Investigation of the Internal Structure of Fiber Reinforced Geopolymer Composite under Mechanical Impact: A Micro Computed Tomography (CT) Study

The internal structure of fiber reinforced geopolymer composite was investigated by microfocus X-ray computed tomography (μCT) under mechanical impact. μCT is a non-destructive, multi approach technique for assessing the internal structures of the impacted composites without compromising their integrity. The three dimensional (3D) representation was used to assess the impact damage of geopolymer composites reinforced with carbon, E-glass, and basalt fibers. The 3D representations of the damaged area with the visualization of the fiber rupture slices are presented in this article. The fiber pulls out, and rupture and matrix damage, which could clearly be observed, was studied on the impacted composites by examining slices of the damaged area from the center of the damage towards the edge of the composite. Quantitative analysis of the damaged area revealed that carbon fabric reinforced composites were much less affected by the impact than the E-glass and basalt reinforced composites. The penetration was clearly observed for the basalt based composites, confirming μCT as a useful technique for examining the different failure mechanisms for geopolymer composites. The durability of the carbon fiber reinforced composite showed better residual strength in comparison with the E-glass fiber one.


Introduction
Over the past 35 years, microcomputed tomography (µCT) has proved to be an effective and standard tool for quantifying structure-function relationships, disease progression, and preclinical models, thus contributing to scientific and bioengineering advancements [1]. µCT is a non-destructive imaging technique for the production of high-resolution three-dimensional (3D) images composed of two-dimensional (2D) trans-axial projections, or "slices", of a target specimen. Its principle is based on the attenuation of X-rays passing through the object or sample being imaged. As an X-ray passes save weight, minimize resin void size, and maintain fiber orientation during the fabrication process. In general, impact events cause combinations of damages [16]. The resulting damage may include significant fiber failure, matrix cracking, delamination, and debonded elements. Damage caused by low-energy impact is more contained but may also include a combination of broken fibers, matrix cracks, and multiple delaminations. Thus, it is essential to study the fiber orientation and the overall damage in fiber at the composite after impact under mechanical load [17].
Based on this assumption, a study was carried out on fiber reinforced geopolymer composite under mechanical load. The inner view of the composite on the point of structural integrity was carried out by investigating the damage of the fiber composite by the micro computed tomography method after mechanical impact. Although theoretical predictions about fiber orientation and porosity are presented, their combination with an experimental approach using micro computed tomography are not thoroughly addressed [18].
In a previous work, impact tests were carried out with a drop weight method on the surface of the composite material, and the residual strength of the composite in correlation with damage areas was explained [19]. Composites have complex characteristics and the presence of defects or damage will decrease their strength, thus initiating their ultimate failure [20].
In this work, we studied geopolymer composites with various fiber reinforcements, namely carbon, E-glass, and basalt. One general outcome (and the outcome of the present study) is the need to perform a detailed observation of the sample inner surface quality without any destructive changes. The geometrical defects that develop due to impact inside the fabric reinforced composite may play a determining role in the strength and life of the composite. The internal defects include fiber pull out, misalignment of the fiber, cracks, and damage to the matrix. In this work, we addressed the detailed statistical characterization of geometrical defects in the impacted fabric reinforced composite with 3D images captured by µCT to visualize the interior structural details with high resolution on a scale of interest for damage evaluation of the composite [21,22]. The relationships between the damaged area/volume and impact energies were established, and the composite was characterized, not only investigating the volume of the damaged area, but also the internal damage area including the internal structure and fiber breakage on the various layers of the composite structure. µCT was used to investigate the damaged area, fiber breakage from the surface to internal layers of the composite, delamination, and rupture of the fibers. Overall sustainability and durability of the composite were predicted by the µCT technique depending on the internal 3D image determined.

