In recent years inorganic sunscreens have become much used especially for fair skinned people and children where high sun protection factors are desirable. The major pigment used is titanium dioxide and the other pigment used is zinc oxide. For cosmetic acceptability these are used as ultra-fine dispersions and re-coated in various ways to minimise unwanted photoactivity. However, methods of assessing the photoactivity of such inorganic sunscreen dispersions are far from satisfactory and those that are available only assess the photo-oxidative capacity of the sunscreen. In this work new tests have been developed which assess the photo-reductive capacity of Ti02 and ZnO. These tests have been established so that they use simple equipment and are suitable for use in an industrial based quality control laboratory.

The development of new porous materials for use in applications such as gas storage and separation processes, catalysis, catalysts supports and the removal of environmentally unfriendly species has increased rapidly over the past decade. Research into the development of these new materials has been dominated by metal organic frameworks, covalent organic frameworks, nanoporous polymers and, most recently, porous organic cage molecules. This thesis describes adsorption studies of a metal organic framework, Zn (TBAPy) and a porous tetrahedral organic cage molecule of ~ 1 nm diameter formed by the condensation reaction of 1,3,5- triformylbenzene with 1,2-ethylenediamine. The development of metal organic frameworks has traditionally involved the formation of rigid network structures, analogous to that of zeolites. More recently the focus has shifted to those of dynamic, flexible framework materials, and the response of these materials to adsorption of gases and vapours. The metal organic framework Zn (TBAPy) is based on a zinc metal centre functionalised with benzoate fragments. The initial two-dimensional structure undergoes rearrangement of the paddlewheel units to form a 3D framework, Zn (TBAPy)' upon desolvation. The ability of this 3D network to separate p-xylene and m-xylene was investigated. It was found that these isomers produced different effects on the framework, with p-xylene producing a typical Type I isotherm, whereas m-xylene induced a structural change within the material, with a much slower rate of m-xylene adsorption at higher pressures. This could potentially lead to the equilibrium separation of these two isomers by the metal organic framework Zn (TBAPy)'. The 1 nm diameter tetrahedral cage molecules formed by the condensation reaction of 1,3,5-triformylbenzene with 1,2-ethylenediamine can exist in a number of stable polymorphs, Cage 1α, Cage 1β and Cage 1γ. These polymorphs can be interconverted by exposure to certain organic vapours/solvents. The conversion of Cage 1β to Cage 1α by adsorption of probe molecules ethyl acetate, 2-butanone, diethyl ether, pentane and methanol was studied. Adsorption of ethyl acetate, 2- butanone and diethyl ether produced unusual adsorption isotherms, which included desorption of adsorbed vapour with increasing pressure during the adsorption isotherms. This desorption is attributed to the structural change from Cage 1β to Cage 1α. The unusual desorption step is not observed for methanol or pentane adsorption. The adsorption of methyl acetate was studied over a wide temperature range in order to assess the thermodynamic and kinetic characteristics of the unusual desorption step. The adsorption of dichloromethane showed the reverse transformation of Cage 1α to Cage 1β, showing that the inter conversion produces stable polymorphs. The kinetics of the structural transformation followed an Avrami model and the mechanism is an activated process. Cage 1α has voids between the cages, which are connected by very narrow constrictions that allow the kinetic molecular sieving of oxygen, carbon dioxide and nitrogen. It was found that oxygen adsorbs approximately ten times faster than nitrogen on Cage 1α, with selectivity and rate constants similar to those observed for carbon molecular sieves. The thermodynamics and kinetic results are discussed in terms of structural characteristics and diffusion into molecular cage materials. The kinetic molecular sieving is not present in the polymorph Cage 1β, which has wider pores.

