Abstract
Continuous fiber-reinforced thermoplastics (CFRTs) can in combination with high-strength metals offer characteristics that cannot be achieved with mono-material parts. One possible example is the combination of locally high-temperature resistance in the metal component with superior weight-related mechanical properties due to the CFRT component. This approach requires a reliable and durable joining technology, which considers the material-specific properties and allows to exploit the full potential of CFRT/metal hybrid parts. A promising approach in the field of CFRT/metal joining is the use of metallic pins, which can be embedded in the locally heated CFRT component to create a form-fitting joint. In the current state of the art, primarily single-pins are investigated and characterized: especially the distinct fiber orientation in the direct pin pressing process is only described for single-pin joints. Behind this background, the present study aims at creating an understanding of the fiber orientation mechanism for multi-pin arrays. Therefore, in the scope of this study, unidirectional reinforced glass fiber/polypropylene samples are joined via direct pin pressing and infrared heating with different 1D and 2D multi-pins arrays with different pin-diameters, spacing and pin distributions. The resulting joint morphology is consequently analyzed using micro-computer-tomography. Based on the performed investigations, a model for the fiber displacement mechanism is proposed, and the first recommendations for the design of fiber-friendly multi-pin joints with unidirectional reinforcements are given. It showed that especially pin-spacing in fiber orientation in dependency of the pin diameter is critical for a fully reconsolidated joint quality, and it is suggested that a pin-offset in the fiber direction is beneficial for a fiber-friendly joining process.
1 Introduction
Lightweight manufacturing is a key technology for resource-efficient transportation and gains in relevance due to the increasing climate change [1]. A common approach to achieve weight reductions is the substitution of traditional engineering materials such as steel and aluminum alloys through lightweight and high-strength fiber-reinforced composites. Especially continuous fiber-reinforced thermoplastics (CFRTs) combine high-weight specific modulus and strength with uncomplicated handling [2]: compared to thermoset-based composites, CFRTs can be easily stored at room temperature. Furthermore, high-volume series production is feasible as the solidification process of the matrix during production is defined through the physical cooling process in contrast to the chemical reaction as is the case for thermoset matrix systems [3]. Despite the good mechanical and processing properties, CFRTs are limited when exposed to high temperatures or high abrasive wear [4]. In such use cases, a combination of a hybrid part with high-strength steel or aluminum components can merge the specific advantages of both material groups allowing to exploit new use cases [5]. This approach, however, requires a reliable and cost-efficient joining technology to be applicable in series production [6]. Due to chemical incompatibilities of the metal component and the polymer matrix, welding processes are not feasible [7] and adhesive bonding, although being well suited for composite applications in general, typically requires extensive process control and for many thermoplastic matrix systems requires surface treatment such as plasma treatment, before being applicable in CFRT/metal joining [8,9]. In state-of-the-art metal/CFRT joining, mechanical fastening via bolts and/or rivets is commonly used and well described in the literature. These joining methods, although often beneficial due to the possibility to disassemble the structures, come with distinct disadvantages: first, it is mandatory to create precise bolt holes, which add additional process steps and, when realized via subtractive processes, destroy load-bearing fibers potentially leading to stress concentrations and consequently weaken the CFRT component. Second, the required auxiliary elements add weight to the structure that contradicts the idea of lightweight construction [10].
A relatively new method for CFRT/metal joining is the utilization of cold-formed metallic pin structures that can be embedded into a CFRT component after local heating above the melting temperature of the matrix to create a form-fitting joint [11]. The use of pin structures protruding from the metal sample’s surface to reinforce thermoset-composite/metal hybrid parts has already been widely described in the literature and can be realized during the part manufacturing with dry fabrics via resin transfer molding [12] or with pre-impregnated sheets [13]. Another joining method is the use of metal [14] or composite [15] Z-pins to reinforce an adhesively bonded joint also before the curing process of matrix systems. In comparison to bolted joints with drilled holes, these joints show superior mechanical properties with an increase of force before failure of 31.8% [16]. Despite these promising properties of pinned joints, a subsequent joining cannot be achieved with thermoset-based composites, which is a major disadvantage from a manufacturing standpoint where the use of CFRTs can be a superior alternative due to the repeatedly meltable polymer matrix. However, because of the different properties of thermoplastic and thermoset matrix systems and consequently underlying mechanisms, such as very limited adhesive bonding between polymer matrix and metal substrate and specific fiber displacement mechanisms during the joining process [17], the transferability of the already existing knowledge from thermoset composites is very limited.
In the current state of the art of CFRT/metal pin joining, primarily single pin joints have been investigated which achieve shear loads of up to approximately 473 N [18] and normal loads of up to approximately 88 N [19] before failure. As these measured loads are too low for most application, it is imperative to increase the total transmittable load of the joint. One apparent way to achieve higher loads is the use of multi-pin arrays, which increase the number of used pins leading to higher reaction forces before failure. In order to be able to design such multi-pin joints, a thorough understanding of the fiber displacement and orientation process is required. Especially, the limitations in means of fiber displacement in dependency of the distance between pin structures and the dimensions of a single structure need to be understood.
