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 33
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot 3 AEA SILVER II™ Technology Package DESCRIPTION OF THE PROCESS The AEA Technology (AEA) SILVER II™ technology is based on the highly oxidizing nature of Ag2+ ions, which are generated by passing an electric current through a solution of silver nitrate in nitric acid in an electrochemical cell similar to those used in commercial electrochemical processes. The electrochemical reactions used in the SILVER II™ technology belong to a class of chemical processes collectively known as mediated electrochemical oxidation (MEO). MEO processes have been offered as an alternative to conventional incineration for destroying hazardous wastes. Because they are also relatively new in development and application, experience with these processes is limited (Chiba et al., 1995). Detailed descriptions of AEA’s original total system solution and its unit operations were provided in the original ACW I Committee’s report (NRC, 1999) and the ACW II Committee’s report on Demo II testing (NRC, 2001b). This report provides an update reflecting major changes in the process identified by the technology provider team (consisting of AEA Technology and CH2MHILL) based on prior testing results and on EDS II tests and studies. Figure 3-1 shows a block diagram of the overall process for the current AEA SILVER II™ total system solution. The first step in the system is a modified reverse-assembly process in which the energetics, agents, and metal parts are separated. Energetic materials are removed, reduced in size, and prepared as a slurry in water before further treatment. The agent and energetics are then destroyed in separate electrochemical processing units. Metal parts and fuzes are thermally decontaminated to a 5X condition in a metals parts treater (MPT). Dunnage and DPE suit material are decontaminated to a 5X condition by a dunnage treatment system (DTS) that is similar but not identical to the continuous steam treater (CST) in the Parsons/Honeywell WHEAT1 technology package proposed for disposing of the mustard agent munitions at Pueblo Chemical Depot (NRC, 2001a). Silver, water, nitric acid, and NOx are recovered, converted to reagents, and reused in the process. Other solid, liquid, and gaseous effluents from various process units are collected, separated, treated, and tested to prepare them for safe discharge from the plant. Table 3-1 summarizes key plant performance requirements used in developing the SILVER II™ technology package. The following sections recap briefly the sequence of process steps beginning with munitions disassembly and ending with treatment of various waste streams for final disposal. They are derived from the complete description of the AEA technology proposed in the EDP 1 Water hydrolysis of explosives and agent technology.
OCR for page 34
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot FIGURE 3-1 AEA SILVER II™ demilitarization process. SOURCE: AEA (2001a).
OCR for page 35
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot
OCR for page 36
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot TABLE 3-1 Key Plant Performance Requirements for SILVER II™ Technology at Blue Grass Army Depot Parameter Quantity Total processing period 2 years (730 days) Peak energetics throughput (lb/hr) 287.8 of M28 propellant (equivalent to 12.8 M55 rockets/hr) Peak agent throughput (lb/hr) 166.7 of mustard agent (equivalent to 14.3 155-mm projectiles/hr) Destruction efficiency (%) Agent feed 99.9999 Energetics feed 99.999 Electrolyte composition Anolyte nitric acid 8 M Anolyte silver 0.5 M Anolyte steady-state TOC 3,000 mg/L in agent and energetics, main or primary circuit, 3,000 to <1 ppm in agent polishing or secondary circuit, 3,000 to <10 ppm in energetics polishing or secondary circuit Catholyte nitric acid 6 M Catholyte silver 1 M SOURCE: Adapted from AEA (2001a). submitted to PMACWA for a full-scale pilot plant at Blue Grass Army Depot (AEA, 2001a), and they highlight changes from earlier reports of the ACW I and II Committees. Disassembly of Munitions Munitions Unpacking Unpacking of the munitions is performed in the unpack area (UPA), which is the same as that in the Army’s baseline incineration system design (NRC, 1999). Projectile Processing The AEA technology package uses the baseline system projectile mortar demilitarization (PMD) machine with modification for burster and agent washout (AEA, 2001a). After removal of the nose closure, fuze adaptor, and fuze cup, the burster tube is removed from the agent cavity and transferred to the burster washout machine (BWM), which uses high-pressure water to remove the energetic material from the burster tube. Water is added to the resulting slurry as it is drained to achieve an energetic concentration of 20 weight percent. The slurry is then pumped through an inline static macerator and into a buffer storage tank for the SILVER II™ energetics destruction system. The agent is then drained from the projectile and the cavity washed out with a water spray. The amount of drained agent is measured and sent to the buffer storage tank for the SILVER II™ agent destruction system. The burster tube and other projectile metal parts are placed in a bin for transport to the MPT for 5X decontamination. Rocket Processing Line The rocket dismantling machine (RDM) proposed by AEA is based on the baseline system design but has significant modifications, including these: a modified punch-and-drain station having a hollow top punch to allow water or low-pressure steam to wash out gelled or viscous agent a tube-cutting machine to cut the rocket in its shipping tube into four segments after removing agent a water jet washout station to wash energetic materials out of the burster tube a mechanical extractor to remove the propellant grain in one piece for subsequent size reduction in a propellant grinder (AEA, 2001a) Tests to demonstrate the operability of these modifications would be done at the vendor’s plant prior to shipment or during systematization of a constructed facility using the SILVER II™ technology. Each rocket in its shipping and firing tube is manually removed from its storage pallet in the UPA and transferred to a metering device to ensure that it is oriented with the warhead assembly heading first into the explosion containment vestibule (ECV). The metering
OCR for page 37
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot device feeds each rocket into the ECV and from the ECV into the RDM. Two clamps on the RDM engage and grip the rocket tube near the fuze and near the connection between the bottom of the warhead and the top of the rocket motor, as in the baseline configuration. Once gripped in place, the rocket is punched once on top and twice on the bottom, and the agent drains through the two bottom punches to the agent weigh tank. Hot water or low-pressure steam is then passed into the agent cavity through the hollow top punch to wash out additional agent and minimize agent loading in the metal parts treater (MPT). The agent that drains from the cavity through the two bottom punches is pumped to the agent weigh tank and then to the agent buffer storage tanks before it is fed to the agent SILVER II™ processing system. Rinse water or steam/ condensate and residual agent go directly to a buffer storage tank for the SILVER II™ process (AEA, 2001a, 2001b). When processing GB rockets, the agent weigh tank control system provides a signal indicating the quantity of agent drained from the warhead. If less than a certain minimum percentage of the agent has been drained, the GB is assumed to be gelled or crystallized inside the warhead. Then, low-pressure, saturated steam (15 psig and less than 249.5°F) is fed through the punch head to rinse out the residual agent. When processing VX rockets, after draining the agent, low-pressure water is fed through the punch head to rinse residual agent out of the warhead. After the agent is washed out, the rocket is rotated 90 degrees about its longitudinal axis to minimize leakage of liquid remaining in the cavity. The rocket is then advanced to the rocket cutting station, where it is clamped into the rocket rotator head and then rotated about its longitudinal axis. The first of three tube cutters cuts through the firing tube and rocket body and separates the fuze from the rocket, exposing the top of the burster well cavity. The severed fuze and associated firing tube section are dropped through a chute and conveyed to a bin for transport to the MPT. A power lance with high-pressure (~15,000 psig) water flowing from multiple nozzles is used to wash the explosive from the burster tube as the remaining section of the rocket in its firing tube is rotated. The washed-out energetic material is pumped through an inline macerator to form a slurry and collected in a tank and mixed with additional water to reduce the energetics content to 20 weight percent. The slurry is then sent to the energetics buffer storage for the SILVER II™ process. Additional tube cutters then cut the warhead and the fin assembly from the motor casing. After both of these cuts have been made, the motor casing is moved to another station, where the attached antiresonance rod assembly is removed from the propellant grain. The antiresonance rod assembly and igniter are then transferred into a metal parts bin for transport to the MPT, and the motor casing section is moved to another station, where the rolled edge on the casing is expanded. A pusher assembly then forces the propellant grain from the motor casing, and the grain drops into the propellant grinder hopper. The warhead, fin assembly, and motor casing, along with associated firing tube segments, are dumped into a metal bin via the collection chute and conveyed to the MPT for 5X decontamination. Preparation of Energetic Materials for the SILVER II™ Process Slurries of energetic materials from two sources are treated in the SILVER II™ process: the explosive in the rocket and projectile burster tubes (which may be either Composition B or tetrytol) and the ground-up M28 propellant grain in the rocket motors. The propellant grain is transferred through two interlocked blast gates into the grinder, where it is ground up under water to produce a slurry with particles having a maximum dimension of 1/4 inch. The slurry is discharged through a pipeline in a manner that limits the content of the pipe to no greater than 20 weight percent propellant. The energetics slurry then flows into the energetics slurry feed tank (AEA, 2001a). The uniformity of the energetics slurry, which may at times be combined propellant and explosive, is assured by using a high-shear vortex mixer and a conventional stirrer in the energetic slurry feed tank. The high-shear vortex mixer completes the size reduction of the energetics from burster washout and propellant grinding in preparation for SILVER II™ treatment. The stirrer ensures that all the slurry in the tank passes through the mixer head. The feed tank and slurry piping are trace-heated to prevent cold spots and precipitation or plate-out of energetics materials (AEA, 2001a). The energetics slurry is transferred from the energetics slurry feed tank into the anolyte feed tank through a hydrocyclone. The concentration of the energetics in the hydrocyclone underflow is nominally 40 weight percent. This underflow drains into the anolyte tank. The overflow, depleted in energetics, goes to a second hydrocyclone. The underflow from the second
