Digital images are widely used in computer applications. Uncompressed digital images require considerable storage capacity and transmission bandwidth. Efficient image compression solutions are becoming more critical with the recent growth of data intensive, multimedia-based web applications. This thesis studies image compression with wavelet transforms. As a necessary background, the basic concepts of graphical image storage and currently used compression algorithms are discussed. The mathematical properties of several types of wavelets, including Haar, Daubechies, and biorthogonal spline wavelets are covered and the Enbedded Zerotree Wavelet (EZW) coding algorithm is introduced. The last part of the thesis analyzes the compression results to compare the wavelet types.

Tese de mestrado. Redes e Serviços de Comunicação. Faculdade de Engenharia. Universidade do Porto, INESC Porto. 2009

Despite the advent of alternative dosage forms, solid dosage forms constitute a major proportion of dosage forms not only on the market, but also in many pharmaceutical companies’ pipelines. This is because of their superior stability and ease of manufacturing relative to other dosage forms. Although solid dosage forms have been the topic of discussion for decades, the process of compaction of these dosage forms is considered an art rather than science because of the empiricism that exists in this area. With the introduction of Quality by Design (QbD), it is imperative that the drug development process is guided by structured scientific principles. It has been hypothesized that crystal structure of organic solids plays a pivotal role in understanding the properties, processing and eventually performance of solids. In this regard, the intermolecular interactions within a solid play a paramount role in dictating the materials response to stress. One important parameter that is weakly addressed in the literature is the concept of strength of intermolecular interactions. In the current work, we utilize the concept of elasticity as a metric for strength of intermolecular interactions. We introduce powder Brillouin light scattering; an inelastic light scattering technique to measure the elasticity of organic solids and try to correlate the mechanical moduli extracted from the spectra to the compaction performance of solids. We hypothesize that any redistribution of intermolecular interactions would be reflected in the BLS spectrum of these materials and the material properties can be used to explain the differences in compaction performance. Before we tested our hypothesis, we validated our powder BLS technique using aspirin as model system. We then applied the same analysis to four model systems that involved different ways of reorganizing the intermolecular interactions upon subtle modifications to the molecular structure. In Chapter 4, we investigated the effect of alkyl chain length and crystal structure on the mechanical properties and compaction performance of p-aminobenzoic acid (PABA) and its esters. For the entire ester series, a similar hydrogen bonding pattern was observed with strong N-H…O (carbonyl) and supportive N-H…N interactions. While the ethyl and butyl esters exhibited a layered structure, the methyl ester displayed a 3-D isotropic structure. The crystal structure for PABA exhibited a three-dimensional, quasi-isotropic distribution of the hydrogen-bonding interactions that connected the PABA dimers. The powder BLS spectra for these materials revealed low velocity shear modes for the layered structures and a spectrum consistent with an isotropic structure for Me-PABA and PABA. This was in good agreement with the compressibility behavior under load, with Et-PABA and Bu-PABA more compressible than PABA. However, due to greater particle-particle adhesion, PABA compacts showed greater tensile strength at higher pressures. The moduli calculated also showed that both Et-PABA and Bu-PABA had lower shear and Young’s modulus relative to the other materials. Attachment energies corroborated the above results. These studies showed that weak dispersive forces play an important role in understanding material properties. In Chapter 5, a series of nitrobenzoic derivatives were used to study the effect of secondary interactions on the crystal reorganization and material properties. The materials used in the study include p-nitrobenzoic acid (4-NBA), m-nitrobenzoic acid (3-NBA, 4-chloro-3-nitrobenzoic acid (Cl-NBA), 4-bromo-3-nitrobenzoic acid (Br-NBA), and 4-methyl-3-nitrobenzoic acid (Me-NBA). Crystal structures of the materials revealed different organization of C-H…O interactions. Two types of C-H…O interactions were prevalent namely C-H…O (nitro) and C-H…O (carboxy). The reorganization of these two types of interactions led to different packing motifs and different mechanical behavior. These structural features were reflected in their mechanical properties assessed by powder Brillouin light scattering. Cl-NBA and Br-NBA displayed an isotropic spectrum similar to polystyrene and aspirin. 