EP0634953A1 - Gas phase degradation of organic compounds - Google Patents

Abstract

The operation of a catalytic metal oxide reactor (24) is described which is capable of the complete mineralization of a gaseous stream including several organic pollutants. Several of the reaction constraints individually improve reaction completion and combinations of these factors permit complete degradation of the input gaseous contaminants in a manner not demonstrated previously. The factors include the nature and physical form of the catalyst, the lowering of the humidity in the input gaseous stream, the heating of the input gas stream and the reactor (24), and increasing the space time of the reactants in the reactor (24) beyond that necessary to degrade the initial contaminants.

Description

GAS PHASE DEGRADATION OF ORGANIC COMPOUNDS

Statement Regarding Federally Sponsored Research

This invention was made with United States Government support awarded by the Department of Energy (DOE), Grant No. PO AX0798826-1. The United States Government has certain rights in this invention.

Field of the Invention

This invention relates to the field of detoxification of inorganic and organic pollutants, and in particular to the field of gas phase, catalytic photo-oxidation of pollutants.

Background of the Invention

Ceramic membranes are used currently in industry and science for a variety of processes and purposes, the most common of which involve separation processes, catalysis, and adsorption. Ceramic membranes have increased in popularity because they offer several advantages over organic membranes. Ceramic membranes have a greater chemical stability since they are resistant to organic solvents, chlorine, and extremes of pH. Ceramic membranes are also stable at very high temperatures which allows for efficient sterilization of process and pharmaceutical equipment often not possible with organic membranes. Because ceramic membranes are inorganic, they are generally quite stable to microbial or biological degradation which can occasionally be a problem with organic membranes. Ceramic membranes are also mechanically very stable even under high pressures. The temperature, chemical, and mechanical stability of ceramic membranes allows them to be cleaned more effectively than other less durable membrane compositions.

One of the desirable attributes of many metal oxides formed of transition metals is that the transition metal can have catalytic properties desirable for certain photochemical or electrophotochemical reactions. These materials are also effective adsorbents. Metal oxide porous ceramic materials thus are potentially attractive candidates for use as catalytic or adsorbent agents in industrial scale chemical reactions.

Metal oxide ceramic materials are generally created through what is referred to as a sol-gel procedure. In such a procedure, the metal alkoxide is initiated into solution in a solvent, in a reaction vessel in which the solvent is rapidly being stirred. Depending on the process, the solvent may be alcoholic or aqueous. Whichever solvent is used, the metal alkoxide in solution is then hydrolyzed to create metal hydroxide monomers, dimers, polymers and/or particles depending on the quantity of water used. The hydrolyzing metal oxide particles in the solution tend to aggregate and to readily precipitate from solution. The hydrolysis process must therefore be strictly limited by the control of one or more aspects of the process to prevent precipitation of insoluble metal oxide solids from the solution. The insoluble metal oxide particles are thus, in essence, maintained in suspension until they are peptized by the addition of an acid, which causes the particles of the metal oxide to have a greater propensity to remain in suspension, presumably due to opposing charges acquired by the particles during the peptizing process. Such stabilization of the formation of particles has also been accomplished sterically by adding surfactant agents. The stable suspension thus produced, referred to as a sol, is then treated by removing the solvent therefrom to create a gel or semisolid material. Such gels may be subject to further solvent removal, and are then sintered or fired to turn the gel from a semisolid into a completely solid, rigid, durable material. Such ceramic oxide materials can be formed typically as coatings, or as supported or unsupported membranes, or as monoliths or other solid densified and vitrified objects. One limitation on previous catalytic or adsorbent processes using transition metals and materials made from them is the need to produce catalytic or adsorbent agents in a physical form that allows for a large degree of surface contact with the reaction substrates. Metal oxide materials are often most readily available in the form of films, solid particles, or crystals, but none of these physical forms has the large degree of surface area desirable for materials used as catalysts or adsorbents. While metal oxide materials can be coated onto relatively porous substrates, clearly it would be advantageous to have the substrate itself in a physical form which would be both stable and capable of convenient handling, and also which would have great surface area so as to make the catalytic or adsorbent agent available to the reaction substrates.

While ceramic membranes are one possible physical form that catalytic or adsorbent metal oxide ceramic materials may take, it is not the only one. One typical form of catalyst or adsorbent used in many industrial scale chemical reactions is a pellet. Such pellets can be loosely packed into beds or reactors. If the pellets are of a sufficient internal porosity, the vapor or liquid pressure drop through a reactor filled with such pellets will be within acceptable bounds . Since such pellets are often used in industrial catalytic or adsorbent processes, it is an advantage of any newly developed catalytic or adsorbent materials that they be capable of manufacture in a form which may readily be accepted by existing industrial applications. It has also been recognized for some time that metal oxide particles can catalytically photodegrade many toxic chemicals, such as volatile organic compounds (VOC), which are widely used as industrial solvents for such tasks as degreasing metals and for dry cleaning. Over the years, many VOC were improperly disposed of, or have leaked from underground storage tanks, polluting the soil and ground water. Urgent research efforts are underway to inexpensively and effectively decontaminate and remediate soil and ground water contaminated with carcinogenic and/or toxic VOC.

Prior research has tended to focus on easy-to-degrade compounds, such as acetate, and on the use of suspended particles to degrade such compounds. There are, for example, teachings in the prior art of the use of suspensions of titanium dioxide powder to degrade organic molecules. However, the use of suspended particles for these processes is a serious limitation, since solid substrates are clearly more convenient to utilize. For example, chlorinated organic compounds such as trichloroethylene (TCE) can be degraded in suspension or over a solid catalyst. When this reaction is performed in an irradiated slurry of Ti0₂, catalyst or water must be removed by filtration after the reaction occurs. In contrast, if volatile organic pollutants are degraded to complete mineralization in the gas phase by passage through a solid catalyst, post-reaction filtration is not necessary. While gas-phase oxidation of volatile organics has this advantage, one known disadvantage is that incomplete gas-phase mineralization of some VOC over a solid Ti0₂ catalyst may yield harmful side reaction by¬ products. One such harmful by-product is phosgene gas (C0C1₂) which can cause a fatal edema hours after exposure. While Ti0₂ in several different forms can catalyze the catalytic photo-oxidation of VOC, what is lacking is an optimized oxidation protocol which fully mineralizes volatile organic compounds such as trichloroethylene to C0 and HC1 in a reasonable amount of time without producing phosgene gas as a side product.

Summary of the Invention The present invention is summarized in that a method for gas-phase, catalytic photo-oxidation of volatile organic compounds to C0₂ and HC1 in a photoreactor involves optimizing several variables to ensure complete mineralization and to avoid production of harmful side reaction by-products. In producing solid, porous Ti0₂ catalytic bodies, advantageous results are obtained by using a very high percentage of anatase form of Ti0₂ and by ensuring that the porosity of the material produced is sufficient to provide catalytic surfaces not only for the reactant gases, but also for the intermediates produced during photo-oxidation.

Further improvements in the complete mineralization of toxic chlorinated organic compounds are achieved by heating the reactant gas input stream and the photoreactor to approximately 60°C before photocatalysis occurs; by allowing reactant gases and intermediate oxidation products sufficient space times in which to contact the catalyst; and by reducing the mole fraction of water in the input gas stream to a level that does not inhibit the photo-oxidation of the harmful organic compounds. It is an object of the present invention to define several improvements to the process of degrading organic compounds on porous solid metal oxide ceramic bodies. It is another object of the present invention to define improvements which singularly or in combination increase the extent of mineralization of chlorinated organic compounds and reduce the evolution of harmful side reaction by-products.

