*2.3. Design of Manufacturing Parameters and Case Studies*

The mechanical characteristics of the pieces produced using AM are very much dependent upon the manufacturing parameters. Figure 4 includes a diagram of the most relevant of those parameters.

**Figure 4.** The geometrical parameters with a direct influence on the mechanical properties of pieces manufactured using FDM.

This paper will study the influence of layer height, the percentage of infill, and construction orientation of the test specimen on the mechanical tensile properties of the two materials as these are the three most influential parameters. One of the main objectives is to check how much the mechanical performance improves when the percentage of the infill is increased, which is why two values (20% and 50%) have been included in the study. By considering one or the other infill value, the volume ratio of material changes due to the variation of the infill density which applies different sizes of cell patterns. On the other hand, the thickness of the piece's shell remains constant, with a value of 0.8 mm. Layer height is a parameter that influences manufacturing time; the precision of details and the finish are significantly improved by using finer layers. However, this study sets out to also analyse the influence of this parameter on mechanical strength. To do so, two layer heights, 0.1 mm and 0.2 mm, will be used. Finally, a comparison will be made to discover how much the mechanical strength varies with respect to the orientation of the piece during its construction. This is a fundamental factor due to the anisotropy conferred upon the piece by the layering manufacturing process. Both in the case of using orientation 1 and in that of using orientation 2, the layers are parallel to the stress, but the orientation of the infill raster varies with respect to the tensile force. Furthermore, in this case, it is necessary to use supports, and this causes a certain degree of roughness on the overhanging surface when removing them, but the results are acceptable for test purposes. In the case of orientation 3, layering in a direction perpendicular to the stress should result in a reduction of mechanical strength. A summary of the parameters whose influence on mechanical performance is going to be analysed is presented in Table 4.


**Table 4.** The definition of case studies for a layer width of 0.4 mm, a printing velocity of 60 mm/min (30 mm/min on wall and bases), a line-type wall pattern, and infill patterns of 45◦ and 135◦.

Finally, the rest of the parameters are specifically defined for each material in Table 5.

**Table 5.** The remaining manufacturing parameters specified for each material.


#### *2.4. Experimental Procedure and Geometrical Dimensions of Test Specimens*

The test procedure that is going to be followed is explained in a previous work [43] and uses the type I general usage test specimen as per ASTM 638-14 [23] (Figure 5, above), as explained in detail in that study. It also includes the principal difficulties encountered during the testing of plastic material specimens obtained using FDM and the fact that the test specimen breaks prematurely due to the concentration of stresses, which have also been reported by other authors, such as Wendt et al. [44]. The ASTM 638-14 standard, indicated by Dizon et al. [45], is the most used for this type of mechanical characterisation. The specimen is fixed by the clamps, aligning its longitudinal axis with the axis of the test machine. The wedges should be tightened enough to avoid the specimen to slide without causing it to collapse. The HOWIN software is used to control the test, to set the input parameters and to gather the values obtained. The speed of the test should be 5 mm/min according to the ASTM D638-14 standard (the minimum value of those established in the standard that causes the break between 0.5–5 min).

According to the ASTM 638-14 standard, the nominal tensile stresses must be used, that is, the tensile load carried by the test specimen at any given moment per unit area of the minimum original cross section. Since this standard is used for determining the tensile properties of solid plastics obtained for traditional processes, the minimum original cross section is computed by multiplying the thickness of the sample and the width of the narrow section.

Two identical test specimens of each type are going to be manufactured and, in the event of the results varying significantly, a third test shall be carried out. The elongations will refer to the variation of the distance between grips and not to the variation of the benchmark length measured using an extensometer, indicated in standard ISO 527-1:2012 [46], which recommends the use of the nominal distortion when extensometers are not used. Finally, a velocity of 5 mm/min is going to be set in accordance with that established in the ASTM standard, so that the rupture occurs between 0.5 and 5 mins. Figure 5 shows examples of the PLA and ABS test specimens used in the tests arranged by case.

**The Geometry of the test specimens according to ASTM 638-14**

**Figure 5.** The test specimens manufactured using PLA and ABS. (**a**) Case 1; (**b**) Case 2; (**c**) Case 3; (**d**) Case 4; (**e**) Case 5.

The variables to be studied are tensile yield stress (*R*p), mechanical tensile strength (*R*m), nominal strain at break (εt), and the modulus of elasticity (*E*t).

#### **3. Results**

#### *3.1. Mechanical Behaviour of the PLA Parts*

Figure 6 shows the PLA test specimens after the test, with a specimen of each type being included. Although all of them fracture in the narrow area, it can be seen that some do so outside the limits of the benchmark length. This is one of the difficulties experienced during the laboratory tests described in a previous study [35].