An important issue in low temperature polymer electrolyte membrane fuel cells (PEMFC) is the lack of alternative catalysts to platinum for the oxygen reduction reaction (ORR). The high cost and potential limited availability of platinum restricts its long term use for large scale applications in PEMFC. Consequently, there is a great interest in alternative catalysts to platinum for PEMFC. In this research a systematic study of the synthesis and optimization of carbon-supported palladium and palladium alloy nanoparticle electrocatalysts is reported. The catalysts investigated were Pd, Pd-Au, Pd- Co, Pd-Fe and Pd-Ti supported on carbon black (Vulcan XC-72R). At least two different atomic metal to metal ratios for bimetallic catalysts were investigated. All catalysts were initially evaluated for the ORR by voltammetry in a three-electrode cell. Different reducing agents, including hydrogen, ethylene glycol (EG), formaldehyde and sodium borohydride were used for the synthesis of Pd nanoparticles. The use of EG led to Pd nanoparticles with the highest ORR activity; this synthetic method was optimised by adjusting the pH of the system. Pd nanoparticles of approximately 6 nm diameter dispersed on carbon black with exchange current densities for the ORR of ca. 1.0 x 10-11 A cm-2 were obtained. Two synthetic procedures were chosen for the preparation of bimetallic catalysts: simultaneous co-deposition of both metals on the carbon support and deposition of the second metal on carbon-supported Pd. Pd-Co alloy with atomic ratio Pd:Co 4:1 exhibited improved ORR activity compared to Pd/C after being heat treated at 300 ºC under H flow. The effect of heat treatment under H flow on 22 the ORR activity and physicochemical properties was also studied. Pure Pd particles exhibited sintering after heat treatment; the presence of Au, Co and Fe decreased the degree of sintering and the presence of Ti did not affect Pd particle growth. Pd and Pd-Co were evaluated in low temperature hydrogen PEMFC, and Pd was tested as cathode catalysts in hydrogen polybenzimidazole (PBI) based high temperature PEMFC, and in direct methanol fuel cells (DMFC). Optimized Pd and Pd-Co catalysts were tested in a hydrogen low temperature PEMFC and the results were compared to those of the state of the art commercial Pt catalyst. With approximately 1.7 times higher metal loading than Pt (still significantly lower cost) the fuel cell with the Pd cathode gave better performance than that with Pt operating with air at 40 ºC. A comparative study of Pd and Pt was carried out in DMFC using different methanol concentrations and under different operating conditions. At methanol concentrations of 5 M and higher, the Pd cathode based cell performed better than that with Pt at 60 ºC with air. A pseudo one dimensional model for Nafion® -based low temperature hydrogen PEMFC was developed to simulate the influence of cathode catalyst, metal loading, electrode thickness and different operating conditions on the cell voltage and current density output. The model considered mass transport through a thin film electrolyte and through porous media but not gas flow along the channels of the cell. The model closely predicted experimental results at 20 and 40 ºC. Above 40 ºC cell performance did not improve experimentally as was predicted by the model; this lack of improvement was attributed to the decrease of oxygen permeability through Nafion® caused by the lower humidity at higher temperatures. Predicted results showed that enhanced fuel cell performance in the whole current density range could be achieved by increasing metal loading in the cathode whilst maintaining the catalyst layer thickness, which could be practically achieved by increasing the metal content of the carbon-supported catalyst.

Foams are ubiquitous in many industrial processes and therefore have been subjected to extensive research in the past few decades. However, a number of fundamental issues relating to foams remain poorly understood including the production and control of bubbles and how they contribute to product structure, appearance, and rheology. The work in this thesis thus sets out to study the foam rheological behaviour and its related structural evolution, to explore a potential method in controlling bubble structure during foam generation, and to introduce a novel non-invasive technique in probing foam microstructure. Foam rheological behaviour was studied by using two approaches. In a rheometric approach, the foam response to external shear and the related microstructural change, its time dependency, thixotropic behaviour and viscoelastic properties were investigated for both detergent (Gillette) and protein (egg white protein) foams. The flow curves constructed were well fitted with four-parameter rheological models, the Cross and Carreau models and revealed the true flow behaviour of foam; there was no yield stress as reported in the literature. On the other hand, the macroscopic flow behaviour of a protein foam (WPC) in straight pipes and through fittings (sudden expansion, orifice plates and perforated plates) were studied. A friction factor analysis approach which treats foam as a non-Newtonian power law fluid was successfully used to model the flow of WPC foam, this has a practical significance in that pressure drop can be calculated using a constant friction factor along a pipe of a constant cross section in any flow regime. The potential of ultrasound in controlling foam structure during the foam generation process was investigated. Depending on the types of surfactant and concentrations, ultrasound reduced the mean bubble size of foams and improved foam homogeneity. Single bubble formation in an ultrasound field managed to shed light onto the possible mechanism involved; ultrasound increased the bubble formation rate and reduced the bubble coalesce events at the nozzle. X-ray micro-computed tomography was used successfully in revealing the true three dimensional structure of some cellular/foam food products, both solid and semi liquid. Both two- and three-dimensional quantitative analyses of structural parameters were done and the spatial bubble size distribution was reconstructed using a stereology technique.

