Click for next page ( 35

The National Academies of Sciences, Engineering, and Medicine
500 Fifth St. N.W. | Washington, D.C. 20001

Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement

Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 34
Worker Protection and Potential for Offsite Release The recovered chemical warfare materiel at the Pine Bluff Non-Stockpile Facility (PBNSF) varies more widely in type, agent content, and physical condition than do the stockpile items at the same location. Many of the non-stockpile muni- tions stored at the Pine Bluff Arsenal (PBA) were recovered from burning pits and are badly corroded and difficult to characterize. This chapter addresses the unusual challenges presented by the handling of recovered chemical warfare materiel, explosively or nonexplosively configured, intended for PBNSF. It will not review nonmunition safety concerns, such as conventional industrial accidents and fires not involving either agent or high explosives. PROTECTING PINE BLUFF NON-STOCKPILE FACILITY PERSONNEL FROM EXPOSURE TO CHEMICAL WARFARE AGENTS PBNSF personnel will be protected from exposure to non- stockpile agents by personal protective equipment (PPE), by barriers, and by engineering controls. The PBNSF process areas are categorized by hazard type, and personnel are required to wear different levels of PPE depending on the hazard category of the process area in which they are work- ing. The allowable concentration of agent in the air in a given area determines the type of PPE required. The air in areas requiring Level A PPE is monitored for agent at the gross detection level (GDL) or the maximum permissible limit (MPL). These areas are usually monitored to the GDL (0.2 mg/m3) but must be below the MPL (100 mg/m3) for H/HD and HN-3 (U.S. Army, 2003a). The air in the agent- free areas is monitored at the 8-hour time-weighted average (TWA) level.) The TWA monitoring level for H/HD and iTWA is the permissible 8-hour airborne concentration of a chemical agent to which a worker may be exposed. A IWA exposure limit is gener- ally set so that workers may be exposed 5 days per week for a working lifetime with minimal risk of adverse health effects. 34 HN-3 is 0.003 mg/m3. The agent-free areas are monitored 24 hours a day, 7 days a week, to assure a safe working environment. Pine Bluff Non-Stockpile Facility Personal Protective Equipment and Characterization of Process Area Hazards PBNSF operating personnel entering known or poten- tially agent-contaminated areas (i.e., hazard Categories A and B) will use an industrial Level A zipper-type PPE suit.2 These suits are totally encapsulating (vapor-tight) chemical protective suits with positive pressure, full facepiece, self- contained breathing apparatus and have been approved by the National Institute for Occupational Safety and Health (U.S. Army, 2002a). Hazard Category A and B areas are under negative air pressure; Category A areas may be contaminated with liquid agent and are assumed to be contaminated with agent vapor; Category B areas may be contaminated with agent vapor. Further details concerning hazard categories, as well as the hazard categories assigned to the various process areas of PBNSF, are provided in Table 3-1. Pine Bluff Non-Stockpile Facility Chemical Agent Monitoring Devices The agents to be monitored are sulfur mustard (H. HS, and HD) nitrogen mustard (HN-3) arsenicals (PD) 2Vivian Graham, Non-Stockpile Chemical Materiel Product, "Pine Bluff Non-Stockpile Facility Discussion: Safety Issues," briefing to the commit- tee on April 22, 2003.

