Advances in asymmetric moist-electric generators with innovative heterogeneous structures

Abstract

Harvesting energy from the ubiquitous moisture induced streaming potential or ion concentration gradients is a promising green technology. Lately, the emergence of high-performance asymmetric moist-electric generators (AMEGs) with an inhomogeneous distribution of their moist-electric active layer has significantly propelled this field forward, as the asymmetric structures can effectively regulate the moisture adsorption and ion migration. Nonetheless, a profound comprehension of the interrelationship between the working mechanisms, materials, and device structures remains essential. In this review, we provide a comprehensive account of AMEGs spanning from their fundamental research to real-world applications. We initially delve into the fundamental mechanisms and material engineering strategies centered on asymmetric architectures. Subsequently, we discuss device design strategies and power applications of AMEGs. Lastly, we provide an analysis of the challenges, potential solutions, and future trajectory for this promising technology.

---

Broader context

Unlike liquid water, moisture has long been underutilized and even recognized as a negative presence. Meanwhile, a large amount of energy is consumed annually for dehumidification. Harvesting energy through moisture-induced potentials is a promising green technology that offers a new approach to efficiently utilize moisture. Fabricating moist-electric devices through asymmetric structures is an effective strategy with the advantage of regulating ion migration. Despite these merits, the working mechanism of moist-electrics is not clear, which limits further development. Consequently, elucidating the intrinsic relationship between device configuration and working mechanism is crucial for the fabrication of asymmetric moist-electric generators (AMEGs). In this work, the authors systematically discuss device configurations, moisture utilization, and ion migration in AMEGs. More specifically, several main structures of AMEGs are further systematically divided, including gradient structure, multifunctional structure and ionic-diode structure, as well as the structure–performance relationship between the working mechanism and the structure. Finally, the potential applications of AMEGs based on the characteristics of their structure and output are further discussed. The ultimate goal of this work is to provide a convenient strategy for the development of advanced AMEGs while facilitating an in-depth understanding of moisture-solid interactions and the ion migration in solid porous materials.

1. Introduction

With the continuous growth of energy demand, developing renewable energy harvesting technologies has become an important way to address the energy crisis and environmental issues. A range of prototypes of generators that harvest energy from the environment have been developed, including solar cells,^1–3 thermoelectric generators (TEGs),^4–6 triboelectric nanogenerators (TENGs),^7–9 piezoelectric nanogenerators,^10,11 and moist-electric generators (MEGs).^12–17 Among these, MEGs can generate electricity from the ubiquitous moisture through streaming potential or ion concentration gradients.^15,16 With the ubiquitous moisture resource, continuous and considerable electricity can be achieved through ionic currents generated by the adsorption, storage and dissociation of moisture within MEGs. MEGs can not only be applied to self-powered technology for humidity response, but also as portable power sources for wearable electronics.^18,19

The electricity generated by MEGs is generally caused by the migration of ions within MEGs, which can be generated by the dissociation of water molecules after moisture adsorption or by the dissociation of functional groups or salts.²⁰ The former is commonly known as the streaming potential, while the latter is known as the ion concentration gradient. They are both based on the charge transfer between water molecules and solid surfaces, as well as the electric double layer (EDL) theory.^19,21 Particularly, the application of asymmetric structures in the energy field has gradually attracted attention in recent years. Especially in the field of MEGs, by partially doping or modifying the moist-electric layers and interfaces or introducing extra functional structures and devices to obtain asymmetric MEGs (AMEGs), the capacity of the moisture adsorption and dissociation or the ion migration can be effectively improved.^22,23

The performance of AMEGs has been progressively improved, especially the voltage output (up to 1.1 V) and stability (for several months); however, challenges remain in device fabrication, mechanism analysis, and output performance for the real-world applications.¹⁸ The primary issue is that even high-performance AMEGs still need to be integrated to stably power daily electronics. Numerous studies have been focused on improving the peak output, while the operational durability at high platform is less shown.²⁴ The internal ion transportation within MEGs lacks intuitive and effective means of characterization, constraining the understanding of the working mechanism. In the part of asymmetric device design, there is a need to deal with the balance between device fabrication complexity and output performance.

In this review, we focus on the construction and application of AMEGs and explore their unique advantages and innovations in working mechanisms and performance. Firstly, we introduce the basic working mechanisms of MEGs. Important materials and structures are then discussed. Asymmetric structures with different and multiple functions for MEGs are analyzed, including gradient-structured AMEGs, multifunctional-structured AMEGs and ionic-diode-structured AMEGs, as well as conformation-mechanism relationships under different asymmetric structures (Fig. 1). Meanwhile, we explore their broader applications in sustainable power sources and self-powered devices. Finally, we summarize and compare the characteristics of the above three structures in AMEGs, including output performance, fabrication complexity, operational stability, and potential application. Current challenges and opportunities for structural engineering and device design of MEGs are proposed, providing a roadmap for future developments in this field.

Fig. 1 A summary of asymmetric moist-electric devices according to the device design, device structure, and simplified working mechanism.

2. Working mechanisms

As an emerging and attractive technology for harvesting energy from the environment, moist-electrics can actively harvest and convert energy from ambient moisture. The fundamental theory of MEGs and the working mechanism of energy conversion are still to be clarified. Here we discuss the moist-electric technology from the perspectives of basic material properties, the theory of solid–liquid interface, and the energy conversion process. For most MEGs, power generation is achieved by asymmetric migration of ions, which is very similar to the mechanisms of osmotic energy harvesters.^25,26 Currently, there are two basic theories used to explain the working mechanisms of MEGs, which are the streaming potential-induced current and the ion gradient-induced current. In some special cases, these two mechanisms may occur simultaneously.^19,21,27,28

2.1 Moisture–solid interface

To better understand the intrinsic and fundamental mechanism of MEGs and the conversion process of moisture to electrical energy, we begin with a discussion of the nature of the water molecule and the interface and interaction between the solid–liquid phase. A single water molecule is composed of one oxygen atom and two hydrogen atoms, with an H–O–H angle of 104.5°. This molecular structure imparts polarity to water molecules and further enables hydrogen bonding between water molecules, and its strong dynamic nature allows protons to also be transported within the hydrogen-bonded medium, either from H₃O⁺ to an adjacent water molecule or from a water molecule to an adjacent OH⁻.^29,30 Atmospheric water exists in the form of clusters rather than individual molecules, and the distribution of the clusters depends on atmospheric humidity or water vapor pressure.³¹

Water clusters in the air can be extracted by hygroscopic materials, which are substances that absorb water into their bulk phases. The enthalpy change between gaseous water and absorbed water provides the energy output of moist-electric devices. However, due to the limitations of the second law of thermodynamics, the enthalpy change cannot be completely converted into electricity. Therefore, without the participation and interference of other factors, the output energy of moist-electric power generation (ΔE_MEG) is always less than the chemical potential difference (ΔE_CP) in the process of phase transformation (Fig. 2a).^32,33

Fig. 2 Two main working mechanisms of moist-electrics. (a) Schematic of the energy conversion of ion gradient-induced current. The chemical potential of moisture (ΔE_CP), which provides the energy for the dissociation of ions, resulting in ion migrations and the generation of electric potential energy in a moist-electric generator (ΔE_MEG). (b) The working mechanism of ion gradient-induced current. (c) Schematic of electrical double layer model of ionic solution approaching the solid surface. (d) The working mechanism of streaming potential-induced current.

2.2 Ion gradient-induced electricity generation

The ion concentration gradient can induce electricity generation in moist-electric devices in ambient humidity, realized by asymmetric moisture adsorption, dissociation and further ion migration based on hydrophilic and hygroscopic functional groups.^33,34 The oxygen-gradient-enabled AMEG was first proposed by Qu's group in 2015.¹² The side of the device with a higher oxygen content can adsorb more moisture and then release more migratable protons from the hydrophilic functional groups (relative to the other side with a lower oxygen content), thereby causing a proton concentration difference across the entire device and forming an ionic current (Fig. 2b).^12,14

Recently, the performance of ion gradient-induced AMEGs has been improved dramatically from the previously pulsed, low-voltage output (<40 mV) employing asymmetric structural designs and material optimizations. Two main strategies have been proposed to improve the performance of ion gradient-induced AMEGs, both by maintaining or expanding the ion concentration gradient. The simplest method is to incompletely encapsulate the device so that one side of the device is exposed to ambient humidity while the other side is encapsulated by using electrodes or an additional substrate.^18,24 The use of fluorinated polyimide film as a substrate for encapsulating ionic hydrogel-based AMEG was reported by Tao, who demonstrated that the device can generate an open-circuit voltage (V_OC) of 0.8 V by relying only on the asymmetric moisture adsorption due to incomplete encapsulation.¹⁸ Another common strategy is to achieve a continuous output by presetting functional group gradients or ion gradients in the moist-electric layer, which can be obtained by heterogeneous treatments such as heating, irradiation, doping, etc.^14,35

2.3 Streaming potential-induced electricity generation

Once a solid with a charged surface comes into contact with water (including not only liquid water, but also gaseous water), ions with opposite charges to those on the surface are drawn to the surface owing to the Coulomb interactions, creating an ionic cloud. Such ionic cloud and the surface charge then form the EDL between the solid–liquid interface, as proposed by Helmholtz.³⁶ On the liquid side, the ions with opposite polarity to the surface charge (counter ions) are attracted to the surface and form a compact layer (Stern layer), while the ions with the same polarity (common ions) are repelled.³⁷ The concentration of counter ions in the liquid phase away from the solid–liquid interface gradually decreases to form a diffusion layer (Fig. 2c).

