6
Gas-Phase Chemical Reduction Technology

Process Description

The gas-phase chemical reduction process reviewed by the AltTech Panel was submitted to the Army by ECO LOGIC, Inc., of Rockwood, Ontario. ECO LOGIC is the developer and TPC for this technology and will be referred to as the TPC. The acronym GPCR will be used in the remainder of this report to refer to the particular process design submitted by this TPC for a gas-phase chemical reduction technology to destroy chemical agents. The process uses hydrogen and steam at elevated temperatures (up to 850°C) and nominally atmospheric pressure to transform organic wastes into simpler substances that are either less toxic or convertible to less toxic materials; these substances are also easier and safer to reuse or to release to the environment. The overall process requires a high temperature reaction vessel, where the chemical reduction occurs, followed by a gas scrubbing train to remove inorganic by-products. The process also includes provisions for removing other by-products and regenerating hydrogen gas through steam reforming. Figure 6-1 is a schematic illustration of the process.

Chlorinated hydrocarbons, such as polychlorinated biphenyls (PCBs), are chemically broken down and reduced to methane (CH4)and HCl with CO and CO2 as by-products. Nonchlorinated aromatic hydrocarbons such as toluene are reduced primarily to methane, with minor amounts of other light hydrocarbons. Carbon and presumably some heavier hydrocarbons are also produced.

The flow through stainless steel reactor has nozzles to accelerate the vaporization or dispersion of liquid wastes, which are injected directly into the reactor mix of hot gases consisting of H2, H2O, CO, and CO2. Within the reactor, radiant-tube heaters heat the mixture to 850°C. The residence time in the reactor is 2 to 6 seconds, although the TPC has stated that reactions occur in less than one second.

The gases exiting the reactor are scrubbed to remove by-products. Water is used as a quench to decrease the gas temperature and absorb water-soluble products, including HCl. These and other acidic products are further scrubbed by caustic scrubbers. A heavy-oil scrubber can be used in the scrubber train to remove some hydrocarbons. A standard monoethanolamine (MEA) scrubbing system removes most of the H2S (produced from sulfur-containing feeds) and CO2 from the gas train. The separated H2S requires further treatment to convert it to elemental sulfur and water.

The TPC has also developed and employed a sequencing batch vaporizer (SBV), which is a high- temperature chamber (up to about 550°C) in which hot gases from the recirculating process stream, including H2, H2O, CO, and possibly CH4, desorb organic contaminants reactively and thermally from drums and bulk inorganic solids. The SBV consists of two autoclave-like chambers that are operated independently in batch mode. The chambers can be fairly large—large enough to hold a ton container. A high temperature thermal reduction mill (primarily a bath of molten tin) can also be used to separate contaminants from soil or solids; the tin is a heat transfer medium to drive off volatile material, leaving inert solids behind. The gases from the thermal reduction mill and SBV are swept into the reactor for treatment. GPCR incorporates equipment for catalytically reforming most of the methane from the reactor to H2, CO, and CO2; the reformed gas is recirculated to the reactor to provide part of the necessary hydrogen.

The TPC has also developed mechanisms for holding gaseous process residuals for analysis prior to release or storage in containers. The overall process is monitored at a number of points using several methods: on line gas chromatography, chemical ionization mass spectrometry, a NOVA® oxygen analyzer, and a NOVA® gas analyzer to monitor H2, CO, CO2, and CH4.

The reactor (Figure 6-2) is constructed of stainless steel with a ceramic lining. The feed stream and hot reactant gases are injected through several ports mounted on the reactor. Special nozzles disperse liquid wastes into the hot gas. The gas mixture is heated further by 18 vertical radiant-tube heaters, which are isolated from the reaction mixture by an atmosphere of CO2.



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 102
--> 6 Gas-Phase Chemical Reduction Technology Process Description The gas-phase chemical reduction process reviewed by the AltTech Panel was submitted to the Army by ECO LOGIC, Inc., of Rockwood, Ontario. ECO LOGIC is the developer and TPC for this technology and will be referred to as the TPC. The acronym GPCR will be used in the remainder of this report to refer to the particular process design submitted by this TPC for a gas-phase chemical reduction technology to destroy chemical agents. The process uses hydrogen and steam at elevated temperatures (up to 850°C) and nominally atmospheric pressure to transform organic wastes into simpler substances that are either less toxic or convertible to less toxic materials; these substances are also easier and safer to reuse or to release to the environment. The overall process requires a high temperature reaction vessel, where the chemical reduction occurs, followed by a gas scrubbing train to remove inorganic by-products. The process also includes provisions for removing other by-products and regenerating hydrogen gas through steam reforming. Figure 6-1 is a schematic illustration of the process. Chlorinated hydrocarbons, such as polychlorinated biphenyls (PCBs), are chemically broken down and reduced to methane (CH4)and HCl with CO and CO2 as by-products. Nonchlorinated aromatic hydrocarbons such as toluene are reduced primarily to methane, with minor amounts of other light hydrocarbons. Carbon and presumably some heavier hydrocarbons are also produced. The flow through stainless steel reactor has nozzles to accelerate the vaporization or dispersion of liquid wastes, which are injected directly into the reactor mix of hot gases consisting of H2, H2O, CO, and CO2. Within the reactor, radiant-tube heaters heat the mixture to 850°C. The residence time in the reactor is 2 to 6 seconds, although the TPC has stated that reactions occur in less than one second. The gases exiting the reactor are scrubbed to remove by-products. Water is used as a quench to decrease the gas temperature and absorb water-soluble products, including HCl. These and other acidic products are further scrubbed by caustic scrubbers. A heavy-oil scrubber can be used in the scrubber train to remove some hydrocarbons. A standard monoethanolamine (MEA) scrubbing system removes most of the H2S (produced from sulfur-containing feeds) and CO2 from the gas train. The separated H2S requires further treatment to convert it to elemental sulfur and water. The TPC has also developed and employed a sequencing batch vaporizer (SBV), which is a high- temperature chamber (up to about 550°C) in which hot gases from the recirculating process stream, including H2, H2O, CO, and possibly CH4, desorb organic contaminants reactively and thermally from drums and bulk inorganic solids. The SBV consists of two autoclave-like chambers that are operated independently in batch mode. The chambers can be fairly large—large enough to hold a ton container. A high temperature thermal reduction mill (primarily a bath of molten tin) can also be used to separate contaminants from soil or solids; the tin is a heat transfer medium to drive off volatile material, leaving inert solids behind. The gases from the thermal reduction mill and SBV are swept into the reactor for treatment. GPCR incorporates equipment for catalytically reforming most of the methane from the reactor to H2, CO, and CO2; the reformed gas is recirculated to the reactor to provide part of the necessary hydrogen. The TPC has also developed mechanisms for holding gaseous process residuals for analysis prior to release or storage in containers. The overall process is monitored at a number of points using several methods: on line gas chromatography, chemical ionization mass spectrometry, a NOVA® oxygen analyzer, and a NOVA® gas analyzer to monitor H2, CO, CO2, and CH4. The reactor (Figure 6-2) is constructed of stainless steel with a ceramic lining. The feed stream and hot reactant gases are injected through several ports mounted on the reactor. Special nozzles disperse liquid wastes into the hot gas. The gas mixture is heated further by 18 vertical radiant-tube heaters, which are isolated from the reaction mixture by an atmosphere of CO2.

