Surface Wettability for Skin-Interfaced Sensors and Devices

Abstract

The practical applications of skin-interfaced sensors and devices in daily life hinge on the rational design of surface wettability to maintain device integrity and achieve improved sensing performance under complex hydrated conditions. Various bio-inspired strategies have been implemented to engineer desired surface wettability for varying hydrated conditions. Although the bodily fluids can negatively affect the device performance, they also provide a rich reservoir of health-relevant information and sustained energy for next-generation stretchable self-powered devices. As a result, the design and manipulation of the surface wettability are critical to effectively control the liquid behavior on the device surface for enhanced performance. The sensors and devices with engineered surface wettability can collect and analyze health biomarkers while being minimally affected by bodily fluids or ambient humid environments. The energy harvesters also benefit from surface wettability design to achieve enhanced performance for powering on-body electronics. In this review, we first summarize the commonly used approaches to tune the surface wettability for target applications toward stretchable self-powered devices. By considering the existing challenges, we also discuss the opportunities as a small fraction of potential future developments, which can lead to a new class of skin-interfaced devices for use in digital health and personalized medicine.

Graphical Abstract

1. Introduction

Compared with bulky diagnostic instruments or wrist-mounted wearables, the thin, soft, conformal sensors and devices ^[1-10] can reliably interface with human skin ^[11-14] to provide bio-integrated devices for health monitoring and disease diagnosis ^[15-17]. However, the practical applications of these skin-interfaced devices in long-term use have to consider device integrity and sensing performance variations ^[18] in varying hydrated conditions, including those from the ambient environment (outside) and the human body (inside) (Figure 1, top). The former includes raining ^[19-21], humid weather ^[22-24], swimming ^{[25, 26]}, and washing ^{[27, 28]}, whereas the latter comes from sweating (sensible and insensible) ^[29-34], skin lipids secreting ^[35-37], breathing ^{[38, 39]}, and wound exuding ^[40-42]. The device performance degradation and failure may result from the penetration of water molecules into the devices ^[43-46] or the contamination of bacteria/viruses in the droplets on the device surface ^[47-50]. Moreover, the accumulated sweat at the device/skin interface after long-term use reduces adhesion strength ^[51] and causes skin inflammation and breakdown ^[52-54]. Regardless of these issues, the bodily fluids (e.g., sweat, saliva, and tears) provide the sensors with a rich reservoir of biomarkers for noninvasive measurements ^[55-59]. Besides the (triboelectric) nanogenerators powered by the water/moisture or chemicals in sweat, the motion of raindrops and humidity difference can also provide renewable mechanical or chemical energies for energy harvesting to power sensors and devices ^[60-63].

Figure 1.

Schematic summarizing the considerations and applications of skin-interfaced sensors and energy harvesters with suitable surface wettability in hydrated conditions from the ambient environment to the human body. The surface wettability design provides the device with waterproof ^[27], self-clean ^[67], moisture management ^[68], sweat collection ^[69], direct sweat transport ^[70], and electrochemical sweat analysis ^[71], as well as droplet- ^[19] and moisture-enabled energy harvesters ^[72]. The representative conditions include wound exude ^[73], interstitial fluid ^[74], sweat ^[63], lipid ^[35], drinking ^[75], washing ^[43], raining ^[76], and swimming ^[25].

As the surface wettability of the devices easily yet efficiently modulates their interaction with liquids ^[64-66], the engineered surface wettability can prepare the devices for the target use in varying hydrated conditions (Figure 1, middle). The hydrophobic surface with water repellence can spontaneously and rapidly spread and roll off liquids, whereas the hydrophilic surface with excellent water affinity promotes surface wetting and liquid wicking. This review will first provide an overview of the commonly used approaches on surface modification, microstructure creation, and wettability patterning to engineer the wettability on the device surface. The surface wettability design can provide skin-interfaced sensors with unique functionalities (Figure 1, bottom), including waterproof, self-clean, moisture management, sweat collection, sweat transport, and electrochemical sweat analysis. These sensors can be potentially integrated with droplet- or moisture-enabled energy harvesting units to result in a self-powered sensing platform. We will highlight the advantages and applications of the resulting devices with engineered surface wettability. Finally, we conclude this review with a brief discussion of the existing challenges and opportunities for future research development in this burgeoning field.

2. General surface wettability

The device surface with varying wettability exhibits different contact angles (CAs) to the liquids, with CA > 90° (or < 90°) for low (or high) surface energy as a hydrophobic (or hydrophilic) state ^{[77, 78]}. The smooth surface with an intrinsic hydrophobic or hydrophilic property can be adjusted by the surface functional groups. The bio-inspiration from many living organisms (e.g., lotus leaves, rose petals, pitcher plants, and Stenocara beetles) ^[79-81] promotes the development of micro-/nano- hierarchical structures for superwetting surfaces (e.g., CA of > 150° for superhydrophobic or < 10° for superhydrophilic states). The hydrophobicity and hydrophilicity resulted from micro-/nano- structures and chemical surface modifications can also be combined to provide patterned surface wettability for patterning and controlling liquids. These commonly used approaches are summarized in Table 1.

Table 1.

The common methods to modulate the surface wettability.

Methods	Advantages	Disadvantages	Ref.
Plasma treatment	Simple, fast, and easily tuned	Needed special equipment and hydrophobic recovery	[82]
UV irradiation	Simple, low cost, and more effective	Surface easily degraded and instability	[83]
Chemical modified	Stable and durable	Prevent subsequent process	[84]
Spray coating	Simple, scalable, and applicable for a variety of substrates and not limited to small areas	Hard to tune the surface morphology and thickness, waste of materials	[85]
Dip coating	Simple, inexpensive process and applicable for complex shapes, inexpensive equipment	Time-consuming, have to fully disperse in organic solution	[86]
Electrospinning	Versatile, effective, and applied to a variety of materials	Needed organic solution and special equipment	[68]
Photolithography	Scalable, ultrahigh-resolution	Needed photomasks, clean environment, and expensive equipment	[87]
Soft lithography	Convenient, effective, low-cost, and without expensive and complex equipment	Have to combine with other techniques	[88]
Laser direct writing	Negligible heat-affected zone, no-contact process, precise ablation threshold, and high resolution.	Time-consuming, limited process ranges, expensive equipment	[89]
Chemical etching	Simple, low cost, and scalable	Hard to turn the surface morphology	[90]
Plasma etching	High resolution and fidelity.	Expensive and complex equipment	[91]
Wax printing	Simple, low-cost fabrication, rapidity, robustness, and lack of organic solvent consumption	Low resolution, temperature sensitivity	[70]
Stamp	Simple, low-cost, and mass production	Low resolution and require various-shaped stamps	[92]

2.1. Surface Modification

Due to the inherent property of materials, every material has its specific wettability. Therefore, choosing materials with suitable wettability would effectively simplify the post-processing steps. In the device systems, functional device components with active materials often reside on a soft elastomeric substrate to match the mechanical properties of the skin. Due to the surface methyl group (-CH₃), the commonly used substrates are hydrophobic, including silicone rubber such as Ecoflex ^[93] and polydimethylsiloxane (PDMS) ^[94], poly (styrene-isoprene-styrene) (SIS) ^[25], and polystyrene-block-poly(ethylene-ran-butylene)-block-polystyrene (SEBS) ^[75]. In contrast, the polymers/plastics with hydrophilic functional groups (e.g., carboxyl or amino) exhibit hydrophilicity, including polyethylene terephthalate (PET) ^[95], polyimide (PI) ^[96], polyurethane (PU) ^[97]. Moreover, the porous substrate materials (e.g., paper ^[98], textile ^[99], and silk ^[100]) can even provide active liquid wicking.

