Effect of alkali–silica reaction on the mechanical properties of concrete with and without pozzolan Open Access

Concrete is a fundamental building material known for its durability and high compressive strength (Matos et al., 2022). However, its longevity and performance can be significantly compromised by internal expansion reactions generated by the interaction between cement alkalis and reactive aggregates. These reactions involve chemical processes that induce material deterioration, a process that depends on the aggregate reactivity – moderate or high (Neville, 2013; Marzouk and Langdon, 2003; Giaccio et al., 2008), quantity, size and particle size distribution (Multon et al., 2009, 2010), the amount of available moisture alkalis (Multon et al., 2009), material porosity and pre-existence of micro cracks (Tosun et al., 2007). The damage imposed by the internal expansions can cause a significant deterioration process in concrete structures throughout their useful life. Among the different types of chemical reactions that can affect the quality and durability of concrete structures, alkaline reactions stand out, such as the alkali-silica reaction (ASR). The occurrence of ASR has the potential to modify the mechanical characteristics of concrete by inducing alterations in its internal structure and giving rise to delayed cracks. Pozzolanic materials are widely used to mitigate the effects of ASR due to their ability to consume calcium hydroxide through secondary hydration reactions, reducing the availability of alkalis in the pore solution and refining the microstructure (Neamat et al., 2025). However, not all pozzolanic cements are equally effective, especially when the overall alkali equivalent content remains high. This study addresses this gap by comparing concrete with and without pozzolan under controlled ASR conditions, analyzing the unexpected deterioration pattern observed in pozzolanic cement with elevated alkali equivalent content.

Several studies have investigated the impact of ASR on various mechanical properties of concrete, including compressive strength, modulus of elasticity, flexural strength (modulus of rupture) and tensile strength (Ramadan et al., 2025; Matos et al., 2022). These studies have consistently shown that ASR can significantly reduce the strength and deformation properties of structural concrete, but the magnitude of the reduction is not uniform across all its characteristics.

The effect of ASR on the compressive strength of concrete is a function of time. It has been found that compressive strength decreases as damage due to the internal expansions increases at the micro-structural level. Many researchers reported that the loss in compressive strength can be as high as 40–60%, with a reduction of 20% likely for expansion rates found in engineering practice (Marzouk and Chen, 1995). The authors showed that compressive strength losses highly depended on several variables, such as mix design, aggregate type and storage conditions. More recently, Sanchez et al. (2018) showed that compressive strength is little influenced by ASR internal expansions, considering high expansion levels (0.30%). The concrete exhibited compressive strength loss of around 0.29%–0.43% when it had delayed ettringite formation (DEF) combined with ASR or only DEF. The author also reported that for low expansion levels (i.e. 0.05%), tensile strength losses ranged from 30% to ≈60%. For expansion levels between 0.20 and 0.30%, tensile strength losses ranged from 55% to ≈60%.

Another important variable associated with ASR is the effect caused by this reaction on the tensile strength of the concrete, which stems from the swelling phenomenon of the ASR gel. Smaoui et al. (2005) showed that this mechanical property of the material is often strongly affected by internal expansion due to gel swelling rather than compressive strength.

Some authors studied the behaviour of normal and high strength concrete specimens in direct tension loading. Uniaxial tests are often used to determine the stress–strain relationship because they describe the mechanical behaviour of concrete in tension, shear and bond more accurately than indirect tension tests (Marzouk and Langdon, 2003). The authors also observed that the direct tensile strength of high-strength concrete was 4–5% of the cylinder’s direct compressive strength (fc0). The direct tensile strength of the normal-strength concrete sample was found to be approximately 8–10% of fc₀. The direct tension test findings showed that the ASR gel was more susceptible to changes than the compressive strength ones.

The damage induced by internal expansion reactions has a more pronounced effect on the direct tensile strength of concrete and can significantly reduce the load-bearing capacity of the material, which depends on this strength (Blight and Alexander, 2011). On the other hand, it should be highlighted that assessing the tensile strength of concrete is not an easy task, given the availability of various methodologies that provide different results.

