4
Catalytic Extraction Process Technology

Process Description

Catalytic Extraction ProcessingTM (CEPTM) is a proprietary technology patented by its developer, Molten Metal Technology, Inc., and licensed to M4 Environmental L.P. for specified U.S. governmental applications. 1 M4 Environmental L.P. joined with several other firms to prepare the submission on CEP in response to the Army request for information on alternative technologies.2 Hereafter in this chapter, M4 Environmental LP. and its supporting firms will be referred to as the technology proponent company (TPC). In addition to processing of HD and VX, the submission included processing of the steel ton containers and all dunnage generated in the course of demilitarization operations at the two sites. Destruction of HD and VX by CEP is accomplished in a series of unit operations after the ton containers have been opened and the contents transferred to interim storage tanks.

CEP has been designated by the U.S. Environmental Protection Agency (EPA) as a nonincineration technology. The distinction between incineration (or combustion) and CEP is based upon reaction mechanisms as well as end products. Combustion, which occurs by means of a series of gaseous, reactive intermediates (free radicals), requires high temperature, intimate mixing, adequate residence time, and excess oxygen to achieve high destruction efficiency. CEP, by contrast, is conducted mainly within a molten metal bath at high temperature and low oxygen potential. The products of combustion are in high oxidation states (e.g., CO 2, H2O), whereas products of CEP are in reduced states (e.g., CO, H2).

Technology Overview

A CEP reactor, which is called a catalytic processing unit (CPU), contains a bath of molten metal, typically iron or nickel. For treating chemical warfare agents, the TPC has decided that two CPUs are required. Each CPU is a steel pressure vessel containing a molten metal bath and an optional slag or flux cover. In CEP, these reactors are typically operated in the temperature range of 1425°C to 1650°C (2600°F to 3000ºF). The vessel is lined with refractory materials selected to provide thermal insulation and resistance to corrosion, erosion, and penetration by components of the bath. An electric induction coil, embedded within the refractory lining surrounding the metal bath, provides the energy to melt the metal charge and maintain the temperature of the bath during processing. The CPU headspace, which is several times the height of the molten metal bath, provides physical space to allow disengagement of the offgas from the molten metal and slag. One or more tapping ports through the vessel sidewall allow recovery of metal and slag phases with minimal interruption of operation. One CPU is fitted with a side chamber that can be heated by its own induction coil to melt ton containers. The molten metal flows from the side chamber into the main bath of the CPU. The TPC plans to feed dunnage, placed in steel containers, directly into the metal bath.

The feed material and the cofeeds of oxygen and methane can be injected into the molten metal bath either through a lance entering the top of the bath or through one or more bottom-entering tuyeres. (The TPC has used top-entering lances in numerous bench-scale CPUs.) A tuyere consists of three concentric metal tubes cast into a removable refractory block that is bolted into the bottom of the CPU. The TPC proposes using the tuyere injection of liquid agent and cofeed gases for chemical demilitarization.

Feed material, which may be liquid, gas, finely divided entrained solids, or a pumpable slurry, is metered, mixed, and pumped through the central tube of the

1  

M4 Environmental L.P. is a 50/50 limited partnership of a subsidiary of Lockheed Martin and a subsidiary of Molten Metal Technology, Inc.

2  

The other firms participating in the submission are Bechtel National, Inc., Fluor Daniel, Inc., and Battelle Memorial Institute.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 25
--> 4 Catalytic Extraction Process Technology Process Description Catalytic Extraction ProcessingTM (CEPTM) is a proprietary technology patented by its developer, Molten Metal Technology, Inc., and licensed to M4 Environmental L.P. for specified U.S. governmental applications. 1 M4 Environmental L.P. joined with several other firms to prepare the submission on CEP in response to the Army request for information on alternative technologies.2 Hereafter in this chapter, M4 Environmental LP. and its supporting firms will be referred to as the technology proponent company (TPC). In addition to processing of HD and VX, the submission included processing of the steel ton containers and all dunnage generated in the course of demilitarization operations at the two sites. Destruction of HD and VX by CEP is accomplished in a series of unit operations after the ton containers have been opened and the contents transferred to interim storage tanks. CEP has been designated by the U.S. Environmental Protection Agency (EPA) as a nonincineration technology. The distinction between incineration (or combustion) and CEP is based upon reaction mechanisms as well as end products. Combustion, which occurs by means of a series of gaseous, reactive intermediates (free radicals), requires high temperature, intimate mixing, adequate residence time, and excess oxygen to achieve high destruction efficiency. CEP, by contrast, is conducted mainly within a molten metal bath at high temperature and low oxygen potential. The products of combustion are in high oxidation states (e.g., CO 2, H2O), whereas products of CEP are in reduced states (e.g., CO, H2). Technology Overview A CEP reactor, which is called a catalytic processing unit (CPU), contains a bath of molten metal, typically iron or nickel. For treating chemical warfare agents, the TPC has decided that two CPUs are required. Each CPU is a steel pressure vessel containing a molten metal bath and an optional slag or flux cover. In CEP, these reactors are typically operated in the temperature range of 1425°C to 1650°C (2600°F to 3000ºF). The vessel is lined with refractory materials selected to provide thermal insulation and resistance to corrosion, erosion, and penetration by components of the bath. An electric induction coil, embedded within the refractory lining surrounding the metal bath, provides the energy to melt the metal charge and maintain the temperature of the bath during processing. The CPU headspace, which is several times the height of the molten metal bath, provides physical space to allow disengagement of the offgas from the molten metal and slag. One or more tapping ports through the vessel sidewall allow recovery of metal and slag phases with minimal interruption of operation. One CPU is fitted with a side chamber that can be heated by its own induction coil to melt ton containers. The molten metal flows from the side chamber into the main bath of the CPU. The TPC plans to feed dunnage, placed in steel containers, directly into the metal bath. The feed material and the cofeeds of oxygen and methane can be injected into the molten metal bath either through a lance entering the top of the bath or through one or more bottom-entering tuyeres. (The TPC has used top-entering lances in numerous bench-scale CPUs.) A tuyere consists of three concentric metal tubes cast into a removable refractory block that is bolted into the bottom of the CPU. The TPC proposes using the tuyere injection of liquid agent and cofeed gases for chemical demilitarization. Feed material, which may be liquid, gas, finely divided entrained solids, or a pumpable slurry, is metered, mixed, and pumped through the central tube of the 1   M4 Environmental L.P. is a 50/50 limited partnership of a subsidiary of Lockheed Martin and a subsidiary of Molten Metal Technology, Inc. 2   The other firms participating in the submission are Bechtel National, Inc., Fluor Daniel, Inc., and Battelle Memorial Institute.