3-NBA, 4-NBA and Me-NBA displayed different spectra from Cl-NBA and Br-NBA with high frequency tailing in the longitudinal mode distribution indicating a specific direction of extended molecular interactions. The Young’s modulus and shear modulus followed the order: 3-NBA < Me-NBA< 4-NBA < Cl-NBA < Br-NBA. The maximum longitudinal modulus Mmax was the highest for 3-NBA and was significantly greater than rest of the materials. From the compaction studies, it was observed that the tabletability followed the rank order 3-NBA > 4-NBA > Me-NBA ≈ Br-NBA ≈ Cl-NBA which is the same order as Mmax. 3-NBA by virtue of its low shear and Young’s modulus to be the most compressible material, but the compressibility rank order was 4-NBA > Me-NBA ≈ 3-NBA > Cl-NBA > Br-NBA. However, 3-NBA by virtue of its greater particle-particle adhesion was the most compactable material. The yield pressures obtained from Heckel plots revealed that 4-NBA and Me-NBA were relatively more plastic when compared to the other materials. This study demonstrated that subtle changes to the molecular structure can result in drastically different crystal packing which in turn would influence the mechanical properties and the compaction performance of organic solids. In Chapter 6, we investigated the effect of cocrystallization on the compaction performance of caffeine(CAF). The series of halo-nitrobenzoic acids (F-NBA, Cl-NBA and 3-NBA) were used as coformers. The cocrystals CAF: F-NBA, CAF: Cl-NBA and CAF: NBA Form 1 adopted a flat layered structure that can undergo deformation with ease. This increased the compressibility of the cocrystals relative to CAF. In addition to the improved compressibility, by virtue of increased particle-particle contacts, the cocrystals also displayed superior tabletability. In contrast to the layered structures adopted by the three cocrystals, the CAF: NBA Form 2 displayed a columnar structure that exhibited resistance to stress. The compressibility and the tabletability of CAF: NBA Form 2 was significantly compromised when compared to that of Form1. All the compaction characteristics of the cocrystals were in good agreement with moduli and parameters obtained from powder BLS spectra. The layered materials showed the presence of low velocity shear modes corroborating the earlier studies. There was a clear difference in the spectra of the polymorphs, indicating that powder BLS can be used for mechanical screening of polymorphs. In Chapter 7, we examined the effects of crystal structure and coformer functionality on the compaction performance of theophylline (THY). The coformers employed include 4-fluoro-3-nitrobenzoic acid (FNBA), acetaminophen (APAP), and p-aminobenzoic acid (PABA). While THY-APAP and THY-FNBA exhibited layered structures, the THY-PABA displayed a interdigitated columnar structure. Powder BLS spectra showed the presence of low frequency shear modes relative to THY for all the three cocrystals. However, the order of frequencies followed: THY-FNBA < THY-APAP < THY-PABA. The shear modulus calculated followed the order THY-APAP≈ THY-FNBA < THY < THY-PABA which is in agreement with the crystal structures discussed. The Young’s modulus followed the order THY-FNBA < THY-APAP < THY < THY-PABA. The two layered structures (THY-APAP, THY-FNBA) showed distinct compaction performance (similar compressibility but different compactability and tabletability). The layered structures were more compressible than THY which is hypothesized to undergo deformation through multiple mechanisms. THY-PABA showed poor compaction properties. This highlights the fact that although the coformer (PABA) is molecularly similar to FNBA, the resultant cocrystals are structurally and mechanically distinct. These observations were well supported by the moduli calculated from powder BLS studies. The order of yield pressures obtained from Heckel analysis followed the same order as shear modulus. The tensile strength of the compacts of the cocrystals level off at around 150 MPa but the tensile strength of THY compacts continues to increase. From a manufacturing perspective the cocrystals can prove to be a better option as they as more compactable at higher porosities or they possess greater tabletability at low compaction pressures. Overall, we have used model systems to demonstrate that the redistribution of intermolecular forces upon point substitution or cocrystallization have a dramatic effect on the material properties. It is also worth noting that elasticity along with plasticity can provide important information about the strength of interactions which would help in understanding the role of weak intermolecular forces in the performance of organic materials. To gain a better perspective of the compaction process and move towards a QbD approach, it is also imperative to understand the link between crystal structures, intermolecular interactions which is possible with the help of novel characterization techniques (BLS, AFM).

