One-step in situ synthesis of single aligned graphene–ZnO nanofiber for UV sensing

Abstract

Here, we report a simple, one-step method for the in situ synthesis and alignment of a single graphene-doped zinc oxide (Gr–ZnO) composite nanofiber fabricated by electrospinning across electrodes and its subsequent use for ultraviolet (UV) detection. The study involves the optimization of the calcination temperature and time-dependent electrospinning for alignment of the Gr–ZnO composite nanofiber across gold electrodes. SEM analysis and Raman studies revealed the presence of ZnO embedded within uniformly distributed Gr flakes throughout the surface of the composite fiber. XRD analysis of Gr–ZnO confirmed a highly crystalline ZnO structure with amorphous graphene. Different weight percentages of graphene flakes were used to synthesize the composite to determine the optimum composition that enabled the synergistic contribution of both ZnO and graphene for UV sensing. I–V measurements under UV illumination showed that the composite to exhibit superior UV sensing performance with a ∼1892-fold increase in conductance for 0.5 wt% of graphene. The underlying mechanism of charge transfer of photogenerated electrons and holes from ZnO to graphene under UV illumination was studied. The method presented in this study provides a simple, effective and economically viable strategy for the alignment of various composite fibers and can therefore be used in a wide range of sensing applications.

---

1 Introduction

Near-ultraviolet detection is important for many applications such as environmental monitoring,¹ military applications,² flame detection³ and industrial quality control.⁴ In this regard, 1D zinc oxide (ZnO) nanostructures find wide usage, mainly due to their extraordinary electronic properties, which include a large room-temperature band gap of 3.37 eV,⁵ a high exciton binding energy of 60 meV,⁶ and good photoelectric properties by means of electron–hole generation or recombination during UV illumination. In addition, UV sensors based on ZnO nanofibers have exhibited a high on–off ratio and fast response.⁷ To further enhance their properties and expand their range of applications, efforts have been made to synthesize ZnO hybrids. Hoa et al.⁸ synthesized 2D NiO sheets and a 1D ZnO composite by a hydrothermal process to enhance the sensitivity of a UV sensor. Li et al.⁹ reported a polymer/ZnO hybrid for a wavelength-selective response in near-UV sensors. Xi et al.¹⁰ made a ZnO/SiO₂ nanofiber mat by electrospinning for a flexible UV sensor. 1D ZnO hybrids exist in different morphologies such as nanorods,¹¹ nanowires¹² and nanofibers,¹³ which can be synthesized using various methods such as chemical vapour deposition,¹⁴ hydrothermal methods,¹⁵ electrochemical deposition¹⁶ and electrospinning.¹⁷ Among the methods that have been mentioned, electrospinning is a simple and economically viable technique for synthesizing 1D ZnO hybrid nanofibers. More importantly, the electrospun nanofibers are continuous and easy to align from a wide range of materials and could be processed for applications. The main challenges in the fabrication of single nanofiber-based devices are their precise positioning between the electrodes and the establishment of an effective near-ohmic or ohmic contact. In most cases, nanofibers are first synthesized and then metal contacts are carefully positioned at their ends by e-beam lithography.¹⁸ However, such techniques are complex, result in low throughput and thus hinder the practical applicability of the sensor.

In this study, we report the one-step in situ synthesis and alignment of a single graphene-doped ZnO nanofiber by the use of electrospinning. Graphene, which is a two-dimensional carbon monolayer, has gained much attention in electronics and optoelectronics because of its excellent charge carrier transport mobility, high electrical conductivity, high optical transmittance and high mechanical stability.¹⁹ As well as these properties, its work function plays an important role in fabricating a UV sensor when used in conjunction with ZnO, as the fact that graphene's work function is lower than the conduction band of ZnO allows the transfer of photogenerated electrons to graphene. We demonstrate that time, which in this case is defined as the duration of operation of electrospinning, is a vital parameter for the alignment of a single nanofiber between pre-patterned electrodes. Our key strategy involves the optimization of the operating time for fabricating a single aligned Gr–ZnO nanofiber device in a controllable and reproducible manner. By varying the operating time, the number of fibers deposited between the electrodes could be controlled. The device was then evaluated for UV detection. Unlike conventional ZnO-based UV sensors, which suffer from issues related to the recombination of charge carriers, this approach utilizes the excellent transport properties and high optical transmittance of graphene to enhance the performance of the UV sensor. To the best of our knowledge, this is the first time that a detailed study of the effect of the electrospinning time on the density of ZnO-based fibers and subsequent in situ alignment of a single graphene–ZnO nanofiber composite for UV sensing has been reported.

