Hyper-Resolution Spatially Resolved Marangoni Flow Characterization Using Optical Coherence Tomography and Machine Learning Regression for Uniform Spin-Coated Films

**Abstract:** Conventional spin-coating processes for thin film deposition are frequently plagued by hydrodynamic instabilities, primarily the Marangoni effect, leading to non-uniform film thickness and compromised device performance. This research introduces a novel and scalable method for characterizing spatially resolved Marangoni flow dynamics during spin-coating using optical coherence tomography (OCT) and machine learning (ML) regression, enabling real-time feedback control for fabrication of uniformly thin films. We demonstrate the capability to predict film thickness across a substrate with >95% accuracy based on dynamically captured flow field data, offering a pathway towards automated process optimization. This technology has immediate commercial applicability in microelectronics manufacturing, solar cell fabrication, and organic electronics, potentially reducing material waste by up to 20% and improving device yield.