Until now, only one article investigates the fiber orientation process in the direct pin pressing process in CFRTs on the example of a single pin joint [17]. The currently known studies investigating multi-pin joints either do not describe the resulting fiber orientation [20] or use ultrasonic vibration to heat the matrix system leading to a very limited heated area and zone of free fiber displacement leading to significant fiber damage [21,22]. Behind this background, this study aims at creating an understanding of the resulting fiber orientations in the direct pin pressing process with infrared pre-heating and suggesting a model for underlying fiber displacement mechanisms.
1.1 Used materials
In the course of this study, a custom-fabricated quasi-unidirectional glass fiber/polypropylene (GF/PP) CFRT was used. The matrix polymer is a polypropylene (PP)-type BJ100HP from Borealis AG (Vienna, Austria), which is a special low-viscosity copolymer for GF/PP composite applications [23]. The fiber component consists of eight layers of a quasi-unidirectional non-crimp fabric from Saertex GmbH (Saerbeck, Germany). The total aerial weight of one fabric layer is 296 g/m2, which is composed of 283 g/m2 of 300 tex glass fiber rovings in the weft direction, 1 g/m2 of 68 tex glass fiber rovings in the warp direction as well as 12 g/m2 of a 76 dtex polyester sewing thread. The stitching length of the sewing is 3.5 mm, and the average distance between individual warp rovings is measured as an average of five measurements to 3.58 mm which corresponds well with the stitching length. Figure 1 shows an exemplary macroscopic image of the used non-crimp fabric. The slight distortion of the vertical sewing thread is a result of the handling during the preparation process of the sample and is not representative of the final composite.
Image of the used non-crimp fabric.
The impregnation process was conducted at the NMF Gmbh (Fürth, Germany) on an interval hot press via the direct-melt process. The final fiber content was measured to a mass fraction of 71% via incineration, which corresponds to a fiber volume content of 47%. This fiber content is in the range of typical commercially available CFRT materials such as Tepex dynalite 104/RG600(x)/47% from Lanxess AG (Colgone, Germany) [24].
1.2 Definition of pin-array geometry
In the course of this study, three different types of pin arrays were investigated: The first type is referred to as 1D array in which the pins are placed in a line with linear spacing. The second type is referred to as 2D symmetric array, in which the pins are placed in a square spacing with equal distances in height and width. The third type is referred to as 2D asymmetric array in which every other line of pins is offset by a half distance between individual pins. Figure 2 gives examples of each array type. Each pin array was manufactured on a base plate with a dimension of 45 mm × 45 mm
Basic types of pin arrays.
Each pin array is defined via two variables: “D” describes the distance between two individual pins while “d” stands for the diameter of each pin. Within each pin array, the diameter of the pin structures is identical. With the distance between and diameter of the pins, the pin density “φ” can be calculated with the following equation:
Thereby, the pin density describes the areal fraction of a given joining area that is occupied with pins. It is expected that due to fiber displacement mechanisms, a certain maximum pin density can be observed from where any additional increase of density leads to negative effects such as fiber damage or insufficient reconsolidation quality of the CFRT component in the joining area.
1D arrays were joined in two different orientations in relation to the fiber orientation of the CFRT component (0 and 90°) while 2D symmetric arrays could only be joined in one orientation due to their symmetry in two axis. Asymmetric arrays were joined in such way that the offset of the pin lines is perpendicular to the fiber orientation. Table 1 summarizes the investigated pin arrays. Two different pin diameters 1.0 and 1.5 mm were investigated. The spacing between the pins was chosen in such a way that for both pin diameters, similar pin density results in order to allow for good comparability of the influence of pin diameter independently from the pin density. Of each described array, one sample was joined and consequently investigated.
