R. E. Haun and R. C. Eschenbach
Retech, Division of LMAES

J. A. Batdorf
SAIC

Plasma systems are attractive for hazardous waste treatment because they can make a very durable waste form, with minimal waste volume. This paper describes important design considerations for such systems, and reports tests with both rotary hearth and fixed hearth systems. The system selected for remediation of Pit 9 at the Idaho National Engineering and Environmental Laboratory was Retech's eight foot Plasma Arc Centrifugal Treatment (PACT-8) system. The results of shakedown tests of this unit and PACT development work at smaller scale will be reported. Also, results from two one-hundred hour tests of SAIC's pilot-scale Plasma Hearth Process (PHP) unit, built by Retech, will be discussed. Finally, imminent new applications and planned tests will be described.

Retech started working on plasma systems for treating hazardous wastes in 1985, using arc devices and furnaces originally developed for melting metals. In the ensuing years, a number of treatment systems have been built and delivered to customers. All of these systems have involved melting the condensed phases, then casting the melt into a mold. Since the melt is typically at a temperature of 1300 to 1600°C, many of the constituents of the feed evaporate into the headspace and pass from the primary melt chamber to a secondary treatment chamber before entering a gas cleanup system.

Typical treatment chamber sizes for lab-scale units are 0.6 to 0.9 m outside diameter. Such systems may treat from a few to a several tens of kilograms per hour. However, production scale equipment needs to treat hundreds or thousands of kilograms per hour. Scaling work by Retech has shown that the melting rate of metal oxides is roughly proportional to bath surface area (i.e. proportional to the square of linear dimension). We have built five systems with the main chamber diameter approximately three meters. Such systems can process up to 1000 kg/hr, needed for larger volume applications.

The first production size system Retech built was for MGC Plasma in Muttenz, Switzerland, and was delivered in July 1990. The original concept (ref. 1) was to charge material in 200 liter drums, which would be rolled into a 5 drum feeding chamber, and picked up by a feeder mechanism for charging to the primary chamber. This feeding system proved to be unsuited for production because of frequent occurrences of drum dropping. An improved feeder mechanism was developed, and the system was used for production operations with a metallurgical waste (ref. 2). The Muttenz Plasma Arc Centrifugal Treatment (PACT) system had an eight-foot (thus PACT-8) diameter centrifuge within which the material charged was melted.

Plasma waste treatment systems which use a transferred arc are roughly twice as efficient for melting condensed phases as non-transferred arcs (as a consequence of limited bath contact area), which is a strong incentive to avoid non-transferred arc gas heaters (ref. 3 discusses this point in more detail). Molten slag baths are electrically conductive, as are molten metal baths. However, solidified slag at room temperature is not electrically conductive, raising an important design element: starting. This is particularly important for production operations, which need a nearly foolproof method for starting. Refractory design for durability and in certain locations, conductivity, is also important for production operations. Recent developments will be reported in the next sections.

As Retech was building a production-size system (PACT-8) for the Pit 9 program (ref. 4), we took a critical look at the Pit 9 PACT-8 design with regard to operational and maintenance issues. The reason for this critical review was to identify areas where improvements could be made to ensure the operational availability of the melter over a one year time span. The Pit 9 inventory was estimated to be 1,060,000 pounds of waste to be treated during the first phase of design in April 1994 (this was labeled the 30 % design phase). This was then planned to be accomplished in eleven 100-hour campaigns. At the 90% design phase submittal in November 1996, the estimate of Pit 9 material to be treated had grown to 3,920,000 pounds. To accomplish this in one year will require forty 100-hour campaigns. It became very clear to the design review team that the operational availability of the melter had become an important factor to successful completion of melter operations in one year.

Two important areas were identified which could significantly impact operational availability. The first was the need for quick and reliable starting of the plasma arc torch given the potentially wide compositional range of the feed material. The schedule could not tolerate long pre-heat cycles prior to starting the plasma arc torch. The second was to confirm the durability of the refractory system selected including the refractory lined slag-pouring throat. Also, the schedule could not withstand a complete replacement of the primary chamber refractory during the one year of melter operations.

The approach taken to address these two important areas was a Retech funded program using our pilot scale PACT-2 in Ukiah, CA. The program consisted of parallel efforts to accumulate hours on a scaled down refractory system and to develop a reliable torch starting technique. A wide range of feed materials were processed in the PACT-2 over a twelve month time span to determine the effect on durability of the centrifuge refractory. Torch starting techniques were also evaluated in relation to component life.

