Experimental and Statistical Analysis for Eco-Friendly Self-Compacting Concrete Incorporating Recycled Coarse Aggregate and Nano-Glass Waste

2.1 Cement, FA and superplasticizer

Ordinary Portland Cement (OPC-Type I), commercially known as ALMASS, was utilized as the main binder. Additionally, low-calcium Class F fly ash (FA), provided from Euro Build in India, was incorporated as a supplementary cementitious material to enhance the sustainability and performance of the SCC-NGRA mixtures. The fineness of the cement used was 340 m²/kg, while the FA exhibited a lower fineness of 290 m²/kg, reflecting the relative specific surface areas of their particles. The specific gravity was 3.15 for the cement and 2.2 for the FA, indicating a notable difference in density that may influence the workability and overall behavior of the SCC mixtures. The chemical compositions of both the OPC and FA were determined through X-ray fluorescence (XRF) spectroscopy, and the results are tabulated in Table 1.

Table 1. Chemical composition of cement and FA determined by X-ray fluorescence

The OPC complies with ASTM C150 standards for Type I cement, ensuring consistent quality and performance in concrete applications. Meanwhile, the Class F FA meets the requirements of ASTM C618, characterized by its low CaO content (< 10%) and high silica and alumina content, which contribute to its pozzolanic reactivity.

A superplasticizer based on polycarboxylic-ether formulation with a specific gravity of 1.08 ± 0.02 and PH-value of 7 ± 1 was used in all mixes to attain the required slump value for fresh.

2.2 Preparation of nano-recycle glass waste

The accumulation of green glass bottles in landfills necessitated the exploration of sustainable recycling methods, leading to the selection of this waste material for the study. The process of NRG preparation shown in Figure 1. The collected glass fragments underwent thorough cleaning to eliminate contaminants before processing. Initial crushing using an electrically powered primary crusher reduced the particle size to a range of 10-50 mm.

tu_pian_1.png

Figure 1. NRG proportion process

Further refinement of the recycled glass was performed using a laboratory high-energy planetary ball mill (500 mL capacity, stainless-steel jar with zirconia balls), operated in controlled cycles to progressively reduce particle size. The ground material was initially sieved through a No. 50 mesh (300 μm), and the retained fraction was subjected to additional milling. This milling-sieving process was repeated with progressively finer sieves (No. 100 mesh, 150 μm; No. 200 mesh, 75 μm) until a substantial portion of the powder achieved sub-75 μm size. The final particle size distribution was verified by standard sieving analysis to confirm attainment of the nano-scale range suitable for subsequent use in SCC mixtures.

To optimize the material for use in cementitious systems, the sieved glass powder was dried in a conventional oven at 120℃ for one hour to remove residual moisture, a critical step to prevent interference with subsequent processing. The dried powder was then reclassified using a sieve shaker to isolate particles below 75 μm, ensuring suitability for fine applications. The selected fraction was transferred to a high-energy ball mill, where it underwent intermittent milling cycles of two hours each, interspersed with drying phases at 120℃ to mitigate agglomeration and maintain powder homogeneity. This cyclical grinding-drying process was repeated over a total milling duration of 10 hours, ensuring gradual and uniform particle size reduction. The resulting NRG powder was characterized for consistency and fineness before incorporation into the experimental concrete mixtures.

Following the production stage, compositional analysis of the NRG was conducted using energy-dispersive X-ray spectroscopy (EDS). The examination revealed a weight composition of 32.6% carbon (C), 24.6% oxygen (O), 5.5% sodium (Na), 31.0% silicon (Si), and 6.3% calcium (Ca). The significant silicon content (31.0 wt.%) confirms silicon dioxide (SiO₂) as the primary phase, consistent with conventional glass composition, see Figure 2(a). The detected sodium (5.5 wt.%) and calcium (6.3 wt.%) concentrations indicate a soda-lime glass structure, commonly employed in commercial glass products. The oxygen content (24.6 wt.%) corresponds stoichiometrically to oxide phases, including SiO₂, Na₂O, and CaO. The significant carbon content (32.6 wt.%) likely originates from surface contamination, residual organic matter during sample preparation, or environmental exposure. Moreover, the prepared NRG exhibited a specific gravity of 2.52.

tu_pian_2.png

tu_pian_3.png

tu_pian_1.png

Microstructural characterization of the NRG was performed using dry-dispersion laser diffraction (Sympatec RODOS T4.1 Particle Size Analyzer) complemented by scanning electron microscopy (SEM) imaging. Figure 2(b) demonstrates a heterogeneous nanoparticle distribution, featuring irregular sharp-edged flakes, fragmented particles, and occasional needle-like crystals spanning nanoscale to microscale dimensions. This morphology arises from high-energy mechanical grinding and milling processes, which enhance specific surface area and impart unique physicochemical properties.

