Flame‐retardant, flexible, and breathable smart humidity sensing fabrics based on hydrogels for respiratory monitoring and non‐contact sensing

Smart sensing fabrics are becoming increasingly attractive in the emerging wearable areas of medical and military so far. Here, for the first time, we present a smart humidity sensing fabric (SHSF) based on a moisture‐sensitive polyacrylamide hydrogel for respiratory monitoring and non‐contact sensing. Fabricated by in situ cross‐linking of the hydrogel precursors on the fibers of the fabric, the flexible SHSF shows excellent sensitivity, outstanding flame retardance, air permeability, water retention capacity, and stability after treatment with lithium bromide solution. Specifically, its conductance increases more than 311 times as humidity increased from 11% to 98%. Besides, the humidity sensor features good repeatability and the ability to work normally under folding due to its flexible nature. As a clothing material, hydrogel–fabric composite exhibits 4.3 times the burning time compared to cotton fabric, illustrating better flame retardance. The SHSF is used to monitor human breathing and non‐contact finger approaching in real time, demonstrating its flexibility in practical applications. This work provides strategies for preparing high‐performance, flame‐retardant SHSF for emerging wearable electronic devices.

Smart sensing fabric, a new type of electronic product with sensing capabilities based on textiles, enables us to take advantage of sensors anytime with minimal interference. [1][2][3][4][5][6] Being flexible and wearable, smart sensing fabric can play an important role in various applications, such as electrocardiograph monitoring in medical field, 7 controlling drug release for wounded soldiers in military, 8 human motion detection, 9 and exercise intensity assessment 10 in commercial field. Among all the environmental stimuli, humidity sensing is a crucial part of smart sensing fabric as it is closely related to people's health and comfort. [11][12][13][14][15] For example, by monitoring humidity under bed sheets, patients requiring prolonged bed rest can prevent pressure ulcers. 16 People can also monitor sweat and respiratory rate while exercising. [17][18][19] In addition, the flame-resistant behavior of textiles is of great importance. Burns are a global public health problem and the World Health Organization estimates that approximately 180,000 people die from burns each year. 20 Most of these fire incidents occur indoors; thus, portable and effective personal protective equipment such as flame-resistant fabric is particularly significant compared to public firefighting equipment. 21 Integrating flame-resistant behavior into smart sensing fabrics not only makes sensing performance apply to fire rescue better but also gives everyday wearers more time to escape, which lessens the casualty toll.
Conventional humidity sensing fabric is based on flexible substrate strips covered with a moisturesensitive sensing layer. Weaving is used to incorporate the flexible substrates into the textile. The sensing material may use cellulose acetate butyrate or polymer electrolyte such as polymethyl methacrylate and poly (3,4-ethylenedioxythiophene) poly(styrene sulfonate) (PEDOT-PSS), 22,23 while flexible substrates are usually made of polyethylene terephthalate and polyimide. 23,24 Nevertheless, with the use of flexible substrates, the differences in the physical properties between flexible substrates and fabric severely affect the comfort of the textile. Meanwhile, the weaving limits the versatility of the weaving process of the fabric 25 and also causes electrical connection problems. 22 The mechanical stresses generated by flexible substrate during weaving degrade the sensing performance. 