OCR for page 38
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot hydrocyclone, still containing substantial quantities of energetics, returns to the energetics slurry feed tank. The overflow from the second hydrocyclone goes to the propellant grinder, where it makes up part of the water needed to slurry the incoming propellant grains. Preparation of Agent for SILVER II™ Treatment The recovered agent from the projectile and rocket processing lines is pumped to the agent feed system, consisting of a tank and a pump (not shown in Figure 3-1). Dilute nitric acid (HNO3) from the NOx reformer is added to the agent tank depending on the water volumes. (Water additions and other reagent additions achieve an 8 M HNO3 concentration and a 0.5 M AgNO3 concentration for the anolyte feed stream.) An agitator is provided to ensure homogeneity of the contents. The tank is maintained at a slight negative pressure relative to the Category A area2 by venting to the anolyte offgas condenser. A nitrogen blanketing system is provided, with a regulator set at a slight negative pressure. Treatment of Agent and Energetics by the SILVER II™ Process Overall Chemistry The SILVER II™ process is based on the highly oxidizing nature of Ag2+ ions in an aqueous HNO3 solution. Ag2+ ions are among the strongest oxidizing agents known. HNO3 is also a strong oxidizing agent and contributes to the overall destruction process, although the amount of HNO3 that reacts appears to be less than 5 percent of the reaction required for total destruction (AEA, 2001b). In each of the SILVER II™ cells, a pair of electrodes (anode and cathode) is housed in a compartment within the cell. A semipermeable membrane is placed between the electrodes. The membrane maintains electrical continuity between the electrodes and prevents mixing of the anolyte and catholyte solutions. The electrochemical cells operate at 190°F and essentially atmospheric pressure. Direct current is applied to the electrochemical cells at a potential of 2 volts, resulting in an electrochemical reaction that generates Ag2+ ions. Since the Ag2+ ions simply mediate the reaction process, the overall reaction across the anolyte and catholyte is essentially the reaction of the organic feed with nitric acid, forming NOx, water, CO2, and inorganic acids: Organic (anolyte) + HNO3 (catholyte) → CO2 (anolyte) + H2O (catholyte) + inorganic acids (anolyte) + NOx (anolyte and catholyte) Ag2+ ions generated at the anode react with the water and HNO3 of the anolyte solution to form a range of radicals (e.g., OH•, NO3 •) that in turn oxidize the organic material in the anolyte solution completely and irreversibly to carbon dioxide, some nitrogen oxides (NOx from the direct reaction with the acid, which proceeds to a moderate extent), inorganic ions, additional hydrogen ions (H+), and small amounts of carbon monoxide. Ag2+ can also react directly with water in the anode compartment to form oxygen gas. Both Ag+ and cationic impurities in the anolyte can migrate across the membrane to the catholyte compartment. To balance the electrochemical reaction in the anolyte, there is a corresponding cathode reaction that involves reducing nitric acid to nitrous acid, which in turn partially decomposes to NOx gases and water. The water balance is complex and involves two countercurrent fluxes. Water is transferred across the membrane from the anolyte to the catholyte in the form of hydrated protons generated as a product of the anode reaction. Water also flows in the opposite direction from the cathode compartment to the anode compartment owing to the lower acidity (higher water concentration) in the cathode compartment. A more detailed discussion of the SILVER II™ reaction chemistry is given in the ACW I Committee report and is not discussed further in this report (NRC, 1999). A summary block flow diagram depicting the SILVER II™ systems applicable to the processing of both agent and energetics is given in Figure 3-2. The flow circuit for AgCl separation (shown by dotted lines) is used only when processing H or HD, which contain chlorine. As shown in Figure 3-2, the SILVER II™ systems incorporate internal recycle streams: (1) from the NOx reformer to replenish nitric acid lost from the anolyte and catholyte solutions and (2) from the catholyte circuit to return silver ions, unreacted organic 2 An area where agent contamination is to be expected. Personnel performing maintenance in Category A areas must be in DPE suits. Category B areas require the highest level of respiratory protection but a lesser degree of skin protection. Category C areas are those where the concentration of airborne substances are known and the criteria for using air-purifying respirators are met.
OCR for page 39
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot FIGURE 3-2 SILVER II™ process system for agent or energetic destruction.
OCR for page 40
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot material, and mineral acid ions that have crossed the membrane from the anolyte to the catholyte circuit. There is also an anolyte purge stream to the impurities removal system (IRS).3 The anolyte purge stream is processed by the IRS in a batchwise sequence through SILVER II™ electrochemical cells in a polishing circuit, where remaining organics are destroyed. After polishing, the purge stream content undergoes AgCl precipitation and separation using hydrocyclones and a centrifuge. The liquid from AgCl precipitation is then sent to an HNO3 evaporation operation for recovery of the nitric acid and to concentrate the impurities in the evaporator bottoms or brine stream. Batchwise processing of the anolyte purge stream is accomplished by collecting the stream in one of three tanks and then processing the contents of that tank through the preceding three steps. The contents collected in each of the other two tanks are similarly processed in sequence. Process Outputs Solids. Treatment of mustard agents by the SILVER II™ process results in the precipitation of silver chloride, which, if not removed, could cause cell plugging. This precipitate is removed from the main circuit in a slurry underflow from hydrocyclones in the anolyte feed stream circuit. Subsequently, this slurry is dewatered in a centrifuge to approximately 50 weight percent AgCl, with the remainder of the slurry consisting of water, nitric acid, silver nitrate, and trace amounts of other materials in the anolyte loop. The SILVER II™ polishing circuits for both agent and energetics systems continuously produce a larger stream of AgCl than what is produced in the main circuit when H or HD feeds are being processed. It should be noted that essentially all silver in the main circuits from processing any agent or energetics is eventually removed in the AgCl precipitation step following the polishing of agent and energetic anolytes. In this operation, hydrochloric acid (35 weight percent) is added to the feed tank containing the polished anolyte to form a slurry that contains AgCl precipitate. The slurry is passed through a hydrocyclone and the underflow is sent to a centrifuge for dewatering to approximately 50 weight percent AgCl. The hydrocyclone overflow of the slurry consists of water, HNO3, trace amounts of other materials in the anolyte loop and about 2 percent unreacted silver nitrate. The AgCl sludge is then combined with AgCl sludge from the main circuit when processing H or HD feeds. This resulting sludge stream is then washed with 18 percent NaOH to neutralize the remaining acid. The neutralized sludge is then thermally treated to a 5X decontamination level in an enclosed auger heated by hot oil. The 5X silver chloride is sent to an off-site silver reclaiming contractor, which then returns the silver to the plant for preparing silver nitrate makeup. The liquid overflow from the hydrocyclones and the AgCl centrifuge in the polishing circuit is sent to an evaporator for recovery of HNO3, which is condensed from the overhead vapor stream and pumped to the NOxreformer. AEA states that the evaporator is expected to recover approximately 70 percent of the HNO3 in the evaporator feed for subsequent recycling to the reformer (AEA, 2001a). The recovery percentage will vary depending on the amount of metals present as impurities in the feed and the final concentrations of metal and sulfuric, phosphoric, and hydrofluoric acids. For agent feed streams, the concentration of acids other than nitric places a limit on the evaporator operation, because carryover of these other acids in the evaporator vapor may occur. Energetic feed streams have no S, P, or F, and the acidic recovery is limited only by the solubility of iron and aluminum nitrates in the strongly acid evaporator bottoms. The bottoms containing mineral acids and metals are then treated, tested for agent, and prepared for off-site disposal. Liquids. Liquid effluents are expected to be limited to excess concentrated HNO3 of undefined purity that will be produced when processing energetics. Gases. The anolyte offgas from the primary and polishing systems contains O2, CO2, and NOx and is saturated with water vapor and HNO3 at 176°F. Small, variable amounts of CO are also formed from the oxidation, the amount depending on the organic feedstock. Most volatile organics from decomposition of feed material are expected to be condensed with water and acid vapor in the main and polishing circuit offgas condensers. These anolyte offgas condensers from the main and polishing circuits are cooled with chilled water to operate at 91°F and 40°F, respectively. The condensates are returned to the anolyte in the main system. The offgas from the catholyte feed tank contains O2 3 The IRS consists of the secondary polishing circuits of the SILVER II™ process and the associated components, including hydrocyclones, AgCl separation equipment, and the HNO3 evaporator.
