Piezoelectric micromachined ultrasonic transducers (PMUTs) are commonly used as acoustic emitters and receivers. Here, we demonstrate that a PMUT can also serve directly as an on-device audio classifier for edge applications. This capability is enabled by physical reservoir computing (PRC), which offers a low-cost way to process and classify time-varying signals by exploiting the naturally evolving, nonlinear transient response of a physical device instead of training large internal weights. The PMUT provides two key ingredients for PRC: (1) geometric (cubic) nonlinearity, well described by a Duffing-type vibration model that produces amplitude-dependent frequency shifts and state diversification; and (2) short-term memory encoded in its ring-down and transient buildup, supplying the fading history required for temporal feature extraction. Using experimentally calibrated parameters, our numerical simulations closely match measured vibration responses, validating the model fidelity. We show that enhancing the effective geometric nonlinearity increases 10-class spoken digit (“0”–“9”) recognition average accuracy from 86.2% to 97.2%. We further raise the accuracy of our approach to 99.0% by optimizing the transient observation window—restricting the readout to the most informative portion of the PMUT’s response. Together, these two validated control-methods offer a simple, tunable way to improve model accuracy, with the potential of enabling compact acoustic systems with embedded edge inference for time-series classification.

This work introduced a novel bridge-electrode Piezoelectric Micromachined Ultrasonic Transducers (PMUT) design, which enabled actively opposite excitation of the diaphragm, achieving a doubled sensitivity in both transmit (TX) and receive (RX) without increasing fabrication complexity, regardless of the shape of the vibrational membrane. Arrays based on the proposed electrode design were fabricated and characterized, and the reconstructed B-mode ultrasound images exhibited sharper resolution.

This paper presents a systematic experimental demonstration of PMUT-based parametric array acoustics. (1) We achieved efficient generation of acoustic harmonics from a MEMS platform; compared to conventional bulk piezo stacks or Langevin transducers, the proposed PMUT array provides improved impedance matching and higher acoustic output efficiency, enabling effective nonlinear interaction in air. (2) By steering and overlapping the two primary beams (with frequency $f_{1}$ and $f_{2}$), we introduced, for the first time, a device to generate a spatially tunable focal region, where the harmonic components are confined, enabling localized acoustic energy projection with suppressed sidelobes and reduced near-field pressure fluctuations. (3) Multiple nonlinear components, including difference-, sum-, and higher-order frequencies, are experimentally detected, extending the functionality of the PMUT array toward air-coupled nondestructive testing and high-resolution 3D obstacle sensing.

We show that a piezoelectric micromachined ultrasonic transducer (PMUT), typically used as an acoustic transmitter and receiver, can also serve as an on-device audio processor for edge computing. Exploiting its intrinsic cubic nonlinearity modeled by the Duffing equation, we implement the PMUT as a physical reservoir computer. Simulations with experimentally calibrated parameters reproduce its nonlinear dynamics and closely match measurements. Using this platform, we perform 10-class audio digit classification (“0”–“9”) and improve recognition accuracy from 86.2% to 97.2% by strengthening the PMUT's nonlinear response. The results show the potential of building compact acoustic systems with the embedded edge processing capability for the time-series classification task.

This work presents a further development of a monolithic air-coupled PMUT array sensing system, operating at 200 kHz, for mid-air obstacle detection. Two significant advancements over state-of-the-art technologies are demonstrated: (1) A boundary-released PMUT structure with a coupled cantilever-plate design is proposed to enhance the sensitivity of the full transceiving loop, implemented in a linear array configuration; (2) enhanced resolution and accuracy achieved through pre-characterized acoustic field distribution and a double-stage delay-multiply-and-sum (DMAS) beamforming method. Additionally, a comprehensive development platform for acoustic MEMS devices is proposed, enabling a system-level optimization to fully exploit the performance of the entire sensing system.

