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.
9 Teledyne-Commodore Solvated Electron Technology Package INTRODUCTION AND OVERVIEW The Teledyne-Commodore team's technology pack- age for the destruction of assembled chemical weapons (summarized in Table 9-1) involves four fundamental technologies: · The chemical agent, energetic materials, and metal parts are separated via ammonia jet cutting and wash-out. · Solvated electron technology (SET) is used to de- stroy chemical agent, deactivate energetic materi- als, and decontaminate metal parts and dunnage. SET solutions, metallic sodium in anhydrous liq- uid ammonia, are highly reducing and are charac- terized by an intense blue color from the presence of partially solvated (i.e., ammoniated) electrons. The blue color provides a visual indicator of the reactivity of the solution for destroying agents and energetics. The solid and liquid residuals of the SET process are treated by water hydrolysis to destroy the excess sodium. · The hydrolysate from agent and energetics de- struction is further treated by oxidation with so- dium persulfate or hydrogen peroxide. Figure 9-1 shows how the four fundamental tech- nologies are linked and identifies the basic process flow. Teledyne-Commodore assigns a separate area for application of each technology to the materials from the munitions. The following section describes each area in detail. 133 DESCRIPTION OF THE TECHNOLOGY PACKAGE Munitions Access and Energetics Deactivation (Area 100) Teledyne-Commodore designed Area 100 to handle three types of munitions: M55 rockets containing ex- plosives, propellant, igniters, and agent; projectiles and mortars containing explosives and agent; and land mines containing explosives and agent. Figures 9-2 through 9-4 outline the approach for each. The pro- posed disassembly processes differ significantly from the baseline processes developed by the Army, princi- pally in the use of ammoniajet cutting and ammonia wash-out. Fluid jets are used routinely in industry for cutting metal and have been used in the demilitarization of conventional munitions. For background information on jet cutting, see Appendix G. Only the particular ap- plication proposed by Teledyne-Commodore is de- scribed below. In both the cutting and wash-out operations, ammo- nia is pressurized to 2,720 aim (40,000 psi) by an in- tensifier pump and delivered to the work area through a 0.01-inch diameter orifice at velocities of about 1,000 m/s. Pre-intensifier booster pumps are used to ensure adequate pressure and volume of ammonia and to prevent flashing of the liquid into gas during the suction stroke of the intensifier. For the cutting opera- tion, 180-micron abrasive particles, normally garnet, are added to the pressurized ammonia stream through a stainless steel venturi mixing section. A pressure
134 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS TABLE 9-1 Summary of the Teledyne-Commodore SET Technology Package Major Demilitarization Operation Approach(es) Disassembly of munitions High-pressure ammonia fluidized jet cutting and ammoniajet wash-out. Treatment of chemical agent SET reduction; hydrolysis of the condensed-phase products; oxidation of the hydrolysate with sodium persulfate or hydrogen peroxide. Treatment of energetics SET reduction; hydrolysis of the condensed-phase products; oxidation of the hydrolysate with sodium persulfate or hydrogen peroxide. Treatment of metal parts Agitation of shredded parts in SET solution to a 3X condition. Treatment of dunnage Grinding or shredding; mixing with SET solution to destroy agent. Disposal of waste Solids. product from process and decontaminated dunnage sent to a suitable permitted landfill; metal parts shipped to Rock Island Arsenal for SX treatment. Liquids. stabilized with cement and shipped to appropriately permitted landfill; oil and hydraulic fluids sent to TSDF. Gases. off-gases from process and vaporized hydrocarbon residuals burned in boiler. vessel surrounds the cutting and wash-out processing equipment, providing an intermediate chamber that is maintained at 10.5 aim (140 psi") inside an Army baseline explosion-containment room. The pressurized fluid from the intensifier pump Is passed through a chiller to reduce the temperature of the liquid "significantly" below room temperature be- cause prechilling is reported by the technology provider to enhance the jet cutting properties of the liquid am- monia. The proposal does not include details on the chiller or the optimum temperature of the fluid jet stream. The following sections summarize Teledyne- Commodore's approach to the disassembly of each munition type. . Rocket Disassemb/y M55 rockets are loaded from the receiving area through an airlock onto a rotary index table located in a pressurized chamber. The index table automatically rotates the rockets to six successive cutting and wash- out stations for separation of agent, energetics, and metal parts as outlined in Figure 9-2. Projecti/e/Mortar Disassemb/y Land Mine Disassemb/y Figure 9-4 is a diagram of the process proposed for disassembling M23 land mines by ammoniajet cutting and wash-out. Fuzes are separated from the muni tions via jet cutting and are initiated in a detonation chamber using a high-voltage electrical charge. The fuze parts are then washed with SET solution in a dedicated reactor. Treatment of Chemical Agent (Areas 200, 400, and 500) SET Reduction fo//owed by Hydro/ysis (Area 200) In Area 200, a mixture of agent and anhydrous am monia, transferred from Area 100, is collected in a car bon steel vessel, 3.5 feet in internal diameter and 10.5 feet in height. A SET solution is generated by mixing liquid sodium and ammonia to form a 4-percent solu tion of sodium in liquid ammonia. The sodium is trans ferred at its melting point of 97.5°C (207.5°F); the liq uid ammonia is transferred at room temperature and 10.5 aim (140 psi") (saturated conditions). The re agents are combined in an in-line static mixer that Teledyne-Commodore describes as "flow splitting de vices, which provide uniform droplet sizes of both streams" (Teledyne-Commodore, 1997~. When the so Projectiles and mortars are processed similarly to dium comes in contact with the ammonia, it dissolves rockets by ammoniajet cutting and wash-out. Figure rapidly, releasing 1,400 calories per mole of sodium 9-3 outlines the sequence of cuts. introduced. Since a 4 wt. percent solution of sodium in ~, ~, .. . . . . .
135 : : ,..~: j ~ ' !a~ | II ·i~ ~* . ----1~---------------------------~-------1--------l---------------------- ~ -------l-- ·~ 4! ~ -~ 1 L _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
136 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS Cut no. ~ Remove fuze; - wash out booster ~ Cut no. ~ Fluid-jet puncture; - remove agent Cut no. ~ Remove aft bulkhead; _ wash out propellant Cut no. A) Cut motor; _ wash out propellant \! 1 ~6 ~ ~ Remove booster; Hi\ wash out burster -A) Remove burster/forward bulkhead; wash out warhead Looser\ {1 11 ~\ 11 11 ~~ ~ \ M417 fuze \ (fluid-jet only cuts \ through safe areas \ containing secondary \ explosives) / \ Burster / / FIGURE 9-2 The sequence of cuts for ammoniajet cutting and wash-out of MSS rockets. Source: Teledyne-Commodore, 1997. liquid ammonia contains 0.17 moles of sodium per hun- dred grams, the heat release on mixing would be 238 calories per hundred grams of solution. The heat of vaporization of ammonia at room temperature and 10.5 atm (140 psi") is 279 cal/g. Therefore, approxi- mately 1 percent of the ammonia would be expected to vaporize in the process of forming the SET solution. The prepared SET solution is fed to a static reactor mixer, and agent from the feed vessel is introduced below the liquid surface. The reaction is carried out at temperatures of 19 to 23°C (66 to 73°F) and pressures of 8.5 to 12.4 aim (110 to 167 psi"). The agent destruc- tion reactions are exothermic, and temperature is con- trolled by the evaporation of ammonia. The evaporated ammonia and off-gases from the SET reaction are col- lected in holding tanks for testing prior to venting to the gas-treatment train (Area 800~. The slurry from the SET reaction, still at elevated pressure, is treated with water to destroy the excess sodium. The resultant sus- pension is transferred at elevated pressure to a centri- fuge where solids and liquids are separated. The solids are transferred to a carbon steel vessel, 3 feet in inter- nal diameter and 8 feet in height, and mixed with wa- ter. The solution is then pumped to Area 500 for oxida- tion. The liquids remaining in the centrifuge are mixed with water and fed to the bottom of the ammonia-re- covery tower in Area 400. Although reactions of organics with solutions of
TELEDYNE-COMMODORE SOLVATED ELECTRON TECHNOLOGY PACKAGE metallic sodium in liquid ammonia have been studied since 1865, the reaction products cannot be predicted. In general, the solvated electrons are attracted to the polar bond between carbon and a more electronegative species, such as chlorine, fluorine, phosphorus, sulfur, or oxygen. The result is cleavage of the covalent bond. The strongly electronegative species leaves as an an- ion. The complementary site on the less electronega- tive atom (usually carbon) may capture a second elec- tron, thereby becoming negative; or the bonds in the remaining carbon skeleton may rearrange themselves and release a gaseous alkane or alkene. Further reac- tion occurs when the condensed-phase products of the SET reaction are hydrolyzed. Teledyne-Commodore's experimental results from treating specific agents with SET in the laboratory are described below. Agent HD (Mustard). The expected initial reaction of mustard in the SET process is cleavage of the car- bon-chlorine bond by a solvated electron to form so ,3 Burster well 2) 137 dium chloride (NaCl). Measured results, scaled up to 100 pounds of HD, are shown in Table 9-2. The gas- eous ammonia released in the SET reaction results from evaporation, which is used to control the temperature of the liquid mixture. In the proposed full-scale sys- tem, the gaseous ammonia from the hydrolysis reac- tion is produced in the agent-ammonia recovery tower (see Area 400~. Teledyne-Commodore was unable to determine the molecular composition of the slurry from the SET re- action prior to hydrolysis. The analytical problem, as described to the committee, was two-fold. One was the tendency for foaming and sudden gas evolution when the pressure was dropped on the SET product slurry. The other was the lack of standard analytical protocols for ammonia systems. The solid and aqueous solution from the hydrolysis step was combined for analysis and was found to contain NaCl, sodium sulfide (Na2S), and a variety of polysulfides. Agent concentrations were below detectable limits in the solid and liquid products, 1 1 1 '- '- 1~- 1 ~ ~ 1 , Burster cup ~ /, 1 ~ / Agent cavity ~an\\ ~1 ~71 ~ Rotating fluid-jet wash-out head Mustard heel (removes heels or scale) or crystallized GB (I Fuze 1 ID Sequence of cuts , ~ Provides access hole 4) into cavity for agent identification (as) Removes fuze ' - Removes burster cup and ( 3 ) accesses burster well for ~' explosive wash out ,~` Accesses agent cavity allowing W fluidjet to wash out all agent and remove heels or scale FIGURE 9-3 The sequence of cuts for ammoniajet cutting and wash-out of projectiles and mortars. Source: Teledyne-Commodore, 1997.
