Piezoelectricity Enhancement of Nanogenerators Based on PDMS and ZnSnO3 Nanowires through Microstructuration.

The current trend for smart, self-sustainable, and multifunctional technology demands for the development of energy harvesters based on widely available and environmentally friendly materials. In this context, ZnSnO3 nanostructures show promising potential because of their high polarization, which can be explored in piezoelectric devices. Nevertheless, a pure phase of ZnSnO3 is hard to achieve because of its metastability, and obtaining it in the form of nanowires is even more challenging. Although some groups have already reported the mixing of ZnSnO3 nanostructures with polydimethylsiloxane (PDMS) to produce a nanogenerator, the resultant polymeric film is usually flat and does not take advantage of an enhanced piezoelectric contribution achieved through its microstructuration. Herein, a microstructured composite of nanowires synthesized by a seed-layer free hydrothermal route mixed with PDMS (ZnSnO3@PDMS) is proposed to produce nanogenerators. PFM measurements show a clear enhancement of d33 for single ZnSnO3 versus ZnO nanowires (23 ± 4 pm/V vs 9 ± 2 pm/V). The microstructuration introduced herein results in an enhancement of the piezoelectric effect of the ZnSnO3 nanowires, enabling nanogenerators with an output voltage, current, and instantaneous power density of 120 V, 13 μA, and 230 μW·cm-2, respectively. Even using an active area smaller than 1 cm2, the performance of this nanogenerator enables lighting up multiple LEDs and other small electronic devices, thus proving great potential for wearables and portable electronics.

Introduction

Nowadays, there is an increasing demand for smart, self-sustainable, and multifunctional technology to enable concepts such as internet-of-things, relying on low-cost, and compact devices. The vision is to expand smart-functionality with embedded electronics, sensors, and connectivity to an ever-increasing number of objects. The development of self-powered sensors that harvest the ambient environmental energy will be the key toward these systems and is mandatory to be in line with the preservation of the global environment and sustained economic growth.[1]

One great renewable electricity source is vibration energy harvesting, which has been explored essentially from electromagnetic, electrostatic, and piezoelectric transduction.[2,3] Electromagnetic energy harvesters are the least reported ones, with an output voltage that is poorer when compared with the other two types, despite being the most suitable for low-frequency stimuli.[4] Piezoelectric nanogenerators (PENG) are the most used transducers for mechanical vibrations, human motion, stream flow, and acoustic noise.[5] This type of energy harvester provides higher output voltages and is more efficient at high frequencies.[1,4,6] Finally, electrostatic transduction is typically based on the triboelectric effect that causes an electric potential difference when a periodic gap or mismatch is induced by friction between two contacted surfaces with opposite triboelectric polarities.[7] In a triboelectric nanogenerator (TENG), this physical contact creates electrostatic charges, driving the electrons to flow back and forth between back-coated electrodes, meaning this phenomenon is more efficient for a low-frequency range.[1,2] Therefore, PENGs and TENGs have a high potential for different applications such as sustainable energy sources, biomedical systems, and smart sensors.[8,9]

