Vertically aligned BaTiO₃ nanowire arrays for energy harvesting

Abstract

Nano-electromechanical systems (NEMS) developed using piezoelectric nanowires (NWs) have gained immense interest in energy harvesting applications as they are able to convert several different forms of mechanical energy sources into electric power and thereby function as reliable power sources for ultra-low power wireless electronics. In this work, a piezoelectric NEMS vibrational energy harvester is fabricated through the development of a synthesis process for vertically aligned barium titanate (BaTiO₃) nanowire (NW) arrays directly on a conductive substrate. These poled ferroelectric NW arrays are characterized through direct vibration excitation and demonstrated to provide efficient harvesting of mechanical vibrational energy producing an average power density of ∼6.27 μW cm⁻³ from 1g acceleration. In order to substantiate the superior energy harvesting performance of the newly developed BaTiO₃ NW arrays, a direct comparison is made with conventional ZnO NW arrays. Here, we clearly report that the ferroelectric BaTiO₃ NW NEMS energy harvester has ∼16 times greater power density than the ZnO NW NEMS energy harvester from the same acceleration input.

---

Broader context

The power generating capacity of piezoelectric energy harvesters composed of aligned ZnO nanowire arrays has been thoroughly investigated in the literature; however, ZnO has a lower electro-mechanical coupling coefficient as compared to several ferroelectric ceramics which limits ZnO's performance for energy harvesting. Amidst ferroelectric ceramics, the environmentally friendly lead-free ferroelectric BaTiO₃ can potentially be utilized to develop high performance energy harvesters. However, vertically aligned BaTiO₃ nanowire arrays had not been developed prior to this study for energy harvesting applications. Here, we report the synthesis method employed for obtaining novel vertically aligned ferroelectric BaTiO₃ nanowire arrays directly on conductive fluorine doped tin oxide (FTO) glass and demonstrate their efficient application in a nano-electromechanical system (NEMS) based energy harvester driven by base vibrations. The NEMS energy harvester is designed to have resonance below 1 kHz for efficient energy harvesting of ambient mechanical vibrations, which usually reside in the 1 Hz to 1 kHz range. It is observed that vertically aligned nanowire arrays composed of BaTiO₃ have 16 times higher power density than the widely used ZnO. This communication also compares the AC power and power density experimentally measured with respect to varying load resistors from NEMS energy harvesters.

Piezoelectric nanowires have attracted tremendous interest in the field of nanotechnology for energy applications as they possess excellent electro-mechanical energy conversion capability which remains intact even with their reduced scale. This scalability provides the ability for piezoelectric nanowires to function as a reliable alternative power source for emerging ultra-low power wireless sensor networks and micro-systems.^1–3 More recently, the creation of nano-electromechanical systems (NEMS) developed using piezoelectric nanowires has gained interest due to their ability to generate electric power by harnessing several different sources of mechanical energy such as sound waves,⁴ ultrasonic waves,^5–7 vibrational energy,^8,9 atomic force microscope tip induced stimuli,^10,11 and biomechanical energy.^12,13 In each of these energy harvesting applications, the ambient mechanical energy sources surrounding the piezoelectric nanowire based NEMS device are directly converted into useful electrical energy to provide power.^14,15 The power generating capacity of energy harvesting devices based on aligned ZnO nanowire (NW) arrays has been thoroughly investigated in the literature.¹⁶ These reports have shown that the mechanical energy conversion efficiency is high enough to produce sufficient electricity to potentially replace a battery and power small electronic devices;^17,18 however, ZnO has a lower electro-mechanical coupling coefficient as compared to many ferroelectric ceramics which may limit ZnO's performance for energy harvesting where the source is ambient mechanical vibrations in the environment.

Among NEMS, those made of ferroelectric perovskite nanostructures and thin films such as PZT (PbZr_xTi_1−xO₃)^9,19–22 and barium titanate (BaTiO₃)^23–25 can produce greater energy transfer due to their higher electro-mechanical coupling coefficients thus providing an efficient means to harvest mechanical energy. Additionally, environmental concerns over the toxicity of lead based piezoelectric materials have posed corresponding limitations over their future use in electronics enabling high performance lead-free BaTiO₃ to potentially serve as an alternative.²⁶ Previously, Wang et al.²⁷ performed a numerical analysis to show that the BaTiO₃ NWs have higher power generating capability and higher energy conversion efficiency as compared to ZnO NWs for the same size. The theoretical reason for this higher power generation is attributed to the higher piezoelectric coupling coefficient of BaTiO₃ NWs.²³ Furthermore, conventional solution grown ZnO NWs have high carrier density which may contribute to the risk of increased conductivity from the mobile free charge carriers (free electrons) as recently reported by Sohn et al.²⁸ and Pham et al.²⁹ Moreover, high dielectric loss at low frequencies from ZnO impedes its performance as a transducer for vibrational energy harvesting applications as ambient vibrations typically exist at frequencies below 1 kHz.³⁰

