Graphene and graphene oxide double decorated SnO₂ nanofibers with enhanced humidity sensing performance

Abstract

We demonstrate humidity sensing with SnO₂@G–GO nanocomposites using three important parameters for a sensing device: sensitivity, response and recovery time, and stability. Here, the SnO₂@G–GO nanocomposites were fabricated by classical electrospinning and solution evaporation. The as-prepared SnO₂@G–GO sensor demonstrated very high sensitivity (up to 32 MΩ/% RH), fast response and recovery time (less than 1 s), and good stability. Pure SnO₂ and SnO₂@G hybrid NFs were prepared as reference materials for humidity sensing. And they showed low sensitivity and slow response to humid air. The performance for the incorporation of graphene and graphene oxide with SnO₂ to greatly improve the humidity sensing properties were discussed in detail.

---

1. Introduction

To meet the demands of future auto-control and data acquisition systems, higher qualified humidity detection and control devices are required.^1,2 In order to further promote humidity sensor properties (sensitivity, selectivity, chemical and thermal stability) intensive efforts have been made to explore new sensing materials, beyond traditional ceramic, semiconducting and polymeric materials.³

Graphene (G), a one-atom-thick two-dimensional (2-D) layers of sp²-bonded carbon, has been extensively investigated due to its unique structure and remarkable electronic properties,^4–8 also it has been proved that water molecules could effectively incorporate into graphene structure even at low humidity,^9,10 indicating that graphene can be a good candidate for novel humidity sensing material.

Moreover, functionalized graphene materials possess exceptional biological and chemical sensing properties.^11,12 Graphene oxide (GO) is a graphene derivative, which possesses many oxygen-containing functional groups, such as hydroxyl and epoxide groups on the basal plane, carbonyl and carboxyl groups at the edges, making it a strongly hydrophilic electrical insulator.^13–16 It is interesting to incorporate graphene with GO in a simple and controllable way for advanced humidity sensing application. Both graphene and GO are flexible and it is difficult to use them directly. Usually they are coated on the surface of metal oxide nanofibers (NFs).¹⁷

Electrospinning is a simple technique for the fabrication of various of 1-D nanomaterials, with diameter ranging from several micrometers down to tens of nanometers.^18,19 1-D SnO₂, which may yield potential applications in areas such as photocatalyst,²⁰ lithium storage capacity,^21–23 sensing material,^24–26 have been extensively studied. The incorporation of flexible graphene and GO with SnO₂ nanostructure can commendably construct conductive self-standing network with high sensitive surface.^5,27,28

In this work, we have used a classical electrospinning technique to successfully fabricate SnO₂@graphene (SnO₂@G) hybrid NFs. The SnO₂@G hybrid NFs were then wrapped by well-stretched GO, forming SnO₂@G–GO nanocomposites via solvent evaporation, as reported previously.^23,29 Industrial SnO₂ powder, pure electrospun SnO₂ NFs and SnO₂@G hybrid NFs were used as reference samples. Both pure SnO₂ based sensors took long time (more than 10 s) to response. With the addition of graphene nanosheets, the response and recovery time were greatly improved (2 s and 4 s). While the decorated GO further enhanced the response behavior and sensitivity. Therefore, it is reasonable to believe that metal oxide composites with graphene and graphene oxide are promising for fabricating devices with advanced humidity sensing properties.

2. Experimental section

2.1 Materials synthesis

All reagents were used without further purification. The graphene (Fig. S1†) has been prepared by micromechanical cleavage in which highly oriented pyrolytic graphite is peeled by Scotch tape,^19,30,31 forming graphene nanosheets under intense ultrasound. GO (Fig. S2†) was prepared by Hummers method.³¹ Both SnO₂@G hybrid NFs and SnO₂ NFs were prepared by electrospinning. As for typical electrospinning process, the spinneret had an inner diameter of 0.6 mm. Grounded steel strip were used as the collector. A distance of 15 cm and a direct current voltage of 20 kV were applied between the tip of the spinneret and the collector. After electrospinning, the fibers were heated from room temperature to 450 °C at a rate of 2.5 °C min⁻¹, and then held at 450 °C for 2 h in air.