OCR for page 41
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot (sparged into the catholyte feed tank for use in the NOx reformer), plus NO and NO2, and is saturated with water vapor and HNO3 at 176°F. This offgas first passes through a condenser to recover water and acid, which is returned to the catholyte feed tank. The anolyte and catholyte offgas streams are combined downstream of their respective condensers and flow to the NOx reformer to produce concentrated nitric acid. Offgas from the reformer system is scrubbed with NaOH/NaOCl solution to remove unreacted NOx. The scrubbed NOx reformer offgas, which contains CO2, volatile organic compounds (VOCs), and small amounts of CO, flows to the gaseous effluent treatment systems (as shown in Figure 3-1)—of the MPT in the case of agent processing and of the dunnage treatment system (DTS) in the case of energetics processing. There, thermal treatment and CATOX units oxidize the CO and organics. The gaseous effluent treatment systems are described in a later section. Process Equipment The electrochemical cell used for the SILVER II™ process is based on a standard industrial cell design. Detailed discussion of the design, installation, and operation of the cells is provided in the EDP report, Volume 2 (AEA, 2001a). For a full-scale facility at Blue Grass, the main processing circuits would consist of six operating cell stacks of 48 cells each for agent and six operating stacks of 42 cells each for the energetics. In addition, the SILVER II™ processes for both agent and energetics each have polishing systems consisting of three cell stacks, each configured identically to the main circuit stacks. The total power requirements for the primary (main) circuit and secondary (polishing) cell stacks for both the agent and energetics systems in the AEA SILVER II™ process are given in Table 3-2. In addition to the operating stacks, both the agent and energetics systems have one installed spare cell stack. Five more cell stacks are stored as spares. All operating stacks are connected in series to a single direct current power supply. The cells within each stack are connected in parallel. The installed spare is electrically shorted and isolated from the anolyte and catholyte circuits by valves. Demo II testing established that the electrodes and membranes can operate at a current of up to 2 kiloamperes per electrode pair (AEA, 2000). The anodes are platinized titanium (titanium plated with 5 microns of platinum); the cathodes are also titanium plated with 2.5 microns of platinum. All gaskets are Viton GF (peroxide grade), and the cell membranes are DuPont Nafion 324. Flow through the cells is in parallel using manifolds with 1/4-inch fluid-flow inlet port orifices to provide equal flow to all cells. As noted in Table 3-2, the secondary polishing system cells require electrical power at 50 percent of the level of the main circuit cells because the electrochemical organic destruction efficiency (i.e., the fraction of the electric current used to oxidize organics) decreases sharply as the organic concentration is reduced from an initial concentration of ~3,000 ppm to 1 to 10 ppm. The average destruction efficiency of the polishing cells is expected to be 2 to 10 percent, with the main competing reaction being the oxidation of water. Consequently, the offgases from the polishing system cells TABLE 3-2 Summary of SILVER II™ Plant Electrochemical Cells Cell Stack Quantities Description Power (kW) Operating Pressure (psig) Material Electrode Pairs/Cell Stack In Service Installed Standby In Storage Agent Primary electrochemical cell 1,120 ~0 PVDF/Ti/Pt 48 6 1 3 Secondary electrochemical cell 560 ~0 PVDF/Ti/Pt 48 3 0 2 Energetics Primary electrochemical cell 991 ~0 PVDF/Ti/Pt 42 6 1 3 Secondary electrochemical cell 495.5 ~0 PVDF/Ti/Pt 42 3 0 2 SOURCE: AEA (2001a), Volume 2.
OCR for page 42
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot contain a high percentage of oxygen, which flows to the NOx reformer with the other offgases from the main circuit. This oxygen, along with the oxygen that was sparged into the catholyte feed tank, is used to generate nitric acid in the reformer. AEA is primarily relying on lining all vessels and piping in the agent SILVER II™ system with polytetrafluoroethylene (PTFE, or Teflon) or PFA for corrosion resistance. To prevent HF from returning to the SILVER II™ system via the nitric acid stream from the evaporator, AEA is evaluating methods of keeping the HF in the liquid phase of the nitric acid evaporator bottoms by using a complexing chemical. The method of trapping the fluoride in the evaporator bottoms had not been selected at the time this report was prepared. Therefore, additional unit operations yet to be defined may be required to deal with this issue. Processing and Treatment of Metal Parts, Dunnage, and Other Solid Waste Metal Parts Processing Metal parts, fuzes, and fiberglass segments of shipping and firing tubes from the munitions disassembly lines, and drums containing particulate matter from the candle filters and cyclones of the gaseous effluent treatment systems are conveyed in bins to one of two MPTs for decontamination to the 5X level. Normally only one MPT is required; however, AEA has identified disposal campaigns where a second MPT would also be used (AEA, 2001a). In the MPT, metal and fiberglass parts are heated to over 1000°F for at least 15 minutes. The MPT is a modified industrial oven with the following four zones: zone 1 for purging air with nitrogen and providing an inert gas barrier for the following zones zone 2 for warming the materials contained in bins to 300°F to volatilize organic materials from the fiberglass and other materials being treated zone 3 for heating the parts above 1,000°F for more than 15 minutes to achieve 5X decontamination zone 4 for cooling and verifying that the parts are acceptable for release Zones 1 and 4 use inert gas atmospheres and are separately exhausted directly to the carbon filters of the MDB HVAC system. Zones 2 and 3 operate with steam atmospheres and are exhausted through a dedicated gaseous-effluent treatment system to remove entrained water and control the level of trace organics released. (The system is described in a later section.) Steam used in zones 2 and 3 is recirculated through electrically heated exchangers or heaters. The cycle time for each batch is approximately 1 hour to ensure that all of the contents of the metal parts bins reach 1,000°F for 15 minutes and that all organic material in the fiberglass shipping tube segments is decomposed. This thermal treatment also is expected to decompose (possibly through initiation) fuzes and igniters contained in the metal parts bins. Temperature markers4 are placed in each batch to verify that the required time at temperature has been achieved for 5X decontamination. Treated parts are moved to zone 4, where they are cooled and the atmosphere is sampled to verify that the parts are free of agent to the limits established for release off-site. If agent is detected above release limits, the bin is recycled through the MPT. After cooling, the metal parts are sent to a metal parts crusher and then off-site for disposal or recycling. Dunnage Processing Wood dunnage, DPE suit material, and spent carbon are treated to the 5X decontamination level using a DTS (AEA, 2001a). The major components of the DTS include the carbon carrier medium silo, the DPE suit shredder, the wood shredder, the feed hopper, and two redundant, full-capacity dunnage treaters. The DTS design is based on a design by Parsons that uses a continuous steam treater (CST) (Parsons, 2001). Wood pallets and DPE suit material are processed in their respective shredders and the shredded material transferred to one of two feed hoppers along with carbon carrier media. The feed hoppers discharge into the dunnage treater through a rotary air lock. Each feed hopper will feed either of the two dunnage treaters. Normally, one dunnage treater is on standby. Each dunnage treater is a 4-ft-diameter horizontal vessel made of Hastelloy C276 material, within which 4 Devices inserted in material being treated that undergo observable physical change after experiencing the required time at temperature.
