Long-term structural behaviour of marine-conditioned geopolymer concrete cast with geopolymer-coated recycled concrete aggregates and basalt fibres Open Access

Geopolymer concrete (GPC), a cement-free alternative, is an emerging construction material replacing Ordinary Portland Cement (OPC) concrete (Davidovits, 1991). By-product materials such as fly ash (Naghizadeh et al., 2023; Yang et al., 2023) and slag (Fan et al., 2023; Garg et al., 2024) serve as binder material in GPC when used with an alkali activator, eliminating traditional cement’s need. A chemical reaction between aluminosilicate material and alkali activator forms the GPC (Ahmad Zaidi et al., 2021; Shehata et al., 2022), resulting in comparable performance with OPC concrete with the advantage of a low carbon footprint. The role of low-carbon concrete in decarbonising the construction industry is crucial (Chen et al., 2023), as concrete is the second most used material globally after water, with approximately 30 billion tonnes consumed annually (Sousa et al., 2023). This is due to cement production for concrete, accounting for 8% of the global CO₂ emissions (Ellis et al., 2020). Producing each kilogram of cement requires 1.5 kg of raw materials and generates 0.8 kg of carbon dioxide (Chen et al., 2010). The construction industry can significantly reduce its environmental impact by replacing OPC concrete with GPC, as it is responsible for 39% of energy-related emissions worldwide (Chen et al., 2021). This change contributes to achieving a net-zero future.

This research investigated the performance of self-compacting geopolymer concrete (SCGC), incorporating recycled concrete aggregates (RCA) as coarse aggregates. Aggregates constitute 60–75%of the volume of the concrete (Wang et al., 2021). Replacing them with RCA offers the significant benefit of conserving natural resources while reusing crushed concrete waste (Knoeri et al., 2013; Nikmehr and Al-Ameri, 2022). The share of crushed concrete waste is 50% of construction and demolition waste and it is the most significant type of such waste (Tam, 2008). Despite these merits, many researchers have reported the adverse impact of using RCA as aggregates in GPC, such as decreased compressive strength (Srinivas and Sukesh Reddy, 2019; Ayub et al., 2021), reduced splitting tensile strength (Al-Jaberi et al., 2021; Ayub et al., 2021; Lim and Pham, 2021; Waqas et al., 2021), lower modulus of elasticity (MoE) (Al-Jaberi et al., 2021; Lim and Pham, 2021; Kanagaraj et al., 2022).

These challenges necessitate exploring methods for modifying RCA to produce high-quality RCA-based concrete (Nikmehr et al., 2024c). Among these methods, Nikmehr et al. (2024a) coated the surface of RCA with geopolymer slurry for being incorporated in a SCGC. They observed a 28% enhancement in compressive strength, a 21.47% improvement in tensile strength, and a 51.9% increase in MoE. Also, Li et al. (2009) reported that geopolymer-coated RCA provides a denser interfacial transition zone (ITZ), enhancing the workability and compressive strength of concrete. In another study, Junak and Sicakova (2017a) used a coating mixture comprising 75% coal fly ash, 9% Na₂SiO₃, 11% 8M NaOH, and 5% water to form a 0.25 mm layer on RCA incorporated into OPC concrete. This treatment resulted in improved compressive strength after 28 days and one year compared to samples cast with uncoated RCA.

Incorporating basalt fibres (BF) is another treatment technique that has had an excellent impact on RCA-based concrete, as reported in the literature (Şahin et al., 2021; Vijaya Prasad et al., 2022; Nikmehr et al., 2024d). At the same time, it is a safe, simple, and energy-efficient treatment technique (Nikmehr et al., 2024c). BF is produced by crushing a volcanic rock called basalt (Vijaya Prasad et al., 2022; Wang et al., 2022), the most sustainable reinforcing fibre due to abundance of basalt rock worldwide, which makes up around 70% of the earth’s surface (Lopresto et al., 2011). The length and content of the fibre have an influential impact on the concrete’s properties, and it’s crucial to use the optimum amount of them to prevent the detrimental effect of excessive content on the initial and long-term performance of the concrete. For example, Heweidak et al. (2022) reported an excellent improvement in the fly ash and slag-based SCGC containing 2% of 30-mm and 12-mm hybrid length BF with the content fraction of 3 to 1. They reported increase of 20% and 62% in 28-day compressive and tensile strength respectively. For slag and fly ash-based SCGC containing only RCA, BF enhanced compressive strength, splitting tensile strength, and MoE by 17%, 47%, and 38%, respectively (Nikmehr et al., 2024b). BF also increased the ultimate load capacity of reinforced beams by 3%, contributing to more ductile behaviour and increased deflection at failure, demonstrating improved toughness characteristics (Nikmehr et al., 2024d). There is a scope for further study on BF containing concrete mixes’ durability and long-term properties. So, this study addresses the gap and examines the long-term properties of SCGC containing RCA and BF in the simulated harsh marine environment.

