Modeling and Characterization of Thermally Oxidized 6H Silicon Carbide

Abstract

Wet oxidation and computer simulation of metal oxide semiconductor (MOS) structures were studied on n‐type (Si‐face and C‐face) and p‐type (Si‐face) 6H‐normalSiC . The effects of thermal oxidation conditions at temperatures between 1100 and 1250°C on the electrical properties of MOS capacitors were studied. Oxidation model parameters for SSUPREM3 are presented, and simulation results are compared to laboratory studies with good agreement. The C‐V characteristics of the MOS capacitors were measured at high frequency in the dark and under illumination at room temperature. Samples prepared by wet oxidation showed accumulation, depletion, and inversion regions, but they did not show inversion regions under dark conditions. The interface trap densities and emission time constants of fast states were determined by ac conductance measurements. The electrical properties of normaln‐normaland p‐normaltype 6H‐normalSiC were compared. It was found that oxidation rate of C‐face was five to seven times faster than that of Si‐face on normaln‐normaltype SiC . The electrical properties of Si‐face and C‐face of normaln‐normaltype 6H‐normalSiC showed that the quality of oxide on the Si‐face was superior to the oxide on the C‐face.

Limiting Step Involved in the Thermal Growth of Silicon Oxide Films on Silicon Carbide

“…However, these excitation curves could be well fitted using a similar profile to the one shown in the inset, only its starting position being deeper in the sample, corroborating the result obtained analyzing the areas under the excitation curves. The results from Si samples confirm the idea that the oxide thermal growth process is limited by diffusion of the oxidant species in the existing SiO 2 layer, since for thicker Si 16 [19]. On the other hand, for SiC the behavior can be explained by an oxide thermal growth mechanism in which the oxidant species reacts in a region near (and including) the oxide=Si interface and not at an abrupt interface as in the case of Si samples.…”

“…From previous works it is known that in the oxide thickness range of the present work the 16 O 2 = 18 O 2 oxidation sequence leads to the incorporation of 18 O in near surface and near-interface regions for both SiC [17,21,22] and Si [20,23,24] indicating that the oxidant species is mobile in both cases. In the case of Si oxidation, 18 O fixation near the surface is due to 16 O- 18 O exchange mediated by defects existing in this region [23]. 18 O fixation near the interface is due to interstitial diffusion of O 2 across the growing oxide without interacting with it and reaction with Si at the interface, forming SiO 2 , while Si was shown to be practically immobile being absent in its nonoxidized form away from the interface region [25].…”

“…Investigations of the thermal oxide film on SiC in its different regions have been pursued [2 -7], indicating that in the surface and bulk regions the oxide is similar to that grown on Si [4,8,9], that the interface is less abrupt [5,10], and that incompletely oxidized C or Si are found near the interface [2,3,6,7,[11][12][13][14]. However, a thorough understanding of the phenomena taking place during the thermal growth of SiO 2 on SiC is still missing.Growth of SiO 2 on SiC is much slower than on Si, presenting more scattered data [15][16][17] and differences along the polar directions [ 0001 and (0001) in the hexagonal polytypes]. Several authors have tried to model the kinetics of the thermal growth of SiO 2 on SiC [15,17,18], mostly within the framework of the Deal and Grove model [19], which satisfactorily describes the thermal growth of SiO 2 on Si in dry O 2 for oxide thicknesses above 20 -30 nm.…”

“…Growth of SiO 2 on SiC is much slower than on Si, presenting more scattered data [15][16][17] and differences along the polar directions [ 0001 and (0001) in the hexagonal polytypes]. Several authors have tried to model the kinetics of the thermal growth of SiO 2 on SiC [15,17,18], mostly within the framework of the Deal and Grove model [19], which satisfactorily describes the thermal growth of SiO 2 on Si in dry O 2 for oxide thicknesses above 20 -30 nm.…”

“…One set of SiC samples was thermally oxidized in a single batch in 16 O 2 (45 h, 1100 C, 100 mbar), etched in a HF water solution (1 nm=s etching rate) for different times (1, 10, 20, and 30 s) and then all oxidized in 18 O 2 (1 h, 1100 C, 100 mbar). The aim was to generate different thicknesses of Si 16 O 2 (due to the different etching times) that the oxidizing species would traverse during the second oxidation step in 18 O 2 , but identical Si 16 O 2 =Si interfaces where it would react to form new oxide, since HF attacked only the external surfaces.…”