Inﬂuence of Thickness on Water Absorption and Tensile Strength of BFRP Laminates in Water or Alkaline Solution and a Thickness-Dependent Accelerated Ageing Method for BFRP Laminates

: This paper ﬁrst presented an experimental study on water absorption and tensile properties of basalt ﬁber-reinforced polymer (BFRP) laminates with di ﬀ erent specimen thicknesses (i.e., 1, 2, and 4 mm) subjected to 60 ◦ C deionized water or alkaline solution for an ageing time up to 180 days. The degradation mechanism of BFRP laminates in solution immersion was also explored combined with micro-morphology analysis by scanning electronic microscopy (SEM). The test results indicated that the water absorption and tensile properties of BFRP laminates were dramatically inﬂuenced by specimen thickness. When the BFRP laminates with di ﬀ erent thicknesses were immersed in the solution for the same ageing time, the water absorption of the specimens decreased ﬁrstly before reaching their peak water absorption and then increased in the later stage with the increase of specimen thickness, while the tensile strength retention sustaining increased as specimen thickness increased. The reason is that the thinner the specimen, the more severe the degradation. In this study, a new accelerated ageing method was proposed to predict the long-term water absorption and tensile strength of BFRP laminates. The accelerated factor of the proposed method was determined based on the specimen thickness. The proposed method was veriﬁed by test results with a good accuracy, indicating that the method could be used to predict long-term water absorption and tensile strength retention of BFRP laminates by considering specimen thickness in accelerating tests. W.Z.; analysis, G.C.; investigation, G.C. and Y.W.; resources, Y.W.; data curation, W.Z.; writing-original and


Introduction
Fiber-reinforced polymer (FRP) composites such as Carbon FRP (CFRP), Glass FRP (GFRP) and Aramid FRP (AFRP) have been widely used in infrastructure construction and other fields [1][2][3][4][5][6][7][8][9][10][11]. However, large-scale applications of FRP composites in infrastructure construction are still limited for some reasons, e.g., high cost of CFRP and AFRP composites, poor chemical stability of GFRP composites [12][13][14][15][16]. Over the past few years, basalt fiber has gradually received more attention as a new inorganic green fiber for its environment-friendly features. The basalt fiber is made of basalt stone after melting at 1450-1500 • C. There is no pollution during its production process [17,18]. The basalt BFRP laminates used in the current study were fabricated using wet lay-up method. The fiber direction of basalt fabrics was properly stacked to ensure that the unidirectional BFRP laminates were fabricated. The laminates with specific number of fabric layers (i.e., 2, 4, and 8 layers) were fabricated. Specially designed molds with the thickness h = 1, h = 2, and h = 4 mm to fabricate the laminates need to be used to ensure the required thickness of the laminates. The fiber volume fraction of the BFRP Appl. Sci. 2020, 10, 3618 4 of 21 laminates was 0.32. In this paper, BFRP laminates with the thickness h = 1, h = 2, and h = 4 mm were represented with B 1 , B 2 , and B 4 , respectively.

Water Absorption Test
According to ASTM D5229 [39], the water absorption specimens were subjected to deionized water and alkaline solution (pH = 13) at a temperature of 60 • C for a duration of 180 days. The setting pH of the alkaline solution is 13, which was used to simulate the pore water of the concrete [40]. The square specimens of water absorption (shown in Figure 1a) were cut from the fabricated BFRP laminates, and the dimensions of the specimens were 60 × 60 × h mm, where h represents the thickness of the BFRP laminates. The initial dry weight (W 0 ) of the specimen was weighed using a precision electronics balance (Model FA2004N, Shanghai Precision Scientific Instruments Co., Ltd., Shanghai, China) with an accuracy of 0.1 mg before immersion. After specific immersion time, the wet specimen was re-weighed. Before re-weighing, the water on the surface of the specimen was wiped using absorbent paper. The water absorption M(t) at specific ageing time is calculated by Equation (1). Ten specimens for each thickness were tested and the average values and standard deviation of water absorption for each thickness at specific immersion time were calculated.
where W t is the weight of wet specimen at specific ageing time t.

Water Absorption Test
According to ASTM D5229 [39], the water absorption specimens were subjected to deionized water and alkaline solution (pH = 13) at a temperature of 60 °C for a duration of 180 days. The setting pH of the alkaline solution is 13, which was used to simulate the pore water of the concrete [40]. The square specimens of water absorption (shown in Figure 1a) were cut from the fabricated BFRP laminates, and the dimensions of the specimens were 60 × 60 × h mm, where h represents the thickness of the BFRP laminates. The initial dry weight (W0) of the specimen was weighed using a precision electronics balance (Model FA2004N, Shanghai Precision Scientific Instruments Co., Ltd., Shanghai, China) with an accuracy of 0.1 mg before immersion. After specific immersion time, the wet specimen was re-weighed. Before re-weighing, the water on the surface of the specimen was wiped using absorbent paper. The water absorption M(t) at specific ageing time is calculated by Equation (1). Ten specimens for each thickness were tested and the average values and standard deviation of water absorption for each thickness at specific immersion time were calculated.

100%
(1) where Wt is the weight of wet specimen at specific ageing time t.

Tensile Test
The requirements for the tensile test of BFRP laminates were conducted in accordance with ASTM D3039 [41]. The rectangular tensile specimens with the dimensions of 250 × 15 × h mm (shown in Figure 1b) were cut from the fabricated BFRP laminates, which were soaked in deionized water and alkaline solution (pH = 13) at 60 °C last up to 180 days. As shown in Figure 1b, aluminum end-tabs were used in two ends to avoid the premature damage of the specimen during the tensile test. According to ASTM D3039 [41], the monotonic tensile test was conducted using a universal machine (Model WDE-200E, Jinan Gold Testing Machines Inc., Jinan, China) through displacement control with a displacement speed of 2 mm/min. The tensile stress  could be calculated according to the Equation (2) and the strain was measured using an extensometer with a gauge length (Model Y50/10, Changchun Sanjing Test Instrument Co., Ltd., Changchun, China) of 50 mm in the test, as shown in Figure 2. It should be noted that the acquired tensile stress was nominal tensile stress, which is more convenient to apply in practice engineering. Five specimens of each thickness were tested at specific duration, the average values and standard deviation of tensile strength retention and tensile elastic modulus were respectively calculated.