The marine environment has detrimental impacts on the durability of concrete structures in terms of the cycling periods of wetting and drying in tidal zones, salt weathering, and changing temperatures (Yi et al., 2020). Moreover, due to the GPC’s porous structure, increasing the concrete’s water absorption, further investigation on the behaviour of the marine-exposed GPC have been recommended by previous studies. Also, in comparison to natural aggregates, RCA increases concrete’s water absorption, as noted by the researchers who observed further penetrating water into the concrete samples (Ayub et al., 2021), higher sorptivity values (Bhardwaj and Kumar 2021), and increased water permeability (Adnan et al., 2008). This presents a unique challenge when studying the impact of chloride concentration and the corrosion of steel reinforcing bars within the concrete, particularly under marine conditions. The durability performance of GPC in harsh environments has not been sufficiently studied (Reddy et al., 2013), mainly when GPC contains RCA. Although several studies have explored the mechanical and structural properties of GPC (Farooq et al., 2021; Nikmehr et al., 2023), this cement-free concrete is still in its early stages of development. It has not thoroughly undergone extensive field testing. Long-term studies are needed to assess its mechanical, structural and durability properties, unlike traditional concrete which has been thoroughly researched and is the foundation for existing design codes (Amran et al., 2021).

This research aims to understand the long-term mechanical properties of newly developed GPC incorporating 100% geopolymer-coated RCA and BF, including compressive strength, splitting tensile strength, and MoE after exposure to marine conditions for 6 and 12 months. Additionally, the structural performance of RCA-incorporated SCGC is evaluated through three-point bending tests on reinforced beam specimens exposed to simulated cyclic drying and wetting in tidal zones for six and twelve months under sustained service loads. The results of the marine-exposed samples are compared to control specimens stored in a normal environment under sustained loading to assess the effects of marine exposure. Furthermore, the microstructural and chemical characteristics of the mixes will be analysed and discussed in this study.

Binder material used in this research includes grade 1 fly ash ^[1], micro fly ash ^[2], and Ground Granulated Blast-Furnace Slag (GGBS) ^[3], with the chemical composition tabulated in Table 1. Anhydrous sodium metasilicate ^[4], in solid form, was the type of activator. In this research, three mixes denoted as M0, M100-C1, and M100-C1BF were cast, and the mix proportions are tabulated in Table 2. Mix M0 was cast following the self-compacting mix design reported by Rahman and Al-Ameri (2021). All coarse aggregates in mixes M100-C1 and M100-C1BF were replaced with coated RCA ^[5]. The RCA used in mixes M100-C1 and M100-C1BF were coated with the geopolymer slurry using the mix proportions shown in Table 2. Mix M100-C1BF contains a 2% hybrid length of 12 and 30 mm BF ^[6].

The process of coating the RCA is depicted in Figure 1, following the method reported by Nikmehr et al. (2024a). The maximum size of the NA and coated RCA are 14 mm, and the gradation curves of the fine and coarse aggregates applied in this study are illustrated in Figure 2.

Self-compacting concrete was prepared by mixing all the solid materials for four minutes. Then, water was included gradually and mixed for more than 8 min, with a resting time of two minutes (Rahman and Al-Ameri, 2021). Regarding mixes M100-C1 and M100-C1BF, RCA was coated with geopolymer slurry one day earlier than casting these mixes, followed the method introduced by previous research by Nikmehr et al. (2024a). This treatment has been reported to enhance compressive strength by 28%, tensile strength by 21.47%, and MoE by 51.9%, indicating its effectiveness in improving the quality of RCA-based SCGC.

