Abstract Not Available.

Nanoindentation is a widely recognized method for characterizing the mechanical properties of thin films and small volumes. This thesis reports the results of an investigation aimed at developing a technique by which the fracture toughness of a thin film or small volume can be determined from low-load indentation experiments. The method relies on the fact that most brittle materials form radial cracks when indented with a sharp indenter such as a Vickers indenter. The lengths of the radial cracks produced during microindentation experiments have been shown to correlate reasonably well with fracture toughness. As a result, simple semi-empirical relations have been developed to calculate fracture toughness based on the measurement of indentation crack length. The one problem encountered in applying the indentation cracking method for measuring fracture toughness to low loads is that there are threshold loads below which most materials do not form radial cracks. For Vickers and Berkovich indenters, the cracking threshold is 25 grams ($\sim$250 mN) or more for most ceramic and glass materials. It is shown that the problems imposed by the cracking threshold can largely be overcome by using an indenter with the geometry of a cube corner. With a cube corner indenter, the cracking threshold of most brittle materials can be reduced to loads as small as 0.1 gram ($\sim$1 mN). In addition, the well-developed relationships between indentation crack length and fracture toughness used for the Vickers indenter at high loads, generally above 1000 grams ($\sim$10 N), can be used with the cube corner indenter at very small loads, 0.1 gram ($\sim$1 mN) and above, provided different empirical constants are used. The Berkovich indenter, which is the three-sided analog of a Vickers indenter, is also found to obey the fracture toughness relations of the Vickers indenter. A model is developed which explains the differences in cracking thresholds for the Vickers, Berkovich, and cube corner indenters. The model is based on the hardness impression itself serving as the precursor flaw from which cracks extend.

Single wall carbon nanotubes (SWNTs) were studied as low shear rate viscosity modifiers in the polypropylene matrix of nanocomposites. To create nanocomposites which did not demonstrate increased low shear rate viscosity when nanotubes were added into the polymer melt, this work focused on modifying the sidewall profile of the SWNTs and manipulating their configuration in the polymer matrix before subjecting them to rheological testing. Fluorinated single-walled nanotubes (F-SWNTs) played a critical part in evaluating how functionalizations could affect the viscosity of the polymer melt. Fibers made from weight percents ranging from 2.5wt% to 10wt% of F-SWNTs in isotactic polypropylene, with Mw = 250,000 g/mol (iPP250,000), were shown to have lower complex viscosity profiles than the neat iPP250,000 sample at low shear rates. These fibers demonstrated decreases in the complex viscosities, in the low frequency range, of 36.5%, 27.8% and 37.5%, for the 2.5wt.%, 5wt.% and 10wt.% fiber samples respectively. F-SWNTs were shown to stimulate in situ initiated reactions in the polymer melt where free radials generated, as a result of spontaneously dissociated fluorine atoms which scavenge hydrogen from the surrounding polymer chains, facilitate the covalent bonding of nanotubes directly to the polypropylene chains. These covalently bonded nanotubes were then forced to align themselves, along with the polymer chains when the bulk composite is spun into a small diameter fiber (dia.130mum). Another method used to create low shear rate viscosity nanocomposites was to capitalize on the low viscosity properties of low molecular weight isotactic polypropylene (Mw = 12,000g/mol) by creating a hybrid nanocomposite system in the iPP250,000 matrix. In these systems, benzoyl peroxide was used as the free radical initiator which could facilitate the covalent bonding of single walled nanotubes to the polymer chains in the melt. Both non-functionalized and functionalized hybrid nanocomposites were evaluated for their rheological properties; and for equivalent weight percents of nanotubes incorporated, the hybrid nanocomposites which were functionalized, using benzoyl peroxide as the initiator, demonstrated lower complex viscosities profiles. A maximum decrease of 54% was observed, at low frequency, for a 0.1wt% non-functionalized hybrid sample compared to its functionalized counterpart.

Due to the accumulated van der Waals attraction, the separation of single-walled carbon nanotubes (SWNTs) has been a challenge. This thesis focus on theoretical modeling and study of separating SWNT array in electrolytes and by chemical functionalization. First, the swelling behavior of SWNT fibers in superacid as well as the sudden collapse of such dispersion with introduction of water are studied by Derjaguin-Landau-Verwey-Overbeek (DLVO) theory. A simple kinetic model for the charging mechanism of SWNT is proposed. Optimization of the controlled parameters of the model has been carried out. Second, we designed a tensegrity structure with three (5,5) SWNTs connected by nine polyethylene chains. One of relative strong connection ways between polyethylene and tube cap is constructed. Structure optimization and molecular dynamics simulation are performed to analyze the stability changes in such and several similar structures by adding some or removal one chain.