The modulation of surface free energy can change the intrinsic wettability (Table 1). The low-surface-tension reagents with the fluorine atom (e.g., fluroroalkanes, fluoropolymers, and fluorosurfactants) decrease the surface tension and lower the surface energy to achieve hydrophobicity. In comparison, the reagents such as silane coupling agents ended with the hydrophilic groups (e.g., -COOH, -OH, -NH₂) can enhance the hydrophilicity of the surfaces ^[101]. Therefore, surface modification with self-assembled functional groups can be used to tune the surface wettability and the process relies on molecular interactions such as intermolecular forces or electrostatic interactions (Figure 2a). Meanwhile, many surface treatment strategies such as plasma, ultraviolet/ozone (UVO), and Tesla coil have also been explored to increase the number of hydroxyl groups for increased wettability (Figure 2b). As one of the commonly explored surface modification methods, the plasma treatment can promote wettability in both directions: hydrophilic with O₂ plasma but hydrophobic with fluorocarbon plasma ^[102]. UVO treatment can also generate reactive atomic oxygen molecules for oxidizing the substrate and eliminating hydrocarbon-containing organic moieties ^[103]. Compared with O₂ plasma treatment, the UVO treated PDMS surface shows a much more stable performance, with lower hydrophobic recovery and almost unchanged contact angle months after treatment ^[83]. Notably, bringing the plasma or UVO activated surfaces can form covalent binding at the interface through siloxane bonds, which has been widely used for device integration and assembly ^{[104, 105]}.

Figure 2.

Common approaches to adjust the surface wettability of materials. Surface modification: (a) Schematic illustration of the polymer substrates modification by chemical agents. (b) Schematic illustration of the polymer substrates modification by oxygen, nitrogen, fluorocarbon plasma, and UVO treatment. Micro-/nanostructure construction: Schematic of the dip-coating (c) ^[86], spray-coating (d) ^[75], and electrospinning method (e) ^[106] to construct random micro-/nanostructure, and soft lithography (f) ^[87] to construct well-defined micro-/nanostructure. Wettability patterned: Schematic of the mask-based modification (g) to construct hydrophilic region in the hydrophobic substrate ^[111].

2.2. Creation of Micro-/nano- Structures

For a surface with either hydrophilic or hydrophobic properties, the increase of the surface roughness enhances its hydrophilicity or hydrophobicity. Therefore, the modulation in the surface energy is often combined with the creation of random or well-defined micro-/nano- structures to realize superwettability (Table 1). The random roughness can be easily created by simple, low cost, and scalable fabrication approaches, including spray-coating, dip-coating, electrospinning, chemical etching, electrochemical deposition ^[79]. For instance, direct immersing the substrates into the nanomaterial solution in dip-coating can simply provide the integration of nanomaterials even on complex surfaces (Figure 2c) ^[86]. Although spray-coating can spray the solution of nanomaterials over the substrate with potential for large-scale industrial production, it may not be directly applied for the complex 3D porous structure (Figure 2d) ^[75]. In contrast, the electrospinning process assembles nanofibers to fabricate the textile with micro-/nano-structures, which can also be combined with other surface modification methods to integrate nanomaterials for increased roughness and functionalization (Figure 2e) ^[106]. It is also interesting to note that besides the changed surface chemical constitution, prolonged plasma treatment with high power can create surface etching of polymeric materials for enhanced surface roughness ^{[66, 107]}.

Different from the random roughness, the generation of well-defined surface structures over a large area relies on the use of relatively expensive and complicated processing steps such as photolithography and soft lithography ^{[87, 88]}, reactive ion etching (RIE) ^[91], or direct laser writing ^{[89, 108]}. As one of the most commonly used methods, photolithography can directly and efficiently produce well-defined micro-/nanostructures on the substrate, which can also be replicated to the soft substrate as templates for use in soft lithography (Figure 2f) ^[87]. Because the removal of the templates may damage the complicated micro-/nanostructures and even the templates themselves, these techniques are often explored for creating relatively simple and regular patterns. In comparison, the recently emerged direct laser writing is attractive in precision manufacturing. Because of the extremely short pulse width and ultrahigh peak intensity, direct laser writing can create well-defined micro-/nanostructures on diverse material surfaces in a programmable manner ^[109].

2.3. Wettability Patterning

The patterned surface wettability can be created by mask-based modification or direct printing (e.g., constructing hydrophobic barriers in hydrophilic substrates) (Table 1) ^[110]. In the former, the O₂ plasma or UV irradiation through the opening of the mask easily induces hydrophilic region in hydrophobic substrates by changing the surface functional groups or surface morphologies (Figure 2g) ^[111]. As for direct printing, it can precisely print the hydrophobic materials onto the superhydrophilic substrate. Compared to screen printing, flexographic printing, and inkjet printing, wax printing and stamping are frequently used as they are low-cost, simple, rapid, and solvent-free ^[112]. However, the patterns are associated with a low resolution due to ink penetration ^[113]. Efforts to address the challenges lead to the development of a mask-free laser scribing strategy to directly create wettability contrast on-demand by tuning both the surface morphology and chemical composition ^[114-116]. For example, the superwettability patterns with superhydrophilic porous graphene and superhydrophobic layered graphite on the PI substrate are prepared by laser scribing the target region of the substrate at low or high power, respectively ^[114]. Moreover, laser-induced heating can remove the low-surface-energy reagent to result in superhydrophilic regions on the superhydrophobic substrate ^[115].

3. Application of Surface Wettability in Skin-interfaced Sensors and Devices

The surface wettability design can provide the skin-interfaced sensors and devices with unique functions that are not easily or readily available with other approaches. The device performance can also be significantly enhanced compared with their counterparts without the designed surface wettability. As a result, this section will briefly summarize the exploration of surface wettability design in skin-interfaced sensors and devices for emerging functions and applications.

3.1. Waterproof

Because the skin-interfaced sensors and devices need to be waterproof in varying hydrated conditions to avoid performance degradation and electrical leakage to the body ^{[46, 117]}, various strategies from polymer encapsulation ^{[118, 119]} to superhydrophobic coating have been explored (Figure 3a). The superhydrophobic coating can explore either heterogeneous or homogeneous material integration. As the most commonly used hydrophobic encapsulation layer, PDMS can encapsulate flexible optoelectronics systems that consist of photodetectors and proximity sensors on the fingertip to exhibit invariant performance in saline solution for 3 h or 1000 cycles of immersion (Figure 3b) ^[28]. However, the insufficient hermeticity of PDMS to water vapor (e.g., ~ 708.1 g/m²/day even with a thickness of 92 μm) ^[120] has led to the use of polymeric materials with lower permeability, including PI ^[121], fluorinated ethylene propylene (FEP) ^[122], and SIS ^[25]. For example, the 125 μm-thick SIS layer has a water vapor permeability of 37 g/m²/day, which provides a much lower water mass change (< 20% for 4 h) in soft microfluidic devices than that with PDMS (100% within 3 h) at 37 °C for precise biomarkers analysis ^[25]. The 50 μm-thick FEP with an even low water vapor permeability of 2.04 g/m²/day ensures the high underwater performance for the crack-based strain sensor (i.e., > 93 % or 72 % of initial sensitivity on the 6th or 18th day) ^[122].

Figure 3.