All mixes comply with the requirements of European Specification and Guidelines for Self-Compacting Concrete (EFNARC, 2002) and VicRoads specifications for Self-Compacting Concrete (VicRoads, 2006). As for the mechanical properties of the mixes, cylinder samples with a diameter of 100 mm and a height of 200 mm were cast based on AS 1012.8.1 (Australian Standard, 2014a). Moreover, reinforced beams of 100 × 100 × 500 mm were designed and cast according to AS 3600 (Australian Standard, 2018), as shown in Figure 3. Four longitudinal deformed steel rebars, each with a diameter of 10 mm and a yield strength of 500 MPa, were used to reinforce the beams.

Table 3 tabulated the number of samples were cast from each mix, as illustrated in Figure 4. All Samples were cured in the ambient condition, following the code of AS 1012.8: 1986 (Australian Standard, 2000a), which defines the ambient temperature for Victoria in Australia as 23 ± 2 °C and a relative humidity of 50%. For each mix, mechanical properties of the 28-day samples were tested based on the following codes: AS 1012.9: 2014 (Australian Standard, 2014b) for compressive strength, AS 1012.10:2000 (Australian Standard, 2000b) for splitting tensile strength, and AS 1012.17: 1997, Method 17: Determination of the static chord MoE and poison’s ratio of concrete specimens (Australian Standard, 1976) for MoE. The test setup is depicted in Figure 5. A minimum of triplicate specimens was tested for each of the mechanical performance test methods.

28-day ambient-cured beams were subjected to a three-point bending test to study the loads and deflection of the mid-span of the beams. The three-point bending test was conducted in accordance with the procedure specified in ASTM C78/C78 M:2022 (ASTM, 2022). To this aim, the loading rate of 33 N/s was introduced to the beams by using a universal flexural frame, as the test set-up illustrated in Figure 6a. The crack patterns during the loading were also analysed and the crack widths were measured using crack width gauge CWG-540 and a high-definition camera, as shown in Figure 6b.

The remaining beams and cylinder samples were conditioned in both ambient and accelerated marine environments to assess the mechanical and structural behaviour of the concrete mixes at six-month and one-year intervals. This timeline was selected as appropriate for long-term study in harsh environments, utilising purpose-built ageing tanks at Deakin University’s Concrete Lab, in alignment with conditions adopted in previous studies (Rahman and Al-Ameri, 2023), despite the absence of a standardised duration for assessing concrete durability in marine environments.

For the microstructural analysis of the hardened concrete, scanning electron microscopy (SEM) images were captured and examined for three mix samples after 28-day, 6 and 12 months of ambient and marine exposure. To prepare the samples, sections were cut from the interior of the specimens, polished with sandpaper, and dried in a 60 °C oven for one week. The dried samples were then mounted on SEM specimen stubs. After painting around the samples with graphite, the sample surfaces were coated with platinum using a Leica ACE600 (see Figure 7 (a)). SEM images were then captured with a JEOL JSM-IT300 microscope (see Figure 7 (b)) for detailed microstructure analysis. Energy-Dispersive X-ray (EDX) analysis was performed with a JEOL JSM-IT300 microscope (see Figure 7 (b)) to examine the chemical composition of M0, M100-C1, and M100-C1BF samples until one year of exposure to ambient conditions and a simulated marine environment.

All beam specimens (tested under ambient curing and accelerated marine exposure) were subjected to sustained loading, simulating the service load using the frame shown in Figure 8. The marine stainless-steel grade 316 was used for the frame, rods, bolts and nuts to prevent corrosion in seawater. According to AS/NZS 1170.0:2002 (Australian Standard, 2002), the long-term service factor equals 0.4–0.6. For simulating the service load in this study, around 50% of the ultimate load capacity in the three-point bending test was applied to the reinforced beams cast from each mix, as tabulated in Table 4. The nuts on the rods were tightened to introduce the loads to the beams, and a digital torque wrench was used to apply the required load. The digital torque wrench was calibrated using dummy hollow steel shapes as beams and a load cell to define the relationship between the applied load and the required torque, as shown in Figure 9. As previously recommended by other researchers, the torque wrench was used periodically to maintain consistent loads throughout the beams’ conditioning period (Rahman and Al-Ameri, 2023; Wei et al., 2023).

