Graphene Annealing: How Clean Can It Be?

Dec 13, 2011 - weak p-doping;11 however, removing the PMMA residue is more difficult than it seems. It cannot be dissolved with any known organic solv...
1 downloads 0 Views 3MB Size
Download PDF
Letter pubs.acs.org/NanoLett

Graphene Annealing: How Clean Can It Be? Yung-Chang Lin,† Chun-Chieh Lu,† Chao-Huei Yeh,† Chuanhong Jin,‡ Kazu Suenaga,‡ and Po-Wen Chiu*,† †

Department of Electrical Engineering, National Tsing Hua University, Hsinchu 30013, Taiwan National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba 305-8565, Japan

‡

S Supporting Information *

ABSTRACT: Surface contamination by polymer residues has long been a critical problem in probing graphene’s intrinsic properties and in using graphene for unique applications in surface chemistry, biotechnology, and ultrahigh speed electronics. Poly(methyl methacrylate) (PMMA) is a macromolecule commonly used for graphene transfer and device processing, leaving a thin layer of residue to be empirically cleaned by annealing. Here we report on a systematic study of PMMA decomposition on graphene and of its impact on graphene’s intrinsic properties using transmission electron microscopy (TEM) in combination with Raman spectroscopy. TEM images revealed that the physisorbed PMMA proceeds in two steps of weight loss in annealing and cannot be removed entirely at a graphene susceptible temperature before breaking. Raman analysis shows a remarkable blue-shift of the 2D mode after annealing, implying an anneal-induced band structure modulation in graphene with defects. Calculations using density functional theory show that local rehybridization of carbons from sp2 to sp3 on graphene defects may occur in the random scission of polymer chains and account for the blue-shift of the Raman 2D mode. KEYWORDS: Graphene, annealing, PMMA, Raman, TEM

G

ease of handling, and wide process latitude. Exposure of PMMA to an electron beam causes scission of the polymer chains, which can be selectively dissolved by a chemical developer, leaving an uncovered area for further processing such as etching or metallization. Graphene nanoribbons and quantum dots have been created using PMMA as an etching mask.9 Graphene field-effect transistors and logic circuits fabricated in laboratories also rely deeply on the PMMA process. Recently, the widespread use of PMMA as a transfer substrate has even become an unavoidable material of choice for isolating graphene sheets grown by chemical vapor deposition (CVD). Our previous studies showed that a thin layer of PMMA residue (1−2 nm) adsorbs on a graphene surface upon contact with graphene.10 An increase of transfer or lithography times causes more PMMA to accumulate on graphene. This thin layer is known to modify graphene’s surface properties and to cause weak p-doping;11 however, removing the PMMA residue is more difficult than it seems. It cannot be dissolved with any known organic solvents. To obtain a clean surface, graphene samples are often thermally treated (150−300 °C) to burn off the PMMA residue after a series of device processes. However, no studies to date address the critical problem of PMMA decomposition and the influences of annealing on graphene, that is, to what level graphene can be cleaned by annealing,

raphene, a two-dimensional honeycomb lattice of carbon, possesses intriguing features that make it attractive for a broad range of applications, particularly involving its electronic structure and transport properties. The charge carriers in graphene can travel with extremely high mobility because of their massless nature in the linear E−k dispersion at low energy. In a free-standing graphene layer, the Fermi energy coincides with the conical points where the conduction and valence bands meet and the electron mobility can reach 200 000 cm2/ (V s) at room temperature, the highest value ever known in pure semiconductors.1,2 For graphene lying on a substrate, however, the electronic properties of graphene are altered,3 and the charge carrier mobility drops accordingly by orders of magnitude. This significant impact on electron mobility occurs because any surrounding medium could act as a dominant source of extrinsic scattering, which effectively reduces the mean free path of carriers.4,5 Likewise, the interaction of graphene with molecules that are physisorbed not only causes doping or scattering but also alters graphene’s electronic structure if adsorption occurs at the edge or at defect sites.6−8 Adsorption-induced scattering becomes evident, especially in quantum Hall transport in which the product of the scattering time and cyclotron frequency must fulfill the condition τωc > 1 if more than one complete revolution without scattering is intended. Poly(methyl methacrylate) (PMMA) is the most popular macromolecule used as an e-beam resist in fabricating graphene devices, possessing excellent film properties, high resolution, © 2011 American Chemical Society