This work presents a two-port "overpass" electrode design for a circular-shaped PMUT that reproduces the anti-phase inner/outer excitation effect of a traditional inner-outer electrode design with three ports. The proposed overpass connections create opposed local electric fields in a uniformly polarized piezoelectric thin film, enabling constructive charge addition while simplifying hardware connections. A circular-shaped PMUT cell with the proposed electrode design was fabricated and compared to an identical PMUT cell using the conventional inner/outer electrode layout. Air-coupled frequency response of the central velocity shows that the overpass design delivers two times higher output sensitivity than conventional electrode design with only single-electrode excitation (inner-only or outer-only), while requiring only two ports. The approach also yields a substantial increase in electromechanical coupling (k(t)(2)), and adds no complexity to the thin-film stack and driving circuitry, improving scalability for large element-count arrays.

We present and experimentally validate a novel framework that leverages physics-informed neural networks (PINNs) for advanced MEMS eigenmode analysis. Our work advances the field in three key ways: (1) we demonstrate the successful application of PINNs to calculate multiple vibration modes in MEMS diaphragms of arbitrary geometry by proposing and implementing three key physical equation as constraints; (2) we validate our PINNs model through digital holographic microscopy (DHM) measurements on a fabricated Piezoelectric Micromachined Ultrasonic Transducer (PMUT) device; and (3) we develop a highly efficient computational platform that combines our PINNs and already existing analytical models to evaluate PMUT array performance using only geometrical configurations of the array.

Industry 5.0 emphasizes smart, connected systems, integrating MEMS-based sensors, IoT, and AI to enable advanced, real-time monitoring solutions. In this study, we present a non-contact, non-destructive ultrasonic inspection method for filtration membranes using air-coupled piezoelectric micromachined ultrasonic transducers (PMUTs) suitable to be deployed in the fabrication line for in-line and real time testing of the production. Filtration membranes play a critical role in respiratory protection and ventilation systems, where ensuring consistent performance—measured via particle filtration efficiency (PFE) and pressure resistance—is essential. We designed and fabricated a novel 200 kHz PMUT array optimized for air-coupled through-transmission measurements of porous materials. Experimental validation was performed on different meltblown polypropylene membranes.. The PMUT results were benchmarked against a broadband air-coupled ultrasonic transducer. The transmission coefficient measured by the PMUT array demonstrated sensitivity sufficient to distinguish between membranes with different filtration properties. These findings suggest the feasibility of using PMUTs for rapid, high-precision quality assessment of filtration materials.

This paper presents an analytical approach to determine optimal offset signals for driving air-coupled piezoelectric micromachined ultrasonic transducers (PMUTs), aimed at effectively reducing ring-down time and broadening bandwidth without compromising transmission sensitivity. To achieve this, a time-domain equivalent circuit simulation platform for PMUTs is developed to quickly obtain and analyze the air-coupled PMUTs response. This platform facilitates to fast obtain the ring-down times for different excitation signals and allows continuous adjustment of parameters for the offset signal waveform. The optimal offset signal waveform is then identified by comparing ring-down times generated across various parameter configurations of offset signals in simulation. The effectiveness of these optimized offset signals achieved through the proposed method is confirmed experimentally with different driving signals. For example, for a PMUT cell with a radius of 360 μm, the ring-down time is decreased by at least 85.49 % with the obtained optimal offset signal for short-period driving signals, while the -6dB bandwidth of the PMUT is increased by more than 3.85 times with the obtained optimal offset signal for long-period driving signals. By suppressing ring-down, the proposed method minimizes the blind zone, sharpens the echo envelope, and enhances positioning accuracy for ultrasound applications. Furthermore, the developed simulation platform has significantly improved the efficiency of time-domain simulations for PMUTs research, providing a solid foundation for future system-level optimizations and studies on PMUTs applications.

Structural health monitoring (SHM) systems often employ Lamb wave transducer arrays for damage detection and localization. The simultaneous excitation of multiple transmitters offers an efficient data acquisition strategy. This article presents a single-channel, multifrequency simultaneous excitation system for Lamb wave-based damage imaging. First, a low-cost bandpass filter circuit module is used to divide the excitation signal into distinct frequency ranges under the single signal generator. Subsequently, the excitation signal is optimized to constrain the actual excitation applied to the transmitter to specific frequency ranges, ensuring nearly orthogonal distribution between the excitation signals of different transmitters. On this basis, the response signals are separated in the frequency domain and matched to their respective transmitters for damage localization. The effectiveness of the proposed method is validated through experiments conducted on an aluminum plate with a hole and simulated damages.

