Session H22: Epitaxial Graphene on Silicon Carbide

Controlled growth of monolayer graphene on silicon carbide

Controlled thermal decomposition of silicon carbide is so far the most effective method for growing graphene epitaxially and at the wafer scale. In this work we study the graphenization of SiC(0001) and SiC($000\bar{1}$) as a function of temperature and buffer-gas pressure in a custom-built ultrahigh vacuum (UHV) induction furnace. In-situ characterization by both Auger electron spectroscopy and low-energy electron diffraction (LEED) was used to determine the pressure-temperature ``phase boundary'' for the formation of monolayer graphene, and the transient growth rate of graphene layers otherwise. Sample quality was further assessed ex-situ using a variety of techniques such Raman spectroscopy and scanning tunneling microscopy. The effect of buffer gas was modeled with kinetic theory.   

The role of carbon surface diffusion on the growth of epitaxial graphene on SiC

Growth of high quality graphene films on SiC is regarded as one of the more viable pathways toward graphene-based electronics. Graphitic films form on SiC at elevated temperature because of preferential sublimation of Si. Little is known, however, about the atomistic processes of interrelated SiC decomposition and graphene growth. We have observed the formation of graphene on SiC by Si sublimation in an Ar atmosphere using low energy electron microscopy, scanning tunneling microcopy and atomic force microscopy. This work reveals that the growth mechanism depends strongly on the initial surface morphology, and that carbon diffusion governs the spatial relationship between SiC decomposition and graphene growth. Isolated bilayer SiC steps generate narrow ribbons of graphene, whereas triple bilayer steps allow large graphene sheets to grow by step flow. We demonstrate how graphene quality can be improved by controlling the initial surface morphology specifically by avoiding the instabilities inherent in diffusion-limited growth. This work is supported by the LDRD program at Sandia Labs, and the US DOE Office of Basic Energy Sciences, Division of Materials Science and Engineering (DE-AC04-94AL85000), and was performed in part at CINT (DE-AC04-94AL85000).   

Insights into the epitaxial growth of graphene on SiC substrate: A computational study

Experimentally, the epitaxial growth of graphene on SiC substrate has been observed for both the Si-terminated $(0001)$ or C-terminated $(000\bar {1})$surface of 4H-SiC or 6H-SiC wafers, respectively at sufficiently high temperatures in ultrahigh vacuum [Surface Science \textbf{600}, 3906 (2006); PRB \textbf{77}, 155303 (2008)]. However, the mechanism of the sublimation of Si atoms and the graphitization of excess C atoms on the surface of 4H-SiC or 6H-SiC wafers that leads to the epitaxial growth of graphene on SiC is still unclear. The purpose of this work is to conduct a temperature-dependent study of the evolution of 4H-SiC surfaces using the molecular dynamics scheme based on the SCED-LCAO Hamiltonian [PRB \textbf{74}, 155408 (2006)] so that the evolution of the surface reconstruction of 4H-SiC including the formation of the interface between the substrate and graphene layers can be understood at the microscopic level.   

First principles calculation of structural and electronic properties of the 5x5- SiC(0001) reconstructed surface

In the process of graphene layer growth via sublimation of Si atoms from SiC crystals, several reconstructions of the underlying SiC surface arise. Tunneling microscopy and spectroscopy experiments have revealed a rich spectrum of surface states in $5\times5$ and $6\sqrt{3}\times6\sqrt{3}-R 30^{\circ}$ reconstructions on the 6H-SiC(0001) surface that are important for the electronic properties of overlaying graphene layers. We have carried out a detailed study of the structural and electronic properties of the $5\times5$ reconstruction using Density Functional Theory methods as implemented in the SIESTA code. The model consists on a carbon adlayer that fits onto a perfect $5\times5$ SiC(0001) surface, with adatoms arranged with specific spatial separations. Preliminary results suggest that these reconstructions are not caused by the existence of particular stacking sequences (known as S1, S2 and S3) appearing along the c-axis during growth of the SiC crystal. The structural and electronic parameters obtained show good agreement with experimental observations and give insight into the electronic properties of graphene layers obtained with these methods.   

Band Engineering and Magnetic Doping of Epitaxial Graphene on SiC (0001)

Advances in the epitaxial growth of graphene films on SiC have the potential to open new classes of device applications that may revolutionize the semiconductor roadmap for future decades. However, this progress will require an in-depth understanding and utilization of the electronic processes that take place at the nanoscale, in particular the role of the interface buffer layer, where most of the electronic properties of the system can be controlled. In analogy with the formation of the Schottky barrier in metal-insulator interfaces (the energetic barrier the electrons have to overcome to go from the valence band of the metal to the conduction band of the insulator) here we demonstrate the ability to tune and control the band alignment and the magnetic doping at the heterojunction between graphene and SiC, a fundamental requirement for improving device efficiency and applicability. Using first principles calculations, we will show how the surface electrostatic distribution can be used to tune the valence band offset by introducing surface impurities such as B, Al, N, and P. Similarly, we will demonstrate how the introduction of magnetic impurities in the buffer layer can tune the spintronic behavior of the epitaxial graphene layer. This work was supported, in part, by the NERC/NIST SWAN-NRI and the DARPA/HRL CERA programs.   