Tensile Test
The requirements for the tensile test of BFRP laminates were conducted in accordance with ASTM D3039 [41]. The rectangular tensile specimens with the dimensions of 250 × 15 × h mm (shown in Figure 1b) were cut from the fabricated BFRP laminates, which were soaked in deionized water and alkaline solution (pH = 13) at 60 • C last up to 180 days. As shown in Figure 1b, aluminum end-tabs were used in two ends to avoid the premature damage of the specimen during the tensile test. According to ASTM D3039 [41], the monotonic tensile test was conducted using a universal machine (Model WDE-200E, Jinan Gold Testing Machines Inc., Jinan, China) through displacement control with a displacement speed of 2 mm/min. The tensile stress σ could be calculated according to the Equation (2) and the strain was measured using an extensometer with a gauge length (Model Y50/10, Changchun Sanjing Test Instrument Co., Ltd., Changchun, China) of 50 mm in the test, as shown in Figure 2. It should be noted that the acquired tensile stress was nominal tensile stress, which is more convenient to apply in practice engineering. Five specimens of each thickness were tested at specific duration, the average values and standard deviation of tensile strength retention and tensile elastic modulus were respectively calculated.
where F represents the measured tensile force. A represents the nominal cross-sectional area of the specimens.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 5 of 21 where F represents the measured tensile force. A represents the nominal cross-sectional area of the specimens.

Scanning Electron Microscopy
Fractured surfaces of tensile failure specimens were observed using scanning electron microscopy (SEM) (Model Nova NanoSEM450, FEI Inc., Oregon, USA) at an accelerated voltage of 15 kV. All the observed specimens were sprayed with gold powder to increase their conductivity for easy observation [22,29]. Figure 3 shows the average values of the water absorption of BFRP laminates soaked in 60 °C deionized water and alkaline solution. It can be seen that the values of error bars were generally small, indicating that the dispersions of results were acceptable. It is noted that the water absorption of all specimens increased first before reaching their peak water absorption and then decreased as the ageing time increased. When soaked in deionized water and alkaline solution at 60 °C, the peak water absorption of the specimens B1, B2 and B4 were 1.31%, 1.16%, 0.96% and 1.60%, 1.52%, 1.50%, respectively. After a 180-day immersion, the water absorption of specimens B1, B2 and B4 decreased to 0.18%, −0.43%, −1.16% for deionized water immersion and to −0.27%, −0.73% and −2.15% for alkaline solution immersion, respectively. Generally, the water absorption of FRP composites increased gradually in the initial ageing duration. After that, the water absorption reached the dynamic moisture equilibrium or kept increasing. However, the water absorption curve of this paper, as shown in Figure 3, was quite different from the general cases reported in the existing literature [42,43]. The decline in water absorption after reaching its peak water absorption was probably caused by the serious degradation of the resin matrix in hygrothermal environment. The water absorption changes of BFFRP laminates were mainly affected by water immersion and hydrolysis of epoxy resin. When the mass gain of epoxy matrix from water immersion was less than the mass loss of epoxy resin due to the hydrolysis, the water absorption of BFRP laminates began to decline. As the degree of hydrolysis of epoxy resin increased, the water absorption of aged BFRP laminates were lower than that without ageing. This abnormal water absorption behavior (i.e., the decline in water absorption after reaching its peak water absorption) was also reported in the other literature [44][45][46].

