Reactive processing methods for functionalisation of polymers and in-situ compatibilisation of poly(ethylene terephthalate)-based blends

One of the main objectives of this study was to functionalise various rubbers (i.e. ethylene propylene copolymer (EP), ethylene propylene diene terpolymer (EPDM), and natural rubber (NR)) using functional monomers, maleic anhydride (MA) and glycidyl methacrylate (GMA), via reactive processing routes. The functionalisation of the rubber was carried out via different reactive processing methods in an internal mixer. GMA was free-radically grafted onto EP and EPDM in the melt state in the absence and presence of a comonomer, trimethylolpropane triacrylate (TRlS). To optinuse the grafting conditions and the compositions, the effects of various paranleters on the grafting yields and the extent of side reactions were investigated. Precipitation method and Soxhlet extraction method was established to purifY the GMA modified rubbers and the grafting degree was determined by FTIR and titration. It was found that without TRlS the grafting degree of GMA increased with increasing peroxide concentration. However, grafting was low and the homopolymerisation of GMA and crosslinking of the polymers were identified as the main side reactions competing with the desired grafting reaction for EP and EPDM, respectively. The use of the tri-functional comonomer, TRlS, was shown to greatly enhance the GMA grafting and reduce the side reactions in terms of the higher GMA grafting degree, less alteration of the rheological properties of the polymer substrates and very little formation of polyGMA. The grafting mechanisms were investigated. MA was grafted onto NR using both thermal initiation and peroxide initiation. The results showed clearly that the reaction of MA with NR could be thermally initiated above 140°C in the absence of peroxide. At a preferable temperature of 200°C, the grafting degree was increased with increasing MA concentration. The grafting reaction could also be initiated with peroxide. It was found that 2,5-dimethyl-2,5-bis(ter-butylproxy) hexane (TIOI) was a suitable peroxide to initiate the reaction efficiently above I50°C. The second objective of the work was to utilize the functionalised rubbers in a second step to achieve an in-situ compatibilisation of blends based on poly(ethylene terephthalate) (PET), in particular, with GMA-grafted-EP and -EPDM and the reactive blending was carried out in an internal mixer. The effects of GMA grafting degree, viscosities of GMAgrafted- EP and -EPDM and the presence of polyGMA in the rubber samples on the compatibilisation of PET blends in terms of morphology, dynamical mechanical properties and tensile properties were investigated. It was found that the GMA modified rubbers were very efficient in compatibilising the PET blends and this was supported by the much finer morphology and the better tensile properties. The evidence obtained from the analysis of the PET blends strongly supports the existence of the copolymers through the interfacial reactions between the grafted epoxy group in the GMA modified rubber and the terminal groups of PET in the blends.

Oxidation and stabilisation chemistry of metallocene-based polyolefins during melt processing

Metallocene catalyzed linear low density polyethylene (m-LLDPE) is a new generation of olefin copolymer. Based on the more recently developed metallocene-type catalysts, m-LLDPE can be synthesized with exactly controlled short chain branches and stereo-regular microstructure. The unique properties of these polymers have led to their applications in many areas. As a result, it is important to have a good understanding of the oxidation mechanism of m-LLDPE during melt processing in order to develop more effective stabilisation systems and continue to increase the performance of the material. The primary objectives of this work were, firstly, to investigate the oxidative degradation mechanisms of m-LLDPE polymers having different comonomer (I-octene) content during melt processing. Secondly, to examine the effectiveness of some commercial antioxidants on the stabilisation of m-LLDPE melt. A Ziegler-polymerized LLDPE (z-LLDPE) based on the same comonomer was chosen and processed under the same conditions for comparison with the metallocene polymers. The LLDPE polymers were processed using an internal mixer (torque rheometer, TR) and a co-rotating twin-screw extruder (TSE). The effects of processing variables (time, temperature) on the rheological (MI, MWD, rheometry) and molecular (unsaturation type and content, carbonyl compounds, chain branching) characteristics of the processed polymers were examined. It was found that the catalyst type (metallocene or Ziegler) and comonomer content of the polymers have great impact on their oxidative degradation behavior (crosslinking or chain scission) during melt processing. The metallocene polymers mainly underwent chain scission at lower temperature (<220°C) but crosslinking became predominant at higher temperature for both TR and TSE processed polymers. Generally, the more comonomers the m-LLDPE contains, a larger extent of chain scission can be expected. In contrast, crosslinking reactions were shown to be always dominant in the case of the Ziegler LLDPE. Furthermore, it is clear that the molecular weight distribution (MWD) of all LLDPE became broader after processing and tended generally to be broader at elevated temperatures and more extrusion passes. So, it can be concluded that crosslinking and chain scission are temperature dependent and occur simultaneously as competing reactions during melt processing. Vinyl is considered to be the most important unsaturated group leading to polymer crosslinking as its concentration in all the LLDPE decreased after processing. Carbonyl compounds were produced during LLDPE melt processing and ketones were shown to be the most imp0l1ant carbonyl-containing products in all processed polymers. The carbonyl concentration generally increased with temperature and extrusion passes, and the higher carbonyl content fonned in processed z-LLDPE and m-LLDPE polymers having higher comonomer content indicates their higher susceptibility of oxidative degradation. Hindered phenol and lactone antioxidants were shown to be effective in the stabilization of m-LLDPE melt when they were singly used in TSE extrusion. The combination of hindered phenol and phosphite has synergistic effect on m-LLDPE stabilization and the phenol-phosphite-Iactone mixture imparted the polymers with good stability during extrusion, especially for m-LLDPE with higher comonomer content.