2 Experimental

2.1 Materials

Polyacrylonitrile (PAN, M_w 150 000), dimethylformamide (DMF), and zinc acetate dihydrate (ZnAc) were purchased from Sigma Aldrich. Graphene flakes were procured from Graphene Supermarket, USA. The positive photoresist S1813 and its corresponding developer were procured from Shipley Inc., USA. All chemicals were analytically pure and were used as received without any further purification. DI water from a Millipore system (∼18.2 MΩ cm) was used throughout the experiment.

2.2 Sensor design and fabrication

Microfabricated gold electrodes were fabricated on a highly doped p-type silicon wafer using the cleanroom facilities available at IIT Hyderabad, India. In brief, a 300 nm SiO₂ layer was grown on a (100)-oriented p-type substrate by a thermal oxidation technique to insulate the substrate. This was followed by spin-coating the positive photoresist S1813 to define the source and drain areas by photolithography. A chromium layer (20 nm) for adhesion and a gold layer (180 nm) were deposited by DC sputtering. Finally, the electrodes were patterned by a lift-off technique by ultrasonication with acetone. The radius and gap of the electrodes were fixed at 100 μm and 50 μm, respectively.

2.3 Preparation of electrospun Gr–ZnO nanofiber

A Gr–ZnO nanofiber was synthesized according to the following procedure. PAN (8 wt%) and ZnAc (10 wt%) were dissolved in DMF. The mixture was stirred for 2 hours at room temperature. Thereafter, different weight percentages of graphene flakes, i.e., 0.2 wt%, 0.5 wt% and 0.8 wt%, were added to the solution and stirred for 24 hours at room temperature to obtain a homogeneous viscous solution that was ready for electrospinning. The electrospinning set-up consisted of a syringe, needle, grounded collector and high-voltage supply. ZnO and Gr–ZnO solutions were electrospun on fabricated gold electrodes that were patterned on a Si–SiO₂ substrate at room temperature. The patterned electrodes were kept 10 cm away from the tip of the 24-gauge needle. The voltage applied between the needle and the grounded collector was maintained at 18 kV with a constant flow rate of 8 μL min⁻¹. In addition to the standard electrospinning parameters, the operating time for electrospinning was optimized to obtain a single aligned nanofiber between the pre-patterned electrodes. The operating time for the current study was optimized to be in the range of 3–5 seconds, which produced a well-aligned single.

Gr–ZnO nanofiber between the gold electrodes. Thereafter, the devices were calcined at various temperatures such as 400 °C, 500 °C and 600 °C to remove organic contents and residues. The complete process, which includes the fabrication of the gold electrode by photolithography and electrospinning, is pictorially shown in Fig. 1.

Fig. 1 (a) Schematic of the microfabrication of gold electrodes; (b) electrospinning with collector as pre-patterned electrode; (c) UV sensing with single aligned Gr–ZnO nanofiber device.

2.4 Device characterization and sensing

Field emission scanning electron microscopy with a ZEISS Ultra-55 scanning electron microscope was used to study the surface morphology of the Gr–ZnO nanofibers and to determine the range of diameters of the fibers at various calcination temperatures. The electrical conductivity of a single nanofiber was measured using a cascade two-probe method. The current through the sample was measured with a Keithley 4200 SCS. The sample was measured four times in different directions by applying a potential of −1.0 to +1.0 V and an average value of the resistance was calculated. Raman studies were carried out on a Senterra inVia opus Raman spectrometer (Senterra, Bruker, UK) using 532 nm excitation and the optical power that was delivered onto the sample was 10 mW cm⁻². XRD studies (X'pert PRO) were performed using Cu Kα radiation. Electrical measurements with UV illumination were performed using a custom-assembled UV light of 12 W.

