Small-scale water energy harvesting for sustainably-powered distributed electronics

Abstract

Small-scale water energy harvesting offers a promising pathway to power the rapidly expanding ecosystem of distributed electronics. The inherent ubiquity and mechanical energy of water make it an ideal resource for sustainable, off-grid power generation. The efficiency and application scope of these energy harvesters are fundamentally governed by their working environment and transduction mechanisms. This review provides a comprehensive overview of small-scale water energy harvesting technologies, covering different forms of water and the corresponding energy transduction mechanisms, including triboelectric, piezoelectric, electromagnetic, and thermoelectric approaches. We highlight the state-of-the-art examples with particular emphasis on recent advances in triboelectric nanogenerators (TENGs)-based of small-scale water energy harvesting, focusing on advances in nanomaterial design, hybrid transduction mechanisms, and interfacial engineering. The diverse applications of these technologies in energy harvesting, self-powered sensors, and self-powered wearable devices are discussed. Finally, we outline future research directions, emphasizing the development of hybrid systems, advanced interfacial engineering, and intelligent power management, aiming to bridge the gap between laboratory innovation and real-world deployment.

Introduction

Water pervades Earth’s systems in a multitude of forms across natural and engineered environments. From oceans and lakes that together cover more than 70% of the planet’s surface¹ to atmospheric precipitation and cloud formations, and from the circulatory fluids that sustain living organisms to condensation and spray processes integral to industrial operations, water underpins ecological function and human activity while mediating diverse energy flows. The exploitation of kinetic and potential energy in moving water represents one of humanity’s earliest and most consequential technological endeavors, driving mills, irrigation, and transport across civilizations². In contemporary energy systems, hydropower persists as the largest single source of renewable electricity globally, accounting for roughly 14% (~4500 TW∙h) of total electricity generation³. Ongoing advancements in turbine engineering, environmental mitigation strategies—such as fish-friendly turbine designs and sediment management practices—grid integration techniques, and marine energy technologies are broadening the role of water-based energy in low-carbon transitions^4,5. Moreover, emergent approaches that integrate reservoir management with climate adaptation, hybrid renewable systems, and nature-based solutions highlight the multifunctional potential of water infrastructure to support energy security, ecosystem services, and resilience to hydrological variability^6,7,8.

Beyond the traditional water energy harvesting, the rapid expansion of the Internet of Things (IoT), with device counts projected in the tens of billions^9,10, is shifting demand patterns toward numerous, spatially distributed, low-power loads. This evolution has renewed interest in localized electrical generation paradigms—such as microgrids, distributed generators, and energy-harvesting systems—that are intended to supplement, rather than supplant, centralized generation. Distributed generation can mitigate transmission losses (commonly reported at 5–10% in advanced grids), enhance operational reliability through fault isolation and localized control, and better align supply with the intermittent, low-power profiles characteristic of many IoT applications^11,12. In this context, small-scale water energy harvesting emerges as a particularly promising approach owing to its adaptability across environments and its potential to exploit ubiquitous moisture and water flows in both built and natural settings.

In the landscape of energy harvesting, the energy scale and portability are critical factors that influence the deployment and utility of various technologies (Fig. 1). As the energy scale increases, devices such as power plant-scale hydropower systems can generate substantial energy outputs, enabling large-scale applications but often at the expense of reduced portability due to their size and infrastructure requirements. Conversely, as energy scale decreases, portable devices offer significant flexibility, facilitating integration into distributed electronics and IoT applications but may be limited by lower power output. For small-scale energy harvesting in distributed electronics, considerations such as energy conversion efficiency, scalability, and environmental adaptability are crucial. Efficient energy capture tailored to specific environmental factors ensures consistent power supply, while scalable designs support diverse application contexts, from wearables to industrial sensors. Furthermore, as devices become smaller and more integrated, the balance between energy density and device footprint becomes paramount. Advanced materials and innovative designs that enhance charge collection and minimize energy loss are vital in meeting the demands of resilient, low-power electronic networks.

Fig. 1: Different types of water energy harvesting technologies.

Energy scale and portability are compared.

In light of these opportunities and challenges, this review provides a comprehensive overview of the rapidly growing field of small-scale water-energy harvesting technologies. Compared with recent reviews on hydrovoltaic technology, which primarily focus on electricity generation driven by water–material interfacial interactions under quasi-static conditions¹³, this review adopts a broader framework organized around different forms of water and the mechanical, thermal, and multiphysical driving forces they provide, enabling cross-mechanism comparison within small-scale water energy harvesting. Beginning with an analysis of water forms and energy transduction mechanisms, this review critically evaluate cutting-edge technologies through the lenses of material innovation, device architecture, and interfacial design. These advancements are further contextualized within their potential applications, spanning energy storage systems, real-time environmental sensing, and next-generation wearable electronics. Lastly, we conclude by summarizing the key challenges that remain and offering perspectives on future research directions, including high-performance interface design, long-term reliability, system integration, and standardized evaluation, with the aim of providing guidance for the further advancement of this field.

Forms of water and energy transduction mechanisms

Water in its gaseous, liquid, and solid forms enables µW–W scale energy harvesting through distinct transduction mechanisms^14,15,16. The choice of water phase and energy harvesting mechanism significantly influences design parameters, including fabrication complexity, power performance, and working environment. This section explores the forms of water and transducers that are employed in the fields of small-scale water energy harvesting technologies, highlighting their operational principles, advantages, and limitations. Comparison of energy transduction mechanisms for small-scale water energy harvesting is summarized in Table 1^{14,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65} and Table 2^{14,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65}.

Table 1 Comparison of energy transduction mechanisms for small-scale water energy harvesting

Table 2 Advantages and disadvantages of energy transduction mechanisms for small-scale water energy harvesting

Gaseous forms

Moist-sorption-driven mechanisms

Phase-change–induced capillary flow

Streaming-potential generators convert capillary- or evaporation-driven fluid flow in charged pores or channels into electric power via electrokinetic coupling^66,67 (Fig. 2a). When an electrolyte flows through a charged surface, the electric double layer (EDL) adjacent to the surface contains excess counterions; flow drags these ions, producing a streaming current^68,69. An open-circuit streaming potential develops as ionic convection is balanced by electromigration and diffusion. For thin EDLs, the Helmholtz–Smoluchowski relation describes the streaming potential:

$${V}_{{stream}}=\frac{\varepsilon \zeta }{\eta \sigma }\Delta P$$

where $\varepsilon$ is the fluid permittivity, $\zeta$ is the zeta potential, $\eta$ is the fluid viscosity, $\sigma$ is the fluid conductivity, and $\Delta P$ is the pressure gradient^68,69. Evaporation sustains capillary flow by imposing a pressure head (Laplace pressure ΔP ≈ 2γcosθ/r, where γ is the surface tension, θ is the contact angle, and r is the pore radius)⁷⁰. Thus, an evaporating wick or porous membrane can continuously drive flow from a liquid reservoir through charged nanostructures⁷¹. The electricity generation and duration of devices driven by this mechanism are limited due to the low ion concentration in the water adsorbed and condensed from humid air and the considerable distance between ions and the surface of nanochannels⁷². Strategies including the reduction of nanochannel diameter⁷³, increase of number of nanochannels⁷⁴, enhancement of zeta potential⁷⁵, and improvement of electrokinetic conversion efficiency through channel wall smoothening⁷⁶ have been demonstrated as effective approaches for achieving higher streaming potential.

Fig. 2: Energy transduction mechanisms in gaseous forms.

a Streaming Potential: Conversion of capillary or evaporation-driven fluid flow in charged pores or channels into electric power via electrokinetic coupling. b Ion Concentration Gradient: Diffusion of charge carriers driven by the concentration gradient. c Sorption-driven deformation: Mechanical energy generated through evaporation and sorption arises from hygroscopic swelling and deswelling of materials. d Electronic Seebeck Conversion: Temperature differential across thermoelectric materials. e Thermo-osmotic Streaming: Temperature-induced flow of solvent and ions through porous media or membranes. f Vapor-driven Electromechanical Conversion: Transforms steam or humid air pressure and momentum into mechanical energy.

Ion concentration gradient

A distinct class of generators relies not on bulk liquid flow but on sorption-mediated ion generation and sustained ionic gradients within hygroscopic materials. The charge carrier diffusion according to the concentration gradient inside an active material is the most fundamental working mechanism (Fig. 2b). The hygroscopic functional layers enriched with oxygen-functional groups (e.g., –OH, –COOH, –NH₂, –SO₃H/–SO₃⁻) begin to ionize by absorbing moisture in the atmosphere, liberating mobile charge carriers (H⁺ and H₃O⁺) while leaving their conjugate anions immobilized on the polymer backbone⁷⁷. According to the Grotthuss mechanism^78,79, as water molecules migrate from a wet state to an arid state, the proton and hydronium populations also drift along the hydrogen-bond network, leading to charge separation and potential generation. Therefore, the chemical potential energy of moisture uptake can be converted into electrical energy through directional ionic transport within the functional layer. A prerequisite for the unidirectional charge carriers transfer is the establishment of an ion concentration gradient during moisture uptake, and there are several strategies: (i) water-content gradients⁷⁹, controlling moisture absorption via exposure conditions of devices, such as one-sided or dual-sided open electrodes; (ii) pre-introduced functional group gradients⁷⁷, forming chemical asymmetry via supplying constant bias⁸⁰, laser writing⁸¹, thermal reduction⁸², electrodeposition⁸³, and UV photoreduction⁸⁴; (iii) salt concentration gradients⁸⁵, utilizing Marangoni-assisted gelation⁸⁵, irradiating a laser⁸⁶, and laminating a cation-selective membrane⁸⁷.

Sorption-driven mechanical deformation

Mechanical energy generated through evaporation and sorption arises from the hygroscopic swelling and deswelling of materials, which induce strains, bending, and oscillations in engineered bilayers and porous films^88,89. These mechanical motions can be transduced into electrical energy via piezoelectric^90,91,92,93 or triboelectric^{94,95,96,97,98} mechanisms. In bilayer actuators, the differential uptake or loss of water between layers with varying hygroscopicity or stiffness results in curvature changes during wet–dry cycles (Fig. 2c)⁹⁹. The integration of piezoelectric layers facilitates the conversion of bending strain into electrical charge¹⁰⁰. Additionally, evaporation-induced airflow or capillary forces can drive the cyclic motion of small flaps or cantilevers, causing repeated contact and separation of triboelectric surfaces. This process generates charge transfer and intermittent electrical pulses^{94,95,96,97,98,101}. In comparison to electrokinetic methods, mechanical transduction typically yields higher instantaneous power in pulsed form, though it necessitates cyclic actuation. Key design considerations include maximizing the strain or force output per unit moisture variation¹⁰², optimizing resonant behavior to enhance energy conversion efficiency¹⁰³, and ensuring material fatigue resistance to endure repeated sorption cycles¹⁰⁴.

Steam-driven mechanisms

Thermoelectric harvesting

The utilization of steam for thermoelectric energy harvesting capitalizes on temperature gradients between a steam source and a cooler sink¹⁰⁵, enabling power generation through the Seebeck effect, which defined as ΔV = S·ΔT where S is the Seebeck coefficient and ΔT is the temperature differential across the thermoelectric material^106,107 (Fig. 2d). The power output is influenced by the material’s power factor (S²·σ) and its figure of merit (ZT = S²·σT/κ), with σ representing electrical conductivity, κ thermal conductivity, and T absolute temperature^108,109. To maximize the temperature gradient in steam-coupled systems, efficient thermal coupling to the vapor stream is essential, complemented by effective heat rejection at the cold side. Techniques such as finned heat sinks¹¹⁰, fluid cooling²⁷, or phase-change condensers¹¹¹ are commonly employed. Practical challenges include mitigating parasitic heat conduction through support structures and minimizing contact resistances, both of which can reduce the effective ΔT¹¹². Optimizing the balance between the active area and thickness of thermoelectric elements is also critical to enhance performance while minimizing thermal losses¹¹³. Low-grade steam, prevalent in industrial waste-heat applications, e.g., near-saturated or modestly superheated steam, offers a promising energy source for thermoelectric harvesting²⁹. Advanced materials such as skutterudites, PbTe, Mg₃Sb₂-based alloys, and segmented legs have been employed in high-performance thermoelectric modules, achieving power outputs ranging from microwatts to watts depending on the temperature gradient and device scale^114,115,116.

Thermo-osmotic and thermoelectrokinetic conversion

Thermo-osmotic and thermoelectrokinetic conversion systems harness temperature-driven solvent and ion flux through porous media or membranes to generate electrical signals^117,118. When a temperature gradient is applied across a charged nanopore or membrane containing an electrolyte, such as the Fig. 2e, two primary phenomena occur: (1) thermo-diffusion (Soret effect), in which species redistribute in response to the thermal gradient, creating concentration gradients¹¹⁹, and (2) thermo-osmotic flow where net solvent transport is driven by temperature-dependent interfacial interactions and surface tension gradients¹²⁰. In charged nanochannels, these effects interact with the EDL, producing a thermoelectric-like open-circuit voltage (ionic Seebeck effect) and a thermally driven streaming current¹²¹. Theoretical models integrate Navier–Stokes hydrodynamics with Poisson–Nernst–Planck transport, incorporating thermal diffusion terms to describe the system’s behavior¹²². Power generation is influenced by factors such as surface charge density, channel confinement (EDL overlap), ionic mobility asymmetry, and the applied temperature gradient^122,123.

Microturbines for vapor-harvesting

Microturbines designed for vapor-harvesting are compact rotary machines that convert the pressure and momentum flux of steam or humid-air flows into mechanical work¹²⁴, which is then transduced into electrical energy via integrated electromagnetic generators (Fig. 2f). At small scales, the design of these turbines deviates from conventional turbine theory due to the dominance of viscous^125,126, tip^127,128, and bearing losses¹²⁹, as well as Reynolds-number effects that influence blade loading and flow separation behavior¹³⁰. The theoretical aerodynamic available power is given by 0.5·ρ·A·U³, where ρ is the fluid density, A is the swept area, and U is the flow speed¹³¹. However, practical shaft power is constrained by the achievable aerodynamic conversion efficiency (η_aero)¹³² and allowable pressure-drop limits in process ducts¹³³. To optimize performance, microturbine designs focus on blade-element adaptations for low-Reynolds-number regimes, impulse versus reaction staging to suit available pressure head¹³⁴, and minimizing leakage and tip-clearance losses. These design principles are well-documented in the literature on small-scale turbomachinery, including actuator-disk¹³⁵ and blade-element theories¹³⁶.

Liquid forms

Hydromechanic-driven mechanisms

Electromagnetic transduction

Electromagnetic microgenerators are highly reliable and well-suited for converting the continuous shaft power of microturbines into electrical energy¹³⁷. Prototypes of duct-mounted microturbines integrated with permanent-magnet alternators have demonstrated milliwatt-level outputs for centimeter-scale rotors operating in moderate flow conditions, as documented in engineering reports and prototypes^33,138,139. Design strategies derived from microturbine offer critical guidance for optimizing magnetic circuits and windings to maximize energy conversion efficiency¹⁴⁰. Electromagnetic generators operate on the principle of electromagnetic induction, as articulated by Faraday’s law¹³⁷. In rotary microturbines, the mechanical rotation of the shaft drives a rotor fitted with permanent magnets past stationary coils (Fig. 3a). The time-varying magnetic flux through the coils induces an electromotive force (EMF) that is directly proportional to the rate of change of the magnetic flux. The instantaneous electrical power generated is influenced by several parameters, including the induced voltage, load impedance, internal winding resistance, and back-EMF, which introduces electromagnetic damping to the mechanical system.

Fig. 3: Energy transduction mechanisms in liquid and ice/snow forms.

a Flow-driven Electromagnetics: Mechanical rotation of a shaft, driven by liquid flow, spins a rotor with permanent magnets past stationary coils to generate electromagnetic energy. b Flow-resonant Electrical Polarization: Mechanical movement of piezoelectric cantilevers induced by fluid flow, leading to electrical polarization and energy generation. c Solid–Solid Contact Electrification: Generation of electrical charge through triboelectric effects during contact and separation of solid materials induced by wave motion. d Liquid-Solid Contact Electrification: Generation of electrical charge when solid and liquid surfaces contact and separate. e Droplet-driven Contact Electrification: Conversion of the kinetic and interfacial energy of falling or sliding droplets into electrical energy through contact and separation processes. f Snow-Geothermal Seebeck Conversion: Generating electrical energy by exploiting temperature differences between geothermal heat sources and snow or ice acting as the cold sink in thermoelectric systems. g Impact-induced Polarization: Generation of electrical signals through mechanical interactions between solid precipitation (like hail or ice) and structural surfaces. h Ice-surface Contact Electrification: The transfer of electrical charge occurs when ice or hail contacts a dielectric surface and separates.

Piezoelectric transduction

Piezoelectric harvesters are particularly well-suited for compact, low-maintenance applications where oscillatory motion is present. Experimental studies of piezoelectric cantilevers in fluid flows have shown that operating near the resonant frequency can yield average power outputs in the micro- to milliwatt range (Fig. 3b)^39,141. In liquid environments, the increased fluid damping lowers the Q-factor and reduces average power unless the design compensates by increasing the forcing amplitude or expanding the active area¹⁴². Piezoelectric transducers operate based on the direct piezoelectric effect, where mechanical strain induces an electrical charge^143,144,145. When stress is applied to a piezoelectric material, its intrinsic polarization state changes, which in turn disturbs the charge equilibrium at the electrodes. As a result, charge transfer occurs in the external circuit, generating an induced current. The relationship between the external stress and the polarization of the piezoelectric material can be expressed by Equation (1):

$${D}_{i}={d}_{{ij}}{\delta }_{j}$$

where D_i is the polarization change along direction i, d_ij is the piezoelectric coefficient and $\delta$_j is the applied stress along the direction j¹⁴⁶. Since the polarization is determined by the intrinsic properties of the piezoelectric material, the performance of piezoelectric harvesters is minimally affected by humidity¹⁴³. Even under extreme conditions involving strong impacts, they can maintain stable output efficiency¹⁴⁷ and a long operational lifespan¹⁴⁸. In hydromechanical energy harvesters, piezoelectric materials are often attached to cantilevers, bimorphs, or unimorphs that deform under fluid forces such as vibration^{149,150,151,152}, vortex excitation^153,154,155, or droplet impact^43,44,156. The efficiency of piezoelectric harvesters is maximized when the mechanical excitation drives the structure near its resonant frequency, amplifying strains and enhancing power density.

