Formatted Title
Lessons Learned Thermally Treating 18 Source Areas on the Velsicol Chemical Superfund Site in Michigan
Background/Objectives
The Velsicol Chemical Superfund Site covers nearly 100 acres along the Pine River in St. Louis, Michigan. Remediation is underway in several operable units (OU) including the Former Plant Site (FPS), which housed a variety of industrial and chemical manufacturing operations dating back to the early 1900s. The site historically manufactured industrial salts, magnesium oxide, rare earth elements, brominated fire retardants, and the pesticide dichlorodiphenyltrichloroethane (DDT). Although chemical manufacturing ceased in 1977, historical operations at this site have left significant local, regional, and statewide environmental impact to both human and ecological receptors. Between 2017 and 2022 in situ thermal treatment (ISTT) in the form of thermal conduction heating was applied in three sequential phases for treatment of 18 discrete source areas in the FPS. ISTT was selected to reduce dense non-aqueous phase liquids (DNAPL) mass and mobility and eliminate potential offsite migration of DNAPL to surface water and sediments of the Pine River which were previously restored by the United States Environmental Protection Agency.
Approach/Activities
Significant amounts of DNAPL were believed present among the collective source zones which spanned more than 120,000 square feet at the FPS. The first application of TCH at the FPS consisted of four individual source areas and included thermal treatment to a depth of 22 feet below ground surface (ft bgs). The second and third phases of TCH treatment targeted 14 additional source areas to depths of up to 32 ft bgs. An adaptive well installation approach was implemented for all TCH phases, to adjust the thermal treatment depth to the elevation of the lower confining layer which varied across the site. While it was initially expected that the majority of the DNAPL mass would be mobilized and extracted from the source areas in the vapor phase, in practice more than 90% of the total DNAPL mass was removed as free product. Adaptive design strategies for DNAPL extraction, treatment and storage were implemented across each operational phase of thermal treatment. In total TCH wellfields included 903 heater wells, 181 multi-phase extraction (MPE) wells, and 122 shallow vapor extraction wells. Remediation progress was monitored and documented at 62 temperature monitoring points and 51 pressure monitoring points.
Results/Lessons Learned
The project team encountered unexpected removal mechanisms during TCH operations. This included the extraction of DNAPL containing 1,2-dibromo-3-chloropropane (DBCP) which required implementation of a complex personnel protection program due to the unique health hazards posed by this contaminant. Although source area contamination was dominated by volatile organic compounds, it became evident during the first TCH phase that DNAPL recovery was the governing mechanism for contaminant removal at the FPS. During the initial phase, more than 90% of the total contaminant mass was recovered as DNAPL by MPE well pumping; operating phases 2 and 3 followed similar extraction trends with 98% and 97% of the total contaminant mass removed as DNAPL, respectively. The ISTT system included provisions for DNAPL recovery; however, the fundamental treatment concept relied on thermal oxidation for the destruction of contaminants vaporized and extracted during heating operations. Accordingly, the presence of DNAPL as the primary contaminant stream triggered significant change in the technical, regulatory, contractual, and financial aspects of the project. Furthermore, unlike most TCH remediation projects, DNAPL mass recovery in all areas of the FPS was observed late in the remedy and after the site had been at boiling temperatures for a substantial amount of time. As a result, the energy input required to effectively remove DNAPL exceeded typical industry observations, with energy densities between 520 and 630 kilowatt-hours per cubic yard treated. Through three sequential phases of ISTT over 380,000 lbs of contaminants were removed from the subsurface, more than seven times what had been estimated to be present in the 18 individual source areas and more than 90% of which was removed as free product.