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 17
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Pueblo Chemical Depot 2 Hydrolysis Tests of Energetic Materials Both the General Atomics and the Parsons/Honeywell technology packages use caustic (base) hydrolysis as the initial step in destroying the energetic materials recovered from chemical weapons (Table 2–1). The technology packages use different secondary treatments of the products resulting from hydrolysis (i.e., the hydrolysates). General Atomics treats energetic materials hydrolysate by SCWO; Parsons/ Honeywell uses biological treatment. In the first ACW I Committee report, the committee expressed concern about the immaturity of the technology base for the hydrolysis of energetic materials. As the committee noted, although alkaline hydrolysis was effective in destroying energetic materials to 99.999 percent, the chemistry of the process was not well understood (NRC, 1999). The committee reiterated these concerns in the supplemental report (NRC, 2000). In response to these concerns, PMACWA initiated a multilaboratory test program during the EDS phase to clarify the chemical and engineering parameters for the efficient, safe alkaline hydrolysis of the energetic materials in assembled chemical weapons. In this chapter, the ACW II Committee briefly reviews disposal practices currently in use for energetic materials, briefly describes the hydrolysis treatment process for energetic materials, and describes PMACWA’s program for engineering design testing of energetics hydrolysis. This is followed by an assessment of the status of the test program at the time this report was prepared, an evaluation of the results until that time, and a reassessment of the ACW I Committee’s original findings. CURRENT PRACTICES FOR THE DISPOSAL OF ENERGETIC MATERIALS In the past, DOD disposed of a large percentage of unwanted munitions and the energetic materials they contained by an open burn/open detonation (OB/OD) process, an environmentally undesirable approach that has already been banned in Europe. DOD has begun to work toward minimizing OB/OD as a means of disposal (JOCG, 2000). When an item is identified as surplus, it is transported to a site at which it will either be destroyed immediately or stored until it can be destroyed. Most of the sites that receive munitions for destruction are Army depots such as Hawthorne Army Depot (Nevada), Sierra Army Depot (Nevada), and Hill Air Force Base (Utah) (Mitchell, 1998). Other sites that sometimes receive munitions for experimental purposes include the Naval Surface Warfare Center (NSWC) Indian Head Division (Maryland) and the Nevada Test Site of the Department of Energy. Frequently discussed alternatives to OB/OD include processes known as resource reclamation and recycling (R3). These R3 processes are designed to either reclaim and recycle valuable metals from obsolete surplus ordnance or TABLE 2–1 Nominal Composition of Energetic Materials Used in Chemical Munitions Energetic Material Composition Tetryl 2,4,6 trinitrophenylmethylnitramine Tetrytol 70% tetryl/30% TNT Composition B 60% RDX/39% TNT/1% wax Composition B4 60% RDX/39.5% TNT/0.5% calcium silicate M28 propellant 60.0% nitrocellulose/23.8% nitroglycerin/ 9.9% triacetin/2.6% dimethylphthalate/2.0% lead stearate/1.7% 2-nitrodiphenylamine M8 propellant 52.15% nitrocellulose/43% nitroglycerin/ 3% diethylphthalate/1.25% potassium nitrate/ 0.6% ethyl centralite M1 propellant 84% nitrocellulose/9% dinitrotoluene/ 5% dibutyl phthalate/1% diphenylamine/1% lead carbonate Source: Bonnett, 2000.
OCR for page 18
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Pueblo Chemical Depot reclaim potentially valuable energetic components such as TNT, RDX, and HMX for reuse (Garrison, 1994). However, many of the R3 processes are still being evaluated for economic feasibility in a number of DOD demonstration programs (Newman et al., 1997; Marinkas et al., 1998; Goldstein, 1999). According to Mitchell (1998), “In 1998, approximately 60% of the 100,000 tons of demilitarization surplus ordnance were disposed in a way which enabled at least some of the material to be recovered and recycled.” Some energetic materials do not lend themselves to recovery and recycling, either because the economics of the process are unfavorable or because the material properties are unfavorable. Nitrocellulose-based propellants and materials containing nitrate ester plasticizers are not suitable feedstocks for an R3 program because of their long-term instability. Compositions containing these ingredients always include a stabilizer to prevent catastrophic self-heating as the materials age. However, the degradation of the propellants and the presence of impurities in aging energetics of this type make them poor candidates for the economical recovery of energetic components. These materials are still destroyed by OB/OD. Research has been done to evaluate the potential of demilitarized gun propellants for a variety of uses, such as sensitizers for commercial slurry explosives and boiler fuels (Machacek, 2000). The demilitarization of small items, such as igniters and fuzes, is routinely accomplished in an APE-1236 furnace, a rotary kiln in which the devices are heated until the energetic material decomposes thermally. The amount of material that could be recovered from these items is small, and the energetic materials themselves, especially detonators, are often quite sensitive. Because of their sensitivity, attempting to disassemble the items would be more hazardous than disassembling main-charge explosives. Therefore, these items are either intentionally “functioned” (i.e., actuated) or thermally decomposed. Alternative technologies to OB/OD for items that contain energetic materials not worth recovering are being explored but are not widely used. Confined burning, a process in which the gaseous and condensed products of combustion can be captured and treated before release, is being used at some sites around the country. Hydrolysis of energetics as a means of disposal is being used at the Hawthorne Army Depot. Several other