LLNL Researchers Discover Real Cause of Spatter During Metal 3D Printing Processes

So much about 3D printing is still a mystery – especially when things go wrong. Anyone who works with 3D printing is well-acquainted with the fact that the technology isn’t perfect, and that there are going to be errors, failed prints, and all manner of frustrations to deal with along the road to completing a project. Sometimes, it’s unclear where those errors and flaws come from, but when it comes to metal 3D printing in particular, finding the source of those defects is a matter of utmost importance. Defects in metal parts tend to mean compromised function and safety issues, so it’s critical to figure out where they’re coming from so they don’t happen again.

Researchers at Lawrence Livermore National Laboratory (LLNL) have spent a lot of time studying the numerous issues that plague metal 3D printing, and recently they made an important discovery related to what’s commonly known as spatter. Spatter happens frequently during powder bed fusion additive manufacturing processes, and it involves tiny particles of liquid metal being ejected from the laser’s path. This can result in contamination of the powder bed and issues such as porosity, roughness and lack of adhesion in the finished parts.

It has been believed that spatter is caused by the laser’s recoil pressure, but the LLNL researchers have discovered that it’s actually due to the entrainment of metal particles by an ambient gas flow. They came to this conclusion after combining high-speed imaging of melt pool dynamics with high-resolution computer simulations.

“People have been assuming that recoil pressure leads to spatter because that’s what the laser welding community has seen,” said Sonny Ly, an LLNL physicist. “We imaged right at the melt pool and you could see particles ejected right from the pool due to recoil, but a majority of particles are swept away and entrained by the gas flow. The entrained particles can go back into the laser beam and are melted, leading to a more dominant form of spatter.”

The video images were taken with three different types of cameras, including a sensor capable of taking up to 10 million frames per second. According to LLNL engineer Gabe Guss, this high-quality imaging allowed the researchers to see not only the wave of pressure created by the laser and the counter-drop of liquid metal, but the gas flow above the powder bed that sucked in the particles, where they either melted or sailed through the laser.

“It turns out only about 15 percent of the ejections of molten particles are caused by splashing in the melt pool, which was the assumed mechanism — the rest is primarily cold particles passing through the laser beam above the melt pool and some other factors,” Guss said. “It’s surprising because when one watches commercial printers, you see the hot ejections and they look like they come from simply outward gas pressure, not the inward entrainment effect.”

The images were compared to high-fidelity simulations that had previously been validated for other additive manufacturing applications, and the researchers discovered that the direction of the spatter was influenced by the incline of the melt pool.

“These cameras can’t show in detail what’s happening below the surface of the melt pool,” said Saad Khairallah, an LLNL computational engineer/physicist who ran the simulations. “The simulations showed a difference in the morphology of the melt pool beneath the laser spot, which allowed us to interpret the experimental observations. This is an example where simulations complement experiments and become a key component in a science story.”

With this research, we can now have a better understanding of powder bed fusion 3D printing, and existing flow models can be improved. Moreover, now that we know what causes spatter, the effects can be better mitigated. The research was published in a paper entitled “Metal vapor micro-jet controls material redistribution in laser powder bed fusion additive manufacturing,” which you can access here. Authors of the paper include Sonny Ly, Alexander M. Rubenchik, Saad A. Khairallah, Gabe Guss and Manyalibo J. Matthews. Discuss in the LLNL forum at 3DPB.com.