A portable triboelectric spirometer for wireless pulmonary function monitoring.

Coronavirus disease 2019 (COVID-19) as a severe acute respiratory syndrome infection has spread rapidly across the world since its emergence in 2019 and drastically altered our way of life. Patients who have recovered from COVID-19 may still face persisting respiratory damage from the virus, necessitating long-term supervision after discharge to closely assess pulmonary function during rehabilitation. Therefore, developing portable spirometers for pulmonary function tests is of great significance for convenient home-based monitoring during recovery. Here, we propose a wireless, portable pulmonary function monitor for rehabilitation care after COVID-19. It is composed of a breath-to-electrical (BTE) sensor, a signal processing circuit, and a Bluetooth communication unit. The BTE sensor, with a compact size and light weight of 2.5 cm3 and 1.8 g respectively, is capable of converting respiratory biomechanical motions into considerable electrical signals. The output signal stability is greater than 93% under 35%-81% humidity, which allows for ideal expiration airflow sensing. Through a wireless communication circuit system, the signals can be received by a mobile terminal and processed into important physiological parameters, such as forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC). The FEV1/FVC ratio is then calculated to further evaluate pulmonary function of testers. Through these measurement methods, the acquired pulmonary function parameters are shown to exhibit high accuracy (>97%) in comparison to a commercial spirometer. The practical design of the self-powered flow spirometer presents a low-cost and convenient method for pulmonary function monitoring during rehabilitation from COVID-19.

Introduction

Within merely one year, coronavirus disease 2019 (COVID-19), an infectious disease with potential for severe lung damage (Fig. 1 A), has infected tens of millions of people worldwide and killed millions of others (Zhao et al., 2020; Fajnzylber et al., 2020). To date, COVID-19 has not yet been effectively treated, and the pulmonary function of surviving patients still requires long-term monitoring after discharge (You et al., 2020). Pulmonary function tests (PFTs) based on spirometry are noninvasive diagnostic tools that monitor lung function through recording volume-flow curves of respiration, thereby providing critical pulmonary function parameters (Dirksen et al., 1998), such as forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC). It has been reported that both FEV1 and FVC values are generally smaller than the lower normality values in discharged COVID-19 patients, whereas FEV1/FVC is above the upper limit of normality values at the initial stage of clinical recovery (Fumagalli et al., 2020). During the rehabilitation period, the above indicators will return to normal, with both the FEV1 and FVC increasing as their ratio decreasing. Therefore, the spirometer is an important instrument recommended for daily wear that monitors and evaluates pulmonary function for rehabilitation care after COVID-19 (Chung et al., 2019). Conventional spirometry facilities are typically restricted to hospitals, and the growing number of COVID-19 patients may potentially cause extreme shortages of medical supplies. In fact, many hospitals are now no longer providing PFT services due to pandemic-related risks, in order to reduce cross-infection rates (Kouri et al., 2020). Moreover, requiring patients to regularly visit the hospital during lockdown is particularly inconvenient and may contribute to growing infection rates as well. An ideal approach would be to develop innovative and cost-effective point-of-care testing instruments featuring portable and wireless systems, through which discharged patients can self-test at home while their physiological signals are transmitted to mobile terminals and monitored on the cloud by physicians (Fig. 1B).

