Thermo-elastic properties of hydrated epoxy-graphene nanocomposites from ensemble-based molecular dynamics simulations

arXiv:2607.15301 · cond-mat.soft, cond-mat.mtrl-sci · Submitted 2026-07-10 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Condensed Matter. The day in condensed matter: materials, phases and the measurements under them.

Kai: We're looking at the paper: "Thermo-elastic properties of hydrated epoxy-graphene nanocomposites from ensemble-based molecular dynamics simulations".

Mira: Here's the gist of the paper. Epoxy-based materials are inherently hygroscopic, absorbing moisture from the environment, which can significantly alter their short and long-term performance.

Kai: Our results reveal a distinct threshold at 3 wt% water content: below this, hydration primarily reduces the glass transition temperature, while mechanical properties remain unaffected.

Mira: Beyond 3 wt%, however, the mechanical properties deteriorate, highlighting a non-linear sensitivity to water uptake.

Kai: Hydration Effects on Water Structure The study investigates the structure of water molecules in the pure epoxy resin through partial radial distribution functions (RDF) and cluster size analysis.

Paper summary: Kai: So we're looking at this paper titled "Thermo-elastic properties of hydrated epoxy-graphene nanocomposites from ensemble-based molecular dynamics simulations." Basically, they’re digging into how water messes with these epoxy materials and how adding graphene changes that.

Mira: Right. The core thesis is that these epoxy materials absorb moisture from the air, and this absorption really shifts their performance over time, both short term and long term. They claim they found a specific threshold at three wt percent water content where things change significantly <ref:2607.15301#pg2>.

Lev: That threshold is important because it’s where hydration stops just subtly changing the material and starts actively breaking its mechanical structure.

Kai: So beyond that three percent, they see a real drop in mechanical properties, not just some slight shift in temperature.

Mira: Exactly. And what's interesting is how the water molecules actually organize themselves inside the epoxy matrix as you add more of them. They use tools like radial distribution functions and cluster size analysis to map out these water structures across different hydration levels, from one percent up to five percent water content.

Kai: I’m curious about the structural findings. What did they find about those water clusters?

Lev: They found that at low hydration, you mostly see small water clusters, less than ten molecules. But as the water content goes up, they start seeing much larger clusters because the polymer sites get saturated and interwater interactions take over.

Mira: That saturation is key. It explains why the glass transition temperature drops down to three hundred eighty Kelvin when they hit that three percent mark, which they attribute to water molecules penetrating in and creating more free volume for the epoxy chains to move around.

Kai: Penetration leading to lower temperature sounds like a classic plasticizing effect. But what about those mechanical properties?

Lev: Well, for the young’s modulus, it stays similar up to two percent water content for both neat epoxy and graphene-epoxy systems.

Mira: That stability up to two percent is a bit surprising because they say beyond that point, the water molecules aren't just sitting on the polar sites anymore; they form these larger clusters that plasticize the whole matrix, reducing all those moduli for both materials.

Kai: And for graphene-epoxy specifically, what happens there?

Lev: The Young’s modulus is similar to pure epoxy up to two percent, but then it drops after that threshold because those larger water clusters are present in both systems. Interestingly, the Poisson ratio actually increases above two percent water content in the graphene system.

Mira: That increase in Poisson ratio is something I hadn't focused on much before. And then there’s the graphene effect on the water structure itself, which seems to be a major part of this paper, "Thermo-elastic properties of hydrated epoxy-graphene nanocomposites from ensemble-based molecular dynamics simulations".

Kai: Right. The paper also looks at how graphene changes that water structure. They found that graphene actually reinforces the aggregation of those water molecules; it doubled the size of the largest water clusters compared to when there was no graphene present.

Paper summary: Lev: That suggests that while graphene might act as a barrier, its presence seems to actually increase the opportunities for those larger water interactions within the matrix, even if it doesn't change how much water gets in overall.

Mira: They also noted that graphene doesn't seem to change the immediate local environment around the epoxy’s polar oxygen atoms; water molecules don't just sit right on top of the graphene sheets. But they did see a longer range order divergence, where the graphene containing systems have more large water clusters compared to pure epoxy.

Kai: So, there's this tension here between graphene acting as a barrier and its presence encouraging larger water organization within the bulk material.

Lev: That’s what they hypothesize—the high surface area of the graphene tactoid might reduce the space available for water dispersion, which in turn increases those water-water interactions.

Mira: It’s a complex interplay between the interface and the bulk structure. And we can't forget how they tackled that uncertainty, because MD simulations give you results, but they aren't perfect predictions.

Kai: That leads us into the next part of this paper about how they handled those uncertainties.

