D. Battleson, M.C. Willis, J.L. Montgomery, S. Babko-Malyi
MSE Technology Applications, Inc.

This paper presents the results of tests on a hot offgas ceramic filter manufactured by Pall, Inc. The thermal driver for the offgas test components was a rotary plasma arc furnace installed at the MSE Technology Applications, Inc. (MSE) facility in Butte, Montana. The tests were sponsored by the U.S. Department of Energy's (DOE) Mixed Waste Focus Area (MWFA) under the Controlled Emissions Demonstration Project (CED) and were conducted from May through October of 1997.

Operating filtration in a thermal treatment system in close proximity to the primary combustion chamber has the advantage of removing carryover particulate close to the source, thus preventing particulate from entering other downstream offgas treatment components (i.e., any wet scrubber section). Successful deployment of a hot offgas filter will minimize downstream contamination of wet scrubbers and possibly capture carryover radionuclides closest to the source in a mixed waste thermal treatment unit.

The Pall filter generally filtered over 99% of particulate emitted from the primary plasma chamber for limited periods of time. Visual inspection of the scrubber blowdown water in the wet scrubber section downstream of the Pall filter showed effluent was much clearer than without the Pall filter. This is indicative of a dramatic reduction in carryover contaminants downstream of the hot filter.

Prior to plugging, the Pall Corporation hot ceramic full-scale filter has operated in the plasma centrifugal furnace offgas system for limited periods. The causes of the plugging seem to be excessive particulate, chemical composition of the feedstock, or a combination of both. A combination of chlorine, lead, and sodium containing compounds in the feedstock seem to cause plugging of the filter in its present configuration.

Some of the transuranic mixed wastes at the Idaho National Engineering Environmental Laboratory (INEEL) and other similar U.S. Department of Energy (DOE) sites are being considered for thermal treatment. Resultant emissions from thermal treatment systems are heavily scrutinized for compliance by the public, government air quality regulators, and the end users of the thermal treatment systems.

The DOE has directed the Mixed Waste Focus Area (MWFA) to conduct the Controlled Emissions Demonstration Project (CED) with the objective of evaluating and selecting leading edge offgas treatment technology for demonstration at MSE Technology Applications, Inc. (MSE) technology development test facility located in Butte, Montana.

The overall goal of the CED Project is to test and demonstrate mature offgas treatment components that will process offgas to meet or exceed the U.S. Environmental Protection Agency's (EPA) new 1997 Maximum Achievable Control Technology (MACT) rule and reduce the total overall offgas emissions and secondary wastestreams. The technologies tested should be applicable to a wide assortment of applications.

The Pall Incorporated Ceramic Vitripore^TM Filter Automated Filtration System, Model#7BBD70002-123, is one of the technologies that was evaluated and selected by a competitive procurement process and tested in 1997.

The Pall filter is located immediately downstream of the 6-foot Plasma Arc Cetrifugal Treatment System (PACT-6) primary chamber and is connected to the primary chamber by approximately 20 feet of 5-inch inner diameter, refractory-lined pipe. Figure 1 is a simplified schematic of the unit, and Fig. 2 shows the filter position in the overall PACT-6.

The filter consists of a carbon steel refractory-lined filter vessel and 40 Vitropore^TM filter elements (candles), filter tubesheet, blowdown receiver vessel, and control subsystem. The filter candles are made of silicon carbide with a sodium aluminosilicate binder.

Tables I and II below, contain process specifications and equipment specifications, respectively, for the filter system.

Offgas enters the filter housing, and the contaminate particulate is collected on the outer surface (dirty side) of the cylindrical ceramic filter elements that are mounted on the tubesheet assembly. Gas flows through the vessel, and particles are deposited to form a cake. The cake becomes less permeable as its’ layer grows and the differential pressure across the filter increases. At a preset differential pressure level, a blowback cycle is initiated. The differential pressure blowback setting can also be overridden by a timed pressure blowback cycle for redundancy. A manual blowback initiation function is also provided. The excess cake is removed during the blowback cycle by injecting dry compressed preheated nitrogen into the interior (clean side) of the ceramic elements through a nozzle on the blowback manifold. Preheating the nitrogen gas prevents thermal shock to the elements. The blowback system is designed for dry caking; wet caking will inhibit efficient operation and may require element removal and cleaning.

