Microstructure Characteristics of GFRP Reinforcing Bars in Harsh Environment

Fiber reinforced polymer (FRP) composites have been suggested as corrosion-resistant alternatives to traditional steel reinforcement in concrete structures. Within this family of composites, glass ﬁber reinforced polymers (GFRPs) have been gaining momentum as the primary selection of FRP for construction applications. Despite being advantageous, its wide adoption by the industry has been hindered due to the degradation of its performance in severe environmental conditions. As such, signiﬁcant studies have been carried out to assess the mechanical properties of GFRP bars subject to diﬀerent conditioning schemes. However, the inconsistencies and wide variations of results called for more in-depth microstructure evaluation. Accordingly, this paper presents a critical review of existing research on the microstructure of GFRP reinforcing bars exposed to various conditioning regimes. The review analysis revealed that sustained load limits set by codes and standards were satisfactory for nonaggressive environment conditions but should be updated to include diﬀerent conditioning regimes. It was also found that conditioning in alkaline solutions was more severe than concrete and mortars, where test specimens experienced irreversible chemical degradation, more hydroxyl group formation, and more intense degradation to the microstructure. The progression of hydrolysis was reported correlatively through an increase in hydroxyl groups and a decrease in the glass transition temperature. While moisture uptake was the primary instigator of hydrolysis, restricting it to 1.6% could limit the reduction in tensile strength to 15%. Further, the paper identiﬁes research gaps in the existing knowledge and highlights directions for future research.


Introduction
Reinforced concrete structures are designed with certain life expectancy under normal service conditions. However, exposure to severe environments may reduce the design service life due to unanticipated durability problems, including corrosion of steel reinforcement. e corrosion of steel reinforcing bars causes cracking and spalling of concrete, creating major reductions in performance and a significant increase in cost for rehabilitation. In fact, Koch [1] reported that 3-4% of each nation's gross domestic product is dedicated to corrosion-related expenditures. As a form of corrosion management, glass fiber reinforced polymer (GFRP) reinforcing bars have been proposed as a replacement to steel equivalents. is alternative reinforcement is characterized by its light weight, high strength-to-weight ratio, and corrosion resistance [2][3][4]. However, conflicting results on the degradation of its properties upon exposure to alkaline or acidic solution, moisture/water, and elevated temperatures have been a major setback in its adoption by the construction industry [2][3][4].
For the last few years, extensive research has been carried out to investigate the durability performance of GFRP reinforcement exposed to different environmental conditions. Research findings have identified the major contributors to deterioration of GFRP as moisture uptake, alkaline environment, and temperature. Moisture penetrates the matrix by the flow of solution into the microgaps between polymer chains and capillary transport through fiber/matrix interfacial microcracks and in microcracks formed during manufacturing [5]. As a result, the matrix exhibits plasticisation, leading to a reduction in strength and glass transition temperature. On the other hand, alkaline solutions break the bond between the oxygen and carbon atoms in the molecular chain of the GFRP matrix. is breakage creates microcracks and fractures in the matrix structure, leading to more moisture uptake and a significant reduction in strength. Further, to expedite investigative work conducted on GFRP reinforcing bars, higher temperatures have been employed. Vijay and GangaRao [6] reported that an increase in the conditioning temperature led to an exponential amplification in the actual age. Nevertheless, it should be noted that the elevated temperatures employed in most work may have affected the physical properties of the bars by increasing the coefficient of thermal expansion [7] and, thus, may have altered the behavior of the GFRP bars, rendering the results, in some cases, unrealistic.
With the majority of past work characterizing the durability performance of GFRP bars, several researchers have reviewed and summarized these publications to highlight the gaps in existing knowledge and propose future direction of research. In contrast, much fewer and more recent investigations have correlated the degradation in GFRP to changes in the microstructure. Yet, there is no such review and summary of these in-depth investigations. Accordingly, this paper aims to present an overview of recent studies examining the effect of conditioning on the microstructure of GFRP reinforcing bars. Research findings obtained help to shed light on the morphological and microstructure changes in GFRP bars subject to different exposure conditions using various microstructure characterization techniques, including differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), Fourier transform infrared spectroscopy (FTIR), and scanning electron microscopy (SEM). A correlation among these microstructural evaluation techniques is reported. e influence of moisture uptake on the microstructure of conditioned GFRP bars is also evaluated.