In this thesis, the kinetic issues in heterogeneous catalysis was widely investigated and developed except chapter 3 as an independent project. The micro-kinetics simulation was employed to investigate the activity as well as the selectivity in heterogeneous catalysis. Then we noticed the adsorbate-adsorbate interaction which could impose significant effect on the kinetic behaviour and was focused to building a bridge to connect the adsorbate-adsorbate interaction with the catalytic activity. In this situation, we first brought out a new micro-kinetics model to include the adsorbateadsorbate interaction on surface and em:ploy it in the reaction of I -butene to 2-butene in chapter 5. Moreover, we have frn1her studied the general trend of coverage effect on the BEP relation as well as the volcano curve which are considered as the key building blocks in theoretical calculation in heterogeneous catalysis. In the last chapter, we created a novel Kinetic Monte Carlo Simulation method that can address the lateral interaction to study the mysterious 'synergic effect' for bimetallic catalyst. It is admitted that coverage effect is a complex and widely existed issue while our method is qualitatively rather than quantitatively to address the impact. However, with our results, the understanding of the general trend as well as the nature behind the coverage effect is more clear and easy to understand.

The catalytic conversion of hydrocarbons over zeolites has been applied in large scale petroleum-refining processes. However, there is always formation and retention of heavy by-products, called coke, which causes catalyst deactivation. This deactivation is due to the poisoning of the acid sites and/or pore blockage. The formation of coke on hydrocarbon processing catalysts is of considerable technological and economic importance and a great deal of work has been carried out to this study. The main aim of this work is to understand the deactivation of zeolite catalysts as a result of coke deposition. The deactivation by coke of USHY zeolite was investigated during catalytic conversion of hydrocarbons – 1-pentene, n-heptane and ethylbenzene – as representatives of olefins, paraffins and aromatics respectively, at different reaction temperatures, time-on-streams and composition. Three novel techniques, coke classification, thermogravimetric method for characterising coke precursors and indirect temperature programmed desorption (TPD) for catalyst acid sites characterisation were developed to further study catalyst deactivation mechanism. Product distribution, coke formation, characterisation of coke precursors, as well as the role of strong acid sites on hydrocarbon reactions are presented and discussed. During catalytic reactions of 1-pentene over USHY zeolite, cracking and hydride transfer were the predominant reactions in initial stage which deactivated rapidly allowing isomerisation to become the main reaction afterwards. Deactivation studies showed that coke formation was very strong initially which is in good correlation with the initial rapid deactivation. The hydrogen freed during this initial time from the formation of high C/H ratio coke components contributed to the formation of hydride transfer products. The amount of coke precursors decrease with increasing reaction temperature due to the higher desorption of coke precursors into gas phase while hard coke amount increased with temperature as expected from an activated process. The coke amount formed was not proportional to the reactant feed composition, because of a strong pseudo-zeroth- order initial coking on strong acidic sites. The thermogravimetric method provides insight into the chemical character of coke precursor components in terms of the mode of their removal and allows further classification of coke precursors into small and large coke precursors. The concentration and strength of acid sites of coked catalysts were studied by the TPD methodology. Besides, characterisation of coke precursors was also revealed. The initial deactivation preferentially on strong acid sites is very fast. The concentration of free acid sites is inversely correlated well with the total concentration of coke rather than individual coke groups. Coke precursors tend to be more stable at higher reaction temperatures. Furthermore, by selectively poisoning strong acid sites of USHY zeolite, it shows conclusively that strong acid sites are responsible for cracking and hydride transfer reactions as well as strong coke formation while weak acid sites can only catalyse double bond isomerisation.

