Low-noise accelerometers are crucial for various applications, including structural health monitoring, machine condition monitoring, inertial navigation, and earthquake early warning. While traditional piezoelectric accelerometers offer superior performance, the demand for lighter, cheaper alternatives has driven the development of low-noise capacitive MEMS accelerometers. These capacitive devices, however, often require vacuum packaging to reduce noise, increasing cost and complexity. They are also susceptible to temperature changes, leading to thermal drift. Piezoelectric accelerometers, conversely, possess high quality factors, reducing the need for vacuum encapsulation and offering good temperature durability and linearity. This study explores the advantages of microfabrication to develop piezoelectric MEMS accelerometers, aiming to provide low-noise, smaller, and more cost-effective solutions. Cantilever beam and symmetric suspension structures are commonly adopted for piezoelectric MEMS accelerometers due to their higher sensitivity. Materials like aluminum nitride (AlN), lead zirconate titanate (PZT), and zinc oxide (ZnO) are commonly used, with PZT offering the advantage of much higher piezoelectric constants. Existing research has explored various combinations of structures, materials, and deposition methods, such as sol-gel and magnetron sputtering, but these methods have limitations in achieving the desired thickness of PZT film for optimal sensitivity. The aerosol deposition method, capable of depositing thick, uniform PZT films at room temperature, offers a superior approach to overcome these challenges. This paper focuses on utilizing aerosol deposition to fabricate a cantilever beam MEMS accelerometer and assesses its suitability for low-noise applications.

Several studies have focused on developing piezoelectric MEMS accelerometers using different structures, materials, and fabrication methods. Wang et al. used sol-gel deposited PZT films with annular diaphragm structures, achieving sensitivities from 0.77 to 7.6 pC/g and resonant frequencies from 35.3 to 3.7 kHz. Hewa-Kasakarage et al. developed PZT cantilever accelerometers using sol-gel deposition, achieving 3.4–50 pC/g charge sensitivities and 60–1.5 kHz resonance frequencies. Saayujya et al. used ZnO in a similar cantilever structure, achieving 1.69 mV/g voltage sensitivity and 2.19 kHz natural frequency. Gesing et al. used AlN in a four-symmetric suspension beam structure and achieved a noise density of 510 µg/√Hz, 0.0981 pC/g charge sensitivity, and 19.1 kHz natural frequency. Trivedi et al., using a four-symmetric structure with a 1 µm PZT layer, obtained a noise density of 5800 µg/√Hz, 8.12 mV/g sensitivity, and 9.62 kHz natural frequency. However, none of these studies used thick PZT film deposition (tens of µm), despite its potential benefits for sensitivity. The commonly used sol-gel deposition is unsuitable for creating thick films due to cracking during heat treatment. Aerosol deposition, a room-temperature process, avoids this issue and enables the fabrication of high-quality, thick films quickly.

This study employed a cantilever beam structure for the 1-axis accelerometer, comprising four layers: a Pt/Ti top electrode, a PZT sensing layer, a stainless-steel bottom electrode, and a tungsten proof mass. The design protects the PZT layer during wet etching and prevents short circuits. The accelerometer operates in the d31 mode. The target application was structural health monitoring, requiring a working bandwidth above 100 Hz and noise density below 30 µg/√Hz. COMSOL Multiphysics software was used for simulations, considering the material properties of PZT, stainless steel, and tungsten. Rayleigh damping with a damping ratio of ζ = 0.006 was used. Boundary conditions were set with the left side of the cantilever fixed, and a uniform harmonic force density was applied to the proof mass. The bottom surface of the PZT layer was grounded, and the top surface acted as the charge output terminal. Simulations determined the first natural frequency (898.9 Hz) and frequency response. The charge sensitivity was determined from the linear relationship between output charge and excitation acceleration. Noise analysis was performed by calculating the thermal mechanical and thermal electrical noise equivalent accelerations. Experimental measurements were conducted to determine frequency response, sensitivity, and noise performance. The frequency response was measured from 10 to 1300 Hz at various accelerations. Transmissibility was calculated by dividing output charge by acceleration. The working bandwidth was determined from multiple frequency response measurements. The relationship between excitation acceleration and output charge was measured to determine charge sensitivity. Noise equivalent acceleration was measured from 10 to 300 Hz. The fabricated accelerometer was tested by measuring the vibrations of a fan and compared to a commercial piezoelectric accelerometer. A comparison with a low-noise capacitive MEMS accelerometer (ADXL1001) was also performed.

The simulation results showed a first natural frequency of 898.9 Hz and a working bandwidth of 0.1 to 200 Hz (within ±5% deviation). The simulated charge sensitivity was 36.8 pC/g, and the noise equivalent acceleration was 8.3, 5.9, and 2.7 µg/√Hz at 10, 20, and 95 Hz respectively. Experimental measurements confirmed a first natural frequency of 867.4 Hz at 0.1g. The working bandwidth was determined to be 10–200 Hz with less than 5% deviation from the transmissibility at 95 Hz. The measured charge sensitivity was 22.74 pC/g. The measured noise equivalent acceleration ranged from approximately 4.5 µg/√Hz at 10 Hz to approximately 6 µg/√Hz at 20 Hz. The experimental results validated the simulation results and demonstrated a good match between the fan vibration measurements from the fabricated sensor and a commercial piezoelectric accelerometer. A comparison with the ADXL1001 revealed significantly lower noise levels for the fabricated sensor.

The results demonstrate the successful fabrication and characterization of a low-noise piezoelectric MEMS accelerometer using aerosol deposition. The close agreement between simulation and experimental results validates the design and fabrication process. The achieved noise levels are significantly lower than those reported in previous studies using alternative methods for PZT film deposition, highlighting the advantages of aerosol deposition in creating thick, high-quality PZT films. The sensor's performance compares favorably to other piezoelectric MEMS accelerometers and commercially available low-noise capacitive MEMS accelerometers. The demonstrated capabilities show its potential for applications requiring high sensitivity and low noise, such as structural health monitoring and other vibration sensing applications.

This study successfully demonstrated a novel approach to fabricating a low-noise piezoelectric MEMS accelerometer using aerosol deposition. The resulting device exhibits excellent performance, with a low noise floor and a broad working bandwidth, suitable for various applications. The aerosol deposition technique proved highly effective for creating thick PZT films, enhancing sensitivity without compromising mechanical strength. Future research could explore optimizing the design parameters, such as beam dimensions and proof mass, for further performance improvements. Investigating the long-term stability and reliability of the sensor under various environmental conditions would also be beneficial.

The current study focused primarily on the sensor's performance in a controlled laboratory environment. Further investigation is needed to assess its performance under real-world conditions, which might include temperature variations, humidity, and external interference. A more comprehensive analysis of the error sources and their impact on the accuracy of the measurements should be performed. While the noise performance was shown to be superior to some existing devices, a more extensive comparison with a wider range of commercially available accelerometers would strengthen the conclusions.