Integrated Design of Alternative Technologies for Bulk-Only Chemical Agent Disposal Facilities (2000)

Monitoring for the presence of agent is a critical step in ensuring worker and public safety, which is the number one priority of the PMCD. Monitoring for agent in the vapor phase is common to baseline incineration system facilities and alternative technology bulk-site facilities. The committee anticipates that the Army's considerable experience in monitoring techniques for both mustard and VX will be directly transferable to the ABCDF and NECDF. The committee has already recommended several potential improvements for vapor-phase agent monitoring that are applicable to the ABCDF and the NECDF (NRC, 1994a, 1996b).

Insofar as the committee can ascertain, the ADPs for the ABCDF and NECDF include adequate agent monitoring of the air in areas where elevated agent concentrations are likely, such as agent storage and neutralization process areas. However, provisions for monitoring for agent in areas downstream of the toxic cubicles may not be adequate. These areas include hydrolysate storage and downstream processing areas (VOC treatment and biochemical treatment at the ABCDF; air stripping and SCWO at the NECDF). The committee considers it prudent for the Army to monitor ambient air and gaseous exhausts to the atmosphere for agent in processing areas downstream of the toxic cubicles where process upset conditions (e.g., operator error) could potentially result in agent contamination.

Recommendation 5-1. Plant ventilation air at each facility should be monitored to verify the absence of agent in processing areas downstream of the toxic cubicles. Gaseous exhausts to the atmosphere should also be monitored to verify the absence of agent. Agent monitoring at each facility should incorporate the lessons learned from agent monitoring at the other disposal facilities.

Monitoring for the presence of agent in process streams is necessary to verify agent destruction after neutralization and to verify the absence of agent prior to final release or disposal. Monitoring process streams and nonprocess waste streams for the presence of agent is also required to ensure worker safety and compliance with environmental regulations. The following main processing streams will require monitoring:

• vapor streams from scrubbing operations or activated carbon treatment (see discussion above on vapor-phase monitoring)

• hydrolysate from neutralization of agent

• ton container clean-out solutions

• aqueous effluents from the secondary treatment processes (biological treatment for HD hydrolysate; SCWO for VX hydrolysate) prior to discharge or disposal (e.g., after crystallization and separation of salts from SCWO effluent)

• solid salts or sludges separated from aqueous streams

• ton containers and metal parts following clean-out operations

Analysis for agent contamination of solid phases will also be required for nonprocess wastes, including spent activated carbon, process sludges, dunnage, and personal protective equipment.

Monitoring for residual agent in solid materials is required for both baseline and bulk-site disposal facilities. Establishing that ton containers, metal parts, and other solid materials (e.g., personal protective equipment) have been decontaminated to a 3X level requires confining the material in an enclosed space and analyzing the vapor for agent concentrations that do not exceed the eight-hour time-weighted average limit for worker exposure. A 3X decontamination level must be demonstrated prior to the shipment of materials for off-site treatment or disposal. In many cases, establishing a 3X decontamination level is an intermediate step prior to

thermal treatment to achieve a 5X condition (either by onsite treatment at baseline facilities or by off-site treatment for 3X decontaminated ton containers at bulk-site facilities).1 Thermal treatment to a 5X standard will not be an option for on-site treatment of solid wastes at the ABCDF and NECDF.

Measuring potential agent contamination in certain solid process streams (e.g., sludges and dried salts) and in spent activated carbon presents challenges unique to the alternative technology sites. For baseline sites, these materials can be thermally treated in one of the incinerators (e.g., the metal parts furnace) to achieve the 5X decontamination level prior to disposal of the ashes. Because this is not an option for the alternative technology bulk sites, measurement techniques must be developed and demonstrated to certify that these materials have no detectable agent prior to off-site shipment for disposal.

Vapor-phase monitoring to meet the 3X specification will most likely not be sufficient to verify agent destruction in the solid materials from the bulk sites because these materials could contain either strongly adsorbed agent or occluded agent that could be released in the future. Because of the unique analytical interferences resulting from the composition of particular waste streams, the measurement methods will have to be specific to each waste stream, and each method will have to be validated for the specific waste matrix. Criteria for determining the detection limit for each method should be based on the hazard and risk evaluations for that waste stream.

Measuring agent at low levels in aqueous phases, such as hydrolysate, decontamination fluids, and final waste streams, is particularly challenging because of the presence of organic and inorganic constituents that interfere with measurement techniques. Measuring for residual agent in the hydrolysate from the neutralization step may present the greatest challenge. Nevertheless, measurement is necessary for the following purposes:

• to verify the efficacy of the primary agent destruction step (e.g., 99.9999 percent destruction)

• to delist the hydrolysate so that it can be managed as a nonhazardous waste in subsequent processing steps (biological treatment for HD hydrolysate, SCWO for VX hydrolysate)

• to clear any hydrolysate to be shipped off site for process development and testing purposes

From operational experience obtained during the development of the neutralization process and pilot operations conducted at the Chemical Agent Munitions Disposal System (near Tooele, Utah), the Army is aware of the potential presence of residual agent in poorly mixed or occluded zones of the hydrolysis reactor. This agent may not be detected during analysis of a batch of hydrolysate following neutralization but may be detected after the hydrolysate has been transferred to downstream storage vessels. Operator error could also result in the transfer of hydrolysate that still contains residual agent to downstream processing. Reactor designs for the ABCDF and NECDF have been carefully developed to avoid poorly mixed or occluded reaction zones, and interlocks are provided to prevent operator errors. However, the committee considers it prudent for the Army to analyze the contents of hydrolysate storage tanks following the transfer of each batch of hydrolysate to the tanks.

