Microstructural evolution and elastic response of graphene oxide–modified concrete: SEM and finite element modelling using COMSOL multiphysics

This study investigates the microstructural attributes and elastic response of graphene oxide (GO)-modified concrete, focusing on relating morphological changes to its macroscale mechanical performance. To meet this aim, three mixes with varying GO dosages (0.00%, 0.04%, and 0.08% by weight of cement) were selected, and FESEM was used to assess concrete morphology, compactness, and microstructural heterogeneity. Finite element analysis (FEA) was then performed in COMSOL Multiphysics to explore the behaviour of elastic displacement, von Mises stress distribution, and reaction-force response of the control and GO-modified concrete under uniaxial compression. The findings from the SEM micrographs comprising 0.04% GO revealed a refined microstructure, denser CSH gel formation, and improved fibrillar crystalline growth, which supported a 30.97% increase in 28-day compressive strength compared to the control concrete. This behaviour is mainly associated with GO nucleation. Conversely, the 0.08% GO mix exhibited localized aggregation, resulting in inconsistent hydration behaviour. On the other hand, under the application of the same displacement inputs in COMSOL Multiphysics, the 0.04% GO mix showed a 14.44% increase in reaction force response compared to the control mix. This leads to improved stiffness, greater resistance to elastic deformation and improved load-bearing capacity. The numerical assessment is thus interpreted as a macroscale analysis of stiffness-dependent elastic behaviour, and not as an independent prediction of compressive strength. Collectively, the combined experimental and finite element observations indicate that 0.04% GO provided a better response among the dosages investigated under the adopted material and testing conditions.

Concrete is one of the most commonly employed construction materials owing to its versatility, accessibility and ability to sustain high compressive loads. However, concrete is inherently brittle in nature and encompasses low tensile resistance, which makes it prone to the initiation and propagation of microcracks under mechanical loading. These cracks lead to microstructural discontinuities that predominantly occur in the cementitious matrix, especially in porous regions and at interfaces between hydration products and unhydrated cement particles. To overcome this, the researchers focused on incorporating nanomaterials into concrete to examine their behaviour within the concrete matrix to meet the current scenario. Furthermore, contemporary developments in nanotechnology have provided specialised techniques to alter the cement hydration at the nanoscale. All these constitute a refinement of the microstructure of concrete (Bhatrola et al., 2023; Anish and Logeshwari, 2024; Mohtasham Moein et al., 2024; Sanchez and Sobolev, 2010). Among several nanomaterials, graphene oxide (GO) has received considerable attention owing to its large specific surface area, oxygen functional groups and strong interaction potential with cement hydration products. Previous studies have noticed that the addition of a small amount of GO, i.e., 0.01%–0.05% of the weight of cement, improves the mechanical properties, including compressive strength, tensile strength and flexural strength, by up to 46%, 67% and 59% (Lin et al., 2022), respectively. Interestingly, a recent study on a UHPC mix employed with GO and various coarse aggregate ratios concluded that the variations in the coarse aggregate shifted the outperformed mix from 0.04% GO (without aggregate) to 0.01% GO (in presence of aggregate) (Mao et al., 2024).

Similarly, studies have reported improvements in the mechanical performance and durability of concrete in the presence of GO (Shang et al., 2015), primarily by accelerated hydration kinetics and densification of the calcium silicate hydrate (CSH) gel. Such improvements are facilitated by the functional groups that enhance interactions between GO sheets and cement hydration products. In addition, a study has reported that GO incorporation can improve compressive strength, reduce pore connectivity, and refine the interfacial transition zone. This improvement is generally attributed to the heterogeneous nucleation effect of GO sheets that accelerates the formation of calcium silicate hydrate (CSH) gel, the principal hydration product responsible for strength development in cement-based systems. Moreover, the GO sheets provide secondary nucleation sites for precipitation of hydration products, which leads to greater concrete matrix compactness with fewer microstructural defects and enhanced continuity of hydration phases. Thus, the nucleation effect of GO sheets contributes to early-stage hydration, leading to a dense, finer microstructure. However, this beneficial response is strongly dosage-dependent (Gladwin Alex et al., 2022; Suo et al., 2022). At relatively low concentrations, the high surface area and oxygen-containing functional groups of GO can promote heterogeneous nucleation and matrix densification, whereas excessive GO content may favour sheet interaction, localized clustering, and non-uniform hydration. Consequently, the dosage associated with maximum mechanical or microstructural improvement is not consistent across different cementitious systems. Although these have been proposed in many studies, detailed interpretation of microstructural changes, particularly the specific morphological evolution of CSH in the presence of GO, especially the changes from outer to inner CSH, the formation of fracture initiation zones, and agglomeration effects associated with GO dosage, remains comparatively limited (Murali et al., 2022). Therefore, a comprehensive study of the mechanism of a heterogeneous material is necessary to be widely used in the structure-related industry.