OCR for page 102
--> Effluent gases leave the reactor through a stainless steel central tube that leads to the scrubber system. GPCR has been under development since 1986 and has progressed from bench-scale testing through commercial-scale operation. A number of organic feed materials, particularly chlorinated wastes, have been tested at bench-scale. Several kinds of feed materials are currently being treated at commercial-scale (tons per day), including pesticides, chlorinated hydrocarbons, and PCBs. A full-scale facility to treat mixtures of toluene and the pesticide dichloro diphenyl trichloroethane (DDT) is operating in Australia. A plant in Canada for PCB destruction, which was visited by an AltTech panel team in January 1996, went on line in the spring of 1996. Scientific Principles Feed-Destruction Chemistry The chemistry by which GPCR destroys organic feed material is much more complex than a simple high temperature reduction with hydrogen of organic compounds to produce methane. The complexity results from the reduction with hydrogen being accompanied by reactions of carbonaceous intermediates, including elemental carbon, with steam to yield the final products. Although the thermodynamic principles of reducing organics with hydrogen to carbon and the resulting reactions of carbon with steam (carbon steam chemistry) have been thoroughly studied and are well understood, the interplay of kinetics and thermodynamics in the GPCR reactor are more difficult to ascertain. The chemical agents HD and VX contain a high proportion of heteroatoms (atoms other than carbon, hydrogen or oxygen, such as chlorine, phosphorus, sulfur, or nitrogen). The reaction products containing these heteroatoms will generate a large volume of inorganic process residuals. HD is 45 percent chlorine, 20 percent sulfur, and 30 percent carbon by weight; VX is 12 percent phosphorus, 5 percent nitrogen, 12 percent sulfur, and 49 percent carbon by weight (hydrogen and oxygen make up the rest of each compound). This heteroatom content raises two unanswered questions. First, what are the final heteroatom products from the reactor? Second, how are they scrubbed or otherwise removed? The acid gases and other inorganic products must first be scrubbed from the reactor effluent gas and then converted to a form suitable for disposal or recycling in commerce. The reactions of organic compounds containing heteroatoms are even more difficult to predict without the same kind of detailed experimental work the TPC has carried out on the feed materials it currently treats successfully. The TPC, which has considerable operational experience treating a number of highly halogenated wastes such as PCBs, hexachlorobenzene, and DDT, has found empirically that a fine balance of hydrogen and steam is necessary to avoid generating substantial amounts of carbon and polyaromatics in the reactor. The TPC has developed empirical models to predict operating parameters that yield optimal product composition: primarily methane, with CO and CO2. Nonetheless, the TPC allows in the design for some production of carbon (as soot), and the panel believes that some high-molecular-weight aromatics are produced. Therefore, carbon and other solids must be managed downstream in addition to the gaseous products (see Appendix F). For simple hydrocarbons, the TPC describes GPCR as a high temperature reduction by hydrogen to produce methane. Simple thermodynamic calculations reveal, however, that considerable amounts of carbon would be expected from the initial reaction with hydrogen. Therefore, carbon must react subsequently with H2O to generate CO, CO2, and, ideally, more hydrogen. Some high-molecular-weight carbon residue is also generated. This postulated pathway is supported by results reported by the TPC. Steam is added to the hot feed gas to react with the carbon to form CO2 and CO; the H2 content of the reactant gas is maintained above 55 percent, a level at which experience indicates the major product will be methane. Feed materials that contain heteroatoms must yield products that contain these elements, products such as acid gases (e.g., HCl) and reduced inorganics (e.g., H2S). The TPC has found that chlorinated wastes yield HCl as a primary product. The clean formation of HCl under the reaction conditions can be understood in terms of simple thermodynamics, given that chlorine probably cannot speciate to many other products under the reaction conditions. For chlorinated hydrocarbons, the overall reaction can then be visualized as:1 Cx Hy Clz + H2 = CH4 + HCl + C + other products 2C + 3H2O = CO + CO2 + 3H2 CH4 + 2H2O = CO2 + 4H2 CH4 + H2O = CO + 3H2 CO + H2O = CO2 + H2 Cx Hy Clz + H2 + H2O = CH4+ CO + CO2 + C+ HCl + other products 1   Not all of the equations shown here and below are balanced.

OCR for page 102
--> Figure 6-1 Schematic diagram of commercial-scale process. Source: ECO LOGIC, 1996a. In principle, the third, fourth, and fifth reactions could occur in the waste destruction reactor to produce all of the H2 needed for the hydrogenation reaction (first reaction). In practice, however, methane remains a major product from the reactor. The methane is converted to H2, CO, and CO2 in the steam reformer to provide enough H2 for the reactor. Another significant factor is that the rate of reaction of carbon with steam (second reaction) is slow, even at 850°C. For example, at 850°C, the time to react 99 percent of the carbon would be 23 days; in the SBV at 550°C, the same completion would require 500 years. Although reactive carbon-containing intermediates might react much faster, it is likely that some carbon will

OCR for page 102
--> be formed and must be managed by the downstream treatment of the reactor effluent. For simple chlorinated hydrocarbons, the TPC has sufficient practical experience to operate the process at conditions that generate the least amount of carbon. Even so, some carbon is produced and must be managed, and additional hydrogen must be regenerated or added. Far less is understood, fundamentally or empirically, about the fate of other heteroatoms—such as sulfur, nitrogen, and phosphorus that are present in the chemical agents HD and VX—in feed streams entering the GPCR reactor. The reactions of these heteroatoms have not been investigated extensively, and the interplay of kinetics and thermodynamics is difficult to predict a

OCR for page 102
--> Figure 6-2 Main reactor in the gas-phase chemical reduction process. Source: ECO LOGIC, 1996a priori. Predictions are necessary both for developing appropriate scrubber systems and for identifying and managing toxic residuals. Predicting the residuals from HD appears to be much more straightforward than predicting the residuals from VX. One can reasonably expect H2S to be the principal sulfur-containing product exiting the reactor from HD destruction. The TPC reports that this expectation is borne out by its experimental and full-scale work on wastes containing small amounts of sulfur. Moreover, the hydrogenolysis of organosulfur compounds to H2S is well known from commercial hydrodesulfurization processes. For HD destruction, the overall reaction can be summarized as: SC4H8Cl2 + H2 = CH4 + C + CO + CO2 + H2S + HCl + other products The TPC's empirical knowledge and operational experience with other feed materials should be sufficient to develop the appropriate conditions for HD destruction. However, provisions will be needed for handling the large sulfur (as H2S) residual stream. Although the TPC has some experience with small amounts of sulfur in feed materials, it will have to scale-up the MEA scrubber to handle the much larger quantities of H2S that would be generated by HD. Adding a new, scaled-up scrubber unit to the flow plan will bring the usual complement of potential problems in both startup and continuing operations. The TPC's plan to use commercially available technologies to convert H2S to elemental sulfur for ultimate disposal seems sound but considerably increases the complexity of the overall process.

