Non-Destructive and Destructive Testing to Analyse the Effects of Processing Parameters on the Tensile and Flexural Properties of FFF-Printed Graphene-Enhanced PLA

: The signiﬁcance of non-destructive testing (NDT) methods cannot be overstated as they help to evaluate the properties of a material without damaging/fracturing it. However, their applicability is dependent on their ability to provide reliable correlation with destructive tests such as tensile and ﬂexural. This correlation becomes more problematic when the material is not homogeneous, such is the case with parts manufactured using a popular additive manufacturing process termed as fused ﬁlament fabrication (FFF). This process also requires optimisation of its parameters to achieve desired results. Therefore, this study aims to investigate the effects of four different nozzle temperatures, print bed temperatures, and print speeds on FFF-printed Haydale’s Synergy Graphene Enhanced Super Tough PLA through three non-destructive (ultrasonic, hardness, strain) and two destructive (tensile, ﬂexural) testing methods. Samples were manufactured using Anet ® ET4 Pro 3D printer and evaluated as per British and International standards. Two non-destructive tests, i.e., ultrasonic and hardness have been associated with evaluating the tensile properties of the manufactured parts. These results were correlated with destructive tensile testing and showed good agreement. The NDT method of strain measurement showed a very good correlation with the destructive three-point ﬂexural test and was able to provide a reliable evaluation of ﬂexural properties as a function of all three processing parameters. The results presented in this work highlight the importance of NDT methods and how they can be used to evaluate different properties of a material.


Introduction
Fused filament fabrication (FFF) is a popular additive manufacturing (AM) process that is based on the principle of material extrusion and makes use of thermoplastics to manufacture products [1][2][3]. Fused deposition modelling (FDM) is another terminology used for this process but is a trademark of Stratasys. The systems working on the principle of FFF/FDM are termed as 3D printers [4,5]. The FFF process has several notable advantages such as ease of operation, wide variety of materials, and cost-effectiveness [6,7]. However, the layer-by-layer nature of the process and anisotropic behaviour of the materials make the detection of defects in FFF-printed parts difficult. These defects could also be introduced if the processing parameters are not optimised for the material being used such as nozzle temperature, print bed temperature, print speed, infill percentage, and infill pattern [8]. In such a situation, non-destructive testing (NDT) is an appropriate choice for detecting and evaluating these defects without damaging the FFF-printed parts or altering their properties as opposed to a destructive test, e.g., tensile, and flexural [9]. These methods have been in use to provide an assessment of a product's integrity, quality, and reliability [10]. The common NDT methods for FFF include visual inspection, ultrasonic testing, thermography, and acoustic emission testing. Mwema et al. [11] used visual GPLA showed significant improvements compared with PLA. These examples highlight the significance of graphene-enhanced PLA material, the importance of understanding the combinational effect of process parameters, and the need for optimisation to achieve desired results. However, the majority of testing undertaken to assess the mechanical properties of materials is destructive. In the case of FFF/FDM, manufacturing parts for testing is quick but bespoke properties require optimisation of processing parameters and having reliable NDT methods could help evaluate such properties without damaging the printed parts.
Therefore, it is evident that NDT methods are extremely useful and can help in evaluating the properties of FFF-printed parts, especially when multiple processing parameters are being optimised to achieve desired results. The effectiveness of NDT methods also depend on the properties being investigated. It is not mandatory that every material property can be evaluated using NDT methods. Therefore, this work aims to analyse three NDT methods (i.e., ultrasonic testing, strain measurement, and indentation hardness testing) for their effectiveness in evaluating the tensile and flexural properties of graphene-enhanced PLA material. Three processing parameters for the material have been investigated, i.e., nozzle temperature, print bed temperature, and print speed. The work analyses the effectiveness of the NDT methods in evaluating the properties of GPLA and whether they can capture the impact of all three processing parameters. The presence of nanoplatelets in the PLA matrix can hinder proper analysis using NDT methods and this work aims to explore that avenue. The experimental methodology for this study is presented in Section 2 with descriptions of the material, standards for manufacture and testing, combinations of processing parameters, and NDT methods. The effects of the processing parameters are discussed in Section 3 using both non-destructive and destructive methods. Section 4 presents a correlation between non-destructive and destructive testing to show the applicability of the former to evaluate tensile and flexural properties for graphene-enhanced PLA. The conclusions of this work are outlined in Section 5.

