▼ The foundation of materials science is the structure-property relationship of materials. The structure of materials can be described at various length scales like macroscopic, microscopic, atomic and subatomic structure. Since the properties of materials are ultimately founded in the underlying atomic structure, the design of novel materials with optimized properties requires an increasing focus on the atomic lengthscale. There, computational methods need to be developed that can quickly and efficiently screen the nearly infinite number of possibilities of mixing different elements into alloys for optimized compositions. In this study we have used atomistic modeling to examine the structure and properties of complex materials with the goal of identifying structures and properties of materials for finding improved and optimized processes and compositions. With the availability of parameters to study almost all the systems, density functional theory (DFT) is probably the most widely used theory to model the atomic and electronic structure of materials. While DFT is very accurate, it is computationally very expensive which limits the system size to ~1000 atoms. Molecular dynamics can be used with empirical potentials to study larger-scale materials which have a less periodic arrangement of atoms such as glasses, cracks and fracture in metals, polymers, proteins etc. In this work we have used both of these atomistic techniques to understand materials such as metallic glasses and processes involved in corrosion. In the first study, we have examined the stability and transformation of metavanadate and decavanadate which is important to understand their role as corrosion inhibitors for aluminum alloys. We have used DFT and ab-initio molecular dynamics to understand their solvation behavior in water. The water cell size effect on the solvation behavior of metavanadate was investigated as well. In this thesis, we have also investigated if the effect of changing the number of electrons can be correlated to the effect of changing the pH value of the solution on the vanadates. In the second study, we have developed an empirical embedded atom method (EAM) potential (which takes into account the effect of electron clouds in metals) of beryllium, which among others is an important element in common bulk metallic glass alloys. This task was necessary since all EAM and modified EAM potentials (which takes the angular nature of the bonding into account in addition to the electron cloud effect) of beryllium available in the literature failed to predict the properties of beryllium to a sufficient degree of accuracy. The developed potential predicts the relative stabilities of various lattices of beryllium correctly when compared to values calculated using DFT. It also predicts the elastic moduli of beryllium with good accuracy and without negative moduli, which was a major drawback of the previously reported potentials. Additionally, it predicts the vacancy and vacancy cluster formation energy and the interstitial formation energy with good accuracy. The third study in this thesis is concentrated on the study of metallic glasses using molecular dynamics with embedded atom method potentials. We have addressed the issue of predicting the glass forming ability in ternary systems without using any experimental input, solely based on a computational study which was still missing in the literature. Our results suggest that the modeled fragility is sufficient to predict the glass forming ability in ternary systems, which we validate by calculating the fragility as a function of composition in the ternary Cu-Zr-Ti system. We also show that the icosahedral fraction of the nearest-neighbor shells, which has been frequently suggested to provide information on glass forming ability, does not correlate well with it while its differential with respect to the composition may correlate well with the glass forming ability. However, due to the limited number of test cases examined in the present study, this needs to be further probed. The deformation behavior of metallic glasses is very different from that of their crystalline counterparts owing to the absence of long-range order. We have also explored the deformation behavior of metallic glasses using molecular dynamics for two glass compositions, Cu65Zr35Ti5 and Cu45Zr45Al10. The effect of system size, annealing temperature, testing temperature and strain rate on the deformation behavior of the glasses was investigated. We found that annealing temperature and strain rate do not affect the elastic modulus to a great extent, while the modulus increases with decreasing the testing temperature from 300 K to 50 K. The compressive and tensile yield strength were also investigated. The results suggest that the yield strength is higher for compressive load. The deformation of glass is discussed in terms of change in icosahedron fraction and von Mises shear strain. Finally, the deformation behavior of metal-metallic glass composites was explored for the cases where the crystal is the stronger and weaker phase, respectively. For this we have studied Cu-(Cu65Zr35Ti5) nano-laminates with copper as the stronger phase and Al-(Cu45Zr45Al10) nano-laminates where the glass is the stronger phase. The effect of orientation of the crystal relative to the interface, the loading direction, the annealing temperature and the crystal volume percent was also investigated. Advisors/Committee Members: Windl, Wolfgang.

▼ Bone is an anisotropic, hierarchically structured material, and as a result, its mechanical behavior is highly statistical in nature. It has been shown for other engineering materials that mechanical testing at the microscale enables characterization of individual microstructural components in an effort to understand their role in the macroscopic mechanical behavior. The application of such microscale testing to bone will permit modeling of the aggregate material to predict effects of age, disease, or injury on the mechanical properties, thus enabling a better understanding of the disease state.In the present work, dual focused ion beam (FIB) and femtosecond (FS) laser micromachining techniques are employed to produce microscale mechanical test specimens of bovine cortical bone on the order of 10 – 30 µm. A FIB is advantageous for micromachining pillars as it is capable of producing small scale features by applying a Ga+ ion beam that penetrates and removes the surrounding material. A FS laser uses ultrashort laser pulses to ablate the material by locally heating it to its vaporization temperature, creating a plasma that is dissipated into a flowing gas. The FS laser is advantageous for micromachining of biological materials because it may be used in ambient, non-vacuum environments, making it a flexible tool for machining the bone surface while preserving its microstructure. The short pulse duration minimizes thermal diffusion and heating of the surrounding material. Prior research suggests that FS laser machining causes very little residual damage to the surrounding bone tissue. Processing parameters and feasible specimen geometries and dimensions are discussed. The fabrication of such pillars allows for micromechanical compression testing of time independent behavior using a modified nanoindenter with a flat punch tip. By achieving successful fabrication of micron scale pillars, it is possible to test the constitutive mechanical properties of mineralized tissue that comprises bone. The present work analyzes the mechanical testing of 20- and 30 µm nominal diameter pillars to small and large strains. Pillars tested to small strains are selectively placed within regions of interest on the bone sample surface (i.e. osteons). Modulus, strength, and modes of deformation are compared between samples. Advisors/Committee Members: Flores, Katharine.

▼ Methane and carbon monoxide are two hazardous gases which require continuous monitoring by gas sensors in underground coal mines for explosion prevention and toxicity, respectively. This work explored implementing miniaturized gas sensors in this area to simultaneously detect both gases for benefits of increased portability and reduced power consumption of the chemiresistive gas sensor device. The focus of this research was to understand how the particle size, morphology, and microstructure of the metal-oxide film affected the gas sensor performance to the two gases of interest on miniaturized gas sensor devices in the form of microhotplate platforms. This was done through three main research studies. The first was conducted by growing SnO2 nanowires from SnO2 particles using an Au-catalyst. Growth conditions including temperature, time, and oxygen partial pressure were explored to determine the formation aspects of the SnO2 nanowires. Gas sensor studies were completed that provided evidence that the SnO2 nanowires increased detection to a fixed concentration of carbon monoxide compared to SnO2 particles without nano-structure formation. A second research study was performed to compare the gas sensor performance of SnO2 nanoparticles, hierarchical particles, and micron-size particles. The nanoparticles were developed into an ink and deposited via ink-jet printing on the microhotplate substrates to control the microstructure of the metal-oxide film. By preventing agglomeration of the nanoparticle film, the SnO2 nanoparticles displayed similar gas sensor performance to methane and carbon monoxide as the hierarchical particles. Both nano-structures had much higher gas sensor response than the micron-size particles which confirms the surface area of the metal-oxide film is critical for reaction of the analyte gas at the surface. The last research study presented in the dissertation describes an oxide nanoparticle array developed for detecting methane and carbon monoxide in the presence of one another. A design of experiments was constructed and principal component analysis was used for determining the optimum temperatures of the metal-oxide elements. A four element array was developed with the SnO2 and TiO2 sensor elements able to detect methane concentrations of interest and the ZnO and NiO sensor elements able to detect the carbon monoxide concentrations. A linear based prediction model was developed and tested for accuracy and reproducibility of the model to a series of random gas concentrations. Advisors/Committee Members: Morris, Patricia.

