4.4.1.3 Long-Term Effectiveness and Permanence -- Alternative 1 The long-term effectiveness of HTTD was assessed in terms of the risk remaining after treatment. Risks may be posed by chemicals not removed from the contaminated material during treatment. Performance data, factors that influence performance, and system limitations are discussed below. High Temperature Thermal Desorption Performance EPA has specified a PCB performance standard of 2 ppm for evaluating whether HTTD systems are equivalent to incinerators in removing PCBs. Full- and pilot-scale applications using a number of HTTD technologies prove that this PCB performance standard can be met. The results of full-scale applications of thermal desorbers at PCB-contaminated sites are discussed below. Table 4-4 summarizes full- and pilot-scale performance data for various HTTD technologies. To date, only three vendors have used thermal desorption technology for full-scale remediation of PCB-contaminated Superfund sites: SoilTech; RUST; and Westinghouse Remediation Services, Inc. (WRS). The SoilTech ATP system was used at the WBD site in Brant, New York, and at the Waukegan Harbor Superfund site in Waukegan, Illinois. At the WBD site, the ATP was used in combination with dechlorination; therefore, the performance of the ATP at the WBD site is discussed in Section 4.4.2.3. Performance data was collected during SITE demonstrations and proof-of-process (POP) tests at the three sites. Approximately 253 tons of soil was treated during the SoilTech ATP SITE demonstration at the Waukegan Harbor Superfund site in June 1992. The average PCB concentration was reduced from 9,761 mg/kg in feed soil to 2.0 mg/kg in treated soil, a 99.8 percent average removal efficiency. An average of 0.84 microgram per dry standard cubic meter (頜/dscm) of PCBs was emitted to the atmosphere from the ATP stack, resulting in a 99.999987 percent DRE. During the Waukegan Harbor site remediation, the ATP treated about 12,800 tons of harbor sediment and sandy soil containing 6,700 to 23,000 mg/kg PCBs. The treated soil PCB concentrations were less than 2 mg/kg. The SoilTech ATP has been used to remediate three additional Superfund sites contaminated with VOCs, SVOCs, and pesticides. The RUST X*TRAX system SITE demonstration was performed in May 1992 at the Re-Solve, Inc., Superfund site in North Dartmouth, Massachusetts. Approximately 215 tons of PCB-contaminated soil was treated. The average PCB concentration in feed soil was 247 mg/kg, and the average PCB concentration in treated soil was 0.13 mg/kg, a 99.95 percent removal efficiency. PCBs were not detected in stack gas samples. During the Re-Solve, Inc., site remediation, the X*TRAXTM system treated over 50,000 tons of soil containing 25 to 13,000 mg/kg PCBs. Treated soil PCB concentrations ranged from 2 to 8 mg/kg. The X*TRAX system has been selected to remediate a PCB-contaminated sludge site in South Carolina. The WRS low temperature thermal stripping (LTTS) system was selected to clean up the Acme Solvents Reclaiming, Inc. (Acme), site in Rockford, Illinois. The site is contaminated with VOCs, phthalates, and PCBs. A POP test was conducted in March and April 1994. Approximately 210 tons of contaminated soil was treated during 10 test runs. Contaminated soil contained up to 20 mg/kg of PCBs, and treated soil contained less than 2.5 mg/kg of PCBs. The vent gas contained less than 1.8 g/dscm of PCBs. During the Acme site remediation, the WRS LTTS treated about 6,000 tons of soil and achieved an average treated soil PCB concentration of 0.6 mg/kg. Limited bench-scale tests on dioxin-contaminated soil indicates that the SoilTech ATP can effectively remove dioxins from soil. Dioxin concentrations were reduced from concentrations as high as 912 頜/kg in the untreated material to 0.007 頜/kg in the treated soil (SoilTech no date). HTTD performance data is not available for sewage treatment sludge. SoilTech reportedly has conducted full-scale tests on municipal solid wastes; however, test resuts were not available for PRC's review. These materials may be treated if they contain or are blended with inert solids. Factors that Influence Performance The degree to which an HTTD system is able to remove contaminants from wastes and remain cost effective depends on certain key characteristics, including the moisture content of the waste, the boiling points of the contaminants, the hydrocarbon content of the waste, and the particle size distribution and soil classification of the waste being treated. These characteristics are described below. Moisture Content In the HTTD process, excess moisture is removed at the expense of excess burner fuel and can affect treatment performance. Depending on the specific HTTD system being used, the feed material should contain 10 to 20 percent moisture before entering the system. HTTD technologies can treat wastes that contain more than 20 percent moisture; however, pretreating wastes that contain more than 20 percent moisture improves process efficiency. Pretreatment methods include filter pressing wastes, air drying wastes, blending wastes with dryer materials, and mixing with treated fines. The moisture content of most of the contaminated material from the six CD sites except the lagoon sludge is not expected to cause problems during HTTD treatment. The lagoon sludge, even when dewatered, will contain a significant amount of moisture. The lagoon sludge should be blended with dryer material for treatment. Controlled air drying and blending should be considered to reduce the moisture content of the PCB-contaminated material. Site wastes may contain significant levels of VOCs; therefore VOC emissions may hamper air drying and blending operations. Site wastes may also contain asbestos and metals, and air drying and blending may result in significant and potentially harmful inorganic