technologies (e.g., molten-salt destruction) are being used at research and development sites (e.g., Eglin Air Force Base and Strauss Avenue Thermal Treatment Plant) to destroy energetic materials, but these technologies are not an integral part of DOD’s plan for the demilitarization of obsolete munitions. CAUSTIC HYDROLYSIS OF ENERGETIC MATERIALS Caustic hydrolysis of energetic materials has been investigated as an alternative technology to the OB/OD method. Newman (1999)1 published a review of the known chemistry of caustic hydrolysis of energetic materials used in assembled chemical weapons, and recent work on the destruction of aromatic nitro compounds (TNT and tetryl) by alkaline hydrolysis has been reported (Bishop et al., 2000). The chemistry of caustic hydrolysis takes advantage of the susceptibility of the functional groups commonly found in energetic materials to attack by hydroxide ion, which yields products that are essentially nonenergetic. Caustic hydrolysis decomposes energetic materials to organic and inorganic salts, soluble organic compounds, and various gaseous effluents. Partial hydrolysis of some energetic materials, particularly materials with aromatic ring systems, may lead to ill-defined oligomeric materials with low solubility in either aqueous or organic solvents. The rate of reaction depends on, among other things, the concentration of the energetic compound in solution or, for heterogeneous reactions, on the surface area of the solids being hydrolyzed. An important factor in determining the rate of destruction is the phase of the compound in the hydrolysis reactor. The compounds of interest may be divided into three classes: compounds that are liquids at normal reactor temperatures (e.g., 2,4,6-TNT and nitroglycerin) compounds that are solids at normal reactor temperatures (e.g., RDX and tetryl) polymeric materials (usually nitrocellulose) TNT has low solubility in aqueous solutions and forms an emulsion with hot caustic solution. Thus, because the TNT is molten, the size of the droplets in the emulsion is determined not by the size of the granules in the original feedstock but by the degree of agitation in the hydrolysis reactor, as well as the presence of any surfactant. For Composition B, the size of the RDX particles in the reactor will reflect the size that was used in manufacturing the Composition B.2 During manufacturing, a small fraction of the RDX dissolves in molten TNT, but the remainder is suspended in the TNT matrix. Therefore, when the TNT is remelted, the original RDX particles can be recovered. Thus, the RDX particle size does not depend on the size of the Composition B pieces fed into the reactor but on the size of the original RDX particles mixed into the TNT, typically between 10 µm and 1 mm. The particle distribution may be skewed toward the larger particles because the smaller particles dissolve more rapidly in the TNT. 1 This information can also be found in condensed form in Appendix E of the initial ACW I Committee report (NRC, 1999). 2 Composition B contains 1 percent wax. Depending on the nature of the wax, some long-chain fatty acids may be present, which act as surfactants. Hydrolysis of the plasticizers in M28 propellant may also release phthalate salts, which can aid in the emulsification of TNT when M28 and Composition B are processed together.
OCR for page 19
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Pueblo Chemical Depot Tetryl, like RDX, is a solid at the temperatures in the hydrolysis reactor. For the neat tetryl in burster charges, the grain size depends on the extent to which the pressed explosive charges have been processed prior to being added to the hydrolysis reactor. The case of tetryl in tetrytol is quite different. TNT and tetryl are very similar chemically, so the solubility of tetryl in molten TNT is quite high (82 g/100 g TNT at 80°C [176°F]; 149 g/100 g TNT at 100°C [212°F]) (Kaye and Herman, 1980). Tetryl in tetrytol is mostly dissolved in the TNT phase, so the rate of dissolution and subsequent reaction in the hydrolysis medium depends mainly on the TNT/tetryl droplet size and not on the particle size of the tetryl that was originally used to make the tetrytol. Polymeric nitrocellulose is solubilized by the degradation of the glycosidic linkages along the polymer chain. Nitrocellulose-based propellant grains contain other components that are released as the nitrocellulose breaks down. Even before it is completely broken down, infiltration of caustic into the partially decomposed propellant grains may allow the nitroglycerin inside the grains to come in contact with the caustic medium and react. In an effort to increase the reaction rate of the energetic materials in the hydrolysis reactor, strong caustic solutions (pH greater than 12), elevated temperatures (60°C to 155°C [140°F to 311°F]), and elevated pressures (up to 14 atm) have been used. The hydrolysis reaction is exothermic, so process controls are necessary to maintain the reactor temperature and respond to thermal excursions in order to prevent a runaway reaction. Because upsets are always possible, the reactors and containment buildings must be designed to contain the maximum credible explosive event.3 A thorough understanding of fundamental requirements for the hydrolysis of energetic materials is essential to the design and operation of a chemical agent demilitarization facility, where high levels of engineering controls are necessary to ensure the safe disposal of chemical agent and the prevention of process upsets throughout the facility. The ACW I Committee report indicated that four issues surrounding the caustic hydrolysis of energetic materials must be addressed (NRC, 1999): determination of the steps required for removing the energetic material from the munitions and reducing it to the appropriate particle size determination of safe operating parameters for heterogeneous mixtures of energetics, metals, and contaminants development of process controls and equipment to minimize the accumulation of precipitates and minimize the effects of an accident determination of the scale-up parameters