Humidity-dependent piezoelectric output of Al–ZnO nanowire nanogenerator and its applications as a self-powered active humidity sensor

Abstract

Al-doped ZnO nanowire arrays are used to fabricate a piezoelectric nanogenerator, and its output is significantly dependent on the humidity in the environment, showing its applications as a self-powered active humidity sensor. The piezoelectric output of the device acts as both the power source and response signal to the humidity.

---

With the rapid development of functional nanodevices, investigating portable, small-size and sustainable power sources for driving these functional nanodevices is becoming more and more important. In the past several years, a new self-powered system that integrates energy generator and functional nanodevice has been proposed, aiming at harvesting energy from the environment to power the functional nanodevice, such as self-powered pH sensors, automobile speedometers, gas sensors and magnetic sensors.^1–4 In our previous work, by coupling the piezoelectric and gas sensing characteristics of ZnO nanowire (NW) arrays, an unpackaged ZnO NW piezo-nanogenerator (NG) as a self-powered active gas sensor has been firstly demonstrated,⁵ in which the piezoelectric output from ZnO NW acts as both the power source and gas sensing signal. After the establishment of this new piezo-gas sensing mechanism, a series of work has been done to improve the performance of the new device. For examples, noble metal catalyst modification on the surface of ZnO NWs can improve the sensitivity;⁶ introducing heterostructures can improve the selectivity.⁷ Although these methods have shown the possibility for high-performance piezo-gas sensing, it is still significant to search a low-cost and simple method for improving the piezo-gas sensing performance. For traditional resistance-type gas sensing materials, doping with proper element is a very effective way to enhance their sensing performance.^8,9 Thus it is highly expected that element-doped ZnO NW can improve the self-powered active gas sensing.

In this paper, Al-doped ZnO NW arrays have been used to fabricate NG as self-powered active humidity sensors, and high sensitivity has been achieved. The piezoelectric output of the device can act not only as a power source, but also as a very sensitive response signal to the relative humidity of the ambience. The self-powered humidity sensing performance of Al-doped ZnO nanoarrays is much higher than that of undoped ZnO nanoarrays. Such high performance can be attributed to the large amount of adsorption sites from the Al doping.

The vertically aligned Al-doped ZnO nanowire (NW) arrays were synthesized by a seed-assisted wet-chemical method. Firstly, a piece of flat Ti foil as the substrate was cleaned with water/alcohol and dried in a nitrogen stream. ZnO seed layer was then deposited on the Ti substrate by a wet-chemical-annealing method reported in our previous work.⁵ After that, 0.8 g of Zn(NO₃)₂·6H₂O and 0.0046 g of Al(NO₃)₃·9H₂O was dissolved in 50 ml of deionized water. After dissolved evenly, 0.3 g HMTA was added into the solution, stirring for 10 min at room temperature. Finally, the Ti foil coated with ZnO seeds was immersed into the above solution. The reaction autoclave was sealed and maintained at 93 °C for 24 h. After cooling down to room temperature, the Ti substrate coated with vertically-aligned Al-doped ZnO NW arrays was removed from the solution, washed with deionized water and ethanol, and dried at 60 °C.

Fig. 1a shows the final device structure of the self-powered active humidity sensor, which is composed of three major components: Al-doped ZnO NW arrays on Ti foil, Al layer and Kapton boards. Such a device structure is similar to a typical NG without packaging. Ti foil actes as both the substrate for Al-doped ZnO NW arrays and the conductive electrode. A piece of Al foil (thickness = 0.05 mm) positioned on the top of Al-doped ZnO NW arrays acts as the counter electrode. Two terminal copper leads are glued on the two electrodes with silver paste for electrical measurements, respectively. To ensure the stability of the device, the finished device is fixed between two sheets of Kapton boards as the frame.

Fig. 1 Fabrication process of the self-powered active humidity sensor based on Al-doped ZnO NW arrays. (a) Schematic diagram showing the structural design of self-powered active ethanol sensor. (b) Schematic image showing the device actively detecting humidity at room temperature. (c) The optical image of the flexible device.

