A self-powered active hydrogen sensor based on a high-performance triboelectric nanogenerator using a wrinkle-micropatterned PDMS film

Abstract

A triboelectric nanogenerator powered room-temperature hydrogen (H₂) sensor was fabricated using ZnO nanorod (NR) arrays decorated with Pd nanoparticles (NPs) and a wrinkle-micropatterned polydimethylsiloxane (PDMS) nanogenerator. A generated open-circuit voltage of 16.2 V and short-circuit current of 0.512 μA were obtained when the device was exposed to a 5.3 N contact force at a fixed pressing frequency of 3 Hz. The instantaneous output power density from the device was 15.81 μW cm⁻² when connected to a load resistor of 1 MΩ. The triboelectric output of the as-fabricated device was attributed to the enhanced triboelectrification of the wrinkle-micropatterned PDMS and the synergistic interplay of Pd/ZnO heterojunctions, which effectively acted as both the energy source and H₂ sensing signal. Upon exposure to 3 vol% H₂ at room temperature under the same applied force and deformation frequency, the triboelectric output voltage of the device decreased from 16.2 V (in dry air) to 1.04 V, through which a response value (sensitivity) up to 1457.69% was obtained. The device also showed an excellent low limit of detection (20 ppm), but a relatively slow response–recovery time (115–126 s). These results suggest that the device can be used in practical applications and can stimulate new research for the development of next-generation portable self-powered active H₂ sensors.

---

Introduction

In recent years, there has been interest in harvesting ambient energy due to the large increases in world-wide energy consumption. Scavenging renewable energy from mechanical vibrations, heat, light, and chemicals have attracted global attention as they can reduce environmental pollution from carbon emissions, provide an uninterrupted and long-term energy supply, and make the world less dependent on costly fossil fuel.^1–4 Among these energy sources, mechanical energy is the most abundant in our environment. Energy harvesters based on electrostatic, piezoelectric, and electromagnetic effects have been developed to utilize ambient mechanical energy. Most recently, economic, simple, compact, and robust nanogenerators (NGs) have been developed by Wang et al. based on the triboelectric effect.^5–7

Typically, when two different (usually polymeric) materials with distinct surface electron affinities are brought into physical contact, electrostatic charges are induced on their surfaces due to the coupling of triboelectrification and electrostatic induction. A potential drop can be generated due to the contact induced triboelectric charges when the surfaces are separated by a mechanical force. Then, electrons flow between electrodes on the top and bottom surfaces of the materials to balance the potential difference. This is the fundamental working principle of a triboelectric nanogenerator (TENG). Since its discovery in 2012, TENGs have attracted a great deal of attention in harvesting energy from human motion, walking, vibrations, mechanical triggering, rotating tires, wind, flowing water, and potential applications in self-powered active sensors.^5–18 Recently, a number of triboelectric-based self-powered active mechanical^10–14 and chemical/gas sensors^15–18 have already been reported by various research groups with the objective of utilizing ambient environmental energy to power the device. Materials that have a strong triboelectrification effect are preferred due to their ability to capture the transferred charges and retain them for an extended period of time. Moreover, the configuration and design of the contact area of the materials play a vital role in enhancing the TENG and TENG-based sensor output, which can be achieved through intentionally generated micro-to-nano scale structures at the contact interface.^19–23 Usually, the structures are composed of densely packed pebbles, arrays, valleys, or hierarchical units, which can increase the effective contact area of a TENG. Recently, surface wrinkling on polydimethylsiloxane (PDMS) has received special attention in various applications due to its outstanding end product flexibility, stretchability, and the ability of patterning in large-scale; furthermore, microstructured PDMS can apparently improve the performance of TENGs.^9,19,24,25 Additionally, wrinkle-micropatterned PDMS surfaces exhibited a higher friction area, which preferentially enhanced the triboelectric output compared to flat-, micropyramid-, hierarchical-, and sponge-like polymer films.²⁶