Common strategies to explore surface wettability for waterproof. (a) Schematic illustrating polymer encapsulation (I), heterogeneous (II) and homogeneous (III) superhydrophobic coating. (b) Hydrophobic PDMS encapsulation on inorganic photoelectronic devices ^[28]. (c) Groove-shape micro-/nanostructured superhydrophobic PDMS encapsulation ^[88]. (d) Superhydrophobic surface with inorganic hydrophobic nanoparticles on the strain sensor ^[125]. (e) Robust superhydrophobic surface by polymer enhancement strategy on the pressure sensor ^[27]. (f) Intrinsically hydrophobic polymer-based strain sensors for underwater motion monitoring ^[75]. (g) Humidity influence of intrinsically superhydrophobic textile-based TENGs ^[85].

However, the hydrophobic polymeric encapsulation with relatively high adhesion (compared to superhydrophobic surfaces) to the water droplets may retain them on the surface for increased risk of water penetration ^[123]. To achieve superhydrophobicity for the surface, the micro-/nano-structure has been introduced to reduce the contact area and increase the contact angle. The wearable organic solar cells encapsulated by a groove-shaped superhydrophobic PDMS film exhibit a high photoelectric conversion efficiency of 85% compared to those of 70% with a flat PDMS encapsulation in a water flush test (8 h, 100 mL/min) for extended operation (Figure 3c) ^[88]. The superhydrophobic coating directly formed on the surface of devices can also be thinner compared to the common polymeric encapsulation strategies.

The simple fabrication of the superhydrophobic surfaces involves dip-coating or spraying-coating of inorganic hydrophobic nanoparticles such as fumed silica nanoparticles (Hf-SiO₂) on the device surface ^[123-125]. For example, the paper-based strain sensor spray-coated with the superhydrophobic Hf-SiO₂ exhibits strong waterproof properties (CA≈164°) to monitor the human motions in harsh water conditions or even detect subtle underwater vibrations (Figure 3d) ^[125]. However, the weak Van der Waals interactions at the nanoparticles/device interface are fragile and easily broken upon external abrasion, resulting in the degeneration of superhydrophobicity ^[126]. Based on a polymer enhancement strategy, the polymer polypyrrole-polydopamine-perfluorodecyltrlethoxysilane (PPy-PDA-PFDS) on the superhydrophobic coating enhances its mechanical stability (Figure 3e) ^[27]. The carbon nanotubes (CNTs) encapsulated by PPy-PDA-PFDS can tightly anchor onto the fibers of the e-textile substrate to prevent the damage by the external abrasion (maintaining CA_water ≈ 165° under machine washing for 3 times or tape-peeling for 60 times). Moreover, a high fluorination degree in PPy-PDA-PFDS reduces surface energy to provide the e-textile with a superhydrophobic property against both water and oil (CA_oil ≈ 160.3°). The resulting wearable e-textile can detect the pulse from the human wrist under both dry and sweaty conditions, as well as finger bending under the water.

As the superhydrophobic encapsulation inevitably changes the mechanical and thermal properties of the wearable sensor to affect its sensitivity ^{[121, 122, 127, 128]}, it is highly desirable to explore the intrinsically superhydrophobic sensors ^{[75, 85, 117, 129]}. For example, the composite of stretchable thermoplastic elastomer (TPE) and conductive multi-walled carbon nanotubes (MWCNTs) treated with ethanol provides the resulting strain sensor with enhanced mechanical stability and robust superhydrophobicity (Figure 3f) ^[75]. The spray-coated ultrathin sensor (1 μm) can easily integrate onto the latex glove for full-range and real-time detection of finger motions, with resistance against water and corrosive (acidic or alkaline) environments. Although the superhydrophobic surface as a waterproof layer is effective, its performance largely degrades in a high humid environment ^[64]. The textile-based triboelectric nanogenerators (TENGs) with spray-coated superhydrophobic CNT/TPE can maintain 80% voltage output after 1 h exercise, but the value decreases rapidly to 22.3 % in a high humid condition (RH ≈ 85 %, though still better than the pristine textile of almost zero) (Figure 3g) ^[85]. This performance degradation results from the absorption, nucleation, and growth of the small water molecules on the superhydrophobic surface to affect the transport of electric charges ^[64].

3.2. Self-clean

Besides liquids, various contaminations (e.g., microorganisms, bacterial, and particles) can also affect device performance. The superhydrophobic surface with low adhesion and strong water repellency can leverage the rolling of the water droplets to remove particles from the surface for self-cleaning) ^{[27, 117, 124, 130-133]}. Though washing can partially remove certain contaminations, the PDMS film placed in a dusty environment still shows dramatically decreased transmissivity (Figure 4b I) ^[88] the contaminated PDMS interlayer in the TENG results in a decrease in electrical output to 51% of its initial value (Figure 4b II) ^[87]. Compared to the common PDMS, cleaning the contaminated superhydrophobic PDMS leads to a higher transmissivity (Figure 4b III), and the one in the TENG allows the device to achieve 88% of its initial electrical output (Figure 4b IV). It should be noted that self-cleaning from the superhydrophobic surface does not work well for particles that are smaller than the surface micro-/nano- structures ^[134]. For the organic contaminants that are challenging to remove by water, it is possible to clean by using photocatalytic materials (e.g., TiO₂, CuO, ZnO) with superhydrophilic surfaces, as the photocatalytic reaction can produce superoxide anions and hydroxyl radicals intermediates to decompose them in water (Figure 4c) ^[135]. Decorated with hydrophilic TiO₂ nanoparticles, the all-nanofiber-based TENGs exhibit an excellent photocatalytic self-cleaning effect to completely degrade the surface pollution of methylene blue dye (Figure 4d) ^[49]. As a result, the decreased current output recovers from 70% to 91% under solar lighting for 25 min. As the hydrophilic self-cleaning functional layer may also result in the penetration of water and sweat into the electrode, an additional hydrophobic polytetrafluoroethylene (PTFE) layer is sputtered on the film to ensure stable performance.

Figure 4.

Self-cleaning in the superhydrophobic surface. Schematics illustrating the superhydrophobic and self-cleaning surface for anti-contamination against (a) particulates and (c) organic pollutants, and (e) anti-biofouling. (b) Optical images of (I) hydrophobic and (III) superhydrophobic PDMS in a dusty environment, ^[88] and output voltages of (II) hydrophobic and (IV) superhydrophobic PDMS interlayer in TENGs in a dusty environment ^[87]. (d) Optical images to show the performance change among original, polluted, and photo catalytically self-cleaned TENGs ^[49]. (f) Superhydrophobic/oleophobic fabric shows high resistance against aerosol filtration and protein adsorption ^[141]. (g) The superhydrophobic laser-induced functional graphene mask with photothermal property for self-sterilization ^[149]. (h) Superomniphobic textile-based TENGs with anti-bacterial properties ^[50]. (i) The combination of the superhydrophobic surface with antibacterial Ag NPs ^[48].