to meet the destruction requirements for diverse munitions OVERVIEW OF THE TEST PROGRAM In response to the challenges listed above, the PMACWA devised an EDS test plan that addresses some of these issues. The start of testing was delayed from August 2000 to December 2000, and the tests were not to be completed until the end of March 2001. The responsibility for coordinating the program was assigned to U.S. Army Tank-Automotive and Armaments Command (TACOM) Armament Research, Development and Engineering Center (ARDEC) at Picatinny Arsenal, New Jersey. The Picatinny Test Plan Requirements describes objectives, planned testing, and team member responsibilities for the EDS test program (Bonnett, 2000). The main objectives of the test plan and some of the organizations involved are summarized below: Picatinny Arsenal and Holston Army Ammunition Plant (HAAP) are responsible for determining and defining optimum operating parameters for the hydrolysis of all energetic materials contained in munitions at the Pueblo Chemical Depot and the Blue Grass Army Depot. Los Alamos National Laboratory (LANL) and NSWC Indian Head Division (Yorktown, Virginia) are responsible for performing bench-scale tests to address ACW I Committee concerns about the solubility of energetics in alkaline solutions, the simultaneous processing of different types of energetics, and the reduction to the proper particle sizes for operation. TACOM ARDEC is responsible for incorporating data generated from bench-scale tests into full-scale production processes at HAAP to demonstrate the hydrolysis operations. The Pantex Plant is responsible for production of tetrytol hydrolysate. The Radford Army Ammunition Plant (RAAP) is responsible for production of M28 simulant hydrolysate. Because the EDS test program had not been completed at the time this report was submitted, only interim results were evaluated. Testing at the Holston Army Ammunition Plant The reactor at HAAP has the following characteristics: a 2,000-gallon, glass-lined, jacketed reactor vessel a recirculation loop, unheated and uninsulated, that 3 A determination of the maximum credible explosive event is made by considering the probabilities of accidental violent reactions (e.g., rapid burning or detonation of energetic materials) and the resulting damage and hazards associated with each event. The resulting hazards define the appropriate hazard mitigation and control strategies that are used to minimize the impact of an accidental explosion in a facility in order to protect people and property.
OCR for page 20
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Pueblo Chemical Depot reenters the reactor vessel at the bottom a dual-flight, Hastelloy C, variable-speed agitator in the reactor vessel a condenser/scrubber for off-gases pH-, temperature-, and flowmeters The energetic feeder is a single-screw, loss-in-weight feeder with a 500 lb/hr capacity. The control system uses programmable logic control (PLC) with remote and local control capabilities. Energetic material will be fed into the reactor as a dry solid and screened as it is added to the feeder hopper to prevent particles larger than 0.5 inches in diameter from entering the reactor. Demonstration runs are being conducted at various operating conditions for seven different energetic materials and mixtures (Bonnett, 2000): M1 propellant M8 propellant tetrytol M28 propellant Composition B Composition B4 M28 propellant/Composition B4 (86/14 weight-percent ratio based on amounts in an M55 rocket) The primary process parameters to be studied are (1) reactor residence time, (2) energetic feed rate, (3) reaction temperature, (4) caustic concentration, and (5) agitation speed. Although the munitions at Pueblo Chemical Depot contain only M1 propellant, M8 propellant, and tetrytol, the EDS tests with M28 propellant and Composition B high explosive are included in this chapter for completeness and because the tests are closely related. Composition B has been studied in anticipation of demilitarization operations at the Blue Grass, Kentucky, storage site; M28 propellant was used for testing the ERH, a component of the General Atomics technology package (see Chapter 3). An acceptable energetic feed rate will be based on (1) control of the heat generated to avoid thermal runaway, (2) completeness of the reaction, (3) achievable and effective settings for agitation speed, (4) elimination/minimization of foam formation, and (5) elimination/minimization of the production of undesirable by-products (Bonnett, 2000). The Army expects the time required for addition of the energetic materials to be inconsequential when compared with the total processing time required for the hydrolysis reaction. Therefore, a broad range of feed rates will be consistent with operational safety. Chemical sensors will be used and control strategies developed to ensure that all energetic materials are fully hydrolyzed and that process and operator safety is maintained. Other objectives of the system performance evaluation are summarized below (Bonnett, 2000): to demonstrate performance of the control sensors and the logic programming for the sensors to determine the efficacy of process cooling and temperature control to determine the efficiency of agitation by the impeller and the recirculation loop to determine the efficacy of steps taken to control foam formation to determine condenser performance to determine maintenance requirements to develop a contingency for sudden shutdown Hydrolysate samples collected during tests will be used in the following ways: to characterize fully hydrolysate products as a function of time and reactor operating parameters (e.g., temperature, residence time, and caustic concentration) to determine if any energetic materials are generated as by-products (e.g., picric acid) to evaluate the stability of hydrolysate for post-treatment processing to determine final product characteristics that influence post-treatment technology (e.g., pH, solid content, particle size, and homogenization) At the conclusion of each test run, the interior surfaces of the reactor will be visually