Lubricants and cooling agents such as oil, ethylene glycol, and water are often used as traditional heat transfer fluids (HTFs) in engines, radiators, heat pumps, and other equipment which require cooling and/or energy maintenance. The United States of America (USA) spends over $80 billion on energy maintenance and increasing the thermal efficiency of HTFs by 25% could annually save over $20 billion. This study was aimed at improving the thermal conductivity (TC) of synthetic and vegetable oil heat transfer fluids (HTFs) by using Single-Walled Carbon Nanotubes (SWNTs) additives that resulted in a semi-solid lubricant that was a nanotube-HTF (n-HTF). The n-HTFs were processed with nitrogen containing additives that coupled with toluene and acetone solvent processing promoted the gelation of the carbon nanotube fibers in the oil. Such processing improved the nanotube dispersion due to hydrogen bonding (H bonding) and micelle formation of the amine groups around the carbon nanotube rod-like fibers and the additive-oil matrix. This allowed for high weight percent (Wt %) loadings of carbon nanotubes in the oil. Characterization via thermal graviometric analysis (TGA) and Fourier Infrared Transform (FTIR) showed that high temperature radical mechanisms breakdown both the oil and nanotubes, and optical microscopy showed that sonication-homogenizing-mixing affects the coagulation-flocculation-aggregation-agglomeration of the nanotubes. Additionally we used two new TC instruments, the Mathis-Hot Disk and KD2 systems, to provide accurate and reproducible data with a 10% and 4% error margin for the Mathis and KD2, respectively. Raman and FTIR spectroscopies suggest that the TC enhancements result from SWNT phonon mechanisms, these being phonon-phonon, phonon-defect, and phonon-interface, all of which are present at room temperature with the absence of ballistic and superconductivity phenomena. Additional vibrations in the oil that occur due to Brownian motion and electron-phonon and H bonding from the additives would have also contributed to the TC mechanism and were evidenced via Raman spectroscopy and FTIR. The optical microscopy, Raman, FTIR, and TC values indicated that the n-HTFs had a three-dimensional (3D) SWNT networked structure due to the inclusion of the oleylamine additive and the toluene-acetone processing. This would have resulted in the formation of oleylamine-nanotube micelles that were suspended in the oil-additive mixture. Raman spectroscopy evidenced a percolation effect that was coupled with fluid vibrations and Brownian motion, which led to the absence of the one- or two-dimensional (1D, 2D) ballistic or superconductivity phenomena that is often associated with aligned and single SWNTs and other 1D or 2D media, and this was reflected in the TC values. Nevertheless, the TC of the resulting n-HTFs was improved by over 80-96% when compared to other HTFs and the dispersion of the nanotubes in the oil-additive mixture was greatly enhanced. Applications of the n-HTFs include: mechanical-frictional damping, semiconductor packaging, thin-films, hydraulic oil lubricant use, thermoelectric power, thermo-sensing, fuel cells, additive-antioxidant-viscosity modifiers, and filtration.

The effect of increased Cu and Si additions and heat treatment on the polarization resistance, oxide-layer/double-layer capacitance, open circuit potential and breakdown potential of Al-Cu-Si alloys was evaluated. Electrochemical impedance spectroscopy and DC polarization techniques were combined with mechanical property evaluations to determine the relationship between the microstructure of Al-Cu-Si alloys and the ability of their oxide layer to provide protection against corrosion in an aqueous 5% NH$\sb4$Cl solution. Alloys were investigated in both the solution-treated and artificially aged conditions. The solution-treatment procedure involved heating the alloys to 550$\sp\circ$C for 30 minutes and subsequently water-quenching. Artificial aging was carried out at 250$\sp\circ$C. Increased aging times of 1, 2, 8 and 32 hours at 250$\sp\circ$C were employed on Al-Cu(2%), Al-Si(2%) and Al-Cu(2%)-Si(2%) alloys in order to correlate the observed electrochemical behavior with the precipitation strengthening reactions which occur during age-hardening. These precipitation reactions govern the resultant microstructure and determine the mechanical behavior, electrochemical behavior and type of corrosion attack observed in Al-Cu-Si alloys. In general, an increase in the susceptibility to pitting corrosion was correlated to a decrease in the protective oxide layer thickness brought about as a result of increased Cu and Si additions in solution-treated alloys and the precipitation of Guinier-Preston zones in age-hardened alloys. Cu and Si atoms at the metal/oxide interface inhibit the diffusion of Al$\sp{+3}$ ions to the oxide/electrolyte interface thereby limiting the oxide-layer thickness. Increased aging times, leading to the formation of Guinier-Preston zones and a concomitant age-hardening peak, also decrease the total number of Al$\sp{+3}$ ions available for passivation and thereby decrease the oxide-layer thickness. Consequently, the oxide-layer/double-layer capacitance and the polarization resistance of the alloys are correlated to changes in the oxide-layer thickness and the loss in continuity of the oxide-layer with overaging. In addition, the open-circuit and breakdown potentials and the type of corrosion attack observed during anodic DC polarization scans was related to the type of alloying element and their distribution.