The X-ray diffraction (XRD) pattern of the prepared NRG, as shown in Figure 2(c), provides critical insight into the structural nature of the processed material. Utilizing Cu-Kα radiation, the XRD analysis reveals a broad diffraction hump, typically centered around 20°-30° 2θ, which is a well-known signature of amorphous silicate-based materials, including soda-lime glass. The absence of sharp, well-defined peaks indicates that the NRG does not contain any significant crystalline phases. This confirms that the glass retained its amorphous network structure throughout the crushing, milling, and drying processes. The preservation of amorphousness is beneficial when NRG is intended for use as a pozzolanic material or a cementitious additive, as amorphous silica tends to be more reactive in alkaline environments than its crystalline counterparts [34]. On the other hand, the broad hump also indirectly confirms that the high-energy ball milling and drying cycles were effective in reducing particle size without triggering devitrification (crystal formation). This is crucial: extended or aggressive thermal treatment or improper milling can sometimes cause partial crystallization in glass powders, which would lower their reactivity [35].

2.3 Aggregate

In this study, two types of coarse aggregate were used: Natural Coarse Aggregate (NCA) and RCA. The NCA consisted of crushed gravel with a maximum nominal size of 20 mm, sourced from the Euphrates River in Anbar Province, Iraq, while the RCA was derived from demolition waste of a local building. To maintain experimental uniformity, the RCA was processed to match the particle size distribution of the NCA. This ensured consistent gradation between the aggregates, eliminating particle size variations as a potential factor influencing concrete performance. The fine aggregate consisted of natural sand procured from the Al-Ekhaider quarry, with: maximum particle size, specific gravity, and water absorption of 4.75 mm, 2.63, and 1.5% respectively. The sieve analysis results with gradation limits of Iraqi Standard (IQS No. 45/1984) or ASTM C136/C136M and physical properties of NCA, RCA and fine aggregate are presented in Table 2. In this regard, the sieve analysis results clearly demonstrate compliance with the Iraqi Standard of coarse and fine aggregate sieve analysis (IQS No. 45/1984).

Table 2. Sieve analysis and physical properties of aggregate

A 30-liter laboratory mixing pan was employed for the preparation of all concrete mixtures, adhering to a standardized batching and mixing protocol. Initially, RCA, NCA, and fine aggregate were introduced into the mixer. Following homogenization, the binder components-OPC, FA, and NRG were incorporated. Approximately one-third of the total mixing water was then gradually added, with mixing continued for an additional minute. Subsequently, the remaining water and high-range water-reducing admixture (HRWRA) were introduced, followed by three minutes of continuous mixing. The mixture was then allowed to rest for two minutes before undergoing a final two-minute mixing cycle to ensure uniformity. The slump flow diameter of the SCC was designed to comply with EFNARC (2005) guidelines, targeting a range of 720 ± 20 mm. Trial batches were cast for each mixture type, with HRWRA dosages adjusted to achieve the specified slump flow. Once the workability of the SCC-NGRA was confirmed, the mixture was cast into molds (cylinders, cubes, and beams) without any compaction, as its self-weight ensured complete filling of voids. All specimens were demolded 24 hours after casting and subsequently cured in a temperature-controlled water tank until testing, see Figure 3.

tu_pian_5.png

Figure 3. Clarify mixing of SCC-NGRA production

The experimental workflow is effectively summarized in Figures 1 and 3. Figure 1 illustrates the sequential preparation of nano-recycled glass (NRG), while Figure 3 depicts the mixing, casting, and curing stages of SCC-NGRA. Together, these figures provide a visual representation of the entire methodology, serving the same purpose as a conventional flowchart and ensuring transparency of the process.