OCR for page 34
WORKER PROTECTION AND POTENTIAL FOR OFFSITE RELEASE The air monitoring equipment being considered for PBNSF includes: MINICAMS3 with halogen-specific detectors for mustard-containing chemical agents Depot Area Air Monitoring System (DAAMS) moni- tors (for chemical agents) MINICAMS adapted for detection of PD (phenyldichloroarsine) when processing GTRs filled with arsenicals . The MINICAMS and the DAAMS will be used to moni- tor for chemical agents at the TWA level throughout the facility. The MINICAMS is also used to monitor for agents at the GDL and MPL levels. The DAAMS is frequently used for perimeter monitoring at facilities, but it is unclear whether perimeter monitoring will be conducted. Category A areas at PBNSF are designed to contain po- tentially high concentrations of airborne agent vapor. Per- sonnel are required to wear Level A protective gear in all areas where concentrations could approach levels greater than the immediately dangerous to life and healths level if there is an accident. This approach protects workers from doses that could lead to acute effects. MINICAMS The MINICAMS is considered a near-real-time auto- mated air sampling system with response times that are typi- cally 4 min for the GDL or MPL levels and 10 min for the TWA level. The MINICAMS captures the agents on sor- bent; the agents are then desorbed into a gas chromatograph. The MINICAMS is configured with a halogen-specific de- tector (the primary chemicals anticipated are H/HD, MN-3, and industrial arsenicals, most of which contain the halogen chlorine). The accuracy of the proposed MINICAMS-XSD (halogen-specific detector) units is around the TWA level. When operating in an atmosphere containing a lower con- centration of agent (<10 TWA), the air is sampled directly onto a solid sorbent sample tube. In environments with higher concentrations of agent (>10 TWA), the flow auto- matically switches to a low volume sample loop before the 3MINICAMS is a low-level, near-real-time monitor typically used to provide early warning of airborne exposure hazards. The MINICAMS-XSD (halogen-specific detector) unit is an automated air sampling system that collects compounds, thermally desorbs them into a capillary gas-chroma- tography column for separation, and detects the compounds with a halogen- specific detector (U.S. Army, 2000). The combined sampling and analysis time for the MINICAMS is 3 to 10 minutes, depending on the agent being examined (U.S. Army, 2003a). 4Immediately dangerous to life and health is the maximum exposure concentration from which an individual could escape within 30 min without experiencing escape-impairing symptoms or irreversible health effects. 35 TABLE 3-1 Hazard Categorization of PBNSF Process Areas Process Area Hazard Categorya Receiving storage area (includes warming area) Unpack area Fill extraction preparation area Fill extraction area Detonation chamber area Decontamination area Holding tank area Agent treatment area Metal parts repackaging and storage area D C A/B A A/B A A A, B. C C/D aHazard categories are defined as follows: A identifies a toxic process area under negative pressure, possibly contaminated with liquid agent and assumed to be contaminated with agent vapor. It is a high-hazard area, requiring the use of PPE. B indicates a toxic process area under negative pressure, possibly contaminated with vapor chemical warfare agent. It is a high hazard area requiring the use of PPE. C indicates work areas under negative pressure and subject to inadvertent chemical warfare agent vapor contamination. It is considered a low agent hazard area; protective gear is not required to be worn unless monitoring indicates a need. Although chemi- cal warfare agent contamination of Category C areas is not expected, PPE must be available for use in Category C areas. D indicates areas under ambient pressure that are not subject to contamination. It is considered a negligible chemical warfare agent hazard area. These areas are typically mechanical and electrical equipment support rooms and the facility perimeter. SOURCE: Adapted from U.S. Army (2002d). solid sorbent tube to allow monitoring at higher levels (MPL). Depot Area Air Monitoring System Monitors In the event of a TWA MINICAMS alarm, the DAAMS is used to confirm the MINICAMS reading. The samples will be collected using a vacuum pump, a sequencer, and DAAMS sample tubes. The DAAMS sample tubes are packed with solid sorbent to trap the airborne chemicals and will be taken to the laboratory for analysis. DAAMS moni- toring is also used for analysis of the air lines of the life support systems to assure personnel they are agent free and for historical monitoring at DAAMS-only sample stations in Category D work areas. The DAAMS monitor is based on solid sorbent preconcentration of the sampled air, followed by thermal desorption and analysis by gas chromatography using a flame photometric detector. Sample vapors are passed directly into the sorbent tube. The preconcentrator tubes are inserted into a heated inlet, where the contents are desorbed into a gas chromatograph. A sulfur band-pass filter and linearizer circuit are used to detect chemical agent. Knowing the amount of chemical agent on the sorbent tube and the total volume of air sampled, the average agent con-

OCR for page 34
