Synthesis and characterisation of homogeneous porous titanosilicate catalyst supports for carbon dioxide conversion

The work presented in this thesis has been oriented towards generating supported monometallic and bimetallic nano-catalysts for the efficient conversion of carbon dioxide, particularly via the Sabatier methanation process. Developing longevity of the catalyst’s life was a starting point for this work, since it is known that metal nanoparticles sinter (aggregate) during high temperature catalysis and thus lead to lower catalytic performance (over time). Supported nano-catalysts are known to be more effective than conventional non-nano alternatives, both because of their inherently higher surface area and because of synergistic electronic effects. Therefore, the use of nano-catalysts is generally desirable, but in order to do so and sustain catalytic activity, sintering resistance must be introduced. Titania has long been known to provide strong metal-support interactions, and consequently less sintering of reduced metal particles. Therefore, titania was employed throughout this study, but not in the bulk form; instead it was incorporated either as titanosilicate or as Ti-rich nano-clusters, which are meant to serve as discrete anchoring sites for catalytic metal nanoparticles. With this rationale in mind, a mesoporous titanosilicate MCM-41 type catalyst support was synthesised according to the conventional method of simultaneously hydrolysing and condensing silicon and titanium alkoxides in an aqueous alkaline medium and in the presence of a cationic surfactant (Chapter 3). It was found via N2 sorption and XRD that a relatively high surface area hexagonal mesoporous material was present. However, further characterisation by electron microscopy showed that the material contained phase and morphological heterogeneities which were clearly visible, both at the particle scale (e.g. 100 - 500 nm) and the bulk (particle aggregate) scale (e.g. 100 - 500 μm). At the particle scale, a hexagonal mesoporous silicate phase (as seen in XRD) and an amorphous porous titanosilicate phases were found to co-exist. At the bulk scale, these particles gathered into aggregates which were segregated into Ti-rich and Si-rich entities. In both cases, particles had atomic scale compositional homogeneity. It is notable that, although clearly revealed by electron microscopy, these heterogeneities were impervious to detection by the traditional N2 sorption and low angle XRD approaches. In early studies (Chapter 3), the hydrolysis rates, condensation rates and the precursor-surfactant assembly conditions of the multiple precursors, were found to be sub-optimal. In subsequent studies, these sub-optimal conditions were scrutinised, paying particular attention to the mechanism of formation of the matrix. Phase and morphological homogeneity was thus improved, gradually. In order to accurately characterise the bulk (aggregate) scale heterogeneities found in Chapter 3, an innovative Laser Ablation – Inductively Coupled Plasma – Time of Flight – Mass Spectrometry (LA-ICP-TOF-MS) protocol was developed (Chapter 4). With this approach it proved possible to achieve quantitative analysis of porous catalytic materials. Concentrations ranging from trace (ppm) to major (high wt %) levels were measurable. The 3D spatial distributions of elements in the materials were also measurable at a μm scale. This facilitated the characterisation of heterogeneity within and between aggregates. In addition to studying the bulk scale heterogeneity of Ti-MCM-41, the extent of and distribution of Ni impregnated into mesoporous silica MCM-41 materials was studied. Ni was chosen because it is known to be a good CO2 conversion catalyst. As part of this project, the ability to anchor metal nanoparticles to discrete anchoring sites was sought. The first approach that was attempted made use of aqueous solutions (Chapter 5) in which the reaction pH was tuned towards the isoelectric point of titania so as to slow the condensation rate relative to silica. In order to do so, a phosphate buffer was employed. The presence of phosphate anions led to its association with positively charged titanium species which, in turn, led to the formation of Ti-rich nano-clusters (50 – 100 nm) (potential anchoring sites for metal nanoparticles). These were found to be distributed uniformly in the matrix (intra-particulate cluster homogeneity). The non-ionic surfactant Triton X-100 was employed as the template in these studies, and this led to the formation of uniform micropores, though not mesopores. Of the conditions probed, the pH 5.1 condition was found to produce the highest degree of compositional homogeneity at all length scales. Having achieved this degree of success, an attempt was made to extend the degree of homogeneity to the intra-particulate (atomic) scale, making use of alcoholic rather than aqueous solutions (Chapter 6). Alkoxide precursors and the surfactant were solubilised in n-butanol in all cases and the product material generated possessed Si and Ti distribution homogeneity at all length scales. However, early attempts led to materials with poor surface areas (25-30 m2/g) and porosities (~ 13 mL of N2 at STP). It was found that hydrothermal treatment of the reaction mixture in the presence of tetramethylammonium hydroxide (TMAH) and surfactant facilitated the development of more surface area (e.g, 69 m2/g) and porosity (28 mL of N2 at STP). Even so, a significant portion of this surface area was external surface area (17.7 m2/g). So, despite the success in generating compositional homogeneity at all length scales, the internal porosity remained to be increased. In the next approach, a lyotropic liquid crystalline gel medium was employed directly (Chapter 7). Initially, the non-ionic surfactant P123 was melted, and then, the liquid alkoxide precursors (dissolved in ethanol) were poured into the melt. The resulting liquid was homogenised by stirring. To this liquid, water was added to create the liquid crystalline gel. In this way, both the liquid crystal and the pore walls were assembled simultaneously. The resulting liquid-crystal—pore wall composite could either be treated in an aqueous/alcoholic medium first (followed by calcination), or be directly calcined. The former approach gave better pore order and more surface area (374 m2/g) compared to the latter (286 m2/g). In both cases internal mesoporosity is the predominant type of porosity present. Importantly, in all cases, morphological and compositional homogeneity was achieved at all length scales – intra-particulate (atomic), inter-particulate and bulk scales. Thus, the hitherto under-appreciated, but critical problem of phase (and morphological) heterogeneity was identified, and was overcome. Therefore, potential opportunities have now been created for sintering resistant and therefore longer lasting catalysts for carbon dioxide conversion (and similar) reactions.