Hybrid In Situ Reinforcement of EPDM Rubber Compounds Based on Phenolic Novolac Resin and Ionic Coagent

Abstract

For the design of stretchable and flexible high-performing materials, the reinforcement of elastomeric grades plays a crucial role. State-of-the-art fillers such as carbon black benefit from a high reinforcement but often negatively affect the processing and mixing properties of rubber compounds. To overcome this drawback, the synergistic properties of hybrid in situ filler systems are studied for EPDM compounds comprising a phenol novolac resin and ionic coagents such as zinc (meth)acrylates (ZD(M)A. With the help of a combined novolac/ZD(M)A system, the compounds could be tailored in a unique way towards higher toughness and enhanced cross-link density. Further, the fracture surface of the EPDM–novolac compounds was analyzed by scanning electron microscopy, revealing a significant change of the morphology from rough and disordered to smooth and homogenous for samples with coagents. In addition, the results clearly showed that the introduction of ionic coagents is able to compensate shares of carbon black filler in the EPDM compound. The toughening of samples with zinc (meth)acrylates is attributed to the synergistic formation of an interpenetrating polymer-filler network by simultaneous covalent and ionic cross-linking.

1. Introduction

Ethylene-propylene-diene rubber (EPDM) is a terpolymer of ethylene, propylene, and a nonconjugated diene, and characterized by a fully saturated, hydrocarbon main chain with low levels of unsaturation in the side groups. The presence of the diene enables sulfur vulcanization of EPDM and enhances peroxide cross-linking efficiency. This low unsaturated, nonpolar rubber polymer is known for its good resistance against ozone, oxygen, irradiation, and heat and against polar media, such as aqueous solutions and polar solvents [1]. In industry, EPDM finds application in the manufacturing of high-performance automotive products. A prominent example is the V-belt, which is mainly used for the power transmission between pulleys. Due to its excellent physical attributes, EPDM is a suitable material for such a demanding application [2]. However, the EPDM matrix needs to be modified to unfold its full potential. To obtain superior vulcanizate properties, most rubbers, including EPDM, need to be reinforced with fillers. Today, carbon black (CB) and silica are the most common reinforcing fillers in the rubber industry [3,4]. The reinforcing effect is not only governed by the chemical nature of the filler but also by several other factors, for example, the size, surface area, shape, and structure of the filler particles, and their dispersion and distribution [5,6,7]. In particular, the dispersion of the filler within the rubber matrix plays a distinctive role in the overall reinforcement. Silica fillers are especially known for their strong filler–filler networking, leading to agglomeration and, thus, poor filler dispersion. Furthermore, the mixing and processing of rubber compounds are aggravated by the viscosity increase upon the introduction of solid fillers [8,9]. EPDM V-belts, on the one hand, require sufficient reinforcement to satisfy the application requirements, but on the other hand, they should also show a good flow for wetting the reinforcing textile during belt production. Achieving a sound balance between these contradictory application and production needs requires some subtle optimization of the rubber compound composition or a specialized approach.

Rigid, reinforcing particles can also be introduced in a soft rubbery matrix by the in situ formation of an interpenetrating network (IPN). An IPN is defined as a combination of two cross-linked polymers, where one polymer network is cross-linked in the presence of a second polymer network [10]. For industrial applications, the generation of simultaneous IPN (SIN) is the most feasible approach. SIN formation occurs within a “one-pot” reaction, where all components are mixed and subsequently cross-linked simultaneously. The advantage of the IPN/SIN approach for EPDM V-belt applications is that on the one hand, the viscosity of the rubber compound is reduced, and on the other hand, the final vulcanized part is reinforced. The key to successfully forming SINs is the combination of two independent and noninterfering reaction mechanisms. This would be applicable for the radical cross-linking of EPDM polymer and the step-growth mechanism of phenolic resin [10,11,12]. Phenolic resins are already of high industrial relevance and are usually applied as tackifiers, cross-linking aids, or reinforcing agents of rubbers [13]. Especially phenol-formaldehyde (PF) novolac resins have been identified as useful reinforcing agents due to their high storage stability. Even though PF is a relatively polar resin, it can be modified with fatty acid tails to blend with nonpolar EPDM [14]. In addition, it has been reported that the further addition of CB and silica has an advantageous effect on the overall reinforcement. The synergistic effect of the filler and PF leads to a more uniform resin distribution with superior mechanical properties [15,16,17]. However, a distinct drawback of PF is their interference with peroxide cross-linking. This effect could be explained by their structural resemblance to radical scavengers such as phenolic antioxidants and pyrogallol [14,18,19].

Another approach suitable for the reinforcement of peroxide-cured rubber is the use of ionic coagents. Coagents are multifunctional, highly reactive monomers that are typically used for the peroxide cross-linking of (EPDM) rubber to enhance the peroxide cross-linking efficiency and provide a higher cross-link density. In particular, (meth)acrylic coagents, such as trimethylolpropane trimethacrylate (TRIM), are known to increase the cross-linking rate due to the formation of very reactive radicals [20,21]. Ionic coagents, based on (meth)acrylic salts, may combine the action of peroxide coagent with that of a reinforcing agent. Zinc (meth)acrylate (ZD[M]A) ionic coagents are extremely polar and do not dissolve in the nonpolar EPDM matrix. As a result, they form small, phase-separated ionic domains during peroxide cross-linking, which are tightly connected to the rubber matrix via multiple covalent bonds [20].

Thus, in the current work, a standard EPDM compound used for V-belt applications was compared with (hybrid) in situ reinforcing routes. The amount of CB was reduced in order to guarantee better compound processability, and the reinforcement was introduced by PF and ionic coagents. The effect on cross-linking rate, cross-link density, morphology, and tensile properties of the related compounds was studied in detail.

