Mechanical Performance Under Various Conditions of 3D Printed Polylactide Composites With Natural Fibers

In the study, polylactide-based (PLA) composites modied with natural particles (wood, bamboo, and cork) and with different levels of inlling (100%, 80%, and 60%) obtained by additive methods were tested. The effect of type ber, inll level and crystallization rate on the mechanical properties were investigated by using tensile, exural, and impact tests. The materials were subjected to mechanical tests carried out at 23 and 80 °C. Furthermore, hydrothermal degradation was performed, and its effect on the properties was analyzed. The addition of natural llers and different level of inlling result in a similar level of reduction in the properties. Composites made of PLA are more sensitive to high temperature than to water. The decrease in Young's modulus of PLA at 80 °C was 90%, while after 28 days of hydrodegradation ~ 9%. The addition of bers reduced this decrease at elevated temperatures. Moreover, the impact strength has been improved by 50% for composites with cork particles and for other lignocellulosic composites remained at the same level as for resin.


Introduction
For more than three decades, there has been a substantial increase in interest in products developed using additive manufacturing (AM). Presently, AM has revolutionized the manufacturing industry as it enables rapid manufacturing of products while simultaneously reducing cost [1] [2]. One of the most commonly used AM techniques is fused deposition modeling (FDM), which involves successive overlapping of layers of laments; the production occurs without additional processing of the material, which allows to reduce waste during production [3]. This technology has enabled to develop prototypes with complex shapes, which is otherwise di cult to obtain by traditional techniques (injection molding) [4]. However, the production by AM requires a continuous ow of materials, which can be disturbed during the lament application process. The lack of isotropy of products can signi cantly affect the mechanical properties of the tested materials; hence, new solutions to improve this method are being constantly investigated, such as nozzle and platform temperature, printing speed, height of layer deposition, raster angle of in ll pattern, build orientation, and addition of particles [5] [6]. To date, many works has been published on the relationship between the lament system and each other 0°/90° or 45°/45° [7].
Depending on the position of the laments in relation to each other, the materials exhibit different properties, for example, materials with a 45°/45° lament system have higher values of mechanical properties than those with a 0°/90° lament system [8]. By optimizing the methods at the production level, it is possible to design a product with superior properties.
The development of AM also led to the development of other agents blended with the materials used in this technique. The most commonly used materials in FDM are thermoplastic polymers: acrylonitrile butadiene styrene, polyetherimide, polylactide (PLA), polycarbonate, and polymethylmethacrylate [9] [10].
Because of concerns related to environmental pollution, PLA is gaining popularity as an alternative to synthetic polymer materials. PLA is a biodegradable, biocompatible polymeric material. This material is not only environmentally friendly, but it also has properties comparable to those of petrochemical polymers [11]. In addition, its melting temperature is relatively low, which makes its use in FDM costeffective.
To change characteristics and additionally to reduce costs, composites with the addition of bers or particles are often formed to reduce the production cost of the material, as is the case in traditional processing methods. Modi cations of laments can also be carried out in AM. It is reasonable to reinforce PLA with natural llers containing lignocellulose such as wood, hemp, ax etc. (the material remains 100% biodegradable). According to previous literature, the addition of natural llers improves the stiffness of composites while reducing tensile or exural strength [12]. However, the addition of lignocellulosic llers to polymer composites generates a material with a relatively high value of mechanical properties to density, while further reducing its cost.
Despite the many advantages of biodegradable polymers, during their processing and service life some disadvantages appear. Biodegradable polymers have faster water absorption capacity than petrochemical polymers. The penetration of water inside the biodegradable polymer composite affects its structural stability, leading to breakage of the polymer chains [13]. This process is further accelerated by factors such as temperature, time of exposure to water and modi cations [14]. These factors are mainly important for materials intended for medical applications (surgical screws, blood vessel prostheses, sutures, bone anastomoses, etc.). Because PLA is widely used in the eld of medicine, it is crucial to conduct tests that measure the impact of the aquatic environment on PLA-based materials. Studies have shown that the level of water absorption increases and the value of mechanical properties decreases with increased duration of contact of composites with the aquatic environment. It was also observed that as the water temperature increases, the water absorption rate of the material increases, which causes a corresponding decrease in the value of mechanical properties [15]. Moreover, because bers such as lignocellulose are hydrophilic in nature, the addition of llers containing these bers accelerates the process of water absorption into biodegradable composites [16]. To date, only a few studies have been published on the negative impact of water on the mechanical properties of 3D-printed polymer materials, including factors such as the degree of crystallization, degree of in lling, and addition of natural llers [17].
In the present study, PLA-based composites reinforced with natural particles (wood, bamboo, and cork) were prepared using 3D printing. In addition, neat PLA materials with varying degrees of in lls (100%, 80%, and 60%) were prepared. The novelty of the present study is the comparison of the impact of the addition of natural llers and the different degree of in lling on the mechanical properties of composite materials in various operating environments such as ambient temperature, elevated temperature (80°C), and immersion in saline solution at 38°C. In addition, two states of composite materials were tested: semi-amorphous and crystalline. For this purpose, various mechanical tests (tensile strength, bending, and impact) were conducted. These studies are important for the development of the eld of materials science and manufacturing techniques. To the best of the authors' knowledge, no direct comparison of all the above mentioned aspects has been performed; thus, the present study may further expand the area of potential applications of these composites. All mechanical tests were performed under vary testing condition:

