Ultra‐fast, selective, non‐melting, laser sintering of alumina with anisotropic and size‐suppressed grains

Abstract

In this paper, we report an ultra‐fast sintering phenomenon of alumina achieved by the scanning laser irradiation method. Using CO2 laser irradiation, we found that micrometer‐sized alumina powder (d50 = 1.2 µm) can be sintered close to full density within a few tens of seconds. The microstructure of laser‐sintered alumina was different from that of the furnace‐sintered alumina. The relative density and grain size of the laser‐sintered alumina gradually decreased from the center of the laser beam to the edge. Anisotropy of the grain size was measured along and perpendicular to the scanning direction. This anisotropy decreased as the scanning speed decreased from 0.1 mm/s to 0.01 mm/s. The sintering master curve of grain size versus relative density, which reflects the sintering mechanism, was found to be affected by the laser scanning speed. When the laser scanning speed was 0.1 mm/s, grain size suppression was found for the almost fully dense alumina. However, at lower scanning speed (e.g., 0.01 mm/s), there was significant grain growth in the regions where the relative density was greater than 90%. These results clearly indicate that alumina can be sintered, in the solid‐state, to a high density in a short time using scanning laser and the microstructure is different from the furnace‐sintered alumina.

Machine Learning-based Microstructure Prediction During Laser Sintering of Alumina: A Conditional Generative Approach

“…Then we use RCWGAN-GP to predict the microstructure of laser-sintered alumina at the laser power of 1.6 W. Predicting the microstructure of laser-sintered alumina is challenging, since the processing mechanism is still unclear. In our previous study, we demonstrated ultrafast, solid-state laser sintering of alumina [25]. In that study, the resulting alumina's microstructure deviates from the typical sintering master curve [30,31].…”

“…In addition, the predicted microstructure features also accurately reflected the trend of the influence of laser power. In our previous study, we showed that a higher laser power resulted in a larger grain size, larger relative density, and lower porosity [25]. The predicted SEM micrographs under 1.6 W of laser had grain size and relative density that are clearly larger than those of 1.4 W, but smaller than those of 1.9 W. We examined both situations of synthesizing SEM micrographs using random seeds and a single seed.…”

“…The processing parameter of interest is the laser power. The experimental procedure is described in our previous study in details [25]. We initially build our ML network based on CGAN, since CGAN allows the combining of SEM data with processing parameter.…”

“…In the above case study, it is difficult for us to have great statistical comparisons of grain and pore sizes between RCWGAN-GP predicted SEM micrographs and experimentally obtained ones. This is because we usually determine the statistics of grain and pore sizes using manual methods on the SEM micrographs [25]. We were not able to characterize the grain size and pore size distributions manually because the datasets contain thousands of SEM segments.…”

“…Sintered alumina was characterized using SEM to study the influence of laser power on microstructure. The detailed processing procedure is reported in our previous work [25]. During a typical laser sintering experiments, alumina paste is casted on a fused silica substrate, using doctor's blade, with a thickness of about 500 μm.…”