Carbon Filtration for Reducing Emissions from Chemical Agent Incineration (1999)

The carbon filter system chosen by the Army for the pollution abatement system filter system (PFS) is a fixed-bed design. The Army also briefly considered two alternative filter processes: a carbon-injection system; and a catalytic oxidation system. Following a discussion of available types of adsorbents; these alternative processes are briefly described.

The activated carbon chosen for evaluation in the experimental work performed by Mitretek was a coconut-shell carbon— the same material used by the Army for ventilation air filters at chemical disposal facilities. The Army has indicated a preference for this type of carbon simply to avoid the need to stock various types. Other types of carbon are made principally from coal. The Army's choice for the PFS was not known at the time this report was prepared, but the committee suspects that it will probably be coconut-shell carbon. Based on the Mitretek analysis and the Army's experience with air filtration, coconut-shell carbon will probably perform satisfactorily. Whether other types of activated carbon might perform better (e.g., with less dust formation or better adsorption) could only be determined on the basis of operating experience.

A number of synthetic adsorbents have been developed over the last 30 years (e.g., several zeolites), and remarkable selectivities have been observed for specific separations that are otherwise difficult. However, carbon has remained the universal choice for treating flue-gas. This is probably because of economics, particularly when the material will be used only once (i.e., with no regeneration). The Army prefers not to attempt to regenerate carbon that is significantly contaminated with agent.

The fixed-bed system uses carbon particles of 8 to 16 mesh (1.2 to 2.4 mm)— particulates large enough to limit the pressure drop through the beds to a reasonable level. An alternative is to grind the carbon to a much finer size (e.g., passing 200 mesh [0.074 mm]) and injecting this "micronized" carbon directly into the gas flow to provide a few seconds of contact time. The system approaches equilibrium in a matter of seconds because of the small particle size. That is, mass transfer rates increase rapidly as particle size is reduced.

The solids are then separated from the gas using either an electrostatic precipitator or a bag filter. In practice, carbon-injection systems are frequently retrofited systems to existing facilities, and the separation device is the one that was installed originally. If bag filters are used, the micronized carbon may thinly coat the surfaces, increasing gas/solid contact as gas flows through this layer. This process works basically the same way as the fixed-bed design chosen by the Army (i.e., with uniform flow of gas through a carbon bed). Adequate contact can be obtained even with very thin beds of fine particles.

The carbon-injection process evaluated by the Army contractor involved a thin layer of carbon coating on a bag filter. The coating was assumed to be 0.2 inches thick, which is comparable to the filter cake maintained on pulse-jet filters. The process was considered to have the following drawbacks compared with a "conventional" carbon bed:

• Handling the micronized carbon introduces problems associated with housekeeping and potential respiratory problems in the event of a failure in a filter element. This problem would be compounded because the carbon dust would be contaminated with substances of potential concern (SOPCs). In addition, finely divided activated carbon dust presents a fire and explosion hazard.

• The process requires that a layer of uniform thickness be built up on the bag filter to provide uniform treatment for all of the gas. If the thickness varies, gas will pass preferentially through the thinner sections. It was not clear how uniform thickness would be achieved.

• More carbon consumption is required in the dry sorbent-injection design than the fixed-bed design (see Appendix E). Thin layers on bag filters would have to be replaced frequently, during which breakthroughs would have to be avoided while acceptable flow and pressure were maintained.

Catalytic oxidation is conceptually a simple process. The output from the mist eliminator of the baseline system would be preheated and passed through a fixed bed of a platinum catalyst chosen for its resistance to poisoning by chlorine, phosphorus, and sulfur. SOPCs and any residual agent would be oxidized in the catalyst bed. The process would operate at 700ºF (371ºC) or higher, so substantial amounts of natural gas would be required to heat the flue gas to reaction temperature. The flow rates to the blowers would also be much higher because of the higher temperature and the addition of more combustion products from the natural gas; therefore, the blowers would have to be much larger, which would increase energy consumption. All of these factors would add to the cost of the facility, making the system more expensive than the fixed-bed carbon filter system.

The catalytic oxidation system's demonstrated very high levels of destruction (e.g., 5 nines) was also questioned by the Army contractor. This destruction level would be extremely difficult to demonstrate during normal operations, when the materials of most concern in the flue gas— dioxins and furans— would already be down to parts per trillion, or nondetectable levels. Measuring outlet concentrations of one-millionth of the already difficult-to-measure values would require sampling times longer than the facility operating life. In the event of an incinerator upset that induced higher flue gas concentrations, a fixed activated carbon bed would be able to handle the problem. The ability of the catalytic oxidation unit to do so was judged to be less certain, although foreseeable upsets should produce inconsequential increases in SOPCs. Another concern about the catalytic oxidation system was the potential for poisoning the catalyst, thus reducing its effectiveness.

Catalytic oxidation has one potential advantage over the fixed carbon bed— it would oxidize a lot of the SOPCs that would not be captured by the carbon because they are not strongly adsorbed. Health risk assessments, however, indicate that most of the risk is associated with chlorinated dioxins and furans, which are strongly adsorbed. The volatile materials that might escape a carbon bed would be at such low concentrations that they would contribute very little to the health risk.

Based on differences in operating experience and costs, the Army's decision to choose carbon bed filtration over catalytic oxidation appears to be sound.