Effect of the ironing process on ABS parts produced by FDM

2.1 Manufacturing of specimens by FDM

Cubic and prismatic specimens (beams and plates) were manufactured on the 3D printer Ultimaker 3 with their brand of ABS filament. The 3D models were created in SolidWorks and exported in a stereolithograpy (STL) file to the slicer software Cura 4.2.1, to generate the g–code meant to be interpreted by the machine. Due to the limitations of the slicer, a Python script was developed to apply the ironing process in any desired layer of the modelled samples and generate a rotation of the same ironing path. The python script reads a conventional g–code file and searches ironing instructions present in a top layer which are then used to apply the process on different layers and manipulate its path and material flow, if needed.

Specimens were produced using an extrusion temperature of 250°C, a nozzle diameter of 0.4 mm and an infill density of 100%. The temperature of the platform was 90°C for cubes and beams. For plates, to induce a controlled level of warping, the temperature of the platform was set to 74°C. The printing speed was 30 mm/s for the first layer, external layers (top and bottom), and contours. A printing speed of 60 mm/s was chosen for internal contours and filling layers. Number of contours is 2 with a thickness of the contours, tops and bottoms equal to 0.8 mm, while the height of the initial layer was set to 0.27 mm.

Cubic samples (Figure 1A) were produced with a layer thickness of 0.2 mm, raster angles of (−45°/45°) and (0°/90°). The ironing process on cubic samples was applied on the top layer of cubes with a zigzag pattern, line spacing equal to 0.1 mm and material flow of 10%. The ironing pattern was applied in parallel to the printing pattern of the top layer. The ironing speed was 20 mm/s. The values for inset were 0.150 or 0.175 mm. ABS beams (Figure 1B) were produced with a layer height of 0.15 mm, a raster orientation of (−45°/45°) and a deposition speed of 60 mm/s. The ironing process on beams was applied in between n alternated layers (n = 4, 8, and 16) without adding material (0% flow), with an inset of 0.150 mm and an ironing speed of 25 mm/s. ABS plates (Figure 1C) were produced with a layer height equal to 0.15 mm, a raster orientation of (−45°/45°), and deposition speed of 60 mm/s. The ironing process on plates was applied after one of the three first layers of the specimens without adding flow material (0% flow), with an inset of 0.150 mm and an ironing speed of 25 mm/s. The geometry of the specimens and the applied ironing scheme are shown in Figure 1.

In the cubic samples, material usage in specimens with ironing increased slightly due to the 10% extrusion flow. Manufacturing time increased by more than the time it would take to print an extra layer since ironing speed is less than printing speed. As for energy consumptions, the increment of final energy consumption of the cubic specimens where ironing was performed is equal to the energy consumption of a normal sample plus the energy of one extra deposition layer. The measurements were obtained with a commercial energy consumption meter (EnergyOT), with a 15-second temporal resolution.

2.2 Optical profilometry measurements

Optical profilometry measurements were done on the top surface of ABS cubic samples. To increase the ABS surface reflectivity, the top surface of the samples was coated with reflective coating. The reflective coating (Au–Pt) was deposited on a sputtering system (Quorum–Q150TES) using a current of 40 mA and evaporation time of 60 seconds. The profilometry measurements based on the white light interferometry method were performed on the profilometer Profilm 3D with a 20× objective. The surface scanned area was 1.0 × 0.85 mm². The raw data were acquired using the envelope peak analysis. The raw data files were exported and postprocessed with the software “Profilm Online.” The average roughness (Ra) of the cubes top surface was determined considering the entire scanned area and along a perpendicular line relative to the deposition path of the final top layer. For each experimental condition, two measurements representative of the cubes surface were considered.

2.3 Flexural tests

Three-point bending tests following the standard ASTM D790 were performed on a universal testing machine (INSTRON 5566) with a load cell of 50 kN, crosshead speed of 0.5 mm/s, a support span of 30 mm, and a rod radius of 6.0 mm, until the failure of ABS beams occurs. The load-displacement curves were recorded. For each experimental condition, five measurements were considered. The ultimate flexure stress (UFS) and the flexural modulus (E_F) were calculated.