Biofouling from the spontaneous accumulation of macromolecules or microorganisms (e.g., proteins, cells, bacteria) on the device surface often results in degenerated performance and increased safety risk (e.g., inflammation and infection) ^{[48, 73, 136-138]}. For example, the respiratory droplets containing the SARS-CoV-2 virus from infected individuals could adhere to the device surface to increase the infection risk ^[139]. Therefore, it is desirable to explore anti-biofouling of the superhydrophobic self-cleaning surfaces that can significantly reduce the contact area between the microorganisms-containing water droplet and device surfaces (Figure 4e) ^[140]. The polyurethane-coated fabric is functionalized with a perfluoro-tert-butanol-hexamethylene diisocyanate (PFtB-HDI) with superhydrophobic/oleophobic properties shows excellent filtration efficiency of 92.66% for aerosols containing 100 nm polystyrene particles (similar in size to SARS-CoV-2) (49.37% for the one without coating) (Figure 4f) ^[141]. Moreover, the resulting superhydrophobic fabric also shows a 5-fold increase in resistance against protein adsorption such as S-protein in virus infection compared to that of the bare fabric. The minimized water droplet wetting and aerosol penetration of the superhydrophobic fabric show great potential as masks to prevent virus infection. It should be noted that a small fraction of the virus could still attach to the superhydrophobic surface ^[142]. Therefore, strategies have been explored to sterilize the superhydrophobic surface, including antiviral agents ^[143], plasmonic heating ^[144], photosensitizers ^[145], and photocatalytic ^[67], photothermal ^{[144, 146]}, electrothermal ^[147], and antibacterial metal nanoparticles ^[148]. The superhydrophobic laser-induced functionalized graphene mask with photothermal properties can heat up to 80°C under sunlight illumination to achieve self-sterilization (Figure 4g) ^[149]. Besides viruses, bacterial growth is also undesirable, as it could cause bacterial infections ^{[47, 48]}. The antibacterial property of the superomniphobic materials coated on the surface of textile-based TENGs allows them to repel and prevent the proliferation of S. aureus (Figure 4h) ^[50]. Similar to the concern over the attached virus, it is also desirable to kill the already attached bacteria on the superhydrophobic surface. By introducing the antibacterial silver nanoparticles (Ag NPs), the superhydrophobic strain sensor shows increased antibacterial and anti-adhesion rates of 86.73 and 99.91% against S. aureus compared with the one without Ag NPs (14.53% and 76.63%) (Figure 4i) ^[48]. The enhanced anti-adhesion property comes from the increased surface roughness contributed by Ag NPs. The antibacterial actions of Ag NPs on the cell wall and membrane involve Ag NPs and Ag ions-induced oxidative stress to damage the intracellular biomolecules and structures ^[150].

3.3. Moisture Management

Effective moisture management through breathable design or direct sweat transport prevents sweat accumulation at the skin/device interface to ensure adhesion for robust device performance ^[151-153] and minimize the risk for inflammation and infection ^{[47, 154]}. The breathable design adjusts the thermal-moisture balance and achieves gas exchange between the human body and the ambient environment ^[155-157]. As a result, the porous structure from the sponge, textile, and nano-network has been exploited to enhance permeability ^{[44, 52, 53, 158-161]}. The porous graphene electrodes mounted onto an elastomeric sponge substrate exhibit a high water vapor permeability of 18 mg cm⁻² h^{−1 [158]}. The water-wicking effect of the hydrophilic porous sponge also facilitates sweat transport and water evaporation to minimize the risk of inflammation. However, these porous devices are easily affected by contamination, especially in hydrated environments (Figure 5a I) ^{[155, 162]}. Therefore, it is natural to combine the porous structure with the superhydrophobic surface to enable waterproofing and self-cleaning properties in the porous structure (Figure 5a II). Integrating the omniphobic (hydrophobic and oleophobic) coating into the textile-based TENGs results in a self-power tough/gesture sensor with waterproof (CA_water ≈ 155°) and breathable (air permeability around 90.5 mm s⁻¹) properties (Figure 5b) ^[50]. Similarly, electrospinning of polyvinylidene fluoride (PVDF), carbon, and PU nanofibers as the sensing, electrode, and substrate layers in the all-fiber structured e-skin exhibits water vapor permeability of 10.26 kg m⁻² day⁻¹ and high hydrophobicity (CA_PVDF ≈ 134°, CA_PU ≈ 147°) (Figure 5c) ^[163]. The triboelectric effect of e-skin allows it to be used for pressure sensing (sensitivity of 0.18 V kPa⁻¹ in the range of 0 - 175 kPa), energy harvesting, and motion positioning.

Figure 5.

Moisture management in the fabric with varying surface wettability. (a) Schematic illustrating the common strategies for moisture management: hydrophilic fabric, superhydrophobic fabric, and hydrophobic/superhydrophilic Janus fabric. (b) The e-textile with omniphobic coating shows waterproof and breathable properties for motion energy harvesting ^[50]. (c) Electrospun all fabric superhydrophobic e-skin with waterproof and breathable properties for self-powered human motion monitoring ^[163]. (d) Hydrophobic/superhydrophilic Janus textile with asymmetric hydrophilic conical micropores for directional sweat transport ^[166]. (e) The integration of the Janus textile with an electrochemical sensor array for spontaneous sweat sampling and simultaneous analysis ^[167]. (f) The Janus textile helps drain wound exudate and provides anti-adhesion against bacteria for lowered infection risk and accelerated wound healing ^[168].

As passive evaporation is less effective for sweat transport from the skin to the outside than active transport, developing directional sweat transport based on surface wettability can reduce sweat retention on the skin to improve the level of comfort (Figure 5a III) ^{[164, 165]}. As one representative example, the Janus polyester/nitrocellulose textile with asymmetric properties on its two sides (hydrophobic inner or skin side; hydrophilic outer side, generates a wettability or pressure gradient to drive directional liquid motion across it (Figure 5d) ^[166]. The Janus textile can not only pump the sweat through the hydrophobic layer to wet the superhydrophilic layer, but also block the external water droplets by spreading them on the superhydrophilic layer. The conical micropores design also minimizes the hydrostatic pressure and maximizes the capillary force to facilitate sweat transport, drying the wet skin within seconds. However, the superhydrophilic outer side is easily saturated to slow down the sweat transport rate. Efforts to address this challenge lead to the development of a superhydrophobic textile with wettability gradient microchannels across the thickness direction prepared by the selective plasma treatment ^[82]. The continuous sweat transport driven by capillary force in the wettability gradient regions would roll off sweat from the fabric to maintain excellent performance and exhibit an ultrahigh water transmission rate up to 159.84 kg m⁻² day⁻¹.

By combining directional sweat transport Janus textile with an electrochemical sensor array, the integrated smart Janus textile band also leverages the high-efficiency sweat transport for unidirectional and spontaneous sweat sampling to detect biomarkers in biofluids (Figure 5e) ^[167]. Besides sweat, the other biofluids such as wound exudate can also be effectively managed by the Janus textile consisting of hydrophobic nanofiber array and hydrophilic microfiber networks (Figure 5f) ^[168]. Compare with the conventional dressing of medical gauze, the self-pumping Janus textile dressing effectively drains excessive biofluid at the wound site to lower the risk of wound infection, while promoting re-epithelialization, collagen formation, and wound closure. The amphiphilic nanofibrous wound dressing with Janus superhydrophilic/superhydrophobic feature also exhibits anti-adhesion against bacteria to lower the risk of wound infection ^[41]. Obviating the need for dressing change can also avoid secondary injuries.

