The electrical transport properties of epitaxial graphene layers are correlated with the SiC surface morphology. In this study we show by atomic force microscopy and Raman measurements that the surface morphology and the structure of the epitaxial graphene layers change significantly when different pretreatment procedures are applied to nearly on-axis 6H-SiC(0 0 0 1) substrates. It turns out that the often used hydrogen etching of the substrate is responsible for undesirable high macro-steps evolving during graphene growth. A more advantageous type of sub-nanometer stepped graphene layers is obtained with a new method: a high-temperature conditioning of the SiC surface in argon atmosphere. The results can be explained by the observed graphene buffer layer domains after the conditioning process which suppress giant step bunching and graphene step flow growth. The superior electronic quality is demonstrated by a less extrinsic resistance anisotropy obtained in nano-probe transport experiments and by the excellent quantization of the Hall resistance in low-temperature magneto-transport measurements. The quantum Hall resistance agrees with the nominal value (half of the von Klitzing constant) within a standard deviation of 4.5 × 10(-9) which qualifies this method for the fabrication of electrical quantum standards.

Precision measurements of the quantum Hall resistance with alternating current (ac) in the kHz range were performed on epitaxial graphene in order to assess its suitability as a quantum standard of impedance. The quantum Hall plateaus measured with alternating current were found to be flat within one part in 10 7 . This is much better than for plain GaAs quantum Hall devices and shows that the magnetic-fluxdependent capacitive ac losses of the graphene device are less critical. The observed frequency dependence of about -8×10 -8 /kHz is comparable in absolute value to the positive frequency dependence of plain GaAs devices, but the negative sign is attributed to stray capacitances which we believe can be minimized by a careful design of the graphene device. Further improvements thus may lead to a simpler and more user-friendly quantum standard for both resistance and impedance.Graphene is probably the most fascinating electronic material discovered in the last decades [1][2][3]. Among its various unique properties, an anomalous 'half-integer' quantum Hall effect (QHE) is most interesting for metrology, where the fact that the Hall resistance is quantized and depends only on fundamental constants is utilized for the representation and maintenance of the resistance unit, the ohm. Typically, twodimensional electron systems (2DES) realized in GaAs/AlGaAs heterostructures [4] are used for this purpose. The required relative measurement uncertainty of better than 1 part in 10 8 is, however, only obtained at strong magnetic fields around 10 tesla and at temperatures of 1.4 kelvin and below. In contrast, in graphene the cyclotron energy splitting between the Landau levels (which is the main factor determining the robustness of the quantized Hall resistance (QHR)) is so large that fingerprints of the QHE are even observed at room temperature [5]. Thus, with graphene a highly precise QHR standard working at low magnetic fields and temperatures above 4 kelvin is conceivable, which would be an enormous advantage for practical metrology. In fact, when measuring with direct current (dc), it has been demonstrated already that the precision of the QHE in high quality graphene devices matches that of GaAs devices [6][7][8][9]. However, in the forthcoming fundamental constant-based redefinition of the Système International d'Unités (SI) [10], also the impedance units (capacitance and inductance) will be traced to fundamental constants [11]. The most direct way to represent the impedance units is to use a quantum Hall resistance measured with alternating current (ac QHR). This has two advantages. Firstly, deriving the resistance and impedance units from the same quantum effect improves the consistency of the SI. And secondly, using the same QHE device at dc and at ac in one and the same cryomagnetic system would constitute a practical and economical advantage. Therefore, the question naturally arises whether graphene can replace GaAs also in the realm of impedance units, leading to an at least equally precise, but more userfriendl...

We investigate the 1/f noise properties of epitaxial graphene devices at low temperatures as a function of temperature, current, and magnetic flux density. At low currents, an exponential decay of the 1/f noise power spectral density with increasing temperature is observed that indicates mesoscopic conductance fluctuations as the origin of 1/f noise at temperatures below 50 K. At higher currents, deviations from the typical quadratic current dependence and the exponential temperature dependence occur as a result of nonequilibrium conditions due to current heating. By applying the Kubakaddi theory [S. S. Kubakaddi, Phys. Rev. B 79, 075417 (2009)], a model describing the 1/f noise power spectral density of nonequilibrium mesoscopic conductance fluctuations in epitaxial graphene is developed and used to determine the energy loss rate per carrier. In the regime of Shubnikov-de Haas oscillations, a strong increase of 1/f noise is observed, which we attribute to an additional conductance fluctuation mechanism due to localized states in quantizing magnetic fields. When the device enters the regime of quantized Hall resistance, the 1/f noise vanishes. It reappears if the current is increased and quantum Hall breakdown sets in.

No abstract

No abstract