OCR for page 102
--> The reduction of VX is much more complex, and the products are more difficult to predict. The speciation of the phosphorus and nitrogen present in VX is considerably more difficult to predict without laboratory bench work. The overall reaction for VX can be summarized by: C11H26SNPO26+ H2+ H2O = CH4 + CO2 + CO + C + H2S + P-products(?) + N-products(?) In contrast to hydrodesulfurization chemistry, the removal of phosphorus from organophosphorus compounds by hydrogenolysis has not been studied extensively. A more thorough understanding or at least empirical knowledge of the fate of nitrogen and phosphorus is clearly necessary for destruction of VX. Identifying the phosphorus and nitrogen products is also necessary for developing appropriate scrubbing systems and delineating the ultimate form and disposal of process residuals. The TPC believes that nitrogen-containing feed materials will yield both N2 and NH3 in the reactor; some HCN is another possibility though not favored thermodynamically. The analogy for phosphorus is tenuous, however. The main issue in heteroatom speciation can be illustrated with phosphorus-containing materials. Phosphorus-steam chemistry is not well understood, nor is reduction of the pentavalent phosphorus [P(V)] compounds found in the environment of a GPCR reactor. [P(V) is the form of phosphorus present in VX.] Although the TPC initially suggested that phosphine (PH3) would be the main phosphorus-containing material exiting the reactor (by analogy to the TPC's experience with methane production from carbonaceous material in the highly reducing steam environment of the reactor), the TPC has not reported detecting or characterizing any phosphorus-containing products from the laboratory-scale tests of VX surrogates. The panel's own thermodynamic calculations suggest that reduction of oxides of P(V) to phosphine is unlikely. From thermodynamic considerations, more likely products are oxyphosphorus acids (e.g., HPO2) and perhaps elemental phosphorus. (Appendix F describes thermochemical calculations made by the panel to understand potential speciation for phosphorus in the reactor.) A cautionary note is that oxyphosphorus materials are probably much less volatile than their carbon analogues, CO and CO2, and therefore might remain in the reactor or foul the exit tube or downstream piping. (A metric ton of VX would yield 375 kg of phosphoric acid.) Experimental work will be necessary to define the phosphorus end products in the reactor and explore these possibilities, particularly because the models used by the TPC are empirical and derived from experimental data for carbon speciation. These speciation issues are serious and will require substantial laboratory testing to resolve them prior to pilot-scale work. The TPC understands these issues and has stated that work is being done on them. The TPC has developed a plan to determine the speciation of phosphorus and design a method of scrubbing phosphorus-containing residuals from the reactor effluent. This aspect of the underlying chemistry is difficult for the panel to assess further without these empirical studies. Reactor Effluent Scrubbing The principal inorganic products of the gas-phase reduction of HD would be HCl and H2S. Both can be managed by conventional scrubbing systems that the TPC has previously employed for other feed materials. Although the large volume of H2S from HD will require scaling up the caustic and MEA scrubbers that the TPC currently uses with other feed materials, doing so should not be arduous. The plan by the TPC to convert H2S on-site to elemental sulfur using conventional commercial technology is preferable to storing and transporting large volumes of H2S, which is highly toxic. However, the conversion will increase the complexity of the overall system. There are two main problems associated with scrubbing inorganic products and acid gases from VX destruction: (1) determining the primary phosphorus-containing products exiting the reactor and (2) developing or implementing the scrubbing systems needed to handle these products in the effluent stream. The TPC has presented proprietary chemistry for scrubbing phosphine, but the technique requires further demonstration and may be inappropriate if phosphine is in fact not produced (see discussion above). For both VX and HD, the scrubbers must also remove not only the elemental carbon formed in the reactor but also any high-carbon-content precursors that may be present, such as aromatics and polycyclic organic compounds. The elemental carbon will probably be present as very finely divided particulates (soot) and will wash out with the initial water quench. The TPC should have experience with this process from its current operations. The TPC has stated that it expects that any polycyclic organic compounds will be present in very small amounts and can be recovered by heavy-oil scrubbing.

OCR for page 102
--> Test work on agents will show whether or not this scrubber will be needed. The TPC has not described the ultimate disposition of this process residual, which conceivably could be managed by recovering the high-carbon material from the oil, by burning it along with the recovery oil, or by other means. Regardless of the status of the scrubbing technologies, the recovery of process residuals containing the speciated products of chlorine, sulfur, phosphorus, and carbon and their conversion for ultimate disposal is clearly a complex process that will require a number of unit operations. Technology Status The TPC has experience treating organic wastes, including PCBs, other chlorocarbons, and hydrocarbons such as toluene, and has been developing GPCR for more than 10 years. It has conducted a significant amount of work at laboratory, pilot, and commercial-scales. Pilot-scale work has been performed since 1991 at several sites in the United States and Canada. The TPC has a laboratory-scale system available for waste treatability studies and has use this system for preliminary tests on agent surrogates, such as the organophosphorus pesticide malathion. Pilot-scale demonstrations have been performed on several materials, including polyaromatic hydrocarbons at Hamilton Harbor, Ontario, in 1991 and PCBs (PCB-contaminated oily water, highly concentrated PCBs in oil, and contaminated soil) at Bay City, Michigan, in 1992. Commercial units are currently deployed in Australia and Canada, and others are in progress. The TPC has been treating a mixture of DDT and toluene on a commercial-scale in Australia. Another system at St. Catherines, Ontario, processes PCBs both as a concentrated material and in PCB-contaminated concrete, for General Motors. For treating these feedstocks, the status of the technology is advanced, since there are commercial facilities in operation. In the judgment of the AltTech Panel, the TPC has considerable experience with these feed streams in all aspects of facility operation, including operational requirements and considerations, mass balances, gas recycling, and management of residual HCl. Although the panel received detailed modeling data from the TPC, it did not receive detailed laboratory data from the agent destruction tests, which were at laboratory-scale. No bench-scale tests have been reported to the panel. Full operational manuals, hazard and operability studies, process and instrument diagrams, and risk analyses have been developed and documented for processing DDT-toluene mixtures and PCBs. The TPC's experience with organic wastes forms a basis for applying the technology to agent destruction, but further development specific to the chemical agents to be treated is still required. For instance, all operations to date have been outdoors. Agent destruction facilities, however, will require containment of all unit processes where agent may be present. Containment of hydrogen gas within a building can be hazardous. Additional hazardous-operation procedures for handling these conflicting safety demands were not addressed in the TPC's submissions. In past and current operations, the TPC has also tested the capability of the SBV to remove and destroy organic wastes from inorganic matrices. This experience qualifies, to some extent, as a demonstration of the SBV's efficacy for treating ton containers and dunnage. The panel notes that the washing, cleaning, decontaminating, and shipping procedures for ton containers that the Army has proposed for neutralization (see Chapter 7) could also be used with GPCR. In summary, the main uncertainty in applying this process to agent destruction centers on identifying and managing the inorganic by-products derived from the sulfur, phosphorus, and nitrogen in the agents. What are the primary inorganic products from the reactor, and how will they be removed and managed downstream? The systems currently used by the TPC should work well for HD destruction after modification of the scrubber train to handle the large load of H2S. The status of the technology for HD treatment is near commercial except for: (1) the lack of demonstrated handling of large volumes of H2S, (2) the overall process demonstration on HD itself (small-scale tests to show the process can destroy agent have been successful), and (3) the resolution of secondary containment and safety issues specific to processing chemical warfare agents and hydrogen gas. Because much less is known on both a fundamental and practical level about the identity and handling of phosphorus-containing residuals from GPCR, the technology for VX destruction is less mature than it is for HD. The TPC has little experience with phosphorus-containing materials, even at bench-scale. Although the