▼ Self-assembled nanoislands in the gadolinia-doped ceria (GDC)/ yttria-stabilized zirconia (YSZ) system have recently been discovered. This dissertation is an attempt to study the mechanism by which these nanoislands form. Nanoislands in the GDC/YSZ system form via a strain based mechanism whereby the stress accumulated in the GDC-doped surface layer on the YSZ substrate is relieved by creation of self-assembled nanoislands by a mechanism similar to the ATG instability. Unlike what was previously believed, a modified surface layer is not required prior to annealing, that is, this modification can occur during annealing by surface diffusion of dopants from the GDC sources (distributed on the YSZ surface in either lithographically defined patch or powder form) with simultaneous breakup, which occurs at the hold temperature independent of the subsequent cooling. Additionally, we have developed a simple powder based process of producing nanoislands which bypasses lithography and thin film deposition setups. The versatility of the process is apparent in the fact that it allows us to study the effect of experimental parameters such as soak time, temperature, cooling rate and the effect of powder composition on nanoisland properties in a facile way. With the help of this process, we have shown that nanoislands are not peculiar to Gd containing oxide source materials on YSZ substrates and can also be produced with other source materials such as La2O3, Nd2O3, Sm2O3, Eu2O3, Tb2O3 and even Y2O3, which is already present in the substrate and hence simplifies the system further. We have extended our work to include YSZ substrates of the (110) surface orientation and have found that instead of nanoisland arrays, we obtain an array of parallel nanobars which have their long axes oriented along the [1-10] direction on the YSZ-(110) surface. STEM EDS performed on both the bars and the nanoislands has revealed that they are solid YSZ-rich solid solutions with the dopant species and are heterogeneous in composition with dopant enrichment at the top of the islands (bars) while their bases are pure YSZ. Finally, we discuss some of the future work directions and possible applications of these nanostructures that are being explored in collaboration with our colleagues Kunal Parikh and Prof. Jessica O. Winter in the Dept. of Chemical and Biomolecular Engineering and Michael Susner and Prof. Michael Sumption in the Dept. of Materials Science and Engineering. Advisors/Committee Members: Akbar, Sheikh.

▼ Structural materials show novel properties with decreasing microstructural feature size, particularly in the nanocrystalline regime to the amorphous limit of grain size reduction. These materials show remarkable yield strength, radiation hardness, and fatigue resistance. At the amorphous limit, they also show strong corrosion resistance. Combining the two elements together to form a composite, these materials show exceptional yield strength and ductility, to significantly improve toughness that is often compromised in high-strength materials. However, plastic deformation at these length scales is difficult to characterize and novel mechanisms for deformation and failure operate in these system. Although some progress has been achieved at visualizing deformation of nanocrystalline materials in-situ, the operating mechanisms are still under active debate. Further, classical continuum-based theories break down at the length of the dislocation core and grain boundary. However, this problem is ideally suited to atomistic simulation using molecular dynamics (MD) to characterize the energetics of dislocation-grain boundary interaction. Two limitations of the MD method, however, are the extreme strain rates and uncertainty regarding thermal activation. To bypass both of these, the approach here is not to focus on dislocation nucleation from a grain boundary under applied stress/strain, but rather the energetics of interaction and the potential barrier or trapping strength of athermal dislocation-grain boundary interactions. Studying the interaction between asymmetric tilt boundary and inclined dislocation, a position dependent response and strong resistance to step creation are observed. Also, a metastable higher energy grain boundary undergoes long-range reordering that reduces system energy in excess of the input dislocation line energy. Nanocrystalline materials are observed to nucleate partial dislocations from grain boundaries, leading to deformation twinning. The energetics of this process is still under discussion. Because continuum theories break down in the region of the dislocation core and grain boundary plane, we must study the interaction energy of these systems using atomistic calculations. Due to the very high strain rates produced in classical MD calculations, the nucleation process is not accessible via this route. Here, we instead focus on a series of energy minimizing atomistic calculations between a low-symmetry boundary and a general dislocation to study the energetics of interaction, the nature of grain boundaries under stress, and what leads to dislocation pile-ups at grain boundaries. We find that boundaries resist dislocations with Burgers vectors normal to the boundary and a strong position dependence within the grain boundary sub-structure. At the amorphous limit, bulk metallic glasses fail by intensely localized shear bands with very little macroscopic ductility, which limits structural service. It is understood through experiment and thermodynamics that a defect of volume from one to several vacancies plays a critical role in initiating and propagating shear, but not the structure of this defect or how it operates. Here we use atomistic modeling and electron density calculations to study defect distribution and migration in a metallic glass, and evolution under shear. We find a critical stress threshold to activate shear banding, and an enhancement of shear localization in the neighborhood of a dissolving void. Advisors/Committee Members: Flores, Prof. Katharine.

▼ Higher performance and durability requirements of gas-turbine engines will require a new generation of thermal barrier coatings (TBCs). This is particularly true of engines operated at higher temperatures, where TBCs are subjected to attack by CaO-MgO-Al2O3-SiO2 (CMAS) glassy deposits. In this work, a new approach for mitigating CMAS attack on TBCs is introduced, where up to 20 mol% Al2O3 and 5 mol% TiO2 in the form of a solid solution is incorporated into Y2O3-stabilized ZrO2 (YSZ) TBCs. The fabrication of such TBCs with engineered chemistries is made possible by the solution-precursor plasma spray (SPPS) process, which is uniquely suited for depositing coatings of metastable ceramics with extended solid-solubilities. In the current work, the TBC serves as a reservoir of Al and Ti solutes, which are incorporated into the molten CMAS glass that is in contact with the TBC. An accumulation of Al concentration in the CMAS glass as it penetrates the TBC shifts the glass composition from the difficult-to-crystallize psuedowollastonite field to the easy-to-crystallize anorthite field. The incorporation of Ti in the glass promotes crystallization of the CMAS glass by serving as a nucleating agent. This combined effect results in the near-complete crystallization of the leading edge of the CMAS front into anorthite, essentially arresting the front. Both of these phenomena will help crystallize the CMAS glass, making it immobile and ineffective in penetrating the TBC. It is shown that incorporation of both Al and Ti in the CMAS glass is essential for this approach to be effective. In this dissertation, results from characterization and testing of these new TBCs are presented, together with a discussion of mechanisms responsible for CMAS-attack mitigation. The penetration of CMAS causes a loss of strain tolerance of the coating. Delamination maps are used to demonstrate the combined effects of CMAS penetration, temperature gradient and cooling inhomogeneity on the coating. Evans and Hutchinsons model has been used to produce delamination maps and predict the durability of novel TBCs. Advisors/Committee Members: Padture, Nitin.

▼ High velocity sheet metal forming was the focus of much research approximately 30 years ago, but has become somewhat dormant since then. Initially developed for fabricating big components, their use continued for the manufacture of small components of intricate shapes as well. However, their potential to improve formability was poorly understood.In the current investigation, electrohydraulic forming was used to generate high workpiece velocity. Electrical energy up to 50 kJ was stored in a bank of capacitors and was discharged rapidly across a pair of electrodes under water. This generated a high intensity shock wave which forced the sheet metal into a die. Comparative conventional forming was done by pressurizing the oil beneath the workpiece with a hydraulic pump.Formability increase of over 400% was observed on 6061 T4 Aluminum sheets and an increase of over 150% on interstitial free iron and OFHC copper. Limit strains were directly measured on the samples. Forming Limit Diagrams were used to compare low and high rate results. The estimated metal velocities at high rates ranged from 130 m/s to 350 m/s and the strain rates were of the order of 101s-1.At such energies and time scales, inertial forces act to diffuse the deformation away from the localized areas, thus postponing the failure. Compressive stresses generated by the impact of sheet with the die wall also increase the ductility. However, the strain rate appeared to be still too low to cause adiabatic localization that might have limited the formability. The estimated velocity was also much lower than the von Kármán;n critical velocity for localization. The terminal hardness of finished components differed from the initial hardness within 20%. Thus, the material constitutive behavior did not appear to have changed.Increased formability makes these high rate processes competitive to traditional superplastic forming. In order to distinguish this effect from superplasticity, the current phenomenon is termed hyperplasticity. Advisors/Committee Members: Daehn, Glenn S.

▼ Electromagnetic forming is a possible alternative to sheet metal stamping. There are multiple limitations to the incumbent stamping methods including: complex alignment, changes to component shapes, and ductility issues, which often limits available formed geometry. Electromagnetic forming allows for the avoidance of some of these issues, but introduces a few other issues.In this thesis, the issues with electromagnetic forming will be discussed in conjunction with the application of the uniform pressure coil. Also, the effects on properties of the electromagnetically formed samples in comparison to the traditional samples will be presented. These properties include hardness, formability and interface issues. Lastly, discussed in this paper is the implementation of the Photon Doppler Velocimetry (PDV) system, a velocity measurement system used to determine the velocity of the workpiece and compare it to physics-based models of the process. Advisors/Committee Members: Daehn, Glenn.

