Article Alleviating Molecular-Scale Damages in Silica-Reinforced Compounds by a Self-Healing Modifier

Bashir Algaily 1,2, Wisut Kaewsakul 3,* , Siti Salina Sarkawi 4 and Ekwipoo Kalkornsurapranee 1,*

1 Science and Technology, Division of Physical Science, Faculty of Science, Prince of Songkla University, Hat Yai Campus, Songkhla 90110, Thailand; [email protected] 2 Department of Physics, Faculty of Science and Technology, Al-Neelain University, Khartoum 11111, Sudan 3 Technology and Engineering, Department of Mechanics of Solids, Surfaces and Systems, Faculty of Engineering Technology, University of Twente, P.O. Box 217, 7522 NB Enschede, The Netherlands 4 Malaysian , RRIM Research Station, Sg. Buloh, Selangor 47000, Malaysia; [email protected] * Correspondence: [email protected] (W.K.); [email protected] (E.K.); Tel.: +31-53-489-7117 (W.K.); +66-84-758-85069 (E.K.)

Abstract: The property retentions of silica-reinforced natural rubber vulcanizates with various con- tents of a self-healing modifier called EMZ, which is based on epoxidized natural rubber (ENR) modified with hydrolyzed maleic anhydride (HMA) as an ester crosslinking agent plus zinc acetate dihydrate (ZAD) as a transesterification catalyst, were investigated. To validate its self-healing efficiency, the molecular-scale damages were introduced to vulcanizates using a tensile stress–strain cyclic test following the Mullins effect concept. The processing characteristics, reinforcing indicators, and physicomechanical and viscoelastic properties of the compounds were evaluated to identify the influences of plausible interactions in the system. Overall results demonstrate that the property re- tentions are significantly enhanced with increasing EMZ content at elevated treatment temperatures, because the EMZ modifier potentially contributes to reversible linkages leading to the intermolecular  reparation of rubber network. Furthermore, a thermally annealing treatment of the damaged vul-  canizates at a high temperature, e.g., 120 ◦C, substantially enhances the property recovery degree, Citation: Algaily, B.; Kaewsakul, W.; most likely due to an impact of the transesterification reaction of the ester crosslinks adjacent to the Sarkawi, S.S.; Kalkornsurapranee, E. molecular damages. This reaction can enable bond interchanges of the ester crosslinks, resulting in Alleviating Molecular-Scale Damages the feasibly exchanged positions of the ester crosslinks between the broken rubber molecules and, in Silica-Reinforced Natural Rubber thus, achievable self-reparation of the damages. Compounds by a Self-Healing Modifier. Polymers 2021, 13, 39. Keywords: filler; intermolecular reparation; polymer failure; rubber crosslink; composite https://dx.doi.org/10.3390/ polym13010039

Received: 14 November 2020 1. Introduction Accepted: 17 December 2020 Published: 24 December 2020 One of the competing features of rubber goods is “durability”. Failures of rubbers originate due to excessive local stresses applied during their operations, leading to matrix Publisher’s Note: MDPI stays neu- breakages starting from a small molecular scale, called micro-damages [1]. Continued or tral with regard to jurisdictional claims repeated operations of the rubbers will result in the progressive failure of rubber molecular in published maps and institutional networks growing to a larger failure scale, called macro-damages [2]. This leads to the affiliations. permanent breakdown of materials, whereby they cannot function to fulfill the desired performance any longer and, thus, need to be disposed of. Earlier disposal or shorter lifespan of a product would contribute to an undesired waste accumulation that in turn requires efficient management to eliminate the wastes [3]. Essential resources, e.g., invest- Copyright: © 2020 by the authors. Li- ment costs, labor, and time, will be consumed to enable the management. In this regard, censee MDPI, Basel, Switzerland. This the improved durability of a product should beneficially reduce the consumption of the article is an open access article distributed aforementioned resources. under the terms and conditions of the Creative Commons Attribution (CC BY) A final rubber product has a permanent form stability due to a three-dimensional license (https://creativecommons.org/ network chemically tailored to rubber chains. This process is commonly known as vul- licenses/by/4.0/). canization or , in which the rubber molecular chains are crosslinked to one another,

Polymers 2021, 13, 39. https://dx.doi.org/10.3390/polym13010039 https://www.mdpi.com/journal/polymers Polymers 2021, 13, 39 2 of 22

giving a macromolecular network structure in the rubber matrix [4]. All conventionally vulcanized rubbers, e.g., treads, rubber mounts, and gloves, are classified as ther- mosetting polymers that cannot be reprocessed or recycled, unlike . The nonrecyclability of vulcanized rubbers is mainly due to this chemical three-dimensional network existent in the materials [5]. There have been attempts to recycle rubber vulcan- izates using the devulcanization process. To date, successes toward a certain extent on this devulcanizing technology have been reported but cannot yet be implemented in a full-scale process [6]. A permanent network can be generated through curatives. Conventional crosslinking agents include sulfur and peroxide, and they are the most commonly used for rubber vulcanization [4]; for instance, are generally cured with sulfur. Once the permanent rubber network is partially damaged, it is unable to be completely recovered or reformed. Furthermore, the failure of the network will progressively occur as a function of use cycles leading to a continually lowered performance and, therefore, reduced durability of the rubber [7]. To make a rubber product endurable toward operating failures, an initiative to pre- vent or even auto-internally repair the molecular damages would be one of the promising solutions. External operating parameters are basically constant for each individual appli- cation, e.g., a set of tires supports a vehicle under a certain range of dynamic mechanical conditions. While the crosslinks and intermolecular interactions in a rubber vulcanizate are rather complex and variable, micro- and macro-damages can possibly be originated at the main chains, crosslinking bonds, and strong and weak interactions [8]. If the interactions or molecular chains are subject to an excess tension force or degrading conditions [9], these elements will inevitably be torn apart, resulting in molecular-scale damages. Therefore, to allay an ongoing process of network failure, self-reparation of the network via a chemical and/or a physical approach would assist in bringing the network back to an acceptable level. Consequently, this could lead to high-performance retention of the rubbers after operating cycles. As stated, enabling vulcanized rubber self-healing seems to be a good route for reduc- ing the micro- and macro-damages. Innovation in multifunctional materials has enabled scientists to increase their efforts to develop such elastomeric materials possessing sig- nificant self-healing properties with acceptable technical performances [10]. Developing self-healing is an interesting issue; however, research has still been on a prelimi- nary scale. Once it becomes practical, it is expected to have a high impact on a vast range of applications. For tires, this concept has the potential to prolong their service life and may reduce accidents caused by the explosion of tires while running on the road [7]. Truck tires are basically based on natural rubber (NR) as the major rubber. Previous works recently reported that the introduction of ester crosslinks to epoxidized natural rubber (ENR) could make the final vulcanizates reprocessable due to the unique characteristic of a dynamic network based on ester bonds [11]. This ester crosslink is able to be thermochemically exchanged at elevated temperatures when there is a transesterification catalyst present in the system [12]. In other words, this ester crosslink can alter its bonding position at a high temperature through the transesterification reaction [13]. Prior evidence [14] clearly demon- strated that the rubber pellets of an ester-crosslinked ENR can be remolded giving again a cohesive rubber sheet. This is because the ester bonds at the interface of the vulcanizate pellets can exchange their bonding sites, making chemical crosslinks among rubber chains at the pellet interfaces. It is worth determining if the adaptability of the ester crosslinks enables the recombination of separate molecular networks. Therefore, it would be highly promising if this thermochemically exchangeable ester crosslinking system can contribute to the self-reparation of molecular-scale damages in a rubber matrix after operations. The present study aims at validating the healing ability of a self-healing modifier based on dicarboxylic acid-modified ENR when used in silica-reinforced natural rubber compounds. The modifier was recently evolved to enable the reprocessability of ENR vul- canizates, as reported elsewhere [15]. The silica NR reference was formulated on the basis of an NR truck tread compound [16]. The self-healing modifier was added to the reference Polymers 2021, 13, 39 3 of 22

compound with varied loading levels. There were two series of samples depending on how the modifier was added, i.e., extra addition or blending. Mixing behavior and compound properties were monitored. Molecular-scale damages were introduced to vulcanized sam- ples using a tensile stress–strain cyclic test, the commonly known “Mullins effect” [17]. The results of immaculate samples were compared with their counterparts which underwent the first cycle test. The capability for self-reparation of the broken bonds or interactions after molecular-scale damages, as well as the correlation between the viscoelastic properties and property retention of vulcanizates, was the focus of this investigation.

2. Materials and Methods 2.1. Materials Natural rubber grade STR5L was used (Chalong Latex Industry Co., Songkhla, Thai- land). Compounding ingredients included conventional precipitated silica grade Ultrasil VN3 (Evonik United Silica (Siam) Ltd., Rayong, Thailand), bis-(3-triethoxysilylpropyl) tetrasulfide or TESPT (Si-69 tradename, Evonik, Essen, Germany), white process oil (PTT Public Co. Ltd., Bangkok, Thailand), diphenyl guanidine or DPG, n-(1,3-dimethylbutyl)- n-phenyl-p-phenylenediamine or 6PPD (both from Vessel Chemical Co. Ltd., Bangkok, Thailand), zinc oxide or ZnO (Thai-Lysaght Co. Ltd., Phra Nakhon Si Ayutthaya, Thailand), stearic acid (Imperial Industrial Chemicals Co. Ltd. Bangkok, Thailand), n-cyclohexyl-2- benzothiazolesulfenamide or CBS, 2,2,4-trimethyl-1,2-dihydroquinoline or TMQ (both from Lanxess Co. Ltd., Bangkok, Thailand), and sulfur (Siam Chemicals Co. Ltd., Samudprakarn, Thailand). The ingredients for the self-healing modifier based on epoxidized natural rubber (ENR) modified with a dicarboxylic acid and zinc acetate dihydrate included the ENR containing 50 mol.% epoxy groups or ENR-50 (Muang Mai Guthrie PCL, Phuket, Thailand), kaolin clay (Siam Chemicals Co. Ltd., Samudprakarn, Thailand), maleic anhydride or MA (Sigma-Aldrich, Shanghai, China), 1,2-dimethylimidazole or DMI (Alfa Aesar, Kandel, Germany), and zinc acetate dihydrate or ZAD (analytical grade, Ajax Finechem Pty. Ltd., New South Wales, Australia). They were used without further purification.

2.2. Sample Preparation 2.2.1. Preparation of the Self-Healing ENR Modifier ENR-50, HMA (hydrolyzed maleic anhydride), DMI, and ZAD at 100, 3, 5, and 5 phr (parts per hundred of rubber by weight), respectively, were mixed in an internal mixer for 15 min with a rotor speed of 70 rpm at a starting mixer temperature of 40 ◦C, following a protocol described in a previous work as concisely illustrated in Scheme1[ 14,15]. The obtainable rubber masterbatch was named “EMZ” (epoxidized natural rubber modified with hydrolyzed maleic anhydride plus zinc acetate dihydrate) as an acronym in this context. Regarding the function of the ingredients used, DMI was employed as accelerator for the esterification reaction between the carboxylic acids of HMA and epoxide groups of ENR. It was manually premixed with kaolin clay, an inert filler, prior to introduction into the mixer to facilitate the incorporation of the DMI into the compound. ZAD functions as a transesterification catalyst with efficiency to promote the transesterification reaction at elevated temperatures, shuffling the formed ester bonds and enabling network rearrange- ment [12,13]. This mechanism allows the crumbled rubber to again reform into a coherent vulcanized sheet [14,15]. Hence, this material is considered a self-healing modifier, as it would potentially promote the intermolecular self-reparation of molecular damages in vulcanizates. The material structure, cure characteristics, and other technical properties of this modifier were recently reported in [14,15]. Polymers 2021, 13, x FOR PEER REVIEW 4 of 24

Polymers 2021, 13, 39 4 of 22 vulcanizates. The material structure, cure characteristics, and other technical properties of this modifier were recently reported in [14,15].

