MoS₂, absolute ethanol (99.7 wt%) were purchased from Aladdin Ltd. (Shanghai, China). Poly (N-vinylpyrrolidone) (PVP, average Mw 40,000) was purchased from Sigma-Aldrich Co., Ltd. (Shanghai, China). LCP substrate of 50 μm was provided from Taiflex Scientific Co., Ltd. All chemicals were of reagent grade.

All electrochemical experiments were conducted on a CHI660E electrochemical workstation. Electrode appearance scanning microscopy was handled using VHX-5000 Ultra-depth three-dimensional microscope (keyence, Japan). Scanning electron microscopy (SEM) images were collected on a Sigma HV (Zeiss, Germany). Fourier transform infrared spectroscopy (FT-IR) spectra were collected on a Nicolet iN10 FT-IR spectrometer (Thermo Fisher, United States).

1.0 g bulk MoS₂ was suspended in a mixed solution of absolute ethanol and deionized water (1:1, v/v) and subjected to ultrasonication with 380 W power. After exfoliating at room temperature for 1 h, 0.2 g PVP was added as surfactants and stabilizer for another 48 h under the same conditions. Then the suspension was centrifuged for 15 min at 7,500 rpm, and the resulting supernatant was preserved as the MoS₂/PVP ink for inkjet printing (Jin X. et al., 2020).

The humidity sensor was fabricated on a 50 μm thick LCP substrate with copper cladding on one side. The copper interdigital electrodes were fabricated on the initial flexible substrate by FPC technology, which involves two main steps as followed: First, optical dry film photoresist lithography (DFPL) was performed by roller lamination of the dry film photoresist (30 μm, Dupont) on LCP copper clad plate at 100°C and a speed of 1.4 m/min, followed by UV-exposure with the energy density of 130 mJ/cm². Second, the exposed LCP copper clad plate was etched by an aqueous solution of 0.5 mol/L FeCl₃ and HCl (1:1).

The as-prepared MoS₂/PVP ink was inkjet-printed on LCP-copper interdigital electrodes which were pre-treated by oxygen plasma treatment (100 W, 2 min) to fabricate humidity sensors. Inkjet printing was conducted using a commercial drop-on-demand Epson L310 printer. The printer cartridge was designed to dispense droplets with a volume of 3 pL based on the orifice of 63.5 μm in diameter (Jin X. F. et al., 2020). MoS₂/PVP ink was inkjet-printed on copper interdigital electrodes to fabricate humidity sensors, repeat 10 times and dry at 42°C for 10 min after each printing. The flexible humidity sensors were left to dry for 12 h under ambient conditions.

The humidity sensor fabrication procedure is illustrated in Figure 1A. The interdigital electrodes were fabricated on a mechanically flexible LCP substrate by FPC technology. As initial substrate, a dimension of 250 mm × 300 mm × 50 μm LCP substrate with 18 μm copper cladding on one side was patterned into 180 pieces of miniature interdigital electrode with 4 mm in width and 6 mm in length. As one of two-dimensional (2D) nanomaterials, MoS₂ possesses an inherently high surface-to-volume ratio which could serve as an ideal candidate for gas sensing (Zhang et al., 2014). Owing to the assisted exfoliation of PVP, MoS₂/PVP ink (Figure 1D) has the more uniform and smaller sizes than MoS₂ ink (Figure 1C) which increases the edge active sites. Inset in the top right corner of Figure 1D was the photograph of MoS₂/PVP ink which is stable for more than 1 month under ambient conditions. As indicated from the SEM image in Figures 1B–D, the size of layered MoS₂ was considerably reduced to nanometer scale after liquid-phase ultrasonic exfoliation. As shown in Figures 1E–G, the LCP-copper interdigital electrodes exhibited extreme flexibility.

Figure 2 gave a comparison of FT-IR spectra between MoS₂ and MoS₂/PVP, which further confirmed the successful preparation of MoS₂/PVP ink. The typical characteristic peaks of PVP were observed in the range of 4,000–500 cm⁻¹. The stretching vibration of C-H was seen at 2,890 cm⁻¹. The characteristic peak at 1,650 cm⁻¹ corresponds to the carbonyl group. The peak at 1,290 cm⁻¹ was assigned to the stretching vibration of C-N. The FT-IR spectrum exhibits a characteristic peak at 780 cm⁻¹, indicating the existence of a molybdenum compound.

The humidity atmospheres were produced by different saturated salt solutions of LiCl for 11% RH, MgCl₂ for 33% RH, Mg(NO₃)₂ for 54% RH, NaCl for 75% RH, KCl for 84% RH and KNO₃ for 94% RH, respectively (Dai et al., 2019). The humidity-sensing performances of sensors were evaluated by the response (R) which was defined as: R = I R H − I 11 / I 11 , where I R H and I 11 are the measured current of the sensor at the given RH and 11% RH levels, respectively.