138 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS \\ Mine unpack/ / , ~ 3` Pressure x, ( area '< / ('at ~: Burster cuttin Cuttitnagn era ~ and wash Ou9 Explosion-containment station Booster cutting and sa ion wash-out station / Unpack area FIGURE 9-4 Schematic diagram of the process proposed for disassembly of M23 land mines. Source: Teledyne-Commodore, 1997. and no agent was detected in the ethylene off-gas, pro- vided that agent had been introduced below the surface of the SET solution. Agent GB. The expected initial reaction of GB in the SET process is cleavage of the phosphorus-fluorine bond by a solvated electron to form sodium fluoride. Measured results, scaled up to 100 pounds of GB, are TABLE 9-2 Measured Results for the SET/Hydrolysis Reaction of HD based on Laboratory Data and Scaled Up to 100 lb of Agent Material NH3 (liquid) Na (liquid) HD C2H4 (gas) H2 (gas) NH3 (gas) Slurry Water Solid Aqueous solution Loss SET Reaction (lb) Hydrolysis (lb) Feed Product FeedProduct 1,656 76 100 22 0.05 0.4 0.04 84 1,495 1,676 1,676 282 50 67 397 TOTAL 1,832 1,832 1,958 1,959 Heat Release 352,800 BTU 119,363 BTU Source: Adapted from Teledyne-Commodore, 1998a. shown in Table 9-3. Release of ammonia gas in the SET reaction again results from evaporation used for temperature control. Release of both ammonia and iso- propyl alcohol to the gas phase in the hydrolysis reac- tion occurs in the agent-ammonia recovery tower in the proposed full-scale system (see Area 400) (Figure 9-5~. TABLE 9-3 Measured Results for the SET/Hydrolysis Reaction of GB based on Laboratory Data and Scaled Up to 100 lb of Agent Material NH3 (liquid) Na (liquid) GB C3H8 (gas) H2 (gas) CH4 (gas) C2H4 (gas) NH3 (gas) Slurry Water Solid Aqueous solution Isopropanol (aqueous) SET Reaction (lb) Hydrolysis (lb) Feed ProductFeed Product 958 41 100 8 0.1 0.1 0.001 1,0471,047 112 TOTAL 1,099 1,099 1,159 8 0.03 897 122 122 9 1,158 Heat Release 29,730 BTU 47,207 BTU Source: Adapted from Teledyne-Commodore, 1998a.
TELEDYNE-COMMODORE SOLVATED ELECTRON TECHNOLOGY PACKAGE The composition of the slurry from the SET reaction was not analyzed prior to hydrolysis. Table 9-4 shows the solid and aqueous products predicted based on nuclear magnetic resonance (NMR) analysis using 3lP spectra. In small-scale tests, GB was not detectable in any of the reaction products. Agent VX. Unlike mustard and GB, the VX molecule does not have a strongly polar bond for the solvated electrons to attack. Teledyne-Commodore postulates that cleavage occurs initially at the phosphorus-sulfur bond, the sulfur-carbon bond, or at both simulta- neously. Measured results, scaled up to 100 pounds of VX, are shown in Table 9-5. Ammonia is vaporized during the SET reaction, and ammonia and a mixture of alcohols and amines are generated as part of the hy- drolysis process in the agent ammonia-recovery tower in the proposed full-scale system (See Area 400) (Fig- ure 9-6~. The composition of the slurry from the SET reaction was not analyzed prior to hydrolysis. Table 9-6 shows the solid and aqueous products predicted based on NMR analysis using 3lP spectra. o 139 Concentration of Hydro/ysate and Recovery of Ammonia (Area 400) In Area 400, the hydrolyzed liquid from the SET process is pumped to the bottom of the agent ammonia- recovery tower. The recovery tower is a packed carbon steel column with a hot oil reboiler and a chilled over- head condenser. The gases that rise to the top of the water are condensed with chilled propylene glycol and transferred to an ammonia-recycle drum. Noncon- densable gases are sent to Area 800 for treatment. The material that collects at the bottom of the recovery tower, which contain the residues from the hydrolysis of the liquid products of SET, is sent to Area 500 for oxidation. Oxidation (Area 500) In Area 500, the hydrolysates from Areas 200 and 400 are oxidized with sodium persulfate. The purpose of this oxidation step is to eliminate CWC Schedule 2 compounds by converting all organic phosphorus to H3C\ 1 1 + ° CH O P O NH4 H3C P O Na H3C I I CH3 0 Na Ammonium isopropyl methyl phosphonate H3C P +Na 1 Disodium methyl phosphonate H3C P-NH O- +NH 0- +NH Sodium ammonium methyl phosphinate o 11 HC P O-+NH 1 O- +NH Ammonium methyl phosphonamide 4 Diammonium methyl phosphonate FIGURE 9-5 Formulas for the more complex reaction products from SET/hydrolysis of GB. aThese compounds are more commonly called phosphonites.
140 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS TABLE 9-4 Predicted Solid and Aqueous Reaction Products of SET/Hydrolysis of GBa Reaction Products Formula Moles/Mole of GB lb/100 lb of GB Sodium fluoride NaF 1.00 30.0 Sodium hydroxide NaOH 0.84 24.0 Isopropanol C3H7OH 0.39 16.7 Ammonium isopropyl methyl phosphonate See Figure 9-5 0.36 39.8 Propane COHN 0.25 7.85 Disodium methyl phosphonate See Figure 9-5 0.25 25.0 Sodium ammonium methyl phosphinate See Figure 9-5 0.16 13.6 Ammonium methyl phosphonamide See Figure 9-5 0.13 10.4 Diammonium methyl phosphonate See Figure 9-5 0.10 7.3 Hydrogen H2 0.84 1.2 aPrediction based partly on quantitative analysis and partly on compounds identified but not quantified by NMR analysis. Source: Adapted from Getman, 1998. inorganic orthophosphate and converting the thiol from VX to sulfonic acid. Teledyne-Commodore is still in the very early stages of investigating persulfate oxida- tion for final treatment of residuals from the SET/hy- drolysis reactions. The pH of the solution is adjusted to around 10 by addition of 10 N sodium hydroxide, and the tempera- ture of the solution is raised to 80 to 85°C (176 to 185°F). Sodium persulfate is then fed to the reactor as a 32.5-percent solution in water. Gaseous products of oxidation are sent to the gas-treatment train (Area 800~. When oxidation is completed, the contents of the reac- tor are evaporated. Water is condensed for reuse, and the solids are packaged for delivery to a landfill. Teledyne-Commodore has not yet identified opti- mum conditions for the oxidation of hydrolyzed resi- dues from SET treatment of GB and VX. The addition of sodium persulfate to the hydrolysates in small-scale tests resulted in highly exothermic reactions, more vig- orous with VX than with GB. The persulfate had to be added very slowly to keep the temperature in the range of 95 to 100°C (203 to 212°F). Experiments character- ized by Teledyne-Commodore as full-scale reactions of VX and GB residues from SET/hydrolysis with so- dium persulfate are described below. In an experiment on VX,25 grams of residuals were mixed with 367 grams (287 cc) of 10 N sodium hy- droxide and heated to 85°C (185°F) with stirring. The first increment of sodium persulfate caused a vigorous exothermic reaction and a rapid increase in tempera- ture to 100°C (212°F). The solution was allowed to cool, and 468 grams (360 cc) of 32.5-percent sodium Material Feed persulfate was added over a period of two hours. In the process, 1,020 cc of gas was evolved, but analysis of the gas is suspect because of air leakage into the sample. Based on a review of published literature pro- vided by Teledyne-Commodore, the committee expects that CO2 and oxygen would be the gaseous reaction products. However, Teledyne-Commodore did not ana- lyze for CO2, and the air leak precluded a determ~na- tion of the amount of oxygen, if any, generated in the oxidation reaction. An analysis of the condensed-phase product showed only 69 percent conversion of phos TABLE 9-5 Measured Results for the SET/Hydrolysis Reactions of VX based on Laboratory Data and Scaled Up to 100 lb of Agent SET Reaction (lb) Hydrolysis (lb) Product Feed Product 486 21 100 2 0.05 0.05 35 554 16 NH3 (liquid) Na (liquid) VX C2H6 (gas) H2 (gas) C2H4 (gas) NH3 (gas) Slurry Water Loss Solid Aqueous solution Alcohols and amines (gas) TOTAL 607 607 657 657 0.003 0.07 443 555 102 17 188 9 Heat Release 52,200 BTU 43,265 BTU Source: Adapted from Teledyne-Commodore, 1998a.