Summary of investigated pin arrays
| Name | D (mm) | d (mm) | φ (%) |
|---|---|---|---|
| 1D-9-1.5-90° | 9 | 1.5 | 2.18 |
| 1D-9-1.5-0° | 9 | 1.5 | 2.18 |
| 1D-6-1.5-90° | 6 | 1.5 | 4.91 |
| 1D-6-1.5-0° | 6 | 1.5 | 4.91 |
| 1D-4-1.5-90° | 4 | 1.5 | 11.04 |
| 1D-4-1.5-0° | 4 | 1.5 | 11.04 |
| 1D-6-1.0-90° | 6 | 1.0 | 2.18 |
| 1D-6-1.0-0° | 6 | 1.0 | 2.18 |
| 1D-4-1.0-90° | 4 | 1.0 | 4.91 |
| 1D-4-1.0-0° | 4 | 1.0 | 4.91 |
| 1D-2.7-1-90° | 2.7 | 1.0 | 10.77 |
| 1D-2.7-1-0° | 2.7 | 1.0 | 10.77 |
| 2D-sym-9-1.5 | 9 | 1.5 | 2.18 |
| 2D-sym-6-1.5 | 6 | 1.5 | 4.91 |
| 2D-sym-4-1.5 | 4 | 1.5 | 11.04 |
| 2D-sym-6-1.0 | 6 | 1.0 | 2.18 |
| 2D-sym-4-1.0 | 4 | 1.0 | 4.91 |
| 2D-sym-2.7-1.0 | 2.7 | 1.0 | 10.77 |
| 2D-asym-9-1.5 | 9 | 1.5 | 2.18 |
| 2D-asym-6-1.5 | 6 | 1.5 | 4.91 |
| 2D-asym-4-1.5 | 4 | 1.5 | 11.04 |
| 2D-asym-6-1.0 | 6 | 1.0 | 2.18 |
| 2D-asym-4-1.0 | 4 | 1.0 | 4.91 |
| 2D-asym-2.7-1.0 | 2.7 | 1.0 | 10.77 |
1.3 Additive pin manufacturing via powder bed fusion (PBF)
The manufacturing of the pin structures was performed at the chair of manufacturing technology via the powder bed fusion process. As a material for both the base sheet and the powder, stainless steel type 316L (1.4404) was used with a sheet thickness of 1.5 mm. The powder was manufactured by LPW Technology Ltd. (Runcorn, United Kingdom) and has a size distribution from d 10 = 19 µm to d 90 = 43 µm. The utilized PBF machine was manufactured by DMG Mori AG (Bielefeld, Germany) and is of the type Lasertec 30 SLM. The temperature of the build platform was set to 200°C in order to reduce heat-induced stresses, the layer thickness was set to 50 µm, the scanning speed was set to 1,000 m/s, the laser power to 254 W, and the hatch distance to 0.1 mm which leads to a volume energy density of 51 J/mm3. These values were derived from ref. [15], where it showed good results with comparably precise reproduction of the defined dimensions with a maximum average deviation of 0.05 mm between manufactured and designed pin diameters.
After the PBF process, the pin arrays were cut from the base sheet with a CO2 laser cutting machine to the sample size of 45 mm × 45 mm. Figure 3 shows an example of a 1D pin array after the laser cutting process which is still partially attached to the base sheet.
Image of pin Array type 1D-sym-6-1.4.
1.4 Joining process and sample preparation
The Joining operation was performed on a custom build heating and joining apparatus which features a short-wave infrared spot from Optron Infrared GmbH (Garbsen, Germany) to locally heat the CFRT component and a pneumatic piston from Festo SE&Co. KG (Esslingen am Neckar, Germany) with a piston diameter of 80 mm. The Infrared spot has a maximum power of 150 W and a nominal focus diameter of 10 mm at a focus distance of 50 mm. Figure 4 shows a schematic image of the used joining apparatus.
Illustration of joining apparatus.
The joining process was thereby performed in three steps:
heating of the CFRT component via infrared radiation;
reposition of positioning device under the pneumatic piston and placing of the steel component onto the CFRT component;
Insertion of pins into CFRT component via the pneumatic piston and cooling and reconsolidation of CFRT component for 30 seconds with applied joining force;
The parameters of the heating process have been defined in pre-trials under the following pre-conditions:
The minimum target diameter of the molten must be 27 mm in order to accommodate the dimensions of the pin arrays;
The irradiated surface must not exceed a temperature of 300°C as thermogravimetric investigations at a heating rate of 10 K/min of the CFRT component showed first mass losses at this temperature;
The sample must be heated above the melting temperature, which has been measured to 163.4°C via differential scanning calorimetric measurements at a heating rate of 10 K/min;
Thereby, it is critical to consider the occurring temperature gradients: Since only one side of the sample is heated and the intensity of the IR radiation is highest in the center of the radiated area, a temperature gradient between top and bottom surface as well as between center and border of the heated zone can be observed. Consequently, it is required to determine heating parameters that ensure heating above the melting temperature at the bottom border of the heated zone while avoiding an overheating of the center top surface of the sample.
Since the nominal focus diameter of the IR spot at a distance of 50 mm measures 10 mm, the required dimension of the molten zone cannot be achieved at the normal focus distance. In order to increase the size of the irradiated area, the distance between the IR spot and CFRT surface is increased.
The parameters of the insertion and reconsolidation process were based on investigations in which a piston speed of 0.1 m/s showed good results. In comparison to the cited study, the joining forces are increased to 3,000 N in order to increase the areal reconsolidation pressure which would, due to the increased overlap of CFRT and steel component compared to [17], be significantly lower. With samples measuring 45 mm × 45 mm in this study, the applied pressure on average measures 1.48 MPa. Table 2 summarizes the chosen heating and joining parameters.
Heating and joining parameters
| Distance spot-CFRT | IR spot power | Heating time | Joining force | Piston speed |
|---|---|---|---|---|
| 120 mm | 67.5 W | 300 s | 3,000 N | 0.1 m/s |
After the joining and cooling process, the 1D samples were separated prior to further investigations. This is possible due to the pin geometry without undercuts, which allows a separation without significant damage to the CFRT component. The advantage of this procedure is that the quality of computed tomography images would be negatively impacted when high-density differences between components (such as steel and polymer matrix) are present. Since the polymer matrix is solid after the cooling process, the fiber orientations are preserved in the CFRT component even after the subsequent pin removal. Consequently, the removal of the pin structures allowed for easier X-ray computed tomography (CT) investigations while not compromising the significance of the CT images. Due to the increased number of pin structures, this procedure is not applicable to 2D arrays without causing significant damage to the CFRT component.