Taxonomy
GFRP reinforcing bars are made of glass fibers impregnated with an organic resin or matrix. Typical diameters range between 8 and 20 mm with smooth surfaces for smaller sizes and ribbed and/or sand-coated surfaces for larger counterparts. GFRP reinforcement falls within the family of FRP composites. e taxonomy on the studies of FRP composites is presented in Figure 1. ey can be classified based on their fiber type, resin type, and application. Different combinations of fibers and resins have been employed in the production of FRPs, of which composites made with glass fibers and vinylester are the most common. Also, FRPs have been used across many industries, including aerospace, automotive, sports, musical, and construction. In the latter industry, they have been employed in new and existing buildings as internal and external reinforcing bars and for repair, strengthening, and retrofitting purposes. e literature has especially focused on the ability of FRPs to replace steel as the main reinforcement in concrete structures [8][9][10][11][12][13][14][15]. However, before being widely adopted by the construction industry, their durability performance was rigorously investigated. Accordingly, GFRP reinforcing bars have been conditioned in various media, including acidic and alkaline solutions, concrete, tap water, and seawater. Elevated temperatures were also employed during conditioning to accelerate the degradation process. Further, the validity of the environmental reduction factors and creep rupture stress limits adopted by the current international guidelines and standards were assessed by applying a sustained load to GFRP bars exposed to severe environments. Such specimens were examined for performance retention by measuring the tensile properties after conditioning. Correlations between the properties and microstructural changes were evaluated using DSC, TGA, SEM, FTIR, and moisture uptake.

GFRP Reinforcing Bars
e properties of GFRP reinforcing bars and conditions to which they are exposed are primary factors affecting their performance and microstructure. As such, a comprehensive review of the literature on the physical and mechanical properties and conditioning of GFRP bars is presented in the sections below. It should be noted that the literature reviewed in this paper focuses on examining the microstructure characteristics of GFRP bars exposed to different conditioning regimes.

Physical Properties.
Prior to conducting any experimental testing on GFRP reinforcing bars, their physical properties were determined. Various tests, shown in Table 1, were typically carried out to measure the diameter, void content, fiber content, moisture uptake, glass transition temperature, and thermal expansion coefficient. Physical properties reported in past studies are summarized in Table 2. e most commonly used glass and matrix materials were E-glass and vinylester, with 90 and 74% of the conducted research having used this fiber and matrix, respectively, as shown in Figure 2. Epoxy, on the other hand, has been only recently employed as the resin of GFRP reinforcing bars, with 14% of published work having explored its combination with E-glass [8,10,11,23,41,42]. e diameter has also been varied between 8 and 44.5 mm, with the majority being approximately 12 mm. Yet, Benmokrane, et al. [24] concluded that the bar diameter had limited influence on the physical and mechanical properties of GFRP reinforcing bars. Furthermore, very limited work reported the thermal expansion coefficients, as it was not typically measured by researchers, but provided by the manufacturer [8,13,14,[24][25][26][27]. In contrast, the moisture uptake has been essential in characterizing the performance and assessing its retention capacity. It ranged between 0.01 and 1.59%, by mass, for unconditioned/control specimens. e fiber content was measured by volume and mass. However, the latter was more commonly reported with the availability of standardized test procedures (ASTM D3171 and ASTM E1868). Measured values ranged between 70 and 83%. e volumetric fiber content, on the other hand, was mainly provided by the manufacturer and could only be determined at the initial stages of production. 2 Advances in Materials Science and Engineering