emissions to the air. All drying and blending operations should be conducted in a controlled environment. Boiling Points of Contaminants For effective desorption of contaminants from soil, an HTTD system should be operated at temperatures exceeding the boiling points of the contaminants in the soil. PCBs are the primary organic contaminants of concern at the CD sites. Commercial PCBs were produced by collecting boiling point fractions during distillation of chlorinated biphenyl mixtures. Table 2-10 presents the molecular weight, boiling point range, vapor pressure, and vaporization rate for various Aroclors. In general, the boiling point range increases and the vapor pressure and vaporization rate decrease with increasing molecular weight. Therefore, it is expected that more highly chlorinated Aroclors would be more difficult to remove from waste than less chlorinated Aroclors. Similarly, more chlorinated isomer groups such as octachlorinated biphenyls are expected to be harder to remove than less chlorinated isomer groups such as trichlorinated biphenyls. The principal Aroclor detected at the CD sites was Aroclor 1248. Therefore, an HTTD system should be operated at temperatures exceeding about 750 澹 to effectively remove PCBs from contaminated materials at the sites. Higher HTTD temperatures will be required to remove dioxins and furans, which have boiling points estimated to be as high as 1,000 澹. Hydrocarbon Content Contaminated sites typically contain "hot spots" where contamination levels are much higher than in surrounding areas. Hot spot soils have elevated heat values and can cause fluctuations in the treatment system temperature if not combined with less-contaminated material. In order to achieve uniform feed material characteristics, excavated material containing high levels of organic or oily wastes typically must be blended before being fed to the primary desorption chamber. Different systems have allowable maximum feed concentrations ranging from 3 to 25 percent hydrocarbons by weight. Daily sampling and analysis of the blended, stockpiled material for contaminant concentration and moisture content is typically performed by an on-site laboratory before processing to ensure uniform feed material. For a given volume of contaminated soil, a specific volume of gases will be created from vaporized moisture, volatilized contaminants, and products of combustion from the primary chamber burners. The concentration of the volatilized contaminants in the gas stream would determine the lower explosive limit (LEL) of the gas. The LEL is a limiting safety factor associated with low-temperature thermal treatment systems and directly relates to processing rates at specific contaminant levels. In practice, vapor concentrations should be limited to 25 percent of the LEL. Higher levels of organics may be allowed in the desorption unit if oxygen is maintained at low levels in the system. For example, a nitrogen purge gas is used in the RUST X*TRAX system to maintain low concentrations of oxygen. The wastes at the CD sites are expected to be very heterogeneous, and the hydrocarbon content of the contaminated materials are likely to vary significantly. Care should be taken to blend hot spot soils to lower the hydrocarbon content of wastes fed to the thermal desorber. Particle Size Distribution and Soil Classification Care must be taken to properly prepare the contaminated material for treatment. For optimum treatment, feed consistency should be as uniform as possible. Preparation of soils may be needed to ensure that the material is properly sized. By screening or grinding, particles can be reduced to a uniform size of less than 0.5 to 2.5 inches in diameter, depending on the requirement of the system used. Large clumps should not be fed into the HTTD system because some contaminants could remain in the soil because of inefficient heat transfer in the soil particle. Soils with high silt and clay content of greater than 20 to 30 percent and gummy solids may reduce process throughput and increase treatment costs. Heavy clays may need to be processed in a mixer with other materials such as dry sand to achieve a semi-flowable solid. In addition, contaminants tend to be adsorbed onto smaller soil particles because such particles have a larger surface area with more active sites available for contaminant sorption and chemical and physical bonding. Because fine-grained soils such as clayey soil have more active sorption sites, they are typically more difficult to treat than coarser soil and sediment that contain an equal concentration of contaminants. Soil samples collected from the CD sites indicate that site soils consist primarily of clay and silty clay. In addition, large pieces of bedrock may be intermingle with the soils. Materials blending would most likely be necessary to improve soil transport through the HTTD system. System Limitations HTTD technology is not effective in removing nonvolatile inorganic contaminants such as metal wastes (except for mercury, which has a boiling point of 674 澹). Studies indicate that metals in the leachate from the TCLP generally do not increase in concentration after treatment. Although nonvolatile inorganic contaminants such as metals are not removed, they do not inhibit the process when treating organic contaminants. The contaminated solid waste may also contain asbestos. HTTD is not effective for treating asbestos-contaminated wastes. Materials that are difficult to treat include tars and substances that form tars at low temperatures and at relatively short exposure times. The technology is most effective on matrixes that are nonadsorptive and of low porosity. Sands are easily treated because contaminants are easily desorbed from the surface. Materials that are difficult to treat include humus, organic decay products, wood, and industrial adsorbents.