Composites laments and sample preparation
-at ambient temperature (+ 23°C ± 2), which was re ected to the control group; -at elevated temperature (+ 80°C) in the temperature chamber (Instron), to asses impact of high temperatures.
Additionally, samples for mechanical tests were heat treated to increase the degree of crystallinity. In this process samples were placed in a furnace at 85°C for 60 min between two glass plates to prevent their deformation. This process is justi ed in the case of PLA because it has a semi-amorphous state during production. Upon heating the material above cold crystallization temperature and its slow cooling, a crystalline phase material is formed.

Hydrolytic degradation
For the calculation of water absorption, samples were immersed into the saline solution (distilled water with 0.9 wt.% of NaCl) at 38°C (physiological temperature) to create human body environment. The measurements were carried out in accordance with the modi ed ISO:62 standard, which was adopted as hydrothermal aging environment, modifying the temperature and water speci cations to the desire aging condition. The samples were kept in saline for 1, 7 and 28 days. The water absorption rate of the samples was calculated by the following equation: percentage of solution content [%] = (W t − W 0 )/(W 0 )×100, where W t stands for the instantaneous weight of the sample and W 0 for the initial weight of the sample.
To determine the in uence of water uptake on the basic mechanical properties of the specimens which were incubated for 28 days in water, a tensile test was performed. This study was conducted to consider to two conditions: physiological temperature and water absorption.

Morphological study
The fracture morphology of all the samples after tensile test were evaluated by scanning electron microscopy (SEM) (JEOL JSM5510LV,Tokyo, Japan) operating at 20 kV in low vacuum. Prior to imaging the specimens were gold-sputtered by using an auto vacuum coater (Cressington,Watford, UK). This study gave information about ller distribution and showed main characteristics of the lament orientation on breakthroughs.

Statistical analysis
At least ve measurements for each material were performed and average values and the relative standard deviation were calculated. The statistical analyses with Analysis ToolPak in Excel (2016) such as one-way analysis of variance (ANOVA) and subsequent t-test at signi cant levels of p < 0.05 were considered to be statistically signi cant. Table 1 summarizes the used abbreviations, density and speci c properties of PLA composites. It seems obvious that with a decrease in the degree of lling in materials, its density decreases. This relationship was also observed in these studies, where the following trend has been noticed:

