This thesis is comprised of studies in the characterisation of monolayer-protected and metal cluster of the structural response of size-selected (bare) clusters to chemical reactions. The technique employed is high-angle annular dark field (HAADF) aberration-corrected scanning transmission electron microscopy (AC-STEM). The effect of chemical reactions on size-selected metal clusters was investigated. The clusters under investigation were imaged with AC-STEM and their structure was assigned by comparing the atomic resolution images with a set of multi-slice STEM image simulation atlases. The effect of vapour-phase 1-pentyne hydrogenation on size-selected Aux (x=923 and 2057) cluster was studied and it was found that the gold nanoclusters demonstrate high stability in both size distribution and structure under the reaction. On the contrary, size-selected Pdx (x=923 and 2057) clusters tended to transform from amorphous to high symmetry structures under the same reaction condition. The gas-phase CO oxidation reaction on size-selected Aux (x=561, 923 and 2057) cluster was studied with regard to cluster size distribution and atomic structure. It was found that under the same conditions of the CO oxidation reaction, two different kind of ripening modes could be identified depended on the cluster size. Smoluchowski ripening, in which clusters diffuse intact and coalescence, is found to occur for Au2057 in the CO oxidation reaction. Ostwald ripening, in which larger clusters grow at the expense of smaller ones, was found to occur for Au561 and Au923 clusters, due to the extra energy generated from catalytic CO oxidation reaction.

This submission for the degree of Doctor of Science describes work initiated, led or contributed to by Dr Graham Machin, a graduate of the University of Birmingham, over the last 25 years. The submission comprises of six main parts, illustrating the different contributions made to the area of temperature measurement. In brief these are: • Development and improvement of non-contact thermometry standards • Comparisons of non-contact thermometry scales, including the development of new methods of comparison • Contributions to the development of high temperature fixed points as new temperature references above the freezing point of copper (1 084.62 °C) • Development of methods for the mitigation of temperature sensor drift at high temperatures • Improvements and innovations in medical thermometry - in a variety of settings • Leadership in international temperature measurement research The thesis begins with an introduction to the subject of the research, including a brief overview of the presented papers. Then a list of the presented papers is given followed by copies of the manuscripts. The thesis ends with an Appendix containing an up to date publication record.

Liquids are the primary medium for chemical and biological reactions. Phase transitions can give rise to phase separation through nucleation or spinodal decomposition. This phase separation may be the reason behind structure in many liquids, which then has an effect on more practical processes. Mesoscopic (1 nm-1 μm) structuring of the liquid in the neighbourhood of a second critical point associated with a liquid-liquid transition (LLT) can cause anisotropic diffusion, which greatly affects the wetting properties of the liquid. It may also segregate reactants and products, affecting reaction rates and outcomes. Some examples of processes, which may be affected by this, include the synthesis and preparation of pharmaceutical drugs (which is sensitive to liquid structure, phase separation, and nucleation), preparation of nanometre scale materials, electrochemistry, dye-based solar cells, separation technology, and heterogeneous catalysis. This thesis concentrates on advancing of the understanding of some of the phase transitions and mesophases occurring in liquids. The first two chapters are introductory and the remaining chapters describe original research. Chapter 3 describes our studies on crystal nucleation in the presence of a liquid-liquid phase separation. The subjects of these studies were two binary liquid systems: nitrobenzene-hexane and water-1,2-trans-dichloroethylene. Our discovery of template-less asymmetric water crystals (water fluff), which we were able to reproduce repeatedly, strongly suggests that, despite common belief, high driving force does not necessarily lead to ill-defined crystalline forms. We also propose that crystal templating induced by liquid-liquid phase separation might be a general phenomenon, not limited to water. In Chapter 4 we explore one of the most interesting and challenging postulates in the field of liquid science, that is the existence of more than one isotropic liquid states in a single-component liquid at constant temperature and an LLT between these different states. By employing a number of experimental techniques, some of them not previously used to study this subject, we provide a new evidence for the existence of the LLT in triphenyl phosphite (TPP). Furthermore we answer an important and fundamental question of what is the order parameter characterising the LLT in TPP. Finally, Chapter 5 shows how our investigations into the existence of the LLT in n-butanol led to the conclusion that the observed phenomenon is in fact a transition between a supercooled isotropic liquid and a liquid-crystalline (LC) state. The LC is in general considered a form of matter “in between” the liquid and the crystal. However, the LC phase in n-butanol geometrically frustrated the formation of the stable crystalline phase. This frustrated phase can be seen as a template for similar ordering in other molecular liquids and is likely to be essential in their supercooling and LLTs.

