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.
58 REVIEW OF ALTERNATIVE TECHNOLOGIES According to the technology provider, Ag(II) ions generated at the anode of the electrochemical cell react with the water and nitric acid of the anolyte solution to form a range of other oxidizing radicals (OH, NOT-. The Ag(II) ions and other oxidizing species then react with the organic material delivered into the anolyte vessel and are reduced to AgfI) ions, nitrate ions, and water. The organic matenal itself is completely oxidized to carbon dioxide, oxides of nitrogen KNOT and traces of carbon monoxide, protons (H+), and inorganic salts. No hydrogen is produced in the process. Off-gas from the reaction passes from the anolyte tank via a chiller (to condense nitric acid vapors) to an NOX reformer. To balance the electrochemical reaction in the anolyte vessel, a cathode reaction reduces nitric acid to nitrous acid and water, while other reduction reactions generate NO2. The evolved gases pass from the catholyte tank to the NOX reformer. The overall process is operated at a temperature of 90°C (194°F) and at atmospheric pressure. During the electrochemical reaction, nitric acid is consumed in the catholyte circuit, and water is transferred across the semipermeable membrane in the electrochemical cell from the anolyte to the catholyte. Some Agile ions are also transferred across the cell membrane. In order to maintain steady-state operating conditions, proportions of the anodyne and catholyte circuits are bled to a nitric acid and silver recovery unit, which includes the previously mentioned NOX reformer. Here, a combination of evaporation and fractional distillation is used to recover the NOx as nitric acid and to generate streams for the return of (~) silver ions together with 16 M nitric acid to the anolyte circuit and (2) 4 M nitric acid without any silver ions to the catholyte circuit. A dilute nitric acid stream is also produced, which is recycled within the plant and used to prepare organic materials before they are fed into the SILVER II process. After leaving He NOX reformer, all off-gas passes through a caustic scrubber to remove residual NOX. Treated gas Dom the scrubber is pumped into one of the two tanks of the hold-test-release system. The tanks have sufficient capacity to hold approximately 15 minutes of generated off-gas at a nominal maximum pressure of ~ atm. When one tank is full, the second tank is switched on line, and the gas in the first tank is analyzed. The gas entering each tank is continuously analyzed for chemical agent vapor content by a standard system, such as the Army's automatic continuous air monitoring system (ACAMS). The proposed concept is to hold the gas and only release it to the atmosphere upon confirmation that it is free of chemical agent. During the development of the SILVER T! process, its applicability to many different organic compounds (including explosives, propellants, rocket fuels, pyrotechnics, and industrial solvents) was investigated. Disassembly of Munitions and the Removal of Agent/Energetics The transport, receipt, and pre-processing of munitions prior to processing with SILVER ~ are done using the Anny baseline technology modified with additional material-handling stages at the end of the process. These stages reduce material size and segregate material for SILVER II processing. The baseline munitions disassembly process is described in Appendix C.
ARCTECH ACTODEMIL TECHNOLOGY PACKAGE TABLE 4-1 Summary of the ARCTECH ACTODEMIL Approach 59 Major Operation Approach(es) Disassembly of munitions Army's existing baseline disassembly method except that rockets are sheared in 10 places instead of seven; water jets may be used for washout of energetics. Treatment of chemical agent Mustard. Hydrolysis at near-neutral pH followed by oxidation with hydrogen peroxide, both steps in the presence of humic acid. GB and VX. Base (KOH) hydrolysis followed by oxidation with Fenton's reagent, both steps in the presence of humic acid. Treatment of energetics Base (KOH) hydrolysis in presence of humic acid. Treatment of metal parts Base (KOH) hydrolysis in presence of humic acid. Treatment of dunnage Base (KOH) hydrolysis in presence of humic acid. Disposal of waste Solids. Stabilized with cement and sent to appropriately permitted landfill; metal parts shipped to Rock Island Arsenal for SX treatment. Liquids. Stabilized with cement and sent to appropriately permitted landfill. Gases. Offgas from process treated in acid scrubber, filtered through activated carbon, and discharged to the atmosphere; no hold-test-release step. in Erickson, 1997) and concluded that the "humin" present in the soil, which is distinct from the humic acid, binds with the contaminants. Studies by Bollog (1992), Pennington et al. (1992), and Caton et al. (1994) concluded that the biodegradation of TNT in soil stops short of mineralization and that reaction products con- taining reduced nitro groups bind strongly to the humic fraction of the soil. According to the technology provider, humic acid is an association of molecules that form aggregates of elongated bundles of fibers at low pHs and open flex- ible structures perforated by voids at high pHs. The voids can trap and adsorb both organic and inorganic compounds if the charges are complementary. Humic acid combines with organic compounds by electrostatic bonding (i.e., attraction of a positively charged