The use of cold atmospheric pressure plasma for applications related to microbial decontamination has grown enormously over the last decade. Non-thermal plasmas generated in ambient air contain a wide variety of reactive oxygen and nitrogen species, or RONS. When such species interact with microorganisms they induce a number of biological changes, ultimately resulting in inactivation of the organism. This thesis focuses on the design, development, optimisation and application of air plasma systems for microbial decontamination. The aim of the work is to gain a better understanding of how RONS are produced in air plasma and how they are transported through different phases of matter, including gases and liquids. It is shown that RONS generation is highly dependent on the discharge conditions and two distinct modes of operation are observed. Downstream of the discharge, the transport of RONS to the sample region is of paramount importance as many highly-reactive species are lost. To address this challenge, the structure of the plasma generating electrodes was systematically studied to optimise the plasma generated air flow and therefore the transport of species downstream. Optimised electrode structures were shown to generate flow velocities in excess of 1m/s which is an order of magnitude improvement over transport by diffusion alone. Using the optimised plasma system, the impact of RONS in real decontamination scenarios linked to food and water security were considered. This included investigation of plasma decontamination of liquid samples, solid surfaces and tissues. It was shown that plasma decontamination can be extremely effective but many factors influence the efficacy of the approach. Microorganisms shielded within a liquid layer or by a complex surface morphology were shown to be particularly difficult to inactivate. Overall, this work has demonstrated that plasma can be a highly effective tool for microbial decontamination but careful consideration of both the discharge parameters and the sample properties is required to achieve the highest level of decontamination.

Wheat is the most popularly grown cereal in the world. Over 400 million tonnes of wheat is milled to flour annually for inclusion in baked products. Flour milling has evolved into an efficient operation; however, process modelling, simulation and optimisation can make significant improvements to the operation. One such improvement is the use of more informative quality testing procedures to supply information about milling potential of a sample. Breakage matrices were used to describe the breakage function that relates the inlet and outlet particle size distributions over 151 Break. Successful prediction of the milled stock particle size distribution for the native (unsegregated) wheat sample showed that wheat grains milled independently during roller milling; this was confirmed using high-speed video. The milling ratio (roll gap over grain thickness) was identified as the parameter determining breakage and enabled a range of roll gaps and inlet particle sizes to be considered simultaneously. The variation in cumulative undersize, P(x), with milling ratio was described by a quadratic function whose coefficients varied in a quadratic manner with particle size (x). Nine breakage coefficients were used to describe the breakage of a wheat variety over a range of roll gaps and inlet particle thickness. The particle size distribution was predicted using the breakage equation and shown to be linear in the range 250-2000 Ilm for 21 different wheats. The slope of the particle size distribution was related to wheat hardness and varied with milling ratio. This variation was shown to indicate the sensitivity of the variety to changes in the milling ratio. Flour millers generally prefer varieties with a high sensitivity to changes in milling ratio. Rank Hovis, sponsors of this project, mill approximately 1.6 million tonnes of wheat per year. The results of this study provide the foundation for savings in milling efficiency (through increased flour extraction), wheat substitution (using lower priced wheat in place of high priced wheat). It is estimated that these measures could save over £1m per annum.