Other objects, advantages, and features of the present invention will become apparent from reference to the following detailed description and drawings. Brief Description of the Drawings Figs. 1-5 are schematic representations of the process for forming the metal oxide ceramic bodies used in the improved process of the present invention. Fig. 6 is an embodiment of a photoreactor useful for carrying out the improved process of the present invention.

Fig. 7 is a schematic depiction of a second embodiment of a photoreactor useful for carrying out the improved process of the present invention.

Figs. 8 and 9 are graphic representations of the dependence of TCE mineralization on space time.

Figs. 10 and 11 are graphic representations of the dependence of TCE mineralization on temperature. Figs. 12-14 are graphic representations of the dependence of TCE mineralization on water vapor mole fraction.

Fig. 15 is a graphic representation of the dependence of TCE mineralization on both space time and temperature. Fig. 16 is a graphic representation of the concentrations of various VOC of a reactant gas stream before catalytic photo-oxidation.

Figs. 17 and 18 are graphic representations of the concentrations of various VOC of a reactant gas stream after catalytic photo-oxidation.

Fig. 19 is a graphical representation of the inlet feed stream for the reaction of Example 7.

Detailed Description of the Invention The present invention is directed to improvements in a process which uses porous metal oxide ceramic bodies to degrade organic molecules. The existing process includes the step of preparing a photocatalytic ceramic body, exposing it to organic compounds in the gas phase, and photodegrading the organic compounds. The improvements to the process further define the characteristics of the ceramic body used in the process and further define the process conditions such that the organic compounds are fully degraded and do not generate toxic phosgene gas. The catalytic material used in the process is generally made by preparing a particulate sol which is dried to a semi-solid gel, extruded to form semi-solid pellets, and fired to achieve a ceramic material of defined pore size. The theory behind this approach is that a particulate metal oxide material is composed of a plurality of discrete metal oxide particles, which are packed during the gelation process, and which are fused in the sintering process to form a unitary material. Therefore the packing density of those particles determines the porosity of the final material. If the particles are packed in a uniform, maximum density, close- packing arrangement, about 30% porosity will be achieved. In this case, the size of the particles above will determine the size of the relatively uniform pores remaining between the particles in the resulting fused membranes. If, on the other hand, a looser packing of the particles can be obtained, then a material having a larger porosity will be achieved. It has been discovered that by controlling the degree of aggregation of particles in the sol stage by adjusting the pH of the sol, the porosity of the resulting ceramic material can be selectively adjusted within a range. Through this technique, the porosity of the resulting material can be increased beyond that which could be obtained by a close packing of the particles in the resulting ceramic material.

It is also advantageous that the particles making up the membrane be small. If the particles in the sol grow larger, potential catalytic and adsorbent surfaces are buried in the interior of the aggregated particles and therefore not available to substrates. By contrast, if small particles are agglomerated into irregular clusters, to make larger clusters, instead of making the particles larger by accretion, maximum surface site availability is maintained. Thus a process for catalytic or adsorbent materials should favor particle-to-particle aggregation over particle size growth.

The theory behind the approach used to achieve this result can be understood by the fact that since most metal oxide particles have high charge in a sol when the sol is a low pH water medium, the particles remain in stable suspension only because of the electric charge repulsion which they have for each other. If the pH is gradually increased by removing protons from the suspension created in this fashion, the particles will lose their electrical repulsion for each other and slowly aggregate into larger structures. The structure of these aggregates is believed to be more branched, or ramified, than similar aggregates which could be formed by adding electrolytes to the sol in order to achieve a similar effect through charge screening. The gradual removal of protons from the sol tends to form aggregations of particles, in contrast to an accretion process triggered by additional electrolytes which not only compresses the particles, but can lead to uncontrolled aggregation, flocculation, and ultimately precipitation of the particles from the solution.

Figs. 1 - 5 generally illustrate this concept. Fig. 1 is intended to illustrate the general relationship of metal oxide particles formed in a sol. In Fig. 1, the particles in the sol are repelling each other due to charge, and are thus spaced randomly about in the solvent (not shown). If those particles are solidified using a close packing model, by removing the solvent, a continuous porous ceramic material, such as a porous ceramic membrane, will be created which has a microstructure much like that viewed in Fig. 2. The material will be reasonably porous, with the size of the pores determined solely by the size of the particles, since the only pores are those formed between the particles in the close packing model.

In Figs. 3 to 5, the process for producing the ceramic bodies used in the present invention is schematically illustrated. In Fig. 3, again the metal oxide particles are indicated as they are formed in a sol. In Fig. 4, the first step of the porosity controlled process of the present invention is illustrated. In the sol at the stage of Fig. 4, the particles are aggregated by the slowly increasing pH into irregular ramified aggregates of particles . The aggregates of particles remain in solution or suspension in the sol, which can become significantly more viscous as a result. The sol containing these aggregates can then be dewatered, and sintered, into a material which resembles Fig. 5. This material as illustrated in Fig. 5 not only has small pores between the particles, but has, in addition, large vacuum or pore areas created between the aggregates of particles. Thus the density of the overall material is lower, and the porosity is higher, even though the particles themselves are the same. In this way, a high surface area, and a large porosity, are both combined in a unitary material formed as an aggregation of small particles.

One of the most useful features of ceramic oxide materials formed by this process is their utility in catalytic and adsorbent processes. It is therefore an advantage to be able to fabricate these materials in a form suitable for use in existing industrial processes which rely on catalysts or adsorbents. Since one common form in which catalysts or adsorbents are generally sold and utilized in industrial processes is in pellets, the fabrication of metal oxide ceramic materials in pellet form is desirable. It has been found that certain sols, such as those described herein, are, in their semi-solid or viscous form, capable of extrusion. If the sols from the sol-gel process are partially dewatered into a viscous but not yet completely gelified material, and the gelation is initiated but not completed, such partially solidified gels may be extruded through a suitable orifice to create elongated forms which may be cut, formed or otherwise reshaped into any desired size. Since the materials are semi-solid during the extrusion process, they retain the shape into which they are cut or formed if they are permitted to remain in that shape during the completion of the gelation process. In other words, gelation is started until the materials have a semi-solid form, then the extrusion or shaping process begins. Since these materials will generally retain their physical shape, other possible shaping or molding operations are also possible. Once the shaping process is completed, the gelation is allowed to continue to completion, to create shape retaining solid xerogel materials. Those solid gelled materials can then be sintered, or vitrified, by firing into ceramic materials which retain the shape of the shaped gels. By this process, continuous extrusions having any desired cross-sections, such as circular ones, can be created. Through the cutting of the continuously extruded ribbon, sticks or pellets of any desired length can be created. Such pellets, when appropriately fired, retain their shape, are rigid and stable, and can be packed in any suitable reactor vessel or container as may be utilized in industrial applications.