Due to the enhanced scope for potential applications of BaTiO₃ nanowires, there has been extensive research in the synthesis of these nanowires with a focus on different chemical methods^31–35 for achieving desired dimensions and orientation. However, aligned BaTiO₃ NW arrays had not been developed prior to this investigation for energy harvesting applications. Here, we report for the first time the synthesis method employed for obtaining novel vertically aligned BaTiO₃ nanowire (NW) arrays directly on a conductive fluorine doped tin oxide (FTO) glass with a length of ∼1 μm and an aspect ratio of ∼12. The aligned array formed by these nanowires enables them to exhibit higher strains as compared to their bulk form and thereby enhances the piezoelectric energy conversion capability of the material.^36,37 Consequently, we experimentally analyze the power harvesting performance of BaTiO₃ NW arrays as compared to conventional ZnO NW arrays by developing a novel NEMS piezoelectric energy harvester that is driven from local variations in acceleration from a vibrating source. Here, the NEMS energy harvester is designed to have resonance below 1 kHz for efficient energy harvesting of ambient mechanical vibrations, which usually reside in the 1 Hz to 1 kHz range. This communication also contains the first report on the comparison of the AC power and power density with respect to varying load resistors from NEMS energy harvesters fabricated using ZnO NW arrays and BaTiO₃ NW arrays. A reference shear accelerometer is utilized to give an accurate measure of the input base acceleration of the vibration source that triggers the voltage response from the NEMS energy harvester. The results reported here clearly demonstrate the potential for high performance NEMS energy harvesters using aligned arrays of BaTiO₃ NWs that can efficiently harvest mechanical vibrations when integrated with a suitable low frequency resonating structure.

Vertically aligned BaTiO₃ NW arrays are directly grown on a conductive FTO glass substrate using a two-step hydrothermal process that is inexpensive. BaTiO₃ NW arrays are synthesized by temperature assisted hydrothermal reaction in a solution containing Ba²⁺ ions with the single crystal vertically aligned titanium dioxide (TiO₂) NW arrays acting as precursors (see Methods). Detailed X-ray diffraction (XRD) analysis of precursor TiO₂ NW arrays on FTO glass used for conversion to BaTiO₃ is carried out to match with the rutile phase (see ESI, Fig. S1†). The resulting BaTiO₃ NWs have a length of ∼1 μm and a diameter of ∼90 nm preserving the morphology of precursor NW arrays following the ion exchange reaction. A detailed analysis of the microstructure of the aligned array of NWs is performed using a scanning electron microscope (SEM) as shown in Fig. 1a. The crystallographic structure of the nanowires is analyzed using X-ray diffraction (XRD) and the XRD pattern in Fig. 1b shows that the NWs are BaTiO₃ (JCPDS no. 5-0626). The high resolution transmission electron microscopy (HRTEM) image of the as-prepared single crystalline BaTiO₃ NWs with clear crystal lattice fringes is shown in Fig. 1c. For the power harvesting performance comparison, aligned ZnO NW arrays with a length of ∼1 μm and a diameter of ∼100 nm were also grown on a conductive FTO glass using a low temperature solution-growth approach as shown in the SEM image in Fig. 1d.^38,39 Both the BaTiO₃ NW arrays and ZnO NW arrays grown on conductive FTO glass were sputter coated with a 1 nm gold (Au) layer on the top surface prior to applying them as NEMS energy harvesters.

Fig. 1 Characterization of piezoelectric nanowires. (a) Cross-sectional SEM image of BaTiO₃ NW arrays with the inset showing the top view. (b) X-ray diffraction spectrum of BaTiO₃ NW arrays showing the majority of peaks to be BaTiO₃ (JCPDS no. 5-0626). (c) HRTEM image of the BaTiO₃ nanowire showing the clear crystal lattice fringes of single crystal structure. (d) Cross-sectional SEM image of ZnO NW arrays with the inset showing the top view.

The BaTiO₃ based NEMS energy harvester is fabricated by bonding a strip of indium foil to the non-conductive edge of the FTO glass substrate and formed into a beam to make contact with the top of the as-synthesized vertically aligned BaTiO₃ NW arrays thereby serving as the top electrode. The conductive FTO glass acts as the bottom electrode with the BaTiO₃ NW arrays sandwiched between the two electrodes. This configuration allows the NEMS energy harvesting device to achieve a low resonant frequency by capitalizing upon the beam's resonance rather than the NW's resonance. The ZnO NW NEMS energy harvester fabrication also utilizes the same indium beam technique to contact the ZnO NW arrays grown on FTO glass as shown in the schematic of the configuration in Fig. 2a. An Au layer (work function of ∼5.1 to 5.47 eV) that was sputter coated on as-synthesized ZnO nanowires (electron affinity of ∼4.1 to 4.3 eV) prior to device fabrication assists to form a Schottky barrier between indium (top electrode) and ZnO NW arrays.^10,40 It is important to form a Schottky barrier for efficiently extracting piezoelectric charge from the nanowire's tip and also for blocking the electron flow through the interface from the metal side to the semiconducting nanowires side. The surface area of the indium top electrode above the NW arrays has dimensions of ∼5 × 4 mm² in both NEMS energy harvesters to ensure similarity for power density comparison. The BaTiO₃ based NEMS energy harvester is then poled with a high DC electric field (∼120 kV cm⁻¹) for 24 hours to ensure that the dipoles of the single crystal BaTiO₃ NWs align in the electric field direction which is normal to the plane of the two electrodes along the orientation of the NWs.⁴¹ This high voltage poling between the two electrodes is essential for the piezoelectric function of ferroelectric BaTiO₃ NW arrays but is not required for ZnO NW arrays as they possess intrinsic spontaneous polarization. The NW energy harvesters are excited through base vibration generated by a permanent magnet shaker while the input base acceleration is accurately measured using a reference shear accelerometer (PCB 352C22). The photographic images of the experimental setup used for the characterization of the NW NEMS energy harvester are shown in ESI Fig. S2.†