3.4. Sweat Collection

The accurate and reliable of sweat analysis relies on efficient sweat collection with minimized sweat contamination or evaporation. Regardless of the use of various materials and epidermal microfluidic devices for sweat sampling, surface wettability can effectively modulate the capillary force for improved performance. For epidermal microfluidic devices, the sweat flow in the microchannels (Figure 6a) is driven by the sweat secretion pressure from the sweat gland and the capillary pressure (P_C) of microchannels that are related to their wettability through the Young-Laplace equation. For a rectangular microchannel with height h and width w, the capillary pressure is related to and the contact angle θ_i (i = top, bottom, left, and right) of microchannel walls ^{[169, 170]}:

$$P_C=-\gamma \left[\frac{\cos {\theta }_t+\cos {\theta }_b}{h}+\frac{\cos {\theta }_l+\cos {\theta }_r}{w}\right],$$ (1)

where γ is the surface tension of the liquid in the microchannel. Equation (1) implies that a negative capillary pressure spontaneously wicks sweat into the hydrophilic microchannel (Figure 6a I), whereas the positive capillary pressure in the hydrophobic microchannel needs to be smaller than the sweat secretion pressure (Figure 6a II) that is generated by the concentration difference (e.g., Na, Cl) between sweat and plasma ^{[171, 172]}. The spontaneous wicking effect in the hydrophilic microchannels can easily collect the sweat without external driven force (Figure 6b) ^[173]. The sweat collection rate of 5.0 μL/min in the hydrophilic microchannels is larger than the typical sweat rate of 2.4 μL/min for a 10 mm-diameter collection area ^[95]. The high-speed sweat collection over a large area can also avoid mixing-induced contamination for improved measurement accuracy ^[172]. In the hydrophobic microchannels, their dimensions need to be carefully designed to ensure a capillary pressure smaller than the sweat secretion pressure (~70 kPa) for effective collection. Varying the channel dimension to change their capillary pressure until the sweat collection occurs can also measure sweat secretion pressure (Figure 6c) ^[171]. Moreover, the use of the sweat secretion pressure requires tight coupling of the microfluidic devices to the skin through strong adhesive materials, which are challenging to remove ^{[151, 152, 174, 175]}. Though plasma treatment and UV irradiation can easily transform hydrophobic surfaces into hydrophilic ones ^{[176, 177]}, hydrophobic recovery often occurs within hours ^[178], leading to the need for further chemical modifications to retain long-term hydrophilicity ^[179].

Figure 6.

Sweat collection modulated by surface wettability in wearable devices. (a) Schematics illustrating the liquid collection in (I) hydrophobic and (II) hydrophilic microchannels. (b) Optical images of water droplets in hydrophobic and hydrophilic microchannels ^[173]. (c) A series of hydrophobic microchannels of different sizes in microfluidic devices to measure sweat secretory pressure ^[171]. (d) Schematic illustrates the liquid collection in sweat-absorbing materials. Soft sweat collection devices based on (e) sponge ^[181], (f) paper ^[182], and (g) fabric ^[183]. Schematic illustrates the liquid collection in (h) a microwell superwettability pattern and (k) an asymmetric superwettability pattern. Superwettability pattern bands based on (i) flexible PET substrate ^[184] and (j) stretchable SEBS substrate ^[185] for sweat sampling and sensing. (l) Superwettability wedge-patterned microchannel for sweat sampling and sensing ^[187].

The preparation of low-cost hydrophilic components in wearable epidermal microfluidic devices often involves the use of sweat-absorbing porous materials such as the sponge, paper, and fabric with the strong capillary force for rapid collection (Figure 6d) ^[180]. Consisted of an inductive coil and a planar capacitor formed with interdigitated electrodes on the hydrophilic sponge substrate, a stretchable and wireless sweat sensor can analyze the sweat during spontaneous collection (Figure 6e) ^[181]. Besides the assessment of the sweat volume from dielectric change with an accuracy of 0.06 μL/mm², the sensor can also use the colorimetric method to detect sweat composition such as OH⁻, H⁺, Cu⁺, and Fe²⁺. The inductive coil integrated with the planar capacitor and interdigitated electrodes also allow wireless measurements. With water-activated dyes located at the tips of radial finger channels, the paper-based epidermal patch can measure the sweat loss as the sweat fully saturates finger channels with varying lengths to change color at the tips (Figure 6f) ^[182]. The paper-based patches can be easily tailored in varying sizes and patterns to handle a broad range of sweat rates from 1.5 to 15.3 μL cm⁻² min⁻¹ for personalized use. The stretchable cotton fabric with improved breathability as a capillary channel can also be used for the collection and passive delivery of sweat from the skin to the sensor electrode for biomarkers detection (Figure 6g) ^[183]. The fractal structure with 8 main branches in the sweat collection unit exhibits the shortest induction time (< 1 min in the 4.9 cm² region after sweating) and highest collection flux (up to 4.0 μL cm⁻² min⁻¹), compared with other fractal structures based on the analysis of minimal flow resistance and nodal distance. Therefore, the resulting device allows fast response to abrupt changes in sweat rates or concentrations for real-time detection of physical conditions during exercise.

To direct sweat from the skin surface to the sensing area, patterned superhydrophobic and superhydrophilic regions in a superwettable band (Figure 6h) enables wetting only in the superhydrophilic microwells with embedded indicators for colorimetric detection of pH, chloride, glucose, and calcium (Figure 6i) ^[184]. The superhydrophobic surface is generated by the roll-to-roll coating of nano-dendritic silica, whereas oxygen plasma through shadow-masked defines the superhydrophilic sensing microwells. Assembling superhydrophilic silica nanoparticles-decorated PU nanofiber textiles on the SEBS-based substrate can also yield an intrinsically stretchable sensing patch (50% stretching) for colorimetric sensing in the superhydrophilic region (Figure 6j) ^[185]. The large wettability contrast can direct sweat collection with reduced sampling time.

The two opposite surfaces of the droplet can form different curvatures on the cactus-inspired asymmetric superwettability patterns, which further induce changes in the Laplace pressure to efficiently direct sweat transport (Figure 6k) ^[186]. For instance, the gradually widened superwettability patterns toward the center in the microchannels increase sweat collection speed to 70 mm s⁻¹ by unidirectional Laplace pressure, compared with 0.05 mm s⁻¹ in the stripe-patterned channels (Figure 6l) ^[187]. The asymmetric wedge-patterned superwettability surface is fabricated by selective spray-coating of PVA-modified hydrophilic SiO₂ NPs onto the superhydrophobic octadecyltrichlorosilane (ODTS)-modified hydrophobic SiO₂ NPs in the background. With the sweat confined inside the channel, the doubled sweat-collecting efficiency (over the stripe-patterned microfluidic channel) reduces the filling time in the sensing area to 5 min after exercise. The patch integrated with a sweat sensor can quickly respond to changes in sweat biochemicals for continuous monitoring.

3.5. Directing Sweat Flow

Though sweat biomarkers offer insights into disease states (e.g., diabetes, cystic fibrosis, and dehydration), the single time-point measurement from colorimetric assays in one reaction reservoir is often not sufficient ^[69]. As a result, multiple reservoirs with microvalves to direct the sweat flow in the microfluidic devices are often needed for chronometric sampling. The capillary bursting valves rely on the change in the geometry of the microchannels (Figure 7a) ^[188], as their bursting pressure (BP) to open is related to surface wettability of the rectangular microchannel with width b and height h in the diverging region ^[171]:

$$BP=-2\gamma \left[\frac{\cos {\theta }_I^{\ast }}{b}+\frac{\cos {\theta }_A}{h}\right]$$ (2)

where θ_A is the critical advancing contact angle of the channel, ${\theta }_I^{\ast }=\text{min}[{\theta }_A+\beta ,{180}^{{}_{°}}]$, and β is the diverging angle of the channel. Compared to their hydrophobic counterparts, the hydrophilic microchannels (or microvalves) of the same dimension are associated with lower bursting pressures. Therefore, the hydrophobic recovery in certain hydrophilic microvalves leads to a gradual increase in the bursting pressure to even exceed the sweat secretion pressure, resulting in function loss of the microvalves. As exposure to oxygen plasma is often used to activate the PDMS surfaces for device layer bonding, it is better to wait for a complete hydrophobic recovery in the as-prepared devices (e.g., 24 h) ^[97]. Combing a series of capillary bursting valves with varying bursting pressures in different locations of microfluidic devices ensures their effective operation (Figure 7b) ^[189]. With the bursting pressure of the first set of capillary bursting valves larger than that of the second set located at the inlet of the reservoirs, reservoirs can be completely filled before the built-up pressure bursts open the second set for continued sweat flow.

