Combined Effects of Steel and Glass Fibres on the Fracture Performance of Recycled Rubber Concrete

Abstract

As an environmentally friendly construction material, recycled rubber concrete (RRC) is commonly used as a road material owing to its excellent flexural strength and crack resistance. Previous studies have shown that the addition of fibres is an effective method for improving the crack resistance of concrete. The purpose of this study is to investigate the fracture performance of RRC reinforced with steel fibres (SFs) and glass fibres (GFs). A total of 28 RRC mixtures were prepared. The results of the fracture test showed that the addition of SFs and GFs significantly enhanced the RRC fracture performance. The maximum increases or decreases in flexural strength, brittleness coefficient, fracture energy, initial fracture toughness, and unstable fracture toughness were 64.9, −34.6, 775.6, 92.0, and 118.4%, respectively. The ideal GF content is usually in the range of 0.4–0.6% and decreases with increasing SF content. In addition, scanning electron microscope (SEM) tests were conducted to explore the mechanism of the effect of hybrid fibres on RRC at a microscopic level. The results show that SFs were always pulled out, while GFs were pulled apart at the initial defects. At the same time, excessive GFs caused more initial defects. These results are expected to provide theoretical direction and experimental support for the practical application of hybrid fibre-reinforced recycled rubber concrete (HFRRRC).

1. Introduction

The widespread use of traditional concrete has led to excessive exploitation of natural aggregates, which exacerbates the depletion of natural resources and environmental damage. Therefore, many researchers propose using recycled resources and abundant sea sand resources as alternatives [1,2]. Meanwhile, with the development of urbanisation, the number of means of transportation is growing rapidly. This not only requires high-quality road materials, but also produces more waste rubber from scrapped car tires. Waste rubber cannot be degraded naturally. Moreover, waste tires are difficult to dissolve at high temperatures; therefore, it is very difficult for waste tires to be recycled into new rubber products. The most common disposal options for waste rubber tires are direct burning and burial, both of which cause irreversible pollution of the natural environment. To overcome the negative impact of waste rubber on the environment, there is an urgent need to develop a harmless method for consuming large quantities of waste rubber [3]. The use of recycled rubber (RR) in the concrete industry is an attempt to move in this direction. The use of RR as a substitute aggregate not only reduces the environmental impact of waste rubber by consuming it in large quantities, but also significantly reduces the consumption of natural aggregates such as sand and gravel [4,5,6]. The use of RR in the concrete industry has been attempted for a long time. However, it has been found that some of the key mechanical properties of concrete, including workability, compressive strength, tensile strength, stiffness, and modulus of elasticity, are significantly reduced when rubber is added to concrete [7,8,9]. Because road materials require a lower modulus of elasticity to improve the comfort of vehicle occupants, the decline in these properties makes recycled rubber concrete (RRC) an ideal road material [10,11]. Furthermore, RRC has a higher energy dissipation capacity than ordinary concrete, which makes it more resistant to vehicle impact loads as a road material [12,13].

Strength, stiffness, and stability are three key factors to consider when designing a structure [14,15]. While it is difficult to improve mechanical properties at the structural level, improving them at the material level is an effective approach. In recent years, more and more new types of high-performance concrete have been studied [16,17,18]. Concrete pavements are typically damaged by cracking owing to excessive bending and tensile stresses, as concrete itself is a material with good compressive properties and poor fracture properties. Therefore, numerous researchers have explored methods to improve the mechanics properties of RRC [19,20,21]. One effective method to improve the fracture properties of concrete is to add fibres within the concrete [22,23,24,25]. From a comprehensive perspective, the incorporation of fibres can significantly enhance the fracture properties of concrete; however, the enhancement effect of different fibres varies. Overall, the improvement in the fracture performance of concrete by fibres mainly depends on the following three aspects: (1) the type and characteristics of fibres, (2) the direction and distribution of fibres, and (3) the bond between the fibres and the cement matrix. Regarding the types and characteristics of fibres, the selection of fibres with a higher tensile strength and elasticity modulus is beneficial for enhancing the fracture properties of concrete [26,27]. Finer and softer fibres mainly limit the development of microcracks into macroscopic cracks, whereas larger and stiffer fibres mainly limit the development of macroscopic cracks and reduce the crack width [28,29]. Therefore, the addition of hybrid fibres is more effective in improving the fracture properties of concrete. With respect to the direction and distribution of fibres, the fracture resistance of each part within the concrete was enhanced when the fibre material was uniformly distributed within the concrete, thus improving the overall fracture resistance of the concrete. When the direction of the fibre was parallel to the direction of the main tensile stress in the concrete, the fibres effectively improved the crack resistance of the concrete. The direction and distribution of the fibre material within the concrete are mainly related to the workability of the fresh concrete, casting process, size of the specimen, and wall effect of the formwork [30]. Wang et al. studied the strengthening effect of four types of fibres on the fracture performance of rubber concrete. Compared with the control group, all fibres significantly enhanced fracture energy and crack propagation after cracking. Among them, steel fibres (SFs) are the best, increasing the fracture energy by about 50 times. The pull-out resistance of different fibres is not affected by the addition of rubber [31]. Gültekin et al. studied the fracture energy, compressive strength, and flexural strength of self compacting concrete, including glass fibres (GFs) or basalt fibres. Although the addition of fibres usually reduces compressive strength, it significantly improves flexural strength and fracture energy. The flexural strength and fracture energy of glass fibre reinforced concrete have been increased by 58.6% and 55.1%, respectively [32]. Muhyaddin studied the hybridisation of GF, micro steel fibre (MSF) and long hooked steel fibre (HSF) to improve the mechanical and fracture properties of ultra-high performance concrete (UHPC). The results show that the performance order is MSF+HSF, MSF+GF, and HSF+GF. The use of a single GF does not improve ductility and may even exhibit brittle failure [33]. In addition, excessive fibre addition not only tends to cause fibre agglomeration, but also weakens the bond interface between the fibres and concrete [34]. Therefore, although the addition of fibres enhances the fracture properties of concrete, the type, characteristics, and content of fibres need to be carefully selected to maximise their role.