While the method and product disclosed below are illustrated in particular with metal oxide ceramic membranes of titanium, it has been described previously that ceramic material fabrication methods proven to be effective with titanium may also be adapted for use with other transition metal oxide materials, as well as those of other metallic elements such as silicon and aluminum. The method and product of the present invention has utility for these other metals as well. The method described below is exemplified with titanium, in particular, because titanium is considered one of the more difficult metals to work with due to its strong tendency to rapidly hydrolyze and precipitate from solution. Titanium also has particularly unique and advantageous catalytic and photocatalytic properties which many other transition metals do not possess to the same degree. In general, the improved process of the present invention begins with the creation of a metal alkoxide. Metal alkoxides are commercially available chemical supplies of reasonable cost. It has been found that such materials may be conveniently used in an aqueous or organic alcohol sol-gel process to make metal oxide ceramics. This has been done conventionally by hydrolyzing the metal oxide so that it precipitates, and then resuspending the metal oxide particles into the solution by peptizing the solution with nitric acid. Alternatively, in very acidic aqueous solution, the metal oxide precipitates are immediately redissolved. In any event, the product created by either of these processes is an acidic metal oxide solution, containing relatively small metal oxide particles in a highly acidic environment. Over time, if not otherwise treated, the metal oxide particles will tend to accrete in size, forming larger and larger particles. By screening the proton-dependent charge on the particles through the addition of inert electrolytes such as NaN0₃, or KC1, one destabilizes the particles and causes aggregation or flocculation. It has been found, however, that increasing the pH in the suspension, by slowly removing protons from the solution, results in the aggregation of these particles into larger aggregates, such as illustrated in Fig. 4, without precipitation or flocculation. This can be done advantageously by dialyzing the acidic solution against pure de-ionized water to slowly remove protons from the solution. The pH of the solution needs to be monitored during dialysis, to control the porosity of the resulting product. Within limits, the higher the pH to which the sol is raised, the larger the porosity that the gel produced from the sol will exhibit. However, the zero point of charge (ZPC), or the pH value at which the metal oxide molecules carry no net charge, for the metal oxide is the limit to which the pH can be increased. Approaching the ZPC point will initiate the gelation of the solution within the dialysis process. As the solution slowly increases in pH, it will increase in viscosity slowly but gradually until the maximum point, at which gelation occurs. Therefore, in practice of the present invention, if maximum porosity is desired, the solution shall be raised to a pH close to, but not quite at, the gelation point of the sol. As a pragmatic limit, 0.5 pH lower than the ZPC is the best condition for practical utility. For a titanium dioxide membrane, such a desirable pH is pH 4. Other metals will require slightly altered pH values depending on the ZPC for the material. Once the dialysis has been conducted to the point at which the appropriate amount of protons are removed from the solution, and it is in a semi-solid or viscous state, the gel can be formed. This can be done by evaporating the remaining water from the sol in a desiccator box, in a rotary evaporator, or in some other fashion. Since the physical shape of the gel will be the same as the resulting metal oxide ceramic material, if membranes or planar layers are required, the gel should be plated or layered onto a support. Alternatively, if other shapes, such as those described below, are to be utilized the gel may be suitably manipulated as desired.

Once the gel has been appropriately dewatered, it then may be fired, at temperatures up to 400 °C, to produce stable porosity-controlled ceramic materials, the porosity of which will be dependent upon the pH to which the sol was raised during the process of proton removal.

It has been found that a particular advantage of this procedure is that during the gelation process, the semi-solid gel can be manipulated into any desired physical shape. The sols are preferably concentrated to a high degree of viscosity. This can be done by using a rotary evaporator. The viscous sol can be removed from the rotary evaporator, and then reshaped using any suitable reshaping device or process. It has been found, in particular, that the viscous sol, or semi-solid gel, can be extruded through a confined orifice under pressure. This has been done utilizing a simple plastic syringe. The solidified, semi-solid gel can be loaded into the syringe, and then manually ejected from the syringe mouth, resulting in a tube or ribbon of soft solidified gelled material having a diameter of the orifice, and which can be cut into any desired length. The shaped, solidified, soft gel can then be desiccated and fired to result in ceramic bodies that are crack free, stable in shape and size, and have a high degree of porosity. The resulting ceramic bodies will have shrunk, during the desiccation process, to a size approximately one half of that in each dimension of the solidified soft gel material from which they were sintered.

Using the porosity-control process for making metal oxide ceramic materials, it has been found that the porosity of the material can be increased from 25 to 50% when the pH of the sol increases from 1.5 to 4.0 by dialysis using titanium dioxide. The mean pore diameter of the titanium dioxide xerogels produced by the procedure is generally within the range of single digit nanometers. It is to be understood that the mean pore diameter is, however, an aggregate of both the small pores contained between the individual particles, and the larger pores created between the aggregates of the particles. The sintering process will reduce the porosity slightly. However, it has been found that larger porosity in the xerogel stage still results in larger porosity in the resultant ceramic product than that obtained from similar xerogelβ having a lower porosity at an analogous stage in the process.

Utilizing the extrusion and cutting technique to make porosity controlled ceramic pellets, it has been found that porosity within the range of 30 to 50% is readily achievable from these sols. The shaped ceramic bodies have high surface area, in the range of 50-400 m²/g of material. The resulting ceramic pellets are translucent and crack free. The translucency is believed to be a result of small particles forming up the materials, which are so small as to be incapable of refracting or interacting with light in a visible wave length.

The resulting titanium ceramic membrane functions as a highly desirable substrate for the photo-catalyzed degradation of organic molecules. The surface of the ceramic bodies are highly porous, thereby readily adsorbing organic molecules . The titanium molecules are readily available for catalytic activity. The catalysis is actuated by UV light, and broad spectrum UV radiation, even sunlight, is usable, although intense artificial UV light may tend to enhance the speed of the degradation. In combination with the improvements of the present invention, such ceramic bodies are able to completely mineralize certain chlorinated organic contaminants to C0₂ and HC1 in the presence of UV light. No other photo- oxidation system using a solid catalyst known to the present inventors is capable of essentially complete mineralization, without producing toxic phosgene gas (C0C1₂) which can cause a fatal edema hours after exposure to the gas.

Others have reported the complete mineralization of TCE in a dilute (50 ppm) aqueous solution by irradiating a 0.1 or 0.3 wt% slurry of Ti0₂ with UV light. However, catalytic photo-oxidation in suspension is difficult because it requires filtering to remove the catalyst or water from the oxidation reaction. To avoid this difficulty, the art has moved in the direction of solid phase, supported catalysts through which polluted liquids or gases may pass. One such approach has been to coat Ti0₂ catalyst onto a fiberglass mesh. A second approach performs gas-phase catalytic photo-oxidation of organic contaminants in a fixed bed photoreactor or in a flat-plate fluidized bed photoreactor using Ti0₂ supported on silica gel. While both of these systems degrade TCE, neither is capable of complete mineralization, always yielding some amount of phosgene gas. Complete mineralization of TCE (C1₂C=CHC1) has been postulated to proceed along two parallel pathways, represented by Equation (1 ) and by the combination of Equations (2) and (3) .

C1₂C-CHC1 + H₂0 + 3/2 0₂ - 2C0₂ + 3HC1 (1) C1₂C-CHC1 + 2H₂0 - ClCHj-COOH + 2HC1 (2)

C1CH₂-C00H + 3/2 0₂ - 2C0₂ + HC1 + H₂0 (3)

As the preceding stoichiometric equations indicate, complete mineralization of 1 mole of TCE generates 2 moles of C0₂. Accordingly, the ratio of C0₂ generated to TCE degraded indicates the extent to which TCE has been fully mineralized. At ratios less than 2, mineralization is incomplete, presumably because intermediates have not been fully converted or because side reaction by-products have been generated from TCE degradation products.

