Cellulose from energy crops or agriculture residues can be utilized as a sustainable energy resource to produce biofuels such as ethanol. The process of converting cellulose into solvents and biofuels requires the saccharification of cellulose into soluble, fermentable sugars. However, challenges to cellulosic biofuel production include increasing the activity of cellulose-degrading enzymes (cellulases) and increasing solvent (ethanol) yield while minimizing the co-production of organic acids. This work applies novel surface analysis techniques and fermentation reactor perturbations to quantify, manipulate, and model enzymatic and metabolic processes critical to the efficient production of cellulosic biofuels. Surface analysis techniques utilizing cellulose thin film as the model substrate are developed to quantify the kinetics of cellulose degradation by cellulase as well as the interactions with cellulase at the interfacial level. Quartz Crystal Microbalance with Dissipation (QCM-D) is utilized to monitor the change in mass of model cellulose thin films cast. The time-dependent frequency response of the QCM simultaneously measures both enzyme adsorption and hydrolysis of the cellulose thin film by fungal cellulases, in which a significant reduction in the extent of hydrolysis can be observed with increasing cellobiose concentrations. A mechanistic enzyme reaction scheme is successfully applied to the QCM frequency response for the first time, describing adsorption/desorption and hydrolysis events of the enzyme, inhibitor, and enzyme/inhibitor complexes. The effect of fungal cellulase concentration on hydrolysis is tested using the QCM frequency response of cellulose thin films. Atomic Force Microscopy (AFM) is also applied for the first time to the whole cell cellulases of the bacterium C. thermocellum, where the effect of temperature on hydrolysis activity is quantified. Fermentation of soluble sugars to desirable products requires the optimization of product yield and selectivity of the cellulolytic bacterium, Clostridium thermocellum. Metabolic tools to map the phenotype toward desirable solvent production are developed through environmental perturbation. A significant change in product selectivity toward ethanol production is achieved with exogenous hydrogen and the addition of hydrogenase inhibitors (e.g. methyl viologen). These results demonstrate compensatory product formation in which the shift in metabolic activity can be achieved through environmental perturbation without permanent change in the organism’s genome.

THE EFFECTS OF CERIA ADDITION ON AGING AND SULFATION OF LEAN NOx TRAPS FOR STAND ALONE AND LNT-SCR APPLICATIONS Model powder and fully formulated monolithic lean NOx trap (LNT) catalysts were used to investigate the effect of ceria on desulfation behavior. Temperature-programmed reduction (TPR) experiments (model catalysts) showed each of the oxide phases present is able to store sulfur and possesses distinct behavior (temperature at which desulfation occurs). La-CeO2 or CeO2-ZrO2-containing samples (monoliths) showed a greater resistance to deactivation during sulfation and required lower temperatures to restore the NOx storage efficiency to its pre-sulfation value. Fully formulated monolithic LNT catalysts containing varying amounts of Pt, Rh and BaO were subjected to accelerated aging to elucidate the effect of washcoat composition on LNT aging. Elemental analysis revealed that residual sulfur, associated with the Ba phase, decreased catalyst NOx storage capacity and that sintering of the precious metals resulted in decreased contact between the Pt and Ba phases. Spatially-resolved inlet capillary mass spectrometry (SpaciMS) was employed to understand the factors influencing the selectivity of NOx reduction in LNT catalysts degreened and thermally aged) containing Pt, Rh, BaO and Al2O3, and contained La-stabilized CeO2. Stretching of the NOx storage and reduction zone (NSR) zone resulted in increased selectivity to NH3 due to the fact that less catalyst was available to consume NH3 by either the NH3-NOx SCR reaction or the NH3-O2 reaction. Additionally, the loss of oxygen storage capacity (OSC) and NOx storage sites, along with the decreased rate of NOx diffusion to Pt/Rh sites, led to an increase in the rate of propagation of the reductant front after aging, in turn, resulting in increased H2:NOx ratios at the Pt/Rh sites and consequently increased selectivity to NH3. Finally, a crystallite scale model was used to predict selectivity to NH3 from the LNT catalysts during rich conditions after a fixed amount of NOx was stored during lean conditions. Both the experimental and model predicted data showed that the production of NH3 is limited by the rate of diffusion from the Ba storage sites to the Pt particles at 200 °C. At 300 °C, the process is limited by the rate at which H2 is fed to the reactor.