This section reports the experimental investigation of SCC-NGRA mixtures and interprets the relevance of the obtained results in the context of sustainable concrete development.

6.1 Fresh properties

In order to characterize the fresh properties of SCC-NGRA, standard procedures were applied, including the determination of slump flow diameter, slump flow time, V-funnel time, and L-box ratio.

6.1.1 Slump flow and workability

An ideal SCC-NGRA mix should exhibit excellent flowability and the ability to pass through congested reinforcement without blockage. Additionally, it must maintain a homogeneous distribution of aggregates while demonstrating high resistance to segregation. These properties are typically assessed through standardized testing procedures [41, 42]. Depending on the intended application, SCC-NGRA performance requirements can be selected from key fresh-state properties (i.e., flowability) and further defined by specific classification thresholds or target values, as outlined in Table 4 [43]. In this study, the inclusion of RCA at 0%, 50% and 100% replacement levels, along with NRG at 0%, 1%, 2%, and 3%, revealed distinct trends across key fresh state parameters - namely slump flow diameter, V-funnel time, slump flow time (T500), and L-box height ratio.

Table 4. Classification of SCC-NGRA properties (slump flow, viscosity, and passing ability) according to BS EN 206:2013+A1:2016 guidelines

As can be seen in Figure 5, the control mixture exhibited a slump flow diameter of 730 mm, which falls within Class SF2 according to BS EN 206:2013 + A1:2016 guidelines [43], suitable for elements with normal reinforcement congestion. Replacing natural aggregate with 50% and 100% RCA reduced the flow diameter to 725 mm and 715 mm, mainly due to the rough surface and irregular shape of RCA, which increases inter-particle friction and lowers mix lubricity. These observations are consistent with findings reported by Güneyisi et al. [44], who noted similar decrements in flow diameter with increasing RCA levels due to altered rheological response under low shear conditions. NRG incorporation further influenced the flow characteristics. Mixtures with 1% and 2% NRG exhibited minor reductions in flow, while 3% NRG mixtures marked a noticeable decline, with flow values approaching the lower boundary of SF2 classification. The angular morphology and high surface area of NRG particles promote water absorption and increase internal friction, contributing to reduced spread ability. This behavior parallels the observations by Assaad and Khayat [45], who linked fine pozzolanic filler addition with lower workability in SCC due to water demand escalation.

tu_pian_12.png

Figure 5. Slump flow diameter of SCC-NGRA mixtures with and without RCA and NRG modification

Simultaneous use of RCA and NRG caused a greater reduction in flowability than when each was added alone (Figure 5). For example, while the fraction of 100% RCA reduced flow diameter by about 2.1% relative to the control, and using 3% NRG alone led to decreased flow diameter by about 2.7%, the combined 100% RCA 3% NRG mixture further decreased flowability by about a 4.1%. This trend suggests a synergistic effect, where the adverse impact of RCA on mix lubricity is amplified by the fine, high surface area NRG particles that further absorb mixing water due to its pozzolanic surface reactivity and restrict cement paste movement. Mechanistically, this interaction can be attributed to the dual increase in internal friction: RCA contributes macro-scale friction due to its angularity and rough texture, while NRG particles influence micro-scale rheology by thickening the cementitious matrix. As a result, the paste’s ability to effectively coat and separate aggregate particles diminishes, thereby restricting outward flow during slump testing. This compounded effect corroborates the rheological behavior reported by Zhang et al. [46], who observed that hybrid substitution with recycled and nanostructured materials often demands higher plasticizer dosages to maintain baseline workability. On the other hand, the complex interaction between RCA and NRG highlights the need for rheological fine tuning in mixes targeting sustainability through dual replacement strategies. As demonstrated in a recent study by Mohseni et al. [47], the combined use of recycled and nanomaterial-based modifiers introduces unique challenges in maintaining flow class compliance while also targeting durability and performance metrics. Notably, the reduction in slump flow across mixed RCA and NRG combinations still kept the values within the SF2 classification.