Fig. 2 NEMS energy harvester configuration and characterization. (a) Schematic diagram of the NEMS energy harvester fabricated using piezoelectric NW arrays. (b) Schematic of voltage generation from the piezoelectric NEMS energy harvester (P denotes the polarization direction, σ denotes the stress and V is the piezo-voltage generated by the nanowires) with the inset at the bottom showing the electrical circuit representation of the piezoelectric energy harvester with the voltage (V_L) measured across the load resistor (R_L) to characterize the AC power delivered to the load (C_p denotes capacitance of the source and R_p is the leakage resistance).

Compressive and tensile stress generated from the inertial force of the vibrating indium beam on the BaTiO₃ and ZnO NW arrays result in charge generation from the direct piezoelectric effect, thus developing an alternating potential difference across the two electrodes as shown schematically in Fig. 2b. This forms the working principle of a vibration-driven NEMS energy harvester. The electrical equivalent circuit for the NEMS energy harvester is shown as an inset in Fig. 2 with the piezoelectric voltage, V, induced from the vibration acceleration in series with the inherent capacitance of the source, C_p, and piezoelectric leakage resistance, R_p, connected in parallel. The voltage, V_L, is measured across the load resistor, R_L, to calculate the AC power dissipation. Here, the piezoelectric leakage resistance, R_p = X_C²/R_S, is not taken into account as it is normally two orders of magnitude higher than the impedance (Z_S) of the source capacitance (in the pF range) where R_S is the series resistance and X_C = 1/jωC_p is the reactance of the capacitor. As a result, the effect of leakage resistance on the overall impedance is negligible. The source capacitance (C_p), which is the capacitance measured between the two electrodes of the NEMS energy harvester, is measured using an Agilent E4980A LCR meter (see ESI Table S1† for properties of the NEMS energy harvester used in power characterization). The impedance measurement showing the series resistance (R_S) and the reactance (X_C = 1/jωC_p) of the BaTiO₃ NEMS energy harvester and the ZnO NEMS energy harvester respectively is also performed (see ESI Fig. S3†). Here, the impedance contributed by the capacitance (Z_S = 1/(ω_nC_p)) of the piezoelectric NWs at resonant frequency, ω_n, is matched using purely resistive loads to determine the AC power.⁴²

All measurements are performed inside a grounded faraday cage to reduce the effects of extrinsic power-line noise (60 Hz harmonic noise) on the NEMS output voltage. The output voltage is measured using a high impedance (1 TΩ) voltage follower with unity gain, and the short circuit current is measured using a high speed electrometer (Keithley 6514). The dynamic response analysis of the NEMS energy harvester is performed using the frequency response function (FRF) characterization that gives the relative magnitude and phase of the ratio of the response signal from the NW arrays to the stimulus input base acceleration. Firstly, the FRF between the open circuit output voltage from the NEMS energy harvester and the input base acceleration measured by the reference shear accelerometer is examined to determine the open circuit resonant frequency when excited with burst chirp and white Gaussian noise signals from the shaker that have flat power spectral density in the test frequency range of up to 1 kHz (see ESI Fig. S4†). Secondly, the FRF between the short circuit current from the NW arrays and the input base acceleration is also characterized using burst chirp and white noise excitation. The piezoelectric open circuit voltage (V_OC) and short circuit current (I_SC) are maximized at the resonant frequency as they correspond to the frequency where the indium beam generates maximum strain on the NW arrays of the NEMS energy harvester. At the resonant frequency, the root mean square (RMS) voltage (V_L) measured across the external resistive load (R_L) will be used to determine the AC power (P_L) experimentally from the NEMS energy harvester as shown in eqn (1).⁴³ The peak AC power is dissipated when the external resistive load (R_L) is matched with the source impedance (Z_S) as per maximum power transfer theorem.