2. Materials and Methods

2.1. Materials

Keltan 6160D (structure depicted in Figure 1a) was used as the EPDM matrix and was received from ARLANXEO (The Hague, The Netherlands). The rubber was peroxide cross-linked with di(tert-butylperoxyisopropyl) benzene (Perkadox 14-40 B-PD; PO) from Nouryon (Amsterdam, The Netherlands). Corax N660 from Orion Engineered Carbons (Senningerberg, Luxembourg) was applied as reinforcing carbon black filler. In addition, the tall-oil modified phenolic novolac resin SP-6701 (PF) was purchased from SI Group (Schenectady, NY, USA), and the corresponding hardener hexamethylenetetramine (Rhenogran Hexa-80; HMT) was obtained from RheinChemie Additives (Cologne, Germany). The chemical structure of the PF resin and hardener are shown in Figure 1b,c. The studied zinc-based coagents Sartomer SR 633 (ZDA) (Figure 1f) and Dymalink 634 (ZDMA) (Figure 1e) were purchased from Cray Valley (Exton, PA, USA), whereas Rhenofit TRIM (Figure 1d) was supplied by RheinChemie (Cologne, Germany). Other additives included the process oil PLI P 460 from Petronas (Kuala Lumpur, Malaysia) and the antioxidants Vulkanox HS and Vulkanox ZMB2/C5, both from Lanxess (Cologne, Germany). The toluene for equilibrium swelling and gel content determination was purchased from Sigma Aldrich (St. Louis, MO, USA). All chemicals have been used without further modification.

2.2. Processing of EPDM Compounds

For the preparation of the EPDM compounds, all components except the curatives (PO, HMT) were transferred to an internal mixer (type GK 1.5 E/1, Werner & Pfleiderer, Stuttgart, Germany). The internal mixer was equipped with intermeshing rotors and a mixing chamber of 1500 cm³, using a fill factor of 72%. The rotor speed was set to 20 rpm and the ram pressure to 8 bar. The initial mixing temperature was set to 30 °C. The compound was mixed for 5 min to ensure a complete homogenization of the components, following the steps stated in

Table 1. Mixing parameters for the EPDM compounds.

Time (min)	Steps
0	Addition of rubber
0–0.5	Crumbling of rubber
0.5–1.0	Addition of filler, oil, reinforcing agent, and antioxidants
1.0–3.0	Mixing (ram down)
3.0–3.2	Sweeping (ram up)
3.2–5.0	Mixing (ram down)
5.0	Dumping of compound

.

The final dumping temperature was around 110 °C. The batch was next transferred onto a two-roll mill (model WNU3, Troester, Hannover, Germany) with roll speeds of 20 and 22 rpm. The mixture was allowed to cool down to 40 °C, and subsequently, the curing agents (PO and HMT) were incorporated. After complete homogenization, the compound was sheeted out for either compression molding (gap width 2 mm) or rheometric analysis (gap of 5–6 mm).

For vulcanization, the rubber mixture was cut into 200 × 200 × 2 mm sheets. All compounds were vulcanized at 180 °C and 120 bar in a hydraulic laboratory press (Model TP1500, Fontijne Presses, Delft, The Netherlands). The vulcanization time was calculated according to the t₉₀ time and set to 2 × t₉₀.

The samples were further cross-linked at 150 °C for 24 h in a drying oven (Model UF55, Memmert, Schwabach, Germany) to study the influence of the postcuring. The abbreviations 0 h and 24 h are added to the sample codes for clarification in the discussion section and denote the time of postcuring.

The compositions of the studied compounds are given in

Table 2. Compositions of in situ reinforced, peroxide-cured EPDM compounds ¹.

Compound	EPDM	CB N660	Process Oil	PO	Vulkanox HS	Vulkanox ZMB2	ZDMA	ZDA	TRIM	PF	HMT
Belt–Ref	100	50	15	7	1.5	1.5	-	-	4	-	-
EPDM–7.5ZDMA	100	30	15	5	1.5	1.5	7.5	-	-	-	-
EPDM–15ZDMA	100	30	15	2.5	1.5	1.5	15	-	-	-	-
EPDM–7.5ZDA	100	30	15	5	1.5	1.5	-	7.5	-	-	-
EPDM–15ZDA	100	30	15	2.5	1.5	1.5	-	15	-	-	-
EPDM–TRIM	100	30	15	7	1.5	1.5	-	-	4	-	-
PF–ZDMA	100	30	15	5	1.5	1.5	7.5	-	-	30	3
PF–ZDA	100	30	15	5	1.5	1.5	-	7.5	-	30	3
PF–TRIM	100	30	15	7	1.5	1.5	-	-	4	30	3
PF–5PO	100	30	15	5	1.5	1.5	-	-	-	30	3
PF–7PO	100	30	15	7	1.5	1.5	-	-	-	30	3

¹ Amount of ingredients in parts per hundred rubber (phr).

.

2.3. Characterization of Nonvulcanized Compounds

The investigation of the vulcanization characteristics of the rubber compounds has been conducted with a rotorless, moving die rheometer (MDR) (Model MDR 2000E, Alpha Technologies, Hudson, OH, USA). The measurements were carried out at a testing temperature of 180 °C for 30 min. For the discussion of the vulcanization characteristics, the minimum elastic torque (M_L), the maximum elastic torque (M_H), the scorch time (t₂), and the time to reach 90% of complete cure (t₉₀) were considered.

