Mask‐Type Acoustic Sensor Featuring a Conventional Disposable Mask Embedded with Electrospun Poly(styrene) Fiber Mats

Disposable masks are widely used, particularly in medicine to prevent the spread of infections, as they afford hygiene and convenience. Recently, sensor‐integrated masks are explored for monitoring human voice and respiration. However, masks equipped with acoustic/respiration sensors containing nano/microfiber mats cannot be used as readily as conventional disposable masks because of their high material and manufacturing costs. To address these issues, in this article, a self‐power‐generating mask‐type acoustic/respiration sensor (MTAS) incorporated with electrospun atactic poly(styrene) (aPS) microfiber‐nonwoven mats is proposed. aPS is integral to the MTAS fabrication strategy as it is inexpensive and amenable to the one‐step production of charged electrospun microfiber mats. When subjected to acoustic waves, the MTAS generates a voltage corresponding to the sound pressure and frequency owing to its ferroelectric properties. Furthermore, when a tablet is connected to the mask worn by a subject who is vocalizing, the uttered words are converted into text on the tablet. Additionally, the MTAS worn by a person can output a voltage in response to breathing. Overall, the devised mask can pioneer the development of systems that can link voice and breathing in virtual spaces, enable real‐time translation of speech, and monitor respiratory diseases.

et al. fabricated a nonwoven mat of electrospun fibers comprising poly(vinylidene fluoride) (PVDF) sandwiched between Ausputter-coated polyethylene terephthalate (PET) film electrodes, which produced a higher voltage than that of commercial PVDF films in response to acoustic waves. [15]Further, Niu et al. developed an acoustic sensor comprising a nonwoven mat of electrospun PVDF fibers with a small amount of nylon-6, which exhibited an output voltage in the frequency range of 230-800 Hz and higher output than that of the pristine-nonwoven mat; this was attributed to the piezoelectric effect of the PVDF nano/microfibers caused by dipolar charges and the triboelectric effect of the nylon-6 nano/microfibers induced by real charges. [16] mask-type acoustic/respiration sensor (MTAS) comprising mats of charged nano/microfibers has also been reported.Chen et al. fabricated a mask embedded with a charged microfiber mat of poly(propylene) (PP) and P(VDF-co-trifluoroethylene) P(VDF-TrFE) produced by melt blowing and poling output voltage in response to the human voice by harnessing the piezoelectric effect.[6] However, employing piezoelectric polymers such as P(VDF-TrFE) as disposable mask materials is challenging because of their high material costs and the additional requirement of poling.
[21][22] For example, Cheng et al. fabricated a mask comprising a poly(ether imide)-nonwoven fiber mat that was subjected to corona charging and suspended between two iron net mesh electrodes, which generated voltage in response to respiration. [19]Although many innovative strategies have been devised to adopt charged nano/microfiber mats as respirationsensing elements, challenges remain in terms of reducing the high material costs associated with the use of polymers such as PVDF for the generator and the requirement of specific manufacturing processes such as poling.These aspects deteriorate attributes of the resulting respiration sensors that are characteristic of conventional disposable masks, such as costeffectiveness and straightforward daily utilization.The high material cost is because of the use of expensive fluoropolymers and piezoelectric polymers, while the high manufacturing cost is attributable to the nano/microfiber mat production, postcharging, and assembly.Therefore, disposable masks equipped with acoustic/respiration sensors comprising nano/microfiber mats with low material and manufacturing costs are desirable.
Our group discovered that electrospun nonwoven microfiber mats of atactic poly(styrene) (aPS), which typically do not exhibit the piezoelectric effect in film form, exhibit direct/converse electromechanical properties that closely resemble the direct/ converse piezoelectric effect. [11,23]The aPS fiber mats exhibited a significantly high apparent piezoelectric constant d app (d app ≤ 2894 pC N À1 and d app > 30 000 pm V À1 , as determined via quasistatic direct and quasistatic converse electromechanical characterization, respectively). [11]The electromechanical properties evidently originate from the ferroelectret behavior of the aPS fiber mats caused simply by electrospinning. [11]Notably, the electrospun aPS fiber mats require no post-charging treatments such as poling, leading to lower manufacturing costs than those of the nano/microfiber mats requiring post-charging.Additionally, aPS is a general-purpose inexpensive polymer.Therefore, aPS fiber mats are promising as inexpensive, easy-to-use acoustic/respiration sensor elements for embedding in disposable masks as compared to those listed in Table 1.
In this study, a self-power-generating MTAS incorporated with electrospun aPS microfiber-nonwoven mats was fabricated to reduce the material and manufacturing costs.The MTAS was embedded with two layers of positively and negatively electrospun aPS microfiber mats that acted as ferroelectrets; as a result, the MTAS output voltage in response to the acoustic wave application was approximately two times higher than that from the sensor embedded with a single-layered fiber mat.The mask enables voice characterization for mask-wearing human speakers through commercially available voice recognition software, respiration monitoring, and voice-based energy harvesting.and a nonwoven PP filter, respectively.The MTAS comprised four layers (A-D; A: nonwoven PP fabric partially coated with conductive poly(3,4-ethylene dioxythiophene)/poly(styrene sulfonate) [PEDOT/PSS] and a negatively electrospun aPS fiber mat; B and D: pure PP-nonwoven fabric; C: nonwoven PP filter partially coated with PEDOT/PSS and a positively electrospun aPS fiber mat).The PEDOT/PSS-coated areas in Layers A and C acted as electrodes.PEDOT/PSS was used in the initial study as a readily usable material in the conductive electrode.To ensure safety during the use of the MTAS, it was reported that the inhalation of a small amount of aPS is harmless for humans, [24] and PEDOT/ PSS is a biocompatible material. [25]To further ensure safety, the PP-nonwoven filter in Layer C prevents the aPS fibers from being inhaled by persons wearing the MTAS; Layer D prevents the person's face from coming in direct contact with PEDOT/PSS.Nevertheless, as ensuring the person's safety is crucial, strict safety testing is necessary in the future.Details concerning the structure and fabrication are provided in Section 4.