organic cation to an ionized carboxylic or phenolic group), hy- drogen bonding, and ligand exchange. In addition, the high concentrations of stable free radicals in humic acid are capable of binding organic compounds that can ion- ize or protonate to the cation form. The mechanisms that have been postulated for the adsorption of organic com- pounds include (1) Van derWaal's attractions, (2) hydro- phobic bonding, (3) hydrogen bonding, (4) charge trans- fer, (5) ion exchange, and (6) ligand exchange. Van der Waals forces, which are involved in the adsorption of non-ionic and nonpolar compounds, are created by short-range dipole-dipole interactions. Van de Waals forces are additive in nature, and these forces between the atoms of the adsorbate and the adsorbent can create considerable attraction for large molecules. However, the data provided by ARCTECH indicate that the po- tential separation effect of this binding action is negli- gible for the chemical agents. DESCRIPTION OF THE TECHNOLOGY PACKAGE Disassembly of Munitions and the Removal of Agent/Energetics Rocket Disassemb/y The baseline disassembly process (described in Ap- pendix C) is used except that rockets are sheared in 10 places to allow better access to agent and energetics. The normal baseline process shears the rocket in seven places. ARCTECH also stated that water jets could be used to aid in the removal of energetics from muni- tions, although details were not provided. After the agent is drained, the rocket parts are sent to a decon- tamination reactor containing a-MAX solution, which is caustic and dissolves the aluminum parts, helping to eliminate residual agent. The solution also attacks the energetics and hydrolyzes them. Projecti/e/Mortar Disassemb/y For the demilitarization of the 155-mm, VX-filled M121A1 projectile, ARCTECH uses the baseline dis- assembly process up to the point at which munitions
60 Land Mine Disassembly REVIEWOFALTERNATIVE TECHNOLOGIES For M23 mines, the Army baseline disassembly system is used to punch and drain the agent and push out the burster charge. In addition to this baseline processing, the mines are sheared into four parts to open the radial initiator charge and expose the energetic components. Treatment of Chemical Agent The chemical agent drained Tom the various components is transferred to an interim storage tank where it is mixed and continually stirred with dilute nitric acid recycled Mom the SAVER ~ process. The dilute nitric acid provides the necessary water balance for the electrochemical oxidation reactions in the agent SILVER IT reaction circuit. It also disperses the agent, thereby increasing the surface area/volume ratio of any gelled agent and provides partial destruction via hydrolysis of the chemical agent. The interim storage provides a buffer volume for the feed of materials to the SILVER I! reactors. The design calls for interim storage capacity for up to 48 hours of SILVER I] operation. The chemical agent/dilute acid mixture is fed continuously to the SILVER I] reaction circuits for agent destruction dunng normal operation. A schematic drawing of the anolyte and catholyte circuits is shown in Figure 3-3. The anolyte circuit consists of a glass-lined anolyte vessel in which the reagents and the agent solution are mixed to start oxidation. The vessel is maintained at 90°C (194°F) through the use of a Herman fluid circulating through a jacket surrounding the vessel. The agent matenal hydrolyzed by Me dilute nitric acid is oxidized in the anolyte vessel by the Ag (~) ions, which are reduced to Ag (~) in the process. Carbon dioxide (CO2), oxygen (02), carbon monoxide (CO), nitrous oxide (N2O), and water are generated in the anolyte vessel. According to the technology provider, NOX is not evolved from the anolyte, but under certain conditions, excess NOX can be present. These gaseous products then flow to the NOX recovery system. The anolyte solution is circulated to the electrochemical eel] where He Aged ions are regenerated. Other reagents added to the anolyte vessel mclude calcium nitrate to precipitate CaSO4 and calcium fluoride, silver nitrate solution to maintain silver concentration (when chloride is present, silver chloride will precipitate), and nitric acid. A hydrocyclone between the anolyte vessel and the electrochemical cell prevents solids Dom entering and Potentially blocking the narrow passages In the cell. A continuous bleed of anolyte solution takes place at the hydrocyclone for recovery of silver chloride and nitric acid. The vented gases from the anolyte vessel are passed Rough a water/glycol-cooled condenser at 0°C (32°F) to condense nitric acid vapor, water vapor, and trace volatile organic compounds. The condensate is returned to the anolyte vessel, and the remaining off-gas is sent to the NOX reformer. The catholyte circuit also consists of a stirred glass-lined vessel from which the catholyte solution is circulated to the cathode side of the electrochemical cell. The catholyte vessel also has a thermal jacket to maintain the temperature at 90°C (194°F). In the cathode side of the cell, nitric acid is reduced to nitrous acid, NOX, and water. The NOX is vented from the catholyte vessel to the BOX refonner. A slipstream of the catholyte solution is continuously bled to the silver and nitric acid recovery system. A make-up input stream of dilute acid to the catholyte vessel maintains the necessary concentrations of nitric acid.