A Mooney viscosimeter (Model MV 2000, Alpha Technologies, Hudson, OH, USA) was used to determine the Mooney viscosity. The measurements were carried out at 100 °C and gave the results in Mooney units (MU). The final Mooney viscosity is reported as ML1+4, where L refers to applying a large rotor, 1 equals the preheating time, and 4 indicates the rotation time.

2.4. Material Characterization of EPDM Vulcanizates

The vulcanized and post-cured sheets were prepared for tensile testing and tear resistance in the lengthwise direction of the rolling during milling with a cutting press (Zwick/Roell, Ulm, Germany). Both tensile and tear testing were performed on a universal material testing machine (Model Z001, Zwick/Roell, Ulm, Germany) at room temperature (23 °C).

The specimens were clamped with 50 mm for tensile testing, and the crosshead speed was set to 200 mm/min according to DIN 53504 [22]. The strain was monitored with the real-time tracking MercuryRT system (Version x64 2.4.2, Sobriety, Kurim, Czech Republic).

Each reported tensile data is a mean value of five specimens. Tear resistance was performed on the nicked angle test specimen. The measurements were conducted at room temperature (23 °C) at 500 mm/min crosshead speed according to ISO 34-1 [23]. Each reported tear strength data is a mean value of three specimens.

The Shore A hardness was determined according to DIN ISO 7619-1 [24]. The hardness was obtained from 6 mm thick samples after 3 s of indentation time of the truncated cone. Each reported Shore A hardness value is the median of three specimens

The compression set (CS) was analyzed according to DIN ISO 8150-1 [25]. The compression was performed at room temperature (23 °C) or 100 °C for 24 h. Finally, the compression set was calculated according to Equation (1), where H_b refers to the thickness before and H_a after compression. H_cs is defined as the compressed height (4.73 mm) of the sample.

$$CS (\%)=\frac{H_b-H_a}{H_b-H_{cs}}\times 100$$

(1)

All compression set results are given as average values obtained from the data of three specimens.

2.5. Dynamic Mechanical Analysis

The storage modulus (E’) and the loss factor (tan δ) of the postcured samples were analyzed via dynamic mechanical thermal analysis (DMTA). The samples were prepared with a cutting press (type 7100.002, Zwick/Roell, Ulm, Germany) equipped with a DIN 53504-S2 die. The narrow part of the multipurpose specimen was analyzed. The experiment was performed on a DMA 8000 (Perkin Elmer, Waltham, MA, USA) in tensile mode with a clamping length of 1 cm, a pre-load of 8 N, and a heating rate of 3 K/min at an oscillation frequency of 1 Hz in the range of −50 to 220 °C.

2.6. Morphology

For morphology measurements, the subsequent fractures of the tear specimen (obtained according to Section 2.4) were analyzed. The obtained fracture surface of the tear specimen at RT was investigated utilizing scanning electron microscopy (SEM; Tescan VEGA-II, Brno, Czech Republic) at an accelerating voltage of 5 kV. One specimen was analyzed for each vulcanizate.

2.7. Equilibrium Swelling and Rubber Gel Content

To determine the swelling degree, a small rubber piece (150–200 mg) was swollen for 24 h at RT in toluene. The swelling degree was calculated according to Equation (2), where the weight of the swollen rubber piece (w_a) was related to its initial weight (w_b). Subsequently, the swollen sample was dried for 24 h at 50 °C, which gave the dried weight (w_d) of the compounds. For the calculation of the gel content, the composition of the vulcanizate was divided into initial rubber weight (w_r), the weight of insoluble (w_insoluble) including fillers, resin, and coagents, and weight of solubles (w_solubles) including process oil and 40% peroxide.

The rubber gel content was calculated according to Equation (3)

$$\mathrm{Swelling}\mathrm{degree}(\%)=\left(\frac{w_{\mathrm{a}}}{w_{\mathrm{b}}}-1\right)\times 100$$

(2)

$$\mathrm{Rubber}\mathrm{gel}\mathrm{content}(\%)=\frac{w_d-w_{insolubles}}{w_r}\times 100$$

(3)

The cross-link density (M_c) was calculated according to a modified Flory–Rehner relationship [26,27]. This relationship includes the consideration of the volume fraction of the rubber in the swollen gel (v_r), which is calculated according to Equation (4), where $\alpha$ refers to the weight loss of rubber compounds during swelling, ${\rho }_s$ to the density of toluene, and ${\rho }_r$ to the density of EPDM rubber. The weight fraction of the elastomer network, comprising EPDM, coagent, and peroxide, is indicated by $\varphi$.

$$v_r=\frac{w_b\times \mathsf{\varphi }\frac{1-\alpha }{{\rho }_r}}{w_b\times \mathsf{\varphi }\frac{1-\alpha }{{\rho }_r}+\frac{w_a-w_d}{{\rho }_s}}$$

(4)

The calculated value for v_r is further used to determine M_c of the rubber compounds according to Equation (5). n_w refers to the network chain density, f to the functionality of cross-links (which is set to 4), V_s the molar volume of the solvent, and $\chi$ the Flory–Huggins interaction parameter between rubber and solvent.

$$M_c\left(\frac{g}{mol}\right)=\frac{1}{n_{sw}}=-\frac{{\rho }_r{V}_s\left(1-\frac{2}{f}\right)\left(v_r^{\frac{2}{3}}-0.5v_r\right)\left(1.3-0.4v_r\right)}{\ln \left(1-v_r\right)+v_r+\chi v_r^2}$$

(5)

The obtained results were collected from three specimens and given as a mean value.

