Experimental Investigation on the Effect of Carbon Fiber Reinforcements in the Mechanical Resistance of 3D Printed Specimens

In the last years, fiber-reinforced polymer composites have been under study for additive manufacturing. For this purpose, it is important to assess the behavior of these materials in terms of mechanical properties. The present experimental study evaluates the mechanical resistance of both PLA and carbon fiber reinforced PLA. The work used a full factorial Design of Experiments (108 tests) selecting as factors the infill density, infill pattern, material, number of perimeters and printing orientation. The main results highlight that the most influential factors on the tensile strength are both type of material and number of perimeters. In this study, the use of reinforcements did not improve the mechanical resistance attained by the corresponding virgin material. Particularly, for some selected specimens, the porosity measured in the fracture section is larger for the reinforced PLA specimens, so they showed a smaller cross-section.


Introduction
Though some precursors of additive manufacturing (commonly known as 3D printing) such as Joseph E. Blanther, Pierre Alfred Leon Ciraud and Hideo Kodama are well known, it is Charles Hull, founder of 3D Systems, who is often identified as the 3D printing inventor in the 1980s due to his patent for the stereolithography process. Nonetheless, Hull developed the STL file format for this manufacturing process, which currently remains the standard format in 3D printing slicers [9,19,32,39]. Since the 1980s, additive manufacturing has been increasing its popularity by means of affordable desktop solutions that have become accessible due to the expiration of original patents [19] and by Adrian Bowyer's RepRap project [18].
Additive manufacturing market forecasts are positive about the growth of the market in the coming years, up to 2023/2025, as reviewed by Altıparmak et al. [2] and Peng et al. [36]. The technology is one of the key components of industry 4.0 or smart manufacturing due to its suitability for mass customization [28] and being applied in applications such as manufacturing, civil engineering, automotive engineering, biomedical engineering, food and clothing [44,49].
Fused filament fabrication (FFF), which Stratasys has been operating commercially as fused deposition modeling (FDM) since 1991, is currently one of the most expanded additive manufacturing technologies [46]. Material extrusion is the basis of the process. A thermoplastic material is heated above its melting temperature at an extruder and deposited onto a bed using a layer-by-layer strategy through a nozzle tip. Then, the material solidifies shortly after deposition and incrementally creates the final shape of the part [3,16]. Polymer materials, usually supplied as thermoplastic filaments, are widely used in additive manufacturing applications. Among others, some of the most common materials used in FFF are the polylactic acid (PLA) and the acrylonitrile butadiene styrene (ABS) [14,45]. Thermoplastics like polyether ether ketone (PEEK) and polyethylenimine (PEI) are also available, but they usually require advanced printers in terms of temperature control. Other additive manufacturing processes such as laminated object manufacturing, binder jetting and selective laser sintering were developed to increase the applications to other materials such as ceramics and metals [4,46].
The development of additive manufacturing goes hand in hand with the development of new materials such as the composite materials [11,12,22,50]. These materials have found a wide number of applications in sectors such as automotive and aerospace [8,40]. It is critical to understand the capabilities and limitations of the composite materials to use them in 3D printing, particularly, to understand the influence of both short and continuous carbon fiber reinforcements [31,47]. Regarding, continuous reinforcements, for instance, Heidari-Rarani et al. [17] showed how tensile and bending strengths of continuous carbon fiber reinforced PLA outperformed those of pure PLA.
Short carbon fiber reinforcement composites are found to be a good material for various additive manufacturing technologies expecting advantages in strength, stiffness, creep resistance, thermal expansion, or toughness, depending on the type of fiber used [47]. For instance, Ning et al. [34] presented a study in which several carbon-fiber reinforced ABS specimens were fabricated and tested, proving that tensile strength and Young's moduli could be improved. The authors also identified the influence of factors such as fiber length and reinforcement percentage. However, issues such as void formation as identified by Tekinalp et al. [41] may compromise the results of the process. Table 1 shows some recent studies on the use of polymers reinforced by short carbon fibers. The work by Tekinalp et al. [41], Ning et al. [34], Ferreira et al. [12] and Blok et al. [5] are focused on aspects regarding mechanical characterization of such 3D printed composites, basically under tensile, flexural and shear tests, quantifying elastic and strength properties. Nevertheless, several printing parameters and 3D printer setting influence the mechanical properties of the 3D printed parts [29,30]. Rao et al. [37] performed a statistical study to identify the influence of layer thickness, infill pattern and extrusion temperature on the tensile strength of a PLA reinforced with short carbon fibers. Kumar et al. [24] performed a statistical study investigating the influence of infill density, printing speed and layer height on a similar composite. The investigation of such influences is relevant to quantify the importance of FFF parameters (as well as their interactions) on properties of 3D printed materials. However, the field still requires new experimental studies to thoroughly evaluate the suitability of these materials, their applications and requirements. The present study shows an experimental investigation on the mechanical properties of 3D printed specimens, manufactured through FFF technique, using both PLA and reinforced PLA with short carbon fibers. The study aims at analyzing and quantifying the influence of different printing parameters on the tensile strength and studying the fracture behavior. The printing parameters investigated were the number of perimeters (1 and 2), infill density (60, 80 and 100%), infill pattern (grid, octa and triangular), printing orientation (edgewise, horizontal and vertical) and type of material (PLA and reinforced by short carbon fiber PLA). The study includes the following sections after this introduction. Section 2 presents the materials and methods used to perform the experimental study. Section 3 shows the obtained results that help to calculate the stress-strain diagram for some selected specimens, to carry out a statistical study to evaluate the influence of the printing conditions and to analyze the fracture behavior and fracture strain for some selected specimens based on the porosity. Finally, Sect. 4 identifies the main conclusions of the research work.