Mechanical Properties.
e performance of GFRP reinforcing bars is typically characterized by its ability to retain its mechanical properties. For this reason, as-received and unconditioned specimens were tested for various performance indicators, as summarized in Table 3, including tensile and flexural strength, strain, and modulus. Shear strength was also reported in relevant work. Experimental results of a number of studies are shown in Table 4. Over 90% of the studies focused on tensile properties, with values of tensile strength, modulus, and strain in the range of 432-1321 MPa, 31-65 GPa, and 1-3%, respectively. Given the large database of results for the former two properties, it was possible to develop a relation between them. Figure 3 presents a correlation between tensile modulus (E t ) and  Advances in Materials Science and Engineering 3 square root of tensile strength ( �� f t ) of GFRP bars. A linear trend line, in the form of Equation (1) could provide a prediction of tensile modulus from the strength with reasonable accuracy. On the other hand, flexural and shear properties were recorded on few occasions, mainly in research work related to bending and shear testing. Yet, the available data have shown a linear relationship between flexural (f r ) and tensile strength (f t ), as shown in Figure 4 and Equation (2).
is linear relation provides a reasonably accurate prediction of the former from the latter without the need for experimental testing. It is also worth noting that tensile properties of unconditioned GFRP samples made with epoxy as the matrix material seem to be, on average, 14% higher than those of vinylester counterparts. Additionally, a correlation was made between the physical and mechanical properties of GFRP specimens made with E-glass and vinylester. From Figure 5, it is clear that the tensile strength and modulus are generally proportional to the fiber content, up to a maximum value of 83%, by mass. Similarly, Micelli and Nanni [44] reported that GFRP bars 3.3. Conditioning. e durability performance of GFRP reinforcing bars exposed to di erent environmental conditions has been evaluated using accelerated aging tests. Four main exposure parameters were investigated: surrounding medium, conditioning duration, conditioning temperature, and presence of a sustained load. Table 5 summarizes the exposure conditions employed by various studies on the microstructure of GFRP bars. It should be noted that only one study reviewed in this work, assessing the microstructure, has investigated the e ect of acidic solutions on the performance of GFRP. Seawater exposure has also received little attention, although it is critical to assess the properties and microstructure of GFRP bars for employment in o shore structures and bridge decks in coastal cities and seaports [10,11,42]. According to Figure 6(a), alkaline and water solutions have been utilized in 39 and 25% of the studies shown herein. It seems that the ease of preparing such media renders them most favourable among researchers, especially that the former is proposed to simulate concrete alkalinity and behavior. Concrete, on the other hand, has been used by 23% of presented research, even though it better represents reallife scenarios. e duration of conditioning has been widely altered in research studies and ranged between 0.02 and 7200 days. Figure 6(b) shows that 28, 18, 10, 23, and 21% of studies have conditioned GFRP bars for a maximum of 0-90, 91 to 180, 181 to 240, 241 to 365, and more than 365 days, respectively. Clearly, the majority of research has focused on examining the e ects of early age and prolonged conditioning. e conditioning temperature has been pivotal in accelerating aging of GFRP specimens. Work by Vijay and GangaRao [6] highlighted the relationship between actual age in calendar days at Morgantown, WV, and chamber conditioning days in the form of Equation (3), where T is the temperature in°F. It is worth noting that in their experiments, they [6] placed the specimens in a weathering chamber, i.e., a controlled temperature environment, and in alkaline solutions. E ectively, an increase in chamber conditioning temperature (T) led to an ampli cation in the actual age.
us, the conditioning temperature has been altered within and among di erent studies. In general, they have ranged between 20°C and 90°C, with the exceptions of [25,32], which investigated the performance of GFRP bars subject to freezing and elevated temperatures. Based on Equation (3), such conditioning temperatures (20°C-90°C) could accelerate the aging by at least four times when specimens are conditioned in alkaline solutions. For instance, alkaline conditioning for 30 days in a chamber at 50°C is equivalent to an actual age of 2660 days (approx. 7 years) at Morgantown, WV. Yet, this equation may not be applicable to specimens encased in concrete or conditioned in regimes other than alkaline solutions. Figure 6(c) highlights the distribution of conditioning temperatures, with 77% of studies utilizing conditioning temperatures above 50°C. Nevertheless, it should be noted that such temperatures may a ect the physical properties of the bars by increasing the coe cient of thermal expansion [7] and, thus,

3.4.
Loading. e degradation of GFRP reinforcing bars could be intensi ed in the presence of a sustained load during conditioning [53][54][55]. ACI 440.1R [2] limits the tensile stress in GFRP under service conditions to 14% of the ultimate tensile strength (UTS), considering the strength reduction factor for environmental exposure. Higher stress levels caused a shift in the degradation mechanism of GFRP from being a ected by the rate of di usion of alkaline solutions to being controlled by solution transport through resin cracks [56]. Nevertheless, it can be noted from Figure 6(d) that only 27% of investigated microstructure studies applied a sustained load to GFRP bars, with values ranging from 0 to 80% UTS. e objectives of these works, nonetheless, were di erent. While some researchers applied sustained loads within the limits speci ed by codes and standards [2,4,[57][58][59][60], others explored the possibility of increasing this allowable limit [11,12,28,[33][34][35]52]. Some of these latter studies concluded that the limits recommended by codes and standards were conservative for speci c reinforcing bars and conditions. In fact, elevated temperatures could accelerate the degradation induced by conditioning. As a result, the limits were found to be      x x 150 to 450 20 to 60 25 [12] x X Up to 1445 -20-65 [13] x 60 to 180 23 to 50 x [14] x x 60 to 240 23 to 50 x [15] x 100 to 660 40 x [23] x 41.67 to 208. 33 22 to 60 x [42] x 90 to 365 23 to 65 x [24] x 90 60 x [25] x 0.0208 to 0.1285 −100 to 325 x [26] x x 60 to 365 23 to 70 x [27] x 30 to 180 23 to 60 x [28] x 30 to 270 23 0-25 [29] x x x x 28 to 50 60 to 80 x [30] x x x 180 to 540 23 to 50 x [31] x 30 to 720 60 x [32] x 0.0417 to 0.125 100 to 300 x [33] x inappropriate and required revision through further research data.

