1 Introduction
Cement manufacturing is considered one of the main sources of greenhouse gas emissions. Among the greenhouse gases, the contribution of CO
2 emitted from fossil fuels and industrial processes to global warming is about 65% [
1]. Cement manufacturing is one of the high CO
2 emitting sources, alongside deforestation and the burning of fossil fuels. Therefore, any attempt to reduce the CO
2 released during cement production or using less cement-based materials will significantly contribute to sustainability by reducing the environmental impact of cement production.
Solid waste generation has increased massively around the world in recent decades, and there are no signs of it decreasing. According to a report, worldwide municipal solid waste generation is expected to have increased by roughly 70 % to 3.4 billion metric tons by 2050 [
2]. Despite this high increase ratio, a small part of it is recycled. In the USA, only 32.1% of the 4.9 pounds of waste produced daily by a single person were recycled in 2018 [
3]. The amount of garbage produced in the USA increased between 1960 and 2018 from 88 to 292 million tons. Out of the 292 million tons, only about 94 million tons of solid waste were recycled, with a percentage of 32.1 [
3]. Additionally, finding a place for waste is a major challenge due to the increasing levels of consumption and urbanization.
Among recyclable materials, glass is considered one of the least recycled, with only 25% of glass products being recycled [
3]. Due to this small percentage of recycled glass waste, many research studies have been conducted to find more practical applications for waste glass (WG). Otherwise, WG causes significant environmental problems due to its large volume and slow decomposition rate. One of the first investigations into using recycled WG in building materials was conducted in 1963 when Schmidt and Saia [
4] transformed glass waste into glass chips for producing wall panels. Since then, extensive research has been carried out to utilize WG in building materials. One of the key findings from these studies is the potential of using recycled glass cullet as a sustainable aggregate source in the building sector [
5].
However, several challenges emerged after the WG was introduced as an aggregate. Early trials were unsuccessful because of alkali-silica reaction (ASR) gel formation and reduced workability and strength properties in concrete. Additionally, using glass bottles limits the size and shape of WG particles, as the maximum size of the particles cannot be bigger than the thickness of the bottle. Crushed bottle wastes can be used as another source of aggregate. However, they tend to produce huge, flat, elongated aggregates that may negatively affect workability and have been linked to declines in compressive strength [
6,
7]. Another challenge is the contamination and uniformity of glass waste. Differences in color lead to diversity in chemical composition for most shattered WGes, making them unrecyclable. Paper and plastic labels, caps, and corks left over from the bottles’ original use and contents, such as sugar, could all cause contamination [
8].
Additionally, seasonal variations in the waste stream should be anticipated. Another disadvantage of using WG as the aggregate is the negative effect on mortars in a fresh and hardened form, such as workability, compressive strength, and flexural strength [
9,
10]. Another negative effect reported by Afshinnia and Rangaraju [
9] was the ASR due to the usage of WG as aggregate. Various chemical and mechanical treatments have been proposed to address durability challenges. One of these treatments was crushing WG into powder, which in turn activates its pozzolanic reaction and reduces ASR.
Reactive silica in some types of aggregates and alkalis from cement (mainly) can react to form ASR gel, a harmful gel that expands with temperature and moisture, causing internal stresses in concrete [
11,
12]. This reaction is considered one of the primary causes, after corrosion, for the deterioration of highway concrete structures [
13,
14]. Several strategies have been proposed to mitigate concrete damage caused by ASR. Out of all these strategies, the use of supplementary cementitious materials (SCMs) at the appropriate dosage level is the most successful and efficient way to stop the harmful effects of ASR [
15]. Fly ash, silica fume, ground granulated blast-furnace slag, and metakaolin are a few SCMs that have been partially substituted for Portland Cement in previous studies to effectively control ASR expansion [
16]. Recently, WG has also been utilized as an SCM to reduce ASR.
Several researchers have investigated the use of WG in cement-based materials, either as a partial replacement for aggregates or Portland Cement [
6–
10,
15,
17–
34]. The primary challenge of using glass is the increased risk of ASR gel formation, as glass contains a high amount of silica. However, many studies have reported successful results using WG in powder form as an SCM [
15,
17,
18]. While the exact mechanism by which glass powder mitigates ASR is not yet clear, it is linked to the pozzolanic reaction and the subsequent effects on the hydration products. WG can be considered pozzolanic when finely ground, typically to a particle size below 75 µm [
34].
The accelerated mortar bar test (AMBT) has been used in various studies to determine the optimal dosages of glass powder required to reduce ASR. Ke et al. [
18] demonstrated that substituting 20% of cement with waste glass powder (WGP) with particle sizes ranging from 38 to 53 µm by weight outperformed fly ash in reducing ASR. Cai et al. [
20] also used WGP as a 10% substitution to test the effect of ASR mitigation on mortar bar samples, finding that WGP achieved a greater reduction in ASR growth compared to other SCMs. Douaissia et al. [
29] studied the impact of gradually substituting glass powder and silica fume for 10% to 50% of the volume of cement on the ASR expansion and mechanical characteristics of hardened mortar samples. They found that incorporating glass powder decreased ASR expansions and mitigated mechanical performance losses caused by ASR. Zheng [
30] claims that ASR mitigation is possible through the glass’s pozzolanic reactivity, which consumes calcium hydroxide (CH) to create calcium silicate hydrates (C-S-H), and a decline in the concentration of ettringite and mono-sulfate in the concrete’s pore solution. He concluded that the mitigating effect is due to the increase in aluminum (Al) concentration in the pore solution, which reduces the dissolution of amorphous silica from reactive aggregates. The increase in Al concentration is explained as a result of the decrease in solid mono-sulfate.