Scanning Electron Microscopy
Fractured surfaces of tensile failure specimens were observed using scanning electron microscopy (SEM) (Model Nova NanoSEM450, FEI Inc., Oregon, USA) at an accelerated voltage of 15 kV. All the observed specimens were sprayed with gold powder to increase their conductivity for easy observation [22,29]. Figure 3 shows the average values of the water absorption of BFRP laminates soaked in 60 • C deionized water and alkaline solution. It can be seen that the values of error bars were generally small, indicating that the dispersions of results were acceptable. It is noted that the water absorption of all specimens increased first before reaching their peak water absorption and then decreased as the ageing time increased. When soaked in deionized water and alkaline solution at 60 • C, the peak water absorption of the specimens B 1 , B 2 and B 4 were 1.31%, 1.16%, 0.96% and 1.60%, 1.52%, 1.50%, respectively. After a 180-day immersion, the water absorption of specimens B 1 , B 2 and B 4 decreased to 0.18%, −0.43%, −1.16% for deionized water immersion and to −0.27%, −0.73% and −2.15% for alkaline solution immersion, respectively. Generally, the water absorption of FRP composites increased gradually in the initial ageing duration. After that, the water absorption reached the dynamic moisture equilibrium or kept increasing. However, the water absorption curve of this paper, as shown in Figure 3, was quite different from the general cases reported in the existing literature [42,43]. The decline in water absorption after reaching its peak water absorption was probably caused by the serious degradation of the resin matrix in hygrothermal environment. The water absorption changes of BFFRP laminates were mainly affected by water immersion and hydrolysis of epoxy resin. When the mass gain of epoxy matrix from water immersion was less than the mass loss of epoxy resin due to the hydrolysis, the water absorption of BFRP laminates began to decline. As the degree of hydrolysis of epoxy resin increased, the water absorption of aged BFRP laminates were lower than that without ageing. This abnormal water absorption behavior (i.e., the decline in water absorption after reaching its peak water absorption) was also reported in the other literature [44][45][46].   Figure 4 shows the degradation mechanism of epoxy resin soaked in deionized water or alkaline solution. It should be noted that the proportion of each part of the schematic is not drawn strictly according to the actual size, but just to better explain what needs to be expressed. The complete epoxy molecular chains of epoxy resin are presented in Figure 4a. When the BFRP laminates were immersed in high-temperature (i.e., 60 °C) deionized water or alkaline solution for a period of immersion time, lots of water molecules and/or OH − accelerate the erosion to epoxy resin, leading to the hydrolysis reaction of epoxy resin. The hydrolysis reaction is that the ether bonds (C-O) of epoxy molecular chains are broken as shown in Figure 4a. At this time, the epoxy molecular chains were broken down and formed some short molecular chains, dissolved in the solution finally. Macroscopically, the hydrolysis process of epoxy resin could be characterized by the appearance of some voids and micro-cracks after the water molecules and/or OH − deteriorated to epoxy resin as shown in Figure 4b. In order to verify the degradation mechanism, SEM observations were also conducted. The representative SEM images were selected from lots of SEM images; similar phenomena were observed in different regions of different samples. Figure 5a shows the surface of unaged specimen that was intact without ageing defects. Figure 5b,c represent the aged specimen in 60 °C deionized water and alkaline solution for 180 days, respectively. It can be seen that some obvious voids and micro-cracks appeared on the surface of the aged specimens, which supported the degradation mechanism of epoxy resin shown in Figure 4. Moreover, it can also be seen that, compared with the micro-cracks soaked in deionized water in Figure 5b, the extended micro-cracks appeared in Figure 5c due to the erosion of OH − in alkaline solution. A more severe hydrolysis reaction occurred on the specimens soaked in alkaline solution compared with the specimens soaked in deionized water. Therefore, the change range of water absorption was greater for alkaline solution immersion. In addition, it should be noted that it is significant to quantify the micro-cracks for a better understanding the hygrothermal ageing on epoxy resin. The related work will be included in a future study.  Figure 4 shows the degradation mechanism of epoxy resin soaked in deionized water or alkaline solution. It should be noted that the proportion of each part of the schematic is not drawn strictly according to the actual size, but just to better explain what needs to be expressed. The complete epoxy molecular chains of epoxy resin are presented in Figure 4a. When the BFRP laminates were immersed in high-temperature (i.e., 60 • C) deionized water or alkaline solution for a period of immersion time, lots of water molecules and/or OH − accelerate the erosion to epoxy resin, leading to the hydrolysis reaction of epoxy resin. The hydrolysis reaction is that the ether bonds (C-O) of epoxy molecular chains are broken as shown in Figure 4a. At this time, the epoxy molecular chains were broken down and formed some short molecular chains, dissolved in the solution finally. Macroscopically, the hydrolysis process of epoxy resin could be characterized by the appearance of some voids and micro-cracks after the water molecules and/or OH − deteriorated to epoxy resin as shown in Figure 4b. In order to verify the degradation mechanism, SEM observations were also conducted. The representative SEM images were selected from lots of SEM images; similar phenomena were observed in different regions of different samples. Figure 5a shows the surface of unaged specimen that was intact without ageing defects. Figure 5b,c represent the aged specimen in 60 • C deionized water and alkaline solution for 180 days, respectively. It can be seen that some obvious voids and micro-cracks appeared on the surface of the aged specimens, which supported the degradation mechanism of epoxy resin shown in Figure 4. Moreover, it can also be seen that, compared with the micro-cracks soaked in deionized water in Figure 5b, the extended micro-cracks appeared in Figure 5c due to the erosion of OH − in alkaline solution. A more severe hydrolysis reaction occurred on the specimens soaked in alkaline solution compared with the specimens soaked in deionized water. Therefore, the change range of water absorption was greater for alkaline solution immersion. In addition, it should be noted that it is significant to quantify the micro-cracks for a better understanding the hygrothermal ageing on epoxy resin. The related work will be included in a future study. Appl. Sci. 2020, 10, x FOR PEER REVIEW 7 of 21   Table 2 gives the average values with standard deviations of the tensile strength (f) and elastic modulus (E) of the specimens at specific ageing time. The average values of the tensile strength retention and elastic modulus retention are plotted against ageing time in Figures 6 and 7, respectively. It can be seen that the variation of error values was small overall, which can be considered that the results of the tensile test were accurate and reasonable. It can be found from Figure 6, in the first 14 days, the tensile strength decreased rapidly by 20%-40%. Afterward, the tensile strength showed a slower decline with the increase of ageing time. After an ageing of 180 days, the tensile strength retention of specimens B1, B2, and B4 in deionized water and alkaline solution were 34.0%, 37.7%, 40.6% and 20.1%, 26.3%, 33.1%, respectively. It can be seen that the    Table 2 gives the average values with standard deviations of the tensile strength (f) and elastic modulus (E) of the specimens at specific ageing time. The average values of the tensile strength retention and elastic modulus retention are plotted against ageing time in Figures 6 and 7, respectively. It can be seen that the variation of error values was small overall, which can be considered that the results of the tensile test were accurate and reasonable. It can be found from Figure 6, in the first 14 days, the tensile strength decreased rapidly by 20%-40%. Afterward, the tensile strength showed a slower decline with the increase of ageing time. After an ageing of 180 days, the tensile strength retention of specimens B1, B2, and B4 in deionized water and alkaline solution were 34.0%, 37.7%, 40.6% and 20.1%, 26.3%, 33.1%, respectively. It can be seen that the  Table 2 gives the average values with standard deviations of the tensile strength (f ) and elastic modulus (E) of the specimens at specific ageing time. The average values of the tensile strength retention and elastic modulus retention are plotted against ageing time in Figures 6 and 7, respectively. It can be seen that the variation of error values was small overall, which can be considered that the results of the tensile test were accurate and reasonable. It can be found from Figure 6, in the first 14 days, the tensile strength decreased rapidly by 20%-40%. Afterward, the tensile strength showed a slower decline with the increase of ageing time. After an ageing of 180 days, the tensile strength retention of specimens B 1 , B 2 , and B 4 in deionized water and alkaline solution were 34.0%, 37.7%, 40.6% and 20.1%, 26.3%, 33.1%, respectively. It can be seen that the thicker specimens had higher tensile strength retention than the thinner specimens. Figure 7 shows elastic modulus retention in deionized water and alkaline solution. The change trends of elastic modulus retention were similar to those of tensile strength retention. However, the degradation rate of elastic modulus is far less than that of tensile strength. After the ageing of 180 days, the maximum decrease of elastic modulus retention of all specimens was around 5%-20%. thicker specimens had higher tensile strength retention than the thinner specimens. Figure 7 shows elastic modulus retention in deionized water and alkaline solution. The change trends of elastic modulus retention were similar to those of tensile strength retention. However, the degradation rate of elastic modulus is far less than that of tensile strength. After the ageing of 180 days, the maximum decrease of elastic modulus retention of all specimens was around 5%-20%.    According to the SEM images of tensile failure surfaces of unaged and aged specimens shown in Figure 8, the degradation of tensile properties of BFRP laminates mainly comes from two parts. One is the degradation of epoxy resin and the corresponding interfacial debonding between the basalt fiber and the epoxy resin. Figure 8a exhibits the tensile failure surface of unaged specimen, where the basalt fiber was perfectly bonded to epoxy resin. It means that the basalt fiber and the epoxy resin play a synergistic role in the process of stress transmission during tension for unaged specimens. At the beginning of ageing, as shown in Figure 8b, the interfacial bond between basalt fiber and epoxy resin was gradually weakened due to the hydrolysis reaction of epoxy resin, leading to partial interfacial debonding. As the increase of ageing time, the hydrolysis of epoxy resin would be more severe as shown in Figure 8c (180 days). The other is the ageing of basalt fiber. SEM images of basalt fiber before and after ageing were shown in Figure 8d-f, respectively. It can be seen that the basalt fiber was also deteriorated due to the chemical reactions. Basalt fibers are composed of SiO2 with tetrahedral network structure, in which silicon and oxygen account for 27% and 44% of the total elements, respectively [47]. When basalt fiber was exposed to the solution, H2O and OH − played a significant role in the deterioration of basalt fiber. The chemical reaction formula of basalt fiber is shown as Equation (3). OH − attacks the molecular chains of SiO2, causing a part of SiO2 chains to relax. After that, the produced SiO − further react with H2O to produce OH − to intensify the reaction of Equation (3) [47,48], as follows in Equation (4). Besides, SiOH produced from SiO − is a kind of white colloid, which can transfer H2O and OH − to approach SiO2 bone chains, and further deteriorate basalt fiber. Since the reaction rate of SiO2 bone chains is relatively slow, the degradation of tensile strength and elastic modulus decreased slowly in the late stage, consistent with the experimental results (shown in Figures 6 and 7). According to the SEM images of tensile failure surfaces of unaged and aged specimens shown in Figure 8, the degradation of tensile properties of BFRP laminates mainly comes from two parts. One is the degradation of epoxy resin and the corresponding interfacial debonding between the basalt fiber and the epoxy resin. Figure 8a exhibits the tensile failure surface of unaged specimen, where the basalt fiber was perfectly bonded to epoxy resin. It means that the basalt fiber and the epoxy resin play a synergistic role in the process of stress transmission during tension for unaged specimens. At the beginning of ageing, as shown in Figure 8b, the interfacial bond between basalt fiber and epoxy resin was gradually weakened due to the hydrolysis reaction of epoxy resin, leading to partial interfacial debonding. As the increase of ageing time, the hydrolysis of epoxy resin would be more severe as shown in Figure 8c (180 days). The other is the ageing of basalt fiber. SEM images of basalt fiber before and after ageing were shown in Figure 8d-f, respectively. It can be seen that the basalt fiber was also deteriorated due to the chemical reactions. Basalt fibers are composed of SiO 2 with tetrahedral network structure, in which silicon and oxygen account for 27% and 44% of the total elements, respectively [47]. When basalt fiber was exposed to the solution, H 2 O and OH − played a significant role in the deterioration of basalt fiber. The chemical reaction formula of basalt fiber is shown as Equation (3). OH − attacks the molecular chains of SiO 2 , causing a part of SiO 2 chains to relax. After that, the produced SiO − further react with H 2 O to produce OH − to intensify the reaction of Equation (3) [47,48], as follows in Equation (4). Besides, SiOH produced from SiO − is a kind of white colloid, which can transfer H 2 O and OH − to approach SiO 2 bone chains, and further deteriorate basalt fiber. Since the reaction rate of SiO 2 bone chains is relatively slow, the degradation of tensile strength and elastic modulus decreased slowly in the late stage, consistent with the experimental results (shown in Figures 6 and 7). According to the SEM images of tensile failure surfaces of unaged and aged specimens shown in Figure 8, the degradation of tensile properties of BFRP laminates mainly comes from two parts. One is the degradation of epoxy resin and the corresponding interfacial debonding between the basalt fiber and the epoxy resin. Figure 8a exhibits the tensile failure surface of unaged specimen, where the basalt fiber was perfectly bonded to epoxy resin. It means that the basalt fiber and the epoxy resin play a synergistic role in the process of stress transmission during tension for unaged specimens. At the beginning of ageing, as shown in Figure 8b, the interfacial bond between basalt fiber and epoxy resin was gradually weakened due to the hydrolysis reaction of epoxy resin, leading to partial interfacial debonding. As the increase of ageing time, the hydrolysis of epoxy resin would be more severe as shown in Figure 8c (180 days). The other is the ageing of basalt fiber. SEM images of basalt fiber before and after ageing were shown in Figure 8d-f, respectively. It can be seen that the basalt fiber was also deteriorated due to the chemical reactions. Basalt fibers are composed of SiO2 with tetrahedral network structure, in which silicon and oxygen account for 27% and 44% of the total elements, respectively [47]. When basalt fiber was exposed to the solution, H2O and OH − played a significant role in the deterioration of basalt fiber. The chemical reaction formula of basalt fiber is shown as Equation (3). OH − attacks the molecular chains of SiO2, causing a part of SiO2 chains to relax. After that, the produced SiO − further react with H2O to produce OH − to intensify the reaction of Equation (3) [47,48], as follows in Equation (4). Besides, SiOH produced from SiO − is a kind of white colloid, which can transfer H2O and OH − to approach SiO2 bone chains, and further deteriorate basalt fiber. Since the reaction rate of SiO2 bone chains is relatively slow, the degradation of tensile strength and elastic modulus decreased slowly in the late stage, consistent with the experimental results (shown in Figures 6 and 7). According to the SEM images of tensile failure surfaces of unaged and aged specimens shown in Figure 8, the degradation of tensile properties of BFRP laminates mainly comes from two parts. One is the degradation of epoxy resin and the corresponding interfacial debonding between the basalt fiber and the epoxy resin. Figure 8a exhibits the tensile failure surface of unaged specimen, where the basalt fiber was perfectly bonded to epoxy resin. It means that the basalt fiber and the epoxy resin play a synergistic role in the process of stress transmission during tension for unaged specimens. At the beginning of ageing, as shown in Figure 8b, the interfacial bond between basalt fiber and epoxy resin was gradually weakened due to the hydrolysis reaction of epoxy resin, leading to partial interfacial debonding. As the increase of ageing time, the hydrolysis of epoxy resin would be more severe as shown in Figure 8c (180 days). The other is the ageing of basalt fiber. SEM images of basalt fiber before and after ageing were shown in Figure 8d-f, respectively. It can be seen that the basalt fiber was also deteriorated due to the chemical reactions. Basalt fibers are composed of SiO2 with tetrahedral network structure, in which silicon and oxygen account for 27% and 44% of the total elements, respectively [47]. When basalt fiber was exposed to the solution, H2O and OH − played a significant role in the deterioration of basalt fiber. The chemical reaction formula of basalt fiber is shown as Equation (3). OH − attacks the molecular chains of SiO2, causing a part of SiO2 chains to relax. After that, the produced SiO − further react with H2O to produce OH − to intensify the reaction of Equation (3) [47,48], as follows in Equation (4). Besides, SiOH produced from SiO − is a kind of white colloid, which can transfer H2O and OH − to approach SiO2 bone chains, and further deteriorate basalt fiber. Since the reaction rate of SiO2 bone chains is relatively slow, the degradation of tensile strength and elastic modulus decreased slowly in the late stage, consistent with the experimental results (shown in Figures 6 and 7).

