Jeffrey P. Kulpa
RMI Environmental Services
Ashtabula, Ohio

Ward E. Best
United States Department of Energy
Ashtabula, Ohio

Michael J. Mann, P.E.
ARCADIS Geraghty & Miller, Inc.
Tampa, Florida

The remediation of contaminated soils is a major contributor to the cost of the environmental restoration effort at DOE sites previously employed in the DOE weapons production complex. At the RMI Decommissioning Project (RMIDP) in Ashtabula, Ohio, soil remediation by traditional ship and bury methods comprised over $44 Million of the $156 Million decommissioning baseline budget. The RMIDP has developed and is preparing to deploy a chemical treatment technology for washing uranium contaminated soil.

Soil washing provides a significant volume reduction alternative to transportation and burial or on site disposal when all life cycle costs of burial are considered. Soil washing at the RMIDP is projected to save the DOE over $13 Million in direct remediation costs of the 40,000 tons of uranium contaminated soil and $12 Million in indirect project savings through schedule reduction.

From August 1996 through February 1997 the RMIDP performed Process Definitive Testing (PDT) to validate initial screening study findings regarding the viability of soil washing. PDT defined the operating process parameter requirements for the soil washing carbonate leaching system, and provided valuable design data to be used in production plant design.

A 2 ton per batch soil washing plant was designed, constructed, and operated in a 6 month period. Over 68 tons of soil were processed to provide operating data at a scale close to production scale. The data derived from pilot operations proved the technical viability of carbonate leaching on RMI soil and provided sufficient plant operating data to allow a cost-benefit assessment of soil washing to be performed. The RMIDP soil washing production facility has been designed and construction is scheduled to begin in early 1998.

Site remediation activities at the RMI Extrusion site in Ashtabula, Ohio require extensive efforts and financial investment to remediate the 40,000 tons of uranium contaminated soil which exist as a result of past operations.

The RMI Titanium Company operated the RMI Extrusion Site in Ashtabula, Ohio as part of the U.S. Department of Energy (DOE) weapons production program. Depleted, normal, and slightly enriched (up to 2.1% weight percent U-235) uranium was extruded as part of the DOE production reactor fuels manufacturing process. Uranium extrusion operations were performed from 1962 through September 1988. RMI also extruded depleted and natural uranium under a Nuclear Regulatory Commission (NRC) license and extruded non-radioactive metals (primarily copper based) for the commercial sector. However, the majority of material processed at the facility was for DOE. All extrusion operations at RMI ceased on October 31, 1990.

In 1990, the NRC identified the RMI Extrusion Site for accelerated cleanup under the Site Decommissioning Management Plan. DOE accepted financial responsibility for the cleanup based on previous site operations, which supported the DOE mission. Since 1990, the site has performed equipment and legacy waste removal in preparation for building demolition, and also performed selected soil remediation, and planning/permitting in support of groundwater remediation. The remediation effort is governed by guidance and regulations promulgated by the NRC for site decommissioning and license termination. In addition, the presence of chemical constituents in groundwater has led to involvement by the U.S. and Ohio Environmental Protection Agency (EPA) in remediation decisions involving a Corrective Action Management Unit (CAMU). RMI Environmental Services is the prime contractor to DOE for completion of site decommissioning. The DOE Ashtabula Environmental Management Project Office provides day to day oversight of the remediation effort. The DOE approved project baseline for the decommissioning project is $156 Million.

The RMI Extrusion site consists of 25 buildings located on 7 acres of the 25 acre site. The site is located in Northeastern Ohio about 50 kilometers north of Cleveland and approximately 1.5 kilometers south of Lake Erie. The area immediately adjacent to the site is sparsely populated. The majority of the site and surrounding area is relatively flat with the maximum elevation variation being 4 feet. The site is on the Lake Plain geological feature, which formed in the later part of the glacial Wisconsin Age. The site soils are very clay-like, approximately 60% fines (<0.038 mm) and 40% sand/silt (<1 mm and >0.063 mm).

The site remediation plan which has been approved by the NRC requires soil be remediated to below 30 pCi/gram total uranium. A total of approximately 40,000 tons (1 ton = 2,000 pounds) of uranium contaminated soil have been identified by characterization efforts to require remediation. On average, the soil at RMI is contaminated to about 125 pCi/gram total uranium with maximum contamination levels of 500 pCi/gram observed in several isolated locations.

Two principal mechanisms exist for the uranium contamination of the soil surrounding the plant area. The predominant mechanism is air deposition resulting from stack emissions of extrusion and machining process waste products during the 26 years of plant operations. These emissions while below regulatory limits resulted in uranium exiting the stacks and settling on the immediate surrounding areas. The second mechanism results in direct run-off deposition from plant maintenance operations such as rinsing of equipment. This mechanism contaminated the soils close in to the plant buildings.