Physico-Mechanical properties
PLA100%>PLA80%>PLA60% and PLA/C > PLA/W > PLA/B. It should be emphasized that the addition of bers had a greater impact on the decrease in density than the lower degree of PLA lling. The density of the tested composites decreased with the addition of natural bers due to the closed morphology of lignocellulosic bers and its lower density compared to the matrix [18]. In addition, as can be seen in the SEM pictures (Fig. 6), the composites are characterized by insu cient adhesion between the ber and the matrix, which increased the number of voids in the material -higher porosity. The effect of this phenomenon is a decrease in the volume fraction of neat polymer and thus decrease the density of composites. Moreover, AM itself causes an increase in discontinuities between individual layers and the extra inclusions increase them even more. However, it should be emphasized that these defects did not affect the repeatability of the results, as the standard deviation was less than 5%.
In the case of composites dedicated engineering applications, not only a single property, but a ratio of two dependencies is taken into account during the selection of the material. Because today's market expects materials of high stiffness and relatively low weight, in this work the ratio of Young's modulus value to density was determined. The addition of natural llers decreased the speci c values of the tested composites. The highest speci c modulus was observed for neat PLA (3.02 MPa/(kg/m 3 ); density 1.23 g/cm 3 ); however, a slightly lower value (2.07 MPa/(kg/m 3 )) was recorded for PLA/B where the lowest density was met (1.13 g/cm 3 ). The decrease in the speci c properties for composites with the addition of lignocellulosic bers was lower than 50% and amounted to: PLA/B -31%; PLA/W -37% and PLA/C -48%. Despite similar density values of composites lled with natural bers and with vary in lling levels, PLA80% and 60% showed a decrease of more than double in speci c modulus (difference in density > 8%). In this case, the result was affected by a signi cant decrease in Young's modulus, caused by discontinuities in the material that generated additional stress concentrations areas.

Tensile properties
The selected stress-strain curves for the tested materials are shown in Fig. 2 (a) composites tested at ambient temperature and (b) at elevated temperature (80°C). At 23°C in the rst stage, the strain was proportional to the stress (Hooke's law) up to the yield stress. At both 23°C and 80°C, neat PLA100% showed the highest stress. At 80°C, a signi cant increase in elongation was observed up to over 30% (PLA 100%), and the stress decreased by more than half compared to that at ambient temperature. This effect was caused by the test temperature above T g of PLA (~ 60°C) where the amorphous regions experience transition from rigid state (solid) to more exible (rubbery) state [19]. With an increase in temperature, the ability to move the polymer chain increased, which resulted in signi cant elongation and reduction in strength values. At high temperatures, the polymer chains are reorganized and arranged along the load. Consequently, the material became more susceptible to plastic deformation [20].
Natural llers had a more negative effect on the values of mechanical properties than lling levels at 80°C. This relationship was most likely caused by partial crystallization of neat PLA during the test (higher crystallization rate, higher mechanical properties). Neat semi-amorphous PLA (immediately after production) has greater ability to form crystals during the temperature annealing process than composites with natural bers, as reported in literature [21]. In addition, as shown in Fig. 2b, no visible differences were noted between neat PLA after heat treatment and neat PLA without heat treatment (as observed at 23°C); this nding also con rms the possibility of the crystallization process for semiamorphous samples during tests conducted at 80°C.
To better discuss the results  In general, three factors have impact on the results obtained during the tensile tests of ber reinforced composites: initial ber/matrix strength, ber length, and ber/matrix adhesion. In the case of ber reinforced composites, it is expected that the bers which usually have a higher stiffness than the polymer matrix will carry the load applied to the matrix. However, the condition of su cient ber/matrix adhesion must be met. As was already reported, the addition of unmodi ed natural bers is leading to decrease in strength [22]. In this study, insu cient ber/matrix adhesion con rmed by SEM (Fig. 6) and empty spaces in the material had a negative effect on the mechanical properties. The rst reason is due to the hydrophilic nature of the bers and the hydrophobic nature of the matrix, which prevents the formation of strong bonds between the components. Furthermore, during the FDM process, the incorporation of bers creates higher empty spaces between the applied layers [23].
The modi cation of PLA also had a negative effect on the stiffness of the tested composites. Lowering the degree of lling in composites had a more negative effect on the Young's modulus values than the addition of natural bers. Moreover, in the case of Young's modulus, the differences in decreases in the values to pure PLA were higher than in the case of tensile strength. The following decrease was observed: PLA/B (38%) > PLA/W (41%) > PLA/C (51%) > PLA80% (58%) > PLA60% (61%).
As PLA is known to occur in various forms from amorphous to the crystalline phase, the more the amount of the crystalline phase, the higher is the value of the mechanical properties of materials. Consequently, various methods have been used to increase the degree of crystallinity: addition of llers or heat treatment [24]. Therefore, heat-treated materials were also considered in the study to achieve higher mechanical results.
As shown by previous research of the authors of the presented work, the PLA after annealing (heat treatment) increases the amount of the crystalline phase and thus increases the value of the mechanical properties [25]. The highest improvement in properties was recorded for neat PLA (over 35% for tensile strength). After the thermal treatment, neat 3D printed PLA had higher values of mechanical properties as those of PLA produced by traditional methods (injection molding) [26]. The addition of natural bers did not signi cantly change the tensile strength after crystallization (max. improvement was 10%). This shows that despite the increase in the degree of crystallization, the negative effect of natural llers still persists. This occurs due to the formation of stresses at the borders of the matrix and ller. Moreover, it is possible that this is by reason of increasing loss of adhesion due to shrinkage of the PLA matrix during crystallization, which could increase the lack of ber/matrix adhesion.
In addition, elongation at break after crystallization was reduced more than doubled. The reduction in elongation indicates an increase in material stiffness, which is con rmed by the improvement in Young's modulus value. All tested materials showed an increase in the modulus of elasticity for heat-treated materials. However, it should be noted that the higher improvement for tensile strength than for Young's modulus was registered.
Negative effects of elevated test temperatures were observed in the present study. Neat PLA showed a decrease in tensile strength by approximately 50%, while composites with the addition of bers recorded a decrease by 3 times (from 24.7 ± 0.5 MPa to 7.9 ± 0.5 MPa for PLA/W). The decrease in Young's modulus was even higher, with approximately 90% for neat PLA and approximately 80% for composites. Because of the relatively low T g , the values of the mechanical properties of PLA decreased drastically as the material became more plastic. At high test temperatures, the increase in the crystalline phase did not affect the results.
The results show that both the introduction of natural llers and the reduction of the lling density have a similar effect on the mechanical properties of 3D printed polymer composites. Although a decrease in mechanical properties was observed, it should be noted that the results are comparable with other commonly used composites (polypropylene, polyethylene, etc.) produced by injection molding [27][28]. In all these cases, the main advantage of the produced composites is the reduction of polymer matrix content, which constitutes both an economic and an environmental bene t. In addition, the use of bio-based materials in combination with 3D printing techniques (waste reduction) further increases the above-mentioned bene ts.