Although catalytic photo-oxidation has been achieved before, the improvements of the present invention enhance the conversion of TCE and intermediate degradation products into C0₂ and HC1. By refining the catalytic photo-oxidation process, the percentage of TCE oxidized increases and the opportunity for generating toxic phosgene gas is minimized. While each improvement factor described below shows advantageous benefits when compared to existing processes, the combination of two or more of these factors apparently show synergistic benefits resulting in complete mineralization more readily than can be achieved with any single improvement to the process. In any event, each improvement factor alone brings about increased oxidation of TCE to C0₂ and HC1 than is possible using existing degradation processes.

The improvement factors of the present invention are directed to several aspects of the original process. One process for degrading complex organic molecules, as set forth in U.S. Patent 5,035,784, involves exposing and adsorbing organic molecules in the liquid or gaseous phase to a porous metal oxide ceramic body, such as a titanium dioxide membrane, and irradiating the adsorbed molecules with ultraviolet light. Two improvement factors go to the process for forming the porous metal oxide ceramic body. The first improvement concerns the anatase form of titanium dioxide incorporated into the catalytic substrate. Titanium dioxide occurs naturally in three different crystal forms known as rutile, brookite, and anatase. Because anatase is believed to be the preferred catalytic form, a titanium dioxide membrane having an increased proportion of anatase is believed to provide a more reactive catalyst. While other solid phase Ti0₂ catalysts contain anatase, ceramic bodies that approach 100% anatase permit faster and more thorough catalytic photo-oxidation because they provide more functional catalytic sites per square centimeter of catalyst than catalysts having less anatase. Titanium dioxide materials constructed in the fashion described in this specification will be essentially all anatase.

The second improvement factor relates to a dramatic increase in catalytic substrate porosity, which apparently also increases the available catalytic surface on which photo-oxidation can occur. In addition, the smaller pores present in the membranes trap reactant, contaminants, and intermediates for longer periods of time than in similar bodies with larger pores. Longer contact time provides more opportunity for catalytic photooxidation of reactants and intermediates and tends to generate more final product and fewer harmful intermediates. The improved ceramic bodies have 50-60% porosity which translates to available surface area of 150-190 m^z/g. The porous ceramic bodies having increased porosity have been shown by the inventors to contain a bi odal distribution of pore diameters, with a first peak of pore diameters in the range of 10 Angstroms and a second peak of pore diameters in the range of 35-50 Angstroms. The bimodal pore size distribution apparently arises during production of the porous ceramic bodies. The smaller pores result from the close packing of small aggregates of titanium dioxide particles. The larger pores may represent the packing of agglomerated aggregates which form during sol-phase dialysis. Because the agglomerated aggregates are larger than the small particles, they cannot pack as closely as the smaller particles, forming larger pore sizes between packed agglomerates.

The following three improvement factors relate directly to the process conditions used in degrading complex chlorinated organic contaminates on a porous solid metal oxide ceramic body. While it is very useful to combine these three factors with the materials described above, these factors may also be used in processes based on other titanium dioxide materials. The third improvement factor goes to increasing the temperature of the catalytic pellets and of the reactant gas stream which passes through the catalytic pellets. Temperatures near 60 °C seem optimal for eliminating phosgene production. In combination with the other improvements of this invention, heating also appears to increase the complete mineralization of TCE. The fourth improvement factor is the reduction, but not elimination, of water from the input gas stream. Excess water may promote certain side reactions that yield small percentages of chloroform and carbon tetrachloride in the presence of perchloroethylene. On the other hand, if water is absent from the input reactant gas stream, the mineralization of TCE quickly trails off, presumably because the OH" ions on the Ti0₂ surface are rapidly consumed and are not replenished as they would be if water was present. The fifth improvement factor is the recognition that significantly longer space times are required to allow catalytic photo-oxidation of the initial TCE-degradation intermediates. Space time is a measure of the length of time in seconds that a mole of reactant gas remains in contact with a gram of catalyst. Space time is expressed as the amount of catalyst, in grams, divided by the molar flow rate of the reactant gas, in mol/s. Expressed another way, space time is weight of catalyst times time divided by mole of contaminant in feed gas. Space times typical in the art of catalytic photo-oxidation in the gas phase are on the order of IO⁶ to 10* g s/mol. In contrast, the present inventors have demonstrated that space times in the range of 10* to 10¹⁰ g s/mol are preferred, because they allow sufficient resident time to accomplish complete mineralization of TCE.

Shown in Fig. 6 is a diagram of one embodiment of a packed bed tubular gas phase photoreactor used to degrade TCE according to the process of the present invention. Photoreactor housing 10 included lamp supports 12 and surrounded a packed bed tubular photoreactor 14 and UV lamps 16. The photoreactor 14 mounted within housing 10. Fluorescent blacklight UV lamps 16 (two shown), such as 4W GE F4T5-BLB bulbs, surrounded photoreactor 14 at the photoreactor's top, bottom, and sides. Four bulbs were used by the present inventors, though by adjusting the output of the bulbs, more or fewer bulbs could be used. Photoreactor 14 was formed of a cylindrical clear glass tube packed with Ti0₂ pellets (not shown). Inlet stream tubing 18 connected to one end of photoreactor 14 and outlet stream tubing 20 connected to the opposite end of photoreactor 14. Inlet stream tubing 18 and outlet stream tubing 20 passed through the walls of photoreactor housing 10 and connected on the inlet side to a source of reactant gas and on the outlet side to analytical instrumentation such as a gas chromatograph (not shown), for analyzing effluent gas. The inventors used a Hewlett-Packard 5890 equipped with a Porapak R column using a flame ionization detector or a thermal conductivity detector, though any analytical equipment able to detect photo-degradation products would suffice.

The inlet stream tubing 18 was wrapped in heating tape which was itself connected to a temperature controller (not shown). Any heating tape and thermostat controller combination capable of monitoring and adjusting the inlet gas temperatures within a narrow range would be acceptable. The combination used by the inventors was Thermolyne heating tape and a Glas Col PL312 Temperature Controller.

In use, the photoreactor 14 of Fig. 6 accepts reactant gas stream through inlet tubing 18. The reactant gas stream was heated to desired temperature and UV lamps 16 were ignited. UV-mediated photo-catalytic degradation of organic compounds in the reactant gas caused the emission through outlet tubing 20 of effluent gas containing reaction products.

To analyze the effluent gases, a portion of the effluent was directed to the gas chromatograph. Helium was used as a carrier gas at 30 ml/min. The oven temperature was held at 40°C for 2.5 minutes, then increased at a rate of 50°C per minute and held again at 185°C for 9 minutes. C0₂ and TCE retention times were 1.5 and 11.6 minutes, respectively. Sampling was carried out using a 6-port automatic gas valve made of Hastalloy C to protect against corrosion by HCl. The total amount of C0₂ produced in some experiments was also measured by gravimetric analysis. Carbon dioxide produced was mixed with barium chloride and sodium hydroxide to obtain barium carbonate which was detectable by gravimetric analysis . Gravimetric analysis is not useful for measuring very small amounts of C0₂ as a result of potential error introduced by atmospheric C0₂.

Fig. 7 schematically shows a second embodiment of a photoreactor apparatus 24 used with the process of the present invention. In this embodiment, the inlet reactant gas stream could flow along an exhaust pathway 26 or along a photo-oxidation pathway 28. The exhaust pathway 26, reached by opening valve VI, allows the inlet reactant gas stream to vent to the atmosphere, after filtering through a granular activated carbon (GAC) column. Before venting, the gas stream may be optionally sampled at inlet sampling port 30, by opening second valve V2. Alternatively, inlet reactant gas stream can flow into the photo-oxidation pathway through a metering valve MV1.