Lev: They stress that to get reliable estimates for elastic moduli, you need a large number of replicas and a long simulation length. For instance, for the five percent water content case, they say that a one thousand replica ensemble shows the Young’s modulus values look like a Gaussian distribution.

Mira: A Gaussian distribution is good—it suggests symmetry around the mean value—but it doesn't guarantee anything about the absolute value itself.

Kai: That uncertainty quantification is crucial for anyone trying to build real hardware, because if you only rely on a single simulation run, you’re going to get wildly different mechanical predictions.

Lev: I agree. And they emphasize that hundreds of replicas are necessary to report converged distributions of elastic properties and by extension reproducible mechanical results. That’s the practical reality for anyone trying to use these models for real-world engineering applications with graphene-epoxy nanocomposites from ensemble-based molecular dynamics simulations.

Mira: So, to wrap up this section, the paper shows that hydration has a non-linear effect on these materials, hitting a mechanical tipping point around three percent water content where things change dramatically.

Kai: And the structural analysis shows that graphene doesn't eliminate the problem; it actually alters how water clusters form in response to that moisture.

Lev: It really highlights how important it is to account for those large water clusters when you’re designing systems like these, because they dictate the softening behavior beyond two percent.

Mira: So, we saw that hydration drives structural changes, and graphene modifies those structural changes by influencing water cluster size distribution. That sets us up perfectly for what the authors are suggesting about how these materials will perform under real environmental stress.

Conclusion: Kai: So, this paper looks at how water messes with epoxy and graphene mixes in those things using molecular dynamics simulations.

Mira: Yeah, it's about those thermo-elastic properties of hydrated epoxy-graphene nanocomposites from ensemble-based molecular dynamics simulations.

Lev: It’s interesting that they are focusing on the ensemble approach for these mechanical property estimates because, honestly, you can't get reliable numbers with just one simulation run when you’re dealing with this kind of complex polymer behavior.

Kai: Right, and the authors are using a lot of simulation work to map out how water molecules actually organize themselves within those systems.

Mira: They show that the biggest structural change they see is the growth of water clusters as moisture gets higher in the epoxy matrix.

Lev: And they make a point about graphene influencing that organization, specifically doubling the size of those largest clusters when graphene is added to it.

Kai: That suggests graphene isn't just sitting there as a barrier; it’s actively changing how water interacts with itself inside the bulk material.

Mira: Exactly, and they highlight that while graphene doesn't change the immediate local environment around the epoxy chains, those larger water clusters are what truly dictate the softening behavior beyond two percent moisture content.

Lev: The authors also stress that to get reproducible mechanical results for these materials, you need hundreds of replicas in their ensemble simulation.

Kai: So, it boils down to hydration causing a structural change that's amplified by the presence of graphene.

Mira: And the main implication is that you have to account for those large water clusters when designing any epoxy system intended for long-term use.

University of Rennes, CNRS, Institut de Physique de Rennes (IMR 6251) · Centre for Computational Science - University College London · Advanced Research Computing Centre - University College London · Hexcel Composites

cond-mat.soft, cond-mat.mtrl-sci

Submitted: 2026-07-10

Updated: 2026-11-12

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 80/100

The gist: Epoxy-based materials are inherently hygroscopic, absorbing moisture from the environment, which can significantly alter their short and long-term performance.

Key concepts

Radial Distribution Function (RDF)
This tool measures how water molecules are distributed around each other and around the epoxy chains. A sharp first peak indicates a stable local structure, while larger peaks show the formation of larger water clusters as more water is added to the system.
Glass Transition Temperature ($T_g$)
The $T_g$ is a measure of when an amorphous material transitions from a hard, glassy state to a softer, rubbery state. Water penetration into the epoxy matrix increases free volume, allowing polymer chains to move more easily, which lowers the $T_g$. This effect plateaus around 3% water content.
Ensemble Size and Uncertainty
Because molecular dynamics predictions have high variability, the study found that hundreds of simulation replicas are needed to get a reliable result. A large ensemble (like 1000 replicas) helps show that the predicted mechanical properties follow a predictable pattern, like a Gaussian distribution.
Water Cluster Size
As moisture increases in the epoxy, water molecules transition from small clusters to much larger ones. This happens because the high-affinity sites on the polymer chains become saturated, leading to increased interwater interactions and larger structural features.