Hydrogen (H₂) gas is widely used in various applications such as in industry, renewable energy, and fuel cells.²⁷ These applications require the efficient, accurate, and swift detection of hydrogen with high sensitivity and selectivity over a wide range of concentrations to protect the environment, ensure human safety, and prevent unexpected explosion and asphyxiation due to the leakage of H₂. Among the various reported H₂ sensing platforms, one-dimensional (1D) ZnO-based resistive H₂ sensors showed promising advancements, while Pt and Pd were considered as the best metal catalysts for enhancing H₂ sensing performance.^28–30 These applications now require a number of additional features such as low-cost device fabrication processes, and the devices must be light weight, flexible, and self-powered to enable smart, wearable, handheld and portable H₂ gas sensors. Rashid et al. reported a resistivity-type highly flexible H₂ sensor with a maximum response magnitude of 91% and a low limit of detection of 0.2 ppm.³¹ Fu et al. reported a 1D ZnO based piezoelectric self-powered active H₂ sensor with a maximum sensitivity of 446 towards 800 ppm gas concentration with a detection range from 10 ppm to 800 ppm.³² However, the need for an external power source still provides motivation for further research.

In the current work, we demonstrated a novel triboelectric nanogenerator-based self-powered active H₂ sensor. The as-fabricated device was composed of a layer of Pd nanoparticles (NPs) decorated on 1D ZnO nanorod (Pd NPs/ZnO NR) arrays (as a sensing surface) and a layer of wrinkle-micropatterned PDMS (as triboelectric nanogenerator) separated by a spacer. The sensing performance of the triboelectric nanogenerator-based H₂ gas sensor (TENG-HS) was evaluated in a contact-separation mode at various H₂ gas concentrations in terms of response magnitude and response time. We believe that the enhanced sensing performance demonstrated herein along with the simple sensor packaging may enable future advancement in self-powered active H₂ sensor research trends.

Experimental section

Micropatterned PDMS and device fabrication

The porous silicon (p-Si) mold was patterned using a metal assisted electroless chemical etching process. In a typical process, 30–50 nm silver nanoparticles (Ag NPs) were uniformly deposited on a 4 cm × 4 cm native-oxide etched p-type Si (100) wafer. After metallization, the wafer was etched for 15 min using the HF + H₂O₂ + H₂O etchant solution, followed by rinsing in a HNO₃ solution for 5 min to remove the unwanted Ag NPs. After cleaning with acetone and isopropanol, the p-Si was functionalized using a piranha solution (H₂SO₄ + H₂O₂) to crosslink a thin epoxy layer on the porous surface. Subsequently, the p-Si master was treated with trimethylchlorosilane (TMCS, Sigma Aldrich) by vacuum phase silanization to avoid adhesion between the PDMS and p-Si mold.

To prepare the micropatterned polymer film, PDMS elastomer and cross-linker (Sylgard 184, Tow Corning) were mixed in a 10 : 1 ratio (w/w). After degassing under vacuum, the elastomer mixture was casted on the p-Si master using spin-coating at 2400 rpm for 40 s (SCS 6800 spin coater, Specialty Coating Systems). After incubation at 80 °C for 2 h, the PDMS film was carefully peeled off from the p-Si mold, and laminated Al was placed as contact electrode on the back-side of the patterned PDMS surface. The sample was then sliced into 2 cm × 1.6 cm pieces and attached to a polyethylene terephthalate (PET) substrate. In addition, Pd NPs-decorated onto vertically aligned ZnO nanorod arrays were grown on Au-coated PET films (the detailed synthesis process of Pd/ZnO NRs is available elsewhere).³¹ Finally, both films were sealed with an adhesive spacer to ensure an adequate gap distance between the Pd/ZnO NRs and patterned PDMS films. To enable the charge transfer circuitry, copper conducting wires were connected to both electrodes using silver paste. The effective size of the final device was 1.85 cm × 1.6 cm, and the total thickness of the device was about 2.37 mm.

