The effect of varying specimens’ printing angles to the bed surface on the tensile strength of 3D-printed 17-4PH stainless-steels via metal FFF additive manufacturing

We fabricated 17-4PH stainless-steel specimens via Fused Filament Fabrication (metal FFF). Then, we studied the effect of varying specimens’ printing angles to the bed surface (part orientation) by analyzing the tensile test and scanning electron microscopy results. We found anisotropy, where specimens printed at 90° exhibited the lowest tensile-stress-at-maximum-load of 440.15 MPa and elongation-at-break of 0.83%, compared to specimens printed at 0° of 947.26 MPa and 2.98%, respectively. We recommend printing angles from 0° to 10° to achieve optimum tensile strengths. This study is significant in ensuring quality for the deployment and scale-up of spare-part production via metal FFF.


Introduction
Additive manufacturing (AM) has the advantage of giving users the ability to create bespoke designs of both different geometric and different materials. AM applications cover prototyping, tooling, performance improvement, personalization, spare-part, maintenance, and repair. At Dubai Electricity and Water Authority (DEWA), we are interested in the 3D printing of stainless-steel metal material for the deployment and scale-up of spare-part application, with an example use-case of replacement of obsolete metal parts.
Some of the additive manufacturing technologies for printing metal parts are powder bed fusion (PBF), directed energy deposition (DED), and fused filament fabrication (FFF). [1] The PBF has the advantage of fine printing resolution but relatively slow deposition. The DED has the advantage of relatively fast printing (high build rate) but achieves only a near-net shape (coarse printing resolution). The FFF process has a balanced performance with a relatively good resolution, an acceptable printing time, and a reasonably low operation cost. FFF is suitable for our specific use-case for the spare-part application. While the FFF additive manufacturing was originally applied for pure polymers (without filler), the literature review showed the prospect of adding carbon composite, ceramic, or metal fillers at high concentrations. [1,2] Herein, we defined the term "metal FFF" as the additive manufacturing of metals utilizing the FFF process. Other similar terms reported in the literature were the fused deposition of metals (FDMet), multiphase jet solidification (MJS), material extrusion additive manufacturing (MEAM), or atomic diffusion additive manufacturing (ADAM) process. Throughout this paper, we used the "metal FFF" term for consistency and to avoid confusion.
Gonzalez-Gutierrez et al. [10] investigated the extrusion temperature, flow rate multipliers, and speed multipliers for FFF of 17-4PH stainless-steel. Proprietary binder consisted of thermoplastic elastomer, polyolefin-based backbone polymer, and compatibilizer were used. [10] Gonzalez-Gutierrez et al. [10] observed an optimized extrusion temperature of 260 °C, flow rate multiplier of 200%, and printing rate multiplier of 100% in their case. Gonzalez-Gutierrez et al. also investigated the tensile strength of 17-4PH stainless-steel using proprietary binder based on grafted polyolefin and thermoplastic elastomer material. [11] The average Young's modulus was 196 GPa, the average maximum stress was 696 MPa, and the strain at break was 4%. [11] Godec et al. optimized the extrusion temperature, flow rate multiplier, and layer thickness of FFF process parameters of 17-4PH stainless-steel. [12] The filament composition was 55 vol% of 17-4PH stainless-steel powder in multicomponent binder systems. [12] Galati et al. investigated the density, roughness, and accuracy of 3D-printed 17-4PH stainless-steel using metal FFF process utilizing a commercial metal 3D printer system from Markforged™ company. [13] The advantage of the commercial metal 3D printer system is the user-friendliness of the process, with ready-to-use proprietary binders, ready-touse proprietary metal filament formulation, and ready-to-use proprietary multi-step sintering temperature recipe.
Our motivation in performing this study is as follows. We have printed metal spare parts for DEWA operations on a pilot basis, using the same commercial metal 3D printer mentioned in reference. [13] As we saw the benefits to the DEWA operations, to scale-up the printing for more quantities and more metal part's designs, we need to establish stringent quality checks. However, to our knowledge, there is only one reported paper [13] utilizing this commercial metal 3D printer system. Therefore, we studied the effect of the printing process on the mechanical performance of the 3D-printed part. Specifically, we investigated the tensile mechanical strength property of 3D-printed 17-4PH stainless-steel using the metal FFF process by varying the specimen printing angles.
The novelty of this study is that, to our knowledge, we reported for the first time the anisotropic behavior of 17-PH stainless-steel specimens manufactured via metal FFF additive manufacturing. This study's impact is significant in ensuring the 3D printing quality for the deployment and scale-up of spare-part production via metal FFF additive manufacturing.