Triboelectric transduction

Since 2012, when Wang’s group first reported energy harvesting through the interaction of two solid materials, extensive research has been conducted in the field of triboelectric nanogenerator (TENG) with laboratory experiments achieving power densities in the µW–mW range¹⁵⁷. The triboelectric effect describes the process by which charges are generated and separated at the interface of two dissimilar materials upon contact and subsequent separation (Fig. 3c, d). Subsequent separation or relative motion induces a flow of current through an external circuit^{49,158,159,160,161,162}. In hydromechanical applications, TENGs can be actuated by flapping motion^159,163 or sliding surfaces driven by fluid forces. Unlike the traditional electromagnetic induction effect, the triboelectric effect mainly uses interface physical processes to achieve energy conversion¹⁵⁷, and therefore shows unique advantages in the collection of decentralized and microscale water flow^164,165 and water wave energy^166,167. In water energy harvesting, TENGs are particularly effective for ultra-low energy-density and intermittent water sources, as contact electrification and electrostatic induction are highly sensitive to weak mechanical stimuli^168,169. Their polymer-based architectures further enable straightforward integration with flexible and wearable devices^46,170,171. The underlying mechanism of the triboelectric transductions can be attributed to the processes of contact electrification (CE) and electrostatic induction.

Electron Transfer: The CE is defined as the phenomenon of electrical charging arising from physical contact, which is observed in all materials and forms one of the most fundamental aspects of electricity generation. The CE process at liquid-solid interfaces can be explained by the electron cloud/potential model^172,173,174. At the atomic scale, electrons orbiting the nucleus are loosely confined within potential well generated by the nucleus, forming the electron cloud. As shown in Fig. 4a(i), before the droplet falls on the solid surface, the electron clouds of these materials remain distinct and separate, with the potential well binding the electrons to avoid escape. Upon contact between the liquid and the solid surface, the potential wells of the two materials overlap, thereby altering the distribution of electron clouds. This results in the transfer of electrons from the material with the shallow potential well to that with the deeper potential well (Fig. 4a(ⅱ)). The separated electrons remain in a stable equilibrium until the contact state is altered as the droplet slides across the surface (Fig. 4a(ⅲ)). As the temperature rises, the energy fluctuations of electrons become increasingly intense. Consequently, electrons are more likely to jump out of the potential well, either returning to the original material atoms or being emitted into the air (Fig. 4a(ⅳ)). This model is highly effective in explaining CE process occurring at interfaces between solids, liquids and gases. Furthermore, when liquid solutions containing ions, such as ordinary water or rainwater, come into contact with solid surfaces, an EDL is formed.

Fig. 4: Working Principle of contact electrification and electric double layer.

a Electron-cloud potential -well model for explaining contact electrification (i) Before contact, atoms have distinct electron clouds (ii) During contact, electron clouds overlap (iii) After contact, electrons transfer between atoms (iv) Higher temperature increases electron transfer. Reproduced with permission from ref. ¹⁷³. Copyright 2022, Wiley. b Two-Step Model for Electric Double Layer formation. First step: formation of initial electrostatic charges layer. (i) Schematic representation of molecules in the solution. (ii) Molecules adsorb onto the surface, involving electron transfer results in a charged surface; and (iii) Adsorbed molecules displaced from the surface due to liquid pressure. Second step: Ion segregation and double layer formation: (iv) the presence of charged surface as a result of the first step and (v) Ions in the solution are attracted and adsorb onto the charged surface, forming an electric double layer (EDL). Reproduced with permission from ref. ¹⁷⁴. Copyright 2019, Elsevier.

Ionic Adsorption and Transfer: EDL is a region of separated charges formed at the interface between a charged surface and an electrolyte solution. The structure of the double layer is composed of an inner layer composed of adsorbed ions and an outer layer formed by diffused ions^175,176. The formation of EDL can be summarized as a two-step process. In the first step, upon contact between a liquid and an unexposed solid surface lacking a surface charge (Fig. 4b(ⅰ)), the electron clouds present in the liquid molecules overlap with those located within the atoms of the solid surface (Fig. 4a(ⅱ)). To achieve electron equilibrium, electron transfer takes place, thus forming an initial layer of static charge on the solid surface (Fig. 4b(ⅱ)). Hereafter, the liquid flow propels the molecules that have undergone electron transfer, as well as other molecules, away from the solid wall, leading to the deposition of a layer of ions on the solid surface (Fig. 4b(ⅲ), b(ⅳ)). Then, if the solution contains free ions, such as H⁺ and OH^-, the counterions dispersed within the solution are attracted to the solid charged surface through the action of the Coulomb force, this establishes the EDL (Fig. 4b(ⅴ))¹⁷⁴. Following the initial electron transfer process, which charges the solid surface, subsequent processes are dominated by ion transfer. During the liquid-solid relative motion, the charged liquid induces a rearrangement of the charge on the solid surface, thereby generating current output.

Droplet-driven mechanisms

Multiple mechanisms have been proposed for droplet-based energy harvesting. One representative approach, analogous to triboelectric transduction, relies on CE and electrostatic induction¹⁷⁷. As illustrated in Fig. 3e, in these systems electrical output is generated through interfacial processes that occur as droplets approach, spread on, and detach from dielectric surfaces such as polytetrafluoroethylene (PTFE), fluorinated ethylene propylene (FEP), or polydimethylsiloxane (PDMS)^177,178. Upon contact, interfacial charge transfer is initiated at the liquid–solid interface due to atomic-scale overlap of electron clouds, leading to the formation of an EDL that subsequently governs electrostatic induction during droplet motion¹⁷². This process establishes an uneven charge distribution, with the dielectric surface becoming negatively charged while the droplet acquires an equal positive charge. Upon detachment, the interfacial equilibrium is disrupted, and the induced potential difference drives electrons to flow through the external circuit, producing a transient alternating current (AC) output. With successive droplet impacts, this charge–discharge cycle repeats periodically, resulting in a stable and continuous energy output that is characteristic of TENGs^57,179. The magnitude of the generated voltage and current is strongly influenced by the surface physicochemical properties—including wettability^180,181, roughness¹⁸², and chemical composition^52,168—as well as the intrinsic charge-retention capacity of the dielectric¹⁸³. Hydrophobic and micro/nanostructured surfaces can minimize water adhesion, prevent charge screening by residual moisture, and maintain high surface potential differences, thereby enhancing both charge density and energy conversion efficiency^184,185. Moreover, optimizing the dielectric thickness, electrode configuration, and droplet kinetic parameters (size, velocity, and frequency) can further improve charge transfer efficiency and ensure long-term operational stability even under high-humidity or continuous rainfall conditions¹⁸⁶. Moreover, to further enhance the energy-harvesting efficiency of droplet-based systems, researchers proposed a transistor-like architecture, in which a metal electrode is introduced between the solid–liquid interface^60,187. This design enables the droplet to form a closed circuit instantaneously upon spreading, thereby triggering bulk charge transport within the system. Consequently, the energy conversion process becomes dominated by the bulk effect, resulting in a substantial enhancement of electrical output—improving the performance by several orders of magnitude.

Recently, Wang et al. proposed a new energy conversion mechanism analogous to the photovoltaic effect, but driven by mechanical contact rather than light irradiation, which they termed the tribovoltaic effect¹⁷⁴. The tribovoltaic effect occurs at the p–n semiconductor junction during mechanical contact. When p-type and n-type semiconductors come into contact, electron cloud overlap and electron transfer occur at the interface, leading to band bending and carrier redistribution. As a result, some electrons migrate from the n-region to the p-region, creating a built-in potential difference and producing a unidirectional current in the external circuit. Lin et al. were the first to couple droplets with semiconductors and realize direct current (DC) generation at a liquid–solid interface through the tribovoltaic effect⁵³. Their study revealed that when a water droplet slides across a silicon surface, continuous voltage and current signals are generated at the interface. This phenomenon originates from the electron transfer and bonding interactions between liquid molecules and semiconductor surface atoms during contact electrification, where the released energy excites electron–hole pairs and drives charge separation across the junction. Since the tribovoltaic effect relies on the electronic band structure and built-in electric field at the liquid-solid interface, the performance of this type of tribovoltaic nanogenerator (TVNG) will be affected by many factors, including semiconductor type and doping concentration^62,188,189, ion concentration of the droplet¹⁸⁸, and interface temperature¹⁹⁰.

Ice/snow forms

Thermoelectric harvesting in ice-dominated environments

Thermoelectric (TE) harvesting in ice-dominated environments utilizes the Seebeck effect to convert naturally occurring thermal gradients at ice–water or snow–air interfaces into electrical energy¹⁹¹. In icy environments¹⁹², the cold reservoir is effectively provided by ice, snow, or sub-surface frozen layers, which remain near 0 °C or below, while the hot side can be established by solar warming of the surface, waste heat from buried equipment, or geothermal/groundwater fluxes. Since TE voltage scales linearly with ΔT but power output depends on both ΔT and the module’s internal resistance and heat flow, practical designs aim to maximize the usable temperature difference across the TE legs while minimizing parasitic thermal conduction and contact resistances¹⁹³. This necessitates careful thermal coupling (e.g., high-conductivity interfacial materials on the hot side), insulation to preserve the cold sink, and mechanical mounting that tolerates frost and expansion stresses. Field concepts for TE harvesting in icy environments include TE-augmented polar buoys and ice-moored sensors that exploit sub-surface water temperatures versus surface ice temperatures to supply low-power electronics (Fig. 3f). Similar ideas apply to instrumentation on ice-covered rivers, where diurnal solar heating produces transient ΔT between sun-warmed surfaces and colder ice layers.

Mechanical interactions with solid precipitation

Mechanical interactions between solid precipitation (ice, hail, snow) and structural surfaces offer intermittent but valuable energy sources that can be harnessed via piezoelectric¹⁹⁴ and triboelectric^195,196,197 mechanisms. Piezoelectric harvesters convert strain from impact, compression, sliding, or bending into electrical charge through the direct piezoelectric effect^144,198. Materials such as PZT ceramics, piezoelectric composites, or polymeric piezoelectrics (such as PVDF) generate charge proportional to applied stress and strain rate¹⁹⁸. In ice- and snow-laden environments, relevant loading events include hail/ice impacts like Fig. 3g, sudden snow shedding from roofs or branches, and freeze-induced cracking. The energy available per event scales with the mechanical impulse or strain energy imparted to the transducer (e.g., impact energy ≈ ½mv² for an ice particle), and peak electrical output depends on modal coupling (placing the piezo element at high-strain locations), electromechanical coupling coefficients, and the dynamic bandwidth (resonant amplification improves yield if impact or shedding excites modal frequencies). In additions, triboelectric nanogenerators (TENGs) operate on the principles of contact electrification and electrostatic induction when two dissimilar materials contact and separate¹⁹⁹. In solid-precipitate contexts, ice or hail contacting a dielectric surface can transfer charge that is collected upon separation (Fig. 3h). Translating these concepts to ice/snow requires surfaces engineered for controlled contact and rapid detachment—superhydrophobic or textured coatings promote clean separation for wet contacts, while low-temperature-stable dielectrics and microtexturing can enhance charge transfer with solid ice¹⁹⁵. Major engineering challenges for TENGs in frozen environments include charge leakage through thin water films during melting, reduced surface resistivity at subzero wetting points, abrasion from ice particles, and maintaining long-term triboelectric contrast¹⁹⁷. Mitigation strategies involve hydrophobic/icephobic coatings, encapsulation of electrodes, and hybridization with mechanical buffers to concentrate impact energy.

State-of-the-art examples of small-scale water energy harvesting

Moist- and steam- driven energy harvestings

Phase-change–induced capillary flow

Moisture electric generators (MEGs) based on the streaming-potential mechanism commonly feature a functional layer of porous or nanostructured materials, which can be classified as carbon-based, inorganic solid oxide-based, bio-based, and polymer-based nanomaterials. Water transport driven by evaporation-induced capillary flow within these materials generates an electrical output through electrokinetic charge separation. However, the power output and operational duration of the early reported devices remain limited due to the confined nanochannels and the unsustainable stream flow^200,201. From a materials perspective, recent studies focus on precise control of the nanochannel size and ion selectivity enhancement of materials to enhance the streaming potential.

Carbon-based materials (e.g., graphene oxide (GO)^79,82,202, carbon black²⁰³, carbon nanotubes (CNT)^204,205) have been widely used in MEGs for their high conductivity, tunable surface chemistry and high surface area. Wu et al.²⁰⁶ used 3D printing and freeze-casting to create a honeycomb-structured reduced graphene oxide (rGO) film with interconnected microchannels, offering a 7-fold larger surface area than layered structures and thus greatly enhancing water/ion interaction, power output, and durability. Inorganic solid oxide-based materials can be readily fabricated into nanostructured networks that provide abundant pathways for water and ion transport. Benefiting from their high surface charge density and efficient charge-transfer properties, materials such as TiO₂²⁰⁰, Al₂O₃^207,208, MoS₂²⁰⁹, and Ni–Al layered double hydroxides (LDH)²¹⁰ have demonstrated outstanding performance in MEG devices.

Owing to the inherent hydrophilicity and high surface charge density, bio-based material could sustain streaming potential under capillary flow over extended periods. Liu et al.²¹¹ fabricated a 7 μm-thick film of Geobacter sulfurreducens protein nanowires that generated a stable output of 0.5 V and an electric current density of 17 μA cm⁻² under ambient humidity for over 2 months (Fig. 5a, b). Cellulose-based biomaterial is widely used in streaming potential generators owing to the abundance of aligned capillaries. Lyu et al.⁷⁴ employed electrospinning followed by densification/annealing to tailor the pore size and porosity of cellulose acetate (CA) membranes (optimal porosity ≈ 52.6%, pore size <250 nm), thereby creating more interconnected nanochannels, and the open-circuit voltage was significantly increased from 0.075 to 0.3 V (Fig. 5c).

Fig. 5: Energy harvesting driven by moisture and steam based on phase change induced capillary flow.

a Schematic diagram of the device architecture(bottom) with the TEM images of the purified nanowire network(right panel) generated by Geobacter sulfurreducens, the microorganism depicted as the dark shape in the left panel. b Continuous voltage (V_o) recordings from a nanowire device over a period of more than two months, accompanied by ambient relative humidity indicated by the blue curve. a, b Reproduced with permission from ref. ²¹¹. Copyright 2020, Spring Nature. c Schematic illustration of the fabrication process for a porous cellulose acetate (CA) membrane-based moist-electric generator (P-CMEG)(left), and the proposed mechanism of ion diffusion along the negatively charged surface of the CA nanofiber nanochannels with SEM image of the as-electrospun CA membranes (M_org)(right). Influence of membrane porosity and pore size on open-circuit voltage (V_oc)(bottom). Reproduced with permission from ref. ⁷⁴. Copyright 2020, American Chemical Society. d The Debye screening effect mechanism induced by PCNH’s nanopores. The proposed mechanism of streaming potential involves ion diffusion along the negatively charged surface of nanopores in PCNH(left), and schematic illustration of the electrical double layer and the electric potential profile normal to negatively charged wall in PCNH (right). e Comparison of open-circuit voltage (V_oc) at 80% relative humidity (RH) between PCNH-MEG and CMCH-MEG. Data are presented as the mean ± standard deviation (n  =  3). f The pore size and the generated open-circuit voltage (V_oc) of PCNH-MEG prepared by filling different volumes of CMC. The V_oc data are present as the mean ± standard deviation (n  =  3). d–f Reproduced with permission from ref. ¹⁷. Copyright 2025, Springer Nature. g Depicts the salting-out induced recrystallization process in preparing freeze-assisted salting-out hydrogels(left), SEM images of the gel’s surface and cross-sectional structure after removal of the salting-out ions. h Pore volume ratios of micropores, mesopores, and macropores in various salting-out gels. i Comparison of the power generation performance of salting-out BHNEG with Hofmeister sequence ions. Inset: Schematic illustration of the structural composition of BHMEG. g–i Reproduced with permission from ref. ²¹⁸. Copyright 2025, Springer Nature. j Schematic of the self-sustained electricity generator (SSEG) (left), and side view diagram illustrating the device structure with different contact angles for hygroscopic layer(I), the inner side(II) and outer side(III) of the evaporative layer are shown (right). k Continuous recording of open-circuit voltage(black) and short-circuit current output(red) from the SSEG over ten days in an ambient environment (25 ± 2 °C, RH 60 ± 5%). Inset: Circuit model of the device. j, k Reproduced with permission from ref. ²²². Copyright 2022, Springer Nature.