▼ Titanium and its alloys are comparatively recent newcomers to the metallurgical market. They are gaining widespread acceptance for use in the recreational, aerospace, biomedical, petro-chemical, and commercial processing industries due to their combination of unique and advantageous properties, including high strength, low density, and superior corrosion resistance to most aggressive agents. The material properties of titanium and its alloys can be optimized and tailored by engineering the microstructure via control of chemistry, processing route, and heat treatment. The morphology of the two crystallographic allotropic phases can be manipulated to refine the structure and produce desirable mechanical property combinations. Microstructural constitution of the titanium alloys is classified according to the dominant phase within the alloy; alpha + beta (α + β) titanium alloys are the most widely used alloys. The temperature of the final heat treatment of the α/β components is governed by the service requirements. In order to evaluate the behavior of these alloys for future applications, it is imperative that the microstructural features and characteristics be quantified and examined on a spatial dimension. The Robo-Met.3D is a high precision robotic serial sectioning device that can fulfill this need. Initially, several months were spent resolving problems with the functioning of the Robo.Met.3D. Two-dimensional (2-D) stereology was done on Timetal 550 using automated batch processing with Adobe Photoshop and Fovea Pro. Images from different locations on the gage were obtained and compared. Final data demonstrated quantitative differences which were the result of the heat treatment. Discrepancies and inconsistencies in the data were identified as limiting factors in the reproducibility of the procedure in future work. Serial sectioning using focused ion beam (FIB) was performed using Timetal 550, and three-dimensional (3-D) reconstruction was done using IMOD. Robo-Met.3D procedures and algorithms were identified for serial sectioning collection for titanium alloys using Ti-6Al-4V. Recommendations for future work include developing more efficient procedures for coloring in the microstructural features in the Adobe Photoshop CS™. A new procedure is needed to mount and polish the sample to prevent sample curvature due to the polishing step. Also, the small size of the secondary alpha (α) presents a challenge when examining microstructural features; however, it is imperative that these features be examined in the future to determine their effect on mechanical properties. Advisors/Committee Members: Fraser, Hamish.

▼ Solid state gas sensors are widely used to detect trace levels of different gases. Micro-hotplates provide a robust substrate for miniaturized solid state sensors; however, no well-established technique exists for depositing metal oxide particles directly onto these microhotplates. A novel picoliter drop deposition technique based on ink-jet printing was developed here to achieve accurate, precise drop placement and reproducible film formation.Oxide nanoparticles are the key to creating high performance gas sensors; therefore, syn-thetic techniques capable of controlling the size, shape, composition and microstructure were explored. Tin oxide nanoparticles were synthesized using a hydrothermal route able to produce crystalline nanoparticles. The diameter of these particles was controlled be-tween 3 and 10 nm by controlling the reaction temperature. Templating of tin oxide nanoparticles into hollow microshells with spherical pores incorporated into the micro-structure was pursued using various assembly techniques. Hollow spheres with a pore diameter of 100 nm – 1 μm with a thin nanostructured shell were realized. In addition a new solvothermal synthesis technique for making crystalline nickel oxide nanoparticles with diameters in the range of 5 – 6nm was devised during the course of this research. Optimal synthesis conditions and the reaction mechanism were elucidated. The picoliter drop deposition technique was utilized for the first time here to fabricate oxide nanoparticle-based films directly on microhotplate substrates. Tin oxide nanoparti-cle-laden suspensions were formulated to reduce the working agglomerate size, enable consistent first drop formation and prevent nozzle clogging. A priming technique was introduced to enable accurate, repeatable drop deposition onto the microsensor platforms. A dynamic drop placement methodology was also developed which resulted in a uniform oxide nanoparticle film formation on microhotplate. Microsensors made using the picoliter drop deposited tin oxide nanoparticles showed high gas sensor responses to reducing gases at concentrations levels of 10 – 100 parts per million. The sensing results verify that the combination of nanoparticles synthesized in this work and the novel picoliter drop deposition technique provides a versatile new path-way for gas microsensor fabrication. Advisors/Committee Members: Morris, Patricia.

▼ Flexible and electrically insulating alumina coatings, on metallic lead wires for high temperature applications, have been made using Electrophoretic Deposition (EPD) and Slurry coating techniques.Experiments were done to study the effect of deposition current, solids content and alumina particle size on the EPD of alumina from aqueous suspensions. The alumina was dispersed using an ammonium salt of polyacrylic acid. A current of 1mA, solids content of 40 wt.%, 0.23 μm alumina particle size gave the best results for aqueous EPD. A pH of 8.17 and conductivity of 16280 X10-6(Ω.cm)-1 gave the highest rate of deposition. However the overall effect of the pH and conductivity is not completely understood. The effect of sintering temperature, coating thickness, wire thickness, wire material and its surface condition on adherence of the coating were also studied. The various parameters were qualitatively observed by studying the variation of "crack density" due to these parameters. Adherence was observed to be better for thinner coatings and for coatings made on thinner wires. Adherence also increased with sintering temperature. Heat treatments were done in air and vacuum to modify the surface condition. The heat treatments increased the surface roughness of the wires, which increased the crack density. Platinum showed the least effect due to the heat treatments, while chromel-p showed the maximum. Aqueous EPD samples showed and pores. Non-aqueous EPD was also studied. Coatings were made from suspensions in mixture of ethanol and acetone. They did not have any drying cracks and showed only transverse cracks on sintering. The rate of coating is also much higher than in the case of aqueous EPD. In the case of slurry coating, a machine was successfully built to continuously slurry coat metal wires. A slurry based on MEK and toluene, as dispersing liquids, and Butvar B-79 binder was used. An optimal slurry composition and rate of drawing were determined. The resistivity of the coatings, made by both EPD and slurry coating was found to be of the order of 108 ohm-cm, even at 1050°C. Advisors/Committee Members: Kreidler, Eric.

▼ Metallic glasses are metallic alloy systems with disordered atomic structure. Due to their unique amorphous structure, they exhibit an extraordinary set of properties that are ideal for a wide variety of applications ranging from electrical transformers, armor-piercing projectiles, sporting goods and fuel cells to precision gears for micromotors. In particular, owing to their exceptional mechanical properties like near-theoretical strength (1-3 GPa), large elastic strain range (2-3%), and unusual formability above the glass transition temperature, metallic glasses have tremendous potential in structural applications. Unfortunately, their unique structure also gives rise to significant limitations, such as limited ductility at room temperature due to rapid localization of plastic flow in shear bands. However, when the test volumes approach the size of a shear band nucleus (~50-500 nm), it is believed that shear band formation and propagation can be constrained, leading to enhanced plasticity and failure strength. This study investigates the phenomenon of strain localization using both experimental and computational techniques. On the experimental front, sample size effects on strength, plasticity and deformation modes were explored in a Zr-based bulk metallic glass via micron- and sub-micron scale compression testing. Specimens with diameters ranging from 200 nm to a few microns were fabricated using Focused Ion Beam technique and were tested under uniaxial compression in a nanoindentation set-up with a flat punch tip. Effect of extrinsic factors like specimen geometry and machine stiffness on deformation behavior was discussed. Shear banding was shown to be more stable at this length scale than in macro-scale testing because of a smaller specimen to load frame stiffness ratio. It was found that as the specimen size is reduced to below 300 nm, the deformation mode changes from being discrete and inhomogeneous to more continuous flow including both localized and non-localized contributions at low strains. Moreover, the magnitude of strain bursts was found to decrease with decrease in specimen size. Furthermore, Weibull statistical analysis was performed to investigate the effect of specimen size on yield strength in this metallic glass. It was revealed that the dispersion in strengths increases dramatically with decrease in sample size, attributed to the size distribution of the defects responsible for shear banding. The findings are crucial in designing systems which promote plasticity in metallic glasses by suppressing the shear-band instability and also in direct application of these materials for structural purposes as small components in micro- and nano-scale systems. On the computational front, Molecular Dynamics (MD) simulations have been employed to generate Zr-Cu metallic glass structures. In order to analyze and better understand and visualize the concepts of “free” volume and flow defects in metallic glasses, an electron density model was developed as an upgrade to the traditional hard sphere approaches. Simple tension and shear modes of deformation were simulated using MD in Zr-Cu system, and role of open volume in deformation was studied using the electron density model. In uniaxial tension simulations, effect of temperature and deformation rate is examined, and the process of accumulation of free volume to the point of catastrophic failure is visualized using the Electron Density model. In shear simulations, we find that the as-quenched glass structures undergo homogeneous deformation and do not exhibit any strain localization. However, it is found that by incorporating a cylindrical void in the glass structure as a source of “free” volume, it is possible to induce strain localization. It was found that a critical void diameter of 8 Angstroms was required to successfully initialize strain localization in this system. Advisors/Committee Members: Flores, Katharine.

