GTR/NBR/Silica Composites Performance Properties as a Function of Curing System: Sulfur versus Peroxides

In this work, conventional sulfur and two types of organic peroxides (dicumyl peroxide (DCP) and di-(2-tert-butyl-peroxyisopropyl)-benzene (BIB)) curing systems were used to investigate the possibility for tailoring of the performance properties of GTR/NBR blends reinforced with a variable content of highly dispersive silica (0–30 phr). The curing characteristics, static mechanical and acoustical properties, swelling behavior, thermal stability, and microstructure of the prepared composites were investigated. The results show that regardless of the curing system used, increasing the content of highly dispersive silica resulted in the improvement of the mechanical properties of the studied materials. It was observed that sulfur-based systems are the best choice in terms of cross-linking efficiency determined based on torque increment and cross-link density parameters. However, further analysis of the physico-mechanical properties indicated that the cross-linking efficiency does not match the performance of specimens, and the materials obtained using organic peroxides show higher tensile properties. This is due to the improved physical interactions between the GTR/NBR matrix and highly dispersive silica when using peroxide systems. It was confirmed using the analysis of the Wolff activity coefficient, indicating the enhanced synergy.


Introduction
Recycling rubber products is time-consuming, energy-intensive, and costly, since the material is obtained by irreversible vulcanization. The secondary compounds obtained from recycling exhibit worse properties than the originals and the raw materials, which makes them comparatively uncompetitive. The combination of the difficulties in recycling plus the properties of the obtained materials is in the origin of the economic and mainly environmental issues.
More than 80% of rubber waste is constituted by worn tires [1]. The process of the degradation of tires can take more than 100 years. The high number of worn tires and their slow degradability are features inherent to this industrial waste. It is estimated that about one billion used tires are generated each year. The accumulation of tires causes fire hazards [2] and their combustion emits toxic products that pollute rivers, lakes, and groundwater, bringing unpleasant and dangerous effects to humans and animals. In addition, the carbon dioxide released during the fire influences the greenhouse effect [3].
The most efficient approach to avoid these issues would be transforming used rubber goods into products presenting interesting properties, ideally, equivalent to those obtained with virgin rubber. To achieve this goal, the industry and academia have proposed the devulcanization of waste rubber, consisting of eliminating the cross-links created during vulcanization. With the waste rubber devulcanization, the structure of the network is broken, which enhances the processing of cross-linked rubber. The raw material, obtained

Materials
GTR obtained by ambient grinding of used tires (a combination of passenger car and truck tires in 50:50 mass ratio) with particles size below 0. Highly dispersive amorphous silica (Zeosil 1165 MP) with a molecular weight of 60.2 g/mol, melting point at above 1700 °C, and density 2.1 g/cm 3 was supplied by Solvay Poland (Włocławek, Poland).
Sulfur curing system consists of the following in phr: ZnO; 3.0, stearic acid; 2, TMTD; 1.0, and sulfur; 2.0. All ingredients in the sulfur curing system were provided by Standard Sp. z o.o. (Lublin, Poland). Organic peroxides were supplied by Pergan GmbH (Bocholt, Germany). The chemical structure and characteristics of used peroxide curing system are presented in Table 1.

Sample Preparation
GTR/NBR/SiO2 compounds were mixed with three curing systems (sulfur-based, DCP, and BIB) using a two-roll mill from Buzuluk (Komarov, Czech Republic). The formulation, displayed in Table 2, was as follows in parts per hundred of rubber (phr): 70.0 GTR; 30.0 NBR; curing system (sulfur-based, DCP, BIB) and variable content of SiO2 (0, 5, 15, 30 phr). The compounds were vulcanized into 2-millimeter sheets and cured at 160 (sulfur curing system) and 180 °C (peroxide curing system) using an electrically heated press PH-90 (Nysa, Poland) under the pressure of 4.9 MPa according to the optimum cure time determined as stated in ISO 3417 standard. In order to facilitate the identification of specific samples, special coding was created according to GTR/NBRA/SiX, where A stands for curing system, while X stands for the amount of silica. 8

Materials
GTR obtained by ambient grinding of used tires (a combination of passenger car and truck tires in 50:50 mass ratio) with particles size below 0. 8  Highly dispersive amorphous silica (Zeosil 1165 MP) with a molecular weight of 60.2 g/mol, melting point at above 1700 °C, and density 2.1 g/cm 3 was supplied by Solvay Poland (Włocławek, Poland).
Sulfur curing system consists of the following in phr: ZnO; 3.0, stearic acid; 2, TMTD; 1.0, and sulfur; 2.0. All ingredients in the sulfur curing system were provided by Standard Sp. z o.o. (Lublin, Poland). Organic peroxides were supplied by Pergan GmbH (Bocholt, Germany). The chemical structure and characteristics of used peroxide curing system are presented in Table 1.

Sample Preparation
GTR/NBR/SiO2 compounds were mixed with three curing systems (sulfur-based, DCP, and BIB) using a two-roll mill from Buzuluk (Komarov, Czech Republic). The formulation, displayed in Table 2, was as follows in parts per hundred of rubber (phr): 70.0 GTR; 30.0 NBR; curing system (sulfur-based, DCP, BIB) and variable content of SiO2 (0, 5, 15, 30 phr). The compounds were vulcanized into 2-millimeter sheets and cured at 160 (sulfur curing system) and 180 °C (peroxide curing system) using an electrically heated press PH-90 (Nysa, Poland) under the pressure of 4.9 MPa according to the optimum cure time determined as stated in ISO 3417 standard. In order to facilitate the identification of specific samples, special coding was created according to GTR/NBRA/SiX, where A stands for curing system, while X stands for the amount of silica.