The piezo-humidity sensing performance of the self-powered active sensor was studied by measuring the piezoelectric output against different relative humidity under constant applied compressive strain. The final device was connected to a low-noise preamplifier (Model SR560, Stanford Research Systems) for the measurement of the piezoelectric output voltage. The compressive force is applied to the device by a hammer hooked to a string connected to a stepper motor. The hammer was actuated by a stepper motor moved along a guide rail (the movement of the motor can be controlled by programming), as shown in Fig. 1b. Actually the area of the hammer (the diameter is about 3 cm) for applying the compressive force was larger than that of the device. The compressive force applied on the device is 34 N at the frequency of 2.7 Hz. Under externally applied compressive deformation, the piezoelectric output of the device can act as both the power source and the humidity sensing signal. All the measurements are conducted under the pressure of 1.01 × 10⁵ Pa. An optical image of the device is shown in Fig. 1c, indicating that the device is flexible and can be easily bent by human hand.

The morphology and microstructure of Al-doped ZnO NW arrays are characterized by scanning electron microscopy (SEM, Hitachi S4800) and transmission electron microscopy (TEM, JEOL JEM-2100F). Fig. 2a is the SEM image of Al-doped ZnO NW arrays on the top view. It can be seen that Al-doped ZnO NW arrays are densely grown on Ti substrate along a consistent growth direction. The average diameter of Al-doped ZnO NW arrays is about 100 nm. The cross-sectional-view of Al-doped ZnO NW arrays in Fig. 2b reveals that the average length of the nanowires is about 2 μm. It further confirms that Al-doped ZnO NWs are vertically aligned on the substrate. It should be noted that the morphology and diameter is dependent on the doping concentration of Al, as shown in Fig. S1.† As the doping concentration of Al increases, the diameter of the NWs increases. This result also confirms the dopant of Al element. From the TEM image of one single Al-doped ZnO NW (Fig. 2c), it can be seen that the whole surface of the NW is very smooth. The lattice fringes can be clearly seen in the high-resolution TEM (HRTEM) image (Fig. 2d). The lattice spacing of 0.52 nm, consistent with (001) crystal plane of wurtzite structural ZnO, indicates that the NW grows along [0001] direction. The select area electron diffraction (SAED) pattern (the inset of Fig. 2d) of the NW shows a good crystalline nature.

Fig. 2 (a) SEM image of Al-doped ZnO NW arrays in a top view. (b) SEM image of Al-doped ZnO NW arrays in a side view. (c) TEM image of one single Al-doped ZnO NW. (d) HRTEM and SAED pattern of the tip region of Al-doped ZnO NW. (e) EDS spectrum of Al-doped ZnO NW arrays. (f) XRD pattern of Al-doped ZnO NW arrays grown on Ti substrate.

Fig. 2e is the energy dispersive spectrometer (EDS) spectrum of Al-doped ZnO nanoarrays, showing the existence of four elements (O, Zn, Al and Ti) in this selected region. Similar EDS results have been obtained at other different areas, which confirm that Al are uniformly distributed in the whole system. By analyzing the EDS results, the atomic percentage of Al element is about 0.5%. The crystal phase of Al-doped ZnO nanoarrays was characterized by X-ray powder diffraction (XRD; D/max 2500 V, Cu Kα radiation, λ = 1.5405 Å). The XRD pattern of Al-doped ZnO NW arrays on Ti substrate is shown in Fig. 2f, and the sharp diffraction peaks indicate the good crystalline quality. The peaks marked by pentagram can be indexed to Ti (JCPDS file no. 44-1294) arising from the Ti foil substrate; the peaks marked by inverted triangle can be indexed to hexagonal wurtzite ZnO (JCPDS file no. 36-1451).

No other phase can be observed, which indicates that Al doping does not apparently change the crystal structures of ZnO NW. Meanwhile, no peaks corresponding to aluminium or aluminium compounds can be detected in the XRD patterns, confirming that Al doping exists in the form of impurity atoms.

