Although different approaches for the reduction of NOx species have been widely investigated in the literature, there are a number of drawbacks to the current industrial processes. In addition, increasingly stringent legislation ensures continued interest in improving these methods. An opportunity to utilise the H2 produced during biomass gasification to treat the subsequent exhaust emissions was identified and a number of novel catalyst systems were prepared and explored in a variety of different deNOx processes, and their performance compared to typical ‘standard’ catalysts. Measurements conducted on a pilot scale gasifier confirmed the presence of between 10 – 17 % H2, depending on the conditions in the gasifier, and validated the approach of this work. An experimental set–up consisting of a 15 mm i.d. stainless steel reactor housed within an electric furnace was constructed and commissioned, and an analytical method for the detection of eight potential deNOx reaction species was developed utilising a Quadrupole Mass Spectrometer (QMS). All of the catalysts explored in this study (o.d. = 14 mm, length = 10 mm) were prepared through impregnation of supplied samples of cordierite substrate monoliths coated with a γ–Al2O3 washcoat. Results of catalytic studies make up the bulk of this thesis, and two such processes were explored: (a) Selective Catalytic Reduction (SCR) – Pt/Al2O3 has already been widely explored for this process and Ag/Al2O3 was also prepared as a potential novel material. Although the Ag/Al2O3 catalyst displayed some conversion of NO at higher temperatures (e.g. 17 – 21 % at ≥ 350 ᵒC), apparent selectivity to the formation of N2 was poor (e.g. 40 – 41 %). (b) NOx Storage and Reduction (NSR) – Ba/Pt/Al2O3 is considered the standard catalyst for this process and was compared to two novel materials: Ba/Ag/Al2O3 and K/Ag/Al2O3. With respect to the standard catalyst, the novel systems demonstrated comparable, and in some cases improved performance depending on the NSR conditions at 400 ᵒC. Ba/Ag/Al2O3 demonstrated a capacity to store 11 – 15 % of the supplied NO and subsequently reduce 6 – 77 % of the stored species to N2. K/Ag/Al2O3 demonstrated a capacity to store 23 – 72 % of the supplied NO and subsequently reduce 60 – 92 % of the stored species to N2.

No description available.

The research described in this thesis was carried out as part of the EPSRC National Centre in Continuous Manufacturing and Crystallisation (CMAC), whose vision is to accelerate the adoption of continuous manufacturing processes for the production of high-value chemical products. Results focus upon the crystallisation of multi-component molecular complexes (MMCs), in particular co-crystals, an important class of potential functional molecular materials. MMCs are commonly discovered through small scale batch solvent evaporation, a method that provides little control over the crystallisation process, often yields only small quantities of crystalline material which may not be phase pure and is not generally scalable. Methods have been developed to help translate traditional small scale batch co-crystallisation of MMCs to selective continuous co-crystallisation processes. The co-crystal system urea-barbituric acid (UBA) has three polymorphs, previously identified in the literature and was selected as a candidate for transfer to continuous crystallisation as its polymorphism offers a development challenge in being able to achieve selectivity of solid products through by various crystallisation processes. Chapter 4 describes small scale crystallisation studies of UBA using evaporation, cooling and mechano-crystallisation techniques, demonstrating that UBAI is the more prevalent polymorph. Solid state and solubility analysis of the isolated UBA polymorphs are detailed; these data provide information on the energetic relationship between the polymorphs highlighting the similarity in energy of UBAI and UBAIII, and the metastable nature of elusive UBAII. In addition the polymorphic nature of the barbituric acid starting material has been investigated with samples from different suppliers proving to consist of different polymorphs. As well as being important in designing subsequent crystallisation processes, results from these studies corroborate work previously reported in the literature Chapter 5 describes how these results have been used in development of continuous crystallisation experiments using Mixed-Suspension Mixed-Product Removal (MSMPR) with a cascade of three stirred tank reactors, while Chapter 6 investigates flow crystallisation of UBA using the Continuous Oscillatory Baffled Crystalliser (COBC) and the Kinetically Regulated Automated Input tubular flow Crystalliser (KRAIC). The KRAIC was designed and constructed as part of this research. The continuous crystallisation experiments presented in Chapters 5 and 6 demonstrate the selective continuous co-crystallisation of UBAI across all three continuous crystallisation platforms. In Chapter 7 a range of novel co-crystalline MMCs is presented, with structural determinations, using single crystal X-ray diffraction, coupled with characterisation of properties and crystallisation characteristics; these studies emphasise the physical property advantages of MMCs and in some cases their potential for future translation to continuous crystallisation.