Received: October 24, 2011 Revised: November 18, 2011 Published: December 13, 2011 414

dx.doi.org/10.1021/nl203733r | Nano Lett. 2012, 12, 414−419

Nano Letters

Letter

Figure 1. (a) Process flow of making free-standing graphene with defined Au patterns. (b) Optical image of free-standing graphene with Au patterns on a TEM grid. (c) Schematic illustration of TEM observation of graphene surface in combination with Raman scattering.

Figure 2. TEM images of graphene surface after air and H2/Ar two-step annealing at 200 °C for 2 h. (a) Low-magnification TEM image of graphene. The striplike patterns consist of leftovers of PMMA-A and Cu nanoparticles after annealing. (b) and (c) are the magnified TEM images, showing details of surface cleanliness. The same images are duplicated and colored in the lower panels to distinguish the different layers of PMMA residue that decomposed differently. The areas free of PMMA are shown in gray in the colored images. The lower-left plot shows the corresponding color maps, in which blue, red, and yellow stand for PMMA-G, PMMA-A, and Cu nanoparticles, respectively.

TEM grids (Quantifoil) or silicon substrates. For the samples that combine TEM with Raman measurements, we made additional gold patterns on graphene using e-beam lithography, followed by the same transfer process to TEM holey grids (Figure 1). Electron microscopy characterizations were conducted using a JEOL 2100F transmission electron microscope equipped with a cold field emission gun and two DELTA correctors. Acceleration voltage of 60 kV was used throughout the measurements with specimens at ambient temperature. The TEM images were recorded by a Gatan CCD (model 894) with a typical exposure time of 1 s. Two different batches of samples were tested in annealing. For the first type, graphene sheets were annealed in a tube furnace in air for 1 h, followed by another hour of annealing in a flow of mixed H2 (200 sccm) and Ar (400 sccm). For the second type, annealing was carried out solely in the flow of mixed H2 and Ar under the same conditions. The TEM results for the second type are provided in the Supporting Information. Figure 2 shows the TEM images of graphene annealed at 200 °C in both air and H2 processes for 2 h. The suspended graphene film is subjected to lengthy annealing without macroscopic damage such as cracks or tears (Figure 2a). Prolonged annealing (>2 h) at the same temperature is of little help in surface cleanliness. Our TEM observations showed that the PMMA is decomposed in a two-step scheme: the PMMA facing the air (PMMA-A) has a lower decomposition

determining the parameters that are more effective in this respect, and whether annealing causes any changes in graphene intrinsic properties. We attempt to close this gap. We employed high-resolution transmission electron microscopy (TEM) to study the decomposition of PMMA residues and to provide detailed information on a graphene surface under different annealing conditions. We show that the PMMA residue decomposes in a two-step scheme, depending on the interaction strength with graphene. Lattice defects and local rehybridization of carbons from sp2 to sp3 increase for annealing over a long period or at a high temperature. We found that annealing in air followed by additional annealing with hydrogen provides the cleanest graphene surface. We also fabricated specific gold scaffolds with windows in different shapes to suspend graphene sheets, enabling the combination of TEM observations with Raman measurements before and after annealing in the proximity of the same location imaged by TEM. A significant blue-shift of the Raman 2D mode up to 23 cm−1 was observed in freestanding graphene after annealing, indicative of an annealinduced band structure modulation. Graphene used in the current study was grown on 25 μm thick Cu foils (Alfa Aesar, 99.8% purity) by the CVD method reported previously.12 A thin PMMA layer (MW = 996 kDa) was spun on the graphene/Cu foils for wet etching in an iron(III) chloride aqueous solution and then transferred to 415