Another challenge with using glass powder in cement-based materials is the reduced early age strengths, although the longer-term strengths were comparable [
20,
34]. On the other hand, many studies have investigated the effect of using nano-silica (NS) as a cementitious replacement [
20,
35–
37]. These studies concluded that adding up to 2% or 3% of NS to cement-based materials can improve mechanical strength, particularly in the early ages, by accelerating the hydration process due to its extreme reactivity. NS accelerates cement hydration and can absorb and convert CH into C-S-H gel, which improves the mechanical properties of concrete. Furthermore, compared to silica fume and other SCMs, the pozzolanic reaction in NS is considered more effective. According to Abhilash et al. [
38], concrete developed long-term strength and better durability due to the filler property, pozzolanic action, and pore structure refinement of NS. NS helps reduce the size and amount of CH crystals, creating a denser interfacial transition zone in the microstructure between aggregates and cement paste. Arif et al. [
39] investigated the early strength of concrete with NS and concluded that the nano-modified concrete achieves significantly higher early-age strength compared to the increase in strength observed at 28 d. In terms of ASR mitigation, NS can be effective at relatively low dosages, as little as 3% of cementitious materials, compared to higher addition rates typically required for microsilica, especially when used in the colloidal form [
37].
One of the disadvantages of using WGP as a cementitious replacement is the reduction in early-age strength. To address this issue, NS can be used together with WGP to accelerate the early-age reactions. It is expected that NS will compensate for the reduction in early-age strength caused by WGP by accelerating the pozzolanic reactions. Although the effects of using NS and WGP individually in cement-based materials have been studied extensively, studies on their combined use are very limited. In one of the rare studies examining the combined effect of WGP and NS, Cai et al. [
20] investigated the impact of 10% WGP, 2% NS, and their mixture on mitigating ASR. While their study provided valuable insights, our research explores a wider range of WGP and NS ratios. This approach allows for a more comprehensive understanding of how varying proportions of these materials can optimize the overall performance of cement-based materials.
Therefore, this study aims to bridge this gap by investigating the combined effects of WGP and NS on mitigating ASR without significant alterations on mechanical properties. To achieve this, mortars with 10%, 15%, 20%, and 25% of WGP, combined with 0%, 1%, 1.5%, and 2% of NS, were prepared and tested using the AMBT to evaluate ASR performance. Mechanical properties such as compressive, and flexural strength were measured, and capillary water absorption was also investigated. Additionally, other key properties such as microstructure, thermal stability, and crystalline structure were analyzed for selected mixes using scanning electron microscopy (SEM), thermo-gravimetric analysis (TGA), and X-ray diffraction (XRD) methods. By partially replacing cement with waste materials, we aim to reduce the carbon footprint associated with cement production and create more eco-friendly, sustainable, and durable construction materials.
2 Materials and methods
2.1 Materials
The WG used in this study was in powder form and obtained by grinding E-glass fibers to an average particle size of 0.1 mm. It was then sieved using a 0.06 mm (60 µm) sieve before being used in the experiments. The density of the WGP was 2.47 g/cm
3. The chemical compositions of WGP and cement were determined using X-ray fluorescence analysis, and the results are presented in Table 1. According to the chemical requirements specified in ASTM C1866/1866M [
40], the E-glass fibers used can be classified as type GE. The total equivalent alkali content was 0.53% (Na
2O
eq), which is well within the maximum recommended Na
2O
eq of 4% specified in the standard. Additionally, the amount retained when wet-sieved on 45 µm was lower than 5%, which was indicated as a maximum ratio in ASTM C 1866/1866M [
40].
The density of Ordinary Portland Cement (OPC) used in this study was 3.14 g/cm
3, and the equivalent alkali content was approximately 1% (Na
2O + 0.658K
2O = 0.32 + 0.658
× 1.1 = 1.04%). The loss on ignition of cement was 3.07%, and it was close to the maximum allowable ratio of 3.0% as specified in ASTM C150 [
41].
XRD results for OPC and WGP are shown in Figs. 1 and 2, respectively. We observed that the majority of the peaks for OPC correspond to alite compounds with the chemical composition of (Ca
3SiO
5). On the other hand, WGP displayed a primarily amorphous structure with non-recognizable peaks. According to ASTM C 1866/1866M [
40], the minimum recommended amorphous content for GE-type glass powder is 95%.
The fine aggregate used in this study was river sand obtained from the Sakarya region of Türkiye. This sand has high quartz content and its density was 2.69 g/cm
3. According to ASTM C 1778 [
42], if mortar bars expand less than 0.10% in the AMBT after 14 d of immersion, the aggregate is classified to be non-reactive in most cases. If the mortar bars expand 0.10% to 0.30%, aggregate is classified as moderately reactive, 0.30% to 0.45% highly reactive, and more than 0.45% very highly reactive. The ASR expansion of this sand measured by AMBT according to ASTM C1567 [
43] was 0.36% at 14 d of immersion. Therefore, the sand used in this study can be considered a highly reactive aggregate.
NS (Levasil CB8) used in the production was an alkaline, aqueous dispersion of colloidal silica containing 50% solids by weight. The silica dispersion was sodium stabilized, with the amorphous silica particles carrying a negative surface charge. The SiO
2 particles were discrete, smooth, spherical, and presented a wide particle size distribution. The physical appearance of the dispersion was a white liquid and slightly more viscous than water. Its PH was 9.5, viscosity was 8 cP, and density was 1.4 g/cm
3 [
44]
.Polycarboxylic ether based high range water reducing admixture (MasterGlenium 51) was used in the productions to enhance the flowability of mortars. It was utilized in varying ratios between 1.1% to 1.5% depending on the requirements determined through trial productions. The density of the admixture was 1.05 g/cm3, and its pH value was (6.0 ± 1).
