Eight-atom clusters, reactivity

To understand the high reactivity of the Nig cluster is a remaining problem. All studies, theoretical and experimental, agree that the monovalent eight atom cluster is a particularly stable cluster, due to its closed shell character. Even though Nig should be a monovalent cluster, the reactivity of this cluster does not seem to be lower than for the other clusters. Work to resolve this problem is currently in progress. 

In a similar fashion to Au, Pt clusters showed an abrupt increase in reactivity for n>8 [54]. For Pt clusters in the range 8200 K. However, an additional CO peak was observed for 15-20 atom clusters at 150 K. The reactivity of the clusters increased from one CO molecule produced per eight atom cluster to six CO molecules produced by clusters containing between 15 and 20 atoms leading to a maximum in activity (on per atom basis) at a cluster size of n= 15. Heiz et al. also noted that a small decrease in activity was observed for Pt,j due to the stable (and therefore less reactive) icosahedral structure. However, if CO was dosed first, Pt clusters were effectively poisoned and no CO was produced. 

Wc will first examplify the above principles on a six atom and a seven atom cluster. These two clusters are the most striking examples of the differences beween nickel and cobalt clusters. These cobalt clusters are the two least reactive of all the cobalt clusters, whereas for nickel these two clusters are as reactive as clusters of other sizes. In the final subsection we will discuss also the four, five, eight, nine and ten atom clusters of which in particular the eight and nine atom clusters have a markedly lower reactivity for cobalt than for nickel. All calculations described in this section used a one- or two- electron ECP level description of the metal atoms. No all-electron atoms were included. 

Other Clusters. The cobalt clusters with six and seven atoms are the ones which differ most dramatically from the corresponding nickel clusters. The eight and nine atom clusters are also much less reactive for cobalt than for nickel, while the Co g cluster is as strongly reactive as the Ni.g cluster. The experimental result for the four and five atom cobalt clusters is not completely clear to us. For the five atom cluster it is on the one hand claimed that it has reacted almost completely but on the other hand the relative reactivity is given in a figure as smaller than for the unreactive COg cluster (1). For all these clusters we have obtained some preliminary results, which are described below. 

The correlation implicit in Eq. (21) is quantitatively best for clusters of iron and vanadium consisting of more than eight atoms. For the case of niobium clusters, the correlation is more qualitative, but a distinct one-to-one correspondence of local minima in reactivity with local maxima in IP is indeed observed for clusters containing eight or more atoms. 

The next possibly non-reactive electronic state of Co has two d s and eight d s atoms. The most stable cluster of this type can be viewed as a double bond between two pyramids with the two d s atoms at the pyramidal tops. For this cluster, however, the pyramids are associated via two triangular faces. An alternative way to picture this cluster is as an octahedron bound to two COg d s systems on opposite sides of the octahedron, thus forming the two pyramids. This cluster state turned out to be 23.9 kcal/mol less bound than the above Co q cluster state and is thus excluded. ... 

---