Materials and Methods
The printing materials used in this study were PLA Ivory white, Smartfil (Smart Materials 3D) and reinforced PLA (rPLA), Proto-pasta (Protoplant). The rPLA included 15% in weight of short carbon fibers. The diameter of both materials was 1.75 ± 0.03 mm and their density of 1.24 and 1.3 g/cm 3 for PLA and rPLA, respectively.
The 3D printer used to manufacture the specimens was a 3D BIBO 2 Touch Dual Extruders. The printer was equipped with 0.6 mm diameter hardened steel nozzles to improve the resistance of the nozzle to the abrasion caused by the reinforcements. Besides, according to Yang et al. [48], the use of large nozzle diameters helps in obtaining higher tensile strength. The maximum nozzle temperature is 270 ºC and the maximum value of material flow is 24 cm 3 /min. Finally, the Ultimaker Cura 4.0 software allowed generating the printing code.
The layer-by-layer process may produce unwanted effects such as anisotropy and residual stresses [38,45]. Because of the anisotropy, the UNE 116,005:2012 standard [42] requires printing the specimens in three different orientations: edgewise, horizontal and vertical. To do that, the standard defines two types of specimens: 1A and 1AV, whose length is 150 and 120 mm, respectively. Moreover, the cross-section of the two specimens in the middle section defined by the standard is different, being 4×10 mm 2 (1A) and 5×10 mm 2 (1AV). Figure 1 shows the three orientations to test.
The selection of the printing parameters can notably affect the mechanical properties of the 3D printed parts [7,10,21,25,29,33]. Those related to the layer deposition process (e.g., layer height, infill density, infill pattern, width and number of perimeters, etc.) are of special importance.
The experimental plan includes as factors: number of perimeters (1 and 2), infill density (60, 80 and 100%) and infill pattern (grid, octa and triangular). Figure 2 shows the three types of infill. Other printing conditions included layer height of 0.2 mm, printing speed of 60 mm/s, bed temperature of 60 ºC and printing temperature of 210 ºC and 250 ºC for PLA and reinforced PLA, respectively [27]. Moreover, the height of both bottom and top layer was 0.6 mm. The experiment used a randomized full factorial design of two factors with three levels and one factor with two levels (18 tests per material and orientation). All the tests were performed using the three orientations: edgewise (E), horizontal (H) and vertical (V), but only one specimen was tested for each specific orientation. Based on this, the experimental plan requires printing 108 specimens. Table 2 shows a generic experimental plan to use for each of the orientations and material.
Strength related quantities were chosen as focus of the present assessment, being the tensile strength for the maximum load, which specimens could stand at testing, the main one. Considering the type of specimens here employed and the type of tests performed (tensile), equivalent stiffness (another quantity that can be assessed for a printed material) is mostly influenced by infill density. The larger the infill percentage, the larger the expected final equivalent stiffness. On the other hand, for strength, it is more complicated To measure the maximum load during tensile testing, the machine used was an MTS Criterion 43 with an LPS 104 cell and Instron 2716-015 clamping system, using the TW Elite software. The strain rate was 0.5 mm/min. Tensile testing of the specimens provided the maximum load and the elongation measured due to the crosshead displacement. To calculate the tensile strength, the maximum load was divided by the initial area of the crosssection in the middle part of the specimen. The area of the cross-section was estimated by using the infill density, number and thickness of the perimeters and thickness of the bottom and top layers and it is listed in Table 3 as a percentage related to a full solid area. The  fracture strain was calculated by means of the relation between the crosshead displacement and the distance between the grips (115 mm for 1A and 85 mm for 1AV) [43]. Image J software is employed to evaluate the porosity of the fractured section of the specimens. The software allows converting images of the fracture section to 8-bit grayscale images, after removing the background, to measure the area occupied by the pores in percentage. This evaluation was done by measuring the area related to white pixels with RGB value of 255 against the total area of the cross-section, after using the threshold feature.

