Thermal and Volumetric Signatures of the Mullins Effect in Carbon Black Reinforced Styrene-Butadiene Rubber Composites

Abstract

This paper investigates the interplay between rubber network damage, carbon black (CB) network damage, heat exchange, and voiding mechanisms in filled Styrene-butadiene rubber (SBR) under cyclic loading. To do so, three carbon black filled SBR composites, SBR5, SBR30 and SBR60 are studied. The study aims to quantify molecular damage and its role in inducing reversible or irreversible heat flow and voiding behavior to inform the design of more resilient rubber composites with improved fatigue life and thermal management capabilities. The study effectively demonstrated how increasing carbon black content, particularly in SBR60, leads to a shift from mostly reversible to irreversible and cumulative damage mechanisms during cyclic loading, as evidenced by thermal, volumetric, and electrical resistivity changes. In particular, we identify a critical mechanical energy of 7 MJ.m⁻³ associated with such transition. These irreversible changes are strongly linked to the damage and re-arrangement of the carbon black filler network, as well as the rubber chains network and the formation/growth of voids, while reversible mechanisms are likely related to rubber chains alignment associated with entropic elasticity.

1. Introduction

Styrene-butadiene rubber (SBR), due to its high mechanical performance, is widely used in various industries, especially in automotive tires and seals [1,2]. This performance is enhanced by reinforcing fillers such as carbon black (CB), resulting in significant improvements in strength, durability, and wear resistance [3]. However, the inclusion of such fillers also introduces complex nonlinear behavior, one of the most prominent of which is the Mullins effect. The Mullins effect, first observed in the 1940s, refers to the stress-softening phenomenon exhibited by filled elastomers during cyclic loading [4]. This behavior is characterized by a reduction in stress under repeated loading and is accompanied by energy dissipation and irreversible changes in the material structure.

While this is not the core of the present paper, it is worth noting that Mullins effect has been widely modeled. The pseudo-elastic framework introduced by Ogden and Roxburgh [5] further interprets these mechanisms via a scalar damage parameter that modulates the strain energy function to reflect the history-dependent stiffness degradation, enabling tractable modeling of such complex behavior. Denora and Marano [6] proposed a constitutive model that incorporates directional softening and permanent set to better predict Mullins-like responses in CB-filled systems.

Experimentally, Mullins effect is found to depend on the nature of the material and on the loading conditions. It is found to remain unaffected by a variation of crosslink density and to increase with an increase of filler amount [7]. Through the measure of directional energy dissipation, biaxial loading revealed the anisotropic character of the Mullins effect [8]. Other experimental set-ups demonstrated that Mullins effect is accompanied by an increase of volumetric strain that attests for the appearance of nanovoids as detected by SAXS [9,10]. At slow strain rate (5%. s⁻¹), such volumetric strain results in the opening and closing of cavities, as well as a delay of novel cavities opening while applying increment of cycles.

Thanks to these modeling and experimental studies, the mechanisms underlying the Mullins effect at macromolecular and chains network scales in filled rubbers is an active area of research [11]. First, chain breakage and crosslink rupture within the polymer network have been identified as key damage sources contributing to permanent set and residual strains [12]. Second, the debonding between filler particles and the rubber matrix plays a critical role, especially in filled rubbers, where strain-induced interfacial decohesion leads to energy dissipation and reduced stiffness on reloading [13]. Third, slippage and disentanglement of chains are relevant mechanisms—each contributing to the observed stress-softening and anisotropic damage evolution under multiaxial loading [11,14]. Fourth, the rupture of filler aggregates has been scrutinized for its contribution to the mechanical hysteresis, particularly in carbon black-filled networks, where the level of percolation of the network formed by fillers can be characterized by conductivity measurements [15]. Electrical conductivity is a powerful technique used for the determination of the percolating network of carbon black fillers, especially in SBR based composites [16,17,18,19,20]. In the case of carbon black-filled SBR, the damage to the filler network during cyclic deformation leads to an irreversible alteration of the mechanical response, as the integrity of the carbon black network deteriorates, resulting in a reduced ability to bear load. Biaxial deformation of SBR filled with CB is found to be associated with a progressive destruction of the filler-filler network and a saturation of this mechanism is observed at the highest tested strain [21]. The filler-filler network damage mechanisms can be related to the Payne effect that describes how increasing strain amplitude disrupts the temporary carbon black (CB) network. This structural damage is most active at a characteristic strain above 10–20%, and better carbon black dispersion lessens the effect by preserving the filler network [22,23,24]. Above this critical strain range, the transition from low to high strain rate, is accompanied by another modification of the damage mechanisms. Indeed, high strain and strain rates imposes heat dissipation. One of the critical aspects of the Mullins effect in carbon black-filled rubbers is the irreversible heat exchange that occurs during cyclic deformation. Such irreversibility is mostly attributed to the progressive damage of filler-filler network, which contributes to the overall thermal response of the material. This heat generation is closely linked to the voiding phenomenon as previously discussed, where the formation and growth of voids within the rubber matrix further exacerbate the energy dissipation, leading to hysteresis in the mechanical behavior.

