Fabrication of monodispersed hollow flower-like porous In₂O₃ nanostructures and their application as gas sensors

Abstract

The solvothermal method using L-lysine as a surfactant and ethylene glycol as an efficient auxiliary was used to prepare mono-dispersed, hollow, flower-like porous (HFP) In₂O₃. The morphology of the materials was characterized through scanning electron microscopy (SEM) and transmission electron microscopy (TEM). In addition the structure and composition were analysed using X-ray powder diffraction (XRD) and high-resolution transmission electron microscopy (HRTEM). The sensing performance of the In₂O₃ material for 500 ppb NO₂ was evaluated at near room temperature 40 °C. The results revealed that the HFP-In₂O₃ exhibited better gas-sensing properties, indicating its practical potential for detecting NO₂ at lower concentrations and temperatures.

---

1. Introduction

Indium oxide (In₂O₃) is an important direct wide-band-gap (E_g = 3.55–3.75 eV), n-type semiconductor material and has various applications. For instance it has been used in displays,^1,2 photocatalysis,^3–5 thermoelectrics,^6–8 solar cells,^9–12 and detectors.¹³ In gas-sensing detection applications the material has been attracting a great deal of attention. In₂O₃ gas sensitive properties are closely related to its morphology and structure. In₂O₃ nanostructures with various morphologies from 0D to 3D have been synthesized and examples include nanoparticles, nanofibers, thin-films, nanowires, nanosheets, and hollow microspheres. The synthetic strategies to fabricate these structures are diverse and include physical methods such as evaporation, physical vapour deposition (PVD), and magnetron sputtering. In addition chemical methods such as chemical vapor deposition (CVD), sol–gel,¹⁴ hydrothermal, and electrospinning^15–17 have been widely reported.^18–20 For the current study results, hierarchical structure, and especially hierarchical hollow structures has an obvious reinforced effect upon In₂O₃ gas sensitive properties. This is because hierarchical structures have a higher surface area that promotes mass transport and gas diffusion in the sensor material. The traditional synthetic strategy for hollow, hierarchical structures utilise hard and soft templates and are multi-purpose as a synthetic strategy among hydrothermal synthesis methods of hollow spheres. The hard template and soft template methods dominate the chemical synthesis methods for hollow spheres. Wenjea J. Tseng et al. synthesized mono-dispersed In₂O₃ hollow spheres using commercial polymeric beads as a colloidal hard template and applied this to photocatalysis and gas sensing.²¹ The soft template method uses polymeride or surfactants. For example, Benxia Li et al. synthesized In₂O₃ hollow microspheres via the “formamide-resorcinol-water” system.²² However, these methods all have their own advantages and disadvantages. On the one hand, uniform-sized hollow spheres can be obtained and on the other, the material morphology may be damaged and introduce impurities when using the hard template method. The soft template method has no disadvantages that are mentioned above but can lead to an inhomogeneous size. Therefore, it is even more important to develop a more reasonable strategy to synthesize uniform and monodispersed hollow spheres for gas sensor applications. A uniform size can work well for material structure and the performance of gas sensitive material that the material is not easy to reunite during the synthesis process, which can make the sensitive material pore richer and facilitate the gas diffusion.

In our work, we used the solvothermal method to get the desired monodispersed hollow flower-like porous (HFP) In₂O₃ nanostructures with L-lysine as the surfactant and ethylene glycol serving as an auxiliary. Fabrication of a gas sensor using these nanostructures allowed greater sensitivity and room temperature operation.