To study the long-term behaviour of concrete in harsh environments, the impact of three parameters of wetting and drying cycles, high temperature and high chloride content have been examined in the literature (Rahman and Al-Ameri, 2023; Wei et al., 2023). The simulated seawater was set up to the specific gravity of 1,020 kg/m³ by adding salt (NaCl). As illustrated in Figure 10, the cycles of 6 h for wetting and drying were set up with the water temperature at 50 °C for the wetting cycle and around 40 °C for drying cycle, as previously reported for the experimental studies on concrete exposed to harsh environments (Rahman and Al-Ameri, 2023; Wei et al., 2023). The elevated temperature is recommended for the better observation of the potential physio-chemical degradation phenomena in marine environments (Le Gué et al., 2024). All samples were fully immersed in the simulated seawater during the wet cycles. During the drying cycles, the samples were subjected to dry air conditions. The images of the samples under sustained load and two types of conditioning are illustrated in  Appendix.

The reinforced beams were exposed to a standard and simulated marine environment for 6 and 12 months while coupled with sustained loading. A three-point bending test was conducted on the beams to evaluate their residual strength. Besides, the cylinder specimens were tested to explore the compressive strength, splitting tensile strength, and MoE during 6 and 12 months while exposed to marine and normal conditions.

The mass of the cylinder samples of each mix under the two exposing environments, ambient and marine, was measured to evaluate the changes in mass and density after 6 and 12 months. As depicted in Figure 11, density reduction was observed for all three mixes, M0, M100-C1, and M100-C1BF, after six months of ambient conditioning. The density of all mixes during the one year and two environmental conditions of seawater and ambient is between 2070 and 2,261 kg/m³, less than the density of OPC concrete of around 2,400 kg/m³ (Australian Standard, 2018). The previous studies confirmed the lower density of GPC than OPC (Gunasekaraa et al., 2016).

RCA incorporation slightly reduced the density of 28-day, 6 and 12 months ambient-conditioned samples compared to their counterparts taken from mix M0 that contained natural aggregates. Reducing density by incorporating RCA has been reported in previous studies (Kanagaraj et al., 2022; Verma et al., 2022). The lower density of RCA-based concrete may be due to voids and cracks in the old paste adhered to the RCA. Additionally, the attached mortar on RCA results in a higher mortar-to-aggregate ratio compared to concrete with natural aggregates, as the mortar has greater porosity and lower density than NA (Wang et al., 2021; Nikmehr et al., 2024a). However, there is no difference between the density of M0 and M100-C1 mix samples after 6 and 12 months of marine conditioning.

A slight density reduction can be observed for the mixes conditioned in ambient for one year by 4.34%, 3.76%, and 3.90% for mixes M0, M100-C1, and M100-C1BF respectively. The cause of density reduction in the long term is the gradual development of geopolymerisation as water dissipates (Nath and Sarker, 2017). However, the density of the marine-conditioned samples increased slightly by approximately 2.08%, 3.57%, and 2.37% for M0, M100-C1, and M100-C1BF, respectively. This increase may be attributed to the ongoing geopolymerisation process, which was facilitated by the continuous saturation in water that prevented evaporation and supported further development of geopolymerisation.

Figure 12 demonstrates the long-term compressive strength of three different mixes during the one-year conditioning in seawater and ambient. Mix M100-C1 with coated RCA has higher compressive strength than M0 samples with natural aggregate after 28 days, 6 and 12 months of both ambient and marine conditioning. Previous studies have reported the positive impact of geopolymer-coated RCA on increasing compressive strength in OPC concrete (Gupta, 2012; Junak and Sicakova, 2017b). This is attributed to the geopolymer layer’s role in enhancing the ITZ between the geopolymer paste and RCA.

Mixes M0, M100-C1, and M100-C1BF show improved compressive strength by 44%, 48%, and 39% after one-year exposure in the simulated tidal zone. This finding is consistent with the research by Rahman and Al-Ameri (2022) that reported a 31% increase in the compressive strength of SCGC with NA after four months of marine exposure. The elevated temperature of 50°C and higher humidity levels aid in geopolymerisation, leading to further strength development (Rahman and Al-Ameri, 2022). Moreover, the increases in compressive strength of GPC in marine conditions may be attributed to the presence of Na⁺ in seawater, as strength development in hardened GPC is related to alumina silicate (N-A-S-H) gel network (Megahed et al., 2019). It has also been reported that the activation of Class F fly ash with low calcium content, as specified in American Concrete Institute (ACI) C 618 (Rahman and Al-Ameri, 2021), similar to the fly ash used in this study, plays a crucial role in improving chloride penetration resistance during the geopolymerisation process (Kupwade-Patil and Allouche Erez, 2013). Additionally, slag contributes to forming a denser Calcium-Alumina-Silicate-Hydrate (C-A-S-H) gel, which enhances chloride transport resistance and durability, even outperforming OPC concrete in this regard (Albitar et al., 2017).

