Kim Martin Nikolajsen
One possibility for the removal of volatile organic compounds (VOCs) is catalytic oxidation. For the removal of VOCs in low concentration (≲ 3000 ppmv for propane) it is not possible to sustain the required oxidation temperature over the catalyst without external heating. An approach to overcome this problem is to concentrate the VOC by adsorption, in a "two-step adsorption-incineration" process. This is an unsteady state process in which the VOC laden effluent gas is cleaned by passing it through an adsorber. Once the adsorber is saturated the VOC is desorbed by heating and purging. The concentrated VOC from the desorber can then be oxidized over the catalyst, producing enough heat to sustain the oxidation temperature. Conventional adsorbents have some drawbacks such as tailing during desorption due to the large particle size, high pressure drop and hot-spot formation. The objective of this work was to develop, characterize and test the adsorbent and catalyst, with low resistance to mass transfer resulting from thin films of active material, for the two-step adsorption-incineration process. Both the adsorbent and the oxidation catalysts were supported on sintered metal fibers (SMF). This is a novel, structured support offering several advantages over conventional randomly packed beds: Low pressure drop, due to the high porosity of the SMF, which is important where large flow rates and/or high quantities of gas must be treated. High thermal conductivity, due to the metallic nature, leading to smaller temperature gradients in the fixed-bed. A very high geometric surface area, due to the small fiber size. On the order of 200,000 m2/m3 with a fiber diameter of 20 µm. Furthermore, a mathematical model was developed to show the feasibility of the coupling of the two steps in the process. The adsorber was made from a thin, homogeneous film of MFI-type zeolite covering the fibers of the SMF. The film was grown by the seed-film method by hydrothermal treatment. A 3 µm film was obtained with 10 wt.% zeolite on the fiber after 24 h synthesis at 125 °C. This material had a specific surface area of 30 m2/g. The adsorption equilibrium of propane was described well by the Langmuir isotherm and the heat of adsorption was found to be ΔH0ads = - 43.1 kJ/mol. Isothermal breakthrough curves were measured as a function of temperature and film thickness. A mathematical model, comprised of a tanks-in-series model with a linear driving force (LDF) mass transfer description, successfully described the breakthrough behavior. The overall mass transfer was found to be solely due to diffusion in the zeolite film. However, the thin films show very low resistance to mass transfer leading to low internal concentration gradients and efficient utilization of the adsorbent. The pressure drop was measured and compared to that of a randomly packed fixed-bed of spheres. The equivalent diameter for a constant volumetric flow rate and superficial cross sectional area was dp = 180 µm, for a fiber diameter of df = 26 µm at 10 wt.% zeolite loading. The oxidation catalysts were all supported on SMF and cobalt oxide was used as the active phase. Three different groups of catalysts were developed. The support was modified by coating the fibers either with a thin film of fine powder (alumina or catalyst powder) or with a MFI-type zeolite film identical to the adsorbent. The third group was made from cobalt oxide impregnated directly on the oxidized fibers. The latter type of catalyst was found to be the most active during the screening experiments. The base support composition was varied (stainless steel (SS), FeCrAlloy and inconel) to investigate the support effect on the kinetics. The stainless steel and FeCrAlloy supports had similar activities, surpassing that of the inconel supported catalyst, which might be attributed to a Co3O4 surface enriched with iron from the support. Between the 1.1 wt.%Co3O4/SMFSS and 1.5 wt.%Co3O4/SMFFeCrAlloy catalyst the FeCrAlloy based had the highest normalized activity at low temperature (< 310 °C), low propane mole fraction (0.13 %) and high oxygen mole fraction (16.7 %). The chemical kinetics were determined from a novel and efficient experimental design. The apparent activation energy for this catalyst was found to be 87.5 ± 2.6 kJ/mol, with reaction orders in propane of 0.38 ± 0.04 and in oxygen of 0.30 ± 0.08. The catalysts were tested up to 350 °C. Mass transfer limitations were absent and the kinetics was described well with the power rate law as compared to the Mars-van Krevelen model. Catalysts based on Co3O4, supported directly on oxidized SMF, are very active catalysts, simple to prepare, withstand deactivation due to hot-spot formation and show great potential in comparison to the industrial reference catalyst (copper and manganese oxide supported on alumina). The adiabatic desorption process was investigated by mathematical modeling. Simply purging the adsorption bed with the hot exhaust gas from the reactor cannot result in the concentration of the VOC due to the high heat capacity of the fixed-bed. Hence, it will be necessary to use either a high heat carrier like steam, with the adverse effects of drying and separation of water and VOC, or an approach in which the desorber is heated prior to purging the bed, changing the adsorption equilibrium in favor of high gas concentrations. For the latter method, the minimum theoretical propane mole fraction for an autothermal process was found to be 350 ppmv to sustain a catalytic incinerator at 250 °C. Finally, it can be concluded that for the first time zeolite films on SMF were used as structured adsorbents, leading to very low internal mass transfer and increased process efficiency. Furthermore, the advantages of the SMF can be exploited in a catalytic oxidizer, which can be combined with the adsorber to annihilate VOCs in low concentrations. A novel method for the deposition of fine powder catalyst on SMF was also developed for the first time.
EPFL2007