Several materials have been used for piezoelectric energy harvesters, the most common being Pb(Zr, Ti)O3 (PZT), P(VDF-TrFE), PTFE, ZnO, ZnSnO3, BaTiO3, and GaN.[2,10−14] Although PZT is the one with the highest piezoelectric constant, its conductivity is low and it is environmentally harmful.[5] When looking for sustainable options it is important to avoid toxic and nonrecyclable materials but also those which are scarce. Within these premises, Zn-based metal oxides such as ZnO and ZnSnO3 are some of the most promising materials. ZnO has been widely used because of its easy synthesis and coupling of piezoelectric and semiconductor properties.[1,15−19] ZnSnO3 appears to be an even better candidate for energy harvesting applications, having a higher polarization (≈59 μC/cm2) along the c-axis.[20] Moreover, the advantages of zinc tin oxide (ZTO) are numerous. Not only are both zinc and tin abundant and recyclable, but ZTO is also a ternary oxide, its properties can be tuned by adjusting the cationic ratio, boosting its multifunctionality. Depending on the synthesis method, different morphologies can be obtained and ZnSnO3 can crystallize in two different structures, rhombohedral and orthorhombic, undergoing a phase transformation into Zn2SnO4 at high temperatures (>700 °C).[21−25] However, the presence of two cations in ZTO makes the synthesis of a pure single phase very hard to achieve. Several reports about the application of ZnSnO3 micro- and nanostructures in PENGs and TENGs were published in the last few years. ZnSnO3 nanocubes have been the most reported structure with piezoelectric and piezoresistive effects for applications in energy harvesting and sensitive human motion sensors.[1,5,26−33] More recently, in 2017 Guo et al. reported piezo-nanogenerators using flat films of polydimethylsiloxane (PDMS) mixed with ZnSnO3 nanoplates, produced by a hydrothermal method, showing the difficulty to obtain a single and pure phase of this material by solution processing.[5] Concerning ZnSnO3 micro/nanowires, there are only a few reports of generators making use of vapor phase synthesis methods. In 2012 and 2013, Wu et al. reported nanogenerators of ZnSnO3 microbelts, produced by thermal evaporation at 900 °C, generating a power output around 11 μW·cm–3.[26,28,34] Furthermore, these ZnSnO3-based nanogenerators are generally based in flat films of a mixture of PDMS or other polymers and oxide nano- or microstructures[5,28,31−33,35−37] or simply use PDMS to encapsulate such structures,[26,34] therefore not taking advantage of the microstructuration of the polymeric film to increase the efficiency in the transduction of a force into an electrical output.[12,38−44]

In a previous work, a seed-layer free hydrothermal route at 200 °C was optimized to produce well-controlled ZnSnO3 nanowires.[24] The microstructuration of the PDMS film by a laser engraving technique was also studied and optimized by our group to produce piezoresistive sensors, where several architectures were developed.[45−47] Here, we combine, for the first time, these two recent techniques to microstructure a robust composite of nanowires mixed with PDMS (ZnSnO3@PDMS) into microcones, thus improving the efficiency of the force delivery to the nanowires and greatly increasing their piezoelectric signal. Because of this effect the device presents a high potential for energy harvesting applications.

Experimental Section

ZnSnO3 Nanowires’ Synthesis

The ZnSnO3 nanowires were produced by a hydrothermal synthesis in a conventional oven using SnCl4·5H2O from Riedel-de Haën (0.01 M), ZnCl2 from Merck (0.02 M), and NaOH from Sigma-Aldrich (0.24 M) as precursors and H2O and ethylenediamine from Sigma-Aldrich in a volume proportion of 7.5:7.5 mL as solvents, at 200 °C for 24 h, based on our previous works.[24,48] The resultant precipitate of nanowires’ syntheses was centrifuged at 6000 rpm and washed several times with deionized water and isopropyl alcohol, alternately. The nanostructures were finally dried at 60 °C, in vacuum, for 2 h.

Nanogenerator Fabrication

Four acrylic molds were used, combining aligned or misaligned microcone cavities with measured gaps between cones of 0 or <100 μm. These acrylic molds were produced on a previous work.[45] A laser engraving machine (VLS3.50, 50 W, Universal Laser System, USA) with a CO2 laser beam, a lens’ focus focal length of 2.0 in., a focal spot of 127 μm in diameter, a power of 25 W, and a speed of 0.1524 m·s–1 was used to engrave acrylic plates (Dagol, 5 mm thick, squares of 5 cm × 5 cm) with designs based in the repetition of symmetrical crosses (size of 100 μm) distributed over an area of 2 cm × 2 cm in an aligned or misaligned way and spaced by a theoretical gap between cones of 150 or 300 μm. The final functional area of the devices is 4 cm2.

