Analysis of Engineering Design Studies for Demilitarization of Assembled Chemical Weapons at Blue Grass Army Depot (2002)

Both the General Atomics and the Eco Logic technology packages use caustic (base) hydrolysis as the initial treatment to destroy the energetic materials recovered from chemical weapons (see Table 2-1). Secondary treatment of the resulting hydrolysate is accomplished by supercritical water oxidation (SCWO). The chemistry of caustic hydrolysis takes advantage of the susceptibility of the functional groups commonly found in energetic materials to attack by hydroxide ion, yielding 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 water or organic solvents.

An understanding of the chemistry 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 both the chemical agents and the energetic materials and the prevention of process upsets throughout the facility.

In its 1999 report, the ACW I Committee expressed concern that the technology base for the hydrolysis of energetic materials was not as mature as that for the hydrolysis of chemical agents, that chemical methods for destroying energetics had only recently been considered, and that the chemistry of the process was not well understood (NRC, 1999). Among the specific issues to be addressed were these:

• the particle size reduction that must be achieved for proper operation

• the solubility of energetics in specific alkaline solutions

• the process design of the unit operation and the identification of processing parameters (such as the degree of 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

Similar concerns were reiterated by the ACW I Committee in a supplemental report evaluating the results of testing conducted during the Demonstration I phase of the ACWA program (NRC, 2000a). In response to these concerns, PMACWA initiated a multi-laboratory test program to identify the chemical and engineering parameters for the efficient, safe alkaline hydrolysis of the energetic materials found in assembled chemical weapons. The project plan and preliminary results of this program were reported in Chapter 2 of the ACW II Committee’s 2001 report on the EDS I phase of the ACWA program for an alternative technology facility at Pueblo Chemical Depot (NRC, 2001a). In this chapter, the ACW II Committee briefly describes the hydrolysis treatment process for energetic materials, reviews PMACWA’s program for engineering design testing of energetics hydrolysis, and assesses the results of the test program.

In response to the challenges listed above, PMACWA devised a test plan to address these issues. This EDS testing began in December 2000, and the tests were completed at the end of March 2001. The U.S. Army Tank-Automotive and Armaments Command (TACOM) Armament Research, Development and Engineering Center (ARDEC) at Picatinny Arsenal, New Jersey, was responsible for coordinating the program.

The Picatinny test plan requirements document described objectives, planned testing, and team member responsibilities in the EDS test program for hydrolysis of energetic materials (Bonnett, 2000). However, the objectives evolved during the execution of the program, and the final retrospective statement of the objectives, dated June 18, 2001 (after the completion of the testing), is given in Appendix N of the final report on the EDS testing program for energetics hydrolysis (Bonnett and Elmasri, 2001). As stated in Appendix N, the objectives were as follows:

• address concerns identified by the National Research Council (NRC) and processing issues that surfaced at Radford AAP [Army Ammunition Plant] and PANTEX, Inc., during the manufacturing of the various hydrolysates used to support the previous demonstration testing;

• determine the optimum process operating parameters to support scale-up of the hydrolysis process and the definitization of the Engineering Design Package (EDP) for the pilot phase for the Pueblo Chemical Agent Disposal Facility scheduled in August of 2001 and for the Blue Grass Chemical Agent Disposal Facility scheduled in August 2002;

• define a hydrolysis process that is safe and environmentally compliant, and that will efficiently produce hydrolysates of energetic materials recovered from the various chemical munitions during the disassembly process.

• produce hydrolysates that will be ready for post-treatment processing using such technologies as SCWO, bio-reactor, etc.

The participants in the program and their principal responsibilities were as follows:

• TACOM-ARDEC was responsible for program management and incorporation of bench-scale results into processes at Holston Army Ammunition Plant (HAAP) to demonstrate hydrolysis operations.

• HAAP and BAE Systems were responsible for design, installation and operation of a hydrolysis reactor system. The design of the hydrolysis system was done by BAE Systems, in consultation with TACOM-ARDEC.

• Los Alamos National Laboratory (LANL) performed bench-scale testing in support of the engineering-scale hydrolysis reactor system.