2.1.1 Synthesis of SnO₂ NFs and SnO₂@G NFs. Electrospinning solution was prepared by adding SnCl₂·2H₂O to ethanol and N,N-dimethyl formamide (DMF) mixture (1 : 1) and stirring for 1 h at room temperature. Graphene was then added into the resulting solution and under intense ultrasound for 20 min. 10 wt% polyvinyl pyrrolidone (PVP, Sigma-Aldrich, M_w ≈ 1 300 000) was added to the solution and stirred vigorously for another 3 h. The as-prepared solution was loaded into a syringe with a stainless needle. The electrospun PVP/SnCl₂/graphene NFs were collected and dried in vacuum. Followed by annealing, the SnO₂@G NFs were obtained. The reference SnO₂ NFs were prepared in similar way without the adding graphene step.

2.1.2 Synthesis of SnO₂@G–GO nanocomposites. The GO was dissolved in deionized water by ultrasonic dispersion and magnetic stirring. The SnO₂@G hybrid NFs and the GO solution (GO : SnO₂@G NFs = 1 : 19 w/w) was added into 200 ml mixed solution of ethanol and deionized water. Keep stirring at 60 °C till the ethanol solution was completely evaporated. The powder was collected and preserved drying in the oven.

2.2 Material characterization

The morphologies and microstructures were characterized by scanning electron microscopy (SEM, Hitachi S-4800) and transmission electron microscopy (TEM, JEOL 2010). The crystal structure of the materials was determined by X-ray diffraction (XRD, RIGAKUD/Max-2550 with Cu Kα radiation). The specific surface area and pore diameter were confirmed by Belsorp-Mini II analyser (Japan). Graphene oxide was characterized by Fourier transform infrared Nicolet-5700 spectrophotometer (American). Raman spectra was acquired with a Labram-010 Raman spectrometer under an excitation of the 633 nm laser. The chemical state of the components was further confirmed by X-ray photoelectron spectroscope (XPS, Surface analysis PHI5600, Physical Electronics).

2.3 Electrochemical measurements

The humidity sensing tests were carried out with the aid of interdigital gold electrodes on a ceramics substrate. The humidity sensors were prepared as follows: firstly, the SnO₂@G–GO nanocomposites were dispersed in ethanol under ultrasound, then the as-prepared solution was spread on the interdigital gold electrodes by drop-cast. Finally, the as-prepared device were dried in the oven. For comparison, industrial SnO₂ powder, electrospun SnO₂ NFs and SnO₂@G hybrid NFs based sensors were prepared in the same way. The voltage applied between the two electrodes was 5 V. The electrical properties of the humidity sensors were measured by a high precision sensor testing system NS-4003 series (China Zhong-Ke Micro-nano IOT Ltd). Humidity sensing measurements were carried out in the ambient atmosphere and in environmental control chamber (Weiss-voetsch Environmental Testing instruments, Tai Cang Co., Ltd), which supplied controlled temperature and relative humidity (RH). The ambient atmosphere showed a humidity of 65% RH and a temperature of 20 °C. And the environmental control chamber applied with varying RH (30–90%) at a constant temperature (20 °C), and the real-time impedance was recorded by the sensor testing system.

3. Results and discussion

The representative synthesis process of the SnO₂@G–GO nanocomposites for humidity sensing is illustrated in Fig. 1. The PVP/SnCl₂/graphene precursor solution was firstly electrospun to form nanofibers. And then followed by annealing, SnO₂@G hybrid NFs were obtained. Finally, the SnO₂@G hybrid NFs were gradually wrapped by the stretched GO during the solution evaporation process. And the SnO₂@G–GO nanocomposites were then fabricated.

Fig. 1 Schematic illustration of the synthesis route for SnO₂@G–GO nanocomposites. (a) As electrospun nanofibers; (b) SnO₂@G NFs; (c) SnO₂@G–GO nanocomposites.

The successful preparation of electrospun SnO₂ NFs and SnO₂@G hybrid NFs were firstly confirmed by XRD (Fig. S3†). The microstructures and morphologies of the SnO₂@G were characterized by SEM, TEM and HRTEM (Fig. 2A and B and S4,†).

Fig. 2A shows TEM images of SnO₂@G NFs. It illustrates that the SnO₂@G NFs are well dispersed, with comparatively narrow diameter distribution, centered at 100–200 nm. The average pore diameter of the SnO₂@G NFs is about 3.8 nm, as determined by the BJH method. In addition, the nitrogen adsorption and desorption isotherms measurements showed that the specific surface area of SnO₂@G NFs and SnO₂@G–GO nanocomposites were 29.34, 193.62 m² g⁻¹, respectively (Fig. S5†). From the inset of Fig. 2A, it is clear that graphene is embedded uniformly in the SnO₂ NFs, indicating that the electrospinning is an effective method to synthesize SnO₂@G NFs nanostructures.