Terminology

Summary

Epoxy-based materials are inherently hygroscopic, absorbing moisture from the environment, which can significantly alter their short and long-term performance. p. 0 Our results reveal a distinct threshold at 3 wt% water content: below this, hydration primarily reduces the glass transition temperature, while mechanical properties remain unaffected. p. 0 Beyond 3 wt%, however, the mechanical properties deteriorate, highlighting a non-linear sensitivity to water uptake. p. 0

Hydration Effects on Water Structure

The study investigates the structure of water molecules in the pure epoxy resin through partial radial distribution functions (RDF) and cluster size analysis. The RDF of water molecules oxygen atom with themselves shows a sharp first peak around 2. p. 58 Å which position remains stable with hydration. p. 5 As water molecules are being added, second and third peak, respectively at 4. p. 55 Å and 6-7 Å in bulk water, appear. p. 7 The cluster size distribution reveals predominantly small water clusters (less than 10 molecules) at low hydration (1 wt%). p. 7 At higher hydration, much larger water clusters are found as the high-affinity polymer site become saturated with water, and interwater interactions begin to dominate. p. 7

Glass Transition Behavior

The glass transition temperature is calculated through a hyperbola fitted across all density measurements, using the method of Patrone et al. The increase of the water content up to 3% reduces the transition temperature of epoxy resins down to 380 K, then the transition temperature remains stable up to 5% water content. p. 9 This reduction is explained by penetration of water molecules into the epoxy matrix creating greater free volume for the epoxy chains to move. p. 9 Since the glass transition temperature plateaus at 3% water, the interrupting of hydrogen bonding within the epoxy chains is likely to be the dominant factor in the reduction at low water contents. p. 9

Mechanical Property Changes

The Young’s modulus is similar for both systems, indicating minimal effect of the graphene sheets. The evolution of mechanical properties with hydration show that both neat epoxy and graphene-epoxy remain stable upto 2% water content. p. 0 Beyond this threshold, the water molecules are no longer primarily bound to the polar sites on the epoxy chains; the larger water clusters plasticize the matrix reducing the Young’s, bulk and shear moduli for both graphene containing and pure epoxy systems. p. 8 Inversely, the Poisson ratio increases above 2 wt% water content, with the graphene system consistently higher. p. 8

Ensemble Size and Uncertainty

The study rigorously determines the parameters required to produce estimates of the elastic moduli of the epoxy resin systems, such as the number of replicas and the length of simulation. The large variability of mechanical properties predictions suggests that hundreds of replicas are necessary to report converged distributions of elastic properties and, by extension, reproducible mechanical results. p. 0 For the highly hydrated epoxy (5% water content), a 1000 replica ensemble shows the distribution of the Young’s modulus values resemble a Gaussian distribution, showing symmetry and centered around the mean value. Graphene-reinforced epoxy resins indeed display higher mechanical properties, but suffer from identical softening as their pristine counterpart. p. 1 p. 1 p. 1

Graphene's Influence on Structure

The addition of graphene reinforced the aggregation of the water molecules, with double the size of the largest water clusters produced in presence of graphene. p. 4 The RDF between water molecules oxygen atoms and various epoxy chains atom types are not affected by the insertion of graphene. p. 7 The presence of graphene does not affect the local hydration environment around the polar oxygen atoms of the epoxy polymer, nor that, as expected, the water molecules do not reside near the graphene sheets. p. 7 However, longer-range order observations from cluster size analysis display a clear divergence as graphene containing systems contain a higher fraction of larger water clusters compared to the pure epoxy. p. 8 The authors hypothesize that the high surface to volume ratio of the graphene tactoid decreases the area available for water dispersion, thereby increasing the water-water interactions. p. 8

Conclusion on Material Performance

Overall, our study emphasizes and clarifies the role of water molecules and hydration on the reduction of elastic properties of epoxy-based resins and nanocomposites. Graphene-reinforced epoxy resins indeed display higher mechanical properties, but suffer from identical softening as their pristine counterpart. The effect of hydration was particularly striking beyond 2% to 3% hydration. p. 1 p. 1 The authors conclude that hundreds of replicas are necessary to report converged distributions of elastic properties and, by extension, reproducible mechanical results. p. 0 p. 0

Acknowledgements

This research was equally co-funded by the Engineering and Physical Sciences Research Council (EPSRC) and Hexcel. We acknowledge funding support from the European Union’s Horizon 2020 research and innovation program under Grant Agreement 800925 (VECMA project, www.vecma.eu), from the UK Engineering and Physical Sciences Research Council under Grant Agreements EP/W007711/1 (SEAVEA project, www. p. 0seavea-project.org), and the European Union’s Horizon 2020 Research and Innovation Programme under grant agreement 823712 (CompBioMed2, compbiomed.eu). We are grateful to the United States Department of Energy (DOE) via the Argonne Leadership Computing Facility for providing access to and node hours on Aurora under a DOE INCITE award in the years 2024-2025. p. 0