A library of graphene-based hybrid materials was synthesized as novel hybrid electrochemical electrodes for electrochemical energy conversion and storage devices and electrocatalytical sensing namely enzymeless glucose sensing. The materials used were supercapacitive graphene-family nanomaterials (multilayer graphene-MLG; graphene oxide-GO, chemically reduced GO-rGO and electrochemical reduced GOErGO) and pseudocapacitive nanostructured transition metal oxides including cobalt oxide polymorphs (CoO and Co3O4) and cobalt nanoparticles (CoNP). These were combined through physisorption, electrodeposition, and hydrothermal syntheses approaches. This project was carried out to enhance electrochemical performance and to develop electrocatalytic platforms by tailoring structural properties and desired interfaces. Particularly, electrodeposition and hydrothermal synthesis facilitate chemically-bridged (covalently- and electrostatically- anchored) interfaces and molecular anchoring of the constituents with tunable properties, allowing faster ion transport and increased accessible surface area for ion adsorption. The surface morphology, structure, crystallinity, and lattice vibrations of the hybrid materials were assessed using electron microscopy (scanning and transmission) combined with energy dispersive spectroscopy and selected-area electron diffraction, X-ray diffraction, and micro-Raman Spectroscopy. The electrochemical properties of these electrodes were evaluated in terms of supercapacitor cathodes and enzymeless glucose sensing platforms in various operating modes. They include cyclic voltammetry (CV), ac electrochemical impedance spectroscopy, charging-discharging, and scanning electrochemical microscopy (SECM). These hybrid samples showed heterogeneous transport behavior determining diffusion coefficient (4⨯10-8 – 6⨯10-6 m2/s) following an increasing order of CoO/MLG < Co3O4/MLG < Co3O4/rGOHT < CoO/ErGO < CoNP/MLG and delivering the maximum specific capacitance 450 F/g for CoO/ErGO and Co3O4/ rGOHT. In agreement with CV properties, these electrodes showed the highest values of low-frequency capacitance and lowest charge-discharge response (0.38 s – 4 s), which were determined from impedance spectroscopy. Additionally, through circuit simulation of experimental impedance data, RC circuit elements were derived. SECM served to investigate electrode/electrolyte interfaces occurring at the solid/liquid interface operating in feedback probe approach and imaging modes while monitoring and mapping the redox probe (re)activity behavior. As expected, the hybrids showed an improved electroactivity as compared to the cobalt oxides by themselves, highlighting the importance of the graphene support. These improvements are facilitated through molecular/chemical bridges obtained by electrodeposition as compared with the physical deposition.

