P.O 42.5 normal Portland cement from Bohai Cement (Huludao) Co., Ltd. was used in this test. The local granite gravel was used as the coarse aggregate which had a continuous gradation in the 5-20 mm particle size range. The local ordinary river sand in Jinzhou was used as fine aggregate which had a fineness modulus of 2.7. Table 1 and Table 2 show the physical properties of coarse and fine aggregate, respectively. The granulometry curves of aggregates from the tests according to GB/T 14685-2022 (28) and GB/T 14684-2022 (29) are shown in Figure 1. Urban tap water from Jinzhou City Tap Water Plant was used as the water. The BF and CF were from Jiangsu Tianlong Continuous Basalt Fibre Co., Ltd. and Taian Tongban FIBER Co., Ltd, respectively. The BFRP bars were 12 mm in diameter and 500 mm in length, with a sand-coated ribbed surface treatment, from Jiangsu Green Material Valley New Material Technology Development Co., Ltd. Table 3 and Table 4 show the material properties of fibers and BFRP bars, respectively. Figure 2 shows the appearance morphology of fibers and BFRP bars.

The concrete mixture proportion was designed according to JGJ 55-2011 (30). The mix proportion of benchmark concrete was mass (cement): mass (sand): mass (coarse aggregate) = 1: 1.46: 2.06, and the water-binder ratio was 0.48. The volume content of BF and CF was set to be 0.15 % and 0.2 % of the concrete volume, respectively, based on existing research (31, 32). The target temperature was set to be 20 °C, 100 °C, 200 °C, 300 °C, 400 °C and 500 °C. The main objective was to investigate the effect of the addition of BF and CF on the bond properties. Therefore, the change in fiber content was not considered.

Table 5 shows the specimen groups. This test consisted of 8 groups of specimens, with 3 specimens in each group. The specimen number was represented by a combination of letters and numbers, where HN and HF respectively represented NC and HFRC, and the number represented the target temperature. For example, HF300 referred to specimens of HFRC after exposure to 300 °C.

The specimen with a 150 mm side length, following the guidelines of ACI440.3R-12 (33). The center of the cube specimen was vertically penetrated by the BFRP bar. Based on ACI 440.1R-06 (34) and the findings of Eduarda et al. (35), the bond length was set to be 5d (60 mm), and the bond area was located in the upper portion of the concrete. The bond anchorage length was precisely controlled by sleeving a 90 mm ceramic sleeve in the unbonded area. The tape was used to attach the ceramic sleeve and the BFRP bar, and epoxy resin for 1 mm sealed both ends to avoid mortar intrusion during the casting process. The free end had a 20 mm length reserved for measuring the relative slip. The loading end was anchored by a galvanized steel tube with a 100 mm length to prevent its shear failure during the test. The specimens underwent curing at 23 ± 2 °C and 95 % humidity for 28 days after casting. Figure 3 shows the pull-out specimen.

The SX2-12-10 box resistance furnace was used as the heating device, with a 10 °C/min heating rate and a 1200 °C maximum temperature. A high temperature protection treatment was applied to the exposed BFRP bars by wrapping them with a combination of fire retardant coating and fireproof asbestos before heating. After reaching the target temperature (100 °C, 200 °C, 300 °C, 400 °C, 500 °C), the specimens were maintained at constant temperature for 5h in the resistance furnace and then cooled to 20 °C. The temperature-time curves in the furnace were measured by the K-type thermocouple inside the resistance furnace, as shown in Figure 4. Before the formal test, conducted a pre-test to measure the temperature-time curve inside the specimen using a K-type thermocouple embedded in the BFRP bar bond area, as shown in Figure 5. After being held at a constant temperature for 5 hours, the internal temperature of each pull-out specimen reached the target temperature basically. At target temperatures of 100 °C, 200 °C, 300 °C, 400 °C, and 500 °C, respectively, HN specimens had an average heating rate of 0.36 °C/min, 0.62 °C/min, 0.95 °C/min, 1.35 °C/min, and 1.82 °C/min, the average heating rates of HF specimens were 0.35 °C/min, 0.60 °C/min, 0.94 °C/min, 1.39 °C/min, and 2.12 °C/min. The internal heating rate of the specimens increased as the target temperature increased, but it was lower than that of the high-temperature furnace. Furthermore, no concrete spalling occurred.