Influence of Thickness on Water Absorption
The measured water absorption was analyzed against specimen thickness at specific ageing days (Figures 9 and 10). Figure 9a,b show that water absorption of the specimens decreased by increasing specimen thickness in the early stage (less than two days) of the immersion in deionized water and alkaline solution, respectively. In other words, the thinner the specimen thickness, the faster the water absorption saturation was achieved. As shown in Figure 9c,d, when the ageing time increased to four and seven days, the water absorption rose and then declined as the increasing specimen thickness. This is because in this ageing stage the change in the weight of specimen B1 began to be dominated by the hydrolysis of the epoxy resin (reducing the weight of the specimen) rather than water absorption (increasing the weight of the specimen). While the changes in the weight of specimen B2 and B3 were still dominated by water absorption. At this time, the mass gain of specimen B1 from water absorption was less than the mass loss of epoxy resin due to the hydrolysis, leading to a decrease of measured water absorption of specimen B1. As shown in Figure  10, the water absorption trend in late stage of the immersion (14-180 days) was opposite to that in early stage, indicating that the mass loss of all specimens due to the hydrolysis of epoxy resin was more than the mass gain of water absorption in this stage. The mass loss of thinner specimen (B1) was still greater than that of thicker specimens (B2, B4) in late stage, meaning that a thinner specimen had a greater hydrolysis degree at the given ageing time.