Charge carrier density and mobility uniformity in SiC(0001) epitaxial graphene

In order to optimize electronic transport in epitaxial graphene-based field effect transistors (FETs), the scattering mechanisms and their limits on charge carrier mobility and saturation velocity need to be understood. We evaluated the effects of charge impurity and phonon scatterings in FETs and Hall bar structures (with and without a thermally deposited SiO$_2$ overlayer). Devices typically featured electron mobilities between 3,000 and 4,000 cm$^2$/Vs at 2 K. With the application of a magnetic field up to 9 T, the emergence of quantum Hall plateaus in the Hall bar structures was apparent. However, we found variation in the resistance values at the plateaus caused by charge concentration inhomogeneity in the graphene. The effect of inhomogeneity on charge transport is further evidenced by a linear dependence of inverse mobility on charge concentration, providing evidence that charged scatterers in the deposited dielectric rather than phonons limit the mobility at all temperatures. Furthermore, we found that charge concentration inhomogeneity due to a dielectric overlay also affects current saturation in epitaxial graphene FETs. We will also present related work on planar tunnel junction studies of bandgap engineering in bilayer graphene.   

Conductance Anisotropy in Epitaxial Graphene Sheets Generated by Substrate Interactions

A Four-Probe UHV STM was used to examine graphene grown epitaxially on SiC. By placing all four probes on a terrace or across multiple terraces, we show that current flows more readily along the terraces than across them. The cause of this anisotropy is the interaction between the graphene and morphology of the substrate. We propose a model where charge buildup at the step edges lead to local scattering of charge carriers. This result highlights the importance of creating large uniform terraces in epitaxially grown graphene on SiC for future use in device applications. More generally, it suggests that no matter what method is used to generate the graphene a careful consideration of the underlying substrate will be required to achieve optimal performance.   

Pronounced half integer quantum-Hall effect in gated epitaxial graphene grown on SiC (0001)

Large-area epitaxial graphene film by thermal decomposition of SiC wafer has provided a promising way to a viable electronics technology. By inserting a fully oxidized nanometer thin aluminum film as a seeding layer, graphene Hall-bar devices fabricated on the Si-face of SiC (0001) with high-$k$ ALD Al$_{2}$O$_{3}$ gate stack were examined. At low temperatures, the half integer quantum-Hall effect in Hall resistance is observed along with pronounced Shubnikov-de Haas oscillations in diagonal magneto-resistance, which confirmed that the electrical properties of epitaxial graphene on SiC are essentially the same as those in exfoliated graphene films [1-4]. With top-gate modulation, the last quantum Hall plateau is especially pronounced and retains well-defined even at temperatures as high as 70K, reaching the temperature limit of the present experimental setup. From weak localization peak fitting, the phase coherence length of the gated epitaxial graphene is determined to be $\sim $ 1$\mu $m at 1K and $\sim $ 200 nm at 70 K. [1] T. Shen et al. Appl. Phys. Lett. 95, 172105 (2009). [2] J. Jobst et al., arXiv:0908.1900v1. [3] X. Wu et al., arXiv:0908.4112. [4] A. Tzalenchuk et al., arXiv:0909.1220.   

Rotational Stacking Order in C-Face Epitaxial Graphene

Multilayer epitaxial graphene (MEG) grown on the $(000\bar{1})$ face of SiC is comprised of a high density of rotational (non-AB-stacked) graphene sheets that each retain the special electronic properties of an isolated graphene layer. Transport measurements indicate that the AB planes in the film can themselves be considered as low density faults but the actual density of AB planes is not known. We present a combination of surface x-ray diffraction (SXRD) and angle resolved photoemission spectroscopy (ARPES) experiments that can measure the AB fraction. Although x-rays can penetrate the entire films, they only give averaged stacking information. As we will show, both the density of the rotational planes and their stacking order are important to calculate the AB fraction. By combining ARPES with the SXRD data, we can give a relatively accurate density and distribution of AB planes in MEG films.   

