J Mater Sci: Mater Electron (2013) 24:4893–4900 DOI 10.1007/s10854-013-1494-6
Randomly oriented graphene flakes film fabrication from graphite dispersed in N-methyl-pyrrolidone by using electrohydrodynamic atomization technique Kyung Hyun Choi • Adnan Ali • Jeongdai Jo
Received: 27 July 2013 / Accepted: 7 September 2013 / Published online: 22 September 2013 Ó Springer Science+Business Media New York 2013
Abstract In this work, we report the deposition of graphene flakes exfoliated through graphite dispersion in N-methylpyrrolidone using non-vacuum electrohydrodynamic atomization (EHDA) technique. Stable cone jet mode of EHDA is used to deposit graphene flakes on silicon substrate. The deposited graphene flakes film is characterized by Raman spectroscopy, microscopy, 3D-Nanomap, scanning electron microscope, and UV–visible spectroscopy. Through characterizations it is evident that a randomly oriented graphene flakes film has shown good transparency, conductivity and suitable work function. For electrical characterization of film, it is employed as cathode in a simple diode indium tin oxide/ (poly(3,4-ethylenedioxythiophene) poly(styrenesulfonate)/ polydioctylfluorene-benzothiadiazole/graphene. It is observed that at voltage of 0.3 V, the current density in device is at low value of 2.67 A/cm2 however as the voltage is increased to a value of 4 V the current density is increased by almost 100 times and reaches up to 2.65 9 102 A/cm2. We believe that by further optimizing parameters of EHDA techniques for graphene deposition, more uniform and defect free graphene film can be obtained.
1 Introduction The novelty of graphene electronic properties have been well accredited now . In graphene the charge carriers K. H. Choi (&) A. Ali Department of Mechatronics Engineering, Jeju National University, Cheju 690-756, Korea e-mail: [email protected]
J. Jo Korean Institute of Machinery and Materials, Yuseong-Gu, Taejon 305-343, Korea
behave as mass-less Dirac fermions , and novel effects such as an ambipolar field effect , a room-temperature quantum Hall effect  and the breakdown of the Born– Oppenheimer approximation  have all been observed. A graphene monolayer has also been demonstrated as a transparent electrode in a liquid crystal device . Graphene suffers a problem of mass production which is common to many novel materials. Earlier the standard procedure used to produce graphene was micromechanical cleavage . This gives the superlative samples to date, with carrier mobilities up to 200,000 cm2/V/s [8–10]. However, the single layers so produced, form a negligible fraction amongst large quantities of thin graphite flakes. Alternatively, growth of graphene is also commonly achieved by annealing SiC substrates; however, these samples are in fact composed of a huge amount of domains, most of them sub micro-meter in scale, and they are not spatially uniform in number or size over larger length scales [11–13]. Later on CVD processes were used for large area fabrication on Ni and Cu substrate using controlled environment [14–17], but this would require transfer of the sample to insulating substrates in order to make useful devices, either by mechanical transfer or through solution processing. In recent times, a large number of papers have described the dispersion and exfoliation of graphene oxide [18–21]. This material consists of graphene-like sheets, chemically functionalized with compounds such as hydroxyls and epoxides, which stabilize the sheets in water [22, 23]. However, this functionalization disrupts the electronic structure of graphene. High-quality monolayer graphene can be produced at significant yields by non-chemical, solution-phase exfoliation of graphite in certain organic solvents. This work erected upon over 50 years of study into chemical
exfoliation of graphite . Previously, intercalated graphite could be partially exfoliated by reactions involving the intercalant , through thermal shock  or by acid treatment of expandable graphite . However, to date, such methods have given thin graphite sheets or graphene flakes rather than large-scale graphene monolayer. To make graphene flakes deposition fast and avoid vacuum processing, electrohydrodynamic atomization (EHDA) has been employed here. This technique has never been reported so far for graphene deposition. This process has the potential of becoming a reliable, cost effective and robust route for graphene flakes film. The EHDA process involves the flow of a solution, containing nano particles or polymers of the material to be deposited and here in this case graphene flakes (as a solute), under the influence of an electric field at ambient temperature [28, 29]. Due to the applied electric field, the liquid forms a cone at the orifice of the capillary and a jet evolves. In fact, an imbalance is induced between the surface forces arising because of the surface tension of the liquid to be sprayed