The reactor used for the tests at HAAP has the following specifications (Bonnett and Elmasri, 2001):

• a 2,000-gallon, glass-lined, jacketed reactor vessel

• a recirculation loop, unheated and uninsulated, that reenters the reactor vessel at the bottom

• a dual-flight, Hastelloy C, variable-speed agitator in the reactor vessel

• a condenser/scrubber for offgases

• meters to monitor pH, temperature, and flow

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 was fed into the reactor as a dry solid and screened through a 1.0-inch × 1.0-inch mesh as it was added to the feeder hopper to prevent large particles from entering the reactor.

Six different energetic materials and one mixture of materials were processed in the study (Bonnett and Elmasri, 2001):

• Composition B

• Composition B-4

• tetrytol

• M1 propellant

• M28 propellant in granular form

• M8 sheet propellant

• M28 propellant/Composition B-4 (86/14 weight-percent ratio, based on amounts in an M55 rocket)

Runs 1 through 5, denoted “commissioning runs,” were used to verify that the systems for operating the reactor and for data collection during the subsequent energetics hydrolysis tests were functioning as intended. A summary of the conditions used in the commissioning runs is given in Table 2-2. The feed used for these tests was Composition B, a material that is readily available at HAAP and very similar in composition to Composition B-4 (see Table 2-1). The working volume in the reactor for all the commissioning runs was about 1,700 gallons, and the reaction temperature was maintained at 87°C for the duration of the hydrolysis. The feed was added continuously through an Acrison loss-in-weight feeder.

The test matrix for the energetics hydrolysis tests in runs 6 through 21 is given in Table 2-3 (Bonnett and Elmasri, 2001). For these runs, the reactor operated at 87°C with 700 gallons of caustic. Each run lasted 12 hours, comprising a 4-hour addition phase (only 3 hours for tetrytol) followed by an 8-hour postaddition processing phase. Throughout the duration of the runs, airflow at a rate of 40 standard cubic feet per minute (scfm) was maintained through the reactor headspace. No material balance calculations were performed. For reasons discussed below, run 14 was extended by 12 hours to give a total run duration of 24 hours. The tests involving Composition B-4, M1 propellant, M28 propellant, and the mixture of Composition B-4 and M28 propellants all used a 50-lb/hr feed rate during the first hour, 100 lb/hr during the second hour, 150 lb/hr during the third hour, and 200 lb/hr during the fourth hour, for a total of 500 pounds of energetic material per run. However, for the tetrytol runs, the addition phase lasted only 3 hours and the total mass was only 350 pounds per run, because of the limited availability of tetrytol. The feed rate profile for tetrytol was 50 lb/hr during the first hour, 100 lb/hr during the second hour, and 200 lb/hr during the third hour.

The progress of the hydrolysis reaction was monitored by periodic sampling and analysis of the hydrolysate and by continuous monitoring of the gases evolved. Sixteen 42-ml samples of the hydrolysate were collected during each run. The sampling system attached to the recirculation loop comprised an insulated cabinet with an ice or dry ice bath in which the

sample bottles were chilled so that the hydrolysate samples would be cooled as soon as the samples were collected. Each sample bottle was charged with 30 ml of 6 N sulfuric acid prior to the start of the run. The combination of the cooling and the neutralization of the caustic by the sulfuric acid quenched any further reaction. At the end of each run, a 4- or 8-L sample of the hydrolysate was collected from the recirculation loop.

Evolved gases were monitored continuously for O₂, CO₂, NO_x, CO, and total hydrocarbons (THC) by a continuous emission monitoring system (CEMS). Batch samples of the offgases were also taken and analyzed for NH₃, HCN, residual energetic materials, volatile organic compounds, and N₂. The batch samples were collected over the entire duration of each run, and the single value reported represents the average concentration for the whole run.

At the conclusion of each test run, the interior surfaces of the reactor were visually inspected and samples of residue collected and analyzed.