1.5 CT investigation
In the scope of this study, for all investigations and explanations, the following nomenclature was used: Y-axis describes the direction of the initial fiber orientation, Z-axis describes the thickness direction of the sample while X-axis is oriented in the plane of the sample perpendicular to the fiber orientation.
The X-ray computed tomography as a method of non-destructive testing offers the possibility to investigate the inner structure of samples and hence can be used to create a thorough understanding of fiber displacement and orientation processes. For the measurements in the industrial computed tomography system Metrotom 1500 (Carl Zeiss AG, Oberkochen, Germany), the 2D multi-pin arrays were positioned individually between the X-ray source and detector while the separated CFRT components of the 1D samples were simultaneously tested in groups of eight samples (Figure 5). The group measurements (Figure 5a) allowed for optimized penetration lengths from all sides, while the single measurements of the 2D arrays with a stainless steel fraction (Figure 5b) avoid long penetration lengths which in this case would be disadvantageous for the image quality. The used detector has a resolution of 2,048 Pixel × 2,048 Pixel. The used parameters were an integration time of 2,000 ms and a signal amplification (gain) of 16. Over a sample rotation of 360° per measurement, a total of 2,050 projections were captured which have been averaged out of three individual projections. The resulting voxel size is 18.08 µm. The mono-material 1D samples have been measured with a tube voltage of 140 kV and a current of 69 µA with a resulting spot size of 7 µm without the use of a pre-filter. The multi-material 2D arrays have been measured with a voltage of 190 kV and a current of 87 µA, which resulted in a spot size of 14 µm. Furthermore, a pre-filter consisting of 0.25 mm of copper has been used. The X-ray projections of the investigated samples were consequently reconstructed to a cubic volume model. Thereby, for every volume pixel (voxel), a greyscale value was calculated.
Setup of an X-ray computed tomography system with (a) source and (b) detector with a stack of separated CFRT components of (a) the 1D samples positioned on the manipulator and a single multi-pin array made of (b) stainless steel and CFRT.
The cross-sectional images allow under the premise of sufficient structural resolution, which according to ref. [25] describes the recognizability of surface structures and in the present case enables the separability of the fibers with their respective orientation from the matrix, to highlight materials with certain radiographic density via the adjustment of the contrast setting as generally described in ref. [26]. Therefore, with the used samples, it is possible to separate fiber, matrix, background, and stainless steel.
Due to the strongly differing energy-dependent X-ray absorption of the CFRT and steel joining partners, the measurement of the multi-material 2D samples is a significant challenge. As mentioned above, a problem primarily arises when the density difference of materials is too high so that material transition cannot be clearly recognized due to artifacts [27]. Consequently, it is expected that the highly differing density of stainless steel (approx. 8.0 kg/dm3), glass fiber (approximately 2.3 kg/dm3), and matrix (0.904 kg/dm3) lead to the above described multi-material artifacts.
2 Results
2.1 CT investigation
For 1D arrays, it is possible to display both the matrix and fiber material with a sufficient structural resolution. This allows for a detailed investigation of the present fiber displacement mechanisms.
For the 2D arrays, the high-density difference between the materials leads to multi-material artifacts which results in a significant limitation of the structural resolution of the CT. Especially horizontally and thus with 0° asymmetric samples, beam hardening artifacts occur in proximity to the steel pins (compare Table 3: asymmetric). One possible way to avoid these artifacts would be a second CT measurement with the samples rotated by 90°, which would locally avoid the artifacts. A second possibility would base on a pseudo-dual-energy approach as described by Butzer et al. [28]. However, the previous knowledge derived from the investigation of 1D arrays allows us to interpret the images of the 2D arrays despite the reduced resolution. Figure 6 gives an overview of the resolution of the created images in the X/Y plane.
Summary of average disturbed zones in 0° and 90°
| Array type | D dist. 0°/0 mm | D dist. 0°/1 mm | D dist. 90°/0 mm | D dist. 90°/1 mm |
|---|---|---|---|---|
| 1D-9-1.5-0°/90° (n = 3) | 11.8 ± 1.22 mm | 8.2 ± 0.90 mm | 2.3 ± 0.05 mm | 4.5 ± 1.04 mm |
| 1D-6-1.5-0°/90° (n = 4) | 8.1 ± 0.72 mm | 6.8 ± 0.65 mm | 2.9 ± 0.48 mm | 3.4 ± 0.15 mm |
| 1D-4-1.5-0°/90° (n = 4) | 9.4 ± 0.47 mm | 6.3 ± 0.29 mm | 3.1 ± 0.17 mm | 3.7 ± 0.71 mm |
| 1D-6-1.0-0°/90° (n = 4) | 6.2 ± 0.48 mm | 5.4 ± 0.69 mm | 1.7 ± 0.24 mm | 2.1 ± 0.19 mm |
| 1D-4-1.0-0°/90° (n = 4) | 7.6 ± 1.18 mm | 7.1 ± 0.39 mm | 2.0 ± 0.14 mm | 3.3 ± 0.52 mm |
| 1D-2.7-1-0°/90° (n = 5) | 4.5 ± 0.68 mm | 4.5 ± 0.50 mm | 1.8 ± 0.19 mm | 2.1 ± 0.21 mm |
Summary of CT-image in the X/Y plane.