This thesis describes the results of a series of investigations examining the operational performance followed by the fundamental re-development of two analytical instruments, namely a Pneumatic Spring Particle Sizer (PSPS) and handheld Spring Particle Sizer (handheld SPS) for size distribution analysis of dry powders in the 100 – 2000 μm size range. Each instrument shares the same basic principle of operation involving the use of a closed coil helical extension spring, which is partly filled with the test powder. Particle size distribution data is obtained by stretching the spring to known lengths and measuring the mass of the discharged particles from the spring’s coils. In the case of the Handheld SPS, aimed at on the spot quality control applications, the test particles are discharged from the spring using manual shaking. In the case of the PSPS on the other hand, the particles are discharged using pulsating pressurised air. The design, development and evaluation of two methods for the in-situ measurement of the sample mass within the PSPS are discussed. These include a full-bridge strain gauge assembly and the investigation of the correlation between the minimum fluidisation velocity and the mass of the test sample within the spring. The strain gauge proved to be a successful method producing a mass resolution of ± 1 % for a total sample mass of 280 g. The second method was unsuccessful as it was found that the minimum fluidisation velocity, in most cases did not follow a clear trend with mass. Detailed investigations are conducted aimed at understanding the processes governing the mass discharge rate from the PSPS and hence the sample analysis time by studying the particle migration behaviour. A pulsating fluidised bed of similar dimensions to the spring is used to mimic the behaviour in the spring. Tests involved pulsating fluidisation followed by particle size distribution analysis at equal distances along the length of the bed containing poly-dispersed and mono-dispersed particles. It was observed that any operating or design parameter that promoted the degree of mixing, for example, increasing the fluidising air pulse frequency would reduce the test analysis time. The analysis time also increased with the sample poly-dispersity. In an attempt to reduce the sampling time, the handheld SPS was rotated using a variable speed tumbler as an alternative to manual shaking. Despite the marked reduction in the sampling time, this method resulted in the discharge of particles larger than spring coil openings thereby producing erroneous results. Calibration experiments for the same types of powders revealed a linear relationship between the discharge sample volume and its mass, independent of the particle size in the range 212 – 1000 μm. This allows in-situ measurement of the discharge sample mass in the handheld unit by reading the sample volume collected in the integrated graduated collection cylinder and reference to a previously generated calibration line.

There have always been problems regarding scale up of a process from laboratory scale to commercial scale plant. In general, scale up is considered in terms of process performance, while safety and environmental issues are considered based on the full scale design. The relation between Safety and Environmental Risk and the size of plant is an important consideration in design but one that is often considered only indirectly. The magnitude of hazards change with scale in ways that depend on their nature, as well as the response time of equipment, process inventories, changes in the ability to control etc. The IMPULSE project aims are to deploy innovative process equipment such as microreactors, thin-film devices and other structured components to attain step-change performance enhancement for whole processes, including intensification, thereby contributing to significant improvements in supply-chain sustainability. One theme pursued is the numbering up (rather than scale-up) of processing devices, so that in IMPULSE the size of commercial devices is the same as in the lab. By this means it is expected to have fewer problems regarding performance change on scale up. Also, it has been widely claimed that process intensification leads to safer manufacture. There is a need of methodology which helps industries to choose the best technology. In this study, CTNT (Comparison of Traditional versus New technology) method is developed. First a list of hazards is provided. The main hazards are fire and explosion, runaway reaction, frequency of leaks, releases of hazardous and toxic material and overpressure. Each of these hazards is quantified for both plant and a comparison is done to see how hazard changes with scale of plant. This methodology is demonstrated for a particular reaction but it is more generally applicable. It seems regardless of running IC plant longer hours compared to batch plant, the hazard declines. The results for pool fires show that the diameter of pool reduces by a factor of 3 by moving from batch to IC for 10000 kg product per hour in hydrogenation unit if the pressure kept at 3 bar by increasing the pressure to lObar the diameter of pool increases by a factor of 2. The probability of death decreases by increasing the distance from the failure point. The Pc (explosion overpressure) in IC plant increases by increasing the scale from 10 kg product per hour to 10000 kg product per hour at 9 bar to 30 bar respectively (considering the distance to failure point 3 m). By increasing the pressure to lObar in IC plant the explosion overpressure increases to 20 bar for 10 kg product per hour. The probability of death increases to 48% and 40 % for batch and IC plant respectively for 10000 kg product per hour with r ==0.5m. It is found by reducing the size of pipe the frequency of leak increases. This sounds reasonable as the frequency of a given hole size is lower in a larger pipe. In general larger pipes tend to be designed and protected more carefully, and tend to have thicker. walls. In part this is a compensation for the fact that the consequences of a given hole size will be greater if the inventory is larger. Conventionally, it is thought that smaller is safer. It is claimed that microstructured plant can be located in a small room. But this study shows that for the hydrogenation example the concentration of hazardous/toxic material is what limits the size of the space required. It is found that in order to get the same concentration of hazardous /toxic material in IC plant as batch the volume of room can be as small as 25 m3 compared to batch at 210m3 It was also found that it is hard to protect the individual microreactors against overpressure.