Researchers at Los Alamos National Laboratory (LANL) studied the hydrolysis reactions at bench scale by analyzing the hydrolysate and gaseous effluents generated by processing of the following energetic materials: Composition B-4 tetrytol, M1 propellant, M8 propellant, and M28 propellant, and mixtures of M28 propellant with Composition B-4 and M28 propellant with tetrytol. Two concentrations, 12 and 20 percent, of caustic were used in the experiments for all materials except the M8 propellant, which was studied using only 20 percent caustic.

The experiments were carried out in a 2-L Parr reactor fitted with an internal cooling coil to cool the reaction mixture if necessary and to prevent runaway reactions. The temperature was controlled by a proportional-integral-differential (PID) controller. All experiments started at ambient temperature, which was raised to the desired reaction temperature, 93°C, in a nearly linear ramp over about 15 minutes for the propellants or 25 and 35 minutes for Composition B-4 and tetrytol, respectively.

The Composition B-4 was provided to LANL in the form of large flakes and was used as received. The tetrytol was in large to medium-size chunks. The M1 propellant and the M28 propellant were provided and used as small grains. The M8 propellant was provided in stacks of three sheets weighing 30-35 g each and sewn together with cotton thread. The sheets were not separated before being manually fed into the reactor.

Gas evolution was monitored by online mass spectrometry. Liquid samples were removed during the experimental runs through a sampling port in the reactor. The samples were immediately diluted by a factor of 10 with cold water to quench the reaction and were kept refrigerated until they were analyzed. Ion chromatography was used to determine the concentrations of nitrite, nitrate, formate, and acetate. Total organic carbon, total inorganic carbon (TOC/TIC), and ammonia, were determined. Differential scanning calorimetry was used to determine whether there were any products in the hydrolysate liquid or the solid residue at the end of the reaction that would decompose exothermically. Hydrolysates from Composition B-4 and tetrytol were also subjected to high-performance liquid chromatography (HPLC) analysis to determine whether any of the starting energetic compounds remained at the end of the hydrolysis.

Sixteen hydrolysis runs were successfully completed at HAAP in the course of this study (see Table 2-3). For every energetic material subjected to hydrolysis by itself (i.e., not mixed with another energetic material), the energetic components were reduced to products that no longer posed an explosion hazard. Although the operational parameters, such as base concentration and agitation rate, were varied only minimally, the tests demonstrated that the destruction of the energetic materials could be carried out at a rate consistent with that of other unit operations in the applicable technology packages and under conditions that permitted control of the rate of heat release of the hydrolysis reaction.

One of the major concerns in processing energetics when an exothermic reaction occurs is the reliable control of temperature to prevent runaway reactions. Under the conditions employed at HAAP, the temperature of the reactor varied by only a few degrees from the 87°C set point for all the energetic materials processed. The formation of foam on the surface of the caustic solution during the processing of Composition B-4 was initially a concern, because the foam could insulate the surface of the liquid, preventing heat from escaping and possibly leading to uncontrollable self-heating. However, the spray system used to break up the foam was effective, and the processing of Composition B-4 proceeded uneventfully.

The feeding of dry energetic material into the reactor involved the greatest hazard in the whole process. Standard precautions employed in the handling of dry energetic powders (grounding and bonding of metal surfaces, minimization of stresses exerted on the energetic material) were followed. No incidents occurred during the handling of any of the dry energetics. It should be noted that none of the technology packages using hydrolysis for the destruction of energetics includes handling them as loose dry powders. Instead they are handled as slurries.

Samples of the gaseous reaction products were taken through a sampling line upstream of the condenser and scrubber through which the offgases were treated prior to release to the atmosphere. The major gaseous products found during all the hydrolysis runs were NH₃, N₂O, CO, and hydrocarbons, which were below flammable limits. Quantities of the gases differed depending on the material being hydrolyzed. The evolution of N₂O occurred primarily during the addition of the feed to the reactor and decreased during the postaddition reaction phase. Ammonia evolved throughout the duration of the run, with the amount evolved increasing after the completion of the addition phase.