The capacitance of the BaTiO₃ NW energy harvester is measured by the LCR meter to be 8.21 pF at 1 kHz. The open circuit voltage V_OC FRF characterized from burst chirp voltage response (see ESI Fig. S5a†) after poling produced a resonant peak at ∼160 Hz as shown in Fig. 3a. The sinusoidal excitation at resonant frequency yielded a high peak to peak voltage V_pp of ∼623 mV from 1g RMS base acceleration input as shown in Fig. 3b. The high voltage response is due to the high dynamic strain on the NW arrays from the beam at resonance inducing an alternating piezoelectric charge accumulation at the two electrodes. It is well known that when measuring the open circuit voltage with a voltage buffer amplifier with high input impedance (1 TΩ), the current is at its minimum (theoretically zero) so the AC power is virtually zero.

Fig. 3 Open circuit voltage and short circuit current analysis from the BaTiO₃ NW NEMS energy harvester. (a) Open circuit voltage (V_OC) FRF showing the resonant peak at ∼160 Hz associated with a 90° phase shift. (b) V_OC from 1g RMS sinusoidal acceleration input near resonant frequency (∼160 Hz) shown in the top panel with the bottom panel showing the detailed shape of voltage and input base acceleration. (c) Short circuit current (I_SC) FRF with a resonant peak near ∼160 Hz. (d) I_SC from 1g RMS acceleration input near resonant frequency shown in the top panel with the bottom panel showing the detailed shape of current and input acceleration.

The short circuit current (I_SC) FRF from the BaTiO₃ NW NEMS energy harvester is also characterized by using burst chirp excitation input with the I_SC magnitude peak at a resonant frequency of ∼160 Hz associated with a 90° phase change as shown in Fig. 3c. The I_SC response to chirp input is recorded as shown in ESI Fig. S5b.† High I_SC values from the NW arrays are observed by exciting with a sine wave at resonant frequency (∼160 Hz) with the peak to peak current (I_pp) of ∼1.8 nA recorded from base acceleration input of 1g RMS (Fig. 3d). High I_SC is directly proportional to the piezoelectric charge production from the poled ferroelectric BaTiO₃ NW arrays when increased strain is applied by the resonating indium beam structure. In short circuit electrical boundary conditions, the voltage is theoretically zero so again the AC power is zero.

The capacitance of the ZnO NW NEMS energy harvester is measured by the LCR meter to be 8.72 pF at 1 kHz. A direct vibration excitation experiment is also carried out on the as-fabricated ZnO based NEMS energy harvester to investigate its performance and compare it with the BaTiO₃ device. The V_OC FRF and I_SC FRF of the ZnO energy harvester are analysed by triggering with white noise and burst chirp excitation and the resonant magnitude peak is observed at ∼190 Hz as shown in Fig. 4a and b. The V_OC and I_SC response from the ZnO NW NEMS energy harvester to the burst chirp input signal used for FRF characterization is shown in ESI Fig. S6a and b.† The peak to peak open circuit voltage V_pp and the peak to peak short circuit current I_pp from sine wave excitation at resonance are measured to be ∼85 mV and ∼0.316 nA from the 1g RMS input acceleration as shown in Fig. 4c and d. The voltage and current levels produced are much lower than the BaTiO₃ NW NEMS energy harvester due to ZnO's lower coupling coefficient.

Fig. 4 Open circuit voltage and short circuit current analysis from the ZnO NW NEMS energy harvester. (a) Open circuit voltage (V_OC) FRF showing the resonant peak near ∼190 Hz associated with a 90° phase shift. (b) Short circuit current (I_SC) FRF with resonance peak near ∼190 Hz. (c) V_OC measured near resonance from 1g RMS acceleration shown in the top panel with the bottom panel showing the detailed shape of voltage and acceleration. (d) I_SC from 1g RMS acceleration input near resonant frequency shown in the top panel with the bottom panel showing the detailed shape of current and acceleration.