The contaminated soil areas of the RMI Extrusion site have been divided into three areas. Area B is the process area, which contains the 25 plant buildings. Area B has been contaminated by air deposition and low volume releases of washwaters. Area B also contains a substantial percentage of non-native materials such as crushed stone used for fire roads. Characterization efforts have revealed contamination levels as deep as 3 feet in Area B. Area C is the land north of the process building. Area C includes a flat grassy area leading to an embarkment which drops 20 feet to Fields Brook at the northern boundary of the site. Area D is a flat grassy area east of the plant processing area. Area D has been contaminated primarily by stack emissions from the old burn area, which disposed of uranium contaminated materials. Some low-lying areas of Area D have been contaminated by surface run off from Area B. Table I shows the matrix characteristics of the soil which have some impact on soil washing uranium removal efficiency.

Table I illustrates the high clay content of the RMI soil. Early screening studies revealed that the uranium contamination distributed itself uniformly across the entire range of the soil matrix. The uniform distribution of contamination made typical soil washing approaches such as physical separation not effective. Chemical extraction technology with its ability to treat the entire range of soil matrix sizes including the fines fraction provided the greatest potential for success in reducing contaminated soil volumes.

Uranium commonly exists in two oxidation states. It was expected that much of the uranium at the RMI site would be in the oxidized or U⁺⁶ state since the principle mechanism for contamination had been stack emissions. Uranium in the U⁺⁶ state is very soluble and lends itself readily to chemical treatment and removal. Uranium contamination associated with processing wastes, such as observed in some soil in Area B, is often seen in the reduced or U⁺⁴ state. This form of uranium is much less soluble and more easily complexes with organic matter in the soil. Therefore the U⁺⁴ form of uranium is much more recalcitrant to chemical treatment. In some cases, oxidizers such as hydrogen peroxide or acids can be used to oxidize the uranium to the U⁺⁶ state and make it more soluble. The use of oxidizers can be costly and often only results in marginal improvements in uranium removal efficiency. Process Definitive Testing at RMI revealed that the use of oxidizers was not required to achieve acceptable levels of uranium removal efficiency.

Soil remediation is the major cost contributor for the RMI Decommissioning Project (RMIDP) baseline. Initial baseline studies in 1989 concluded that soil washing would not be effective since physical separation was the only proven technology at that time. The initial baseline concluded that ship and bury was the method necessary to complete soil remediation. Over $45 Million of the $156 Million baseline are associated with soil remediation. In 1995, a project initiative to reduce baseline costs and accelerate project schedule was undertaken. The initiative included an effort to re-examine soil washing based on promising results using chemical extraction being reported at other DOE sites. In August 1995, a two week screening study was completed which concluded that chemical extraction using a carbonate based or acid based approach was technically feasible on RMI soil (Reference 1). A cost benefit assessment based on initial uranium removal efficiencies and projected volume reduction capabilities as reported in the screening study concluded that the project could save up to $25 Million by implementing soil washing.

In 1996, the RMIDP undertook an effort to clearly define the exact approach to be undertaken in implementing soil washing. Process Definitive Testing is the process which defines the process to be employed and generates real cost and performance data at a scale close to production scale. The RMIDP tested over 150 different combinations of soil, process temperature, and reaction times in order to generate the technical data being used to design the production facility.

The effort at RMI included initial bench scale testing to define pilot plant design parameters. The pilot plant was designed, constructed, and operated on over 70 tons of soil in a six month period. The pilot plant operation results provided essential performance data upon which the full scale production facility design is based.

Process Definitive Testing (PDT) revealed that RMI Soils could be treated effectively using a 0.2 Molar sodium carbonate/bi-carbonate solution at a temperature of approximately 115^oF and a retention time of 1.5 hours (Reference 2). The bench scale testing showed that carbonate extraction achieved removal efficiencies of up to 90% and would be effective in meeting the site treatment standard of 30 pCi/gram for approximately 95% of the site soils.

The pilot plant was designed to process 2-ton batches of contaminated soil using the key parameters identified in the bench scale testing. The equipment used in the pilot plant construction was assembled from several sources including DOE-Ashtabula, DOE-Fernald, RMI Titanium Company, and Geraghty & Miller (Alternative Remedial Technologies). A block diagram of the pilot plant is shown in Figure 1. Table II describes major plant components of the pilot plant. The key plant components include the rotary batch reactor in which a heated carbonate solution is contacted with the feed soils, a liquid/soils separation unit to remove the soluble uranium, a de-watering system for the soils, and an ion exchange system to allow removal of the uranium from the liquid. The pilot plant was installed in a 100 foot by 100 foot area of an existing plant building.