Geometrical and Electrical Properties
Scanning electron microscopy (SEM) images of the positively and negatively electrospun aPS fiber mats (Figure 1d, e, respectively) revealed the deposition of uniform microfibers without beaded inhomogeneous structures; the average diameters of these fibers were 3.55 AE 0.21 and 4.18 AE 0.27 μm, respectively (all errors reported herein represent standard deviation.)Additionally, each fiber mat comprised randomly oriented microfibers, like those in filters, and it had a highly sparse structure that could improve breathability.Thus, the fabricated fiber mats were suitable for being embedded into the designed mask-type sensor.
The surface potential was measured for each fiber mat deposited on electrically grounded PEDOT/PSS-coated fabric/filter electrodes; thus, the estimated average surface potentials of the positively and negatively electrospun fiber mats were 235.0 AE 2.4 and À186.2AE 15.9 V, respectively.This confirmed that both electrospun aPS fiber mats were charged, with the positively and negatively electrospun fiber mats primarily storing positive and negative charges, respectively.

Quantitative Characterization of Output Charges
Electrospun aPS fiber mats evidently exhibit ferroelectret behavior. [11]Therefore, when the counter electrode approaches the surface of the fiber mats, charges are induced in the approaching electrode.The amount of induced charge varies with the distance between the electrode and fiber mat surface.When the counter electrode and the electrode under the fiber mat are connected through a load, the change in the amount of generated charges flowing through the load is detected as an output current.Thus, the relationship of the output charge from the fiber mat with the gap between the fiber mat and electrode must be characterized to determine the output charge properties of the MTAS.The relationship of the output charge from the electrode with the gap between the counter electrode and fiber mat was ascertained using the sequential approaching-loading technique (Figure 2). [26]The aPS fiber mats deposited on the nonwoven fabric and the filter were fixed onto a glass substrate for the measurements.
An 8.0 mm diameter disk probe, which was assumed to act as the MTAS electrode, was initially placed on the surface of the aPS fiber mat.The disk probe was moved away from the surface, and its position, h, and the amount of output charge flowing from it were measured.The initial probe position, h = 0 mm, was defined as the location where the probe applied a pressure of %2.0 kPa to the surface of the fiber mat (Figure 2a).The disk probe was then lifted until the pressure reached 0 Pa (Figure 2b) and elevated further to a height of 5 mm away from the fiber mat (Figure 2c; see photograph of the apparatus in Figure 2d).The amount of charge generated when the disk probe was moved from h = 0 mm to h = 5 mm was monitored (Figure 2e).The charge output from each aPS fiber mat in the absence of an external power source is similar to that of electrospun aPS fiber mats deposited on an indium tin oxide (ITO) substrate. [27]The positively and negatively electrospun fiber mats generated negative and positive charges, respectively; the underlying mechanism is explained theoretically in Section 2.4.The output charge increased significantly up to h = 0.1 mm and escalated gradually thereafter for both the positively and negatively electrospun fiber mats.This indicated that even small changes in the fiber mat-electrode gap, caused by phenomena such as acoustic-wave-induced vibrations, could generate voltage.In contrast, the fiber-mat-free nonwoven fabric did not produce any charge (Figure 2e), indicating that the charge output was due to the electrospun fiber mat.

