Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
2 M55 Rocket Storage Condition Assessments RECENT ASSESSMENTS OF STORAGE CONDITIONS The M55 rocket storage conditions reviewed in this chapter refer to a number of factors related to the agent, the munitions, and the time in storage at the depot. These factors, which affect the risk of storage and/or the conditions of processing in the disposal facility, are as follows: . Leakage of agent from a rocket, primarily because its aluminum casing becomes corroded by acid decomposition products of the agent. · Autoignition of the rocket propellant as a result of internal heat generated by the decomposition of the propellant stabilizer, the leakage of agent into the propellant, and the overpacking of leak- ing rockets. · Ignition of the stored rockets by external factors such as lightning, earthquakes, and aircraft crashes. Gelling of the agent in the rocket in certain more acidic GB agent lots. Gelling prevents removal of the agent in the rocket shear machine (RSM) be- fore shearing the rocket. Introducing the sheared rocket pieces with the fuze, burster, propellant, and gelled agent com- ponents into the deactivation furnace system (DFS ). . . This report does not deal with the risk of storage asso- c~ated with acts ot terrorism or sabotage. . . . . ,% . 11 Monitoring conducted by the Army since 1973 to track the condition of the entire U.S. stockpile of chemical munitions indicated the presence of 4,781 leaking munitions as of mid-2002 (Studdert, 2002~. Munitions found to be leaking require overpacking or other remedial actions. Of the 2,102 M55 rocket leakers, almost all contained GB, with leaking attrib- uted to corrosion of the aluminum rocket casing by ac- ids formed during agent decomposition. According to a General Accounting Office (GAO) report (GAO, 1994), 1992 Army records showed that GB M55 leakers constituted about 0.25 percent of the M55 stockpile. The GAO report also noted that the incidence of leaks in other munitions was only about 0.02 per- cent. More recent data have indicated that GB-filled munitions, primarily M55 rockets, continue to exhibit higher incidences of leakage than other munitions, as shown in Table 2-1.~ Along with leakage concerns, the Army was con- cerned that energetic materials (propellant, burster charges, fuzes ) might deteriorate and cause autoignition, particularly in M55 rockets contain- ing GB agent. In 1997, the Program Manager for Chemical Demilitarization (PMCD) published a re- port that analyzed the stability of the M28 propel- lant used to launch M55 rockets (U.S. Army, 1997a). M28 propellant is a double-base propellant iAlthough the percentage of leakers for munitions containing mustard agent H has been higher (0.46) than the percentage for munitions containing GB (0.25), there were approximately 20 times more GB munitions in the original stockpile than H munitions.
12 ASSESSMENT OF PROCESSING GELLED GB M55 ROCKETS AT ANNISTON TABLE 2-1 Munition Leaks by Type of Agent Agent GB H VX HD HT No. of munitions in stockpile 1,546,387 77,498 497,175 931,945 270,135 No. of leakers 3,851 360 273 236 61 Percentage of leakers 0.25 0.46 0.05 0.03 0.02 Note: About half the GB and VX leakers are M55 rockets (U.S. Army, 2002a). Source: Adapted from Studdert (2002~. containing nitroglycerin (NG) and nitrocellulose (NC); both NG and NC degrade slowly under stor- age conditions to generate heat and release nitro- gen oxides. A stabilizer, 2-nitrodiphenylamine, is included to react with the nitrogen oxides. If the stabilizer becomes exhausted, and if the degrada- tion rate and heat generation rate are sufficiently high, there is a possibility of autoignition. The 1997 PMCD report examined this question in detail and concluded as follows: The calculated results show that probability of autoignition for nonleaking rockets is extremely small, and is, in fact, below the minimum probability for inclusion in the QRAs [quantitative risk assessments]. Similarly, the results show that the safe storage life for nonleaking GB rockets extends well beyond the time required for dem~litanzation of the rockets. (U.S. Army, 1997a) When the PMCD report was written, the demilitari- zation program for the U.S. stockpile of chemical mu- nitions was expected to be completed in 2007. How- ever, the risk factors calculated for the year 2012 the deadline extension allowable under the Chemical Weapons Convention, which the Army currently plans to meet are also sufficiently small to support