The ZnSnO3 nanowires were mixed with PDMS with a weight ratio of 20 wt %. Then, the curing agent was added to the previous mixture in a 1:10 w/w ratio of curing agent to elastomer. The mixture was then spin-coated onto each mold at 250 rpm for 90 s, and degassed before curing the ZnSnO3@PDMS composite in an oven at 85 °C. After 30 min of curing, a commercial polyethylene terephthalate substrate covered with indium tin oxide (PET/ITO) substrate (Kintec Company) was placed on top of composite films, which were left in the oven until completing the curing. This PET/ITO substrate played the role of the bottom electrode in the device. After peeling off the composite bounded to PET/ITO, another PET/ITO substrate was placed on top of the microstructured face to work as a top electrode. Both substrates were tightly secured to each other by using kapton tape (DuPont). For an easier connection to external equipment, copper tape (3M) was used as an extension of each electrode of the nanogenerator.

ZnSnO3 Nanowire, Composite, and Nanogenerator Characterization

The structural characterization of the produced ZnSnO3 nanowires and ZnSnO3@PDMS composite was done using a PANalytical’s X’Pert PRO MRD diffractometer with Cu Kα radiation. The X-ray diffraction (XRD) data were acquired in the 10–90° 2θ range with a step size of 0.033°. The morphology and element analysis [scanning electron microscopy/energy-dispersive X-ray (SEM/EDS)] were performed with a Carl Zeiss AURIGA CrossBeam (FIB-SEM) workstation. To investigate the properties of the composite, cured films of the ZnSnO3@PDMS composite without microstructuration were spin-coated with the same parameters as previously mentioned and aluminum electrodes (100 nm) were evaporated on their surface. As a comparison, the same was performed for only PDMS. A constant force of approximately 10 N was applied by a home-made machine (Figure S1) with a linear motor at a frequency of 2 pushes per second to all composites and devices. The contact area for these tests was 0.3 cm2. Besides, the home-made machine, in order to test the devices with different forces, a pen was used to apply human force between 12 and >100 N in an area of 0.7 cm2. The output voltage was collected by a digital oscilloscope (Tektronix TDS 2001C, 10 MΩ input impedance), while the current output signals were acquired using a Keysight B1500A system. The force applied in the produced devices was estimated using a commercial force sensing resistor from Interlink Electronics (ref. SEN05003). The capacitance of PDMS and the ZnSnO3@PDMS composite was measured using a Keysight B1500A system. In order to compare the results of ZnSnO3 nanowires with well-known structures, the same characterization was performed for synthesized ZnO nanowires based on ref (49) (Supporting Information, Figure S2). Piezoelectric force microscopy was performed with a NX10 Park-Systems setup using Micromash NSC36CrAu probes. Sparse nanowires on ITO substrates were detected in a noncontact mode. Piezoresponse measurements were subsequently performed by contacting a single point on the nanowire surface and applying an AC signal of 1 to 4 V to the tip at 17 kHz.

Results and Discussion

Characterization of ZnSnO3 Nanowires and ZnSnO3@PDMS Films

Figure shows the XRD patterns and morphology of ZnSnO3 nanowires. These structures have an average length of 600 nm and diameter of 65 nm. Concerning the crystalline structure identification, although the #28-1486 card was deleted from the ICDD database, this card fits well with a ZnSnO3 orthorhombic phase. In ref (24), this identification was proven. Moreover, this card has been also widely reported as a ZnSnO3 orthorhombic phase, emphasizing the identification.[30,50−53]

Figure 1

(a) XRD patterns of PDMS, ZnSnO3@PDMS film, and ZnSnO3 nanowires before mixing with PDMS. The identification was following ICDD card #28-1486 as explained in ref (24). (b) SEM image of ZnSnO3 nanowires.

Electromechanical properties of individual and unpolarized nanowires were characterized by atomic force microscopy (Figure a,b). Piezoresponse curves were obtained by contacting a single spot on the nanowire surface and are shown in Figure b for ZnO and ZnSnO3. For ZnO, the response is slightly above the noise level and leads to an effective piezoelectric constant of d33 = 9 ± 2 pm/V, in agreement with the values reported for ZnO nanostructures.[54] The ZnSnO3 structures instead yield almost three times stronger responses leading to d33 = 23 ± 4 pm/V. This value compares fairly well with what has been reported in the literature for one-dimensional (1D) nanostructures produced by a hydrothermal synthesis, as assumed by Ghasemian et al. and presented in Table S1.[55] In fact, the obtained d33 (23 ± 4 pm/V) is surpassed only by the well-known BaTiO3 (31.1 pm/V) and LiNbO3 (25 pm/V) which contain critical raw materials, highlighting the interest in ZnSnO3 for piezoelectric applications. Additionally, electrostatic force microscopy (see Figure S3) shows that the relaxed nanowires do not exhibit static surface charges which could impact on their electromechanical response. Thus, the findings demonstrate the piezoelectric properties of the individual nanowires.