2. Experimental

2.1 Synthesis of materials

All chemical reagents (analytical-grade purity) used during the experiment were not modified or subject to further purification. In(NO)₃·4.5H₂O, absolute ethyl alcohol, L-lysine, and the ethylene glycol obtained from Sigma Aldrich, which are based in the USA. The synthetic method was as followed. In(NO)₃·4.5H₂O (0.195 g) and L-lysine (0.041 g) were successively dissolved in ethylene glycol whilst continuously stirring for 3 min. Next, anhydrous ethanol solution was added into the above transparent solution and stirring continued for a further 2 min. The volume ratio of anhydrous ethanol and ethylene glycol is 9 : 1. Finally, the 80% mixture was sealed in a Teflon-lined stainless-steel autoclave of 44 mL capacity and kept at 120 °C for 18 h. After heating it was allowed to cool down to room temperature naturally. After the hydrothermal procedure, the primrose brownness precipitates were collected by centrifugation (8000 rpm, 5 min), washed several times with distilled water and ethanol, and dried in air at 60 °C for 24 h. The sintering was performed by slowly heating to 400 °C at a rate of 1 °C min⁻¹. The hollow flower-like porous In₂O₃ (HFP-In₂O₃) structures were obtained after maintaining the precipitates at 400 °C for 2 h in an air atmosphere.

2.2 Characterization

The morphologies were directly investigated by SEM (SU70 Hitachi, Japan), at an accelerating voltage of 15 kV and TEM (FEI Tecnai TF20), working at 200 kV. The structures were measured using XRD analysis (D/max-2600 PC, Rigaku, Japan) with Cu Kα radiation (λ ≈ 1.5418 Å) over a 2θ range of 10–70°. HRTEM was also performed. The gas sensor characteristics were measured using a static test system.

2.3 Sensor fabrication

The gas sensor device includes two parts: a ceramic tube in which a pair of gold electrodes was covered and a Ni–Cr heating wire that was passed through the tube. The gas sensor device preparation of the hollow flower-like porous In₂O₃ nanostructures is as followed. The as-prepared powder was mixed together with deionized water producing a paste, which was placed onto the ceramic tube.

The gas sensor properties were influenced for external environmental conditions (10% RH), and measured by a static test system. The volume of the test system was 10 L. The response of the device is defined to be:

S = R_g/R_a for which test gases were oxidizing gases and R_a/R_g for reducing gases. Here, R_a and R_g represent the resistances of the sensor in the air and target gas, respectively. The response and recovery times of sensitive performance were defined as the time taken by the sensor to approach to 90% of the last stable resistance in the situation of adsorption and desorption, respectively.

3. Result and discussion

3.1 Structure and morphology

The fabricated products were measured using XRD and SEM techniques. XRD was performed to investigate the crystal phases of the as-fabricated indium oxide powders resulting from heat treatment (HT) of the precursors and sintering at 400 °C for 2 h (Fig. 1a). It is clear from the results that the as-prepared In₂O₃ is consistent with the standard data file (JCPDS file no. 71-2194). After calcification at 400 °C, the precursors were converted into a pure cubic structure of In₂O₃ with space group Ia3̄ (no. 206) and lattice parameters a = 10.117 Å with characteristic peaks at (2 2 2), (4 0 0) and (4 4 0). No diffraction peaks from any other impurities were observed, indicating the high purity of the products. This suggests that the precursor is being completely transformed into the In₂O₃ phase and the high intensity of the diffraction peaks in XRD pattern suggests that the sample is of high crystallinity.

Fig. 1 (a) XRD pattern and (b) morphological characterization of hollow flower-like porous In₂O₃ nanostructures after annealing the precursors at 400 °C for 2 h.

Fig. 1b shows a SEM image detailing the morphology of the as-prepared indium oxide (400 °C for 2 h). The low magnification SEM image (Fig. 1b) clearly displays flower-like In₂O₃ nanostructures with a uniform sphere size of ∼800 nm and are assembled with nanosheets that form a highly agglomerated structure. The inset shows a higher magnification image and indicates that the prepared In₂O₃ nanostructures are hollow and the nanosheets are constructed from a few nanometers of nanoparticles (5–10 nm). The morphology of the In₂O₃ was relatively unified, showing a uniform size of the HFP-In₂O₃.