The (1-x)BiFeO3-(x) (K0.5Bi0.5)TiO3 (BFO-KBT) system has been investigated with respect to composition and temperature. Powder samples have been synthesised by the traditional solid state method at regular intervals across the compositional range. High resolution powder x-ray diffraction measurements have been made on these powders between room temperature and up to 950ºC. The data from these measurements have been analysed by the Rietveld method, and the results of these refinements have been used to construct a phase diagram of the BFO-KBT system. It was found that there are no sharp transitions in the phase diagram, either with composition or with respect to temperature, and that aside from the BFO end member, all samples investigated were best fitted with a mixed phase model. At low mol% KBT, it was found that the mixed phase model that best described the system was a mix of rhombohedral R3c and cubic Pm3m. As the mol% KBT was increased, the cubic phase increased, becoming dominant between 15% KBT and 20% KBT. The rhombohedral phase diminished with increasing mol% KBT and was no longer a component in the mixed phase system beyond 60% KBT. At 70% KBT, the room temperature system could still not be modelled with a single phase. A mix of cubic Pm3m and monoclinic P1m1 was used to model the system. At 90% KBT, the model was again found to be best fitted by a different mix of phases, specifically a mix of monoclinic P1m1 and tetragonal P4mm phases were used to model the data. The tetragonal phase was found to become more dominant from this point with increasing mol% KBT, but even the KBT end member was found to be best modelled with a mix of monoclinic P1m1 and tetragonal P4mm phases. It was found that all samples become more cubic with increasing temperature. It was found that the temperature at which the change to a cubic state occurred decreased with increasing mol% KBT. It was also found that at high mol% KBT and high temperature, there was phase separation into a two-cubic mixed phase model instead of the expected single Pm3m cubic phase. This phase separation has been linked to the reported morphotropic phase boundary in the material. The BFO-KBT system was found to change little with respect to temperature below 500ºC, which makes it an attractive material for high temperature device applications if doped with a material with a stronger piezoelectric response. In addition, single crystal x-ray diffraction of BFO-KBT crystals has been undertaken with laboratory-based and synchrotron systems. It was found that the latter was necessary to obtain good results from a refinement of the data in SHELXL due to the absorption of the BFO-KBT system, minimised through the use of a much higher energy x-ray beam. It was found in JANA2006 that the BFO-KBT system was best fit with anharmonic atomic displacement parameters, which was linked to the difference between the short range and long range order in the material.

In this study, molecular dynamics (MD) simulations have been performed to study rheological properties of liquid alkanes under a range of pressures up to several GPa and high shear rates. The liquid alkane considered in this study is 2,4- dicyclohexyl-2-methylpentane (DCMP), which is a highly viscous fluid. Two further molecular motifs, octadecane (Linear) and 1,6-dicyclohexyl-hexane (Dumb- bell) were chosen as comparison. The rheological properties of DCMP under high pressure were studied using molecular dynamics simulations. A wide range of pressure (1 atm ≤ P ≤ 10 000 atm) and shear (5.47 x 10 7 s-1 ≤ y ≤ 5.47 x 10 11 s-1) conditions have been considered. Simulation parameters have been carefully chosen from preliminary simulations. MD allows access to understanding not always available experimentally. The present simulation results confirm that the density of all three motifs increase as pressure increases. The results also show that the mean squared dis- placements of molecules decrease as pressure increases. At pressures higher than 3000 atm, the movement almost ceases for DCMP and Dumbbell implying a solid- like behaviour at very high pressures. The viscosity of DCMP is higher than that of Dumbbell, but comparable to Linear. As molecules tend to adopt compact shapes at high pressures, this affects their rheological properties accordingly. The viscosity of Dumbbell is found to be the lowest among the three. Viscosities in- crease with pressure for all three molecules with larger changes with DCMP and Linear. It is found that the pressure of the systems increase when the shear is applied. The viscosity of DCMP and Dumbbell increase as pressure increases. The viscosity of DCMP is found to be higher than that of Dumbbell, especially in the high pressure region. DCMP and Dumbbell show the shear thinning behaviour. The start of a plateau is observed for DCMP and the viscosity in the plateau is at the same order of magnitude as the zero shear viscosity estimated from the Stokes-Einstein relation for low pressure systems.