16 A better alternative to integrating the sensing layer into the fabric is an additive manufacturing process, which means covering the sensor directly on the surface of the fabric. It is an effective method for avoiding the problems caused by weaving flexible substrate strips into the fabric. Traditional additive manufacturing processes employ screen printing, 24 inkjet printing, 25 or spray coating 26 to deposit two materials, that is, conductive electrodes and sensing layers, on the fabric surface. To date, flexible pressure, temperature, and humidity sensors are constructed by integrating several non-stretchable and non-transparent sensing materials, such as graphene oxide (GO), 3,18,27 reduced graphene oxide (rGO), 28 single/multi-walled carbon nanotube (SMCNT/MWCNT), 29,30 polymers, 31,32 and micro/nanostructured transition-metal carbides and carbonitrides. [33][34][35] Nevertheless, most of the reported sensors lack air permeability or fire retardancy, bearing limited non-intrusiveness and changing the appearance of the fabrics, among which inorganic or graphene materials have weak processing and film-forming capabilities as electrical devices. Also, their flame retardancy has not been evaluated, either. Recently, hydrogels have been widely studied in the fields of flexible sensors, electronic skin, wearable devices, soft robots, and so on due to their ionic conductivity, stretchability, transparency, and biocompatibility. [36][37][38][39][40][41][42][43][44][45][46][47] The density of the polymer network and the hindering effect on ions will change at different humidity levels. This swelling effect makes the resistance of the hydrogel vary with environmental humidity. 43,48 The properties of ionic conductivity, high stretchability, and humidity sensitivity make it possible that hydrogels can be used as both the humidity sensing material and the substrate, which makes the sensor substrate-free.
Despite the recent significant exploration of hydrogelbased sensors, the hydrogel-based wearable sensors fabricated in situ on fabric without substrate have not yet been reported. If the hydrogel-based humidity sensing fabric can be created, it should have characteristics of a normal fabric, such as being breathable, non-toxic, comfortable, and flexible. In addition, after cross-linking on the fabric fiber, the hydrogels can function as a flame retardant. When exposed to an open fire, it absorbs heat by evaporating water, dilutes oxygen by releasing CO 2 , and insulates oxygen by laminating on the fabric surface. 49 Compared to conventional halogen flame retardants and phosphorus F I G U R E 1 Schematic illustrating the properties of the smart humidity sensing fabric (SHSF) and its applications. flame retardants, hydrogels produce little toxic hydrogen halide gas and offer reduced smoke production, lower volatility, and higher efficiency. 50,51 To our knowledge, smart humidity sensing fabrics (SHSF) based on hydrogels have not been reported so far. For the first time, we present a breathable SHSF with a flame-retardant property based on stretchable polyacrylamide (PAM) hydrogels ( Figure 1). This SHSF achieves the same tactile sensation as ordinary fabrics and good breathability, which still exhibits high sensitivity toward humidity despite deformation. These characteristics enhance its non-intrusiveness and wearability. In order to improve the water retention capacity and stability of the SHSF, we treated them with lithium bromide (LiBr) solution. 52 Moreover, we systematically investigated the effect of the concentrations of LiBr solution on the sensitivity of the fabrics. As for flame retardance, the SHSF does not burn at the slight touch of fire. When continuously exposed to open flames, their flame retardancy is verified by comparing with ordinary cotton fabrics. Furthermore, the SHSF can be applied to monitor physiological activities such as human breathing and finger approach (Figure 1), which is attributed to the capability to sensitively capture subtle changes in environmental humidity.

