Ultrathin hierarchical hydrogel–carbon nanocomposite for highly stretchable fast-response water-proof wearable humidity sensors†

Bingqi Pan , Peipei Su , Minghui Jin , Xiaocheng Huang , Zhenbo Wang , Ruhao Zhang , He Xu , Wenna Liu and Yumin Ye *
Department of Materials Science and Engineering, Faculty of Materials Science and Chemical Engineering, Ningbo University, Ningbo 315211, P. R. China. E-mail: yeyumin@nbu.edu.cn

First published on 13th September 2023

---

---

New concepts Hydrogels are ideal materials for wearable humidity sensors owing to their good humidity responsiveness, flexibility, and biocompatibility. Conventional bulk hydrogel humidity sensors, however, suffer from rather slow response/recovery due to the large moisture uptake and prolonged water molecule penetration, as well as a small specific surface area. Here we overcome this problem by designing wearable humidity sensors using a unique ultrathin hydrogel–carbon nanocomposite with a hierarchical micro–nano architecture. Synthesized via a solvent-free initiated chemical vapor deposition (iCVD) method, the nanocomposite has both a microscale periodic wavy structure and a nanoscale highly porous surface, which significantly enlarges its specific surface area. The submicron thickness and large specific surface area contribute to high sensitivity and a fast response/recovery (13/0.48 s) of the sensor over a wide humidity range (11–96%). The hierarchical architecture of the nanocomposite also enables excellent sensing stability against repeated bending (180°), stretching (100% strain), and scratching. This study thus presents a simple and effective strategy to realize hydrogel sensors with ultrathin thickness yet large specific surface area, thereby overcoming the limitations faced by bulk hydrogel humidity sensors. It also sheds light on the future preparation of high-performance sensors with miniature size, fast response/recovery, and durability.

Introduction

Flexible and wearable sensors have garnered increasing attention for their potential applications in healthcare, electronic skin, and human–computer interactions.^1–3 Humidity represents one of the key indicators that carries important human physiological information.^4,5 Real-time monitoring of humidity in respiration and on skin allows the diagnosis and timely treatment of many respiratory and skin diseases, as well as tracking body metabolism, skin health, and wound recovery. Although various humidity sensors have been developed for health monitoring,^6–8 fabrication of wearable humidity sensors that can be used for real-time monitoring on a daily basis remains a challenge, as it not only requires high sensitivity and stability of the sensor, but also fast response/recovery for rapid detection of subtle humidity variation, excellent flexibility and stretchability that conforms with skin, along with durability and resistance to external media such as water, sweat, and bacteria.

To achieve humidity sensors with high sensitivity and fast response/recovery, sensing elements with strong moisture interaction and high specific surface area are of vital importance. Among which, functional inorganic nanomaterials, such as carbon and metal oxides, and polymers are most widely used.^9–11 Inorganic nanomaterials are advantageous due to their abundant surface active sites, good chemical stability, and electrical properties, but they suffer from poor flexibility and stretchability. Polymers, especially hydrogels, have excellent biocompatibility, elasticity, and humidity responsiveness; therefore, are ideal materials for wearable humidity sensors.^12–14 Various hydrogel-based humidity sensors have been reported with good sensing performance and flexibility.^15–18 However, conventional hydrogel humidity sensors usually employ hydrogels with large thickness.^19–22 These bulk hydrogels could uptake a large amount of moisture before their resistivity reaches equilibrium and the diffusion of water molecules within the hydrogel could take a long time, which seriously impedes the adsorption and desorption process of water molecules, resulting in a rather slow response and recovery. Bulk hydrogel also has a rather small specific surface area for water molecule interaction, resulting in limited sensitivity.

Great efforts have thus been devoted to achieving thinner hydrogel films with a large specific surface area for improved response/recovery time and sensitivity.^23–25 For example, Gao et al. reported flexible humidity sensors with 100 μm hydrogel films to achieve humidity sensing within the relative humidity (RH) range of 20–90% and fast detection of human respiration.²⁶ Liang et al. developed stretchable and transparent humidity sensors using polyacrylamide-based hydrogel films with a thickness as thin as 54 μm.²⁷ The resultant sensor exhibited good sensitivity and stretchability, but both the response and recovery of the sensor still required more than 200 s over the 11–98% RH range, which indicates that even thinner hydrogel films are required for a faster response. However, thinner hydrogel films lack the mechanical robustness for repeated stretching and have rather low conductance. An effective strategy is to coat submicron hydrogel films on flexible substrates, such as polydimethylsiloxane (PDMS), and create surface micro- and nano-structures for enlarged specific surface areas. Traditional solution-based methods, however, have difficulties in achieving ultrathin hydrogel films with precise thickness control and good preservation of the micro- and nano-structures on the substrate surface. In addition, solvent treatment may disrupt delicate surface microstructures and leave impurities and residues within the sensor, affecting its performance.

Vapor deposition represents an effective alternative in obtaining ultrathin conformal hydrogel films. Initiated chemical vapor deposition (iCVD) is a mild polymer thin film fabrication method that allows in situ free radical polymerization on a substrate surface, achieving highly conformal nanoscale coatings with precise thickness control and excellent retention of surface microtopography.^28,29 Hydrogel nanocoatings have been synthesized using iCVD and successfully deposited on surfaces with micro- and nanostructures, such as fabrics,^30,31 microporous membranes,^32,33 and carbon nanotube forests,^34,35 without disrupting the original surface micro- and nano-structures. However, iCVD hydrogel thin films usually have a chemically crosslinked structure, which could limit their stretchability and affect their ion transportation ability, and therefore their conductivity.³⁶

Herein we designed a highly-stretchable fast-response humidity sensor employing a unique ultrathin hydrogel–carbon nanocomposite with a micro–nano hierarchical architecture via a facile two-step solvent-free method. The nanocomposite film was synthesized by depositing a submicron iCVD hydrogel coating on a stretched PDMS film pre-deposited with a thin layer of candle soot (CS) nanoparticles. CS nanoparticles share many merits with graphene and carbon nanotubes, such as a large specific surface area and good electrical conductivity, but are much easier to fabricate and more cost-effective; therefore, they were adopted to both create a surface nanostructure and enhance film conductivity.^37–39 The deposited CS film was conformally encapsulated by the hydrogel nanocoating, which largely retained its nanotopography and enhanced its mechanical robustness. Upon release of PDMS, microscale periodic wrinkles were generated, which along with the hydrogel-coated CS nanoparticles constituted hierarchical surface topography. The wrinkle structure gave the nanocomposite excellent stretchability, while the hierarchical topography significantly enlarged its specific surface area, enabling a humidity sensor with high sensitivity and a response/recovery time of 13/0.48 s over a wide RH range of 11–96%. Encapsulating the nanocomposite with an elastic iCVD-coated superhydrophobic fabric further imparted the sensor with potent water-proof, self-cleaning, and antibacterial abilities, while retaining water vapor permeability. Finally, the nanocomposite sensor demonstrates successful real-time monitoring of various human respiratory behavior and humidity on different parts of skin.