Fig. 3 shows the room-temperature photoluminescence (PL) spectra of undoped ZnO and Al-doped ZnO NWs. The spectra of both the NWs mainly consist of a strong visible emission, a weak visible emission and a very weak UV emission. The strong emission, locating around ∼519 nm, usually results from the radiative recombination of a photo-generated hole with an electron occupying the oxygen vacancy on the surface of the NWs. The weak emission around ∼434 nm is the result of interstitial zinc in the ZnO nanowires.¹⁰ The very weak emission around ∼390 nm is attributed to band-edge emission of ZnO.¹¹ From the inset of Fig. 3, a blue shift of the band-edge emission of Al-doped ZnO NWs (374 nm) can be observed (undoped ZnO NWs: 390 nm). It is also believed to be Burstein–Moss effect¹² because of the doping of Al in the ZnO NWs. The enhanced intensity of the deep-level emission (∼519 nm) of Al-doped ZnO NWs confirm that a large amount of defects have been introduced on the surface of the NWs by Al substitution.¹¹

Fig. 3 PL spectra of the pure ZnO and Al-doped ZnO NW arrays.

Fig. 4a shows the piezoelectric output voltage of Al-doped ZnO NW arrays upon exposure to different relative humidity at room temperature under constant applied deformation. The compressive force in these measurements keeps the same (34 N, 2.7 Hz). The enlarged views of the piezoelectric output are shown in Fig. 4b–k. As the relative humidity is 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, and 60%, the piezoelectric output voltage of the device is 0.819, 0.763, 0.707, 0.611, 0.538, 0.460, 0.347, 0.146, 0.067 and 0.050 V, respectively. Fig. 4l is the relationship between the piezoelectric output voltage and the relative humidity of the devices based on undoped ZnO NW arrays and Al-doped ZnO NW arrays. The piezoelectric output of the devices is dependent on the outside relative humidity, and the piezoelectric voltage dramatically decreases as the relative humidity increases. Similar to the traditional definition of the sensitivity of resistance-type gas sensors (, where R_a and R_g represent the resistance of the sensor in dry air and in the test gas, respectively),¹³ the response R of the self-powered/active humidity sensor under the same deformation conditions can be simply presented as:

where V₀ and V_t are the piezoelectric output voltage of the device at 15% RH and the test relative humidity, respectively. Fig. 4m shows the response–concentration curves of Al-doped ZnO NW arrays (Al-doping concentration is 0.5%) and undoped ZnO NW arrays. It can be seen that the response of Al-doped ZnO NW arrays is very much higher than that of undoped ZnO NW arrays. The response R of Al-doped ZnO NW arrays as self-powered humidity sensor against 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, and 60% RH at room temperature is 7.36, 15.90, 34.00, 52.19, 78.24, 135.92, 460.64, 1115.28 and 1521.98, respectively. While the response R of undoped ZnO NW is merely 2.76, 11.54, 25.69, 50.98, 82.06, 107.64, 166.58, 227.06 and 337.52, respectively.

Fig. 4 (a) The piezoelectric output voltage of the device at different humidity under the same applied strain at room temperature. (b)–(k) are enlarged views of the piezoelectric output at 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55% and 60% RH, respectively. (l) The relationship between the piezoelectric output voltage and relative humidity of Al-doped ZnO and pure ZnO NW arrays. (m) The response of Al–ZnO NW arrays and pure ZnO NW arrays at different relative humidity.

The response–concentration curves of self-powered humidity sensors based on different Al-dopant is shown in Fig. 5a. It can be seen that the response of 0.5% Al-doping concentration is higher than that of 0%, 0.3% and 1% The optimum doped level of Al element in ZnO NW system is about 0.5%. Fig. 5b is the response–concentration curves of Al-doped ZnO NW arrays (0.5% Al-doping concentration) under different applied force, showing that the humidity response of sensor can be influenced by the applied force. Future theoretical work needs to be done on explaining this phenomenon.

Fig. 5 (a) The response of the device with different Al-dopant. (b) The response of Al-doped ZnO NW arrays (0.5% Al-doping concentration) under different applied force.