OCR for page 25
--> tuyere at moderately high pressure, less than 10 atmospheres. Oxygen, in stoichiometric proportion to convert all carbon in the feed and the methane cofeed to carbon monoxide, is metered into the next annulus at high velocity to induce turbulence, mixing with the feed stream, and formation of a jet that rapidly breaks up into small bubbles. A small amount of methane is fed through the outer annulus to cool the tuyere. An inert gas is injected automatically into each of the feed lines as needed to make up the difference between the total flow required in each line and the set-point flow of each feed component (agent, oxygen, and methane). During startup and shutdown, the inert gas alone is pumped through all feed lines to prevent molten metal from entering and plugging the tuyere. According to the TPC's description of the process, when feed material is injected into the bath along with oxygen and methane, the molecular entities in the feed material are decomposed by catalysis into their component elements. These elements dissolve in the metal and form intermediates by bonding chemically with the metal. By appropriate selection of process conditions, the dissolved elements with high solubility in the metal (e.g., carbon, sulfur, and phosphorus) can either be retained in the metal bath up to their saturation limit or induced to react with less soluble elements (e.g., hydrogen, oxygen, and chlorine) to form gaseous products—principally H2, CO, HCl, and H2S with minor amounts of H2O, and CO2. These gaseous products then form bubbles, which ascend and exit the bath. According to the TPC, because CEP is carried out at low oxygen potential and decomposes feed molecules to elements regardless of their starting molecular structure, the process provides neither pathways nor precursors for the formation of oxides of nitrogen or sulfur or the formation of dioxins and furans. The TPC has reported that it expects the process residuals from treating VX or HD, the ton containers, and dunnage to be ferrous alloys, aqueous hydrochloric acid, elemental sulfur, and a synthesis gas. The TPC also has reported that markets for the alloys, hydrochloric acid, and sulfur have been identified. The synthesis gas is combusted, along with natural gas, in an on-site gas turbine generator to provide electricity used in the process. A small amount of slag or ceramic (less than 5 percent of total solid product mass) is also produced and must be disposed of as waste. The panel agrees with the TPC that this slag is likely to pass the U.S. Environmental Protection Agency Toxic Characteristic Leaching Procedure (TCLP) test. (Unless it is delisted, however, it could still be classified as hazardous waste because it is derived from agent.) Chemical Demilitarization Process According to the submitted design, chemical demilitarization operations are to be conducted in a central processing building of approximately 13,000 square feet. The building is partitioned into distinct areas by function (Figure 4-1). Precautionary safety measures confine agent to small areas, reduce the possibility of cross contamination, and reduce requirements for heating, ventilation, and air conditioning (HVAC); high efficiency particulate-arresting filters; carbon filters; and agent monitoring equipment. Ton containers are opened in area 100 and, if necessary for interim storage, cleaned to 3X condition. Dunnage from daily operations is compacted and packaged in small metal containers in the same area. The equipment and techniques used to handle ton containers, including the punch-and-drain process, vacuum transfer of agent and decontamination liquids to interim storage tanks, safe airlock passage, cascaded HVAC, double-containment envelopes, and low pressure injection are based on the equipment and techniques used in the baseline system facilities at Johnston Atoll in the Pacific Ocean and at Tooele, Utah. The only significant change is the addition of an aspirated, self-cleaning gland surrounding the punch to mitigate spillage of agent when the container is penetrated. The two CPUs, designated CPU-1 and CPU-2, are located in area 200. The gas handling train (GHT) and facilities for product recovery are located in Area 300. Area 500 is devoted to product gas utilization; products of CEP are stored in area 700; utilities are located in area 800; and area 1000 houses the emergency relief system. The CPUs and the equipment in the product recovery areas are of modular design, which will allow the TPC to use the same CPUs and product recovery equipment at the Aberdeen site to process HD and, afterward, at the Newport site to process VX. For processing either agent, CPU-2 contains molten iron and processes all ton containers and dunnage. Emptied ton containers are fed by horizontal indexing conveyors and coordinated, double-door, cascade-ventilated airlocks to the premelting side chamber of CPU-2. The steel ton containers melt, and the organics, including all remaining gels, solids, and surface agent residuals, are pyrolyzed. Pyrolysis products and molten

OCR for page 25
--> Figure 4-1 Primary agent and residue process flows for a chemical demilitarization CEP facility. Area 700 (product storage), Area 800 (utilities), and Area 1000 (emergency relief system) are not shown. Source: M4 Environmental L.P., 1996b. metal then enter CPU-2 through a side chute above the level of the molten bath. The TPC states that dunnage canisters will be fed directly into CPU-2. If the ton containers are melted as they are emptied, at the proposed processing rate of VX (169 kg/hour) they will add about 725 kg of metal to the bath every 5 hours. This quantity of metal will increase the bath height about 8 cm, necessitating tapping the bath at approximately 10-hour intervals to maintain an optimum level. The metal tap, which will probably be located at the desired bath height, will be opened by heating it to melt the metallic or slag plug. The tap will be closed by cooling it to solidify a metal or slag plug. Different strategies are required for processing HD (Figure 4-2) and VX (Figure 4-3). In the HD strategy, liquid agent is injected by tuyere into CPU-1, which uses a molten nickel bath to reduce the formation and carryover of metal chlorides. Chlorine is released from the bath as HCl. Sulfur from the HD accumulates in the bath to a concentration of about 27 percent, a concentration at which sulfur is released from the bath as HS. The offgas from CPU-2, which originates from processing the ton containers, any residue in them, and dunnage is quenched with water, pressurized, and injected into CPU-1 to ensure complete reaction of any products of incomplete conversion. Product gases from CPU-1 are quenched with water, filtered, and scrubbed with water to recover aqueous HCl. At this point, the offgas consists primarily of H2S, CO, and H2. The H2S is subsequently converted to elemental sulfur using the commercial SulFeroxTM process. The remaining gases, principally H2 and CO, form the synthesis gas, which is pressurized and stored in one of three

OCR for page 25
--> Figure 4-2 High level block diagram for the destruction of HD by CEP. Source: M4 Environmental L.P., 1996b. tanks with a capacity of 4 m3 each. After a filled tank has been analyzed for agent and other toxic, the gas is combusted in a gas turbine electric generator. In the VX strategy, CPU-2 is the primary reactor for processing agent. Both sulfur and phosphorus from the VX are held in solution in the molten iron and recovered as an Fe-S-P alloy when CPU-2 is tapped to control the bath level. The offgas from CPU-2 is conditioned as described above for HD and injected into CPU-1, which in this case contains an iron bath and functions as a polishing reactor to ensure the destruction of remaining agent or other organics. The offgas from CPU-1 is quenched with water and filtered to yield the synthesis gas of CO and H2. Trace amounts of HCN in the product gas are decomposed by catalysis to H2, N2, and carbon. The VX strategy uses the same approach as the HD strategy for storing and analyzing the synthesis gas prior to combustion. In both treatment strategies, aqueous cleaning and decontamination solutions, including particulates and condensates recovered as water-base slurries from cooling and cleaning the CPU offgases, will probably be injected into CPU-2 for destruction, so that all slag-forming components are kept in the same CPU. Slag formed by the interaction of debris entering with the emptied ton containers, lime-based decontamination solutions, and dunnage can be removed in the same way molten metal is removed. Should the need arise, the facility design includes the capability of opening a ton container with a high pressure water-jet containing abrasive particles. A water spray then removes the gels, residues, and remaining agent, and calcium-based decontamination solution is used to clean the container to 3X condition. The resulting finely divided aqueous slurry can be removed from the cleaning area by aspiration, transported by vacuum pumping to temporary storage, and injected into one of the CPUs for processing to the same residuals as other cleaning solutions and slurries. The use of a water-jet, of course, would require suitable enclosure and capture/treatment of effluent from the spray operation. If a situation arises in which liquids or gases from vessels, piping, or either CPU are vented by means of pressure relief devices, the facility design includes standby equipment to quench the vented material and absorb acid gases. Any residual agent or HS is combusted in a standby boiler prior to releasing the gaseous residual to the atmosphere. Scientific Principles The TPC and the developer of CEP describe the molten metal bath as a dissociation catalyst for molecular entities in feed materials, a solvent for elemental

OCR for page 25
--> Figure 4-3 High level block diagram for the destruction of VX by CEP. Source: M4 Environmental L.P., 1996b. fragments, and a medium for product synthesis. The TPC divides the process conceptually into stages comprising catalytic dissociation of the feed, formation of elemental intermediates with the solvent metal, product synthesis by interaction of elemental intermediates, and partitioning of products among metal, slag, and gas phases. A recent publication by technologists who work for the developer of CEP states, "the CEP unit is not acting as a thermal treatment device in that temperature is not the primary means to change the physical and chemical composition of the feed material . . ." (Nagel et al., 1996, p. 2158). The above description does not address initial thermal and gas-phase reactions in the overall sequence of events between the introduction of feeds and the release of final products. Although bench-scale tests of the process have demonstrated that the process can destroy agent as required by the Army, analysis by the AltTech Panel indicates that the actual conditions are probably more complex than this description implies. The panel's review indicates that a complete description of the scientific principles underlying CEP requires discussion of several additional phenomena, including gas-phase reactions among agent, oxygen, and methane in the inlet jet immediately following tuyere injection; interactions of these gases and intermediate products with metal vapor inside bubbles; and boundary reactions between bubble components and the surrounding metal. Accordingly, the following discussion attempts to provide a more detailed description of the probable scientific principles and further develops details of the probable processes involved. The TPC notes that the submitted design reflects many years of experience in the steel industry with injecting gases into molten steel baths by the use of similar tuyere inlets. However, experience in the steel industry relates primarily to the injection of gases for the purpose of changing the composition of the bath. The escape of a small surplus of these gases from the bath surface is of little concern other than as an economic loss. Thus, there is no long-established precedent from industrial experience for the complete reaction of injected gases with a molten metal bath to the very low-level of residuals required for agent destruction. The panel is not aware of industrial experience with injecting liquids into a molten metal bath. Dissociation and Reaction of Tuyere-Injected Materials In the CEP, a liquid agent or other feed to be destroyed, inert carrier gas, oxygen in stoichiometric proportion to oxidize all carbon in feeds and cofeeds to CO,