Characterization and sensor test

A JSM-6500F field emission scanning electron microscope (FESEM) was used to characterize the surface morphology of p-Si and the micropatterned PDMS film. The water droplet contact angles were measured using contact angle goniometry (Kruss DSA 100 easy drop) via the sessile drop method at room temperature. A DC servomotor (TS3738N4E11) was used to supply the contact force on the device. A functional oscilloscope (Lecroy Wave Runner 610) and a low noise output characterization system (Keithley SCS-4200, with a plug in preamplifier Keithley 4200-PA-1) were used to acquire the electrical signals.

Results and discussion

To increase the contact area and to facilitate the triboelectric output, a surface modified PDMS film was used to fabricate the TENG-HS instead of using a relatively flat polymer sheet. Here, we used a relatively economical, simple, and rapid formation process for patterning PDMS films as compared to the preparation of Si wafer molds by traditional photolithography and laser interference lithography-based etching or ultraviolet-ozone (UVO) treatment of micropatterned polymer films. FESEM analysis of the as-prepared p-Si and the resulting PDMS pattern arrays reveal the formation of high density wrinkle-shaped microstructures across the whole area of the base Si wafer mold as a result of the adopted fabrication procedure (Fig. 1). Fig. 1a and the inset of Fig. 1a show in-plane and cross-section micrographs of the p-Si template, respectively. The pore diameters were 250–350 nm, while the pore lengths were 700–800 nm. The shape and the density of the PDMS microstructures were well controlled by the initial porous patterns on the surface of the Si wafer mold. These patterns in the p-Si mold can be varied by varying the etching time. In order to obtain desired micropatterns with sufficient PDMS surface roughness, we prepared three kinds of p-Si wafer molds using etching times of 10 min, 15 min, and 20 min. We then fabricated the replica of the wrinkle-micropatterns of the PDMS surface. Among the fabricated PDMS micropatterns, the 15 min etched p-Si wafer mold assisted replica exhibited a higher surface roughness, dense and undistorted micropatterns, and was easy to peel-off of the template Si wafer. Fig. 1b shows the as-fabricated wrinkled-micropatterned PDMS (w-PDMS) films. This figure demonstrates that this procedure is an efficient method for preparing large-scale polymer microstructures. High density wrinkle patterns are beneficial for increasing the friction area and the ultimate power generation efficiency of the TENG. A significant advantage of the entire preparation process is that hundreds of replicas of the micropatterned PDMS films can be produced from one single mold and may be scaled-up for large-scale production and practical applications.

Fig. 1 In-plane FESEM micrographs of (a) p-Si wafer mold after 15 min etching process and (b) as-prepared w-PDMS. Inset of (a) shows the cross-section image of the p-Si wafer mold.

We performed contact angle measurements to determine the surface properties of the as-fabricated w-PDMS film. The contact angle (θ_w) of the flat PDMS and w-PDMS were measured to be 101° and 128°, respectively, as shown in Fig. 2. The higher contact angle for w-PDMS was attributed to the heterogeneity, roughness, or conformational disorder, which suggests a reduced surface energy of the w-PDMS compared to the flat PDMS film. It has been reported that surface roughness is directly proportional to contact angle and inversely proportional to surface energy.^33,34 The surface energy of the w-PDMS was calculated to be 0.0076 mN m⁻¹, which was much lower than the flat PDMS film (∼2.60 mN m⁻¹). This phenomenon demonstrates the higher surface roughness of the w-PDMS in comparison to the flat PDMS film.

Fig. 2 Water droplet contact angle images of (a) flat PDMS and (b) w-PDMS films.