Fig. 1

Schematic illustrations of the wireless spirometer. (A) Diagram of a patient suffering from COVID-19. (B) An ideal approach to monitor the patient's pulmonary function through a wireless spirometer. (C) Photograph and (D) vertical section of our proposed PPF monitor based on a BTE sensor. Scale bar, 5 cm. (E) Photograph of the circuit module including a signal processing circuit and a Bluetooth communication unit. (F) Process flow chart of our PPF monitoring system, displaying the paths from signal acquisition (green), signal processing (yellow), wireless transmission (light red) to data analysis and results display (blue) through a mobile terminal (along the arrowhead direction). (G) Photograph and (H) diagram of the BTE sensor. Scale bar, 1 cm. (I) SEM image of the PTFE micro/nanoparticles sprayed onto both surfaces of the PTFE membrane. Scale bar, 10 μm. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Conventional spirometers are categorized into volume-measurement and flow-measurement types (Graham et al., 2019; Leocadio et al., 2019). Compared to volume spirometers, flow spirometers are more compact, allowing for better adaptability to any minimization of test instruments. Typically, various sensors are employed to monitor the velocity of airflow and integrate the volume over time. These sensors are generally classified into four categories according to their individualized working mechanisms: differential pressure pneumotachometers, hot wire anemometers, rotating vane spirometers, and ultrasonic flowmeters (Zhou et al., 2019). However, the wide-range adoption of these spirometers has been overshadowed by the following limitations. Firstly, because these large devices require frequent maintenance, they are typically only available in hospitals and therefore inaccessible for the home-quarantined patient during the COVID-19 pandemic. Moreover, using the same equipment to take measurements of different patients inherently possesses a large probability for cross-infection. Even if commercial home-based instruments are used instead, they are expensive due to their sophisticated structures and mechanisms (West and Theron, 2015), which largely increases the financial burden for COVID-19 patients. In this regard, developing a portable and low-cost spirometer based on novel sensing principles is imperative for pulmonary function monitoring during recovery from COVID-19.

Self-powered sensors based on emerging triboelectric nanogenerator technology have been demonstrated to possess promising sensing capabilities with numerous advantages, such as simple structure, high robustness, and low cost (Wang, 2014; Pu et al., 2017; Zhang et al., 2017; Xiao et al., 2021; Zhao et al., 2021; Conta et al., 2021; Peng, 2021; Mahmud et al., 2021; Chen et al., 2016; Yan, 2020; Xu et al., 2020). Such sensors are capable of generating significant electrical signals in response to ambient stimuli without external power supply (Huang et al., 2021; Su et al., 2020a; Tat et al., 2021; Zou et al., 2020; Zou, 2021). Moreover, their output signals can be directly used as sensing signals to reflect the dynamic changes of their sensing elements (Zhang et al., 2021; Su et al., 2020b; Jiang et al., 2020; Wang et al., 2018). Herein, we report a portable spirometer based on triboelectric technology. It primarily consists of a small and light-weight breath-to-electrical (BTE) sensor (with a volume of 2.5 cm3 and a weight of 1.8 g), a signal processing circuit, and a Bluetooth communication unit. When airflow is conducted into the mouthpiece, the BTE sensor is able to convert the breathing airflow into vibrations of the triboelectric membrane and thereby generate the electrical signals. The frequency of the electrical current signal was chosen as the primary sensing parameter because of its linear relationship with airflow speed. The detectable flow rate of the BTE sensors can reach as low as 0.65 L/s (corresponding to an airflow speed of 1.3 m/s). Moreover, the sensor exhibits good stability (>93%) under varying humidity conditions (from 35% to 81%). To effectively collect the electrical data and achieve wireless transmission, the circuit board is rationally designed to include a bridge rectifier, a micro-controller unit (MCU), and a Bluetooth chip, which greatly improves the portability of the whole system. Furthermore, we developed an APP that is capable of displaying the pulmonary function parameters FVC, FEV1, and FEV1/FVC ratio in a user-friendly manner in real-time. The obtained results have a relative error of less than 3% in comparison to the commercial spirometer, demonstrating the practical potential of the wireless spirometer for point-of-care and home-based PFTs for rehabilitation care after COVID-19.

Results and discussion

Structure design

The portable pulmonary function (PPF) monitor is primarily composed of a mouthpiece featuring an internal BTE sensor, a black surface enclosure (Fig. 1C), and a circuit module positioned in the lower part of the monitor that collects and transmits electrical signals (Fig. 1D and E, and Supplementary Video S1). The whole device has a volume of 233 cm3 and a weight of 95.4 g, demonstrating its extremely high portability. As one exhaled into the mouthpiece, an alternating current was generated from the air flow induced membrane vibration, which was eventually collected via the circuit module. The following signal processing procedure includes three main steps. Firstly, the rectifier bridge transforms the alternating signals into direct current signals that are recognized by the analog-to-digital converter (ADC). Secondly, the ADC converts analog signals into digital signals that can be transmitted wirelessly. Finally, when the Bluetooth low energy (BLE) module receives digital signals, it will switch on and transmit the data to a mobile terminal. A flow diagram of the PPF monitor system is shown in Fig. 1f.