Results and discussion

Fabrication and characterization of the nanocomposite

The hierarchical hydrogel nanocomposite was prepared in a two-step solvent-free process (Fig. 1a). PDMS films were briefly skimmed over a candle flame to deposit a thin layer of CS nanoparticles on the surface, followed by stretching at certain strains and fixing in the iCVD chamber. The iCVD of hydrogel coating was then carried out by feeding vapors of initiator tert-butyl peroxide (TBP), monomer vinyl pyrrolidone (VP), and crosslinker ethylene glycol diacrylate (EGDA) into the chamber, where TBP was decomposed using a heated filament, generating free radicals that polymerized monomers and creating a conformal ultrathin poly(vinyl pyrrolidone-co-ethylene glycol diacrylate) (pVE) coating on the CS and PDMS surface (pVE/CS/PDMS) (Fig. 1b). After deposition, the pVE/CS/PDMS composite film was released and retracted, generating periodic microscale wavy wrinkles. Nanocomposite films with 0%, 50%, 100%, and 150% prestrains were called pVE/CS/PDMS, pVE/CS/PDMS-50%, pVE/CS/PDMS-100%, and pVE/CS/PDMS-150%, respectively. To give water-proof and antibacterial properties, an elastic spandex fabric was coated with a conformal ∼200 nm layer of hydrophobic poly(perfluorodecyl acrylate-co-ethylene glycol diacrylate) (p(PFDA-co-EGDA)) to encapsulate the nanocomposite sensor (Fig. 1c).

Hydrogels with a low crosslinking degree are generally more hydrophilic and have a stronger affinity toward water molecules, but low crosslinking also induces poor stability.⁴⁰ We thus synthesized pVE films with different compositions (pVE1–pVE3) and examined their surface hydrophilicity and stability (Table S1, ESI†). Fig. 2a shows the Fourier transform infrared (FTIR) spectra and water contact angle (WCA) of pVE films along with homopolymer films of poly(vinyl pyrrolidone) (pVP) and poly(ethylene glycol diacrylate) (pEGDA). It is apparent that the increased VP content in the hydrogel films of pVE-1 to pVE-3, evidenced by the strengthened CO stretching intensity at 1665 cm⁻¹ from the VP moiety and weakened CO stretching intensity at 1733 cm⁻¹ from the EGDA moiety,⁴¹ corresponds to the decreased crosslinking degree (from 33.3% to 16.7%) and increased surface hydrophilicity, thus lower WCA. Film pVE-3 has the highest VP content, the lowest crosslinking degree, and minimum WCA of 34° with relatively good stability upon immersion in water; therefore, it was chosen as the model hydrogel film for sensor fabrication.

Fig. 2b reveals the morphology of CS nanoparticles deposited on PDMS. Scanning electron microscopy (SEM) observations revealed clusters of CS nanoparticles with sub-100 nm diameters dispersed on the surface with relative uniformity. The hydrogel coating wrapped around each particle and created a highly porous composite with stacked hydrogel/CS core–shell nanoparticles with increased diameter (Fig. 2c). The retention of the nanoporous structure is of paramount importance for sensing, as it provides a significantly enlarged specific surface area for water molecule adsorption/desorption, which could enable high sensitivity and fast response/recovery of the sensor.⁴² The iCVD method is thus advantageous since it not only produced a conformal nanocoating that preserves the surface nanostructure, but also avoided the use of any solvent which easily destroys the delicate stacked nanoparticle architecture.

X-ray photoelectron spectroscopy (XPS) was then employed to probe the surface composition of the nanocomposite film. High-resolution C 1s XPS spectra reveal that the carbon atoms under a C–Si environment on the PDMS surface have mostly been replaced by C–C/C–H after the deposition of CS (Fig. 2d and e). The altered bonding environment is due to the abundant alkyl groups on CS, which confirms the coverage of the PDMS surface by nanoparticles. The surface composition changed substantially after deposition of pVE hydrogel (Fig. 2f). Six carbon environments can be identified by resolving the C 1s peak (Table S2, ESI†), and the peak areal ratio reveals that the surface is composed of VP and EGDA moieties with the corresponding ratio of approximately 10:1, which agrees well with that from FTIR analysis (ESI†). This indicates that the iCVD process perfectly preserved the stoichiometry of the hydrogel coating and the deposited hydrogel film completely covered the CS/PDMS surface.

Periodic surface wrinkle patterns occur when films composed of both relatively stiff and soft layers are under lateral compression.⁴³ Upon release of the pre-stretched pVE/CS/PDMS composite film, the contraction of PDMS induced wrinkling of the relatively stiff pVE coating. The wavelength and amplitude of the wrinkle periodicity are related to the exerted stress, suggesting possible control of wrinkle topography.⁴⁴ Here we applied different prestrains to the composite film to induce different wavy topography. Fig. 2g–i show successful generation of ordered wrinkle structures with distinct wrinkle patterns resulting from different prestrains. As the prestrain increased from 50% to 150%, the wrinkle wavelength decreased from ∼10 μm to ∼1 μm. This trend is in accordance with literature in which the wrinkle wavelength decreases with the increased prestrain (ESI†).^45,46Fig. 2j shows the cross section of typical wrinkle structure of the nanocomposite film obtained from 50% prestrain. Each wrinkle has a width of about ∼10 μm and height of ∼6 μm. An enlarged view reveals a hydrogel coating of about 400 nm thickness conformally covering the entire wrinkle. A thickness of 400 nm was adopted because thinner coatings may not be able to completely cover the CS surface and thicker coatings may planarize the nanoparticle topography and reduce the specific surface area (Fig. S1, ESI†). The iCVD hydrogel coating was thus able to both retain the nanoscale topography of CS and induce microscale wrinkles, achieving micro–nano hierarchical surface architecture.