Harvesting and converting energy from the streaming potential of a liquid flowing along a charged surface in the solid channel has been studied for more than a century.^38–40 Common streaming potential can be generated not only by bulk water, and evaporation-induced capillary water, but also indirectly by moisture–solid interactions.³⁹ When water is confined to a solid channel of comparable size to the Debye length of the water–solid interface, the EDL overlaps to a certain extent (related to the ion valence and ion concentration, Fig. 2c and d). When water moves along the charged nanochannel under a pressure gradient, or in response to a diffusive flow due to a difference in ion concentration, the ions located outside the shear plane will be transported along with the water molecules inside the charged channel (Fig. 2d). This net ion migration can generate electricity along the nanochannel.⁴¹ Since most MEGs are not configured with a stable liquid water source but directly use liquid water adsorbed and condensed from air, streaming potential-induced MEGs require not only good hygroscopicity of the moist-electric materials but also a certain amount of porosity to produce an electrokinetic effect.⁴² Besides, some evidence also suggests that it is not only the streaming potential that causes the generation of electricity, but also the potential, unidentifiable evaporation potential that may also contribute to the electricity generation in some special evaporation-integrated AMEGs.^43,44

2.4 Other mechanisms

AMEGs have also been designed through redox reactions between moisture and metal electrodes.¹⁷ Yang reported an AMEG with carbon-based electrodes and liquid metal electrode for the top and bottom electrodes, respectively.¹⁷ As shown in Fig. 3a and b, the moisture-induced ionic current is converted into usable electricity through a reaction similar to that of a metal–air battery.^17,45 Chen proposed a moisture-enabled cell (Fig. 3c and d), which utilized ‘‘working moisture’’ to achieve redox reactions through the asymmetric Au/Fe counter electrodes to obtain a stable V_OC over 1.0 V.⁴⁶ As shown in Fig. 3e and f, Qu introduced C-Al composite electrode in AMEG to reduce the diffusion of H⁺ ions into the PDDA layer by reacting H⁺ ions dissociated from the PSSA layer with Al to generate Al³⁺ ions, which can boost the voltage of the control AMEG from 0.3 to 1.1 V.⁴⁷ In addition, the redox reactions provided by the C–Al composite electrode may also enhance the output performance.

Fig. 3 Redox reaction-mediated moist-electric devices. (a) and (b) Redox-enhanced AMEGs using the liquid metal electrode and carbon nanotube-based electrode as counter electrodes. (a) and (b) Reproduced with permission.¹⁷ Copyright 2022, Springer Nature. (c) and (d) Redox-based AMEG using dissimilar metal electrodes as counter electrodes. (c) and (d) Reproduced with permission.⁴⁸ Copyright 2024, Wiley-VCH. (e) and (f) Redox-based AMEG using carbon–metal composite electrodes as counter electrodes. (e) and (f) Reproduced with permission.⁴⁷ Copyright 2023, Cell.

3. Materials

Moist-electric technology is based on moisture–solid and liquid–solid interactions, so the sizes and dimensions of the materials are important considerations for the designs of AMEGs. With the continuous development of nanotechnology, moist-electric materials have covered various dimensions of materials with different functions. Besides, we summarize and divide the candidate materials into five categories: carbon-based materials, inorganic nanomaterials, biomaterials, polymer materials, and hydrogel materials, according to materials composition and morphology.

3.1 Dimensions of materials

The dimensions and sizes of materials affect liquid–solid and moisture–solid interactions.³⁹ For instance, smaller dimensions lead to stronger surface and domain-limiting effects, which are typically beneficial for moist-electric power generation (related to the phenomena of EDL and moisture adsorption). After several years of development, the materials used for moist-electrics have crossed almost all dimensions.

Carbon-based dots have been used as zero-dimensional (0-D) materials for the preparation of moist-electric active layers due to the ease of surface modification and modulation, e.g., carbonized polymer dots containing abundant phosphate groups (Fig. 4a).⁴⁹ In addition, carbon dots can also be introduced into the moist-electric layer in the form of doping to increase the overall functional group content and hydrophilicity of the moist-electric layer.⁵⁰

Fig. 4 Different dimensional materials for AMEGs. (a) 0-D carbon-based dots with a large number of functional groups on the surface. Reproduced with permission.⁴⁹ Copyright 2023, Wiley-VCH. (b) 1-D nanowires for stacking to construct networks. Reproduced with permission.¹⁵ Copyright 2020, Springer Nature. (c) 2-D nanowires for stacking to construct nanochannels. Reproduced with permission.⁵¹ Copyright 2023, American Chemical Society. (d) 3-D MXene aerogel with better hydrophilicity than film-based structure. Reproduced with permission.⁵² Copyright 2023, American Chemical Society.

Both one-dimensional (1-D) nanowires (such as bionanowires) and two-dimensional (2-D) nanosheets (such as graphene oxide nanosheets, GO nanosheets) can be stacked and lapped to produce a network structure or stacked structure containing a large number of pores, which can improve the porosity and moisture adsorption of the active layer when applied to a moist-electric device, as shown in Fig. 4b and c.^12,15,51 Moreover, such stacked structures may also facilitate the directional transport or migration of moisture or ions. Park reported MXene-based three-dimensional (3-D) aerogel for AMEG, which has better hydrophilicity compared to the ordinary MXene film material, contributing to water transport and streaming potential (Fig. 4d).⁵²

3.2 Types of materials

Carbon-based materials were firstly reported, such as GO-based AMEG (GO-AMEG), due to their well-established synthesis method and wide range of sources.¹² When exposed to moisture, the GO nanosheets display fast adsorption and desorption kinetics of water molecules as the functional groups are extremely sensitive to water molecules. O^δ−–H^δ− bonds of the functional groups in GO can be weakened by water molecules adsorbed on the GO surface, releasing the mobile protons. Most carbon-based materials exhibit good electric conductivity and a large specific surface area, and their use as an active layer in AMEGs can reduce the internal resistance of the generator and obtain better hygroscopicity. By controlling the synthesis of carbon materials, materials with various pore sizes and structures can be obtained, such as modified carbon nanofibers with a hierarchical porous structure, which can contribute to the transport of water and the generation of streaming potential, as shown in Fig. 5a.⁴⁵ Considering that inorganic materials usually possess a stable lattice structure, good water stability, and humidity tolerance, inorganic-based moist-electric devices usually show good operation stability in theory.⁵³ Since the first successful report of titanium dioxide nanowires for efficient power generation from ambient humidity,¹³ a large number of inorganic materials have been synthesized and modified for moist-electric energy harvesting, including inorganic semiconductors,^54–56 inorganic oxides,^{17,45,57–59} and inorganic layered or 2-D materials (Fig. 5b).^{51,52,60–62}

Fig. 5 Different types of materials for AMEG. (a) Hydrophilic GO and carbon nanofiber with rich functional groups for carbon materials-based AMEGs. Reproduced with permission.^12,45 Copyright 2015, 2022 Wiley-VCH. (b) Inorganic and stable 2-D MXene and aluminium oxide for AMEGs. Reproduced with permission.¹⁷ Copyright 2022, Springer Nature. Reproduced with permission.⁶¹ Copyright 2023, Royal Society of Chemistry. (c) Microbial film and silk fibrion for biomaterials-based AMEGs. Reproduced with permission.⁶³ Copyright 2019 Elsevier. Reproduced with permission.⁶⁴ Copyright 2021 Wiley-VCH. (d) PSSA film and PAN/PSSA fiber for flexible AMEGs. Reproduced with permission.²⁰ Copyright 2019, Royal Society of Chemistry. Reproduced with permission.⁶³ Copyright 2021, Royal Society of Chemistry. (e) Humidity-tolerant water-storage hydrogel for integrated AMEG. Reproduced with permission.⁵² Copyright 2023, American Chemical Society. (f) Hygroscopic moisture-adsorption hydrogel for stretchable AMEG. Reproduced with permission.²⁴ Copyright 2023, Wiley-VCH.

To make AMEGs more biocompatible and environmentally friendly, biomaterials are gradually being used to fabricate green and sustainable AMEGs, such as protein nanowires, microorganisms, natural wood, and biological nanofibers (Fig. 5c).^15,63 For instance, biodegradable protein films extracted from the collagen of animal bodies, such as –NH₂ and –COOH groups, have a high concentration of amino acids and can easily absorb water molecules from the surrounding air.⁶⁴ Polymers are one of the most widely used materials in moist-electric technology, due to their ease of modification, good ion transport capacity and ionic conductivity (Fig. 5d).^65,66 Generally, polymer-based materials can be stretchable and flexible, or light transparent, which endow the AMEGs with a wide range of applications, such as skin electronics.^67,68

Hydrogel is a hydrophilic, cross-linked, three-dimensional network structure capable of storing large amounts of water or adsorbing moisture from the environment, and is therefore highly compatible with moist-electric technology.^59,69–72 There are two main roles for hydrogels in AMEGs: one is to act as a water-storage layer and the other is to adsorb moisture and participate directly in the formation of the streaming potential or ion gradient.^18,23,24,52 AMEGs that use only hydrogel as a water-storage layer have lower ambient humidity requirements, and the moist-electric layer can obtain water directly from the water-storage hydrogel (Fig. 5e).^52,61,72 Besides, by combining the moisture-adsorbing hydrogel with the moist-electric layer or evaporation layer, such dual-function AMEGs can be used to continuously generate electricity due to their cycle capacity for sustainable moisture adsorption and water evaporation (Fig. 5f).^44,52,61 The direct introduction of hydrogels into the hygroscopic layer of AMEGs requires dealing with the problems of dehydration or poor hygroscopicity, which can be achieved by the addition of glycerol or hygroscopic salts for moisture retention or moisture adsorption during the preparation of desired hydrogels.^18,24

4. Device structures

Here, we focus more on the relationship between the moist-electric layers, or between the electrode and the moist-electric layer, and cases like the use of only dissimilar metal electrodes in device design are not included. Based on the classification of materials, structures and functions adopted in moist-electric devices, we broadly classify the device designs of AMEGs into three categories: gradient-structured AMEGs fabricated mainly by inhomogeneous preparation and modification of materials, multifunctional-structured AMEGs obtained by integrating or in situ preparing different functional layers, and ionic-diode-structured AMEGs with special structures and functions or properties.

4.1 Gradient structures

There are two main electrical output-related problems faced by MEGs: insufficient moisture adsorption resulting in low device output, and rapid saturation of moisture adsorption or ion gradients resulting in low stability.^28,73 Constructing asymmetric and inhomogeneous gradient structures in MEGs is a common and effective method to solve the above problems (Fig. 6a).²⁷ As shown in Fig. 6b, Deng reported an aerogel-based AMEG with a structural gradient. By adjusting the surface roughness and micro-nanostructure of the carbon nanotube-based composite aerogel, a structural gradient with differences in hydrophilicity and water absorption can be achieved.⁷³ The difference in moisture adsorption within the generator can be spontaneously generated by this well-designed structural gradient to improve its output. Another type of structural gradient is the direct construction of wet-dry asymmetric gradients, which are semi-coated to the functionalized carbon layer by a hygroscopic ionic hydrogel (Fig. 6c).²³ Even long-term exposure to air as a whole cannot break the asymmetry.