Han et al. [24] and Yang et al. [25] demonstrated that the bound rubber structure and filler dispersion significantly influence energy dissipation mechanisms during cyclic loading, especially at higher filler contents where interfacial stress transfer and cavitation dynamics are more pronounced. Zhan et al. [26] further emphasized the need to distinguish between reversible (if the rubber is heated after tensile test) and irreversible components of stress-softening, correlating these behaviors to viscoelasticity and internal damage, respectively.

In previous investigations, we used a combination of diverse experimental techniques such as digital image correlation, infrared thermography and electrical conductivity to understand molecular mechanisms associated with cyclic loading and Mullins effect in carbon black filled EPDM under various deformation conditions. The studies showed that volumetric changes mainly resulted from internal damage, while thermal expansion is limited [27]. At high strain (above 200% of deformation), damage in such materials is characterized by permanent voids and chains scission [28]. At high filler content (at and above 40 phr of CB) [29,30] and at high strain rate [31], damage mostly occur in filler network and at rubber-filler interface, while the rubber matrix is preserved. In these materials, the Mullins effect is characterized by stress softening and upturns in voiding rate and heat sources beyond previously reached maximum deformation, attributed to the (re-) activation of dissipative cavitation mechanism. Our ambition is to implement such techniques to carbon black filled SBR, that are often used separately, but their combination is expected to provide an in-depth discussion on the deformation mechanisms at macroscale and molecular scale.

Based on the approach used for EPDM materials, the present paper aims to investigate the interplay between damage at molecular scale, heat exchange, and voiding signatures in filled SBR under cyclic loading. We aim to establish the thermal and volumetric strain behavior when filled SBR is stretched at relatively high strain rate where both parameters can be measured simultaneously. We will focus on the damage at filler and rubber networks and their role in inducing irreversible heat flow and voiding behavior during cyclic tests that are used to characterize the Mullins effect. Insights gained from this study are expected to inform the design of more resilient rubber composites with improved fatigue life and thermal management capabilities.

2. Materials and Experiments

2.1. Materials

Vulcanized SBR samples were kindly provided by Michelin (Compagnie Générale des Établissements Michelin SCA, Clermont-Ferrand, France). The filler was CB N347 and was introduced into the rubber matrix at 5, 30, and 60 phr (parts of weight per hundred parts of rubber) level corresponding to a mass fraction of 4.8, 24.1 and 37.5% respectively. The type of CB has been chosen as a standard CB of the pneumatic industry and also for its well characterized properties from the scientific literature, as shown by recent works [32,33]. The density of the matrix and the filler are ρ = 0.92 g/cm³ and ρ = 1.8 g/cm³. The composition of the materials used is summarized in

Table 1. Material compositions.

Sample Code	SBR5	SBR30	SBR60
SBR	100	100	100
CB N347 (phr)	5	30	60
CB N347 (wt.%)	4.8	24.1	37.5
6PPD (phr)	1.9	1.9	1.9
Stearic acid (phr)	2	2	2
ZnO (phr)	2.5	2.5	2.5
Sulfur (phr)	1.6	1.6	1.6
CBS (phr)	1.6	1.6	1.6
Density (g.cm−3)	0.95	1.12	1.24

. Samples were prepared in the following way: all ingredients except vulcanizing agents were first mixed in a lab scale Haake internal mixer. The vulcanization agent was then incorporated in a calendaring machine following a procedure developed by Michelin. The sheet was then cut into pieces into a mold and put in an oven at 150 °C for vulcanization. The vulcanization time for each material was optimized with a rheological measurement at 150 °C. They were found to be equal to 35, 20, and 15 min for the materials with 5, 30, and 60 phr, respectively. Finally, samples were removed from the molds and air-cooled.