2.2 Mix design
The replacement ratios were 10%, 15%, 20%, and 25% for WGP, and 0%, 1%, 1.5%, and 2% for NS by weight of cement. Twenty different combinations of these ratios of WGP and NS were used in the productions. For mortar coding, the first number given in front of GP indicates the replacement ratio of WGP, while the second number in front of NS refers to the replacement ratio of NS with cement. For example, a mixture coded as 10GP1NS contains 10% WGP and 1% NS. The mortars produced solely with OPC without any replacement of admixtures were labeled as “Control” in order to measure the expansion resulting from the reactive aggregate and to understand the effectiveness of mortars made with WGP and NS. Tables 2 and 3 summarize the mix design of mortars produced for ASR tests and mechanical tests.
2.3 Methods
2.3.1 Flowability and air content
In fresh mortars, both flowability and air content were tested. The addition of WGP did not significantly affect the mortar flowability. However, adding NS decreased the flowability. The flowability level was found to be higher in mortars with 0% NS and showed a decreasing trend with increasing NS content. This trend was observed consistently in mortars containing 10%, 15%, 20%, and 25% WGP. The highest flow rate measured was 24 cm, while the lowest was 17 cm.
When it comes to air content, the amount of WGP did not contribute much to the air content percentage. However, NS significantly influenced the air content in most specimens. The air content ratio increased with the increase of NS content. For example, using 2% NS increased the air content from 4.3% to 6.1% in specimens coded as 20GP0NS and 20GP2NS, respectively.
2.3.2 Accelerated mortar bar test
AMBTs were conducted according to ASTM C1567 [
43]. A mixture of cement, sand, and water was prepared according to specific proportions. The dry materials for the test mortar were proportioned using 1 part of cement to 2.25 parts of graded aggregate by mass. The quantities of dry materials to be mixed at one time in the mortar batches for producing three specimens were 440 g of cement and 990 g of sand. A water-to-binder ratio equal to 0.47 by mass was used in all mixes. The mix design of the mortars for AMBT tests is presented in Table 2.
The mixing process was conducted according to BS EN 196-1 [
45]. The mortars were then molded into prismatic bars, using molds specially designed for the AMBT test. The bars were typically 25 mm × 25 mm × 285 mm prisms having a gauge length of 250 mm. The mortar bars were cured in a moist room at (23 ± 2) °C and a relative humidity of at least 95% for 24 h. After curing, the mortar bars were removed from the mold and stored in water at (80 ± 2) °C for 24 h, and then the initial reading was taken. Then, they transferred to containers, which had a capacity of 14 L. The containers were filled with NaOH solution to accelerate the forming of ASR gel during the test duration. The containers were then placed in an oven and stored at (80 ± 1) °C for 14 d. Comparator readings were taken at 0, 3, 7, and 14 d. The results of the tests were reported as the percentage change in the length of the mortar bars.
2.3.3 Flexural and compressive strength
Table 3 shows the mix design used for producing mortars for compressive and flexural strength tests. The mortars were prepared twice for each combination: one for 7-d tests and the other for 28-d tests. Prismatic samples with dimensions of 40 mm × 40 mm × 160 mm were prepared and kept in molds, with their surfaces covered by stretch films. After 24 h, the specimens were placed in water for curing at a temperature of (20 ± 3) °C until the test days. The water-to-binder ratio was kept at 0.5 in all productions, with OPC partially replaced with WGP and NS in varying amounts.
Three-point flexural strength tests were conducted on the prismatic specimens according to BS EN 196-1 [
45]. The samples were placed on two supports 100 mm apart from each other. The load was applied at the center of the sample at a rate of (0.5 ± 0.1) MPa/s until the sample fractured, and the flexural strength was calculated. The relationship between different combinations and their strength will be discussed in Subsection 3.2.
After breaking the samples into two pieces in the flexural strength test, one piece from each sample was tested for compressive strength, while the other piece was for the capillary water absorption test. The sample was placed on a flat surface, centered under the compression machine, and a compressive force was applied at a rate of (0.5 ± 0.1) MPa/s until the sample fractured. The maximum strength at which the sample fractured was recorded as the compressive strength of the mortar.
2.3.4 Capillary water absorption
Capillary action can significantly impact the durability and performance of building materials. Therefore, a capillary water absorption test was conducted on mortars according to BS EN 1015-18 [
46] to determine the mean coefficients of water absorption due to capillary action. For this test, the halves of the prismatic specimens that remained from the flexural test were used. The side surfaces of specimens were sealed with impermeable material to prevent water absorption from these sides. Water ingress was allowed only from the bottom surface, which was cut and smoothed. The weights of the specimens were measured at different intervals. The test results are presented in Subsection 3.3.
2.3.5 Scanning electron microscopy
SEM is a powerful tool used in various fields of science and engineering to characterize the microstructure of a wide range of materials at high magnifications. Six selected combinations were used for SEM analyses after immersing the mortar specimens in NaOH solution for 14 d to accelerate ASR gel formation. After the ASR test, the samples were stored, and SEM analyses were conducted approximately on the 90th day. The tested combinations were: C (control), 2NS, 10GP2NS, 15GP2NS, 20GP2NS, and 25GP2NS. At least four mortar pieces from each combination were used for testing after being crushed into small sizes of approximately 1 cm × 1 cm. The samples were analyzed at different zoom levels ranging from 100× to 10000×. The aim was to understand the effect of using WGP and NS in reducing the expansions of mortars due to ASR and to identify the different hydration compounds, such as C-S-H, CH, and ettringite.
2.3.6 Thermo-gravimetric analysis
TGA is a technique used to analyze various properties of materials such as thermal stability, composition, purity, and reaction kinetics. In our study, two specimens were chosen for analysis: a specimen with zero WGP content (Control) and a specimen with the highest WGP content without NS (25GP0NS). From this analysis, we aim to compare the mortars with and without WGP and determine the consumption of Ca(OH)2, which can be considered as a sign of pozzolanic reaction.