At present, there is limited research on the fracture performance of SF and GF reinforced concrete, especially on RRC. Compared to ordinary concrete, RRC often has more internal defects, and the mechanism of the fibre-combined effect in RRC needs to be explored. This is significant for improving the crack resistance performance of rubber concrete pavement, thereby increasing the service life of the pavement. Therefore, this study investigated the fracture performance of RRC containing SFs and GFs, and analysed the influence mechanism of SF and GF on the fracture performance of RRC through microscopic experiments. The results of this research provide important guidelines for producing and using RRC.

2. Experimental Setup

2.1. Raw Materials

In this study, the strength grade and specific gravity of ordinary Portland cement were 42.5R and 3.11, respectively. The properties of cement are shown in

Table 1. The properties of cement.

Contents	CaO	SiO2	Fe2O3	Al2O3	SO3	MgO	Fineness	Loss on Ignition
Composition (%)	63–67	19–23	4–6	3–7	1.9	1	1.1	1.7

. The experimental water was tapping water from Guangdong University of Technology, with a specific gravity of 1.00. River sand with a maximum particle size of 5 mm and 20 mesh waste rubber were used as fine aggregates.

Table 2. The basic performance of fine and coarse aggregates.

Aggregate Type	Particle Size (mm)	Apparent Density (kg/m3)	Bulk Density (kg/m3)	Water Absorption (%)
Sand	<5	2636	1543	0.5
Recycled rubber	<2.5	750	-	-
Coarse aggregate	5–16	2641	1344	2.1

shows the river sand parameters obtained according to GB/T14684-2011 [35]. The parameters of the RR provided by the recycled rubber production company are shown in Table 2. The coarse aggregate was made of granite crushed stone with a particle size of 5–16 mm. Table 2 shows the coarse aggregates parameters obtained according to GB/T14685-2011 [36]. The gradation distributions of the sand, RR, and coarse aggregates obtained through the screening test are shown in Figure 1. A straight copper-plated SF with a length of 12 mm was used according to the parameters provided by the supplier. Two lengths of GFs were used in this study, which were 6 and 12 mm. The physical parameters of the fibres are listed in

Table 3. Properties of SFs and GFs.

Fibre Type	Length (mm)	Specific Gravity	Equivalent Diameter (μm)	Tensile Strength (MPa)	Elastic Modulus (GPa)
Steel fibres (SFs)	12	7.8	200	3000	200
Glass fibres (GFs)	6/12	2.68	14	1700	72

. Finally, the superplasticiser (SP) was used to enhance the workability of RRC. The solid content and specific gravity of the SP were 9% and 1.02, respectively.

2.2. Concrete Mixtures

To explore the anti-cracking and toughening effects of SF and GF on the fracture performance of RRC, 28 mix proportions were designed in this study, as shown in

Table 4. Test mixture (kg/m³).

Mix Number	Cement	Water	GFs	SFs	Coarse Aggregate	Sand	Recycled Rubber	SP
S0G0	554.10	245.30	0.00	0.00	966.30	531.20	17.10	2.80
S0.4G0	551.90	244.40	0.00	31.20	962.40	529.10	17.00	2.80
S0.8G0	549.70	243.40	0.00	62.40	958.50	527.00	16.90	2.80
S1.2G0	547.50	242.40	0.00	93.60	954.70	524.80	16.90	2.70
S0G0.2L6	553.00	244.90	5.40	0.00	964.30	530.20	17.00	2.80
S0.4G0.2L6	550.80	243.90	5.40	31.20	960.50	528.00	17.00	2.80
S0.8G0.2L6	548.60	242.90	5.40	62.40	956.60	525.90	16.90	2.70
S1.2G0.2L6	546.40	241.90	5.40	93.60	952.70	523.80	16.80	2.70
S0G0.4L6	551.90	244.40	10.70	0.00	962.40	529.10	17.00	2.80
S0.4G0.4L6	549.70	243.40	10.70	31.20	958.50	527.00	16.90	2.80
S0.8G0.4L6	547.50	242.40	10.70	62.40	954.70	524.80	16.90	2.70
S1.2G0.4L6	545.20	241.40	10.70	93.60	950.80	522.70	16.80	2.70
S0G0.6L6	550.80	243.90	16.10	0.00	960.50	528.00	17.00	2.80
S0.4G0.6L6	548.60	242.90	16.10	31.20	956.60	525.90	16.90	2.70
S0.8G0.6L6	546.40	241.90	16.10	62.40	952.70	523.80	16.80	2.70
S1.2G0.6L6	544.10	240.90	16.10	93.60	948.90	521.70	16.70	2.70
S0G0.2L12	553.00	244.90	5.40	0.00	964.30	530.20	17.00	2.80
S0.4G0.2L12	550.80	243.90	5.40	31.20	960.50	528.00	17.00	2.80
S0.8G0.2L12	548.60	242.90	5.40	62.40	956.60	525.90	16.90	2.70
S1.2G0.2L12	546.40	241.90	5.40	93.60	952.70	523.80	16.80	2.70
S0G0.4L12	551.90	244.40	10.70	0.00	962.40	529.10	17.00	2.80
S0.4G0.4L12	549.70	243.40	10.70	31.20	958.50	527.00	16.90	2.80
S0.8G0.4L12	547.50	242.40	10.70	62.40	954.70	524.80	16.90	2.70
S1.2G0.4L12	545.20	241.40	10.70	93.60	950.80	522.70	16.80	2.70
S0G0.6L12	550.80	243.90	16.10	0.00	960.50	528.00	17.00	2.80
S0.4G0.6L12	548.60	242.90	16.10	31.20	956.60	525.90	16.90	2.70
S0.8G0.6L12	546.40	241.90	16.10	62.40	952.70	523.80	16.80	2.70
S1.2G0.6L12	544.10	240.90	16.10	93.60	948.90	521.70	16.70	2.70