ARCTECH ACTODEMIL TECHNOLOGY PACKAGE it chelates metals; and it reduces oxidized metal spe- cies. Humic acid can easily be precipitated by lowering the pH to less than 2.0. ARCTECH believes that the humic acid exists as a colloid at a pH of less than 2.0, and reducing the pH to below 2.0 causes the colloidal particles to agglomerate. ARCTECH claims that the a-MAX solution effec- tively and irreversibly hydrolyzes chemical agents and energetic materials, converting them into nontoxic compounds. In the case of VX, the reaction is expected to destroy (or convert to other materials) 99.9999 per- cent of the agent in six hours. EA-2192 may be present temporarily as an intermediate but is not present in the product stream. Neutralization, or hydrolysis, of chemical agents with a strong base has been documented extensively in the open literature and is the method the Army has se- lected for the destruction of HD and VX stored in bulk containers at Aberdeen, Maryland, and Newport, Indi- ana (see Appendix D). The most significant study of agent hydrolysis using ARCTECH's approach was done by GEOMET in early 1998 to assist ARCTECH in responding to the Army's list of data gaps. The study includes tests of the hydrolysis of GB, VX, H. HD, and HT. Two-liter glass flasks were used as reactor vessels. For GB and VX, initial experiments involved 100 g of agent treated in the presence of large quantities (240 g) of KOH, 280 ml of a-MAX solution, and 650 ml of added water. Temperatures of 90°C (Run 1) and 70°C (Run 2) were used. In a third set of experiments, the 650 ml of added water was replaced with 280 ml of reaction liquor from the first set of reactions and 280 ml of reaction liquor from the hydrolysis of energetics. The purpose of this replacement was to simulate steady-state plant operations in which the residual re- action liquor from one batch would be used for the next batch. In a final set of experiments (Run 4), only KOH and fresh water were used (no humic acid). Based on results from previous studies, the hydroly- sis reactions for mustard, including H. HD, and HT, were run at near neutral conditions. KOH was added only to keep the pH above 5. For H. only one reaction, at a temperature of 90°C, was carried out. For each run, agent was continuously pumped into the reaction flask at a rate of 8.6 ml/hr until a total of 100 g of the agent had been added over a period of 12 hours. The reactions were allowed to continue for a 61 maximum of 24 hours. The experimental conditions for the tests at 90°C are shown in Table 4-2 and the results are shown in Table 4-3. The appearance of an organic phase in the GB and VX hydrolysates has been reported by others (see Appendix D) but was not no- ticed by GEOMET. Destruction of all agents was com- plete after 24 hours to below detection limits shown in Table 4-3. ARCTECH precipitates the humic acid by adding nitric acid to reduce the pH to less than 2.0, then de- strays the reaction products using oxidation with Fenton's reagent (i.e., hydrogen peroxide plus ferrous ion) for GB and VX, and hydrogen peroxide for mus- tard. Some of the humic acid, perhaps 10 percent, is destroyed by the hydrogen peroxide during this step. ARCTECH claims that the presence of hump acid in- hibits the formation of gaseous reaction products dur- ing the destruction of GB and VX. In the full-scale process, two jacketed, 4,000-gallon stainless steel reactors are used in a two-day cycle. The temperature is held at 90°C, and the operating pressure is atmospheric. The reactors are fitted with internal and external mixing systems, with the external system us- ing a static mixer. The reactors also have external cool- ing loops and reflux condensers. Vented gases from the TABLE 4-2 Summary of Experiments by ARCTECH with Agents at 90°C. The agent addition rate was 8.6 mVhr. Agent Reactants Mass or Volume VX 100g water 650 ml dilute a-MAX (with 240 g KOH) 280 ml KOH water dilute 3% humic acid 66 g (to maintain pH > 5) 280 ml 560 ml Source: ARCTECH and ICE Kaiser, 1998.
62 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS TABLE 4-3 Results of Analysis for Residual Agent during ACTODEMIL Neutralization Process at 90°C Agent Sample Designation Residual Agent Concentration (mg/L) Residual Hydroxide Concentration Gas Evolved Conclusions GB L1 at t = 0 hr< 20 4.43 M None (agent < 0.5 ppb) Reaction complete by t = 6 hr< 20 3.63 M 12 hours; no gas evolved; t= 12hr< 20 2.61 M all agents destroyed; t = 24 hr< 20 2.50 M KOH left at end of reaction. VX L1 at t = 0 hr< 20 4.95 M None (agent < 0.5 ppb) Reaction complete by t = 6 hr128 4.54 M 12 hours; no gas evolved; t= 12hr145 3.97 M all agents destroyed; excess t = 24 hr< 20 4.29 M KOH left at end of reaction. HD L1 at t = 0 hr< 20 No hydroxide used None (agent < 0.5 ppb) Complete destruction of t = 6 hr200,000 in the tests except agent at end of 24-hour t = 12 hr140,000 for pH adjustment. period; no gas evolved. t=24hr<20 HT L1 at t = 0 hr< 20 No hydroxide used None (agent < 0.5 ppb) Complete destruction of t = 6 hr230,000 in the tests except agent at end of 24-hour t = 12 hr130,000 for pH adjustment. period; no gas evolved. t=24hr<20 H L1 at t = 0 hr< 20 No hydroxide used 75 ml (agent < 0.5 ppb) Complete destruction of t = 6 hr420,000 in the tests except agent at end of 24-hour t = 12 hr310,000 for pH adjustment. period; minimal gas evolved; t = 24 hr< 20 no unreacted agent in gas stream. Source: ARCHTECH, and ICF Ka~ser, 1998. reactors and ventilated air from the reaction area are directed to a gas scrubbing system that contains both acid and caustic scrubbers. From these scrubbers, the gases are directed to activated carbon filters for further treatment. The reactors are initially filled with an aqueous re- action solution prepared by mixing humic acid with either fresh water or recycled reaction liquor from which the humic acid has been precipitated. KOH is added for the hydrolysis of GB and VX. Agent is fed into one of the reactors as it is drained from the muni- tions. At the end of the shift, the feed is secured, and the reaction is allowed to proceed to completion. A six- hour reaction time will ensure the destruction of any residual EA-2192 from VX treatment. On