Thin film flows are of central importance to a variety of industrial and scientific processes such as heat and mass transfer, micro- and nanotechnological settings such a lab-on-a-chip devices, and geophysical flows. Understanding their behaviour is thus of central importance. While they have been subjected to a great deal of analysis, there are still regimes in which this behaviour is only partially understood, especially in highly nonlinear free surface flows, and particularly in the presence of other competing physical effects. This dissertation focuses on the behaviour of nonlinear flows in the presence of electric fields. By introducing a potential difference between the surface electrode on which the fluid rests and a second electrode above the fluid, an interfacial stress is induced due to the disparity in electrical properties between the liquid and gas media. It is found via direct numerical computations in both linear and nonlinear regimes that for a broad range of parameters the induced disturbances have wavelengths which are large compared to the thickness of the film layer. This may be exploited in the guise of the lubrication approximation in order to reduce the Navier-Stokes equations to inordinately more tractable forms; these `low-order models' form the core of the study of this thesis. Particular attention is paid to studying these films in a variety of geometries. It turns out that the introduction of curvature to a substrate renders the asymptotic analyses somewhat more delicate than is the case for planar films. The generalised analysis presented herein is applied to several different particular situations, including geometries where the surface is curved in the streamwise or spanwise direction. The resultant models incorporate the effects of capillarity, viscous stress, electrostatically-induced Maxwell stress, and inertia, as well as interfacial charge transport effects. It is demonstrated via comparison to direct numerical computations that this low-order modelling is highly accurate in a wide array of circumstances, where particular attention is given to flows on planes and on the surface of fibres. It is demonstrated that disturbances of particular wavelengths can be either excited or suppressed in a spatiotemporally-specific manner. This gives an enormous degree of control with many practical ramifications. Perhaps most significantly it is demonstrated that this method allows one to overcome the formerly ubquitous assumption that low-order models were confined to use for thin film flows: it is now possible to model `thick' flows. This has only been verified in a small set of geometries, but stands to have significant ramifications as the method could be used to improve any existing model in the literature that is predicated on the thin-film assumption.

In industry, solvents are an important processing aid. They often make up the bulk medium in liquid-phase reactions and have a large impact on process performance, from reaction rate to environmental impact. Traditionally solvents are chosen based on heuristics, experience and knowledge of the reaction. Recently systematic methodologies for choosing a solvent which can maximise the rate of a reaction have started to emerge (R. Gani, C. Jiménez-González and D.J. Constable, Comp. Chem. Eng., 29, 1661-1676, 2005 and M. Folić, C. S. Adjiman and E. N. Pistikopolous, Am. Inst. Chem. Eng. J., 53 (5), 1240-1256, 2007). In this work, a systematic experimental protocol for gaining insight into a specific reaction by monitoring the kinetics of this reaction in a variety of solvents using in situ 1H NMR spectroscopy and obtaining the reaction rate constants by nonlinear parameter estimation is presented. The methodology has been applied to an SN2 reaction, namely a Menschutkin reaction, in which a tertiary amine is added to a primary haloalkane to form a quaternary ammonium salt. This reaction has been chosen because it is a classic reaction to study solvent effects as this type of reaction is very sensitive to solvents and there is evidence that an empirical model of solvent effects can be developed. Reaction rate constants have been successfully determined for 10 deuterated solvents at ambient temperature (25 °C) and pressure (1 atm) using in situ 1H NMR spectroscopy. The reaction rate constants have been found to vary over two orders of magnitude depending on the solvent used. This systematic methodology has also been applied to the alkylation reaction of sodium β-naphthoxide and benzyl bromide which forms the products 1-benzyl-2-naphthyl ether and 1-benzyl-2-naphthol. It is known from the literature that depending upon the solvent used the ratio of these products vary accordingly (N. Kornblum, R. Selzer and P. Haberfield, J. Am. Chem. Soc., 85 (8) 1148-1154, 1963). In highly dipolar aprotic solvents such as DMSO and DMF the product favoured will be 1-benzyl-2-naphthyl with percentage yields as high as 95 % and 97 % respectively, whereas in polar protic solvents for example, water and 2,2,2-trifluroethanol (solvents that are capable of forming hydrogen bonds) the other product 1-benzyl-2-naphthol is more favoured in yields as high as 84 % and 85 % respectively. The kinetics and selectivity of this reaction have been studied in 9 solvents by in situ 1H NMR spectroscopy at ambient temperature (25 °C) and pressure (1 atm). This experimental study shows the strong variation of selectivity and the differences in the reaction rates, which are all dependent on the solvent.