Abdominal aortic aneurysms are a focal dilation of the aorta which can be potentially life threatening if left untreated. Endovascular aneurysm repair (EVAR) is a noninvasive treatment that can reduce the mortality rate when compared to the standard open repair. Yet, EVAR is associated with other complications that can arise such as migration, endoleaks, or device related failures. These complications drive the need for reinterventions which have been shown to occur more frequently with EVAR than with open repair. The long term fixation and sealing characteristics of these devices is likely related to the nature of its apposition to the aortic wall. Currently there is little understanding of these mechanics and factors in how the device performs at the time of deployment. A computational model that reflects the compressive nature of an endovascular graft is beneficial in investigating these mechanics. The aims of the study are; 1) formulate an experimental methodology that captures the radial compressive nature of the stent graft, 2) develop a parameterized finite element model of the stent structure, and 3) compare the compressive behavior this model against the acquired experimental data. A 2 mil polyethylene sleeve was used to transfer a compressive vacuum pressure from the sleeve onto 10 independent stent grafts. The loading-unloading pressure was cycled from 0 to -50 mmHg (complete collapse) over 5 minutes. A pressure transducer and optical micrometer were used to capture the vacuum pressure and diameter relationship. All ten grafts compressed in a similar elliptical shape configuration. Commercial software was leveraged to construct a parameterized model of the stent geometry. All crest and trough vertex locations of the sinusoidal stent structure were validated within 1 mm of a measured value. A dynamic quasi-static computational simulation was completed that included large deformations and contact between the sleeve and stent as well as self-contact in the sleeve. Our results show that the model is representative of the experiments and can be used to interrogate how a stent graft will perform during certain stages of deployment and immediately after deployment with some caution in regard to the stated limitations.