This thesis presents an experimental and computational fluid dynamics (CFD) study of spray-drying, spray-freezing and spray-freeze-drying of whey proteins. The effects of varying feed concentration (20 - 40% w/v) and outlet temperature (600 - 120 degrees Celcius) on whey protein denaturation (determined by DSC) and solubility at pH 4.6 (by Kjeldhal and RP-HPLC methods) have been investigated in a pilot-scale cocurrent spray dryer. The study confirms that low outlet gas temperatures (60 and 80 degrees Celcius) produce the lowest amount of denaturation, with almost complete denaturation observed at 120 degrees Celcius. Slightly more denaturation was found with a 40% feed concentration. A reversed phase HPLC technique has been applied to measure the loss of solubility of a-lactalbumin and P-lactoglobulin. Significantly higher losses in solubility were observed for P-lactoglobulin compared to alactalburnin. Increasing the feed concentratioq at higher outlet temperatures also caused noticeable increases in insolubility. The reversed phase HPLC results were consistent with those obtained from total protein nitrogen content (Kjeldhal) analyses. This comparative study suggests that the protein solubility can also be calculated from RP-HPLC technique. Spray-freeze drying is an alternative approach to spray drying, which is less likely to cause protein denaturation and loss of solubility. Conventional freeze-drying involves high capital and operating costs, due to the low temperatures, high vacuum and very long drying times. One solution to this problem is to reduce the dimensions of the material being dried. This is the basis of the spray-freeze-drying technique, involving atomisation of a liquid to form droplets, freezing the droplets and subliming off the ice at low temperature and atmospheric pressure in a fluidised bed. However, the quantities of gas required for atmospheric freeze-drying are prohibitively expensive. A pilot-scale spray-freeze-drying process was investigated, in which fluidisation was performed at sub-atmospheric pressures, allowing rapid freezedrying (in about one hour) but using much less gas. This was demonstrated using whey protein which yields a product with a highly porous structure, with little loss of protein solubility. This process has potential to produce high-value-added food and pharmaceutical products more quickly and cheaply than is currently possible by commercial vacuum freeze-drying processes. CFD simulations were developed for short and tall form spray-dryers to study the particle velocity, temperature and residence time during drying. These simulation results agreed well with the published experimental data. The tall-form spray dryer model predictions showed that more particles impact on the cylindrical wall position and this may affect the protein denaturation and solubility. This study suggests that a short form dryer with a bottom outlet is more suitable for drying of proteins. Similarly, a CFD simulation for the spray-freezing operation was developed to study the gas flow pattern and particle trajectories and histories. This CFD model also includes the latent heat effects during the phase change. The simulation predictions agreed reasonably well with the experimental results. A comparison of simulations for a solid and a hollow cone spray suggested that the latter yields lower particle temperatures with low particle collection efficiency. A modified chamber geometry was proposed and the simulation showed that the new design could achieve higher particle collection efficiencies.

A critical analysis has been made of raw beef to establish the nature of the components which particulate in the thermal reactions resulting in cooked beef odour. Extraction of beef with water or aqueous alcohol gave an extract which, when heated, produced a cooked beef odour whereas the extracted residue did not. The water soluble material was fractionated into broad groups of constituants by dialysis, gel filtration and ion exchange chromatography, and the various fractions assessed for their cooked beef odour potential. The more interesting fractions were analysed qualitatively by thin layer chromatography and quantitatively by ion exchange chromatography, colourimetry, U.V. spectrometry, and gas-liquid chromatography. The water soluble high molecular weight fraction alone did not give cooked beef odour when heated whereas the water soluble low molecular weight material did, Furthermore the components responsible for this aroma were widely distributed in the fractions obtained by gel filtration. The application of the results to experiments with model system showed that the precursors of cooked beef odour should include a source of hydrogen sulphide as well as the relevant amino acids and sugars. The relative weakness of cooked beef odour given by the water soluble material and its fractions, as compared with whole beef, may be related to the virtual absence of the hydrogen sulphide precursor. As a result of this observation and the evidence obtained from a study of model systems it is felt that sulphur may be provided largely by the meat protein.