OCR for page 43
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot is mounted a screw auger to transport wood dunnage, spent carbon, or a mixture of carbon carrier media and either wood chips or shredded DPE suits from one end to the other. The treater is operated under a slight negative pressure, and high-temperature, low-pressure steam sweeps offgases generated by thermal decomposition into the offgas treatment system and contributes to thermal decomposition of the dunnage. The residence time of the dunnage material inside the treater is sufficient to ensure that all of the material contained in the enclosed auger is heated to the conditions required for 5X decontamination. In EDS testing of the CST, residence times of 1 hour were found to be adequate (Parsons, 2001). Actual residence times in the dunnage treater would be determined by using temperature markers during start-up operations. At the discharge end of the dunnage treater, the treated material is discharged through double rotary valves to prevent air backflow into the treater. The treated dunnage drops into a discharge hopper, where it is quenched with a water spray to extinguish any residual embers. Steam generated from the quench water is returned to the dunnage treater as part of the sweep steam. The hopper atmosphere is then sampled to ensure agent levels are below the limits set for discharge. The treated dunnage char is conveyed outside the MDB to a storage silo and transported to off-site disposal. Each dunnage treater is electrically heated using multiple, equally spaced high-performance mineral-in-sulated band heaters.5 The entire dunnage treater, including heaters, is insulated by ceramic fiber insulation blankets contained within a steel shroud. Since the dunnage treater is operated under only slightly negative pressure, all openings are equipped with seals to prevent fugitive fumes from escaping. The slight negative pressure ensures proper flow of steam and the gases generated in the treater. The dunnage treater operates at temperatures from 1,000 to 1,100°F. Using start-up data from thermal markers, the rotational speed of the screw conveyor or auger is adjusted to provide the residence time required for 5X decontamination. As the screw rotates, the solid material is conveyed through the length of the treater. Offgases and volatile materials are primarily VOCs, CO, CO2, and H2O generated by thermal decomposition and, to a lesser extent, steam reforming reactions with the dunnage material (Parsons, 2001). When processing DPE suit material, HCl will also be present in the sweep gas stream. The process gases are treated in a gaseous effluent treatment system (described in the following section). Processing and Treatment of Gaseous Effluent Streams Two emissions control systems are used prior to discharging the gaseous effluents to the carbon filters in the MDB HVAC system. One emissions control system is provided for the gaseous effluents from the agent SILVER II™ system and the MPTs. The other emissions control system handles gaseous effluent from the energetics SILVER II™ system and the DTS (see the process flow diagram in Figure 3-1). Each system is redundant in that a system is associated with each of the two MPTs and each of the two DTSs in the full-scale design. The two emissions control systems are very similar. In order of occurrence, the flow paths for both systems contain the following elements. Where an element is unique to one of the systems, it is so noted. Cyclone. The first element in the system for DTS gaseous effluent is the cyclone provided to separate large particulate matter that could rapidly plug the downstream candle filters. This element is not required in the system for MPT gaseous effluent. Separated particulate matter is discharged into a steel container through a rotary air lock. The particulates are processed in the MPT to ensure 5X decontamination. Candle filters. These consist of multiple high-temperature ceramic or sintered-metal candles fitted within a plenum and a nitrogen supply for online periodic back flushing. Captured particulate matter is discharged into a steel container through a rotary air lock. The particulates are processed in the MPT to ensure 5X decontamination. 5 Both the Departments of the Army and the Navy have used mineral-insulated band heaters for various superheating applications. For example, the Department of the Navy uses similar equipment to test chemical stability of components within artillery shells (AEA, 2001a). In addition to high watt densities, these heaters provide even temperature profiles and fairly precise temperature control.
OCR for page 60
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot committee is concerned that the enclosed auger design may be more prone to blockage or plugging. Catalytic Oxidation Units Both the MPT and DTS use CATOX units to destroy VOCs in the gaseous effluent streams. The CATOX units use a Pt/Pd oxidation catalyst. AEA uses a scrubber and filter upstream of the CATOX units to remove phosphorus, fluorine, and chlorine compounds that could poison the catalyst. With regard to dioxin/furan formation, AEA states that the use of carbon and HEPA filters downstream of the CATOX units should adequately control emissions of these materials. As noted earlier, testing of the CATOX units of the Parsons/Honeywell WHEAT process has revealed the need for changes in the design and operating conditions of both the CST and the CATOX units in order to reduce the formation of dioxins and furans when processing DPE suit material (Parsons, 2001). Thus, while the CATOX units appear to be capable of operating reliably, data from existing EDS testing of the Parsons CST and CATOX units indicate that there is still a significant risk of added delays and cost to achieve acceptable emissions performance (Parsons, 2001). According to the AEA’s EDP, carbon dioxide emissions from a full-scale facility at Blue Grass would be significant, 90 ton/yr, even with the anticipated 98-99 percent efficiency of the CATOX units (AEA, 2001a). The carbon dioxide arises primarily from the treatment of carbon and wood and DPE suits in the dunnage treatment CST and of fiberglass in the MPT. ASSESSMENT OF INTEGRATION ISSUES Component Integration The committee notes that the SILVER II™ technology package is a combination of many continuous and batch-processing steps with buffering capabilities between most of the processing steps. The throughput and availability of each process step, in combination with equipment redundancy and sufficient buffer storage capacity between process steps, must result in the specified destruction rate. Integrating the individual processing steps will require effective process monitoring and control to ensure that appropriate materials are fed at appropriate rates at each step and that all material discharged from the plant meets safety and environmental specifications. In addition, attainment of the required process availability depends on the durability of all materials of construction and on the effectiveness of the plant operating and maintenance force. Process Operability Destruction of the Blue Grass stockpile within the time specified by the CWC treaty requires that the overall process achieve the required throughput levels and process availability (i.e., the fraction of time that the plant can operate). The SILVER II™ process has been designed to enable disposal of the entire stockpile at Blue Grass in 475 days. This includes 112 days of zero production for mandated periods (holidays, maintenance, and external causes like power outages). Shake-down/ramp-up and other activities that must be performed when switching between rockets and projectiles, and when switching between agent types, would require additional time. AEA believes that the SILVER II™ technology package is a conservative design; that is, it has sufficient equipment redundancy and overcapacity to readily meet the required destruction rate at the assumed equipment reliabilities. To the committee’s knowledge, although large numbers of electrochemical cells are operated in parallel in the chemical industry, such operations have not faced the materials handling challenges that AEA’s SILVER II™ agent and energetics destruction systems face. Hence, assumed equipment reliabilities may be optimistic. Materials of Construction Even in the relatively short-duration Demo II tests, it was apparent that corrosion from hydrogen fluoride (HF) in the anolyte and catholyte circuits would be a serious problem (NRC, 2001b). Results of Demo II coupon tests showed significant weight losses during exposure to simulated operating environments. Therefore, AEA conducted additional material coupon tests in the EDS II testing in addition to testing conducted with PTFE-lined piping segments. Specifically, these tests addressed the performance of two of the proposed lining materials, PTFE and PFA, and an alloy, Inconel 690. PTFE-lined piping segments were also tested during the 12-kW engineering tests. Piping with PTFE lining only 1/16-inch thick performed acceptably in the severe environment of the evaporator tests. AEA also
OCR for page 61
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot TABLE 3-6 Materials Selection for Key Sections of a SILVER II™ Full-Scale Plant Section or Unit Operation Provisional Materials Selections Anolyte circuits Agent: generally all process-fluid-wetted parts to be lined with PFA, PTFE, or PVDF. Piping to be PFA-lined carbon steel throughout. Pumps, valves, etc. to be PFA-lined, ductile cast iron. Agitators etc. to be PTFE-coated carbon steel. Heat exchangers to use PFA for all wetted surfaces. Seals and gaskets to be Kalrez, PTFE, or PTFE-clad; hydrocyclones to be PFA-lined carbon steel. Energetics: generally all process-fluid-wetted parts to be PFA, PTFE, PVDF, or titanium; piping to be stainless steel with PFA lining; pumps to be titanium; valves to be PFA-lined ductile cast iron; agitators to be PTFE-coated or PVDF-clad carbon steel; heat exchangers to be titanium with PFA or PVDF-wetted surfaces; seals and gaskets to be Kalrez, PTFE, or PTFE-clad; hydrocyclones to be titanium. Electrochemical cells End plates and ports to be PVDF-lined ductile cast iron. Gaskets to be Viton for agent and Viton GF for energetics. Electrodes to be platinized titanium, pinhole-free, to a thickness of 5 micrometers for the anodes and 2.5 micrometers for the cathodes. Catholyte circuits Agent: same as anolyte circuit. Energetics: piping to be PFA-lined carbon steel; pumps to be titanium; valves to be PFA-lined ductile cast iron; agitators to be PTFE-coated stainless steel; seals and gaskets to be Kalrez, PTFE, or PTFE-clad. Heat exchangers to be 304L stainless steel. Catholyte evaporator to be Inconel 690/625; hydrocyclones to be titanium. Polishing circuits Agent: same as main anolyte and catholyte circuits. Energetics: same as main anolyte and catholyte circuits. NOx reformer Agent: piping to be PFA-lined stainless steel; pumps and valves to be PFA-lined ductile cast iron; heat exchangers to be PFA or PVDF on wetted surfaces; seals and gaskets to be Kalrez, PTFE, or PTFE-clad; column sections to be PFA-lined carbon steel; packing to be PVDF. Energetics: piping to be PFA-lined stainless steel; pumps to be PFA-lined carbon steel; valves to be PFA-lined ductile cast iron; heat exchangers to be PFA or PVDF on wetted surfaces; seals and gaskets to be Kalrez, PTFE, or PTFE-clad; column sections including trays to be Inconel 690/625; packing to be Inconel 690/625. Offgas scrubber Agent and energetics: piping, column, and packing to be 