and the Maxwell stresses because of the electric field that pulls the liquid downwards in the capillary into a stable jet that further disintegrates into smaller droplets because of coulomb forces; hence, a cloud of charged, mono-dispersed and extremely tiny droplets is achieved. The droplets generated from the spray are deposited on a substrate to form a uniform film of the solute. The jet breakup of a pendant droplet under the influence of an electric field was explored by Huang et al.  and then by Zeleny . Many modes of atomization, such as micro-dripping, spindle, cone-jet and multi-jet exists . Apart from other atomization modes, the stable cone-jet mode of atomization is the most important mode as near-mono-dispersed droplets of few micrometers in size are generated . The stable cone-jet mode occurs when there is comparative impartiality in the electric field strength and the surface tension of the flowing medium. The Applied voltage, the flow rate and the physical properties of the liquid, namely, density, viscosity, surface tension, relative permittivity and electrical conductivity, controls the droplet size generated from EHDA under the cone jet mode. Jaworek and Krupa  has studied EHDA phenomenon to classify the modes of EHD atomization. EHDA has been used by Samarasinghe et al.  to deposit gold films on silicon wafer substrates. Muhammad et al.  deployed EHDA for the deposition of CIS absorber layers for second generation solar cells. Chen and Schoonman  studied the use of EHDA and chemical vapor deposition for the fabrication of thin ceramic films for lithium batteries, organic photovoltaic cells and solid-oxide fuel cells. Jaworek et al.  used EHDA to form metal-oxide coatings on micro-fibers. In this work, a comprehensive description of graphene flakes film fabrication is described i.e. direct deposition of
J Mater Sci: Mater Electron (2013) 24:4893–4900
exfoliated graphene flakes from solution on to substrate by using EHDA process. For EHDA, graphene flakes dispersion is synthesized and its physical properties are measured to determine processing parameters. The graphene film deposited by using EHDA process is cured at 200 °C for 1 h. After curing, randomly oriented graphene flakes film has been characterized by using FE-SEM, UV–vis–NIR spectroscopy, X-ray diffraction (XRD), Raman Spectroscopy and 3D-Nanomap. Electrical behavior of the film is investigated by depositing graphene as cathode onto F8BT in diode device i.e. indium tin oxide (ITO)/(poly (3,4-ethylenedioxythiophene) poly(styrenesulfonate) (PEDOT:PSS)/polydioctylfluorene-benzothiadiazole (F8BT)/ graphene.
2 Experimental procedure 2.1 Materials and methods Graphite powder and N-methyl-pyrrolidone (NMP) solvent were purchased from Sigma-Aldrich. The graphite powder (0.02 g) was dispersed in NMP (8 ml) solvent. The solution was then bath sonicated for 30 min at room temperature. After sonication, vacuum filtration was done to remove large un-exfoliated graphite flakes. For further purification, filtrate was subjected to centrifugation. It was carried out for 30 min at 4,000 rpm. Supernatant was separated from sediment. The viscosity of the ink was measured to be 15.8 mPa by using Viscometer VM-10A system. The surface tension of the ink was measured to be 57–59 mN/m by using Surface-electro-optics (SEO)’s contact angle analyzer. The electrical conductivity of the ink was 12.2 lS/cm, measured by using conductivity meter (Cond6? meter). 2.2 Film deposition The deposition of graphene flakes was carried out using EHDA deposition system [39, 42]. A schematic diagram of the setup is shown in the Fig. 1. For processing, the ink is placed in an ink chamber (Nano NC Nozzle adapter) through a syringe pump (Hamilton, Model 1001 GASTIGHT syringe), and a constant flow rate is provided through the syringe pump. A metal nozzle of 110 lm internal diameter [Havard 33G] is used as the anode, and it is connected to a Trek Model 610E high voltage source for the generation of the required electric field. Ground is provided by connecting the ground terminal of the power source to a moving stage that holds the substrate. A highspeed camera, along with a light source, is used in conjunction with a portable computer to capture images of the events taking place at the capillary’s orifice. Before the
J Mater Sci: Mater Electron (2013) 24:4893–4900
Fig. 1 Schematic of the experimental setup
graphene deposition, silicon substrate is cleaned with acetone for 1 min using ultrasonication and then dried. The flow rate with a combination of varying potentials is used in order to observe atomization modes. The speed of moving stage is kept at 5 mm/s. Then, the ink is subjected to electrostatic atomization at an applied flow rate of 150 ll/h and a potential of 3.9 kV in the stable cone jet mode. For covering the (2 9 2) cm2 silicon wafer, 10 mm optimized stand-off is kept constant along with the deposition speed.