Methodology Materials
We used a Metal-X 3D printer system (Markforged™, USA) to conduct this work. Commercially available 17-4PH stainless-steel filaments (Markforged™, USA) with a diameter of 1.75 mm were used to print all the specimens. The filaments consist of stainless-steel material with a proprietary binder. During storage, the 17-PH stainless-steel filament spools were stored vertically inside packaging boxes as per manufacturer recommendation and due to the spool material's fragility at room temperature.

3D printing processes
Dog-bone-shaped tensile test specimens were used. The geometry of the specimens is shown in The printing processes consist of four phases: fused filament fabrication followed by washing, drying, and sintering. The schematic image can be found in Fig. S2 (Electronic Supplementary Material; Online Resource). The temperature parameters' values (chamber temperature, print bed temperature, and sintering temperatures) were not provided due to the proprietary or non-open-source characteristics of this equipment. The chamber temperature was warmer than room temperature in order for the spool material to be relatively flexible. The print bed consists of a temperature-controlled and vacuum-capable stage coupled with a disposable print sheet.
We investigated ten different print orientations, varying from 0° to 90° with an increment of 10° for each print orientation angle. The Y 0° specimen was printed with the configuration of 0° printing angle to the print bed in X, Y, and Z axes. Throughout this work, the printing angle in the X and Z axes remained zero constantly, while the printing angle orientation in Y-axis was varied. An example of the specimen name nomenclature: when the printing angle was set at 10° rotation with respect to Y-axis, the specimen name would be Y 10°. The details of the printed specimens were summarized in TABLE S1 (Electronic Supplementary Material; Online Resource). Figure 1 shows the illustration of varying the printing angle. For each printing angle, four specimens were fabricated for tensile testing purposes, and one specimen was fabricated for the investigation of the morphology.

Mechanical tests
The tensile test was carried out at room temperature using the Instron 3367 apparatus with Bluehill LE 3.77 software. The load cell was rated at a maximum of 30 kN. The tensile rate was set to 10 mm/min. The desired outputs obtained were: (i) elongation-at-break, (iii) tensile-stress-at-maximum-load (iii) tensile-stress-at-yield, and (iv) Young's modulus. Four specimens from each batch were used, in which their obtained values were averaged for each printing angle.

Mass-reduction measurements
The metal FFF technology has three states during the process. The "green part" is the state after the layered deposition, while the "brown part" is the state after the part goes through the debinding or washing stage. Finally, the "silver part" is the final state after the part is sintered and has metallic characteristics. 14 The mass-reduction-after-washing measurement was conducted by calculating the difference in the mass of the specimens before washing (green mass) and after washing (brown mass), refer to Eq. 1. The relative density was calculated from the mass-reduction-after-washing data, refer to Eq. 2. (1)

Imaging (top-view) of as-sintered specimens
Scanning electron microscopy (SEM) was conducted using Tescan VEGA XM SEM to examine the morphology of the layers of printed dog-bone specimens. The operation was conducted under the following conditions: 50 × magnification, 20 kV accelerating voltage, secondary electron mode, and high vacuum.

Imaging (cross-section) of as-sintered specimens
The internal structure of the printed specimens before tensile testing was observed by cutting the specimen for a brittle fracture. Obtaining a brittle fracture is necessary to preserve the internal geometry. [15] Specimens were held for 1 min in liquid nitrogen until they were cooled to -196 °C. BS Charpy Impact equipment was used to break as-sintered specimens in half. After cutting, the morphology of the fracture was imaged using the FEI-ThermoFischer Quattro S SEM. The operation was conducted under the following conditions: 80 × and 500 × magnifications, 20 kV accelerating voltage, secondary electron mode, and high vacuum.