. Fibres were added according to the volume percentage of concrete. The content of SFs was designed as 0, 0.4%, 0.8% and 1.2%, and the content of GFs was designed as 0, 0.2%, 0.4% and 0.6%. In this study, rubber content was set to 10%. RR was added according to the volume percentage of the fine aggregate. The SP was added according to the mass percentage of the cement, which was 0.5% in this study. The water-cement ratio was 0.4, and water consumption was adjusted based on the moisture content and water absorption of the aggregates. Each mix proportion is marked in the SxGyLz format. Sx indicates that the SF content is x%. Similarly, Gy indicates that the GF content is y%. Lz is the length of the GFs, which is z mm. If the mix number is without Lz, this implies that no GFs have been added.

2.3. Preparation of Test Specimens

Beams with lengths (l), widths (t), and heights (h) of 400, 100, and 100 mm, respectively, were used as fracture specimens [37]. The height (a₀) and width of the pre-cracks were 30 and 1 mm, respectively. A schematic diagram of the size of the specimen is shown in Figure 2. Three beams with pre-cracks were prepared for each mix proportion as fracture specimens, for a total of 84. To obtain the compressive strength, three cylinders with diameters of 100 mm and heights of 200 mm were prepared for each mix proportion as axial compression specimens.

The preparation process of specimens is shown in Figure 3. To evenly distribute the GFs and SFs within the RRC, the preparation process of hybrid fibre-reinforced recycled rubber concrete (HFRRRC) is divided into eight steps. (1) First, cement, sand, and RR were added to a mixer for dry mixing. After starting the mixer, SFs and GFs were uniformly added, which lasted for 60 s. (2) After water and SP were fully mixed, around 70% of the mixture was poured to the mixer and operated for 60 s. (3) The remaining 30% mixture and coarse aggregates were poured into the mixer, and operated the mixer for 180 s to thoroughly mix it. (4) Fresh concrete was poured into the moulds and vibrated for 1 min. The vibration time should not be excessively long to prevent the SF from sinking to the bottom and the rubber particles from floating. (5) Aluminium sheets with a thickness of 1 mm were vertically inserted into the fresh concrete to form a pre-crack. (6) Fresh concrete was smoothened along the surface of the test mould. To prevent the evaporation of internal water and to affect the hydration rate of the cement, the specimens were covered tightly with plastic and wet linen cloths. (7) After the initial setting of the fresh concrete, the aluminium sheets were carefully pulled out with iron tongs. (8) After curing for 24 h, the specimens were demoulded and cured in water for 28 days. Finally, the specimens were removed and wiped dry before mechanical testing.

2.4. Testing Schemes

Axial deformation in the axial compression test was measured using two linear variable differential transformers (LVDTs). The LVDTs were mounted in the middle of the specimen at a distance of 80 mm. The loading procedure for the compression experiments was displacement controlled, which was 0.18 mm/min. The compressive strength of each specimen was determined according to the standard ASTM C39/C39M [38].

An electro-hydraulic servo static and dynamic universal testing machine with a maximum capacity of 500 kN was used for the fracture tests. The fracture test setup is illustrated in Figure 2. The support span, s, was 300 mm. Two LVDTs were symmetrically placed in the mid span of the test specimens to measure the deflection. Four strain gauges were used to test the crack development in each specimen. Crack mouth opening displacement (CMOD) was measured using an extensometer. To fix the extensometer, two aluminium sheets with a thickness (h₀) of 1 mm were symmetrically bonded to both sides of the pre-crack using an adhesive. To eliminate load eccentricity, the top and bottom of the specimens were levelled with high-strength gypsum before the test. A load was applied with a displacement control of 0.2 mm/min. During the test, strain gauges, LVDTs, extensometer, and load data were recorded at a frequency of 1 Hz.

The analysis indices of the fracture test included the flexural strength f_t, fracture energy G_f, initial fracture toughness Kⁱⁿⁱ, and unstable fracture toughness K^un. The flexural strength of the specimens was calculated as follows:

$$f_{\mathrm{t}}=\frac{3P_{\max }s}{2t{\left(h-a_0\right)}^2}$$

(1)

where P_max is the maximum load during the testing. The equation for calculating the fracture energy G_f is as follows:

$$G_{\mathrm{f}}=\frac{{\displaystyle \int \begin{array}{c}{\delta }_0\\ 0\end{array}P\mathrm{d}\delta +mg\frac{s}{l}}{\delta }_0}{t\left(h-a_0\right)}$$

(2)

where P is the load recorded during the test; δ₀ is the deflection required for calculating the fracture energy and is determined as 2 mm; m represents the mass of the specimen; g is the gravitational acceleration, i.e., 9.8 N/kg.