The AC power from the energy harvester is calculated by measuring the voltage, V_L, across several load resistors, R_L, ranging from 1 MΩ to 500 MΩ. The source impedance, Z_S, of BaTiO₃ NW arrays with a capacitance of ∼8.21 pF at natural frequency (ω_n = 2πf_n where f_n = ∼160 Hz) was evaluated to be ∼121 MΩ. The AC power from the BaTiO₃ NW NEMS energy harvester increased rapidly as R_L increases up to 50 MΩ reaching a uniform peak value of ∼125.5 pW at an optimal R_L of 120 MΩ and then reduces as R_L is traced up to 500 MΩ since voltage across the increasing load resistors starts saturating towards the V_OC. The peak power density across the optimal R_L is calculated to be ∼6.27 μW cm⁻³ from 1g RMS base acceleration (Fig. 5a). For the ZnO NW NEMS energy harvester, the source impedance, Z_S, at resonant frequency (ω_n = 2πf_n where f_n = ∼190 Hz) is measured to be ∼96 MΩ and the peak AC power dissipated across the optimal R_L of 50 MΩ is only ∼8 pW from the same input base acceleration of 1g RMS. The peak power density from the ZnO based NEMS energy harvester is calculated to be ∼0.4 μW cm⁻³ as shown clearly in Fig. 5b. This power density is ∼16 times lower than the peak power density (∼6.27 μW cm⁻³) recorded from the BaTiO₃ based NEMS energy harvester driven by the same base acceleration of 1g RMS and hence substantiates the superior vibrational energy harvesting performance from BaTiO₃ NW arrays. In addition, the voltage magnitude of the FRF from BaTiO₃ NEMS is characterized across several load resistors and the magnitude peak at resonant frequency is found to increase with the increase in the load resistors with the maximum peak being that of the open circuit voltage FRF (1 TΩ) as shown in Fig. 5c. Moreover, the magnitude of the voltage FRF from ZnO NEMS is also characterized across several load resistors (R_L) to demonstrate a similar increase in magnitude with the highest peak at resonance from the V_OC FRF (Fig. 5d). The voltage V_L across the optimal R_L with a RMS value of ∼123 mV provides the maximum peak power density from the BaTiO₃ energy harvester (see ESI Fig. S5c†). For ZnO NEMS, the voltage V_L across optimal R_L has a RMS value of ∼20.2 mV to provide the lower peak power density as compared to BaTiO₃ NEMS from the same base acceleration input (see ESI Fig. S6c†). In addition, a switching polarity test is performed to confirm that the measured signal responses from the NEMS energy harvester are generated by the nanowires (ESI Fig. S7†). Moreover, this reported power density of the BaTiO₃ NEMS energy harvester (∼6.27 μW cm⁻³) is comparable to several meso-scale and MEMS-scale resonant cantilever based energy harvesters driven by base vibration.^44–46

Fig. 5 Power characterization of NEMS energy harvesters. (a) AC power and power density of the BaTiO₃ NW NEMS energy harvester with varying load resistor (R_L) showing a peak power of ∼125.5 pW and a peak power density of ∼6.27 μW cm⁻³ at an optimal R_L of 120 MΩ from 1g RMS acceleration. This peak power levels are much greater than the peak power from the ZnO NW NEMS energy harvester. (b) AC power and power density of the ZnO NW NEMS energy harvester with varying load resistor (R_L) showing a lower peak power of ∼8 pW and a peak power density of ∼0.4 μW cm⁻³ at an optimal R_L of 50 MΩ from 1g RMS sinusoidal base acceleration. (c) Voltage magnitude of FRF from the BaTiO₃ NW NEMS energy harvester measured for various load resistors shows an increasing magnitude peak as the load resistance increases. (d) Voltage magnitude of FRF from the ZnO NW NEMS energy harvester measured for various load resistors shows an increasing magnitude peak as the load resistance increases with the maximum peak from V_OC FRF (1 TΩ).

Conclusions

In summary, a novel approach to fabricate NEMS vibrational energy harvesters with resonant frequencies below 200 Hz using vertically aligned ferroelectric BaTiO₃ NW arrays (∼1 μm long) and conventional ZnO NW arrays (∼1 μm long) that are synthesized directly on a conductive FTO glass has been demonstrated. This low resonant frequency is very important for the implementation of vibration based energy harvesters since spectral energy is rarely available beyond 1 kHz in ambient vibration sources. Power characterization analysis of NEMS energy harvesters has been performed under direct vibration excitation from a vibration shaker equipped with a shear accelerometer that can give an accurate measure of the input base acceleration. This is the first report on the AC power and power density comparison of novel ferroelectric BaTiO₃ NW array and conventional ZnO NW array based NEMS through experimental testing. The results demonstrate the superior vibrational energy harvesting capability of the BaTiO₃ NW arrays that have a peak power density of 6.27 μW cm⁻³ by resistive impedance matching from input base acceleration of 1g RMS. This power density of BaTiO₃ NW arrays is ∼16 times higher than the peak power density from ZnO NW arrays of ∼0.4 μW cm⁻³ measured across the optimal load resistor from the same sinusoidal base acceleration of 1g RMS at resonant frequency. In addition, the peak open circuit voltage and peak short circuit current levels at resonant frequency measured from the BaTiO₃ NW based NEMS energy harvester are demonstrated to be more than 5 times greater than the response recorded from a ZnO based NEMS energy harvester. This reported power density of the BaTiO₃ NEMS energy harvester is comparable to many meso-scale and MEMS-scale resonant vibrational energy harvesters.^44–46