Fig. 2 (A) TEM images of SnO₂@G hybrid NFs. The graphene was embedded in single SnO₂ NF. (B) HRTEM images of SnO₂@G NFs. (C) TEM and (D) SEM images of SnO₂@G–GO nanocomposites.

Fig. 2B is the HRTEM image of SnO₂@G NFs. As shown on the top part of the figure, the interplanar distance of the crystalline material is 0.34 nm, which corresponds well to the lattice spacing of the (110) plane of SnO₂. Fig. 2C shows the TEM images of SnO₂@G–GO nanocomposites. Fig. 2D and Fig. S6† are SEM images of SnO₂@G–GO nanocomposites. It is clear that the SnO₂@G hybrid NFs are wrapped by well-stretched GO. Fig S7† shows the Raman spectra of SnO₂@G NFs and SnO₂@G–GO nanocomposites, which confirms the existence of GO. Fig. S8† shows the XPS spectra of SnO₂@G–GO nanocomposites. The peaks of tin (Sn 3p, 3d, 4s, 4p, 4d) emerge, which are expected from SnO₂, while the peak of C 1s is attributed to graphene sheets and GO.

The humidity sensing tests were carried out with interdigital gold electrodes and an environmental control chamber. Fig. 3A shows macrograph and SEM of interdigital gold electrodes with the sensing material. It is clear that sensing material was evenly coated on the surface of interdigital gold electrode. Dynamic testing procedures were carried out, which provided information on three important parameters for a sensing device: sensitivity, response and recovery time, stability.²¹

Fig. 3 (A) The macrograph and SEM of interdigital electrodes with the sensing materials. (B) The response and recovery behavior for the sensors. (C) The impedance response of the SnO₂@G–GO based sensor between ambient atmosphere (65%) and different RH (30–90%). (D) The dynamic impedance response of the SnO₂@G NFs based humidity sensor.

As shown in Fig. 3B, the response and recovery behaviors of four sensors based on industrial SnO₂ powder, pure electrospun SnO₂ NFs, SnO₂@G hybrid NFs and SnO₂@G–GO nanocomposites were measured by switching the sensors from ambient atmosphere (20 °C, 65% RH) to the environmental chamber (20 °C, 30% RH), respectively. Response and recovery time is defined as the time required to reach 90% of the final equilibrium value.^21,28 The response and recovery behavior of each sensor was normalized to the impedance values measured at low humidity (30% RH) and ambient atmosphere (65% RH): I_30%/I_65%. As shown in Fig. 3B and the inset, the ratio of impedance at 30% RH and impedance at 65% RH of four different sensors changed over time, respectively.

Both industrial SnO₂ powder and electrospun SnO₂ NFs based sensors took long time (more than 10 s) to response. As comparison, the response and recovery time of SnO₂@G hybrid NFs based sensors were measured to be 2 s and 4 s, respectively. This is due to the fact that response and recovery process was directly associated with the water adsorption–desorption, and the regeneration was likely to have higher activation energy.^32,33 Obviously, the addition of graphene significantly reduced the response and recovery time. It was reported that water molecules adsorbed on graphene would form water clusters through hydrogen bond, and the water cluster link was a donor.³¹ Therefore, water molecules adsorbed on graphene could increase the conductivity. As for SnO₂@G–GO nanocomposites based sensor, the response and recovery behavior was further improved. The response and recovery time of SnO₂@G–GO based sensor were even shorter (less than 1 s). This is because that the rich hydrophilic groups on the surface of GO made it easy for the nanocomposites to obtain water molecules from the ambient.¹³

Further dynamic testing of the sensor based on SnO₂@G–GO nanocomposites was carried out by applying a humidity pulse between ambient atmosphere (20 °C, 65% RH) and different RH (30, 40, 50, 60, 70, 80, and 90% RH) in the environmental chamber. In the experiment, the sensor was firstly exposed to ambient atmosphere, of which the initial stage was recorded as a baseline.

Fig. 3C shows a series of transient impedance response of the as-prepared sensors to switch from ambient RH (65% RH) to different RH (30–90% RH) at ambient temperature (20 °C). When the device switched from 65% RH in the ambient atmosphere to 30% RH in environmental chamber, the impedance promptly increased, and then gradually reached a relatively stable value. The device then was switched into the ambient and the impedance fell rapidly. The measurement was performed repeatedly. During the three cycles at different humidity, the sensor showed repeatable response, indicating good stability. The sensitivity of the sensor can be calculated according to the following formula:^34,35

Sensitivity = Δ(impedance)/Δ% RH

where Δ(impedance) is different value between the instant measuring impedance and the baseline impedance, and Δ% RH is the corresponding different value of the RH.