Measurements
The vulcanization process of the samples was investigated via Monsanto R100S rheometer (St. Louis, MO, USA) with the oscillating rotor, following ISO 3417. To determinate the cross-linking rate, the cure rate index (CRI) was calculated according to Formula (1) as follows [26]: where t 90 : optimum vulcanization time, min; t 2 : scorch time, min. The determination of the R 300 parameter allowed us to investigate the thermal aging resistance of the prepared samples during curing at elevated temperature. R 300 is obtained from the time at which torque reaches the maximum value (M max. ) and it describes the percentage of reversion degree after a period of 300 s [27]. It was calculated according to Equation (2) as follows: where M max. : maximum torque, dNm; M 300s : torque 300 s after maximum torque, dNm. The tensile strength and elongation at break were measured in accordance with ISO 37. Tensile tests were carried out on a Zwick Z020 machine (Ulm, Germany) at a constant speed of 500 mm/min. Direct extension measurements were conducted using an extensometer with sensor arms. The reported results are an average of five measurements for each sample. Shore hardness type A was assessed using a Zwick 3130 durometer (Ulm, Germany) according to ISO 7619-1.
The density was determined based on the Archimedes method, as explained in ISO 1183. Measurements were carried out at room temperature in a methanol medium, without exception.
The swelling degree of the vulcanized samples (0.2 g) was estimated via a swelling test, carried out in toluene at room temperature. The swelling degree was calculated according to Equation (3) as follows: where Q: swelling degree, %; m t : a mass of the sample swollen after time t, g; and m o : an initial mass of the sample, g. Sol fraction was calculated in accordance with Formula (4) as follows: where W 1 : mass of the vulcanized sample before swelling, g; and W 2 : mass of the vulcanized sample after extraction, g. According to the following Flory-Rehner Equation (5) [28], cross-link density can be determined by equilibrium swelling in toluene: where ν e : cross-link density, mol/cm 3 ; V r : gel volume in the swollen sample; V 1 : solvent molar volume (toluene = 106.2, cm 3 /mol [12]); and χ: polymer-solvent interaction parameter (in the calculations, it was assumed to be 0.472 [12]).
In the case of rubber compounds modified with GTR, which includes an active filler such as carbon black, and for filled compounds, the Krause correction can be applied [29]. Actual cross-link density can be calculated with the Krause correction according to Equations (6) and (7) as follows: where ν e : the measured chemical cross-link density, mol/cm 3 ; ν after correction : the actual chemical cross-link density, mol/cm 3 ; K: constant characteristic of the filler but independent of the solvent; φ f : the volume fraction of filler in the sample that is calculated; ρ r : the density of studied compound, g/cm 3 ; m 0 : the weight of the sample before extraction, g; ρ f : the density of filler, g/cm 3 ; and m dry : the weight of the sample after extraction, g. For this research, the density of carbon black was selected to be 1.85 g/cm 3 and the K constant was selected to be 1.17 [30].
In order to understand the interaction between polymer matrix and silica, the reinforcing activity of the filler was assessed using Wolff activity coefficient: α F [31]. Correlation between ∆M SPEC and α F was calculated according to Equations (8) and (9) as follows: where ∆M x : the torque increment of the vulcanizate containing x phr of filler during vulcanization, dNm; ∆M 0 : the torque increment of an unfilled vulcanizate, dNm; m x : the weight of added filler, g; and m p: the weight of the polymer in vulcanizate, g. TGA studies were carried out using Mettler Toledo TGA/SDTA 851. Initially, measurements were made under nitrogen at a flow rate of 60 mL/min of gas at a heating rate of 20 • C/min. At 550 • C, airflow was introduced, and the heating rate was reduced to 10 • C/min. ATR-FTIR spectra were acquired using an Avatar 320 spectrometer from Nicolet equipped with a CsI optical collimated and horizontal attenuated total reflectance (HATR) accessory. The internal reflecting element (IRE) was a ZnSe crystal set at an incidence angle of 45. ATR spectra were collected in the range of 4000-650 cm −1 by averaging 40 scans at 4 cm −1 of resolution. To avoid local differences in the composition of the samples produced by variations in a dispersion of the components, samples for FTIR were extracted from selected representative regions of the test specimen.
The acoustic properties had been measured using a two-microphone impedance tube Bruël and Kjaer type 4206 in the frequency range 100-6500 Hz, according to the specification ASTM E 1050, which describes the standard test method for impedance and absorption of acoustical materials.
Scanning electron microscopy (SEM) was used to qualitatively examine the fracture surface of the samples broken by the mechanical tests to study the compatibility with silica at the GTR/NBR interface. Several images of the samples were taken in a JEOL Materials 2021, 14, 5345 6 of 21 5610 microscope. Prior to the observations, the samples were covered with a fine layer of gold-palladium in order to increase their conductivity.

