Surface plastic flow of three-dimensional printed polylactic acid in the tribological study of surface patterned polymer

The potential of three-dimensional (3D) printing in polymer tribology is poorly explored. Material alignment and composition play vital roles in altering the friction and wear characteristics of 3D printed materials. In the current study, 3D patterns made by fused deposition modelling are used to print advanced tribo-composites. Two different surface patterns (line and circular) are provided through 3D printing using white and silver polylactic acid (PLA). The deformation and distribution of white and silver PLA over the samples surface are observed after a wear test. Results showed that the coefficient of friction is not influenced by changes in the surface pattern. However, the wear rate increased for samples with line patterns on the contact surface, since plastic flow was more significant in this case. Moreover, the filling factor exhibited an influence on increasing the plastic flow of the contact surface for samples with a line pattern.


Introduction
In the past few decades, interest in polymer composites has greatly increased for several tribological applications. Tribology in polymers is a complicated topic, since friction and wear behaviour are highly influenced by the type of polymer and its additives used in the tribo-system. The main advantage of using a polymer is its potential for formation of a transfer Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. film (TF) on the counterface. A TF provides a protective layer to a hard counterface, preventing severe damage to the soft polymer as well as helping to stabilize the coefficient of friction (CoF). This protective layer acts by covering the hard asperities of the counterface [1]. However, although a TF is beneficial in some cases (protective/lubricating) it can be detrimental (due to high adhesion) in others. The process of formation and the characteristics of TFs have been previously studied by several authors [2][3][4][5][6]. The proper selection of additives in polymers has become a common practice for manufacturers when high wear resistance is desired in polymer tribo composites (PTCs). Other additives, such as polytetrafluoroethylene (PTFE), bronze and MoS 2 , are commonly used as solid lubricants. The effect of addition on the bulk material may lead to a reduction in the mechanical strength of the polymer composite.
The top surface of the bulk material is responsible for making contact with the counterface. For this reason, the characteristics of the top surface are related to the tribological properties of the polymer composite. In general, the additives are added to the complete bulk material rather than just to the top surface. However, some researchers (e.g. [7,8]) have come up with solutions for adding the additives only to the top surface. Subramanian et al [8] fabricated what they called hybrid polymer tribo composites (hPTCs) of basalt-polyethelene with a PTFE blend on the top surface. According to the authors, the tribological performance of the hPTC matched that of commercially available PTCs used in large-scale bearing applications, with the benefit of not losing structural strength [8]. Producing PTCs using conventional technologies such as hand laying is rather well established. However, the process by itself is tedious, and very limited combinations can be made using those methods. Besides, conventional manufacturing methods do not properly place the additives in the desired locations on the PTC bulk material.
One of the recent and advanced methods for producing hPTCs is three-dimensional (3D) printing. This technique has the advantage of placing a particular material (additive) precisely in the matrix at predetermined locations. 3D printed plastics are found worldwide since they can be used in different tribological applications, ranging from domestic products to medical implants and engineering components. Besides the potential for multi-material composition, optimization of the topography and functionally graded structures are benefits of 3D printed plastic, which allows weight reduction without loss of mechanical strength [9]. There are a few 3D printing techniques that can be used for multi-material 3D printing: fused deposition modelling (FDM); selective laser sintering (SLS; solidification of powders); stereo lithography (SLA; solidification of liquid in a bath); polyjet multihole head inkjet 3D printing; and laminated object manufacturing [10][11][12][13][14].
So, designing a hPTC with good tribological and mechanical properties may be a challenge. Bearing this in mind, the present work offers a more comprehensive study of the tribological and surface behaviour of macro-geometrical structures (3D placed multi-materials) for 3D printed polymers. Two different surface patterns (dots and lines) and their corresponding deformation/modification after wear tests were investigated. Mechanical and tribological characteristics of the hPTC materials selected in this study are discussed throughout the paper.

