Implementation of a Parametric Ultrasonic Receiver Using Multilayer Lead Zirconate Titanate for a Feasibility Study of an Ultrasonic-Beam-Focused Hearing Aid

Abstract

We demonstrated that focusing an ultrasonic beam on the eardrum can overcome the high-frequency sensitivity limitations and acoustic distortion of conventional hearing aid receivers. Multilayer PZT was used for an ultrasonic receiver that operates at low voltage and enters the external auditory canal, and a 3 mm radius radiator was designed to radiate the focused parametric acoustic signal to the center of the eardrum based on an acoustic analysis according to the frequency. To this end, an ultrasonic earphone consisting of a radiator attached to multilayer PZT and a 130 kHz parametric ultrasonic modulator was implemented; vibration and sound pressure were measured using a laser vibrometer and a tube-type microphone. The proposed parametric ultrasonic receiver generates an average sound pressure of 70 dB SPL at a frequency of 1~10 kHz with a 10 V_peak applied voltage; this was implemented to provide a higher output in the range of 5 kHz and above, which is difficult to cover with existing receivers.

1. Introduction

Traditional hearing aids, such as behind-the-ear (BTE), in-the-ear (ITE), and completely implantable (CIC) hearing aids, amplify sound and transmit it to the eardrum via a hearing aid receiver. Digital hearing aids can compensate for hearing loss based on individual hearing data, but these still encounter problems. First, it is difficult to compensate for hearing loss due to the frequency characteristic limitations of the receiver. Patients with sensorineural hearing loss have significant hearing loss in the frequency range above 3 kHz, so a receiver with excellent high-frequency characteristics is required [1,2,3,4]. However, it is difficult for conventional receivers driven by small electromagnets to achieve sufficient sound pressure in the high-frequency range because they utilize the resonance characteristics of the vibration system [5,6,7]. Second, the sound from the receiver reaches the eardrum at different times due to reflections and resonances in the ear canal, resulting in loss of speech recognition. Third, due to ventilation problems, if the ventilation hole is enlarged, then the feeling of ear blockage is reduced, but the sound pressure in the ear canal decreases and sound leaking from the ear canal may cause acoustic feedback [8].

Puria et al. devised a light-driven contact hearing aid that can be installed in the ear canal without surgery and can provide the same effect as a middle-ear implant, with a uniform sound pressure of 110 dB SPL (sound pressure level) in the range 100 Hz to 10 kHz [9,10]. This method modulates sound into light and sends it to a solar cell and signal processing module installed near the tympanic membrane; then, a micro actuator in contact with the tympanic membrane vibrates to transmit acoustic vibration. However, the solar cell and signal processing unit requires surgery.

Direct transmission of sound to the center of the tympanic membrane using ultrasound is advantageous because it does not require a solar cell or a micro actuator. Recently, Seok et al. reported the results of an experiment using a parametric ultrasound array that can transmit sound of approximately 68 dB SPL to a virtual tympanic membrane at a distance of 2.5 cm. This compact ultrasonic device consists of 12 planar PZT (lead zirconate titanate) actuators and is designed to fit into the ear canal. However, the authors did not analyze the frequency characteristics to obtain the acoustic gain at high frequencies [11].

In this paper, we propose a single-radiator ultrasonic hearing aid receiver (serving as the output emitter) that increases power efficiency by increasing the ultrasonic radiation area and lowering the radiation frequency in the ear canal. We implemented an ultrasonic receiver using multilayer PZT devices for low-voltage driving and measured the directivity and frequency response by driving a SRAM (square root amplitude modulation) parametric ultrasonic modulator. The results revealed the potential benefits of an ultrasonic hearing aid.

2. Materials and Methods

2.1. Implementation of Ultrasonic Receiver

The human ear canal is 7 to 9 mm in diameter and the receiver for an ultrasonic hearing aid must fit inside it [12]. We used multilayer PZT vibrator (5 mm length, 2 × 3 mm cross-section; AE0203D04DF, KEMET, New Taipei City, Taiwan) as the core of the transducer. A 3 mm thick brass backing (acoustic impedance ~37 Mrayl, similar to PZT’s 34.5 Mrayl) was attached to one side using a thin epoxy adhesive (impedance~2.5 Mrayl, thickness < 2 μm). The adhesive used is an epoxy resin with low acoustic impedance (z = 2.51 Mrayl) and a thickness of less than 2 μm. A silicone cap for the receiver with an outer diameter of 9.5 mm was attached to the backing material body with the radiator side as the inlet. When the silicone cap is inserted, the thin sleeve rubber of the cap bends, making it possible to insert it into an ear canal with a diameter of 8 mm. When inserting the receiver, a plastic guide ring with an inner diameter of 6.2 mm was inserted to prevent the vibrating radiator from coming into contact with the silicon cap. The left side of Figure 1c shows an ultrasonic hearing aid receiver, and the right side shows a BA (balanced armature) receiver (FK-23451, Knowles, Itasca, IL, USA) for comparison, placed in a silicon cap of the same size.