2.2. Tensile Tests Combining In-Situ Infra-Red Thermography and Digital Image Correlation

Cyclic uniaxial tension tests have been performed to three types of SBR named SBR05, SBR30 and SBR60. For all the samples, the gauge length is estimated to 20 mm. A bone shape has been chosen to allow homogeneous loading in the central zone of the gauge length. The experimental setup is presented in Figure 1. The testing machine is an electromechanical Instron 5967. It was initially equipped with a load sensor of 2 kN. During the mechanical tests, the local displacements (according to the three cartesian coordinates) were measured by digital image correlation (VIC-3D 10 software from Correlated Solutions, Irmo, SC, USA) using a stereovision system comprising 2 monochrome cameras (Pike F505 AVT 5 megapixel). These cameras were equipped with Schneider Kreuznach 50 mm lenses (Schneider Kreuznach, Bad Kreuznach, Germany). The temperature field on the specimen surface has been recorded with the thermal camera PI450 from Optris Company (Optris GmH & Co. KG, Berlin, Germany). The frame rate of image recording was 6 Hz. The sensor resolution is 382 × 288 pixel. For all the samples, the imposed travel speed was 5 mm.s⁻¹, and the maximum imposed displacements were 20, 40 and 60 mm, corresponding respectively to 100, 200 and 300% nominal strain. For each of these displacements, 3 cycles were applied. Since the data originate from separate devices, they have been interpolated into a common temporal base. Such process has been done into the testing machine temporal base because of its higher time discretization. The post-processing of thermal images enabled the determination of the temperature at the center of the samples along with the position of the moving clamp. The moving clamp position was determined by finding the maximum peak of the derivative thermal signal along the loading axis, as illustrated in Figure 1. The displacement was also used to take marks in the temporal base to define the starting and ending points of the interpolation domain. With the VIC 3D 10 software, the local true strains of a region of interest taken at the center of the sample have been extracted. The resolution for the strain fields is given by:

$$res=\left(filtersize-1\right)\times step\times scale$$

(1)

The subset, step, and filter size (in pixels) had been chosen in the ranges $\left[29;47\right]$, $\left[3;7\right]$, and $9$, respectively. λ₁ and λ₂ are the longitudinal and transversal deformation, respectively. It is assumed that the transversal deformation in the thickness of the specimen, λ₁, is equal to that in the specimen width (assumption of transversal isotropy). Within this assumption, the volumetric deformation is given by:

$${\displaystyle \frac{∆V}{V_0}}={\lambda }_1{{\lambda }_2}^2-1$$

(2)

2.3. Analysis of the Heat Sources

To estimate heat sources/sinks, only the temperature at the specimen surface in the central part of the tensile bars is considered, where the temperature field is spatially homogeneous (Figure 1). Hence, the heat source distribution is seen as uniform within the considered area of interest. A linearization of heat losses in the heat equation results in the following expression of the heat source [34,35,36]:

$$S=\rho C_p(\dot{\theta }+{\displaystyle \frac{\theta }{\tau }})$$

(3)

S is the heat source expressed in MW.m⁻³, $\rho$ is the bulk density (g.cm⁻³) of the tested material, and C is its heat capacity (J.g⁻¹K⁻¹). θ = T − T₀ represents the temperature variation above (positive) or below (negative) the equilibrium room temperature, T0. θ is the rate of heating (K.s⁻¹). The time constant τ (s) characteristic of the heat exchange (during heating or cooling) along the specimen thickness was calculated using the following equation:

$$\theta =\left(T_{t=0}-T_0\right)e^{-{\displaystyle \frac{t}{\tau }}}$$

(4)

with T_t=0 and T₀ being the temperatures at the beginning and at the end of the relaxation step, respectively, after the specimen has been heated (during loading) or cooled (during unloading). The thermal relaxation times is expected to decrease with the applied deformation, that is explained by the reduction of the specimen thickness when applying a longitudinal strain. Indeed, if we assume a volume conservation in our materials subjected to loading, the material’s thickness would be equal to e₀/$\sqrt{\lambda }$, with e₀ being the thickness of the material in the undeformed state, and λ the stretching ratio, equal to 1 + ε, ε being the strain. The dependence of the thermal diffusion time on the applied deformation can be evaluated assuming that this time decreases similarly with the thickness reduction upon tensile deformation:

$$\mathsf{\tau }={\tau }_0/\sqrt{\lambda }$$

(5)