OCR for page 102
--> TPC has developed a plan for addressing these issues, the time line for doing so is unclear. Operational Requirements and Considerations Process Operations GPCR consists of a number of sequential subsystems (e.g., feed system, SBV, reactor, scrubber train, methane reformer) that must be tightly controlled and integrated. Because the process is tightly integrated, provision must be made for safety cutoffs or mechanisms that recirculate excess materials back to earlier stages in the process. At least one scrubber is needed to manage each heteroatom-containing product that exits the reactor. The recovery subsystem—that is, the scrubbers and the subsequent unit operations for generating final process residuals from the scrubber effluents, such as conversion of H2S to elemental sulfur—consists of standard unit operations (with the possible exception of operations to scrub and handle the phosphorus-containing products). Nevertheless, this subsystem adds considerable complexity to the overall process for destroying agent. The conversion of H2S to sulfur alone will require four or five unit operations and a compressor. Mass Balance The panel received a flow sheet for the processing operations and some material balance predictions. The TPC has developed an empirical model, based on past experience, from which it made the following predictions: Methane is the predominant hydrocarbon produced, as long as the H2 content of the circulating gas stays above 55 volume percent (dry basis). The model uses 60 percent H2. Steam is necessary to limit the production of elemental carbon or high-molecular-weight aromatic material. The steam content is not specified in the model but in practice has been 20 to 80 percent of the dry gas. The model assumes gaseous carbon species in the reactor effluent gas occur in the ratios of 42 percent CH4, 34 percent CO2, and 24 percent CO (see Appendix F for details). In addition, 10 percent of the carbon in the feed materials is assumed to exit as elemental carbon (soot). The hydrogen/ methane ratio in the reactor effluent gas is assumed to be 2.0. Energy Balance The TPC did not provide a complete energy balance, but the electric energy required for the reactor heating elements was stated. For HD destruction, the reactor requires 5,019 kW·h/day, or 346 kW·h per ton of HD destroyed. The total heat input required is much larger, and the rest is supplied by burning fuel gas: either the gas produced in the operation or LPG (liquefied petroleum gas, primarily propane). Process Residuals Some of the process residual streams are reasonably well defined; some are not. They are discussed below in terms of the fate of the carbon and the heteroatoms in the feed material. Carbon. Some carbon (as methane) and hydrogen are eventually burned in a steam boiler, which supplies the steam for the gas reformer that converts methane to CO2 and H2. The residuals from the combustion exit the process primarily as CO2 and steam from a stack. Some carbon (as CO2) in the reactor effluent gas is scrubbed out with the H2S in the MEA scrubber. The MEA will scrub out most of the CO2 in the reaction gas; there is somewhat more CO2 than H2S expected in the gas. Some carbon (10 percent of the feedstock C) is estimated to exit the process as carbon soot. Some may also exit as hydrocarbon products that would be scrubbed by oil. Heteroatoms. The heteroatoms are scrubbed from the reactor effluent gas in various ways. Chlorine exits the reactor as HCl, which is scrubbed out first with water, which eventually produces a residual of concentrated aqueous hydrochloric acid. After the water quench, a caustic scrubber completes the removal of HCl and other acidic components with sodium hydroxide solution. The residual stream from the caustic scrubber is a solution of sodium salts (sodium chloride and salts of other scrubbed acids).

OCR for page 102
--> The sulfur exits the reactor primarily as H2S gas, which is scrubbed out with MEA, a standard commercial treatment. The H2S is stripped from the MEA solution with hot steam. The TPC plans to convert the H2S to elemental sulfur using a commercial process called the SulFerox process. The H2S can be converted to sulfur even in the presence of the carbon dioxide that is stripped with it. The SulFerox process uses oxygen from intake air to oxidize H2S to water and sulfur, with the oxidation mediated by an iron chelate that is regenerated with air. The sulfur is filtered from the SulFerox liquor. The gas from the subsequent sulfur plant, which is one of the effluents of GPCR, would have three components: (1) the CO2 remaining after the SulFerox treatment removes the H2S, (2) the water vapor from the water produced with the sulfur from oxidation of the H2S, and (3) the spent air (lower in oxygen) from regenerating the iron chelate. The flow sheet shows some sulfur, as sulfate and sulfite, appearing in the scrubber liquors from scrubbers preceding the MEA scrubber, although the major sulfur-containing product is expected to be H2S. Some H2S would be expected to dissolve in scrubber liquors (e.g., the sodium hydroxide solution in the caustic scrubber), which will make them unacceptable for discharge without further treatment. The issues associated with inorganic by-products from VX have been addressed above and in Appendix F. The panel's calculations, based on thermodynamic equilibrium, suggest that the following products will form: chlorine will yield HCl; sulfur will form primarily H2S with trace amounts of SO2; phosphorus will yield higher-valent oxides such as P4O6 and perhaps elemental phosphorus; nitrogen will form N2, NH3, and possibly nitrogen oxides. If nitrogen reactions are kinetically controlled rather than reaching equilibrium, HCN may be a minor product. Presumably the fate of phosphorus under the process conditions will be determined by the experimental work planned by the TPC. The exact nature of the scrubber effluent, in fact the type of scrubber to be used, will not be known until this work is done. Ton Container Treatment In the design as submitted, the emptied ton containers are cleaned in the SBV. The maximum residual agent in the SBV effluent gas stream, which is fed to the main reactor, cannot be predicted quantitatively prior to pilot-testing, although the Army's experience with container cleanout at existing facilities suggests it will be very low under the SBV process conditions (described below). The SBV will also be used to heat the solid carbon residual that collects in the water quench effluent. The heating will drive off and react any volatile material associated with the carbon. Ton containers will be placed in the SBV, two at a time, and treated with hot gas at 540°C (1000°F) for several hours (probably around 6 hours). The TPC has stated that this relatively long period at high temperature will be at least equivalent to the Army 5X cleaning conditions of heating for 15 minutes in an oxidizing atmosphere at about the same temperature. Evaporation and thermal cracking are stated to be the most important processes that occur, and they are independent of the composition of the circulating gas. Testing must be done to verify these statements. If regulatory approval can be obtained, the TPC has stated that it would prefer to recycle the ton containers. An alternative for treating the emptied ton containers in the SBV would be to use the Army process (hot water wash, followed by steam cleaning), described in Chapter 7, to clean them sufficiently to allow shipping them to Rock Island Arsenal, Illinois, for melting. Materials and Energy Balance The TPC provided material balance data (feed rates with product compositions) for the reactor, the scrubbers, the product gas boiler, and the catalytic reformer when treating HD. Details on feed streams and products for these unit operations are given in the tables and discussion in Appendix F. The TPC provided a similarly detailed material balance for VX destruction, but that balance has not been reproduced for this report because the panel believes that the reaction chemistry is still too uncertain, as explained above. Most of the chemical bonds in HD (or VX) should break at the temperature of the reactor. If some feed material were to exit the reactor without reaching reactor temperature, some agent might not be completely broken down into the simple residuals expected. According to the TPC, its experience with other feed materials suggests that this will not happen. However, based on the information available to the