Results and Discusions
The experimental plan included the realization of 108 tensile strength tests. Table 4 and Table 5 list the results of the tests in terms of the tensile strength and fracture strain for PLA and rPLA, respectively. The tables show the results classified by means of the orientation used for printing.
The results of tensile strength and fracture strain are close as shown in Tables 4 and 5. In fact, shortly after attaining the maximum load, the specimen suddenly breaks after a short additional elongation, a behavior associated to brittle fracture.

Stress-Strain Diagram
By means of the initial geometry of the specimen and the maximum load and elongation results, it is possible to plot the stress/strain diagram. For clarity, the diagram depicts only one test for both materials and for the three orientations. Figure 3 shows the results obtained for test 2 (infill density of 80%). In the figure, it is possible to appreciate how the PLA specimens admit higher tensile stress magnitude than that of the reinforced PLA.
Particularly, the reinforced PLA shows a more fragile behavior, being the measured strains  smaller than those obtained for PLA. The tensile stress that withstands the reinforced material is clearly lower than that of the PLA.

Statistical Analysis
The calculated results for the tensile strength were analyzed by means of the Analysis of Variance (ANOVA) including as factors: orientation, number of perimeters, infill density, infill pattern and material. Table 6 lists all these results. The results were also analyzed by means of the Shapiro-Wilk tests to check the normality of the residuals (W = 0.99145, p-value = 0.7366 < W).
As the significance level for the ANOVA, a p-value of 0.05 is selected [6]. Thus, based on Table 6, only the type of material, the number of perimeters and the orientation are significant sources of variation of the tensile strength. Alafaghani et al. [1] studied the influence of the infill pattern for infill density of 100%. In their paper, the authors stated that the influence of the infill pattern produced no significant influence on the mechanical properties. Regarding the infill density, Alafaghani et al. [1] and Gunasekaran et al. [15] indicated that the larger the infill density, the larger the tensile strength expected during tensile testing due to an improvement in bonding. However, based on the ANOVA, the variation of tensile strength depending on the infill density is minimal in this experimental study and cannot be considered as a statistically significant source of variation.
The contribution of the significant sources of variation to the total variability is not uniform. In this sense, in Fig. 4, it is possible to appreciate the percent contribution (%) to the total variability for each of the sources, particularly, the residuals that accounted for 43.5% of the variability. This may be due to the absence of interactions in the analysis. The other sources of variation explained more than 50% of the variability. After the residuals, the type of material (24.5%) and number of perimeters (26.4%) are the most important sources of variability.
The use of reinforcements produced a reduction of the tensile strength attained by the specimen. Particularly, 24.5% of the measured variability depended on the type of material. This result agrees well with that presented by Ferreira et al. [12] and Kovan et al. [23]. The type of carbon fiber used as reinforcements (short) can be an adequate explanation for these results. In this case, the polymer matrix mainly supported the load while the reinforcements did not play an important role on that. On the contrary, the reinforcements seem to create a more fragile material. Liu et al. [26] identified that adding short carbon fibers to virgin PLA considerably worsen the mechanical properties. It should be noted that it is still possible to improve the results of rPLA by optimizing the printing configuration based on the selected parameters, and by performing additional research in other factors such as the bed and printing temperature or nozzle size. Among the significant sources of variation, the factor with one of the highest influence on the tensile strength was the number of perimeters (Fig. 4). Particularly, the resistance was increased when two perimeters were used for both PLA and rPLA materials as shown in Fig. 5. The number of perimeters influences the amount of material in a particular crosssection. This result agrees well with the findings presented by Lanzotti et al. [25]. Similarly, Fountas et al. [13] showed also how tensile strength increases when the shell thickness (i.e., related to the number of perimeters) increases. When evaluating the influence of the material, the reinforced PLA showed lower values for the tensile strength at both one and two perimeters.  The orientation used in printing is expected to condition the tensile strength. Particularly, the maximum strength should be obtained when the printing orientation is the same as the direction used in the tensile testing [45]. Kamaal et al. [20] showed results in which the tensile strength of carbon fiber reinforced PLA is slightly lower when printing in vertical orientation than when using the other two orientations. Although, the ANOVA highlighted this factor as statistically significant, the contribution to the variability was minimal. By way of caution, it should be considered that some previously published studies related to the orientation are based on one single specimen, while, in the present study, two different specimens (i.e., 1A and 1AV) were used. Moreover, it should be noted the effect of the influence of the bottom and top layers of 0.6 mm thickness in the edgewise and horizontal orientations, when using two perimeters. This creates a slightly weaker contour and, thus, a weaker cross-section than that of the vertical orientation (see Table 3).