the follow- ing day, the contents are tested. If the reaction is com- plete, the reactor contents are drained into a spent reac- tion solution tank, and small plastic and metal parts are removed. The drained reactor is then filled with fresh reaction solution. Thus, on any given day, one of the two reactors is being used for agent destruction, while the other is being tested, drained, and refilled. Treatment of Energetics The technology provider has conducted substantial tests on bulk energetic materials, and test results indi- cate that all of the energetic materials can be completely destroyed. Hydrolysis of nitrocellulose, cyclotetra- methylene-tetranitramine (HMX), RDX, TNT, nitro- glycerin, and other aliphatic nitrate esters, nitramines, and nitoaromatic type propellants and explosives has been demonstrated. In laboratory-scale studies, complete denitration of the nitrocellulose was obtained in a matter of minutes at optimum operating conditions. No solid residue re- mained. The product was analyzed for a broad range of volatile and semivolatile compounds, and no hazard- ous products were found at levels of concern. The prod- uct was also characterized and formulated into a fertil- izer product, which was evaluated in tests of plant growth. Pilot-scale tests using 100 lb batches have been car- ried out on single, double, and triple base propellants. All of the energetic compounds were destroyed, yielding
ARCTECH ACTODEMIL TECHNOLOGY PACKAGE products that complied with RCRA, TCLP, and Uni- versal Treatment Standards (UTS) levels. ARCTECH conducted these pilot-scale tests at the Hawthorne Army Depot using a stirred tank reactor under a con- tract with the U.S. Army Industrial Operations Com- mand. The jacket of the reactor was used to circulate either steam or cold water, and the reactor was equipped with an anchor agitator. The double-base propellant tested was type M26 from 105-mm shells. The M26 propellant was suc- cessfully denitrated, with greater than 99.9999 percent removal of nitro groups within 24 hours. Tests of the products for TNT, RDX, dinitrotoluene, nitroglycer- ine, and nitroguanadine were negative. Ammonia was evolved during the reaction. Propellants tested in the laboratory included IMR (a single-base, small-arms propellant), WC 830 (double- base, small-arms propellant), Hercules 2400 (double- base, small-arms propellant), Hercules RS nitrocellu- lose (neat powder), and M30 (triple-base propellant). Laboratory studies showed that the reaction was com- plete in a matter of minutes for nitrocellulose; all of the nitrocellulose was consumed, leaving no solid residue. The cyanide concentration in the product was below the detection limit of 100 mg/kg. Similar results were achieved for other explosives and propellants, includ- ing RDX, HMX, nitroglycerine, TNT, dinitrotoluene, and tetryl. Energetics are removed from the munitions and handled differently depending on the particular muni- tion. If the energetic material is in an aluminum con- tainer (e.g., some bursters), the container and energetic material are fed to the hydrolyzer. The concentrated KOH solution dissolves the aluminum and then attacks the energetic material. (See Appendix E for possible problems with burster material, such as tetrytol, where the reaction could be slow unless the energetic material is cut up further.) If the energetic material is in a steel container, it is washed out as a slurry, using KOH solu- tion. The slurry is then fed to the hydrolyzer. In the proposed full-scale process, two double- walled stainless steel reactors are used in a two-day cycle. The temperature is held at 80°C, and the operating iThe shells were described as follows: extruded cylindrical grains, 7 per- forations, 0.219" diameter, 0.518" long; nitrocellulose 67.25 percent; nitro- glycerin 25 percent; ethyl centrallite 6 percent; barium nitrate 0.75 percent; graphite 0.3 percent. 63 pressure is atmospheric. The reactors are fitted with mixers and reflux condensers. One reactor is initially filled with concentrated a-MAX solution, and energet- ics are fed into it as they are removed from the muni- tions. At the end of the shift, the feed is secured, and the reaction is allowed to proceed to completion. On the following day, the contents are tested and drained into the spent a-MAX tank, and small plastic and metal parts are removed. The reactor is then re- filled with fresh concentrated a-MAX solution. On any given day, one of the two reactors is being used for the destruction of energetics, while the other is being emp- tied and refilled. Treatment of Metal Parts Decontamination of metal parts, along with the dun- nage (but not DPEs), is carried out in double-walled, jacketed, stainless-steel reactors. High pH (20 percent KOH) a-MAX solution is used for parts contaminated with GB and VX. A near-neutral aqueous solution is used for parts contaminated with mustard. The process- ing temperature is 90 tolO0°C, and the operating pres- sure is atmospheric. The reactors are fitted with mix- ers, cleaning lances, and reflux condensers. Two process trains comprising three reactors each are installed, with each train sized for one days output of metal. First, one tank is filled with the appropriate decon- tamination solution. For GB- and VX-contaminated materials, concentrated a-MAX solution is used for rocket parts, and dilute a-MAX solution is used for other parts. Water/humic acid is used for mustard-con- taminated parts. Metal parts from munitions are con- veyed to and then immersed in the tank using charge cars in a remote operation. At the end of the shift, the feed is secured, and the reaction is allowed to proceed to completion. A reaction time of 6 to 12 hours is planned. The parts are removed after treatment and combined with the small metal and plastic parts re- moved from the agent and energetic destruction tanks and rinsed with an unspecified amount of clean water. In concept, the treated metal parts/dunnage should comply with the 3X standard for decontamination and will be transported to the Rock Island Arsenal for final treatment to the 5X level. Small-scale tests carried out by ARCTECH indicate that this approach is technically feasible.