the con- clusion of the 1997 report of the PMCD. A report from the Edgewood Research, Develop- ment and Engineering Center concluded that if agent has leaked into the propellant, stabilizer degradation may be accelerated (U.S. Army, 1996a). Leaking rock- ets are placed in sealed containers ("overpacked") to prevent agent from escaping. This, however, reduces the rate of heat transfer from the rocket to the igloo and leads to more internal heating, which in turn increases autoignition probabilities. PMCD sponsored an exten- sive analysis of heat transfer from overpacked leaking rockets and concluded that autoignition at any site was extremely unlikely (approximate frequency of 1 x 1 O-5 to 3 x 1O-5 per year) and that "autoignition of stored M55 rockets is not a significant contribution to public health risk" (U.S. Army, 2002b).2 While this appears quite reasonable, the results of the study are being evaluated in detail in a National Research Council (NRC) report on stockpile degradation, due to be re- leased later in 2003. AGENT GELLING IN GB-FILLED M55 ROCKETS During the processing of GB-filled M55 rockets at the Tooele Chemical Agent Disposal Facility (TOCDF), some rockets could not be drained because of agent gelling. This gelling has been correlated with certain manufactured lots of GB agent and has been attributed to the presence of diisopropylcarbodiimide (DICDI), used as a stabilizer in some lots of GB. When GB was manufactured from 1953 to 1957 at the Rocky Mountain Arsenal, it was stored in bulk tanks and each lot was identified by an agent lot num- ber. The production methods for these lots differed, as did their subsequent treatment. These differences are documented in Army records. In the 1960s, bulk agent was loaded into a variety of containers and munitions, each identified by a munitions lot number. Thus, each item in the stockpile is identified by both an agent lot number and a munitions lot number. These procedures gave rise to four main subtypes of GB agent PRO, PRO-AS, RO-RS, and RD-RS. The 2The median site-specific annual autoignition probability for overpacked rockets at Anniston is 3 x 10-s per year, which is equivalent to about one chance in 33,000 per year that an autoignition will occur. For the other three sites where overpacked M55 rockets are stored, the autoignition probability is 1 x 10-s per year, which is equivalent to one chance in 100,000 per year. The median site-specific annual autoignition probability for nonoverpacked (undetected) leaking rockets at Anniston is approxi- mately 1.4 x 10-6 (about one chance in 700,000 per year). The lower frequency estimate for nonoverpacked leaking rockets is due to the lower peak heat generation and slightly higher heat losses for nonoverpacked rockets compared with overpacked rockets.
MSS ROCKET STORAGE CONDITION ASSESSMENTS acronyms used to describe these subtypes are explained as follows: . . . . PRO (preroundout agent). Agent lots of GB manufactured from 1953 to 1955 to meet a 92 per- cent purity specification. Tributylamine (TBA) was added as a stabilizer. Subsequent testing of these agentlots showed purifies ranging from 81 to 94 percent and indicated that the TEA was mostly in the form of (C4Hg)3NH+F-, suggesting possible production of HE (U.S. Army, 1985, 1988~. Two GB acidic degradation products, diisopropyl methylphosphonate (DIMP) and methyl phos- phonofluoridic acid (MPFA), were each detected at levels from 2 to 10 percent by weight. RO (roundout agent). Agent lots of GB manufac- tured from 1955 to 1957 to meet a modified pu- rity goal of 88 percent. A final distillation step was eliminated in the processing. Over the next few years, the Army continued to test the RO lots and found that they were showing significant acidity. Since some of the agent was intended for use in aluminum M55 rockets, where acidity would cause corrosion, some preventive measures were explored. RD (redistilled RO). RO lots of GB were redis- tilled over the next 6 years to improve purity and were redesignated as RD. In addition, TEA sta- bilizer was replaced by DICDI to reduce the acidity and allow the agent cavity of M55 rock- ets (constructed with aluminum casings) to be loaded with GB. RS. Agent lots of GB that were restabilized with DICDI were identified by adding RS to the basic agent subtype. RO-RS lots have the highest percentage of leakers, 0.273. One GB M55 rocket lot filled with PRO-RS and stored at Anniston has a leaker percent of 0.13. Other PRO-RS lots have a leaker percent