304L stainless steel; pumps to be polypropylene-lined carbon steel. Reagent feed systems Agent and energetics: process water, silver nitrate, sodium hydroxide, and sodium hypochlorite feed system tanks, pumps, pipes, and valves all to be 304L stainless steel; gaskets to be EPDM or similar. Agent: nitric acid and hydrochloric acid feed system tanks, pumps, pipes, and valves all to be PFA-lined carbon steel; gaskets to be PTFE, Viton GF, or similar (e.g., Kalrez); hydrochloric acid pumps and valves to be PFA-lined ductile iron. Energetics: nitric acid and hydrochloric acid feed system tanks, pumps, pipes, and valves all to be 304L stainless steel; gaskets to be PTFE, Viton GF, or similar (e.g., Kalrez). Fasteners, coatings, etc. All fasteners (nuts, bolts, etc.) and other unpainted structural elements to be 316L or 304L stainless steel depending on the application. Protective coating for structural steelwork and carbon steel or ductile cast iron plant items—to be developed. All spill containment to be 304L stainless steel. Wiring and instrumentation Wetted instrument parts and pockets generally in accordance with wetted plant items in applicable area. Transmitter housings generally coated as structural steelwork above or constructed from 304L stainless steel according to availability and life-cycle cost. Cable insulation to be PVC, and carbon steel conduit (coated in accordance with guidance above) to be employed throughout. SOURCE: Adapted from AEA (2001b). reports that PTFE is the preferred material in the semiconductor industry, where aqueous HF is used extensively and leaching from HF containers would be unacceptable for maintaining product purity (AEA, 2001o). Based on this information and additional research, AEA has made the provisional selection of the materials of construction for the SILVER II™ process systems shown in Table 3-6. This selection of materials should be adequate, assuming that AEA addresses the issue of assuring reliability in achieving leaktight joints after repeated disconnection and reconnection of piping and other system components. The committee also notes that AEA had not resolved corrosion of the platinized titanium electrodes observed during EDS II fluoride transport tests but expected to achieve a satisfactory solution. Maintenance Issues The full-scale plant provides for maintenance to be performed in DPE suits in areas where agent may be present. To minimize downtime, AEA has developed a detailed spare parts policy covering installed and stored
OCR for page 62
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot spares. All pumps and hydrocyclones within the plant will have an installed spare (i.e., a standby) to allow remote changeover. The pumps have an automatic changeover system initiated by flow instruments and motor drives and a manual override. Hydrocyclones have manually initiated changeover only, since failure would be detected by a combination of installed instrumentation and operator judgment. An additional spare for each pump and hydrocyclone is held in storage. Isolating valves and flushing lines are provided to allow repair or replacement without requiring a plant shutdown. One spare electrochemical cell stack is installed in the primary anolyte circuit. Manual intervention is required to connect the spare cell stack and disconnect a faulty cell stack. Five spare cell stacks are kept in storage, allowing replacement of all primary or secondary electrochemical cell stacks (but not both at once) in the case of common-mode failure, e.g., severe blockage. The inventory of spare cell stacks was not deemed necessary to cover common-mode failure of both primary and polishing (secondary) electrochemical cells, because their anolyte circuits are separate and the catholyte circuit is much less likely to be the source of failure (AEA, 2001a). The use of installed spares and rotating equipment out of service for periodic maintenance should reduce the impacts on production rate from failures in equipment. This conclusion assumes that the resulting complexity of piping layout and increased manipulation of piping connections does not negate the goal of assuring process reliability. Process Safety Preliminary hazard analyses (PHAs) have been conducted for the SILVER II™ process at various stages of design and have served as building blocks for the EDP PHA effort (AEA, 2001a). These PHAs use the Failure Modes and Effects Analysis (FMEA) technique in accordance with the following regulations and standards: MIL-STD-882D, Department of Defense Standard Practice for System Safety (DoD, 2000) American Institute of Chemical Engineers (AIChE) Guidelines for Hazard Evaluation Procedures, Second Edition with Worked Examples (AIChE, 1992) Occupational Safety and Health Administration (OSHA) Regulation 29 CFR Part 1910.119, Process Safety Management of Highly Hazardous Chemicals Environmental Protection Agency Regulation 40 CFR Part 68, Risk Management Program The FMEA is used to identify single equipment and system failure modes and each failure mode’s potential effect(s) on the system or plant. The analysis generates recommendations for increasing process safety and equipment reliability. All equipment failure modes are assigned a risk assessment code (RAC) based on frequency and consequence severity categories assigned to the failure mode in accordance with MIL-STD-882D (DoD, 2000). Those failure modes with RACs of 1 or 2 require recommendations for design or procedural reductions in risk to RAC 3 or 4. After proposed fixes are identified, the failure mode is reranked as if the recommendation or fix had been implemented. This process is repeated until all items have an acceptable RAC (3 or 4). The approach described is appropriate for assuring process safety for the preliminary design and is commonly used in industry where hazardous materials are handled. The committee further notes that a full-scale quantitative risk assessment (QRA) will be required in conjunction with the completion of the final design to assure that all process safety issues have been fully addressed. An operational safety aspect of particular concern is the control of the oxidation reactions in the SILVER II™ agent and energetics systems. Since the process depends primarily on the generation of Ag2+ ions, the electrochemical oxidation reaction stops immediately when the power is switched off. Some chemical oxidation by nitric acid reaction also occurs, but AEA has reported that this reaction is a minor (less than 5 percent) contributor to the agent or energetic oxidation (AEA, 2001b). Since the technology provider has designed the process to operate on process streams with low concentrations of organics, the SILVER II™ process can be safely and quickly shut off at any time (e.g., from safety interlocks at other stages of the overall process) without any consequences associated with thermal or chemical inertia. Precipitation of trinitrobenzene or other sensitive energetic materials in the energetics slurry handling and SILVER II™ system was identified as a concern during Demo II testing. Design changes identified in the final EDP and demonstrated in EDS II 12-kW tests appear to adequately prevent precipitation.
OCR for page 63
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot Worker Health and Safety As in the baseline system facilities, much of the operation for a SILVER II™ plant is controlled from remote locations that protect workers from explosions and exposure to agent. ECVs and explosion containment rooms (ECRs) are used in the same manner as in the baseline technology. An area of possible concern is the technology provider’s decision that all SILVER II™ cells processing energetics will be operated in Category C areas in the MDB layout. AEA indicated that the energetic slurry would normally be agent-free. The committee is concerned that cross-contamination of energetic materials with agent could occur during planned agent/energetics accessing and removal operations using modified disassembly machines. This cross-contamination may require reclassification of Category C areas to Category A areas for SILVER II™ energetics processing since normal maintenance might expose internal cell parts that have become agent-contaminated. Since DPE suits were used in conjunction with the 12-kW tests, the committee concludes that they will be suitably safe for use in the possible presence of concentrated (60 percent) nitric acid. Other hazardous chemicals that would be found in an AEA SILVER II™ plant are typical of those in any large chemical or electrochemical plant and appropriate worker safety equipment and procedures would be used. The SILVER II™ NOx reformers require oxygen (introduced through the catholyte vessels). Preliminary material balances indicate the quantity of oxygen consumed in a disposal campaign for GB rockets at Blue Grass would be approximately 1.6 million pounds. The oxygen would be supplied as liquefied oxygen and vaporized to a gas on-site. Although liquefied oxygen is a hazardous material, it is routinely used in chemical plants and in the NASA space shuttle program without significant safety problems. Public Safety Accidental releases of agent or other hazardous materials are expected to be no more likely for a SILVER II™ plant than for plants using other technologies for destruction of agent. The most significant source of accidental exposure to the public remains the rockets in storage and transport to the chemical demilitarization facility. Human Health and the Environment Characterization of Effluents In its Demo II report, the committee noted that the following gaseous effluents from the SILVER II™ process were analyzed (NRC, 2000a): Anolyte gas was measured for CO, SO2, VOCs, semivolatile organic compounds (SVOCs), agent, and Schedule 2 decomposition compounds. Prereformer and postreformer gas was measured for O2 and NOx. Discharged offgas was measured for CO2, O2, CO, N2, N2O, H2, SOx, NOx, VOCs, SVOCs, agent, and Schedule 2 decomposition compounds. The Demo II tests revealed the presence of VOCs in the offgas stream and, as described earlier, the technology provider has included thermal treatment and a CATOX unit in the final design for the combined offgas streams. Full characterization of the gases from the combined SILVER II™ agent process and MPT process streams has not been possible since such a stream was not produced in testing. The same is true for the combined SILVER II™ energetics process and DTS process streams. However, testing of the Parsons/ Honeywell WHEAT CST technology has allowed characterizing the results for treatment of dunnage and DPE suit material. This testing, while not directly representative of SILVER II™ DTS operation, indicates the need for changes in design and operating conditions of the DTS and CATOX units to reduce the formation of dioxins and furans when processing DPE suit material. As noted earlier, CO emissions at a SILVER II™ facility at Blue Grass would be significant, 90 ton/yr, even with the anticipated 98-99 percent efficiency of the CATOX units. The liquid discharges from the anolyte circuit, catholyte circuit, NOx reformer, and caustic scrubber were sampled and analyzed for metals and organics. At the time this report was prepared, not all of the data were available to the committee. As noted in the Demo II test report, the Army has concluded that the characterization of the products from agent and propellant destruction showed that most hazardous intermediates were encountered at relatively low levels that could be effectively destroyed with additional treatment steps (NRC, 2001b).