Flexural properties
The properties obtained during the three-point bending test are very important, because they combine different stresses, i.e. compression and tension. Phenomena occurring during the bending test depends on the position: the upper layer is subjected to compression while the lower layer to tension. Therefore, in the bending test, good ber/matrix adhesion is not as signi cant as the ber orientation. The more parallel the bers are arranged in the matrix, the higher the exural properties.
Similar relationships as in the tensile test were noticed in the three-point bending test (Fig. 3). However, the results obtained during the bending test are twice as high as during the tensile test. This fact is related to the aforementioned mechanisms that occur during tests. In addition, as can be seen in the SEM pictures (Fig. 6), the bers are distributed mostly parallel in the matrix, which means that the bers are more resistant to applied load. A higher decrease in properties was noted for composites with a lower degree of lling, for exural modulus it was up to 3 times lower. At elevated temperatures, the decrease in all values remained at the same level for each of the composites. This indicates a greater in uence of temperature on mechanical values than the amount of lling or type of ller.
The in uence of the degree of crystallization of composites is also visible. Along with the increase in the degree of crystallization, the mechanical properties increased at 23°C, the highest differences can be seen in the case of neat PLA, the increase was about 20%. At 80°C composites reported lower properties after heat treatment than before. As with tensile properties, this is due to the formation of crystallites when tested at high temperatures.