The current military respirator provides protection from contaminants using a cartridge packed with adsorbent activated carbon particles treated with metal salts to provide protection from toxic gases. However, the user of this respirator is subject to a physiological burden as a result. One component of this burden is the pressure drop, which makes breathing through the respirator filter difficult, with the burden becoming more severe at higher breathing rates. This project investigates the reduction of pressure drop and hence burden in respirator cartridges by using adsorbent hollow fibres. These are made up of adsorbent powder held together with a polymer binder to replace the conventional adsorbent particles. Adsorbent hollow fibres have a number of advantages, including lower pressure drop, the ability to operate in any orientation, no special filling requirements and customisability against emerging threats, such as toxic industrial chemicals. Dynamic challenges were performed using ammonia, hydrogen sulphide and cyclohexane as candidate gases, as each typifies a particular category of toxic industrial chemicals. Adsorbent hollow fibres were customised by treating with metal salts, metal organic frameworks and pore forming agents, and by replacing the hollow fibre polymer binder with a novel microporous polymer. In addition, the pressure drop of these adsorbent fibres was compared to granular beds. Pressure drop was then modelled using the Hagen-Poiseuille equation, and the breakthrough time was modelled with the Wheeler-Jonas equation, enabling the prediction of pressure drop and breakthrough time for new designs of adsorbent hollow fibre cartridges.

The application of multi-component crystallisation as a route to tuneable and switchable colour properties is investigated. Molecular complexes of haloanilines with nitro- and dinitro-substituted benzoic acid derivatives are presented, which enable determination of the crucial structural properties required for colour in these systems, in addition to demonstrating possible approaches to achieving tuneable and temperature-dependent colour. The findings offer a framework for future design and development of this class of functional materials. Two possible mechanisms for thermochromism have been established in the series of molecular complexes. The first mechanism is disorder-facilitated structural rearrangement; molecular disorder prevails in many of the molecular complexes, due to competition between donor and acceptor sites, and inherently results in an inefficient crystal packing in the structures. In some of the molecular complexes, the increased crystal space surrounding the molecules facilitates a structural rearrangement, which is accompanied by a change in colour, and allows the phase transition to occur in a single-crystal to single-crystal manner. The thermochromic transition temperature is tuneable through modification of the haloaniline molecule, including the reversibility of the transition. The second thermochromic mechanism is proton transfer, between the haloaniline and dinitrobenzoic acid components in metastable systems; the phase transition occurs with a distinct colour change but does not occur in a single-crystal to single-crystal manner. Similarly to the disorder-facilitated thermochromism, the transition temperature is tuneable through exchange of the halogen atom. One molecular complex combines the two thermochromic mechanisms, with the crystal space generated through disorder allowing proton transfer to occur with preservation of the single crystal. Proton transfer in the series of molecular complexes is key to the generation of both colour and switchable colour, with the transfer of a proton acting as a colour switch; a salt-cocrystal continuum region has been established for the series, based on ΔpKa values. The colour in the solid-state is induced through formation of a mixed stack arrangement of aromatic electron-donor and electron-acceptor molecules, which themselves are colourless or pale-coloured. Where at least one of the molecular components is in their neutral form, colour prevails, ranging between red and yellow; a number of possible stoichiometries and stacking types of the donor and acceptor have been observed. Tuneable colour has been achieved through simple modifications of the molecular components, through exchange of the halogen substituent or introduction of a methyl group, for example.

An experimental and theoretical investigation has been carried out into the evaporation of organic solvents (drying) from thin films of ink. By means of the infra-red technique, experimental drying curves have been recorded from stationary inked specimens during various conditions of forced convective drying. In the constant-rate period, the effect on the drying time of changes in air velocity, air temperature, coating thickness and percentage solvent in the ink, have been measured. Heat and mass transfer theory has been used to develop a criterion for the relative rate of drying of a range of solvents. Theoretical predictions of constant-rate drying time show good agreement with experimental measurements. In the falling-rate period, both the theoretical and experimental evidence appears to show that the rate of drying is limited by the rate at which solvent can diffuse through the polymer residue to the free surface. Analysis of experimental drying curves indicates that the diffusion coefficient is concentration dependent. To provide a theoretical solution of this process, numerical methods were used to solve the diffusion equation, with the diffusion coefficient exponentially dependent upon solvent concentration. The concept of the characteristic drying curve has been successfully used to correlate experimental drying rates in the falling-rate region. Non-dimensional results from the numerical solution of the diffusion equation have also been expressed in the form of a characteristic drying curve. Qualitative agreement between theory and experiment is good.