36 centration in the air can be calculated. By increasing the sample time or flow rate, the average concentration sensitiv- ity can be increased. The DAAMS is much more sensitive, and therefore more accurate, than the MINICAMS method (the sensitivity of the DAAMS monitors is <0.0006 mg/m3~; however, the response time is 1 to 12 hours.5 Arsenicals Monitoring The Army plans to use MINICAMS adapted for detection of phenyldichloroarsine (PD) for air monitoring when pro- cessing German Traktor rockets (GTRs) filled with arseni- cals (U.S. Army, 2003c). PD is the only arsenical to be con- tinuously monitored because it is the only one with significant vapor pressure and blistering properties. Other industrial arsenicals either have very low vapor pressures or lack vesicant properties. Monitoring of airborne arsenic for historical purposes will be carried out by drawing ambient air through filters that will collect PD and the less volatile arsenical agents as well as arsenic-containing particulates (U.S. Army, 2003a). Af- ter sampling, the filters will be digested to convert the ar- senic-containing materials into an aqueous solution. The arsenic content of the solution will be determined by con- ventional means such as atomic absorption spectroscopy or inductively coupled plasma analysis (U.S. Army, 2003a). PROTECTION OF PINE BLUFF NON-STOCKPILE FACILITY PERSONNEL FROM ACCIDENTAL DETONATIONS PBNSF must be designed to withstand the accidental detonation of a munition undergoing treatment while mini- mizing the release of toxic chemical agents to the atmo- sphere. Adequately withstanding accidental detonations will be defined in terms of the explosion containment require- ments of the individual areas at PBNSF where accidental detonations could occur (Chapter 2~. Munitions in specially designed overpack containers are brought from the storage igloos to the PBNSF receiving dock and then into PBNSF at the receiving/storage areas, where they will be warmed during cold weather. After warming, the overpacked munitions are moved to the unpack room, where they are checked for leaking agent. Leakers will be kept in their overpack and returned to storage for treatment at the end of the destruction campaign for the agent being processed. Nonleakers are removed from the overpack, marked as either explosively or nonexplosively configured, and transferred to the fill extraction preparation area. In the fill extraction area, munitions are emptied either in an explosive containment chamber (ECC) or by the projec- 5A variance in response time can be due to the use of different analytical techniques and/or different sampling times. lions: ASSESSMENT OF THE ARMY PLAN FOR THE PINE BLUFF NON-STOCKPILE FACILITY tile washout system (PWS), depending on whether or not they are explosively configured. After draining the agent and washing out the munition in the ECC or PWS, the munition may be placed in a heel-dissolving tank before being placed in the detonation chamber (DET), where shaped explosive charges are used to access and detonate emptied munitions that are explosively configured. Accidental detonations are possible at the following loca- Receiving dock. This area is the outer entry point for munitions coming from the storage igloos and is out- side of the designed containment of PBNSF. There- fore, the receiving dock could be the site of the most significant dispersion of agent from an accidental deto- nation. However, munitions brought to the receiving dock will be overpacked in containers that are spe- cially designed to contain leaks from the munitions within, which will reduce the possibility of a release of agent to the atmosphere. Accidental detonations, how- ever, would not be contained by the overpack contain- ers and would result in the release of explosive force and agent to the atmosphere. Receiving/storage area. Overpacked munitions are held in this area and, if necessary, warmed prior to further processing. This area is inside PBNSF and the effects of an accidental detonation would be greatly mitigated by facility containment and the munition overpack. Unpack area. Once munitions are removed from their overpacks in the unpack area, the potential for agent release during an accidental detonation is dependent on the containment design of PBNSF. Fill extraction preparation area and fill extraction and detonation area, including ECC-1, ECC-2, DET, PWS, auxiliary processing vessels, and DET staging area. Munitions are handled and moved through the various emptying and cleaning processes in this area and are subject to accidental detonation as a result of handling errors and mishaps. Significant amounts of agent could be released if a detonation occurs outside an ECC. Process hazard analyses are management tools used to examine the accidents that could happen in segregated areas of PBNSF. These analyses describe the accidents that could happen during the processes that take place in the different areas of PBNSF. Design changes or engineering controls are instituted to reduce the probability of their occurrence or to minimize the impacts if they do occur. Another manage- ment tool, job hazard analyses, describes accidents that could occur during individual operations in the distinct process areas. The process and job hazard analyses were being per- formed as this report was being prepared and were not avail- able for review by the committee.