of 0.009. RD-RS lots have a mean leaker percent of 0.053. All other lots have lower percentages of leakers (SAIC, 2002a). M55 rockets were loaded with GB from various agent lots during the 1960s. From analyses of leaker data since 1973, it appears that the more acidic GB agent lots are more prone to causing leakage, probably because they corrode the aluminum (U.S. Army, 1985, 1995b). Gelling problems in GB-filled M55 rockets were first encountered during the GB rocket disposal cam- paign at TOCDF. GB gelling had previously been en- countered in a few 155-mm GB-filled projectiles at the Johnston Atoll Chemical Agent Disposal System (JACADS). Gelling is identified during processing when the agent fails to drain adequately after the agent cavity of the rocket has been punched open. The de- gree of gelling can vary greatly from a thickening that increases viscosity and slows the draining process to semisolid or crystalline states. During the GB M55 rocket disposal campaign at TOCDF, almost 29,000 rockets were destroyed through the DFS at rates of up to 33 per hour, in accordance with regulatory permit allowances. Three restabilized munitions lots (5,287 rockets) were found to be gelled and were processed differently, as described in Chap- ter 3 (EG&G, 2002a). The gelling originally was observed to have taken place in certain GB lots that had been restabilized with DICDI because of their inadequate purity and high acidity. It is known that DICDI can react with residual water in the GB and form 1,3-diisopropyl urea, which is insoluble in GB and forms the urea crystals that were sometimes observed during the original GB rocket and projectile filling operations that used restabilized (-RS) agent lots (U.S. Army, 2002c). Urea crystals of- ten are observed in gelled agent lots as well. However, although gelling also seems to occur preferentially in -RS lots, the gelling mechanism now appears to be related to GB hydrolysis, which produces acidic spe- cies that react with the aluminum casing to produce aluminum phosphonate species, which, in turn, serve to link hydrolyzed GB molecules and form a viscous gel (Wagner, 2001~. Implications for Processing For the Resource Conservation and Recovery Act (RCRA) permit for TOCDF, the Utah Department of Environmental Quality allowed a processing rate of 1.6 gelled rockets per hour through the furnace based on a simple scaling of the approved limit for agent loading in the DFS. Because there are more gelled rockets at Anniston, proceeding in this manner would significantly extend the disposal schedule. As will be discussed in Chapters 3 and 4, the DFS kiln may be able to process gelled rockets at a higher rate. An analysis that includes considerations for determining a safe rate will be pre- sented. Of course, the DFS kiln system would have to be tested at the accelerated rate to prove its performance capabilities and to satisfy regulatory requirements.
14 STOCKPILE RISK CONSIDERATIONS Quantitative risk assessments (QRAs) are developed for each stockpile site to quantify the storage and dis- posal risks (SAIC, 2002b). The major storage risks were found to be associated with earthquakes and light- ning strikes. Risks from terrorist threats are handled separately by the Army and were not included in the public risk assessment. The frequencies for lightning- induced ignition of M55 rockets in a site stockpile range from 6 x 10 - to 2 x 10-3 per year and for earth- quake-induced ignition from 1 x 10 - to 8 x 10 - per year (U.S. Army, 2002b).3 Both event ranges are 3The range 6 x 10~ to 5 x 10-3 is equivalent to about one chance in 1,700 per year to one chance in 500 per year; 1 x 10~ to 8 x 10~ is equivalent to about one chance in 10,000 per year to one chance in 1,250 per year. ASSESSMENT OF PROCESSING GELLED GB M55 ROCKETS AT ANNISTON slightly higher than the estimated risk of autoignition mentioned previously, but still relatively low. None- theless, the frequencies for these natural occurrences indicate that prompt disposal is the proper course of action. As disposal operations progress, storage risk de- creases. The risk from processing is less than the stor- age risk, and storage risk can decline rapidly as rockets are eliminated. Chapter 4 addresses risk implications for four pro- cessing schedule options at ANCDF. These implica- tions are a consequence of the fact that a significant number of GB M55 rockets at Anniston contain gelled agent and of the Army's desire to process them as fast as safety allows.