Abstract
Abstract
An experimental CMR study is presented on the catalytic oxidation of ethane. Two different alumina membranes were investigated, the first with V loading close to monolayer coverage, the second with a significant higher V dopant degree, meaning with bulk-like V2O5 as the active phase. Catalytic testing is performed under a wide variety of parameters to identify suitable regions of operation conditions. Results are compared with previous PBMR and FBR experiments in order to draw conclusions regarding the potential of the different membrane reactor configurations. After testing the membranes have been be investigated by spatially resolved SEM and XPS to evaluate the reaction induced modifications in the chemical composition and structure of the membranes.
Introduction
Membrane reactors are attractive in the selective oxidation of light hydrocarbons. This is due to changed reactant concentration and contact time profiles compared to the conventional fixed bed reactor (FBR) [1,2]. Among the different membrane reactor configurations the catalytic membrane reactors (CMR) with impregnated ceramic porous membranes have several advantages. First, they can be operated with very low pressure drops because of the absence of a packed bed causing a high resistance. Second, there are significant lower permeability limitations for the oxidant e.g. compared to electrochemical or other membrane reactors with gas dense ion conducting membranes. Third, in comparison to the latter ones, operation temperature is 200-300 K lower, what allows to save energy costs. Despite of an intensive research in this field, catalytic membrane reactors are still far from industrial application. One of the open questions is the possibility of reversible and irreversible changes of the catalytic layer yielding modifications in the catalytic activity during the operation under local oxygen deficit conditions.
Experimental
Two catalytic membrane series were investigated within this study. Commercially available tubular alumina ceramic composite membrane tubes (length 250 mm, outer diameter 10 mm, inner diameter 7 mm) were used as support. The separation layer (γ-alumina, 2 µm thickness) had a pore diameter of 10 nm. The membranes were impregnated by dipping into a hot saturated solution of vanadyl acetylacetonate in acetone. Membranes were fully impregnated to suppress possible deep oxidation reactions caused by blank alumina [3]. In case of the first membrane series the dipping was performed one time for 4 hours. In case of the second series the membranes were dipped for 2 hours in the impregnation solution, then the membranes were dried at 100 °C for 30 minutes followed by a second impregnation step. After impregnation all membrane tubes were washed in cold acetone to prevent inhomogenities of the impregnated layer. Than they were dried overnight at 130 °C and calcinated at 700 °C for 4 hours. Finally, both ends of the calcinated membranes were vitrified by dipping into molten glassware in that way, that a permeable zone of 60 mm length results in the middle of the membranes. For characterisation purposes one membrane of each series was cut up. The achieved V loadings were found to be 0.003 and 0.15 % V, respectively. BET measurements give values of 0.21 m²/g for both membrane series. This corresponds to a value of 42 m²/g for the separation layer. The V density (atoms/nm²) is 1.7 for series 1 and 84 for series 2, meaning the catalytic active phase is present as well dispersed vanadate in case of series 1 and as bulk-like V2O5 in series 2.
Catalytic testing of membranes in the selective oxidation of ethane with varying temperatures (450 - 650 °C), feed compositions (0.7 % ethane, 0 - 2 % O2, balance N2), W/F ratios (0.103 and 0.062 g cat . h/l) and shell to tube side ratios (0.11 – 9) to identify best operation regions is presently underway. Additionally, the performance of one unimpregnated membrane was evaluated to check for wall reactions under real operation conditions. The results will be compared to previous PBMR and FBR experiments [2-4] in order to draw conclusions regarding the potential of the different membrane reactor configurations.
After the catalytic testing the used membranes were cut up to check for changes due to their use inside the reactor. Cut membranes have been analyzed using SEM and XPS. The first technique provides information on structural changes and a possible V migration within the membrane. The latter one provides the details of the chemical composition and, particularly, of the effective oxidation state of vanadium (Vef.ox) in the surface layer [5]. Our spatially resolved XPS measurements provide the axial and circular mapping of the Vef.ox along the curved membrane surface thus allowing the evaluation of Vef.ox profiles during CMR operation. Also these results are compared to the classical FBR operation.
References
1. A. L. Y. Tonkovich, J. L. Zilka, D. M. Jimenez, G. L. Roberts, J. L. Cox,
Chem. Eng. Sci, 51(5) (1996) 789
2. F. Klose, T. Wolff, S. Thomas, A. Seidel-Morgenstern, Catal. Today 82 (2003) 25
3. F. Klose, M. Joshi, C. Hamel, A. Seidel-Morgenstern, Appl. Catal. A: General 260 (2004) 101
4. F. Klose, T. Wolff, S. Thomas, A. Seidel-Morgenstern, Appl. Catal. A: General 257 (2004) 193
5. Y. Suchorski, L. Rihko-Struckmann, F. Klose, Y. Ye, M. Alandjiyska, K. Sundmacher, H. Weiss, Appl. Surf. Sci., 2005, in press.