Materials
The most commonly used FDM method for 3D printing was used to print polylactic acid (PLA). In the current research, two different coloured materials (PLA extrafill traffic white and PLA silver) were used. PLA filament is made of natural ingredients and is easily biodegradable. Its material density is 1.24 g cm −3 and its glass transition temperature is 55 • C-60 • C based on the datasheet provided by the manufacturer (Prusa Research a.s.). Nevertheless, the literature [15][16][17][18] highlights that the mechanical properties of the finished product are highly dependent on the machining parameters. In the case of the tribological test specimens, all standard test specimens (EN ISO 20753:2014 A2 test specimen, figure 1) were prepared with the same parameters. The parameters of the process were 0.4 mm diameter extruder nozzle, 0.2 mm layer thickness, 0.8 mm contour width, 100% fill and 45 • raster angle with the orientation flat on the build platform, parallel to the x-axis.
The tensile tests were performed using a Zwick Roell Z100 Tensile Test Machine corresponding to ISO 527-1:2012 standard (tensile speed 10 mm min −1 ). The tensile results were evaluated using the textXpert software. The results (average of three tensile tests) were compared with the material data sheet from the manufacturer (see table 1). Since the tensile modulus could not be clearly determined on the measurement curves of the specimens, it was calculated from the slope of the line fitted to the points of the curves measured at 0.0005 and 0.0025, as recommended by the standard. However, due to the manufacturing technology, some air gaps always remain between the plastic strands, creating voids in the cross section. To calculate the tensile strength, the nominal cross section was used, as the rate of these gaps is relatively small.

3D printing
For the production of the test specimens, a Prusa i3 MK2S printer (Prusa Research a.s.; figure 2) was used with Multi Material 2S, enabling us to use several materials at once to orient a single composite. The system is operated with a single extrusion head where the material is automatically replaced. The machine lays the required amount of the first material per layer and then unloads the filament from the extrusion head, loading the second material. Then a small amount is extracted with the wipe tower, which ensures that the residue is removed from the first material and a smooth flow is formed from the second material. A complete replacement cycle takes about 45 s. The printer always starts printing with a larger mass of material, in this case with the silver material. Later, the process optimizes itself, so it can start the next layer with the material it used to finish the previous layer.

Test specimens
Cylindrical specimens with a diameter of 16 mm and a height of 24 mm were printed for the pin-on-disc tests. Two surface patterns, lines and dots, were prepared with different filling factors. It is worth mentioning that the study aims to better understand the plastic flow of the multi-material selected in this investigation, comparing its integrity when interacting with the steel counterface. This means the distribution of both silver and white PLA in the contact region will be mapped after the wear tests. This will allow us to clearly understand the plastic flow behaviour for different surface patterns. The patterns are formed on the top, flat surface of the specimens with a depth of 3 mm. In the case of line patterns, filling factors of 25%, 50% and 75% was established. For example, a filling factor of 25% means that the white PLA covers 1/4 of the top surface, whereas the remaining 3/4 is silver PLA. The dotted surface pattern was prepared with filling factors of 25%, 37% and 50%. Pure materials were also evaluated for reference purposes, that is, specimens with only white PLA (W specimen) or silver PLA (S specimen). All specimens were rubbed against a metal disc made of SAE 1020 steel through a tribometer. A description of the test rig is provided in the next section. This disc has a diameter and thickness of 160 mm and 12 mm, respectively (figures 3(b) and (c)).
A K-type thermocouple was embedded at 1.5 mm from the disc surface in order to monitor the disc temperature during the wear tests. Figure 2 shows an example of both specimens (pins), as well as a photo and a drawing of the metal disc (counterface) used in the wear tests. Table 2 summarizes all the specimens selected in this study. In order to simplify the nomenclature of the specimens, the letters L and D are used to represent the surface patterns of lines and dots, respectively. A number along with these letters is used to identify the respective filling factor of white PLA. For instance, L25 represents the specimen with a surface of line patterns containing 25% white PLA.
Previous investigations [19][20][21][22][23] examined the effect of surface texture on the friction and wear of polymer materials. Those studies reported a strong influence of the type of textured surface on tribological results. By 'textured surface' those researchers defined a surface that was not smooth but which had a particular texture, or a combined form (e.g. in figure 4(a)). However, this is not the case for the specimens selected in the present study, where a surface pattern was used rather than a textured surface (e.g. in figure 4(b)).