The objective of this thesis was to study the effect of heat treatment and glycation on five industrially important functional properties of soy protein, namely solubility, emulsifying ability, water holding ability, acid gelation ability in soy yogurts and the heat stability of soy protein emulsions. These objectives were achieved completing three tasks: 1) The physicochemical properties were studied of Soycomil K, a commercially available, insoluble soy protein concentrate. Differential scanning calorimetry (DSC) analysis showed that it was 28% denatured compared to native, laboratory SPC. Further analysis showed that the aggregated structure is more hydrophobic and there are more disulfide bonds than found in commercially available soy protein isolates (SPI). Heat treatment at alkaline pH and low protein concentration increased its solubility. Heat treatment at 100 °C increased SoyComil’s solubility significantly compared to heat treatment at 70 °C. Glycation of SoyComil K with glucose at 70 °C increased solubility compared to the control, whereas glycation at 100 °C decreased solubility. Heat treatment of Soycomil increased its solubility more than glycation. 2) A soy yogurt with smooth texture and high water holding capacity (WHC) was developed with SPI heated in the presence of pectin and glucose, followed by homogenization with the oil ingredient of the yogurt recipe. The texture matched that of commercially available yoghurt manufactured with soymilk. A yogurt made with SPI only showed low WHC. These results provide evidence that combined heat treatment and glycation of SPI improved its functional properties. A study of the structure of the yogurt showed that the majority of bonds were hydrophobic bonds, whilst electrostatic and disulfide bonds played a small part in maintaining the yogurt structure. 3) A study of the rate of aggregation of SPI stabilized emulsions at pH4.5 showed that the heat stability of SPI emulsions was strongly dependent on protein concentration and temperature. A second study showed that the presence of polysaccharides either improved or had little effect on heat stability depending on the concentration of polysaccharide added. This was attributed to interactions between soy proteins and specific polysaccharides. A study of the heat stability of mixed whey protein concentrate (WPC) and SPI emulsions showed that WPC dominated the oil droplet destabilization behaviour, and that low proportions of WPC were able to slow down the heat-induced breakdown of SPI/WPCemulsions containing a high proportion of SPI.

A series of distillery process waters was characterised by pyrolysis-GC/MS for humic substances derived from soils and vegetation in water in conjunction with dissolved organic carbon analysis. Additional analyses were carried out for phenolic compounds by HPLC, anions and cations by ion chromatography, and metals and colour by spectrophotometry. Over 70 organic compounds were identified, with waters from north and western areas of Scotland containing higher levels of peat-derived compounds, with organic carbon determined by the length of watercourse traversed. Ionic concentration of waters was related to the geology of the surrounding areas. Little variation was found in other measurements of water chemistry. Using these samples for mashing, spirits were produced with notable differences in character, particularly in terms of spirit ‘heaviness’. Chemical analysis of these spirits by GC/MS highlighted significant differences in levels of volatile higher alcohol and sulphur-containing compounds, although these did not correlate with sensory character. The influence of organic and inorganic constituents of water in isolation was considered by preparing a series of artificially-spiked waters produced for use in mashing, representing the range of variation present in industrial process waters. The presence of peat-derived compounds in mashing water caused increased ester and reduced higher alcohol production by yeast, whereas the presence of metal ions inhibited yeast activity. The resulting new-make spirits from ionic-supplemented waters showed minimal differences in sensory character, whereas those containing peat-derived compounds exhibited greater variation, with heavier, more complex spirit character.

Standard analysis of malt flour can mask the heterogeneity of hydrolytic enzyme activity. Kernel heterogeneity can lead to brewhouse problems and a product with unpredictable nitrogen and fermentable sugar content. The variability between individual grains of important malting characteristics was measured in malt samples produced under aerobic and anaerobic conditions. Various parameters (including germinative energy, germinative capacity, moisture content, β-glucanase activity, friability and homogeneity) were measured to ensure that the 5 d aerobic Optic malt, produced in Heriot-Watt university micromaltings, was viable and of commercial quality. The 5 d aerobic malt kernels produced at Simpsons Maltings in Berwick-Upon- Tweed were heavier than the micromalt. Commercially produced malt kernels had higher levels of fermentable sugars and soluble nitrogen than the micromalt despite lower α-amylase, β-amylase and ‘total’ limit dextrinase activity. Differences between the 5 d aerobic micromalt and the 5 d aerobic commercially produced malt are indicative of why micromalting cannot always be used as a model system for what is happening industrially and must be modelled on commercial practice. Subjecting 5 d aerobic micromalt to 24 h anaerobic incubation resulted in increased levels of fermentable sugars per l wort. 24 h anoxia also resulted in increased α-amylase and limit dextrinase activities. There are potential industrial applications for this anaerobic wort. Limit dextrinase inhibitor protein present in crude extract prepared from mature barley, eluted from a gel filtration column at a higher molecular weight than expected. The limit dextrinase inhibitor protein either aggregates or binds to other proteins in a high molecular weight complex.