Influence of Thickness on Water Absorption
The measured water absorption was analyzed against specimen thickness at specific ageing days (Figures 9 and 10). Figure 9a,b show that water absorption of the specimens decreased by increasing specimen thickness in the early stage (less than two days) of the immersion in deionized water and alkaline solution, respectively. In other words, the thinner the specimen thickness, the faster the water absorption saturation was achieved. As shown in Figure 9c,d, when the ageing time increased to four and seven days, the water absorption rose and then declined as the increasing specimen thickness. This is because in this ageing stage the change in the weight of specimen B 1 began to be dominated by the hydrolysis of the epoxy resin (reducing the weight of the specimen) rather than water absorption (increasing the weight of the specimen). While the changes in the weight of specimen B 2 and B 3 were still dominated by water absorption. At this time, the mass gain of specimen B 1 from water absorption was less than the mass loss of epoxy resin due to the hydrolysis, leading to a decrease of measured water absorption of specimen B 1 . As shown in Figure 10, the water absorption trend in late stage of the immersion (14-180 days) was opposite to that in early stage, indicating that the mass loss of all specimens due to the hydrolysis of epoxy resin was more than the mass gain of water absorption in this stage. The mass loss of thinner specimen (B 1 ) was still greater than that of thicker specimens (B 2 , B 4 ) in late stage, meaning that a thinner specimen had a greater hydrolysis degree at the given ageing time.
It is noteworthy that the water absorption of BFRP laminates was studied by considering the epoxy resin and the basalt fiber as a whole in this study. It is significant to determine water absorptions of the resin and basalt fiber respectively for a better understanding of the water absorption mechanism. In a future study, the related work (i.e., separately studying the water absorption of the resin and the basalt fiber) will be incorporated in the authors' study target.  It is noteworthy that the water absorption of BFRP laminates was studied by considering the epoxy resin and the basalt fiber as a whole in this study. It is significant to determine water absorptions of the resin and basalt fiber respectively for a better understanding of the water absorption mechanism. In a future study, the related work (i.e., separately studying the water absorption of the resin and the basalt fiber) will be incorporated in the authors' study target.

Influence of Thickness on Tensile Strength
The tensile strength retention vs. specimen thickness under different ageing durations was plotted in Figure 11. It can be observed that the specimens with thinner thickness had smaller  It is noteworthy that the water absorption of BFRP laminates was studied by considering the epoxy resin and the basalt fiber as a whole in this study. It is significant to determine water absorptions of the resin and basalt fiber respectively for a better understanding of the water absorption mechanism. In a future study, the related work (i.e., separately studying the water absorption of the resin and the basalt fiber) will be incorporated in the authors' study target.

Influence of Thickness on Tensile Strength
The tensile strength retention vs. specimen thickness under different ageing durations was plotted in Figure 11. It can be observed that the specimens with thinner thickness had smaller