▼ The basic formation of in-situ MgB2, and how variations in the formation process influence the electrical and magnetic properties of this material was studied. Bulk MgB2 samples were prepared by stoichiometric, elemental powder mixing and compaction while strand samples were prepared by a modified PIT with subsequent reaction. The influence of various reaction schedules on the formation reaction was studied. Two different optimum reaction-temperature windows were indentified, namely, low-temperature heat-treatment (<650°C) and high-temperature heat-treatment (>650°C) for the preparation of MgB2. XRD was used to confirm phase formation and microstructural variations were studied with the help of SEM. Following this, the focus turned to critical field enhancement via doping with various compounds targeting either Mg or B sites. Large increases in irreversibility field, and upper critical field, of bulk and strand superconducting MgB2 were achieved by separately adding SiC, amorphous C, and selected metal diborides (NaB2, ZrB2, TiB2) in bulk samples and different sizes of SiC (~200 nm, 30 nm and 15 nm) in strand samples. Lattice spacing shifts and resistivity measurements were consistent with dopant introduction to the lattice. The increases in the Bc2 were also complimented by an increase in the transport Jcs, especially for the SiC doped samples. Flux pinning analysis performed on SiC doped samples showed that while some small level of particulate-enhanced pinning was present, the majority of the pinning was associated with a grain boundary mechanism, suggesting that transport Jc increases were predominantly Bc2 related. Normal-state resistivities were measured for various binary and doped MgB2 samples as a function of temperature and were modeled based on the Bloch-Gruneissen equations. This allowed extraction of the residual resistivities, Debye temperatures and current carrying volume fractions, well as providing information on the electron-phonon coupling constant. The residual resistivity was found to increase by a factor of three, Debye temperature decreased and the electron-phonon coupling constant increased marginally for the SiC doped samples as compared to the binary sample. This change in ρ0 and θD confirmed the XRD evidence that the dopants were increasing µoHirr and Bc2 by substituting on the B and Mg sites of the crystalline lattice. Advisors/Committee Members: Dregia, Suliman A.

▼ In the present work multifilamentary Tube type and distributed barrier Rod-in-Tube (RIT) type Nb3Sn composites were studied in detail. Tube type composites consisting of subelements of Nb-7.5 wt% Ta alloys with simple Cu/Sn binary metal inserts were studied in an attempt to enhance the performance boundaries of these conductors. We focused on correlating the composition and morphology of the intermetallic A15 to the transport and magnetic properties for varying heat treatments. In particular, lower temperature HTs were studied, specifically 625°C and 635°C as a function of time. The extent of A15 formation, the ratio of the coarse/fine grain areas, and the amount of untransformed 6:5 phase were then observed as a function of time –temperature. A15 stoichiometry was investigated and compared for two different Nb3Sn strand designs, specifically Tube type and high performance RIT type strands. Transport measurements were performed on both categories of conductors for various conditions. The objective of the study was to investigate the limits of tube type conductor performance and to compare this to that of RIT conductors. Specifically, the Sn stoichiometry and A15 grain size for RIT and Tube type conductors were compared and to correlated with the transport properties of the two strand types. Tube type conductors were compared to RIT conductors, each after the application of single step and two-step HTs with plateaus ranging from 615°C to 675°C for various times. The influence of strand geometry and reaction route was related to the resulting A15 stoichiometries. Fractography was performed to investigate the effect of a two-step reaction on the morphology, the ratio of coarse/fine grain area and grain size of fine grain A15. The effect of Ti doping on superconducting properties of RIT type Nb3Sn strands was also studied. Advisors/Committee Members: Sumption, Micheal.

▼ The relationship between the texture and the microstructure of both beta processed and alpha/beta processed Ti alloys has been examined in this study. In the beta-processed microstructures, it has been shown that two sets of alpha colonies sharing a common {0001} plane and rotated by ~10.5° from each other may have growth directions which have very large angles of about ~80.7° between them. Moreover, it was observed that alpha laths growing from certain prior beta grain boundaries sometimes shared common basal planes. In some special cases, the alpha laths growing into two different prior beta grains from the grain boundary between them had almost exactly the same orientation, although they had vastly different growth directions. Additionally, there were some cases in which alpha laths growing into different prior beta grains not only had the same crystallographic orientation, but also had the same growth direction. Scanning Electron Microscopy (SEM), Orientation Imaging Microscopy (OIM) and Transmission Electron Microscopy (TEM) have been used to investigate these phenomena and the existing theories of growth directions have been used in conjunction with the results obtained to explain them. In alpha/beta-processed alloys, the phenomenon of globularization of alpha laths breaks down the beta-processed microstructure and modifies the texture of these alloys. Samples of Ti-6Al-4V having colony and basketweave microstructures were hot deformed in the high alpha/beta temperature range and their microstructure and microtexture were examined by the use of SEM and OIM. It is shown that the samples which had a colony microstructure had greater “clustering” of grains with similar orientations than those having a basketweave microstructure. The mode of transformation on heating from the alpha to the beta phase was investigated by measuring the texture of both phases at different temperatures, in situ, using the HIPPO instrument at the Los Alamos Neutron Science Center. A comparison of the pole figures for both phases has allowed an insight into the mode of transformation of the alpha to the beta phase, and it appears that the beta phase forms by the growth of the preexisting beta, and not by fresh nucleation of beta in the alpha phase. Advisors/Committee Members: Fraser, Hamish L.

▼ Biofilms are communities of bacteria that live on substrates. In many cases, biofilms have a detrimental effect on the surfaces to which they adhere. For example, biofilms negatively impact dental enamel, metals, and polymers. One of many challenges in biofilm research is analyzing the interface between a biofilm and its substrate, to determine how the biofilm adheres to, and affects, the substrate. This analysis can be challenging due to the limitations of methods such as SEM, confocal microscopy and ultramicrotomy. Indeed, SEM can provide topographical information of the sample’s surface but is fundamentally a two dimensional technique. Confocal microscopy is a useful tool for characterizing large areas of the biofilm but lacks the spatial resolution that TEM provides. Traditionally, ultramicrotomy has been the preferred method of sample preparation for TEM analysis of biological systems despite the lack of site-specificity and compromising effect on structure from mechanical sectioning with a knife. Recently, dual-beam focused ion beam (DB-FIB) technology has revolutionized the field of metals, ceramics, semiconductor materials, and TEM sample preparation; however this technique has not been widely applied to soft material characterization largely due to rapid degradation of these materials under ion beam irradiation. In this study, we have developed novel processes to reduce ion damage to the biofilm during milling, and successfully used DB-FIB to analyze Pseudomonas fluorescens biofilms and the interface between the biofilm and its substrate, polyurethane paint. An FEI Helios Nanolab 600 DB-FIB was used to create serial section data sets for three-dimensional reconstruction and also to excise thin foils for (Scanning) Transmission Electron Microscopy ((S)TEM) and Electron Energy Loss Spectroscopy (EELS). Serial sectioning of the biofilm provided novel structural information about surface adherence mechanisms of the bacteria. (S)TEM imaging and EELS characterization was applied to investigate compositional differences across the interface as well as view the internal structures (or inclusions) of the bacteria. Advisors/Committee Members: Fraser, Hamish.

▼ The beneficial influence of boron additions on processing, microstructure, physical and mechanical properties of various titanium alloys has been recognized since 1950’s. However, boron additions to titanium alloys to obtain specific microstructures and mechanical properties for several niche applications, including automotive and aerospace, have been actively studied during the past 25 years. The addition of boron concentrations greater than 0.05 wt.% to titanium alloys creates a dispersion of TiB. The presence of TiB enhances the tensile and fatigue strengths as well as the wear resistance as compared to the original titanium alloy. Although these improvements in mechanical properties are attractive, there are still two major obstacles in using these alloys: (1) relationship of microstructure and mechanical properties in Ti-B alloys needs further investigation to optimize the alloys for specific commercial applications; and (2) cost to benefit ratio of producing these alloys is high for a given application(s). The Armstrong process is a novel process that can produce commercially pure (CP) titanium and titanium alloy powder directly from TiCl4 (and other metal halides or as required, to obtain the desired alloy composition). The Armstrong process uses sodium as a reducing agent, with similar reactions as the Hunter process using sodium as a reducing agent and Kroll process using magnesium as a reducing agent. The Armstrong process forms CP-Ti and titanium alloyed powder, which can be directly consolidated or melted into the final product. In comparing the downstream processing steps required by the Kroll and Hunter processes with direct consolidation of Armstrong powder, several processing features or steps are eliminated: (1) restriction of batch processing of material, (2) blending of titanium sponge and master alloy material to create titanium alloys, (3) crushing of the sponge product, (4) melting, and (5) several handling steps. The main objective of this research was to characterize structure and properties of CP-Ti and Ti-B alloys produced by the Armstrong process. Particular emphasis has been placed on improved understanding of the strengthening mechanisms associated with the addition of boron to titanium alloys. Advisors/Committee Members: Williams, James C.

