This research forms part of a programme of work at the University of Canterbury investigating the production of liquid fuels from biomass. The drivers for this research are the plentiful supply of woody biomass in New Zealand as well as the necessity for a reduction in the use of fossil fuels. Fischer-Tropsch synthesis has been chosen as the base conversion method for syngas to liquid fuels. While Fischer-Tropsch plants are traditionally very large, the low geographical density of the biomass feedstock necessitates a shift from a traditional economies of scale approach. In this research a sawmill integrated polygeneration scenario is proposed that recognises the synergy between the heat and electrical requirements of a mill and the Fischer-Tropsch process that can supply both as well as liquid fuels. Techno-economic modelling of variations to this polygeneration arrangement were performed using a traditional Fischer-Tropsch slurry reactor as the basis. The breakeven price of syncrude produced in the process based on a 30 year plant life and 10% discount factor was as low as $US 167 per barrel. This arrangement is coupled with development of and experimentation with a microchannel reactor in a further attempt to overcome economies of scale disadvantages. The lab scale microchannel reactor consisted of a shim with 50 channels of 37mm length with 0.2mm height and 0.3mm width. The microchannel reactor was tested with shorter run periods to compare different catalyst washcoats consisting of neat cobalt, cobalt on titania and a combustion synthesis method over a temperature range of 210-240°C at 20 bar. Comparison was also made to a lab scale fixed bed reactor with a powdered cobalt on titania catalyst. The neat cobalt washcoat proved to have the best performance per unit mass of catalyst of the three washcoats. The performance of the microchannel reactor was 32-40 times better per unit catalyst mass than the fixed bed reactor. From data based on the shorter runs the neat cobalt washcoat and the cobalt on titania washcoat were selected for further analysis over longer runs at a range of pressures from 2-20 bar and temperatures from 210-240°C. These runs were each approximately 70 hours long and provided a better analysis of the narrowed catalyst choice. The productivity results of the catalysts were fitted to established kinetic equations from literature with an excellent correlation. More accurate Anderson-Schultz-Flory selectivity values were also obtained ranging between 0.72 to 0.82. This is certainly an area that would warrant further attention as a higher selectivity has a very positive affect on plant economics. Establishment of the kinetic equations for the catalyst performance allowed modelling of reactors with greater volume along with investigation of mass transfer limitations to assist in scale up of the technology. It was found that under 4-5mm hydraulic diameter channel dimensions the mass transfer limitation from the bulk gas phase to the catalyst interface is negligible. A scaled up microchannel reactor concept design is proposed utilising stainless steel mesh folded into 2mm channels to increase catalyst surface area compared to straight shim. A costing correlation was produced per unit of reactor volume to allow a full scale cost of the microchannel reactor to be estimated for inclusion into the techno-economic model. The revised techno-economic model was optimised through pressure variation to give a breakeven syncrude value of $US118 per barrel at Fischer-Tropsch reaction conditions of 10 bar and 240°C. This brings the value well within historical crude price trends.

Fischer-Tropsch (FT) synthesis is a reaction which entails the conversion of synthesis gas, also known as syngas (a mixture of H2 and CO gases), to liquid hydrocarbon fuels, oxygenated hydrocarbons, chemicals and water. This syngas mixture is obtained from natural gas, coal, petroleum, biomass or even from organic wastes. In this study cobalt catalysts supported on novel carbon spheretitania (CS-TiO2) composite materials were synthesized and tested for their performance in the FT process. Initially carbon spheres (d = 80-120 nm) were prepared in a vertical swirled floating chemical vapour deposition reactor without the use of a catalyst. The rate of production was controlled and the highest production rate of about 195 mg/min was obtained at an acetylene (C2H2) flow rate of 545 mL/min at 1000 °C. The produced carbon spheres (CSs) had a narrow size distribution with a uniform diameter size. Purification and functionalisation of the CSs improved the total surface area, due to the removal of PAHs which blocked the CS pores. The introduction of functional groups to the CSs was achieved and these changed the wetting properties of the