OCR for page 102
--> panel, tests have not yet been done for some process residuals that could result from partial breakdown of HD or VX. The least detectable concentration for most process substances and residuals from HD destruction will be very low (parts-per-billion range) because the compounds of analytical interest will be in liquids that can be stored for hours or days, allowing ample time for detailed analysis. Feed Streams There are six feed streams to the reactor: the liquid agent feed, the SBV effluent gas, steam, hot waste water recycled for processing, reformer gas, and recycled reactor effluent gas. The reformer gas is the largest of these streams (about 85 percent of total moles fed) and has the composition shown in Table 6-1. The reformer gas is already at high temperature when it reaches the reactor; the electric heaters in the reactor are needed only to provide enough incremental heat to reach the reaction temperature. The two agent-derived feed streams are the liquid chemical agent from the ton containers and the effluent gas stream from the SBV. One ton container at a time is drained into a holding tank, and the liquid agent is pumped directly from this holding tank into the reactor. The ton container with its residual liquid or gel (the heel) is moved into the SBV, where remaining material is vaporized or reacted with the hot circulating gas (hydrogen, methane, steam, and CO). The effluent gas from the SBV is fed to the reactor. According to the TPC, the precise details of safety containment for holding tank, pumps, and lines have not been worked out. The scale of the equipment the handles the agent-derived feed streams (pumps, reactor, and SBV) is the same as the equipment in operating plants in Australia TABLE 6-1 Composition of Reformer Gas Component g-mols/m3 Volume % H2 755 74.0 CH4 15.3 1.5 CO 35.3 3.5 CO2 55.3 5.4 H2O 159.8 15.7 and Canada. The exact mechanical layout and protective housing to be used when a ton container is punched and drained have not been designed. However, the system proven by the Army (see Chapter 7) for the punch-and-drain unit operation can be used. No pretreatment of gels or solids is needed if they are sent to the SBV inside the ton containers. Process Residual Streams Bulk Agent The residual streams consist of the combustion gases from a steam boiler and a number of products from the scrubbing system. For HD, the overall reaction (agent in, residuals out) will be:2 SC4H8Cl2 + 5O2 → 18.8 N2 + 4CO2 + 2 H2O + 18.8 N2 + 2HCl + H2S + H2O + 1/2 O2 + HCl (in solution) + H2O The material balance for HD destruction in Appendix F is based on an agent feed rate of 5 metric tons per day and the TPC's model for assigning products of reaction. There appears to be some flexibility possible in the material flows through the process. The product HCl is scrubbed out in the water quench, along with the solid carbon from the reactor. The carbon, which is filtered out, eventually goes to the SBV for drying and the removal of any adsorbed products of incomplete reduction. The residual carbon is a final process residual and will have to be disposed of. Most of the H2S is recovered as elemental sulfur in a multistep process. H2S and CO2 are scrubbed from the gas by MEA. Efficient removal of HS from the reactor effluent gas is required for two reasons. First, most of the gas will be steam-reformed, and the reforming catalyst may be sensitive to sulfur. Second, the remainder will be burned, and the combustion exhaust gases will have to meet regulatory limits on sulfur. The MEA scrubber is a two-vessel system, with H2S and CO2 absorbed in one column and regenerated by steam-stripping in a second. The circulating MEA must be alternately cooled and heated as it flows from one vessel to the other. The effluent gas from the MEA stripper consists of H2S and CO2 (about 70 volume 2   The caustic scrubber will also produce a small amount of NaCl. Some H2S and NaSH/Na2S may also be present in the quench and scrubber solutions, respectively.

OCR for page 102
--> percent CO2). The TPC proposes to oxidize the H2S to sulfur using the SulFerox process (which is also proposed for use with the CEP, see Chapter 4). The CO2 is released to the atmosphere. The form of the phosphorus-containing residual (or residuals) must be determined by further experimental work. Presumably the phosphorus would be converted to its most stable form as a phosphate salt. In a reducing atmosphere, P(III) (phosphorus with a valence of 3) is the stable form of phosphorus. At this time, the TPC has not identified commercial facilities for receiving any of the liquid or solid process residual streams described above. Nonprocess Wastes Nonprocess wastes will be treated like process wastes. Solids (such as personal protection suits) will be treated in the SBV. Solid residuals from the SBV will probably be classified as toxic waste. Liquids (such as used decontamination fluid) will be sprayed directly into the reactor, along with the liquid agent feed stream. Products from treating these nonprocess wastes will become part of the same residual streams as the process residuals described above. Process Instrumentation and Controls The GPCR design states that the instrumentation and the control system for agent destruction applications will be based on the ones used in the operating facilities in Australia and Canada, with appropriate modifications. The following instrumentation is being used. Chemical Ionization Mass Spectrometer. This is a sensitive, soft-ionization mass spectrometer capable of monitoring, by visual display and recording, selected organic compounds at concentrations of parts per billion. Results are obtained in seconds. The gas outlet of the reactor will be sampled and monitored for compounds selected as indicators of the completion of the destruction process. Preliminary tests on the reaction chemistry will be used to select the compounds to be monitored. For HD, as an example, unreacted HD and simple sulfides or mercaptans might be the selected indicator compounds. M200 Gas Chromatograph. This on line gas chromatograph will also be used to analyze samples from the reactor effluent gas, for the same purpose as the mass spectrometer. NOVA Process Gas Analyzer. A high-precision infrared detector will be used to monitor concentrations of methane, carbon monoxide, and carbon dioxide at various locations in the process gas flows. Hydrogen will be monitored by a thermal conductivity cell. As a safety measure, oxygen will be monitored continuously by a NOVA oxygen analyzer with an electrochemical cell. Gaseous residual streams coming out of the process will also be monitored, particularly for residual agent. Pressure measurement and control are important for the process; to preserve a safe hydrogen atmosphere, negative pressure anywhere in the recirculating gas circuit must be avoided. The system pressure is maintained by continuous feed gas inputs to the reactor and by adjusting the rate of removal of reactor effluent gas; the latter is controlled by a variable-speed blower with a gas bypass around the blower. As an added safety feature, gas from the high pressure storage subsystem can be fed back to the reactor if the system pressure becomes negative. An aspect of measuring and controlling pressure that the TPC did not specifically address in its written submissions or in dialogue with the panel is proper ventilation of the building that will serve as a secondary containment for the recirculating gas circuit. (Secondary containment is required as a backup line of control to prevent an accidental release of agent to the atmosphere.) The system circuit must be maintained at a slight positive pressure relative to the ambient building air, to prevent oxygen from leaking in. Any leakage will therefore be process gas leaking out. In current operations in Australia and Canada, there is no explosion hazard if hydrogen leaks out of the system because the entire system is outdoors (no secondary containment). A small hydrogen leak that causes a small, controllable flare in an unconfined system could lead to an explosion if hydrogen accumulates in an air-filled secondary- containment building. For example, the experience of the National Aeronautics and Space Administration in handling hydrogen, which it uses routinely and in large quantity, has been that all leaks in enclosed systems lead to fires. Therefore, any secondary containment for GPCR will require substantial ventilation to prevent the accumulation of hydrogen. At the same time, the ventilation system must prevent a release of agent from the secondary containment in case of a leak.