▼ Efforts to develop automotive glasses with enhanced solar control characteristics have been motivated by the desire for increased consumer comfort, reduced air-conditioning loads, and improved fuel-economy associated with a reduction in the total solar energy transmitted into the automotive interior. Prior innovations in solar control automotive glasses resulted from rather limited compositional modifications to the base soda-lime-silicate glass and neglected to exploit a multitude of non-linear mechanistic interactions among batch constituents resulting in only marginal improvements in solar-control properties.In the current investigation, the base soda-lime-silicate glass (72.7 wt.% Si02, 14.2% Na2O,10.0% CaO, 2.5% MgO, 0.6% Al2O3 with 0.3 Na2SO4 added to the batch as a fining agent) was modified with Fe2O3 (0.0 to 0.8 wt. %), NiO (0.0 to 0.15 wt. %), CoO (0.0 to 0.15 wt. %), V2O5 (0.0 to 0.225 wt. %), TiO2 (0.0 to 1.5 wt. %), SnO (0.0 to 3.0 wt. %), ZnS (0.0 to 0.09 wt. %), ZnO (0.0 to 2.0 wt.), CaF2 (0.0 to 2.0 wt. %), and P2O5 (0.0 to 2.0 wt. %) to exploit reported non-linear mechanistic interactions among the dopants by which the solar-control characteristics of the base glass can be modified. Dopants to the base glass were classified as primary dopants (Fe203, NiO, CoO, V205); that is dopants which in-of-themselves manifest absorption in some portion of the solar spectrum, and as interactive dopants (Ti02, SnO, ZnS, ZnO, CaF2, P205) which in-of-themselves do not manifest absorption bands but which, nonetheless affect the absorption characteristics of the primary dopants.Prior investigations have utilized a Lambert-Beer Law of Absorption to relate the molar concentration of dopants in the glass to the observed transmission (optical density) characteristics of the doped glass. In complex glass systems, interactions among batch constituents in which the redox state of transition metal oxide additions are modified during the melting process, results in the inability to predict the exact redox state and molar concentrations of the transition metal oxides in the final glass upon cooling below the glass transition temperature. In this case, only the initial batch composition and redox state can be controlled by the experimenter and for this reason, the batch composition was properly chosen as the independent variables in the current investigation. A modified Lambert-Beer law of absorption utilizing the linear and quadratic levels of the primary dopants, and a series of binary and ternary interaction terms among primary and interactive dopants was utilized to related batch composition to the observed optical densities in the current investigation.In addition to the specification of the independent variables based on batch composition, proper definition of the dependent variable to be modeled was of critical importance. Initially, the calculated solar and visible transmittances obtained from the measured solar transmission curves were the obvious choices for the dependent variables to be modeled. However, the solar and visible transmittances were realized to integrated response variables; that is the value of the solar and visible transmittances were functionals of the transmission curve. Integrated responses variables are problematic to model as infinitely many transmission curves can have the same integrated response value. For this reason, the optical density at each wavelength necessary for the calculation of solar and visible transmittances were selected as the discrete response variables from which the integrated response values could be obtained.Upon specification of the independent and dependent variables, as well as the model form relating the two, an experimental design strategy was specified. For simple systems with few independent variables and limited interactions among these variables, one-at-a-time or full-factorial design strategies are often utilized. However, in complex systems, full or fractional-factorial design strategies are impractical due to the sheer number of experiments which must be conducted. For this reason, computer-assisted D-Optimal Design Of Experiments (DOE) utilizing the Haller Information Technology Systems (HITS®) software package was utilized to design an initial set of 64 experimental melts by which the optical response of the system could be modeled based on batch composition.500 gram batches in which the dopants were added as reagent grade Fe2O3, NiO, Co(NO3)2-6H20, V2O5, TiO2, SnO, ZnS, ZnO, CaF2 and CaHPO4 were melted under ambient conditions of pO2 at 1500°C for 3hrs in platinum crucibles by INTEGREX® Product Testing Systems. The melts were then cast into stainless steel molds of approximate diameter of 1.5" at a variable thickness of 0.33 to 0.5" generating 3 ingots per melt. The glass ingots were subsequently annealed at 621°C for 1 hour and then furnace cooled to room temperature in order to eradicate any effects of thermal history. The ingots were then inspected for optical homogeneity and shipped to S&S Optical for grinding to a fixed thickness of 3.3 ± 0.1 mm with a surface finish corresponding to 80/40 as specified by the ANSI standard. The samples were then measured for total solar transmittance in accordance with the specifications of ASTM E903-96 by Optical Data Associates utilizing a Cary-500 UV/VIS/NIR spectrophotometer equipped with an integrating sphere. CIE-LAB color and visible transmittance were measured in accordance with the specification of ASTM E308. Inductively Coupled Plasma Spectrophotometry was utilized to measure the concentration of V2O5, CoO, NiO, Fe203, P205, TiO2, SnO, and ZnO in each sample. F‾ concentration was determined by utilizing an ion selective electrode. Total sulfur as SO3 was determined utilizing a LECO®; combustion-titration instrument. Iron redox state was determined by measurement of the ferrous iron content utilizing HF/H2SO4 sample digestion followed by colorimetric measurement of the Fe(ll)-Phenanthroline complex at 510nm.Multiple linear correlation analysis utilizing the HITS® Multiple Correlation Analysis (MCA) software module was performed correlating the measured optical density at each wavelength necessary for the calculation of solar and visible transmittances to the batch composition. The MCA module utilizes a Root Mean Square Deviation (RMSD) or Sy.x as the measure of the fit of the model to the data set as the model selection criterion. Model terms were selected based on the t-value which established the statistical significance of the correlation between the specified model term and the data set. It was found that model specification was complicated by co-linearity among model terms which provided for no clear choice of optimum model among competing models with comparable fit. For this reason, a model selection algorithm designated as the exchange method was utilized to select a self-consistent series of model across the wavelength range. The algorithm consisted of a forward selection mode in which model terms were brought into the model in order of decreasing t-value until no terms remained with a t-value greater than or equal to 2 indicating the term is statistically significant at a confidence level of 95%. At this point, due to co-linearity among the selected model terms, some terms that had been brought into the model have altered their t-values upon the inclusion of latter model terms to a t-value less than two. At this point, a backwards elimination mode was initiated in which all terms included in the model which have t-values less than two were deleted from the model. This process then alternated between the forward selection and backwards elimination mode until the model converges to a solution in which all terms included in the model are statistically significant at a 95% confidence level and all terms excluded from the model are statistically insignificant at a 95% confidence level. Once this algorithm had generated models for every wavelength necessary for the calculation of solar and visible transmittances, the full model was selected by taking the subset of all model terms which seemed to be consistently significant in some part of the full wavelength range.Utilizing the developed model, exceptional fit in terms of both the discrete response (the transmission curves) and the integrated response (visible and solar transmittance, L, a*, b* color coordinates) were realized. Validation experiments have established the utility of the models towards optimization of solar control automotive glasses. Glasses utilizing Fe2O3, CoO, NiO, V2O5, ZnO and P2O5 have generated innovative solar control glasses with substantial improvements in solar control characteristics. This glass compares quite favorably with PPG Gl-20 solar control van glass which has a visible transmittance of 24%, a solar transmittance of 23%, and a solar-IR transmittance of 18% at 3.3mm. The glass produced in the current investigation has the required visible transmittance of 24% but with substantially lower values of both solar and solar-IR transmittance of 11 % and 3% respectively. In addition to the models utility in optimization of solar control glasses, the model also provides insight into the mechanistic interactions which are operable in solar control glasses. Advisors/Committee Members: Drummond, Charles H.