Figure 2

Atomic force microscopy to characterize the piezoresponse of individual ZnO and ZnSnO3 nanowires: (a) nanowire topographies of ZnO and ZnSnO3 obtained in noncontact mode (the red spot indicates the area that was contacted in piezoresponse measurements); (b) contact mode tip oscillation as a function of tip-bias ac-voltage to extract the effective piezocoefficient d33. (c) Piezoelectric response of ZnSnO3@PDMS and only PDMS for a pushing force of 10 N. The inset shows the output of ZnSnO3@PDMS with a greater detail, highlighting the peaks corresponding to the pressing and releasing of the composite.

In order to investigate the piezoelectric effect of the ZnSnO3@PDMS composite and to compare it with a film of only PDMS, electrodes of aluminum were directly evaporated on both surfaces of these samples to avoid any air gaps between the films and the electrodes, thus preventing any triboelectric effect. Subsequently, the samples were subjected to a mechanical stimulus of 10 N. In Figure c, it is clear that PDMS has barely any voltage generation upon pushing, which is expectable given that PDMS is not a piezoelectric material. The small signal recorded is coming from the interference with the electrodes.

The piezoelectric coefficient (d33) value of a material can be determined according to: d33 = Q/F, where Q is the electric charge and F is the force. Knowing that the charge can be obtained by: Q = C × V, where C is the capacitance and V is the voltage, the expression for d33 becomes: d33 = C × V/F.[56] To clarify that the voltage output differences for PDMS and ZnSnO3@PDMS are not coming from different capacitance values, the capacitance of these materials was measured, achieving values of 7.3 and 10 pF, respectively, for samples with the same area (≈3 cm2). Given that the capacitance of the ZnSnO3@PDMS composite is higher than PDMS, and assuming similar d33 coefficient for both, the ZnSnO3@PDMS composite output voltage should be smaller. However, as shown in Figure c, the output voltage for the composite is much higher than for PDMS, suggesting a larger d33 of ZnSnO3@PDMS, which was determined as ≈2 pm/V. Therefore, the ZnSnO3@PDMS composite presents an evident piezoelectric response which must be attributed to the ZnSnO3 nanowires.

Optimization of the Devices’ Architecture

Herein, two different nanogenerator architectures (an unstructured and a microstructured one) were explored to study the influence of the microcones of the ZnSnO3@PDMS composite in the devices’ performance. The fabrication process of the microstructured nanogenerators is illustrated in Figure , where the microcavities with the shape of inverted cones are engraved by laser in acrylic plates. Four different engraving designs were explored to obtain aligned or misaligned microcones with different gaps between cones (0 or <100 μm). The microstructured films with the dimensions of the microcones can be seen in Figure S4. As it is further explored in ref (45), the measured gaps between cones (0 or <100 μm) are smaller than the gaps between cones that are used in the design imported to the laser engraving machine (150 or 300 μm, respectively) because of the melting of the mold’s material (acrylic) during the engraving of the cavities. After molds’ engraving, the ZnSnO3@PDMS composite is spin-coated on the molds, being the composite capped by PET/ITO substrates in both sides, to work as bottom and top electrodes. A photograph of a final device is shown in the inset of Figure . For the unstructured ZnSnO3@PDMS device, instead of spin-coating the composite over molds, the composite is spin-coated directly over the bottom electrode, and the top electrode is placed over the free surface of the composite after a partial curing.

Figure 3

Fabrication steps of a microstructured nanogenerator. The inset is a photograph of one nanogenerator.

