NR/EPDMd/SiO 2 elastomeric compounds: mechanical, microstructural and interfacial insights into a silica reinforcement mechanism

1 Introduction

The growing demand for more sustainable polymeric materials and the tightening of international climate and waste policies have intensified research into the recycling and valorization of vulcanized rubbers, particularly in high-volume sectors such as automotive and civil construction (United Nations Climate Change Conference COP30, 2025; Markets and Markets, 2025; Chem Analyst, 2026; Fazli and Rodrigue, 2020). Among the industrially most relevant elastomers are natural rubber (NR) and ethylene–propylene–diene monomer rubber (EPDM) (Markets and Markets, 2025; Chem Analyst, 2026). NR is valued for its high tensile strength, resilience and excellent dynamic behavior, derived largely from its ability to crystallize under strain, whereas EPDM offers outstanding resistance to oxidation, ozone, UV radiation and weathering, which makes it the material of choice for roofing membranes, seals and automotive weather-strips. The end-of-life management of these elastomers is, however, problematic: the three-dimensional sulfur network that provides their useful properties also prevents them from being simply re-melted and reprocessed (Fazli and Rodrigue, 2020; Wiśniewska et al., 2022).

In this context, devulcanization is a particularly promising technological route, since it allows the selective scission of the sulfur-sulfur (S–S) and carbon-sulfur (C–S) crosslinks of the network while largely preserving the main polymer backbone (Fazli and Rodrigue, 2020; Dorigato et al., 2023). A variety of devulcanization technologies have been developed and scaled up, including thermomechanical and mechanochemical treatments, ultrasonic and microwave-assisted processes, and continuous twin-screw extrusion (Brunella et al., 2022; Colom et al., 2021; Dorigato et al., 2023; Gschwind and Jordan, 2023; Yun et al., 2004). Microwave and combined thermomechanical/microwave procedures are especially attractive for EPDM waste because the polar additives and carbon black present in the compound couple efficiently with the electromagnetic field, enabling rapid and relatively selective de-crosslinking (Colom et al., 2021). The reincorporation of the resulting devulcanized rubber into virgin elastomeric matrices, followed by partial revulcanization, closes the material loop and supports circular-economy targets (Colom et al., 2024; Wiśniewska et al., 2022; Ramarad et al., 2015).

Nevertheless, blending a devulcanized phase with a virgin matrix modifies the final properties in ways that depend strongly on the compatibility between the phases, the resulting morphology and the degree of co-vulcanization (Brunella et al., 2022; Colom et al., 2021; Colom et al., 2023). Devulcanized EPDM (EPDMd) retains a residual crosslink density and a distinct chemical nature with respect to NR, so that at high loadings it tends to interact preferentially with its own chains and to form rigid aggregates and microcavities that impair the mechanical response (Colom et al., 2023; de Sousa et al., 2020). Understanding and controlling this phase behavior is therefore essential to the design of NR/EPDMd compounds with useful and reproducible properties.

A widely used strategy to improve the modulus, strength and dynamic behavior of elastomeric formulations is the addition of reinforcing fillers such as silica (SiO₂). The polymer-filler interaction and the state of silica dispersion decisively influence the mechanical and thermal response of the compound (Azizi et al., 2020; Lee et al., 2020). In a two-phase system that combines a virgin and a devulcanized elastomer, silica can play a dual role: as a conventional reinforcing filler and as a morphological modulator able to localize around the EPDMd aggregates and to mediate their integration into the matrix. The simultaneous presence of two elastomeric phases with different characteristics, however, may also promote incompatibility, segregation and interfacial phenomena—such as the migration of activators (zinc stearate) — that affect both structural and functional performance (Cañavate et al., 2011; Kaczewiak et al., 2025).