The photo-oxidation pathway includes a first coiled copper tubing (1/4" o.d.) 32 immersed in an ice bath 34 and terminating in a glass container (not shown) at the bottom of coil 32. A second copper coil 34, connected at its first end to the glass container, extends out of the ice bath and is connected at its second end to 1/4" o.d. 316 stainless steel tubing 36. Valves V3 and MV2, a pressure gauge 38, flow meter 40, and thermometer 42 are attached onto the stainless steel tubing 36. Liquid crystal temperature indicators (not shown), such as Omega RLC-10, were attached to the outside of photoreactor cells 44. A:relative humidity meter (Fisher Scientific, not shown) was used to measure the humidity of the gas stream. The various gauges and meters permit accurate monitoring of the temperature, pressure, flow rate and humidity of the gas entering the photoreactor. The stainless steel tubing 36 terminated at a 4-way inlet manifold 46 designed to split the single inlet gas stream into 4 parallel inlet streams. The inlet manifold 46 was made of 1/2" o.d. 316 stainless steel tubing sealed at both ends and arrayed with silver soldered inlet and outlet nipples (not shown). Short pieces of Norprene plastic tubing (3/16 ") (not shown) connected the inlet manifold outlet nipples to the photoreactors 44. On the photoreactor outlet side, a 4-way 3/4" o.d. 316 stainless steel outlet manifold 48 constructed like the inlet manifold combined the four photoreactor outlet streams into a single effluent path.

As in the embodiment of Fig. 6, inlet gas temperature control was achieved by a thermostat control (Glas Col PL312, not shown) attached to heating wire (Cal-Cord) (not shown) which was wrapped around the second copper coil 34, stainless steel tubing 36, and the inlet manifold 46. Heat loss was minimized by covering the photoreactor cells and inlet lines with insulation (not shown). As noted above, four identical annular packed pellet photoreactors were operated in parallel in this embodiment. Any number of photoreactors may be operated in parallel, as long as the inlet manifold is modified to evenly divide the inlet gas stream among the photoreactors. Each photoreactor 44 of Fig. 7 is an annular packed-pellet ring surrounding a UV light source. The glass ring was blown from Corning 7740 borosilicate glass tubing. The inner glass tubing had a diameter of 2.8 cm and a thickness of 1.5 mm. The outer glass tubing had a diameter of 3.8 cm and a thickness of 2.0 mm. The cross-sectional area of photoreactor 44 was 2.9 cm² and had a height of 13.4 cm. In the center of each annular photoreactor was a single 4W fluorescent blacklight lamp (not shown), such as a GE F4T5-BLB with a photon flow rate of 1.24 x 10"' einstein per second using uranyl oxalate as an actinometer.

Short pieces of Norprene plastic tubing (3/16 ") connected the outlet manifold inlet nipples to 316 stainless steel tubing 50. A photoionization detector (not shown) provided rough estimates of the amount of organic contaminants at the inlet and outlet sampling ports. Tedlar or Teflon bags with stainless steel valves and a septum port (not shown) could be attached to inlet and outlet sampling ports for sample storage and subsequent sensitive GC/MS monitoring of VOC in the inlet and outlet gas streams.

The components downstream of second copper coil 34, including the four parallel photoreactors, were installed in a portable suitcase (13" by 21" by 7.5") along with a switch and ballasts for the lamps (not shown).

The catalyst packed into the annular ring of photoreactor 14 and photoreactor 44 was prepared by a sol- gel method as outlined in Example 1, below. Pellets were fired at 300°C and had a final diameter of 1 mm. As measured by BET analysis, the porosity of the pellets was within a range of 50-56% with a specific surface area of 160-194 m²/g. The pellets were pure anatase as determined by powder x-ray analysis.

An inlet gas stream directed along photo-oxidation pathway 28 through V1 was dehumidified by condensing water into the glass -container at the bottom of ice bath 34. Dehumidified gas was heated to a desired temperature and passed through tubing 36 into inlet manifold 46 where it was split into four identical streams. Each stream passed into a photo-reactor 44 where VOC in the stream were photocatalytically degraded by UV light from the UV lamp at the center of annular photo-reactor 44. Effluent gas passed into outlet manifold 48 where it was combined into a single effluent gas stream. The effluent gas stream was then either released to the atmosphere after filtering through a granulated activated carbon filter or was sampled for later analysis at outlet sampling port 52. In a third, larger embodiment similar to the second embodiment, the size of the annular photoreactor is increased both in inner diameter and in length. By increasing the diameter and length of the annular photoreactor, the intensity of light reaching the catalytic material is increased, thereby increasing the catalytic efficiency of the photoreactor.

Each larger photoreactor is an annular packed-pellet ring surrounding a UV light source. The glass ring was blown from Corning 7740 borosilicate glass tubing. The inner glass tubing had a diameter of 5.4 cm. In the center of each annular photoreactor was a single fluorescent black light lamp with a photon flow rate of 33.5 x IO⁶ Einstein per second using Uranyl oxalate as an actinometer.

The larger photoreactor, integrated into the gas stream components of the previous embodiment, provides a more intense light than that of the smaller version. As determined from the diameter and length of the light and from its increased output over a wider area, the larger photoreactor provides a light intensity of about 1.62 x 10* E/cm²s. In contrast, the intensity of the smaller photoreactor is about 1.23 x 10* E/cm²s. It has been determined by the inventors that even though the photoreactors can hold large amounts of catalyst (e.g., on the order of 350-500 g in the larger photoreactor), the vast majority of that catalyst absorbs no UV light. Almost all of the UV light is absorbed in the top few microns closest to the UV light (e.g., 2-10 μ). Therefore, it would be more efficient to provide a much thinner layer of catalyst that is just a few microns thick. In such an arrangement, virtually all of the catalyst would be exposed to UV light. It would be preferred that the thin catalyst layer be coated onto a surface to a thickness of between about 0.1 and 2 μ. The surface could be any non-UV light-absorbing porous or non- porous support material, including but not limited to glaβs beads and glass fiber optic fibers. By reducing the thickness of the catalyst bed, it is believed that the amount of catalyst required can be reduced by perhaps a factor of between about 300 and 1000. These numbers are based upon the assumption that the existing catalyst bed

HEET RULE 26 is about 3 mm thick while light can penetrate the catalyst only to about 3-10 μ.

Practically speaking, the reduction in amount of catalyst will offer dramatic improvement in reaction efficiency which is typically expressed in terms of the space time needed to degrade a particular compound. Space time takes the units of gs/mol and is a measure of the amount of catalyst required to degrade 1 mol of a compound in the feed stream in a fixed time. A typical calculation of space time presumes that the entire catalyst bed absorbs UV light. Since the inventors have demonstrated that this is not the case, it is believed that dramatic reductions in space time are achieved by increasing the penetration of light into the catalyst bed. For example, in a photoreactor containing 379 g of catalyst, PCE gas in the feed stream at 500-1000 ppm requires 6.2 to 12.4 x 10* gs/mol at a flow rate of 1 liter per minute. At faster or slower flow rates, the space times are proportionally larger or smaller, respectively. Based on the calculations described above, and assuming a 1000 fold reduction in amount of catalyst, the same degradation could be achieved in 6.2 to 12.4 x 10^s gs/mol. By increasing the flow rate, it is reasonable to believe that space times as small as 1 x IO⁵ are possible. At flow rates in the range of up to 4 liters per minute, PCE at concentrations between 500 and 1000 ppm are degraded at space times between about 1.6 and 25 x 10* gs/mol using the larger annular photoreactor.

