NGUYEN ET AL. VOL. 5 ’ NO. 6 ’ 5273–5279 ’ 2011 www.acsnano.org 5273 May 18, 2011 C 2011 American Chemical Society Temperature and Gate Voltage Dependent Raman Spectra of Single-Layer Graphene Khoi T. Nguyen, Daner Abdula, Cheng-Lin Tsai, and Moonsub Shim * Department of Materials Science and Engineering, University of Illinois, Urbana, Illinois 61801, United States E xceptionally high carrier mobilities and thermal conductivity observed in single-layer graphene have generated much excitement in multiple areas includ- ing low-dimensional physics, high-speed electronics, and thermal management ap- plications. 1À3 Central to most electronics applications is the understanding and con- trolling doping/charging and thermal re- sponses. Raman spectroscopy has been a powerful characterization tool for graphene and related materials especially with re- spect to doping and thermometry. 4À13 For example, Fermi level dependent electronÀ phonon coupling arising from the Kohn anomaly, “charge impurities”, and heat dis- sipation processes in graphene have been examined with the aid of Raman spectro- scopy. 5,6,9,14 In particular, changes in the G-band optical phonon energy and line width as well as G/2D intensity ratio have been frequently used to estimate doping levels and to infer temperature in graphene. 3,5,6,15 There are two key underlying physical pro- cesses that allow such measurements. Strong coupling of G-band phonons to carrier single particle excitations leads to phonon softening and broadening near the charge neutrality point or the Dirac point (DP). 16 Hence the G-band phonon energy and line width are strongly dependent on the Fermi level posi- tion or doping. With respect to heating when the Fermi level position is fixed at the DP, the decrease in the G-band line width is also dominated by the same electronÀphonon coupling process, that is, thermal smearing of electron population distribution near the DP reducing the degree of electronÀhole pair generation that is coupled to the optical phonon transition. 17 The downshift in the G-band phonon energy with increasing tem- perature, on the other hand, is dictated by an- harmonic coupling mainly to acoustic phonons rather than electronÀphonon coupling. 16,17 This downshift in the G-band phonon energy with temperature has often been used to esti- mate the temperature of graphene which has allowed for the use of Raman spectroscopy as a simple yet elegant means of studying thermal conductivity 3,7,18 and energy dissipation in graphene. 9,19 However, the effects of the local chemical environment, especially with respect to interactions with ambient molecules and substrates that can lead to charging, can com- plicate the interpretation of the temperature dependence of the Raman G-band energy and line width. Here, we examine Fermi level posi- tion dependent Raman spectra of single-layer graphene with varying temperature to separate out doping and temperature-induced effects. RESULTS AND DISCUSSION Raman spectra of electrically contacted graphene at the indicated temperatures with applied gate voltage (V g ) ensuring Fermi level position to be at the DP are shown in Figure 1. As expected, both the G and 2D modes downshift with increasing temperature. However, this temperature dependence can be significantly different when the Fermi level position is not ensured to be the same. Figure 2a shows V g dependence * Address correspondence to mshim@illinois.edu. Received for review April 29, 2011 and accepted May 18, 2011. Published online 10.1021/nn201580z ABSTRACT Raman spectra of electrostatically gated single-layer graphene are measured from room temperature to 560 K to sort out doping and thermally induced effects. Repeated heating cycles under Ar led to convergent first-order temperature coefficients of the G-band (χ G = À0.03 cm À1 /K) and the 2D-band (χ 2D = À0.05 cm À1 /K) frequencies, which are independent of doping level as long as the Fermi level does not shift with temperature. While the intrinsic behavior may be different (e.g., χ G ∼ À0.02 cm À1 /K near room temperature), these values appear more appropriate in describing responses of most graphene samples on SiO 2 substrates. The more negative χ G value than theoretical expectations may be explained by interactions with the substrate reducing the lattice thermal expansion contribution to the temperature dependence of G-band frequency. Enhanced interactions with the substrate may also be responsible for zero-charge, room- temperature G-band line width increase and 2D-band frequency downshift. KEYWORDS: graphene . electrostatic gating . Raman spectroscopy . thermometry . substrate ARTICLE