Impact properties
For ber-reinforced composites during loading, the impact strength component is affected by ber pull out or breaking of bers. As reported in the literature, composites with longer bers that pull out have higher impact strength values, while composites with short bers have signi cantly reduced impact strength values [29]. In composites with the same ber content, shorter bers occur more than long ones, which implies means that short bers generate more stress concentration areas at the ber ends. Impact strength results for composites were presented in Table 3. In the present study, composites with the addition of wood and bamboo bers showed similar values of mechanical properties as those of pure PLA; this nding indicates no negative impact of natural bers on the impact properties. According to the literature, the addition of natural llers to polymer composites results in a reduction of impact strength [30]. It should be emphasized that in this paper the results of impact strength are higher or at a similar level as the matrix. The impact strength of PLA/C composites was improved by more than twice (from 10.87 kJ/m 2 to 22.1 kJ/m 2 ).
Similar to the values of properties obtained during the tensile and bending tests, the impact strength values increased after thermal treatment. The most obvious gain effect caused by the increase in the amount of the crystalline phase was noted for neat PLA 100%.
The reduction in the degree of lling decreased the impact strength value by approximately 30% for both PLA 80% and 60%; however, no signi cant differences in the results (p > 0.05) were noted for the different levels of lling (80% and 60%).

In uence of hydrolytic degradation on mechanical properties
In the eld of biodegradable polymers, hydrolytic degradation and its impact on changing mechanical properties of materials have a very important role. Several factors determine water absorption of composite materials: type of polymer matrix, type and content of llers, adhesion matrix/ ller, and test temperature [31]. In this study synergic effect of water and temperature was studied. Figure 4 shows the change in weight of the composites during immersion in saline at 38°C.
As observed for modi ed composites, the highest mass increase occurred in the rst stage; subsequently, saturation occurred after 7 days (when the daily weight gain of the samples was less than 0.01%), and the mass remained constant for up to 28 days.
The mechanism of water absorption in these studies is not in accordance with Fick's law of diffusion.
There is no three-stage sorption course distinguished: rapid increase in the rst phase and sorption release followed by saturation [32]. This is most likely caused by capillarity phenomena, which was caused by ber swelling. In addition, as other researchers indicate, this phenomenon may be associated with the leaching of polymer particles, which leads to weight loss [33].
The highest water absorption was shown by PLA 60% (8.1%) and PLA 80% (7.8%). This was because of a lack of material continuity due to the bulk region in the material, as observed in Fig. 6. PLA 100% showed the greatest density of material and hence the lowest water absorption value (~ 1%). Because PLA is hygroscopic in nature, it can absorb approximately 1% water horizontally [35]. According to previous literature, lignocellulose-containing additives are susceptible to water absorption due to their hydrophilic nature (presence of many hydroxyl groups). The results of the present study also con rmed this fact. PLA/B (5.7%) showed the highest water absorption in the group of lignocellulose ber-reinforced composites. This is not only because of the strong hydrophilicity of PLA/B that has a signi cant impact on water absorption, but also due to the bers around which additional microchannel spaces are formed through which water can ow. In the group of composites reinforced with llers, PLA/C (4.28%) had the lowest water absorption rate. This relationship results from the cellulose content in the bers. The higher the cellulose content, the higher the water absorption capacity. The content of cellulose in individual llers used in the research is: 40-50% for wood, 40-55% for bamboo and 12-25% for cork, which corresponds to the results obtained in this study PLA/B > PLA/W > PLAC [34]. Moreover, higher water uptake capacity by wood than by cork has already been reported by other studies [27]. Figure 5 shows the results of tensile strength and Young's modulus after hydrodegradation of the composites. The tensile strength value for composites immersed for 28 days at + 38°C decreased by approximately 5% and 20% for neat PLA and composites with natural particles, respectively. Lower decrease in the mechanical properties of neat polymer are associated with a greater ability to plastic deformation than for bers-reinforced composites. Furthermore, natural llers containing lignocellulose have a high water absorption capacity, which causes them to swell (increase in size), which contributes to the formation of cracks in the matrix [35]. In addition, water owing into the material caused the bers to detach from the matrix, which further contributed to the reduction of ber/matrix adhesion. During the tensile test of materials subjected to water immersion, lignocellulosic llers induce additional stress that reduces the values of mechanical properties (expansion and contraction of bers during water absorption).
However, in the case of Young's Modulus the decrease was about 10% for all tested materials. As mentioned earlier, composite stiffness is not as dependent on ber adhesion as it is in tensile strength.
Additionally, a slight decrease in the stiffness of the materials could be caused by an increase in the degree of crystallinity of the material, because of immersion in water at 38°C.
A similar tendency was observed for samples after heat treatment. By comparing the results obtained during tests at 80°C, it can be concluded that biodegradable composites are more susceptible to high temperatures than to water. This is particularly evident for Young's Modulus, where neat PLA subjected to high temperature showed a decrease of approximately 90% at elevated temperatures.
3.6. Microstructure Figure 6 shows the fractured surface of the composites after the tensile test. As shown in the gure, the addition of natural llers changed the characteristic of the breakthrough from ductile to fragile. As already mentioned, the decrease in the values of mechanical properties for composites with 80% and 60% in lling was caused by a lack of material continuity and gaps between the applied laments (higher porosity of materials) what can be seen in Fig. 6. Each natural ber used in the present study has a different structure. Wood bers are the widest, and the width reaches approximately 800 µm; additionally, the adhesion between the matrix and the ber is poor. Bamboo bers are the longest, and their lamentary structure can be observed in the gure. A black rings around the bamboo bers occur, which indicates poor adhesion and deformation. However, because of the length of bamboo bers, the value of the properties of composites with these bers are the highest compared to those of the other tested materials. In addition, the differences among the distances between the laments are the lowest for PLA/B, because of the higher viscosity of the material and the predominant elastic fraction, which causes impeded ow and directly affects the interlayer bonding.