6.1.2 Viscosity indicators

The results of T500 time (indicative of flow rate) and V-funnel time are presented in Figures 6(a) and (b), respectively. As shown in Figure 6(a), the T500 time increased progressively with higher RCA and NRG content. Control mixture T500 was recorded at 2.01s, while mixes made with 100% RCA and 3% NRG reached up to 3.02 s, reflecting increased viscosity. On the same line, Figure 6(b) shows V-funnel time followed a similar trend, increasing from 6.51 s in the control mixture to 11.0 s in the RCA 100% NRG 3% mix. These findings suggest that the combined inclusion of RA and nano additives increases flow resistance, particularly due to the cumulative effect of surface roughness and particle fineness. This response is primarily rooted in the rheological incompatibility introduced by dual-scale surface heterogeneity: RCA disrupts uniform particle packing and fluid film continuity, while NRG alters paste rheology through microstructural densification. Together, they hinder flow by limiting particle dispersion and reducing lubrication. Such dual-material interactions were also documented by Liu et al. [48], who noted a nonlinear rise in plastic viscosity when coarse recycled particles were paired with mineral fines exhibiting high surface activity. Moreover, a subtle but notable shift in the relative contribution of NRG becomes apparent when RCA levels are high. At 100% RCA, the incremental addition of NRG (from 1% to 3%) yields sharper increases in both T₅₀₀ and V-funnel times compared to mixes without RCA. This suggests that NRG’s influence is amplified in coarse, high-void frameworks, where paste continuity is already compromised. It highlights the importance of mix synergy analysis rather than isolated component optimization in sustainable SCC-NGRA design.

tu_pian_24.png

(a)

tu_pian_25.png

(b)

Figure 6. Time-dependent rheological characteristics of SCC-NGRA mixtures: (a) T₅₀₀ and (b) V-funnel test results

According to BS EN 206:2013+A1:2016 guidelines, these T500 and V-funnel time values correspond predominantly to higher viscosity classifications, necessitating careful admixture tuning to maintain acceptable workability. Notably, even at elevated NG dosages of current study, the mixes remained within functional viscosity limits lower than 12 s for V-funnel, ensuring usability for vertical and horizontal placements. Recent literature also supports this trend; Altayawı et al. [49] emphasized that fine pozzolans, while enhancing long-term performance, often reduce flow speed and necessitate mix optimization.

6.1.3 Passing ability

The L-box test, which assesses passing ability by simulating reinforcement conditions, showed a consistent decrease in the H₂/H₁ ratio as the content of RCA and NRG increased (see Figure 7). The control exhibited near ideal flow (0.99), whereas the 100RA30NG mixture dropped to 0.86. All mixes, however, remained above the minimum threshold of 0.80 prescribed by BS EN 206:2013 + A1:2016 guidelines confirming adequate passing capacity. In this regard, such trend can be attributed to a dual-phase hindrance effect: RCA imposes geometric blockages due to its angularity and non-uniform size, while NRG thickens the interstitial paste, diminishing the ability of aggregates to reorient and move freely during narrow passage flow. The compounding nature of these factors restricts the mixture’s ability to pass through reinforcement simulations without segregation or blockage. These findings align with observations made by Liu et al. [48], who highlighted the trade-off between pozzolanic enhancement and passing ability in nanomaterial-modified SCC. Similarly, Altayawı et al. [49] reported a decline in the passing ability of ternary SCC mixes incorporating RA and nano-silica blends.

tu_pian_26.png

Figure 7. Variation of L-box height ratio for SCC-NGRA mixtures

6.1.4 Flow-viscosity-passing correlations

To gain deeper insight into the interdependency of fresh-state parameters, a composite figure was developed to illustrate the relationship between T500, V-funnel flow time, and H2/H1 ratio (Figure 8). The plotted data reveal a clear trend: increasing T500 correlates with a simultaneous rise in V-funnel time and a gradual decline in L-box ratio across all mixtures. This tri-directional interaction underscores a shared rheological underpinning where higher internal viscosity whether induced by RCA angularity or NRG powder fineness slows flow and restricts movement through confined spaces. Notably, mixes containing both high RCA and NRG levels cluster toward the higher end of the V-funnel axis and the lower end of the L-box axis, reflecting the most significant resistance to flow and obstruction in passing. The inverse trend between L-box and the other two parameters emphasizes the sensitivity of passing ability to changes in viscosity and yield stress, as thicker pastes impede aggregate mobility under restricted conditions. Despite these reductions, all mixtures remained within the performance thresholds prescribed by BS EN 206:2013+A1:2016, affirming their suitability for structural applications with careful proportioning. The combined figure thus provides a compact yet comprehensive visual synthesis of how modifications to material composition affect flow characteristics, aligning with findings reported in recent multivariable SCC rheology study by Wang et al. [50].