The wrinkle surface topography and nanoporous hydrogel/CS coating greatly improved the specific surface area and hydrophilicity of the nanocomposite. From BET analysis, even without the wrinkle structure, the pVE/CS/PDMS nanocomposite exhibits a significantly enlarged specific surface area of 0.51 m² g⁻¹, which is more than 26 times that of the planar pVE-coated PDMS film (Fig. S2, ESI†). Wrinkle topography further increased the specific surface area. The pVE/CS/PDMS-50%, pVE/CS/PDMS-100%, and pVE/CS/PDMS-150% nanocomposite films exhibit specific surface areas of 1.18, 1.61 and 2.04 m² g⁻¹, respectively. The pVE/CS/PDMS-150% nanocomposite shows a 107-fold increase of specific surface area compared to that of the planar pVE/PDMS film. The increased specific surface area of the nanocomposite with a higher prestrain is due to the decreased wavelength and heightened amplitude from the larger prestrain.

The hydrogel nanocoating and the large specific surface area greatly promoted surface wettability. The CS-covered PDMS is superhydrophobic with a WCA of 156° (Fig. 2k). The pVE hydrogel coating changed the surface to hydrophilic and lowered the WCA of pVE/CS/PDMS to 18° through the hydrophilic lactam groups from the VP moiety. Due to the enlarged surface area, the wrinkled pVE/CS/PDMS films achieved superhydrophilicity, while the films with 100% and 150% prestrains even reached a WCA of 0°. The superhydrophilic nature of the nanocomposite films confirms that the composite surface has a strong affinity to water molecules, and their large specific surface area can accommodate vast and fast moisture adsorption and desorption.

Sensing performance of the nanocomposite sensor

Using the hierarchical nanocomposite film as a sensing element, we constructed humidity sensors and evaluated their sensing performance with the setup shown in Fig. S3 (ESI†). For comparison, sensors made from nanocomposite films without wrinkle structure (pVE/CS/PDMS) and pVE-coated PDMS film without CS were also tested. Fig. 3a exhibits the current response of the sensors with the increase of RH at the logarithmic scale. The current in all sensors increases with the rise of RH, exhibiting a good logarithmic linear relationship. Incorporation of CS in the film greatly enhanced the current response (Fig. S4, ESI†) due to the improved electric conductivity and emergence of surface nanotopography. Wrinkling further triggered a larger response due to the enlarged specific surface area. A nanocomposite sensor with the largest prestrain (pVE/CS/PDMS-150%) shows the best linear fit (R² = 0.994).

The sensitivity of a humidity sensor is usually defined as the change in response, e.g., current, per change of RH, but its value may differ at high and low humidities, so different calculation methods are used.^47,48 Here we calculate the sensitivity as the total change of current in the RH range of 11–96% and compare it between different composite sensors. From Fig. 3b, sensors with hierarchical topography, thus a larger specific surface area, all exhibit higher sensitivity than that without the wrinkle structure, and larger prestrain induces higher sensitivity. A sensor made from pVE/CS/PDMS-150% nanocomposite shows the highest sensitivity of 42.35 nA/%. The high sensitivity of the nanocomposite sensor is attributed to the hydrophilicity of the pVE surface and the large specific surface area from the hierarchical topography that provides maximal water molecule adsorption capacity. In addition, the nanoporous structure tends to induce capillary condensation at high RH, which could also accelerate water molecule adsorption and further enhance the response of the sensor.⁴⁹

Real-time monitoring of human physiological behavior requires rapid detection of humidity change from the human body; fast response/recovery of the sensor is thus essential. The response/recovery time is often defined as the time required for the current change to reach 90% of its total current variation over a specific RH range. Evidently, all nanocomposite sensors exhibit fairly fast responses below 40 s, and even more swift recoveries all within 1.2 s at a wide RH range of 11–96% (Fig. 3c). Similar to the trend of sensitivity, the nanocomposite with a larger prestrain responds faster to RH change, although their recovery times are similar. The pVE/CS/PDMS-150% sensor has the shortest response/recovery time (13/0.48 s) when RH increased from 11% to 96% and then decreased again (Fig. 3d). The fast response/recovery of the hierarchical nanocomposite sensor represents one of the shortest over such a wide RH range among the reported hydrogel-based humidity sensors from the literature.^50–53 The superior performance of the nanocomposite sensor is a result of its hierarchical-structure-enabled large specific surface area and ultrathin thickness of the hydrogel, which substantially accelerated the adsorption and desorption of water molecules upon a change in RH.

More sensing performance tests were implemented on the pVE/CS/PDMS-150% nanocomposite sensor for a comprehensive evaluation. From the I–U curve of the pVE/CS/PDMS-150% humidity sensor, the slope increases with the increase of RH, indicating the conductivity increases and the resistance decreases at higher RH (Fig. 3e). This is a typical characteristic of a hydrogel humidity sensor, since more water molecules are adsorbed in the hydrogel at higher RH, promoting its ion conductivity. Fig. 3f shows the adsorption–desorption step curve of the nanocomposite sensor. As RH increases, the composite film adsorbs more water molecules and the current rises accordingly. The adsorbed water molecules escape from the composite as RH decreases and the current gradually declines, exhibiting a step-by-step incline/decline with good symmetry. Under the same RH, the sensor exhibits similar current values during both the adsorption and desorption processes, demonstrating very small hysteresis (deviation of the current during adsorption and desorption). Indeed, the largest hysteresis during adsorption/desorption is only 1.38% at the RH of 56%, possibly due to the very few remaining water molecules within the wrinkled nanoporous structure.

The stability of the humidity sensor is another important feature required for long-term use. During the cyclic response/recovery test, the pVE/CS/PDMS-150% nanocomposite sensor generates a stable current response in the 11–96% RH cycling process, indicating its good stability towards RH change (Fig. 3h and Fig. S5, ESI†). The sensor performance is also stable with a steady current response upon exposure to different RH environments for up to 30 days (Fig. 3i). Humidity sensors may accumulate excessive water adsorption at high humidity which condenses and causes the failure of the sensor.⁵⁴ The stable performance of the nanocomposite sensor even after long-term exposure to this high RH environment suggests that the micro–nano hierarchical structure of the nanocomposite facilitates not only adsorption of sufficient water molecules at high RH, but also their evaporation to prevent large-scale condensation, maintaining a good dynamic balance.