The central pull-out test was applied to the pull-out specimens in accordance with GB/T 30022-2013 (36). Before the test, the specimens should be placed in the reaction force test device which was shown in Figure 6(a). The device was primarily composed of two thick backing plates, a center retainer bar, and four long retainer bars. Adjusted the nuts of the four long retainer bars in the reaction force test device to make the bottom surface of the pull-out specimen tightly adhere to the steel plate to avoid eccentric loading. After adjusting the pull-out specimen, using the upper and lower fixtures of the WDW-300 KN electro-hydraulic servo testing machine to clamp the center retainer bar of the reaction force test device and the steel tube of the pull-out specimen respectively, as shown in Figure 6(b). A linear variable displacement transducer (LVDT) was placed between the free end and the end of the center retainer bar of the reaction force test device to measure the slip value. Afterward, it was loaded at a rate of 1 mm/min, which was controlled by displacement.

The average bond stress τ for each group of specimens was calculated according to ASTM D7913/D7913M-14 (37). The formula is shown in Equation [1]:

Where τ was the average bond stress between the BFRP bar and concrete, MPa; P was the ultimate failure load of the specimen, N; d was the diameter of the BFRP bar, mm; l _a was the bond anchorage length of the BFRP bar, mm.

The TESCAN MIRA LMS and ZEISS Sigma 300 scanning electron microscopes were used to observe the internal microstructure of concrete. The SEM samples were taken from the concrete in the bonding interface.

Figure 7 shows the appearance morphology changes of BFRP bars with the increase in temperature. Significant changes in surface color were caused by temperature increase. Before 200 °C, the surface color remained almost unchanged, and only a few sand particles fell off. The surface color turned to carbon black after 300 °C, and all the sand particles came off. The matrix resin of the BFRP bars began to decompose after reaching the melting point at 400 °C, and a golden brown color appeared. The fibers of the bars became loose after 500 °C, and the matrix resin completely decomposed and lost its viscosity. The severe decomposition of the matrix resin after 500 °C eliminated the tensile strength, which prevented the subsequent pull-out test from being conducted. Hence, the temperature range of the tensile test in this research was from 20 °C to 400 °C.

Table 6 shows the tensile test results. The BFRP bar had glass transition temperature (T_g) and thermal decomposition temperature (T_d) in a high-temperature environment (38). With rising temperatures, the tensile strength dropped. When exposed to 100 °C, the tensile strength was 1.6% lower than when it was at 20 °C. The tensile strength decreased by 6.3 % after exposure to 200 °C, where the BFRP bar matrix resin reached the T_g. The tensile strength partially recovered after the matrix resin solidified upon cooling to 20 °C. The tensile strength dropped by 41.8 % at 300 °C and reached only 13.2 % of its initial value at 400 °C. This was due to the resin matrix reaching the T_d at 300 °C and undergoing irreversible oxidation and decomposition after being heated at a constant temperature for 5h. The tensile strengths were not significantly influenced by temperatures below 300 °C compared to it at 20 °C. However, the tensile strength declined rapidly at 300 °C. The low resistance of FRP bars to high temperatures was mainly attributed to T_g and T_d of the resin. Only the specimens exposed to temperatures of 20 °C, 100 °C, 200 °C, and 300 °C conducted pull-out tests, as the tensile strength of the bar was severely deteriorated at 400 °C.

Figures 8(a) and 8(b) show the SEM images of HF specimens at 20 °C. Figure 8(a) shows that the concrete mainly contained C-S-H gel, acicular ettringite (AFT), and hexagonal Ca(OH)₂ (CH). The microstructure was dense, and the pores were less. The aggregate and the solidified cementitious material were bonded together to form a uniform and dense continuous phase, which gave the cement paste a strong bond performance. Figure 8(b) shows that the surface of BF was covered by dense hydration products (C-S-H gel or CH crystal group). This indicated that there were good bonds in fiber and could be well embedded into the concrete matrix. This strong bond could enhance the force transfer of fiber, reduce internal damage, and then improve the bond properties. Figures 8(c) and 8(d) show that there were many voids in the HN specimen, and there was no substance inside the voids. However, the voids in the HF specimen were filled with a large amount of hydration products, which divided the large voids into small ones. The hydration products at the bonding interface in the HF specimen were significantly more than those in the HN specimen.

