Versatile NBR/EPDM reinforced Cu–Al–Zn alloy composites with inherent soft magnetic behavior

To formulate the elastomeric composites, two base rubbers were employed. Ethylene–propylene–diene rubber (EPDM, Vistalon 650S) was procured from ESSO Chemie, Germany; it contained 55% ethylene and 9% ethylidene norbornene as the diene, with a Mooney viscosity ML(1 + 8) at 127 °C of 48–52, and a density of 0.86 g cm⁻³. Acrylonitrile–butadiene rubber (NBR, Bayer AG, Germany), with a bound acrylonitrile content of 32% and a specific gravity of 1.17 ± 0.005 g cm⁻³, was purchased for use as the second elastomeric component.

In addition, zinc oxide (ZnO) and stearic acid (StAc) were acquired as activators, while N-cyclohexyl-2-benzothiazole sulfenamide (CBS) was supplied as the primary accelerator. Elemental sulfur was obtained as the vulcanizing agent, and polymerized 2,2,4-trimethyl-1,2-dihydroquinoline (TMQ) was sourced as the antioxidant. Furthermore, the metallic filler, a Cu–Al–Zn ternary alloy (Devard’s Alloy Extra Pure, LOBA CHEMIE PVT. LTD.), was selected for its multifunctional reinforcement potential. Moreover, the coupling agent, 3-(trimethoxysilyl)propyl methacrylate (TMSPM, Sigma-Aldrich, Germany), was procured for use as a silane-based compatibilizer to enhance polymer–filler interactions. All chemicals were used as received without further purification.

Composites were fabricated by incorporating Cu–Al–Zn alloy at loadings of 2.5, 5, 10, 15, and 20 phr into an NBR/EPDM blend mixed with a ratio of 1:1 (Table 1). Mixing was performed on a laboratory two-roll mill with rolls 470 mm in diameter and 300 mm in width, operating at a slow-roll speed of 24 rpm and a gear ratio of 1:1.4, in accordance with ASTM D3182-07 (2012). The morphology and particle size distribution of the Cu–Al–Zn alloy filler used in the formulations are presented in Fig. 1, while the detailed formulations are listed in Table 1. A constant amount of coupling agent, poly(3-trimethoxysilylpropyl methacrylate) (TMSPM, 4 phr), was incorporated in all filled composites, whereas the neat blend was prepared without coupling agent.

Full size image

To relieve internal stress and ensure uniform filler distribution, the rubber blends were conditioned overnight at ambient temperature after compounding. Vulcanization of the blends was carried out in a compression-molding hydraulic press at 152 ± 1 °C under a pressure of 40 kg cm⁻². The curing characteristics of the NBR/EPDM–Cu–Al–Zn composites were determined using an oscillating disk rheometer (R-100 MDR One, TA Instruments) at 152 ± 1 °C in compliance with ASTM D2084-11 (2011).

Additionally, the Cu–Al–Zn alloy was characterized before being incorporated into the blends. ATR-FTIR spectroscopy confirmed the presence of O–H stretching (~ 3433 cm⁻¹), C–H vibrations (~ 2976–2924 cm⁻¹), and metal–oxygen bands (619–1049 cm⁻¹), indicating surface hydroxylation and oxide formation, which promote interaction with the polymer matrix. The alloy composition was further validated by energy-dispersive X-ray (EDX) analysis, which confirmed the presence of aluminum, copper, and zinc.

The coupling agent, poly(3-trimethoxysilylpropyl methacrylate) (TMSPM), was synthesized through free-radical solution polymerization of 3-(trimethoxysilyl)propyl methacrylate (TMSPM) in acetone, initiated with azobisisobutyronitrile (AIBN), under a nitrogen atmosphere. The preparation procedure was followed as reported in the reference [23]. The resulting polymer was characterized by ATR-FTIR spectroscopy (Fig. 2). The spectrum, baseline-corrected in OriginPro 2025, exhibited characteristic absorption bands at 3534 and 3379 cm⁻¹ (O–H stretching), 3068 and 2925 cm⁻¹ (aliphatic C–H stretching), 1709 cm⁻¹ (ester C=O stretching), 1395 cm⁻¹ (C–H bending), and 1170 and 1057 cm⁻¹ (Si–O–C vibrations). Additional peaks at 1262, 976, and 823 cm⁻¹ further confirmed the successful polymerization of TMSPM.

