw. 2003. Personal communication from Ron Versaw, Foster Wheeler Environmental Corporation, to Neil Bingert, Tetra Tech EM Inc. April 14.