image: a, (a) Two-wavelength holographic set-up for LBM monitoring, (b) laser fabrication machine (BE1-3: beam expanders, PBS1-3: polarizing beam splitters, M1-7: mirrors, C1-3: beam splitter cubes, D1: chromatic filter, D2: dichroic plates, W: protection window, L1-2: lenses). b, Key-hole in melt pool on 316L substrate. (a) Bottom view of melt pool with key-hole at 88 ms after the beginning of the recording, (b) bottom view of melt pool with key-hole at 88 ms after the beginning of the recording, (c) Illustration of the Rosenthal key-hole regime (adapted from R. Fabbro, “Developments in nd:yag laser welding,” in Handbook of laser welding technologies, S. Katayama, ed., Woodhead Publishing Series in Electronic and Optical Materials, pp. 47–72, Woodhead Publishing, 2013). view more
Credit: by Matthieu Piniard, Beatrice Sorrente, Gilles Hug, and Pascal Picart
Additive manufacturing is a key technology for industry that allows building structures by progressive addition of material, with complex geometries. One of the main techniques is the laser beam melting process which consists of successively melting metal or ceramic powder layers to manufacture 3D mechanical structures. However, due to complex physical phenomena involved in the laser-matter interaction, melt pool morphology instabilities can affect the final quality of the structure and remain challenging to foresee by simulation. Several optical monitoring approaches, based on the melt pool process radiation or with a secondary illumination source, have been developed to control in real-time the process. Unfortunately, they cannot measure the 3D shape of the metallic melt pool.
French scientists from ONERA and the Laboratoire d’Acoustique de l’Université du Mans have developed a two-wavelength holographic system dedicated to metal additive manufacturing able to in-situ measure the topography of the melt pool during laser beam melting. The experimental results are presented in a paper published in Light: Advanced Manufacturing and successful tests were realized on a 316L substrate. A deep depression zone characteristic of the key-hole regime was observed at the bottom of the melt pool in agreement with the Rosenthal model. During the recording sequence, instabilities of the melt-pool regime which oscillated between the key-hole mode and the transition mode without depression were observed.
The key-hole regime is a crucial defect that cannot be identified from thermal radiation monitoring techniques, but it can corrupt the manufactured object and reduce drastically the productivity of metal additive manufacturing. Therefore it is strategic to be able to detect this defect.
The holographic module includes two linearly polarized lasers emitting in the visible range and providing a large synthetic wavelength. These lasers generate a large illumination which allows to explore the melt pool and the surrounding area which is still hot. The probe beam is co-axially aligned with the high power fusion laser emitting at 1080 nm of the simplified laser beam melting machine. The surface to be analyzed is installed on a motorized stage that permits the scanning of the processing laser. These first tests were carried out at a speed of 100 mms-1. Holograms are then obtained from the reflected probe beams and the reference beams. Then, a high-speed camera can record the spatially multiplexed two-color holograms at both the temporal and spatial scales of the laser melting process. Thermal noise which is an incoherent noise is naturally filtered in off-axis digital holography since it is localized only in the zero-order of the hologram.
To the best of our knowledge, these are the first results of in-situ full-field melt pool and track topography measurements in laser beam melting. Although the proof of concept is demonstrated, several improvements are required to increase the performance of the set-up. The main item is related to the increase of the laser power of the two probe beams. This would permit collecting more photons and thus to address non-cooperative surfaces such as 316L powders which are less reflective than the 316L substrate. Since the motion blur contributes to degrade the quality of the measured 3D shape, a more powerful laser would also minimize this effect so that the exposure time could be reduced and the velocity increased up to that of industrial processes, 1 m.s-1 typically. We plan also to simplify the optical set-up to allow its implementation in an industrial additive manufacturing machine.
This work opens the way to new strategies to optimize in-situ additive manufacturing processes and develop reliable numerical models to apprehend the dynamics of the melting process. Conclude the scientists.
Journal
Light: Advanced Manufacturing