In the first two cases, a rupture with an irregular breakage line can be observed. This is due to the layers being oriented in the direction of the stress and the raster angle of the infill being 45◦. In the third case (50% infill), the test specimen behaves almost as though it were solid, as can be observed in its transversal cross-section. In the fourth case (orientation 2), the line of fracture occurs along a line perpendicular to the stress and located at a point where the infill joins with the walls. In the last case (orientation 3), the fracture occurs as a result of the layers separating because they were built up perpendicular to the stress due to the construction orientation.

**Figure 6.** The PLA test specimens tested (one of each type).

The graphs in Figure 7 show the test results for the different parameters. To help analyse these data more clearly, a graph is provided showing the test data for each type of test specimen grouped together along with the properties of the filament (Figure 7f). In each case, the results showing the intermediate values of each series have been chosen.

**Figure 7.** *Cont.*

**Figure 7.** The stress–strain curves (PLA). (**a**) Case 1; (**b**) Case 2; (**c**) Case 3; (**d**) Case 4; (**e**) Case 5; (**f**) the comparison of the results with the intermediate values of each series of test specimens.

The corresponding mechanical properties of the PLA test specimens have been calculated based on the information gathered during the tests. The data corresponding to the yield stress, tensile strength, nominal strain at break, and the modulus of elasticity are shown in Table 6.



∼= (similar to the benchmark case), + (higher than the benchmark case), ++ (much higher than the benchmark case), − (lower than the benchmark case), −− (much lower than the benchmark case).

Firstly, by comparing the results with respect to the benchmark case (Case 1), it is found that, indeed, the variation of the manufacturing parameters causes a variation in the tensile behaviour of the PLA test specimen, with its general performance being on the fragile side compared with that of the filament, which presents a high degree of ductility.

The slope of the elastic area (and, consequently, Young's modulus) remains practically unmodified by the change in layer height (Case 2), while it clearly diminishes when the manufacturing orientation is modified (Cases 4 and 5), and rises on the infill percentage being increased (Case 3).

The nominal strain at break (and, therefore, the ductility) clearly increases in those cases with a higher infill percentage (Case 3) and also with orientation 2 (Case 4), while it is reduced with manufacturing orientation 3 (Case 5) and remains more or less invariable with the increase of layer height (Case 2).

In all the cases, Young's modulus and the nominal strain at break are lower than in the filament.

The mechanical strength rises with the increase of infill percentage (Case 3) with respect to the initial conditions (Case 1). Meanwhile, a reduction thereof is observed in the cases of increased layer height (Case 2) and variation of the construction orientation (Cases 4 and 5), with this reduction being especially noteworthy in the case of orientation 3 (Case 5), as was to be expected.

Finally, the yield stress presents a performance similar to the tensile strength, bearing in mind that they practically coincide in many cases given the generally fragile performance of the PLA test specimens.

Considering mechanical strength to be the parameter of greatest interest due to its greater repeatability across laboratories and to its importance as a mechanical property from the design and functionality point of view for certain applications, Table 7 shows the percentage variation of the average mechanical strength in terms of the parameters used and the variation with respect to the material of the filament (data provided by the manufacturer).

**Table 7.** The comparison of the mechanical properties of PLA test specimens with those of the filament.


These data reveal that increasing the infill up to 50% (Case 3) greatly improves the mechanical strength (27%) as this makes the part more solid and reduces the number of cavities (its weight increases by 16%). The increase of layer height (Case 2) causes the maximum tensile strength to fall by 11%. Although this effect is less pronounced than that of varying the infill and the manufacturing orientation, it can be seen that finer layers lead to better results as far as both the finish and mechanical properties are concerned. With respect to the manufacturing orientations, orientation 2 (Case 4), with the layers in a direction parallel to the stress, but perpendicular to those of orientation 1, reduces the mechanical tensile strength by 22% (Figure 8a). On observing the fractured part (Figure 8b), it can be seen that this behaviour might be due to the different orientation of the infill with respect to the direction of the stress, which is what causes the part to rupture at the point where the infill joins the wall.

**Figure 8.** The detail of the infill and fracture surface in terms of orientation. (**a**) Orientation 1; (**b**) Orientation 2; (**c**) Orientation 3.

In the case of orientation 3 (Case 5), the overlap of layers perpendicular to the direction of the tensile stress causes the tensile strength to fall by 28%. This reduction of strength caused by varying the manufacturing orientation is due to the anisotropy of the pieces manufactured in the layers. In Case 5, this orientation causes the stress to revert to the interface between the layers, namely the weakest area of the pieces produced using additive manufacturing (Figure 8c).