Fig. 6 Gradient structures for AMEGs. (a) Schematic of the structural gradient for AMEGs. (b) Differences in hydrophilicity–hydrophobicity due to differences in micro-surface structure for moist-electric energy harvesting. Reproduced with permission.⁷³ Copyright 2023, Royal Society of Chemistry. (c) Planar AMEG with wet-dry asymmetric structure. Reproduced with permission.²³ Copyright 2022, Wiley-VCH. (d) Schematic of functional group gradient for AMEGs. (e) Schematic illustration of GO-AMEG with functional group gradient. (f) SEM image and O/C ratio of GO film with functional group gradient. (e) and (f) Reproduced with permission.¹⁴ Copyright 2018, Springer Nature. (g) Schematic of preset ion gradient for AMEGs. (h) In situ fabrication of nanowire arrays containing Na⁺ ion gradient for moist-electrics. Reproduced with permission.³⁵ Copyright 2018, Elsevier. (i) Schematic of the hybrid gradient for AMEGs. (j) and (k) Schematic and SEM image of the hybrid gradient for GO-AMEG. (j) and (k) Reproduced with permission.⁷⁴ Copyright 2023, Elsevier.

For ion gradient-induced MEGs, the construction of the functional group or ion gradients is a more straightforward way to fabricate AMEGs, according to the working mechanism of ion concentration gradient-induced current (Fig. 6d and g).^14,19,75 Partial reduction of oxygen-containing functional groups is undoubtedly an easy way to form functional group gradients, such as pristine GO and reduced GO with fewer oxygenated functional groups, which can be achieved by uneven heating or irradiation (Fig. 6e and f).^14,76 Chen achieved a longitudinal sodium ion gradient structure in the growth direction by precisely controlling the growth of polypyrrole nanowires (Fig. 6h).³⁵ Moreover, the application of double gradient or multi-gradient to AMEG has also been successfully demonstrated (Fig. 6i). Liu used a unique solid molding method to fabricate GO-AMEG, which mainly consists of multilaminate GO and reduced GO.⁷⁴ As shown in Fig. 6j, multilaminate GO has more oxygenated functional groups, and higher roughness compared to reduced GO, and thus this AMEG consists of not only functional group gradient, but also structural or conformational gradient, which ultimately can achieve a voltage output of more than 1.1 V (Fig. 6k).⁷⁴ Both structural gradients and preformed ion gradients are relatively easy strategies for device optimization and both have been extensively demonstrated to enhance the output of moist-electric devices (Table 1). Nonetheless, there are some issues with such methods, such as the difficulty of controlling the accuracy of the constructed gradient and how the gradient affects the output in detail.

Table 1 Summary and comparison of gradient-structured AMEGs

Device structure	Device feature	Stability^a	V OC (V)	J SC (μA cm^−2)	Ref.
a The index of stability is the time taken for the open-circuit voltage (VOC) or short-circuit current (JSC) of the moist-electric device to drop to 80% of its peak value.
Gradient-rGO/GO	Gradient of oxygen-containing groups and Schottky junction	<5 s	1.5	∼0.15	14
Gradient-GOF	Gradient of oxygen-containing groups	<5 s	∼0.02	∼5	12
P-rGO/GO	Gradient of oxygen-containing groups	>100 h	∼0.45	—	76
Heterogeneous mrGO	Double gradient	>12 h	1.39	120	74
GO/GO-rGO	Gradient of hydrophilic groups	∼24 h	0.76	73	77
CSWNT/NSWNT-aerogel	Double gradient	—	1.45	117	73
AHS fabric	Preformed wet-dry gradient	168 h	0.65	—	23
Gradient-doped PPyNWs	Preformed Na^+ ion gradient	<1 s	0.072	—	35
Polarized-G-3D-PPy	Preformed ClO4^− ion gradient	<5 s	0.06	12	78
g-POMs-GO	Phosphotungstic acid gradient	<250 s	0.015	6.2	79
Al3^+ gradient-NWs	Preformed Al3^+ ion gradient	<5 s	0.13	0.342	80
1T-2H MoS2 nanosheets	Phase gradient (1T-2H)	>500 s	0.019	6.24	81
SA/MWCNT fibers	Radial oxygen gradient	>36 h	0.38	1.74	82
ORC/ARC-aerogel	Gradient of functional groups	>24 h	1.07	—	83

---

4.2 Multifunctional structures

The directed migration and transportation of the mobile ions is essential to the moist-electric energy conversion process. In such a process, the directed migration of ions is thought to be primarily or only driven by differences in ion concentrations, which are due to the conformation gradient or moisture gradient.^33,84 The former is usually realized by the above-mentioned structural gradients and functional group gradients, while the latter can be achieved through specific functional materials and structural design. As shown in Fig. 7a, the most common moisture gradient is typically achieved with a semi-open device design, where one part of the device is sealed (moisture proof) while the other part is well breathable (moisture adsorption). The desired moisture gradient can be easily created inside the device to generate electricity by employing additional substrate or special electrode combined with counter breathable electrode (Fig. 7b).^18,22,24

Fig. 7 Multifunctional structures and integrations for hybrid AMEGs. (a) Schematic of semi-sealed design for AMEGs. (b) Semi-sealed functional structure for moist-electric device. Reproduced with permission.¹⁸ Copyright 2022, Wiley-VCH. (c) Schematic of interface modification for AMEGs. (d) Interface modification for Schottky junction-based AMEGs. Reproduced with permission.¹⁴ Copyright 2018, Springer Nature. Reproduced with permission.⁸⁵ Copyright 2019, Royal Society of Chemistry. (e) Schematic of integrated functions for AMEGs. (f) The integration of LiCl-coated hygroscopic layer and evaporation layer for AMEGs. Reproduced with permission.⁴⁴ Copyright 2022, Springer Nature. (g) Decreasing V_OC with increasing relative humidity. Reproduced with permission.⁵² Copyright 2023, American Chemical Society. (h) The schematic of phyto-inspired AMEG via moisture adsorption–evaporation cycles. Where P(VDF-TrFE) stands for poly(vinylidene fluoride-trifluoroethylene), PAN stands for polyaniline, and CNW/LiCl stands for cellulose nonwoven/LiCl. Reproduced with permission.²⁷ Copyright 2024, American Association for the Advancement of Science. (i) The left part is the reported hybridized generators based on MEGs and their general output performance. The right part is the reported MEGs based on other technologies. (j) Schematic illustrations of the reported structures of AMEG-based hybridized generators. Where TEG stands for thermoelectric generator, PV stands for photovoltaics, and TENG stands for triboelectric generator. (k) Structure and electricity generation of the AMEG-based hybridized generator integrated with a triboelectric generator. Reproduced with permission.⁶¹ Copyright 2023, Royal Society of Chemistry. (l) V_OC (tested under different environmental conditions) of the hybrid moist-thermoelectric generator. Reproduced with permission.⁸⁶ Copyright 2023, Wiley-VCH. (m) Current performance of light-coordinated AMEG. Reproduced with permission.⁸⁷ Copyright 2021, Wiley-VCH.

There are also some problems of ion or electron migration (such as energy level mismatch and interfacial barriers) at some specific electrode/moist-electric-layer interfaces, which can be improved by interfacial modulation (Fig. 7c). For specific GO-AMEGs, the use of Au–Ag electrode pairs can achieve better performance than Au–Au electrode pairs and Ag–Ag electrode pairs, due to the presence of the Ag–GO Schottky junctions that promote carrier separation (Fig. 7d).^85,88 The nonlinear response caused by the Ag–GO Schottky diode has also been demonstrated by the current–voltage characteristic curves, with no significant nonlinear current–voltage characteristic curves observed in Ag–Ag and Au–Au electrode-based control devices.^14,85

Integrating different functional moisture-adsorption layers or water-storage layers with the evaporation layer is another effective moisture management strategy (Fig. 7e). Guo reported self-sustained AMEGs consisting of LiCl-coated fiber paper and carbon black-coated fiber paper (Fig. 7f).⁴⁴ The hydrophilic and hygroscopic LiCl-coated fiber paper is capable of adsorbing a large amount of moisture and pumping the moisture to the hydrophobic carbon black-coated fiber paper with rich nanochannels. The captured water then evaporates naturally to complete the sustainable cycle of moisture adsorption and evaporation.²⁸ The electricity is usually generated by the evaporation-induced streaming potential in this type of AMEG, which is integrated with an evaporation layer, thereby necessitating a certain size and quantity of nanochannels within the evaporation layer.^39,43,72,89 Water-rich, water-retaining, but poorly hygroscopic hydrogels can also be integrated with evaporation layers for AMEGs, which have a broader range of applications with lower humidity requirements for the working environment. However, high humidity or low temperature may degrade the performance of the device (Fig. 7g), and the water in the generator cannot be efficiently evaporated and transported to generate the streaming potential due to the slowed evaporation rate.⁵² Inspired by the roots (water absorption), stems (water transportation) and leaves (water evaporation) of plants,^27,90 Wang group proposed a similar but optimized fibric-based AMEG equipped with a good capacity for moisture transport and evaporation. In this tri-layer device, the evaporation layer was optimized into a bilayer structure integrated with the moist-electric layer, or evaporation layer (Fig. 7h). The evaporation layer consists of polyacrylonitrile and poly(vinylidene fluoride-trifluoroethylene) membranes, allowing for enhanced evaporation and unidirectional water transportation due to the hydrophobic gradient.²⁷ Besides the design of the evaporation layer introduced in AMEG, power generation can also be achieved by a similar process, moisture desorption. Qu reported a well-designed organic/inorganic composite membrane (sodium alginate/SiO₂/reduced GO) that can adsorb moisture at high humidity. This membrane can produce an asymmetric distribution of Na⁺ ions inside the device and then generate electricity. While at low humidity, the interior of the water-containing sodium alginate/SiO₂/reduced GO layer can gradually dehumidify, and the originally asymmetric Na⁺ ions begin to migrate in the opposite direction, generating the opposite electrical signals.⁹¹ Compared with previous AMEG by harnessing a single moisture adsorption process, the full cycle AMEG integrates adsorption and desorption enabled power generation into a closed-loop process, thus affording repeatable electricity-generating performance.