Methods

Synthesis of aligned BaTiO₃ NW arrays

Synthesis of the vertically aligned BaTiO₃ nanowire (NW) arrays is performed on a conductive substrate using a two-step hydrothermal reaction. First, the precursor TiO₂ nanowire arrays were grown on conductive fluorine doped tin oxide (FTO) glass (Pilkington, TEC7 coated, 2.2 mm thick, 7 Ω sq⁻¹) through an acidic hydrothermal reaction process.⁴⁷ Initially, FTO glass was cut into square dimensions (∼10 mm × ∼10 mm) using a laser ablator (Epilog Laser) and was cleaned by sonication for 30 minutes in a 1 : 1 : 1 volume ratio solution of deionized water, acetone, and 2-propanol. After sonication, the FTO glass substrate was rinsed with methanol and water, and placed vertically inside a high pressure reactor containing 10 mL of deionized water, 10 mL of hydrochloric acid (Fisher, 37%) and 1 mL of titanium isopropoxide (Fisher, ACS). The reactor was then heated at 200 °C for 3 hours. Following the first hydrothermal process, the reactor was cooled down to room temperature and the resultant FTO glass substrate with an array of vertically aligned TiO₂ nanowires was rinsed with deionized water and dried in ambient air. The substrates were then put into a solution containing Ba²⁺ ions and converted to BaTiO₃ by a second hydrothermal reaction which was carried out at temperatures between 150 °C and 240 °C for 4 to 8 hours.^48–50 The Ba²⁺ ion solution concentration and temperature (150–240 °C) of the ion exchange procedure were optimized to enable shape retention of the precursor TiO₂ NW arrays during conversion resulting in an aligned BaTiO₃ NW arrays. Lastly, the samples were rinsed again with deionized water and dried in ambient air to yield BaTiO₃ NW arrays on a conductive FTO glass substrate. The as-synthesized BaTiO₃ NW arrays were heat treated at 600 °C for 30 minutes to remove any hydroxyl defects before their use as NEMS energy harvesters.^51,52

Synthesis of aligned ZnO NW arrays

The ZnO NW arrays were also synthesized on a FTO glass substrate (∼10 mm × ∼10 mm, 2.2 mm thick) using a low temperature solution growth approach.³⁸ The FTO glass substrate was initially cleaned in ethanol and acetone (1 : 1) solution by sonication for 10 min. It was then removed and ultrasonicated in DI water for 2 min followed by drying at 100 °C for 5 min. The conductive side of the FTO glass substrate was then seeded with 2 mM zinc acetate (Zn(O₂CCH₃)₂, Alfa) in ethanol solution by dip coating and then performing thermal decomposition at 300 °C for 20 min. The growth solution (fill factor: 40%) was prepared using 25 mM zinc nitrate hexahydrate (Zn(NO₃)₂·6H₂O, 99%, Sigma-Aldrich), 25 mM hexamethylenetetramine (HMTA, Sigma-Aldrich) and 5–7 mM polyethylenimine (PEI, Aldrich).^39,53 The FTO glass was immersed on the top surface of the growth solution with the seeded conductive side facing down so that ZnO particles do not precipitate on the NW arrays. The reaction is carried out at 85 °C for 3 hours in a convection oven. The resulting substrate with the aligned ZnO NW arrays on the FTO glass substrate was rinsed with DI water and dried at room temperature.

Characterization of BaTiO₃ NW arrays and ZnO NW arrays

The morphological properties which include the orientation and dimensions of the BaTiO₃ NW arrays and ZnO NW arrays were examined using an ultra-high resolution field-emission scanning electron microscope (FESEM) FEI Nova NanoSEM 430. The crystal structure of the as-prepared BaTiO₃ NWs and ZnO NWs was examined using an X-ray diffractometer (XRD) equipped with a curved position sensitive detector (CPS120, Inel) with Cu K_α radiation. The crystal structure and lattice parameter of individual BaTiO₃ nanowires were studied using an FEI (Philips) Tecnai F30 high resolution transmission electron microscope (HRTEM) that operates at 300 kV accelerating voltage provided by a field-emission electron gun (FEG).

Fabrication of the NEMS vibrational energy harvester

The NEMS energy harvester using aligned BaTiO₃ NW arrays was fabricated by sputtering a 1 nm Au layer on top of the as-prepared NW arrays grown on the FTO glass substrate using a PELCO SC-7 Auto Sputter Coater. A malleable indium (Alfa-Aesar, 99.9%, 0.127 mm thick) foil was then bonded to the base of the non-conductive glass substrate and formed into a beam to make contact with the top of the NW arrays to serve as the top electrode. The Au layer (work function ∼5.1 to 5.47 eV) that was initially coated on top of the BaTiO₃ NW arrays improves the contact with the indium top electrode and also assists to form a barrier to minimize leakage as reported by McCormick et al.⁵² The indium beam served as the top electrode while the conductive side of the FTO glass substrate served as the bottom electrode with the BaTiO₃ NW arrays in between to form a sandwich configuration. The FTO glass substrate's edge was insulated using a Kapton polyimide (Dupont) film to ensure there is no resistive electrode contact to cause shorting. The above fabricated BaTiO₃ NW NEMS energy harvester was poled at room temperature by supplying a high DC voltage of ∼120 KV cm⁻¹ (TREK 477A Supply/Amplifier) across the two electrodes for 24 hours to ensure that the dipoles align in the electric field direction.

