Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
5 Burns and Roe Technology Package INTRODUCTION AND OVERVIEW The Burns and Roe team's technology package is shown schematically in Figure 5-1. This package uti- lizes the Army's baseline disassembly technology to separate the components of munitions (e.g., chemical agent, fuzes, bursters, metal casings, etc.), with subse- quent treatment by high-temperature plasma to decom- pose the chemical agents, propellants, and wooden, fi- berglass, or plastic packing materials. All metals, including the munition casings, are melted by the plasma. An explosion chamber is used to deactivate explosive components by energetic initiation (detona- tion or deflagration). Debris and gas from the explo- sion chamber are then also treated using high-tempera- ture plasma. The technology provider's approach for performing the required major demilitarization opera- tions is summarized in Table 5-1. Because plasma waste treatment, which is integral to the proposed system, is a unique technology, plas- mas and their characteristics are discussed first. The Burns and Roe technology package is then described in detail. Background on Plasma Electric arcs and discharges have been of interest to scientists and engineers for decades because they in- volve high-temperature, conductive gases (plasmas). Typically, an arc can be established between two con- ducting electrodes (e.g., graphite or metal) in a variety of atmospheres. The plasma is comprised of molecules, atoms, ions, and electrons at temperatures of 1,000°C 71 to 20,000°C (1,832°F to 36,032°F) depending on the current and voltage, the gaseous environment, and the pressure of the constricting gas. Either physical or mag- netic constriction can be used to increase temperatures. Because plasma arcs between electrodes generally involve voltage drops of 100 V or more, chemical bonds (whose strengths range from 2 to 10 electron volts LeV]) will be broken, and ionization processes (at 4 eV to 25 eV) will occur. Thus, material exposed to a plasma environment will be transformed into atoms, ions, and electrons, with only a few molecules remain- ing. This makes the potential use of plasma arcs, torches, melters, and other plasma devices attractive for destroying undesirable molecules (e.g., hazardous wastes). High-temperature plasmas can also produce endothermic neutral species (e.g., C2H2, C2N2, and NO) or gaseous molecular ions (e.g., SiO+ and CO+. When the plasma is cooled to room temperature, most of the molecules are thermodynamically stable, but some metastable species (that are stable at higher tempera- tures but unstable at lower temperatures) might sur- vive. In addition, metastable species could be formed during cooling, which could also be present at room temperature. DESCRIPTION OF THE TECHNOLOGY PACKAGE Disassembly of Munitions and the Removal of Agent/Energetics The technology provider proposes using the baseline approach to disassemble munitions and segregate the agent, energetic materials, and munition bodies. (See
72 Munitions | Baseline ~Agent ~ disassembly - Munition unpack '1 Fuzes and bursters Energetic destruction chamber Rocket motors Dunnage Metal parts Plasma waste converters C1 and C2 · Plasma waste converter · B Residue from fuzes and bursters Plasma waste converter D FIGURE 5-1 Schematic diagram of the Burns and Roe technology package. Appendix C for a description of the baseline disassem- bly system.) The only modifications to the baseline dis- assembly process, which occur subsequent to removal of agent and energetics, are: (1) limiting the number of munition bodies per tray to nine, and (2) modifying the conveyors to accept smaller trays. Description of the Plasma Waste Converter The Burns and Roe technology package uses spe- cialized plasma waste converters (PWCs) to treat all materials, including chemical agent. Six PWCs of four different types are proposed: two to treat agent and munition bodies (PWC A1 and PWC A2), one to treat pieces of rocket motors containing propellant (PWC B), two to treat metal parts to the 5X standard (PWC C1 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS Plasma waste Pollution converters · abatement system A1 and A2 A Brine to treatment Pollution · abatement system Brine to treatment · 5X metals to off-site facility Pollution abatement system B Brine to treatment _ 5X metals to off-site facility Pollution · abatement system · Brine to treatment Gas to hold-test-release system (may be burned in boiler) and PWC C2), and one to treat dunnage (PWC D). The basic operation of all the PWCs is the same. A typical PWC (Figure 5-2) is a cylindrical, refrac tory-lined vessel with an opening in the roof through which a plasma torch is inserted. (For larger PWCs, more than one torch may be used through more than one opening.) There are no airtight seals between the plasma torch and the vessel roof, and the PWC is oper- ated at slightly negative pressure to prevent gas from exiting through the opening. Thus, air is always leak- ing into the PWC. Each plasma torch is a non- transferred torch1 consisting of a cylindrical pipe con- taining water-cooled copper electrodes. The plasma IIn a nontransferred torch, both the anode and cathode are contained within the torch.
BURNS AND ROE TECHNOLOGY PACKAGE TABLE 5-1 Summary of the Burns and Roe Approach Major Demilitarization Operation Approach(es) Disassembly of munitions Army baseline disassembly process. Treatment of chemical agent Thermal destruction using plasma waste converter (PWC). Treatment of energetics Initiation of explosives in explosion chamber (residual passed through PWC); destruction of propellant in PWC. Treatment of metal parts SX treatment of metals (complete melting) in PWC. Treatment of dunnage SX treatment of dunnage in PWC. Disposal of waste Solids. Slag for recycling; salts from scrubbers to appropriately permitted landfill. Liquids. None. Gases. Hold and test; feed to boiler or thermal oxidizer (combustion) if test results are acceptable. feed gas passes through the torch, and the plasma is formed in the torch between the anode and cathode. No other materials, such as agent or energetics, are intro- duced into the plasma torch. The torch creates a plasma with a temperature, as reported by the technology provider, in the range of 15,000°C (27,032°F).2 The plasma exits the torch into the PWC chamber and impacts onto solid and liquid material (e.g., metal from weapons) at the bottom of the chamber. In an agent-treatment PWC, agent is in- troduced into the hot plasma near the bottom of the PWC chamber. Steam is introduced with the agent at a controlled rate to convert elemental carbon or soot (cre- ated by dissociation of the feed stream molecules) to CO. The plasma exiting the torch cools very quickly in the chamber by a combination of the following mechanisms: · mixing with infiltration air · sensible heat required to heat the waste feed to the PWC · the endothermic chemistry of degradation of the agent and other organic waste materials introduced into the PWC · decomposition of the steam introduced for soot control · decomposition of the CO2, if used as a plasma feed gas 2Because plasmas can contain molecules, atoms, ions, arid free electrons, several plasma temperatures can be defined. The committee considers the temperature listed here to be reasonably representative of the very high temperatures of the plasma components. 73 · formation of NOX from nitrogen in the weapons material feed, the plasma feed gas (if N2), or infil- tration air · heat losses through the PWC shell For the demonstration system, the technology pro- vider indicates that the air in-leakage rate is approxi- mately 30 standard cubic feet per minute (SCFM) (Burns and Roe, 1999), compared to a plasma feed gas flow rate of approximately 20 SCFM and a PWC total gas outflow on the order of 140 SCFM (Burns and Roe, 1998b). The temperature of the exit gas for the PWC is Air infiltration \, 1 Mixed bulk gases Rae Chemical agent and steam Plasma feed gas ~1,100°C product gas (Ar, CO2, or N2) (to pollution abatement Plasma torch I // I\ Plasma at ~L:~ Molten slap pool (if present) FIGURE 5-2 Schematic diagram of a typical plasma waste converter (PWC) for treating agent.