The sensing mechanism of the hydrogel nanocomposite sensor is illustrated in Fig. 4. The nanoporous hydrogel/CS and the periodic microscale wrinkles constituted a micro–nano hierarchical topography, which combines with the high surface energy of the hydrogel accommodating a large interface for intimate water molecule interaction. The nanoporous surface also triggers capillary condensation at high RH, which accelerates the penetration of water molecules within the nanopores and greatly enhances the sensitivity and responsiveness. Meanwhile, the ultrathin thickness of the hydrogel layer avoids the prolonged ion transportation commonly occurring in bulk hydrogels, facilitating fast adsorption and desorption of water molecules. Sensing of the nanocomposite at different RH follows the Grothuss mechanism:^55,56 water molecules are chemisorbed at low RH, forming hydrogen bonds with the polar groups of hydrogel, so their movement is restricted, exhibiting rather poor conductivity; physisorption of water occurs at high RH, and a hydration layer is formed, where protons and hydronium ions generated by ionization are freely transferred between adjacent water molecules, resulting in greatly improved conductivity.

Flexibility, stability, and durability of the sensors

In addition to good sensing performance, the practical application of wearable sensors also requires a number of other important features, including high flexibility and stretchability, stable performance under deformation, as well as good mechanical durability and resistance to possible liquid contamination. Since deformation causes structural disruption in sensing elements, maintaining a stable sensing performance under strain is critical for wearable sensors. We first conducted a bending test by bending the pVE/CS/PDMS-150% humidity sensor at different angles and measured its sensing performance (Fig. 5a). The nanocomposite sensor exhibits excellent bendability and maintains a good sensing performance in the bending range of 0°–180° (Fig. 5b). The good maintenance of the sensing ability against bending is due to the wrinkle structure on the surface. Upon deformation, the wrinkles are flattened, counteracting the effect of stress by raising the wavelength and decreasing the amplitude of the wrinkle; thereby preserving the intactness of the nanocomposite structure on the surface. Slight weakening of the current response of the sensor was observed under large bending angles. This is possibly due to the stretching of the nanocomposite upon bending, which lengthens the conducting pathway and reduces the cross-sectional area of the nanocomposite, leading to slightly increased resistance. It is also possible that some adjacent buckles are in contact under high RH at the initial state, which reduces its resistance. These buckles would separate upon bending; therefore, the resistance is relatively larger at the same RH.⁵⁷ Under the same bending angle, the response current increases with the increase of RH. Even in the extreme case of bending at 180°, a rather good sensing performance is retained with the sensitivity of 32.94 nA/% and response/recovery time of 25.2/0.4 s (Fig. 5c).

We then explored the stretchability and the corresponding sensing performance of the nanocomposite sensor under different strains. The sensor shows good stretchability enduring more than 100% strain (Fig. 5d). Generally, stretching would generate microcracks of the sensing elements and elongate and even disrupt the conduction pathway of ions, which increases the total electrical resistance. However, due to the prestrain-induced wrinkle structure and the interconnected hydrogel/CS nanoparticle network that facilitates ion conduction, the nanocomposite sensor experiences little influence on its sensing ability even under strain as large as 100% (Fig. 5e). Similar to the bending test, the overall current response slightly decreases under large strain, but the difference is not substantial. The impact of strain on the current response is relatively more obvious at lower RH, which is possibly due to the swelling of hydrogel at higher RH that partially compensates the deformation. The sensitivity of the nanocomposite sensor only drops from 42.35 nA/% to 35.2 nA/%, while the response/recovery time changes from 13/0.48 s to 28.8/0.35 s (Fig. 5f), suggesting outstanding performance of the sensor even under large strain.

The stability and durability of the nanocomposite sensor under constant deformation were also examined. The sensor first underwent repeated cycling of RH between 11% and 96% under the initial, 180° bending, and 100% strain states. The current response is stable both under the initial and deformed states (Fig. 5g). The hysteresis of the sensor under bending and strain are 1.14% and 1.30%, respectively, both smaller than that under the initial state (Fig. S6, ESI†). This is explained by the fewer water molecules retained in the hierarchical structure when the wrinkles are flattened under deformation. We then investigated the mechanical durability of the sensor by measuring sensor performance under repeated bending to 180° and stretching to 100% for up to 1000 times under ∼60% RH (Fig. 5h). During the cyclic bending and stretching, the response current stably cycles between 34 to 38 nA and 33 to 39 nA, respectively. The excellent stability of the current response suggests that the current pathway was not disrupted even under extreme and repeated deformations, proving its superior mechanical durability. In fact, the strain at break is up to 139% for the pVE/CS/PDMS-150% sensor from the stress–strain test (Fig. S7, ESI†). The sensor is also durable against repeated scratching by a pair of tweezers (Fig. S8, ESI†). We attribute the observed mechanical durability of the sensor to the wrinkle structure and an interpenetrating network between the hydrogel coating and PDMS. During iCVD, monomers and radicals could penetrate into the PDMS surface and polymerize to form a crosslinked hydrogel network.⁵⁸ This thin layer of crosslinked hydrogel at the near surface of PDMS film constitutes an interpenetrating network with PDMS, which generates strong bonding at the hydrogel/PDMS interface. The interpenetrating network along with the conformal wrapping by the pVE coating also anchors CS nanoparticles, preventing potential collapse and detachment from the surface.

Although many studies on wearable humidity sensors have been reported, limited flexibility and stretchability and rather slow response/recovery over a wide RH range are the commonly encountered problems. These issues stem from the stiffness, rather large thickness, and relatively small specific surface area of the sensing elements. Table 1 compares the performance of the nanocomposite sensor with some of the flexible humidity sensors from the literature with the best reported sensing and mechanical flexibility performance. Humidity sensors made from various sensing materials and their composites have been reported with high sensitivity and good stretchability. Their stretchability usually relies on the elasticity of the hydrogel or other polymers in the sensing elements, but relatively thick polymer components and limited specific surface area often lead to rather slow response/recovery of the sensor. Faster response/recovery has been reported on a partial range of RH or during respiration monitoring, but not within the whole RH spectrum. Mechanical durability is another important feature required by practical application of wearable sensors, which, however, was often ignored in previous studies. The reported pVE/CS/PDMS-150% sensor thus stands out with its high sensitivity, fast response/recovery over a wide RH range, and excellent stretchability and mechanical durability. These merits are attributed to the unique micro–nano hierarchical structure and the ultrathin hydrogel/CS nanocomposite coating, which accommodate fast adsorption and desorption of abundant water molecules. The sensor also endures repeated bending, stretching, and scratching while maintaining a stable sensing performance, indicating its great potential in practical applications.