2.2 Fiber displacement phenomena for 1D and 2D arrays
2.2.1 1D arrays
When investigating the computed tomography images of the 1D pin arrays in the Y–Z plane, it becomes apparent that the fiber reorientation is primarily characterized by a displacement in the direction of pin insertion in the Z-axis.
Figure 7 gives an example of a 1D array oriented in 0° to the fiber orientation: In the X–Z plane, it can be seen that a certain area is present, in which the fibers are dragged along the pin insertion leading to a distortion in the Z-axis. A similar phenomenon can be seen in Hufenbach et al. where the process of thermoactivated pinning was reported [29]. A likely explanation for this behavior is shear stresses as a result of the pin insertion motion causing the matrix to flow in the same direction and dragging the fibers along the same direction. In the X–Y plane, it can be seen that a disturbed zone of the fibers spans from the pin array perpendicular to the fiber orientation proceeding from the pin array. In fiber orientation, it can also be seen that a matrix-rich wedge protrudes from the first and last pin, similar as it could be seen for single-pin joints in ref. [17].
Example of 1D pin array with 0° array orientation.
When investigating a sample with a pin array at 90°, different phenomena can be observed: Figure 8 gives an example of the same pin array as Figure 7, but with a different fiber orientation: In the Y–Z plane, a significant fiber displacement in Z can be seen which is accompanied with an area of fiber compaction under the pin. In the X–Z plane, a similar displacement in the direction of pin insertion can be seen as in the X–Z plane in Figure 7, but with a higher number of pins. The X–Y plane shows distinct matrix-rich zones in direct proximity to the pin holes in the fiber orientation. It is noticeable that in this image, a thin warp roving runs horizontally through the image from left to right. It can be seen that this roving is displaced toward the bottom of the image and that the dimensions of matrix-rich disturbed zones are not significantly decreased through this glass fiber thread. It is assumed that the loose weave of the thread allows it to be displaced easily while not significantly disturbing the fiber displacement.
Example of 1D pin array with 90° array orientation.
In order to measure the influence of the dimensions of the pin array (pin diameter and pin spacing), the dimensions of the disturbed zones are measured. For 0° arrays, the disturbed zones are measured perpendicular to the fiber orientation. For 90° arrays, the measurement was conducted along the fiber orientation. For this, the disturbed zones were manually measured at the surface of the CFRT sample as well as in the middle of the sample at a depth of 1 mm. The sample size differs between three and five measurements between the different samples because the number of individual pins in an array is dependent on the spacing between pins (Table 3).
Figure 9 gives an overview of the measured disturbed zones in dependency on the fiber orientation and the depth of the used cross section for the measurement. It can be clearly seen, that the dimension of the disturbed zones in relation to both fiber directions is significantly larger for arrays with pins with 1.5 mm in relation to 1.0 mm in diameter. In 0°, the average disturbed zone for 1.5 mm over all three array types at a depth of 0 mm is 9.8 mm, and for 1.0 mm, it is 6.1 mm, which is an increase of 60%. At 90° and a depth of 0 mm, the average disturbed zone measures 2.8 mm for 1.5 mm pins and 1.8 mm for 1.0 mm pins, which is an increase of 51%. This increase in the size of the disturbed zones corresponds well with the increase in diameter, which is 50%.
Measured dimension of disturbed zones at the surface and at a depth of 1 mm in 0° and 90° in relation to the initial fiber orientation.
Furthermore, it is noticeable that the disturbed zone in 0° is larger at the surface than it is at a depth of 1 mm. One explanation for this behavior lies in the fiber displacement along the Z-axis during pin insertion. When looking at the Z–Y plane in Figure 8, a fiber bundle can be seen that is initially located at the surface of the sample and is displaced by the pin structure forming an approximately triangular displacement pattern, which corresponds with the dimension of the disturbed zone in the X–Y plane. Since the width of the triangular displacement pattern decreases with increasing depth, the dimension of the displaced zone in 0° also decreases with the depth.
The disturbed zone in 90°, however, increases with the depth, in which it is measured. This phenomenon can not only be seen in the X–Y plane but can also be reconstructed in the Y–Z plane (Figure 7). One possible explanation for this is that a lateral matrix flow caused by the pin insertion of the pin structures into the CFRT component is more pronounced at a higher depth causing an increased fiber displacement further from the CFRT surface.