In this study, fracture toughness was calculated using the double-K fracture criterion proposed by Xu and Reinhardt [39]. According to this criterion, the initial fracture toughness and unstable fracture toughness are the key indicators for determining the fracture performance of concrete. When the stress intensity factor of concrete was less than the initial fracture toughness, cracks did not develop. When the stress intensity factor was greater than the initial fracture toughness and less than the unstable fracture toughness, cracks developed stably. Crack instability develops when the stress intensity factor exceeds the unstable fracture toughness. Therefore, the crack toughness can be calculated as follows:

$$K^{\mathrm{ini}}=\frac{3P_{\mathrm{ini}}s}{2h^{2}t}\sqrt{a_0}F\left(\frac{a_0}{h}\right)$$

(3)

$$F\left(\frac{a_0}{h}\right)=\frac{1.99-\frac{a_0}{h}\left(1-\frac{a_0}{h}\right)\left(2.15-\frac{3.93a_0}{h}+2.7{\left(\frac{a_0}{h}\right)}^2\right)}{\left(1+2\frac{a_0}{h}\right){\left(1-\frac{a_0}{h}\right)}^{\frac{3}{2}}}$$

(4)

where F(a₀/h) is calculated using Equation (4). P_ini denotes cracking load. Determining the initial load is key to processing the test data, and there are generally two methods. The first is the strain gauge method [40,41]. The strain gauges are pasted on both sides of the pre-crack end, and the retraction point of the strain value of the load–strain curve is considered as the cracking point. This method is simple, easy to implement, and widely used. The second method is the test curve method [37,42]. As stated in the standard and other studies, the load corresponding to the turning point of the rising section of the load–CMOD curve from a straight section to a curved section is the cracking load. This method can overcome some of the disadvantages of the strain gauge method (e.g., the appearance times of surface cracks and internal cracks are inconsistent). In this study, the second method is used to determine the cracking point.

To calculate the unstable fracture toughness, the initiation load P_ini and pre-crack height a₀ in Equations (3) and (4) were replaced with the maximum load P_max and critical crack length a_c, respectively. The critical crack length was calculated as follows:

$$a_{\mathrm{c}}=\frac{2}{\pi }\left(h+h_0\right)\arctan \sqrt{\frac{t{E}_{\mathrm{f}}}{32.6P_{\max }}{\mathrm{CMOD}}_{\mathrm{c}}-0.1135}-h_0$$

(5)

where h₀ is the thickness of the aluminium sheet; CMOD_c is the CMOD corresponding to the maximum load P_max; E_f is the rupture modulus, which was calculated as:

$$E_{\mathrm{f}}=\frac{1}{t{c}_{\mathrm{i}}}\left[3.70+32.6{\tan }^2(\frac{\pi }{2}\frac{a_0+h_0}{h+h_0})\right]$$

(6)

where c_i denotes the initial compliance. This was calculated as the ratio of P to CMOD at any point on the straight segment of the P–CMOD curve. P_ini and its corresponding CMOD were selected uniformly for this study.

Finally, samples were obtained from the fracture surface of the specimen for the scanning electron microscope (SEM) test.

3. Result and Discussion

3.1. Failure Mode Analysis

The final morphologies of the fractured specimens are shown in Figure 4. For the specimens without SFs, as shown in the left column of Figure 4, the cracks mostly developed vertically upward after they appeared, and the crack widths were not large. When the GF content was 0.6%, as shown in Figure 4m,y, the development path of the cracks became more tortuous. This indicates that when SFs were not added, the macroscopic cracks were not effectively confined and always developed along the high-stress position. Further, a small crack width indicates that the deformation of the specimen was small when its failure and the deformation ability was poor. Specimens with different SF contents are shown in the rows in Figure 4. With the increase in SF content, the crack above pre-crack changes from vertical to tortuous. This indicates that the presence of SFs can enhance the crack resistance of RRC so that crack development avoids these enhanced areas. In addition, the crack width increased with increasing SF content. This indicates that the deformation during the RRC fracture failure was larger, and the deformation capacity was better with the increase in the SF content.

According to the failure mode analysis, the effects of the two fibres on the fracture performance of the RRC are as follows: When only the GFs were added, the cracks in the specimens were relatively vertical. This shows that GF had little effect on the restriction of macroscopic cracks, and the cracks appeared at the position of the greatest stress. The GFs that appeared in the macroscopic cracks were either pulled out or broken. Therefore, the GFs mainly acted before the appearance of macroscopic cracks. Whether it was a 6 mm or 12 mm GF, it had no significant effect on the final failure mode of the fracture specimen. In contrast, SFs had a greater influence on the final failure mode. The main reason for this was that the bridging effect of SFs can continue after the appearance of macroscopic cracks. Steel fibres reduce the maximum stress at the crack tip by altering the strain field at the crack tip. Studies have shown that as long as the crack width does not exceed 1/4 of the length of the SF, SF-reinforced concrete still has a load-bearing capacity [43]. Therefore, SFs can improve the deformation ability and toughness of RRC.

Figure 5 shows the crack development in fracture tests. At the beginning, as shown in Figure 5a, the fibres are uniformly distributed in the RRC and there is a number of initial defects. As the load increases, microcracks first appear at initial defects and the interface between the aggregates and matrix, as shown in Figure 5b. At this point, GFs can suppress the development of microcracks. When the load continues to increase to the cracking load, microcracks connect to form macroscopic cracks. Cracks usually avoid fibres and appear in weak areas of RRC. Therefore, the presence of SFs makes macroscopic cracks more tortuous, as shown in Figure 5c. Finally, as shown in Figure 5d, the load continued to increase until it approached the ultimate load of the specimen. At this point, most GFs break, some of SFs are pulled out and fail. Only a portion of SFs play a bridging role. The development time from Figure 5c to 5d is relatively long. In other words, the SFs gradually improve the fracture toughness of RRC during the process of being pulled out.