The relationship between mechanical dissipation and thermal dissipation during cyclic loading is provided by the calculation of the corresponding dissipated energies. The mechanical energy (W_def) is split into elastic (W_e) and anelastic (W_an) parts. The former part is the recoverable elastic energy per unit of volume. The latter part, i.e., the total anelastic energy per unit volume, is made of intrinsic dissipation (ϕ₁) and stored energy per unit volume (W_s). In the following, the elastic component per cycle remains close to zero and it can be considered that the mechanical and anelastic energy per unit volume per cycle are approximately the same. The deformation energy is then written:

$$W_{def}={\phi }_1+W_s$$

(6)

The hysteresis loops that characterizes the deformation energy loss is essentially due to material dissipation. In the absence of microstructural transformations, the area of the surface of the hysteresis loop equals the amount of energy dissipated in the material over a complete loading cycle. The mechanical energy per cycle was estimated using the expression below:

$$W_{def}=\int {\displaystyle \frac{Fd\epsilon }{e_0{L}_0}}$$

(7)

The mean dissipation energy rate over a cycle (~ϕ1) was then calculated via the zero-dimensional approach of the local heat diffusion equation:

$${\phi }_1=\rho C_p\int (\dot{\theta }+{\displaystyle \frac{\theta }{\tau }})dt$$

(8)

2.4. Electrical Conductivity Measurements

Resistivity is measured with a four-point contact method (Kelvin method) creating an arbitrary potential difference on the specimen via a generator. The current (I) and voltage (V) are measured by an ammeter in series (Keithley 6514 Electrometer, (Keithley Instruments, Solon, Ohio, United States of America) and a voltmeter in parallel (Promax PD352, PROMAX Test & Measurement SLU, L’Hospitalet de Llobregat, Spain), respectively, providing the specimen electrical resistance. The power supply is an HEINZINGER HNC 6000-1pos (Heinzinger electronic GmbH, Peiting, Germany). The resistivity is calculated using Ohm’s law:

$$\rho = {\displaystyle \frac{VS}{l}}$$

(9)

with S as the cross section of the specimen and l as the distance between voltmeter connectors. The electrical conductivity (in Siemens/m) is then calculated as the inverse of the resistivity.

3. Results and Discussion

3.1. Mechanical, Thermal and Volumetric Strain Responses

The Mullins effect has been characterized on SBR5, SBR30 and SBR60 by performing incremental cyclic tests consisting of three levels of deformation, each level comprising 3 cycles (see Figure 2). During the tensile tests, both the temperature changes at the surface of rubber specimen and the volumetric strain were measured. The temperature changes measured by in-situ infrared thermography attest for the presence of reversible and/or irreversible dissipative mechanisms induced by the cyclic loading. A heat source treatment will be provided to quantify the elastic and inelastic contributions. In both SBR5 and SBR30, the temperature varies in a range above and below room temperature, suggesting a certain reversibility of the heat exchange. Conversely, in SBR60, the temperature rise is higher during loading phases and never decreases down to room temperature, which indicates the occurrence of cumulative and irreversible dissipative mechanisms within the time of the experiments. The volumetric strain confirms the occurrence of damage induced by deformation that could reflect various mechanisms such as voids nucleation/growth, and crack propagation. The volumetric strain is found to progressively increase with both deformation and carbon black content, expectedly due to a higher level of damage by voiding induced by the strain amplification in the rubber phase in presence of fillers, and stress concentration at the interface between the matrix and the fillers. One may note that, in the experimental conditions chosen (room temperature and a nominal strain rate of 56%.s⁻¹), the volumetric strain is found to be reversible (it does not surpass more than 1% for all cycles and all tested materials) upon unloading, suggesting a closing of cavities. Nonetheless, the shape of volumetric strain during loading and unloading, especially for the SBR60, suggests an eventual delay between opening and closing of cavities, and will be studied in more details in the next sections.