74 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS estimated to be approximately 1 ,1 00°C (2,01 2°F) (Burns and Roe, 1997~. Treatment of Chemical Agent After being drained from the munitions, chemical agent is pumped to storage tanks and, subsequently, into one of two PWCs designated Al or A2, identical units approximately 7 It in diameter and 9 It high (ex- ternal dimensions) designed for liquid feed of up to 1,200 lb/hr. Agent and steam are injected into the PWC where they mix with the hot plasma. Product gas then passes through a pollution abatement system (PAS) consisting of four major components: · a vertical down-flow duct with water/caustic spray to quench the gas · a countercurrent, multistage acid-gas scrubber with vertical upward gas flow and gravity-driven, . downward flow of scrubbing liquid · a mist eliminator a cartridge filter for the removal of fine particulates Upon exiting the PAS, the product gas enters a hold- test-release system consisting of a compressor, a con- denser, and storage tanks. The tank contents are sampled for chemical agent and other components (not yet specified). If no agent is detected, the technology provider plans either to (1) burn the gas in an on-site boiler or oxidizer or (2) sell it as a fuel. If agent is detected, the gas is recycled to PWC Al or A2 for reprocessing. Treatment of Energetics Two general types of energetic materials will be treated: M55 rocket propellant (designated M28), which is configured to burn rather than detonate; and burster and fuze materials, which are intended to deto- nate upon initiation. MSS Rocket Propellant After being sheared, the severed M55 rocket motor pieces containing the M28 propellant are fed into PWC B 7 It in diameter and approximately 9 It high (exter- nal dimensions) designed to destroy energetic materials at rates of up to 1,500 lb/hr. The unit is fabri- cated of 2.5-inch-thick carbon steel to contain any ac- cidental explosions. PWC B also includes a feed chute and hydraulic ram for introducing the waste material into the vessel. The waste materials mix with the hot plasma, and product gas is discharged to the same PAS and hold- test-release system described previously. Some rocket- component materials (e.g., metal parts and fiberglass shipping and firing-tube pieces) do not remain in the plasma field long enough to vaporize but melt, forming a molten pool at the bottom of the vessel. The metals are tapped and drained to form ingots. Nonmetallic slag is also periodically tapped and drained. Other Energetics To lower the risk of detonations in the PWC, the technology provider proposes deactivating explosive material in an explosion chamber. This chamber is commercially available (designed and manufactured by Bofors) and is used by both the military and industry to deactivate small quantities of explosives. It is made of thick high-strength steel and is designed (1) to with- stand multiple detonations of a specific mass of TNT, and (2) to contain the product gas from the detonations. Both the quantity and type of energetic material treated per batch must be known in advance to ensure that the unit's explosive rating is not exceeded. In this application, the bursters and fuzes from rock- ets, projectiles, mortars, and land mines are fed into the explosion chamber, where they are thermally initiated. The gas from the explosion chamber is then slowly vented to PWC B for further treatment. The solids are removed and are also fed to PWC B. The gaseous and molten products from PWC B are treated in the same way as the M55 rocket products discussed in the pre- ceding section. Treatment of Metal Parts Bodies of projectiles and land mines, drained of agent and emptied of energetic materials, are placed in trays and conveyed to PWC C1 or C2 identical PWCs designed to decontaminate and melt metal parts. The proposed units are 7 It in diameter and 9 It high (exter- nal dimensions), with a peak capacity of 6,000 lb/hr of
BURNS AND ROE TECHNOLOGY PACKAGE metal munition bodies. Trays of munition bodies are moved by roller conveyor to a scissor lift that raises the trays to a feed chute. The tray is conveyed into the feed chute, and a ram pushes the tray containing the muni- tion bodies into the PWC where both are consumed. Gaseous products flow through the PAS and hold-test- release system described previously. Molten metal col- lects in the bottom of the PWC, which is tapped peri- odically and the metal cast into ingots. This metal will be considered as treated to 5X condition because it will meet the criterion of "heated to at least 1,000°F for at least 15 minutes." Treatment of Dunnage Dunnage is gravity-fed to a shredder for size reduc- tion and fed to PWC D (approximately 5 It in diameter and 5 It high [external dimensions]) designed to de- stroy dunnage at rates of up to 1,200 lb/hr. Gaseous products flow though the PAS and hold-test-release system described previously. Molten materials col- lected in the bottom of the PWC are periodically recov- ered. The technology provider proposes to control the formation of graphite, soot, and other carbonaceous material by adding steam to form CO. Process Instrumentation, Monitoring, and Control Monitoring of "traditional" process variables (e.g., temperatures, liquid and gas flow rates, etc.) is accom- plished using standard, off-the-shelf, chemical-process equipment and instrumentation. Monitoring for agent is accomplished using the ACAMS and depot area air monitoring system (DAAMS) developed by the Army. Agent feed to the PWC is monitored for flow rate and pressure, with real-time signals relayed to the con- trol room. Automatic feed cutoff valves are employed if operational ranges (yet to be established) are vio- lated. Sensors in the energetic feed chute and the muni- tion metal-body conveyor detect blockage of feed ma- terial and initiate appropriate action, which could include PWC shutdown. (Because the feeds are not critical to PWC operation, automatic shutdown because of a feed blockage is not included.) Inside the PWCs, power feed, plasma feed gas flow, vessel pressure and temperature, and steam flow are 75 monitored and controlled to within established operat- ing ranges. Appropriate pressure, temperature, and flow-sensing instrumentation is used to gather and transmit the information to the control room. Any de- viation from the established limits of any of the param- eters cited above results in automatic PWC shutdown. The product gas leaving the PWC is cooled, com- pressed, and collected in tanks downstream of the PAS. This gas is then sampled and analyzed for chemical agent using both ACAMS and DAAMS agent monitors. Feed Streams In addition to the munitions and packing materials, the following materials are fed to the system: · the plasma feed gas, which acts as the plasma me- dium to the PWCs (argon was specified in the technology provider's proposal; N2 and CO2 are being used in the demonstration unit) · steam to the PWCs to convert elemental carbon to CO · caustic to the PAS to support the quenching and scrubbing operations ~ make-up water to the PAS to support the quench- ing and scrubbing operations The technology provider has generated a mass balance for the proposed system using the maximum possible feed rates to each PWC. This results in a "mix-and- match" configuration that is not representative of any particular munition campaign. Nevertheless, the infor- mation reflects the sizes of anticipated flows. Process inputs are summarized in Table 5-2. Waste Streams The waste streams from the system will be either gaseous or solid. There is no liquid waste stream. (Wet scrubbers are used to absorb and neutralize products like HC1, HE, SO2, and P4O~o using a caustic solution, but the scrubber liquid is subsequently evaporated, leaving a salt cake.) Gases The product gas from the PWCs is quenched in a vertical down-flow duct with water/caustic spray. It
