1. Introduction

Capacitive micromachined ultrasonic transducers (CMUTs) have emerged as a pivotal technology in the field of underwater acoustics, particularly in the development of hydrophones. Hydrophones, specialized acoustic sensors designed for underwater applications, are essential tools for detecting and converting underwater sound waves into electrical signals with high precision^[1−6]. They play a crucial role in diverse fields such as marine biology, underwater navigation, and seismic monitoring. Traditionally, hydrophones have relied on piezoelectric materials to convert sound waves into electrical signals^[7−10]. However, with the advent of MEMS technology, CMUTs offer a compelling alternative. CMUT-based hydrophones provide numerous advantages over conventional piezoelectric hydrophones, including smaller size, lower manufacturing cost, and better environmental adaptability^[11−13]. These benefits are particularly significant in the challenging underwater environment where pressure and density can vary widely.

CMUTs operate on the principle of electrostatic transduction, where a membrane and a fixed electrode form a capacitor^{[14, 15]}. When an acoustic wave causes the membrane to vibrate, the capacitance changes, producing an electrical signal. This method offers several advantages: CMUTs can achieve higher bandwidth and sensitivity, and their working principle results in lower harmonics, reducing imaging artifacts and enhancing signal quality.

In marine biology, CMUT-based hydrophones can be used to study the communication and behavior of marine mammals, capturing vocalizations that provide insight into species' social structures and habits. In underwater navigation, they are integral to advanced sonar systems that map the ocean floor and ensure the safe passage of vessels. For seismic monitoring, CMUT hydrophones can detect and analyze subtle seismic activities, contributing to our understanding of tectonic movements and the assessment of earthquake risks beneath the ocean floor.

This paper introduces the design and implementation of a CMUT-based hydrophone array for underwater acoustic applications. A systematic design study was conducted to achieve an optimal sensing structure and overall performance. The CMUT array was fabricated using a sacrificial release process, with sensing cells designed in an innovative multi-ring nested structure. The performance of the CMUT array was thoroughly evaluated against industry-standard piezoelectric hydrophones. The results indicate that the CMUT array demonstrates superior performance in several key areas, including bandwidth, sensitivity, and non-linearity. These findings highlight the potential of CMUT-based hydrophones to revolutionize underwater acoustic sensing, offering significant improvements over traditional technologies. The enhanced capabilities of CMUT hydrophones promise to expand our understanding of underwater environments and improve the effectiveness of applications ranging from marine research to seismic monitoring and beyond.

2. Theory and design

Fig. 1 presents a schematic cross-sectional view of a CMUT unit fabricated using a sacrificial release process. The silicon nitride vibration film is situated above the vacuum cavity, while below the vacuum cavity are the aluminum (Al) electrode layer, the silicon dioxide (SiO₂) sacrificial layer, and the silicon (Si) substrate layer.

Fig. 2 shows the equivalent receiving circuit model for a CMUT sensing cell, which includes a voltage-mode amplification circuit. The equivalent circuit model includes the acoustic, mechanical, and electrical domains. In the acoustic domain, ${P_{{\text{out}}}}$ is the output pressure, ${V_{\text{v}}}$ is the volume velocity, ${{\text{Z}}_a}$ is the acoustic impedance. The ${{\text{Z}}_a}$ comprising of a damping load ${R_{\text{a}}}$ and a mass load ${m_{\text{a}}}$ is described as the medium resistance and inductance. In the mechanical domain, ${F_{\text{e}}}$ is the electrostatic force, ${V_{\text{p}}}$ is the velocity at the center, ${m_{\text{m}}}$ and ${k_{\text{m}}}$ are derived to represent the effective stiffness and the effective mass of the sensing cell for a particular vibration mode, respectively. In the electrical domain, ${V_{{\text{ac}}}}$ is the input voltage, ${i_{\text{e}}}$ is the current, ${C_0}$ is the active capacitance corresponding to the electromechanical transformer, ${C_{\text{p}}}$ is the parasitic capacitance, $- {C_0}$ is the so-called electrostatic spring softening effect. ${A_{{\text{eff}}}}$ is effective area, which is usually equal to 1/3 the surface area of the sensing diaphragm. $n$ is the electromechanical transformer ratio, which represents the energy conversion efficiency.

