ABSTRACT
CONTENTS 20.1 Introduction ...................................................................................................................... 575 20.2 SiC MOSFET and the Channel Mobility Problem ...................................................... 577 20.3 Oxidation of SiC............................................................................................................... 580 20.4 Oxide and Interface Composition ................................................................................. 583
20.4.1 Ion Scattering Spectrometry .............................................................................. 584 20.4.2 TEM and EELS .................................................................................................... 585
20.5 SiO2=SiC Interface Trap Density Measurements......................................................... 586 20.6 Interface Trap Passivation in SiO2=4H-SiC.................................................................. 589
20.6.1 Interface Trap Passivation by NO.................................................................... 589 20.6.2 Interfacial Nitridation......................................................................................... 592 20.6.3 SiO2=SiC Interface Trap Passivation Using Hydrogen ................................. 595 20.6.4 Alternate Wafer Orientations ............................................................................ 597 20.6.5 Oxidation in Sintered Alumina Environment ................................................ 602
20.7 Atomic-Level Defects at the SiO2=SiC Interface and Passivation Mechanisms ..... 603 20.7.1 Carbon-Related Defects at or Near the SiO2=SiC Interface .......................... 604 20.7.2 Defects in the Oxide ........................................................................................... 607
20.8 Remaining Issues at the SiO2=SiC Interfaces and Concluding Remarks................. 608 Acknowledgments ..................................................................................................................... 609 References.................................................................................................................................... 610
Research and development throughout the last decade has led to the emergence of silicon carbide (SiC) electronics. SiC’s potential for high-temperature, high-power, and highfrequency electronics arises out of attractive material properties such as wide band gap, high critical breakdown field, high thermal conductivity, and high electron saturation velocity. These properties, coupled with extreme chemical inertness and mechanical hardness, make SiC extremely attractive for electronics operating under extreme conditions. In general, it has been widely recognized that replacement of conventional Si by a wide
band gap material would lead to substantial socioeconomic gains in niche application such
in vehicles, and sensing technology. A 1995 report by the National Research Council [1] addressed many of the issues related to the development of ‘‘next-generation’’ semiconductors that could replace Si-based devices and operate reliably under extreme conditions. Since then, significant advances have taken place with regards to wide band semiconductor materials such as SiC, GaN, and diamond (see Appendix). Among all these competing materials, SiC technology is currently the most mature with respect to device processing. Currently, SiC Schottky diodes and metal-semiconductor field-effect transistors (MESFETs) are available commercially. Power metal-oxide-semiconductor field-effect transistors (MOSFETs) are expected to be available within the next couple of years. The market for SiC diodes and transistors is expected to exceed $50 million by the year 2009 [2]. SiC has numerous polytypes determined by the stacking sequence of the Si-C bilayers in
the crystal structure, with each polytype possessing significantly different electronic properties [3,4]. The 4H and 6H polytypes are most common from a commercial standpoint. Some of the properties of 4H-SiC and 6H-SiC compared to Si are summarized in Table 20.1. 4H-SiC has a wider band gap and a higher and more isotropic bulk electron mobility compared to 6H-SiC. These characteristics make 4H the preferred polytype for high-power and high-temperature applications. From a technological standpoint, the success of any novel metal-oxide-semiconductor (MOS) technology primarily depends on the following factors-the fourth is a particular advantage for SiC over many competing technologies:
1. Availability of large area single crystal wafers of the and high-quality epitaxial layers
2. Effective doping processes for producing n-and p-type materials
3. Suitable metallization schemes for ohmic contact formation
4. Formation of a defect-free and reliable gate dielectric
SiC meets each of these demands as a result of significant breakthroughs and continual development in each of these areas. Although SiC device processing has not yet attained the degree of perfection of Si processing, rapid developments are underway for each of the above issues. Wafer size and quality has been steadily growing while the cost per unit area has been decreasing. At present, 400 diameter wafers with low defect densities (<15 micropipes cm2) are commercially available. Ion implantation of N and P (for n-type) and B and Al (for p-type) are used for producing extrinsic SiC, and various metallization schemes are available for reliably producing ohmic and rectifying contacts. Many issues remain regarding wafer quality, implantation, activation annealing processes and ohmic contact formation, particularly on p-SiC. Each of these topics defines separate and extensive research problems, well beyond the scope of discussion in this chapter
TABLE 20.1
Selected Properties of Si, 4H-SiC, and 6H-SiC
i.e., formation of high-quality gate dielectrics for SiC MOS devices. In Section 20.2, we will introduce the classic mobility problem in 4H-MOSFETs. We will discuss in detail the impact of the dielectric-semiconductor interface in relation to SiC MOSFET operation. Next we will take a closer look at the oxidation process in SiC, followed by a discussion of bulk and interfacial composition of oxides grown on SiC. This will be followed by a basic discussion of interface characterization in SiC MOS devices using electrical measurements. In Section 20.6, we will review state-of-the-art interface-trap passivation processes. This will be followed by a discussion on the current understanding of the origin of interface defects and atomic-scale passivation mechanisms. Finally, we will comment on the most relevant problems for state-of-the-art oxides grown on SiC with regards to future device development. We also note that studies of the SiO2=SiC interface on different polytypes with different
band gaps and employing different crystal faces provide a unique ‘‘laboratory’’ for interface physics. There is no other semiconductor=dielectric system that allows these broad variations, particularly involving SiO2, the most widely used gate dielectric. Historically this ability to vary the properties of the semiconductor has played a large role in formulating models of the interface and the nature of the interfacial defects.