Influence of Thickness on Tensile Strength
The tensile strength retention vs. specimen thickness under different ageing durations was plotted in Figure 11. It can be observed that the specimens with thinner thickness had smaller retention than that of specimens with thicker thickness at given ageing time. For example, the tensile strength retention of specimen B 1 was 5%-20% lower than that of specimen B 4 . Combining the test results of tensile test and water absorption, it can be concluded that at the beginning of ageing (around 30 days), the specimen with a thinner thickness had a faster deterioration rate than that of the specimen with a thicker thickness. After that, the deterioration rate was not sensitive to the thickness, while the ageing degree of the specimen with a thinner thickness was more severe than that of the specimen with a thicker thickness during a long period of ageing. Figure 12 shows the ageing of the specimens with different thicknesses through the cross section. It should be noted that the aged areas shown in Figure 12 only represent the aged areas of the specimens at a certain ageing time and the aged areas will gradually increase with the increasing ageing time. Additionally, note that the proportion of each part of the schematic is not drawn strictly according to the actual size, but just to better explain what needs to be expressed. Since the thickness of the specimen is much smaller than its length and width, it can be assumed that the deterioration of the specimen is carried out along the direction of the thickness. The aged area gradually expands to the interior of the specimen with the increase of ageing time. Although the thicknesses of B 1 , B 2 , and B 4 are different, the diffusion coefficient of the immersion solution is identical due to the same fabrication materials [49,50]. Therefore, the aged areas of B 1 , B 2 and B 4 are the same after a same certain ageing time. It can be seen clearly that the thicker specimen has a larger unaged area, i.e., S 1 < S 2 < S 4 . As the ageing time increases, the relationship between the unaged area S of BFRP laminates with different thicknesses was still established, i.e., S 1 < S 2 < S 4 . Therefore, a thinner specimen is inferior to a thicker one on long-term hygrothermal properties due to a faster degradation rate.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 12 of 21 retention than that of specimens with thicker thickness at given ageing time. For example, the tensile strength retention of specimen B1 was 5%-20% lower than that of specimen B4. Combining the test results of tensile test and water absorption, it can be concluded that at the beginning of ageing (around 30 days), the specimen with a thinner thickness had a faster deterioration rate than that of the specimen with a thicker thickness. After that, the deterioration rate was not sensitive to the thickness, while the ageing degree of the specimen with a thinner thickness was more severe than that of the specimen with a thicker thickness during a long period of ageing. Figure 12 shows the ageing of the specimens with different thicknesses through the cross section. It should be noted that the aged areas shown in Figure 12 only represent the aged areas of the specimens at a certain ageing time and the aged areas will gradually increase with the increasing ageing time. Additionally, note that the proportion of each part of the schematic is not drawn strictly according to the actual size, but just to better explain what needs to be expressed. Since the thickness of the specimen is much smaller than its length and width, it can be assumed that the deterioration of the specimen is carried out along the direction of the thickness. The aged area gradually expands to the interior of the specimen with the increase of ageing time. Although the thicknesses of B1, B2, and B4 are different, the diffusion coefficient of the immersion solution is identical due to the same fabrication materials [49,50]. Therefore, the aged areas of B1, B2 and B4 are the same after a same certain ageing time. It can be seen clearly that the thicker specimen has a larger unaged area, i.e., S1 < S2 < S4. As the ageing time increases, the relationship between the unaged area S of BFRP laminates with different thicknesses was still established, i.e., S1 < S2 < S4. Therefore, a thinner specimen is inferior to a thicker one on long-term hygrothermal properties due to a faster degradation rate.  retention than that of specimens with thicker thickness at given ageing time. For example, the tensile strength retention of specimen B1 was 5%-20% lower than that of specimen B4. Combining the test results of tensile test and water absorption, it can be concluded that at the beginning of ageing (around 30 days), the specimen with a thinner thickness had a faster deterioration rate than that of the specimen with a thicker thickness. After that, the deterioration rate was not sensitive to the thickness, while the ageing degree of the specimen with a thinner thickness was more severe than that of the specimen with a thicker thickness during a long period of ageing. Figure 12 shows the ageing of the specimens with different thicknesses through the cross section. It should be noted that the aged areas shown in Figure 12 only represent the aged areas of the specimens at a certain ageing time and the aged areas will gradually increase with the increasing ageing time. Additionally, note that the proportion of each part of the schematic is not drawn strictly according to the actual size, but just to better explain what needs to be expressed. Since the thickness of the specimen is much smaller than its length and width, it can be assumed that the deterioration of the specimen is carried out along the direction of the thickness. The aged area gradually expands to the interior of the specimen with the increase of ageing time. Although the thicknesses of B1, B2, and B4 are different, the diffusion coefficient of the immersion solution is identical due to the same fabrication materials [49,50]. Therefore, the aged areas of B1, B2 and B4 are the same after a same certain ageing time. It can be seen clearly that the thicker specimen has a larger unaged area, i.e., S1 < S2 < S4. As the ageing time increases, the relationship between the unaged area S of BFRP laminates with different thicknesses was still established, i.e., S1 < S2 < S4. Therefore, a thinner specimen is inferior to a thicker one on long-term hygrothermal properties due to a faster degradation rate.

Theoretical Model of Accelerated Factor
The above experimental results indicated clearly that the water absorption and tensile strength retention of BFRP laminate were dramatically influenced by specimen thickness. As illustrated in Figures 3 and 6, it could be observed that the ageing degradation (i.e., the changing trends of water absorption and reduction of tensile strength retention) of BFRP laminates with thinner specimens was more severe than that of thicker specimens when the BFRP laminates were soaked in the solution for a same long time. Therefore, there is a possibility to deduce an accelerated ageing model based on specimen thickness for water absorption and tensile strength retention of BFRP laminates under hygrothermal ageing environment. In this section, a theoretical accelerated ageing model based on specimen thickness was developed and the accelerated factors (AFs) related to specimen thickness were theoretically deduced.

Accelerated Factor of Water Absorption
Considerable researches show that the water absorption of FRP composites can be described by two models as shown in Figure 13, i.e., the Fick's model and Two-stage model [39,42,51]. The Fick's model assumes that the initial phase of water absorption increases linearly with t 1/2 in the initial phase, and then increases non-linearly until the water absorption reaches a dynamic equilibrium without obvious changes. For the Two-stage model, the water absorption is identical to Fick's model in the initial phase. But the water absorption cannot reach the equilibrium stage due to the water immersion constantly and the degradation of resin matrix. In the current study, the test results of water absorption conform to the Two-stage model.

Theoretical Model of Accelerated Factor
The above experimental results indicated clearly that the water absorption and tensile strength retention of BFRP laminate were dramatically influenced by specimen thickness. As illustrated in Figures 3 and 6, it could be observed that the ageing degradation (i.e., the changing trends of water absorption and reduction of tensile strength retention) of BFRP laminates with thinner specimens was more severe than that of thicker specimens when the BFRP laminates were soaked in the solution for a same long time. Therefore, there is a possibility to deduce an accelerated ageing model based on specimen thickness for water absorption and tensile strength retention of BFRP laminates under hygrothermal ageing environment. In this section, a theoretical accelerated ageing model based on specimen thickness was developed and the accelerated factors (AFs) related to specimen thickness were theoretically deduced.

