Structure of effective catalyst layers around bubbles in a fluidized catalyst bed

https://doi.org/10.1016/j.cej.2006.06.017Get rights and content

Abstract

In a fluidized catalyst bed, the reactant gas transfers from the bubble phase to the emulsion phase and reactions proceed in the emulsion phase. The catalyst particles around the bubbles should contact the gases containing a high concentration of the reactants. Therefore, the effect of the catalysts around the bubbles is very important for estimating the conversion and selectivity in the reactor. In order to study the role of these catalysts, the hydrogenation of carbon dioxide was carried out in a fluidized catalyst bed. Based on the results, the amount of the catalyst that was effective for the reaction was calculated. In addition, the shape of the bubbles ascending in the fluidized catalyst bed was observed using a fast X-ray computer tomography (CT) scanner. The structure of the bubbles in the fluidized catalyst bed was very complicated and the surface area of the bubbles was much greater than the obtained when assuming spherical shaped bubble. By assuming that effective catalysts existed around the bubbles, the thickness of catalyst layer was obtained. Finally, the 3-dimensional images of the catalyst layers around the bubbles were reconstructed.

Introduction

Fluidized catalyst beds have been used in chemical and refinery processes and extensively studied for more than 50 years. However, the features of this reactor are still not perfectly understood [1]. In a fluidized catalyst bed, since the reaction mainly occurs in the emulsion phase, the mass transfer rate between the bubbles and emulsion phase affects the overall reaction rate. Therefore, the bubble size, bubble shape, interface area and bubble ascending velocity are the important parameters for the design of a fluidized bed reactor.

Many reactor models have been proposed for fluidized catalyst beds. Almost all of these models considered the mass transfer between the bubble and emulsion phases. When the reaction rate becomes high and mass transfer is the rate-determining step, conversion becomes independent of the reaction rate. However, some experimental results showed that the conversion increased with the reaction rate even when the reaction rate was very high. These results can be attributed to the region just above a distributor [2], [3], the catalysts directly contacting with the bubble phase gas [3], [4], [5], [6], [7], and the dilute phase above the dense phase [3], [4], [5], [6], [7], [8], [9].

In the previously proposed reactor model, some mixing states in the emulsion phase have been assumed. However, the radial concentration distribution is not considered in these models. Based on the results of the direct observation of the bubbles and particle movements [10], there are many particles that seem to not directly contact the reactant gases. The lateral holdup of gas bubbles in a fluidized catalyst bed was almost parabolic [3]. The bubble frequency in the core part is high, but it is lower near the wall. Since catalysts are porous materials, the mixing of the catalyst particles contribute to the transport of the gaseous components in the emulsion phase. The mixing of particles was enhanced by the passage of bubbles and the reactant gases were also well mixed in the core part of the bed. This effect is insignificant near the wall. When the reaction rates are high and the first order reaction rate k = 5 s−1, 50% of the reactants is converted to products within 0.14 s. In this case, the main components of the emulsion phase gas near the wall are the products and the catalysts in this region are not effectively used for the reaction. On the other hand, the catalyst particles around the bubbles come in contact with the gas containing a high concentration of reactants. The reaction will almost completely proceed in this region when the reaction rate is much higher than the mass transfer rate between the two phases.

Therefore, the reactor model should consider the reaction rate, particle movement and bubble behavior. In the present study, the contact efficiency between the reactant gases and catalyst in a fluidized catalyst bed was investigated by carrying out the hydrogenation of carbon dioxide. Since the catalyst particles around the bubbles directly contact the reactants in the bubbles, they should contribute more to the reaction than the catalyst isolated from the bubbles. The amount of the effective catalyst was calculated from the reaction results based on some assumptions. In addition, an X-ray computer tomography (CT) scanner has been used for the analysis in the present study.

Grohse [11] has measured the variation in density of a bed using X-rays for the first time. Rowe and Partridge [12] observed the bubble shape by X-ray photographs and discussed the frequency of bubble splitting and coalescence. The studies on X-ray absorption have been conducted [13], [14] by the group at University College, London since this publication. X-ray computed tomography systems have been applied to determine the local solid distribution in fluidized beds in the 1990s. Kantzas [15], [16], [17] obtained images of the density and holdup using a fourth generation modified medical scanner.

Recently, Kai et al. [18], [19] have reconstructed 3-dimensional image of bubbles using a fast X-ray CT scanner. In the present study, the shape of the bubbles ascending in fluidized catalyst beds was observed using the X-ray CT scanner. Hence, the thickness of the effective catalyst layer could be obtained by using the surface area and the amount to effective catalyst. Finally, the 3-dimensional images of the bubbles and the effective catalyst layer were reconstructed.

Section snippets

Hydrogenation of carbon dioxide

Two types of Ni-La2O3/γ-Al2O3 catalysts were used. The mean particle diameter and the particle density of the catalyst were 54 μm and 660 kg m−3 for CAT-1 and 95 μm and 1090 kg m−3 for CAT-2, respectively. These catalysts were prepared by impregnating porous γ-Al2O3 with an aqueous solution of nickel nitrate. After being dried at 373 K, the impregnated powders were oxidized by air for 2 h at 523 K and reduced by hydrogen for 2 h at 623 K in a fluidized bed. Fig. 1 shows the apparatus used for the

Reaction analysis

The reaction rate equation obtained from the rate analysis in a fixed bed reactor was complicated. Since the molar fraction of hydrogen was excessive compared to the stoichiometric relation, a simplified equation was used by assuming a 1/3-order reaction with regard to CO2 concentration:UGdCAdz=k1/3CA1/3

Fig. 3 shows the relation between the reaction temperature and conversion in the fluidized catalyst bed for CAT-1 and CAT-2. As the reaction temperature increased, the reaction rates became

Conclusions

The hydrogenation of carbon dioxide was carried out in a fluidized catalyst bed. Since the activation energy of the apparent reaction rate constant was almost the same as that obtained in a fixed bed reactor, the overall reaction rate in a fluidized catalyst bed was controlled by the reaction rate. The conversion in a fluidized catalyst bed was successfully estimated by the model assuming plug-flow contact and the effective catalyst in the bed. This catalyst probably contacted the bubble phase

References (23)

  • S. Furusaki et al.

    Mass transfer coefficients and amounts of direct-contact-catalyst in fluidized catalyst beds

    Can. J. Chem. Eng.

    (1984)
  • Cited by (4)

    • Experimental investigation of overlapping bubbles based on laser scanning

      2020, Experimental Thermal and Fluid Science
      Citation Excerpt :

      Generally, limited by the scanning speed, the resolution between the slice layers at the scanning direction is inconsistent with the resolution on the slice plane. In order to match the both, image interpolation method from CT or MRI technology [23,24] is implemented between slices in the paper. Based on the gradual and continuous surface of the bubbles in the bubbly flow, a morphology interpolation method based on image erosion and dilation is developed for slices.

    • Ultrafast X-ray tomographic imaging of multiphase flow in bubble columns - Part 1: Image processing and reconstruction comparison

      2018, International Journal of Multiphase Flow
      Citation Excerpt :

      To extract information such as phase distinguishment, the images are segmented or binarized using a constant threshold. Kai et al. (2007) used a threshold of 135 for their 8-bit or 256 gray scale images of a fluidized bed using FBP. Gray values less than the threshold value are attributed to the dense particle phase and values greater than the threshold value correspond to the bubble phase.

    View full text