64 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS ARCTECH is considering using waterjet cutting for accessing mine initiator energetics. Fuzes are alumi- num-bodied and will dissolve in high pH solutions. The technology provider was not certain whether or not azide would be destroyed by a-MAX solution. Treatment of Dunnage (and Protective Suits) Contaminated dunnage is treated to a 3X level using the metal parts decontamination lines. ARCTECH shreds or cuts DPE suits and contaminated pallets and treats these materials by immersion in a reactor con- taining dilute a-MAX solution. The reactor is a double- walled stainless-steel tank held at 90°C and operated at atmospheric pressure. It is fitted with an agitator and a reflux condenser. Nonmetal materials (including DPE suits) are sent to a hazardous-waste landfill for disposal, and cleaned metal is sent to the Rock Island Arsenal for treatment to 5X. Process Instrumentation, Monitoring, and Control Operating conditions, such as temperature, pH, pres- sure, and gas and liquid flow rates will be monitored using standard industrial equipment. Feed Streams In addition to the chemical munitions, which con- tain agent, energetics, and metal parts, several other major chemical feed streams are involved in the pro- cess: KOH for hydrolysis; humic acid; nitric acid for pH control (to precipitate humic acid); Fenton's re- agent, an aqueous solution of hydrogen peroxide and ferrous sulfate, for final oxidation of most of the organ- ics; and cement for stabilization of most of the final product. The required weight of these chemicals greatly exceeds the weight of agent treated. For example, the chemicals required to treat a sample mass of 100g of VX are shown in Table 4-4. The quantities of other materials required to treat (1) agents other than VX and (2) energetics will be different from those listed in the table, but the amounts will be large. (No plan has been presented for handling grit and fine solids from waterjet cutting.) TABLE 4-4 Materials Required for Processing 100 g of VX Using the ACTODEMIL Process Material VX KOH Humic acid HNO3 H2O2 Cement Mass Required (g) 100 234 28 22s 1,740a 1 o,ooo+b a30% solution (estimated). bEstimated assuming a concrete mixture of 50% cement and 50% water. Waste Streams The major waste stream produced by the ACTO- DEMIL treatment process is the product of the stabili- zation operation, incorporating all of the effluent from hydrolysis/hydrogen peroxide treatment. Two products result from hydrolysis: precipitated humic acid containing adsorbed salts and organics; and a solution of salts and organics. Both are treated with Fenton's reagent to oxidize the remaining organics completely without oxidizing more than 10 percent of the humic acid. Spent activated carbon from the air purification system is added, as is spent scrubber solu- tion from off-gas treatment systems. Both the liquid and solid streams are then stabilized with cement. Nei- ther the peroxide treatment nor stabilization with ce- ment has been demonstrated. Other waste streams are: decontaminated DPE suits and other dunnage; metal parts cleaned to a 3X condi- tion and packaged for off-site shipment; and solids from waterjet cutting (if used). Air and fumes from various parts of the process are collected and treated prior to discharge. Sources in- clude the explosion-containment vestibule, explosion- containment room, toxic cubicle, demilitarization room, and reactors. The gases evolved during the ener- getic destruction process, including ammonia, amines, and nitrogen oxides, are removed from the gas stream by the acid scrubber. The gas treatment is similar to the treatment planned for the Aberdeen/Newport neutral- ization processes in which gaseous emissions are first scrubbed with dilute nitric acid solution to remove any amines and then scrubbed by caustic solution to re- move any residual agent. Mustard contains volatile low
ARCTECH ACTODEMIL TECHNOLOGY PACKAGE molecular weight chlorinated hydrocarbons that are difficult to hydrolyze or to remove with caustic or acid scrubbers or with activated carbon absorbers. ARCTECH does not address the management of these compounds in their process. The gas then passes through a chiller to reduce the water vapor content and a gas heater to elevate the tem- perature above the dew point. Treatment with activated carbon filters follows. The filtered air then passes through a baseline heating, ventilating, and air-condi- tioning filtration system for final treatment and is then discharged to the atmosphere. The mass balances vary with the munition being treated, and the process is still under development. A sample mass balance based on small-scale laboratory tests is shown in Table 4-5. This sample covered only the hydrolysis step and did not include the final oxida- tion with hydrogen peroxide (still to be demonstrated) or stabilization with cement. ARCTECH did not estimate the amount of cement required to stabilize the entire product (salts plus wa- ter). However, the amount will be large (e.g., 10,000 parts cement per 100 parts original VX). Start-up and Shutdown The processes unique to ARCTECH are relatively simple. The hydrolysis reactors are the most complex because the contents are first heated with steam to the operating range (80 to 100°C, depending on the reac- tion) and then maintained at the desired temperature with chilled water. Shutdown involves only draining the reactors and turning off the heating/cooling system. If a process upset occurs, the reactions in the hydroly sis reactors will continue, so the process cannot be quickly shut down. EVALUATION OF TH E TECH NOLOGY PACKAG E Process Efficacy Effectiveness Laboratory and pilot-scale tests have shown that the agents and energetic materials are completely de- stroyed. Hydrogen peroxide or Fenton's reagent is used to complete the destruction of agents. 65 Like the Army's neutralization process selected for the Aberdeen, Maryland, and Newport, Indiana, sites, the ACTODEMIL technology utilizes base hydrolysis. The Army has established the efficacy of the neutral- ization technology at both laboratory and pilot scales and is now moving forward with implementation of a full-scale system for the destruction of bulk agents at those two sites. The 1998 GEOMET study (ARCTECH and ICE Kaiser, 1998) showed that agent is effectively de- stroyed using ARCTECH's hydrolysis approach. Moreover, no Schedule 1 compounds were found in reaction product streams. However, several Schedule 2 compounds were detected in the liquid and solid frac- tions after humic acid precipitation. For GB, the com- pounds detected included isopropyl methylphosphonic acid, methylphosphonic acid, diisopropyl methyl- phosphonate, and dimethyl methylphosphonate. Simi- lar results were obtained for VX, with EMPA, methyl- phosphonic acid, and diisopropyl methylphosphonic acid detected. Thiodiglycol, 1,4-oxathiane and 1,4-dithiane were found