The hydrolysis of Composition B-4 produced N₂O and CO rapidly upon addition of the energetic feed. For all three runs involving Composition B-4 (runs 6, 7, and 8), the maximum concentration of CO in the offgas was 400-450 ppm, peaking near the end of the addition phase. By the end of the runs, the concentration dropped below 50 ppm for runs 6 and 7 and below 100 ppm for run 8. The evolution of N₂O paralleled the evolution of CO. It peaked at a concentration of about 20,000 to 25,000 ppm and dropped off to less than 5,000 ppm at the end of the run. Ammonia gas was measured only in the batch samples. The concentration ranged from 12.9 to 23.3 g/m³.

Small amounts of the energetic constituents of Composition B-4 were detected in the offgas stream during the addition phase. RDX and HMX (an impurity in RDX) were measured at a few hundreds and a few tens of micrograms per cubic meter, respectively. TNT, which is more volatile than RDX and HMX, was detected at a few milligrams per cubic meter. Minor amounts of other energetic impurities found in TNT, such as 2,4- and 2,6-dinitrotoluene, were also detected. This was attributed to the entrainment of some of the dust from the energetic material in the offgas stream. No energetic compounds were detected in the offgas at the end of the runs.

Three concentrations (11.3, 20, and 25 weight percent) of caustic were used in the hydrolysis runs of M1 propellant. In runs 9, 10, and 11, the THC level in the offgas continued to increase during the postaddition reaction phase, rising to 1,600-2,500 ppm by the end of the 12-hour runs. The runs at base concentrations of 20 and 25 percent showed lower hydrocarbon evolution than the one at 11.3 percent concentration. The runs in 20 and 25 percent caustic showed THC levels of 1,600 ppm in the offgas at the end of the run, whereas the run at 11.3 percent showed 2,500 ppm. Run 14, a replicate of run 10, was carried out for 24 hours instead of the standard 12 hours. It produced THC results very similar to those of run 10 for the first 12 hours. Near the end of the first 12 hours, the THC concentration began to decrease and had reached a level of about 800 ppm as the duration of the run approached 24 hours. Just before termination of the run, the agitation rate was increased and a sudden spike appeared in the THC level, reaching a maximum of about 2,200 ppm. The final report for the project states that these data supported the hypothesis that there is a significant amount of dissolved gas in the hydrolysate (Bonnett and Elmasri, 2001).

Several compounds were identified in the offgas stream from M1 propellant that probably arose from the solvents used in processing the propellant. Acetone (~50-90 ppm), ethyl ether (~30-130 ppm), and toluene (~30 ppm) were all detected in runs 9, 10, and 11. Cyclohexanone was reported for all these runs at 70-120 ppm. No batch analyses of the offgases from run 14 were reported.

There was also a dependence on caustic concentration in the evolution of N₂O in the runs. The runs at 20 and 25 percent caustic produced higher levels of N₂O during the addition phase, peaking at about 800-900 ppm, then tailing off to less than 100 ppm at the end of the run. For the run at 11.3 percent caustic, the level of N₂O did not exceed 500 ppm at any time and decreased slowly after the addition phase to about 100 ppm by the end of the run.

M28 propellant contains about 24 percent nitroglycerin. Nitroglycerin is a liquid at ambient temperature and pressure, with a significant vapor pressure at the reactor temperature. During the addition phase of the hydrolysis of M28 propellant, about 7,000 ppm of nitroglycerin was detected in the offgases, but shortly after completion of the addition phase, the level dropped to about 400 ppm and remained nearly constant for the duration of the run.

Acetone and toluene were detected in the offgases at levels comparable to those in the M1 propellant offgas. No diethyl ether was reported for M28 propellant. A few ppm of 2-butanone (methyl ethyl ketone), another common solvent, were detected.

Ammonia evolved over the course of the reaction, starting at an average level of about 100 mg/m³ during the addition phase and rising to about 1,000 mg/m³ during the postaddition reaction. Some N₂O was generated, although at levels an order of magnitude lower than for Composition B-4. Oxides of nitrogen were detected at a few parts per million during both the addition and reaction phases.

Tetrytol, like Composition B-4, contains a matrix of TNT in which a second ingredient is incorporated. In tetrytol, the second component is tetryl (2,4,6-trinitrophenylmethylnitramine). The volume and content of the offgases differ markedly from those of Composition B-4. Whereas Composition B-4 generates significant quantities of ammonia, tetrytol generates two to three orders of magnitude less ammonia. Composition B-4 produced N₂O in the range of 15,000 to 20,000 ppm during the addition phase, whereas tetrytol produced only one-tenth that amount. However, the THCs were 10 times higher for tetrytol than for Composition B-4. Neither explosive produced more than a few parts per million of NO_x.

