Epoxy resins are the quiet workhorses of modern industry. The carbon fiber panels in aircraft wings, the blades of wind turbines, and countless high-performance structural components all depend on these thermosetting polymers, whose densely crosslinked molecular networks deliver exceptional strength, chemical resistance, and thermal stability. Yet that same crosslinked architecture is a curse at the end of a product’s life: once cured, conventional epoxy cannot be melted, reshaped, or meaningfully recycled, and the dominant petroleum-derived resin, bisphenol A diglycidyl ether, adds a petrochemical footprint to the disposal problem. A new study published in Advanced Science reports a bio-based alternative that appears to dissolve this long-standing trade-off, using an unexpected design lever—nanoscale empty space within the polymer network—to achieve strength, toughness, and recyclability in a single material.
The research team, led by Di Zhao, Kai Dong, and Chengji Zhao, constructed what are known as vitrimers: crosslinked polymers that contain dynamic covalent bonds capable of exchanging with one another when heated. Unlike conventional thermosets, whose networks are permanently frozen, vitrimers can rearrange their topology, allowing them to be self-healed, welded, reprocessed, and chemically degraded on demand. Dynamic bonds such as disulfides, Schiff bases, and siloxane linkages have been incorporated into epoxy systems before, but a persistent problem has haunted the field: making a network more dynamic usually makes it weaker. High crosslinking density delivers stiffness and strength but immobilizes the polymer chains, producing brittleness and sluggish bond exchange. Lowering the crosslink density restores mobility but sacrifices mechanical robustness. Most attempts to escape this dilemma have relied on rubber toughening, thermoplastic blending, or inorganic nanofillers—additives that complicate the formulation, disrupt network homogeneity, and can themselves impair recyclability.
The team’s strategy was different. Rather than adding anything, they engineered the intrinsic free volume of the network—the fraction of space within the material that is not occupied by polymer chains. Free volume is often treated as an incidental byproduct of molecular packing, but the researchers proposed treating its size and spatial distribution as a tunable structural parameter in its own right. To test the idea, they designed two renewable-resource-derived epoxy monomers: a bifunctional monomer called VAN-EP, built from vanillin, and a trifunctional monomer called PCA-EP, derived from protocatechualdehyde. Both carry rigid conjugated Schiff base scaffolds formed by amine-aldehyde condensation, and both were cured with 1,3-bis(3-aminopropyl) tetramethyldisiloxane, a flexible siloxane-containing amine, producing rigid-flexible networks held together by two types of dynamic bonds: imine and siloxane.
Molecular dynamics simulations revealed how profoundly the monomer functionality shaped the internal architecture. The bifunctional VAN-BAS system exhibited a fractional free volume of 38.3 percent, roughly 29 percent higher than the 29.7 percent of the trifunctional PCA-BAS network. The denser trifunctional system packs its chains more tightly because its excess crosslinking points impose localized confinement. The consequences rippled through every property measured. The diffusion coefficient of VAN-BAS, a measure of translational molecular mobility, was 12 percent higher than that of PCA-BAS, and its van der Waals interaction energy was substantially more negative, indicating that moderate free volume actually allows chains to arrange more efficiently and stabilize intermolecular interactions rather than simply loosening the structure.
Mechanically, the two materials diverged in exactly the way the free-volume framework predicts. PCA-BAS, with a crosslink density of 3880 moles per cubic meter—41 percent higher than VAN-BAS—showed greater stiffness, with a storage modulus of 5115 megapascals and a glass transition temperature of 127 degrees Celsius. But its restricted segmental mobility made it brittle. VAN-BAS, by contrast, achieved a tensile strength of 94.3 megapascals, a flexural strength of 168.7 megapascals, and an impact strength of 39.1 kilojoules per square meter, surpassing PCA-BAS by 26, 31, and 46 percent respectively. Fracture surface analysis told the story visually: VAN-BAS broke with a rough, fish-scale morphology studded with fine wrinkles, the signature of extensive plastic deformation and energy dissipation, while PCA-BAS displayed the smooth, flat river-pattern characteristic of brittle fracture. The nanoscale voids in VAN-BAS act simultaneously as shock absorbers, accommodating local chain deformation and relieving stress concentrations, and as molecular lubricants that enable efficient energy dissipation during crack propagation.
Thermal performance remained competitive despite the differences in network density. The thermal decomposition temperatures at five percent weight loss were 294 degrees Celsius for VAN-BAS and 307 degrees for PCA-BAS, a modest 13-degree gap given the substantial difference in crosslinking. The researchers attribute this resilience to the strong intermolecular interactions of the conjugated imine bonds, which suppress thermal molecular motion. In PCA-BAS, the restricted chain mobility even promotes formation of a stable carbonaceous char layer at high temperature, which acts as a thermal barrier and raises the char yield.
The most striking results concerned dynamics. Counterintuitively, PCA-BAS, despite containing a higher total concentration of dynamic covalent bonds, relaxed stress more slowly than VAN-BAS. When heated from 140 to 200 degrees Celsius, the stress relaxation time of VAN-BAS dropped from 426 to 45 seconds, while PCA-BAS required 786 to 65 seconds over the same range. The activation energy for topological rearrangement was 59.6 kilojoules per mole for VAN-BAS versus 70.5 for PCA-BAS, and the extrapolated topological freezing temperature was nearly 19 degrees lower for the freer network. The lesson is that bond concentration alone does not govern network reorganization kinetics; the dense crosslinking of PCA-BAS imposes steric constraints that raise the energy barrier for bond exchange and outweigh the benefit of having more dynamic bonds to exchange. In self-healing tests, a scratch on VAN-BAS closed almost completely within 60 minutes at 200 degrees Celsius, while PCA-BAS managed only a 77 percent width reduction in the same time.