OCR for page 25
--> and methane are injected by tuyere at moderately high pressure (less than 10 atmospheres) and high velocity into the molten metal bath. The injected materials form a jet that extends several tuyere diameters into the bath. The high velocity of the oxygen gas stream causes turbulence and contributes to entrainment of metal vapor and droplets within the jet. These effects of the initial momentum quickly dissipate, and the jet breaks into bubbles that rise through the molten metal because of their buoyancy. Subdivision of larger bubbles increases the total surface-contact area and increases the collision frequency between gas molecules and the molten metal. As the bubbles rise to the surface, they continue to change in size for several reasons. They tend to increase in size as the ferrostatic head decreases; they tend to decrease as gaseous intermediates are absorbed into the molten metal; and they tend to increase as product gases released from the molten metal migrate back into them. Some very small bubbles may also form through the nucleation of gases produced in the molten metal and then grow as they agglomerate with other bubbles or accumulate more gas released from metal. Radiant heat transfer from the hot metal to the aspirated liquid droplets and gas bubbles is extraordinarily rapid at the high temperature of the bath because the rate of radiant heat transfer is proportional to the fourth power of the absolute temperature. For example, a hypothetical sphere 100 µm in diameter will receive energy at 1600°C at the rate of 5 × 10-3 calories per second, which is sufficient to vaporize a like volume of liquid agent and heat the resultant vapor, as multiple 100-µm bubbles, to 1000°C in less than 50 milliseconds. The panel's judgment is that partial degradation of agent and gas-phase reaction between agent or agent fragments and oxygen is very likely under these circumstances. A significant fraction of the feed probably undergoes partial oxidation, and the products of partial oxidation then interact with the molten metal to form intermediates. The panel also concludes that oxidation is probably not complete and should not be termed combustion, even though reactions proceed stepwise by molecular collisions among gas-phase intermediates. Increasing the effective pressure of the bubbles increases the gas density and therefore the collision frequency between bubble contents and the molten metal. Thus, increasing the operating pressure of the CPU or increasing the bath depth increases the rates of reactions in the bubbles. The TPC has ascertained that the processing rate for a given reactor increases significantly with an increase in operating pressure. An important issue is whether there is opportunity for back reactions to form complex organic compounds from intermediates. The assumption that the opportunity is negligible is important to the TPC's statement that no detectable recombinant dioxins or furans are produced. However, it is possible and thermodynamically feasible to produce HCN in the conditions of the CEP bath when processing VX. In the original submission from the TPC, the inert gas was specified to be nitrogen. The TPC has subsequently considered using argon for this purge/make-up gas. For processing HD at least, using argon instead of nitrogen would resolve the issue of HCN formation by removing any source of nitrogen. Although the extent of HCN production can be controlled to very small concentrations, the fact that it does occur indicates that the claim that no detectable recombinant dioxins or furans (i.e., complex compounds) are produced does not apply to simple compounds like HCN. Dissolution kinetics are also important to the formation of intermediates. For example, hydrogen is sparingly soluble in molten iron, and when organic compounds containing hydrogen are injected into molten iron, hydrogen gas evolves from the bath while the carbon dissolves in the metal. It is also reasonable to expect that the initial bubbles formed by the break-up of the jet contain H2. (If nitrogen were used as the inert make-up gas, N2 would also be a significant component of the initial bubbles.) Catalysis by the Bath and the Formation of Intermediates There is ample evidence in the peer-reviewed literature to support the TPC's position that the molten metal bath serves as a true catalyst by decreasing the activation energy for dissociation of organic molecules, participating in the formation of intermediates, and increasing the efficiency of product formation without itself undergoing change (Satterfield, 1991). Given the formation of intermediates, their relative solubilities in the metal are another factor to consider, particularly for the VX strategy, in which some elements are to be retained in the bath while others exit as offgas. The panel estimated the solubility of VX components in the bath and the time required to saturate the bath under processing conditions of 1600°C and the proposed feed rate (Table 4-1). Columns 2 and 3 list the saturation solubility (in parts per million by weight) and the total weight of elements in the bath, based on a reasonable

OCR for page 25
--> TABLE 4-1 Calculated Solubility of VX and Cofeed Elements in Iron at 1600°C and Time to Saturate the Iron Bath at Processing Conditions   Solubility in Bath     Element ppm kga Feed Rate kg/hb Time to Saturate Bath h C 54,000c 442 87.4 5.05 H2 25d 0.20 17.8 0.011 P2 110,000c 892 19.6 45.5 O2 1,290c 10.6 130.7 0.081 S2 110,000c 892 20.3 43.9 N2 88d 0.14 8.9 0.016 Notes a Bath assumed to contain 8,163 kg iron; contribution of dissolved elements was not considered. b Feed rates: 169 kg/h VX agent; 110 kg/h oxygen; and 5 kg/h methane. c From Massalski, 1986, pages 842 (C), 1746 (P), and 1762 (S). d From Rao, 1985, pages 438 (H2) and 463 (N2). assumption of the partial pressures of the gases derived from the feeds. Column 5 lists the time required to saturate the bath at the elemental feed rate given in column 4, which is derived from the molecular composition of the feed and cofeeds and their feed rates. These values are only computational estimates; numerous simplifying assumptions were needed, and interactions among bath components were ignored. However, the calculations do illustrate the following points. Bath Saturation Point for Retained Elements. Because the solubilities in molten iron of carbon, phosphorus, and sulfur are significant, amounting to 5.4, 11, and 11 wt pct, respectively, considerable time is required to saturate the bath with these elements. The TPC's strategy for VX calls for controlling the release of phosphorus and sulfur gases (preventing breakthrough) by keeping the bath below saturation. The strategy is to remove alloyed bath metal at intervals by tapping, while adding molten iron by processing ton containers. Once the bath reaches saturation for phosphorus or sulfur, the ton containers must be processed at a rate sufficient to supply enough new iron to alloy all the phosphorus and sulfur in the agent feed. The calculated values in column 2 of the table indicate that the amount of iron in a ton container, 636 kg, will dissolve only about 69 kg of sulfur and a similar quantity of phosphorus. The 682 kg of VX within a ton container contains about 82 kg of sulfur and 79 kg of phosphorus. Although these calculations are based on numerous simplifying assumptions, they indicate that synchronizing the addition of iron to the bath with the agent feed rate will be critical in avoiding the breakthrough of sulfur and phosphorus into the offgas. In particular, these computations indicate that the TPC's suggestion of stockpiling ton containers for treatment at a later date while processing VX is not an option unless there is a significant alternative iron feed. Hydrogen and Nitrogen. The solubilities of hydrogen and nitrogen in molten iron are extremely low, and Table 4-1 suggests that the bath will become saturated with these elements in less than 1 minute. Although the bath, when in continuous operation for processing VX, is likely to be saturated with hydrogen and nitrogen, the kinetics indicate that significant proportions of hydrogen and nitrogen in the feed may not pass through metallic intermediates but may form gas bubbles directly. Supersaturation of the bath as a whole with these and other sparingly soluble elements is likely because the feed materials are introduced into the bath at the bottom, where the ferrostatic head is greatest.