CSs. Functionalising the CSs for longer than 17 h in HNO3 destroyed the morphology of the CSs. After successful preparation of functionalised CSs, the interactions between CSs and TiO2 were studied by in the TiO2 composite using two different sol-gel methods, namely the conventional sol-gel and the surfactant wrapping sol-gel method. The surfactant wrapping sol-gel method entailed the modification of the CSs by dispersing them in a surfactant, in this case hexadecyltrimethylammonium bromide or CTAB [(CH3(CH2)15N(CH3)3Br]. This introduced alkyl “tails” which eased the dispersability of the CSs before coating them with Ti[O(CH2)3CH3]4 (a source of TiO2) to produce a homogeneously coated CS-TiO2 composite material (defined as ASW3). It should be mentioned that many, many experiments were performed to develop an efficient and reliable method to make homogeneously coated CS-TiO2 composites since it was found to be very difficult to achieve an interaction between carbonaceous materials and TiO2 especially by sol-gel procedures. The traditional sol-gel method was used to prepare CS-TiO2 composites with different ratios viz. 1CS-1SG, 1CS-2.5SG, 1CS-5SG, 1CS-10SG, 1CS-25SG and 1CS-50SG. These composites showed weak interactions between CSs and TiO2 even at high TiO2 loading ratio. Interestingly the surface area of these composites showed high values of 80 and 85 m2/g for 1CS-5SG and 1CS-10SG, respectively. At lower TiO2 ratios the measured surface area was similar to that of CSs, i.e 10 m2/g for 1CS-1TiO2. At high TiO2 ratios the measured surface area was similar to that of TiO2, i.e 49 m2/g for 1CS-50TiO2. The TEM images of CS-TiO2 (ASW3) composites prepared by surfactant wrapping methods showed a successful TiO2 coating of CSs. The TiO2 grain size was 8.0 nm with both anatase and rutile phases. High surface areas (up to 98 m2/g) of composite materials were achieved by employing this procedure. The high surface areas achieved suggest that the interaction between CSs and TiO2 was homogeneous and the increase was due to the “bridge” formed between CSs and TiO2. A series of cobalt catalysts (10% by weight) supported on these materials was carried out by the deposition precipitation method using Co(NO3)2·6H2O as the metal precursor. After appropriate drying and calcination the catalysts were characterized using traditional characterisation techniques and tested in the FT reaction using a fixed bed reactor. The the 10%Co/CS catalyst produced a CO conversion of 15.2% while the catalyst had a low total BET surface area (6 m2/g) compared to non-carbonaceous catalysts with higher BET surface areas. This observation suggests that the surface area did not necessarily play a role in the CO conversion, but that other properties (reducibility and dispersion) of CSs influenced the catalyst activity. After coating CSs with TiO2 and loading cobalt to produce 10%Co/ASW3 both the BET surface area of the catalyst and the CO conversion increased to 83 m2/g and 20.1%, respectively. CO-TPD of 10%Co/ASW3 showed a large amount of strongly adsorbed CO. This increased CO was due to the interaction between CSs and TiO2 which developed CO adsorptive sites.

In this study the synthesis and use of carbon nanofibers (CNFs) as catalysts supports for Fischer Tropsch synthesis is reported. The synthesis of carbon nanofibers with two distinct morphologies was optimized based on the reports in the literature that the straight (SCNF) and helical (CCNF) carbon nanofibers grow on Cu catalysts with different particle sizes. To selectively grow CNFs with a single morphology Cu catalysts were designed using different synthesis procedures (by using unsupported, coated and silica supported catalysts). The prepared copper oxide (CuO) nanoparticles were characterized by techniques such as TEM, XRD and nitrous oxide chemisorption. These techniques showed that the unsupported and coated CuO catalyst precursors has large particle sizes (range 100-300 nm) and thus had low Cu atomic surface area, while the supported CuO catalysts displayed low particles sizes in the nanoscale regime (<20 nm) and hence had high atomic surface area. Preparation of CNFs was carried out 300 using acetylene (C2H2) gas as the carbon source. Cu catalysts with large particle sizes resulted in straight CNFs and the small supported Cu nanoparticles grew helical CNFs because of the change in the nanoparticle surface energy during adsorption of the acetylene gas and the silica (SiO2) support effects that limited Cu nanoparticles from sintering (i.e. final particles size 60 nm). Soxhlet extraction proved to be an invaluable step in removing adsorbed polycyclic aromatic hydrocarbons. Because of the low thermal stability of these CNFs the materials were then annealed at higher temperatures ranging from 