76 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS TABLE 5-2 System Inputs for the Burns and Roe So/ids Mass Balance Scrubber Brine Wastes. The brine from the PAS StreamFlow (lb/hr) scrubber is processed to recover salts using the baseline brine reduction system. This system dries the brine with Inputs to PWCs Al and A2a HD1,170 rotary-drum dryers to yield low water-content salts (5 Decontamination solution, oil, laboratory waste650 to 15 percent by weight). The water vapor is discharged Steami78 to the atmosphere; the water is not recycled. The salts are sent to an appropriately permitted landfill for dis Inputs to PWC Bb M28 propellant386 posal. (Consideration is also being given to recycling Decontamination solution25 the brine through a PWC.) Comp B464 Fiberglass200 Metal Munition Bodies. The 5X-treated molten Metal parts270 Steam46 metal is drained from the PWCs, cast into ingots, and Argon400 sold for scrap. Inputs to PWCs C1 and C2a HD585 Residues from the PWC Bottoms. Nonmetallic ma Metal parts8,600 serials that collect in the bottom of the PWCs are mixed stegaOm26s with sand in the unit and recovered as a vitrified mate rial (slag). Input to PWC D Miscellaneous dunnage1,165 The process outputs from the technology provider's Steam450 mass balance are summarized in Table 5-3. The pre Argon200 dieted compositions of the gaseous effluent streams Input to pollution abatement systemsfrom the PWCs are shown in Table 5-4 and the pre Make-up water7,997 ' Caustic (30 percent NaOH)5,867 dieted compositions after scrubbing are shown in Table 5-5. (The compositions would be different if a plasma Total mass input to system30,118 aloe HD-filled 155-mm projectiles per hr b20 M55 rockets per hr Source: Burns and Roe, 1998a. TABLE 5-3 Mass Outputs for the Burns and Roe System StreamFlow (lb/hr) then passes through a packed tower for acid-gas scrub bing (using caustic) and a mist eliminator for water Outputs from PAS for PWCs Al and A2a removal. Fine particulates are removed by a cartridge product gasi,822 filter. (A venturi scrubber is included in the demonstra- Outputs from PAS for PWC Bb lion system for particulate removal, just prior to the product gasi,036 packed tower [Burns and Roe, 1998b].) Brine83 Following treatment with the pollution control Outputs from PAS for PWCs C! and C2a equipment Just described, the product gas Is held In a product gasi,214 pressure vessel where it is sampled and analyzed for Brine4,864 agent and other components (still to be determined). If Outputs from PAS for PWC Dc the gas is agent-contaminated, it is recycled to the PWCs product gas539 for further treatment. If no agent is detected, (1) the gas Metals/silicates/bottom materials from PWCs9,236 Is used as fuel for an on-s~te boiler, (2) the gas Is shipped off site as fuel for other applications, or (3) the Total mass output from system30,166 gas is burned in an on-site oxidizer. If the gas is burned a 100 HD-f~lled 155-mm progechles per hr in a boiler or oxidizer, gaseous effluents from the boiler b20 M55 rockets per hr or thermal oxidizer are scrubbed and released to the Cmiscellaneous dunnage waste environment via a traditional stack. Source: Burns and Roe, 199Sa.
BURNS AND ROE TECHNOLOGY PACKAGE TABLE 5-4 Predicted Composition of Product Gas from the Plasma Waste Converters (Prior to Scrubbings TABLE 5-5 Predicted Composition of Product Gas from the Plasma Waste Converters after Scrubbinga PWC PWC PWC PWC PWC PWC PWC PWC Compound A1 end A2b BC C1 and C2b Dd Compound A1 end A2b BC C1 and C2b Dd 77 COS 22.08 9.57 COS 22.08 9.57 CO 706.01 385.43286.61 783.42 CO 706.01 385.43286.61 783.42 C2H2 89.98 36.56 C2H2 89.98 36.56 H2CO 19.65 7 99 H2CO 15.72 6.39 CH417.39 7.827.0715.89 CH417.39 7.827.0715.89 H2O95.41 36.5544.068.22 H2O5.28 2.672.725.23 H290.79 29.6134.40102.64 H290.79 29.6134.40102.64 Ar800.00 400.00800.00200.00 Ar800.00 400.00800.00200.00 Total2,797.68 1,080.901,650.011,449.46 Total1,822.18 1,035.841,214.111,400.05 aPlasma feed gas is argon; temperature is approximately 1,100°C (2012°F); all quantities in lb/hr. bloo HD-filled 155-mm projectiles per hr C20 M55 rockets per hr dmiscellaneous dunnage waste Source: Burns and Roe, 1998a. feed gas other than argon were used; CO2 and N2 are being tested in the demonstrations.) Start-up and Shutdown PWC start-up procedures involve first initiating the plasma feed gas flow, the torch cooling-water flow, and the plasma-torch power. The vessel is then allowed to reach the prescribed operating temperature, which usu- ally takes approximately two hours from a "cold" start- up. Weapons material feed and steam flow (to convert elemental carbon to CO) are then initiated. The tech- nology provider estimates that once the vessel reaches the desired temperature, steady operation at capacity can be achieved in approximately five minutes. Upon shutdown, the weapons material feed is stopped, the steam flow is cut off, the power to the torch is turned off, and the plasma feed gas flow is stopped. The torch cooling water continues for approxi- mately 30 minutes to protect the electrodes during cool down of the vessel. aall quantities in lb/hr bloo HD-filled 155-mm projectiles per hr C20 M55 rockets per hr dmiscellaneous dunnage waste Source: Burns and Roe, 1998a. EVALUATION OF TH E TECH NOLOGY PACKAG E Process Efficacy Effectiveness Because assembled chemical weapons contain mainly the elements C, O. H. N. S. P. halogens, and various metals (e.g., A1, Fe, Co, Ni), one can predict from standard thermodynamic calculations that CO, H2, CO2, H2O, H2S, HC1, HF, N2, NOX, SOx, and vari- ous metal oxides will be formed at ambient tempera- tures (e.g., 20°C to 35°C; 68°F to 95°F), depending on the availability of oxygen. Compositions at higher temperatures can also be calculated, and Tables 5-6 and 5-7 show the mole fractions of the equilibrium prod- ucts at 2,227°C (4,040°F) for various agent and ener- getic feed materials predicted by the technology pro- vider, assuming argon is the plasma feed gas (Burns and Roe, 1997~. These results were not generated via testing but were calculated by a chemical-equilibrium computer program (IVTANTHERMO) that uses stan- dard thermodynamic data (Burns and Roe, 1997). Most