---

Furthermore, wearable sensors may constantly be exposed to water, sweat, and bacteria, which could not only undermine the sensing performance but damage the sensor structure.⁶⁵ Accumulation of bacteria on the surface may also harm human health. Here we address this issue by encapsulating the nanocomposite sensor with an elastic fabric coated with hydrophobic p(PFDA-co-EGDA) via iCVD. Upon encapsulation, the sensor exhibits superhydrophobicity and even superoleophobicity, manifested by contact angles of >150° towards water, HCl and NaOH solutions, peanut oil, glycerol, and common drinks such as milk, cola, and tea (Fig. 6a and b). The superhydrophobic sensor also exhibits a good self-cleaning ability (Fig. 6c).

Upon exposure to both Gram-negative Escherichia coli (E. coli) and Gram-positive Staphylococcus aureus (S. aureus) bacterial solutions, the sensor demonstrates potent antibacterial capability with an antifouling rate of more than 97% for both (Fig. 6d and e). SEM observation reveals accumulation of bacteria adhered on fibers of pristine fabric but not on the coated fiber surface. Due to the solvent-free process and conformal nature of the iCVD technique, the morphology and porosity of the fabric were not altered after deposition (Fig. S9, ESI†); therefore, liquid cannot penetrate through the fabric, but water vapor can be freely transmitted. Indeed, when dripping both water and artificial sweat on the sensor, the response current is not altered, indicating the good water-proof ability of the sensor (Fig. 6g). In addition, the water-proof performance of the sensor is also not affected under repeated bending and stretching (Fig. 6h and Movie S1, ESI†). These results demonstrate excellent resistance toward water, sweat, and bacteria of the encapsulated sensor, suggesting applicability towards real-world applications.

Human physiological monitoring

The high sensitivity, fast response/recovery, and excellent flexibility and durability of the pVE/CS/PDMS-150% nanocomposite sensor suggest its great potential in real-time human physiological monitoring. We first used the sensor for human respiration monitoring by embedding it in an oxygen mask (Fig. 7a). The variation in the water vapor concentration of exhaled and inhaled air changes the resistance of the nanocomposite, which leads to a fluctuation in current. The humidity increases during exhalation, resulting in a rise in current; and the humidity decreases during inhalation, resulting in a drop in current. The respiratory rate of normal adults is 12–20 BPM (Breaths Per Minute). A rapid breathing rate greater than 24 BPM is common in hyperthyroidism and pneumonia pulmonary embolism; while a slow breathing rate less than 10 BPM is common in conditions of shock and increased intracranial pressure. The response/recovery time required for such high-frequency breathing is therefore much shorter than that needed for environmental humidity monitoring. In Fig. 7b, the respiratory frequency can be clearly distinguished between slow breathing (6 BPM), normal breathing (16 BPM), and rapid breathing (30 BPM), proving the capability of the nanocomposite humidity sensor in real-time human respiration monitoring.

We further demonstrate the use of the pVE/CS/PDMS-150% sensor for real-time respiratory monitoring during different motion states of the human body (Fig. 7c). According to the breathing frequency and intensity, the three motion states of lying down, walking, and running can be easily identified. The drift of the baseline during slow breathing is due to the gradual increase of the water molecule content in the oxygen mask. As the respiratory rate increases, the rapid inhalation aids the desorption of water molecules, allowing the baseline to drop and stabilize. The sensor was then used to detect sleep apnea syndrome (SAS), a sleep disorder in which breathing stops during sleep resulting in serious potential health problems. The nanocomposite sensor exhibits a fast response and stable electric signal during suffocation (Fig. 7d), suggesting the possibility of timely intervention when this status is detected and fed back to the alarm system. The fast response/recovery and high sensitivity of the nanocomposite sensor thus indicate its potential as an effective means in real-time monitoring of human respiratory behavior and assessment of human health status.

The humidity of human skin is an important indicator for many health conditions, ranging from daily skincare to different dermatitis and wound healing status. Owing to its excellent flexibility and stretchability, the pVE/CS/PDMS-150% humidity sensor can be comfortably attached to different parts of human skin during motion (Fig. 8). By encapsulating the nanocomposite within the superhydrophobic fabric, sweat and other liquids can be blocked from reaching the sensor surface while water vapor can still be freely transmitted and collected; thereby ensuring the intactness and unaffected sensing performance of the sensor (Fig. 8a). Fig. 8b and c depict the current change of the sensor upon washing and applying cream on cheek and neck skin, respectively. Both the response currents on the cheek and neck skin exhibit an obvious trend with an initial rise and then fall after washing and applying cream. The short increase in the current after washing is due to the evaporation and accumulation of residue water on the skin surface. As the residue water evaporates, the current gradually drops and even falls below the initial value eventually, indicating loss of moisture after washing without applying cream. Cream obviously maintains much more moisture on the skin surface, as it not only generates a higher current response after application, but also maintains the response current above the initial current after more than 2 min.

During exercise, the human body dissipates heat by evaporating water from the skin surface to regulate the body temperature. Upon attaching the nanocomposite sensor on the forehead, elbow, and palm, the evaporation of water from these parts during and after exercise (squat) can be clearly observed from their response current. Sensors on all parts detected the rise in current during exercise and a short period after the exercise stopped; the current then gradually falls and finally approaches the initial value. The moisture accumulation at different parts is also different during exercise, which leads to different changes of scale in the current. Water evaporates less at the elbows, so the current change is relatively small. The distribution of sweat glands on the forehead and palms is denser; therefore, the signal received by the humidity sensor is stronger. The successful and in-time capture of humidity variation is largely due to the high sensitivity and fast response/recovery of the nanocomposite sensor. It is worth noting that humidity variation in both human respiratory behavior and skin condition occur within seconds; it is therefore essential to maintain the response and recovery time of humidity sensors within a similar timescale. These results thus demonstrate that the prepared pVE/CS/PDMS-150% humidity sensor can be effectively applied to monitor human physiological behavior in real time.