Hydrogels as a class of polymer-based materials have attracted much attention recently due to their excellent hygroscopicity and are particularly suitable for coupling with nanostructured functional layers to induce power generation. Lin et al.¹⁷ employed a delignified pomelo peel skeleton filled with carboxymethyl cellulose (CMC)-based hydrogel with tuned nanopores below the Debye length (Fig. 5d). This causes the overlap of EDLs, allowing cation-selective transport (H⁺/Cu²⁺) and enhancing the flow potential. The resulting device delivered a voltage of 1.32 V at 80% RH which is approximately 3 times higher than those of conventional delignified pomelo peel respectively (Fig. 5e)¹⁷. Figure 5f further illustrates that the voltage enhancement is due to the Debye screening effect, which is enabled by the sub-Debye-length pore size of the hydrogel¹⁷.

For streaming potential-based MEG, rational functional layer microstructure design enables continuous charge transfer via asymmetric hydration and steady ion flux. In vertically aligned microchannel architectures, pore orientation dictates directional moisture and ion transport via capillary flow²¹². Recent representative strategies include hydrophilic coating on vertical channels, such as carbon-black impregnation within natural wood²¹³, LiCl-immersion of Juncus effusus with axially aligned channels²¹⁴, and anisotropic Al³⁺-loaded CNF aerogels fabricated by the freeze-drying method²¹⁵. In these samples, anisotropic microchannels combined with surface ionic modification enable directional capillary-driven water flow and enhanced electrokinetic charge separation, leading to enhanced streaming-potential outputs. Another kind of porous structure is foam-type porous architecture, such as aerogels²¹⁶, fibrous mats²¹⁷, and hydrogels^218,219, where the interconnected microstructure maximizes moisture transmission and internal surface area, acting as a fast water exchange and continuously renewing electroactive interfaces while buffering RH fluctuations. By inserting biomaterial and forming a covalently crosslinked porous network, enlarging the EDL active area and preventing wet collapse, thereby stabilizing the stream paths of water and charge carriers²²⁰. The foam-type porous architecture can be further optimized through a salting-out strategy using various anions (citrate⁻, SO₄²⁻, CO₃²⁻, acetate⁻, and Cl⁻), enabling precise control over nano/mesopore distribution and porosity (Fig. 5g)²¹⁸. As exemplified by Wang et al.²¹⁸, this approach was employed to construct a poly(vinyl alcohol) (PVA)/MXene bilayer hydrogel membrane featuring micro-meso-macroporous directional channels with asymmetric charge characteristics, where the multi-scale pore network greatly enhanced moisture transport and ionic mobility. Among the anions tested, the citrate⁻-treated hydrogel the highest proportion of micropores and mesopores (92.1%) and thus generate the optimal electrical signal output. This structural superiority ultimately contributed to a peak power density of 146 µW cm⁻² (Fig. 5h and Fig. 5i)²¹⁸.

Streaming-potential generators employ asymmetric device architectures to regulate moisture adsorption and desorption kinetics, thereby inducing continuous water and ion transport for sustained electricity generation. A typical planar asymmetric structure induces in-plane potential difference via uneven EDL distribution between dry and wet regions of a charged surface—Zhang et al.²²¹ asymmetrically deposited a hygroscopic ionic hydrogel on functionalized carbon, maintaining >0.6 V via spontaneous EDL formation at the interface. In addition, heterostructural designs combining moisture-absorbing and evaporative layers have been proposed to modulate water adsorption–desorption dynamics, ensuring continuous water flow and ion migration. Tan et al.²²² designed a bilayer heterostructured generator composed of a hygroscopic LiCl–cellulose paper layer and an evaporative carbon-black–cellulose layer, which delivered a stable output /;;pof ~0.78 V and 7.5 μA (25 °C, 60% RH) for over 10 days (Fig. 5j and k).

Ion concentration gradient

Graphene oxide (GO) can dissociate protons under humid conditions and was among the first materials used for MEGs owing to its abundant hydroxyl and carboxyl functional groups⁷⁷. Thermal reduction⁸², electrodeposition⁸³ and ultraviolet (UV) photoreduction⁸⁴ were used to create functional group gradient within the GO to generate electricity based on the ion concentration gradient. Due to the complex fabricating process, bio-based materials and polymers (e.g., poly(4-styrenesulfonic acid) (PSSA), polydopamine (PDA), poly(vinyl alcohol) (PVA), and polyacrylic acid (PAA)) have since been widely explored for stable MEG output.

Apart from conventional polymers, polymeric hydrogels are considered promising MEG functional materials owing to their excellent water absorption, ionic conductivity, biocompatibility, and flexibility²²³. Upon exposure to humid air, they readily absorb moisture through abundant hydrophilic groups, forming water-content gradients that drive ion migration. Incorporating ionic functional groups or salts can significantly enhance the charge transport of hydrogels^224,225. Yang et al.¹⁴ demonstrated that substituting protons with other cations (Na⁺) in a PVA–sodium alginate (SA) hydrogel leads to improved conductivity, achieving a short-circuit current density of 1.3 mA cm⁻² and 1.30 V voltage output for over 120 h from a single device (Fig. 6a, b). Furthermore, the MEG could be enhanced by protonation-doping of otherwise polymer networks. For example, by introducing abundant H⁺ carriers, a phytic-acid-doped PVA hydrogel achieved an output of 0.8 V and 0.24 mA cm⁻² at 80% RH (Fig. 6c, d)²²⁶. The semi-crystalline PVA matrix, plasticized with glycerol, ensured continuous proton transport pathways and endowed the gel with high mechanical strength and anti-drying properties. In addition, double network (DN) hydrogel was proposed to enhance the mechanical robustness and electrical performance because it can maintain high water activity and provide continuous 3D porous pathways, enabling rapid proton migration and stable humidity-driven gradients (Fig. 6e)²²⁷. A DN hydrogel composed of sulfated cellulose nanofibrils and PVA generated an open circuit voltage of 0.9 V and a short circuit current density of 92 µA cm⁻² at 80% RH²²⁷.

Fig. 6: Energy harvesting driven by moisture and steam based on ion concentration gradients.

a Schematic illustration of a single moisture-electric generator (MEG) unit featuring AlgCa/Na network embedded within a PVA gel. The water gradient is defined as the difference in moisture content between upper and lower sides of the MEG sample. bThe open-circuit voltage (V_oc) (red) of an MEG device over time measured under open environment with fluctuating RH with synchronously recorded ambient RH(black). a, b Reproduced with permission from ref. ¹⁴. Copyright 2024, Springer Nature. c Structure of a single ionic hydrogel moisture-electric generator (IHMEG) device features asymmetric moisture penetration layers. d Continuous DC open-circuit voltage (V_oc) output (red) of an IHMEG device over time under ambient environmental conditions, with ambient relative humidity (blue) and temperature (black) recorded synchronously. c, d Reproduced with permission from ref. ²²⁶. Copyright 2022, Wiley. e Structure of a nanocellulose-based hydrogel moisture-electric nanogenerator (N-HMEG) with asymmetric moisture penetration layers. Reproduced with permission from ref. ²²⁷. Copyright 2024, Elsevier. f Schematic illustration depicts moisture-enabled electric generation in the bilayer of polyelectrolyte films (BPFs), which enables spontaneous water adsorption and ion dislocation from moist air and the schematic of the HMEG device. Reproduced with permission from ref. ²⁴⁵. Copyright 2021, Springer Nature. g Schematic diagram of a single Ion-Exchange Membrane Moisture-Electric Generator (IEM-MEG) (left) and working principle of IEM-MEG (right). Reproduced with permission from ref. ⁸⁷. Copyright 2025, Wiley. h Schematic illustration of the moisture-activated electricity generator. i Current density measurement of MEGs under different flow conditions. Synchronized flow (left), water flow only (center) and desynchronized flow (right). h, i Reproduced with permission from ref. ⁸⁵. Copyright 2025, Wiley.

By incorporating nanofillers such as carbon-based materials, inorganic particles, or nanofibers into polymer hydrogel networks, nanocomposite hydrogels can provide additional ion transport pathways while enhancing the mechanical strength, water retention, and ionic conductivity of the hydrogels²²⁸. Huang et al.²²⁹ proposed that SiO₂ and rGO fillers in SA hydrogels can significantly increase the ionic transport and generate a current output of 2.24 mA, which was significantly higher than that of the filler-free analog. An organic–inorganic hybrid generator with a hygroscopic polyacrylamide hydrogel on silicon nanowire arrays produced 1.28 V at 60% RH and 35 °C, retaining 60% of its output after 800 h—about 3 times higher than comparable devices²³⁰.

This type of MEG typically comprises a pair of charge-extracting electrodes and a sandwiched hygroscopic functional layer, relying on hygroscopic moisture-exposure and gradient-driven architectures for ion diffusion-based power generation. The most typical strategy to create water content gradient within the functional layer is one-sided moisture exposure. The top surface of the functional layer is directly exposed to air, while the bottom surface is sealed. This forms an internal unidirectional gradient which drives free ion transport to geneloading enhances optical absorptiorate external potential differences and currents. The morphology and hygroscopicity of the top electrode critically affect the MEG electric performance. In order to induce a strong potential gradient, heterogeneous electrodes are preferred to be utilized in MEG devices. As a key component for enhancing the electrical performance, the metal faces a contradiction where the electrode area (optimizing current collection) and the exposure area (ensuring water absorption to maintain the gradient) cannot be maximized simultaneously. Zhu et al.²³¹ reported that a top metal electrode covering 30% of the top surface balances current collection and water absorption, achieving a maximum current density of 1.00 mA cm⁻². In addition, porous graphite papers²³², porous metal films^14,233,234, and metal-coated textiles²³⁵ were utilized as electrodes for efficient moisture absorption. However, MEG with one-sided exposure architecture always reaches the water content equilibrium eventually, resulting in a decrease in its power generation. Dual-side moisture exposure architecture consisting of absorption and evaporation layers was proposed to achieve the continuous cycle of moisture adsorption and evaporation. The evaporation layer (e.g., PVDF-coated polymer fiber membranes^236,237, carbon-based porous materials^{216,238,239,240}, and MXene aerogels²⁴¹) typically features pore connectivity and low surface energy, facilitating rapid moisture release.

Gradient-driven architectures typically refer to the functional layer structure design, encompassing pre-introduced functional gradients and salt concentration gradients. Zhao et al.⁷⁷ first demonstrated the construction of an oxygen-content gradient in GO membranes using the moisture-electric annealing technique. The side with higher oxygen content adsorbs more water, releasing more H⁺ ions from the functional groups and thereby creating an ion concentration difference across the membrane. However, these GO-centric modifications are confined to films/membranes and involve complex manufacturing processes. By contrast, forming a functional-group charge gradient is intrinsically more straightforward in bilayer polyelectrolytes, where opposing fixed charges are segregated into distinct layers, yielding a built-in interfacial asymmetry without elaborate surface modification^242,243,244. A typical example is proposed by Wang et al.²⁴⁵, they combine polycation (polydiallyl dimethyl ammonium chloride (PDDA)) and polyanion (polystyrene sulfonic acid (PSS) and polyvinyl alcohol (PVA) hybrid film (PSSA)), which spontaneously adsorb water from humid air and dissociate Cl⁻ from PDDA and H⁺ from PSSA to establish the ion concentration gradient between two layers (Fig. 6f). Furthermore, salt-concentration-gradient can be created directly within the functional layer during fabrication or self-maintained under asymmetric sorption, thereby establishing a persistent chemical potential for directional ion diffusion⁸⁵.

Zhang et al.⁸⁷ proposed a laminated structure that comprising a cation-selective Nafion membrane sandwiched between hydrogels with disparate LiCl concentrations (4 M/0.01 M) (Fig. 6g), which can establish a Li⁺ gradient and selective Li⁺ diffusion, enabling directional cation migration. Additionally, Kim et al.⁸⁵ employed Marangoni-assisted gelation to fabricate a LiBr-rich-to-lean gradient from the top to the bottom of the Fumion FAA-3 hydrogel (FAA gel) (Fig. 6h). By synchronizing this gradient with the direction of water flow, their MEG delivered a current density 15-fold higher than that of pure FAA gel under 80% RH (Fig. 6i). Even under conditions of limited moisture, the device can still generate electricity through its inherent salt concentration gradient.

The power generation duration of ion concentration gradient-based MEGs typically ranges from seconds to days^{14,17,18,19,20,21,22,23,24}. However, a key challenge is that the ion gradient progressively weakens or even dissipates entirely as directional ion migration continues. This results in saturation of the water content and substantial decay in electrical output, thereby limiting the long-term operational stability^{14,17,18,19,20,21,22,23,24}. Recently, various effective mitigation strategies have been proposed to address this issue. One notable approach is the previously mentioned construction of dual-sided moisture exposure architecture, which consists of an absorption layer and an evaporation layer. This design can generate a persistent, running-water circle, thereby counteracting the depletion of the ion concentration gradient^222,246. As reported by Tan et al.²²², MEGs based on this structure can produce a continuous voltage of 0.78 V and a current of 7.5 μA (25 °C, 60% RH) for more than 10 days. Another strategy is based on ionic diode rectification, which selectively filters ions using negatively charged or positively charged nanochannels. It allows ions to migrate in one direction while substantially suppressing reverse diffusion, thereby maintaining an asymmetric charge distribution over extended periods of operation. Zhang et al.²¹² implemented an ionic diode–type hybrid membrane composed of carbon nanotubes (CNT) and anodic aluminium oxide (AAO) in MEGs, forming a PN-junction-like built-in electric field and thus enabling selective ion separation and steady-state unidirectional ion transport. The resulting devices can generate continuous electrical output for at least one month. Building upon this strategy, Fu et al.²⁴⁷ further integrated electrode regulation to develop a durable MEG that combines directional ion transport with interfacial charge management, enabling continuous power generation for up to 1240 h under ambient humidity. In this design, redox-active electrodes were employed to consume accumulated ions via ion intercalation or redox reactions, rather than allowing passive charge buildup, thereby effectively suppressing electrostatic screening and current decay. Moreover, Duan et al.²⁴⁸ proposed a photocatalysis-assisted ion gradient reconstruction strategy by introducing a photocatalytic layer beneath the moisture-absorption layer. Under light illumination, photogenerated charge carriers drive the hydrogen evolution reaction, consuming accumulated H⁺ ions and releasing H₂, thereby reconstructing the ion gradient and extending continuous current output to 650 h.

Thermoelectric harvesting

Steam-driven thermoelectric harvesting exploits a temperature differential between a vapor source and a cooler reservoir to drive Seebeck-effect conversion of heat into electricity^106,107. A vertical solar cogeneration arrangement integrates a photovoltaic (PV) module, a thermoelectric (TE) module, seawater reservoirs, highly absorptive evaporating paper, and a hollow partition that channels vapor to the TE hot junction (Fig. 7a). In this configuration the PV layer supplies energy for vaporization, the hollow channel directs steam to the TE hot side, and seawater serves as the cold sink that sustains the thermal gradient required for Seebeck-based power generation²⁴⁹. Material selection and device geometry govern thermal coupling and parasitic losses in the upright layout. Evaporative paper and the PV stack should exhibit favorable thermal conductivity and vapor diffusivity to promote evaporation and conductive/convective heat transfer to the TE element. Reducing the air gap between PV and TE lowers convective resistance and evaporation impedance, increasing steam-mediated heat flux to the TE hot side and thereby enlarging the temperature differential; the referenced system produced ≈3.6 W m^–2 thermoelectric power under the reported conditions (Fig. 7b, c)²⁴⁹.

Fig. 7: Energy harvesting driven by moisture and steam utilizing the thermoelectric effect.

a Structural schematic diagram and spectrum energy distribution of the solar-driven photovoltaic-steam-thermoelectric-steam (PV-S-TE-S) system. b Evaporation rate of the PV-S-TE-S system under real outdoor solar irradiance conditions. c Variation of circuit voltage and output power of the TE devices over time. (a-c) Reproduced with permission from ref. ²⁴⁹. Copyright 2024, Elsevier. (d) Schematic diaphragm of the air–water interface TE with coated CNT-CNC nanocomposite (PCC sponge)(left), and temperature profiles of air–water interface with and without PCC sponge under light irradiation for 30 min at an optical density of 1 kW m⁻². e Evaporation rates of the PCC sponge adapted in different molds and self-containing configurations under an optical density of 1kWm⁻². f Temperature difference between two sides of TE module at various solar irradiations. d–f Reproduced with permission from ref. ²⁷. Copyright 2019, Wiley. g Schematic diagram of air–water interface device using MnO₂-decorated cotton as hot layer of device(left), and the change of evaporation rates and conversion efficiency for MCx membranes under 1 kW m⁻² irradiation(right). h Surface temperature of the TE device with lower surface of 25 °C under various solar irradiation intensities. i Open-circuit voltage of the TE device with lower surface of 25 °C under various solar irradiation intensities g–i Reproduced with permission from ref. ²⁸. Copyright 2023, Wiley.

Air–water interface TE architectures exploit the temperature contrast between a photothermally heated evaporating surface and bulk water, mitigating adverse effects of large air gaps^27,28. An example mounts a PDMS sponge coated with a CNT–CNC nanocomposite (PCC sponge) onto the TE hot side to serve as a broadband absorber and thermal conduit, creating a temperature difference between sponge and the bulk water. (Fig. 7d)²⁷. Increased CNT/CNC loading enhances optical absorption, surface thermal conductivity, and evaporation flux: at 2 mg mL^–1 the evaporation rate reached 1.35 kg m^–2 h^–1 (≈ 1.02 kg m^–2 h^–1 above the uncoated baseline) and the sponge surface temperature rose to ≈43.7 °C (≈10.9 °C higher than the control), yielding markedly improved light-to-steam conversion efficiency (≈87.4% versus 19.1% without the sponge) and enhanced TE output. Geometric design of the sponge modulates effective evaporation area; larger effective areas (parallelogram, circular, or curved-square) correlate with higher evaporation rates (Fig. 7e). With increasing solar energy (optical density in the experiment), the temperature difference between the TE module (Fig. 7f) and the voltage output increases.