Theoretical Model for Output Charge
In our previous study, the positively electrospun aPS fiber mat was found to hold real charges as a ferroelectret, with positive and negative real charges typically held in the upper and lower parts of the fiber mat, respectively. [11]Additionally, the fiber mat deposited on a material that acted as the bottom electrode was found to output charge when the upper electrode approached the surface of the electrode and indented the fiber mat. [28]To explain the charge output theoretically, the aPS fiber mat was assumed to be a ferroelectret with an infinitely wide rectangular structure and thickness T F0þ , whose upper and lower surfaces were charged with effective surface charge densities of identical amplitudes (þσ EFF0þ and Àσ EFF0þ , respectively) [28] (Figure 3a; h: the position of the upper electrode; h = 0: the upper electrode position where the electrode applied a constant pressure onto the fiber mat, e.g., the pressure in the experiment described in Section 2.3 [%2.0 kPa]).The upper electrode was then raised to h 0 , that is, the electrode position where the pressure applied to the electrode was 0 Pa and equal to the surface of the fiber mat (Figure 3b).Using this ferroelectret model, the amount of charge generated by the positively electrospun fiber mat for different Figure 3. Schematics depicting a theoretical model of charge generation from fiber mats a-d) positively and e-h) negatively electrospun onto the PEDOT/ PSS-coated nonwoven mat: a,e) the upper electrode pressing on the fiber mat; b,f ) the upper electrode being lifted to a height corresponding to the fiber mat surface; c,g) increasing the distance between the fiber mat and upper electrode; d,h) creating a sufficiently large distance between the fiber and electrode, leading to no charge generation on both electrodes.i-l) Schematics illustrating a theoretical model of charge generation from the positively and negatively electrospun fiber mats with an intermediate poly(propylene) (PP)-nonwoven layer: i) the upper electrode pressing the fiber mats, j) the pressurization-free scenario, k) increasing the distance between the fiber mat and upper electrode, and l) creating a sufficiently large distance between the fiber mat and electrode.
where S is the area of the electrode and ε F0þ is the permittivity of the positively charged fiber mat.Equation (1) describes ΔQ(h) for the electrode position changing from that in Figure 3a to that in Figure 3b, which causes ΔQ(h) to increase significantly in the negative direction.Equation (2) expresses ΔQ(h) for the electrode moving away from the surface of the fiber mat.When h is small, that is, when the electrode is close to the surface of the fiber mat, the change in ΔQ(h) is large.As h increases, that is, when the electrode moves farther away from the surface of the fiber mat (Figure 3c), the change in ΔQ(h) decreases.Furthermore, when h becomes significantly large (Figure 3d), ΔQ(h) converges to Àσ EFF0þ S. Compared with the experimental data (Figure 2e), the amount of charge generated by the positively electrospun fiber mat shows a large, negative increase for h values close to 0 mm and a large change up to %0.1 mm.The change in ΔQ(h) diminishes with increasing h, with almost no change observed at h = 5 mm.Thus, the theoretical model indicates the amount of charge generated by the fiber mat.
The negatively and positively electrospun fiber mats were considered to be similar (Figure 3e-h).In the theoretical model for the negative fiber mat, the terms σ EFF0þ , T F0þ , h 0þ , and ε F0þ in Equation ( 1) and (2) were replaced with σ EFF0À , T F0À , h 0À , and ε F0À .Because the effective surface charge density of the negatively electrospun fiber mat was negative, the output charge was positive.Moreover, the amount of output charge ΔQ(h) increased with increasing h, similar to that of the positively electrospun fiber mat in the positive direction.
Models of the positively and negatively electrospun fiber mats were then assessed (Figure 3i-l).The positive and negative fiber mat models contained the positively and negatively electrospun fiber mat surfaces facing each other.The positive charge on the surface of the lower positively electrospun fiber mat induced a negative charge on the electrode of the upper negatively electrospun fiber mat and vice versa.As the upper and lower fiber mats induced charges on the other electrode, more charges could be generated on the upper and lower electrodes than only on the positively or negatively electrospun fiber mats.Consequently, the charge induced at the upper and lower electrodes flowed to the load when the distance between the positively and negatively electrospun fiber mats was large (Figure 3j,k).Hence, a voltage higher than that of only the positively or negatively electrospun fiber mat was generated between the upper and lower electrodes.When the distance between the fiber mats was significantly large, no charge was generated by the electrodes; this is similar to the case of the positively or negatively electrospun fiber mats (Figure 3l).

