Superhydrophilic CuO nanowire QCM humidity sensor with horsefly-inspired self-cleaning ability applied for non-contact detection

Abstract

Humidity sensors are crucial for a variety of applications, such as biomedicine and semiconductor fabrication. It is still a formidable challenge to achieve a humidity sensor with high sensitivity and anti-pollution ability. Inspired by the horsefly eye structure, we propose a superhydrophilic CuO nanowire quartz crystal microbalance humidity sensor covered by a micro-pit array, which can achieve outstanding self-cleaning properties. The physicochemical property of the sensing materials was characterized in situ using SEM, XRD, and XPS. The sensing characteristics of the specimen towards water molecules were studied using quartz crystal microbalance with dissipation monitoring, showing a high sensitivity of 82.5 ± 7.7 Hz/%RH. Molecular dynamics simulation and Brunauer–Emmett–Teller nitrogen sorption analysis were conducted for revealing the superhydrophilicity mechanism of CuO nanowires. Subsequently, a CuO nanowire humidity-sensing model was established, which was verified by both theoretical calculations and experiments. A smart non-contact sensing system based on the sensor was constructed, and the humidity intensity of human breath and fingers were detected. This work demonstrates the brilliant properties of superhydrophilic CuO nanowires for non-contact sensing applications, providing new solutions for medical health monitoring, industrial environment detection, and human-computer interaction.

---

1. Introduction

Humidity sensors are vital for medicine,¹ food storage, agriculture,² semiconductor fabrication, and electronics industries.³ Recently, it was reported that a humidity sensor can also be applied to human–machine interaction for non-contact sensing.⁴ The non-contact sensing mode can reduce the risk of bacterial and virus transmission infections in conventional touch screens. From this perspective, further exploration on smarter non-contact sensing and control systems based on humidity sensors is of great significance. In practical application, humidity sensors are invariably exposed to air and become dusty. Achieving anti-pollution ability and high sensitivity⁵ remains central issues in the research of humidity sensors. These properties mostly depend on the detecting mechanism, sensing material, and the surface structure of sensors.

One of the ultrasensitive humidity detector is based on the quartz crystal microbalance (QCM). The detection mechanism of QCM is based on piezoelectric effect and can provide nanogram precision.⁶ It possesses multiple merits such as superior stability and self-recovery ability when being wet. For providing high sensitivity, coating the sensor with proper sensing materials is particularly important. To this end, numerous materials containing two-dimensional materials (such as MoS₂⁷ and graphene oxide⁸), metallic oxides (such as ZnO,⁹ SiO₂,¹⁰ and SnO₂¹¹), polymers,¹² metal–organic frameworks,¹³ and fiber composites (such as CuO/reduced graphene oxide,¹⁴ CuO/ZnO,¹⁵ and Cu(OH)₂/graphene oxide¹⁶) have been introduced and studied. Researchers have made some progress in humidity sensing using these materials. Nevertheless, in some cases, plentiful sensing materials (such as polymers and salts) exhibit swelling when being wet,¹⁷ resulting in the damage of the sensor. This problem can be solved preliminarily by endowing the sensing material with superhydrophilicity. Droplets will spread and self-evaporate quickly in several seconds on a superhydrophilic surface, which is called self-healing ability.⁶ For example, Cu(OH)₂ nanowires exhibit excellent superhydrophilicity, possessing self-recovery ability when being wet, and offer high-humidity-sensing performances. However, they suffer from the drawback of thermal decomposition when the temperature is over 150 °C,¹⁸ making them impractical for high temperatures and hampering their wider practical applications. Therefore, developing sensing materials with both superhydrophilicity and high-temperature stability is necessary for high-performance humidity sensors.

On account of the superior thermal stability of metal oxides, a CuO-based resistive humidity sensor has been fabricated.^19–21 The traditional fabrication methods of CuO include spraying,²² syringe dispensing,²³ laser sintering,²⁴in situ growth,²⁵ and annealing from Cu(OH)₂ nanowires.¹⁴ Among them, the in situ growth can obtain superior mechanical stability and maintain the unique nanostructure morphology,²⁶ obtaining a high aspect ratio. However, these CuO nanostructures are always fabricated based on Cu foil²⁷ or in the form of sediment²⁸ in the solution. The in situ growth of CuO nanostructures on QCM has never been achieved. Obviously, relying on copper substrate²⁹ and exfoliation of CuO nanowires from the substrate are the main problems.¹⁵ Thus, an appropriate fabrication process for in situ fabricating CuO nanostructures on non-Cu substrates with satisfied adhesion is highly deserved. To address this issue, exploring how to grow CuO nanostructures in situ on QCM is essential for constructing a highly sensitive and stable humidity sensor.

Thus far, the influence of annealing temperature on the humidity sensing properties of CuO nanostructures¹⁹ has been studied. However, the superhydrophilicity and humidity sensing mechanism of CuO nanostructures have not been clearly revealed. Moreover, to further improve the stability of humidity sensors, anti-pollution ability is expected.³⁰ Thus, it remains a huge challenge to realize elaborate control of the morphology of CuO nanostructures on QCM.

Herein, we demonstrate a CuO nanowire QCM sensor prepared by a liquid–solid reaction and annealing method. This sensor is easy to prepare and cost-effective. Multi-harmonic quartz crystal microbalance with dissipation monitoring (QCM-D) was used to detect the transition of the sensors at various relative humidity (RH) levels. The sensing results revealed that the CuO nanowire humidity sensor exhibits small swelling changes even at high RH levels, demonstrating satisfactory humidity sensing properties. Furthermore, a micro-pit array self-cleaning layer was constructed on the sensor surface for improving the anti-pollution ability. When a droplet drops down on the surface, it will bounce and take the dust away from the sensor surface. The superhydrophilicity and humidity sensing mechanism were revealed by theoretical calculations and experiments, shedding new light on the interaction between water molecules and CuO nanowires. Strikingly, human breath and fingers can be detected in real time by a homemade smart non-contact sensing and control system, demonstrating the prospect of the CuO nanowire QCM humidity sensors for detecting physiological activity.

2. Experimental section

2.1. Fabrication of CuO nanowires

Absolute ethyl ethanol (C₂H₅OH) and acetone (C₃H₆O) were purchased from Sinopharm Chemical Reagent Co. Ltd (Shanghai, China). Sodium hydroxide (NaOH) and ammonium persulfate ((NH₄)₂S₂O₈) were supplied by Shanghai Aladdin Bio/Chem Technology Co., Ltd (Shanghai, China). QCM with a fundamental resonance frequency of 8 MHz, consisting of two silver electrodes with a diameter of 5 mm on both sides of the AT-cut quartz crystal (Fig. S1a, ESI†), was provided by Hi-trusty Co., Ltd, China. Another QCM with a fundamental resonance frequency of 5 MHz, consisting of a gold electrode with a diameter of 10 mm of the AT-cut quartz crystal (Fig. S1b, ESI†), was furnished by Biolin Scientific Co., Ltd, China.