in products from the hydroly- sis of HD, HT, and H. The 1998 GEOMET study also evaluated the use of hydrogen peroxide or Fenton's reagent to destroy Schedule 2 compounds, and the results were inconclu- sive. Thiodiglycol was destroyed to below detection limits. The Schedule 2 compounds from hydrolysis of VX and GB were reduced in concentration but not to below detection limits. This issue must be addressed. ARCTECH's research on the treatment of energet- ics has ranged from laboratory-scale tests on a variety of energetic materials to larger scale testing on bulk propellants and energetics. Large-scale testing on single-, double-, and triple-base propellants at Hawthorne Army Depot had favorable results. With the exception of single-base propellant, the reactions were apparently rapid and resulted in the complete destruc- tion of the energetics. No solid residues remained, and the reaction products were nonenergetic and nontoxic after 24 hours residence time. Ammonia was evolved during the reactions. To test for treatment of PCBs, a 10,000 ppm sample of Aroclor 1253 oil was placed on a sample of alumi- num. The sample was washed with a-MAX solution for 30 seconds. The a-MAX solution was then acidified with nitric acid, which precipitated the humic acid. No
66 TABLE 4-5 Humic Acid ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS Feed and Product Masses and Concentrations for Hydrolysis of 100 g of VX in the Presence of Product Streams Total Feed Liquid Solid VX 100 g EMPA46.4 g 3.21% EMPA 1.20 g 2.51% KOH 234 g Thiol28.8 g 1.99% Thiol 0.06 g 0.13% HA 28 g MPA4.2 g 0.29% MPA 0.11 g 0.23% Water 908 g "Pentanami de"15.8 g 1.09% Chloride 0.34 g 0.70% HNO3 225 g Water + Salts1,351.8 g 93.42% HA 28.00 g 58.33% Water 18.29 g 38.10% Total 1,495 g Total1,447.0 g 100% Total 48.00 g 100% Source: ARCTECH and ICE Kaiser, 1998. PCBs were found in the remaining liquid. Chloride ion concentration in the reaction mixture increased during the reaction, providing evidence of PCB dechlorina- tion, but the completeness of dechlorination was not measured. The technology provider concedes that ad- sorption of PCBs on the precipitated humic acid is pos- sible and that these adsorbed PCBs could be thermally Resorbed. But this has not been tested. The unique aspect of this technology is the use of humic acid to separate products of agent hydrolysis from solution. The separation factors2 for the products (i.e., concentrations on humic acid vs. concentrations remaining in solution) are, therefore, important param- eters for determining the effectiveness of humic acid. These concentrations can be obtained from material balances. One separation discussed in the literature is for atra- zine/humic acid in ethyl alcohol (Sanjay and Fataftah, 1997~. The separation factor, 375, calculated from the data is very large and represents a very selective sepa- rat~on. The concentration of atrazine was very low, however, only about two ppm (by weight) in the solu- tion. The separation factor would be expected to de- cline as the concentrations increased. Tests on agents with preliminary mass balances have been presented in the ARCTECH data-gap resolution technical report (ARCTECH and ICE Kaiser, 1998~. Separation factors can be calculated from the reported data. The concentration levels of the materials of inter- est are not very high (e.g., 2 to 3 percent) (see Table 4-5), 2 The separation factor is defined as a=[y/(l-y)]/[x/(l-x)], where y and x are concentrations of the material of interest in the two phases. For this application, y represents the solid phase, and x represents the liquid phase. although they are much higher than in the "atrazine" case noted above. There were two product streams from the experiment: a wet humic acid (solid) stream from the centrifuge containing hydrolysate product; and a liquid solution. The total feed and two product streams are summarized in Table 4-5. (Some minor components have been omitted). None of the products normally expected from the hydrolysis reaction appears to have concentrated on the HA-solid. For example, EMPA concentration is 3.2 percent in the liquid, 2.5 percent on the solid. If the EMPA concentration is recalculated assuming that it is simply dissolved in the water associated with the wet solids, its concentration in the liquid is then 2.9 percent (virtually the same as in the bulk solution). The data for other major components is hard to ex- plain. The thiol (a major product of VX hydrolysis) appears in the product liquid but is reported to be al- most nonexistent on the solid. The committee postu- lated that perhaps this is because thiol was not extracted by the analytical procedure. It appears that a relatively small amount of humic acid will be used in the process (the a-MAX solution is 3 percent humic acid). As a consequence, the final solid product solid humic acid with any adsorbed salts and water is a very small fraction of the total. For ex- ample, as Table 4-5 shows, the liquid product is 1,447 g (96.8 percent of the total product), and the amount of EMPA in the liquid is 46.4 g. In contrast, the mass of solid product is only 48 g (3.2 percent of the total orod- uct), of which 1.2 g is EMPA. Two conclusions can be drawn from these data: (1) separation factors appear to be very modest for
ARCTECH ACTODEMIL TECHNOLOGY PACKAGE some important components; the separation factor for EMPA, for example, is or ~ 1 (essentially no separa- tion); and (2) even if the separation factor were large, there is not enough HA-adsorbent present to make a significant difference (i.e., the solid product is a very small fraction of the effluent stream). The final process step is oxidation with Fenton's re- agent (H2O2~. Reported data show EMPA is reduced to either the ppm range or below the quantifiable limit. Unfortunately, the amount of H2O2 was not given, but, it can be estimated assuming that the organics are oxi- dized to CO2, H2O, etc. The theoretical requirement for VX oxidation is 522 g of H2O2 for 100 g of VX (i.e., for the example mass balance in Table 4-5~. This corre- sponds to 1,740 g of 30 percent H2O2 solution. Thus, hydrogen peroxide will be a major input to the process. Sampling and Analysis ARCTECH has not provided a definitive sampling and analysis plan. However, the proposal states that agent will be monitored at the outlet of the off-gas vent from the overall process, at the air intake for the opera- tion, and in worker activity and observation corridors. The agent monitoring techniques are the same as those in the Army's baseline process and at Aberdeen and Newport. Monitoring of agent in aqueous reaction ma- trices mimics techniques used at Aberdeen and New- port. The committee does not anticipate that new ana- lytical techniques will be necessary. ARCTECH's proposal states that the reactors used for hydrolysis of agent and energetics will be operated batch-wise on a 24-hour cycle. The reaction mixture from the previous batch will be sampled at the begin- ning of this cycle, which provides 6 to 12 hours of re- action time after the last addition of agent or energetics to the reactor. However, the process was changed after the proposal was submitted. Hydrogen peroxide or Fenton's reagent could now be added to the hydrolysis reactors after the hydrolysis reaction is completed. Whether or not this would lengthen the reaction cycle or change the sampling plan is not clear. ARCTECH's proposal also mentions measurements of pH and oxidation/reduction potential and states that decontaminated dunnage and DPE will be tested for agent before being sent off site for disposal. 