different magnitude of voltages showed different atomization modes such as the dripping to the multi-jet [43–45]. Figure 2 represents high-resolution and high-speed images of different atomization modes captured at 150 ll/h flow rate for increase in applied voltages. At 0 V, stable meniscus was observed. When the voltage was gradually increased from zero up to 2.3 kV, only the dripping occurred. At 2.8 kV, the micro-dripping mode and at 3.3 kV the pulsating cone-jet modes were observed, respectively. But when voltage increased to 3.8 kV, the stable cone-jet mode and with further increase to 4.3 kV, the multi-jetting were observed, respectively. Figure 3 provides the observed operating envelop of the graphene flakes dispersion for EHDA. A stable cone-jet region is shaded in operating envelop to emphasize the possible flow rate and voltage combinations. For film deposition, a flow rate of 150 ll/h was used throughout the experiments. The occurrence of the classical hydrodynamic atomization requires the hydrodynamic relaxation time Th to exceed the charge relaxation time Tq . The two quantities are given as:
2.3 Taylor cone and atomization modes Graphene flakes dispersion was remained stable during the experiments and no sedimentation was observed at any point. The EHDA experiments were initially performed with flow rates using from 50 to 1,000 ll/h in order to determine the optimum spraying conditions with the standoff distance fixed at 10 mm. At each flow rate, the strength of the electric field was increased by gradually increasing the potential difference between the anode nozzle and the grounded stage [38–42]. At each flow rate step, applying
Fig. 3 Operating envelop for different EHDA zones. The stable cone jet region is shaded to emphasize upon the possible flow rate and voltage combinations for optimized atomization conditions (standoff = 10 mm)
Fig. 2 Electrohydrodynamic atomization modes of graphene flakes dispersion observed during the deposition process, a stable meniscus b dripping, c micro-dripping, d pulsating unstable cone-jet, e stable cone-jet, f multi-jet
4896 Fig. 4 Optical microscopic images the deposited graphene flakes film (scale bar 50 lm)
J Mater Sci: Mater Electron (2013) 24:4893–4900
(a) Bare Silicon
(b) Deposited Graphene flakes film
Th ¼ LD2 =Q
Tq ¼ e0 er =K
here, L denotes the axial length of the jet, D is the jet diameter, Q is the flow rate and K is the electrical conductivity. Using the following values for these quantities, L & 800 lm, D & 10 lm, Q = 150 ll/h = 4.1667 9 10-11 m3/s, er = 2.3, e0 = 8.85 9 10-12 F/m, K = 0.0000122 S/m. It implies that Th & 0.00192 s which is much greater than Tq & 1.67 9 10-6 s. Thus, the required condition for classical EHDA is being fulfilled.
3 Results and discussions 3.1 Optical microscopy Figure 4a, b shows optical images of graphene flakes film deposited onto silicon substrate. It is observed that by using EHDA for deposition, the graphene flakes are randomly oriented on substrate due to which surface morphology of the film is non-uniform. In next section, SEM analysis of the film has further confirmed it. 3.2 SEM analysis Figure 5a–e shows the SEM images of the film taken at different resolutions i.e. from low to high resolutions. It is observed that the surface morphology of the deposited film is not uniform throughout. High resolution images have
confirmed randomly oriented deposited graphene flakes, which has resulted in non-uniform surface morphology of the film. For quantitative surface analysis of the film, 3DNanomap has been used for surface profilometry. 3.3 3D Nanomap The 3D surface profile, 2D surface profile and the X-profile of the graphene film deposited on Si substrates is shown in Fig. 6. It shows that its Ra value is 1.28 nm. X-profile of the film reflects that the film surface has many nano-level bumps (peak to valley), which means that graphene film deposited by using EHDA process has different number of graphene layers i.e. single layer, bilayer, few layer and multilayer morphology. A depiction of graphene flakes before and after deposition has been shown in Fig. 7. 3.4 X-ray diffraction analysis The crystal structure and crystalline size of the deposited graphene flakes film is determined by Rigaku XRD operated at 36 kV and 36 mA with Cu Ka radiation in the range of 10°–65° with a step of 0.02°. Figure 8 shows XRD profile of deposited graphene flakes film, two broad peaks at 2h = 24.7° and 2h = 42.3° are observed with an inter˚ which are assigned to layer spacing of 37 and 39 A graphene (002) and (100) plane, respectively. In the XRD pattern of graphene, a typical diffraction peak is broadened which indicates smaller crystalline size of graphene. The calculated d-spacing, suggesting that the distance between
