-thick silicon oxide gate films grown at using modified reactive ion beam deposition with pyrolytic-gas passivation
Schematic diagram of the modified reactive ion beam deposition (RIBD) and cleaning (RIBC) system with an in situ pyrolytic-gas passivation (PGP). The system comprises an electron-cyclotron-resonance (ECR) plasma-type ion source, a cleaning and growth chamber, and a substrate exchange chamber (not described). The energy of ionized species is controlled by adjusting the extraction voltage supplied between the plasma chamber in the ion source and the substrate holder in the cleaning and growth chamber. Paralytic gases (PG’s) are blown around the substrate surface through the preheating unit, which is the same one used in the furnace-type PGP method.
High-resolution transmission-electron-microscope cross-sectional image ( magnification) of the -thick RIBD-with-PGP-grown silicon oxide films on -type Si(100) wafers covered with in situ P-doped RIBD-grown polycrystalline Si films. The layer indicated by the arrow is the RIBD-with-PGP-grown amorphous film sandwiched between the lower single crystal Si(100) and the upper polysilicon.
Current density -electric field intensity characteristics of the 1.5- and -thick RIBD-with-PGP silicon oxide films grown at with ion energy and growth rate using source gas and -base and PG’s, whose gas flow rates ( and ) in the total material gas flow are 5% and 0.2%, respectively.
Fowler-Nordheim (FN) plot of the -thick RIBD-with-PGP silicon oxide films grown at with ion energy and growth rate using source gas and -base and PG’s of and . The potential barrier height energy of the oxide for FN tunneling electrons was derived under the assumption that the electron effective mass in the oxide is half the electron mass in free space .
Correlations between the number densities of N, F, and O atoms (, , and , respectively) and for the 1.5-, 3.0-, and -thick RIBD-with-PGP silicon oxide films grown at with ion energy and growth rate using source gas and -base and PG’s of and . The results of the least-squares-fitted experimental linear equations of and as a function of are also shown. The offset values of and extrapolated at and , respectively, were and , respectively.
Article metrics loading...
Full text loading...