tu_pian_27.png

Figure 8. Integrated assessment of viscosity and passing ability in SCC-NGRA: correlation between T₅₀₀, V-Funnel time, and cc ratio

To further elucidate the rheological behavior of the modified SCC-NGRA mixtures, a correlation plot was developed linking slump flow diameter with V-funnel time and H₂/H₁ ratio (Figure 9). The resulting relationships illustrate that as slump flow diameter decreases, V-funnel time increases, indicating a consistent reduction in flowability accompanied by a rise in mix viscosity. Concurrently, the H₂/H₁ ratio trends downward, reflecting diminished passing ability. This inverse correlation suggests that mixes with limited spread ability tend to resist flow under gravity and are less capable of navigating congested formwork, likely due to paste stiffening and aggregate interlock.

tu_pian_28.png

Figure 9. Relationship between flowability, viscosity, and passing ability in SCC-NGRA

This combined plot effectively demonstrates the co-dependence of fresh-state properties in SCC-NGRA systems incorporating RCA and NRG powder. The decrease in slump diameter is accompanied by reduced mobility of the concrete, indicating that the mix becomes more resistant to movement. This is likely due to the combined effects of the rough surface texture of the RCA, which increases internal friction, and the fine particles of NRG powder, which make the paste denser and less able to flow freely around the aggregates. This graphical analysis aligns with findings by Lee et al. [17], who emphasized the necessity of multi-parameter assessment when designing SCC for hybrid material systems with different replacement levels.

6.2 Hardened properties

The mechanical response of the investigated SCC-NGRA mixtures was evaluated in terms of compressive strength, splitting tensile strength, and flexural strength (modulus of rupture). In addition to direct experimental testing, the study further explored the predictive reliability of existing international models and proposed regression-based correlations. Comparative assessments were performed across all strength parameters to highlight the influence of mix modifications and substitution levels.

6.2.1 Compressive strength development

The compressive strength of all concrete mixtures was evaluated at 28 and 90 days, and the results are presented in Figure 10. The control mixture reached 39.31 MPa at 28 days and 47.41 MPa at 90 days. Incorporating RCA and supplementary cementitious materials led to notable variations in strength development. For instance, the mixture containing 50% RCA showed a reduction in 28-day strength by approximately 11.9% compared to the control, which is attributed to the weaker ITZ and higher porosity typically associated with RCAs. However, extending the curing age to 90 days mitigated this reduction, indicating a positive long-term pozzolanic contribution from the additives. The general trend suggests that the compressive strength of mixtures incorporating RCA and NRG improves with curing time, progressively narrowing the performance gap with the control mixture. This enhancement can be attributed to the continued hydration of unreacted cement particles and the progressive densification of the matrix. The finely ground glass, acting as a pozzolanic material, reacts with calcium hydroxide to form additional C-S-H, a secondary product which contributes to filling micro-voids and refining the pore structure [51]. This effect becomes more apparent with increasing NRG content, as the additional C-S-H enhances internal packing and strengthens the ITZ [52]. The positive impact is particularly noticeable at later ages, highlighting the long-term reactivity and microstructural benefits provided by NRG.

tu_pian_29.png

Figure 10. Development of compressive strength in RCA, NRG modified SCC-NGRA. Variability indicated by ACI 214R-1, based error bars. Normalized 28 and 90 days deviations (%) are labeled