Supplementary video related to this article can be found at https://doi.org/10.1016/j.bios.2021.113329

The following is the supplementary data related to this

article:

The structure of the BTE sensor is schematically illustrated in Fig. 1g. The device is principally composed of an acrylic framework, two aluminum (Al) plates forming an angle of 15°, and a polytetrafluoroethylene (PTFE) membrane that can flutter in between (Fig. 1H). Each lateral edge of the Al plates was chamfered into a wedge structure, acting as a wind guiding channel, as shown in Supplementary Fig. S1. This design allows the PTFE membrane to flutter even at very low flow rates, down to 0.65 L/s (corresponding to an airflow speed of 1.3 m/s), resulting in electricity generation based on contact electrification and electrostatic induction (Zhu et al., 2014; Cui et al., 2020; Gong et al., 2020). The PTFE and Al were chosen as triboelectric materials due to their large different tendency to gain or lose electrons in contact electrification (Zou et al., 2019). PTFE micro/nanoparticles were sprayed onto both surfaces of the PTFE membrane as a nanoscale surface modification to enhance the effective friction area and hence the triboelectric effect (Zhu et al., 2013; Zhou et al., 2020; Chen et al., 2017, 2020). As shown in Fig. 1i, PTFE micro/nanoparticles were distributed densely and uniformly across the membrane which can greatly enhance the surface area of PTFE to increase the number of the surface tribo-charges. If micro/nanoparticles are sparse, the triboelectric effect will be weakened and the electric output will decrease significantly. The BTE sensor has a volume of 2.5 cm3 (1.3 cm × 0.8 cm × 2.4 cm) and a weight of 1.8 g, which are 26 and 11 times, respectively, smaller than the sensor of a commercial spirometer (Contec SP10) (Supplementary Fig. S2 and Table S1). The small size of the BTE sensor allows it to be easily installed inside the mouthpiece, acting as a self-powered airflow speed sensor for PPF monitoring in spirometry.

Working mechanism of breath-to-electrical (BTE) conversion

When airflow is exhaled into the mouthpiece and passes through the PTFE membrane, vortices are generated at the left and right end of the membrane, due to the Kármán vortex (Xu et al., 2017), forcing the PTFE to flutter periodically and contact alternatively with the two Al plates (Supplementary Fig. S3 and Supplementary Video S2). After several periods of contact and separation, both sides of the PTFE membrane become negatively charged and the Al plates become positively charged as a result of the triboelectric effect. When the PTFE membrane is in the middle stage of the device as shown in Fig. 2 A(I), there is no potential difference between the two Al electrodes due to electrostatic equilibrium. If the PTFE membrane approaches one of the Al electrodes, such as the left electrode as shown in Fig. 2A(II), a current flow from right to left through an external load due to electrostatic induction (Zhou et al., 2020; Meng et al., 2020; Zhou et al., 2020; Jin et al., 2020; Zhou et al., 2020; Deng et al., 2020). When the PTFE membrane moves towards the right electrode (Fig. 2A(III)-(IV)), which is clearly revealed by the high-speed camera (Fig. 2B and C), a reverse current flows back from the left to the right electrodes. Therefore, cyclic electricity generation is accomplished with the respiration-induced fluttering of the PTFE membrane.

Fig. 2

Working mechanism and electrical outputs of BTE Sensor. (A) Working mechanism of the BTE sensor. (B), (C) Vibration patterns of the PTFE membrane at different positions. Scale bar, 3 mm. (D) Photograph of a testing platform for investigate the electrical signals of the BTE sensor under different conditions. Scale bar, 5 cm. (E) QSC, (F) VOC and (G) ISC of the BTE sensor at varying airflow speeds.

