2.1. Fiber Deposition System
In this work, a long fiber deposition system for integration in 3D printers recently developed by the authors [
17] is presented. The development of this system was based on the concepts and technology developed by the RepRap movement, founded in 2005 by Dr. Adrian Bowyer. The FFF prototype equipment was constructed from extruded aluminum profiles and an ISEL brand XYZ axis system with trapezoidal spindle movement, as shown in
Figure 1.
The equipment systems of control and monitoring were developed using an Arduino Mega 2560 controller, 3 external drives MSD542-V2.0 and in NEMA 23 step motors, as shown in
Figure 2. As both the hardware and the adopted software are open source, the FFF prototype developed can be considered a low-cost system. The development of this coextrusion system is intended to promote the melting of the polymer fibers present in the TPFL
® filament and their controlled deposition.
Since the carbon fiber filaments mixed with PA12 do not have enough stiffness to be compressed by the extrusion rollers, enabling their passage through the heated nozzle, it is necessary to use a flow of molten polymer capable of generating shear stresses on the surface of the fibers, thus promoting its displacement/drag towards the heated nozzle outlet. The strategy selected to join the fiber filament with the thermoplastic filament was to feed the fiber filament before thermal cutting in a cold zone, where the polymer is still in a solid state, as shown in
Figure 3, detail b).
The developed heat break component has as its function to: (i) feed the fibers to the thermoplastic filament and (ii) allow the thermal cut between the hot zone and the heatsink to avoid the melting of the polymer outside the heated zone which would cause clogging in the system. This component must, additionally, guarantee the maintenance of the hot zone temperature, by minimizing heat transfer by conduction, and finally allow for the protection of the polymeric components and the adjacent electronic systems. The material selected was stainless steel AISI 316L, which is a refractory steel, and has a low coefficient of thermal conductivity of 16.3 W/m·K at 100 °C.
The developed heat break described above was produced using a SLM Selective Laser Melting (SLM Solutions SLM 125hl model) metal additive manufacturing equipment. The metallic powder used was 316L stainless steel supplied by SLM Solutions. The selection of this AM technique allowed for manufacturing reduction time of the different prototype parts by the simultaneous production of the same component with slight variations of geometry. However, the SLM requires pre- and post-processing operations and holes rectification.
Figure 4 shows the components produced, still fixed to the construction platform by the supports, prior to carrying out the post-processing operations.
The solution adopted allows the introduction of the fiber filament into the extrusion channel before the thermal cut of the extrusion system, as shown in
Figure 5.
In addition, a fiber-cutting system was developed and applied at the outlet of the heated nozzle. This system allows for the automatization of the fiber-cutting operation. The cutting of the fibers is performed by a diamond disc coupled to a direct current (DC) motor. The motor backing is coupled to a hinged mechanism that allows for the reciprocating movement of the disc between the fiber cutting position and the standby position. The drive mechanism of the disc is composed of a NEMA 8 stepper motor, a V-spindle, a connecting rod and a guiding assembly consisting of a chute and bushings, as shown in
Figure 6. Since it is a prototype of a device, the polymeric materials selected for the construction of the chassis, connecting rod and bushes were ABS and PA6, respectively. In the manufacture of these components, a FFF polymer additive manufacturing equipment was used to model uPrintSE from the Stratasys. The cutting system is triggered at the end of the deposition of each layer of fiber. This action is controlled by the inclusion of a routine in the Gcode file of the component manufacture. This routine allows for the deposition pause by moving the printhead vertically, thus activating the cutting mechanism. After cutting, the disc collects and the deposition of the fibers starts in the next layer, as exemplified by the sequence of images in
Figure 6. During this routine the temperature control of both the extrusion and the deposition layer is maintained.
In order to integrate the two extrusions and cutting systems, a specific support was developed and applied to the X-axis carriage of the printing equipment. The geometry of this support and its dimensions have been designed, considering the specific system of the axes used.
Figure 7 shows the assembly of the system. The arrangement of the axle system and the high dimensions caused constraints, preventing a greater compactness of the support. This support should therefore be reengineered with a purpose of its integration into small FFF systems. For the extrusion of PA12, a commercially available 1.75 mm E3D-v6 extrusion system was used. This extrusion system fulfills the exact specifications required using PA12, namely the high melting temperatures of the polymer (~300 °C) and its geometry to avoid clogging [
40].
Adjustments and calibrations were made to the system prior to the functioning of the developed printing system. The first procedure, after loading the firmware in the controller of the equipment, was the verification of the axes’ direction of movement and the logic of the signal of the limit switches according to the firmware. Then, the maximum displacement length of the three axles was calculated and the construction platform leveled. Finally, printing tests were performed to adjust deposition parameters such as velocity, acceleration and its derivative (jerk). It was also necessary to adjust the percentage of polymer added to the fiber strand in the coextrusion.
The composite additive manufacturing, through the FFF technology needs to respect and adapt the procedures to overcome the limitations of the materials and the process, aiming to optimize the mechanical properties of the components produced. To pursue this objective, some rules were followed from the modeling phase of the components to the preprocessing of the file generation of the trajectories, as well as the processing parameters, such as the processing temperature, the deposition velocity, and the height and width of the layers.
Table 2 summarizes the values of the processing parameters assumed after an optimization process. Samples with dimensions and geometries, according to the standards required, were printed for to carry out tensile and three-point bending tests. In the following section, these and other tests are carried out with the aim of characterizing printed composites.