Reduction
Factors. e e ects of service time, relative humidity, and temperature are incorporated into an environmental reduction factor for GFRP bars used as internal reinforcement in concrete structures. ACI 440.1R [2] relates the design (f * fu ) and guaranteed tensile strength (f fu ) using the proposed reduction factors (C E ). e relationship is shown as Eq. (4). e reduction factors for environmental conditions, shown in Table 6, ranged between 0.3 and 1.0. Lower values were associated with higher relative humidity, service time, and temperature, while a reduction factor of 1 was typically attributed to less severe conditioning [61]. Another load reduction factor was introduced by several researchers and codes to account for the presence of a sustained load [2,4,[58][59][60]. Its value ranged between 0.2 and 1.0, with more recent codes suggesting lower reduction factors. is is associated to the wide variations of results reported and somewhat concerning impact of accelerated aging, even though extensive research has been conducted in this area. e combined e ect of both factors has also been noted, as shown in the last column of Table 6, to vary between 0.13 and 0.50.

Microstructure Evaluation Methods
Microstructural characterization was employed to evaluate changes to GFRP reinforcing bars after conditioning and provide explanations to the degradation in mechanical properties. Of the many procedures available, TGA, SEM, FTIR, and DSC have been the most commonly used. Table 7 summarizes the types of microstructure tests conducted in each of the studies presented herein. While moisture uptake is not a typical microstructure evaluation method, it has been proven pivotal in understanding the causes of degradation in the properties of GFRP specimens [10]. Clearly, SEM has been the most utilized of all tests. In fact, it has been employed in 92% of the studies reviewed, while DSC, FTIR, and moisture uptake have been used in only 40-43%, as shown in Figure 7. is is mainly due to SEM being able to highlight and identify morphological changes due to conditioning, which are typically characterized by matrix degradation, interfacial deterioration, and/or ber etching. TGA, on the other hand, has been used in 5% of the studies reviewed herein. Its primary use was to determine the thermal degradation of the polymer matrix [12,25]. Nevertheless, it is worth noting that the majority of researchers have resorted to more than one microstructure test to con rm experimental results and ndings. It should also be noted that the results reported hereafter for conditioned samples are associated to the most severe conditioning scheme utilized in each study.

Moisture Uptake.
Moisture penetrates into a composite material by the ow of solution into the microgaps between polymer chains, capillary transport through ber/matrix interfacial microcracks, and in microcracks formed during manufacturing [5]. e di usion process mainly depends on the type of matrix and surrounding environment. It can be measured as per the procedure of ASTM D570 [20]. Typical moisture absorption/uptake tests involve weighing samples before and after immersion. However, the hydrolysis reaction that may have occurred during conditioning could cause mass dissolution and, thus, a ect the results. To account for this mass loss, samples are oven-dried for 24 hours at 100°C. is oven-dried mass would then be used to nd the moisture uptake: moisture uptake(%, by mass) (conditioned mass − oven dried mass) (initial mass) × 100. Figure 8 is a typical moisture uptake curve that presents the increase in mass over conditioning time. Clearly, the di usion rate of moisture is di erent for temperatures below   Advances in Materials Science and Engineering and above 60°C. At temperatures of 60°C and below, the majority of the uptake occurs within the rst 30 days of immersion. At higher temperatures, the uptake increases at a steady rate up to 50 days. Robert, et al. [7] explained that such behavior is owed to the ampli cation of thermomechanical degradation mechanisms at a temperature of 80°C. Further, the moisture uptakes for conditioned and unconditioned GFRP reinforcing bars tested in the past literature are summarized in Table 8. ese values correspond to the most severe conditioning regime employed in each study. Results for conditioned samples ranged between 0.06 and 5.1, with vinylester-GFRP presenting the widest variations. Past researchers have discussed several factors that a ected the uptake, namely, presence of a sustained load, void content, pH of conditioning medium, conditioning duration, and conditioning temperature [8-11, 24, 30, 32, 35, 38, 47, 56, 64, 65]. Each factor will be discussed in the sections below. x [8] x x x x [9] x x x [10] x x x [11] x x x x [12] x x x [13] x [14] x x x [23] x x x [42] x x [24] x x [25] x x x [26] x x x [27] x x x [28] x [29] x [30] x [31] x x x x [32] x [33] x [34] x x x x [35] x x x [36] x [37] x [38] x x x [39] x x x [40] x x [44] x [45] x [46] x x [47] x [48] x x x [49] x [50] x [51] x