Reprocessing confirmed the practical payoff. Pulverized VAN-BAS hot-pressed at 200 degrees Celsius and 15 megapascals for 30 minutes reformed into smooth, coherent specimens with retained chemical structure and stable thermomechanical properties, though tensile strength fell to 64 percent of the original after one cycle and 57 percent after two—better retention than PCA-BAS, which dropped to 61 and then 50 percent, but a reminder that repeated thermal remolding still accumulates damage. The team suggests that optimized reprocessing conditions, additional dynamic mechanisms, or nanofillers could improve retention in future work.
The real-world test came with carbon fiber-reinforced composites, the application where epoxy’s recycling problem is most acute. Large composite panels measuring 300 by 300 millimeters were fabricated by compression molding, and the CF/VAN-BAS laminates achieved a tensile strength of 605 megapascals, a flexural strength of 628 megapascals, and an interlaminar shear strength of 68 megapascals—exceeding the CF/PCA-BAS equivalents by 11, 21, and 24 percent. The dynamic network enabled adhesive-free welding of overlapping laminates with a lap shear strength of 31.3 megapascals, competitive with or superior to recent advanced welding methods, and the welded joints could sustain a 3-kilogram load, roughly a thousand times their own weight. Laminates were also thermoformed into a stable W-shaped three-dimensional component and then flattened back without delamination, with the reconfigured parts bearing compressive peak loads of 748 newtons.
Perhaps most consequentially for sustainability, the free-volume-tailored matrix degraded gently under solvent. Immersion in dimethylformamide at 120 degrees Celsius caused the network to swell and fragment within about four hours, releasing the carbon fibers intact. The recovered fibers showed smooth, defect-free surfaces under electron microscopy and retained 96 percent of their original tensile strength, with surface chemistry and graphitization degree essentially unchanged according to X-ray photoelectron and Raman spectroscopy. At least 90 percent of the solvent was recoverable by rotary evaporation, and the process required no catalyst, though the authors note that full life-cycle and techno-economic assessment remains future work. Taken together, the study establishes fractional free volume as a genuine design principle for covalently adaptive polymers—one that decouples strength from toughness and dynamics from density, and points toward wind blades, aerospace panels, and structural adhesives that can be welded, reshaped, and unmade rather than landfilled.
Subject of Research: Free-volume engineering of bio-based epoxy vitrimers for recyclable carbon fiber composites
Article Title: High‐Performance Bio‐Based Epoxy Vitrimers and Composites: Synergizing Robustness and Reprocessability via Fractional Free Volume Regulation
Article References: Dong, K., Zhao, D., & Zhao, C. (2026). High‐Performance Bio‐Based Epoxy Vitrimers and Composites: Synergizing Robustness and Reprocessability via Fractional Free Volume Regulation. Advanced Science, Article e77917. https://doi.org/10.1002/advs.77917
Image Credits: AI Generated
DOI: 10.1002/advs.77917
Keywords: epoxy vitrimers, bio-based polymers, free volume, carbon fiber composites, self-healing, recyclability, dynamic covalent bonds, Schiff base, siloxane, vanillin, stress relaxation, sustainable materials
Cite Scienmag News
APA MLA Chicago
Denise Maddox. (October 2, 2026). Empty Voids, Full Potential: Free-Volume Engineering Unlocks Recyclable Bio-Based Epoxy That Rivals Petroleum Plastics. Scienmag. https://scienmag.com/empty-voids-full-potential-free-volume-engineering-unlocks-recyclable-bio-based-epoxy-that-rivals-petroleum-plastics/
Denise Maddox. “Empty Voids, Full Potential: Free-Volume Engineering Unlocks Recyclable Bio-Based Epoxy That Rivals Petroleum Plastics.” Scienmag, 2 October 2026, https://scienmag.com/empty-voids-full-potential-free-volume-engineering-unlocks-recyclable-bio-based-epoxy-that-rivals-petroleum-plastics/. Accessed 2 October 2026.
Denise Maddox. “Empty Voids, Full Potential: Free-Volume Engineering Unlocks Recyclable Bio-Based Epoxy That Rivals Petroleum Plastics.” Scienmag. October 2, 2026. https://scienmag.com/empty-voids-full-potential-free-volume-engineering-unlocks-recyclable-bio-based-epoxy-that-rivals-petroleum-plastics/
Copy citation Download RIS
Tags: advanced polymer network designbio-based polymersbio-derived alternatives to petroleum plasticscarbon fiber compositeschemical degradation of crosslinked polymersdynamic covalent bondsenvironmentally friendly epoxy materialsepoxy vitrimersfree volumefree-volume engineering in polymersnanoscale empty space in epoxy networksrecyclabilityrecyclable bio-based epoxyreshaping and reprocessing of thermosetsSchiff baseself-healingsiloxanestrength and toughness in recyclable epoxiesstress relaxationsustainable epoxy resinssustainable materialsthermosetting polymers with recyclabilityvanillinvitrimers with dynamic covalent bonds
