Real-time identification of clad morphology and deposition regimes in SS 316 L DED via optical emission spectroscopy

• Plasma emission intensity is established as a quantitative indicator of melt-pool geometry during laser directed energy deposition. • Distinct intensity thresholds reveal fundamental transitions between incomplete fusion, stable conduction melting, transition, and keyhole regimes. • Optical emission signals capture subsurface melt-pool evolution beyond surface-visible observations, enabling deeper process insight. • A physics-informed intensity–morphology relationship enables real-time identification of deposition state and process stability. Real-time monitoring in laser-based directed energy deposition (DED) is critical for maintaining melt-pool stability, consistent clad morphology, and controlled deposition regime transitions. However, conventional real-time monitoring-techniques such as infrared thermography and high-speed imaging suffer from emissivity sensitivity, environmental interference, and large data volumes, limiting their suitability for continuous monitoring. In this study, optical emission spectroscopy (OES) is employed as the primary real-time diagnostic at a 1 kHz acquisition rate to establish quantitative correlations between plasma-plume activity, clad morphology, and deposition regimes during SS 316 L DED. A CMOS camera is used as a secondary tool to support visual interpretation of plume behaviour. Systematic variation of laser power and scan-speed to powder-feed-rate (SS/PFR) ratios demonstrates that plasma-plume intensity at the Fe I 520.79 nm line, used as a proxy metric for plume activity, exhibits a threshold-dependent relationship with depth aspect ratio, enabling discrimination among lack-of-fusion, conduction, transition, and keyhole regimes. Importantly, the OES signal captures both surface melt-pool width and subsurface penetration depth, providing insight into melt-pool dynamics beyond surface observations. Piecewise regression enables real-time estimation of clad width (R 2 = 0.99) and depth (R 2 = 0.92) from a single-wavelength signal, establishing a quantitative, regime-aware diagnostic framework for SS 316 L DED.