The buildup of microscopic residual stresses that occur during the process of 3D printing metal parts can lead to deformation and even cracking in the part, complicating the printing process and resulting in defects that cause damage and part failure.
To develop a strategy for mitigating these tiny stresses, Lawrence Livermore National Laboratory (LLNL) researchers have combined synchrotron X-ray diffraction with computer modeling to better understand the link between residual stresses and the mechanical properties of 3D-printed 316L stainless steel. The study was published in the journal Nature Communications on Sept. 25.
Using Argonne National Laboratory’s Advanced Photon Source and LLNL’s advanced electron microscopy facility, researchers looked at how “marine grade” stainless steel parts built through the laser powder-bed fusion metal 3D-printing process strained under tension and compression at a grain-level scale, comparing the experiments to computer models.
They found that microscale stresses have a profound impact on the yielding and work-hardening behavior (the ability of a metal to deform elastically and plastically to accommodate applied stresses) of 3D-printed 316L stainless steels, and that residual stresses cause differences in the material’s responses to tension and compression, an important factor for load-bearing applications and component designs.
“Right now, the assumption is that the yield stresses in compression or tension are the same, but what we found is they’re actually not the same — the build material has intrinsic residual stress and that causes the asymmetry,” explained LLNL staff scientist and principal investigator Morris Wang. “It’s about a 10 percent difference, which is very important because when you design materials and components you actually have to take that into account.”
To alleviate residual stress, engineers typically release it by exposing the printed parts to heat after a build in a process called thermal annealing. The LLNL team used scanning and transmission electron microscopes to examine the microstructure of the steel at different length scales before and after heat treatment, noting that while low-temperature annealing does not change the microstructure, most of the tension/compression asymmetry went away after annealing.
“We carefully chose the heat treatment temperature to reduce the strength anisotropy down to a minimum without altering the microstructure,” said co-lead author Wen Chen, a former LLNL postdoctoral research scientist who is now an assistant professor of mechanical engineering at the University of Massachusetts at Amherst. “The as-processed microstructure makes this material outstanding — we want to keep it while removing unwanted process-induced artifacts by proper heat treatment. Whereas low-temperature heat treatment can help reduce residual stresses, it may not fully eliminate them.”
Wang, co-lead author and LLNL staff scientist Thomas Voisin and fellow Lab co-author Jean-Baptiste Florien used the synchrotron at Argonne to perform the high-energy X-ray diffraction, allowing the team to probe large samples of the 3D-printed stainless steel in situ, providing them with unique insights into the material’s mechanical deformation. Along with collaborators at the Georgia Institute of Technology, they also developed a crystal plasticity finite element computer model to simulate the impact of the stresses on mechanical behavior and explain the physics behind the differences seen between tension and compression. The models were matched to experimental results of lattice strain responses and found to be consistent.
“Looking at the spacing between the atomic planes using the synchrotron gave us information on how much residual stress and strain we had in the material, and the very interesting part is the fact that the modeling that was done at Georgia Tech actually could very accurately capture those features of the 3D-printed materials and explain this asymmetry,” Voisin said. “Besides showing this asymmetry, we’re also showing that we can model it, and that means basically we can better control it. This provided a good example of why we need to keep collaborating between experiments and simulations to really go in-depth into understanding the material behavior.”
The ability to accurately reproduce experiments through computer modeling will allow future engineers or designers working with stainless steel to build the models into their design criteria and examine the specific mechanisms that cause deformations, the researchers said. However, more studies will be needed to find out how to further minimize and control residual stress, including testing the models with other common materials such as titanium-64 and aluminum alloys, they added.
LLNL’s Center for Engineered Materials and Manufacturing Director Chris Spadaccini, and Yin Zhang, Ting Zhu and David McDowell of Georgia Tech co-authored the study. The work was funded by the Dynamic Material Properties Campaign in the National Nuclear Security Administration’s Office of Experimental Science.