3. Results and Discussion

3.1. Influence of Ionic Coagents on Reinforcement

A standard V-belt has to withstand high stresses during operation. Thus, EPDM compounds used for V-belt production typically contain a high amount of carbon black to meet the demands for high tensile strength. It should be noted that for the production of V-belts, various textile fabrics are applied to further increase the strength of the final products [2,28]. Herein, a reference EPDM compound was prepared using 50 phr CB (Belt–Ref). At this filler concentration, the mixing and processing performance of the EPDM compound is already negatively affected by the high viscosity of the filler system. In addition, due to the high viscosity of the compound, challenges with efficiently impregnating the reinforcing fabrics are expected to arise. To overcome the high viscosity without losing mechanical strength, it was evaluated if 20 phr of CB can be removed and replaced by the introduction of ionic coagents. Coagents are well known for their boost of cure rate and improved cross-link density. However, in nonpolar rubbers such as EPDM, they can in situ form small domains and generate reinforcing particles in the rubber matrix [20]. For the standard system Belt–Ref, the performance of a nonionic coagent TRIM was compared with ionic zinc di(meth) acrylates (ZDMA and ZDA). The used coagents are depicted in Figure 1 and are characterized by a different functionality. Whereas TRIM exhibits three methacrylate moieties, ZD(M)A only contributes to two functional groups. Therefore, the ZD(M)A content was set to higher levels than the TRIM compounds in the current study. Further, it should be noted that the peroxide content was adapted to lower contents for the ZD(M)A compounds to avoid reversion during rubber cure.

A unique property of ZD(M)A coagents is the introduction of ionic bonds due to the present zinc ions. This yields a synergistic reinforcing system comprising covalent and ionic cross-links. The cure characteristics of nonvulcanized compounds containing coagents were determined via moving die rheometer (MDR) analysis and are depicted in Figure 2a and summarized in

Table 3. MDR rheometer characteristics and Mooney viscosity of reinforced EPDM compounds under investigation.

Sample	t90 (min)	t2 (min)	MH (dNm)	ML (dNm)	ΔM (dNm)	ML1+4 (MU)
EPDM–7.5ZDMA	5.3	0.66	13.5	1.55	12.0	81
EPDM–15ZDMA	5.9	0.92	11.4	1.30	10.1	83
EPDM–7.5ZDA	4.2	0.50	15.8	1.59	14.2	76
EPDM–15ZDA	4.1	0.67	16.7	1.51	15.2	76
EPDM–TRIM	5.4	0.68	16.0	1.62	14.4	75
PF–ZDA	6.5	0.72	12.6	1.11	11.6	67
PF–ZDMA	6.2	0.77	11.4	1.11	10.3	84
PF–TRIM	8.4	0.72	14.3	1.14	13.2	59
PF–7PO	11.0	1.04	10.7	1.19	9.5	73
PF–5PO	11.6	1.33	8.4	1.19	7.2	74
Belt–Ref	5.4	0.61	19.6	1.95	17.6	89

. Several critical parameters, including the optimum cure time (t₉₀), scorch time (t₂), minimum torque (M_L), and maximum torque (M_H), were studied.

The rheogram of the studied EPDM compounds shows the influence of the coagent on cure rate and cross-link density. All samples comprise a decent cure rate, which is attributed to the presence of the respective coagents. However, the introduction of ZDA results in a significant improvement of the optimum cure time. Methacrylates are known to form tertiary radicals, which are much more stable compared with the secondary radicals of acrylates. Therefore, ZDA reduces the optimum cure to a greater extent than TRIM or ZDMA [20]. In addition, the scorch time is influenced, which is defined as the period before the vulcanization process starts. For EPDM–7.5ZDA, the vulcanization is initiated at an earlier stage than the other samples due to the less stable acrylate radicals [20].

In general, the maximum torque is proportional to the cross-link density of compounds. Belt–Ref exhibits by far the highest maximum torque due to the reinforcing power of CB. As a result, the reduced maximum torque value of the EPDM compounds containing the selected coagents reflects the loss of 20 phr of CB. The cross-link density is strongly affected by the reactivity and functionality of the coagent. Thus, whilst compounds containing TRIM and ZDA exhibit a decent cross-link density, compounds with ZDMA suffer from a low cross-link density.

In a further step, the content of the ionic coagents was increased from 7.5 to 15 phr to evaluate how a higher amount of ionic reinforcing domains influences the cure behavior and the viscosity of the compounds. The results revealed that both Mooney viscosity and cross-link density increase with increasing ZDA content, which is related to a higher number of available functional moieties. In contrast, ZDMA does not provide a suitable reinforcing capacity, and 15 phr even contributes to inferior properties.

The cross-link density can also be qualitatively assessed by the swelling degree of compounds. A low swelling generally relates to a higher cross-link density due to the higher resistance of the network against swelling.

Table 4. Swelling degree, gel content, and cross-link density of EPDM compounds prior to and after post-curing for 24 h at 150 °C.

Sample	Swelling Degree (%)		Gel Content (%)		Mc (mol/g)
	0 h	24 h	0 h	24 h	0 h	24 h
EPDM–7.5ZDMA	158 ± 0.6	161 ± 0.4	100 ± 0.1	100 ± 0.1	740 ± 3	730 ± 3
EPDM–15ZDMA	196 ± 1.0	187 ± 0.4	96 ± 0.9	98 ± 0.1	1089 ± 300	926 ± 7
EPDM–7.5ZDA	142 ± 0.1	142 ± 0.6	100 ± 0.1	100 ± 0.1	734 ± 225	603 ± 2
EPDM–15ZDA	138 ± 0.1	137 ± 0.1	99 ± 0.1	99 ± 0.2	786 ± 176	564 ± 2
EPDM–TRIM	162 ± 0.6	164 ± 0.3	100 ± 0.1	100 ± 0.1	766 ± 4	763 ± 4
Belt–Ref	137 ± 0.4	135 ± 0.2	100 ± 0.1	100 ± 0.1	703 ± 15	674 ± 3

summarizes the obtained values for EPDM vulcanizates reinforced with coagents. The results are in good agreement with the MDR data. The reduction of 20 phr CB results for all samples in an increase of swelling degree, indicating a lower cross-link density. Interestingly, 15 phr ZDMA show increased swelling, which leads to the assumption of an inferior curing behavior. Similarly, the gel content of the rubber is reduced for EPDM-15ZDMA, revealing an inferior cure compared with the other vulcanizates.