Incorporating CV-based error bars enhances the clarity and technical reliability of the data by quantifying inherent testing uncertainty and supporting the visual comparison of mix performance. It enables researchers and practitioners to discern whether observed differences between mixtures are likely due to material modifications or fall within acceptable experimental variability. Furthermore, such estimation allows for benchmarking the quality control level of the mixes and reinforces the reproducibility of results, which is critical for both academic validation and field application. Error bars shown in Figure 10 were calculated following the guidelines of ACI 214R-11 [53, 54], based on an assumed CV of 7.5%, which is appropriate for concrete strengths within the 35-55 MPa range. In addition, the normalized values that represent the percentage deviation from the control mixture are marked above 28 and 90 days. These values serve as a critical lens to assess the influence of the incorporation of RCA and NRG on mechanical performance. At 28 days, a clear reduction in compressive strength was observed in the mixtures containing RCA. The 50% RCA and 100% RCA mixtures exhibited strength losses of approximately 11.9% and 20.3%, respectively, in comparison to the control. This decline can be attributed to the inherent porosity and weaker ITZs commonly associated with RCA. However, the incorporation of NRG demonstrated a compensatory effect. The 1% NRG mixture exceeded the control’s performance by 5.3%, indicating enhanced hydration and matrix refinement due to pozzolanic activity. Moreover, the hybrid mixture (50%RCA1%NRG) achieved a more moderate reduction of 6.4%, showcasing the synergistic interaction between partial RCA replacement and NRG supplementation. At 90 days, the positive trend continued across all formulations. The 50% RCA mixture increased its performance to a percent of 9.0%, while the 100% RCA mixture showed a slight positive deviation of 1.6%. Notably, the mixture with 3% NRG exhibited a 38.7% strength gain, reinforcing the long-term reactivity of NRG. On the other hand, the 100%RCA1%NRG mix exceeded the control mixture, with a deviation of 21.8%, suggesting that prolonged curing allows NRG to fully engage in secondary hydration reactions that contribute to improved packing density and strength.

6.2.2 Splitting tensile strength

The splitting tensile strength results are presented in Figure 11 for both 28 and 90 days curing periods. Overall, tensile performance exhibited more sensitivity to compositional changes than compressive strength, reflecting the higher dependency of tensile failure on microstructural continuity and interfacial quality. At 28 days, mixtures containing RCA exhibited notable reductions in tensile strength relative to the control. For example, the 50% RCA mixture experienced an approximate 12.2% drop, which aligns with the expected decline in tensile capacity due to increased porosity, weaker particle bonding, and potential residual micro-cracks introduced during aggregate recycling [55].

Interestingly, mixtures incorporating NRG demonstrated an improved tensile strength recovery over time. At later curing ages, particularly in specimens containing 2% and 3% NRG, the observed tensile gains suggest the development of a more integrated and cohesive internal structure. This can be attributed to the pozzolanic reactivity of the NRG, which contributes to the formation of additional C-S-H gel that not only densifies the matrix but also bridges microcracks [56, 57]. This crack-bridging effect appears to be especially beneficial under tensile loading conditions, where localized weaknesses can lead to brittle failure. In this regard, there is a strong agreement between the current experimental trends and previous studies that used recycled glass and other pozzolans in the production of conventional concrete. Their results showed that the incorporation of such supplementary cementitious materials confers significant splitting tensile strength benefits to recycled aggregate concrete by tightening its internal structure, especially at the aggregate paste interface, providing agents for micro-crack bridging, and facilitating continued strength development beyond the standard 28 days period [58, 59].

tu_pian_30.png

Figure 11. Development of splitting tensile strength in RCA, NRG modified SCC-NGRA. Variability indicated by ACI 214R-11, based on error bars. Normalized 28 and 90 days deviations (%) are labeled

As with the compressive strength results, the error bars depicted in Figure 11 were estimated using a coefficient of variation (CV) of 8.0%, consistent with the variability range recommended by ACI 214R-11 for tensile strength in conventional concrete [53, 54]. These statistical bounds support a more transparent interpretation of data reliability and experimental consistency across all tested mixtures. Integrating CV-based variability into the visualization strengthens the analytical depth, especially when comparing the performance of recycled and modified concretes against the control mix. Moreover, normalized deviation values annotated above each 28 and 90 days bar serve to highlight the relative performance of each formulation. At 28 days, the 50% RCA mixture showed an 18.4% reduction in tensile strength compared to the control, while 100% RCA exhibited a greater decline of 19.2%. Conversely, the inclusion of 1% NRG alone improved the tensile strength by 4.3%. The hybrid SCC-NGRA mixture, namely 50%RCA1%NRG, demonstrated a moderated reduction of 10.4%, reflecting partial compensation by the pozzolanic effect of NRG. A similar trend was observed at 90 days, where strength gains were recorded across all mixtures, reducing the performance gap with the control. Specifically, the normalized loss for 50% RCA improved to 15.3%, while 100% RCA reduced its deficit to 29.7%. The mix with 3% NRG alone further enhanced its performance, reaching a 15.9% increase over the control, whereas the 50%RCA3%NRG mixture yielded a deviation of 12.4% compared to the reference SCC-NGRA mixture. These findings suggest that long-term curing allows for the progressive reactivity of NRG to manifest more fully, contributing to matrix densification and improved tensile integrity. When combined with CV-derived error bands, this dual-layer visualization approach delivers a balanced view of strength development, uncertainty, and material innovation in sustainable concrete systems.

