The Effects of Build Parameters and Strain Rate on the Mechanical Properties of FDM 3D-Printed Acrylonitrile Butadiene Styrene ()
1. Introduction
The ever-increasing demand for the development of new performance materials is attracting unprecedented renewed research interests in materials science. However, the rapid production of critical machine components and complex structures with high precision, at minimal cost, and with the required service properties may not be feasible by the conventional manufacturing techniques [1] - [8] . This was a strong motivation for the emergence of Additive Manufacturing (AM). Research concerns in AM have risen in a number of distinct ways [9] . It has gained prominence both in the popular media as well as the scientific journals [10] [11] . Research activities on the industrial applications of AM have led to an astronomical expansion in market for components produced through AM [12] . In addition, there is the prediction of a global compound annual growth of 27% in AM [13] and the nearly $11 billion industry in 2015 will increase to $26.7 billion by 2019. Furthermore, it was reported that the aggregate share of global AM by the United States, West European and Asian markets is expected to rise from 59.2% recorded in 2014 to an estimated 70% by 2019. According to [14] , AM market has equally been projected to grow to $3.5 billion in 2017 and to approximately $10 billion in 2022. These exciting market reports may not be realized unless researchers redouble their efforts in the field of AM.
Additive manufacturing can be described as the process of adding material layer by layer to form a part [15] . This contrasts with the traditional method, often referred to as Subtractive Manufacturing (SM), in which material is removed from the bulk to form the desired component. It is a technique that surpasses SM as it opens up opportunity for handling complex shapes with great design flexibility, and reducing waste in the production of machine components [16] [17] [18] [19] [20] . This method has become very popular for industrial and domestic purposes as it provides opportunities for reducing lead time and maximizing inventory strategies [21] . Among different AM technologies available, FDM remains one of the most versatile techniques due to its inherent flexibility and rapid prototyping [4] [7] [12] [22] - [29] . In FDM process, a thermoplastic filament is heated and extruded using a robotically controlled head where the material is deployed layer by layer on a printing surface in a temperature-controlled environment [27] [30] [31] [32] .
The existing and numerous potential applications of thermoplastic based materials such as ABS in manufacturing may be responsible for the renewed acceptance of FDM for production of structural components (Bumper bars, clarinets, automotive interiors and valves, pipes and fittings, bathtubs, shower stalls, connectors, etc.) from polymers. Several research efforts have focused on FDM for manufacturing ABS based components. FDM process with laser-assisted heating was adopted by Jun et al. [33] for improving the forming quality and shape accuracy of ABS. Results showed that the laser-assisted heating in the forming of ABS thin-walled parts has enormous effect in increasing the temperature of the local forming regions than the pre and post laser-assisted heating. The significant benefit of this temperature rise at the local forming regions is the remarkable improvement in the tensile strength, shape accuracy and effective bonding width of thin-walled fabricated parts using laser-assisted heating. As part of efforts to improve the performance of 3D-printed ABS, graphene nanoplatelets were incorporated in ABS matrix by Dul et al. [27] using a completely solvent-free process and then extruded in filaments suitable for FDM. Although the incorporation of graphene nanoplatelets improved tensile modulus, substantial reductions in the ultimate tensile strength, strain at break, coefficient of thermal dilation and creep compliance of the ABS were observed.
The influence of build processing parameters of 3D using FDM on the quality of parts and their functionality has been documented in the literature [34] . Guessama et al. [35] investigated the anisotropic damage effect of 3D printed ABS from FDM. The authors adopted severe compression conditions to describe anisotropy induced when printing under different build orientations. The outcome of the study revealed inconsistent plastic damage. It was concluded in their work that printing orientation is a key build parameter to be considered in order to optimize service properties of FDM 3D printed ABS. While Griffiths et al. [36] studied the influence of FDM build parameters on processing efficiency and performance of build parts, Srivastava et al. [37] employed response surface methodology for optimizing FDM process parameters such as (contour width, raster width, air gap, raster angle, slice height and orientation). From the foregoing and other similar studies [38] - [44] , it has been reported in literature that FDM build parameters have enormous effects on cost optimization and performance properties of ABS and other polymeric based materials. Despite that mechanical properties of polymers material are critical for their structural applications, available studies on the influence of FDM build parameters on the mechanical properties of 3D-printed polymer materials are limited.
