Build Surface Topography Determination in Electron Beam Powder Bed Fusion using Electron Optical Imaging

Process monitoring technologies are a key cornerstone in future accelerated validation routes of high-risk high-value additive manufacturing applications in the aerospace, defense, energy or medical technology sector. Powder bed fusion processes are the most mature among the many additive manufacturing processes. Electron beam powder bed fusion utilizes an electron beam gun, as energy source, to processes metallic alloys under medium-to-high vacuum conditions. Typical optical, contactless process monitoring approaches are difficult to implement since extensive thermal radiation, metallization and X-rays plague optical ports. Among multiple important monitoring objectives, the quasi-static state of the build surface topography after preheating and after melting is of utmost importance. The melted part surface topography is a direct marker of a successful strategy. The applied and preheated powder layer topography defines the local powder thickness with the melted part surface. In summary, both quantities provide ample process-related information which can be used in process optimization or layer-based control scenarios. The development of surface topography measurements which tolerate the electron beam powder bed fusion process environment is therefore highly relevant. Electron Optical imaging is a resilient process monitoring method with many mature applications based on single-detector systems. The build surface is scanned with high scan speeds by a focused, low-power electron beam. Emitted electrons are recorded by robust single- or multi-detector systems consisting simply of conductive metal plates. Amplification of the detector currents is necessary which requires dedicated measurement equipment. Detector voltages time series data is mapped to 2D images, similar as in scanning electron microscopy. The aim of this work is to demonstrate in situ build surface topography measurement in electron beam powder bed fusion. A large-area photometric stereo like method is developed using a prototype multi-detector Electron Optical imaging system. A gradient computation chain is formulated, based on existing multi-detector works in the scanning electron microscopy community. The gradient data is input to modern, highly optimized normal or surface reconstruction algorithms which compute the build surface topography. To facilitate multi-detector Electron optical imaging system design a ray tracing based model of the image formation process is formulated and validated. The outcomes of this work enabled the design of a process-ready multi-detector system which enabled build surface topography reconstruction every layer of a build process. In addition, this work illustrates the importance of thorough electron beam characterization to further advance Electron Optical imaging. The minimal resolvable defect sizes provides the detection limit of the method and depends on the transformation of the defect by the point spread function of the imaging system which is the electron beam profile. Also, the accuracy of the reconstructed build surface topography depends on the electron beam profile. Therefore, two complementary beam characterization methods are employed to characterize the electron beam current density profile across the complete build plane. Furthermore, the contrast performance of the developed multi-detector Electron Optical imaging is determined using a newly developed test artifact.