The observed improvements in compressive and tensile strength with NRG addition can be attributed to microstructural modifications revealed by SEM. The nano-glass particles, due to their pozzolanic reactivity and filler effect, contribute to a denser matrix by refining pore structure and reducing voids within the ITZ. This densification enhances load transfer across the paste-aggregate interface, mitigating the common weaknesses associated with RA. Similar mechanisms have been reported by Dahamm et al. [23] and Onaizi et al. [24], who found that finely ground glass powders improved matrix homogeneity and bond strength. Compared with these studies, the present work demonstrates that even at low replacement levels (≤ 3%), NRG can partially offset the reduction in mechanical properties caused by RCA, thereby highlighting its synergistic role in sustainable SCC production.

6.2.3 Flexural tensile strength

Flexural tensile strength is a critical indicator of a concrete mixture’s ability to resist bending and crack propagation under transverse loading, making it particularly relevant for structural applications such as beams, pavements, and slabs [60]. Unlike compressive or splitting tensile strength, flexural performance captures the combined influence of matrix cohesion, interfacial bonding, and post-cracking ductility [61]. Assessing this property at later ages is essential for understanding the long-term mechanical behavior of concretes incorporating RA and supplementary cementitious materials, especially in systems where microstructural refinement and delayed pozzolanic activity play a significant role [62].

The evaluation of flexural tensile strength at 90 days offers further insight into the structural integrity of concrete mixtures modified with RCA and NRG. As shown in Figure 12, a descending trend in strength is evident with increasing RCA content, while NRG incorporation yields notable improvements. The control mixture recorded a flexural strength of 3.78 MPa, against which all other mixtures were benchmarked. The 50% RCA and 100% RCA mixtures showed reductions of approximately 22.0% and 40.2%, respectively, relative to the control. These losses are likely attributed to diminished matrix continuity and stress-transfer efficiency due to higher porosity and residual microdefects in RA [54, 63-68]. On the contrary, the 1% NRG mixture outperformed the control, registering a 22.0% increase in flexural strength. The improvement stems from NRG’s pozzolanic reactivity, which refines pores, bridges cracks, and strengthens tensile resistance. The hybrid formulation (50%RCA1%NRG) mitigated much of the RCA induced degradation, showing only a 7.4% reduction relative to the control, indicating a partial recovery of tensile integrity through NRG supplementation. To effectively convey these trends, a grouped bar chart with superimposed normalized deviation annotations is recommended. Each bar would represent the average flexural strength, while the normalized percentage differences would be labeled above the bars. Error bars can be added based on an assumed coefficient of CV of 9.0%, which aligns with ACI 214R-11 guidelines for flexural tests in normal-strength concrete [53].

tu_pian_31.png

Figure 12. Development of flexural strength in RCA, NRG modified SCC-NGRA. Variability indicated by ACI 214R-11, based error bars. Normalized 90 days deviations (%) are labeled

6.3 Strength correlation and benchmarking

To extend the understanding of mechanical behavior beyond individual strength metrics, this section presents a comprehensive evaluation of 90 days hardened properties through statistical correlation and comparative benchmarking. It is noteworthy that flexural strength was exclusively measured at 90 days; therefore, this section intentionally focuses on 90 days test results only. The evaluated concrete mixtures are all based on SCC-NGRA, which is characterized by its flowability, high paste volume, and reduced compaction energy requirements actors that may influence mechanical correlations and performance trends.