Curing Characteristics
The effect of different curing systems on GTR/NBR/SiO 2 compounds' curing characteristics is presented in Figure 1 and summarized in Table 3. Two curing temperatures were used when analyzing the results, 160 • C for sulfur and 180 • C for peroxides. This is due to the nature of the systems used.
The acoustic properties had been measured using a two-microphone impedance tube Bruël and Kjaer type 4206 in the frequency range 100-6500 Hz, according to the specification ASTM E 1050, which describes the standard test method for impedance and absorption of acoustical materials.
Scanning electron microscopy (SEM) was used to qualitatively examine the fracture surface of the samples broken by the mechanical tests to study the compatibility with silica at the GTR/NBR interface. Several images of the samples were taken in a JEOL 5610 microscope. Prior to the observations, the samples were covered with a fine layer of goldpalladium in order to increase their conductivity.

Curing Characteristics
The effect of different curing systems on GTR/NBR/SiO2 compounds' curing characteristics is presented in Figure 1 and summarized in Table 3. Two curing temperatures were used when analyzing the results, 160 °C for sulfur and 180 °C for peroxides. This is due to the nature of the systems used.   Analyzing the data, it is obvious that the minimum torque, which is a numerical value that determines the processability of the material, for the sulfur cross-linked samples increases with the increasing amount of filler (10.5, 14.9, 24.0, and 45.6 dNm for 0, 5,  15, and 30 phr loadings of highly dispersive silica, respectively), which is in accordance with many independent studies about silica-filled elastomers [32][33][34]. The same trend is observed analyzing the maximal torque values (49.8, 52.0, 59.1, and 75.3 dNm for 0, 5, 15, and 30 phr loading of silica, respectively), indicating a drastic increase in the stiffness of the tested materials. The opposite phenomenon was noted for torque increment values (39.3, 37.1, 35.1, and 29.7 dNm for 0, 5, 15, and 30 phr loading of silica, respectively). This parameter is correlated to the cross-link density of the samples. The same volume of sample is always analyzed during the oscillating disc rheometer test; therefore, as the amount of filler increases, the volume of NBR and GTR that can cross-link with the sulfur system decreases. The scorch time (1.2, 1.2, 1.0, and 1.1 min for 0, 5, 15, and 30 phr loading of silica, respectively) and the optimum cure time (2.9, 2.7, 2.4, and 2.7 min for 0, 5, 15, and 30 phr loading of silica, respectively) do not change significantly for all the samples in the series. This means that the cross-linking system used is equally effective for all the samples, and the presence of silica only affects the values related to the stiffness and processability of the material. Parameter R 300 increases with an increasing silica content (0.3, 0.6, 1.2, and 1.8% for 0, 5, 15, and 30 phr loading of silica, respectively). A similar observation was described by Kazemi et al., who investigated natural rubber hybrid composites filled with maple/silica/carbon black [35]. The authors indicated that a high silica content (20 phr) in the studied systems resulted in an agglomeration formulation, which was related to hydrogen bonding between the silanol groups present on the silica surface.
For peroxide curing systems, the M min. and M max. values' trend is the same as for the sulfur-cross-linked samples. Regarding the comparison of the values between the types of peroxides used, it is that the M max. value for the samples cross-linked with BIB peroxide is higher (38.1, 39.5, 54.6, and 75.9 dNm-DCP; 45.0, 49.3, 60.5, and 82.8 dNm-BIB). This is related to the structure of BIB peroxide [36,37], which generates a high number of free radicals compared to DCP and, as a consequence, results in a higher cross-link density and stiffness of the tested material.
It is confirmed by ∆M values, which are higher for BIB peroxide than for DCP (23.2, 24.4, 27.3, and 20.3 dNm-DCP; 31.2, 33.0, 35.0, and 28.5 dNm). It is interesting that, with 30 phr of silica, the value drops for 25.7% (DCP) and 18.6% (BIB) compared to 15 phr of the filler. A similar tendency was observed for the samples with the sulfur-based system. This phenomenon can be related to the agglomeration of silica filler in the GTR/NBR matrix when the 30 phr is used.
An analysis of the optimal vulcanization time and scorch time has shown that, despite the higher active oxygen content, the number of radicals generated, and the similar half-life temperature, BIB shows lower efficiency in terms of curing time. This is due to the higher number of radicals generated by BIB compared to DCP; therefore, for BIB, the formulation of a three-dimensional network, stabilization, and the termination of free radical takes more time.
The R 300 parameter values of DCP cross-linked materials increase for 5 phr of silica, while it starts to decrease after this value (15 and 30 phr of silica). However, in the case of BIB, these values are constant regardless of the amount of filler used. As mentioned earlier, BIB exhibits longer curing times, making it not as rapid as DCP. A more intense free radical generation process can not only lead to cross-linking but also to elastomer chain degradation, as indicated by the higher values of the R 300 parameter. Moreover, with the increasing content of the filler, the values shift towards 0 value, which corresponds to better resistance for reversion during curing. This phenomenon can be related to the partial devulcanization of GTR. It is well known that the efficiency of rubber devulcanization is strongly affected by temperature [38,39], and usually increases with a higher temperature (in the studied case, 180 • C for peroxide curing system and 160 • C for curing system temperature). The partial devulcanization of GTR has a significant impact on the interfacial interaction in matrix-GTR. This is due to the additional physical interactions and also the possible migration of components between matrix and GTR, which was comprehensively described in work [40].
High temperature and pressure also enhanced the sintering of GTR particles without using additional additives [41,42]. This is related to the combined effects of the devulcanization, degradation, and secondary vulcanization (revulcanization) of GTR (or GTR with NBR) supported by free radicals' reactions.
Moreover, it should be pointed that during peroxide curing, it is also possible that the peroxide-induced degradation of GTR might occur. Sabzekar et al. [43] and Colom et al. [44] studied the effect of BPO on the thermo-mechanical devulcanization of waste rubber, but due to a similarity in the peroxide decomposition mechanisms for other peroxides, similar effects can be expected. In the case of BIB and DCP, some differences are related to their affinity of GTR and NBR.
Furthermore, it should also be pointed out that the components also present in GTR (e.g., carbon black, curing agents' residue, etc.) might have an impact on the efficiency sulfur curing system [45,46] and peroxides [47,48] during the formulation of materials modified with GTR.
As already noted, the amount of filler has a critical effect on the cross-linking characteristics of the sample. The vulcanization mechanism is affected mainly by two factors, first, the physical inhibition caused by the silica, which reduces the mobility of the rubber chains and prevents part of the rubber from participating in the vulcanization. This phenomenon is also present during the curing with sulfur. Second, the effect of the hydroxyl surface groups of the silica, which promotes the radical formation, but also enhances the agglomeration of silica. Depending on the amount of silica and curing system type, and the competition of physical and radical promotion effects, vulcanization can be affected greatly. The obtained results (changes of ∆M parameter) are confirmed by the calculation of the Wolff activity coefficient as presented in Figure 2. Interaction decreases for the sulfur system, while for peroxides, it slightly rises and then drops for 30 phr. This phenomenon can be related to the possible migration of carbon black from partially devulcanized GTR to the NBR matrix during the processing of the studied materials [49,50]. It should be highlighted that during the calculation of the Wolff activity coefficient, the carbon black present in GTR was not included as a filler phase (assumption based on the fact that the content of carbon black in GTR is constant).