OCR for page 25
--> Oxygen. The solubility of oxygen in molten iron is much greater than hydrogen or nitrogen but far less than carbon, sulfur, or phosphorus. The calculated time of less than 5 minutes for the bath to become saturated reflects the high feed rate. The solubility of oxygen favors the formation of an iron-oxygen intermediate. These calculations indicate all components in the feeds and cofeeds are soluble enough to support the TPC's description of the formation of elemental intermediates. Given the formation of elemental intermediates, product synthesis can occur by chemical reaction among those intermediates. Partitioning of Products among Metal, Slag, and Gas Phases To some extent, the process residuals from CEP can be customized by adding appropriate cofeeds or controlling operating conditions. As noted above, the design specifies that oxygen cofeed is provided in stoichiometric proportion to convert carbon in the feed material and the methane cofeed to CO at the desired carbon concentration and temperature of the bath. The oxygen stoichiometry determines the ratio of CO to CO2 in the product gas, and this ratio is monitored as a process control on the oxygen feed rate. Hydrogen appears as H2 in the product gas because the oxygen potential in the bath is less than the potential required to form significant amounts of H 2O. Similarly, SO2 and NOx formation are thermodynamically unfavorable. For processing HD, sulfur can be recovered in the gas-phase by allowing sulfur in the bath to increase to a saturation concentration above which the formation of H2S from H2 and the Fe-S intermediate is thermodynamically favored. Or, sulfur can be recovered as an alloy element by tapping bath metal from the CPU before the saturation concentration is reached, as the TPC proposes to do for processing VX. The chemistry of phosphorus, although more complicated, is similar in that phosphorus can be obtained as an iron alloy by tapping the metal before the saturation concentration is reached. The panel notes, however, that although CEP has been performed extensively with iron baths containing carbon, sulfur, and chlorine, to the panel knowledge it has not been performed with iron baths containing phosphorus in addition to carbon and sulfur. Metals such as aluminum, calcium, and silicon that form oxides that are more stable than CO at the operating temperature will be oxidized and will accumulate in the slag phase (as Al2O3, CaO, and SiO2, respectively). Cofeeds may be required to ensure the slag is sufficiently fluid. For example, silica and lime are appropriate cofeeds if the feed material contains appreciable aluminum or alumina. Metals whose oxides are less stable than CO will either accumulate in the molten metal (Co, Cr, Cu, Ni, Mn) or exit the bath as vapor (Cd, Pb, Zn). Iron is the preferred bath metal for processing VX. However, if iron were used to process HD, there would be substantial formation and carryover of FeCl2 vapor, which would form a dust in the downstream systems, requiring a more extensive dust removal strategy than the particle filters included in the current design. The use of a nickel bath for processing HD reduces this problem because NiCl2 is less stable than HCl and does not form to a significant extent. Nearly all of the chlorine from the HD forms HCl and is recovered in the aqueous scrubber. Under the same processing conditions, a nickel bath will become saturated with sulfur in about the same time as an iron bath of equal mass and will become saturated with carbon in less than half the time of an iron bath. Process Modeling The most important consideration to the panel, in light of the short residence time of bubbles in the bath, is whether agent or significant fragments of agent can avoid decomposition by remaining in or migrating to a bubble and passing unreacted through the bath. An analysis of the probability and consequences of the requisite reactions at the molecular level would involve complicated computations dependent on numerous assumptions. Instead, it is customary in such circumstances to use engineering models that work from both basic principles and experimental data to provide an approximation adequate for design purposes. The TPC has done extensive experimentation and modeling to understand bubble formation, break-up dynamics, and the operating limits of CEP performance. The models used by the TPC indicate that the process depends heavily on three factors: (1) bubble size, with the critical largest-bubble diameter being on the order of a fraction of an inch (the actual size is proprietary); (2) residence time, with the typical single-path residence time being a fraction of a second (actual time is proprietary); and (3) an energy dissipation term that reflects the degree to which metal vapor and droplets inside the bubbles increase the gas-metal contact.

OCR for page 25
--> Although these models were developed and used by the TPC, the panel did not review or evaluate them in detail for this report. Rather, the panel has relied upon the TPC's representations that the model results correlate well with the very high DRE (destruction removal efficiency) values that were achieved in the experimental and commercial-scale demonstration reactors to which the models were applied. The TPC has stated that it intends to use a residence time that provides a design safety factor of at least 10 to assure the destruction of VX or HD agent to at least the required six 9's DRE (99.9999 percent). Conclusions on the Underlying Science The TPC's explanation of CEP performance is based upon accepted free energy principles.3 The panel believes the engineering design models used to design the system have been based upon solid scientific data. The panel did not, however, review these models in detail. The TPC' s original submission did not include equipment for holding the synthesis gas until analysis had ensured the complete destruction of agent or other toxic components prior to combusting the gas in a gas turbine or using it in some other way. However, in response to the concerns of communities near the storage sites, the TPC has subsequently changed the design to include three 4-m3 storage tanks, in parallel, in the synthesis gas line prior to the gas turbine. Each tank has the capacity to store 15 minutes of anticipated output of synthesis gas pressurized to 20 atmospheres, gauge (300 psig). This storage capacity allows the synthesis gas to be analyzed before it is used as a fuel and the emissions are released to the atmosphere. The proposed design for a chemical demilitarization facility is undergoing continuous development as the TPC accumulates operating experience in other applications. The opinion of the AltTech Panel is that the process is adequately understood and satisfactorily engineered at this time to process either HD or VX successfully and safely, when operated properly, to meet the required six 9's DRE. Technology Status The information available to the panel on CEP operational units is summarized in Table 4-2. As of early 1996, the TPC reported more than 15,000 hours of molten metal test experience with its reactors. Much of this experience was in tests on the 10 to 15 bench-scale units at the TPC's Fall River site. The nominal bath size of these units is 4 to 9 kg. Fall River Demonstration Unit The Fall River Demonstration Unit (Demo Unit) is the largest operational CPU. As of April 1996, the longest period of continuous, commercial-scale operation in this unit while processing liquid or gaseous organics was 120 hours, during which 1,680 kg of feed was processed. The associated on-stream factor was between 50 and 80 percent, depending on experimental requirements.4 The TPC plans to use an on-stream factor of about 82 percent for the CPUs for destroying HD and VX at Aberdeen and Newport. The TPC also reports that the Demo Unit was used to demonstrate the long-term operability, reliability, and product performance of CEP as a contractual milestone prior to an agreement with a major chemical manufacturer to build a commercial facility. The 93-hour test included a switch-over from injecting solid feed material (biosludge) to injecting heavily chlorinated liquid organic material (RCRA waste F024). The TPC reports that the results of this test surpassed more than 40 performance criteria (for environmental protection, product quality, reliability, operability, feed injection, etc.) established by the customer, Hoechst Celanese. The reported test results included an on-stream factor up to 90 percent, mass balance closures at 100 percent, and feed injection rates that met commercial-operation requirements. The TPC reported that steady-state operational requirements were met and surpassed (validated by on-site customer evaluations), as demonstrated by the steady-state production of high-quality synthesis gas that met the customer's on-site recycling requirements. 3   The panel wishes to thank Dr. Nev A. Gokcen, former supervisor (retired), Thermodynamics Laboratory, Albany Research Center, Bureau of Mines, for his help in discussing the applicability of the free-energy equations used by the TPC as taken from Table C-3 (p. 892) of Stoichiometry and Thermodynamics of Metallurgical Processes (Rao, 1958). The text identifies the equations as the "standard free energy change between the Raoultian to the 1-wt.% standard state." 4   The on-stream factor, or availability, is defined for this chapter as the number of days per 360-day year a facility is fully operational.

OCR for page 25
--> TABLE 4-2 Status of CEP Units from Bench Scale to Commercial Scale a Location Reactor Units Nominal Metal Bath Size (kg molten metal) Development Scale Comments Fall River, Massachusetts 10-15 CPUs 4-9 bench Much of TPC's bath operating experience is with these experimental units.   APU-10 450 pilot Repeated continuous runs of >100 hours each. Tuyere injection of liquid chlorinated organic feed.   Variable Pressure Reactor 68 pilot Demonstrated hot metal operation for >700 hours. Automated heating to maintain bath temperature.   Demo Unit 2,700 commercial size Used for demonstrating CEP at commercial-scale. Quantum-CEP Oak Ridge, Tennessee RPU-1 45 bench Used for depleted uranium hexafluoride. Panel observed unit in operation.   RPU-2 (2 units) ~9 (per unit) bench Used for treatability studies. Panel observed unit in operation.   RPU-3 450 pilot Has performed more than 15 small-scale tests and a 27-hour pilot-test.   RPU-4 "Combo" 1,360 commercial size Bath size expandable to 3,200 kg. Under construction for summer 1996 startup. To be used to demonstrate CEP at commercial-scale. SEG-Q-CEP Oak Ridge, Tennessee 2 units up to ~900 commercial size For batch-mode volume reduction of radioactive ion-exchange resins. Processed >27,000 kg of resins as of May 1996. a Table data based on information from Valenti, 1996, and M4 Environmental L.P., 1996b. Oak Ridge Facilities The Quantum-CEP reactor units at the TPC's Oak Ridge site are referred to as RPUs (radioactive processing units). Members of the AltTech Panel observed the bench-scale units at Oak Ridge in operation during site visits. The SEG/Quantum-CEP units are located at a separate site in Oak Ridge and are designed for batch-mode commercial operations. Each campaign will consist of a 36-hour startup, 3 to 5 days of injection of radioactive ion-exchange resins, and a 36-hour shutdown, for a total campaign duration of 6 to 8 days. During the panel's site visit in March 1996, the SEG facility at Oak Ridge was still in scale-up activities using nonradioactive resins, prior to commercial operation. As of May 1996, the facility was reported to have processed more than 27,000 kg of ion-exchange resins. The TPC reported that a peak throughput rate of 150 percent of design had been achieved and that equipment upgrades were being made.