Materials
Geopolymer binder was prepared by mixing 49% alumina-silicate powder, 44.12% alkali activator containing NaOH/KOH, and 6.88% metakaolin powder. The combination was stirred for 10 min by a home-made mixer machine for a complete homogenous mixture.
The assembled fabric reinforced geopolymer composites were placed in a vacuum bag and cured under 0.003 MPa at room temperature for 24 h. The bag with the composite was then placed in a curing oven at 70 • C for 12 h [12]. To maintain a thickness of 3 mm, the fabrics were arranged in different layers, such as 7 layers of E-glass, 10 layers of carbon, and 15 layers of basalt fabric. The volume fraction of the fiber, the matrix, and the voids of the three laminates were calculated according to the formulas reported in Equations (2)-(4): where V f , V m , V v are the volume fractions of fibers, matrix, and voids in a sample, n is the number of fiber layers in a composite sample, and t, L, and b are the thickness, length, and width of a composite sample, respectively. The volume fractions of fiber, matrix, and voids of the three laminates are reported in Table 1.
characterized, not only investigating the volume of the damaged area, but also the internal damage area including the internal structure and fiber breakage on the various layers of the composite structure. µCT was used to investigate the damaged area, fiber breakage from the surface to internal layers of the composite, delamination, and rupture of the fibers. Overall sustainability and durability of the composite were predicted by the µCT technique depending on the internal 3D image determined.

Materials
Geopolymer binder was prepared by mixing 49% alumina-silicate powder, 44.12% alkali activator containing NaOH/KOH, and 6.88% metakaolin powder. The combination was stirred for 10 min by a home-made mixer machine for a complete homogenous mixture.
Composite samples (100 × 100 × 3 mm 3 ) were prepared by hand lay up technique with fabrics of carbon (Figure 1a), E-glass (Figure 1b), and basalt ( Figure 1c) reinforcement in alumino-silicate geopolymer matrices with metakaolin binders using piles of fabrics in the 0-90° direction [12].  The geopolymer matrix composition was 2.04, 31.80, 0.08, 15.15, 1.74, 0.63, 0.24, and 48.32% in Al, Si, P, K, Zr, Na, Ca, and O, respectively. The samples were impacted by the drop weight method, and the damage was created at the center of the material ( Figure 2). The assembled fabric reinforced geopolymer composites were placed in a vacuum bag and cured under 0.003 MPa at room temperature for 24 h. The bag with the composite was then placed in a curing oven at 70 °C for 12 h [12]. To maintain a thickness of 3 mm, the fabrics were arranged in different layers, such as 7 layers of E-glass, 10 layers of carbon, and 15 layers of basalt fabric. The volume fraction of the fiber, the matrix, and the voids of the three laminates were calculated according to the formulas reported in Equations 2, 3, and 4: where Vf, Vm, Vv are the volume fractions of fibers, matrix, and voids in a sample, n is the number of fiber layers in a composite sample, and t, L, and b are the thickness, length, and width of a composite sample, respectively. The volume fractions of fiber, matrix, and voids of the three laminates are reported in Table 1. The geopolymer matrix composition was 2.04, 31.80, 0.08, 15.15, 1.74, 0.63, 0.24, and 48.32% in Al, Si, P, K, Zr, Na, Ca, and O, respectively. The samples were impacted by the drop weight method, and the damage was created at the center of the material (Figure 2). The indenter was struck from the height of 0.50 m on the surface of the composite to create a notch. Five specimens for each material system with dimensions of 100 × 130 × 3 mm 3 were tested with an in-house built drop weight factor. A hardened steel striker with a hemispherical tip of 16 mm diameter was impacted on the sample from a chosen drop height. The weight of the impactor was considered to be 3.072 kg. The energy, 6.2 J, was achieved from the chosen drop height of 0.50 m. A schematic diagram of the system is reported in Figure 3. The samples were examined in detailed damage areas using the µCT techniques before and after the impact test. The specimens for the µCT investigation were cut with a dimension of 50 × 50 × 3 mm 3 around the central damage area. The distance of the outer edge was 25 mm towards the boundary from the center of impact and the notch generated on the composite surface.

