Catalyst dispersion on supported ultramicroporous inorganic membranes using derivatized silylation agents

Abstract

Two approaches were utilized to disperse rhodium metal particles onto the ultramicroporous silica separation layer of a composite inorganic membrane prepared by modification of a commercial alumina support. In one approach, the silica membrane surface was first amine-derivatized using the silylation agent H 2N(CH 2) 2NH(CH 2) 3Si(OCH 3) 3, followed by reaction with the metal-organic compound [(1,5-COD) RhCl] 2. This approach led to uniformly dispersed ca. 6 nm rhodium particles located only on the surface of the membrane (as revealed by cross-sectional transmission electron microscopy) after a hydrogen reduction at 200°C. The size of the [(1,5-COD) RhCl] 2 molecule combined with a reduction of the pore size due to silylation apparently prevented significant penetration into the membrane separation layer. Modest reductions (ca. 50%) in helium and nitrogen permeability resulted from this treatment. An alternative procedure involved wet impregnation of the membrane with a (Rh (acac) 3)/THF solution, followed by air calcination and hydrogen reduction. This approach produced a much lower coverage of ca. 4 nm rhodium particles. Helium and nitrogen permeability measurements indicated that considerable pore blockage existed after the calcination step, and that the reduction step led to defect or crack formation in the SiO 2 membrane layer as evidenced by a sizeable increase in permeability. It was proposed that the silylation approach led to the [(1,5-COD) RhCl] 2 precursor being located primarily on the surface with little pore penetration, while the Rh (acac) 3 was able to penetrate the pores to a great extent, leading to damage of the membrane layer during decomposition of the ligands. The silylation approach appears to be a general strategy to control the size, surface coverage, and location of catalyst in catalytic membrane design.