More information can be found in Fig. 2. The hollow structures were characterized by TEM and shown in Fig. 2a. This results in an image where the edge is black and the center has a light color. TEM gives a more accurate size of the sphere at approximately 867 nm. The edge of the spherical shell can be seen to be made of nanosheets and some nanoparticles, which is consistent with results obtained from the SEM (Fig. 1b). Selecting nanosheets from the edge of the spherical shell and taking a further magnified image as shown in Fig. 2b reveals the nanoparticle (approximately ranged from 5–10 nm) make-up of the nanosheets. The high-resolution transmission electron microscopy (HRTEM) image (Fig. 2c) of the nanosheet shows fringe distances of 0.29 nm and 0.21 nm, corresponding to the lattice distances of the (222) and (332) planes of cubic In₂O₃. So, it can be clearly seen that nanosheets are made up of single crystal particles. The HRTEM (Fig. 2b and c) confirms that the porous In₂O₃ flower-like structures are polycrystalline in nature.

Fig. 2 (a) TEM image of HFP-In₂O₃ nanostructures. (b) High-magnification TEM images of the nanosheets. (c) HRTEM image of part of the nanosheets.

3.2 Materials synthesis mechanism

Fig. 3a shows that with no added L-lysine the material consists of 50 nm particles of relatively uniform size and no sheet structure. The product is a size-inhomogeneous aggregate and composed of nanosheets when the solution only has anhydrous ethanol with no ethylene glycol (Fig. 3b). This is consistent with the result of Kwon-Il Choi et al.²³ Fig. 3c shows that the materials are uniformly sized hollow spheres assembled from nanosheets and nanoparticles. By comparing Fig. 3a and b it can be seen that L-lysine plays a crucial role of connecting indium. Comparison of Fig. 3b and c reveals that ethylene glycol has dispersive effects upon the system. The probable mechanism is shown in Fig. 4. The positive NH³⁺ and negative COO⁻ may come from NH₂ and COOH radicals derived from lysine. Thanks to the electrostatic interaction, the negative end of the lysine absorbs the In³⁺ ion that constructed a unit subassembly and these units form sheets because of oriented absorption among ions as a result of irregular reunion (when adding no ethylene glycol). Growth and integration between other sheet layers results in disorderly reunion of sheets. This can be explored from a thermodynamic point of view.^24–26 The lower surface energy is more stable than the higher one and the surface energy is higher for single particles or sheets. Therefore the aggregate of particles and sheets decrease the surface energy to a stable condition. The ethylene glycol increases surface energy of units, which are composed of indium and lysine in order to influence their aggregate. This leads to the material growing to form relatively mono-disperse microsphere structures. Fig. 3c further shows that the ethylene glycol not only influences the aggregate of spheres but also the units in Fig. 3c. Both pieces and particles (they do not assemble into sheets) can be seen because of the property of ethylene glycol which seems to oil encase the water droplet and prevents it to grow.²⁷

Fig. 3 SEM images of various precursors: (a) no L-lysine; (b) no ethylene glycol; (c) anhydrous ethanol and ethylene glycol.

Fig. 4 The synthetic mechanism of the materials.

In addition, this HFP-In₂O₃ material synthesis mechanism also conforms to Ostwald ripening of porous secondary particles and the Kirkendall effect. The outer primary units will swallow up the inner ones through Ostwald ripening when the outer ones are bigger or packed in a thicker manner than those on the inner parts. Zeng²⁸ reviewed the main factors in the synthesis of hollow structures by Ostwald ripening. Meanwhile, during the oxidation of dense and crystalline indium particles, hollow structures can be formed by the Kirkendall effect when the outward diffusion of indium cations is faster versus the inward diffusion of oxygen to the metal core.^29–31 Generally, the development of the shell layers is mainly due to the Kirkendall effect rather than Ostwald ripening.