Conclusions

In summary, we developed an ultrathin hydrogel/CS nanocomposite film on PDMS with a unique micro–nano hierarchical topography using a two-step solvent-free CVD process. The nanocomposite consists of candle soot nanoparticles conformally deposited with iCVD-deposited ultrathin hydrogel coatings with an ordered microscale wrinkle structure. Due to the large specific surface area from the nanoporous hierarchical topography, humidity sensors made from the nanocomposite exhibit high sensitivity and fast response/recovery. The wrinkle structure enables high stretchability of the sensor, while the conformal hydrogel nanocoating and interpenetrating network between the hydrogel and PDMS anchor CS nanoparticles impart good durability. Encapsulating the nanocomposite sensor with a superhydrophobic fabric provides the sensor with water-proof and antibacterial properties. The sensor is successfully used for both real-time human respiratory and skin humidity monitoring, indicating its practical potential for application as a durable wearable multifunctional humidity sensor.

Experimental section

Materials

Monomer VP was purchased from TCI, and PFDA was purchased from Jinan Guochentaifu Chemical. Crosslinker EGDA and initiator TBP were purchased from Aladdin. Silicon wafer (MEMC Electronic Materials) was used as the reference substrate for both hydrophilic and hydrophobic iCVD coatings, while PDMS films (SYLGARD^TM) were used as the substrate for hydrophilic coating and spandex fabric was used as the substrate for hydrophobic coating.

Preparation of hierarchical hydrogel–carbon nanocomposites

Homemade PDMS films (thickness: ∼1 mm) were quickly skimmed over a candle flame to obtain a light gray CS coating. The CS/PDMS film was then stretched to different strains (0%, 50%, 100%, and 150%) and fixed in a custom-made iCVD reaction chamber^66,67 along with a reference silicon wafer. During deposition, an alloy filament array (Ni80/Cr20) was placed on top of the substrate and heated to ∼220 °C. The initiator TBP, crosslinker EGDA, and monomer VP were vaporized at 30 °C, 60 °C, and 80 °C, respectively, and fed into the reaction chamber. Their flow rates were controlled using needle valves (Swagelok). The pressure of the chamber was maintained at 300 mTorr through a butterfly valve (VAT). Substrate temperature was kept at 35 °C via a circulating water cooling system at the backside of the stage. The thickness of the coating on the reference Si wafer was monitored in situ by a 633 nm He–Ne laser (JDS Uniphase) interferometry system. After coating of the pVE hydrogel, the PDMS was released and retracted to obtain wrinkled pVE/CS/PDMS nanocomposite films.

Hydrophobic nanocoating on fabric

p(PFDA-co-EGDA) coating was deposited on an elastic spandex fabric using the same iCVD reactor. During deposition, the filament and substrate temperatures were maintained at 220 °C and 44 °C, respectively. The flow rates were controlled at 0.24 sccm (PFDA), 0.06 sccm (EGDA), and 0.4 sccm (TBP). The deposition pressure was controlled at 200 mTorr. The coating thickness on the reference wafer was approximately 200 nm as estimated from the interferometer.

Preparation and test of the humidity sensor

Silver paste was applied on the prepared nanocomposite films as electrodes with an electrode spacing of 1 cm. The electrodes were heated to 50 °C in an oven for 30 minutes to solidify. The resultant sensor was then encapsulated under the superhydrophobic fabric using adhesive for water-proofing and self-cleaning purposes. The humidity sensing performance was measured using an electrochemical workstation (CS2350H, Corrtest). Written consent was obtained from all participants in the respiration and skin humidity sensing tests prior to the research.

Characterization

FTIR spectra of the polymer coatings were collected using a FTIR spectrometer (Nicolet 6700). SEM images were obtained using field emission SEM (Nova Nano SEM 450). XPS experiments were carried out using a Thermo Scientific K-Alpha instrument. BET experiments were carried out on a fully automated BET specific surface analyzer (Micromeritics ASAP 2460). A contact angle goniometer (Kruss DSA 100) was used to measure the static contact angles. The stress–strain test was conducted using a universal testing machine (KD Series) at a speed of 50 mm/min.

Antibacterial test

E. coli and S. aureus were used as the model bacteria for the antibacterial test of the sensor. Pristine and coated fabrics with a size of 4 × 4 cm² were exposed to 10 mL of bacterial solution (∼10⁶ CFU mL⁻¹) and incubated at 37 °C for 2 h. The unadhered bacteria were rinsed off from fabric with phosphate buffer solution (PBS). The resultant fabrics were then placed in a centrifuge tube containing 10 mL of PBS and ultrasonicated for 10 s to release the adhered bacteria. 5 μL of the resultant bacterial solution was taken out and scraped onto Luria broth (LB) agar medium with a spatula and incubated at 37 °C for 12 h. The antibacterial rate of each sample was calculated as (U–C)/U × 100%, where U and C are the bacterial colony counts for uncoated and coated fabrics, respectively. The error bars indicate the standard deviation from at least three measurements.

Conflicts of interest

The authors have no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (51873093, 62104121), the Natural Science Foundation of Zhejiang Province (LY20E030003, LQ22E020002), and the Ningbo Science & Technology Innovation Research Program (2021Z058).