Marzouk and Chen (1995) found that after 12 weeks of exposure, the tensile strength of normal-strength concrete specimens containing highly reactive aggregates or moderately reactive aggregates decreased. The tensile strength decreased by 37% and 31% in specimens made with extremely and moderately reactive aggregates, respectively, after being exposed to a NaOH solution. Tensile strength was not reduced in specimens made with moderately reactive aggregates and placed in deionized water for 12 weeks. Specimens made with highly reactive aggregates, on the other hand, showed a 7% loss in tensile strength over a 12-week test period. This small drop can be attributed to the significant variability of the concrete’s tensile strength values.

Sanchez et al. (2020) and Gautam et al. (2015) showed that when DEF occurs combined with ASR in concrete, stiffness can decrease by 50%–65%, and in cases where only ASR is present, the reduction in static modulus of elasticity can reach up to 67%. Additionally, reactive aggregates can contribute to concrete expansion when DEF is involved. Gautam et al. (2015) studied the effect of multiaxial stress on the expansion and mechanical property degradation of ASR-affected concrete and reported average decreases of 20% in the modulus of elasticity of the concrete in the test specimens.

Expansions caused by interactions between reactive aggregates and cement alkalis play a significant role in the ageing and degradation of concrete structures. Understanding how these expansions affect the strength and deformation properties of structural concrete remains an outstanding question in scientific research. The study focused on evaluating the effects of internal expansions, specifically due to ASR, on the mechanical and deformation properties of structural concrete, with and without pozzolanic materials known for enhancing durability and long-term performance, to analyze the influence of these reactions on the concrete’s structural integrity.

The study focused on evaluating the effects of internal expansions, specifically due to ASR, on the mechanical and deformation properties of structural concrete, with and without pozzolanic materials known for enhancing durability and long-term performance, to analyze the influence of these reactions on the concrete’s structural integrity. ASR mechanism involves a chemical reaction between the hydroxyl ions (OH⁻) in the alkaline pore solution of cement paste and reactive silica found in some aggregates. This reaction produces an expansive alkali–silica gel, which absorbs water and swells. The expansion generates internal tensile stresses that lead to the initiation and propagation of microcracks within the cement matrix. These microcracks, particularly in the interfacial transition zones, may not always result in compressive and tensile strengths loss but drastically reduce the material’s stiffness. As a result, the modulus of elasticity is often more sensitive to early-stage ASR than other mechanical properties.

An extensive experimental campaign was performed to assess the effects of internal expansion reactions on the strength and deformation properties of structural concrete due to ASR. A total of 240 standardized cylindrical concrete specimens (200 mm height and 100 mm diameter) were produced with cement of two types: one with added pozzolanic material and the other produced primarily by clinker and gypsum. These samples were subjected to diametrical compressive strength, axial tensile strength and axial compressive strength tests under normal conditions (60 samples of PC-IV and 60 of PC-V), and in an environment with conditions to accelerate the ASR process (60 of PC-IV and 60 of PC-V). Figure 1 illustrates the process described (120 concrete samples were used as reference samples for the mechanical tests performed).

The fine aggregate utilized in the concrete mixture was natural, non-reactive sand. For the physical characterization, the following tests were performed: (1) maximum characteristic size: 4.75 mm; (2) determination of fine material passing through the 75 µm sieve: 2.04%; (3) determination of clay content in clods and friable materials: 0.57%; (4) fineness module: 2.74; (5) pulverulent material: 1.41 and (6) silica reactivity: Innocuous.

To assess the reactivity of the fine aggregate, an accelerated test on mortar bars was performed following the recommendations of the Brazilian standard NBR 15577-4 (2008). The aggregate was prepared according to usual standards. The procedure involved molding three mortar bars with a water–cement ratio of 0.47 by mass, using 440 g of cement and 990 g of fine aggregate, and the specimens were cast in prismatic molds measuring 25 × 25 × 285 mm³. The test results demonstrated that the fine aggregate used did not exhibit any reactive potential and could be classified as innocuous concerning internal expansion reactions, such as ASR.

The coarse aggregate used comes from a quarry located on the BR-232 highway, in Vitória de Santo Antão, Pernambuco. For the physical characterization, particle size and specific mass tests were carried out, with the following results obtained: (1) Maximum characteristic size: 19.0 mm; and (2) Unit mass in the dry compacted state: 1.379 kg/l.