For a better comparison between the devices and to achieve more reproducible results, a home-made machine (Figure S1 and Supporting Information Video S1) was used with a pushing force of 10 N and a pushing area of 0.3 cm2. Despite the limitation on the force intensity applied through this machine, it allows to correctly compare all the samples, guaranteeing that the output differences between them are due to their different compositions, instead of different forces in the case of manual stimulation. Unless otherwise stated, this machine was used to mechanically stimulate all the samples analyzed in this work. Figure a,b shows the output voltage and current of devices with the different configurations previously described, highlighting the reproducibility of the output for each repeated stimulus. The unstructured device has a smaller piezoelectric output than the one presented in Figure c (for which an evaporated electrode was employed) because of an irregular contact between the top PET/ITO substrate and the ZnSnO3@PDMS composite. Nevertheless, the results evidence that the presence of microstructuration gives rise to a higher output when compared to unstructured devices. Even though the microstructuration of the films reduces the effective contact area between the composite and the electrode, it is thought that it concentrates the compressive forces into the cones (Supporting Information Video S2), and so upon their compression and deformation, the nanowires dispersed in the microstructures undergo a greater compression/bending. This translates into more charges being induced at the electrode, resulting in a greater piezoelectric signal. The compressive forces do not actuate so effectively upon the nanowires in unstructured devices, which is illustrated in Figure e and Supporting Information Video S3. Therefore, in the unstructured nanogenerators the nanowires suffer less mechanical deformation resulting in lower charge induction and consequently in smaller outputs, even though their contact area with the electrode is larger.

Figure 4

Output (a) voltage and (b) current generated from the ZnSnO3@PDMS device with different configurations: unstructured films, microstructured films with aligned or misaligned microcones with a gap between cones of 0 or <100 μm. Output (c) voltage and (d) current generated from only PDMS and ZnSnO3@PDMS microstructured films devices with aligned microcones with a gap between cones of <100 μm. The circles represent average values with standard deviations (of 4–6 measurements) for positive and negative peaks. Abbreviations: A—aligned, M—misaligned, G—gap. (e) Proposed schematics of force deformation for a microstructured and unstructured device and photographs showing the cross section of the devices before and after pushing force. Note that the schemes are not at scale. T (1.2 mm) and T′ (940 μm) and t (782 μm) and t′ (724 μm) are the thicknesses of the microstructured/unstructured ZnSnO3@PDMS films before and after pushing, respectively.

The microstructuration of the composite also induces the presence of air gaps between the basis of the cones and the top electrode, which contributes to the induction of charges by a triboelectric effect. This is verified in Figure S5, where a microstructured device of only PDMS shows a higher output voltage and current than an unstructured one. To investigate the relevance of the triboelectric effect contribution to the performance of the ZnSnO3@PDMS composite nanogenerators, microstructured devices of only PDMS and ZnSnO3@PDMS with aligned microcones with a gap between cones of <100 μm were tested under the same conditions as previously described, with the results shown in Figure c,d. If the triboelectric effect was the main effect contributing to the performance of the nanogenerators, then ZnSnO3@PDMS would have an output similar to that of only PDMS, where its output can only come from the triboelectricity generated between PDMS and ITO. However, the PDMS device shows a peak-to-peak voltage (0.7 V) and current (95 nA) output much smaller than the respective values for the ZnSnO3@PDMS nanogenerator with the same configuration (8.7 V and 1.25 μA). Therefore, the output results for the ZnSnO3@PDMS nanogenerator point toward an enhanced piezoelectric effect resultant from the structuration of the composite into microcones, which is the dominant effect dictating the performance of this type of nanogenerators. Based on these findings, a proposed charge generation and displacement mechanism for both the piezoelectric and triboelectric effects in microstructured devices is illustrated in Figure S6.