The principal findings of this one year program were very encouraging. Through appropriate refractory selection, geometric design, and operational technique using the centrifuge and plasma arc torch, a sacrificial refractory layer (i.e. a slag skull) was able to be built up to enhance refractory life in the bottom and on the walls of the centrifuge. This slag skull protects the electrically conductive refractory in the centrifuge bottom from significant deterioration during waste processing. It also helps to shield the hot refractory lined throat from direct arc impingement thereby enhancing the throat life. Operational parameters were developed for the plasma arc torch which utilized standard torch parts. System start-up and shut-down procedures were also developed to enhance starting the plasma arc torch without the need for an auxiliary device such as a grounding rod, second plasma arc torch or an oxy-fuel burner. This program provided very useful insights for operating and optimizing a production scale system.

In July 1997 the pilot scale development work was confirmed during a shakedown test of the Pit 9 PACT-8. During this test we demonstrated the ability to start the plasma arc torch without an auxiliary device. During about 60 hours of torch operation, we demonstrated the ability to start on a slag layer and confirmed that throat erosion was minimal. We also successfully poured four drums of vitrified waste to a desired height and weight within 41 gallon capacity molds.

This operating experience is being incorporated into design and procedures of other systems.

The Plasma Hearth Process (PHP) is a fixed hearth plasma processing system that was developed by SAIC, Retech, and the Department of Energy (DOE). The PHP technology has been developed to provide efficient treatment of DOE's stored mixed waste. The PHP technology is applicable to the treatment of a wide range of waste types that are currently being generated and stored throughout the DOE complex. The development program has included the construction of four PHP systems – two 200 kW bench-scale systems, one 1,200 kW demonstration test unit, and one 1,200 kW pilot-plant system. The bench-scale systems process feed materials contained in 1-gallon steel cans or introduced through an auger feed system. The demonstration test unit processed feed materials contained in standard 30-gallon steel drums. The pilot-plant system is designed to process feed materials contained in standard 55-gallon steel drums. The PHP systems have processed a wide range of simulated wastes including combustible waste, non-combustible scrap, inorganic sludge, organic sludge, cemented sludge, and heterogeneous debris. Whole, unopened containers are slowly fed to the system and "drip melted" to form a pool of molten slag and molten metal. Organic materials and volatile species are gasified and fully oxidized in a secondary combustion chamber. The resulting offgas is cooled, filtered, and cleaned in an air pollution control system.

The technology is currently transitioning from the research and development phase to the implementation phase. During this transition, Department of Energy support has decreased and private sector support has increased. Currently, SAIC and BNFL Inc. are conducting tests on the PHP Pilot Plant located at Retech's facility in Ukiah, CA. During June of 1997, the system was operated for 101 hours (plasma torch operating time) and processed 11,600 kg of simulated waste (ref. 5). The simulated waste consisted of drums of soil; soil and metal; soil, metal, and wood; and simulated debris waste. During this test, 39 drums of simulated waste and 17 empty drums were processed. A total of 11 drums of slag were produced. The slag had a total mass of 7200 kg and a volume of 2.7 m3. The average time to process a single drum was 1 hour. The feed material throughput ranged from 135 to 375 kg/hr with an average feed rate of 250 kg/hr. During this test the offgas was sampled to obtain baseline data for the emissions of cadmium, chromium, lead, nickel, particulate, hydrogen chloride (HCl), carbon monoxide, and nitrogen oxides. The emissions for all of these regulated compounds were very low. The system consistently removed 99.98% of the HCl which resulted in emissions of less than 0.0007 kg/hr. The particulate emissions were below the method detection limit, which was two orders of magnitude below the regulatory limit. The metals emissions were all at least an order of magnitude below the regulatory limits. Carbon monoxide emissions were fairly constant at 4 ppmv. This indicated excellent combustion (previous testing had indicated greater than six 9's DRE for hazardous organic compounds). The nitrogen oxide emissions varied depending on the feed material but were below regulatory limits.

During December of 1997, the system was operated for 102 hours (plasma torch operating time) and processed 11,600kg of simulated waste. The feed material consisted of drums of simulated debris waste. During this test, 89 drums of simulated waste and 23 empty drums were processed. A total of 16 drums of slag were produced. The slag had a total mass of 8300 kg and a volume of 3.3 m³. The average time to process a single drum was 37 minutes. The feed material throughput ranged from 85 to 390 kg/hr with an average feed rate of 200 kg/hr. Although the throughput rate of this test was lower than the previous test, the feed material was less dense and the volumetric throughput was doubled. It is expected that with minor system changes the volumetric throughput can be doubled again, resulting in 55-gallon drum processing times on the order of 15 minutes.

