Theoretical Study On Acrolein Hydrogenation By Gold Catalysis And The Support Effect

Since the findings of Haruta and Hutchings that gold nanoparticles are catalytically active for CO oxidation at low temperatures and hydrochlorination of ethyne, respectively, gold catalysts have been applied for a wide range of reactions. Claus et al. show that gold nanoparticles have good activity as well as selectivity for the hydrogenation ofα,β-unsaturated aldehyde toα,β-unsaturated alcohols, which are important intermediates for production of perfumes, flavoring, and pharmaceuticals. There are two functional groups (C=C and C=O) inα,β-unsaturated aldehyde. The C=C hydrogenation leads to saturated aldehydes and C=O hydrogenation to unsaturated alcohol. For theα,β-unsaturated aldehyde with substitutes on the C=C bond, the selective hydrogenation of the C=O group can be easily realized due to the steric effects on the C=C group. For the simplest acrolein (AC) where there are no substituents on the C=C bond, however, traditional catalysts are hardly efficient for the selective hydrogenation of the C=O bond to produce allyl alcohol (AyOH). For example, on Pt/ZrO2 catalyst the relative yield of AyOH via hydrogenation of AC is less than 4%, compared to 42% on Au/ZnO.Despite the various experimental studies, there are still uncertainties for the active sites of gold particles for H2 activation. First, Corma et al. proposed that the active sites toward H2 dissociation must be neutral or have a net charge close to zero. However, the recent experiments demonstrated that the perimeter interfaces between gold and oxide are the potential active centers for H2 dissociation. We investigated the H2 adsorption and dissociation on gold clusters supported on ZnO surface, and find that depending on the size and shape of the clusters on a specific oxide surface, the active sites of the gold clusters are either low-coordinated neutral atoms or those on the perimeter interface. These gold atoms on the active sites contribute the most to the frontier molecular orbitals.No definite conclusions have been drawn about the active sites for C=C and C=O hydrogenations. Claus et al. argued that the low-coordinated Au atoms on the nanoparticles are the active sites for C=O hydrogenation. However, we computationally studied the hydrogenation of AC on AU20 cluster and on Au(110). Comparing the results from different models, we demonstrated that the active sites for C=O hydrogenation are on the high coordinated flat surfaces whereas the low-coordinated sites are favorable for the C=C hydrogenation. Small sized gold nanoparticles have more low-coordinated sites while large ones possess more surface sites. Indeed, the experiments show that the yield of AyOH increases with the increase of the gold particle size, which is in nice accordance with our point of view about the active sites.Compared with plenty of computational studies on the platinum surface, few studies have been conducted on the acrolein hydrogenation at the gold catalysts. We have systematically studied the mechanisms of acrolein hydrogenation in gas phase, on supported and unsupported gold clusters as well as on flat surface. Different from Pt in which the kinetics of C=O hydrogenation is controlled by the desorption of the produced AyOH from the metal surface, both H-addition of C=O bond and the desorption of AyOH determines the whole kinetics of C=O hydrogenations on the gold catalyst. Unlike the situation on Pt surface where both kinetic and thermodynamic effects favor the C=C hydrogenation, C=O hydrogenation is more favorable than that of C=C group, though the latter is thermodynamically favored. Therefore, good selectivity towards AyOH on gold catalysts than on Pt catalysts is well rationalized.Metal-support interactions play important roles in the metal catalysis. Less is known for the mechanism of support effects on the catalytic reactions. We use the combined Au/ZnO models and investigate the nature of metal-oxide interaction as well as its effects on the AC hydrogenation. It is revealed that the interaction leads to the charge transfer from the gold catalysts to the ZnO support. The positively charged gold clusters change the stability of H atom during acrolein hydrogenation. Because the interaction of acrolein and its derivatives with the substrates hardly changes, the stability variation of H atom from the initial to the transition states contributes mostly to the hydrogenation barrier changes from the unsupported to supported catalysts. For the first time we proposed a H-mediated mechanism of support effects. This finding should be significant for further studies on the metal oxide support effect on catalysis.Carbon nanotubes (CNT) have been paid much attention for the potential applications in metal catalysis. Most studies are about the promising properties for catalytic improvements by CNT confinement effects. We investigate both the encapsulation and the deposition interaction between small gold clusters and CNT, as well as the H2 dissociation on the CNT supported Au clusters. Different from the popular viewpoints that confinement effects improve the metal catalysis, we find that H2 dissociation on the gold clusters is hindered by the confinements effects of CNT, which indicates that the CNT confinement effect do not always promote the catalytic performance of a catalyst.In a word, focusing on the hydrogen dissociation and hydrogenation of acrolein on gold catalysts, this dissertation elucidates the hydrogenation mechanism of acrolein on gold catalyst and clarifies the active sites of gold catalysts for both hydrogen activation and acrolein hydrogenation. For the first time we propose an H-mediated mechanism of support effects. Finally we demonstrate that CNT confinement effect may inhibit some reactions.