Regarding the different designs explored herein, devices with aligned cones present a slightly higher voltage than devices with misaligned cones, while for a current output the alignment does not seem to produce a significant influence. Table S2 presents the peak-to-peak output voltage and current values of all the devices shown in Figure a,b. The configuration with aligned microcones with a gap between cones <100 μm has the best performance with a peak-to-peak voltage of 8.7 V and a peak-to-peak current of 1.25 μA, possibly because the density of cones in this configuration is smaller than in the configurations with a gap between cones of 0 μm, and so the forces get more concentrated in the first configuration, leading to a greater deformation of the microcones. This effect has been shown by other groups in capacitive sensors.[57,58] Additionally, the microcones for a gap between cones <100 μm are higher than the microcones with a gap between cones of 0 μm because the latter ones suffer from a fusion effect as a consequence of their fabrication process (Figure S4, please check ref (45) for further details). Therefore, there is a greater volume to be compressed in the configuration with a gap between cones <100 μm, and so a greater number of nanowires may be bent, thus increasing their piezoelectric signal.

Pushing Force Dependence

To stimulate the device with the optimized configuration with larger forces, different levels of tapping force using a pen (contact area of 0.7 cm2) were tested. Figure shows the peak-to-peak voltage and current outputs obtained in these experiments. The results highlight that the nature of the stimuli applied by the home-made machine is different from these manual stimuli: while the first resembles a continuous stimulus, the latter are more impulse like. Therefore, for similar forces applied manually (between 12 and 15 N), the device has a higher voltage and current output with values of 31.2 V and 1.8 μA, corresponding to an instantaneous power density of 8.2 μW·cm–2. The maximum output of the device with this type of stimuli was ≈120 V and ≈13 μA, corresponding to an instantaneous power density of ≈230 μW·cm–2.

Figure 5

Output (a,b) voltage and (c,d) current of the nanogenerator with the optimized configuration, generated by applying a human force from 15 to 50 N and over 100 N using a pen (inset) to deliver the stimulus.

Device Stability

In order to study device stability over time, 12,000 cycles were performed for 100 min without interruption on a nanogenerator with the optimized ZnSnO3@PDMS microstructured films. As shown in Figure a,b, despite a slight variation in the output, mostly at initial cycles because of the unoptimized assembly of the PET/ITO top contact with copper tape, the device presents a stable performance even after being pushed over 12,000 cycles. The microcones present no signs of degradation after this test, as shown in Figure S7. After one month of the initial measurements (Figure a,b), the same level of performance is observed, as presented in Figure S8.

Figure 6

Output (a) voltage and (b) current generated from the nanogenerator with the optimized configuration for 12,000 cycles (100 min).

Proof of Concept

To test the device for a practical application, it is relevant to investigate the electric power that the device can provide with a load. Therefore, the output voltage, current, and instantaneous power of the device with the optimized configuration were evaluated under several load resistances, as shown in Figure a,b. The output current decreases with an increasing load resistance while the output voltage increases. The instantaneous power increases with the load resistance, reaching a maximum peak of 1 μW with 15 MΩ, and then decreasing for higher load resistances.

Figure 7

Output (a) voltage and current and (b) instantaneous power generated from the ZnSnO3@PDMS microstructured films with the optimized configuration under several load resistances. Photographs of a device directly connected to (c) a single white LED and (d) five blue LEDs in series, with the respective insets showing the LED/LEDs off and on (driven by the energy of the nanogenerator). Photographs of (e) an electronic humidity sensor and (f) a digital watch being powered up by a 10 μF capacitor previously charged by the nanogenerator.

The device with the optimized configuration was then directly connected to one white light-emitting diode (LED) or five blue LEDs in series. The I–V curves of the single and five LEDs are shown in Figure S9 and highlight that the single white LED needs at least 2.3 V and 1.56 μA to light up, while the five blue LEDs in series need 11.4 V and 0.13 μA. Figure c,d shows that after manually pushing the device either the single white LED (Supporting Information Video S4) or the multiple blue LEDs in series (Supporting Information Video S5) could light up.

Powering of small electrical devices was also tested with the optimized nanogenerator. For this end, the nanogenerator was stimulated to charge a 10 μF capacitor through a full-wave rectifier over 10–15 min. The capacitor was then connected to an electronic humidity sensor, a digital watch, and a dial watch to successfully switch them on, as shown in Figure e,f and Supporting Information Video S6. The circuit schematic and the capacitor charging curve are presented in Figure S10.

