Page 44 - MSAM-4-2
P. 44

Materials Science in Additive Manufacturing                        Measuring the porosity of AM components



            Another explanation could be that the porosity rate at the   Individual component layers are not recognizable in any
            sample surface or near the surface is generally higher, which   of the designs.
            is almost not considered in the micrograph analysis.    The sintered ceramic components have an uneven
                                                         38
            A special geometry-dependent influence may exist but   surface quality (Figure  18D-F). Areas with material
            cannot be explained comprehensibly with the micrograph   adhesions and small cracks are clearly recognizable, but
            analysis, as the component areas under consideration do   there are also good areas without irregularities, which
            not include the component geometry.
                                                               are relatively smooth. The components have also shrunk
            4.1.4. LCM                                         considerably as a result of the sintering process and the
                                                               burning out of the plastic content.
            In the LCM process, the various component designs
            were produced individually in three different print jobs   (A) Green parts
            (Figure 18A-C). The same batch of material was always   The pore size of zirconium oxide green parts printed using the
            used. The resulting part quality of all components   LCM process is typically less than 50 μm. In principle, open and
            was ultimately  comparable, without  defects  or cracks.   closed pores can occur, but the number of pores in the green
            However, the  surface  quality  of the  cuboid  and the   parts is minimal in this process and closed pores predominate.
            cylinder is occasionally characterized by superficial   The measured density values and the resulting porosities of
            material adhesions. These irregularities did not occur   all measurement methods for the LCM green parts made
            with the ball head. However, it was necessary to print   of zirconium oxide are listed in Figure 19 (accuracies of the
            with a raft so that the components adhered to the building   measurement methods are displayed in Figure 8).
            platform. This raft then had to be removed manually,
            which was not possible without leaving any residue.   Using the gas pycnometer, nearly true densities were
                                                               measured for the ceramic green parts made of zirconium
                                                               oxide. These density values correspond approximately
             A             B             C
                                                               to  the reference density of 3.72  g/cm³ determined  for
                                                               an LCM-printed green part with the same printing
                                                               parameters and the same material batch. The maximum
                                                               deviation is 0.12 g/cm³ equivalent to 3.22% and lead to
                                                               porosity values of 0.23 – 3.28%, which are caused by
                                                               the few open pores that are present as a result of the
             D             E             F
                                                               printing or cleaning process of the green parts. A slight
                                                               dependence on geometry can be recognized (maximum
                                                               deviation of 0.11 g/cm³ which is equivalent to 2.96%). It
                                                               can be seen that the simpler the component geometry
                                                               and the smaller the resulting component surface, the
                                                               higher the density.
            Figure 18. Ceramic parts printed with zirconium oxide and LCM process.
            (A-C) Green parts directly after printing; (D-F) sintered ceramic parts   The same situation results from the Archimedes
            Abbreviation: LCM: Lithography-based ceramic manufacturing   measurements with an apparent density. From this,






















            Figure 19. Measured density and porosity values of additively manufactured ceramic green parts
            Volume 4 Issue 2 (2025)                         18                        doi: 10.36922/MSAM025090010
   39   40   41   42   43   44   45   46   47   48   49