Accelerated Factor of Water Absorption
Considerable researches show that the water absorption of FRP composites can be described by two models as shown in Figure 13, i.e., the Fick's model and Two-stage model [39,42,51]. The Fick's model assumes that the initial phase of water absorption increases linearly with t 1/2 in the initial phase, and then increases non-linearly until the water absorption reaches a dynamic equilibrium without obvious changes. For the Two-stage model, the water absorption is identical to Fick's model in the initial phase. But the water absorption cannot reach the equilibrium stage due to the water immersion constantly and the degradation of resin matrix. In the current study, the test results of water absorption conform to the Two-stage model. According to the ageing mechanism reported in the literature [49,50], the following assumptions can be used to establish the water absorption model. First, the specimens with different thicknesses have the same diffusion rate for the same immersion solution because the diffusion rate is controlled by the concentration gradient of the immersion solution. Second, although the water absorption at the regular time varies with specimen thickness, the ageing mechanism of the composite remains unchanged. Third, the specimens with different thicknesses fabricated by the same material will eventually reach the same water absorption rate under the same immersion solution.
The Fick's and Two-stage models can be used to describe the relationship between specimen thickness and ageing time in stage I (mass gain due to swelling of FRP composites) and stage II (mass loss due to relaxation of FRP composites), respectively. In the two models, the stage I of water absorption is identical. The Fick's model is shown in Equation (5), but it is difficult to find the According to the ageing mechanism reported in the literature [49,50], the following assumptions can be used to establish the water absorption model. First, the specimens with different thicknesses have the same diffusion rate for the same immersion solution because the diffusion rate is controlled by the concentration gradient of the immersion solution. Second, although the water absorption at the regular time varies with specimen thickness, the ageing mechanism of the composite remains unchanged. Third, the specimens with different thicknesses fabricated by the same material will eventually reach the same water absorption rate under the same immersion solution.
The Fick's and Two-stage models can be used to describe the relationship between specimen thickness and ageing time in stage I (mass gain due to swelling of FRP composites) and stage II (mass loss due to relaxation of FRP composites), respectively. In the two models, the stage I of water absorption is identical. The Fick's model is shown in Equation (5), but it is difficult to find the relationship among different parameters in this expression, which is needed to be simplified as presented in Equation (6) [36,40].
where M ∞ is the effective moisture equilibrium content, D is the diffusion coefficient. When two specimens with the thicknesses of h 1 and h 2 soaked in the same solution reach the same water absorption at ageing time t 1 and t 2 , respectively. The equilibrium equation, i.e., M(t 1 ) = M(t 2 ) can be expressed as Equation (7) according to Fick's model, which can be simplified to Equation (8).
According to the previous assumptions, the diffusion coefficient D is identical when the specimens are soaked in the same solution, yields Therefore, the AF of water absorption in stage I of Two-stage model can be expressed in After the above derivation of AF in Stage I, it can be found that there is a certain relationship between ageing time t and specimen thickness h indeed. To further explore the relationship between water absorption, ageing time t and specimen thickness h in stage II of Two-stage model, the abscissa of water absorption curves can be changed from t 1/2 to t/h 2 , as shown in Figure 14. According to the literature [48], the changing the abscissa (from t 1/2 to t/h 2 ) does not change the trends of water absorption in Stage I and Stage II. The water absorption curve of Two-stage model in Stage II was regarded as declining linearly that is determined by the slope of the descending section, tanα, which is only related to the materials of FRP composite and conditional environment [42,51]. It is known clearly that the tanα of water absorption curves in Stage II is identical for the specimens B 1 B 2 and B 4 due to the same materials of BFRP laminates and conditional environment. Therefore, in the whole Stage II of the Two-stage model, the relationship between water absorption M(t) at any given time and the maximum water absorption M m can be expressed as When the specimens with two thicknesses h 1 and h 2 reach the same water absorption at t 1 and t 2 , respectively, the equilibrium equation can be expressed as Equation (12) Equation (12) can be simplified as It has been known by the derivation of Stage I in the two water absorption models that the Equation (10) is applicable when the value of water absorption increases to M m , i.e., Thus, the AF of stage II in Two-stage model can be expressed by Equation (10).
Thus, the AF of stage II in Two-stage model can be expressed by Equation (10 Therefore, the accelerated factor based on specimen thickness for water absorption is deduced theoretically, which can be applied in Fick's model and Two-stage model.

Accelerated Factor of Tensile Strength Retention
When the tensile specimens are soaked in the solution, the tensile strength decreases due to the gradual increase of the ageing area. The ageing condition of tensile specimens can be represented by the change of the cross section as shown in Figure 15, which is assumed that the specimen is uniformly aged along the thickness direction on the upper and lower sides. Thus, the actual tensile strength  can be expressed by Equation (15).
where u and a are the initial (unaged) and residual (aged) tensile strength of the specimen, respectively. Au and Aa are the unaged area and aged area of the specimen, respectively. b and h are the width and thickness of the specimen, respectively. x is the total ageing depth along the thickness of the laminates. It has been reported by study [49,52,53] that the total ageing depth x of the laminates is proportional to the square root of the immersion time t, which is as follows regarded as a function of Therefore, the accelerated factor based on specimen thickness for water absorption is deduced theoretically, which can be applied in Fick's model and Two-stage model.

Accelerated Factor of Tensile Strength Retention
When the tensile specimens are soaked in the solution, the tensile strength decreases due to the gradual increase of the ageing area. The ageing condition of tensile specimens can be represented by the change of the cross section as shown in Figure 15, which is assumed that the specimen is uniformly aged along the thickness direction on the upper and lower sides. Thus, the actual tensile strength σ can be expressed by Equation (15).
where σ u and σ a are the initial (unaged) and residual (aged) tensile strength of the specimen, respectively. A u and A a are the unaged area and aged area of the specimen, respectively. b and h are the width and thickness of the specimen, respectively. x is the total ageing depth along the thickness of the laminates.
Thus, the AF of stage II in Two-stage model can be expressed by Equation (10 Therefore, the accelerated factor based on specimen thickness for water absorption is deduced theoretically, which can be applied in Fick's model and Two-stage model.

Accelerated Factor of Tensile Strength Retention
When the tensile specimens are soaked in the solution, the tensile strength decreases due to the gradual increase of the ageing area. The ageing condition of tensile specimens can be represented by the change of the cross section as shown in Figure 15, which is assumed that the specimen is uniformly aged along the thickness direction on the upper and lower sides. Thus, the actual tensile strength  can be expressed by Equation (15).
where u and a are the initial (unaged) and residual (aged) tensile strength of the specimen, respectively. Au and Aa are the unaged area and aged area of the specimen, respectively. b and h are the width and thickness of the specimen, respectively. x is the total ageing depth along the thickness of the laminates. It has been reported by study [49,52,53] that the total ageing depth x of the laminates is proportional to the square root of the immersion time t, which is as follows regarded as a function of It has been reported by study [49,52,53] that the total ageing depth x of the laminates is proportional to the square root of the immersion time t, which is as follows regarded as a function of where α is a constant of the FRP composite that is irrelevant to specimen thickness h. Then Equation (15) can be transformed as follows According to the literature [50,54], for the given resin matrix, fiber, fabrication and immersion condition, the specimens with different thicknesses will decrease to the same strength when the ageing time is long enough. As a result, the equilibrium equation can be expressed as Equation (18) when the tensile strengths of the specimens when two thicknesses, h 1 and h 2 , decrease to the same value at ageing time t 1 and t 2 , respectively.
Thus, the AF of tensile strength retention of BFRP laminates can also be written as Equation (10). From the previous theoretical derivation, for BFRP laminates with two thicknesses, when the water absorption or tensile strength retention of the BFRP laminates with two different thicknesses reaches the same value, the ratio of the ageing time of the BFRP laminates with the two thicknesses is proportional to the square of the ratio of the corresponding two thicknesses. The square of the ratio of the two thicknesses is considered as accelerated factor (AF). As a result, it is feasible to accelerate ageing by reducing the specimen thickness. For example, if the water absorptions or tensile strength retentions of two BFRP laminates with different thicknesses reach a certain same value, the predicted ageing time of thicker specimen (t 1 ) can be calculated/accelerated by multiplying the real ageing time of thinner specimen (t 2 ) by the corresponding accelerated factor, AF = (h 1 /h 2 ) 2 , based on the two thicknesses of the thicker specimen (h 1 ) and thinner specimen (h 2 ).

