44045-AC5
Electronic Structure Studies of Transition Metal Phosphides
The catalysts employed in traditional commercial hydrodesulfurization (HDS) and hydrodenitrogenation (HDN) processes are sulfides of Mo or W promoted by Co or Ni (often designated as CoMoS or NiMoS catalysts) The need to develop ever more active HDS/HDN catalysts has led to research focused on the search for new catalytic materials. A number of studies have shown that MoP, WP, CoP, Co2P, Ni2P, CoMoP, and NiMoP are all active HDN and HDS catalysts, and several of the phosphides have been reported to exhibit higher HDS and/or HDN activity than the commercial CoMoS or NiMoS catalysts.
Although reports have appeared discussing certain aspects of the electronic structure of some of the phosphides there has been no comprehensive study of the electronic structure and bonding in this whole group of catalytically active phosphides. We have carried out such a comprehensive study. MoP, WP, CoP, Co2P, Ni2P, CoMoP, and NiMoP exhibit several different, but in some cases related, crystal structures. The results of our Fenske-Hall band structure calculations have enabled us to understand the nature of the metal-metal, metal-phosphorous, and phosphorus-phosphorous bonding in these materials, and to compare the electronic structure of the phosphides which have similar or related crystal structures.
Our recent work has focused on the surface electronic structures of Ni2P and NiMoP. Comparisons of the electronic structures of Co2P, CoMoP, Ni2P, and NiMoP are particularly important, because, unlike the MoS2 based catalysts where addition of Co or Ni "promotes" the activity of the catalyst, the mixed Co/Mo and Ni/Mo phosphides show a marked decrease in activity compared to Co2P or Ni2P. The structures of Co2P, CoMoP, Ni2P, and NiMoP are similar, though not identical. In Co2P and Ni2P the metal atoms occupy an equal number of tetrahedral and square pyramidal sites. In CoMoP and NiMoP, the square pyramidal metals are replaced by Mo. The decrease in activity when Mo substitutes for the square pyramidal metals suggests either that the square pyramidal Ni atoms provide active sites in Co2P and Ni2P or that introduction of Mo alters the electronic structure of the tetrahedral Ni in CoMoP or NiMoP. For the bulk materials, substitution of Mo alters the nature of the states around the Fermi level so that in NiMoP the majority of these states are Mo in character and in CoMoP about half are Mo in character. Since the catalytic activity is determined by the surface electronic structure, however, surface calculations are necessary to explain the change in activity brought about by the substitution of Mo. Ni2P, and NiMoP are particularly attractive candidates for surface calculations because they clearly exhibit surfaces that expose only tetrahedral metals or only square pyramidal metals. Experimental evidence has suggested that both these surfaces are stable. Our results show that the electronic structure of the surface which exposes the square pyramidal metals is very different in Ni2P and NiMoP. This is not surprising, since the surface atoms are different. What is surprising is that the electronic structure of the surface which exposes the tetrahedral Ni atoms is very similar in Ni2P and NiMoP. While the introduction of Mo changes the bulk electronic structure and the electronic structure of the square pyramidal surface, it has almost no effect on the electronic structure of the tetrahedral surface. These results, taken together with the experimental findings that substitution of Mo leads to a decrease in activity, indicate that it is the square pyramidal Ni atoms in Ni2P are responsible for the catalytic activity.