Scanning electron microscopy, as an effective characterization resource, has been extensively used to assess hydration morphology and to determine the evolution of cement hydration products at the microscale. Particularly, it enables to visualize hydration features such as outer CSH, inner CSH, voids, fractures and crystalline growth patterns, thereby offering valuable insight into the densification behaviour of cementitious materials. Therefore, through SEM observations, it is possible to identify the qualitative evidence of alterations in morphological structure, pore characteristics and other hydration-driven products that are visible in the micrographs. Although, as discussed earlier, several studies have reported improved microstructural compactness in GO-incorporated systems, detailed interpretations correlating hydration morphology with crack-initiation regions and pore distribution remain relatively limited. Addressing this limitation is essential to clearly understand the microstructure and mechanical performance of GO-modified concrete (at various dosages) and to support its shift toward widespread applications. In this context, recent improvements in research and development by graphene and GO manufacturers have enabled scale-up production and reduced costs, making their use in concrete more practical and scalable (Li and Zheng, 2022; Liu et al., 2017).

Numerical simulation, together with experimental performance, has emerged as an effective framework for assessing the mechanical response and stress distribution behaviour of cementitious materials (Dong et al., 2026). Finite element analysis offers a visualization of stress concentration zones, distribution of displacement and characteristics of deformation under loading conditions and thus provides complementary information on the macroscale structural response under adopted numerical conditions. Among the available computational platforms, COMSOL Multiphysics provides a versatile environment for modelling deformation behaviour and stress response in concrete systems subjected to compressive loading. Existing numerical investigations of GO-doped concrete have ranged from molecular models to mesoscale models and permeability and corrosion behaviour of concrete. Few studies have combined SEM-based morphological observations with a macroscale finite element assessment of the elastic response of GO-reinforced concrete within the same experimental framework. Based on the assumptions and attributes adopted in this study, the microstructural features observed by SEM are not explicitly represented; instead, the numerical response reflects the equivalent macroscopic properties assigned to the concrete.

In this context, a gap remains in the combined interpretation of GO dosage, experimentally observed microstructural morphology, and the corresponding macroscale elastic response of concrete, as presented in Table 1. This work addresses this gap through complementary experimental, microstructural, and finite element assessments of concrete containing 0%, 0.04%, and 0.08% GO. SEM is employed to examine variations in matrix morphology, compactness, pore characteristics, localized defects, and crystalline features associated with different GO dosages, while COMSOL Multiphysics is independently used to evaluate the total displacement, von Mises stress distribution, localized stress concentrations, and stiffness-dependent reaction-force response under compressive loading. The novelty of the study lies in interpreting these microstructural and numerical observations within the same investigation while maintaining their distinct roles; the SEM observations provide qualitative microstructural insight, whereas the finite element model provides a macroscale assessment of the elastic response without explicitly representing the observed microstructural features. Thus, this work aims to evaluate the effect of GO dosage on microstructural parameters and macroscale elastic behaviour under the adopted experimental and numerical conditions.

Materials

Ordinary Portland Cement (OPC), conforming to IS 12269 (BIS, 2013), fine aggregate, and coarse aggregate of IS sieve sizes 20 mm and 12.5 mm were used in accordance with IS 383. Portable water was used for mixing and curing. Graphene oxide (GO), provided by the supplier, was used as the nanomodifier in the concrete. The graphene oxide powder was examined using scanning electron microscopy (SEM), wherein GO exhibited folds and wrinkled surfaces, and is comparable to the literature (Reddy and Prasad, 2023; Reddy and Prasad, 2022). These folds and wrinkles of GO sheets could be noticed in Figure 1. Figure 2 shows the elemental composition of GO. The GO sheets were approximately 500 nm in lateral dimension, about 5 nm thick, and had an oxygen content of about 21%.

