Membranes are vital and touch every aspect of our daily lives. For instance, they ensure that clean or pure drinking water can be produced, or they can purify our air. However, there is no one-membrane material that can perform all these objectives, and they must be catered for each specific application. As more traditional membrane materials are either polymer based or inorganic/ceramic based, each application will require a different material that can use one or both of these components. Currently, much research is dedicated to combining both polymer and inorganic based materials to form hybrid membranes that could "have the best of both worlds". The most common form of hybrid membranes are mixed matrix membranes, or MMM. However, their synthesis normally involves the direct addition of inorganic particles into an existing polymeric solution, leading to composite materials without strong interaction between organic and inorganic moieties. In this work, we explore the reversal of this process; that is, the formation of the polymeric component directly from the surface of these particles using surface-initiated atom-transfer radical polymerization, or si-ATRP. Such a method intends to provide a high intimacy between polymers and ceramics, and strong interconnection via covalent bonds. Using a latex based solution of these functionalized titania particles, we present two pathways, known as "Coating Onto" and "Grafting From" pathways, in which these latexes can be applied to larger, tubular membrane coatings. These latexes were first analyzed by TGA to determine their functionalization densities and degree of polymerization. SEM was then employed to visually assess the membrane surfaces to determine the final optimal conditions used for preliminary gas separation tests. (C) 2016 Elsevier B.V. All rights reserved.< Réduire