The routine flow protocols for the fabrication of CuO nanowire QCMs are shown in Fig. S2 (ESI†). In this work, we mainly focused on the fabrication of the sensor with a micro-pit array self-cleaning layer, which was inspired from horsefly eye structure (Scheme 1). Before the lithography and deposition of the Cu layer, the original QCMs were pretreated with acetone, ethanol, and deionized water for 5 min in an ultrasonic bath followed by drying in a stream of nitrogen (Scheme 1a). The precursor Cu layers with diverse thicknesses were deposited on the upside electrode of QCMs using an electron beam evaporator (Ebeam-500S) at a speed of 1 Å s⁻¹ under 2.5 × 10⁻⁴ Pa (Scheme 1b). Then, Cu(OH)₂ nanowires were grown in situ on the patterned Cu layer by immersing the sensor in NaOH (0.375 M) and (NH)₄S₂O₈ (0.015 M) aqueous solutions at 25 °C (Scheme 1c), followed by rinsing with ethanol and deionized water sequentially (Scheme 1d). The samples were annealed at different temperatures (100–900 °C with an interval of 100 °C) with a series of heating rates (1, 2.5, 5, and 10 °C min⁻¹) for several annealing durations (0, 1, 3, and 5 h) (Scheme 1e). Then lithography was employed to fabricate the micro-pit array self-cleaning layer on the QCM surfaces, as shown in Scheme 1f. Ultimately, superhydrophilic CuO nanowire QCMs with self-cleaning ability were obtained.

Scheme 1 Protocols of the fabrication of CuO nanowire QCMs: (a) cleaning bare QCMs, (b) preparing a patterned Cu layer via photolithography and deposition, (c) in situ growing of the Cu(OH)₂ nanowires on the as-prepared patterned Cu layer in solution, (d) cleaning the Cu(OH)₂ nanowires, (e) in situ annealing of the Cu(OH)₂ nanowires for obtaining CuO nanowires in air, and (f) fabrication of a micro-pit array via photolithography.

2.2. Characterization of the materials

In situ continuity characterization. X-ray diffraction using CuKα (XRD, Smartlab 9 kW PIGAKV) was exploited to identify the crystal texture annealed in situ. X-ray photoelectron spectroscopy (XPS, AXIS-ULTRA DLD-600W) with a vacuum of 2 × 10⁻⁸ Torr was used to study the chemical composition transformation by annealing in situ. A typical in situ XPS observation took about 48 h to increase the temperature from 25 °C to 400 °C, during which the data were recorded. Another Bruker V70 with a gas line controller was applied to characterize the functional groups of the sensing materials at various RH levels in situ. Thermogravimetric analysis was performed using a TGA8000 PerkinElmer in situ.

A scanning electron microscope (SEM, FEI Tecnai G280-300) operating at an accelerating voltage of 10 kV with an electric current of 0.17 nA was used for visualizing the morphologies of the sensing materials and micro-pit array self-cleaning layer. Transmission electron microscopy (TEM, Talos F200X) was implemented to analyze the lattice configuration. The element distribution was analyzed by energy-dispersive X-ray (EDX) elemental mapping using a single nanowire (in TEM). An XRD-7000 was used to study the crystal texture of the sensing materials. Fourier transform infrared (FTIR) spectra under 0.1 Pa were analyzed using a Vertex 70v with a vacuum-pumping system to clarify the original functional groups of the sensing materials. Raman spectroscopy was performed using a Via-Reflex Renishaw. Brunauer–Emmett–Teller (BET) nitrogen sorption analysis of the sensing materials was performed using a Micromeritics ASAP2020HD88. The water contact angle (WCA) was measured using a Dataphysics OCA35 contact-angle system at room temperature with a droplet volume of 0.4 μL. For the interaction among the droplet, dirt and sensor, the contact line motion (top-view) was recorded using an optical microscope (VHX-1000E) on an optical platform (HGZZ0906Z Wuhan Red Star Yang Technology Co. Ltd), while the shape evolution (axonometric-view) was captured using Acuteye High-Speed Image System V4.0 at 5000 fps.

2.3. Testing of the humidity sensor

A schematic diagram of the humidity testing system is illustrated in Fig. S3 (ESI†), which consists of humidity sources (from ∼0% to 97% RH), a CuO nanowire QCM humidity sensor, a QCM-D analyzer, and a PC. The QCM sensor was placed in a hermetic container in which constant humidity conditions were achieved by placing various saturated salt solutions inside the container. Frequency shift (Δf) accompanied by dissipation (ΔD) was obtained using QSense (Biolin, QVH401) with several overtones simultaneously. These change data from all overtones were used to gain more information on the viscoelastic properties upon swelling of the sensing materials exposed to different RH levels. The sensitivity of the sensor is defined as the ratio of Δf and relative humidity change (ΔRH). The relative sensitivity is defined as the ratio of the sensitivity of the sensor after annealing to that before annealing. The response/recovery time are defined as the time when the sensor frequency shift reaches 90% of the total shift. All the measurements were implemented at 25 ± 1 °C except for the QCM temperature resistance test.

2.4. Methods of the simulations

2.4.1. Molecular dynamics simulation. The system consists of 23 500 atoms (10 000 Cu, 10 000 O, 3000 H, and 1500 O) in an orthorhombic box with dimensions 127 × 121 × 23 ų containing eight CuO surface types under periodic boundary conditions. The simulations were conducted in the NVT ensemble at 298 K for 1 ns with an integration time step of 0.5 fs.

2.4.2. DFT calculations. A 4 × 4 surface supercell of CuO (16 Cu and 16 O) with a thickness of 10 Å was calculated with one or more H₂O under six positions (Fig. S5, ESI†). Water molecules and the surface CuO atoms were allowed to relax in three dimensions during geometry optimization by FORCITE. The adsorption energy E_ads was obtained as the difference between the total energy of the interacting system E(CuO + H₂O) and those of the separate sensing materials E(CuO) with the target molecule E(H₂O) after geometry optimization using the Cambridge Sequential Total Energy Package (CASTEP).