2.3.7 X-ray diffraction
XRD was used to analyze the crystalline structure of the mortars. The specimens were prepared by crushing and converting them into a powdered form before testing. Analyses were conducted using a powder diffractometer, and the patterns were quantitatively analyzed using computer software (Malvern Panalytical ver. 5.2).
In this study, various combinations of mortar specimens containing WGP and NS, along with the control specimen, were tested. Two combinations from each WGP content were chosen: WGP with no NS and the highest NS content (2%). This selection aimed to differentiate between the hydration products of each WGP content with and without NS. The selected combinations were; Control (C), 1NS, 2NS, 10GP0NS, 10GP2NS, 15GP0NS, 15GP2NS, 20GP0NS, 20GP2NS, 25GP0NS, and 25GP2NS.
3 Results and discussion
3.1 Accelerated mortar bar test
3.1.1 Effect of waste glass powder on mitigating alkali-silica reaction
Figure 3 shows the expansion results of mortar bars with varying WGP contents. The green line represents the acceptable expansion limit of 0.1% for mortar bars stored in NaOH solution for 14 d, as specified in ASTM C1778 [
42]. The WGP replacement ratios were 0%, 10%, 15%, 20%, and 25% of cement. The control specimen (0% WGP) exceeded the acceptable limit by around day 5 and reached 0.36% expansion on day 14, demonstrating the high reactivity of the utilized aggregates. In contrast, the expansions recorded for the WGP-containing specimens were 0.20% for 10GP0NS, 0.19% for 15GP0NS, 0.10% for 20GP0NS, and 0.05% for 25GP0NS. Notably, two of these specimens, the 20% and 25% replacement levels, reduced the expansion to acceptable levels. Overall, we can conclude that WGP showed an effective mitigation level, reaching an 86% reduction in expansion with 25% of WGP. Similarly, Ke et al. [
18] reported that the expansion decreased from 0.085% to 0.025% with a 30% glass powder replacement level, resulting in a 71% reduction at 14 d.
3.1.2 Effect of nano-silica on mitigating alkali-silica reaction
Figure 4 shows the relationship between ASR expansion and the NS content in the specimens. The control specimen (0% NS) exhibited an ASR expansion of 0.36%. On the other hand, specimens with 1%, 1.5%, and 2% NS showed slightly reduced expansions of 0.33%, 0.31%, and 0.32%, respectively. This corresponds to an 8% reduction with 1% NS, a 14% reduction with 1.5% NS, and an 11% reduction with 2% NS compared to the control. Although the reductions are modest, NS may contribute to ASR mitigation by refining pore structure and enhancing pozzolanic reactivity. However, the trend does not show a consistent decrease with increasing NS content, suggesting that its stand-alone effect is limited under the current conditions. A broader study with additional NS levels may reveal a more gradual trend in expansion reduction.
3.1.3 Combined effect of waste glass powder and nano-silica on mitigating alkali-silica reaction
Figure 5 shows the reduction in ASR expansion for various combinations of WGP and NS compared to the control specimen. The highest reduction was achieved with 25% WGP replacement and 2% NS (25GP2NS), reaching an 88% reduction. These results indicate that WGP effectively mitigates ASR expansion, while the effect of NS was minimal.
To further examine this outcome, the average expansion for 0, 1%, 1.5%, and 2% NS was calculated for each WGP content (0%, 10%, 15%, 20%, and 25%). The average expansion values were 0.36%, 0.24%, 0.19%, 0.10%, and 0.05%, respectively. Notably, the most effective results were achieved at 20% and 25% WGP replacement, where the expansion fell below 0.1%. This indicates that the mortar bars reached an unreactive expansion level at approximately 20% WGP replacement. In summary, replacing cement with 0%, 10%, 15%, 20%, and 25% WGP resulted in expansion reductions of 8%, 34%, 47%, 73%, and 86%, respectively. Overall, these findings clearly demonstrate the substantial inhibition of ASR expansion by WGP, while NS provides a marginal but potentially synergistic effect. Moreover, the results appear more efficient than those reported in earlier studies by Jani and Hogland [
33], and Zheng [
30]. The observed mitigation aligns with the mechanisms proposed by Gökşen et al. [
47], where the synergistic effect of WGP and fly ash significantly suppressed ASR expansion through pozzolanic reactivity and silica consumption mechanisms and developed better results than when used individually.
Several studies have attempted to explain why glass powder minimizes ASR, especially given that ASR expansion is generally expected to increase when WG is used as an aggregate due to its high silica content. According to Zheng [
30], WGP undergoes a pozzolanic reaction similar to that of other SCMs, reacting with portlandite (Ca(OH)
2) to form C-S-H gel. The high silica content in glass powder, which is the primary element in glass in terms of chemical composition, was the primary cause of the pozzolanic reaction. Our XRD and TGA results support this mechanism by confirming the pozzolanic activity of WGP. The amorphous silica in the glass promotes a faster reaction with Ca
2+ in the solution, resulting in increased formation of C-S-H and a reduction of free Ca
2+ within the pore solution. Consequently, the reaction between glass powder and Ca(OH)
2 leads to enhanced pozzolanic reactivity to produce more C-S-H and improved mitigation of ASR [
18].