▼ Modeling sheet metal forming operations requires understanding of the plastic behavior of sheet alloys along complex strain paths. In most materials, plastic deformation in one direction will affect subsequent deformation in another direction. For one-dimensional deformation, this phenomenon is known as the Bauschinger effect. The Bauschinger effect in heat-treatable aluminum alloys is heavily dependent on the presence of hardening precipitates. A new method was developed to test sheet materials under uniaxial reversed loading to compressive strains greater than 0.20. Studies of commercial aluminum alloys 2524 and 6013, show a larger Bauschinger effect for materials that have been artificially aged past peak strength, where the precipitates are semi-coherent or incoherent. Not only is there a substantial reduction of the reverse yield stress, the period of the transient behavior following the load reversal is also lengthened. This effect is seen in over-aged materials after prestrains as small as 0.4%. Materials with less aging had shorter transient periods but showed some permanent softening after the reversal, which was a function of the prestrain. A constitutive model was developed, based upon the nonlinear kinematic hardening model, which is capable of describing the reduction in the reverse yield stress and the hardening transient observed after the reversal. The ability to model the permanent offset was introduced by the addition of a new term into the formulation, which can be defined as a function of plastic strain. This model has been implemented in a finite element program that successfully reproduces the main features of the experimental results for both uniaxial and more general strain paths. Simulations of draw-bead forces showed significant reductions, up to 25%, because of the Bauschinger effect. Simulations of the springback angle using the new model were within 1.5 degrees of the experimental results for the 2524 over-aged and peak-aged tempers that show a large Bauschinger effect. Advisors/Committee Members: Wagoner, Robert H.

▼ In this work, the low temperature synthesis of MgB2 from Mg/B and MgH2/B powder mixtures was studied using Differential Scanning Calorimetry (DSC). For the Mg/B powder mixture, two exothermic reaction events were observed and the first reaction event was initiated by the decomposition of Mg(OH)2 on the surface of the magnesium powder. For the MgH2/B powder mixture, there was an endothermic event at ~375 °C (the decomposition of MgH2 into H2 and Mg) and an exothermic event ~600 °C (the reaction of Mg and B). The Kissinger analysis method was used to estimate the apparent activation energy of the Mg and B reaction using DSC data with different furnace ramp rates. The limitations of MgB2 low temperature synthesis led to the development of a high pressure induction furnace that was constructed using a pressure vessel and an induction heating power supply. The purpose was to not only synthesize more homogeneous MgB2 samples, but also to determine whether MgB2 melts congruently or incongruently. A custom implementation of the Smith Thermal Analysis method was developed and tested on aluminum and AlB2, the closest analogue to MgB2. Measurements on MgB2 powder and a high purity Mg/B elemental mixture confirmed that MgB2 melts incongruently and decomposes into a liquid and MgB4 at ~1445 °C at 10 MPa via peritectic decomposition. Another measurement using a Mg/B elemental mixture with impure boron suggested that ~0.7 wt% carbon impurity in the boron raised the incongruent melting temperature to ~1490-1500 °C. Lastly, the solubility limit for carbon in MgB2 was studied by making samples from B4C and Mg at 1530 °C, 1600 °C and 1700 °C in the high pressure furnace. All three samples had three phases: Mg, MgB2C2, and carbon doped MgB2. The MgB2C2 and carbon doped MgB2 grain size increased with temperature and the 1700 °C sample had needle-like grains for both phases. The presence of the ternary phase, MgB2C2, suggested that the maximum doping limit for carbon in MgB2 had been reached. The 1530 °C sample was characterized by Electron Probe Microanalysis at the University of Oregon and the average carbon concentration was estimated to be ~5.9 at%. Further investigation using TEM found MgO inclusions in the 1530 °C sample which were not detected with X-ray diffraction. Advisors/Committee Members: Sumption, Michael.

▼ Molecular electronics is the study of charge transport through single molecules or molecular ensembles. Molecular electronic junctions consist of single molecules or an ensemble of molecules positioned between two conducing contacts. To fabricate and measure the electronic properties of molecular junctions, several techniques have been employed such as scanning tunneling microscopy, conducting probe atomic force microscopy, and vapor deposition of top contacts. Charge transport observed through molecular junctions has been shown to exhibit technologically important phenomena such as rectification, conductance switching, and orbital gating. The primary focus of the field of molecular electronics is to understand the effect of molecular properties, such as structure and molecular orbitals, on charge transport mechanisms through molecular junctions. In this dissertation, the various techniques to fabricate and characterize molecular junctions are discussed, along with an introduction to charge transport mechanisms expected to control transport through molecular junctions. More specifically, this dissertation is primary focused on the fabrication and characterization of molecular junctions fabricated through the formation of an electronic contact on a molecular layer through physical vapor deposition. A common problem with this technique is structural damage to the molecular layer or metal penetration through the molecular layer during the contact formation. To overcome these limitations, a novel fabrication technique was developed and employed to fabricate reproducible molecular junctions through a physical vapor deposition technique without molecular damage or metal penetration. Termed surface diffusion mediated deposition (SDMD), the technique remotely deposits a metallic contact adjacent to and about 10 – 100 nm away from the molecular layer. Surface diffusion causes the metallic contact to migrate towards and onto the molecular layer to form an electronic contact. With SDMD, single molecule and many-molecule junctions are fabricated and electronically characterized. To probe electronic states and molecular structure in molecule/oxide junctions, an in-situ optical absorbance spectroscopy technique was developed and employed to monitor bias induced molecular redox events in solid-state molecular junctions. Correlation of the observed spectral changes with molecular redox events allows characterization of the electronic properties of molecules which are critical in understanding charge transport through molecules. In a related application, the developed in-situ optical absorbance spectroscopy technique was used to probe doping events in polypyrrole/oxide junctions. Doping reactions in polypyrrole are shown to strongly depend on the surrounding environment. For application to both molecular and conjugated polymer junctions, in-situ absorbance spectroscopy is shown to be a useful analytical tool to determine charge transport mechanisms. Finally, a thermal oxidation technique is introduced to increase the resolution of nanoimprint lithography to fabricate nanogap electrodes for molecular junctions. The advantage of this technique is the ability to use a simple, fast, and reliable oxidation process to increase the resolution of standard nanofabrication techniques. Advisors/Committee Members: Frankel, Gerald S.

▼ Recently, phase field computational models have been developed as tools for the investigation of microstructure formation at length scales, time scales, and levels of detail which would otherwise be unaccessible. These robust simulation techniques have been applied to study a wide variety of problems. This work details their application to the study of surface roughening and irradiation hardening. First, motivated by recently reported experimental observations, the role of coupling between elasticity, composition and structure during island formation in thin films will be examined. Localized stress centers (such as those caused by morphological or chemical variations in a film) and their influence on the formation of surface islands by the Asaro-Tiller-Grinfeld instability is investigated via the phase field method. Such stress centers are found to cause a wave-like marching of surface islands away from the centers during their formation, and these results are compared to experimental observation. Coupling between composition and stress is then introduced, and its influence on surface evolution and morphology is studied. Secondly, a phase field model of dislocations is employed to examine irradiation hardening in Zirconium. After neutron irradiation Zr-alloys develop a characteristic damage microstructure: populations of dislocation loops with -type Burgers vectors, believed to be formed by the condensation of point defects. These defect populations have been correlated with significant hardening in Zr-alloys, while the specific causal mechanism remains unknown. The phase field method is employed in order to study the specifics of damage loop and glide dislocation interaction at the mesoscale. Long-range elastic interactions are considered via the phase-field microelasticity theory, and an ad hoc methodology is developed to consider contact interactions. Dislocation glide is simulated in the presence of a representative population of damage loops, and the effect of interactions on critical-resolved shear stress is measured. These results are compared with existing experimental literature. Advisors/Committee Members: Wang, Yunzhi.

▼ Titanium and it alloys are extensively utilized in critical applications that require materials with high strength to weight ratios, rigidities, and toughnesses. This being the case, over 70 years of research have been devoted to the measurement, understanding, and tailoring of the mechanical properties of these alloys. Despite these efforts, surveys of the current knowledge base and understanding of the mechanical responses of Ti alloys demonstrate that numerous mechanical behaviors have yet to be investigated and explained. It has been noted, but generally not appreciated, that commercially important materials display modest strength differentials near room temperature when deformed under quasi-static loading conditions at modest rates (~10-5 to 10-3 1/s). Under static loading, subtle variations in plastic flow behavior leads to dramatically weaker materials when loaded in tension versus compression. The asymmetric material responses of single and two-phase alloys deformed under monotonic constant rate and creep conditions have been investigated and related to the fundamental slip behavior observed in single crystalline materials. Two-phase titanium alloys containing a majority volume fraction of the alpha (HCP) phase have long been known to undergo creep deformation at lower temperatures (T < Tm) and stresses (σ < σys). The time dependence of this plasticity, stemming from a-type slip in the alpha-phase, has been found to be sensitive to microstructural condition. The nature of low temperature creep in heat-treatment modified beta-annealed Ti-6Al-2Sn-4Zr-2Mo has been investigated. Microstructural features, particularly primary alpha-lathe and beta-rib structure (secondary alpha morphology and content), were systematically modified, and the resulting structures were deformed under both creep and constant strain rate loading conditions. Variations in plastic response are discussed in terms of strain hardening and strain rate sensitivity parameters. The mechanical performance of engineering Ti alloys has long been known to be sensitive to the to nature of applied load waveform. A review of the open literature elucidates several gaping holes in the current understanding of waveform induced plastic response of cyclically loaded materials. This study addresses several of those issues. Sustained load hold time effects during the fatigue an alpha-Ti alloy is investigated with respect to loading conditions and slip planarity and compared to cyclic fatigue and creep responses at room temperature. Advisors/Committee Members: Mills, Michael J.