Petrographic analyses were performed (NBR 15577-3, 2008) on thin blades with a thickness of approximately 30 µm, prepared from the rock that constituted the aggregate, and also in mortar bars, examined using an integrated Olympus BX41 microscope. Mesoscopic analysis of the material showed a dark grey rock with light grey to white spots, represented by prismatic and/or rounded feldspar phenocrysts, most of which are aligned according to the rock’s foliated structure, defined not only by the presence of feldspar phenocrysts but also by dark mica lamellae (Figure 2), that is the rock analyzed falls into the category of “Potentially Reactive Rocks Containing Quartz” (Fettes and Desmons, 2007).

In this research, two types of Portland cement were used: PC-V, and PC-IV, with a higher concentration of pozzolanic additives, which can include natural or industrial by-products such as fly ash, granulated blast furnace slag and active silica, among others (Table 1). The incorporation of pozzolanic materials offers several advantages for concrete, such as pore refinement, improved mechanical properties such as compressive strength and increased cohesion of the cement paste by filling empty spaces that would otherwise be occupied by water.

The design compressive characteristic strength of the concrete was set at 35 MPa, a representative value of the practice of concrete construction in the Pernambuco region (Brazil). The concrete mix ratio (1.0:1.1:2.1) was determined using the ACI Committee 211 method (2012), and the variable values used to make all concrete batches (PC-IV and PC-V) are listed below: (1) slump range value: 10 ± 20 mm; (2) maximum characteristic size of the aggregate: 19 mm; (3) mixing water consumption and air content: 224 kg/m³ with 1% of air content; (4) water-cement ratio (w/c): 0.45; (5) cement consumption: 497 kg/m³; (6) coarse aggregate consumption: 1.043 kg/m³ and (7) fine aggregate consumption: 546 kg/m³.

To create a favorable environment for the initiation and propagation of ASR, a tank made of high-temperature-resistant material was used to store all concrete specimens. A high-density polyethylene plate with holes was placed at the bottom to prevent direct contact between the specimens and the tank surfaces. A heating element and thermostat were installed to maintain a consistent temperature of 60 ± 2°C (Figure 3). The procedure to simulate internal expansions due to the alkali–silica reaction in the concrete specimens involved the following steps: (1) after casting, the specimens were immersed in a 1 N NaOH solution at 80°C for 30 days; (2) after this period, the specimens were submerged in a metal tank containing a 1 N NaOH solution at a constant temperature of 60°C for 242 days, accelerating the onset of expansion in the concrete prisms (RILEM AAR-2, 2016).

A high-density polyethylene shield was placed at the bottom of the tank structure to prevent the specimens from coming into contact with the material in the tank. Regular checks on the level of the sodium hydroxide solution inside the tank and the temperature of the internal environment were performed every two weeks. This was necessary to ensure that the initial conditions of the tests would not change over time. Finally, after a curing process of 242 days at 60°C, it was possible to observe that the ASR pathology affected approximately 49 PC-IV specimens and 51 PC-V specimens.

Inspections of the concrete specimens showed visual evidence of cracks and silica gel deposits, indicative of the start of ASR. The minimal cracking patterns observed suggest a potential scenario of moderate reaction progression, that is the result of the C-S-H gel occupying the voids, generating tensile stresses that induce the micro-cracks observed (Figure 4).

Mortar bar tests were conducted on both fine and coarse aggregates, and the experimental setup involved fully immersing the mortar bars, which contained the fine and coarse aggregates combined with a high-alkali reference cement, in a 1 N NaOH solution at 80°C. The purpose of this immersion was to simulate ASR conditions. Periodic length measurements of the mortar bars were taken, starting 24 h after immersion. Additional readings were recorded every two days until day 35, at which point the accelerated mortar bar test was completed.

For each aggregate type, three prismatic mortar bars were prepared. Each bar had a square cross-section measuring 25.0 ± 0.7 mm on each side and a length of 250.0 ± 2.5 mm, ensuring consistent dimensions across samples. Ordinary Portland Cement was used in the mix, with a Na₂O equivalent (Na₂O + 0.658 K₂O) of at least 1%, ensuring a high alkali content to promote ASR reactions. The fine and coarse aggregates were processed by crushing and sieving to obtain the appropriate particle size distribution, following the guidelines provided by RILEM AAR-2 (2016). Both fine and coarse aggregates were subjected to controlled preparation to meet the standard grading requirements and ensure uniformity in the experiment.

The results indicate that only the coarse aggregates showed a potential reactivity level of R1, which implies expansions in the range of 0.19%–0.4% in 30 days. In contrast, the fine aggregate showed no reactivity potential. Consequently, the only source of chemical reactivity in the concrete produced comes from the coarse aggregates used (Figure 5).

