Controlling Epitaxial Graphene Growth
Monday, December 13, 2010 – 1:30pm
Advanced SiC Epitaxial Research Laboratory, U.S. Naval Research Laboratory, Washington, DC 20375
In the last two years, tremendous progress has been realized in the fabrication of epitaxial graphene (EG) RF devices . However, improving the control of EG growth – thickness and doping uniformity ‐ is crucial for continued success. It is natural to think that understanding the initial steps of graphene growth is paramount to future development of control strategies. In this regard, we present recent results in two areas of EG growth synthesis: graphene island formation on (000‐1) 6H‐SiC and single layer graphene growth on (0001) 4H‐SiC step‐free mesas (SFMs). Through control of temperature and Ar pressure, graphene epitaxy can be slowed, resulting in island growth on the C‐face of SiC . The islands are thought to represent early stages in graphene growth, and in all cases examined, originated from threading screw dislocations associated with the substrate. We followed the growth process from island to complete film using optical and scanning electron microscopies and Raman spectroscopy. Island centers were generally thicker than the edges. During island expansion, coalescence of adjacent islands naturally occurred, which resulted in thickness inhomogenieties. The island crystallite length scale was extracted using the ratio of the Raman D and G band intensities; this length scale was found to increase with increasing island lateral dimension, eventually achieving the values associated with complete films. To aid in understanding EG growth without the impact of substrate defects, we investigated growth on SFMs. SFMs were formed by a kinetically‐controlled lateral step‐flow SiC growth process at 15800C on (0001) 4H‐SiC substrates patterned with mesas . When threading screw dislocations are not present, the SiC growth process results in atomically flat mesas. Subsequently, EG was grown in a 100 mbar Ar ambient at 1620°C on an array of SFMs with side lengths ranging from 40μm to 200μm. For short growth times, partial graphene coverage of SFMs was observed suggesting a growth mechanism limited, in part, by C surface diffusion. For long growth times, complete EG mesa coverage was established and the step bunching morphology typically observed on conventional basal plane substrates was not found. For the long growth time case, analysis of the 2D Raman signal implied two layers of graphene were present in the central area of the mesa. In contrast, analysis of the 2D Raman signal from mesa edges suggested at least 3 layers were present. In addition, the graphene in the central region of the SFMs was approximately strain‐free whereas the graphene at the edges was more strained. These latter properties differ substantially from EG grown on conventional basal plane substrates which exhibit significant strain.
 J.S. Moon, et al, IEEE Electron Device Letters 31, 260, (2010).
 J.L.Tedesco, et al, Appl. Phys. Lett. 96, 222103 (2010).
 P.Neudeck, et al, Mater. Res. Soc. Symp. Proc. 911, 85 (2006).
Host: Paola Barbara