An alternative air–water interface implementation employs MnO₂-decorated cotton (MC_x) as the photothermal hot layer with bulk water as the cold sink²⁸. MnO₂ nanoparticles are formed on cotton via KMnO₄ dipping and in situ reduction by cellulose constituents; immersion time controls MnO₂ loading. Higher MnO₂ coverage significantly improves broadband absorption and photothermal conversion, increasing evaporation rate and surface temperature (Fig. 7g): the high-loading MC10 sample reached 2.24 kg m^–2 h^–1 and ≈70.2 °C under 1 Sun after 8 min, compared with 1.13 and 0.86 kg m^–2 h^–1 (and ≈40 °C for lower-load samples), demonstrating the direct influence of photothermal loading on performance (Fig. 7h, i).

Steam-based TE systems couple photothermal evaporators to TE hot sides while using bulk water as the cold reservoir; performance is dictated by absorber optical/thermal properties, evaporative flux, thermal coupling, and evaporator geometry. Optimization of absorber composition, porosity, thermal conductance, and module spacing is therefore essential to maximize temperature differentials and improve thermoelectric conversion efficiency for steam-driven energy harvesting.

Thermo-osmotic and thermoelectrokinetic conversion

Thermo-osmotic electricity harvesting exploits ion thermal diffusion induced by a temperature gradient, principally the Soret effect, where ionic species migrate from hot to cold regions and generate a thermovoltage^117,118. In steam-based implementations, water vapor establishes a thermal bias across an ionic medium; evaporation amplifies the temperature gradient, thereby enhancing ionic thermal diffusion and enabling conversion of thermal to electrical energy. The resulting thermoelectric potential arises from combined ionic Soret transport and redox contributions when redox couples are present²⁵⁰. Materials for efficient thermo-osmotic conversion therefore require large ionic Seebeck coefficients, which depend on differences in heat transport and entropy between anions and cations; ionic diffusion coefficients and migration entropy govern the thermal migration rate²⁵¹. Ionic conductors can exhibit substantially larger Seebeck coefficients than electronic counterparts^31,32, for example, values approaching ~66.7 mV K^–1 have been reported²⁵².

Thermo-osmotic materials are commonly classified as liquid, quasi-solid, or solid state, with quasi-solid and solid matrices preferred for encapsulation and leakage avoidance²⁵¹. One representative architecture employs polyethylene oxide (PEO) end functionalized with hydroxyl groups and doped with NaOH to form alkoxide end groups (-C–O–Na⁺), yielding a pronounced ionic Seebeck response (≈10–11 mV K^–1) in a sandwich cell (Fig. 8a)²⁵³. In that configuration the PEO–NaOH electrolyte occupies a confined chamber bounded by planar Au electrodes; measured thermovoltage scales linearly with the imposed temperature difference, indicating the central role of temperature on device output.

Fig. 8: Energy harvesting driven by moisture and steam through thermos-osmotic and thermoelectrokinetic processes.

a Schematic illustration of an ionic thermoelectric supercapacitor device of PEO doped with NaOH, equipped with an Au/CNT electrode. Reproduced with permission from ref. ²⁵³ Copyright 2016, Royal Society of Chemistry. b Molecular structures and fabrication process of PAA-PEO-NaCl ionic hydrogels. c Thermopower of PAA-PEO-NaCl hydrogels with varying salt concentration under different temperature gradients. d Ionic conductivity and Seebeck coefficient of the PAA-PEO-NaCl ionic hydrogels with different NaCl content. Inset: schematic diagram illustrating the setup for measuring ionic thermovoltage. b–d Reproduced with permission from ref. ²⁵⁴. Copyright 2023, Wiley. e Schematic of charging and discharging cycle of the Ionic Thermoelectric Supercapacitor (ITESC). (i) establishing the temperature gradient(16 K) to induce thermovoltage, (ii)charging an electrochemical capacitor through an external load, (iii)removing the gradient and isolating the device, and (iv)discharging (both with R_load = 100 kΩ). Reproduced with permission from ref. ²⁵³. Copyright 2016, Royal Society of Chemistry. f Seebeck coefficient of PVA-based materials with different concentration (PVA 10% (gray), PVA 15% (red) and PVA 25% (blue)) at different humidity. Reproduced with permission from ref. ²⁵⁵. Copyright 2024, Elsevier. g Design and synthesis of the moist thermoelectric generator (MTEG): polyelectrolyte membrane was prepared via copolymerization of AMPS and SSS under UV light, then encapsulated between carbon cloths coated with carbon nanotubes and assembled into the MTEG, electricity generated by the dislocation and diffusion of Na⁺ and H⁺ ions under hygrothermal synergistic system.(top), and the voltage output of the MTEG remain stable for 118 h under 60% ΔRH and 15 K temperature difference (Environment temperature at 70 °C) and the image of the LED bulb light up by using waste steam(bottom). Reproduced with permission from ref. ²⁹. Copyright 2023, Wiley. h Schematic illustration of the energy harvesting using asymmetric SPEEK/PES blend membrane subjected to temperature gradient(top), and the output power densities of the membrane under different concentration gradient ranging from 5 to 1000, with top side facing the low concentration (LiBr) solution (0.01 M) and output power densities of the asymmetric SPEEK/PES blend membrane with a 50-fold concentration difference under various ΔT from −30 to 30 °C (bottom). Reproduced with permission from ref. ²⁵⁶. Copyright 2021, Springer Nature.

Polyelectrolytes markedly enhance ionic Soret responses: mixed polyelectrolyte–salt systems can present Soret coefficients orders of magnitude larger than simple salt solutions²⁵¹. For instance, an ionic hydrogel composed of poly(acrylic acid) (PAA) and PEO blended with NaCl shows temperature-dependent thermovoltages (Fig. 8b) and reveals a strong dependence of ionic Seebeck coefficient on salt concentration (Fig. 8c)²⁵⁴. The Seebeck coefficient increases with NaCl concentration up to an optimum (~3 M, ~3.26 mV K^–1), beyond which ion-pairing and screening reduce thermopower. Concurrently, excessive salt loading diminishes ionic conductivity—attributable to reduced water content and suppressed ion dissociation and mobility (Fig. 8d). Polymer matrices also enable beneficial properties such as self-healing and adhesion; electrode architectures (e.g., Ni-foam/CNT) can further suppress noise while improving capacitance and interfacial resistance.

Ionic thermo-osmotic systems can be combined with energy storage to form thermally driven supercapacitors (Fig. 8e). Operation proceeds by (i) establishing a temperature gradient that induces a thermovoltage, (ii) charging an electrochemical capacitor through an external load, (iii) removing the gradient and isolating the device, and (iv) discharging the stored energy to an external circuit.

Salt-free designs also exist: a moisture-activated PVA hydrogel exhibited an ionic Seebeck coefficient of ~9.26 mV K^–1 by harnessing ambient humidity, with optimal performance at ≈60% RH and PVA mass fractions of 10–15 (Fig. 8f)²⁵⁵. Excess humidity or excessive PVA packing reduces free water and charge mobility, degrading stability and Seebeck response.

Hybrid devices that couple thermoelectric and moisture-driven mechanisms can extend operation under steam-rich conditions. A moist thermoelectric generator (MTEG) employing a crosslinked copolymer membrane of AMPS and PSSS permits concurrent H⁺ and Na⁺ migration under combined temperature and humidity gradients, producing sustained voltages (≈1.8 V V_oc for ∼27 h) by balancing evaporation and moisture transport (Fig. 8g)²⁹. In a complementary approach, a LiBr solution paired with an asymmetric sulfonated PEEK/PES membrane and LiMn₂O₄/CNT electrodes yielded enhanced ionic currents; isopropyl alcohol (IPA) treatment reorganized the SPEEK/PES membrane microstructure and further increased ion transport (Fig. 8h)²⁵⁶, also with large concentration and temperature gradient, which can further increase the power density of the device.

Thermo-osmotic energy conversion from steam couples Soret-driven ionic migration, material-dependent ionic Seebeck responses, and device architectures that manage water, ion concentration, and interfacial redox to maximize thermopower. Key optimization levers include polymer matrix selection, salt/polyelectrolyte composition and concentration, electrode architecture, and membrane microstructure engineering to balance high Seebeck coefficient, sufficient ionic conductivity, and mechanical/operational stability.

Water-flow- and droplet-driven energy harvestings

Electromagnetic transduction

Water-flow electromagnetic (EM) energy harvesters transduce hydrodynamic motion into electricity by driving a relative translation or rotation between a permanent magnet and a coil, thereby varying magnetic flux through the winding^{33,34,35,36,37,38,137,138,139,140,257}. High performance follows from maximizing flux linkage and dΦ/dt: tight gaps between magnet and coil, concentrated magnetic paths, and adequately high turn counts increase the induced EM field. Because the transduction pair consists of a permanent magnet and a coil, the material choices are accordingly focused. Neodymium-iron-boron (NdFeB) magnets provide high remanence, magneto-crystalline anisotropy, and the highest maximum energy product (>400 kJ m⁻³) among commercial magnets²⁵⁸, enabling compact sources of strong flux, while copper coils offer low resistivity and nonmagnetic behavior for minimal magnetic drag³⁴. Therefore, solenoids with hundreds to thousands of turns or lithographically defined micro-planar coils are used to realize dense windings and compact gaps.

Furthermore, practical electrical power depends on hydrodynamic coupling (e.g., rotor torque or wake amplitude from vortex vibration) and is maximized by regulating the local through-flow velocity, and by coupling the vortex shedding frequency to the resonance of the device structures¹³⁷. In pipeline-type devices^34,35,36, a turbine or impeller converts through-flow into rotor motion of a magnet that sweeps a time-varying EM field across stator coils with multi-phase layouts (Fig. 9a). By utilizing a water-meter housing, Hoffmann et al. implemented a two-pole ring magnet with three iron-core coils, achieving up to 720 mW at 20 L min⁻¹ and still ~2 mW at 3 L min⁻¹ by reducing the threshold flow through optimization of the impeller and magnetic circuit (Fig. 9b)³⁵. With micro-planar coils patterned on the inner surface of the cubic holder (Fig. 9c), Ahmad et al. showed that series connection increased V_OC and power to 65 mV and 548 µW at 7 L min⁻¹, demonstrating efficacy at low flow (Fig. 9d)³⁶.

Fig. 9: Energy harvesting from water flow and droplets via electromagnetic transduction.

a A turbine or impeller converts the fluid flow into rotational motion of a rotor containing a magnet, which then move past stator coils to generate a time-varying electromagnetic field. b Structure of a rotational radial-flux energy harvester featuring a two-pole ring magnet(left), with the magnetic field distribution within the energy harvester structure(center), and the design of a rotational energy harvesting based on a three-phase generation principle (right). Reproduced with permission from ref. ³⁵. Copyright 2013, IOP. c An exploded view of the micro planar coil and magnet housing of the MPC-based hydrokinetic energy harvester(left), and the cross-section view of the water inlet and outlet of the device(right). d Output power as a function of load resistance for combined coils. c, d Reproduced with permission from ref. ³⁶. Copyright 2024, Wiley. e Sketch of vortex-induced design, with an obstacle inside the water flow, harvesting the periodic lift or pressure oscillations. f Optical image(top), and working principle(bottom) of an electromagnetic energy harvester. g Experimental result of induced voltage of the coil(left), and power spectral density corresponding to the induced voltage of the coil(right). f, g Reproduced with permission from ref. ³⁸. Copyright 2012, Elsevier. h Schematic diagram of the water energy harvester (left). Photo of experimental setup of harvester, magnetic core and coil part of the prototype(right). i Peak to peak output voltage and power as a function of load resistance at a flow velocity 0.409 m/s. Inset: a typical voltage waveform across the matched load resistance. h, i Reproduced with permission from ref. ³³. Copyright 2017, Wiley.

In addition, vortex-induced designs, placing an obstacle inside the water flow, harvest the periodic lift or pressure oscillations of a von Kármán vortex street, with output maximized when the shedding frequency set by Strouhal scaling locks to a device resonance^33,37,38,137 (Fig. 9e). In diaphragm-type implementations^37,38, wake pressure deflects a thin membrane coupled to a small magnet, moving it relative to a fixed coil. Wang et al. reported ~11 mV and 0.4 µW under 0.254 kPa excitation at 30 Hz using a polyester diaphragm³⁷. Moreover, as shown in Fig. 9f, with a trapezoidal bluff body upstream of a PDMS diaphragm to tune excitation frequency, amplitude, and transducer placement, his team obtained ~20 mV and 1.77 µW at 0.3 kPa and 62 Hz, notwithstanding the compact volume (37.9 cm³) of the device (Fig. 9g)³⁸. Alternatively, Lin et al. directly exploited the vortex with an asymmetric arc-shaped elastic beam, inducing controlled flow separation that intensifies the pressure difference across the beam under low volumetric flux (Fig. 9h)³³. By affording tunable natural frequencies, a prototype with compact size (152.4 cm³) can realize a V_OC of 1.44 V and a maximum power of 0.503 mW at 0.409 m s⁻¹and an external load of 110 Ω (Fig. 9i). This outcome underscores the impressive potential of resonance-matched, low-flow EM harvesters for application in wireless sensor networks.

Piezoelectric transduction

Piezoelectric energy harvesters used for small-scale water energy harvesting are commonly engineered in the form of thin films^{40,155,259,260,261} or cantilever structures^{39,262,263,264}, encapsulated with a hydrophobic protective layer to ensure environmental stability. When subjected to fluid oscillations or droplet impacts, the piezoelectric material undergoes mechanical deformation, leading to a change in polarization and the generation of an electrical current. This design enables efficient mechanical-to-electrical energy conversion even under dynamic or wet conditions, making it well-suited for small-scale water energy harvesting applications^141,146.

In the harvesting of flow-induced vibration energy, vortex-induced vibration provides an effective mechanism to drive piezoelectric structures into resonance for efficient energy conversion^39,153,265. Sherrit et al. introduced a spline nozzle configuration by placing two semicircular obstacles on both sides of the flow channel³⁹. As the fluid passes through, the flow field periodically separates, forming a stable vortex street in the downstream region and generating cyclic pressure fluctuations (Fig. 10a). These fluctuations act on a commercial piezoelectric bimorph cantilever (V21BL) mounted within the channel, driving it into resonance at specific flow velocities and enabling efficient mechanical-to-electrical energy conversion. However, the formation of stable vortices requires specific hydrodynamic conditions, limiting the practical deployment of such systems in variable or low-Reynolds-number flows^155,266. Consequently, researchers have explored alternative excitation mechanisms such as turbulence-induced vibration^150,267, cavity-flow oscillations¹⁴⁹, and continuous hydrodynamic impacts^{268,269,270,271,272} to enhance the adaptability of piezoelectric harvesters under diverse flow conditions. Moreover, due to the low output power and poor durability of commercial piezoelectric bimorph materials, researchers have increasingly turned to flexible polymer-based piezoelectric materials, particularly poly(vinylidene fluoride) (PVDF)^{40,41,272,273,274}, as the core component of piezoelectric transducers. The fabrication of PVDF is highly customizable, allowing its structural design to adapt to various excitation modes of water flow. Thus, it can effectively harvest energy from wave motion^273,275,276, turbulence-induced vibrations^277,278, or continuous hydrodynamic pressure impacts²⁷². Moreover, by incorporating functional piezoelectric nanofillers into the PVDF matrix, the β-phase content and dipole orientation can be significantly enhanced, thereby improving its piezoelectric performance. Mankuni et al. developed a flexible PVDF–Ti₃C₂ MXene nanocomposite piezoelectric nanogenerator using a simple solvent-casting method, achieving a self-poled high β-phase structure without any external poling treatment⁴⁰ (Fig. 10b). The composite film containing 1 wt% MXene exhibited the highest β-phase content (~72.5%) and enhanced polarization properties. During underwater flow energy harvesting tests at a flow velocity of 3 m/s, the device generated a peak-to-peak voltage of 21 V and a short-circuit current of 0.9 μA, corresponding to a maximum power of approximately 36.8 μW, while maintaining excellent flexibility and stability. Furthermore, piezoelectric harvesters based on nanowires or nanofibers have demonstrated superior electromechanical coupling and scalability. Their one-dimensional morphology and high aspect ratio enable larger strain responses and improved dipole alignment under low mechanical stress. Wang et al. integrated electrospun PVDF nanofibers into microfluidic systems to develop a self-powered viscosity and pressure sensor that simultaneously harvested flow energy and monitored droplet dynamics⁴¹ (Fig. 10c). When microfluid flowed across the suspended PVDF nanofiber layer, periodic fiber deformation produced transient voltage pulses up to 1.8 V, sufficient for droplet and bubble counting within the channel. The output voltage amplitude exhibited a negative correlation with both the fluid viscosity and input pressure, allowing the device to function as a self-powered microfluidic sensor for real-time flow characterization. The system maintained stable operation over 1000 s of continuous flow, confirming its durability and reliability in microscale liquid environments. Meanwhile, hybrid energy harvesting has recently emerged as an important strategy for enhancing output performance. By integrating piezoelectric transducers with other energy conversion mechanisms such as TENGs^{279,280,281,282} or EMG^283,284, the overall energy utilization efficiency can be significantly improved, allowing simultaneous harvesting of multiple forms of mechanical energy and achieving higher output power and stability.