Figure 5 also shows that at 16 days, an expansion of 0.19% was reached, indicating that the coarse aggregates used can be classified as rapidly reactive. Typically, an expansion of approximately 0.10% at 14 days is sufficient to categorize an aggregate as rapidly reactive, as noted by Alejos et al. (2014). Therefore, the coarse aggregate used in the research demonstrates very rapid reactivity. Some standards allow for the continuation of accelerated mortar bar tests up to 30 days when expansions at 14 days fall within the range of 0.10%–0.20%. In such cases, an expansion threshold of 0.19% is commonly used to classify an aggregate as reactive. This further validates the coarse aggregate’s classification as highly reactive, which requires attention to potential alkali–silica reactivity risks in concrete applications.

In contrast, the mortar bars containing fine aggregates displayed minimal linear length variations of 0.02% at 14 days and 0.06% at 28 days. These slight expansions are considered non-detrimental and do not pose a significant risk when these fine aggregates are used in concrete structures.

The compressive strength tests for the concrete samples, affected and not affected by the internal expansion reactions due to ASR, were carried out following the NBR 5738 (2016) and NBR 5739 (2018) standards. The height/diameter ratio of the specimens was respected in all the tests carried out. Tensile tests were performed following the recommendations of NBR 7222 (2011). The method, also known as the split-cylinder test, is a common and widely used approach to evaluating the tensile strength of concrete.

Finally, the static modulus of elasticity of the test specimens was determined using static tests (NBR 8522-1, 2021), with the initial tangent modulus being taken into account in this assessment.

Figure 6 and Table 2 present the compressive strength results obtained and the dispersion measures for the samples analyzed, as well as the number of specimens tested in each condition. In addition, Table 2 also shows the lower and upper limits of each sample, calculated using the interquartile range (IQR) concept, a method for identifying outliers. In all the samples analyzed, only one outlier was observed.

Table 3 presents a comparative summary of compressive and tensile strength losses reported in the literature under ASR exposure, together with the findings from this study. It can be observed that the compressive strength reductions for PC-IV and PC-V fall within the lower range of reported values for similar expansion levels. Conversely, the tensile strength reduction in PC-V is comparatively higher, suggesting variability depending on cement composition. These findings reinforce the moderate expansion level induced in this study and highlight the relevance of cement chemistry in ASR progression.

Using the average compressive strength values of the tested specimens, the reductions attributed to ASR were −18.58% for PC-IV and −6.82% for PC-V, indicating that the loss in compressive strength for PC-IV was approximately 2.7 times greater than for PC-V. At first glance, this result may appear counterintuitive, as PC-V contains a higher proportion of calcium oxide than PC-IV (Table 1). However, a key distinction lies in the alkaline oxide content of the cements, which leads to significantly different alkaline equivalents values: 1.487% for PC-IV and 0.792% for PC-V. The alkaline equivalents in PC-IV likely intensified ASR, resulting in more severe microstructural damage and, consequently, greater strength loss. Additionally, the hydration dynamics of the two cements may have contributed to this outcome. PC-V exhibits a faster initial hydration rate, enhancing early-age strength development and potentially increasing resistance to ASR-related deterioration. In contrast, the slower early hydration of PC-IV may have made it more susceptible to internal expansion effects during the accelerated exposure period. Together, these factors help explain the higher compressive strength reduction observed in the pozzolanic cement specimens.

Previous studies have explored the impact of ASR on the compressive strength of concrete. Ahmed et al. (2003) investigated two highly reactive aggregates, including fused silica, along with a low-reactivity aggregate used as a control. Results showed a decrease in compressive strength with increasing expansion at both 20 and 38°C. Notably, at 38°C, a marked reduction in compressive strength was observed, especially in specimens containing the highly reactive fused silica aggregate.

Sanchez et al.’s (2017) results regarding compressive strengths were quite similar to those obtained in this work. Cope and Slade (1992) reported an increase in compressive strength in a concrete mix containing reactive fused silica. It also concluded that curing concrete with a slowly reactive aggregate at elevated temperatures does not significantly impact compressive strength, either at an early age or after the full curing period. Jones and Clark (1998) similarly concluded that ASR expansion leads to a reduction in compressive strength. However, Saint-Pierre et al. (2007) have indicated that ASR does not significantly affect compressive strength. These conflicting findings may be due to variations in specimen type, testing conditions and material properties (Mohammadi et al., 2020).