As was the case with Composition B-4, some energetic materials were detected in the offgases during the addition phase, but none of these persisted throughout the run. No energetic materials were found in the offgases during the postaddition reaction phase.

The main constituents of the offgases from the M8 propellant hydrolysis were CO, N₂O, THCs, and NH₃. Both CO and N₂O were generated mainly during the addition phase, reaching levels of 150-250 and 4,000 ppm, respectively. The THCs rose slowly during the addition phase, reaching a peak about 1 or 2 hours after completion of the addition, then declining slightly from

a peak of about 700-800 ppm to an end-of-run level of a little over 500 ppm. The ammonia level averaged 500 ppm during the addition phase and then rose to an average level of about 1,800 ppm for the reaction phase.

The mixed M28 propellant and Composition B-4 produced offgases that were the combined results of the separate treatment of the two energetic materials. The evolution of N₂O was high during the addition phase (peak of ~12,000 ppm), as was characteristic of Composition B-4, and the evolution of THCs maintained a level of about 150 ppm during the entire postaddition reaction phase. Low levels of NO_x were observed. Ammonia was evolved at levels lower than those produced by Composition B-4 and higher than those produced by M28. The only oddity in the offgas data for the mixed M28 and Composition B-4 run was the presence of tetryl during the addition phase of the hydrolysis. This suggests that there was contamination at some point, possibly in the feed system, because tetrytol was the last energetic processed through the Acrison loss-in-weight feeder in the three runs just preceding the mixed M28 and Composition B-4 run.

Since the hydrolysis of energetic materials in the Eco Logic and General Atomics technology packages (see Chapters 4 and 5) produces hydrolysates that would be subjected to some subsequent treatment (e.g., SCWO), the main concern is the continued presence of an energetic species in the hydrolysate that might constitute an explosion hazard. The data collected from midrun samples are not reliable because the results are scattered, and it is not clear whether the suspended solids were included in the analysis. The most that can be determined from these data is that the concentration of energetic material is decreasing with time during the processing phase, as expected.

The end-of-run hydrolysate analyses provide more useful data on the presence of energetic species, because they are based on a larger homogeneous sample. The end-of-run hydrolysate analyses indicate that, in half of the cases, the level of energetic material had been reduced below the detection limit. The exceptions are listed in Table 2-4. None of these low levels of energetic material pose any explosion hazard, because even slurries with 10 to 30 percent of strong explosives

such as RDX or HMX are not detonable. The explosion hazard of the hydrolysates containing only a few parts per million of any of the compounds in Table 2-4 is nil (EDE, 2001a).

The main identifiable carbon-containing compounds in the hydrolysates of all the runs were acetate and formate salts. The fraction of the total organic carbon in the hydrolysate that was accounted for by these two species ranged from about 10 percent for tetrytol to about 40 percent for the M28 propellant. Thus, a considerable amount of the carbon in solution remains unidentified. Because the hydrolysate will undergo further processing, this is not necessarily a problem, provided the explosion hazard has been removed in the hydrolysis. The concentration of organic carbon in the hydrolysate is about 2 percent (16.375 g/L in run 12), and slurries of high explosives such as RDX are not detonable at concentrations below about 30 percent (see above), so the assumption that the hydrolysate would not be detonable even if all the organic carbon in the hydrolysate were present as an explosive remains valid.

The presence of cyanide in the hydrolysate does raise some concern. The levels of cyanide in the end-of-run analyses ranged from 29 mg/L (M1 propellant, 11.3 percent caustic) to 705 mg/L (tetrytol, 26 percent caustic). While the solution is alkaline, the cyanide will remain in solution as aqueous cyanide ion. However, if the solution were made neutral or acidic, which would reduce the hazards associated with handling caustic solutions, the cyanide would be converted to hydro-

cyanic acid (HCN), which would present a toxic gas hazard (DoD, 1999).

