Triboelectric pixels as building blocks for microscale and large-area integration of drop energy harvesters

Article Triboelectric pixels as building blocks for microscale and large-area integration of drop energy harvesters Graphical abstract Highlights dTrecxel: a device architecture for one-step conversion of droplet energy dA layout compatible with micrometric and submillimetric scales dAn increase of 67% of peak power by harvesting with four- Trecxel one-drop impact dLife cycle reaches beyond 18,000 Milli-Q droplet impacts Authors Ali Ghaffarinejad, Xabier Garcı ´a-Casas, Fernando Nu ´n ˜ez-Ga ´lvez, ..., Juan Ramo ´nSa ´nchez-Valencia, A ´ngel Barranco, Ana Borra ´s Correspondence [email protected] (X.G.-C.), [email protected] (A.B.) In brief We present the first triboelectric nanogenerator configuration to simultaneously convert the energy from a single drop impact by several devices. Trecxel’s characteristic dimensions are designed to be compatible with a wide range of scales and configurations. Whether it is micrometric or submillimetric scales, bulk or thin-film triboelectric layers, or transparent or flexible devices, Trecxel can adapt to all. This versatility of Trecxel reassures us of its potential to be a game-changer in energy harvesting of solid-liquid interactions. Ghaffarinejad et al., 2025, Device 3, 100566 February 21, 2025 ª2024 The Author(s). Published by Elsevier Inc. https://doi.org/10.1016/j.device.2024.100566 ll Article Triboelectric pixels as building blocks for microscale and large-area integration of drop energy harvesters Ali Ghaffarinejad, 1,3 Xabier Garcı ´a-Casas, 1,3, *Fernando Nu ´n ˜ez-Ga ´lvez, 1,2 Jorge Budagosky, 1,2 Vanda Godinho, 1 Carmen Lo ´pez-Santos, 1,2 Juan Ramo ´nSa ´nchez-Valencia, 1 A ´ngel Barranco, 1 and Ana Borra ´s 1,4, * 1 Nanotechnology on Surfaces and Plasma Laboratory, Consejo Superior de Investigaciones Cientı ´ficas (CSIC), Materials Science Institute of Seville (CSIC-US), C/Ame ´rico Vespucio 49, 41092 Seville, Spain 2 Departamento de Fı ´sica Aplicada I, Universidad de Sevilla, C/Virgen de Africa 7, 41011 Seville, Spain 3 These authors contributed equally 4 Lead contact *Correspondence: xabier.garci[email protected] (X.G.-C.), [email protected] (A.B.) https://doi.org/10.1016/j.device.2024.100566 SUMMARY Triboelectric nanogenerators (TENGs) are the most promising technology for harvesting energy from low-frequency liquid flows and impacts such as rain droplets. However, current drop energy harvester technologies suffer from low output power due to limitations in triboelectric materials, suboptimal device designs, and the inability to fully capture the kinetic energy of falling drops. This article introduces a microscale TENG capable of efficiently converting drop impact energy into electrical power in a single, rapid step. The device features a capacitive structure that enhances energy conversion through instantaneous capacitance changes when the drops contact the submillimetric top electrodes. This compact design establishes a path toward the development of dense arrays and rain panels and is adaptable to various liquids, scales, triboelectric surfaces, and thin-film configurations, including flexible and transparent materials. THE BIGGER PICTURE Water stores immense potential energy through low-frequency movements in waves, rivers, rain, and tides. Solid-liquid triboelectric nanogenerators have emerged to harness efficiently the kinetic energy in moving or falling drops. In regions such as Europe, where most significant cities experience prolonged rainy forecasts, the potential for such energy harvesting parallel to photovoltaic panels is critical to compliance with the decarbonization agenda. Beyond rain panels, these nanogenerators hold promise for powering small-scale liquid sensors and microfluidic devices. However, the main factor preventing the full realization of the latter relies on the non-optimal and centimeter-scale device architecture applied so far, which involves a two-step energy-conversion process that inherently wastes the kinetic and potential energy stored in the drops. Here, we disclose the first compact triboelectric nanogenerator architecture to harvest the high electrical power density from a single water drop through a single-step process. The Trecxel converts the complete kinetic energy of a falling drop into microsecond-long electrical pulses by creating instantaneous capacitance variation due to the high-speed contact of the drop with the reduced active area of the top electrodes. The size of the active electrodes can be adapted to the mean droplet size to allow kinetic energy harvesting from a single droplet by several Trecxels simultaneously, providing a straightforward demonstration of submillimetric TENG arrays. Trecxel’s compatibility with supported and self-supported rigid, flexible, and transparent configurations enables record values of maximum peak power up to 2,000 W/m 2 from first interactions with salty drops. Through a proof of concept utilizing CMOS-compatible technology, the Trecxels based on SiO 2 thin films showcase instantaneous power-conversion rates exceeding 5 mW per droplet interaction, highlighting its potential for practical applications and compatibility with low environmental impact and high-yield production approaches. Device 3, 100566, February 21, 2025 ª2024 The Author(s). Published by Elsevier Inc. 1 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Please cite this article in press as: Ghaffarinejad et al., Triboelectric pixels as building blocks for microscale and large-area integration of drop energy harvesters, Device (2024), https://doi.org/10.1016/j.device.2024.100566 ll OPEN ACCESS INTRODUCTION The generation of electrical energy from low-frequency water movement is an emerging concept compared to wind turbines and solar panels. 1–6 Thus, a narrow variety of energy scavenger architectures have been developed during the last decade to harvest from drops and liquid movements available in nature, such as ocean waves, river flows, and rainfall, to produce electrical energy as an alternative to the centuryold large-scale hydraulic power generation. 7,8–11 These harvesters work following the electrokinetic effect or as solidsolid, solid-liquid, or liquid-liquid (slippery induced porous surface) triboelectric nanogenerators (TENGs). 7,12–16 Interestingly, most examples for drop-TENGs (D-TENGs) share a common two-step energy generation process that brings two major problems: first, a falling water drop impacts the charged dielectric (triboelectric) interface and loses part of the kinetic energy, then slips down toward an exposed electrode or slips over a buried electrode for electricity generation. 6,15,17–19 Following this, several novel designs of D-TENGs have been devised. However, attaining substantial instantaneous power output remains a formidable challenge. In 2022, a ground-breaking advancement in the field emerged by introducing a next-generation D-TENG featuring a ‘‘transistor-like’’ structure. 20 This innovation achieved an extraordinary instantaneous power output of 50 W/m 2 . Also, in this architecture an essential element is the extended bottom electrode in contact with the bottom side of the triboelectric layer. Very recently, Wang and co-workers have reported on the benefits of using arrays of bottom electrodes to increase the responsivity and, therefore, the energy density harvested from multiple-droplet impact events. 21 In such a ‘‘solar panel-like’’ D-TENG, a bridge reflux structure is used to independently control each D-TENG, avoiding the movement of charges between the lower electrodes and reducing the inherent capacitance between the bottom electrode and the ground, and by the implementation of a vertical diode structure, leading the charges in a unidirectional path. However, all these proposed architectures still impose relatively large areas for the energy harvesters (typically on the centimeter scale), which become unnecessarily large when dealing with energy coming from falling drops, as in the case of rainfalls, with a typical diameter of a few millimeters. This large form factor results in a significant reduction of the generated power density. In this article, we aim to solve this issue by demonstrating an alternative drop energy harvester with submillimetric dimensions that converts most of the kinetic energy of falling drops and simultaneously improves the output power density by several orders of magnitude. 