3. Results and discussion

3.1. Material characterizations

A couple of Cu(OH)₂ nanowires were in situ sintered at different temperatures to obtain CuO nanowires, and the corresponding SEM images are plotted in Fig. 1a and Fig. S6 (ESI†). The results indicate that the CuO nanowires almost keep the morphology of Cu(OH)₂ nanowires, while the width becomes narrower, when the annealing temperature is lower than 300 °C (Fig. S7, ESI†). A higher annealing temperature results in significant roughness on the CuO nanowire surfaces. When the annealing temperature exceeds 800 °C, the original nanowire structure no longer exists. In contrast, the furnace heating rate (1–10 °C min⁻¹) and the annealing durations (0–5 h) do not have a strong effect on the morphology of the CuO nanowires (Fig. S8, ESI†). Therefore, well-defined CuO nanowires can be produced by thermal dehydration of the Cu(OH)₂ nanowires in air directly, which is in proximity to the literature.²⁹ The nanowire surfaces possess obvious ravine-liked vein (Fig. 1b). Both the natural end (Fig. 1c) and the cut section (Fig. 1d) of the nanowire show that it is assembled by some smaller nanowires, creating the vein that enlarges the superficial roughness.

Fig. 1 Morphological characterizations of CuO nanowires: (a) SEM images of the nanowires annealed at different temperatures (200–800 °C), (b) enlarged SEM images of the nanowire surface, (c) natural end face of the nanowire, (d) cut section of the nanowire sliced by Focused Ion Beam; (e) TEM image of CuO nanowires annealed at 200 °C, (f) HRTEM image of the nanowires, (g) SAED images, elemental mapping of (h) original CuO nanowires, (i) Cu element, and (j) O element.

A series of TEM images of an individual nanowire before and after annealing at 200 °C are displayed in Fig. S9 (ESI†), and the EDS images are also given in Fig. S10 (ESI†). After annealing, the nanowire gets thinner in accordance with the SEM observations. Fig. 1e gives a typical TEM image of the annealed nanowires, where the vein can also be seen clearly. The HRTEM image shows a lattice fringe spacing of 2.33 Å assigned to (111) of CuO (Fig. 1f). The corresponding selected area electron diffraction (SAED) image indicates that the CuO nanowire is polycrystalline while possessing a trend of monocrystal (Fig. 1g). The elemental mappings of the nanowires are shown in Fig. 1h–j, verifying that the uniform CuO nanowires were successfully obtained by the dehydration of Cu(OH)₂ nanowires.

The in situ observation of the nanowires at increasing annealing temperatures by XRD is plotted in Fig. 2a. Before being annealing, the strong peaks centered at 2θ = 16.7°, 23.7°, 33.9°, 35.7°, 38.1°, and 53.2° are from the (002), (021), (002), (111), and (022) planes of orthorhombic Cu(OH)₂ (JCPDS card No. 03-0310), respectively. With the increase in the annealing temperature over 150 °C, all the peaks of Cu(OH)₂ nanowires disappear except the peaks of the substrate. After annealing at a higher temperature, the high-intensity peaks observed at 2θ = 35.45° and 38.73° correspond to (002), and (111) of CuO nanowires (JCPDS card No. 05-0661), respectively. Besides, the results showed that higher annealing temperatures, shorter annealing durations (Fig. 2b) and higher heating rates (Fig. 2c) do not have an obvious impact on the crystallization of CuO nanowires. It means that CuO nanowires are obtained successfully by annealed at 150 °C at a heating rate of 5 °C min⁻¹ for 1 hour. In a previous report,³¹ a small amount of Cu₂O was observed to be formed during the preparation of Cu(OH)₂ nanowires. The XRD peaks of the nanowires before annealing were compared with the standard cards of Cu(OH)₂ and Cu₂O, as shown in Fig. S11 (ESI†). The peaks corresponding to Cu(OH)₂ are significantly stronger than that of Cu₂O, and trace amounts of Cu₂O can be ignored.

Fig. 2 Material characterizations of CuO nanowires: XRD patterns of the nanowires before and after annealing at (a) different temperatures, (b) different times, and (c) different rates. XPS patterns of (d) copper element and (e) and (f) oxygen element in nanowires annealed at different temperatures. (g) FTIR spectra of the CuO nanowires in air and vacuum; (h) and (i) Raman spectra of the nanowires annealed at different annealing temperatures.

High-resolution XPS spectra of the sensing materials at different annealing temperatures are shown in Fig. 2d and e. With the increase in annealing temperatures, the peaks centered at 933.9 eV and 934.9 eV, accounting for the Cu 2p_3/2 signal of Cu(OH)₂ shift to 932.4 eV and 933.6 eV, which are related to the Cu 2p_3/2 signal of CuO. Similarly, the O 1s spectrum is resolved into two components centered at 528.5 eV, corresponding to the oxygen of CuO. All the detailed peak curves are plotted in Fig. S12 and S13 (ESI†). A typical XPS spectrum of the O element of the nanowires annealed at 200 °C is shown in Fig. 2f. The O 1s spectrum is resolved into two components centered at 528.5 eV, which corresponds to the oxygen of CuO.

The FTIR spectra of the CuO nanowires in air and a vacuum are illustrated in Fig. 2g. Compared to that in air, the curve in a vacuum is more smooth and distinctly due to less H₂O and CO₂, especially in the range of the gray rectangle marked. The band at 3500 cm⁻¹ stems from the Cu–O vibration. The bands at 3309.4 cm⁻¹ and 1631.2 cm⁻¹ correspond to the stretching and bending modes of the hydroxyls of adsorbed water. The band at 423.3 cm⁻¹ can be assigned to the Cu–O stretching mode and prove that CuO is formed. Typical Raman spectra of the CuO nanowires as a function of annealing temperatures (25 °C, 200 °C, 300 °C, 400 °C, and 500 °C) are shown in Fig. 2h and i. The Raman spectra reveal three main phonon modes of the CuO nanowires and the three peaks at 280.6 cm⁻¹, 332.3 cm⁻¹ and 623 cm⁻¹ correspond to the A_g, B_g¹, and B_g² of CuO nanowires, respectively.^23,32 The weight of the sensing materials as a function of temperature (30–1000 °C) was also recorded, as demonstrated in Fig. S14 (ESI†). It validated that ∼150 °C is the phase transformation point of the Cu(OH)₂ nanowires.