When the distance between the receiver and the eardrum is 6 mm and the diameter of the eardrum is 8 mm, the divergence angle for the sound from the receiver to reach the 4 mm diameter area from the center of the eardrum is approximately 36°. When the diameter of the sound radiator is 5 mm, in order to maintain a divergence angle of 36°, a carrier frequency of approximately 130 kHz is required according to Equation (1).

$$\theta =2{\sin }^{-1}{\displaystyle \frac{0.61\lambda }{D}}$$

(1)

where θ is the divergent angle in radians, λ is the wavelength of the sound, and D is the diameter of the radiator.

2.2. Parametric Audio Modulation Method for the Hearing Aid

In the parametric audio system, the audio is generated because of the interference of the carrier frequency and sideband signals frequency due to the nonlinearity of the air when the strong ultrasonic wave modulated by the audio signal is radiated from the transducer. To generate a focused beam in the hearing aid receiver, it is necessary to select an ultrasonic modulation method suitable for this. Parametric audio is a secondary sound generated by air nonlinearity when a strong ultrasonic beam travels, with the characteristics of a sharp audio beam. So far, several parametric audio-modulation schemes have been developed for directional loudspeakers [13,14]. Since the DSBAM (double-sided band-amplitude modulation) method was first implemented by Yoneyama et al. [15], SRAM (square root amplitude modulation), SSBAM (single-sided band-amplitude modulation), pulse width modulation (PWM), etc., have been proposed to lower THD and implement an effective system. PWM is inexpensive and easy to implement, but it is not suitable for hearing aids because of its high THD. In this study, the authors decided to use the SRAM method, which is easy to implement and has low THD. In the SRAM approach, the audio signal’s envelope is pre-processed by taking its square root before modulating the ultrasonic carrier. This square root pre-modulation, based on Berktay’s far-field solution, effectively linearizes the demodulation process and thereby reduces undesirable harmonic distortion in the audible output.

That is, the audible signal spectrum, ${\omega }_m$, among the various signals in which the nonlinear density modulation of air occurs, corresponds to the difference frequency component between the carrier, ${\omega }_s$, and the sideband’s signal spectrum, ${\omega }_s\pm {\omega }_m$. Parametric audio has a characteristic in which the audio power gradually increases with distance from the radiator and then attenuates, and is transmitted with high directivity. At this time, when the sound pressure of the SRAM wave is $p_0$, the sound pressure, $p\left(t\right)$, of the resulting output audio signal can be described by the following Equation (2), which is well known as far-field solution, as described in [16]:

$$p\left(t\right)={\displaystyle \frac{\beta p_0^2{S}_0}{16\pi {\rho }_0{c}_0^{4}r{\alpha }_0}}{\displaystyle \frac{d^2}{d{t}^2}}{E}^2(\tau )$$

(2)

where β is the nonlinearity coefficient of the air; p₀ is the amplitude of the ultrasonic carrier wave (primary pressure); S is the cross-sectional area of the ultrasonic beam; α is the acoustic absorption coefficient of the air; ρ₀ is the air density; c is the sound velocity in the air; r is the propagation distance from the source; m(t) is the modulation envelope of the audio signal.

Here, when the distance, $r$, from the sound source is very small, $\tau =t$ can be substituted. If the constant term of Equation (2) is summarized and expressed simply as $K$, it can be seen that the output sound pressure, $p\left(t\right)$, is proportional to the second-order time-derivative value of the squared envelope, ${E(\mathsf{\tau })}^2,$ as shown in the following equation:

$$p\left(t\right)\propto {\displaystyle \frac{d^2}{d{t}^2}}{E}^2(\tau )$$

(3)

In general, hearing aid receivers are designed to have the ability to provide an audio sound of up to 100 dB SPL to the hearing-impaired people using them. At this time, it is necessary to know how much primary sound pressure, $p_0$, is required to generate a parametric sound pressure of 100 dB SPL, as an ultrasonic radiator of a size that can fit inside the receiver. For this, assuming 80% modulation at 1 kHz, we must calculate the $p_0$ required in Equation (3).