Scheme 1. Preparation of the self-healing modifier used in this study and the possible ester and Scheme 1. Preparation of the self-healing modifier used in this study and the possible ester and ionic crosslink structures generated through the hydrolyzed maleic anhydride after the modifier ionic crosslink structures generated through the hydrolyzed maleic anhydride after the modifier underwent a vulcanization process [14,15]. underwent a vulcanization process [14,15].

2.2.2.2.2.2. Compound Compound Preparation Preparation TheThe formulation formulation of of silica-reinforced silica-reinforced natural natural rubber rubber compounds compounds used used as a rubberas a rubber ma- matrixtrix is is shown shown inin TableTable1. 1 Two. Two sets sets of the of compounds the compounds were prepared were prepared with two with different two different EMZEMZ additions: additions: Set Set A, extraA, extra addition addition of EMZ of EMZ at 0 phr, at 0 5 phr, phr, 105 phr, phr, and10 phr, 15 phr and to 15 the phr to the compounds coded as the reference, E-05, E-10, and E-15, respectively; Set B, blending of compounds coded as the reference, E-05, E-10, and E-15, respectively; Set B, blending of NR/EMZ (i.e., the total amount of rubber was 100 phr) at the ratios of 100/0, 95/5, 90/10, / ( . . ) / / / andNR 85/15,EMZ giving i e , the the compoundstotal amount coded of rubber as the reference, was 100 B-05,phr B-10,at the and ratios B15, respectively.of 100 0, 95 5, 90 10, Theand reason 85/15, for giving designing the compounds these two different coded seriesas the for reference, the present B-05, investigation B-10, and wasB15, thatrespectively. theThe EMZ reason modifier for designing is based on these ENR. two Thus, different it can be series considered for the as eitherpresent a compatibilizerinvestigation was that orthe a secondary EMZ modifier elastomer; is based a compatibilizer on ENR. Thus, is commonly it can be added considered as an extra as either ingredient a compatibilizer to compounds,or a secondary while elastomer; rubber is added a compatibilizer by blending with is commonly the base rubber added of compounds.as an extra ingredient to compounds, while rubber is added by blending with the base rubber of compounds. Mixing of the compounds was performed using an internal mixer (Charoen Tut Co. Ltd., Samut Prakan, Thailand) with a mixing chamber of 300 cm3. The mixer was operated at a rotor speed of 60 rpm and a fill factor of 70%. An initial temperature of mixing was adjusted at 100 °C. The dump temperature of all compounds was controlled to be in the range of 140–150 °C, such that an optimum of the hydrophobation or silanization reaction between silica and silane coupling agent was obtainable. To start the mixing process, NR was first masticated on a two-roll mill for 2 min to reduce its viscosity. Then, it was loaded into the internal mixer along with the EMZ modifier and mixed for 2 min. After that, half

Polymers 2021, 13, x FOR PEER REVIEW 5 of 24

of the silica and silane coupling agent was added and mixed for 5 min prior to incorpora- tion of the second half of the silica and silane together with white process oil. Other ingre- dients, i.e., ZnO, stearic acid, TMQ, 6PPD, and half of DPG were added, mixed for 3 min, and finally discharged. The obtainable compound was then sheeted out on a two-roll mill and kept overnight before adding CBS, the other half of the remaining DPG, and sulfur Polymers 2021, 13, 39 on a two-roll mill. The final compounds were sheeted and kept overnight prior to5 oftheir 22 property characterizations.

Table 1. Compound formulations used in this study. Table 1. Compound formulations used in this study. Dosage (Parts per Hundred Rubber, phr) Dosage (Parts per Hundred Rubber, phr) Ingredients Set A: Extra Addition of EMZ a Set B: Blending of EMZ b Ref. a b Ingredients - Set A: Extra- Addition of- EMZ - Set- B: Blending of EMZ- Ref. E 05 E 10 E 15 B 05 B 10 B 15 NR c (STR 5L d) 100 100E-05 100 E-10100 E-1596 B-0592 B-10 88 B-15 NR cEMZ(STR e 5L d) -100 5 10010 10015 1005 9610 92 15 88 EMZSilicae 55 -55 555 1055 1555 555 10 55 15 Silane couplingSilica agent (TESPT f) 5 55 55 55 55 55 55 55

Silane coupling agent (TESPT f) 5 Processing oil 8 Processing oil 8 Zinc oxide 3 Zinc oxide 3

StearicStearic acid 1 1 6PPD6PPD gg 1 1 The quantities of these compositions are the same h as those added to the reference (Ref.) compound. TMQTMQ h 1 1 DPG i 1 DPG i 1 CBS j 1.5 j SulfurCBS 1.5 1.5 . a The total rubberSulfur contents of these compounds1 5 were higher than 100 phr due to the particular extra amount of EMZ masterbatch added to the compounds; b The total rubbera The total contents rubber of these contents compounds of these were compounds at 100 phr; wec naturalre higher rubber; than d100standard phr due Thai to rubberthe particular grade 5L; e epoxidized natural rubberextra modified amount with hydrolyzedof EMZ masterba maleic anhydridetch added plus to the zinc compounds; acetate dihydrate; b Thef totalbis-(3-triethtoxysilylpropyl) rubber contents of these tetrasulfide; g n-(1,3-dimethylbutyl)-compoundsn-phenyl- werep-phenylenediamine; at 100 phr; c naturalh 2,2,4-trimethyl-1,2-dihydroquinoline; rubber; d standard Thai rubber igradediphenyl 5L; e guanidine; epoxidizedj n -natu- cyclohexyl-2-benzothiazolesulfenamide.ral rubber modified with hydrolyzed maleic anhydride plus zinc acetate dihydrate; f bis-(3-tri- ethtoxysilylpropyl) tetrasulfide; g n-(1,3-dimethylbutyl)-n-phenyl-p-phenylenediamine; h 2,2,4- trimethylMixing-1,2 of-dihydroquinoline; the compounds wasi diphenyl performed guanidine; using j n an-cyclohexyl internal mixer-2-benzothiazolesulfenamide. (Charoen Tut Co. Ltd., Samut Prakan, Thailand) with a mixing chamber of 300 cm3. The mixer was operated at2.3. a rotorCharacterizations speed of 60 rpm and a fill factor of 70%. An initial temperature of mixing was adjusted at 100 ◦C. The dump temperature of all compounds was controlled to be in the 2.3.1. Cure Characteristics and Mooney Viscosity range of 140–150 ◦C, such that an optimum of the hydrophobation or silanization reaction betweenThe silica cure andcharacteristics silane coupling of the agent compounds was obtainable. were determined To start theusing mixing a moving process, die NRrhe- wasometer first or masticated MDR (MDR2000, on a two-roll Alpha mill Technologies, for 2 min to reduce Hudson, its viscosity.OH, USA Then,) with ita wascuring loaded tem- intoperature the internal of 150 mixer °C for along 30 min with. The the optimum EMZ modifier cure andtime mixed (Tc90) forand 2 min.rheometer After that,cure halftorque of thewere silica reported and silane and couplingused for agentpress- wasvulcanizing added and the mixed compounds for 5 min into prior 1.5 tomm incorporation thick sheets. ofThe the investigated second half ofcompounds the silica and were silane tested together for their with Mooney white processviscosity oil. using Other a Mooney ingredients, vis- i.e.,cometer ZnO, ( stearicMV2000, acid, Alpha TMQ, Technologies, 6PPD, and Alpha half of Technologies, DPG were added, Ohio, mixed USA) forat 100 3 min, °C with and a finallylarge rotor discharged. according The to obtainableASTM D1646 compound. The value was of ML then (1 sheeted + 4) 100 out °C onwas a reported two-roll.mill and kept overnight before adding CBS, the other half of the remaining DPG, and sulfur on2.3.2. a two-roll Apparent mill. Crosslink The final Density compounds were sheeted and kept overnight prior to their propertyMeasurements characterizations. of the apparent crosslink density of vulcanizates were carried out via a swelling method with the vulcanizates in toluene as solvent according to ASTM D471- 2.3.12a Characterizations. The cured test specimens of 15 × 15 × 1.5 mm3 were immersed in 30 mL of toluene for 2.3.1.7 days Cure at room Characteristics temperature and. The Mooney samples Viscosity were removed from the toluene, and then the weightThe of cure the characteristicsswollen samples of thewas compoundsmeasured. The were specimens determined were using dried a movingin a vacuum die rheometeroven at 75 or °C MDR for 24 (MDR2000, h, and the Alphadry weig Technologies,ht of the specimens Hudson, was OH, finally USA) measured with a curing. The ◦ temperatureapparent crosslink of 150 Cdensity for 30 ( min.ν) was The calculated optimum using cure time the (Tmodifiedc90) and Flory rheometer–Rehner cure equation torque wereas shown reported in Equation and used ( for1) [18] press-vulcanizing. the compounds into 1.5 mm thick sheets. The investigated compounds were tested for their Mooney viscosity using a Mooney viscometer (MV2000, Alpha Technologies, Alpha Technologies, Ohio, USA) at 100 ◦C with a large rotor according to ASTM D1646. The value of ML (1 + 4) 100 ◦C was reported.

2.3.2. Apparent Crosslink Density Measurements of the apparent crosslink density of vulcanizates were carried out via a swelling method with the vulcanizates in toluene as solvent according to ASTM D471-12a. The cured test specimens of 15 × 15 × 1.5 mm3 were immersed in 30 mL of toluene for 7 days at room temperature. The samples were removed from the toluene, and then the weight of the swollen samples was measured. The specimens were dried in a vacuum oven Polymers 2021, 13, 39 6 of 22

at 75 ◦C for 24 h, and the dry weight of the specimens was finally measured. The apparent crosslink density (ν) was calculated using the modified Flory–Rehner equation as shown in Equation (1) [18]. 2 −(ln(1 − Vr ) + Vr + χV ) ν = 0 0 r0 (1)  1  3 Vr0 2Vs Vr0 − 2

where v is the molar volume of the toluene (i.e., 106.9 cm3/mole), χ is the Huggins interaction constant (i.e., 0.38 for silica-filled NR) [19], and Vr0 is the volume fraction of the swollen rubber, which could be computed using Equations (2) and (3) [20].

V θ r0 = 1 − m (2) Vr f 1 − θ

 1  3 m = 3c 1 − Vr0 + Vr0 − 1 (3)

where c is the parameter of silica–rubber interaction (i.e., 1.17) [19], θ is the volume fraction of silica, and Vrf is the volume fraction of rubber in the swollen filled rubber, which can be calculated using Equation (4). w2 = ρr Vr f ( − ) , (4) w2 + w1 w2 ρr ρs

where w1 is the weight of the swollen sample, w2 is the weight of the sample after drying, 3 and ρs and ρr are the densities of the solvent used (i.e., toluene, 0.886 g/cm ) and of the base rubber (i.e., 0.92 g/cm3 for NR), respectively [19].