3 Results and discussions

Electrospinning of the Gr–ZnO composite was carried out for different time intervals so as to observe the time-dependent behaviour of the alignment of fibers between the electrodes. The operating time of electrospinning was varied from 15 seconds to 2 seconds. Fig. 2 shows SEM images of the device where the nanofibers were electrospun for 15, 12, 7 and 4 seconds. As the operating time for electrospinning was decreased from 15 seconds to 4 seconds, randomly arranged high-density nanofibers were converted into a single perfectly aligned nanofiber between the electrodes. At 15 seconds, a dense network of fibers that were randomly arranged between the electrodes was observed. With a decrease in the electrospinning time to 12 seconds, there was a significant reduction in the number of fibers between the electrodes, which still retained a random orientation. Fibers that were electrospun for 7 seconds displayed a further improvement in the arrangement and alignment of the fibers. Finally, at 4 seconds we observed a single fiber that was aligned perfectly between the electrodes. The procedure was repeated several times (7 times) and a 3–5 second period was found to be the optimum period for aligning a single fiber. A suitable explanation for this phenomenon of variation in the optimized time can be attributed to several physical parameters such as temperature, humidity etc. which affect the evaporation rate and viscosity of the solvent and thereby change the working parameters of electrospinning. This result was in close agreement with the work done by Sharma et al.,²⁰ where they reported the alignment of 1–5 carbon nanowires with a time in the range of 3–5 seconds with the help of a rotating drum. The use of a rotating drum for alignment on pre-patterned electrodes is not recommended as the position of the pre-patterned electrodes changes continuously, thus making it difficult to obtain an aligned fiber between the pre-patterned electrodes. Furthermore, in that study a focused ion beam was used to discard extra carbon fibers, if any. In the present study we optimized the parameters by carefully studying the density of fibers between the pre-patterned electrodes that were electrospun for various operating times. For all the above-mentioned operating times, the optimized electrospinning parameters as mentioned in the Experimental section were kept constant. Fig. 3 shows the near-linear relationship between the number of fibers between the electrodes and the operating time for 0.5 wt% graphene. As the electrospinning time increased we observed gradual curling behaviour of the fibers, which is in agreement with reports where the operating time of electrospinning was kept high.

Fig. 2 SEM images of the device electrospun for various time periods: (a) 15 seconds, (b) 12 seconds, (c) 7 seconds, and (d) 4 seconds. As the time decreased from 15 seconds, randomly arranged composite nanofibers were converted into a single perfectly aligned composite nanofiber between the pre-patterned gold electrodes.

Fig. 3 Graph demonstrating the near-linear relationship between no. of fibers and operating time of electrospinning for 0.5 wt% graphene.

Composite nanofibers that were synthesized at 4 seconds electrospinning time were calcined at different temperatures to study the morphology of the fibers with increasing temperature. Fig. 4a and b show high-magnification SEM images of pure ZnO and Gr–ZnO nanofiber, respectively, both calcined at 400 °C.

Fig. 4 High-magnification SEM images showing the morphology of fibres calcined at various temperatures: (a) pure ZnO calcined at 400 °C; (b) Gr–ZnO calcined at 400 °C; (c) Gr–ZnO calcined at 500 °C; (d) Gr–ZnO calcined at 600 °C. As the calcination temperature increased, it was observed that the nanofiber diameter and flake size decreased.

A pure ZnO nanofiber exhibits a smooth surface morphology, whereas the composite displays flake-like structures uniformly all over the surface of a ZnO nanofiber. The morphology of the composite, as seen in Fig. 4b, is a ZnO nanofiber embedded in graphene flakes. Furthermore, as can be clearly seen, the flakes are transparent, which is an inherent property of graphene. The diameter and morphology of the composite fibers changed based on the composition of the solution used for electrospinning. Upon the addition of graphene, the conductivity of the solution changed. As the synthesis of the composite was achieved by electrospinning, the conductivity of the solution is an important parameter in determining the diameter of nanofibers. In electrospinning, stretching of the polymer solution is caused by repulsion of charges on its surface. Hence, when the conductivity increases more charges are carried by the electrospinning jet, which causes the polymer solution to stretch more, thereby decreasing the diameter. In our work, upon the addition of graphene the conductivity of the solution increased and this in turn decreased the overall diameter and morphology of the composite. This observation is in agreement with various studies that are reported on electrospinning.^21,22 In addition, the initial dispersion of graphene is an important parameter that affects the diameter and morphology of the nanofiber.^23–26 It is evident from Fig. 4a and b that the initial graphene concentration in the electrospinning solution affected both the diameter and the morphology of the nanofiber. As can be seen in Fig. 4b, graphene is uniformly dispersed all over the surface of ZnO, which proves that 0.5 wt% graphene is sufficient to form a uniform dispersion. The dispersion of higher wt% graphene was not studied, as a UV sensor requires dark current to be low. An increase in the graphene content increases the dark current, thereby affecting the sensitivity of the sensor. This is explained in the latter part of the discussion, where optimization studies to select the optimum Gr/ZnO composition for UV sensing are discussed.

