Catalytic Extraction Processing (CEP) is an innovative Elemental RecyclingTM technology that converts organic, organometallic, and inorganic waste, byproduct, or process streams into marketable commercial products:  industrial gases, metal alloys, and ceramics (e.g., inorganic oxides, halides, sulfides). Feed materials are injected into a molten metal bath where dissociation of molecular entities to their respective elements and reaction of these dissolved elemental intermediates to form products occur. Process chemistry is driven by reaction thermodynamics, solution equilibria, and metal catalysis, which allow specific partitioning of elements and conversion of feed materials to the desired products as predicted. Experimental results demonstrate CEP's capabilities for waste minimization (e.g., product formation) and environmental performance (e.g., minimum emissions). Examples include the processing of highly toxic toluene diisocyanate production wastes (EPA RCRA listed waste K027), chlorinated organics (EPA RCRA listed waste F024), and mixed metallic, plastic, and inorganic wastes (weapon componentry). Synthesis gas, hydrogen chloride, ceramic, and metal products were manufactured from these waste materials. CEP consistently demonstrates destruction removal efficiencies (DREs) exceeding 99.9999%, NO x and SO x below detection limits (typically 1 ppm to 100 ppm), and dioxins/furans nondetectable to the targeted regulatory limit of 0.1 ng/Nm3 2,3,7,8 TCDD toxicity equivalent (TEQ). Condensed-phase environmental quality was verified by ceramic phases passing toxicity characteristic leaching procedure (TCLP) and the absence of hazardous organic constituents in both metal and ceramic phases. CEP's environmental performance has been validated by the U.S. EPA's best demonstrated available technology (BDAT) equivalency designations, while its manufacturing capabilities have been confirmed by Massachusetts Department of Environmental Protection (MADEP) recycling certifications.

Fluidized-bed bioreactors were developed and operated at three scales (diameters of 0-1, 0.2, and 0.5 m) by the Chemical Technology Division. The performance of these reactors in denitrification was simulated using the following modified form of Monod kinetics to describe the reaction kinetics: NO" rate = WlTTT«0T)(%bl ' omass) In the fluids-movement portion of the simubtion the tanks-in-series approximation to backmixing was used. This approach yielded a V max of 3.5 g/m 3-min [% biomass) and a K s of 163 g/m 3 for the 0.5-m bioreactor. Values of V mx and K s were also determined for data derived from the 0.1-m bioreactor, but inadequate RTD data reduced the confidence level in these results. A complication in denitrification is the multi-step nature of the reduction from nitrate to nitrite to hyponitrite and finally to nitrogen. An experimental study of the effect of biomass loading upon denitrification was begun. It is recommended that the experimental work be continued. I F ; v Contents Page developed a biodenitrification process for wastewater, using pseudaaonas denitrificans. In this process bacteria are immobilized on coal particles, which are fluidized by the nitrate wastewater stream. The bacteria catalyze the denitrification of nitrates in the presence of a carbonaceous energy source to N^ and C0 2. The direct application of this technology is in the nuclear-fuel-processing cycle where nitrates are was*.e products of a num ber of steps. The technology may be applied to a wide variety of other waste streams and to the production of certain chenicils such as methane and ethanol. A simulation of the fluidized-bed bioreactor was required for the de sign of a larger-scale plant. Existing pilot-plant data and a review of the literature indicated that a modified Monod kinetics model of tht form: rate = V max (lTTW)(* biomass) was found to describe the denitrification reaction adequately. In the operating range of pilot-plant data (pH values between 7.0 and 8.6 and C/K > 1.7), pH and ethanol concentration had little effect on the reaction rate. An earlier residence-time-distribution study was used to describe the backmixing in the 0.5-m reactor. Two reactor models were used to simulate the contacting pattern in the bioreactor: the tanks-in-series model and the plug-flow-in-parallel model. The simulation models were derived from material balances incorporating the rate axpression and the biomass loading profile in the reactor. A non-linear regression model, based on Marquart's algorithm, was used to determine optimal values of v max (maximum rate of reaction) and K s (saturation constant) from both models. No significant difference between the plug-flow-in-parallel model and the tanks-in-series model was observed other than the large standard error in K s *n the case of the plug-flow model. The fitted parameters were in sensitive to the shape of the biomass loading profile within the bioreactor; hence a tanks-in-series model with a linear biomass loading profile was chosen to simulate the bioreactor. Values for ...

No abstract