2.3.3. Bound Rubber Content as Indicative for –Rubber Interaction Bound rubber content was measured using an uncured sample without curatives. The measurement was carried out under an ammonia atmosphere in order to cleave the physical linkages in the samples, giving only the amount of chemically bound rubber content, i.e., filler–rubber interactions, as well as the lightly crosslinked rubber network [21]. The procedure started with preparing an uncured sample of about 0.2 g. Then, the sample was cut into small pieces and placed in a filter bag made from a metal mesh, i.e., 400 mesh. A bag containing the sample was immersed into 20 mL of toluene at room temperature for 72 h, and the solvent was renewed every day. Afterward, the sample was removed from the toluene and dried for 24 h at 105 ◦C. Then, the sample was again immersed in 20 mL of toluene at room temperature for 72 h and placed in an ammonia atmosphere; the solvent was renewed every day. Finally, the sample was dried for 24 h in a vacuum hot-air oven at 105 ◦C. The bound rubber content of each sample was calculated according to the following equation [22]:

m − m w Bound rubber content (%) = 2 1 2 , (5) m1w1

where m1 is the initial weight of a sample, m2 is the final weight of the dried sample, w1 is the rubber fraction in the compound, and w2 is the silica fraction in the compound.

2.3.4. Tensile Properties Vulcanizates with a thickness of about 1.5 mm were die-cut into dumbbell-shaped specimens. The tensile tests were carried out using a universal testing machine (Instron 3365, Instron Co. Ltd., Bangkok, Thailand) with a crosshead speed of 500 mm/min according to ASTM D 412 at room temperature. Five specimens per vulcanizate were performed, and the average value was reported. Polymers 2021, 13, 39 7 of 22

2.3.5. Payne Effect as Indicative for Filler–Filler Interaction The Payne effect was measured using a rubber process analyzer (RPA, Alpha Tech- nologies, Ohio, USA). An uncured sample was first vulcanized directly in the RPA at 150 ◦C ◦ for its respective Tc95 and subsequently cooled to 100 C prior to the Payne effect test. The test was set with a strain sweep in the range of 0.56–100% at a frequency of 0.50 Hz. The Payne effect was calculated using the data of storage moduli measured, i.e., the difference between the storage shear moduli at 0.56% (G’0.56) and 100% strain (G’100).

2.3.6. Mechanical Stress Relaxation Analysis The stress relaxation test of vulcanizates was carried out using a universal testing machine (Instron 3365, Instron Co. Ltd., Bangkok, Thailand) with a crosshead speed of 100 mm/min. In this experiment, the deformation of a sample was set constant at a specific strain of 70% of the ultimate elongation at break of the sample. This means that each sample was predetermined to know its ultimate tensile properties, particularly elongation at break, prior to conducting the test. The decline in force over 60 min while keeping a constant deformation of the samples was monitored. The decay exponential stress relaxation curves of vulcanizates could be computed for their slope using Equation (6) [23].

σ(t) = σ0 − kln(t) (6)

where σ0 is the initial stress subjected to a specimen at the starting stage, and k represents the slope of the line.

2.3.7. Assessing the Self-Reparation of Molecular Damages in Vulcanizates In general, the “damage” of a material is defined as an undesirable failure occurring in a system that affects its performance; hence, an understanding of the damage nature and the damage progression is essential to suppress the damages, maintaining high reten- tion of the product properties. The damages do not necessarily lead to a suddenly entire loss of material performance [10], but rather a continual deterioration in overall proper- ties, once operated under actual or simulated service conditions. The degree of damage can be evaluated by comparing the two states of a material, i.e., pristine or undamaged versus damaged. The Mullins effect measurement is considered as an approach for monitoring the dam- age and recovery degree, since it induces the rupture of physical and chemical bonds [17] inside a material matrix during loading cycles; the rupture immediately progresses after the first loading cycle [24]. This rupture mechanism occurs in all cases, i.e., unfilled and filled or reinforced rubbers; this phenomenon is more pronounced in reinforced rubbers [17]. However, in this study, the damage was observed as a reduction in the material stiffness as a result of microscale defects, e.g., polymer chain and network breakages [25]. By utilizing the Mullins effect concept for the present experiment, the molecular network damages were introduced by applying multiple stress/strain cycles (i.e., 10 cycles) to samples [26]. The vulcanizates with a thickness of about 1.5 mm were die-cut into dumbbell-shaped specimens. A tensile testing machine was used; the initial free length between the clamps was 40 mm. A sample was stretched to its maximum extension ratio λ of 70% relative to the elongation at break of the vulcanizate predetermined with a different sample at a constant crosshead speed of 100 mm/min in ambient conditions. After 10 stretching cycles, the sample was rested at room temperature without pressure for 30 min. Then, the sample was treated for 30 min at different temperatures, i.e., room temperature, as well as 40, 80, and 120 ◦C, in a vacuum hot-air oven and cooled down to room temperature for 60 min. The stress–strain cyclic test was repeated for the treated samples. The obtained results were compared with that of the pristine counterparts to evaluate the self-reparation degree. In addition, hysteresis calculated from the area of the first stress–strain cycle was assessed to indicate the change in energy dissipation of vulcanizates [27]. To quantify Polymers 2021, 13, 39 8 of 22

the recovery degree, the recovery ratio (R) was defined as the value of the hysteresis of a healed sample at a specific temperature (Ahealed) to that of the original (Apristine) one, expressed as follows: A R = healed × 100 (7) Apristine

Polymers 2021, 13, x FOR PEER REVIEW3. Results and Discussion 8 of 24

3.1. Relationship among Mixing Data, Compound Viscosity, and Filler Interactions The mixing fingerprints of the investigated compounds are shown in Figure1. The 3. Results and Discussion final mixing torques, i.e., from about 700–900 s of mixing time, more or less leveling 3.1. Relationship among Mixing Data, Compound Viscosity, and Filler Interactions off (see Figure1a). These data technically confirms that the particles of additives, in The mixing fingerprints of the investigated compounds are shown in Figure 1. The particularfinal mixing silica torques, filler, i.e. are, from incorporated about 700–900 and s of brokenmixing time, down more to or their less leveling smallest off sizes that cannot be(see reduced Figure 1a) any. These further data technically under this confirms mixing that condition.the particles of Thus, additives, the in mixings particular were considered tosilica be completed,filler, are incorporated as the optimumand broken down dispersion to their degreesmallest sizes of the that particulate cannot be re- compositions is achievable.duced any further In general, under this the mixing mixing condition is the. Thus, most the criticalmixings were processing considered step to be determining the completed, as the optimum dispersion degree of the particulate compositions is achieva- qualityble. In general, of final the compounds mixing is the and most vulcanizates. critical processing These step mixingdetermining data the show quality that of all compounds reachfinal compounds an optimum and level vulcanizates in terms. These of themixing dispersion data show degree. that all compounds A noticeable reach difference an between theseoptimum compounds level in terms is the of the mixing dispersion torque degree intensity,. A noticeable especially difference at thebetween final these mixing stage. compounds is the mixing torque intensity, especially at the final mixing stage.

Figure 1. Mixing fingerprints based on (a) torque and (b) temperature of silica-filled NR compounds with various EMZ contents added via either extra addition (E05-E15) or blending (B05-B15).

Polymers 2021, 13, 39 9 of 22

Polymers 2021, 13, x FOR PEER REVIEW 9 of 24 The dump or discharge temperatures (Figure1b) and mixing torques of the com- pounds could be extracted and plotted as a function of EMZ loading with the two different additionFigure 1. Mixing procedures, fingerprints based as on depicted (a) torque and in (b Figure) temperature2. Uponof silica-filled increasing NR com- the content of EMZ, the pounds with various EMZ contents added via either extra addition (E05-E15) or blending (B05- dumpB15). temperatures of the silica-filled NR compounds increased, reflecting that the mixes generated a higher shear rate that in turn led to a higher thermal accumulation in the The dump or discharge temperatures (Figure 1b) and mixing torques of the com- compounds.pounds could be extracted Considering and plotted the as a fu factor(s)nction of EMZ causing loading with this the consequence,two different the dump tempera- turesaddition corresponded procedures, as depicted well within Figure the 2. Upon mixing increasing torques the content (Figure of EMZ,2) as the a result of the increase in dump temperatures of the silica-filled NR compounds increased, reflecting that the mixes compoundgenerated a higher viscosities shear rate duringthat in turn mixing. led to a higher Therefore, thermal accumulation during mixing, in the these compounds must generatecompounds different. Considering degreesthe factor(s) of causing intermolecular this consequence, networking the dump temperatures reactions and/or interactions amongcorresponded reactive well with sites, the mixing which torques contribute (Figure 2) to as thea result higher of the increase viscosities. in com- In addition, the EMZ can pound viscosities during mixing. Therefore, during mixing, these compounds must gen- acterate as different a coupling degrees agentof intermolecular which shieldsnetworking the reactions silica and surface/or interactions and better among interacts with NR chains, Polymers 2021, 13, x FOR PEER REVIEWcontributingreactive sites, which to an contribu increasete to the in higher the Mooneyviscosities. In viscosity, addition, the as EMZ discussed can act as a later 10 for of 24 the bound rubber coupling agent which shields the silica surface and better interacts with NR chains, con- results.tributing Theto an mixingincrease in torques, the Mooney dump viscosity, temperatures, as discussed later andfor the Mooney bound rubber viscosities showed relatively goodresults. correlations, The mixing torques, asillustrated dump temperatures, in Figure and Mooney3. viscosities showed rela- tivelyThis goodreaction correlations, becomes as moreillustrated pronounced in Figure 3under. elevated temperatures. The silanization of the silane hydrophobizes the polar surface of silica, resulting in less polarity, which provides better compatibility with the nonpolar rubber matrix, i.e., NR in this case. Thus, the dispersion of silica in NR is much better if compared with a system without a silane coupling agent [28]. The amelioration in filler dispersion should reduce the viscosity of the compounds, as discussed later for the results regarding the filler–filler interaction or Payne effect. However, when the TESPT molecule couples with silica clusters, there is a possibility that its molecule splits off into two halves, creating reactive sulfur radicals at the end half-molecule. This site is highly reactive toward NR (diene rubber) molecules and can form a covalent bridge between silica and rubber during elevated temperatures, e.g., this mixing condition [16], leading to a so-called higher “filler–rubber interaction”. This interaction contributes to a higher viscosity of the compounds. In addition, TESPT has a side effect of releasing free sulfur into the system during mixing at high temperature, since it has four sulfur atoms on average in the middle of its molecule. When it snaps during mixing, it can be assumed that each half of the silane molecule carries one sulfur atom, meaning that there are two atoms of sulfur on average for each TESPT molecule liberated into the system. This free sulfur can partially crosslink NR molecules during mixing, contributing to the polymer network and, thus, increasing the compound viscos- ity [16]. All investigated compounds contained the same dosages of silica, silane, and cu- rative. Therefore, the effects of silica–silane–rubber coupling and lightly crosslinked rub- Figureber were 2. Dump assumed temperatures to be and identical final mixing among torques these of the compounds silica-filled NR. Thesecompounds unique as a phenomena of Figurefunction 2.of EMZDump content temperatures added via either ex andtra addition final (Set mixing A) or blending torques (Set B) of. the silica-filled NR compounds as a the silica–TESPT-filled NR compound further revealed that the reactive sites originating function of EMZ content added via either extra addition (Set A) or blending (Set B). fromAs silica mentioned and silane above, were important most intermolecular likely to interact reactions/interactions with the self-healing must have EMZ oc- modifier. curred during mixing, which led to an increase in compound viscosities, as indicated by the mixing torques and Mooney viscosities. This resulted in increased mixing tempera- tures (see Figure 3). Inside these compounds, there were several possible simultaneous interactions due to the reactive ingredients and a rather high mixing temperature, i.e., 130–150 °C. These ingredients included natural rubber, silica, TESPT silane, accelerators, activators, and the EMZ modifier. The investigated compounds had different qualities of the EMZ masterbatch, while the dosages of other compositions remained identical. Hence, this EMZ component could be the key influencing factor. Regarding the reference compound which was based on a silica–silane-filled NR sys- tem, silica can react with bifunctional organosilane, i.e., bis-triethoxysilylpropyl tetrasul- fide or TESPT, under temperatures as low as 120 °C, known as the “silanization reaction”.