Micro Computed Tomography
The µCT analysis was carried out in a SkyScan 1272 (Bruker, Kontich, Belgium) with a source voltage of 100 kV and a source current of 100 µA using a camera pixel size of 16 µm. Slices of the µCT images were collected from the impacted composite by scanning their images with the rotation of 0.28°. 3D images were built using the CT Vox graphics software (Bruker, Kontich, Belgium; parameters in Table 2). µCT imaging not only affords advantages over the conventional approach but also eliminates the tedious sample preparation.
Besides analyzing internal structures of the composite, µCT provides a better understanding of the behavior and damage of the composite materials undergoing impact loading. The µCT method helps to detect and measure the fiber breakage, which leads to the basic understanding of the mechanisms of crack growth in composites. Three samples were tested for each individual specimen by the µCT method before and after damage. µCT slices were obtained and examined at each degree of composites rotation to observe differences in fiber rupture and matrix debonding. More than 450 slices were produced by µCT for each sample.  The samples were examined in detailed damage areas using the µCT techniques before and after the impact test. The specimens for the µCT investigation were cut with a dimension of 50 × 50 × 3 mm 3 around the central damage area. The distance of the outer edge was 25 mm towards the boundary from the center of impact and the notch generated on the composite surface.

Micro Computed Tomography
The µCT analysis was carried out in a SkyScan 1272 (Bruker, Kontich, Belgium) with a source voltage of 100 kV and a source current of 100 µA using a camera pixel size of 16 µm. Slices of the µCT images were collected from the impacted composite by scanning their images with the rotation of 0.28 • . 3D images were built using the CT Vox graphics software (Bruker, Kontich, Belgium; parameters in Table 2). µCT imaging not only affords advantages over the conventional approach but also eliminates the tedious sample preparation. Besides analyzing internal structures of the composite, µCT provides a better understanding of the behavior and damage of the composite materials undergoing impact loading. The µCT method helps to detect and measure the fiber breakage, which leads to the basic understanding of the mechanisms of crack growth in composites. Three samples were tested for each individual specimen by the µCT method before and after damage. µCT slices were obtained and examined at each degree of composites rotation to observe differences in fiber rupture and matrix debonding. More than 450 slices were produced by µCT for each sample.
For the carbon fabric composite, a selection of the slices through the damaged area was done ( Figure 4). For the carbon fabric composite, a selection of the slices through the damaged area was done ( Figure 4). Using all of the slices, a 3D reconstruction image was created with the Vox software ( Figure 5).

Results and Discussion
An area of 6.1 × 6.5 mm 2 was measured in carbon fabric reinforced composite (Figure 6a1-c1), showing a circular deformation on this composite surface. In longitudinal and transverse directions, broken fiber bundles in carbon fabric reinforced geopolymer composite were seen (Figure 6a2-c2), and the gap between the fibers became larger near the back side of the impact.

Results and Discussion
An area of 6.1 × 6.5 mm 2 was measured in carbon fabric reinforced composite (Figure 6a1-c1), showing a circular deformation on this composite surface. In longitudinal and transverse directions, broken fiber bundles in carbon fabric reinforced geopolymer composite were seen (Figure 6a2-c2), and the gap between the fibers became larger near the back side of the impact.