Catalysis is one of the pillars of the chemical industry. While the use of catalyst is typically recognized in the automobile industry, their impact is more widespread as; catalysts are used in the synthesis of 80% of the US commercial chemicals. Despite the improved selectivity provided by catalyst, process inefficiencies still threaten the sustainability of a number of synthesis methods, especially in the pharmaceutical industry. Recyclable solid supported catalysts offer a unique opportunity to address these inefficiencies. Such systems coupled with continuous synthesis techniques, have the potential to significantly reduce the waste to desired product ratio (E-factor) of the production techniques. This research focuses developing sustainable processes to synthesize organic molecules by using continuous synthesis methods. In doing so, solid supported metal catalyst systems were identified, developed, and implemented to assist in the formation of carbon-carbon bonds. Newly developed systems, which utilized metal nanoparticles, showed reactivity and recyclability, comparable to commercially available catalyst. Nanoparticles are emerging as useful materials in a wide variety of applications including catalysis. These applications include pharmaceutical processes by which complex and useful organic molecules can be prepared. As such, an effective and scalable synthesis method is required for the preparation of nanoparticle catalysts with significant control of the particle size, uniform dispersion, and even distribution of nanoparticles when deposited on the surface of a solid support. This project describes the production of palladium nanoparticles on a variety of solid supports and the evaluation of these nanoparticles for cross coupling reactions. This report highlights novel synthesis techniques used in the formation of palladium nanoparticles using traditional batch reactions. The procedures developed for the batch formation of palladium nanoparticles on different solid supports, such as graphene and carbon nanotubes, are initially described. The major drawbacks of these methods are discussed, including limited scalability, variation of nanoparticle characteristics from batch to batch, and technical challenges associated with efficient heating of samples. Furthermore, the necessary conditions and critical parameters to convert the batch synthesis of solid supported palladium nanoparticles to a continuous flow process are presented. This strategy not only alleviates the challenges associated with the robust preparation of the material and the limitations of scalability, but also showcases a new continuous reactor capable of efficient and direct heating of the reaction mixture under microwave irradiation. This strategy was further used in the synthesis of zinc oxide nanoparticles. Particles synthesized using this strategy as well as traditional synthesis methods, were evaluated in the context industrially relevant applications.

Fixed bed reactors are widely applicable in a range of chemical process industries. Their ease of use and simplified operation make them an attractive and preferred option in reactor selection, however the geometric complexities within the bed as a result of the unstructured packing has made the design of such beds historically based on pseudo-homogenous models together with correlation-based transport parameters. Low tube-to-particle diameter ratio (N) beds, in particular, are selected for highly exothermic or endothermic reactions, such as in methane steam reforming or alkane dehydrogenation. Due to the large fraction of tube to catalyst particle contact in these low N beds, wall effects induce a mass transfer boundary layer at the wall, and in the case of thermal beds, a simultaneous resistance to heat transfer. Computational Fluid Dynamics (CFD) has been shown to be an accurate tool for experimental validation and predictive analysis of packed beds, and may be used to derive more accurate design parameters for fixed bed reactors. More specifically, the elucidation of dispersion, or the transport of reactant and product within the bed due to molecular diffusion and convective flow is of fundamental interest to the design of fixed beds. Computational Fluid Dynamics was used in this research to study solute dispersion in eight beds of varying N at a range of particle Reynolds numbers in the laminar flow regime. In the first stage of research, flow development was simulated in three-dimensional packed beds of spheres. Then, the reactor wall was sectioned to include a boundary condition of pure methane, from which the solute could laterally disperse into the bed. In the second stage, a two-dimensional representation of the bed was created using the commercial Finite Element Analysis software COMSOL Multiphysics. In these models, axial velocity profiles and radial methane concentration profiles taken from the 3-D models were supplied, and a fitting procedure by use of the Levenberg-Marquardt Least-Squares optimization algorithm was completed to fit radial dispersion coefficients and near-wall mass transfer coefficients to the CFD data. These optimization runs were conducted for all N at a number of bed depths in each case. Two sub-studies were conducted in which a constant velocity profile and a local velocity profile were supplied to the 2-D model, and the optimization re-run. It was found that this two parameter model did not fully account for various mechanisms of dispersion in the bed, namely the increasing rate of dispersion from the tube wall boundary layer up to the bed center, but only accounted for a diffusive-dispersion at the wall and a constant-rate, convective-dispersion everywhere else in the bed. Length dependency of dispersion coefficients were also noted, particularly in the developing sections of the bed. Nevertheless, the combined CFD and optimization procedure proved to be an accurate and time-efficient procedure for the derivation of dispersion coefficients, which may then lend themselves to the standard design of packed bed reactors.