Output Voltage from Differently Structured MTASs
The acoustic properties of the fabricated MTAS were evaluated by attaching it to the face of a mannequin (Figure 4a-c), which was equipped with a piezo buzzer in the mouth to ensure the emission of an acoustic wave with a controlled frequency and sound pressure level from the mouth.The fabricated MTAS was attached to the mannequin such that it covered the nose and the chin, similar to a mask worn by a person, and the voltage output from the MTAS was directly measured using an oscilloscope, without any amplification circuit or other devices.
First, the output voltage from the MTAS comprising the positively and negatively electrospun fiber mats was evaluated for acoustic waves with frequencies of 120 and 240 Hz and a sound pressure level of 110 dB (Figure 4d,e).Furthermore, the average peak output voltage was calculated from 100 datapoints of absolute values of the peak output voltages (Figure 4f ).Notably, the frequencies of 120 and 240 Hz were selected because they were representative of the Japanese male/female voice. [29]To determine the background noise voltage, the output voltage from the MTAS was measured in the absence of the acoustic wave (Figure 4d,e).The Fourier transform of the background noise waveform indicated that the peak frequency of the background noise was 60 Hz (Figure S3, Supporting Information); this can be attributed to the noise derived from commercial power supply because 60 Hz is the commercial power frequency in western Japan.The MTAS generated a significantly higher voltage than the background noise voltage upon being exposed to the acoustic waves (Figure 4d-f ), demonstrating its self-power-generative acoustic-sensing capability.In the MTAS, the applied acoustic wave causes the fiber mat and/or the inner and outer electrodes to vibrate, altering the distance between each fiber mat and the electrodes, thereby modifying the amount of electrostatically induced charges; consequently, the charge output from the electrodes flows to the oscilloscope probe and is detected as a voltage.
To verify the effect of employing two (positively/negatively) electrospun fiber mat layers, MTASs with only the positively or negatively electrospun fiber mats were also evaluated (Figure 4d,e).The MTAS containing only the positively electrospun fiber mat comprised Layer C with the positively electrospun fiber mat and Layer A without any electrospun fiber mat (only the PEDOT/PSS coating), whereas that with only the negatively electrospun fiber mat comprised Layer A with the negatively electrospun fiber mat and Layer C without any electrospun fiber mat (only the PEDOT/PSS coating).The MTASs with only the positively or negatively electrospun fiber mats generated a voltage that was approximately half that of the MTAS with two positively and negatively electrospun fiber mats (Figure 4d-f ).This doubling of the output voltage generated by the MTAS with two electrospun fiber mat layers is consistent with the explanation provided in Section 2.4.
Triboelectrification could have occurred in the fabricated MTAS.Thus, to verify the presence of triboelectrificationinduced output voltage, two types of MTASs were examined under acoustic wave application: an MTAS without any electrospun fiber mats, that is, both Layer A and Layer C with no electrospun fiber mats (only the PEDOT/PSS coating), and an MTAS with positively/negatively electrospun fiber mats discharged by isopropyl alcohol spraying (Figure 4d-f ).The results indicated that the output voltage from the two MTASs was nearly identical to the background noise voltage.Therefore, the output voltage originating from triboelectrification in the fabricated MTAS under acoustic wave application was negligible.Essentially, the output voltage from the MTAS emanated from the positively and negatively charged electrospun fiber mats.