Advances in Materials Science and Engineering
e effect of sustained load can be noted by comparing the results of [10,11]. e only difference between these two researches is the addition of a sustained load of 25% UTS in [11]. Results, presented in Table 8, show that even with an average 26% shorter conditioning durations, the uptake was, on average, 68% higher. It is thus obvious that the addition of a sustained load can aggravate the moisture uptake.
Recent work by El-Hassan, et al. [10] compared two types of GFRP bars under similar exposure conditions. e void contents were measured as 0.10 and 0.23%. e authors reported higher moisture uptake in the latter GFRP bar with higher void content. However, with limited information on the void content or fiber content in the existing literature, it is difficult to create a general conclusion. Further work would be needed to investigate the effect of void content on performance retention and microstructure changes.
To simulate severe environments, researchers have utilized alkaline solutions with different pH. Values of pH ranged between 12 and 13.6. A recent study by D'Antino, et al. [66] reported that higher pH resulted in less tensile strength retention. Yet, immersion in simulated concrete pore (alkaline) solution is not a proper reproduction of reallife concrete. e main difference is the degree of contact of alkaline solution with the GFRP reinforcing bars. In the former case, the surface of the bar is entirely coated by the solution, resulting in maximum contact area, while in the latter case, alkaline solutions remain in the concrete pores, causing limited contact with the imbedded GFRP bar. In the case of moist concrete, representing underwater concrete, the degree of contact is intermediary between the dry concrete and wet alkaline solution.
Conditioning temperature and duration have been widely investigated. Figure 8 shows the effect of each of these conditions on the moisture uptake. It is clear that higher temperatures and longer exposure durations resulted in more uptake. Yet, it is worth noting that the former was the more influential factor of the two, thus, causing a more exacerbated degradation phenomenon. Robert, et al. [7] concluded that a thermomechanical degradation mechanism would occur at an elevated temperature of 80°C, owing to higher solution diffusion in a more porous resin matrix. Nevertheless, such conditioning temperature does not correspond to service life conditions of GFRP bars, justifying the maximum aging temperature recommended by ACI 440.3 [3] and CSA S806 [60] being 60°C.
ACI 440.6 [62] and CSA S807 [63] propose durability limits in the form of maximum allowable moisture/water uptake. ese limits are 1 and 0.75%, by mass, respectively, regardless of the exposure conditions. Table 8 shows the comparison of conditioned GFRP reinforcing bars with each limit. It is clear that more than half of the bars tested were not classified of high durability, based on moisture uptake. In fact, Figure 9 shows that 56 and 62% of bars, tested under the worst conditions in each study, failed to fall within the high durability limits of ACI 440.6 [62] and CSA S807 [63], respectively. In general, samples conditioned in alkaline solution at temperatures beyond 50°C were deemed unacceptable [8,23,27,39,46]. GFRP bars imbedded in concrete resulted in uptakes within the acceptable range, except for El-Hassan, et al. [11], where a sustained load of 25% UTS was applied, which is higher than the maximum allowable load recommended by ACI 440.1R [2]. A general comparison between alkaline and concrete conditioned GFRPs shows higher uptake in the former case and lower degree of satisfactoriness by the codes. erefore, it should be noted that the moisture uptake of GFRP bars fell within the acceptable limits if conditioned in concrete and for shorter durations, lower temperatures, and/or without a sustained load. Nevertheless, in real-life scenarios, where GFRP bars are imbedded in concrete, it is possible to limit the moisture uptake by creating an impermeable concrete.
is can be achieved by adding supplementary cementitious materials, as fly ash, ground granulated blast furnace slag, silica fume, or others.

TG Analysis.
TG analysis was employed in the past literature to assess the decomposition of the polymer matrix under heat exposure. ASTM E1868 [19] was used to examine the change in mass as a function of temperature (20°C-800°C) and evaluate the thermal stability of the matrix. A typical mass variation curve as a function of temperature measured by TGA is shown in Figure 10. It is clear that a major mass loss of 18% occurred in the range of 300°C-450°C. is is attributed to the thermal degradation of the polymer matrix, as the molecular chains broke, leading to microcrack formation at the ber/matrix interface and matrix itself. Such mass loss also indicates a critical loss of mechanical properties due to an irreversible degradation mechanism.