3.2. Sensing and self-cleaning properties

3.2.1. Sensing properties. The frequency shifts of CuO nanowire QCMs with various Cu layer thicknesses (100, 150, 200, 250, and 300 nm) and reaction times (5, 10, 15, 20, 25, and 30 min) are drawn into a contour map, as given in Fig. 3a. Taking the Cu layer of 200 nm as an example, as the reaction time increases, the resonance frequency shifts first increase to reach a maximum and then decrease (Fig. 3a1), regardless of whether the QCM is covered with sensing materials on a single side, both sides, or in growing types (Fig. S15, ESI†). For the reaction time of 15 min, the maximum frequency shift occurs when the Cu layer is 300 nm (Fig. 3a2). However, the sensor with these parameters is not preferred because of unstable production. It should be pointed out that the QCM deposited with Cu with 50 nm, 250 nm, and 300 nm may stop vibrating sometimes due to the destruction of devices or the heavy mass of sensing materials. In view of this, the Cu layer with 200 nm is preferred on account of large Δf with low standard deviation among other samples. Furthermore, the sensitivity enhances with the increase in annealing temperature up to 200 °C, while decreases along with the continuous increase in the temperature (Fig. 3b). Though the highest sensitivity occurs at 500 °C, the fabrication processes are not stable because the QCM may suffer the destruction of high temperature. Therefore, the QCM with 200 nm Cu layer for a reaction time of 15 min, annealed at 200 °C, is preferable for fabricating superior CuO nanowire QCM humidity sensors, which were selected for subsequent experiments to characterize the humidity sensing properties by QCM-D.

Fig. 3 Sensing properties of the CuO nanowire QCM sensor: (a) frequency shift (Δf) contour map of sensors fabricated under different reaction times and Cu layer thicknesses. (b) Pattern of relative sensitivity vs. annealing temperature of CuO nanowires. (c) Curve of ln Δf vs. relative humidity. (d) Graph of the sensor changes from 11% RH to 56% RH and repeats inversely for several times, where F is the frequency (Hz), D is the dissipation, and S represents the smooth curve. (e) Short-time stability of the sensor. (f) Frequency shift of the sensor when being wet with droplets. (g) Frequency shift of the sensor at different temperatures. The inset picture of the sensor at (g1) 200 °C and (g2) and (g3) 0 °C.

The ln Δf vs. relative humidity curve for water molecules adsorbed onto CuO nanowires is shown in Fig. 3c. The regression coefficient is 0.995, exhibiting a superior linear ln Δf to RH. The average sensitivity is 82.5 ± 7.7 Hz/%RH, which is superior to many previously reported works, as shown in Table S1 (ESI†). The fastest response and recovery times are 0.1815 s and 16.1732 s, respectively (Fig. S16, ESI†). No obvious fluctuations were observed when the sensor changes from 11% to 56% and repeats in reverse for several cycles (Fig. 3d), suggesting that the sensor possesses satisfying stability. The maximum frequency variation under 11% over 1 h is R = 16.0388 (x̄ = 69.9398, σ = 2.6698) and dissipation variation is R = −3.6593 (x̄ = −3.0775, σ = 0.4479), indicating that the sensor possesses acceptable short-time stability (Fig. 3e). The response curves of the QCM humidity sensor after storage for over two weeks are shown in Fig. S17 (ESI†). Compared with the curve of 2 days, the curves of 9 days and 15 days show a slight upward shift. This may be due to the adsorption of gases such as CO₂ in the air by CuO. After 15 days, the sensitivity only decreased by about 4.66%. The QCM humidity sensor exhibits good long-term stability. Besides, the variation in basic frequency concerning time is recorded, suggesting that the CuO QCM can be a preferable candidate for practical applications.

Another salient advantage of the CuO nanowires QCMs is its self-recovery ability when being wet by droplets, as shown in Fig. 3f. It can be seen that the resonant frequency decreases quickly after the droplet touching the sensor and returns to its original state due to the evaporation of the outspreading droplet, which is attributed to the superhydrophilicity of CuO nanowires. Moreover, the sensor shows excellent temperature resistance. When the temperature increases from room temperature to 200 °C, decreases reversely to 0 °C, and then increases to 200 °C, though the frequency shifts show a fluctuation resulting from the RH changes, the sensor will come back to its basic resonant frequency, as demonstrated in Fig. 3g. These results indicate that CuO nanowire QCM humidity sensor has a good performance.

3.2.2. Self-cleaning properties. The humidity sensors used in open-air for a long term always suffer from dust, leading to a loss of efficiency. A surface with the self-cleaning property may relieve this contradiction. Different from the self-cleaning mechanism of the traditional superhydrophilic surfaces such as TiO₂,³³ superhydrophilic CuO cannot degrade organic impurities under sunlight. Therefore, using a water droplet rolling off and taking the dust away from the surface may be a superior method. Here a micro-pit array was set on the sensor surface for constructing a droplet bouncing surface (Fig. 4a), achieving self-cleaning ability. As a proof of concept of self-cleaning, the micro-pit array could be controllably assembled on the sensor surface by lithography (Fig. 4b). The simulation results (Fig. 4d) show that the droplet completely suspends on the micro-pit's array on the surface. The air in the micro-pits generates capillary force, which prevents the droplet from penetrating into the micro-structure on the surface. The micro-structure changes the apparent solid–liquid contact into a solid–liquid–gas three-phase composite contact, leading to an increase in the WCA on the surface, and becomes more hydrophobic. In addition, according to the pressure distribution diagram (Fig. 4e), the pressure in some pits will increase during the contact between the droplet and the surface. This is caused by the droplet compressing the air in the pits, and the compressed air behavior is beneficial to the droplet rebound. It is known that a larger WCA can lead to a water droplet bouncing on the surface, especially for hydrophobic surfaces with low adhesive force. Herein, we mainly focus on the impact of size on the WCA of the self-cleaning layer. For the surface without micro-pit arrays, the droplet drops down and spreads around on the surface (Fig. 4g). The process will carry the dust to the edge of the surface while not carrying the dust away from the surface. On the contrary, for a surface fabricated with a micro-pit array self-cleaning layer, if a droplet drops on the surface, the main droplet will bounce and leave from the surface. The leaving droplet will take the dust away from the surface, realizing the self-cleaning ability, whether the surface was placed horizontally or obliquely. To validate the self-cleaning performance of the designed structure, we scatter the model contaminants on the 0° and 30° tilted surfaces and deposit a water droplet (1 m s⁻¹) at the top edge of the surface. From the successive images, it can be observed that the dropped water droplet carries the dust when rolled off from the surface (Fig. 4h–k and Movie S1–S3, ESI†). Most worthy of mention is that the WCA of the surface after hydrophobic treatment is just 120° (Fig. 4c), which is lower than the reported works.³⁴ The reason for this extraordinary phenomenon is the air in the micro-pit array. When the droplet encounters the surface, the droplet blocks the open window of the micro-pit array and the air could only be compressed without escaping. The existing air reduces the adhesive force between the nanowires and the droplet, leading to the bouncing of the droplet on the sensor surface. The frequency shift of the sensor after hydrophobic treatment shows a good linear relationship with humidity (Fig. 4f). Especially in high-humidity environments, the sensor frequency does not experience an exponential drift. The sensor after hydrophobic treatment still has satisfactory humidity sensing performance. These results indicate that a self-cleaning humidity sensor can be obtained by constructing a micro-pit array.