dx.doi.org/10.1021/nl203733r | Nano Lett. 2012, 12, 414−419

Nano Letters

Letter

Figure 3. TEM images of graphene after air and H2/Ar two-step annealing at 250 °C for 2 h. (a) and (b) show the details of surface cleanliness, with the same scale in Figure 2b,c for ease of comparison. The same images are duplicated and colored in the lower panels to distinguish the different layers of PMMA residue that decomposed differently. The coloring is the same as that in the schematics of Figure 2. The areas free of PMMA are shown in gray in the colored images. (c) Atomic resolution of graphene clean surface with PMMA residue shown piecewise at the bottom corner after annealing.

annealing time is long (>30 min). As shown in Figure 4, numerous tears in nanoscale appear after annealing in air for

temperature, whereas the scission reaction occurs at a higher temperature for PMMA facing the graphene (PMMA-G). PMMA-A consists of 3−5 layers of PMMA after a typical ebeam lithography or PMMA-assisted transfer process. PMMAG is the bottom layer that is in direct contact with graphene. After annealing, PMMA-A forms a striplike pattern with wrapped nanoparticles (Figure 2a). These particles were identified as CuOx in an amorphous state, which was found to mitigate the mobility reduction due to suppressed charged impurity scattering.13 It accounts for the observed higher mobility after annealing. The PMMA-A can start initial scission at temperatures as low as 160 °C, but 200 °C is more effective for PMMA-A removal. The PMMA-A cannot be entirely cleaned up even under a harsh annealing condition, for example, 350 °C for 5 h. Zooming in the TEM image shows the very thin PMMA-G network close to the percolation threshold (Figure 2b,c). The initial scission of PMMA-G starts at a higher temperature (∼200 °C) than that of PMMA-A, as is evident in the color panels of Figure 2b,c, which highlight the distribution of PMMA-G over the graphene surface. Note that most of the surface is still obscured by PMMA-G after the 200 °C annealing. Figure 3 shows the representative TEM images of graphene under the same annealing conditions with the temperature rising to 250 °C. The density of the striplike PMMA-A is approximately the same as that annealed at 200 °C. However, most of the PMMA-G has been burned out, leaving clean flatlands surrounded by residual PMMA-G networks (Figure 3b,c). Similar to the PMMA-A, the PMMA-G networks fail to decompose further even at higher temperatures to which the graphene sheet is susceptible before breaking (Figure S1 in Supporting Information). Higher-temperature annealing (>250 °C) does not really yield a much cleaner surface, but at the risk of structural damages when graphene is free-standing. We found that the cleanliness of the graphene surface remains far from satisfactory even after annealing at up to 700 °C in a TEM vacuum chamber. Figure 3 presents the typical cleanest graphene surface that we could obtain after systematic tests of different annealing conditions. Annealing in air at 250 °C followed by another annealing in H2/Ar mixture is found to be a more effective recipe for burning both PMMA-A and PMMA-G. The clean area spans from 102 to ∼8 × 103 nm2. However, annealing in the presence of oxygen is in favor of defect formation if the

Figure 4. TEM images of graphene after air and H2/Ar two-step annealing at 200 °C for 2 h. Annealing in the presence of oxygen helps to clean up PMMA, but at the cost of defect formation. The annealinginduced tears can be easily found.

over 1 h at 200 °C. To avoid degradation of graphene quality, single-step annealing in diluted hydrogen may be recommended, although the clean areas are considerably smaller. TEM images of such annealed graphene are provided for comparison in Figures S2 and S3 of the Supporting Information. The PMMA residue on graphene can also be traced by the Xray photoelectron spectroscopy (XPS) spectra. Figure 5 shows the evolution of C 1s core-level spectra with annealing temperature. For peak fittings, the C 1s background spectra are removed using Shirley algorithm, and the sp2 component of C−C bonding is obtained using an asymmetric Gaussian− Lorentzian formula. The main peaks (gray) in Figure 5 are essentially composed of the sp2 and sp3 hybrids. The binding energy of the sp3 hybrids is shifted with respect to the sp2 hybridized carbon by ∼0.7 eV, similar to the value between graphite and diamond reported in previous works.14,15 Some other tails (green, red, and blue) in the spectra correspond to the bonding energy of species in PMMA or functional groups on graphene. Chemical shifts of +1.4, +2.4 ± 0.1, and +4.3 ± 0.1 eV are consistently obtained in each spectrum, assigned to the different chemical environments of carbon atoms in PMMA.16−18 Figure 5b−d shows a clear reduction in the 416