The electrostatic force in the mechanical domain ${F_{\text{e}}}$ can be expressed as:

$${F_{\text{e}}} = \frac{{{\varepsilon _0}A{V^2}}}{{2{{({d_0} - x)}^2}}},$$

(1)

where ${\varepsilon _0}$ is permittivity of free space. $V$ is applied DC voltage. $x$ is diaphragm displacement. ${d_0}$ is initial gap height.

The transformer ratio $n$ can derived as:

$$n = V\frac{{{\varepsilon _0}A}}{{{{({d_{{\text{eff}}}} - x)}^2}}}.$$

(2)

Generally, a transducer's sensitivity is governed by the transformer ratio, inversely correlated with the square of the cavity depth. CMUT offers the potential for high transformation ratios due to surface micromachining capabilities, enabling the fabrication of relatively narrow gaps, sometimes as small as 50 nm, which enhances its performance.

In the equilibrium position, the electrical force is counterbalanced by the restoring force, with the mechanical force exerted by the membrane being dependent on the spring constant $k$.

$${F_{{\text{mech}}}} = kx.$$

(3)

By equating the relevant equations, the relationship between membrane displacement and the applied DC voltage can be determined.

$$V = \sqrt {\frac{{2kx}}{{A{\varepsilon _0}}}} ({d_0} - x).$$

(4)

The displacement at the point of collapse can be identified by setting $\rm{d}V/\rm{d}x$ to zero, indicating the collapse occurs at that specific displacement.

$$x = \frac{{{d_0}}}{3}.$$

(5)

The pull-in collapse voltage is

$${V_{{\text{coll}}}} = \sqrt {\frac{{8kd_0^3}}{{27{\varepsilon _0}A}}} ,$$

(6)

$${d_{{\text{eff}}}} = \frac{{{d_{\text{m}}}}}{{{\varepsilon _{\text{r}}}}} + {d_0}.$$

(7)

Here ${d_{\text{m}}}$ is the diaphragm thickness. ${\varepsilon _{\text{r}}}$ is the relative dielectric constant of the diaphragm material.

Furthermore, the capacitance of the diaphragm can be determined using the following expression:

$${C_x} = \frac{{A{\varepsilon _0}}}{{{d_{{\text{eff}}}} - x}}.$$

(8)

The maximum receiving sensitivity V/μPa of the CMUT array denoted as ${S_{\text{R}}}$ is defined as follows:

$${S_{\text{R}}} = \frac{{n{A_{{\text{eff}}}}}}{{{R_{\text{m}}}}}.$$

(9)

Consequently, achieving high sensitivity values comparable to those of piezoelectric transducers is feasible. The maximum receive sensitivity ${S_{\text{R}}}$ is influenced not only by the transformer ratio but also by the medium damping, which is dictated by the target application. Therefore, enhancing the transformer ratio remains the only viable method to improve receive sensitivity.

The current generated by the CMUT can be expressed as:

$${i^2} = {\left(\frac{{{V_{{\text{DC}}}}{C_0}}}{g}\right)^2}\int_0^{{w_{\text{H}}}} {\frac{{4kT{w_0}/(mQ)}}{{{{(w - {w^2}/w)}^2} + {{({w_0}/Q)}^2}}}\rm{d}w} ,$$

(10)

where ${w_0}$ is the resonant frequency, k is Boltzmann's constant, $Q$ is the quality factor, $T$ is the absolute temperature, ${V_{{\text{DC}}}}$ is the applied DC bias voltage, $g$ is the gap between plates and ${C_0}$ is the capacitance between the plates.