In Fig. 3C, the measured impedance of SnO₂@G–GO based sensor at different RH (30, 40, 50, 60, 70, 80 and 90%) is (1075, 1990), (1065, 1750), (1035, 1400), (1010, 1120), (970, 825), (960, 575), and (955, 380) (MΩ, MΩ), respectively. And the corresponding sensitivity at 30, 40, 50, 60, 70, 80, and 90% RH is 26.1, 25.4, 24.3, 22, 32, 26.7 and 23 MΩ/% RH. Overall, the impedance of SnO₂@G–GO based sensor decreased with the increase of RH. In the lower range (<65% RH) or in the higher range (>65% RH), the sensitivity decreased as the RH increasing. The sensitivity can reach highest value when the sensors switched from ambient air to 70% RH, which is 32 MΩ/% RH. The high sensitivity enabled facile practical application.

Fig. 3D demonstrates the impedance response of the SnO₂@G hybrid NFs without being wrapped by GO to dynamic switch between ambient atmosphere (20 °C, 65% RH) and different RH (30, 40, 50, 60, 70, 80 and 90%) in the environmental chamber. In Fig. 3D, the response of the SnO₂@G hybrid NFs based sensor at 30% and 40% RH was relative slow and hard to reach stable values. Also, it is clear that the SnO₂@G NFs based sensor showed little impedance change at 80, 90% RH when compared with that of 70% RH, indicating that the adsorption of moisture on SnO₂@G NFs would approach saturation at 70% RH. As the RH further increased (from 70% to 90% RH), the moisture absorption remained almost the same while it took much longer time for SnO₂@G based sensor to approach the basal level. This is because prolonged exposure to humid environment caused the gradual formation of stable chemisorption on the surface, which led to a progressive drift in the impedance of the SnO₂@G based sensor.³⁶

When compared with SnO₂@G NFs based sensor, at low RH (30%, 40% RH), SnO₂@G–GO nanocomposites based sensor responded much faster. This is due to the fact that the large amount of hydrophilic groups on GO were easily accessible and were ready to absorb water molecules from the ambient. At relative high RH (from 70% to 90%), SnO₂@G NFs based sensor had reached a saturation while SnO₂@G–GO based sensor still exhibited high sensitivity. The larger surface to volume ratio of the SnO₂@G–GO nanocomposites enhanced diffusion rate of water molecules into or out-off the porous nanostructure; and therefore the reaction would be efficient, which led to larger change of impedance. These results further confirmed that GO played a critical role in this humidity sensing process.

Based on previous studies and the analysis of the experimental data, the possible humidity sensing mechanism of the SnO₂@G–GO nanocomposites was proposed as the following. Because of the low operating temperature (20 °C), chemisorption could hardly happen.^36,37

When the SnO₂@G–GO based sensor was exposed to dry air, some O₂ molecules would adsorb at the interface and grain boundaries of porous SnO₂@G NFs. Adsorbed oxygen molecules captured free electrons from n-type SnO₂ to form O₂.^33,34,37 At low RH, when the sensor was exposed to humid air, water molecules replaced O₂⁻ and physisorbed on the active sites of the SnO₂@G NFs, with molecular form. In order to achieve charge balance, electrons were released from O₂⁻ (Fig. 4A).^36–38 Electrons were attracted to the surface of SnO₂@G NFs by the preferential alignment of water dipoles.^33,36 At the same time, water molecules formed clusters on graphene via hydrogen bonds and cluster link acted as a donor.³¹ As result, water molecule adsorption on graphene enhanced the conductivity of the nanocomposites. Improved electron conduction caused more water molecules to participate the adsorption–desorption process, resulting in a large change of impedance.

Fig. 4 Schematic illustration of processes of the possible humidity sensing mechanism of the as-prepared SnO₂@G–GO nanocomposites based sensors, (A) at 30–40% RH, (B) at 50–60% RH, (C) at 70–90% RH.