3.3 Gas sensing properties

The working temperature apparently influenced the response of a sensor material,^32–34 regulation of adsorption/desorption processes at equilibrium and the competition for chemisorption between the test gas and atmospheric oxygen O₂ for active surface sites. Thus, the response corresponding to the operating temperature curve was measured. Fig. 5a shows the variation of the response of the In₂O₃ sensors that were exposed to 500 ppb NO₂ gas at different operating temperatures. The response is confirmed to be very strongly dependent on the operating temperature. When the operating temperature is greater, the gas response value increases sharply, but as the operating temperature gets higher than 40 °C, the response slowly decreases. The maximum response value to NO₂ is approximately 50.5 at 40 °C. Therefore the optimal operating temperature of 40 °C was chosen for the sensor based HFP-In₂O₃ to further measure the characteristics of the gas sensors. The response result increases with temperature initially due to the acceleration of the surface reaction rate with operating temperature. However, in contrast very high temperatures may be detrimental since it may lead to low diffusion depth and therefore decrease the overall response to NO₂.^35–39 With the optimal temperature, the diffusion depth and reaction rate to NO₂ reaches a maximum.^40,41 The relationship between the total time for a complete test (response and recovery) and the working temperature was explored (Fig. 5b). The figure illustrates that faster response can be obtained by increasing the operating temperature because of the enhanced activation energies which is associated with an increased temperature. Moreover, Fig. 5a and b illustrates that the hollow structure is more suitable for working under lower temperature. This also can be illustrated in Table 1 which comparison in term of operating temperature, sensitivity and selectivity between results of the literature and the present work.^43–48

Fig. 5 (a) Correlations between the operating temperature and the gas response; (b) correlations between the operating temperature and operating time (response and recovery time at one cycle time).

Table 1 Comparison in term of operating temperature, sensitivity and selectivity between results of the literature and the present work (hierarchical: H_I; hollow: H_O)

Materials	Temp. (°C)	Sensor response	Selectivity	Con. (ppm)	Ref.
In2O3 (HI)	150	74	NO2	0.5	43
In2O3 (HO)	350	0.91	C3H6O	50	44
In2O3 (HI)	140	∼41.1	NO2	0.2	45
In2O3 (HO)	∼200	7	CO	1	46
WO3 (HIHO)	140	∼18	NO2	0.1	47
WO3 (HIHO)	100	89	NO2	1	48
In2O3 (HIHO)	40	50.5	NO2	0.5	Present work

---

The response transient characteristics of the HFP-In₂O₃ nanostructures sensors to 500 ppb NO₂ were investigated at a lower temperature of 40 °C (10% RH) (Fig. 6a). The response and recovery times to 500 ppb NO₂ are approximately 11 and 18 min, respectively. Fig. 6b shown the three times repeated response transients curves to 500 ppb for the sensor NO₂ at 40 °C and the three times response/recovery times were basically consistent indicating a repeatable and stable characteristic of sensor based on HFP-In₂O₃. The sensitivity behaviors were further investigated by exposing the sensor to different concentrations of NO₂ at room temperature (10% RH). The results were displayed in Fig. 6c. When the sensor was exposed to 10 ppb NO₂, the response was about 3. With increasing concentration of NO₂, the responses greatly increased and no saturation phenomenon. The responses were about 6.3, 15, 25, 50.5, 62, 73.2, 159 and 220 to 30, 80, 200, 500 ppb, 800 ppb, 1000 ppb, 3000 ppb and 5000 ppb NO₂, respectively.

Fig. 6 (a) Response transients resistance of the sensor. (b) The repeated response transients curves to 500 ppb for the sensor NO₂ at 40 °C. (c) The gas responses to different concentrations of NO₂ for the sensor using the HFP-In₂O₃ at 40 °C. (d) Comparison of responses of In₂O₃ sensors to various gases at 40 °C (10% RH).

Selectivity is a necessary parameter for gas sensors in both practical and research applications. Fig. 6d shows the cross-sensitivity of a sensor that is coated in flower-like In₂O₃ nanostructures. In this case the gases and volatile liquids had concentrations of 500 ppb and 100 ppm respectively. The molecules tested included CO, H₂, NH₃, C₂H₆O, CH₂O and C₃H₆O. All of the gasses were tested at an operating temperature of 40 °C. It is expected that the In₂O₃ sensor will exhibit a higher response to NO₂ among the operated gases and lower response to the other gases. The results did indeed show that the sensor using the HFP-In₂O₃ nanostructures exhibits an excellent selectivity to NO₂ and little to the other tested gases.