OCR for page 25
--> Agent Testing Battelle/Columbus Laboratory (a member of the team that prepared the TPC submissions) has tested agent destruction in a bench-scale CEP unit. The TPC has issued a news release reporting a "destruction percentage" of eight 9's (99.999999 percent) for processing HD and VX (M4 Environmental L.P., 1996a). From the AltTech Panel's preliminary review of the full report on these tests, the panel concludes that the tests demonstrated that the CEP technology can destroy agent to at least the six 9's DRE required by the Army. Further implications of the test results for a full-scale operation are discussed below in the section on Scale-Up. Summary of Technology Status The development of the various subsystems required for a chemical demilitarization facility has been demonstrated by successfully injecting feed materials, generating process products, and achieving high on-stream factors at developmental facilities. A wide range of materials has been processed, including polystyrene with graphite, ion-exchange resins, acetone, industrial biosolid waste, chlorotoluene with heavy organics, chlorobenzene, fuel oil with chlorotoluene, dimethyl acetamide with heavy organics, benzonitrile, diazinon, diazinon with sulfur, and surplus metal components. These materials have been in various physical forms, including liquids, slurries, fine solids, and bulk solids. Various feed-addition systems, including configurations with a top-entering lance or a bottom-entering tuyere, have been studied. Successful tuyere injections of liquids, slurries, and fine solids have been demonstrated in which the injection rates and the reactor design were optimized for steady-state operations. Injection rates comparable with commercial levels have been demonstrated at both the demonstration-scale and advanced processing units. Bulk additions of metal components, scrap metals, and wood have been demonstrated at feed rates comparable to commercial- scale and with successful conversion of materials. The TPC's design for processing bulk solids uses two reactors. The receiving unit includes a premelting chamber for melting and volatilization. The second unit is used to polish the offgas from the first unit. Panel Summary of Technology Status As of May 1996, the TPC has accumulated considerable test experience with CEP technology, as described above, and is gaining commercial experience. However, the TPE does not yet have extended, continuous commercial experience with CPUs of commercial size. Process Operation Process Description The TPC provided the following process diagrams, which will be referred to in this and subsequent sections as needed: Block flow diagram for CEP facility (Figure 4-4) CEP process flow diagram for VX feed injection system into CPU-2 with premelting chamber for ton containers (Figure 4-5) CEP process flow diagram for VX CPU-2 offgas treatment (Figure 4-6) CEP process flow diagram for VX CPU-1 gas handling train (Figure 4-7) CEP process flow diagram for VX relief system (Figure 4-8) CPU block diagram and material balances for HD treatment (Figure 4-9) CPU block diagram and material balances for VX treatment (Figure 4-10) CEP heat and material balances for VX gas handling (Table 4-3) Agent Detoxification Residual Agent Based on tests using HD, VX, and agent surrogates as CEP feed materials, the TPC anticipates a DRE for each agent in excess of six 9's (99.9999 percent). If, as the result of equipment failure, operator error, or some other circumstance, residual agent remains in the synthesis gas emerging from the gas handling train (see Figure 4-7), it can be detected in the hold-up tanks before the gas is released to the energy recovery system for combustion. If analysis of a tank detects the presence of agent above the six 9's DRE limit, the contents can

OCR for page 25
--> will have double containment walls. The space between the primary and secondary containment walls will be monitored with DAAMS tubes, which will enable maintenance personnel to identify and repair leaking valves, fittings, etc., in the primary containment before the leak allows agent to escape the second containment. The TPC plans to develop a maintenance control document as part of the detailed design phase. This document will include equipment maintenance schedules; parts lists for routine maintenance; lubrication requirements for each item of equipment; and maintenance procedure summaries specifying the frequency, purpose, references, prerequisites, and listings of all tasks and reviews. The documents will also include an instrument index and spares list, as well as preventive maintenance procedures for instruments, and will serve as a source book for miscellaneous maintenance items required for startup. Software will be used to record maintenance schedules and provide daily reminders and reports. Operations and Maintenance Operational Safeguards All important variables such as temperatures, pressures, flow rates, and levels are measured, recorded, and alarmed throughout the system. Critical controls are provided with automatic alternatives if there is a safety risk or the possibility of damage to equipment. In areas of the plant that handle agent, the interstitial space in double-walled piping and equipment will be continuously monitored for agent, as a means of detecting leaks in the primary containment. In the gas handling train, the quench water source has assured backup water sources, such as the firewater system. The backup water source ensures that hot offgas from the CPUs is cooled to prevent damage to the gas handling train. The entire system is designed for operation via remote instrumentation, controls, and video cameras from a control center separate from the central building. The architecture of the DCS uses a centrally integrated executive protocol, which includes an emergency process-shutdown that is hard-wired and completely independent of the control computers and requires no human intervention. The plant design adheres to approved safety principles for operations involving hazardous chemicals, including the following: All operations are designed to keep agent and agent-contaminated fluids inside the ton container, storage tank, or process piping at all times. Agent and agent-contaminated fluids are transferred from the collection point to nearby storage tanks by vacuum pumping techniques. The capacity and number of storage tanks for agent and agent-contaminated fluids are set to the minimum needed for the design throughput. Each tank is contained within a separate cell, and all cells are located together in the same area. Pumps for pressurizing the agent feed are located as close to the reactor as possible to minimize the length of piping that conveys pressurized agent to the CPU. The pump pressure is as low as possible consistent with maintaining reliable feed conditions under all operating conditions. Liquid agent and agent-contaminated fluids are transferred only through double-wall piping. The annulus is purged continuously with inert gas and monitored to detect the presence of agent. Pipes and ducts are welded and fully inspected. Bolted and sealed connections are used only where they are essential. In the event of a transfer-pump failure, agent or agent-contaminated fluid in the piping drains back into the source tank. All agent-involved pipes are sized and routed to allow unimpeded flow and minimize the chance of contamination traps. All components involved in pumping, storage, or piping of agent are mounted to be readily accessible for corrective maintenance and area housekeeping by personnel wearing appropriate safety gear. The areas around the CPUs are designed for convenient and secure access and are maintained at ambient temperature, to permit immediate emergency response via multiple routes for personnel in full protective clothing. The central building is partitioned in such a way that air monitors placed throughout the process areas can detect and verify agent leaks quickly and effectively. Failure and Hazards Analysis The TPC has performed several hazard and operability studies of CEP technology for the demonstration and

