Appendix A Selected Destruction Techniques

facility, minute quantities of agent were detected in the CHEMICAL NEUTRALIZATION: brine. At the neutralization THE ARMY’S EXPERIENCE brine was considered agent-free if a 5 percent excess sodium hydroxide level was achieved. At CAMDS a more In a 1969 report, the National Research Council (NRC) strict criterion was used of less than 20 parts per billion recommended chemical neutralization for the destruction (ppb) agent (the Army’s soldier drinking water standard). of the chemical weapons (CW) agent GB and incineration Difficulties in certifying this level of destruction at for mustard agents H and HD. After research and CAMDS may have come from occlusion of GB in rust or development work on chemical neutralization at the other particulate, formation of GB during analysis, or Chemical Agent Munitions Disposal System (CAMDS) false positives resulting from some unidentified interfer- (Tooele, Utah) and Rocky Mountain Arsenal (, ence in the complex neutralization mixture (l). Agent Colorado) in the 1970s, the Army concluded that inciner- emissions at the facility during brine spraying at Rocky ation was the best method for the destruction of all Mountain Arsenal often exceeded the action level (0.0003 chemical weapons. The NRC Committee on Chemical 3 milligram per cubic meter (mg/m )) and occasionally the Weapons Disposal has recently been asked by the Army 3 shutdown level (0.003 mg/m ). These levels were promul- to reevaluate incineration and alternatives for CW dis- gated by the Department of Health and Human Services posal. (DHHS) and the Army’s Surgeon General. However, perimeter monitors showed that the emission standard for Chemical Neutralization of Nerve Agents the general population was not exceeded (l). Most of the Army’s experience with large-scale chemi- cal neutralization of chemical weapons was with the After being drained of GB, the empty munition bodies organophosphorus ester agent GB. This agent was suc- were moved to a deactivation furnace where explosives cessfully neutralized using aqueous sodium hydroxide on and propellants were incinerated and metal parts ther- a scale compatible with destruction of the current U.S. mally decontaminated, i.e., incinerated. Empty ton con- CW stockpile. Approximately 8.4 million pounds of tainers were similarly incinerated in separate furnaces. GB-taken from underground storage tanks, GB ton Thus, “chemical neutralization” actually applied only to containers, M139 bomblets (Honest John Warhead), M34 the drained agent and treatment of the remaining waste cluster bombs, rockets, and 155/105-mm projectiles, depended on incineration. However, disposal of the M34 were neutralized at Rocky Mountain Arsenal from 1974 cluster bomblets and ton containers used a caustic to 1976 and at CAMDS between 1979 and 1982. By (aqueous sodium hydroxide) wash to treat the drained weight, this represents 17 percent of the 25,000 tons of container by neutralizing any residual agent. For some agent to be destroyed in the current program. reason—possibly a lack of confidence in the efficacy of GB is stable at neutral pH but is hydrolyzed rapidly at this process-the caustic wash treatment was also fol- alkaline pH. The half-life of GB at 300 C in aqueous lowed by thermal decontamination (incineration). solution is 146 hours at pH 7 (neutral conditions) but decreases to 0.4 hour at pH 9 (alkaline conditions) (l). From 1979 to 1981,13,951 M55 rockets containing GB Presumably at higher pH and temperature, the hydrolysis (2.9 percent of the current M55 rocket stockpile) were rate would be even more rapid. The suggestion has been destroyed by this combined chemical neutralization/ made that the addition of a catalyst could speed up this incineration process (l). The Army also reported prob- hydrolysis reaction even more (2). lems with re-formation of GB during the brine drying process, although it is not clear why the corrective actions As part of the Army’s program, after GB neutralization described above were not applied to solve the problem in was determined to be complete, the resulting brine was evaporated by spray drying and the salts were packed into this instance. The reaction was also reported to take longer drums for disposal. There were some problems with the than expected. Adding excess sodium hydroxide to spray-drying process, including the possibility that GB accelerate the reaction created a larger amount of salt for might re-form under certain conditions. This re-formation disposal. Given the intrinsically rapid hydrolysis rate of could be successfully avoided by adjusting the pH and GB under alkaline conditions (corresponding to a short brine flow rate, and by reducing the operating temperature half life), the apparent slow reaction encountered in this (l). situation may have been due to problems associated with the large scale of the demonstration such as complete and Difficulties were also encountered in confirming that thorough mixing of the organic material with the aqueous the brine was agent-free. Particularly at the CAMDS sodium hydroxide. –31- 32 ● Disposal of Chemical Weapons: Alternative Technologies