Acoustic Properties of MTAS
The relationship between the applied sound pressure and output voltage was determined for the MTAS with the positively/negatively electrospun fiber mats by using the setup shown in Figure 4a-c.The sound pressure was varied from 0.3 to 11.6 Pa at 120 Hz and from 0.3 to 15.9 Pa at 240 Hz (Figure 5a; the relationship between sound pressure level and output voltage is shown in Figure S4, Supporting Information).The output voltage from the MTAS increased with increasing sound pressure.
The average noise level voltage was 7 mV, and voltages exceeding the noise level were output from the mask at both frequencies when the sound pressure exceeded 1.1 Pa.Furthermore, higher voltages were output when the MTAS was subjected to a 120 Hz acoustic wave compared with those generated by the MTSA exposed to a 240 Hz acoustic wave.The output voltage increased almost proportionally with sound pressure at both investigated frequencies, validating the sound-pressure-sensing ability of the devised MTAS.
A first-order linear approximation (represented by the dotted lines in Figure 5a) indicated that the voltage generation by the MTAS was initiated at sound pressures of 0.383 and 0.546 Pa at frequencies of 120 and 240 Hz, respectively.The slope varied with the frequency, with a lower frequency resulting in a larger voltage output.
The sensitivity of the MTAS, S, was determined as follows: [30] S ¼ where V and P are the output voltage from the MTAS and the sound pressure, respectively.By using the data in Figure 5a, the sensitivity values were calculated to be 6.40 and 4.82 mV Pa À1 at 120 and 240 Hz, respectively; these values are notably lower than those of numerous previously reported materials (Table 1).This was because of the difference in the resistance of the oscilloscope probe used to measure the output voltage.The oscilloscope probe resistance was determined to be 10 MΩ by using a generic resistance oscilloscope.The output voltage corresponding to the current was low because a highimpedance probe (>100 MΩ) was not employed.Moreover, the amplitude of the waveform was lower than that measured using a high-impedance probe owing to the time constraint of the probe impedance.The shape of the mask also possibly contributed to the low sensitivity.Because the fabricated MTAS had a pleated shape, the gap between the electrode and fiber mat increased and the electrostatic induction of the electrode decreased when the MTAS was attached to the mannequin, thereby decreasing the voltage output.Nevertheless, the person's speech detected by the MTAS was recognized by the commercial speech recognition software Siri on an iPad as shown later in this section, demonstrating that the developed MTAS exhibited sufficient sensitivity.Subsequently, the output voltage from the MTAS was measured under the application of acoustic waves of different frequencies (40-1000 Hz; Figure 5b).For two fixed sound pressures (3.6 and 6.3 Pa), the output voltage from the MTAS decreased with increasing frequency.The output from the mask subjected to an acoustic wave depended on the change in the fiber mat-electrode distance owing to the vibration of the fiber mat and/or the electrode; however, the fiber mat could not move quickly owing to its light weight and flexibility.Therefore, the mask did not respond to vibrations caused by high-frequency sound waves, and the vibration of the fiber mat diminished, resulting in an inability to output a high voltage at high frequencies.Notably, the voltage output increased slightly at 600 Hz, similar to resonance, which was presumably due to resonance between the fiber mat and the electrode vibration.In particular, voltage was generated at %120 and %240 Hz-the frequencies of the average Japanese voice.Moreover, voltage was generated by the MTAS up to a frequency of 1000 Hz, which is the limit of the frequency range for Japanese speech.Overall, these results indicate that the mask functions effectively as an acoustic sensor for Japanese speech.
The output voltage from the MTAS was subsequently measured while a 39-year-old male subject wore it (Figure 6a,b).When the person vocalized "Ah", an output voltage was detected in response (Movie S1, Supporting Information).The sound pressure level of the voice was %110 dB, and the average peak voltage was %0.234 V.The output voltage produced by the human voice was approximately six times higher than that generated using the piezo buzzer at a frequency of 120 Hz and a sound pressure level of 110 dB.This increased output voltage was likely due to the larger irradiation area of sound waves from the mouth of the person than that of the mannequin; this is because the sound waves produced by the person were not as directional as those of the piezo buzzer.Thus, human voice detection with the fabricated MTAS evidently exhibited significantly superior output characteristics than those of the piezo buzzer.This indicated that the shape of the MTAS and deposition area of the fiber mat were suitable for monitoring human voice.
The Fourier-transformed frequency spectrum of the Figure 6b data (Figure 6c) revealed a main frequency of %149 Hz and overtones two to six times the main frequency.Thus, the devised MTAS enabled the simultaneous detection of not only a specific frequency but also multiple frequencies.When the person vocalized the Japanese phrase "Kyoto Kogei Sen'i Daigaku" (Kyoto Institute of Technology in Japanese), the waveform of the output voltage changed according to the words vocalized (Figure 6d).
The MTAS was connected to a tablet via a phone cable/plug (Figure 7a).No amplification circuits or devices were included in the MTAS or cable/plug, and a load of 1.0 kΩ was introduced to help the tablet recognize the cable/plug connection.Voice recognition software on the tablet was used to characterize the output voltage from the MTAS while the person vocalized.When the MTAS-wearing 39-year-old male subject vocalized "Kyoto Kogei Sen'i Daigaku", the software displayed the same in Japanese (Figure 7b and Movie S2, Supporting Information).This demonstrated that the human-voice-induced voltage output from the devised MTAS could be recognized by commercial voice recognition software.We checked the voice recognition software when the person removed the MTAS and vocalized, thereby confirming that the tablet's internal microphone did not work during the aforedescribed experiment.In other words, the tablet characterized the words through the MTAS.
To evaluate the reproducibility of the output voltage from the developed MTAS, the output voltage was measured when the MTAS was continuously subjected to the acoustic waves (frequency of 120 Hz and sound pressure level of 110 dB) generated by the piezo buzzer for 12 h (Figure 8).The initial average peak output voltage from the MTAS was 22.7 mV; the average peak output voltage after acoustic wave subjection was 23.7 mV.The output voltage demonstrates no significant decrease even after the 12 h subjection.Therefore, the MTAS shows stable output reproducibility for continuous use for at least 12 h.