These assumptions are backed by the calculated cross-link density M_c. The given results refer to the average molecular weight between cross-links, meaning a lower M_c refers to higher cross-link density [29]. Unsurprisingly, Belt–Ref exhibits the lowest cross-link density due to the reinforcing effect of 50 phr CB. The other samples exhibit reduced M_c, which reflects the loss of 20 phr CB. However, it is most prominent for EPDM–15ZDMA, indicating a suppressed network formation.

The impact of the coagents on the mechanical and dynamical properties was evaluated and is depicted in Figure 2b–d and summarized in

Table 5. Mechanical properties of reinforced EPDM compounds under investigation.

Sample	TS (MPa)	εb (%)	σ100 (MPa)	σ300 (MPa)	Tear (N/mm)
EPDM_7.5ZDMA-0 h	25.4 ± 1.1	617 ± 16.7	2.6 ± 0.1	8.5 ± 0.2	16 ± 1.0
EPDM–7.5ZDMA-24 h	24.5 ± 0.3	587 ± 19	2.7 ± 0.1	8.8 ± 0.1	17 ± 0.4
EPDM–15ZDMA-0 h	16.0 ± 0.3	617 ± 16	2.4 ± 0.1	6.4 ± 0.1	26 ± 0.2
EPDM–15ZDMA-24 h	15.8 ± 0.4	619 ± 11	2.5 ± 0.1	6.5 ± 0.2	28 ± 0.1
EPDM–7.5ZDA-0 h	22.0 ± 0.8	492 ± 8	3.0 ± 0.1	10.7 ± 0.2	14 ± 0.8
EPDM–7.5ZDA-24 h	22.4 ± 1.2	476 ± 9	3.2 ± 0.1	11.5 ± 0.2	15 ± 1.2
EPDM–15ZDA-0 h	23.2 ± 1.2	560 ± 15	3.2 ± 0.1	10.2 ± 0.3	19 ± 1.5
EPDM–15ZDA-24 h	21.6 ± 1.2	535 ± 13	3.2 ± 0.1	10.2 ± 0.3	18 ± 0.3
EPDM–TRIM-0 h	16.3 ± 0.7	452 ± 9	2.4 ± 0.1	8.3 ± 0.1	12 ± 0.3
EPDM–TRIM-24 h	16.6 ± 0.5	459 ± 9	2.5 ± 0.1	8.5 ± 0.2	10 ± 0.7
PF–ZDA-0 h	12.9 ± 0.7	742 ± 14	3.1 ± 0.1	6.4 ± 0.1	22 ± 0.4
PF–ZDA-24 h	11.7 ± 0.9	715 ± 24	2.8 ± 0.1	6.5 ± 0.3	22 ± 0.2
PF–ZDMA-0 h	11.4 ± 1.1	688 ± 34	3.0 ± 0.1	6.6 ± 0.3	29 ± 1.4
PF–ZDMA-24 h	12.4 ± 0.7	658 ± 34	3.3 ± 0.1	7.7 ± 0.3	28 ± 0.5
PF–TRIM-0 h	10.0 ± 0.3	509 ± 33	3.1 ± 0.2	7.2 ± 0.3	32 ± 0.3
PF–TRIM-24 h	11.3 ± 0.3	442 ± 17	3.4 ± 0.1	8.3 ± 0.2	20 ± 1.0
PF–7PO-0 h	10.0 ± 0.7	880 ± 16	2.0 ± 0.1	4.2 ± 0.2	28 ± 1.2
PF–7PO-24 h	9.8 ± 0.5	672 ± 23	2.2 ± 0.2	5.4 ± 0.5	26 ± 0.8
PF–5PO-0 h	8.4 ± 0.3	966 ± 40	1.9 ± 0.2	3.5 ± 0.2	32 ± 0.6
PF–5PO-24 h	9.3 ± 0.3	768 ± 95	2.3 ± 0.1	5.1 ± 0.2	30 ± 0.5
Belt–Ref-0 h	18.0 ± 1.1	386 ± 33	3.2 ± 0.1	12.6 ± 0.2	15 ± 0.3
Belt–Ref-24 h	17.4 ± 0.7	386 ± 16	3.2 ± 0.1	12.2 ± 0.2	14 ± 0.5

and

Table 6. Tan δ, Shore A haradness, and compression set properties of reinforced EPDM vulcanizates.

