3D Printed Part Annealing
Fused deposition modeling, known as 3D printing, has an abundance of benefits in being able to quickly make a variety of accurate models cheaply and quickly, but effective 3D printing has some constraints. In practical applications, a large constraint is the directional weakness that comes from fusing layers of plastic together. These fused layers are weak joints compared to the homogenous plastic found in injected-molded models, and will often result in brittle fracture under tensile stress. If the model is undergoing stress in only one direction, the model can sometimes be printed in an orientation to accommodate the stress, but that can also lead to increased printing time and more material use if supports are required.
One method to achieve 3D printed objects with uniform strength and ductility is through annealing the parts to "reflow" the plastic to create more bonds across layer lines. Due to the low glass transition temperature of typical plastic filaments, this can be done in a home oven. The main issue with this method is deforming the model and losing part accuracy, so to try to alleviate this, I encased the model in baking soda, then packed the backing soda around the model, and weighed it down.
In the video, you can see three of the same model all printed in PETG plastic with 100% infill. The first model was printed "flat," creating large layers bonded together, which allowed to model to undergo plastic deformation and not fracture. The next two models were printed "upright," meaning they had a large number of very small layers with little surface area to bond together. As expected, the model fractures almost immediately under stress, and the fracture is directly between layers. However, the last model was annealed in the oven, and underwent nearly the same plastic deformation as the first model, suggesting more bonds had been created and strength had been increased. Unfortunately, the model did shrink a bit in the X and Z directions and grow a bit in the Y direction, but this change was less than 2%. In the future, I will make more models to gain a larger test set and better determine how the dimensions change during annealing. I would also like to anneal models with more defined details to determine if the baking soda packing method can fully protect models during annealing.