Lactic acid bacteria (LAB) comprise a major part of the natural microflora of Scotch malt whisky fermentations, due to their tolerance of heat and elevated ethanol concentrations. In this study, their effects on the organoleptic properties of the spirit were investigated. Samples from late (>70 h) fermentations were obtained from whisky distilleries throughout Scotland. Bacteria of varying colony morphologies were isolated, purified, and characterised initially using random amplification of polymorphic DNA – polymerase chain reaction (RAPD-PCR). Isolates with differing RAPD patterns were retained and their ability to produce 10-hydroxystearic acid (10- HSA) from oleic acid was determined qualitatively using high performance thin layer chromatography. 10-HSA is the primary precursor of γ-dodecalactone, which is an important flavour compound in malt whisky responsible for the desirable “sweet and fatty” characteristic of the spirit. Thirty-nine isolates had strong or weak bioconversion activity while 89 isolates displayed negligible or no activity. Forty-two strains, largely from the former category were identified using partial 16S rRNA gene sequences. Lactobacillus paracasei was the predominant organism but L. brevis and L. plantarum were also identified. These 42 strains were assessed for their bioconversion capacity in a semi-quantitative manner using gas chromatography – mass spectrometry (GC-MS) and five isolates, comprising L. brevis, two strains of L. paracasei, and two strains of L. plantarum were selected for further study. These isolates were used in laboratory–scale, simulated whisky fermentations with Saccharomyces cerevisiae. Fermentation liquor (wash) was distilled to produce new-make spirit, which was analysed organoleptically by quantitative descriptive analysis. Spirit from fermentations inoculated with L. brevis had an enhanced “sweet” character, probably due to the higher γ-lactone levels detected in this whisky, as well as increased “sulfury” and “meaty” notes, most likely due to yeast autolysis. L. paracasei enhanced the “green/grassy” notes of new-make spirit, while also adding a “sour” aroma probably resulting from the elevated levels of lactic acid detected in the wash. Like L. paracasei, L. plantarum increased the “green/grassy” notes of new-make spirit. Further fermentations were carried out in which L. brevis, one strain of L. paracasei, and one strain of L. plantarum were inoculated into fermentations with yeast comprising 90% S. cerevisiae and 10% Torulaspora delbrueckii, which had been iii isolated previously from Scotch whisky fermentations and shown to enhance the concentration of γ-lactones in new-make spirit. Co-fermentation of L. brevis with S. cerevisiae and T. delbrueckii resulted in a spirit with increased “green/grassy”, “sweet”, and “oily” notes, with decreases in “sulfury” and “meaty” observed when the wild yeast was not present. Spirit derived from co-fermentations of L. paracasei and T. delbrueckii exhibited increased “soapy”, “sour”, and “sulfury” notes. Cofermentation of L. plantarum and T. delbrueckii caused increases in “green/grassy”, “soapy”, “sweet”, “sour”, and “sulfury” notes. Increased concentrations of γ-lactones were detected in new-make spirit distilled from fermentations inoculated with L. brevis, presumably contributing to the enhanced sweet character of this spirit. This effect was further amplified by the inclusion of T. delbrueckii in the laboratory–scale fermentations.