3D finite element modeling of the drop and the triboelectric device interaction for different liquid conductivities confirms the sudden capacitance variation and charge induction as principal mechanisms for instantaneous power generation. The dimensions, architecture, and compatibility with materials make this triboelectric harvester a game-changer in the long-awaited dream of large-scale energy conversion from falling drops, paving the way for rain panels and thin-film-compatible selfpowered liquid sensors. RESULTS Trecxel design and working mechanism Figure 1 presents the design and working mechanism of the proposed triboelectric drop energy harvester, where the capacitive and microscale structure of the device resembles the pixels in a charge-coupled image sensor. Hence, we name it a triboelectric pixel or, in short form, Trecxel (Figure 1A). In the first approach, the Trecxel has a top square cross-section and an active area of L3Lwhich includes the back and two top electrodes (dark gray in Figure 1A). The system is compatible with any type of substrate, rigid or flexible, polymeric or inorganic, consisting of or decorated by a triboelectric layer. The dynamics of the falling drop that impacts the triboelectric surface and the charging state of the surface determine the voltage generated by the Trecxel. In the case of a hydrophobic surface, such dynamics follow a highly asymmetric path, with a very short spreading time (in the range of microseconds) when landing on the surface and much longer recoiling and sliding times (on the order of a few milliseconds). 22–27 To fully elucidate the working mechanism of the device, we have included a detailed description of the spreading, recoiling, and sliding dynamics in Note S1. Figure 1B schematically presents the complete cycle of electricity generation caused by the impact of a drop on the active area of the Trecxel. At instant t 1 , where the output voltage is zero, a high-speed drop is approaching to make a sudden contact withthe Trecxel. At t 2 , the highkinetic energy of the droplet that impacts the surface rapidly deforms it. The generated voltage in this instance is the result of two simultaneous processes: first, charge induction in the drop when it comes into contact with the triboelectric layer, and second, the sudden increase in capacitance of the Trecxel due to the change of effective overlapping area (A e )between the back and top electrodes, which generate a high-power output. Here, the power of such potential difference (V Load as indicated in Figure 1A) is measured by connecting a resistive load (R L ) to the Trecxel. The fast dynamic of the spreading process can be modeled according to the analytical model of Pasandideh-Fard et al., 26,28 which predicts the instantaneous drop diameter in contact with the surface (Note S1 and Video S1). Figure 1Balsoshows a comparison of the normalized voltage pulse generated by a single Trecxel (for a charged triboelectric surface, gray curve, see experimental details below) and the time derivative of the area (dotted curve, as will be discussed below; the change in the contact area of the dropaffects the deliveredvoltage) according to the spreading dynamics after the impact ofa droplet on a hydrophobic surface. It can be noticed that there is perfect agreement in the spreading dynamics regarding the timescale, which predicts a pulse duration of 17.2 ms that accurately matches the observed voltage pulse. Once the active area is covered by the droplet at t 3 , the power generation is zero as the spreading dynamic stops contributing to the harvesting by this individual Trecxel. After t 3 , the drop starts to recoil and slide from the hydrophobic Trecxel, which reduces the contact area of the water-solid interface. When the active area of the Trecxel starts to be reduced due to the recoiling and sliding process (between t 3 and t 4 ), a negative voltage signal is developed. The detachment process continues up to instance t 4 , when the overlapping area of the droplet with the active area of the Trecxel is almost zero, which is the last Please cite this article in press as: Ghaffarinejad et al., Triboelectric pixels as building blocks for microscale and large-area integration of drop energy harvesters, Device (2024), https://doi.org/10.1016/j.device.2024.100566 2Device 3, 100566, February 21, 2025 Article ll OPEN ACCESS instant when the device delivers a negative voltage. The droplet recoiling and sliding process before rebound is much slower than spreading, 22,24 and the output voltage is much smaller (and in most cases barely visible) than the generated at instance t 2 .Therefore, in this first approach dealing with highly hydrophobic surfaces, we will focus our attention onthe positive and instantaneous peak voltage generation during impact and spreading. After a complete separation between the drop and active electrodes at t 5 the drop bounces or slides away, and the output current drops to zero. As detailed in Note S1, this model mechanism includes not only the effect of the drop dynamics but also the dependency on the electrode dimensions and position with respect to the drop contact. 6 Figure 1C presents the circuit model of the Trecxel considering the density of surface charges (which develops a surface potential, V sp ), the variation in capacitance (C var ), and the effect of the drop conductivity (R drop ). The model is inspired by previous research works in the field of triboelectric nanogenerators 29,30 and is updated to include the dynamic behavior of Trecxel. In this model, V sp =Q s /C 0 , which is usually referred to as the surface potential of the dielectric layer, and Q s is the charge stored on the triboelectric surface (C). These surface charges can be injected by artificial (here we use an electron beam) or by natural 6,20 means (such as contact with the droplets or by friction with the wind). As detailed below, the injection of electrical charges into the triboelectric layer enhances the instantaneous conversion of energy. In our case, C 0 is the equivalent capacitance of the two dielectric layers (substrate and triboelectric layers in Figure 1A, C0=Sε0 d1 ε1 +d2 ε2, with thicknesses and relative dielectric constants of d 1 ,d 2 and e 1 ,e 2 , respectively, with an area of S=L 2 , and ε0is the dielectric permittivity in vacuum). C var is the time-dependent capacitance of the Trecxel existing between the two top active electrodes and the back electrode which changes between two values, minimum (C min ) and maximum (C max ) as the drop makes contact with Trecxel (Figures 1B and 1D). Q var is the variable charge of the C var with the relation of V var =Q var /C var,31,32 and R drop is proportional to the resistivity of the drop. The simulation results for the given dimensions of Figure 1. Trecxel design, working mechanism, and modeling (A) Schematic of the Trecxel actuated by a raindrop. (B) Dynamics of drop-Trecxel interactions and electricity generation time frame. Inset: comparison between theoretical (adapted from the analytical model by Pasandideh-Fard et al. described in Note S1) 28 and experimental voltage peak generation during spreading in a microscale Trecxel. (C) Electrical circuit model that embeds the effect of surface charges (V sp ), variation of Trecxel capacitance (C var ), and drop resistance (R drop ). (D) Simulation of capacitance variation during a complete cycle of electricity generation. See also Video S1. Please cite this article in press as: Ghaffarinejad et al., Triboelectric pixels as building blocks for microscale and large-area integration of drop energy harvesters, Device (2024), https://doi.org/10.1016/j.device.2024.100566 Device 3, 100566, February 21, 2025 3 Article ll OPEN ACCESS the Trecxel area concerning the droplet impact dynamic show that the C var value more than doubles in less than 20 ms(Figure 1D and Note S1, see also Figure S3). Instantaneous power generation in the Trecxel is highly affected by rapid variations in Trecxel capacitance. C var is the result of the coupling between top and back electrodes and can be defined as Cvar =ε0AeðtÞ d1 ε1 +d2 ε2=C0AeðtÞ S, where A e (t) is the effective overlapping area between the top and back electrodes as a function of time. The value of C var is at its minimum before the drop reaches the Trecxel and drop-electrode contact starts to form (Figures 1B and 1D), which is equally formulated as Cmin =C0Amin eðtÞ Swith Amin e=2LW (the overlapping area between the top and back electrodes). As contact forms between the drop and the electrodes, C var starts to increase until the drop fully covers the entire surface of Trecxel and fills the entire gap between the two electrodes (Figures 1B and 1D), which makes Cmax =C0Amax eðtÞ Swith Amax e=L2. Eventually, as the drop slips away from the Trecxel, the drop-electrode contact breaks apart, and C var reduces to C min again. It is worth stressing once again that the reduction in effective area when the droplet recoils and detaches from the surface is a much slower event that produces a reverse current in the load. In our first realization of the Trecxel, this current would be almost negligible compared to most of the previous D-TENGs reported in the literature given the microscale size of the electrodes. 