Environmental degradation of polyethylene-based plastics

The criteria involved in the degradation of polyethylene-based degradable polymer samples have been investigated, with a view to obtaining a clearer mechanism of photo-biodegradation. The compatibility of degradable polymer samples during materials recycling was also studied. Commercial and laboratory prepared degradable polymer samples were oxidised in different environments and the oxidation products formed were studied using various analytical chromatographic and spectroscopic techniques such as HPLC, FT-IR and NMR. It was found that commercial degradable polymer samples which are based on the ECO systems, degrade predominantly via the Norrish II process, whereas the other degradable systems studied (starch-filled polyethylene systems, transition metal systems, including metal carboxylate based polyethylene systems and the photoantioxidant-activator systems) photodegrade essentially via the Norrish I process. In all cases, the major photoxidation products extracted from the degradable polymer samples were found to be carboxylic acids, although, in the polymer itself a mixture of carbonyl containing products such as esters, lactones, ketones and aldehydes was observed. The study also found that the formation of these hydrophilic carbonyl products causes surface swelling of the polymer, thus making bioerosion possible. It was thus concluded that environmental degradation of LDPE is a two step process, the initiation stage being oxidation of the polymer which gives rise to bioassimilable products, which are consequently bioeroded in the second stage, (the biodegradation step). Recycling of the degradable polymer samples as 10% homogeneous and heterogeneous blends was carried out using a single screw extruder (180°C and 210°C) and an internal mixer (190°C). The study showed that commercial degradable polymer samples may be recycled with a minimal loss in their properties.

The chemical modification of polymer blends by reactive processing

The primary objective of this research was to examine the concepts of the chemical modification of polymer blends by reactive processing using interlinking agents (multi-functional, activated vinyl compounds; trimethylolpropane triacrylates {TRIS} and divinylbenzene {DVD}) to target in-situ interpolymer formation between immiscible polymers in PS/EPDM blends via peroxide-initiated free radical reactions during melt mixing. From a comprehensive survey of previous studies of compatibility enhancement in polystyrene blends, it was recognised that reactive processing offers opportunities for technological success that have not yet been fully realised; learning from this study is expected to assist in the development and application of this potential. In an experimental-scale operation for the simultaneous melt blending and reactive processing of both polymers, involving manual injection of precise reactive agent/free radical initiator mixtures directly into molten polymer within an internal mixer, torque changes were distinct, quantifiable and rationalised by ongoing physical and chemical effects. EPDM content of PS/EPDM blends was the prime determinant of torque increases on addition of TRIS, itself liable to self-polymerisation at high additions, with little indication of PS reaction in initial reactively processed blends with TRIS, though blend compatibility, from visual assessment of morphology by SEM, was nevertheless improved. Suitable operating windows were defined for the optimisation of reactive blending, for use once routes to encourage PS reaction could be identified. The effectiveness of PS modification by reactive processing with interlinking agents was increased by the selection of process conditions to target specific reaction routes, assessed by spectroscopy (FT-IR and NMR) and thermal analysis (DSC) coupled dichloromethane extraction and fractionation of PS. Initiator concentration was crucial in balancing desired PS modification and interlinking agent self-polymerisation, most particularly with TRIS. Pre-addition of initiator to PS was beneficial in the enhancement of TRIS binding to PS and minimisation of modifier polymerisation; believed to arise from direct formation of polystyryl radicals for addition to active unsaturation in TRIS. DVB was found to be a "compatible" modifier for PS, but its efficacy was not quantified. Application of routes for PS reaction in PS/EPDM blends was successful for in-situ formation of interpolymer (shown by sequential solvent extraction combined with FT-IR and DSC analysis); the predominant outcome depending on the degree of reaction of each component, with optimum "between-phase" interpolymer formed under conditions selected for equalisation of differing component reactivities and avoidance of competitive processes. This was achieved for combined addition of TRIS+DVB at optimum initiator concentrations with initiator pre-addition to PS. Improvements in blend compatibility (by tensiles, SEM and thermal analysis) were shown in all cases with significant interpolymer formation, though physical benefits were not; morphology and other reactive effects were also important factors. Interpolymer from specific "between-phase" reaction of blend components and interlinking agent was vital for the realisation of positive performance on compatibilisation by the chemical modification of polymer blends by reactive processing.

