Direct Writing of Patterned, Lead‐Free Nanowire Aligned Flexible Piezoelectric Device

A high‐performance flexible piezoelectric nanogenerator (PNG) is fabricated by a direct writing method, which acquires both patterned piezoelectric structure and aligned piezoelectric nanowires simultaneously. The voltage output of the as‐prepared PNG is nearly 400% compared with that of the traditional spin‐coated device due to the effective utilization of stress. This facile printing approach provides an efficient strategy for significant improvement of the piezoresponse.

Recently, printing technology is experiencing a striking development in diverse architecture manufacturing and device fabricating. [45][46][47][48] Among them, direct writing which deposits continuous fi lament serves as a fl exible approach to pattern materials for both planar and freestanding 3D structure in a layer-by-layer sequence. [48][49][50]  nanoparticles (NPs) are also fabricated to compare with their corresponding nanowires ( Figure S3, Supporting Information). Figure 1 e demonstrates the cross-sectional SEM image of a KNN NWs based direct-written (NWs-DW) PNG with a sandwiched structure along the direct writing direction. A magnifi ed cross-sectional SEM image (Figure 1 f) shows that the KNN NWs adopt strongly preferential orientations in the poly(dimethylsiloxane) (PDMS) matrix. XRD pattern ( Figure S4, Supporting Information) of the NWs-DW PNG without top electrode is also measured to verify the preferential orientation. Detailed description and explanation can be referred to the Supporting Information. Piezoelectric layer's cross-sectional SEM graphs of a KNN NWs based spin-coated (NWs-SC) PNG and a KNN NPs based spin-coated (NPs-SC) PNG are also taken for comparison ( Figure S5, Supporting Information). The crosssectional SEM image of the NWs-SC PNG exhibits random oriented KNN NWs in the PDMS matrix due to the inconstant direction of the centrifugal force during the spin coating process, which is completely different from the KNN NWs' arrangement in the NWs-DW PNG. For illustration of the direct-written structure, an optical image of the piezoelectric layer with paralleled pattern (Figure 1 g) is demonstrated. The black line which shows a stereo-morphology is the sample, while the golden line is the uncovered Au/Cr-coated polyethylene terephthalate (PET) bottom electrode substrate. The paralleled pattern is chosen to guarantee nanowires in different fi laments aligned in the same  Figure 1 h presents a photograph of a bendable PNG with an effective area of 2 cm × 2 cm, which indicates that the device is very fl exible. Figure 2 presents the printing behavior of the KNN NWs based ink and its corresponding printed morphology. The rheology of the ink should be optimized to print fi laments with moderate aspect ratio, which ensures the entity of the patterned piezoelectric layer architecture and keeps the underlying layers with minimal deformation. The ink used in the study compromises KNN NWs and PDMS matrix (Figure 2 a inset). The rheological behavior of the inks with solid loading of different concentration ( c ) ranging from 10%-40% is studied (KNN NWs can hardly well disperse in PDMS when c ≥ 50%). From Figure 2 a, which provides the apparent viscosity ( η ) as a function of shear rate, the viscosity is in proportion to the ink concentration. The viscosity is dramatically increased at low shear rate when the ink concentration reaches 40%. At the same time, it also exhibits highly shear thinning behavior, which guarantees the ink fl owing smoothly through the nozzle during printing. Figure 2 b demonstrates their storage modulus ( G ′) and loss modulus ( G ″) as a function of shear strain. For inks with concentration below 40%, they demonstrate liquid-like response ( G ′ < G″). On the contrary, the ink with solid loading of 40% exhibits a storage modulus plateau that exceeds loss modulus by almost an order of magnitude at strain lower than shear yield strain (≈0.05%), which therefore ensures a solidlike nature in the quiescent state. Thus it can be concluded that among the tested inks, the one with solid loading of 40% exhibits the most desired rheological behavior due to its high viscosity under low shear rate, the shear thinning behavior and the solid-like response ( G ′ > G ″) under low shear strain, which is distinct compared with the inks with concentration lower than 40%.