TELEDYNE-COMMODORE SOLVATED ELECTRON TECHNOLOGY PACKAGE H3C P Na O +NH4 Sodium ammonium methyl phosphinate H3C P Na 1 O Na Disodium methyl phosphinate o 11 H. CH C O P O +Na 1 CH3 Sodium ethly methyl phosphonate . . - 1 H3C P O NH4 1 O +NH 4 Diammonium methyl phosphonate S 11 + Et 0 P O Na CH 3 Sodium ethyl methyl phosphorothiolate o 11 H3C P NH2 1 O +NH Ammonium methyl phosphonamide (CH3 )2 CH \oH/ CH 2 CH 2 \oH/ CH(CH 3~2 / \ - '/ \ CH(CH 3~2 (CHAT CH CH2 CH2 / Dihydroxy N. N. N', N', - tetraisopropyl piperazine S 11 + H3C P O Na O +NH4 Sodium ammonium methyl phosphorothiolate FIGURE 9-6 Formulas for the more complex reaction products from SET/hydrolysis of VX. aThese compounds are more commonly called phosphonites. phorus to orthophosphate. Some methylphosphonic acid was found to be present in the oxidized product by NMR analysis, but the amount was not quantified. In an experiment on GB, 21.7 grams of residuals were mixed with 200 cc (250 grams) of 10 N sodium hydroxide and heated to 90°C (194°F); 375 cc of gas was released during the heating process. The Teledyne- Commodore report does not explain why the tempera 141 ture increase alone, prior to the addition of persulfate, resulted in a gaseous release. Moreover, the gas was not identified. Over the next hour, 487.5 grams (375 cc) of 32.5-percent sodium persulfate was added in drops to maintain the temperature in the range of 95 to 100°C (203 to 212°F). An additional 925 cc of gas was evolved during the oxidation process. This analysis of the gas is also suspect because of apparent air leakage
142 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS TABLE 9-6 Predicted Solid and Aqueous Reaction Products of SET/Hydrolysis of VXa Reaction Products Formula Moles/Mole of VX lb/100 lb of VX Sodium bis-diisopropylamino ethyl mercaptide Sodium hydroxide Ethyl alcohol Hydrogen Sodium ammonium methyl phosphinate Sodium ethyl methyl phosphorothiolate Disodium methyl phosphinate Ethane Dihydroxy N,N,N',N'- tetraisopropyl piperazine Ammonium methyl phosphonamide Sodium ethyl methyl phosphonate Diammonium methyl phosphonate Sodium ammonium methyl phosphorothiolate See Figure 9-6 NaOH C2HsOH H2 (g) See Figure 9-6 See Figure 9-6 See Figure 9-6 C2H6 (g) See Figure 9-6 See Figure 9-6 See Figure 9-6 See Figure 9-6 See Figure 9-6 0.73 0.59 0.55 0.48 0.36 0.24 0.16 0.16 0.135 0.12 0.05 0.04 0.03 50.0 8.8 9.5 0.4 16.0 14.5 7.4 1.8 14.7 5.0 2.7 1.5 1.3 aPrediction based partly on quantitative analysis and partly on compounds identified but not quantified by NMR analysis. Source: Adapted from Getman, 1998. into the sample. For the condensed phase, Teledyne- Commodore reports, "Within experimental error, all phosphorus in the product was present as inorganic phosphate and NMR showed no C-P bonds." Treatment of Energetics (Areas 300, 600, and 700) SET Reduction fo//owed by Hydro/ysis (Area 300) The slurry produced from the wash-out of energetic material in Area 100 is collected in a carbon steel ves- sel, approximately 2.5 feet in internal diameter and 7 feet in height, and diluted with ammonia to a "stan- dard composition." A solution of sodium in liquid am- monia is formed, pumped to a reactor vessel, and mixed with the diluted energetic slurry. The reaction time for destruction of energetics is reported by Teledyne-Com- modore to be as long as 30 minutes. In contrast, decon- tamination of agent by the SET process is reported to be almost instantaneous. For M28 propellant and RDX, the longer reaction time may be attributable to their relatively low solubility in SET solutions. Also, when any energetic material is reacted, side reactions may consume solvated electrons to form polymers. The products formed in the SET reaction are trans- ferred to a holding vessel, approximately 3 feet in in- ternal diameter and 8 feet in height. Water is added, and the hydrolysate is pumped to the bottom of the energetics-ammonia recovery tower in Area 600. The products of SET/hydrolysis are generally complex polyaromatic hydrocarbons that Teledyne-Commodore believes are formed from the polymerization of the aro- matic or heterocyclic ring structures common to most energetics and from small quantities of nitrate, nitrite, and cyanide salts. Teledyne-Commodore conducted research on SET destruction of explosives, fuzes, and propellants over a nine-month period in the laboratories of the Southwest Research Institute. Most were bench-scale tests con- ducted at-33.4°C (-28°F), the boiling point of ammo- nia at atmospheric pressure. One series of tests was conducted at ambient temperature and at a pressure of about 10.5 atm (140 psi"), the conditions proposed for a full-scale plant. No major differences were observed in the reactions carried out under these two very differ- ent sets of conditions. Energetics tested were TNT, RDX, tetryl, PETN (pentaeryhritol tetranitra), Comp B, picric acid, nitrocellulose, and M28 propellant. In the laboratory tests, a weighed sample of the en- ergetic was dissolved in liquid ammonia, and sodium was added incrementally until the characteristic blue color of the SET solution was observed. At the end of the reaction, isopropyl alcohol or water was added to destroy excess sodium, and the ammonia was evapo- rated prior to collecting a sample for analysis. Details of tests on different types of energetics are provided below.