Two-dimensional Fourier transform ion cyclotron resonance mass spectrometry (2D FT-ICR MS) is a data independent tandem mass spectrometry technique that allows direct correlation between precursor and fragment ions without the necessity of any sort of isolation prior to fragmentation. Two-dimensional mass spectrometry (2D MS) experiments were conceived in the 1980s from inspiration of two-dimensional nuclear magnetic resonance (2D-NMR), but the development of the technique was stopped because of the insufficient computational capabilities of that time. In 2010, with the progress in computer technology and the improvements in FT-ICR MS instrumentation and methods, the interest in developing 2D MS was renewed. Since then, 2D FT-ICR MS showed to be a valuable platform for the analysis of small molecules, macromolecules such as peptides and glycopeptides, and complex mixtures deriving from tryptic digestion of proteins of increasing sizes. All the proteomics studies performed with 2D MS used the bottom-up (BUP) approach. The main goal of this Ph.D. research work was to develop two-dimensional mass spectrometry for the challenges of contemporary mass spectrometry, therefore pushing the limits of the technique towards the most interesting, cutting-edge research grounds, by optimising the technique, developing 2D MS analysis methods, and finally constructing improved instrumentation act to better perform 2D MS experiments. On this purpose, the rising importance of top-down proteomics (TDP) led to the development of 2D FT-ICR MS for the analysis of intact proteins. To realise this first project, 2D MS was optimised for TDP analysis using calmodulin (CaM) as a model, and infrared multiphotondissociation (IRMPD) as fragmentation technique. The study compared the two-dimensional BUP and TDP analysis of CaM with standard one-dimensional tandem mass spectrometry, and showed that the use of 2D MS allowed to obtain comparable cleavage coverages, with a consistent saving in sample and time. The study affirmed the suitability of 2D MS for top-down proteomics, still leaving space for more developments allowing its optimisation. The total cleavage coverage of CaM was later implemented with the use of BUP 2D electron-capture dissociation (ECD) MS. In order to improve the cleavage coverages obtained in top-down two-dimensional mass spectrometry, another technique was developed, called MS/2D MS. The technique makes use of an extra fragmentation step before two-dimensional mass spectrometry, achieved through external isolation of a charge state of interest, and collisional fragmentation in a hexapole-based collision cell before the ions are sent to the ICR-cell. The output of MS/2D MS is a single 2D mass spectrum with information equivalent to MS3 about the entire mixture. The technique has been successfully developed using ubiquitin (Ubi) and calmodulin as models, and showed new uses of the features exclusive of 2D MS, such as the use of precursor ion scans to assign the protein terminus of many precursors ions at once, and the analysis of internal fragments formed by the first dissociation. Furthermore, as long as extra fragmentation steps can be performed, a deeper investigation of the analytes can be achieved, in an experiment collectively called MSn /2D MS. After the development of 2D MS for TDP, another method was developed, implementing the use of two-dimensional mass spectrometry for another type of complex mixtures: polymers. An extensive 2D MS study was performed on a simple mixture of homopolymers such as PMMA, and on two complex mixtures commonly used as drug excipients, polysorbate 80 and TPGS. The study demonstrated again that single 2D MS experiments give results equivalent to multiple MS/MS experiments targeted at the different precursors. Furthermore, it allowed the high resolution tandem mass spectrometry of polysorbate 80, representing the first kind of study of that type because of the challenges of isolating single analytes from the polymer mixture without the use of a previous separation, such as chromatography. The use of neutral-loss lines was particularly useful in this study, allowing to discriminate between oleated, linoleated, and non-esterified species. Finally, 2D mass spectra of polymers showed characteristic diagonal lines, specific for each polymeric mixture, representing some sort of fingerprint of the mixture. The last part of the thesis is dedicated to the optimisation of 2D FT-ICR MS for ECD MS, the application of the technique to another class of compounds, glycans and glycan-derivatives, and finally to the development of a more capacious ICR-cell. The research performed in this thesis showed that 2D FT-ICR MS is a viable data independent mass spectrometry technique for top-down proteomics and polymer analysis; that the technique is fast, and it possesses unique features that improve ion assignments, such as precursor ion scans and neutral-loss lines. Finally, it opens for two-dimensional mass spectrometry the path to glycomics and to an improved 2D MS analysis of even more complex mixtures, thanks to the development of a novel ICR-cell.