Figures 8(e) and 8(f) show the SEM images of HF and HN specimens at 200 °C, respectively. A lot of capillary water, free water, and bound water evaporated from the concrete at 200 °C, creating pores in the previously uniform and dense continuous medium. BF filled the pores and was slightly worn but not broken in HF, as shown in Figure 8(f). The gaps and micro-cracks in the concrete could be bridged. The secondary hydration reaction on the surface of BF produced many C-S-H gels and fine spherical particles that adhered to the fiber surface, which enhanced the mechanical interaction of fiber, and restrained the expansion of microcracks. The high thermal conductivity of BF could effectively transfer temperature and reduce the temperature gradient inside and outside the concrete.

Figure 8(g) ~ (h) show the SEM images of HF specimens after exposure to 300 °C. At 300 °C, the CF melted, and pores formed inside the matrix. Although the pores caused porosity to increase, they also provided a release channel for pore pressure and uneven thermal expansion stress. The decomposition of CH reduced the bond between cementitious materials and aggregates, resulting in a loose and porous internal structure of concrete. The surface of BF became smooth and had tiny cracks, indicating that the interface bonding force was weakened and some debonding occurred between fiber and matrix. Fiber fracture was the primary failure mode of BF, indicating that BF and matrix bear stress together before fracture. The cracks between BF and matrix further extended and connected with the pores left by the melting of CF to form the penetrating cracks.

Figure 9 illustrates the failure modes, which were classified as splitting failure, pull-out failure, and pull-out splitting failure.

In the case of splitting failure, a slight slip occurred on the loading end at the onset of the test loading, whereas the free end remained stable. As the load increased, the slip at the loading end increased, and the free end also began to slip. Splitting failure of the concrete occurred when the tangential tensile stress induced by the radial force exceeded the ultimate tensile strength of the concrete, resulting in an apparent brittle failure.

Pull-out failure specimens maintained good integrity, as no obvious cracks on the front and side. The bars were gradually pulled out with the increase in load, due to the shear strength of bars being much smaller than the tensile strength of concrete. When the free end was pulled out to a distance of roughly 20 mm, the loading was stopped. All specimens below 300 °C of HN and HF exhibited pull-out failures.

After 300 °C, the HN specimens experienced splitting failure, and the concrete was split into two pieces, whereas the HF specimens experienced pull-out splitting failure. The bars were gradually pulled out as the load increased, and obvious cracks emerged on the specimen surface in pull-out splitting failure, while integrity was preserved.

Figure 12 shows the bond-slip curves. The change rule of the HN and HF specimens was consistent. The bond stiffness and bond strength of each specimen gradually decreased, and the peak slip increased, with the increase in temperature. Four stages could be divided into the curves: the micro slip stage, the slip stage, the pull-out stage, and the residual stress stage. However, there were only the first three stages on the curve of the splitting failure specimen (HF300). In the beginning, the curve was in the micro slip stage. There were micro slip occurrences on the loading end, and the bond stress increased linearly. The chemical adhesion mainly provided the bond stress. The curve entered the slip stage with the loading continued, where the relative slip began to appear, and the bond stress rose nonlinearly. The chemical adhesive force disappeared, and the mechanical interaction and friction force between the transverse rib of the BFRP bar and the concrete interface mainly provided the bonding force. After the bond stress reached peak value, the curve entered the pull-out stage. Due to the further wear of the ribs of the bars, the concrete between the ribs being cut into pieces, the mechanical interaction weakened sharply, resulting in a gradual decrease in the bond stress. A nonlinear decrease was shown in the bond stress in the residual stress stage. Due to the slow pull-out of BFRP bars, the free end reserved outside entered into the concrete, and the mechanical interaction and friction force between the transverse ribs on the free end and the concrete interface was generated, which made the specimen regain some bonding force.