In Griffiths et al. [36] , experiments were designed and conducted to quantify the effects of FDM build parameters on the tensile strength and Young modulus of 3D-printed parts. It was found that a maximum infill resulted to an improvement in the tensile strength. This is attributed to the fact that a maximum infill helps the production of solid build parts that are stronger than those produced using a honeycomb structure on the same material. A combination of infill level and a side slicing orientation were found to enhance the Young modulus of build parts. A 100% infill, 0.4 mm layer height and a slicing orientation were effective for the optimization of the tensile strength and Young modulus. Wu et al. [41] investigated the effects of raster angle and layer thickness parameters on the mechanical properties of 3D-printed ABS and Polyether-ether-ketone (PEEK). It was found that raster angle and layer thickness have significant effects on their tensile, compressive and bending properties. The combination of 300 µm layer thickness and [0˚/90˚] raster angle led to an optimal improvement in their mechanical properties. Although FDM build parameters demonstrated better mechanical properties for PEEK in comparison to ABS, the challenge of pore formation and poor interlayer bonding were reported by the authors as research gaps for future works. In Luzamin et al. [42] , the influence of layer thickness, deposition angle and infill parameters on Flexural force in FDM built specimens was studied. It was shown that combinations of 10% infill and 60˚ deposition angle reduced the total build time and improved the maximum flexural strength. The drawback of their works [42] is the issue of high minimum detectable effect size (MDES). Thus, for higher flexural strength at this combination of FDM build parameters, there is need for further research efforts towards reducing the MDES.
Kay [45] studied the effect of raster orientation on the structural response of 3D-printed ABS based material using FDM. The study showed that specimens with raster oriented in the direction of loading exhibited the highest strength. In a study performed by Roberson et al. [46] , 3D-printed impact ABS test specimens were considered in order to compare the effect of stress concentrator fabrication on the impact test data from printing and machining approaches. This was done in four different build orientations. There was an evidence of sensitivity to build orientation in terms of impact resistance, impact energy and break energy. An attempt was made by Patel et al. [47] to examine the influence of crack length and layer orientation on 3D-printed ABS specimens. It was found that crack length and layer orientation dictate the fracture properties of the specimens. An increase in crack length enhances the stress intensity factor, but decreases the required fracture load. Riddick et al. [48] investigated the effects of build direction and orientation on the mechanical response and failure mechanism of the fabricated ABS specimens. The tensile strength, the elongation-at-break and the tensile modulus were characterized along with failure surfaces at different build direction and raster orientation [−45˚/+45˚], [0˚/90˚], [90˚]. Their results showed that the tensile strength, elongation-at-break and tensile modulus were highly dependent upon raster angle and build direction.
The elasticity and yielding responses of ABS material created by 3D printing were investigated by Zou et al. [49] . The effect of printing orientation on the mechanical properties was quantitatively evaluated. It was found that printing orientation determines the precision or accuracy of results obtained. The correlation between the mechanical properties of part manufactured out from ABS using FDM and parameter such as layer thickness and orientation was investigated by Rankouhi et al. [50] . Statistical analysis of the data showed that thickness and raster orientation significantly influenced its mechanical properties. Mishra [51] reported that FDM build parameters such as raster fill pattern enormously influences the mechanical and wear resistance behaviors of 3D-printed ABS specimen. Effort was made to minimize the anisotropic behavior by controlling the raster fill pattern during part buildings. It was found that an increase in part orientation and layer thickness increases the surface roughness and presence of residual stresses. Dawoud et al. [52] investigated the influence of selected FDM build parameters on the mechanical response of ABS. The authors explored variations in raster angle and gap as possible criteria for improving the mechanical behavior of ABS material. Maximum tensile and impact strengths were obtained in the study with raster angle of [−45˚/+45˚] whereas [0˚/90˚] favored the attainment of the most enhanced flexural strength. It was however reported that a positive gap drastically reduced the investigated mechanical properties.