Sample	Tan δ (-)	Hardness (Shore A)	CS@RT (%)	CS@100 °C (%)
EPDM–7.5ZDMA-0 h	-	58 ± 1.2	17 ± 0.5	32 ± 1.0
EPDM–7.5ZDMA-24 h	0.63	60 ± 0.6	-	-
EPDM–15ZDMA-0 h	-	62 ± 1.5	24 ± 0.7	47 ± 0.6
EPDM–15ZDMA-24 h	0.60	62 ± 0.6	-	-
EPDM–7.5ZDA-0 h	-	60 ± 1.2	26 ± 0.8	37 ± 1.2
EPDM–7.5ZDA-24 h	0.61	58 ± 0.6	-	-
EPDM–15ZDA-0 h	-	60 ± 0.6	25 ± 0.9	52 ± 1.9
EPDM–15ZDA-24 h	0.57	59 ± 1.2	-	-
EPDM–TRIM-0 h	-	56 ± 1.0	24 ± 0.6	26 ± 0.9
EPDM–TRIM-24 h	0.63	57 ± 0.6	-	-
PF–ZDA-0 h	-	72 ± 0.6	34 ± 0.6	58 ± 1.8
PF–ZDA-24 h	0.50	72 ± 0.6	-	-
PF–ZDMA-0 h	-	71 ± 0.6	33 ± 0.6	54 ± 1.4
PF–ZDMA-24 h	0.52	71 ± 1.2	-	-
PF–TRIM-0 h	-	65 ± 0.6	19 ± 0.9	32 ± 2.1
PF–TRIM-24 h	0.49	66 ± 0.5	-	-
PF–7PO-0 h	-	63 ± 0.6	25 ± 0.4	45 ± 1.2
PF–7PO-24 h	0.48	68 ± 1.2	-	-
PF–5PO-0 h	-	63 ± 1.0	22 ± 0.6	37 ± 1.7
PF–5PO-24 h	0.50	68 ± 1.2	-	-
Belt–Ref-0 h	-	63 ± 0.6	14 ± 0.3	16 ± 0.4
Belt–Ref-24 h	0.54	62 ± 0.6	-	-

. A stiff material behavior characterizes the stress–strain of Belt–Ref due to the presence of 50 phr CB. The reduction to 30 phr is reflected by lower maximum stress but enhanced elongation at break for EPDM–TRIM. Due to the present ionic and covalent bonds, the molecular chains of ZD(M)A-reinforced compounds can orient along the direction of the external force. Thus, higher elongation at break and improved stress at break than EPDM–TRIM is observed. A higher ZD(M)A content leads to a deterioration of the tensile properties, which is only minor for EPDM–15ZDA but significant in EPDM–15ZDMA. This behavior is also explicable with the low cross-link density of EPDM–15ZDMA. However, the effect of post-curing is negligible, indicating good thermal stability for 24 h at 150 °C.

Tear strength is not significantly affected by the reduction of CB. However, a high ZD(M)A content increases tear strength, attributed to the agglomeration of ionic domains at higher concentrations. Such agglomerations require more energy for the crack to propagate [30]. The loss of CB leads to a minor deduction in Shore A hardness, which does not seem to be influenced to a great extent by the actual coagent type.

A compression set (CS) is a good benchmark to evaluate the elastic recovery of compounds. Upon adding coagents, the CS values at room temperature slightly increase due to the introduction of rigid particles. At elevated temperatures, the CS further increases attributed to the physical and chemical changes in the sample matrix. Whereas the change for EPDM–TRIM is negligible, the ionic coagents significantly affect the elastic recovery at 100 °C. This phenomenon is attributed to the dissociation of the Zn²⁺ bonds at elevated temperatures [31].

The dynamic properties show a superior E’ loss for samples with reduced CB content. This finding is related to a lower cross-link density of the samples with only 30 phr CB. Interestingly, all samples show a stable modulus with increasing temperature, except EPDM–15ZDMA. Peak tan δ is slightly shifted to higher values for reduced CB amount. However, the respective coagents hardly influence the peak height, yielding similar values for all compounds. In addition, the glass transition temperature (T_g) is in a similar range for all compounds.

The fracture surface of the tear specimen was analyzed via SEM and is depicted in the supplementary information (Figure S1). All samples are characterized by a rough surface with finely dispersed particles.

The obtained results highlight the power of ionic coagents to replace 20 phr of CB. Especially 7.5 phr ZDA showed promising mechanical properties by simultaneously enhancing the processing parameters. However, high concentrations of ZD(M)A yield deteriorated properties, making them unsuitable for further investigation.

3.2. Rheological Characterization and Curing Behavior of Hybrid Reinforced EPDM Compounds

Resins are suitable for introducing reinforcing, rigid particles in a soft rubber matrix. PF has already been employed as a reinforcing agent for EPDM, but a distinct drawback is the scavenging effect of PF on peroxide-cured vulcanizates. In the previous section, it was demonstrated that ionic coagents are a suitable way to introduce reinforcement into the EPDM matrix. However, they also positively affect the curing efficiency of peroxide-cured rubbers, making them attractive for hybrid filler systems. In the following, the hybrid reinforcement (ionic and nonionic coagents in combination with PF) is compared to Belt–Ref and compounds only reinforced with the PF resin.

The obtained curing characteristics for hybrid reinforced compounds are depicted in Figure 3a and are summarized in Table 3. The rheogram of Belt–Ref is characterized by a plateau, whereas solely PF reinforced compounds PF–5PO and PF–7PO exhibit a marching modulus. In addition, M_H decreases in the presence of PF, which can be only partly compensated by increasing the PO content from 5 to 7 phr. The results confirm the radical scavenging effect of PF on peroxide cure and give rise to incomplete curing of the network. However, the cure rate is significantly improved in the presence of ionic coagents, which yields a plateau of the curve and reduces t₉₀.

M_H can be further increased by incorporating TRIM as a coagent, which comprises a higher functionality than ZD(M)A. Although the compound PF–TRIM benefits from a high M_H, a higher concentration of (meth)acrylate moieties is also known to increase the cure time, which is evident for the prolonged t₉₀ in PF–TRIM [20]. Again, the scorch time is reduced by the introduction of coagents to the base matrix. However, PF seems to act as a retarder which results in a higher scorch safety for the hybrid compounds. Therefore, a longer time frame for processability is given. A further distinctive parameter is ΔM, which is related to the cross-link density of the compounds. The introduction of PF results in a lower cross-link density compared with Belt–Ref. However, the introduction of the coagents positively impacts the overall cross-link density of EPDM–PF compounds, following the trend of PF–TRIM > PF–ZDA > PF–ZDMA.