The hydrolysis tests at HAAP of M8 propellant proceeded as expected, with the typical gaseous and condensed products being formed as noted previously and with the M8 being completely consumed in the course of the hydrolysis. A problem was encountered, however, because the layered sheets of the propellant are sewn together with cotton thread. The cotton threads did not decompose or disintegrate and became entangled on the lower agitator and shaft. Thus, they could clog pumps, pipes, and valves. However, the committee expects that simple engineering solutions are available to prevent this from happening. For the tests at HAAP, the sheets were not treated in any way to disassemble them or reduce the size of the feed introduced into the reactor. Instead, the sheets were added to the caustic solution manually, bypassing the loss-in-weight feeder.

The studies at LANL were designed to provide complementary data on the hydrolysis reaction of energetic materials under similar conditions to those employed at HAAP. Gas production was monitored by mass spectroscopy over the course of the reaction, but only N₂, N₂O, NH₃, and NO were reported. No measurements were made of total hydrocarbons (THC) in the offgas.

Some experiments on the hydrolysis of Composition B-4 were reported prior to this program (see Appendix C in Bonnett and Elmasri, 2001). Composition B-4 at 10 percent concentration was treated with 3 M (12 weight percent) caustic at 90°C in a total solution of 50 ml. However, the temperature was not well controlled and oscillated over about 10°C with a periodicity of approximately 10 minutes. The results were consistent with the observations made at HAAP:

• The main organic products were acetate and formate, with formate being present at about 5 to 10 times the concentration of acetate.

• Most of the carbon in solution was in the form of organic compounds, not inorganic.

• The main nitrogen species in solution was nitrite, with a minor amount of nitrate.

The experiments undertaken at LANL were carried out in a 2-L Parr reactor charged with 1 kg of caustic solution and 10 percent energetics at 93°C (versus 87°C at HAAP). In contrast to the experiments at HAAP, the energetic material was introduced into the caustic at ambient temperature before heating. The heating from ambient to 93°C typically took about 40 minutes. The reaction temperature was maintained with vigorous stirring set at 720 rpm until the total reaction time reached 100 min.

Hydrolysis of Composition B-4 and tetrytol under these conditions using 12 or 20 percent caustic destroyed the energetic materials. Analysis of the Composition B-4 and tetrytol hydrolysate by high-performance liquid chromatography (HPLC) and differential scanning calorimetry did not detect picric acid or any other known energetic material. However, because detection limits for the method were not reported, it is not possible to determine whether amounts of picrate or other energetics below the detection limits might have been present in these small samples (Bonnett and Elmasri, 2001).

In the LANL experiments, all of the energetic material was introduced into the vessel at the beginning of the run, when the caustic was still at ambient temperature; in the HAAP runs, the caustic was heated to the reaction temperature before the energetic feed was introduced. Therefore, the rate of destruction from the LANL data is not comparable to that from the HAAP data. It appears that the evolution of gas commenced at about 65-70°C for both Composition B-4 in 20 percent caustic and tetrytol in 12 percent caustic. No temperature data were presented for Composition B-4 in 15 percent caustic, and tetrytol appeared to begin generating gas at a somewhat higher temperature (~80°C) in 20 percent caustic.

The hydrolyses of all the propellants tested (M1, M8, and M28) formed some solid residue, the analysis of which was not pursued. However, differential scanning calorimetry analysis showed no significant exotherms in the thermograms of the residues from any of the propellants.

The evolution of gas from the propellants began at a somewhat higher temperature than for Composition B-4 and tetrytol. The first observation of N₂O did not occur for M1 propellant until the reaction temperature reached about 90°C. The other propellants began evolving gas at around 80°C.

At the request of the PMACWA, LANL also looked at the degradation of the rayon cloth used to contain the M1 propellant charges. Swatches of the cloth were heated in caustic solutions (6, 12, and 20 weight percent NaOH) at 93°C for 340 minutes. Although the swatches were damaged by the treatment, they were not consumed nor did they disintegrate. The final weights were not determined, because some crystalline material adhered to the swatches and could not be removed by repeated rinsing with water. This result is consistent with the results at HAAP. However, the committee believes that simple engineering solutions are available to address the problem posed by the persistence of the rayon bags.