M100-C1BF obtained the highest strength at all stages while exposed in marine conditions for one year by 39%, from 34.5 MPa for 28-day samples to 47.88 MPa. As the fibres are pulled out and break, energy is absorbed to overcome the frictional stresses required for crack propagation, which helps enhance compressive strength (Guo et al., 2019). For the samples conditioned in ambient, the increase in the compressive strength was also observed, i.e. 5.45 MPa (24%), 3 MPa (12%), and 3.62 MPa (10%) for M0, M100-C1, and M100-C1BF respectively. This observation is aligned with the literature that the strength development of GPC continued with time (Wardhono et al., 2017).

The splitting tensile strength of three concrete mixes was measured after 6 and 12 months of ambient and marine conditioning, and the results were compared to their 28-day tensile strength, as shown in Figure 13. The use of RCA lowered the 28-day splitting tensile strength of mix M100-C1 by 14% compared to the control mix with natural aggregates. This reduction in tensile strength in RCA-based concrete is attributed to weaknesses at the interfaces between the RCA and the residual mortar, as well as between the RCA and the new geopolymer paste (Zhang et al., 2021). Similar reductions in indirect tensile strength from RCA incorporation have been reported in previous studies (Zhang et al., 2021).

The tensile strength of the mix M0 containing sole NA increased after 12 months of marine conditioning by 43% and reached 3.65 MPa, while there was no significant increase in the tensile strength of the samples in the ambient chamber for one year. Splitting tensile strength of the mix M100-C1 increased by 15.4% and 25.4% after six months of ambient and marine conditioning. An additional six months of ambient conditioning further raised the tensile strength of M100-C1 samples by 8.7%, while a slight decrease of 4.3% was observed in the marine-conditioned samples.

The tensile strength of the mix M100-C1BF decreased during one year when both marine and ambient conditions were conditioned. However, the tensile strength of this mix is still higher than that of plain samples cast with coated RCA without using fibres. This is because BF’ bridging action helps slow crack propagation while limiting the opening and expansion of macro-cracks in the concrete (Branston et al., 2016; Guo et al., 2019). After 12 months of conditioning, the M100-C1BF samples in both marine and ambient environments showed decreases of 8.2% and 11.1%, respectively, compared to the 28 day samples. Jiang et al. (2014) reported that the positive effects of BF decrease significantly after 90 days. This might be due to some degradation in the chopped fibres and loss of bonding between fibres and paste (Branston et al., 2016). The degradation of BF due to seawater has also been reported by Le Gué et al. (2024).

The development of the MoE for M0, M100-C1, and M100-C1BF concretes from 28 to 360 days for both marine and ambient conditioning is presented in Figure 14. This property represents the ratio between a specific applied stress range and the corresponding strain within the elastic limit (Wardhono et al., 2017). Over the 28-to-360-days period, the MoE for M0, M100-C1, and M100-C1BF concretes ranged from 8.35 to 24.42 GPa, 8.2–22.85 GPa, and 9.5–19.97 GPa respectively. M0 concrete reached a higher modulus of 15.9 GPa at 28 days, compared to 14.2 GPa of M100-C1 concrete. Reducing MoE by incorporating RCA promotes the bonding of residual paste to the RCA, leading to a closer match in deformation behaviour between the RCA and geopolymer paste, thereby reducing the risk of fractures (Liu et al., 2019). Previous studies have documented a lower MoE in concrete made with RCA (Al-Jaberi et al., 2021; Lim and Pham, 2021; Kanagaraj et al., 2022). The MoE for M100-C1BF was 13.53 GPa at 28 days, as adding Basalt fibre slightly reduces the MoE, confirmed by the available literature (Kowalik and Ubysz, 2021).