Electron Paramagnetic Resonance Studies of Multi-layer Epitaxial Graphene Formed on Semi-Insulating 4H- and 6H-SiC Substrates

Electron paramagnetic resonance (EPR) experiments were performed at 9.5 GHz on a set of multi-layer epitaxial graphene (MEG) samples. The films ($\sim $20-30 layers) were formed via desorption of Si from the C-faces of SI 4H- and 6H-SiC substrates at $\sim $1350 $^{o}$C under vacuum in a commercial SiC epitaxy reactor. Additional characterization included room-temperature Raman measurements and Hall effect measurements at both 77 and 300 K. EPR between 4.2 and 50 K revealed a single paramagnetic resonance line with an isotropic Zeeman splitting g-value of 2.003 and FWHM of $\sim $4.5 G. Most notably this feature was not observed from EPR of the parent SiC substrates alone. Electron paramagnetic resonance was also performed on a highly-oriented pyrolytic graphite (HOPG) reference sample and the usual anisotropic feature (g$_{\bot }$=2.003, g$\vert \vert $=2.050; where $\vert \vert $ refers to the c-axis) associated with charge carriers in crystalline graphite was found. Work to determine if the EPR feature in these MEG samples is associated with charge carriers or defects within the films or at the film/substrate interfaces will also be discussed.   

Selective Graphitization of Silicon Carbide: Effect of Argon Background Pressure and Transport Measurements on the Epitaxial Graphene

Electronic quality epitaxial graphene has been selectively grown on SiC in areas not capped by aluminum nitride (AlN). The argon (Ar) pressure during growth is an important parameter for the selectivity of the graphitization process. Atmospheric pressure inhibits growth even in non-capped regions, while high vacuum allows growth over the entire surface. With an intermediate Ar pressure of 100 Pa for 20 min, the molecular-beam epitaxial (MBE) AlN withstands high graphitization temperatures of 1420$^{\circ}$C inhibiting graphene growth under it. Graphene hall-bars were successfully fabricated using this method with no exposure of the graphene to external chemicals, such as resists and etchants that deteriorate the performance of graphene. The hall-mobility measured is about 600 cm$^{2}$/Vs, which can be further enhanced by fine-tuning the Ar pressure and improving the quality of SiC surface prior to graphitization.   

Thermoelectric characterization of large area graphene grown on SiC

The thermoelectric power (TEP) of the epitaxially grown large area graphene was studied. Graphene multi-layers were obtained on C- terminated surfaces of 4H-SiC while monolayers of graphene were obtained from Si- terminated surface of 4H-SiC by thermal decomposition. Electrostatic deposition technique was sequentially used to reduce the number of graphene layers from the C-face samples to a monolayer. We measured and compared the thermoelectric power of monolayer graphene on both Si-face and C-face of SiC. All investigated multi-layer graphene samples showed a positive Seebeck coefficient in ambient conditions and turned negative after vacuum-annealing at 550 K in a vacuum of 2 x 10$^{-7}$ Torr. In contrast, monolayer graphene for both Si- and C- faces showed a relatively small negative Seebeck coefficient in ambient conditions and saturated at a greater negative value after vacuum-annealing. We also measured the response of the TEP while exposing the degassed graphene to various gases. Charge transfer effects were seen for both acceptor and donor type gasses. Finally the gate dependence of the TEP of the large area graphene was studied using a polymer top gate.   

A New Method to Prepare Large High Quality Epitaxial Graphene Samples

One of the roadblocks to graphene technology is the difficulty to produce large uniform samples. We have developed a new way to produce large high quality graphene. The structure and morphology of the epitaxial sample are characterized by angle-resolved photoemission spectroscopy (ARPES), low-energy electron diffraction (LEED), Raman spectroscopy and atomic force microscopy (AFM). The results demonstrated significant improvement of surface smoothness and an increase of terrace size, compared to graphene prepared by normal vacuum annealing.   

Nano-Objects Developing at Graphene/Silicon Carbide Interface

We use scanning tunneling microscopy and spectroscopy to study epitaxial graphene grown on a 4H-SiC(000-1)-C-face substrate. The results reveal amazing nano-objects at the graphene/SiC interface leading to electronic interface states. Their height profiles suggest that these objects are made of packed carbon nanotubes confined vertically and forming mesas at the SiC surface. We also find nano-cracks covered by the graphene layer that, surprisingly, is not broken, with no electronic interface state. Therefore, unlike the above nano-objects, these cracks should not affect the carrier mobility.   

Thermally Induced Folds in Exfoliated Graphene

Graphene samples were prepared on various substrates via mechanical exfoliation. The samples were then annealed in vacuum at temperatures from 200 \r{ }C to 1000 \r{ }C. AFM images showed that folds were generated on graphene's surface above a critical temperature after annealing. The top of folds were enveloped by wave-like structure. The mechanism for the fold formation is believed to be due to differences in thermal expansion between the graphene and the substrate. We will discuss the dependence of fold formation on annealing temperature, graphene-substrate interaction, graphene thickness, and presence of graphene defects. We believe these results are relevant to the understanding of similar fold formation in both CVD and SiC epitaxial growth of graphene.