OCR for page 64
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot Completeness of Effluent Characterization The effluents from the integrated SILVER II™ processes had not been completely characterized when this report was prepared. Prior committee assessments of effluent characterization (NRC, 1999, 2001b) are still considered valid. Effluent Management Strategy The AEA effluent management system proposal is to send dilute nitric acid that is not used as process water makeup as a waste stream to a publicly owned treatment works (POTW) under a pretreatment exemption. An analysis of the characteristics of effluent from the Demo II tests shows that it could be disposed of in a POTW; however, it is not known whether such a facility exists. At the time this report was prepared, the solid products from the MPT, DTS, and the evaporator had not been characterized. Tests of silver recovery from silver chloride demonstrated that economic silver recovery was possible (AEA, 2001m). Although the silver chloride used in the tests was not generated from purged anolyte, impurities in the silver chloride effluent produced from purged anolyte are not expected to alter this conclusion, which is based on EDS II tests using a simulated anolyte purge solution (AEA, 2001n). Off-site Disposal Operations The anolyte in agent and energetics SILVER II™ systems will be constantly purged so that metals and mineral acids that accumulate in the catholyte and anolyte circuits can be removed. These contaminants will be collected in the evaporator bottoms and disposed of using standard industrial waste treatment techniques. The contaminants will include iron, lead, and aluminum from the munitions as well as sulfur, phosphorus, and fluorine from the agent treatment; small amounts of silver not recovered during the AgCl precipation; and trace amounts of organic carbon remaining after the polishing step. The occurrences of all species but silver and organic carbon are generic to all processes that destroy agents and/or energetics. Some concentrated nitric acid may be produced as a by-product in the SILVER II™ process when processing energetics compounds, which contain nitrogen. The material will be collected and sold to off-site users. During Demo II testing, analysis of nitric acid samples by the Ensign-Bickford Company determined that the acid could be used in the manufacture of energetics at its Kentucky facility (NRC, 2001b). While it is currently planned to process uncontaminated dunnage in the DTS, the committee notes that this material could also be disposed of off-site. Environmental Compliance and Permitting The permitting process for the SILVER II™ technology package is expected to be substantially more complex than for current baseline facilities. The permitting protocol is well established for the baseline facilities. SILVER II™ will require a Subpart X permit, for which there is no precedent. The committee also notes that the SILVER II™ total solution technology package for Blue Grass entails a very large scale-up from 3 cells during Demo II and EDS II testing to 432 cells for agent destruction and 378 cells for energetics destruction, which will be a concern for permit writers.9 In seeking an RD&D permit,10 for example, permit writers would need to be assured that the applicant knows where the materials, particularly the metals, are going, which in turn demands that permitters have confidence in the mass balances for the technology. Mass balances currently in the EDP do not fully characterize the process streams, particularly the effluent streams. Although the committee sees no reason to conclude that permitting cannot be accomplished, not-withstanding the current absence of a full characterization of all process effluents, it believes the permitting will be very complex. There is also a need to develop a very comprehensive set of performance data to allow operating conditions to be established even after a facility has been designed and built. Further insight into the permitting challenge is given by considering the following list of candidate permitting requirements and issues being developed by the Department of Energy for the SILVER II™ process that it is considering for destruction of spent organic solvent from its plutonium-uranium solvent extraction processes (DOE, 2001). These requirements and issues include: 9 The 12-kW plant used for Demo II and EDS tests uses three cells operating at 2,000 amps maximum per cell at 2 V. 10 RD&D stands for research, development, and demonstration. This is a special type of RCRA permit that allows testing of treatment units for development and demonstration purposes. These types of operations require a permit even though it is much less stringent and demanding than a full RCRA operating permit.
OCR for page 65
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot Establishing regulatory-based performance standards for the technology covering regulatory parameters of interest such as these: A DRE for organics in offgas of 99.99 percent and a risk-based, site-specific analysis. Site-specific risk analysis of VOC emissions in vent gases. Levels of NOx emissions in relation to Clean Air Act ambient standards and the risk-based alternative acute emissions guidance levels from Volume I of the Hazardous Waste Combustion Risk Analysis Guidance Document (EPA, 1998). This will address NOx as potentially convertible to nitric acid. Characteristics of solid and liquid residuals in relation to land disposal restrictions requirements using the toxic constituent leaching procedure and analysis of the underlying hazardous constituents such that implementation of proper disposal options can be ensured. Site-specific risk analysis of oxidation by-product emissions, including dioxins and nitrated polyaromatic hydrocarbons. Metals emissions, e.g., mercury emissions, in relation to EPA’s hazardous waste combustion maximum achievable control technology (MACT) standards and a site-specific, risk-based analysis that is particularly focused on silver and other metals impacted by the formation of chlorinated and nitrated volatile metals. Particulate matter emissions in relation to hazardous waste combustion standards. Establishing operating parameters of interest to permit writers (compliance assurance): SILVER II™ oxidation conditions Current flow to cell Silver content of anolyte circuit (AgNO3) makeup rate Waste feed cutoff tied to current flow Temperature of anolyte feed tank (main concern is precipitation of by-products) Anolyte mixer operation Anolyte circulating pump flow rate Total organic level in anolyte circuit Waste feed characteristics Waste feed rate to anolyte cell Maximum particle size Maximum chlorine feed rate (AgCl precipitation) Tramp metal and other inorganic impurity concentration Hydrocyclone Minimum and maximum pressure drop NOx reformer operating conditions Condenser maximum temperature Oxygen feed rate Distillation operating temperature Offgas condenser Maximum operating temperature Cooling liquid flow Scrubber operating conditions Caustic flow rate Scrubber pH Catalytic oxidizer operating conditions Minimum and maximum operating temperature CO destruction efficiency continuously or at set intervals Confirmatory dioxin tests HEPA filter operating conditions (separate analysis under way) Minimum and maximum pressure drop Installation testing Maximum shelf life Maximum temperature Relative humidity limits Carbon filter operating conditions Maximum operating temperature Carbon replacement rate Offgas flow rate Emissions monitoring Continuous emissions measurements of CO/CO2/O2 in vent gases Continuous measurement of NOx emissions in vent gases Continuous emissions measurements of total hydrocarbons Chlorine emissions monitoring Hold-test-release protocols It is clear from the preceding list that permitting will be challenging for the AEA SILVER II™ technology package for Blue Grass. ASSESSMENT OF OVERARCHING TECHNICAL ISSUES Overall Engineering Design Package The overall EDP offered for AEA’s SILVER II™ technology has grown in complexity from what was presented during the Demo I and Demo II phases of the ACWA program. Some of this growth in complexity was to be expected as the level of design developed. The EDP design does appear to reflect most of the configuration and equipment necessary to implement a full-scale pilot plant final design for the current state of understanding of the SILVER II™ technology. The complexity of steps identified for this technology has continued to increase as greater understanding of the detailed requirements of the process is developed. These new requirements arise not just from more detail in design, but also from the identification of additional unit operations that must be included for satisfactory operation, particularly in the agent and energetics destruction systems. Some of these requirements were identified in Demo II testing, others have continued to be identified as EDS II testing proceeds. These changes appear to arise from the fact that the basic process, while understood in principle, is not well understood in detail. This reflects a significant immaturity in the SILVER II™ technology for this very demanding application and raises concerns that more complexity would arise as the design is further developed. For example, the current assumption that all energetics destruction by SILVER II™ will take place in a Category C area may not be valid if the current methods of agent and energetics separation cannot provide satisfactory assurance that the energetics going to SILVER II™ processing are agent-free. Should the current
OCR for page 66
