| A major goal of diamond thin film research is to achieve control over diamond crystallite nucleation. This dissertation examines the causes of nucleation, primarily for low C flux rates in the absence of diamond seeding, intentional adsorbates, or electrical bias. Both hot filament and microwave enhanced chemical vapor deposition (ECVD) techniques were employed.;This work begins with an experimental comparison of diamond nucleation to the capillarity theory of nucleation. The capillarity theory is derived from thermodynamic arguments, and predicts nucleation events both to occur in concave substrate features, and to be favored as the substrate temperature is lowered.;To isolate and test the effect of substrate topography, ;Non-random nucleation patterns observed on convex features imply that other factors may work in concert with the topography. The factors of dangling bonds and carbide formation were subsequently investigated.;Silicon substrates were studied by electron spin resonance (ESR) to investigate the role of dangling bonds induced in the near surface region. The results indicated that substrate topography plays a more significant role in nucleation than does dangling bond density.;Transmission electron microscopy (TEM), secondary ion mass spectrometry (SIMS), and X-ray photoelectron spectroscopy (XPS) were used to investigate the formation of a carbide layer on Si substrates. The results indicated that ;The elimination of these two factors suggests that particular atomic arrangements of the substrate surface or chemical contaminants are responsible for nucleation. These considerations and others indicate the necessity for a cyclohexane-like nucleation kernel, and highlight the need for understanding the kinetics leading to its formation on the substrate. |