inspected and samples of residue, if present, will be collected and characterized. A detailed preventive maintenance program will be developed to minimize the possibility of incidents during the cleanup of accumulated precipitates. Materials of construction will be investigated for alkaline and acid resistance. The rate of buildup of potentially energetic by-product salts will be assessed. The type and frequency of maintenance will be determined. Bench-scale Tests at Los Alamos National Laboratory Researchers at LANL are characterizing the hydrolysis reactions at bench scale by analyzing the hydrolysate and gaseous effluents generated by the processing of energetic materials. The concentration of caustic will be varied (12, 20, 25, and 30 weight percent) in these tests to determine its influence on destruction efficiency and residence time. LANL will also investigate the feasibility of hydrolyzing mixtures of energetics, the effects of particle size on reaction rates, the formation and growth of crystals in the hydrolysate, and the feasibility of mixing various energetic hydrolysates (Bonnett, 2000). These data will be used in developing requirements for size-reduction equipment and the methodologies for handling incoming energetic materials during full-scale processing. LANL and NSWC are also investigating the feasibility of
OCR for page 21
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Pueblo Chemical Depot processing mixtures of energetic materials found in a single type of munition: M1 propellant/tetrytol (105-mm M2 cartridge) M8 propellant/tetryl (4.2-inch, M2 cartridge) M28 propellant/Composition B4 (115-mm, M55 rocket) The actual ratios of propellant to burster explosive in the munitions will be used. Hydrolysate products are also being analyzed for picrate, as recommended in the ACW I Committee report (NRC, 1999). LANL has hydrolyzed the following combinations with no major perturbations: tetrytol (TNT and tetryl); cyclotol (RDX and TNT); octol (HMX and TNT); and nitrocellulose, nitroglycerin, nitroguanidine, triple-base propellant, and HMX (Bishop, 2000). Processing perturbations such as foaming were managed and controlled using well-known engineering techniques. Bench-scale Tests at the Pantex Plant Hydrolysis experiments at Pantex have shown that cyclotol (70 percent RDX and 30 percent TNT) and tetrytol (70 percent tetryl and 30 percent TNT) reacted within 1 hour and 3 hours, respectively, in 6 to 12 percent caustic. The metric for the completion of the reaction is the disappearance of solid material in the reactor (Belcher, 2000). The reaction time for the tetrytol was probably less than 3 hours, because the functional groups in the tetryl molecule, which are similar to those in TNT and RDX, should react at similar rates. However, only a lower bound on the rate of tetrytol destruction is available, because no observations were made before 3 hours had elapsed. Observations made after 1 hour for cyclotol showed that all solids had been consumed. Bench-scale Tests at the Naval Surface Warfare Center The NSWC will conduct calorimetric studies to determine the heat of reaction for hydrolysis reactions with various concentrations of caustic. This information will be used to develop strategies for reaction controls and to prevent runaways and upsets (Bonnett, 2000). The following reaction parameters are being determined for each energetic material by accelerating rate calorimetry: temperature of the maximum self-heating rate dependence of reaction rate on pressure and temperature rate of pressure and temperature increase heat of reaction moles of gas evolved per unit mass of energetic material activation energy of the reaction reactor cooling requirements This information will be useful for numerical modeling and simulation of the hydrolysis reaction process. Hydrolysate Production at the Radford Army Ammunition Plant Prior to Demonstration I, some attempts had been made to hydrolyze energetic materials on a large scale. RAAP (along with the Pantex Plant) produced the hydrolysates used during Demonstration I and the EDS tests in the ACWA program. RAAP was also expected to produce hydrolysate from M28 surrogate for the EDS program. RAAP had manufactured M28 surrogate propellant specifically for the preparation of hydrolysate. For environmental reasons, the surrogate did not contain lead stearate, which is normally included in M28 propellant as a burn-rate modifier. The propellant was prepared in grains in the shape of right circular cylinders, 1/16 inch in diameter by 1/16 inch long. Some of the problems that might be encountered in a large-scale operation were illustrated by a recent upset at RAAP. On October 14, 2000, hydrolysate from M28 surrogate propellant was being prepared when the piping of the recirculation loop ruptured, causing significant damage to the equipment (described later in this chapter). PROGRAM STATUS Results of Tests at the Holston Army Ammunition Plant Because the start of the EDS test program on energetics hydrolysis was delayed, the testing at HAAP had generated only limited data at the time this report was prepared. Energetic materials representative of the materials in the Pueblo stockpile had not yet been tested in the full-scale reactor at HAAP. As of January 1, 2001, only two Composition B hydrolysis runs had been completed. In one run, 200 lb of Composition B were hydrolyzed, and in the other, 500 lb were hydrolyzed. A detailed analysis of the hydrolysate composition as a function of time during these runs had not been completed by February 1, 2001; however, useful information was generated about the systemization (preoperational testing) of a full-scale hydrolysis reactor. At this point, the committee can comment only on the test plan and the preliminary results from these two runs. The committee believes that the test plan is well designed to determine acceptable parameters on the full-scale reactor at HAAP. Systemization of the 2,000-gallon hydrolysis reactor at HAAP was completed within 4 months. Composition B, which is produced at HAAP and is readily available, was chosen for the initial experiments. The disposal of hydrolysate is covered under existing permits for handling waste from the production of Composition B. Because foaming is difficult to control in Composition B, the two test runs with this material provided a good test of the efficacy of measures designed to control foaming. The problems that occurred were typical of any start-up operation. For example, the Acrison feeder, which is used to
