Investigating Degree of Particle Melt at the Centre for Advanced Additive Manufacturing (AdAM)
Over this past summer IGERT fellow Anthony D’Amico work in the Centre for Advanced Additive Manufacturing (AdAM) at the University of Sheffield in Sheffield, UK. Selective laser sintering (SLS) which is a widely commercially available additive manufacturing (AM) method applicable to both polymeric and metallic powders. As with almost all AM methods, SLS builds up parts layer by layer. At each layer, a laser scans across a powder bed to melt a pre-determined pattern. The melt then re-solidifies, forming a layer. In high speed sintering (HSS), print heads lay down ink in a pre-determined pattern and a lamp then heats the powder bed. Because the inked powder absorbs more radiation than un-inked powder, the inked pattern melts and then re-solidifies to form a layer. For semi-crystalline polymers the crystallinity of the built part and the powder used can be different. Ideally the built part would consist only of powder particles that had melted completely and recrystallized, leading to uniform crystallinity throughout the part; however this is not the case in practice. A portion of many of the powder particles remains un-melted throughout the build process. Upon recrystallization, this leads to a bulk phase with one crystallinity and un-melted cores with a different crystallinity. The mismatch in properties can weaken parts and cause failure.
To study this phenomenon, the idea of degree of particle melt (DPM) has been developed. DPM is the mass fraction of the melted phase of the final part. For parts with multiple semi-crystalline phases, different melting occurs at multiple temperatures and with multiple enthalpies of melting. By analyzing the melt peaks observable in a differential scanning calorimetry curve (DSC), the DPM can be determined. Anthony’s work over this past summer looked at how DPM varies both with characteristic length of parts printed (to examine the impact of variance in heat transfer with part size) and with changes in several process parameters. In particular he varied the amount of ink deposited, the temperature the build bed was held at, and the layer height used for the build. He found that DPM does not vary significantly with characteristic length. This indicates sacrificial test samples built during a larger part build can be tested and will be representative of the properties of the part being built. For variations in process parameters, DPM varies significantly only with changes in layer height. This indicates the current baseline parameters used provide resilience in properties even if variations occur during the build.