In addition to combining and integrating with different functional layers, AMEGs can be designed as hybridized generators capable of harvesting other types of energy simultaneously, such as integrating with photovoltaic devices, TEGs and TENGs (Fig. 7i).^61,86,92,93 As shown in Fig. 7i–k, due to the output characteristics, AMEG-based hybridized generator (AMEG-HG) with the integrated TENG is usually prepared as two modules whose output power is managed independently (the TENG part requires a rectifier bridge, while the MEG part does not).⁶¹ Since the behavior of ion migrations in the specific thermoelectric mechanism is similar to that in the moist-electric devices, thermoelectrics-integrated AMEG-HG can be prepared by combining the thermoelectric power generation part and the moist-electric power generation part as a whole (Fig. 7j). Owing to the integrated design, the thermoelectrics-integrated AMEG-HG is not required to deal with the problem of power matching, while the existing photovoltaics-integrated AMEG-HG requires further consideration of its power matching due to its reported immature design, where the photovoltaic power generation can produce several milliamps of current, while the MEG power generation can only generate a few microamps of current (Fig. 7i).⁹³

Thermoelectrics-integrated AMEG-HG have been reported to concurrently harvest energy from moisture, since the carriers in both devices primarily exist in ionic form.⁹⁴ Despite the relative difficulty of monolithically integrating thermoelectrics and moist-electrics and the unclear working mechanism, which involves complex ion migrations induced by both humidity and temperature differences, great interest has been aroused.^86,92,95 Yan proposed a two-in-one strategy to prepare a hybridized generator by combining the working mechanisms of the thermoelectrics and moist-electrics. The proton transport within the generator can be significantly enhanced under the influence of both temperature and humidity. The polyelectrolyte membrane consisting of a crosslinked copolymer of poly(2-acrylamide-2-methylpropane sulfonic acid) and poly(sodium styrene sulfonate) can dissociate to provide migratable H⁺ and Na⁺ ions, and even after humidity saturation, the migration of Na⁺ ions can be driven by temperature difference to generate electricity continuously.⁸⁶ Through such strategy and design, the hybridized generator can generate a stable, V_OC close to 2 V in an environment with a 60% relative humidity difference and 15 K of temperature difference (Fig. 7l).

Light can excite photosensitive materials in addition to radiating heat, and based on this, Qu's group reported a light-coordinated AMEG capable of simultaneously harvesting solar energy, which consists of a hybrid film of hygroscopic polyelectrolytes and photosensitive materials (Fig. 7m).⁸⁷ The gradient distribution and migration of H⁺ ions can reduce the recombination of photogenerated carriers, while the separated photogenerated carriers are also able to improve the ionic conductivity of the device. Under the mediation of light, the short-circuit current density (J_SC) generated by moist-electric energy conversion can be increased from 0.15 mA cm⁻² to 0.5 mA cm⁻². In addition to the concept of light coordination, Zhou proposed a hybridized moist-electric-photovoltaic generator for all-weather energy harvesting based on the existence of microbial photovoltaic systems (Geobacter sulfurreducens and photosystem II particles). This hybridized generator produces a photovoltaic electric field in the same direction as the moist-electric field, and can generate a V_OC of 0.45 V and a J_SC of 4.2 μA cm⁻² without light, whereas they be increased to 0.7 V and 6.4 μA in the presence of light (1.5 mW cm⁻²).⁹⁶ A summary and comparison of the output performance for the currently reported multifunctional-structured AMEGs are listed in Table 2.

Table 2 Summary and comparison of multifunctional-structured AMEGs

Device structure	Device feature	Stability^a	V OC (V)	J SC (μA cm^−2)	Ref.
a The index of stability is the time taken for the open-circuit voltage (VOC) or short-circuit current (JSC) of the moist-electric device to drop to 80% of its peak value.
Printed-GO/MIS	Asymmetric moisture adsorption	∼50 s	0.7	0.3	22
GO/PAAS	Extra moisture adsorption and Schottky contact	120 h	0.6	∼1.2	85
PVA-PA-Glycerol hydrogel	Asymmetric moisture adsorption	>1000 h	∼0.8	240	18
PAM-AMPS-LiCl hydrogel	Asymmetric moisture adsorption	>3 h	0.81	480	24
LiCl/CB-Cellulon paper	Integration of evaporation and moisture adsorption	>240 h	0.78	∼7.5	44
C/P(AMPS-SSS0.5)/C	Integration of thermoelectrics and moist-electrics	118 h	1.81	>200	86
GO/ZnO heterojunction	Interfacial modification	8 h	∼0.4	0.15	56
PSSA/(Rose Bengal)	Sunlight-coordinated system	∼47 h	0.92	1550	87
PSSA-PEDOT:PSS-FeCN^4−/^3−	Integration of thermoelectrics and moist-electrics	—	0.9	800	95
SiNWs/PDDA	Extra moisture adsorption	70 h	1	8.2	55
Ti3C2Tx/PAM-hydrogel	Extra moisture storage	∼120 h	0.6	1160	52
Ti3C2Tx/ionic-hydrogel	Extra moisture storage	>27 h	0.3	∼1600	61
LiCl/CBs@fabric	Integration of evaporation and moisture adsorption	—	0.3	50	72
PSSA-Kc	Integration of light-trapping and interfacial modification	>180 h	0.8	1600	97
ZnAl-LDHs	Integration of tribovoltaics and moist-electrics	—	0.69	6500	62

---

4.3 Ionic diode structures

Recently, inspired by p–n junctions that regulate photogenerated carrier separations in solar cells or by ionic diode-like structures that modulate ion migrations through transmembrane potentials in cell membranes (e.g., electroactive cells in electric eels), AMEGs with ionic diode structures have been developed.^{16,17,57,69,98–100} The main practical problem faced in the moist-electric conversion process is the inability of the device to deliver a long-term and stable output, which is mainly due to the lack of sustainable ion gradients inside the device. The ionic diode can modulate the distributions of anions and cations and maintain a certain ion concentration difference, which is mainly achieved by two heterogeneous moist-electric layers with opposite surface potentials.^16,17,98Fig. 8a and b illustrate two main structures of ionic-diode-structured AMEGs, respectively, according to their working mechanisms. In the former, ion modulation of the ionic diode is mainly achieved by ions dissociated by external moisture (including H⁺ and OH⁻, Fig. 8a), while in the latter, it is mainly achieved by the dissociation of functional groups (e.g., H⁺ ions from –SO₃H group and Cl⁻ ions from –NCl group, Fig. 8b) in the moist-electric layer itself.^16,17 Moreover, both types of ionic diode structures typically require positively and negatively charged moist-electric materials with high porosity and hygroscopicity for sufficient moisture adsorption. The former also requires that the positively and negatively charged layers on rich nanochannels facilitate the generation of the streaming potential (under the influence of the built-in electric field from the ionic diode). The most distinctive features of AMEGs with ionic diodes are the nonlinear voltammetric curve and the behavior of ion rectification, which mainly originated from the diode structure, i.e., the differentiated responses to applied positive and negative bias voltages (Fig. 8c and d).^16,17,101

Fig. 8 Ionic diodes for high-stability AMEGs. (a) and (b) Schematic of ionic diode-type AMEGs based on (a) streaming potential and (b) ion concentration gradient. (c) Differential responses to opposite bias voltages in diode-based AMEG. (d) The nonlinear voltammetric curve in ionic diode-based AMEG. (e) Preparation of ionic diode-based AMEG by a simple casting and spraying strategy. Reproduced with permission.¹⁶ Copyright 2021, Springer Nature. (f) Schematic illustration of the modified protein with different treatment and Zeta potentials of the different protein dispersions. Reproduced with permission.⁶⁴ Copyright 2023, Royal Society of Chemistry. (g) Schematic of the electrode-based diode for AMEG with ultra-long and stable working performance. Reproduced with permission.¹⁷ Copyright 2022, Springer Nature. (h) Schematic of the anti-freezing AMEGs based on hydrogel ionic diode. J–V curves of ionic diode-based AMEG in the ±2 V bias interval and output performance at low temperature. Reproduced with permission.⁵⁹ Copyright 2023, Royal Society of Chemistry.

Inspired by the asymmetric lipid bilayer structure, Qu's group proposed a bilayered AMEG with a direct-assembled ionic diode, consisting of sequentially sprayed negatively charged polyanion film and positively charged polycation film (Fig. 8e).¹⁶ The bilayer structure exhibits a nonlinear response with significant and differentiated current responses under positive and negative bias voltages. This AMEG can adsorb moisture and dissociate the polyanion and polycation films to generate H⁺ and Cl⁻, respectively, and produce differences in anion and cation concentrations, which can work continuously for 250 hours with a peak voltage of 1.0 V.¹⁶ Another approach is to modify neutral materials with differentiated treatments and modifications.^64,102 Chu presented a protein-based AMEG in which the proteins are modified by acid and alkali treatments, respectively, resulting in the formation of positive groups (–NH₃⁺) and negative charges (–COO⁻) in the protein structure, respectively. The acid-treated proteins are positively charged while the alkali-treated proteins are negatively charged (Fig. 8f).⁶⁴ In addition to acid and alkali treatments, the carboxylation and quaternization of neutral materials can also be used to construct the required ionic diode structure for AMEGs.¹⁰²

Considering the complexities of the structures and preparations of ionic diode-based devices, a fabrication-simplified AMEG that directly uses carbon nanotube electrodes as a negatively charged layer in contact with positively charged anodic aluminium oxide has been reported. This method is also effective in constructing the ionic diode junction that produces significant nonlinear current (between ±1.2 V bias voltages) and this diode-structured AMEG still maintain its voltage at 1.1 V after 700 hours of operation (Fig. 8g).¹⁷ Yang also reported an ionic-diode-structured AMEG that uses a highly hygroscopic hydrogel as a negatively charged layer (Fig. 8h). By regulating the content of LiCl in the hydrogel polymer network, moisture adsorption and anti-freezing are successfully achieved simultaneously. Even at −20 °C, the ionic-diode-structured AMEG still has a good rectification ratio as well as a voltage.⁵⁹ A summary and comparison of the output performance for the currently developed ionic-diode-structured AMEGs are listed in Table 3. Generally, ionic-diode-structured AMEGs show better stability and maintain higher voltages than gradient-structured or multifunction-structured AMEGs due to their better regulation of ion transport and distribution.^25,103,104

