Facile and Controllable Ultrasonic Nebulization Method for Fabricating Ti3C2Tx‐Based Strain Sensor and Monitoring of Human Motion and Sound Wave

Flexible and wearable electronic devices hold great potential in electronic skins, health monitoring systems and soft robotics. Among them, flexible strain sensors with high performance are key components for wearable health monitoring devices. However, the facile and controllable preparation of highly sensitive sensors still faces significant challenges. By virtue of excellent conductivity of 2D transition metal carbids (MXenes), this work reports a facile and low‐cost fabrication strategy for large‐scale production of strain sensors. The sensitive layer is deposited on flexible interdigital electrodes by ultrasonic nebulization of Ti3C2Tx nanosheets. By controlling the nebulization time, different thicknesses of Ti3C2Tx films has a great influence on the performance of strain sensors. The Ti3C2Tx‐based strain sensor exhibits good sensing performances such as high GF (19.1) in the low strain range (≈0.25%–1.14%), short response time (0.7 s), and stable durability (over 1000 cycles). In practice, the potential applications of the strain sensor in sound frequency detection, human physiological signal monitoring and facial expression recognition are demonstrated. Finally, this work integrates the strain sensor with a miniaturized analyzer to assemble a wearable motion monitoring device for mobile healthcare. This study provides a facile strategy for fabricating flexible strain sensors in the field of wearable electronics.


Introduction
In recent years, flexible electronic devices in which strain sensors as important components have attracted much attention due to their huge potential applications in artificial electronic The ORCID identification number(s) for the author(s) of this article can be found under https://doi.org/10.1002/admi.202300030. tension and compression, the sliding and stacking of Ti 3 C 2 T x nanosheets in strain sensors change the number and length of conductive paths, as well as the resulted resistance variation, [39] which makes Ti 3 C 2 T x have broad application prospects in strain sensors.
The polar functional groups which introduce to Ti 3 C 2 T x surface during the preparation process make them hydrophilic and easily to disperse into aqueous suspensions or inks. [40] Therefore, Ti 3 C 2 T x colloidal dispersions can be deposited by digital printing (inkjet printing, 3D printing), nondigital printing (screen printing, transfer printing), ink writing, and a variety of coating techniques (spray coating, rod-coating, vacuum filtration, drop-casting, etc.). [41] Although the printing methods have great advantages in resolution and patterning, it has strict requirements on the rheological and morphological properties of inks, [42,43] which limited their practical applications. In this regard, coating methods are low-cost technology without being reliant on the properties of inks. For instance, spray coating is a widely used deposition technique in industry. It works by jetting inks under the pressure of high-speed air flow which can cover a large area quickly. However, the quality and uniformity of films are susceptible to influence by airflow velocity, inject distance, and substrate temperature. In addition, spray coating requires a large amount of raw materials. [44] Other coating techniques also have different limitations in applications of 2D materials for flexible devices. For example, rod coating is not good at preparing thin films. [45] Drop casting inevitably generate coffee rings which leading to inhomogeneous film morphology. [46] These coating methods are more difficult to achieve controllability in terms of tunable thickness and film uniformity.
Ultrasonic nebulization spray coating is a relatively mild and thickness controllable deposition technique. The transducer in the ultrasonic nebulizer generates an ultrasonic frequency in the liquid. When a certain power is reached, capillary waves are formed on the surface of the liquid droplets, and the waves are ejected to form a fine mist. The droplets in the fine mist are small enough to avoid the coffee ring effect. [47] Ultrasonic nebulization can deposit films over large areas and is expected to be mass-produced in industrial applications without causing any damage to the material. Moreover, the deposited films are relatively uniform, because the droplets produced by ultrasonic nebulization are homogeneous. [48][49][50][51][52][53][54][55][56] Taking into account the factors mentioned above, in this paper, we report a flexible Ti 3 C 2 T x -based strain sensor with high sensitivity. For this sensor, the conductive Ti 3 C 2 T x film and the Ag/polyimide (PI) interdigital electrode were used as the sensitive layer and the electrode layer, respectively. Compared with other methods of preparing the Ti 3 C 2 T x film of strain sensors, such as vacuum filtering, [57] printing, [58] and dip-coating, [59] this strategy is even simpler. The electrical-mechanical performance of Ti 3 C 2 T x -based strain sensors was investigated by real time electrical signal test system. The results showing excellent sensing performances for strain sensors, such as high GF (19.1) in the low strain range (≈0.25%-1.14%), short response time (0.7 s), and stable durability (over 1000 cycles). In addition, when the strain sensor bending outward/inward, Ti 3 C 2 T x sheets adopt a crack/overlap conformation thus increasing/decreasing the contact resistance, resulting in different gauge factors and coefficients of determination. Benefiting of the above performances, we demonstrated the strain sensors used for detecting different frequencies and decibels of sound wave, monitoring the signals of human artery pulse, and distinguishing different facial expressions. We also integrated the flexible strain sensor with a miniaturized analyzer to assemble a wearable motion monitoring device for mobile healthcare.