78 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS TABLE 5-6 Theoretical Equilibrium Composition of Product Gas from Plasma Treatment of Agentsa Mole Fraction Species Produced GB Feed VX Feed TABLE 5-7 Theoretical Equilibrium Composition of Product Gas from Plasma Treatment of Energeticsa Mole Fraction HD Feed Species Produced TNT Feed Tetryl Feed NC Feed HE PO N2 SH S.Ox 10-2 3.9x 10-2 l.Sx 10-2 1.1 x 10-2 6.8 x 10-3 N2 SH H 1.2x 10-2 3.2 x 10-3 2.2x 10-2 3.1 x 10-3 3.3 x 10-3 1.8 x 10-2 1.3 x 10-3 3.1 x 10-2 S SO P2 PO2 4.4 x 10-3 1.3 x 10-3 2.5x 10-3 1.7 x 10-3 1.2 x 10-3 PN PH2 P2O3 1.3 x 10-4 6.5x 10-5 4.1 x 10-5 2.9 x 10-5 3.6x 10-5 C2H2 COOH H2CO C1 C2H CH4 CH3 2.1 x 10-2 9.3 x 10-3 6.1 x 10-3 3.6 x 10-3 9.1 x 10-5 3.5x 10-5 2.9x 10-5 2.5x 10-5 aPlasma feed-gas is argon; temperature is 2,230°C (4,040°F). Source: Burns and Roe, 1997. of the higher temperature species should revert to CO2, H2O, CO, and H2 when cooled. However, the product distribution and composition might be partially kineti- cally controlled rather than thermodynamically deter- mined. Thus, species that are stable at higher tempera- tures but metastable at lower temperatures might be present after cooling. C3H C2H HCN CN 9.0X 10-5 1.6x 10-5 4.0x 10-3 3.9x 10-5 a~l r 1 · · O O 5.5 X 10-2 rlaSma reea-gas 1S argon; temperature 1S 2,230 C (4,040 F). Source: Burns and Roe, 1997. The plasma temperature in the PWCs is estimated by the technology provider to be in the range of 15,000°C (27,032°F). Although one would expect very high destruction efficiencies at this temperature, much of the agent may not be exposed to such a high tem perature in the PWCs. The plasma arc is created in an enclosed torch through which only the plasma feed gas (e.g., argon, N2, or CO2) flows. The arc heats the gas, which ionizes, dissociates, and then flows into the chamber surrounding the torch. The chemical agent is injected into the side of the chamber, not through the torch (see Figure 5-2). Inside the chamber, the agent mixes with the plasma. The maximum temperature to which each agent molecule is exposed is unknown, but the temperature gradient within the chamber is very large, as is evidenced by the estimated gas exit tem perature of 1,100°C (2,012°F) (Burns and Roe, 1997 ). The heterogeneous conditions in the PWC could cause organic intermediates to form. The prevailing view is that organic intermediates3 are formed during an initial vaporization and pyrolysis phase, prior to 3In combustion systems, these organic intermediates are usually referred to as products of incomplete combustion (PICs). Typical examples include benzene, toluene, and naphthalene.
BURNS AND ROE TECHNOLOGY PACKAGE oxidation (Dellinger et al., 1986; Glassman, 1987; Dempsey and Oppelt, 1993~. If complete mixing of the intermediates and stoichiometric quantities of oxidant occurs, all of the intermediates would be oxidized. However, in practice, the mixing will not be perfect, and some intermediates may bypass the plasma zone without being complete oxidized, even if sufficient oxidant is present. In addition, larger particles may not have sufficient time to decompose completely in the plasma zone. This imperfect mixing might also allow a small fraction of the original fuel or waste to pass down- stream intact, without being pyrolyzed or oxidized. The oxidant in the PWCs may come from the plasma feed gas (if air or CO2 is used), or it may come from air in-leakage into the PWC from the surrounding room. For the demonstration system, the technology provider has indicated that the air in-leakage rate is approximately 30 SCFM (Burns and Roe, 1999), compared to a plasma feed gas flow rate of approximately 20 SCFM and a PWC total gas outflow on the order of 140 SCFM (Burns and Roe, 1998b). Thus, the air in-leakage forms a significant fraction of the total gas flow for the demonstration unit. If the available oxygen is less than the amount theoretically required for the complete oxidation of organics (includ- ing organic intermediates), then the lack of oxygen, coupled with incomplete mixing, could lead to signifi- cant quantities of organic intermediates being formed and passed downstream from the PWCs to the PAS. For the reasons described above, the committee doubts that all of the chemical agent feed would actu- ally be exposed to the ultra-high temperature plasma and believes that some residual toxic materials would remain or form. Nonhomogeneous temperature distri- butions, gas turbulence, and incomplete mixing may limit the absolute effectiveness of this process (i.e., the target destruction efficiencies would not be achievable at the required throughput rates). Actual chemical agents and munition components must be processed in the proposed PWCs (or other units of similar size and design) to prove the efficacy of the process and to opti- mize design parameters, such as flow rates, reduction/ oxidation conditions, and residence times. Design pa- rameters will be very sensitive to equipment configu- ration, scale-up, plasma feed gas, and the type of chemical weapon and feed rate. The results of the dem- onstration tests, which are being performed at a reduced scale, could provide some (but not all) of the data to address these concerns. 79 The technology provider supplied a comparison be- tween the theoretical products and the products mea- sured by GC (gas chromato~ranhv) for a mixed feed of . . . .. . ", ~ ,, polyethylene, cellulose, water, and air to a PWC of unspecified size (Burns and Roe, 1997~. This is shown in Table 5-8. The technology provider states that "due to the limitations of gas chromatography, the samples were analyzed only for the primary components of the [product gas]." Thus, approximately 6 percent of the gas was not accounted for by the GC analysis, and there are discrepancies between the theoretical and measured concentrations. Regarding these discrepancies, the technology provider observed that "the computer tends to underestimate hydrocarbon species and oxygen, while overestimating carbon monoxide and hydrogen" (Burns and Roe, 1997~. This statement reinforces the committee's concern that hydrocarbon species not pre- dicted by equilibrium calculations (including trace or- ganic species that are of environmental concern) could be present in the product gas from the PWC. The technology provider has indicated that soot for- mation in the PWC could be significant. This phenom- enon is predicted by the thermodynamics. If CO2 is TABLE 5-8 Comparison of Experimental and Predicted Gas Compositions Subsequent to Plasma Treatmenta Mole Fractions From Gas Chromatography Predicted by Equilibrium Calculations CH4 co2 C2H2 C2H4 NH3 N C Other Total 3.12 3.87 0.61 Coo 2.12 Coo Coo na na na na 93.8s l.Ox 10-6 1.0x 10-7 2.0 x 10-12 0.8s 00.00 aPlasma feed-gas is air; waste/air feed is approximated by C2sHlo4o44Nl7 bThe technology provider did not analyze for these compounds. Source: Burns and Roe, 1997.