In Equation (2), except for the $p_0^2$ term and the $E(\tau )$ term, all of the terms can be treated as proportional constants, so the proportional constant is calculated as $\mathrm{K}\approx 4.903\times {10}^{-15}$ using the parameter values shown in

Table 1. Parameters for the calculation of primary beam sound pressure.

Target Pressure	Modulation Index	Envelope Freq.	Air Density	Sound Velocity	Absorption Coefficient	Non- Linearity	Radiator Radius	Object Distance
$p(t)$	$\mathrm{m}$	$f_m$	${\rho }_0$	$c_0$	${\alpha }_0$	$\beta$	$a$	$r$
100 dB SPL	0.8	1 $\mathrm{k}\mathrm{H}\mathrm{z}$	1.21 $\mathrm{k}\mathrm{g}/{\mathrm{m}}^3$	343 $\mathrm{m}/\mathrm{s}$	$1.37 {\mathrm{N}}_{\mathrm{p}}/\mathrm{m}$	1.2	0.3 cm	0.5 cm

. A sound pressure of 100 dB SPL corresponds to $=2$ ${\mathrm{N}/\mathrm{m}}^2$, and since the carrier is AM-modulated, the amplitude of the envelope, $E(\tau )$, is substituted with $E(\tau )={\displaystyle \frac{m{P}_O}{2}}\sin (2\mathsf{\pi }{f}_m$t). As a result, $p_0\approx 131.5$ dB SPL is obtained from Equation (3). Then, the first ultrasonic pressure, $p_0$, must be supplied to obtain an audible intermediate sound pressure, p(t), of 100 dB SPL. This is achieved by applying a carrier ultrasonic wave of 130 kHz to a vibrating radiator with an ultrasonic radius of 3 mm. This is calculated to be approximately 132 dB SPL.

2.3. Measurement of Frequency Characteristics

To measure the frequency characteristics of the ultrasonic receiver using the proposed ultrasonic parametric audio modulation approach, a data acquisition system (DAQ, NI PXI-4461 in NI PXI-1042; National Instruments, Austin, TX, USA), LDV (laser Doppler vibrometer) (OFV-5000 and OFV-545, Polytec, Karlsbad, Baden-Württemberg, Germany), and a probe microphone system (ER-7C, Etymotic Research, Elk Grove Village, IL, USA) were connected to form the device shown in Figure 2a. After taking the root square of the audio signal generated by the DAQ system, the SRAM signal was obtained by producing it with the carrier signal. For this, two multiplier ICs (MPY634, Burr Brown, Tucson, AZ, USA) were used. The modulation rate was set to m = 0.8, and the characteristics of the parametric modulator, in which THD (total harmonic distortion) increases when it is too high, were considered. The output signal of the 1 Vrms SRAM modulator was amplified to a level capable of generating a parametric audio signal and supplied to the PU receiver. For this purpose, RF power with a voltage gain of 30 dB, an output of 5 W, and a cutoff frequency of 1 MHz was used. For ease of experimentation, the amp used was a push–pull amplifier composed of 2SA1012 and 2SC1814 power transistors, without a class D amp, which requires a switching stage and low-pass filter. Figure 2b shows the configuration of the experimental setup installed in a soundproof chamber with a noise level of less than 50 dBA. It shows the tube of the laser beam head and the probe microphone of the LDV installed to measure the sound pressure and vibration amplitude generated from the radiator by supplying the SRAM signal to the PU receiver. Figure 2c presents the oscilloscope waveform captured at this time. The first image is a diagram of the SRAM voltage (yellow) applied to the receiver, the second presents the ER-7 microphone output voltage (cyan), and the third presents the LDV output waveform (magenta).

Figure 3 shows the sensitivity characteristics of the PU receiver measured at distances of 3, 6, and 10 mm, and is plotted along with the frequency characteristics of the BA receiver at a distance of 10 mm for comparison. A 130 kHz SRAM signal with a modulation rate of 80% was applied to drive the PU receiver. In order to avoid changes in vibration characteristics due to the limitations of the amplifier and the heat generation of the PZT element used, a 10 V_peak SRAM signal was applied for the test, and the power supplied at this time was 1.3 mW, and the same power was supplied to the BA receiver. When applying 10 V peak to the receiver, the voltage amplitude required to obtain 100 dB SPL sound was calculated as 33 V; to avoid changes in the vibration characteristics due to the limitations of the amplifier and heat generation of the PZT element used, a 10 V_peak was applied and tested.