TELEDYNE-COMMODORE SOLVATED ELECTRON TECHNOLOGY PACKAGE TNT. Sodium was added incrementally to a solution of one gram of TNT in liquid ammonia. After 1.957 grams of sodium had been added, the blue color char- acteristic of SET remained, and the test was terminated. The residue remaining after the addition of water and evaporation of the ammonia was a paste, consisting of "nondescript polymeric organic materials." The only reaction products identified were those listed in Table 9-7, which total less than 1 wt. percent of the TNT treated. No products of SET/hydrolysis could be iden- tified by either NMR or liquid chromatography/mass spectrometry (LC/MS)analysis. No TNT or other ni- trated toluenes were found in the residues at Na/TNT ratios greater than 0.5. The residue from SET reactions with TNT were sen sitive to electrostatic energy as evidenced by the pres- ence of smoke and the presence of combustion prod- ucts during testing. The residue can be considered electrostatically sensitive because the minimum igni- tion energy was measured to be as low as 90 my in at least one of five trials. The tests were conducted at con- stant arc gap of 0.25 inches (the distance from the end of the electrode to the inside bottom surface of the an- ode sample-holder cup). Teledyne-Commodore cites Batz et al. (1997) for a discussion of the reaction of TNT with SET solutions. Batz observed a white precipitate upon reaction of TNT with a SET solution and noted that, "Evaporation of the ammonia leaves an off-white solid which explodes upon agitation." In the Batz experiments, a polymeric fraction was also formed that was insoluble in water and very sensitive to explosion when agitated. RDX. One gram of RDX added to 100 ml of liquid ammonia did not dissolve completely after stirring for 1 hour and 16 minutes prior to the addition of sodium. TABLE 9-7 Identified SET Reaction Products of Treatment of TNT Products Quantity Nitrite Nitrate Ionic cyanide Gases releaseda 1 8 mg/g 98 mglg 0.595-2.8 mg/g 13 mug aMainly hydrogen, with ppm quantities of C~-C3 hydrocarbons. Source: Adapted from Teledyne-Commodore, 1998a. 143 The reaction was judged to be complete after sodium additions totaled 1.3 grams, when the blue color of the SET solution persisted for more than six minutes. The residue after hydrolysis and ammonia evaporation was an off-white dry flaky material. The only reaction prod- ucts identified after the addition of sodium and subse- quent hydrolysis are shown in Table 9-8. The products identified account for less than 1 percent of the RDX treated. No reaction products could be identified by ei- ther NMR or LCIMS. Semivolatile analysis identified small quantities of hexamethylenetetramine, a possible impurity in the RDX that survived SET treatment. NMR analysis detected no RDX in the residue. How- ever, the residue was sensitive to electrostatic energy, as indicated by the presence of smoke and combustion products under test conditions. The minimum measured ignition energy was 50 my. Tetryl. Tetryl reacted similarly to TNT. Both ener- getics are very soluble in liquid ammonia, and the resi- due in both cases was a pasty, nondescript polymeric species. Total gases released from the SET reaction with tetryl ranged from 32 to 214 ml/g and averaged 3 percent hydrogen, 9 percent oxygen, 10 percent nitro- gen, 901 ppm methane, and 117 ppm C2 hydrocarbons. No tetryl was identified in the residue by high perfor- mance liquid chromatography analysis. In one test, after the water quench to destroy excess sodium, a strong exothermic reaction occurred upon drying of the residue from SET treatment. Teledyne- Commodore notes, however, that if the residue is mixed during the drying process, the temperature does not in- crease sufficiently to cause carbonization. Comp B. The addition of Comp B to liquid ammonia turned the solution a deep cranberry-red color. Sodium was added incrementally, and at a Na/Comp B ratio of TABLE 9-8 Identified SET Reaction Products of Treatment of RDX Products Quantity Nitrite Tonic cyanide Gases releaseda 3. 1-26 mg/g 0.016-15.8 mg/g 55-140 mug aO.21 percent hydrogen; 11-455 ppm methane; and 2-115 ppm ethane. Source: Adapted from Teledyne-Commodore, 1998a.
144 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS TABLE 9-9 Identified SET Reaction Products of Treatment of Comp B Products Quantity Nitrite Nitrate Inorganic cyanide 33 mglg 2 mglg 3.4 mglg Source: Adapted from Teledyne-Commodore, 1998a. 1.15:1.0, the SET blue color persisted. The residue af- ter hydrolysis and evaporation of ammonia was a paste that was soluble in both water and methanol. The only reaction products identified (summarized in Table 9-9) account for less than 1 percent of the weight of the Comp B. No reaction products could be identified by either NMR or LC/MS. Neither RDX nor TNT was identified in the residue. However, the residues were sensitive to electrostatic energy, as evidenced by the presence of smoke and combustion products during testing. The minimum measured ignition energy was 100 mJ. M28 Propellant. M28 propellant dissolves very slowly in liquid ammonia. To accelerate dissolution, the propellant was ground to a sawbust-like consistency with a wood file. The residue after SET treatment, hy- drolysis, and evaporation of ammonia was a soft sticky paste. The only identified reaction products (summa- rized in Table 9-10) account for less than 1 percent of the weight of the M28. No reaction products could be identified by either NMR or LC/MS analysis. None of the target explosive compounds was identi- fied in the residue, as determined by high pressure liq- uid chromatography analysis. A TCLP analysis of the residue showed lead at a concentration of 75.6 mg/L. The nonhydrolyzed, dried, ammoniated residue was sensitive to electrostatic energy, as indicated by flaming. The minimum measured ignition energy was 175 mJ. In a larger-scale test conducted at ambient tempera- ture and elevated pressure, 97.9 grams of ground M28 and 97.9 grams of sodium were added in increments to 6 liters of liquid ammonia. The test took close to seven hours to complete. The reaction mixture was allowed to evaporate in the reactor, covered with a sheet of plastic, and a slow flow of helium was introduced to exclude ambient air. After 17 hours, the residue had the appearance of drying mud. The next morning, approximately 15 hours later, the plastic had been destroyed and part of it had hardened and turned dark. The residue was a dry, black, carbonaceous material with a charcoal-like odor. In a second simi- lar test, 20 ml of isopropyl alcohol were added to destroy the unreacted sodium prior to evaporating the ammonia in an argon stream. This test was com pleted without incident. In a test conducted to determine the effect of operat- ing at 2.5 wt. percent ground M28 at 9.2 aim (120 psi"), Teledyne-Commodore observed a spontaneous reac- tion in the reactor while pressure checking the system at 6.8 aim (85 psi"). The undissolved M28 Propellant appeared to have melted. 1 1 Hydrolysate Concentration and the Recovery of Ammonia (Area 600) The energetics-ammonia recovery tower operates identically to the agent-ammonia recovery tower. The aqueous bottoms are collected in a carbon steel vessel, 4.5 feet in internal diameter and 12.6 feet in height, and transferred in batches to Area 700 for oxidation. Oxidation (Area 700) According to the technology provider, the hydro- lyzed residues of SET treatment of energetics contain cyanides at levels that exceed regulatory limits. Alka- line chlorination, the cyanide destruction method in most common use, is not applicable to the hydrolysates because hypochlorite reacts with ammonia to form ni- trogen bichloride, a strong irritant. The problem arises from incomplete removal of ammonia from the SET/ hydrolysis residual. After conducting several screen- ing tests to find an oxidizer that would destroy cyanide TABLE 9-10 Identified SET Reaction Products of Treatment of M28 Propellant Products Quantity Nitrite Nitrate Tonic cyanide 9.1 mg/g 5.9 mg/g 0.6 mg/g Source: Adapted from Teledyne-Commodore, 1998a.
TELEDYNE-COMMODORE SOLVATED ELECTRON TECHNOLOGY PACKAGE in the presence of significant quantities of other re- duced species, the technology provider selected alka- line, copper-catalyzed persulfate/hydrogen peroxide oxidation. However, this system has not been tested sufficiently to identify optimum operating conditions. Gas Treatment (Area 800) Gases and ammonia vapors from areas 200 and 600 pass through reflux condensers to remove most of the ammonia. Residual ammonia vapors and noncon- densable gases are collected in parallel holding tanks for testing. After verifying that no agent is present, the gases are vented through a pressure let-down valve to a water scrubber to remove ammonia as ammonium hy- droxide, which is sent to the ammonia recovery tower (Area 400~. Noncondensable gases are chilled, reheated to a relative humidity of about 50 percent, and passed through a deep-bed activated carbon adsorber to re- move any trace agent; low-boiling hydrocarbons and hydrogen are passed through the bed and recycled as boiler fuel. SET Reduction and the Hydrolysis of Metal Parts and Dunnage (Area 900) Treatment of Meta/ Parts Munitions metals separated in Area 100 are shred- ded in Area 900 and subsequently transferred to a metal SET reactor for treatment to 3X condition. Treatment involves agitation with a SET solution in a tumbler re- actor operated in batch mode. The metal SET tumbler reactor is a double cone mixer without baffles. The shredded metal parts are introduced through the top of the tumbler, which is then sealed, vented, and filled with ammonia. Liquid sodium is injected to form a SET solution, and the sealed ves- sel is rotated slowly to ensure that all surfaces of the metal parts are wetted by the solution. Upon comple- tion of the SET reaction, water is added to the reactor to destroy the remaining sodium. The 3X metal parts are shipped to Rock Island Arsenal for 5X treatment. The proposed system has not been tested, but in tests on small metal coupons, agent was destroyed to below detectable limits, and none of the metals was found to be reactive with SET solution. 145 Treatment of Dunnage and Abrasives Many of the materials that constitute dunnage are porous to agent and, therefore, are decontaminated in SET solution. Teledyne-Commodore proposes reduc- ing the particle size of dunnage and nonmetal wastes other than activated carbon to three-eighths inch or smaller by means of a size-classifying fine shredder with internal recycling. The shredded particles are transferred to a rotary-plow mixer for treatment with SET solution. Carbon, which is already smaller than three-eighths of an inch, is fed directly to the mixer. The rotary-plow mixer is a horizontal cylindrical vessel with a central rotating shaft. Plow-shaned heads if, rotate through the vessel to agitate the mixture. After the finely divided solids have been introduced, the ves- sel is sealed, ammonia and liquid sodium are added, and the mixer is started. Following decontamination, water is added to destroy excess sodium. The proposed system has not been tested, but small- scale tests have shown that agent was destroyed to be- low detection limits. In addition activated carbon, DPE, waste oils, and silicone rubber were found to be reac- tive with the SET solution. A slurry of ammonia and abrasives from jet cutting is also treated in SET solution. This occurs in one of two abrasive SET reactors, which are tumblers de- signed to treat slurries. Following decontamination, water is added to destroy excess sodium. Process Instrumentation, Monitoring, and Control The ammoniajet cutting system incorporates sen- sors for pressure, temperature, and ammonia leakage; these sensors are integrated with control systems. When deviations from normal pressures or temperatures oc- cur, the pressurized fluid is automatically diverted to the supply tank and its recirculation system until the problem is corrected. A control-system failure triggers redundant pressure-relief valves. On-line electric sen- sors detect ammonia leakage into secondary-contain- ment piping; excessive leakage automatically activates a shutdown and maintenance request. The rotary index tables are equipped with sensors that monitor the position of munition parts and the op- eration of the fluidjet cutting components. Additional . · .
