**2. Materials**

In this study, the mechanical behavior of concrete was enhanced by adding three types of chopped carbon fibers (normal carbon fiber, carbon fiber without coupling agent, and recycled carbon fiber). This section introduces the materials used in the preparation of CFRC, which include the material characteristics of carbon fiber, carbon fiber recycling technology, and coupling agen<sup>t</sup> removal process.

#### *2.1. Carbon Fiber*

The lightweight polyacrylonitrile (PAN-based) carbon fibers, which have a high tensile and specific strength, have been applied to the aerospace industry, wind turbine, sports equipment, and automotive parts. The carbon fiber was obtained from Tairylan Division, Formosa Plastics Group; then, the fiber was chopped at Sheng Peng Applied Materials Co., Ltd. [30]. The material properties are listed in Table 1 [31].


**Table 1.** Material properties of chopped normal carbon fiber.

#### *2.2. Removal of Coupling Agent on the Surface of Carbon Fiber*

The carbon fibers were immersed in pure water for one day, and then we performed GC-MS testing on the immersion solution to identify the type and boiling range of the coupling agen<sup>t</sup> and other substances. The coupling agen<sup>t</sup> was found to be C26H48O3Si with a molecular weight of 436.74 g/mol under the boiling point of 507.1 ± 50.0 ◦C at a pressure of 760 mm-Hg [18]. The removal of the coupling agen<sup>t</sup> process is shown in Figure 1; the carbon fibers were wrapped with aluminum foil and then placed into a Muffle furnace (PF-40, Chuan-Hua Precision, New Taipei City, Taiwan) at a high temperature of 550 ◦C for 3 h.

**Figure 1.** Coupling agen<sup>t</sup> removal processes. 1. Carbon fibers were wrapped with aluminum foil; 2. Placed the carbon fibers with aluminum foil into a Muffle furnace; 3. Set the furnace temperature at 550 ◦C for 3 h.

Generally, the coupling agen<sup>t</sup> is an inorganic compound, and it reduces the adhesion between the fiber and cement. The differentiation of the coupling agen<sup>t</sup> presence and absence on the surface of carbon fiber is shown in Figure 2. The carbon fiber is easier to distribute in the concrete by removing the coupling agent, compared with normal carbon fiber.

**Figure 2.** The appearance of chopped carbon fiber: (**a**) Normal carbon fiber, (**b**) Carbon fiber without coupling agent.

#### *2.3. Recycled Carbon Fiber*

Carbon fiber tow is the thread used to weave carbon fiber fabrics. As a standalone product, it can be used to make wound parts, in pultrusion, or chopped as a local reinforcement. This 12 k tow (or yarn) comprises 12,000 individual carbon filaments, which boast the highest ultimate tensile strength in the industry. The microwave-assisted pyrolysis (MAP) is different from the traditional heating method. It uses microwaves to quickly rub the molecules in the substance, prompting the molecules to rotate quickly to generate heat, and then quickly decompose. The characteristics of microwave heating are: Microwaves only aim at the materials that can absorb them, so energy use is more concentrated. The electric power system is used to generate microwaves, it can be quickly heated, the process is easy to control, and automatic control can be realized. Microwaves penetrate to the absorbable substances, so they can heat the entire object uniformly [32,33].

In the MAP technology, the microwave heat radiations can transfer through the carbon fiber from inner to outer. It can remove the resin from the CFRP bicycle frame scrap to transfer into recycled carbon fiber. Figure 3 shows the microwave machine (PYRO 260, Milestone, Sorisole, Italy) at the Department of Molecular Science and Engineering Laboratory of the National Taipei University of Technology, Taipei, Taiwan. The recycled carbon fiber was obtained from the carbon fiber scrap by using the MAP process in the microwave machine at 950 ◦C for 1 h, and then a single filament tensile test was carried out.

**Figure 3.** The microwave machine.

In this study, both carbon fiber and recycled carbon fiber groups were tested in the single filament tensile test and carried out in accordance with ASTM D3379 [34]. Each group of carbon fibers had 30 individual filaments. The load–displacement relationships of the normal and recycled carbon fibers are shown in Figure 4.

**Figure 4.** The load–displacement curves of the single filament tensile test: (**a**) Normal carbon fiber, (**b**) Recycled carbon fiber.