To study the effect of the calcination temperature on the composite, fibers were calcined at elevated temperatures of 500 °C and 600 °C. Fig. 4c and d show high-magnification images of the Gr–ZnO composite calcined at 500 °C and 600 °C, respectively. As can be clearly seen, there is a decrease in the fiber diameter (∼250 nm) and also the flake size. The decrease in the nanofiber diameter can be attributed to the decomposition and evaporation of organic components and residues. The decrease in the flake size of graphene was caused by the burning of carbon content at high temperatures, which is evident from the Raman spectra of the composite nanofiber calcined at 600 °C (Fig. S1 in ESI†), which further decreased the overall diameter of the nanofiber.²⁷ Hence, the optimized composite nanofiber calcined at 400 °C was chosen for UV sensing applications because of its higher graphene content.

Fig. 5 shows a high-magnification SEM image of the contact of a gold electrode and a Gr–ZnO composite nanofiber. The image was taken at a 45° tilt to ascertain that uniform contact had been established between the fiber and the metal electrode, which minimized the contact resistance between the two, thereby enhancing the transfer of electrons from the composite nanofiber to the gold contact and vice versa. This was achieved by calcination after in situ alignment of the fiber, which stabilized the contact. I–V curves that were recorded further confirmed the near-ohmic contact. The in situ alignment of fibers is advantageous in forming a near-ohmic contact, which involves heating both the electrode and the nanofiber simultaneously, thus avoiding the additional heating steps that are often required when the electrodes are fabricated after the electrospinning process.

Fig. 5 SEM image of contact between Gr–ZnO and gold electrode.

Fig. 6 shows X-ray diffraction (XRD) patterns of graphene, pure ZnO nanofibers and the Gr–ZnO nanofiber composite calcined at 400 °C in the range of 20° < 2θ < 60°. The 2θ peaks at 31.8° (100), 34.4° (002), 36.3° (101), 47.5° (102) and 56.6° (110) indicate the hexagonal wurtzite crystal structure of ZnO nanofibers. The diffraction peaks at 26.1° and 54.71°, which are denoted by G (002) and G (004), respectively, indicate the crystal planes of graphene. The broad graphene peak at 26.1° in the Gr–ZnO XRD pattern indicates amorphous carbon, which is due to defects that were induced by the addition of ZnO. Also, the low intensity of the peak can be attributed to the low concentration of graphene in the composite. The XRD data thus confirm the high crystallinity of Gr–ZnO composite nanofibers.

Fig. 6 XRD patterns of graphene, pure ZnO nanofibers and 0.5 wt% Gr–ZnO nanofiber composite calcined at 400 °C.

Detailed chemical/molecular analysis was performed in terms of Raman spectroscopy to confirm the formation of the composite. Fig. 7 presents the Raman spectrum of the Gr–ZnO composite with 0.5 wt% graphene flakes. The well-known G peak (∼1574 cm⁻¹), which is characteristic of an sp² hybridized carbon peak, is evident. The other peak, which is at ∼1359 cm⁻¹ and is denoted as the D band, which originates from structural imperfections, is also prominent. The intensity of the D peak is greater than that of the G peak, which also confirms the formation of the composite due to an increase in disorder caused by the addition of ZnO. Also, the addition of ZnO to graphene made it amorphous, which caused the intensity of the G peak to be low and not uniform.¹³ With 0.5 wt% graphene in the composite, the typical Raman peaks of ZnO disappear. This is due to the large number of graphene flakes that were embedded in the ZnO nanofiber, which concealed the Raman signals of ZnO.

Fig. 7 Raman spectrum of Gr–ZnO composite with 0.5 wt% Gr calcined at 400 °C.