OCR for page 22
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Pueblo Chemical Depot deliver Composition B to the reactor, failed to operate because of a software error when it was loaded with 500 lb of material. Water inadvertently added to the reactor overflowed into the dump tank and the secondary containment. The mass flowmeter in the recirculation loop malfunctioned. Despite these problems, feed rates of more than 490 lb/hr were maintained during the addition phase of the run, and the exothermicity was controllable. No components of Composition B (RDX and TNT) were detected at the end of the run in samples taken from the recirculation loop. Although complete characterization is still required, preliminary results to date have indicated that caustic hydrolysis is an acceptable technology for destroying energetic materials in a full-scale operation. The energetics hydrolysis reactor at HAAP is larger than the reactors in either EDP for Pueblo. The proposed systems will be suitable for the small quantity of energetic materials to be treated at Pueblo, but conclusions based on the HAAP experience (a 2,000-gallon reactor) must be reconsidered to account for the intended feed rate and energetic materials loading in pounds/gallon/hour for the 200-gallon reactor at Pueblo. Results of Tests at Los Alamos National Laboratory Characterizing the progress of the reaction is a prerequisite for designing a process for destruction of energetic materials by means of hydrolysis. The progress of the reaction can be followed by observing the disappearance of the initial feed. The studies at LANL are designed to assess the risk of untreated energetic materials in solution after hydrolysis has been completed. The solubilities in water of RDX and TNT at 60°C (140°F) have been reported in the literature to be approximately 300 ppm and 600 ppm, respectively (Gibbs and Popolato, 1980). LANL has determined the solubility of HMX at 90°C (194°F) to be approximately 150 ppm (Bishop, 2000). Because any dissolved energetic material is rapidly hydrolyzed (RDX half-life <1 s, HMX half-life of 0.92 min at 25°C (72°F) and 1.5 M NaOH), the absence of any solids is a valid indicator that all of the original energetic material has dissolved and reacted. The solubilities of the hydrolysate products and their cumulative effects are being investigated at LANL. These findings will guide the pilot-plant and full-scale demonstration runs. Test results at LANL indicate that reliable mass balances for the hydrolysis reactions of aromatic nitro compounds, such as TNT and tetryl, are difficult to obtain. A significant fraction of the carbon in the hydrolysis of TNT and tetryl ends up as a mixture of high molecular weight compounds (mol. wt.=3,000 to 30,000) that do not appear to be energetic (as evaluated by a simple hammer test) but may present other problems in subsequent processing steps. For example, because they are viscous, they might clog pumps or pipes. Also, they might separate from the aqueous phase of the hydrolysate and affect the reaction time for complete hydrolysis. The researchers at LANL recognize the concerns of the ACW I Committee about the formation of picric acid in the presence of lead, particularly the lead stearate in the M28 propellant, which could lead to the formation of lead picrate, a very sensitive explosive (Bishop, 2000). LANL characterized TNT and tetrytol hydrolysate for picric acid by high-pressure liquid chromatography and differential scanning calorimetry and found no evidence of picric acid or any other known energetic material (Bishop, 2001). Although the primary concern was the lead stearate in M28 propellant, General Atomics proposes hydrolyzing other lead-containing components, such as lead azide in fuzes, simultaneously with M28 propellant in the ERH. (In the Parsons/Honeywell design, fuzes would not be treated in the hydrolysis reactors but would be heated in an energetics rotary deactivator [ERD] until they deflagrate or detonate [see Chapter 4].) The LANL researchers believe that the lead stearate will likely form products such as lead hydroxide, which is more insoluble than lead picrate (Bishop, 2001). Based on the studies at LANL, the committee believes secondary treatment of the hydrolysate will have to treat any insoluble products that may form during hydrolysis or subsequent processing of the waste streams from the demilitarization plant. The toxicity of products formed during the entire processing cycle of the waste stream must be considered. The EDS hydrolysis testing at LANL is still under way. The researchers will continue to investigate the hydrolysis of all of the energetics for the neutralization of the assembled chemical weapons stockpile. They also plan an online analysis of the gas-phase products of hydrolysis, an analysis of the solid residue (including testing for the presence of energetic materials), and an analysis of the gases evolved from the hydrolysate during storage. Further studies of pH neutralization and an analysis of lead products may also be undertaken. The results will add to the technology base for the hydrolysis of energetic materials. Results of Tests at the Naval Surface Warfare Center The experimental database generated thus far includes completed accelerating rate calorimetry runs of neat energetic materials and runs of a mixture of tetrytol in a 30 weight percent solution of NaOH (Bishop, 2000). The results of these experiments are intended to provide a better understanding of the controls necessary to prevent runaway reactions. It would be premature for the committee to draw conclusions based on the limited results available so far. Analysis of an Incident at the Radford Army Ammunition Plant The M28 surrogate was hydrolyzed in an existing tank that had previously been used only for the preparation and