Marine conditioning also enhances the MoE in all three mixes. After six months of exposure, the MoE increased by 54%, 61%, and 48% for mixes M0, M100-C1, and M100-C1BF, respectively. This increase may be attributed to the elevated temperatures in the marine conditioning environment, which accelerates the hydration process of unreacted aluminosilicate materials. However, after an additional six months of marine conditioning, the MoE decreased. Despite this, marine conditioning still resulted in higher stiffness for the mixes compared to normal conditioning after one year. The reduction in MoE over time for alkali-activated GPC has also been reported by Wardhono et al. (2017).

Load and mid-span deflections for beam samples cast from three mixes of M0, M100-C1, and M100-C1BF are depicted in Figure 15. A steep slope is observed in the elastic stage for beams exposed to seawater. This is due to the higher strength of the mixes after one year of sea conditioning. However, after the elastic stage, a further gradual incline with various tendencies was observed, attributed to the multiple materials used to cast each mix.

As tabulated in Table 5, the retained ultimate load capacity of the beam M100-C1 after six months of marine conditioning increased by 23% compared to the 28-day M100-C1 beam and increased by 35% after one year of marine exposure. This increase aligns with the rise of the tensile strength of the cylinder samples. However, the retained load capacity of the beam M100-C1 after six months of ambient is less than that of the marine-exposed samples, and a drop of 11% was observed after one year of ambient conditions while it was under the sustained load. A similar trend was also observed for the beams containing natural aggregates, so replacing coated RCA as the coarse aggregates had no adverse impact on the load capacity of the reinforced beams. However, Bhoopesh and Jithin (2017) reported that untreated RCA used as coarse aggregates reduced the ultimate load capacity of reinforced beams by 9% compared to those made with only natural coarse aggregates.

Inclusion BF enhanced the ultimate load capacity after six months of ambient and marine exposure, which has a similar trend as the tensile strength, which is enhanced by incorporating BF, confirmed by the literature (Al-Rousan et al., 2023). However, a drop of 5.6% can be seen after 12 months of marine conditioning for beam M100-C1BF compared to samples after six months. This might be due to some degradation in the BF and loss of bonding between fibres and paste (Branston et al., 2016). Moreover, the role of the basalt fibre in the increase of porosity and adverse effect on the chloride penetration resistance of the concrete (Guo et al., 2019) leads to the ingress of further chloride of the seawater into the conditioned samples in marine weathering tanks.

As shown in Figure 15, all three mixes exhibited a significant reduction in deflection at failure load after one year under sustained loading, both in marine and ambient conditioning. This resulted in a smaller area beneath the load-deflection curves, corresponding to toughness (Low and Beaudoin, 1994; Noushini et al., 2019). The deflection before reaching the ultimate load (peak) is more reflective of the beam’s rigidity, strength, and micro-crack formation, while ductility is derived from the post-peak deflection (Low and Beaudoin, 1994; Noushini et al., 2019). The brittle behaviour of the beams after one year of sustained loading is attributed to microcracking within the material, which weakens its internal structure. Consequently, the material becomes more brittle and loses its ability to undergo significant deformation before failure, thereby reducing its ductility that has been confirmed by current literature (Dong et al., 2018).

Incorporating coated RCA led to reduced toughness and a more brittle response than natural aggregates across 28-day and 12-month conditioning periods. The mix M100-C1BF, which includes BF, showed a slight improvement in toughness, consistent with previous research highlighting the benefits of BF in enhancing the toughness of reinforced beams (Noushini et al., 2019). However, after 12 months, M100-C1BF beams displayed brittle behaviour, potentially due to the fracture of the BF at the point of failure (Branston et al., 2016).

Figure 16 to Figure 21 display the crack patterns and failure mechanisms after a three-point bending test conducted on all beam samples after 6 and 12 months of ambient and marine conditioning compared to the 28-day beam specimens. The samples underwent shear failure caused by the short beam span and the concentrated support reactions. This led to shear failure as primary diagonal cracks propagated through the shear zone.

For the 28-day samples, during the initial loading phase between 1 kN and 3 kN, the initial flexural cracks appeared at the mid-span beneath the loading point. As the load increased, shear cracks formed within the shear zone, extending diagonally toward the centre of the beams due to the shear stress in that region. The cracks widened and deepened at this stage, resulting in shear failure. Beams M0 and M100-C1, which lacked fibres, exhibited brittle failure. In contrast, as shown in Figure 16, BF significantly reduced crack formation during the bending tests by enhancing load transfer, leading to slower crack development and more efficient stress redistribution.