As a part of optimization studies to select the optimum Gr/ZnO composition, 3 different weight percentages of graphene flakes, i.e., 0.2 wt%, 0.5 wt% and 0.8 wt%, were used to synthesize the composites and subsequent UV sensing was performed. It was found that for the sample containing 0.2 wt% graphene, the value of the dark current that was recorded was the lowest; however, upon UV exposure the resulting current became saturated after 30 minutes, thus yielding a ∼586-fold change in conductance as opposed to a ∼1892-fold change for the sample with 0.5 wt% graphene. For 0.8 wt% graphene, the dark current that was recorded was high, making it unsuitable for UV sensing applications. Hence, in order to achieve better sensitivity, a 0.5 wt% graphene composition was selected to synthesize the composite. Figures that show values of dark resistance measured as a function of different compositions and the I–V response of a typical device with 0.2 wt% graphene can be found in the ESI (Fig. S2 and S3, respectively†). Fig. 8 shows the current–voltage (I–V) response of a typical single aligned Gr–ZnO composite nanofiber device with 0.5 wt% Gr calcined at 400 °C. The reason for choosing the sample calcined at 400 °C was the greater visibility of the flake size in the SEM images, which indicates the presence of more graphene, which helps enhance the transport properties of the photogenerated electrons. Moreover, graphene decomposes above 500 °C, which is evident from Raman spectra, as discussed earlier in this section. The I–V response under UV illumination was measured at 5, 15, 30, and 45 minutes to study the response of current with time. Under UV illumination, electron–hole pairs are generated when the energy of the illumination is greater than or equal to the band gap of ZnO. For pure ZnO, photogenerated electrons tend to recombine, which decreases the carrier lifetime and hence the current. Hence, the need for graphene as a transport material arises for capturing photogenerated electrons, thereby preventing them from recombining and ultimately increasing the carrier lifetime. In the case of the Gr–ZnO nanofiber composite, as the work function of graphene is lower than the conduction band of ZnO²⁸ photogenerated electrons are readily transferred to graphene and the higher electron mobility of graphene enables faster collection of charge carriers at the gold electrodes, thereby increasing the carrier lifetime of the photogenerated electrons. For pure ZnO nanofibers, in the absence of UV light a depletion layer is formed due to the oxidation of n-type ZnO nanofibers by adsorption of atmospheric oxygen molecules, thereby decreasing the conductivity.²⁹ Before starting to measure the current under UV illumination, samples were kept in the dark for 12 hours so as to stabilize them. Under exposure to UV light for 5 minutes there was a 69% decrease in resistance for 0.5 wt% graphene, which indicates a quick response to UV illumination for both devices. After 15 minutes, a 98.67% decrease in resistance for 0.5 wt% graphene was observed. This increase in conductance can be attributed to the faster collection of photogenerated electrons by graphene. As the time of illumination was increased to 30 minutes and finally 45 minutes, 99.8% and 99.9% decreases in resistance were observed, respectively. As time increased, the amount of photogenerated electrons that were transferred to graphene increased, thereby increasing the current. After 30 minutes of UV exposure, the decrease in the resistance was less compared to the decrease in 15–30 minutes for 0.5 wt% graphene and it became saturated after 40 to 45 minutes of UV illumination. The above phenomena can be attributed to the decomposition of graphene due to photogenerated holes and electrons, which react with atmospheric oxygen and moisture, thereby forming reactive oxygen species that oxidize graphene.³⁰

Fig. 8 I–V curves of a typical Gr–ZnO device with 0.5 wt% Gr with and without UV illumination at various time instants.

Fig. 9 provides a schematic representation of the process. Photogenerated electrons tend to be transported towards the gold electrode due to the very high charge transport mobility of graphene, but few electrons react with atmospheric oxygen to form O₂ radicals. Also, photogenerated holes react with atmospheric oxygen and form hydroxyl (OH) radicals. These reactive oxygen species that are generated on the surface of graphene are responsible for the oxidation of graphene, thereby reducing the carbon content and increasing the oxygen content and hence leading to saturated conductivity. There have been few studies that reported the use of ZnO/graphene composites for UV sensing. Dang et al.³¹ fabricated a phototransistor with a channel based on vertical ZnO nanorods and graphene. Boruah et al.³² demonstrated a UV sensor based on ZnO nanowires grown on graphene foam. Fu et al.³³ reported a high-performance single ZnO nanowire produced using chemical vapor deposition and sandwiched between graphene sheets with an on–off ratio of 800.

Fig. 9 (a) Schematic of the charge separation of photogenerated electrons and holes and their role in the oxidation of graphene under UV illumination; (b) band structure of Gr–ZnO nanofiber composite with gold electrodes.

