Effect of different parameters on the tensile properties of printed Polylactic acid samples by FDM: Experimental design tested with MDs simulation

Fused Depositional Modeling (FDM) is one of the common methods for 3D printing of polymers, which is expanding in various industrial applications, scientic researches, and engineering applications due to its ability to make complex parts. In this research, molecular dynamics (MDs) simulation has been used to predict the physical and mechanical properties. Then, the mechanical properties of the printed parts are obtained. The mechanical properties of 3D printed parts strongly depend on the correct selection of processing parameters. In this study, the effect of three important parameters such as inll density, printing speed, and layer thickness are investigated on the tensile properties of PLA specimens. For this purpose, standard specimens with four inll densities of 20%, 40%, 60% and 80%, two speeds of 20 mm/s, and 40 mm/s, and two thicknesses of 0.1 mm and 0.2 mm are printed and tested under quasi-static tensile test. In all printed specimens, the print angle is ± 45°. The experimental results show that the inll density in comparison to the other two parameters has signicant effect on mechanical properties such as modulus of elasticity, ultimate strength, and failure strain. According to these results, by increasing the inll density, the stiffness and strength of the specimens increases considerably. At inll density of 80%, the specimens has the highest stiffness and strength, but it exhibits a brittle behavior. Moreover, it can be deduced that by reducing the layer thickness although the modulus of elasticity increases a little, ductility is greatly affected.


Introduction
Three dimensional (3D) printing is one of the fast prototyping methods that produces the product using a 3D model created by a computer [1][2][3][4][5]. 3D printers are used in various industries such as mechanical, automotive, aerospace, civil, and medical engineering. For example, a 3D printer can be used to produce prototypes, molds, screws or gears used in robots, plastic parts for some machines, veneers for some prostheses, ber composites, and more [6,7]. The production process of 3D printing is based on the additive manufacturing process [8]. In the additive manufacturing process, the layers are stacked one after the other, eventually turning a computer-generated 3D le into a physical product. Today, the use of additive manufacturing technology has increased due to its exibility, and design advantages [9][10][11][12][13]. For example, this technology allows the creation of complex shapes and structures with less weight, reduces development time and tool costs, and dramatically simpli es the production process [14][15][16][17][18]. Also, using this method, parts that are made of several assembly parts are made as a single object without assembly. Also, by using this method, parts that are made of several assembly parts are made as a single object without assembly [19][20][21][22][23]. 3D printers are capable of printing a variety of materials such as metals, ceramics and polymers. Types of 3D printers for polymer printers include: Fused deposition modeling (FDM), Stereo lithography (SLA), Digital light processing (DLP), Selective laser sintering (SLS), and laminated objective manufacturing (LOM), Liquid Deposition Modeling (LDM) (ceramic 3D printer) [24,25]. 3D printers of FDM are one of the most common 3D printing methods in which extruded laments of thermoplastic polymers are used to produce the part [26][27][28][29][30]. The extruder is placed on a three-dimensional CNC table that moves in the x, y and z directions and places the molten polymer inside the extruder on the part [31,32]. After nishing one layer, the extruder moves upwards as thick as one layer [33][34][35][36][37]. It is important to note that many parameters affect the quality and nal properties of a sample printed with FDM [38][39][40][41][42]. Construction parameters can be divided into three general categories: 1) Slicing parameters. The most important of these parameters are: Thickness of layers, Number and walled thickness, Print speed, Nozzle diameter, Printing angle, Feed rate, In ll density, and In ll (raster) pattern. 2) Building orientation: Samples can be printed and made horizontally, vertically or transversely.