3.2 Flexural and compressive strength tests
Figure 6 shows the variation of flexural strength at 7 d for different combinations of WGP and NS used specimens in this study. The overall flexural strength ranges from 6.5 to 8.2 MPa for 15GP1NS and control specimens, respectively. Results indicate that increasing WGP content alone has a reduction effect on flexural strength at an early age (7 d). Specifically, decreases of approximately 9% and 18% were observed for 10GP0NS and 15GP0NS, respectively, compared to the control. However, decreases were almost identical for 20GP0NS and 25GP0NS, at 13% and 12%, respectively. On the other hand, increasing only the NS ratio slightly decreased the flexural strength at early ages if the specimens did not contain WGP. Reductions in strength were 4%, 9%, and 5% for 1NS, 1.5NS, and 2NS, respectively, compared to the control specimen.
Figure 7 shows the flexural strength of specimens at 28 d. At this age, the flexural strength values changed between 6.5 to 10.0 MPa for 20GP2NS and 2NS, respectively. Adding only WGP decreased the flexural strength as WGP content increased. The highest decreases were observed at 22% and 19% for 20GP0NS and 25GP0NS, respectively. Adding only NS showed a similar trend to the 7-d results, with slight decreases up to 1.5NS and an increase at 2NS. Decreases were 6% and 2% for 1NS and 1.5NS, respectively, while an increase of 4% was observed for 2NS.
Figures 8 and 9 show the variation in compressive strength at 7 and 28 d, respectively. Compressive strength values range from 32.5 to 57.1 MPa for 7 d. Results indicate that higher WGP content in mixes results in lower compressive strength, similar to the flexural strength results. Decreases in compressive strength were 11%, 15%, 20%, and 37% for 10GP0NS, 15GP0NS, 20GP0NS, and 25GP0NS, respectively, compared to the control specimen.
Compressive strength values range from 49.8 to 72.0 MPa for 28 d. The trend of decreasing compressive strength with increasing WGP content continued, but decreases were generally less pronounced than 7 d. Specifically, decreases were 7%, 18%, 15%, and 25% for 10GP0NS, 15GP0NS, 20GP0NS, and 25GP0NS, respectively, compared to the control specimen. While most samples showed smaller strength decreases at 28 d compared to 7 d, the 15GP0NS sample did not. This anomaly was possibly due to variations in the microstructure. Additionally, the small difference observed for the 15GP0NS sample can be considered normal when taking into account the standard deviation of cement-based materials. Overall, the lower decrease in compressive strength after 28 d indicates lower reactivity at early ages and the late pozzolanic effect of WGP.
On the other hand, the role of NS in increasing the compressive strength was also observed, although the results were very close to each other. Increases in compressive strength were 3%, 9%, and 10% for 7 d and 1%, 2%, and 8% for 28 d for 1NS, 1.5NS, and 2NS, respectively, compared to the control specimen. These results confirm the effect of NS at early ages.
3.2.1 Combined effect of waste glass powder and nano-silica on flexural and compressive strength
NS contributed to the slight increases generally and balanced the reductions caused by WGP at 7 d. For example, the results for flexural strength with 25% WGP and 0, 1%, 1.5%, and 2% NS were 7.2, 7.6, 7.3, and 7.4 MPa, respectively. These results indicate 6%, 2%, and 3% increases according to 25GP0NS. However, the results are relatively close to each other. At other WGP replacement levels, slight increases in flexural strength were generally observed with the addition of NS. The highest increase was observed for 15GP1.5NS, which exhibited a 15% increase compared to 15GP0NS. This suggests that the presence of NS helps to improve the overall flexural strength, with the optimal improvement seen at specific replacement levels.
A similar trend is observed at 28 d. For example, flexural strength results were 8.3, 9.5, 9.4, and 8.0 MPa for 15GP0NS, 15GP1NS, 15GP1.5NS, and 15GP2NS, respectively. These results indicate 14% and 13% increases for 15GP1NS and 15GP1.5NS, and a 4% decrease for 15GP2NS compared to 15GP0NS. Using NS together with WGP positively impacts the flexural strength results in general, similar to findings in the literature that investigate the effect of NS on compressive and flexural strength [
48].
In terms of compressive strength test results, slight increases or almost similar results were recorded, generally with the addition of NS. With increasing NS ratio in mortars, containing both WGP and NS, compressive strength increased slightly at 0%, 10%, and 15% WGP replacement levels at 7 d. With 20 and 25% of WGP levels, increasing NS did not lead to significant changes in compressive strength. The average increases in compressive strengths with the addition of NS were 6%, 3%, and 4% for 0%, 10%, and 15% WGP, while 1% decrease and increase were observed for 20% and 25% of WGP, respectively. Similarly, the role of NS in increasing compressive strength was observed at 28 d. The average increases in compressive strength with the addition of NS were 3%, 3%, 8%, 2%, and 7% for 0%, 10%, 15%, 20%, and 25% of WGP, respectively. Although the effect of NS was minimal, slight increases were observed in almost all combinations.
The improvements can be attributed to better particle packing and the accelerated hydration process facilitated by NS, leading to increased formation of C-S-H gel and an overall enhanced microstructure. These findings align with some literature reviews investigating the effect of NS on the initial setting time and initial hardening process [
36,
49]. According to Nigam and Verma [
49], adding NS forms more H
2SiO
2 in the hydration process, which then reacts with more Ca
2+ ions, generating more C-S-H.
However, some inconsistent results were observed with increasing NS content, mainly due to the agglomeration effect of colloidal NS, especially at higher concentrations (2%). This phenomenon causes the agglomeration of NS particles at specific points rather than being dispersed uniformly throughout the mortar, creating weak points. Similar observations were reported in other studies where NS agglomeration was observed during the production process, and some leftover NS was noticed to be collected at the bottom of the container during mixing [
50]. These factors have slightly affected our results.