80 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS used as the plasma feed gas, it is also possible for soot to form if the resulting CO disproportionates to C(gr) + CO2. The technology provider has proposed adding steam to the feed to convert elemental carbon to CO. This approach has been demonstrated in various plasma systems used for processing municipal waste. How- ever, careful control of the steam-to-carbon ratio would be required to control the formation of soot. Otherwise, there could be significant emissions of soot. In prac- tice, this control could be difficult to achieve with waste streams of various compositions. The committee was informed by the technology pro- vider that argon was not used for the demonstrations because it was too expensive (Burns and Roe, 1999~. Burns and Roe had planned to use air as the plasma feed gas until it was suggested that members of the Dialogue might consider the use of air to be akin to incineration (Hindman et al., 1999~. The technology provider then tested CO2 as the plasma feed gas but ultimately decided to use N2. Because different plasma feed gases have different thermodynamic and chemical properties, the choice of the plasma feed gas can have a significant impact on the performance of the system. For example, the power requirements will vary with the plasma feed gas. Electrode wear may also depend on the type of gas, and the composition of the product gas will certainly vary. Therefore, tests performed with one plasma feed gas may not be indicative of perfor- mance with a different gas. The volatile low molecular weight chlorinated hy- drocarbons in mustard that can be difficult to destroy are not expected to pose a special difficulty for the plasma treatment units, although this has yet to be demonstrated. Sampling and Analysis From the responses to the data-gap questions (Burns and Roe, 1998a), it appears that the technology pro- vider expects to use the same sampling and analytical procedures being used in the baseline incineration sys- tem. These are probably adequate. Maturity Research on plasma-arc technology dates back to the early 1900s, and many practical industrial applications have been developed, including arc melting of metals; electric arc welding; plasma processing of ores; plasma spraying of metallic or oxide powders; and plasma gen- eration of atomic, ionic, and molecular spectra for ana- lytical systems. Plasma arcs have also been used for treating hazardous wastes. According to the General Accounting Office (GAO), research plasma-arc furnaces have ranged from 2 to 8 It in diameter, with power levels of 150 kW to more than 1 MW (GAO, 1999~. Wastes treated include solvents, paint, batteries, incin- erator ash, and radioactive materials. Of the research initiatives by the U.S. Department of Energy and DOD over the past 10 years on plasma treatment of hazardous waste, two have reached the implementa- tion stage: (1) a Navy project to destroy hazardous materials on shore (scheduled for operation in 2000), and (2) an ongoing asbestos destruction project at Port Clinton, Ohio. Other projects are still in the research phase. Although organic wastes have been destroyed using plasma-arc furnaces, much of the research to date has focused on the vitrification of inorganic substances within wastes (e.g., radioisotopes) rather than on the destruction of organic wastes. A subgroup of the committee visited an Ontario Hydro Technologies site (Toronto, Ontario) on April 5, 1998, to observe a prototype PWC and to be updated on progress in the development of equipment for use with real chemical agents and energetics during the ACWA demonstration phase (see Appendix B). The visiting team was shown a basic version of the technol- ogy provider's PWC system. The items observed in- cluded (1) a long pipe with water-cooled copper elec- trodes that operated as a nontransferred DC-plasma torch emitting hot plasma, and (2) a cylindrical furnace system about 6 It tall and 3 to 4 It in diameter (external dimensions) with an opening on the top for batch feed- ing. The electric power requirements (DC) for this unit were on the order of 100 kW to 500 kW, depending on the type of waste and processing rates (e.g., 100 to 500 lb/hr). During the subcommittee's visit, the small DC- plasma torch that fed the plasma into a refractory-lined furnace was used to demonstrate the treatment of the following materials: · a simulated, double-base Propellant (nitro~lvcerin and nitrocellulose) 1 1 ~7 ~
BURNS AND ROE TECHNOLOGY PACKAGE · metals · plastics · household materials All of these materials, which were hand loaded by members of the visiting team, were melted or decom- posed in the plasma discharge, as viewed on a TV monitor. No spectroscopic monitoring devices- optical or mass spectrometric were in operation. A ram feeder and a refractory trough on the device were apparently intended for removing slag. The proposed units for scrubbing SOx and HC1 and for extracting desired products (e.g., CO, H2, and metals or silicate products), which are shown schematically in the sys- tem diagrams, were not present at the Ontario Hydro site. The demonstration system being tested at Edge- wood, Maryland, is designed to perform these func- tions (Burns and Roe, 1998b). The PWCs proposed by the technology provider have never been tested with actual munitions or chemi- cal agents. According to Burns and Roe (1997), tests conducted by Acurex Environmental Corporation at the EPA's Air Pollution Prevention and Control Division of the National Risk Management Research Labora- tory showed that a PWC could destroy simulants of nerve agents, blister agents, and energetics. The PWC tested was a refractory-lined stainless steel vessel sized to process 25 lb/hr of material. It was equipped with a 50 kW to 100 kW, nontransferred, water-cooled DC torch, and the plasma feed gas was argon. The total amount of material destroyed was not reported. In all cases, the destruction efficiencies were stated to be in excess of 99.9999 percent. Work by MSE, Inc., of Butte, Montana, in 1993 un- der Department of Energy sponsorship and using a plasma centrifugal furnace made by RETECH, Inc., was cited to validate the effective use of a plasma fur- nace for the destruction of MK 72 Mod 5 fuzes. A Startech report on PWC processing of pyrotechnic-con- taminated materials for Ensign-Bickford in October 1995 gave no details about the PWC used or about the quantities of material processed (Burns and Roe, 1998a). Major products were identified by GC. A unit developed at Drexel University used an inductively- coupled argon-based plasma device to process energet- ics and agent simulants (Burns and Roe, 1997~. The full-scale units proposed by the technology 81 provider have not been produced yet. A smaller proto- type, larger than the units at Ontario Hydro, is being tested between February and May of 1999 at an Army test facility in Edgewood, Maryland. Robustness Based on the many practical applications of plasma technology in the industrial sector, it can be considered a robust technology. However, robustness for destroy- ing chemical weapons, and especially large segments of energetic rocket propellants, remains to be demon- strated. Adaptation of plasma devices for the destruc- tion of chemical weapons will require (1) special han- dling equipment for the safe introduction of shells, rockets, and land mines; and (2) further development of the torch and chamber designs to ensure the destruc- tion of agent and the production of effluents that can be scrubbed and burned or converted to slag and sent to a landfill. Meeting these requirements will entail a much more extended development and testing program than the one being undertaken for the ACWA demonstra- tion phase. The program would have to ensure that the energy released from the processing of rocket propel- lants can be controlled. Monitoring and Contro/ The committee was not given detailed design pa- rameters for the full-scale units. A smaller demonstra- tion unit has been installed and is being operated. The technology provider plans to use the monitoring and control systems currently in use at DOD chemical-dis- posal facilities and laboratories. Demonstration testing may show that the proposed monitoring and control strategies are effective, but the committee does not have sufficient information to make an evaluation at this time. Applicability Conceptually, plasma technology is applicable to all assembled chemical weapons and could be used at any of the chemical weapons storage sites. However, the proposed process would have to overcome the engi- neering hurdles described above to treat the various components of assembled chemical weapons.