Fixed bed reactors are widely applicable in a range of chemical process industries. Their ease of use and simplified operation make them an attractive and preferred option in reactor selection, however the geometric complexities within the bed as a result of the unstructured packing has made the design of such beds historically based on pseudo-homogenous models together with correlation-based transport parameters. Low tube-to-particle diameter ratio (N) beds, in particular, are selected for highly exothermic or endothermic reactions, such as in methane steam reforming or alkane dehydrogenation. Due to the large fraction of tube to catalyst particle contact in these low N beds, wall effects induce a mass transfer boundary layer at the wall, and in the case of thermal beds, a simultaneous resistance to heat transfer. Computational Fluid Dynamics (CFD) has been shown to be an accurate tool for experimental validation and predictive analysis of packed beds, and may be used to derive more accurate design parameters for fixed bed reactors. More specifically, the elucidation of dispersion, or the transport of reactant and product within the bed due to molecular diffusion and convective flow is of fundamental interest to the design of fixed beds. Computational Fluid Dynamics was used in this research to study solute dispersion in eight beds of varying N at a range of particle Reynolds numbers in the laminar flow regime. In the first stage of research, flow development was simulated in three-dimensional packed beds of spheres. Then, the reactor wall was sectioned to include a boundary condition of pure methane, from which the solute could laterally disperse into the bed. In the second stage, a two-dimensional representation of the bed was created using the commercial Finite Element Analysis software COMSOL Multiphysics. In these models, axial velocity profiles and radial methane concentration profiles taken from the 3-D models were supplied, and a fitting procedure by use of the Levenberg-Marquardt Least-Squares optimization algorithm was completed to fit radial dispersion coefficients and near-wall mass transfer coefficients to the CFD data. These optimization runs were conducted for all N at a number of bed depths in each case. Two sub-studies were conducted in which a constant velocity profile and a local velocity profile were supplied to the 2-D model, and the optimization re-run. It was found that this two parameter model did not fully account for various mechanisms of dispersion in the bed, namely the increasing rate of dispersion from the tube wall boundary layer up to the bed center, but only accounted for a diffusive-dispersion at the wall and a constant-rate, convective-dispersion everywhere else in the bed. Length dependency of dispersion coefficients were also noted, particularly in the developing sections of the bed. Nevertheless, the combined CFD and optimization procedure proved to be an accurate and time-efficient procedure for the derivation of dispersion coefficients, which may then lend themselves to the standard design of packed bed reactors.

As redes poliméricas são materiais amplamente estudados, pois suas propriedades especiais permitem que sejam aplicadas em áreas como indústria de fertilizantes, medicina, bioquímica, análises químicas dentre outras. A microestrutura de uma rede polimérica, em geral, exerce grande influência sobre as propriedades macroscópicas desses materiais e o interesse da influência dessa microestrutura nas propriedades finais são de interesse estratégico. As reações de ciclização influenciam no controle da microestrutura das redes poliméricas, é sabido que um aumento na diluição do sistema aumenta a incidência deste tipo de reações. A modelagem matemática da copolimerização do estireno-divinilbenzeno é um assunto amplamente estudado, porém poucos estudos foram realizados considerando as reações de ciclização com uma cinética definida e não um problema tipo caixa-preta. Este trabalho teve como principal objetivo o estudo da copolimerização de estireno-divinilbenzeno em solução e sua modelagem matemática com a inclusão das reações de ciclização intramoleculares. Sendo assim, reações de copolimerização de estireno-divinilbenzeno em soluções com baixas concentrações de monômeros foram realizadas em batelada em um reator de vidro, inicialmente foram utilizados dois modelos matemáticos para estudar o comportamento do sistema nestas condições, denominados: Modelo A e Modelo B. O Modelo A foi desenvolvido através do balanço de massa de todas as espécies no meio reacional e inclusão das reações de ciclização. O tamanho máximo dos polímeros mortos considerados neste modelo foi de 300 unidades monoméricas, pois devido à diluição acreditava-se que este tamanho máximo abrangesse todos os tamanhos de polímeros mortos, porém sua comparação com dados experimentais mostrou o contrário. O Modelo B foi baseado no modelo desenvolvido por Aguiar (2013) e utiliza o balanço de massa para as espécies não poliméricas e método dos momentos para as espécies poliméricas (radicais poliméricos e polímeros mortos). Este modelo utiliza também o Fracionamento Numérico para determinação das massas moleculares e ponto de gel, as reações de ciclização foram incluídas através do Método dos Caminhos. Quando comparados aos dados experimentais, o Modelo B mostrou-se mais realista com menores tempos de simulação e com menores problemas numéricos que o Modelo A, portanto este foi utilizado para o estudo do sistema em questão. Os resultados apresentados através do Modelo B indicam que o parâmetro atribuído à cinética das ligações cruzadas (Cp) foi de 0,05 e o valor do parâmetro de ciclização do menor segmento ciclizável (3 unidades monoméricas) foi de 130 s-1 para a temperatura de 90ºC, os valores para os demais tamanhos foram calculados através da equação de Rolfes e Stepto. Este trabalho é uma continuação ao trabalho de Aguiar (2013) e seus resultados mostraram que as simulações das variáveis: concentração de duplas ligações pendentes, Massa Molecular Mássica Média (Mw) e polidispersidade aproximaram-se mais dos dados experimentais quando as ciclizações são incluídas no modelo quando comparadas à abordagem sem a inclusão das reações de ciclização. / Polymer networks are widely studied materials; their especial properties allow them to be applied in areas such as the fertilizer industry, medicine, biochemistry, chemical analysis among others. In general, the polymer network microstructure has influence in macroscopic properties of materials, hence the interest of such microstructure in final properties are of strategic interest. The cyclization reactions influence in the microstructure control of polymer networks. It is known that an increase in systems dilution can increase the cyclization reactions incidence. Mathematical modeling of copolymerization of styrene-divinylbenzene is a widely studied subject, but few studies have been conducted considering the cyclization reactions with a defined kinetic and not a problem black-box type. This work aimed to study the styrene-divinylbenzene copolymerization solutions and their mathematical modeling with the inclusion of intramolecular cyclization reactions. Thus, solution copolymerization of styrene and divinylbenzene was carried out at low concentration of monomers in batch reactor. Two mathematical models were initially used to analize the behavior of the system, which were called: Model A and Model B. The Model A was developed by molar balance of species in the reaction medium and includes cyclization reactions, which were considered to happen in polymer chains with 300 or less monomer units. Due the dilution was believed that this number of units covering all sizes of dead polymers, but comparison between Model A an experimental data proved otherwise. The Model B was based in model of Aguiar (2013), and uses the mass balance for non-polimerics species and moments methods for polimerics species. Model B also uses numerical fractionation for average molecular weight and gel point determination, and the method of paths to approach cyclization reactions. When compared to experimental data, Model B proved more realistic, presenting shorter simulation times and less numerical problems than Model A. Therefore Model B was chosen to represent the system. The results presented by Model B indicate that the parameter assigned to the kinetics os crosslink (Cp) was fitted at 0,05 and cyclization rate constant for paths with 3 monomer units was fitted 130 s-1 at temperature of 90°C. The cyclization rate constants for longer paths were calculated trough Rolfes and Steptos equation. This work is a follow up to Aguiars work (2013) and the results showed that the simulation of variables: concentration of pendant double bonds, average molecular weight and polidispersity better predicted when the cyclization rate constants are greater than zero.