Fig. 5 a–c Low magnification SEM surface view of the deposited films of graphene using deposition speed 5 mm/s. d, e High magnification SEM micrographs showing surface characteristics and very high magnification showing grain size
J Mater Sci: Mater Electron (2013) 24:4893–4900
Fig. 6 a 3D surface profile, b 2D surface profile and c X-profile of the graphene film deposited on Si substrate Fig. 7 Schematic depiction of the graphene flakes in dispersed form and after deposition on silicon substrate. a Graphene flakes dispersed in NMP. b Schematic of deposited graphene film
graphene layers is close to the theoretical value 0.34 nm . The crystallite size of graphene is 2.1 nm which is calculated by Scherer equation by using XRD radiation of wavelength k (nm) and measuring full width at half maximum of peaks (b) in radian located at any 2h = 24.7° in the pattern. 3.5 Raman spectra analysis Raman spectroscopy of the deposited graphene flakes film is carried out by using the LabRam HR800 micro Raman spectroscope (Horiba Jobin–Yvon, France). The Raman system is operated at the 10 mW laser power and an excitation wavelength of 514 nm with Ar? ion laser. Figure 9 shows the Raman spectrum of the deposited graphene flakes film on silicon substrate. The major features, commonly observed in all chemically processed graphene are D band at 1,354/cm, G band at 1,580/cm and 2D band at 2,725/cm. The G band at 1,580/cm corresponds to an E2g mode which is related to the sp2-bonded carbon atoms vibration in a 2D hexagonal lattice. And the D band at 1,354/cm arises from a breathing mode of k-point phonons of A1g symmetry. The high intensity of D band indicates the presence of sp2 C with defects. It is also reported that
the D band arises from the reduction in size of in-plane sp2 domains as well as the larger surface-to-volume ratio. Deposited graphene flakes film has strong in-plane sp2 bonds in two-dimensional system. The in-plane size of the crystallites is deduced from the intensities ratio between the G and D band integrated intensities (ID/IG). Using the expression given by Pimenta et al.  La ðnmÞ ¼ 560= E4l ðID =IG Þ ð3Þ where El(laser energy) = 3.86 eV and ID/IG = 0.8, therefore the in plane graphene crystallite is 3 nm. 3.6 UV/vis/NIR analysis Figure 10 shows the transmittance spectra of EHDA the deposited graphene flakes film in ultra-violet (UV), visible and near-infra-red (NIR) spectrum. The transparency of the film is recorded by a UV/vis/NIR spectrometer (Shimadzu UV-3150) with a range of 200–800 nm. Transmittance of the deposited thin film is observed 92–93 % in the visible range. This percentage transmittance is in accordance to our schematic interpretation of Raman spectroscopy of graphene film. The transmittance spectrum reveals that the
graphene film is more suitable for optoelectronic device applications. For band gap calculation, an absorbance spectrum of the film is taken. The optical band gap of
J Mater Sci: Mater Electron (2013) 24:4893–4900
graphene is estimated by fundamental relation given by equation aht = B(ht - Eg)n,where a is the absorption coefficient, hm is the energy of absorbed light, n = 1, for direct allowed transition and B is proportionality constant . Energy gap (Eg) was obtained by plotting (ahm)2 versus hm and extrapolating the linear portion of (ahm)2 versus hm to zero, as shown in Fig. 10b. The bandgap of graphene was estimated to be 3.95 eV by using this method. 3.7 Electrical behavior analysis
Fig. 8 XRD patterns of the graphene flakes film
Fig. 9 Raman spectra of the deposited graphene flakes film Fig. 10 a UV–vis spectrum of the graphene flakes film b (ahm)2 versus hm plot for bandgap estimation of graphene flakes film
In electronic devices contact between two adjacent layers is very much important for adequate electrical behavior. Therefore, to analyze the deposited graphene flakes film electrical behavior, Agilent B1500A semiconductor device analyzer is used. The current density–voltage (J–V) curve of the organic diode device with structure ITO/PEDOT:PSS/F8BT/graphene in linear scale recorded by semiconductor analyzer is shown in Fig. 11, in which graphene is used as cathode. The J–V plot confirms the proper contact between adjacent layers in diode device. At voltage of 0.3 V, the current density in organic structure is at low value of 2.67 A/cm2 and after that as further voltage was applied, the device current increased by almost 