500-1400 in an inert environment (passing N2 gas). The helical CNFs snapped under high temperature annealing ( 900 ) resulting in shorter lengths in comparison to the straight CNFs. BET analysis of the annealed CNFs indicated that the CNFs annealed at 500 and 900 have increased surface area and have a mesoporous pore structure with the surface area ranging from 200-350 m2/g. Raman and Fourier transform IR spectroscopy indicated that the CNFs annealed at 500 and 900 , (which were the main material of interest because of their high surface area and thermal stability) had different hybridized carbon content. CNFs annealed at 500 contained both sp2 and sp3 hybridized carbon while annealing the CNFs at 900 resulted in a complete rehybridization of the carbon content to sp2. The carbon sp3 content in the CNFs annealed at 500 therefore implied that CNFs annealed at this temperature are more defective in comparison to the CNFs annealed at 900 . Since it is well known that material functionalities are affected by the amount of defects present inside the different CNFs were then applied as catalyst supports for Fischer Tropsch synthesis (FTS) to compare the support effects on cobalt active sites. The CNF surfaces were first modified by functionalization using concentrated HNO3 solution. The preparation of the catalyst systems was performed by a simple HDP method using urea. The CNFs and the FT catalysts were characterized using different techniques such as XRD, TEM, BET, TPR and Raman spectroscopy. Reactor studies performed at 220 (P = 8 bar, GHSV= 1200 mL.h-1. ) showed the catalysts had activities with CO conversion ranging from 25-45%. It was observed that catalysts supported on CNFs annealed at 500 displayed higher average activities of about 15% (based on the CO conversions) in relation to the catalysts supported on CNFs annealed at 900 . Catalysts showed minimal water gas shift reaction and high methane selectivity (i.e. 20-30%) which can be attributed to the small Co crystallite sizes and low pressure reaction conditions.

The effect of the addition of ethanol (2% and 6%) during the Fischer-Tröpsch (FT) synthesis has been investigated using a 10%Co/TiO2 catalyst in a stirred basket reactor (T = 220°C, P = 8 bar, H2/CO = 2). The transformation of ethanol vapour (2% and 6% in nitrogen) over the Co/TiO2 catalyst was also studied in the absence of the synthesis gas under FT reaction conditions. Ethanol was observed to be incorporated in the growing chain and was found to (i) increase the selectivity to light products, (ii) increase the olefin to paraffin ratio and (iii) significantly decrease the catalyst activity. These effects were almost completely reversed when the ethanol in the feed was removed. Thermodynamic predictions, TPR and XRD analysis have shown that cobalt metal particles were oxidised to CoO by ethanol but that re-reduction to Co metal was possible when ethanol was removed from the feed stream allowing the catalyst to recover most of its initial performance, in particular when high flow rates were used. The effect of the cobalt carboxylate chain length (C2, C5 and C9) used in the preparation of alumina supported cobalt catalysts has been studied by TPR, XRD and hydrogen chemisorption techniques. The activity and selectivity of the prepared catalysts have been evaluated for the Fischer-Tröpsch (FT) reaction in a stirred basket reactor. It is shown that for catalysts with Co content of 10 wt.% the activity increases as the carboxylate chain length increases while the selectivity towards methane and light hydrocarbons decreases with the carboxylate chain length. The catalyst prepared using cobalt acetate was found to present the highest metal-support interaction and the poorest performance for the Fischer-Tröpsch reaction. When the metal content was increased to 15 wt.% Co and 20 wt.% Co respectively, the metal-support interaction for the catalyst prepared from cobalt acetate significantly decreased making it a better catalyst for the FT reaction compared to the catalysts prepared from C5 and C9 cobalt carboxylates. The effect of the addition of Au to a Co FT catalyst supported on titania, alumina and silica respectively, has been investigated by varying the amount of Au (0.2 to 5 wt.%) added to the catalyst. The catalysts were characterized by atomic absorption spectroscopy, XRD, XPS and TPR analysis. The catalyst evaluation for the Fischer- Tröpsch reaction activity and selectivity was achieved in a fixed bed micro-reactor (H2:CO = 2; 20 bar; 220°C). Addition of Au to supported Co catalysts improved the catalyst reduction and the cobalt dispersion on the catalyst surface. The catalyst activity for the FT reaction and the methane and light product selectivity increased with Au loading in the catalyst.