Preparation and Characterization of MXene Ti 3 C 2 T x
Ti 3 C 2 T x sheets were obtained by etching the Al atom layer from Ti 3 AlC 2 powders with LiF/HCl. [60] And it required further sonication to obtain a large number of monolayers Ti 3 C 2 T x (Figure 1a) which were utilized as sensitive elements. Figure 1b shows the preparation process of the Ti 3 C 2 T x -based strain sensor. First, the Ag interdigital electrodes on polyimide (PI) substrate were prepared by screen printing. Subsequently, Ti 3 C 2 T x sheets were uniformly nebulized on interdigital electrodes and finally packaged to meet further testing requirements. The successful synthesis of Ti 3 C 2 T x sheets is crucial for our strain sensors. After LiF/HCl etching, the Al atoms layer in MAX phase (Ti 3 AlC 2 , Figure S1a, Supporting Information) was removed, and the layered structure of Ti 3 C 2 T x was formed ( Figure S1b, Supporting Information). Nacre-like layered stacking of Ti 3 C 2 T x can be seen from the section SEM image as shown in Figure 1c. The transmission electron microscope (TEM) image of single-layered Ti 3 C 2 T x sheets in Figure 1d confirms the 2D structure of Ti 3 C 2 T x . The lateral size of Ti 3 C 2 T x sheets is generally in the range of several hundred nanometers to one micron. The selected area electron diffraction (SAED) pattern (inset in Figure 1d) indicates the hexagonal crystal structure of Ti 3 C 2 T x . [57] The XRD spectra of Ti 3 AlC 2 and Ti 3 C 2 T x are shown in Figure 1e. The (002) diffraction peak shifts to a low angle, and the strong 2θ diffraction peak at 39° disappears after etching, indicating that Al atoms layer were removed from the Ti 3 AlC 2 structure by selective etching, which are consistent with previous studies. [61][62][63][64] The EDS element maps ( Figure S2a, Supporting Information) of the Ti 3 C 2 T x sheet illustrate the uniform distribution of Ti, C, O, and F elements. The thickness of Ti 3 C 2 T x sheets were investigated by atomic force microscope (AFM) ( Figure S2b, Supporting Information). From the height cutaway view, the thickness is ≈1.35 nm, which indicates that Ti 3 C 2 T x sheet is monolayer. [65]

Optical and Electrical Properties of Ti 3 C 2 T x Films
The efficient and convenient ultrasonic nebulization technique was used to deposit Ti 3 C 2 T x sheets on flexible Ag interdigital electrodes printed on PI substrate. The Ti 3 C 2 T x on the substrate is distributed uniformly and firmly bonded. The concentration of the delaminated Ti 3 C 2 T x colloidal solution used for nebulization was 0.2 mg mL −1 . The deposition thickness of Ti 3 C 2 T x sensitive layers can be controlled by adjusting nebulization time. Figure 2a shows optical photos of Ti 3 C 2 T x nebulized on quartz glass with different nebulization time (1-6 min).