100 times and reaches up to 2.65 9 102 A/cm2 at voltage 4 V which indicates an increase in charge carrier injection. For investigation of current conduction mechanism in diode device i.e. ITO/PEDOT: PSS/F8BT/graphene, the I– V characteristic curve of the diode in log I-log V scale is shown Fig. 12. Three major regions have been found in the Fig. 12. At low voltages (i.e. 0.01–0.1 V) the current density is increased linearly with increasing the bias voltage. For voltages between 0.1 and 1 V, it is observed that the current density is directly proportional to the V1.74. With further increase of bias voltage, the current density shows V2.3 dependence. The transport in the second and third regions with slope C2 and this is like space-charge– limited-current (SCLC) mechanism. The first region is ohmic with slope about unity. The second region having a
J Mater Sci: Mater Electron (2013) 24:4893–4900
have shown that graphene flakes are randomly oriented and having defects but yet it has shown good electrical performance in diode device as cathode. A high current density of 2.65 9 102 A/cm2 at voltage 4 V is obtained. It shows that electronic transportation in randomly oriented graphene flakes film fabricated by EHDA technique is as good as other fabrication techniques. Therefore, it can be used to create a myriad of depositions and geometries necessary for electronic devices. Furthermore, it can be used to fabricate graphene-based composites which are a key requirement in many applications, such as thin-film transistors, optoelectronics and energy storage devices. The efficiency of film can further be enhanced by optimizing EHDA processing parameters and controlling film morphology i.e. to obtain less randomly oriented graphene flakes film. Fig. 11 I–V characteristics of ITO/PEDOT:PSS/F8BT/graphene diode device
Acknowledgments This study was supported by a grant from the cooperative R&D Program (B551179-08-03-00) funded by the Korea Research Council Industrial Science and Technology, Republic of Korea.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
Fig. 12 logI–logV characteristics of ITO/PEDOT:PSS/F8BT/graphene
slope equal to 1.74 indicates SCLC . The third region has a slope C2 which can be attributed to trap charged current limited mechanism. The presence of traps might be due to impurities, various defects and randomly oriented deposited graphene flakes.
4 Conclusion We report that the randomly oriented graphene flakes film is successfully fabricated by a non-vacuum, solution processible process EHDA, which has significant characteristics such as moderate voltage requirements, room temperature deposition and simple processing. An operating envelop is explored for graphene flakes dispersion. Characterizations
11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.
A.K. Geim, K.S. Novoselov, Nat. Mater. 6, 183 (2007) K.S. Novoselov et al., Nature 438, 197 (2005) K.S. Novoselov et al., Science 306, 666 (2004) Y.B. Zhang, Y.W. Tan, H.L. Stormer, P. Kim, Nature 438, 201 (2005) S. Pisana et al., Nat. Mater. 6, 198 (2007) P. Blake et al., Nano Lett. 8, 1704 (2008) K.S. Novoselov et al., Proc. Natl. Acad. Sci. USA 102, 10451 (2005) K.I. Bolotin et al., Solid State Comm. 146, 351 (2008) S.V. Morozov et al., Phys. Rev. Lett. 1, 016602 (2008) X. Du, I. Skachko, A. Barker, E.Y. Andrei, Nat. Nanotechnol. 3, 491 (2008) C. Berger et al., Science 312, 1191 (2006) T. Ohta et al., New J. Phys. 10, 023034 (2008) S.Y. Zhou et al., Nat. Mater. 7, 259 (2008) S. Marchini, S. Gunther, J. Wintterlin, Phys. Rev. B 76, 075429 (2007) A.L.V. de Parga et al., Phys. Rev. Lett. 1, 056807 (2008) P.W. Sutter, J.-I. Flege, E.A. Sutter, Nat. Mater. 7, 406 (2008) Y. Pan et al., Adv. Mater. 21, 2777 (2009) G. Eda, G. Fanchini, M. Chhowalla, Nat. Nanotechnol. 3, 270 (2008) D. Li et al., Nat. Nanotechnol. 3, 101 (2008) S. Stankovich et al., Nature 442, 282 (2006) X. Wang, L.J. Zhi, K. Mullen, Nano Lett. 8, 323 (2008) S. Stankovich et al., Carbon 45, 1558 (2007) I. Jung et al., Nano Lett. 7, 3569 (2007) M.S. Dresselhaus, G. Dresselhaus, Adv. Phys. 30, 139 (1981) L.M. Viculis, J.J. Mack, R.B. Kaner, Science 299, 1361 (2003) G.H. Chen et al., Carbon 42, 753 (2004) X.L. Li et al., Science 319, 1229 (2008) M.S. Oh, K. Lee, K.H. Lee, S.H. Cha, J.M. Choi, B.H. Lee, M.M. Sung, S. Im, Adv. Funct. Mater. 19, 726 (2009) D.K. Hwang, K. Lee, J.H. Kim, S. Im, J.H. Park, E. Kim, C.S. Kim, H.K. Baik, Electrochem. Solid-State Lett. 10, H117 (2007) Z.M. Huang, Y.Z. Zhang, M. Kotaki, S. Ramakrishna, S. Compos, Sci. Technol. 63, 2223 (2003)
4900 31. 32. 33. 34. 35. 36. 37. 38.