A thesis submitted to the Faculty of Science, University of the Witwatersrand; Johannesburg, in fulfilment of the requirements for the degree of Doctor of Philosophy. July 1991 / Precipitated-iron Fischer- Tropsch catalysts were sulfided in the range 500 - 20000 ppm S/Fe with an aqueous sulfide source (Na2S, (NB4)zS, (NB4)zS5) during the precipitation process. Sulfidation was performed at pH 10.75, 8.5 and 6.9. Sodium ions were removed by centrifugation, and atomic absorption analysis confirmed low sodium levels (0-51 ppm). Based on solution speciation models, ferrous sulfide (FeS) which formed from aqueous HS' species, was found to influence the iron-oxyhydroxide crystallite morphology. It is proposed that, when sulfide was added at pH 10.75, FeS molecules functioned as nuclei for crystallite growth, while a pH 6.9 they assisted 'with the aggregation of particles. The processes of nucleation and aggregation appeared to be in competition following sulfidation at pH 8.5, resulting in a composite morphology that produced an inactive catalyst. The bulk structure of the catalysts was elucidated using XRD, SEM and nitrogen porosimetry, All sulfided catalysts exhibited enhanced BET surface areas and total pore volumes with a maximum at 2000 ppm S (surface area = 166 m2/g,total pore volume = 0.254 cm3/g) compared to an unsulfided catalyst (surface area = 58 m2/g, total pore volume = 0.184 cm3/g), Furthermore, for any series of catalysts at the same level of sulfidation, the BET surface areas were observed to decrease as the pH of sulfide addition decreased. Increasing levels of sulfidation (to 20000 ppm) brought about an increase in crystallite size and therefore, improved crystallinity as determined by XRD measurements. Materials with larger crystallites possess smaller surface areas, and thus the crystallinity was found to increase as the pH of sulfidation decreased. Surface characterisation by XPS after calcination at 400°C and reduction (400°C), revealed sulfate species (169.4 eV) on catalysts sulfided with 500-2000 ppm, while sulfide species (162.O eV) emerged at higher sulfide content. No sulfates were observed on reduced catalysts following calcination at 200 C. [Abbreviated Abstract. Open document to view full version]

MSc (Chemistry), Faculty of Science, University of the Witwatersrand, 2002

This Dissertation investigates the kinetics of CO2 methanation over nickel and cobalt catalysts. Methanation was studied for both Ni/γ-Al2O3 and Co/ZrO2 catalysts, which were synthesised using an incipient wetness impregnation technique and subsequently characterised using analyses based on gas adsorption, XRD, TPR and thermogravimetry. Separately a CO hydrogenation reaction, the Fischer-Tropsch process, was modelled numerically to examine the influence of mass transfer in practical, commercial pellets of catalyst. The kinetics of methanation was investigated for Ni/γ-Al2O3 over a wide range of reactant partial pressures using a gradientless, spinning-basket reactor operated in batch mode and in a laboratory-scale, continuous fixed-bed reactor. Langmuir-Hinshelwood kinetic models were developed to represent the observed kinetics in each reactor: these models were then compared. For the batch reactor, a rate expression based the dissociation of a chemisorbed CO intermediate being the rate-limiting step was found to be consistent with the experimental results. However, results from the fixed-bed suggested that the hydrogenation of an adsorbed C atom determined the rate of reaction. These differences in the kinetics on Ni/γ-Al2O3 between the fixed-bed and batch reactors suggest that a Langmuir approach using a single, rate-determining step may not be representative across all conversions. The rate over the Co/ZrO2 catalyst was characterised in the fixed-bed reactor over a range of reactant partial pressures at temperatures between 433 K and 503 K. The rate was observed to be dependent on hydrogen partial pressure and temperature, with the rate increasing with both. Previous research has reported a wide range of values of the apparent activation energy, with a study suggesting it was sensitive to pressure. Accordingly, the apparent activation energy was investigated for pressure sensitivity over a range of pressures between 5 and 15 barg: it was found to be constant. The values determined (~88-91±8 kJ/mol) were notably consistent with those reported for CO hydrogenation on cobalt. Kinetic schemes based on Langmuir-Hinshelwood and power law equations were evaluated, with the results best described by a reaction scheme based on the carbide pathway, with a rate-determining step of CH hydrogenation. A reaction-diffusion model of the Fischer-Tropsch process in a 2-D hollow cylinder was developed and analysed across a range of Thiele moduli and the extents of error in both effectiveness factor and selectivity were quantified relative to one-dimensional sphere and slab analogues. The errors between 2-D and 1-D analogues were found to be most significant between Thiele moduli of ~0.25 and ~3. Hollow cylinder effectiveness factors were bounded by those of sphere and slab above and below Thiele moduli of ~0.75 and ~1.15 respectively for the conditions examined, with the effectiveness factors exceeding those of both sphere and slab models between these moduli. A comparison of the hollow cylindrical pellets against spheres of equivalent volume demonstrated that hollow cylinders provided improved fixed-bed performance, with improved effectiveness factors and selectivities due to the lowered diffusion lengths of the hollow cylindrical geometry.