A programmable logic controller is the heart of the control system and is primarily used to control the blowback sequence. The controls sequentially open and close four solenoid valves when a blowback cycle is initiated. Blowback gas pressure, system differential pressure, and process temperature are displayed during normal operation by a liquid crystal diode (LCD). Pilot lights indicate alarm conditions, such as excess filter differential pressure or low blowback gas pressure.

Target differential pressure for this process was 10-15 inches water column (wc) and the alarm was set at 30 inches wc.

Because the particulate emitted from the furnace is very fine, a precoat of tabular alumina was applied to improve the permeability of the permanent dustcake. The precoat material aids in the filtration process by becoming part of the filter element itself and filtering smaller diameter particulate. The particle size of the precoat was chosen to try to form a permanent cake that has a somewhat larger particle size than the particulate evolving from the furnace.

The reasons for installing the Pall filter are two fold: 1) to remove particulate from the gas stream prior to the gas being quenched and scrubbed for acid gases in a water-based system and 2) to protect the downstream, electrically heated experimental afterburner from particulate plugging. Removing a high amount of particulate as close as possible to a thermal mixed waste (radionuclides present) treatment system primary combustion chamber will minimize downstream migration of radioactive and other hazardous solids to the remaining downstream air pollution control (APC) equipment.

The primary specific test objective for the Pall filter was to determine the particulate and metals removal efficiency at high offgas temperatures (i.e, 1,600 °F.) The initial selected feedstock composition is shown in Table III. This feedstock was developed by the MWFA and is deemed typical of their wastes. Secondary objectives included documenting reliability, availability, and maintainability data on the unit.

The Pall filter was operated a total of 244 hours under various process feed conditions. A total of 1,000 hours of run time was planned; however, PACT-6 system equipment failures and subsequent time and budget constraints prevented running the total 1,000 hours. Table IV summarizes tests where PACT-6 offgas was processed through the Pall filter.

Due to filter plugging and subsequent high filter differential pressure drop, specific suspect feed ingredients were eliminated successively, and testing proceeded to determine if that particular ingredient was causing the filter to plug. Automatic blowback was increased to the maximum frequency, and manual blowback was initiated during filter plugging conditions. These actions did not have any effect on the filter plugging. Once the filter plugged, it remained plugged and the test operation had to be stopped because of unallowable positive pressure increases in the plasma unit primary combustion chamber.

Late FY97 testing has shown that the Pall filter differential pressure does not become excessive when using a relatively benign feedstock. The benign feedstock constituents included sand, hydraulic oil, carbon and stainless steel, Portland cement, lead, MicroCel E^®, and ceric oxide. It appears the filter plugs when the feedstock contains plastic or compounds containing sodium (i.e., sodium carbonate.)

Operations were secured for operational safety reasons when the differential pressure increased to the point that the primary PACT-6 chamber gauge pressure went positive.

Figure 3, test 97-ORGANIC-1, is typical of a run where the Pall filter plugged. Differential filter pressure is plotted against time.

Table V summarizes the isokinetic particulate sampling results. Sampling was conducted using EPA Reference Methods 1, 2, and 12 for particulate loading and volatile metals. The sampling was performed simultaneously at two locations upstream and downstream of the Pall ceramic filter. The stack typically operates from a negative 15 inches wc to a negative 60 inches wc, and these relatively high vacuums and also nonuniform offgas velocities and compositions create sampling problems not usually encountered when sampling positive pressure stacks. The EPA sampling methods were designed for sampling offgases at positive pressures and fairly steady-state velocities and offgas compositions. The sampling was generally successful after sampling techniques were instituted and some modifications were made to sampling hardware.