82 Process Safety ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS The unique equipment proposed by the Burns and Roe team includes an explosion chamber, the PWCs, a PAS, and a product-gas collection and storage system. The pressure and temperature in the explosion cham- ber vary cyclically. When energetic material is being fed, the chamber is at ambient conditions. When the energetic material is initiated, the peak temperature and pressure associated with the deflagration or detonation of up to 2,000 grams of TNT are reached. When the product gas is vented, the pressure and temperature re- turn to ambient conditions. The PWCs operate at very high temperatures but at slightly negative pressure. Although the plasma tem- perature is estimated by the technology provider to be in the range of 15,000°C (27,032°F), the interior PWC wall surface is expected to be much lower (1,650°C t3,000°F]; the temperature of exterior PWC surface will be lower still (about 66°C t150°F]~. The PWC dis- charges molten material into ladles for cooling. The full-scale PAS receives gas from the PWCs at about 1,100°C (2,012°F) and at rates up to 2,000 lb/hr. The gas is quenched, scrubbed, and filtered before be- ing pressurized to 100 psig for storage at ambient tem- perature. The explosion chamber and the PWCs oper- ate in a batch mode, and the product-gas collection system includes a hold-test-release step before the gas is stored in bulk. Thus, despite the uncertainties about adequate exposure of the various input materials to the high temperature of the plasma described earlier, the presence of any chemical agent can, in principle, be determined prior to the gaseous effluent being released to storage for subsequent use. Worker Health and Safety ACWA demonstration tests are planned to confirm that excessive energy will not be released when rocket propellant reacts in PWC B and that the propellant will not detonate. The results of these tests should be thor- oughly analyzed. The explosion chamber and the PWCs, including the feed systems, dunnage shredder, and molten material discharge systems, are operated remotely and are inter- locked. Ease of maintenance should be integral to the PWC design, particularly the replacement of the torch electrodes, the repair of the PWC refractory liner, access to the interior of the PWC, and operation of the molten-material discharge valve. Although worker in- teractions with high-temperature equipment/material or rotating equipment should he minimized hv the con . . . . , _ , trots ant' cosign features, worker hazards will probably be higher in the presence of high-temperature systems than in the presence of lower temperature systems. Other worker hazards include the use of a large amount of argon, CO2, or N2 (all asphyxiants), the pro- duction of pressurized flammable gas, and an electrical power system of 440 to 700 V and 800 amps. None of these hazards is unique, however, and the risks can be minimized with proper precautions. Worker interac- tions with hazardous chemicals will be limited to caus- tic for the off-gas scrubber and acid for the neutraliza- tion of scrubber brine. Public Safety A substantial amount of flammable gas will have to be stored, whether the product gas is burned on site or shipped off site. A large explosion or deflagration in- volving this gas could cause an on-site hazard and, po- tentially, an off-site hazard from the direct thermal ef- fects or overpressure forces. A greater concern is the potential damage from explosions to containment struc- tures that could lead to a release of agent. Explosion hazards are common in industry and can be minimized by good design and operation. Cooling water is circulated through the plasma torch to keep it from melting at the high plasma tempera- tures. A leak in the cooling system could spray water into the plasma. If the leak is sudden, rapid vaporiza- tion could cause a pressure pulse that might overload the downstream gas-handling equipment. Then, un- treated agent could be released into the surrounding room through the torch opening in the top of the PWC. Similar "puffing" hats been Nerved in comhuLstion . . . . equipment when excessive back pressure occurs. If the leak is gradual, the resulting steam would dissociate in the plasma forming hydrogen and oxygen gas that could recombine and explode if the mixture is in the flammable range above its autoignition temperature. The effect of liquid water introduced into a plasma in the presence of other species present in PWCs must be determined before larger scale experiments are per- formed. The normal PWC operating conditions appear
BURNS AND ROE TECHNOLOGY PACKAGE to be outside the flammability range for hydrogen, but the effect of the additional water (from a leak in the torch cooling system) could create an explosive com- bination. (For example, the presence of a steam diluent would raise the autoignition temperature; whereas, an argon diluent would decrease the autoignition tempera- ture EKumar and Koroll, 19951~. If some water is not completely vaporized, it would fall into the molten material at the bottom of the PWC, and a metal-water reaction could create a pressure pulse. These mecha- nisms should be investigated further, unless the prob- ability of the failure of the torch is determined to be very low. The technology provider is aware that torch failure is a concern, and the potential for an explosion has been reduced by the torch design and by redundant flow and pressure controls that would actuate fast-closing valves on the water feed as well as the waste feed in the event of a failure. Testing is planned to validate that agent does not reform and that other hazardous materials (e.g., Sched- ule 2 compounds and dioxins) do not form as the PWC effluent cools. The potential formation of metastable species (e.g., C2H2, HCN, C2N2) that could be quenched from the rapid cooling of product gas should be thor- oughly investigated. Human Health and the Environment Burns and Roe states that there will be no gaseous air emissions and no liquid discharges from the inte- grated system and that the solid waste will consist only of metal ingots, vitrified material, and possibly scrub- ber salts. The technology provider has also indicated that the scrubber salts might be recycled to the PWC and vitrified with sand to produce a very stable solid waste. Thus, it is claimed in the proposal that there would be virtually no impact on human health or the environment. However, the committee has identified some issues that must be addressed during the develop- ment of the integrated process. Effluent Characterization and Impact on Human Health and Environment The primary solid-waste streams include fly ash material caught in filters, scrubber salts from the PAS, 83 and metal ingots and slag from the PWCs. The treat- ment temperature for metal parts is expected to exceed the required 5X conditions; therefore, metal parts treated in the PWCs should receive a 5X designation. The technology provider plans to explore the option of recycling liquid scrubber effluent with fluxing agents, such as lime and sand, in a PWC to generate a vitrified solid waste. The treatment of scrubber liquor by vitrification in the plasma unit has not been proven. The committee's concerns relate to the behavior of salts at high temperature and whether the acid components could be incorporated into the melt without being re- leased. For example, NaC1 salts could react with SiO2 at high temperatures to form gaseous SiCl4; also, NaF salts could react with SiO2 to yield SiF4. The PWCs produce the primary gaseous discharge. The technology provider proposes that this gas will be passed through a PAS to a holding tank. The commit- tee is concerned that a PAS designed for fully oxidized gas may not be as effective for gas generated under reducing conditions in the PWCs. Whether reducing conditions exist will depend on the plasma feed gas. The performance of the PAS must be evaluated at the design operating conditions and for the actual product gas. The committee concluded that some significant de- sign changes may be required to the baseline PAS to optimize its performance for the product gas. For ex- ample, the PAS being used by Burns and Roe in the ACWA demonstrations includes a venturi scrubber to control particulates. Because the product-gas flow rate for the PWC being demonstrated is only on the order of 140 SCFM, the gas velocities may not be sufficient for the venturi to remove particulates effectively. The technology provider presented no data on the effluent characterization if the product gas is burned as boiler fuel. Contrary to the technology provider's claim that there would be no air emissions, the committee concluded that small amounts of acid gases, including NOX, SOx, HC1, and HE, may be generated during the burning of plasma-generated product gas. The NOX and SOX originate from the high-temperature burning of the gas; HC1 and HE are produced with the chlorine and fluorine (originally in the agent) that may not be com- pletely removed by the scrubber (typical scrubber efficiencies are 99.9 percent or greater). In addition, the boiler burner systems must be designed to burn