Advisors/Committee Members: Konya, Calvin.

▼ Composite α/β stereographic projections and slip system misorientation diagrams were developed and used to establish a new understanding regarding anisotropic deformation behavior of alpha and beta phases aligned according to the Burgers orientation relationship. Application of these tools showed that a 2-fold, maximum common crystal symmetry existed between single variants of Burgers oriented alpha and beta phases, which placed specific requirements on alignment and response of mating slip systems in alpha and beta phases. Of particular import were implications for crystallographic elements of mating slip systems that were found to be oriented within 180° of each other, i.e. anisotropic deformation behavior was predicted according to requirements of 2-fold maximum common crystal symmetry. Inspection of misorientation between mating slip systems established new requirements for determining the breadth of anisotropic slip behavior for a-basal, a-prism, a-pyramidal and c+a pyramidal slip systems. It was shown testing at 3, 6, 6, and 12 maximum Schmid factor loading axes was required to fully measure anisotropic deformation behavior in these slip systems, respectively. This meant each mated slip system would have 1, 2, 2 and 4 unique responses when loaded at orientations of maximum Schmid factor on the alpha slip system. The root cause of these anisotropic responses was shown to result from large changes in Schmid factor on beta slip systems when loading conditions on alpha slip systems were held constant, i.e. at maximum Schmid factor. Changes in Schmid factor on mating beta slip systems were shown to be a natural consequence of differences in crystal symmetry between alpha and beta phases oriented according to the Burgers orientation relationship. Extension of this information to titanium microstructures consisting of multiple alpha variants in a single beta grain showed that 144 possible combinations of two adjacent alpha variants could be grouped according to six characteristic crystallographic and morphological misorientations. A reference α/β variant VIa replicated crystallographic orientation of eleven other alpha variants through single rotations of 0°, 10.52° @ {011} ‖ (0001), 60° @ <111> ‖ <2-1-10> and 90° @<010> Δ 5.26° <-12-10>; double rotations of 10.52° @ {011} ‖ (0001) + 60° @ <111> ‖ <2-1-10>; and triple rotations of 10.52° @ {011} ‖ (0001) + 60° @ <111> ‖ <2-1-10> + 10.52° @ {011} ‖ (0001). Growth directions between pairs of α/β variants were shown to have interface misorientations of (0°/180°), (24.9°/155.1°), (54.5°/125.5°), (60.76°/119.3°), (77.9°/102.1°) and (80.6°/99.4°). This information, combined with that in slip system misorientation diagrams, was used to make qualitative predictions of relative ease and difficulty of slip transmission between combinations of α/β/α variants based on in-plane and out-of-plane misorientations of slip planes and slip directions. For example, slip transmission between variants of near parallel alignment (Δ = 1.38°) between (0001) and {10-11} with a common {011}<111> slip system was predicted, then confirmed through Transmission Electron Microscopy (TEM) imaging techniques. Six variations of the production “Triplex” Ti-62222S heat treatment process were developed and applied to effect large changes in room temperature tensile deformation behavior. Large changes in morphology of primary alpha features and the presence of α2-Ti3Al in the alpha promoted large reductions in fracture stress and fracture strain capability of beta solution heat-treated and aged Ti-62222S. These effects were shown to be additive, so that the combination of high amounts of aligned colony alpha + high presence of α2-Ti3Al in the alpha produced the lowest fracture stress and fracture strain. Work hardening in all six microstructures was relatively low, e.g. n ranging between ~0.05 and 0.1. As with fracture behavior, increasing the amount of aligned or colony microstructure and/or increasing the presence of α2-Ti3Al in the alpha were shown to lower work hardening. Examination of fracture surfaces and polished cross sections perpendicular to fracture surfaces revealed an associated transition from transgranular to intergranular fracture with increasing amount of aligned or colony alpha and if the final aging treatment increased the presence of α2-Ti3Al in the alpha. An important factor distinguishing the two fracture modes was the presence of fine voids at α/β interfaces in sub-fracture surface regions just below the main crack. Material aged to have less presence of α2-Ti3Al in the alpha, i.e. no aging or aging at 538°C/8hrs/AC showed these fine pores, while material aged to increase the presence of α2-Ti3Al in the alpha, i.e. aging at 621°C/8hrs/AC, did not. More detailed inspection of deformed materials using OIM and TEM imaging techniques showed several interesting and previously unreported behaviors for Ti-62222S. These included the presence of deformation twins in the alpha, emission of dislocations from α/β interfaces, very strong fringe contrast between pairs of dislocations gliding on all slip systems and dislocations in the beta with b = [010] in pile-ups associated with dislocations gliding on an unaligned a2-basal slip system. Information gleaned from composite α/β stereographic projections and slip system misorientation diagrams proved critical in developing proper interpretations of these observations. Advisors/Committee Members: Fraser, Prof. Hamish.

▼ Jelly-roll type Nb-Al multifilamentary wires were rapidly heated to temperatures above 1800°C followed by quenching to room temperature, using experimental setups designed and built for this purpose. The ductile Nb-Al bcc solid solution retained by quenching is transformed by an 800°C/10h transformation heat treatment to superconducting Nb3Al of A15 structure. In the range of compositions investigated (23.5 to 35.0at% Al) the bcc phase was found to form by an endothermic reaction taking place during rapid heating. Optimal critical current densities (at 4.2K in magnetic fields exceeding 15T) were obtained for samples heated to maximum temperatures in a range of less than 200°C above the bcc formation reaction. At higher temperatures the bcc solid solution starts to melt, leading to inhomogeneity that degrades the critical current density. Samples quenched from fully liquid state exhibit large pores that can reduce the critical current density. X-ray diffraction and scanning electron microscopy on samples quenched from temperatures above and below the bcc formation temperature showed that at 23.5at% Al, an A15 structure Nb3Al formed in the earlier stages of the rapid heating massively transforms to bcc solid solution. A massive transformation is also found at 35.0at% Al, sigma-type Nb2Al converting to bcc solid solution. At 27.5at% Al, Nb3Al reacts eutectoidally with Nb2Al to form bcc solid solution. All these transformations were found reversible and heating rate independent. The bcc solid solution formation and the extended Al solubility reported could not be explained based on published Nb-Al equilibrium phase diagrams so it is concluded that none of them is correct and a revised phase diagram is proposed. All superconducting properties improved with increase in the initial heating rate associated with the transformation heat treatments. X-ray diffraction and transmission electron microscopy studies revealed that at low heating rates the spacing between certain planar faults is reduced which leads to poorer superconducting properties. At slow heating rates, prior to the transformation to A15, a B2 ordering of the bcc phase was found to proceed to a higher extent than for slow heating, which suggests a connection with the planar fault formation. Advisors/Committee Members: Snyder, Robert L.