3) Temperature conditions: includes ambient temperature, nozzle temperature and desktop temperature [43][44][45]. Recently, many studies have been performed on the mechanical properties of 3D polymer samples printed by FDM method, and these studies show the importance of paying attention to 3D printing parameters. Most recent research has focused on the effect of second-and third-group fabrication parameters on the behavior of samples printed with Acrylonitrile Butadiene Styrene (ABS) lament, while Polyactic Acid (PLA) lament, despite its unique properties, has received less attention from researchers [46][47][48][49][50]. For example, a group of researchers studied the effect of print parameters such as layer thickness, direction and printing angle on mechanical properties using ABS and PLA laments. Croccolo et al. [51] studied the direction and angle of printing of ABS specimens. Tymrak et al. [52] investigated the effect of printing angle parameters and the thickness of the sample layers with PLA.
Chacon et al. [53] studied the printing direction, layer thickness and lament exit rate and performing three-point tensile and exural tests, they found that the PLA samples which printed by the FDM printer in the horizontal direction has the best mechanical properties by decreasing the layer thickness and increasing the lament exit rate. Kozior et al. [54] investigated the mechanical properties of PLA samples produced by FDM printers in various printing directions by pressure testing. Mahmood et al. [55] studied the tensile properties of ABS specimens by examining the parameters of layer width and thickness, sample lling density, number of walls and printing direction. Alafaghani et al. [56] investigated the effect of the mechanical properties and increasing the accuracy of PLA samples printed by FDM printer by researching the parameters of layer thickness, printing speed, nozzle temperature, sample internal network lling density, internal network lling pattern, and printing direction. Tronvoll et al. [57] tried to increase the dimensional accuracy, and smoothness of the surface of printed specimens with helical, and sloping surfaces (such as screws) by changing the layer height parameter and using a linear pattern.
Popescu et al. [58] reviewed existing articles on the effect of FDM printer parameters on the mechanical properties of polymer samples. Khandan et al. [59] printed a mitral heart valve with Polyurethane polymer by FDM method and checked its properties by MDs simulation. They also examined tensile strength, strain at fracture, permeability, and the ultimate tensile strength to monitor the mechanical property of this arti cial heart valves.
A review of the above sources shows that most recent research has focused on the effect of print angle and direction and lling density on mechanical properties. While the printing speed and layer thickness can have a great impact and also for predicting the mechanical and physical properties by MDs simulation that is another novelty of this article. On the other hand, in these studies in the literature review, only the modulus of elasticity and strength (among the mechanical properties) is considered and the effect of changing the parameters on fracture strain, weight of the produced part, printing time, and consequently costs are not discussed. Also, among the researches, only two researchers have studied the effect of printing speed on mechanical properties, which in both studies, high speeds have been considered. Since the three parameters of lling the internal network of the sample, printing speed and layer thickness have a signi cant effect on the fabrication time of the part and so far these three parameters have not been studied together, in this study, the effect of these three fabrication parameters on the properties of PLA polymer specimens including stiffness, ultimate strength and nal strain of fracture are investigated and nally the most suitable conditions in terms of weight and printing time are introduced and simulated.