OCR for page 102
--> The most crucial temperature to control is the reactor temperature. The final reaction temperature is achieved by electric heaters, and small adjustments are made by controlling the electric power. (A major drop in temperature would trigger a cutoff of the feed.) The overall process in the submitted design for GPCR is tightly integrated. Effluent gas from the reactor must be scrubbed in a series of scrubbers. The final, MEA scrubber must be regenerated continuously and the H2S converted to elemental sulfur. A portion of the reactor-effluent gas stream must be fed continuously to the steam reformer to maintain hydrogen concentration. The effluent must be preheated for this process. Another part of the effluent gas stream is simply recycled to the reactor after preheating. Some of the effluent gas is compressed and stored and then used to even out flow and pressure requirements. The control program used in existing operations appears satisfactory and will have had many hours of operation on a commercial- scale before a facility for treating chemical agent is ready for testing. The TPC has assessed several failure modes and has developed control strategies for them. Process parameters have been identified that are critical for process control and safety. The monitoring and control operations, which are based on treating materials other than chemical agent, are well documented. The complete scrubbing system required for processing chemical agents has not yet been settled. (The TPC is considering incorporating a multistage sulfuric acid scrubber with other more conventional scrubbing towers, but the plan may change depending on the results of further laboratory work.) New controls may be needed for new scrubber equipment. Certain control requirements will place stringent limits on the operation, particularly controls related to the safe handling of high temperature hydrogen—namely, the control and monitoring of pressure and the control of oxygen. Close control of temperature will be required for the chemical reactor units: the main reactor, the steam reformer, the MEA stripper and regenerator, the sulfur recovery system, and possibly the HCl scrubber. In summary, GPCR technology involves a complex chemical plant that must be operated carefully. Process Stability, Reliability, and Robustness Stability Of the reactions that are presumed to take place in the GPCR main reactor, hydrocarbon cracking followed by hydrogenation to CH4 is highly exothermic. The steam-forming reactions to yield CO and H2 are highly endothermic. For HD and VX, the net heat effect in the reactor appears to be very slightly exothermic. By monitoring the temperature and controlling the agent feed and the heat for bringing the feed streams to the reaction temperature, the process can easily be controlled. The operation of the main reactor should be stable because the reactor can operate reasonably well over a wide range of temperature above the minimum needed for the reduction reactions to occur. Operation of the steam reformer will require closer temperature control to maintain catalyst activity. Scrubbers should operate satisfactorily within modest deviations from design conditions. Reliability Because GPCR is a continuous process rather than a batch operation, multiple reactions in multiple reactors must proceed in continuous balance. The main reactor has had considerable commercial operation and has proven reliable. The entire system, however, consists of a number of sequential unit operations that must be tightly integrated and controlled. Agent destruction in the reactor is followed first by a sequence of chemical scrubbers and a sulfur recovery reactor, then by a catalytic steam reformer whose catalyst would be poisoned by breakdown of the scrubbing system. Removal of the process residuals from the recirculating gas stream for treatment and disposal, as well as steam reforming of H2 from methane, are carried out in a continuous loop. In a system operating in this continuous recirculating mode, a failure in one unit operation could significantly effect the others. For example, any reaction products containing sulfur or phosphorus that are carried past the scrubbers to the steam reformer are likely to poison the catalyst in the reformer. The system as a whole therefore must be carefully controlled at many points. The information provided by the TPC indicates that the overall system has been stable and reliable in operations that treat simple chlorocarbons ("simple" meaning compounds containing just chlorine as a heteroatom). The reliability of the more complex system required to treat one or more chemical agents will require demonstration. Tighter controls will certainly have to be implemented. The stored energy that might be of concern is the hot H2, which obviously must be protected from air.

OCR for page 102
--> Scenarios can be envisioned in which air leaks into the process stream because of negative pressure. Multiple controls are built into the system to prevent negative pressure in the system. The panel expects that the mechanical equipment will be highly reliable because it is generally standard in the chemical industry. However, existing plants have not been operating long enough to give a definitive answer on long-term operability and performance. Robustness The reactor can operate satisfactorily over a considerable range in temperature—±100°C from the design temperature. The catalytic steam reformer will require close temperature control and a "clean" feed; that is, no catalyst poisons in the input gas, such as sulfur or phosphorus. The scrubbers will require constant monitoring but should be able to handle a modest range of feed rates. The system response to upsets will depend on the upset. A sudden drop in a feed stream will trigger the pressure control system, which will add recycle gas to maintain positive system pressure. A drop in energy input to the main reactor (electric power to the heaters) will presumably produce a temperature drop. The panel believes that the time constant for such a temperature change is on the order of several seconds. Drop in reactor temperature below a set-point would shutdown the agent feed as well as trigger the pressure control. A change in the feed material will probably require major changes in the scrubbing system. A change in the feed material or feed rate may also require modifying the amount of gas circulating to the steam reformer, but the system should adjust for this by responding to the monitoring data provided by the gas analyzer. Materials of Construction Satisfactory performance of the overall GPCR system depends upon containing the reaction and the products of reaction. Designing an effective containment requires understanding how chemicals in the process environment interact with materials of construction to degrade them. In general, it appears that the unit processes of this technology are similar to or the same as unit processes for which substantial experience exists and for which there are conventional and satisfactory materials of construction. To understand the possibility for premature degradation that would reduce the integrity of important materials in this system, it is necessary to consider specific environments to which they are exposed, the properties of the materials, the design features of design that affect degradation, and the possible failure modes. Environmental Definition Three operating environments are expected in the GPCR technology: The reactor and the immediate downstream piping contain gas mixtures at temperatures as high as 900°C. This gas consists mostly of hydrogen and steam, but other corrosive species are present, such as HCl, H2S, phosphorus oxides, and carbon. Particularly in the reactor itself, this gas may also contain some agent in the process of being destroyed. In the scrubber systems, room temperature acid solutions are produced that contain various acidic species, including chloride, sulfur acids, and phosphorus acids. Products from the scrubber systems contain H2S, CO2, and steam. Although these are the general steady-state operating environments in the GPCR process, other environmental conditions that will occur intermittently can be important to material degradation. First, shutdown conditions may lead to aerated acidic or other corrosive conditions on component surfaces. For example, in fossil-fueled power systems the chemicals formed by the reaction of humidity with deposits during shutdown are often more corrosive than the deposits. Second, environments on the outside of component surfaces may be corrosive during operation or shutdown because of humidity, dripping water, or industrial gases. Third, accidents or out-of-specification conditions may occur during normal operation. Finally, the degradation of materials of construction usually involves subtle processes dependent on the relative amounts of chemical species. If a generally analogous system in terms of anticipated species present in the technology has proven satisfactory, a common presumption is that a similar system will also perform satisfactorily. Whether this argument by analogy is valid depends on

