Non-invasive microscale electric field measurements using LIBS technology

Abstract

This paper proposes a non-invasive micrometer-scale electric field measurement method using Laser-Induced Breakdown Spectroscopy (LIBS) combined with machine learning. With the increasing demand for higher voltage in power systems, ensuring insulation performance at the microscale is crucial to prevent breakdowns caused by high local electric field strengths. Traditional electric field sensors are limited in their ability to measure at the micron level and are invasive.The proposed method leverages the advantages of LIBS for fast, real-time, non-contact measurements. It investigates how the applied electric field influences laser-induced plasma dynamics using emission spectrum, Mach–Zehnder interferometry, and imaging. The results show that increasing electric field strength weakens atomic spectral lines, enhances ionic spectral lines, and induces a redshift in ionic wavelengths. These changes are attributed to enhanced collisional ionization and the Stark effect. Additionally, the electric field slows plasma expansion and reduces imaging intensity. A machine learning algorithm is used to build an electric field prediction model, and the incorporation of laser energy fluctuations significantly improves the electric field prediction. The model's performance is significantly improved by incorporating laser energy fluctuations, achieving R² = 0.97 and RMSE = 5.36. This method offers a promising solution for precise, non-invasive electric field measurement at the micrometer scale, with potential applications in UHV system optimization.