Materials And Method
2.1. Introduction of PLA and its molecular structure PLA is a linear aliphatic thermoplastic polyester that can be completely extracted from renewable sources such as corn [60]. Of course, its use has been limited to medicine such as sutures. But today this polymer is widely used in many industries [61]. PLA is very similar to plastic [62]. This thermoplastic lm is biodegradable and a combination derived from the renewable plant sources of corn and sugar beet [63].
PLA is one of the few polymers whose molecular structure can be controlled by the ratio of isomers L and D to obtain a crystalline or amorphous polymer with high weight as shown in Fig. 1 [64].
PLA is known as a food safety agent and can come into contact with food. This polymer decomposes without the need for a catalyst by hydrolyzing ester bonds. The rate of decomposition depends on the shape and size of the polymer object, the isomer ratio and the hydrolysis temperature as shown in Fig. 2.
In general, the physical and chemical properties of lactic acid are given in Table 1 [65][66][67]. Vapor pressure at 25mmHg 0.0813 In this paper, as shown in Fig. 3, the simulation of PLA was performed by MDs method. The element of PLA as shown in Fig. 4 is simulated by Materials studio software to predict the mechanical and physical properties before fabrication.

Simulation Steps, Force eld, and software
In this article, the MDs method is applied to predict and obtain the signi cant mechanical and physical properties of simulated PLA [68][69][70]. The following steps have been taken to complete the molecular dynamics method as shown in Fig. 5.
The initial atomistic models are created in Materials Studio software and Condensed-phase Optimized Molecular Potentials for Atomistic Simulation Studies (COMPASS27) force eld is determined for modeling inter-and intra-atomic interactions. The COMPASS parameters for covalent molecules are completely con rmed using various calculation techniques including extended molecular dynamics simulations of liquids, crystals, and polymers. COMPASS is the rst ab-initio force eld able to foretelling thermo-mechanical properties of polymers like polymer nanocomposites perfectly. NVE explains that the sum of kinetic (KE) and potential energies (PE) is conserved, T and P are unregulated and N, V, and E denote a constant number, volume, and energy, respectively. At this step, the simulation box is placed at a temperature of 298 °K under NVE. The simulation time is considered 100 ps.
NVT represents that temperature (T) is regulated via a thermostat, which typically adds a degree of freedom to the conserved Hamiltonian; KE and PE are included in the Hamiltonian; P is unregulated. At this part, the simulation box is set at the temperature of 298 °K under NVT. The initial density of the system (0.9gr / cm³) is assumed to allow molecules and atoms to be displaced to move towards optimal mode. The simulation time considered 100 ps.
NPT is similar to NVT, but the pressure (P) is regulated. Density is one of the physical properties that is considered in atomic modeling. It determines the accuracy of the density of the atoms in equilibrium. If the atomic modeling path is followed correctly, the density of the atomic system is expected to be close to the actual density of the system in comparison to the macro. Furthermore, it is assumed that after the simulation time, the amount of any quantity attributed to the system of atoms, including the converged density of the solution uctuations, will decrease over time. At this point, the system is pressurized at atmospheric pressure 1 at a temperature of 298 °K under a constant NPT to close the system density to the actual density. NPT can also eliminate system tensions. The simulation time at this stage considered 100 ps.

Experimental work
PLA lament is one of the most widely used laments in 3D FDM printers. This lament has unique properties compared to other laments. These properties include: 1) PLA is a thermoplastic polymer that is generally produced from lactic acid from corn and sugar beet therefore, it is recyclable. 2) In addition to being recyclable, low environmental impact and it is highly compatible with the human body. 3) An important advantage of this lament is that it does not emit gas during printing. Therefore, it allows printing to be done at home, in classrooms, and factories without special ventilation. 4) These laments do not need a warm bed. Therefore, there will be no problems in the cooling process such as partial shrinkage and cracking. 5) The melting point of PLA lament is between 180-230 ˚C. Given these advantages, this study focuses on the mechanical behavior of PLA laments. It should be noted the diameter of the lament that used in this research is 1.75 mm.

Sample preparation
First, the standard sample of tensile test as shown in Fig. 6 is modeled according to Type I standard ASTM D638 in Solid works software and then save as extension of *.STL [71].
Then the modeled *.STL le is converted to G-Code using Cura software. It should be noted that in most cases, during the process of converting the model to STL and then G-code, the geometry of the part loses its dimensional accuracy slightly. Cura software is one of the most comprehensive and practical 3D printing software. This software uses a very advanced graphical environment that is equipped with a layering simulator. Using this simulator, all different manufacturing parameters can be applied to create *.STL model. Therefore, in the next step, different parameters of model making are selected in this software. Table 2 shows the technical speci cations of the 3D printer used to print SP1-SP6 samples. Figure 7 shows the overview of 3D printer that printed specimens (SP1-SP6) in the present work. In order to investigate the effect of the internal network density parameters of the sample, the printing speed, and the thickness of the sample layer were made according to the speci cations given in Table 3 and during the time that presented in this table. According to the standard, at least ve samples must be made and tested for each case to prove the reproducibility of the results. Due to changes in manufacturing conditions and adverse events that may occur during the manufacturing process, and the scattering of data in printed samples may be high in some cases. Therefore, in this study, the results of the three samples that had the highest repeatability were reported. It should be noted that the rest of the construction parameters such as the thickness of the initial layer, the width of each layer, walled thickness are xed in all samples and their values are listed in Table 4. Also, according to the melting point of PLA lament the nozzle temperature for printing all parts, was set at 210°C. The printing angle of ± 45° also gives better mechanical properties to the printed samples [72,73]. Because the printing time of a layer with an angle of 45° is less than other angles, and this causes the bottom layer not to cool completely and as a result, more adhesion is created between the layers.
Finally, give the prepared G-code to the FDM printer to print the sample using PLA lament. An example of a printed PLA specimen is shown in Fig. 8.