Supplementary video related to this article can be found at https://doi.org/10.1016/j.bios.2021.113329

The following is the supplementary data related to this

article:

To obtain a quantitative understanding about the working mechanism, a three-dimensional COMSOL finite element simulation was employed to simulate the electrical potential distributions between two electrodes under open-circuit conditions (Supplementary Fig. S4 and Supplementary Video S3). A tribo-charge density of 3 nC/cm2 was assumed on the surfaces of negatively-charged PTFE and positively-charged Al. The potential difference between two electrodes is defined as: , where V is the open circuit voltage, with U and U being the electrical potentials of the left and right electrodes, respectively. The simulation covers the entire electricity generation cycle as discussed above, with the PTFE membrane first approaching the right electrode and then moving towards the left electrode. The simulated result of V associated with the position of PTFE membrane is summarized in the Supplementary Fig. S5.

Supplementary video related to this article can be found at https://doi.org/10.1016/j.bios.2021.113329

The following is the supplementary data related to this

article:

A testing platform was designed and established to investigate the BTE signal conversion under different circumstances (Fig. 2D). To quantitatively characterize the conversion process, an air hose was employed to control the input airflow speed to the mouthpiece, and a commercial hot-wire anemometer recorded the airflow speed in real time. The whole measurement setup was then placed inside an acrylic box with a hydro-thermograph to monitor the temperature and humidity. With the above platform, the relationship between electrical outputs and airflow speeds was systemically studied. The temperature and the humidity inside the acrylic box during the test were 16.7 °C and 54.5%, respectively. Fig. 2E shows the results of transferred charge (Q SC) between electrodes under short-circuit condition. As airflow speed increases from 2.9 m/s to 22.1 m/s, the peak Q SC gradually increases from 0.9 nC to 3.2 nC, resulting from increasing contact areas as the PTFE membrane (with micro/nanoparticles on the surface) collided intensively and closely with the Al plates at larger airflow speeds. However, when airflow speed is further increased, Q SC remains almost unchanged due to the saturation of surface tribo-charges, as shown in Supplementary Fig. S6a. A similar trend is observed in V OC as shown in Fig. 2F and Supplementary Fig. S6b, which can be explained by the following classical equation of TENGs (Wang et al., 2015, 2016):where C TENG is the device capacitance determined by the TENG structure, which remains constant during the experiment. For the short circuit current (I SC), it increases with the airflow speed from 0.2 μA at 2.9 m/s to 4.5 μA at 22.1 m/s (Fig. 2G). Unlike the changing trend of Q SC and V OC, I continues to increase to 6.2 μA with the airflow speed increasing to 24.5 m/s (Supplementary Figs. S6c and S7). For TENG, I SC is determined by the following equation:where t is the time. Hence, I is determined by two aspects: the transfer of Q SC and the transferred time. As the airflow speed begins to increase, both Q SC and the fluctuating frequency of the PTFE membrane increase and contribute to increasing I SC. After Q SC reaches its saturated value, the oscillation frequency continues to grow with the airflow speed, leading to the continuous enhancement in I (Supplementary Fig. S8). According to the results discussed above, airflow speed is observed to affect Q SC, V OC, and I SC in different ways, with both Q SC and V OC saturated at high speed, while I SC does not reach any saturation. Therefore, I SC is chosen as a key signal for sensing the breath airflow speed.

Performance of BTE sensor

Frequency and amplitude are the two main parameters of the I signal, both of which were studied for sensing performance. The obvious nonlinear relationship between the current amplitude and the airflow speed is shown in Supplementary Fig. S9. Moreover, large amplitude fluctuations at high speeds due to the irregular flutter of the PTFE membranes exacerbate the unreliability of sensing performance. The current frequency (f), on the other hand, exhibits a good linear relationship with the airflow speed (v) described by f = 31.0v - 35.0 (R  = 0.981) as shown in Fig. 3 A. The linearity may be concluded from the equation of Strouhal number (Xu et al., 2017):where St is Strouhal number, f is the vortex-induced fluttering frequency, l is the length of the fluttering membrane, and v is the incoming airflow speed. Since St is approximately constant under the same test conditions, the frequency will be linearly correlated to the airflow speed. It is worth mentioning that changing l will affect the slope of the equation between f and v. More specifically, the smaller the length of the PTFE, the larger the slope will be.