Imaging (cross-section) of specimens after tensile testing
The fracture cross-sections after the tensile test were investigated using Tescan VEGA XM SEM to observe the internal pattern and the fracture morphology. The operation was conducted under the following conditions: 50 × and 500 × magnifications, 20 kV accelerating voltage, secondary electron mode, and high vacuum.

Results and discussion
The anisotropy in the mechanical property of polymer (without additive) using FFF is well-known; for example, Acrylonitrile Butadiene Styrene (ABS) 3D-printed part has higher strength when printed horizontally than vertically due to relatively weak adhesion strength between layers. [15] We are interested in answering the questions: whether metal FFF (e.g., 17-4PH stainless-steel) has similar anisotropy behavior compared to polymer FFF (e.g., ABS) and whether the trend is dominated by the filament material or by the FFF characteristic.
The morphology of the specimens was checked using SEM; see the location in Fig. S3 (Electronic Supplementary Material; Online Resource). At 50 times magnification, the top surfaces of the middle-section of dog-bone-shaped printed specimens were observed, as shown in Fig. 2. It can be seen that varying the printing angle changed the appearance of the 3D-printed specimens. At a print angle of 0 o (Y 0°) , there was a visible difference in the appearance of the surface, which attributed to the toolpath functionality of the printing software at that particular angle; see Fig. S4 (Electronic Supplementary Material; Online Resource). As the printing angle was increased, the produced specimens had perpendicularly defined deposited layers with the height of 0.72,0.37,0.25, 0.19, 0.17, 0.14, 0.13, 0.13, and 0.12 mm for Y 10°, Y 20°, Y 30°, Y 40°, Y 50°, Y 60°, Y 70°, Y 80°, and Y 90°, respectively, as shown in Fig. 2b to j. Figure 3 provides the graphical comparison of tensile properties. The histograms showed the average value, while the whiskers showed the maximum and minimum value for each batch, as seen in Fig. 3b to e. Tabulated data can be found in TABLE S2 (Electronic Supplementary Material; Online Resource). There was a noticeable difference in the tensile properties. As an example [see Fig. 3(a)], the tensile stress-strain curve shows that Y 0° and Y 10° recorded the highest values of tensile-stressat-maximum-load, which were 947.26 and 949.87 MPa, respectively. However, larger printing angles such as 60°, 70°, 80°, and 90° exhibited significantly lower stress values of 418.12, 415.28, 488.88, and 440.15 MPa, respectively. Specimen Y 0° and Y 10° exhibited the highest elongation-at-break of approximately 2.98% and 2.2%, respectively, however specimen Y 60°, Y 70°, Y 80°, Y 90° showed relatively low elongation-at-break of 0.87%, 0.60%, 1.07%, and 0.83%, respectively. The relative standard deviation of tensile-stress-at-maximum-load was 32.77 while being 54.82 for elongation-at-break. This explained the extent of high anisotropy of the tensile-stress-at-maximum-load and elongation-at-break parameters of the specimens. The FFF characteristic dominated the tensile-stress-at-maximum-load and elongation-at-break. Several factors affect the quality of the fabricated specimens using material extrusion FFF additive manufacturing: extrusion temperature, infill rate, material flow, cooling rate post printing, and build orientation. [16][17][18][19] These factors influence the geometric precision of the 3D-printed metal specimens, thus affecting the mechanical properties. [16][17][18][19] We observed that the tensile-stress-at-yield and Young's modulus had no anisotropy trends. Tensile stress at yield and Young's modulus of all specimens showed close readings, with small values of the relative standard deviation of 16.13 and 9.15, respectively.  The morphology of the specimens at 5,000 times magnification showed the presence of microscopic pores in all specimens of Y 0° to Y 90°, see Fig. S5 (Electronic Supplementary Material; Online Resource). The diameter of microscopic pores ranged from 1.2 μm to 3.4 μm. It could be attributed to the debinding of the binder that induced pores. [1,7] Fig. S6 (Electronic Supplementary Material; Online Resource) shows the average range of mass-reduction-after-washing varied between 4.13% to 4.25% among the specimens; i.e., the relative density varied between 95.75% to 95.87%. We found little or no correlation between the mass-reduction-after-washing and the tensile properties (elongation-at-break, tensile-stress-at-maximum-load, tensilestress-at-yield, Young's modulus). It indicated that in this case, other FFF printing parameter(s) had a more substantial influence on the tensile properties, which was the air-gap parameter as explained in the next paragraph. Figure S7 (Electronic Supplementary Material; Online Resource) shows the location of the cross-section. The crosssection of the specimens was checked using SEM. Figure 4 shows the cross-section view of the fracture (without undergoing tensile test) of the specimens revealed triangular-shaped or diamond-shaped openings (air gaps) with the approximate size of ~ 50 μm to ~ 100 μm. These air gaps were attributed to the underfill between deposited toolpath lines of the walls. We observed that different printing angles produced different airgap directions with respect to the tensile test loading direction. Different printing angles also produced different shapes of air gaps. At low printing angles, such as 0° and 10°, the air gaps have a triangular shape, spatially separated, and having a relatively small area, see Fig. 4a and b. At intermediate printing angles, such as 20° to 50°, the air gaps have a diamond shape, spatially separated, and having a relatively intermediate area, see Fig. 4c and f. At large printing angles, such as 60° to 90°, the air gaps have a line or rectangular shape, spatially connected, and having a relatively large area, see Fig. 4g and j. The effective cross-section of the specimens decreased with increasing printing angles due to the increase in the area of the air gaps. In addition, the air gaps provided sites for crack initiation, which were more prominent for specimens printed at high angles, such as 60° to 90°. It explained the specimens printed at low printing angles had better tensile-stress-at-maximum-load than specimens printed at high printing angles.
We propose that the specimens' tensile properties can be improved in the future if the air gaps are covered. Reference [10,15] showed that increasing the flow rate multiplier (extrusion multiplier) or decreasing the layer thickness could minimize the presence of air gaps. Figure S8 (Electronic Supplementary Material; Online Resource) shows the cross-section view of the post-mortem specimens after the tensile test. It also showed the presence of air gaps, relevant to the finding in Fig. 4.