When projecting these findings to the design of 2D multi-pin arrays, it is assumed that especially the disturbed zones in 0° going out from the pin structures would lead to the consistent matrix-rich zones along the fiber orientation if the pins are spaced in a distance below the dimension of the disturbed zone. A continuous matrix-rich zone would potentially lead to poor mechanical performance of the joint as it is hardly possible to introduce the load into the fibers. Consequently, when designing a pin array for practical applications, the pin spacing in 0° should be larger than the dimension of the disturbed zones leading for the used CFRT material to a minimum pin spacing of approximately 12 mm for 1.5 mm pins and 8 mm for 1 mm pins.
In 90° to the fiber orientation, an overlap of the disturbed zones could potentially lead to high fiber compaction and potential fiber damage and thus should be avoided, too. This would result in a minimum pin spacing of 90° of approximately 5 mm for 1.5 mm pins and 3.5 mm for 1 mm pins. This would lead to the following required joining area per pin: 60 mm2 for 1.5 mm pins and 28 mm2 with 1.0 mm pins leading to pin densities of approximately 2.9% for 1.5 mm pins and 2.8% for 1.0 mm pins, respectively. These values lie very close, and under consideration of the significant standard deviations in the measurements of the disturbed zones, it can be assumed that the achievable pin densities are almost identical.
A clear assumption about whether smaller or larger pins are beneficial for high transmittable loads in the joint cannot be given at the current state. It is possible, that smaller pins leading to the reduced undulation of the fibers could result in better mechanical performance under the assumption of identical achievable pin densities. However, this assumption has to be investigated in future studies.
2.2.2 2D arrays
The investigation of 2D arrays is more difficult due to the above-mentioned beam hardening artifact. Consequently, an in-detail investigation of the 1D arrays could not be performed in this study.
However, when comparing symmetric and asymmetric pin arrays, it is clearly visible that the offset is beneficial in means of fiber displacement and avoidance of overlapping matrix-rich zones. The asymmetric arrangement effectively doubles the distance between two pins in direction of the initial fiber orientation (Figure 10).
Comparison of 2D arrays with symmetric and asymmetric arrangement.
Based on these findings, it is assumed that an asymmetric arrangement could also double the achievable pin densities, which could largely improve the mechanical performance of the joints. Furthermore, the alternating undulation between the offset pin lines could improve the load introduction into the CFRT laminate when the joint is loaded at 0° to the initial fiber orientation due to a “locking effect” of the displaced fibers. Future studies are planned to investigate these assumptions in means of mechanical joint performance.
2.3 Model for fiber displacement mechanism
As could be seen, symmetric pin arrays show a displacement pattern that basically represents a stack of 0° 1D pin arrays that are joined simultaneously. Consequently, in the following section, the focus will lie on the fiber displacement mechanism for a 0° 1D pin array whereby the results from the previous sections as well as from the findings of single pin arrays in ref. [15] are used (Figure 11).
Model of fiber displacement.
The joining process is structured in three stages:
The CFRT component is heated above the melting point of the matrix and the fibers are loosely embedded in the matrix and can be rearranged. Furthermore, a certain deconsolidation of the sample can be expected, leading to slight surface defects and irregularities.
The pin array is pressed into the CFRT component, leading to a displacement of the fibers located under the main axis of the pin array. Due to the high fiber volume content in the samples, a certain amount of matrix is displaced with the fibers leaving a narrow void groove along the axis of the pin array in the CFRT component. The displaced material is accommodated through a thickening of the CFRT sample as it could also be seen for single pins in ref. [17].
The reconsolidation pressure, which in this case is approximately 1.5 MPa (area of the sample is 45 mm × 45 mm, reconsolidation force is 3,000 N), leads to a leveling of the CFRT surface and a filling of void groove with matrix and following the matrix flow also a certain amount of displaced fibers. This displacement of fibers leads to characteristic concave fiber paths as can also be seen in ref. 17 and Figure 7.
3 Conclusion
This study shows a feasible process route of the direct pin pressing process of multi-pin arrays to create continuous fiber-reinforced thermoplastics-steel hybrid parts with metallic pins and infrared heating of the CFRT. In the scope of this study, distinct fiber displacement phenomena could be shown: A major fiber displacement in Z-direction can be observed with all investigated pin arrays as it could also be shown for single pin arrays in several studies [15,17]. Furthermore, distinct disturbed zones, in which the initial fiber orientation of the CFRT laminate is distorted due to the pin insertion, could be shown in the direction of the fiber as well as perpendicular to the initial fiber orientation. These disturbed zones increase with larger pin diameter and are highly dependent on the depth in which they are measured: the size of the disturbed zone in the fiber direction decreases with increasing depth while the size of the disturbed zone perpendicular to the fiber direction increases when measure in 1 mm depth compared to measurements at the surface. Despite the larger size of the disturbed zone for 1.5 mm pins, the expected achievable pin density for 2D arrays under the assumption that pins are spaced in such a way that disturbed zones do not touch is almost identical for both investigated pin diameters and with 2.9% for 1.5 mm and 2.8% for 1.00 mm, respectively.