Submitted to School of Chemical and Metallurgical Engineering, Faculty of Engineering and the Built Environment, University of the Witwatersrand, Johannesburg, South Africa in Fulfillment of the Award of the Degree of MSc (ENG) in Chemical Engineering 25 October 2016 / The motivation behind this study is the need to manage and reduce wastes, in particular waste tyre and biomass, while in turn recovering energy from these carbonaceous materials. These wastes were gasified to produce synthetic gas which served as a feed to the Fischer-Tropsch Synthesis process to produce hydrocarbons. The formed hydrocarbons can be used as fuels for different purpose like transportation, domestic and industrial heating systems. Cobalt supported on zeolite catalysts are used because of their high acidic sites present in the zeolite that can break the Anderson-Schultz-Flory polymerization kinetics and also because cobalt-based catalysts are preferred for low temperature Fischer-Tropsch (LTFT) synthesis process due to their negligible water and carbon dioxide formation as well as stability and life span. In this research, a bi-functional Co/H-ZSM-5 catalyst was synthesized, characterized and evaluated for direct production of hydrocarbons at different process conditions. The bi-functional catalyst was prepared by incipient wetness impregnation method of an aqueous cobalt solution as the source of cobalt metal onto an H-ZSM-5 zeolite support, thereafter dried at 120 °C and calcined at 400 °C to obtain the finished Co/H-ZSM-5 catalyst. Physicochemical analyses performed included, Nitrogen Physisorption at 77 K to determine the surface area, pore volume and size of the synthesized catalyst. Also the N2 adsorption was used to determine the adsorptive properties of the catalyst. X-ray diffraction at 2θ region between 10 to 90 ° by using Co-Kα radiation (λ=1.79026 Å) was used to determine the material crystallinity, structure and composition. For the morphology and elemental composition of the catalyst, a Scanning Electron Microscopy coupled with an Energy Dispersive X-ray Spectroscopy was used. Thermal stability of the catalyst was checked using a Thermal Gravimetric Analyzer to determine how the catalyst degraded with time when temperature was increased uniformly. Reducibility of the catalyst was determined by using Temperature Programmed Reduction equipment in a hydrogen environment from room temperature to 900 °C. Transmission Electron Microscopy was used to check the catalyst morphology, and the dispersion of the metal-oxide particles within the catalyst support. The bi-functional zeolite supported catalyst was found to possess a surface area of 292 m2/g, pore volume of 0.18 cm3/g and pore size of 2.83 nm. The catalyst morphology was found to be irregular and aggregated-circular shape with a particle size of about 2.5 ± 0.5 μm. The embedded cobalt-oxide particles were obtained to be about 8 ± 3 nm located closer to the surface of the support and were reduced to metallic cobalt of 25% composition, at 330 °C in a hydrogen rich environment with an expected hydrogen consumption of 133 %. The process conditions under study involved flow rate, pressure and temperature and synthetic gas of different H2/CO ratio. The Synthetic gas mixture was purchased from Afrox and prepared in a way to mimic or simulate the syngas mixture expected from gasification of the waste tyre and biomass. However the study mainly focused on Hydrogen, Carbon Monoxide and Carbon dioxide as the dominant constituents of a waste tyre produced syngas. The bi-functional, Co/H-ZSM-5 performance evaluation was compared to commercial Co/SiO2 catalyst under similar conditions. The performance evaluation and comparison was made based on conversion and selectivity at different conditions. The process conditions considered were a flow rate of 1200, 2400 and 3600 GHSV (ml/gcat.hr), a pressure of 2, 8 and 15 bar, Low Temperature Fischer-Tropsch (LTFT) process at 220 and 250 °C was used, with a syngas composition that included H2/CO ratio of 1.5, 2.5 and 2.5 with 5 % of CO2 present in the reactant feed. The combination of 2 bar, 1200 GHSV and temperature of 220 °C and 1.5 of ratio was considered as low process conditions. While the combination of 15 bar, 1200 GHSV, 250 °C and ratio of 2.5 was considered as high