The hydrodynamic behaviour and mass transfer characteristics of single droplets in a pulsed sieve plate column

The literature relating to the performance of pulsed sieve plate liquid-liquid extraction columns and the relevant hydrodynamic phenomenon have been surveyed. Hydrodynamic behaviour and mass transfer characteristics of droplets in turbulent and non-turbulent conditions have also been reviewed. Hydrodynamic behaviour, i.e. terminal and characteristic velocity of droplets, droplet size and droplet breakup processes, and mass transfer characteristics of single droplets (d≤0.6 cm) were investigated under pulsed (mixer-settler & transitional regimes) and non-pulsed conditions in a 5.0 cm diameter, 100 cm high, pulsed sieve plate column with three different sieve plate types and variable plate spacing. The system used was toluene (displaced) - acetone - distilled water. Existing photographic techniques for following and recording the droplet behaviour, and for observing the parameters of the pulse and the pulse shape were further developed and improved. A unique illumination technique was developed by which a moving droplet could be photographed using cine or video photography with good contrast without using any dye. Droplet size from a given nozzle and droplet velocity for a given droplet diameter are reduced under pulsing condition, and it was noted that this effect is enhanced in the presence of sieve plate. The droplet breakup processes are well explained by reference to an impact-breakup mechanism. New correlations to predict droplet diameter based on this mechanism are given below.vskip 1.0cm or in dimensionless groups as follows:- (We)crit= 3.12 - 1.79 (Eo)crit A correlation based on the isotropic turbulence theory was developed to calculate droplet diameter in the emulsion regime.vskip 1.0cm Experimental results show that in the mixer-settler and transitional regimes, pulsing parameters had little effect on the overall dispersed phase mass transfer coefficient during the droplet formation and unhindered travel periods.

The mechanisms of melt stabilisation of polyolefins

The effects of melt stabilisers on the oxidative degradation of polyolefins (polypropylene, low density polyethylene) have been studied under a variety of processing conditions . The changes in the both chemical and physical properties of unstabilised polymers occurring during processing were found to be strongly dependent on the amount of oxygen present in the mixer. 2 ,6 ,3' ,5' -tetra-tert-butyl-4'-phenoxy-4-methylene-2, 5-cyclohexadiene-1- one (galvinoxyl), iodine, nitroxyl radicals and cupric stearate were found to be very efficient melt stabilisers particularly when processed in a restricted amount of air. The mechanisms of their melt stabilising action have been investigated and a common cyclical regenerative mechanism involving both chain-breaking electron acceptor (CB-A) and chain-breaking electron donor (CB-D) antioxidant activity was found to be involved in each case. 2,6,3',5'-tetra-tert-butyl-4'-hydroxy phenyl-4-rrethylene-2,5-cyclohexadiene- 1-one (hydrogalvinoxy1), 4-hydroxy, 2,2,6, 6-tetra methyl-N-hydroxy piperidine and hydrogen iodide were formed together with olefinic unsaturation in the substrates during the melt processing of the polymers containing galvinoxyl, 4-hydroxy, 2,2,6, 6-tetra methyl piperidine oxyl and iodine respectively. No bonding of the melt stabilisers to the polymers was found to occur. Cupric stearate was found to undergo a similar redox reaction during its action as a melt stabiliser with the formation of unsaturation in the polymer. Evidence for the above processes is presented. The behaviours of melt stabilisers in the subsequent thermal and photooxidation of polyolefins have also been studied. Galvinoxyl which is very effective under both mild and severe processing canditions has been found to be an effective antioxidant during thermal oxidation (oven ageing) and it is also moderately good. as a photo-stabiliser. Iodine and cupric stearate acted efficiently during melt stabilisation of polymers, however they were both ineffective as thermo-oxidative antioxidants and UV stabilisers. Although the melt stabilisation effectiveness of stable nitroxyl radicals (e.g. 4-hydroxy, 2,2,6,6-tetra methyl piperidineoxyl and Bis- (2,2,6 ,6-tetra methyl-4- piperidinyl-N-oxyl) sebacate) is not as high as that of galvinoxyl during processing particularly in excess of air, they have been found to be much more efficient as UV stabilisers for polyolefins. The reasons for this are discussed.