Fabrication of SHSF
The PAM hydrogel was synthesized using a facile onepot polymerization method. Acrylamide (AM), potas-sium chloride (KCl), N,N-methylenebisacrylamide (MBA), and ammonium perchlorate (AP) were utilized as the monomer, ionic charge carrier, cross-linker, and thermal initiator, respectively, for PAM hydrogel synthesis. Specifically, chemicals including 7.5 g AM, 0.005 g MBA, 0.09 g KCl, and 0.0375 g AP were mixed in deionized water to prepare a 50 g hydrogel precursor. Subsequently, the precursor was dropped onto the spread textile until the whole textile was immersed in the mixed precursor. After being heated at 90 • C for an hour, the obtained SHSF based on hydrogel-fabric was immersed into LiBr solution for 30 min to incorporate LiBr in the hydrogel via a facile salt infiltration strategy, and the extra LiBr solution on the surface of SHSF was removed with a filter paper ( Figure 2A). Conventional hydrogels suffer from the intrinsic problem of water evaporation-induced instability. Here, LiBr, a deliquescent salt, can play a key role in enhancing the moisture retention capacity of the hydrogel by promoting the water absorption.

Deformability characterization
The resistance/conductance of the SHSF was monitored using a Keithley analyzer (Model 2400) after applying a fixed voltage of 1 V on the SHSF. To apply desired strains to the SHSF for the electromechanical test, the two ends of the samples were fixed on a homebuilt motorized stretching stage, which was controlled by a Zolix SC300-3A motion controller and corresponding software. Two metal clips were attached to the two ends of the SHSF and served as the electrodes. The relative resistance change (ΔR/R 0 [%]) during deformation is used to evaluate the electromechanical property of the SHSF, where ΔR is the change of resistance with respect to its original resistance R 0 .

Humidity sensing characterization
The SHSFs based on hydrogel-fabric and untreated fabric that were both soaked with LiBr were cut into sheets of 21 mm × 11 mm in size. An AC bias voltage of 1 V with a frequency of 500 was applied at both ends. The changes in resistance and capacitance were measured by an LCR meter (a meter that measures inductance, capacitance and resistance, Tonghui, TH2832). Different relative humidities (RH) were provided by gas collection bottles containing different saturated salt solutions. Specifically, saturated potassium sulfate (K 2 SO 4 ), KCl, sodium chloride (NaCl), sodium bromide (NaBr), potassium carbonate (K 2 CO 3 ), magnesium chloride (MgCl 2 ), and lithium chloride (LiCl) solutions provided 98%, 85%, 75%, 59%, 43%, 33%, and 11% RH, respectively (calibrated by digital thermo-hygrometer [CEM, SA615]). The duration of each RH detection cycle consisted of fixed 300 s of placing sensors in collection bottles containing different saturated solutions, and the rest time was spent by placing sensors in collection bottles containing saturated LiCl solution with 11% RH for recovery.

Characterization of SHSF
By adjusting the amount of precursor adsorbed by the fabric, the precursor was only absorbed into the fiber, instead of covering the pores, thus forming a porous network. The improved composite had no difference with the unmodified fabric either to the naked eye, under optical microscope, under electron microscope, or to the touch (Figures 2B,C, S1 and S2).

Breathability
The air permeability of fabric is closely related to the comfort of wearing. We evaluated the permeability of SHSF by covering the bottle containing 10 g deionized water with different materials, such as SHSF and the unmodified fabric. Airtight plastic and an open bottle were set as a control group. As shown in Figure 2D, the SHSF showed the same vapor transmission rate as the open bottle, leaving 96% of the water left in the bottle, and the water in the bottle covered with SHSF evaporated even faster. For bottles covered with airtight plastic, almost all the water remained. It can be seen that the SHSF has good performance of permeability.
The good air permeability of the SHSF is partly due to the intrinsically good air permeability of the cotton. This is also because hydrogels are gelled at the fiber level and do not block the interwoven fiber gaps, as shown in Figure 2B. PAM hydrogels are rich in hydrophilic groups, such as hydroxyl and ─NH 2 , which play a crucial role in the absorption of water molecules. Consequently, the water absorption capacity of SHSF is stronger relative to the untreated fabric, resulting in faster water fall in the container in which the SHSF is wrapped.

Deformability
Deformability is an important property of wearable devices. We cut the SHSF into a size of 21 mm × 11 mm × 1 mm. Firstly, the sample was pre-stretched to 5% of the original length at most. Figure S3A shows a typical relative resistance variation versus strain curve when different strains are applied. The resistance increases monotonically from 15.4 to 17.2 KΩ with increased strain up to 5%, which is attributed to the geometrical effect, that is, the increased length and decreased cross-sectional area of the sensing material. The plot of response versus strain displays a linear region with the gauge factor of 2.36 ( Figure S3B). Although the tensile property of SHSF is limited, the deformation such as bending and torsion seems fantastic and favors wearing comfort. After rubbing, it returns to the original shape, reflecting the good elasticity, as indicated in Figure S4 and Movie S1.

Flame retardance
Hydrogel covered in cotton fabric makes the fabric flame-retardant. Smart sensing fabric integrated with flame-retardant performance can be better used in fire rescue. To the daily wearer, they get more escape time, which effectively reduces casualties caused by fire. The fire resistance of the unmodified cotton fabric is compared with that of the hydrogel-fabric composite by being treated with ignition in air. As shown in Figure 2E and Movie S2, the cotton fabric burned through within a few seconds when exposed to the flame, while the composite withstood the flames for a much longer period of time, with only a small portion of the edge areas caught fire after exposure to fire for 15 s. The cotton fabrics burned out completely after 6 s of continuous combustion, while the composite burned out after 26 s of continuous combustion. For additional reference, the flame retardancy of the hydrogel-gauze composite was also treated by ignition in air as in Figure 2F and similarly, the composite also had good flame-retardant properties and good structure after ignition. As shown in Figure 2G, the untreated fabric will continue to burn once ignited until they burn out, while hydrogel-fabric composite will only burn when in contact with an open flame and will not continue to burn when not in contact with a flame. Therefore, the fabric treated with PAM hydrogel has an evident flame-retardant property. By modifying fabrics with hydrogels, it is possible to design fire-resistant materials that perform better than either fabric or hydrogel, providing sufficient protection to the wearers. The flame-retardant mechanism of the fabric covered with hydrogel is mainly due to the absorption of a large amount of energy when water is heated and evaporated in the hydrogel. In the meanwhile, the hydrogel will release water vapor, carbon dioxide and other non-combustible gases to dilute the oxygen concentration when exposed to fire. After water evaporates, the hydrogel acts as an air-insulating layer on the surface of the fabric, further inhibiting combustion. In addition, to improve the water retention capacity of flame-retardant hydrogels, highly hydrated salts (calcium chloride, CaCl 2 ) can be added to hydrogel composite to extend the water retention time while improving the flame-retardant properties. 49

F I G U R E 3 (A)
Real-time conductance response of fabric soaked in lithium bromide (LiBr) solution to relative humidity (RH) ranging from 98% to 33%. (B) Experimentally obtained (dots) and exponentially fitted (red lines) conductance responses versus RH curves for the fabric sensors with and without polyacrylamide (PAM). Real-time capacitance response of hydrogel-fabric humidity sensor to RH ranging from 98% to 33% at an operating frequency of (C) 500 Hz, (D) 1000 Hz, and (E) 2000 Hz. (F) Experimentally obtained (dots) and exponentially fitted (red lines) capacitance responses versus RH curves for 500, 1000, and 2000 Hz.

Humidity sensing properties of SHSF
To verify that the moisture-sensitive property of SHSF is attributed to hydrogel, we compared the humidity responses of the fabric sensors with and without PAM. From Figure 3A, we can see that fabric soaked in LiBr solution without PAM modification also responds when exposed to different humidities, which is attributed to the positive correlation between the number of ions dissolved in water and the RH. As shown in Figure 3B, the conductance response of the fabric-hydrogel sensor at each RH was 5.5-138 times larger than that of the unmodified fabric. It can be observed that there was a linear relationship between the conductance response and RH for the fabric immersed in LiBr only, while there was an exponential relationship for the fabric-hydrogel sensor. Therefore, the hydrogel modification boosts the humidity sensitivity of the fabric, which should be attributed to the moisture adsorption ability of the hydrogel.
Then, we explored the effect of frequency of AC current applied to the hydrogel-fabric composite on the humidity sensing characteristics. The relationship between the capacitance response and the RH of the hydrogel-fabric humidity sensor was measured at 500, 1000, and 2000 Hz ( Figure 3C-E). It can be observed in Figure 3F that the capacitance responses to 98% RH reach 87,967%, 65,800%, and 41,526% at 500, 1000, and 2000 Hz, respectively, and the response decreases with increasing operating frequency. In addition, the sensor operating at 500 Hz exhibited a capacitance change close to five orders of magnitude with the variation of RH from 11% to 98%, reflecting the high sensitivity. Therefore, 500 Hz was chosen as the optimal operating frequency for the remaining experiments.
To investigate the effect of the concentration of LiBr on the performance of the humidity sensor, we measured the conductance changes ( Figure 4A-C) and the capacitance changes ( Figure S5A-C) of the fabric-hydrogel sensors at different RH after soaking in different concentrations of LiBr for 0.5 h and then calculated the corresponding responses. According to Figure 4A-C, it can be seen that the higher the RH is, the larger the conductance. Under the environment of 98% RH, the conductance of the sensor immersed in 1, 2, and 3 M LiBr solutions reached stable responses of 31,197.8%, 1972.4%, and 1483.2%, respectively ( Figure 4D). Likewise, at 98% RH, the capacitance of the sensor immersed in 1, 2, and 3 M LiBr solutions attained stable responses of 88,293.2%, 938.5%, and 759.6%, respectively ( Figure S5D,E). Among them, the conductance and the capacitance responses of the sensor immersed in 1 M LiBr solution increased by more than 288 and 429 times when the RH increased from 33% to 98%, respectively. The sensitivity of the sensor immersed in LiBr solution with a concentration of 1 M is the largest, followed by 2 and 3 M. Therefore, we chose the fabric-hydrogel immersed in LiBr solution with a concentration of 1 M for other experiments.
During the experiments, it was found that the response (the conductance change ΔG/the initial conductance G 0 ) and the sensitivity (the slope of response vs. humidity correlation) of the humidity sensor, were greatly influenced by G 0 of the sensor. The amount of ΔG of the humidity sensors immersed in LiBr at concentrations of 1, 2, and 3 M are of the same order of magnitude. However, the initial stable G 0 values of different soaking concentrations differ greatly (Figure S6). The G 0 for the soaked LiBr solution concentration of 1 M is in the order of 10 −9 , while that is 10 −6 for 2 and 3 M. Therefore, the sensitivity of the humidity sensor for soaking concentration of 1 M is much greater than that for soaking concentrations of 2 and 3 M. In addition, as the concentration of LiBr solution increases, more lithium ions and bromine ions are introduced to make the sensor have a greater conductance value at the same humidity, that is, better conductivity.
The humidity sensing mechanism is the swelling effect of water absorption by the hydrogels (Figure 4E). When the environmental RH rises, water vapor condenses on the hydrogel. Hydrophilic groups in PAM (e.g., -NH 2 ) ( Figure 4F) connect water molecules through a large number of hydrogen bonds, resulting in the swelling of the hydrogel and a decrease in the density of the polymer network. As a result, the hindering effect of the polymer network on the ion migration is alleviated, leading to an increase in conductivity. In addition, the increase in the number of ions dissolved at high RH increases the number of charge carriers. These two effects result in higher conductance at high RH. 43,48 If the difference between the RH of the two environments is greater, the conductivity change will also be more pronounced, and the response will be greater. In conclusion, it is the abundant hydrophilic groups on the PAM hydrogel on the fabric that play a crucial role in the absorption of water molecules, leading to the increased ionic conductivity and enabling the manufacture of humidity sensors with high sensitivity. Meanwhile, the increased capacitance of the hydrogel-fabric at higher RH is attributed to the increased permittivity of hydrogel upon adsorption of moisture.
As shown in Figure 4G,H, the response time of the sensor becomes longer as the concentration of LiBr solution increases. Meanwhile, introducing more conductive ions makes a greater conductance value and better conductivity at the same humidity level. In addition, the recovery time of the sensor becomes longer as the concentration of LiBr solution increases. In hydrogels, water molecules bound to ions must break the bonds between water molecules and ions to evaporate, while free water molecules do not require additional energy to break the bonds. Therefore, free water molecules evaporate more easily, as shown in Figure 4I. The higher degree of hydration of dissolved salt ions means that the bond strength between cations and water molecules is stronger, and the cations bind more water molecules. It makes water molecules evaporate harder and provides the sensor with better water retention capacity. For hydrogels containing the same salt, the steady-state conductivity increases with increasing initial dissolved salt concentration and decreases with decreasing ambient RH. 52 To further study the repeatability of the sensor, we continuously tested the conductance and capacitance response five times at 33%, 59%, and 98% RH, respectively. The conductance and capacitance responses to the same RH are almost the same in the cycling test (Figures 5A-C and S7), indicating that the humidity sensor has good repeatability. In addition to the ability to operate in a relaxed state, the sensor can also work properly under deformation, such as folding. We folded the hydrogelfabric in half, reduced the width by half with the size of 10 mm × 2.5 mm, and evaluated the humidity response ( Figure 5D). As illustrated in Figure 5E, after folding, the humidity sensor still displays a noticeable response in the range of 33%−98% RH, and the response to 98% RH reduced from 31,198% to 13,165%. The reduced sensitivity of the composite humidity sensor after folding is because the interaction surface area of the composite material with the environment is decreased ( Figure 5F). As there are fewer sites to absorb water molecules, the ability of the sensor to adsorb moisture is weakened. This sensor is moderately immune to its own folding deformation, which facilitates a stable humidity response in practical wearable applications.

Application
The hydrogel-fabric humidity sensor will activate quickly and attain a relatively high response in the first few seconds. Therefore, although it takes a few minutes for the signal to settle, the sensor is able to practically applied in real-time humidity switch detection, such as respiratory monitoring and non-contact sensing.

Respiratory monitoring
It is worth noticing that the humidity sensor is sensitive enough to detect human respiration and its frequency. Attributed to the humidity difference between the exhaled gas and the environment, the conductance of the sensor increases abruptly during exhalation, followed by a rapid recovery during inhalation ( Figure 5G). During the experiments, the sensor placed under the nose was able to acutely capture the subtle RH changes generated by the different breathing frequencies, from 15 to 56 breaths/min. This property enables the sensor to be applied to the biomedical wearables for the monitoring of the respiratory status of patients suffering from respiratory disorders such as dyspnea and asthma.

Non-contact sensing
Another application is the non-contact sensing. The skin surface of the human finger contains moisture, which leads to the changes in RH around the hydrogel-fabric sensor when the finger is at a certain distance from the surface of the sensor. The humidity sensor captures the changes occurring on its surface in the form of conductance changes, as shown in Figure 5H. The conductance response decreases with the increase in finger distance due to the declined RH. The sensitivity (slope) of the sensor is calculated to be 5.8 from the linearly fitted response versus finger height curves ( Figure 5I). Based on this property, the hydrogel-fabric sensor is capable of monitoring the finger distance at the surface based on changes in conductivity and is highly potential in non-contact human-machine interaction.

CONCLUSIONS
In summary, a flame-retardant, stretchable, breathable, and high-sensitivity humidity sensing fabric was developed by the in situ synthesis of PAM hydrogel on cotton fabric. Commercially available fabrics gain the advantages of humidity responsiveness and fire retardance after facile modification with hydrogel while preserving the excellent breathability and flexibility, which enriches the functionality of traditional fabrics. To address the intrinsic instability problem of conventional hydrogels, LiBr solution is used to treat the SHSF to improve the water retention capacity and stability. Notably, the conductance of the SHSF increased more than 311 times as the RH increased from 11% to 98%, showing excellent sensitivity. Even when the SHSF was folded, the conductance response was as high as 13,164.6%. The combination of hydrogel and fabric avoids the utilization of substrate as employed by traditional humidity sensors and smart sensing fabric, which greatly simplifies the device structure and fabrication process. SHSF will not burn as soon as it is ignited. When continuously exposed to open fire, it has a much better flame-retardant ability than ordinary cotton fabrics. The use of hydrogel as a flame retardant also avoids the release of toxic gases and decreases the amount of smoke produced by the burning of traditional halogen and phosphorus flame retardants. In addition, it shows good humidity sensing reproducibility and breathability. Based on the above characteristics, this hydrogel-based SHSF has a broad application prospect in the fields of flexible sensors, wearable electronic devices, health detection, etc.