Osmium-copper clusters

Ruthenium-copper and osmium-copper clusters (21) are of particular interest because the components are immiscible in the bulk (32). Studies of the chemisorption and catalytic properties of the clusters suggested a structure in which the copper was present on the surface of the ruthenium or osmium (23,24). The clusters were dispersed on a silica carrier (21). They were prepared by wetting the silica with an aqueous solution of ruthenium and copper, or osmium and copper, salts. After a drying step, the metal salts on the silica were reduced to form the bimetallic clusters. The reduction was accomplished by heating the material in a stream of hydrogen. 

The results of the EXAFS studies on osmium-copper clusters lead to conclusions similar to those derived for ruthenium-copper clusters. That is, an osmium-copper cluster Is viewed as a central core of osmium atoms with the copper present at the surface. The results of the EXAFS investigations have provided excellent support for the conclusions deduced earlier (21,23,24) from studies of the chemisorption and catalytic properties of the clusters. Although copper is immiscible with both ruthenium and osmium in the bulk, it exhibits significant interaction with either metal at an interface. 

The ruthenium-copper and osmium-copper systems represent extreme cases in view of the very limited miscibility of either ruthenium or osmium with copper. It may also be noted that the crystal structure of ruthenium or osmium is different from that of copper, the former metals possessing the hep structure and the latter the fee structure. A system which is less extreme in these respects is the rhodium-copper system, since the components both possess the face centered cubic structure and also exhibit at least some miscibility at conditions of interest in catalysis. Recent EXAFS results from our group on rhodium-copper clusters (14) are similar to the earlier results on ruthenium-copper ( ) and osmium-copper (12) clusters, in that the rhodium atoms are coordinated predominantly to other rhodium atoms while the copper atoms are coordinated extensively to both copper and rhodium atoms. Also, we conclude that the copper concentrates in the surface of rhodium-copper clusters, as in the case of the ruthenium-copper and osmium-copper clusters. 

Ruthenium-copper and osmium-copper clusters are examples of bimetallic clusters in which one component is from Group VIII and the other from Group IB of the periodic table. These clusters are of particular interest because copper is virtually completely immiscible with either ruthenium or osmium in the bulk (7). Copper has the face-centered cubic structure in the metallic state, whereas ruthenium and osmium both exhibit hexagonal close-packed structures (8). 

The discussion of EXAFS on ruthenium-copper clusters in the previous paragraphs emphasizes the qualitative aspects of the data analysis. A quantitative data analysis, yielding information on the various structural parameters of interest, has also been made and published (31). Of particular interest was the finding that the average composition of the first coordination shell of ruthenium and copper atoms about a ruthenium atom was about 90% ruthenium, while that about a copper atom was about 50% ruthenium. Details of the methods involved in the quantitative analysis of EXAFS data on ruthenium-copper clusters are not considered here, since the technique is described in the following discussion of EXAFS studies on osmium-copper clusters. 

EXAFS Studies of Osmium-Copper Clusters (32). Results of EXAFS studies on osmium-copper clusters dispersed on silica lead to conclusions similar to those derived for ruthenium-copper clusters. The discussion here is concerned with a catalyst with a 1 1 atomic ratio of copper to osmium. The clusters constituted 2.66 wt% of the total catalyst mass (2% Os, 0.66% Cu). The average diameter of the clusters was estimated to be about 15 A. 

Plots of the function K xUO vs. K at 100°K for the extended fine structure beyond the osmium edge for pure metallic osmium, and for the osmium-copper clusters in the catalyst containing 2 wt% Os and 0.66 wt% Cu, are shown in the left-hand sections of Figure 4.13. The associated Fourier transforms of the functions are shown in the right-hand sections of the figure. As previously noted, the Fourier transform yields the function n(R), the peaks of which are displaced from the true interatomic distances because of the phase shifts. Similar plots for the extended fine structure beyond the copper K edge for pure metallic copper and for the osmium-copper catalyst are given in our 1981 paper (32). 

Phase shifts determined in this manner for OsOs and CuCu were employed to be consistent with the use of adjusted theoretical phase shifts for the osmium-copper atomic pair, which will be considered subsequently in the analysis of data on silica-supported osmium-copper clusters. For the osmium-copper pair, two phase-shift functions are necessary, depending on which of the atoms is the absorber atom and which is the backscattering atom. The two situations are distinguished by using the designation OsCu for the... 

In summary, the results of the EXAFS studies on osmium-copper clusters indicate that the osmium atoms in the clusters are coordinated predominantly to other osmium atoms, while the copper atoms are extensively coordinated to both copper and osmium atoms. The results are thus very similar to those obtained for ruthenium-copper clusters. Since the copper atoms appear to have essentially equal numbers of copper and osmium atoms as nearest neighbors, it seems reasonable to conclude that the osmium-copper clusters consist of small patches or multiplets of copper atoms located on the surface of the osmium. 

The size of the osmium-copper clusters of interest in the catalyst considered here is such that the number of metal atoms which could be present in a full surface layer is significantly higher than the number that would be located in the interior core. For a stoichiometry of one copper atom per osmium atom, there are, then, too few copper atoms to form a complete surface layer around the osmium. It should be realized that parameters derived from the EXAFS data on the osmium-copper clusters are average values, since there is very likely a distribution of cluster sizes (9) and compositions in a silica-supported osmium-copper catalyst. 

For the catalyst containing osmium alone on silica, the osmium clusters behave as if they are more electron deficient than pure metallic osmium that is, there appear to be more unfilled d states to accommodate the electron transitions from the 2pin core level of the absorbing atom. In the silica-supported osmium-copper clusters, however, the osmium atoms appear to be less electron deficient than they are in the pure osmium clusters dispersed on silica. The presence of the copper thus appears to decrease the number of unfilled d states associated with the osmium atoms. This observation is the first that we have made regarding the electronic interaction between the components of a bimetallic cluster catalyst. Further studies of such interactions are currently in progress on other bimetallic catalysts. 

---