The behaviour of antioxidants in polyolefins under conditions of U.V. exposure

The effects of antioxidants and stabilizers on the oxidative degradation of polyolefins (low density polyethylene [LDPE] and polypropylene [PPJ have been studied after subjecting to prior high temperature processing treatments. The changes in the both chemical and physical properties of unstabilized polymers occurring during processing were found to be strongly dependent on the amount of oxygen present in the mixer. Subsequent thermal and photo-oxidation showed very similar characteristics and the chromophore primarily responsible for:both thermo and photooxidative degradation of unstabilized polymers was found to be hydroperoxide formed during processing. Removal of hydroperoxide by heat treatment in an inert atmosphere although increasing ketonic carbonyl concentration, markedly decreased the rate of photo-oxidation, introducing an induction period similar to that of an unprocessed sample. It was concluded that hydroperoxides are the most important initiators in normally processed polymers during the early stages of photo-oxidation. Antioxidants such as metal dithiocarbamates which act by destroying peroxides into non-radica1 products were found to be efficient melt stabilizers for polyolefins and effective UV stabilizers during the initial photo-oxidation stage, whilst a phenolic antioxidant, n-octadecyl-3-(3',5'-di-terbutyl 4'hydroxypheny1) propionate (Irganox 1076) retarded photo-oxidation rate in the later stages. A typical 'UV absorber' 2-hydroxy-4-octyloxy-benzophenone (HOBP) has a minor thermal antioxidant action but retarded photo-oxidation at all stages. A substituated piperidine derivative, Bis [2.2.6.6-tetramethylpiperidlnyl-4] sebacate (Tinuvin 770) behaved as an pro-oxidant during thermal oxidation of polyolefins but was an effective stabilizer against UV light. The UV absorber, HOBP synergised effectively with both peroxide decomposing antioxidants (metal dithiocarbamates) and a chain-breaking antioxidant (Irganox 1076) during photo-oxidation of the poymers studed whereas the combined effect was additive during thermal oxidation. By contrast, the peroxide decornposers and chain-breaking antioxidant (Irganox 1076) which were effective synergists during thermal oxidation of LDPE· were antagonistic during photo-oxidation. The mechanisms of these processes are discussed.

Reactive processing of polymers: structural characterization of products by 1H and 13C NMR spectroscopy for glycidyl methacrylate grafting onto EPR in the absence and presence of a reactive comonomer