In the table, it can also be seen that with the best of the combinations of parameters (Case 3) it is possible to achieve a strength that is only 4% less than that specified for the filament by the manufacturer.

#### *3.2. Mechanical Behaviour of the ABS Pieces*

Figure 9 shows the ABS test specimens after the test, with a specimen of each type being included. It can be observed that all the test specimens rupture where they are narrowest (some of them outside the benchmark length) except the one manufactured with orientation 3 (layers perpendicular to the stress), which ruptured with an extremely low stress due to a separation of layers in the area of the agreed radius.

**Figure 9.** The ABS test specimens tested (one of each type).

In Cases 1 and 2, a rupture with a jagged line can be observed. This is due to the layers being oriented in the direction of the stress and the raster angle of the infill being 45◦ (Figure 10a). In Case 3, given that it has 50% of infill, the test specimen is almost solid, which is why the distortion is more uniform (Figure 10b). In Case 4, the fracture occurs along lines perpendicular due to the stress and located at the points where the infill joins the walls (Figure 10c). In Case 5, the fracture occurs at low distortions as a result of the separation of layers because, due to the construction orientation these build up perpendicular to the stress, which causes the test specimen to break prematurely as the interface between the layers in this material is extremely weak (Figure 10d).

**Figure 10.** The fracture cross-section. (**a**) Case 1; (**b**) Case 3; (**c**) Case 4; (**d**) Case 5.

The graphs of Figure 11 show the results of the tests with the different parameters for the test specimens manufactured using ABS, where a lower variability in the results than in the case of the PLA test specimens can be observed.

**Figure 11.** The stress–strain curves (ABS). (**a**) Case 1; (**b**) Case 2; (**c**) Case 3; (**d**) Case 4; (**e**) Case 5; (**f**) The comparison of the results with the intermediate values of each series of test specimens.

The data corresponding to the yield stress, tensile strength, nominal strain at break, and modulus of elasticity are shown in Table 8.

**Table 8.** The mechanical properties of the ABS test specimens and variation with respect to the benchmark case (Case 1).


∼= (similar to the benchmark case), + (higher than the benchmark case), ++ (much higher than the benchmark case), − (lower than the benchmark case), −− (much lower than the benchmark case).

It can be seen how the change of manufacturing parameters also causes a variation in the tensile performance of the ABS test specimens.

The slope of the elastic area (stiffness) remains almost the same when orientation 2 is used (Case 4), while said incline (and, therefore, Young's modulus) increases significantly with an infill of 50% (Case 3), and less noticeably with an increase in layer height (Case 2). As regards orientation 3 (Case 5), the test specimen breaks so quickly that it is hardly possible to draw conclusions as to its tensile performance except for the fact that its deformation capacity is practically nil.

The nominal strain at break is slightly affected by layer height (Case 2), nor by the change to orientation 2 (Case 4) or by the increased percentage of infill (Case 3). However, the strain at the break diminishes significantly with respect to Case 1 when orientation 3 is used (Case 5).

In all the cases, Young's modulus and strain at break are lower than in the filament, as would occur in the case of PLA.

On the other hand, with the exception of the cases of vertical orientation (Case 5) and the increase of the infill (Case 3), the differences with regards to mechanical strength are not too pronounced in all the other cases. Both the use of thicker layers (Case 2) and the change of the layer orientation to keep them parallel to the stress (Case 4) reduce the maximum strength, but only slightly. Increasing the infill to 50% (Case 3), with the resulting 18.5% rise in weight, causes the tensile strength to increase by 25.19%, while orienting the layers perpendicularly to the stress (Case 5) causes the premature rupture of the test specimen with a strength reduction of 88.07% (Table 9). Another interesting fact shown by the table is that with the different combinations of parameters, the best results that can be obtained are 22% worse than those of the filament.



The change of infill results in a more solid test specimen and a reduction of cavities, once again result in an increase of mechanical strength. As regards layer orientation, if this is perpendicular to the stress, it causes the strain to revert to the interface between layers, namely the weakest area in the pieces obtained using additive manufacturing and especially in the case of ABS. The weakness of this interface can be seen in the graph, which shows that the piece breaks under a far lower strain than the pieces manufactured using other orientations due to layer separation.

#### *3.3. Comparative Analysis*

In the stress-strain graph of Figure 12, the tensile performance of both materials is shown together to facilitate the comparative analysis.