Liang et al.³⁴ demonstrated the one-step synthesis of a SnO₂/graphene nanocomposite by a solvothermal method for application in Li-ion batteries, but in our approach we demonstrated in a single step not only the in situ synthesis of a ZnO/graphene nanofiber but also the alignment of a single composite fiber across the electrodes, thus integrating it into a device. This controlled approach to the synthesis and precise positioning of the fiber can be performed simultaneously across multiple electrodes on a wafer, thus making this technique suitable for large-scale commercial applications.

Zhang et al.³⁵ synthesized a ZnO nanowire using an electrodeposition method with a transparent graphene contact. Chang et al.³⁶ demonstrated the use of a ZnO nanorod/graphene composite produced by a solution growth method, which was then drop-cast between metal electrodes. Wang et al.³⁷ fabricated a flexible UV sensor with a reduced graphene oxide/ZnO composite synthesized by a hydrothermal method. Unlike most of the above-mentioned works, which involve working with high-density nanofibers in the form of mats or nanorods, or complex fabrication techniques such as defining electrodes using e-beam lithography, a transfer process or drop-casting, which even in ideal conditions will lead to device-to-device variation, our work involves a simple, reproducible approach for the one-step in situ alignment of graphene–ZnO using electrospinning. This method is advantageous as it allows precise manipulation of the resistance of a sensor upon UV exposure by aligning a single composite nanofiber across the electrodes. In addition, the current work employs a two-terminal device configuration, which can be easily integrated further for an all-electronic readout signal for real-time applications.

4 Conclusions

In summary, a facile, one-step in situ method for producing a UV sensor based on a single aligned graphene–ZnO nanofiber composite was developed. The time-dependent behaviour of the alignment of nanofibers was studied and was optimized for the alignment of a single fiber across the pre-patterned electrodes. Structural and chemical characterization revealed that the uniform morphology of the composite consists of ZnO embedded in graphene flakes throughout the surface of the fiber. It was observed that approximately 4 seconds of electrospinning time is required to obtain a perfectly aligned single graphene–ZnO fiber. Under UV illumination for 45 minutes, a significant decrease in the resistance of the device was observed. Our proposed approach to an in situ aligned composite provides a simple, cost-effective alternative for developing an integrated UV sensor and can be extended to a variety of other sensing applications.

Acknowledgements

A part of the reported work (characterization) was carried out at the IITBNF, IITB under INUP which is sponsored by DeitY, MCIT, Government of India.