Unlike compressive strength, adding NS did not significantly affect flexural strength. Similarly, Turkmenoglu et al. [
48] reported that adding NS showed a noticeable enhancement in compressive strength and, at the same time, did not significantly affect flexural strength. They investigated the synergistic effect of nanosilica and microsilica and their combination, finding that NS had a negative effect on the brittleness of the concrete. They recorded this effect by calculating the ratio between flexural strength and compressive strength. A reduction in the ratio means an increase in the brittleness of the concrete. Similar results were found in our study. Figure 10 shows the ratio of flexural strength to compressive strength at 28 d. We notice that adding NS slightly affected brittleness by causing decreases in this ratio. This ratio increased with the increase in the WGP replacement ratio up to 15%, then started to decrease. In summary, these results suggest that WGP may modestly enhance ductility, while NS has an adverse effect. This influence corresponds to the physical structure of this type of glass and has nothing to do with the hydration process.
The pozzolanic activity index (PAI) of mortars was also calculated using the compressive strength test results. PAI is recommended as 75% at 28 d according to BS EN 450-1 [
51] with a 25% replacement of cement. In our study, PAI was calculated as 75% for the 25GP0NS, barely meeting the minimum requirement. However, PAI values increased with the addition of NS. PAIs of the mortars produced with 25% WGP and 1%, 1.5%, and 2% NS were 77%, 82%, and 88% at 28 d, respectively (for 25GP1NS, 25GP1.5NS, and 25GP2NS).
3.3 Capillary water absorption
The mean coefficients of water absorption values are determined according to BS-EN 1015-18 [
46], as shown in Fig. 11. Results indicate that the least amount of water was absorbed by the specimens with 2% nano-silica (2NS) and the control specimens. Specimens with 1% and 1.5% NS absorbed slightly more water through the test’s duration, but at the same time, the ability to absorb water decreased as the NS percentage increased.
In specimens with WGP content, the initial addition of 10% WGP increased capillary water absorption compared to the control specimen. This high water absorption with 10% WGP can be attributed to the initial disruption in the mortar matrix caused by the incorporation of WGP particles. The introduction of 10% WGP may lead to increased porosity and reduced density in the mortar, allowing more water to be absorbed initially.
As the proportion of WGP increased beyond 10%, the effects on capillary water absorption changed. When the WGP content increased from 10% to 25%, the capillary water absorption coefficient decreased from 0.163 to 0.123 kg/(m
2∙min
0.5). Increasing the percentage of WGP decreases the amount of water absorbed. This reduction can be attributed to the filling effect of additional WGP particles within the mortar matrix, which reduced pore size and increased compaction, thus diminishing water absorption. Higher proportions of WGP may have also improved particle packing and interfacial bonding, further reducing water ingress. A comparable outcome was observed in another study when marble powder was used. Priyadarshini et al. [
52] reported a sudden decrease in water absorption when 10% of the cement was replaced with marble powder. In their study, an initial increase in filler percentage led to increased void content in the polymer composite, facilitating water intake. However, further increasing the filler percentage to 10% resulted in reduced water absorption due to improved particle packing and reduced porosity.
Moreover, combining WGP with NS reduced capillary water absorption more effectively than using WGP alone. For example, the capillary water absorption coefficient decreased from 0.163 to 0.133 kg/(m2∙min0.5) from 10GP0NS to 10GP2NS. A similar trend is observed in other mixes as well. This indicates that NS helps reduce the increased capillary water absorption coefficient caused by WGP. The addition of NS improved the microstructure of the mortars, possibly by decreasing the porosity and increasing the pozzolanic reaction with CH produced during the hydration of cement. The small particle sizes of NS filled the voids between larger cement particles, leading to a denser packing of the particles in the mortar and reducing the overall porosity. Additionally, the generation of more C-S-H gel further refined the pore structure and reduced the connectivity of capillary pores. WGP had a similar effect, and the best combination of WGP and NS was observed in the 25GP2NS mix, which used the highest amount of additives. This proves the synergistic effect of WGP and NS when combined.
3.4 Scanning electron microscopy
SEM images of the mortars, referred to as control, 2NS, 10GP2NS, 15GP2NS, 20GP2NS, and 25GP2NS, at 90 d of age can be seen in Figs. 12 to 17. From these figures, we can notice a reduction in permeability by adding WGP and NS. This reduction in permeability can be attributed to a denser structure with fewer pores and more C-S-H.
In Fig. 12, the control sample shows cracks as evidence of ASR gel formation. Mortar with 2% NS presented a denser structure with more C-S-H formation, but cracks were still found (Fig. 13). These cracks likely occur because the control and 2% NS specimens do not have sufficient amounts of WGP to effectively mitigate the ASR gel formation. It should be noted that the origin of these cracks may also be due to the crushing process of the sample.
The blended mortar with 10% WGP and 2% NS showed a slight increase in hydration products and a relatively dense internal structure with the appearance of CH crystals and ettringite (Fig. 14). This observation is further supported by the XRD results shown in Subsection 3.6, where the highest CH peak intensity was recorded for the 10GP2NS sample. When the WGP content reached 15%, the internal structure became denser, and fewer cracks were found with more C-S-H formation. Also, ettringite and CH crystals started to disappear, which contributes to higher strength (Fig. 15).
Figure 16 shows the microstructure of the 20% WGP and 2% NS used specimen, which is somewhat similar to the microstructure of 15GP2NS as seen in Fig. 15. The observed three-dimensional C-S-H gel contributes to mitigating ASR expansion. Additionally, the number of pores and cracks decreased, indicating improved compactness in the cement mortar. An increase in WGP content promotes cement hydration. Moreover, the content of C-S-H gel increased significantly in the hydration product of mortar with 25% WGP (Fig. 17), and the structure exhibits relatively less porosity.
