Raman study on the effects of annealing atmosphere of patterned graphene

Research article Received: 15 March 2017 Revised: 26 September 2017 Accepted: 27 September 2017 Published online in Wiley Online Library: 9 November 2017 (wileyonlinelibrary.com) DOI 10.1002/jrs.5280 Raman study on the effects of annealing atmosphere of patterned graphene Jangyup Son, a,c Minkyung Choi, b Jongill Hong a * and In-Sang Yang b * Despite extensive research on graphene, there are still lacks of understanding the structural changes under harsh stress environments such as high current in uncontrolled atmosphere. Investigating the structural changes of graphene patterned into device at high temperature would be important as the electrical current path of graphene becomes narrower and thus stronger in heat dissipation. In this paper, we performed a comparative study of the structural and electronical changes of graphene for as-grown graphene and patterned graphene in a microbridge shape heated up to 500 C in air or Ar. While the as-grown graphene heated in air or Ar was nearly free from the structural changes, the microbridge graphene exhibited strong structural changes after annealing in Ar, i.e. the broadening in the G and the 2D peaks remained even after cooling back to 30 C. On the other hand, the microbridge graphene heated in air was observed to release stress probably due to formation of vacancies by oxygen adsorption. The different behavior of microbridge graphene heated in Ar from that of as-grown graphene in the same condition is obvious. This means that graphene in microscopic devices should be treated differently from the as-grown graphene. Copyright 2017 John Wiley & Sons, Ltd. Keywords: patterned graphene; annealing effect; deformation; defects; voids Introduction Graphene, a two-dimensional (2D) monolayer consisted of carbon atoms arranged in a honeycomb lattice, has attracted immense attention because of its unusual and exceptional electronic properties [1 3] since it was discovered by micromechanical cleavage. [4] Because graphene is a zero gap material, regarded as semimetal, with a linear dispersion at the Fermi energy, its electronic characteristics, such as electric-field effects, high carrier mobility, and outstanding conductance, allow one to address basic questions of quantum phenomena in 2D materials and also open a new avenue in carbon-based electronics. [5 10] Moreover, the 2D atomic structure makes graphene an ideal candidate for applications of not only nanoelectronic devices but also chemical and biological sensors. [11 14] For various applications, a surface state or a structural change of graphene is the most important factors in modifying the physical and chemical properties of graphene, and its effect on the properties of graphene is substantially intensified because of a purely 2D nature of graphene. In particular, as graphene is patterned to nanostructured devices and experiencing high current density in various atmosphere, the thermal effect can be lethal to a device operation because the elevated temperature resulted from Joule heating can damage the originally designed surface structures of graphene. [15 18] When the temperature of graphene increased under air atmosphere, the exposure of a surface can bring about unwanted p-doping in graphene due to oxygen adsorption. [16] In addition, the structure can be changed because of the difference in thermal expansion between substrate and graphene. [17,19] Therefore, an understanding of the thermal effect on graphene is essential and indispensable to design graphenebased devices operating at high temperature. In this study, we address such a thermal effect on the structural and chemical changes in graphene under air or Ar gas as a function of temperature. In particular, we adopt in-situ Raman spectroscopy to observe the evolution of chemical states and the structural changes. We focus not only on the different annealing atmosphere but also on the difference in the size and the shape of graphene samples: One is unpatterned graphene of large area (as-grown graphene) and another is a graphene sample patterned into a microbridge shape (microbridge graphene). Annealing in air introduced p-doping for both graphene samples of as-grown and microbridge shape, which is consistent with our previous finding. [16] What is surprising in this work is that the behavior of microbridge graphene heated in Ar is quite different from that of as-grown graphene in the same condition, meaning that the graphene in microscopic devices should be treated differently from the as-grown graphene. Experimental Graphene was synthesized by a chemical vapor deposition method [20,21] on a high purity copper catalyst (Alpha aecer, * Correspondence to: Jongill Hong, Department of Materials Science and Engineering, Yonsei University, 50 Yonsei, Seodaemun-gu, Seoul 03722, Korea. E-mail: hong.jongill@yonsei.ac.kr In-Sang Yang, Department of Physics, Ewha Womans University, Daehyeon-dong, Seodaemun-gu, Seoul 03760, South Korea. E-mail: yang@ewha.ac.kr a Department of Materials Science and Engineering, Yonsei University, 50 Yonsei, Seodaemun-gu, Seoul 03722, South Korea b Department of Physics, Ewha Womans University, Daehyeon-dong, Seodaemungu, Seoul 03760, South Korea c Department of Mechanical Science and Engineering, University of Illinois at Urbana-Champaign, Urbana, IL, 61801, USA 183 J. Raman Spectrosc. 2018, 49, 183 188 Copyright 2017 John Wiley & Sons, Ltd.

J. Son et al. 99.999%) under H 2 atmosphere (70 mtorr, 3 sccm) with methane used as a hydrocarbon source (650 mtorr, 30 sccm). As-grown graphene on copper catalyst was spin-coated with polymethylmethacrylate and back-side graphene was etched using oxygen plasma. The copper catalyst was finally etched in 1.5 wt% ammonium persulfate solution. After several rinsing processes in distilled water, the graphene of 1 1 cm 2 was transferred on a 300 nm-thick SiO 2 /heavily p-doped Si wafer (as-grown graphene). The same batch of the chemical vapor deposition graphene was separately transferred on the electrodes-patterned pads and patterned into a microbridge shape of 4 70 μm 2 (microbridge graphene) by electron beam lithography. Subsequently coatedpolymethylmethacrylate layer was removed by acetone, and the microbridge graphene was annealed at 350 C with a mixture of Ar and H 2 condition to remove the polymer residues remained on the surfaces. Raman measurements [16,17] were carried out under 760 torr atmospheric pressure, and the humidity was about 45%. The Raman spectra were recorded with a Horiba Jobin-Yvon LabRam HR spectrometer and detected with a liquid-nitrogen-cooled charge-coupled device (CCD) detector. The 514.5 nm line of an Ar-ion laser was used as the excitation source, and a laser power on the sample were kept around 100 μw, to avoid heating of the sample during the Raman measurement with the incident laser beam. The grooves/mm grating was 600 gr/mm, and the cm 1 between each charge coupled device pixel was 1.6 cm 1. The Raman scattered light signal was collected in a backscattering geometry using a long-working distance microscope objective lens ( 50, 0.5 N.A.). To minimize the uncertainties of peak position or intensity, the wavenumber of the Raman lines was calibrated using the c-si wafer at 521 cm 1 throughout the measurements. As a result, we could get consistent G-peak signal, and we used integrated intensity of G-peak to normalize the Raman signal. The samples were put into the sealing cage having a window for laser exposure, and Ar gas was flowed during the heating. For the heating under air, all the gas ports were kept open. The temperature was increased with the step of 50 100 C, and each Raman measurement was carried out in situ for 1 min in 2 min after the target temperature was reached. The Raman spectra were taken from near the center of each graphene samples by focusing the Raman excitation beam onto an ~1 μm diameter spot. We tried to observe the difference of the structural changes and the chemical doping in graphene samples of as-grown and microbridge-patterned, which were annealed in two different types of atmosphere, namely, air and Ar. Results and discussion Figure 1(a) shows the changes of Raman spectra of the as-grown graphene measured in situ with increasing temperature in air. All the Raman spectra are normalized against the G-peak intensity. From 30 to 300 C, G-peak and 2D-peak are found to gradually shift to lower wavenumbers (red shift) as the temperature increased, which is consistent with the behavior of graphene during heating. [22 24] However, unusual broadening starts to appear in both G-peak and 2D-peak above 350 C (Table 1). In our previous study, it was suggested that the broadened G-peak had two components of both chemical and physical origins; an irreversible broadening due to the adsorbed oxygen on the graphene and the resulting defects, in addition to physical deformation such as wrinkle and crumpling, part of which was found to be thermodynamically reversible as observed in the reversible component of the Raman broadening. [17] After cooling down to the 30 C, the peak-widths restored back to the original values (Table 1), and it indicates that the structural stress was released by the formation of the vacancies, or voids collection of vacancies on graphene surface due to the chemisorption of the oxygen atoms as in our simulation of oxidation of graphene. [17] The increase of the D-peak after cooling back to the 30 C is consistent with the increased number of defects. In contrast, when the as-grown graphene was heated in Ar gas, as shown in Fig. 1(b), G-peak and 2D-peak were gradually redshifted with increase of temperature from 30 to 184 Figure 1. The changes of in-situ Raman spectra of the as-grown graphene with increasing temperature in different atmosphere. The Raman spectra changes in air (a) and in Ar (b). Figures (c) and (d) show the changes in peak wavenumber and shape of the G-peak of the as-grown graphene samples heated in air and in Ar, respectively; 30 C before heating (bottom spectra), 350 C during heating (middle spectra), and 30 C after annealing (top spectra). wileyonlinelibrary.com/journal/jrs Copyright 2017 John Wiley & Sons, Ltd. J. Raman Spectrosc. 2018, 49, 183 188

Effects of different annealing atmosphere Table 1. The values of the peak wavenumber and the FWHM in the parenthesis of G-peak on the first row, and those of 2D peak in the second row in each cell, from the Raman spectra measured in-situ at several annealing temperatures As-grown graphene Microbridge graphene Temperature ( C) Air Ar Air Ar 30 (before annealing) 1599.6 (11.7) / 2697.9 (33.5) 1599.1 (11.0) / 2699.4 (34.5) 1593.1 (11.1) / 2690.6 (32.2) 1593.2 (11.2) 2690.8 (32.4) 350 1584.6 (15.1) / 2674.5 (42.4) 1577.9 (16.1) / 2673.1 (43.2) 1589.6 (21.7) / 2683.3 (44.3) 1579.6 (28.5) / 2683.4 (48.0) 500 1582.9 (16.7) / 2667.2 (45.5) 1572.9 (20.0) / 2664.3 (43.9) 1584.6 (20.4) / 2668.7 (49.8) 1577.9 (70.3) / 2671.6 (57.1) 30 (after annealing) 1609.6 (10.0) / 2709.6 (32.5) 1597.9 (13.4) / 2696.2 (39.8) 1609.9 (15.3) / 2709.8 (34.9) 1594.0 (85.1) / 2693.9 (49.5) All the values are in units of cm 1. FWHM, full width at half maximum. 500 C, but the peak widths do not restore to the original values. (Table 1.) This manifests that the structural stress was not released by the heating in Ar because Ar atoms are inert and they are not reacted with carbon atoms of graphene surface. After the as-grown graphene cooled down from 500 to 30 C, the D-peak was also observed as shown in Fig. 1(b). The formation of vacancies should not be responsible for the D-peak because of inactivity of Ar. Rather, the D-peak in this case indicates the structural changes due to stress, consistent with the broad shoulder peak in the G-peak. The increase of surface roughness after annealing in Ar is confirmed in our atomic force microscopy (AFM) measurements (Fig. S1). Figure 1(c) and 1(d) compare the G-peaks with the as-grown graphene before and after heating, and after cooling from 500 to 30 C in air and Ar, respectively. For the as-grown graphene heated in air [Fig. 1(c)], the G-peak was blueshifted after cooling by 10 cm 1 (Table 1). It indicates graphene was p-doped by the chemisorption of oxygen on the graphene surface. [16] In contrast, graphene heated in Ar [Fig. 1(d)], the G-peak returned back to the initial wavenumber, i.e. there was no significant chemical changes in graphene. Values of the peak position and the width of the G-peak and 2D-peak of the Raman spectra measured in situ at several annealing temperatures are summarized in the Table 1. It is noteworthy that the Raman spectra of the graphene patterned into microbridge shapes before heating [bottom spectra of Fig. 2(a) and 2(b)] are not any different from those of the as-grown graphene before heating [bottom spectra of Fig. 1(a) and 1(b)]. Even the integrated area ratio of 2D-peak to G-peak (A 2D /A G ) of the microbridge graphene before heating is nearly the same to that of the as-grown graphene (Fig. 3). It means that the processing steps during the microbridge patterning did not leave defects, deformation, or stress on the graphene itself. However, in-situ Raman measurements on microbridge graphene show surprisingly different results from those observed for the as-grown graphene. As the microbridge graphene sample is heated in air from 30 to 500 C as shown in Fig. 2(a), the G-peak and 2D-peak shift to lower wavenumbers during the heating as usual (Table 1). But a broad shoulder peak just below the G-peak appear strong up to 350 C, and gradually disappear from 400 to 500 C. Vacancies or collection of vacancies, voids, are optically observed in the microbridge graphene annealed at high temperatures. By forming such voids, the structural stress may be released making the broad peaks disappeared at high temperatures above 400 C. [17] After the microbridge graphene was cooled down back to the room temperature, the broad shoulder peak entirely disappeared, and the Figure 2. The changes of in-situ Raman spectra of the microbridge graphene with increasing temperature in different atmosphere. The Raman spectra changes in air (a) and in Ar (b). Figures (c) and (d) show the changes in peak wavenumber and shape of the G-peak of the microbridge graphene samples heated in air and in Ar, respectively; 30 C before heating (bottom spectra), 350 C during heating (middle spectra), and 30 C after annealing (top spectra). 185 J. Raman Spectrosc. 2018, 49, 183 188 Copyright 2017 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jrs

J. Son et al. Figure 3. The ratio of the area of 2D-peak to that of G-peak (A 2D /A G ) in the Raman spectra of (a) the as-grown graphene and (b) the microbridge graphene samples during heating and cooling back to 30 C in two different atmospheres (air and Ar). The temperatures on the x-axis are in the annealing sequence (from the left to the right). integrated area ratio of 2D-peak to G-peak (A 2D /A G ) also decreased. Note that the ratio is closely related to the structural quality of graphene, such as defects, thickness, and grain sizes (discussed later). [25,26] The most dramatic structural changes were observed in the microbridge graphene heated in Ar. As shown in Fig. 2(b), the G-peak and 2D-peak were shifted gradually to lower wavenumbers as temperature increased from 30 to 500 C (Table 1). The broad shoulder peak around the G-peak started to grow strong at temperature over 200 C. This broad shoulder peak dominates and takes over the G-peak at higher annealing temperatures. The observation reveals that the structural stress in graphene is generated by heating and cannot be released without the vacancies or voids formed by the chemisorption of oxygen. Therefore, the broad peak did not disappear even after cooling back to the room temperature, which is clearly different result from that of the microbridge graphene heated in air. In Fig. 2(c) and 2(d), while the microbridge graphene heated in Ar shows no significant change in chemical states after cooling, the p-doping is still observed after cooling in air (Table 1). This result indicates that chemical changes in graphene by the adsorption of oxygen is not relevant to the size or the shape of graphene. Figure 3(a) and 3(b) show the ratio of the area of 2D-peak to that of G-peak (A 2D /A G ) in the Raman spectra of the as-grown graphene and microbridge graphene samples, respectively, during heating and cooling back to 30 C. The area ratio was calculated with the values from fitting the Voigt function. The A 2D /A G value is believed to be positively correlated with the crystalline quality of graphene. [25,26] For as-grown graphene heated under air [Fig. 3(a)], the A 2D /A G ratio decreased with increase of temperature, and it kept the tendency even after cooling, which can be interpreted as the 186 Figure 4. The electric field effect of the microbridge graphene heated in air and Ar. The source-drain currents (I sd ) measured with the sweep of the back gating bias (V bg ) at each temperature as indicated for microbridge graphene samples heated in air (a) and in Ar (b), respectively. The source-drain voltage (V sd ) of 100 mv was applied. The insets show the mobility at different temperatures. wileyonlinelibrary.com/journal/jrs Copyright 2017 John Wiley & Sons, Ltd. J. Raman Spectrosc. 2018, 49, 183 188

