High Speed Thermal Imaging of Melt Pools Reveals Underlying Thermophysics in Additive Manufacturing Processes

At its heart, additive manufacturing (AM) of metals is a heat transfer problem that depends on many nano and microscale thermophysical processes. Laser powder bed fusion, the most popular AM process uses a moving heat source (e.g. laser or an electron beam) to sinter metal powder together in predefined patterns, to build parts layer by layer. Temperature fields control the stability of the melt pool and common defect structures in laser powder bed fusion (L-PBF) parts. Experimental measurements of the melt pool temperature using conventional infrared imaging techniques or pyrometry lack the temporal and spatial resolution needed to determine melt pool temperature profiles. We have developed an alternative experimental method to measure melt pool and surrounding temperatures using a high-speed color camera (&gt;50,000 frames per second). This method creates real-time thermal imaging AM tool leveraging the principle of two-color (a.k.a. dual-wavelength) pyrometry, where each pixel acts as a two-color pyrometer. Relative to conventional thermal imaging approaches, two-color pyrometry is advantageous because it is less sensitive to melt pool emissivity, plume transmissivity, and the camera’s view factor. Our initial two-color experiments were performed using a Photron AX200 high speed color camera where the temperature given by the ratio of the red to green pixel values was validated against the NIST blackbody source.<br/>Temperatures from melt pools of Ti-6Al-4V, Inconel, and 316L-SS were taken as a function of laser power and scan velocity offer new information about the thermophysical processes occurring in the melt pool. Our high resolution temperature images are amongst the first taken of the melt pool and suggest that temperatures peak between 3500-4000 K in these materials. These temperatures, which are directly under the laser, can surpass the metal's vaporization temperature causing the vaporization of material into a metal vapor plume. The mass loss due to vaporization, which is modified by the accommodation coefficient, results in a downward pressure on the melt pool, inducing convective flows towards the back of the melt pool. In addition, a vapor depression is formed such that the laser reflects internally, increasing the effective absorptivity for higher energy density processing conditions. Surface tension gradients can also induce convective currents in melt pools, known as Marangoni convection. Temperature-dependent specific heat, thermal conductivity, and density also impact the melt pool and surrounding temperatures. Computational fluid dynamics packages such as FLOW-3D incorporate these melt pool physics but require several input parameters that are uncharacterized. Surface temperatures measured with our high-speed two-color thermal imaging system are a starting point to fit uncharacterized parameters and validate these complex multi-physics models.