화학공학소재연구정보센터
Applied Catalysis A: General, Vol.433, 58-68, 2012
Fischer-Tropsch synthesis using H-2/CO/CO2 syngas mixtures: A comparison of paraffin to olefin ratios for iron and cobalt based catalysts
Fischer-Tropsch synthesis (FTS) using H-2/CO/CO2 syngas mixtures over cobalt and iron based catalysts was carried out in a fixed-bed reactor. CO2 rich feeds produce products that are mostly light hydrocarbons with higher molar paraffin to olefin (P/O) ratio, whereas CO rich feeds shift the product composition to an FT type product with higher olefin product selectivity over both iron and cobalt based catalysts. Although the P/O ratio for FTS is strongly dependent on the operating conditions, the experimental evidence shows that the linear relationship between P(n+1)/O(n+1) and P-(n)/O-(n) holds for a large number of experiments. It is also shown to be independent of the type of the reactor, the composition of the syngas, reaction conditions and the kind of catalyst. Two features about the ratio of xi = [P(n+1)/O(n+1)]/[P-(n)/O-(n)] for the FT products have been identified: (1) with n > 2, the experimental values of xi(n>2) are higher than 1, fairly constant and independent of chain length n; (2) with chain length n = 2, the ratio of P-3/O-3 to P-2/O-2 (xi(n-2)) is significantly different, and shows that xi(n-2) << xi(n>2). An equilibrium hypothesis is considered in an attempt to explain this experimental phenomenon. A simple vapor-liquid equilibrium (VLE) model indicates that the ratio of P(n+1)/O(n+1) to P-(n)/O-(n) changes in a range of (1, 1/beta), where beta is the variation of the vapor pressure coefficient, which is related to the incremental energy of vaporization per CH2 unit of the hydrocarbon chain. Our experimental results support the expression when the chain length n > 2. But with chain length n = 2, this expression is unable to explain the relationship between P-3/O-3 and P-2/O-2. Another model, based on quasi reaction equilibrium, is developed to explain the linear relationship between P(n+1)/O(n+1) and P-(n)/O-(n). We assume that the reaction of Cn+1H2n+2 + CnH2n+2 = Cn+1H2n+24 + CnH2n reaches quasi-equilibrium. Because the experimental results are quite close to the equilibrium calculations, we postulate that the product distribution might be determined by considering reaction equilibrium. (C) 2012 Elsevier B.V. All rights reserved.