Results and Discussion
An area of 6.1 × 6.5 mm 2 was measured in carbon fabric reinforced composite (Figure 6a1-c1), showing a circular deformation on this composite surface. In longitudinal and transverse directions, broken fiber bundles in carbon fabric reinforced geopolymer composite were seen (Figure 6a2-c2), and the gap between the fibers became larger near the back side of the impact. For E-glass reinforced composites (Figures 7 and 8), the notch formation in the central region was much more pronounced than in the carbon fibers. Comparing Figures 4 and 7, it is clear that the damage stretched over a much wider area. The depth of the hole was about half the thickness of the specimen and seemed to be thicker at the center of the impact.
For E-glass reinforced composites (Figures 7 and 8), the notch formation in the central region was much more pronounced than in the carbon fibers. Comparing Figures 4 and 7, it is clear that the damage stretched over a much wider area.   Even in the top layer, the fiber rupture was more pronounced, which was evident from the slices in a longitudinal and transverse directions (Figures 9a1-c1 and a2-c2). Fiber sliding was observed in the composite due to the slippery behavior of E-glass fabrics and the weak bonding between the fiber and the matrix of the composite. Fiber delamination on the Even in the top layer, the fiber rupture was more pronounced, which was evident from the slices in a longitudinal and transverse directions (Figure 9a1-c1,a2-c2). Even in the top layer, the fiber rupture was more pronounced, which was evident from the slices in a longitudinal and transverse directions (Figures 9a1-c1 and a2-c2). Fiber sliding was observed in the composite due to the slippery behavior of E-glass fabrics and the weak bonding between the fiber and the matrix of the composite. Fiber delamination on the Fiber sliding was observed in the composite due to the slippery behavior of E-glass fabrics and the weak bonding between the fiber and the matrix of the composite. Fiber delamination on the various layers was observed in the composite in the area range of 50 mm towards its longitudinal direction. The slipping behavior of fibers within the matrix of the composite, as well as the loosening of the matrix within the damage zone, was very prominent. The depth of the hole was about twice the thickness of the sample. On the backside, fiber bundles were broken. Thus, the number of layers decreased when one approached the center of impact. This explains why the composite seemed to be thinner in the region of the impact.
The impact on the basalt geopolymer composite (Figure 10) was penetrated through its surface. There was extensive damage in the central area that spread out until the edge (Figure 10, slice 450). The specimen was completely deformed, and even the regions outside the impact zone remained permanently twisted due to the impact, as is clearly shown from the top images in Figure 10 where the two "flat" parts no longer lay in the same plane. various layers was observed in the composite in the area range of 50 mm towards its longitudinal direction. The slipping behavior of fibers within the matrix of the composite, as well as the loosening of the matrix within the damage zone, was very prominent. The depth of the hole was about twice the thickness of the sample. On the backside, fiber bundles were broken. Thus, the number of layers decreased when one approached the center of impact. This explains why the composite seemed to be thinner in the region of the impact. The impact on the basalt geopolymer composite (Figure 10) was penetrated through its surface. There was extensive damage in the central area that spread out until the edge (Figure 10, slice 450). The specimen was completely deformed, and even the regions outside the impact zone remained permanently twisted due to the impact, as is clearly shown from the top images in Figure 10 where the two "flat" parts no longer lay in the same plane. As a result, the remainder of the composite could only carry very negligible loads compared to the original value before the damage [23]. The overall image of the basalt fiber reinforced geopolymer composite is shown in Figure 11. The impact area was observed with a fiber rupture region from the top along the direction of the central impact rupture. As a result, the remainder of the composite could only carry very negligible loads compared to the original value before the damage [23]. The overall image of the basalt fiber reinforced geopolymer composite is shown in Figure 11. The impact area was observed with a fiber rupture region from the top along the direction of the central impact rupture.     For the E-glass and the carbon fabrics, the delamination was visible, but the accompanying expansion in thickness was limited. The thickness of the basalt composite was more than double in the affected area. This behavior was attributed to the strong chemical interactions and the mineral exchange among the basalt fabrics and the geopolymer matrix [24]. Thus, strong chemical bonding transforms the composite-like cement structure, and above all, it weakens the fibers, showing the fragile behavior of the composite structure. The carbon fiber pull out around the impact zone was well explained by the concept of the laser scanner, and these observations match well with ours. The unidirectional carbon fibers in an epoxy matrix and the arrangement of the fibers in the laminar plane were studied using µCT [16], highlighting different failure modes of the composite and internal damage [25,26] that agreed with the fiber damage concept presented in this study.
The central damage areas of the carbon, the E-glass, and the basalt fabric composites got larger in that order (Figure 13a-c). Fiber breakage and matrix cracks were observed in all cases [SEM micrographs of the damaged samples for E-glass ( Figure S1) and basalt ( Figure S2) fabric reinforced composites are reported as examples in the Supplementary Materials], however, for carbon composites, the area outside the range of 25 mm showed no change in appearance. The fiber damage within the central region was observed while the rest of the composite was intact with good adhesion of fibers within the matrix. The severely damaged area of the carbon geopolymer composite was 15 ± 2 mm 2 (calculated from the picture of the damaged area at the impact side), but the damage stretched over a larger circumferential area. For the E-glass and the carbon fabrics, the delamination was visible, but the accompanying expansion in thickness was limited. The thickness of the basalt composite was more than double in the affected area. This behavior was attributed to the strong chemical interactions and the mineral exchange among the basalt fabrics and the geopolymer matrix [24]. Thus, strong chemical bonding transforms the composite-like cement structure, and above all, it weakens the fibers, showing the fragile behavior of the composite structure. The carbon fiber pull out around the impact zone was well explained by the concept of the laser scanner, and these observations match well with ours. The unidirectional carbon fibers in an epoxy matrix and the arrangement of the fibers in the laminar plane were studied using µCT [16], highlighting different failure modes of the composite and internal damage [25,26] that agreed with the fiber damage concept presented in this study.
The central damage areas of the carbon, the E-glass, and the basalt fabric composites got larger in that order (Figure13 a,b,c). Fiber breakage and matrix cracks were observed in all cases [SEM micrographs of the damaged samples for E-glass ( Figure SM1) and basalt ( Figure SM2) fabric reinforced composites are reported as examples in the supplementary materials], however, for carbon composites, the area outside the range of 25 mm showed no change in appearance. The fiber damage within the central region was observed while the rest of the composite was intact with good adhesion of fibers within the matrix. The severely damaged area of the carbon geopolymer composite was 15 ± 2 mm 2 (calculated from the picture of the damaged area at the impact side), but the damage stretched over a larger circumferential area. This can be seen in Figure 5, where the affected zone was larger on the surface. The E-glass and the basalt composites showed only damage in the clearly distorted area with no effective changes in other parts of the composite. The pull out mechanism of the carbon fiber effectively impacted other parts of the composite, probably due to the inter shear stress mechanism [27]. The sliding mechanism of the E-glass fibers reduced the inter shear effect, while the basalt composite was penetrated suddenly, almost as if no fibers were present. The E-glass composites presented a severely damaged area in the range of 42 ± 3 mm 2 . The basalt reinforced composite had a severely damaged area of 97 ± 5 mm 2 . The area of damage is also represented in Figure 13, where the green part represents the non-affected area and the black one is the heavily distorted area. Not only was the affected area much larger for the basalt reinforced composites, but the shape of the impact was different as well. The non-spherical deformation was caused by the orientation of the textile fibers, which can also be seen in Figure 12c, where the deformation of the fiber bundles is very clear. The theoretical and experimental approach of woven fiber in the carbon composite at various axis of damage during dynamic loading was described in literature [28] and well matches with the damage behavior observed here, such as matrix cracking and fiber debonding on the composite. The overall fiber orientation and rupture through the composite from the edge towards the center is supported by the supplementary file Video S1.
The parameters for the 3D image construction and some results were compared to the carbon, the E-glass, and the basalt reinforced geopolymer composites ( Table 3). The pixel size was fixed for all group images to 16 µm. The value of open porosity in the case of the carbon fabric reinforced composite showed 17%, while the E-glass fabric showed 43%, and the basalt composite showed the extreme value of 68% porosity. The video shows the breakages of the fiber from the center to the edge and consists of more than 450 slices.