The impact of colour adjustment on the flavour stability of a portfolio of locallybrewed pale lager beers with a range of colouring agents such as specialty malts, roasted barley, colouring beer and artificial caramel colorant was investigated. All brewing control parameters and beer specifications were defined and monitored under a rigorous regime in order to avoid processing factors that might interfere with or modify the two parameters under investigation. The colour appearance parameters of the beer samples at distinct ages (fresh, forced aged and 12 month-aged) were psychophysically assessed by means of sensory viewing method (magnitude estimation) by an expert panel of the Colour Imaging Group at the Department of Colour Science, University of Leeds. Likewise, the aforementioned samples were physically measured by tele-spectroradiometry and digital imaging system at two different environments. Significant differences between the beer samples at sameand distinct ageing conditions were detected in terms of lightness, colourfulness, hue angle, opacity and clarity, although all of the samples were colour-adjusted to the same colour units according to conventional procedures (EBC colour units). In addition, good agreement between the sensory viewing (magnitude estimation) method and telespectroradiometry was observed. In contrast, some discrepancies between the aforementioned methodologies and the digital imaging technology were detected. Flavour stability was assessed by the detection and quantification of fifteen flavour-active beer ageing compounds (10 aldehydes and 5 non-aldehydes compounds) by GC-MS using headspace-solid phase microextraction (HS-SPME) with on-fibre PFBOA [O-(2,3,4,5,6-pentafluorobenzyl)hydroxylamine] derivatisation for aldehydes compounds and on-fibre DVB-CAR-PDMS [divinylbenzene-carboxenpolydimethylsiloxane 50/30_m] derivatisation for non-aldehyde compounds. The results were correlated with the determination of the endogenous anti-oxidative potential (EAP) of the beer samples by electron spin resonance (ESR) using N-tert-butyl-α-(4- pyridyl)nitrone N’-oxide (POBN) and the sensory assessments provided by the I.C.B.D. sensory panel. Additionally, the quantification of organic radicals of the specialty malts, the roasted barley (whole intact kernel and milling fraction measurement) and the artificial caramel colorant were conducted by ESR. Based on the results of this holistic approach, a colouring agent was selected for improving the flavour stability of pale lagers according to its physicochemical-, sensorial and psychophysical effects as colour appearance.

Forecasting packaging material demand is crucial to effective future planning by capital intensive material manufacturers. Yet several considerations introduce great uncertainty over the future packaging mix. Foremost is a highly heterogeneous and dynamic end-use marketplace subject to multitudinous technological, economic, consumer and legislative change-forces. These act at all levels upon a diverse and complex supply chain that suffers data paucity and, hence, opacity of cause and effect. A wide range of future-oriented decision technologies was examined to meet these challenges. None promise competitive advantage over commissioned forecasts of aggregate demand. At the sector level, the petfood market is relatively homogeneous and simple, but nonetheless significant. Neural network analysis of its causal relationships led to rich results and a simple, workable causal-forecasting model. Data paucity inspired three key development paths. First, a weakness in current implementations of genetic algorithm model input selection was exposed - result variability due to training data set division. Novel software invoked genetic algorithm input selection over exhaustive permutations of training cases to generate a result distribution, thereby partially automating model specification for wider application. Second, the neural networks were implemented in a scenario-planning spreadsheet, to isolate the more certain and less certain factors in scenario forecasting. Third, several unprecedented factors change past relationships and can undermine even the most accurately specific model. Accordingly, a Delphi survey was conducted to develop scenarios of the potential impact of remote retailing upon packaging demand. Consequently, although the five-year outlook for tinplate petfood packaging is open to interpretation, the most likely scenario is stable demand. Petfood and human food cans exhibit clear strengths in the remote-retailing scenario, but high uncertainty is envisaged for the remaining packaging applications. Such unprecedented forces should be continually monitored, and marketing activities should emphasise the strengths of tinplate in the scenarios thereby envisaged.