Authors: J.R. Grim, Marek Skowronski, W.J. Everson, V.D. Heydemann
Abstract: The selectivity, material removal rate, and the residual subsurface damage of colloidal
silica (CS) chemi-mechanical polishing (CMP) of silicon carbide substrates was investigated using
atomic force microscopy (AFM) and plan view transmission electron microscopy (TEM). Silica
CMP, in most process conditions, was selective. In the damage region surrounding remnant
scratches, the vertical material removal rate exceeded the planar material removal rate, which
resulted in an enhancement of the scratches over the duration of the polishing process. The material
removal rate was low, about 20 nm / hr. In addition, the selectivity leads to a slow removal of
residual subsurface damage from mechanical polishing. The silica CMP polished surface exhibits
significant subsurface damage observed by plan view TEM even after prolonged polishing of 16
hours.
1095
Authors: W.J. Everson, V.D. Heydemann, Rick D. Gamble, David Snyder, G. Goda, Marek Skowronski, J.R. Grim, E. Berkman, Joan M. Redwing, J.D. Acord
Abstract: A new chemical mechanical polishing process (ACMP) has been developed by the Penn
State University Electro-Optics Center for producing damage free surfaces on silicon carbide
substrates. This process is applicable to the silicon face of semi-insulating, conductive, 4H, 6H, onaxis
and off-axis substrates. The process has been optimized to eliminate polishing induced
selectivity and to obtain material removal rates in excess of 150nm/hour. The wafer surfaces and
resultant subsurface damage generated by the process were evaluated by white light
interferometery, Transmission Electron Microscopy (TEM), Atomic Force Microscopy (AFM), and
epitaxial layer growth. Residual surface damage induced by the polishing process that propagates
into the epitaxial layer has been significantly reduced. Total dislocation densities measured on the
ACMP processed wafers are on the order of the densities reported for the best as grown silicon
carbide crystals [1]. Characterization of high electron mobility transistors (HEMTs) grown on these
substrates indicates that the electrical performance of the substrates met or exceeded current
requirements [2].
1091
Authors: Timothy Bogart, W.J. Everson, Rick D. Gamble, Ed Oslosky, David Snyder, Eugene Furman, Steve Perini, M. Lanagan
Abstract: Semi-insulating silicon carbide (SiC) wafers are important as substrates for high
frequency devices such as AlGaN-GaN HEMT’s. A nondestructive characterization technique has
been developed to measure the dielectric properties of SiC wafers in the GHz frequency range
where the devices will operate in order to validate wafers for high yield working devices. The
dielectric loss is measured at approximately 16 GHz in a split microwave cavity. Initial results
show a correlation where the dielectric loss decreases as the resistivity increases, where the
resistivity was measured using a Contactless Resistivity Mapping system (COREMA). The
uniformity of dielectric loss across SiC wafers was evaluated using a split post dielectric resonator
cavity fixed at 5.5GHz to measure the dielectric loss at five points on a wafer. Dielectric loss as a
function of temperature from room temperature to 400°C was also studied.
733
Authors: Mark A. Fanton, R.L. Cavalero, R.G Ray, B.E. Weiland, W.J. Everson, David Snyder, Rick D. Gamble, Ed Oslosky
Abstract: The effects of growth conditions, diffusion barrier coatings, and hot zone materials
on B incorporation in 6H-SiC crystals grown by physical vapor transport (PVT) were evaluated.
Development of high purity source material with a B concentration less than 1.8x1015 atoms/cm3,
was critical to the growth of boules with a B concentration less than 3.0x1016 atoms/cm3.
Application of refractory metal carbide coatings to commercial graphite to serve as boron diffusion
barriers and the use of very high purity pyrolytic graphite components ultimately led to the growth
of SiC boules with boron concentrations as low as 2.4x1015 atoms/cm3. The effect of growth
temperature and pressure were closely examined over a range from 2100°C to 2300°C and 5 to 13.5
Torr. This range of growth conditions and growth rates had no effect on B incorporation. Attempts
to alter the gas phase stoichiometry through addition of hydrogen gas to the growth environment
also had no impact on B incorporation. These results are explained by considering site competition
effects and the ability of B to diffuse through the graphite growth cell components.
47
Authors: V.D. Heydemann, W.J. Everson, Rick D. Gamble, David Snyder, Marek Skowronski
805
Authors: W.J. Everson, David Snyder, V.D. Heydemann
837
Authors: David Snyder, V.D. Heydemann, W.J. Everson, Donovan L. Barrett
9