3.4 Gas sensing mechanism

The response of the sensor is a reflection of the change in the material resistance. The change of resistance for HFP-In₂O₃ sensors is due to the adsorption and desorption of test gas molecules on the surface of the materials. Its gas sensing mechanism can be explained using the modulation model of the depletion layer.⁴² At room temperature, sensors located in the air adsorb oxygen on the surface of material which, with the associated electrons of the material, form O₂⁻ (adsorbed oxygen).

When the sensor is exposed to an oxide gas such as NO₂, it takes the electrons from the adsorbed oxygen because the oxidizability of NO₂ is higher than that of oxygen. Furthermore, the test gas captures electrons from the conduction band of In₂O₃ and increases the material surface depletion layer thickness. Therefore the resistance rapidly increases during the adsorption process. When the In₂O₃-based sensor comes back to air gases, the NO₂ partial pressure is reduced and the NO₂ molecules desorb from the material leading to a reduction in the depletion layer thickness and the sensor returns to its original state. The desorption and absorption of test gases are accompanied by oxidation–reduction reactions, which require energy. Therefore is likely that a low working temperature may be the cause of the slower desorption. The surface of HFP-In₂O₃ is rich in nanoparticles. These nanoparticles possess nanosize quantum confinement effects that results in producing many active sites and these active sites could improve the sensitivity of materials. The HFP-In₂O₃ hierarchical nanostructures and the accumulation of nanosheets can form the pore passageways that allow the gas quickly access and break away from active sites on the inner and outer surface of the hollow spherical material. The abundant pore passageways are beneficial to NO₂ rapidly transmitting and the response speeded up. Therefore this special geometry is also helpful to in order to improve the gas sensing properties.

The morphology and structure of material can clearly affect the performance of gas sensitivity. The response of different In₂O₃ structures is shown in Fig. 7. The experimental investigation shows the sensing properties of materials at 500 ppb NO₂ and 40 °C with (a) no L-lysine (NL) and (b) no ethylene glycol (NEG) and (c) anhydrous ethanol/ethylene glycol (AE/EG). The response values of NL, NEG and AE/EG are 1.08, 7.5 and 50.5 respectively which correspond to a response time of 33 min, 32 min, 11 min respectively. The recovery times associated with NL, NEG and AE/EG are 82 min, 45 min, and 18 min respectively. The response of the materials with NL almost had no change because these particles aggregate to reduce the surface area. In addition the NEG system had a little response with its sheets structure increasing the surface area. The response of the AE/EG (HFP-In₂O₃) system has a greatly improved response due to its hollow structure with larger surface area, which can provide more active sites. The surface of these materials is composed of a sheet each of which is made up of particles with active sites providing many passageways and chances to gas sensitivity. In addition, it is obvious that a hollow structure allows efficient diffusion of gas into and out of the structure allowing a fast response. In Fig. 7 the HFP-In₂O₃ material has a higher response and shorter response/recovery time to NO₂.

Fig. 7 The normalized response with (a) NL; (b) NEG; (c) AE/EG.

4. Conclusion

Solvothermal synthesis is used to fabricate uniform size HFP-In₂O₃ with L-lysine, ethylene glycol and annealing of the precursors at 400 °C/2 h. Its gas properties were investigated through cross-comparison and showed that 500 ppb NO₂ at 40 °C with a response value of about 50.5 is far higher than other gases. The material exhibited a much higher selectivity to NO₂. This method of ethylene glycol assisting L-lysine at a relatively low temperature to prepare HFP-In₂O₃ offers an effective strategy to prepare uniform size hollow flower-like porous spheres.

Acknowledgements

This work was partially supported by the Natural Science Foundation of China (No. 61403110).