The corrosion susceptibility of Al thin-film metallizations in an aqueous solution containing chlorine was measured. DC polarization techniques were used to determine the corrosion behavior of Al thin-film metallizations, Al-alloy thin-film metallizations and Al thin-film metallizations deposited on CVD W and sputtered W-Ti(10%) barrier layers. A galvanic series was developed by listing the stable open-circuit potentials for several metallizations, barrier metals and metallization/barrier-layer couples immersed in a 2000 ppm NH$\sb4$Cl electrolyte. Barrier layers of CVD W and sputtered W-Ti(10%) influenced the corrosion behavior of Al metallizations deposited on their surface. Results were discussed in terms of the galvanic couple formed between the metallization and the barrier layer, as well as the difference in surface topographies of CVD W and sputtered W-Ti(10%). The rough surface inherent on CVD W causes localized variation in metallization thickness and may provide sites where the W layer is exposed to the electrolyte, thereby acting as an efficient cathode. These results were substantiated with SEM photomicrographs. Electrochemical Impedance Spectroscopy (EIS) and DC polarization techniques were used to study the effect of increasing film thickness and grain size on the corrosion behavior of Al thin films. In general, as the film thickness is increased, the overall resistance to corrosion improves. The increase in corrosion resistance is attributed to an enlargement of the grain size, a result of the thin-film deposition process. As the grain size is increased, there is a reduction in the effective grain boundary area, which is the most prominent place were defects can be created or accumulated, thus rendering the oxide layer weak and promoting corrosion. The effect of grain size on the corrosion resistance of the films could not be determined, since the films were pitting upon immersion in the electrolyte.

Carbon buckytubes are widely known to be made on the cathode of a DC arc. The key aspect of the technique is an extremely high electric field (10$\sp7$ Vcm$\sp{-1}$). Generalizing from this method, we attempted the growth of buckytubes by pyrolysis of hydrocarbons in the high electric field established between an 8 mm commercial carbon fiber, later in the experiment thinner carbon fibers, and an opposing electrode. This research led us to conclude that the continuous growth of buckytubes could only start with a "seed crystal"--a perfect nanotube. We also observed field emission from buckytubes at a low onset voltage (60 Volts) and with a high current density (up to 10$\sp6$ A/cm). The onset voltage for emission depends on the history of the nanotube. The field emission from buckytubes after laser annealing versus after micro amp emission shows the difference that could be attributed to two distinct states of the tube tips: closed tips and open tips.

Nanoindentation techniques have been used to explore the hardness and elastic modulus of Nb/MoSi$\sb2$ microlaminates and stressed aluminium and stainless steel alloys. The elastic moduli of the microlaminates fall on the high end of range set by the rule-of-mixtures and increase slightly above it at small wavelengths. The hardness of the microlaminates increases with decreasing wavelength in a manner consistent with Hall-Petch strengthening. It is shown that average mechanical properties of the microlaminates can be measured when the plastic penetration depths are as small as one bilayer thickness. Studies in the aluminium alloy and stainless steel show that while hardness is affected by applied stress, the standard methods by which hardness and modulus are determined from nanoindentation data may lead to inaccuracies when large stresses exist in the material.

Various thin film metallizations were tested in order to gain insight into their performance as VLSI interconnects. Biaxial stress-strain test results exhibited limited ductility, even for pure FCC metals such as copper and aluminum. This may be caused by the columnar structure which does not present the same resistance to crack propagation as a random polycrystalline structure. A Hall-Petch relationship determined for bulge testing of pure aluminum is comparable to another worker's results for films tested by the substrate curvature method. Al-Cu(2%) alloys are shown to have a response to precipitation hardening similar to that of bulk material of the same composition. Uniaxial tension testing of thin films is problematic because of wrinkles and edge defects associated with sample preparation. Electromigration damage in Al-Si(1%) is accelerated by the presence of a tensile stress. Activation energies for creep of aluminum and Al-Cu(2%) are nearly identical to those determined in long term electromigration testing of 5 $\mu$m interconnect stripes.