The behaviour of a catalyst is intrinsically linked to its structure and, therefore, accurate structural refinements are desired to tune their overall function. Higher functional demands on catalysts systems, require more complex disordered materials which are inherently difficult to characterise with conventional analytical techniques. Solid state NMR is an excellent probe of local order and, hence, is utilised in this thesis for the structural determination of several catalytic related materials. The first direct105 Pd solid state NMR measurements of diamagnetic (K2PdCl6, (NH4)2PdCl6 and K2PdBr6) complexes is reported, thereby introducing an effective 105 Pd chemical shift ranges with respect to the newly proposed 105 Pd chemical shift reference (0.33 M H2PdCl6(aq)). The enormous 105 Pd quadrupolar moment, makes the interaction very sensitive to small structural distortions as demonstrated by the measurable quadrupolar parameters for the three complexes, despite the high symmetry octahedral Pd coordination. The detected deviation from a cubic symmetry, was corroborated by XRD PDF analysis. The 105 Pd quadrupolar parameters are shown to be more sensitive to minute disorder than conventional XRD and other quadrupolar nuclei NMR. Ambient temperature 105 Pd NMR observations of Pd metal determined the Knight shift as K = −3.205 ± 0.006 %, where variations in the 105 Pd Knight shift allowed for detection of defects in the cubic metal structure and for differentiation of Pd nanoparticle sizes. The developed 105 Pd NMR methodology was then applied in a multi-technique structural investigation of doped Pd catalysts, that confirmed the interstitial location of the dopants. The use of the newly developed structure-generation software, supercell, in combination with GIPAW-DFT calculations and solid state NMR, is shown to be a thorough tool for structural determination of disordered materials. The methodology is applied to two phases of the aluminosilicate mullite (3:2 mullite and 2:1 mullite), and provides complete assignment of the 17O,27 Al and 29 Si NMR spectra. The distribution of AlO4/SiO4 sites in the mullite structure is shown to be random, proving the presence of SiO4 moieties in the tritetrahedral (T3O) environments. The observation of said moieties directly contradicts Loewenstein’s avoidance principle. Additionally, a quasi-tetrahedral site with an additional long bond ((Al/Si)O4+1) is discovered and a vacancy adjacent three-coordinated Al site (AlO3[]) is proposed. The findings from this investigation are then applied to the 27 Al MQMAS study of boron doped mullites, providing additional evidence for the AlO3[]motif. A thorough 11 B and 27 Al solid state NMR investigation was undertaken on three series of aluminium borate phases (A9B2, A2B and metastable Al6-xBxO9 (where 1 ≤ x ≤ 3)) with varying Al/B ratios. A solid solution of AlO4 and BO4 tetrahedra was discovered in all three phases (to varying extents), justifying the conflicting compositional/structural reports present in the literature. Differences in the crystallinity of commercially available, sol-gel synthesized, and solid state synthesised A9B2 samples were documented. The solid state NMR study of the disordered phases, A2B and metastable Al6-xBxO9, utilised multiple fields and MQMAS measurements to constrain the simulation of the 1D spectra, which allowed for complete assignment of the structure and corrected previous erroneous reports. An AlO4+1 site, analogous to the discovered mullite environment, is found in both disordered phases.

It was known highly dispersed Co (HDC) catalysts for Fischer-Tropsch applications could be synthesised from either Co metal or CoCO3 precursors. Both synthesis routes were known to precipitate Co oxide nanoparticles from solution onto a support via Co(III) ammine complexes. The synthesis pathway was not known. Using solid state NMR and single crystal X-ray diffraction (XRD) a suite of Co(III) complexes have been characterised. The HDC catalyst synthesis from Co metal produced the Co(III) complex which was identified as [(Co(NH3)3)2(μ-OH)2(μ-CO3)](CO3)∙5H2O via single crystal XRD. The 59 Co solid state NMR then confirmed the precipitant from the CoCO3 route was the same complex. The characterisation of a suite of Co(III) complexes of differing oxo coordination has been obtained. It was observed that each oxo coordinating ligand increased the NMR chemical shift by ~800 ppm. For a single oxo coordination a disparity of ~1000 ppm was also observed due to effects from the anion, crystallographic water and type of oxo coordination. Spin-spin coupling with the 59 Co nucleus was observed to occur for bidentate carbonate and ammine ligands yielded ǀ2 J( 59 Co, 13 C)ǀ of 1-10 Hz and ǀ1 J( 59 Co, 15 N)ǀ of 56-75 Hz. The bonding of monodentate and bridging carbonate ligands was observed to be ionic in nature lacking the spin-spin coupling. It was hypothesised Co oxide nanoparticles precipitated from Werner’s complex, [Co((μ-OH)2Co(NH3)4)3](SO4)3∙9H2O. This was observed in the solid state NMR studies at 14518 ppm. From the time resolved 59 Co solution state NMR measurements of the HDC catalyst syntheses a resonance was observed at 14300 ppm. This was assigned to the central Co oxo coordinated environment of Werner’s complex. In conjunction with observations of chemical shifts coinciding with known precursors to Werner’s complex, it has been possible to suggest a reaction pathway. It was concluded that [Co(NH3)5CO3]+ and [Co(NH3)4CO3]+ eventually become Werner’s complex before precipitating out of solution as Co oxide nanoparticles. The study of Pt and Ni hydrogenation catalyst precursors on α-alumina, γ-alumina, silica and titania have been undertaken. From the Pt materials, it was observed oligomeric Pt oxide structures deposited on the supports. The 195 Pt NMR measurements of these oligomeric structures showed disorder due to a mix of hydroxyl and O bridges in addition to preferential orientations. These orientations were thought to arise from the formation of sheet like structures. All measurements were compared to PtO2, the 195 Pt measurement of which produced an axial symmetric lineshape due to chemical shift anisotropy yielding a Ω of 3916(100) ppm, κ of -1.00(8) and a δiso of 4607(100) ppm. Finally, 27 Al measurements of the γ-alumina support showed reactions with nitric acid occurred. This formed an aluminium nitrate species on the surface which was removed with heat treatment. It was also concluded AlO(V) sites were preferential binding points for the Pt oxide. The AlO(V) 27 Al NMR resonance was seen to disappear with heat treatment, it has been proposed AlO(V) converts to AlO(VI) with the addition of Al–O– Pt bonds.