Model Validation and Discussions
In this study, the ageing accelerated method was taking the BFRP specimen with the thickness h = 4 mm as the standard specimen. The AFs of BFRP specimens with h = 1 mm, h = 2 mm, and h = 4 mm are calculated by Equation (10) and the results are shown in Table 3. The obtaining of accelerated ageing days is transformed by multiplying the actual ageing days of the thicker specimen by the corresponding AF based on the two thicknesses of the thicker specimen and the standard specimen. According to the existing test results, the accelerated time is up to eight years on the water absorption trend and tensile strength retention of BFRP specimen with h = 4 mm. It should be noted that the transformed results of the specimens with 1 and 2 mm represent the predicted results of the standard specimen h = 4 mm. Figure 16 shows the long-term prediction of water absorption of BFRP specimens with h = 4 mm according to the transformed results of 1 and 2 mm specimens in deionized water and alkaline solution. The predicted curves were fitted by using Matlab (2018a Version, MathWorks, Inc., Massachusetts, USA, 2018). It can be seen from Figure 16 that the predicted curves are fitted well with a Two-stage model in whole. It reveals that the predicted law of water absorption in deionized water was better with Two-stage model than that in alkaline solution. Figure 17 shows the long-term prediction of tensile strength retention of BFRP specimens with h = 4 mm according to the transformed results of 1 and 2 mm specimens in deionized water and alkaline solution, which is also fitted well with Phani-Bose's model [55] in whole. Compared with the traditional temperature-dependent accelerated ageing method, there are two main advantages of the proposed new specimen thickness-dependent accelerated ageing method in this study. First, the proposed thickness-dependent accelerated ageing method is easy to apply because this method does not need the activation energy, which must be required in the temperature-dependent method. In order to obtain the activation energy in the temperature-dependent method, at least three different temperature environment tests must be conducted [56], which increases the test difficulty. In contrast, the accelerated factor calculating in the proposed method is only dependent on specimen thickness, which would be easy to be conducted. Second, the AF of the temperature-dependent accelerated ageing method is limited due to the limitation of Tg of FRP composites, leading to the acceleration times being relatively small. The AF of the proposed method is only dependent on specimen thickness and thus eliminates the limitation of Tg. Therefore, greater AF can be obtained in the proposed method.

Conclusions
In this paper, the BFRP laminates with the thickness of h = 1, h = 2, and h = 4 mm were fabricated by wet-layup method and the influence of thickness on their water absorption and tensile properties were experimentally studied under hygrothermal environment as well as degradation mechanism. A specimen thickness-dependent accelerated ageing method was proposed. The following conclusions can be drawn from the testing results and discussions in this study: The long-term properties of BFFRP laminates were greatly affected under hygrothermal environment. The water absorption trend of BFRP laminates soaked in both 60 C deionized water and alkaline solution increased first before reaching their peak water absorption and then decreased Compared with the traditional temperature-dependent accelerated ageing method, there are two main advantages of the proposed new specimen thickness-dependent accelerated ageing method in this study. First, the proposed thickness-dependent accelerated ageing method is easy to apply because this method does not need the activation energy, which must be required in the temperature-dependent method. In order to obtain the activation energy in the temperature-dependent method, at least three different temperature environment tests must be conducted [56], which increases the test difficulty. In contrast, the accelerated factor calculating in the proposed method is only dependent on specimen thickness, which would be easy to be conducted. Second, the AF of the temperature-dependent accelerated ageing method is limited due to the limitation of Tg of FRP composites, leading to the acceleration times being relatively small. The AF of the proposed method is only dependent on specimen thickness and thus eliminates the limitation of Tg. Therefore, greater AF can be obtained in the proposed method.

Conclusions
In this paper, the BFRP laminates with the thickness of h = 1, h = 2, and h = 4 mm were fabricated by wet-layup method and the influence of thickness on their water absorption and tensile properties were experimentally studied under hygrothermal environment as well as degradation mechanism. A specimen thickness-dependent accelerated ageing method was proposed. The following conclusions can be drawn from the testing results and discussions in this study: The long-term properties of BFFRP laminates were greatly affected under hygrothermal environment. The water absorption trend of BFRP laminates soaked in both 60 C deionized water and alkaline solution increased first before reaching their peak water absorption and then decreased

Conclusions
In this paper, the BFRP laminates with the thickness of h = 1, h = 2, and h = 4 mm were fabricated by wet-layup method and the influence of thickness on their water absorption and tensile properties were experimentally studied under hygrothermal environment as well as degradation mechanism. A specimen thickness-dependent accelerated ageing method was proposed. The following conclusions can be drawn from the testing results and discussions in this study: The long-term properties of BFFRP laminates were greatly affected under hygrothermal environment. The water absorption trend of BFRP laminates soaked in both 60 • C deionized water and alkaline solution increased first before reaching their peak water absorption and then decreased with the increase of immersion duration, which was caused by the hydrolysis of the epoxy matrix. The tensile properties of BFRP laminates degraded apparently after ageing, especially in alkaline solution. SEM images show basalt fiber deteriorated due to the solution immersion.
Specimen thickness had a significant influence on the water absorption and tensile strength of BFRP laminates after ageing. When the BFRP laminates with different thicknesses were immersed in the water or alkaline solution for the same ageing time, the water absorption decreased in early stage of immersion and then increased in late stage of immersion as the specimen thickness increased, while the tensile strength retention kept increased during the whole ageing process. The reason is that the ratio of aged area to the total area in the thinner specimens was larger than that of the thicker specimens, leading to the more severe ageing degree of the thinner specimens.
An innovative thickness-dependent accelerated ageing method for water absorption and tensile strength retention of BFRP laminates was proposed, in which the accelerated factors were theoretically deduced based on specimen thickness. The proposed method is in good agreement with test results. Compared with the traditional accelerated ageing method based on temperature, the proposed method is much easier to be conducted and has the potential to obtain a greater accelerated factor.

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