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot assumption prove invalid, major changes in MDB design and costs might be required. Likewise, the MPT design is based on removal of 99 percent of agent from metal parts before MPT processing. If this efficiency is not achieved, further changes in the metals washing and/or MPT processes may be required. The committee believes that further testing during systemization of a full-scale facility would reveal a need for more than the usual number of design changes. Some aspects of the current design that point to the likelihood of further change are the following: The RDM agent cavity washout scheme and the use of a tube cutter to cut shipping tube and rocket is a significant departure from the baseline process and AEA’s previously specified process for removing firing tubes and cutting the rockets with a high-pressure water jet using garnet grit (AEA, 2001b). The committee notes that the method of cutting is unproven for this application, and there are significant uncertainties associated with differential movement between the tube and the rocket body as both are rotated and cut. In addition, the proposed method of washing or steaming agent out of the rocket agent cavity has not been demonstrated and, if this method is not as effective in removing agent as planned, it is unclear that the MPT and its associated gaseous effluent system would be capable of handling the additional agent loading. The destruction of fuzes and supplemental charges has changed. Instead of using a standard detonation chamber, they are simply placed in the MPT and allowed to decompose or detonate. At the same time, the MPT design has changed—it now uses a steam-heated industrial oven with internal steam recirculation, which has not been demonstrated for destruction of fuzes and supplemental charges. The technology provider now plans to use electrical steam heaters and internal steam recirculation for the MPT. Although this configuration seems feasible and is a variant of the MPT technology tested for the Parsons/Honeywell WHEAT technology package, the committee notes that it has not been demonstrated in its proposed configuration and that the steam-reactant environment may pose significant materials technology challenges and result in different chemical loadings in the gaseous effluent treatment system. Also, the operating duty of the recirculation blowers, which are required to cycle between near ambient and over 1,000°F every 4 to 6 hours, will challenge their reliability. (When the MPT is recharged every 4 to 6 hours, the high-temperature zones are exposed to cooler temperatures as a result of charging and discharging.) The DTS is based on the CST of the Parsons/ Honeywell WHEAT technology package (Parsons, 2001). However, the DTS design uses an enclosed auger to move the dunnage as it is being decontaminated to a 5X condition. This enclosed auger was selected to prevent dust and debris from accumulating in the bottom of the dunnage treater. The ability of the enclosed screw to operate without significant plugging and with adequate flow of the steam sweep gas has not been demonstrated. Many major design changes have taken place in the agent and energetics SILVER II™ process systems. For example, High-shear vortex mixers are included in all feed tanks to provide final size reduction and mixing of particulates/precipitates that may exist in electrolyte slurries. These mixers replace inline mixers that had been proposed at the end of Demo II testing. Also, conventional stirrers will be installed in large-diameter tanks to ensure that the high-shear vortex mixers contact the entire contents of the tank (AEA, 2001a). The basis for this combination of mixing devices is experience in a 3-foot-diameter vessel and will require revalidation in the geometry of the 10-foot-diameter full-scale vessel. A catholyte recycle circuit was added following Demo II testing because of greater than anticipated flow of metal ions, mineral acids, and organic material from the anolyte to the catholyte in the electrochemical cells. EDS II testing revealed the need to monitor nitric acid concentration as well as total acid concentration in the catholyte circuit to prevent formation of silver deposits that led to arcing to the cell membrane and the development of holes in the membrane. More changes may be necessary to maintain the higher nitric acid concentrations in the catholyte. Demo II testing identified the necessity of an
OCR for page 67
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot anolyte polishing step in the IRS. The design used is based on limited rundown tests at the end of the EDS II 12-kW testing. The IRS has been modified to incorporate a batch step for removal and recovery of silver ions accumulating in the IRS circuit. This removal operation involves the addition of 35 percent HCl and subsequent collection and treatment of the AgCl precipitate for off-site removal. The process has not been demonstrated on purged anolyte solution with metal, mineral acid, and trace organic impurities. The purge of anolyte to the IRS also requires use of an evaporator to recover nitric acid and produce an anolyte-derived brine waste containing mineral acids and metals from anolyte (i.e., the impurities removal operation). Concern for cavity plugging of the cell electrodes from slurry particulates has resulted in the use of hydrocyclones on all cell feed streams in both the main circuit and the polishing circuit. AEA states (AEA, 2001o) that the anolyte and catholyte circuits use a ‘multi-clone’ arrangement of hydrocyclones with their overflows combining to a single process line, which is then fed to the electrochemical cells. Their underflows are combined to a common discharge line and returned to the respective anolyte or catholyte tanks. . . . This arrangement is common in industry where it is necessary to achieve a high flow capacity and a particular particle cut size. The proposed multi-clone arrangement has been designed with slurry handling and trace heating considerations to prevent any premature settling and blockage formation. In addition, the full-scale plant employs an additional multi-clone as a stand-by spare so that any changes in performance due to blocking or plugging (observed by pressure or flow changes) can be rectified by switching to the installed spare. Integrated operation of these hydrocyclones has been demonstrated in a three-cell configuration. However, scale-up to 432 cells may present new challenges in flow and pressure management to sustain satisfactory hydrocyclone operation. Reevaluation of Steps Required for Implementation In 1999 the ACW I Committee identified several key steps that would have to be implemented before the AEA SILVER II™ technology could be fully implemented (NRC, 1999). These steps were first reevaluated by the ACW II Committee following the Demo II tests (NRC, 2001b). These steps are again reevaluated here on the basis of the results of the EDS II testing and EDP. 1. Modified shearing locations for M55 rockets and a new shearing machine must be tested to show routine segregation of components and reduction in particle sizes to less than 1/4 inch in diameter. This still needs to be evaluated. The technology provider plans to test the new RDM at the vendor’s facility prior to shipment. No proof of principle prior to this time is anticipated, especially demonstration of the efficacy of the tube-cutting concept for concurrent cutting of firing tubes and contained rockets. Also, the capability of the punch-and-drain and rocket-cutting operations to prevent agent contamination of energetics has not been demonstrated. Reduction of propellant size by grinding under water appears to be feasible with the current concept. However, it is noted that this feasibility depends very much on successful propellant grain extraction using the untested RDM concept. 2. The modified mine shearing approach must be tested. Since no mines will be processed at Blue Grass, there is no need to test mine shearing. 3. The dissolution of fuzes and mine bodies in nitric acid and SILVER II™ solution must be evaluated. This is no longer relevant, because fuzes and any supplemental charges will be fed to the MPT for deactivation. 4. All effluents must be characterized in detail when treating agents contaminated with metals from disassembled chemical weapons (i.e., potential trace species and reaction by-products, such as nitrated hydrocarbons, partially oxidized products, and metals, must be identified) and their environmental impacts evaluated. The Demo II test evaluated all of the major effluent streams for a full suite of trace species and reaction by-products. At the time the committee was preparing its Demo II report, not all of the data were available and the impact of trace species, particularly in brines and atmospheric releases, on facility permitting remains to be determined. This information was still unavailable to the committee as of the time the present report was being prepared. 5. Demonstrations of the scale-up, development, and integration of hardware with real materials of construction must focus on the robustness of the parallel flow in multiple-cell reactors. The issues of cell blockage, hydrocyclone performance, and NOx reformer performance must be addressed. The committee again stresses the importance of this
OCR for page 68
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot step in light of the increased complexity of the process after the changes proposed to address problems revealed during the Demo II tests. As noted in the earlier discussion on design changes, the technology provider has included hydrocyclones on all cell feed streams. While the hydrocyclones will surely improve particulate management, the committee remains concerned about the robustness of cell flow control in the presence of particulates that pass the hydrocyclones, especially considering the large array of parallel flow paths (up to 432 on the catholyte circuit that is common to both the main anolyte and polishing circuit for agent destruction). 6. The efficacy of high-pressure jet washout of agent and gelled agent from M55 sheared pieces must be tested. This recommendation was not addressed in the Demo II tests or EDS II tests. It remains valid even with the proposed RDM design that uses tube cutting and a hollow upper punch with steam/hot water washout. The technology provider plans to conduct testing at the vendor’s facility, but this testing can only be conducted under simulated conditions, and it may be insufficient to reveal design deficiencies with respect to removal of agent heels to the levels required by the MPT design. 