In means of pin arrangement in 2D arrays, it could be shown that for unidirectionally reinforced CFRT components, an asymmetric pin placement leads to favorable fiber displacement patterns which are expected to allow for a doubled pin density, assumedly leading to improved mechanical properties of the CFRT/metal hybrid joint. A detailed investigation of the dimensions of the disturbed zones was not possible due to the limited resolution of the CT images which is a result of the high-density difference between pin, fiber, and matrix.
Based on the findings of the CT measurements, a three-stage model of the fiber displacement process is proposed. This model is described by a significant fiber displacement in the direction of the pin insertion motion leading to a linear void created by the displaced fibers in the matrix. These voids are subsequently filled during the reconsolidation stage via a lateral matrix flow which drags fibers with it leading to characteristic concave fiber orientation patterns.
In future studies, it is planned to improve the quality of the computed tomography images via studies for the ideal choice of the pre-filter, especially its thickness and combination to realize a pseudo-dual-energy approach based on ref. [28] as well as CT measurements with an energy-resolving detector to improve the material separation as suggested in ref. [30].
Furthermore, mechanical analysis of different 2D pin arrays must be conducted in order to validate the assumptions made in this study. Additionally, it is planned to use cold-formed pin structures, which would allow for significantly reduced manufacturing effort as well as offer the potential to benefit from the superior mechanical strength of the pin structures due to work hardening processes. Thereby, a smoother pin structure can be expected, which could lead to reduced fiber distortion perpendicular to the initial fiber orientation to lower shear force during the joining process.
-
Funding information: Funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – TRR 285 – Project-ID 418701707.
-
Conflict of interest: Authors state no conflicts of interest.
References
[1] Siebenpfeiffer W. Leichtbau-Technologien im Automobilbau. ATZ/MTZ-Fachbuch. 1, SpringerVieweg. ISBN: 978-3-658-04024-6 2014.10.1007/978-3-658-04025-3Search in Google Scholar
[2] Agarwal BD, Broutman LJ, Chandrashekhara K. Analysis and performance of fiber composites. 4th edn. Hoboken: John Wiley & Sons, Inc; 2017. p. 5–12. ISBN: 978-81-265-3636-8.Search in Google Scholar
[3] Sommer M, Edelmann K, Wöginger A, Christmann M, Mack J, Medina L. Thermoplastische Prepregs und Halbzeuge. In: Neitzel M, Mitschang P, Breuer U, editors. Handbuch Verbundwerkstoffe. Munich: Carl Hanser Verlag; 2014. p. 147–99. 978-3-446-43696-1.Search in Google Scholar
[4] Barkoula N-M, Karger-Kocsis J. Effects of fibre content and relative fibre-orientation on the solid particle erosion of GF/PP composites. Wear. 2002;252:80–7.10.1016/S0043-1648(01)00855-9Search in Google Scholar
[5] Meschut G, Merklein M, Brosius A, Drummer D, Fratini L, Füssel U, et al. Review on mechanical joining by plastic deformation. J Adv Join Process. 2022;5:2022.10.1016/j.jajp.2022.100113Search in Google Scholar
[6] Martinsen K, Hu SJ, Carlson BE. Joining of dissimilar materials. CIRP Ann –Manuf Technol. 2015;64:696–9.10.1016/j.cirp.2015.05.006Search in Google Scholar
[7] Dawai Z, Qi Z, Xiaoguang F, Shengdung Z. Review on joining process of carbon fiber reinforced polymer and metal: Methods and joining process. Rare Met Mater Eng. 2018;47(12):3686–96.10.1016/S1875-5372(19)30018-9Search in Google Scholar
[8] Schürmann, H. Klebeverbindungen. Konstruieren mit Faser-Kunststoff-Verbunden. Springer Verlang, Berlin-Heidelberg; 2005. p. 569–604. ISBN: 978-3-540-72189-5.Search in Google Scholar
[9] Brockmann W. Adhesive bonding of polypropylene. Karger-Kocsis J. Polypropylene: An A-Z Reference. London: Kluwer; 1999. p. 1–6. ISBN: 978-94-010-5899-5.10.1007/978-94-011-4421-6_1Search in Google Scholar
[10] Mitschang P, Velthuis R, Rudolf R. Fügeverfahren für FKV. in Neitzel M., Mitschang P. Handbuch Verbundwerkstoffe. 2. Auflage, Carl Hanser, München-Wien; 2014. p. 469–82. 978-3-446-43696-1.Search in Google Scholar
[11] Kraus M, Frey P, Kleffel T, Drummer D, Merklein M. Mechnical joining without auxiliary element by cold formed pins for multi-material-systems. AIP Converence Proceedings; 2019. p. 2113.10.1063/1.5112570Search in Google Scholar