**Figure 12.** *Cont.*

**Figure 12.** The comparison of the stress–strain curves for both materials. (**a**) Filaments vs. Reference case, (**b**) Reference case vs. + layer height, (**c**) Reference case vs. + infill density, (**d**) Reference case vs. layer orientation 2, (**e**) Reference case vs. layer orientation 3.

The variability of the results can be compared for both materials through Tables 6 and 8 through the standard deviation. The variability of the modulus of elasticity is very similar for PLA and ABS, although, in general, it can be stated that results with ABS showed a lower variability than in the case of PLA. To help quantify the effect of the manufacturing parameters, Figure 13 shows the data corresponding to the modulus of elasticity, nominal strain at break, yield stress, and tensile strength obtained following the tests for the 5 cases analysed.

**Figure 13.** The comparison of mechanical properties for PLA and ABS. (**a**) Modulus of elasticity; (**b**) Nominal strain at break; (**c**) Yield stress; (**d**) Tensile strength.

Both figures show that the test specimens manufactured using PLA generally present a more rigid performance associated with the higher values of Young's modulus found, in line with the values obtained for the filament. The influence of the manufacturing parameters presents a similar trend for both materials.

As far as the strain at break values obtained are concerned, it can be said that there is no clear trend given that, for example, in Cases 1 and 2, a greater ductility is obtained for ABS, while in Cases 3 and 5 the PLA's ductility is clearly higher, with very similar values being obtained for both materials in Case 4.

On the other hand, in light of the values obtained, the pieces manufactured using PLA have a greater tensile strength than their ABS counterparts, just as would occur with the filament. Furthermore, the change of parameters has less of an influence on mechanical strength in the case of ABS, with the exception being Case 5, where the reduction of mechanical strength is more significant. Figure 13d shows that PLA undergoes more significant reductions of maximum strength than ABS when layer height is increased (Case 2) and transversal orientation is used (Case 4). Under all circumstances, the infill percentage (Case 3) is the factor with the greatest influence on the results. The performance of the tensile yield stress is identical to that of the maximum strength.

Finally, the differences with respect to the starting material and the decrease of mechanical strength in the case of orienting the layers perpendicular to the stress (Case 5) are less in the PLA, which would appear to indicate a significantly stronger bond of the layers in this material (Table 10).


**Table 10.** The comparison of mechanical strength with the benchmark case and the filament.

Based on this data, it can be observed that increasing the infill up to 50% (Case 3) causes a marked increase in mechanical strength, with this being of a similar magnitude in both materials (27% PLA + 16% weight and 25% ABS + 18% weight). This is because the piece is more solid and has fewer cavities.

As has already been explained, the layer height effect (Case 2) is more significant with PLA as it causes an 11% reduction in tensile strength, whereas in the case of ABS this is only 8%. It can thus be seen that finer layers lead to better results not only where the finish is concerned, but also in mechanical properties, although the effect is more noticeable in the PLA than in the ABS.

As far as the manufacturing orientations are concerned, orienting the layers in the direction parallel to the stress, but perpendicular to Case 1 (Case 4), greatly reduces tensile strength in PLA, (−22%) while its effect is not remarkable in ABS (−6%).

In the case of vertical orientation (Case 5), the overlapping of layers perpendicular to the direction of the tensile force reduces tensile strength by 28% in PLA, a much lower figure than in the case of ABS, which fractures prematurely (with an 88% reduction of maximum strength) due to its weak bond between layers.

The variations with respect to the filament lead to the conclusion that the material which undergoes a lower reduction of its mechanical properties during the additive manufacturing process is PLA (a mere 25% as against the 38% of ABS under benchmark conditions), which would appear to indicate the bond it achieves between layers is better than that of ABS, thus, making it an ideal thermoplastic for use in FFF/FDM technologies.

Finally, in Table 10 it can be observed that with the best combination of parameters it is possible to achieve a strength of only 4% less than that specified by the manufacturer for the PLA material, namely, figures that are far better than those returned by ABS (−22%). This indicates that the bond between layers in this material is far better than in ABS, at least with the manufacturing parameters used in this study, and that with an optimum combination of parameters (a 100% infill, the alignment of layers in the direction of the stress, a layer height of 0.1 mm, and optimum raster and pattern, temperature and velocity parameters), it could be possible to achieve strength results very similar to those of the starting filament or injection or compression moulded pieces.

#### **4. Conclusions and Future Work**

This work has studied the influence of the principal manufacturing parameters of an FDM process on the mechanical properties of the two thermoplastic materials most widely used in this type of additive technique: ABS and PLA. These parameters are infill percentage, layer height and manufacturing orientation. The results with ABS show a lower variability than in the case of PLA.