References

1. S. Li, Y. Zhang, X. Liang, H. Wang, H. Lu, M. Zhu, H. Wang, M. Zhang, X. Qiu, Y. Song and Y. Zhang, Nat. Commun., 2022, 13, 5416 CrossRef CAS PubMed .
2. Y. Wang, C. Zhao, J. Wang, X. Luo, L. Xie, S. Zhan, J. Kim, X. Wang, X. Liu and Y. Ying, Sci. Adv., 2021, 7, 4553 CrossRef PubMed .
3. M. Lin, Z. Zheng, L. Yang, M. Luo, L. Fu, B. Lin and C. Xu, Adv. Mater., 2022, 34, 2107309 CrossRef CAS PubMed .
4. C. Raymond, T. Matthews and R. M. Horton, Sci. Adv., 2020, 6, 1838 CrossRef PubMed .
5. F. Gueder, A. Ainla, J. Redston, B. Mosadegh, A. Glavan, T. J. Martin and G. M. Whitesides, Angew. Chem., Int. Ed., 2016, 55, 5727–5732 CrossRef CAS PubMed .
6. S. Li, Y. Zhang, Y. Wang, K. Xia, Z. Yin, H. Wang, M. Zhang, X. Liang, H. Lu, M. Zhu, H. Wang, X. Shen and Y. Zhang, InfoMat, 2020, 2, 184–211 CrossRef CAS .
7. J. Yang, R. Shi, Z. Lou, R. Choi, K. Jiang and G. Shen, Small, 2019, 15, 1902801 CrossRef PubMed .
8. Y. Lu, G. Yang, Y. Shen, H. Yang and K. Xu, Nano-Micro Lett., 2022, 14, 150 CrossRef CAS PubMed .
9. C. Wang, K. Xia, H. Wang, X. Liang, Z. Yin and Y. Zhang, Adv. Mater., 2019, 31, 1801072 CrossRef PubMed .
10. X. J. Liu, M. Gao, J. Y. Chen, S. Guo, W. Zhu, L. C. Bai, W. Z. Zhai, H. J. Du, H. Wu, C. Z. Yan, Y. S. Shi, J. W. Gu, H. J. Qi and K. Zhou, Adv. Funct. Mater., 2022, 32, 2203323 CrossRef CAS .
11. J. Wu, Z. Wu, H. Xu, Q. Wu, C. Liu, B.-R. Yang, X. Gui, X. Xie, K. Tao, Y. Shen, J. Miao and L. K. Norford, Mater. Horiz., 2019, 6, 595–603 RSC .
12. C. Jung, S.-J. Kim, J. Jang, J. H. Ko, D. Kim, B. Ko, Y. M. Song, S.-H. Hong and J. Rho, Sci. Adv., 2022, 8, 8598 CrossRef PubMed .
13. Z. Bai, X. Wang, M. Zheng, O. Yue, M. Huang, X. Zou, B. Cui, L. Xie, S. Dong, J. Shang, G. Gong, A. M. Blocki, J. Guo and X. Liu, Adv. Funct. Mater., 2023, 33, 2212856 CrossRef CAS .
14. S. Zeng, J. Y. Zhang, G. Q. Zu and J. Huang, Carbohydr. Polym., 2021, 267, 118198 CrossRef CAS PubMed .
15. S. Hasan, A. Z. Kouzani, S. Adams, J. Long and M. A. P. Mahmud, Sens. Actuators, A, 2022, 335, 113382 CrossRef CAS .
16. B. B. Ying, Q. Y. Wu, J. Y. Li and X. Y. Liu, Mater. Horiz., 2020, 7, 477–488 RSC .
17. Z. R. He, Z. X. Zhou and W. Z. Yuan, ACS Appl. Mater. Interfaces, 2022, 14, 38205–38215 CrossRef CAS PubMed .
18. X. L. Zou, X. C. Wang, Z. X. Bai, O. Y. Yue, C. Wei, L. Xie, H. J. Zhang and X. H. Liu, Chem. Eng. J., 2023, 463, 142349 CrossRef CAS .
19. P. He, R. Guo, K. Hu, K. Liu, S. Lin, H. Wu, L. Huang, L. Chen and Y. Ni, Chem. Eng. J., 2021, 414, 128726 CrossRef CAS .
20. J. Yu, Y. Feng, D. Sun, W. Ren, C. Shao and R. Sun, ACS Appl. Mater. Interfaces, 2022, 14, 10886–10897 CrossRef CAS PubMed .
21. Z. Lei and P. Wu, Nat. Commun., 2018, 9, 1134 CrossRef PubMed .
22. Z. Lei, Q. Wang and P. Wu, Mater. Horiz., 2017, 4, 694–700 RSC .
23. Z. Wu, Q. Ding, Z. Li, Z. Zhou, L. Luo, K. Tao, X. Xie and J. Wu, Sci. China: Mater., 2022, 65, 2540–2552 CAS .
24. T. Kim, J. W. Lee, C. Park, K. Lee, C. E. Lee, S. Lee, Y. Kim, S. Kim, S. Jeon, D. Y. Ryu, W.-G. Koh and C. Park, Nano Energy, 2022, 92, 106688 CrossRef CAS .
25. J. Wu, Y. Liang, Z. Zhou, H. Ding, W. Huang, K. Tao, W. Shi, B.-R. Yang and X. Xie, Sens. Actuators, B, 2021, 344, 130259 CrossRef CAS .
26. Y. Gao, F. Jia and G. Gao, Chem. Eng. J., 2022, 430, 132919 CrossRef CAS .
27. Y. Liang, Q. Ding, H. Wang, Z. Wu, J. Li, Z. Li, K. Tao, X. Gui and J. Wu, Nano-Micro Lett., 2022, 14, 183 CrossRef CAS PubMed .
28. H. Moon, H. Seong, W. C. Shin, W.-T. Park, M. Kim, S. Lee, J. H. Bong, Y.-Y. Noh, B. J. Cho, S. Yoo and S. G. Im, Nat. Mater., 2015, 14, 628–635 CrossRef CAS PubMed .
29. H. S. Suh, D. H. Kim, P. Moni, S. Xiong, L. E. Ocola, N. J. Zaluzec, K. K. Gleason and P. F. Nealey, Nat. Nanotechnol., 2017, 12, 575–581 CrossRef CAS PubMed .
30. C. Su, X. Huang, L. Zhang, Y. Zhang, Z. Yu, C. Chen, Y. Ye and S. Guo, Nano Energy, 2023, 107, 108095 CrossRef CAS .
31. D. Soto, A. Ugur, T. A. Farnham, K. K. Gleason and K. K. Varanasi, Adv. Funct. Mater., 2018, 28, 1707355 CrossRef .
32. X. Shi, Y. Ye, H. Wang, F. Liu and Z. Wang, ACS Appl. Mater. Interfaces, 2018, 10, 38449–38458 CrossRef CAS PubMed .
33. P. Chen, J. Lang, Y. Zhou, A. Khlyustova, Z. Zhang, X. Ma, S. Liu, Y. Cheng and R. Yang, Sci. Adv., 2022, 8, 8812 CrossRef PubMed .
34. Y. Ye, Y. Mao, H. Wang and Z. Ren, J. Mater. Chem., 2012, 22, 2449–2455 RSC .