Fig. 4 Structure and self-cleaning properties of micro-pit array: SEM images of the sensor self-cleaning surface (a) micro-pit array and (b) enlarged view of the micro-pit array. (c) WCA contour map of sensors fabricated with different side lengths and spacing. Successive simulation images of the droplet bouncing on the surface of the micro-pit array: (d) morphology evolution and (e) pressure distribution. (f) Humidity sensing performance of sensors before and after hydrophobic treatment. (g) Droplet rapidly spreading around on the superhydrophilic CuO surface. Droplet bouncing on the horizontal CuO surface with micro-pit: (h) morphology evolution and (i) self-cleaning process. Droplet bouncing on the obliqued CuO surface with micro-pit: (j) morphology evolution and (k) self-cleaning process.

3.3. Superhydrophilicity and sensing mechanisms

3.3.1. Superhydrophilicity mechanisms. The surface of CuO is known for its hydrophilicity, and sometimes, it exhibits a superhydrophilic nature, as nanowires are present.³⁵ However, its superhydrophilic mechanism has rarely been revealed clearly. Regarding nanostructures, the wetting property is governed by the surface chemical composition and roughness. Here, molecular dynamics simulations and characterizations are both performed to study the wetting property of CuO nanowires.

On the one hand, eight models including flat and rough CuO surfaces with different crystal orientations were conducted, as shown in Fig. S18 (ESI†). A water droplet with 1500 water molecules was set on the top of the CuO surfaces, including flat-Cu (Fig. 5a), flat-O, groove, and pillar array surface (Fig. 5c). After optimization, for surfaces with (002) in crystal orientations, the droplet spreads out on the surfaces, exhibiting WCAs of 73.288° (Fig. 5b), 72.745°, 83.347°, and 80.164° (Fig. 5d and Table S2, ESI†), respectively. Similar performances can be observed from the (111) surface (Fig. S18, ESI†). These results show that the CuO surface has intrinsic hydrophilicity in theory.

Fig. 5 Molecular dynamics simulations and characterizations: water droplet on the flat CuO surface (a) before optimization and (b) after optimization. Water droplet on the groove CuO surface (c) before optimization and (d) after optimization. Roughness of CuO: (e) AFM images of the single nanowire and (f) light interferometer patterns for a large surface (50 × 50 μm).

On the other hand, AFM was employed to characterize the single CuO nanowire surface morphology and roughness, as displayed in Fig. 5e and Fig. S19 (ESI†). Its surface is relatively smooth with a roughness of R_a = 9.22 nm for a certain area. Beyond that, a light interferometer was conducted to measure the large surface area roughness (50 × 50 μm), as shown in Fig. 5f. The surface has a roughness of R_a = 18 nm. According to the Wenzel theory, cos θ₀ = γ cos θ, the roughness can amplify the wettability.³⁶ It can be roughly calculated that here γ is over 5, according to the light interferometer. In addition, γ is lower than 66, according to the BET results. Therefore, if θ is lower than 89°, the surface should be superhydrophilic. Since 72.745°< θ < 83.347° was estimated by molecular dynamics simulations, CuO nanowires surface should exhibit superhydrophilicity. This theoretical speculation fits with the experimental finding, which is also in accordance with the literature.²⁷ In sum, the CuO surface is immanent hydrophilic, while the unique morphology and significant roughness have certainly enhanced its wettability from hydrophilic to superhydrophilic.

3.3.2. Sensing mechanisms. Humidity sensing depends on the effects of H₂O molecules adsorbed on CuO nanowires. Molecular dynamics simulation, DFT calculations, QCM-D, BET, XPS, in situ FTIR spectroscopy and STEM were all performed to reveal the mechanism.

The free diffusion of water molecules (H₂O) on the surface of CuO nanowires in a nitrogen atmosphere was simulated by the molecular dynamics method. As time goes by, only water clusters appearing on the CuO surface by water molecules aggregate, which indicates that CuO nanowires have excellent adsorbability and good selectivity to H₂O (Fig. 6a). DFT calculations of H₂O on copper oxide nanowires were used to clarify the microscopic principle of adsorption (Fig. S20, ESI†). When a single H₂O molecule is adsorbed in different ways at different positions on the CuO surface, the ΔE_ads is about −1 eV (Table S3, ESI†), which means that the adsorption interaction between them is very strong, and the adsorption process is a spontaneous exothermic process, which is consistent with the reported work.²³ Finally, at the top site (Fig. 6b), the average population is 0.18, and the average bond length is 2.02 Å, which is less than 3 Å. Large populations and short bond lengths are typical covalent interaction features. At some bridge sites and hollow sites, the average population number was only 0.07, and there was no covalent interaction (Fig. 6c). In the two adsorption configurations of electron density difference of CuO with H₂O (Fig. 6e and f), while charges are transferred from Cu to O, the blue color around Cu at the adsorption site of CuO is deepened with charge losing, while the red color around O in H₂O is deepened with charge increasing, and there is an interaction between them. Charge transfer occurs in both adsorption modes, but their effect modes are different. Comparing the differential charge density distributions, at the top adsorption site, Cu in CuO and O in H₂O are wrapped by multiple electronic equipotential surfaces (Fig. 6g). This simultaneous wrapping indicates that there is a common electron pair to form a covalent bond between Cu and O. On the contrary, in other bridge and hollow adsorption sites (Fig. 6h), there are no wrapped or shared electron pair, which is just electrostatic adsorption. The analysis of partial density of states (PDOS) of O and Cu before and after adsorption can further reveal the interaction mechanism between them (Fig. 6i and j). The results show that electrons in the s and p orbitals of oxygen atoms transition to a low energy level due to adsorption, and the adsorbed H₂O is in a more stable state. The energy levels of s orbitals and p orbitals of Cu split at −23 eV and −7 eV, respectively, and then they hybridize with O in H₂O to produce a covalent interaction. However, the overlapping degree is weaker than that between Cu–O bonds in CuO, meaning forming a weak chemical bond between CuO and H₂O.

Fig. 6 Humidity sensing mechanism of CuO nanowires: (a) successive images of the molecular dynamics simulation of the water and nitrogen molecules on the CuO surface. (b) and (c) Covalent adsorption and electrostatic adsorption. (d) Sketch map of CuO nanowires humidity sensing mechanism. (e) and (f) Differential charge density distributions of covalent adsorption and electrostatic adsorption. (g) and (h) Charge density distributions of covalent adsorption and electrostatic adsorption. (i) and (j) PDOS of covalent adsorption and electrostatic adsorption of CuO with H₂O. (k) In situ FTIR spectra of CuO under different humidity levels from 0% RH to 100% RH. (l) Sensitivity of the sensor and influential parameters (BET surface, Cu/O mass ratio, and O_a/O₁) vs. annealing temperature.