Respiration Monitoring
Analysis of the output voltage from the MTAS, when the MTASwearing 39-year-old male subject breathed (Figure 9 and Movie S3, Supporting Information), indicated that the MTAS detected both normal and fast/intense breathing.The average peak voltages, which were the averaged absolute values of the maximum and minimum values per cycle for five instances of normal breathing and 15 instances of fast/intense breathing, were 0.214 and 1.290 V, respectively.A higher voltage than that of the voice was generated because the breathing deformed the shape of the entire MTAS.Additionally, the output voltage for intense/fast breathing was approximately six times higher than that for normal breathing.The distance between the inner and outer electrodes varied with the breathing pressure, leading to a higher voltage for intense breathing.
For exhalation, a positive voltage was output because of the reduced distance between the inner and outer layers (C and A, respectively), similar to the positive charges that were output during the transition from the state shown in Figure 3b to that in Figure 3a.For inhalation, a negative voltage was output as the distance between the inner and outer layers increased, similar to the negative charges that were output during the transition from the state in Figure 3b to that in Figure 3c.Moreover, the amplitude of the positive voltage exceeded that of the negative voltage, and the pressure and flow rate during exhalation were higher than those during inhalation. [31]Therefore, the MTAS could distinguish exhalation and inhalation.
The frequencies for normal and fast/intense breathing were 33 and 156 breaths per minute, respectively.Therefore, the MTAS was able to detect the intensity and speed of breathing, thereby highlighting its potential to monitor respiratory diseases such as sleep apnea [32] and chronic obstructive pulmonary disease. [33]

Energy Harvesting
The energy-harvesting capability of the MTAS was investigated while the MTAS-wearing 39-year-old male subject vocalized.The output voltage from the MTAS during vocalization was an alternating voltage (Figure 6b); thus, it was converted to a direct voltage using the full-wave rectifier circuit shown in Figure 10a.The output voltage from the MTAS was converted to a direct voltage when the person vocalized "Ah" with a sound pressure level and duration of approximately 110 dB and 6 s, respectively (Figure 10b).The direct output voltage increased when the person started vocalizing and then saturated after its rise delay owing to the charging period of the 100 nF capacitor.The voltage decreased after the vocalization was terminated.The output current was calculated to be 10 nA by dividing the voltage output from the MTAS during vocalization (%0.1 V) by the resistance of the oscilloscope probe (10 MΩ).Consequently, 1.0 nW of energy was estimated to be harvested by the mask by multiplying the voltage and current.Additionally, the output power density was calculated to be 160 nW m À2 on the basis of the deposit area of the 90 mm diameter aPS mats.These results highlight the energy-harvesting capability of the fabricated MTAS.

Charge Retention Properties of the Fiber Mats
To investigate the charge retention properties of the positively and negatively electrospun aPS fiber mats, the surface potential of each fiber mat was measured over time.The fiber mats were stored in two different atmospheres: ambient conditions (see the humidity-time and temperature-time plots shown in Figure S5, Supporting Information) and high-humidity conditions (relative humidity, %90%; temperature, %30 °C).The high-humidity conditions corresponded to the human subject wearing the MTAS.The surface potentials of the fiber mats were measured for 29 d.
Defining the day after 2 d of post-electrospinning drying as Day 0, the average surface potentials of the positively and negatively electrospun fiber mats at Day 0 were 203.5 AE 32.1 and À171.2AE 40.5 V, respectively.Normalized surface potential-time plots were constructed for the fiber mats (Figure 11), with the surface potentials on Day 0 considered as 1 and À1 for the positively and negatively electrospun fiber mats, respectively.
The normalized surface potentials of the mats stored under atmospheric conditions deceased gradually.However, those under the high-humidity conditions decreased significantly up to Day 7 and then diminished gradually.Under the highhumidity conditions, charges trapped at a shallow level among the charges in the fiber mats tended to transition to an intermediate level of water molecules and leak through them, leading to a rapid, significant decrease in potential.Subsequently, the charges captured in the deep trap were retained, thereby gradually decreasing the surface potential. [34]However, because the humidity was low under ambient conditions, the chargetrapping-related shallow trap leakage through the water molecules was diminished, and the surface potential decreased gradually.
Disposable masks are typically used for only 1 d; therefore, the normalized surface potential of the positively and negatively electrospun fiber mats stored under high-humidity conditions for 1 d was targeted for analysis.The obtained values remained greater than 80%, indicating the possibility of achieving sufficient performance even in daily use.Additionally, the surface potentials were maintained at 50% even after 29 d under atmospheric conditions, emphasizing the ability of the mats to  perform even after approximately one month.Therefore, the aPS microfiber mat, which is an inexpensive material that can be manufactured solely by electrospinning without any post-charging processes such as poling, is suitable as the sensor material for disposable masks.