3.2. Load–Deflection Curves

Figure 6 illustrates the load–deflection curves of different series. The load–deflection curves for each series were divided into ascending and descending sections. The influence of SFs and GFs on the load–deflection curves of each series was mainly reflected in the peak load and deformation capacity.

From the perspective of the peak load, when the SF content was less than 1.2%, the peak load of the curves increased with the increase in GF content. In this case, when the GF contents of 6 and 12 mm was 0.6%, the peak load was maximised. When the SF content was 1.2%, the peak load of the curves first increased and then decreased as the GF content increased. When the GF content of 6 and 12 mm was 0.4%, the peak load reached its maximum. Therefore, an appropriate amount of GF can increase the peak load. This was because the GFs had a bridging effect and improved the mechanical properties of concrete. However, excessive GFs reduce the workability of fresh concrete and increase the void ratio inside the concrete, which was harmful to the mechanical properties of concrete. When comparing the load–deflection curves for different SF contents, the peak load increases with the increase in SF content. This was because the SFs can play a bridging role after the concrete had cracked so that the concrete on both sides of the crack can continue to carry loads.

The inclusion of fibres can effectively improve the deformation ability of RRC. The deflection corresponding to the peak load was defined as the peak deflection. As shown in Figure 5a,b, when no SF was added, the peak deflection first increases and then decreases with the increase in GF content. The peak deflections of the specimens with GFs were greater than those of the S0G0 group. Both GF lengths had the highest peak deflection when the content was 0.2%. After the peak deflection, the load value decreased rapidly, indicating brittle failure, and the slope of the descending curve was similar to that of the S0G0 group. This indicates that an appropriate number of GFs can improve the deformation capacity of the RRC before the peak load, but there was not much improvement in the deformation capacity after the peak load. This was due to the fact that GFs can limit the formation and development of microcracks in the ascending section of the load–deflection curve. However, excessive GFs can increase the initial defects of the RRC. During the descending section of the load–deflection curve, there were evident macroscopic cracks in the RRC, and the GFs broke or were pulled out to fail. Therefore, the deformation ability of the RRC after the peak load was not significantly improved when only GFs were added. As shown in Figure 5c–f, for the specimens with SFs, the descending section of the load–deflection curve can be divided into two sub-phases: a rapid descent phase and a stabilisation phase. As the SF content increases, the transition zone between these two subphases becomes smoother. The load corresponding to the transition zone was greater. This indicates that SFs can effectively limit the speed of the development of macroscopic cracks. The transition from the rapid descent phase to the stabilisation phase at higher load levels increased the energy consumed and improved the toughness of the RRC. By comparing the 6 mm GFs series with the 12 mm GFs series, it can be observed that the existence of SFs better improves the deformation capacity of the 12 mm GFs series after peak loading.

3.3. Index Analysis

Table 5. Test results.

Mix Number	Compressive Strength		Flexural Strength		Brittleness Coefficient		Fracture Energy		Initial Fracture Toughness		Unstable Fracture Toughness
	RV * (MPa)	PC * (%)	RV (MPa)	PC (%)	RV	PC (%)	RV(J/m2)	PC(%)	RV(MPa·m1/2)	PC(%)	RV(MPa·m1/2)	PC(%)
S0G0	34.36	-	5.81	-	5.91	-	228	-	0.60	-	1.24	-
S0.4G0	36.93	7.5	6.13	5.5	6.03	1.92	405	77.3	0.79	30.2	1.74	40.3
S0.8G0	35.06	2.0	6.87	18.3	5.10	−13.7	917	301.7	0.88	46.1	2.29	84.8
S1.2G0	38.09	10.9	8.03	38.1	4.74	−19.8	1531	570.8	0.93	54.2	2.32	87.8
S0G0.2L6	34.24	−0.3	6.53	12.3	5.25	−11.3	297	30.3	0.72	19.0	1.79	44.5
S0.4G0.2L6	36.57	6.4	7.18	23.6	5.09	−13.9	480	110.4	0.91	50.0	1.94	57.1
S0.8G0.2L6	38.57	12.3	8.26	42.2	4.67	−21.0	1063	365.7	0.90	49.4	2.44	97.6
S1.2G0.2L6	42.55	23.8	8.78	51.0	4.85	−18.0	1755	669.2	0.96	58.3	2.63	112.8
S0G0.4L6	34.77	1.2	7.27	25.1	4.78	−19.1	333	46.0	0.80	32.2	1.71	38.5
S0.4G0.4L6	34.77	1.2	7.40	27.4	4.70	−20.6	736	222.4	0.91	51.4	1.88	51.8
S0.8G0.4L6	40.05	16.6	8.48	45.9	4.72	−20.1	1175	415.0	0.89	47.7	2.31	87.0
S1.2G0.4L6	36.87	7.3	8.84	52.0	4.17	−29.4	1737	661.2	0.98	61.6	2.70	118.4
S0G0.6L6	35.26	2.6	7.31	25.7	4.83	−18.4	351	53.6	0.90	49.5	1.95	57.8
S0.4G0.6L6	39.53	15.0	7.59	30.5	5.21	−11.9	903	295.6	1.08	79.5	2.07	67.5
S0.8G0.6L6	37.58	9.4	8.75	50.5	4.30	−27.3	1283	462.1	1.09	80.3	2.39	93.4
S1.2G0.6L6	33.73	−1.8	8.73	50.1	3.87	−34.6	1798	687.9	1.07	77.8	2.69	117.7
S0G0.2L12	35.68	3.8	6.09	4.8	5.86	−0.9	298	30.6	0.91	51.3	1.31	5.6
S0.4G0.2L12	37.95	10.4	7.37	26.7	5.15	−12.8	755	230.9	1.01	67.4	1.63	31.6
S0.8G0.2L12	39.81	15.9	7.99	37.5	4.98	−15.8	1331	483.3	0.94	56.3	2.25	82.2
S1.2G0.2L12	41.59	21.0	9.06	55.9	4.59	−22.4	1727	656.8	1.02	68.6	2.24	80.8
S0G0.4L12	36.36	5.8	6.71	15.4	5.42	−8.3	267	16.9	0.90	49.3	1.33	7.2
S0.4G0.4L12	38.49	12.0	7.88	35.5	4.89	−17.4	901	294.7	0.99	63.2	1.71	38.6
S0.8G0.4L12	38.99	13.5	8.07	38.8	4.83	−18.3	1403	514.6	0.95	56.7	2.26	82.3
S1.2G0.4L12	43.60	26.9	9.59	64.9	4.55	−23.1	1998	775.6	1.16	92.0	2.57	107.6
S0G0.6L12	37.18	8.2	6.83	17.4	5.45	−7.9	325	42.5	0.95	58.0	1.43	15.6
S0.4G0.6L12	39.90	16.1	8.16	40.4	4.89	−17.3	942	312.9	0.97	61.1	1.73	39.9
S0.8G0.6L12	38.02	10.7	9.01	55.0	4.22	−28.6	1357	494.7	1.09	80.1	2.05	65.5
S1.2G0.6L12	40.28	17.2	9.19	58.1	4.38	−25.9	1702	645.9	1.13	86.7	2.29	85.5