146 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS sensors monitor clearance and torque on the table to prevent jamming or overload. Controls embedded in the index tables only allow operations to proceed if control algorithm requirements are satisfied. A mis- handled munition causes the index table to stop until the workstation is cleared. If a workstation is empty, the control system prevents operation of the fluidjet components until munitions are properly loaded into the work nests. SET solutions are prepared by mixing predetermined quantities of liquid sodium and ammonia in an in-line static mixer. Mass-flow controls and interlocks control the process precisely. Two independent mass-flow meters control the flow of agent and energetics into a static-mixer reactor containing the prepared SET solu- tion. Interlocks stop agent flow if SET solution is un- available. Measurements of conductivity are used to indicate destruction of agent or energetics and the avail- ability of excess SET solution. However, the technol- ogy provider has also noted that some SET products of energetics destruction are ionic and contribute to the conductivity of the reacting SET solution. The headspace in the reactor-product vessel is moni- tored continuously for traces of agent in the gas stream. If agent is present in the off-gas, the gas is held and reworked to ensure complete destruction. TABLE 9-11 Process Inputs for the Teledyne Commodore Technology Package for VX-filled M55 Rockets Processed at a rate of 20/hr Process Inputs Mass Flow (lb/hr) Munitions Agent VX Propellant Other energetics Metal parts Abrasive Sodium (for agent) Sodium (for energetic) Sodium (for decontamination) Water (make-up) Oxidants (for agent) Oxidants (for energetic) Cement Dunnage Total 200.0 384.0 64.0 896.0 132.0 39.6 277.7 80.0 230.2 3,540.5 3,139.2 3,279.7 440.0 12,700.9 Source: Adapted from Teledyne-Commodore, 1998b. Hold-test-release provisions for the abrasive reac- tor, the munition tumbler, and the rotary-plow mixer ensure that all metal parts and dunnage are tested for decontamination before release for packaging and disposal. Feed Streams Tables 9-1 1 and 9-12 list the feed streams entering the Teledyne-Commodore system when processing VX-filled M55 rockets and HD-filled 155- mm oroiec- tiles, respectively. Waste Streams The Teledyne-Commodore system generates at least 17 process-waste streams for disposal. Those identi- fied by Teledyne-Commodore are listed in Table 9-13. Tables 9-14 and 9-15 provide mass flows for the consolidated waste streams leaving the system when processing VX-filled M55 rockets and HD-filled 155- mm projectiles, respectively. Start-up and Shutdown Start-up and shutdown procedures for the full-scale system are still being developed. TABLE 9-12 Process Inputs for the Teledyne- Commodore Technology Package for HD-filled 155 mm Projectiles Processed at a rate of 100/hr Process Inputs Mass Flow (lb/hr) Munitions HD ~. 1 170.0 ~nerget~cs4 1.0 Metal parts8,600.0 Abrasive154.0 Sodium (for agent)778.1 Sodium (for energetic)22.6 Sodium (for decontamination fluid)132.6 Oxidants (for agent)7,575.8 Oxidants (for energetic)398.3 Cement150.0 Dunnage525.0 Total 19,547.4 Source: Adapted from Teledyne-Commodore, 1998b.
TELEDYNE-COMMODORE SOLVATED ELECTRON TECHNOLOGY PACKAGE TABLE 9-13 Process Waste Streams Released to the Environment 147 Waste Source Waste Treatment FinalDisposalMethod Ferrous metal parts from projectiles SET treatment to 3X Shipment to Rock Island for SX processing Land-mine and non-PCB MSS None Landfill shredded metal parts Residual sodium salts from oxidation Evaporation, filtration, dewater~ng, RCRA Subtitle C landfill of SET-treated agent and energetics and packaging Ethanol liquid from oxidation of Integrated with off-gas Burned for energy recovery SET-treated VX N,N=-tetrisopropylpiperazine Unknown Unknown dihydroxide from oxidation of SET-treated VX Sodium nitrate and polymeric organics (no lead) from oxidation of SET-treated energetics Packaged for shipment RCRA Subtitle C landfill Noncondesable off-gases Reheated to 50% relative humidity and passed through a deep-bed carbon ads orber Released to air Source: Adapted from Teledyne-Commodore 1997, 1998a. EVALUATION OF TH E TECH NOLOGY PACKAG E Process Efficacy Effectiveness of Disassembly of Munitions Teledyne-Commodore proposes to use a completely new disassembly process based on ammoniajet cut- ting. Although waterjet cutting offers many potential benefits for the demilitarization of ordnance (see Ap- pendix G), Teledyne-Commodore proposes substitut- ing ammonia for water in the fluidjet cutting operation mainly because the primary treatment process is car- ried out in anhydrous liquid ammonia. Teledyne-Commodore also mentions several advan- tages of ammonia over water. They point out that many of the hazardous materials that are washed out of
148 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS TABLE 9-14 Process Outputs for the Teledyne Commodore Technology Package for VX-filled M55 Rockets Processed at a rate of 20~r Process Outputs Mass Flow (lb/hr) Metal parts Stabilized abrasive Stabilized propellant Product slurry (from agent) Product slurry (from energetic) Gases (from agent) Gases (from energetic) Dunnage Total 896.0 396.0 8,593. 1 1,861.0 399.2 32.2 3.4 520.0 12,700.9 Source: Adapted from Teledyne-Commodore, 1998b. munitions have very low solubility in water and con- tain surfactants that aid in the formation of stable emul- sions. These emulsions have presented serious safety and maintenance problems in full-scale conventional- munition disassembly facilities. Anhydrous ammonia, in contrast, dissolves most of the explosive materials and forms nondetonable solutions. Teledyne-Commodore successfully demonstrated ammoniajet cutting at Redstone Arsenal on M60 (in- ert) and M61 (live) rockets, a 4.2-inch M2 inert mortar, and a 105-mm M60 inert projectile (Teledyne-Com- modore, 1998c). Successful wash-out of Comp B and M28 propellant was demonstrated during the M61 test series. Teledyne-Commodore claims that ammoniajet cutting was 25 percent faster than waterjet cutting, but no details were provided on how the comparison was made. TABLE 9-15 Process Outputs for the Teledyne Commodore Package for HD-Filled 155 mm Projectiles Processed at a rate of looter Process Outputs Mass Flow (lb/hr) Metal parts Stabilized abrasive Salts (from agent) Salts (from energetic) Water (recycled) Gases (from agent) Gases (from energetic) Dunnage Total 8,600.0 454.0 5,218.8 227.7 3,980. 1 436.5 0.3 630.0 19,547.4 Source: Adapted from Teledyne-Commodore, 1998b. Ammoniajet cutting would require a major change in the baseline munition-disassembly areas. Specifi- cally, Teledyne-Commodore plans to conduct these op- erations at room temperature and at a pressure of around 10.5 aim (140 psi"), which would require pres- surized enclosures for the cutting workstations. Effectiveness of Detoxification of Chemica/ Agents the technology provider has conducted more than 250 tests of portions of the proposed system on 15 dif- ferent chemical agents, 9 energetic materials or com- positions, and 21 different metal or process-waste (i.e., dunnage) combinations. The largest quantities of agent tested in a single batch were 1.4 pounds of HD, 1.1 lb. Of HT, 1.1 lb. of VX, and 1.3 lb. of GB. The concentra- tion of agent in the SET solution after completion of the laboratory tests was below detection limits in all cases (less than 200 ppb for HD and HI; and less than 20 ppb for VX and GB. ) The technology provider con- ducted four tests of post-treatment with sodium persulfate: one with VX, one with TNT/HD/Lewisite, one with TNT/ HD, and one with TNT/Lewisite. (Lewisite is a chemi- cal agent but is not included in the ACWA program. ~ Teledyne-Commodore has demonstrated that the SET process, followed by hydrolysis, can destroy chemical agents to a destruction efficiency of at least 99.9999 percent. However, considerably more testing and analysis will be required to determine the exact molecular composition, phase distribution, and quan- tity of reaction products. This lack of information com- pounds the difficulty of developing optimum condi- tions for the final oxidation step. The data in Tables 9-3 through 9-6 illustrate this problem. The data in Table 9-3 were measured experimentally. The data in Table 9-4 were predicted by the technol- ogy provider, based partly on quantitative analysis and partly on compounds identified but not quantified by NMR analysis. Teledyne-Commodore adjusted the molecular compositions shown until they were able to get a mass balance. The committee noted several in- consistencies in the two tables. The total measured hy- drogen release in Table 9-3 is 0.13 lb/100 lb GB. The technology provider predicted hydrogen release in Table 9-4 is 1.2 lb/100 lb GB, almost 10 times larger. The measured isopropyl alcohol output in Table 9-3 is 9 lb/100 lb GB; the predicted output in Table 9-4 is