dx.doi.org/10.1021/nl203733r | Nano Lett. 2012, 12, 414−419

Nano Letters

Letter

In Figure 6, we compare the Raman spectra of CVD graphene before and after annealing. Before annealing, graphene surface is covered by a thin layer of PMMA residue due to the wet transfer. No discernible Raman signals of PMMA (typically at 1450 and 1530 cm−1 using 633 nm excitation) are seen in Figure 6a when graphene is lying on a silicon substrate. However, a clear background signal spanning from 1100 to 1700 cm−1 appears when graphene becomes freestanding (Figure 6b). The low-frequency Raman signal can be deconvoluted into four components: the sharp graphene G (1584 cm−1) and D (1324 cm−1) modes (green fits) embedded in two broad peaks (yellow fits) at similar positions originating from the G and D modes of amorphous carbon (mixture of sp2and sp3-bonded carbon).19,20 The overall spectrum in Figure 6b is much less intense than that in Figure 6a because of the removal of the strong substrate reflection, and as a result, the weak signal of amorphous carbon can be highlighted on freestanding graphene. Increasing surface cleanliness by annealing leads to a decreasing amorphous signal and overall spectrum intensity in free-standing graphene (Figure 6d). The spectrum exhibits the lowest intensity for the cleanest graphene surface due to the lack of multiple reflections and interference of light in samples.10 However, none of these prominent features are observed for graphene lying on a substrate (Figure 6c). A remarkable Raman feature that exists in both free-standing and supported graphene sheets is the clear blue-shift of the 2D peak after annealing (Figure 6e,f). The blue-shift is observed consistently in the Raman spectra of all 16 samples examined and spans widely from Δ2D = 3 to 23 cm−1, depending on the annealing temperature. For free-standing graphene, the Δ2D population means are 6, 11, and 12 cm−1 for 200, 250, and 300 °C annealing, respectively (Figure 6g). The monotonic increase of the Δ2D with annealing temperature is also found in the

Figure 5. XPS spectra of PMMA-transferred CVD graphene: (a) before annealing and annealing at (b) 200, (c) 250, and (d) 300 °C. The annealing conditions are the same as the samples shown in the TEM images. The binding energy of the sp2 C−C bonding (gray) is assigned at 284.4 eV, and chemical shifts of +1.4, +2.4 ± 0.1, and +4.3 ± 0.1 eV are assigned for the different carbon atoms in PMMA, as indicated by the schematic inset in (a).

intensity of the C 1s states associated with PMMA after annealing. The presence of PMMA leftovers after annealing is in good agreement with the TEM observations. Note that only little differences are seen in the spectra between 250 and 300 °C annealing, indicating that removing PMMA by annealing becomes inefficient after the temperature reaches a critical point. This coincides well with the TEM results.

Figure 6. (a) and (b) are the Raman spectra of Si-supported and free-standing CVD graphene, respectively. Different Raman spectra are seen in the two different cases. (c) and (d) are the Raman spectra of Si-supported and free-standing CVD graphene after annealing at 250 °C, respectively. (e) and (f) are the comparisons of the 2D peak positions before and after annealing for Si-supported and free-standing CVD graphene, respectively. (g) Histogram of the Δ2D as a function of annealing temperature for both Si-supported and free-standing graphene. (h) Schematic illustration of the electronic structure near the Dirac points of pristine (linear) and annealed (parabolic) CVD graphene. The arrows indicate the double-resonance process in the pristine and annealed CVD graphene. 417