Naturally occurring enzymatic pathways enable highly specific, rapid thiophenic sulfur cleavage occurring at ambient temperature and pressure, which may be harnessed for the desulfurization of petroleum-based fuel. One pathway found in bacteria is a four-step catabolic pathway (the 4S pathway) converting dibenzothiophene (DBT), a common crude oil contaminant, into 2-hydroxybiphenyl (HBP) without disrupting the carbon-carbon bonds. 2’-Hydroxybiphenyl-2-sulfinate desulfinase (DszB), the rate-limiting enzyme in the enzyme cascade, is capable of selectively cleaving carbon-sulfur bonds. Accordingly, understanding the molecular mechanisms of DszB activity may enable development of the cascade as industrial biotechnology. Based on crystallographic evidence, we hypothesized that DszB undergoes an active site conformational change associated with the catalytic mechanism. Moreover, we anticipated this conformational change is responsible, in part, for enhancing product inhibition. Rhodococcus erythropolis IGTS8 DszB was recombinantly produced in Escherichia coli BL21 and purified to test these hypotheses. Activity and the resulting conformational change of DszB in the presence of HBP were evaluated. The activity of recombinant DszB was comparable to the natively expressed enzyme and was competitively inhibited by the product, HBP. Using circular dichroism, global changes in DszB conformation were monitored in response to HBP concentration, which indicated that both product and substrate produced similar structural changes. Molecular dynamics (MD) simulations and free energy perturbation with Hamiltonian replica exchange molecular dynamics (FEP/λ-REMD) calculations were used to investigate the molecular-level phenomena underlying the connection between conformation change and kinetic inhibition. In addition to the HBP, MD simulations of DszB bound to common, yet structurally diverse, crude oil contaminates 2’2-biphenol (BIPH), 1,8-naphthosultam (NTAM), 2-biphenyl carboxylic acid (BCA), and 1,8-naphthosultone (NAPO) were performed. Analysis of the simulation trajectories, including root mean square fluctuation (RMSF), center of mass (COM) distances, and strength of nonbonded interactions, when compared with FEP/λ-REMD calculations of ligand binding free energy, showed excellent agreement with experimentally determined inhibition constants. Together, the results show that a combination of a molecule’s hydrophobicity and nonspecific interactions with nearby functional groups contribute to a competitively inhibitive mechanism that locks DszB in a closed conformation and precludes substrate access to the active site. Limitations in DszB’s potential applications in industrial sulfur fixation are not limited to turnover rate. To better characterize DszB stability and to gain insight into ways by which to extend lifetime, as well as to pave the way for future studies in inhibition regulation, we evaluated the basic thermal and kinetic stability of DszB in a variety of solvation environments. Thermal stability of DszB was measured in a wide range of different commercially available buffer additives using differential scanning fluorimetry (DSF) to quickly identify favorable changes in protein melting point. Additionally, a fluorescent kinetic assay was employed to investigate DszB reaction rate over a 48 hr time period in a more focused group of buffer to link thermal stability to DszB life-time. Results indicate a concerningly poor short-term stability of DszB, with an extreme preference for select osmolyte buffer additives that only moderately curbed this effect. This necessitates a means of stability improvement beyond alteration of solvation environment. To this end, a more general investigation of glycosylation and its impact on protein stability was performed. Post-translational modification of proteins occurs in organisms from all kingdoms life, with glycosylation being among the most prevalent of amendments. The types of glycans attached differ greatly by organism but can be generally described as protein-attached carbohydrate chains of variable lengths and degrees of branching. With great diversity in structure, glycosylation serves numerous biological functions, including signaling, recognition, folding, and stability. While it is understood that glycans fulfill a variety of important roles, structural and biochemical characterization of even common motifs and preferred rotamers is incomplete. To better understand glycan structure, particularly their relevance to protein stability, we modeled and computed the solvation free energy of 13 common N- and O-linked glycans in a variety of conformations using thermodynamic integration. N-linked glycans were modeled in the β-1,4-linked conformation, attached to an asparagine analog, while O-linked glycans were each modeled in both the α-1,4 and β-1,4-linked conformations attached to both serine and threonine analogs. Results indicate a strong preference for the β conformation and show a synergistic effect of branching on glycan solubility. Our results serve as a library of solvation free energies for fundamental glycan building blocks to enhance understanding of more complex protein-carbohydrate structures moving forward.

This study explores the reactions and related species of NOx pollutants in methane flames in order to understand their production and consumption during the combustion process. To do this, several analytical simulations were run to explore the behavior of nitrogen species in the pre-flame, post- flame, and reaction layer regions. The results were then analyzed in order to identify all "steady-state" species in the flame as well as the determine all the unnecessary reactions and species that are not required to meet a defined accuracy. The reductions were then applied and proven to be viable.