7–10,12–15 Additionally, as we will show below, this negative branch will strongly depend on the wetting properties of the surface (wetting contact angle and slippery character of the surface). To investigate the effect of rapid variations of A e (t) on the output signal of the Trecxel, the dissipated power over a resistive load is calculated. When a resistive load R L is connected to the Trecxel and a water drop hits the surface, a current I L will pass through it and the instantaneous power of P=R L I L2 is dissipated by the load. Ignoring all sources of energy loss in the circuit, Pis equal to the power generated by the Trecxel. The passing current is dependent on C var and V var with the equation IL=CvardVvar dt +VvardCvar dt . 33,34 Considering the dimensions of the Trecxel on the microscale together with the two dielectrics used in our first realization (glass substrate and triboelectric perfluoroalkoxy alkane [PFA] sheet), C var magnitude would be in the range of a few femtofarads; therefore, by approximation the first term of the equation describing I L can be ignored compared to the second term producing IL=VvardCvar dt . The power is then calculated as PðtÞ=RLC0 2 S2Vvar2dAeðtÞ dt 2 :(Equation 1) This equation shows that the instantaneous output power is strongly dependent on the change of effective surface area with time and, consequently, with the capacitance variation as dCvar dt zdAeðtÞ dt . Therefore, the faster the change in effective area in time, the higher the instantaneous output power. Figure S2 shows a picture of a first physical realization of the triboelectric pixel concept (see Video S1, a slow-motion video showing drop velocity and voltage generated by a single Trecxel). As gathered in experimental procedures, this first Trecxel was produced by a simple shadow-mask processing of sputtered electrodes on a commercially available polymer. In this case, silver was utilized as a conductive material, but the architecture is equally compatible with the application of transparent conducting oxides or other transparent solutions with reduced dimensions as presented below. The Trecxel has a square top cross-section with dimensions of L= 900 mm, W= 200 mm, d 2 = 600 mm, d 1 =50mm, and h= 300 nm. In this example, the substrate is borosilicate glass covered by a triboelectric layer made of PFA. After the microscale assembly of the system, the surface of the triboelectric layer is charged using a beam of 20 kV/20 mA generated by a field-emission electron gun. This procedure provides a high density of stored electric charge on the surface (see experimental procedures) granted by the high electronegativity of fluorine atoms in the PFA polymer. 35,36 Such a Trecxel was used in the experiments to evaluate the energy generated by single drops with a volume in the range of 37 mL impinging on the triboelectric layer from a height on the order of 30 cm. Large droplets have been chosen to simulate a realistic ‘‘rain’’ scenario. 37,38 The graphs in Figure 2A depict the time evolution of the voltage (V Load vs. time) recorded at different optimal loads at 65 kU,82kU, and 260 kUfor rain, 60 mS/cm NaCl saltwater, and Milli-Q drops. In all cases, the instantaneous power generation is enough to light a commercially available light-emitting diode (LED) directly connected to the Trecxel, as recorded in Video S2.InFigure 2A, the duration of the voltage pulse is approximately 50 ms for rain and saltwater drops. This transition time is longer than presented in Figure 1B (see also Figure S3) obtained for a freshly charged surface. We attribute this effect to a lower surface charge in the steady state (see experimental procedures), which makes the initial charge induction in the drop significantly slower. In fact, the transition time is significantly longer (ca. 900 ms) for Milli-Q water drops (Figure 2A). Since the dynamics of the drop impact on the surface should not be modified by the different water drops used here, we attribute such a longer transition time to the low conductivity of the Milli-Q water drops, which affect the charge induction time drastically. The maxima at optimal loads correspond to harvested peak powers (and peak instantaneous output power density) of 134 mW (scaled to 174 W/m 2 ) for salty water, 61.5 mW (76 W/m 2 ) for rainwater, and 3.96 mW (4.9 W/m 2 ) for Milli-Q water (considering an active area of 900 3900 mm 2 )(Figure 2B). Such a huge difference in power density for the case of saltwater and rainwater compared to Milli-Q water drops is related again to the much higher conductivity (60 mS/cm, 10 mS/cm, 3–6 mS/cm for salty, rain, and Milli-Q water, respectively) of the former two compared to the latter (see finite element modeling and simulations). Moreover, the energy provided for a single Trecxel is on the order of 1 nJ (1.36, 1.04, and 0.47 nJ for salty, rain, and Milli-Q water, respectively). Although this energy is low compared to the potential energy of the droplet (which falls from a height of 30 cm with a mass of 37 mg, EP=m$g$h= 109 mJ), as will be shown below, the optimizations developed in this article bring the efficiency to competitive state-of-the-art values for drop energy harvesting. Please cite this article in press as: Ghaffarinejad et al., Triboelectric pixels as building blocks for microscale and large-area integration of drop energy harvesters, Device (2024), https://doi.org/10.1016/j.device.2024.100566 4Device 3, 100566, February 21, 2025 Article ll OPEN ACCESS Higher conductivity of non-pure water drops allows for rapid exchange of electrical charges between the active electrodes and the drop and faster change of the Trecxel capacitance value. Thus, experiments carried out with nonpolar, nonconductive liquids such as diiodomethane did not produce any power generation. Figure S4 shows the output characteristic of a control device with single top and bottom contacts on the millimeter to centimeter scale. This architecture has recently been at the center of attention by researchers working on the development of solid-liquid TENGs as an early example of instantaneous power generation by this type of energy harvester. 5,6,15,21,39 The experimental methods and materials used for the fabrication and test of the control device are the same as those used for the Trecxel. The instantaneous power obtained with the single-top-electrode configuration for salty, rain, and Milli-Q water is 5.97 mW(R L = 82 kU), 1.67 mW(R L =65kU), and 0.06 mW(R L = 260 kU), respectively, corresponding to a scaled power density of 9.9 mW/m 2 , 2.7 mW/m 2 , and 105 mW/m 2 for an active area of 600 mm 2 . Compared to the above results for the Trecxel architecture, the performance of our system improves the instantaneous power density of the control device to 1.75 310 4 , 2.8 310 4 , and 4.6 310 4 for salty, rain, and Milli-Q water, respectively. The calculated ratios indicate an improvement of more than four orders of magnitude in the output power density when a microscale Trecxel is used for harvesting the energy of a single drop instead of a device with millimeter-scale (top-electrode) and centimeter-scale (bottom electrode and triboelectric layer) dimensions. Moreover, the transition time of the three output curves shown in Figure S4 is in the same range, varying between 2 ms and 5 ms, whereas in the case of the Trecxel (Figure 2A) it is on the order of a few tens of microseconds and a few milliseconds for the conductive (saltwater and rainwater) and Milli-Q water drops, respectively. These results and the analysis carried out for different types of salts below show that the Trecxel device is extremely sensitive to the conductivity of the drops, an operational advantage for the development of self-powered water sensors compared to the reported single-top-electrode devices. 