OCR for page 23
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Pueblo Chemical Depot storage of aqueous NaOH solution. Although the tank had an agitation mechanism, it was suspected to be inadequate for complete and efficient hydrolysis. Additional agitation was provided by an external recirculation loop that accepted material from the bottom of the tank and returned it to the top. The external recirculation loop had steam trace heating and two parallel pathways fitted with closed impeller pumps that were not designed to handle slurries. In preparation for the hydrolysis, the contents of the tank were brought to the proper NaOH concentration (12 weight percent) by diluting the appropriate amount of 50 percent caustic. The temperature was adjusted to 91°C (195°F) using internal steam (40 psig) coils. This temperature was maintained overnight. On October 13, particles of M28 were added incrementally over several hours through a hopper that directed them onto a slanted screen, where they were supposed to remain until they were digested. However, the agitation suspended the propellant grains in the caustic solution, and the particles later settled on the bottom of the tank at the intake for the recirculation loop. At 8:15 A.M. on October 14, 2000, one of the two pumps in the recirculation loop was turned on, but it stalled and had to be shut down. This pump had frequently stalled even when only caustic NaOH solution was present, so nothing seemed unusual. At 8:25 A.M., the other pump was started; at 9:50 A.M., it was found to have been pushed apart from overpressurization resulting from a blockage. Both valves on the recirculation line, one near the bottom of the tank and one at the pump suction, were then closed off, effectively stopping the recirculation. The steam trace line heating was continued. At 11:05 and 11:30 A.M., the recirculation loop piping was overpressurized, thereby causing the pipe to split and the flanges to separate. Partially digested propellant spilled onto the floor, but the propellant did not ignite. The system was cooled; contents of the tank were left in place, pending a decision on how to remove them safely. The chain of events leading up to the pressure rupture of the pipes appears to have started with the ingestion of partially decomposed propellant grains into the recirculation line. The recirculation pumps became clogged with the grains, and when the valves were closed at the intake and pump ends of the lower leg of the recirculation loop, the mixture of caustic and propellant was trapped in the piping and there was no way for the gases generated by the decomposition of the propellant to escape. The reaction was sustained by the combination of heat generated by the decomposition and heat supplied by the steam lines. The buildup of pressure eventually led to separation of the flanges and rupture of the piping. This is an important illustration of the potential hazards of reactions with energetic materials. These hazards must be controlled and accident scenarios must be considered. Some of the lessons learned from the RAAP that are germane to the HAAP experiments include pump design, recirculation intake location, screen mesh size for controlling migration of undesired energetics, and heat transfer. Pumps need to have an open design capable of handling slurries. Any recirculation loop needs to be located away from the bottom of the reactor vessel (or any other dead zone) so that undesired chunks and precipitates are not entrained through small-flow geometries. Screen mesh size and screen location need to be chosen carefully to ensure that chunks are controlled and the screen is kept open enough to maintain flow. Heat transfer data (i.e., inputs and outputs) need to be integrated into a control system that links process perturbations with variables subject to manipulation. Sensors should be identified that can confirm the following: the flow in narrow geometries (such as a nonintrusive flowmeter on the recirculation loop) the rate of temperature rise in locations that correspond to anticipated hot spot formation (such as thermocouples located at or close to the minimal clearance points between the wall and agitator) the agitator speed in the reactor (i.e., a tachometer) Process control setpoints and limits for acceptable operation should then be established, along with control algorithms that implement corrective action if process upsets or perturbations are detected. The severity of the incident might have been mitigated if consideration had been given to the reaction that was taking place between the propellant and the caustic. Failure to stop the steam trace heating on the recirculation loop helped to sustain the temperature needed for the reaction to continue, and closing the valves at both ends of the segment of the loop below the tank ensured that the gases produced would build up pressure. Although the incident at RAAP is unlikely to occur at Holston because the intake position and the type of pumps used are different, blockage could be caused by something else. For example, one of the compounds to be hydrolyzed at Holston is TNT, which has a melting point of 82°C (180°F). The intended reactor temperature for the hydrolysis experiments is between 85°C (185°F) and 95°C (203°F). Because the recirculation loop is neither insulated nor heated in any way, TNT might cool and crystallize in the piping when the reactor is being run at the lower temperature. None of the munitions at Pueblo contains M28 propellant; therefore, this incident has no direct bearing on planned disposal activities at Pueblo. However, the incident does show that even though the maximum credible event may not result from every process upset, sound engineering practices must always be used. This will reduce the likelihood of an accident and mitigate the consequences of accidents that do occur. The incident also highlights the need for training personnel involved in such operations to become aware of all possible hazards.
