1. Introduction
Additive manufacturing is currently one of the most popular topics of different research papers [
1]. There are many publications connected with different types of analysis. Among these, the authors consider material properties, the numerical analysis of additively manufactured parts, and case studies of specific parts. Laser powder bed fusion (L-PBF) technologies are characterized by using an energy source to melt metallic powder grains [
2] where the volume of the material is constituted on melt pools which create a specific, layered structure [
3,
4]. Additively manufactured parts were used for the authors’ own research connected with static, dynamic, and fatigue behavior [
5,
6,
7,
8] with additional process modification [
9] and heat treatment [
10]. Regardless of the process parameters selected, different material behavior was observed during the testing of monolithic parts and thin-walled or lattice-structured elements. The mentioned phenomenon was observed by Sienkiewicz et al. [
2], where geometrical accuracies and microstructures revealed some weaknesses of the selected manufacturing process, such as the deviation of the dimensions of lattice struts measured on the frontal plane. Additionally, the authors of the mentioned work revealed the presence of imperfections such as porosity, voids, and unmelted powder grains. These kinds of issues must be taken into account during finite element analysis (FEA).
One of the most popular approaches is using the plastic strain value adopted in FEA; a good example was used by Kucewicz et al. [
11], where the authors implemented a meshless method of modeling the orthotropic properties of the material. This method represents a kind of compromise between microscopic and macroscopic levels. In other research papers [
12], it was stated that increasing the value of relative density causes a growing sensitivity of the structure to strain rate effects during the testing of thin-walled honeycomb structures made of Ti6Al4V alloy. In some research papers, it was observed that the honeycomb structures are characterized by brittle fracture in dynamic tests, which resulted in significantly worse energy-consuming parameters for that kind of structure [
13]. Different material behavior after the L-PBF processing of thin-walled parts has to be taken into account by designers, especially when constructing parts dedicated to military, aircraft, and medical applications [
14,
15,
16,
17,
18,
19].
Design assumptions dedicated to selectively laser melted (SLM) parts state that it is common to design self-supporting structures to minimize the use of support structures [
20,
21]. In many cases, the use of bridge structures, which are not self-supporting, is recommended. That type of structure is used two connect points without any support structure from below. These kinds of short bridges can be obtained without any support structures, and it is assumed they will be used to save material and manufacturing process time.
Geometrical complexity significantly affects heat transfer into the material volume during the manufacturing process. The phenomenon of heat transfer is strictly related to molten pool behavior during the L-PBF process, which in turns affects void generation. It has been described by Gusarov et al. [
22], where authors point to transfer of the laser radiation in powder as being responsible for the production of volumetric heat. Created in this way, a melt pool’s flow is driven by its surface tension. Movement of the melt pool along the exposure line allows for its contact with the substrate in its central part, and with loose powder particles in its lateral part (where surface energy is reduced).
On the other hand, Liu et al. [
23] described the influence of powder particles on heat transmission, where the local overheating of powder particles elicits material evaporation and generates recoil pressure on the molten pool. Further analysis of material behavior during L-PBF was performed by Matthews et al. [
24], where the authors analyzed melt pool dynamics and vapor flow using the Lagrangian-motion component; they determined it was responsible for the adiabatic movement of the material in response to forces present in the considered volume of the melt pool.
During our preliminary research connected with the numerical analysis of honeycomb structures (not yet published), it was observed that there were significant differences between experimental results and numerical analysis. The main issue was connected with significant porosity growth in the area of the thin wall visible in
Figure 1.
The presence of significant differences between the properties of monolithic and thin-walled parts obtained using the same material and additive manufacturing technology is considered in this paper. It determines how the manufacturing process of thin-walled parts affects the microstructure and porosity of that type of element. Additionally, five different process parameter groups were tested to analyze how to minimize the negative influence of manufacturing thin-walled structures on parts’ structural properties. The influence of the parameters used for thin-walled structures on their surface roughness was also taken into account.
2. Materials and Methods
2.1. Material
The material used for sample manufacturing was gas-atomized 316L stainless steel powder characterized by a spherical shape supplied by SLM Solutions AG company (SLM Solutions, Lubeck, Germany) with the grain size in the range of 15–45 µm and a flowability of 14.6 s/50 g. The chemical composition of the material is shown in
Table 1.
The material properties of SLM-processed 316L steel (based on SLM Solutions’ Data Sheet) using 50 μm layer thickness are shown in
Table 2.
The powder particles’ size distribution cumulated mass had values as follows: D10 = 18.22 μm, D50 = 30.50 μm, D90 = 55.87 μm. Scanning electron microscopy (SEM) images of the powder grains are shown in
Figure 2.
2.2. Additive Manufacturing Process
A selective laser melting process was performed using an SLM 280HL machine (SLM Solutions, Lubeck, Germany) in an argon atmosphere. For sample manufacturing, five process parameters were used and are shown in
Table 2. Abbreviations used in
Table 3: L
p—laser power, ev—exposure velocity, h
d—hatching distance, l
t—layer thickness, ρ
E—energy density. Energy density is the main process parameter that creates a relationship between laser power, exposure velocity, hatching distance, and layer thickness.
These parameters were selected based on previous research experience [
25] where significant differences during hatching distance changes were observed. Additive manufacturing of thin-walled structures is characterized by a higher cooling rate than occurs in monolithic parts, so it is very important to keep stable and slight change which can be controlled to a certain extent. That kind of characteristic was registered when changing laser power. For this research, five parameter groups were used and only laser power was changed to reach a stepwise movement in energy density equal to 10 J/mm
3. That approach allowed the observation of material behavior after using different values of laser power.
2.3. Material Properties Analysis
Powder particles analyses were performed using a Jeol JSM-6610 (Jeol, Tokyo, Japan). The evaluation of porosity, microstructural observations, and roughness analysis using 3D laser measurement was performed using an Olympus LEXT 4100 confocal microscope (Olympus Corporation, Tokyo, Japan,). For porosity analysis, MountainsMap 6 software (version 7.2, Digital Surf, Besançon, France) was used. To reveal the microstructure, an “acetic glyceregia” solution heated to 40 °C was used as the etchant (6 mL HCl, 4 mL HNO3, 4 mL CH3COOH, and 0.2 mL glycerol). The etching time was 5 s.
For each manufactured part, Vickers hardness distribution measurements were taken using a Struers DuraScan 70 (Struers, Copenhagen, Denmark) hardness tester. All measurements were conducted in the monolithic part of the samples (shown in
Figure 3). The samples’ geometry was designed in this way in order to analyze porosity in the monolithic and thin-walled parts and to check the material condition in the areas of the “bridges”. Their dimensions were as follows: 100 mm length, 10 mm width, and 12 mm height. Thin-walled parts (area “A” in
Figure 3) were characterized by a thickness of 0.75 mm and 1 mm spacing. The monolithic part of the samples in the middle (area “B” in
Figure 3) had dimensions of: 10 mm length, 10 mm width, and 12 mm height.
To reduce surface roughness, a PK 1200 E sandblaster with silica sand (Virgo, Siedlce, Poland) was also used. For the additional surface treatment of the additively manufactured (AM) parts, a Struers LectroPol (Struers, Copenhagen, Denmark) was used for electropolishing.
For surface-treated samples, each surface was electropolished three times in A3 Struers electrolyte using 35 V voltage, 5 A amperage, and 40 s polishing time.