Conclusions
µCT reveals the internal geometry of the fabric reinforced composite materials, showing the overall image of the composite from the impact point towards the edge of the sample and the fiber breakage, notch, or depth of the impact areas. The advantages of this technique are that it not only shows the impact area but also very clearly shows the fiber rupture in the different slice images. Furthermore, other parameters-like porosity and sample position-can be obtained from the 3D image. The mechanism of damage of the composite is well explained on the basis of the µCT investigation of the damaged area. The following conclusions can be deduced from the obtained results: 1. Fabric reinforced composites were much less affected by the mechanical impact than the E-glass and the basalt reinforced composites.
2. As examined for basalt composites, µCT was found to be one of the most useful techniques for examining the different failure mechanisms for geopolymer composites.
3. µCT provided the internal geometry (including 3D image) of the fabric reinforced composite materials, including parameters like porosity and sample position. This is the prime advantage of µCT over other techniques. 4. Our study showed that the durability of the carbon fiber reinforced composite showed better residual strength in comparison with the E-glass fiber one.
5. The carbon fiber reinforced composite had a much better impact resistance, with fiber damage only in the central area with a limited amount of fiber sliding in the surrounding areas.
Supplementary Materials: The following are available online at http://www.mdpi.com/2076-3417/9/3/516/s1, Figure S1: SEM images of E-glass geocomposite after damage, showing matrix rupture, fibre pull out and non-adhesion. Figure S2. SEM images of Basalt reinforced geocomposite after impact test, showing the fiber breakage, matrix cracks, non-adhesion of fabric-matrix interface, Video S1: 3D analysis of basalt fiber from center to edge in basalt fiber reinforced composite

Conflicts of Interest:
The authors declare no conflict of interest.