TELEDYNE-COMMODORE SOLVATED ELECTRON TECHNOLOGY PACKAGE 16.7 lb/100 lb GB. The amounts of propane released are the same in both tables. In Table 9-3, the amount of ammonia vaporized during the SET reaction is much lower than expected, based on the heat release shown. The technology provider attributes the difference to heat losses to the vessel and surroundings. The committee is not convinced that this explains the discrepancy. The committee expected stoichiometric agreement between NaOH and H2. Presumably, NaOH is formed by the reaction of water with excess sodium in the SET product solution, which would yield 0.5 moles of H2 per mole of NaOH. The technology provider hypoth- esizes a molar ratio of 1:1 (see Table 9-4~. This higher ratio is possible because some hydrogen is generated in the SET reaction prior to hydrolysis. However, the H2 to NaOH ratio actually measured was less than 0.5. According to the data in Table 9-3, 112 lb water/100 lb GB was added to the SET reaction products in the hydrolysis step; and 122 lb aqueous solution/100 lb GB was generated. The committee calculated that dissolved solids in the aqueous solution would be 10 lb/100 lb GB, far less than the quantities hypothesized by the Technology Provider in Table 9-4 for NaOH and NaF, both of which the committee expected to partition largely to the aqueous phase. The committee identified similar inconsistencies be- tween the experimental data reported in Table 9-5 and the predicted product mix in Table 9-6 for VX. In Table 9-5, the measured quantity of hydrogen released was 0.12 lb/100 lb VX; the hypothetical quantity in Table 9-6 is 0.4 lb/100 lb VX. Moreover, the quantity of hy- drogen hypothesized in Table 9-6 does not bear a rea- sonable relationship to the quantity of NaOH hypoth- esized. The quantity of ethane reported is the same in both tables. As with GB, the amount of ammonia va- porized was less than expected, based on the heat re- lease shown in Table 9-5. Considering the quantity of water added in the hydrolysis step and the quantity of aqueous solution produced, the committee calculated a dissolved solids content of 86 lb/100 lb VX. Therefore, some of the organic compounds listed in Table 9-6 would have to partition to the aqueous phase. Teledyne- Commodore did not analyze the aqueous liquid and the solids separately. One additional point is worthy of note. Mustard con- tains volatile low molecular weight chlorinated hydro- carbons that are difficult to hydrolyze or to remove with 149 caustic or acid scrubbers or with activated carbon absorbers. The SET/oxidation process does not specifi- cally address the management of these compounds, and their treatment remains to be demonstrated. Effectiveness of Decomposition of Energetic Materials Teledyne-Commodore has not demonstrated that the SET process, followed by hydrolysis, has the capacity to decompose the energetic materials. In laboratory experiments, the condensed-phase products were gen- erally pasty, difficult to handle, of unknown polymeric composition, and sensitive to electrostatic ignition. Moreover, the reaction products identified represent less than 1 wt. percent of the material treated. The com- mittee was troubled that in several instances, appar- ently spontaneous exothermic reactions occurred. The root causes of the exotherms have yet to be identified. These results, which were obtained relatively early in the technology provider's sequence of tests, indicate that SET reactions with energetics in excessive amounts of sodium can produce sodium salts that are sensitive energetic materials. Expected sodium salts may include sodium amide, sodium azide, sodium amido-peroxide, sodium nitramidate, and possibly even sodium salts of hydrazine. Depending on the consump- tion rate of the solvated electron, some sodium picrate or diazo or azoxy derivatives of TNT may also be present. In a more recent report, Teledyne-Commodore (1998b) indicates that some of these problems may be solved if the sodium content is reduced. In small-scale laboratory experiments, the accumulation of problem- atic precipitates was eliminated by destroying excess sodium in the SET solution with isopropanol prior to the evaporation of ammonia. However, considerable additional testing will be required to demonstrate that this procedure will be effective in a full-scale system. Effectiveness of Fina/ Treatment to Produce Wastes Suitable for Disposa/ The SET/hydrolysis products from the treatment of GB and VX include Schedule 2 compounds that must be further treated. Teledyne-Commodore has tested so- dium-persulfate oxidation as the method of secondary treatment, but only on a laboratory scale. A substantial
150 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS amount of additional testing will be needed to validate the process and determine optimum conditions for full- scale operation. In the few laboratory tests that were completed, the reactions were highly exothermic and led to bubbling and rapid rises in temperature. To main- tain the temperature in the range of 90 to 100°C (194 to 212°F), the sodium-persulfate solution had to be added in extremely small increments, which, of course, in- creased the reaction time. The technology provider has not yet developed a complete analysis of either the gaseous-phase or con- densed-phase reaction products. In the treatment of VX, the piperazine formed by SET/hydrolysis appears to undergo no change during the oxidation process. Teledyne-Commodore believes that the sodium bis- diisopropylamino ethyl mercaptide formed in the pri- mary reaction is converted to the corresponding sulfite by the sodium persulfate. Other products are inorganic phosphates and sulfates, although some methanol may also form. In the treatment of GB, Teledyne-Commo- dore believes the principal oxidation products are so- dium bisulfate, sodium sulfate, sodium phosphate, am- monium hydroxide, and acetone. The SET/hydrolysis products from the treatment of energetics include inorganic cyanide at levels that ex- ceed regulatory limits. Other products of SET/hydroly- sis have not been identified. Alkaline chlorination, the method most commonly used for the destruction of cya- nide in aqueous systems, is not applicable to the solu- tions formed in SET/hydrolysis because the hypochlo- rite reacts with ammonia to form nitrogen bichloride, a strong irritant. The alternative copper-catalyzed alka- line persulfate/hydrogen peroxide system developed by Teledyne-Commodore has not been adequately tested. In assessing their own efforts to date, Teledyne- Commodore concluded, Additional work is required to validate the oxidation pro- cess. For realistic process residues, the residues will have significantly higher concentrations, affecting the process conditions. Oxidative exotherms will also be greater in higher concentrations. The oxidation process appears ap- proximately equally effective for the nitroaromatic com- pounds, but less effective for M28 propellant (Teledyne- Commodore, 1998c). In the view of the committee, the process for the final oxidation of energetics residues is still in the early re- search stage. Tests showed that hydrogen and low molecular weight gaseous hydrocarbons form during hydrolysis. In the current conceptual design, however, no provi- sion is made for venting gases released in the hydroly- sis steps in Area 200 to the gas-treatment train. This issue must be addressed if the design is taken further. Sampling and Analysis No standard generally accepted sampling and analy- sis methods are available for systems based on liquid ammonia. The condensed-phase products of SET can only be analyzed for specific chemical components af- ter hydrolysis. The hydrolysis process changes the composition of the SET products, in addition to con- verting excess sodium to sodium hydroxide and pro- ducing hydrogen gas. It may not be necessary to analyze the products of SET and the products of hydrolysis separately because hydrolysis occurs immediately after SET treatment, and the hydrolysis products are passed on to the final oxidation step. Nevertheless, the identity of hydrolysis products is still uncertain. When chemical agents are treated, color changes provide a visual indication of the completion of the SET reduction process, and measurements of conduc- tivity provide a quantitative indication. When energet- ics are treated, the persistence of the blue color charac- teristic of solutions of sodium in liquid ammonia can be masked by the intense colors of solutions of ener- getics in liquid ammonia. Measurements of conductiv- ity may also be difficult to interpret because of the for- mation of inorganic nitrates, nitrites, and cyanides, which contribute to conductivity. Gases are released in both the SET process and the hydrolysis process. Teledyne-Commodore found it dif- ficult to collect gas samples for analysis. However, there are well developed methods for collecting gas, so this problem should be relatively easy to overcome. Maturity The committee is not aware of any full-scale appli- cations of the type proposed by Teledyne-Commodore. The system is quite complex and has never been oper- ated as a totally integrated package that includes oxida- tion. The system involves at least 16 unit operations. Five are for SET treatment of agents, energetics, shredded
TELEDYNE-COMMODORE SOLVATED ELECTRON TECHNOLOGY PACKAGE dunnage, metal parts, and fuzes; the other eleven are for hydrolysis of agents and energetics, fluidjet cut- ting, oxidation of energetic residue, oxidation of agent residue, evaporation of oxidized residues, recovery of ammonia, stabilization of oxidized residues of cement, detonation of fuzes, decontamination of abrasives, and stabilization of fuzes/abrasives. The technology provider tested SET destruction of malathion, an agent simulant, at the Teledyne-Com- modore pilot-scale plant in Marengo, Ohio, where many of the unit processes and operations proposed for the ACWA program were demonstrated. The largest quantity of malathion tested in a single run was 100 pounds. The agent simulant was introduced at a rate of 125 pounds per hour. The SET solution was made up of elemental calcium in liquid ammonia. Malathion was reduced to below detectable limits. The committee notes that the Marengo facility did not include a mod- ule for testing the oxidation of residuals with persulfate. Robustness The SET process appears to be capable of destroy- ing agents with a wide range of feedstock composi- tions, temperatures, and pressures. The reactions seem to be most sensitive to the sodium/feedstock ratio, which must be high enough to ensure complete reac- tion. However, based on the data the committee re- ceived, the capability of the SET/hydrolysis process to deactivate energetics does not appear to be satisfac- tory, and optimum operating conditions have not been established. In tests conducted at Redstone Arsenal, Teledyne- Commodore encountered and was able to resolve sev- eral operating problems, including plugging of the liq- uid-sodium feed lines, which was resolved by adding an ammonia wash to dissolve solidified sodium in the piping. The technology provider plans to modify the design to maintain a uniform sodium temperature throughout the flow system. Another problem was the formation of sodium-oxide plugs in the mass-transfer system caused by small amounts of air in the sodium line. Controls were added to prevent this. Monitoring and Contro/ Teledyne-Commodore uses conductivity as the main indicator that the reaction is complete. However, because 151 some of the products of the SET reaction contribute to conductivity, additional control algorithms are being developed that incorporate temperature, pressure, and feed composition. The technology provider also intends to improve the mass-flow monitor and control system for all process