Les transistors bipolaires à hétérojonctions (TBH) de la filière InP offrent aujourd’hui des fréquences de coupure supérieures à 400GHz pour le système InP/GaAsSb. Grâce à ces fréquences, les transistors bipolaires sont utilisés pour la réalisation de circuits performants dans des applications millimétriques telle que les communications optiques. Ainsi, pour atteindre ces performances remarquables, les dimensions verticales et latérales des TBH ont été considérablement réduites, entraînant l’auto-échauffement dans les TBHs aux densités de courant élevées. Cette thèse a donc pour objet l’étude et le développement de TBH InP/GaAsSb reportés sur un substrat hôte de silicium en vue de l’amélioration de la dissipation thermique. Une technique de transfert des couches épitaxiales a d’abord été présentée. Nous étudions ensuite les problématiques liées à la technique choisie et les paramètres de report par thermo-compression à basse température ont été optimisés. Le développement de la technologie InP/GaAsSb sur silicium a ensuite été effectué en partant d’une technologie classique de TBH non reportés. La réalisation du contact de collecteur, notamment, a fait l’objet d’une attention particulière. La réduction de l’épaisseur des couches actives ainsi que la technologie employée ont permis d’atteindre une fréquence de transition Ft supérieure à 400GHz. L’étude du comportement thermique des TBH a enfin été présentée grâce à l’extraction de la résistance thermique. Des valeurs très faibles ont été obtenues sur la technologie reportée de 800 à 1300W/K.m selon les dimensions des transistors ; ces valeurs sont très proches de celles simulées pour la même technologie. Elles constituent les premières mesures effectuées sur des TBH InP/GaAsSb transférés sur un substrat de silicium à haute conductivité thermique. Le report des TBHs sur silicium a ainsi permis une amélioration de la résistance thermique de 70% par rapport à une technologie standard de TBH non reportés. Ces résultats permettent de conclure quant à l’efficacité du report pour la réduction drastique de l’auto-échauffement dans les transistors bipolaires. / The InP heterojonctions bipolar transistors (HBT) offer today cut-off frequencies larger than 400GHz for the InP / GaAsSb system. Thanks to these performances, these transistors are used for the realization of successful circuits in millimeter-wave applications such as the optical communications. So, to reach these remarkable performances, the HBT are subject to a notorious self-heating phenomenon due to high current density of collector. This thesis thus has for object the study and the development of InP / GaAsSb HBT transferred on a host substrate of silicon with the aim of the improvement of the thermal behavior. We report first of all the principles of the bipolar transistor as well as the state of the art of the various materials used for fast transistors. A transfer technique of epitaxial layers was then presented. We study bounding problems resulting from the chosen technique and transfer parameters for valid thermo-compression at low temperature were optimized. The development of InP / GaAsSb transferred technology on silicon was then made. In particular, the collector contact realization has needed particular attention. Active layers thickness reduction as well as device fabrication process technology allowed reaching transition frequency Ft higher than 400GHz. The study of HBT thermal behavior was finally presented with thermal resistance extraction. Very low values were obtained on the transferred technology, from 800 to1300 W/K.m according to transistors size; these values are very close to those obtained by TCAD simulation for such a technology. It is the first measurement on InP / GaAsSb transferred-HBT on high thermal conductivity silicon substrate. This transfer technology has so allowed thermal resistance improvement of 70 % compared with that of standard HBT technology. This work leads to the influence of transferred-substrate for the severe reduction of self-heating in bipolar transistors technology.

Nearly one-third of all pharmaceutical substances on the market are able to sorb water into their crystal lattices to form hydrates, which can often compromise stability during processing and/or storage[1]. The tendency of a hydrate to lose its water of crystallization during the manufacturing process of tablet compression is of particular concern to formulation scientists. The amount of water freed as a function of increasing compaction pressure can be explained by the mobility of water within the compact. The mobility of water is determined by the size and shape of the crystal lattice, the numbers and strengths of the hydrogen bonds, and the presence of high-energy sites of disorder[2]. Due to their differing crystal structures, theophylline monohydrate (THM), citric acid monohydrate (CAM), theophylline-water-citric acid cocrystal hydrate (CATHP hydrate), and dicalcium phosphate dihydrate (DCPD) make for interesting model systems to examine the dehydration under mechanical load. The thermal dehydration of both powders and tablets was carried out via thermal gravimetric analysis (TGA). By comparing the temperatures required to start removal of water loss from the powder to that of the tablet, the average amount of water of crystallization that is freed by the compaction process may be quantified. The average amount of water freed by the compaction process results from a competition between the mechanically-induced disorder of the crystal structure that increases the molecular mobility of water within the tablets, and the trapping of water within the interparticulate void spaces at high compaction pressures. The compressibilities, compactabilities, and tabletabilities of the materials were calculated as a function of increasing compaction pressure. The consolidation of the powder bed under pressure was modeled by out-of-die Heckel Analysis which demonstrated the ease of deformation of the model compounds. XRD was utilized to show the decrease in overall order of the crystal lattice as a result of compression as well as anisotropy within the tablets. Crystallographic approaches were utilized to demonstrate the compactness of the crystal structure, and how it affects water mobility. Relaxation pulse experiments (T1, T2) utilizing solid-state NMR were used to directly probe the mobilities of the water molecules within the crystal lattice of THM. The results from T1 and T2 relaxation experiments directly measure the change in molecular mobility of water within the tablets as a function of compaction pressure. This provided independent verification of the trends in molecular mobility and average water freed as a function of compaction pressure observed during TGA dehydration. Raman spectroscopy was used to indirectly measure the polarizability and vibrational motions of THM, and these results corroborate those obtained from ssNMR and TGA dehydration experiments. Overall, this work highlights the potential impact that tablet compression can have on API hydrate stability. 1. Hilfiker R (editor). 2006. Polymorphism in the Pharmaceutical Industry. Weinheim, Germany: Wiley-VCH Verlag GmbH & Co, KGaA. 2. Byrn SR, Pfeiffer RR, Stowell JG. 1999. Solid-state chemistry of drugs. SSCI, Inc.