Despite being under sustained load for 6 and 12 months, cracks decreased over time, particularly in samples conditioned in the marine environment. For instance, only two diagonal cracks were observed in the 12-month marine-conditioned M100-C1 sample (Figure 21(n)), compared to six diagonal cracks in the 28-day M100-C1 sample (Figure 21(b)).

Figure 22 to Figure 26 illustrate the width of initial cracks and cracks after failure in samples from three different mixes after 28 days of ambient curing, followed by 6 and 12 months of conditioning in ambient and marine environments.

Regarding 28-day samples (refer to Figure 22), during the three-point bending test on beam M0, cracks initiated at 0.08 mm under initial loads and expanded to 1.4 mm at failure. In contrast, the cracks in beam M100-C1 widened to 1.95 mm after failure. However, the cracks in beam M100-C1BF were narrower compared to M100-C1 due to the role of fibres in controlling crack propagation.

Regarding the 6- and 12-months conditioned samples (refer to Figure 23 to Figure 26), the initial cracks are generated by sustained loads applied via the load frames. The width of cracks in samples subjected to higher tension loads was greater than in those with lower loads. For instance, initial cracks in the 6-months ambient-cured M100-C1 samples were wider than in the M0. Additionally, cracks in ambient-cured samples were generally wider than those in marine-conditioned samples, consistent with the higher strength observed in the marine-conditioned samples. After 12 months, cracks continued to widen in ambient and marine-conditioned samples due to prolonged exposure to sustained loads. However, the cracks under sustained load in all samples remain within the acceptable crack width of 0.3 mm (Australian Standard, 2018) for both marine and ambient conditioning periods.

An analysis of post-failure crack widths revealed that samples containing RCA exhibited wider cracks than those cast with natural aggregates. Due to their crack-arresting capabilities, BF helped reduce crack width compared to plain beams cast with natural and coated RCA. The effectiveness of BF in minimising crack width and propagation has also been well-documented in the literature (Şahin et al., 2021; Hossain et al., 2023). However, only a small number of BF were visible in the post-failure cracks, indicating that a combination of fibre pull-out and rupture caused the failure of the samples. This observation is consistent with findings reported by Branston et al. (2016), who anticipated potential durability issues for BF-concrete mixes over extended periods or under more harsh environmental conditions.

Three mixes’ microstructure was analysed using SEM images of the specimens of 28-day, six-month, and twelve-month ambient and marine conditions under 1,000x magnification, as illustrated in Figure 27. In the samples of three mixes, a considerable amount of Calcium-Silicate-hydrate (C-S-H) gel can be seen, which contributes to the strength of the concrete. No unreacted fly ash or GGBS are present, indicating that the aluminosilicate materials have reacted appropriately. For the samples of 6 months marine-conditioned, crystalline salts can be seen, which means the penetrates of seawater. Regarding the 12-month samples, there is a substantial amount of ettringite (Aft) and salt crystalline in their SEM images, indicating the influence of the seawater. Also, partially reacted fly ash, high porosity, and fractured BF can be seen inside the weak bound of basalt fibre and paste in the ITZ between basalt fibre and geopolymer paste in 12 months marine-conditioned sample of mix M100-C1BF.

In comparison, a very proper bond can be observed between the basalt fibre and geopolymer in the M100-C1BF 12-month ambient-conditioned sample. This aligns with the drop in the tensile strength and ultimate load capacity of mix M100-C1BF after 12 months of marine conditioning compared to 6 months of marine-conditioned M100-C1BF sample. This could also define the brittle behaviour of the M100-C1BF samples in the long term.

EDX analysis was also performed to study the chemical composition of the mixes after 12-month conditioning in ambient and simulated tidal zone of the seawater. To this aim, EDX analysis was performed in the M0, M100-C1, and M100-C1BF samples after 12 months of ambient conditioning, as depicted in Figure 28. EDX results indicate that a high concentration of Calcium (Ca), Aluminium (Al), and Silicon (Si) in the mixes of M0, M100-C1, and M100-C1BF due to the proper reaction between fly ash and slag in the mixes. Figure 29 depicts the EDX analysis on the M0, M100-C1, and M100-C1BF samples after 12 months of marine conditioning. Again, high peaks of Ca, Al, and Si show proper reaction among the alumina silicate material in the mixes. The presence of chloride can be observed in the EDX graphs of three mixes that specify the sea water’s ingression in the marine-conditioned samples.