Full size image

To assess particle morphology and dispersion, scanning electron microscopy (SEM, JEOL JX 840, Japan) equipped with an energy-dispersive X-ray spectroscopy (EDS) analyzer was employed. The Cu–Al–Zn alloy and the fractured surfaces of NBR/EPDM vulcanizates containing the alloy were examined. Before imaging, specimens were sputter-coated with a thin conductive layer of gold to ensure surface conductivity and high-resolution micrographs.

Furthermore, the chemical structure of the alloy and the elastomeric composites was probed by ATR-FTIR spectroscopy (JASCO FTIR-6100E, Japan), operated in absorption mode over the spectral range of 4000–400 cm⁻¹. The samples were finely ground, homogenized with spectroscopic-grade KBr, and pressed into transparent pellets. The spectra were recorded at a resolution of 4 cm⁻¹ to enable identification of characteristic functional groups.

The thermal stability of the composites was assessed using thermogravimetric analysis (TGA, Perkin Elmer, USA). Measurements were carried out in a nitrogen atmosphere over the temperature range of 50–1000 °C at a constant heating rate of 10 °C min⁻¹, allowing the identification of characteristic mass-loss steps and decomposition temperatures.

Complementary differential scanning calorimetry (DSC) measurements were performed on an SDT-Q600 (USA) to evaluate thermal transitions and enthalpic changes. Samples were heated from − 1.1 to 260 °C under nitrogen at a rate of 10 °C min⁻¹. The total enthalpy change (ΔH_total) was determined as the sum of the melting enthalpy (ΔH_fusion) and crystallization enthalpy (ΔH_{crystallization}) [27]:From these data, the apparent specific heat capacity (Cp) of each sample was estimated according to [28]:where ΔH_total is expressed in J g⁻¹ and ΔT corresponds to the transition temperature interval (K). This approach enabled the quantification of both enthalpic and specific heat changes associated with the incorporation of the alloy.

The curing behavior of NBR/EPDM composites containing varying loadings of Cu–Al–Zn alloy in the presence of the coupling agent TMSPM was evaluated using a Moving Die Rheometer (MDR One, TA Instruments, New Castle, DE, USA). Measurements were performed at 152 ± 1 °C in accordance with ASTM D2084-11. The rheometric data provided scorch time (ts2), optimum cure time (tc90), and torque development parameters, which are essential for assessing the vulcanization process.

The cure rate index (CRI), a widely accepted rheological parameter for describing curing kinetics, was calculated from the torque–time curves using the following relation [29]:where t_c90 represents the optimum cure time, time required to reach 90% of the maximum torque, and t_s2 denotes the scorch time; the time at which samples reach 2 dNm above the minimum torque.

Tensile properties of the NBR/EPDM composites were assessed in accordance with ASTM D412-06a (2013) using a Zwick Roell Z010 universal testing machine (Ulm, Germany). Stress–strain responses were recorded to extract key parameters, including tensile strength, elongation at break, and modulus, providing quantitative insight into the reinforcement efficiency of the incorporated alloy. For statistical reliability, each formulation was tested in quintuplicate, and the mean values were reported.

The magnetic behavior of the Cu–Al–Zn alloy and the prepared NBR/EPDM composites was investigated at ambient temperature using a Vibrating Sample Magnetometer (VSM, Lake Shore Model 7410, USA). Magnetic hysteresis loops were obtained to determine the principal magnetic parameters, including saturation magnetization (Ms), remanent magnetization (Mr), coercive field (Hci), and squareness ratio.

To enable quantitative comparison between the alloy and the composites, retention and relative changes were calculated using the following relations [30]:These calculations provided a direct measure of the extent to which the magnetic performance of the alloy was preserved or modified following incorporation into the elastomeric matrix.