The possible formation of lead picrate when nitroaromatic compounds such as TNT and tetryl are hydrolyzed in the presence of lead has been an ongoing concern of the two NRC ACW committees (NRC, 2000a, 2001a, 2001b). The perceived problem is that solid lead picrate, precipitated from solution and dried, might detonate. For this to happen, the following conditions must be met: (1) both lead (II) ions and picrate ions must be present in the solution, (2) the lead picrate must precipitate because its solubility product is exceeded, and (3) the solid must be present in a condition that is detonable with a stimulus that can be achieved at some point during the processing cycle. The processing cycle includes the hydrolysis, storage, and subsequent treatment of the hydrolysate (if the solid is transferred along with the hydrolysate to the post-hydrolysis treatment step) and, when appropriate, cleanup of the hydrolysate reactor and system.

It has been shown that picrate ion is formed during the hydrolysis of tetrytol. As shown in Table 2-5, picrate was observed in the midrun analyses for all the hydrolysis runs of tetrytol and in the midrun analyses for the mixed M28/Composition B-4 run. For two of the three tetrytol runs, there was a low level of picrate remaining at the end of the run, but no picric acid was reported in any of the hydrolysate samples from any of the Composition B-4 runs that preceded the tetrytol runs. From the midrun concentrations of picric acid in the tetrytol runs, it appears that the formation of picrate decreases with increasing base concentration. This observation, and the nonzero values of picric acid in the end-of-run hydrolysate samples for runs 16 and 17, would lead one to expect to find some picrate remaining in the hydrolysate at the end of run 15. However, none was reported. The level of picric acid reported during the postaddition phase of run 15 varied between 10 and 30 mg/L, yet none is reported in the end-of-run analysis of the hydrolysate.

The presence of picrate in run 21 (M28/Composition B-4) might be attributable to contamination. The offgas analysis during the addition phase of the mixed energetic indicated some tetryl in the offgas. This may have been due to contamination of the feed system, since tetrytol was processed in the runs that used the loss-in-weight feeder immediately preceding the mixed energetic run. (The M8 sheet propellant was hydrolyzed in runs 18, 19, and 20, between the tetrytol and the mixed M28/Composition B runs, but the M8 sheets were fed into the reactor manually, bypassing the loss-in-weight feeder.) If some tetrytol was inadvertently introduced into the reaction vessel from a contaminated feeder, it might have led to the formation of a small amount of picrate. No picrate was reported in the end-of-run analysis of the hydrolysate from run 21.

The results from the work done at LANL (see below) failed to find any picrate in the hydrolysates from either Composition B-4 or tetrytol. However, because the detection limit for the method used was not reported, it is not possible to determine whether or not the LANL results are consistent with those reported in the Holston study (Bonnett and Elmasri, 2001).

The two studies, taken together, suggest that tetryl, not TNT, leads to the formation of picrate, since neither the LANL results nor the Holston results showed the formation of picrate in the absence of tetryl. If this is correct, the formation of lead picrate cannot occur when Composition B and M28 propellant are coprocessed, provided no tetryl contamination is present. The solubility of lead picrate in the reaction medium under consideration is not known. Because lead in solution at high pH is predominantly in the form of plumbite ion (HPbO₂^-), the concentration of free Pb²⁺ ions should be very low, and the solubility product of lead picrate would have to be exceedingly small for

any lead picrate to precipitate. Even if some lead picrate were to be formed, the risk of an explosion from initiation of the lead picrate would be small. If good engineering practices are followed and the hydrolysis reactor is designed to eliminate dead-legs in the piping and other places where solids might accumulate, any solid that is formed would be distributed through the reactor and would be dispersed in the surrounding aqueous phase. The main hazard due to the formation of any dry lead picrate would occur if the reactor were allowed to dry out (thus removing the water that desensitizes the lead picrate) or during disassembly of the apparatus.¹

Although precipitation of lead picrate is not likely to be a hazard during normal operation of a well-designed hydrolysis reactor, some data are still unavailable. During the hydrolysis runs at Holston, the material balance for lead was extremely poor. None of the runs of M1 propellant, each of which contained a little less than 4 pounds of lead in the feed, indicated any lead in the end-of-run hydrolysate analysis. Also, no lead was identified in the solids remaining at the end of any of the hydrolysis reactions. The inability to account for the lead dissolved or suspended in solution at high pH is troubling. At a minimum, the lack of data leaves open the possibility that lead picrate could be formed until such time as the speciation of the lead can be explained or the precipitation of lead can be fully dismissed on the basis of the solubility of lead picrate at these pHs.