FTIR Analysis.
A number of past studies explored the degradation mechanism of GFRP reinforcing bars using FTIR analysis. Samples were placed in an FTIR spectrometer and analysed from 400 to 4000 cm −1 in transmittance or absorbance mode. Figure 11 shows typical FTIR spectra for conditioned and unconditioned samples. Five zones could be identi ed: 900-1200, 1400-1600, 1600-1800, 2800-3100, and 3200-3600 cm −1 . ese spectral zones are, respectively, attributed to O-H bending, C-O stretching, C O stretching, C-H stretching, and O-H stretching. e past literature studies have identi ed and reported these peaks, as shown in Table 9. It is clear that most conducted research has focused on the O-H and C-H stretching bonds. is is owed to the ability to evaluate the hydrolysis reaction by measuring the peak in the OH band, which increases due to hydrolysis and relating it to the constant peak of CH. As such, the relative quantity of hydroxyl groups is characterized by the ratio of maximum peaks of OH-to-CH [13,14,27,34]. Table 10 presents the OH-to-CH ratio of conditioned and control specimens of numerous studies. It should be noted that the reported ratio for conditioned samples is a result of the most severe conditioning scheme utilized in each study. is is a typical approach adopted by several researchers [8,9,13,14,24,26,31,34]. e ratio for control GFRP bars ranged between 0.21 and 2.60, while that of conditioned counterparts was between 0.25 and 14.30. With such wide variations in the degradation, i.e., OH-to-CH ratio, the percentage increase was calculated and reported in Table 10. It ranged between 0 and 450%. us, to provide a more conclusive interpretation of the results, a more isolated investigation is performed. GFRP bars made with epoxy and vinylester are separated and each correlated to its associated moisture uptake. Results of Figure 12(a) show that the higher the moisture uptake, the higher the OH/CH increase due to conditioning of epoxy-GFRP bars. ese test samples were all conditioned at 60°C, but for di erent durations. In fact, the left marker (0.74% moisture uptake increase) was conditioned in alkaline solution for up to 720 days, but did not show any increase in OH/CH ratio, indicating no degradation due to hydrolysis reaction [31]. is shows that 43 [8] 3200-3650 -2800-3000 -- [9] 3300-3600 1600-1800 2800-3100 1400-1600 900-1100 [10] 3200-3600 -2800-3000 -- [11] 3200-3600 -2800-3000 -- [13] 3300-3600 -2800-3100 -- [14] 3300-3600 -2800-3000 -- [23] 3200-3650 -2800-3000 -- [24] 3300-3600 1600-1800 2800-3100 1400-1600 900-1200 [26] 3200-3500 -2800-3000 -- [27] 3300-3600 -2850-3100 -- [31] 3400 -3026 -- [34] 3200-3400 -2800-3100 -- [35] 3430 -2900 -- [38] 3430 -2900 -- [39] 3540 -2900 -- [40] 3334 1557 2900 1420 1019  the degradation only initiated after 1% moisture uptake. Further increase in uptake up to 2% did not cause signi cant impact on the change in OH/CH ratio, after which it dramatically increased up to 115%. us, it can be concluded that epoxy-GFRP bars could experience limited degradation if the moisture uptake is maintained below 2%, by mass. Figure 12(b) presents a relation between OH/CH percent increase and moisture uptake in conditioned GFRP bars made with vinylester matrix. e OH/CH increase remained below 20%, indicating a lower degree of hydrolysis reaction, when the moisture uptake was below 1.5%. Beyond 1.5%, degradation was much more signi cant due to conditioning, with OH/CH increasing by up to 95% compared to the control samples. In comparison and based on the studies revised in this paper, GFRP bars made with vinylester resin have been found to be more generally susceptible to hydrolysis than their epoxy counterparts. e e ect of the initial tensile strength of unconditioned GFRP bars on their resistance to degradation was investigated in [10,11]. In these studies, two di erent types of GFRP bars with dissimilar tensile strengths were compared. FTIR analysis results of [10], presented in Table 10, show that the GFRP bar with lower void content had an increase in OH/CH ratio of 15.9% compared to 94.9% for the sample with the higher void content. Obviously, as the void content increased, the hydrolysis reaction progressed, leading to higher OH/CH ratios. e degradation was further intensi ed upon the application of a sustained load, as shown in the results of [11].

DSC Analysis.
e thermal behavior and characteristics of conditioned and control GFRP samples is analysed using DSC.
e main properties obtained from DSC are glass transition temperature (T g ), curing process, melting temperature, crystallinity, relaxation, and thermal stability [25]. Of these properties, glass transition temperature is the most reported in the past literatures. It is determined by performing two scans, each from 20°C to 250°C, in accordance with ASTM E1356 [21]. e rst scan compares the T g of the conditioned and control samples. A reduction in T g is indicative of matrix plasticization. e second scan identi es the degradation mechanism, whereby a similar T g for conditioned and control samples indicates a reversible plasticizing e ect, while a lower T g for the former signi es an irreversible chemical degradation. Figure 13 illustrates typical DSC curves for alkaline-conditioned and unconditioned GFRP reinforcing bars. e step shown in the range of 90°C-110°C is used to nd the T g . While it appeared to be an endothermic or melting peak, there was no melting involved in this work [31]. In fact, the authors associated this peak to "the thermal stress relaxation phenomenon of polymer chains of the GFRP rebar during long-term ageing".
Furthermore, a summary of T g temperatures and cure ratio of various studies is presented in Table 11. ese temperatures are results of the GFRP samples subject to the most severe conditioning scheme utilized in each study. It is clear that most studies experienced some decrease in the T g of the rst scan, ranging from 0 to 46%. is wide variation in results is owed to the di erent curing regimes, ber content, matrix type, and presence of sustained loading, which all contribute to lowering the T g and the deterioration of the GFRP bars. However, it is possible to isolate some values by selecting GFRP bars made with epoxy matrix and correlating them to other parameters. Figure 14 shows that higher changes in OH/CH ratios led to more intense reductions in the T g . is shows a clear correlation between results of FTIR and DSC, signifying the progression of a hydrolysis reaction. In other terms, the factors that a ect the OH/CH ratios have a similar e ect on the T g .
In the second scan, studies that employed concrete as the conditioning medium did not note any change in the T g between conditioned and control samples [10,11,14,34,38]. On the other hand, specimens conditioned in alkaline solutions and water showed some signs of irreversible chemical degradation, with the former conditioning scheme being more detrimental [25,27,42]. is provides evidence to the severity of conditioning in alkaline solutions, rendering results somewhat unrealistic and unreliable.
Additionally, the cure ratio is shown in Table 11. It is the ratio of the T g of the rst to second scan for the control samples. Other than [13,14,34], which are conducted by the same authors and probably using the same GFRP bars, the reported cure ratios were above 95%. Lower values signify that the samples were not fully cured and that postcuring occurred during the second scan.

