The Graphene Squeeze-Film Microphone

Abstract

Most microphones detect sound-pressure-induced motion of a membrane. In contrast, we introduce a microphone that operates by monitoring sound-pressure-induced modulation of the air compressibility. By driving a graphene membrane at resonance, the gas, that is trapped in a squeeze-film beneath it, is compressed at high frequency. Since the gas-film stiffness depends on the air pressure, the resonance frequency of the graphene is modulated by variations in sound pressure. We demonstrate that this squeeze-film microphone principle can be used to detect sound and music by tracking the membrane's resonance frequency using a phase-locked loop. The squeeze-film microphone potentially offers advantages like increased dynamic range, lower susceptibility to pressure-induced failure and vibration-induced noise over conventional devices. Moreover, microphones might become much smaller, as demonstrated in this work with one that operates using a circular graphene membrane with an area that is more than 1000 times smaller than that of MEMS microphones.

Keywords: gas pressure; graphene; membrane; microphone; resonance frequency; squeeze-film effect.

Figure 1

Schematic drawings of (a) static pressure sensor, (b) condenser microphone, (c) squeeze-film pressure sensor, and (d) squeeze-film microphone. The squeeze-film devices distinguish them from conventional devices by having a pressure below the membrane that is always equal to the external pressure, caused by the relatively short equilibration time τ_eq. Instead of measuring pressure by monitoring the deflection Δz of the membrane, the squeeze-film devices determine gas pressure by monitoring its effect on the membrane’s resonance frequency f_res. (e) Illustration of a graphene drum over a cavity in a SiO₂/Si substrate. A red (633 nm) laser beam passes a polarized beam splitter (PBS) and a quarter-wave plate (λ/4) such that the reflected beam is diverted into a photodetector (PD) for analysis by a phase-locked loop (PLL, implemented in UHFLI) or vector network analyzer (VNA). A blue (405 nm) laser beam enters the red laser beam path through a dichroic mirror (DM) and excites the drum at a given frequency. The objective (Obj) focuses both laser beams on the drum. A speaker modulates the air pressure P(t) by sound waves, which in turn modulate the spring constant k of the drum and thus its resonance frequency f_res. (f) An optical image of a typical device shows a suspended drum that is connected to the environment by a venting channel. (g) The magnitude and phase as a function of frequency measured by the VNA at a pressure of 10 mbar. The continuous red line indicates the −π/2 phase shift at the resonance frequency indicated by the dashed red line. (h) Magnitude as a function of frequency and pressure as recorded with the VNA. Red circles represent the extracted resonance frequencies and the black dashed lines fits to eq 1.

Figure 2

(a) A time-snapshot of the shift in resonance frequency of the drum (blue) and the output of the reference microphone (red) due to a 1 kHz sound wave at 100 dB_SPL. (b) Waterfall plot of typical spectral responses of the drum around frequencies f_out = f_in with f_in ranging from 1 kHz to 20 kHz. All curves are offsetted by 15 dB from one another. (c) The raw spectral response of the drum (blue) and of the speaker (red). After compensating for the speaker response, the graphene drum (yellow) shows a flat response for all audible frequencies. All curves are offsetted by 25 dB from one another. In panels (b) and (c), the dB-scale is obtained by using the amplitude of 68 Hz received at f_in = 1 kHz and 100 dB_SPL (see panel (a)) as reference. (d) Response of the graphene drum (blue) and reference microphone (red) to a 1 kHz sound wave for sound pressures ranging from 84 dB_SPL to 42 dB_SPL. The black line indicates the response expected from eq 1.

Figure 3

Selected normalized spectrograms of a song recorded with (a) the graphene squeeze-film microphone and (b) the reference microphone. The red squares enable direct comparison between data in panels (a) and (b).

Figure 4

(a) Signal-to-noise ratio and (b) total harmonic distortion of the measured squeeze-film microphones (cyan circles) as a function of their resonance frequency. The SNR increases approximately linearly with the logarithm of the resonance frequency (cyan line in a). The red line in a indicates the SNR limit by thermo-mechanical noise. (c) Thermo-mechanical SNR limit at a sound pressure level of 80 dB (red) and 94 dB (blue). The orange point marks the performance of the MP23DB01HP of STMicroelectronics. The black dashed lines indicate the possible miniaturization by squeeze-film microphones at constant SNR. (d) Comparison of the mechanical compliance and thickness of the presented squeeze-film microphones (red) to other MEMS microphones (blue), to the MP23DB01HP of STMicroelectronics (orange), and different types of graphene microphones (cyan,⁻ magenta,⁻ black). Panel (d) is adapted from ref (26). Available under a CC-BY 3.0. Copyright 2023 RSC Publishing.