Fig. 3

Performance of the BTE sensing module. (A) Relationship between current frequency and airflow speed (R = 0.981). (B) Current frequency and amplitude of the BTE sensor under different ambient humidity (from 35% to 81%) at an airflow speed of 13 m/s. (C) Stability of the BTE sensor at an airflow speed of 9.5 m/s. (D) Results of a comparative test about the sensing performance between the BTE sensor and a hot-wire anemometer under random airflow speeds.

Considering that human exhaled air contains large amounts of moisture, which then affects contact electrification, the electrical signals under different humidity conditions were further investigated. The humidity was controlled through a humidifier (whose outlet was connected into the acrylic box) with adjustable humidity levels, while the airflow speed was fixed at around 13 m/s. As shown in Fig. 3B, when the environmental humidity inside the acrylic box is adjusted from 35% to 81%, the current amplitude decreases significantly by 25%, from about 0.8 μA to 0.6 μA. This indicates that the electrical output is prone to deteriorate under humid environments, due to the triboelectric effect being weakened. Nevertheless, the current frequency remains almost unchanged at about 361 Hz, regardless of humidity variation, exhibiting a stability greater than 93%. The current frequency depends on the cycles of the contact-separation of the PTFE membrane, which varies little with the breathing humidity. Therefore, although the humidity reduces the amplitude of the TENG current output, the frequency remains almost unchanged for further biomonitoring. Moreover, a stability investigation of the output performance was carried out for a period of 30 days. As shown in Fig. 3C, the current signal remained at almost 264 Hz under an airflow speed of 9.5 m/s. As a result, the current frequency with a higher linearity, stability, and resistance to humidity variation is chosen as the crucial sensing parameter. Comparative tests regarding sensing performance between the BTE sensor and the hot-wire anemometer under random airflow speeds are also shown in Fig. 3D. It was found that the airflow speed obtained from the BTE sensor is consistent with that of a commercial device, indicating the reliability, accuracy, and feasibility of using the BTE sensor design as a core component for real-time breath airflow speed sensing.

Rehabilitation care after COVID-19

Being able to calculate the flow-volume curve is an important function for a spirometer to evaluate pulmonary rehabilitation progress after COVID-19 treatment. The flow (F) and the volume (L) are determined by the following equations: where S is the cross-sectional area of airflow channel, v is the airflow speed derived from the current frequency of the BTE sensor, and t is the passing time of airflow.

Following the above discussion, we developed a PPF monitor based on the BTE sensor (Fig. 4 A, Supplementary Fig. S10 and Supplementary Video S4). The basic framework of the spirometry device is illustrated in Fig. 4b. Firstly, a volunteer tester exhaled forcefully into the mouthpiece to drive the BTE sensor to generate an alternating analog signal. After the rectification of a full wave rectifier, the ADC module converts the analog signal into digital signals. Then, the MCU processes the digital signals and sends to the BLE for wireless transmission. At this stage, digital signals are a list of magnitudes of current. The detailed circuit diagrams are presented in Supplementary Fig. S11. When a mobile terminal receives the data through Bluetooth connection with the PPF monitor, it starts to process the signal through a signal processing component which converts the magnitudes of current signals into frequency information for the following operation. The flow chart of the program is specifically elaborated in Supplementary Fig. S12, including noise filtering, frequency extraction, airflow speed calculations, volume integration, and result display. Concerning that the power line interference is always inevitable during our measurement, an algorithm program combining the output amplitude and frequency is developed to exclude the interference of a noise signal. The logic of this algorithm program is filtering the noise signals through setting reasonable amplitude thresholds and operating sensing signals by contracting frequency values. The detailed process is presented in Supplementary Note S1. Using the above testing system, we performed actual PFTs on a volunteer. Firstly, we evaluated the normal status of the tester's pulmonary function as reference group. We also extracted information from the current signal (Supplementary Fig. S13a) and converted it into the breath airflow speed (Supplementary Fig. S13b), flow speed (Supplementary Fig. S13c), volume (Supplementary Fig. S13d), and then the flow-volume curve. As a result, the normal FVC value for the volunteer tester is 2.99 L, which could be directly obtained from the wireless spirometer (Fig. 4C). In addition, the FEV1 value can be derived from the integration of the flow curve with respect to time for the initial time period of 1 s. It was calculated to be 2.12 L and the FEV1/FVC ratio was found to be 71%, both of which are listed in Table 1 in detail. Afterwards, the volunteer tester conducted the second test (described as Stage 1) immediately after 15 min of vigorous exercise to simulate abnormal pulmonary function of COVID-19 patients (Custovic et al., 1994; Zavorsky et al., 2019). These results suggest that while FVC and FEV1 decline significantly, FEV1/FVC increases markedly compared to the reference group (Fig. 4D, Table 1, and Supplementary Fig. S14), which is similar to the clinical feature of the discharged COVID-19 patients (Fumagalli et al., 2020). After a 5-min break, the tester took the third test (described as Stage 2) to simulate the rehabilitation of pulmonary function for COVID-19 patients. In this test, both FVC and FEV1 improved to some extent, but still failed to return to the reference group status (Fig. 4E, Table 1, and Supplementary Fig. S15). About an hour later, the tester performed the last test to simulate recovered pulmonary function. It is evident that the indicators are almost identical to the reference group, indicating a complete recovery of pulmonary function (Fig. 4F, Table 1, and Supplementary Fig. S16). The above tests demonstrate the feasibility of using this wireless spirometer to monitor the pulmonary rehabilitation process for discharged COVID-19 patients.

Fig. 4

Demonstration of the wireless spirometer. (A) Photograph of the pulmonary function test using our wireless PPF monitor for a volunteer tester. (B) Flow diagram of the test process and the visualization of the test results. (C), (D), (E), (F) Results of PFTs at four different stages: reference group, Stage 1, Stage 2, and recovered group, respectively. The upper parts are the results of our PPF monitor and the lower parts are the comparison charts of a commercial spirometer.

Table 1

Comparative results obtained from CS (commercial spirometer) and PPFm (portable pulmonary function monitor).

Tester	FVC (L)		FEV1 (L)		FEV1/FVC (%)
	CS	PPFm	CS	PPFm	CS	PPFm
Normal	3.05	2.99	2.15	2.12	70	71
Stage 1	1.88	1.85	1.63	1.59	87	86
Stage 2	2.47	2.50	1.88	1.90	76	76
Recovered	3.01	3.02	2.17	2.18	72	72

Supplementary video related to this article can be found at https://doi.org/10.1016/j.bios.2021.113329

The following is the supplementary data related to this

article:

The effectiveness of the PPF monitor was further validated in the comparison tests, where a commercial spirometer (CONTEC SP10) was employed to record the forced expiratory flow-volume curves of each test from the same volunteer tester. Table 1 also presents the comparative results between the two spirometers. The relative error (% Err), defined as the value difference between the PPF monitor and the commercial spirometer, are all below 3% (Supplementary Table S2). These errors may have resulted from the following two aspects. First, spirometers with different sensing principles may not have the same sensitivity. Second, some minor physiological state differences between the same volunteer tester may occur for different trials. However, it is important to note that the errors are within an acceptable range (Ambastha et al., 2016), especially for the FEV1/FVC ratio, which fully demonstrates the clinical applicability of our PPF monitor for point-of-care PFTs.

It is worthwhile mentioning that spirometry is also of great significance for chronic obstructive pulmonary disease (COPD). COPD, characterized by airflow limitation or airway obstruction with respiratory symptoms such as chronic and progressive dyspnea (Choate et al., 2020; Mathers and Loncar, 2006), is becoming the fourth leading cause of death globally (Supplementary Fig. S17). Although the disease is treatable at its early stage, it is largely under-diagnosed and under-treated, even when it becomes clinically significant or enters moderately advanced stages (Choate and Mannino, 2017; Vogelmeier et al., 2017). The main criterion for diagnosing COPD is a FEV1/FVC ratio of less than 70% (Cruz, 2007), which can be effectively and accurately diagnosed through spirometer. Additionally, spirometry is recommended once COPD is identified, and long-term monitoring is required to reduce morbidity or mortality in patients (Chung et al., 2020). From the aforementioned discussion, our PPF monitor is capable of providing crucial parameters such as FEV1, FVC, and their ratio, which are sufficient for primary diagnosis and monitoring of COPD.