Beyond their mechanical and thermal behavior, elastomers intended for civil-engineering applications can benefit from additional functionalities such as acoustic absorption. Heterogeneous morphologies and the presence of microcavities or porous structures favor energy-dissipation mechanisms through internal friction and viscoelastic losses, which are of interest in potential damping and acoustic applications (Swift et al., 1999; Colom et al., 2014). The relationship between microstructure, recycled-phase content and silica incorporation is thus a key aspect in the design of multifunctional elastomeric compounds. Recent advances in elastomeric composites highlight that controlling the interfacial interactions between silica and rubber phases is critical to achieving effective stress transfer and dispersion, particularly in complex multi-phase or recycled rubber formulations (Cañavate et al., 2000; Bakly et al., 2025). Tailoring the filler surface or interphase architecture can significantly mitigate domain incompatibility and promote physical reinforcement across phase boundaries (Cañavate et al., 2000).

Within this framework, the present work addresses the mechanical, spectroscopic and morphological characterization of compounds based on NR and devulcanized EPDM, with and without silica, in order to evaluate the influence of the recycled-phase content and of the filler on the overall material performance. By correlating internal structure and interfacial phenomena with macroscopic properties, this study aims to establish design criteria for sustainable elastomers.

The valorization of NR/EPDM compounds is further motivated by the breadth of their current and emerging applications. EPDM-based materials dominate automotive sealing systems, weather-strips, door and glass-run seals, hoses and cooling-system components, as well as body and engine anti-vibration mounts; this demand is being reshaped by the electrification of vehicles, which introduces new sealing, thermal-management and noise-vibration-harshness (NVH) requirements. In construction, EPDM is the reference elastomer for single-ply roofing and waterproofing membranes, facade gaskets, structural and bridge bearings, and anti-vibration or railway under-sleeper pads. Beyond these established uses, devulcanized NR and EPDM open a route to sustainable, upcycled products: recycled-rubber compounds are increasingly employed in acoustic and vibration-damping panels (Swift et al., 1999; Colom et al., 2014), sports and playground surfaces, geomembranes and thermoplastic vulcanizates (Cañavate et al., 2011), and in engineering rubber compounds in which a devulcanized phase partially replaces the virgin polymer (Colom et al., 2024; Ramarad et al., 2015). More recent, multifunctional directions like electromagnetic-interference (EMI) shielding, thermal-interface materials, flexible sensing elastomers and green-tyre components incorporating recycled rubber - make the ability to reincorporate a controlled fraction of devulcanized EPDM, without sacrificing mechanical performance (Fazli and Rodrigue, 2020; Wiśniewska et al., 2022). All of these applications share a common requirement: a reliable balance between the recycled-content target and the mechanical and functional performance of the final compound, which is precisely the trade-off addressed in the present study. In this context, this work aims to evaluate the structure–property relationships of natural rubber/devulcanized EPDM waste blends reinforced with silica, proposing a physical interphase mechanism that explains filler localization, chain mobility restriction, and zinc stearate exudation at the phase boundaries.

2 Experimental

2.1 Materials

The EPDM used, employed in roofing membranes, was supplied by Firestone Building Products (Terrassa, Spain). According to the manufacturer, the material contains 100 phr of EPDM, 15 phr of aluminium silicate, 35 phr of carbon black, 15 phr of a non-polar plasticizer and 2 phr of a curing system based on ZnO, N-cyclohexyl-2-benzothiazole sulfenamide (CBS) and sulfur.

These inherent constituents actively influence the final NR/EPDMd blend. The pre-existing fillers (carbon black and aluminum silicate) reinforce the EPDMd phase, making these microdomains stiffer than the surrounding virgin NR matrix. The non-polar plasticizer facilitates EPDMd processability and dispersion during melt compounding. Furthermore, the residual curing agents (ZnO, CBS, and sulfur) can participate in the subsequent revulcanization step, promoting interfacial co-vulcanization and the formation of crosslinks between the NR and EPDMd phases.

Precipitated silica (Ultrasil 7000 GR, Evonik Industries) with a BET specific surface area of 175 m²/g, average primary particle size of 15–20 nm, and purity 98% was used as reinforcing filler.

Natural rubber (NR), benzoyl peroxide (BPO), N-330 carbon black (CB), SiO₂, the vulcanization accelerators tetramethylthiuram disulfide (TMTD), N-tert-butyl-2-benzothiazole sulfenamide (TBBS) and N-cyclohexyl-2-benzothiazole sulfenamide (CBS), stearic acid, zinc oxide (ZnO) and sulfur (S) were supplied by Vigar (Spain). Stearic acid and ZnO act as activators, facilitating and improving the vulcanization process.