Physico-Mechanical Properties of GTR/NBR/SiO2
The physico-mechanical properties of the samples are shown in Table 4. The analysis of the results, for the sulfur system, shows an increase in tensile strength (5.2 ± 0.7, 6.1 ± 0.6, 7.5 ± 0.9, and 8.5 ± 1.5 MPa), modulus at 100% (3.2 ± 0.9, 3.6 ± 0.8, 4.6 ± 0.6, and 6.8 ± 0.9 MPa), and hardness (69 ± 1, 69 ± 1, 74 ± 2, and 81 ± 2 ShA) as the amount of filler increases, while elongation at break decreases (168 ± 17, 167 ± 19, 162 ± 20, and 131 ± 21%). The second one shows better behavior and this is reflected in the strength of the material. The results may be due to restricted SiO 2 diffusion due to the high cross-link density, as well as the vulcanization kinetics of a peroxide-based system by the presence of silica. When the amount of silica increases to 30 phr, the interaction decreases because the amount of SiO 2 is too high, and agglomerates can be formed.

Physico-Mechanical Properties of GTR/NBR/SiO 2
The physico-mechanical properties of the samples are shown in Table 4. The analysis of the results, for the sulfur system, shows an increase in tensile strength (5.2 ± 0.7, 6.1 ± 0.6, 7.5 ± 0.9, and 8.5 ± 1.5 MPa), modulus at 100% (3.2 ± 0.9, 3.6 ± 0.8, 4.6 ± 0.6, and 6.8 ± 0.9 MPa), and hardness (69 ± 1, 69 ± 1, 74 ± 2, and 81 ± 2 ShA) as the amount of filler increases, while elongation at break decreases (168 ± 17, 167 ± 19, 162 ± 20, and 131 ± 21%). Those results simply show that an increasing amount of silica rises the stiffness of the material resulting in higher tensile strength, modulus at 100%, and hardness, while limiting the flexibility and rotation of the elastomer chains (decrease in elongation at break). Those assumptions are in accordance with M max. , ∆M, and the Wolff activity coefficient. Interesting relationships were observed when peroxide systems were used. In terms of tensile strength, modulus at 100%, and hardness, the increasing trend was maintained. However, in terms of elongation at break, the parameter behaves according to the value of ∆M and the Wolff activity coefficient (increase up to 15 phr of silica, then the value drops), indicating the enhancement and physical interactions occurred between the elastomeric matrix and silica due to the application of peroxides and the higher temperature during curing. As mentioned in the previous section, this observation can be related to the following three factors: (i) the partial devulcanization or revulcanization of GTR during the compression of composites at 180 • C; (ii) the possible migration of carbon black from the GTR phase to NBR; (iii) the tendency of silica filler to agglomerate.
There are also important differences based on the type of peroxide used. DCP has a lower tensile strength (5.1 ± 0.3, 5.6 ± 0.3, 7.2 ± 0.6, and 10.0 ± 1.0 MPa), modulus at 100% (2.0 ± 0.8, 1.9 ± 0.3, 2.6 ± 0.7, and 4.3 ± 0.8 MPa), and hardness (54 ± 2.57 ± 1.65 ± 1, and 78 ± 1 ShA) compared to BIB (6.0 ± 0.4, 6.8 ± 0.5, 8.4 ± 0.5, and 10.2 ± 0.6 MPa; 2.4 ± 0.3, 2.5 ± 0.5, 3.1 ± 0.8, and 5.0 ± 0.6 MPa; 58 ± 1.61 ± 1.68 ± 1, and 81 ± 1 ShA, respectively, for tensile strength, modulus at 100%, and hardness). The higher values are due to the higher stiffness of the material, which, in turn, is due to the amount of generated free radicals cross-linking the system of studied samples. In the case of elongation at break, the higher values were obtained for DCP (up to 15 phr), which only confirms the lower efficiency of the peroxide. It is interesting that the value (240 ± 23%) for 30 phr of silica is very similar to the BIB elongation at break value (225 ± 23%). Figure 3 shows graphically the differences in the strain-stress curves. The courses of the stress-strain curves for the three curing systems are similar in shape. They show only the differences discussed above in terms of modulus and elongation values. This is related to similar behavior in the development of effective interactions between the matrix and filler, supported by the formulated three-dimensional network.
the BIB elongation at break value (225 ± 23%). Figure 3 shows graphically the differences in the strain-stress curves. The courses of the stress-strain curves for the three curing systems are similar in shape. They show only the differences discussed above in terms of modulus and elongation values. This is related to similar behavior in the development of effective interactions between the matrix and filler, supported by the formulated three-dimensional network.  In order to further analyze the data and more easily determine the correlations between the results obtained as a function of the type of system and amount of filler, spider charts were prepared and are presented in Figure 4. The charts allow the observation of the coherence of the values obtained for the different properties, taking into account the previous discussions related to the cross-linking, microstructure, and their relationship with the tests performed on the samples. The general trend of a determined type of curing system is reflected in a deviation of the scheme towards one of the axes.