67 Maturity The ACTODEMIL technology is at an early overall stage of development. The hydrolysis of agent with a-MAX has been done only on a small bench scale. Hydrolysis of certain energetics has been done on a much larger scale (i.e., hundreds of pounds of energet- ics). Many questions about the effect of using humic acid have not been answered. The proposed oxidation step using H2O2 has not been demonstrated. Although the water balance is still uncertain, continued develop- ment would most likely show the generation of a large aqueous waste stream, which would have to be stabi- lized and shipped to a landfill. Robustness Robustness the ability of a process to accom- modate wide variations in the composition and oual . ,~ ,~ . . fly or reeastocK Is difficult to assess at this early stage of development. Because the complete re- moval of agent and energetics from munitions might be difficult, the technology provider proposes shear- ing rockets in 10 places instead of the seven places in the baseline technology. Water jets may be used to help remove energetics from munitions. The opened munitions would be immersed in an a-MAX solution to hydrolyze the energetics and any remain- ing agent. Preliminary tests of this procedure were successful, but whether or not this approach pro- vides adequate and sufficiently rapid destruction of agent and energetics under all expected operating conditions remains to be demonstrated. Tests indicate that once the agent and energetics have been removed from the munitions, they can be successfully hydrolyzed. The ACTODEMIL process is a batch process, and the reaction time can be extended if agent and energetics are not destroyed in the normal operating period. Therefore, the robustness of the hy- drolysis step is expected to be adequate. The remainder of the process, comprising humic acid precipitation, chemical oxidation, and solidifica- tion or stabilization, is not sufficiently developed for a meaningful assessment of robustness. Studies on a larger scale and studies under off-normal conditions should be done.
68 Monitoring and Contro/ Because of the mild reaction conditions and the con- ventional nature of the equipment for this technology, monitoring and control should be straightforward and conventional. App/icabi/ity The process as currently developed can destroy all types of agent without generating other Schedule 1 compounds. Certain Schedule 2 compounds will be produced, however, and their complete destruction re- mains to be demonstrated. Complete destruction of several types of explosives and propellants has been demonstrated. The process can also decontaminate agent-contaminated metal parts and DPEs. Dissolution of aluminum parts, but not steel, has also been demon- strated. The ability of humic acid to adsorb reaction residues selectively was not demonstrated. Overall, the concept appears to be applicable to all munition types, but because data are lacking in several key areas (e.g., destruction of Schedule 2 compounds, feasibility of recycling process liquor), it is impossible to draw definitive conclusions about the applicability of the process. Process Safety The unique equipment for this process with neutralization, waterjet cutting (if used), oxida- tion, and solidification. Reverse assembly, waterjet washout and cutting, as well as energetics and agent neutralization, will take place in an explosion-contain- ment room to minimize the potential effects of explo- sions during the handling and processing of agent and energetics. The hydrolysis processes will operate at low temperature (80 to 100°C) and ambient pressure, and hydrolysis and oxidation will be performed in a batch mode. Thus, the effectiveness of treatment can be ascertained prior to release for solidification. So- lidification is a routine industrial process that occurs downstream of the primary and secondary detoxifica- tion processes, which include a hold-test-release op- eration. The solidification system should pose no unique hazards. is associated ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS Worker Health and Safety The hazardous materials unique to this process are potassium hydroxide, hydrogen peroxide, and nitric acid, all of which are widely available and widely used industrial chemicals. If a process upset occurs, the in- complete hydrolysis products will be very hazardous. Procedures are expected to be established for safe shut- down and restarting of the process. The air effluent during an upset will continue to be treated by the scrub- bers and the activated carbon filters. The disassembly processes are derivatives of the baseline processes and are not considered to create new or increased risks to workers. Waterjets are being con- sidered for washing out energetics, and waterjet cut- ting may be used to make mine-initiator energetics more accessible. As discussed in Appendix G. water- jet washout and cutting are commonly used systems and should not pose unique hazards in this application because they will be used in the explosion-containment room. Public Safety Release of agent or other regulated substances in plant effluents is expected to be extremely unlikely. The destruction of agent and energetics will be verified after Fenton's reagent or hydrogen peroxide is added to the hydrolysate in a hold-test-release operation. The air effluent from all processes and areas will be con- tinuous and will be cleaned using baseline acid and caustic scrubbers and activated carbon adsorption. No hold-test-release operation is contemplated for the air or gaseous effluents. Instead, these streams will be monitored continuously. ARCTECH claims that any flammable gases gener- ated in the neutralization process will be adsorbed by the humic acid, but this has not been thoroughly dem- onstrated. Therefore, the vent system should be de- signed and operated as though flammable gases will be present. The technology provider claims that the reactions between the a-MAX solution and the agents are mildly exothermic. The reactors will have water-filled jackets for heating or cooling. Off-gases will pass through re- flux condensers cooled with chilled water. The Army's studies on VX have shown that the heat released during