OCR for page 24
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Pueblo Chemical Depot SUMMARY ASSESSMENT The EDS energetics hydrolysis test program was instituted to address the ACW I Committee’s findings and recommendations on the hydrolysis of energetic materials. The test program was delayed and had not progressed very far by the time this report was submitted for review. Consequently, only limited information was available. The plan for the program, however, addresses all of the findings and recommendations of the ACW I Committee (NRC, 1999, 2000) with the exception of the hydrolysis of energetics contaminated by agent. The data from pilot-scale hydrolysis of contaminated energetics confirm that common energetic materials (e.g., nitro compounds, nitramines, and nitrate esters) can be effectively and safely converted to nonenergetic products by the action of caustic at elevated temperature (Bonnett, 2001). However, the colocation of operations involving chemical agents and explosives increases the safety concerns raised by any operation involving energetics. Under normal circumstances, process upsets and deviations from standard operating procedures occasionally lead to unplanned explosions or detonations. For this reason, stringent regulations have been promulgated to isolate other operations from explosives-handling operations. The value of these safeguards was apparent in the incident at RAAP. Similarly, consideration should be given to the proximity of operations involving agent and operations involving energetics in facility designs. Even as benign an event as a leak or spill of caustic solution charged with energetics could seriously disrupt the ingress and egress of personnel at nearby facilities that have nothing to do with the energetics operation until the spill has been cleaned up and an incident investigation completed. Process upsets and unplanned events that can be tolerated in a nonchemical-weapons environment are not acceptable in a chemical-weapons destruction facility. In a facility for demilitarization of assembled chemical weapons, otherwise minor upsets cannot be tolerated. Therefore, a thorough understanding of all aspects of the energetics hydrolysis process will be essential. Experience thus far with tests at HAAP suggests that the concerns involving the immaturity of the hydrolysis process for energetics, which were identified in the findings and recommendations of both ACW I Committee reports, have not been fully addressed. The ACW II Committee’s concern is focused on the possible impact of process upsets on the facility, rather than on the adequacy of the hydrolysis process to neutralize energetic materials. The completion of the EDS test program is expected to provide a much more complete understanding of the hydrolysis process, control systems, maintenance requirements, and other considerations necessary for determining the applicability of this technology to assembled chemical weapon demilitarization. However, EDS experiments being conducted with tetrytol, which contains TNT, will not be representative of the hydrolysis of neat tetryl. The TNT, which is molten at the reaction temperature, will change the physical state of the tetryl when tetrytol is the feedstock. The relationship between the rate of destruction of tetryl and the granulation of the feedstock material from neat tetryl boosters has not yet been established. The ACW II Committee will continue to monitor the evolving test data and results from the various locations conducting ACWA EDS tests on energetics hydrolysis, including such issues as the characterization and treatment of off-gases. Further evaluation of this developing information will be made by the committee in a forthcoming report (to be published in 2002) on the EDS program for proposed alternative technologies for the Blue Grass site. Previous Findings and Recommendations of the ACW I Committee In this section, the findings and recommendations on energetic hydrolysis from the two ACW I Committee reports are reviewed to determine the extent to which they are still valid as a result of EDS testing (NRC, 1999, 2000). Review of Findings and Recommendations from the 1999 Initial ACW I Committee Report Finding GA-2. Hydrolysis of energetics at the scales proposed by the technology provider is a relatively new operation. Chemically, it is possible to hydrolyze all of the energetic materials; however, the rate of hydrolysis is limited by the surface area and, therefore, depends on particle size. (Smaller particles are more desirable because they have a higher surface-to-volume ratio.) The proposed method of removing and hydrolyzing the energetics appears to be reasonable, but further testing is required to determine the hydrolysis rates and to confirm that throughput rates can are achieved. This finding is being addressed by the ACWA EDS program. General Finding 2. The technology base for the hydrolysis of energetic materials is not as mature as it is for chemical agents. Chemical methods of destroying energetics have only been considered recently. Therefore, there has been relatively little experience with the alkaline decomposition of ACWA-specific energetic materials (compared to experience with chemical agents). The following significant issues should be resolved to reduce uncertainties about the effectiveness and safety of using hydrolysis operations for destroying energetic materials: the particle size reduction of energetics that must be achieved for proper operation the solubility of energetics in specific alkaline solutions process design of the unit operation and the identification of processing parameters (such as the degree of
OCR for page 25