References

[1] R. J. Morgan, J. E. O’neal, D. L. Fanter, The effect of moisture on the physical and mechanical integrity of epoxies, Journal of Materials Science 15 (1980) 751–764. p. 0

[2] L.-h. Tam, D. Lau, Moisture effect on the mechanical and interfacial properties of epoxy-bonded material system: An atomistic and experimental investigation, Polymer 57 (2015) 132–142. p. 1

[3] Seung Geol Lee, Seung Soon Jang, Jongman Kim, G. Kim, Distribution and Diffusion of Water in Model Epoxy Molding Compound: Molecular Dynamics Simulation Approach, IEEE Transactions on Advanced Packaging 33 (2010) 333–339. p. 1

[4] S. Yamamoto, R. Kuwahara, K. Tanaka, Dynamic behaviour of water molecules in heterogeneous free space formed in an epoxy resin, Soft Matter 17 (2021) 6073–6080. p. 1

[5] S. Pandiyan, J. Krajniak, G. Samaey, D. Roose, E. Nies, A molecular dynamics study of water transport inside an epoxy polymer matrix, Computational Materials Science 106 (2015) 29–37. p. 0

[6] D. Zhang, K. Li, Y. Li, H. Sun, J. Cheng, J. Zhang, Characteristics of water absorption in amine-cured epoxy networks: a molecular simulation and experimental study, Soft Matter 14 (2018) 8740–8749.

[7] J. Hou, L. Zou, Z. Han, L. Zhang, Molecular Dynamics Simulation of Key Physical Properties of Cross-Linked Epoxy Resins Saturated with Water Absorption, in: 2024 IEEE 7th International Electrical and Energy Conference (CIEEC), IEEE, Harbin, China, 2024. p. 2

[8] T. Frömbgen, E. Surzhikova, J. Dölz, J. Proppe, B. Kirchner, C. R. Jacob, Uncertainty quantification for in silico chemistry, Chemical Reviews 126 (2026) 4189–4236.

[9] C. Sheng, G. Wu, X. Sun, S. Liu, Molecular Dynamics Investigation of the ThermoMechanical Properties of the Moisture Invaded and Cross-Linked Epoxy System, Polymers 14 (2021) 103. p. 2

[10] S. Peretz Damari, L. Cullari, D. Laredo, R. Nadiv, E. Ruse, R. Sripada, O. Regev, Graphene and boron nitride nanoplatelets for improving vapor barrier properties in epoxy nanocomposites, Progress in Organic Coatings 136 (2019) 105207. p. 0

[11] D. G. Papageorgiou, I. A. Kinloch, R. J. Young, Mechanical properties of graphene and graphene-based nanocomposites, Progress in Materials Science 90 (2017) 75–127. p. 0

[12] J. L. Suter, M. Vassaux, P. V. Coveney, Large-scale molecular dynamics elucidates the mechanics of reinforcement in graphene-based composites, Advanced Materials 35 (2023) 2302237. p. 1

[13] H.-K. Liu, Y.-C. Wang, T.-H. Huang, Moisture effect on mechanical properties of graphene/epoxy nanocomposites, Journal of Mechanics 32 (2016) 673–682. p. 0

[14] W. Li, L. Zhang, M. Zhang, S. Chen, Structures of Graphene-Reinforced Epoxy Coatings and the Dynamic Diffusion of Guest Water: A Molecular Dynamics Study, Industrial & Engineering Chemistry Research 59 (2020) 20749–20756. p. 0

[15] S. Kwon, M. Y. Lee, S. Yang, Molecular dynamics approach on the hygroelastic behavior of epoxy/graphene nanocomposites, Journal of Mechanical Science and Technology 33 (2019) 741–747. p. 0

[16] S. Yang, S. Kwon, M. Y. Lee, M. Cho, Molecular dynamics and micromechanics study of hygroelastic behavior in graphene oxide-epoxy nanocomposites, Composites Part B: Engineering 164 (2019) 425–436. p. 0

[17] X. Zhang, G. He, H. Yao, X. Wang, G. Ma, J. Li, Z. Yu, G. Lu, Z. Gao, Effect of graphene on the properties of epoxy in hygrothermal environment by molecular dynamics method, Electronic Research Archive 31 (2023) 3510–3533. p. 0

[18] S. Wan, R. C. Sinclair, P. V. Coveney, Uncertainty quantification in classical molecular dynamics, Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 379 (2021) rsta. p. 22020.0082, 20200082.