Although the agent VX, which is structurally similar to lead to re-formation of the agent. In some situa- GB, can also be chemically neutralized, this was never tions, the rate of neutralization in large-scale tests of demonstrated by the Army on a large scale. Acid chemical agents was much slower than had been chlorinolysis (chlorination in an aqueous acidic medium predicted. Very large amounts of impurities in followed by caustic neutralization) rather than alkaline certain grades of mustard agent also made neutrali- hydrolysis was suggested by the Army as the best method zation difficult to monitor adequately. However, it for chemical destruction of VX (l). However, VX was is not clear that the problems encountered with shown to be neutralized on a small scale by hydrolysis ‘‘industrial scale-up” of chemical neutralization are with sodium hydroxide. The problems encountered with insurmountable, and the scale of a chemical neutral- the neutralization of GB led the Army to abandon plans ization program is similar to or smaller than that of to test the large-scale neutralization of VX. The Army industrial large-scale processes. indicated that a poor water volubility, mixing problems, 3. The quantity and nature of the waste produced by and the presence of a “bis” impurity (with unspecified neutralization are more problematic than those susceptibility to alkaline hydrolysis) made alkaline hy- produced by incineration. Calculations by the Army drolysis of VX difficult (l). As with GB, neutralization indicated that 1 pound of GB will produce 1.5 was apparently intended only for the liquid chemical pounds of salt, compared to a salt yield of 1.4 agent. Incineration was to be used for destruction of the pounds from incineration. In practice, the excess explosives and propellant components, and for thermal caustic added to speed up the reaction led to 2.6 decontamination of munition cavities and metal parts. In pounds at Rocky Mountain and 3 to 6 pounds at addition, lack of a reliable low-level monitoring capabil- CAMDS of salt per pound of agent hydrolyzed. The ity for VX in the neutralization brine at the time of the Army speculated that the sometimes heterogeneous Army’s research program led it to conclude that DHHS form of some agents (partially gelled, mixed with would never approve chemical hydrolysis (l). solid particles such as rust) may have contributed to the variation in results obtained with chemical Chemical Neutralization of Mustard Agents neutralization. These types of technical problems Mustard agent has also been shown in the Army’s encountered in the transformation of industrial research to be hydrolyzed under alkaline conditions on a processes from bench-scale demonstrations may not small scale, although only slowly at ambient temperature. be insurmountable given sufficient motivation to The corresponding reaction rates with alkaline hydrolysis reach a solution. Also, it has been argued that at elevated temperature were not reported. The products producing larger amounts of salts from neutraliza- of mustard hydrolysis with sodium hydroxide were not tion may be relatively more acceptable than some of identified and their toxicities were not assessed. Alkaline the perceived problems of incineration, such as the hydrolysis of mustard agents on a pilot-plant scale was formation of dioxins. reported, using the base monomethanolamine instead of 4. The capital and operating costs of chemical neutral- sodium hydroxide, to produce a homogeneous nontoxic ization were estimated to be higher than those of organic waste. Calcium hypochlorite slurry or aqueous incineration (1). This cost comparison might have bleach (sodium hypochlorite) was used to oxidize rather to be considerably revised now in light of the than hydrolyze mustard agents, but there was “uncer- unanticipated cost increases in the Army’s inciner- tainty about the completeness of the reaction” (l). ation program, which are due in part to technical problems encountered by the Army after 1987 (3). Summary of the Army’s Neutralization 5. The analytical problems encountered in certifying Process Experience that the waste materials of chemical neutralization were agent-free. This must be compared to the In the Army’s summary of its experience with chemical analytical problems faced in demonstrating that neutralization, the following reasons were given for incineration products of incomplete combustion abandoning the process in favor of incineration (l): (PICs) are dioxin or agent-free, etc. 1. The perceived complexity of neutralization com- After reviewing the Army’s experience with chemical pared to incineration. Caustic reactants used for neutralization and incineration of chemical weapons, the chemical neutralization had to be handled safely in NRC in 1984 supported the Army’s decision to abandon bulk quantities. However, many industrial large- chemical neutralization in favor of incineration. In light of scale processes routinely use such caustic agents, the current political opposition to incineration, and after e.g., the manufacture of soap, The major safety issue considerably more experience with this technology, it is in handling chemical weapons will always be the not clear that the same endorsement would be made today. agents themselves. In principle, with appropriate conditions, alkaline hy- 2. The sensitivity of neutralization to a number of drolysis could be a means to chemically neutralize the variables that could slow the reaction and possibly CW agents GB, VX, and mustards. Problems encountered Appendix A-Selected Chemical Weapon Destruction Techniques ● 33 with alkaline hydrolysis of all types of CW agents might removal efficiencies” (DREs) of 99.99999 percent and today appear to be surmountable in view of new has proposed, but not begun, a demonstration of the techniques as well as increased pressure to exploit technology for DARPA with actual agents. Although existing techniques for use with CW destruction. The Modell has not yet worked with actual CW agents, it has NRC’s Committee on Review and Evaluation of the demonstrated the oxidation of some explosive materials Army Chemical Stockpile Disposal Program and Com- such as the effluent from TNT manufacture (’red water’). mittee on Alternative Chemical Demilitarization Tech- nologies, were recently requested by the Army to The formation and environmental release of dioxin, a reevaluate incineration and alternatives for CW disposal. concern in the incineration of organic materials, may not be significant for SCWO. Modell has demonstrated that the dioxin congener (TCDD) was oxidized to below SUPERCRITICAL WATER detectable levels when introduced at 500 parts per million OXIDATION (ppm). The Modell technology has also been demon- strated to work for the destruction of dioxin in pulp mill Supercritical water refers to water that has been heated waste streams. Wood, when suitably pulverized and and pressurized to a transition point between gas and converted to a flowable form, can also be treated. liquid phases, and thus has some of the properties of both. In the supercritical phase the solvent properties of water Gaseous effluents from SCWO are carbon dioxide and change, and organic materials becomes soluble, whereas oxygen with traces of carbon monoxide (10 to 15 ppm inorganic salts become insoluble and tend to precipitate with optimized operation). In Modell’s process, these (4). Organic materials in solution with supercritical water effluent gases are expanded and cooled; the carbon can be oxidized by oxygen introduced from air. This is a dioxide is solidified and removed, and the oxygen is broad-spectrum oxidation procedure for organic com- recycled through the system. This process might be pounds, and at elevated temperature (4000 C), even considered a “closed” system in comparison with refractory compounds such as coal are oxidized (5). incineration. Oxidation of organic materials containing Supercritical water oxidation (SCWO) is similar to hetero atoms such as fluorine, , sulfur, or incineration in that it involves oxidation of organic phosphorus produces the corresponding hydrofluoric, compounds to carbon dioxide and inorganic acids or salts. hydrochloric, sulfuric, and phosphoric mineral acids in However, SCWO operates at much lower temperatures solution. These can be neutralized, precipitated from the than incineration. SCWO is under commercial develop- SCWO reactor water solution, and removed from the ment as a general technology for the destruction of many SCWO reaction vessel. different organic hazardous materials, and the destruction Although Modell believes that its SCWO technology of CW agents would at most constitute only a small could be developed for the Army’s CW destruction portion of its use. program, it has been unable to interest the Army in this SCWO may have certain advantages over incineration work. From Modell’s perspective, the Army’s commit- for the oxidation of organic waste. Compared to inciner- ment to incineration technology has led it to dismiss ation, SCWO has no requirement for a large airflow. viable alternatives such as SCWO. On the other hand, the SCWO carries out oxidation at lower temperature, and the U.S. Department of Energy (DOE) has apparently ex- reaction medium (water) can be contained