On the other hand, though ZnO is not a tribo-series material, it exhibits triboelectric properties due to its finite conductivity characteristics.²⁵ In addition, a Pd NPs/ZnO NR sensing network is promising for H₂ detection and was fabricated on a PET substrate to allow for a flexible sensor interface. In fact, the as-fabricated device has two functions: (1) generation of triboelectric output and (2) measurement of the surface adsorbed H₂ gas molecules using triboelectric signals. The typical structure and an optical image of the TENG-HS are schematically shown in Fig. 3. Generally, the potential drop occurs in triboelectric devices in an alternating way through intermediate positions between pressed and released positions.²¹ The as-fabricated TENG-HS was operated in a contact-separation mode. This means that when Pd/ZnO/Au and PDMS/Al surfaces were brought into contact, localized charge transfer occurred due to contact electrification (or a triboelectric charging mechanism), and an electric potential difference was generated due to the surface separation. In the TENG-HS device, the wrinkled micropatterns on the PDMS surface played a vital role in enhancing the contact surface, resulting in a large triboelectric charge density due to the compressional deformation between the films. During the triboelectric event, PDMS captures electrons and is negatively charged, whereas Pd/ZnO donates electrons and is positively charged, forming a dipole state with a zero net charge. When the surfaces separated due to elastic force, the positive and negative charges were separated. Then, the negatively charged PDMS surface repelled electrons from the bottom Al electrode to the top Au electrode through an external load based on the coulombic force to screen the positive triboelectric charges on the Pd/ZnO film until an electrostatic equilibrium was reached. At the same time, an instantaneous voltage was generated on the load resistance during the electron transfer process. After complete separation, the direction was reversed because the electric charges moved to provide neutralization. More importantly, the contact process was faster than the separation process. Hence, a relatively higher amount of instantaneous power was generated when the surfaces were in contact; however, the net charge transfer during the contact and separation process was the same.

Fig. 3 (a) Schematic diagram and (b) photographic image of the as-fabricated TENG-HS device.

Fig. 4a shows the device functionality at different external pressing frequencies from 0.5 to 5 Hz, which revealed that a reliable output voltage could be obtained from the as-fabricated TENG-HS. As the frequency was increased from 0.5 to 3 Hz, the output voltage also increased gradually. However, a dense and slight increase in output voltage signal was observed at 5 Hz. This phenomenon might be attributed to the pushing force that was subsequently applied on the device before the separation of the layers, thus impeding the smooth recovery. A typical electrical output power from the device was extracted by measuring the voltage across the load resistors ranging from 0.01 to 10 MΩ (0.01, 0.05, 0.1, 0.5, 1, 2, 4, 5, 8, and 10 MΩ) as shown in Fig. 4b. The voltage across the load follows an increasing trend with increasing load resistance. The effective electrical power density of the device was closely related to the external load and reached a maximum value of 15.81 μW cm⁻² (∼18.84 μW cm⁻² from V_OC and I_SC) at a load resistance of 1 MΩ.

Fig. 4 (a) Measured output voltage of the TENG-HS under different external pressing frequencies (0.5–5 Hz) and (b) dependence of the power density output on external load resistance (0.01–10 MΩ).

For gas sensing characterization, TENG-HS was placed in a fully sealed lab-made gas chamber composed of two inlets for gas supply and vacuum, respectively, and one outlet for air. The gas mixture of H₂ and synthetic air was supplied inside the chamber at a constant flow rate of 50 sccm (standard cubic centimeters per minute) with different H₂ concentrations. The gas chamber was purged with synthetic air between each H₂ pulse for flushing. A contact force of around 5.3 N at a fixed frequency of 3 Hz was applied for contact and separation processes using the servomotor (at a constant rotation speed of 250 rpm).