The results shown in Figure 2 are presented as a function of the strain (Figure 3). Stress strain curves show an increasing mechanical hysteresis by increasing both the strain and CB content. This is expectedly due to dissipative viscoelastic and damage mechanisms, as previously mentioned. Volumetric strain—strain curves show distinct behaviors as a function of the CB content: SBR5 and SBR30 show almost completely reversible volumetric strain variations while the SBR60 shows some hysteretic behavior. The temperature changes are complex to discuss based on such parameter. We propose an analysis of the heat source and sink whose treatment are detailed in Section 2.3.

3.2. Heat Source and Dissipation Energy Rate

The thermal response at the sample surface is complex, as the CB network is being modified during the sample deformation, but the characteristic times of heat exchange with strain and cycles had been estimated in a prior work [37]. From the knowledge of such characteristic time and from heat source analytical equations proposed in Section 2.3, heat sources are calculated. Heat source is represented as a function of the time and of the strain for the SBR5, SBR30 and SBR60 (Figure 4). Heat sources are positives during loading and heat sink negatives during unloading, suggesting that the energy dissipated during the loading is recovered (totally or partly) during the unloading. A significant difference is observed between SBR5 and SBR30 from one side and SBR60 on the other side. In the SBR60, the heat source during the first loadings is always found higher as compared to the subsequent cycles, suggesting higher level of energy dissipation during this phase. During unloading, no differences are observed in the heat sources with the cycle accumulations. This leads to positive values of mean energy dissipation rates for the first cycles of all series, for the three deformations tested (see Figure 4f). The results are consistent with those found by Le Cam et al. in highly filled SBR [38,39].

Figure 5a–c shows the evolution of dissipated mechanical energy (mechanical hysteresis) with the application of cycles. Mechanical dissipation increases with strain and appears to stabilize rapidly at cycle 2. The symmetrical heat source and heat sink in SBR5 and SBR30 is confirmed by the absence of any mean dissipation energy rate (integral of the heat source during the cycle over time), as shown in Figure 5d,e. The difference between reversible and irreversible voiding in the case of SBR5, SBR30 and SBR60 (Figure 5f) may be related to the presence or absence of a percolating network for the SBR60. In the following, an investigation of the damage mechanisms in highly filled SBR (SBR60) that comprises a percolating network of carbon black, will be conducted.

Figure 6 shows the mean dissipation energy rate as a function of the dissipated mechanical energy. A critical energy dissipation at around 7 MJ.m⁻³ is associated with the appearance of mean dissipation energy rate. Below this value, heat sources and sink are found symmetric, suggesting reversibility of the dissipative mechanisms. These latter are mostly due to entropic elasticity that are then predominant over viscoelastic mechanisms. Above such critical value, the mean energy rate seems proportional to mechanical energy dissipation. The detection of this mean dissipation energy rate suggests the preponderance of irreversible dissipative mechanisms. Such mechanisms are likely associated with increased viscoelasticity as well as possible filler-filler damage and/or rubber chains network alteration that will be discussed in the following by discussing on the change in volumetric strain hysteresis or electrical resistivity.

3.3. Role of Voiding in Dissipative Mechanisms in Highly Filled SBR650

As a preliminary remark, one must note that the measure of volumetric stain has been evaluated assuming transverse isotropy of the deformation, as the deformation field in the face along the thickness had not been measured. Such assumption may result in discrepancies of the obtained results depending on the applied cycles. This is due to the anisotropy of the cavities formed that may change with the accumulation of cycles, as shown in the literature [9]. This suggests that the absolute values of the obtained volumetric strain may not be accurate. Hence, only the qualitative evolution with longitudinal strain is discussed.