In the case of density, sol fraction, swelling degree, and cross-link density, all the samples show an identical trend in the function of silica content. The density and cross-link density increase, while the sol fraction and swelling degree decrease with an increasing content of silica. What is interesting is that the shapes of the figures obtained from plotting the analyzed data on the charts are very close to each other when peroxides are used, which confirms the identical cross-linking pattern of the compounds. In the case of sulfur, the chart has several key differences, particularly in terms of strength parameters (as discussed in the previous section of this subsection), sol fraction, swelling degree, and cross-link density. From the data obtained, it can be seen that sulfur is the most effective system studied (lower degree of swelling, lower sol fraction, and higher value of cross-link density); however, the sulfur-cured composites showed the lower mechanical properties (e.g., tensile strength) compared to composites cured by peroxides. It can be explained by the interaction between GTR, NBR, and the silica phase, supported by peroxide curing at a high temperature, hence, the enhanced interactions between the used components (as the combined effects of GTR devulcanization and silica dispersion level in the GTR/NBR system). This shows that the effect of reinforcement by highly dispersed silica is more important in terms of physico-mechanical properties than the degree of cross-linking of the sample.
tween the results obtained as a function of the type of system and amount of filler, spide charts were prepared and are presented in Figure 4. The charts allow the observation o the coherence of the values obtained for the different properties, taking into account th previous discussions related to the cross-linking, microstructure, and their relationship with the tests performed on the samples. The general trend of a determined type of curing system is reflected in a deviation of the scheme towards one of the axes.   In the case of density, sol fraction, swelling degree, and cross-link density, all the samples show an identical trend in the function of silica content. The density and cross link density increase, while the sol fraction and swelling degree decrease with an increas ing content of silica. What is interesting is that the shapes of the figures obtained from plotting the analyzed data on the charts are very close to each other when peroxides are used, which confirms the identical cross-linking pattern of the compounds. In the case o sulfur, the chart has several key differences, particularly in terms of strength parameter (as discussed in the previous section of this subsection), sol fraction, swelling degree, and cross-link density. From the data obtained, it can be seen that sulfur is the most effective system studied (lower degree of swelling, lower sol fraction, and higher value of cross link density); however, the sulfur-cured composites showed the lower mechanical prop erties (e.g., tensile strength) compared to composites cured by peroxides. It can be ex plained by the interaction between GTR, NBR, and the silica phase, supported by peroxide curing at a high temperature, hence, the enhanced interactions between the used compo nents (as the combined effects of GTR devulcanization and silica dispersion level in the GTR/NBR system). This shows that the effect of reinforcement by highly dispersed silica is more important in terms of physico-mechanical properties than the degree of cross linking of the sample. Figures 5 and 6 show the spectra of GTR/NBR compounds without and with silica cured with the three systems studied in the 700-3850 cm −1 range. The absorbance maxima in 2965 cm −1 is related to the aromatic C-H bonds vibrations (e.g., styrene-butadiene rub ber and carbon black present in GTR). The signals in the rank of 2850-2950 cm −1 are at  Figures 5 and 6 show the spectra of GTR/NBR compounds without and with silica cured with the three systems studied in the 700-3850 cm −1 range. The absorbance maxima in 2965 cm −1 is related to the aromatic C-H bonds vibrations (e.g., styrene-butadiene rubber and carbon black present in GTR). The signals in the rank of 2850-2950 cm −1 are attributed to the C-H vibrations of CH 2 groups present in the structure of elastomers. As can be observed, the intensity of these signals for composites cured with peroxides are much higher compared to sulfur-cured samples (the same tendency was observed for samples with 5 phr of silica, see Figure 6). This can be due to the following two factors: (i) during the decomposition of peroxides, low molecular by-products are formulated, and some residual is left on the material surface; (ii) in the GTR/NBR blends, a possible migration of the carbon black (strong absorber of infrared radiation [51]) present in GTR occurs and its efficiency is correlated with the progress of GTR devulcanization. Regardless of the curing system type, the level of the signal in 2850-2950 cm −1 is reduced with the increasing content of silica. tained in GTR and 964 cm (trans) RCH=CHR'.