The reduction in strength observed in the concrete samples tested aligns closely with the findings of Sanchez et al. (2018) and Diab et al. (2021). These results can be compared with expansion levels ranging from 0.05% to 0.2%, where Sanchez et al. (2017) reported a 5%–25% reduction in strength and Alejos et al. (2014) indicated an 18% reduction at 0.2% expansion or after 6 months. This consistency confirms that the method used to induce ASR in the specimens led to concrete expansion classified as low to medium growth. Furthermore, Islam and Ghafoori (2018) demonstrated that compressive strength is generally unaffected by ASR at early ages, consistent with the minimal expansion seen in our samples. The reductions in strength were statistically significant for both cements, although to different degrees. PC-IV specimens exhibited a highly significant loss (t = 5.45, p < 0.0001; F = 29.66, p < 0.00001), while PC-V specimens also showed a significant, but smaller, reduction (t = 2.48, p = 0.021; F = 6.38, p = 0.017). The two-way ANOVA highlighted strong main effects of cement type (p < 0.0001) and ASR condition (p < 0.0001), with a marginal interaction (p = 0.056). Although pozzolanic additives are generally expected to mitigate ASR, this result can be explained by two key factors: the alkali equivalent content and the hydration dynamics of the cements. First, PC-IV presented a considerably higher alkali equivalent (1.487%) compared to PC-V (0.792%). Since the alkali equivalent accounts for the total reactive alkalis available to trigger ASR (expressed as Na₂Oeq = Na₂O + 0.658K₂O), this difference is critical. In PC-IV, the higher availability of soluble alkalis likely intensified the reaction between alkalis and reactive silica in the aggregates, accelerating gel formation and producing greater internal swelling pressures. By contrast, the lower alkali content in PC-V limited the overall reactivity, contributing to its lower strength reduction.

Second, the hydration dynamics of the two cements also help explain the outcome. PC-V, a high-CaO clinker-rich cement, is characterized by a faster hydration rate, which favors early densification of the microstructure. This rapid formation of calcium silicate hydrate (C–S–H) reduces pore connectivity and makes it harder for ASR gel to migrate and accumulate at critical interfaces. PC-IV, on the other hand, contains significant pozzolanic additions that react more slowly, leading to delayed strength development and higher early porosity. Under the accelerated exposure conditions of the present study, this slower hydration may have left PC-IV more vulnerable to ASR-induced expansion, as the porous microstructure facilitated gel accumulation and cracking before the long-term benefits of pozzolanic refinement could take effect.

Together, the combination of higher alkali availability and slower hydration-driven microstructural densification explains why PC-IV, contrary to expectations, experienced a more severe reduction in compressive strength than PC-V.

The first objective of this research was to assess whether internal expansion in concrete, caused by ASR, significantly impacts the material’s properties. The results indicate that ASR does not lead to a strong and consistent reduction in compressive strength. This outcome suggests that a decrease in compressive strength is not a reliable indicator of ASR occurrence in concrete.

Figure 7 presents the tensile strength results of all the specimens tested, affected and unaffected by ASR, and Table 4 shows the dispersion measures for the samples in each group investigated, as well as the number of specimens tested in each condition, and the lower and upper limits of each sample, calculated using the IQR concept, a method for identifying outliers.

The results showed values of the coefficients of variation (CV) of the specimens tested in tension much higher (CVs ranged from 15.23% to 19.10%) than those obtained in the compression tests (CVs < 8.42%). This behaviour reflects the recognized difficulty of assessing the tensile strength of concrete, due to the intrinsically fragile nature of this material. Concrete has high compressive strength, but its tensile strength is significantly lower. In addition, the heterogeneous distribution of aggregates within the cement matrix contributes to the formation of micro-cracks, which can adversely affect tensile strength. The presence of factors such as moisture, variations in curing conditions and the quality of the materials used in the concrete mix can lead to significant variations in test results.