Conclusion

In summary, we propose a PPF monitor based on a triboelectric BTE sensor for evaluating the pulmonary function of COVID-19 patients. The BTE sensor is capable of generating substantial electrical signals without any external power supply through using the energy of airflow during respiration. It is apparent that the frequency of the current signals exemplifies a good linear relationship to airflow speed, with the sensor maintaining stable under varying humidity conditions. After data collection, the subsequent circuits can immediately convert electrical signals and transmit data wirelessly. Furthermore, a specialized program is able to analyze and display pulmonary function parameters on the mobile terminal. Based on the aforementioned principles, a PPF monitor system with signal acquisition, signal processing, and wireless transmission modules was developed to monitor human pulmonary function in real time and provide critical evaluation parameters for pulmonary rehabilitation of COVID-19 patients. We expect that the PPF monitor system will present a more convenient and low-cost approach than current commercial methods for rehabilitation care after the COVID-19.

Methods

Materials

PTFE membrane was purchased from DuPont. Aluminum foil (0.025 mm thick, 99.45%) was purchased from Alfa Aesar. PTFE nanoparticle (Kontaflon 85) was purchased from Kontakt Chemie.

Fabrication of breath-to-electrical (BTE) sensor

The framework (acrylic) of BTE sensor was cut to a size of 13 mm, 8.0 mm, and 24 mm by a laser cutting machine (GY-D460C). Aluminum foils were attached on the inner surfaces of the framework whose edge sides have been chamfered into a wedge structure. A piece of PTFE membrane was placed at the middle of the framework. The BTE sensor was installed inside a commercial mouthpiece for the following tests.

Fabrication of wireless transmission module

The alternating signal generated by BTE sensor was converted into a DC signal through a commercial bridge rectifier (DB207s). A resistance (1 MΩ) was connected in parallel to divide the rectified voltage. A control chip (ATmega328) was selected to collect the electrical signal in real time through the analog port “A5”. The BLE module (HC-05) was connected to the “RXD” and “TXD” ports of the control chip. All the above components were soldered onto a circuit board.

Experimental setup for airflow speed tests

The mouthpiece with BTE sensor was placed inside a homemade testing platform (acrylic box). Airflow was generated from an air blower (Stanley STPT600) with speed control and guided to the BTE sensor through an air hose. An outlet of a humidifier was connected into the testing platform to adjust humidity. A hot wire anemometer (Smart sensor AR866A) was adopted to monitor the airflow speed in real time. The temperature and humidity are recorded by a hygrothermograph (Mijia).

Characterizations and measurements

The surface of the PTFE membrane was characterized by scanning electron microscopy (Hitachi S5500). Both Q SC and V OC were measured by a Keithley 6514 electrometer (Tektronix). I SC was measured by a SR570 current preamplifier (Stanford Research System). The fluttering behavior of the PTFE membrane in the BTE sensor was recorded using a high-speed camera (Revealer 5F01) at a frame rate of 2000 f/s.

Experimental setup for pulmonary function tests

Volunteer testers exhaled into the wireless PPF monitor for pulmonary function tests. A commercial spirometer (Contec SP10) was employed to monitor pulmonary function parameters with the same process for comparison. The volunteer testers participating in PFTs are also co-authors of this manuscript. All participants have confirmed the details of the experiment and no ethical approval was required in this case.