Fig. 5 represents the open-circuit voltage (V_OC) (triboelectric output) of TENG-HS at room temperature in dry air and various concentrations of H₂. In dry air, the peak-to-peak open-circuit voltage was found to be around 16.2 V (Fig. 5a and b), while the value decreased to 15.72 V (Fig. 5b) after 0.002 vol% H₂ exposure and gradually decreased with increasing H₂ concentrations. The measured output voltage in air was attributed mainly to the triboelectric charge generation using a w-PDMS film because triboelectric charges would rarely be generated by Pd/ZnO NRs. In addition, a small amount of piezoelectric charges would be generated when the Pd/ZnO NRs were bent and in contact with a w-PDMS surface. Conversely, during operation in dry air, the Pd/ZnO film did not acquire sufficient surface energy to screen the triboelectric effects; therefore, it showed a maximum output voltage. However, when the TENG-HS was exposed to H₂ at a certain concentration, H₂ molecules reacted with the chemisorbed oxygen species (O_n(ads)⁻) on the Pd/ZnO interface, released free electrons back to the conduction band of ZnO, and caused an enhanced surface charge density, which effectively screened the triboelectric effects leading to a voltage drop.^32,35,36 Moreover, this phenomenon was accelerated with increasing H₂ concentration, resulting in a further decrease in output voltages. Two single positive and negative signal peaks are shown in Fig. 5c. These are related to the triboelectric behavior and electric charge flow, respectively, at the moments of pressing and releasing. The interval time of the TENG-HS was defined as the time taken to generate positive and next negative peaks and was calculated to be 0.26 s, which was mainly attributed to the enhanced discrete surface and rigidness of w-PDMS and Pd/ZnO NRs films. The enlarged views of the triboelectric output voltages in dry air and various concentrations of H₂ are shown in Fig. 5d–k. Upon exposure to 0.01, 0.05, 0.1, 0.5, 1, 2, and 3 vol% H₂, the triboelectric output voltages of TENG-HS were about 13.1, 8.72, 5.74, 5.12, 4.24, 2.16, and 1.04 V, respectively.

Fig. 5 (a and b) Output voltage (V_OC) variation with and without H₂ exposure. (c) A single pressing and releasing signal peak of the TENG-HS. Enlarged output voltage signals of the TENG-HS: (d) in air and (e–k) under 0.01 to 3 vol% H₂ exposure, respectively.

The line graphs depicted in Fig. 6a and b show the dependency of the triboelectric output voltage and short-circuit current density (J_SC) on the H₂ concentrations, respectively. It is clearly observed that the as-fabricated device can effectively detect H₂ within the concentration range between 0.002 and 3 vol%. The inset of Fig. 6a depicts an enlarged view of the triboelectric output voltage behaviors of the device at lower H₂ concentrations, which revealed a possible detection limit of the device down to 0.002 vol%. The current density and the short-circuit output current (I_SC) (inset of Fig. 6b) also showed descending behavior similar to V_SC. In dry air, a J_SC value of 1.16 μA cm⁻² was obtained (I_SC = 0.51 μA), while the value was reduced to 0.21 μA cm⁻² (I_SC = 0.09 μA) when exposed to 3 vol% H₂. Fig. 6c represents the calculated response magnitude of the TENG-HS based on H₂ concentrations. The sensor response (or sensitivity) of the device against H₂ under the same applied force can be defined as follows:

Fig. 6 TENG-HS output properties in terms of (a) open-circuit voltage (V_OC), (b) short-circuit current density, and (c) response magnitude variations with respect to H₂ concentrations. Insets: (a and c) the magnified view of the output at lower H₂ concentrations and (b) short-circuit output current versus H₂ concentration curve.

Here, S denotes the sensor response in percentage, V_a and V_g represent the open-circuit voltage of the TEHS in air and at a certain concentration of H₂, respectively, while ΔV denotes the difference between V_a and V_g. A maximum response of about 1457.7% was obtained at 3 vol% H₂. The inset of Fig. 6c depicts a magnified view of the response behaviors of the device at lower H₂ concentrations, in which zero percent (0%) response denotes no response at 0 vol% H₂ (as no target gas was supplied).

The response time (τ_res) and recovery time (τ_rec) characteristics of the sensor at 1 vol% H₂ are shown in Fig. 7. The response–recovery time of the sensor was defined as the time to reach 90% of the total output voltage change. When the atmosphere of the test chamber was changed from dry air to H₂, the triboelectric output voltage decreased with time and then reached saturation within 115 s (response time, τ_res). Afterward, the test chamber was purged rapidly with dry air, resulting in the recovery of the initial baseline value within 126 s (recovery time, τ_rec). In addition, similar response–recovery time values were observed for 0.01 to 3 vol% H₂ concentrations. These relatively long response–recovery times can be attributed to the following factors: (1) delay times to acquire sufficient surface energy to screen the triboelectric effect, (2) delay times for surface polarization, and (3) disturbances in adsorption–desorption during the redox process due to frequent external forces. The relatively slow response–recovery time characteristics may restrict practical applications; hence, extensive work is required to improve the device performance by accelerating the response–recovery process.