Fig. 10: Energy harvesting from water flow and droplets via piezoelectric transduction.

a High-speed camera snapshots (frame rate = 1200fps) capturing the motion of a standard V2IBL actuator within a flow harvester at a flow rate of 14 L/min(left), and a photo of the Armored (2 and 5 mil) QuickPack (QP21B) cantilevers with steel shim(right). Reproduced with permission from ref. ³⁹. Copyright 2015, SPIE. b Schematic of the experimental setup (left), alongside an illustration depicting the interaction between PVDF chain and the MXene nanosheet (right). Reproduced with permission from ref. ⁴⁰. Copyright 2024, American Chemical Society. c Top-view schematic and image showing the motion of PVDF NFs with flowing droplets or bubbles(top), and the output voltage and input pressure curve(bottom). Reproduced with permission from ref. ⁴¹. Copyright 2017, American Chemical Society. d The oscillating motion of a droplet causes continuous bending of the piezocantilever generating electricity(left), and the experimental results show the bending displacement of voltage output of the piezocantilever actuated by acoustically oscillating droplets at various frequencies and distances(right). Reproduced with permission from ref. ⁴². Copyright 2015, Elsevier. A. Physical. e Schematic diagram of the piezoelectric beam on moment before impact (top) and after impact(right). Reproduced with permission from ref. ⁴³. Copyright 2017, SAGE. f Model of the substrate structure (left) and snapshots at different impact velocity, showing rebounding height of droplet with an impact velocity of 1.5 m/s and splash of droplet with an impact velocity of 2.55 m/s under the film tension effect (right). Reproduced with permission from ref. ⁴⁴. Copyright 2021, Elsevier.

In droplet energy harvesting, since the kinetic energy of the droplet is primarily transferred through its impact on the material surface, cantilever-based structures are particularly suitable for piezoelectric harvesters^{42,43,44,285,286}, as they can effectively convert the transient impact force into mechanical deformation and thus electrical energy. Lee et al. further demonstrated that when liquid droplets are acoustically excited near their natural oscillation frequencies, their periodic deformation can continuously drive the bending of a piezocantilever, generating electrical power through the piezoelectric effect⁴². The amplitude of oscillation and resulting voltage output were found to depend strongly on the applied frequency, droplet size, and distance from the acoustic source, achieving a maximum power output of approximately 80 μW for a 6 μL droplet (Fig. 10d). Complementarily, Wong et al. modeled and experimentally analyzed the transient dynamics of a piezoelectric beam under successive raindrop impacts, accounting for water-layer accumulation effects⁴³. Their results revealed that the added water mass reduces the beam’s natural frequency and voltage output, while the optimal energy-harvesting performance occurs before the water layer spreads to the beam’s width ends (Fig. 10e). Compared with the traditional single-sided cantilever configuration, Hao et al. proposed an innovative fixed–fixed hybrid piezoelectric structure that combines PVDF beams with an elastic film substrate, aiming to enhance the mechanical adaptability and electrical output of droplet-based harvesters⁴⁴. This composite design effectively integrates the advantages of cantilever and fixed–fixed configurations, allowing both large deformation and high-frequency response under droplet impact (Fig. 10f). The elastic film serves as a dynamic coupling medium, transferring and amplifying the transient impact energy to the piezoelectric beams at both ends. Similar to flow-induced vibration energy harvesting, droplet-based piezoelectric harvesters primarily rely on internal mechanical vibrations to convert the kinetic energy of droplet impacts into electricity. However, during the energy conversion process, contact and separation between droplets and the material surface are accompanied by significant interfacial charge transfer and electrostatic induction. When relying solely on the piezoelectric material, the surface electrostatic potential generated by contact electrification remains underutilized¹⁷⁷. To address this limitation, researchers have proposed the concept of a hybrid piezoelectric–triboelectric energy harvester^156,287,288. This configuration integrates a piezoelectric transducer with a TENG, enabling simultaneous energy harvesting through the strain-induced polarization of the piezoelectric layer and the electrostatic induction at the liquid–solid interface during droplet impact²⁸⁹. Through this coupled mechanism, the kinetic energy of a single droplet can be simultaneously converted into both bulk piezoelectric energy and surface triboelectric energy, thereby significantly enhancing the overall energy conversion efficiency and output power, and achieving more efficient and stable droplet energy harvesting.

Triboelectric transduction

Triboelectric transducion convert the kinetic energy carried by small-scale water sources—such as droplets, flow-induced vibrations, or fluid impacts—into electrical energy through electrostatic induction and contact electrification^{47,54,290,291,292,293}. In the energy harvesting process, TENGs generally operate in two modes. The first is the direct-contact mode, in which the water flow or droplets directly interact with the TENG surface, generating a periodic contact-separation process^{56,294,295,296,297}. The second is the indirect-contact mode, where the liquid acts as a mechanical excitation source, applying pressure or impact to deform structural components thereby inducing contact and separation between internal triboelectric layers^{45,298,299,300,301}.

In the direct-contact mode, CE occurs at the liquid–solid interface when the liquid first touches the solid surface (Fig. 11a)⁵⁶. At this moment, overlapping electron clouds at the atomic scale induce electron transfer from the liquid to the solid, resulting in surface charging¹⁷². To maintain electrostatic equilibrium, counterions in the liquid then redistribute and accumulate near the charged surface, forming the EDL. The EDL consists of a tightly bound surface charge layer and a diffuse ionic layer within the liquid, together establishing the interfacial potential that governs subsequent electrostatic induction and energy conversion³⁰². For conventional EDL-based TENGs, the typical configuration places the electrode beneath a dielectric layer, where the sliding contact and separation of flowing water induce charge transfer and generate electrical output signals^{54,295,303,304}. Among the factors affecting performance, the water flow rate plays a crucial role. Cheng et al. designed a hybrid triboelectric nanogenerator (TENG) with a waterwheel-like structure, in which the blades were covered with nanostructured PTFE thin films⁵⁶. When flowing water impacted the superhydrophobic PTFE blades, it not only generated electrostatic energy through the water–solid contact electrification effect but also drove the rotation of the wheel to harvest mechanical kinetic energy (Fig. 11b). Under a flow rate of 54 mL s⁻¹, the water-TENG achieved an open-circuit voltage of 72 V and a short-circuit current of 12.9 μA. The overall system exhibited an instantaneous power density of 0.59 W m⁻², sufficient to light up 20 commercial LEDs or charge a 4.7 μF capacitor to 13 V within 326 seconds.

Fig. 11: Energy harvesting from water flow and droplets via triboelectric transduction.

a Schematic of the direct contact mode triboelectric transduction, where liquid contact with the solid surface induces contact electrification. b Structural diagram of the hybrid device(left), schematic of the operational mechanisms for the water-based TENG and disk-based TENG component in the hybrid system(center), and the voltage (V_oc) and current (I_sc) versus time curve for the disk-TENG(right). Reproduced with permission from ref. ⁵⁶. Copyright 2014, American Chemical Society. c Exploded diagram of the DE-TENG(left), with a schematic illustrating water level variations relative to the device and corresponding transferred charge during a single operational cycle(right). Reproduced with permission from ref. ³⁰⁷. Copyright 2023, Wiley. d Diagram of a water-tube-based triboelectric nanogenerator (WT-TENG) and its operational mechanism(top). The ‘WAVE’ pattern demonstrates WT-TENG powering 103LEDs in seesaw mode, and water waves lighting up 150 LEDs in Victoria Harbor, Hong Kong(bottom). Reproduced with permission from ref. ⁴⁹. Copyright 2021, Wiley. e Schematic showing the operating mechanism of the PC-TENG tube, which relies on the formation and modulation of the electrical double layer (EDL). Reproduced with permission from ref. ³¹³. Copyright 2022, Elsevier. f Operating mechanism of the water TENG, where triboelectricity generation is primarily generated through contact electrification(left), and the output current produced by water TENG from water droplet, with a fixed volume of 30 μL and falls from a height of 90 cm(right). Reproduced with permission from ref. ⁵⁷. Copyright 2014, Wiley. g Schematic diagram of switchable TENGs featuring adjustable air gaps between the triboelectric layer and electrode, enabling control over the direction of electron flow(left), and the rectification efficiency varies with maximum spreading diameters(right). Reproduced with permission from ref. ⁵⁸. Copyright 2024,Wiley h Schematic illustration(left), and photo of its stretch state(right) of SH-TENG. Reproduced with permission from ref. ⁵⁹. Copyright 2023, Elsevier.

In addition, optimizing the design of the electrodes and the surface morphology of the dielectric material can further enhance the energy-harvesting capability of the TENG and increase the strength of its output signals²⁹⁵. Despite significant progress, traditional structures still suffer from problems such as fixed charge and sluggish interface response, making it difficult to achieve efficient energy conversion in complex hydrodynamic environments^305,306. To address this limitation, Liang et al. proposed a liquid-solid TENG based on a dynamic electrical bilayer, employing an asymmetric structure to achieve efficient water wave energy capture by constructing a relative dynamic EDL³⁰⁷. The device consists of a positively charged polypropylene (PP) membrane and a negatively charged fluorinated ethylene propylene (FEP) membrane (Fig. 11c)³⁰⁷. When water waves periodically immerse and detach the device, the EDL at the interface is continuously constructed and disrupted, driving free charges to flow back and forth between the electrodes, forming a stable AC output. Under optimized parameters (frequency 0.5 Hz, movement height 6 cm), it achieved a peak power density of 30.95 W m⁻³ and an average power density of 4.94 W m⁻³. Compared with open small-scale water flow energy harvesting, triboelectric transduction is also well suited for encapsulated water-flow energy harvesting systems^{49,160,308,309,310,311,312}. In such configurations, the liquid is typically enclosed within a sealed container, where it undergoes reciprocating motion under external mechanical excitation. Wu et al. encapsulated deionized water inside an FEP tube and wound two copper electrodes side-by-side on the outer surface of the tube, forming a water-tube-based TENG (WT-TENG)⁴⁹ (Fig. 11d). When the water inside the tube reciprocates due to external mechanical stimulation, the relative displacement between the water and the tube wall continuously changes the solid–liquid contact area, thereby generating alternating current. The advantage of this encapsulated liquid design lies in its ability to operate efficiently in multiple modes—including rotation, swing, seesaw, and horizontal linear motion—allowing it to harvest energy from various low-frequency mechanical sources such as ocean waves, human motion, and environmental vibrations. However, due to the one-dimensional linear characteristics of the tubular structure, the liquid flow is directionally constrained, leading to a significant decline in energy conversion efficiency under multidirectional or nonlinear excitations. In particular, during irregular wave motion or multi-axis vibration, the system cannot effectively capture kinetic energy from all directions. To overcome this limitation, researchers have proposed spherical³¹² or UFO-shaped^310,311 encapsulated triboelectric nanogenerators. Unlike tubular designs, these multidimensional encapsulations allow the internal liquid to flow freely in three-dimensional space, enabling effective energy harvesting under external forces from arbitrary directions. In addition, researchers also studied the continuous flow of water within the pipes. Munirathinam et al. designed a single-electrode tubular TENG that utilizes the continuous flow of tap water to achieve sustainable energy generation³¹³ (Fig. 11e). The results showed that as the flow rate increased from 0 to 2.5 L min⁻¹, both the output voltage and current rose correspondingly. This enhancement was attributed to the higher flow velocity allowing more charged water molecules to enter and leave the PTFE surface per unit time, thereby inducing a greater amount of charge on the electrode.

As another form of small-scale water, droplets are also well suited for energy harvesting based on triboelectric transduction^{53,57,60,242,314}. Lin et al. were the first to use the principle of TENGs to harvest mechanical energy from water droplets via a combination of contact electrification and electrostatic induction⁵⁷. In their design, when a droplet impacts an electrode surface coated with a PTFE film, an instantaneous charge transfer occurs at the solid–liquid interface, generating an alternating electrical signal through the external circuit (Fig. 11f). Experimental results showed that a single droplet could generate approximately 9.3 V of voltage and 17 μA of current, corresponding to a power density of about 20 mW m⁻². Building upon this foundation, researchers systematically investigated the influence of dielectric layer thickness³¹⁵, ionic concentration^63,316, and surface chemical properties on the charge-transfer dynamics^{64,65,290,317}, while developing various structural designs to more efficiently capture and convert the kinetic and interfacial energy carried by droplets. However, due to the inherent characteristics of conventional TENGs, which primarily generate AC signals, they are not directly compatible with wearable electronic devices that require direct current (DC) power sources, thereby limiting their practical applications^173,318. To overcome this limitation, Zhou et al. proposed a switchable TENG (s-TENG) based on the dielectric breakdown effect⁵⁸. In this design, a microscale air gap was introduced between the FEP triboelectric layer and the ITO electrode (Fig. 11g). By regulating the surface charge density and electric field intensity, localized air ionization occurs once the electric field exceeds the dielectric strength of air, triggering breakdown discharge and inducing unidirectional electron flow. Through this mechanism, a self-rectified DC output was achieved without the need for an external rectification circuit. Experimental results demonstrated that the device reached a rectification ratio of ≈133 and an energy conversion efficiency of about 26%, maintaining stable DC output under low-frequency (0.8–2.5 Hz) and random mechanical excitations. Furthermore, to better adapt to the needs of flexible and wearable electronics, Yang et al. developed a highly stretchable PTFE particle–reinforced droplet-based TENG⁵⁹ (Fig. 11h). The device was fabricated by embedding PTFE microparticles in a flexible Ecoflex substrate and combining it with liquid-metal electrodes, which significantly improved both mechanical flexibility and hydrophobic stability while maintaining excellent electrical conductivity. Even under a 500% strain, the device retained over 80% of its output performance, demonstrating outstanding mechanical robustness and environmental adaptability.

In the indirect-contact mode, since the driving force of TENG comes from the mechanical action of external water flow or waves, the friction material is no longer limited to the dielectric layer that is in direct contact with the liquid, but is extended to achieve energy conversion through solid-to-solid contact^{158,319,320,321,322,323,324}. Wang et al. first proposed a fully enclosed rolling spherical TENG³¹⁹. The device consists of a closed spherical shell containing a freely rolling nylon ball (Fig. 12a). The inner wall of the shell is covered with a polyimide film and connected to two electrodes. When water waves or external vibrations cause the device to oscillate, the nylon ball rolls back and forth between the two electrodes, repeatedly contacting and separating from the Kapton film, thus creating a periodic potential difference between the two electrodes and outputting alternating current. However, since the nylon balls used in this type of rolling TENG are made of a hard material, they only form point contact with the dielectric layer surface during the rolling process, resulting in a limited actual contact area and low triboelectric efficiency, which in turn limits the overall charge density and output performance. To address this issue, researchers proposed a soft-contact spherical TENG structure^158,320,321. This involves replacing the soft structure, such as silicone¹⁵⁸ or a liquid-containing flexible sphere³²¹, with the rigid sphere. This creates a deformable soft contact interface between the sphere and the electrode. This design significantly increases the actual contact area, allowing for the generation of more triboelectric charge in a single contact. The soft-contact spherical TENG using a liquid/silicone soft core exhibited an open-circuit voltage of approximately 288 V and a short-circuit current of up to 5 μA in a low-frequency (approximately 2 Hz) water wave environment (Fig. 12b), with a peak power density more than eight times higher than that of a rigid spherical inner liner structure¹⁵⁸.

Fig. 12: Energy harvesting from indirect-contact mode of water via triboelectric transduction.

a Schematic of the freestanding-triboelectric-layer-based nanogenerator (RF-TENG) enclosed rolling Nylon ball(left). Photograph displays 70 green LEDs powered by the rocking TENG within a wave system(right). Reproduced with permission from ref. ³¹⁹. Copyright 2015, Wiley. b Structural schematic of soft-contact model spherical triboelectric nanogenerator (SS-TENG) featuring a flexible rolling sphere, along with a photograph of a SS-TENG (scalar bar = 1.5 cm) (top), with a typical electrical output waveforms of the SS-TENG and LEDs illuminated by the SS-TENG submerged in water(bottom). Reproduced with permission from ref. ¹⁵⁸. Copyright 2019, Elsevier. c Schematic of a single TENG and multilayer stacking process and the conversion of cam rotation into lever translation. Reproduced with permission from ref. ⁴⁵. Copyright 2023, Wiley. d Schematic diagram of P-TENG components (top). Reproduced with permission from ref. ³⁰¹. Copyright 2019, Elsevier. And photograph of the pendulum triboelectric layer and the fully assembled P-TENG device(bottom). Reproduced with permission from ref. ¹⁵⁹. Copyright 2021, Cell Press.

In addition to spherical structures, researchers have also explored swing-based configurations for water energy harvesting to more effectively utilize the mechanical energy of water flow^45,48,325. Zhou et al.⁴⁵ designed a multi-layer stacked TENG (MLS-TENG), in which the fluid-driven rotational motion is converted into reciprocating linear motion of sliders, enabling efficient energy conversion from hydrodynamic motion to electricity (Fig. 12c). Benefiting from the multilayer-stacked design, the open-circuit voltage of the MLS-TENG increased significantly from 860 V (single-layer) to 2410 V, and an energy harvesting rate of 2 mJ min⁻¹ was achieved under real river flow conditions. Han et al. proposed a wave-driven linkage-mechanism TENG that converts the vertical oscillation of waves into rotational motion of a rotor through a floating plate and linkage system, thereby greatly enhancing energy transfer efficiency⁴⁸. In addition, the use of rabbit-hair brushes on the rotor effectively protected the dielectric surface from wear, significantly extending the operational lifetime of the device. Within swing-type designs, another approach is the pendulum-type TENG, which suspends a weight at the center of the device to generate periodic oscillations under external disturbances or wave excitations^159,301. This design can directly convert the reciprocating displacement of waves into periodic angular motion, driving contact–separation or sliding friction between triboelectric layers for energy harvesting (Fig. 12d)^159,301.