Figure 3. Correlations among final mixing torques, dump temperatures, and Mooney viscosities of Figuresilica-filled 3. Correlations NR compounds among with different final mixing amountstorques, of EMZ added dump via temperatures,either extra addition and (E05– Mooney viscosities of silica-filledE15) or blending NR compounds(B05–B15). with different amounts of EMZ added via either extra addition (E05–E15) or blending (B05–B15). The self-healing EMZ modifier contains hydroxyl groups from the opened structure of epoxide or oxirane rings on ENR molecules [15]. A direct condensation reaction be- tween the silanol groups on the silica surface and the radical ions at the opened epoxide positions on ENR molecules [21] could occur during mixing due to their acidic character

Polymers 2021, 13, 39 10 of 22

As mentioned above, important intermolecular reactions/interactions must have oc- curred during mixing, which led to an increase in compound viscosities, as indicated by the mixing torques and Mooney viscosities. This resulted in increased mixing temperatures (see Figure3). Inside these compounds, there were several possible simultaneous interac- tions due to the reactive ingredients and a rather high mixing temperature, i.e., 130–150 ◦C. These ingredients included natural rubber, silica, TESPT silane, accelerators, activators, and the EMZ modifier. The investigated compounds had different quantities of the EMZ masterbatch, while the dosages of other compositions remained identical. Hence, this EMZ component could be the key influencing factor. Regarding the reference compound which was based on a silica–silane-filled NR system, silica can react with bifunctional organosilane, i.e., bis-triethoxysilylpropyl tetrasul- fide or TESPT, under temperatures as low as 120 ◦C, known as the “silanization reaction”. This reaction becomes more pronounced under elevated temperatures. The silanization of the silane hydrophobizes the polar surface of silica, resulting in less polarity, which provides better compatibility with the nonpolar rubber matrix, i.e., NR in this case. Thus, the dispersion of silica in NR is much better if compared with a system without a silane coupling agent [28]. The amelioration in filler dispersion should reduce the viscosity of the compounds, as discussed later for the results regarding the filler–filler interaction or Payne effect. However, when the TESPT molecule couples with silica clusters, there is a possibility that its molecule splits off into two halves, creating reactive sulfur radicals at the end half-molecule. This site is highly reactive toward NR (diene rubber) molecules and can form a covalent bridge between silica and rubber during elevated temperatures, e.g., this mixing condition [16], leading to a so-called higher “filler–rubber interaction”. This interaction contributes to a higher viscosity of the compounds. In addition, TESPT has a side effect of releasing free sulfur into the system during mixing at high temperature, since it has four sulfur atoms on average in the middle of its molecule. When it snaps during mixing, it can be assumed that each half of the silane molecule carries one sulfur atom, meaning that there are two atoms of sulfur on average for each TESPT molecule liberated into the system. This free sulfur can partially crosslink NR molecules during mixing, contributing to the polymer network and, thus, increasing the compound viscos- ity [16]. All investigated compounds contained the same dosages of silica, silane, and curative. Therefore, the effects of silica–silane–rubber coupling and lightly crosslinked rubber were assumed to be identical among these compounds. These unique phenomena of the silica–TESPT-filled NR compound further revealed that the reactive sites originating from silica and silane were most likely to interact with the self-healing EMZ modifier. The self-healing EMZ modifier contains hydroxyl groups from the opened structure of epoxide or oxirane rings on ENR molecules [15]. A direct condensation reaction between the silanol groups on the silica surface and the radical ions at the opened epoxide positions on ENR molecules [21] could occur during mixing due to their acidic character and thermal acceleration. This led to linkage formation between EMZ and silica. The increased degree of this EMZ–silica reaction induced the filler–rubber interactions in the system, as clearly indicated by the bound rubber contents shown in Figure4, resulting in a higher viscosity and, thus, elevated shear rates, which in turn contributed to a higher mixing temperature (Figures2 and3). Therefore, these results essentially hint that the chemical interactions between EMZ modifier and silica and/or the self-association of the EMZ itself, i.e., the reactions between HMA and zinc acetate with ENR-50, can be activated by an effective high temperature during mixing, i.e., 135–145 ◦C. The increase in compound viscosities with a higher loading of EMZ, as shown in Figure3, can be attributed not only to the enhancement in silica–rubber interactions, but also other parameters contributing toward increased compound viscosity. In general, a better silica dispersion resulting from a lower filler–filler interaction plays a key role in reducing the viscosity of the compounds [29]. In this study, TESPT was used at its optimum level (i.e., silica/TESPT: 55/5 phr) [16]; this TESPT content was basically sufficient to hydrophobize the silica surface, giving reduced compound viscosities. However, the free Polymers 2021, 13, 39 11 of 22

sulfur released from TESPT during mixing could partially crosslink the rubber chains, leading to increased compound viscosities [16]. Moreover, the addition of EMZ could contribute to higher compound viscosities due to reactive interactions. There were two possible reactions occurring during mixing responsible for the rise in compound viscosity upon increasing the EMZ level: (1) The self-association of EMZ through hydrogen bonding or polar interactions between HMA and ENR-50 or even the possible self-crosslinking of ENR [30,31], and (2) chemical interactions generated between silica and ENR [32,33]. These two intermolecular interactions seem to involve in the formation of extra networks, Polymers 2021, 13, x FOR PEER REVIEW 12 of 24 contributing to the increase in viscosities of silica-filled NR compounds.

Figure 4. Influence of EMZ content on chemically bound rubber contents and Payne effect of sil- Figureica-filled 4. NRInfluence compounds of. EMZEMZ was content added onvia chemicallyeither extra addition bound (Set rubber A) or blending contents (Set and B). Payne effect of silica-filled NR compounds. EMZ was added via either extra addition (Set A) or blending (Set B). 3.2. Cure Characteristics and Apparent Crosslink Density The filler–filler interactions of the compounds monitored using a Payne effect analysis The cure characteristics of silica-filled NR compounds, i.e., scorch time (Ts2), opti- weremum evaluated, cure time (Tc90 as), shownminimum in torque Figure (S’4Min.), The and Payne maximum effect torque of silica-filled(S’Max), are shown NR in compounds slightlyTable 2. decreasedWith increasing with EMZ increasing loading, the EMZ scorch content. and cure This times result considerably implies decreased that the. interfacial interactionsIt was previously between reported the silicathat the and silica rubber surfaces matrix are reactive were toward enhanced curative after absorption, the addition of EMZ, givingparticularly a greatly accelerators reduced and filler–filler ZnO, since interaction.these chemicals As are mentioned polar, which above, in turn the results intermolecular . interactionsin cure retardation originatedHence, through the interactions strong covalentbetween EMZ and/or and hydrogenthe silanol groups bonds of between silica the ENR led to reduced absorption of the curatives on the silica surface and, thus, a faster cure rate andof the silica. compounds This also [34] corresponded upon increasing well EMZwith loading the. Additionally, improvement the inENRs bound contained rubber contents (Figureepoxide4), groups indicating along atheir better chains silica–rubber that could activate interaction. the adjacent Therefore, double the bonds formation to be of strong interactionsmore reactive between toward creating the silica a radical and at ENR-50 an allylic consequentlyposition on the rubber accounted chains for. Thus, a better silica dispersionthe epoxide in group thecompound could also increase with increasing the vulcanization EMZ loading.rate and reduce the scorch and cure times [35]. The compounds of Set B possessed a higher quantity of EMZ or ENR 3.2.compared Cure Characteristics to their counterparts and Apparent of Set A Crosslink. Therefore, Density the cure and scorch times of Set B compounds were slightly faster than those of Set A compounds. The cure characteristics of silica-filled NR compounds, i.e., scorch time (Ts2), optimum cureTable time 2. Cure (Tc90 characteristics), minimum of silica torque-filled NR (S’ Mincompounds), and with maximum different EMZ torque loadings (S’Max. ), are shown in Table2. With increasing EMZ loading, the scorch and cure times considerably decreased. Cure Characteristics It wasCompounds previously reported that the silica surfaces are reactive toward curative absorption, TS2 (min) TC90 (min) S′Min (dN·m) S′Max (dN·m) ΔS′, S′Max − S′Min (dN·m) particularlyReference accelerators1.01 7 and.16 ZnO,0 since.59 these12 chemicals.26 are polar,11.67 which in turn results in cureE-05 retardation. 1.08 Hence,6.29 the interactions0.63 between12.33 EMZ and the11.70 silanol groups of silica led toE reduced-10 absorption0.54 5. of56 the curatives0.67 on the12. silica79 surface and,12.12 thus, a faster cure rate of theE compounds-15 0.52 [34] upon5.26 increasing0.77 EMZ loading.13.21 Additionally,12.44 the ENRs contained epoxideB-05 groups along0.51 their5.54 chains that0.70 could activate12.95 the adjacent12 double.25 bonds to be more - . . . . . reactiveB 10 toward 0 creating46 5 a14 radical at0 79 an allylic13 position13 on the rubber12 34 chains. Thus, the B-15 0.41 5.05 1.08 13.76 12.68 epoxide group could also increase the vulcanization rate and reduce the scorch and cure The rheometer minimum torques slightly increased with increasing EMZ content, which corresponded to the increasing trends of Mooney viscosities and the final mixing torque (see Figure 3). The rise in the maximum torque values with increasing content of

Polymers 2021, 13, 39 12 of 22

times [35]. The compounds of Set B possessed a higher quantity of EMZ or ENR compared to their counterparts of Set A. Therefore, the cure and scorch times of Set B compounds were slightly faster than those of Set A compounds.

Table 2. Cure characteristics of silica-filled NR compounds with different EMZ loadings.