Local heat transfer coefficients were measured during the forced convective flow vaporization of a 60% v/v methyl- cyclohexane/40% v/v n-hexane mixture in a 48 Kw electrically heated horizontal furnace tube (15 feet long, 3/4 inch I.D.). Experimental single-phase heat transfer coefficients were a factor of 1.18 higher than those predicted by the Dittus-Boelter equation. In two-phase flow a large variation occurred in circumferential heat transfer coefficients (up to 250%) and large departures were observed from values predicted by the Chen correlation. For example, in the subcooled nucleate boiling regime, measured heat transfer coefficients were as much as three times the predicted values. Mandhane et al's flow pattern map only provided an indication of observed flow regimes. In an attempt to predict fouling a single-phase model was developed based upon a two step reaction in which all diffusional and kinetic resistances were included. Removal was considered to be by mass transfer and fluid shear. Considerable success was obtained in applying the model to Watkinson's (2) data on gas oils. The model was applied in a semi-quantitative manner to data on two-phase heat transfer coefficients to assist in the prediction of fouling in vaporizers. Even a slight variation in coefficients had a significant effect on fouling.

Extremum seeking controllers have been widely studied in recent years, but mainly with regard to systems having . stationary optima, and when moving extrema have been considered the investigations have frequently been purely theoretical. The present study investigates both experimentally and theoretically the possibility of improving the simple on-line optimizer which uses a sinusoidal perturbation signal to deal more adequately with a system which is being disturbed. The system used was a small experimental water gas shift reacter and the disturbances ranged from random to second order integrated. The method was based on Box and Jenkins' technique which combines a stochastic model of the disturbances and a dynamic model of the system to formulate a predictor controller. Difficulties encountered in applying this method to optimization arise mainly from non-linearities in the dynamics which are introduced by the optimizer, and to a lesser extent by the non-quadratic nature of the objective function. Two parameters, the amplitude and the frequency of the perturbation signal, were studied by means of a two by two factorial design, while the other parameters were maintained at best values found by Price and Kisiel. Because of difficulties in operating the experimental apparatus the practical results were not entirely satisfactory, although they did show the necessity of operating with the lower frequency perturbation signals. The simulation studies confirmed that better following was obtained with a slower perturbation signal, and in addition they showed that the ability of the predictor optimizer improved considerably over that of the simple optimizer as the frequency of the disturbances decreased.

Many industrial processes are shown to have a common underlying form involving transport of energy and matter by flow and dispersion or diffusion. Mathematical models using the Laplace transform and root-locus methods produce space parameter dependant transfer functions, and explain resonance like phenomena characteristic of distributed forcing of flow systems. Harmonic or functional analysis depends on having eigen functions for the differential operators for the system. The number of sections required for a lumped parameter model produced by spatial quantization was found (on a digital computer) to depend on the disturbances being considered and a parameter characterizing the system. A cheap, simple, special purpose electronic analogue was developed. Control design by conventional methods yields a useful standard of comparison. The absolutely optimal solutions from the calculus of variations (etc.) are shown to present major computational difficulties especially when the theory is extended to partial differential and integral equations. Practical use of sub-optimal control design methods and the analytical development of a direct feedback controller all depend on having a state-space of low dimensionality. A correlation coefficient criterion for instrumentation gives a method for specifying instrumentation for protection purposes but not for control or performance measure. Control based on instantaneous computation on a measure of state is shown to need only small amounts of instrumentation but sensitivity to parameter changes has to be taken into account. Spatially distributed control can deal with disturbances arising anywhere in the system, and sensitivity to parameter changes is reduced at the cost of greater complexity The structure of the control scheme and its instrumentation is largely determined by the spatial location of the measure or measures used for performance assessment and the relationship between spatial displacement and time delays in the distributed systems.

Surface patterning is important for a range of engineering applications, including controlled wetting and spreading of liquids, adhesion and assembly of smart coatings. There is therefore a need of simple, cost-effective and scalable techniques for pattern formation over a wide range of scales. Conventional methods fail to comply with these requirements, as costs and complexity increase in the attempt to impress nm to µm scale features. By contrast, wrinkling of bi-(multi- )layers is inherently inexpensive, scalable and robust, and has the potential for soft matter patterning from the nano- to the macro-scale. This work investigates the controlled multi-layer generation of polydimethylsiloxane (PDMS) glassy skins via surface oxidation using plasma exposure and/or ultraviolet ozonolysis (UVO). Uniaxial mechanical compression is then employed to induce pattern formation via a well-known wrinkling instability. Topographies with wavelengths down to 45 nm are achieved, for the first time, as well as features with characteristic lengthscales of ∼ 10s µm. Moreover, simple design routes for double frequency nested pattern formation, imposed by compression of tri-layer laminate films, are established. The work concludes by exploiting wrinkling as a method for the mechanical characterisation of thin drying films. A time-resolved wrinkling interrogation during film drying is established as a simple and reliable approach to determining evolving mechanical properties of films, overcoming the difficulties associated with handling very thin free-standing films and the limited sensitivity of conventional methods, with potential applications extending to coatings, personal care items, and foods.