On the other hand, since oxygen-rich groups (hydroxyl and epoxy groups) caused GO to have strong hydrophilicity, well-stretched GO could easily capture water molecules from the ambient.³⁶ This contributed greatly to the fast response and recovery behavior of the humidity sensing. With an increase in RH, subsequent layers of water molecules were physisorbed onto the first water layer, attaching with hydrogen double bonds through hydroxyl groups (Fig. 4B). This process caused an increase of spacing among the interlayer of GO sheets,^31,40,41 which could be sufficient to accommodate water molecules to form a water layer. The formation of a liquid water layer provided a conduction path between GO films and electrode.^39,42,43 At relative high RH (70–90% RH), the physisorption of water molecules onto SnO₂@G NFs reached saturation. And then the GO started to dominate the humidity sensing process. With further increase in RH, a water molecule could form a single bond to a hydroxyl group, and the ions might have more freedom to move through the water layer (Fig. 4C). These ions transported charges between physisorbed water molecules on the surface of SnO₂@G NFs and those on the surface of GO. Ions generated via the reaction of water molecules with the GO functional groups on the surface led to the decrease in the impedance.⁴⁴ This process was identical to the conduction in pure water. Aside from the fact that ionic conduction in adsorbed water layers, the ionization of carboxyl groups of GO would also greatly contribute to ionic conduction at relative high RH.^39,44 All these processes enhanced the humidity sensing capability of SnO₂@G–GO nanocomposites.

4. Conclusions

In summary, we have synthesized the nanocomposites of SnO₂@G hybrid NFs wrapped by GO, through a simple and well-controlled physical method. This approach is non-toxic and ready for massive production at large scale. The as-prepared SnO₂@G–GO nanocomposites based humidity sensors demonstrated enhanced humidity sensing properties: high sensitivity, fast response and recovery, good repeatability. Moreover, the experiment was performed at room temperature (20 °C) and ambient humidity (65% RH), the as-prepared sensors could still deliver optimal sensing behavior, which is important in practical applications and could meet the demand of the reduction of energy consumption in sensor devices.

Acknowledgements

B. Lu thanks to financially supported by National Natural Science Foundation of China (No. 21303046), China Scholarship Council (File No. 201308430178), Hunan University Fund for Multidisciplinary Developing (No. 531107040762).