Dispersion of graphene oxide

The procured GO powder was weighed using a balance accurate to 0.001 g, as required, and slowly added to the measured potable water. The powder content and water were measured according to the corresponding mix proportions. GO powder was slowly poured into the water and stirred with a glass rod to prevent clumping. The mix was then sonicated using a bath sonicator (manufacturer: Citizen Scale (India) Pvt. Ltd., now Aczet Pvt. Ltd., India) for 20 min in an ice bath. This bath sonicator provides built-in time and temperature monitoring accurate to 1 s and 1°, respectively. The ice bath prevented overheating during sonication, keeping the maximum temperature at 30 °C. During sonication, the beaker was covered with aluminium foil to prevent evaporation and spillage. 10 mL of the sonicated solution was assessed for Zeta potential measurements using a Brookhaven instrument. The zeta potential of −39.16 mV indicated that the sonicated material was highly stable and had excellent dispersibility.

Mixing and sample preparation

Concrete mixes were designed for M25 grade as per IS 10262:2019 (BIS, 2019). The resulting mix proportions were 1:1.60:2.74 (C: FA: CA), which were taken from our previous work (Vishnu vardhan and Mohan ganesh, 2025). The cement content obtained was 425 kg/m 3, with a water-to-cement ratio of 0.46. The dry ingredients (cement, fine aggregate and coarse aggregate) were weighed according to the mix. Initially, the coarse and fine aggregates were poured into the pan, followed by cement and water, and mixed properly to ensure uniformity. This dry mix was then mixed with the gradual addition of water dispersed with GO to obtain a homogeneous concrete mixture. The mixture was then poured into the specimens and cured until the required testing age. The cubes (150 mm × 150 mm × 150 mm) were prepared for compression testing and tested using CTM at 7, 14, and 28 days (3 specimens per day) to assess compressive strength. Three mixtures were prepared with GO dosages of 0%, 0.04% and 0.08% by weight of cement. For microstructural analysis, fractured fragments, approximately 10 mm in size, were collected from the core portion of the tested specimens.

Microstructural characterisation

Microstructural analysis was performed using Scanning Electron Microscopy (SEM) to observe morphological changes, primarily the CSH gel structure, gel pores, crack initiation zones, and the distribution of hydration products. Examination was carried out at multiple magnifications to obtain the morphology of the outer and inner CSH. These images were captured at an accelerating voltage of approximately 20 kV. This facilitated the clear observations in the morphological structure of the concrete specimens. However, to overcome the non-conductive nature of concrete specimens, it has been coated with gold, eliminating bright charging spots during examination.

COMSOL multiphysics

To twin the experimental specimen, a finite element model (3D) of a 150 mm × 150 mm x150 mm cube was modelled, using COMSOL Multiphysics, to examine the elastic behaviour of concrete under uniaxial compressive loads, as shown in Figure 3a. The solid mechanical interface was used to analyse the created model under the assumptions that the material is homogeneous and isotropic. This assumption was made to represent the complete response of concrete within the elastic range. In the developed model, concrete behaviour was assumed to be linearly elastic to assess its macroscale elastic response. Hence, the finite element analysis was employed as a complementary macroscale assessment to evaluate the displacement and equivalent stress, localized stress concentrations, and the comparative reaction-force response of the investigated mixtures under identical geometry, loading, and boundary conditions. Despite this, the present model does not explicitly represent pores and the interfacial transition zone as individual phases; instead, it reflects their collective contribution through equivalent macroscopic material properties assigned to each mixture. Furthermore, explicitly simulating the heterogeneous material and failure mechanisms would require a mesoscale model incorporating an appropriate nonlinear damage or fracture formulation. Thus, the parameters required for the material to be assigned are as follows: Young’s modulus, Poisson’s ratio, and density. The Elastic modulus for the control mix was calculated from the experimental compressive strength using standard empirical relations given in IS: 456–2000 (B. of Indian Standards, 2000). Similarly, for the GO-modified concrete, the elastic modulus was assigned from its corresponding compressive strength values. The Poisson’s ratio and the density of concrete were assumed to be 0.2 and 2,400 kg/m 3, respectively. These values are consistent with typical ranges for normal-weight concrete.