FTIR Analysis
As reported in a previous work [52], there is a weak FTIR signal at 1540 cm −1 , which is present in the samples cured with BIB and DCP. This is attributed to the first vulcanization of GTR, which has been made using the sulfur system and, therefore, also contains some ZnO and stearic acid. FTIR spectra of GTR/NBR samples filled with 5 and 30 phr of silica and the spectrum of SiO2 are presented in Figure 6. In this figure, the increase in the band present in the SiO2 spectrum is comparatively evident in the studied samples according to their composition. A wideband at 1090 cm −1 increases with the content of SiO2 overlapping other bands. There is also a comparative diminution of the intensity of all the bands with the increase in SiO2, observed especially at the contents of SiO2 around 30 phr. These spectra were obtained using ATR, which measures the surface of the samples; the results are indicative of a migration of the particles of SiO2 from the core of the sample to the surface. In our previous work [52], we also observed that the materials cured with the sulfur system present a higher cross-link density than the ones cured with DCP and that produces a lower signal in the ATR spectroscopy of NR/GTR. This phenomenon is also produced here. The samples cured using the sulfur process present less migration from the core to the surface of the SiO2 and this is observed as a smaller increase in the related band. The data are coherent with the previous results in cross-linking.
FTIR analysis demonstrated that the silica filler and type of curing system used have a significant influence on the chemical structure of the obtained GTR/NBR-based composites, which is due to the difference in the three-dimensional network formulated by the sulfur or peroxide system and also the level of physical interactions between GTR/NBR and silica filler supported by GTR devulcanization.

Thermogravimetric Analysis
The results of the thermogravimetric analysis of GTR/NBR/SiO2 samples are presented in Figure 7 and summarized in Table 5. It was observed that the type of curing system and the amount of SiO2 influence TGA thermograms, resulting in a different thermogram for every sample studied. This is coherent with the results presented previously. To avoid the combined effect of thermal and thermo-oxidative degradation, initially, the measurements were made under nitrogen, although to compile the thermo-oxidation of the samples at 550 °C, airflow was used. This experiment was carried out until the degradation of the organic components and carbon black was complete, to determine the stability of the composites and the number of inorganic residua (mainly SiO2 and other compounds generated in the thermo-oxidation process). Different data have been obtained at 2, 5, 10, and 50% of weight loss, which corresponded to the T−2%, T−5%, T−10%, and T−50% temperatures. The values corresponding to the 2% were higher for the sulfur curing system than in the samples cured by peroxides. This is due to the nature of cross-linking in the presence of peroxides. As a result of the generated free radicals, not only crosslinking but The sample cured with sulfur presents two significant FTIR bands at 1538 and 1398 cm −1 assigned to zinc stearate formed during the reaction between ZnO and stearic acid during rubber compounding [6]. These bands are only present in the samples cured with the sulfur curing system because samples with DCP or BIB do not include this component. The bands at 1260 and 1150 cm −1 , also specific to the sulfur samples, are assigned to C-O products obtained in reactions of sulfur systems.
Other relevant bands, common to the presented spectra are 1450 cm −1 related to the CH 2 − stretching GTR/NBR bands; a small band at 1375 cm −1 , corresponding to the CH 3 symmetric bend of GTR; a small band at 1018 cm −1 related to the C-C of carbon black contained in GTR and 964 cm −1 (trans) RCH=CHR'.
As reported in a previous work [52], there is a weak FTIR signal at 1540 cm −1 , which is present in the samples cured with BIB and DCP. This is attributed to the first vulcanization of GTR, which has been made using the sulfur system and, therefore, also contains some ZnO and stearic acid.
FTIR spectra of GTR/NBR samples filled with 5 and 30 phr of silica and the spectrum of SiO 2 are presented in Figure 6. In this figure, the increase in the band present in the SiO 2 spectrum is comparatively evident in the studied samples according to their composition. A wideband at 1090 cm −1 increases with the content of SiO 2 overlapping other bands. There is also a comparative diminution of the intensity of all the bands with the increase in SiO 2 , observed especially at the contents of SiO 2 around 30 phr. These spectra were obtained using ATR, which measures the surface of the samples; the results are indicative of a migration of the particles of SiO 2 from the core of the sample to the surface.
In our previous work [52], we also observed that the materials cured with the sulfur system present a higher cross-link density than the ones cured with DCP and that produces a lower signal in the ATR spectroscopy of NR/GTR. This phenomenon is also produced here. The samples cured using the sulfur process present less migration from the core to the surface of the SiO 2 and this is observed as a smaller increase in the related band. The data are coherent with the previous results in cross-linking.
FTIR analysis demonstrated that the silica filler and type of curing system used have a significant influence on the chemical structure of the obtained GTR/NBR-based composites, which is due to the difference in the three-dimensional network formulated by the sulfur or peroxide system and also the level of physical interactions between GTR/NBR and silica filler supported by GTR devulcanization.