One to two tons of composited soil was batch loaded into the rotary batch reactor (RBR), and a 0.2 Molar sodium carbonate/bi-carbonate solution was added to achieve a 30% solids slurry. For the majority of batches which were run at elevated temperatures, the carbonate solution was heated before being added to the RBR.

After the initial leaching period was complete (usually 2 hours), the slurry was pumped to a wet screen to remove oversize material. The oversize material received no further treatment. The oversize fraction for pilot operations was approximately 5%. The slurry containing particles less than 1 millimeter was transferred to the two sequential thickeners. The thickeners performed solids separation. Fresh leach solution was added to produce a slurry containing approximately 10% solids. The fresh leach solution rinsed the uranium rich leach solution from the soil and permitted additional uranium extraction. Flocculant was added to allow suspended fines to settle out of solution. Dewatering of the treated soil was performed by means of a plate and frame filter press. The filter cake produced by the filter press was staged as clean soil.

The uranium rich supernatant received further processing. Uranium dissolved within the leach solution was recovered using dual upflow ion exchange columns. A small filter press was used upstream of the ion exchange columns in order to remove fine particles from the pregnant leachate and protect the ion exchange columns from being fouled. In passing through the ion exchange resin, the dissolved uranyl carbonates were adsorbed onto the ion exchange resin. The leachate was recycled back into the system. Once loaded with uranium, resin efficiency decreased. The resin was regenerated to permit resin re-use. The resin was regenerated by backwashing the columns with 1.5 Molar sodium chloride and 0.05 Molar sodium carbonate solution.

Uranium eluted from the resin during regeneration was recovered by a two stage chemical precipitation reaction. First, hydrochloric acid was used to lower the pH (<2) and drive off excess carbonate as carbon dioxide. This step was followed by the addition of hydrogen peroxide and a pH adjustment with sodium hydroxide. These compounds brought about the precipitation of uranium as uranyl peroxide or "yellow cake." An extremely small volume, less than 0.1% of initial feed volume, is collected as precipitated yellow cake.

The following process parameters were varied during the performance of the 38 batch runs:

• Feed soils (various site locations and feed contamination concentrations),
• Reaction temperature (ambient, 110^oF to 120^oF, and 140^oF), and
• Leaching time (1 to 2 hours).

The performance of the pilot plant operation was closely monitored by the completion of an extensive sampling and analysis campaign (Reference 3). Sampling and analysis was performed to assess:

• Leaching performance
• Ion Exchange system performance
• Resin Loading
• Resin Regeneration
• Uranium Precipitation

Radioactive constituents were measured in the soil and in process water. X-Ray Fluorescence (XRF) was used as the primary uranium measuring tool for in-process parameters. The quality assurance program required independent analysis using gamma spectroscopy and alpha spectroscopy to validate XRF results. Kinetic Phosphorescence Analysis (KPA) was the primary analysis tool for measuring uranium concentrations in process water. These analyses were used for assessing resin performance and leaching kinetics. Wet process chemistry measurements of parameters such as pH, carbonate concentration, and chloride concentration were performed on all batches at various operational locations to allow for control of the process and assessment of process performance. Table III presents typical results from the batch pilot operations.

The key technical findings of pilot plant operations can be summarized as follows:

• Batches performed on soil from Area C and Area D resulted in uranium removal efficiencies in the range of 87% to 91%, averaging 90%. Treated soil product activity consistently measured below the treatment standard with a measured activity in the range of 12 to 14 pCi/gram.
• Batches performed on Area B material resulted in removal efficiencies in the range of 65% to 84%. The lower efficiencies in Area B probably result from different deposition mechanisms, and different feed material physical composition. Area B contains relatively large amounts of non-native oversize materials, which will require special attention during production operation. During full scale production operations, identification and selective excavation of untreatable hot spot soils within Area B will be required. However, the percentage of untreatable hot spot soils is estimated to be less than 5% of total contaminated soil volume.
• Leaching kinetics analysis revealed that an effective contact time is 75 minutes and 115^oF is an effective treatment temperature. Longer leach times and higher temperatures resulted in marginal improvements in uranium removal efficiency.
• The effectiveness of oxidant addition using 1000 ppm hydrogen peroxide in bench scale testing of several batches revealed that oxidants had no beneficial effect on Area C and Area D soil and a small positive effect on Area B soil.
• The ion exchange system performance was very good for the entire pilot project. Ion exchange performance averaged 91% uranium removal efficiency from the leachate based on average input concentration of 16 ppm and output concentration of 1.7 ppm uranium.
• The resin was successfully regenerated several times during pilot operations to restore the resin to an acceptable removal efficiency. Resin loading efficiencies observed during pilot operations were slightly lower than achieved in laboratory testing. The slightly lower resin loading was probably due to organic fouling of the resin during plant operations. Overall, resin performance was deemed satisfactory for full scale operations.
• A mass balance of uranium in the feed compared to uranium removed in the ion exchange columns, process waste streams, and regenerant solutions was conducted. The mass balance revealed that 15.3 pounds of uranium entered the plant and 85% or 13 pounds were in the output streams. The difference is attributed to the multiple number of data points and the variability of measurement at these data points. Overall, the 85% closure of the mass balance was deemed acceptable.