Owolabi et al. [32] carried out a study to understand the high strain rate dynamic behavior of 3D-printed ABS produced using FDM. Results obtained showed multiple stages of contraction and expansion during impact loading. The ring-like formation observed around layers of the specimen was noted to lead to the significant manifestation of multistage deformation behavior in 3D-printed ABS. This multistage collapse is an indication of potentials for novel energy absorption mechanism that can be explored at lower strain rates. The manner of fabrication using FDM which absorbed and released the energy thus acting as a multistage spring was attributed to the potential energy absorption mechanism. From the authors’ findings, the choice of FDM as a production technique for 3D-printed ABS can be considered as beneficial if capability for energy absorption is highly paramount. An investigation on the thermo-mechanical creep properties of polymeric materials such as ABS using FDM was conducted by Turk et al. [53] . The outcome of this work showed that FDM build orientations significantly influences the mechanical properties of ABS. Huang et al. [30] investigated the effects of FDM build fiber orientation, filament dimensions, and chemical composition on the mechanical properties of ABS-printed components. Results of 3D-printed material manifested anisotropic properties at different filament extrusion directions. In the study carried out by Aliheidari et al. [54] on the fracture resistance measurement of FDM 3D-printed ABS, it was inferred that the degree of interlayer bonding influences the fracture resistance properties of ABS materials.
Despite the immeasurable prospects of AM such as FDM, its absolute exploitation in industries is still hindered due to unreliability in the mechanical response behaviors of ABS based components [55] . Parts processed through FDM typically have lower mechanical properties when compared to those processed via conventional manufacturing techniques. From the foregoing, inappropriate choice of FDM build parameters has been shown to be a strong limitation to maximizing the mechanical properties potentials of ABS based components. Apart from the fact that studies on the effects of various combinations of FDM build parameters on mechanical properties of 3D-printed ABS are very sparse, investigations on multiple build styles of three parameters combinations for each build style have not been sighted. In view of these research gaps, there is need for continuous rigorous research efforts in order to further elucidate the influences of different build parameters on the mechanical behaviors of ABS. In this work, an investigation is made into the effects of raster angle, layer thickness and interior fill style at different strain rates on the modulus of toughness and modulus of resilience of 3D-printed ABS which have been rarely reported in literature. In addition, the effects of these parameters on the yield strength and the ultimate tensile strength were studied to complement what have been earlier reported. The printer available for this study only had discrete layer thicknesses available making continued exploration of this parameter impossible. The attempt made on investigating the effects of eight distinct build styles (with three FDM build parameters combinations for each build style) and exploration of different strain rates can be considered as the novelty of the present work. The outcome of this study may provide more useful information for predicting mechanical properties of FDM build 3D-printed ABS.
2. Materials and Methods
The Specimens used in this work were manufactured from commercially available ABS P430 filament. This filament was designed for use with the procured 3D printer, the Stratasys Dimension 1200es. The machine’s build envelop was 254 mm × 254 mm × 305 mm. Full factorial design experiment was performed. This experiment incorporated a 2-level, 3-factor design with raster angle, layer thickness and interior fill style as illustrated in Table 1. Response variables under investigation include the yield strength, ultimate tensile strength (UTS), modulus of resilience and modulus of toughness. There was a total of 8 build combinations as indicated in Table 2.
The ASTM dog bone test specimen was designed using Siemens NX 10 CAD software. The test specimens were designed to conform to ASTM D-638, Standard Test Method for Tensile Testing of Plastics, as illustrated in Figure 1. The CAD file was then exported as an STL file such that it could be interpreted by the FDM software, CatalystEX. Prior to testing, specimens were visually inspected for any physical damage and labelled per their build style and production date. Specimens were then conditioned for a minimum of 48 hours at 77 degrees Fahrenheit to comply with the ASTM D638 standard. Upon completing the conditioning period, uniaxial tensile tests were carried out using the 50 kN Instron 5569A tensile testing machine equipped with Bluehill data acquisition software, as shown in Figure 2. Testing was carried out at four different strain rates: 0.127 cm/min, 0.5 cm/min, 5 cm/min and 10 cm/min.
For each build combination in Table 2, five specimens were produced, 40 per strain rate and 160 specimens in total. The data was recorded and analyzed using
Table 1. FDM process parameters and associated levels.
Table 2. Full factorial experimental design.
Figure 1. ASTM D638 Dog-Bone Test Specimen [56] .