The single filament tensile test results show that the tensile strength of recycled carbon fiber treated with MAP is similar to that of the normal carbon fiber. This proves that MAP technology can effectively recycle CFRP waste and retain its mechanical strength.

#### *2.4. SEM Surface Morphology of Carbon Fiber*

In order to understand the cleanliness of the resin on the surface of the recycled carbon fiber, in addition to observing the samples before and after the MAP treatment with a scanning electron microscope (SEM), energy dispersive X-ray spectrometry (EDS) was also used to observe the amount of removal. Figure 5a–d shows the photos and SEM images of the waste (recycled) CFRP before and after MAP technology. Using EDS to measure the carbon content of the carbon fiber, the carbon content of the recycled carbon fiber after MAP technology was 99.8%. The SEM image is shown in Figure 5d.

**Figure 5.** Photos and SEM images of fragment samples of recycled carbon fiber composite materials. (**a**) Fragment samples of recycled CFRP before MAP technology; (**b**) SEM image before MAP technology; (**c**) fragment samples of recycled CFRP after MAP technology; (**d**) SEM image after MAP technology.

The surface morphology of the chopped carbon fiber with and without coupling agen<sup>t</sup> was analyzed by scanning electron microscope (model: JSM-7610F, JEOL, Tokyo, Japan) in the Department of Molecular Science and Engineering Laboratory of the National Taipei University of Technology, Taipei, Taiwan. The surfaces of normal carbon fiber (with coupling agent) and carbon fiber without coupling agen<sup>t</sup> were observed, and their SEM images are shown in Figure 6. Using EDS to measure the carbon content of the carbon fiber, the carbon contents of the normal carbon fiber with and without furnace heating process were 100% and 99.2%, respectively. The coupling agen<sup>t</sup> on the surface of the carbon fiber might have interfered with the bonding strength between the carbon fiber and cement by the furnace heating method.

**Figure 6.** SEM observation results of the surface of normal carbon fiber and carbon fiber without coupling agent. (**a**) SEM image of the chopped carbon fiber without furnace heating; (**b**) SEM image of the chopped carbon fiber with furnace heating.

#### *2.5. Carbon Fiber-Reinforced Concrete*

Carbon fiber is a non-corrodible material. The CFRC can reduce the cracks during the service life of concrete structures, it can preserve the steel rebar from corrosions, and mitigate impact loading. In this study, Portland cement was obtained from the Taiwan Cement Corporation [35]. The concrete (benchmark) and CFRC specimens were tested under compressive, three-point bending, and impact tests. The concrete water–cement ratio was 0.6; the cement, sand, fine aggregate, and coarse aggregate ratio was 1:1.05:1.5:0.75. The fineness modulus of the fine aggregate (3/8) and coarse aggregate (6/8) were 3.03 and 7.33, respectively. The fineness modulus (F.M.) of both aggregates for the concrete specimen was 6.01, as shown in Table 2. The concrete and CFRC specimens were cured at 28 days.

**Table 2.** Fineness modulus of aggregates.


#### **3. Experimental Methods and Setups**

In this study, the CFRC specimens were prepared using three kinds of carbon fiber (recycled carbon fiber, normal carbon fiber, and carbon fiber without coupling agent). The compressive test, flexural test, and impact test followed the ASTM and ACI standards.

#### *3.1. Experiment Planning*

The three types of carbon fiber (normal carbon fiber, recycled carbon fiber, and carbon fiber without coupling agent) were planned to be used to prepare CFRC specimens with three different fiber weight proportions (5 ‰, 10 ‰, and 15 ‰), and the specimen names and descriptions, and the planning of the specimens are shown in Tables 3 and 4, respectively. For example, specimen C-R05 represents the compression test of recycled carbon fiber with the addition of 5‰ weight proportion.

**Table 3.** The naming and descriptions of CFRC specimens.


**Table 4.** Planning of CFRC specimens.


#### *3.2. Slump Test*

The workability of the fiber-reinforced concrete has a significant impact on the construction quality, and good workability will prevent honeycombs in the concrete that reduce its strength. In this study, the slumps of CFRC of different types of carbon fiber and different weight proportions were tested according to ASTM C143/C143M −20; the slump fluidity range was about 15–230 mm, respectively [36].