It is also herein demonstrated that the ability to degrade particular compounds is optimized at particular temperatures. Heating to temperatures near 60°C seems to increase the complete mineralization of TCE and to eliminate phosgene production. Higher temperatures, around 100°C are preferred for eliminating PCE. Another factor governing the space time requirements of this photocatalytic system is the relative difficulty or ease with which particular compounds are degraded. The degradation of each organic molecule falls somewhere along the spectrum of difficulty. Carbon tetrachloridβ is considered a very difficult molecule to degrade while formic acid is considered relatively easy. PCE is considered a difficult molecule, though less difficulty than carbon tetrachloride. Benzene is more difficult to degrade than PCE, owing to its conjugated structure. It is believed that by increasing the space time, even very difficult molecules can be degraded. As is detailed below, the inventors have demonstrated complete photodegradation of benzene in the presence of oxygen and a small amount of water at a space time of 2.2 x IO⁹ gs/mol. The process described herein can generally be used with halogen substituted alkyls, aromatic hydrocarbons, with or without halogen substitutions, and sulfonated and nitrogen substituted hydrocarbons having between 1 and 12 carbon atoms.

The inventors believe that a 1-2 order of magnitude reduction in space time is likely when the molecules being degraded are easier to degrade than TCE or PCE. This range of improvement is reported in liquid phase photocatalysis. Matthews reported about a 1 order of magnitude difference between a hard-to-degrade organic such as carbon tetrachloride and an easily degraded molecule such as formic acid. Similar results would be expected in the gas phase. Moreover, the rate of photocatalysis is much faster in the gas phase than the liquid phase. Whereas liquid phase degradation occurs over minutes, hours, or days, gas phase degradation takes place in seconds or less. The benefits of the improvements of the present invention will become clear upon consideration of the following examples.

EXAMPLES

Example 1 Preparation of Catalytic Photoxidative

Metal Oxide Ceramic Bodies

A titanium dioxide porosity controlled membrane was created using a peptized sol approach. 180 ml of distilled deionized water was mixed with 1.3 ml of concentrated ACS certified HN0₃. Then 15 ml of titanium tetra-isopropoxide (TifOPr¹) ) was added drop-by-drop into the acidic water while stirring. Care was taken to minimize exposure of unused Ti ( 0Pr-J « to water, especially high humidity, by handling the Ti ( 0Pr-J ₄ in a glove box under dried nitrogen or by quickly recapping the bottle. Alternative, TifOPr¹)* may be dispensed from a sealed separatory funnel. Precipitation of particles occurred immediately after addition of the metal alkoxide. The precipitant was then stirred continuously for three days at room temperature until peptized, producing a metal oxide sol. The material was transparent, slightly milky and bluish. The peptized sol was then dialyzed using SpectraPor 3500 M.W.-cutoff dialysis tubing (54 mm flat width) against Milli-Q water for 3-4 days until the resulting partially gelled material had a final pH near 4. To form porous titanium dioxide ceramic bodies, such as pellets, useful in the process of the present invention, a controlled porosity titanium dioxide sol was created through the process described above. The sol was then concentrated in a rotary evaporator at 50^βC just until the sol became viscous. * The viscous sol was then removed from the rotary evaporator flask and placed into a plastic syringe mold, in which it was allowed to stand at 30-40°C for 2-3 days. At the end of that time, the sol in the mold had gelled and solidified. The syringe mold had an orifice of about 2 mm. The soft solidified gel was extruded through the syringe mouth in a continuous ribbon onto plastic weighing boats or, preferably sheets of

Teflon. The ribbon was allowed to dry in a desiccator box at room temperature for about 1 week. The ribbon cracked into small fragments during this drying. The xerogel pieces, thus created were fired at a temperature of 300 °C for 4 hours. The pellets thus created were transparent and could easily be perceived to be porous due to their relatively light weight and low density. The fired pellets experienced about a 50% shrinkage in size from the xerogel pieces. The pellets were cut further to obtain a diameter of about 1 mm. The pellets were measured by nitrogen adsorption, and found to have a porosity of 50- 56% and a specific surface area of 160-194 m²/g. The pore 5 diameter distribution in the pellets was bimodal, with peak diameters around 10 Angstroms and 35-50 Angstroms.

Example 2

Effect of Porosity, Composition and Space Time on TCE Mineralization

10 To determine the effect on TCE mineralization of space time, an inlet stream of reactive gases having 445 ppm TCE, 3 x IO'³ mole fraction H₂0 and 0.02 mole fraction 0₂ was passed through fresh Ti0₂ pellets prepared according to Example 1 for a variety of space times in the apparatus

15 of Fig. 6. As shown in Fig. 8, the vast majority of TCE is degraded at space times as low as 8 x 10* g s/mol. This piece of information, however, tells only part of the mineralization story. When the ratio of [C0₂] produced/[TCE] degraded is measured over the same range of

20 space times, Fig. 9, it is apparent that mineralization is far from complete at 8 x IO⁶ g s/mol, even though little or no TCE remains detectable. At space times greater than 8 x 10* g s/mol, additional C0₂ is evolved, though even at 8 x IO⁷ g s/mol, the ratio reaches only 0.8. Complete

25 mineralization of 1 mole of TCE would generate 2 moles of C0₂. Hence, some portion of the input TCE is presumably converted to a more stable intermediate which requires a significantly longer space time to itself convert to C0₂ and HC1, or some degradation products have been converted

30 into other products, such as phosgene gas, in side reactions.

A likely intermediate was identified by degrading input TCE at a space time of 1.1 x 10⁷ g s/mol and scrubbing the outlet gas stream in deionized water. By

35 titrating the scrubbing water with silver nitrate, the inventors measured free chloride ions in solution. GC measurement of TCE degradation and silver nitrate titration of CI" ions together indicated that 2.1 +/- 0.2 moles of HC1 were produced per mole of TCE degraded. Since the TCE starting material contains three chloride ions per mole, the production of approximately 2 moles of chloride ions per mole of TCE degraded suggested that the catalyst-bound intermediate retained a single chloride ion or that a volatile compound containing a chloride ion was formed and not detected. Later exposure of these pellets to illumination in a humid air stream with no TCE present generated considerable C0₂, thereby implying that material had been retained on the pellets. Furthermore, since C0₂ production is quite low at the space time used in this experiment, the intermediate probably contained 2 carbons, like TCE. By analyzing the data derived from other chlorinated organic oxidations, the inventors believe that the intermediate is likely to be monochloroacetic acid (MCA) . Diffuse reflectance FTIR spectroanalysis of the Ti0₂ pellet surface was also consistent with the notion that the intermediate bound to the catalyst was MCA.

Example 3

Effect of Porosity, Composition and Temperature on TCE Mineralization

An input reactant gas stream having 350 ppm TCE, 0.02 mole fraction 0₂ and 0.003 mole fraction H₂0 was passed through 0.06 g of Ti0₂ pellets prepared according to Example 1 at a flow rate of 270 ml/ in through the apparatus of Fig. 6. As shown in Fig. 10, the rate of disappearance of TCE did not vary over a temperature range of 23 to 62^βC. In contrast, as shown in Fig. 11, there was a marked increase in evolution of C0₂ per mole of TCE degraded at higher temperatures. At 62°C, approximately 1.1 moles of C0₂ were produced for each mole of TCE degraded. While this does not represent the theoretical maximum of 2 moles of C0₂ per mole of TCE, it is significantly higher than the 0.4 moles evolved at lower temperatures. This result suggests that the intermediate MCA identified in Example 2 and other possible intermediates degrade at a higher rate at increasing temperatures.