Generally speaking, the pieces manufactured with PLA behave more rigidly than in the case of ABS, just as with the starting material, with the influence of the manufacturing parameters being very similar for both materials. However, there is no clear trend regarding the influence of the manufacturing parameters on the strain at fracture values obtained, unlike the starting material, where the ABS showed a higher degree of ductility than the PLA.

Regarding ABS, the mechanical strength results barely vary with respect to layer height (the maximum strength only falls by 7.57% when the layer height is increased from 0.1 to 0.2 mm) and when using another orientation in which the layers remain parallel to the stress (a reduction of tensile strength by an average of 5.87%). However, increasing the infill up to 50% (with an 18.5% weight increase) causes the tensile strength to rise by 25.19% given that the test specimen is more solid and the cavities are smaller. On the other hand, the vertical orientation of the layers perpendicular to the stress causes the premature rupture of the test specimen with a strength reduction of 88.07%. This is due to the strain reverting to the interface between layers, namely, the weakest area in the pieces obtained using additive manufacturing, especially in the case of ABS.

Another interesting point to emerge from this study is that, with the different combinations of parameters chosen, the best results that can be obtained are 22% worse than those of the starting material. This, combined with the lack of strength of the test specimens whose layers are oriented perpendicularly to the stress, confirms that the ABS pieces manufactured using this technology and these parameters show an extremely weak bond compared with the strength of the material.

In the case of PLA, the increased layer height causes the tensile strength to diminish by 11%, which is notably greater than in the case of ABS, which is why it can be concluded that finer layers lead to better results not only where finish is concerned, but also in mechanical properties, although the effect is more noticeable in the PLA than in the ABS. On the other hand, it can be seen that increasing the infill up to 50% (with a 16% rise in weight), greatly improves mechanical strength (27%), which is very similar to that of ABS (25.19%).

With respect to the manufacturing orientations of the layers in a direction parallel to the stress, but perpendicular to those of the initial orientation, the tensile strength is reduced by 22%, far and above that of ABS (5.87%). This behaviour can be attributed to the different orientation of the infill with respect to the direction of the stress, which is what causes the part to fracture at the point where the infill joins the wall.

The orientation with the overlapping of layers perpendicular to the tensile force the reduces maximum tensile strength by 28%, a much lower figure than in the case of ABS, which fractured prematurely (with an 88.07% reduction of the maximum strength) due to its weak bond between layers.

Regarding PLA it can be concluded that, with the best of the combinations of parameters a strength of only 4% less than that specified by the manufacturer for the material can be achieved, which is far better than that returned by the ABS (−22%). This indicates that the bond between layers in this material turns out to be extremely strong and is, therefore, highly suitable for use in additive

technologies given that with an optimum combination of parameters it is possible to achieve strength results very similar to those provided by injection or compression moulded pieces.

In general, the infill percentage is the factor of greatest influence in the results. The performance of the yield stress is identical to that of the maximum strength.

The methodology proposed is a reference of interest in studies involving the determination of mechanical properties of polymer materials manufactured using these technologies. Specifically speaking, these results can be extremely useful for the selection of suitable materials and parameters in FDM design and manufacturing processes.

The inclusion of other factors of interest such as the direction of the raster and manufacturing speed, as well as the influence of the interaction of these factors on the results, is proposed as a line of work for the future. Additionally, the influence of other advanced parameters that are closely related to more specific thermophysical and/or chemical phenomena (the local temperature of the filament during deposition, the molecular diffusion at the polymer interface, and the deposition pressure/force, among others) will be also addressed in future studies; this kind of analysis will require specialised equipment in most of the cases due to its complex nature.

**Author Contributions:** Conceptualization, J.C. and A.M.C. Funding acquisition, J.C. and A.M.C.; Investigation, A.R.-P. and J.C.; Supervision, A.M.C.; Writing—Original draft, A.R.-P.; Writing—Review & editing, A.M.C.

**Funding:** The APC was funded by the Annual Grants Call of the E.T.S.I.I. of UNED through the projects of reference [2018-ICF04] and [2018-ICF10].

**Acknowledgments:** This work has been developed within the framework of the project DPI2016-81943-REDT of the Ministry of Economy, Industry and Competitiveness. The authors also want to acknowledge Miguel Ángel Sebastián for his valuable comments and support in the experimental tests developed in this work. The authors would like to take this opportunity to thank also the Research Group of the UNED "Industrial Production and Manufacturing Engineering (IPME)".

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.

#### **References**


© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