#### *3.3. Compressive Test*

To explore the compressive strength of the three types of CFRC with different fiber additions, the dispersed chopped carbon fiber was mixed into concrete at a cement weight ratio of 5 ‰, 10 ‰, and 15 ‰. The diameter of the CFRC cylindrical specimen was 10 cm and the height was 20 cm. According to ASTM C39/C 39M-01 [37], the universal testing machine (HT-9501 Series. Hong-Ta, Taipei, Taiwan) was used for testing.

#### *3.4. Three-Point Bending Test*

According to ASTM C293-02 [38], the bending test of 28 cm × 7 cm × 7 cm CFRC specimens was carried out by a universal testing machine (HT-9501 Series. Hong-Ta, Taipei, Taiwan) with a load cell (WF 17120, Wykeham Farrance, Milan, Italy). The bending

specimen was tested in the material laboratory of the Department of Civil and Disaster Prevention Engineering, National Taipei University of Technology. The three-point bending test is shown in Figure 7.

**Figure 7.** CFRC specimen flexural test setup.

## *3.5. Impact Test*

According to ACI 544-2R [39], the impact test of *φ*150 mm × 64 mm CFRC specimens was carried out by impact equipment (SP-006, Sheng Peng, Yunlin, Taiwan). The impact energy was 25 J as the interval, and the number of repeated impacts was measured from 50 to 150 J. Figure 8a shows the impact specimen and impact device, and Figure 8b shows the impact test equipment.

**Figure 8.** Impact test: (**a**) CFRC with impact device, (**b**) impact test equipment.

#### **4. Experimental Results and Discussions**

Three types of CFRC specimens were prepared from different fibers (recycled carbon fiber, normal carbon fiber, and carbon fiber without coupling agent). The test results of compressive, bending, and impact performance were obtained with the different kinds of carbon fiber in CFRC specimens. The recycled carbon fibers were fabricated by a scrap of CFRP bicycle frame with a MAP technology. The recycled carbon fiber was adopted from Thermolysis Co., Ltd. (Kao-Hsiung, Taiwan) [40]. The recycled carbon fiber was chopped to a length between 20 and 30 mm. The microwave conditions were established according to the process in Section 2.3.

#### *4.1. Slump Test Result*

The workability depends on the w/c ratio and weight proportions of fibers. The slump values of CFRC with different types of carbon fiber and different weight proportions are shown in Table 5. The test results showed that the slump value was not affected by the types of carbon fibers but affected by different weight proportions. The recycled carbon fiber with a length of 20–30 mm had a slight variation in the slump according to carbon fiber coupling agen<sup>t</sup> presence and absence. The CFRC with 5‰ fiber weight proportion of the carbon fiber had the best workability, and its slump was about 160 mm. The CFRC with 15‰ fiber weight proportion of the carbon fiber became stickier, making the CFRC wet mixing hard to mix and sometimes diminished its mechanical strength. The CFRC mixture with more than 15‰ weight proportion had less workability under 0.60 w/c ratio.

**Table 5.** Slump test of different fiber weight proportions.


#### *4.2. Compressive Test Result*

In this study, the three different kinds of carbon fiber-reinforced concrete were subjected to compression test by uniaxial loading. Figure 9 shows the average compressive strength of CFRC and benchmark under different proportions. Compared with other fiber weight proportions, adding 10‰ fiber weight proportion of CFRC can increase the maximum compressive strength.

**Figure 9.** Average compressive strength of CFRC and benchmark specimens. (Note: C—Compression test; B—Benchmark; R—Recycled carbon fiber; N—Normal carbon fiber; W—Carbon fiber without coupling agent).

Table 6 shows the compressive strength of different CFRC and benchmark under different fiber weight proportions. Under three different carbon fiber weight proportions, the recycled carbon fiber-reinforced concrete and normal carbon fiber-reinforced concrete did not exhibit a greater enhancement effect on compressive strength than the carbon fiberreinforced concrete by removing the coupling agent. For instance, the C-W10 specimen exhibited the highest compressive strength (33.19 MPa) compared with C-B specimen, C-N10 (30.69 MPa), and C-R10 (30.49 MPa), as shown in Table 6.


**Table 6.** Compressive strengths of CFRC and benchmark under different proportions.

Note: C—Compression test; B—Benchmark; R—Recycled carbon fiber; N—Normal carbon fiber; W—Carbon fiber without coupling agent.

