Jianhua Zhang,
Zhangfu Chen,
Xiaoxue Xu,
Wei Liao and
Lianqiao Yang*
Key Laboratory of Advanced Display and System Applications, Ministry of Education, Shanghai University, Yanchang Road 149, Shanghai 200072, China. E-mail: yanglianqiao@i.shu.edu.cn
First published on 14th November 2017
In this paper, a novel and scalable method to fabricate graphene/carbon nanotube (CNT) hybrid transparent conductive films on Cu substrates, which combines electroplating and chemical vapor deposition (CVD) is proposed and demonstrated. The Cu substrate was electroplated with electrolyte containing conductive CNTs; then, a uniform graphene film was grown on Cu. After a commonly utilized polymethyl methacrylate assisted transfer process, a hybrid graphene/CNT transparent conductive film was obtained at the target substrate. Conventional graphene grown on electropolished Cu was used as the reference sample. The comprehensive characterization using scanning electron microscopy (SEM), Raman microscopy system, and transmission electron microscopy (TEM) selected area electron diffraction pattern show that the CNTs are uniformly covered by a monolayer graphene with comparable quality to graphene grown on electropolished Cu. The hybrid thin films exhibit outstanding surface morphology (RMS of 1.26), obviously enhanced electrical properties (the square resistance decreases from 490 to 394 Ω sq−1), better surface wettability (7° decrease in contact angle), and a negligible transmittance loss (1.3% reduction at 550 nm) compared to CVD graphene that was grown on electropolished Cu. It is anticipated that the graphene/CNT hybrid films that are fabricated using the proposed approach can be a promising alternative to ITO to realize the emerging, particularly flexible optoelectronic devices.
Graphene, a two-dimensional material that consists of a single sheet of hexagonal carbon atoms, has attracted tremendous interest because of its unique optical, electrical, mechanical and physical properties.6–9 Carbon nanotube (CNT), a one-dimensional allotrope of graphene, is also attractive to researchers because of the outstanding physical properties in the longitudinal direction.10 To combine the advantages of the two-dimensional graphene and one-dimensional CNT, many researchers have devoted to fabricate hybrid nanostructures and attempted their applications in optoelectronic devices.11–14 For example, Tung et al. demonstrated a hybrid nanostructure of reduced graphene oxide (rGO) and CNT for transparent conductors.11 Using the chemical vapor deposition (CVD) method, Qiu et al. prepared a nanocomposite, where freestanding aligned CNT arrays grew on a large-area single-layered graphene sheet.12 Wang et al. fabricated three-dimensional few-layer graphene and CNT foam architectures.13 Using a two-step CVD method, Dong et al. synthesized the 3D graphene/CNT hybrid structures to apply to electrochemical sensing.14 However, these hybrid nanostructures either use non-ideal rGO as the substitute of graphene or require complex fabrication processes. In addition, the CNTs are usually spin-coated or printed onto the target substrate and physically attached to graphene, which cannot offer an effective electron transport network between CNT and graphene.
Here, this paper reports a simple method to form graphene/CNT hybrid thin films using one-step CVD growth on a Cu foil substrate pre-electroplated with conductive CNT. The performance of hybrid TCF is significantly enhanced, and the possible reasons of this enhancement are explored.
The surface condition is an important factor to elucidate the quality of graphene, including the pre-treatment, fabrication and transfer processes. Fig. 2 shows the surface morphology of three types of TCFs. The SWCNT with a diameter of approximately 3 nm is clearly observed in Fig. 2a. It is easy to understand that the Cu on the anode has a smoother surface and less impurities than that on cathode, and the surface morphology of the Cu substrate is directly proportional to the graphene surface condition. However, the root-mean-square (RMS) of hybrid TCF is only approximately 1.26 nm, which is lower than graphene(c) (1.58 nm) and graphene(a) (2.02 nm). The good surface morphology of hybrid films likely indicates that the SWCNT acted as a “filler” and flattened the rugged Cu surface, thereby increasing the flat level of the catalyst in the CVD process. The SWCNT also played the role of a “soft” buffer that decreased the mismatch in surface morphology between the soft Cu substrate and the rigid SiO2/Si substrate in the transfer process. Thus, a low surface roughness and a small value of RMS on the hybrid films were obtained.20 In order to evaluate the thickness of hybrid film, an artificial scratch was made and the thickness was found to be about 0.8 nm as shown in Fig. 3a and b. The improvement in RMS can be more useful and applicable for its utilization in organic optoelectronic devices because they commonly require good surface morphology.
Fig. 2 Surface roughness of (a) SWCNT/graphene hybrid thin films, (b) graphene(a), (c) graphene(c), transferred on Si/SiO2 substrate. |
Fig. 3 (a) AFM image of SWCNT/graphene hybrid thin films with an artificial scratch, (b) profile plot of the height along the line in (a). |
To further investigate the surface condition, SEM images of graphene/SWCNT hybrid TCFs on the SiO2/Si substrate were captured, and the results are shown in Fig. 4a. Fig. 4a shows that SWCNT was perfectly wrapped by the graphene layer, and no damage was observed on the SWCNT, which demonstrates that SWCNT remains stable during the electroplating and CVD processes. Furthermore, no impurities and defects such as wrinkles, folds, cracks and ripples were observed on the graphene layer of hybrid TCF, which also indicates that the TCFs possess a good surface morphology.