OCR for page 34
WORKER PROTECTION AND POTENTIAL FOR OFFSITE RELEASE Finding 3-1: The segregation of operations in the PBNSF appears to be appropriate; however, the likelihood of acci- dental detonation cannot be estimated until the process haz- ard analyses and job hazard analyses have been completed. Recommendation 3-1: The Army should complete process hazard analyses and job hazard analyses to provide critical information before finalizing the design of the Pine Bluff Non-Stockpile Facility. PROTECTION OF THE PUBLIC AND THE ENVIRONMENT The PBNSF is not designed to contain an occurrence of the maximum credible event (MCE).6 The MCE at PBNSF is the accidental detonation of a com- plete GTR (full warhead and rocket motor), resulting in an energetic force of approximately 17 lb of trinitrotoluene equivalent and the dispersion of 7 lb of agent. The PBNSF building, however, is not designed to contain the effects of the MCE.7 Rather, it is designed with blowout panels, which will vent internal pressure in the event of the MCE. This design seems not to be consistent with the mandate to provide maxi- mum protections for the environment, the general public, and the personnel who are involved in the destruction of the lethal chemical agents and munitions (50 U.S.C. Section 1521(c)~1~(A)~. In the committee's view, the only methods of providing maximum protection are to redesign the building to contain the current MCE or to take action to reduce the MCE so that the current design is adequate to contain it. This and other issues should be resolved in the systematic design inte- gration review (see Recommendation 2-3~. The committee notes that the Army's calculations for the no-effects distances for the MCE show that it extends far beyond the boundaries of PEA (U.S. Army, 2002b). How- ever, no deaths outside the boundaries of PEA are calculated to be caused by an MCE at PBNSF (U.S. Army, 2002b). The committee did not peer review the air dispersion modeling performed for the no-effects distance calculation, 6The MCE is defined as the worst single event that could occur at any time with the maximum release of a chemical agent from a munition, con- tainer, or process as a result of unintended, unplanned, or accidental occur- rence (U.S. Army, 1999). 7Peter Wells, Task Engineer, Shaw Environmental, Inc., "PBNSF Bounding Challenge to HVAC Filters," briefing to the committee on Au- gust 1, 2003. The term "maximum protection" is defined in the Defense Appropria- tion Act of 1996. Such terms are generally applied on a case-by-case basis. See Appendix D in (NRC, 1999) for a compilation of Army definitions and to learn how this term may fit into general regulatory risk management policy. 9The no-effects distance is the downwind distance beyond which no adverse human health effects (em., excessive contractions of the pupil of ~ "a- 7 the eye, muscle tremors, airway tightening, nausea, vomiting, and diarrhea) would be expected to occur (U.S. Army, 2002b). 37 although members of the committee did perform some lim- ited confirmatory calculations that included meteorological data; these calculations suggested that the Army's calcula- tions were conservative (i.e., more likely to overestimate concentrations than to underestimate them). The committee recognizes that the Army is investigating possible methods for removing the rocket motors from the warheads so that only the warheads will enter PBNSF, and it encourages that effort. Based on the information reviewed by the committee, one or more methods of removing the motors appear feasible. Finding 3-2: The committee finds that the safety of personnel outside the PBNSF may be compromised because the building is not designed to contain the release of agent from the MCE. Separating the warhead from the rocket motor and process- ing only the warhead in PBNSF will increase the safety of operations inside PBNSF by eliminating the only situation where the energetic capacity of the munition exceeds the containment capacity of the building. Recommendation 3-2: The German Traktor rocket war- heads should be separated from the rocket motors and only the warheads should be allowed to enter the Pine Bluff Non- Stockpile Facility so as to reduce the maximum credible event to a level that can be fully contained by the structure. The Army should continue to investigate thoroughly the fea- sibility of separating the German Traktor rocket motors from their warheads to determine how and where these operations can be accomplished safely. External Monitoring The heating, ventilation, and air conditioning system will be monitored at the TWA level using both MINICAMS and DAAMS placed at the midpoint of the carbon filters, at the filter-housing vestibules, and at the effluent stack of the heat- ing, ventilation, and air conditioning system. It is unclear whether perimeter monitoring will be performed. The PEA perimeter monitors (DAAMS) for the stockpile disposal fa- cility might be employed to fulfill this function. Once the systems contract for PBNSF is awarded, a site monitoring plan will be finalized in coordination with the PEA, the Pine Bluff Chemical Activity responsible for stock- pile storage, and the Centers for Disease Control. This moni- toring plan presumably will encompass the ability to distin- guish the point from which any agent is being emitted. Sampling and Analysis of Liquicl and Solicl Secondary Wastes at the Pine Bluff Non-Stockpile Facility Liquicl and Solicl Waste Streams As noted in Chapter 2, the agent contained in each muni- tion is sampled and analyzed, first at the Pine Bluff muni-

