Fundamental Studies of the Silicon Carbide MOS Structure

Mrinal Kanti Das, Purdue University

Abstract

The ability to thermally grow SiO2 on silicon carbide (SiC) has made Metal Oxide Semiconductor (MOS) devices feasible in SiC. Critical in their development is a thorough understanding of the SiC MOS structure itself. Three important steps have been accomplished toward this goal. First, tools to characterize the oxide/SiC interface have been developed. The room temperature photo C-V, the high-low, and conductance-frequency methods constitute quite an arsenal of interface characterization techniques. However, due to their inability to measure across the wide bandgap and the need for the MOS capacitor test structure, the current arsenal needs to be complemented with the charge pumping (CP) method which provides DIT across midgap and is measured directly on MOSFETs. Second, the interface has been optimized. Work to in this area has lowered DIT to the mid-1010 cm−2eV−1 range near midgap and Q F to a record low 5 × 1011 cm−2. But, we have made direct measurements of a high density of interface states near the conduction band. Hence, future work must concentrate on the conduction band states before SiC enjoys a true device quality interface. The bottom line is that SiC is still missing the analog of the hydrogen anneal from silicon MOS technology. Finally, in order to transfer the technology from test structures to devices, device behavior must be better understood, namely the inversion channel mobility, which currently is the limiting agent in MOS device development. Mobilities on 6H-SiC approach a record high 100 cm2/Vs while the corresponding values on 4H-SiC are at least an order of magnitude lower for simultaneously processed MOSFETs. Process optimization of the 4H has resulted in mobilities of 25 cm2/Vs. The low mobilities on these devices can be explained with the high density of interface states near the conduction band. They lower mobility by trapping and Coulombically scattering the mobile carriers in the channel. Clearly, the future of SiC MOS hinges on the reduction of the bandtail states near the conduction band.

Degree

Ph.D.

Advisors

Cooper, Purdue University.

Subject Area

Electrical engineering|Electromagnetics|Materials science

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS