Growth, photoluminescence and thermal conductance of graphene-like nanoflakes grown on copper foils in methane environment

Materials Science in Semiconductor Processing 27 (2014) 97–102

Contents lists available at ScienceDirect

Materials Science in Semiconductor Processing journal homepage: www.elsevier.com/locate/mssp

Growth, photoluminescence and thermal conductance of graphene-like nanoflakes grown on copper foils in methane environment C. Dang a,d, Q.L. Che b, B.L. Gao a, L. Li c, B.B. Wang d,n a

College of Mathematics and Physics, Huaiyin Institute of Technology, 1 Meicheng Road, Huaian, Jiangsu, Huaian 223002, PR China National Key Laboratory of Science and Technology on Surface Engineering, Lanzhou Institute of Physics, PO Box 94, Lanzhou 730000, Gansu, PR China c College of Chemistry, Chongqing Normal University, Chongqing 401331, PR China d College of Chemistry and Chemical Engineering, Chongqing University of Technology, 69 Hongguang Rd, Lijiatuo, Banan District, Chongqing 400054, PR China b

a r t i c l e i n f o

Keywords: Graphene-like nanoflakes Chemical vapor deposition Photoluminescence Thermal conductance

abstract In this paper, we reported a simple and effective synthesis method of graphene-like nanoflakes (GNFs) on the copper foils by hot filament chemical vapor deposition in methane environment. The structure and composition of GNFs were studied by field emission scanning electron microscope, micro-Raman spectroscope, and Fourier transform infrared spectroscope, respectively. According to the characterization results and the growth process, the formation mechanism of GNFs was investigated, which was based on the formation of carbon particles and the diffusion and assembly of carbon atoms. The photoluminescence (PL) of GNFs was measured in a Ramalog system and the PL spectra show a weak and a strong PL bands centered at about 411 and 515 nm, respectively. The measurement results of thermal conductance of GNFs indicate that the thermal conductivity of GNFs is up to 480 W/mK. Our results can enrich the knowledge on the synthesis, optical and thermal properties of graphene-based nanomaterials and contribute to the development of graphene-based devices. & 2014 Elsevier Ltd. All rights reserved.

1. Introduction Recently, low-dimensional carbon-based nanomaterials such as graphene, carbon nanotubes, nanodots and silicon carbide nanofilms have received much attention due to their unique structures and properties [1–4]. For the graphene and graphene nanoribbons, they have fascinating physical properties such as quantum electronic transport, a tunable band gap, giant intrinsic mobility, high elasticity and electromechanical modulation [5–7]. The graphene

n

Corresponding author. Tel./fax: þ86 23 62563221. E-mail address: [email protected] (B.B. Wang).

http://dx.doi.org/10.1016/j.mssp.2014.06.031 1369-8001/& 2014 Elsevier Ltd. All rights reserved.

nanoflakes (GNFs) are composed of multilayer graphene nanosheets, which have similar structure of graphene and graphene nanoribbons. In particular, the GNFs can vertically grow and form the thin graphitic edge planes. The results of Shang et al. indicate that the thin graphitic edge planes are basically responsible for the electrocatalytic and biosensing properties, which imply that the GNFs have extensively potential applications in the areas of electrochemistry and biology such as the fabrication of electrodes of electrocatalytic and biosensing devices [8]. Moreover, the GNF films can form graphene network [7], thus they can be used to fabricate the pressure sensors because the strain response of graphene network mainly depends on the overlap area and contact resistance of neighboring graphene sheets [9].

98

C. Dang et al. / Materials Science in Semiconductor Processing 27 (2014) 97–102

This is the possible reason that the GNFs have become the spotlight of extensive research effort. Since the reactive plasma was extensively used to fabricate various nanomaterials [10,11], the GNFs were synthesized by plasmaenhanced chemical vapor deposition and magneticallyenhanced arc discharge [7,12]. The properties of GNFs are related to the layer number of graphene and their stackingassembly order [7], which activates us to synthesize the GNFs using different methods and study their properties. When the graphene is synthesized by chemical vapor deposition (CVD), a mixture of methane and hydrogen is usually employed [1,5,7]. However, the atomic hydrogen can etch the graphite carbon [13,14], thus the GNFs were grown using methane as the reactive gas in this work to reduce the etching effect of hydrogen on the GNFs. Since copper can catalyze the growth of carbon-based materials such as graphite, carbon nanotubes and graphene [15], we used the copper foils as the substrate to grow the GNFs in hot filament chemical vapor deposition (HFCVD) system. To date, the CVD growth of graphene on copper foils has been extensively studied [1,14,16,17]. For the formation of graphene on copper surface, it is accepted that the formation of graphene on copper surface results from the surface reactions and that the hydrogen-assisted dehydrogenation of hydrocarbon is responsible for the formation of graphene [14,16,17]. However, can graphene be formed on copper surface in methane environment? It is found that the amorphous carbon and GNFs are formed depending on the growth time in the methane environment in this work. The perfect graphene lacks a bandgap so that it is difficult to generate photoluminescence (PL) [18], while the graphene oxide with the bandgap generates blue light [19]. Simultaneously, the results of Han et al. indicate that the graphene nanoribbons possess the electronic bandgap [6]. Because of the structural similarity between the GNFs and graphene nanoribbons, the GNFs should have a bandgap and generate the PL, thus the PL properties of GNFs were studied. In addition, the suspended graphene has a very high thermal conductivity of 5300 W/mK near 350 K [20] so that we were interested in the thermal properties of GNFs and also studied their thermal conductivity. 2. Experimental To grow the GNFs, the surfaces of copper foils were polished with SiC paper to remove the impurities, and then the foils were cleaned in the ultrasonic solution of ethanol for 15 min. The GNFs were synthesized on copper foils in HFCVD system described in Ref. [21]. Briefly, the CVD chamber in the HFCVD system contains a heating system constructed by three tungsten filaments which were heated to about 1800 1C. The distance between the filaments and copper foils was about 8 mm. Since the copper foils were exposed to the filaments, the copper foils were fast heated to above 850 1C by the hot filaments. After the chamber was evacuated to lower than 2 Pa, hydrogen with a flow rate of 50 sccm was inlet into the CVD chamber, and then the pressure was adjusted to about 2  103 Pa. Simultaneously, the filaments were heated by AC current in hydrogen

environment. Once the temperature of a copper foil reached about 850 1C, methane was substituted for hydrogen in the same flow rate as hydrogen to synthesize the GNFs. In this work, three specimens were prepared for different growth time. It is found that a lot of GNFs are formed in 10 and 15 min while the amorphous carbon is formed when the growth time is 5 min. The morphologies of GNFs were investigated using a Hitachi S-4800 field emission scanning electron microscope (FESEM), which was operated at 15 kV. The composition of GNFs was determined by a T64000 micro-Raman spectroscopy using a 514 nm line of Ar þ laser and a 8400S Shimadzu Fourier transform infrared (FTIR) spectroscopy, respectively. The PL spectra of GNFs were recorded in a SPEX 1403 Ramalog system using a 325 nm He–Cd laser as an excitation source. The thermal conductivity of GNFs was measured in a LTI-237CM system using the laser-flash method. During the measurement, an Nd: YAG laser was employed, and the excitation wavelength, pulse energy and duration are 10.6 μm, 75 mJ and 10 ns, respectively.

3. Results and discussion 3.1. Structure and composition of GNFs on copper foils Fig. 1 is the FESEM image and Raman spectrum of specimen grown for 5 min (the Raman spectrum is obtained from HR 800 micro-Raman spectroscopy using the 325 nm line of semiconductor laser). From Fig. 1(a), one can see that the carbon particles are formed and few GNFs are formed. The Raman spectrum in Fig. 1(b) shows that the D, G and 2D peaks of carbon materials are located at about 1349, 1589 and 2705 cm  1, respectively [7]. The very weak 2D peak relative to the G peak indicates that the specimen is composed of amorphous carbon particles. Fig. 2 is the FESEM images of specimens with the growth time of 10 and 15 min (the specimens are named as A and B), respectively. As shown in Fig. 2, the vertical GNFs are grown on the copper foils. From Fig. 2, one can see that the area of GNFs is increased with the increase of the growth time. Fig. 3 is the Raman spectra of specimens A and B. In Fig. 3, both the Raman spectra show that the D, G and 2D peaks are located at about 1349, 1583 and 2699 cm  1, respectively. The strong 2D peaks indicate that the GNFs are composed of multi-layer graphene sheets [1,22,23] The weak peak at about 2938 cm  1 is attributed to the Dþ G peak [7]. As shown in Fig. 3, the D peaks for two specimens are strong and the strong D peak of GNFs is observed in Ref. [7]. For the perfect graphene films, the D peak is very weak [1,24] and it mainly results from the edge defects of graphene films [23]. Our specimens are composed of the vertical GNFs so that the edge defects can be detected by Raman spectroscopy, thus the strong D peaks appear in the Raman spectra. Simultaneously, the strong D peak in Ref. [7] is due to the edge defects of GNFs. According to Fig. 3, the intensity ratio of 2D to G peaks is about 0.9 and 0.5 for the spectra (1) and (2), which means that the GNFs of specimen A are thinner than that of specimen B because

C. Dang et al. / Materials Science in Semiconductor Processing 27 (2014) 97–102

99

Fig. 1. (a) FESEM image and (b) Raman spectrum of the specimen grown for 5 min.

Fig. 3. Raman spectra of specimens (1) A and (2) B.

Fig. 4 are related to OH groups [28]. From Fig. 4, one can see that the peaks of spectrum (b) are stronger than those of spectrum (a), which possibly results from the strong absorption of specimen B due to the thick GNFs. 3.2. Formation mechanism of GNFs on copper foils

Fig. 2. FESEM images of specimens (a) A and (b) B.

the ratio is reduced with the increase of graphene layers [1]. To confirm the chemical groups in the specimens, these specimens are studied by FTIR and the results are shown in Fig. 4. As shown in Fig. 4, every spectrum shows the peaks at about 624, 1383 and 1621 cm  1, which are attributed to the C–H, C–O–H and C ¼C groups, respectively [25–27]. In addition, the peaks at about 3405–3550 cm  1 shown in

In the previous works [1,14,16,17], the CVD growth of graphene on copper surface was studied, which indicates that the formation of graphene results from the hydrogenenhanced reactions of hydrocarbon on copper surface. Our results shown in (Figs. 1 and 2) indicate that the formation of graphene in methane environment is different from the results reported in Refs. [1,14,16,17] and that the formation of graphene in methane environment starts from the carbon particles. In this section, the formation of graphene on copper surface in methane environment is studied based on the preparation process of GNFs. Before the growth of GNFs, the copper foil was heated in hydrogen environment. Due to the high temperature (  1800 1C) of hot filaments, hydrogen molecules are

100

C. Dang et al. / Materials Science in Semiconductor Processing 27 (2014) 97–102

Fig. 4. FTIR spectra of specimens (a) A and (b) B. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

decomposed into the atomic hydrogen, H2 -2H:

ð1Þ

The atomic hydrogen can further clean the surface of copper to remove the impurities such as oxygen and carbon [29]. After hydrogen is replaced by methane, methane is decomposed into various hydrocarbon radicals and hydrogen by the hot filaments [30]. For example, the methyl radicals can be formed through the reaction of thermal decomposition, CH4 -CH3 þ H:

ð2Þ

The methyl radicals are easily adsorbed on the copper surface due to the clean surface, and then they diffuse on the surface. Because the copper surface was polished by the SiC paper, a number of micro-pits were produced on the copper surface. As a result, the methyl radicals are aggregated in the micro-pits owing to the blocking effect of micro-pits. Under the assistance of atomic hydrogen formed by the reaction (2), some carbon atoms are formed from the aggregated methyl radicals through the elimination reaction [21]. However, the elimination process is slow because of a small number of hydrogen atoms, thus these carbon atoms can form sp2 carbon clusters since the

graphite phase is stable in our growth time and pressure [31]. Consequently, the aggregated methyl radicals in the micro-pits are transformed into the carbon particles composed of sp2 carbon and hydrocarbon radicals, i.e., the carbon particles are amorphous. Due to the formation of carbon particles, the methyl radicals deposited on the surfaces of carbon particles are catalytically decomposed to form carbon atoms [32]. As a result, the carbon atoms diffuse towards the copper surface under the high potential of carbon on the carbon particles [31]. When they meet the methyl radicals on the cooper surface, it is difficult to move on the copper surface due to the increase of mass. In this case, there are more carbon particles to form on the copper surface, which are confirmed by Fig. 1. With the increase of growth time, the hydrocarbon radicals deposited on the copper surface are fast decomposed by the carbon particles to form carbon atoms [32]. The carbon atoms depend on the carbon particles to form GNFs via the diffusion and assembly of carbon atoms, which are studied in Ref. [31]. Due to the high potential of carbon on the carbon particles, the carbon atoms diffuse towards to the copper surface. From Ref. [33], it is found that the surface energy of graphite (54.8 mJ/m2) is larger than that of graphene (46.7 mJ/m2), thus the atoms arrange in a mode of layer by layer to form the graphene nanosheets [34]. Due to the difference in thermal expansion coefficients of the graphene nanosheets and copper, a stress is produced in the graphene nanosheets. As a result, the graphene nanosheets are fractured to form the GNFs. In Ref. [35], the results indicate that the negative expansion coefficient of graphene remains up to 2300 K, which makes the GNFs shrink at the growth temperature and the ends of fractured GNFs bend upward. As a result, the vertical GNFs are formed. Because there are a number of dangling bonds on the ends of GNFs, the hydrocarbon radicals are bonded on the ends of GNFs to enlarge the area of GNFs, This is the reason that Fig. 2 shows the increase of GNF area with the increase of growth time. 3.3. Photoluminescence of GNFs on copper foils Fig. 5 is the PL spectra of specimens A and B at room temperature. As shown in Fig. 5, the PL spectrum (1) shows a weak blue PL band at about 411 nm and a strong green PL band at about 515 nm while the spectrum (2) exhibits the two PL bands at about 414 and 522 nm, respectively. By comparing the spectrum (1) with spectrum (2), one can see that every PL band in the spectrum (2) has a red shift relative to the corresponding PL band in the spectrum (1). Fig. 4 indicates that the GNFs contain the C–H groups, thus the blue PL band originates from the sp2 C–H groups because the blue PL can be emitted from aromatic or olefinic molecules [36]. For the green PL bands at 515 and 522 nm, the corresponding energy is about 2.38 and 2.43 eV, respectively. The results of Eda et al. indicate that the bandgap of sp2 carbon clusters with a diameter of  3 nm is about 0.5 eV [19]. From Fig. 2, the thickness of GNFs is about 5 nm, thus the bandgap of GNFs is lower than 0.5 eV because of the shrinking of bandgap with the size increase of carbon clusters [19]. It is difficult to

C. Dang et al. / Materials Science in Semiconductor Processing 27 (2014) 97–102

101

are reduced for the specimen B relative to the specimen A. As a result, the scattering of phonons in the specimen B is weakened, thus the thermal conductivity of specimen B is larger than that of specimen A. 4. Conclusion

Fig. 5. PL spectra of specimens (1) A and (2) B. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

interpret the green PL emission using the bandgap emission mechanism. According to the interband transition mechanism of carbon materials [37], the transition with energy more than 2 eV occurs between πn and π bands, thus the two green PL bands at 515 and 522 nm are related to the interband transition between πn and π bands. The red shift of PL bands in the spectrum (2) is associated with the thickness of GNFs. In Section 3.1, the low intensity ratio of 2D to G peak of spectrum (2) in Fig. 3 and the high absorbance intensity in Fig. 4(b) indicate that the GNFs in specimen B are thicker than that in specimen A, which means that the size of sp2 clusters in GNFs of specimen B increases [19]. According to the calculation results of Ref. [19], the energy gap of π–πn transition is reduced with the increase of sp2 clusters, i.e., the energy difference between π and πn bands is reduced with the increase of sp2 cluster. As a result, the energy of emitting photons is lowered with the size increase of sp2 clusters, thus the PL bands in the spectrum (2) have a red shift. 3.4. Thermal conductance of GNFs on copper foils The thermal conductivity obtained in LTI-237CM system is 88 and 480 W/mK for the specimens A and B, respectively. The values of thermal conductivity of our specimens approach the thermal conductivity of SiO2 supported graphene nanoribbons (50–500 W/mK) [38], but they are much lower than the value of suspended single-layer graphene (5300 W/mK) [20]. It is possible that the structural similarity between the GNFs and graphene nanoribbons results in the small difference in the thermal conductivity of them. Because the GNFs are supported by copper foil, a number of phonons are leaked through the GNF-support interface [39], which is the main reason that the thermal conductivity of our specimens is much lower than that of the suspended single-layer graphene. Moreover, some phonons are scattered by the edge defects of GNFs so that the thermal conductivity is further lowered [40]. Fig. 2 shows that the GNFs of specimen B have larger area than that of specimen A, i.e., the edge defects of GNFs

In summary, the GNFs were synthesized by HFCVD in methane environment. The structural and compositional properties of GNFs were studied using FESEM, microRaman spectroscopy, and FTIR, respectively. According to the characterization results, the formation mechanism of GNFs was investigated, which was related to the formation of carbon particles and the diffusion and assembly of carbon atoms. The PL of GNFs was studied in SPEX 1403 Ramalog system and the PL results indicate that the GNFs can emit weak blue light and strong green light, which are related to the sp2 C–H groups and interband transition between πn and π bands, respectively. The thermal conductance of GNFs was measured and the results indicate that the thermal conductivity of GNFs is 88–480 W/mK, which is much lower than the thermal conductivity of suspended graphene due to the leakage of phonons and the scattering of phonons by the edge defects of GNFs. Our results can enrich the knowledge on the synthesis, optical and thermal properties of graphene-based nanomaterials and contribute to the development of graphene-based devices.

Acknowledgments This work is supported by National Natural Science Foundation of China (NSFC) (Grant no. 60676056). References [1] X. Li, W. Cai, J. An, S. Kim, J. Nah, D. Yang, R. Piner, A. Velamakanni, I. Jung, E. Tutuc, S.K. Banerjee, L. Colombo, R.S. Ruoff, Science 324 (2009) 1312–1314. [2] S. Kumar, H. Mehdipour, K. Ostrikov, Adv. Mater. 25 (2013) 69–74. [3] Z. Yang, Z. Li, M. Xu, Y. Ma, J. Zhang, Y. Su, F. Gao, H. Wei, L. Zhang, Nano-Micro Lett. 5 (2013) 247–259. [4] Q. Cheng, S. Xu, J. Long, K. Ostrikov, Chem. Vap. Deposition 13 (2007) 561–566. [5] K.S. Kim, Y. Zhao, H. Jang, S.Y. Lee, J.M. Kim, K.S. Kim, J.-H. Ahn, P. Kim, J.-Y. Choi, B.H. Hong, Nature 457 (2009) 706–710. [6] M.Y. Han, B. Özyilmaz, Y. Zhang, P. Kim, Phys. Rev. Lett. 98 (2007). (206805-4). [7] D.H. Seo, S. Kumar, K. Ostrikov, Carbon 49 (2011) 4331–4339. [8] N.G. Shang, P. Papakonstantinou, M. McMullan, M. Chu, A. Stamboulis, A. Potenza, S.S. Dhesi, H. Marchetto, Adv. Funct. Mater. 18 (2008) 3506–3514. [9] J. Zhao, G.-Y. Zhang, D.-X. Shi, Chin. Phys. B 22 (2013) 057701–057709. [10] Q. Cheng, S. Xu, K. Ostrikov, J. Mater. Chem. 20 (2010) 5853–5859. [11] K. Ostrikov, E.C. Neyts, M. Meyyappan, Adv. Phys. 62 (2013) 113–224. [12] I. Levchenko, O. Volotskova, A. Shashurin, Y. Raitses, K. Ostrikov, M. Keidar, Carbon 48 (2010) 4570–4574. [13] K.-J. Lee, J.-G. Koh, J.-S. Shin, K.-H. Kwon, C.-H. Lee, J. Korean Phys. Soc. 48 (2006) 141–145. [14] I. Vlassiouk, S. Smirnov, M. Regmi, S.P. Surwade, N. Srivastava, R. Feenstra, G. Eres, C. Parish, N. Lavrik, P. Datskos, S. Dai, P. Fulvio, J. Phys. Chem. C 117 (2013) 18919–18926. [15] C. Mattevi, H. Kim, M. Chhowalla, J. Mater. Chem. 21 (2011) 3324–3334. [16] H. Mehdipour, K. Ostrikov, ACS Nano 6 (2012) 10276–10286.

102

C. Dang et al. / Materials Science in Semiconductor Processing 27 (2014) 97–102

[17] W. Zhang, P. Wu, Z. Li, J. Yang, J. Phys. Chem. C 115 (2011) 17782–17787. [18] L. Cao, M.J. Meziani, S. Sahu, Y.-P. Sun, Acc. Chem. Res. 46 (2013) 171–180. [19] G. Eda, Y.-Y. Lin, C. Mattevi, H. Yamaguchi, H.-A. Chen, I.-S. Chen, C.-W. Chen, M. Chhowalla, Adv. Mater. 22 (2010) 505–509. [20] A.A. Balandin, S. Ghosh, W. Bao, I. Calizo, D. Teweldebrhan, F. Miao, C.N. Lau, Nano Lett. 8 (2008) 902–907. [21] B.B. Wang, K. Ostrikov, T. van der Laan, K. Zheng, J.J. Wang, Y.P. Yan, X.J. Quan, J. Mater. Chem. C 1 (2013) 7703–7708. [22] A.C. Ferrari, J.C. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri, S. Piscanec, D. Jiang, K.S. Novoselov, S. Roth, A.K. Geim, Phys. Rev. Lett. 97 (2006) 187401–187404. [23] C. Casiraghi, S. Pisana, K.S. Novoselov, A.K. Geim, A.C. Ferrari, Appl. Phys. Lett. 91 (2007). (233108-3). [24] A. Gupta, G. Chen, P. Joshi, S. Tadigadapa, P.C. Eklund, Nano Lett. 6 (2006) 2667–2673. [25] B. Schwenzer, S.N. Kerisit, M. Vijayakumar, RSC Adv. 4 (2014) 5457–5464. [26] E. Fuente, J.A. Menéndez, M.A. Díez, D. Suárez, M.A. Montes-Moran, J. Phys. Chem. B 107 (2003) 6350–6359. [27] C.M. Luk, L.B. Tang, W.F. Zhang, S.F. Yu, K.S. Teng, S.P. Lau, J. Mater. Chem. 22 (2012) 22378–22381.

[28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39]

[40]

P.-C. Hsu, H.-T. Chang, Chem. Commun. 48 (2012) 3984–3986. J.T. Wolan, G.B. Hoflund, J. Vac. Sci. Technol. A 16 (1998) 2546–2552. C.-J. Chen, M.H. Back, R.A. Back, Can. J. Chem. 54 (1976) 3175–3184. B. Wang, K. Zheng, Q. Cheng, L. Wang, C. Chen, G. Dong, Chem. Vap. Deposition (in press, http://dx.doi.org/10.1002/cvde.201407105). S.C. Lee, H.J. Seo, G.Y. Han, Korean J. Chem. Eng. 30 (2013) 1716–1721. S. Wang, Y. Zhang, N. Abidi, L. Cabrales, Langmuir 25 (2009) 11078–11081. M. Kusunoki, T. Suzuki, T. Hirayama, N. Shibata, K. Kaneko, Appl. Phys. Lett. 77 (2000) 531–533. N. Mounet, N. Marzari, Phys. Rev. B: Condens. Matter 71 (2005). (205214-14). F. Bonaccorso, Z. Sun, T. Hasan, A.C. Ferrari, Nat. Photonics 4 (2010) 611–622. Y. Iwano, T. Kittaka, H. Tabuchi, M. Soukawa, S. Kunitsugu, K. Takarabe, K. Itoh, Jpn. J. Appl. Phys. 47 (2008) 7842–7844. A.D. Liao, J.Z. Wu, X. Wang, K. Tahy, D. Jena, H. Dai, E. Pop, Phys. Rev. Lett. 106 (2011) 256801–256804. J.H. Seol, I. Jo, A.L. Moore, L. Lindsay, Z.H. Aitken, M.T. Pettes, X.S. Li, Z. Yao, R. Huang, D. Broido, N. Mingo, R.S. Ruoff, L. Shi, Science 328 (2010) 213–216. W. Jang, Z. Chen, W. Bao, C.N. Lau, C. Dames, Nano Lett. 10 (2010) 3909–3913.