Conclusion
A self-power-generating MTAS embedded with two-layered ferroelectrets of positively/negatively electrospun microfiber mats of an inexpensive polymer (aPS) was fabricated without any post-charging process, including poling.This MTAS also consists of the conventional disposable mask used as the substrate for embedding the fiber mat, resulting in satisfactory filtering performances required for a face mask without requiring other materials and direct usability as a mask.The inexpensive polymer, manufacturing process, and structure contribute to achieving this inexpensive disposable mask-type sensor.
The MTAS generated voltage in response to acoustic waves without an external power supply.In particular, the MTAS, which contained two-layered positively/negatively electrospun fiber mats, generated a voltage that was approximately two times higher than that of either positively or negatively electrospun fiber mats.The MTAS output voltage was nearly proportional to the sound pressure, and acoustic waves were detected at frequencies ranging from 40 to 1000 Hz.When an MTAS-wearing human subject vocalized, the MTAS generated a voltage in response.Furthermore, when the MTAS was connected to a commercially available tablet without any amplification circuits or devices, the tablet software recognized the spoken words.Additionally, the MTAS could output voltage in response to the breathing of the MTAS wearer, for example, distinguishing between normal and intense breathing.These results highlight the viability of the MTAS as an acoustic and breathing sensor that can achieve the low costs required for disposable masks.
More than 50% and 80% of the initial surface potential of the fiber mats were retained upon storage under room-temperature and high-humidity conditions for 29 and 1 d, respectively.This performance retention underscores the feasibility of storing and utilizing the MTAS.Therefore, the MTAS can facilitate improvements in communication through voice characterization, especially for the hearing impaired; establish voice-breathing links in virtual spaces; enable real-time recording of meeting minutes; permit real-time translation of speech; and allow the monitoring of respiratory diseases.

Experimental Section
Framework of MTAS: A widely and commonly used disposable threelayered nonwoven mask was used as the framework for the designed MTAS.The mask was disassembled into three sheets-two PP-nonwoven fabrics and one PP-nonwoven filter-and used as the MTAS framework.The nonwoven fabrics were used as Layers A, B, and D, whereas the nonwoven filter was used as Layer C. A 0.5%-1% PEDOT/PSS aqueous solution (high-conductivity grade, Merck KGaA) containing 0.5 vol% ethanol (≥99.5%,Nacalai Tesque) relative to the PEDOT/PSS solution was drop-cast at the centers of Layers A and C over a 90 mm diameter area.Ethanol was added to improve wettability by reducing the surface tension of the solution.After drop-casting, Layers A and C were dried under vacuum for 2 d.SEM images of the nonwoven PP fabric and nonwoven PP filter coated with PEDOT/PSS were acquired (Figure S2a,b, Supporting Information).To measure the voltage generated at the external PEDOT/PSS-coated electrodes, conductive yarns were sewn at Layers A and C from the center of the PEDOT/PPS electrode to the outside of each layer and used as electrical cables.
Materials and Fabrication of Electrospun Fiber Mats: aPS (M w % 280 000; Merck KGaA) was dissolved in N,N-dimethylformamide (DMF; Nacalai Tesque) to a concentration of 30 wt% at room temperature.Then, the solution was loaded into a glass syringe equipped with a 23-G stainless-steel needle and subsequently ejected at a rate of 1.4 mL h À1 with a syringe pump.The nonwoven PP fabric or filter partially coated with PEDOT/PSS was anchored onto an electrically grounded aluminum plate (300 Â 200 Â 1 mm) and then placed 20 cm away from the tip of the needle.Subsequently, positively or negatively electrospun aPS fiber mats were produced by applying þ9.0 or À9.0 kV to the needle, respectively.The electrospinning period for each fiber mat was fixed at 7 min.The temperature and humidity during the electrospinning were 32.4-35.4°C and 24.4%-28.4%,respectively.After electrospinning, the fiber mat deposited outside the PEDOT/PSS area was removed using a knife.
Fabrication of the Mask-Type Sensor: Immediately after electrospinning, the mask-type sensor was assembled from Layers A to D, with the fiber mats on Layers A and C facing each other.For the initial study, Layers A-D, ear pads, nose pads of the upper part, and the lower part were integrated using a stapler.Although it was reported that the as-electrospun aPS fiber mats hardly contained residual DMF, [11] the assembled mask sensor was nevertheless dried to evaporate DMF under vacuum for 2 d as a precautionary measure to ensure safety.The signal outputs from the mask-type sensor were positive and negative for the Layer-A and Layer-C cables, respectively.
Morphological Characterization: SEM images were obtained using a TM4000PlusII device (Hitachi High-Tech, Japan) in backscattered electron detection mode with an acceleration voltage of 10 kV.A 4 nm thick gold layer was coated onto each fiber mat prior to obtaining the SEM images.The average fiber diameter was calculated from 100 diameters measured by SEM.
Surface Potential Measurements: The surface potential of each fiber mat was measured using a digital low-voltage static meter (KSD-3000, Kasuga Denki, Japan) over an area of 20 Â 20 mm with a 10 mm gap between the probe and the surface of each fiber mat.For the surface potential measurements, each fiber mat produced on the nonwoven fabric/filter was flattened/cut and fixed onto a %0.7 mm thick glass substrate (30 Â 30 mm in area) coated with %150 nm thick ITO to reduce the bending-induced inhomogeneity.During the measurements, the PEDOT/PSS bottom electrode of each fiber mat was electrically grounded.The average surface potential of each fiber mat was calculated using five measured surface potentials.Subsequently, three positively electrospun and three negatively electrospun fiber mats were stored under room-temperature conditions in a closed box in an office and under high-humidity conditions in a temperature and humidity controller (STC-V, SANPLATEC, Japan); the surface potential of each fiber mat was measured for up to 29 d.Temperature and humidity under ambient conditions were measured using a thermo recorder (TR-72wb, T&D, Japan), and those under the high-temperature, high-humidity conditions were controlled at 30 °C and 90% using the aforementioned controller.
Output Charge Amount Characterization: The h-dependent output charge amount was assessed using the sequential approaching-loading technique [26] and the setup shown in Figure 2a (LRFB-01, Lead Techno, Japan).An 8.0 mm diameter metallic disk probe was placed in contact with the surface of each fiber mat at a pressure of %2.0 kPa (%0.10 N load) and regarded as the initial position (h = 0 mm).Then, the probe was moved vertically up and down with respect to the surface of the fiber mat at a regulated speed of 4.5 mm s À1 and up to the position h = 5.0 mm.The speed and position were controlled using an automatic Z-stage, and the load applied to the fiber mat was measured using a load cell.When the probe was in motion, its position (h) and charge output were simultaneously measured using a laser displacement meter (LK-G152/LK-G3000V, Keyence, Japan) and charge amplifier, respectively.