References

1. M. Razeghi and A. Rogalski, J. Appl. Phys., 1996, 79(10), 7433–7473.
2. M. H. Huang, S. Mao, H. Feick, H. Yan, Y. Wu, H. Kind and P. Yang, Science, 2001, 292(5523), 1897–1899.
3. S. Basu and A. Dutta, Sens. Actuators, B, 1994, 22(2), 83–87.
4. Q. Wan, Q. H. Li, Y. J. Chen, T. H. Wang, X. L. He, J. P. Li and C. L. Lin, Appl. Phys. Lett., 2004, 84(18), 3654–3656.
5. B. E. Sernelius, K. F. Berggren, Z. C. Jin, I. Hamberg and C. G. Granqvist, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37(17), 10244.
6. A. Yousef, N. A. Barakat, T. Amna, A. R. Unnithan, S. S. Al-Deyab and H. Y. Kim, J. Lumin., 2012, 132(7), 1668–1677.
7. S. Bai, W. Wu, Y. Qin, N. Cui, D. J. Bayerl and X. Wang, Adv. Funct. Mater., 2011, 21(23), 4464–4469.
8. L. T. Hoa, H. N. Tien and S. H. Hur, Sens. Actuators, A, 2014, 207, 20–24.
9. H. G. Li, G. Wu, H. Z. Chen and M. Wang, Sens. Actuators, B, 2011, 160(1), 1136–1140.
10. M. Xi, X. Wang, Y. Zhao, Z. Zhu and H. Fong, Appl. Phys. Lett., 2014, 104(13), 133102.
11. O. Akhavan, ACS Nano, 2010, 4(7), 4174–4180.
12. J. O. Hwang, D. H. Lee, J. Y. Kim, T. H. Han, B. H. Kim, M. Park and S. O. Kim, J. Mater. Chem., 2011, 21(10), 3432–3437.
13. S. An, B. N. Joshi, M. W. Lee, N. Y. Kim and S. S. Yoon, Appl. Surf. Sci., 2014, 294, 24–28.
14. X. Dong, Y. Cao, J. Wang, M. B. Chan-Park, L. Wang, W. Huang and P. Chen, RSC Adv., 2012, 2(10), 4364–4369.
15. X. Zhou, T. Shi and H. Zhou, Appl. Surf. Sci., 2012, 258(17), 6204–6211.
16. Z. Yin, S. Wu, X. Zhou, X. Huang, Q. Zhang, F. Boey and H. Zhang, Small, 2010, 6(2), 307–312.
17. S. Xu and Z. L. Wang, Nano Res., 2011, 4(11), 1013–1098.
18. A. Stafiniak, B. Boratyński, A. Baranowska-Korczyc, K. Fronc, D. Elbaum and M. Tłaczała, J. Am. Ceram. Soc., 2014, 97(4), 1157–1163.
19. M. J. Allen, V. C. Tung and R. B. Kaner, Chem. Rev., 2009, 110(1), 132–145.
20. S. Sharma, A. Sharma, Y. K. Cho and M. Madou, ACS Appl. Mater. Interfaces, 2012, 4(1), 34–39.
21. C. J. Angammana and S. H. Jayaram, IEEE Trans. Ind. Appl., 2011, 47(3), 1109–1117.
22. S. H. Tan, R. Inai, M. Kotaki and S. Ramakrishna, Polymer, 2005, 46(16), 6128–6134.
23. V. A. Davis, A. N. G. Parra-Vasquez, M. J. Green, P. K. Rai, N. Behabtu, V. Prieto and M. Pasquali, Nat. Nanotechnol., 2009, 4(12), 830–834.
24. T. Xu and V. A. Davis, Langmuir, 2014, 30(16), 4806–4813.
25. C. Jiang, A. Saha, C. C. Young, D. P. Hashim, C. E. Ramirez, P. M. Ajayan and A. A. Martí, ACS Nano, 2014, 8(9), 9107–9112.
26. C. Jiang, A. Saha, C. Xiang, C. C. Young, J. M. Tour, M. Pasquali and A. A. Martí, ACS Nano, 2014, 7(5), 4503–4510.
27. K. Pramoda, S. Suresh, H. R. Matte and A. Govindaraj, Bull. Mater. Sci., 2013, 36(4), 585–590.
28. P. Roy, A. P. Periasamy, C. T. Liang and H. T. Chang, Environ. Sci. Technol., 2013, 47(12), 6688–6695.
29. H. Kind, H. Yan, B. Messer, M. Law and P. Yang, Adv. Mater., 2002, 14(2), 158.
30. D. H. Mun, H. J. Lee, S. Bae, T. W. Kim and S. H. Lee, Phys. Chem. Chem. Phys., 2015, 24(17), 15683–15686.
31. V. Q. Dang, T. Q. Trung, L. T. Duy, B. Y. Kim, S. Siddiqui, W. Lee and N. E. Lee, ACS Appl. Mater. Interfaces, 2015, 7(20), 11032–11040.
32. B. D. Boruah, A. Mukherjee, S. Sridhar and A. Misra, ACS Appl. Mater. Interfaces, 2015, 7(19), 10606–10611.
33. X. W. Fu, Z. M. Liao, Y. B. Zhou, H. C. Wu, Y. Q. Bie, J. Xu and D. P. Yu, Appl. Phys. Lett., 2012, 100(22), 223114.
34. J. Liang, W. Wei, D. Zhong, Q. Yang, L. Li and L. Guo, ACS Appl. Mater. Interfaces, 2012, 4(1), 454–459.
35. H. Zhang, A. V. Babichev, G. Jacopin, P. Lavenus, F. H. Julien, A. Y. Egorov and M. Tchernycheva, J. Appl. Phys., 2013, 114(23), 234505.
36. H. Chang, Z. Sun, K. Y. F. Ho, X. Tao, F. Yan, W. M. Kwok and Z. Zheng, Nanoscale, 2011, 3(1), 258–264.
37. Z. Wang, X. Zhan, Y. Wang, S. Muhammad, Y. Huang and J. He, Nanoscale, 2012, 4(8), 2678–2684.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra15351d

---