process condition. Three pre-calibrated GCs (two online and one offline) were used to analyze the reaction products and the feed and the integrated peak-data analyses was captured by the use of a Data Apex Chromatograph software package known as Clarity ® (v. 2.5). The captured and analyzed data was used to calculate conversion and selectivity according to the methods reported in literature. With regard to the effect of process conditions, at low process conditions, the bi-functional catalyst, Co/H-ZSM-5, resulted in a 3 % CO conversion, while the commercial Co/SiO2 catalyst, resulted in 15 % of CO conversion. However the bi-functional catalyst was more selective to gasoline range products and 16 % selectivity to C5 hydrocarbons was obtained and 79 % to C6+, as compared to selectivities of 4 and 75 % for C5 and C6+ respectively, for Co/SiO2 catalyst. Also Co/SiO2 was found to be more selective to Olefins, the undesired products, with a selectivity of about 91 % to C6+ hydrocarbons as compared to a selectivity of 87 % for C6+ hydrocarbon obtained by using the bi-functional Co/H-ZSM-5 catalyst. Methane production was high for the Co/SiO2 catalyzed reaction, (about 13 % selectivity) with some quantity of water produced, as compared to 3 % methane selectivity for Co/H-ZSM-5 catalyst with no water produced during the reaction. At low process condition, both catalysts were less prone to middle distillates hydrocarbon production. At high process conditions, a CO conversion of about 54 and 68 % was obtained by Co/H-ZSM-5 and Co/SiO2 catalyst respectively. At these conditions the H-ZSM-5 supported catalyst was observed to produce more methane, about 53 % selectivity while for Co/SiO2 catalyst it was obtained to be 35 % selective to methane, with 66 and 7 % of C6+ olefin and paraffin selectivity respectively. Co/H-ZSM-5 offered 9 % selectivity to C6+ per olefin and paraffin hydrocarbons. The commercial catalyst showed an orderly manner of distributing products at these conditions while the bi-functional catalyst randomly distributed the formed products with a high selectivity to middle olefin distillates. In terms of CO2 co-feeding in the reactant feed, both CO and CO2 were hydrogenated to hydrocarbons. A CO conversion of about 73 % was obtained by Co/H-ZSM-5 catalyzed reaction while for Co/SiO2 catalyzed reaction a conversion of 70 % was obtained. About 63 and 75 % of CO2 conversion was obtained by H-ZSM-5 and SiO2 supported catalyst. These results were obtained at high process conditions. No change in paraffin selectivity was observed when comparing a state in which CO2 was present and absent, however olefin selectivity is significantly affected by the presence of CO2. Thus, an increase in olefin selectivity is observed with Co/SiO2, achieving 76 % of C6+ Olefin from 66 % and Co/H-ZSM-5 increasing middle olefin distillated from 25 to about 30 % of selectivity. Based on the performance evaluation the bi-functional catalyst was proven to yield higher hydrocarbons from a simulated waste-tyre synthetic gas with no requirement of downstream hydrocracking, since the bi-functional catalyst cut-off higher hydrocarbons due to its acidic sites. While the metallic sites of the catalyst, catalyzes the reaction of synthetic gas to hydrocarbons. This type of catalyst with both metallic sites and acidic sites is a hybrid-catalyst commonly known as bi-functional catalyst (Kang et al., 2014). At low process conditions the bi-functional Co/H-ZSM-5 catalyst is found to be more preferred while at higher process conditions the commercial catalyst was found to be more preferred, however in the presence of CO2 co-feeding, either catalyst can be used, but if water elimination is required the bi-functional catalyst is more suitable for the process. / MT2017