Table 3 Summary and comparison of ionic-diode-structured AMEGs

Device configuration	Positive layer^a	Negative layer^b	Stability^c	V OC (V)	J SC (μA cm^−2)	Ref.
a The positive layer generally refers to a material with a positive Zeta potential. b The negative layer generally refers to a material with a negative Zeta potential. c The index of stability is the time taken for the open-circuit voltage (VOC) or short-circuit current (JSC) of the moist-electric device to drop to 80% of its peak value.
C/PDDA/PSSA/C	PDDA	PSSA	>50 h	1.38	1.0	16
CNT/AAO/In-Ga	AAO	CNT	>30 d	1.1	11.3	17
HPCNF/AAO/EGain	AAO	HPCNF	—	1.1	27	45
CNT/Al2O3/EGain	Al2O3	CNT	>30 h	1.03	47.77	57
C/Hydrogel/AAO/In-Ga	AAO	Hydrogel	>60 h	1.25	300	59
CNT/AAO/CaCl2/EGain	AAO	CNT	—	1	350	58
C-Al/PSSA/PDDA/C	PDDA	PSSA	> 13 h	1.1	∼4	47
Au/PDDA/PSSA/Fe	PDDA	PSSA	∼15 h	1.08	50	46
(GO)PANI/F-Nafion(PDDA)	(GO)-PANI	F-Nafion (PDDA)	∼10 h	0.9	8	105
Cu/PDDA/PSSA/Cu	PDDA	PSSA	—	0.8	∼500	106
MXene-PDACl/MXene-PSSNa	MXene-PDACl	MXene-PSSNa	—	∼0.07	—	60
Acid-protein/Base-protein	Acid-protein	Base-protein	>90 d	1.45	113	64
Pt/Anion-wood/Cation-wood/Pt	Cation-wood	Anion-wood	∼25 h	0.57	77	102
Ni/NSWNT/CSWNT/Ni	NSWNT	CSWNT	—	1.45	117	73
Pt/C-SilkNF/N-SilkNF/Pt	Cationic-SlikNF	Negative-SilkNF	—	0.12	∼0.1	107
PC-Hydrogel/PA-Hydrogel	PC-hydrogel	PA-hydrogel	∼25 h	0.05	—	108

---

5. Applications

The enormous progress in the development of AMEGs has spawned potential applications that can be distinguished into two categories depending on their functions: (1) self-powered functional devices and (2) sustainable power sources.

5.1 Self-powered functional devices

As the moist-electric device is highly sensitive to ambient humidity, one of the most straightforward applications of AMEG is a self-powered humidity sensor. Living organisms are capable of generating a changing humidity field within a certain range, which can help to enable moisture-based non-contact sensing (Fig. 9a).^109–111 Yang reported a sensing unit consisting of an ionic-diode-structured AMEG that can be used to achieve sensing and perception by arraying. Sensing of living and non-living things in non-contact mode can be achieved with the introduction of a humidity-sensitive aluminium oxide/carbon nanotube-based diode (Fig. 9b).¹¹¹ Besides, a single-component multimodal sensor based on GO-AMEG is used for multimodal sensing and monitoring, which avoids most of the problems faced by integrated multi-sensors, such as complex circuitry and cumbersome structures. Through a well-designed machine learning model, the coupled response of GO-AMEG to multiple stimuli such as light intensity, temperature, humidity, and pressure change can be decoupled, and skin temperature, pulse, and specific gesture signals can be monitored and recognized at the same time (Fig. 9c).¹¹²

Fig. 9 Self-powered functional devices based on AMEGs. (a) The response curves of the current over time when the finger is at different distances from the device. (b) The sensing array selectively recognizes the approach of fingers. (a) and (b) Reproduced with permission.¹¹¹ Copyright 2024, Elsevier. (c) Schematic illustrating of multimodal sensing based on GO-AMEG to simultaneously monitor light intensity, temperature, humidity, and pressure change empowered by the machine-learning model. Reproduced with permission.¹¹² Copyright 2022, Wiley-VCH. (d) Schematic illustration of AMEG-based memory cell array. (e) AMEG-powered reading of binary codes and corresponding demodulated display for the word “BIT” according to the ASCII. (d) and (e) Reproduced with permission.⁹⁴ Copyright 2016, Wiley-VCH. (f) Schematic illustration of reading information encrypted into AMEG by moving electrode. (g) Schematic illustrating of the programmable AMEG for the desired purpose and the according surface-potential maps under changing relative humidity. (h) Surface-potential maps of a QR code-integrated AMEG featuring a reversal information-encryption strategy. (f)–(h) Reproduced with permission.¹¹³ Copyright 2022, Wiley-VCH.

In addition to sensing and perception, AMEG can also be used for information storage, encryption, and display.¹¹⁴ Qu reported an AMEG-based memory device for moisture-powered information storage. When moisture enters the device, it creates a continuous ionic gradient within the device, analogous to writing data into a memory device (Fig. 9d).¹¹⁵ The switching effect of this novel memory device is entirely dependent on the humidity difference, and the switching ratio (ON/OFF ratio, up to 10⁶) is much higher than that of conventional resistive memory diodes with an ON/OFF ratio of only ≈10⁴.⁹⁴ The pre-programmed word “BIT” can be displayed by the standard eight-bit code under the surroundings of human-breathing operation (Fig. 9e). Tan reported a self-sustained hygroelectronic interface based on AMEG to pattern the hygroscopic layer to achieve persistent concealment and display of information through surface potential differences. As shown in Fig. 9f, persistent, inhomogeneous water distribution is achieved by distributing the LiCl/poly(vinyl alcohol) hydrogel unevenly on the carbon-based bulk material, and this designed inhomogeneous distribution can be extracted by applying an electrical signal to obtain encrypted information.¹¹³ Monitoring the electrical signals generated by the difference between dry and wet on the surface of the AMEG employing moving electrodes enables the display of information, due to the moisturized potential difference (from −100 mV to 20 mV) generated by uneven adsorption of moisture on the AMEG (Fig. 9f).^23,113,116 AMEG-based information technology has the potential to be a promising technology that can decode and encrypt information at specific humidity conditions through patterned designs in AMEG (Fig. 9g and h).

5.2 Sustainable power sources

The sustainable and stable power generation from well-designed AMEGs can support a series of mobile electronics and high-power applications. Meanwhile, self-powered system based on AMEGs are capable of minimizing external energy dependency and offering an exploration of power-free electronics.^87,117,118 Chen reported a long-cycle-life AMEG using a metal electrode with self-charging behavior, which allows the Fe electrode to be continuously oxidized to obtain the required Fe³⁺ ions utilizing a charging moisture such as H₂O₂ (Fig. 10a). As long as specific oxidation conditions are met (e.g. in HNO₃, KMnO₄, and HClO), the device can be capable of self-charging and self-discharging by replacing the charging and working moisture (Fig. 10b).⁴⁸ As the AEMGs generate direct-current electricity, the generated power can also be stored in energy storage devices.^16,119 Zhang's group reported a sandwich AMEG integrated with a polymer-based supercapacitor in which AMEGs and the supercapacitor share the same bottom gold electrode.¹¹⁹ Besides, as shown in Fig. 10c, the supercapacitors have also been reported to be used directly for integration with the AMEG, with one electrode of the AMEG and one electrode of the supercapacitor forming the positive and negative electrodes of the self-charging system.¹²⁰ As shown in Fig. 10d, we also concluded and compared several existing AMEG-based hybrid power sources. The most common is the conventional discrete system where the power supply unit (AMEG) and the energy storage unit (supercapacitor) are connected only by wires, which is not beneficial for device integration. The above problem can be solved by applying a co-electrode or adopting a tandem structure to improve the integration rate of the whole system, but complicated device fabrication needs to be solved (Fig. 10d). The co-electrode power source systems of AMEGs are to some degree easier to prepare, compared to tandem systems, but require additional electricity or operation to switch between charging and discharging for specific applications. Meanwhile, the preparation process is more complicated for tandem systems, where the supercapacitor is directly connected in series with the power generation part (during the preparation process), and the whole system can provide a more stable and sustainable voltage output.

Fig. 10 Sustainable power sources based on AMEGs. (a) The charging mechanism of rechargeable AMEG by the introduction of H₂O₂. (b) The corresponding output voltage curves of polymer-based AMEG with different charging moistures (HNO₃, KMnO₄, and HClO). (a) and (b) Reproduced with permission.⁴⁸ Copyright 2024, Wiley-VCH. (c) A hybrid system based on the integration of GO-AMEG and GO-based supercapacitor. Reproduced with permission.¹²⁰ Copyright 2024, Springer Nature. (d) From top to bottom are a schematic of a conventional discrete system based on MEGs, a schematic of a self-charging system with a common electrode between the MEG and supercapacitor, and a schematic of a MEG-based self-charging system with a tandem structure, respectively. (e)–(g) Schematic illustrations of large-scale integration of AMEG units by (e) laser processing¹⁶ (Copyright 2021, Springer Nature) (f) 3D printing¹²¹ (Copyright 2024, Wiley-VCH), and (g) Miura-ori folding strategy¹⁶ (Copyright 2021, Springer Nature). (h) Schematic of using a protein-based AMEG to power the Ge–Si nanowire transistor. Reproduced with permission.¹⁵ Copyright 2020, Springer Nature. (i) Electro-deposition of nickel millimeter-level structure by 5 × 5 hydrogel-based AMEGs. Reproduced with permission.¹⁸ Copyright 2022, Wiley-VCH. (j) The outdoor tent window is stitched with 8 × 240 series-parallel fabric-based AMEGs to directly supply power to the mobile phone. Reproduced with permission.²⁷ Copyright 2024, American Association for the Advancement of Science.