The formulations of the NBR/EPDM blends (Table 1) were prepared with increasing concentrations of Cu–Al–Zn alloy (2.5–35 phr) in the presence of a constant coupling agent. At the same time, the base elastomer ratio and curatives were kept unchanged. This systematic variation enabled direct evaluation of the alloy’s effect on curing dynamics, crosslink density, and viscoelastic balance (Figs. 3 and 4, Table 1).

Full size image

Full size image

The rheological characteristics of the NBR/EPDM composites provide key insight into how the Cu–Al–Zn alloy influences curing dynamics, crosslink density, and viscoelastic balance (Fig. 3, Table 1). The torque–time profiles (Fig. 3A) revealed a distinct non-linear influence of the alloy on crosslink formation. The neat blend (NE0) displayed the lowest maximum torque (MH = 12.37 dNm) and torque difference (ΔM = 11.28 dNm), confirming its limited baseline crosslink density. At moderate filler levels, NE2 and NE3 achieved higher M_H and ΔM values, 12.87 and 13.85 dNm, respectively, that demonstrated more efficient filler–rubber interactions and improved network reinforcement. Nevertheless, higher alloy loadings (NE4–NE8) led to comparable or diminished torque, indicating that excessive filler disrupted optimal crosslinking and reduced network integrity [16, 40, 41].

In contrast, NE2 and NE3 showed rapid tan δ decay and lower equilibrium values, confirming enhanced elasticity and restricted chain mobility (Fig. 3B). Other filled samples (NE5–NE8) exhibited intermediate behavior, consistent with their reduced torque and less favorable modulus evolution.

With respect to curing kinetics, the induction period or scorch time (t_s2) and optimum cure time (t_C90) increased upon alloy addition. For example, NE0 and NE2 displayed t_C90 values of 20.01 min and 23.31 min, respectively. It was assumed from the literature that the filler delayed vulcanization was due to competitive interactions with accelerators [42]. Nevertheless, NE2 and NE3 maintained acceptable cure rate indices (CRI ≈ 5 min⁻¹), confirming efficient curing under moderate loading. In contrast, higher filler contents (NE5–NE8) extended cure times without yielding improvements in ΔM or tan δ, suggesting that saturation and agglomeration hindered further reinforcement [43].

In parallel, the time-dependent evolution of dynamic moduli (Fig. 4) corroborated these trends, revealing that alloy incorporation accelerated gelation and shifted the balance toward elastic reinforcement. The storage modulus (G′) of NE3 rose more steeply and attained higher equilibrium values than NE0, reflecting superior elastic reinforcement and increased crosslink density. Conversely, the loss modulus (G″) of NE3 declined relative to NE0. This reduction signified diminished viscous dissipation due to restricted chain mobility and enhanced stress transfer through the polymer–filler interface, promoted by effective alloy dispersion and the coupling agent [16, 44, 45]. Importantly, the crossover points of G′ and G″ occurred earlier in NE3 than in NE0, marking an accelerated gelation process and faster establishment of an elastic-dominated network. This behavior highlights that well-dispersed filler domains reinforce the matrix without sacrificing processability [46].

By contrast, NE8 exhibited a markedly different response. The storage modulus (G′) increased only modestly and remained below NE0 at longer curing times, indicating that excessive filler loading impeded efficient network development. The loss modulus (G″) of NE8 displayed a pronounced initial rise and reached its maximum later than NE0, signifying delayed relaxation dynamics and elevated viscous contributions [47]. Moreover, the slower decline of G″ demonstrated persistent energy dissipation and a lack of elastic recovery [48]. Hence, the viscoelastic profile of NE8 confirmed that high alloy concentrations promoted filler–filler interactions and agglomeration, which disrupted uniform stress transfer and reduced crosslink homogeneity. Consequently, excessive loading reintroduced viscous losses and diminished the reinforcement advantages observed at moderate alloy incorporation.