Results of molecular dynamics simulation for PLA
The mechanical and physical properties of PLA are shown in Fig. 11 and Table 5 by following the steps that mentioned in Sect. 2.2 are derived the mechanical, and physical properties. To draw the density diagram, rst, the ensemble of NVT to maximize the energy of system is done. Next the ensemble of NPT is done to plot the density, as shown in Fig. 11, the density converges to 1.23 g/cm 3 . Table 5 shown the comparison of mechanical, and physical properties of PLA with MDs modeling and experimental analysis.

Diagram of Radial Distribution Function
One of the important quantities to evaluate the equilibrium validation of the system in molecular dynamics is the radial distribution function (RDF). The RDF expresses the mass distribution of the system over atomic distances that is expressed by relation (1).
To validate the simulation results, the elastic stiffness matrix and elastic constants of the PLA were determined using a constant strain method. The elastic stiffness matrix components were de ned for PLA, under a strain of 60.003 and at a pressure of 1 atm that this matrix is shown as follows: As can be observed, because of the isotropy of the material, the diagonal elements are nearly similar and the matrix is approximately symmetric.

Mechanical testing evaluation
To investigate the effect of in lling density, print speed, and layer thickness on the mechanical properties of PLA samples made by FDM printer, rst-four groups of samples SP1, SP2, SP3 and, SP4 were compared, then SP2 and SP5 samples and SP3 and SP6 samples were compared with each other. Usually, due to the long printing time and high consumption of lament, the parts are not printed completely solid. In lling density is a parameter that is displayed as a percentage and indicates how much of a solid model should be lled with material when printing. Therefore, this quantity directly depends on the weight and construction time of the sample. Fig. 13 shows the stress-strain diagram of SP1, SP2, SP3, and SP4 with different density percentages. Fig. 14 shows that the mechanical properties of printed specimens such as stiffness, ultimate strength and fracture strain are strongly dependent on the in lling density and they are directly proportional to the modulus of elasticity and strength. Fig. 14 also shows the broken specimens SP1 to SP4 after the tensile test. As can be seen, most of the specimens are broken in the failure range.
Also, by comparing the printing time of SP1 to SP4 parts in Table 3, it is proved that the lling density is directly related to the printing time. Table 6 shows the detail and average mechanical properties of SP1 -SP6 specimens. Fig. 14 also shows the broken specimens SP1 to SP4 after the tensile test. As can be seen, most of the specimens are broken in the failure range. By comparing the values in Table 6 and the stress-strain diagrams in Fig. 13, it can be seen that as the internal network in lling density increases, its stiffness and strength increase dramatically. But from 60% density upwards the strain of failure is reduced. This means that as the lling density increases too much, the specimen becomes brittle and deforms less under a quasi-static load. If the failure strain of the SP4 part is ignored, this part can be considered as the best sample. As previously mentioned, the printing process in FDMs is based on melting the lament in the extruder and cooling the material at ambient temperature, and this is done layer by layer. These rapid heating and cooling, despite creating residual stress and internal defects in the part, cause it to become extremely brittle and increase its modulus of elasticity and strength and brittleness. Excluding SP4 specimens due to low fracture strain, by considering the modulus of elasticity the ultimate strength and fracture strain of SP3 with 60% density have higher mechanical properties among parts with lling density of 40, 20 and 60%. Also, between these three parts, SP2 sample (40% part) has a higher strength and weight than the other two parts with a small difference. Therefore, it can be concluded that in cases where the strength of the part is less important than its form and appearance, a density of 40% is more appropriate in terms of cost, material and time savings.