Cure Characteristics Compounds 0 0 0 0 0 TS2 (min) TC90 (min) S Min (dN·m) S Max (dN·m) ∆S ,S Max − S Min (dN·m) Reference 1.01 7.16 0.59 12.26 11.67 E-05 1.08 6.29 0.63 12.33 11.70 E-10 0.54 5.56 0.67 12.79 12.12 E-15 0.52 5.26 0.77 13.21 12.44 B-05 0.51 5.54 0.70 12.95 12.25 B-10 0.46 5.14 0.79 13.13 12.34 B-15 0.41 5.05 1.08 13.76 12.68

The rheometer minimum torques slightly increased with increasing EMZ content, which corresponded to the increasing trends of Mooney viscosities and the final mixing Polymers 2021, 13, x FOR PEER REVIEW 13 of 24 torque (see Figure3). The rise in the maximum torque values with increasing content of EMZ supported the higher overall network densities in the system, as indicated by the results of bound rubber contents (Figure4) and apparent crosslink densities (Figure5). ForEMZ filled support compounds,ed the higher the overall cure network torque densities difference in the was system affected, as indicated not only by by the the crosslink results of bound rubber contents (Figure 4) and apparent crosslink densities (Figure 5). density of the vulcanizate but also the filler–filler interaction. However, according to the For filled compounds, the cure torque difference was affected not only by the crosslink Paynedensity effect of the shownvulcanizate in Figure but also4, the there filler was–filler a clearinteraction decrease. However, in filler–filler according interaction.to the Thus, itPayne is obvious effect shown that in the Figure network 4, there densities was a clear of decrease Set B vulcanizates in filler–filler interaction were greater. Thus, than those of theirit is obvious Set A counterparts. that the network The densities self-healing of Set B EMZvulcanizates modifier were showed greater itsthan unique those of characteristic oftheir increasing Set A counterparts the network. The self density,-healing i.e., EMZ polymer modifier crosslinks showed its unique and filler–rubber characteristic interactions, inof thisincreasing silica-filled the network NR system. density, i.e., polymer crosslinks and filler–rubber interactions, in this silica-filled NR system.

FigureFigure 5. EffectEffect of ofEMZ EMZ contents contents on the on apparent the apparent crosslink crosslink density of densitysilica-filled of NR silica-filled vulcanizates NR. vulcanizates. EMZ was added via either extra addition (Set A) or blending (Set B). EMZ was added via either extra addition (Set A) or blending (Set B). 3.3. Reinforcing, Mechanical, and Viscoelastic Properties The reinforcement index, i.e., M300/M100 [29], is commonly used to indicate the re- inforcing efficiency of reinforced rubber compounds, especially in tire technology. In gen- eral, this reinforcement index corresponds well with tan δ at 60 °C derived from a dynamic mechanical analysis (DMA), whereby a higher M300/M100 or lower tan δ at 60 °C denotes a higher reinforcing efficiency, which theoretically leads to the better rolling resistance of a tire [36]. According to Figure 6, the reinforcement indices of silica-filled NR vulcanizates showed an increasing trend with increasing EMZ loading. This result is not surprising, since more filler–rubber interactions were generated by EMZ (as discussed in Figure 6) in this system, giving better homogeneity of the mix and stronger linkages between silica and rubber, consequently leading to enhanced reinforcement power. Figure 6 shows the tensile strength and elongation at break of silica-filled NR vul- canizates containing various EMZ contents. It is obvious that, upon increasing the con- centration of EMZ, the elongation at break slightly decreased for the vulcanizates with EMZ from 5 to 10 phr and significantly decreased for the vulcanizates with EMZ at 15 phr (approximately 50 units different), attributed to an increase in apparent crosslink density. The improvement in tensile strength with increasing content of EMZ was consistent with the improved reinforcement index and could also be attributed to a higher filler–rubber interaction, as indicated by the bound rubber content and apparent crosslink density (see Figures 4 and 5).

Polymers 2021, 13, 39 13 of 22

3.3. Reinforcing, Mechanical, and Viscoelastic Properties The reinforcement index, i.e., M300/M100 [29], is commonly used to indicate the rein- forcing efficiency of reinforced rubber compounds, especially in tire technology. In general, this reinforcement index corresponds well with tan δ at 60 ◦C derived from a dynamic mechanical analysis (DMA), whereby a higher M300/M100 or lower tan δ at 60 ◦C denotes a higher reinforcing efficiency, which theoretically leads to the better rolling resistance of a tire [36]. According to Figure6, the reinforcement indices of silica-filled NR vulcanizates showed an increasing trend with increasing EMZ loading. This result is not surprising, since more filler–rubber interactions were generated by EMZ (as discussed in Figure6) in Polymers 2021, 13, x FOR PEER REVIEWthis system, giving better homogeneity of the mix and stronger linkages between silica14 and of 24 rubber, consequently leading to enhanced reinforcement power.

FigureFigure 6. 6.Effect Effect of of EMZ EMZ content content on on tensile tensile strength, strength, elongation elongation at break,at break, and and reinforcement reinforcement index index of - . silica-filledof silica filled NR vulcanizates. NR vulcanizates EMZEMZ was addedwas added via eithervia either extra extra addition addition (Set A)(Set or A) blending or blending (Set B). (Set B). Figure6 shows the tensile strength and elongation at break of silica-filled NR vul- canizatesFigure containing 7 displays various the decay EMZ of contents. tensile stress It is obviousas a function that, of upon time increasing for the vulcanizates, the con- centrationcharacterized of EMZ, by a the stress elongation relaxation at test break. The slightly slope decreasedof each stress for therelaxation vulcanizates line indicates with EMZthe fromviscoelastic 5 to 10 phrproperties, and significantly i.e., the elastic decreased and forviscous the vulcanizates ratio, of a vulcanized with EMZ atelastomer 15 phr . (approximatelyThe stress relaxation 50 units is different),associated attributed with the co tontinually an increase progressive in apparent breakage crosslink of networks, density. Theinteractions, improvement and/or in tensile molecular strength chains, with asincreasing well as the content movement of EMZ of viscous was consistent segments, with e.g., thesmall improved molecules reinforcement and uncrosslinked index and short could chains, also be under attributed applied to adeformation higher filler–rubber [37]. The interaction,slopes of the as vulcanized indicated bycomposites the bound with rubber a higher content EMZ andcontent apparent were slightly crosslink steeper density than (seethose Figures of the4 and samples5). without or with a lower loading of EMZ. This reflects that the EMZ promotedFigure7 adisplays greater thestress decay relaxa of tensiletion rate stress of vulcanizates, as a function implying of time for that the an vulcanizates, increment in characterizedEMZ content by in a the stress vulcanizates relaxation contributed test. The slope to a ofhigher each viscous stress relaxation proportion, line as indicates well as to thea potentially viscoelastic increased properties, extent i.e., of the intermolecular elastic and viscous interactions ratio, and of a/or vulcanized short chains elastomer. [38,39]. A Thedetailed stress relaxationdiscussion ison associated the involved with interactio the continuallyns in this progressive system is breakageprovided ofin networks,Section 3.5. interactions,Furthermore, and/or the molecular molecular weight chains, of as NR well could as the be movement reduced due of viscousto the EMZ segments, modifier e.g.,. It smallis commonly molecules known and uncrosslinked that hydrolyzed short maleic chains, acid under or dicarboxylic applied deformation acid, used [as37 ].an The ester slopescrosslinking of the vulcanized agent, can composites shorten the with diene a higherrubber EMZchains content due to were their slightly acidic characteristics steeper than . thoseThis of leads the samplesto a greater without extent or withof shorter a lower molecular loading of chains, EMZ. Thisa lower reflects average that themolecular EMZ promotedweight, or a greatera more viscous stress relaxation segment, ratethereby of vulcanizates, increasing the implying stress relaxation that an incrementrate of vulcan- in EMZizates, content as indicated in the vulcanizates by a higher contributedabsolute value to aof higher the stress viscous relaxation proportion, slope. as well as to a potentiallyConsidering increased the initial extent stress of intermolecular of each sample, interactions it increased and/or withshort increasing chains EMZ [38,39 con-]. Atent detailed (see Figure discussion 7). This on the was involved attributed interactions to the extra in crosslinks this system generated is provided by intheSection self-healing 3.5. modifier EMZ (see Figures 4 and 5). These additional crosslinks were formed through ester, ionic, or hydrogen bonds. The functionalities of EMZ are also beneficial for enhanc- ing silica–rubber interactions and, thus, improving reinforcing efficiency [40]; more de- tails are discussed in Section 3.5.

Polymers 2021, 13, 39 14 of 22

Furthermore, the molecular weight of NR could be reduced due to the EMZ modifier. It is commonly known that hydrolyzed maleic acid or dicarboxylic acid, used as an ester crosslinking agent, can shorten the diene rubber chains due to their acidic characteristics. This leads to a greater extent of shorter molecular chains, a lower average molecular weight, Polymers 2021, 13, x FOR PEER REVIEWor a more viscous segment, thereby increasing the stress relaxation rate of vulcanizates,15 of 24 as indicated by a higher absolute value of the stress relaxation slope.

Figure 7. Stress relaxation curves as a function of EMZ loading levels of silica-filled NR vulcanizates. EMZ was added via Figure 7. Stress relaxation curves as a function of EMZ loading levels of silica-filled NR vulcanizates. EMZ was added via either extra addition (Set A) or blending (Set B). either extra addition (Set A) or blending (Set B).

3.4. PropertyConsidering Retention the initial after Molecular stress of each Damages sample, it increased with increasing EMZ content (see FigureThe stress7). This–strain was cyclic attributed test, often to the known extra crosslinksas the Mullins generated effect, by was the performed self-healing to verifymodifier the EMZ reversibility (see Figures of 4molecular and5). These interactions, additional e. crosslinksg., ionic and were hydrogen formed through bonds, ester,inter- molecularionic, or hydrogen forces, and bonds. loosely The physical functionalities entanglements, of EMZ after are alsointroducing beneficial molecular for enhancing dam- agessilica–rubber through interactionsa deformation and, process thus, improvingto a vulcanizate reinforcing. Samples efficiency were stretched [40]; more to details a certain are straindiscussed of 70% in Sectionof their 3.5 predetermined. ultimate elongation at break. At this applied strain, some bonds, molecules, or intermolecular linkages were likely ruptured, leading to micro- damages3.4. Property inside Retention the vulcanizate after Molecular samples Damages. The possible micro-damages could have oc- curredThe at stress–strainfiller–filler interactions, cyclic test, often loose known rubber as– thefiller Mullins interactions, effect, wasshort performed rubber molecules to verify betweenthe reversibility crosslinking of molecular points, and interactions,/or entanglements e.g., ionic [24] and. hydrogenAs evidenced bonds, in previous intermolecular work [17],forces, an and obviously loosely sharp physical decrease entanglements, in the instan aftertaneous introducing tensile molecular stress of the damages material through in the seconda deformation test cycle process was reported to a vulcanizate.. This is a Samplescommon werephenomenon stretched of to a a vulcanizate certain strain. It ofcan 70% be seenof their that predetermined the second cycle ultimate had a much elongation smaller at hysteresis break. At ( thissee Figure applied 8a–c) strain,. This some is because bonds, themolecules, major breakages or intermolecular of molecules/intermolecular linkages were likely interactions ruptured, occurred leading toin micro-damagesthe first stress– straininside cycle the. vulcanizate However, the samples. values derived The possible from the micro-damages following cycle could could have imply occurred some im- at mediatefiller–filler reversible interactions, interactions loose rubber–filler in the system interactions,. The hysteresis short of rubber vulcanizates molecules can betweenbe char- acterizedcrosslinking by the points, difference and/or between entanglements the integral [24 areas]. As of evidencedthe loading in and previous unloading work stress [17–], strainan obviously curves sharp[17] (see decrease Figure in 8d) the. instantaneousAn obtainable tensile value stress indica oftes the the material energy in dissipation the second fromtest cycle the rupture was reported. of bonds This/interactions is a common under phenomenon a large deformation of a vulcanizate.. It can be seen that the second cycle had a much smaller hysteresis (see Figure8a–c). This is because the major breakages of molecules/intermolecular interactions occurred in the first stress–strain cycle. However, the values derived from the following cycle could imply some immediate reversible interactions in the system. The hysteresis of vulcanizates can be characterized by the difference between the integral areas of the loading and unloading stress–strain curves [17] (see Figure8d). An obtainable value indicates the energy dissipation from the rupture of bonds/interactions under a large deformation.