A high-sensitivity air-coupled PMUT (piezoelectric micromachined ultrasonic transducer) integrated with an advanced electrical control module is presented, surpassing state-of-the-art performance in multiple aspects. 1) An optimized electrical module with precise phase shift and attenuation control is introduced, effectively eliminating the ringdown effect with clean ripples in transmission. 2) A force feedback control loop (FFCL) is integrated, enabling tunable bandwidth performance while also suppressing ringdown effects during reception. By combining the excitation control module and the FFCL, the PMUT system significantly mitigates the ringdown effect across the full transceiving loop. More importantly, the control module resolves the inherent trade-off between bandwidth and sensitivity, achieving a remarkable reduction in the full-loop ringdown effect with minimal sensitivity loss. This work establishes a foundation for advanced air-coupled ultrasonic sensing systems with enhanced performance and versatility.

As MEMS (Micro Electromechanical Systems) structures grow more prevalent in microelectronics, cost reduction becomes increasingly important. The packaging of today's MEMS is a significant cost factor. For complete quality functioning, many MEMS structures need to operate at a low pressure, and some may even require vacuum. MEMS packaging is still mostly done with wire-bonding technique. However, one of the challenges with this technique is to protect the MEMS mechanical structures throughout the packaging process. Other than wire bonding, Flip-chip bonding is a popular integration technique. This study aims to study the effect of two different flip chip bonding techniques on the electrical performance of the PMUT (Piezoelectric micromachined ultrasonic transducers.

In this study, we introduce a novel Piezoelectric Micromachined Ultrasonic Transducer (PMUT) structure that leverages multimode vibration to significantly expand bandwidth at high frequencies. The research presents a unique “chainring”-shaped PMUT cell configuration, and an array design engineered to minimize crosstalk. Verified by simulation, the proposed design demonstrated a fractional bandwidth exceeding 100%, centered at 6 MHz. The design was fabricated and validated with experiments.

This paper proposes a new 3D spatial sensing approach via compressed sensing (CS) by using a single-channel air-coupled piezoelectric micromachined ultrasonic transducer (PMUT) operated with multi-frequency. Our study focuses on a single-channel transducer with a PMUT array composed of several diaphragms with different radius sizes. It is known that small variations in the radius size can cause distinct transmission signals of all diaphragms that are excited by the same excitation signal. In this way, the acoustic field distribution of a region of interest (ROI) can be distorted especially in the direction perpendicular to the wave propagation, which could help to obtain more distinctive information about the scatterers at different locations in any 3D ROI. Therefore, a compressed 3D spatial sensing approach is proposed and used for acquiring measurements of the designed single-channel transducer. The information of any object in a 3D ROI can be mapped onto a collection of basis functions constructed via the nearly mutual orthogonal echo signals from all scatterers in the ROI. Furthermore, the proposed approach is verified with simulated acoustic measurements obtained from the established PMUT equivalent circuit model and the K-Wave acoustic propagation model via an obstacle-sensing application. Based on the sparsity nature of objects in the ROI, the reconstruction of 2D/3D images of objects can be accomplished via a CS-based algorithm. The obtained image reconstruction results show that the proposed approach allows not only for detecting localization but also for reconstructing descriptive features of an object.

A novel PMUT is presented, which outperforms the state-of-the-art in two aspects 1) a force feedback control loop (FFCL) is integrated, allowing tunable bandwidth performance, 2) An electrically controlled feedback force is used to mimic the geometric and material nonlinearities of the structure to realize the hardening spring effect in small vibration amplitude conditions, which improves the working flexibility and stability of the device. The abovementioned novelties ensure the dynamic tunability of the frequency response of the PMUT, which increases the versatility of the device to meet the requirement of different applications.