Differential scanning calorimetry (DSC) was employed to assess the thermal transitions and enthalpic behavior of NBR/EPDM–Cu–Al–Zn composites (NE0, NE2, NE3, NE5, NE8). The neat blend (NE0), prepared without a coupling agent, displayed the highest total enthalpy change (ΔH_total) and a positive specific heat capacity (Cp) with values of − 24.9 J g⁻¹ and 0.392 J g⁻¹ K⁻¹, respectively (Table 2). In turn, this confirmed the presence of inherently ordered crystalline domains and relatively mobile polymer chains [49, 50]. With alloy incorporation, total enthalpy change (ΔH_fusion) and ΔH_total decreased in NE2 and more markedly in NE5 and NE8, where strongly negative Cp values (− 0.129 and − 0.347 J g⁻¹ K⁻¹) indicated exothermic crystallization dominating over endothermic melting (Table 1). Such a negative Cp reflected reduced thermal reversibility, as filler-induced ordering and restricted chain mobility promoted rigid, less deformable packing.

In contrast, NE3 exhibited the highest ΔH_fusion (142.7 J g⁻¹) and a positive ΔH_total (+ 44.4 J g⁻¹) with a small positive Cp (0.165 J g⁻¹ K⁻¹), signifying partial recrystallization and improved structural cohesion at moderate loading. Through its silane functionality, the coupling agent bonded the alloy surface to the elastomeric matrix, ensuring filler stabilization and chain mobility. Consequently, premature rigidification was prevented. Moreover, the transition temperatures (ΔT = 447–464 K) varied only slightly among the samples, indicating that the main thermal events occurred within a narrow range. However, the enthalpic differences emphasize the joint role of filler concentration and coupling chemistry in dictating chain dynamics [45].

In parallel, thermogravimetric analysis (TGA) reinforced these observations (Fig. 5A). From the results, all samples decomposed in a single major step, with onset temperatures ranging between 447 and 464 °C. The neat matrix (NE0) degraded earlier than alloy-filled composites, confirming the barrier effect of the metallic particles. Additionally, it was assumed that the coupling agent strengthened filler–matrix adhesion, which delayed heat transfer and restricted volatile release.

Full size image

Correspondingly, kinetic evaluation through the Coats–Redfern model revealed a rise in activation energy (Ea), from ~ 71.4 kJ mol⁻¹ in NE0 to ~ 105.5–151 kJ mol⁻¹ in NE2 and NE3 (Fig. 5B). Thus, moderate filler levels stabilized the network by raising the energy barrier to chain scission. The coupling agent further amplified this stabilization by reducing interfacial defects, thereby hindering premature degradation. However, NE5 and NE8, despite showing relatively high Ea values, exhibited strongly negative ΔH_total, reflecting competing recrystallization–degradation processes that undermined structural reversibility.

In summary, DSC and TGA confirmed that Cu–Al–Zn alloy has exerted a non-linear effect on thermal behavior. Moderate incorporation (NE2–NE3) coupled with surface functionalization has been achieved by exploiting a silane agent, enhanced molecular packing, and kinetic stability. This, in turn, produced the most balanced performance. By contrast, excessive loadings (NE5–NE8) restricted chain mobility and promoted filler-driven crystallization, which compromised thermal reversibility despite delayed degradation.

The tensile behavior of the NBR/EPDM–Cu–Al–Zn composites is shown in Fig. 6A. The neat blend (NE0) exhibited the lowest tensile strength (1.9 MPa) and elongation at break (≈ 250%), reflecting a relatively loose network’s crosslinking. With the incorporation of alloy filler and TMSPM coupling agent, both tensile strength and elongation at break increased significantly, peaking at NE2 and NE3 with values of 5.9 MPa, 850% and 5.5 MPa, 800%, respectively. This synergistic improvement highlights the role of the coupling agent in anchoring alloy particles to the polymer backbone, which promoted homogeneous stress transfer and delayed crack propagation [51]. Beyond this optimum range, further increases in filler content (NE5–NE8) led to a decline in tensile properties, attributed to filler agglomeration and associated stress concentration sites that disrupted chain extensibility [52].

Full size image

The swelling analysis provided complementary evidence of these network modifications (Fig. 6B). Equilibrium swelling (Q_m) decreased substantially for NE2 and NE3 compared to the neat matrix, confirming enhanced solvent resistance and denser crosslinking. Simultaneously, the soluble fraction (SF) increased moderately with filler addition, reaching its lowest values at NE2–NE3 and rising sharply in NE6–NE8. This trend suggested that moderate filler concentrations reduced the proportion of unbound or extractable chains, whereas excessive filler disrupted homogeneity and promoted incomplete crosslinking.