* RV: representative value; PC: percentage changes.

lists the values of each index. The values of each index in Table 5 represent the average values of the three specimens. If the following two situations occur, they were considered separately: (1) if both the maximum and minimum values differ from the median value by more than 15% of the median value, the median value was considered as the representative value; and (2) if only one of the maximum and minimum values differs from the median value by more than 15% of the median value, the value with a large difference was discarded, and the average of the other two values was considered as the representative value. In addition, to better reflect the impact of SFs and GFs on the indexes, Table 5 also lists the change rates of each index. The rates of change were relative to the S0G0 group. Positive values indicate an increase and negative values indicate a decrease. In addition, because we had previously studied the compressive strength and elastic modulus, this study focuses on the effects of SFs and GFs on the fracture index of RRC.

3.3.1. Compressive Strength

As shown in Table 5, when the SF contents were 0 and 0.4%, the compressive strength increased with the GF content. When the SF content was greater than 0.4%, the compressive strength first increased and then decreased with increasing GF content. Specifically, when only GFs were added at 0.6%, the 6 mm and 12 mm GFs resulted in the largest increase in the compressive strength, which was 2.6 and 8.2%, respectively. When only 1.2% SFs were added, the greatest gain in compressive strength was 10.9%. However, when the 12 mm GF content was 0.4% and SF content was 1.2%, the largest gain in compressive strength was 26.9%. Therefore, although the separate addition of the two fibres can improve the compressive strength of RRC, the combination of GFs and SFs resulted in the largest compressive strength.

3.3.2. Flexural Strength

The flexural strength of the fracture test was calculated using Equation (1) and is shown in Table 5 and Figure 7. When the SF content was less than 1.2%, the flexural strength of each series increased with the GF content; however, the growth rate decreased. In this case, the flexural strength of the RRC reached its maximum at a GF content of 0.6%. When the SF content was 1.2%, the flexural strength first increased and then decreased with an increase in the GF content. The flexural strength of the RRC reached a maximum at a GF content of 0.4%, which was similar to 0.3–0.6% in the literature [44]. Specifically, when only GFs were added at 0.6%, the 6 mm and 12 mm GFs resulted in the largest increase in the flexural strength, which was 25.7 and 17.4%, respectively. For the specimens with SFs, 6 mm and 12 mm GFs make the highest improvements in flexural strength, which were 52.0 and 64.9%, respectively, when the SF content was 1.2% and the GF content was 0.4%.

In addition, when only 1.2% SF was added, the greatest increase in flexural strength was 38.1%. For the specimen with only 12 mm GF, the growth rate of the flexural strength reached a maximum of 17.4% when the GF content was 0.6%. However, in the case of hybrid fibres, the growth rate of the flexural strength reached a maximum of 64.9% when the SF content was 1.2% and the 12 mm GF content was 0.4%. Therefore, the separate addition of SFs and GFs can improve the flexural strength of the RRC. The combination of GFs and SFs resulted in the largest flexural strength. The positive mixing effect was caused by the different properties of the SFs and GFs. At the initial stage, a larger quantity of GFs effectively reduces the stress concentration at the microcrack tip. Therefore, the stress can be redistributed and more matrix materials can be used, improving the strength of the RRC specimen. Due to the significant impact of GFs on the workability of RRC, it is not advisable to exceed 0.4% of GF content when the total volume fraction of hybrid fibres is large. Due to the high strength and elastic modulus of SFs, a skeleton was formed inside the concrete. Simultaneously, the SFs played a bridging role when a macroscopic crack appeared until the SFs were pulled out and failed.