OCR for page 25
--> commercial facilities described above. In addition, the TPC contracted with a third party to perform a hazard analysis specifically to support its submission for the chemical demilitarization program (M4 Environmental L.P., 1996e). This analysis, which used a failure modes and effects analysis (FMEA) approach, identified 1,129 failure events. Of these, 17 unique events for both facility sites were assigned a risk assessment code of 2, indicating that the risk was not acceptable. None of these code 2 risks involved exposure to chemical agent, and only one involved personal injury. The remaining 16 involved only a possible loss of processing capability because of damage to critical components in the gas handling train. The TPC plans to conduct additional safety and hazards reviews during the design, engineering, and facility commissioning phases of development. The TPC states that, for these reviews, it will use methodologies and techniques developed by E.I. DuPont de Nemours and Company, Imperial Chemical Industries, and the Chemical Process Safety Institute that meet or exceed the requirements specified in the Occupational Safety and Health Administration regulations, Process Safety Management of Highly Hazardous Chemicals (29 CFR 1910.119). The TPC also plans to implement a comprehensive health and safety program to establish best practices for ensuring safety. These practices include emergency response plans, plans for communicating information on chemical and radiological hazards, ALARA (as low as reasonably achievable) review procedures, safety training requirements, procedures for change management, and standard industrial safeguards. The TPC intends to document all operational procedures and practices, incident investigation reports, and compliance audits. Maintenance Routine Maintenance Requirements For the feed preparation systems, feed systems, and balance of the plant (Areas 100 and 900), most of the routine maintenance after startup involves checking and adjusting for wear and tear of mechanisms and stops and replacing pressure seals and glands to prevent leakage of fluids and gases. Critical elements of the feed preparation equipment such as the punch tools, the probes for extracting liquid agent, and the water-jet cutting nozzles and cleaning heads need frequent replacement because they have high rates of wear. Because operations at the two sites will be of short duration (about one year each) and the number of process cycles to be completed is fairly low (1,700 ton containers at each site, plus miscellaneous discrete items), the wear on the process equipment should be within acceptable limits. An important aspect of routine maintenance will be calibration of instruments such as the ACAMS (automatic continuous air monitoring system) or MINICAMS. Because both of these instruments are gas chromatographs, they require a significant level of routine calibration and maintenance. The experience of one of the TPC partners in working with the instrumentation at the Tooele Chemical Disposal Facility gives the TPC team experience in setting up and operating a calibration and maintenance program for these and other agent monitoring instruments. Maintenance Manuals and Procedures The TPC provides maintenance manuals and operating procedures for all its operating CEP units. Because the CEP facility for chemical demilitarization is still in the conceptual design phase, no facility-specific manuals or procedures have been developed yet. The TPC plans to develop a project maintenance manual covering preventive maintenance, lubrication, scheduled checks and inspections, cold test plans, and integrated test plans for startup. The manual will be prepared as the detailed design nears completion and will contain detailed procedures, checklists, and valve line-ups. Documented Record of Performance The feed preparation systems, feed systems, and most of the balance-of-plant systems (Areas 100 and 900) use equipment that is the same as or similar to equipment used in the Army baseline incineration system. Records of performance probably exist for this equipment, and one can reasonably assume that similar levels of operation and maintenance will apply when the equipment is used in the proposed CEP system. Downtime Experience Based on the TPC's experience to date, the TPC has allowed for approximately 60 days of maintenance and

OCR for page 25
--> 300 days of continuous operation per operating year for each site (Aberdeen and Newport). Utility Requirements Table 4-7 summarizes the TPC's stated utility requirements for a CEP agent destruction facility. The numbers in the table represent steady-state processing of agent at the design rate (upper bound) of one ton container (liquid agent to CPU-1, empty container to CPU-2) approximately every 4 hours. The principal utility requirements are natural gas and electric power. Note that the total electric power load of 1,510 kW shown in the table is a net load and includes a load-reducing contribution of 3,525 kW from cogeneration. Of the 33.35 x 106 Btu/hr (9,767 kW equivalent) of natural gas required at steady-state operation, 30.6 x 106 Btu/hr (8,962 kW equivalent), or 92 percent, is used for cogenerating electric power. The energy contribution to cogeneration from the synthesis gas is estimated at about 2 x 106 Btu/hr (586 kW equivalent). For electric power, the maximum operating load of about 7,500 kW (not shown in Table 4-7) occurs when starting up the two CPUs together and lasts a maximum of 2 days. During CPU startup, there is also additional demand for natural gas to fuel the headspace heaters. The water requirement is minor, consisting of makeup for a small offgas scrubber, makeup for a small cooling tower, and use by personnel. The total average requirement is estimated at 10 gallons (38 liters) per minute. Scale-Up Requirements The discussion of scale-up requirements for CEP is divided into issues related to scaling up the equipment and issues related to how processes are likely to perform when carried out at a larger scale. Equipment Scale-Up Front End and Back End Equipment The development of all process operations and equipment at the front-end of the process, as well as the back end of the plant, is well advanced. The same or similar equipment is used either in the Army's baseline program or in industry at the scale required for an agent destruction facility. For example, the punch-and-drain equipment for ton containers has operated successfully at the JACADS chemical demilitarization facility. CPU Equipment The state of development of the CPU and related equipment is described above in the Technology Status section. The Demo Unit is a commercial-scale reactor with a metal bath size of 2,700 kg. The three iron CPUs in the CEP conceptual design submitted to the Army are about 8,200 kg each; the nickel bath is about 5,350 kg. Based on these preliminary estimates of nominal bath size, a scale-up of approximately 3:1 from the largest CPU in operation is required. In the judgment of the panel, the TPC has sufficient experience and understanding of CEP technology to perform the scale-up of bath size successfully. The TPC has told the panel that it plans to use multiple tuyeres in each of the CPUs. Basic oxygen furnaces in the steel industry use many more tuyeres than are under consideration for this process. (At a meeting with the panel in January 1996, a TPC representative said that 16 to 20 tuyeres per furnace is common in the steel industry.) The TPC is continuing to validate the use of multiple tuyeres in an agent destruction CPU, and confirmation on an appropriate number of tuyeres will be part of a final engineering design. The design concepts for the premelting chamber to melt ton containers and for the system for feeding dunnage (in steel canisters) into the CPU-2 bath do not, to the panel's knowledge, have similarly close industrial counterparts. The TPC has conducted a demonstration program to test the processing of scrap metal, as a surrogate for some solid-waste feed streams of interest to the U.S. Department of Energy. However, the premelting chamber as suggested for the chemical demilitarization facility will require extensive development and demonstration. The TPC's reported experience to date includes a demonstration test in which six marine-location markers supplied by the Department of Defense were enclosed in cylindrical steel containers 0.8 m long and 9 cm in diameter. The containers were fed one by one into a molten metal bath through a gland in the top of the CPU. This test lends some credence to the submitted method for processing dunnage by loading it into cylindrical steel canisters 1 m long by 30 cm in diameter and feeding the canisters into CPU-2.

OCR for page 25
--> TABLE 4-7 Summary of Utility Requirements for a CEP Facility                 Fuels   Chemical     Unit Description Plant Air (scfm) Instrument Air (scfm) Breathing Air (scfm) Nitrogen (scfm) Oxygen (scfm) natural Gas (Btu x 106h)   Fuel Oil (Btu x 106h) Type Gallon Water Make-U (gal./ min.) 100 Feed preparation 120 @ 200 psig 10 @ 90 psig 100 @ 90 psig 40 @ 50 psig         10% HTH decontam. soln. 10 to 40 per ton container 0.5     + 30 @ 90 psig                 = 5 to 20 gal./h   200 Catalytic processing 20 @ 90 psig 20 @ 90 psig 100 @ 90 psig 150 @ 200 psiga 50 @ 200 psig 0.6           300 Gas handling train 20 @ 90 psig 10 @ 90 psig 50 @ 90 psig           10% HTH decontam. soln. 600a   500 Power generation 10 @ 90 psig 10 @ 90 psig       30.6           700 Product storage 20 @ 90 psig 20 @ 90 psig                   900 Infrastructure 100 @ 90 psig 5 @ 90 psig 50 @ 90 psig 10 @ 50 psig   2 1.2b   10% HTH decontam. soln. 600a 9.5 1000 Relief system 10 @ 90 psig 75 @ 90 psig 150 @ 90 psig     0.15   10%HTH decontam. soln.   600a   Totals   120 @ 200 psig + 210 @ 90 psig 75 @ 90 psig 300 @ 90 psig 150 @ 200 psig + 50 @ 50 psig 50 @ 200 psig 33.35 1.2b       10.0     Steamc Boiler Feed Water Condensate               Unit Description High Pressure (lb./h) Low Pressure (lb./h) High Pressure (lb./h) Low Pressure (lb./h) High Pressure (lb./h) Low Pressure (lb./h) Electrice Power (KW) Coolingd Water (gal./ min.) Demineralized Water (gal./min.) Domestic Water (gal./ min.) Sanitary Sewage (gal./ min.) 100 Feed preparation   20 @ 15 psig   0.05               200 Catalytic processing             4,000 750       300 Gas handling train             225 210 5     500 Power generation (1,720) @ 435 psig   4.2       (3,525) 151       700 Product storage             5         900 Infrastructure   780 @ 15 psig   1.95   0.5 800 379   5 5 1000 Relief system             5         Totals   (1,720) @ 435 psig 800 @ 15 psig 4.2 2 0 0.5 1,510 1,500 5 5 5 a For intermittent use. b Additional fuel oil will be required for a short period, to power diesel generators in case of an electrical power outage. c For steam ( ) indicates quantity produced. d Cooling water supplied at 80ºF and returns at 100ºF. e Connected electric load is 6000 KW. essential load is 3500 KW, UPS load is 150 KW.

