co hydrogenation on cobalt-based fisher-tropsch catalysts: chlorine poisoning reveals the nature of the most active sites
Résumé
1.Introduction
Synthetic fuels and base chemicals can be obtained from both fossil and renewable sources through the conversion of synthesis gas (syngas). Biomass-derived syngas contain significant concentrations of chlorine that may contaminate downstream catalysts. Rytter and co-workers reported that the presence of Cl introduced ex situ during cobalt-based catalyst preparation decreased the metal surface without, surprisingly, affecting activity, thus increasing the turnover frequency [1]. Davis and co-workers investigated the effect of alkali and chloride addition to a Fischer-Tropsch slurry phase reactor using a Co-based catalyst and noted a marked deactivation by halides and potassium [2]. The authors related the metallic cobalt poisoning to a blocking effect of metal sites by halides and potassium in agreement with results of H2 chemisorption obtained ex situ on used cobalt FT catalysts [2]. We report herein an operando diffuse reflectance FT-IR spectroscopy (DRIFTS) study in which the hydrogenation of CO is monitored in the presence of 150 ppm of chlorine. DRIFTS enables monitoring the state of the cobalt surface, which is typically covered with various carbonyl species under reaction conditions [3].
2.Experimental
The preparation of the 14 wt.% Co/Al2O3 catalyst, experimental setup, DRIFTS cell and procedures are described in details elsewhere [3]. The hydrogenation of CO was performed at atmospheric pressure using a modified high temperature DRIFT cell. The reactor effluent was quantified using a 2 m-pathlength FT-IR gas cell. Methane, ethane, propene and methanol were the main reaction products in the present conditions and were quantified, while higher hydrocarbons were not monitored. The chlorine was introduced via a saturator held at -70°C containing trichloroethylene (TCE). Differential conditions were used, to ensure that the catalyst bed remained gradientless.
3.Results and discussion
The rates of formation of the main products sharply decreased upon the introduction of the chlorinated compound, typically showing a 10-fold drop. The IR signal of CO(ads) could be essentially decomposed into three bands, whether chlorine was present or not. The effect of chlorine was strongest for a surface carbonyl “B” characterized by a band located at ca. 1881 cm-1, likely a bridged species formed on edges, steps and defect sites. The corresponding C-O bond strength increased by about 15 ± 8 kJ/mol in the presence of chlorine, based on a calibration curve presented elsewhere [3]. Two other linear carbonyl species (noted L1 and L2) were affected by chloride addition, albeit to a lesser extent.
It must be stressed that the relative decrease of the overall carbonyl band area was far lower than that of the activity, suggesting that each chlorine atoms poisoned several surface sites and/or selectively poisoned the most active sites. The most active sites were likely those associated with the bridged CO(ads) “B”, which displayed the highest intensity loss.
The effect of Cl was partly and slowly reversible. Surprisingly, the chlorine poisoning hardly affected the position of the main on-top CO(ads) IR band, in contrast to the effect of surface coverage when CO was removed from the feed.This effect is consistent with chlorine decreasing the electronic density at the cobalt metallic centers, which in turn limited the electronic back-donation from cobalt to the antibonding molecular orbitals of the adsorbed CO [3].
Assuming that the rate-determining step of the hydrogenation reaction was the C-O bond dissociation of CO(ads), the operando DRIFTS data indicated that the loss of activity was in part related to the decrease of the concentration of the carbonyl species. In addition, the electronic density removed by chlorine from the metallic cobalt limited the electronic back-donation to CO(ads) and thus limited the weakening of the C-O bond, therefore providing another cause for the observed rate decrease
4.Conclusions
Our operando DRIFTS investigation showed that chlorine strongly poisoned the activity of cobalt for CO hydrogenation. The effect is both site poisoning and electronic.The most active sites are suggested to be those associated with bridged CO(ads) absorbing at ca. 1881 cm-1, the heat of adsorption of which is increased by about 15 kJ/mol in the presence of chlorine.
5. References
[1] Ø. Borg, N. Hammer, B. C. Enger, R. Myrstad, O. A. Lindvåg, S. Eri, T. H. Skagseth, E. Rytter, J. Catal. 279 (2011) 163.
[2] M. K. Gnanamani, V.R.R. Pendyala, G. Jacobs, D.E. Sparks, W.D. Shafer, B.H. Davis, Catal. Lett. 144 (2014) 1127.
[3] A. Paredes-Nunez, D. Lorito, Y. Schuurman, N. Guilhaume, F.C. Meunier, J. Catal. 329 (2015) 229–236.