[19] M. Vassaux, S. Wan, W. Edeling, P. V. Coveney, Ensembles Are Required to Handle Aleatoric and Parametric Uncertainty in Molecular Dynamics Simulation, Journal of Chemical Theory and Computation 17 (2021) 5187–5197. p. 2

[20] J. L. Suter, W. A. Müller, M. Vassaux, A. Anastasiou, M. Simmons, D. Tilbrook, P. V Coveney Rapid, Accurate and Reproducible Prediction of the Glass Transition Temperature Using Ensemble-Based Molecular Dynamics Simulation Journal of Chemical Theory and Computation 21 (2025) 1405–1421. p. 2

[21] H. Sun, S. J. Mumby, J. R. Maple, A. T. Hagler, An ab Initio CFF93 All-Atom Force Field for Polycarbonates, Journal of the American Chemical Society 116 (1994) 2978–2987. p. 2

[22] J. R. Maple, M. Hwang, T. P. Stockfisch, U. Dinur, M. Waldman, C. S. Ewig, A T Hagler Derivation of class II force fields Journal of Computational Chemistry 15 (1994) 162–182. p. 2

[23] M. J. Hwang, T. P. Stockfisch, A T Hagler Derivation of Class II Force Fields CFF93 for the Alkyl Functional Group and Alkane Molecules Journal of the American Chemical Society 116 (1994) 2515–2525. p. 2

[24] K. K. Bejagam, C. N. Iverson, B. L. Marrone, G. Pilania, Molecular dynamics simulations for glass transition temperature predictions of polyhydroxyalkanoate biopolymers, Physical Chemistry Chemical Physics 22 (2020) 17880–17889. p. 0

[25] Materials Design MedeA, 2024 URL: https://www. p. 0materialsdesign.com/medeasoftware.

[26] J.-P. Ryckaert, G. Ciccotti, H. J. Berendsen, Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes Journal of Computational Physics 23 (1977) 327–341.

[27] P. N. Patrone, A. p. 4 Dienstfrey, A. R. Browning, S. Tucker, S. Christensen, Uncertainty quantification in molecular dynamics studies of the glass transition temperature, Polymer 87 (2016) 246–259. p. 9

[28] H. Ribeiro, W. M. Silva, M.-T. F. Rodrigues, J. C. Neves, R. Paniago, C Fantini, H D R Calado, L M Seara G G Silva Glass transition improvement in epoxy/graphene composites Journal of Materials Science 48 (2013) 7883–7892. p. 0

[29] S.-C. Shiu, J.-L. Tsai, Characterizing thermal and mechanical properties of graphene/epoxy nanocomposites, Composites Part B: Engineering 56 (2014) 691–697. p. 1

[30] D. Freedman, P. Diaconis, On the histogram as a density estimator:L 2 theory, Zeitschrift für Wahrscheinlichkeitstheorie und Verwandte Gebiete 57 (1981) 453–476.

[31] W. Edeling, M. Vassaux, Y. Yang, S. Wan, S. Guillas, P V Coveney Global ranking of the sensitivity of interaction potential contributions within classical molecular dynamics force fields npj Computational Materials 10 (2024) 87. p. 0

Summary of Key Findings

[No specific summary text provided in the prompt, so this section will be based on the paper's findings as requested.]

The main structural consequence of hydration was the growing size of water clusters in the resin and the nanocomposite. The addition of graphene reinforced the aggregation of the water molecules, with double the size of the largest water clusters produced in presence of graphene. Graphene-reinforced epoxy resins indeed display higher mechanical properties, but suffer from identical softening as their pristine counterpart. The effect of hydration was particularly striking beyond 2% to 3% hydration.

Uncertainty Quantification

The study demonstrates that hundreds of replicas are necessary to report converged distributions of elastic properties and, by extension, reproducible mechanical results. For the highly hydrated epoxy (5% water content), a 1000 replica ensemble shows the distribution of the Young’s modulus values resemble a Gaussian distribution, showing symmetry and centered around the mean value. The authors conclude that hundreds of replicas are necessary to report converged distributions of elastic properties and, by extension, reproducible mechanical results.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Ensemble-based uncertainty quantification for mechanical properties allows AI systems to produce reproducible mechanical results by demonstrating that hundreds of replicas are necessary to report converged distributions of elastic properties and, by extension, reproducible mechanical results.

  2. The ability to quantify aleatoric uncertainty through ensemble averaging enables the development of more reliable predictive models for heterogeneous materials where the stress tensor values, and by extension, the elastic moduli, calculated from finite deformations have significantly high variance and act chaotically.

  3. AI systems can predict structural consequences of hydration by analyzing water cluster size distributions, specifically learning that The addition of graphene reinforced the aggregation of the water molecules, with double the size of the largest water clusters produced in presence of graphene.

Related papers