An analysis of the dispersion measures, particularly the average values, indicates that the tensile strength of PC-IV samples decreased by 9.95% after exposure to ASR, in contrast with PC-V samples that exhibited a more pronounced reduction, 30.49%. These results differ from those obtained by Sanchez et al. (2018), who reported a rapid and significant reduction in tensile strength during the early stages of ASR. However, the results obtained with PC-V samples are following the study of Diab et al. (2021), which recorded a 27% reduction after six months corresponding to a 0.2% expansion, as well as Sobrinho (2012), which reported reductions between 21% and 35%, and Doran (1992) who showed that, at expansions of 0.05%, 0.10%, 0.25% and 0.50%, the specimens lost 15%, 25%, 45% and 60% of their splitting tensile strength after 28 days, results consistent with those obtained in this research. Losses in tensile strength measured with gas pressure tests (Komar et al., 2013) indicated significant reductions, 12%–50%, for low levels of expansion, values similar to those found by Sanchez et al. (2018). The tensile strength results showed that specimens with PC-V cement presented a reduction in tensile strength 3.06 times greater than those with PC-IV cement. Statistical significance was assessed through both independent-samples t-tests and one-way ANOVA. For PC-IV, the comparison between ASR-affected and non-affected specimens yielded t = 1.51 (p = 0.145) and F = 2.20 (p = 0.152), indicating no statistically significant difference at the 5% significance level. Conversely, for PC-V, the difference was highly significant, with t = 4.58 (p = 0.00031) and F = 19.89 (p = 0.00027), confirming that ASR-induced tensile strength reduction in this cement type is statistically robust. The two-way ANOVA further supported these findings: while cement type alone was not significant (F = 0.79, p = 0.380), ASR condition had a strong main effect (F = 18.40, p < 0.001), and the interaction between cement type and ASR condition was also significant (F = 6.29, p = 0.016). These results corroborate the descriptive statistics, showing that while PC-IV exhibited a modest mean reduction (≈10%) within the variability expected for tensile tests, PC-V experienced a substantially greater reduction (≈30%), consistent with its higher susceptibility to ASR and with the mitigating role of pozzolanic additions in PC-IV.

Figure 8 and Table 5 summarize the modulus of elasticity test results of all the specimens tested, affected and unaffected by ASR. This table also includes the dispersion measures for the samples in each group investigated, that is the lower and upper limits of each sample, calculated using the IQR concept.

The coefficients of variation in all the sample groups were very similar, ranging from 9.90% to 12.32%, indicating low variability in the sample data. This consistency highlights the stability of the static modulus of elasticity of concrete. As observed in previous studies, the static modulus is a reliable property with significant applicability for predicting the degree of damage and the rate of deterioration in concrete elements affected by ASR (Reinhardt and Mielich, 2011). Table 4 shows that, for both types of cement used, the reduction observed in the static modulus of elasticity of the tested specimens was very close (−35.52% for PC-IV and −28.69% for PC-V cement).

The reduction in the modulus of elasticity is strongly associated with the mechanism of ASR. The expansive alkali-silica gel formed by the reaction between alkalis in the cement paste and reactive silica in aggregates absorbs moisture, leading to swelling and generation of internal tensile stresses. These stresses result in the formation of microcracks throughout the cement matrix. Unlike compressive and tensile strengths, which may be marginally affected in early stages, the modulus of elasticity is directly influenced by these microcracks, as they compromise the material’s stiffness. This makes the modulus of elasticity a particularly sensitive indicator for detecting early ASR damage, even before significant strength loss occurs.

The results obtained with the static modulus of elasticity are consistent with other results presented in the literature. Diab et al. (2021) reported reductions in the modulus of elasticity between 36 and 39% for similar expansion levels and up to 41.54% over a 12-month testing period, while Sanchez et al. (2018) documented a similar behaviour, consistent with the patterns observed in this experimental campaign, showing a comparable rate of reduction across different types of cement. This supports the use of the modulus of elasticity as an effective tool for detecting ASR, as the reaction presents similar impacts on various cement types.