M_L is decreased significantly upon the addition of PF to the matrix. However, the type of coagent seems negligible for M_L, which appears to be mainly governed by the type of resin. Resins are already reported as suitable plasticizing agents and improve the general flow behavior of the compounds. Low values for M_L and Mooney viscosity are therefore seen as beneficial for the processability of rubber compounds [32]. Especially PF–ZDA and PF–TRIM exhibit advantageous processing performance compared with Belt–Ref.

For a better understanding of the overall network composition, swelling measurements were performed. Further, the swelling degree is also a suitable method to determine the influence of the post-curing process on the EPDM/PF compounds. The obtained values for hybrid-reinforced EPDM vulcanizates are summarized in

Table 7. Swelling degree of hybrid-reinforced EPDM compounds prior to and after postcuring for 24 h at 150 °C.

Sample	Swelling Degree (%)		Gel Content (%)		Mc (g/mol)
	0 h	24 h	0 h	24 h	0 h	24 h
PF–ZDA	154 ± 0.8	142 ± 2.6	97 ± 0.1	99 ± 0.1	953 ± 4	821 ± 22
PF–ZDMA	166 ± 1.5	148 ± 0.4	97 ± 0.1	99 ± 0.1	1075 ± 16	872 ± 2
PF–TRIM	147 ± 5.2	141 ± 0.6	97 ± 0.2	99 ± 0.1	896 ± 15	823 ± 4
PF–7PO	192 ± 3.4	180 ± 1.1	96 ± 0.3	99 ± 0.2	1433 ± 48	1230 ± 11
PF–5PO	226 ± 0.8	206 ± 0.2	93 ± 0.3	97 ± 0.2	1944 ± 15	1571 ± 7
Belt–Ref	137 ± 0.4	135 ± 0.2	100 ± 0.1	100 ± 0.1	703 ± 15	674 ± 3

.

In general, highly cross-linked networks show lower swelling due to the tightly locked interface and high solvent resistance [33,34]. As expected, EPDM–PF compounds without coagent exhibit the highest swelling degree due to their overall lower cross-link density. In addition, the swelling degree declines in the order of PF–ZDMA > PF–ZDA > PF–TRIM, which indicates the highest restriction of swelling for PF–TRIM. PF–ZDA and PF–TRIM are even in a similar range as Belt–Ref, indicating a suitable method for replacing 20 phr CB. Postcuring of the samples yields a further reduction of the swelling for all EPDM–PF compounds. Further, the gel content is reduced for vulcanizates without coagents. This is attributed to the scavenging effect of PF on the rubber cross-linking and therefore reduced curing efficiency. The curing process continues during the post-curing step, resulting in enhanced gel content for all samples.

As expected, the calculated cross-link density is lowest for solely PF-reinforced vulcanizates. The addition of coagents results in a significant enhancement of the cross-link density. However, similar to the previous section, ZDMA results in low cross-link density for the nonpostcured samples. Postcuring significantly impacts the PF-reinforced compounds, which indicates further curing of the novolac resin.

These results demonstrate the positive influence of coagents on the state of cure and are in good agreement with the MDR analysis. However, postcuring is assumed to yield a further increase of cross-link density for all compounds. This increase can be attributed to aging-induced chain scission, which contributes to the formation of new cross-links of the EPDM matrix or further curing of the PF resin [35,36].

3.3. Mechanical Properties of EPDM–PF Compounds

For the characterization of the reinforcing efficacy, several mechanical tests were conducted, which are summarized in Table 5 and Table 6. As visualized in Figure 3a, PF and the type of coagent significantly influence the stress-strain behavior of the compounds. Belt–Ref is characterized by fast growth in modulus, resulting in high maximum stress. By replacing 20 phr CB with PF, the stress-strain curves significantly change towards lower modulus and maximum stress but increased elongation at break (ε_b). Upon adding the (meth)acrylate-based coagents, a stiffening of the compounds can be observed. This is explicable with the increase in cross-link points due to the introduction of the coagents. The mobility of the chain is restricted, yielding a decrease in elongation at break (ε_b). Furthermore, the increased cross-link density is also reflected by the enhanced modulus at 300% strain (σ₃₀₀) for all hybrid filler systems.

While the stiffening of PF–TRIM is pronounced, the other coagents yield stress-strain curves, which are characterized by enhanced elongation at break. The increased stiffening of PF–TRIM could be attributed to the trifunctionality of the coagent, which results in a tight network structure. Unlike TRIM, ZD(M)A can additionally introduce ionic bonds, contributing to a synergistic system of covalent and physical cross-links within the rubber matrix. Especially in saturated rubbers, the physical absorption dominates filler–rubber interactions, resulting in divergent stress-strain curves compared with PF–TRIM [37,38]. Therefore, ZD(M)A seems suitable for further toughening of EPDM–PF compounds. In addition, postcuring leads to further stiffening of all compounds, which is most pronounced for sole PF-reinforced compounds PF–7PO and PF–5PO. Such change indicates the proceeding curing reaction of PF.

The tear strength is significantly enhanced upon the introduction of phenol novolac resin compared with Belt–Ref. Due to the presence of in situ formed rigid particles in EPDM–PF compounds, the crack needs to propagate around them and requires higher energy. In the absence of PF, the cracks can propagate more easily, yielding inferior tear properties.