dx.doi.org/10.1021/nl203733r | Nano Lett. 2012, 12, 414−419

Nano Letters

Letter

CO2, CH4, C2H4, C2H6, HCOOCH3) in trace amounts.31 In the typical random scission illustrated in Figure 7b, macroradicals undergo β-scissions to release the monomer, unzipping the polymer chains. However, two simultaneous bond scissions to release a monomer is less probable.32 Under such a circumstance, many macroradicals will be generated in the process of random scissions before the monomers and other stable species are given off. Some of the macroradicals may thus form covalent bonds with graphene on the defect sites (Figure 7c), leading to a local rehybridization of carbons from sp2 to sp3. The rehybridization modifies graphene’s band structure near the Fermi level, and the reduced Fermi velocity can account for the 2D blue-shift after annealing. Notably, no covalent bonds are formed between the ideal graphene and macroradicals. Physisorption of PMMA residue on graphene causes neither band structure modulation nor discernible charge transfer (Supporting Information). The factors with observable effects on the decomposition temperature include molecular mass, polydispersity,33 tacticity,34 and sample dimensions.35 In our experimental findings, the PMMA residue on graphene decomposes in two stages, resembling the multistep decomposition of silica-adsorbed PMMA thin layer in thermogravimetry measurements.36 It indicates that interaction with graphene might interrupt the chain depropagation; that is, a higher fraction of interaction sites would mean a higher decomposition temperature.37 Annealing provides an easy means to remove polymer contaminants on graphene though it remains far from satisfactory for gaining clean surface in large areas. In the random scission of PMMA, generated radicals may react with graphene defects, leading to local rehybridization of carbons from sp2 to sp3 on graphene. On the other hand, the partially formed radical sites can be attached to long and heavy molecular fragments that interact with the surrounding polymer chains, making the PMMA residue even harder to remove at a susceptible temperature before graphene breaks. The development of a polymer-free transfer technique is desirable. Moreover, we also suggest deposition of a removable protective layer prior to the lithography process to keep the graphene surface from being contaminated without the need for postannealing.

supported graphene. Different mechanisms can be applied to explain the observed 2D blue-shift. Compressive strain which results in mode stiffening is ruled out here because of the freestanding structure of our graphene samples. We also exclude any appreciable contribution to the observed shifts by charges. Free-standing graphene is known to be charge neutral, and either type (electron or hole) of doping would lead to an explicit G peak upshift and reduced I2D/IG.8,21−24 Over 80% of the samples examined show no G peak shifts, while the rest exhibit shifts of 1−3 cm−1, which are considered as experimental errors or unintentional doping by gas-molecular adsorption. Another scenario that could account for the 2D blue-shift is proposed in Figure 6h, which illustrates the difference in the double-resonance scattering near the K points before and after annealing. It is anticipated that annealing of graphene with defects such as those in domain boundaries may result in a strong interaction between the graphene and the adsorbates, Band structure modulation near the Fermi energy may occur, leading to a reduced Fermi velocity at low energy, which in turn shifts the 2D peak to a higher frequency, resembling two-layer graphene stacked in a turbostratic order.25,26 Thermal decomposition of PMMA is a complex radical chain reaction and generally accepted that the process involves initiation, depropagation, transfer, and termination reactions.27 Kashiwagi et al. first showed that the thermal degradation of PMMA proceeds in three steps of weight loss.28 The leaststable step (∼165 °C) is initiated by scissions of head-to-head linkages (H−H) because the bond-dissociation energy of H−H linkages is less than that of a C−C backbone bond. The second step is initiated by scissions at unsaturated ends, involving a hemolytic scission β to the vinyl group (Figure 7a). The last

■

ASSOCIATED CONTENT

S Supporting Information *

Additional TEM images of CVD graphene annealed under different conditions and the calculations of MMA−graphene complex band structures by density functional theory. This material is available free of charge via the Internet at http:// pubs.acs.org.