Example 4

Effect of Porosity, Composition and Humidity (Water Vapor Mole Fraction) on TCE Mineralization

An input reactant gas stream having 340 ppm TCE, 0.02 mole fraction 0₂, and a range of water vapor mole fractions between 4.2 x 10"* and 8.6 x IO"³ to 0.010 was passed through the Ti0₂ pellet catalyst of Example 1 in the apparatus of Fig. 6. The disappearance rate of TCE was independent of the water vapor mole fraction across this range, as shown in Fig. 12. Indeed, a similar reaction performed at 65 °C with water vapor mole fractions ranging between 0.007 and 0.027 was similarly independent of the water vapor mole fraction, although the absolute TCE disappearance rate was somewhat higher at the higher temperature, as shown in Fig. 13. Despite this seeming independence, it does appear that some water is necessary to successfully degrade TCE. As shown in Fig. 14, when the input reactant gas stream contained no water, and when the Ti0₂ catalyst is heat-dried to remove surface water, the TCE degradation reaction slows quickly, presumably for lack of necessary OH" ions on the Ti0₂ catalytic surface. Apparently because OH" ions are used up during TCE degradation, the catalytic activity of the Ti0₂ catalyst diminishes over time.

Example 5

Effect of Porosity, Composition, Space Time and Temperature on TCE Mineralization Drawing upon the observation in Example 2 that C0₂ production increases at space times large enough to allow degradation of the MCA intermediate within the reactor, and upon the observation in Example 3 that more C0₂ is produced at higher temperatures, the inventors monitored the ratio of C0₂ production per mole of degraded TCE at two temperatures and at a range of high space timeβ in the Ti0₂ pellets of Example 1 in the apparatus of Fig. 6. As shown in Fig. 15, the theoretical maximum amount of C0₂ production per mole of TCE degraded was reached at approximately 6.1 x IO⁷ g s/mol when the input reactant gas stream and photocatalysis apparatus were held at 63.1°C.

Example 6

Effect of Porosity, Composition, Space Time,

Humidity (Water Vapor Mole Fraction) and

Temperature on TCE Mineralization VOC-containing air samples drawn from a soil vapor extraction well at the Department of Energy Savannah River facility in Aiken, SC were introduced via an oil-less sample pump (Gast) into the photoreactor apparatus of Fig. 7 containing Ti0₂ pellets of Example 1. Because the photoreactor had a maximum flow rate smaller than the sample pump, a portion of the sample was diverted to a granular activated carbon column before atmospheric release while a second portion was catalytically photo- oxidized. The mole fraction of 0₂ in the reactant gas stream was 0.2 and the humidity was 40% at 34 °C, indicating a mole fraction of water of 0.022. That portion of the inlet gas stream which would pass through the photoreactors was first dehumidified by passage through a water trap which condensed most of the humidity from the sample into a glass container at the bottom of a vessel containing ice-water. The mole fraction of water actually entering the photoreactors was 0.007.

Detectable VOC in the input sample included 1,1,1- trichloroethane (1,1,1-TCA, 64-160 ppm), trichloroethylene (TCE, 222-1100 ppm), tetrachloroethylene (PCE, 1900-7000 ppm) and 1,1-dichloroethylene (1,1-DCE, 30-55 ppm). While the gas concentrations varied widely over time at the inlet, as shown in Fig. 16, the concentration ratio of each gas to TCE remained constant. These fluctuations did not seem to have any relationship to the ambient temperature and may have been due to diurnal changes in barometric pressure. To establish adsorption equilibrium between the bulk gas stream and adsorbed species on the Ti0₂ pellet catalyst before photo-oxidation, the reactant gas stream was passed through the system at 1100 ml/min without irradiation until the amount of organic contaminants at the outlet was almost equal to the amount at the inlet, as monitored by a photoionization detector (H-Nu model PI- 101). After equilibration, the lamps were ignited and the flow rate was reduced to 100 ml/min. Fig. 17 reveals the dramatic decrease in VOC concentrations measured at the outlet for the VOC present in the input gas stream, except for 1,1,1-TCA which was not degraded in this process. In addition, small amounts of carbon tetrachloride and even smaller amounts of chloroform were also detected after catalytic photo- oxidation of the VOC, as indicated in Fig. 18. Because these two compounds were at their highest concentration just after photo-catalysis began, it is possible that they were generated in side reactions driven by the presence of water desorbing from the catalyst. The catalyst is hydrophilic and has great surface area onto which water could adsorb. After the reaction proceeded for about 20 hours at nominal temperature, the water concentrations at the inlet and outlet were both at a mole fraction of 0.007 as desired, and the level of CC1* had dropped to 40-100 ppm from a high of 277 ppm.

The products of the reaction included C0₂ and HC1, which corroded the stainless steel fittings on the outlet gas sampling bags. The space times for TCE was calculated to be in the range of 1.3 x IO⁹ - 9 x 10' g s/mol, which are higher even than those values successfully used at elevated reaction temperatures in Example 5 to fully mineralize TCE.

Samples taken from the inlet and outlet sampling ports of the photoreactor apparatus of Fig. 7 were tested for the presence of phosgene gas which is produced during incomplete mineralization of TCE. Initially, samples were passed through a methanol trap to produce methyl- chloroformate and were then injected into the GC/MS equipped with a 30 m by 0.53 mm by 3 μm VOCOL column. No phosgene gas (or its detectable derivative methylchloroformate) was detected from the catalytic photo-oxidation reaction.

Example 7 VOC-containing air samples were drawn from the same soil vapor extraction well described in Example 6. The gases were diluted with ambient air, but not otherwise treated, before entering the reactor so that the concentration of PCE was about 500 to 1000 ppmv (parts per million by volume). Two photoreactors were used; both were of the larger embodiment described above. The first, termed Reactor A, contained 494 g of catalyst. Reactor B contained 379 g of catalyst. The catalyst was prepared as described in Example 1.

At flow rates of 2 liters per minute (1.6 to 3.1 x 10* gs/mol PCE), greater than 99% of the VOC molecules were converted; between 10 and 30 ppmv of hexachloroethane were detected in the outlet gas stream. No products other than HCL CL₂ and C0₂ were found.

By increasing the space time to 12.4 to 25 x 10^s, hexachloroethane was decreased to below 1 ppmv. All samples were analyzed within 2 hours of collection by gas chromatography on a 60 m VOCOL column with an EC detector. With smaller space times, the process produced other unwanted bi-products such as carbon tetrachloride, phosgene and greater amounts of hexachloroethane.

The composition of the inlet feed stream is depicted in Fig. 19. At about 4 days, the inlet stream was diluted with outside air, thereby reducing the concentration of VOCs to below 1000 ppm. Generally, the concentration of PCE was in the range of 500-1000 ppm in the diluted feed stream. Beginning at about day 3.5 of the test, the flow rate through both reactors A and B was 2 liters per minute. At about day 6.5, the flow rate of the feed stream through reactor B was reduced to 1 liter per minute. On day 8, the flow through reactor B was again reduced to 0.5 liters per minute. On day 9, flow through reactor B was again increased to 2 liters per minute. The flow through reactor A was maintained at 2 liters per minute from day 3.5 through day 9.5, at which time it was increased to 4 liters per minute.