constituents. Teledyne-Commodore is not far enough along in the development of a full-scale system for the committee to assess start-up and shutdown procedures. Even for the SET component, which has been in- vestigated in greatest detail, the firm notes, "Due to the intrinsic reactivity of ammonia and to kinetic ef- fects, the mixing protocol can significantly impact re- action mechanisms and pathways." In batch experi- ments, Teledyne-Commodore has tested two mixing protocols: one called "forward addition," in which the material to be treated is added to a premixed SET solu- tion; the other called "backward addition," in which sodium is added to a premixed solution of the material to be treated in liquid ammonia. The sodium consump- tion and the products of reaction are different in the two cases. For the full-scale plant, Teledyne-Commo- dore proposes using a hybrid protocol called "con- trolled stoichiometry," in which reagents are added and mixed in "carefully controlled ratios" via a static mixer. In the few bench-scale tests on SET/hydrolysis of energetics and oxidation of the residues from the treat- ment of both agents and energetics, exotherms were observed that are not yet well understood. Therefore, the possibility of runaway reactions cannot be ruled out at this time. Applicability The SET process has been applied at a reduced scale to a wide range of agents, energetics, agent/energetic combinations, and solid coupons contaminated with agent and energetics and is conceptually applicable to all assembled chemical weapons types in the U.S. stockpile. Process Safety A number of pieces of equipment and processes are unique to the SET system. These must be taken into account in an evaluation of process safety:
152 . . ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS an extraction subsystem: - a fluidjet cutting machine using high-pressure (2,720 atm; 40,000 psi") liquid anhydrous am monia streams (1 8pm) and garnet abrasive for opening and accessing chemical weapons in side a pressurized containment vessel (the mu nitions access vessel) a high-pressure (28.2 atm; 400 psi") fluidjet wash-out using liquid anhydrous ammonia to remove remaining agent and energetics from the munition cavities inside the munitions ac cess vessel an abrasives-removal system for separating garnet or other abrasives from the ammonia cutting fluid a fuze-removal machine for extracting intact fuzes after agent and energetics have been removed an explosion-containment chamber in which fuzes, removed from the munitions in the mu nitions access vessel, are initiated to effect their destruction by detonation a system for removing munitions metals and packaging a shredder system for rocket and land-mine bodies and packaging for these items a destruction system: six SET reactors for the destruction of agent and energetics and the decontamination of dun nage, shredded rockets and land mines and their packaging, abrasives used in cutting, and fuze parts; the reactors use a solvated-electron solution consisting of sodium dissolved in liq uid anhydrous ammonia ammonia-recovery systems for the agent and energetics SET reactors separate oxidation reactors for reaction prod ucts from the agent and energetic SET reactors water wash-out chambers to remove remain ing sodium from all solids (e.g., dunnage, metal parts, fuze parts, etc.) in preparation for their disposal as waste or for shipment as 3X metal parts to a U.S. Army thermal-treatment facility The SET process operates at ambient temperatures. . . ~. . ~. using evaporation of ammonia to remove heat from the mildly exothermic reactions between the SET solution and the agent and energetics in their respective reac- tors. Pressures as high as 2,720 aim (40,000 psi") are used for the ammoniajet cutting solution during disas- sembly operations. With the exception of the small sec- tion of piping and the intensifier pump for the fluid- cutting solution, most of the systems operate at near atmospheric pressure to 10.5 aim (140 psi"). Where liquid ammonia is used, vessels will be operated at ap- proximately 10.5-atm pressure, which is the vapor pres- sure of ammonia at typical ambient temperatures. Worker Health and Safety The Teledyne-Commodore technology package is basically an adaptation of existing technologies used for the destruction of conventional munitions and other hazardous chemicals. Assuming careful design and operation, the general technology of fluid cutting and fluid wash-out can achieve acceptable worker-safety levels. The aspects of the process that require special attention from a worker-safety standpoint are listed below: · use of liquid ammonia-cutting fluid at 2,720 aim (40,000 psi") fluidjet wash-out with liquid ammonia SET reaction with agent and energetics by com- bining the ammonia slurry containing these mate- rials with liquid sodium in an in-line mixer operation of mechanical equipment in ammonia atmospheres at 10.5 aim (140 psi") · processing of agent at elevated pressure ( 10.5 atm; 140 psi") · addition of water to metallic sodium A unique aspect of this technology from the stand- point of worker safety is the hazard associated with the primary chemicals used to destroy the agent and en- ergetics. Liquid anhydrous ammonia boils at -33.4°C (-28°F) and becomes a toxic gas capable of burning in air. Liquid sodium (melting point = 98°C; 208°F) is a reactive and pyrophoric metal that burns violently upon exposure to air or other oxidizing media and requires special firefighting methods and materials. Sodium persulfate and hydrogen peroxide solutions are reac- tive chemical oxidizers as well as health hazards. (For example, because of the high reactivity and toxicity of
TELEDYNE-COMMODORE SOLVATED ELECTRON TECHNOLOGY PACKAGE hydrogen peroxide at concentrations above 52 percent, it is treated as a regulated substance by the Occupa- tional Safety and Health Administration and the State of California Office of Emergency Services.) Although all of these chemicals are widely used in industry and releases and accidents are infrequent, be- cause of the hazardous and reactive nature of these chemicals, the SET systems must be carefully designed to ensure worker safety (e.g. from toxic gas exposures and fires). This includes minimizing the occasions when workers are in contact with the areas in which these hazardous chemicals may be present. The disassembly processes differ from the baseline system in that fluidjet cutting is used to access agent and energetics in all of the munitions. Indexing tech- nology now used in the disassembly of munitions will be used to control the location of the cuts. The cutting machines in a pressurized ammonia environment may be less reliable than anticipated and may require more maintenance to achieve desired throughputs. The added maintenance would increase opportunities for worker exposure. Most of the experience with cutting and fluidjet wash-out has been with fluids other than ammonia (see Appendix G). The SET reactions have been well docu- mented, but the destruction efficiencies over a pro . . . . ~ .. ... ... .. ... tonged period of operation with cutting operations W111 have to be carefully monitored. The use of liquid anhydrous ammonia, with atten- dant ammonia vapors, is a significant hazard to work- ers because of the toxicity and flammability of ammo- nia (although the flammability range is only 4 to 15 percent). The committee' s concern is that ammonia is not only flammable but can also become a fuel if a fire is initiated by other sources. The committee recognizes that anhydrous ammonia is a widely used industrial chemical, is the fluid of choice in large refrigeration systems, and is extensively used as a fertilizer. The widespread use of anhydrous ammonia by workers with varying levels of familiarity suggests that ammonia could also be used in the proposed systems without significant risks to worker safety. The most significant issue is the necessity of wear- ing DPE suits for maintenance of the equipment in both the extraction and destruction subsystems. The equip- ment should be designed for easy access for the change- out and safe removal of contaminated parts. The design 153 of tumblers and stirred reactors in the destruction sub- systems will be especially important. In addition, the durability and reliability of DPE material in the pres- ence of ammonia and residual SET solution is a con- cern. Although small-scale laboratory tests indicated that DPE degradation was manageable, additional tests will be required to ensure the safety of maintenance personnel. Another safety concern is the ability of the fuze de- struction (detonation) chamber to withstand a large number of explosions. The technology for fuze destruc- tion is well established, but the area of concern is that fuzes will be handled and detonated in an ammonia- vapor atmosphere by remote control. These operations will have to be safe and the maintenance and repair requirements minimal; at the same time, high through- puts will have to be maintained. The ammonia-agent and ammonia-energetic slurries will be mixed with liquid sodium to provide the sol- vated electrons for reducing chemical bonds. The long- term durability of the equipment for performing this mixing must be proven to ensure that maintenance re- quirements are low and opportunities for exposure are . . · . mmlmlzecl. Dunnage and other agent-contaminated materials will be shredded and then reacted with SET solution in a tumbler-type reactor. This technology has not been demonstrated in the presence of these materials. The committee is concerned that worker risk may be in- creased if maintenance requirements are high. Public Safety The committee believes that the likelihood of re- leases of agent or other regulated (hazardous) sub- stances to the atmosphere or the groundwater system at the facility is small. Hold-test-release systems are used for all effluent streams except the containment ventila- tion system, which uses baseline air-cleaning technol- ogy. The primary cause for a release of material con- taining agent or other regulated substances would be a disruptive explosion. The likelihood of such an event is expected to be extremely small at the conclusion of the design process for the full-scale facility. (This de- sign process is understood to include the completion of a quantitative risk assessment.) Ammonia at ambient temperature readily dissipates.