Polymers 2021, 13, 39 15 of 22 Polymers 2021, 13, x FOR PEER REVIEW 16 of 24

( ) ( .) - FigureFigure 8. (aa)Stress–strain curves of the reference sample (Ref.)Ref and B B-1515 vulcanizate under a cyclic tensile deformation for 10 cycles. Integral area of hysteresis loop as a function of cycle number of the vulcanizates from (b) Set A, with extra for 10 cycles. Integral area of hysteresis loop as a function of cycle number of the vulcanizates from (b) Set A, with extra addition of EMZ, and (c) Set B, with blending of EMZ. (d) Schematic illustration of how to calculate the values of hysteresis addition of EMZ, and (c) Set B, with blending of EMZ. (d) Schematic illustration of how to calculate the values of hysteresis from the loading and unloading stress–strain curves. from the loading and unloading stress–strain curves. As a function of the hysteresis values derived from the integral areas under the cyclic As a function of the hysteresis values derived from the integral areas under the cyclic stress–strain curves of samples shown as examples in Figure 9, the degree of hysteresis stress–strain curves of samples shown as examples in Figure9, the degree of hysteresis recovery after micro-damage could be calculated. Figure 10 shows the resulting recovery recovery after micro-damage could be calculated. Figure 10 shows the resulting recovery degree of the investigated samples. Upon increasing the amount of the EMZ modifier, degree of the investigated samples. Upon increasing the amount of the EMZ modifier, - togethertogether with with elevating elevating the the annealing annealing temperature temperature after after a micro a micro-damage,damage, the theproperty property re- . ( tentionretention was was substa substantiallyntially enhanced enhanced. A A higher higher recovery recovery (aboutabout 84%) comparedcompared toto the the refer- ref- erenceence (Ref.) (Ref.) sample sample was was shown shown for for the the B-15 B-15 sample sample after after undergoing undergoing a thermala thermal treatment treatment at at120 120◦C. °C This. This result result essentially essentially suggests suggests that that there there were were plausibly plausibly reversible reversible elements elements in the in thesystems. systems On. On the otherthe other hand, hand, for the for reference the reference sample, sample, the recovery the recovery ratio showed ratio showed a signifi- a significantlycantly smaller smaller value value due to due a lesser to a lesser extent extent of reversible of reversible segments segments in the in vulcanizate. the vulcanizate Thus,. Thus,it is obvious it is obvious that the that micro-damage the micro-damage could be could alleviated be alleviated by adding by theadding self-healing the self- modifierhealing modifierEMZ. In addition,EMZ. In elevatedaddition, temperatures elevated temperatures could significantly could significantly accelerate the accelerate recovery the perfor- re- coverymance. performance This is common. This for is acommon chemical for reaction/interaction, a chemical reaction/interaction, as higher temperatures as higher tem- raise peraturesthe reaction raise kinetics. the reaction Moreover, kinetics this. system Moreover, of silica-filled this system NR of plus silica ENR-based-filled NR self-healingplus ENR- basedmodifier self- containedhealing modifier rather complexcontained interactions, rather comple sincex interactions, there were since various there reactive were vari- sites ouswhich reactive could sites link which to one another,could link particularly to one another, through particularly chemical mechanisms.through chemical mecha- nisms.

Polymers 2021, 13, x FOR PEER REVIEW 17 of 24

Polymers 2021, 13, 39 16 of 22 Polymers 2021, 13, x FOR PEER REVIEW 17 of 24

Figure 9. Examples of stress–strain cyclic curves of (a) the reference and (b) B-15 vulcanizates from the first cycle of a cyclic tensile test. The pristine sample of each vulcanizate after its first-round test was thermally treated at different tem- peratures prior to the repetition of the stress–strain cyclic test.

Regarding the similar trend of recovery ratio observed for the reference sample, it is commonly known that, in a silica-filled rubber compound, there are always reversible interactions in the system. These interactions by nature originate from the reactive ingre- dients formulated in the compound such as active fillers (silica), as well as from polar Figure 9. Examples of stress–strainspecies cyclic in the curves system, of (a )e the.g., referenceprotein inand natural (b) B- 15rubber vulcanizates molecules from andthe firstthe cyclelike. ofMoreover, a Figure 9. Examples of stress–strain cyclic curves of (a) the reference and (b) B-15 vulcanizates from the first cycle of a cyclic cyclic tensile test. The pristinedis sample- and of re each-entanglement vulcanizate afterof the its rubber first-round chains test can was also thermally play a treatedrole in atthe different property tem- recovery tensile test. The pristine sample of each vulcanizate after its first-round test was thermally treated at different temperatures peratures prior to the repetition[41] of. theFurther stress– informationstrain cyclic test on. the possible intermolecular interactions is provided in the prior to the repetition of the stress–strain cyclic test. next section. Regarding the similar trend of recovery ratio observed for the reference sample, it is commonly known that, in a silica-filled rubber compound, there are always reversible interactions in the system. These interactions by nature originate from the reactive ingre- dients formulated in the compound such as active fillers (silica), as well as from polar species in the system, e.g., protein in natural rubber molecules and the like. Moreover, dis- and re-entanglement of the rubber chains can also play a role in the property recovery [41]. Further information on the possible intermolecular interactions is provided in the next section.

( ) - FigureFigure 10. 10.Recovery Recovery of of the the hysteresis hysteresis after after a tensilea tensile cyclic cyclic test test (10 10 cycles) cycles ofof silica-filled silica filled NR NR vulcan- vulcan- izates with different contents of EMZ modifier added via either extra addition (Set A) or blending izates with different contents of EMZ modifier added via either extra addition (Set A) or blend- (Set B). ing (Set B).

Regarding the similar trend of recovery ratio observed for the reference sample, it is commonly known that, in a silica-filled rubber compound, there are always reversible interactions in the system. These interactions by nature originate from the reactive ingre- dients formulated in the compound such as active fillers (silica), as well as from polar speciesFigure 10. in Recovery the system, of the e.g., hysteresis protein after in a natural tensile cyclic rubber test molecules (10 cycles) andof silica the-filled like. NR Moreover, vulcan- dis-izates and with re-entanglement different contentsof of theEMZ rubber modifier chains added can via also either play extr aa roleaddition in the (Set property A) or blending recov- . ery(Set [B)41 ]. Further information on the possible intermolecular interactions is provided in the next section.

Polymers 2021, 13, x FOR PEER REVIEW 18 of 24

3.5. Relevant Intermolecular Interactions in this System The possible intermolecular interactions could be summarized into a graphical scheme, as illustrated in Scheme 2. These interactions could be categorized into two main types, permanent and reversible linkages, as described below. Permanent interactions.—These include rubber–rubber crosslinks derived from monosulfidic (C–S–C), disulfidic (C–S2–C), polysulfidic (C–Sx–C), carbon–carbon (C–C), ether, and ester linkages. The sulfidic crosslinks were the majority in this case since the compounds were vulcanized with an accelerated sulfur cure system. In addition, free sul- fur liberated from the silane coupling agent “TESPT” could crosslink the rubber mole- cules, contributing to the network density [42]. For the ENR-based self-healing modifier, direct silica–epoxide coupling could have occurred, forming strong interactions between Polymers 2021, 13, 39 ENR molecules and silica particles in addition to the coupling through the silane bridging 17 of 22 mechanism [32,43]. The network derived from ENR-based self-healing modifier was con- sidered as the minority in this system; ether crosslinks (C–O–C) can be formed through the self-crosslinking of hydroxyl groups from the opened epoxide structure of ENR 3.5.[30,43,44], Relevant and Intermolecular ester linkages Interactionscan also be generated in this System from carboxylic groups of the hydro- lyzedThe maleic possible anhydride intermolecular [14,15]. Moreover, interactions tightly could entangled be summarized rubber chains into are a graphical also con- scheme, . assidered illustrated permanent in Scheme crosslinks2. These These interactionspermanent interactions could be are categorized harder to be into damaged two main by types, applied forces compared to ionic, hydrogen bond, polar, and other weak chemical and permanent and reversible linkages, as described below. physical interactions.

SchemeScheme 2. PostulatedPostulated intermolecular intermolecular interactions interactions in silanized in silanized silica-filled silica-filled natural natural rubber rubbercom- compounds pounds with the presence of the self-healing modifier EMZ. with the presence of the self-healing modifier EMZ. Reversible interactions.—The compounds prepared in this study were combined Permanent interactions.—These include rubber–rubber crosslinks derived from mono- with an ENR-based self-healing modifier, i.e., EMZ. This modifier likely contributed a sulfidicgreat extent (C–S–C), to the disulfidicreversible linkages (C–S2–C), to the polysulfidic rubber network (C–S.x –C),As illustrated carbon–carbon in Scheme (C–C), 2, ether, andthe possible ester linkages. reversible The interactions sulfidic include crosslinks hydrogen were bonding, the majority as well in as this rubber case– sincerubber the com- poundsand rubber were–filler vulcanized polar interactions with an. As accelerated described in sulfur Scheme cure 1, the system. EMZ modifier In addition, has the free sulfur liberatedability to generate from the ester silane crosslinks coupling between agent ENR “TESPT” molecules could with crosslinkthe opened the oxirane rubber struc- molecules, contributingture. Despite tothe the ester network bonds, densityionic interactions [42]. For could the ENR-based have occurred self-healing through Zn modifier,2+ and direct silica–epoxidecarboxylic groups coupling from the could dicarboxylic have occurred, acids modifying forming the strong ENR molecules interactions [12,14,15] between. ENR moleculesMoreover, andENR silicaand silica particles can inform addition reversible to the hydrogen coupling bonding through between the silane the bridgingsilanol mech- anismgroups [ 32of ,silica43]. Theand networkthe epoxide derived groups from of EN ENR-basedR [21,32–34,43]. self-healing These interactions modifier could was consideredbe as the minority in this system; ether crosslinks (C–O–C) can be formed through the self- crosslinking of hydroxyl groups from the opened epoxide structure of ENR [30,43,44], and ester linkages can also be generated from carboxylic groups of the hydrolyzed maleic anhy- dride [14,15]. Moreover, tightly entangled rubber chains are also considered permanent crosslinks. These permanent interactions are harder to be damaged by applied forces com- pared to ionic, hydrogen bond, polar, and other weak chemical and physical interactions. Reversible interactions—The compounds prepared in this study were combined with an ENR-based self-healing modifier, i.e., EMZ. This modifier likely contributed a great extent to the reversible linkages to the rubber network. As illustrated in Scheme2, the possible reversible interactions include hydrogen bonding, as well as rubber–rubber and rubber–filler polar interactions. As described in Scheme1, the EMZ modifier has the ability to generate ester crosslinks between ENR molecules with the opened oxirane structure. Despite the ester bonds, ionic interactions could have occurred through Zn2+ and carboxylic groups from the dicarboxylic acids modifying the ENR molecules [12,14,15]. Moreover, ENR and silica can form reversible hydrogen bonding between the silanol groups of silica and the epoxide groups of ENR [21,32–34,43]. These interactions could be reversed after their breakage. Thermal treatment could accelerate the reversible mechanism of these interactions. Silica clusters have a strong affinity among themselves. They tend to form larger clusters via hydrogen bonding due to a high concentration of silanol groups on Polymers 2021, 13, 39 18 of 22