A direct-written printer is utilized to print the well-performed ink through a 200 µm cylindrical nozzle (Figure 2 d inset). In order to acquire a patterned PNG with enhanced piezoelectric lamina thickness, a 3D multilayer structure is printed. SEM images show the side views (Figure 2 c) of 1 layer, 3 layers, and 5 layers of the piezoelectric lamina obtained through a layerby-layer printing sequence, respectively. Figure 2 d presents their corresponding height of fi laments as a function of the printed layer number. It is obvious that the structure height enhances with the layer number. Whereas, their width remains nearly constant when the layer number increases, which is concluded from their corresponding top views and the relationship between width and the layer number ( Figure S6, Supporting Information).
Generated output of the PNG device during the periodic bending/unbending tests is carried out employing a bending stage executed at a horizontal displacement of 2 cm with a bending frequency of 0.5 Hz under the 0.44% strain. It should be noted that the bending trajectory is parallel to the pattern fi lament's direction ( X -axis), which is also consistent with the nanowires' orientation. Figure 3 a displays three states of the mechanical motion, i.e., original, bending, and release state, and their corresponding power generation mechanism. In the bending state, the current fl ow is generated due to dipoles movement from equilibrium position and  charges accumulation at two opposite side surfaces of the material. In the release state, the accumulated charges return to their original state, which results in output signals in the opposite direction. The 1-layer KNN NWs-DW PNG generates an open-circuit voltage of ≈21.0 V and a short-circuit current of ≈0.5 µA under mechanical deformation, respectively ( Figure 3 bi), and a well-behaved, periodic alternation of negative and positive peaks of electric signals are observed. To confi rm that the measured output signals are purely generated by the PNG sample, a widely used switching-polarity test is conducted, and the inversion of voltage and current signals is demonstrated in Figure 3 bii in the reverse connection.   The output performance of the control device fabricated by directly writing the PDMS matrix without the KNN NWs is also measured. As shown in Figure S7 in the Supporting Information, there is no reliable signal generated from the direct-written device containing only pure PDMS layer. The low electrical signals, which are presumably caused by several electrostatic charges at the electrodes, can be ignored in comparison with the NWs-DW PNG.
The output performance ( Figure 4 a and Figure S8, Supporting Information) of the KNN NWs-SC PNG with same working area and piezoelectric lamina thickness of the 1-layer KNN NWs-DW PNG is also measured for comparison. As shown in Figure 4 a, the voltage output difference of the NWs-SC PNG and the 1-layer NWs-DW PNG is apparently appreciable under identical stress. Specifi cally, the open-circuit voltage of the direct-written device (≈21.0 V) is nearly 400% compared with that of the spin-coated one (≈5.3 V), which verifi es the effect of the direct-written structure features on the output performance of PNGs. For further confi rmation, a corresponding simulation is carried out via a fi nite element method, and the numerical modeling result is approximately in agreement with the experimental result. The calculated piezoelectric voltage output of the NWs-SC PNG and the 1-layer NWs-DW PNG is presented in Figure 4 b, where piezopotential is depicted by color code. The piezoelectric lamina of the NWs-SC PNG is a fl at layer with nanowires random oriented, while the NWs-DW PNG consists of aligned oriented nanowires located in the patterned PDMS matrix, and the same number of nanowires is included in both simulations for a fair comparison. For visual recognition of the inner nanowires' arrangement inside the PDMS matrix, half-transparent photographs corresponding to the model in calculation are illustrated as displayed in Figure S9 (Supporting Information).