Effects of different annealing atmosphere decrease in the grain size due to the formation of vacancy defects caused by the oxygen adsorption. In contrast, in the case of heating in the Ar atmosphere, the A 2D /A G ratio increased with temperature, as shown in Fig. 3(a). The ratio did not decrease at all and rather it sharply increased after cooling from 500 to 30 C. This indicates that as-grown large area graphene is nearly free from the structural changes resulting from the difference in the thermal expansion between the SiO 2 substrate and the graphene sample. In addition, because Ar atoms are inert chemically, the heating of as-grown graphene in Ar showed the thermal annealing effect [27] and the enhancement of the crystal quality of graphene. On the other hand, the structural changes in the microbridge graphene were strongly affected by the heating as we discussed (Fig. 2). The A 2D /A G ratio decreased in the microbridge graphene irrespective of annealing atmosphere, as shown in Fig. 3(b). It is worth to note that the A 2D /A G ratio of the microbridge graphene heated in Ar also decreased above 300 C, although it started to increase at low temperatures like the as-grown graphene heated in Ar, which may be caused by the stress built up. Moreover, the structural changes in the microbridge graphene appear much stronger than that in the as-grown graphene [Figs 1(b) and 2(b)], which resulted in the decrease of the A 2D /A G ratio implying poor crystalline quality. [25] Finally, we investigated the relation between structural change and carrier electrical mobility in the microbridge graphene by the electrical field effect measurements. The source-drain currents (I sd ) were measured with the sweep of back gating bias (V bg ) at a source-drain voltage (V sd ) of 100 mv at each temperature. The mobility was calculated from the equation μ =[L/(W C i V sd )] [d(i sd )/d(v bg )], [28] where L (W) is length (width) of graphene channel. C i is the gate capacitance, which was 1.08 10 8 F/cm 2 in our SiO 2 dielectric materials (thickness was 300 nm and ε r is 3.9). The transconductance [d(i sd )/d(v bg )] was calculated from the transfer curves at the left side of the Dirac point. All field effect curves shown in Fig. 4(a) were intentionally upshifted one by one to clearly show the shift of the Dirac point as a function of temperature. The microbridge graphene heated in air showed that the Dirac point moved to the positive values of V bg, and the magnitude of the shift became large with temperature, which means that the microbridge graphene heated in air is chemically p-type doped states and the doping becomes significant with temperature, which is consistent with our previous results. In Fig. 4(a) (inset), the mobility showed a small decrease up to 350 C, and then an increase again up to 500 C. We can explain such a tendency of change in mobility by the structural changes in the former temperature range and by the stress release in the latter temperature range. In the case of the microbridge graphene heated in Ar [Fig. 4(b)], the Dirac point was not changed as the temperature increased, i.e. no chemical reactions occurred on the graphene surface. However, the mobility gradually decreased up to 500 C, as shown in Fig. 4(b) (inset), indicating that the structural changes irrelevant to the stress release reduced the mobility, which is consistent with the result shown in Fig. 2(b). Conclusion We investigated the structural and chemical changes in the as-grown and the microbridge graphene samples heated in air or Ar atmosphere by in-situ Raman spectroscopy. The as-grown graphene heated in air and Ar were not significantly different from each other. In contrast, the microbridge graphene sample heated in Ar showed strong structural changes during annealing at elevated temperatures possibly due to the stress built up between the graphene and the SiO 2 substrate. The different behavior of the microbridge graphene heated in Ar from that of the as-grown graphene in the same condition is clearly apparent, so that graphene in microscopic devices