Researchers at Lawrence Livermore National Laboratory (LLNL) have developed an all-optical ultrasound technique capable of performing on-demand characterization of melt tracks and detecting formation of defects in the laser powder bed fusion (LPBF) metal 3D-printing process. The team proposed a diagnostic using surface acoustic waves (SAW), generated by laser-based ultrasound, that can reveal tiny surface and subsurface defects in LPBF metal 3D printing.
The team reported that its system can effectively and accurately evaluate laser melt lines — the tracks where the laser liquifies metal powder in LPBF printing — by scattering acoustic energy from melt lines, voids, and surface features that can be quickly detected.
The team validated its findings using optical microscopy and x-ray computed tomography (CT).
An LLNL team demonstrated that a diagnostic incorporating surface acoustic waves — generated by laser-based ultrasound — could effectively and accurately evaluate laser melt lines and find defects in laser powder bed fusion metal 3D printing by scattering acoustic energy from melt lines, voids, and surface features that can be quickly detected. Courtesy of David Stobbe/LLNL.
“We hope that this work demonstrates the potential for an all-optical ultrasound system capable of rapid, on-demand, in situ characterization of LPBF processes and powders,” said LLNL engineer and principal investigator David Stobbe. “The demonstrated laser-based ultrasound, surface acoustic wave system showed excellent sensitivity to surface and near-surface features, including breaks in the LPBF melt line, metal surface splatter, and subsurface air voids.”
SAWs have historically been used to characterize surface and near-surface features such as cracks, pits, and welds in engineering materials, and they are used in geology — at a much larger length scale — for detecting subterranean features such as caves. Due to their surface and near-surface sensitivity, SAWs are well-suited for characterizing melt lines in LPBF printing, according to the researchers.
The LLNL researchers produced laser melted lines using a fiber laser directed into a vacuum chamber and produced samples of titanium alloy for analysis with 100-, 150-, and 350-W powered lasers. Next, they developed a method for producing and detecting surface acoustic waves, using a pulsed laser to generate ultrasound and measured the displacement with a photorefractive laser interferometer.
The researchers also performed simulations to inform the experimental measurements and to help interpret the results. They simulated and measured the displacement from the pulsed laser and showed scattering from the melt line, as well as breaks in the melt line, metal splatter adjacent to the melt line, and subsurface air voids under the melt line. The team measured the same features experimentally and observed excellent agreement between simulation and experiment.
The results from laser-based ultrasound (LBU) experiments were validated with optical microscopy for the surface features and x-ray CT for the subsurface features. Researchers reported that in comparison with x-ray CT, the LBU system is “better posed to perform real-time inspection and can acquire and process data at a faster rate.”
“Utilizing the laser-based ultrasound significantly shortened the time for subsurface void detection compared to conventional x-ray CT from days to minutes,” said LLNL engineer and lead author Kathryn Harke. “While more development would need to be done before implementation of this diagnostic for in-process monitoring, our team is excited by these initial findings.”
The method is well suited for in situ implementation in LPBF printing, though there are limits on the size and depth of detectable voids, and in situ monitoring or post-build inspection would require further development, the researchers said.
“A system like this may find use for rapidly qualifying new LPBF machines and in-service machines after changes to metal powder feedstock or modifications to the melt laser power or scan speed,” Stobbe added.
The work was funded by the Laboratory Directed Research and Development program.
The research was published in Scientific Reports (www.doi.org/10.1038/s41598-022-07261-w).