The aligned nanowires and the micropatterned morphology which are both integrated in the PNG device via the direct writing method, tend to explain in-depth the enhanced energy harvesting property. KNN NWs in the PDMS matrix adopt strongly preferential orientations (Figure 1 f) after extrusion from the nozzle because of the applied shear force induced alignment. [ 51 ] The identity of the nanowires' longitudinal direction and the force direction leads to an enlargement of the nanowires' deformation, thus results in exaggerated dipole displacement and piezoelectric property. Apart from the advantage caused by the nanowires' alignment, the micropatterned morphology with paralleled fi laments may also result in the performance enhancement. When the force is applied along the nanowires' longitudinal direction ( X -axis), the PDMS matrix tends to deform, which lays a compression on the piezoelectric component, namely the KNN NWs. The stress would inevitably lead to strain not only in the parallel direction ( X -axis) but also in the perpendicular direction ( Y -axis) of the nanowires' orientation. However strain in Y -axis is not wanted because force in the radial direction of the piezoelectric nanowires leads to less deformation than force in the longitudinal direction, thus impairing the PNG output performance. If there is a gap between fi laments, the structure would restrict stress relaxation  Adv. Sci. 2016, 3, 1600120 in Y -axis more effectively, which accordingly increases the strain in X -axis. The consistency of the nanowires' direction, pattern fi laments' direction, and force direction results in an enlargement of strain in a cooperative manner. Thereby, the micropatterned piezoelectric lamina with nanowires in aligned arrangement fabricated by direct writing would signifi cantly enhance the degree of stress sensitivity and the effi ciency of mechanical impact transfer, which inevitably contributes to the strain enlargement. The open-circuit voltage ( V out ) generated by the fl exible PNG is interpreted as [ 36,52 ] where l is the perpendicular distance between the adjacent electrodes, ε( l ) is the function of the strain along the direction of l , Y is the Young's modulus, and g 31 is the piezoelectric voltage coeffi cient. The increasing of strain inside the composite inevitably leads to an enhanced voltage output according to the equation. The inner mechanism of the relationship between strain variations and the piezoelectric output can be referred to the Supporting Information. Energy harvesting property of KNN NPs-SC PNG is also investigated ( Figure S10, Supporting Information). It can be concluded that piezoelectric nanowires play a superior effect than nanoparticles in energy harvesting, which might be due to their effective transport of charge carriers, high piezoelectricity, and responsiveness to tiny random mechanical disturbances. Thus, the nanowires based PNG provides feasibility for scenarios where only small triggering forces are available and contributes to higher sensibility. The cooperative effect of material's morphology, alignment of nanowires, and the pattern of piezoelectric lamina would lead to the exaggerating performance of the piezoelectric property.
To further improve the voltage output of this well-performed direct-written device, PNGs with increasing piezoelectric lamina thickness realized by 3D multilayer direct writing are fabricated. The open-circuit voltage output (Figure 4 a) and the short-circuit current output ( Figure S8, Supporting Information) of NWs-DW PNGs with 3 layers and 5 layers of the piezoelectric lamina under dynamic bending-unbending cycles are measured. As expected, it shows that the electric performance increases with respect to the layer number and consequently the distance between electrodes ( d ). The device generates high voltage output up to ≈72.2 V compared with that of previous reported lead-free PNGs (Table S1, Supporting Information) when a 5-layer PNG is made. Similarly, the calculated voltage output as a function of printed layer number is also performed (Figure 4 b). To be mentioned, further increasing layer of the PNG is not studied because of the piezoelectric layer fracture induced by thickness enhancement.
Depending on the exceptional piezoelectric characteristic of the direct-written PNG device, applications of sensing the mechanical movement and harvesting energy are demonstrated. When a fi nger with the PNG wrapped on is mechanically bent ( Figure 5 a), apparent voltage signals and a huge amount of voltage are generated as illustrated in Figure 5 b. As for the deviation of the output peaks, it is attributed to the fi nger's compression impacted on the device, which results in the voltage output enhancement, and the irregular speed of the fi nger mechanical motion. Apart from sensing mechanical movement, the power generated by the PNG is also applied to operate commercial electronic units. Commercial light-emitting diodes (LEDs) are connected in series for power generation demonstration (Figure 5 c). To store the electrical energy from periodically bending the PNG device, it is incorporated in a circuit comprising a full-wave bridge rectifi er with four diodes and a 2.2 µF capacitor connected in series as demonstrated in top inset of Figure 5 d. During the charging process, the total stored voltage in the capacitor reaches up to 37.7 V (bottom inset of Figure 5 d) by continual bending/unbending deformation of the PNG. The stored energy is suffi cient to light up 12 commercial LED arrays in series as depicted in Figure 5 d. These results demonstrate that the direct-written PNG can successfully harvest electric energy for generating remarkable signals and driving electronic units.