Knowledge of structural stability theory is of paramount importance to the practicing structural engineer. In many instances, buckling is the primary consideration in the design of various structural configurations. The first chapter introduces a simplified treatment of the elastic stability of a battened column using classical methods without getting involved with lengthy and complicated mathematical operations. In chapter two, a treatment of the elastic stability of a frame is presented, including effects of elastic restraints. In this study, a theoretical treatment is given which although approximate, is believed to constitute a satisfactory solution of the structure.

The primary objective of this study was to find out if changes in chemistry could be used to quantify Pinus radiata compression wood severity or degree of compression wood development. Basic chemical composition and the lignin structure was assessed for a range of different compression wood samples sourced from juvenile wood, mature wood, earlywood, latewood, branches, knots, 2-year and 1-year old Pinus radiata. Fluorescence microscopy was used as the reference method to assess the degree of compression wood development. Lignin structure of compression wood was studied by thioacidolysis, size exclusion chromatography, and thioacidolysis/31P NMR spectroscopy. Variation in the basic chemical composition and lignin structure with compression wood severity was ascertained. Results showed that, as the severity of compression wood changed, progressively from normal through mild to severe, all chemical parameters commonly associated with compression wood changed concurrently. With increasing severity lignin and galactose levels increased while glucose and mannose levels decreased. Lignin structural changes were also associated with changing severity of compression wood. Levels of p-hydroxyphenyl (H) releasable β-ethers increased and guaiacyl (G) releasable β-ethers decreased. Similarly, levels of uncondensed p-hydroxyphenyl units increased, while uncondensed guaiacyl units decreased. Similar proportions of condensed guaiacyl units were present in compression wood and normal wood. Similar trends in chemical composition were observed between the compression wood and related opposite wood in branches, knots and young wood of Pinus radiata. A number of chemical parameters changed linearly with compression wood severity. They were: the amount of lignin and galactose, the galactose/glucose ratio and p-hydroxyphenyl content in lignin. Parameters based on the p-hydroxyphenyl unit content in lignin, the H/G releasable β-ether ratio, releasable p-hydroxyphenyl β-ether units and uncondensed p-hydroxyphenyl C9 units are most suitable indicators of compression wood severity as they spanned a larger range relative to the normal wood levels and were not influenced by the morphological origin of wood samples. Chemical methods for quantifying compression wood severity should focus on the detection and measurement of these parameters. Galactan present in Pinus radiata compression wood was isolated and characterised. Structural investigation by methylation analysis and NMR spectroscopy revealed that this galactan was largely composed of (1→4)-linked β-D-galactopyranose residues. No evidence was found to indicate the presence of any branches. Characterisation of lignin in cell wall fractions of Pinus radiata normal wood revealed that middle lamella lignin has a higher lignin content, a lower amount of releasable β-ethers and a more condensed lignin than the secondary wall lignin. Levels of releasable p-hydroxyphenyl units were not higher in middle lamella lignin. A new method based on thioacidolysis and 31P quantitative NMR spectroscopy for estimation of the degree of lignin condensation of the phenolic and etherified C9 units in in situ wood lignin is described. Using this method it was found that phenolic C9 units in in situ lignin were considerably less condensed than etherified C9 units in both compression wood and normal wood.

"October 2002 (Revised May 2003)"--T.p. Includes bibliographical references (leaves 256-264).

Thesis (M.S.)--University of Nevada, Reno, 2005. / "August 2005." Includes bibliographical references (leaves 118-127). Online version available on the World Wide Web.