the silica surface. Large silica clusters/network are easy to break down under an applied force [28]. However, they can be recovered when a vulcanizate returns to its original shape. These reversible interactions are considered to be reformed to a certain extent after micro- damages under an unloading condition. In this system, it is most likely that the reversible interactions played a major role in the property retention of the samples, particularly with a higher amount of the healing modifier and an elevated temperature. Moreover, van der Waals forces are additional interactions that can cause the reversibility of a vulcanizate. Large or small molecules possessing similar chemistry have a good affinity between each other, which leads to their good compatibility; these interactions are relatively weak but could confer materials with coherent characteristics. Physical chain entanglements—In a rubber network, there are tightly and loosely entangled chains. Tight chain entanglements are hardly loosened and are, thus, considered as similar to chemical crosslinks. A high extension results in the removal of loose entangle- ments, thereby lowering the stress on the second and subsequent loading–unloading cycles, as evidenced in Figure8a. This would result in a change in material entropy [ 45]. How- ever, it could partially be recovered by a high temperature. The recovery is supported by thermal motions which produce new physical chain entanglements [46]. It was remarked that thermal treatment can only accelerate the return of rubber chains to their equilibrium state but cannot assist in recovering the ruptured irreversible bonds [47]; this type of bond breakage is regarded as permanent molecular damage. However, the calculated apparent recovery ratio clearly confirmed good performance after three experimental repetitions. The recovery ratios slightly decreased after the first repetition of the experiment, as shown in Figure 11. In this study, the recovery process was dominated by the reversibility of crosslinked networks in the system, permitting the crosslinked network to be rearranged during thermal treatment. Upon heating, the fractured portions of the network diffused and re-entangled between the molecular chains, causing the ruptured crosslinks to restore their integrity, most possibly due to the ex- changeable transesterification interaction and the thermo-reversible hydrogen and/or ionic bonds. As reported in our previous work [15], the transesterification reaction of ester crosslinks/bonds can be accelerated by thermal treatment, whereby a higher treatment temperature (e.g., from room temperature to 200 ◦C) results in a higher bond-interchange degree, leading to a greater potential for intermolecular self-reparation of the rubber net- work. In addition, this bond-exchange reaction requires a transesterification catalyst, which was ZAD in this case, whereas the epoxide groups on ENR molecules also needed to be opened, providing radical ions and OH groups to the system, which could react with Polymers 2021, 13, x FOR PEER REVIEW 20 of 24 the carboxylic groups in the self-healing modifier. A chemical reaction regarding this bond-interchange mechanism was proposed in previous works [12,13,15].

Figure 11. The recovery ratio of silica-filled NR vulcanizates with different EMZ loading levels treated at (a) room tem- Figure 11. The recovery ratio of silica-filled NR vulcanizates with different EMZ loading levels treated at (a) room perature and (b) 120 °C for three rounds of the cyclic test. temperature and (b) 120 ◦C for three rounds of the cyclic test. 3.6. Potential Role of the Thermochemically Exchangeable Ester Bonds In addition to the self-association of the aforementioned reversible interactions that highly contributed to the intermolecular self-reparation of the molecular damages in the silica-filled NR vulcanizates, it is highly promising that the thermochemically exchange- able ester bond from the EMZ self-healing modifier significantly promoted this self-repa- ration. The possible ester crosslink structure of the modifier is described in Scheme 1. This is because, when the sample was treated at elevated temperatures, particularly at 120 °C, the hysteresis retention showed its highest value (Figure 9), meaning that the elasticity of the material was highly recovered. The reasons for the highly maintained elasticity of a vulcanizate are the rubber network density and a low extent of molecular damages. Treat- ment of the vulcanizates at 120 °C seemed to be crucial for the self-reparation of the mo- lecular damages through a transesterification reaction of the ester crosslinks. It is most likely that, at the interfaces of the molecular damages, which could have been in the form of micro-voids, the ester crosslinks at one interface of the void could rearrange and ex- change the bonding positions to another side of the interface. This generated new ester crosslinks bridging the ruptured interfaces of the void, leading to the micro-damages be- ing repaired. This elucidation is consistent with the findings from previous work [12,14,15]. Scheme 3 graphically illustrates the mechanism of self-reparation of molecular damages through a transesterification reaction of the thermochemically exchangeable es- ter crosslinks.

Polymers 2021, 13, 39 19 of 22

3.6. Potential Role of the Thermochemically Exchangeable Ester Bonds In addition to the self-association of the aforementioned reversible interactions that highly contributed to the intermolecular self-reparation of the molecular damages in the silica-filled NR vulcanizates, it is highly promising that the thermochemically exchangeable ester bond from the EMZ self-healing modifier significantly promoted this self-reparation. The possible ester crosslink structure of the modifier is described in Scheme1. This is because, when the sample was treated at elevated temperatures, particularly at 120 ◦C, the hysteresis retention showed its highest value (Figure9), meaning that the elasticity of the material was highly recovered. The reasons for the highly maintained elasticity of a vulcanizate are the rubber network density and a low extent of molecular damages. Treatment of the vulcanizates at 120 ◦C seemed to be crucial for the self-reparation of the molecular damages through a transesterification reaction of the ester crosslinks. It is most likely that, at the interfaces of the molecular damages, which could have been in the form of micro-voids, the ester crosslinks at one interface of the void could rearrange and exchange the bonding positions to another side of the interface. This generated new ester crosslinks bridging the ruptured interfaces of the void, leading to the micro-damages being repaired. This elucidation is consistent with the findings from previous work [12,14,15]. Scheme3 Polymers 2021, 13, x FOR PEER REVIEWgraphically illustrates the mechanism of self-reparation of molecular damages through21 of 24 a

transesterification reaction of the thermochemically exchangeable ester crosslinks.

Scheme 3. Plausible self-reparationself-reparation mechanism of molecular damages of rubber network through thethe transesterificationtransesterification reaction ofof thethe thermochemicallythermochemically exchangeableexchangeable esterester crosslinks.crosslinks.

4. Conclusions 4. ConclusionsThe mixing torques, mixing temperatures, and compound viscosities indicated that addingThe EMZ mixing contributes torques, to mixing higher temperatures, intermolecular and interactions compound inside viscosities the compounds indicated.that Ex- traadding polymer EMZ networks contributes and to/or highersilica–rubber intermolecular interactions interactions with the insidepresence the of compounds. EMZ were confirmedExtra polymer by the networks results of and/or chemically silica–rubber bound rubber interactions contents with and the apparent presence crosslink of EMZ wereden- sitiesconfirmed. Furthermore, by the results the EMZ of chemicallycould reduce bound the filler rubber–filler contents interactions and apparentof silica clusters, crosslink as indicateddensities. Furthermore,by the lower thePayne EMZ effect could values reduce. The the addition filler–filler of interactionsEMZ via blending of silica with clusters, the primaryas indicated rubber by thehad lower a slightly Payne stronger effect values. impact The on additionthese properties of EMZ than via blending that via withits extra the additionprimary rubberto the compounds, had a slightly attributed stronger impactto a higher on these concentration properties of than EMZ that relative via its to extra the totaladdition rubber to thecontent compounds, for the blending attributed approach to a higher. The strength concentration properties of EMZ and relativereinforcement to the efficienciestotal rubber of content vulcanizates for the blendingwere enhanced approach. with The increasing strength propertiesEMZ content and. reinforcementThe property retentionefficiencies of ofvulcanizates vulcanizates was were significantly enhanced improved with increasing with increasing EMZ content. EMZ The content property and withretention a thermally of vulcanizates annealing was treatment significantly at high improved temperatures, with e increasing.g., 120 °C, EMZ used content as the max- and with a thermally annealing treatment at high temperatures, e.g., 120 ◦C, used as the imum temperature in this study. This was attributed to the EMZ contributing to ester crosslinks and an increased concentration of chemically reversible linkages. The ester bonds could have potentially been interchanged via a transesterification reaction, which enabled the rearrangement of chemical ester crosslinks in the rubber network, thereby leading to self-reparation of the damaged network. This work clearly confirms that the EMZ self-healing modifier significantly improved the property retention of vulcanizates. The intermolecular self-healing mechanism of this modifier requires a thermal treatment and a transesterification catalyst after unloading conditions. Nonetheless, there are still a number of open questions that remain to be explored in order to explicitly elucidate the effects and mechanisms behind these findings. Further proof of the self-reparation of the molecular damages using advanced characterizations is highly desired.

Author Contributions: Conceptualization, W.K. and B.A.; methodology, B.A. and W.K.; micro- structural analysis, B.A.; investigation, W.K. and E.K.; writing—original draft preparation, B.A.; writing—review and editing, W.K., E.K., and S.S.S. All authors read and agreed to the published version of the manuscript. Funding: This research was funded by the Higher Education Research Promotion and Thailand’s Education Hub for the Southern Region of the ASEAN Countries Project Office of the Higher Edu- cation Commission, as well as by the Natural Rubber Innovation Research Institute (Grant No. SCI6201170s), Prince of Songkla University.

Polymers 2021, 13, 39 20 of 22

maximum temperature in this study. This was attributed to the EMZ contributing to ester crosslinks and an increased concentration of chemically reversible linkages. The ester bonds could have potentially been interchanged via a transesterification reaction, which enabled the rearrangement of chemical ester crosslinks in the rubber network, thereby leading to self-reparation of the damaged network. This work clearly confirms that the EMZ self-healing modifier significantly improved the property retention of vulcanizates. The intermolecular self-healing mechanism of this modifier requires a thermal treatment and a transesterification catalyst after unloading conditions. Nonetheless, there are still a number of open questions that remain to be explored in order to explicitly elucidate the effects and mechanisms behind these findings. Further proof of the self-reparation of the molecular damages using advanced characterizations is highly desired.

Author Contributions: Conceptualization, W.K. and B.A.; methodology, B.A. and W.K.; microstruc- tural analysis, B.A.; investigation, W.K. and E.K.; writing—original draft preparation, B.A.; writing— review and editing, W.K., E.K., and S.S.S. All authors read and agreed to the published version of the manuscript. Funding: This research was funded by the Higher Education Research Promotion and Thailand’s Education Hub for the Southern Region of the ASEAN Countries Project Office of the Higher Education Commission, as well as by the Natural Rubber Innovation Research Institute (Grant No. SCI6201170s), Prince of Songkla University. Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