The array stacking constructed series-parallel integration of AMEGs enables their high-voltage applications such as electrodeposition and household appliances.^16,18 Due to the excellent automation and programmability of 3D printing technology, large-scale integration of AMEGs can be achieved by directly printing conductive interconnects between separate generator units, which is hardly achieved by low-precision laser cutting or complex manual assembly (Fig. 10e and f). This printing technique has been used in the fabrication of AMEG arrays, which can obtain voltages over 180 V with only a small linear deviation of the current after paralleling (Fig. 10f).¹²¹

Directly using AMEGs through series and parallel connections always face the problem of undesired large size and poor flexibility, which is expected to be solved by layer-by-layer printing and origami folding.⁹⁸ For example, a folding strategy using Miura-ori has been reported to stack 2D structures into 3D structures for the folded fabrication of GO/reduced GO-structured AMEGs on a paper substrate (Fig. 10g).⁹⁸Miura-ori folding not only optimizes device integration, but also reduces non-operating self-discharge through self-registered mechanical contact. Yao reported the use of a single protein-based AMEG for direct powering of Ge/Si nanowire transistors, with good logic behavior due to the good stability of this well-designed AMEG and without the need for other external power sources (Fig. 10h).¹⁵ Typically, it is difficult for a single AMEG to be used directly as a high-power power source, so arrays of AMEGs with different sizes need to be designed for power sources by above integrated methods. For example, the integration of hydrogel-based AMEG on a substrate by a simple coating method can be used for the electroplating of millimetre-sized nickel metal (Fig. 10i).¹⁸ In addition, large-scale paralleling of AMEG can be used directly for milliampere-level or even ampere-level applications, such as charging mobile phones (Fig. 10j).²⁷

6. Conclusions and outlook

Overall, the advances, including but not limited to the emergence of novel asymmetric structures, the mechanism of the enhanced performance, and the potential of wider application in regard to AMEGs are reviewed in depth. Research on AMEGs has flourished in recent years, with encouraging results greatly improving the performance of AMEGs and progressing it towards practical applications. Some typical output voltages and currents are summarized and plotted in Fig. 11a, and factors such as output performance, fabrication complexity, operation stability, and wider application have also been taken into account for a comprehensive assessment (Fig. 11b). For instance, hybridized AMEGs integrated with thermoelectric materials can be designed to achieve an impressive output of close to 2 V. AMEGs based on ionic diodes generally have higher output stability and V_OC due to their built-in electric field for ion modulation. Despite the significant progress in output power density and stability, AMEGs still face grand challenges, such as the preparation of materials and device structures. Due to the variety of materials and structures used for AMEGs, we briefly summarize the preparation and construction of AMEGs in Challenge Part (Fig. 11c). Gradient-structured AMEGs generally place high demands on the construction of functional group gradients or structural gradients, and require precise regulation of synthesis conditions.^12,73,122 For multifunctional-structured AMEGs, complicated preparation processes need to be solved, such as layer-by-layer growth and assembly of different functional layers.^16,120 The preparation of ionic-diode-based AMEGs requires overcoming the poor hygroscopicity of the positively and negatively charged materials, as well as optimizing the interfacial contacts between the positively and negatively charged materials and between the functional materials and the electrodes.^45,59

Fig. 11 Summary of the structures and output performance of AMEGs. (a) V_OC and J_SC of AMEGs with different structure configurations. (b) Horizontal comparisons of AMEGs with different asymmetric structures in terms of output performance, complexity, stability, as well as wider application. (c) Characteristics, challenges and applications of the high performance AMEGs.

6.1 Fundamental understanding of the working mechanism

Multiple theories have been proposed to explain the working mechanisms of MEGs, but understanding their mechanisms fundamentally and intuitively remains a challenge.^13,21 The adsorption and migration process of water molecules, either in gaseous or liquid state, and ions within the asymmetric device involve several processes, and sometimes chemical redox reactions occur between different interfaces. Powerful characterization technologies such as surface potential microscopy,¹⁶ nuclear magnetic resonance spectroscopy,¹²³ and scanning electrochemical microscopy,¹²⁴ which can be utilized to investigate these complex chemical/electrochemical processes as an electrochemical surface measurement and analysis method. It may be feasible to use a reference electrode with high stability and reversibility commonly used in electrochemical systems, such as Ag/AgCl electrodes, to implement the research of moist-electric devices.

6.2 Balance of fabrication complexity and performance benefits

Asymmetric structures can indeed improve the performance of MEGs.^22,23,69 However, the benefits of device performance improvement over the preparation cost are not ideal. Constructing asymmetric hygroscopic structures is a simple method.²³ Ionic diodes can modulate ion migration and distribution, which is an effective solution to improve device output and stability simultaneously. However, the fabrication process needs to be further optimized, such as using doping, dip-coating, and other methods.

6.3 Enhancement of electrical performance

The output of AMEGs is generally not high enough to directly power electronics. A single AMEG usually generates a voltage of around 1 V and a current of a few μA. The main drawback is their relatively low power density (usually at μW cm⁻²). Designing materials with improved hygroscopicity and moisture dissociation is currently a mainstream strategy. However, the materials should also be efficient in ion transportation, regarding the enhancement of intrinsic ionic conductivity or creation of physical ion-conductive pathways. Integration with other energy techniques, such as a thermo-electric generator, is a good solution to enhance the performance of AMEGs and overcome the output fluctuation problem of a single type device. Besides, the electrodes for AMEGs still present challenges, with the need to consider their effect on moisture adsorption and the conversion of ionic current to electronic current.²⁸ The use of active metal electrodes does enhance device performance, especially the current density, while their use in the research of AMEG devices and mechanisms should be cautiously considered. At the application level, the use of active metal electrodes may be an option if it makes the AMEG competent for its life cycle (several months, for instance).

6.4 Environmental adaptability and stability

AMEGs generally need to work at specific humidity levels and may not be able to effectively output electricity in extremely arid regions.¹² The introduction of hydrogels for water harvesting or storage is expected to expand the application prospects.²³ Although higher temperatures can accelerate the migration of ions within the device and improve the output performance, excessive heat may cause dehydration. It could be hybridized with thermoelectrics and the recently reported water-evaporation-based water-electricity co-generation systems, ensuring sustainable outputs under high temperatures.⁴⁴ Besides, to improve the output performance of AMEGs, moisture adsorption and ion dissociation are typically enhanced by increasing the hygroscopicity of the system; however, excessive hygroscopicity may lead to poor stability of AMEG, and there is a need to trade off hygroscopicity and stability at the application level. The integration of a hydrophobic evaporation layer within the hygroscopic AMEG device can facilitate the continuous circulation of moisture within the generator, which can lead to better operational stability.⁴⁴

6.5 Further applications

AMEGs are expected to be used directly as power sources and self-powered systems because of their considerable direct-current output.¹²⁵ Due to the limitations of output power density and output sustainability, AMEGs need to be further designed for practical applications. Integration of the AMEG and supercapacitor in the form of tandem has been reported to be able to self-charge to maintain a stable voltage during moisture adsorption, and its integrated supercapacitor can also maintain a high stability under constant self-charging/discharging.¹²⁰ There is also a great potential to directly integrate AMEGs with other water-related technologies beyond energy harvesting, such as water splitting and water treatment, to expand their applications.⁸⁶

The mechanisms of most AMEG devices involve ion-based complex systems (e.g., EDL). As a to-be-developed technology, AMEG devices have the potential to be used as bio-inspired sensing, intelligent iontronic devices in the future, in addition to their applications as power sources. It has been reported that AMEG can be used to detect and perceive the changing humidity field generated by the proximity or removal of living organisms (the proximity of non-living organisms cannot generate a changing humidity field) to achieve non-contact smart sensing.¹¹¹ Meanwhile, some AMEGs with special configurations, such as ionic diode-based generators, could concurrently be applied in logic circuits and field-effect transistors, analogue synapses and data storage.^126–128 Based on the responsiveness of different material components to moisture, the potential for AMEGs to be used for encryption and display of information is promising, while the error in the response of the device to different humidity levels needs to be reduced to improve the accuracy of the encryption.¹¹³

Author contributions

R. L. and K. N. conceived the project. K. N. conducted the literature search and wrote the manuscript. K. N., Q. R., and S. L. discussed and revised the manuscript. Y.-C. L., B. S., X. Z., and R. L. supervised all writing.

Data availability

No primary research results, software or code have been included and no new data were generated or analysed as part of this review.

Conflicts of interest

The authors declare no competing interests.

Acknowledgements

This work was supported by National Natural Science Foundation of China (52103306), Natural Science Foundation of Jiangsu Province (BK20210719), and Natural Science Foundation of Suzhou (ZXL2023181).

