Zeolite- and MgO-supported rhodium complexes and rhodium clusters: Tuning catalytic properties to control carbon–carbon vs. carbon–hydrogen bond formation reactions of ethene in the presence of H 2 Pedro Serna a,b , Bruce C. Gates b,⇑ a Instituto de Tecnología Química, Universidad Politécnica de Valencia-Consejo Superior de Investigaciones Científicas, Avda. de los Naranjos s/n, 46022 Valencia, Spain b Department of Chemical Engineering and Materials Science, University of California, Davis, CA 95616, USA article info Article history: Received 1 March 2013 Revised 5 July 2013 Accepted 9 July 2013 Available online 30 August 2013 Keywords: Catalysis by rhodium Zeolite-supported catalysts Carbon–carbon bond formation Metal/acid cooperative mechanisms abstract Essentially molecular rhodium catalysts were made from Rh(C 2 H 4 ) 2 (acetylacetonate) on zeolite HY and on MgO and characterized by infrared and X-ray absorption spectroscopies. The supported rhodium spe- cies anchored to the zeolite, initially in the form of Rh(C 2 H 4 ) 2 , selectively catalyzed ethene dimerization, typically at 298 K and 1 bar, but when the catalyst was poisoned by CO, or the support was changed to MgO or zeolite NaY, or the rhodium was converted into small clusters, the ethene underwent predomi- nantly hydrogenation. The preciseness of the synthesis of the supported rhodium species facilitated determination of structure-catalyst performance relationships that led to a schematic representation of how the dimerization proceeds by a mechanism involving both the rhodium complexes and zeolite sur- face OH groups. The reaction is facilitated by H 2 and proceeds as one ethene molecule is activated by an isolated rhodium complex and another by a weakly acidic Si–OH–Al group. Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction Most supported metal catalysts consist of particles that are large enough to resemble pieces of bulk metal, but when the dimensions of the particles become less than about a nanometer, their properties become essentially different. As the number of me- tal atoms in a particle approaches one, the supported species take on a molecular character, with the metal often being cationic when bonded to an oxide or zeolite support. Thus, there are excellent opportunities to tune the catalytic properties of highly dispersed supported metals, and the design variables are the number of me- tal atoms per particle and the ligands bonded to the metal, which include the support. The opportunities to tailor the catalytic properties of supported metals are evidently maximized when the metals are chosen to be those that are catalytically active in a variety of forms for numer- ous reactions. Gold and rhodium are thus exemplary; single-me- tal-atom (mononuclear) complexes and small clusters of these metals are catalytically active for many reactions, including those involving alkenes [1,2], acetylenic compounds [3,4], and CO [5– 7], among others. Elucidation of the patterns of catalytic activity and selectivity as a function of the structure of highly dispersed supported gold has been slow to develop because of the difficulty of stabilizing gold in the form of highly dispersed species—these are readily aggregated under even mild reaction conditions. Rho- dium seemingly offers better opportunities to elucidate the cata- lytic properties of highly dispersed supported metal species because it is possible to prepare stable catalysts not only in the form of mononuclear rhodium complexes but also in the form of small clusters, including dimers [8,9] and clusters of only a few Rh atoms each [10,11]. Reactions catalyzed by rhodium in various highly dispersed forms include alkene hydrogenation, alkene hydroformylation, and other reactions of CO and H 2 [2,6]. In the form of metallic par- ticles, rhodium is active for many reactions, exemplified by hydro- genation of alkenes and aromatics [12]. Our group has been investigating strategies for the synthesis, characterization, and control of the catalytic properties of supported rhodium catalysts for a number of years, refining the approaches to identifying and assessing the catalyst design variables with the goals of making catalysts that are highly uniform in structure and that allow tuning of the activity and selectivity by controlling the structures of the surface species. The progression from early [13] to recent [8] work indicates an evolution in the choice of metal precursors, classes of supports, and techniques for characterizing structure and catalyst function, [14–17] among others. The precursors are organometallic compounds because they offer control of the reactivity with 0021-9517/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcat.2013.07.005 ⇑ Corresponding author. E-mail address: bcgates@ucdavis.edu (B.C. Gates). Journal of Catalysis 308 (2013) 201–212 Contents lists available at ScienceDirect Journal of Catalysis journal homepage: www.elsevier.com/locate/jcat