Experimental Methodology
HDPlas ® PLA-GNP-A filament from 3D Haydale Ltd., (Loughborough, UK) was used to manufacture samples to analyse the effects of nozzle temperature, print bed temperature, and print speed. The introduction of the HDPlas ® functionalised graphene nanoplatelets of a planar size between 0.3-5 µm helps to improve dispersion and bonding within the PLA polymer. It provides the material with improved operating temperature performance, high rigidity, good impact strength, and excellent interlayer adhesion for smooth printing [27]. The characteristics of GPLA are shown in Table 1. Anet ® ET4 Pro desktop 3D printer was used to manufacture two sets of samples. BS EN ISO 527-2:2012 [28] was followed to manufacture dog-bone samples for tensile testing whereas rectangular samples for flexural testing were manufactured as per BS EN ISO 178:2019 [29]. The dimensions for the manufactured samples are shown in Figure 1. Ultimaker Cura 4.11.0 [30] was used to generate G-code files based on the different combinations of the three processing parameters, as shown in Table 2. Five samples for each combination of the processing parameters were tested while all other parameters were kept constant and were taken from literature [22]. The infill density was set at 100% and infill pattern of lines was used. The flow percentage and layer height were set to 100% and 0.2 mm, respectively.     The dog-bone samples were first subjected to ultrasonic testing using a Proceq PUNDIT ® PL-200 (Test Equipment Center, Zürich, Switzerland) comprising two 54-kHz transducers [31]. This method utilises high-frequency sound waves to detect flaws and defects in products. The samples were tested at three points along their length to ascertain an average value of transmission time. After ultrasonic testing, the dog-bone samples were subjected to indentation hardness testing as per BS EN ISO 868:2003 [32] using a Shore D durometer. The indentation was measured at five different points to obtain an average hardness value for all the samples. These two non-destructive tests are linked to the destructive tensile testing that was subsequently undertaken as per BS EN ISO 527-2:2012 [28] on a TIRAtest 2810 Universal Testing Machine at a crosshead speed of 1.5 mm/s according to the standard.
The NDT method of strain measurement is more suited to assess bending in the rectangular samples. Two metallic strain gauges were bonded to both sides of the samples in a half bridge configuration because it instils more sensitivity by measuring both tensile (positive) and compressive strain (negative). BF350-3 AA type metal foil resistance strain gauges were used with a resistance of 350 ± 0.1 ohms, sensitivity factor of 2.0-2.20, and precision level of 0.02. They were bonded on the surfaces of the rectangular samples with cyanoacrylate adhesive [33]. A static load of 1000 g (1 kg) was applied to the samples and the resistance values from the strain gauges were recorded using a HBM Data Acquisition System QuantumX MX 1615B system running Catman DAQ software with a half bridge configuration [34]. After strain measurements, flexural testing was undertaken on a TIRAtest 2810 Universal Testing Machine with a speed of 2 mm/min as per BS EN ISO 178:2019 [29]. The diameter of the former and supports was 10 mm, and the samples were placed in such a way to ensure a support span length of 60 mm for the test. The reason for testing commercially available GPLA is because of its unique properties and incorporation of graphene nanoplatelets that can make evaluating the mechanical properties difficult through NDT methods. The work also focuses on assessing the effectiveness of the NDT methods in capturing the impact of all three processing parameters.

Ultrasonic Testing
This NDT is useful in detecting defects in FFF-printed parts such as voids and gaps [8,12,15,22]. The presence of a void or poorly welded layer interface would result in a higher value as opposed to a sample that is properly packed with good layer adhesion. The test was conducted on all the dog-bone samples and three measurements were taken along their length. The results of the transmission times for the four different nozzle temperatures are shown in Figure 2.
It is evident from Figure

Hardness Testing
The second non-destructive test in this work is simple and inexpensive. It does not completely destroy a material such as in a tensile or flexural test. However, indentation hardness testing leaves behind marks or imprints on the tested parts. Shore hardness is a non-destructive testing method that can determine how effectively a material resists indentation, providing insight into how it will perform over time. Although hardness testing cannot usually find defects, it can show how materials are affected by stress and how components will wear during their operational life. This test can shed some light on the strength, ductility, and wear resistance of GPLA as there is a very close relationship between hardness and tensile strength [35,36]. The dog-bone samples were subjected to indentation Shore D hardness testing and the results are shown in Figure 3. Similar to ultrasonic testing (Section 3.1), there is not a significant difference in the hardness values. The maximum hardness values were observed for nozzle temperature of 190 • C (Figure 3b) whereas the lowest were shown by 210 • C (Figure 3d), indicating that the hardness values decreased with the increase in nozzle temperature. Furthermore, as the print bed temperature increased, the hardness values increased until 70 • C for 180 • C ( Figure 3a) and 200 • C (Figure 3c) before gradually decreasing. For nozzle temperature of 190 • C, the hardness values started high at the lowest print bed temperature of 60 • C before dropping at 70 • C; then, it reached a peak value at 80 • C and dropped again at 90 • C. For nozzle temperature of 210 • C, the hardness values started at their peak at the lowest print bed temperature of 60 • C before dropping until 80 • C and then rising slightly at 90 • C. It can also be seen that as the nozzle temperature increased, the difference in hardness values at different printing speeds also increased with the most deviations at different print bed temperatures being observed at higher nozzle temperatures. Similar to ultrasonic testing (UT), the print speed does not play a significant role in impacting the hardness values of GPLA and they are quite consistent at different nozzle temperatures. of 210 °C, the hardness values started at their peak at the lowest print bed temperature of 60 °C before dropping until 80 °C and then rising slightly at 90 °C. It can also be seen that as the nozzle temperature increased, the difference in hardness values at different printing speeds also increased with the most deviations at different print bed temperatures being observed at higher nozzle temperatures. Similar to ultrasonic testing (UT), the print speed does not play a significant role in impacting the hardness values of GPLA and they are quite consistent at different nozzle temperatures.

Tensile Testing
The results from the tensile testing for the different nozzle temperatures are shown in Figure 4. The highest values of load were observed at 180 • C and, as the nozzle temperature increased, the average load values decreased. Nozzle temperatures of 190 • C ( Figure 4b) and 210 • C (Figure 4d) showed a decrease in average load values as the print bed temperature increased.
For 180 • C (Figure 4a) and 200 • C (Figure 4c), the average load values increased until 70 • C and then decreased with the increase in print bed temperature. Furthermore, all the samples showed a decrease in average load values with an increase in print speed [37,38], indicating that high print speeds affect the structural integrity of the samples. Linking these results to the ultrasonic testing (Section 3.1), it becomes evident that the same conclusions cannot be drawn. Average fracture loads are clearly affected by print speed whereas UT did not capture significant differences in transmission times as a result of changing print speeds. UT should be able to evaluate the porosity of the printed samples [8,15,22] and lower porosity should lead to higher fracture load values. However, it is to be noted that the differences in the average transmission times and the average loads at different nozzle temperatures is not significantly high. This could be the reason for the UT not being able to capture the impact of changing print speeds.

Tensile Testing
The results from the tensile testing for the different nozzle temperatures are shown in Figure 4. The highest values of load were observed at 180 °C and, as the nozzle temperature increased, the average load values decreased. Nozzle temperatures of 190 °C ( Figure  4b) and 210 °C (Figure 4d) showed a decrease in average load values as the print bed temperature increased.  [37,38], indicating that high print speeds affect the structural integrity of the samples. Linking these results to the ultrasonic testing (Section 3.1), it becomes evident that the same conclusions cannot be drawn. Average fracture loads are clearly affected by print speed whereas UT did not capture significant differences in transmission times as a result of changing print speeds. UT should be able to evaluate the porosity of the printed samples [8,15,22] and lower porosity should lead to higher fracture load values. However, it is to be noted that the differences in the average transmission times and the average loads at different nozzle temperatures is not significantly high. This could be the reason for the UT not being able to capture the impact of changing print speeds.

Strain Measurement
Strain is a crucial factor in determining the strength of a material. In this study, two 350-ohm strain gauges were bonded to all the rectangular samples to measure strain upon the application of a static load of 1000 g (1 kg) at room temperature. The samples were placed on two rollers of 10 mm-diameter, and the strain values were recorded using the HBM QuantumX MX 1615B system. The upper strain gauge recorded the compressive strain whereas the lower gauge measured the tensile strain as the load was being applied in the middle of the sample where the gauges were bonded [33]. The results of the static loading are shown in Figure 5.
Two aspects are evident from  (Figure 5c) before rising sharply. Nozzle temperatures of 180 • C and 210 • C showed slightly different behaviour. With 180 • C, the lowest strain value was observed at 80 • C before rising sharply (Figure 5a) whereas 210 • C nozzle temperature showed the lowest value at the lowest print bed temperature of 60 • C and then gradually rose (Figure 5d). These results indicate that strain measurements represent the effect of the three processing parameters (i.e., nozzle temperature, print bed temperature, and print speed) more effectively compared to ultrasonic testing and hardness testing as both tests were not significantly affected by the print speed. It is to be noted that several factors could affect the strain measurements. They include the size and placement of the strain gauges as well as ambient noise and temperature. This non-destructive test forms the basis for evaluating the flexural properties of the GPLA samples.

Strain Measurement
Strain is a crucial factor in determining the strength of a material. In this study, two 350-ohm strain gauges were bonded to all the rectangular samples to measure strain upon the application of a static load of 1000 g (1 kg) at room temperature. The samples were placed on two rollers of 10 mm-diameter, and the strain values were recorded using the HBM QuantumX MX 1615B system. The upper strain gauge recorded the compressive strain whereas the lower gauge measured the tensile strain as the load was being applied in the middle of the sample where the gauges were bonded [33]. The results of the static loading are shown in Figure 5.

Three-Point Flexural Testing
This is a destructive test that measures the force required to bend a beam under threepoint loading conditions. It is used to determine the flex or bending properties of a material and for selection of parts that will support loads without flexing [6,24,39]. The results from the three-point flexural testing are shown in Figure 6. As the nozzle temperature increased, the average load required to break the samples also increased by reaching its peak at 190 • C but then falling sharply as the temperature increased. With the increase in print bed temperature, the maximum load was observed at 70 • C for nozzle temperatures of 190 • C (Figure 6b) and 200 • C (Figure 6c) before falling sharply. For nozzle temperatures of 180 • C (Figure 6a) and 210 • C (Figure 6d), the maximum load values were observed at 80 • C before falling rapidly. It is also evident from Figure 6 that as the print speed increased, the load values decreased for all the samples, indicating that high print speeds affect the structural integrity of the samples (similar to tensile testing, as discussed in Section 3.3). print bed temperature, the maximum load was observed at 70 °C for nozzle temperatures of 190 °C (Figure 6b) and 200 °C (Figure 6c) before falling sharply. For nozzle temperatures of 180 °C (Figure 6a) and 210 °C (Figure 6d), the maximum load values were observed at 80 °C before falling rapidly. It is also evident from Figure 6 that as the print speed increased, the load values decreased for all the samples, indicating that high print speeds affect the structural integrity of the samples (similar to tensile testing, as discussed in Section 3.3).

For Tensile Testing
This work focuses on analysing the impact of nozzle temperature, print bed temperature, and print speed on the tensile and flexural properties of graphene-enhanced PLA material. It also analyses the effectiveness of three non-destructive methods to evaluate the strength of GPLA. It is important to note that NDT methods should be chosen based on their applicability to assess a certain material property. In this study, the NDT methods of ultrasonic and hardness testing are linked to the tensile properties [8,12,15,22,35,36] of GPLA. These NDT methods can help assess the tensile strength of GPLA samples without breaking them. To verify their effectiveness, Figure 7 shows the correlation between the ultrasonic results and the tensile strength of all the dog-bone samples. It is evident that there is a good correlation between the two test results with the average transmission time falling with rising tensile strength [8,15,22]. This is because samples with higher strength are more closely packed and possess better adhesion of their layers that allow the sound waves to travel quickly through them. It can also be seen in Figure 7 that the transmission times did not significantly change at higher nozzle temperatures, indicating a good correlation as the same aspect was observed from tensile testing results (Section 3.3). However, it is to be noted that ultrasonic testing could not capture all the variations as the values did not show a strong correlation with print speed. Nonetheless, these results help in evaluating the tensile properties of GPLA and ultrasonic testing is a viable NDT method for this purpose.
falling with rising tensile strength [8,15,22]. This is because samples with higher strength are more closely packed and possess better adhesion of their layers that allow the sound waves to travel quickly through them. It can also be seen in Figure 7 that the transmission times did not significantly change at higher nozzle temperatures, indicating a good correlation as the same aspect was observed from tensile testing results (Section 3.3). However, it is to be noted that ultrasonic testing could not capture all the variations as the values did not show a strong correlation with print speed. Nonetheless, these results help in evaluating the tensile properties of GPLA and ultrasonic testing is a viable NDT method for this purpose. Similar to ultrasonic testing, hardness also correlates well to tensile properties of a material [35,36]. As can be seen in Figure 8, the hardness values increased with the increase in tensile strength of GPLA and vice versa. This is a useful NDT that can help ascertain the properties of a material without fracturing it. However, imprints are left after a hardness test due to the indentation and can cause issues if surface roughness is a concern. In short, both ultrasonic and hardness testing have shown their capability in correlating well with the tensile properties of GPLA as three of its processing parameters (nozzle temperature, print bed temperature, and print speed) were modified. Similar to ultrasonic testing, hardness also correlates well to tensile properties of a material [35,36]. As can be seen in Figure 8, the hardness values increased with the increase in tensile strength of GPLA and vice versa. This is a useful NDT that can help ascertain the properties of a material without fracturing it. However, imprints are left after a hardness test due to the indentation and can cause issues if surface roughness is a concern. In short, both ultrasonic and hardness testing have shown their capability in correlating well with the tensile properties of GPLA as three of its processing parameters (nozzle temperature, print bed temperature, and print speed) were modified.

For Flexural Testing
The strain values observed through the non-destructive test (Section 3.4) showed a larger range compared to transmission times and hardness values. This could be attributed to the fact that the setup for strain measurements is quite similar to the three-point flexural testing as a load, albeit static, is being applied at the middle of the rectangular sample [33]. The correlation between strain measurements and flexural strength is shown in Figure 9.
It is evident that the effect of all three processing parameters, i.e., nozzle temperature, print bed temperature, and print speed, has been captured with this correlation for flexural testing. The strain values show a clear dip upon an increase in flexural strength as samples exhibiting higher resistance to bending show lower strain values [40]. It can be observed that all the strain values match very well with the flexural strength values for GPLA. They fall with an increase in flexural strength and rise with a decrease in flexural strength, indicating that this NDT method has captured the impact of all three processing parameters better than the methods used to evaluate the tensile properties of GPLA.

For Flexural Testing
The strain values observed through the non-destructive test (Section 3.4) showed a larger range compared to transmission times and hardness values. This could be attributed to the fact that the setup for strain measurements is quite similar to the threepoint flexural testing as a load, albeit static, is being applied at the middle of the rectangular sample [33]. The correlation between strain measurements and flexural strength is shown in Figure 9.

For Flexural Testing
The strain values observed through the non-destructive test (Section 3.4) showed a larger range compared to transmission times and hardness values. This could be attributed to the fact that the setup for strain measurements is quite similar to the threepoint flexural testing as a load, albeit static, is being applied at the middle of the rectangular sample [33]. The correlation between strain measurements and flexural strength is shown in Figure 9.

Conclusions
Optimisation of process parameters is critical for the fused filament fabrication process to achieve desired results. It is important to understand how different parameters interact and the resulting material properties. One way to evaluate such properties is to use non-destructive testing methods and verify their effectiveness through correlation with destructive testing. In this work, such a correlation has been presented for grapheneenhanced PLA material manufactured using the FFF process at four different nozzle temperatures (180 • C, 190 • C, 200 • C, 210 • C), four different print bed temperatures (60 • C, 70 • C, 80 • C, 90 • C), and four different print speeds (50 mm/s, 60 mm/s, 70 mm/s, 80 mm/s). The presence of graphene nanoplatelets in the PLA matrix can adversely affect the effectiveness of the NDT methods in evaluating the mechanical properties of GPLA with changes in the processing parameters. The results from the mechanical testing showed that the tensile strength of the GPLA samples decreased with an increase in nozzle temperature. Nozzle temperatures of 190 • C and 210 • C showed a decrease in tensile strength as the print bed temperature increased. For 180 • C and 200 • C, the tensile strength increased until 70 • C and then decreased with the increase in print bed temperature. Furthermore, all the samples showed a decrease in average load values with an increase in print speed. Flexural testing showed that as the nozzle temperature increased, the flexural strength increased by reaching its peak at 190 • C, but then falling sharply as the temperature increased. With the increase in print bed temperature, the maximum flexural strength was observed at 70 • C for nozzle temperatures of 190 • C and 200 • C, before falling sharply. For nozzle temperatures of 180 • C and 210 • C, the maximum flexural strength was observed at 80 • C before falling rapidly. Similar to tensile testing, the flexural strength decreased with an increase in print speed.
NDT methods of ultrasonic and hardness testing have been correlated with the tensile properties of GPLA. Both the NDT methods exhibited a limited range of values and did not show significant variations due to changes in print speed. However, their correlation was still valid and highlighted their applicability to evaluate tensile strength obtained through destructive tensile testing. The NDT method of strain measurement showed a wide range of values corresponding to the changes in the three process parameters. These measurements also showed a better correlation with the destructive three-point flexural test because the application of load, albeit static, was similar to how the load is applied during flexural testing. The results presented in this work show a good correlation between non-destructive and destructive tests to highlight the effectiveness of these practices in evaluating the properties of different materials manufactured using the fused filament fabrication process. They also indicate that the presence of graphene nanoplatelets in the PLA matrix did not adversely affect the results of the NDT tests and such tests can be used to evaluate the properties of composite FFF filaments effectively. Furthermore, these results hold significant scientific, technological, and industrial merit as they can help manufacturers identify and understand how different processing parameters interact and which NDT methods are effective in analysing the required material properties. Furthermore, the quantitative values for the mechanical properties of GPLA can support complex simulations for the optimisation of intricate geometries to be printed.