[12] Graham DP, Rezai A, Baker D, Smith PA, Watts JF. A hybrid joining scheme for high strength multi-material joints. 18th International Conference on Composite Materials; 2011.Search in Google Scholar
[13] Parkes PN, Butler R, Meyer J, de Oliveria A. Static strength of metal-composite joints with penetrative reinforcement. Compos Struct. 2014;118:250–6.10.1016/j.compstruct.2014.07.019Search in Google Scholar
[14] Mirzaei A, Darbandi AH, Ramzaninezhad M, Rezvaninasab M, Alaei MH. Experimental and numerical assessment of the effect of transverse reinforcing of AL-GFRP single lap adhesive joint using Z-pins under tensile loading. J Adhes. 2021;98:2662–8.10.1080/00218464.2021.1988577Search in Google Scholar
[15] Dhilipkumar T, Rajesh M. Influence of manufacturing process and GFRP pin loading on shear and dynamic behaviour of composite joints. J Adhes. 2022:2022:538–57.10.1080/00218464.2022.2025782Search in Google Scholar
[16] Eberl L, Avila Gray L, Zaremba S, Drechsler K. The effect of fiber undulation on the strain field for pinned composite/titanium joints under tension. Compos Part A. 2017;103:148–60.10.1016/j.compositesa.2017.09.015Search in Google Scholar
[17] Popp J, Kleffel T, Römisch D, Papke T, Merklein M, Drummer D. Fiber orientation mechanism of continuous fiber reinforced thermoplastic hybrid parts joined with metallic pins. Appl Compos Mater. 2021;28:951–72.10.1007/s10443-021-09892-0Search in Google Scholar
[18] Popp J, Kleffel T, Drummer D. Influence of pin geometry on the joint strength of CFRT-metal hybrid parts with metallic pins. Join plastics/Fügen von Kunststoffen. 2021;3:177–83.Search in Google Scholar
[19] Popp J, Drummer D. Joining of continous fiber reinforced thermoplastic/steel hybrid parts via undercutting pin structures and infrared heating. J Adv Join Process. 2022;5.10.1016/j.jajp.2021.100084Search in Google Scholar
[20] Ucsnik S, Scheerer M, Zaremba S, Pahr DH. Experimental investigation of a novel hybrid metal-composite joining technology. Compos Part A. 2010;41:369–74.10.1016/j.compositesa.2009.11.003Search in Google Scholar
[21] Feistauer EE, Amancio-Filho ST. Ultrasonic joining of lightweight alloy/fiber-reinforced polymer hybrid structures, joining of polymer-metal hybrid structures: Principles and applications. 2018;307–33.10.1002/9781119429807.ch11Search in Google Scholar
[22] Feistauer EE, Ebel T, Santos JM, Amancio-Filho ST. Ultrasonic joining of through-the-thickness reinforced Ti-4 Al-6 V and polyetherimide hybrid joints. Society of Plastics Engineers (Hrsg.): SPE ANTEC Anaheim 2017. Anaheim, USA: Society of Plastics Engineers; 2017. p. 1718–24.Search in Google Scholar
[23] Borealis AG. Product Data Sheet Polypropylene BJ100HP. 28, 2021. p. 12. https://www.borealisgroup.com/storage/Datasheets/bj100hp/BJ100HP-PDS-REG_EUROPE-EN-V7-PDS-EUR-27865-10014097.pdf.Search in Google Scholar
[24] Lanxess AG: Tepex Material Data sheet, https://lanxess.com/-/media/Project/Lanxess/Corporate-Internet/Products-and-Solutions/Brands/Tepex/Data-sheets/MDS104RG600x47.pdf, 2017.Search in Google Scholar
[25] Bartscher M, Neuschaefer-Rube U, Illemann J, Borges de Oliveira F, Stolfi A, Carmignato S. Qualification and testing of CT systems. In: Carmignato S, Dewulf W, Leach R, (eds). Industrial X-Ray Computed Tomography. Springer, Cham; 2018. 10.1007/978-3-319-59573-3_6.Search in Google Scholar
[26] DIN EN ISO 15708-3:2019-09, Zerstörungsfreie Prüfung - Durchstrahlungsverfahren für Computertomographie - Teil 3: Durchführung und Auswertung (ISO 15708-3:2017); 2019.Search in Google Scholar
[27] VDI/VDE 2630 Part 1.4:2010-06, Computed tomography in dimensional metrology - Measurement procedure and comparability; 2010.Search in Google Scholar
[28] Butzer J, Döring A, Gondrom-Linke S, Harbecke M. Dual-energy in industrial CT, How simple math can reveal structures hidden by artefacts. Proceedings of the 6th International Conference on Image Formation in X-Ray Computed Tomography; 2020. p. 94–7.Search in Google Scholar
[29] Hufenbach AW, Kupfer R, Hornig A. Thermoactivated pinning – A novel joining technique for thermoplastic composites. Solid State Phenom. 2012;188:176–81.10.4028/www.scientific.net/SSP.188.176Search in Google Scholar
[30] Bhayana R, Parakh A, Kambadakone A. Material decomposition with dual- and multi-energy computed tomography. MRS Commun. 2020;10:558–65.10.1557/mrc.2020.86Search in Google Scholar
© 2023 the author(s), published by De Gruyter
This work is licensed under the Creative Commons Attribution 4.0 International License.