OCR for page 25
--> Performance Scale-Up Front End and Back End Performance All the processes in areas 100 and 900 have been demonstrated in the Army baseline system with live agent at scales similar to the scale for an operational CEP facility, except for the optional high pressure water-jet systems for cutting open and cleaning ton containers. The panel expects the water-jet systems will work as proposed because they are commercial systems that have worked well on similar materials under extremely harsh conditions over long periods of time. CPU Performance The TPC has done extensive experimentation and modeling of CPU performance to understand bubble formation, breakup dynamics, and the operating limits of molten metal baths. As described in the Process Modeling section, this modeling work has identified three key factors in CPU performance to be bubble size, residence time, and energy dissipation by gas-metal mixing and gas-metal contact within gas bubbles. The TPC states that the modeling results correlate well with DRE values achieved in actual tests. The design for the full-scale baths is stated to provide a residence time with at least a tenfold safety factor over the residence time required to meet the requirement of at least six 9's (99.9999 percent) DRE. Testing Agent Surrogates in CEP The TPC tested destruction of an HD surrogate, half-mustard gas (HMG, 2-chloroethyl ethyl sulfide). The result was a DRE of at least nine 9's for conversion of HMG to synthesis gas, HCl, Fe-S alloy, and H2S. The DRE calculation was limited by the amount of agent processed and the lower detection limit of the analytical method. In another test, diazinon, which is structurally similar to VX, was reported to have been converted to synthesis gas, with the phosphorus and sulfur from the diazinon retained in the metal phase as an Fe-S-P alloy. Analysis of the offgas was conducted in accordance with EPA method TO-14. By this method, no C2 or higher hydrocarbons were detected at the lower detection limits, which are in the part-per-billion range. Third party analyses confirmed that no hazardous organic constituents were present in the ceramic or metal alloy products, which also passed the TCLP test for RCRA metals. The TPC states that the results verify that these solid products are nontoxic and potentially marketable. The AltTech Panel agrees with the TPC's interpretation of these tests as showing that the technology can destroy agent. The AltTech Panel sees no reason to expect the qualitative aspects of these test results to be different when the process is scaled up. The major conversion products and the partitioning between gaseous and condensed-phase are expected to be the same. The panel also believes the tests provide a strong preliminary indication that the residuals from a carefully designed CEP process to destroy chemical agents are likely to be nontoxic and safe for release to the environment or to commercial use, as the TPC anticipates. However, the panel cautions that the particular quantitative results obtained in these tests on surrogates, such as a particular DRE value or the nondetection of trace products at parts-per-billion concentrations in residuals, should not be directly extrapolated to full-scale operation unless information on certain key scaling parameters is provided. In the case of CEP test results, an important scaling parameter is one that the panel has named the specific processing rate, which for convenience can be defined as the amount of agent (in kilograms) processed per hour, per unit size of the bath (measured, for example, in 1,000 kg of molten metal). The closer the specific processing rate of a test is to the specific processing rate projected for a full-scale operation, the more confidence one can place in extrapolating quantitative test results. In the case of the tests on agent surrogates, the panel did not receive data from which specific processing rates could be calculated. Therefore, the quantitative results obtained under full-scale operation could be better or worse than these bench-scale test results with agent surrogates. Testing Actual Agent in CEP As noted in the Agent Testing section of Technology Status, the TPC has tested actual HD and VX agent in a bench-scale CPU at Battelle/Columbus Laboratories. The panel received the full report on these tests in early June 1996. The report states the agent destruction efficiency of the bench unit as eight 9's (99.999999 percent) for HD and VX. Based on the panel's preliminary review of the report, it appears to be more accurate to

OCR for page 25
--> call this result a DRE because the offgas passed through at least one filter before it was tested. The panel obtained sufficient data on the tests on actual agents to calculate specific processing rates for comparison with the rates for the full-scale system (Table 4-8). (The latter were computed from the design feed rates of agent and the nominal bath size.) Of several bath compositions tested for each agent, the panel used the results from the bath composition closest to that of the full-scale bath under steady-state operation. The bench-scale tests used a single top-entering lance to feed agent into the bath, whereas the design for a full-scale facility has bottom-entering tuyeres. As the table shows, these bench-scale tests of agent destruction were run at significantly lower specific processing rates than the rates the TPC has designed for a full-scale facility. In the panel's judgment, with the admonition stated above about extrapolating quantitative results from small-scale tests to performance of a full-scale operating facility, the implicit scaling factor in the specific processing rate for VX of 2.6:1 is within acceptable engineering practice. In making this judgment, the panel has taken into account the TPC's stated design safety margin of 10:1 in bath residence time and the reported test result of eight 9's DRE, which implies a performance margin beyond the required six 9's DRE. The panel cautions that the implicit scaling factor in the specific processing rate for HD of 5.4:1 leads to even greater uncertainty in extrapolating the bench-scale DRE to full-scale performance. The panel believes that the TPC understands the complexity of scaling quantitative performance measures such as DRE from bench-scale tests to full-scale operations. However, the panel would prefer DRE data for VX and especially for HD from bench-scale tests conducted at specific processing rates closer to the rates for the full-scale design. Unit Operations This section summarizes the unit operations in CEP treatment of chemical agents for the Aberdeen and Newport sites, including unit operations required to treat secondary process streams and residuals prior to disposal. A unit operation is a combination of equipment that accomplishes one specific step in a process. Table 4-9 lists the unit operations for CEP by process area. Process Safety Process safety risk factors for a CEP agent destruction facility can be divided into two categories: factors related to handling agent prior to its introduction into the CPUs and factors related to the molten bath technology. The risk factors inherent in the handling of agent prior to entry into the CPUs include storage risk, transportation risk, and the risk from the punch-and-drain operation. These risk factors are common to all the agent destruction technologies reviewed in this report, but they can be exacerbated or ameliorated by aspects of a specific technology. For example, how quickly a facility using the technology can reach operational status or the rate at which the agent can be processed with that technology can alter the storage risk by changing the length of time that the agent must be stored. The CEP technology is well advanced, and the design calls for processing the agent at each site in one year. Both of these technology-specific features help in reducing storage risk. As another example, the capability in the CEP design for treating emptied ton containers to the equivalent of 5X condition by melting and processing them immediately reduces the risk from handling the containers. The process safety risk factors inherent in CEP include issues associated with high temperature molten TABLE 4-8 Specific Processing Rates of Bench Tests Relative to Full-Scale Design Rates     Specific Processing Rate (kg agent/hour/1,000 kg bath metal)   Agent Tested Bath Composition Bench Test Full-Scale (design) Scaling Factor (full-scale/bench) HD Ni + 2% C 7 38 5.4 VX Fe + 7% P+ 7% S + C 8 21 2.6

OCR for page 25
--> TABLE 4-9 CEP Unit Operations by Process Area Area 100, Container and Dunnage Feed storage (ton containers) Transportation and Handling Punch-and-drain station Ton container wash and preparation Dunnage handling and preparation Liquid (agent and container-washout) storage and feed Area 200, CPUs CPU-1 Premelting chamber to CPU-2 CPU-2 CPU-2 offgas quench, scrub, particulate removal, and compressor Area 300, Gas Handling Train Gas quench and particulate removal HCl recovery Sulfur recovery Area 500, Synthesis Gas Utilization Gas compression and retention/analysis Power generation Steam-methane reformer (option for methanol recovery)a Methanol production (option for methanol recovery)a Area 700, Products Storage Sulfur product storage HCl product storage Methanol product storage (option for methanol recovery)a Area 800, Utilities Inert gas storage and feed Oxygen storage and feed Natural gas feed Air-plant air and instrument air Water-plant, potable, cooling, boiler feed, and chilled Steam-generation and condensate handling Electricity Diesel power backup Area 1000, Relief and Scrubber System Scrubber (decontamination solution) Boilers a These unit operations are only present if synthesis gas is converted to methanol instead of being burned to generate power. Under the methanol option, the power generation unit process would not be installed. baths such as the integrity of the refractory confinement, the proximity of the molten bath to water cooling coils (raising the possibility of steam explosions), the behavior of the tuyeres, and the instrumentation for monitoring the refractory confinement. In the panel's judgment, none of these factors presents an insurmountable impediment to the safety of the process. Many of the risk factors have already been addressed by the TPC in the hazard analysis it conducted for design of a chemical demilitarization facility (discussed above under Failure and Hazards Analysis) or on the basis of the TPC's research and operational experience with CEP. The panel was satisfied that the TPC had adequately addressed several issues the panel had raised during site visits regarding integrity of the refractory. The panel found no evidence of scenarios involving a loss of electrical power, loss of cooling, failures of pumps or valves, breaks in agent lines from inadvertent over-pressurization, or inadvertent temperature transients that would lead to off-site releases of agent or toxic process products. Pessimistic scenarios for a coincident loss of normal power, loss of backup power, and loss of cooling result in the solidification of the molten metal bath in place without significant release to the atmosphere.