should be treated differently from the as-grown unpatterned graphene. On the other hand, both the as-grown and microbridge graphene heated in air showed a nonstructural change at high temperature due to the formation of vacancies by oxygen chemisorption, which resulted in the stress release as well as p-type doping. We believe that our Raman results on the structural changes of graphene, either of unpatterned large area or microbridge shape, heated to high temperature in different atmosphere could suggest crucial information for future graphene devices that would eventually be patterned on a nanometer scale and thus subject to high temperature by Joule heating. Acknowledgements I. S. Yang acknowledges the support by the National Research Foundation of Korea (NRF) grant funded by the Korean government (grant number 2017R1A2B2009309). Research at Yonsei University was supported by Creative Materials Discovery Program (2015M3D1A1070465) through the National Research Foundation of Korea (NRF) funded by the Korean government. References [1] A. K. Geim, K. S. Novoselov, Nat. Mater. 2007, 6, 183. [2] A.H.C.Neto,F.Guinea,N.M.R.Peres,K.S.Novoselov,A.K.Geim,Rev. Mod. Phys. 2009, 81, 109. [3] K.S.Novoselov,V.I.Fal ko, L. Colombo, P. R. Gellert, M. G. Schwab, K. Kim, Nature 2012, 490, 192. [4] K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva, A. A. Firsov, Science 2004, 306, 666. [5] K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, M. I. Katsnelson, I. V. Grigorieva, S. V. Dubonos, A. A. Firsov, Nature 2005, 438, 197. [6] Y.Zhang,Y.Tan,H.L.Stormer,P.Kim,Nature 2005, 438, 201. [7] A. K. Geim, Science 2009, 324, 1530. [8] F. Schwierz, Nat. Nanotechnol. 2010, 5, 487. [9] F. Bonaccorso, Z. Sun, T. Hasan, A. C. Ferrari, Nat. Photon. 2010, 4, 611. [10] H. Liu, Y. Liu, D. Zhu, J. Mater. Chem. 2011, 21, 3335. [11] Y. Shao, J. Wang, H. Wu, J. Liu, I. A. Aksay, Y. Lina, Electroanalysis 2010, 22, 1027. [12] M. Pumera, Mater. Today 2011, 14, 308. [13] T. Kuila, S. Bose, P. Khanra, A. K. Mishra, N. H. Kim, J. H. Lee, Biosens. Bioelectron. 2011, 26, 4637. [14] Y. Song, Y. Luo, C. Zhu, H. Li, D. Du, Y. Lin, Biosens. Bioelectron. 2016, 76, 195. [15] X. Li, B. D. Kong, J. M. Zavada, K. W. Kim, Appl. Phys. Lett. 2011, 99, 233114. [16] M.Choi,J.Son,H.Choi,H.J.Shin,S.Lee,S.Kim,S.Lee,S.Kim,K.R.Lee, S. J. Kim, B. H. Hong, J. Hong, I. S. Yang, J. Raman. Spectrosc. 2014, 45, 168. [17] J. Son, M. Choi, H. Choi, S. J. Kim, S. Kim, K. R. Lee, S. Vantasin, I. Tanabe, J. Cha, Y. Ozaki, B. H. Hong, I. S. Yang, J. Hong, Carbon 2016, 99, 466. [18] C. Durkan, Z. Xiao, Appl. Phys. Lett. 2015, 107, 243505. [19] D. Yoon, Y. W. Son, H. Cheong, Nano Lett. 2011, 11, 3227. [20] K. Kim, Y. Zhao, H. Jang, S. Lee, J. Kim, K. Kim, J. Ahn, P. Kim, J. Choi, B. Hong, Nature 2009, 457, 706. [21] S. Bae, H. Kim, Y. Lee, X. Xu, J. Park, Y. Zheng, J. Balakrishnan, T. Lei, H. Kim, Y. Song, Y. Kim, K. Kim, B. Özyilmaz, J. Ahn, B. Hong, S. Lijima, Nat. Nanotechnol. 2010, 5, 574. [22] I. Calizo, A. A. Balandin, W. Bao, F. Miao, C. N. Lau, Nano Lett. 2007, 7, 2645. [23] A. A. Balandin, Nat. Mater. 2011, 10, 569. [24] J. E. Weis, S. Costa, O. Frank, M. Kalbac, J. Phys. Chem. Lett. 2014, 5, 549. [25] T. Wu, H. Shen, L. Sun, J. You, Z. Yue, RSC Adv. 2013, 3, 9544. [26] S. Hussain, M. W. Iqbal, J. Park, M. Ahmad, J. Singh, J. Eom, J. Jung, Nanoscale Res. Lett. 2014, 9, 546. 187 J. Raman Spectrosc. 2018, 49, 183 188 Copyright 2017 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jrs

J. Son et al. [27] W. Choi, Y. S. Seo, J. Y. Park, K. B. Kim, J. Jung, N. Lee, Y. Seo, S. Hong, IEEE Trans. Nanotechnol. 2015, 14, 70. [28] A. D. Bartolomeo, F. Giubileo, F. Romeo, P. Sabatino, G. Carapella, L. Iemmo, T. Schroeder, G. Lupina, Nanotechmology 2015, 26, 475202. Supporting information Additional Supporting Information may be found online in the supporting information tab for this article. 188 wileyonlinelibrary.com/journal/jrs Copyright 2017 John Wiley & Sons, Ltd. J. Raman Spectrosc. 2018, 49, 183 188