Figure 2. Instron Universal Tester Experimental Setup.
the Bluehill data acquisition software. The yield strength was calculated using the 0.2% offset method, while the modulus of resilience and the modulus of toughness were calculated by finding the area under the stress-strain curve up to the yield point and the point of fracture, respectively. Statistical significance was determined using IBM’s SPSS statistics software. A factorial Analysis of Variance (ANOVA) with four independent variables (raster angle, layer thickness, interior fill style, and strain rate) was conducted at a 95% confidence interval (the significant or risk level α was set at 0.05).
3. Results and Discussion
3.1. Comparison of the Effect of Build Parameter on the Mechanical Properties of ABS
In Griffiths et al. [36] , the ultimate tensile strength that was obtained with 100% infill build parameter is similar to that of the present work where the highest ultimate tensile strength was obtained with solid interior fill style. The layer thickness values in the present work and that of [36] differs by 36.5%, in a build style consisting of interior fill and layer thickness parameters combination. Among the [45˚/−45˚], [0˚], [90˚] and [0˚/90˚] raster angles investigated by Riddick et al. [48] , the raster angle [45˚/−45˚] in the vertical build specimens of ABS exhibited the highest ultimate tensile strength of 19.80 MPa. Whereas the present study achieved optimum ultimate tensile strength with raster angle of [0˚/90˚]. The optimal ultimate tensile strength of 27.30 MPa was achieved in this work with build style 8 (from optimum combination of build parameters: [0˚/90˚] raster angle, 0.254 mm layer thickness and solid interior fill style). In Rankouhi et al. [50] and the present study, the choice of low layer thickness seemed to favor the attainment of the highest ultimate tensile strength. Their tensile test results showed that ABS samples printed with 0.200 mm layer thickness exhibited higher ultimate tensile strength compared with 0.400 mm layer thickness. This is very much similar to the results of the present study where 0.254 mm layer thickness displayed higher ultimate tensile strength than the layer thickness of 0.3302 mm. The maximum ultimate tensile strength was obtained in the study of Dawoud et al. [52] with raster angle of [−45˚/+45˚] contrary to [0˚/90˚] obtained in the present work.
Based on the comparisons made to prior research efforts [36] [48] [50] [52] , the results presented in this study on modulus of toughness and modulus of resilience of 3D-printed ABS using different build styles (combinations of raster angle, layer thickness and interior fill style for each style) and the effect of strain rate on their mechanical properties can be considered as novel contributions to the existing literature.
3.1.1. Raster Angle Observations
The raster angle had a primary influence on the modulus of toughness. This is seen most markedly when viewing Figure 3. There is a prominent difference between the modulus of toughness of all specimens built with the [45˚/−45˚]
Figure 3. Effect of Raster Angle on Modulus of Toughness.
raster angle when compared to specimens built with the [0˚/90˚] raster angle. This observation was consistent throughout the experiment when comparing specimens built with the crisscross [45˚/−45˚] raster angle versus specimens built with the cross [0˚/90˚] raster angle. This difference in raster angle can lead to an increase of up to 200% in modulus of toughness in favor of specimens built with the [45˚/−45˚] raster angle. The results for the other material responses were not as clear. For example, Table 3 shows that build style 1, which was fabricated with the cross [0˚/90˚] raster angle, had a higher mean yield strength and modulus of resilience than build style 7, which was fabricated using the crisscross [45˚/−45˚] raster angle while holding the other parameters constant. However, build style 7 had a higher mean UTS and modulus of toughness when compared to build style 1. As the strain rate is increased in Tables 4-6, some overlaps were observed for the mean values of yield strength, UTS, and even modulus of resilience for build style 1 and build style 7. The constant overlaps of error bar and trend reversals in Figures 4-6 make it difficult to identify the relationship between the raster angle and the investigated mechanical properties.
3.1.2. Layer Thickness
Results obtained showed that the smaller layer thickness of 0.254 mm had superior mechanical properties, only for specimens built with the cross [0˚/90˚] raster angle (Factor 1: +). Specimens built with both 0.254 mm layer thickness and the cross [0˚/90˚] raster angle had superior mechanical properties when compared to those built with the 0.3302 mm layer thickness and cross [0˚/90˚] raster angle. When the 0.254 mm layer thickness is combined with the crisscross [45˚/−45˚] raster angle, this trend is either reversed or not distinct. This suggests that there may be key interaction between the layer thickness and the raster angle that dictated the mechanical properties of the specimens. This observation is illustrated in Figures 7-10.