2.4 Sectional profiles geometry measurements with a scanner

The measurements of the profile geometry of the ABS plates were performed with a flat bead scanner from a commercial printer (HP OfficeJet 4630) (Figure 2A). The contact image sensor (CIS) from the scanner is a 216 × 297 mm area sensor with an optical resolution of 600 dpi (0.04 mm). ABS plates were placed on the top of the scanner (Figure 2B), and for each profile, an image was collected and then processed in a sequence of steps. First, the image is converted to binary colour scheme (Figure 2C) and then imported to SolidWorks. After, in the program console, several points (30) were collected along the perimeters of each specimen, defining its geometry by a polynomial extrapolation. The warping (δ) was calculated by the average maximum difference, in z direction, between the extracted value and the theoretical dimension (Figure 2D). The experimental error of the measurement was computed with the pixel width of the acquired images. For each experimental condition, 10 measurements were considered.

3.1 Optical profilometry

Figure 3 displays the magnified views of the coated top surface specimens together with the applied ironing schemes and top surface profiles of the ABS cubes estimated by optical profilometry.

From the data present in Figure 3, it is possible to observe that the Ra of the surface of samples decreased when the ironing process was applied. The global results show a maximum reduction from 23 to 9 μm (variation of 60%) of the surface Ra of ABS cubes when the ironing process was applied. It was also noted that the results show low dependency of the Ra value with the estimation method (area or perpendicular profile line), with a maximum variation of 10% between them. The obtained value of 9 μm for the Ra is comparable with the best reported values of 10 μm obtained by optimization of the FDM parameters for ABS,²⁴ being inferior to the values of Ra reported for PLA (more than 20 μm)²⁵ and superior to the optimal solutions (4 ± 1.5 μm) founded on Akande.²⁶ Postprocessing methods involving 3D chemical melting can achieve Ra values of 10 μm,²⁷ while the Ra values for posttreatments with acetone are inferior to 2.5 μm.²⁸ The post treatments with CO₂ lasers lead to Ra values in the submicron range.²⁹ Considering that the results were obtained adding a 10% material flow during the ironing process (default value in the slicer software), in future studies, the influence of material increment should be addressed, and the effect of a purely thermal procedure in the surface roughness of FDM parts quantified.

3.2 Flexural tests

Figure 4A displays the UFS and E_F of the ABS beams, highlighting that samples where the ironing process was applied in alternated layers have lower UFS and E_F, when compared with the samples without ironing. This result reveals that the thermal effect of the ironing process was not effective in increasing the ILS of ABS samples. The failure mode of specimens shows that the fracture occurs on stress concentration areas induced by the ironing process on the lateral surfaces of specimens, possibly due to retraction of the ironed layers and selected process parameters (Figure 4B).

Although the ironed samples were subjected to an additional heating treatment that potentially could increase ILS,¹⁹ the potential benefits of this thermal treatment vanish with the presence of more severe geometric irregularities (stress concentration factors) on the samples surface. For comparison, the elimination of stress concentration factors (by a melting process) can increase almost two times the deflection at break and the maximum tensile strength of ABS parts produced by FDM.²⁷

3.3 Warping

Figure 5A displays the photographs of the ABS plates without ironing and with ironing process applied on the top of the first and third layers. It is possible to observe that ABS plates that were not subjected to the ironing process present a higher level of distortion than the ABS plates subjected to the ironing process (Figure 5B). This severe level of distortion occurs in the middle of the ABS plates profile. When the ironing process is applied, there is a relocation of the maximum warping from the profile's middle (specimens without ironing) to the profile's tip (specimens with ironing).

On average, in samples that were subjected to the ironing process, the warping was reduced from 0.61 (specimens without ironing) to 0.41 mm (specimens with ironing on the first layer), which represents a 33% reduction of the distortion levels. The effectiveness of ironing on reducing the levels of distortion decreases when ironing is applied on the second or third layers. The positive effect of ironing on the reduction of warping of ABS plates can be related with an increase of adhesion between layers and the printer platform (through the reduction of internal stresses), induced by the thermal energy delivered during the ironing process.^{18, 19, 22, 23}