Fig. 7 Response–recovery time characteristics of the TEHS to 1 vol% H₂ concentration.

Conclusions

In summary, we have, for the first time, successfully fabricated a triboelectric nanogenerator-based H₂ sensor by combining a uniformly grown Pd NPs/ZnO NRs array with a wrinkle patterned PDMS. The triboelectric output generated by the TENG-HS acted not only as a power source, but also as a response signal in terms of H₂ detection. A maximum output power density of 15.81 μW cm⁻² was obtained at a load resistance of 1 MΩ under a contact force of 5.3 N with a fixed pressing frequency of 3 Hz. In dry air, the device generated an open-circuit output voltage of 16.2 V and a short-circuit output current of 0.512 μA. Upon exposure to 1 vol% H₂, the device generated an open-circuit output voltage of 4.24 V and a short-circuit output current of 0.108 μA. Moreover, the device can effectively detect H₂ from 0.002 to 3 vol% at room temperature. The as-fabricated device exhibited a relatively slow response–recovery time (115–126 s). Even so, our successful demonstration illustrates the potential of this method for various applications and may inspire research in the field of portable self-powered active gas sensors for efficient detection of hazardous gases in complex environments.

Acknowledgements

This work was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF), funded in 2014 by the Ministry of Science, ICT and Future Planning (NRF-2014R1A2A2A01002668).