Fracture Surfaces of Selected Specimens
In Fig. 6, it is possible to observe the fracture surface of the specimen for tests 4, 9 and 13 that correspond to octa, grid and triangular infill patterns, respectively. The parameters used to perform these tests were layer height of 0. perimeters and vertical orientation. The true cross-section of the fracture zone is higher for the PLA specimens. This is due the occurrence of undesired porosity. Tekinalp et al. [41], when using carbon fiber reinforcements, also identified the occurrence of porosity in their experimental study. As Özen et al. [35] stated, the bonding is not perfect, so partially bonded filaments and voids compose these 3D printed composite specimens. Particularly, as shown in Table 7, the porosity related to the fractured section, calculated by means of the Image J software, notably increases when using reinforcements, and changes from 1-8% to 15-20%. Besides, from Fig. 6, it is possible to notice how the bonding between the two perimeters is even stronger when using PLA (because some grooves that mark the limits between perimeters are visible in rPLA specimens and are not for PLA).
The presence of porosity is especially important when studying the fracture during testing. In this sense, Fig. 7 shows the fracture strain calculated by means of the crosshead displacement against the distance between the grips. In the figure, it is possible to appreciate how the strain for PLA is larger than that of rPLA. Thus, the reinforcement clearly provides brittleness to the specimens. It is also highlighted the possible influence of the used nozzle. The diameter of the nozzle might not have had a suitable size to let the short carbon fibers align with the material flow. It is also highlighted the possible influence of the carbon fiber percentage in the results. Ning et al. [34] showed how when increasing the carbon fiber percentage using ABS matrix from 5 to 10%, the tensile strength diminished. In the present study, the percentage of carbon fiber was even higher, 15% in weight. Moreover, the PLA itself has a mechanical behavior closer to brittle in comparison to ABS.  Fig. 7 Failure strain for the different studied orientations Furthermore, by observing the results in this subsection, it is possible to hypothesize that the perimeters had a higher influence on specimens' strengths because they seem to be areas of deposited material with improved adhesion, in comparison to infill patterns. Perimeters are of simpler and more straightforward deposition in comparison to infills, which could assure more consistent material flow and areas of improved adhesion. Moreover, in the present case, two perimeters fill a considerably bigger portion of the specimen cross section with a region of good adhesion, in comparison to the one perimeter case.

Conclusions
In recent years, 3D printing is increasing its applications in a wide range of industrial sectors. New materials are being developed and tested for 3D printing purposes, for instance reinforced polymers. The present study shows an experimental investigation of the mechanical resistance of 3D printed parts manufactured of PLA and short carbon fiber reinforced PLA. The main conclusions of the study are as follows: • The stress-strain diagram of a selected group of specimens showed how the virgin material outperformed the reinforced one in terms of attained tensile stress and strain. • The full factorial experiment carried out to evaluate the influence of the orientation, infill density, infill pattern, number of perimeters and type of material on the tensile strength identified as statistically significant factors both the type of material and number of perimeters. • The effect of the number of perimeters was to increase the tensile strength as their values were increased. Both tested materials followed this trend. • Regarding the type of material, the use of short carbon fiber reinforcements in the PLA did not produce improvements in the mechanical strength of tested specimens, as expected from previous literature results. This behavior might have been favored due to a higher level of porosity of the here studied samples cross-section.
Funding Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature.

Data Availability
Authors can confirm that all relevant data are included in the article and/or its supplementary information files. Source data for Fig. 3 are available upon reasonable request to the authors.
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/.