Fischer-Tropsch synthesis reaction to produce sulphur free hydrocarbons has enjoyed a resurgent in interests due to increases in world oil prices. In this work, the suitability of Mo and Mo-W carbides has been investigated as a possible cost-effective alternative to noble metals in Fischer-Tropsch synthesis. The molybdenum and tungsten monometallic and bimetallic carbides were prepared through precipitation from homogeneous solution to the sulphide followed by carburization with a mixture of propane and hydrogen to produce the resulting metal carbide. A 23 factorial design strategy was employed to investigate the effect of three carburizing variables, namely, time, temperature and gas ratio on the resulting catalyst. In particular, the effect of supports was also examined through four common semiconductor oxide supports, namely: Al2O3, SiO2, TiO2 and ZrO2. Thermogravimetric analysis of the carburization reactions showed that the conversion from metal sulphide to the metal carbides is a multistep process producing different phases of carbides, namely ??-MoC1-x, ??-Mo2C, ?? -WC1-x and ??-W2C, depending on heating rate and temperature. The rate determining step of the carburising reaction is the diffusion of carbon atoms into the metal matrix, hence giving relatively low activation energy values. Statistical analysis of the factorial design revealed that all three carburizing variables affect the final physiochemical makeup of the catalyst. SEM analysis showed that the carbides are well dispersed on the surface of the support and catalyst particles produced are nanoparticles in the range of 25 to 220 nm depending on the support. Fischer-Tropsch activity test showed that monometallic molybdenum carbide is active under Fischer-Tropsch conditions while tungsten carbide is inactive for the conditions studied in this project. However, bimetallic carbide catalyst, consisting of the two mentioned metals gave overall higher reaction rates and decreased methane selectivity. Steady state analysis revealed that there are two active sites on the surface of molybdenum carbide catalyst resulting in two chain growth propagation values when analysed via the Anderson-Schulz-Flory kinetics. Overall, ZrO2 support appeared to be the most suitable support followed by SiO2, TiO2 and Al2O3. Finally, kinetic modelling of data showed that methanation and higher hydrocarbons formation path occurs via combination of the oxygenated intermediate and Eley-Rideal mechanism.

In this thesis, a range of TiO2-based photocatalysts and cobalt-based Fischer-Tropsch (FT) catalysts were developed and synthesised via the one-step Flame Spray Pyrolysis(FSP). The work starts with the demonstration of bare TiO2 nanoparticles synthesis with controlled characteristics such as specific surface areas, crystallite sizes and anatase content. A comparative study was carried out by benchmarking with commercial Degussa P25 TiO2. The FSP TiO2 was shown to be more efficient in mineralising pollutants requiring direct charge transfer such as the saccharides, while P25 was better for mineralising alcoholic and aromatic compounds. Both catalysts were equally as active in mineralising carboxylic acids. Upon identifying the optimal synthesis of bare TiO2, an in situ co-precipitation of highly dispersed Pt on TiO2 was carried out in the flame. Deposition of Pt resulted in enhanced photocatalytic performance as a result of efficient charge trappings. It is highlighted here the inter-relationship between Pt oxidation states and the TiO2photocatalysis of carboxylic acid, alcohol and aromatic compounds. Depending on the mineralisation path adopted by the model organic compounds, they were shown to have direct influence on the Pt oxidation states. These oxidation states in turn affect the mineralisation rates of the organic compounds. Substitutional-doping of TiO2 with Fe(III) with tunable bandgap was also possible by FSP synthesis. The high temperature synthesis coupled with rapid quenching resulted in 5 times higher solubility limit (Fe/Ti = 0.05) than that previously reported in the literature. Under visible light irradiation, FSP-made Fe-TiO2 improved the photocatalytic mineralisation of oxalic acid by more than 6 times, with respect to P25 and FSP TiO2. Furthermore, the photocatalyst was reusable over a number of repetitions with minimal leaching or loss in activity. The last part of the work concerns the development of bare and Ru-doped Co-ZrO2 catalysts, where cobalt was finely dispersed within the zirconia matrix. Doping of Ru enhanced significantly the reducibility of cobalt, reducing even the embedded cobalt beneath the zirconia surface. It also increased the extent of CO-chemisorption and as such, enhanced the FT activity. This is the first time, catalysts of such type is synthesised and tested for FT reaction.