2.2 EPDM devulcanization

EPDM devulcanization was carried out by a two-step combined thermomechanical/microwave procedure: a thermomechanical treatment in a Brabender mixer followed by a microwave treatment in a prototype microwave oven fitted with a motorized PTFE stirring system, according to a protocol previously established and validated by our research group (Colom et al., 2021).

2.3 Sample preparation

NR/EPDMd elastomeric compounds were prepared at 70 °C in a Brabender internal batch mixer at a rotor speed of 100 rpm. The total mixing time was 8 min, comprising 2 min of preliminary mastication of the rubber, 4 min of mixing with devulcanized EPDM (0, 20, 40, 50 and 60 phr), 30 phr of carbon black and, in the corresponding series, 15 phr of SiO₂, and a final 2 min with the sulfur curing system.

The same curing system was used for all samples: ZnO 5.0 phr, stearic acid 3.0 phr, TBBS 1.0 phr and a sulfur-based curing agent.

The resulting compound was calendered into a thin film and subsequently compression-moulded into 3 mm-thick plates at 160 °C for 12 min under 4.9 MPa in a Dr. Collin P 200E laboratory press (Germany). Samples were coded as NRXEPDMd(SiO₂), where X is the EPDMd content and the SiO₂ suffix indicates the presence of silica; the reference compound (0 phr EPDMd) was coded NR100. The overall workflow is summarized in Figure 1 and Table 1.

TABLE 1. Composition of the samples.

2.4 Characterization techniques

Tensile properties of the vulcanized NR/EPDMd compounds were determined according to ISO 37 in a high-precision Instron 3366 testing machine (USA) fitted with pneumatic grips (20 kN load cell), at a constant crosshead speed of 500 mm/min, 50% ± 5% relative humidity and 21 °C ± 2 °C.

Hardness was measured with a Zwick 3130 Shore-A durometer (Germany) according to ISO 7619-1. Both tensile and hardness values are the average of five measurements per sample prepared from a single, carefully controlled processing batch for each formulation, ensuring a consistent thermal and processing history across specimens.

The chemical structure was analysed by attenuated-total-reflectance Fourier-transform infrared spectroscopy (FTIR-ATR) using a Perkin-Elmer Spectrum Two spectrometer (USA) equipped with a diamond-crystal ATR accessory; spectra were recorded at 2 cm^-1 resolution and 40 scans over 500–3,500 cm^-1. Surface morphology was examined with a JEOL 5610 scanning electron microscope (Japan); prior to observation, samples were sputter-coated with a thin gold–palladium layer to increase conductivity.

3 Results and discussion

3.1 Mechanical characterization

Table 2 shows the mechanical properties: tensile strength (TS), strain at break (ε_B), Young’s modulus (E) and Shore-A hardness of the NR/EPDMd compounds (0–60 phr EPDMd, with and without 15 phr silica), reported as mean ± standard deviation. Tensile values were obtained from tensile testing (ISO 37) and hardness from Shore-A measurements (ISO 7619-1).

TABLE 2. Mechanical properties and Shore-A hardness of the NR/EPDMd compounds.

The tensile strength (TS) shows two distinct regimes as a function of EPDMd content. At low EPDMd contents, the silica-free samples exhibit higher TS, whereas for EPDMd contents of 40 phr and above the silica-containing samples improve. In both series the TS increases up to 50 phr EPDMd and then falls suddenly at 60 phr. The increase up to 50 phr is attributed to the additional crosslinking developed between the NR and the EPDMd phases: a higher network density on adding EPDMd raises the TS. The drop at 60 phr, however, is governed by different phenomena.

EPDMd interacts with the matrix not only through sulfur bridges formed during revulcanization but also through physical entanglements and secondary chemical forces between the olefinic chains of both NR and EPDMd. When the EPDMd fraction becomes too high, the NR/EPDMd interactions are reduced because EPDMd tends to interact preferentially with its own chains and to form aggregates (see SEM micrographs), which impairs the mechanical behavior (Colom et al., 2023). The linear NR macromolecules, free of bulky side groups, are better able to develop interactions with EPDMd and to form S-S bridges at their unsaturations (Colom et al., 2026; de Sousa et al., 2020); this is consistent with previous reports showing that EPDMd tends to agglomerate (Lee et al., 2020; Colom et al., 2024).

The presence of silica in high-EPDMd samples increases the TS because SiO₂ encapsulates the EPDMd aggregates that form. Although these silica-enveloped aggregates are not numerous enough to constitute a continuous phase, they resist fracture more effectively than the bare aggregates present in the silica-free, high-EPDMd samples. In addition, because part of the silica is consumed in surrounding the aggregates, the amount of silica dispersed in the matrix reduces the number of stress-concentration sites, which further contributes to the higher TS relative to the silica-free counterpart.

The elongation at break (EaB) remains stable up to 50 phr EPDMd. Higher EaB values (600%) are obtained in the silica-containing samples, whereas in the silica-free samples the EaB decreases (from 500% towards 400%); as with TS, the EaB collapses at 60 phr EPDMd. The development of EPDMd aggregates within the NR matrix explains these values: the EPDMd microdomains are crosslinked mainly through di- and polysulfidic bridges, forming rigid regions that are stiffer than the surrounding NR matrix. When silica is incorporated, the EPDMd aggregates become enclosed by SiO₂ and better integrated into the NR matrix, which favors deformation and preserves the elongation.

The Young’s modulus (E) increases with both EPDMd content and silica addition: the silica arranged around the EPDMd aggregates creates a rigid interphase that severely restricts chain mobility at the phase boundary and transforms the comparatively soft EPDMd particles into stiff, reinforced nodules. As for TS and EaB, the modulus decreases at 60 phr EPDMd in both series. Finally, Shore-A hardness increases progressively with EPDMd content (from 0 to 60 phr) in both series, because the residual crosslink density of EPDMd exceeds that of virgin NR; the incorporation of silica raises the hardness by up to 10 Shore-A units at every EPDMd content, through a volumetric/mechanical reinforcement that restricts chain mobility. Unlike the tensile strength, where silica can be counter-productive at low EPDMd content by interfering with NR strain induced crystallization, hardness benefits from silica across the whole composition range, so that the 60 phr EPDMd + SiO₂ compound is the stiffest and most indentation-resistant variant. The main trends are summarized in Table 3.

TABLE 3. Summary of the observed mechanical-property trends of the NR/EPDMd compounds as a function of EPDMd content and silica addition.

3.2 Microstructural analysis by FTIR-ATR

Figure 2 shows the 1880–600 cm^-1 fingerprint region of the NRXEPDMd samples with and without silica. The band at 1725–1735 cm^-1 (C=O) corresponds to the carbonyl group of free stearic acid; since the samples are not oxidized, this peak does not indicate degradation but rather the residual acid that did not react during vulcanization. The band at 1,538–1,540 cm^-1 (symmetric COO⁻) is associated with zinc stearate (ZnSt₂), the compound formed by the reaction between stearic acid and zinc oxide (the activators) (Cañavate et al., 2011; Kaczewiak et al., 2025). In addition, two other prominent absorption bands are observed in the lower-frequency region: the peak at 795 cm^-1 corresponds to the symmetric Si–O–Si stretching vibrations of silica and mineral silicates (Azizi et al., 2020; Lee et al., 2020), while the band at 720 cm^-1 is assigned to the in-phase rocking deformation of long methylene sequence chains -CH₂- characteristic of the ethylene units in the EPDM backbone (Colom et al., 2021; Colom et al., 2023). Across the series, the intensity of the 1,538 cm^-1 band is very high, indicating a massive exudation of stearate crystals towards the surface. These crystals may act as an interfacial lubricant that prevents silica from adhering properly to the NR matrix, explaining why the tensile strength is lower in these samples when SiO₂ is added at low EPDMd content. In the 1,150–950 cm^-1 region, some non-linear intensity variations arise from the superposition of asymmetric Si–O–Si stretching (1,070–1,100 cm^-1) with residual silicates and carbon black from EPDMd (approx 1,021 cm^-1). These spectral differences reflect surface filler segregation and local variations in optical contact within the shallow (approx 1–2 μm) ATR penetration depth. The broad, intense band at 1,070–1,100 cm^-1 (Si–O–Si, asymmetric siloxane stretching) reflects the presence of silica; its width can be related to the degree of agglomeration. In high-EPDMd samples the silica is located preferentially around the EPDMd aggregates and appears in the surface region, whereas the silica-free samples still show a weaker Si–O–Si band arising from the silicate contained in the EPDM itself. This surface-located, rigid “shell-type” interphase also accounts for the increase in Young’s modulus, since it compensates for the mechanical weakness of the NR matrix.

3.3 Microstructural characterization by SEM

The SEM analysis of the 50 phr EPDMd samples (Figure 3) reveals a complex morphology with cracks and microcavities indicative of limited phase compatibility. The incorporation of SiO₂ modifies this microstructure: the silica localizes preferentially around the EPDMd aggregates, creating rigid interphases that restrict polymer-chain mobility and act as a bridging layer that integrates the EPDMd aggregates into the NR matrix. This encapsulation mechanism would be what allows the macroscopic mechanical properties, tensile strength and Young’s modulus, to increase in contents up to 50 phr, masking the intrinsic heterogeneity of the blend through an effective physical and chemical reinforcement. The rough EPDMd morphology would also help to retain the massive exudation of zinc-stearate crystals detected by FTIR-ATR, preventing this lubricant from excessively weakening the interfacial adhesion until the critical limit of 60 phr is reached, where compatibility finally fails and the properties decay.

3.4 Proposed mechanism

Our proposed mechanism rationalizes the differential behavior by the competition between three phenomena that depends on EPDMd content: (i) strain-induced crystallization and self-reinforcement of the NR matrix; (ii) the formation of rigid, residually-crosslinked EPDMd aggregates; and (iii) the location and bonding state of the silica, modulated by the massive surface exudation of zinc stearate (ZnSt₂) revealed by FTIR-ATR. Figure 4 illustrates the resulting microstructure for the three composition regimes, with and without silica.

3.4.1 Low EPDMd (20 phr)

Here the compound is essentially reinforced NR, whose strength derives largely from strain-induced crystallization (Huneau, 2011). The added silica lacks an effective silane coupling agent and the interface is contaminated by the massive exudation of zinc-stearate crystals detected by FTIR-ATR (intense 1,538 cm^-1 band) (Lee et al., 2020; Kaczewiak et al., 2025). ZnSt₂ acts as an interfacial lubricant that prevents silica from adhering to the NR matrix, so the silica behaves as a weakly bonded, stress-concentrating phase and disrupts NR crystallization. Consequently, adding silica lowers the tensile strength and hardness, while the poorly bonded, slippery filler allows a larger strain at break.

3.4.2 Intermediate EPDMd (40–50 phr)

At these contents the EPDMd forms discrete, rigid aggregates (di-/polysulfidic residual network) that are stiffer than the NR matrix (Colom et al., 2023). The silica now migrates to and encapsulates these aggregates, forming a rigid shell that (a) stiffens the aggregate, turning a soft elastomeric particle into a reinforced nodule, (b) integrates the aggregate into the NR matrix and transfers stress efficiently across the interface, and (c) consumes part of the silica in the shell, reducing the number of dispersed-silica stress concentrators. The result is a simultaneous increase in tensile strength, Young’s modulus and hardness, with the strain at break preserved or even improved. The 50 phr EPDMd + SiO₂ compound is therefore the optimum (TS 11.8 MPa, E 15.4 MPa, EaB 635%) (considering optimum as “balanced” combination of tensile strength, stiffness, and flexibility). This proposed encapsulation behavior aligns with findings in similar filled rubber systems, where localized filler networks at the interphase restrict chain mobility and facilitate stress transfer between dissimilar elastomeric domains (Cañavate et al., 2000; Bakly et al., 2025).

3.4.3 Excess EPDMd (60 phr)

When the EPDMd fraction is too high, the rigid aggregates coalesce and the interfacial area to be covered exceeds the available silica; the SEM micrographs reveal microcavities and cracks. The encapsulation mechanism can no longer bridge the now quasi-percolating rigid phase, so every property falls sharply with respect to 50 phr. Silica still keeps the values above those of the silica-free compound (TS 8.26 vs. 5.86 MPa; E 9.57 vs. 3.66 MPa; hardness 72.2 vs. 65.2 Shore A), but the loss of interfacial compatibility dominates.

4 Conclusion

The incorporation of devulcanized EPDM (EPDMd) derived from roofing waste into natural rubber (NR) matrices provides a highly effective strategy for formulating sustainable elastomeric compounds with tunable mechanical properties and controlled structural behavior up to a content of 50 phr. Within this compositional range, the system maintains adequate values of tensile strength and strain at break while progressively increasing Young’s modulus and hardness. At 60 phr EPDMd, however, a critical threshold is reached: the over-saturation of recycled rubber causes severe phase separation, aggregate coalescence, and the formation of microcavities and cracks, leading to a sharp collapse of both tensile strength (falling to 5.86–8.26 MPa with and without SiO₂) and Young’s modulus (falling to 3.66–9.57 MPa with and without SiO₂) due to a loss of interfacial compatibility.

The macroscopic response and interfacial adhesion of these compounds are governed, according to our proposed mechanism, by a competition between three distinct compositional regimes and the dual role of silica as a reinforcing filler and a morphological modulator. At low EPDMd contents (20 phr), where mechanical performance is dominated by the strain-induced crystallization of the NR matrix, silica addition is counter-productive. Spectroscopic analysis by FTIR-ATR revealed a pronounced surface migration of zinc stearate (band at 1,538 cm^-1) that acts as an interfacial lubricant, hindering polymer–filler bonding, disrupting NR crystallization, and consequently lowering both tensile strength (from 8.81 to 5.77 MPa) and Shore-A hardness (from 59.5 to 53.3). Conversely, at intermediate EPDMd contents (40–50 phr), silica acts as a morphological modulator through an aggregate-encapsulation mechanism. The silica localizes around the rigid, residually crosslinked EPDMd particles to form a stiff siloxane shell (evidenced by FTIR-ATR at 1,096 cm^-1 and SEM), transforming soft elastomeric particles into reinforced nodules and bridging them into the NR matrix. This mechanism partially compensates for the structural heterogeneity of the blend, enabling the NR50EPDMdSiO₂ compound to achieve good overall performance (in terms of mechanical properties), with a maximum tensile strength of 11.83 MPa, Young’s modulus of 15.43 MPa, strain at break of 635%, and hardness of 66.8 Shore A. Beyond 50 phr EPDMd, the available silica becomes insufficient to encapsulate the quasi-percolating rigid phase, and aggregate coalescence dominates.

In conclusion, the correlation established between internal microstructure, interfacial migration phenomena, and macroscopic mechanical response confirms that a effective strategy lies at intermediate EPDMd contents combined with silica reinforcement. These findings may be useful as technical criteria for upcycling EPDM waste into sustainable elastomers suitable for structural engineering applications. Furthermore, the microstructural heterogeneity and presence of microcavities at elevated EPDMd contents may find utility in multifunctional uses, such as potential acoustic and vibration damping in civil construction.

From a sustainability perspective, valorizing devulcanized EPDM waste into NR composites offers a viable circular economy pathway to reduce landfilling and virgin polymer consumption In this line, a prospective direction for future research could involve integrating dynamic mechanical analysis (DMA), acoustic testing, and Life Cycle Assessment (LCA) to further evaluate their potential applicability in structural or dampening systems.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Miroslav Slouf, Institute of Macromolecular Chemistry CAS, Czechia

Reviewed by: Sivaprakasam Palani, Addis Ababa Science and Technology University, Ethiopia

Ali Bakly, University of Babylon, Iraq