Conclusion
We have investigated the effect of varying specimens' printing angles on the tensile strength of 17-4PH stainless-steel parts via metal FFF additive manufacturing. The results showed that the specimens exhibited anisotropy. The ultimate tensile strength of the specimens varied with the changing of the printing angles. Specimens printed at 90° to the print bed showed an ultimate tensile strength of 440.15 MPa, which was less than 50% of the ultimate tensile strength obtained of specimens printed at 0° to the print bed. We also observed a similar trend for the specimens' elongation at break. Specimens printed at 90° to the print bed showed ~ 3.5 times lower elongation at break, compared to specimens printed at 0° to the print bed.
The anisotropy was attributed to the presence of air gaps, which existed due to extrusion underfill. Moreover, as the specimens' printing angle changed, the toolpath, number of layers, and the direction of load relative to the layers changed accordingly. While metal FFF additive manufacturing promises a balanced performance of good resolution, acceptable printing time, and reasonably low operation cost, further developments are required to reduce the air gaps between the extruded lines.
We recommend using printing angles of 0° to 10° to the print bed for 17-4PH stainless-steel material metal FFF additive manufacturing. This is because optimum tensile strengths were obtained for metal FFF when the printing angles were between 0° to 10°. The Y 0° specimens had an average of 947.26 MPa of ultimate tensile strength and 2.98% of elongation at break, while Y 10° specimens had an average of 949.87 MPa of ultimate tensile strength and 2.20% of elongation at break. This study highlighted the importance of considering the printing angles to avoid any potential mechanical failure or deformation when the part is in operation.

Data Availability
Data to be made available on a reasonable request.

Conflict of interest
All authors declare no conflict of interest.

Open Access
This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.