35. H. Sojoudi, S. Kim, H. Zhao, R. K. Annavarapu, D. Mariappan, A. J. Hart, G. H. McKinley and K. K. Gleason, ACS Appl. Mater. Interfaces, 2017, 9, 43287–43299 CrossRef CAS PubMed .
36. N. J. Trujillo, Q. Wu and K. K. Gleason, Adv. Funct. Mater., 2010, 20, 607–616 CrossRef CAS .
37. M. R. Mulay, A. Chauhan, S. Patel, V. Balakrishnan, A. Halder and R. Vaish, Carbon, 2019, 144, 684–712 CrossRef CAS .
38. Z. Wei, K. Yan, H. Chen, Y. Yi, T. Zhang, X. Long, J. Li, L. Zhang, J. Wang and S. Yang, Energy Environ. Sci., 2014, 7, 3326–3333 RSC .
39. D. Wang, Q. Sun, M. J. Hokkanen, C. Zhang, F. Y. Lin, Q. Liu, S. P. Zhu, T. Zhou, Q. Chang, B. He, Q. Zhou, L. Chen, Z. Wang, R. H. A. Ras and X. Deng, Nature, 2020, 582, 55–59 CrossRef CAS PubMed .
40. M. Sun, Q. Y. Wu, J. Xu, F. He, A. P. Brown and Y. M. Ye, J. Mater. Chem. B, 2016, 4, 2669–2678 RSC .
41. C. Su, Y. Hu, Q. Song, Y. Ye, L. Gao, P. Li and T. Ye, ACS Appl. Mater. Interfaces, 2020, 12, 18978–18986 CrossRef CAS PubMed .
42. P. Su, W. Liu, Y. Hong, Y. Ye and S. Huang, Mater. Today Chem., 2022, 24, 100786 CrossRef CAS .
43. P.-Y. Chen, M. Liu, Z. Wang, R. H. Hurt and I. Y. Wong, Adv. Mater., 2017, 29, 1605096 CrossRef PubMed .
44. N. Liu, Q. Sun, Z. Yang, L. Shan, Z. Wang and H. Li, Adv. Sci., 2023, 10, 2207210 CrossRef CAS PubMed .
45. F. Xu, X. Wang, Y. Zhu and Y. Zhu, Adv. Funct. Mater., 2012, 22, 1279–1283 CrossRef CAS .
46. Z. Niu, H. Dong, B. Zhu, J. Li, H. H. Hng, W. Zhou, X. Chen and S. Xie, Adv. Mater., 2013, 25, 1058–1064 CrossRef CAS PubMed .
47. Y. Yang and Z. D. Deng, Appl. Phys. Rev., 2019, 6, 011309 Search PubMed .
48. D. Toan, N. Thanh, P. Hoang-Phuong, N. Nam-Trung, D. Dzung Viet and J. Bell, Biosens. Bioelectron., 2020, 166, 112460 CrossRef PubMed .
49. T. Li, L. Li, H. Sun, Y. Xu, X. Wang, H. Luo, Z. Liu and T. Zhang, Adv. Sci., 2017, 4, 1600404 CrossRef PubMed .
50. Z. Wu, X. Yang and J. Wu, ACS Appl. Mater. Interfaces, 2021, 13, 2128–2144 CrossRef CAS PubMed .
51. X. Guan, Y. Yu, Z. Hou, K. Wu, H. Zhao, S. Liu, T. Fei and T. Zhang, Sens. Actuators, B, 2022, 358, 131438 CrossRef CAS .
52. W. Jeong, J. Song, J. Bae, K. R. Nandanapalli and S. Lee, ACS Appl. Mater. Interfaces, 2019, 11, 44758–44763 CrossRef CAS PubMed .
53. J. Dai, H. Zhao, X. Lin, S. Liu, Y. Liu, X. Liu, T. Fei and T. Zhang, ACS Appl. Mater. Interfaces, 2019, 11, 6483–6490 CrossRef CAS PubMed .
54. Z. Zhen, Z. Li, X. Zhao, Y. Zhong, L. Zhang, Q. Chen, T. Yang and H. Zhu, Small, 2018, 14, 1703848 CrossRef PubMed .
55. J. Qin, X. Yang, C. Shen, Y. Chang, Y. Deng, Z. Zhang, H. Liu, C. Lv, Y. Li, C. Zhang, L. Dong and C. Shan, Nano Energy, 2022, 101, 107549 CrossRef CAS .
56. Y. Yi, C. Yu, H. Zhai, L. Jin, D. Cheng, Y. Lu, Z. Chen, L. Xu, J. Li, Q. Song, P. Yue, Z. Liu and Y. Li, Nano Energy, 2022, 103, 107780 CrossRef CAS .
57. L. Li, H. Xiang, Y. Xiong, H. Zhao, Y. Bai, S. Wang, F. Sun, M. Hao, L. Liu, T. Li, Z. Peng, J. Xu and T. Zhang, Adv. Sci., 2018, 5, 1800558 CrossRef PubMed .
58. K. Jeong, Y. Lee, Y. Kim, H. Mun, K.-U. Kyung and S. G. Im, Chem. Eng. J., 2022, 429, 132250 CrossRef CAS .
59. S. Li, T. Wan, H. Wei, S. Wang, B. Wang and B. Cheng, Sens. Actuators, B, 2022, 362, 131806 CrossRef CAS .
60. Z. He, G. Zhou, Y. Oh, B. M. Jung, M. K. Um, S. K. Lee, J. I. Song, J. H. Byun and T. W. Chou, Mater. Today Nano, 2022, 18, 100214 CrossRef CAS .
61. Y. Komazaki and S. Uemura, Sens. Actuators, B, 2019, 297, 126711 CrossRef CAS .
62. H. Xing, X. Li, Y. Lu, Y. Wu, Y. He, Q. Chen, Q. Liu and R. P. S. Han, Sens. Actuators, B, 2022, 361, 131704 CrossRef CAS .
63. L. Ma, R. Wu, A. Patil, S. Zhu, Z. Meng, H. Meng, C. Hou, Y. Zhang, Q. Liu, R. Yu, J. Wang, N. Lin and X. Y. Liu, Adv. Funct. Mater., 2019, 29, 1904549 CrossRef CAS .
64. L. Zhu, X. Li, T. Kasuga, K. Uetani, M. Nogi and H. Koga, J. Mater. Chem. C, 2022, 10, 3712–3719 RSC .
65. Y. Liu, X. Li, H. Yang, P. Zhang, P. Wang, Y. Sun, F. Yang, W. Liu, Y. Li, Y. Tian, S. Qian, S. Chen, H. Cheng and X. Wang, ACS Nano, 2023, 17, 5588–5599 CrossRef CAS PubMed .
66. X. Huang, M. Sun, X. Shi, J. Shao, M. Jin, W. Liu, R. Zhang, S. Huang and Y. Ye, Chem. Eng. J., 2023, 454, 139981 CrossRef CAS .
67. Q. Song, R. Zhao, T. Liu, L. Gao, C. Su, Y. Ye, S. Y. Chan, X. Liu, K. Wang, P. Li and W. Huang, Chem. Eng. J., 2021, 418, 129368 CrossRef CAS PubMed .

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3mh01093g

This journal is © The Royal Society of Chemistry 2023