When the fiber weight proportion was 10‰, the compressive strength of specimen C-W10 was the highest, followed by specimen C-N10 and specimen C-R10. Using EDS to measure the carbon content of the carbon fiber, the carbon contents of the normal carbon fiber with the furnace heating process, the recycled carbon fiber with the MAP process, and the normal carbon fiber were 100, 99.8, and 99.2%, respectively. The test results showed that the higher the carbon content of the carbon fibers, the higher the compressive strength of the CFRC specimen. The surface of the carbon fiber without coupling agen<sup>t</sup> or resin residual, the adhesion force between carbon fiber and cement increased. Therefore, the CFRC specimen without coupling agen<sup>t</sup> (C-W10) has the highest compressive strength.

#### *4.3. Three-Point Bending Test Result*

In this Subsection, the flexural strength of three kinds of CFRC with different fiber weight proportions are compared with the benchmark specimens. As shown in Figure 10, the CFRC with 10 ‰ of fiber weight proportion increased its flexural strength more than the other fiber weight proportions, such as 5 ‰ and 15 ‰. From the slump test results, the 15 ‰ fiber weight proportion of CFRC was close to the lower range compared with the ASTM C143/C143M-20 standard requirement (mention in Section 3.2), respectively. Therefore, the carbon fibers were not easy to distribute uniformly in the FRC, and the flexural strength was reduced. Additionally, the CFRC with 5 ‰ fiber weight proportion of carbon fiber was not enough to increase the flexural strength because the amount of carbon fiber was too low.

**Figure 10.** Average flexural strengths of CFRC and benchmark specimens. (Note: F—Flexural; B—Benchmark; R—Recycled carbon fiber; N—Normal carbon fiber; W—Carbon fiber without coupling agent).

Table 7 shows the flexural strength of CFRC and benchmark under different fiber weight proportions. The F-W10 specimen increased its flexural strength up to 50.4% compared with F-B specimen. Similarly, the flexural strength of F-R10 and F-N10 had higher strength than the benchmark specimen at 46.3 and 39.6%, respectively. In Section 2.4, the SEM images with the corresponding EDS showed the carbon content of carbon fiber without coupling agent, recycled carbon fiber, and normal carbon fiber were 100, 99.8, and 99.2%, respectively. As seen from the flexural test results and the EDS measurement, the flexural strength increased with the carbon contents of the carbon fiber; the fewer residuals increased the flexural strength of the CFRC specimens.

ġ


**Table 7.** Flexural strengths of CFRC and benchmark under different proportions.

Note: F—Flexural; B—Benchmark; R—Recycled carbon fiber; N—Normal carbon fiber; W—Carbon fiber without coupling agent).

#### *4.4. Impact Test Result*

The impact numbers of benchmark and CFRC specimens under different impact energies are shown in Table 8. From the test results, the I-N10, I-R10, I-W10 specimens resisted repeated impact at low energy of 50 J and the average impact numbers were about 285, 356, and 410, respectively. Compared with the I-B specimen, the I-N10, I-R10, and I-W10 specimens and the impact number increase percentages were about 1828%, 2305% and, 2669%, respectively. From the impact test result, the CFRC specimens with recycled carbon fiber and carbon fiber without coupling agen<sup>t</sup> exhibited enhancement effects compared to the CFRC specimen with normal carbon fiber.


**Table 8.** Impact test results of CFRC specimens under different impact energies.

Note: I—Impact; B—Benchmark; N—normal carbon fiber; R—recycled carbon fiber; W—carbon fiber without coupling agent.

> Figure 11 shows the impact energy/number curve of CFRC specimens. It can be clearly seen that the repeated impact number of CFRC specimens under low impact energy is greater than that under high impact energy.

> As seen in Figure 11, the I-W10 specimen had the highest average number of repeated impacts when the impact energy was 50 J, followed by I-R10 and I-N10. The EDS measurement results and impact test results show that the higher carbon contents of the carbon fiber had higher repeated impact capability of CFRC.

**Figure 11.** The curves of impact energy/number of CFRC specimens. (Note: I—Impact; B— Benchmark; N—Normal carbon fiber; R—Recycled carbon fiber; W—Carbon fiber without coupling agent).