OCR for page 34
38 lions assessment system and again at PBNSF, to ensure that the subsequent processing and monitoring operations are appropriate. In general, processing of the munitions gener- ates three secondary liquid waste streams: neutralents spent decontamination solutions miscellaneous process fluids (hydraulic fluids, sol- vents, etc.) The first two liquid waste streams will be monitored for chemical agent at various points during processing at PBNSF. All three will be sampled and analyzed before re- lease from Army control. Munitions processing generates several types of second- ary solid wastes: metal scrap from munition bodies spent carbon from filters miscellaneous solid wastes (wipes, personal protective equipment, dunnage, overpacks, etc.) The metal scrap may either be decontaminated to the 3Xi level and sent to another Army site or government contractor for recycling or thermally decontaminated to the 5Xii level on-site and released to civilian recyclers. The spent carbon and other solid wastes will be sent to a treatment, storage, and disposal facility (TSDF) for treatment and disposal after an analysis of the vapors in the headspace of the waste con- tainers to ensure that the waste conforms to the 3X release standard. Planning for the closure of PBNSF will require analytical procedures that can certify the suitability of materials such as soil, concrete, and metal for recycling or disposal. A pre- vious National Research Council (NRC) report (NRC, 2001 c) noted that agent in soil or concrete "is a potential problem during cleanup and closure operations when these materials must be certified as agent free." The Armv' s cro- gram for stockpile demilitarization is developing experience on the necessary analytical procedures. Research on the fate of chemical agents in the environment may also be relevant (Rosenblatt et al., 1996~. i3X refers to a level of decontamination at which solids are suitable for transport for further processing. ii5X refers to a level of decontamination at which solids may be re- leased for general use or sold (e.g., as scrap metal) to the general public in accordance with applicable federal, state, and local regulations. A common misconception is that 5X means simply that the solid has been placed in a temperature zone of 1000F or higher for 15 minutes. In fact, a 5X condi- tion indicates that the solid has been completely heated to and then held at a temperature of at least 1000F for 15 minutes. ASSESSMENT OF THE ARMY PLAN FOR THE PINE BLUFF NON-STOCKPILE FACILITY Liquids Sampling and Analysis Neat agent. As previously noted, PBNSF operations will include a reconfirmation of the contents of each munition to be processed. Vapor samples will be drawn from within the auxiliary processing vessel containing the munition body af- ter initially accessing the munition fill. Similarly, when muni- tions are being processed in the PWS, a hole is drilled in the projectile body to permit removal of a vapor sample for analy- sis. The agent vapor will be analyzed by gas chromatography and mass spectrometry according to Army procedures. Aqueous solutions. The neutralents generated in the chemical processing trailer reactors as well as other aqueous streams (e.g., water and 10 percent sodium hydroxide rinses) likely to be contaminated with agents will be analyzed for agent concentration before release. These streams will be sampled by drawing liquid from the neutralization reactors, the waste retention tanks, or the spent decontamination solu- tion tanks. The samples will be tested to ascertain that agent concentrations are below the Army's established 50 ppm release standard for blister agents. Procedures similar to those used to characterize wastes from the EDS or the legacy Munitions Management Device mobile systems will be used (U.S. Army, 2001a). HD in neutralent will be detected by coupled gas chromatography and mass spectrometry. The detection limit of this method is below 5 ppm, which pro- vides a substantial margin for certifying that the HD concen- tration is 50 ppm or less. (This method is described in the Utah Resource Conservation and Recovery Act (RCRA) permit application for the MMD [DEQ, 19991.) The gas chromatography/mass spectrometry method for HD in hy- drolysate is a routine operation, and the committee antici- pates that the Army can reliably and in a timely manner make these measurements. The committee did not receive or re- view information on whether all of the measurements could be made; such a review is beyond the scope of the charge to the committee. The contents of liquid waste storage tanks will be sampled to ascertain compliance with RCRA standards for hazardous wastes prior to shipment off-site. Analysis of residual agent was covered in some detail in at least three earlier NRC re- ports. The report Integrated Design of Alternative Technolo- gies for Bulk-Only Chemical Agent Disposal Facilities (NRC, 2000a) discusses possible reasons for the presence of residual agent in hydrolysate even though adequate residence time is provided. The report Occupational Health and Work- place Monitoring at Chemical Agent Disposal Facilities (NRC, 2001 c) shows in depth the mustard hydrolysis path- ways and discusses the high toxicity of some of the mustard degradation products. Another report, Evaluation of Alter- native Technologies for Disposal of Liquid Wastes from the Explosive Destruction System (NRC, 2001a), discusses the possible presence of residual agent in suspended solids in EDS hydrolysates and in the cracks and crevices of metal