References

1. A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J. Am. Chem. Soc., 2009, 131, 6050–6051.
2. R. Liu, S. T. Lee and B. Sun, Adv. Mater., 2014, 26, 6007–6012.
3. J. Park, J. Kim, H.-S. Yun, M. J. Paik, E. Noh, H. J. Mun, M. G. Kim, T. J. Shin and S. I. Seok, Nature, 2023, 616, 724–730.
4. O. Bubnova, Z. U. Khan, A. Malti, S. Braun, M. Fahlman, M. Berggren and X. Crispin, Nat. Mater., 2011, 10, 429–433.
5. Q. Zhang, Y. Sun, W. Xu and D. Zhu, Adv. Mater., 2014, 26, 6829–6851.
6. X.-L. Shi, J. Zou and Z.-G. Chen, Chem. Rev., 2020, 120, 7399–7515.
7. J. Chen, Y. Huang, N. Zhang, H. Zou, R. Liu, C. Tao, X. Fan and Z. L. Wang, Nat. Energy, 2016, 1, 16138.
8. R. Liu, Z. L. Wang, K. Fukuda and T. Someya, Nat. Rev. Mater., 2022, 7, 870–886.
9. B. Shao, M.-H. Lu, T.-C. Wu, W.-C. Peng, T.-Y. Ko, Y.-C. Hsiao, J.-Y. Chen, B. Sun, R. Liu and Y.-C. Lai, Nat. Commum., 2024, 15, 1238.
10. C. R. Bowen, H. A. Kim, P. M. Weaver and S. Dunn, Energy Environ. Sci., 2014, 7, 25–44.
11. Z. L. Wang and J. Song, Science, 2006, 312, 242–246.
12. F. Zhao, H. Cheng, Z. Zhang, L. Jiang and L. Qu, Adv. Mater., 2015, 27, 4351–4357.
13. D. Shen, M. Xiao, G. Zou, L. Liu, W. W. Duley and Y. N. Zhou, Adv. Mater., 2018, 30, 1705925.
14. Y. Huang, H. Cheng, C. Yang, P. Zhang, Q. Liao, H. Yao, G. Shi and L. Qu, Nat. Commum., 2018, 9, 4166.
15. X. Liu, H. Gao, J. E. Ward, X. Liu, B. Yin, T. Fu, J. Chen, D. R. Lovley and J. Yao, Nature, 2020, 578, 550–554.
16. H. Wang, Y. Sun, T. He, Y. Huang, H. Cheng, C. Li, D. Xie, P. Yang, Y. Zhang and L. Qu, Nat. Nanotechnol., 2021, 16, 811–819.
17. Y. Zhang, T. Yang, K. Shang, F. Guo, Y. Shang, S. Chang, L. Cui, X. Lu, Z. Jiang, J. Zhou, C. Fu and Q.-C. He, Nat. Commum., 2022, 13, 3484.
18. S. Yang, X. Tao, W. Chen, J. Mao, H. Luo, S. Lin, L. Zhang and J. Hao, Adv. Mater., 2022, 34, 2200693.
19. T. Xu, X. Ding, H. Cheng, G. Han and L. Qu, Adv. Mater., 2024, 36, 2209661.
20. T. Xu, X. Ding, Y. Huang, C. Shao, L. Song, X. Gao, Z. Zhang and L. Qu, Energy Environ. Sci., 2019, 12, 972–978.
21. H. Lim, M. S. Kim, Y. Cho, J. Ahn, S. Ahn, J. S. Nam, J. Bae, T. G. Yun and I. D. Kim, Adv. Mater., 2024, 36, 2301080.
22. Y. Liang, F. Zhao, Z. Cheng, Y. Deng, Y. Xiao, H. Cheng, P. Zhang, Y. Huang, H. Shao and L. Qu, Energy Environ. Sci., 2018, 11, 1730–1735.
23. Y. Zhang, S. Guo, Z. G. Yu, H. Qu, W. Sun, J. Yang, L. Suresh, X. Zhang, J. J. Koh and S. C. Tan, Adv. Mater., 2022, 34, 2201228.
24. H. Zhang, N. He, B. Wang, B. Ding, B. Jiang, D. Tang and L. Li, Adv. Mater., 2023, 35, 2300398.
25. Z. Zhang, L. Wen and L. Jiang, Nat. Rev. Mater., 2021, 6, 622–639.
26. K. Fan, S. Zhou, L. Xie, S. Jia, L. Zhao, X. Liu, K. Liang, L. Jiang and B. Kong, Adv. Mater., 2024, 36, 2307849.
27. Y. Hu, W. Yang, W. Wei, Z. Sun, B. Wu, K. Li, Y. Li, Q. Zhang, R. Xiao, C. Hou and H. Wang, Sci. Adv., 2024, 10, eadk4620.
28. J. Tan, X. Wang, W. Chu, S. Fang, C. Zheng, M. Xue, X. Wang, T. Hu and W. Guo, Adv. Mater., 2024, 36, 2211165.
29. M. Eigen, L. De Maeyer and J. D. Bernal, Proc. R. Soc. London, Ser. A, 1997, 247, 505–533.
30. N. Agmon, H. J. Bakker, R. K. Campen, R. H. Henchman, P. Pohl, S. Roke, M. Thämer and A. Hassanali, Chem. Rev., 2016, 116, 7642–7672.
31. E. Brodskaya, A. P. Lyubartsev and A. Laaksonen, J. Phys. Chem. B, 2002, 106, 6479–6487.
32. P. Guan, R. Zhu, G. Hu, R. Patterson, F. Chen, C. Liu, S. Zhang, Z. Feng, Y. Jiang, T. Wan, L. Hu, M. Li, Z. Xu, H. Xu, Z. Han and D. Chu, Small, 2022, 18, 2204603.
33. J. Xu, P. Wang, Z. Bai, H. Cheng, R. Wang, L. Qu and T. Li, Nat. Rev. Mater, 2024 DOI:10.1038/s41578-023-00643-0.
34. Q. Li, F. Wang, Y. Zhang, M. Shi, Y. Zhang, H. Yu, S. Liu, J. Li, S. C. Tan and W. Chen, Adv. Mater., 2024, 36, 2209479.
35. X. Nie, B. Ji, N. Chen, Y. Liang, Q. Han and L. Qu, Nano Energy, 2018, 46, 297–304.
36. D. C. Grahame, Chem. Rev., 1947, 41, 441–501.
37. S. L. Carnie and D. Y. C. Chan, Chem. Phys., 1983, 78, 3348.
38. H. A. Stone, A. D. Stroock and A. Ajdari, Annu. Rev. Fluid Mech., 2004, 36, 381–411.
39. Z. Zhang, X. Li, J. Yin, Y. Xu, W. Fei, M. Xue, Q. Wang, J. Zhou and W. Guo, Nat. Nanotechnol., 2018, 13, 1109–1119.
40. J. Yin, J. Zhou, S. Fang and W. Guo, Joule, 2020, 4, 1852–1855.
41. H. Daiguji, Chem. Soc. Rev., 2010, 39, 901–911.
42. D. Shen, W. W. Duley, P. Peng, M. Xiao, J. Feng, L. Liu, G. Zou and Y. N. Zhou, Adv. Mater., 2020, 32, 2003722.
43. S. Fang, J. Li, Y. Xu, C. Shen and W. Guo, Joule, 2022, 6, 690–701.
44. J. Tan, S. Fang, Z. Zhang, J. Yin, L. Li, X. Wang and W. Guo, Nat. Commum., 2022, 13, 3643.
45. X. Lu, T. Yang, C. Fu, Z. Jiang, Y. Zhang, K. Shang, C. He, J. Zhou and Q. C. He, Adv. Energy Mater., 2022, 12, 2202634.
46. M. Shi, Y. n Yang, Y. Han, J. Wang, Y. Wang, D. Li, J. Lv, W. Wu, Z. Wang, X. Wei and N. Chen, Adv. Energy Mater., 2023, 14, 2303815.
47. T. He, H. Wang, B. Lu, T. Guang, C. Yang, Y. Huang, H. Cheng and L. Qu, Joule, 2023, 7, 935–951.
48. M. Shi, Y. Yang, Y. Han, J. Wang, Y. Wang, D. Li, J. Lv, W. Wu, Z. Wang, X. Wei and N. Chen, Adv. Energy Mater., 2024, 14, 2303815.
49. Q. Li, Y. Qin, D. Cheng, M. Cheng, H. Zhao, L. Li, S. Qu, J. Tan and J. Ding, Adv. Funct. Mater., 2023, 33, 2211013.
50. Z. Yan, N. Li, Q. Chang, C. Xue, J. Yang and S. Hu, Chem. Eng. J., 2023, 467, 143443.
51. Y.-M. Cao, Y. Su, M. Zheng, P. Luo, Y.-B. Xue, B.-B. Han, M. Zheng, Z. Wang, L.-S. Liao and M.-P. Zhuo, ACS Nano, 2024, 18, 492–505.
52. K. Zhao, J. W. Lee, Z. G. Yu, W. Jiang, J. W. Oh, G. Kim, H. Han, Y. Kim, K. Lee, S. Lee, H. Kim, T. Kim, C. E. Lee, H. Lee, J. Jang, J. W. Park, Y.-W. Zhang and C. Park, ACS Nano, 2023, 17, 5472–5485.
53. B. Lu, H. Cheng and L. Qu, ACS Nano, 2024, 18, 2730–2749.
54. D. Shen, M. Xiao, Y. Xiao, G. Zou, L. Hu, B. Zhao, L. Liu, W. W. Duley and Y. N. Zhou, ACS Appl. Mater. Interfaces, 2019, 11, 14249–14255.
55. Y. Wu, B. Shao, Z. Song, Y. Li, Y. Zou, X. Chen, J. Di, T. Song, Y. Wang and B. Sun, ACS Appl. Mater. Interfaces, 2022, 14, 19569–19578.
56. Y. Yang, J. Wang, Z. Wang, C. Shao, Y. Han, Y. Wang, X. Liu, X. Sun, L. Wang, Y. Li, Q. Guo, W. Wu, N. Chen and L. Qu, Nanomicro Lett., 2023, 16, 56.
57. Y. Yao, X. Lu, C. Fu, Y. Zhang, J. Fang, J. Qin, Q. C. He and T. Yang, Adv. Funct. Mater., 2024, 34, 2311465.
58. C. Fu, J. Zhou, X. Lu, H. Feng, Y. Zhang, K. Shang, Z. Jiang, Y. Yao, Q. C. He and T. Yang, Adv. Sci., 2024, 11, 2305530.
59. J. Fang, X. Zhang, P. Duan, Z. Jiang, X. Lu, C. Fu, Y. Zhang, Y. Yao, K. Shang, J. Qin, Y. Liu and T. Yang, Mater. Horiz., 2024, 11, 1261–1271.
60. J. Yin, N. Liu, P. Jia, Z. Ren, Q. Zhang, W. Lu, Q. Yao, M. Deng and Y. Gao, SusMat, 2023, 3, 859–876.
61. G. Kim, J. W. Lee, K. Zhao, T. Kim, W. Kim, J. W. Oh, K. Lee, J. Jang, G. Zan, J. W. Park, S. Lee, Y. Kim, W. Jiang, S. Li and C. Park, Energy Environ. Sci., 2024, 17, 134–148.
62. S. H. Sohn, G. J. Choi, B. D. On and I. K. Park, Adv. Energy Mater., 2024, 14, 2304206.
63. M. Li, L. Zong, W. Yang, X. Li, J. You, X. Wu, Z. Li and C. Li, Adv. Funct. Mater., 2019, 29, 1901798.
64. R. Zhu, Y. Zhu, L. Hu, P. Guan, D. Su, S. Zhang, C. Liu, Z. Feng, G. Hu, F. Chen, T. Wan, X. Guan, T. Wu, R. Joshi, M. Li, C. Cazorla, Y. Lu, Z. Han, H. Xu and D. Chu, Energy Environ. Sci., 2023, 16, 2338–2345.