An added complication in the destruction of the chemical weapons at Blue Grass is the imprecision with which the composition of the munitions is known. The original ACWA RFP stipulated that the burster charges in the 155-mm rocket warheads contained Composition B (U.S. Army, 1997a). However, the MIDAS database, which is maintained by the Army to describe the entire munitions stockpile, states that those bursters may contain either Composition B or tetrytol (U.S. Army, 1997b). Surveillance data from the Army indicate that the bursters in the Blue Grass stockpile contain only Composition B, even though certain munitions lots in the overall stockpile were manufactured with both Composition B and tetrytol. The type of burster is controlled by lot at manufacture and can be deduced from the lot number information. If a comprehensive inspection of the stockpile at Blue Grass reveals that some of the bursters contain tetryl, then it will be necessary for the technology providers to employ disassembly processes that keep the rocket propellant segregated from the burster explosive.

The EDS energetics hydrolysis test program was instituted to address the ACW I Committee’s concerns about the hydrolysis of energetic materials. The test program has now been completed, and the results indicate that the hydrolysis of the energetic materials present at Blue Grass can be conducted safely. The program addressed all of the findings and recommendations of the ACW I Committee (NRC, 1999, 2000a) with the exception of testing on the hydrolysis of energetics contaminated by agent. The committee believes that such testing is no longer necessary because the EDS program has clearly indicated that both agents and energetics are destroyed by caustic. The probability of any interaction between the agent and the energetics that would prevent complete destruction is unlikely. Moreover, the committee again notes that integration concerns such as the safe handling and destruction of agent-contaminated energetics should be addressed as soon as practicable to minimize delays during systemization of the disposal facility (NRC, 2001a).

Finding (Blue Grass) EH-1. Alkaline hydrolysis is an effective and safe method for destroying the energetic materials present in the weapons at Blue Grass Army Depot. There appear to be no insurmountable obstacles to using this technology to destroy the energetics in assembled chemical weapons.

Finding (Blue Grass) EH-2. Results from the energetics hydrolysis test program have shown that hydrolysis rates are consistent with the proposed throughput rates necessary to meet the current disposal destruction schedule for the Blue Grass stockpile.

Finding (Blue Grass) EH-3. The presence of environmentally hazardous compounds in the offgases of the hydrolysis reactor requires that an emission control system capable of removing these products be employed in any technology package that uses caustic hydrolysis for the destruction of energetic materials.

Finding (Blue Grass) EH-4. Cyanide was present in the alkaline hydrolysate from all the energetic materials tested at HAAP. If the hydrolysate is neutralized or made acidic, HCN will be liberated.

Finding (Blue Grass) EH-5. The cotton threads used to hold sheets of M8 propellant together and the rayon bags used to hold M1 propellant grains are not destroyed by the caustic under the conditions that were used in this study. This may cause clogging or entanglement in pipes, pumps, or valves.

Finding (Blue Grass) EH-6. Analysis of the hydrolysates from lead-containing energetic materials failed to account for most of the lead present in the feed.

Recommendation (Blue Grass) EH-1. Tetrytol should not be processed simultaneously with any lead-containing energetic material until such time as it can be confirmed that lead picrate will not precipitate from the hydrolysate. The solubility of lead picrate in the reaction medium should be determined as a function of temperature, and the lead in the reaction should be accounted for in a mass balance.

Recommendation (Blue Grass) EH-2. Steps should be included in the processing of M1 propellant and M8 sheet propellant to prevent the persistent rayon bags or cotton threads from clogging or jamming the hydrolysis equipment.