DFT calculations were used for the adsorption behavior of several water molecules on the CuO surface, which proves that water clusters were formed in a high humidity environment. When the amount of H₂O increases, the number of Cu–O covalent bonds will increase first. While it increases to four, hydrogen bonds will occur between H₂O. When it continues to increase, only the hydrogen bonds will increase. In the whole process, the adsorption energy is decreasing, as shown in Table S4 (ESI†). The adsorption energy decrease is not obvious with the increase in covalent bond, but after hydrogen bond appear, the adsorption energy declines obviously with the increase in hydrogen bonds. According to the bonding situation and adsorption energy change trend, there are chemical bonds at low humidity and two types of bonds at high humidity. That is, hydrogen bond plays a leading role in adsorption process at high humidity, and covalent bond plays a leading role at low humidity. The PDOS of hydrogen bond was analyzed (Fig. S21, ESI†). The electrons on the s and p orbitals of O all migrate to a low energy level, with a shift of about −4 eV, which indicates that the adsorption product is more stable. The corresponding electronic peaks in the bonding orbitals of oxygen atoms are broadened and reduced, which is characterized by covalent bonds. Some oxygen atoms with hydrogen bonds have energy level splitting on the p orbitals and hybridize with hydrogen atoms, while the other O with hydrogen bonds have no energy level splitting at the same position of the p orbitals, indicating that there are hydrogen bonds with different strengths. According to the analysis of electronic density diagram, strong hydrogen bonds share more electrons than weak hydrogen bonds, which also indicate that there are strong and weak adsorption between hydrogen bonds. In a word, due to hydrogen bonds, water molecules are adsorbed continuously, thus forming water clusters on the surface of CuO.

The adsorption mechanism of H₂O molecules on the CuO surface is a process of chemical adsorption and physical adsorption (Fig. 6d). First, on the surface of CuO, Lewis acid–base interaction occurs between H₂O and CuO, and charges are transferred from the CuO surface to H₂O molecules, forming covalent bonds between Cu–O atoms, which is the chemical adsorption of water. Then, the H₂O molecule of chemisorbing on the CuO surface continues to adsorb other H₂O molecules by forming hydrogen bonds, which is a kind of physical adsorption, which is also the main role of H₂O clusters, appearing on the CuO surface under a high-humidity environment.

In the experiment, QCM-D qualitatively evaluates the viscoelastic changes of CuO nanowires under different humidities through the change of dissipation value (ΔD). CuO nanowires exhibit a gradual transition from rigid to viscoelastic behavior, as humidity levels increases. In a relatively low humidity environment (<56% RH), ΔD is small, CuO nanowires do not swell much. When the humidity level goes up to 84%, the ΔD changes greatly, the sensing materials swell intensively due to the absorbed and condensed water. In addition, in in situ FTIR results, the characteristic peak of water will appear on the surface of CuO nanowires with the increase in humidity, showing the liquid water appeared on the surface (Fig. 6k). The first two characterization experiments prove that H₂O is adsorbed onto the surface of CuO nanowires in the form of molecules at low humidity and condensed water at high humidity, which is consistent with the DFT results. The sensitivity increases first reaching the maximum at an annealing temperature of 200 °C and decreases with the increase in the annealing temperature. The majority of works have shown that a larger sensing area can achieve a higher sensitivity. The BET results show that with the increase in annealing temperature, the specific surface area of the sensing material decreases, while the sensitivity partially increases (Fig. 6l). This means that the surface area is not the main factor, affecting the sensing performance. As a known semiconductor material, a large number of oxygen vacancies were produced during the annealing of CuO, which provide more adsorption sites, and the stronger the chemical adsorption process, which may be one of the main influencing factors. The ratio between lattice oxygen species (O_l) and surface adsorbed oxygen (O_a)³⁷ indicates that the number of surface oxygen vacancies increases with the increase in annealing temperature, which is consistent with sensitivity.

To sum up, the decrease in the resonant frequency is governed by the adsorbed water molecules, and the number of absorbed water molecules depends on surface area, oxygen vacancy and defects. The solution growth preparation and annealing of CuO nanowires increase the availability of edges, defects and oxygen vacancies, and promote the adsorption of water molecules. When the oxygen vacancy of CuO particles increases, the surface atoms will be in a relative relaxation state owing to the loss of adjacent atoms, followed by a reduction in the corresponding vibration frequency and scattering of vibration frequency signals.³⁸ Thus, the high sensitivity of CuO nanowire-based QCMs is mainly ascribed to the surface area, the hydrophilic surface-terminating functional group (O), oxygen vacancy and hydrogen bond, which play a dominant role in the absorption of water.

3.4. Human–machine interaction applications

Production and daily life are inseparable from sensors.^39–41 To demonstrate the excellent sensing performance of the superhydrophilic CuO nanowire-based humidity sensor, breath (normal and after sporting) and touch switch (high and low distance) tests were conducted.

3.4.1. Non-contacting sensing and control system applications. Highly sensitive humidity sensors of CuO are desirable for non-contact human–machine interaction (HMI) applications. Noticeably, the humidity level of the finger is associated with the distance. It can be seen that the resonant frequency decreases with the increase in the finger distance (Fig. 7a). Particularly, the response time is only 0.17 s (Fig. 7b), suggesting that superhydrophilic CuO nanowires are appropriate for non-contact detection. According to the results, we designed a touchless switch based on CuO nanowire QCMs, as shown in Fig. 7c. When the humidity level exceeds a certain threshold value, the LED will show different colors. Furthermore, a non-contact human–machine interaction (HMI) system using the Bluetooth technology base on the smartphone (Fig. 7d) was explored to expand the application scenarios of the humidity sensor, such as health monitoring, breathing testing, and other physiological studies.

Fig. 7 Non-contacting human–computer interaction application: (a) frequency shift of human finger vs. distance, (b) response time of the sensor; (c) touchless switch based on CuO nanowire QCMs; (d) non-contact human–machine interaction and sensing system with the Internet of Things; (e) smart human breath detecting system; human breath (f) before sporting and (g) after sporting.