The pilot plant operation provided technical data upon which the production plant design is based. Pilot operations revealed the following key information points, which require special attention:

• Flocculants added to the process slurry were marginally effective in facilitating solid/liquid separation. Flocculation difficulties were attributed to changing water chemistry and varying weight percent of solids during batch operations. Production plant design is for a continuous flow process with special attention being paid to flocculant/slurry mixing effectiveness.
• Carry-over of fine solids occurred downstream of the thickener tanks and the de-watering filter press. Carry over of fines is a concern with respect to plugging of ion exchange input filters or resin columns. Design of the production plant has concentrated efforts at minimizing carryover of fines and establishing an effective filtering technique prior to the ion exchange columns.
• Foaming occurred in the Rotary Batch Reactor, vibrating screen, thickener tanks, precipitator tanks, and evaporator tanks. Anti-foaming agents were tested and were somewhat effective in reducing the amount of foaming. Foaming considerations are an important part of production plant design with the effort concentrating on selection of anti-foaming materials and design of processing equipment to minimize air/process water entrainment.
• Material selection issues existed in the fact that certain PVC valves and pipe fittings were exposed to excessive loads and vibration which resulted in fatigue failure. In addition, mechanical seals supplied with carbon/ceramic/EPDM construction materials failed showing breakdown of the stationary EPDM elastomer due to chemical incompatibility. The replacement seal was upgraded to viton and exhibited greater chemical resistance. Several other mechanical seals failed due to abrasion causing production plant design to weigh the pros and cons of mechanical seals versus packing glands.

The RMIDP Pilot Plant Project included plant design, equipment procurement, plant construction, plant operation, data analysis, and plant decontamination. The cost of the pilot plant project was minimized by reusing existing equipment from various sources and by leasing the main filter press, air compressor, and boiler. The total cost of pilot operations was $638,670 which was primarily composed of plant operational costs and laboratory analysis efforts.

Pilot plant operations validated earlier results which concluded that soil washing at a production scale could be successfully performed at a cost of between $250 and $350 per ton. The data obtained during pilot operations has been used to design a simple, inexpensive production facility. Pilot operations helped define the expected chemical, utility, equipment, and labor needs associated with treating site soils.

The design of the full scale production facility has been completed with procurement of plant components expected to begin in March 1998. The plant will be constructed in a new production facility building to be constructed in the west area of Area B. The Area B soil in the location of plant construction has been remediated, backfilled, and compacted in preparation for construction activities.

Based on the expected volume of soil to be processed, a 10 ton per hour continuous feed soil treatment facility that uses standard soil handling and processing equipment has been designed. Excavated soil will be staged on a concrete staging pad and a heated soil storage building will be constructed to allow year round soil processing in Northeastern Ohio. The production plant will require three operators and a foreman. Operations are planned for two shift operation, five days per week. Plant production operations will be conducted for 14 hours per day, allowing up to 140 tons of soil to be processed daily.

Soil treatment at RMI is projected to save DOE approximately $300 per ton over traditional ship and bury approaches. In addition to the direct savings resulting from chemical treatment, schedule reduction will result in additional savings to the RMIDP. Total savings to DOE resulting from the performance of soil washing on the 40,000 tons of contaminated soil will be approximately $25 Million.

Soil washing at the RMI Decommissioning Project has been verified to be technically viable and economically feasible. From August 1996 through February 1997, Process Definitive Testing was conducted to define operating process parameters and to provide design data for the design of a production facility. A pilot plant was designed and operated at close to production scale to provide detailed information regarding process performance. The data derived from pilot plant operations has been used to complete the design of the RMIDP Soil Washing Production Plant.

1. Soil Washing Technology Screening Study for the RMI Extrusion Plant, Ashtabula, Ohio; Alternative Remedial Technologies, Inc.; September 1995.
2. Soil Washing Treatability Report for the RMI Titanium Company Extrusion Plant Site, Ashtabula, Ohio; RMI Environmental Services, November 1996.
3. Soil Washing Pilot Project Report for the RMI Titanium Company Extrusion Plant Site, Ashtabula, Ohio; Volumes I & II, RMI Environmental Services, April 1997.

BACK