The first question addressed in this thesis is whether the cumulative triboemission from the abrasion of alumina by a diamond indenter (repeat-pass sliding) correlates with the volume of material removed and, in particular, whether transitions in the rate of material removal are mirrored in the cumulative triboemission rate. As a function of load and number of diamond passes, several wear regimes are observed that are characterized from SEM micrographs by different relative proportions of plastic flow material and fractured surface in the wear scar. In all but one wear regime, the correlation between the wear volume and cumulative triboemission is modest (linear regression coefficient R2 = 0.71); including the one atypical wear regime worsens the correlation. The wear volume and cumulative triboemission are shown to be random variables with normal and lognormal distributions, respectively. Again, excluding the atypical wear regime, the correlation between the logarithms of the estimated population means is significantly better (R2 = 0.91) than the correlation between wear volume and cumulative triboemission for individual samples. In addition to the overall correlation between wear volume and cumulative triboemission, transitions from one wear regime to another are marked by changes in the slope of the mean cumulative triboemission versus pass number. These transitions correlate with the relative fraction of plastic flow debris in the wear scar. The second question addressed in this thesis is whether the introduction of the chemical vapor aluminum tri-sec butoxide, [C2H5CH(CH3)O]3Al, ATSB, into the boundary layer of an alumina-on-alumina sliding contact can reduce wear and friction. A split-plot factorial experiment was conducted; the factors tested, in addition to the presence or absence of ATSB, were normal load, sliding speed, and surface roughness. The main conclusions of the experiment are that ATSB has no statistically significant effect on specific wear, but that the presence of ATSB reduces friction by 21% at low sliding speed (0.02 m/s) and increases friction by 26% at high sliding speed (1.2 m/s). / Ph. D.

Lubrication of ceramics is a difficult and not completely solved problem. Ceramics do not respond to conventional lubricants which are designed to function by a chemical reaction with the surface. There is, accordingly, increased interest in the development of lubrication alternatives for ceramics, and in understanding the tribochemical fundamentals by which new lubrication processes can be designed and controlled. In particular, the mechanism of tribopolymerization for some addition-type monomers is thought to be initiated and controlled by triboemitted low-energy electrons. This Ph.D. dissertation presents the experimental work carried out at the Virginia Polytechnic Institute and State University for the characterization of charge intensity and energy distribution of charged-particles triboemitted from sliding contacts of ceramics. A review is presented of research work on tribochemistry and, in particular, on tribopolymerization as a lubrication mechanism. Relevant literature is also reviewed on the phenomena of exoemission, triboemission and fractoemission of charged-particles. The design, construction and development of a new instrument and data acquisition system to carry out triboemission measurements under high vacuum and for controlled load, sliding speed and retarding grid-voltage is described. The charge intensity is characterized of the particles triboemitted from two related ceramics, alumina and sapphire, and from one metallic material, i.e., aluminum, when scratched by a diamond pin. In the case of alumina, triboemitted-charge intensity also is studied by sliding contact of an alumina ball. Burst-type negatively-charged particle triboemission was observed from diamond-on-alumina, diamond-on-sapphire, and alumina-on-alumina sliding contacts. The different crystalline structure, i.e., of alumina and sapphire, does not appear to be a factor in electron triboemission. In general, large bursts of electron triboemission may appear superimposed on a constant lower level of small-burst emission. This constant level, being higher than background-noise, does not vary between different ceramic specimens, while maximum levels of triboemission-bursts differ by two orders of magnitude between different specimens. The characteristic decay-time of the triboemission bursts is found of about 100ms. Lower-level decaying post-contact emission of negatively-charged particles from ceramics is observed. Low negatively-charged triboemission was observed from diamond-on-aluminum sliding contacts. The positively-charged triboemission from these sliding material systems was also measured. Low positive-ion emission, barely above background level, was observed for the diamond-on-ceramics and alumina-on-alumina systems. The retarded-energy spectra of the negatively-charged particle triboemissions from ceramics were also obtained. Such spectra show decaying rates of triboemission for increasing minimum energy of the triboemitted particles: an important fraction of the total electron triboemission is produced in the zero to 5eVolts energy-range, with a decaying tail extending beyond the test maximum level of 48 Volts. These experimental measurements are discussed with a focus on the possible role of triboemitted charged-particles in tribopolymerization as a mechanism of ceramic lubrication. It is concluded that low-energy electrons are emitted in bursts from ceramics under sliding contact, the essential first step in the hypotheses of tribopolymerization of certain addition-type monomers, while positively-charged emission is negligible. These findings strongly support tribopolymerization results from previous research. A frequency domain analysis of the triboemission data is carried out. For the electron-triboemission outputs, a characteristic pattern is found for the experimentally estimated frequencies of occurrence of the triboemitted particles. A new probability distribution, called "Convoluted Poisson" is developed to describe this triboemission data. Good agreement is found between the probabilities of triboemitted-particle occurrence, as predicted by such distribution, and the experimental probabilities estimated from triboemission outputs. The significance of the two parameters defining this "Convoluted Poisson" distribution is explored and discussed with a focus on basic surface-change phenomena. / Ph. D.