This research explored the mechanical, structural, microstructural, and chemical properties of three geopolymer mixes incorporating natural aggregates, RCA, and BF. Reinforced beams and cylinder samples were cast for the experiments. Various tests were conducted after 28 days, 6 and 12 months of conditioning the samples in ambient and marine environments to evaluate density, compressive strength, splitting tensile strength, MoE, and the structural behaviour of the reinforced beams under three-point bending tests. Additionally, the microstructure was examined using a scanning electron microscope, and the chemical composition was analysed through EDX. The following conclusions are drawn from the study:

• (1)

  The samples containing coated RCA and BF exhibit the highest compressive strength at both ambient conditioning and marine conditioning after one year. The ambient-conditioned samples’ compressive strength gains further through one year by 24%, 12%, and 10% for mixes M0, M100-C1, and M100-C1BF respectively. Similarly, one year of marine after conditioning, the compressive strength improved further by 44%, 48%, and 39%, respectively.

• (2)

  Both mixes of M0 and M100-C1 demonstrate the increase of tensile strength for one year of conditioning in the marine environment. Adding BF significantly increased the tensile strength of the mix with coated RCA, reaching 3.41 MPa at the age of 28 days. A drop of 8.2% and 11.1% was observed for one year marine and ambient-conditioned M100-C1BF samples due to the potential degradation of BF and weakening of the bond between fibres and geopolymer paste. Still, the splitting tensile strength of 12-month marine-conditioned beams with BF is higher than that of their counterparts with sole-coated RCA.

• (3)

  Including RCA and BF in M100-C1 and M100-C1BF mixes resulted in slightly lower MoE. Marine conditioning notably enhanced the MoE in all mixes during the first six months, likely due to accelerated hydration processes in the elevated temperatures of the aquatic environment. However, prolonged exposure led to a reduction in MoE. Despite this decline, marine-conditioned samples maintained higher stiffness than ambient-conditioned samples after one year.

• (4)

  Incorporating coated RCA increased the ultimate load of reinforced beams after a three-point bending test for both ages on 28 and 365 days of marine and environment conditioning, while the post-failure deformation and toughness decreased with wider cracks in failure. BF improved the toughness and increased the ultimate load capacity due to the role of fibres in arresting the cracks throughout the increase in shear force.

• (5)

  Marine conditioning for 12 months significantly enhanced the mechanical and structural behaviour of all three mixes, indicating the suitability of SCGC mixes for marine and underground structures.

• (6)

  The SEM analysis of the three mixes revealed a considerable amount of C-S-H gel contributing to strength, with no unreacted fly ash or GGBS. Crystalline salts appeared in six-month marine-conditioned samples, indicating seawater penetration, while ettringite and salt crystals were prominent in twelve-month marine-conditioned samples. In the twelve-month marine-conditioned M100-C1BF mix, weak bonding at the Basalt Fibre–Interfacial Transition Zone (BF-ITZ) bonding, high porosity, and fractured BF were observed, explaining the reduced tensile strength and brittle behaviour. In contrast, a stronger bond was seen in the twelve-month ambient-conditioned M100-C1BF sample.

• (7)

  The EDX analysis of the M0, M100-C1, and M100-C1BF samples after 12 months of ambient conditioning showed high peaks of Ca, Al, and Si, indicating effective reactions among the aluminosilicate materials in the mixes. Additionally, chloride presence was detected in the EDX graphs of all marine-conditioned samples, confirming seawater ingress.

In summary, this study investigated cement-free concrete’s durability and long-term properties as a sustainable construction material containing crushed waste concrete as a total mass of coarse aggregate. It highlights the substantial impact of environmental factors on the properties of the GPC mixes and improvement in the mechanical and structural behaviours of RCA-based SCGC under marine conditions.

The authors gratefully acknowledge the valuable support provided by Deakin University throughout this study. They also extend their sincere thanks to Lube Veljanoski, Michael Shanahan, Dr. Reza Parvizi, Dr. Ehsan Bahrami Motlagh, Dr. Andrew Sullivan, Leanne Farago, and Uli Bauer for their assistance during the experimental phase. Appreciation is also given to Fly Ash Australia for supplying micro fly ash and to Independent Cement for providing GGBS as binding materials.

Figure A1