65. H. Wang, H. Cheng, Y. Huang, C. Yang, D. Wang, C. Li and L. Qu, Nano Energy, 2020, 67, 104238.
66. D. Zhang, X. Meng, W. Hou, W. Hu, J. Mo, T. Yang, W. Zhang, Q. Fan, L. Liu, B. Jiang, L. Chu and M. Li, Nano Res. Energy, 2023, 2, e9120050.
67. L. Li, Z. Chen, M. Hao, S. Wang, F. Sun, Z. Zhao and T. Zhang, Nano Lett., 2019, 19, 5544–5552.
68. R. Liu, Y. Lai, S. Li, F. Wu, J. Shao, D. Liu, X. Dong, J. Wang and Z. L. Wang, Nano Energy, 2022, 95, 107056.
69. X. Pan, Y. Jin, Y. Zhou, X. Q. Wang, W. Lu, J. He and G. W. Ho, Adv. Energy Mater., 2023, 13, 2204095.
70. X. Pan, Q. Wang, D. Benetti, L. Jin, Y. Ni and F. Rosei, J. Mater. Chem. A, 2023, 11, 19506–19513.
71. C. Zhang, M. O. G. Nayeem, Z. Wang, X. Pu, C. Dagdeviren, Z. L. Wang, X. Zhang and R. Liu, Prog. Mater. Sci., 2023, 138, 101156.
72. X. Wen, Z. Sun, X. Xie, Q. Zhou, H. Liu, L. Wang, X. Qin and S. C. Tan, Adv. Funct. Mater., 2024, 34, 2311128.
73. X. Zhang, Z. Dai, J. Chen, X. Chen, X. Lin, S. Yang, K. Wu, Q. Fu and H. Deng, Energy Environ. Sci., 2023, 16, 3600–3611.
74. K. Fan, M. Peng, J. Liu, X. Li, X. Wang, B. Kong and X. Liu, Chem. Eng. J., 2023, 461, 142097.
75. B. Shao, Y. Song, Z. Song, Y. Wang, Y. Wang, R. Liu and B. Sun, Adv. Energy Mater., 2023, 13, 2204091.
76. H. Cheng, Y. Huang, F. Zhao, C. Yang, P. Zhang, L. Jiang, G. Shi and L. Qu, Energy Environ. Sci., 2018, 11, 2839–2845.
77. F. Chen, S. Zhang, P. Guan, Y. Xu, T. Wan, C.-H. Lin, M. Li, C. Wang and D. Chu, Small, 2023, 2304572,  DOI:10.1002/smll.202304572.
78. J. Xue, F. Zhao, C. Hu, Y. Zhao, H. Luo, L. Dai and L. Qu, Adv. Funct. Mater., 2016, 26, 8784–8792.
79. J. Liu, Y. Qi, D. Liu, D. Dong, D. Liu and Z. Li, J. Mater. Sci., 2018, 54, 4831–4841.
80. N. Chen, Q. Liu, C. Liu, G. Zhang, J. Jing, C. Shao, Y. Han and L. Qu, Nano Energy, 2019, 65, 104047.
81. D. He, Y. Yang, Y. Zhou, J. Wan, H. Wang, X. Fan, Q. Li and H. Huang, Nano Energy, 2021, 81, 105630.
82. R. Zhang, H. Wang, M. Qu, S. Li, Y. Ma, X. Li, P. Tang and Y. Bin, Chem. Eng. J., 2023, 473, 145325.
83. L. Huang, Y. Zhang, X. Song, D. Li, X. Chen and Q. Yuan, Nano Energy, 2023, 118, 108973.
84. Y. Cao, B. Xu, Z. Li and H. Fu, Adv. Funct. Mater., 2023, 33, 2301420.
85. Y. Huang, H. Cheng, C. Yang, H. Yao, C. Li and L. Qu, Energy Environ. Sci., 2019, 12, 1848–1856.
86. M. Yang, Y. Hu, S. Zheng, Z. Liu, W. Li and F. Yan, Adv. Sci., 2023, 10, 2206071.
87. J. Bai, Y. Huang, H. Wang, T. Guang, Q. Liao, H. Cheng, S. Deng, Q. Li, Z. Shuai and L. Qu, Adv. Mater., 2022, 34, 2103897.
88. A. Nourbakhsh, M. Cantoro, A. Hadipour, T. Vosch, M. H. van der Veen, M. M. Heyns, B. F. Sels and S. De Gendt, Appl. Phys. Lett., 2010, 97, 163101.
89. W. Duan, B. Shao, Z. Wang, K. Ni, S. Liu, X. Yuan, Y. Wang, B. Sun, X. Zhang and R. Liu, Energy Environ. Sci., 2024, 17, 3788–3796.
90. X. Zhang, M. Wang, Y. Wu, X. Chen, K. Wu, Q. Fu and H. Deng, Adv. Funct. Mater., 2023, 33, 2210027.
91. H. Wang, T. He, X. Hao, Y. Huang, H. Yao, F. Liu, H. Cheng and L. Qu, Nat. Commum., 2022, 13, 2524.
92. H. Shen, K. Xu, Y. Duan, P. Wu, Z. Qian, Y. Chen, Y. Luo, C. Liu, Y. Li, J. Cui and D. Liu, Adv. Sci., 2023, 10, 2206483.
93. J. Tan, J. Dou and Q. Tang, Chem. Commun., 2023, 59, 4608–4611.
94. F. Zhao, L. Wang, Y. Zhao, L. Qu and L. Dai, Adv. Mater., 2017, 29, 1604972.
95. J. Bai, Y. Hu, T. Guang, K. Zhu, H. Wang, H. Cheng, F. Liu and L. Qu, Energy Environ. Sci., 2022, 15, 3086–3096.
96. G. Ren, Q. Hu, J. Ye, A. Hu, J. Lü and S. Zhou, Research, 2022, 2022, 9873203.
97. J. Bai, Q. Liao, H. Yao, T. Guang, T. He, H. Cheng and L. Qu, Energy Environ. Sci., 2023, 16, 3088–3097.
98. L. Yang, F. Yang, X. Liu, K. Li, Y. Zhou, Y. Wang, T. Yu, M. Zhong, X. Xu, L. Zhang, W. Shen and D. Wei, Proc. Natl. Acad. Sci. U. S. A., 2021, 118, e2023164118.
99. J. Hao, B. Bao, J. Zhou, Y. Cui, X. Chen, J. Zhou, Y. Zhou and L. Jiang, Adv. Mater., 2022, 34, 2203109.
100. K. Ni, B. Xu, Z. Wang, Q. Ren, W. Gu, B. Sun, R. Liu and X. Zhang, Adv. Mater., 2023, 35, 2305438.
101. J.-C. Feng, S.-X. Li, Z.-P. Zhang, Y. An, Q.-S. Gao, Z. Sun and H. Xia, Nano Energy, 2024, 119, 109103.
102. T. Cai, L. Lan, B. Peng, C. Zhang, S. Dai, C. Zhang, J. Ping and Y. Ying, Nano Lett., 2022, 22, 6476–6483.
103. Q. Ma, Y. Li, R. Wang, H. Xu, Q. Du, P. Gao and F. Xia, Nat. Commum., 2021, 12, 1573.
104. C. Wang, E. Seo and J. Park, Biomicrofluidics, 2024, 18, 011503.
105. D. Maity and M. Fussenegger, Adv. Sci., 2023, 10, 2300750.
106. G. Liu, Z. An, Y. Lu, Y. Wu, Z. Shi, X. Li, J. Lv, H. Wen, Z. Peng, R. P. S. Han, Y. He, Q. Ye, Q. Chen, F. Zhang, J. Liu and Q. Liu, Nano Energy, 2024, 119, 109098.
107. W. Yang, L. Lv, X. Li, X. Han, M. Li and C. Li, ACS Nano, 2020, 14, 10600–10607.
108. W. Lu, T. Ding, X. Wang, C. Zhang, T. Li, K. Zeng and G. W. Ho, Nano Energy, 2022, 104, 107892.
109. J. Yang, R. Shi, Z. Lou, R. Chai, K. Jiang and G. Shen, Small, 2019, 15, 1902801.
110. L. Lu, C. Jiang, G. Hu, J. Liu and B. Yang, Adv. Mater., 2021, 33, 2100218.
111. Y. Zhang, D. Long, H. Feng, K. Shang, X. Lu, C. Fu, Z. Jiang, J. Fang, Y. Yao, Q.-C. He and T. Yang, Biosens. Bioelectron., 2024, 247, 115922.
112. C. Yang, H. Wang, J. Yang, H. Yao, T. He, J. Bai, T. Guang, H. Cheng, J. Yan and L. Qu, Adv. Mater., 2022, 34, 2205249.
113. Y. Zhang, Z. Yu, H. Qu, S. Guo, J. Yang, S. Zhang, L. Yang, S. Cheng, J. Wang and S. C. Tan, Adv. Mater., 2024, 36, 2208081.
114. Y. Liang, F. Zhao, Z. Cheng, Q. Zhou, H. Shao, L. Jiang and L. Qu, Nano Energy, 2017, 32, 329–335.
115. C. Tan, Z. Liu, W. Huang and H. Zhang, Chem. Soc. Rev., 2015, 44, 2615–2628.
116. T. G. Yun, J. Bae, A. Rothschild and I.-D. Kim, ACS Nano, 2019, 13, 12703–12709.
117. Y. Tao, Z. Wang, H. Xu, W. Ding, X. Zhao, Y. Lin and Y. Liu, Nano Energy, 2020, 71, 104628.
118. R. Liu, M. Takakuwa, A. Li, D. Inoue, D. Hashizume, K. Yu, S. Umezu, K. Fukuda and T. Someya, Adv. Energy Mater., 2020, 10, 2000523.
119. Y. Han, B. Lu, C. Shao, T. Xu, Q. Liu, Y. Liang, X. Jin, J. Gao and Z. Zhang, Chem. Eng. J., 2020, 384, 123366.
120. L. Wang, H. Wang, C. Wu, J. Bai, T. He, Y. Li, H. Cheng and L. Qu, Nat. Commum., 2024, 15, 4929.
121. Y. Huang, K. Zhou, H. Cheng, T. He, H. Wang, J. Bai, C. Yang, T. Guang, H. Yao, F. Li, G. Hou, Z. Xu and L. Qu, Adv. Funct. Mater., 2024, 34, 2308620.
122. Y. Wang, X. Yuan, K. Ni, Y. Song, X. Li, X. Zeng, B. Shao and B. Sun, Energy Environ. Sci., 2024, 17, 4780–4793.
123. X. Liu, D. Lyu, C. Merlet, M. J. A. Leesmith, X. Hua, Z. Xu, C. P. Grey and A. C. Forse, Science, 2024, 384, 321–325.
124. E. Pastor, Z. Lian, L. Xia, D. Ecija, J. R. Galán-Mascarós, S. Barja, S. Giménez, J. Arbiol, N. López and F. P. García de Arquer, Nat. Rev. Chem., 2024, 8, 159–178.
125. X. Wang, F. Lin, X. Wang, S. Fang, J. Tan, W. Chu, R. Rong, J. Yin, Z. Zhang, Y. Liu and W. Guo, Chem. Soc. Rev., 2022, 51, 4902–4927.
126. H. J. Kim, B. Chen, Z. Suo and R. C. Hayward, Science, 2020, 367, 773–776.
127. H. Ma, J. Liang, J. Qiu, L. Jiang, L. Ma, H. Sheng, M. Shao, Q. Wang, F. Li, Y. Fu, J. Wang, E. Xie, Y. Chai and W. Lan, Adv. Mater., 2023, 35, 2301218.
128. S. H. Han, S. I. Kim, M.-A. Oh and T. D. Chung, Proc. Natl. Acad. Sci. U. S. A., 2023, 120, e2211442120.

---

Footnote

† These authors contributed equally to this work.

---