www.advmatinterfaces.de
As seen in the photographs of Ti 3 C 2 T x nebulized on quartz glass with different nebulization time (1-6 min) in Figure 2a, the quartz glasses appear the greyish black and gradually deepen over the nebulization time, indicating more Ti 3 C 2 T x sheets deposited on the substrate. In ultraviolet-visible (UV-vis) spectra (Figure 2b), the Ti 3 C 2 T x films obtained at different nebulization time have obvious absorption in the UV region of 350-500 nm, and a broad absorption peak between 700-800 nm. In the visible light region, the curve is relatively flat, which are consistent with previous research. [66][67][68] As shown in Figure 2c, with increasing of the nebulization time, the transmittance of films at 550 nm gradually decreases (82.8% for 1 min to 12.03% for 6 min). The electrical conductivities of Ti 3 C 2 T x films were measured by the four-point probe method. As the nebulization time increases, the distributions of sheet resistances are more uniform ( Figure S3, Supporting Information). As shown in Figure 2d, the sheet resistances of Ti 3 C 2 T x films decrease with increasing nebulization time. After 3 min of nebulization, the sheet resistances gradually stabilize at about 500 Ω sq −1 . The thickness of Ti 3 C 2 T x films gradually increases from 13 (1 min nebulization) to 313.7 nm (6 min nebulization) (Figure 2e), which measured by the crosssectional SEM images in Figure S4, Supporting Information. It is necessary to balance these parameters (transmittance, sheet resistance and thickness) for different application, so an appropriate nebulization time should be selected.

Electrical-Mechanical Performance of the Strain Sensor
The current-voltage (I-V) characteristic of the ultrasonic nebulization-assisted Ti 3 C 2 T x -based strain sensor with different nebulization time is shown in Figure 3a. With the voltage increases from −1 to 1 V, the symmetrical I-V curves confirm that the typical ohmic contact between Ag electrodes and Ti 3 C 2 T x sheets. [69] This indicated the reliability of the ultrasonic nebulization-assisted Ti 3 C 2 T x -based strain sensor, which has great potential in human motion monitoring. To study the electricalmechanical response of strain sensors, an electrical signal test system including a high-precision tensile machine which can accurately control the deformation of the sensor and an electrical signal detection device linked with a computer for signal collection was used (Figure 3c). The strain sensor was fixed on the tensile machine in a natural state to test the electricalmechanical characteristics quantitatively. By Controlling the movement of the tensile machine, the strain-sensing properties of the sensor can be tested by bending outward and inward. When the downward distance of the tensile machine is dL, the bending radius (r) can be calculated with equation (1): [70,71] 2 ( / ) /12

www.advmatinterfaces.de
where L and h represent the initial length and the thickness of the sensor, respectively (Figure 3b). Suppose the arc length l is the conductive path length which defined as l = θ × (r + d), d = h/2 , where θ is the central angle. Therefore, the strain ε can be calculated as equation (2) : where θ 0 , θ t , r 0 , and r t denote the central angle and the bending radius before and after bending, respectively. Because r 0 >> d, θ t r t ≈ θ 0 r 0 , Equation (2) can be converted to As the sensor is flat before bending, r 0 tends to infinity, ε = d/r t = h/2r t . Similarly, the strain can be calculated as ε = − h/2r t when bending inward. Table S1, Supporting Information, shows the downward distance (dL) setting in the experiments, the corresponding bending radius (r) and the strain (ε).
The electrical-mechanical performance of the Ti 3 C 2 T x -based strain sensor with different nebulization time was investigated to determine the appropriate nebulization time. Figures S5  and S6, Supporting Information, shows the changes of normalized resistance (ΔR/R 0 ) by gradually increasing the degree of strains (from 0.25% to 1.14%) when outward ( Figure S5, Supporting Information) and inward ( Figure S6, Supporting Information) bending. It can be seen that when the nebulization time is less than 3 min, the normalized resistances do not change significantly under the strain ranges from 1.02% to 1.14%, indicating that the strain sensor can work under less than 1.02% strain. Nevertheless, the strain sensor with 4 min nebulization time can recognize different degrees of bending. In addition, when the strain sensor was bent outward and inward, the responses of resistance changes were positive and negative, respectively. Notably, all the sensors that nebulized for 2-6 min can recognize strains in the full range from 0.25% to 1.14%. Moreover, the stable response can be obtained by repeating the test three times under each strain, indicating the stability of the device (Figures S5 and S6, Supporting Information). We selected the strain sensor with 4 min nebulization time for testing the ability of continuous response under both outward and inward bending, as shown in Figure 3d. It can be seen that, the strain sensor responded in succession with increasing the strain ranges from 0.25% to 1.14%, indicating the good continuous responsiveness and a small response time of less than 0.7 s ( Figure S7, Supporting Information), allowing accurate and continuous measurements with the strain change.

www.advmatinterfaces.de
Generally, two characteristics, namely linear working window and strain sensitivity, were used to evaluate the performance of strain sensors. The linear working window is the strain sensing range where the resistance increased linearly with the strain (R 2 ≥ 0.95, where R 2 is the coefficient of determination which normally ranging from 0 to 1). R 2 represents a statistical measure of the degree of replication of the experimental results by the linear regression model. The strain sensitivity is usually expressed by the gauge factor (GF), defined as GF = (ΔR/R 0 )/ε, where ΔR = R − R 0 (R 0 is raw resistance, R is resistance under deformation), ε is the applied strain. The GF values of the different strain sensors can be obtained by calculating the slopes of the linearly fitted lines which were plotted the normalized resistance changes (ΔR/R 0 ) versus axial strains (ε), [72,73] which are shown in Figure 3e (outward) and Figure 3f (inward). The statistical results of the GF and coefficients of determination the strain sensor with different nebulization time are shown in Figure S8, Supporting Information. It is critical for strain sensors to achieve high GF without sacrificing data linearity. As a result, when bending outward, the strain sensor with 4 min nebulization time balanced the two factors (GF = 19.1, R 2 = 0.993), when bending inward, the strain sensor with 2 min nebulization time shows the best performance (GF = 12.6 and R 2 = 0.992). Then, the stability and durability of the nebulized Ti 3 C 2 T x -based strain sensors were tested by bending outward and inward more than 1000 times. Figure 3g is the real time response when the strain sensors bending outward (4 min nebulization time) and inward (2 min nebulization time) with strain 0.43%, respectively. It can be seen that there is no significant change in waveform and resistance g) The cycling stability test at a strain rate of 200 mm min −1 for outward (4 min nebulization time) and inward (2 min nebulization time) bending, respectively. h) Schematic illustration of the sensing mechanism and photographs of the strain sensor as a variable resistor to control the brightness of light-emitting diode.

www.advmatinterfaces.de
during the whole test, indicating that the stable durability and long-term stability of the strain sensors when bending repeatedly at high frequency. In order to ensure optimal practical applications, the strain sensor with 4 min nebulization time was selected for following test, due to its superior performance when subjected to outward bending and an acceptable range of performance (GF = 10.3 and R 2 = 0.977) when subjected to inward bending. Figure 3h reveals the working mechanism of the Ti 3 C 2 T x -based strain sensor, which is attributed to the resistance change originating from the reversible contacts among Ti 3 C 2 T x sheets. Specifically, when the sensor was bent outward, the distance between the Ti 3 C 2 T x sheets increases, leading to an increase in the carrier transfer path distance, resulting in the increase of the contact resistance. Conversely, when the strain sensor was bent inward, the Ti 3 C 2 T x sheets had more overlap which creates more conductive channels, so the resistance becomes lower. [32] As a validation, we used the strain sensor as a variable resistor to control the brightness of light-emitting diode (LED) (photographs in Figure 3h). When bending inward, the resistance decreased leading to the LED brighter (bottom photograph) than the original state (middle photograph). On the contrary, when bending outward, the LED became dimmer (top photograph) as the resistance increased.

Sound Wave Detection
By virtue of the high sensitivity and short response time, the strain sensors hold potential for vibration detection. First, the utility of the strain sensors as a flexible device for monitoring the frequency of sound wave was examined, because the sound waves can vibrate the strain sensors slightly. We attached the strain sensors on a thin membrane and covered on a loudspeaker to test the different frequencies of sound wave by pronouncing an English word "sensor." We observed that the time difference between the peak of signals is consistent with the frequency of the sound wave as displayed in Figure 4a. Besides this, there are significant changes in the peak values of the word "sensor" pronouncing at different decibels (Figure 4b), indicating that the strain sensor can identify the slight and rapid changes during the vibration. In addition, when the loudspeaker pronouncing different words or phrases, the sensor can determine the number of syllables in the words or phrases (Figure 4c). The above results illustrate that our strain sensors could be as the sound frequency detector, based on the signals originating from the strain caused by sound vibration.

Real-Time Monitoring of Various Human Facial Expressions and Physiological Movements
The excellent performances of the nebulized Ti 3 C 2 T x -based strain sensor also enable it to be used as a wearable device for c) The signals of the strain sensor when the loudspeaker pronouncing words or phrases containing different syllables.

www.advmatinterfaces.de
real-time human facial expressions and physiological movements monitoring. The human pulse wave is an important physiological signal that can reflect health conditions such as blood pressure, heart rate, arteriosclerosis, etc.. [315,74,75] The utility of our strain sensors as a flexible device for monitoring the pulse condition of human in real time is shown in Figure 5a. The characteristic peak positions can be identified from the pulse waveform, including P (Percussion), T (Tidal), and D (Dicrotic) waveforms. According to the peak value of P and T, the arterial stiffness can be calculated, which becomes reference data for disease diagnosis, [76] suggesting the strain sensor has the potential to contribute to disease diagnosis and medical detection. In addition, the strain sensor also can monitor human micro-expression such as smiling (Figure 5b) and frowning (Figure 5c). The strain sensors were placed on the laughing muscle and eyebrows, and the corresponding resistive responses were recorded in real time during repeated smiles and frowns. The normalized resistance values of the strain sensor increased first and then returned to the initial state during the monitoring of micro-expression. When the model made the same expression repeatedly, the response signal waveform remained initial state, but different expression showed different waveforms, which makes it possible to monitor human micro-expressions. Figure 5d shows that the strain sensor can clearly recognize the swallowing action when it is attached to the throat. The collected signals for the nebulized Ti 3 C 2 T x -based strain sensor which as wearable device resulted from different degrees of stretching or shrinking strain of human skin. In addition, the strain (ε) of the strain sensor for detecting physiological movements can be calculated according to the linear fitting equation of the response (ΔR/R 0 ) versus the strain (ε) when bending outward ( Figure S10 and discussion, Supporting Information). Taking the example for monitoring the pulse, the response of strain sensor is less than 0.15%, corresponding to 0.23% of the strain (ε). Because the bending radius (r t ) and the strain (ε) have the relationship denoted by ε = d/r t = h/2r t when bending outward, the bending radius (r t ) of the strain sensor can be calculated which is 43.4783 mm.

www.advmatinterfaces.de
We summarized the bending radiuses (r t ) of different physiological movement's detection in Table S2, Supporting Information. Obviously, the bending radius (r t ) decreases with the intensity of physiological movements, suggesting different degrees of deformation of the strain sensor. The development of wearable electronic devices has important implications for healthcare and disease prevention. Therefore, we connected the strain sensor with a micro-analyzer ( Figure S9, Supporting Information) to construct a portable wearable electronic device for monitoring human motion. Figure 5e shows the circuit diagram of the wearable system and the joint motion signals in real time (Supplementary video 1), indicating that the strain sensor could be a convenient method for human motion monitoring. Therefore, the nebulized Ti 3 C 2 T x -based strain sensor exhibits broad application prospects in vibration detection and human physiological movement monitoring.

Conclusion
In summary, we have successfully developed a cost-effective, easy-manufactured, mass-produced, and highly sensitive Ti 3 C 2 T x -based strain sensor. The uniform Ti 3 C 2 T x film was formed on Ag electrodes by ultrasonic nebulization spray coating deposition onto a flexible polymeric support, yielding reversible resistance sensitive layer among Ti 3 C 2 T x sheets, so that the strain sensors exhibited good performance when both bending outward and inward. The nebulized time is a key factor influence on the GF and coefficients of determination. The Ti 3 C 2 T x -based strain sensor exhibits excellent sensing performances such as high GF (19.1) in the low strain range (≈0.25%-1.14%), short response time (0.7 s), and stable durability (over 1000 cycles). The strain sensor has great potential for vibration detection and human physiological movement monitoring. The different frequencies and decibels of sound wave could be distinguished by time and intensity of the response. The signals of human artery pulse and different facial expressions could be collected by simply placing the strain sensor onto the body to extract human physiological parameters for the evaluation of health. Therefore, this work demonstrates a promising strategy for assembling the flexible wearable strain sensor with good performance for various practical applications, such as sound frequency detection, human motion detection, and smart E-skin.

Experimental Section
Synthetic Procedures of Ti 3 C 2 T x : Ti 3 C 2 T x nanosheets were prepared by removing Al atom layer from Ti 3 AlC 2 powder. First, 15 mL of concentrated HCl (12 m) was mixed with distilled water (5 mL) to obtain 9 m of diluted HCl solution (20 mL). Second, 1.6 g LiF powder was added into the diluted HCl solution with magnetic stirring for 10 min. Third, 1 g Ti 3 AlC 2 powder was slowly added to the mixture, then the mixture was kept at 35 °C for 24 h under constant stirring. Finally, the synthetic product after the reaction was repeatedly washed with DI water until the supernatant reached a pH ≈7. Then exfoliated by sonication and centrifuged at 3500 rpm to obtain few or single layers of MXene Ti 3 C 2 T x .
Fabrication of Strain Sensor: First, the PI substrate was cleaned with acetone, and then the interdigital electrode was printed on the PI substrate by screen printing. Add the delaminated Ti 3 C 2 T x colloidal solution with a concentration of 0.2 mg mL −1 into the self-made ultrasonic nebulizer. Then, place the prepared PI-based interdigital electrodes on the heating plate with 80 °C, turn on the atomizer, and the nebulization periods from 1 to 6 min. In the nebulization process, the utilization of an ultrasonic nebulizer possessing consistent power and performance was employed, and all devices utilized the same batch of aqueous MXene solution. The nebulizer was securely positioned at an optimal height to prevent movement, and it is paramount to note that all devices were situated at the same nebulization center during the nebulization procedure. After the electrode was dried, the copper wires were connected with silver paste, and then the device was sealed with polydimethylsiloxane (PDMS).
Electrical-Mechanical Characteristics Test: The electrical-mechanical characteristics of the Ti 3 C 2 T x -based strain sensor were tested with an electrical signal test system which including a tension machine (mark-10) and an electrochemical workstation (CHI660E) linked with a computer for signal collection.
Sound Wave and Human Motion Test: In practice, the strain sensor was attached on preservative film and covered on a loudspeaker to distinguish frequencies and decibels of a word or phrase. In addition, the strain sensor was also fixed on the skin of different parts of human body (such as wrist, cheek, procerus, and throat) to real-time monitor human physiological signals and facial expressions. The experiments involving the human motion monitoring of the sensors were performed with the full, informed consent of all participants, who are also authors of the manuscript. In addition, ethics committee approval was not required to perform these experiments.
Statistics Analysis: All sensing date are presented as normalized responses: where R is the response resistance and R 0 is the resistive baseline. In Figure 2c-e, the mean value represents the average transmittance, sheet resistances and thicknesses of the MXene films with different nebulization time, respectively. Results are depicted as mean values ± standard deviation derived from three independent experiments.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.