References

1. K. S. Park, W. Sato, G. Grause, T. Kameda and T. Yoshioka, Thermochim. Acta, 2009, 493, 105–108.
2. X. G. Xu, J. Chen, S. Z. Deng, N. S. Xu and J. Lin, J. Vac. Sci. Technol., B: Microelectron. Nanometer Struct.--Process., Meas., Phenom., 2010, 28, 490–494.
3. L. Y. Chen, Y. Liang and Z. D. Zhang, Eur. J. Inorg. Chem., 2009, 7, 903–909.
4. B. X. Li, Y. Xie, M. Jing, G. X. Rong, Y. C. Tang and G. Z. Zhang, Langmuir, 2006, 22, 9380–9385.
5. H. G. Sun, W. L. Fan, Y. L. Li, X. F. Cheng, P. Li, J. C. Hao and X. Zhao, Phys. Chem. Chem. Phys., 2011, 13, 1379–1385.
6. M. Ohtaki, D. Ogura, K. Eguchi and H. Arai, J. Mater. Chem., 1994, 4(5), 653–656.
7. W. Xu, Y. Liu, B. Chen, D. B. Liu, Y. H. Linc and A. Marcellide, Phys. Chem. Chem. Phys., 2013, 15, 17595–17600.
8. Y. Liu, W. Xu, D. B. Liu, M. Yu, Y. H. Lin and C. W. Nan, Phys. Chem. Chem. Phys., 2015, 17, 11229–11233.
9. T. O. L. Sunde, M. Lindgren, T. O. Mason, M. A. Einarsrud and T. Grande, Dalton Trans., 2014, 43, 9620–9632.
10. S. Calnan and A. N. Tiwari, Thin Solid Films, 2010, 518, 1839–1849.
11. R. A. Ismail, O. A. Abdulrazaq and K. Z. Yahya, Surf. Rev. Lett., 2005, 12, 515–518.
12. R. Sharma, R. S. Mane, S. K. Min and S. H. Han, J. Alloys Compd., 2009, 479, 840–843.
13. H. Meng, W. Yang, X. H. Yan, Y. G. Zhang, L. Feng and Y. F. Guan, Sens. Actuators, B, 2015, 216, 511–517.
14. X. S. Niu, H. X. Zhong, X. J. Wang and K. Jiang, Sens. Actuators, B, 2006, 115, 434–438.
15. Z. X. Cheng, X. B. Dong, Q. Y. Pan, J. C. Zhang and X. W. Dong, Mater. Lett., 2006, 60, 3137–3140.
16. A. Qurashi, E. M. El-Maghraby, T. Yamazaki and T. Kikuta, J. Alloys Compd., 2009, 480, 9–12.
17. T. Hyodo, H. Inoue, H. Motomur, K. Matsuo, T. Hashishin, J. Tamaki, Y. Shimizu and M. Egashira, Sens. Actuators, B, 2010, 151, 265–273.
18. C. Y. Wang, M. Ali, T. Kups, C. C. Rohlig, V. Cimalla, T. Stauden and O. Ambacher, Sens. Actuators, B, 2008, 130, 589–593.
19. T. Hyodo, H. Inoue, H. Motomur, K. Matsuo, T. Hashishin, J. Tamaki, Y. Shimizu and M. Egashira, Sens. Actuators, B, 2010, 151, 265–273.
20. M. Q. Huang, Z. D. Cui, X. J. Yang, S. L. Zhu, Z. Y. Li and Y. Q. Liang, RSC Adv., 2015, 5, 30038–30045.
21. W. J. Tseng, T. T. Tseng, H. M. Wu, Y. C. Her and T. J. Yang, J. Am. Ceram. Soc., 2013, 96, 719–725.
22. B. X. Li, Y. Xie, M. Jing, G. X. Rong, Y. C. Tang and G. Z. Zhang, Langmuir, 2006, 22, 9380–9385.
23. K. I. Choi, H. R. Kim and J. H. Lee, Sens. Actuators, B, 2009, 138, 497–503.
24. C. Li, D. H. Zhang, S. Han, X. L. Liu, T. Tang and C. W. Zhou, Adv. Mater., 2003, 15, 143–146.