Such a highly-sensitive self-powered humidity sensing performance can be attributed to the coupling of the piezoelectric effect and the humidity sensing properties of Al-doped ZnO NWs. For undoped ZnO NWs, it is well known that they have a high density of point defects, such as oxygen vacancy, which provide n-type carriers (electrons) for their conductivity.^14,15 At the same time, ZnO NWs have high piezoelectric output under applied deformation. When the c-axis of ZnO NW is under external strain, a piezoelectric field can be created on the surface that can drive the electrons in the external circuit flowing forward and back (the output of NG); at the same time, the free electrons in ZnO NWs will transfer and partially screen this piezoelectric field and decrease the piezoelectric output (piezoelectric screening effect).¹⁶ Previous theoretical and experimental works have confirmed that the change in free-carrier density can affect the piezoelectric output of ZnO NW NG.¹⁷ Our previous work on the self-powered active gas sensing of ZnO NWs has shown that the electron concentration of ZnO NW arrays can be affected by the gas adsorption on the surface, which greatly changes the screening effect and thus affects the piezoelectric output of the device.⁵ In this paper, Al doping in ZnO NWs leads to the introduction of more oxygen vacancy-related defects in the NWs, which is confirmed by the enhanced intensity of the deep-level emission of Al–ZnO NWs. The surface of Al–ZnO NWs becomes highly active for reaction as more adsorption sites for water molecules are provided by these oxygen vacancies.¹⁸ Al doping can probably lead to the adsorption of a large amount of water molecules on the surface, which can result in conductive H₃O⁺, enhance the screening effect and decrease the piezoelectric output.

The detailed working mechanism of the self-powered active humidity sensor based on Al-doped ZnO NW arrays is shown in Fig. 6. Since the change in free-carrier density can affect the piezoelectric output of the NWs, the piezoelectric output of Al-doped ZnO nanoarray NG can act not only as a power source, but also as a response signal to humidity at room temperature. Small inositol of Zn²⁺ and Al³⁺ produces high local charge density and strong electrostatic field and represent good sites for chemisorption of water molecules.¹⁹ When the device is at low relative humidity without any applied force (Fig. 6a), water molecules quickly occupy the available sites under exposure to the atmosphere. Initially, water vapour is chemisorbed on the surface of the Al-doped ZnO NW arrays, and then hydroxyl groups can form on the surface. After the first layer of chemisorbed water forms, subsequent layers of water molecules are physically adsorbed. The physisorbed water dissociates into H₃O⁺ and OH⁻ ions because of the high electrostatic field in the chemisorbed layer. A charge transport by a Grotthuss chain reaction occurs when H₃O⁺ releases a proton to a neighboring water molecule which accepts it while releasing another proton. H₃O⁺ appears in the physisorbed water and serves as a charge carrier in H₂O–Al–ZnO NWs.²⁰ Under the compressive deformation, both the conductive H₃O⁺ in the water layer and the free electrons inside of the Al-doped ZnO NWs can have directional movement and screen the piezoelectric polarization charges in the NWs, thus the piezoelectric output voltage of the humidity sensor is lowered, as shown in Fig. 6b.

Fig. 6 The working mechanism of the self-powered active humidity sensor based on Al-doped ZnO NW arrays. (a) Schematic illustration showing the device based on Al-doped ZnO NW arrays at low relative humidity without applied force. (b) The piezoelectric output of the device under mechanical deformation at low relative humidity. (c) Al-doped ZnO NW without compression at high relative humidity. (d) The piezoelectric output of the device under mechanical deformation at high relative humidity.

At high relative humidity (Fig. 6c), continuous water adsorption on the material surface will give rise to physisorbed water multilayers, which are less affected by the underlying chemosorbed layer. Consequently, the protons will gain freedom to move randomly inside the physisorbed water multilayers according to the Grotthuss mechanism.²¹ When the device is under a compressive strain at high relative humidity (Fig. 6d), the high density of protons in the water layer, play a crucial role in decreasing the piezoelectric output of the device. Thus the piezoelectric output voltage of the device is significantly lowered.

In summary, highly-sensitive piezo-humidity sensing has been realized from Al-doped ZnO NW NG as self-powered active humidity sensors. The piezoelectric output of the device acted as both the power source and the response signal to the relative humidity of the ambience. The large amount of adsorption sites from the Al element doing resulted in a high density of conductive H₃O⁺, enhancing the piezoelectric screening effect and significantly decreasing the piezoelectric output. Our study demonstrated that introducing element dopant into the self-powered active gas sensor was a very effective way to enhance their piezo-humidity sensing performance.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (51102041 and 11104025), the Fundamental Research Funds for the Central Universities (N120205001 and N120405010), and Program for New Century Excellent Talents in University (NCET-13-0112).