The analysis was then followed by constraining the developed model against translation and rotation to simulate the experimental conditions for compression testing. Therefore, the bottom surface of the modelled cube was constrained, and the top surface was loaded with prescribed displacement in the vertically downward direction. This configuration represents a simplified experimental compression setup and ensures identical loading conditions for comparison among the investigated mixtures. However, the model does not explicitly incorporate platen–specimen contact or friction; therefore, the localized stress concentrations observed near the bottom edges and corners are associated with the adopted fully constrained boundary condition. This magnitude was attained from the strain level based on the fundamental definition of engineering strain, i.e., ε = ΔL/L. This ensures that the analysis was conducted in the elastic region of the concrete. The completely developed model underwent the discretisation with a tetrahedral mesh (finer) to ensure numerical convergence while maintaining computational efficiency and to acquire the stress and displacement, as shown in Figure 3b. Additionally, a stationary solver was used to obtain the response of the concrete model under compression. After performing the analysis, the outcomes were recorded and represented as total displacement (solid.disp), von Mises stress (solid.mises), and reaction force at the constrained boundary. The reaction force was calculated by surface integration at the fixed boundary. Furthermore, contour plots and slicing were developed to better visualise the deformed patterns and stress distribution within the specimen.

Mesh sensitivity analysis

A mesh-sensitivity analysis was done on the 0.04% GO model using three progressively refined free-tetrahedral meshes: Normal, Fine, and Finer, as shown in Table 2. This analysis was particularly adopted for the 0.04% GO model because it was the better-performing mix. The outcomes demonstrated that the reaction force varied negligibly with successive mesh refinement. The corresponding differences were less than 0.01%, indicating convergence of the global reaction-force response. On the other hand, the maximum von Mises stress varied for various mesh types and mesh elements, demonstrating comparatively greater sensitivity of the localized peak stress to mesh refinement. Based on this assessment, the Finer mesh was adopted for the subsequent numerical analyses.

Compressive strength

The results of the GO-reinforced concrete and control mix, i.e., without GO, are shown in Figure 4. It has been observed that the strength of concrete increases with the dosage of the graphene oxide. The maximum strength was 40.6 MPa at 28 days of testing for 0.04% GO mix. From Figure 4, concrete containing 0.04% GO outperformed the other mixes of all concrete ages being tested. The rate of strength gain was about 44% and 31% at 7 days and 28 days, respectively, for 0.04% GO-reinforced concrete compared to the control mix. The results demonstrated that the rate of strength gain in the initial days is faster than in the latter stages. This is due to the nucleation effect of the GO sheet in the matrix (Vishnu vardhan and Mohan ganesh, 2025). Although the strength decreased for concrete specimens containing 0.08% GO compared with those at 0.04% GO, it remained above that of the control mix. This reduction in the strength of concrete reinforced with 0.08% GO is attributed to agglomeration, which is further discussed in section 3.2.

Microstructure of concrete

Figures 5a–f show the 7-day concrete morphology with and without GO, revealing clear distinctions between them. In the control mix (Figures 5a,b), the microstructure is characterised by loosely packed outer CSH, visible gel pores, and partially hydrated cement particles. The outer CSH resembles a cloud-like mass and the inner CSH around the cement grains appears denser. Microvoids and cracks are visible within the matrix and the fracture regions are primarily found in the porous zones. Unhydrated cement grains surrounded by CSH indicates that the hydration is in progress at this stage. Fractures in the internal CSH gel have been observed and could be a result of the application of stress to the specimens. These are usually observed in the porous zone which indicates microstructural discontinuities due to incomplete hydration at early ages. However, 0.04% GO doped concrete (Figures 5c,d) exhibits a much denser CSH gel. The fibrillar crystalline network appears to be more interconnected indicating a more rapid hydration. It is observed that the separation of CSH on cement particles (indicated by an oval) appears to be less, indicating that graphene oxide sheets augment the nucleation activity. Also the crystalline mass indicates to bridge surrounding hydration regions and thus to reduce the visible connectivity of pores. This could be due to the presence of GO, which improves the growth rate of the CSH gel at an early age and enhances the concrete’s load-bearing capacity. Furthermore, notably, crack formation within the CSH gel is comparatively limited, indicating improved early-stage matrix integrity. Altogether, the improvement in compressive strength of the GO-doped concrete was greater than that of the control mix. For the 0.08% GO mixture (Figures 5e,f), needle-like ettringite and monosulfate hydrates around 1 µm are clearly visible. However, localized agglomeration regions are observed, characterized by clustered hydration products. These agglomerates are likely due to heterogeneous regions from the hydration reactions resulting from a higher GO dosage. Such clustering may induce local micro-stress concentrations, potentially explaining the reduction in structural uniformity at elevated GO content.

The SEM images of concrete at 14 days (Figure 6) depict hydration and formation of CSH gel. In the control specimen (Figure 6a), the hydration products are noticeable within the matrix. The concrete incorporating 0.04% GO (Figures 6b,c) demonstrates a reticular inner CSH network interwoven with crystalline fibres. These thread-like crystalline patterns serve as growth patterns for a further CSH gel, as seen in Figure 6c. This observation demonstrates that the presence of GO promotes the evolution of loosely packed outer CSH into denser inner CSH (Qureshi and Panesar, 2019). Additionally, the effect of GO facilitates the reaction between water and anhydrous C 3 S or C 2 S. The intergrowth of crystalline fibres and CSH gel reduces pore spacing and improves the continuity of the cement matrix. In the specimen containing 0.08% GO (Figure 6d), regions of agglomeration are observed. The clustered hydration products create morphological irregularities in the matrix, indicating that excessive GO could hinder uniform hydration. However, the specimens tested after 14 days showed a lower level of agglomeration than the 7-day specimens (as shown in Figure 6d).

At 28 days, the control concrete (Figure 7a) shows an outer CSH layer surrounding the cement particles, arranged in a layered pattern. This is because the outer layer of cement particles reacts first, followed by subsequent layers that occupy the space between unhydrated cement particles, gel pores, and voids. Thereby, the hydration product could be compacted into a solid heterogeneous material. The concrete incorporating 0.04% GO (Figure 7b) exhibits more pronounced CSH expansion and the formation of crystalline fibrous structures. Further, Figure 7c shows CSH gel growth and distribution from the tip of the crystalline fibrous mass to its top. The distribution and growth of the crystalline structure are random; hence, the CSH distribution follows the same pattern. The hydration products appear more compact and uniformly distributed, indicating improved microstructural densification. The crystalline fibres create a network that promotes CSH gel formation throughout the matrix, thereby improving structural integrity. Conversely, irregular CSH deposition and non-uniform growth behaviour are observed in the specimen with 0.08% GO (Figure 7c). These features are associated with the agglomeration of graphene oxide at higher dosages, which disrupts the homogeneous distribution of hydration products. Taking the complete analysis, GO serves as a nucleation site for the hydration products. Therefore, the formation of calcium silicate hydrated gel grew denser than graphene oxide-free concrete. The sketched irregular shape (white dotted portion) in Figure 7c illustrates the de-bonded CSH gel under loading. This occurred adjacent to the space within the matrix and the region where the outer CSH is present. From this, we can conclude that the crack initiation was in the areas of pores, voids, and the location of the outer CSH gel.

Finite element analysis - COMSOL multiphysics

The numerical results evaluate the mechanical response of 150 mm concrete cubes under displacement-controlled loading of 0.165 mm, as explained by comparing GO-doped and GO-free concrete specimens. This explanation includes total vertical displacement along with reaction forces, von mises stress with its contour plots.

As illustrated in Figures 8a,b, both the control (0% GO) and the reinforced (0.04% GO) specimens exhibit a linear vertical displacement gradient. The maximum displacement is localized at the top loading surface, transitioning uniformly to zero at the fixed base, which is in line with the prescribed boundary conditions and exhibits the absence of rigid body motion. The smooth, continuous variation in displacement along the cube’s height reflects the deformation pattern as anticipated from the adopted linear-elastic material model for both concrete mix models in COMSOL. Although the displacement field remains unchanged under the prescribed boundary conditions, the 0.04% GO concrete model shows a higher response; however, variations in this response arise from the different elastic moduli assigned to the two concrete mix models. Further, the displacement spectra are uniform in shape, primarily resulting from the identical prescribed displacement applied to both models. In a displacement-controlled regime, the structural resistance is quantified by the reaction forces generated at the base. As shown in the simulation results, the control specimen (0% GO) produced a reaction force of 697.88 kN, whereas the 0.04% GO-reinforced specimen exhibited a significantly higher reaction force of 798.65 kN, followed by the 0.08% GO mix (770.62 kN). The reaction force of the 0.04% GO model was approximately 14.44% higher than that of the control model. However, this difference should be interpreted in light of the elastic modulus used as an input in the finite element model. Since the elastic modulus assigned to each mixture was derived from its experimentally measured compressive strength, the higher reaction force obtained for the 0.04% GO model is primarily a consequence of its higher assigned elastic stiffness under the same prescribed displacement. Therefore, the reaction-force increase should not be regarded as an independent prediction or validation of the experimentally observed strength enhancement. Therefore, the reaction-force increase could be interpreted as complementary numerical evidence of the stiffness-dependent elastic response associated with the material properties adopted from experimental analysis. Thus, the finite element analysis serves as a complementary macroscale assessment technique, wherein differences in elastic stiffness are reflected in displacement, stress distribution, and reaction-force response under identical loading and boundary conditions.

Figure 9 presents the von Mises stress nephograms, which are scalar visualizations of the equivalent internal stress state within the 150 mm cubes. Under the same geometric boundary conditions for all specimens, the peak stress of the 0.04% GO specimen (approximately 7.0 × 10 7 N/m 2) is significantly higher than that of the control (approximately 6.0 × 10 7 N/m 2) and 0.08% GO mix (approximately 6.8 × 10 7 N/m 2). The stress distribution is relatively uniform throughout the cube in all specimens (Figures 9a–c) with localized stress concentrations at the bottom boundary. These ‘hotspots’ are characteristic of the confinement effect at the fixed support. The localized stress concentrations around the contact points are associated with the adopted boundary condition and geometric discontinuities at the specimen edges. Moreover, the higher density and magnitude of these contours (7.1×10 7 N/m 2) suggest that these are in line with the elastic moduli derived from experimental results. The stress gradients of all specimens appear smooth and uniform, which indicates that the model was a numerically stable representation of the adopted elastic behaviour without numerical irregularities. Additionally, the stress colour profile of the 0.04% GO sample remains consistent at higher force magnitude. The stress contour shows a relatively higher concentration of stress near the bottom edges and corners of the specimen. This stress concentration at the edges is due to geometric discontinuities and boundary effects, while the uniform stress distribution in the core region suggests efficient load transfer and a macroscale elastic response under the adopted displacement-controlled loading condition.

The distribution of internal stress was illustrated using von Mises iso-contour line plots as shown in Figure 10, which provides a deep insight into a macroscale representation of the elastic stress response. For all three specimens, the contour lines are symmetrically distributed with respect to the vertical loading axis, ensuring numerical stability and a consistent stress-field pattern. In the control specimen (Figure 10a), the contour values range from a minimum at the upper regions to a maximum of 6.2 × 10 7 N/m 2 localized at the base. On the other hand, the 0.04% GO specimen (Figure 10b) exhibits a remarkable increase in stress with a maximum magnitude of 7.1 × 10 7 N/m 2. The corresponding stress response of the 0.08% GO specimen is also presented in Figure 10c, where in the magnitude is noted as 6.85 × 10 7 N/m 2. The sudden rise in contour density for the 0.04% GO specimen demonstrates the model’s stiffness-dependent response. This ‘stiffening effect’ is visually evident in the sharper, more compressed iso-stress lines compared with the dispersed pattern of the control specimen. Nonetheless, this could be due to the adopted elastic modulus. Also, the concentration of contours on the base boundary indicates the localized stress concentration due to the frictional constraints of the fixed support. Furthermore, the GO-reinforced specimen endures 14.5% higher internal stress density while maintaining geometric symmetry.

Based on the experimental and simulation analyses, the following conclusions were elucidated:

The incorporation of graphene oxide significantly influenced the hydration kinetics and altered the microstructure of concrete, thereby enhancing the formation of calcium silicate hydrate (CSH) gel and improving the compactness of concrete with respect to the control mix. Notably, a concentration of 0.04% GO increased the tendril-like structure of the CSH gel, with its growth and distribution aligning with the trajectory of these tendril-like crystal structures. Following this path, inner CSH gel formation was observed along with improved fibrillar crystalline interconnectivity, potentially leading to a reduction in pore continuity. Furthermore, the thread-like crystalline mass bridged the openings or voids, thereby promoting the continuous growth of the CSH gel structure. As a result of these observations, it was authenticated that graphene oxide acts as a nucleation site for the regulation of hydration products. These observed results were well corroborated by the findings of the existing work, wherein a 30.97% increase in the 28-day compressive strength of GO-doped concrete was documented over the control mix. Agglomeration-type defects were visually noticed at a GO dosage of 0.08%, which reduced the effectiveness of microstructural refinement.

Finite element analysis showed that, under the same boundary conditions and prescribed displacement, the maximum stress of the 0.04% GO model exhibited approximately 7.0 × 10 7 N/m 2, which is greater than that of the control specimen that showed approximately 6.0 × 10 7 N/m 2. This higher stress response in the 0.04% GO model reflects the higher elastic modulus assigned to this mixture and therefore represents a stiffness-dependent numerical response. Furthermore, the model was intended to capture the elastic response of concrete (linear analysis) rather than its failure or ultimate strength. Hence, the simulation provided a complementary macroscale assessment of the displacement and stress distribution associated with the experimentally derived elastic properties under identical numerical conditions.