154 Human Health and the Environment Eff/uent Characterization and Impact on Human Health and Environment Teledyne-Commodore has not characterized the ef- fluents from their total system in sufficient detail for the committee to assess potential impacts on human health and the environment. The chemical composition of the products of final oxidation is unknown, and the oxidation processes themselves are still under develop- ment. The technology provider has demonstrated that agent is unlikely to be present in any of the process effluents but has not demonstrated the absence of other chemical compounds of concern. Gaseous Effluents. Teledyne-Commodore proposes scrubbing and recycling all ammonia vapors. Other gases produced in the system include hydrogen, low molecular-weight aliphatic hydrocarbons (propane, methane, ethane, ethylene, and propylene), and nitro- gen. These gases will be passed through carbon-filter beds to remove impurities and burned for use as supple- mental fuel to supply some of the energy for the pro- cess. The gases released to the environment should, therefore, consist primarily of carbon dioxide, water vapor, and nitrogen. Should the scrubbing process not remove all of the ammonia, this ammonia will pass through the carbon beds and be burned as part of the supplemental fuel, creating additional NOx in the com- bustion gases. Carbon beds must be impregnated with phosphoric acid or a similar material to enable adsorp- tion of ammonia. If this type of carbon is used, it must be tested for its ability to remove trace amounts of chemical agent. Liquid Effluents. Teledyne-Commodore proposes recycling all wastewater for reuse in the process. Other anticipated liquid effluents include isopropanol, etha- nol, diisopropyl-(amino) ethane sulfonic acid, and 2,2~- hydroxy diethylether. The first two will be separated in the ammonia-recovery towers and burned with the noncondensable gases as supplemental fuel. The rest will be mixed with cement and sent to a hazardous- waste landfill. The cement stabilization process for these wastes has not been tested. Solid Wastes. The process produces many solid wastes, including metal parts, dunnage, sodium salts, ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS polymeric materials, lead salt, and various organics. These wastes are the end products of final treatment methods that have not yet been optimized. The compo- sition and properties of the products are, therefore, not certain and final disposal methods have not been tested. The salt that Teledyne-Commodore anticipates will be produced in the largest quantity is sodium sulfate- 5.2 lbAb GB; 3.5 lbAb HD; 2.4 lbAb of M28; and addi- tional quantities from the oxidation of VX, TNT, tetryl, tetrytol, and Comp B. Sodium sulfate in itself is not a listed or hazardous waste as defined by the EPA. How- ever, as a product derived from the treatment of haz- ardous wastes (i.e., agents and energetics), sodium sul- fate is subject to RCRA hazardous-waste regulations. Sodium sulfate is quite soluble in water and may have to be stabilized prior to disposal in a landfill. How- ever, sodium sulfate has been known to retard settling and cause spilling of the most common cement-based stabilization agents. A special cement may, therefore, have to be used to prevent leaching of the salt. Completeness of Eff/uent Characterization Final effluents from the oxidation process and the gas-treatment train have not been fully characterized. Chemical compositions and quantities are largely un- known and are likely to change as process conditions are refined. Eff/uent Management Strategy The technology provider has estimated the compo- sition of the process effluents and has proposed dis- posal plans for them. The committee believes that these plans are reasonable. However, further work will be necessary to characterize the effluents accurately. Resource Requirements The resource requirements (including electrical power) for the proposed system are not unusual. Environmenta/ Compliance and Permitting Commodore Remediation Technologies, a predeces- sor of Teledyne-Commodore, has received a nation- wide permit under the TSCA to use solutions of cal- cium in liquid ammonia (known as Agent 313) to
TELEDYNE-COMMODORE SOLVATED ELECTRON TECHNOLOGY PACKAGE remove PCBs from contaminated soils and metal sur- faces. However, the SET technology proposed here will be permitted under RCRA, which is considerably more complicated than permitting under TSCA. Further- more, the proposed technology package has many more components than the permitted PCB system, including SET reactors for many different types of contaminants, both solids and liquids; oxidation reactors for treatment of SET residuals; and processes for preparing the prod- ucts of oxidation for final disposal. Because the pro- cess as a whole (and the individual components) are unique, they will have to be permitted under Subtitle X of RCRA. Because federal and state regulatory agen- cies have limited experience in issuing Subtitle X per- mits, the permitting process will almost certainly be prolonged. Another aspect of the process that may lead to per- mitting delays is the use of cleaned off-gas as a boiler fuel. Extensive testing may be required to characterize this stream to ensure its adequacy for that purpose. STEPS REQUIRED FOR IMPLEMENTATION The technology provider will have to take the fol- lowing steps prior to implementation: · verity products of bibl/hydrolys~s of agents through experimentation · establish optimum conditions for SETlhydrolysis of agents through laboratory tests, followed by pilot-plant demonstration identify the unknown precipitates of SET-energet- ics reactions establish optimum conditions for the oxidation of residuals from SETlhydrolysis of both agents and energetics through laboratory tests, followed by pilot-plant demonstration test waste-disposal methods · pilot test methods for decontaminating metal parts . and dunnage revise the preliminary design for the hypothetical system, especially the interfaces between the unit processes and operations, and demonstrate the re- vised design at pilot-scale 155 FINDINGS Finding TC-1. The use of ammoniajet cutting in the munitions disassembly process could solve some of the problems encountered in baseline disassembly. How- ever, the process must be thoroughly tested to address reliability and maintenance issues. Finding TC-2. Conditions for SET destruction of agents have been reasonably well established but dem- onstrated only on a small scale. Finding TC-3. Conditions for SET deactivation of en- ergetics have not yet been determined. Moreover, en- ergetics have not been completely deactivated in labo- ratory tests, which raises concerns about explosions or other upsets. Finding TC-4. The reaction chemistry is not yet fully understood for either SET destruction of agents or SET deactivation of energetics. Finding TC-5. The products of SETlhydrolysis of agent and energetics have not been adequately charac- terized. Thus, the technology provider cannot be cer- tain that all of the SETlhydrolysis products can be oxi- dized by the persulfate step. Furthermore, the products of oxidation of the SET/hydrolysis products have not been adequately characterized. Finding TC-6. The solid wastes produced by the over- all process have not been characterized well enough to establish whether they are suitable for safe disposal by existing methods, such as landfill, or whether pretreat- ment methods (i.e., stabilization) to convert them to an acceptable form for disposal will be necessary. Finding TC-7. The use of cleaned off-gas as a boiler fuel poses unique permitting challenges. Any process demonstration must characterize this stream to ensure that this off-gas can be permitted as a boiler fuel. Finding TC-X. The full-scale system for hydrolysis of the SET products will differ significantly from the sys- tems used in the reduced-scale tests. Because further chemical reactions occur during hydrolysis, the com- ponents of the full-scale system must be tested.