Table 3. Results for tests conducted at strain rate of 0.127 cm/min.
Table 4. Results for tests conducted at strain rate of 0.5 cm/min.
Table 5. Results for tests conducted at strain rate of 5 cm/min.
Table 6. Results for tests conducted at strain rate of 10 cm/min.
Figure 4. Effect of Raster Angle on Yield Strength.
Figure 5. Effect of Raster Angle on Ultimate Tensile Strength.
3.1.3. Interior Fill Style
The effect of the interior fill style on the yield strength is illustrated in Figure 11.
Figure 6. Effect of Raster Angle on Modulus of Resilience.
Figure 7. Effect of Layer Thickness on Yield Strength.
It is shown that build 8 (Solid) had higher average yield strength at all strain rates when compared to build 1 (High Density). This observed increase from 7% to about 53% is consistent across build comparisons in Figure 11 and at all strain rates. The effect of the interior fill style on the UTS is described using Figure 12. It was revealed that build 8, with the solid interior fill style, had a higher UTS for all strain rates in comparison to build 1. This observation is consistent across all build comparisons in Figure 8 and at all strain rates with an increase ranging from 15% to 29%. Similar observations were noticed on the effect of the interior fill style on the modulus of resilience and modulus of toughness as
Figure 8. Effect of Layer Thickness on Ultimate Tensile Strength.
Figure 9. Effect of Layer Thickness on Modulus of Resilience.
shown in Figure 13 and Figure 14, respectively. These observations were deemed reasonable since the solid interior fill style can accommodate more stress due to its denser structure and higher effective cross-sectional area than that of the high density interior fill style.
Figure 10. Effect of Layer Thickness on Modulus of Toughness.
Figure 11. Effect of Interior Fill Style on Yield Strength.
Figure 12. Effect of Interior Fill Style on Ultimate Tensile Strength.
Figure 13. Effect of Interior Fill Style on Modulus of Resilience.
Figure 14. Effect of Interior Fill Style on Modulus of Toughness.
3.2. Statistical Significance
The summary of ANOVA for modulus of toughness, modulus of resilience, UTS and yield strength is illustrated in Tables 7-10. The modulus of toughness results indicated a dominant, statistically significant main effect of raster angle (A) (with a p-value = 0.000) followed by a significant main effect in interior fill style (C) (p-value = 0.000) and strain rate (D) (p-value = 0.000). The main effect for layer thickness (B) was deemed insignificant. Analysis for the modulus of resilience implies a dominant, statistically significant main effect of strain rate (D) (p-value = 0.000) followed by a significant main effect in interior fill style (C) (p-value = 0.000) and layer thickness (B) (p-value = 0.000). The main effect for raster angle (A) was considered insignificant with a p-value of 0.858.
The ANOVA summary for UTS showed a dominant, statistically significant
Table 7. ANOVA summary table for modulus of toughness.
Table 8. ANOVA summary table for modulus of resilience.
Table 9. ANOVA summary table for UTS.
Table 10. ANOVA summary table for yield strength.
main effect of strain rate (D) (with a p-value = 0.000) and interior fill style (C) (with a p-value = 0.000). This is followed by significant main effects in layer thickness (B) (p-value = 0.000) and raster angle (A) (p-value = 0.000). Also for yield strength, results showed a dominant, statistically significant main effect of strain rate (D) (p-value = 0.000) and interior fill style (C) (p-value = 0.000). This is followed by significant main effects in layer thickness (B) (p-value = 0.000) and raster angle (A) (p-value = 0.000).
3.3. Stress-Strain Curves
Figure 15 and Figure 16 illustrate the typical stress-strain behavior of specimens built with the cross [0˚/90˚] and crisscross [45˚/−45˚] raster angles at a strain rate of 0.5 cm/min in accordance with ASTM D638. The stress-strain curves for
Figure 15. Stress Strain Curves for [0˚/90˚] Raster Angle Specimens at 0.5 cm/min.