39. 40. 41. 42.
J. Zeleny, Phys. Rev. 3, 69 (1914) G.I. Taylor, Proc. R. Soc. Lond. A 313, 453 (1969) M. Cloupeau, B. Prunet-Foch, J. Electrostat. 25, 165 (1990) A. Jaworek, A. Krupa, J. Aerosol Sci. 30, 873 (1999) S.R. Samarasinghe, I. Pastoriza-Santos, M.J. Edirisinghe, L.M. Liz-Marzan, Appl. Phys. A 91, 141 (2008) N.M. Muhammad, S. Sundharam, H.W. Dang, A. Lee, B.H. Ryu, K.H. Choi, Curr. Appl. Phys. 11, S68 (2011) C.H. Chen, J. Schoonman, J. Ind. Eng. Chem. 10, 1114 (2004) A. Jaworek, A. Krupa, A.T. Sobczyk, M. Lackowski, T. Czech, S. Ramakrishna, S. Sundarrajan, D. Pliszka, Solid State Phenom. 140, 127 (2008) A. Khan, K. Rahman, M.T. Hyun, D.S. Kim, K.H. Choi, Appl. Phys. A Mater. Sci. Process. 104(4), 1113 (2011) N.M. Muhammad, A.M. Naeem, N. Duraisamy, D.S. Kim, K.H. Choi, Thin Solid Films 520, 1751 (2012) N.M. Muhammad, S. Sundharam, H.W. Dang, A. Lee, B.H. Ryu, K.H. Choi, Curr. Appl. Phys. 11, S68 (2011) K.H. Choi, M. Mustafa, K. Rahman, B.K. Jeong, Y.H. Doh, Appl. Phys. A 106, 165 (2012)
J Mater Sci: Mater Electron (2013) 24:4893–4900 43. A. Khan, K. Rahman, D.S. Kim, K.H. Choi, J. Mater. Process. Technol. 212, 700 (2011) 44. K. Rahman, J.B. Ko, S. Khan, D.S. Kim, K.H. Choi, J. Mech. Sci. Technol. 24(1), 307 (2010) 45. K.H. Choi, A. Ali, A. Rahman, N.M. Mohammad, K. Rahman, A. Khan, S. Khan, D. Kim, J. Micromech. Microeng. 20, 075033 (2010) 46. K. Rahman, A. Khan, N.M. Nam, K.H. Choi, D.S. Kim, Int. J. Precis. Eng. Man J. 4, 663 (2011) 47. J. Liu, H. Bai, Y. Wang, Z. Liu, X. Zhang, D.D. Sun, Adv. Funct. Mater. 20, 4175–4181 (2010) 48. M.A. Pimenta, G. Dresselhaus, M.S. Dresselhaus, L.G. Canc¸ado, A. Jorio, R. Saito, Phys. Chem. Chem. Phys. 9, 1276 (2007) 49. J. Tauc (ed.), Amorphous and Liquid Semiconductors (Plenum, New York, 1974), p. 173 50. S. Baranovski, Charge Transport in Disordered Solids with Applications in Electronics (Wiley, Chichester, 2006)