Figure 7. (a) Terminal scission of PMMA. MC stands for methoxycarbonyl (MC = COOCH3). (b) Random scission of PMMA, yielding MMA as the major product. (c) Formation of a covalent bond between a PMMA repeat unit and graphene with 5−7− 7−5 pair defect (marked in yellow). The adsorption configuration is optimized. C, O, and H atoms are shown in gray, red, and white colors, respectively.

■

AUTHOR INFORMATION

Corresponding Author *E-mail: [email protected].

■

step, which has highest activation energy, is initiated by random scission within the polymer chain (Figure 7b). The random C− C scission has shown to be the dominant mechanism that dissociates PMMA, with monomers as the major product.29,30 The generation of monomers is accompanied by the formation of a number of low-molecular-weight stable species (H2, CO,

ACKNOWLEDGMENTS C.J. and K.S. acknowledge the supports from JST-CREST and Grant-in-Aid for Scientific Research from MEXT (#19054017). The other authors also acknowledge the support of the Taiwan National Science Council under Contract NSC 97-2112-M007-016-MY3. 418

dx.doi.org/10.1021/nl203733r | Nano Lett. 2012, 12, 414−419

Nano Letters

■

Letter

(34) Kashiwagi, T.; Horil, H.; Hatada, K.; Kitayama, T. Polym. Bull. 1989, 21, 433. (35) Manring, L. E. Macromolecules 1989, 22, 2673. (36) Zhang, B.; Blum, F. D. Polym. Prepr. 2002, 43, 484. (37) Morgan, A. B.; Antonucci, J. M.; Vanlandingham, M. R.; Harris, R. H. Jr.; Kashiwagi, T. Polym. Mater. Sci. Eng. 2000, 83, 57.