Thermogravimetric Analysis
The results of the thermogravimetric analysis of GTR/NBR/SiO 2 samples are presented in Figure 7 and summarized in Table 5. It was observed that the type of curing system and the amount of SiO 2 influence TGA thermograms, resulting in a different thermogram for every sample studied. This is coherent with the results presented previously. To avoid the combined effect of thermal and thermo-oxidative degradation, initially, the measurements were made under nitrogen, although to compile the thermo-oxidation of the samples at 550 • C, airflow was used. This experiment was carried out until the degradation of the organic components and carbon black was complete, to determine the stability of the composites and the number of inorganic residua (mainly SiO 2 and other compounds generated in the thermo-oxidation process). Different data have been obtained at 2, 5, 10, and 50% of weight loss, which corresponded to the T −2% , T −5% , T −10%, and T −50% temperatures. The values corresponding to the 2% were higher for the sulfur curing system than in the samples cured by peroxides. This is due to the nature of cross-linking in the presence of peroxides. As a result of the generated free radicals, not only crosslinking but also the scission of the main bonds can occur, resulting in the formation of low molecular weight compounds. Moreover, it should be mentioned that during the decomposition of peroxides, low molecular by-products are also formulated. These compounds generally have a lower degradation temperature than the rubber itself, and for this reason, a faster onset of degradation was observed for the peroxides.
composites. In sulfur-cured samples, some inorganic compounds generated by reaction in the thermo-oxidative process also appear. It is also worth remarking that the DTG plot shows that thermal stability increases with silica and it is slightly higher in samples cured with peroxides. In these DTG curves, three peaks are observed, assigned to (a) the thermal decomposition of the NR and BR of GTR at 410 °C, (b) the thermal decomposition of the SBR of GTR at 465 °C and NBR at 470 °C, and (c) a broad peak at 630 °C that corresponds to the thermal decomposition of carbon black [53].

Acoustic Properties
The sound absorption coefficient as a function of the frequency of the studied samples is presented in Figure 8. All the materials presented showed similar curves, regardless of the amount of filler and the type of cross-linking system. The only noticeable difference is the peak height in the 3000-5000 Hz range for sample GTR/NBRDCP/Si30. The greatest influence on the acoustic properties of materials is their physical and chemical structure, as well as being strongly related to the density of the test samples [54].   Figure 7 shows the thermograms (TGA and DTG) for all the samples. The increase in the char residue of the samples is due to the presence of a high proportion of silica in the composites. In sulfur-cured samples, some inorganic compounds generated by reaction in the thermo-oxidative process also appear. It is also worth remarking that the DTG plot shows that thermal stability increases with silica and it is slightly higher in samples cured with peroxides. In these DTG curves, three peaks are observed, assigned to (a) the thermal decomposition of the NR and BR of GTR at 410 • C, (b) the thermal decomposition of the SBR of GTR at 465 • C and NBR at 470 • C, and (c) a broad peak at 630 • C that corresponds to the thermal decomposition of carbon black [53].

Acoustic Properties
The sound absorption coefficient as a function of the frequency of the studied samples is presented in Figure 8. All the materials presented showed similar curves, regardless of the amount of filler and the type of cross-linking system. The only noticeable difference is the peak height in the 3000-5000 Hz range for sample GTR/NBR DCP/Si30 . The greatest influence on the acoustic properties of materials is their physical and chemical structure, as well as being strongly related to the density of the test samples [54]. The sound absorption is related to sample structure, properties, thickness, and surface conditions, as well as to the incident angle and frequency of the sound waves [56].
During the test, analysis is performed over a wide range of frequencies (500-6000 Hz), and the values change depending on the frequency value. Therefore, the sound-absorbing property can be determined by analyzing specific frequency values (500, 1000, 2000, and 4000 Hz). When the average frequency equals or is higher than 0.2, the tested material can be called "sound-absorbing". In Table 6, the sound absorption coefficient for the selected regions and its average value is presented. The results show that none of the Therefore, the acoustic properties result directly from the type of material, particle size distribution, methods of its preparation, and thickness [55]. Taking that into consideration, and the fact that all the samples were characterized with a relatively high degree of crosslinking, one may conclude that is the reason for the similar sound absorption behavior.
The sound absorption is related to sample structure, properties, thickness, and surface conditions, as well as to the incident angle and frequency of the sound waves [56].
During the test, analysis is performed over a wide range of frequencies (500-6000 Hz), and the values change depending on the frequency value. Therefore, the sound-absorbing property can be determined by analyzing specific frequency values (500, 1000, 2000, and 4000 Hz). When the average frequency equals or is higher than 0.2, the tested material can be called "sound-absorbing". In Table 6, the sound absorption coefficient for the selected regions and its average value is presented. The results show that none of the tested materials are good enough to be a sound-absorbing material.