5,6,15,21,39 Furthermore, these comparative short times for power generation and the reduced dimensions of the device allow for a high-frequency harvesting response to multiple drop events, as shown in Video S3, where the peak voltage generated by sequential drops can be followed, depicting high voltage values after successive events. As mentioned, the impact dynamics cannot explain the differences in the output signal for the different water types employed. Thus, we investigated the underlying mechanisms that govern the relationship between voltage peaks and the conductivity of droplets, employing finite element modeling. Specifically, we conducted simulations using the electrostatics module within the COMSOL Multiphysics package (for detailed methodology, please refer to Figures S5 and S6 and experimental procedures). These simulations involve a 3D model of the substrate/droplet system (Figure 3). The shape of the drop is described as a 3D expanding ‘‘pancake,’’ thereby increasing its contact surface while thinning to keep its volume (37 mL) constant. To describe approximately the effect of the tilted surface of the Trecxel, the center of the contact surface moves along the surface at a constant speed (vslidez0:5 m/s, see supplemental information) during the simulation. In these simulations, the drop and air media are modeled as a single domain and the difference between both is established by a position-dependent relative permittivity εðx;y;zÞ(Figure 3A). Electrostatic simulations are performed under open-cir- cuit (OC) conditions, where the bottom electrode is grounded while the top electrodes are left at floating potential. On uncharged surfaces, contact electrification between the pure water drop (Milli-Q in the experiments) and the surface may provide almost 90% of charges by an electron-transfer process from water molecules to the polymer surface, while the remaining 10% relies on ion transfer. 40 On the contrary, the dominant process is ion transfer in the case of charged surfaces. 40 In our case, the PFA surface has previously been negatively charged within the region between the top electrodes. Additionally, we have considered that the rest of the PFA surface is slightly residually charged due to the interaction with the environment and the manipulation of the device. Consequently, the primary mechanism taken into account in our model is the screening of the negative charges on the PFA surface via the transfer of positive ions from the drop to the surface in contact with it (see Figure S5). Specifically, we have assumed a position-dependent surface Figure 2. Characterization of the Trecxel (A) Trecxel output voltage (L=900mm, W= 200 mm, d 2 = 600 mm, d 1 =50mm, and h= 300 nm) in three different optimal resistive loads for different water drops, namely raindrop, salty, and Milli-Q water. (B) Maximum instantaneous power-load graph showing the output power of the Trecxel for the different water drops and resistive loads. Please cite this article in press as: Ghaffarinejad et al., Triboelectric pixels as building blocks for microscale and large-area integration of drop energy harvesters, Device (2024), https://doi.org/10.1016/j.device.2024.100566 Device 3, 100566, February 21, 2025 5 Article ll OPEN ACCESS charge density, Qsðx;yÞ, equal to 50 mC/m 2 within a small rectangular region between the top electrodes (due to the electronbeam charging process) and 0.8 mC/m 2 in the rest of the PFA surface (related to parasitic/residual charges at the surface) (Figure 3A). 29 These values are lower and higher than the surface charge density measured experimentally for the bare PFA in the active area before and just after the charging process (0.3 and 95 mC/m 2 ), respectively. We have assumed these higher/ lower charge density values considering that the uncharged/ charged PFA surface will naturally charge/discharge with the impact of droplets, as described below for the Trecxel device case. Following our hypothesis, the accumulation of positive Figure 3. 3D model of the working mechanism of a single Trecxel (A) Geometry of the 3D model implemented in COMSOL and surface charge density distribution around the top electrodes, the latter showing the evolution of the drop from t= 0 (when the drop just touches the surface at a single point) to the instant when the drop reaches its maximum diameter (t max ). Specific details are given in supplemental information and experimental procedures. (B) Open-circuit voltages as a function of time for three values of NaCl concentration, corresponding to the cases of salty, raindrop, and Milli-Q water cases. (C) Load voltages as a function of time for the same three types of drops. The load resistances used in the simulations of each case are the optimal obtained experimentally. The inset shows a zoom-in of the spreading time region, where the experimental curves have been added for comparison. (D) Load voltages as a function of the surface charge density for three types of drops. (E) Dependency of the load voltage with the triboelectric layer thickness. Please cite this article in press as: Ghaffarinejad et al., Triboelectric pixels as building blocks for microscale and large-area integration of drop energy harvesters, Device (2024), https://doi.org/10.1016/j.device.2024.100566 6Device 3, 100566, February 21, 2025 Article ll OPEN ACCESS Figure 4. Trecxel arrays on rigid and flexible configurations (A) Schematic of a ‘‘rain panel’’ that includes an n3marray of Trecxels. (B) Schematic of 1 32 array of Trecxels actuated by a single raindrop. (C) Output voltage of the two Trecxels connected to a load of 65 kUand activated using a raindrop (Trecxels charged and on-supported configuration). (D) Schematic of 2 32 array of Trecxels actuated by a single raindrop. (E) Optical image of the 2 32 array used in the experiment in a flexible self-supported configuration. (legend continued on next page) Please cite this article in press as: Ghaffarinejad et al., Triboelectric pixels as building blocks for microscale and large-area integration of drop energy harvesters, Device (2024), https://doi.org/10.1016/j.device.2024.100566 Device 3, 100566, February 21, 2025 7 Article ll OPEN ACCESS charges in the drop as it expands on the PFA surface may originate from two sources. (1) Accumulation of Na + ions at the PFA/ droplet interface: a fraction of Na + ions, the closest to the double layer, will tend to diffuse to the interface. This diffusion process will occur within the Debye time, tNa, whose value depends on the concentration of NaCl. (2) Accumulation of H + protons, with Debye time tH. The time dependence of these accumulated positive charges is assumed to be similar in shape as for the case of a charging (parallel plate) capacitor, i.e., 1 ð1fNaÞet=tH fNaet=tNa (where fNa is the fraction of Na + closest to the interface that participates in the screening of negative charges on the surface). However, the Debye time for protons, tH, is around four orders of magnitude smaller than tNa, so one can assume ð1fNaÞet=tHz0. Finally, the expression used to quantify the number of positive charges within the drop during contact with the surface is QðtÞ=Qsðx;yÞScðtÞ1fNaet tNa ;(Equation 2) where ScðtÞis the time-dependent contact area between the drop and the PFA surface (i.e., excluding the electrodes region). This expression guarantees that, for t/N, the charges accumulated on the contact surface of the drop will fully compensate for the negative charges on the PFA surface. In this model, the positive charges within the drop are located only at the drop surface in contact with the PFA. For the simulations shown here, we have considered the spreading and recoiling stages up to a maximum time (see panel 8inFigure S5). This