Open Access

Inconel 718 is a precipitation strengthened, nickel-based super alloy of interest for the Additive Manufacturing (AM) of low volume, complex parts to reduce production time and cost compared to conventional subtractive processes. The AM process involves repeated rapid melting, solidification and reheating, which exposes the material to non-equilibrium conditions that affect elemental segregation and the subsequent formation of solidification phases, either beneficial or detrimental. These variations are difficult to characterize due to the small length scale within the micron sized melt pool. To understand how the non-equilibrium conditions affect the initial solidification phases and their critical temperatures, a multi-length scale, multi modal approach has been taken to evaluate various methods for identifying the initial phases formed in the as-built Inconel 718 produced by laser-powder bed fusion (L-PBF) additive manufacturing (AM). Using a range of characterization tools from the bulk differential thermal analysis (DTA) and x-ray diffraction (XRD) to spatially resolved images using a variety of electron microscopy tools, a better understanding is obtained of how these minor phases can be properly identified regarding the amount and size, morphology and distribution. Using the most promising characterization techniques for investigation of the as-built specimens, those techniques were used to evaluate the specimens after various heat treatments. During the sequence of heat treatments, the initial as-built dendritic structures recrystallized into well-defined grains whose size was dependent on the temperature. Although the resulting strength was similar in all heat treated specimens, the elongation increased as the grain size was refined due to differences in the precipitated phase distribution and morphology.

The crystallization proceeds by the advance of the crystal faces into the disordered phase at the expense of the material excess, the supersaturation. Using a conservation constraint for the transformation ratio α∈[0,1] as complementing the rescaled supersaturation to 1 and a kinetic law for the normal growth velocity as function of the supersaturation raised to power g, the growth order, we derive an equation for the rate of transformation dα/dt. We integrate it for the six combinations of the three spatial dimensions D = 1, 2, 3 and the two canonical values of g = 1, 2 towards obtaining expressions for αDg. The same equation, with g = 1 and D = n (n is the so called Avrami exponent) is obtained when taking only the linear in α term from the Taylor’s expansion around α = 0 of the model equation of Johnson-Mehl-Avrami-Kolmogorov (JMAK). We verify our model by fitting datasets of α21 and α31 (from α = 0 to αupper = 0.999) with JMAK to obtain from the fit n = 1.725, 2.43, resp. We show further how the values of n depend on the value of αupper to which the fit is performed starting always from 0. Towards building a validation protocol, we start with validating α21 with published results.

Up to now, the mechanisms of surface nucleation and surface-induced texture formation are far from being understood. Corresponding phenomena are discussed hypothetically or even controversial, and related studies are restricted to very few glasses. As a main drawback, the vast majority of previous work on oriented growth of surface crystals do not consider the possible effects of glass surface treatments (like polishing defects, adhered particles and cracks, …) or ambient annealing conditions (vacuum or atmospheric pressure) which both can influence crystal nucleation and growth. Moreover, very few observations of crystal orientation related to surface crystallization were focused on separately growing crystals. In conclusion, up to now, no systematic studies on initially oriented crystal growth or nucleation from defined active nucleation sites have been pursued. Therefore, the main objective of the proposed project is to gain a basic understanding of the mechanisms of surface-induced microstructure formation in glass ceramics. We shall answer the question whether preferred orientation of surface crystals is the result of oriented nucleation or caused by other orientation selection mechanisms acting during early crystal growth. In both cases, crystal orientation may be caused by the orientation of the glass surface itself or the anisotropy and orientation of active surface nucleation defects. New techniques of experimental characterization of surface-crystallized glasses shall be applied to find answers to the questions posed above. Based on a controlled preparation and annealing of the glass sample surfaces, that includes the defined creation of active nucleation defects, and followed by the microstructural characterization especially during the very early growth stages of separated surface crystals, the orientation relations between defect and initiated surface crystal will be analyzed. In order to gain new insights, a couple of entirely new microstructure diagnostics techniques, including chemically contrasted XRM, laser-based sample preparation for XRM, and 3D electron backscatter diffraction shall be developed or applied. This way, using X-ray microscopy as a pivotal technique for the non-destructive 3D characterization of relatively large samples at superior spatial resolution (but also complemented by other microstructure diagnostics techniques, e.g. TEM elemental mappings at cross-sectioned samples) unprecedented insights into the microstructure evolution upon surface crystallization of model glasses will be gained.

Up to now, the mechanisms of surface nucleation and surface-induced texture formation are far from being understood. Corresponding phenomena are discussed hypothetically or even controversial, and related studies are restricted to very few glasses. In this talk the state of the art on mechanisms of surface nucleation are summarized. On one hand, mechanical damaged surfaces show high nucleation activity, at which the nucleation occurs at convex tips and edges preferentially. On the other hand, solid foreign particles are dominant nucleation sites at low damaged surfaces. They enable nucleation at temperatures even far above Tg. The nucleation activity of the particles is substantially controlled by their thermal and chemical durability. But no systematic studies on initially oriented crystal growth or nucleation from defined active nucleation sites have been pursued, so far. Therefore, the main objective of a just started project is to advance the basic understanding of the mechanisms of surface-induced microstructure formation in glass ceramics. We shall answer the question whether preferred orientation of surface crystals is the result of oriented nucleation or caused by other orientation selection mechanisms acting during early crystal growth. In both cases, crystal orientation may be caused by the orientation of the glass surface itself or the anisotropy and orientation of active surface nucleation defects. As a first attempt we focused on possible reorientation of separately growing surface crystals during early crystal growth. First results show clear evidence that separately growing crystals can reorient themselves as they are going to impinge each other.

The crystallization of 11BaO-25CaO-64SiO2 mol% glass powders and related effects of glass powder milling have been studied by X-ray diffraction (XRD), light and scanning electron microscopy (SEM) and differential thermal analysis (DTA). Glass powders were made by ZrO2 crushing with subsequent sieving or sedimentation or alternatively made by milling with agate, corundum and tungsten carbide. Although surface crystallization was detectable during late sintering, all powders fully sintered. Wollastonite was found as primary crystal phase followed by cristobalite and sanbornite precipitation. The latter crystallization steps caused a second DTA-peak for fine powders. DTA-peak temperatures shifted to lower temperature with decreasing particle size in any case. No significant effects of milling materials on DTA-peak temperatures were evident. As an exception, however, milling in agate caused strongly decreased temperatures of the second DTA-peak. This observation can be explained by the progressive mechanical damaging, which results from the required prolonged milling time with agate due to its small materials density, and the fact that diamond made scratches did promote surface crystallization of cristobalite. The present study thus shows that even for glass powders with excellent sinterability, milling can affect microstructure evolution.