Abstract
Introduction
Sulfated zirconia (SZ) is a highly active catalyst for n-alkane isomerization. Typically, SZ first passes through a yet unexplained period of increasing activity and then, unless special measures are taken, deactivates rapidly. Investigations correlating the surface state and the catalytic performance are rare [1], and the nature of surface deposits is usually studied post mortem and ex situ. Here, we have monitored the surface of SZ throughout all stages of the reaction profile, i.e. during increasing and decreasing activity. In situ UV-vis spectroscopy was used to follow n-butane or n-pentane isomerization and, after the catalytic experiment, to identify the products formed from reaction of surface deposits with O2 or H2O. Additionally, we have interrupted the alkane isomerization at different stages of catalyst activity and then probed the surface in calorimetry and temporal analysis of products experiments.
Experimental
Two types of sulfated zirconia were used. An already sulfated commercial precursor (MEL Chemicals XZO 682/01) was calcined for 3 h at 823 K, the sulfate content of this sample was about 4.5 wt %. The second sample was produced through precipitation from ZrO(NO3)2, sulfation with ammonium sulfate, and calcination for 3 h at 873 K. This sample contained about 9 wt% sulfate. In situ UV-vis spectroscopy was performed using a PerkinElmer Lambda 9 spectrometer equipped with an integrating sphere and a reactor with a quartz window which holds about 1.2 g of SZ [1,2]. Reaction conditions were either 5 kPa n-butane and 378 K or 1 kPa n-pentane and 298 K. After 14 h, samples were cooled to 298 K and exposed to oxygen, then to water vapor (saturated at 298 K), and the spectra were recorded parallel to gas phase analysis by on-line MS. Temporal analysis of products (TAP) was conducted in a TAP-2 system with a Hiden QMS [3]. In order to adjust different surface states, the TAP reactor was used as a fixed bed flow reactor (0.1 - 0.4 g of sample, 5 kPa alkane in He, 423 K) with on-line GC. The flow was interrupted after distinct time intervals, the sample was evacuated for 1 min, and n-butane, isobutane, CO2, or H2 were pulsed (1013-1017 molecules). SETARAM Calvet calorimeters in combination with a dosing system were used to obtain adsorption isotherms and differential heats of adsorption, qdiff [4,5]. The calorimeter cell was used as a flow reactor, and the isomerization reaction was performed with 500 mg of sample at 1 kPa n-butane at a total flow of 50 ml/min and 378 K. Reaction products were monitored by on-line GC. The reaction was stopped at different stages of activity, loosely adsorbed species were removed by evacuation, and n- or isobutane was adsorbed at 313 K.
Results and Discussion
In situ UV-vis experiments did not show any changes to the spectra during the periods of increasing activity, neither in n-butane or n-pentane isomerization, although in some cases not all carbon was recovered in the effluent stream, i.e. some hydrocarbon was withheld by the catalyst. Surface species were either too few to be detected or did not absorb in the UV-vis range. Adsorption isotherms of isobutane recorded of catalysts in their phase of rising activity showed fewer sites than the fresh catalysts. Mass spectra taken during evacuation for the TAP experiment revealed species larger than C4 after a few minutes on stream. A series of TAP experiments taken between the minute long intervals of reaction in flow revealed a steady decrease of the number of sites available for the interaction with n-butane. The experiments demonstrate that the number of sites for adsorption of reactant or product is decreasing, already at early times on stream. The nature of the species blocking the sites is so far unknown - it could also be water, as a contaminant or side product - but the species are not removable by evacuation and are thus strongly attached.
In situ UV-vis spectra recorded during n-butane or n-pentane isomerization invariably showed the formation of bands at 310 nm (butane) or 330 nm (pentane). Such bands have been observed before in post mortem investigations of SZ catalysts and have been attributed to allylic species [6]. It was possible to now correlate the formation of these bands with the catalytic performance. Surface deposits became first detectable at the point of highest conversion and lowest selectivity, where Cn-1 and Cn+1 alkanes are formed as side products. The formation of gas phase products and surface species appeared to be competitive, as the conversion declined rapidly, the bands indicating deposits grew in an opposite way. The allylic surface species are thus definitely not intermediates of the reaction, and maybe not even the cause but the consequence of catalyst deactivation, i.e. the result of a change in selectivity of the catalyst.
After a 14 h run, the deactivated catalysts - still in the in situ UV-vis cell - were exposed to oxygen and then to water vapor, to use derivatization for better identification of the surface species and to elucidate potential routes of regeneration. Exposure to oxygen resulted only in minute changes to the UV spectra. Reaction with water vapor reduced the intensity of the band at 310 (330) nm, produced new bands at 380-390, 455-460 and 550-560 nm. Bands at 430-455 nm have been assigned to quinone-type compounds previously [6], the other bands are not yet identified. The band at 380-390 nm lost intensity with time in the moist stream while the other two bands gained in intensity. Fewer gas phase species were generated than during oxygen treatment. Nevertheless, the surface deposits clearly react with the components of ambient air under formation of volatile compounds. Indeed, a freshly deactivated catalyst turned brownish within minutes after being taken out of the reactor but the color faded within an hour. Surface deposits on sulfated zirconia catalysts can thus not be studied ex situ, as they may be altered and volatilized by contact with air.
The investigations demonstrate that so far unknown but stable surface species are formed during the initial stages of reaction, i.e. the period of increasing activity. These species block sites for reactant and product alkanes. Unsaturated species are formed later on stream at and after the maximum in conversion has been reached, their formation is competitive to the gas phase products and related to the bimolecular reaction channel that yields side products. None of the so far identified surface species can thus be seen as a carrier of activity.
Financial support through DFG SPP 1091 is kindly acknowledged.
References
1. R. Ahmad, J. Melsheimer, F.C. Jentoft, R. Schlögl, J. Catal. 218 (2003) 365.
2. M. Thiede, J. Melsheimer, Rev. Sci. Inst. 73 (2002) 394.
3. J.T. Gleaves, G.S. Yablonski, Y. Schuurman, Appl. Cat. A-Gen., 160, (1997) 55.
4. E.N. Coker, H.G. Karge, Rev. Sci. Inst., 68 (1997) 4521.
5. L.C. Josefowicz, H.G. Karge, E.N. Coker, J. Phys. Chem., 98, (1994) 8053.
6. D. Spielbauer, G.A.H. Mekhemer, E. Bosch, H. Knözinger, Catal. Lett. 36 (1996) 59.