84 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS relatively low heating-value gas efficiently to prevent the release of unburned organics and other trace spe- cies of environmental concern. . Semivolatile and volatile metals, such as lead, cad- m~um, and arsenic, are expected to volatilize in the PWC and condense downstream as a fine fume of sub- micron-sized particles. Thus, the particulate control device will be challenged by a fine metal fume with particulate sizes that are difficult to capture. In addi- tion, this metal fume would be pyrophoric. The tech- nology provider indicated during meetings with the committee that high-efficiency particulate air (HEPA) filters would be added to the PAS to control the metal fume. This addition will have to he evaluated to deter . . .~ ~. . in-. . . mine ~ a non~amman~e bitter would be required. Also, the pressure drop associated with the additional filter would have to be accounted for in the design of the system. Completeness of Effluent Characterization The technology provider has stated that the operat- ing temperature in the plasma zone is about 15,000°C (27,032°F) and that molecules subjected to this tem- perature will be dissociated into atomic components. However, all components may not be subjected to the temperature of the plasma. The specific characteristics of the product gas depend on the constituents in the waste material being fed to the chamber and the tem- peratures of the plasma and bulk gas. The technology provider currently has no data on the actual character- istics of the PWC effluent gas for the feeds expected from the disassembly of chemical munitions. The tech- nology provider has calculated the thermodynamic equilibrium constituents of the gas with the assump- tion that the gas reached the plasma temperature (or at least a very high temperature). These calculations do not address product constituents from material that does not reach the high temperatures of the plasma zone because of bypass or because of kinetic limitations that allow metastable molecules to persist. A more com- plete chemical analysis of product gas generated at pi- lot scale will be necessary, including both major con- stituents and trace species of environmental concern. After being treated in the PAS, gas from the plasma unit will be passed to a pressurized holding tank. Based on the data in Table 5-5, and assuming compression to 100 psi", cooling to 25°C (77°F), and a holding time of 1 hour, the required tank volume will be approximately 14,300 ft3. The gas will be sampled and analyzed for agents by ACAMS and DAAMS. The technology provider also proposes using a continuous emissions-monitoring sys- tem consisting of a Fourier transform infrared (FTIR) analyzer and an unspecified particulate-matter moni- tor. Once the product gas has been certified to be agent free, it may be used as a boiler fuel. The proposed gas analysis will not analyze for trace organic by-products or metals. Although the FTIR system can measure some species to ppb levels, it may not be sensitive enough to characterize fully trace organic and metallic species of environmental concern on a continuous basis. The committee is also concerned about the deposi- tion of materials on the walls of the hold-up chamber that could be vaporized or resuspended when the cham- ber is evacuated. A rigorous characterization protocol for the hold-test-release system must be developed and validated prior to implementation, regardless of the fi- nal disposition of this gas stream. Effluent characterization and chemical analysis of the product gas are scheduled as part of the ACWA demonstration phase. A careful, meticulous study of the effluent gas will be critical to the evaluation of this technology. The data should include the identification of any organic intermediates that would be included in an HRA. The committee is concerned that the effluent characterization when CO2 or N2 is used as the plasma feed gas will not be valid when argon is used. Testing should be done with the specific plasma feed gas pro- posed for the full-scale system. - - Eff/uent Management Strategy The strategy for effluent management proposed by the technology provider is designed to eliminate all liq- uid discharges and hazardous-waste discharges. Bulk metals are melted in the plasma units and turned into solid ingots, which are expected to meet the 5X decon- tamination criteria. Dunnage and miscellaneous solid waste will be treated on site in a separate PWC. Scrubber discharge, which includes aqueous waste containing salts, may be treated on site in a PWC or processed in a brine reduction system prior to disposal
BURNS AND ROE TECHNOLOGY PACKAGE in a landfill. If a PWC is used, the scrubber discharge liquid will be mixed with sand or other fluxing agents and then fed to the plasma unit where the water will be evaporated and the solids vitrified. No data were pro- vided to the committee on this process, and the com- mittee questions whether the material can be melted without releasing acid components of the scrubber ef- fluent during vitrification as SiCl4 (gas) or SiF4 (gas). If the baseline brine reduction system is used instead, the technology provider will have to demonstrate that excessive particulate emissions from the system will not occur. The Tooele Chemical Agent Disposal Sys- tem brine reduction system failed its environmental testing because of such emissions, and the brine is cur- rently being shipped off site for disposal in a landfill. Management of trace metals is also a potential con- cern. The technology provider did not characterize the fly ash that would be collected in the filters of the PAS, which could contain a large amount of the trace metals volatilized in the plasma unit. This effluent stream will be a hazardous waste that will have to be solidified prior to final disposition. If the formation of soot is not prevented by steam injection, a significant amount of finely divided carbon could be present in the effluent as graphite or soot. In general, the product gas from the PWCs will con- sist of a variety of organic compounds of uncertain composition. The high temperatures and oxygen defi- cient (or even reducing) conditions (depending on the plasma feed gas) lead the committee to believe that many of the compounds that can be present in trace quantities in the emissions from combustion systems will probably be present in higher concentrations in the gaseous streams from this process. Although the tech- nology provider proposes capturing and holding this stream for analysis, the committee believes that this will be difficult. The technology provider has presented no data to demonstrate the feasibility of this type of gas capture, containment, and characterization. The committee also questions the feasibility of burn- ing this gas in a boiler. The high chlorine, sulfur, phos- phorus, and nitrogen content of the raw materials will result in a complex mixture of compounds that will have to be removed from the gas stream prior to burn . 