The fatigue performance (Fig. 7A) revealed a general improvement in endurance with alloy incorporation compared to the neat NE0. Samples NE1, NE3, and NE4 exhibited the highest mean fatigue lives, reflecting enhanced filler–matrix interaction and improved stress transfer efficiency. Despite NE1 and NE4 showing the single highest cycle value, their large deviation indicated less uniform reinforcement. In contrast, NE3 demonstrated both high fatigue life and low data scatter, signifying superior structural integrity and uniform dispersion of the Cu–Al–Zn alloy phase within the elastomer network.

Full size image

In the same vein, the hardness results (Fig. 7B) followed a similar enhancement trend with increasing alloy content. The Shore A hardness rose from 47.7 for NE0 to a maximum of 61.3 for NE3, reflecting increased rigidity due to restricted segmental mobility and higher crosslink density. A slight decline at higher filler loadings (NE4–NE8) suggested that excessive alloy may act as a stress concentrator, partially offsetting the reinforcing effect. Overall, the combination of increased fatigue endurance and hardness confirmed the synergistic reinforcement imparted by the metallic alloy within the elastomeric matrix.

In parallel, the Flory–Rehner analysis further clarified these findings (Fig. 8). For NE2–NE3, the molecular weight between crosslinks (Mc) reached minimum values (≈ 1.3–1.5 × 10⁷ g mol⁻¹), while the corresponding crosslink density (νe) attained its maximum. These results are consistent with the observed improvement in tensile strength and elongation, indicating that optimum filler dispersion produced a compact, elastically reinforced network. Conversely, higher loadings (NE5–NE8) exhibited increasing Mc and decreasing νe, in line with the higher Q_m and SF, confirming the detrimental effect of filler agglomeration on network continuity.

Full size image

The magnetic response of the obtained composites filled with Cu–Al–Zn alloy was comprehensively examined using a Vibrating Sample Magnetometer (VSM) (Fig. 9). All composites exhibited narrow S-shaped hysteresis loops with markedly low coercivity (Hci), characteristic of soft magnetic materials and efficient domain reversibility (Fig. 9A). The pristine Cu–Al–Zn alloy displayed distinct magnetic parameters (Ms = 3.41 × 10⁻² emu g⁻¹, Mr = 2.22 × 10⁻³ emu g⁻¹, Hci ≈ 53.9 G) (Fig. 9B, Table 3). However, upon incorporation into the elastomer matrix, these values decreased substantially, with the composites retaining only a fraction of the alloy’s magnetization. For instance, NE3 preserved approximately 25% of Ms and 8% of Mr. This indicated that the alloy effectively induced magnetic responsiveness; however, the viscoelastic nature of the rubber constrained domain alignment and thereby limited long-range magnetic interactions (Fig. 10A). In context, the curve fitting NE3 to the Langevin function confirmed the efficiency of its magnetic domain response. Through moderate alloy incorporation, dispersion was promoted while agglomeration was minimized. To put it differently, it is currently clarified that several intrinsic mechanisms contribute to the significant drop in Ms and Mr upon loading the alloy within the elastomer matrix. Firstly, the diamagnetic character of NBR/EPDM introduces a matrix dilution effect that reduces the effective magnetic phase per gram of composite [53]. Secondly, polymer encapsulation around the metallic particles restricts domain wall mobility and suppresses coherent domain rotation, thereby decreasing both saturation and remanent magnetization [54]. Finally, interfacial interactions influence domain coupling. In this regard, the well-dispersed particles (as in NE3) exhibit limited exchange interactions, whereas agglomerates at higher loadings weaken uniform magnetization pathways and increase magnetic pinning. These combined effects explain the reduced magnetization values relative to the behavior of the free alloy.