Set-up of the wear tests
The tribological tests were performed using a pin-on-disc tribotester (UFRGS, Brazil) ( figure 5). This test rig permits operation over a wide range of temperatures (up to 600 • C), sliding velocities (from 1.8 to 16 000 mm s −1 ) and normal forces where T is the torque measured in the machine's shaft (N m), F N is the normal force (N) and R is the sliding radius (m).
Precision/resolution of the CoF measured by the tribotester is ±0.013, as shown in [24]. The wear tests consisted of the continuous sliding of the specimen (which works as a pin) against the steel discs. They were carried out at constant sliding velocity and normal force. The wear tests performed with each surface pattern were repeated three times. In order to avoid thermal degradation of the polymer, each test corresponded to 10 runs of 180 s each, carried out under air cooling. By doing this, the maximum bulk temperature measured beneath the surface of the counterface did not exceed 26 • C. The data acquisition sampling rate was set to 200 Hz for all tests. Table 3 summarizes the operating parameters of the wear tests.
Wear of the specimens was measured from the mass loss by subtracting the final mass of each specimen after the wear tests from the initial mass measured before wear test. An electronic balance with a precision of ±0.1 mg was used for weighing the polymer test specimens. Disc wear is not discussed in the present paper, since this metal counterpart did not show significant wear during the tests. Moreover, images of the surface of the counterface (disc) were captured using a 15 megapixel digital camera (Canon) in order to assess the formation of a polymer TF on the counterface.
Macrographs of the worn surfaces of the specimens were obtained through a stereoscope (Zeiss, model Stemi 508). Then, those macrographs were processed using a specially designed algorithm (in MatLab Mathworks®) in order to   The analysis was based on the area of white PLA that was displaced from its original position due to distortion ( figure 6). Then, the area missing from the initial (or original) position is divided by the total area of the image (11 mm × 7 mm), providing a result that represents the fraction of the area (in %) which left the original position. This is a way to quantify the 'integrity' of the specimen, i.e. the change in geometry caused by plastic deformation. Figure 7 shows the average results of wear and CoF for all specimens with different surface patterns. Analysis of variance (ANOVA), with a significance level α = 0.05, was performed to evaluate the effect of the surface pattern and the percentage of white PLA on friction and wear. CoF was relatively stable for all three repetitions (low standard deviation, expressed by the error bar). ANOVA did not show any significant difference in CoF between the specimens with different surface patterns and percentage of white PLA. For the dot pattern, the increase in the filling factor (white PLA) showed a subtle increase in friction. But this behaviour was not seen in  the line pattern specimens, where sample L25 exhibited the highest friction. As presented in table 4, the Pearson correlation coefficient seems to indicate that there is a strong correlation (1) between the mean CoF and the filling factor for the dot pattern and a weak correlation (−0.5) for the line pattern. However, the difference between the average CoF values is not significant because it is within the precision/resolution of the tribotester. Regarding wear, it was observed that wear of the pure white PLA (W) specimen is significantly lower than that of the other specimens. ANOVA also showed that there are no statistical differences in wear among all other specimens (S, L25, L50, L75, D25, D37 and D50). This means that different surface patterns caused no significant difference in wear under the applied conditions. It is noteworthy that the dynamicity involved in the TF characteristics may bring uncertainty in long-duration tests. Hence it is important to investigate the wear process to understand how this affects the integrity of the polymer material (deformation and plastic flow at the surface). Aziz et al [19] examined the friction behaviour of PLA composites made by FDM with three circular surface textures (varying the diameter) through the pin-on-disc test (ASTM G-99 standard). The textures studied by those researchers were  at a depth of 1 mm from the top surface. The authors reported different friction behaviour for the different textures with the intermediate diameter exhibiting the lowest friction. This is different from what was observed in the present study, where CoF was similar among the different surface patterns. The authors explained that the lowest friction in case of the intermediate diameter was due to the formation of a more efficient TF on the disc. Figure 8 exhibits the surface of specimens D25 and L25 for the conditions before and after the wear tests. Overall, those surfaces represent the behaviour seen on the surface of the other specimens. There is some clear plastic