Investigating the effect of print speed
The print speed is the same as the nozzle speed during printing. High printing speeds cause discontinuity of production in printing. The reason for this is the lack of time to melt and extrude the lament, which is very much related to the type of printer used and the mechanical speci cations of the device, and precisely for this reason, any high speed cannot be printed with any printer. As mentioned in previous studies, low speeds were not considered at all. Therefore, in this study, two low print speeds were selected and the samples were printed with the same fabrication parameters. Fig. 15 shows the stress-strain diagram of the tensile test for two selected samples from the SP2 and SP5 groups. Table 7 compares the numerical values of the mechanical properties of SP2 and SP5. As can be seen from the SP5 group diagram in Fig. 15, the reduction in print speed at low speeds has also led to an increase in mechanical properties such as stiffness, elongation-to-break, and ultimate strength. According to the printing time of each group of parts in Table 3, it can be seen that due to doubling the manufacturing time, the increase in properties has not been signi cant, but in cases where increasing the mechanical properties is important for the user, the specimen can be printed as quickly as possible to achieved the best properties. Fig. 16 also shows the broken specimens of SP2 and SP5 after the tensile test.

Investigating the effect of layer thickness
Since 3D printing technologies make the specimens layer by layer, the thickness of each of these layers in the print determines the quality of the specimens. In this study, two groups of SP3 and SP6 pieces with a difference in layer thickness of 0.1 mm were tested to investigate the effect of layer thickness on the mechanical properties of printed parts. Fig. 17 shows the obtained stress-strain diagram from the tensile test of one of the components of SP3 and SP5 groups. Numerical values of the mechanical properties are also summarized in Table 8. The results show that reducing thickness of the layers leads to a slight increase in weight and stiffness of the part while greatly reducing the fracture strain. Therefore, the thickness of the layer has the greatest effect on the fracture strain of the piece and is inversely proportional to it. On the other hand, many researchers have shown that the thickness of the layers also has a signi cant effect on the dimensional accuracy and surface quality of the printed part, and as shown in Table 9, the dimensional accuracy of the SP6 sample group is closer to the standard dimensions of ASTM. Fig. 18 also shows images of broken specimens SP3 and SP6 after tensile testing. Comparing the fabrication time of SP3 and SP6 parts according to Table 3, it can be seen that although the layer thickness decreases, the dimensional accuracy and surface quality increase, the printing time becomes longer because the thickness of each layer becomes thinner and the printer has to print more layers to make the whole part, which takes more time to complete the manufacturing process.

Conclusion
In this study, the effect of different parameters on the tensile properties of printed PLA samples by fused depositional modeling with MDs method and experimental analysis were investigated. For this purpose, in addition to the effect of three parameters of layer thickness, print speed, and internal network in lling density on the tensile properties of PLA polymer parts made by FDM printer, the most cost-effective parameters (in terms of cost and time) have also been extracted. The obtained experimental results from tensile experiments show that with increasing the in lling density, the mechanical properties of the parts increase signi cantly. However, at very high in lling densities, the components behave more brittle and have lower strain fractures, so in cases where the strength of the part is less important than its shape and appearance, a density of 40% is more appropriate in terms of cost, material and time savings. It was also observed that the printing speed has less effect on the mechanical properties of PLA parts than other parameters and it has increased the mechanical properties. It was also observed that reducing the thickness of the layer, while slightly increasing the stiffness of the parts, makes the part extremely brittle and on the other hand, it causes the increases the dimensional accuracy and surface quality of the specimens. Therefore, if the beauty and appearance accuracy of the part is a priority, the thickness of the lower layers is more suitable for this, otherwise the thickness of the higher layer causes faster and cheaper construction and also gives better mechanical properties to the specimen.

Declarations
Spatial isomers of lactic acid Figure 2 Spatial chemistry of lactic acid and lactide molecules  Simulation atomistic model of PLA in materials studio software Figure 5 Page 22 /29 Steps to obtain the mechanical, and physical properties using Materials studio software Dimensions of standard specimen (in millimeters) for plastics Figure 7 3D printer that printed specimens (SP1-SP6) in the present work Overview of 3D printed specimen Density convergence during simulation for PLA at the temperature of 298 °K Figure 12 Diagram of the radial distribution of atoms that convergence to the number 1 Figure 13 Page 27/29 The stress-strain curves of PLA specimens with different in lling density Figure 14 Samples of SP1 to SP4 after tensile test Figure 15 Stress-strain curves of PLA specimens with different printing speeds Figure 17 Stress-strain curves of PLA specimens with different layer thicknesses Figure 18 Samples of SP3 and SP6 after tensile test