3.4.2. Breath detection application. High sensitivity with the rapid response of the sensor is expected to be employed in human physiological actives such as breath detection (Fig. 7e). When people breathe under normal conditions, the resonant frequency exhibits a dynamically abrupt decrease during exhalation and comes back to the original value during inspiration (Fig. 7f). After sports, the breath becomes deeper than that under normal condition (Fig. S22, ESI†). As time goes on, the resonant frequency exhibits an increase due to the rest, as shown in Fig. 7g. In practical application, the robust humidity sensing capability of the CuO nanowire QCM humidity sensor is shown in Fig. S23 (ESI†), exhibiting highly stable testing ability.

4. Conclusions

In summary, we demonstrated a superhydrophilic CuO nanowire-based QCM humidity sensor with high sensitivity and self-cleaning ability for non-contact human–machine interaction applications. The unique CuO nanowires were obtained from the thermal decomposition of the in situ-grown Cu(OH)₂ nanowires on a substrate deposited with a Cu layer, which is environmentally benign and cost-effective. The CuO nanowire-based QCM delivers a high sensitivity of 82.5 ± 7.7 Hz/%RH due to the large surface area with lots of oxygen vacancy. The sensor can keep its original sensing properties after being wetted because of its superhydrophilicity, showing an excellent self-recovery ability. Besides, the theoretical calculation and experiment results reveal that the high performance is attributed to the coupling effect of abundant adsorption sites such as oxygen vacancy and physisorption of water molecules. Since the sensor can detect human breath, speaking, and finger motion, a non-contact human–machine interaction system was constructed and applied in alarm and switch. Our work provides new perspectives for the humidity sensor with self-cleaning ability in fields such as medical health monitoring and industrial environment detection.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

There is no conflict of interest to declare.

Acknowledgements

This work is supported by the National Natural Science Foundation of China (Grant No. 52275562). The authors acknowledge Nano Fabrication and Measurement Laboratory of Collaborative Innovation Center for Digital Intelligent Manufacturing Technology and Application and engineers in the center of Micro-Fabrication and Characterization (CMFC) of Wuhan National Laboratory for Optoelectronics for the support in the SEM and XRD tests, respectively. The authors appreciate Mr Jing Luo for developing the circuit and software of the sensing and control system. The authors thank the Analytical and Testing Center of Huazhong University of Science and Technology for providing SEM, TEM, XRD, and FTIR facilities. The authors thank Prof. Bin Shan for his support in theoretical calculations. The authors also thank Aigtek (Xi’an, China) engineers for their support and provision of amplification equipment (ATA-7000).