Notes and references

1. Z. L. Wang, Nano Res., 2008, 1, 1–8.
2. X. Wang, Nano Energy, 2012, 1, 13–24.
3. Y. Yang, W. Guo, K. C. Pradel, G. Zhu, Y. Zhou, Y. Zhang, Y. Hu, L. Lin and Z. L. Wang, Nano Lett., 2012, 12, 2833–2838.
4. C. Xu and Z. L. Wang, Adv. Mater., 2011, 23, 873–877.
5. W. Yang, J. Chen, G. Zhu, X. Wen, P. Bai, Y. Su, Y. Lin and Z. L. Wang, Nano Res., 2013, 6, 880–886.
6. W. Yang, J. Chen, G. Zhu, J. Yang, P. Bai, Y. Su, Q. Jing, X. Cao and Z. L. Wang, ACS Nano, 2013, 7, 11317–11324.
7. J. Yang, Y. Yang, J. Chen, H. L. Zhang, W. Yang, P. Bai, Y. Su and Z. L. Wang, Adv. Energy Mater., 2014, 4, 1301322–1301330.
8. Z. H. Lin, G. Cheng, L. Lin, S. Lee and Z. L. Wang, Angew. Chem., Int. Ed., 2013, 52, 12545–12549.
9. S. Wang, L. Lin and Z. L. Wang, Nano Energy, 2015, 11, 436–462.
10. J. Chen, G. Zhu, W. Yang, Q. Jing, P. Bai, Y. Yang, T.-C. Hou and Z. L. Wang, Adv. Mater., 2013, 25, 6094–6099.
11. Y. Yang, H. Zhang, Z.-H. Lin, Y. S. Zhou, Q. Jing, Y. Su, J. Yang, J. Chen, C. Hu and Z. L. Wang, ACS Nano, 2013, 7, 9213–9222.
12. B. Meng, W. Tang, Z.-H. Too, X. Zhang, M. Han, W. Liu and H. Zhang, Energy Environ. Sci., 2013, 6, 3235–3240.
13. L. Lin, Y. Xie, S. Wang, W. Wu, S. Niu, X. Wen and Z. L. Wang, ACS Nano, 2013, 7, 8266–8274.
14. H. Zhang, Y. Yang, Y. Su, J. Chen, K. Adams, S. Lee, C. Hu and Z. L. Wang, Adv. Funct. Mater., 2014, 24, 1401–1407.
15. Z.-H. Lin, G. Zhu, Y. S. Zhou, Y. Yang, P. Bai, J. Chen and Z. L. Wang, Angew. Chem., Int. Ed., 2013, 52, 5065–5069.
16. H. Zhang, Y. Yang, Y. Su, J. Chen, C. Hu, Z. Wu, Y. Liu, C. P. Wong, Y. Bando and Z. L. Wang, Nano Energy, 2013, 2, 693–701.
17. Z.-H. Lin, Y. Xie, Y. Yang, S. Wang, G. Zhu and Z. L. Wang, ACS Nano, 2013, 7, 4554–4560.
18. Z.-H. Lin, G. Cheng, W. Wu, K. C. Pradel and Z. L. Wang, ACS Nano, 2014, 8, 6440–6448.
19. M.-L. Seol, J.-H. Woo, D.-I. Lee, H. Im, J. Hur and Y.-K. Choi, Small, 2014, 10, 3887–3894.
20. X.-S. Zhang, M.-D. Han, R.-X. Wang, F.-Y. Zhu, Z.-H. Li, W. Wang and H.-X. Zhang, Nano Lett., 2013, 13, 1168–1172.
21. S. Wang, L. Lin and Z. L. Wang, Nano Lett., 2012, 12, 6339–6346.
22. F.-R. Fan, L. Lin, G. Zhu, W. Wu, R. Zhang and Z. L. Wang, Nano Lett., 2012, 12, 3109–3114.
23. K. Y. Lee, J. Chun, J.-H. Lee, K. N. Kim, N.-R. Kang, J.-Y. Kim, M. H. Kim, K.-S. Shin, M. K. Gupta, J. M. Baik and S.-W. Kim, Adv. Mater., 2014, 26, 5037–5042.
24. J. Luo, F. R. Fan, T. Zhou, W. Tang, F. Xue and Z. L. Wang, Extreme Mechanics Letters, 2015, 2, 28–36.
25. Y. H. Ko, G. Nagaraju, S. H. Lee and J. S. Yu, ACS Appl. Mater. Interfaces, 2014, 6, 6631–6637.
26. Z. L. Wang, J. Chen and L. Lin, Energy Environ. Sci., 2015, 8, 2250–2282.
27. T. Hubert, L. B. Brett, G. Black and U. Banach, Sens. Actuators, B, 2011, 157, 329–352.
28. M. Kashif, M. E. Ali, S. M. U. Ali, U. Hashim and S. B. A. Hamid, Nanoscale Res. Lett., 2013, 8, 68–76.
29. D.-T. Phan, A. S. M. I. Uddin and G.-S. Chung, Sens. Actuators, B, 2015, 220, 962–967.
30. D.-T. Phan and G.-S. Chung, Sens. Actuators, B, 2014, 199, 354–360.
31. T.-R. Rashid, D.-T. Phan and G.-S. Chung, Sens. Actuators, B, 2013, 185, 777–784.
32. Y. Fu, W. Zang, P. Wang, L. Xing, X. Xue and Y. Zhang, Nano Energy, 2014, 8, 34–43.
33. D. K. Owens and R. C. Wendt, J. Appl. Polym. Sci., 1969, 13, 1741–1747.
34. S. M. Chiu, S. J. Hwang, C. W. Chu and D. Gan, Thin Solid Films, 2006, 515, 285–292.
35. B. Yu, Y. Fu, P. Wang, Y. Zhao, L. Xing and X. Xue, Phys. Chem. Chem. Phys., 2015, 17, 10856–10860.
36. R. Khan, H.-W. Ra, J. T. Kim, W. S. Jang, D. Sharma and Y. H. Im, Sens. Actuators B, 2010, 150, 389–393.

---