Other droplet-driven mechanisms

In addition to triboelectric transduction, a range of alternative mechanisms has been explored for harvesting energy from droplets^{51,53,60,62,186,326,327,328,329,330}. Xu et al. proposed a droplet-based electricity generator (DEG) with a bulk-effect transistor-like architecture⁶⁰. In this design, a conductive electrode (such as aluminum) was placed on the surface of the dielectric layer, enabling conversion of interfacial charge induction into a bulk effect and thereby achieving significant current amplification (Fig. 13a). Experimental results showed that the output voltage increased to 143.5 V, the current reached 270 μA, and the power density was approximately 2600 times that of conventional interfacial TENGs, sufficient to light up more than 400 commercial LEDs simultaneously. Building on this, researchers systematically analyzed how factors such as droplet impact position^187,331, dielectric material type^326,327,328, and layer thickness^60,186 affect the bulk-effect droplet electricity generator’s output performance. Meanwhile, Wu et al.⁶¹ employed high-speed imaging synchronized with electrical signal measurements to construct an electrohydrodynamic model describing the interaction between droplet impact and surface charging (Fig. 13b). Combined with an equivalent RC circuit analysis, this approach quantitatively described the energy conversion process and revealed the coupling relationship between droplet impact dynamics and electrical response⁶¹.

Fig. 13: Energy harvesting from water flow and droplets via non triboelectric transduction.

a Schematic diagram of the DEG(left), and comparison of the output current generated by the DEG(red) with that of a control device(black) under continuous impact of individual droplets(right). Reproduced with permission from ref. ⁶⁰. Copyright 2020, Springer Nature. b High-speed camera imaging and electrical signal measurement capturing the interaction between droplet impact and surface charging. Reproduced with permission from ref. ⁶¹. Copyright 2020, American Physical Society. c Schematic illustration of the Droplet-Based (DB-TENG) unit. Reproduced with permission from ref. ⁵⁰. Copyright 2021, American Chemical Society. d Schematic diagram (top) and optical image (bottom) of the floating-droplet-activated electricity generator (FDEG). Reproduced with permission from ref. ⁵². Copyright 2025, Cell Press. e Experimental setup for the tribovoltaic measurement along with the external circuit schematic(left), and the open-circuit voltage generated by DI water droplet slides on a P-type silicon wafer (0.1Ω cm) at a speed of 20 mm/s, with a static contact diameter of 2.5 mm(right). Reproduced with permission from ref. ⁵³. Copyright 2020, Elsevier. f Schematic illustration of a static M-S junction activated by light and water, alongside a dynamic M-S junction driven by sliding friction(left). Current–Voltage characteristics of a static Schottky MSM structure and electron cloud potential model depicting the electron transfer process from sliding state to separated state(center). Carrier transport behavior within the static MSM structure, governed by triboelectric potential and positive current response of MSM structure from impinging water droplet. Reproduced with permission from ref. ⁶². Copyright 2025, Wiley.

Beyond the kinetic energy of continuously falling droplets, researchers have also explored non-dropping droplet motions—such as oscillation, sliding, and vibration—as potential sources for energy harvesting. Wei et al. designed an all-weather droplet-based TENG, in which droplets were exposed to an open wave environment, enabling stable energy generation under conditions of high humidity, high salinity, and even acidic or alkaline media⁵⁰ (Fig. 13c). Building upon this concept, Zhou et al. introduced a transistor-inspired encapsulated Floating Droplet Electricity Generator that features a UFO-shaped floating body with asymmetric ring–disc electrodes⁵². In this configuration, the periodic rolling of droplets across the FEP dielectric layer alternately closes and opens the internal circuit (Fig. 13d). Under asymmetric capacitance conditions, differential charge accumulation occurs, substantially enhancing the electrical output. Operating at a wave frequency of 0.28 Hz, the system achieved a current output of 107.6 mA, a power density of 1.2 kW m⁻³, and a charge density of 31.2 mC m⁻³, while maintaining stable power generation under multidirectional wave excitations.

In the field of sliding-droplet energy harvesting, another effective mechanism is the tribovoltaic effect^{53,62,190,329,330}. Lin et al. were the first to apply this principle to a liquid–semiconductor interface, achieving DC generation through the frictional interaction between water droplets and a silicon surface. In their experiments, deionized water droplets sliding over silicon produced a voltage of approximately 0.4 V and a current of 0.3 μA, with the output polarity determined by the semiconductor doping type (p-type or n-type)⁵³ (Fig. 13e). Furthermore, Zheng et al. provided additional experimental verification of the carrier separation mechanism arising from the combined effects of interfacial electron transfer and the built-in electric field at the liquid–semiconductor interface³²⁹. With the continued advancement of research on the tribovoltaic effect, structural optimization of semiconductor–metal junctions has enabled remarkably enhanced performance. Recent designs have achieved an output charge of up to 25,500 nC and an energy output of 5.8 × 10⁻⁶ J from a single sliding droplet, demonstrating the strong potential of interface engineering for high-efficiency droplet-based direct current generation⁶² (Fig. 13f). Their findings established a solid theoretical foundation for liquid-based DC energy harvesting driven by the tribovoltaic effect.

Ice- and snow-driven energy harvesting

As a widely distributed form of water in the cryosphere, snow holds untapped mechanical energy potential through triboelectrification. Ahmed et al. presented the first all printable snow-based TENG, operating in single-electrode mode by friction and impact between snow particles and a UV curable silicone triboelectric layer with PEDOT:PSS electrodes (Fig. 14a)¹⁹⁶. This device delivers an instantaneous output power density of up to 0.06 and 0.2 mW m⁻², an open circuit voltage of 6 and 8 V, and a current density of 20 and 40 μA m⁻² in tapping and sliding scenarios, respectively (Fig. 14b, c)¹⁹⁶. Beyond energy harvesting, it functions as a self-powered multifunctional sensor for monitoring snowfall rate, snow accumulation depth, wind direction, and speed. Its remarkable stretchability (125%), lightweight, and conformability further enables wearable applications, such as tracking human motions in snow-related sports. The ice and snow environment of polar regions provides an ideal application scenario for thermoelectric power technology. Traditional solar and energy harvesters in these environments require bulky batteries and wind generators to harvest energy during the days with little sunshine (Fig. 14d) ³³². Jones et al. developed an upstand Seebeck Effect generator on the sea-ice by converting the temperature difference between air and ice water into electricity. With the device having the evaporator, intermediate, and the condenser section. The power output reached the highest voltage of 4.4 V and 360 mW with the evaporator immersed in the seawater and after 9 min, which shows the ability of energy harvesting during the winter period. Yuan et al. developed a road thermoelectric generator system (RTEGSs) based on Seebeck effect, harvesting thermal energy from roads to produce electricity³³³. Figure 14e shows the structure of the thermoelectric module. Three types of RTEGSs with different technologies methods were designed in this work. In winter, the pavement-subgrade system melted snow 60 min faster than conventional surfaces (Fig. 14e(i)). Summer road temperatures were reduced by 10 °C at approximately 17:00 with the pavement-ambient system (Fig. 14e(ii)). The pavement-flowing water system achieved a power density of 4.28 mW cm⁻³, while the highest power was 197.23 mW under operating condition 70-5-2.5 (Fig. 14e(iii)).

Fig. 14: Energy harvesting driven by ice and snow.

a Schematic diagram of snow triboelectric nanogenerator (snow-TENG), featuring the surface composed of UV-curable silicone and SEM image showing micropattern at various magnifications with scale bars of 100 μm and 50 μm respectively(top). Snow-TENG at different stretching conditions(bottom). b Schematic illustration of snow-TENG tapping testing setup, which includes a vertical linear motor, snow layer and fabricated snow-TENG device, along with the measured open circuit voltage and short-circuit density during tapping conditions. c Photograph of a snow-TENG device unit mounted on a bicycle with recorded open-circuit voltage and short-circuit current density during sliding operation. a–c Reproduced with permission from ref. ¹⁹⁶. Copyright 2019, Elsevier. d Traditional solar and wind power generator used in Ice environment(top). Power output of the Seebeck generator prototype device in ice water under quiescent conditions(bottom). Reproduced with permission from ref. ³³². Copyright 2011, Elsevier. e Overview of principle behind thermoelectric technology (top left). Power of RTEGS under different conditions of road surface, flowing water temperature and flowing water velocity (bottom left). RTEGS technology based on Seebeck effect, including (i) Pavement-subgrade RTEGs, (ii) Pavement-ambient RTEGS, and (iii) Pavement-flowing water RTEGS(right).Reproduced with permission from ref. ³³³. Copyright 2023, Elsevier.

Application of the small-scale water energy harvesting

Following the systematic analysis of the working principles, material selection, and structural design of different types of small-scale water energy transductions, it is evident that researchers have made significant progress in recent years. However, to achieve the leap from experimental verification to practical application, further exploration of its functional realization and potential in different scenarios is needed. Therefore, this section will primarily explore the applications of small-scale water energy in several important field, including energy harvesting, self-powered sensors, self-powered wearable devices, and other novel applications.

Energy harvesting

Static water energy harvesting

Energy harvesting from static natural water primarily involves extracting energy from non-flowing or phase-change water systems^{14,29,85,238,334,335,336}. In these systems, the water remains relatively still, and energy conversion mainly stems from heat exchange between the water and its surroundings, as well as the continuous ion gradient generated by moisture within the hygroscopic material^305,337. These energy transducers do not rely on water movement but rather utilize inherent thermodynamic and electrochemical processes to generate electricity. By combining solar water heating systems with thermoelectric generators (TEGs), the latent heat of water can be effectively converted into electrical energy. Kadohiro et al. designed a all-day energy harvesting system based on hydrothermal energy storage³³⁵ (Fig. 15a). This system uses the temperature difference of hot water stored during the day to drive the TEG, and can continue to generate electricity at night, thus achieving continuous energy output of day and night. The results show that when the hot water temperature is maintained at 60–80 °C and the ambient temperature is about 25 °C, the TEG module can achieve a thermoelectric conversion efficiency of about 2.5% and output a stable DC signal. Amagai et al. combined a thermoelectric generator (TEG) with a natural spring water system, using the stable temperature difference between the spring water and the ambient air to achieve continuous collection and power generation of static water heat energy³³⁶ (Fig. 15b). In this study, a flexible thermoelectric generator module (f-TEG) was used to tightly wrap copper heat exchange tubes, enabling efficient heat conduction without the need for additional heat exchange devices. In actual outdoor testing, the f-TEG can still achieve a power output in the range of 10 mW under conditions where the temperature difference is less than 20 K, with a peak power of up to 19.6 mW and an average power of about 14.7 mW, which is sufficient to drive a Bluetooth wireless data logger to achieve water temperature monitoring and real-time data transmission.

Fig. 15: Static water energy harvesting.

a Schematic illustration of an all-day thermoelectric energy harvesting system utilizing solar irradiation and cold water during daytime, and hot and cold water at nighttime.Reproduced with permission from ref. ³³⁵. Copyright 2020, MDPI. b Prototype design schematic featuring a heat exchanger on the spring-water side, a flexible thermoelectric generator (f-TEG) and a heat exchanger on the air side (left), along with a photograph of the assembled system (right). Reproduced with permission from ref. ³³⁶. Copyright 2024, Elsevier. c Schematic and fabrication process of the asymmetric wood-based ionic hydrogels(left), optical image of 3D-SMEG assembled on acrylic sheet and PET film matrix(center), and 3D-SMEG systems applications (right). Reproduced with permission from ref. ²³⁸. Copyright 2025, Royal Society of Chemistry. d Schematic of MEG featuring FAA gel sandwiched by asymmetric porous and blocking electrodes with SEM image of the freeze-dried FAA hydrogel for the measurement(left), and schematic of vertical stacking strategy(center). Demonstration of the MEG-activated smart window in ambient air and screen QR code by the activated window(right). Reproduced with permission from ref. ⁸⁵. Copyright 2025, Wiley. e Power density curves of MTEG with a series of loads at different humidity and temperature differences (left), and integration of the scalable electrical generator(right). Reproduced with permission from ref. ²⁹. Copyright 2023, Wiley. f Schematic of solar-driven water-electricity cogeneration device, and voltage and current output of SEG for dripping water drops during one sun irradiation (left), schematic of SEG connection in series and parallel, and the solar-driven device for power and clean water generation(right). Reproduced with permission from ref. ³⁴⁰. Copyright 2025, Elsevier.

In addition to liquid water, moisture in the air can also be used to generate electricity. Chen et al.²³⁸ demonstrated that a 10 × 10 array of moisture electricity generator (MEG) units could achieve an output of 13.8 V and 1 mA, which was sufficient to power a streetlight for continuous nighttime illumination, presenting an alternative to conventional solar-powered lights (Fig. 15c). They further developed a self-powered system comprising a 3 × 100 array of MEG units, which delivered an output of 4.12 V and 10 mA, enabling direct charging of a mobile phone using ambient moisture. This is particularly suitable for outdoor activities or emergency power supply. Moreover, the stacking of multiple power panels could potentially yield a power output exceeding 34 MW, indicating its promising potential for supplying electricity to future residential buildings. For energy-harvesting systems, ensuring long-term operational stability and reliability is a crucial prerequisite for practical implementation. Kim et al. designed a MEG based on a salt concentration-gradient cationic hydrogel, which can operate stably for over 50 days even under low-humidity conditions (30% RH) while maintaining excellent electrical output performance⁸⁵ (Fig. 15d). The core mechanism originates from the synergistic effect between the salt concentration gradient and water molecule migration, which induces a sustained directional ion flux within the hydrogel, thereby generating a stable direct current (DC) in the external circuit. The MEG achieved a power density of 13.8 mW m⁻² with an internal resistance of only ~4 kΩ, significantly outperforming previous humidity-driven energy devices. Furthermore, modular stacking through series and parallel integration enabled scalable voltage and current outputs, capable of continuously powering various low-power electronics such as smart watches, electronic clocks, and liquid crystal displays, demonstrating great potential for long-term energy supply and self-powered systems in arid environments.

To further enhance energy-harvesting efficiency, researchers proposed an integrated moist–thermoelectric generator (MTEG) that couples thermoelectric and moisture-electric conversion for efficient recovery of steam energy^{29,338,339,340}. The core design principle of such systems lies in the spatial and functional synchronization of thermal and moisture gradients to minimize internal impedance and maximize carrier flux. Yang et al. addressed the challenge of underutilized waste-steam energy in industrial environments by developing a flexible MTEG that employs a polymer electrolyte copolymer membrane (PAMPS/PSSS) as the core functional layer²⁹ (Fig. 15e). The hybrid performance enhancement mechanism is based on a synergistic enhancement loop between the temperature gradient (∆T) and the moisture gradient (∆RH). On the one hand, high-humidity conditions induce extensive water adsorption, promoting the dissociation of functional groups (-SO3H) and increasing the concentration of mobile ions. This provides a significantly higher density of charge carriers for the Soret-effect-driven thermodiffusion, thereby amplifying the thermoelectric output. On the other hand, the established ∆T accelerates localized water evaporation and transport dynamics, which, in turn, sustain and widen the ∆RH. This cyclic synergy ensures a continuous supply of carriers and a sustained driving force. Consequently, this “1 + 1 > 2” synergistic energy-conversion effect resulted in an open-circuit voltage of 1.81 V and a power density of 4.75 μW cm⁻² under a ∆T of 15 K and a ∆RH of 60%, far exceeding that of conventional single-function TEG or MEG systems. Moreover, Li et al. further extended the application of static water energy harvesting to solar-driven systems by proposing a solar evaporation-enhanced generator (SEG) based on the solar-driven interfacial evaporation (SDIE) effect³⁴⁰ (Fig. 15f). In this system, a photothermal conversion layer and an electrical generation layer were integrated into a porous sponge substrate, enabling simultaneous coupling of solar energy absorption, water evaporation, and electricity generation. Under standard solar irradiation (1 sun), the SEG achieved a water evaporation rate of 1.75 kg m⁻² h⁻¹ and a power density of approximately 20.6 μW cm⁻². This design effectively realizes the synergistic utilization of solar, thermal, and moisture energy, providing a sustainable pathway for converting low-grade water energy into electricity.

Kinetic water energy harvesting

Kinetic water sources in nature generally include raindrops, flowing rivers, and ocean waves. By converting their inherent kinetic energy into electrical energy, it is possible to efficiently harvest the widely distributed and renewable mechanical energy present in the environment^146,177,341. Xu et al. developed a droplet electricity generator (DEG) based on a transistor structure, which can efficiently convert the impact energy and interfacial charge transfer process of falling droplets into electrical energy³¹⁴ (Fig. 16a). When a 53 μL droplet lands on the device surface, it can generate an open-circuit voltage of up to 260 V and a short-circuit current of 270 μA, with an output power density exceeding 50.1 W m⁻². Through a self-made integrated water collection and power generation system, stable power generation under rainfall conditions was achieved, along with the efficient collection and storage of rainwater. Furthermore, this research demonstrated the system’s application potential in forest environments and outdoor rainfall conditions, validating its real-time collection and conversion capability of raindrop energy in natural environments. In addition, due to the complexity and diversity of outdoor scenarios, the researchers further explored the applicability of the droplet energy harvesting system to different environments and material surfaces. By placing the electrodes on top of the dielectric layer rather than beneath it, Zhang et al. designed a universal single-electrode DEG (SE-DEG)³⁴² (Fig. 16b). The SE-DEG successfully harvested water droplet kinetic energy directly from various artificial and natural surfaces (such as PTFE, glass, wood, leaves, cicada wings, and stone) without relying on a specific substrate. This innovative structure avoids the interface shielding effect in traditional TENGs, allowing charge to be efficiently conducted through the bulk effect, thus achieving stable output in various complex environments. Experimental results show that the SE-DEG assembled on the PTFE surface can generate an open-circuit voltage of 62.2 V and a short-circuit current of 86.5 μA, representing a performance improvement of over 1000 times compared to traditional structures. Furthermore, researchers explored more applications based on droplet energy harvesting, such as integrating transducers into everyday items like umbrellas^343,344 or installing them on building rooftops³⁴⁵. These designs not only enhance the device’s environmental adaptability but also provide new directions for the development of IoT devices.