OCR for page 34
WORKER PROTECTION AND POTENTIAL FOR OFFSITE RELEASE parts. It concludes that any nonincineration technology used to treat the hydrolysate must be robust and able to deal with these issues. The various categories of liquid waste will be analyzed for toxicity characteristic metals and organics us- ing the toxicity characteristic leaching procedure.l2 Signifi- cant concentrations of some metallic corrosion products can be expected in the GTRs because the chemical agents have been in contact with the steel casings for nearly 60 years. It is likely that similar release standards will be applied to the neutralents and rinses generated from arsenical agents con- tained in the GTRs. PD, which has vesicant properties in addi- tion to being a vomiting agent, will probably be held to a 50 ppm release standard like HD and MN-3. Procedures to analyze for arsenicals are still being developed. The aqueous streams generated from the arsenical agents will contain toxic arsenic salts. Arsenic and other toxicity characteristic metals (as well as toxicity characteristic organics) will be analyzed under RCRA protocols before the solutions are released to a TSDF for treatment and disposal. No specific release standard for total arsenic concentration has been set.l3 Solicis Sampling and Analysis Decontamination of metal scrap to the 3X level will be established by analysis of the headspace vapor in the metal decontamination units (MDUs) in accordance with require- ments outlined in Army regulations and Department of the Army document 385-61 (U.S. Army, 2002a). In this proce- dure, the MDU is sealed and the headspace vapor is ana- lyzed using MINICAMS systems adapted to the particular agent being handled in each munitions campaign. Typically, the MDU is held at 70F for 4 hours before headspace sam- pling. The 3X decontamination level requires a headspace vapor concentration below 0.003 mg/m3 for sulfur mustard agent. Other categories of solid wastes are sampled and ana- lyzed by similar procedures. Typically, the wastes are packed in a drum, and the sealed drum is allowed to stand at 70F i2The toxicity characteristic leaching procedure is discussed at greater length in Chapter 5. Conformation obtained at a meeting of the committee, National Research Council staff, Army personnel, and Stone & Webster staff, Boston, Mass., May 21-22, 2003. 39 for 4 hours before headspace sampling. Again, the require- ment for a 3X decontamination level in campaigns dealing with HD is a vapor concentration of less than 0.003 mg/m3. A question has been raised about the suitability of this pro- cedure for release of packages containing spent carbon or other filter material. A previous NRC report (NRC, 2001 c) noted that activated carbon "has a high adsorptive capacity and could therefore give a very low agent vapor pressure from headspace sampling even if a substantial loading of agent remained in the carbon. If the temperature of the car- bon were raised, this agent could be released, posing a dan- ger to anyone not properly prepared or equipped." Nonvolatile Agents in Sorbent Materials A potentially troublesome problem is the handling of solid materials containing chemical agents having little vapor pressure. One example might be the handling of wipes and dunnage from the processing of GTRs containing diphenyl- chloroarsine (DA). The soiled materials may contain signifi- cant quantities of DA, which would not be detectable by the standard vapor test used to certify 3X level decontamina- tion. Given proper handling and packaging, these materials might not be a major hazard to current workers but could be of concern during subsequent treatment and disposal, par- ticularly in the event that the Army sends these materials to off-site TSDFs, as planned. Looking ahead to closure of PBNSF, similar concerns may apply to concrete or soil on which DA has been spilled. Recent work in Japan has shown that DA sorbed on celluloid or pumiced has persisted for more than 60 years if not exposed to hydrolytic conditions (Science Council of Japan, 2002~. Even after hydrolysis, ar- senic-containing residues remained. Research done by the Japanese Chemicals Evaluation and Research Institute in- cludes development of methods for analysis of DA and other arsenicals in the presence of solids (Science Council of Japan, 2002~. i4Among the Japanese munitions abandoned in Manchuria after World War II were cylinders of Agent Red containing mixtures of DA and DC sorbed on celluloid and pumice as well as projectiles containing neat Agent Red (a DA/DC mixture).