Notes and references

1. K. Sreejivungsa, N. Thanamoon, N. Phromviyo, W. Jarernboon, M. Takesada and P. Thongbai, Advanced humidity sensing properties of CuO ceramics, Sci. Rep., 2024, 14, 9726.
2. E. Ganbold, E. S. Kim, Y. Li, F. Yin, P. K. Sharma, J. Jeon, J. Oh, D. N. Lee and N. Y. Kim, Highly Sensitive Interdigitated Capacitive Humidity Sensors Based on Sponge-Like Nanoporous PVDF/LiCl Composite for Real-Time Monitoring, ACS Appl. Mater. Interfaces, 2023, 15, 4559–4568.
3. S. Tachibana, Y. Wang, T. Sekine, Y. Takeda, J. Hong, A. Yoshida, M. Abe, R. Miura, Y. Watanabe, D. Kumaki and S. Tokito, A Printed Flexible Humidity Sensor with High Sensitivity and Fast Response Using a Cellulose Nanofiber/Carbon Black Composite, ACS Appl. Mater. Interfaces, 2022, 14, 5721–5728.
4. B. Xie, H. You, H. Qian, S. Hao, Z. Li, Y. Qin, H. Wen, S. Xia, Y. Xiang and J. Hu, High-Performance Flexible Humidity Sensor Based on MoOx Nanoparticle Films for Monitoring Human Respiration and Non-Contact Sensing, ACS Appl. Nano Mater., 2023, 6, 7011–7021.
5. Z. Ma, T. Fei and T. Zhang, An overview: Sensors for low humidity detection, Sens. Actuators, B, 2023, 376, 133039.
6. J. Lin, N. Gao, J. Liu, Z. Hu, H. Fang, X. Tan, H. Li, H. Jiang, H. Liu, T. Shi and G. Liao, Superhydrophilic Cu(OH)2 nanowire-based QCM transducer with self-healing ability for humidity detection, J. Mater. Chem. A, 2019, 7, 9068–9077.
7. H. Song, X. Zhu, E. Yang and Z. Han, A QCM Humidity Sensor Based on MoS₂ Nanoflower-Modified Chitosan Composite With High Quality Factor, IEEE Trans. Electron Dev., 2023, 70, 1849–1853.
8. D. Zhang, X. Song, Z. Wang and H. Chen, Ultra-highly sensitive humidity sensing by polydopamine/graphene oxide nanostructure on quartz crystal microbalance, Appl. Surf. Sci., 2021, 538, 147816.
9. X. Cha, F. Yu, Y. Fan, J. Chen, L. Wang, Q. Xiang, Z. Duan and J. Xu, Superhydrophilic ZnO nanoneedle array: Controllable in situ growth on QCM transducer and enhanced humidity sensing properties and mechanism, Sens. Actuators, B, 2018, 263, 436–444.
10. Z. Ma, Y. Yu, K. Wu, Y. Song, S. Liu, X. Yang, T. Fei and T. Zhang, Au loaded mesoporous SiO₂/gelatin hydrogel: Detecting low humidity and NH₃, Chem. Eng. J., 2023, 471, 144788.
11. N. Gao, H. Li, W. Zhang, Y. Zhang, Y. Zeng, H. Zhixiang, J. Liu, J. Jiang, L. Miao, F. Yi and H. Liu, QCM-based humidity sensor and sensing properties employing colloidal SnO2 nanowires, Sens. Actuators, B, 2019, 293, 129–135.
12. P. Qi, Z. Xu and T. Zhang, In Situ Growth of Dopamine on QCM for Humidity Detection, Chemosensors, 2022, 10, 522.
13. D. Zhang, N. Luo, Z. Xue, Y. Bai and J. Xu, Effect of Open Metal Sites in Cobalt-Based Bimetallic Metal–Organic Framework Nanoparticles-Coated Quartz Crystal Microbalance (QCM) for Humidity Detection, ACS Appl. Nano Mater., 2022, 5, 2147–2155.
14. J. Wang and W. Zeng, Research Progress on Humidity-Sensing Properties of Cu-Based Humidity Sensors: A Review, J. Sens., 2022, 7749890.
15. A. Zainelabdin, G. Amin, S. Zaman, O. Nur, J. Lu, L. Hultman and M. Willander, CuO/ZnO Nanocorals synthesis via hydrothermal technique: growth mechanism and their application as Humidity Sensor, J. Mater. Chem., 2012, 22, 11583–11590.
16. H. Fang, J. Lin, Z. Hu, H. Liu, Z. Tang, T. Shi and G. Liao, Cu(OH)2 nanowires/graphene oxide composites based QCM humidity sensor with fast-response for real-time respiration monitoring, Sens. Actuators, B, 2020, 304, 127313.
17. Y. Zhang, Y. Wu, Y. Fu, Q. Jia and Z. Zhang, Sulfonated hypercross-linked porous organic polymer based humidity sensor, Sens. Actuators, B, 2024, 401, 134997.
18. C. Li, Y. Yin, H. Hou, N. Fan, F. Yuan, Y. Shi and Q. Meng, Preparation and characterization of Cu(OH)2 and CuO nanowires by the coupling route of microemulsion with homogenous precipitation, Solid State Commun., 2010, 150, 585–589.
19. C. Wu, D. Liu and Z. D. Cui Yuting, Influences of annealing temperature on humidity sensing properties of nanostructured CuO, J. At. Mol. Phys., 2019, 36, 581–587.
20. K. Malook, H. Khan and M. Ali Ihsan-Ul-Haque, Investigation of room temperature humidity sensing performance of mesoporous CuO particles, Mater. Sci. Semicond. Process., 2020, 113, 105021.
21. R. Nitta, H. Lin, Y. Kubota, T. Kishi, T. Yano and N. Matsushita, CuO nanostructure-based flexible humidity sensors fabricated on PET substrates by spin-spray method, Appl. Surf. Sci., 2022, 572, 151352.
22. H. Ben Saâd, M. Ajili, S. Dabbabi and N. T. Kamoun, Investigation on thickness and annealing effects on physical properties and electrical circuit model of CuO sprayed thin films, Superlattices Microstruct., 2020, 142, 106508.
23. D. Li, J. Hu, R. Wu and J. G. Lu, Conductometric chemical sensor based on individual CuO nanowires, Nanotechnology, 2010, 21, 485502.
24. X. Zhou, W. Guo, J. Fu, Y. Zhu, Y. Huang and P. Peng, Laser writing of Cu/CuxO integrated structure on flexible substrate for humidity sensing, Appl. Surf. Sci., 2019, 494, 684–690.
25. J. Cao, L. Deng, X. Wang, W. Li, Y. Xie, J. Zhang and S. Cheng, Stable Lithium Metal Anode Achieved by In Situ Grown CuO Nanowire Arrays on Cu Foam, Energy Fuel, 2020, 34, 7684–7691.
26. X. Jiang, T. Herricks and Y. Xia, CuO Nanowires Can Be Synthesized by Heating Copper Substrates in Air, Nano Lett., 2002, 2, 1333–1338.
27. S. John, S. S. Vadla and S. C. Roy, High photoelectrochemical activity of CuO nanoflakes grown on Cu foil, Electrochim. Acta, 2019, 319, 390–399.
28. H. Alsweefe, S. K. Al-Hayali and A. Al-Janabi, Efficient humidity sensor based on an etched no-core fiber coated with copper oxide nanoparticles, J. Nanophotonics, 2018, 12, 46018.
29. O. Baranov, K. Bazaka, T. Belmonte, C. Riccardi, H. E. Roman, M. Mohandas, S. Xu, U. Cvelbar and I. Levchenko, Recent innovations in the technology and applications of low-dimensional CuO nanostructures for sensing, energy and catalysis, Nanoscale Horiz., 2023, 8, 568–602.
30. Q. Li, H. Liu, S. Zhang, D. Zhang, X. Liu, Y. He, L. Mi, J. Zhang, C. Liu, C. Shen and Z. Guo, Superhydrophobic Electrically Conductive Paper for Ultrasensitive Strain Sensor with Excellent Anticorrosion and Self-Cleaning Property, ACS Appl. Mater. Interfaces, 2019, 11, 21904–21914.
31. R. Yamamoto, D. Kowalski, R. Zhu, K. Wada, Y. Sato, S. Kitano, C. Zhu, Y. Aoki and H. Habazaki, Fabrication of superhydrophobic copper metal nanowire surfaces with high thermal conductivity, Appl. Surf. Sci., 2021, 537, 147854.
32. M. H. Chou, S. B. Liu, C. Y. Huang, S. Y. Wu and C. L. Cheng, Confocal Raman spectroscopic mapping studies on a single CuO nanowire, Appl. Surf. Sci., 2008, 254, 7539–7543.
33. R. Wang, K. Hashimoto, A. Fujishima, M. Chikuni, E. Kojima, A. Kitamura, M. Shimohigoshi and T. Watanabe, Light-induced amphiphilic surfaces, Nature, 1997, 388, 431–432.
34. D. C. Manatunga, R. M. de Silva and K. M. N. de Silva, Double layer approach to create durable superhydrophobicity on cotton fabric using nano silica and auxiliary non fluorinated materials, Appl. Surf. Sci., 2016, 360, 777–788.
35. K. Tang, X. Wang, W. Yan, J. Yu and R. Xu, Fabrication of superhydrophilic Cu2O and CuO membranes, J. Membr. Sci., 2006, 286, 279–284.
36. J. Bico, U. Thiele and D. Quéré, Wetting of textured surfaces, Colloids Surf., A, 2002, 206, 41–46.
37. K. Ye, K. Li, Y. Lu, Z. Guo, N. Ni, H. Liu, Y. Huang, H. Ji and P. Wang, An overview of advanced methods for the characterization of oxygen vacancies in materials, TrAC, Trends Anal. Chem., 2019, 116, 102–108.
38. H. Ming, K. Pan, Y. Liu, H. Li, X. He, J. Ming, Z. Ma and Z. Kang, Electrochemical fabrication of Cu(OH)2 and CuO nanostructures and their catalytic property, J. Cryst. Growth, 2011, 327, 251–257.
39. R. Liao, X. Zhao and M. Liu, Inverted Molding with Porous Skeleton Nickel Foam for Preparing Flexible Multi-Wall Carbon Nanotubes Pressure Sensors, Sensors, 2023, 23, 9560.
40. C. Chen, J. Li, Z. Tang, G. Liao and L. Nie, In-situ co-precipitation of Bi-MOF derivatives for highly sensitive electrochemical glucose sensing, Microchem. J., 2024, 199, 109897.
41. L. Nie, Y. Huang, Y. Yin, M. Liu, L. Wu and H. Yang, Internal defect identification method of TSV 3D packaging based on built-in integrated sensor, Adv. Mech. Eng., 2022, 14, 1669534664.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5tc00676g

‡ Co-first authors. These authors contributed equally to this work.

---