The synthesis of microporous silica materials for the production of highly gas selective membranes is achievable through sol-gel chemistry. These materials can be tailored to impart certain functionalities, such as molecular sieving pore size control, or the creation of preferential gas adsorption sites by embedding specific metals or metal oxides on the nanoscale. However, to date only a small number of single metal dopants have been investigated, which leaves significant scope for further exploratory work. As such, this thesis explores the incorporation of novel metals and metal combinations and investigates how the metal-silica and metal-metal interactions affect material formation as well as membrane performance.
The first contribution of this thesis found that lanthanum does not behave like conventional transition metal dopants (Co, Ni, Fe etc.). The formation of a lanthanum silicate phase changes the pore size distribution of the silica xerogel. At low lanthanum concentrations (La/Si molar ratio ≤ 0.15) the PSD remains in the microporous region, though pore volume is reduced with increasing concentration. This was due to the conversion of microporous silica into a dense lanthanum silicate phase. At high lanthanum concentrations (La/Si > 0.15) minimal microporosity remains and the PSD shifts into the mesoporous region, with larger pore sizes achieved at higher concentrations. This mesoporosity manifests through the creation of voids between the lanthanum silicate nano-domains, a consequence of the consumption of the microporous silica.
Continuing from this study, lanthanum cobalt co-doped silica membranes were prepared to investigate both the functionalities generated by, and the effect of metal-metal interactions on lanthanum silicate formation. This lead to the second contribution of this thesis; the He/CO2 permselectivity significantly increased from 80 to 196 over 6 days of permeation testing at 500˚C. This is a novel result and suggests the formation of the silicate phase interacted with both silica matrix and cobalt nano-domains in a way where larger pore sizes preferentially closed. This is contrary to conventional metal oxide silica matrices, where the smaller pore sizes tend to collapse under heat treatment.
Based on these results, cobalt-palladium (transition metal-transition metal) doped silica materials and membranes were prepared to compare the metal-metal interactions with the cobalt-lanthanum (transition metal-lanthanide) materials. This study provided the third contribution of this thesis; the preferential reduction of palladium oxide alters the gas transport through the doped silica membrane. The reduction of palladium oxide nanoparticles to metallic palladium results in a 40% decrease in volume. The phase change caused a decrease in the He/N2 gas permselectivity from 70 to 30 which was postulated to be due to the opening of a molecular gap in the membranes porosity.
The preceding LaCo and PdCo silica membrane studies showed that the reduction of oxidised nanoparticles can open molecular porosity gaps, the larger the amount of reducible phase, the larger the deviation in permeation. To further probe the interaction between the metal nanoparticle and the silica matrix, palladium doped silica membranes were prepared under reducing conditions (H2). Palladium was chosen due to its propensity to form highly crystalline nanoparticles which enable clear characterisation. In addition, the preparation of the membrane in H2 enabled the membrane performance to be monitored when no molecular porosity gap exists to accommodate the oxidation (expansion) of the nanoparticle. This lead to the fourth contribution of the thesis; the oxidation of metallic palladium nanoparticles prepared in reducing conditions deteriorates the performance of silica membranes. This was observed through a decrease in the He/N2 permselectivity from 30 (before oxidation) to 6 (after oxidation). This permselectivity loss was not recoverable upon re-reduction. The drop in separation performance is due to the stress imparted on the surrounding silica matrix from the expansion of the palladium nanoparticles, since no molecular gap exists to accommodate their expansion. This lead to the formation of micro-cracks throughout the silica matrix which impacted separation performance.
The studies undertaken within this thesis, and as a whole throughout the associated literature, focus on the metal dopant species; how the metal dopant directly interacts with the silica material, how the metal dopant interacts with the permeating gas molecules. These studies assume that the metal dopant does not affect the formation of silica structures that are removed from the metal-silica interface. Thus the final study in this thesis investigates the influence of metal dopants (palladium chloride) on the silica sol-gel reactions through the initial sol-gel, drying and calcination steps. This study provides the fifth and final contribution of this thesis; the interaction of the PdCl2 precursor with water inhibits silica hydrolysis and condensation reactions before gelation. Thereafter, the hydrolysis and condensation reaction rates and extents are enhanced relative to the non- doped silica xerogels. This is due to both the dehydration of the aqua palladium complex and the low quantity of ethanol remaining in the gel to inhibit the reaction.