Recommendation 5-2. After the transfer of each batch of hydrolysate from the reactors to hydrolysate storage tanks, the hydrolysate should be sampled from the hydrolysate storage tank and analyzed before subsequent processing to verify that it is free of detectable agent.

The hydrolysate testing requirements are distinctly different from the requirements for the verification of agent destruction for final release of process effluents because the hydrolysate is an “in-process ” stream that requires further treatment (in accordance with both the CWC and environmental regulations). The Army's initial specifications for testing hydrolysate, which were based on standards for battlefield drinking water, were unrealistic for three reasons: (1) hydrolysate will not be discharged without further treatment; (2) human consumption is highly unlikely; and (3) the complexity of the analytical techniques makes measurements unusually time consuming and unreliable. More appropriate criteria for detection limits would be based on the required destruction efficiency and a risk-based evaluation of subsequent hydrolysate management. Analysis of the hydrolysate should also include measurements of potentially active species, such as sulfonium ions from HD hydrolysis and EA2192 from VX hydrolysis.

Criteria for detection limits to verify the absence of agent in the aqueous waste streams after final on-site treatment should be based on a risk evaluation that assumes either discharge to public waters or another final disposal method. Risk evaluations must address both worker and public risks. Evaluations of public risks from the accidental release of hydrolysate at less than 330 ppb (99.9999 percent destruction level) of agent have shown that public risk is minimal (Garrick and Gekler, 1999). However, an evaluation of worker risk and potential mitigation measures has not been completed.

Recommendation 5-3. The criteria for releasing VX hydrolysate to subsequent processing steps at the Newport

facility should be based on the 99.9999 percent agent destruction level required by the state of Indiana and an evaluation of worker and public risks associated with releases of agent in hydrolysate equal to or less than this destruction level (i.e., 330 ppb). Assessments of worker risk should include a thorough evaluation of risk mitigation measures and additional design and operating requirements necessary to satisfy this release criterion.

Measuring very low levels of VX in hydrolysate has been extremely difficult because of the presence of reaction products from the neutralization of VX, organic impurities, stabilizers present in the stockpile grade VX that remain after neutralization, and the high concentrations of excess sodium hydroxide necessary for neutralization. The hydrolysate from the neutralization process thus remains a highly reactive matrix. Representative sampling of the hydrolysate is also difficult because it consists of two phases (aqueous and organic) that must be maintained in a thoroughly mixed condition during sampling, and the organic phase can wet the sample container walls. Thus, the Army has had difficulty reliably meeting the initial target of 20 ppb (the battlefield drinking water standard). The required VX destruction efficiency for the neutralization process (> 99.9999 percent) necessitates a detection limit of less than 330 ppb. By contrast, a detection limit of less than 20 ppb has been demonstrated for VX in effluents from SCWO treatment of hydrolysate. Measurement of VX in SCWO effluents is much simpler because concentrations of organic constituents are low and the pH is near neutral.

The following approaches have been developed for measuring VX in hydrolysate (NIVA Consultants et al., 1999):

• nuclear magnetic resonance (NMR) carried out directly on hydrolysate

• chloroform extraction followed by gas chromatographyion trap mass spectroscopy (GC-ITMS)

• methylene chloride extraction followed by high-performance liquid chromatography-mass spectroscopymass spectroscopy (HPLC-MS-MS)

• solid-phase extraction followed by gas chromatographymass selective detection (GC-MSD)

• hexane/acid liquid-liquid extraction followed by GC-ITMS-MS

Multiple variations of these approaches have been investigated. NMR requires approximately 16 hours to achieve a detection limit of 500 –1,000 ppb. Because of the long analytical time and high detection limit, this method is considered unsuitable. The chloroform extraction method has been found to achieve a detection limit between 80 and 200 ppb and requires approximately four hours to complete. Methylene chloride extraction followed by HPLC-MS-MS has been found to achieve a detection limit of approximately 100 ppb and requires less than one hour to complete. The solid-phase extraction technique has been found to achieve a detection limit of between 4 and 35 ppb and requires several hours. However, analytical recoveries from this method have been erratic, ranging from less than 50 percent to greater than 350 percent. Evaluation of this approach suggests the possible formation of VX under the analysis conditions from the original agent impurities.

Recent results obtained during analysis of the VX hydrolysate produced at the Chemical Agent Munitions Disposal System for the planned SCWO EST indicate that the hexane extraction method can repeatably achieve a detection limit of less than 20 ppb (NIVA Consultants et al., 1999). However, this method not only requires approximately six hours to complete under optimum conditions, it also appears to be sensitive to both the technician performing the analysis and the laboratory (e.g., instrumentation model) where the analysis is carried out. These limitations suggest that further specification and validation will be necessary to achieve reliable results for routine use.

Recommendation 5-4. The Army should develop and demonstrate methods of chemical analysis to confirm the destruction of VX in the hydrolysate at the Newport facility. These methods should include procedural specifications and provisions for training so that confirmation at the required detection limits can be confirmed by testing by different analysts at multiple laboratories.

Verification of the destruction of mustard in HD hydrolysates has not presented the same technical challenges as VX, but it does require the use of NMR analysis, which takes four to six hours to measure both mustard and sulfonium ions (U.S. Army, 1998b). Verification of agent destruction also constitutes a critical path item in the operational cycle of each facility. Currently, analysis of each batch of hydrolysate takes six hours, provided that reliable analytical results are obtained from the first analysis. Thus, reducing the time required to verify agent destruction in process streams would significantly improve the overall processing efficiency and schedule.

Recommendation 5-5. The Army should develop and demonstrate innovative analytical techniques that require significantly less time than present techniques to verify agent destruction for both VX and HD hydrolysates at the required detection limits.