SEM Analysis.
SEM was employed to evaluate the microstructure and morphological changes to the GFRP reinforcing bars after conditioning. e ber-matrix interface is considered one of the most vulnerable areas in GFRP reinforcing bars. It is merely a few microns thick and plays a critical role in transfer of the load between ber and  matrix [67]. In the presence of moisture and alkalis, the bond between ber and matrix is weakened, causing damage to the interface itself and the GFRP as a whole. Micrographs typically highlight di erent damages or deteriorations that may be induced by conditioning, including matrix degradation, ber/material interface degradation, microcrack formation, and ber etching and leaching. Table 12 shows a summary of deteriorations that have been identi ed in di erent studies using SEM. Clearly, matrix degradation, ber/matrix interfacial degradation, and microcrack formation are interlinked, as they are reported simultaneously in multiple studies; they are denoted collectively hereafter as damage A. Fiber etching and leaching, on the other hand, is designated as damage B. Of the literature reviewed in this work, 58, 15, and 27% reported damage A, damage B, and no damage, respectively, as shown in Figure 15. is demonstrates that the matrix and the ber/ matrix interface are indeed the most vulnerable components of GFRP reinforcing bars, with much lesser deterioration of the ber. Figure 16(a) shows a micrograph from past work that identi ed damage A in a conditioned GFRP reinforcing bar [10]. Samples in this study were placed in seawatercontaminated concrete for 15 months at 60°C. Clearly, the ber/matrix interface weakened to the extent that ber delamination occurred. is is primarily owed to the progression of hydrolysis reaction.
Fiber etching and leaching were only noticed in research work that examined the durability performance of GFRP reinforcing bars made with E-glass/vinylester and conditioned in alkaline environment; they are represented as damage B. Figure 16(b) presents a micrograph of a GFRP bar that su ered damage B [28] with circumferential damage to the ber. In this work, GFRP samples were conditioned in an alkaline solution under a 25% UTS sustained load. e severity of the testing conditions resulted in all samples failing within 20 days of conditioning. No tensile strength could be retained at that point. e degree of deterioration was too harsh that it caused matrix degradation, ber/matrix interfacial debonding, microcrack formation, and ber damage simultaneously.
In some of the reviewed studies, no damage or deterioration was noticed. ough some degradation may have been reported in the microstructure study, the mechanical properties of GFRP reinforcing bars were not signi cantly a ected. It is thus critical for researchers to correlate the tensile strength, moisture uptake, OH/CH ratio, and SEM micrographs in a distinct study to provide a better understanding of the e ect of the microstructure on the mechanical properties. Yet, while the scope of this work is to focus on microstructure characteristics of GFRP bars exposed to di erent conditioning regimes, a correlation based on the reviewed studies has been developed. For this matter, results from [10,11,13,14,27,35] are employed. It is worth noting the GFRP samples were conditioned di erently in these studies. e SEM micrographs of Figures 17(a)-17(f ) show little to no cracks at the ber/matrix interface, indicating limited degradation. e GFRPs of [10,11,27] are    [11,35] were conditioned under sustained loads. Accordingly, it is possible to accept an increase in moisture uptake and OH/CH ratio up to a threshold of 1.6 and 18%, respectively, while anticipating up to 15% reduction in tensile strength.