is the time when the drop surface contact area returns to zero (see t 4 in Figure 1B). Figure 3B shows the OC voltage, VOC, as a function of time considering the indicated NaCl concentrations diluted in water simulating the three types of water utilized in the experiments. The drop of the OC potential is evidence of the progressive screening of the negative charges on the surface of the PFA. Note also on the close-up inset of Figure 3C how the term et tNa allows us to reproduce a delay in the screening as time (and contact surface) progresses. As mentioned above, this delay in the spreading process of the droplet manifests itself as a decrease in the peak voltage and an important increase in the width of the signal. Hence, Figure 3C shows the voltage peaks obtained from the simulations by considering resistive loads (R L ) for salty, raindrop, and Milli-Q water, respectively. The inset shows a zoom-in of the spreading period, where an excellent agreement with the experimental voltage signals can be noted, which have been added for comparison. This fair agreement of the model and the experimental results allows us to analyze the effect of different features on the voltage-generation mechanism. On the one hand, we have artificially charged the triboelectric surface between the two top electrodes with an e-beam. However, the voltage peak generation is compatible with the surface homogeneous charging with a lower charge density, mimicking the natural charging of the surface by natural means (e.g., manipulation, friction with air, or drops). In Figure 3D, we compare the results obtained for these two situations. It is worth noting that the dependence on the type of water is maintained regardless of the surface charging, although the peak intensities decrease for the lowcharged Trecxel. In this comparison, we have maintained the optimal load resistance for both charging conditions, which might vary slightly, as we will see experimentally in the following section. On the other hand, the power-conversion mechanism proposed for the Trecxel is based on capacitance variation, which makes the system dependent on the triboelectric layer thickness. Simulations in Figure 3E demonstrated the dependence on the thickness for three different PFAs, ranging from 25 to 100 mm, showing that the lower the thickness, the higher the load voltage obtained. This enhancement can be easily understood taking into account that the output voltage can be expressed as (following the same considerations that led to Equation 1): VðtÞ=RLC0Vvar SdAeðtÞ dt . Considering that Vvar and dAeðtÞ dt are not altered by the difference in the thickness, and since C0=Sε0 d1 εd1 +d2 εd2, the decrease in d 1 explains the increase in the C 0 and, consequently, the enhancement of the output voltage. First steps toward high-density power conversion by Trecxel arrays Referring to Figure 2, the selected dimensions for the design and fabrication of the Trecxel (such as Land W) and the number of top electrodes are limited to the applied mask fabrication process. Thus, it is important to stress that the Trecxel energy-har- vesting mechanism is compatible with a design to integrate the smallest possible active area so that a single drop covers the whole surface of the device at once upon contact or, ultimately, several Trecxels at the same time. Indeed, to push forward the power delivery of drop energy harvesters and to address their integration in large-area devices, it is critical to demonstrate an approach for energy conversion from multiple events, which in this context involves multiple drops and drop impacts, as it would be required to harvest the energy from rainfall. A straightforward approach for large-scale energy harvesting from multiple water drops is to form a densely packed n3marray of Trecxels as schematized in Figure 4A. Figure 4B presents the schematic of a 1 32 array of Trecxels packed together with a gap distance of g= 200 mm and d= 2 mm. The materials used in the fabrication of the array, experimental conditions, and the dimensions of Trecxels are the same as those used for a single Trecxel in Figure 2.Figure 4C shows the output characteristics of each Trecxel connected to a load of R L =65kU, produced by the impact of a raindrop. The temporal dynamics of the (F) Output voltage of the four Trecxels in (E), each of them connected to a load of 10 MUand activated by Milli-Q, rain, and salty droplets of 37 mL. (G) Comparison of the energy per droplet obtained as a function of the load for different droplet heights impacting a centered self-supported uncharged Trecxel. The inset shows the response for droplets impacting near the Trecxel. (H) Dependence of the energy conversion on the number of top electrodes of the devices. (I) Comparison of the output voltage of the four different Trecxels, each of them connected to a load of 10 MUand activated by a Milli-Q droplet (top) and the corresponding output for the four Trecxels connected in parallel (bottom). Please cite this article in press as: Ghaffarinejad et al., Triboelectric pixels as building blocks for microscale and large-area integration of drop energy harvesters, Device (2024), https://doi.org/10.1016/j.device.2024.100566 8Device 3, 100566, February 21, 2025 Article ll OPEN ACCESS PFA sheet was used as the substrate, directly supporting the electrodes on both sides. For the case of the thin-film-compatible Trecxel, fabrication started from a double-sided polished intrinsic silicon wafer or fused silica slips, which were cut into a 20 mm 320 mm square and ultrasonically cleaned with ethanol and then acetone. Next, a thin layer of silicon oxide was deposited by PECVD at room temperature following the protocol shown elsewhere. 49 Chlorotrimethylsilane from Sigma-Aldrich was used as the delivered Si precursor. The oxygen and Si precursor were supplied in the reactor using a mass flow controller. The base pressure of the chamber was lower than 2 310 5 mbar, and the total pressure during deposition was around 10 2 mbar. The plasma was generated in a 2.45-GHz microwave electron-cyclotron resonance SLAN-II operating at 800 W, and the deposition was carried out in the downstream region of the reactor. The thin-film thickness was settled at 300 nm. After the deposition, chemical derivatization of the surface was carried out by exposing the sample at 80C to 1H,1H,2H-perfluorooctyltrietoxisy- lane (PFOTES) 98% vapor on a static low vacuum (obtained with a rotatory pump) for at least 2 h. 52 Thus, in this case, the silicon plays the role of the substrate, and the perfluorinated silicon oxide plays the role of the triboelectric layer. The bottom and top electrodes were produced by masking and metal sputtering. First, non-adhesive Kapton films or stainless-steel sheets were patterned with the geometry of the top and back electrodes through a laser ablation process (ExplorerOne, 355 nm wavelength). Silver electrodes were deposited over the triboelectric layers by DC sputtering through both aligned top and bottom masks. This deposition was carried out in an argon atmosphere with a total pressure of 1.6 310 1 mbar and a voltage of 600 V for plasma generation. The base pressure of the chamber was lower than 10 3 mbar. The deposition time was 2 h for each electrode to obtain fully conductive layers with a thickness of ca. 300 nm. Gold-titanium electrodes Figure 9. Thin-film Trecxel configuration (A) Photograph of the thin-film devices for four Trecxels for ITO (transparent) electrodes. (B) Voltage output obtained for the impact of multiple Milli-Q droplets on four different Trecxels fabricated on fused silica. (C) Voltage output obtained for the impact of a raindrop on a single Trecxel fabricated on intrinsic Si. were deposited by evaporation. In a vacuum chamber (total pressure lower than 10 5 mbar), a thin film of 2–3 nm titanium (Ti) was deposited directly by evaporation on the samples through the shadow mask. Just after deposition, gold (Au) evaporation was performed from another evaporator in the same chamber. Finally, the 80-nm gold thin film was deposited on top of the previous titanium layer through the same shadow mask. ITO thin films as bottom and top electrodes were deposited by RF magnetron sputtering using a 200 ITO target (indium-tin oxide 90–10 wt %), 99.99% pure, from Testbourne. The base pressure of the chamber before deposition was lower than 5 310 6 mbar, and a working pressure of 5 310 3 mbar of pure Ar was employed. The substrates were heated up to 350C, which was kept constant during deposition. The power applied to the target was 50 W, and the distance between the target and the substrates was 10 cm. Setup for measurements and characterizations For all experiments except the characterization at different heights in Figure 4, water drops are released from a distance of between 30 and 50 cm by a 5-mL polyethylene pipette onto the Trecxel surface. The pipette dispenses water drops with a volume of 37 mL. The pipette is gripped and rigidly fixed by a laboratory stand and clamps. Trecxel is fixed onto a tilted flatbed, which makes a 30tilting angle with the horizon. The output voltage of the Trecxel is measured using a Tektronix 3024 oscilloscope. A high impedance probe with 10 MUof internal resistance is used to measure the voltage across different loads. 