Figure 7.

Directing sweat flow in wearable devices by surface wettability design. Schematics illustrating the control of sweat flow by (a) capillary bursting valves, (c) hydrophobic valves, (f) wettability-switchable valves, and (h) patterned wettability (e.g., hydrophobic barrier in hydrophilic porous absorbent materials). Time-sequence images of sweat collection by a series of (b) capillary bursting valves ^[189], (d) hydrophobic valves ^[190], (e) one-opening reservoirs with hydrophobic valves ^[173], and (g) wettability-switchable hydrophobic valves in wearable microfluidic devices ^[192]. Patterned wettability with (i) the wax-printed hydrophobic barrier in the 2D paper ^[70], (j) the wax-printed hydrophobic barrier in the 3D paper ^[197], and (k) embroidered hydrophilic threads in hydrophobic fabric substrate for wearable microfluidic devices ^[198].

Selective hydrophilic treatment can create hydrophilic microchannels with hydrophobic valves in the untreated area (Figure 7c). As the advancing front of the sweat is blocked by the hydrophobic valve, the sweat preferentially flows along the hydrophilic reservoir wall to gradually complete the filling. After the sweat fills each reservoir from the microchannel, the built-up pressure from sweat secretion bursts opens the hydrophobic valves to start filling the next reservoir (Figure 7d) ^[190]. Moreover, the hydrophobic valves are placed in the bridge channel between reservoir and microchannel to allow air release during sweat filling, thereby enabling the reservoir with one opening design to reduce sweat evaporation (Figure 7e) ^[173]. The tunable surface wettability can provide an actively controlled valve to sample, route, and compartmentalize biofluid (Figure 7e) ^[191]. The on-demand electrowetting valve is demonstrated with an electrode modified by a hydrophobic self-assembled monolayer (1H, 1H, 2H, 2H-perfluorodecanethiol, PFDT) downstream to a pair of inkjet-printed Ag electrodes on a hydrophilic PET substrate. As a low voltage is applied to the Ag electrodes, the hydrophobic monolayer becomes hydrophilic to allow continued fluid through the microchannel. Integrating four independent electrowetting valves in the hydrophilic microchannels achieves the on-demand sweat collection upon the applied voltage for each valve (Figure 7g) ^[192]. Moreover, the pH-responsive superwetting surface possesses great potential for active sweat control, resulting from the synergy between the organosilanes (3[2-(2-amino ethylamino) ethylamino] propyl trimethoxy silane (AEPTMS) and octyl trimethoxy silane (OTMS) ^[193]. The protonation of the pH-sensitive amino groups in AEPTMS in acidic solutions renders the surface hydrophilic, whereas the unprotonated AEPTMS and OTMS are superhydrophobic. Besides the pH, the sensing surface can also respond to urea and glucose in sweat due to the generated free hydroxyl and hydrogen ions from enzyme-catalyzed reactions.

To realize multiplexed sweat analysis with porous absorbent materials, the hydrophobic barrier is patterned on the hydrophilic substrate to define the sweat transport channels and detection regions (Figure 7g) ^{[112, 194-196]}. The hydrophilic filter paper-based wearable microfluidic device with various hydrophobic shaped channels and functional zones patterned by wax printing is capable of guiding the sweat flow for multiplexed sensing of sweat loss, pH, glucose, and lactate concentrations (Figure 7h) ^[70]. A 3D paper-based wearable microfluidic device with a vertical flow channel fabricated by folding the pre-defined 2D channels allows continuous sweat flow driven by evaporation of sweat, which ensures fresh sweat flow across the electrodes (Figure 7i) ^[197]. Stitching or embroidering hydrophilic cotton threads into the hydrophobic-treated cotton fabric substrate also creates fabric-based microfluidic devices with the hydrophobic barrier to effectively direct the sweat flow (Figure 7j). With the colorimetric treatment of the hydrophilic thread in the detection zones, the resulting fabric-based microfluidic devices can be used to detect electrolytes and metabolites (e.g., pH, chloride, and glucose concentrations) and local sweat loss ^[198].

3.6. Electrochemical Sweat Analysis

Different from colorimetric analysis, the electrochemical sweat sensor with functionalized electrodes reversibly transduces sweat analyte concentration into electrical signals (current or voltage) to provide real-time sweat analysis with high sensitivity ^{[16, 59]}. Though the electrochemical sensors do not need sophisticated microfluidic systems, their efficiency highly depends on the surface wettability of electrodes for interacting with sweat, leading to extensive efforts on the synthesis and design of electrode materials ^[199]. Because of good electrocatalytic performance, stability, and electrical conductivity, conventional electrode materials (e.g., metal nanoparticles, metal oxides, and carbon nanomaterials) are still widely used ^[12]. However, their hydrophobicity results in limited electrolyte diffusion onto the electrode surface to compromise their electrochemical performance in sweat analysis especially when insufficient sweat is secreted, which calls for the development of hydrophilic electrodes (Figure 8a) ^[200-204]. The hydrophilic poly(urethane-acrylate) (HPUA) mixed with the Ag flakes can be printed as electrodes on the textile to provide improved diffusion over the hydrophobic Ag-SEBS electrode (Figure 8b) ^[205]. The redeposited Ag from dissolved Ag⁺ also increases the contact between adjacent Ag flakes and to the target sweat analytes for enhanced sensitivity. Compared with the hydrophobic Ag-SEBS electrode, the hydrophilic Ag-HPUA electrode with reduced resistance and enhanced affinity from the hydrophilic surface facilitates sweat diffusion, which is particularly important for drastically reduced sweat volume. Introducing the nanodendritic structure to the hydrophilic Au electrode further increases the specific surface area and enhances response signals (e.g., 3 times in peak current of the bare gold electrode) ^{[167, 206]}. The hydrophilic nitrogen-doped carbon textile (SilkNCT) substrate also contributes to the rapid sweat diffusion by capillary force to enhance the electrochemical reaction efficiency on the electrode for improved sensing of glucose, lactate, ascorbic acid, uric acid, Na⁺, and K⁺ (Figure 8c). For example, the response of SilkNCT to ascorbic acid (AA) from differential pulse voltammetry shows up to 5 times increase compared to that of the carbonized cotton fabric ^[207].

Figure 8.

Electrochemical sweat sensing facilitated by surface wettability design. Schematics illustrating the (a) liquid and (d) oxygen diffusions in the hydrophilic or hydrophobic electrode. (b) Hydrophilic Ag-HPUA electrode for enhanced affinity to sweat ^[205]. (c) Hydrophilic carbonized silk fabric electrode for rapid sweat transport ^[207]. (e) Superhydrophobic triphase electrodes to supply sufficient oxygen for high enzyme activity ^[209] and (f) their use in wearable sweat sensors for real-time sweat analysis ^[210].