Conclusions and Remarks
In this paper, studies examining the microstructure of GFRP reinforcing bars exposed to di erent environmental conditions and sustained loading were reviewed. Experimental results and ndings were collected from past literature, analysed, and compared to provide conclusive remarks on the e ect of conditioning on the microstructure of GFRP reinforcing bars. Based on the conducted analysis, the following conclusions can be drawn: (i) Mechanical properties from the past literature were used to provide correlation equations for unconditioned GFRP bars. Tensile strength and modulus of as-received GFRP bars were correlated in a linear equation. ese relationships could  [8] x x x [9] No deterioration reported [10] x x x [11] x x x [12] x x x [13] No deterioration reported [14] No deterioration reported [15] x x x x [23] x x x [42] x x [24] No deterioration reported [25] x x x [26] No deterioration reported [27] x x x [28] x x x x [29] x x x [30] x x x [31] No deterioration reported [32] x x x [33] x x x [34] No deterioration reported [35] x x x [36] x x x x [37] x x x [38] No deterioration reported [39] x x x [44] x x x [45] x x x [46] x x x x [47] x x x [49] x x x [50] x x x x [51] No deterioration reported    [11]. (c) Adapted from [13]. (d) Adapted from [14]. (e) Adapted from [27]. (f ) Adapted from [35].
provide reasonably accurate prediction of the tensile modulus from the tensile strength. (ii) Only 14% of reviewed studies investigated the performance and microstructure of GFRP bars made with E-glass and epoxy. On average, the tensile properties of such unconditioned bars seemed to be 14% superior to those of vinylester counterparts. Discrepancies in the testing procedures did not allow for a direct microstructure comparison between the two types. Further testing is needed prior to the adoption of these GFRP reinforcing bars. (iii) In the majority of reviewed studies, GFRP specimens were conditioned in alkaline and water solutions due to the ease of the experimental setup. Concrete was only employed in 23% of the studies. Conditioning durations varied greatly, with 0-90 and more than 365 days being most commonly used. (iv) e application of a sustained load to GFRP reinforcing bars during conditioning was adopted in 28% of the studies reviewed. Research studies that exceeded the limits stated by codes and standards concluded that the set limits were only valid for specific nonaggressive environment conditions. In fact, elevated temperatures promoted the degradation mechanism. More research is needed to update these limits to include different conditioning regimes. (v) Higher temperature, prolonged conditioning, presence of a sustained load, and conditioning in alkaline solution caused an increase in moisture uptake. Most uptake values extracted from the literature exceeded the limits set by ACI 440.6 [62] and CSA S807 [63]. Only specimens that were conditioned in concrete at low temperatures, for short durations, and/or without a sustained load were within the acceptable moisture uptake durability limits. is raises a concern to the aggressiveness of the conditioning utilized in the literature. (vi) A correlation was developed between the progression of hydrolysis (OH/CH increase) and the decrease in T g (%). Within an increase of 20% OH/ CH, the T g decreased by only 10%. However, greater surges in OH/CH caused much more dramatic reductions in the glass transition temperature. Nevertheless, limiting the moisture uptake could control the increase in hydroxyl groups. (vii) Conditioning in alkaline solution was found to be too severe and did not accurately represent real-life scenarios. Specimens conditioned in alkaline solutions experienced irreversible chemical degradation, more hydroxyl group formation, and more intense degradation to the microstructure. (viii) Four types of damages were reported in conditioned GFRP bars: matrix degradation, fiber degradation, fiber/matrix interface degradation, and microcrack formation. Most of concrete-encased GFRP specimens did not experience excessive degradation, as opposed to alkaline conditioning that severely damaged the GFRP. (ix) e increase in moisture uptake and OH/CH ratio and formation of microcracks does not necessarily have a severe effect on the mechanical performance. In fact, it is possible to allow an increase in moisture uptake and OH/CH ratio up to 1.6 and 18%, respectively, while anticipating up to 15% reductions in tensile strength.
e conclusions presented above were based on the past literature reviewed in this work. Further studies are needed to fill the research gaps before GFRP reinforcing bars could be widely adopted by the construction industry. e following are suggestions for future research direction: (i) A standard testing procedure for GFRP reinforcing bars capable of providing consistent and reliable results that represents real-life scenarios (ii) Comparative performance and microstructure evaluation of GFRP reinforcing bars in concrete and subjected to different environment conditions and levels of sustained load (iii) Microstructure characterization of GFRP bars exposed to acidic solution (iv) e effect of void content on performance retention and microstructure changes of GFRP reinforcing bars (v) Performance and microstructure comparison between GFRP bars made with different resin, i.e., epoxy and vinylester, subject to the same conditioning schemes (vi) Performance and microstructure comparison between concrete-encased GFRP specimens exposed to elevated temperatures with an applied sustained load and actual in-service specimens (vii) GFRP reinforcing bars in concretes made with different supplementary cementitious materials to limit the moisture uptake

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this paper.