Notes and references

1. Y. Noboru and S. Yasuhiro, Sens. Actuators, 1986, 10, 379.
2. C. Y. Lee and G. B. Lee, Sens. Lett., 2005, 3, 1.
3. Z. Chen and C. Lu, Sens. Lett., 2005, 3, 274.
4. P. L. Kian, Q. L. Bao, K. A. Priscilla and J. X. Yang, J. Mater. Chem., 2010, 20, 2277.
5. S. Stankovich, D. A. Dikin, G. H. B. Dommett, K. M. Kohlhaas, E. J. Zimney, E. A. Stach, R. D. Piner, S. T. Nguyen and R. S. Ruoff, Nature, 2006, 442, 282.
6. A. B. Alexander, G. Suchismita, W. Z. Bao, C. Irene, T. Desalegne, F. Miao and C. N. Lau, Nano Lett., 2008, 8, 902.
7. Z. S. Bonaccorso, T. Hasan and A. C. Ferrari, Nat. Photonics, 2010, 4, 611.
8. D. Xu, S. Ivan, B. Anthony and Y. A. Eva, Nat. Nanotechnol., 2008, 3, 491.
9. X. F. Feng, S. Maier and M. Salmeron, J. Am. Chem. Soc., 2012, 134, 5662.
10. F. Yavari, C. Kritzinger, C. Gaire, S. Li, H. Gulapalli, T. Borca-Tasciuc, P. M. Ajayan and N. Koratkar, Small, 2010, 6, 2535.
11. D. F. Jesse, J. A. Matthew, C. T. Vincent, Y. Yang, B. K. Richard and H. W. Bruce, ACS Nano, 2009, 3, 301.
12. A. K. Geim, Science, 2009, 324, 1530.
13. R. D. Daniel, P. Sungjin, W. B. Christopher and S. R. Rodney, Chem. Soc. Rev., 2010, 39, 228.
14. G. H. Lu, L. E. Ocola and J. H. Chen, Nanotechnology, 2009, 20, 445502.
15. G. N. Cristina, R. T. Weitz, M. B. Alexander, S. Matteo, M. Alf, B. Marko and K. Klaus, Nano Lett., 2007, 7, 3499.
16. Y. W. Zhu, M. Shanthi, W. W. Cai, X. S. Li, W. S. Ji, R. P. Jeffrey and S. R. Rodney, Adv. Mater., 2010, 22, 3906.
17. B. A. Lu, T. Li, H. T. Zhao, X. D. Li, C. T. Gao, S. X. Zhang and E. Q. Xie, Nanoscale, 2012, 4, 2978.
18. B. A. Lu, Y. J. Wang, Y. X. Liu, H. G. Duan, J. Y. Zhou, Z. X. Zhang, Y. Q. Wang, X. D. Li, W. Wang, W. Lan and E. Q. Xie, Small, 2010, 6, 1612.
19. B. A. Lu, Y. M. He, H. G. Duan, Y. J. Zhang, X. D. Li, C. Q. Zhua and E. Q. Xie, Nanoscale, 2012, 4, 1003.
20. S. S. Wu, H. Q. Cao, S. F. Yin, X. W. Liu and X. R. Zhang, J. Phys. Chem. C, 2009, 113, 17893.
21. J. Zhu, G. H. Zhang, X. Z. Yu, Q. H. Li, B. A. Lu and Z. Xu, Nano Energy, 2014, 3, 80.
22. T. Yang and B. A. Lu, Phys. Chem. Chem. Phys., 2014, 16, 4115.
23. J. Zhu, D. N. Lei, G. H. Zhang, Q. H. Li, B. A. Lu and T. H. Wang, Nanoscale, 2013, 5, 5499.
24. Q. Kuang, C. S. Lao, Z. L. Wang, Z. X. Xie and L. S. Zheng, J. Am. Ceram. Soc., 2007, 129, 6070.
25. Y. Liu, K. Erik and M. L. Liu, Chem. Mater., 2005, 17, 3997.
26. Y. J. Chen, L. Nie, X. Y. Xue, Y. G. Wang and T. H. Wang, Appl. Phys. Lett., 2006, 88, 083105.
27. K. Sungjin, H. K. Sook, Y. L. Seong, H. K. Jae and B. P. Chan, Adv. Mater., 2011, 23, 2009.
28. C. Cheng and D. Li, Adv. Mater., 2013, 25, 13.
29. D. J. Xue, S. Xi, Y. Yan, K. C. Jiang, Y. X. Yin, Y. G. Guo and L. J. Wan, J. Am. Chem. Soc., 2012, 134, 2512.
30. C. N. R. Rao, A. K. Sood, K. S. Subrahmanyam and A. Govindaraj, Angew. Chem., Int. Ed., 2009, 48, 7752.
31. J. Berashevich and T. Chakraborty, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 80, 033404.
32. Z. Imrana, S. S. Batoola, H. Jamila, M. Usmana, M. Israr-Qadir, S. H. Shah, S. Jamil-Rana, M. A. Rafiq, M. M. Hasan and M. Willander, Ceram. Int., 2013, 39, 457.
33. R. S. Niranjan, S. D. Sathaye and I. S. Mulla, Sens. Actuators, B, 2001, 81, 64.
34. M. K. Bernard, J. Am. Ceram. Soc., 1991, 74, 697.
35. N. K. Pandey and T. Karunesh, Sens. Transducers J., 2010, 122, 9.
36. T. Enrico, Sens. Actuators, B, 1995, 23, 135.
37. M. K. Bernard, J. Am. Ceram. Soc., 1991, 74, 697.
38. Z. Chen and C. Lu, Sens. Lett., 2005, 3, 274.
39. B. Stefano, W. Richard, W. Di, A. Michael, H. Samiul, S. Elisabetta, H. Nadine, K. Jani and R. Tapani, ACS Nano, 2013, 7, 11166.
40. Y. Yao, X. D. Chen, H. H. Guo, Z. Q. Wu and X. Y. Li, Sens. Actuators, B, 2012, 161, 1053.
41. R. Schaub, P. Thostrup, N. Lopez, E. Lægsgaard, I. Stensgaard, J. K. Nørskov and F. Besenbacher, Phys. Rev. Lett., 2001, 87, 2066104.
42. V. M. Nikhil, R. Ashwin, S. R. Rodney and B. S. Vivek, ACS Nano, 2010, 4, 2300.
43. Y. Yao, X. D. Chen, J. F. Zhu, B. Q. Zeng, Z. Q. Wu and X. Y. Li, Nanoscale Res. Lett., 2012, 7, 363.
44. Y. L. Guo, B. Wu, H. T. Liu, Y. Q. Ma, Y. Yang, J. Zheng, G. Yu and Y. Q. Liu, Adv. Mater., 2011, 23, 4626.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra10571d

---