The NEMS energy harvester using ZnO NW arrays was also fabricated using a 1 nm Au layer sputtered on the as-synthesized aligned ZnO NW arrays on the FTO glass substrate with the same procedure as discussed above with the indium beam to serve as the top electrode. The Au layer assists to form a Schottky barrier between the indium electrode and the semiconducting ZnO nanowires.⁵⁴ Similarly, sufficient insulation at the FTO substrate edge was needed so a polyimide film was used to ensure that there was no shorting between the two electrodes.

Electrical measurement

The capacitance and impedance measurements from the NEMS energy harvesters were conducted using an Agilent E4980A high precision LCR meter. Mechanical vibration was generated from a Miniature Permanent Magnet shaker (Labworks, Inc. ET-132) and the voltage measurements from the NEMS energy harvester under vibration excitation were performed using a voltage follower/buffer amplifier with unity gain constructed using Linear Technologies (LTC6240CS8 CMOS Op Amp) which was chosen for its high input resistance (1 TΩ), low input bias current (0.2 pA) and low noise (voltage noise < 10 nV Hz^−1/2).⁵⁴ The short circuit current measurement from the NEMS energy harvester was performed using a high-speed electrometer (Keithley 6514, up to 1200 readings per s). The grounded faraday cage used as a noise shield from electromagnetic interference (EMI) surrounded the NEMS energy harvester thus attenuating the extrinsic noise and preserving the piezoelectric NW linear characteristics. The burst chirp signals for FRF characterization were generated using the Spectral Dynamics Siglab data acquisition (DAQ) system (Model 50-21) from a virtual function generator (vfg) in the MATLAB environment. All other signals were generated and acquired through a DAQ board (NI USB 4431) operated using NI SignalExpress software. All signals were re-examined for accuracy during data acquisition using an oscilloscope (Tektronix, DPO 3014 Digital Phosphor Oscilloscope).

Acknowledgements

The authors would like to gratefully acknowledge support from the Air Force Office of Scientific Research (AFOSR) under contract number FA9550-12-1-0132 and the direction of Dr Douglas Smith.