Full size image

Full size image

Moreover, Arrott plots (Fig. 10B) reinforced the composites’ paramagnetic character and provided further insight into the magnetic ordering nature of the composites. For NE3, the linear region of the plot intercepted the M² axis positively, signifying the presence of ferromagnetic-like ordering and spontaneous magnetization arising from cooperative magnetic domains. The slope was relatively steep, confirming soft magnetic characteristics dominated by reversible domain movement. In contrast, NE8 exhibited a flatter slope and shifted intercept toward the origin, characteristic of weakened ferromagnetic coupling and increased paramagnetic contributions. This indicated that excessive alloy content suppressed cooperative exchange interactions due to particle agglomeration and polymer shielding effects. In this context, the hysteresis loops with finite Mr and Hci confirmed that the composites did not exhibit classical paramagnetism. Rather, they displayed a diluted soft ferromagnetic behavior, in which the intrinsic soft ferromagnetic properties of the Cu–Al–Zn alloy were preserved but significantly attenuated by matrix dilution, polymer shielding, and constraints on domain rotation. Therein, the Arrott plots indicated suppressed long-range magnetic ordering rather than true paramagnetism. NE3 maintained weak ferromagnetic-like ordering due to improved particle dispersion and partial exchange coupling. However, NE8 demonstrated dominant dipolar interactions and reduced magnetic coherence arising from particle agglomeration and enhanced polymer encapsulation. This interpretation reconciled the Arrott plot behavior with the observed soft hysteresis loops.

In parallel, a quantitative comparison with the free alloy has been displayed. It highlighted the novelty of this approach (Table 3). The incremental alloy incorporation resulted in progressive increases in saturation and remanent magnetization (Ms and Mr), thereby confirming the contribution of the metallic filler to the overall composite response (Table 3). Moreover, remanence and coercivity followed similar trends. Notably, NE3 exhibited stronger Mr and moderate Hci, indicating efficient reversible magnetization and effective particle dispersion (Scheme 1). Conversely, higher loadings (NE5–NE8) induced pronounced agglomeration, which in turn elevated Hci while depressing Mr, impeding uniform magnetization pathways. Accordingly, squareness ratios were negligible for NE0 and NE2 yet high at higher loadings (NE5–NE8). Thus, such behavior suggested persistent residual alignment despite diminished overall magnetization (Scheme 1).

Full size image

In line with these findings, the first derivative of magnetization curves (dm/dh) provided a sensitive measure of magnetic reversibility and domain interaction (Fig. 11). The peak shape and width of the dm/dh signal directly reflect the coercivity distribution and magnetic domain uniformity within the composite. The NE3 composite exhibited a sharp and symmetric derivative peak centered around zero field, reflecting a narrow coercivity distribution and facile domain wall motion (Fig. 11A). This behavior evidenced the existence of an effective magnetic coupling between well-dispersed alloy particles, consistent with the soft magnetic nature observed in its hysteresis profile. In contrast, NE8 presented a broader, less intense peak, despite the systematic and symmetric nature of the plot that had been observed. As a result, this behavior reflected uniform domain alignment, yet reduced differential susceptibility, sluggish domain reversal, and magnetic pinning features (Fig. 11B). Additionally, the systematic and symmetric nature of the NE8 response suggested that at higher alloy concentrations, magnetic coupling among adjacent domains became more pronounced, leading to cooperative magnetization and greater magnetic ordering. Furthermore, the reduced derivative amplitude, ~ 2.5 × 10^–7 emu g⁻¹ Oe⁻¹, and peak broadening confirmed that high filler loading promotes dipolar interactions and domain entrapment. Therefore, this resulted in a compromised magnetic reversibility and elevated energy dissipation. Noteworthy, the increased homogeneity came at the expense of magnetic softness, as the restricted domain wall mobility yielded slightly higher coercivity.

Full size image

At room temperature, the ε′ and ε″ values of the composites NE0, NE3, and NE8 remained suppressed at lower frequencies due to electrode polarization, yet rose abruptly when the frequency exceeded the frequency threshold value (3 × 10⁶ Hz). This rise indicated the onset of interfacial polarization and enhanced dipole alignment with the applied field. With the inclusion of the Cu–Al–Zn alloy, both ε′ and ε″ were elevated in the order NE8 > NE3 > NE0, confirming that metallic fillers strengthened Maxwell–Wagner–Sillars polarization and expanded the interfacial dipole regions within the elastomeric matrix. Likewise, σ′ climbed consistently across the frequency range, with NE8 attaining the highest values owing to its larger filler fraction and enhanced polarization fields. At this stage, the room-temperature results established the baseline behavior of the composites and demonstrated that filler concentration governed dielectric strength and charge transport efficiency (Fig. 12).

Full size image

At elevated temperatures, the same trends persisted but became more pronounced. As the temperature reached 40, 65, and 90 °C, ε′ for all samples rose. As a result, higher thermal energy enabled segmental mobility and reduced chain rigidity, allowing the dipoles to follow the alternating field more effectively. Correspondingly, ε″ was amplified with temperature, reflecting intensified energy dissipation as polarization processes accelerated. For NE0, the changes remained modest because conduction and polarization originated solely from the polymer matrix. In contrast, NE3 displayed a more substantial elevation in both ε′ and ε″, as thermally assisted interfacial polarization strengthened at the polymer–metal interfaces. For NE8, the effect became particularly robust, and the high dielectric permittivity reflected pronounced dipole buildup within densely packed filler domains (Fig. 13).

In parallel, σ′ rose steadily with temperature for all compositions. At higher temperatures, charge carriers more readily overcame potential barriers, resulting in enhanced hopping or tunneling transport. While the order NE8 > NE3 > NE0 remained consistent, the temperature-dependent data revealed additional detail. NE3 displayed the most pronounced high-frequency enhancement at 90 °C, suggesting that moderate filler loading created more efficient conduction pathways under thermal activation. Despite exhibiting the highest ε′ and ε″ values, NE8 experienced a slight decrease in σ′ at the highest temperature, as particle agglomeration hindered charge mobility at high frequencies. As the temperature approached 90 °C, all samples demonstrated conductivity values characteristic of a transition toward semiconductor-like behavior, particularly at high frequencies where thermally activated charge transport dominated (Fig. 13).

Full size image

Overall, the integration of room-temperature and elevated temperature data revealed a coherent pattern in which both filler concentration and thermal activation dictated the dielectric and conductive performance of the composites. At room temperature, interfacial polarization and filler dispersion shaped the response, while at elevated temperatures, enhanced dipole motion and thermally activated transport amplified these effects. From the integrated results, it was clear that moderate alloy loading (NE3) yielded the most favorable balance of dielectric and conductive properties over the examined temperature range. However, the high loading (NE8) intensified polarization but introduced agglomeration effects that moderate conductivity at high frequencies. This integrated result confirmed that the Cu–Al–Zn alloy imparted a temperature-sensitive dielectric and electrical response that could be tailored through appropriate adjustment of filler content.

The SEM micrographs (Fig. 13) provide morphological evidence for the influence of the Cu–Al–Zn alloy filler on the NBR/EPDM blend. The neat matrix (N0, Fig. 14A) displayed a relatively smooth fracture surface with limited microstructural heterogeneity, typical of an unfilled elastomer blend. Upon the addition of 2.5 phr alloy (N2, panel B), localized bright regions corresponding to metal particles are clearly observed, with improved dispersion and intimate interaction with the surrounding polymer matrix.

Full size image

This correlates with the rheological findings where N2 showed increased torque (ΔM) and reduced tan δ, as well as enhanced tensile strength and elongation at break, confirming the reinforcing contribution of the filler at moderate loading. A further increase to 5 phr (N3, panel C) reveals a more uniform distribution of alloy domains, with fewer voids at the polymer–filler interface, suggesting improved adhesion. The morphology of N3 explains its superior mechanical properties, especially the gains in tensile strength and toughness, which align with its higher crosslink density observed in the rheological analysis. However, at excessive loading (N8, panel D), large filler agglomerates are evident, accompanied by microvoids and superficial fractures, which act as stress concentrators. These features explain the deterioration in torque, higher tan δ, and reduced elongation at break observed in the mechanical tests. Therefore, the SEM analysis substantiates the conclusion that while moderate filler levels enhance dispersion and matrix–filler interaction, excessive alloy addition induces agglomeration that undermines the reinforcing efficiency and compromises the overall performance.