RESUMEN
The understanding, at the atomic level, of the role played by additives (dopants or promoters) in the chemistry of an industrial catalyst is a very complex and difficult task. We succeeded in this goal for the ethylene oxychlorination catalyst (CuCl(2)/gamma-Al(2)O(3)), used to produce dichloroethane, a key intermediate of the polyvinyl chloride chemistry (PVC). Among the most used additives for both fluid and fixed beds technologies (LiCl, KCl, CsCl, MgCl(2), LaCl(3), CeCl(4)) we have been able to highlight that KCl, and CsCl, forming in reaction conditions a mixed phase with CuCl(2), strongly modify the catalyst behaviour. In particular, these additives are able to displace the rate determining step from the CuCl oxidation (undoped catalyst) to the CuCl(2) reduction. This results from the decrease of the rate of the latter reaction, thus the overall activity of the system. For all remaining additives the rate determining step remains the CuCl oxidation, as for the undoped catalyst. These results have been obtained coupling the catalyst activity monitored with a pulse reactor working in both non-depletive and depletive modes with time resolved XANES spectroscopy performed under in operando conditions (i.e. coupled with mass spectrometry). Formation of CuK(x)Cl(2+x) and CuCs(x)Cl(2+x) mixed phases has been proved monitoring the Cu(II) d-d transitions with UV-Vis spectrometer and the CO stretching frequency of carbon monoxide adsorbed on reduced catalyst by in situ IR spectroscopy. Finally, of high relevance is the observation that the fully oxidized catalyst is inactive. This unexpected evidence highlight the role of coordinatively unsaturated Cu(I) species in adsorbing ethylene on the catalyst surface indicating that copper, in the working catalyst, exhibits a (I)/(II) mixed valence state.
RESUMEN
The comprehensive understanding of the composition, behaviour and reactivity of a catalyst used inside industrial plants is an extremely hard task that is rarely achieved. It requires the use of different spectroscopic techniques, applied under in situ or in operando conditions, and combined with the investigation of the catalyst activity. Often the operating experimental conditions are different from technique to technique and the different results must be compared with care. In the present contribution, we combined in situ XANES/EXAFS, IR spectroscopy of adsorbed CO, CO chemisorption and catalytic tests performed using a pulse reactor in depletive mode. This multitechnical approach resulted in the understanding of the role that dopants (LiCl, KCl, CsCl, MgCl(2) LaCl(3)) have in the nature, relative fraction, reducibility and dispersion of Cu-phases on CuCl(2)/gamma-Al(2)O(3) catalysts for oxychlorination reaction, a key step of the PVC chemistry. In the undoped catalyst two Cu phases coexist: Cu-aluminate and supported CuCl(2), being the latter the only active one [J. Catal., 2000, 189, 91]. EXAFS and XANES highlighted that all dopants contribute more or less efficiently in increasing the fraction of the active copper species, that reaches a value of almost 100% in the case of MgCl(2) or LaCl(3). EXAFS directly, and IR indirectly, proved that the addition of KCl or CsCl (and less efficiently of LiCl) results in the formation of mixed CuK(x)Cl(2+x) or CuCs(x)Cl(2+x) phases, so altering the chemical nature of the active phase. XANES spectroscopy indicates that addition of MgCl(2) or LaCl(3) does not affect the reducibility by ethylene (under static conditions) of the active CuCl(2) phase and that the reducibilility of the new copper-dopant mixed chloride are in the order CuCl(2) > CuLi(x)Cl(2+x) > CuK(x)Cl(2+x) > CuCs(x)Cl(2+x). However, when reduction is done inside a pulse reactor, a more informative picture comes out. The last technique is able to differentiate all samples, and their ability to be reduced by ethylene resulted in the order: La- > Mg- > Li-doped > undoped > K- > Cs-doped catalyst. To understand this apparent discrepancy the dispersion of the active phase, measured by CO chemisorption, was needed: it has been found that addition of LiCl increases enormously the dispersion of the active phase, LaCl(3) significantly and MgCl(2) barely, while addition of both KCl and CsCl results in a decrease of the surface area of the active phase. The mechanism of the enhancing effect of La and Mg on catalytic activity is still not clear, but it could be associated to the modification that they induce to the support surface: the Cu is so highly dispersed that almost all is in direct contact with support surface. It is finally worth noticing that the previous EXAFS and XANES study allowed us to refer the chemisorption data to the active phase only, while the IR study allowed us to fix the Cu(+)/CO surface stoichiometry. Summarizing the use of a multidisciplinary approach has been the conditio sine qua non (mandatory condition) to understand the complex role that the different additives have on the active phase of the CuCl(2)/gamma-Al(2)O(3) catalysts for ethylene oxychlorination.