Fig. 4 (a) SEM and (b) Raman spectra of TCFs on SiO2/Si substrate at laser excitation wavelength of 532 nm. |
Raman spectroscopy, which is a typical and effective method to evaluate the structure and defects of carbon-based low-dimensional materials, was used to evaluate the TCF, and the results are shown in Fig. 4b. Fig. 4b displays the average Raman spectra of 5 points in TCF with excitation wavelength of 532 nm on the SiO2/Si substrate. The low D peak at ∼1350 cm−1 of graphene, which corresponds to the degree of crystallinity and defect, indicates that the CVD graphene has high quality. The two prominent peaks at ∼1570 and 2700 cm−1 can correspond to the G and 2D bands, respectively.21 The G band is created by the E2D phonon at the central area of Brillouin zone, and the 2D band represents double-resonant Raman scattering.22 The peak ratio of G/2D is 0.85, and the symmetric 2D band is centered at 2701 cm−1 with a full width at half maximum of 33 cm−1, which demonstrates that the graphene sheet is monolayer dominated.23,24 The intensities of the G and 2D band peaks of SWCNT are distinctly higher than those of the graphene film possibly because of the SWCNT has a significantly stronger scatter signal. For the hybrid TCFs, the peak G/2D ratio of 0.65 is between those of SWCNT (0.35) and graphene (0.85), which indicates the presence of both CNT and graphene. In addition, the peak at ∼2450 corroborates the combination of CNT and graphene, although its interpretation was still subject to debate.25
To further confirm the crystallinity and the layers of graphene, SAED was conducted on the graphene transferred to TEM grid and the results are shown in Fig. 5. The clear diffraction pattern (Fig. 5a) identifies the good crystallinity of graphene. The line profile of the diffraction patterns (Fig. 5b) demonstrates the monolayer nature of the graphene domain.
Fig. 5 (a) SAED pattern of graphene, (b) profile plots of the diffraction peak intensities along the arrows in (a). |
Optical transmittance is an important factor for TCFs. Fig. 6a shows the optical transmittance spectra of TCFs transferred to glass substrates. The optical transmittance of graphene(c) was 96.5%, which was slightly lower than that of graphene(a) (97.2%) at the wavelength of 550 nm. The transmittance difference was mainly caused by the surface impurities of Cu at the anode and cathode. In contrast, the transmittance of hybrid TCF was 95.9%, which is slightly lower than the two others as anticipated. However, the transmittance of hybrid TCF remains higher than 94% in the visible light range of 400–800 nm, whose performance is more than acceptable for its application as TCF. The inset graph of Fig. 6a shows the scene through the SWCNT/graphene hybrid film on glass; the red dotted rectangular shows the area of the sample with the size of 2 × 2 cm2. No vision disparity is observed in the graph, which also confirms the uniformity and transparency of the hybrid films.
As an alternative of ITO, the electrical conductivity is the foremost issue of TCF in addition to the optical performance and surface condition. The sheet resistance of three types of TCFs that were transferred onto the SiO2/Si substrates was measured by Hall effect measurement system, and the results are shown in Fig. 6b. Fig. 6b shows that the sheet resistance of the hybrid TCF is 394 Ω sq−1 with a transmittance of 95.9% at the wavelength of 550 nm, which is significantly lower than those of graphene(a) and graphene(c). In particular, few numbers of SWCNT from Fig. 6a can distinctly decrease the sheet resistance of the SWCNT/graphene hybrid films, which confirms that few SWCNTs can add conductive channels and repair line defects to decrease the electrical resistance of CVD-grown graphene. The graphene/SWCNT hybrid films were investigated by Yun and Yang, and it has been shown that the decrease in sheet resistance significantly depends on the CNT concentration.11,26 In this study, the concentration of the SWCNT solution is 0.0004 mg mL−1, which is far below the concentration in the literature with the identical range of transmittance and sheet resistance. In other words, the electroplating method of the CNT to the Cu substrate shows better enhancement effect than the commonly used approach to fabricate graphene/CNT hybrid TCF. For the CNTs that are spin-coated or printed on the target substrate, the physical adhesion of CNT to graphene cannot form effective chemical bonding, which hinders the carrier transport between CNT and graphene. In this study, the CNT was partially embedded in Cu, which acted as the catalyst of graphene; during the deposition process, the chemical bond might form between graphene and CNT. Thus, the contact resistance between graphene and CNT significantly decreased; then, the carrier transport was enhanced compared to conventional methods. Because of the extremely low concentration of the SWCNT solution, good balance between sheet resistance and transmittance for the nanoscale thickness TCF was obtained and further demonstrates that the possible formation of the chemical bond between graphene and CNT strengthens the carrier transport and conductivity.
The wetting properties of TCF to function materials directly affect the final application in optical devices. The Contact Angle Instrument was used to evaluate the wetting performance, and the results are shown in the inset graphs of Fig. 6b. The water contact angles (WCAs) of graphene(a) and SWCNT/graphene are 88.9° and 82.3°, respectively, which indicates that SWCNT/graphene is more hydrophilic than graphene(a). The wetting property of graphene-based TCF is a widely concerned issue, and the WCAs vary from ∼40° to ∼90° in the literature.27,28 It has been widely accepted that the WCA is related to the physicochemical composition, surface roughness, surface functional groups and doping level of graphene-based TCF, which eliminates the equipment error.28 Here, we would not strive to distinguish the specific reason of hydrophilic improvement because it is not the main point of this study. Instead, we would like to emphasize that the increase in wetting property of SWCNT/graphene can be helpful in the application of transparent conductive electrodes in optoelectronic devices.
In this study, we only investigated the SWCNT at a certain concentration with the electroplating target of Cu and corroborated the quality enhancement effect of hybrid TCF. In fact, we believe that the proposed approach is applicable to any conductive CNT and graphene catalyst as long as it requires electro-polishing pre-treatment.
This journal is © The Royal Society of Chemistry 2017 |