Figure 2 .
Figure 2. Schematics explaining the sequential approaching-loading technique, in which a) the upper disk probe pressed the fiber mat at %2.0 kPa in the initial state (h = 0 mm), b) the upper disk probe position was identical to the fiber mat thickness, and c) the upper disk probe was moved up from the fiber mat.d) Photograph of the setup used for the analysis.e) Output charge amount from the fiber mats positively and negatively electrospun onto a PEDOT/ PSS-coated nonwoven mat measured for varying h (with flattened Layers C and A).

Figure 4 .
Figure 4. a,b) Front-and c) side-view photographs of a mannequin a) without and b,c) with the MTAS.Output voltage from various types of MTASs subjected to a buzzer sound wave with frequencies/sound pressure levels of d) 120 Hz/110 dB and e) 240 Hz/110 dB.(ES: electrospun).f ) Average peak output voltage from various types of MTASs; the gray area represents the averaged background noise voltage and its standard deviation range.

Figure 5 .
Figure 5. Relationship between the peak output voltage from the MTAS and the a) sound pressure and b) frequency.The black line and gray area represent the averaged background noise voltage and its standard deviation range, respectively.

Figure 6 .
Figure 6.a) Photograph of the output voltage measurement system for the MTAS when an MTAS-wearing subject vocalizes "Ah".b) The resulting output voltage-time plot and c) its Fourier-transformed frequency spectrum.d) Output voltage-time plot when the subject vocalized "Kyoto Kogei Sen'I Daigaku".

Figure 7 .Figure 8 .Figure 9 .
Figure 7. a) Circuit connecting the MTAS to a tablet, with a load resistance of 1.0 kΩ connected in parallel to the MTAS.b) Photograph of the tablet displaying a phrase vocalized by a human subject.

Figure 10 .
Figure 10.a) Full-wave rectifier circuit featuring a 100 nF capacitor connected in parallel to the output for converting the alternative voltage from the MTAS into a direct voltage to determine the energy harvested.b) Direct voltage derived from the MTAS when the mask-wearing person vocalized.

Figure 11 .
Figure 11.Time-dependent normalized surface potentials of the positively/negatively electrospun aPS fiber mats.The fiber mats were stored under a) low-humidity and room-temperature and b) high-humidity (90%) and room-temperature conditions (30 °C).

Table 1 .
Representative previously reported acoustic/respiration sensors comprising nonwoven mats of charged nano/microfibers.