OCR for page 25
--> Based on the panel's preliminary and qualitative evaluation, the most significant off-site risk appears to be associated with risk factors inherent in handling agent prior to the CEP process. In particular, the principal risk factors appear to involve mishaps during the punch-and-drain operation or damage from airplane crashes or other external events to holding tanks where agent is stored before being fed to the main reactor. The subsections on process safety below address the risk factors specific to CEP technology. However, the panel believes that none of these factors seriously challenges the safety of the facility. Safety Issues Related to Off Site Releases The following issues should be addressed fully and clearly in a final CEP process design. Integrity of the Refractory. The work by the TPC on the integrity of the refractory must be included in the safety documentation for a final CEP design. The TPC has done much work to avoid gas-jet impingement on the refractory lining of the CPU and to select refractory materials for the lining that resist gas permeation, thermal degradation, corrosion, erosion, and penetration by components of the molten metal and slag. Integrity of the Agent-Bearing Components. This issue was explored briefly by the panel, and no significant issues were uncovered. However, because certain parts of the design are still preliminary, the panel encourages the TPC to pursue its stated plans for continuing, comprehensive safety and hazard analyses as part of the development process. Particularly important is further exploration of scenarios involving failures of piping or components. (Failure could be caused by thermal attack by molten material, system overpressure, subtle system interactions, or other causes.) Cooling Offgas Piping. Scenarios involving a failure to cool the offgas piping should be explored. This is probably not an issue, but at the time of the panel's review, the consequences of such scenarios were not clear. Buildup of Combustible Gases. The TPC's design as submitted prevents a buildup of combustible gases in the vicinity of the system by maintaining a high ventilation rate. Assurances should be made that combustible gas buildup cannot occur and that the high ventilation rate does not compromise the design capability to contain leakage of agent. Worker Safety Issues There are a number of worker safety issues associated with high temperature molten baths, high temperature corrosives in the scrubbers, and secondary containment (concerning both inadvertent leaks and maintenance activities). These risk factors need to be addressed in the final operational design, and realistic emergency responses need to be spelled out. Specific Characteristics that Reduce Risk Inherent in the Design Because of the natural temperature gradient in the CPU refractory material, the molten material will solidify before it gets very far into the refractory. This self-sealing feature helps keep the molten metal away from the water-filled induction coils and thus reduces the possibility of a steam explosion. A loss of electrical power, of cooling water to the heat exchanger, or of the cooling for pumps could result in the molten metal solidifying in place. Although solidification would be an operational problem if it were to occur, it is not a safety issue. Schedule Figure 4-11 is the latest schedule submitted to the AltTech Panel from the TPC for the major activities and milestones in a chemical demilitarization program to use CEP technology at the Aberdeen and Newport sites. Table 4-10 is the panel's analysis, based on the TPC schedule, of activities on the critical path to completion of the program, their duration, and the cumulative time from start of the program to the end of that activity. An important aspect of the TPC's concept as submitted to the Army is that the same CEP equipment would be installed first at Aberdeen for HD destruction, then moved to Newport and installed there for VX destruction. Advantages and disadvantages of this approach are discussed below. Another key aspect of the design is that the TPC's preferred approach, after a go-ahead from the Army to begin work, is to move directly to design of a facility

OCR for page 25
--> Figure 4-11 CEP program schedule and phasing concept. Source: M4 Environmental L.P., 1996d. with full-scale CPUs for the next stage of development. A facility at that scale is more conventionally referred to as a demonstration plant than a pilot plant. To indicate how the schedule relates to the Defense Acquisition Board's decision to proceed with pilot-scale development, the panel will refer to this next stage as pilot/ demonstration. The facility for this pilot/demonstration phase at each site will be equipped with enough gas handling capability to ensure protection of human health and the environment, but the full gas handling train will not be installed until full-scale operation. The TPC foresees no scale-up effort required to move from pilot-testing to full-scale processing. The panel cautions, however, that although use of full-scale equipment at the pilot/demonstration stage means that no equipment scale-up will be required, whether performance scale-up is needed depends on how closely the final stages of pilot-testing resemble the process conditions for full-scale, continuous operation. The pilot/demonstration activities will entail a good deal of work, including systemization with agent surrogates, preoperational surveys, an operational readiness evaluation, and similar requirements prior to full-scale operation. Provided that the TPC continues testing and develops an adequate design basis prior to construction of the pilot/demonstration facility (that is, resolves remaining issues such as demonstrating the premelting chamber, scaling the bath to the larger size required, resolving the number and placement of tuyeres, and demonstrating process performance at the design

OCR for page 25
--> TABLE 4-10 Critical Activities in the Program Schedule Activity Duration (months) Cumulative (months) 1. Prepare and obtain regulatory permits, etc., for Aberdeen 15 15 2. Aberdeen construction (site prep. and installation) 12.2 27.2 3. Aberdeen pilot/demonstration (startup, test, and system modifications) 8 35.2 4. Aberdeen full-scale HD operations 12 47.2 5. Newport construction (site prep. and installation)a 14.25 60.45 6. Newport pilot/demonstration (startup, test, and system modifications) 8.25 68.7 7. Newport full-scale VX operation 13 81.7 a Newport construction overlaps one month with Aberdeen full-scale operation. specific processing rates), the panel believes that 8 months can suffice for performance scale-up and required startup activities. The full-scale operation at each site is designed to be continuous, 24 hours per day, at the agent feed rates specified above in the Feed Streams section. The scrubbed offgas is either combusted with natural gas in a gas turbine generator to produce electricity for the plant or converted to methanol. At this stage, process residuals would be placed on the commercial market. The design as submitted is not clear about how process residuals would be handled during the earlier pilot/ demonstration stage. The TPC has stated that the submitted design provides sufficient throughput to allow all agent, ton containers, and dunnage to be destroyed in 12 months from the start of full-scale operation at Aberdeen and in 13 months from the start of full-scale operation at Newport (M4 Environmental L.P., 1996d). Assuming that construction at Aberdeen can be approved by January 30, 1998, the TPC anticipates that the program for both sites will be completed before the end of 2003, more than a year before the Army deadline of December 31, 2004. The AltTech Panel believes that the TPC's goal of completing the destruction of each stockpile in 12 to 13 months after commencing full-scale operation is achievable, if the throughput rates assumed in the submission can be sustained for the duration of the operation. In the panel's judgment, the time allotted for pilot/demonstration activities at Newport is essential. The VX configuration uses the same equipment but a different set of processing parameters and constraints, as well as handling a different agent and a different partitioning of chemical elements to product phases. After processing HD at Aberdeen has been completed, the CEP systems will be decontaminated, decommissioned, and relocated to Newport for processing VX. The TPC believes this plan for reusing equipment is a cost-effective and time-saving solution for destroying agent stockpiles at multiple sites. The panel agrees that there are advantages to sequential operations but cautions that there are also risks to the schedule. A significant delay in the Aberdeen schedule could delay the agent destruction schedule at Newport. In fact, any delay in one of the activities along the critical path can delay subsequent activities. For example, the submitted schedule reflects early and vigorous efforts to complete the required reviews and secure necessary approvals. The TPC estimates that a permit for construction of a plant producing atmospheric emissions can be obtained in Maryland within 15 months of project start. The panel notes that this relatively short time for permitting may depend on the TPC acquiring a recycle waiver from RCRA permitting requirements. If the permitting process takes longer and construction is delayed, the schedule does have about

OCR for page 25
--> 15 months of slippage time at the end to still meet the Army deadline. The panel notes in passing that the time shown in Figure 4-11 for decontamination and decommissioning is probably only the time required to decontaminate and decommission the CEP systems. (The schedule refers to the activity as phase 1 of decontamination and decommissioning.) Additional time will probably be required for decontaminating and decommissioning the central building and the associated infrastructure.