until it is tested pressed an interest in using the technology to treat to be safe. A major selling point of this technology is that radioactive mixed waste. In general, Modell considers potential PICs are entrained in solution rather than that the greatest commercial opportunities and market for emitted in stack gases. The apparently superior control of this technology are not in CW destruction but in other emissions is an attractive feature of SCWO technology. areas of organic waste disposal. The Modell Corp. is The effluents from SCWO, in contrast to exhaust stack currently operating a bench-scale SCWO facility with a gases from incineration, may be collected, analyzed, and capacity of 30 gallons per day and hopes soon to construct even recycled to achieve more complete destruction. an SCWO facility in Germany with a capacity of 5 to 10 tons a day. Modell estimates a cost of $200 per ton for In discussions with Modell Corp. and General Atomics, sludge disposal. two companies that are involved in the development of SCWO for CW agent destruction, it was apparent that General Atomics’ (GA) engineers who are involved currently there are both advantages and limitations with with SCWO development are considerably more cautious this technology. Modell has completed the initial phase of about the future uses of this technology for hazardous a research program on the use of SCWO for the treatment waste and CW agent disposal (6). Although a wide range of CW agents (5). It has successfully demonstrated for the of organic compounds have been shown by GA to be Defense Advanced Research Project Agency (DARPA) oxidized by SCWO technology on a bench scale, it has the destruction of “simulants” (analogs) of GB, VX, and never been demonstrated with actual CW agents. GA is HB on a bench scale. Modell achieved “destruction and currently under contract to DARPA and the Office of 34 ● Disposal of Chemical Weapons: Alternative Technologies

Naval Research to perform the necessary research and water. Since steam gasification is a reducing process build a prototype SCWO system that would be capable of rather than an oxidation process such as incineration PICS processing relatively small amounts of chemical agents, will not be formed. This reaction is a temperature- propellants, and other U.S. Department of Defense controlled equilibrium process so, in principle, any level wastes. SCWO as it is presently conceived is not designed of destruction may be achieved by selecting an appropri- to handle solid CW agent disposal waste forms such as ate temperature. drained munitions or dunnage. The research program, The current developer is interested in testing its steam which will be conducted with support from the University of Texas, the IIT Research Institute (Chicago, Illinois), gasification system for CW destruction. It has demon- strated the device with CW simulants but not with actual and EcoWaste Technologies (Austin, Texas), will ini- CW agents. The machine is designed to handle bulk tially focus on two areas of importance-corrosion and solids handling. Corrosion with SCWO is a significant objects, including 55-gallon drums filled with waste, and issue especially because of the mineral acids formed from may therefore be suitable for handling properly prepared the oxidation of compounds containing fluorine, chlorine, munitions and CW containers. Drummed wastes are sulfur, and phosphorus hetero atoms. The handling of gasified in the drums, one drum at a time, by placing each solids, such as the salts formed and precipitated during drum in a chamber operated at 300 to 700 ‘C. In its current form, halogens are removed from organic materials by SCWO, will require careful attention because such inorganic salts can become sticky under SCWO condi- pretreatment with alkali, and the remaining dehaloge- tions and could foul the walls of the reactor. nated organic material is destroyed by steam gasification. Sulfur- and phosphorus-containing materials are con- The DARPA contract calls for a 15-month research verted to salts of their reduced forms hydrogen sulfide and phase to address these and other issues and to develop phosphine. In contrast to conventional incineration, steam data, initially on simulants and then on actual CW agents. gasification does not use an airflow so the gas flow output Concurrently, a 3+-year pilot-plant development effort is to the environment is minimized. planned to lead to the construction and testing of a prototype SCWO unit with a capacity of 1,000 to 1,500 The developer is offering a “premarket” machine for gallons per day. Full demonstration of the practical customer evaluation. The machine is approximately 5 by 6 by 7 feet and costs about $700,000 when supplied with application of this technology to CW agent destruction is therefore estimated by GA