Grafted GMA on EPR samples were prepared in a Thermo-Haake internal mixer by free radical melt grafting reactions in the absence (conventional system; EPR-g-GMA(CONV)) and presence of the reactive comonomer divinyl benzene, DVB (EPR-g-GMA(DVB)). The GMA-homopolymer (poly-GMA), a major side reaction product in the conventional system, was almost completely absent in the DVB-containing system, the latter also resulted in a much higher level of GMA grafting. A comprehensive microstructure analysis of the formed poly-GMA was performed based on one-dimensional H-1 and C-13 NMR spectroscopy and the complete spectral assignments were supported by two-dimensional NMR techniques based on long range two and three bond order carbon-proton couplings from HMBC (Heteronuclear Multiple Bond Coherence) and that of one bond carbon-proton couplings from HSQC (Heteronuclear Single Quantum Coherence), as well as the use of Distortionless Enhancement by Polarization Transfer (DEPT) NMR spectroscopy. The unambiguous analysis of the stereochemical configuration of poly-GMA was further used to help understand the microstructures of the GMA-grafts obtained in the two different free radical melt grafting reactions, the conventional and comonomer-containing systems. In the grafted GMA, in the conventional system (EPR-g-GMA(CONV)), the methylene protons of the GMA were found to be sensitive to tetrad configurational sequences and the results showed that 56% of the GMA sequence in the graft is in atactic configuration and 42% is in syndiotactic configuration whereas the poly-GMA was predominantly syndiotactic. The differences in the microstructures of the graft in the conventional EPR-g-GMA(CONV) and the DVB-containing (EPR-g-GMA(DVB)) systems is also reported (C) 2009 Elsevier Ltd. All rights reserved.

Reactive processing of polymers:effect of in situ compatibilisation on characteristics of blends of polyethylene terephthalate and ethylene-propylene rubber

Ethylene-propylene rubber (EPR) functionalised with glycidyl methacrylate (GMA) (f-EPR) during melt processing in the presence of a co-monomer, such as trimethylolpropane triacrylate (Tris), was used to promote compatibilisation in blends of polyethylene terephthalate (PET) and f-EPR, and their characteristics were compared with those of PET/f-EPR reactive blends in which the f-EPR was functionalised with GMA via a conventional free radical melt reaction (in the absence of a co-monomer). Binary blends of PETand f-EPR (with two types of f-EPR prepared either in presence or absence of the co-monomer) with various compositions (80/20, 60/40 and 50/50 w/w%) were prepared in an internal mixer. The blends were evaluated by their rheology (from changes in torque during melt processing and blending reflecting melt viscosity, and their melt flow rate), morphology scanning electron microscopy (SEM), dynamic mechanical properties (DMA), Fourier transform infrared (FTIR) analysis, and solubility (Molau) test. The reactive blends (PET/f-EPR) showed a marked increase in their melt viscosities in comparison with the corresponding physical (PET/EPR) blends (higher torque during melt blending), the extent of which depended on the amount of homopolymerised GMA (poly-GMA) present and the level of GMA grafting in the f-EPR. This increase was accounted for by, most probably, the occurrence of a reaction between the epoxy groups of GMA and the hydroxyl/carboxyl end groups of PET. Morphological examination by SEM showed a large improvement of phase dispersion, indicating reduced interfacial tension and compatibilisation, in both reactive blends, but with the Tris-GMA-based blends showing an even finer morphology (these blends are characterised by absence of poly-GMA and presence of higher level of grafted GMA in its f-EPR component by comparison to the conventional GMA-based blends). Examination of the DMA for the reactive blends at different compositions showed that in both cases there was a smaller separation between the glass transition temperatures compared to their position in the corresponding physical blends, which pointed to some interaction or chemical reaction between f-EPR and PET. The DMA results also showed that the shifts in the Tgs of the Tris-GMA-based blends were slightly higher than for the conventional GMA-blends. However, the overall tendency of the Tgs to approach each other in each case was found not to be significantly different (e.g. in a 60/40 ratio the former blend shifted by up to 4.5 °C in each direction whereas in the latter blend the shifts were about 3 °C). These results would suggest that in these blends the SEM and DMA analyses are probing uncorrelatable morphological details. The evidence for the formation of in situ graft copolymer between the f-EPR and PET during reactive blending was clearly illustrated from analysis by FTIR of the separated phases from the Tris-GMA-based reactive blends, and the positive Molau test pointed out to graft copolymerisation in the interface. A mechanism for the formation of the interfacial reaction during the reactive blending process is proposed.

Coupling of Heck and hydrogenation reactions in a continuous compact reactor

A continuous multi-step synthesis of 1,2-diphenylethane was performed sequentially in a structured compact reactor. This process involved a Heck C-C coupling reaction followed by the addition of hydrogen to perform reduction of the intermediate obtained in the first step. Both of the reactions were catalysed by microspherical carbon-supported Pd catalysts. Due to the integration of the micro-heat exchanger, the static mixer and the mesoscale packed-bed reaction channel, the compact reactor was proven to be an intensified tool for promoting the reactions. In comparison with the batch reactor, this flow process in the compact reactor was more efficient as: (i) the reaction time was significantly reduced (ca. 7 min versus several hours), (ii) no additional ligands were used and (iii) the reaction was run at lower operational pressure and temperature. Pd leached in the Heck reaction step was shown to be effectively recovered in the following hydrogenation reaction section and the catalytic activity of the system can be mostly retained by reverse flow operation. © 2009 Elsevier Inc. All rights reserved.