6.3.1 Correlation of compressive strength with splitting tensile strength and flexural strengths

The relationship between compressive strength and tensile-related properties is critical for predicting concrete performance. In this regard, splitting tensile and flexural tensile strengths were analytically correlated with compressive strength through regression modeling, enabling the derivation of empirical relationships characterizing the mechanical behavior of the SCC-NGRA mixtures (see Figure 13). A pronounced linear correlation was observed between compressive strength and splitting tensile strength, revealing a strong predictive linkage across the SCC-NGRA mixtures. This trend suggests that compressive strength serves as a reliable indicator of tensile performance, likely due to shared dependence on matrix integrity and internal load distribution mechanisms. The strength correlation is consistent with previous reports in the literature. For example, Şahmaran et al. [64] found similar proportionality between compressive and tensile strength in self-consolidating systems incorporating supplementary cementitious materials. In contrast, the relationship between compressive strength and flexural strength followed an exponential trend. This may reflect the increasingly brittle to ductile transition behavior of SCC-NGRA as compressive strength increases, particularly when nano-sized pozzolanic additives such as NRG are present. The exponential model also highlights the enhanced post cracking performance observed in the strength of SCC-NGRA, corroborating findings by Mustapha et al. [65], who demonstrated similar trends when nano-silica was used to improve flexural capacity. Notably, the higher R² value for flexural strength correlation implies a more predictable response under bending conditions, possibly due to the beneficial matrix densification and crack-bridging effects imparted by NRG particles. A similar trend was reported by Li et al. [66], who demonstrated that incorporating nano-silica into SCC mixtures enhanced flexural capacity through matrix densification and crack-bridging mechanisms, even when improvements in compressive strength remained modest.

tu_pian_32.png

Figure 13. Correlation of compressive strength with splitting tensile and flexural tensile strengths in SCC-NGRA mixtures

Overall, these correlations provide useful predictive tools and validate the composite behavior of SCC mixtures modified with sustainable materials. Their high degree of fit confirms the internal consistency of the experimental data and sets a foundation for engineering level design estimations in future SCC applications.

6.3.2 Normalized mechanical index

To quantitatively benchmark the overall mechanical performance of each SCC-NGRA mixture, a Normalized Mechanical Index (NMI) was developed based on compressive strength, splitting tensile strength, and flexural tensile strength results. Each value was normalized relative to the control, and the NMI was computed as the arithmetic mean of the three normalized parameters (see Figure 14).

tu_pian_33.png

Figure 14. Normalized mechanical index

The presented behavior reveals a distinct stratification among SCC-NGRA mixtures incorporating RCA, NRG, or their combination. Mixtures containing only RCA showed markedly reduced performance. Notably, the 100% RCA mixture yielded the lowest index (70.4%), followed by 50% RCA at 83.3%, confirming that higher substitution levels of NCA with RCA diminish matrix continuity and microstructural cohesion. This degradation is consistent with the findings of Etxeberria et al. [67], who reported that RCA weakens the ITZ and increases porosity in SCC matrices. In contrast, mixtures incorporating NRG demonstrated significant performance enhancement. The 3% NRG mixture achieved the highest NMI (126.7%), with 2% NRG closely behind at 124.7%, clearly surpassing the control. These improvements are attributed to the pozzolanic reactivity and filler effect of NRG, which promote matrix densification and secondary CSH formation. Similar enhancements have been documented for nano-silica systems in SCC by the study [65]. Mixtures combining RCA and NRG exhibited intermediate performance. For instance, the inclusion of 1% NRG raised the NMI of 100% RCA from 70.4% to 79.3% (100%RCA1%NRG), while 3% NRG increased it further to 87.7% (100%RCA3%NRG). This suggests that NRG can partially mitigate the mechanical drawbacks of RCA, though the improvement is proportional to the NRG dosage and does not fully restore strength to control levels.

It is also noteworthy that NRG enriched mixtures exhibited relatively higher improvements in flexural strength than in direct tensile strength, highlighting NRG’s stronger influence on crack bridging, post peak ductility, and energy absorption. This observation is in agreement with earlier findings in Section 3.2. In addition, a comparable enhancement in flexural performance was reported by Ahmed et al. [68], who demonstrated that incorporating nano-silica into lightweight structural concrete mixtures significantly improved flexural strength. This improvement was attributed to the densification of the cement matrix and enhanced microstructural integrity, even when utilizing RA.