CRediT authorship contribution statement

Qinghao Xu: conceived the idea, designed the experiment, Writing – original draft, Formal analysis, Data curation, analyzed the data, drew the figures, and composed the manuscript. All authors made technical comments on the manuscript. The manuscript was written through contributions of all authors. All authors have given the approval to the final version of the manuscript. The volunteers participating in the PFTs are also co-authors of this manuscript. All authors have seen the paper, agree to its content, and approve submission. Yunsheng Fang: Writing – original draft, conceived the idea, designed the experiment, Formal analysis, Data curation, analyzed the data, drew the figures, and composed the manuscript. All authors made technical comments on the manuscript. The manuscript was written through contributions of all authors. All authors have given the approval to the final version of the manuscript. The volunteers participating in the PFTs are also co-authors of this manuscript. All authors have seen the paper, agree to its content, and approve submission. Qingshen Jing: Writing – original draft, conceived the idea, designed the experiment, Formal analysis, Data curation, analyzed the data, drew the figures, and composed the manuscript. All authors made technical comments on the manuscript. The manuscript was written through contributions of all authors. All authors have given the approval to the final version of the manuscript. The volunteers participating in the PFTs are also co-authors of this manuscript. All authors have seen the paper, agree to its content, and approve submission. Ning Hu: Writing – original draft, conceived the idea, designed the experiment, Formal analysis, Data curation, analyzed the data, drew the figures, and composed the manuscript. All authors made technical comments on the manuscript. The manuscript was written through contributions of all authors. All authors have given the approval to the final version of the manuscript. The volunteers participating in the PFTs are also co-authors of this manuscript. All authors have seen the paper, agree to its content, and approve submission. Ke Lin: Writing – original draft, All authors made technical comments on the manuscript. The manuscript was written through contributions of all authors. All authors have given the approval to the final version of the manuscript. The volunteers participating in the PFTs are also co-authors of this manuscript. All authors have seen the paper, agree to its content, and approve submission. Yifan Pan: Writing – original draft, All authors made technical comments on the manuscript. The manuscript was written through contributions of all authors. All authors have given the approval to the final version of the manuscript. The volunteers participating in the PFTs are also co-authors of this manuscript. All authors have seen the paper, agree to its content, and approve submission. Lin Xu: Writing – original draft, All authors made technical comments on the manuscript. The manuscript was written through contributions of all authors. All authors have given the approval to the final version of the manuscript. The volunteers participating in the PFTs are also co-authors of this manuscript. All authors have seen the paper, agree to its content, and approve submission. Haiqi Gao: Writing – original draft, All authors made technical comments on the manuscript. The manuscript was written through contributions of all authors. All authors have given the approval to the final version of the manuscript. The volunteers participating in the PFTs are also co-authors of this manuscript. All authors have seen the paper, agree to its content, and approve submission. Ming Yuan: Writing – original draft, All authors made technical comments on the manuscript. The manuscript was written through contributions of all authors. All authors have given the approval to the final version of the manuscript. The volunteers participating in the PFTs are also co-authors of this manuscript. All authors have seen the paper, agree to its content, and approve submission. Liang Chu: Writing – original draft, All authors made technical comments on the manuscript. The manuscript was written through contributions of all authors. All authors have given the approval to the final version of the manuscript. The volunteers participating in the PFTs are also co-authors of this manuscript. All authors have seen the paper, agree to its content, and approve submission. Yanwen Ma: Writing – original draft, All authors made technical comments on the manuscript. The manuscript was written through contributions of all authors. All authors have given the approval to the final version of the manuscript. The volunteers participating in the PFTs are also co-authors of this manuscript. All authors have seen the paper, agree to its content, and approve submission. Yannan Xie: Writing – original draft, guided the whole, Project administration, conceived the idea, designed the experiment, Formal analysis, Data curation, analyzed the data, drew the figures, and composed the manuscript. All authors made technical comments on the manuscript. The manuscript was written through contributions of all authors. All authors have given the approval to the final version of the manuscript. The volunteers participating in the PFTs are also co-authors of this manuscript. All authors have seen the paper, agree to its content, and approve submission. Jun Chen: Writing – original draft, guided the whole, Project administration, conceived the idea, designed the experiment, Formal analysis, Data curation, analyzed the data, drew the figures, and composed the manuscript. All authors made technical comments on the manuscript. The manuscript was written through contributions of all authors. All authors have given the approval to the final version of the manuscript. The volunteers participating in the PFTs are also co-authors of this manuscript. submitted the manuscript and was the lead contact. All authors have seen the paper, agree to its content, and approve submission. Lianhui Wang: Writing – original draft, guided the whole .

Declaration of competing interest

The authors declare no competing interests.