Notes and references

1. S. Xu, Y. Qin, C. Xu, Y. G. Wei, R. S. Yang and Z. L. Wang, Nat. Nanotechnol., 2010, 5, 366.
2. R. Zhang, L. Lin, Q. S. Jing, W. Z. Wu, Y. Zhang, Z. X. Jiao, L. Yan, R. P. S. Han and Z. L. Wang, Energy Environ. Sci., 2012, 5, 8528.
3. Y. F. Hu, C. Xu, Y. Zhang, L. Lin, R. L. Snyder and Z. L. Wang, Adv. Mater., 2011, 23, 4068.
4. N. Y. Cui, W. W. Wu, Y. Zhao, S. Bai, L. X. Meng, Y. Qin and Z. L. Wang, Nano Lett., 2012, 12, 3701.
5. X. Y. Xue, Y. X. Nie, B. He, L. L. Xing, Y. Zhang and Z. L. Wang, Nanotechnology, 2013, 24, 225501.
6. Y. J. Lin, P. Deng, Y. X. Nie, Y. F. Hu, L. L. Xing, Y. Zhang and X. Y. Xue, Nanoscale, 2014, 6, 4604.
7. W. L. Zang, Y. X. Nie, D. Zhu, P. Deng, L. L. Xing and X. Y. Xue, J. Phys. Chem. C, 2014, 118, 9209.
8. X. Y. Xue, Y. J. Chen, Y. G. Liu, S. L. Shi, Y. G. Wang and T. H. Wang, Appl. Phys. Lett., 2006, 88, 201907.
9. L. C. Jia, W. P. Cai and H. Q. Wang, Appl. Phys. Lett., 2010, 96, 103115.
10. B. Djurišić, W. C. H. Choy, V. A. L. Roy, Y. H. Leung, C. Y. Kwong, K. W. Cheah, T. K. Gundu Rao, W. K. Chan, H. F. Lui and C. Surya, Adv. Funct. Mater., 2004, 14, 856.
11. X. Q. Meng, D. Z. Shen, J. Y. Zhang, D. X. Zhao, Y. M. Lu, L. Dong, Z. Z. Zhang, Y. C. Liu and X. W. Fan, Solid State Commun., 2005, 135, 179.
12. Z. Q. Xu, H. Deng, Y. Li and H. Cheng, Mater. Sci. Semicond. Process., 2006, 9, 132.
13. E. Comini, G. Faglia, G. Sberveglieri, Z. W. Pan and Z. L. Wang, Appl. Phys. Lett., 2002, 81, 1869.
14. M. W. Allen and S. M. Durbin, Appl. Phys. Lett., 2008, 92, 122110.
15. Q. J. Yu, C. L. Yu, J. Z. Wang, F. Y. Guo, S. Y. Gao, S. J. Jiao, H. T. Li, X. T. Zhang, X. Z. Wang, H. Gao, H. B. Yang and L. C. Zhao, RSC Adv., 2013, 3, 16619.
16. Y. F. Hu, L. Lin, Y. Zhang and Z. L. Wang, Adv. Mater., 2012, 24, 110.
17. F. Zhang, Y. Ding, Y. Zhang, X. L. Zhang and Z. L. Wang, ACS Nano, 2012, 6, 9229.
18. M. G. Zhao, X. C. Wang, J. P. Cheng, L. W. Zhang, J. F. Jia and X. J. Li, Curr. Appl. Phys., 2013, 13, 403.
19. W. P. Tai and J. H. Oh, J. Mater. Sci.: Mater. Electron., 2002, 13, 391.
20. X. Q. Fu, C. Wang, H. C. Yu, Y. G. Wang and T. H. Wang, Nanotechnology, 2007, 18, 145503.
21. A. Zainelabdin, G. Amin, S. Zaman, O. Nur, J. Lu, L. Hultman and M. Willander, J. Mater. Chem., 2012, 22, 11583.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra10216a

---