25. J. Y. Liu, T. Luo, F. L. Meng, K. Qian, Y. T. Wan and J. H. Liu, J. Phys. Chem. C, 2010, 114, 4887–4894.
26. X. M. Xu, D. W. Wang, W. B. Wang, P. Sun, J. Ma, X. S. Liang, Y. F. Sun, Y. G. Ma and G. Y. Lu, Sens. Actuators, B, 2012, 171–172, 1066–1072.
27. Z. Q. Sun, T. Liao, Y. H. Dou, S. M. Hwang, M. S. Park, L. Jiang, J. H. Kim and S. X. Dou, Nat. Commun., 2014, 5, 3813.
28. H. C. Zeng, J. Mater. Chem., 2006, 16, 649–662.
29. Y. Yin, R. M. Rioux, C. K. Erdonmez, S. Hughes, G. A. Somorjai and A. P. Alivisatos, Science, 2004, 30, 711–714.
30. H. J. Fan, Y. Gosele and M. Zacharias, Small, 2007, 3, 1660–1671.
31. B. Liu and H. C. Zeng, J. Am. Ceram. Soc., 2004, 126, 16744–16746.
32. T. Hyodo, H. Inoue, H. Motomur, K. Matsuo, T. Hashishin, J. Tamaki, Y. Shimizu and M. Egashira, Sens. Actuators, B, 2010, 151, 265–273.
33. P. C. Xua, Z. X. Cheng, Q. Y. Pan, J. Q. Xu, Q. Xiang, W. J. Yu and Y. L. Chu, Sens. Actuators, B, 2008, 130, 802–808.
34. C. Q. Wang, D. Chen, X. L. Jiao and C. L. Chen, J. Phys. Chem. C, 2007, 111, 13398–13403.
35. N. Yamazoe, G. Sakai and K. Shimanoe, Catal. Surv. Asia, 2003, 7, 63–75.
36. G. Sakai, N. Matsunaga, K. Shimanoe and N. Yamazoe, Sens. Actuators, B, 2001, 2, 125–131.
37. A. M. Andringa, N. Vlietstra, E. C. P. Smits, M. J. Spijkman, H. L. Gomes, J. H. Klootwijk, P. W. M. Blom and D. M. Leeuw, Sens. Actuators, B, 2012, 171–172, 1172–1179.
38. N. Matsunaga, G. Sakai, K. Shimanoe and N. Yamazoe, Sens. Actuators, B, 2003, 96, 226–233.
39. N. Yamazoe and K. Shimanoe, Sens. Actuators, B, 2008, 128, 566–573.
40. P. Bogdanov, M. Ivanovskaya, E. Comini, G. Faglia and G. Sberveglieri, Sens. Actuators, B, 1999, 57, 153–158.
41. J. Ding, T. J. McAvoy, R. E. Cavicchi and S. Semancik, Sens. Actuators, B, 2001, 77, 597–631.
42. Z. H. Jing and J. H. Zhan, Adv. Mater., 2008, 20, 4547–4551.
43. X. M. Xu, P. L. Zhao, D. W. Wang, P. Sun, L. You, Y. F. Sun, X. S. Liang, F. M. Liu, H. Chen and G. Y. Lu, Sens. Actuators, B, 2013, 176, 405–412.
44. H. X. Dong, Y. Liu, G. H. Li, X. W. Wang, D. Xu, Z. H. Chen, T. Zhang, J. Wang and L. Zhang, Sens. Actuators, B, 2013, 178, 302–309.
45. X. M. Xu, X. Li, H. J. Zhang, C. H. Feng, C. Wang, F. M. Liu, Y. F. Sun, P. Sun and G. Y. Lu, RSC Adv., 2015, 5, 30297–30302.
46. A. Shanmugasundaram, P. Basak, S. V. Manorama, B. Krishna and S. Sanyadanam, ACS Appl. Mater. Interfaces, 2015, 7(14), 7679–7689.
47. J. W. Li, X. Liu, J. S. Cui and J. B. Sun, ACS Appl. Mater. Interfaces, 2015, 7, 10108–10114.
48. C. Wang, C. H. Feng, M. Wang, X. Li, P. F. Cheng, H. Zhang, Y. F. Sun, P. Sun and G. Y. Lu, RSC Adv., 2015, 5, 29698–29703.

---