3.3.3. Brittleness Coefficient

The brittleness coefficient, defined as the ratio of the compressive strength to the flexural strength, is an important index for RRC subjected to tensile failure. RRC with larger brittleness coefficients exhibited greater brittleness. The brittleness coefficients of the RRC are listed in Table 5 and Figure 8. For the specimens without adding SFs, when the GF content was 0.4%, 6 and 12 mm GFs make the maximum decrease in brittleness coefficient, which were −19.1 and −8.3%, respectively. For the specimens with SFs, when the 6 mm GF content and SF content was 0.6% and 1.2%, respectively, the maximum decrease in the brittleness coefficient was −34.6%. When the 12 mm GF content and SF content was 0.6% and 0.8%, respectively, the maximum decrease in brittleness coefficient was −28.6%. Overall, the addition of fibres can reduce the brittleness coefficient of the RRC. It can be noted that when 12 mm GFs and SFs were added together, the brittleness coefficient only slightly decreased with the increase in GF content, when the content of GF was larger than 0.4%. Therefore, the relationship between the brittleness coefficient and the GF content is quite complex. For 6 mm GF, the optimal content was 0.4–0.6%. For 12 mm GF, the optimal dosage was 0.4%.

3.3.4. Fracture Energy

As the most useful material parameter in the analysis of cracked concrete structures, the fracture energy was calculated using Equation (2) [45]. It is shown in Table 5 and Figure 9. The fracture energy reflects the amount of energy dissipated by the RRC fracture. When only GFs were added, the fracture energy increased slightly with increasing GF content. When only 0.6% GFs were added, the 6 and 12 mm GFs resulted in the highest improvements in the fracture energy, which were 53.6% and 42.5%, respectively. When only SFs were added, the fracture energy increased significantly with increasing SF content, with the greatest improvement being 570.8%. When the 6 mm GF content and SF content was 0.6% and 1.2%, respectively, the maximum increase in the fracture energy was 687.9%. When the 12 mm GF content and SF content was 0.4% and 1.2%, respectively, the maximum increase in fracture energy was 775.6%. Overall, the fracture energy increased with the increase in the content of two types of fibres, especially SFs. The reason was that even at lower fibre content, the softening part of the load–deflection curves of RRC containing SFs were longer [33]. The 12 mm GFs caused a greater increase in fracture energy than the 6 mm GFs. This was because the propagation of cracks typically avoids the area where the fibres were located, and longer GFs make it more difficult for cracks to avoid them.

In addition, when only 1.2% SFs were added, the greatest gain in fracture energy was 570.8%. For the specimen with only 12 mm GF, the growth rate of the fracture energy reached a maximum of 42.5% when the GF content was 0.6%. However, in the case of hybrid fibres, the growth rate of the fracture energy reached a maximum of 775.6% when the content of SF was 1.2% and the content of 12 mm GF was 0.4%. Therefore, although both SFs and GFs can increase the fracture energy of RRC, SFs play a major role. According to Figure 9, without adding SFs, the lines became gentler with increasing GF content. However, when SFs were added, the slope of the lines increased, indicating that the two fibres had a positive mixing effect. This was because brittle GFs fail quickly after the appearance of macroscopic cracks. GFs alone cannot effectively limit the crack development after peak loads. When SFs were added, SFs prevented the rapid development of macroscopic cracks. The specimens containing SFs had more tortuous cracks, indicating that more GFs and RRC played a role in increasing the energy consumed for crack development. The failure modes of the RRC confirmed this, as described in Section 3.1.

3.3.5. Initial Fracture Toughness

The initial fracture toughness results are presented in Table 5 and Figure 10. Equation (3) shows that the initial fracture toughness is proportional to the cracking load. Therefore, the effect of SFs and GFs on the cracking load was equal to their effect on the initial fracture toughness. In general, the initial fracture toughness of the RRC was increased by the inclusion of fibres. When only 1.2% SFs was added, the highest gain in the initial fracture toughness was 54.2%. When only 0.6% GFs were added, the 6 mm and 12 mm GFs resulted in the highest improvement in the initial fracture toughness, which was 49.5 and 58.0%, respectively. When the 12 mm GF content and SF content was 0.4% and 1.2%, respectively, the initial fracture toughness gained a maximum increase of 92.0%. Before the RRC cracked, the RRC and fibres were synergistically deformed by the adhesive force. Compared with RRC without fibres, the high elastic modulus of SFs allows RRC to withstand greater loads under the same deformation, thereby increasing the cracking load. The increase in initial fracture toughness was not significant with the increase in SF content. Literature [41] introduced 1% SF into concrete, which only increased the initial fracture toughness by 2.6%. Simultaneously, because of the higher quantity of GFs in the same content, the formation and development of microscopic cracks were limited by the GFs better. Therefore, GF seems to contribute more significantly to the initial fracture toughness than SF. Compared with single fibres, hybrid fibres were more effective in the enhancement of the initial fracture toughness of RRC.

3.3.6. Unstable Fracture Toughness

The unstable fracture toughness results are listed in Table 5 and shown in Figure 11. In general, the unstable fracture toughness of RRC increased with an increase in the SF content but only slightly increased through increasing the content of GF. When only 1.2% SFs were added, the greatest gain in the unstable fracture toughness was 87.8%. When only 0.6% GFs were added, both lengths of GFs resulted in the highest improvements in the unstable fracture toughness, which were 57.8% and 15.6%, respectively. When the 6 mm GF content and SF content was 0.4% and 1.2%, respectively, the maximum increase in the unstable fracture toughness was 118.4%. Compared to the initial fracture toughness, SFs had a more significant improvement in unstable fracture toughness. This conclusion can also be drawn from literature [41]. For the RRC specimens reinforced with the two types of fibres, the increase in the unstable fracture toughness was not significant with the increase in the GF content. This was because unstable fracture toughness is the critical point at which cracks develop from stability to instability. The bridging effect of the GFs mainly occurs in the microcrack stage. After the appearance of macroscopic cracks, most of GFs quickly fail. Therefore, the amount of GF content has little effect on the unstable fracture toughness. At this time, the SFs play a major bridging role, slowing down the development of macroscopic cracks, and thereby improving the unstable fracture toughness of the RRC.