In summary, the SEM images confirm that the cement paste is densest when the WGP and NS content increase. This trend aligns perfectly with our capillary water absorption results. The number of hydration products significantly increased, with the presence of three-dimensional C-S-H gel, lamellar CH crystals, and needle-like ettringite. These hydration products overlap, and there is a substantial reduction in the number of holes and cracks, resulting in a dense network structure. This demonstrates the effectiveness of WGP and NS in different contents in mitigating ASR gel formation. However, the relationship between C-S-H gel formation and strength is more complex. It can be influenced by factors such as the Ca/Si ratio, as discussed by Zheng [
30]. He explained that two types of C-S-H are created: the outer layer and the inner layer. The inner layer can be seen directly around glass particles in their SEM images. The low Ca/Si ratio in the glass powder that makes up the outer layer affects the mechanical properties. The ASR gel’s qualities are similar to those of C-S-H gel due to the decrease in Ca/Si ratio, which also affects the gel’s swelling and expansion characteristics.
Although the morphology of hydration products such as C-S-H gel, CH, and ettringite was interpreted based on SEM images, it should be noted that no EDS analysis was performed to confirm their chemical composition. Therefore, the identification of these phases remains morphological and indicative, not definitive. To support the interpretation, XRD results presented in Section 3.6 provide complementary evidence for the presence of these hydration products. A similar approach was adopted by Kaya et al. [
53] in their study on slaked lime-based alkali-activated mortars, where SEM morphology was used to infer the presence of hydration products such as C-S-H and ettringite, while phase confirmation relied on XRD analysis. Accordingly, the needle-like structures interpreted as ettringite and the fibrous or honeycomb-like textures attributed to C-S-H gel are consistent with hydration morphologies reported in previous studies, but should be considered tentative identifications in the absence of direct compositional analysis.
3.5 Thermo-gravimetric analysis
TGA was performed to investigate the relationship between temperature and weight loss in each specimen. Figure 18 illustrates the overall weight loss as the temperature increases, while Fig. 19 depicts the rate of weight loss in response to gradual temperature changes. The rate of weight loss in the control specimen (about 0.07%) was larger than that of the 25GP0NS specimen (0.03%) in intervals between 400 and 500 °C. The loss in weight in this interval is attributed to the content of Ca(OH)
2 (CH), as mentioned by Jiang et al. [
31].
In TGA, the specimens undergo three distinct phases as the temperature rises. In Phase 1, temperature increases to 100–150 °C, causing the dehydration of hydrated silicates and the evaporation of water. When the temperature reaches 400–500 °C during Phase 2, the decomposition of Ca(OH)
2 and the formation of CaO are the primary causes of weight loss. In Phase 3, the decomposition of CaCO
3 accounts for most of the weight loss when the temperature reaches 700–800 °C [
31]. The higher rate of weight loss in the control specimen indicates a larger amount of Ca(OH)
2 present in the control specimen relative to the 25GP0NS specimen.
To evaluate the change in Ca(OH)
2 content, the section of the TGA curve ranging from 400 to 600 °C was used to calculate the Ca(OH)
2 by Eq. (1) as given by Aliabdo et al. [
32].
In this context, WT represents the total weight loss percentage between 400 and 600 °C, WCH represents the molecular weight of CH, which is 74 g/mol, and WW represents the molecular weight of water, equal to 18 g/mol. The weight loss percentages (WT) were 2.68% for the control and 1.822% for 25GP0NS, and Ca(OH)2%s were calculated as follows:
• For Control: CH % = 2.68 × (74 / 18) = 11.02%,
• For 25GP0NS: CH % = 1.822 × (74 / 18) = 7.49%.
These results indicate a reduction in Ca(OH)
2 content when cement is substituted with WGP. Similar reductions were reported by Aliabdo et al. [
32] with the addition of WGP, which can be attributed to the pozzolanic effect of WGP or the reduction in cement content.
3.6 X-ray diffraction
The primary objective of the XRD test is to monitor the presence of various mineral compounds and hydration products. Figure 20 provides a visualization of the different peaks identified in the XRD test within the intervals of 17° to 50° of 2θ. Notably, the most intense peak observed is attributed to quartz around 26.6°, followed by C3S (alite) around 29.3°. The main peaks for CH in the XRD spectra were found at 18.00° 2 (with a corresponding d-spacing of 4.9092 Å), 34.10° 2 (d = 2.6270 Å), and 47.12° 2 (d = 1.9268 Å). These peaks indicate the presence and relative quantities of these compounds in the different mortar samples.
Figures 21 and 22 show the results of the XRD test within the intervals of 12°–20° and 30°–35° of 2θ, respectively. Peaks are labeled with “CH” at various points around 18° and 34°, showing the presence of CH. Figure 23 presents the CH percentage, while Fig. 24 shows the quartz percentage in various mortar specimens according to quantitative analysis of the XRD test. The significant reduction in CH content from 15.3% to 4.8% with the addition of 2NS highlights the effectiveness of NS in enhancing pozzolanic reaction. The results are discussed in two stages: first, focusing on specimens containing only WGP, and second, specimens with both WGP and NS.
3.6.1 Mortars with waste glass powder
Examining all mortars in Fig. 23, a descending trend in CH’s crystal content can be observed from control mortars up to 25% WGP without NS (25GP0NS). CH’s crystal content decreased from 15.3% in the control specimen to 12.1% and 11.1% with 10% and 15% WGP replacement, respectively. Although this increased to 19.5% for the 20% WGP used specimen, CH crystal content reduced again to 13.3% for the 25% WGP used specimen. This anomaly can be explained with the excess SiO2 in the system (as can be seen in Fig. 24), which may cause re-crystallization of CH crystals, leading to a sudden increase in intensity at 20% WGP content. Further increase of the WGP content to 25% may prompt additional pozzolanic reactions with CH, causing a decrease in the intensity of CH crystals once again.