Fig. 16: Kinetic water energy harvesting.

a Photograph of the water collector and dispenser for fixed position droplet generation (left), schematic of the coplanar-electrode droplet energy harvesting (DEH) panel(center), and the self-powdered wireless forest monitoring system(right). Reproduced with permission from ref. ³¹⁴. Copyright 2022, Royal Society of Chemistry. b The measured transferred charges (Q_T) and the charges carried by the water droplet (Q_D) for eight different material surfaces(top), and the procedures for construction of SE-DEG on the stone surface and optical image of SE-DEG harvesting water droplet outdoors(bottom). Reproduced with permission from ref. ³⁴². Copyright 2021, Elsevier. c Schematic of the water-through triboelectric nanogenerator (WT-TENG)(top), with the maximum voltage and current obtained from water, tap water or NaCl solution and the WT-TENG (bottom). Reproduced with permission from ref. ³⁴⁸. Copyright 2018, Springer Nature. d Schematic of rotating contact triboelectric nanogenerator (RC-TENG) and optical image of thermohygrometer driven by RC-TENG at 0.7 m/s flow rate(left). Circuit diagram and RC-TENG charging 100 μF capacitor(Right). Reproduced with permission from ref. ³⁴⁹. Copyright 2024, Elsevier. e Schematic diagram of the isotropic triboelectric–electromagnetic hybrid nanogenerator(iTEHG)(left), with photos showing an array of 320 green power LEDs lighted instantly and continuously by a single iTEHG in Sanya Bay, China. Reproduced with permission from ref. ³¹¹. Copyright 2023, Elsevier. f Schematic of the floating droplet activated electricity generator (FDEG)(left). Photo of the 4 FDEGs driven by the water wave and powering an electronic clock(right). Reproduced with permission from ref. ⁵². Copyright 2025, Cell Press.

For flowing water, TENG is an important approach for harvesting kinetic water energy^346,347,348. Park et al. designed a water-through TENG (WT-TENG) based on a titanium mesh electrode for efficiently harvesting mechanical energy from flowing water in pipelines³⁴⁸ (Fig. 16c). The device operates in a direct-contact mode, where a titanium mesh coated with PTFE nanowires serves as the active electrode. When water flows through the mesh, the contact electrification effect between the water and PTFE nanowires induces charge separation, generating an alternating current output in the external circuit. Experimental results showed that at a deionized water flow rate of 5 mL s⁻¹, the WT-TENG produced an open-circuit voltage of approximately 9.4 V and a short-circuit current of 5.1 μA, with a peak power of 39.7 μW. Notably, the device could effectively harvest energy not only from deionized water but also from tap water and even 0.6 M NaCl solution (simulating seawater), demonstrating strong environmental adaptability and practical potential. The researchers further installed the WT-TENG in flowing water pipes to directly harvest energy from domestic wastewater and river flow. In addition, Liu et al. developed a rotational contact-mode TENG (RC-TENG) based on an indirect-contact mode for harvesting river-flow energy³⁴⁹ (Fig. 16d). Inspired by the structure of a water wheel, multiple contact–separation TENG units were mounted inside a rotating cylinder. Water impact on the external blades generated torque to drive the rotation of the central shaft, enabling periodic contact and separation between triboelectric layers and thus converting hydrodynamic energy into electricity. Experimental results showed that a single TENG unit achieved a peak power density of 157 mW m⁻² at 1 Hz, while the integrated system, operating at a flow velocity of 0.7 m s⁻¹, could stably power a temperature–humidity sensor and an LED display module.

In the field of ocean wave energy harvesting, Xu et al. designed a guided-liquid-based isotropic triboelectric–electromagnetic hybrid nanogenerator (iTEHG) for efficiently capturing wave energy from arbitrary directions³¹¹ (Fig. 16e). The device adopts a tilted satellite-dish-shaped substrate with concentric circular electrode pairs, allowing the liquid to flow along guided paths under gravity. This configuration effectively addresses the problems of severe wear and insufficient contact in traditional solid–solid triboelectric interfaces, significantly extending device lifetime. Field tests conducted in natural marine environments demonstrated that a single iTEHG, even without an external energy storage unit, could continuously light up 320 high-luminosity LEDs and charge a 0.1 F supercapacitor to 3.1 V, powering a wireless temperature–humidity sensor for over 26 min. Moreover, Zhou et al. proposed a floating droplet electricity generator (FDEG), which incorporates asymmetric capacitance between circular and ring-shaped electrodes to greatly enhance the electrical output per unit volume⁵² (Fig. 16f). The device operates by the rolling motion of droplets on the FEP surface, which induces rapid charge transfer between asymmetric capacitors through intermittent circuit closure and disconnection, generating high-amplitude voltage and current outputs. Under a wave frequency of 0.28 Hz, a single FDEG produced a current of 107.6 mA, a voltage of 310.8 V, and a power density of 1.19 kW m⁻³, far exceeding conventional droplet-based generators. The device was tested in a self-made “WAVE” bath that mimicked realistic ocean wave conditions, and it maintained stable energy generation under irregular wave motion. When four FDEG units were connected in parallel, the system successfully powered a humidity sensor, demonstrating its great potential for self-powered marine systems and blue energy networks.

Self-powered sensors

Self-powered physical sensors

Sensors, as a core component of Internet of Things (IoT) systems, are key devices for collecting physical parameters such as ambient temperature^200,350,351, humidity^352,353, and motion^{354,355,356,357,358}. In the field of self-powered physical sensors, small-scale water energy harvesting technology not only serves as an energy supplier^242,359 but also achieves a “signal-for-sensing” functional mode through the inherent coupling relationship between its output signal and the information to be measured in the environment^291,360. For example, droplet impacts, fluctuations in air humidity, and changes in water flow velocity can all alter the output signal of transducers, enabling them to directly serve as sensitive units for sensing changes in the external environment^360,361,362. Another strategy is to temporarily store the harvested energy in capacitors or micro-energy storage devices to provide a stable power supply for commercial sensor modules, achieving real-time monitoring without the need for an external power source^363,364.

In agricultural irrigation scenarios, self-powered physical sensors offer significant advantages: the impact of irrigation drips and air humidity can serve as natural energy sources for real-time monitoring of key parameters such as soil moisture and temperature changes, thereby improving the response speed and management accuracy of crop growth processes^347,365. LAN et al. utilized a PVDF-HFP/fluorinated carbon nanotube composite nanofiber membrane as a friction layer and carbon nanotubes as electrodes to form a lightweight, flexible, and strongly adhesive waterproof and breathable TENG (WB-TENG)³⁶⁶ (Fig. 17a). By attaching it to plant surfaces, a self-powered agricultural system was constructed. This WB-TENG does not affect the plant leaf surface and thus does not affect its transpiration and growth. Simultaneously, by harnessing wind-induced vibrations and raindrop impacts experienced by the plant, the WB-TENG can continuously harvest ambient mechanical energy and provide a stable power supply for wireless sensors around the plant, enabling real-time monitoring of key agricultural parameters such as soil moisture, light intensity, and temperature. The biocompatibility and the reliability of the WB-TENG have also been investigated. After 55 days under moderate sunlight and in an indoor environment, new leaves and roots have grown, indicating that the device does not affect the plant’s normal growth. The output voltage of the WB-TENG shows only a slight difference (~500 V to ~450 V) after long-term testing. Furthermore, Zhang et al. constructed a self-powered humidity sensor (SPHS) using MEG to enable real-time monitoring of humidity in agricultural environments³⁵² (Fig. 17b). This device employs a PSS/PVA/LiCl electrolyte membrane, which has strong hygroscopic capacity, as its core functional layer. It drives water-molecule adsorption and directional ion diffusion along the ambient humidity gradient, coupled with a metal-air redox reaction, thereby generating a stable DC signal without an external power source. Experimental results show that the sensor’s response range covers 33–91% RH, with the output current increasing from 2.6 to 76.1 μA, exhibiting excellent linearity (R² = 0.988) and high sensitivity to changes in ambient humidity. The researchers further integrated this device with a wireless transmission module and auxiliary sensors, such as temperature/light intensity, to construct a long-term, operationally robust agricultural microenvironment monitoring system, demonstrating its broad application potential in precision agriculture, smart farmland management, and maintenance-free IoT sensor networks. The device undergoes a long-term reliability test in a high-humidity environment (91%) to demonstrate its capability in an agricultural environment, achieving an average short-circuit current (I_sc) of more than 75 μA for 10000 s.

Fig. 17: Self-powered physical sensors.

a Optical image of the WB-TENG attached to different plants(left), and voltage output of WB-TENG during leaf-to-leaf contact separation induced by artificial wind and water droplet contacts on rainy days(center), and the optical image of the plant sensor in the soil and connected to the mobile phone(right). Reproduced with permission from ref. ³⁶⁶. Copyright 2021, American Chemical Society. b Schematic of a flexible self-powered humidity sensor SPHS(left), and its current output in response to soil moisture before and after watering(right). Reproduced with permission from ref. ³⁵². Copyright 2024, Elsevier. c Schematic of liquid–solid tubular triboelectric nanogenerator (LST-TENG) designed for ship draft measurement(left), along with its structural configuration(right). Reproduced with permission from ref. ³⁶⁹. Copyright 2019, Wiley. d Schematic of an electrode-grounded droplet-based electricity generator (EG-DEG) sensor embedded in the pipe(left), and the fluid flow rate monitoring system(right). Reproduced with permission from ref. ³⁷⁰. Copyright 2023, Wiley. e 3D graph illustrating the output frequency of the liquid-triboelectric microfluidic sensor (TMS) across various capillaries under different external flow rates(left) Liquid flow sensor for infusion monitor and gas flow sensor for gas detection(right). Reproduced with permission from ref. ³⁷¹. Copyright 2016, American Chemical Society.

By utilizing the instantaneous response characteristics of TENG to external mechanical stimuli, its output electrical signal can not only be used for energy harvesting, but also as a sensitive indicator of changes in distance²⁸³, speed^291,367, or position³⁶⁸. Zhang et al. proposed a self-powered water level sensor based on a tubular liquid-solid TENG (LST-TENG)³⁶⁹ (Fig. 17c). This device uniformly distributes multiple sets of copper electrodes on the outer wall of a PTFE tube. The rise and fall of water within the tube induces charge transfer at the liquid-solid interface, enabling water level detection without an external power source. Research shows that when the water level crosses different electrode regions, the rate of change of the open-circuit voltage (dV_OC/dt) exhibits corresponding positive peaks or negative troughs. These characteristic signals correspond one-to-one with the electrode distribution and can serve as accurate indicators of water level. Based on this principle, they applied the LST-TENG to ship draft monitoring, achieving a water level identification accuracy of 10 mm resolution, approximately ten times higher than traditional measuring tapes. Furthermore, the system maintains stable output under different salinity levels (0–35 mg mL⁻¹) and different water level change frequencies, demonstrating excellent environmental adaptability and practical application potential. LST-TENG also shows stable long-term results, with the output voltage peak remain stable (~12 V) before and after 2000 cycles at a frequency of 0.9 Hz. Moreover, Yang et al. developed a grounded electrode-based droplet generator (EG-DEG)³⁷⁰ (Fig. 17d). Unlike traditional single-electrode DEG, the EG-DEG employs a three-electrode structure: two strip-shaped upper electrodes deposited on a PTFE surface are grounded, while the bottom GSEC electrode is connected to a current amplifier. As a droplet slides along the device surface, it sequentially crosses the two upper electrodes, generating two clearly distinguishable current pulses. By measuring the time interval between these two pulses and combining it with the known electrode spacing, the EG-DEG can achieve millisecond-level measurement of the average droplet sliding velocity, thus enabling high-precision monitoring of droplet dynamics. To further verify its application potential, the authors integrated the EG-DEG into a liquid pipeline to achieve real-time monitoring of fluid flow velocity. As continuous liquid flows through the device, the resulting periodic disturbances generate a stable AC signal, and the frequency of the signal pulses increases with the flow velocity. Therefore, by analyzing the frequency of the output signal, the EG-DEG can quantitatively monitor the fluid flow velocity within the pipeline without any external power supply. The stable current output (~ 6.5 μA) of the EG-DEG during continuous water droplet impinging for 10,000 cycles shows the repeatability and reliability of the device in water droplet energy harvesting. Furthermore, Chen et al. proposed a self-powered Triboelectric Microfluidic Sensor (TMS), further expanding the application scenarios of sensing based on small-scale water energy³⁷¹ (Fig. 17e). This device utilizes the contact electrification and electrostatic induction signals generated when droplets or bubbles pass through the PTFE triboelectric interface in a microchannel to achieve real-time fluid monitoring without an external power source. Experiments show that the TMS can cover a wide dynamic range for liquids (~3–95 μL/s) and gases (~7–280 μL/s), and the output signal exhibits a good linear relationship with the flow rate. More importantly, this sensor demonstrates high application potential in real-world scenarios, such as for monitoring intravenous infusion drip rates and detecting airflow in industrial pipelines, realizing a power-free, low-cost, and easily integrated microfluidic monitoring solution.

Self-powered chemical sensors

As mentioned earlier, the chemical properties of water—such as ionic strength (e.g., salinity) and the types and concentrations of dissolved gases (e.g., O₂, CO₂) or dissolved organic/inorganic substances—significantly affect the charge transfer behavior of transducers, thereby altering their energy harvesting efficiency^372,373. Researchers have utilized this highly sensitive electrical response to liquid composition to develop various self-powered chemical sensors based on small-scale water-powered transducers, which can be used to detect different chemical information such as pesticide residues, heavy metal ions, solution pH, ionic contamination, and liquid component differentiation^374,375,376. By using a direct current TENG (DC-TENG), Wang et al. systematically studied the differences in triboelectricity at the solid–liquid interface of different liquids (e.g., ethanol, acetone, hexane, isopropanol, etc.) based on the DC-TENG structure of the liquid-dielectric interface³⁷⁷ (Fig. 18a). They found that liquid polarity, contact angle, and intermolecular forces significantly affect the amount of charge transfer. Based on this mechanism, they further developed a self-driven detection platform for chemical composition analysis. By comparing the systematic response of the output voltage to changes in liquid composition, they achieved quantitative detection of organic solvent ratios and water content. Zhang et al. proposed a droplet-based TENG based on spatially distributed electrodes, which can distinguish the charge transfer process information as the droplet slides across different positions, providing experimental evidence for exploring the electron transfer mechanism at the liquid-solid interface³⁷⁸ (Fig. 18b). The study shows that the charge accumulation process when a droplet slides on a solid surface is mainly dominated by electron transfer, rather than free ion migration; the higher the ion content in the droplet, the stronger the shielding effect, and the weaker the resulting triboelectric response. This work not only clarifies the nature of solid–liquid triboelectricity but also demonstrates the ability to identify solvent types, solvent ratios, and trace organometallic compounds (such as ferrocene diluted to 0.05 mg/mL) using droplet-TENG, proving its potential as a highly sensitive self-energized chemical probe. Furthermore, Zhang et al. constructed a novel triboelectric spectroscopy (TES) platform utilizing the triboelectric effect at the liquid-solid interface, enabling in-situ characterization of liquid chemical properties³⁷⁹ (Fig. 18c). In this system, as a droplet slides along the FEP surface covering an electrode array, characteristic interfacial charge transfer signals are generated at different locations. The research team combined these signals into a chemically distinguishable triboelectric spectrum by recording the instantaneous charge changes at each electrode position along the droplet’s spatial trajectory. This spatially distributed charge response reflects the electron transfer behavior of ions and molecules at the interface, changes in the electric bilayer structure, and adsorption/desorption processes in the liquid. Therefore, different chemical components exhibit identifiable peak positions, peak intensities, and waveform differences in the spectrum. Experimental results show that TES can distinguish various liquid chemicals, including acids, bases, metal salts, and small organic molecules, within sub-second timeframes, with qualitative and quantitative accuracy reaching approximately 93% and detection limits down to the ppb level. Since the system requires no light source or expensive instruments, and no complex sample preparation, TES shows significant potential as a portable, low-cost, self-powered chemical analysis tool.

Fig. 18: Self-powered chemical and biomedical sensors.

a Schematic of the ring-tube TENG mounted on a rotating acrylic base with support spokes, featuring two electric brushes on an outer stationary acrylic base and the charge density difference of water, ethanol, hexane, and FEP interfaces at equilibrium(left). Relationship between V_oc, contact angle, and polarity for the different liquids (Hexane (1); isopropanol (2); ethanol (3); acetone (4); ethylene glycol (5); and DI water (6))(center). Optical images of the TENG device for chemical composition analysis and moisture-content detection(right). Reproduced with permission from ref. ³⁷⁷. Copyright 2019, American Chemical Society. b Schematic of droplet-TENG, composed of PMMA as the base and top layer is the dielectric polymer film for droplet contact with copper electrode in between(top left). Working mechanism of the droplet-TENG: The charge transfer between liquid and solid surface when the droplet flows on the polymer surface(bottom left). The transferred charges for different droplets (ethanol, acetone, hexane, THF and benzene) sliding on the FEP surface(top right),and the transferred charges of ethanol/ferrocene droplet sliding on the surface as a function of the concentration of ferrocene(bottom right). Reproduced with permission from ref. ³⁷⁸. Copyright 2021, American Chemical Society. c Schematic of the design with the droplet moving on the dielectric surface separating the sample from an array of copper band electrodes(top).Schematic of the working mechanism of TES for chemical analysis(bottom). Reproduced with permission from ref. ³⁷⁹. Copyright 2024, American Chemical Society. d Schematic of the superhydrophobic TENG and procedure for preparing the tubular drop sensor(top). Schematic of the real-time infusion monitoring system(bottom). Reproduced with permission from ref. ¹⁸⁵. Copyright 2020, American Chemical Society. e Schematic of the smart toilet system utilizing a self-powered triboelectric nanosensor array(TENSA) that harvests energy from urine droplets. Reproduced with permission from ref. ³⁸⁷. Copyright 2024, Wiley.