However, the hydrophilic electrode completely wetted by sweat has limited oxygen permeability (2.1 × 10⁻⁵ cm² s⁻¹ in liquid versus 2.0 × 10⁻¹ cm² s⁻¹ in air). As a result, the sensing performance (e.g., the range of linearity, sensitivity, and accuracy) of enzyme-based electrochemical sensors is affected due to the oxygen-mediated enzymatic reactions ^[208]. To address the challenge in traditional solid-liquid two-phase electrodes, a superhydrophobicity air-liquid-solid triphase electrode allows the air layer to form at the liquid/electrode interface (Figure 8d). The flexible triphase enzyme electrode is fabricated by assembling an oxidase enzyme layer and Pt electrocatalysts on a hydrophobic CNT film/porous PVDF substrate (Figure 8e) ^[209]. With facilitated oxygen diffusion to the active oxidase sites, the triphase enzyme electrode maintains stabilized oxidase kinetics to detect glucose, lactate, and sucrose with a wide linear range, a low detection limit, high sensitivity, and selectivity. In addition to higher oxidase kinetics of 487.5 min⁻¹ (vs. 43.1 min⁻¹ for diphase electrode) for glucose detection, the triphase electrode has a detection limit of 5 μM and a linear detection range up to about 30 mM (vs. 0.7 mM for biphase electrode). Compared with the biphase electrode, the triphase electrode also exhibits an enhanced linear detection range for the lactate (32 times from 2.5 mM to 80 mM) and sucrose (25 times from 0.8 mM to 20 mM) detection. In a different design, the triphase electrode consisting of ultrathin porous CNTs-intercalated Ti₃C₂T_x/PB film on a conducting superhydrophobic carbon fiber layer with air holes also allows oxygen to easily diffuse through the substrate to the enzymatic reaction zone (Figure 8f) ^[210]. The resulting wearable sweat-based biosensors exhibit a large linear range (10-1.5×10³ μM, 0-22 mM), ultra-high sensitivity (35.3, 11.4 μA mm⁻¹ cm⁻²), and accuracy (0.33 μM, 0.67 μM) to monitor glucose and lactate, respectively.

3.7. Droplets Energy Harvesting

Droplets triboelectric nanogenerators (droplets-TENGs) that combine triboelectricity and electrostatic induction at the liquid/device interfaces present a low-cost and highly efficient water energy harvesting solution. The droplets-TENGs based on either the single electrode or free-standing mode suffer from drastically reduced electrical output due to the remaining liquid on the surface (Figure 9a) ^{[62, 211]}. By adjusting the surface wettability from hydrophilic to hydrophobic, the electrical output is enhanced by two orders of magnitude (Figure 9b) ^[212]. The superhydrophobic surface can repel water at the surface of droplets-TENGs to avoid charge rebalance and enhance the electrical output. Moreover, the slightly increased (but not exceedingly high) roughness of the superhydrophobic surface can also promote electron transfer and the water contact-separation process for enhanced electrification. Although the superhydrophobic surface with moderate roughness is easier to be damaged by high-speed water impact than that with high roughness ^[213], it exhibits a 14-fold increase in the output current and a 9-fold increase in the output voltage (Figure 9c) ^[214]. Furthermore, the TENGs with a nanostructured superhydrophobic surface shows significantly decreased output voltage even upon low-velocity water impact due to degeneration of hydrophobicity from local water film formation on the surface. Meanwhile, the waterproofing and self-cleaning properties can ensure the long-term operation of droplets-TENGs in various complex hydrated environments. Therefore, the roughness of droplets-TENGs should be carefully designed for desirable output performance. In the all-fabric-based droplets-TENGs with hydrophobic NPs spray-coated onto the hydrophilic PET fabric, the output voltage is increased from 5 to 15 V for a water flow rate of 6 ml/s, as the surface changes from hydrophobic (127.1°) to superhydrophobic (162.1°) ^[215]. Its combination with contacts-TENGs in a cotton glove can collect both the triboelectricity and mechanical energy of water flow, increasing the maximum instantaneous output power density from 0.14 mW m⁻² to 0.3 mW m⁻². An amphiphobic self-charging power raincoat consisting of a hydraulic TENG (with a maximum peak power density of 245.2 mW m⁻²) and several fiber-supercapacitors can light up the light-emitting diode (LED) for more than 300 s after showering with water for 100 s (Figure 9d) ^[76]. By introducing a top electrode on the hydrophobic surface to periodically contact with water droplets, the resulting all-fabric TENGs show a 7-fold enhancement in output voltage over the conventional single-electrode mode TENGs (Figure 9e) ^[86]. As a result, the TENGs with a maximum output power of 0.11 mW can power 25 LEDs under one water droplet. The self-repairing superhydrophobic surface also avoids the degeneration of the superhydrophobic surface for extended service life.

Figure 9.

Wearable triboelectric nanogenerators devices with surface wettability. (a) Schematics illustrate the working principle of droplets-TENGs based on (I) single electrode mode and (II) free-standing mode with and without superhydrophobic surface. The influence of (b) wettability ^[212], and (c) roughness ^[214] on the output performance of the superhydrophobic droplets-TENGs. (d) Superhydrophobic droplets-TENGs combined with contacts-TENGs were integrated into a cotton glove for water energy harvest ^[215]. (e) Superhydrophobic droplets-TENGs were devised to be a power raincoat for water energy harvest ^[76]. (f) All-fabric TENG with the superhydrophobic self-repairing ability for water droplet energy harvesting ^[86].

3.8. Moisture-enabled Energy Harvesting

To harvest energy from moisture or sweat for self-powered devices, the hydrovoltaic generator explores evaporation-induced water flow in the porous carbon micro-/nanochannels to induce a streaming potential and form an electrical double layer at the electrolyte/channel wall interface (Figure 10a) ^{[61, 216, 217]}. In general, the hydrophilic treatment further facilitates the flow in micro/nanochannels to affect the dynamics of ion migration and enhance the steaming current. Therefore, different functional groups on the surface strongly affect electron depletion when in contact with water molecules. For example, density functional theory indicates that the graphene sheets with a hydrophilic C-O-C group can deplete 0.7 e electron charges, but the value drops to 0.0003 e without it ^[218]. Therefore, the hydrovoltaic generator was great potential to harvest energy from human sweat. A self-powered wearable sweat analyzer can be driven by the sweat flow in hydrophilic porous carbon nanochannels during natural sweat evaporation and capillary pumping to generate an output voltage ^[219]. As the process is mediated by the enzyme reaction between lactate oxidase on the porous carbon film and lactate in sweat, the change in sweat lactate concentration during exercise can be wirelessly measured by the output voltage (Figure 10b). However, the high concentration of ions in sweat drastically decreases evaporation-induced electricity generation. To address this challenge, the evaporation-induced generator is combined with a primary battery as a dual-mode electricity generator (DM-ENG), exhibiting stable performance even in high salinity conditions (Figure 10c) ^[220]. In the porous carbon black/PVA film with wettability patterns, the sweat flows in the hydrophilic region to continuously generate a streaming potential, whereas the hydrophobic region that prevents the liquid transport forms a primary battery system. The decrease in streaming potential can be compensated by the primary battery system for the DM-ENG to maintain stable performance in a wide range of ion concentrations across 10 orders of magnitude. Therefore, the voltage output only slightly increases from 0.8 to 0.9 V, though NaCl concentration increases from 0.046 ± 0.037 M to 0.41 ± 0.08 M by almost ten times.

Figure 10.

