Reverse hydrogen spillover

It is evident that the supported clusters have a strong affinity for hydride ligands provided by the support. The process by which the support delivers these ligands is referred to in the catalysis literature as reverse hydrogen spillover. The opposite process (spillover), well known for supported metals [36], is shown by the theoretical results to be a redox process in reverse spillover, the support hydroxyl groups oxidize the cluster. 

Scheme 9.1 A mechanism for metal-catalysed hydrogen spillover, shown by the exchange of support hydroxyls with deuterium. The process can extend to the whole surface (A), but HD is formed by reverse spillover (B), followed by desorption.

Reverse spillover or back-spillover is observed to proceed by surface migration of the spiltover species from the accepting sites to the metal, where it desorbs as H2 molecules or reacts with another hydrogen acceptor such as 02, pentene, ethylene, etc. Reverse or back-spillover (primary as well as secondary) is hindered by H20 (11), whereas secondary spillover is promoted by H20 (case B in Fig. 1). Hydrogen spillover depends on the acceptor surface it is thought to be easier on silica than on alumina (45) for hydrogen-molybdenum-bronze preparation. 

FiC5. 14. NMR spectra of Ru/SiOi. The initial number of silanol protons has been reduced by exchange with deuterium. Both traces are difference spectra with respect to the state after initial evacuation. The continuous line represents a sample under 20 Torr of H, gas, and the dashed line represents a sample after pumping away the reversible hydrogen. There is both reversible and irreversible spillover to the support (signal at. 3 ppm), and ther e is rever sible and irreversible chemisorption on the metal (sigiral at 65 ppm). [Reproduced with permission from Uner et al. (47). ... 

Fig. 15. llrac evolution of hydrogen chemisorption and spillover on the catalyst of Fig. 14. At time t = 0 the sample was exposed to 20 Torr of II2 gas. The solid circles show thaf the quantity of H/metal (intensity of the peak at —65 ppm in Fig. 14) equilibrates rapidly, whereas the total quantity of hychogen (the sum of both intensities in Fig. 14 open squares) continues to evolve in time. This pattern is confirmed by a separate volumetric experiment (solid squares). At time I = 40 min, the pumping away of the reversible hydrogen starts, lire H/metal signal again equilibrates rapidly (solid circles), whereas the total quantity of diminishes more slowly (open squares). A single volumetric experiment (cross) confirms the value found by the open squares. [Reproduced with permission from Uner et al. (47).]...

Fig. 17. A plot similar lo that of Fig. 16 but for Rh/SrTiO,j. The open symbols are the adsorption after the initial evacuation. After the highest pressure point was taken, the reversible hydrogen was pumped away, and the intensities Ia and /b decreased to the level indicated by vac. Next, the adsorption experiment was repeated (solid symbols). The increase of Ia with pressure (the spillover) is less marked than indicated in Fig. 16. [Reproduced with permission from Rojo el al (51). CopyTight 1994 American Chemical Society.]...

Hence the 30 % that are lost could be on the support and slowly react with 1-hexanol to form HA. However when aniline is reacted there is no significant loss of material, which suggests that aniline cannot interact directly with the surface hydroxyls. This suggests that the interaction between aniline and the support hydroxyls is not as simple as shown above, rather it is more likely that the reaction operates via a spillover mechanism involving an intermediate in the nitrobenzene hydrogenation sequence rather than aniline. The alkylation reaction between aniline and 1-hexanol takes place on the metal function, therefore the reaction with the missing aniline associated with the support will be slow as it requires a reverse spillover and a diffusion across the support surface. 

For reactions that are the same on metal and other catalytic sites (e.g., hydrogenation or total oxidation), the reaction may seem to proceed in a similar fashion on the metallic source of spillover and on the diluent support. Some careful studies may be able to discriminate between activity on the metal and the spillover-induced sites. As an example, hydrogenation of ethylene occurs on Pt (or Ni) and on silica or alumina activated by spillover. The product (i.e., only ethane) is the same, as the kinetics often are (rate = /c[C2H4]°[H2]1), but the specific mechanism is different. Deuteration is able to discriminate between the relative rate of alkyl reversal. Deuteration of ethylene on an activated silica produces d2-ethane as the initial product (137), contrary to the results for metal-catalyzed ethylene hydrogenation (2). 

---