As shown in Figure 3A, the dynamic response current decreases with humidity increasing when the as-fabricated flexible humidity sensors were exposed to humidity gas from low to high level. The increase in current response could be due to the electrical resistance of the sensors decreasing with the increase of relative humidity. Owing to the great surface-area-to-volume ratio and high carrier mobility, MoS₂ has been demonstrated to be an effective nanomaterial for humidity sensing. Our previous study has proven the better humidity sensing performance of MoS₂/PVP sensor compared to the pure MoS₂ sensor, possibly due to better hydrophilicity of MoS₂/PVP hybrid nanocomposite, which may contribute to the reduction of mass transfer resistance and consequently promote the reaction kinetic (Jin X. et al., 2020). There are two main kinds of adsorption processes between water molecules and sensing materials, including chemisorption and physisorption. The first adsorption process is chemisorption which is difficult to remove after chemisorbed layer is formed. The second adsorption process is physisorption which will happen above the chemisorbed layer with the increase of the adsorbed water molecules (Ze et al., 2017). The peptide groups of lactams in PVP structure are active binding sites for hydrogen donors. The mechanism of adsorption/desorption of water molecular and PVP is schematically described as follows: (1)

(1)

At low relative humidity, chemisorption of water molecules occurs on the surface of the sensing layer. These chemisorbed water molecules act as electron donors for the sensing materials, allowing for decreased electrical resistance. However, due to the discontinuous of the water layer at this process, the hydronium ion (H₃O⁺) transfer is difficult to occur under the electrostatic field. With relative humidity increasing to high level, more water molecules are physiosorbed above the chemisorbed layer by hydrogen bonding to form continuous water multilayer. According to the Grotthuss chain reaction, the protons hopping-transfer is generated and used to ionic conductivity, leading to a rapid decrease in electrical resistance and increase in response current.

Response/recovery time is one of the most crucial parameters to consider for fabricating humidity sensors (Bahuguna et al., 2020). The response and recover times are defined as the time taken from 10% to 90% of the total current change and vice versa (Guo et al., 2017). As shown in Figure 3B, it can be measured from the breathing pattern that the as-fabricated humidity sensors have an extremely short response and recovery time of 0.8 and 0.7 s, respectively.

Repeatability and Long-term stability are two core properties to consider for commercial applications of humidity sensors. Considering the application for continuous monitoring of respiration rate, repeatability was evaluated by measuring the time-dependent current response from 55% RH to above 90% RH (breathing pattern). The consistent baseline (current response of device under 45% RH) and the current changes of the device under 30 cycles of exposure to the high relative humidity (above 90% RH) indicate good repeatability of the proposed humidity sensor (Figure 3C). Long-term stability of the humidity sensors was evaluated by measuring the current response towards various relative humidity for every few days. As shown in Figure 3E, the current responses toward various relative humidity remain stable for 30 days, thus to ensure the reliability of respiratory frequency monitoring application. The mechanical robustness of the sensors was evaluated by a bending test which was performed based on the procedure reported by Wu et al. (2019). Figure 3D shows that no noticeable response degradation was observed at bending angle of 60°, indicating that bending tests do not adversely affect the performance of the proposed humidity sensors.

The normal respiration of an adult lasts 3–5 s and exhaled air humidity is above 90% RH. In consideration of the high performance of the humidity sensors fabricated in this study, we further integrated the flexible copper interdigital electrode components, sensing layer, and miniaturized PCB board into a wearable patch sensor for respiration monitoring.

The PCB board of electrochemical workstation with miniaturized format (0.8 cm × 1.8 cm) consists of a signal conditioning part, a Bluetooth transceiver and a programmable electrochemical chip. The signal is wireless communication between wearable humidity sensor and smartphone-based platforms. The system is powered by a 150 mAh rechargeable lithium battery. The applied potential was set as 200 mV during operation. Due to a low soldering temperature of LCP (260°C–310°C, 10 s), conductive silver paste was used for the connection between wire and LCP-copper interdigital electrode.

In order to evaluate the real-time breath monitoring application of the fabricated humidity sensor, the device was attached under a volunteer’s nose about 5 mm. As shown in Figure 4, the device can distinguish different respiration patterns such as normal breathing, fast breathing and slow breathing with good stability and sensitivity, holding great promise for realizing continuous monitoring of breath. In addition, the fabricated humidity sensor can realize skin moisture monitoring in a touch-less way. At a distance of 2 mm, the sensor showed high sensitivity towards the moisture content of a volunteer’s fingertip, exhibiting its potential applications in non-contact switch, skin examination and spatial localization monitoring (Figure 3F).