Notes and references

1. Z. L. Wang, Adv. Mater., 2012, 24, 280–285.
2. C. Falconi, G. Mantini, A. D'Amico and Z. L. Wang, Sens. Actuators, B, 2009, 139, 511–519.
3. A. Nechibvute, A. Chawanda and P. Luhanga, Smart Mater. Res., 2012, 2012, 13.
4. S. N. Cha, J. Seo, S. M. Kim, H. J. Kim, Y. J. Park, S. Kim and J. M. Kim, Adv. Mater., 2010, 22, 4726–4730.
5. X. Wang, J. Song, J. Liu and Z. L. Wang, Science, 2007, 316, 102–105.
6. J. Liu, P. Fei, J. Zhou, R. Tummala and Z. L. Wang, Appl. Phys. Lett., 2008, 92, 173105.
7. X. Wang, J. Liu, J. Song and Z. L. Wang, Nano Lett., 2007, 7, 2475–2479.
8. M. Choi, D. Choi, M. Jin, I. Kim, S. Kim, J. Choi, S. Y. Lee, J. M. Kim and S. Kim, Adv. Mater., 2009, 21, 2185–2189.
9. X. Chen, S. Xu, N. Yao and Y. Shi, Nano Lett., 2010, 10, 2133–2137.
10. Z. L. Wang and J. Song, Science, 2006, 312, 242–246.
11. P. X. Gao, J. Song, J. Liu and Z. L. Wang, Adv. Mater., 2007, 19, 67–72.
12. R. Yang, Y. Qin, C. Li, G. Zhu and Z. L. Wang, Nano Lett., 2009, 9, 1201–1205.
13. B. J. Hansen, Y. Liu, R. Yang and Z. L. Wang, ACS Nano, 2010, 4, 3647–3652.
14. S. Xu, Y. Qin, C. Xu, Y. Wei, R. Yang and Z. L. Wang, Nat. Nanotechnol., 2010, 5, 366–373.
15. A. Yu, P. Jiang and Z. L. Wang, Nano Energy, 2012, 1, 418–423.
16. B. Kumar and S. Kim, Nano Energy, 2012, 1, 342–355.
17. Y. Hu, L. Lin, Y. Zhang and Z. L. Wang, Adv. Mater., 2012, 24, 110–114.
18. G. Zhu, A. C. Wang, Y. Liu, Y. Zhou and Z. L. Wang, Nano Lett., 2012, 12, 3086–3090.
19. S. Xu, B. J. Hansen and Z. L. Wang, Nat. Commun., 2010, 1, 93.
20. Y. Qi, N. T. Jafferis, K. Lyons, C. M. Lee, H. Ahmad and M. C. McAlpine, Nano Lett., 2010, 10, 524–528.
21. Y. Qi, J. Kim, T. D. Nguyen, B. Lisko, P. K. Purohit and M. C. McAlpine, Nano Lett., 2011, 11, 1331–1336.
22. Y. Lin, Y. Liu and H. A. Sodano, Appl. Phys. Lett., 2009, 95, 122901.
23. Z. Wang, J. Hu, A. P. Suryavanshi, K. Yum and M. Yu, Nano Lett., 2007, 7, 2966–2969.
24. K. Park, S. Xu, Y. Liu, G. Hwang, S. L. Kang, Z. L. Wang and K. J. Lee, Nano Lett., 2010, 10, 4939–4943.
25. Z. Lin, Y. Yang, J. M. Wu, Y. Liu, F. Zhang and Z. L. Wang, J. Phys. Chem. Lett., 2012, 3, 3599–3604.
26. P. K. Panda, J. Mater. Sci., 2009, 44, 5049–5062.
27. C. Sun, J. Shi and X. Wang, J. Appl. Phys., 2010, 108, 034309.
28. J. I. Sohn, S. N. Cha, B. G. Song, S. Lee, S. M. Kim, J. Ku, H. J. Kim, Y. J. Park, B. L. Choi, Z. L. Wang, J. M. Kim and K. Kim, Energy Environ. Sci., 2013, 6, 97–104.
29. T. T. Pham, K. Y. Lee, J. Lee, K. Kim, K. Shin, M. K. Gupta, B. Kumar and S. Kim, Energy Environ. Sci., 2013, 6, 841–846.
30. N. Setter, Electroceramic based mems: fabrication-technology and applications, Springer, 2005.
31. P. M. Rørvik, T. Grande and M. Einarsrud, Adv. Mater., 2011, 23, 4007–4034.
32. J. J. Urban, J. E. Spanier, L. Ouyang, W. S. Yun and H. Park, Adv. Mater., 2003, 15, 423–426.
33. J. J. Urban, W. S. Yun, Q. Gu and H. Park, J. Am. Chem. Soc., 2002, 124, 1186–1187.
34. U. A. Joshi, S. Yoon, S. Baik and J. S. Lee, J. Phys. Chem. B, 2006, 110, 12249–12256.
35. S. O. Leontsev and R. E. Eitel, Sci. Technol. Adv. Mater., 2010, 11, 044302.
36. H. D. Espinosa, R. A. Bernal and M. Minary-Jolandan, Adv. Mater., 2012, 24, 4656–4675.
37. Z. Yan and L. Jiang, J. Phys. D: Appl. Phys., 2011, 44, 075404.
38. L. E. Greene, B. D. Yuhas, M. Law, D. Zitoun and P. Yang, Inorg. Chem., 2006, 45, 7535–7543.
39. U. Galan, Y. Lin, G. J. Ehlert and H. A. Sodano, Compos. Sci. Technol., 2011, 71, 946–954.
40. C. S. Lao, J. Liu, P. Gao, L. Zhang, D. Davidovic, R. Tummala and Z. L. Wang, Nano Lett., 2006, 6, 263–266.
41. W. J. Merz, Phys. Rev., 1954, 95, 690.
42. N. Kong, D. S. Ha, A. Erturk and D. J. Inman, J. Intell. Mater. Syst. Struct., 2010, 21, 1293–1302.
43. S. Priya, J. Electroceram., 2007, 19, 167–184.
44. V. R. Challa, M. G. Prasad and F. T. Fisher, Smart Mater. Struct., 2009, 18, 095029.
45. W. J. Choi, Y. Jeon, J.-H. Jeong, R. Sood and S. G. Kim, J. Electroceram., 2006, 17, 543–548.
46. D. Shen, J.-H. Park, J. Ajitsaria, S.-Y. Choe, H. C. Wikle and D. Kim, J. Micromech. Microeng., 2008, 18, 055017.
47. B. Liu and E. S. Aydil, J. Am. Chem. Soc., 2009, 131, 3985–3990.
48. B. Im, U. A. Joshi, K. H. Lee and J. S. Lee, Nanotechnology, 2010, 21, 425601.
49. Z. Zhou, Y. Lin, H. Tang and H. A. Sodano, Nanotechnology, 2013, 24, 095602.
50. H. Tang, Y. Lin and H. A. Sodano, Adv. Energy Mater., 2012, 3, 451–456.
51. N. Bao, L. Shen, G. Srinivasan, K. Yanagisawa and A. Gupta, J. Phys. Chem. C, 2008, 112, 8634–8642.
52. M. A. McCormick and E. B. Slamovich, J. Eur. Ceram. Soc., 2003, 23, 2143–2152.
53. C. Xu and D. Gao, J. Phys. Chem. C, 2012, 116, 7236–7241.
54. ‘Linear Technology’, LTC6240/LTC6241/LTC6242 Single/Dual/Quad 18 MHz, Low noise, Rail-to-Rail Output, CMOS Op Amps, 2011.

---

Footnote

† Electronic supplementary information (ESI) available: Additional figures and table are supplied to support the synthesis and power characterization performed on BaTiO₃ and ZnO nanowire array based energy harvesters. See DOI: 10.1039/c3ee42540a

---