7. The treatment of burster charges and M28 propellant in the SILVER II™ reactor must be tested, and the material preparation required to ensure reasonable treatment times with no energetic events must be evaluated. This testing must also determine what happens to the lead stearate in the propellant during SILVER II™ treatment. The tests with the 12-kW system in Demo II successfully confirmed that this technology is capable of destroying the components (nitrocellulose and nitroglycerine) of M28 propellant. For tetrytol, the destruction of TNT and tetryl was good. However, recalcitrant intermediate products were formed during the treatment of tetrytol, which AEA was still evaluating at the time the Demo II report was prepared (NRC, 2001b). Subsequent EDS II testing in the 12-kW system established that a better design that eliminates dead legs, along with trace heating that removes cold spots in the piping and vessels, can eliminate the precipitation of recalcitrant intermediate products. Examination of the 12-kW system following test runs with energetics mixtures containing M28 propellant (containing ~0.5 weight percent lead as lead stearate) also showed that the lead in the lead stearate oxidizes to lead oxide and that much of this material is deposited in the electrode cavity (AEA, 2001d). The committee notes that a material balance was not performed on lead, so it is unclear whether lead oxide will also be carried to the IRS with the anolyte purge stream. If so, it would be expected to be removed in the evaporator brine as a salt. AEA says that it has demonstrated in laboratory-scale tests that lead oxide can be removed from the cells as lead formate using off-line flushing with formic acid solution (AEA, 2001a). The EDP includes provision for removing one cell stack at a time and performing a formic acid wash or flush to remove the accumulated lead oxide. An installed spare cell stack would be put online during the formic acid flush, and the flushed cell would become the installed spare. While this operation was not demonstrated during EDS II testing, it appears to be implementable. Therefore, the committee concludes that energetics destruction by the SILVER II™ process has been satisfactorily demonstrated under planned design and operating conditions. 8. The process must be developed and tested for the efficacy of submerged-bath dilute nitric acid treatment for metals parts, including the effects of agitation and temperature. This recommendation is no longer relevant because the current design now uses an MPT to decontaminate metal parts to a 5X level using high-temperature (>1,000°F) steam at near atmospheric pressure. 9. The treatment of shredded dunnage material must be tested in a prototype-scale SILVER II™ reactor. This is no longer relevant because the shredded wood and DPE suit material will be treated in a dunnage treater similar in design to the CST of the Parsons/Honeywell technology package demonstrated in EDS I testing, and which achieved 5X decontamination. The enclosed auger design of the dunnage treater used in AEA’s DTS remains to be demonstrated. 10. Techniques for controlling particulate matter to prevent plugging of SILVER II™ electrolytic cell channels must be developed and demonstrated. The ACW I Committee’s concern at the time this step was listed was plugging in the cell channels from the formation of AgCl precipitate. AEA has since provided hydrocyclones to remove this precipitate. These hydrocyclones were used in EDS II 12-kW test operations and were found to perform well. However, the committee still remains uncertain about successful operation in a full-scale plant (where there are up to 432 parallel paths vs. the 3 parallel paths that have been
OCR for page 69
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot tested), since particulates, albeit less than 100 microns, still remain in the feed to the cells and can cause plugging if not managed properly. 11. Materials of construction must be evaluated under corrosive and oxidizing conditions. While the committee recognizes the viability of the liner materials chosen following Demo II and EDS II corrosion tests to deal with the severe corrosion conditions, it remains concerned about the ability to readily achieve leaktight connections routinely after both scheduled and unplanned maintenance activities. It is also concerned about AEA’s conclusion that stainless steel can be used in the agent NOx reformers based on the EDS II fluoride transport tests. From these tests, it did not appear that fluorides existed in the gas stream to the reformer; however, the volatility of HF suggests that such an outcome would not be unexpected. More long-term testing of fluorine-containing electrolyte derived from organic materials containing fluorine would be prudent. 12. The realistic potential for off-site recycling/reuse of silver salts and concentrated nitric acid must be evaluated, including recyclers’ ability to accept, handle, and treat these materials. During EDS II testing, it was demonstrated that AgCl decontaminated to a 5X level could be sent to an off-site silver recycler and that the silver could be economically recovered and returned to the plant for silver nitrate makeup. The committee notes that this testing used a simulated anolyte purge that was doped with metals and minerals acids but did not include lead or the organic carbon that remains after polishing. Thus, while recycling of silver seems achievable, there may be further challenges to process efficacy when material from actual anolyte purge streams is used. FINDINGS AND RECOMMENDATIONS Findings Finding (Blue Grass) AEA-1. The SILVER II™ systems for processing agent and energetics have become increasingly complex since Demo II. The increasing complexity appears to be driven primarily by the impacts of phenomena not previously considered important. This indicates that AEA’s EDP design is still too immature for implementation. Test experience pertinent to the stage of maturity of the process includes the following: During Demo II and EDS II testing, only a 3-cell system was operated, which is much simpler than a full-scale 432-cell flow system with all cells in a parallel flow path and fluids containing significant levels of suspended solids. Cell membrane failures are not predicted by cell membrane life tests, and there has been no long-term testing on feeds containing Cl and F. There are plans for the removal of lead oxide using formic acid wash without prior demonstration in a multicell configuration. Tests with a 3-ft-diameter vessel will be assumed to apply to a 10-ft-diameter, full-scale vessel without development of demonstrated scaling parameters for high-shear vortex mixers and stirrers working together. Electrolyte chemistries continue to be discovered and need to be carefully controlled, e.g., there is a need to track nitric acid molarity as well as total acid molarity in the catholyte to avoid silver deposition, and the deposition of lead dioxide on cell electrodes and in electrode cavities, which has required the development of a formic acid wash. The full-scale design has effectively increased cell power requirements by 50 percent to achieve the required destruction efficiencies. These increases were not identified until EDS II testing was performed. There was excessive corrosion of platinum-plated titanium electrodes in the presence of fluorine-containing anolyte feed streams. Improvements in plating techniques are expected to solve this problem, but they have not been demonstrated. Finding (Blue Grass) AEA-2. Although the RDM proposed by AEA follows the basic steps used in the baseline system, it implements the steps with a suite of new equipment that has not been built or demonstrated. No evidence has been provided of such equipment (e.g., tube cutting, burster washout, and grain extraction) being used in the manner required for an integrated RDM for the SILVER II™ technology package. Past experience suggests the design will have to be
OCR for page 70
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot modified as testing occurs, and it is not clear that modifications will ensure that energetics feeds are agent-free, thereby allowing operation and maintenance in a Category C area. Finding (Blue Grass) AEA-3. Long-term operation of the SILVER II™ cell membrane on all the required feed streams has not been demonstrated. Because the SILVER II™ process is the primary treatment process for agent and energetics, the committee is particularly concerned that the life of the cell membranes, on which the efficacy of the SILVER II™ process depends, has not been demonstrated. Reasons for this finding include the following: Long-term operation on H and GB feeds containing chlorine and fluorine has not been demonstrated. The agent simulant, DMMP, does not contain these elements. The laboratory-scale fluoride transport tests revealed severe attack on the Pt plating of the electrodes. AEA is now seeking resolution of this problem with electrode manufacturers. Pt and Ti released in these attacks may plug electrode cavities or impact membrane performance. The slurry flow management scheme to the cells has large numbers of parallel flow paths through the hydrocyclones and through individual electrode cavities. Upsets in these paths can lead to upsets in the quality and quantity of slurry flowing to the electrode cavities, with possible impact on membrane operation. Finding (Blue Grass) AEA-4. The efficacy of using candle filters such as those that AEA has proposed has not been demonstrated. Experience with fouling of the CATOX unit during the Parsons/Honeywell CST tests suggests that the particulates are sticky and may not be readily removed from the ceramic candle filters by back flushes with nitrogen gas. The committee notes that the tests conducted on the CST system (which used cyclones rather than candle filters) are the only tests pertinent to the design proposed for AEA’s SILVER II™ DTS, and there are differences in the latter design that present untested/undemonstrated challenges to successful operation. Finding (Blue Grass) AEA-5. Existing CATOX unit EDS I tests to date indicate that there is still a significant risk of added delays and cost to achieve reliable operation, particularly with respect to meeting acceptable dioxin and furan levels (Parsons, 2001). This finding is based on EDS I CATOX tests by Parsons/Honeywell and also on the differences between the design tested and the offgas effluent treatment system proposed for the SILVER II™ process. Since the AEA CATOX unit is modeled after the Parsons/ Honeywell design, it can be expected to produce similar amounts of dioxin and furan. In addition, the gas treated by the AEA CATOX unit has a different source and therefore a different composition. No CATOX units have been tested on gases from this source. Finding (Blue Grass) AEA-6. The committee believes that a very complex permitting process will be required for the SILVER II™ technology package for Blue Grass, and that there is a need to develop a very comprehensive set of performance data to allow the operating conditions to be established even after a facility has been designed and built. Recommendation Recommendation (Blue Grass) AEA-1. Based on the above findings, the committee recommends that the AEA SILVER II™ process not be implemented at Blue Grass.
Representative terms from entire chapter: