Atomic-Level Insights into Epoxy-Modified Nanofiller Adsorption and Wetting on Calcium Silicate Hydrates: Principles for Optimizing Interfacial Properties
Cheikh Makhfouss Fame
,
Tamon Ueda
,
Marc A. Ntjam Minkeng
,
Eskinder D. Shumuye
,
Yi Wang
,
Chao Wu
Atomic-Level Insights into Epoxy-Modified Nanofiller Adsorption and Wetting on Calcium Silicate Hydrates: Principles for Optimizing Interfacial Properties
a COMAC-Beihang Commercial Aircraft Innovation Center, Hangzhou Innovation Institute of Beihang University, Hangzhou 311115, China
b Guangdong Provincial Key Laboratory for Durability of Marine Civil Engineering, College of Civil and Transportation Engineering, Shenzhen University, Shenzhen 518060, China
c School of Architecture, Building and Civil Engineering, Loughborough University, Loughborough LE11 3TU, UK
d Institute for Advanced Study, Shenzhen University, Shenzhen 518060, China
e School of Civil Engineering, Central South University, Changsha 410075, China
f Department of Civil and Environmental Engineering, Imperial College London, South Kensington Campus, London SW7 2AZ, UK
Cementitious composites are ubiquitous in infrastructure, but their durability is often limited by interfacial weaknesses, particularly when bonded to epoxy polymers. These organic-inorganic interfaces are critical in applications such as fiber-reinforced polymer repair, strengthening, and additive concrete manufacturing. However, debonding and cracking at the epoxy-calcium silicate hydrate (CSH) interface under conditions such as elevated temperature, moisture, or fatigue-loading remain long-standing challenges. A fundamental understanding of the atomic-level interactions governing macroscopic interfacial properties is essential to optimize the performance of these composites; however, these remain largely unexplored by conventional analytical techniques owing to the buried nature of the nanoscale epoxy-CSH interface. In this study, we employed classical molecular dynamics (MD) simulations to elucidate the molecular mechanisms associated with epoxy-modifier-incorporated (nanofillers-incorporated) epoxy prepolymer (i.e., liquid epoxy) adsorption, wetting, and adhesion to CSH surfaces. We investigated three epoxy systems: epoxy prepolymer, acrylamide-modified (AM-modified) epoxy prepolymer, and nanosilica-reinforced (nano-SiO2-reinforced) epoxy prepolymer, with their pre-curing interactions with CSH simulated for the first time. Key interfacial parameters, including interphase formation, wetting ratio, atomic-density profile, monomer-concentration profile, interfacial ionic and hydrogen bonding, interaction energy, and work of adhesion, were rigorously analyzed. Our MD simulations revealed that the diethyltoluenediamine (DETDA) epoxy hardener preferentially adsorbs over bisphenol F diglycidyl ether (DGEBF) resin on the CSH surface, leading to a hardener-rich interphase. Both resin and hardener components were found to concentrate within approximately 5 Å of the CSH surface and exhibit molar ratios significantly higher than that of the bulk, suggestive of a potentially stiffer interphase. The incorporation of AM was found to significantly enhance interfacial adhesion and improve adsorption and wetting through strong electrostatic and hydrogen bonding, resulting in a 16.43% increase in the work of adhesion compared to neat epoxy. The nanosilica filler, while improving interfacial interactions, aggregated and yielded a modest enhancement of 4.45% in the work of adhesion. These findings provide fundamental and novel molecular-level insights for the rational design of epoxy-modified cementitious composites with enhanced interfacial-bonding, durability, and mechanical properties. This study underscores the critical role played by nanofiller selection in tailoring interfacial properties and offers valuable guidelines for material formulations that optimize the interphase region in high-performance polymer-cement hybrid materials.
Cheikh Makhfouss Fame, Tamon Ueda, Marc A. Ntjam Minkeng, Eskinder D. Shumuye, Yi Wang, Chao Wu.
Atomic-Level Insights into Epoxy-Modified Nanofiller Adsorption and Wetting on Calcium Silicate Hydrates: Principles for Optimizing Interfacial Properties.
Engineering, 2026, 61 (6) : 73-89 DOI:10.1016/j.eng.2025.12.032
Cement-based materials are cornerstones of modern infrastructure and provide essential structural frameworks for buildings, bridges, and transportation networks [1,2]. While inherently exhibiting a low embodied carbon footprint per unit mass, the sheer scale of global cement production contributes significantly to anthropogenic CO2 emissions, accounting for approximately 5%-8% of total emissions [3]. Enhancing the mechanical performance and durability of a cementitious structure is therefore crucial for mitigating its environmental impact and fostering sustainable construction practices [4,5]. Furthermore, given the substantial economic and environmental costs associated with structural demolition and reconstruction, extending the service life of existing infrastructure through repair, retrofitting, and strengthening (often termed structural intervention) has become a crucial strategy for sustainable infrastructure management [1,6].
Epoxy-based adhesives are increasingly vital in structural interventions, including concrete repair and strengthening [6], fiber-reinforced polymer (FRP) bonding [7], and in advanced applications, such as self-healing concrete [8,9], multilayered concrete composites [10,11], protective coatings [12], and three-dimensional (3D)-printed concrete [13]. The performance of these applications depends critically on interactions at the epoxy-inorganic interface [14], [15], [16] that occur at various stages before, during, and after the epoxy curing process [14], resulting in a structurally unique interphase region [17] widely considered to be a critical factor in establishing polymer/solid-surface adhesion [18] and interfacial fracture toughness [19]. However, inadequate bonding at epoxy/concrete interfaces remains a significant challenge, leading to debonding, stress concentration, and premature material failure in service environments. Therefore, achieving strong interfacial adhesion between the epoxy and calcium silicate hydrate (CSH), which is the primary binding phase in cement, is essential for long-term performance. The challenge here lies in understanding and controlling the complex molecular interactions that govern interfacial behavior.
Characterizing the epoxy/CSH interface presents substantial experimental challenges. The inherent inaccessibility of this buried interface, coupled with the nanoscale dimensions of the interphase region [14,15] raises fundamental questions regarding its structure and properties. Indeed, available experimental analysis techniques often lack the required sensitivity and spatial resolution at such small scales. For instance, X-ray photoelectron spectroscopy (XPS), which is useful for studying the binding modes of individual epoxy and amine molecules on metal-oxide surfaces [20,21], cannot fully capture the complex interplay between molecular adsorption, wetting, and network formation that define a realistic interphase [14,19].
Molecular dynamics (MD) simulations offer a powerful complementary approach for investigating buried interfaces at the atomic level by providing insights that are difficult to obtain experimentally [22], [23], [24]. Despite not capturing reactive bond formation, classical MD simulations are able to elucidate adsorption mechanisms, dynamic behavior, interphase formation, and interfacial energetics [24], [25], [26], [27]. Indeed, combining MD simulations with experimental characterization techniques offers a holistic and powerful approach for elucidating the complex mechanism associated with polymeric interphase formation on a solid surface [28,29].
While recent MD simulations have focused on the cured epoxy/CSH interface, examining its mechanical properties and debonding mechanisms [30], [31], [32], the critical pre-curing stages that involve liquid epoxy (prepolymer) adsorption, wetting, and adhesion to the CSH surface, have received surprisingly little attention. This oversight is significant because these initial interactions fundamentally dictate the subsequent structure of the interface, interphase development, and composite performance. Addressing this knowledge gap is essential for rationally designing improved epoxy formulations and adhesion strategies. Indeed, understanding the pre-curing behavior of liquid epoxy components on CSH is crucial for controlling the early stage of interface formation, which influences polymer network homogeneity and the durability of systems such as FRP-concrete repairs and additively manufactured cementitious structures. Furthermore, strategically incorporating a filler, such as nanosilica (nano-SiO2) or acrylamide (AM), as an interfacial modifier [28] is a promising strategy for tailoring adsorption, wetting, and interfacial bonding, and the development of sustainable and high-performance polymer-cement composites.
In this study we leveraged comprehensive MD simulations to investigate the molecular origins of pre-cured epoxy monomer adsorption, wetting, and adhesion on a CSH surface, with a specific focus on the influence of nanofiller modification. To this end, we constructed detailed atomistic models of CSH, an epoxy prepolymer (EP), an EP modified with AM filler (EPAM), and an EP with a nano-SiO2 filler, after which they were subjected to all-atom equilibrium classical molecular dynamics (EMD) simulations. These simulations yielded detailed descriptions of the adsorption, wetting, and adhesion mechanisms of each EP system on CSH, including its wetting ratio, local density, molar-ratio distribution of the epoxy monomer near the CSH surface, interfacial ionic and hydrogen-bond formation, interaction energies, and work of adhesion. These simulation results were used to establish design principles for enhancing the interfacial properties of polymer-cementitious composites. This study provides a fundamental understanding of the molecular mechanisms that govern epoxy-CSH interactions and provides a pathway for the development of advanced cement-based materials with superior performance characteristics, such as improved epoxy/concrete interfacial bonding properties.
2. Computational methods
In this section, we describe the computational modeling procedures necessary to accurately capture the adsorption mechanism and wetting behavior of the EP and epoxy-modified nanofillers on the CSH surface (Fig. 1). MD simulations require accurate molecular structures and appropriate force fields to reproduce atomistic interactions and interfacial characteristics. Section 2.1 presents details of how molecular models for the EP, EPAM, EP modified with nano-SiO2 (EPSiO2), CSH, and the interfaces between them (i.e., EP/CSH, EPAM/CSH, and EPSiO2/CSH) were constructed, while Section 2.2 details the force fields and simulation settings used to examine microscopic details and interaction mechanisms. Materials Studio® version 2023 (Dassault Systèmes, France) was used to construct and equilibrate epoxy and CSH models. All simulations were conducted using the open-source large-scale atomic/molecular massively parallel simulator (LAMMPS) code (Sandia National Laboratories, USA) [33], with data analyzed and visualized using Python (Python Software Foundation, USA), Excel (Microsoft Corporation, USA), visual molecular dynamics (VMD; Theoretical and Computational Biophysics Group, University of Illinois at Urbana-Champaign, USA) [34], and the open visualization tool, OVITO (OVITO GmbH, Germany) [35].
2.1. Model construction
CSH is the primary binding phase in a cement-based material and is crucial for its strength and durability. At the molecular level, CSH structures represent the core building blocks of cement mortar and concrete; however, they are complex. Creating a realistic CSH model for investigating polymer adsorption performance is essential to ensure that adsorption properties are realistically reproduced. Drawing on previous CSH molecular-modeling research [36], [37], [38], we constructed a CSH slab based on the crystalline structure of 11 Å-tobermorite (a = 11.6 Å, b = 7.39 Å, and c = 22.77 Å, where a, b, and c are the lattice parameters defining the unit cell lengths along the x-, y-, and z-axes, respectively; and α = β = γ = 90°, with α, β, and γ denoting the interaxial angle; here, 1 Å = 1 × 10−10 m) (Fig. 2(a)) [39,40]. Firstly, the ordered water molecules present in the interlayer of the 11 Å tobermorite structure were removed. Secondly, the silicate chains of tobermorite 11 Å were cleaved from its bridging sites to match the real distribution of the silicate chains in the CSH structure obtained using nuclear magnetic resonance (NMR) spectroscopy [41]. Thirdly, the CSH gel structure was constructed by adding water molecules to the dry disordered tobermorite structure via a grand canonical Monte Carlo simulation at 300 K, with the conjugate gradient algorithm then used to optimized the CSH gel structure by minimizing its energy. The CSH slab was subsequently cleaved along the [0 0 1] direction to expose the interlayer region and its surface hydroxylated [42,43]; this cleavage direction exposes the interlayer calcium ions and the terminal oxygen atoms or hydroxyl groups of the silicate chains, which are expected to interact strongly with the functional groups of the polymer and are consequently critical for understanding adsorption phenomena on the CSH surface. The resulting CSH surface was equilibrated for 1000 ps (1 ps = 1 × 10−12 s) at room temperature (300 K) and atmospheric pressure in the NPT ensemble (i.e., constant number of atoms (N), pressure (P), and temperature (T)), after which a supercell was created by expanding the amorphous CSH surface in the x- and y-directions. The supercell was then relaxed for 1000 ps under NPT conditions at room temperature to achieve a steady-state system. The final CSH model exhibited the following lattice parameters: a = 90.87 Å (x), b = 88.50 Å (y), c = 20.78 Å (z), and α = β = γ = 90°, which is endowed with sufficient space for epoxy molecules to form a film with a flat-on conformation on the CSH surface. Notably, the equilibrated CSH model has a density of 2410 kg·m−3 and a Ca/Si ratio of 1.49, which are consistent with previous experimental and MD simulation results [31].
Three epoxy systems were modeled to investigate the molecular mechanisms associated with polymer adsorption, wetting, adhesion on the CSH surface: ① EP (i.e., neat liquid epoxy), ② EP reinforced with an AM filler, and ③ EP reinforced with nano-SiO2 particles. AM and nano-SiO2 were chosen as epoxy modifiers with the aim of assessing their potential for enhancing interfacial epoxy/CSH adhesion via migration of the nanofiller toward the CSH surface. Previous studies have demonstrated that AM efficaciously improves interfacial bonding. For instance, research by Amornsakchai et al. [44] revealed that grafting AM molecules onto polyethylene (PE) fibers enhances the strength of the bond between the epoxy and PE fibers by a factor of approximately 3.5 compared to untreated fibers. Jiao et al. [28] used XPS and MD simulations to show that the AM molecules used as filler in liquid vinyl ester (VE) migrate to carbon fiber (CF) surfaces. This migration facilitates the formation of chemical bonds between the AM and both the CF surface and VE matrix, which enhances the interfacial and interlaminar shear strengths of the CF/VE composite by 126.98% and 81.86%, respectively, as a consequence, which suggests that they may also be able to enhance the epoxy/CSH interfaces in this study. On the other hand, the nano-SiO2 filler is widely recognized for its ability to enhance the mechanical properties of epoxy resins, including toughness, tensile strength, and elastic modulus [45]. In addition to improving the bulk properties of epoxy, nano-SiO2 has been shown to enhance the strength of epoxy-glass/carbon-fiber interfaces and their delamination resistances [46,47], which suggests that epoxy/CSH interfaces may experience analogous benefits.
The EP model comprised 60 bisphenol F diglycidyl ether DGEBF units and 30 diethyltoluenediamine DETDA molecules, which corresponds to a 2:1 stoichiometric resin-to-hardener functional-group ratio. DETDA, an aromatic diamine bearing two primary amino groups, can react with up to four epoxy functionalities (each amino group provides two reactive hydrogen atoms), thereby contributing four reactive sites, whereas DGEBF contributes two reactive epoxide sites. The initial structure was built with a target density of 1170 kg·m−3, consistent with experimental data for epoxy-amine prepolymer systems [48]. For the EPAM, 18 AM molecules were randomly dispersed within the bulk EP, which corresponds to a filler content of 5 wt%. This concentration was selected based on previous studies that showed that similar filler loadings led to improved interfacial properties [49,50]. Similarly, the nano-SiO2-reinforced epoxy model was prepared by incorporating five nano-SiO2 particles into the bulk EP, which also afforded a filler content of 5 wt%. The nano-SiO2 particles were derived from α-quartz and thermally quenched from 3000 K to ambient conditions to form amorphous nanoparticles approximately 6 Å in diameter. The surface was hydroxylated by protonating the surface oxygen atoms to generate hydroxyl (-OH) groups that increase nanoparticle reactivity and improve compatibility with the epoxy monomers and the CSH surface. Ten amorphous cells were constructed in each case, and their energies were minimized using the gradient-descent optimization method for 10 000 steps to prevent atomic contact. The configuration with the lowest energy was selected for subsequent MD simulations to afford relaxed and equilibrated structures. The system was initially subjected to a 100 ps NVT simulation at 300 K, followed by energy minimization. The system was then annealed by cycling the temperature between 300 and 650 K for five cycles (250 ps NVT simulations each) to overcome energy barriers and explore a broader configurational space, after which its energy was minimized under NPT conditions at 1 atm (1 atm ≈ 0.1 MPa) and 300 K for 500 ps until the cell volume had equilibrated. The system was subsequently further relaxed in the NVT ensemble for 500 ps at 300 K, and its energy further minimization to yield relaxed and stable EP, EPAM, and EPSiO2 structures for use in subsequent simulations [51]. The equilibrated EP, EPAM, and EPSiO2 amorphous cells have dimensions of 33.71 Å × 33.71 Å × 33.71 Å, 32.74 Å × 32.74 Å × 32.74 Å, and 33.95 Å × 33.95 Å × 33.95 Å, respectively. Three interfacial models, namely EP/CSH, EPAM/CSH, and EPSiO2/CSH were created after the CSH surface and various epoxy models had been built by placing the epoxy molecules at a distance of ∼5 Å above the CSH surface, after which a vacuum layer at least 50 Å thick was created above the epoxy layer (in the z direction) to prevent interactions between the epoxy molecules and the periodic image of the CSH surface. The EP/CSH, EPAM/CSH, and EPSiO2/CSH systems contained 16 762, 16 942, and 16 877 atoms, respectively. Such large models ensure that possible finite-size effects ascribable to periodic boundary conditions are avoided.
2.2. Force fields and simulation parameters
The interface force field (IFF) was used to describe bonding and non-bonding interactions among atoms within the various epoxy/CSH models with energies expressed using the polymer consistent force field (PCFF). Referred to as “IFF-PCFF,” the force field uses the 9-6 Lennard-Jones (LJ) parameters [52]. IFF-PCFF was developed using consistent descriptions for chemical bonding in silicates (such as CSH) and clay minerals, while also incorporating PCFF parameters for polymers such as epoxy [53]. This choice was motivated by the accuracy exhibited by the IFF when modeling and simulating CSH/polymer interactions, which provided superior accuracies in terms of predicting properties across a range of structural, energetic (e.g., cleavage, hydration, and organic binding), and mechanical scenarios compared with other force fields and alternative approaches (e.g., cement force field (CementFF), clay force field (CLAYFF), calcium silicate hydrate force field (CSH-FF), condensed-phase optimized molecular potentials for atomistic simulation studies (COMPASS) force field, reactive force field (ReaxFF), universal force field (UFF), and off-the-shelf density functional theory (DFT) [52], [53], [54], [55]). For instance, IFF-predicted adsorption energies of cement components are typically within 10% of the experimental values, which is approximately ten-times more accurate than the results obtained using other force fields, while eliminating uncertainties that exceed 50% when other force fields are used [54]. The IFF-PCFF potential-energy expression detailed by Heinz et al. [52] includes contributions from: ① bonded interactions (i.e., functional terms for bond stretching Ebond, angle bending Eangle, dihedral-angle torsion Etorsion, and improper torsion Eimproper), ② cross-coupling interactions (i.e., bond-bond Ebond-bond, bond-angle Ebond-angle, angle-angle Eangle-angle, bond-bond-angle Ebond-bond-angle, torsion-bond Etorsion-bond, and torsion-angle Etorsion-angle cross-terms), and ③ non-bonded interactions (i.e. van der Waals EvdW with the 9-6 LJ potential and Coulombic or electrostatic interaction ECoul terms). Detailed information on the IFF-PCFF and force field parameters are provided by Heinz et al. [52].
The adsorption and wetting mechanisms associated with the various epoxy/CSH systems were examined by subjecting epoxy/CSH interfacial models to MD simulations in the NVT ensemble for at least 2000 ps, which was sufficiently long to observe dynamic equilibrium in these models. Simulations were repeated three times under different initial conditions (different starting conformations and atomic velocities) for each system to ensure that the results are statistically significant. A spherical cutoff distance of 15 Å was imposed for the pairwise summations of LJ interactions and short-range Coulombic interactions, while the particle-particle particle-mesh solver (PPPM) was applied with a high accuracy of 10−4 kcal·mol−1 for long-range electrostatic interactions. The velocity Verlet algorithm [56] with a time step of 0.001 ps was used to integrate Newton’s equations of motion, and periodic boundary conditions were imposed in all directions. The temperature was set to 350 K and controlled using a Nosé-Hoover thermostat [57] with a damping constant of 0.1 ps [58]. This temperature represents a trade-off between a system with excessive kinetic energy, in which the epoxy molecules desorb from the surface, and one with insufficient kinetic energy for the epoxy molecules to adsorb and diffuse on the CSH surface. The pressure was maintained at 101.325 kPa (i.e., atmospheric pressure) and controlled using a Nosé-Hoover barostat [57] with a damping constant of 1 ps [58]. Four invisible LJ walls were placed at the vertical faces of the simulation box to prevent the epoxy molecules from escaping. A snapshot of the trajectory, which corresponds to 20 000 frames, was stored every 0.1 ps, and the last 5000 frames of the trajectory were considered for subsequent structural analyses.
3. Results and discussion
3.1. Adsorption and wetting mechanism for epoxy on CSH
Fig. 3 shows snapshots of the progressive formation of the EP/CSH, EPAM/CSH, and EPSiO2/CSH interfaces at various simulation times. The epoxy molecules were initially placed approximately 5 Å above the CSH surface to create a thin vacuum gap prior to the commencement of the adsorption process (Figs. 3(a), (f), and (k)). The epoxy molecules became reoriented and coalesced into circular droplets during the early stages (0-30 ps) of adsorption, which is understood as an energetically favorable arrangement that minimizes the total surface free energy [59]. Droplet formation is governed by van der Waals forces that operate between epoxy molecules and internal dipolar interactions involving the resin/hardener/filler functional groups. Attractive electrostatic and van der Waals interactions between the epoxy molecules and the polar CSH surface drove the droplets toward the surface, where the epoxy initially contacted the surface, as the simulation proceeded. The adsorbed monomers then spread laterally, which increased the wetted area over time (Figs. 3(b), (g), and (l)). This behavior is consistent with previous MD studies, which showed that polymer films reorganize themselves on polar substrates to maximize interfacial adhesion [54,60]. However, the epoxy monomers spread more slowly over the CSH surface by ∼1500 ps and the interfacial region began to stabilize, suggestive of the start of epoxy/CSH interfacial dynamic equilibration (Figs. 3(c), (d), (h), (i), (m), and (n)). Stable and equilibrated interfaces were observed at the end of the simulation process (t = 2000 ps, t is the simulation time), with the EPAM molecules observed to adopt a flatter conformation over the CSH surface than either EPSiO2 or EP (Figs. 3(e), (j), and (o)). That is, epoxy wettability on the CSH surface is enhanced by the AM and nano-SiO2 fillers. Notably, both the AM and nano-SiO2 fillers appeared to migrate toward the CSH surface to form thin interphases that bridge the epoxy molecules and the atoms at the CSH surface (Figs. 3(j) and (o)).
Fig. 4 displays the final conformations of the DGEBF resin and DETDA hardener molecules on the CSH substrate. Notably, Fig. 4(a) reveals that DGEBF molecules tend to cluster or aggregate on the CSH surface. The top-down view (Fig. 4(b)) confirms this clustering by showing a region of high DGEBF concentration orthogonal to the CSH surface. The resin molecules do not appear to have spread uniformly; rather, they appear to form a somewhat layered arrangement in which molecules stack or align in strata parallel to the CSH surface with limited direct contact. This stratification is highlighted in the inset of Fig. 4(a), which shows that a subset of DGEBF molecules form two adjacent layers (orange and tan). This behavior suggests that DGEBF molecules interact more strongly with other DGEBF molecules than with the CSH surface, which is possibly ascribable to steric hindrance stemming from the relative largeness and complex structure of DGEBF, which may hinder its ability to effectively diffuse and adsorb at binding sites on the CSH surface. In contrast, the DETDA molecules appear to be significantly more dispersed on the CSH surface. As depicted in Fig. 4(c), DETDA molecules are more likely to lie flat and be scattered on the surface, thereby maximizing their contact area with superficial CSH atoms (Fig. 4(c)). The top-down view (inset) confirms this dispersion, with DETDA exhibiting a strong affinity for CSH. This behavior suggests that DETDA molecules interact more strongly with the CSH surface than with each other, which is likely ascribable to the smaller molecular weight and simpler structure of DETDA that enables it to easily migrate toward the interface and interact with a greater number of polar sites on the CSH surface.
Fig. 5 shows the migration of the AM (cyan) and nano-SiO2 (orange) fillers toward the CSH surface, with the epoxy resin (DGEBF) and hardener (DETDA) molecules omitted for clarity. Side-view snapshots of AM migration are shown in panels (a)-(e), with the corresponding top views depicted in panels (f)-(j). Similarly, panels (k)-(o) and (p)-(t) display the side and top views of nano-SiO2 migration, respectively. Both fillers were initially uniformly dispersed within the epoxy matrix (Figs. 5(a), (f), (k), and (p)). Both fillers apparently overcome intermolecular forces binding them to the resin/hardener network as the simulation proceeds, and migrate to the CSH surface (Figs. 5(b)-(e) and (k)-(o)). The fillers exhibit distinct distribution patterns following adsorption and interface formation; the AM molecules are evenly dispersed across the CSH surface, indicative of strong affinity that is likely driven by hydrogen bonding and/or electrostatic interactions between the functional groups of AM and the polar sites of the CSH substrate. In contrast, nano-SiO2 particles tend to aggregate and form localized clusters, despite being well-dispersed within the epoxy matrix at the start of the simulation; this clustering behavior suggests that nano-SiO2 particles are more attractive to each other than to the CSH surface. The distinct migration and distribution patterns of AM and nano-SiO2 underscore the different roles that they play in modulating interfacial properties: AM promotes a uniform and homogeneous interface that exhibits enhanced adhesion, whereas nano-SiO2 aggregation creates regions of high filler concentration.
Probing the contact area between the epoxy molecules and the CSH surface is a straightforward way of monitoring the formation of the epoxy/CSH interface during the simulation. Contact area was calculated using the following equation:
Where A is the contact area, t is the simulation time, SASAEpoxy(t) and SASACSH(t) are the solvent-accessible surface areas (SASAs) of the epoxy alone and CSH alone, respectively, at time t, and SASAEpoxy+CSH is the SASA of the combined epoxy/CSH system. Note that all SASAs were computed at a solvent probe radius r of 1.4 using a developed TCL (i.e., the embedded scripting language used in VMD) script implemented in the VMD software. Fig. 6(a) shows the temporal evolution of the epoxy/CSH-surface contact area for the EP/CSH, EPAM/CSH, and EPSiO2/CSH systems, which reveals three distinct stages of interface formation: ① adsorption, ② wetting, and ③ interfacial equilibrium. The contact area increases rapidly during the adsorption stage (0-200 ps), with contact times of approximately 110, 60, and 50 ps recorded for EP/CSH, EPSiO2/CSH, and EPAM/CSH respectively, which suggests that the AM or nano-SiO2 present in the epoxy phase accelerates kinetics associated with the initial adsorption of epoxy molecules onto the CSH substrate. The epoxy continues to spread across the CSH surface during the subsequent wetting phase (200-1000 ps), which further increases the contact area. This process is driven by attractive forces between the atoms on the CSH surface and epoxy molecules, which collectively outcompete the cohesive forces that operate within the bulk epoxy, thereby promoting wetting. EPAM/CSH exhibited a particularly large and uniform contact area owing to the strong affinities of the AM functional groups for the CSH surface. In contrast, EPSiO2/CSH showed more localized wetting behavior, which we attribute to nano-SiO2 aggregation hindering uniform spreading, whereas the EP/CSH system consistently exhibited the smallest contact area among the three systems. Finally, interfacial equilibrium phases were observed from approximately 1000 ps to the end of the simulation (2000 ps). The contact area between the epoxy and CSH surface stabilized during this phase, as clearly evidenced by the plateaus in Fig. 6(a). Sustained plateauing in the contact area provides key quantitative evidence that each interface had reached dynamic equilibrium within the 2000 ps simulation timeframe.
Wettability, which describes the ability of the epoxy molecules to spread over the CSH surface, is critical for interfacial bonding (Fig. 6(b)). Accordingly, we calculated the wetting ratio for each system to assess how each nanofiller influences this property. Wetting ratio was determined from the time-averaged contact area during the final 500 ps of the simulation (i.e., the interfacial equilibrium stage) and normalized against the total CSH surface area (SCSH) [61]. The values presented in Fig. 6(b) are averages from at least three independent simulations, with error bars corresponding to standard deviations.
Fig. 6(b) shows that EPAM delivered the highest wetting ratio (71.05%), followed by EPSiO2 (66.18%) and EP (62.15%). These quantitative results are consistent with the visual data provided by the simulation snapshots presented in Fig. 3, where EPAM and EPSiO2 exhibit larger contact lines and smaller contact angles than EP, indicative of enhanced wetting. Notably, nano-SiO2 appears to enhance epoxy wettability on CSH, although more modestly than AM. This improvement is attributable to two primary contributions from the nano-SiO2 particles. Firstly, their migration to the interface increases interfacial polarity via surface silanol (Si-OH) groups that attract epoxy molecules and enhance molecular interactions, particularly hydrogen bonding, with the CSH surface. Secondly, nano-SiO2 particles alter the surface-energy characteristics of the epoxy system and potentially create a more-favorable interfacial tension that facilitates spreading over inorganic surfaces [47]. However, nano-SiO2 aggregation at the interface (Fig. 5) limits wetting uniformity compared to AM. Despite this, nano-SiO2 is still able to foster more-favorable wetting than neat epoxy, although further treatment to reduce aggregation may enhance its effectiveness. These findings highlight the importance of both filler functionality and dispersibility when optimizing interfacial adhesion, with the organic AM filler providing superior wetting performance compared to its inorganic nano-SiO2 counterpart.
3.2. Structural properties of the epoxy/CSH interfaces
3.2.1. Local density distributions
Fig. 7 shows equilibrium atomic number density profiles of the three epoxy systems (EP, EPAM, and EPSiO2) along the z-direction (orthogonal to the CSH surface). Atomic number densities were computed by binning the simulation box into successive 1 Å-thick slabs using the VDM plugin developed by Giorgino [62]. The profiles depict the combined atomic density of all epoxy components (DGEBF, DETDA, AM, and SiO2 where applicable) in each system as a function of the distance from the CSH surface, which were time-averaged over the last 500 ps of the simulation. The number density of each system exhibited a sharp peak at the interface before experiencing a relatively constant decrease with increasing distance from the CSH surface. The EPAM system exhibited the highest peak intensity (∼0.043 Å−3), with its peak located at ∼4 Å from the surface. This significantly more-intense peak strongly suggests that the addition of AM promotes the adsorption of epoxy molecules, leading to a more densely packed and well-defined interfacial layer. This observation is attributable to several factors: ① The polar AM amido group (carbonyl (C=O) connected to nitrogen (N)) strongly hydrogen bonds to both the CSH surface and epoxy molecules, thereby acting as a bridge between the epoxy monomers and CSH, and ② AM promotes a more-favorable epoxy packing arrangement near the interface. The peak observed for the EPSiO2 system is slightly further from the surface (at 5 Å) and is less intense (0.036 Å−3) compared with that recorded for EPAM. The less-pronounced peak close to the CSH surface is possibly explained by nano-SiO2 aggregation at the interface that may hinder direct contact and create a less-ordered arrangement of epoxy molecules. The EP (neat epoxy) system exhibited the least-intense peak (0.0348 Å−3) at 4 Å from the CSH surface, indicative of weaker adsorption in the absence of a filler. This observation underscores the importance of using an appropriate filler to enhance epoxy/CSH interactions.
These results have significant implications for the design of epoxy-modified CSH composites. The high density of epoxy components at the interface, particularly in the EPAM/CSH system, suggests strong interfacial interactions that are expected to translate into superior mechanical properties.
3.2.2. Concentration profiles
We analyzed the relative concentration profile of each monomer (DGEBF, DETDA, AM, and nano-SiO2) along the z-direction (orthogonal to the CSH surface) to shed further light on the molecular interactions that govern the interfacial behavior of the epoxy systems on CSH. The profiles shown in Figs. 8(a)–(c) for the EP/CSH, EPAM/CSH, and EPSiO2/CSH systems, respectively, provide valuable insight into how each species is distributed at the interface and their preferential adsorption onto the CSH substrate. We divided the simulation cell into successive 1 Å-thick slabs, with the local relative concentration (C) of monomer i in each slab defined as its molar ratio in a given slab normalized against its corresponding molar ratio in the entire bulk epoxy system [26,27,63]. That is, $ C=\frac{N_{\text {slab }}}{V_{\text {slab }}} \cdot \frac{V_{\text {bulk }}}{N_{\text {bulk }}}$, where Nslab is the number of atoms of that monomer in the slab, Vslab is the volume of the 1 Å-thick slab, Nbulk is the total number of atoms of that monomer in the entire liquid epoxy, and Vbulk is the total liquid epoxy volume in the simulation cell. Accordingly, a relative concentration of 1.0 indicates that the concentration of component i is the same in the slab as in the bulk (horizontal dash-dot black line in Figs. 8(a)–(c)); values greater than 1.0 signify enrichment, while values less than 1.0 indicate depletion at a given location.
The DGEBF resin in the neat epoxy system (Fig. 8(a)) exhibits a concentration peak at 4 Å from the CSH surface; this peak corresponds to approximately twice the concentration observed for the bulk EP, which is suggestive of resin enrichment at the interface. In contrast, the DETDA hardener exhibited more significant enrichment at the interface, with a peak relative concentration of ∼3.2 recorded at 4 Å, which indicates that DETDA has a stronger preference for adsorption on the CSH surface than DGEBF. This observation is consistent with previous reports on the preferential adsorption of the MXDA amine hardener over the DGEBA resin on iron-oxide surfaces before [64] and after curing [29]. The observed differences in the DGEBF and DETDA distributions suggest that DETDA may play a more dominant role in the initial stages of interfacial adhesion by anchoring the epoxy to the CSH, whereas DGEBF contributes to overall film formation and bulk properties but with less specific surface interactions. The local relative concentrations of both DGEBF and DETDA converge towards their respective bulk concentrations with increasing distance from the CSH interface (z > ∼10 Å), which suggests that attractive forces involving the CSH surface influence epoxy monomers less.
The addition of AM significantly altered the interfacial composition, as shown in Fig. 8(b). AM is very strongly enriched at the CSH surface, with a peak relative concentration of ∼8.7 occurring at 3 Å from the CSH surface atoms, a value that greatly exceeds its bulk concentration. This peak is observed to be 1 Å closer to the CSH surface than to the epoxy atoms in the EP/CSH system, which suggests that AM filler molecules migrate within the epoxy/CSH van der Waals gap, indicative of a tighter interface. This observation confirms the strong affinity of AM for CSH. Interestingly, the AM present on the system significantly enhances DETDA adsorption while slightly increasing DGEBF adsorption. As a result, DGEBF exhibited a broader peak relative concentration that is about 2.5 times higher than its bulk value, and is located between 5 and 12 Å from the CSH surface, whereas DETDA demonstrated a sharp peak with a value of ∼5.4 at 4 Å, which is approximately five-times higher than its average concentration in the bulk resin. This behavior is possibly rationalized by the AM nanofiller competing with DGEBF for CSH surface-binding sites, while synergistically promoting DETDA adsorption. These findings suggest that AM acts as a coupling agent, with a stronger preference for adsorption on the CSH surface than either DETDA or DGEBF, thereby effectively bridging the epoxy components and the CSH surface. AM exhibited a sharp decrease in concentration (from ∼8.7 to ∼0.4) at z > 3 Å, indicative of the formation of an AM monolayer on the CSH surface. The formation of this monolayer is further supported by the very low concentration of AM in the bulk region (i.e., z > 10 Å), where its relative concentration approaches zero, as evidenced by snapshots shown in Figs. 5(a)-(e).
The nano-SiO2 particles in the EPSiO2 system (Fig. 8(c)) are moderately enriched near the CSH surface when compared to the AM system, with a peak relative concentration of approximately 3.8 at 3 Å. This distribution is likely influenced by several factors, including the aggregation tendencies of the nanoparticles, the balance between particle-particle and particle-surface interactions, and the surface chemistry of the nanoparticles (i.e., silanol groups interacting with the CSH surface). The nano-SiO2 in the system appears to have affected both the DGEBF and DETDA distributions little when compared to the AM in the EPAM system. DGEBF exhibited a peak relative concentration of ∼2.0 at 5 Å, while DETDA showed a peak with a value of ∼2.6 at 4 Å; both relative concentrations gradually decreased and converged toward their respective bulk concentrations with increasing distance from the CSH surface (z > 10 Å). These results suggest that nano-SiO2 influences interfacial properties through a different mechanism to AM. Nano-SiO2 was indeed found to exhibit a more-localized effect (Fig. 5), in which regions containing concentrated particles are primarily affected, with the overall distributions of DGEBF and DETDA influenced less than AM. Nano-SiO2 exhibited a relative concentration that approached zero at distances greater than 15 Å, confirming that most of the nano-SiO2 particles also migrate to the CSH surface, as expected.
Based on the observed concentration profiles of the individual epoxy monomers, we further investigated interfacial compositions by analyzing the relative concentration ratios of DETDA (hardener) to DGEBF (epoxy resin) as functions of distance from the CSH surface (Fig. 8(d)). The dashed line at a ratio of 0.5 depicts the stoichiometric molar ratio for the bulk epoxy formulation. This analysis is critical as it rationalizes deviations from the ideal stoichiometry in the interfacial region; such deviations may significantly impact local curing behavior and resultant interfacial properties. Fig. 8(d) reveals that all three systems exhibit significant hardener (DETDA) enrichment relative to that of the epoxy resin (DGEBF) within the immediate vicinity (0-5 Å) of the CSH surface. Specifically, the neat EP prepolymer exhibited a DETDA/DGEBF peak ratio of approximately 1.6 at 3 Å. This hardener segregation is considerably more pronounced in the EPAM system, which displayed a peak ratio of approximately 3.0 at 3 Å. Such substantial DETDA accumulation in the EPAM system is suggestive of significantly faster interfacial curing and the formation of a distinct, potentially highly crosslinked interphase. In contrast, the DETDA in the EPSiO2 system is modestly enriched, with a peak ratio of approximately 1.5 at ∼4 Å, indicative of a less pronounced change in the interfacial epoxy-monomer composition compared to that of the EPAM system. All three systems exhibit hardener-to-resin ratios that progressively decrease and subsequently fluctuate around the bulk stoichiometric value with increasing distance from the CSH surface (i.e., beyond 5 Å). These overall stoichiometric deviations, particularly the hardener enrichment observed near the CSH surface, are attributable to differential adsorption of DETDA and DGEBF onto the CSH substrate. Moreover, the CSH surface is Lewis acidic, which means that it can accept electron pairs. Conversely, the amino nitrogen in DETDA acts as a Lewis base and is capable of donating electrons [65]. Therefore, the DETDA enrichment observed at the interface is expected to significantly improve interfacial adhesion by donating a substantial number of electrons to the cement surface. This analysis underscores the significant impact of fillers, particularly AM, on the molecular conformations and reactivities of epoxy monomers near the CSH surface, and highlights the importance of understanding these interfacial phenomena when designing tailored epoxy-modified cementitious materials.
3.2.3. Interfacial ionic bonds and hydrogen-bonding networks (H-bonds)
Interfacial epoxy/CSH bonding is significantly influenced by ionic bonds involving Ca2+ on the CSH surface and the electronegative polar atoms of the epoxy, which include Ca-ODGEBF, Ca-NDETDA, Ca-OAM, Ca-NAM, and Ca-OSiO2 ionic bonds, in which ODGEBF, NDETDA, OAM, NAM, and OSiO2 refer to the ether oxygen atom of DGEBF, the amino nitrogen of DETDA, the carbonyl oxygen of AM, the amino nitrogen of AM, and the silanol oxygen of nano-SiO2, respectively. We analyzed their radial distribution functions (RDFs) in order to gain insight into the local microstructures of these interfacial ionic bonds, the results of which are shown in Fig. 9. An RDF is defined as the probability density g(r) of finding atom w at a certain distance from the central atom u, which provides insight into the spatial arrangements and packing efficiencies of the atoms, as well as intermolecular spacings at the interfaces [51]. Fig. 9(a) shows Ca-ODGEBF RDFs for the EP/CSH, EPAM/CSH, and EPSiO2/CSH systems; each RDF shows a sharp peak at 2.85 Å that corresponds to the characteristic Ca-O ionic bond length involving the ether oxygen in DGEBF. However, the various systems show slightly different peak intensities, with values of ∼2.85, ∼3.58, and ∼4.37 for EP/CSH, EPAM/CSH, and EPSiO2/CSH, respectively. A more-intense RDF peak usually signifies a greater probability of finding ether oxygen atoms around Ca2+ at that specific distance. However, slight increases in peak intensity in the presence of fillers may not result in any significant differences in the microstructures of the Ca-O bonds in the neat epoxy. Fig. 9(b) displays Ca-NDETDA RDFs, which reveals that all three systems exhibit an average Ca-NDETDA length of 2.85 Å. Notably, this peaks was much more intense for the EPAM/CSH system (∼11.05) compared to that recorded for EP/CSH (∼7.26) and EPSiO2/CSH (∼4.88), confirming that the interface contains a higher concentration of hardener molecules in the presence of AM compared to the other systems, as shown in Fig. 8. Although Ca-N bonds tend to be slightly weaker than Ca-O bonds [13,66], the significantly higher intensity of the Ca-N peak compared to the Ca-O peak in the presence of AM indicates that the DETDA molecules adopt certain conformations and form local environments that favor strongly locally ordered nitrogen atoms around Ca2+ sites upon adsorption. One plausible mechanistic explanation involves the strong affinity of AM for CSH, coupled with its own polar interactions, influencing the DETDA conformation such that ionic Ca-N binding becomes more energetically favorable than Ca-O binding. Consequently, more amino nitrogen atoms cluster in the immediate Ca2+ coordination shell, resulting in a significantly higher DETDA molar ratio at the interface, as shown in Fig. 8(b). Fig. 9(c) compares the RDFs of Ca2+ and the filler polar atoms in the EPAM/CSH and EPSiO2/CSH systems. The RDF for Ca-OAM exhibits a peak at 2.55 Å with a strikingly high intensity of 21.24, indicative of strong, well-ordered, and localized coordination of the carbonyl oxygen in AM to Ca2+. The amino nitrogen of AM also exhibits a well-defined peak (Ca-NAM), although its bond distance (2.65 Å) is longer and its intensity of 9.20 at that distance is less than that of Ca-OAM, which suggests that AM primarily adsorbs onto the CSH surface via its carbonyl oxygen. These findings are consistent with the concentration profile shown in Fig. 8(b), where AM is sharply enriched near the surface. In comparison, the Ca-OSiO2 bond is 2.85 Å long on average, with a peak intensity of 8.79, which indicates that nano-SiO2 also binds strongly via Ca-O bonds but interacts more weakly with the CSH surface compared to AM.
In addition to ionic bonding, the network of hydrogen bonds (H-bonds) between the epoxy components and the CSH surface plays a crucial role in determining the strength of the interfacial bond in an epoxy-modified CSH system. The CSH surface contains several polar sites capable of hydrogen bonding, including the dangling oxygen atoms of the silicate chains (Si-O−), surface silanol groups (Si-OH), and surface water molecules. Two criteria were used to assess the formation of interfacial H-bonds: ① Whether or not the distance between neighboring donor and acceptor atoms is less than the 3.5 Å cutoff, and ② if the donor-hydrogen-acceptor angle is greater than 120° [51]. The number of interfacial H-bonds (nHB) was calculated over the last 500 ps for three replica simulations for each system to determine the average number of H-bonds between the epoxy molecules and the CSH substrate. Fig. 10 compares total nHB values for the interfaces between the three epoxy systems (EP, EPAM, and EPSiO2) and the CSH substrate, along with the contributions from the individual epoxy components. Notably, the EPAM/CSH system exhibited the highest total nHB (55), which significantly exceeded those of the EPSiO2/CSH (38) and EP/CSH (21) systems. This substantially higher level of hydrogen bonding is directly attributable to the AM present in the EPAM/CSH system. AM contributes 28 hydrogen bonds, which corresponds to approximately 51% of the total nHB in this system and underscores its strong propensity to form hydrogen bonds with the CSH surface; this is likely due to the carbonyl oxygen (C=O) and amino (N-H) groups present in AM that act as hydrogen-bond acceptors and donors, respectively, to form strong hydrogen bonds with dangling oxygen atoms, silanol groups, and water molecules on the CSH surface. This observation is consistent with our previous RDF findings (Fig. 9(c)) that revealed very strong interactions between the oxygen atoms of AM and the calcium ions on the surface. These results further support the notion that AM acts as an effective interfacial coupling agent that promotes the formation of a robust network of hydrogen bonds with the CSH substrate. In a similar manner, nano-SiO2 contributes 20 hydrogen bonds in the EPSiO2/CSH system, which corresponds to approximately 52% of the total interfacial nHB (38) in this system, and highlights the significant role played by nano-SiO2 in enhancing interfacial hydrogen bonding. The hydrogen-bonding capability of nano-SiO2 likely stems from the silanol (Si-OH) groups present on its surface that act as both hydrogen-bond donors and acceptors, to form hydrogen bonds with the dangling oxygen atoms, silanol groups, and water molecules on the CSH surface. The DETDA hardener is the primary contributor to H-bond connections in the absence of a nanofiller, whereas DGEBF contributes relatively few hydrogen bonds in all three systems (three, five, and four bonds in the EP/CSH, EPAM/CSH, and EPSiO2/CSH systems, respectively), consistent with the monomer distributions across their interfaces, as shown in Fig. 8. These observations are also consistent with a previous experimental study [65], in which cement paste coated with either an epoxy resin (DGEBA) or an amine hardener was subjected to XPS. The spectrum of the cement coated with the amine hardener provided evidence of interfacial hydrogen bonding between the hardener and cementitious substrate, whereas that of the cement coated with DGEBA resin did not reveal the existence of H-bonds.
3.2.4. Interaction energies
The interaction energy between the epoxy system and the CSH surface quantifies the adhesion strength at the interface and provides valuable insight into the adsorption behavior and adhesive properties of the system. Eint was calculated using the “compute group/group” command in LAMMPS, which evaluates pairwise interaction energies between specified groups of atoms. Interaction energy is expressed as [31,61,67]:
where Eint is the total energy of the epoxy/CSH interfacial system, ECSH is the energy of the isolated CSH substrate, and Eepoxy is the energy of the isolated epoxy components. A more negative Eint value is indicative of stronger adsorption; hence Eint helps to elucidate the binding mechanisms that operate at the epoxy/CSH interface. Interaction energies were further deconvoluted into electrostatic (Eelec) and van der Waals (EvdW) contributions in order to gain deeper insight into the forces driving epoxy adsorption at the CSH surface. The electrostatic contribution provides a measure of the ionic and polar interactions that operate between the polar epoxy groups and the calcium ions and silanol groups on the CSH surface. The van der Waals component captures the contributions of dispersion forces that operate between the nonpolar regions of the epoxy matrix and the CSH substrate. It should be noted that our analysis focused on the electrostatic and van der Waals contributions as defined within the IFF-PCFF. Readers are referred to the work of Yoshizawa et al. [68] for a more detailed quantum mechanical deconvolution of the interaction energies, including explicit terms such as exchange-repulsion and charge-transfer.
Fig. 11 reveals that all three interfacial systems (EP/CSH, EPAM/CSH, and EPSiO2/CSH) exhibit negative Eint values, which confirms that the epoxy components are spontaneously adsorbed onto the CSH surface. As expected, EP/CSH exhibited the least negative Eint (−4523.74 kJ·mol−1) among the three systems. Notably, the electrostatic (Eelec ≈ 51% of the Eint) and van der Waals (EvdW ≈ 49% of the Eint) components are comparable in magnitude, suggestive of a balanced synergy between polar and nonpolar interactions. The ether oxygen and amino nitrogen functionalities in DGEBF and DETDA provide moderate ionic and H-bond interactions with the calcium and silanol groups in CSH and are complemented by dispersion forces that further promote adsorption. EPSiO2/CSH exhibited a more negative Eint (−5045.20 kJ·mol−1) than EP/CSH, consistent with stronger adsorption that stems from a higher Eelec contribution (≈ 61% of the Eint) and a lower EvdW component (≈ 39% of the Eint) compared to those of the neat epoxy system. This shift in the relative contributions of the interaction components suggests that the enhanced adsorption and wetting of epoxy on the CSH surface in the presence of nano-SiO2 are primarily driven by electrostatic forces, although van der Waals forces still play a significant role. Although the nano-SiO2 additive aggregates at the interface, the silanol groups still engage in favorable ionic and hydrogen bonding with the CSH surface (Fig. 9, Fig. 10) leading to stronger electrostatic attraction at the interface. EPAM/CSH evidently exhibited the most negative interaction energy of −6074.75 kJ·mol−1, which indicates that this system is expected to show the strongest rate of epoxy-component adsorption onto the CSH surface in the presence of the AM filler. This strong adsorption is driven predominantly by electrostatic interactions (Eelec ≈ 64% of the Eint), which stems from the highly polar amido groups in AM that engage in strong ionic and hydrogen bonds with charged species on the CSH surface (mainly calcium ions and silanol groups), as previously discussed in Section 3.2.3. The smaller relative van der Waals contribution, which accounts for 36% of the Eint in the EPAM/CSH system, further supports the notion that enhanced adsorption and wettability are predominantly governed by electrostatic interactions that operate between the epoxy molecules and the polar groups on the CSH surface. Overall, the interaction-energy analysis provides strong quantitative evidence for enhanced interfacial adhesion between the epoxy and CSH in the presence of both organic (AM) and inorganic (nano-SiO2) fillers. Specifically, the significantly higher adsorption energy, particularly the electrostatic component, highlights the crucial role played by AM in strengthening epoxy/CSH bonding through strong electrostatic interactions that include both ionic and hydrogen bonds. This finding is consistent with the observed migration of AM toward the CSH surface, its strong hydrogen-bonding and ionic interactions, and its ability to enhance the adsorption of other epoxy components, particularly DETDA. Although nano-SiO2 also improves the interaction energy when compared to the neat epoxy system, it is less effective than AM.
3.2.5. Work of adhesion
We evaluated the work of adhesion (W) for each interfacial system with the aim of further quantifying the effectiveness of AM and nano-SiO2 in strengthening the epoxy/CSH interface, the results of which are depicted in Fig. 12.W was determined by normalizing the absolute value of the interaction energy against the interfacial contact area; that is, $ W=-\frac{E_{\text {int }}}{A}$ [69], [70], [71], where Eint is the total interaction energy (Eq. (2)) and A is the interfacial contact area (Eq. (1)). Conceptually, W is widely used to quantify adhesion at a fundamental molecular level [69], as it represents the reversible thermodynamic work required to separate the epoxy/CSH interface to infinite distance, thereby serving as a measure of the strength of the interfacial bond [71]. A larger W value is indicative of stronger interfacial adhesion, which implies that more energy is required to separate the surfaces or create new surfaces, and correlates with a higher resistance to interfacial failure [71]. Deconvoluting W into its electrostatic (Welec) and van der Waals (WvdW) contributions further helps to elucidate the relative importance of polar/ionic interactions and dispersive forces toward governing adhesion strength.
Fig. 12 reveals that EP/CSH exhibited the lowest total work of adhesion (Wtotal = 149.02 mJ·m−2), with Welec and Wvdw contributing nearly equally to the interfacial bond strength. The addition of nano-SiO2 raises the total work of adhesion to 155.66 mJ·m−2, which corresponds to a modest 4.45% increase. This slight improvement in adhesion strength is primarily driven by enhanced electrostatic interactions between the silanol groups of nano-SiO2 and the CSH surface, while the van der Waals contribution remains relatively unchanged compared to that of the neat epoxy system. In contrast, the addition of AM to the epoxy significantly increases the work of adhesion (from 149.02 to 173.50 mJ·m−2), which corresponds to a substantial (16.43%) increase relative to that of the EP/CSH system. Such a significant increase in Wtotal highlights the potential for marked improvements in the mechanical performance of the interface, and underscores the pivotal role played by the polar amido groups of the AM; these groups form strong ionic bonds with calcium ions and hydrogen bonds with silanol sites on the CSH substrate, leading to a dominant electrostatic contribution, as discussed in 3.2.3 Interfacial ionic bonds and hydrogen-bonding networks (H-bonds), 3.2.4 Interaction energies, 3.2.5 Work of adhesion. Overall, these results confirm that, while nano-SiO2 provides a slight (∼5%) increase in interfacial adhesion, the AM filler enhances it significantly more (∼17%). Consequently, AM emerges as a more effective additive for improving the strength of the epoxy/CSH interface and, by extension, the mechanical properties and durability of the composite interface. This insight highlights the importance of leveraging polar functionalities, particularly amido groups, when designing next-generation epoxy-modified cementitious materials with superior interfacial performance.
3.3. Implications of the study results
The findings of this study provide critical atomic-level insight into the adsorption, wetting, and interfacial interactions associated with epoxy systems on CSH surfaces in the presence of fillers (epoxy modifiers). This insight has significant scientific and engineering implications for the design and optimization of polymer cementitious composites, particularly for applications such as structural repair, enhancing the adhesion in 3D printed concrete layers, and improving durability. The most important implications of our study results are discussed below.
3.3.1. Selective monomer adsorption and interphase formation
The epoxy hardener (DETDA) adsorbs onto the CSH surface in preference to the resin (DGEBF) despite the 1:2 DETDA/DGEBF molar ratio in the bulk phase prior to adsorption, which is a key finding of this study. This observation is consistent with previous research findings [29,54,60,64] that demonstrate that smaller, more polar molecules tend to preferentially adsorb onto polar substrates. This selective adsorption is particularly pronounced in the presence of AM, leading to the formation of a “hardener-rich” interphase that constitutes a distinct region close to the CSH surface where the concentration of DETDA significantly exceeds that found in the bulk epoxy. The formation of this interphase has significant implications for adhesion, mechanical performance, and durability. Given that local curing reaction rates depend on hardener availability, the elevated DETDA concentration at the interface, particularly within the EPAM/CSH system, suggests that the curing kinetics near the CSH surface differ markedly from that of the bulk epoxy, which is expected to accelerate the polymerization process [72] at the interface, potentially resulting in a highly crosslinked interphase proximal to the CSH surface with mechanical properties distinct from those of the epoxy regions further from the CSH surface. This highly crosslinked interphase is expected to significantly enhance the strength of the epoxy/CSH interface and its resistance to delamination, while increasing the glass transition temperature of the epoxy near the cement-substrate surface [15]. These properties lead to the formation of a stiff interphase [73], [74], [75], whereas the bulk region (i.e., the epoxy region further from the CSH surface) remains more flexible. However, an excessively high crosslinking density within the interphase relative to that of the bulk region may lead to interfacial brittleness and diminish the capacity to dissipate energy [76], which promotes microcrack formation under applied stress [77,78] or leads to interfacial debonding under conditions of dynamic load [79,80]. Therefore, the potential tradeoff between strength and toughness must be carefully considered for each specific application. For example, Ueda [6] noted that both high strength and stiffness are important for non-interventional structures, whereas high fracture strains and low stiffnesses are crucial for structural-intervention materials. Thus, a more brittle interphase may be detrimental because ductility and toughness are paramount. On the other hand, any unreacted amines remaining within the interphase following curing may form a protective coating on the CSH surface, fill surface pores, or even trigger intrinsic self-healing reactions by providing reactive amino curing agents for additional crosslinking when the existing network is disrupted under applied load [81].
3.3.2. Impact of the filler on interfacial properties
The addition of the organic (AM) or inorganic (nano-SiO2) filler to the epoxy matrix significantly alters interactions involving CSH and the resultant interfacial properties. However, these two nanofillers exhibit notable differences in migration behavior, dispersion, and ability to enhance adhesion. AM is more affinitive for the CSH surface and forms a uniform monolayer that effectively bridges epoxy molecules and the CSH substrate. This migration is driven by strong hydrogen-bonding and electrostatic interactions between the amido group of AM and the polar sites of CSH, as evidenced by the RDF and H-bonding analyses (Fig. 9, Fig. 10). The AM-rich monolayer not only enhances interfacial adhesion (by ∼17% compared to neat epoxy) but also promotes the adsorption of the epoxy monomer (particularly DETDA) at the interface. These findings highlight the potential use of functional additives, such as AM, for not only improving adhesion but also for promoting selective monomer adsorption at the interface, thereby enabling better control over curing kinetics and interphase formation. In contrast, nano-SiO2 exhibited a tendency to aggregate while enhancing interfacial adhesion (by ∼5% compared to neat epoxy), leading to a less-uniform interphase layer. Taken together, these results suggests that AM is a more-effective interfacial adhesion promoter, whereas nano-SiO2 is more suitable for mechanical reinforcement rather than directly enhancing interfacial adhesion [82].
The contrasting behavior of AM and nano-SiO2 highlights three critical requirements for fillers to effectively improve epoxy/CSH interfacial bonding:
(1) Complete migration and bridging: The filler must migrate to the epoxy/CSH interface during adsorption and wetting to effectively bridge the CSH and epoxy phases. Both AM and nano-SiO2 meet this requirement (Fig. 5).
(2) Uniform dispersion on the CSH surface: The filler should become uniformly dispersed over the CSH surface to maximize the contact area and promote the homogeneous distribution of stress. AM becomes highly dispersed, unlike nano-SiO2 whose aggregation limits its ability to spread uniformly.
(3) Strong interfacial interactions: The filler must interact strongly with both the CSH surface and epoxy components. AM was found to interact strongly with CSH (Figs. 9(c) and 10) and potentially crosslink with DGEBF. By contrast, nano-SiO2 interacts more weakly with the CSH surface, and previous experimental studies [83] have revealed that epoxy molecules and nano-SiO2 particles interact weakly.
The superior performance of AM over nano-SiO2 in terms of enhancing the work of adhesion underscores the importance of fulfilling all three requirements. While nano-SiO2 fulfills the first requirement and, to some extent, the third, its failure to uniformly disperse impedes its overall effectiveness.
3.3.3. Implications for current design codes and standards
Current engineering design codes and standards for polymer-reinforced cementitious materials primarily focus on macroscopic properties, such as bond strength, durability, and failure mode. However, they often overlook critical molecular-level interfacial interactions, such as selective monomer adsorption, interphase formation, and nanofiller dispersion at the polymer-cement interface, which may govern the performance of these materials, especially when nanofillers are added. This study highlighted the need to integrate molecular-level criteria when designing and optimizing polymer-cementitious materials. Future standards may include specific guidelines for nanofiller selection based on their role in enhancing adhesion, mechanical strength, and interfacial wetting. In addition, performance metrics beyond traditional bulk properties, such as the work of adhesion, wettability, and nanofiller-dispersion uniformity, should be considered. Incorporating these parameters is expected to enable design codes to more effectively account for factors that govern interfacial performance at the nanoscale. Furthermore, polymer-cementitious composite designs may follow a more application-specific approach, in which polymer components and nanofiller incorporation are optimized based on specific structural demands. Integrating this molecular-level insight into construction-material guidelines is expected to improve material reliability and reduced premature failures in infrastructure applications.
3.3.4. Limitations of existing MD simulations of epoxy/CSH interfacial systems
Most, if not all, previous MD simulations of epoxy/CSH systems overlooked the critical influence of interphase formation due to preferential epoxy monomer adsorption, as demonstrated in the current study. Many studies performed epoxy crosslinking independently of modeling the CSH surface [30,32,84,85], thereby failing to capture the enrichment phenomenon of the hardener observed in this study. Such simplifications can lead to inaccurate predictions of interfacial properties and performance. An improved approach involves performing the crosslinking process in-situ, in which the epoxy monomers are first allowed to adsorb and wet the CSH surface before initiating the crosslinking simulation. This procedure reflects interfacial bonding more realistically, and captures how preferential adsorption influences the curing process and the resulting interphase structure. Adopting in-situ crosslinking in future MD simulations is expected to facilitate the development of more-realistic models that better predict epoxy/CSH system behavior, thereby providing a pathway to more accurate computational designs of epoxy-modified cementitious materials and their optimization that ultimately lead to improved performance and durability in real-world applications.
3.4. Design principles that afford superior epoxy/CSH interfacial properties
The new molecular-level insights gained in this study provide a framework for optimizing epoxy/CSH interfacial properties through judicious choice of filler, precise control over the interfacial structure, and tailoring epoxy formulations. Three key principles aimed at guiding the designs of high-performance polymer-cementitious composites are proposed, as summarized in Table 1 [54], [60]: ① optimizing filler-CSH and filler-epoxy interactions, ② ensuring effective filler migration and uniform dispersion, and ③ fine-tuning interfacial characteristics by tailoring the epoxy composition.
A critical factor in enhancing epoxy/CSH adhesion involves selecting a functionalized filler or additive that demonstrates strong interfacial interactions with both the epoxy matrix and the CSH surface, thereby ensuring that the nanofiller acts as an effective bridge between the organic and inorganic phases, thereby mitigating delamination risks, and bolstering the overall bonding strength. This study revealed that fillers with polar functional groups, such as amides (-CONH2), markedly enhance adhesion via hydrogen bonding and electrostatic interactions with CSH. Other organic fillers with strong polar functional groups, such as hydroxyl (-OH) and carboxyl (-COO−), are also good choices [54]. Engineered additives should be selected based on their capacities to enhance ionic and hydrogen bonding with CSH, prevent adhesion-weakening interfacial defects, and improve stress transfer by forming well-integrated interphases. Kim et al. [86] provides an avenue for improving interactions between an inorganic filler, such as nano-SiO2, surface-modified SiO2 (e.g., silanized SiO2), and both epoxy and CSH compared to untreated SiO2, thereby mitigating its tendency to aggregate while still contributing to mechanical reinforcement. Other inorganic fillers (e.g., silicon nitride (Si3N4) nanoparticles) that interact strongly with epoxy polymers [50] may be effective targets for future material-development studies. MD simulations provide an excellent tool for screening suitable fillers during material selection, thereby avoiding time-consuming and resource-intensive trial-and-error laboratory experiments.
Moreover, the epoxy formulation needs to be carefully tailored to balance adhesion strength and toughness. Optimizing the resin-to-hardener ratio in a manner that prevents excessive variations in crosslinking density at the interface is also crucial. For instance, the introduction of controlled amounts of a plasticizing agent or hybrid polymer matrix can further enhance fracture toughness without compromising strong adhesion.
3.5. Limitations of the study
Although this study offers valuable molecular insight into how nanofiller-modified EPs interact with CSH, the findings must be interpreted within the context of specific computational and scope-related limitations. Specifically, this work focused on the critical initial stages of adsorption, wetting, and interphase formation involving epoxy monomers and nanofillers on the CSH surface and did not model the epoxy curing process (polymerization). Although our findings on preferential monomer adsorption suggest that it influences local stoichiometry and curing kinetics, directly simulating curing and the mechanical performance of the cured nanofiller-modified epoxy/CSH interface using reactive force fields is an essential future study that will complement the current study. The quantitative results reported herein (e.g., work of adhesion, degree of monomer segregation) are specific to the DGEBF/DETDA epoxy with 5 wt% AM and nano-SiO2 filler systems, and an idealized 11 Å tobermorite-derived CSH model; variations in epoxy chemistry, filler type and concentration, or CSH characteristics (surface defects, C/Si ratio) may provide different outcomes, although the fundamental molecular-interaction principles are expected to be applicable to a broad range of similar materials. Furthermore, while the CSH model assumed a defined hydrated state (Section 2.1), the amount of water adsorbed on a hydrophilic silicate surface, depends considerably on environmental humidity, and variations in humidity have been shown to influence polymer adhesion and interfacial behavior [87]. Therefore, how the water on the CSH surface influences nanofiller-modified epoxy-system adhesion warrants further investigation. Finally, the observed enrichment of DETDA over DGEBF was primarily attributed to enthalpic factors (e.g., specific interactions with the CSH surface) during our analysis of monomer adsorption; however, entropic factors, including molecular size, shape, and conformational freedom, can also promote monomer segregation at solid surfaces in multicomponent epoxy systems, as reported by Yamamoto and Tanaka [88]. While a holistic thermodynamic analysis in which the free energy of adsorption is deconvoluted into its enthalpic and entropic contributions was beyond the scope of the current study, it represents an important future research direction.
4. Conclusions
In this study, we used MD simulations to provide fundamental insight into the molecular mechanisms associated with adsorption, wetting, and interfacial interactions involving EP systems on a CSH surface, the primary binding phase in concrete. Importantly, this study uniquely focused on the behavior of EPs modified with fillers before the onset of curing, thereby providing a novel perspective that complements existing studies that primarily focused on cured epoxy systems. We elucidated the critical role played by the filler in modulating interfacial behavior and enhancing adhesion by investigating neat epoxy, epoxy modified with AM, and epoxy modified with nano-SiO2. The following conclusions are drawn:
(1) DETDA, the epoxy hardener component, adsorbs on the CSH surface in preference to DGEBF, the resin component, leading to the formation of a hardener-rich interphase that affects the local structure and bonding at the interface.
(2) Interfacial interactions are dramatically altered by the addition of a filler, particularly AM, which migrates to the CSH surface to form a uniform and strongly bound monolayer through a combination of hydrogen bonding and electrostatic interactions. This AM-rich monolayer further promotes the adsorption of DETDA and DGEBF. While nano-SiO2 also enhances interfacial interactions, its tendency to aggregate reduces its effectiveness in promoting uniform wetting and the adhesion of epoxy molecules on the CSH surface compared to AM.
(3) Quantitative analyses of the interaction energies and the work of adhesion confirmed that AM effectively enhances the strength of the interfacial bond, primarily through stronger electrostatic interactions. The work of adhesion of the EPAM/CSH system was calculated to be ∼17% higher than that of the neat epoxy system, which highlights the potential of AM as a highly effective coupling agent for organic-inorganic interfaces such as those found in epoxy-cementitious materials. In contrast, nano-SiO2 only provided a modest improvement in the work of adhesion (4.45%) when compared against neat epoxy.
Based on the findings presented above, we proposed a set of design principles for optimizing the strength of the epoxy/CSH interface through the use of a filler. These principles emphasize the importance of ① selecting a filler with appropriate functionalities that can interact strongly with both the CSH surface and epoxy components; ② promoting filler migration to the interface and ensuring its uniform dispersion; and ③ tailoring the epoxy composition, including the resin-to-hardener ratio, to maximize favorable interactions and control the distribution of components at the interface. Specifically, the use of fillers with polar functional groups, such as the amido group in AM or carboxyl groups is recommended.
The insights gained from this study have significant implications for the design and development of high-performance concrete materials for use in strengthening and repair applications. Engineering epoxy systems that exhibit enhanced interfacial adhesion is possible by gaining an understanding of the fundamental mechanisms governing epoxy/CSH interactions, particularly in the liquid state prior to curing, and applying the proposed design principles, thereby potentially leading to concrete structures that exhibit superior performance and durability. We believe that the development of a more comprehensive framework for the design of next-generation polymer-reinforced cementitious materials with tailored interfacial properties and superior performance is possible by combining advanced nanoscale-level computational modeling with experimental testing
Data availability statement
Data will be made available upon reasonable request.
UedaT, ShimomuraT, MatthewsSL. Conservation and intervention in MC2020. Struct Concr2023; 24(4):4464-75.
[2]
ScrivenerKL, JohnVM, GartnerEM. Eco-efficient cements: potential economically viable solutions for a low-CO2 cement-based materials industry. Cem Concr Res2018; 114:2-26.
[3]
BazliM, HeitzmannM, HernandezBV. Hybrid fibre reinforced polymer and seawater sea sand concrete structures: a systematic review on short-term and long-term structural performance. Constr Build Mater2021; 301:124335.
[4]
ScrivenerKL, KirkpatrickRJ. Innovation in use and research on cementitious material. Cem Concr Res2008; 38(2):128-36.
[5]
LiK, UedaT. Durability-based design: Asian perspectives. Sustain Resilient Infrastruct2023; 8(2):158-68.
[6]
UedaT. Material mechanical properties necessary for the structural intervention of concrete structures. Engineering2019; 5(6):1131-8.
[7]
Brandtner-HafnerM. Evaluating the bonding effectiveness of CFRP patches in strengthening concrete structures. Constr Build Mater2024; 436:136966.
[8]
LvL, GuoP, LiuG, HanN, XingF. Light induced self-healing in concrete using novel cementitious capsules containing UV curable adhesive. Cem Concr Compos2020; 105:103445.
[9]
WangX, ChenJ, LiuY, ShengM, LongW, LuoQ, et al. Experimental study on bending fatigue properties of microcapsule self-healing cementitious composites. J Build Eng2025; 101:111873.
[10]
DaneshvarD, BehnoodA, RobissonA. Interfacial bond in concrete-to-concrete composites: a review. Constr Build Mater2022; 359:129195.
[11]
DaneshvarD, DeixK, RobissonA. Effect of casting and curing temperature on the interfacial bond strength of epoxy bonded concretes. Constr Build Mater2021; 307:124328.
[12]
PortelaGM, PachecoF, EhrenbringHZ, ChristR, TutikianB, MancioM. Effects of different coatings on concrete elements due to chloride ion penetration. Coatings2025; 15(1):46.
[13]
HosseiniE, ZakertabriziM, Habibnejad KorayemA, ZakerZ, ShahsavariR. Orbital overlapping through induction bonding overcomes the intrinsic delamination of 3D-printed cementitious binders. ACS Nano2020; 14 (8):9466-77.
[14]
MorschS, WandCR, GibbonS, IrwinM, SipersteinF, LyonS. The effect of cross-linker structure on interfacial interactions, polymer dynamics and network composition in an epoxy-amine resin. Appl Surf Sci2023; 609:155380.
DjouaniF, ConnanC, ChehimiMM, BenzartiK. Interfacial chemistry of epoxy adhesives on hydrated cement paste. Surf Interface Anal2008; 40(3-4):146-50.
[17]
PalmeseGR, McCulloughRL, SottosNR. Relationship between interphase composition, material properties, and residual thermal stresses in composite materials. J Adhes1995; 52(1):101-13.
[18]
StarostinaIA, NguyenDA, BurdovaEV, StoyanovOV. Adhesion of polymers: new approaches to determination of surface properties of metals. Polym Sci Ser D Glues Sealing Mater2013; 6(1):1-4.
[19]
MorschS, LiuY, MalaninM, FormanekP, EichhornKJ. Exploring whether a buried nanoscale interphase exists within epoxy-amine coatings: implications for adhesion, fracture toughness, and corrosion resistance. ACS Appl Nano Mater2019; 2(4):2494-502.
[20]
SalginB, ÖzkanatÖ, MolJMC, TerrynH, RohwerderM. Role of surface oxide properties on the aluminum/epoxy interfacial bonding. J Phys Chem C Nanomater Interfaces2013; 117(9):4480-7.
[21]
TaheriP, TerrynH, MolJMC. An in situ study of amine and amide molecular interaction on Fe surfaces. Appl Surf Sci2015; 354:242-9.
[22]
BüyüköztürkO, BuehlerMJ, LauD, TuaktaC. Structural solution using molecular dynamics: fundamentals and a case study of epoxy-silica interface. Int J Solids Struct2011; 48(14-15):2131-40.
[23]
LauD, JianW, YuZ, HuiD. Nano-engineering of construction materials using molecular dynamics simulations: prospects and challenges. Compos Pt B Eng2018; 143:282-91.
[24]
HouD, YangQ, WangP, JinZ, WangM, ZhangY, et al. Unraveling disadhesion mechanism of epoxy/CSH interface under aggressive conditions. Cem Concr Res2021; 146:106489.
[25]
HueKY, LewJH, MatarOK, LuckhamPF, MüllerEA. Parametric studies of polyacrylamide adsorption on calcite using molecular dynamics simulation. Molecules2025; 30(2):285.
NouranianS, JangC, LacyTE, GwaltneySR, ToghianiH, Pittman JrCU. Molecular dynamics simulations of vinyl ester resin monomer interactions with a pristine vapor-grown carbon nanofiber and their implications for composite interphase formation. Carbon2011; 49(10):3219-32.
[28]
JiaoW, LiuW, YangF, JiangL, JiaoW, WangR. Improving the interfacial strength of carbon fiber/vinyl ester resin composite by self-migration of acrylamide: a molecular dynamics simulation. Appl Surf Sci2018; 454:74-81.
[29]
MorschS, LiuY, HarrisK, SipersteinFR, Di LulloC, VisserP, et al. Probing the nanostructure and reactivity of epoxy-amine interphases. ACS Appl Mater Interfaces2024; 16(50):70097-107.
[30]
HouD, YuJ, LiuQ, DongB, WangX, WangP, et al. Nanoscale insight on the epoxy-cement interface in salt solution: a molecular dynamics study. Appl Surf Sci2020; 509:145322.
[31]
LuoQ, XieY, NieQ, YangQ, ZhangX. Impact of pore sizes and defects on C-S-H/epoxy bonding in wet conditions: a molecular dynamics analysis. Next Mater2024; 5:100253.
[32]
BahraqAA, Al-OstaMA, Al-AmoudiOSB, SalehTA, ObotIB. Atomistic simulation of polymer-cement interactions: progress and research challenges. Constr Build Mater2022; 327:126881.
[33]
ThompsonAP, AktulgaHM, BergerR, BolintineanuDS, BrownWM, CrozierPS, et al. LAMMPS—a flexible simulation tool for particle-based materials modeling at the atomic, meso, and continuum scales. Comput Phys Commun2022; 271:108171.
StukowskiA. Visualization and analysis of atomistic simulation data with OVITO—the open visualization tool. Model Simul Mater Sci Eng2010; 18 (1):015012.
[36]
PellenqRJM, KushimaA, ShahsavariR, van VlietKJ, BuehlerMJ, YipS, et al. A realistic molecular model of cement hydrates. Proc Natl Acad Sci USA2009; 106(38):16102-7.
[37]
HouD, MaH, ZhuY, LiZ. Calcium silicate hydrate from dry to saturated state: structure, dynamics and mechanical properties. Acta Mater2014; 67:81-94.
[38]
Abdolhosseini QomiMJ, KrakowiakKJ, BauchyM, StewartKL, ShahsavariR, JagannathanD, et al. Combinatorial molecular optimization of cement hydrates. Nat Commun2014; 5(1):4960.
[39]
MurraySJ, SubramaniVJ, SelvamRP, HallKD. Molecular dynamics to understand the mechanical behavior of cement paste. Transp Res Rec2010; 2142(1):75-82.
[40]
SelvamRP, SubramaniVJ, MurrayS, HallK. Potential application of nanotechnology on cement based materials. Report. Fayetteville: Mack Blackwell Rural Transportation Center (MBTC); 2009.
[41]
ChenJJ, ThomasJJ, TaylorHFW, JenningsHM. Solubility and structure of calcium silicate hydrate. Cem Concr Res2004; 34(9):1499-519.
[42]
BahraqAA, Al-OstaMA, Baghabra Al-AmoudiOS, ObotIB, AdesinaAY, MaslehuddinM. A nanoscale adhesion mechanism of cement-epoxy interface under varying moisture conditions: a molecular dynamics study. Surf Interfaces2022; 35:102446.
[43]
HouD, LuZ, LiX, MaH, LiZ. Reactive molecular dynamics and experimental study of graphene-cement composites: structure, dynamics and reinforcement mechanisms. Carbon2017; 115:188-208.
[44]
AmornsakchaiT, DoaddaraO. Grafting of acrylamide and acrylic acid onto polyethylene fiber for improved adhesion to epoxy resin. J Reinf Plast Compos2008; 27(7):671-82.
SprengerS. Improving mechanical properties of fiber-reinforced composites based on epoxy resins containing industrial surface-modified silica nanoparticles: review and outlook. J Compos Mater2015; 49(1):53-63.
[47]
KarnatiSR, AgboP, ZhangL. Applications of silica nanoparticles in glass/carbon fiber-reinforced epoxy nanocomposite. Compos Commun2020; 17:32-41.
[48]
OdegardGM, PatilSU, DeshpandePP, KanhaiyaK, WinetroutJJ, HeinzH, et al. Molecular dynamics modeling of epoxy resins using the reactive interface force field. Macromolecules2021; 54(21):9815-24.
[49]
SaraçI, AdinH, TemizS. Investigation of the effect of use of nano-Al2O3, nano-TiO2 and nano-SiO2 powders on strength of single lap joints bonded with epoxy adhesive. Compos Pt B Eng2019; 166:472-82.
[50]
AlhabillFN, AyoobR, AndritschT, VaughanAS. Influence of filler/matrix interactions on resin/hardener stoichiometry, molecular dynamics, and particle dispersion of silicon nitride/epoxy nanocomposites. J Mater Sci2018; 53(6):4144-58.
[51]
FameCM, UedaT, Ntjam MinkengMA, WuC. Durability of epoxy and vinyl ester polymers in wet, seawater, and seawater sea sand concrete environments: molecular dynamics simulations. Constr Build Mater2024; 451:138645.
[52]
HeinzH, LinTJ, Kishore MishraR, EmamiFS. Thermodynamically consistent force flelds for the assembly of inorganic, organic, and biological nanostructures: the INTERFACE force field. Langmuir2013; 29(6):1754-65.
[53]
HeinzO, HeinzH. Cement interfaces: current understanding, challenges, and opportunities. Langmuir2021; 37(21):6347-56.
[54]
JavadiA, JamilT, Abouzari-LotfE, SoucekMD, HeinzH. Working mechanisms and design principles of comb-like polycarboxylate ether superplasticizers in cement hydration: quantitative insights for a series of well-defined copolymers. ACS Sustain Chem & Eng2021; 9(25):8354-71.
[55]
MishraRK, MohamedAK, GeissbühlerD, ManzanoH, JamilT, ShahsavariR, et al. A force field database for cementitious materials including validations, applications and opportunities. Cem Concr Res2017; 102:68-89.
[56]
ComputerVL. “experiments” on classical fluids. I. Thermodynamical properties of Lennard-Jones molecules. Phys Rev1967; 159(1):98-103.
[57]
NoséS. A molecular dynamics method for simulations in the canonical ensemble. Mol Phys1984; 52(2):255-68.
[58]
MishraRK, DarouichS, in’t VeldPJ, FlattRJ, HeinzH. Understanding hydration reactions, mechanical properties, thermal expansion, and organic interfacial interactions of calcium sulfate hydrates from the atomic scale. Cem Concr Res2025; 189:107740.
[59]
IsraelachviliJN. Intermolecular and surface forces. 3rd ed. San Diego: Academic Press; 2011.
[60]
JamilT, JavadiA, HeinzH. Mechanism of molecular interaction of acrylate-polyethylene glycol acrylate copolymers with calcium silicate hydrate surfaces. Green Chem2020; 22(5):1577-93.
[61]
WangS, YanX, ChangB, LiuS, ShaoL, ZhangW, et al. Atomistic modeling of the effect of temperature on interfacial properties of 3D-printed continuous carbon fiber-reinforced polyamide 6 composite: from processing to loading. ACS Appl Mater Interfaces2023; 15(48):56454-63.
[62]
GiorginoT. Computing 1-D atomic densities in macromolecular simulations: the density profile tool for VMD. Comput Phys Commun2014; 185(1):317-22.
[63]
WeiT, XiaoJ, ChengX, GongP, MeiK, HouZ, et al. Improving the mechanical properties of cement-based materials under high temperature: reducing the C3S/C2S ratio. Constr Build Mater2024; 421:135741.
[64]
HarrisK, WandCR, VisserP, SipersteinFR. Segregation in epoxy/amine systems on iron oxide surfaces. RSC Appl Interfaces2024; 1(4):812-20.
[65]
BenzartiK, PerruchotC, ChehimiMM. Surface energetics of cementitious materials and their wettability by an epoxy adhesive. Colloids Surf A Physicochem Eng Asp2006; 286(1-3):78-91.
[66]
KaiMF, JiWM, DaiJG. Atomistic insights into the debonding of epoxy-concrete interface with water presence. Eng Fract Mech2022; 271:108668.
[67]
HueKY, LewJH, Myo ThantMM, MatarOK, LuckhamPF, MüllerEA. Molecular dynamics simulation of polyacrylamide adsorption on calcite. Molecules2023; 28(17):6367.
[68]
YoshizawaK, SemotoT, HitaokaS, HiguchiC, ShiotaY, TanakaH. Synergy of electrostatic and van der Waals interactions in the adhesion of epoxy resin with carbon-fiber and glass surfaces. Bull Chem Soc Jpn2017; 90(5):500-5.
[69]
KisinS, Božović VukićJ, van der VarstPGT, de WithG, KoningCE. Estimating the polymer-metal work of adhesion from molecular dynamics simulations. Chem Mater2007; 19(4):903-7.
[70]
ZhouS, Vu-BacN, ArashB, ZhuH, ZhuangX. Interface characterization between polyethylene/silica in engineered cementitious composites by molecular dynamics simulation. Molecules2019; 24(8):1497.
[71]
NikkhahSJ, MoghbeliMR, HashemianzadehSM. Dynamic study of deformation and adhesion of an amorphous polyethylene/graphene interface: a simulation study. Macromol Theory Simul2016; 25(6):533-49.
[72]
OdagiriN, ShirasuK, KawagoeY, KikugawaG, OyaY, KishimotoN, et al. Amine/epoxy stoichiometric ratio dependence of crosslinked structure and ductility in amine-cured epoxy thermosetting resins. J Appl Polym Sci2021; 138 (23):50542.
[73]
JeyranpourF, AlahyarizadehG, ArabB. Comparative investigation of thermal and mechanical properties of cross-linked epoxy polymers with different curing agents by molecular dynamics simulation. J Mol Graph Model2015; 62:157-64.
[74]
GuhaRD, IdolorO, GraceL. An atomistic simulation study investigating the effect of varying network structure and polarity in a moisture contaminated epoxy network. Comput Mater Sci2020; 179:109683.
[75]
LiC, StrachanA. Molecular dynamics predictions of thermal and mechanical properties of thermoset polymer EPON862/DETDA. Polymer2011; 52 (13):2920-8.
[76]
WuF, ZouH, ZhangQ, ZhangT, YuJ. Combining machine learning and molecular dynamics to predict strength-toughness and energy dissipation mechanisms of hybrid double-crosslinked CNT networks. Comput Mater Sci2025; 246:113403.
FameCM, WuC, FengP, TamL. Numerical investigations on the damage tolerance of adhesively bonded pultruded GFRP joints with adhesion defects. Compos Struct2022; 301:116223.
[79]
WuC, ZhaoXL, ChiuWK, Al-MahaidiR, DuanWH. Effect of fatigue loading on the bond behaviour between UHM CFRP plates and steel plates. Compos Pt B Eng2013; 50:344-53.
LiQ, SiddaramaiahKNH, HuiD, LeeJH. Effects of dual component microcapsules of resin and curing agent on the self-healing efficiency of epoxy. Compos Pt B Eng2013; 55:79-85.
[82]
DittanetP, PearsonRA. Effect of silica nanoparticle size on toughening mechanisms of filled epoxy. Polymer2012; 53(9):1890-905.
[83]
BallerJ, BeckerN, ZiehmerM, ThomasseyM, ZielinskiB, MüllerU, et al. Interactions between silica nanoparticles and an epoxy resin before and during network formation. Polymer2009; 50(14):3211-9.
[84]
ZhouY, HouD, ManzanoH, OrozcoCA, GengG, MonteiroPJM, et al. Interfacial mechanismsconnection in calcium-silicate-hydrates/polymer nanocomposites: a molecular dynamics study. ACS Appl Mater Interfaces2017; 9(46):41014-25.
[85]
BahraqAA, ObotIB, Al-OstaMA, Baghabra Al-AmoudiOS, MaslehuddinM. Molecular-level investigation on the effect of surface moisture on the bonding behavior of cement-epoxy interface. J Build Eng2022; 61:105299.
[86]
KimT, KwonSH, KimHJ, LimCS, ChungI, LeeWK, et al. Effect of the surface modification of silica nanoparticles on the viscosity and mechanical properties of silica/epoxy nanocomposites. Compos Interfaces2022; 29(13):1573-90.
[87]
NakamuraS, TsujiY, YoshizawaK. Molecular dynamics study on the thermal aspects of the effect of water molecules at the adhesive interface on an adhesive structure. Langmuir2021; 37(50):14724-32.
[88]
YamamotoS, TanakaK. Entropy-driven segregation in epoxy-amine systems at a copper interface. Soft Matter2021; 17(5):1359-67.
PDF
(5071KB)