4. Mechanism Analysis

Through the analysis of the failure modes and various indices, it can be inferred that the incorporation of SFs is useful for improving the fracture performance of RRC. Before the cracking load, there was a strong bond between the SFs and RRC. The high elastic modulus and tensile strength of SFs can improve the bearing capacity of RRC, thereby improving the peak and cracking loads. As illustrated in Figure 12a, after the peak load, because the tensile strength of the SFs was greater than the bonding strength with the RRC, the SFs were always pulled out rather than broken. Previous literature has indicated that, until the tensile deformation of SF-reinforced concrete reaches 1/4 of the fibre length, the tensile stress decreases to 0 [37]. Therefore, the SFs located at the tip of the crack can still play a bridging role so that the concrete on both sides of the crack can continue to bear the load. As the SFs were pulled out, the neutral axis and the tip of the crack slowly developed upward. The energy consumed by this process is characterised by an increase in fracture energy.

The influence of the GFs on the fracture performance of RRC mainly occurred before the cracking load. As RRC is a heterogeneous material, there may be certain initial internal defects. When RRC is subjected to external loads, microcracks usually develop from the initial defect and eventually interconnect into macroscopic cracks. In particular, when excessive GFs were added, the poor workability increased the initial defects of the RRC, as shown in Figure 12b. Therefore, the GF content should not be excessively high. When an appropriate GF content is used, the fracture performance of the RRC can be effectively improved, and the workability of the RRC is less affected. Because there were significantly more GFs than SFs at a given fibre content, the formation and growth of microcracks may be efficiently suppressed. Because GFs are brittle materials and microcracks are present on the glass surface, their strength is generally lower than the theoretical strength. Therefore, most of the GFs located in the macroscopic cracks broke after the cracking load. Therefore, the GFs had no significant bridging effect on the descending segment after the peak load.

The fracture properties of the RRC were maximised by the combined addition of SFs and GFs. This indicates that the SFs and GFs had a positive mixing effect on the fracture properties of RRC. In the preliminary loading period, the GFs was useful for inhibiting the formation and growth of microcracks. As the load continued to increase, the microcracks interconnected to form macroscopic cracks. At this point, the bridging effect of the SFs slowed the development of cracks. The two parts bridged by the SFs could continue to bear the load, allowing more RRC and GFs to work. This was verified using fracture energies. As shown in Figure 9, after the addition of SFs, the fracture energy increased more significantly with an increase in GF content. However, the variation in the experimental results is complex because the fibre content has an important impact on the workability of fresh RRC. Further studies are required to elucidate the mechanisms underlying HFRRRC.

5. Conclusions

Fracture experiments were conducted on 28 different concrete mixes in order to examine the effects of SFs and GFs on the fracture characteristics of RRC. Analysis was performed on the failure mechanism, load–deflection curves, flexural strength, fracture energy, brittleness coefficient, initial fracture toughness, and unstable fracture toughness. The test findings supported the following deductions:

(1)

The fractured specimens without SFs were damaged by vertical cracks and their crack widths were smaller. The two GF lengths had no significant effect on the final morphology of the cracks. With the addition of the SFs, the cracks became more curved as the crack width increased. The SFs primarily played a bridging role after RRC cracking. At this point, the GFs located in the cracks had failed.

(2)

In general, the peak load increased with increasing dosages of both fibres. However, with 1.2% SFs, the excessive addition of GFs decreased the peak load. For specimens with only GFs, the peak deflection increased and then decreased, indicating brittle damage. The addition of SFs significantly improved the post-peak deformation capacity and smoothed the descending section of the load–deflection curves. The specimens of the 12 mm GFs series exhibited a greater enhancement of the deformation capacity when SFs were added.

(3)

When only SFs were added, the maximum increases or decreases in flexural strength, brittleness coefficient, fracture energy, initial fracture toughness, and unstable fracture toughness with increasing SF content were 38.1, −19.8, 570.8, 54.2, and 87.8%, respectively. When only GFs were added, the maximum increases or decreases in the above-mentioned indices with increasing GF content were 25.7, −19.1, 53.6, 58.0, and 57.8%, respectively. When SFs and GFs were added together, the maximum increases and decreases in the above-mentioned indices with increasing GF content were 64.9, −34.6, 775.6, 92.0, and 118.4%, respectively. Thus, although the addition of SFs or GFs alone also improved the mechanical performance indices of the RRC, the combination of SFs and GFs produced the most significant improvements.

(4)

In this study, the optimum GF content was found to be 0.4–0.6%. It decreases with increasing SF content ensuring the workability of the RRC.

(5)

The mechanism for the positive mixing effect of SFs and GFs was that they acted at different stages during the fracture test. A large number of GFs restricted the development of microcracks in the early stages of the test. After the RRC cracked, which caused the GFs to fail, the SFs still acted as bridges.

(6)

To prevent the cracking of rubber concrete pavement caused by external factors, adding SFs and GFs together is an effective solution. Before practical application, more performance indicators need to be quantitatively analysed. The durability and impact resistance of HFRRRC still needs to be studied.