At slow strain rate (quasi-static conditions), there is no consensus on the evolution of the volumetric strain and its hysteresis (area under the volumetric strain curve versus longitudinal strain) in carbon black filled SBR. By using DIC technique, Le Cam found that filled SBR that contained 34 phr of CB showed hysteresis for the first cycles, while this hysteresis disappeared for the subsequent cycles [40]. In the present paper, none of the SBR5 and SBR30 show hysteresis (cf. Figure 3b,e), suggesting that all mechanisms of opening and closing of cavities at the macroscopic level are elastic and reversible. This is consistent with weak viscoelasticity as suggested by low mechanical dissipation (see Figure 5a,b). This does not indicate, however, that such mechanisms are reversible at the macromolecular level (i.e., they are not necessarily accompanied by re-entanglement of broken chains). Figure 7a shows the detail of the evolution of the volumetric strain during the three cycles performed at the third maximum deformation for the SBR60. Data from SBR5 and SBR30 are not shown as neither exhibits hysteresis in terms of volumetric strain nor any change in electrical conductivity. The evolution with longitudinal strain of the volumetric strain in SBR60 during the loading phase is found to accelerate just above the maximum deformation of the second series of cycles performed at 150% (see Figure 3h), suggesting that the reaching of highest levels of deformation promote the formation of higher fraction of cavities (higher number and/or volume). In an in-situ SAXS study, Creton et al. suggested a process of creation of nanovoids in confined rubber domains that have not previously cavitated, rather than to the reopening of the previously created cavities [10]. They found that such observations were consistent with the acceleration of volumetric strain [41]. In Figure 7b, the evolution of the volumetric strain hysteresis is shown with the number of cycles for the three successive applied deformations. All cycles show a hysteretic behavior that seem to decrease with their accumulation. The volumetric strain hysteresis is found to evolve regularly with dissipated mechanical energy (Figure 7c), even below the critical value detected in Figure 6. This suggests that voiding mechanisms in early stages of deformation can be associated with low dissipative mechanisms. The latter can be early stage cavitation where the rubber matrix remains mostly elastic and where entropic elasticity dominates. At higher strains, above the critical energy value, cavitation is likely dominated by large strain mechanisms such as cavitation nucleation/growth, decohesion at filler/rubber interface or damage of filler-filler and rubber networks.

To further analyze the SBR composite damage, particularly that associated with the network of carbon black, measures of electrical resistivity of the SBR60 sample were conducted. One may note that, in a previously published paper, the SBR5 and SBR30 do not show any electrical conductivity, while SBR60 is electrically conductive, suggesting the existence of a percolating CB network [37]. The evolution of the electrical resistivity (see Equation (9) in Section 2.4) is shown as a function of the cycles (Figure 8a) and as a function of the strain (Figure 8b). Consistent with a progressive increase of volumetric strain as well as its associated hysteresis, the electrical resistivity regularly increases with strain, suggesting progressive damage induced by the deformation of the CB network. In addition, no specific change of the electrical resistivity with cycles accumulation has been observed, suggesting that the damage of the CB network mostly occurs after the application of the 1st cycle. It is worth noting, however, that in-situ investigation of resistivity of CB filled SBR have shown that the percolating network possibly reform partially during unloading. This could be ascribed to the rearrangement of the CB network via the recreation of filler-filler interactions. Such mechanisms may be dominant, instead of irreversible mechanisms in the rubber matrix within the experimental time, such as slippage at filler-rubber interface [42,43]. While we do not have access to such evolution during her cycles, it is clear that most damage occur in the first cycle and that the maximum strain has deep impact on it (filler-filler damage is usually attributed to Payne effect occurring at a deformation between 0.1 and 10%).

4. Conclusions

This study provides a comprehensive thermomechanical analysis of the Mullins effect in carbon black-filled SBR composites. We highlight the interconnected roles of heat generation, volumetric strain, and filler network damage under cyclic loading. By combining in-situ digital image correlation, infrared thermography, and electrical conductivity measurements, we have shown that increasing filler content intensifies both mechanical hysteresis and irreversible dissipative phenomena. Particularly in SBR60, irreversible heat sources, hysteresis in volumetric strain, and increased resistivity after deformation cycles reveal significant damage of the percolating carbon black network. Our results suggest that beyond a critical threshold of mechanical dissipation (~7 MJ·m⁻³), a transition occurs from reversible to irreversible mechanisms. Reversible mechanisms involve entropy-driven mechanisms such as thermoelasticity, as well as early stage cavitation when the rubber matrix is still highly elastic. Irreversible mechanisms likely involve viscoelasticity, cavitation, filler-filler and rubber-filler alteration, as shown by the progressive increase of volumetric strain and electrical resistivity. These findings underscore the importance of considering both microstructural and thermal effects when evaluating fatigue behavior in highly filled rubbers. The observed energy dissipation patterns and network breakdown offer valuable insights for designing more resilient elastomeric materials with improved thermal management and structural integrity. Future work should explore the reversibility of network reformation in real-time and across broader strain rates, enabling a better understanding of damage recovery mechanisms in complex rubber composites.