In summary, the modulus of elasticity reduction due to ASR is particularly valuable for identifying the reaction’s presence because it directly reflects the internal damage and stiffness loss within the concrete. This stiffness loss tends to increase consistently as ASR progresses, providing a quantifiable metric that can be tracked over time. Additionally, the sensitivity of the modulus of elasticity to ASR damage is observed across various types of cement, making it a reliable and versatile diagnostic tool. An explanation for the higher sensitivity of the static elastic modulus in concrete samples affected by ASR-induced internal expansion is based on the fact that it is a mechanical property very sensitive to microcracks in the concrete matrix, that is the first signs of damage caused by ASR. Even small cracks, with little impact on compressive and tensile strengths, reduce the material’s stiffness and, therefore, its modulus of elasticity. The statistical significance analyses performed indicated that, for PC-IV, the comparison between ASR-affected and non-affected specimens yielded t = 11.58 (p < 0.0001) and F = 126.24 (p < 0.0001), indicating a very strong and statistically robust reduction in modulus of elasticity. For PC-V, the reduction was also highly significant, with t = 14.55 (p < 0.0001) and F = 211.84 (p < 0.0001). The two-way ANOVA confirmed these findings: both cement type (F = 70.03, p < 0.0001) and ASR condition (F = 338.05, p < 0.0001) were highly significant main effects, and their interaction was also significant (F = 7.88, p = 0.006), indicating that the extent of ASR-induced stiffness loss differed by cement type. Using average values, the reductions attributed to ASR were −35.52% for PC-IV and −38.69% for PC-V, showing that both cements experienced substantial degradation in elastic modulus. These results emphasize that the modulus of elasticity is more sensitive to ASR-induced microstructural damage than tensile or compressive strength, and they highlight the importance of evaluating stiffness-related properties when assessing the structural implications of ASR.

This study analyses the impact of ASR-induced expansions on the mechanical properties of concrete, a key factor in concrete aging and degradation, through an experimental campaign with 240 concrete specimens subjected to accelerated ASR conditions, during 4 months. Two cement types were tested: PC-V (additive-free) and PC-IV (with pozzolan), and the reactive component was coarse aggregate containing microcrystalline quartz.

The experimental results showed that internal expansion due to ASR significantly affects various mechanical properties of concrete, especially the static modulus of elasticity. The results showed, consistently, a marked reduction in the modulus of elasticity in different types of cement, regardless of pozzolanic additions. This reduction highlights the modulus’ sensitivity to ASR-induced micro-cracks, making it a reliable early indicator of ASR damage.

The compressive strength results, on the other hand, did not exhibit a strong and consistent reduction as a result of ASR, suggesting that compressive strength alone is not a definitive indicator of ASR in concrete. Nevertheless, the study identified a significant compressive strength reduction in specimens using PC-IV compared to PC-V, likely due to the different alkali content and hydration characteristics of these types of cement.

The tensile strength tests further emphasized the detrimental impact of ASR on concrete’s mechanical integrity, with PC-IV specimens showing greater reductions than PC-V. This behaviour highlights the complex interaction between cement composition and ASR, pointing to the importance of material selection in environments susceptible to ASR. The high variability observed in tensile strength measurements also reinforces the inherent challenges in assessing concrete’s tensile properties, particularly under ASR conditions.

This research presents valuable information on the effects of ASR on the mechanical properties of concrete. The results support the use of modulus of elasticity as a sensitive and consistent indicator of ASR-induced damage while suggesting that changes in compressive and tensile strength, while relevant, can vary widely based on specific conditions. Monitoring these properties together can provide a more comprehensive understanding of the impact of ASR, aiding in the early diagnosis, maintenance and management of ASR-affected structures.

The findings of this study have direct implications for the design, maintenance and diagnosis of concrete structures exposed to internal swelling pathologies such as ASR. By confirming the sensitivity of the modulus of elasticity to ASR-induced deterioration, this research offers a practical tool for early-stage damage detection, allowing engineers to take timely preventive or corrective actions before structural safety is compromised. The results are particularly relevant for infrastructure located in regions with known reactive aggregates and adverse environmental conditions. Additionally, this study advances knowledge in concrete durability and may inform future technical guidelines and policies on sustainable material use. Its experimental approach supports engineering education in durability and structural pathology. Improving concrete resilience enhances public safety and infrastructure longevity, with economic and environmental benefits.

This work was financially supported by Base Funding – UIDB/04708/2020, with DOI: 10.54499/UIDB/04708/2020; Programmatic Funding – UIDP/04708/2020, with DOI: 10.54499/UIDP/04708/2020 of the CONSTRUCT funded by national funds through the FCT/MCTES (PIDDAC); and FCT through the individual Scientific Employment Stimulus 2020.00828.CEECIND/CP1590/CT0004, DOI: 10.54499/2020.00828.CEECIND/CP1590/CT0004.