REFERENCES

(1) Morozov, S. V.; Novoselov, K. S.; Katsnelson, M. I.; Schedin, F.; Elias, D. C.; Jaszczak, J. A.; Geim, A. K. Phys. Rev. Lett. 2008, 100, 016602. (2) Bolotin, K. I.; Sikes, K. J.; Jiang, Z.; Klima, M.; Fudenberg, G.; Hone, J.; Kim, P.; Stormer, H. L. Solid State Commun. 2008, 146, 351− 355. (3) Ni, Z. H.; Ponomarenko, L. A.; Nair, R. R.; Yang, R.; Anissimova, S.; Grigorieva, I. V.; Schedin, F.; Blake, P.; Shen, Z. X.; Hill, E. H.; Novoselov, K. S.; Geim, A. K. Nano Lett. 2010, 10, 3868−3872. (4) Ponomarenko, L. A.; Yang, R.; Mohiuddin, T. M.; Katsnelson, M. I.; Novoselov, K. S.; Morozov, S. V.; Zhukov, A. A.; Schedin, F.; Hill, E. W.; Geim, A. K. Phys. Rev. Lett. 2009, 102, 206603. (5) Lafkioti, M.; Krauss, B.; Lohmann, T.; Zschieschang, U.; Klauk, H.; Klitzing, K. V.; Smet, J. H. Nano Lett. 2010, 10, 1149−1153. (6) Schedin, F.; Geim, A. K.; Morozov, S. V.; Hill, E. W.; Blake, P.; Katsnelson, M. I.; Novoselov, K. S. Nature Mater. 2007, 6, 652−655. (7) Berashevich, J.; Chakraborty, T. Phys. Rev. B 2010, 81, 205431. (8) Medina, H.; Lin, Y. C.; Obergfell, D.; Chiu, P. W. Adv. Funct. Mater. 2011, 21, 2687−2692. (9) Ponomarenko, L. A.; Schedin, F.; Katsnelson, M. I.; Yang, R.; Hill, E. W.; Novoselov, K. S.; Geim, A. K. Science 2008, 320, 356−358. (10) Lin, Y. C.; Jin, C.; Lee, J. C.; Jen, S. F.; Suenaga, K.; Chiu, P. W. ACS Nano 2011, 5, 2362−2368. (11) Pirkle, A.; Chan, J.; Venugopal, A.; Hinojos, D.; Magnuson, C. W.; McDonnell, S.; Colomobo, L.; Vogel, E. M.; Ruoff, R. S.; Wallace, R. M. Appl. Phys. Lett. 2011, 99, 122108. (12) Li, X.; Cai, W.; An, J.; Kim, S.; Nah, J.; Yang, D.; Piner, R.; Velamakanni, A.; Jung, I.; Tutuc, E.; Banerjee, S. K.; Colombo, L.; Ruoff, R. S. Science 2009, 324, 1312−1314. (13) McCreary, K. M.; Pi, K.; Kawakami, R. K. Appl. Phys. Lett. 2011, 98, 192101. (14) Chen, C. T.; Sette, F. Phys. Scr. 1990, T31, 119−126. (15) Morar, J. F.; Himpsel, F. J.; Hollinger, G.; Jordan, J. L.; Hughes, G.; McFeely, F. R. Phys. Rev. B 1986, 33, 1340−1345. (16) Gross, Th.; Lippitz, A.; Unger, W. E. S.; Wöll, Ch.; Hähner, G.; Braun, W. Appl. Surf. Sci. 1993, 68, 291−298. (17) Chiang, T. C.; Seitz, F. Ann. Phys. 2001, 10, 61. (18) Yumitori, S. J. Mater. Sci. 2000, 35, 139. (19) Ferrari, A. C.; Robertson, J. Philos. Trans. R. Soc. London, A 2004, 362, 2477−2512. (20) Capano, M. A.; McDevitt, N. T.; Singh, R. K.; Qian, F. J. Vac. Sci. Technol., A 1996, 14, 431−435. (21) Das, A.; Pisana, S.; Chakraborty, B.; Piscanec, S.; Saha, S. K.; Waghmare, U. V.; Novoselov., K. S.; Krishnamurthy, H. R.; Geim, A. K.; Ferrari, A. C.; Sood, A. K. Nature Nanotechnol. 2008, 3, 210−215. (22) Casiraghi, C. Phys. Rev. B 2009, 80, 233407. (23) Mohiuddin, T. M. G.; Lombardo, A.; Nair, R. R.; Bonetti, A.; Savini, G.; Jalil, R.; Bonini, N.; Basko, D. M.; Galiotis, C.; Marzari, N.; Novoselov, K. S.; Geim, A. K.; Ferrari, A. C. Phys. Rev. B 2009, 79, 205443. (24) Lin, Y. C.; Lin, C. Y.; Chiu, P. W. Appl. Phys. Lett. 2010, 96, 133110. (25) Ni, Z.; Wang, Y.; Yu, T.; You, Y.; Shen, Z. Phys. Rev. B 2008, 77, 235403. (26) Poncharal, P.; Ayari, A.; Michel, T.; Sauvajol, J. L. Phys. Rev. B 2008, 78, 113407. (27) Madras, G.; Smith, J. M.; McCoy, B. J. Ind. Eng. Chem. Res. 1996, 35, 1795. (28) Kashiwagi, T.; Inaba, A.; Brown, J. E.; Hatada, K.; Kitayama, T.; Masuda, E. Macromolecules 1986, 19, 2160−2168. (29) Arisawa, H.; Brill, T. B. Combust. Flame 1997, 109, 415. (30) Costache, M. C.; Wang, D.; Heidecker, M. J.; Manias, E.; Wilkie, C. A. Polym. Adv. Technol. 2006, 17, 272. (31) Madorsky, S. L. Thermal Degradation of Organic Polymers; Interscience Publishers: New York, 1964. (32) Stoliarov, S. I.; Westmoreland, P. R.; Nyden, M, R.; Forney, G. P. Polymer 2003, 44, 883−894. (33) Inaba, A.; Kashiwagi, T. Macromolecules 1986, 19, 2412. 419

dx.doi.org/10.1021/nl203733r | Nano Lett. 2012, 12, 414−419