Biomass steam-reforming in a low throughput fluidized bed reactor

This work presents the results from biomass gasification tests in a low throughput fluidized bed membrane reactor (FBMR) and compares them with two-phase model predictions. The operating temperature was decreased to 550C, while an equimolar steam/cellulose ratio was used. The profiles of kinetics limitation and biomass conversion were analysed when a palladium (Pd) membrane for hydrogen separation was used. The produced gases were analysed by gas chromatography, while steam composition and temperature profiles were retrieved by near-infrared imaging (NIRI) technique. As a single membrane tube did not show high permeability rates, the changes in conversions and temperatures profiles were not significant when a Pd membrane was incorporated. The experimental design and procedure allowed us to visualize low radial dispersions in the FBMR. 1Introduction The miniaturization of chemical processes is one of the most promising technologies with remarkable growth in the chemical industry. In contrast to the conventional objective in chemical industry, feasibility and applicability of such small equipments have been the prior objective than their productivity and efficiency. The large surface to volume ratio permits multiple design arrangements and analysis under safety conditions. Different types of reactors have been investigated during the last decade with a special concern on those showing several functions. Undoubtedly, hybrid reactors of reaction and separation will occupy the largest domain in chemical reactor design [1, 2]. In this work, we intend to experiment a bubbling fluidized bed reactor (BFBR) with a low throughput and combined with a selective membrane. A dual purpose is aimed beyond this investigation (1) in situ visualization of temperature and vapour distribution during the steam reforming of biomass; (2) comparison between experimental data, the model results and visualization maps in a simplified small size fluidized bed reactor. To our knowledge, this objective has not been experienced yet. A higher homogenisation, low radial dispersions and less carbon deposition are expected. Assadulah et al. [3, 4, and 5] claimed that Rh/CeO2/SiO2 was showing a high selectivity with a low carbon deposition for cellulose gasification than the conventional catalysts (dolomite and Nickel supported). In this work, less affordable catalyst is prepared by adding Nickel to Rhodium. The changes of the catalyst activity are investigated by gas chromatography and NIRI technique. We are not aware if this technique has been used previously. However, the spatially resolved infrared imaging (IRI) technique was used by Luss et al. [6] for self-oscillations on supported catalysts and Wolf et al. [7] for monitoring the linear distribution of adsorbed CO coverage on Rhodium supported catalysts. Infrared methods by thermography are still suffering from background interference, accurate emissivity values and detector sensitivities. In this work, near infrared camera is used to map OH band distribution in gas phase and adsorbed on catalyst surface as well. Applying this method may partially circumvent some limiting constraints in the IRI technique : (1) In low throughput BFBR, a negligible NIR beam scattering and high transmittance through the bed are expected due to low thickness of the optical path; (2) A cheaper window in quarts is used while the IRI requires sapphire or ZnSe materials; (3) Compared with OH band intensities, CO and CO2 bands show negligible intensities according to HITRAN data bank [8]; (4) For similar detector sensitivity, Accuracy of the temperature obtained by NIR spectroscopy is higher than that from IR spectroscopy [9]. On the other hand, the typical spectrum of NIR spectroscopy of vapour deposition on Rh/Ni/CeO2/SiO2 catalyst during hydrogen reduction is shown in Figure 1 [10]. The increase of intensity of peaks with time shows OH interaction with catalyst surface, especially ceria due to its weaker reducibility as shown in equation 2. The 1380 nm band is attributed to OH vapour phase while several peaks are overlapped and which are attributed to strong second overtone bands of adsorbed OH with ceria and nickel active sites. Partial shift bands are observed at 1220 due to their interactions with OH of silica matrix. Then, vapor phase map in the BFBR can be obtained by using a NIR camera with a band filter near 1380 nm, while the remaining NIR intensity corresponds to adsorbed OH species. In low throughput reactor, the steam reforming of cellulose will be investigated as follows: (1) In situ OH band visualization; (2) Bimetallic RhNi/CeO2/SiO2 activity; (3) An example of comparison of experimental data of conversion, product composition and temperature in bubble and emulsion phases, with and without membrane, with those obtained by the two-phase flow model. 110