Noninvasive and real-time flowrate measurement for small pipes is crucial in medical diagnostics and aviation hydraulic system health monitoring. However, current bulk PZT-based ultrasonic flowmeters are unsuitable for these scenarios due to their large size, high power consumption, and poor integration with ICs. This paper proposes a highly accurate and universal ultrasonic flowrate measurement technique for pipes with small diameters (<= 15 mm) using piezoelectric micromachined ultrasonic transducers (PMUTs) by taking advantage of their miniaturized size, low power consumption, and superior integration capability. A clamp-on testing method is proposed for ultrasound transmitting and receiving. The inside liquid flowrate is measured based on the propagation time difference between the upstream and downstream ultrasound waves. To enhance accuracy, a unique PMUTs chip with a center circular element and an outer annular element is developed to improve acoustic pressure. Additionally, the incident angle and working frequency of ultrasound waves are optimized to reduce the interference caused by ultrasound wave superposition at the PMUTs and pipe wall interface. This optimization also improves acoustic energy transmission and response. The flowrate of water in a steel pipe is successfully measured, demonstrating excellent linearity (99.69%) between flowrate and time difference, with a high accuracy of 2%. Furthermore, flowrate testing for pipes with different diameters and pipe wall materials is experimentally demonstrated, showing widespread usability. The developed PMUTs-based flowmeter is versatile and compatible with common pipes, allowing compact integration in tight spaces, making it highly promising for practical applications.

Efficient implementation of ultrasound (US) imaging systems is crucial for US applications. Previous studies have shown the utilization of compressive sensing (CS) to simplify US systems by reducing the channels of receiver arrays. We introduce a CS-based optimization approach tailored for the design of receiver arrays without any prior information about target objects. Simulation results demonstrate the benefit of the proposed optimization approach compared to previous uniform or random selection scheme based receiver channel selection in terms of location estimations and shape estimations of objects from reconstructed images.

This study investigates the efficacy of multiple machine learning (ML) strategies for optimizing the design of freeform Piezoelectric Micromachined Ultrasonic Transducers (PMUTs) by leveraging a data-centric methodology. We devise a comprehensive four-stage optimization framework comprising a freeform PMUT shape generator, a feature extractor, a finite element analyzer, and ML estimators. The ML evaluation compared to the finite element analysis reveals that the leading ML estimator accomplished over 95% prediction accuracy with notably low error rates. With this framework, a dataset comprising 30,000 samples was processed within 6 seconds, facilitating the rapid selection of optimal PMUT configurations. Our findings highlight the potential of ML methods to significantly accelerate and optimize PMUT design, resulting in improved sensitivity and precise operational frequency control.

In this paper, we propose a novel piezoelectric micromachined ultrasonic transducer (PMUT) for airborne object detection, which provides high sensitivity compared to the conventional round-shaped PMUT structure. This paper reports the concepts, fabrication and characterization of the new structure. The proposed structure shows a 13.3% improvement in transmitting (TX) sensitivity compared to the conventional one.

This paper presents a broadband ultrasonic rangefinder, with merits of a small blind area and high accuracy, based on a multi-frequency piezoelectric micromachined ultrasonic transducer (MF-PMUT) array. The coupling bandwidth between six types of resonant cells with adjacent resonant frequencies was studied and optimized, and these cells were then integrated into an array to broaden the bandwidth of the device. The array device showed a wide −6 dB fractional bandwidth of 108% in silicon oil. A wide bandwidth can contribute to obtaining the maximum steady-state with less excitation and reducing its residual ring-down, resulting in a small blind area. Pulse-echo ranging experiments demonstrated that the blind area was effectively reduced to 5 mm due to the multi-frequency array. For short distance measurements (<250 mm), the error was well controlled within ±0.3 mm and the 3 σ accuracy was 0.445 mm. Compared with earlier reported PMUTs designed for bandwidth enhancement, the proposed broadband MF-PMUT array is much simpler in design and fabrication, revealing it as a promising candidate for short-distance high-accuracy measurements.