Precision MEMS Timing: Microheater and Machine Learning Compensation for 21 PPT Stability at 7.5 Hours
This paper presents a dual-mode ovenized MEMS resonator clock that achieves a fractional frequency stability of 21 parts per trillion at 7.5 hours of averaging time. The device incorporates a micro-oven in-chip heater that thermally isolates and maintains the suspended MEMS resonator at its turnover temperature. The ultra-stable performance is enabled by leveraging a machine-learning algorithm to dynamically compensate for residual environmental perturbations within the frequency tracking electronics.
Authors
- Gabrielle D. Vukasin (ORCID: https://orcid.org/0000-0002-4301-1880)
- Shwetabh Verma (ORCID: https://orcid.org/0000-0003-0787-6606)
- Gaurav Bahl (ORCID: https://orcid.org/0000-0001-7801-2739)
- Jiheng Jing (ORCID: https://orcid.org/0000-0002-8645-7890)
- Ryan Kwon
- Jintark Kim
- Jie Yan (ORCID: https://orcid.org/0000-0003-1691-4002)
- Thomas W. Kenny
- Jiawei Yang (ORCID: https://orcid.org/0009-0006-4259-7847)
- Manaka Gomi
Institutions
- University of Illinois Urbana-Champaign (US)
- Stanford University (US)
Publication Details
- Published
- 2025-06-29
- DOI
- https://doi.org/10.1109/transducers61432.2025.11110317
- Citations
- 2
- Primary Topic
- Analytical Chemistry and Sensors
- Type
- article
- Field-Weighted Citation Impact
- 1.47