Figure 16. Stress Strain Curves for [45˚/−45˚] Raster Angle Specimens at 0.5 cm/min.
build styles 1, 2, 4, and 8 (shown in Figure 15) represent those for a brittle material with the [0˚/90˚] raster angle being the only common process parameter among all the aforementioned build styles. Whereas, the stress-strain curves for build style 3, 5, 6, and 7 (shown in Figure 16) show a more ductile material as they undergo a much larger amount of plastic deformation prior to fracture. Build styles 3, 5, 6, and 7 shared the [45˚/−45˚] raster angle as the only common process parameter. Furthermore, build styles produced with the [−45˚/45˚] raster angle tends to favor the viscous/energy damping properties of ABS. This improves the modulus of toughness and ability to tolerate plastic deformation when compared to their [0˚/90˚] counterparts.
3.4. Strain Rate Sensitivity
For all build styles, most mechanical properties investigated in this study exhibited modest sensitivity to the strain rate with the mean value of each property increasing as the strain rate is increased. From Figure 17 and Figure 19, a clear distinction between the UTS at the lower strain rates of 0.127 cm/min and 0.5 cm/min was observed. However, this difference is less pronounced at the higher strain rates of 5 cm/min and 10 cm/min. For specimens produced with the [0˚/90˚] raster angle, the mode of failure tended to be more brittle in nature as the strain rate is increased. Surprisingly, this was not the case for specimens produced with the [45˚/−45˚] raster angle as the strain rate seemed to have little effect on the mode of failure and its propensity for plastic deformation.
The effects of strain rate on the other mechanical properties are described using Figure 18-21 were generated with the mean values for each mechanical property over the four strain rates. Similar to the UTS, there is a distinctive increase
Figure 18. Effect of Strain Rate on Yield Strength.
Figure 19. Effect of Strain Rate on Ultimate Tensile Strength.
Figure 20. Effect of Strain Rate on Modulus of Resilience.
Figure 21. Effect of Strain Rate on Modulus of Toughness.
in the mean yield strength and modulus of resilience as the strain rate is increased from 0.127 cm/min to 0.5 cm/min. However, there is significant overlap in the mean yield strength values at the higher strain rates of 5 cm/min and 10 cm/min. Based on Figure 21, the effect of strain rate on the modulus of toughness seemed to depend on the build style. For specimens produced with the [45˚/−45˚] raster angle, there was a noticeable increase in mean toughness as the strain rate increased. However, specimens produced with the [0˚/90˚] raster angle exhibit significant overlap in the mean toughness at all four strain rates.
4. Conclusions
This study focused on investigating the influence of build parameters on the mechanical properties of 3D-printed ABS manufactured using FDM. The following specific conclusions can be drawn based on this study:
1) The effect of raster angle was highly significant on modulus of toughness. Specimens built with the [45˚/−45˚] raster angle have potentials for improving the modulus of toughness by 200% when compared to specimens built with the [0˚/90˚] raster angle. At a strain rate of 5 cm/min and [45˚/−45˚] raster angle, a substantial modulus of toughness 2.044 MPa was achieved.
2) The modulus of resilience is strongly influenced by strain rate values. At any build style combinations of parameters, a significant improvement in modulus of resilience is achievable at higher strain rates. A very high modulus of resilience was achieved to be 0.414 MPa with specimens built with the [45˚/−45˚] raster angle and solid interior fill style.
3) The UTS and yield strength properties were observed to be dominantly affected by strain rate and interior fill style. High UTS of 27.44 MPa and yield strength of 23.79 MPa were obtained with specimens built with the [45˚/−45˚] raster angle and solid interior fill style at a strain rate of 10 cm/min.
4) Although specimens built with both 0.254 mm layer thickness and the cross [0˚/90˚] raster angle had superior mechanical properties to those built with the 0.3302 mm layer thickness and cross [0˚/90˚] raster angle, the effects of the layer thickness on the mechanical properties was found to be inconsistent. The implication of this is that the interaction between the layer thickness and the raster angle is significant.
5) Specimens built with the solid interior fill style have greater yield strength and UTS than those built with high density. This is mostly due to the added density in the solid fill style that allows the accommodation of more stress than the high density fill style.
Future study will focus on dynamic behavior and wear resistance of the 3D-printed polymers to increase their potential applications in industry.
Acknowledgements
The authors gratefully acknowledge the financial support of Consortium for Advanced Manufacturing. The contributions of Wes Everhart (KCP) are also well appreciated.
Data Availability Statement
Data will be made available on request.