SEM
SEM micrographs (magnification ×500) of tensile fractured samples are presented in Figure 9. Figure 9A includes the GTR/NBR compounds cured by sulfur with 5 and 30 phr silica, Figure 9B analog compounds cured with DCP, and Figure 9C with BIB. The structure of the samples in Figure 9B, corresponding to the samples cured with DCP, stands out because of its rough surface with many gaps and voids. The pictures labeled as 9C corresponding to the samples cured with BIB also present a rough surface, though they are not as notorious as the samples in Figure 9B.
The differences observed in the structure of the presented samples can be related to several factors. One of them is the curing temperature used in the preparation of the samples combined with the higher cross-link found in the samples cured with sulfur. The sulfur system is cured at 160 • C and peroxides are cured at 180 • C. These samples, as discussed above, differ in mechanical properties, and the samples cured with sulfur show a more fragile fracture than the samples leading to a cleaner surface, while in the samples cured with peroxides, the surface appears more strained and plastically deformed.
As discussed in the curing characteristics section, the higher temperature used during peroxide curing enhanced interfacial adhesion in the GTR/NBR system reinforced by silica, due to the combined effects of the devulcanization, degradation, and secondary vulcanization (revulcanization) of GTR (or GTR with NBR) supported by free radicals' reactions. These observations corresponded with tensile properties measurements, which were higher for the composites cured by peroxides than the sulfur-based system.
Moreover, the tendency of silica to agglomerate also affects the interfacial compatibility in the studied materials. De et al. [57] investigated the impact of reclaimed rubber (GTR with a particle size up to 0.1 mm was used) on silica reinforcement (content: 11.6-16.9 wt.%) of SBR/reclaimed blends (ratio from 80/20 to 40/60 wt.%). The SEM micrographs (prepared in the same magnification as in the present study) showed that silica filler disperses uniformly in both phases: SBR and reclaimed rubber. In the present study, the GTR particles are much bigger compared to the images presented by De et al. [57], which confirms only the partial devulcanization of the GTR used (with an average particle size below 0.8 mm). Comparing the mechanical properties, De et al. [57] showed that for unfilled SBR/reclaimed vulcanizates, the tensile strength and elongation at break were 2.8 MPa and 377% for SBR/reclaimed rubber in a ratio 80/20 wt.%, while for SBR/reclaimed rubber, the ratio values for these parameters were 5.0 MPa and 455%, respectively. This confirms the possible migration of carbon black from reclaimed rubber to fresh rubber matrix. In this study, regardless of the curing system type used, for the unfilled GTR/NBR system (70/30 wt.%), the tensile strength and elongation at break were in the range of 5.1-6.0 MPa and 168-304%.

Conclusions
The study presented a complex analysis of GTR/NBR blends reinforced with different amounts of silica and cured with three various curing systems. The curing characteristic was served to give detailed information about the influence of the applied variables. Regarding the amount of silica used, deviations from the linearity of results (decrease in ΔM and Wolff coefficient) were observed after exceeding the value of 15 phr. In the case of a decreasing ΔM and Wolff coefficient, one can also expect a decrease in the mechanical properties; however, a constant increase in tensile strength with an increasing filler amount was observed. This phenomenon is due to the partial devulcanization of GTR, which enhances the interfacial interactions between the components used in the GTR/NBR/silica system. The curing characteristics indicate that the most efficient curing system is the sulfur-based one; however, the opposite outcome can be read from physicothe mechanical properties (higher tensile strength and elongation at break for peroxides). It indicates the occurrence of an enhanced interaction between GTR, NBR, and silica supported by the peroxides. These data are also in accordance with the FTIR spectra, which show a significant difference in the chemical structure of the three-dimensional network, The authors [57] also indicate that for SBR/reclaimed rubber in a ratio of 40/60 wt.% filled with 11.6 wt.%, the tensile strength, elongation at break, and hardness parameters were 8.1 MPa, 403%, and 68 ShA. For comparison, in the present study, regardless of the curing system type used, the values of tensile strength, elongation at break, and hardness of the studied GTR/NBR system samples (70/30 wt.%) with 15 phr of silica (11.9 wt.% for sulfur-based system and 12.8 wt.% for peroxide system, see Table 2) were in the range of 7.2-8.4 MPa, 162-313%, and 65-74 ShA, respectively. The presented results confirmed the significant improvement of tensile properties by the incorporation of silica into the rubber compound modified by ground tire rubber, which fully justifies the future development of research in this field.

Conclusions
The study presented a complex analysis of GTR/NBR blends reinforced with different amounts of silica and cured with three various curing systems. The curing characteristic was served to give detailed information about the influence of the applied variables. Regarding the amount of silica used, deviations from the linearity of results (decrease in ∆M and Wolff coefficient) were observed after exceeding the value of 15 phr. In the case of a decreasing ∆M and Wolff coefficient, one can also expect a decrease in the mechanical properties; however, a constant increase in tensile strength with an increasing filler amount was observed. This phenomenon is due to the partial devulcanization of GTR, which enhances the interfacial interactions between the components used in the GTR/NBR/silica system. The curing characteristics indicate that the most efficient curing system is the sulfur-based one; however, the opposite outcome can be read from physico-the mechanical properties (higher tensile strength and elongation at break for peroxides). It indicates the occurrence of an enhanced interaction between GTR, NBR, and silica supported by the peroxides. These data are also in accordance with the FTIR spectra, which show a significant difference in the chemical structure of the three-dimensional network, as well as in the level of physical interactions. The thermogravimetric analysis reflects the mentioned results. Peroxides may partially degrade the structure of blends resulting in the generation of VOCs shifting T −2% toward lower values. It indicates a potential risk of hazardous substances being transferred to the environment with a harmful effect on human life. This aspect requires further detailed analysis with a focus on VOCs' detection and classification. The analysis of acoustic properties allowed us to determine the suitability of the obtained materials as sound absorbing products. For this purpose, the sound absorption coefficient was studied for four different frequencies and the average was determined. When the value is ≥0.2, the material is considered to be sound absorbing. However, the results indicate that none of the materials obtained meet the expected minimum, with values oscillating around 0.06. This means that in the case of the analyzed samples, the main factor affecting the discussed parameter is the matrix material. Despite the use of different crosslinking systems, which were characterized by different degrees of cross-linking (sulfur vs. peroxides), the coefficient remained the same. Moreover, this conclusion is supported by SEM analysis, showing differences in the morphology of the samples when different cross-linking systems and filler amounts were used.