Conclusion
The mechanical behavior of PLA based composites in different states and conditions were investigated in this study. Studies have shown that both the effect of changing the amount of lling and the addition of natural llers have a similar effect on mechanical properties. Heat treatment of manufactured composites increase their mechanical performance. Despite the decrease in mechanical properties after hydrolytic degradation, mechanical properties remain at a high level, which generates their potential use in long term applications. In addition, the composites were characterized by high impact strength values comparable with neat PLA, and in the case of cork composites, improvement constituted 50%.
In summary, the work presents the results obtained through the use of various techniques for the modi cation of biodegradable 3D printed materials. The obtained data will allow to design material with the expected properties at an early stage of engineering design. As a results of the work, the impact of lignocellulosic particles has a different effect on the mechanical properties than in the case of injectionmolded polymer composites. The addition of lignocellulosic bers to injection-molded polymer composites generally improves their Young's modulus, while in the case of 3D printed composites it lowers their mechanical properties. However, the relatively high obtained mechanical properties compere to the density of composites with natural llers and the low decrease in mechanical properties after hydrolytic degradation indicate their potential application in various industry sectors (medicine, furniture, decorations, automotive industry and so on).
The presented results suggest in what direction should be guided further research on 3D printed polymer composites. It is important to increase the ber adhesion to the matrix. This effect can be achieved by introducing smaller natural particles (nano or micro scale) or by adding a plasticizer. With the addition of a plasticizer, one should consider the use of natural composition (leaving 100% biodegradable material) or the nal use of material for engineering products whose 100% biodegradation is not required.
Taken together, additive techniques are a dynamically developing group of manufacturing processes. However, the methodology of research on polymeric composites produced using the FDM is not well systematized [38]. Mechanical tests presented in this paper can provide predictions regarding the in uence of different modi cation of composites (level of in ll, llers and pre-post-processing techniques) on mechanical behavior. In addition, our study took into account the behavior of composites in various environments (temperatures and hydrolytic degradation). Such an extensive and systematic presentation concerning a "di cult" polymer such as PLA (low Tg, biodegradation, high production cost, etc.) is the reference for detailed evaluation of other materials. The modi cations and methodology contained in this work will not only serve for the investigation of model PLA system, but also for other petrochemical matrices that are less sensitive to the aquatic environment and elevated temperatures.
To sum up, the research presented in the paper shows how 3D printed polymer composites can be modeled by changing various components, i.e. type of ller, degree of lling or rate of crystallization (pre or post processing).  Water absorption of PLA composites with vary in lls and llers Change of mechanical properties of samples immediately after printing (control) and after 28 days of hydrolytic degradation: (a) tensile strength and (b) Young's Modulus