For use as a practical PU receiver, a flat frequency characteristic from low frequencies to 10 kHz is desirable. However, for the measured frequency characteristics, sound pressure below 500 Hz was very low; from this frequency to 10 kHz and higher frequencies, the sound pressure increased dramatically and then decreased, reflecting the frequency characteristics of a typical parametric audio system. Regardless of l, various peaks were observed between 0.5 and 10 kHz, and the range of peak values was also larger than that of the BA receiver. A section in which the sound pressure measured at a distance greater than 3 mm from the radiator surface exceeded the sound pressure measured at 3 mm was also observed. Additionally, the average sound pressure in the range of 1 to 10 kHz was about 70 dB SPL.

The BA receiver has a maximum sound pressure at 2~3 kHz and slope of −20 dB/decade, and the gain decreases as it goes to high frequencies. However, the proposed PU receiver has an increase in gain as it goes to high frequencies and exceeds that of the BA receiver at frequencies higher than 3 kHz, as predicted. This finding implies that the receiver can be used by people with impaired sensorineural hearing who require a large high-frequency gain above 2 kHz. It is necessary to add a pre-emphasis filter to enhance the low-frequency gain because the sensitivity is lower than that of a BA-type receiver at 3 kHz or less. To be satisfactory as a practical hearing aid output device, an output sound pressure level of 100 dB at 1 kHz is required. However, the experimental result reached only 60 dB SPL. The main reason for this shortcoming is likely the inability to conduct closed-field experiments, such as those in the external auditory canal, due to limitations in the experimental setup. Additionally, the failure to apply the required high voltage may have contributed to the lower SPL. Nevertheless, the frequency characteristics of the output device are expected to remain unchanged, regardless of the experimental setup.

2.4. Directivity Measurement

Although this receiver uses ultrasonic waves, our goal was to focus an audible-band sound beam, so we measured the directivity of the parametric audio generated by the ultrasonic waves using an audible-band sound pressure meter. We measured the directivity of a beam a very short distance from a radiator with a diameter of only 6 mm. The microphone aperture of the audio level meter must be very small to measure the directivity of the beam. Therefore, in this experiment, a probe microphone system with a 0.5 mm diameter tube was used to ensure sufficient beam measurement resolution.

We also used the small directivity measuring device shown in Figure 4a to measure the directivity within a few centimeters. The silicone tube connected to the measurement microphone is flexible, so this directivity-measuring instrument involved inserting a rigid guide tube with an inner diameter of 1 mm. We used a slider with a groove depth of 7 mm and groove width of 2 mm that can be moved by inserting a protractor or ruler to measure the directivity by moving in circles or straight lines.

The diameter of the radiator was only 6 mm, so it was difficult to measure the directivity of an audio beam radially at a distance of 6 mm or less from the radiator. Therefore, at 6 mm, we measured the sound pressure of the beam cross-section across a straight line. At a distance of 10 mm, we measured directivity in the radial direction. Measurements in the radial direction were determined by dividing the sound pressure measured by moving the slide at 5° intervals in the angular range of ±90° by the value of the sound pressure at the center of the ultrasound beam (0°). Figure 4b presents the measurements in the radial direction at 1 and 9 kHz; the results revealed that the measured sound pressure level was in good agreement with the calculated directional pattern value. Figure 4c presents the measured sound pressure profile, i.e., the axial sound pressure distribution of the cross-section of the audio beam of the PU and BA receivers at 1, 5, and 9 kHz. At the time of measurement, a 100 mV signal was applied to a BA receiver; it was normalized to the maximum sound pressure at the center of the beam and compared with the PU receiver.

During measurement, 10 V was applied to the PU receiver and 100 mV to the BA receiver because the BA receiver is driven by a low-voltage current due to its low internal resistance, while the PU receiver is a potential driving element with a very large internal resistance. The results confirmed that the sound generated by the SRAM modulated ultrasound signal focused more than 95% of the energy within the diameter of the eardrum at 8 mm. To characterize the intrinsic beam pattern, directivity measurements were performed under free-field conditions (i.e., without a baffle or ear). This setup follows standard practices in transducer characterization [13]. Previous studies also evaluated ultrasonic hearing devices in free-field or quasi-free-field setups [17].