OCR for page 102
--> subtle chemical differences, such as the following, which may apply to gas-phase reduction of chemical agents HD and VX. Sulfur Valency. Sulfur species with valences less than +6 are generally corrosive, depending on the pH and the presence of other species. They are corrosive over a wide range of compositions for nickel and chromium alloys, although the susceptibility of an alloy to corrosion depends on how heat treatments have affected its local composition. Sulfur-Chloride Ratio. Corrosiveness of the gas mixture changes greatly, depending on the relative ratios of sulfur and chloride species. Hydrogenation. The presence of lower-valence sulfur species, as well as of phosphorus and cyanide species, influences various modes of hydrogen-related damage, such as cracking and blistering. Materials to be Used The TPC's submission did not define the materials of construction except to note that they will be materials that have performed satisfactorily in analogous industrial systems. The preliminary design suggests that a major structural material would be stainless steel. However, it is necessary to specify which of many stainless steels would be used, the fabricated condition of the material, the conditions of welding (heat-affected zone, weld metal, weld passes), and the residual stresses. It is also necessary to define how the properties of materials change with exposure to processing conditions, especially to temperatures. For example, in the range of 800°C, the microstructures of certain stainless steels change in ways that increase the likelihood of corrosion-induced failure. Design Features Certain design features in any new system affect whether accelerated degradation occurs. Although the materials of construction may be typical of the materials used in analogous systems that perform satisfactorily, often a system-specific configuration of these materials promotes degradation. Among configurations and conditions that may lead to premature degradation in GPCR are the following: Operation at 800°C suggests that the system will be subject to high thermal stresses. Because this system has no prototypes, it will probably be subjected to thermal cycling because of the intermittent runs conducted during the pilot-testing and demonstration phases prior to long-term, continuous operation. Crevices in general, and heat-transfer crevices in particular, can accelerate degradation. The design as submitted does not specify whether crevices are prohibited or have been otherwise considered. The formation of deposits at the bottom of reaction vessels in a scrubbing system generally accelerates degradation processes. Other conditions that accelerate degradation include liquid-vapor interfaces, especially when the vapor contains oxygen. Modes of Degradation As a result of chemical and design conditions, a number of modes of degradation may occur either in the main reactor or in the scrubbing system. In general degradation, the material may "rust away" if surface corrosion increases as the environment becomes increasingly acidic or alkaline. General degradation can be minimized by using alloys that contain a larger percentage of nickel and chromium (high alloys). Localized corrosion, such as pitting and intergranular corrosion, may occur even in highly alloyed materials. In fact, localized corrosion is often more aggressive in highly alloyed materials and can be especially affected by operating temperatures. Stress corrosion cracking and hydrogen embrittlement can occur regardless of the alloy composition. These modes of degradation are particularly aggressive in alloys that are more resistant to general degradation. Stress cycling, such as thermal cycling, may interact with the process environments to produce fatigue cracks. Failure Modes In addition to the modes of degradation described above, there are more general modes of degradation and

OCR for page 102
--> failure to consider. These are especially important given the toxicity of the agent and of some of the principal products of reaction, such as H2S. Among the failure modes that need to be considered are (1) oxygen leakage mixing with hydrogen during startup, shutdown, or operation, (2) release of reactor effluent gases from piping defects, (3) leaks in the pregasifier, and (4) system problems caused by thermal variations. Operations and Maintenance Operations According to the TPC, the system would be operated continuously (24 hours a day, 7 days a week) and would employ four separate shift-teams during treatment operations (12-hour shifts, 4-day rotations). Each shift-team of 10 to 11 people would consist of a shift supervisor (Professional Engineer [P.E.] certified or equivalent training), two process control engineers (P.E. or B.S.), two regular technicians for process maintenance and monitoring, one maintenance/boiler operator-engineer, one logistics coordinator, and three or four trained laborers, as required for handling material. In addition, an operational staff of five would be assigned to the project for a normal work week (5 days per week, 8 hours per day). This staff would consist of a project manager (P.E.), a project administrator, a quality assurance officer, a health and safety officer, and a senior member of the technical support staff (Ph.D.), as required. The TPC states that the normal staff (i.e., the shift-team and the operational staff as needed) would be capable of, and have the required training for, initiating shutdown and restart during normal operations or emergency situations. The system operates on a 24-hour continuous basis. Although the system can be operated on an intermittent basis (e.g., an 8-hour day), cost-effectiveness would decrease because of increased startups, shutdowns, heating costs, etc. Standby mode is similar to operational mode but with reduced utility requirements. For long standby periods, the system would be purged with nitrogen. At ready mode, all system components are operating and up to temperature. Staffing for standby and ready modes is the same as for operational mode, with a provision for reduced staffing in the event of a long standby. The TPC has complete operational manuals, hazardous-operations procedures, process and instrument diagrams, and risk analyses for its commercial operations. The process control system, which was described in the response to the panel's questionnaire, consists of a Moore Advanced Process Automation and Control System interfaced to a microcomputer. Upgrades to the control room (e.g., more screens and monitors) will probably be needed for an agent destruction plant. Although the TPC has considerable experience in pilot-scale and commercial-scale processing of dilute wastes for extended periods of time, it has stated that it is only in the early stages of commercial-scale operations and therefore does not have sufficient operational history to quantify the ratio of downtime to operational time. The TPC estimates about 20 percent downtime, based on its models, and will provide further information as it accumulates experience with current projects. Startup and Shutdown Procedures The panel had the following concerns about the startup and shutdown procedures provided by the TPC: Primary precautions are for keeping oxygen away from hydrogen. The TPC plans to purge the system with N2 and monitor for O2 during startup and shutdown to ensure that hydrogen and oxygen do not mix. The procedure for monitoring for hydrogen leakage out of the system during startup and shutdown was less clear. Gradual startup and shutdown appear to be necessary because of the thermal stresses. The possibility of surface deposits under moist shutdown conditions should be addressed. No criteria were provided for monitoring or assessing the stresses on materials or other damage to the system from an emergency shutdown. The reformer startup must follow a particular procedure to maintain an active catalyst. Maintenance The TPC appears to have considerable operational, and therefore maintenance, experience with full-scale treatment plants. For this technology, a full maintenance plan for an agent-treatment plant will require: (1) consistent implementation of routine maintenance or inspection; (2) objectives for maintaining barriers; (3) controls to prevent release of poisonous downstream gases; (4) attention in all procedures to conditions that might allow oxygen and hydrogen to mix; (5) process-

OCR for page 102
--> specific maintenance manuals; and (6) understanding of failure modes and specification of appropriate inspections to prevent them. Utility Requirements The TPC stated that the electrical energy requirement for the reactor heaters is 5,019 kW-h/day, that is, average power of 209 kW. The general electrical requirement for pumps, other heaters, lighting, etc. was given as 20,000 kW-h/day, or 833 kW. Total average electrical power required is therefore 1,042 kW. The fuel required during operation consists of propane plus part of the effluent gas from the reactor. In addition, the reaction is expected to be very mildly exothermic. The approximate energy requirements for treating 9 metric tons of HD per day are summarized in Table 6-2. below. The TPC's preferred rate of operation is somewhat lower, at 5 metric tons per day. Fuel will also be required for startup and for operation of the SBV. The amounts of fuel required have not been estimated. The TPC stated that the water requirement is 100 gallons per minute of clean water. For HD, this appears to be on the high side. Some water will be used in the steam feed to the catalytic reformer, and some will be used in the caustic scrubber. However, these are small requirements (a few gallons per minute). A larger amount will be needed for cooling (perhaps 20 gallons per minute). Scale-Up Requirements The GPCR process can be broken down into six subsystems: the agent (or waste) feed system, the SBV, TABLE 6-2 Daily Energy Requirements to Process HD at 9 Metric Tons Per Day Energy Source Rate of Use (MJ/h) Percentage of Total Electric power (833 kW) 3,000 28.4 Burn product gas 3,876 36.7 Burn propane 3,517 33.3 Heat of reaction 160 1.5 the reactor, the scrubber system, the catalytic reformer, and the evaporative-cooling/air-water treatment systems. Each of these subsystems consists of a number of unit operations. The various subsystems have not all had the same demonstrations of operability with scale-up. For example, the scrubber system required for HD will be quite different from the demonstrated system for chlorinated materials. The catalytic reformer, on the other hand, should be the same. As noted, the process has been demonstrated at pilot and commercial-scales for processing aromatic hydrocarbons and chlorocarbons. The TPC has said that all system components pertinent to the treatment of chemical agents are demonstrated commercial technologies. The panel believes, however, that demonstration is lacking in the following areas: handling and disposition of high concentrations of sulfur-containing products (primarily H2S) in reactor effluent gas, although commercial scrubbing technologies are available speciation and management of phosphorus-containing products in the reactor effluent gas, including scrubbing technologies or other methods for managing phosphorus-containing reaction products, as well as the final form and mode of disposal of high-volume process residuals containing phosphorus effects of reactor products containing sulfur and phosphorus on the catalytic reformer and mechanisms to avoid poisoning if these products are not fully recovered in the scrubber system (The TPC's experience has been with chlorine-containing materials, which do not present the same problem.) With only preliminary information derived from VX surrogate tests using malathion, and without complete information from the initial agent tests conducted late in the study process, the panel is unaware of the fate of the phosphorus in VX. Moreover, the TPC provided little detail on the scrubbing system, and it was difficult for the panel to verify some of the necessary oxidation-reduction chemistry in the TPC's proposed technology for scrubbing phosphorus compounds from the reactor effluent gas. The TPC provided little detail the on treatment of process residuals for ultimate disposal. Although the capability to cleanout emptied ton containers has not been demonstrated, the SBV has been demonstrated on inorganic matrices. The panel believes that the SBV hold-up times, temperatures, and reactant