2.2. Materials
In this work the additive manufacturing of reinforced thermoplastic composites was carried out using a 1.75 mm polyamide 12 (PA12) filament supplied by eSUN and a hybrid filament of carbon fibers (CF) and polyamide 12 (PA12), TPFL
® by Schappe Techniques (France). The TPFLs are dry prepregs that homogeneously combine reinforcement and matrix filaments. This filament combines long CF and PA12. The matrix is composed of PA12 produced by the company EMS CHEMIE AG (Switzerland) and it is marketed under the name Grilamid 12. The selection of this matrix took into consideration criteria such as adhesion, processing temperatures, performance, durability and cost. The properties of the matrix are shown in
Table 3. The PA12 is a semi-crystalline thermoplastic polymer with a crystallinity of about 41% for solidification rates between 1 and 100 °C/min [
27].
For the hybrid filament manufacturing, discontinuous PA12 fibers which have a mean diameter of 20 μm are used. The reinforcement of the composite material used in this work is composed of long CF supplied by Toho-TenaxFibers GmbH (Germany). Carbon fibers are produced from continuous fiber filaments which are drawn by means of rollers to promote their breakage in distinct sections. This technique allows for the elimination of the weakest points of the fibers and for the improvement of their characteristics, increasing their tenacity and spinning capacity. Thereafter, the long fiber filaments are blended with the PA12 fibers, and may contain 1000 or 3000 carbon filaments, which are defined as 1 K and 3 K, respectively. This mix is carried out to keep the carbon and PA12 fibers parallel to each other and oriented towards the filament. According to Schappe Techniques the fiber volume fraction including the outer winding is 53% with the metric number of 0.84 Mc (km/kg). Some properties of the TPFL
® filament are shown in
Table 4.
The methods used to obtain the composite mechanical characterization and the adhesion between yarns—in the same layer and between yarn layers—were both uniaxial tensile tests, and the three-point bending/flexural tests were performed. The fiber deposition path was made in longitudinal from the middle of the sample to the outer region. After the production, the test samples were manually polished with 220 grit sandpaper to eliminate burrs and edges and to standardize their dimensions. The specimens were printed at an ambient temperature of 23 °C. Polyamide filament coils and carbon fibers/PA12 yarn, prior to use, were kept in an oven for 5 h at a temperature of 70 °C. This procedure aimed to eliminate the moisture present in the polyamide [
41]. The uniaxial and three-point bending flexural tensile tests were performed on a Zwick electrochemical test equipment, model Z100. The specimens were taken to the breaking point and the data obtained was analyzed and processed.
The calculation of longitudinal modulus (
E1) of composites reinforced with fibers were theoretically determined and the values obtained are presented in
Table 5. For the reinforcement with unidirectional fibers, Equation (1) can be applied to calculate
E1 [
29]:
where,
EF and
EM are the modules of carbon fibers and PA12 matrix, respectively, and
Vfc is the carbon fiber volume fraction. For the reinforcement with long fibers, the longitudinal modulus (
E1disc) can be estimated using Equation (2) [
29]:
where,
ξ and
η are given by Equation (3):
From the results of
Table 5 and for the 3 K yarn, the value of volume fraction (
Vfc) and the diameter of carbon fibers (
DF) are 0.1173 and 7 μm respectively, whereas the longitudinal elasticity modules of carbon fibers and PA12 (
EF) = 240 GPa and (
EM)= 1.1 GPa with an average length of carbon fibers being 100 mm. The theoretical calculation of the longitudinal module, through the Equations (1) and (2), is 29 GPa in both cases. Regarding the 1K yarn, the value of volume fraction (
Vfc) and the diameter of carbon fibers (
DF) is 0.0345 and 7 μm respectively, whereas the longitudinal elasticity modules of carbon fibers and PA12 (
Efc) = 240 GPa and (
EM) = 1.1 GPa with an average length of carbon fibers being 100 mm. The theoretical calculation, through the Equations (1) and (2), is around 9 GPa in both cases. These results mean that, theoretically, although the carbon fibers are discontinuous, the mechanical properties are like those of composites reinforced with continuous fibers. This is due to the high aspect ratio—the ratio between length and diameter—exhibited by the carbon fibers.
2.2.1. Uniaxial Tensile Tests
The tensile tests were performed only for a four-layer sample with 3 K yarn, the dimensions of the test samples were selected according to the maximum build dimensions of the prototype equipment.
Figure 8 illustrates the geometry and dimensions of the uniaxial test sample used called the “dog bone”. The average thickness obtained in the test pieces was 4 mm. The uniaxial tensile tests were performed in position control until the specimen rupture, with a constant displacement velocity of 1 mm/min and a preload of 2 N, at room temperature. Three tests were carried out to minimize possible interferences resulting from the processing and testing of test pieces.
2.2.2. Three-Point Bending Tests
The three-point bending test was performed according to ASTM D790 [
42]. The specimens were cut according to a rectangular geometry with the dimensions of 60 × 15 mm, as shown in
Figure 9a and later sanded with 220 grain sandpaper to eliminate burr and surface softening, as shown in
Figure 9b. The thickness and average width of each specimen were then checked. Three samples each of thickness 2 and 4 mm were tested and, for the two types of yarn 1 K and 3 K, the average was then calculated to minimize possible human or material faults.
2.2.3. Morphological and Microstructural Analyses
Scanning electron microscopy (SEM) (Merlin-61-50 FE-SEM, from Carl Zeiss) was used to characterize the microstructure. The cross-sectional images were used to demonstrate the distribution of the long fibers, any possible void content, and the bonding details. The surface of the SEM samples was first sputter coated with gold with a thickness of 20 nm. Then, SEM micrographs were observed at an acceleration voltage of 5 kV and an emission current of 20 pA.