Broadband RF current detector

The problems to be solved in this thesis were 1) development of a broadband RF preamplifier to be used with non-ferrous current probes so that the amplified signal exceeds the errors due to cable pickup, no detection is needed in this application, and 2) development of a self-contained device that amplifies and detects the output from a nonferrous current probe, providing a digital readout of the current. These instruments have been completed and are being tested for use by the National Institutes of Occupational Safety and Health (NIOSH). The self-contained current meter operates at frequencies up to 600 MHz, and detects currents as low as 8 mA . At these current magnitudes, the probe (pick-up coil) will output a voltage of 500μV (-53 dBm on 50Ω) which will have to be raised above 0 dBm. The final circuit uses a RF mixer as a variable attenuator in order to increase the dynamic range, two Monolithic Microwave Integrated Circuits (MMIC) for preamplification, a final broadband amplifier to raise the output compression point, a Schottky diode detector, a sample and hold circuit, and a liquid crystal digital panel meter.

Method to Improve the Linearity of Active Commutating Mixers Using Dynamic Current Injection

A double balanced (DBM) CMOS mixer providing high linearity is presented in this paper. A cross-coupled pair used in the IF stage of the mixer to dynamically inject current into the to mixer provide a high linearity. The proposed DBM was fabricated using a standard 130-nm CMOS process and was tested on-wafer. The double balanced mixer delivers 10 dB conversion gain, 9.5 dBm IIP3, and input P1dB of -2.4 dBm. RF bandwidth of the proposed mixer is 6 GHz, covering 0.5 GHz to 6.5 GHz with IF bandwidth of 300 MHz. RF to IF and LO to IF isolation are also better than 59 dB in the whole frequency band. The circuit uses an area of 0.015 mm2 excluding bonding pads and draw 4.5mW from a 1.2V supply.

Method to Improve the Conversion Gain Flatness of Transformer-Coupled Mixers

A 4-10 GHz, on-chip balun based current commutating mixer is proposed. Tunable resistive feedback is used at the transconductance stage for wideband response, and interlaced stacked transformer is adopted for good balance of the balun. Measurement results show that a conversion gain of 13.5 dB, an IIP3 of 4 dBm and a noise figure of 14 dB are achieved with 5.6 mW power consumption under 1.2 V supply. The simulated amplitude and phase imbalance is within 0.9 dB and ±2◦ over the band.

Energy drinks in Ireland - a review

Energy drinks have risen in popularity in recent years and are now sold in over 165 countries worldwide. On the island of Ireland, energy drinks advertising accounted for 20% of the total soft drinks market advertising in 2014. In the United States, sales increased by 60% between 2008 and 2012, and in 2006, a staggering 500 new brands of energy drinks were released worldwide. In the UK, the energy drinks market is worth £491 million and is growing by 7% year on year. This report has found an eightfold increase in the number of energy drinks available in 2015 compared to 2002. While no standard definition of an energy drink is used in the scientific literature, it is commonly understood to be a non-alcoholic drink that contains caffeine (usually its main ingredient), taurine, vitamins and sometimes a combination of other ingredients (such as guarana and ginseng, among others), and it is marketed for its perceived or actual benefits as a stimulant, for improving performance and for increasing energy. As this report will highlight, there is some confusion amongst the public as to what the term "energy drink" means, as some soft and sports drinks, while containing little or no caffeine, use the term ‘energy’ in the product label, for example, Lucozade. Both the scientific community and the public have raised health concerns about the caffeine and calorie intakes associated with energy drinks and the use of these drinks as a mixer with alcohol. These concerns are disputed by the energy drinks industry.

Investigation of the Nucleation and Growth of Colloidal Indium Phosphide: from Molecular Precursors to Semiconductor Nanocrystals through In37P20(O2CR)51 as a Magic-Size Intermediate

Thesis (Ph.D.)--University of Washington, 2016-06