Profilometry, Modal Analysis and Q-Factor Measurement on PiezoMUMPS

This work presents a cantilever-based piezoelectric MEMS resonator array, developed using the PiezoMUMPs process, for passive frequency decomposition of ultrasonic acoustic emission (AE) signals in the 90–250 kHz range. The array, consisting of 16 silicon cantilevers with varying lengths, aims to provide low-power, in-sensor frequency analysis for AE-based non-destructive testing. Experimental results demonstrate high quality factors (Q = 350–900), open-circuit voltage sensitivities of 0.1–0.3 mV/Pa, and charge sensitivities of 7–25 fC/Pa, supporting the feasibility of integrated MEMS front-end filtering for AE source characterization.

Role of the PICOSCALE Vibrometer

1. Static Profilometry:
   • The PICOSCALE Vibrometer was used to measure curvature of stationary cantilevers.
   • A static upward deflection of ~300 nm was detected in 300 µm long beams, attributed to residual compressive stress in the thin-film AlN/Al electrode stack.
2. Modal Response Imaging:
   • The PICOSCALE system enabled high-resolution modal analysis of cantilever resonances
   • It confirmed coupling effects between adjacent elements due to lateral blow-out during DRIE etching, visible as overlapping modal shapes in neighboring beams (see Fig. 3c, page 3).
3. Frequency-Swept Displacement Measurements:
   • A calibrated wideband shaker stage, also driven via the PICOSCALE system, was used to excite the MEMS chip with a controlled sinusoidal sweep.
   • Displacement amplitudes were recorded at each cantilever tip, revealing distinct resonant peaks across the designed frequency scale (90–250 kHz) for individual array elements (see Fig. 3d, page 3).
   • These measurements validated the designed eigenfrequency distribution and revealed slight deviations (up to ~22 kHz) from FEM predictions, attributed to fabrication-induced variations.
4. Assessment of Q-Factors:
   • The mechanical Q-factors were extracted from -3 dB bandwidths of the resonance peaks, ranging from 350 to 900 in air, highlighting the suitability of the resonators for narrowband frequency selection.

Conclusion

These results demonstrate the effectiveness of the PICOSCALE Vibrometer in precisely characterizing MEMS resonator behavior, validating the design for passive ultrasonic frequency decomposition. The insights gained support future integration of such MEMS arrays into low-power AE sensing systems.

Acknowledgment

We thank Dinko Oletić and his team at the University of Zagreb for their valuable cooperation and for providing access to their work and experimental data.

Download Link

Designing cantilever-based resonator arrays in PiezoMUMPs technology for in-sensor ultrasonic frequency decomposition of acoustic emissions | IEEE Conference Publication | IEEE Xplore