53 Measurements were performed in a lab environment with humidity of 65% and temperature of 20C. For Figures 2,3,4B, and 6C, the rainwater was collected on April 26 and 27, 2021 in Seville, Spain. For the rest of figures, the rainwater was collected on May 18, 2023 in Seville, Spain. Rainwater conductivity was around 6 mS/cm. Salty water was prepared as a NaCl solution in Milli-Q water with a concentration of 0.585 g of NaCl per 100 mL of water to yield a conductivity of 60 mS/cm. For the experiments in Figures 4,5,6,7,8, and 9, the load impedances were changed using a resistance box, and the Trecxel output voltage was measured using a PicoScope oscilloscope. Energy and peak power conversion vs. load curves were characterized from high to low load-resis- tance values. Sodium chloride, potassium chloride, and ammonium chloride (all from Sigma-Aldrich) were used to obtain the different solutions we used to test the Trecxels under different liquid conductivity conditions. To prepare the high-conductivity solutions, we mixed the salt with Milli-Q water to control the weight of the salt, volume of the solution, and conductivity. Water conductivity measures were performed using a conductivity probe, 731-ISM Mettler Toledo InLab, as the conductivity module for the Mettler Toledo SevenExcellence multiparameter. Measurement of water contact angle A WCA measurement system (OCA20 goniometer from DataPhysics) was used to measure and compare the hydrophobicity of the PFA and SiO 2 surfaces with respect to raindrops (extended data in Figure 1). The WCA system Please cite this article in press as: Ghaffarinejad et al., Triboelectric pixels as building blocks for microscale and large-area integration of drop energy harvesters, Device (2024), https://doi.org/10.1016/j.device.2024.100566 Device 3, 100566, February 21, 2025 15 Article ll OPEN ACCESS consists of a peristaltic pump (Perimax 12/4-SM from Spetec) to control the water flow and a set of calibrated micropipette tips that ensure a certain size of the falling droplets. Surface charging protocol The triboelectric layers of the Trecxels were charged using an electron beam in a scanning electron microscope. All samples were exposed for 10 min to a beam acceleration voltage and current settled at 20 kV and 20 mA, respectively, over the active area of the Trecxel (0.9 30.9 mm 2 ). Recently charged surfaces present abrupt voltage curves. To have a more realistic scenario, the surfaces (otherwise stated) were measured several hours after charging and impacting with more than 100 droplets to reach a steady state. Surface electrostatic charges were characterized by employing a non-contacting electrostatic voltmeter (ESVM 1000, Wolfgang Warmbier) with an active area of 40 mm 2 , measured at 10 mm from the surface. Before the charging procedure, the measured potential is (mean value) 33 V, corresponding to a surface charge density of s uncharged = 0.3 mC/m 2 . After the charging protocol for the device with a single Trecxel, the potential is increased to 270 V. If we assume that the charge is located in the active area defined by the beam used in the microscope of 0.9 30.9 mm 2 , and considering that the surface charge density out of this square is equal to the charge density measured before the charging protocol, the surface charge density of the active area is s charged =96mC/m 2 . For the two-Trecxels device, the measured potential before and after the charging protocol results in an increment from 61 V to 177 V. These values correspond to a surface charge density before (also outside the active area) and after the charging protocol (in the active area, which now is 0.9 3 2mm 2 )ofs uncharged = 0.5 mC/m 2 and s charged =21mC/m 2 . For the four-Trecxels PFA flexible device, the measured potential before charging was 22 V, and it increased to 750 V, which corresponds to a surface charge density before and after the charging protocol (in the active area, which now is 4 30.9 3 0.9 mm 2 )ofs uncharged = 0.2 mC/m 2 and s charged =61mC/m 2 . Calculation of average and instantaneous power densities and energy per droplet Assuming that V Load is the measured voltage across a resistive load R L , the instantaneous power is calculated as PLðtÞ=V2 LðtÞ=RL, where Pmax Lis the maximum value of PLðtÞ. Assuming that the active area of the drop energy harvester is A, the maximum power density is defined as Pmax L=A. Proprietary software (NanoDataLyzer) developed by the authors facilitated statistical analysis of the maximum power conversion and energy per droplet. This software is available via open access in Zenodo. 53,54 Each point in the graphs in Figures 4,5,6,7, and 8has been obtained from an average of at least 16 droplet impacts. COMSOL modeling of the Trecxel The V Load was obtained from V OC and C var by numerical simulations using the electrical circuit module of the COMSOL package. We have found that to reproduce the experimental load voltages, it is necessary to consider a maximum capacitance at least one order of magnitude larger than the one obtained theoretically by geometric considerations. The origin of this mismatch is unclear, but previous works 54,55 point out the imperfections and roughness of the dielectric/electrode surfaces as the main origin of the increase in the experimentally observed capacitance with respect to what can be estimated theoretically from the geometry. It is important to note that the dependence of this increase in Cvar on roughness is nonlinear, so it seems reasonable to find increases of up to two orders of magnitude due to roughness and imperfections, electrodes edge effects, and so forth, on the surface in combination with the dielectric characteristics of the water drop. Assuming these features as the origin of the capacitance mismatch, we have multiplied our theoretical capacitance by a factor Nc.Wealsoconsideredtheoverlapbetweenthedrop+topelectrodes and all the bottom electrodes in the simulation of the time-dependent capacity (see Figure S6). Additionally, we have included the possibility of a ‘‘parasitic capacitance,’’ Cp, in the equivalent TENG circuit (see Figure S9). According to Dai et al., 56 these parasitic capacitances are unavoidable and usually range between a few picofarads and hundreds of picofarads. These are assumed to come from a combination of sources such as external metal wires, capacitances in the measurement equipment, and external circuit boards. Since the parasitic capacitance degrades the average power for all values of the load resistor, its main effect is to reduce the optimal load but also reduce the power. In addition to the electrodes, the modeled Trexcel consists of three layers PFA/ Kapton/borosilicate with thicknesses of 25, 50, or 100 mm/70 mm/130 mm and relative permittivities of 2.1/4/4.8. RESOURCE AVAILABILITY Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Ana Borra ´s, anaisabel. [email protected]. Materials availability dThis study did not generate new unique reagents. Data and code availability dThe dataset of this article has been deposited at digital.csic under the DOI https://doi.org/10.20350/digitalCSIC/16536 and is publicly available as of the publication date. dAll original code for the NanoDataLyzer has been deposited at Zenodo under the DOI https://doi.org/10.5281/ZENODO.13225385. dAny additional information required to reanalyze the data reported in this paper is available from the lead contact upon request ACKNOWLEDGMENTS The authors thank the projects PID2022-143120OB-I00, PCI2024-153451, and TED2021-130916B-I00 funded by MCIN/AEI/10.13039/501100011033 and by ‘‘ERDF (FEDER) A way of making Europe, Fondos NextgenerationEU and Plan de Recuperacio ´n, Transformacio ´n y Resiliencia.’’ C.L.-S. thanks the University of Seville through the VI PPIT-US and ‘‘Ramo ´n y Cajal’’ program funded by MCIN/AEI/10.13039/501100011033. J.R.S.-V. acknowledges project ANGSTROM selected in the Joint Transnational Call 2023 of M-ERA.NET 3, an EU-funded network of about 49 funding organizations (Horizon 2020 grant agreement no. 958174). X.G.-C. acknowledges the FPU program under grant number FPU19/01864. FNG acknowledges the ‘‘VII Plan Propio de Investigacio ´n y Transferencia’’ of the Universidad de Sevilla. The project leading to this article has received funding from the EUH2020 program under grant agreement 851929 (ERC Starting Grant 3DScavengers). AUTHOR CONTRIBUTIONS X.G.-C. fabricated all the devices with support from F.N.-G. and V.G. A.G. carried out the characterization of the liquid-solid TENG for single-droplet events of precharged devices. X.G.-C. and F.N.-G. carried out the statistical analysis and characterization of the devices. X.G.-C. and J.R.S.-V. were responsible for the simulations and the drop dynamics vs. power analysis. A.G. is responsible for Equation 1. J.B. carried out the simulations and modeling by FEM. C.L.-S. assisted in the contact angle and device characterization. J.R.S.-V., A. Barranco, and A. Borra ´s devised the conceptualization. A.G. and A. Borra ´s wrote the original draft of the article. All authors contributed to the final version of the manuscript. DECLARATION OF INTERESTS A.G.’s current affiliation is Hanzle UAS, the Netherlands. The content of this article is related to two patent applications: (1) A.G., X.G.-C., A. Barranco, J.R.S.-V., A. Borra ´s: WO 2023/025965 A1. (2) X.G.-C., A. Barranco, J.R.S.-V., A.G., A. Borra ´s: PCT/EP2024/053180. SUPPLEMENTAL INFORMATION Supplemental information can be found online at https://doi.org/10.1016/j. device.2024.100566. Please cite this article in press as: Ghaffarinejad et al., Triboelectric pixels as building blocks for microscale and large-area integration of drop energy harvesters, Device (2024), https://doi.org/10.1016/j.device.2024.100566 16 Device 3, 100566, February 21, 2025 Article ll OPEN ACCESS Received: June 14, 2024 Revised: August 8, 2024 Accepted: September 6, 2024 Published: October 3, 2024 REFERENCES 1. Laajimi, M., and Go, Y.I. (2021). Energy storage system design for largescale solar PV in Malaysia: techno-economic analysis. Renewables: Wind. Water, and Solar 8, 1–23. https://doi.org/10.1186/s40807-020- 00064-5. 2. Zhang, Y., Ren, J., Pu, Y., and Wang, P. (2020). Solar energy potential assessment: A framework to integrate geographic, technological, and economic indices for a potential analysis. Renew. Energy 149, 577–586. https://doi.org/10.1016/j.renene.2019.12.071. 3. Hao, S., Kuah, A.T.H., Rudd, C.D., Wong, K.H., Lai, N.Y.G., Mao, J., and Liu, X. (2020). A circular economy approach to green energy: Wind turbine, waste, and material recovery. Sci. Total Environ. 702, 135054. https://doi. org/10.1016/j.scitotenv.2019.135054. 4. Trovato, V., Conenna, D., Dicorato, M., Forte, G., and Trovato, M. (2020). Economic and financial benefits for wind turbines providing frequency response exploiting the kinetic energy or operating part-loaded. IET Generation, Transmission and Distribution 14, 4371–4387. https://doi.org/10. 1049/iet-gtd.2020.0010. 5. Wu, H., Mendel, N., Van Den Ende, D., Zhou, G., and Mugele, F. (2020). Energy Harvesting from Drops Impacting onto Charged Surfaces. Phys. Rev. Lett. 125, 078301. https://doi.org/10.1103/PhysRevLett.125.078301. 6. Xu, W., Zheng, H., Liu, Y., Zhou, X., Zhang, C., Song, Y., Deng, X., Leung, M., Yang, Z., Xu, R.X., et al. (2020). A droplet-based electricity generator with high instantaneous power density. Nature 578, 392–396. https:// doi.org/10.1038/s41586-020-1985-6. 7. Lin, Z.-H., Cheng, G., Lee, S., Pradel, K.C., and Wang, Z.L. (2014). Harvesting Water Drop Energy by a Sequential Contact-Electrification and Electrostatic-Induction Process. Adv. Mater. 26, 4690–4696. https://doi. org/10.1002/adma.201400373. 8. Zhang, N., Gu, H., Lu, K., Ye, S., Xu, W., Zheng, H., Song, Y., Liu, C., Jiao, J., Wang, Z., and Zhou, X. (2021). A universal single electrode dropletbased electricity generator (SE-DEG) for water kinetic energy harvesting. Nano Energy 82, 105735. https://doi.org/10.1016/j.nanoen.2020.105735. 9. Tang, W., Chen, B.D., and Wang, Z.L. (2019). Recent Progress in Power Generation from Water/Liquid Droplet Interaction with Solid Surfaces. Adv. Funct. Mater. 29, 1901069. https://doi.org/10.1002/adfm. 201901069. 10. Lee, J.H., Kim, S., Kim, T.Y., Khan, U., and Kim, S.-W. (2019). Water droplet-driven triboelectric nanogenerator with superhydrophobic surfaces. Nano Energy 58, 579–584. https://doi.org/10.1016/j.nanoen.2019. 01.078. 11. Yuan, Z., and Guo, L. (2024). Recent advances in solid–liquid triboelectric nanogeneratortechnologies, affecting factors, and applications. Sci Rep 14, 10456. https://doi.org/10.1038/s41598-024-60823-y. 12. Jiang, D., Xu, M., Dong, M., Guo, F., Liu, X., Chen, G., and Wang, Z.L. (2019). Water-solid triboelectric nanogenerators: An alternative means for harvesting hydropower. Renew. Sustain. Energy Rev. 115, 109366. https://doi.org/10.1016/j.rser.2019.109366. 13. Rodrigues, C.R., Alves, C.A., Puga, J., Pereira, A.M., and Ventura, J.O. (2016). Triboelectric driven turbine to generate electricity from the motion of water. Nano Energy 30, 379–386. https://doi.org/10.1016/j.nanoen. 2016.09.038. 14. Lin, Z.-H., Cheng, G., Lin, L., Lee, S., and Wang, Z.L. (2013). Water-Solid Surface Contact Electrification and its Use for Harvesting Liquid-Wave Energy. Angew. Chem. Int. Ed. Engl. 52, 12545–12549. https://doi.org/ 10.1002/anie.201307249. 15. Niu, J., Xu, W., Tian, K., He, G., Huang, Z., and Wang, Q. (2020). Triboelectric Energy Harvesting of the Superhydrophobic Coating from Dropping Water. Polymers 12, 1936. https://doi.org/10.3390/polym12091936. 16. Zhou, J., Tao, Y., Liu, W., Sun, H., Wu, W., Song, C., Xue, R., Jiang, T., Jiang, H., and Ren, Y. (2021). Self-powered AC electrokinetic microfluidic system based on triboelectric nanogenerator. Nano Energy 89, 106451. https://doi.org/10.1016/j.nanoen.2021.106451. 17. Chung, J., Heo, D., Cha, K., Lin, Z.-H., Hong, J., and Lee, S. (2021). A portable device for water-sloshing-based electricity generation based on charge separation and accumulation. iScience 24, 102442. https:// doi.org/10.1016/j.isci.2021.102442. 18. Wang, Y., Gao, S., Xu, W., and Wang, Z. (2020). Nanogenerators with Superwetting Surfaces for Harvesting Water/Liquid Energy. Adv. Funct. Mater. 30, 1908252. https://doi.org/10.1002/adfm.201908252. 19. Yu, J., Ma, E., and Ma, T. (2017). Harvesting energy from low-frequency excitations through alternate contacts between water and two dielectric materials. Sci. Rep. 7, 17145. https://doi.org/10.1038/s41598-017- 17522-8. 20. Xu, X., Li, P., Ding, Y., Xu, W., Liu, S., Zhang, Z., Wang, Z., and Yang, Z. (2022). Droplet energy harvesting panel. Energy Environ. Sci. 15, 2916– 2926. https://doi.org/10.1039/D2EE00357K. 21. Li, Z., Cao, B., Zhang, Z., Wang, L., and Wang, Z.L. (2023). Rational TENG arrays as a panel for harvesting large-scale raindrop energy. iEnergy 2, 93–99. https://doi.org/10.23919/IEN.2023.0015. 22. Kumar, S., Namura, K., Suzuki, M., and Singh, J.P. (2021). Water droplet bouncing on a non-superhydrophobic {Si} nanospring array. Nanoscale Adv. 3, 668–674. https://doi.org/10.1039/D0NA00544D. 23. Guan, Y., Fu, J., Wu, S., Chen, X., and Zhou, C. (2021). The post-impact dynamics of drop rebound on inclined hydrophobic surfaces of various wettabilities. Phys. Fluids 33, 72108. https://doi.org/10.1063/5.0048805. 24. Yun, S. (2017). Bouncing of an ellipsoidal drop on a superhydrophobic surface. Sci. Rep. 7, 17699. https://doi.org/10.1038/s41598-017-18017-2. 25. Wang, H., Liu, C., Zhan, H., and Liu, Y. (2019). Droplet Asymmetric Bouncing on Inclined Superhydrophobic Surfaces. ACS Omega 4, 12238–12243. https://doi.org/10.1021/acsomega.9b01348. 