Overall, the trend indicates a reduction in CH content with increasing WGP content. The reduction in CH’s crystal content can be attributed to enhanced pozzolanic activity with increasing WGP content, during which more CH is consumed to produce more C-S-H. However, the C-S-H gel is amorphous and hard to determine its peak intensity in XRD spectra. These results are aligned with our TGA results and similar conclusions have been reported by Li et al. [
54].
3.6.2 Mortars with waste glass powder + nano-silica
As shown in Fig. 23, specimens with 1% and 2% NS exhibited a reduction in CH’s crystal content gradually from 15.3% in the control mortar to 13.8% and 4.8%, respectively. Specimens containing both WGP and NS generally showed a greater reduction in CH content compared to specimens with the same amount of WGP only, except for 15GP2NS. For example, 10GP0NS had 12.1% of CH, and 10GP2NS had 9.9% of CH content. A similar trend was observed for 20GP and 25GP specimens with and without NS. These results indicate that NS enhances pozzolanic activity by consuming CH in specimens containing WGP.
The effect of NS in increasing the quartz formation as a cementitious replacement appeared in Fig. 24. In specimens with 1% and 2% of NS, quartz formation significantly increased from 46.9% in the control specimen to 51.5% and 82.0%, respectively. For the specimens with both WGP and NS, the formation level of quartz increased in 10GP2NS, 20GP2NS, and 25GP2NS compared to specimens with the same replacement level of WGP only. For example, 10GP0NS had 49.2% quartz in comparison with 58.3% quartz content for 10GP2NS. Quartz is known for its hexagonal crystalline structure, which indicates the formation of well-organized silica. This can be considered as an explanation for mitigating ASR, as the formation of quartz consumes the excess amorphous silica present in the hydration products. Consequently, there is insufficient free amorphous silica to form ASR gel, which helps in reducing the potential for ASR-related expansion and cracking in the concrete.
In addition to the quantitative analysis of CH and quartz phases, the XRD results also provide indirect support for the presence of hydration products such as C-S-H gel and ettringite, which were morphologically interpreted in SEM images. Although these phases are typically amorphous or poorly crystalline and therefore difficult to detect directly via XRD, their formation can be inferred from the progressive consumption of CH and the evolution of silica-related peaks. This approach is consistent with the methodology adopted by Kaya et al. [
53], who interpreted SEM morphologies of C-S-H and ettringite in slaked lime-based alkali-activated mortars, while relying on XRD data to confirm phase development. In our study, the reduction in CH content and the increase in quartz formation, particularly in specimens containing NS, suggest enhanced pozzolanic activity and the likely formation of secondary hydration products. However, due to the absence of EDS analysis, these identifications remain indicative and should be interpreted with caution.
4 Conclusions
This study investigated the effectiveness of WGP and NS as cementitious replacements in mitigating ASR in mortars and examined their effects on mechanical and durability properties. Twenty mortar mixtures were formulated with varying combinations of WGP and NS, and their mechanical properties, ASR mitigation potential, and pozzolanic reactivity were systematically evaluated. The combined use of WGP and NS demonstrated effective ASR mitigation while maintaining acceptable compressive and flexural strength properties. Further research is needed to investigate potential agglomeration effects, especially with smaller WGP particle sizes combined with NS, and to explore the effect of variations in particle sizes of WGP. The following key findings and conclusions can be drawn from this research.
1) WGP demonstrated significant effectiveness in mitigating ASR-induced expansion. Adding more WGP reduced the expansion induced by ASR. Mortars incorporating WGP at replacement levels of 20% and 25% exhibited a notable reduction in expansion from 0.36% to 0.1% and 0.05%, respectively. While NS did not exhibit a direct mitigation effect on ASR gel formation, it accelerated the pozzolanic reaction in the early ages, suggesting its potential for enhancing early-age strength development in concrete mixes.
2) Mortars with WGP replacements showed lower pozzolanic reactivity at early ages, as evidenced by the difference between compressive and flexural strength results at 7 and 28 d. Increasing the WGP replacement level tended to reduce flexural and compressive strength, particularly evident in the 7-d test results. The mortar containing 25% WGP exhibited 37% and 25% less compressive strength than the control specimen at 7 and 28 d, respectively. Conversely, NS incorporation led to slight increases in compressive and flexural strength at both 7 and 28 d. The mortar containing 2% NS exhibited 10% and 8% higher compressive strength than the control specimen at 7 and 28 d, respectively.
3) TGA and XRD results confirmed the pozzolanic reaction of WGP, as indicated by the reduction in CH content and the appearance of WGP-related compounds. SEM analysis revealed a reduction in cracks in mortars containing WGP and showed a denser structure in mortars with WGP replacement, highlighting its potential for enhancing durability. The pozzolanic activity of WGP is the main reason to mitigate ASR.
4) Although the observed pozzolanic reaction indicated by TGA and XRD’s results led to successful mitigation of ASR expansion, a reduction in compressive strength with increasing WGP content was recorded, especially at early ages. This reduction in strength may be attributed to the depletion of portlandite in the specimens and retarding effect of WGP in the hydration reactions as a cementitious replacement, similar to other SCMs. The enhancement of compressive strength at 28 d compared to 7 d is explained by the effect of the pozzolanic reaction of WGP at later ages.
5) The addition of NS helped to mitigate the increase in the capillary water absorption coefficient caused by WGP. As also indicated by SEM results, NS improves the microstructure of mortars by decreasing porosity and increasing the pozzolanic reaction with CH produced during cement hydration.
Incorporating WGP and NS in the production of cement-based materials may create sustainable and effective solutions by reducing the cement content in the built environment. Subsequent research should focus on determining the most effective methods for obtaining, processing, and utilizing waste materials in cement-based materials while also assessing their long-term performances and environmental impacts.
PDF
(8297KB)
