5 Heat effects in a membrane reactor for the water gas shift reaction

M. E. ADROVER; E. LÓPEZ; D. BORIO; M. N. PEDERNERA

STUDIES IN SURFACE SCIENCE AND CATALYSIS

Elsevier

Año: 2007 p. 183 - 188

0167-2991

Most of the hydrogen is produced industrially by steam reforming of hydrocarbons (mainly natural gas) or alcohols (e.g., for fuel cell applications). The process gas stream coming from the steam reformer is composed by H2, CO, CO2, H2O and small amounts of unconverted reactants (CH4). The CO concentration of the gas leaving the reformer must be reduced up to a specified level, with two main goals: 1) increase the H2 production rate and 2) purify the process stream. To these ends, the Water Gas Shift Reaction (WGSR) is widely used: CO+ H2O=CO2+ H2               (1) Reaction (1) is moderately exothermic and strongly controlled by the chemical equilibrium, which is favored at low temperatures. For this reason, in large-scale processes (e.g., ammonia plants) the WGSR is carried out at two different temperature levels, called high temperature (HTS) and low temperature (LTS) shift reactors, using different catalysts. In smaller scale processes, such as the fuel processing for fuel cells (e.g., PEM cells) normally the WGSR is carried out at an intermediate temperature level, requiring large reactor volumes to reach the necessarily high CO conversions (Giunta et al, 2005). An attractive alternative to increase the CO conversion is the membrane reactor (MR). The main idea of this design is the selective permeation of reaction products (e.g., H2) to shift the equilibrium towards products and consequently decrease the CO outlet concentration, or reduce the amount of catalyst for a desired conversion level. The MRs constitute a kind of multifunctional reactor, which combines two unit operations in a single step: chemical reaction and products separation. For this reason, MRs have deserved considerable attention in the scientific literature (Coronas and Santamaría, 1999).   The H2 removal can be carried out by means of selective dense membranes of Pd or its alloys. Although dense membranes offer high selectivity to specific gases, their low permeability and high cost of Pd have limited their application to small-scale processes. In order to increase the permeation, composed membranes became an alternative. In these membranes a substrate of high porosity and low resistance to flow is covered by a metallic layer, which provides the selectivity (Oklany et al., 1998; Hou et al., 1999; Kikuchi et al., 2000; Tosti et al., 2000; Ling and Chang, 2005). The advantages of the MR to perform the WGSR have been demonstrated by Criscuoli et al., (2000). However, the heat effects are neglected and the reactor operation is supposed to be isothermal. This is a common assumption in the modeling of most of the MRs, which agrees with the temperature measurements inside the reactor at laboratory scale. The isothermal temperature profiles are due to the high ratio between the heat transfer area and the reactor volume, i.e., small reactor diameters are normally selected. However, this condition may not be true when higher process scales are necessary; for example: several membrane tubes installed in parallel within a shell with flowing the sweep gas. At this process scale, the usual assumption of isothermal MR should be revised and the heat effects taken into account. This is the purpose of the present work, where the performance of the MR is simulated and compared with that of the conventional fixed bed reactor (CR). To compare both types of reactors, two extreme operating conditions are selected: isothermal and adiabatic operations. The simulation study is carried out by means of a pseudo-homogeneous mathematical model, subject to the following hypotheses: a) axial and radial mass and heat transfer dispersions are neglected; b) Isobaric and isothermal conditions are assumed for the permeation side; c) infinite selectivity for the membrane (H2 is the permeation gas) and d) the permeation flow through the membrane follows Sievert´s law. The kinetic model proposed for the WGSR by Criscuoli et al. (2000) is included. To simulate both the MR and CR, the design parameters and operating conditions given in Table 1 are selected. The CR is modeled by assuming zero for the H2 flow through the membrane. Table 1. Geometric parameters and operating conditions used in the simulations of CR and MR (Criscuoli et al., 2000). Reactor length,L 0.21 m (CR)- 0.15 m (MR) Tube internal diameter, dti 6.7 mm (CR)- 8 mm (MR) Tube external diameter, dte 13.4 mm (MR) Catalyst mass, W 9.4 g  Thickness of Pd film, δ 75 µm Pre-exponential factor of the Sievert permeability coefficient,Q0 2.95 10-4 mol s-1m-1Pa-1/2 Activation energy of the hydrogen permeability, Ep 48499.72 J/mol Feed pressure, Po 1 atm Sweep gas volumetric flow rate, Fsweep 436 ml/min Space time , t 7.5 103 gcat min/mol Feed temperature, To 322 ºC Sweep gas temperature, Tsweep 322 ºC Sweep gas pressure, Psweep 1 atm Inlet CO, dry basis % 12.27 Inlet CO2, dry basis % 11.49 Inlet H2, dry basis % 75 Inlet CH4, dry basis % 1.24 Conclusions    Heat effects in membrane reactors should be considered. Although the WGSR is only moderately exothermic, the temperature changes inside the reactor are not negligible. The only reason to neglect the heat effects in MRs is the small scale of laboratory designs. When intermediate (or large) scales are under consideration, the temperature variations will affect the kinetics and the chemical equilibrium. References Giunta, P.D., Amadeo, N. E., Laborde, M.A.(2005). XIV Congreso Argentino de Catálisis. Santa Fe, Argentina. Coronas, J., Santamaría, J. (1999). Catalysis Today, 51, 377-389. Criscuoli, A., Basile, A., Drioli, E. (2000). Catalysis Today, 56, 53-64. Kikuchi, E., Nemoto, Y.,  Kajiwara, M., Uemiya, S., Kojima, T. (2000). Catalysis Today, 56, 75. Lin, W-H. Chang, H-F. (2005). Surface & Coating Technology, 194, 157. Oklany, J.S., Hou, K., Hughes, R. (1998). Applied Catalysis A: General, 170 ,13-22. Tosti, S., Bettinali, L., Violante, V. (2000). International Journal of Hydrogen Energy, 25, 319-325.