Additive manufacturing (AM) processes may introduce large residual stresses in the as-built part, in particular the laser powder bed fusion process (L-PBF). The residual stress state is an inherent consequence of the heterogeneous heating and subsequent cooling during the process. L-PBF has become renowned for its “free complexity” and rapid prototyping capabilities. However, it is vital to ensure shape stability after the component is removed from the build plate, which can be problematic due to the residual stress inducing nature of this manufacturing process. Residual stresses can be analyzed via many different characterization routes (e.g. X-ray and neutron diffraction, hole drilling, etc.), both quantitatively and qualitatively. From an industrial perspective, most of these techniques are either prohibitively expensive, complex or too slow to be implementable during the early prototyping stages of AM manufacturing.

In this work a deformation based method employing a specific geometry, a so called “keyhole”-geometry, has been investigated to qualitatively evaluate the effect of different stress relief annealing routes with respect to macroscopic part deformation, mechanical properties and microstructure. Previous published work has focused on structures with open geometry, commonly referred to as bridge-like structures where the deformation required for analysis occurs during removal from the build plate. The proposed keyhole-geometry can be removed from the build plate without releasing the residual stresses required for subsequent measurement, which enables bulk manufacturing on single build plates, prior to removal and stress relief annealing. 

Two L-PBF manufactured austenitic stainless steel alloys were studied, 316L and 21-6-9. Tensile specimen blanks were manufactured and the subsequent heat treatments were carried out in pairs of keyhole and tensile blank. Both a contact (micrometer measurement), and a non-contact (optical profilometry) method were employed to measure the residual stress induced deformation in the keyholes. The annealing heat treatment matrix was iteratively expanded with input from the deformation analysis to find the lowest temperature at which approximately zero deformation remained after opening the structure via wire electrical discharge machining. The lowest allowable annealing temperature was sought after to minimize strength loss. 

After stress relief annealing at 900 ℃ for 1 hour, the 316L keyhole-geometry was considered shape stable. The lateral micrometer measurement yielded a length change of 1 µm, and a radius of 140 m (over the 22 mm top surface) was assigned from curve fitting the top surface height profiles. The complementary microstructural characterization revealed that this temperature corresponded to where the last remains of the cellular sub-grain structures disappears. Tensile testing showed that the specimen subjected to the 900 ℃ heat treatment had a marked reduction in yield stress (YS) compared to that of the as-built: 540 MPa → 402 MPa, whereas ultimate tensile strength (UTS) only reduced slightly: 595 MPa → 570 MPa. The ductility (4D elongation) was found to be ~13 % higher for the specimen heat treated at 900 ℃ than that of the as-built specimen, 76% and 67% respectively. 

For alloy 21-6-9 the residual stress induced deformation minimum (zero measurable deformation) was found after stress relief heat treatment at 850 ℃ for 1 hour. Slight changes in the microstructure were observable through light optical microscopy when comparing the different heat treatment temperatures. The characteristic sub-grain features associated with alloy 316L were not verified for alloy 21-6-9. Similar to the results for 316L, UTS was slightly lower for the tensile specimen subjected to the heat treatment temperature required for shape stability (850 ℃) compared to the as-built specimen: 810 MPa → 775 MPa. The measured ductility (4D elongation) was found to be approximately equal for the as-built (47%), and heat treated (48%) specimen. As-built material exhibited a YS of 640 MPa while the heat treated specimen had a YS of 540 MPa. For alloy 21-6-9, the lateral micrometer deformation measurements were compared with stress relaxation testing performed at 600 ℃, 700℃ and 800 ℃. Stress relaxation results were in good agreement with the results from the lateral deformation measurements. 

The study showed that for both steel alloys, the keyhole method could be successfully employed to rapidly find a suitable stress relief heat treatment route when shape stability is vital.