The results of these tests were very encouraging. The system performed very well and problems encountered in the first test were eliminated in the second test. The pilot-plant system is providing the required data and the next series of tests is expected to confirm the remaining design and operational issues. The results of these tests (and the test series planned for early 1998) are being used to finalize a conceptual design for a PHP mixed waste treatment system and to provide additional operating cost data. These data will be used to complete a life cycle cost model for treatment of mixed waste. At this time, the parameters that have a significant cost impact and are still uncertain or unknown are system throughput, system availability, and hearth refractory life.

BNFL and SAIC are currently finishing tests on the pilot-plant system that will provide data to support the next generation design and a life cycle cost model. These costs and the associated business case are being used to evaluate several mixed waste treatment projects for PHP applicability. BNFL is also considering internal applications for intermediate level wastes.

A Retech PACT system for remediating soil contaminated with military wastes is in the final checkout stages in Munster, Germany. The site is contaminated with both explosive materials and a variety of substances used in chemical weapons. Retech built and delivered a PACT-8 treatment system which uses a 1200 kW transferred plasma torch.

Retech is currently assembling a PACT-8 which will be delivered to the U.S. Naval Shipyard at Norfolk, VA to treat their waste (ref. 6). This will treat wastes at only 250 kg/hr, so it will have a lower power plasma system (750 kW) for the primary chamber. It will also have a non-transferred 750 kW arc gas heater to provide supplemental heat to the secondary treatment chamber, thus allowing permitting in Virginia under EPA's Subpart X, rather than subpart O, regulations.

We are currently completing the design of a PACT-8 system to treat production wastes generated at the Tsuruga power station of the Japan Atomic Power Company (ref. 7). The prime contractor for the project is the Toyo Engineering Corporation of Japan. The system will handle 200 liter drums filled with grouted waste or with heavy components, shredded organic waste, and miscellaneous inorganic waste drums which will be charged horizontally, and cut up during the charging process.

Retech has also designed and will build a system for nuclear power plant waste, to be installed in Switzerland (ref. 8). This will be somewhat similar to the one for Tsuruga, except that all the waste to be treated will arrive at the melter in drums, so that no cutting will be done externally. Also the Swiss waste will include some drums of liquid, which will be pumped into the primary chamber to improve uniformity of off-gas evolution.

1. M.R. FÜ NFSCHILLING, R.C. ESCHENBACH, "A Plasma Centrifugal Furnace for Treating Hazardous Waste, Muttenz, Switzerland," Electrotech '92, Montreal, Canada (1992).
2. R.C. ESCHENBACH, "Plasma Arc Systems for Waste Treatment & Metal Recovery," JOM, a publication of The Minerals, Metals & Materials Society, pp. 49-52 (1996).
3. R.C. ESCHENBACH, "Plasma Arc Centrifugal Treatment (PACT) of Hazardous Waste," 8th SPSM, Tokyo University (1995).
4. R.C. ESCHENBACH, M.B. ARNDT, and G.D. PIERCE, "An Integrated Chemical/Thermal Treatment System for Mixed Waste," Waste Management '95, Tucson, AZ (1995).
5. Science Applications International Corporation, Pilot-Scale PHP Limited Source Test Report, SAIC-97/2704 (1997).
6. F.H. GEHRMAN, and DR. BRUCE SARTWELL, "Development of a Plasma Arc System for the Destruction of U.S. Department of Defense Hazardous Waste," Environmental Stewardship - Ships and Shorelines, Virginia Beach, VA, (1997).
7. K. YAMAZAKI, S. KARIGOME, Y. AKAGAWA, Y. DOUGAKI, Y. NAKAYAMA, K. OBARA, A. TSUCHIYA, Y. TSUJI, "Plasma Arc Melting Treatment Process of Low Level Dry Active Waste," WM'97, Tucson, AZ (1997).
8. W. HOFFELNER, TH. MÜ LLER, M. FÜ NFSCHILLING, A. JACOBI, R. C. ESCHENBACH, "New Incineration & Melting Facility for Treatment of Low Level Radioactive Wastes in Switzerland," 1994 Incineration Conference, Houston, TX (May 1994).

BACK