The temperature in reactor A was about 75°C between days 1 and 3. On day 3, the temperature was raised to 100°C for the duration of the test. On about day 3.5, the temperature of reactor B was set at 80*C. This temperature was maintained until about day 5.5 at which time it was increased to 110°C for 1 day, at which time the temperature was reduced to 100°C for the duration of the test.

The outlet stream from reactor A was analyzed beginning on about day 4.5. During this entire test, the temperature of reactor A was 100°C. During the time in which flow was at 2 liters per minute, PCS was virtually undetectable in the outlet stream. At most, about 10 ppm of PCE were detected. Otherwise, PCS concentration was less than 5 ppm. Carbon tetrachloride, chloroform, and phosgene were not detected under these conditions. TCE was observed twice at less than 0.5 ppm. Hexachloroethane was observed at levels at or below 30 ppm. However, it is believed by the inventors that the hexachloro structure of this molecule is overestimated by the EC detector used. Therefore, it is believed that the amount of hexachloroethane was more likely in the range of about 10 ppm.

The results obtained with the reactor A demonstrate that full activity remained, and greater than 99% conversion of PCS was observed for a period of 6 days. This suggests that the catalyst used is quite stable and active under high temperature conditions without loss in activity over time. When the flow rate through reactor A was increased to 4 liters per minute, breakthrough of undegraded PCE was observed. This demonstrates that at this flow rate, there was insufficient space time to per it full degradation.

Reactor B was monitored at 3 temperatures and at several flow rates. At 80°C chloroform and TCE were not detected and PCE was generally present at about 2 ppm, phosgene was also detected at between 1000 and 6000 area counts (EC detector). On about day 5.35, no PCE or carbon tetrachloride were observed, although phosgene was present at about 1900 area counts.

Between days 5.6 and 6.5, the temperature of reactor B was adjusted to 110°C. Flow through the column was at 2 liters per minute. Under these conditions, break through of PCE was observed, although levels of hexachloroethane were below about 10 ppm. Also, phosgene was observed between about 5000 and 10000 EC area counts. Chloroform at 1.4 ppm was also observed once. These conditions are not preferred, because of the accumulation of bi-products.

The temperature of reactor B was adjusted to 100°C and the flow rate varied between 0.5 and 2 liters per minute. Between day 6.9 and 7.8 flow through reactor B was at 1 liter per minute. No chloroform, TCE or carbon tetrachloride were detected. TCE and hexachloroethane were observed at levels below 2 ppm. Phosgene was observed twice during this period at levels below 1000 EC area counts. Between days 8.5 and 9, at a flow rate of 0.5 liters per minute, the level of PCE in the outlet stream was between 0 and about 2 ppm. Hexachloroethane in the outlet stream was at between 1 and 3 ppm. No chloroform, carbon tetrachloride, or TCE were observed. Phosgene was observed once at a level below 1000 EC area counts. Between days 9.4 and 9.8, the flow was adjusted to 1 liter per minute of inlet feed stream plus 1 liter per minute of pure oxygen. Under these conditions, less than 1 ppm of PCE and hexachloroethane were observed. No carbon tetrachloride, chloroform, TCE or phosgene was observed.

These results demonstrate that TCE can be fully mineralized without producing a detectable amount of phosgene gas by using a process having all of the improvements of the present invention (use of metal oxide ceramic bodies made of anatase and having high porosity, and reaction conditions of low humidity, increased space times and increased reaction temperature). These results further demonstrate that under certain conditions, full mineralization is possible with a process having fewer than all five improvements in combination. It is also believed that it is possible to combine the improvements of the present invention in various ways to identify conditions under which 1,1,1-TCA, CC1₄, and other VOC may be degraded.

Claims

1. A method of degrading organic chemicals in a gaseous stream, the method comprising the steps of lowering the moisture content of the gaseous stream and heating the gaseous stream to at least about 60°C; passing the gaseous stream through a reactor such that the gaseous stream is exposed to a catalyst of a transition metal oxide material in a physical form having a high surface area, the reactor being heated to at least about 60°C; while the gaseous stream is exposed to the catalyst, irradiating the catalyst and the gaseous stream from a source of ultraviolet light; and recovering the reaction products from the reactor.

2. A method as claimed in claim 1 wherein the moisture content of the gaseous stream is lowered to less than 0.010 mole fraction.

3. A method of degrading organic chemicals in a gaseous stream, the method comprising the steps of lowering the moisture content of the gaseous stream; passing the gaseous stream through a reactor such that the gaseous stream is exposed to a catalyst of a transition metal oxide material in a physical form having a high surface area; while the gaseous stream is exposed to the catalyst, irradiating the catalyst and the gaseous stream from a source of ultraviolet light; and recovering the reaction products from the reactor only after the gaseous stream has dwelt in the reactor for a space time in excess of 10^s gram-seconds per mole.

4. A method as claimed in claim 3 wherein the space time is in excess of 6 x 10⁷ gram-seconds per mole.

5. A method as claimed in claim 3 wherein the catalyst is a porous transparent titanium oxide ceramic material.

6. A method as claimed in claim 3 wherein the moisture content of the gaseous stream is lowered to less than 0.010 mole fraction

7. A method as claimed in claim 3 wherein the space time is in excess of 1 x IO⁹ gram-seconds per mole.

8. A method of degrading organic chemicals in a gaseous stream, the method comprising the steps of heating the gaseous stream to a temperature of at least about 60°C; passing the gaseous stream through a reactor such that the gaseous stream is exposed to a catalyst of a transition metal oxide material in a physical form having a high surface area, the reactor being heated to at least about 60°C; while the gaseous stream is exposed to the catalyst, irradiating the catalyst and the gaseous stream from a source of ultraviolet light; and recovering the reaction products from the reactor only after the gaseous stream has dwelt in the reactor for a space time in excess of IO³ gram-seconds per mole.

9. A method as claimed in claim 8 wherein the space time is in excess of 6 x 10⁷ gram-seconds per mole.

10. A method as claimed in claim 8 wherein the catalyst is a transparent form of a titanium oxide ceramic material.

11. A method as claimed in claim 8 wherein the space time is in excess of 1 x IO⁹ gram-seconds per mole.

12. A method as claimed in claim 8 wherein the reactor is heated to about 100°C.

13. A method of degrading organic chemicals in a gaseous stream, the method comprising the steps of lowering the moisture content of the gaseous stream and heating the gaseous stream to at least about 60°C; passing the gaseous stream through a reactor such that the gaseous stream is exposed to a catalyst of a titanium dioxide material which is substantially all in the anatase form and which is in the form of a porous material having a porosity of over 50% and a mean pores size of less than 50 Angstroms, the reactor being heated to at least about 60°C; while the gaseous stream is exposed to the catalyst, irradiating the catalyst and the gaseous stream from a source of ultraviolet light; and recovering the reaction products from the reactor only after the gaseous stream has dwelt in the reactor for a space time in excess of 6 x 10⁷ grams-seconds per mole.

14. A method as claimed in claim 13 wherein the moisture content of the gaseous stream is lowered to less than 0.010 mole fraction.

15. A method as claimed in claim 13 wherein the space time is in excess of 1 x IO⁹ gram-seconds per mole.

16. A method as claimed in claim 13 wherein the porous titanium material has a bimodal distribution of pore sizes, with one peak of pore sizes being in the range of 20 to 50 Angstroms and the other peak of pore sizes being in the range of 5 to 20 Angstroms.

17. A method as claimed in claim 13 wherein the reactor is heated to about 100°C.