Regarding Shore A hardness, the loss of CB can be compensated by adding PF. With the further introduction of ionic coagents, the Shore A hardness is even increased for the hybrid filler compounds. In general, the Shore A hardness is linked to the cross-link density of the rubber compounds, which would suggest the highest hardness for Belt–Ref. However, the rigid resin particles significantly enhance Shore A hardness. Further, Costin et al. reported that ZD(M)A is prone to result in high hardness values [21]. However, the compression set slightly deteriorates upon the introduction of a hybrid filler system. The addition of ionic coagents yields a reduced elastic recovery, which is more pronounced at 100 °C. As previously mentioned, the ionic clusters dissociate at elevated temperatures resulting in inferior CS values [31]. Contrary to EPDM–TRIM, the sample PF–TRIM also exhibits a worsening of the CS at elevated temperatures. This finding is attributed to the further cross-linking of PF, resulting in the development of a tight and interlocked network. The formation of such a network in the compressed state reduces the overall elastic recovery of the compounds.

The dynamic mechanical properties were determined from 24 h postcured samples and are shown in Figure 3c,d. The temperature corresponding to the maximum of tan δ (Figure 3d) was taken as the T_g of the samples, which was in a similar range for all compounds under investigation. Above T_g, the storage modulus (E’) showed a significant decrease due to the enhanced mobility of the polymer chain. E’ loss is superior for compounds containing the phenol novolac resin, which highlights the reinforcing effect of PF. At 100 °C, E’ is highest for PF–TRIM-24 h due to the enhanced cross-link density of the respective compounds. This finding confirms again that hybrid filler systems are a possible route to tune rubber reinforcement.

The maximum height of the tanδ curve refers to the dampening effect of the material. After the incorporation of rigid PF particles, the peak is significantly decreased. Therefore, a higher dampening effect for Belt–Ref-24 h is indicated. However, there is only a minor difference in the peak height between the compounds with or without coagent. This suggests that the damping properties are mainly governed by the addition of PF and not the actual coagent.

3.4. Microscopic Analysis of Tear Fractures

The morphology of the fracture surfaces of the tear specimen was determined by SEM measurements and is displayed in Figure 4. Belt–Ref exhibits a relatively smooth fracture surface with finely dispersed particles under 1 µm attributed to CB. However, some larger fragments are visible, which have a size of around 6 µm. The introduction of PF results in the formation of small particles in the micron size, attributed to the cross-linking of PF. Contrary to Belt–Ref, the PF-reinforced vulcanizates are characterized by a rough and heterogeneous surface. However, the border between the EPDM and PF matrix is not well defined without coagents, and larger PF domains are formed. With the help of coagents, the distinction becomes more apparent, resulting in a defined heterogeneous system. The hybrid reinforced vulcanizates show well dispersed and evenly distributed novolac particles averaging around 1 µm or lower.

In general, the morphology of vulcanizates is a crucial parameter for its final mechanical properties. The introduction of PF creates rigid particles, which results in a deviation from the tear path. Such a deviation is attributed to a rough tear and, therefore, higher tear strength [39]. This assumption is confirmed by the obtained tear strength results, which are lowest for the non-PF-reinforced compounds.

3.5. Proposed Structure of Ionic Hybrid Systems

A schematic representation of the proposed network structure of ionic hybrid systems is given in Figure 5. The rubber (black) and the PF network (green) are generated simultaneously during vulcanization. This results in the formation of an interpenetrating network, where the two networks are interlaced with each other [10]. However, the addition of ionic coagents contributes to a multiple network structure governed by covalent and ionic cross-links. It has been reported that ZD(M)A is introduced in the rubber matrix in several ways, namely as (i) homo-polymer, (ii) grafted-polymer, or (iii) residual ZD(M)A monomer [40]. However, highly saturated rubbers have a low amount of grafted poly-ZD(M)A due to the low number of residual double bonds [41,42]. The applied EPDM matrix is highly saturated. Therefore, a high amount of poly-ZD(M)A is expected, which contributes to one side as reinforcing particles and introduces ionic cross-links. These ionic bonds are expected to contribute to the superior mechanical performance of the ionic hybrid systems [43]. Though this study focused on an EPDM grade development for industrial V- belts, this in situ filler loading and supportive ionic coagents seemingly have many other applications, especially in many elastomeric components, which require sophisticated material properties, for example, as sealing applications in high-pressure gas conditions. With these developments, materials can be developed, having stiffer material combinations with higher ultimate strength and elongation at break to obtain tougher materials, which are very useful for those specific applications to avoid permeation and rapid gas decompression [44].

4. Conclusions

The introduction of ionic coagents can partly compensate for the reduction of 20 phr CB. These novel formulations result in enhanced curing properties with superior processing performance, reducing M_L up to a third compared with Belt–Ref. However, PF resin was added to influence the processing behavior further. While EPDM–PF compounds suffer from a cure rate twice t₉₀ of Belt–Ref and low cross-link density, these deficiencies were overcome by introducing TRIM and ZD(M)A to generate a hybrid filler system. By adding TRIM, the mechanical properties were changed towards higher stiffening and hardening the material, reflected by a 55% increase in σ100 compared with PF–7PO. These findings were mainly attributed to the enhanced cross-link density in the presence of TRIM. However, the zinc-based coagents proved a further toughening of the compounds by introducing a synergistic reinforcing mechanism governed by ionic and covalent cross-links. Compared with Belt–Ref, PF–ZDMA and PF–ZDA exhibit a nearly twice-fold εB with similar σ₁₀₀ values. SEM measurements revealed the presence of finely dispersed PF particles for EPDM–PF compounds. However, the dispersion was significantly improved for the hybrid systems.