1ng, a difficult, if not daunting, task. The elemental moieties will also create a gas stream with a composi- tion very different from traditional gaseous fuels. The "7 1 85 predicted composition after scrubbing (Table 5-5) in- cludes several toxic compounds listed in the Clean Air Act Amendments. The committee, therefore, believes that this technology may encounter significant difficul- ties in satisfying the risk-assessment and risk-minimi- zation requirements for boilers and industrial furnaces. Resource Requirements The major resource requirements for this process are water, power, argon (or other plasma feed gas), and caustic. During operation, 40 gallons per minute of water, 600 SCFM of argon, and 8500 lb/hr of caustic will be required (Burns and Roe, 1997~. Although the annual consumption of these materials was not esti , .. .. . . . · . .. mated oy tne recnno~ogy provider, tne committee esti- mates that 5,000 hours of operation per year would re- quire 12 million gallons of water, 180 million cubic ft. of argon, and 42.5 million lb of caustic. The technol- ogy provider estimates that the plasma torches will re- quire 6 MW of power. Each PWC requires 0.5 kWh of electrical energy per pound of material feed. The tech- nology provider has used this value to estimate energy requirements for all of the feed types, including agents, energetics, metal components, DPE suits, and dunnage. Environmenta/ Compliance and Permitting Only a few plasma units have received permits for waste processing in the United States to date. The technology provider has not provided a definitive permit- ting strategy for the unit beyond declaring that the sys- tem would not be permitted under RCRA incinerator- permitting procedures. The regulatory definition of an incinerator includes plasma-based treatment systems that burn waste with oxygen in enclosed chambers or uses afterburners. In the Code of Federal Regulations 40 CFR 260.10 Definitions, a plasma or incinerator is defined as "any enclosed device using a high intensity electrical discharge or arc as a source of heat followed by an afterburner using controlled flame combustion and which is not listed as an industrial furnace." Be- cause boilers are industrial furnaces, the proposed con- figuration would probably not be interpreted as an in- cinerator unless oxygen is used as the plasma feed gas. The committee identified two alternative permitting routes that might be followed for a plasma-treatment
86 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS process that generates gases that are subsequently burned in boilers. First, the permitting process could follow 40 CFR 264, Subpart X, Procedures for Miscel- laneous Treatment Units: A miscellaneous unit must be located, designed, con- structed, operated, maintained, and closed in a manner that will ensure protection of human health and the envi- ronment. Permits for miscellaneous units are to contain such terms and provisions as necessary to protect human health and the environment, including, but not limited to, as appropnate, design and operating requirements, detec- tion and monitoring requirements, and requirements for responses to releases of hazardous waste or hazardous constituents from the unit. Permit terms and provisions shall include those requirements of other rules that are appropriate for the miscellaneous unit being permitted. With Subpart X permitting, the permit writer uses the relevant rules as a guide, and permitting authorities are likely to use the most recent incinerator standards as . · . tne appropriate rules, as they generally do with ther- mal-treatment units. If the plasma-generated product gases are burned in a boiler, regulatory authorities could also opt to impose the boilers and industrial furnace (BIF) permitting pro- cedures. In this case, the authorities could regulate the unit as a boiler burning hazardous waste. BIF rules have been developed, and the permitting procedures have been well defined. The EPA has announced plans to develop Clean Air Act Maximum Achievable Control Technology (MACT) standards for boilers burning haz- ardous waste over the course of the next few years. These new standards will probably be in place prior to the construction of a full-scale PWC system. Thus, al- though the PWC system is not likely to be regulated as an incinerator, the permitting procedures would be . . similar. STEPS REQUIRED FOR IMPLEMENTATION The full-scale implementation of this technology will require demonstration with actual chemical agents and weapons. Some of these studies are scheduled for the ACWA demonstration. There is little doubt that the highest plasma-torch temperatures will destroy mustard, GB, and VX, but no testing has been done to demon- strate that the agents remain in the plasma zone long enough to be destroyed. Nor are there detailed analytical data to indicate side reactions or unpredicted products · ~ ~ ~ ~ tnat couth result te.g., dioxins, SOxFy, and OFT. A more thorough evaluation of the proposed tech- nology will be possible when a full-scale PWC design is available for modeling gas flow rates and evaluating the exact placement of nozzles and ports through which munition materials would be introduced into the hot plasma zone. The following list includes the most criti- cal steps the technology provider must take before pro- ceeding to implementation: 1. Determine the effect of sudden water injection into the plasma torch in the presence of argon, nitrogen, carbon dioxide, and other species present in the plasma system. Include an evalua- tion of the effect of gases present in the PWC on the flammability range of hydrogen gas. 2. Determine the likelihood of the release of un treated agent and other hazardous contaminants from the PWC if the gas generation rate is unex- pectedly high (e.g., due to a cooling-water leak, the inadvertent introduction of explosive mate- rial into the chamber, or a rapid deflagration of propellant). 3. Conduct a thorough analysis of the product gas generated from each PWC using the plasma feed gas proposed for full-scale operation. This analy- sis should include the identification of organic in- termediates that would be of concern in an HRA. 4. Establish the efficacy of pollution-control equip- ment in removing hazardous compounds (e.g., NOx, SOx, HC1, and metals) from the product gas. -, 5. Perform a larger-scale demonstration of PWC op eration, that includes the hold-test-release step. FINDINGS Finding BR-1. No tests have been done involving ac . . · . O dual cnemlca~-agent or propellant destruction in a PWC. Tests with agent and M28 propellant were planned for the demonstrations being conducted between February and May of 1999, but no data were available to the committee at the time of this writing. D 1 . . . Finding BR-2. Scale-up from the small PWC units in existence to the very large units proposed is likely to present significant scientific and engineering challenges. .
BURNS AND ROE TECHNOLOGY PACKAGE Finding BR-3. Tests performed with one plasma feed gas may not be indicative of PWC performance with a different gas. Because different plasma feed gases have different thermodynamic and chemical properties, the choice of the plasma feed gas could have a significant impact on the performance of the system. For example, the electrical power requirements will be determined, in part, by the plasma feed gas. Electrode wear may also depend on the type of gas, and product gas compo- sition will vary. Finding BR-4. The technology provider's proposal for recycling the liquid-scrubber effluent through the PWC to vitrify the salts may not be practical. If scrubber li- quor is fed to a PWC, some of the contaminants may simply revolatilize. In addition, NaC1 and NaF salts could react with SiO2 at high temperatures to form gas- eous SiCl4 and SiF4, respectively (both hazardous materials). Finding BR-5. The maintenance of negative pressure within the PWC has not been demonstrated under mu- nition-processing conditions. Pressure excursions that produce positive pressure in the PWC vessel could re- lease product gas to the surrounding room. Some up- sets that could result in moderate to severe pressure excursions are listed below: · A leak in the torch-cooling system could release water into the PWC, and rapid steam formation could pressurize the vessel. Water leakage might 87 also lead to more severe pressure excursions or even explosions. · Energetic material that remained in a mortar or projectile and was introduced into a PWC could detonate upon heating, which would generate a pressure pulse. The severity of this pulse would depend on the type and quantity of explosive. · An improper cut of the rocket motor could allow a larger-than-design piece of propellant to be intro- duced into the PWC. If the gas production rate from the propellant exceeds the capacity of the downstream PAS, the vessel could overpressurize. The technology provider should investigate the likeli- hood of such events and determine their potential im- pacts on the operation of the PWCs. Finding BR-6. Combustion of plasma-converted gas in a boiler faces three major hurdles: (1) to avoid being permitted under RCRA as a boiler burning hazardous waste, the gas may have to be delisted; (2) the gas may require significant scrubbing to remove compounds that are unsuitable as boiler feedstock; and (3) the boiler will have to be configured to burn gas that has a low heating value efficiently in order to avoid generating unacceptable emissions. Finding BR-7. Although a PWC may not be consid- ered to be an incinerator by permitting authorities, the most likely permitting procedures for a PWC would be similar to those used for incinerators.
