LLNL team develops real-time diagnostic for Liquid Metal Jetting 3D printing

As 3D printing continues to grow and evolve, diagnostics capable of monitoring builds in real-time have become essential tools for producing quality parts, particularly in emerging printing technologies such as Liquid Metal Jetting (LMJ).

In LMJ, tiny molten metal droplets are ejected from a nozzle at high speeds to 3D print a part in layers, similar to inkjet printers on the consumer market. Unlike laser-based metal 3D printing processes, the technology doesn’t require metal powder, which can be hazardous to handle and wasteful of material. Diagnostics used today to ensure high-quality LMJ prints largely rely on high-speed videography, which requires expensive equipment, can be difficult to set up and generates large amounts of data. While adequate for evaluating a few seconds of a test printing, it isn’t a feasible solution for longer builds.

Researchers at Lawrence Livermore National Laboratory (LLNL) are attempting to solve the problem with a new diagnostic tool that can determine the quality of metal droplets and monitor LMJ prints in real time. The approach uses low frequency, electromagnetic near-field detection to capture metal droplet dynamics that, when combined with simulation, provides information on droplet features based on signal amplitude and phase alone.

Researchers said the ability to characterize the droplets using just one parameter significantly reduces the amount of necessary data, making processing and feedback of longer-term LMJ prints possible. The work was published online by the Journal of Applied Physics, where editors selected it as a “Featured Article.”

“Our results demonstrate that in situ monitoring of Liquid Metal Jetting is possible with millimeter-wave detection methods,” said lead author and LLNL engineer Tammy Chang. “This is exciting because it means we could replace computationally expensive high-speed, high-resolution optical diagnostics to enable real-time performance evaluation and feedback control, to ensure high quality printed metal parts.”

With the new technique, they discovered large-scale trends such as print variation and anomalies at the nozzle, as well as micro-level attributes about the droplets, including size, position and dynamics. The team used electromagnetic simulations to model droplet properties, allowing the team to understand the physics of the electromagnetic scattering and how variations in the magnitude and phase of the detected signals affected droplet features, Chang said.

The result of the research is a “compact and non-invasive” diagnostic able to distinguish droplets in LMJ machines at higher print rates than possible before, as well as the capability to detect additional features of the print system, researchers said. The ability to capture droplet properties based on one parameter alone ultimately shows promise for feedback systems in which rapid, real-time processing can be used to adjust print settings and guarantee part quality, they concluded.

“Getting a clean ejection of a single drop that falls straight down is key to achieving good print quality,” said LLNL engineer Andy Pascall, a co-author on the paper. “High-speed videography works well in a lab-scale environment where we are testing new print parameters, but will never work in production. The millimeter-wave diagnostic is a huge improvement because it can be integrated into the printer, doesn’t require optical access and provides data that can be analyzed in real time to determine if high-quality droplets are being generated. This type of diagnostic will be very useful in a production environment.”

In the future, researchers said signal processing techniques could be used to correlate optical and millimeter-wave measurements and predict droplet properties based on millimeter-wave results alone. Processing efforts developed by the LLNL researchers and their collaborators are currently under review.

The work is part of a three-year Laboratory Directed Research and Development (LDRD) project aimed at developing acoustic and electromagnetic monitoring approaches for metal additive manufacturing. It supports LLNL’s droplet-on-demand Liquid Metal Jetting work led by Pascall and physicist Jason Jeffries.

Lab staff scientist and Nondestructive Evaluation Group Leader Joe Tringe devised the original foundational idea for the diagnostic, as well as the broader LDRD Exploratory Research project for acoustic and electromagnetic diagnostics for metal additive manufacturing.

Co-authors on the paper included LLNL scientists and engineers Saptarshi Mukherjee, Nicholas Watkins, Edward Benavidez, Abigail Gilmore and principal investigator David Stobbe.