deformation or plastic flow on the worn surfaces of the specimens. This can be mainly attributed to the shear force caused by the sliding condition. Considerable scratch marks or ruptured areas could not be observed. This means that abrasion was not a significant wear mechanism in the present study. The plastic flow on the polymer surface originated from a weak mechanism of adhesion. The interfacial bonding between polymer and metal was not strong enough to break the cohesive strength of the polymer, generating deformation instead of ploughing or pulling out material. Aziz et al [19] found that the wear of PLA composites is mainly due to adhesion and abrasion, and that there is no ploughing due to effective TF formation between the surfaces, caused by the increase in temperature. Figure 9 presents the results of the distortion (due to plastic flow) on the surface of the polymers. Comparing the dot and line patterns, distortion appears to be prominent for the latter (line pattern). L50 exhibited the highest percentage of polymer outside the original area (5.3%). Comparatively, the highest distortion for the dot pattern was 2.5% (equally for both D25 and D50). Since the lines are perpendicular to the sliding direction, the white PLA tends to flow outside the original region when sheared. This results in a larger distortion when compared with the dotted surface ( figure 10). On the other hand, dot patterns are more spread out all over the surface. For instance, specimen D50 has 19 dots of different sizes with white PLA. This makes distortion difficult and the structure of the specimen becomes more consistent. In other words, distortion is favored for line patterns since white PLA is concentrated in fewer regions (four lines). So, when a specific location on the surface begins to suffer plastic deformation, it influences the nearby regions to suffer the same effect. Also from the viewpoint of filling factor, a 50% mix provides a higher plastic flow. Figure 11 shows the disc surface after wear tests. The formation of two different types of TF was observed for all specimens: a heterogeneously ( figure 11(a)) and a homogeneously distributed film ( figure 11(b)). This TF is formed due to an adhesive wear mechanism, where material from the polymer    is torn off from the pin and 'sticks' on the metal counterface (metal disc). The heterogeneous TF is characterized by lumps of polymer material poorly distributed on the disc surface. The homogeneous TF, on the other hand, is a more concise and well-distributed layer of polymer material on the disc surface. Table 5 shows the two typical types of TF provided by each specimen on the disc surface. No correlation was found between CoF and wear with the TF. This contradicts the research finding of Aziz et al [19], where the formation of a more efficient TF on the disc led to a lower CoF. As a non-significant difference was seen in terms of mass loss for the specimens, a second test was performed with more severe conditions (normal load = 120 N, sliding velocity = 1.5 m s −1 ). In this case, the PLA specimen suffered considerable plastic deformation, with severe damage ( figure 12), in just a few runs of 180 s each. This is associated with the weak resistance of the PLA to the shear force caused by the sliding. The low shear resistance may be due to the presence of voids in the material, which is an effect commonly reported in 3D filament printing processes [25,26]. So, on the one hand the operating parameters used in this work (see table 3) caused little wear. On the other hand, a slight increase in load and velocity caused the polymer material to suffer severe damage. It is evident that none of the specimens failed from delamination. Hence, the PLA specimens selected in this study seem to be good at preventing delamination and they rather fail due to relatively weak mechanical properties, i.e. severe damage on the surface caused by plastic flow.

Conclusions
The current research addresses the tribological performance of 3D printed multi-material PLA with specific surface patterns (line and dots). The following outcomes from the current study can be listed: • Neither the line pattern nor the dot pattern made any significant difference to the tribological properties (friction and wear). • A homogeneous TF, which is often preferred, dominated in both line and circular surface patterns. However, this was not the case for the reference materials. Hence, the pattern clearly facilitates a homogeneous TF. No correlation was seen between CoF and wear with the type of TF. • Surface plastic flow is the preferable form of failure for all the specimens studied in the present work. Comparing the line and dot surface patterns, the latter showed more resistance to plastic flow. This occurred because dot patterns are more spread out all over the polymer surface, which makes distortion difficult and the structure of the specimen becomes more concise.

Data availability statement
The data that support the findings of this study are available upon reasonable request from the authors.