OCR for page 102
--> gases are likely to suffice for this purpose. Laboratory-scale tests have demonstrated desorption efficiencies in excess of 99.9999 percent for organic residues in enclosed containers such as PCB-contaminated lamp ballasts. Process Safety As noted in the section on Process Safety in Chapter 4, the risk factors for process safety in all the alternative technologies can be divided into two categories: factors related to handling agent prior to its introduction into the specific technology and factors related to the agent destruction technology and associated system elements. The process safety risk factors related to the handling of agent prior to entry into this technology, which are common to all the agent destruction technologies, include storage risk, transportation risk, and the risk from the punch-and-drain operation. These factors can be exacerbated or ameliorated by unique aspects of a technology. For example, if the SBV treatment of ton containers is implemented, the risks in handling ton containers will differ somewhat from the risks for a technology that uses hot water and decontamination solution to bring the containers to a 3X condition. The process safety risk factors inherent in GPCR include safety issues associated with high temperature hydrogen, hot water and corrosives in the scrubbers, and secondary containment. Many of the risk factors that are not specific to chemical agent have been addressed by the TPC in safety analyses and in hazard and operability reports. The panel found no failure scenarios involving a loss of electrical power, loss of cooling, failures of pumps and valves, inadvertent overpressurization, or inadvertent temperature transients that would lead to off-site releases of agent or toxic process products. Based on the panel's preliminary and qualitative evaluation, the most significant off-site risk appears to be associated with handling agent prior to the agent destruction process. The principal risk factors appear to involve mishaps in 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 following subsections on process safety address risk factors specific to the GPCR technology. The panel expects that the safety issues discussed below can be resolved through further design and demonstration prior to constructing a full-scale facility. Off-Site Safety Issues The following issues should be addressed fully and clearly in a final GPCR process design. Hydrogen and Other Combustible Gases The process uses hydrogen. In addition to the hydrogen circulating in the process gas stream, most of which is produced in the steam reformer from methane produced in the main reactor, compressed hydrogen is stored in tube trailers and in the product gas tank. Other combustible gases (carbon monoxide, propane) are also present and must be considered. Hydrogen is commonly used in industry and can be used safely. Recent industrial accidents involving hydrogen are rare because of the care taken to handle it properly. The issue is mentioned here only because of the potential for a hydrogen explosion or fire to cause grave damage to personnel and structures if the hydrogen is not managed properly. Also, a hydrogen explosion could lead to a release of chemical agent. Although leaks of flammable gases are a risk factor for worker safety, they are not currently an off-site risk factor in either of the TPC's two current commercial operations. This was discussed above under Process Instrumentation and Controls. When a GPCR system is housed in secondary containment (as is required for agent destruction facilities), the potential increases for buildup of an explosive concentration of hydrogen. The potential increases for damage to agent-bearing structures from an explosion or fire. The containment of the system for an agent destruction facility will need to be designed so that the hydrogen will neither stratify nor build up locally to a combustible concentration. A large detonation or burn near the containers that store the agent could damage containment structures and cause a release of agent. This risk factor should be considered when designing component locations and shielding. The proximity of the hydrogen tube trailer, product gas tank, or any other combustible storage area to the agent-containing components (holding tank, reactor, SBV) is very important.

OCR for page 102
--> Another risk factor that must be considered is combustible mixtures of air and hydrogen inside the circulating gas system, which could result from air leaking into the system combined with flow imbalances. The current system does have design features and controls in place that address this risk in appropriate ways. Decontamination of Ton Containers in the SBV Extremely large doors are needed on the SBV to insert and remove ton containers. These large doors must be sealed tight to prevent leakage of agent and hydrogen. Proper sealing of the SBV at high temperatures can be an engineering challenge. In past operations, two types of seals have been used: glass-fiber gaskets and U-shaped silicon rubber seals with nitrogen gas pumped into the seal. The TPC has stated that it will probably redesign the seals for additional reliability in applying the SBV technology to agent destruction. Design of the Reactor Vessel The design of the reactor vessel needs to consider thermal stresses, welding problems, crevices, and local design problems. These issues, however, are no different for an agent processing facility than for the waste treatment facilities that the TPC has already piloted and run commercially. Worker Safety Issues There are a number of worker safety issues associated with high temperature hydrogen, high temperature steam, hot water and 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. The status of the technology with respect to these risk factors and the nature of the risks have been discussed above. Specific Characteristics that Reduce Risk Inherent in the Design The system operates at low pressure, and it appears to be extremely difficult to overpressurize the system inadvertently. Upon slight overpressure, the reactor is relieved to the caustic scrubber through an 8-inch pipe. The SBV chambers are relieved in a similar manner. There are no apparent ways for the reactor or SBV to fail because of overpressure; there are no valves between either the reactor or the SBV and the pressure relief mechanisms. Loss of electrical power, failure of cooling water to the heat exchanger, or failure of cooling to pumps will result in a "graceful" shutdown of the system. The integrity of the system does not appear to be threatened in any realistic failure scenarios. Schedule The TPC has stated, "the schedules for design, construction, testing and evaluation of a pilot-scale system have been requested by the Army and will be provided according to their requirements." The TPC states that the time for facility construction is about 6 months, with systemization taking another three months. In its submission, the TPC assumed that the Army will provide the secondary containment building and ancillary nonprocess facilities. Although the reactor, feed systems, and steam reformer have been deployed commercially, the lack of details on scrubbers and on handling phosphorus-containing materials could mean further development is necessary. Other than the increases in monitoring requirements, the design of secondary containment, and the engineering necessary for managing the sulfur and phosphorus wastes, this technology is at the point where a unit like the existing commercial systems could serve as the pilot operation for agent destruction. Still to be assessed are the effects on the schedule of designing the secondary containment and any associated reengineering. The effect on schedule is likely to be more severe for VX than for HD because the need for identifying and managing phosphorus-containing reaction products applies only to VX.