References 1. Awaja, F.; Zhang, S.; Tripathi, M.; Nikiforov, A.; Pugno, N. Cracks, microcracks and fracture in polymer structures: Formation, detection, autonomic repair. Prog. Mater. Sci. 2016, 83, 536–573. [CrossRef] 2. Hernández, M.; Grande, A.M.; Dierkes, W.; Bijleveld, J.; Van der Zwaag, S.; García, S.J. Turning vulcanized natural rubber into a self-healing polymer: Effect of the disulfide/polysulfide ratio. ACS Sustain. Chem. Eng. 2016, 4, 5776–5784. [CrossRef] 3. Forrest, M.J. Introduction to recycling and the re-use of rubber. In Recycling and Re-use of Waste Rubber, 2nd ed.; Forrest, M.J., Ed.; De Gruyter: Berlin, Germany, 2019; pp. 1–8. 4. Coran, A.Y. Vulcanization. In The Science and Technology of Rubber, 4th ed.; Mark, J.E., Erman, B., Roland, C.M., Eds.; Elsevier Academic Press: Boston, MA, USA, 2013; pp. 337–381. 5. Asaro, L.; Gratton, M.; Seghar, S.; Aït Hocine, N. Recycling of rubber wastes by devulcanization. Resour. Conserv. Recycl. 2018, 133, 250–262. [CrossRef] 6. Wemyss, A.M.; Bowen, C.; Plesse, C.; Vancaeyzeele, C.; Nguyen, G.T.M.; Vidal, F.; Wan, C. Dynamic crosslinked rubbers for a green future: A material perspective. Mater. Sci. Eng. R Rep. 2020, 141, 100561. [CrossRef] 7. Das, A.; Sallat, A.; Bohme, F.; Suckow, M.; Basu, D.; Wiessner, S.; Stockelhuber, K.W.; Voit, B.; Heinrich, G. Ionic modification turns commercial rubber into a self-healing material. ACS Appl. Mater. Interfaces 2015, 7, 20623–20630. [CrossRef][PubMed] 8. Hernández Santana, M.; Den Brabander, M.; García, S.; Van der Zwaag, S. Routes to make natural rubber heal: A Review. Polym. Rev. 2018, 58, 585–609. [CrossRef] 9. Zaghdoudi, M.; Kommling, A.; Jaunich, M.; Wolff, D. Scission, cross-linking, and physical relaxation during thermal degradation of elastomers. Polymers 2019, 11, 1280. [CrossRef] 10. Tripathi, R.; Sharma, P.; Saini, A.; Verma, G. Self-healed materials from elastomeric composites: Concepts, strategies and developments. In Smart Polymer Nanocomposites; Ponnamma, D., Sadasivuni, K.K., Cabibihan, J.-J., Al-Maadeed, M.A.-A., Eds.; Springer International Publishing AG: Cham, Switzerland, 2017; pp. 219–242. 11. Xu, C.; Cui, R.; Fu, L.; Lin, B. Recyclable and heat-healable epoxidized natural rubber/bentonite composites. Compos. Sci. Technol. 2018, 167, 421–430. [CrossRef] 12. Imbernon, L.; Norvez, S.; Leibler, L. Stress relaxation and self-adhesion of rubbers with exchangeable links. Macromolecules 2016, 49, 2172–2178. [CrossRef] 13. Gamardella, F.; De la Flor, S.; Ramis, X.; Serra, A. Recyclable poly(thiourethane) vitrimers with high Tg. Influence of the isocyanate structure. React. Funct. Polym. 2020, 151, 104574. [CrossRef] 14. Algaily, B.; Kaewsakul, W.; Kalkornsurapranee, E. Enabling reprocessability of ENR vulcanizates by thermo-chemically ex- changeable crosslinks. In Proceedings of the 10th International Polymer Conference of Thailand, Virtually Organized Conference, Bangkok, Thailand, 6–7 August 2020; pp. 217–220. 15. Algaily, B.; Kaewsakul, W.; Sarkawi, S.S.; Kalkornsurapranee, E. Enabling reprocessability of ENR-based vulcanizates by thermochemically exchangeable ester crosslinks. Plast. Rubber Compos. 2020. under review. Polymers 2021, 13, 39 21 of 22

16. Kaewsakul, W.; Sahakaro, K.; Dierkes, W.K.; Noordermeer, J.W.M. Optimization of mixing conditions for silica-reinforced natural rubber tire tread compounds. Rubber Chem. Technol. 2012, 85, 277–294. [CrossRef] 17. Diani, J.; Fayolle, B.; Gilormini, P. A review on the Mullins effect. Eur. Polym. J. 2009, 45, 601–612. [CrossRef] 18. Flory, P.J.; Rehner, J. Statistical mechanics of cross-linked polymer networks II. swelling. J. Chem. Phys. 1943, 11, 521–526. [CrossRef] 19. Nimpaiboon, A.; Sriring, M.; Rojruthai, P.; Sakdapipanich, J. Improved compatibility between silica and natural rubber by the use of carbonyl-terminated low molecular-weight natural rubber. KGK Rubberpoint 2018, 71, 39–45. 20. Kim, D.Y.; Park, J.W.; Lee, D.Y.; Seo, K.H. Correlation between the crosslink characteristics and mechanical properties of natural rubber compound via accelerators and reinforcement. Polymers 2020, 12, 2020. [CrossRef][PubMed] 21. Kaewsakul, W.; Sahakaro, K.; Dierkes, W.K.; Noordermeer, J.W.M. Factors influencing the flocculation process in silica-reinforced natural rubber compounds. J. Elastomers Plast. 2016, 48, 426–441. [CrossRef] 22. Fu, W.; Wang, L. Research on Payne effect of natural rubber reinforced by graft-modified silica. J. Appl. Polym. Sci. 2016, 133. [CrossRef] 23. Abu-Abdeen, M. Single and double-step stress relaxation and constitutive modeling of viscoelastic behavior of swelled and un-swelled natural rubber loaded with carbon black. Mater. Des. 2010, 31, 2078–2084. [CrossRef] 24. Marckmann, G.; Verron, E.; Gornet, L.; Chagnon, G.; Charrier, P.; Fort, P. A theory of network alteration for the Mullins effect. J. Mech. Phys. Solids 2002, 50, 2011–2028. [CrossRef] 25. Laraba-Abbes, F.; Ienny, P.; Piques, R. A new ‘Tailor-made’ methodology for the mechanical behaviour analysis of rubber-like materials: II. Application to the hyperelastic behaviour characterization of a carbon-black filled natural rubber vulcanizate. Polymer 2003, 44, 821–840. [CrossRef] 26. Hernandez, M.; Grande, A.M.; Van der Zwaag, S.; Garcia, S.J. Monitoring network and interfacial healing processes by broadband dielectric spectroscopy: A case study on natural rubber. ACS Appl. Mater. Interfaces 2016, 8, 10647–10656. [CrossRef][PubMed] 27. Song, F.; Li, Z.; Jia, P.; Zhang, M.; Bo, C.; Feng, G.; Hu, L.; Zhou, Y. Tunable “soft and stiff”, self-healing, recyclable, ther- madapt shape memory biomass polymers based on multiple hydrogen bonds and dynamic imine bonds. J. Mater. Chem. A 2019, 7, 13400–13410. [CrossRef] 28. Sarkawi, S.S.; Dierkes, W.K.; Noordermeer, J.W.M. Elucidation of filler-to-filler and filler-to-rubber interactions in silica-reinforced natural rubber by TEM Network Visualization. Eur. Polym. J. 2014, 54, 118–127. [CrossRef] 29. Sarkawi, S.S.; Kaewsakul, W.; Sahakaro, K.; Dierkes, W.K.; Noordermeer, J.W.M. A Review on reinforcement of natural rubber by silica fillers for use in low-rolling resistance tyres. J. Rubber Res. 2015, 18, 203–233. 30. Ng, S.C.; Gan, L.H. Reaction of natural rubber latex with performic acid. Eur. Polym. J. 1981, 17, 1073–1077. [CrossRef] 31. Kalkornsurapranee, E.; Vennemann, N.; Kummerlowe, C.; Nakason, C. Novel natural rubber based on thermoplastic polyurethane blends: Influence of modified natural rubbers on properties of the blends. Iran. Polym. J. 2012, 21, 689–700. [CrossRef] 32. Sarkawi, S.S.; Aziz, A.K.C.; Rahim, R.A.; Ghani, R.A.; Kamaruddin, A.N. Properties of epoxidized natural rubber tread compound: The hybrid reinforcing effect of silica and silane system. Polym. Polym. Compos. 2016, 24, 775–782. [CrossRef] 33. Manna, A.K.; De, P.P.; Tripathy, D.K.; De, S.K.; Peiffer, D.G. Bonding between precipitated silica and epoxidized natural rubber in the presence of silane coupling agent. J. Appl. Polym. Sci. 1999, 74, 389–398. [CrossRef] 34. George, K.M.; Varkey, J.K.; Thomas, K.T.; Mathew, N.M. Epoxidized natural rubber as a reinforcement modifier for silica-filled . J. Appl. Polym. Sci. 2002, 85, 292–306. [CrossRef] 35. Sadequl, A.M. The effect of accelerator/sulphur ratio on the cure time and torque maximum of epoxidized natural rubber. Int. J. Polym Mater. 2000, 46, 597–615. [CrossRef] 36. Sarkawi, S.S. Nano-Reinforcement of tire Rubbers: Silica-Technology for Natural Rubber. Ph.D. Thesis, University of Twente, Enschede, The Netherlands, 2013. 37. Da Rocha, E.B.D.; Linhares, F.N.; Gabriel, C.F.S.; De Sousa, A.M.F.; Furtado, C.R.G. Stress relaxation of nitrile rubber composites filled with a hybrid metakaolin/carbon black filler under tensile and compressive forces. Appl. Clay Sci. 2018, 151, 181–188. [CrossRef] 38. Tang, Z.; Huang, J.; Guo, B.; Zhang, L.; Liu, F. Bioinspired engineering of sacrificial metal–ligand bonds into elastomers with supramechanical performance and adaptive recovery. Macromolecules 2016, 49, 1781–1789. [CrossRef] 39. Liu, J.; Li, D.; Zhao, X.; Geng, J.; Hua, J.; Wang, X. Buildup of multi-ionic supramolecular network facilitated by in-situ intercalated organic montmorillonite in 1,2-polybutadiene. Polymers 2019, 11, 492. [CrossRef][PubMed] 40. Geethamma, V.G.; Pothen, L.A.; Rhao, B.; Neelakantan, N.R.; Thomas, S. Tensile stress relaxation of short-coir-fiber-reinforced natural rubber composites. J. Appl. Polym. Sci. 2004, 94, 96–104. [CrossRef] 41. Fu, W.; Wang, L.; Huang, J.; Liu, C.; Peng, W.; Xiao, H.; Li, S. Mechanical properties and Mullins effect in natural rubber reinforced by grafted carbon black. Adv. Polym. Tech. 2019, 2019, 4523696. [CrossRef] 42. Kaewsakul, W.; Sahakaro, K.; Dierkes, W.K.; Noordermeer, J.W.M. Mechanistic aspects of silane coupling agents with different functionalities on reinforcement of silica-filled natural rubber compounds. Polym. Eng. Sci. 2015, 55, 836–842. [CrossRef] 43. Kaewsakul, W.; Sahakaro, K.; Dierkes, W.K.; Noordermeer, J.W.M. Cooperative effects of epoxide functional groups on natural rubber and silane coupling agents on reinforcing efficiency of silica. Rubber Chem. Technol. 2014, 87, 291–310. [CrossRef] 44. Poh, B.T.; Lee, K.S. FTIR study of thermal oxidation of ENR. Eur. Polym. J. 1994, 30, 17–23. [CrossRef] Polymers 2021, 13, 39 22 of 22

45. Houwink, R. Slipping of molecules during the deformation of reinforced rubber. Rubber Chem. Technol. 1956, 29, 888–893. [CrossRef] 46. Hanson, D.E.; Hawley, M.; Houlton, R.; Chitanvis, K.; Rae, P.; Orler, E.B.; Wrobleski, D.A. Stress softening experiments in silica-filled polydimethylsiloxane provide insight into a mechanism for the Mullins effect. Polymer 2005, 46, 10989–10995. [CrossRef] 47. Cheng, B.; Lu, X.; Zhou, J.H.; Qin, R.; Yang, Y.L. Dual cross-linked self-healing and recyclable epoxidized natural rubber based on multiple reversible effects. ACS Sustain. Chem. Eng. 2019, 7, 4443–4455. [CrossRef]