ARCTECH ACTODEMIL TECHNOLOGY PACKAGE the hydrolysis reaction is moderate and that failure of the cooling system to remove the heat of reaction will cause only a 5 to 8°C temperature excursion. Because of the similarity of ACTODEMIL to the Army's hy- drolysis technology in terms of the basic reaction dur- ing the destruction of chemical agents, ACTODEMIL is not expected to result in any catastrophic thermal excursions. This same analysis also applies to the hy- drolysis of other agents. However, reactions involving hydrogen peroxide and organic compounds are known to be violent and exothermic at times. This issue has not yet been addressed by the technology provider. The maximum amount of agent or energetic that can be treated in a single batch in the reactor will be speci- fied and will comply with the design basis require- ments. An upset in the flow rate of a feed stream could cause minor changes in the reaction and less efficient destruction. This condition can be countered by hold- ing a batch until it has been checked and extending the reaction time if necessary. Based on laboratory-scale, pilot-scale, and demon- stration tests conducted with energetics, ARCTECH has extensive data on the amount of heat released dur- ing their destruction. The ranges of values for heat re- leased from single-, double-, and triple-base propellants have been established. One of the design features is that enough a-MAX solution is used for the reaction with energetics to ensure that the total heat energy re- leased by destruction of the propellants and explosive compounds will not cause a thermal runaway. This will have to be verified experimentally. ARCTECH plans to hydrolyze different types of energetic materials simultaneously in the same reac- tors. As discussed in Appendix E, the committee is concerned that this could lead to the formation of com- pounds that are both energetic and sensitive. There- fore, energetic materials should be processed in sepa- rate reactors unless testing shows that the formation of sensitive compounds does not occur. Human Health and the Environment ARCTECH claims that the process is environmen- tally sound because there are no large process gas streams. Formation of dioxin and furans is very un- likely because processing temperatures are less than 100°C. However, ARCTECH has not completely 69 characterized the products of agent hydrolysis, and de- velopment of processes to completely destroy Sched- ule 2 compounds in agent hydrolysate has not been completed. Characterization of Effluent and Impact on Human Health and Environment Metal parts will be decontaminated to a 3X condi- tion and sent to Rock Island Arsenal for further treat- ment. This will pose no unusual threat to human health and the environment. The process is likely to produce large solid and aque- ous waste streams that will be stabilized and shipped to a hazardous-waste landfill. Although similar waste streams are routinely handled in a manner that poses no threat to human health or the environment, a waste stream this large (about 1,200 tons per day) is likely to be viewed negatively by the public and by regulators. Completeness of Effluent Characterization The experimental results show that parts can be cleaned to 3X condition. Other effluents, especially those containing agent hydrolysate, are still poorly characterized. Effluent Management Strategy Sending the 3X decontaminated metal parts to Rock Island Arsenal is an acceptable procedure. Stabilizing a large aqueous waste stream and disposing of the sta- bilized waste in a landfill may raise concerns among regulators and the public. Resource Requirements The amount of water and hydrogen peroxide could potentially be large. Nevertheless, the projected amount of hydrogen peroxide is expected to be available from commercial sources. Environmenta/ Compliance and Permitting This process could potentially produce large amounts of stabilized aqueous effluent and humic acid solids, which could be problematic when applying for
70 ALTERNATIVE TECHNOLOGIES FOR DEMILITARIZATION OF ASSEMBLED CHEMICAL WEAPONS a permit. Also, if the technology provider indicates that unusually large amounts of hydrogen peroxide will be required, the permitting issues surrounding the trans portation of this material could be significant. STEPS REQUIRED FOR IMPLEMENTATION In order to develop and operate a full-scale ACTODEMIL system, the following steps will have to be taken: 1. Verify the advantages of using humic acid and complete the fundamental definition of the process. 2. Establish the quantity and concentration of hy- dro~en peroxide necessary to complete the de for providing access to agent, energetics, and metal parts for treatment. Finding AR-2. Hydrolysis is anticipated to be effec- tive for destroying chemical agents. Finding AR-3. Destruction of Schedule 2 compounds in the product streams from agent hydrolysis using Fenton's reagent or hydrogen peroxide has not been demonstrated. Finding AR-4. The presence of humic acid during the hydrolysis of agent appears to have little or no effect on agent destruction, with the possible exception of decreasing the amount of gaseous products formed. Finding AR-5. Hydrolysis is expected to be effective ----lo--- a--------- ----------a -- - ----a---- ---- --- r ~ for destroying energetic materials because of the long struct~on of hydrolysis products. 3. Run repeated hydrolysis reactions using recycled residence times. reaction liquor to determine the amount of reac tion liquor that can be recycled under steady-state conditions. A-. Establish a water balance based on previous steps. Then, establish procedures and prepare mass bal ances for the disposal of precipitated humic acid, aqueous effluent, and the stabilized or solidified waste stream. 5. Perform larger scale demonstrations of all steps. To date, only energetics hydrolysis has been dem onstrated at a substantial scale. Then, operate an integrated system that includes all steps of the process. FINDINGS Finding AR-1. The use of the Army's baseline disas sembly method, with minor modification as pro posed by ARCTECH, is anticipated to be effective Finding AR-6. ACTODEMIL requires hydrogen per- oxide in unknown but potentially large quantities to destroy the products from the hydrolysis reactions. This is expected to lead to large quantities of aqueous wastes, which will be stabilized/solidified and, there- fore, will generate a large solid-waste stream (almost certainly RCRA hazardous) that will have to be dis- posed of in a landfill. This could cause extensive per- mitting delays. Finding AR-7. The technology provider claims that humic acid preferentially binds with the by-products of the destruction reactions, allowing recycling of the re- action liquor. This claim was not substantiated by the data provided by the technology provider. Finding AR-X. The ACTODEMIL technology for chemical weapons demilitarization is immature, and will require a good deal more to be done to define and demonstrate the process.