This thesis details the chemical synthesis and nanocharacterisation of magnetic nanoparticles (MNPs). Throughout this thesis, solution-based synthesis methods are used to fabricate MNPs with a variety of shapes, from spheres and cubes to wires and octapods. Chapter 3 assesses the role of polyol solvent in autoclave-based magnetite NP synthesis. It was found that solvents containing functional groups (especially primary alcohols) afforded the greatest control over final MNP morphology and that the presence of additional alkyl substituents could disrupt the packing of surfactant molecules around a particle surface, giving rise to more complex ‘compound’ MNPs. Magnetic measurements show the particles to be superparamagnetic, with saturation magnetisation values close to that of the bulk. In Chapter 4 It was found that by replacing the polyol solvents used in Chapter 3 with a large excess of bulky surfactant molecules it was possible to form high aspect ratio lanthanide oxide (LnOx) nanowires and ribbons. It was found that the nanowires formed via an intriguing 3-stage ‘oriented assembly’ mechanism, in which individual NPs form, before aligning anisotropically and recrystallising into a more crystallographically homogeneous product. The magnetic properties of Gd2O3 nanowires and Dy2O3 nanoribbons are very similar to those of discrete particles, implying that no long-range ordering exists in the nanowires. Optical characterisation of the Eu2O3 nanowires showed that there is an increase in fluorescence lifetime going from bulk Eu2O3 to the nanoscale. In Chapter 5 the high temperature reflux approach is used in the synthesis of manganese oxide (as MnO) particles, using manganese-containing polynuclear carboxylate complexes as precursors. It was found that control over MnO NP size and shape is dependent on many factors, such as the carboxylate ligand present in the precursor (which can disrupt particle stabilisation by surfactant molecules) and the heating regime, which ii determines if the particles grow under thermodynamic or kinetic growth regimes. The magnetic behaviour of MnO MNPs was found to be dependent on the relative number of surface spins present in the particles. In Chapter 6 new routes to gadolinium-doped magnetite particles are discussed. It was found that decomposition of a single-source precursor yielded doped magnetite particles which had a gadolinium content of 2–4%. Magnetic characterisation of these particles showed them to be superparamagnetic, with a reduced saturation magnetisation compared to bulk magnetite. Preliminary investigations showed that the gadolinium doped particles (after being rendered water soluble by a ligand exchange reaction) were readily taken up by human fibroblast cells and exhibited low toxicities.

This thesis focuses on understanding radiation da mage processes du e to a-decay events in xenotime (YPO.). Novel atomistic potentials were derived using a reverse Monte Carlo fitting routine. These potentials were validated using morphological predictions , showing good agreement with natural samples. Intrinsic defects were modelled and binding energies were calculated to investigate their clustering behaviour. Solution energies were calculated for the substitution of lanthanide elements into the lattice, which were found to b e relatively unfavourable. Molecular dynamics simulations were utilised to model damage cascades within xenotime and zircon. Clear differences were observed in the distributions of defects and the polymerised tetrahedral units, which may contribute the improved radiation resistance of xenotime over zircon. Threshold displacement energies were calculated for xenotime utilising the Fibonacci lattice. These showed good agreement with literature sources and may be of use in future experimental studies. Finally swift heavy ion irradiated samples of xenotime were investigated using the X-ray pair distribution function (PDF). A fission track simulation was performed to help inform the experiment, however the damage could not be characterized. The ordering of strontium atoms in mixed fluorapatite samples were also investigate using X-ray PDF, however alterative modelling techniques may be required to analyse this data.