26. Lv, C., Hao, P., Zhang, X., and He, F. (2016). Drop impact upon superhydrophobic surfaces with regular and hierarchical roughness. Appl. Phys. Lett. 108, 141602. https://doi.org/10.1063/1.4945662. 27. Josserand, C., and Thoroddsen, S.T. (2016). Drop Impact on a Solid Surface. Annu. Rev. Fluid Mech. 48, 365–391. https://doi.org/10.1146/annurev-fluid-122414-034401. 28. Pasandideh-Fard, M., Qiao, Y.M., Chandra, S., and Mostaghimi, J. (1996). Capillary effects during droplet impact on a solid surface. Phys. Fluids 8, 650–659. https://doi.org/10.1063/1.868850. 29. Ghaffarinejad, A., and Yavand Hasani, J. (2019). Modeling of triboelectric charge accumulation dynamics at the metal–insulator interface for variable capacitive structures: application to triboelectric nanogenerators. Appl. Phys. A 125, 259. https://doi.org/10.1007/s00339-019-2495-y. 30. Niu, S., Wang, S., Lin, L., Liu, Y., Zhou, Y.S., Hu, Y., and Wang, Z.L. (2013). Theoretical study of contact-mode triboelectric nanogenerators as an effective power source. Energy Environ. Sci. 6, 3576. https://doi.org/10. 1039/c3ee42571a. 31. Hinchet, R., Ghaffarinejad, A., Lu, Y., Hasani, J.Y., Kim, S.-W., and Basset, P. (2018). Understanding and modeling of triboelectric-electret nanogenerator. Nano Energy 47, 401–409. https://doi.org/10.1016/j.nanoen.2018. 02.030. 32. Ghaffarinejad, A., Hasani, J.Y., Hinchet, R., Lu, Y., Zhang, H., Karami, A., Galayko, D., Kim, S.-W., and Basset, P. (2018). A conditioning circuit with exponential enhancement of output energy for triboelectric nanogenerator. Nano Energy 51, 173–184. https://doi.org/10.1016/j.nanoen.2018. 06.034. Please cite this article in press as: Ghaffarinejad et al., Triboelectric pixels as building blocks for microscale and large-area integration of drop energy harvesters, Device (2024), https://doi.org/10.1016/j.device.2024.100566 Device 3, 100566, February 21, 2025 17 Article ll OPEN ACCESS 33. Wang, Y., Yang, Y., and Wang, Z.L. (2017). Triboelectric nanogenerators as flexible power sources. NPJ Flex. Electron. 1, 10. https://doi.org/10. 1038/s41528-017-0007-8. 34. Basset, P., Blokhina, E., and Galayko, D. (2016). Electrostatic Kinetic Energy Harvesting (Wiley). 35. Zhang-Fu, X., Hai, D., Guo-Mao, Y., Ting-Ji, L., and Sun, X.-M. (1991). Constant-current corona charging of Teflon PFA. IEEE Trans. Elect. Insul. 26, 35–41. https://doi.org/10.1109/14.68224. 36. Xia, Z. (1991). Corona charging and charge decay of Teflon-PFA. IEEE Trans. Elect. Insul. 26, 1104–1111. https://doi.org/10.1109/14.108147. 37. Porcu `, F., D’Adderio, L.P., Prodi, F., and Caracciolo, C. (2013). Effects of Altitude on Maximum Raindrop Size and Fall Velocity as Limited by Collisional Breakup. J. Atmos. Sci. 70, 1129–1134. https://doi.org/10.1175/ JAS-D-12-0100.1. 38. Villermaux, E., and Eloi, F. (2011). The distribution of raindrops speeds. Geophys. Res. Lett. 38.https://doi.org/10.1029/2011GL048863. 39. Wu, H., Mendel, N., van der Ham, S., Shui, L., Zhou, G., and Mugele, F. (2020). Charge Trapping-Based Electricity Generator (CTEG): An Ultrarobust and High Efficiency Nanogenerator for Energy Harvesting from Water Droplets. Adv. Mater. 32, 2001699. https://doi.org/10.1002/ adma.202001699. 40. Nie, J., Ren, Z., Xu, L., Lin, S., Zhan, F., Chen, X., and Wang, Z.L. (2020). Probing Contact-Electrification-Induced Electron and Ion Transfers at a Liquid–Solid Interface. Adv. Mater. 32, 1905696. https://doi.org/10.1002/ adma.201905696. 41. Peng, J., Kang, S.D., and Snyder, G.J. (2017). Optimization principles and the figure of merit for triboelectric generators. Sci. Adv. 3, eaap8576. https://doi.org/10.1126/sciadv.aap8576. 42. Zargari, S., Rezania, A., Koozehkanani, Z.D., Veladi, H., Sobhi, J., and Rosendahl, L. (2022). Effect of the inherent capacitance optimization on the output performance of triboelectric nanogenerators. Nano Energy 92, 106740. https://doi.org/10.1016/j.nanoen.2021.106740. 43. Zhao, Z., Dai, Y., Liu, D., Li, X., Zhou, L., Zhang, B., Wang, J., and Wang, Z.L. (2023). Achieving high contact-electrification charge density on inorganic materials. Nano Energy 114, 108616. https://doi.org/10.1016/j. nanoen.2023.108616. 44. Sahu, M.,  Safranko, S., Hajra, S., Padhan, A.M.,  Zivkovi c, P., Joki c, S., and Kim, H.J. (2021). Development of triboelectric nanogenerator and mechanical energy harvesting using argon ion-implanted kapton, zinc oxide and kapton. Mater. Lett. 301, 130290. https://doi.org/10.1016/j.matlet. 2021.130290. 45. Adamovich, I., Agarwal, S., Ahedo, E., Alves, L.L., Baalrud, S., Babaeva, N., Bogaerts, A., Bourdon, A., Bruggeman, P.J., Canal, C., et al. (2022). The 2022 Plasma Roadmap: low temperature plasma science and technology. J. Phys. D Appl. Phys. 55, 373001. https://doi.org/10.1088/ 1361-6463/ac5e1c. 46. Kim, Y.Y., Yang, T.-Y., Suhonen, R., Kemppainen, A., Hwang, K., Jeon, N.J., and Seo, J. (2020). Roll-to-roll gravure-printed flexible perovskite solar cells using eco-friendly antisolvent bathing with wide processing window. Nat. Commun. 11, 5146. https://doi.org/10.1038/s41467-020- 18940-5. 47. Zabihipour, M., Lassnig, R., Strandberg, J., Berggren, M., Fabiano, S., Engquist, I., and Andersson Ersman, P. (2020). High yield manufacturing of fully screen-printed organic electrochemical transistors. npj Flex. Electron. 4, 15. https://doi.org/10.1038/s41528-020-0078-9. 48. Castillo-Seoane, J., Gil-Rostra, J., Lo ´pez-Flores, V., Lozano, G., Javier Ferrer, F., Espino ´s, J.P., Ostrikov, K.K., Yubero, F., Gonza ´lez-Elipe, A.R., Barranco, A ´., et al. (2021). One-reactor vacuum and plasma synthesis of transparent conducting oxide nanotubes and nanotrees: from single wire conductivity to ultra-broadband perfect absorbers in the NIR. Nanoscale 13, 13882–13895. https://doi.org/10.1039/D1NR01937F. 49. Barranco, A., Cotrino, J., Yubero, F., Espino ´s, J.P., Benı ´tez, J., Clerc, C., and Gonza ´lez-Elipe, A.R. (2001). Synthesis of SiO2 and SiOxCyHz thin films by microwave plasma CVD. Thin Solid Films 401, 150–158. https:// doi.org/10.1016/S0040-6090(01)01632-7. 50. Garcı ´a-Casas, X., Ghaffarinejad, A., Aparicio, F.J., Castillo-Seoane, J., Lo ´- pez-Santos, C., Espino ´s, J.P., Cotrino, J., Sa ´nchez-Valencia, J.R., Barranco, A ´., and Borra ´s, A. (2022). Plasma engineering of microstructured piezo – Triboelectric hybrid nanogenerators for wide bandwidth vibration energy harvesting. Nano Energy 91, 106673. https://doi.org/10.1016/j. nanoen.2021.106673. 51. Yarin, A.L. (2006). DROP IMPACT DYNAMICS: Splashing, Spreading, Receding, Bouncing. Annu. Rev. Fluid Mech. 38, 159–192. https://doi. org/10.1146/annurev.fluid.38.050304.092144. 52. Montes, L., Roma ´n, J.M., Garcı ´a-Casas, X., Castillo-Seoane, J., Sa ´nchez- Valencia, J.R., Barranco, A ´., Lo ´pez-Santos, C., and Borra ´s, A. (2021). Plasma-Assisted Deposition of TiO 2 3D Nanomembranes: Selective Wetting, Superomniphobicity, and Self-Cleaning (Adv. Mater. Interfaces 21/2021). Adv. Mater. Interfac. 8, 2170122. https://doi.org/10.1002/ admi.202170122. 53. Ghaffarinejad, A., Lu, Y., Hinchet, R., Galayko, D., Hasani, J.Y., Kim, S.W., and Basset, P. (2018). Bennet’s doubler working as a power booster for triboelectric nano-generators. Electron. Lett. 54, 378–379. https://doi. org/10.1049/el.2017.3434. 54. Garcia-Casas, X., Moreno-Martinez, G.P., Nun ˜ez-Galvez, F., Czermak-A ´lvarez, T., Krishna, H., Aparicio-Rebollo, F.J., Sanchez-Valencia, J.R., Barranco, A., and Borras, A. (2024). NanoDataLyzer. https://doi.org/10.5281/ zenodo.13225385 Version 1.0. Zenodo. 55. Zhao, Y.-P., Wang, G.-C., Lu, T.-M., Palasantzas, G., and De Hosson, J.T.M. (1999). Surface-roughness effect on capacitance and leakage current of an insulating film. Phys. Rev. B 60, 9157–9164. https://doi.org/10. 1103/PhysRevB.60.9157. 56. Dai, K., Wang, X., Niu, S., Yi, F., Yin, Y., Chen, L., Zhang, Y., and You, Z. (2017). Simulation and structure optimization of triboelectric nanogenerators considering the effects of parasitic capacitance. Nano Res. 10, 157–171. https://doi.org/10.1007/s12274-016-1275-7. Please cite this article in press as: Ghaffarinejad et al., Triboelectric pixels as building blocks for microscale and large-area integration of drop energy harvesters, Device (2024), https://doi.org/10.1016/j.device.2024.100566 18 Device 3, 100566, February 21, 2025 Article ll OPEN ACCESS