to be more than 3 years away. a drum feeder and flash vaporizer. The company is already working on chemical waste disposal problems for In discussing the likely success of this application of DOE, such as destruction of chlorinated solvents and SCWO, GA tended to be cautious and to emphasize the organic wastes from weapons dismantlement, that do not need for the data that will be developed during the involve chemical weapons. It has proposed a system that research phase of this work would use one or more mobile steam gasification devices directly in a CW storage igloo or on a truck positioned GA considers the recent bilateral, as well as current and next to the igloo. In this concept, a modified version of the future multilateral, agreements on CW destruction and Army’s current M55 rocket shearing system could be nonproliferation to be an ideal opportunity for SCWO coupled to the drum feeder. Passage of hot steam around technology. For example, the United Nations has an- and through the sheared, drained munition would decon- nounced its intention to destroy the Iraqi CW stockpile, taminate the empty shell. This configuration would have which is considerably smaller than that of the United the advantage of avoiding the risk associated with States. Use of mobile SCWO-based machines for destruc- transporting CWs from the igloo to a larger stationary tion of this type of small stockpile could be an ideal destruction device. application. Unfortunately, the technology may not be developed in time to be applicable to this particular The developer is currently negotiating a contract to situation. destroy napalm bombs located at Camp Pendleton, California, by use of steam gasification. These napalm STEAM GASIFICATION bombs consist of a 15-foot-long aluminum container filled with napalm and an explosive. The company plans TECHNOLOGY to modify its device to contain an entire bomb during the A process known as steam gasification (re-forming) steam gasification destruction process. treats organic materials with steam at 1,000 to 1,300 ‘C under reducing conditions to produce carbon monoxide, carbon dioxide, and hydrogen (7). The steam/organic PLASMA ARC TECHNOLOGY materials stream is recirculated through the high tempera- A newly developed Plasma Arc Technology System ture reactor to control destruction efficiency. The carbon can process nearly 10 pounds per minute of solid waste or monoxide and hydrogen can be reacted over a suitable 55 gallons per hour of liquid waste (4). The developer is catalyst to form either methanol or carbon dioxide and currently testing plasma arc pyrolysis (PAP) technology, Appendix A-Selected Chemical Weapon Destruction Techniques ● 35 although it has not specifically tested the system with CW Chemical Demilitarization, Aberdeen Proving agents. After waste has been atomized in the plasma Ground, MD, September 1987. pyrolysis chamber the resulting elements are cooled in a 2. Picardi, A., International Environmental Assessment second portion of the chamber and recombine to form Group, Alexandria, VA, “Alternative Technologies hydrogen, carbon monoxide, and hydrochloric acid. Such for the Detoxification of Chemical Weapons: An plasmas can reach temperatures of 5,000 to 15,0000 C. Information Document,” prepared for Greenpeace The resulting gases are passed through a wet caustic International, Washington, DC, May 1991. scrubber for removal of particulate and hydrochloric 3. U.S. Congress, General Accounting Office, Stockpile acid. The remaining gases are combusted with air. The Destruction Delayed at the Army’s Prototype Disposal entire system fits on a 45-foot-long transportable tractor- Facility, GAO/NSIAD-90-222 (Washington, DC: July trailer bed. 1990). 4. U.S. Congress, Office of Technology Assessment, The most significant limitation of PAP treatment is that Dioxin Treatment Technologies-Background Paper, only liquids can be treated. Contaminated soil and viscous OTA-BP-O-93 (Washington, DC: U.S. Government materials cannot be processed by the system (4). Thus Printing Office, November 1991). PAP technology in its current form would not be suitable 5. Modell, M., Modell Development Corp., Framing- for the treatment of contaminated drained munitions, ham, MA, personal communication, Nov. 14, 1991. containers, explosives, or propellants, and the dunnage 6. Davison, W., Director, Advanced Processing Systems, associated with chemical weapons. General Atomics, San Diego, CA, personal communi- Appendix A References cation, Nov. 19, 1991. 7. Galloway, T., Synthetic Technologies Inc., 1. U.S. Department of the Army, “Chemical Agent and Richmond, CA, and Sprung, J. L., Sandia National Munition Disposal; Summary of the U.S. Army’s Laboratories, Albuquerque, NM, personal communi- Experience,” Office of the Program Manager for cation, Dec. 20, 1991.