Table shows a comparison between the device with the best performance produced in this work with the literature related to the use of ZnSnO3 nanostructures in piezo/triboelectric harvesters. Although a direct comparison is not possible because of the lack of some information from other works, the device produced herein shows a high performance for energy harvesting from stimuli with a moderate force, compatible with everyday activities. For example, the nanogenerator could be attached to the bottom of shoes to get energy from walking or running. The instantaneous power density reached herein is also one of the highest reported so far, and only the ZnSnO3 nanocubes@PDMS composite reported by Wang et al.[35] has a higher response, but due to the lack of information and quantification about the applied force, a direct comparison is not feasible. Nevertheless, these results highlight the potential for powering portable electronics. Furthermore, given the higher flexibility of the wire morphology when compared with cubes or other structures, the optimization of the weight ratio between ZnSnO3 nanowires and PDMS is expected to result in devices with a greater output for the same stimulus, when compared to those presented in Table .

Table 1

Performance Comparison of ZnSnO3 Nanostructure-based Piezo/Triboelectric Harvesters

material	stimulus/applied force	output voltage	output current	instantaneous power density	year/refs
single-ZnSnO3 microbelt@high temperature process	compressive strain (0.33–1.20 Hz) N/A	110 mV	80 nA	N/A	2012/[26]
single-ZnSnO3 microbelt@high temperature process	compressive strain (0.30–1.66 Hz) N/A	5.3 V	0.13 μA	11 μW·cm–3	2012/[59]
ZnSnO3 nanocubes@PDMS	motion of vehicle tires N/A	20 V	1 μA·cm–2		2014/[33]
ZnSnO3/MWCNTs	strain N/A	40 V	0.4 μA	10.8 μW·cm–3	2015/[32]
ZnSnO3 nanocubes@PDMS	linear motor/bending (1.5–3 Hz) N/A	400 V	28 μA	280 μW·cm–2	2015/[35]
ZnSnO3 nanocubes@PVDF	fatigue testing system (1–7 Hz) 489 N	520 V	2.7 μA·m–2		2016/[36]
ZnSnO3 nanocubes@PVC	human finger (1–10 Hz) N/A	40 V	1.4 μA	≈3.7 μW·cm–2	2016/[37]
ZnSnO3 nanoplates@PDMS	bending (2 Hz) N/A	20 V	0.6 μA		2017/[5]
ZnSnO3 nanocubes@PDMS + TENG	hand stabbing mechanism N/A	50 V (PNG) 300 V (hybrid)		10.4 mW·cm–2 (hybrid)	2017/[60]
ZnSnO3 nanocubes@PDMS	footstep N/A	18 V	8 mA		2019/[61]
ZnSnO3 nanowires@PDMS	pen >100 N	≈120 V	≈13 μA	≈230 μW·cm–2	this work

Conclusions

PENGs were fabricated using a composite film of PDMS and ZnSnO3 nanowires produced by a hydrothermal synthesis at only 200 °C. The ZnSnO3 nanowires clearly showed piezoelectric properties through PFM measurements, reaching a d33 of 23 ± 4 pm/V, considerably larger than the d33 recorded for ZnO nanowires. Several microstructuration designs for the composite films were also studied to enhance the performance of the nanogenerators. The microcones helped to improve the piezoelectric signal coming from the nanowires through an increased efficiency in the force transmission. Such an effect results in a better performance achieved for the configuration with aligned microcones with a gap between cones <100 μm, showing a peak-to-peak voltage around 8.7 V and a peak-to-peak current of 1.25 μA (instantaneous power density of 3.24 μW·cm–2) for a 10 N stimulus applied by a home-made machine. When stimulated manually with a pen (a force over 100 N), a maximum output of 120 V and 13 μA was achieved, with an estimated instantaneous power density of 230 μW·cm–2. This instantaneous power density is enough for practical applications, such as lighting up LEDs and charging capacitors to levels enabling powering up small electronic devices, showing great potential for energy harvesting in wearables.