In this thesis investigations are carried out on two research topics in context to spray drying. The first research topic is the production of dried particles having uniform characteristics. The second research topic is the development of new applications of the reaction engineering approach which, in recent times, has emerged as an effective tool to formulate drying kinetics models. The reaction engineering approach is also implemented to simulate the drying of monodisperse droplets corresponding to the experimental work in the first research topic. Manufacturing micron- and nano-sized particles having uniform characteristics has recently become a popular research area due to the unique functionalities of these kinds of particles in biomedical, drug delivery, functional foods, nutraceuticals, cosmetics and other valuable applications. Spray drying has been a common and economical route to produce dried particles. A typical characteristic of spray dried products is the existence of a significant variation in particle properties such as size and morphology. One possible idea to restrict this product non-uniformity is to achieve a good control over the droplet’s behaviour and characteristics inside the drying chamber. The current thesis has investigated an innovative spray drying technique, i.e. a single stream drying approach in order to restrict product non-uniformity. In this drying approach, identical sized droplets having vertical trajectories are dried under controlled gas flow conditions. The piezoelectricity-driven monodisperse droplet generator is used as the atomizer to disperse liquid droplets. A prototype single stream dryer was assembled based on the single stream drying approach using various components designed in the laboratory and several parts purchased from the market. Experiments were carried out using aqueous lactose solutions as a model system in order to check the practicability of manufacturing uniform-sized spherical particles. Preliminary results were found to be positive and reported in this thesis. Mathematical models on the drying of monodisperse droplets were developed in order to predict important droplet and gas parameter profiles during single stream drying. These models serve as a platform for design, optimization and scale-up purposes. Several important advantages and drawbacks of single stream drying are also reported. Problems encountered during the experimental work and future recommendations are presented in detail so that a more robust and effective drying research tool can be developed in future. Recently the reaction engineering approach (REA) has emerged as a simple and reliable technique to characterize the drying of various food and dairy materials. In this thesis two new applications of the REA are described for the first time in context to convective drying of aqueous droplets. The REA is used in this study to formulate the drying kinetics model for the drying of aqueous sucrose and maltodextrin (DE6) droplets. The effect of initial moisture content was explicitly demonstrated. The development of a new ‘composite’ REA which aimed to model the drying of aqueous droplets containing multiple solutes has been described. The composite REA was found to be suitable to characterize the drying behaviour of aqueous sucrose-maltodextrin mixtures of different proportions. The second new application of the REA is the development of a procedure to estimate surface properties of aqueous droplets during drying. In literature various droplet characteristics such as surface moisture contents were normally estimated using the diffusion-based drying kinetics model or the receding interface model. Surface moisture content and surface glass transition temperature profiles were evaluated here using a lumped-parameter model (REA) during the drying of aqueous sucrose, maltodextrin (DE6) and their mixtures. The same experimental data used for the development of the composite REA were used to yield predictions. The procedure was found to be useful in estimating surface moisture contents and understanding the stickiness behaviour of sugar droplets during drying. During the formulation of the REA-based drying kinetics model in this thesis, the assumption of uniform temperature within droplets was used. In most studies published in literature the uniform temperature assumption was justified by calculating the heat-transfer Biot numbers at the beginning and end of drying. However, the conventional Biot number concept does not take into account the evaporation effect and therefore would not be suitable to drying scenarios. In this thesis, a new approximation procedure is developed to estimate surface-centre temperature differences within materials following the entire drying process. This new procedure was helpful to check the extent of temperature non-uniformity within skim milk droplets under isothermal laboratory conditions as well as industrial spray drying conditions. Both conventional and drying-based Biot numbers are calculated and compared. Predictions showed that temperature gradients within the droplets were negligible during the drying of suspended droplets under laboratory drying conditions (slow drying), whilst the gradients were small and existed only for a short drying period for small droplets under industrial spray drying conditions (fast drying). Furthermore, it was observed that the maximum temperature gradient within the droplets did not exist at the starting or end points of the drying process, and therefore the estimation of Biot numbers at the starting and end point does not reflect temperature non-uniformity under drying conditions. This is a significant theoretical development in the area.