▼ Carbon steel is susceptible to stress corrosion cracking (SCC) in fuel grade ethanol. Dissolved oxygen and corrosion potential have been identified as the critical factors. The threat of SCC prevents the use of the cost-efficient pipeline system for long distance transport of ethanol. Simulated fuel grade ethanol (SFGE) was used in the laboratory. Due to the high electrical resistivity of SFGE, adding non-complexing supporting electrolyte is considered to be the most practical method for accurate potential control. TBA-TFB was found to be the best suitable supporting electrolyte in deaerated FGE among the salts that were tested. Carbon steel exhibits passivity and high corrosion potential in aerated SFGE. Deaerated conditions are of interest so that the effects of high potential on SCC susceptibility can be determined separately from other possible effects of dissolved oxygen. In slow strain rate (SSR) tests, cracking was reproduced at applied potential without oxygen in the deaerated SFGE + TBA-TFB, which indicates that the role of oxygen in ethanol SCC might be a simple oxidizing agent. However, the passivation effect of oxygen is also required to prevent lateral corrosion at crack tip. A potential range for ethanol SCC was determined by SSR tests at different potentials. No experimental evidence was found to support the proposed role of oxygen to react with ethanol to form an aggressive oxidation product. The presence of chloride causes decreasing corrosion potential and enhanced pitting corrosion. The chloride effect on ethanol SCC was investigated both in aerated SFGE at open circuit and deaerated SFGE at applied anodic potentials over a wide range of chloride concentration. A minimum concentration of chloride is required for SCC of carbon steel, but it is not the controlling factor for crack growth. There is no upper limit of chloride concentration for cracking in aerated SFGE, but a window of chloride concentration in deaerated SFGE at applied potential. The dissolution based SCC mechanism has been identified for carbon steel in ethanol environment. SSR testing with periodic potentiodynamic scans at different stains and strain rates show that the anodic current difference between plastic and elastic region has a peak in the cracking potential range. The notched SSR testing is more sensitive to ethanol SCC susceptibility, and it generates more consistent results of current evolution. A high R-ratio and low cyclic frequency crack growth rate (CGR) test was intended to mimic the actual loading condition of pipelines in service at accurately-controlled fracture mechanics conditions. The measured CGR at applied potentials matches the cracking susceptible potential region. An oxygen depletion induced dissolution model and the traditional film rupture induced dissolution model were proposed to explain the mechanism of ethanol SCC. A few models based on oxygen diffusion and consumption were considered in an attempt to explain the differences of the two proposed mechanisms. Complete oxygen depletion is less likely to occur at crack tip. Advisors/Committee Members: Frankel, Gerald.

▼ Gas-phase reactions between H 2and SnO 2surfaces led to novel nanostructures. Nanosheets of SnO 2were produced by a gas-phase reaction of solid SnO 2sintered disks in a reducing atmosphere between 700 and 800 degrees Celsius. The surface morphology was controlled by varying the reaction time and partial pressure of oxygen. It was found that SnO 2was etched by a reaction between H 2and lattice oxygen and a loss of SnO to the vapor phase. Single-crystalline nanofibers of SnO 2were synthesized by a gas-phase reaction of solid SnO 2sintered disks in a reducing atmosphere between 650 and 750 degrees Celsius. The resulting nanostructures grew on regions of the sample that were coated with gold, which acted as a collector of SnO vapor. The nanofiber length was controlled by varying the reaction time and by the sintering agent used to densify the SnO 2. SnO 2thin films were manufactured using DC reactive magnetron sputtering from a tin target in an Ar/O 2atmosphere. After exposure to H 2gas at temperatures between 600 and 680 degrees Celsius, nanofibers were grown on the surface with the aid of gold particles. The nanofiber growth was controlled by varying the reaction time and the orientation of the SnO 2films. The gold was shown to be necessary for both initiation and continuation of growth. Sensing tests were conducted with pure TiO 2and mixed oxide samples of TiO 2and SnO 2sintered samples having different surface areas. Nanostructures of solid solutions and spinodally decomposed samples of mixed oxides of SnO 2and TiO 2were compared to the oriented nanofibers formed in pure TiO 2created by the same heat treatment. Thin films with SnO 2nanofibers were also tested for their sensing response to H 2. Comparisons between the sensing characteristics of the samples were made to determine that those samples with high surface areas and those with a gold coating were more sensitive to H 2in the presence of O 2. Advisors/Committee Members: Akbar, Sheikh.

▼ Modeling the defect structure and mechanical properties of metallic multilayer thin films requires estimates of dislocation parameters such as interfacial line energy, interfacial barrier strength, and resistance to confined layer slip (CLS). A method is presented to estimate these parameters using experimental measurements of hardness and internal stress vs. individual layer thickness, h. Cu/Ni multilayers of varying bilayer thickness (20 nm < BT < 60 nm) and volume fraction (25 < %Ni < 60%) were fabricated via sputtering and characterization performed using x-ray diffraction (XRD), transmission electron microscopy (TEM), and scanning electron microscopy (SEM). Internal stresses for the samples were calculated via peak positions from in-plane XRD and second order elastic constants. The experimental techniques of nanoindentation and micropillar compression data were used to look at flow stresses, hardness, and strain rate sensitivities for the Cu/Ni multilayers. Internal stress was seen to increase with decreasing bilayer thickness and decreasing layer thickness for both layer types. In addition, hardness was seen to increase with decreasing bilayer thickness and decreasing Cu layer thickness. The data acquired via characterization and experimentation was used as inputs within a CLS model in order to extract quantities for interfacial properties. It was seen that that separate values of line energy operate in the Cu and Ni layers and that a single effective line energy for the multilayer is inappropriate. This indicates that dislocation loops will encounter a different resistance at the shared interface depending on whether the dislocation loop originates in Ni or Cu. Analytical models for line energy overestimated the line energy in Cu and underestimated the line energy in Ni. It was seen that the interfacial barrier to dislocation motion increased with increasing bilayer thickness and misfit strain. A maximum value of 444 MPa was extrapolated for the interfacial barrier to dislocation transmission. Pinning stresses increased initially with increasing layer thickness due to the increasing misfit strain. The pinning stresses reached a maximum of 240 MPa in Ni before decreasing for layer thicknesses > 20 nm. Finally, it was seen that coherency stresses derived from lattice parameter mismatch did not completely explain the hardness found in the multilayers. Atomistics had predicted that strength was dominated by the contribution of coherency stresses while these experiments indicated that some contribution from pinning and interfacial stresses was required to fully explain the multilayer strength. The results of the modeling in this dissertation suggest that multilayer strength can be increased by methods other than reducing bilayer thickness. These methods include selecting a seed or buffer layer to place the multilayer in a state of net in-plane compression and increasing the barrier to dislocation transmission by fabricating the compressive phase in thinner quantities compared with the tensile phase. Advisors/Committee Members: Anderson, Peter.

▼ Deformation in polycrystalline nickel-based superalloys is a complex process dependent on the interaction of dislocations with both the intra-granular γ′ particles and the grain boundaries. An extensive body of work exists on understanding the interaction between dislocations and the γ′ particles, but understanding the interaction between dislocations and grain boundaries has been historically hindered by the experimental techniques. In this work a full field strain mapping technique was developed and utilized to explore surface strain accumulation at grain boundaries of René 104 samples with different microstructures. The full field strain mapping technique utilized Correlated Solutions VIC-2D software for digital image correlation to measure strain accumulation from secondary electron images taken during constant load tests at elevated temperature. This technique indicated that the two different microstructures of René 104, one with microscopically flat grain boundaries and the other with serrated grain boundaries, accumulate strain by different methods. Analysis of discrete offsets in grid lines placed prior to deformation indicate that grain boundary sliding (GBS) is an active deformation mechanism at these temperature and strain rate regimes, and that the development of serrated high angle grain boundaries can decrease the activity of this mechanism by 30%. Slip transmission parameters, which mathematically assess the ease of slip transmission across a grain boundary, were calculated based on grain boundary misorientation and grain boundary trace. These parameters proved unsuccessful at predicting strain localization sites in these materials, indicating that slip transmission is not the only factor dictating strain localization sites. Full field strain maps were used to site-specifically extract grain boundaries of interest to study dislocation interaction and sub-surface grain boundary neighborhood. Representative from each of four types of boundaries was selected for scanning transmission electron microscopy (STEM) analysis: high angle grain boundaries that either did or did not experienced strain accumulation, high angle grain boundaries that experienced GBS, and special annealing twin boundary that experienced both strain accumulation and grain boundary sliding. The STEM analysis indicates that high angle grain boundaries that experienced strain accumulation showed increased dislocation content near the grain boundary, while a grain boundary that did not show accumulation and a grain boundary that experienced grain boundary sliding showed no indication of increased dislocation content near the grain boundaries. The STEM analysis also indicated that grain boundary surface roughness and the sub-surface grain boundary neighborhood was substantially more complex for the grain boundaries that experienced strain accumulation as compared to the boundary that experienced GBS. Since STEM foils only provide a two dimensional representation of the grain boundary surface, serial sectioning data sets were reconstructed from stacks of 2D images acquired using the Focused Ion Beam (FIB). These reconstructions confirm that indeed the surface roughness of a boundary that experienced strain accumulation was an order of magnitude greater than a grain boundary that experienced GBS. The observations from the STEM and serial sectioning work indicate that grain boundary neighborhood and grain boundary topography also need to be considered if models are to predict strain localization and GBS sites. Advisors/Committee Members: Mills, Michael.