3. Discussion

Ultrasound hearing aids are similar in principle to middle-ear implants or light-driven contact hearing aids in the sense that they amplify sound and transmit it directly to the eardrum, avoiding the effects of the external auditory canal. More in-depth research is needed to explore whether ultrasonic hearing aids have the performance and effectiveness of typical hearing aids. This area of research is in its infancy, and the main focus here was to determine how much sound pressure can be delivered to the eardrum with a parametric modulated ultrasound beam and how much power the PU receiver should supply at that time.

Our results confirmed that the frequency-specific directivity of the audible sound obtained after SRAM modulation at 130 kHz was within the spread angle of the theoretical beam pattern of the ultrasound beam, and that all the acoustic power was concentrated within 8 mm—the size of the eardrum—at a distance of 6 mm. To clarify, the ultrasonic transducer emits a 130 kHz carrier that is amplitude-modulated by a 0.5–10 kHz audible signal. As the modulated wave propagates through air—a nonlinear medium—the ultrasonic carrier component undergoes progressive attenuation, while the audible signal is recovered via self-demodulation. Therefore, although the emitted waveform is ultrasonic, the effective acoustic output perceived and measured at the eardrum is in the audible range. The frequency characteristics revealed a larger-than-expected difference between the low- and high-frequency gain, and revealed that the high-frequency gain with gain fluctuation after 3 kHz was higher than that of the BA receiver. Additionally, the transient time of the ultrasonic receiver for burst sound is significantly shorter than that of the BA receiver, and this is expected to improve the sound intelligibility of hearing aids and benefit individuals with sensory neural hearing impairments.

The transient time of the ultrasound receiver for the tone burst experiment was significantly shorter than that of the BA receiver, which is an important advantage. In general, however, the sound generated through parametric modulation was not very good in terms of THD, and this needs to be improved.

The receiver consumes more power than an EM receiver at the same sound pressure; to produce sound greater than 80 dB SPL, it generates heat despite the use of multilayer PZT with high vibration efficiency. To solve this problem, a PZT device with a smaller dielectric loss and more layers will be required. One possibility is PMN-PT (lead magnesium niobite–lead titanate), which has a piezoelectric constant (d_33) of more than twice that of PZT, and a dielectric loss (tan δ) that is only 27% of that of PZT [18,19]. We expect that a multilayer element with a piezoelectric constant greater than that of PZT and generating less heat will be commercialized. In parallel with the power consumption problem, the voltage applied to the PZT device must be reduced to 5 V or less for practical use in hearing aids because the battery voltage used in a hearing aid must be considered.

Another advantage of middle-ear implants or light-driven hearing aids is that the feedback sound is very minimal, so ringing can be avoided even if the ear canal is opened. The PU receiver is expected to yield this advantage because it can reduce the sound leakage compared to typical hearing aid receivers, due to the characteristics of its ultrasonic waves with small diffraction. Therefore, it will be necessary to design a cap for the PU receiver to take advantage of these characteristics and prevent the occlusion effect.

The most important issue remaining is the biosafety of ultrasonic hearing aids, which requires further research. Most published guidelines about the maximum output of hearing aids range from 100 to 115 dB SPL, and no specific data are available beyond that [13,17,20]. However, almost all commercial parametric loudspeakers easily exceed 120 dB SPL within a distance of 1 m. In fact, the sound pressure required for conversation by the general public is about 60 dB SPL: hearing-impaired individuals need louder sound, but an ultrasonic carrier sound pressure of about 140 dB SPL is not always required. In the future, more research will be needed to develop parametric sound techniques that can generate greater sound pressure at lower carrier ultrasound levels. More work will also be needed to establish strict safety rules and guidelines for the manufacture and use of ultrasound hearing aids.

4. Conclusions

We developed an ultrasonic hearing aid receiver that is capable of delivering broadband sound directly to the eardrum without relying on external auditory canal resonance. Driven by a 130 kHz SRAM-modulated signal at low voltage (~10 Vpeak), the prototype receiver achieved an average output of about 97 dB SPL in the 3–10 kHz range at a distance of 6 mm from the eardrum. The focused ultrasonic approach offers significant benefits for high-frequency hearing aid performance. The ultrasonic parametric receiver provided higher sound pressure output at frequencies above ~5 kHz compared to a conventional BA receiver. Going forward, we plan to further enhance the ultrasonic hearing aid receiver and address remaining challenges, including lowering distortion, minimizing drive voltage, ensuring safety, and optimizing the design for various ear canal geometries.