Laser-based powder bed fusion of metals (PBF-LB/M) is a complex process that is strongly determined by the dynamic properties of the rapidly moving melt pool. Investigating spatter formation mechanisms, melt pool oscillations, as well as depth, size, and shape of the melt pool is necessary for an in-depth understanding of process behavior. However, the small scale of the melt pool, high scanning velocities, and bright thermal and metal vapor emissions complicate direct observations of the process zone often requiring complex equipment such as x-ray systems that are only accessible for limited applications. Therefore, in this paper, an advanced in situ process monitoring system consisting of a narrow bandwidth laser illumination combined with a high-speed camera is implemented into an experimental laboratory PBF-LB/M machine. This system enables the visualization of melt pool behavior independent of process emissions at 50 000 FPS. The image processing algorithms for determining geometrical features like melt pool size, shape, and aspect ratio are described. Furthermore, the frequencies of the dynamic melt pool surface oscillations are extracted and correlated with the processing conditions. This technique is then applied on the PBF-LB/M-process of titanium in a monosilane-doped extreme high vacuum adequate atmosphere with an ultralow oxygen content of 10−15 ppm. The investigations indicate a strong correlation between the energy input and the oscillation frequency ranging from 4.1 kHz at 2 W/mm to 10 kHz at 0.5 W/mm as the oscillatory mass decreases. The results demonstrate the potential for monitoring the melt pool depth by observing surface oscillations.
May et al. (Sun,) studied this question.