Self-powered biomedical sensors

In recent years, self-powered transducers based on small-scale water energy conversion have also shown significant potential in the field of biosensing^380,381. Human body fluids (such as sweat, blood, and urine) contain abundant ions, metabolites, and biomarkers. Their physicochemical properties, such as conductivity, ionic composition, viscosity, and pH, significantly influence charge transfer behavior at the liquid-solid interface^382,383,384. Researchers have utilized this to achieve passive monitoring of various biological parameters by analyzing changes in electrical signals generated by body fluids in triboelectric or liquid-solid interface generators^185,385,386. Hu et al., based on the superhydrophobic liquid-solid contact TENG, systematically evaluated the differences in electrical signals at the contact interface of different body fluids (including blood, urine, acid and alkali solutions, etc.) and successfully constructed a self-powered drainage bottle droplet sensor and intelligent infusion monitoring system¹⁸⁵ (Fig. 18d). This system can monitor the dripping behavior and infusion status of body fluids in real time, demonstrating the practical application value of droplet characteristic differences in biological monitoring. Compared to blood, urine is more non-invasive and convenient to obtain, and the metabolic information it contains (such as ions, carbohydrates, proteins, urea, etc.) can reflect the body’s health status. Mondal et al. proposed a self-sustaining triboelectric nanosensor for real-time urine analysis in smart toilets, specifically designed for real-time urine analysis³⁸⁷ (Fig. 18e). This sensor integrates a superhydrophobic nanostructure with a flexible triboelectric layer, enabling urine to generate a stable triboelectric output signal during the contact-separation process. The amplitude and charge characteristics of this signal exhibit discernible shifts depending on the ionic strength, pH, specific gravity, viscosity, and pathological components of the urine. In experimental data, normal urine samples typically showed low charge output, while pathological samples containing glucose, protein, or high ion concentrations exhibited significantly enhanced voltage and current signals due to differences in conductivity and interfacial contact characteristics. Through systematic comparative experiments, the researchers successfully identified passive electrical signals associated with various physiological abnormalities, including diabetes, proteinuria, and electrolyte imbalances. The device was then exposed to a toxic experimental environment for more than 270 days and consistently performed, demonstrating its long-term reliability even in extreme conditions.

Self-powered wearable devices

Due to the broad availability of small-scale water sources and their inherently low energy scale, water-based energy harvesters are highly suitable for integration into flexible and wearable electronic systems^200,388. These devices can serve not only as self-powered energy modules for wearable applications but also as self-driven sensing units. Maity et al. developed a heterogeneous moisture-driven electrical generator (MODEG) based on a (GO)PANI/F-Nafion (PDDA) bilayer film, which can continuously harvest energy from ambient moisture and human breath²⁴⁴ (Fig. 19a). They further integrated the MODEG into a commercial face mask, enabling the device to continuously capture the humidity from exhaled air and convert it into electrical power. The mask-integrated MODEG produced 450–600 mV during normal breathing and up to 650 mV during exercise, demonstrating its capability to directly power portable medical devices and wearable electronics, while simultaneously functioning as a self-powered platform for real-time respiratory monitoring. However, the output power is still limited by the humidity gradient and the membrane structure itself, making the MODEG difficult to support higher energy consumption wireless sensing functions. Liu et al. further increased the power generation based on humidity gradients and developed a high-output flexible MEG membrane²⁴² (Fig. 19b). By introducing a metal-air reaction into the bilayer polyelectrolyte structure, they significantly improved the carrier diffusion rate, increasing the device’s output power density to 286.5 μW cm⁻². Building upon this, the authors constructed a fully self-powered wireless wearable monitoring system, integrating energy harvesting, power management, wireless communication, and biochemical/vital sign detection modules. This system not only enables self-driven detection of urine pH and uric acid but can also be used in complex health management scenarios such as posture monitoring and fall detection.

Fig. 19: Self-powered wearable devices and other applications.

a Schematic of a moisture-driven electric generator (MODEG) face mask designed for power supply and real-time breathing monitor(left). Voltage output of a 4 × 4 cm² MODEG integrated with a commercial FDA-licensed FFP2 face mask during individual breaths(center), and single output voltage at normal and exercise breathing conditions(right). Reproduced with permission from ref. ²⁴⁴. Copyright 2023, Wiley. b Schematic of the self-powered system based on MEG with the ion diffusion of the bionic MEG membrane(left), MEG membrane-driven biomedical applications(center), and the self-powered wearable system on the diaper(right). Reproduced with permission from ref. ²⁴². Copyright 2024, Elsevier. c Schematic of an F-TENG, with the components and static contact angle of the hydrophobic surface(top), and the schematic of the F-TENG for harvesting droplets on a rainy day(bottom). Reproduced with permission from ref. ¹⁷⁰. Copyright 2021, American Chemical Society. d Schematic of an underwater wireless multi-site human motion monitoring system based on bionic stretchable nanogenerator (BSNG)(top), and the volunteer swam in different strokes with BSNG (bottom). Reproduced with permission from ref. ³⁸⁹. Copyright 2019, Springer Nature. e Schematic of solid–liquid TENG using PTFE filtration membrane for energy harvesting and image of the PTFE membrane showing the wettability (top). Schematic of the solid–liquid TENGs array for cathodic protection of A3 carbon steel in 3.5 wt% NaCl solution and open circuit potential with and without PTFE-based TENGs(center) and the images of A3 carbon steel immersed in the solution for 1–3 h with and without PTFE-based TENGs(bottom). Reproduced with permission from ref. ³⁹². Copyright 2021, Elsevier. f Schematic of hydrogel-based MEG devices with E-PTFE film for underwater moisture harvesting (left) and images of self-powered underwater wearable MEGs(right). Reproduced with permission from ref. ²²⁴. Copyright 2024, Wiley. g Schematic and optical image of solar-water hybrid electricity generator with common-electrode (HEG-CEA)(left). Schematic of HEG-CEA at large-scale future application(center), and optical images on lighting up a 20 W bulb and an integrated sensor by HEG-CEA on droplet energy harvesting(right). Reproduced with permission from ref. ³⁹⁷. Copyright 2023, Elsevier.

While the flexibility and stretchability of thin films can be well utilized to fabricate wearable electronic devices, their breathability and durability remain significant challenges affecting their long-term wearability. Ye et al. reported a hydrophobic, self-repairing all-fabric TENG specifically engineered for raindrop energy harvesting¹⁷⁰ (Fig. 19c). Through a hierarchical SiO₂/PVDF-HFP/FDTS coating, the textile exhibits superhydrophobicity (157° contact angle), excellent air permeability, and thermally triggered self-repair capability. Notably, the fabric-based TENG delivers a seven-fold enhancement in output voltage relative to conventional textile TENGs and maintains stable operation under repeated droplet impact, making it highly suitable for smart raincoats, umbrellas, and outdoor wearable systems. In addition to addressing underwater health monitoring and rescue systems, Zou et al. developed a bionic stretchable nanogenerator (BSNG) inspired by the ion channel structure of electric eels³⁸⁹ (Fig. 19d). This device combines liquid-solid triboelectricity and electrostatic induction technologies, employing a PDMS-silicon mechanically gated channel structure that enables controlled liquid flow under mechanical deformation. Due to its excellent flexibility, ductility, and robust encapsulation, the BSNG operates stably in underwater environments and provides an output voltage exceeding 10 V during human movement. By integrating multiple BSNG units with a wireless transmission module, the researchers demonstrated multi-point underwater motion monitoring, capable of capturing joint movements during swimming in real time. The system also enables practical underwater safety functions: when used in conjunction with an energy storage module, the BSNG embedded in a diving suit can power a wireless distress signal transmitter and remotely activate warning lights. As the self-powered wearable device will have the chance to contact with the skin, the safety performance of the materials is one of the important fields need to be study. Some common TENG materials had been studies on the biocompatibility, such as (GO)PANI, Yan et al. with the using of the mouse fibroblast cell line L929³⁹⁰, a commonly used cell to test the cytotoxicity. MTT colorimetric assay, a method to test the cell viability and proliferation is done to test the cytocompatibility of the samples, which shown the survival rate on the (GO)PANI is 5% higher than GO after 96 h, reaching 95%, showing a high biocompatibility. Another commonly used material, polydimethylsiloxane(PDMS) and polyvinylidene fluoride(PVDF) are also studied by Li et al.³⁹¹ With the PDMS and PVDF nanogenerators implanted inside female mice for six months. Histological analysis of the skin, muscles and tissue, blood and serum test had been done to test the biocompatibility and with no signs of toxicity or incompatibility were observed.

Other applications

In addition to the typical applications discussed above, such as energy harvesting, self-powered sensing, and self-powered wearable device, high-performance small-scale water energy transducers have also demonstrated significant value in some other scenarios^{392,393,394,395,396}, expanding their functional boundaries in real-world environments.

Corrosion protection

Sun et al. addressed the challenge of metal corrosion in humid and marine environments by integrating liquid–solid TENGs with cathodic protection (CP) technology, proposing a self-powered cathodic protection system based on an array of liquid–solid TENG units³⁹² (Fig. 19e). In their design, PTFE ultrafiltration membranes were used as the liquid–solid triboelectric layer, and multiple water-sealed TENG cells were encapsulated and connected in parallel to form an energy-harvesting array capable of continuously extracting mechanical energy from wave-induced vibrations. The array delivered a short-circuit current of up to 2.68 mA and an output voltage of 105 V in a 3.5 wt% NaCl solution, and successfully provided an impressed cathodic bias for A3 carbon steel. As a result, the open-circuit potential of the steel shifted negatively by approximately 330 mV, markedly suppressing its corrosion tendency compared with unprotected samples.

Remote wireless communication

Remote wireless communication represents an important emerging application of small-scale water energy harvesting, particularly in scenarios where wired power supply or frequent battery replacement is impractical. Shen et al. utilized a MEG based on ionized hydrogels to achieve stable operation in an underwater environment and continuously power a wireless transmitter, thus enabling real-time underwater signal transmission²²⁴ (Fig. 19f). When integrated into an underwater device, the device can continuously output a voltage of approximately 1.0 V and successfully power the wireless communication module, thereby achieving real-time underwater signal transmission without any auxiliary power source. Their research results demonstrate that small hydroelectric harvesters are not merely passive energy converters; rather, they can act as active communication nodes, driving wireless electronic devices, transmitting environmental information, and supporting autonomous operation in complex environments, especially where batteries are unreliable or difficult to obtain.

Solar-water energy harvesting hybrid system

To further expand the application scenarios of small-scale water-energy harvesters, researchers have begun to explore hybrid energy harvesting systems that couple water-based energy conversion with other energy sources, enabling stable and efficient power generation under diverse environmental conditions, with solar panels merging with the droplet electricity generators (DEGs), receiving solar energy during sunlight and water energy during rainfall. Compared with other hybrid solar water energy harvester simply stacking the solar panel and the DEG with separate electrodes, which will lead to transmittance, conversion efficiency, low instantaneous peak power output and manufacturing problems, Liao et al. proposed a hybrid solar–water energy generator with a common-electrode architecture (HEG-CEA)³⁹⁷ (Fig. 19g). In this design, the solar cell itself serves simultaneously as the electrode for the DEG, with both solar panel and DEG sharing the same electrode, taking advantage of its high electrical conductivity and avoiding the optical losses typically associated with stacked configurations. By directly covering the solar cell with an FEP film and introducing a micro top-electrode, the HEG-CEA achieves synergistic conversion of both solar and droplet energy without sacrificing light transmittance. Experimental results show that, compared with conventional systems without a shared electrode, the HEG-CEA with the absence of the barrier of additional conductive layer, improves solar energy conversion efficiency by ~10% and enhances droplet energy conversion efficiency by more than two orders of magnitude. Other optimization of the device includes the control of FEP thickness, which the open circuit voltage and short circuit are quasi-proportional to the thickness, reaching the optimum at thickness 70 µm. The salt concentration of the water droplet, the droplet volume and release height are other parameter that affect the output performance of the device. In experimental results, even enables a single droplet to directly power a 20 W light bulb at the end. Furthermore, in real outdoor environments, the system can simultaneously harvest sunlight and high-velocity droplet impact energy to charge capacitors and power wireless temperature–humidity sensors, demonstrating reliable all-weather energy supply capability.

Other than the shared electrode design of the solar water energy harvesting hybrid system. Guo et al. proposed a design that integrates the water droplet-TENG and the photovoltaics (PV) using a copper electrode, ethylene tetrafluoroethylene (ETFE) layer and bottom silicon cell³⁹⁸. The design is to improve the light absorption by using surface modification techniques, the ETFE layer, which is an anti-reflection layer to increase the light transmission to the PV cell and a hydrophobic material for the TENG. With the textured ETFE layer with pyramid pattern on the silicone PV cell, which increases the optical distance and light absorption.

Outlook

Small‑scale water energy harvesting encompasses a broad class of transduction strategies that convert local environmental stimuli—mechanical (triboelectric, piezoelectric), thermal (thermoelectric, thermo‑osmotic), and water‑mediated (sorption, evaporation, capillarity, ionic streaming)—into usable electrical power. These approaches target distributed low‑power applications such as wearable and implantable electronics, autonomous sensors, and soft robotics, where wired power or bulky batteries are impractical. Recent progress has demonstrated functional devices based on hygroscopic ionic membranes, photothermal evaporators coupled to TE/thermo‑osmotic elements, nanocomposite hydrogels for moisture‑driven ionic generation, and microstructured tribo/piezo interfaces that produce burst power for intermittent loads. Key approaches include engineered material chemistries that sustain ionic gradients, hierarchical porosity for rapid fluid and heat transport, multifunctional composites that combine conductivity with mechanical compliance, and scalable fabrication routes (printing, roll‑to‑roll, laser patterning) that permit gradient and geometry control. Despite the extensive work that has been conducted on small-scale water energy harvesting, several bottlenecks are yet to be overcome before the translation of laboratory demonstrations into field-ready modules:

• (i) Many hygroscopic functional materials suffer from performance degradation caused by repeated humidity cycling, salt crystallization, mechanical fatigue, or biofouling. Future research should develop materials-by-design strategies that balance ion dissociation and selective transport with controlled water retention, mechanical resilience, and biocompatibility, for example through polyelectrolytes, zwitterionic chemistries, nanofiller-reinforced networks (e.g., MXenes, rGO, SiO₂), and dynamic or supramolecular bonding for self-healing and fatigue tolerance.

• (ii) Water-based harvesters generate low-frequency, pulsed, or environment-dependent outputs that are poorly matched to the steady power requirements of electronics. Therefore, the adaptive and hybrid architectures that reconcile burst power with continuous energy supply, including multimodal platforms that combine complementary mechanisms (e.g., hygroscopic ionic layers with thermoelectric modules, or triboelectric interfaces coupled to microstorage) should be focused on in the future.

• (iii) Without efficient power management, a substantial fraction of harvested energy is lost before reaching the load, particularly given the low-amplitude, high-impedance, and often intermittent output characteristics of small-scale water energy harvesters. Based on this, future research should co-design harvesters with ultra-low-loss power electronics, impedance-matched interfaces, and microscale buffers (e.g., microcapacitors, thin-film batteries, and hybrid microsupercapacitors) that are specifically tailored to low-frequency, pulsed energy inputs. In addition, energy-aware control strategies—such as event-driven operation, duty-cycled sensing, and adaptive load management—are needed to minimize standby losses and enable energy-neutral sensing and local processing under realistic environmental conditions.

• (iv) Scalability and manufacturability also constitute a major challenge. Many reported devices rely on bespoke fabrication processes or fragile architectures that are difficult to translate into industry. Multimaterial additive printing, roll‑to‑roll deposition, laser‑patterned gradient writing, and solution processing can be considered for future research, which can translate laboratory concepts to volume production as well as retain gradient and geometry control. Post‑fabrication tuning methods (laser or thermal writing) can also be considered to optimize devices without complex process chains.

• (v) The lack of standardized metrics currently hinders meaningful comparison among harvesting technologies, as performance is often reported under disparate environmental conditions and timescales. To address this issue, several priority metrics should be standardized. First, electrical output normalized by device area or volume is recommended to enable comparison across different architectures. Second, key test conditions—including humidity (e.g., 10–90%), temperature (e.g., 25 °C), flow velocity, and droplet parameters—should be explicitly specified and standardized. Finally, durability metrics, such as time to a defined level of performance degradation under controlled aging and environmental cycling, are needed to assess long-term stability.

Overall, small‑scale water energy harvesting stands at a transition from proof‑of‑concept demonstrations to practical, system‑level solutions. Success requires simultaneous progress in materials innovation, multiscale understanding of coupled transport, adaptive multimodal architectures, integrated energy conditioning, and manufacturable processes validated by standardized protocols. When materials selection, device architectures, power management electronics, and application requirements are considered in an integrated design framework, small-scale energy harvesters can deliver reliable and low-maintenance power for distributed sensing, autonomous soft robotic systems, and other self-sustaining technologies.