3. CMUT array fabrication

Table 1 shows the material properties used in the simulations. The detailed design parameters of CMUTs array are shown in Table 2. The CMUT array are fabricated using a silicon nitride sacrificial release process in this work. The cross-sectional overview of each main step of the fabrication process is illustrated in Fig. 3. The process starts with a silicon wafer substrate covered with a 1 μm thick oxide, and a 0.4 μm thick aluminum layer is deposited and patterned to form the electrode connection layer (Fig. 3(a)). Following 1 μm thick oxide deposition, a 0.4 μm thick Al layer is deposited and patterned to form the bottom electrodes (Fig. 3(b)). Then, another 0.05 μm oxide is deposited as the sacrificial layer and is patterned to define the cavity. Over the patterned oxide sacrificial layer, 1 μm silicon nitride is deposited as the device membrane (Fig. 3(c)). Next, contact holes are opened to access the electrode connection layer for electrical connection (Fig. 3(d)). A 1.0 μm thick Al layer is deposited and patterned to form the top electrode (Fig. 3(e)). The oxide sacrificial layer are etched to form the cavity (Fig. 3(f)). A 1.8 μm thick silicon nitride passivation layer deposition to form device diaphragm (Fig. 3(g)). Finally, holes are opened to access the electrode connection layer for electrical connections (Fig. 3(h)).

The optical microscope image of the CMUT array and the SEM image of its sensing cell are shown in Fig. 4. The size of the CMUT array is 1.5 mm × 1.5 mm and the sensing cells are arranged in a multi-ring nested manner. The CMUT array integrates driving electrodes and detection electrodes, as shown in the Fig. 4. This innovative design helps to realize the miniaturization of the imaging system. Fig. 5 showed the optical microscope image of the CMUT wafer.

4. Transducer characterizations

An industrial standard piezoelectric transducer is used as a reference to measure the receiving sensitivity of our CMUT array. As shown in Fig. 6, the CMUT array exhibits a maximum response frequency at 1.06 MHz, with a receiving sensitivity of −231.44 dB (re: 1 V/μPa) at room temperature. This high sensitivity, combined with the optimal operating frequency, makes the CMUT array an ideal choice for underwater imaging applications. The CMUT array's superior performance ensures precise and reliable detection of acoustic signals, which is critical for high-resolution underwater imaging.

The temperature stability of the receiving sensitivity of the CMUT array is investigated. The ceramic heater and the packaged CMUT array were placed in a small experimental water tank, and the real-time temperature in the water tank was measured with a commercial thermocouple thermometer (Tronovo-TR6602). During the test process, the change of the receiving sensitivity of CMUT array with temperature in the range of 0−60 °C was recorded, and the result shows that the measured receiving sensitivity variation of the CMUT array is 0.87 dB, as shown in Fig. 7. The small variation of the sensitivity with temperature validates its good stability in working state, which also means high reliability in the applications of medical imaging equipment.

The relationship between the measured output voltage and the incident acoustic pressure is shown in Fig. 8. The acoustic pressure is increased from 0 to 200 kPa when CMUT array works at 1 kHz, and the maximum non-linearity is about 0.44% by linear fitting calculation. The non-linearity of the CMUT array is obviously better than other reported products in the same frequency range.

5. Conclusion

This paper presents the design, fabrication, packaging, and characterization of a high-performance transceiver Integrated CMUT array. The transceiver Integrated CMUT array is through drive electrode and detection electrode to separate the ultrasound transmission and reception process instead of switch circuit. The designed CMUT array achieves a receiving sensitivity of −231.44 dB (re: 1 V/μPa) at a resonance frequency, the measured receiving sensitivity variation of the CMUT array is 0.87 dB at the range of 0−60 °C, and the maximum nonlinearity of output voltage at 1 kHz is about 0.44%. The reported high-performance CMUT array shows great commercial value to portable underwater imaging devices.

Acknowledgments

This work was supported in part by the National Natural Science Foundation of China under Grant 61927807, 62320106011, and 62304208, in part by the China Postdoctoral Science Foundation under Grant 2023M733277 and 2024T170848.