Due to some difficulties with the PACT-6 system, some deviations from the EPA Reference Methods were necessary. The deviations were as follows.

• All sampling was performed with the probe positioned in the center of the offgas flow due to sampling hardware limitations.
• The high vacuum experienced in the offgas flow required installation of a shutoff valve in each sampling port, resulting in a deviation from standard procedures.
• Due to high moisture content in the offgas flow, two impingers with desiccant were used instead of one.
• Because of high particulate loading, an S-type pitot tube was used instead of a standard pitot tube.

Some plugging of the sample filter on the upstream side of the Pall filter occurred. The plugging could have been caused by the high particulate concentration at this point in the stream.

Metallic species removal efficiencies were high during the series of tests that incorporated full use of the Pall filter, with the vast majority greater than 99%. This may indicate that the metallic species are solid particles at the test process conditions rather than metal vapors at the inlet to the filter.

The removal efficiencies are displayed as a bar graph in Fig. 4. The graph is based on the data from June 26, July 10, and November 5, when a nonzero amount of particulate was collected downstream of the filter and the elemental analysis of both upstream and downstream particulate was performed.

As tested on the PACT-6 System with the specified feedstocks and resulting process conditions, the Pall filter seems to be very sensitive to either flow excursions, feed chemical makeup, or both. It appears the PACT-6 system plugs the filter with particulate overload in general. Another possible chemical reason is discussed below.

Water vapor condensation was probably not the cause of plugging because the temperatures of the filter were in the ranges of 800 °F to 1,200 °F at the inlet and 400 °F to 600 °F at the outlet. With extended operation, this temperature difference may be reduced.

Chemical analysis of the particulate taken from the candles after the filter plugged during the May 1997 testing showed the main difference in compositions of the particulate, compared to that of normal conditions, was an elevated concentration of lead (27% by weight at plugging; trace amounts at normal conditions) and chloride and sulfate anions (4.1%, 5.1% by weight respectively, trace amounts at normal conditions). In all probability, the main cause of the plugging in this case was condensation of a mixture of metal salts and/or oxides with a formation of a glassy film on the surface of the precoating layer of the filter. For example, lead chloride has a melting temperature of 930 °F and can condense on the candle surface because the concentration of lead and chlorine in the vapor phase of the offgas is sufficiently high, 160 parts per million by volume (ppmv) average for the lead vapors and even higher for chlorine.

Operating with the filter immediately downstream of the PACT-6 primary chamber causes back pressure excursions (i.e., positive pressure in the primary chamber) that could be unacceptable in a system processing hazardous or nuclear waste, especially in a configuration similar to the MSE test train. The advantage of the filter in the process train is that the scrubber blowdown water from the quencher and acid gas scrubber contains at least 99% less particulate carryover; therefore, it is less likely to contain hazardous metals. This could significantly reduce or eliminate present hazardous scrubber blowdown wastestreams.

A test should be conducted with the filter installed downstream of a gas-fired secondary combustor that uses oxygen instead of air for combustion. Excessive back pressure excursions may be mitigated by the dampening effect of the increased volume and by dropping out most of the larger particulate prior to entering the filter. Also, the possibility of any products of incomplete combustion (PICs) carryover to the filter would be reduced. Some of these PICs could be dioxin, which are difficult to test for and which require disposal.

Another recommendation is to route the present offgas in a water-cooled vertical pipe directly above the chamber. The particulate could drop back into the primary chamber where it would adhere to the side wall and be periodically removed. This would lessen the total particulate solids loading on the Pall filter. The configuration design of this pipe would be based on appropriate preliminary modeling of particulate in offgas. To further remove larger particulate, a high temperature cyclone could also be installed downstream of the water-cooled riser pipe and upstream of the Pall filter.

Another idea for future hot filter testing is to substitute different filter materials or media into the Pall filter. This could be performed in conjunction with the different feedstocks.

Work was conducted through the DOE-EM Office of Science and Technology at the Western Environmental Technology Office under DOE Contract Number DE-AC22-96EW96405.