A high-performance fl exible KNN NWs based PNG is fabricated by a direct writing method, which acquires both patterned piezoelectric structure and aligned piezoelectric nano wires simultaneously. The voltage output of the asprepared PNG is nearly 400% compared with the traditional spin-coated one due to the consistency of nanowires' direction, pattern fi laments' direction, and force direction, which exaggerates the effective utilization of stress and the induced high degree of strain. Moreover, by further increasing printed layers, signifi cant performance improvement can be achieved with a maximum voltage output up to ≈72.2 V via the multilayer printing. This simple, effective, and nonlithographic printing approach applied to various materials provides an effi cient strategy for signifi cant improvement of the piezoresponse, allowing for usage in fi elds such as energy harvesting and pressure sensing. Furthermore, it can also be potentially adopted as a strategy for enhancing performance of the recently developed triboelectric nanogenerator [53][54][55] by patterning the contact surface.

Experimental Section
Preparation of the Piezoelectric Nanocomposite Based Ink : Piezoelectric nanocomposite-based inks for direct writing and spin coating were prepared by dispersing 40 wt% KNN NWs (or KNN NPs) in PDMS precursor which was the mixture of PDMS base and curing agent in the proportion of 10:1 by weight. The inks were mechanically stirred at 800 rpm for 0.5 h in an icing bath for homogenization.
Direct Writing of the Microptterned Piezoelectric Layer : The substrate used for direct writing was the bottom electrode, namely an Au/Cr-coated PET fi lm (150 µm in thickness) whose conducting layer was deposited by a high vacuum thermal evaporation system (PATOR, ATT010). The KNN NWs/PDMS ink was loaded in a syringe and extruded through a micronozzle with inner diameter of 200 µm onto the substrate under an applied pressure of 30 psi at a speed of 10 mm s −1 . The micropattern was created using a 3-axis micropositioning stage. The micropatterned piezoelectric layer was fully cured at 150 °C for 0.5 h to retain its morphology.
Spin Coating of the Flat Piezoelectric Layer : The KNN NWs/PDMS (or KNN NPs/PDMS) ink was spin-coated onto the bottom electrode at the rate of 2500 rpm (2000 rpm) for 30 s and cured at 80 °C for 2 h.
PNG Fabrication : Then, an Au/Cr-coated polyimide fi lm (50 µm in thickness) used as top electrode was attached to the surface of the piezoelectric layer with PDMS precursor served as an adhesive agent. The PDMS adhesive agent was spin-coated onto the top electrode at the rate of 3000 rpm for 30 s and procured at 80 °C for 5 min. After fully hardened at 80 °C for 2 h, the packaged device was poled at 150 °C by applying an electric fi eld for 12 h. 1-layer and spin-coated PNG were poled at 0.4 kV. 3-layer and 5-layer PNG were poled at 1.2 and 2.0 kV, respectively.
Characterization and Measurements : The SEM image was characterized using a Hitachi S-4800 scanning electron microscope operating at 5 kV accelerating voltage. The TEM image was taken with a JEOL TEM-2100 transmission electron microscope operating at 200 kV accelerating voltage. XRD pattern was carried out using a Rigaku D/MAX 2500 X-ray powder diffractometer equipped with a 18 kW Cu Kα radiation. The weight percent of elements was determined using an Oxford Instruments INCA EDS operating at an accelerating voltage of 15.0 kV. The optical image of the micropatterned piezoelectric layer was investigated by a Nikon ECLIPSE LV100ND optical microscope which was coupled with a chargecoupled device camera. The ink rheology was measured using an Anton Paar Modular Compact Rheometer. A bending stage was utilized to apply a periodic deformation to the PNG at a desired displacement and speed. The electrical signal of the PNG device was obtained using a KEITHLEY Model 2450 Interactive SourceMeter SMU Instrument. The simulation investigation was conducted by using COMSOL multiphysics software.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.