Femtosecond laser micromachining (FLμM) has emerged as a powerful technology for microfabrication due to its ultrafast laser-material interaction, which minimizes thermal diffusion and allows localized energy deposition. While FLμM is capable of producing features in the micrometer range, true machining precision is determined not only by the ultrashort pulse interaction itself but also by the accuracy and repeatability of positioning and process control during machining. Precision in FLμM results from the combined influence of ultrashort pulse ablation physics, machine stability, and advanced monitoring strategies. Despite ongoing technological advances, fs laser ablation remains difficult to control due to the highly dynamic nature of the process. Within a few hundred femtoseconds, electrons absorb energy far faster than it can be transferred into the lattice, resulting in an electron-lattice non-equilibrium state. This can lead to plasma formation, phase explosion, or subsurface melting depending on fluence and pulse accumulation. These interactions are strongly threshold-dependent and highly sensitive to variations in pulse energy, focus position, or material properties, making consistent material removal challenging without in situ monitoring. Currently, FLμM relies heavily on ex-situ characterization methods such as confocal microscopy and interferometry, which increase cycle time and offer no in-process feedback or control. This thesis focuses on the development and implementation of in-situ monitoring systems for FLμM, aiming to enhance process repeatability, quality control, and defect detection. The research explores the integration of multi-sensor approaches, including acoustic emission (AE) and optical emission (OE) monitoring, to capture critical process signatures during FLμM. The work is structured around three key contributions, each addressing specific challenges in the FLμM process monitoring and control. A multi-sensor system was developed to monitor FLμM processes using AE sensors and off-axis OE detection. Firstly, high-frequency structure-borne AE signals (up to 1.5 MHz) were analyzed using physics-based methods (e.g., root mean square (RMS), mean absolute relative squared error (MARSE), short time Fourier transform (STFT)) and machine learning (ML) techniques. The results demonstrated strong correlations between AE signal characteristics and key laser parameters, such as pulse energy, scanning speed, and focal position. ML-based feature importance analysis identified critical frequency bands (350, 469, 547, and 664 kHz) for detecting process instabilities and material removal regimes, including ablation and melting. Secondly, a monitoring system based on off-axis ultra-high-speed photodiodes (UPDs) was integrated into the FLμM setup to detect optical process emissions. The system utilized three photodiodes with specific wavelengths (visible: 500-900 nm, laser beam reflection: 1030 nm, and infrared: 1100-1700 nm) to capture reflected process radiation. Experiments involving single-line ablation at varying pulse energies and scanning passes revealed the relationship between feature depth and photodiode signals. Additionally, a high-speed spectrometer was implemented to analyze the spectral distribution of plasma emissions. The results highlighted the potential of OE-based monitoring for in situ process control and quality assurance. Thirdly, the integration of AE and OE monitoring systems, combined with machine vision techniques, confocal displacement sensors, and beam monitoring, provides a comprehensive solution for in-situ quality diagnosis and process optimization in FLμM. Furthermore, the combination of AE and OE monitoring systems demonstrated the feasibility of detecting critical material removal regimes, such as ablation and melting. The proposed methodologies offer a pathway toward automated process control, repeatability, and defect detection, with potential applicability across a wide range of materials and microfabrication processes. Tight tolerances in micromanufacturing necessitate automated, controlled, and regulated processes, and the integration of sensors and monitoring units is a critical component of current research and development. This approach provides a non-invasive solution for improving the precision and reliability of FLμM. This research advances the field of FLμM by addressing critical challenges in process monitoring. The developed multi-sensor systems enable in-situ detection of process instabilities, material removal regimes, and defects, ultimately enhancing productivity and product quality. The findings contribute to the broader goal of achieving tight tolerances and high precision in micromanufacturing, paving the way for the industrial adoption of FLμM in flexible production environments. By combining advanced sensing technologies with data analytics and machine learning, this work provides a scalable framework for improving the reliability and efficiency of laser-based microfabrication processes.
Kerim Yildirim (Fri,) studied this question.