Moisture-enabled nanogenerators with enhanced performance from surface wettability design. Schematics illustrating the moisture-enable nanogenerators based on (a) steaming current and (d) ion gradient diffusion. (b) Self-powered wearable sweat-lactate analyzer based on steaming current ^[219]. (c) A flexible dual-mode electricity nanogenerator consists of a steaming current and a primary battery for stable output during sweating ^[220]. (e) Self-powered wearable sensor system based on ion gradient diffusion ^[222]. (f) Self-powered wearable textile-based power generator system based on ion gradient diffusion ^[224].

Driven by a concentration gradient that is created by the liberated anions or cations from the dissociation of the hydrophilic surface functional groups (-OH or -COOH) upon water absorption, the ion diffusion can also be leveraged for energy harvesting (Figure 10d) ^{[24, 221]}. These mobile ions over the surface act as charge carriers and move in response to the concentration gradient, producing an ionic current ^[61]. By creating a gradient of oxygen-containing groups on the porous polydopamine layer between pencil-drawn interdigital graphite electrodes, a moisture-enabled nanogenerator (MENG) can provide sustained output voltage and power density of 0.52 V and 0.246 mW/cm² in 90% RH. The exposure to moisture from artificial exhalation allows the generator to power the flexible pressure sensor for real-time monitoring of human physiological signals such as breathing and pulse (Figure 10e) ^[222]. Additionally, the MENG created on office-copy papers generates a sustained open-circuit voltage of up to ~480 mV for more than 2 h in 95% RH, which can drive the iontophoretic system for transdermal drug delivery ^[223]. The dual asymmetric structure design further enhances the concentration difference of charge carriers for efficiently driving the diffusion of ions in the MENG, resulting in a high open-circuit voltage of up to 1 V and a short-circuit current of 1.7 μA (Figure 10f) ^[224]. The power source integrated into the T-shirt is sufficient to drive commercial wearable electronics such as electronic watches and display screens.

4. Conclusion & Outlook

In summary, this mini-review first provides a brief overview of a few representative methods to engineer surfaces with target wettability, where the hydrophobic (or hydrophilic) surface reduces (or facilitates) fluid wetting. Next, the surface wettability design has been shown to enhance the performance of various skin-interfaced sensors and devices to ensure measurement accuracy and improve the level of comfort in varying hydrated environments. The energy harvesters with the surface wettability design also exhibit increased harvesting efficiency to power the on-body electronics. Despite significant strides that have been made, there is still a critical gap between the devices at the proof-of-concept stage and the eventual practice applications in real environments. However, the existing challenges also present a small fraction of opportunities for future developments.

Stable wettability is desirable for the devices to operate in diverse harsh and demanding hydrated environments. However, hydrophobicity degeneration occurs in high humid or water vapor environments due to nucleation and condensation of small-sized water molecules within the micro-/nano-structured surface ^[225-227]. To provide sustained superhydrophobicity in these challenging environments, the superhydrophobic surface can be combined with efficient low-voltage Joule heating to prevent the condensation and nucleation of water molecules ^[228]. Although reliable long-term use of the device relies on durability, the surfaces with micro-/nano-structured superwettability are often highly susceptible to abrasion and extrusion. Therefore, it is highly desirable to explore advanced design strategies such as a microstructured surface frame “armor” to prevent the removal of the nanostructures by abradants for improved reliability (Figure 11a) ^[229].

Figure 11.

Opportunities for future developments of skin-interfaced sensors and devices with the surface wettability design. (a) Microstructured surface frame “armor” to prevent the removal of the nanostructures by abradants for improved reliability ^[229]. (b) Temperature-responsive smart Janus textile for moisture/thermal management ^[230]. (c) Nanogenerators connected in series or parallel to achieve enhanced output voltage and current for powering common electronic devices ^[220]. (d) Hydrovoltaic power generator with superabsorbent hydrogel for sustained electricity generation ^[216]. (e) Sustained power supplies consisting of micro-supercapacitor arrays and nanogenerators to drive stretchable sensors in a standalone platform ^[234].

As the wettability pattern depends strongly on the precision of the surface treatment, it is important to explore high-precision patterning techniques. Furthermore, exploring tunable wettability surfaces can result in intelligent surface wettability to functionally operate on-demand in dynamically changing environments. The need to facilitate heat/moisture dissipation in hot weather but maintain a warm and wet microclimate in cold weather presents challenges for conventional Janus textiles, as the wetting gradient cannot be reversibly switched. Fortunately, integrating temperature-responsive polymers on both sides of the textile can provide reversible gradients to modulate the direction of water transport for thermal adjustment upon external temperature changes (Figure 11b) ^[230]. The materials and structures with intelligent wettability have been gaining momentum to provide expanded opportunities for skin-interfaced sensors and devices.

While energy harvesting with varying device designs from hydrated environments presents a high potential for self-powered sensors and devices ^{[212, 214, 224]}, the output power density is often not sufficient to drive key electronics such as wireless transmitters. As a simple solution, these nanogenerators can be connected in series or parallel to provide higher output voltage or current for powering common electronic devices (Figure 11c) ^[220]. However, the specific output voltage or current for target applications such as the cardioverter defibrillator may not be readily obtained. Therefore, innovation in advanced functional materials and device layouts is still needed.

The envisioned self-powered electronic systems with energy harvesting from hydrated environments often only work in these challenging environmental conditions (e.g. raining, high humidity, or sweating). One effort to reduce the dependence of energy harvesters on environmental conditions results in the use of superabsorbent hydrogel that provides a sustained water source for the hydrovoltaic power generator (Figure 11d) ^[216]. The stable power supply allows the strain sensor to detect human motions with reliable current readings. It is also possible to integrate the other energy harvesting units ^[231-233] and further store the harvested intermittent energy in supercapacitors or other energy storage modules toward a standalone device system (Figure 11e) ^[234]. The multiple components in this new class of standalone device systems also present challenges in low-cost and scalable manufacturing, which need to be addressed before successful commercializ.

Biographies

Prof. Xiufeng Wang is a professor in the School of Materials Science and Engineering, Xiangtan University, China. He received his Ph.D. (2011) from Xiangtan University, China. From 2016 to 2017, he worked as a visiting scholar at Northwestern University, USA. His research interests focus on wearable microfluidic devices and flexible electronics.

Liu Yangchengyi is a PH.D. candidate at the School of Materials Science and Engineering, Xiangtan University, under the supervision of Prof. Xiaoping Ouyang and Prof. Xiufeng Wang. He is also currently a visiting student at Nanyang Technological University in Singapore, funded by China Scholarship Councip (CSC). His current research interests include wearable mechanical sensors and microfluidic devices.

Prof. Huanyu Cheng earned a Ph.D. degree from Northwestern University in 2015 and a Bachelor’s degree from Tsinghua University in 2010. After his doctoral study, he was appointed as the Dorothy Quiggle Assistant Professor of Engineering Science and Mechanics at The Pennsylvania State University. Dr. Cheng has worked on standalone stretchable sensing systems for biomedicine with over 110 peer-reviewed publications, and his work has been recognized through the reception of numerous awards. He also serves as an associate editor for IEEE Internet of Things Journal, Computers in Biology and Medicine, and reviewer for 196 international journals.

Prof. Xiaoping Ouyang is a professor at the School of Materials Science and Engineering, Xiangtan University, China. He obtained his Ph.D. from Fudan University in 2002. He was elected as an academician of the Chinese Academy of Engineering in 2013. His current research includes the development of functional materials and their application in pulse radiation, human healthcare monitoring, and energy storage.