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Pueblo Chemical Depot agitation and reactor residence time) necessary for complete hydrolysis the characterization of actual products and by-products of hydrolysis as a function of the extent of reaction the selection of chemical sensors and process control strategies to ensure that the unit operation following hydrolysis can accept the products of hydrolysis development of a preventative maintenance program that minimizes the possibility of incidents during the cleanup of accumulated precipitates. This finding is being addressed by the ACWA EDS program. The assessment of the particle size reduction needed for proper operation was limited to a single particle size, which was acceptable for the operation. No attempt to identify an optimum particle size is included in the program. General Finding 3. The conditions under which aromatic nitro compounds, such as trinitrotoluene (TNT) or picric acid, will emulsify in the aqueous phase and not be completely hydrolyzed are not well understood. Therefore, this type of material could be present in the output stream from an energetic hydrolysis step. Complete destruction has now been demonstrated. General Finding 4. The products of hydrolysis of some energetic materials have not been characterized well enough to support simultaneous hydrolysis of different kinds of energetic materials in the same batch reactor. This finding is being addressed by the ACWA EDS program. General Recommendation 5. Whatever unit operation immediately follows the hydrolysis of energetic materials should be designed to accept emulsified aromatic nitro compounds, such as TNT or picric acid, as contaminants in the aqueous feed stream (see General Finding 3). This recommendation is still valid until the EDS testing program is completed and the results indicate otherwise. General Recommendation 6. Simultaneous processing of different types of energetic materials should not be performed until there is substantial evidence that the intermediates formed from the hydrolysis of aromatic nitro compounds will not combine with M28 propellant additives or ordnance fuze components to form extremely sensitive explosives, such as lead picrate (see General Finding 4). This recommendation has been addressed to the extent that the LANL study shows the absence of picrate in hydrolysis products. As yet, mixtures of energetics have not been effectively addressed in the EDS test program, although such tests are planned. Until those tests are completed, mixtures of energetic materials should not be hydrolyzed on a large scale. Review of Findings from the ACW I Committee Supplemental Report (NRC, 2000) Finding (Demo I) GA-1. Testing on the hydrolysis of energetic materials contaminated with agent will be necessary before a full-scale system is built and operated. This finding is not being addressed by the ACWA EDS program. The committee notes that integration concerns such as this should be addressed as soon as practicable to minimize delays during systemization of the disposal facility (see General Finding [Pueblo] 3 in Chapter 5). New Findings and Recommendations Finding (Pueblo) EH-1. Alkaline hydrolysis can be an effective and safe method for destroying energetic materials at Pueblo Chemical Depot. There appear to be no insurmountable obstacles to using this technology to destroy the energetics in assembled chemical weapons. Finding (Pueblo) EH-2. Results from the energetics hydrolysis test program thus far have shown that hydrolysis rates are consistent with the proposed designs for overall throughput rates necessary to meet the current disposal schedule for the Pueblo stockpile. Finding (Pueblo) EH-3. The hydrolysis of neat tetryl from burster charges is not being tested. Tests with tetrytol, which contains TNT, will not be representative of the hydrolysis of neat tetryl. Finding (Pueblo) EH-4. Although the EDS energetics hydrolysis test program addresses many of the issues related to effective destruction of energetic materials from assembled chemical weapons, the tests are based on a predetermined granulation of the feedstock and will not provide information for determining the optimum granule size for disposal operations at Pueblo. The tests will provide information for only one granulation size and will not show the relationship between destruction rate or efficiency and particle size. Finding (Pueblo) EH-5. Mass balances for most of the data from bench-scale hydrolysis experiments on aromatic nitro compounds are incomplete, mainly because of the formation of ill-defined, high-molecular-weight organic compounds. A thorough understanding and more complete characterization of the products of the hydrolysis of TNT and tetryl is still lacking. The complexity of the intermediates may preclude any more exact identification than one based on elemental analysis and functional group identification. Finding (Pueblo) EH-6. Process parameters and process control strategies (e.g., energetic feed rates, caustic concen-
OCR for page 26
Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Pueblo Chemical Depot tration, and reactor temperature) have not yet been characterized in enough detail to ensure a smooth transition to full-scale operation. Recommendation (Pueblo) EH-1. A bench-scale comparison of the rates of hydrolysis of tetryl and tetrytol should be undertaken before any process for the destruction of tetryl is planned. The rates should not be based only on data from tests with tetrytol. Recommendation (Pueblo) EH-2. Any post-hydrolysis treatment technology selected for the Assembled Chemical Weapons Assessment program must be capable of accommodating the possible presence of high-molecular-weight organic compounds generated from aromatic nitro compounds. Recommendation (Pueblo) EH-3. Insofar as possible, the particle size, feed rate, nature of the feedstock (e.g., dry or slurried), and solids loading in the reactor at Holston should be matched with the operating conditions expected at Pueblo to verify the efficacy and safety of the hydrolysis process for energetic materials. Recommendation (Pueblo) EH-4. Hydrolysis reactions at pilot scale and full scale must be remotely operated.
Representative terms from entire chapter: