Synergistic hybrid reinforcement effects of carbon black,silica,and multiwalled carbon nanotubes in self-healing natural rubber composites

Abstract

This study investigated the influence of hybrid fillers such as carbon black (CB), silica, and multiwalled carbon nanotubes (MWCNTs) on mechanical properties and self-healing capability of the natural rubber (NR) composites based on metal thiolate ionic network. The Taguchi method was employed to optimize the hybrid filler composition using a reduced number of experiments, enabling efficient parameter evaluation, improved reliability of statistical analysis, and reduction in experimental cost and time. It was also used to optimize filler loadings and analyse their effects on key mechanical parameters including tensile strength, elongation at break, tear strength, hardness, compression set and crosslink density. Results suggest that MWCNTs contribute to enhanced mechanical performance through their high aspect ratio and interaction with the rubber matrix, followed by silica and carbon black. ANOVA analysis identified a filler combination of 15 phr CB, 20 phr silica, and 8 phr MWCNTs that represents a trade-off, where improved mechanical performance is achieved at the expense of reduced healing efficiency compared to unfilled NR.

1. Introduction

Materials that can autonomously repair damage and restore functionality are commonly referred to as self-healing materials. They represent a transformative advancement in materials science and engineering. ¹ These systems are designed to detect and respond to damage through intrinsic or extrinsic healing mechanisms, thereby prolonging service life and reducing both downtime and maintenance costs. ² Within the spectrum of self-healing polymers, natural rubber (NR), composed primarily of cis-1,4-polyisoprene, stands out for its remarkable elasticity, high extensibility, and resilience under cyclic stress. Its molecular architecture consists of repeating isoprene units (C₅H₈) with carbon–carbon double bonds, enabling reversible deformation and superior elastic recovery ³ NR exhibits outstanding mechanical adaptability. Owing to its renewable origin, low environmental footprint, and performance versatility, NR is extensively employed in tires, seals, tubing, and vibration-damping systems. ⁴ These products typically undergo vulcanization, a process that introduces covalent crosslinks and significantly enhances their mechanical, thermal, and chemical durability. ⁵

Carbon-based nanofillers such as carbon black, silica, and multiwalled carbon nanotubes (MWCNTs) have been extensively used to enhance the performance of polymer composites including thermosetting resins (epoxy and phenolic) and elastomeric matrices.^6,7 These fillers improve the mechanical strength, stiffness, thermal stability, and electrical conductivity of polymer systems through strong interfacial interactions and efficient stress transfer between the filler and polymer matrix. In elastomers, carbon black and silica are widely used reinforcing fillers that significantly enhance tensile strength, abrasion resistance, and durability. ⁸ Meanwhile, nanofillers such as MWCNTs provide high aspect ratio and large surface area, enabling the formation of interconnected filler networks that improve mechanical performance and functional properties. Recent studies have also reported that carbon-based nanofillers can promote crack bridging and energy dissipation mechanisms, thereby improving the self-healing efficiency and structural stability of polymer composites.

In efforts to introduce self-healing functionality into NR, researchers have explored reversible crosslinking strategies, notably ionic interactions.^9,10 While these dynamic bonds enable autonomous healing, they are often associated with a reduction in mechanical strength compared to covalently crosslinked systems. ⁹ This trade-off presents a challenge in contexts where both damage recovery and load-bearing capacity are essential. Zhang et al., (2021), ¹¹ incorporated CB to BIIR rubber and found that the self-healing capability of the rubber increased with the increase in temperature. 18.3 MPa was found to be the best tensile strength for rubbers with 40% of carbon black. Similarly, L Zhang et al., (2019), ¹² used EPDM rubber and investigated the effect of CB. It was found that the rubber can be recycled more than 3 times with high recovery ratio. Tensile strength was more than 20 MPa. Sallat et al., (2018), ¹³ found that the tensile strength of rubber composites recovered 40% after 16 hours of healing at 70˚C when silica was added to BIIR. The tensile strength of unhealed samples was 18.9 MPa. Another study investigated effects of carbon nanotubes on SBR rubber compounds. It was reported that the self-healing capacity increased efficiently up to 90% after healing at 100˚C for 5 hours and tensile strength and Young’s modulus was enhanced by 3.5 MPa. ¹⁴

Although researchers have investigated the use of individual reinforcing agents such as carbon black or silica, limited studies have explored the synergistic effects of hybrid filler systems that incorporate carbon nanotubes alongside traditional fillers. These multi-scale reinforcements offer potential improvements in stress transfer, energy dissipation, and matrix-filler interaction, yet the optimal formulation strategy for balancing self-healing and reinforcement remains underexplored.

This study aims to address these limitations by developing a novel self-healing NR composite reinforced with a hybrid filler system comprising carbon black, silica, and multi-walled carbon nanotubes (MWCNTs). The aim is to harness the synergistic interaction among these fillers to enhance mechanical integrity without compromising the material’s self-healing performance. A statistically robust optimization approach is adopted using the Taguchi method, complemented by Analysis of Variance (ANOVA), to systematically evaluate the effects of individual and combined filler contributions. The Taguchi method provides a more manageable number of experiments while still identifying the most influential factors and optimizing performance. It is hypothesized that the integration of hybrid fillers will significantly improve the reinforcement efficiency while retaining the autonomous healing behaviour. The expected outcome is an optimized, multifunctional elastomeric system capable of delivering high mechanical performance alongside repeatable self-repair, paving the way for its application in high-demand environments requiring reliability, sustainability, and minimal downtime.

2. Experimental

2.1. Materials

Standard Malaysian Rubber L (SMR L) is used as a natural rubber. Zinc oxide (ZnO) and stearic acid were supplied by Zarm Scientific & Supplies Sdn. Bhd, Malaysia. Zinc Thiolate was used as the healing agent, which is a combination of zinc oxide (ZnO) and thiol functionalized compounds and Dicumyl peroxide (DCP) were supplied by Sigma Aldrich (M) Sdn. Bhd. Malaysia. Carbon black (CB) grade N330 was purchased from Zarm Scientific & Supplies Sdn. Bhd. Malaysia. Silica was supplied by Bayer Co. (M) Sdn. Bhd. Multiwalled carbon nanotubes (MWCNTs), toluene and chloroacetic acid were supplied by Kumpulan Saintifik (KSFE), Malaysia.

2.2. Experimental design

2.2.1. Selection of factor and levels

Taguchi method was used to design the experiment for this work. Three factors were considered: carbon black, silica and carbon nanotubes. These factors were varied at five filler loading levels as shown in Table 1. The level L1 to L5 were chosen based on the previous literature that investigated the effect of filler loadings on material properties.^11,12

Table 1.

Experimental control factors and their respective levels.

Factors (F)	Symbol	Unit	L1	L2	L3	L4	L5
CB (F1)	CB	phr	0	5	10	15	20
Silica (F2)	Si	phr	0	5	10	15	20
Multi-walled Carbon Nanotubes (F3)	MWCNTs	phr	0	2	4	6	8

*Part per hundred rubbers.

Although previous studies often report agglomeration of multiwalled carbon nanotubes at concentrations above ∼2 wt.% due to strong inter-tube interactions, higher loadings were considered in the present work to systematically evaluate their influence within a hybrid filler system consisting of carbon black and silica in a natural rubber matrix. The rubber compounding process and the presence of additional reinforcing fillers assist in improving nanotube dispersion, while the Taguchi experimental design enables identification of the optimal filler composition affecting the healing efficiency of the composites.

2.2.2. Selection of orthogonal array (OA)

Given the three factors and five levels considered in this work, the array L25 was chosen for Taguchi method. The L25 orthogonal array (OA) is shown in Table 2. It’s made up of 25 experiment samples, which correspond to 25 rows and three factors assigned to the respective columns along their levels.

Table 2.

Experimental layout of L25 array according to Taguchi method.

Experimental sample	Factors and their levels
	Carbon black	Silica	Carbon nanotubes
1	1	1	1
2	1	2	2
3	1	3	3
4	1	4	4
5	1	5	5
6	2	1	2
7	2	2	3
8	2	3	4
9	2	4	5
10	2	5	1
11	3	1	3
12	3	2	4
13	3	3	5
14	3	4	1
15	3	5	2
16	4	1	4
17	4	2	5
18	4	3	1
19	4	4	2
20	4	5	3
21	5	1	5
22	5	2	1
23	5	3	2
24	5	4	3
25	5	5	4

2.3. Preparation of master batch

The raw materials for this study were prepared based on the formulation shown in Table 2. It was used to compute composition for 25 compounds according to the experimental layout from Taguchi method. The masterbatch was prepared by mixing 100 phr natural rubber, 5 phr zinc oxide, 1 phr stearic acid and 30 phr of zinc thiolate using internal mixer at 135 °C for 5 minutes (Haake TM). The specific composition, mixing parameters and amount of zinc thiolate is taken from our previous study. ¹⁵

2.4. Rubber compounding

Compounding of the self-healing natural rubber was carried out by using laboratory two roll mill as the formulation given in Table 3.

Table 3.

Formulation of self-healing natural rubber.

Composition (phr)
Ingredient	L1	L2	L3	L4	L5
Natural Rubber (SMR L)	100	100	100	100	100
Zinc Oxide	5	5	5	5	5
Stearic Acid	1	1	1	1	1
Zinc Diethyldithiocarbamate	30	30	30	30	30
Dicumyl Peroxide	1.5	1.5	1.5	1.5	1.5
Carbon Black (F1)	0	5	10	15	20
Silica (F2)	0	5	10	15	20
Multi-walled Carbon Nanotubes (F3)	0	2	4	6	8

The average time for one mixing cycle was around 30 minutes. The sequence of addition of compounding ingredients is given in Table 4.

Table 4.

Sequence of mixing.

No.	Step of mixing	Duration (minutes)	Accumulative (minutes)
1.	The mill opening is set at 0.2 mm and the masterbatch is passed through the rolls twice without banding.	4	4
2.	The rubber is banded on the front roll with nip gap of 1.4 mm. Two ¾ cuts are made from each side.	2	6
3.	The mill opening is increased to 1.7 mm and the zinc oxide is added. Two ¾ cuts are made from each side.	3	9
4.	Stearic acid is added. Few ¾ cut from each side.	2	11
5.	Carbon black, silica and MWCNTs hybrid fillers are added.	3	14
6.	Finally, the dicumyl peroxide is added. Two ¾ cuts are made from each side.	3	17
7.	The stock from the mills is cut. The opening is set at 0.8 mm and the roll stock is passed endwise through the mills 6 times.	5	22
8.	The mills are opened to give a minimum stock thickness of 6 mm and the stock is passed through the roll four times by folding it back in itself each time.	6	28
9	The stock is made to 2.5 mm sheet and then stored in freezer.	2	30

2.5. Cure characteristics

The cure characteristics of the rubber compounds were determined using a Mosanto rheometer (MDR 2000). The rubber compounds were removed from the freezer and left at room temperature for an hour to unfreeze the compounds prior to testing. 4g of sample was prepared for each compound. Each sample was then tested at 150˚C for 30 minutes. The rheographs obtained from the cure characterization were used to determine the curing time for each rubber compound.

2.6. Curing and sample preparation

Hot presses were used to cure the rubber compounds and prepare samples for characterization. Curing temperature, pressure and time are important for the rubber compound to flow and fill the shape of a mould (124mm x 70mm x 1mm). The hot press was initially preheated to 150˚C with the pressure of 1000 psi. The sample of each rubber compound was then weighed based on the requirements for each test and added to the heated mould for curing. The shaped sheets were taken out of the mould after hot pressed and they were ready for additional testing.

3. Characterization

3.1. Tensile test

A tensile test was carried out using an Instron 3366 according to ASTM D412. The tensile properties of dumbbell samples with a 1 mm thickness were assessed at a 50 mm/min crosshead speed at room temperature. The thickness of each sample was determined using a thickness gauge. Three samples were tested for each rubber compound.

First, the sample was stretched in a uniaxial direction until failure to determine the tensile strength and the elongation at break of the samples. After failure, the fracture surfaces were brought into contact with each other with a slight gentle press and then allowed to recover for 10 minutes. Using the information gathered on tensile strength before and after the healing process, the mechanical characterization of self-healing natural rubber was obtained. Equations (1) and (2) were used to calculate the healing effectiveness.

${\text{Healing}\hspace{0.17em}\text{efficiency}}_{\text{TS}}=\frac{{\text{Tensile}\hspace{0.17em}\text{strength}}_{\text{healed}}}{{\text{Tensile}\hspace{0.17em}\text{strength}}_{\text{original}}}\mathrm{X}100\%$

(1)

${\text{Healing}\hspace{0.17em}\text{efficiency}}_{\text{EB}}=\frac{{\text{Elongation}\hspace{0.17em}\text{at}\hspace{0.17em}\text{break}}_{\text{healed}}}{{\text{Elongation}\hspace{0.17em}\text{at}\hspace{0.17em}\text{break}}_{\text{original}}}\mathrm{X}\hspace{0.17em}100\%$

(2)

3.2. Tear test

Tear test was conducted according to ASTM D624 using Instron 3366. Three specimens for each rubber compound were prepared where the vulcanized rubber sheets were cut into trousers test pieces. The thickness of each sample was determined using a thickness gauge. The crosshead speed was adjusted to 50 mm/min at room temperature. The maximum force per unit thickness required to propagate a rip or tear until a totally torn rubber sample achieved is called tear strength, measured in N/mm. Equation (3) was used to compute the tearing energy and equations (4) and (5) were used to calculate the healing effectiveness.

$T=\frac{F}{w}$

(3)

${\text{Healing}\hspace{0.17em}\text{efficiency}}_{\text{Tear}}=\frac{{\text{Tear}\hspace{0.17em}\text{strength}}_{\text{healed}}}{{\text{Tear}\hspace{0.17em}\text{strength}}_{\text{original}}}\mathrm{X}100\%$

(4)

${\text{Healing}\hspace{0.17em}\text{efficiency}}_{\text{EB}}=\frac{{\text{Elongation}\hspace{0.17em}\text{at}\hspace{0.17em}\text{break}}_{\text{healed}}}{{\text{Elongation}\hspace{0.17em}\text{at}\hspace{0.17em}\text{break}}_{\text{original}}}\mathrm{X}\hspace{0.17em}100\%$

(5)

The healing efficiency of the composites was evaluated based on the recovery of tensile strength, elongation at break, and tear strength after the healing process. While tensile strength recovery provides a useful indication of the reformation of the polymer network across the damaged interface, elongation at break and tear strength recovery may also be influenced by factors such as filler loading, crack geometry, and testing conditions. Therefore, these parameters should be interpreted as indicators of the recovery of mechanical performance rather than a direct measure of complete network healing. Nevertheless, such metrics are commonly employed in self-healing elastomer systems to provide a practical comparison of healing performance under mechanical loading conditions.

3.3. Swelling test

A swelling test was conducted to measure the crosslink density of the rubber compounds. The sample for the swelling test was 30 mm × 5 mm × 2 mm. Three samples were prepared for each compound. The samples were immersed in toluene for 72 hours after their initial weight was recorded. The swollen sample’s weight was immediately measured after being drained and dried from the toluene. The swollen samples were then immersed in a chloroacetic acid mixed toluene for 5 days. The samples were again weighed after drained and dried. After that, the samples were soaked in distilled water for 60 minutes, followed by toluene again for about 72 hours and the weight of each sample was measured afterwards. Equation (4) was the Flory-Rehner equation used in this study to determine the total crosslink density.

Use Flory-Rehner equation to obtain the total crosslink density.

$\left[X\right]=\frac{\left[-\mathit{ln}\left(1-{\Phi }_r\right)+{\Phi }_r+\chi {\Phi }_r^2\right]}{V_0\left({\Phi }_r^{1/3}-\frac{{\Phi }_r}{2}\right)}$

${\Phi }_r=\frac{m_1/{\rho }_1}{\left(m_1/{\rho }_1\right)+\left(m_2-m_1\right){\rho }_2}$

(6)

${\mathrm{\Phi }}_{\mathrm{r}}$ = volume fraction of the swollen self-healing rubber,

${\mathrm{V}}_0$ = molar volume of toluene (106.2 cm³),

${\mathrm{m}}_1$ = mass of dry rubber ${\mathrm{m}}_2$ = mass of the swollen rubber.

${\mathrm{\rho }}_1$ is the density of rubber $\left({\mathrm{\rho }}_1=0.910\mathrm{g}/{\text{cm}}^3\right)$

$\mathrm{\rho }2$ is the density of toluene ( ${\mathrm{\rho }}_2=0.865\mathrm{g}/{\text{cm}}^3)$ .

$\mathrm{\chi }$ is the interaction parameter (0.393) of rubber and toluene and $\left[\mathrm{X}\right]$ represents the crosslink density in $\text{mol}/{\text{cm}}^3$ .

To obtain the total covalent crosslink and ionic network formed in the developed materials. The sample was immersed in a mixture of toluene and chloroacetic acid in a ratio (95:5) for 120 hours, followed by 60 minutes in distilled water, and finally another 72 hours in toluene. The equation below was employed to calculate the actual ionic interaction in the sample.

$\left[\mathrm{X}\right]\text{Ionic}=\left[\mathrm{X}\right]\text{total}-\left[\mathrm{X}\right]\text{covalent}$

(7)

Where [X]total is the total crosslink density of the materials, [X]covalent is the covalent crosslink density of the materials, and [X]ionic is the metal ionic crosslink density of the materials.

3.4. Hardness test

The hardness of the compounds was measured using a Rapid Shore A durometer, according to ASTM D2240.

3.5. Compression set

ASTM D395 Compression set testing determines an elastomeric material’s capacity to retain its elastic characteristics after being subjected to sustained compressive stress. It evaluates the permanent deformation of the specimen after it has been subjected to compressive stress about 25% of its original size for a period of 22 hours, revealing the percentage of the original height that the sample recovers after unloading. The compression test was performed at 70˚C. For each sample, three specimens were tested. The initial thickness of each specimen (3-button test piece) for all the compounds was measured before being placed in an oven at 70˚C for 22 hours. The thickness of the specimens of each compound was assessed after 22 hours.

3.6. Scanning Electron Microscopy

The surface of testing specimens that had been cut and joint back was examined for morphology of the joining line using a Hitachi TM3000 Tabletop Scanning Electron Microscope (SEM). Due to the non-conductive nature of rubber, the samples were coated with a tiny layer of gold prior to observation.

3.7. Analysis of signal-to-noise (SN) ratio

The SN ratio is calculated for each response variable (e.g., tensile strength, elongation at break, compression set) based on the type of response and the goal (whether it is larger the better, smaller the better, or nominal the best). The SN ratio for each response variable is calculated based on the observed data for each filler combination using the following formulas.

For Larger-the-Better (e.g., Tensile Strength, Elongation at Break), the formula for SN ratio for larger-the-better responses is

$\text{SN}\hspace{0.17em}\text{ratio}=-10\times {\log }_{10}\hspace{0.17em}\left(\frac{1}{n}{\displaystyle \sum }_{i=1}^n\frac{1}{\left(y_i^2\right)}\right)$

where:

$y_i$ represents the measured value of the response variable (e.g., tensile strength).

$n$ is the number of observations for that response.

A higher SN ratio corresponds to higher values of the response, indicating better performance.

For Smaller-the-Better (e.g., Compression Set), the formula for S/N ratio for smaller-the-better responses is

$\text{SN}\hspace{0.17em}\text{ratio}=-10\times {\log }_{10}\hspace{0.17em}\left(\frac{1}{n}{\displaystyle \sum }_{i=1}^n\frac{1}{\left(y_i^2\right)}\right)$

where:

$y_i$ represents the measured value of the response variable (e.g., compression set).

$n$ is the number of observations for that response.

3.8. Analysis of variance (ANOVA)

The effects of the factors influencing the response will be computed using ANOVA to gather information on the sum of squares, degrees of freedom, Confidence level (%) and percentage of contribution of each factor. Sum of squares (SS), total variability of the observed data which will be calculated using the following equation.

$SS=\frac{{\displaystyle \sum }_{i=0}^n{Xa}^2}{n}-\frac{T^2}{N}$

(8)

where SS is the sum of squares, n is the number of observations, a is the factor, X _a is the sum of observed data at factor a, T = Σ X _a is the sum of all data and N is the total number of data points. Degrees of freedom are the number of levels for each factor that can vary independently and equals to the number of levels of each factor minus one. Statistical significance was evaluated using analysis of variance (ANOVA). Factors with p < 0.05 were considered statistically significant. The percentage of confidence level represents the probability of occurrence and the reliability of the data. The percentage contribution can be calculated as follows.

$Percentage\hspace{0.17em}contribution=\frac{{SS}_{\mathrm{a}}}{SS\mathrm{T}}\times 100\%$

(9)

where SS_a is the sum of the squares of factor and SST is the total sum of squares. The optimal combination of the factors and the factor that contributed the most in the performance of the self-healing natural rubber will be predicted using SN ratio and the information from ANOVA.

4. Results and discussion

4.1. Analysis of Taguchi method

Table 5 shows the data of the healing efficiency for tensile strength, elongation at break, tear strength along with hardness, compression set and total crosslink values obtained from various testing.

Table 5.

Experimental results used to calculate SN and ANOVA.

Compound	Tensile strength healing (%)	Elongation at break healing (%)	Tear strength healing (%)	Hardness (shore A)	Compression set (%)	Total crosslink density (mol/cm3)
1	96.2 ± 1.5	93.2 ± 2.3	112.2 ±1.8	15.3 ± 0.6	7.9 ± 0.4	6.4 ± 0.1
2	54.9 ± 1.2	81.8 ± 1.4	64.8 ± 1.5	32.0 ± 0.7	22.5 ± 0.6	6.8 ± 0.2
3	74.9 ± 1.1	61.9 ± 1.0	86.1 ± 1.2	38.7 ± 0.8	25.9 ± 0.5	7.0 ± 0.3
4	61.6 ± 0.8	65.0 ± 1.2	46.9 ± 0.9	42.8 ± 0.5	33.5 ± 0.7	9.9 ± 0.4
5	40.4 ± 0.7	8.6 ± 0.3	35.4 ± 0.8	44.1 ± 1.0	37.4 ± 0.5	10.1 ± 0.2
6	87.3 ± 1.5	19.7 ± 2.3	65.9 ± 1.2	40.8 ± 0.9	27.0 ± 0.9	8.8 ± 0.3
7	38.6 ± 1.2	70.7 ± 0.5	59.7 ± 1.5	34.2 ± 0.8	21.9 ± 0.8	9.0 ± 0.2
8	50.7 ± 1.4	61.8 ± 1.4	43.2 ± 1.4	44.0 ± 0.9	28.2 ± 0.7	9.2 ± 0.4
9	47.8 ± 1.2	31.9 ± 1.5	43.3 ± 1.2	45.1 ± 0.9	31.2 ± 1.0	9.6 ± 0.3
10	69.1 ± 1.3	31.0 ± 1.0	32.4 ± 1.2	32.2 ± 1.0	18.8 ± 0.6	6.7 ± 0.3
11	54.9 ± 1.2	77.8 ± 1.4	63.9 ± 1.1	41.8 ± 0.7	25.5 ± 0.7	9.8 ± 0.2
12	43.1 ± 1.0	74.2 ± 1.0	57.1 ± 1.4	45.4 ± 0.6	24.2 ± 0.7	10.0 ± 0.2
13	80.2 ± 1.1	44.0 ± 1.2	56.7 ± 1.2	45.0 ± 0.8	31.1 ± 0.6	9.5 ± 0.3
14	64.0 ± 0.8	33.9 ± 0.9	45.0 ± 1.2	29.4 ± 0.6	20.4 ± 0.6	7.5 ± 0.3
15	64.3 ± 1.0	33.0 ± 1.2	27.8 ± 1.0	34.3 ± 0.5	32.3 ± 0.4	8.9 ± 0.2
16	55.6 ± 0.4	49.5 ± 1.2	74.0 ± 0.9	39.2 ± 0.7	32.6 ± 0.7	10.3 ± 0.2
17	62.7 ± 0.4	63.4 ± 0.8	54.6 ± 1.0	39.2 ± 0.9	38.4 ± 0.9	10.1 ± 0.2
18	58.1 ± 1.2	51.9 ± 1.0	48.9 ± 0.8	25.4 ± 0.6	28.8 ± 0.5	7.5 ± 0.2
19	60.9 ± 0.7	36.7 ± 1.0	49.8 ± 1.0	32.4 ± 0.6	28.1 ± 0.6	8.1 ± 0.1
20	68.8 ± 1.2	44.8 ± 1.0	48.6 ± 1.1	38.9 ± 0.7	30.9 ± 0.6	9.8 ± 0.2
21	70.7 ± 0.0	33.5 ± 0.0	75.7 ± 1.4	37.2 ± 0.8	22.6 ± 0.5	9.5 ± 0.3
22	69.3 ± 0.0	45.2 ± 0.7	60.4 ± 1.4	46.0 ± 0.8	36.0 ± 0.6	9.3 ± 0.2
23	42.9 ± 1.0	39.8 ± 0.3	68.9 ± 0.9	35.0 ± 0.6	17.3 ± 0.6	8.7 ± 0.3
24	59.4 ± 0.9	21.4 ± 0.2	34.6 ± 1.4	46.5 ± 1.1	33.8 ± 0.9	11.1 ± 0.2
25	40.8 ± 0.7	25.7 ± 0.3	25.3 ± 0.6	49.3 ± 1.0	34.4 ± 0.7	12.4 ± 0.4

From the table, each value representing an average from the samples tested for each compound and these values were used to calculate the SN and ANOVA. The bolded values showed the highest value obtained for each observation.

4.2. Effect of hybrid fillers on healing efficiency

In this study, Zinc thiolate is selected as healing agents due to their ability to promote dynamic and reversible interactions within the elastomer network. A detail investigation on the self-healing mechanism based on metal thiolate ionic network in the self-healing rubber used in this work previously has been extensively studied and reported in our previous work. ¹⁵ In general, the vulcanized zinc thiolate on rubber chains consists of ion pair between Zn2+ and two S− which is presumed to be responsible for constructing the self-healing network. Interaction among ion pairs induces strong electrostatic interaction between neighbouring Zn2+ ion intermediates and lead to formation of ionic multiplets, which then further aggregate to form ion clusters and restrict the flexibility and mobility of the natural rubber chains. The increased size of the ionic cluster aggregates leads to the formation of a larger ion network, which in turn leads to a higher intermolecular binding capacity between the rubber molecular chains. These ionic networks are reversible and act as physical cross-linking points that generate dynamic systems and induce self-healing mechanisms. Zinc oxide acts as a vulcanization activator and facilitates ionic interactions in rubber systems, while thiol functional groups participate in reversible exchange reactions with sulfur, enabling restoration of the network structure and improving the mechanical properties and in turn the self-healing efficiency of natural rubber composites. ¹⁵

The incorporation of fillers such as carbon black, silica, and multi-walled carbon nanotubes (MWCNTs) leads to the formation of filler networks, occluded rubber, and trapped rubber, which contribute to the improved mechanical strength and stiffness of the material. However, these structures can also restrict polymer chain mobility, particularly in the vicinity of the damaged interface. Since effective self-healing relies on polymer chain diffusion and interfacial rearrangement, the reduced chain mobility may limit the reconstruction of the polymer network after damage. Therefore, the influence of fillers on the self-healing mechanism can be attributed to their hindrance effect on molecular chain rearrangement as well as on the reformation of ionic networks responsible for recovery of the materials. In our previous work, ¹⁵ it was found that the material able to recover 100% after the broken samples were brought into contact with each other for 10 min at ambient temperature and the fractured contact surfaces of the was found to well-adhered to the minor trace line. The intermolecular surface diffusion reconstructed the metal ionic crosslinks to form a reversible ionic network that recovered the material performance. However further work needed to assess the intermolecular diffusion within the recovery area, and the recovery could be represented by the interfacial reattachment rather than genuine molecular reconstruction.

4.2.1. Tensile test

The effect of hybrid fillers, i.e., carbon black (CB), silica, and multiwalled carbon nanotubes (MWCNTs), on the self-healing efficiency of natural rubber (NR) compounds was analysed through tensile tests, with results shown in Figure 1. The Signal-to-Noise (SN) ratio analysis, where larger values indicate better performance, revealed that the unfilled NR sample exhibited the highest healing efficiency. The healing process occurs when the fractured surfaces are brought into contact, enabling ionic interactions between Zn²⁺ ions and two S^- groups from the vulcanized zinc thiolate on the rubber chains. These ion pair interactions generate strong electrostatic attractions between neighbouring Zn²⁺ intermediates, leading to the formation of new network structures. Over time, this chain rearrangement and diffusion restored the material’s original mechanical properties. However, the addition of fillers such as carbon black, silica, and MWCNTs introduce rigid filler networks that restrict polymer chain mobility, thereby limiting the extent of ionic network reconstruction during the healing process.

Figure 1.

Main effect plots for the SN ratios of healing efficiency of the tensile strength: (a) Effect of CB, (b) effect of silica, and (c) effect of multiwalled carbon nanotubes.

The highest healing efficiency among the filler-loaded samples was observed at 15 phr of carbon black, 10 phr of silica, and 2 phr of MWCNTs, suggesting that these filler concentrations optimize both mechanical reinforcement and self-healing capacity. As filler loading increased beyond these optimal levels, healing efficiency decreased, likely due to filler agglomeration that restricts the ionic interactions for effective self-healing. ¹⁶

The reinforcement mechanisms of each filler type contribute differently to the tensile healing efficiency of NR compounds. Carbon black enhances the material through physical adsorption and chemisorption. Its heterogeneous surface, with numerous active sites, interacts with the hydrogen atoms on the rubber chains via Van der Waals forces, leading to reinforcement. ¹⁷ Additionally, chemisorption occurs through hydrogen abstraction from quinonic groups on the CB surface, forming rubber grafts that immobilize rubber segments around the filler and reinforce the composite. ¹⁸ Silica works through hydrogen bonding which enhances its mechanical properties. While silica itself does not form hydrogen bonds in the same way that, for example, water or alcohol molecules do, its surface hydroxyl groups (OH) forms hydrogen bonds with the polar groups present on the NR chains. Specifically, NR contains polar functional groups such as hydroxyl groups (OH) and ether linkages (–O–) in its polyisoprene structure, which forms weak hydrogen bonds with the hydroxyl groups on the surface of silica particles. These interactions help in reinforcing the polymer-filler interface, leading to better dispersion of the silica within the rubber matrix and enhanced mechanical properties. Additionally, Silica is capable of participating in some chemical interactions with peroxides in natural rubber (NR) systems, but this is not the main pathway by which it reinforces the rubber. Its primary reinforcing effect comes from improved filler dispersion and strong interfacial interactions with the NR matrix. While limited chemical bonding between silica and the peroxide may occur, it does not change the fundamental peroxide crosslinking mechanism of NR. Instead, it contributes indirectly by strengthening the filler–rubber interface and enhancing the overall mechanical properties. ¹⁹

Multiwalled carbon nanotubes (MWCNTs), on the other hand, interact with NR through a combination of van der Waals forces, π-π interactions and surface functional groups if present. MWCNTs have a large surface area and high aspect ratio, which allow them to reinforce the rubber matrix by facilitating load transfer and stress distribution. The surface of MWCNTs is typically hydrophobic, but it can be modified to enhance interaction with NR.

In this study, MWCNTs were used without surface functionalization. However, MWCNTs interact with rubber matrices through π-π stacking between the graphene sheets in the MWCNTs structure and the unsaturated bonds in NR. Additionally, if any oxygenated groups (such as COOH, OH, or –C=O) are present on the surface of MWCNTs, they can further interact with NR via hydrogen bonding or covalent bonding. However, in the absence of such functional groups, the interaction is mainly physical and relies on van der Waals forces to promote dispersion and reinforcement. MWCNTs, known for their high aspect ratio and rigidity, provide mechanical reinforcement at low concentrations, but at higher loadings, their tendency to form agglomerates impairs their dispersion and effectiveness in self-healing. ²⁰

Table 6 presents the results of the Analysis of Variance (ANOVA) for the effect of fillers i.e., carbon black (CB), silica, and multiwalled carbon nanotubes (MWCNTs) on the healing efficiency of tensile strength in NR compounds. The analysis revealed that MWCNTs and Silica have significant influence on healing efficiency. These findings highlight the distinct contributions of each filler to the self-healing mechanism in NR compounds.

Table 6.

ANOVA results for the effect of fillers on the tensile strength healing efficiency.

Factor	DF	SS	p-value	Contribution (%)
Carbon Black (phr)	4	226.8	0.44	4.38
Silica (phr)	4	1094.6	0.18	21.15
Multiwalled CNTs (phr)	4	1120.7	0.10	21.65

The relatively greater contribution of MWCNTs and Silica on healing efficiency, suggests that the addition of MWCNT and Silica imposes less hindrance on molecular chain rearrangement and the reformation of ionic networks during the healing process. This may be attributed to MWCNT high aspect ratio and unique nanoscale properties, which promote stronger interactions with the polymer chains and reinforce the material without significantly compromising its self-healing capability. This finding is consistent with the previous work which reported reinforcement properties of MWCNTs, promoting better dispersion and interaction within the rubber matrix. ²¹ The reinforcement mechanism of silica is primarily based on hydrogen bonding and chemical crosslinking with the rubber matrix. Silica increases the crosslink density within the NR compounds, improving the mechanical properties, including tensile strength. Although the particle size of silica is larger than that of MWCNTs, the presence of hydroxyl (OH) groups on the silica surface may facilitate the reformation of ionic bonding between Zn²⁺ and OH^- ions during the healing process. ²²

In contrast, CB exhibits a more modest contribution to healing efficiency. CB primarily enhanced the mechanical properties of the composite through physical adsorption and chemisorption with the rubber matrix. However, its contribution to the self-healing mechanism in which relies on the reformation of ionic networks was limited. This could be explained due to the presence of CB particles hinders the neighbouring ionic interactions between Zn²⁺ and S^- within the broken ionic network. While CB particles increased crosslink density and stiffness, this does not facilitate the effective reformation or re-establishment of broken polymer chains needed for self-healing. Thus, CB reinforcement is effective in improving tensile strength, but it restricts polymer chain movement which reduces the material’s ability to heal autonomously. ²³ The physical interactions between CB and the rubber matrix do not promote the same level of molecular mobility and healing at the fracture sites compared to MWCNTs or silica. Therefore, while CB is effective in reinforcing the composite, its contribution to self-healing efficiency is relatively lower, as indicated by the 56% contribution in Table 6.

Additionally, to investigate the relationship between mechanical performance and self-healing efficiency in the hybrid filler-reinforced NR composites, the tensile strength recovery and healing efficiency for different filler combinations was assessed. While the healing efficiency was inferred from the recovery of mechanical properties such as tensile strength, it is important to consider the effect of filler loading on both the tensile strength and the healing efficiency. Fillers, such as carbon black (CB), silica, and multiwalled carbon nanotubes (MWCNTs), have been shown to enhance the mechanical properties of the composite but also hinder self-healing due to restricted ionic network reconstruction. In the supplementary data, Figure (S)D1, shows a comparison of the original tensile strength, healed tensile strength, and the healing efficiency for various filler combinations. Based on the tensile properties of the samples, it can be analysed that the composite with filled with 5phr of CB, 0 phr of silica and 2 phr of MWCNTs exhibit higher healing efficiency compared to the other filled compounds. The self-healing efficiency of the composite is 87.32%, which indicates that more than 50% of the tensile properties is healed.

The figure clearly demonstrates the trade-off between mechanical reinforcement and self-healing efficiency, where tensile strength improves with increased filler content, while healing efficiency tends to decrease, highlighting the limitations imposed by the fillers on the healing process. The healing efficiency shows a recovery pattern where the healing efficiency initially increases and then levels off as the filler loadings increase. This provides valuable insight into the optimization of filler loadings, which must balance mechanical strength with healing capability for practical applications.

4.2.2. Elongation at break

Figure 2 depicts the influence of hybrid filler loadings on the healing efficiency of elongation at break in NR compounds, with the highest healing efficiency observed at 10 phr carbon black (CB), 5 phr silica, and 6 phr multiwalled carbon nanotubes (MWCNTs) and the values of the original elongation at break, healed and the healing efficiency is presented in supplementary data (Figure (S)D2). These optimal loadings reflect a critical balance between reinforcement and polymer chain mobility essential for effective self-healing. The healing mechanism under extension can be explained by the intermolecular diffusion of reversible ionic bonds and the formation of a new ionic network across the fractured surfaces brought into contact. When the material is stretched, the cross-sectional area decreases, bringing the rubber molecular chains closer together. As the strain increases, the mobile Zn²⁺ ions can re-establish interactions with Zn²⁺ ions in the adjacent network. This promotes chain slippage, allowing greater elongation until the applied strain eventually overcomes the ionic bonds and leads to failure.

Figure 2.

Main effect plots for the SN ratios of healing efficiency of the elongation at break: (a) Effect of CB, (b) effect of silica, and (c) effect of MWCNTs.

For CB, moderate loading (10 phr) improved mechanical integrity while maintaining sufficient polymer flexibility. When CB loading in a rubber matrix is high, it tends to form a filler network that significantly increases the composite’s viscosity. This rise in viscosity is primarily due to hydrodynamic effects associated with the filler particles. ²⁴ The increased viscosity not only makes processing the rubber composite more difficult but also further limits the mobility of the elastomer chains caused by the filler presence. Since the self-healing capability of the rubber depends heavily on the segmental mobility of its polymer chains, these mobility restrictions slow down the kinetics of the healing process. ²⁵ Silica at 5 phr enhanced crosslink density through hydrogen bonding, strengthening the matrix without overly constraining chain mobility, however, at silica loadings exceeding the optimal threshold, particle agglomeration becomes increasingly significant. These aggregated silica clusters contribute to a marked increase in the stiffness of the rubber composite. As a result, the material’s elongation recovery and overall flexibility are adversely impacted, demonstrating a trade-off between reinforcement and elastic performance.^26,27 MWCNTs at 6 phr provide nanoscale reinforcement via crack bridging and load transfer, facilitating self-healing. At higher concentrations, aggregation disrupts uniform stress distribution and limits healing by reducing polymer mobility, consistent with findings that well-dispersed nanofillers enhance healing whereas agglomerates hinder it. ²⁸

This trend highlights the delicate interplay between filler dispersion and polymer dynamics, confirming that optimal filler content maximizes both mechanical strength and healing capacity. Excess filler loading restricts chain mobility, which is crucial for the reformation of polymer networks during healing. Table 7 presents the ANOVA results evaluating the contribution of hybrid fillers to the healing efficiency of elongation at break in NR compounds. Silica shows the highest contribution to healing efficiency, followed by carbon black and MWCNTs.

Table 7.

ANOVA results for the effect of fillers on the healing efficiency of elongation.

Source	DF	SS	p-value	Contribution (%)
Carbon Black (phr)	4	2299	0.044	20.80
Silica (phr)	4	4376	0.004	39.58
Multiwalled CNTs (phr)	4	1459	0.47	13.20

Silica has strong capability to form hydrogen bonds and create physical crosslinks with the rubber chains, enhancing elasticity and flexibility while maintaining network integrity. Its relative uniform dispersion and surface chemistry enable silica to promote polymer chain mobility and facilitate elongation recovery during healing. ²⁹ The present of OH group on silica surface allows it to interact favourably with Zn²⁺ on self-healing rubber chains, which enhances ionic interfacial adhesion without excessively restricting chain dynamics, thus supporting elongation. ³⁰ CB’s considerable but slightly lower contribution is linked to its reinforcing action primarily through physical and chemical adsorption. CB enhances stiffness and tensile strength, yet it tends to reduce polymer chain mobility more than silica due to stronger filler–filler networking and localized rigidity, slightly limiting elongation during the healing process. ³¹ Nonetheless, its wide surface area still contributes to maintaining structural cohesion during elongation recovery, justifying its intermediate confidence level.

In contrast, MWCNTs contribute minimally to elongation healing efficiency despite their known mechanical reinforcement properties. Their high aspect ratio and tendency to form entangled agglomerates create rigid zones within the matrix that significantly impede chain mobility and elasticity, essential for elongation recovery. ³² These agglomerates act as stress concentrators, diminishing the composite’s capacity to stretch and recover after damage, thus lowering MWCNT’s influence on elongation healing efficiency. ³³

Furthermore, the values of elongation at break (EB) decreases with increasing loading of all types of fillers. A decrease in elongation at break is explained in terms of adherence of the filler to the rubber polymer matrix leading to the stiffening of the polymer chain and hence resistance to stretch and restriction to the mobility of the rubber chain when the strain is applied. Therefore, the elongation at break of unfilled compound is higher than the compounds reinforced with CB/Si/MWCNTs hybrid fillers which is shown in supplementary data Figure (S)D2.

The presence of CB, silica and multiwalled carbon inhibit polymer chain movement, decreasing the material’s capacity to stretch and extend before breaking. This decrease in chain mobility and thus resulted in a decrease in elongation at break for filled compounds. Moreover, the filler-filler interactions in the rubber resulted in the formation of filler aggregates or networks that acts as physical barriers. These barriers inhibit the movement of the polymer chains, limiting the total elongation capability of the NR compound.

Generally, higher elongation at break can imply better material flexibility and stretchability prior to failure. A larger elongation at break in the context of self-healing materials can provide more opportunities for the reconstruction of broken polymer chains and the reestablishment of intermolecular connections. This increased chain mobility and polymer chain reassociation may contribute to improved healing efficiency.

4.2.3. Tear strength

Figure 3 illustrates the impact of hybrid filler loadings on the healing efficiency of tear strength in NR compounds, measured via Signal-to-Noise (SN) ratios and the values of the original tear strength, healed tear strength, and the healing efficiency is presented in supplementary data (Figure (S)D3). The data reveals that the optimal healing efficiency for tear strength occurs at 15 phr for CB, 10 phr for silica, and 2 phr for MWCNTs, reflecting distinct reinforcing roles and mechanisms for each filler type. In the tear failure mechanism, the region near the crack tip undergoes significant deformation under stress. When the tearing energy exceeds a critical level, both covalent bonds and ionic networks begin to break down. During healing, when the ruptured surfaces are brought back into contact, the remaining ionic networks at the crack tip dissociate into smaller aggregates that are not completely damaged. The presence of ionic dangling chains and brush-like structures formed during tearing facilitates reaggregation and reassociation of the ionic networks, thereby enhancing crack resistance and heal the tear damage.

Figure 3.

Main effect plots for the SN ratios of healing efficiency of the tear strength: (a) Effect of CB, (b) effect of silica, and (c) effect of MWCNTs.

CB’s superior performance at 15 phr can be attributed to its well-known ability to form strong filler–polymer networks that effectively dissipate stress and hinder crack propagation, critical factors for tear resistance. ³⁴ The particle morphology and surface chemistry of CB facilitate strong interfacial adhesion with rubber chains, enhancing the energy required to propagate tears. At this loading, CB achieves an optimal balance between reinforcement and dispersion, providing a robust network that improves tear strength and its healing efficiency. Exceeding this concentration often leads to filler agglomeration, which can create stress concentration points and reduce toughness, but 15 phr remains within the effective reinforcement range.

Silica’s peak healing efficiency at 10 phr results from its capacity to enhance crosslink density via hydrogen bonding and chemical interactions with the rubber matrix, improving the composite’s resistance to crack growth. ³⁵ Silica’s smaller particle size and surface polarity allow it to interact strongly with polymer chains, contributing to improved tear strength by restricting polymer chain slippage and stabilizing crack tips. The reduction in healing efficiency beyond 10 phr can be linked to increased rigidity and filler aggregation, which impair the polymer’s ability to recover post-tear.

MWCNTs exhibit optimal healing efficiency for tear strength at a low loading of 2 phr. This is due to their high aspect ratio and nanoscale reinforcement, which promote effective stress transfer and crack bridging at low concentrations. ³⁶ At higher loadings, however, MWCNTs tend to agglomerate, disrupting the homogeneity of the rubber matrix and leading to localized stress concentrations that facilitate crack initiation rather than suppression. Thus, 2 phr represents the threshold at which MWCNTs provide maximal reinforcement without compromising the matrix integrity or healing potential.

Overall, CB with its extensive surface activity and network-forming capability, excels in resisting tear propagation at moderate loadings. Silica’s contribution centers on enhancing the crosslink density and interfacial bonding, while MWCNTs deliver nanoscale reinforcement that is highly effective only when well dispersed at low concentrations.

Table 8 presents the ANOVA results assessing the contribution of hybrid fillers to the healing efficiency of tear strength in NR compounds. Silica exhibits the highest contribution to the self-healing under tear failure mode. This could again be explained due to the present of OH group on silica surface allowing it to interact favorably with Zn2+ on self-healing rubber chains, which enhances ionic interfacial adhesion and reformation of ionic crosslink network. ³⁷

Table 8.

ANOVA results for the effect of fillers on the tear strength healing efficiency.

Source	DF	SS	p-value	Contribution (%)
Carbon Black (phr)	4	1317.9	0.14	14.54
Silica (phr)	4	5834.7	0.001	64.35
Multiwalled CNTs (phr)	4	357.8	0.27	3.95

The tear test assesses a material’s capability to resist crack initiation and propagation under strain. At optimal levels, filler incorporation enhances matrix–filler interactions, suppressing crack growth and improving tear strength. Beyond this optimum, however, further filler addition leads to a decline in tear performance. This is attributed to inadequate dispersion and wetting of the filler by the rubber phase, which impairs stress transfer and results in reduced tensile strength.

As shown in supplementary data Figure (S)D3, the tear strength of all compounds decreases after healing. This reduction is likely due to incomplete restoration of the molecular network, insufficient bonding at the healed interface, or the formation of healing-induced defects that act as stress concentrators. The presence of CB/Si/MWCNTs hybrid fillers likely limits effective filler–rubber interactions, thereby constraining the healing response.

Among the tested materials, the unfilled NR compound exhibits the highest healing efficiency. This can be attributed to the unrestricted mobility of polymer chains, which promotes interdiffusion and network reconstruction during healing. In contrast, filled compounds show reduced healing efficiency, particularly at higher filler loadings, due to increased crosslink density and restricted chain mobility, both of which impede the reformation of entanglements and crosslinks across the damaged interface.^38,39

4.3. Effect of hybrid fillers on total crosslink

Figure 4 illustrates the effect of increasing loadings of carbon black (CB), silica, and multiwalled carbon nanotubes (MWCNTs) on the total crosslink density of NR compounds, measured through Signal-to-Noise (SN) ratios. The data shows a clear upward trend, with the highest SN ratios at 20 phr for both CB and silica, and at 6 phr for MWCNTs, indicating that higher filler content enhances the crosslink density within the composite.

Figure 4.

Main effect plots for the SN ratios of the crosslink density: (a) Effect of CB, (b) effect of silica, and (c) effect of MWCNTs.

The increase in crosslink density with CB loading up to 20 phr can be explained by its extensive surface area and chemically active sites, which promote strong physical and chemical interactions with rubber chains. These interactions facilitate additional crosslinking points, effectively restricting polymer chain mobility and increasing network density. ⁴⁰ This densification of the polymer network improves the mechanical strength and thermal stability of the composite, justifying the significant contribution of CB at high loadings.

Similarly, silica at 20 phr contributes to the enhanced crosslink density through its polar surface groups that form hydrogen bonds and covalent linkages with the rubber matrix. The presence of abundant silanol groups on silica surfaces fosters strong filler–matrix adhesion, promoting more crosslink points and reinforcing the network structure. ⁴¹ As filler content increases, these interactions intensify, leading to the observed rising trend in crosslink density. For MWCNTs, the peak effect at 6 phr corresponds to an optimal dispersion state, where nanotubes interact effectively with the polymer matrix, acting as nano-scale crosslinking agents that bridge polymer chains and enhance the network integrity. ⁴² Beyond this concentration, agglomeration may limit further increases in crosslink density, explaining the plateau observed at higher loadings.

The overall ascending trend in crosslink density with increasing filler loadings arises from the progressive establishment of physical and chemical interactions between fillers and rubber chains. These interactions restrict polymer mobility and create a more tightly crosslinked network, which enhances the composite’s mechanical properties and durability. ⁴³ Such reinforcement mechanisms are critical in balancing strength and elasticity in self-healing rubber systems.

Table 9 presents the ANOVA results for the contributions of multiwalled carbon nanotubes (MWCNTs), carbon black (CB), and silica to the total crosslink density in NR compounds. It can be seen that MWCNTs demonstrate the highest contribution, followed by Carbon Black and Silica.

Table 9.

ANOVA results for the effect of fillers on the total crosslink.

Source	DF	SS	p-value	Contribution (%)
Carbon Black (phr)	4	12.581	0.01	24.86
Silica (phr)	4	4.017	0.25	7.94
Multiwalled CNTs (phr)	4	26.965	0.001	53.28

Incorporation of MWCNT’s at small amount compared to others types of filler has enough pronounced effect on total crosslink density arises from its exceptional surface area and tubular nanostructure, which facilitate strong physical entanglements and potential chemical interactions with polymer chains. These features enable MWCNTs to act as without much interruption on reformation of ionic molecular network during healing process. ⁴⁴ The physical adsorption and chemisorption mechanisms reinforcement mechanism of carbon black immobilize rubber segments around filler particles and limit the reformation of the ionic bonding during healing process. ⁴⁵ The lower contribution percentage reflects the difference in morphology, spherical CB particles. While.

Silica exhibits the lowest contribution to total crosslink density despite its chemical activity. Its reinforcement mainly depends on hydrogen bonding and filler–matrix adhesion rather than acting as an effective crosslinking bridge within the polymer network. ⁴⁶ The lower surface compatibility and tendency for agglomeration at higher loadings reduce silica’s effectiveness in enhancing network connectivity compared to the other fillers.

Collectively, these findings indicate that MWCNT’s unique nanostructure and interaction capability make it the most effective filler for increasing crosslink density in NR compounds. CB provides moderate enhancement through surface interactions, whereas silica’s contribution is limited by its surface chemistry and dispersion challenges.

4.4. Effect of hybrid fillers on hardness

Figure 5 demonstrates the effect of carbon black (CB), silica, and multiwalled carbon nanotubes (MWCNTs) loadings on the hardness of NR compounds, with the Signal-to-Noise (SN) ratio increasing consistently as filler content rises. The highest SN ratios are observed at 20 phr for both CB and silica, and at 6 phr for MWCNTs, reflecting the progressive stiffening of the composite with increased filler incorporation.

Figure 5.

Main effect plots for the SN ratios of hardness: (a) Effect of CB, (b) effect of silica, and (c) effect of MWCNTs.

The pronounced influence of CB at 20 phr on hardness can be attributed to its well-known reinforcing nature, where its large surface area and aggregated structure create a rigid filler network that significantly restricts polymer chain mobility. ⁴⁷ This network formation effectively increases the resistance to deformation, thereby enhancing hardness. At higher loadings, CB particles form interconnected clusters that amplify stiffness and improve load-bearing capacity, which is critical for hardness improvement in elastomer composites.

Silica’s similar optimal loading at 20 phr relates to its ability to establish strong hydrogen bonding and covalent interactions with the rubber matrix, thereby elevating crosslink density and restricting chain flexibility. ⁴⁸ The polar surface chemistry of silica facilitates firm interfacial adhesion, which contributes to increased hardness by reducing polymer chain mobility and enabling the material to better resist indentation and deformation. MWCNTs reach their peak effect on hardness at 6 phr due to their high aspect ratio and excellent load transfer capabilities at relatively low concentrations. ⁴⁹ At this loading, MWCNTs are well-dispersed within the matrix, forming an effective reinforcing network that enhances stiffness. Beyond 6 phr, agglomeration tends to occur, diminishing their reinforcing efficiency and limiting further hardness improvement.

The upward trend in hardness with increasing filler loading is fundamentally linked to the progressive restriction of polymer chain mobility caused by the filler–matrix interactions and filler networking. As filler content increases, the composite’s microstructure becomes more constrained, resulting in greater resistance to deformation. Notably, the optimum filler loadings for hardness correspond to higher SN ratios compared to those for healing efficiency of tensile strength, elongation at break, and tear strength. This disparity arises because hardness primarily depends on stiffness and resistance to surface deformation, which increase with filler content, whereas healing efficiency requires a delicate balance between reinforcement and chain mobility, often compromised at very high filler loadings. ⁵⁰

In essence, the enhanced hardness with increasing filler loadings reflects the transition from a flexible rubber matrix to a more rigid, filler-dominated composite structure. This transformation underlines the importance of filler selection and concentration in tailoring mechanical properties for specific applications where surface hardness is critical.

Table 10 reveals that multiwalled carbon nanotubes (MWCNTs) have largest contribution of 48.25% to the hardness of NR compounds. Carbon black (CB) follows with 16.29 %, while silica shows the lowest 5.78% percent contribution.

Table 10.

ANOVA results for the effect of fillers on hardness.

Source	DF	SS	p-value	Contribution (%)
Carbon Black (phr)	4	231.92	0.16	16.29
Silica (phr)	4	82.22	0.26	5.78
Multiwalled CNTs (phr)	4	686.73	0.001	48.25

Properties of MWCNTs arise from its high aspect ratio and nanoscale dimensions, enabling the formation of an interconnected rigid network within the polymer matrix that drastically restricts chain mobility and enhances load transfer. ⁵¹ This network effect translates into significant stiffness and increased hardness. The extensive surface area of well-dispersed MWCNTs facilitate strong interfacial adhesion, contributing to efficient stress transfer and reinforcing the composite beyond what particulate fillers can achieve.

Carbon black provides notable reinforcement by immobilizing polymer chains through physical and chemical adsorption, enhancing hardness but to a lesser degree than MWCNTs due to its spherical morphology and relatively lower aspect ratio. ⁵² This results in a substantial but secondary contribution to hardness.

Silica’s minimal contribution is linked to a smaller surface area and a predominant role in chemical crosslinking rather than direct mechanical reinforcement. While silica improves crosslink density and thermal stability, limited ability to form a rigid filler network reduces its impact on hardness compared to MWCNTs and CB. ⁵³ Additionally, silica’s tendency to agglomerate at higher loadings may impair reinforcing efficiency.

In summary, hardness of NR compounds is primarily governed by the ability of fillers to create a rigid, well-dispersed network that restricts polymer chain motion. The nanostructure and network formation of MWCNTs make it the most effective hardness enhancer, followed by the aggregated particulate reinforcement of CB, with silica playing a lesser mechanical role.

4.5. Effect of hybrid fillers on compression set

Figure 6 shows the effect of hybrid filler loadings i.e., carbon black (CB), silica, and multiwalled carbon nanotubes (MWCNT), on the compression set of NR compounds, measured through Signal-to-Noise (SN) ratios, where smaller values are better. The highest SN ratios are observed at 15 phr for CB, 20 phr for silica, and 8 phr for MWCNTs, indicating the optimal loading of each filler for minimizing compression set. The observed SN ratio for compression set is negative. Generally, for larger-the-better responses (e.g., tensile strength, elongation at break), the SN ratio increases with increasing values of the measured property. For smaller-the-better responses (e.g., compression set, where lower values are preferred), the SN ratio decreases as the measured value increases. This is because compression set refers to the permanent deformation after compression, and a lower compression set (i.e., less permanent deformation) is desirable. Thus, when the compression set increases (i.e., the material does not recover as well), the SN ratio becomes negative due to the smaller-the-better characteristic. Therefore, the negative S/N ratio observed for compression set is not an error but rather reflects the nature of the “smaller-the-better” characteristic in this property. A higher S/N ratio for compression set corresponds to a lower compression set which means that the material recovers more effectively from compression and exhibits less permanent deformation.

Figure 6.

Main effect plots for the SN ratios of compression set: (a) Effect of CB, (b) effect of silica, and (c) effect of MWCNTs.

In comparison to other properties such as tensile strength, elongation at break, and tear strength (all of which are larger-the-better properties), the SN ratio behaves in the opposite manner. For these properties, a higher SN ratio corresponds to higher values of the measured property (e.g., higher tensile strength or greater elongation at break), as increased values are desired. This means that the composite’s mechanical properties improve as the filler loadings are optimized. In contrast, the compression set follows the smaller-the-better approach, where the goal is to minimize permanent deformation. As such, a lower compression set results in a higher SN ratio. When the compression set increases (i.e., the material does not recover well), the SN ratio becomes negative, indicating poorer performance.

For CB, the peak SN ratio at 15 phr suggests that this concentration is most effective in enhancing the rubber’s ability to recover from compressive stress. CB contributes to compression set by forming a rigid network within the polymer matrix, which enhances its resistance to permanent deformation under stress. ⁵⁴ At 15 phr, CB particles create a balanced network that imparts stiffness to the composite, yet allows enough flexibility for the material to recover after compression. Higher loadings tend to create filler agglomeration, reducing the material’s ability to revert to its original shape effectively, hence worsening the compression set.

Silica shows the highest SN ratio at 20 phr, indicating that its contribution to reducing compression set is maximized at this loading. Silica reinforces the matrix through hydrogen bonding and chemical crosslinking, increasing the rubber’s resistance to permanent deformation under compression. ³⁵ The increased crosslink density at 20 phr enhances the material’s elasticity, helping it to recover more effectively after being compressed. However, at higher silica concentrations, the material becomes more rigid, potentially impeding full recovery and increasing the compression set, as the matrix becomes more rigid and less able to deform and recover.

MWCNTs reach its peak effectiveness at 8 phr due to its unique reinforcing properties. The high aspect ratio and stiffness of MWCNTs contribute to the composite’s rigidity, which helps resist permanent deformation under stress. ⁵⁵ At lower loadings, the MWCNTs are well-dispersed and form an effective reinforcing network that enhances the composite’s ability to recover. However, at higher concentrations, MWCNTs tend to agglomerate, which disrupts the uniform distribution and limits their contribution to improving recovery, hence the lower contribution at higher loadings. The fillers contribute differently to compression set. CB is most effective at 15 phr due to its balance between reinforcement and flexibility. Silica provides the best performance at 20 phr by enhancing crosslink density, while MWCNTs are most effective at 8 phr, with higher concentrations causing agglomeration and reducing recovery efficiency.

Table 11 presents the ANOVA results for the contribution of multiwalled carbon nanotubes (MWCNT), carbon black (CB), and silica to the compression set of NR compounds. The data reveals that MWCNTs contribute 25.29% to the compression set. This is followed by silica at 14.91%, with CB showing the least contribution at 11.95%.

Table 11.

ANOVA results for the effect of fillers on the compression set.

Source	DF	SS	p-value	Contribution (%)
Carbon Black (phr)	4	144.9	0.12	11.95
Silica (phr)	4	180.7	0.06	14.91
Multiwalled CNTs (phr)	4	306.6	0.01	25.29

MWCNT’s dominant role in enhancing compression set can be attributed to its unique nano structural properties, including its high aspect ratio and high surface area. These characteristics enable MWCNTs to form an interconnected network within the rubber matrix, providing significant resistance to permanent deformation. MWCNTs’ ability to bridge polymer chains and distribute stress uniformly results in higher stiffness and reduced compression set. At optimal loading (6 phr), the dispersion of MWCNTs is sufficient to create an effective reinforcing network, while avoiding agglomeration, which could impede its performance. ⁵⁶ This ability to form a rigid, well-dispersed network enhances the material’s recovery after compressive forces, leading to a significant reduction in compression set.

Silica, with a contribution of 14.91%, plays an important role in enhancing the rubber’s resistance to permanent deformation, primarily through chemical crosslinking and hydrogen bonding. Silica’s surface chemistry promotes strong interactions with the polymer matrix, increasing the crosslink density and thereby improving the stiffness and recovery of the material. ⁵⁷ However, its contribution to compression set is secondary to MWCNTs, as silica does not form the same interconnected, rigid network that MWCNTs create. Silica’s role is more focused on enhancing the material’s mechanical properties, such as tensile strength and elasticity, but its influence on reducing compression set is limited compared to MWCNTs.

CB contributes the least to compression set reduction, with an 11.95% contribution. While CB provides mechanical reinforcement to NR through physical adsorption and chemisorption, its contribution to the compression set is minimal. The spherical morphology of CB particles limits their ability to form a rigid, interconnected network within the rubber matrix, which reduces their effectiveness in preventing permanent deformation. CB primarily enhances the composite’s mechanical strength and toughness but does not significantly improve the material’s ability to recover after compression. ⁵⁸

MWCNTs are the most effective filler for reducing compression set in NR compounds due to their nanoscale properties, which enable efficient network formation and stress distribution. Silica follows with moderate influence, enhancing crosslink density and stiffness but contributing less to compression set reduction. CB’s contribution is the smallest, as it primarily reinforces the rubber matrix without significantly improving the ability to recover from deformation.

4.6. Morphological characteristics

The self-healed surfaces of unfiller NR and NR containing hybrid fillers of carbon black (CB), silica, and multiwalled carbon nanotubes (MWCNTs) were examined using Scanning Electron Microscopy to assess morphological features after healing. The selected formulation (5 phr CB, 15 phr silica, and 8 phr MWCNTs) represents moderate filler loadings, where reinforcement effects are expected while still allowing observation of potential influences on the healing process.

For the unfilled NR sample (Figure 7(a) and (b)), the SEM images show a relatively continuous interface at the rejoined fracture surfaces. The contact between the surfaces appears more uniform, with fewer visible gaps, suggesting that polymer chains are able to reestablish interfacial contact after damage. This observation is consistent with previous reports that unfilled NR exhibits effective recovery behavior due to its inherent chain mobility and ionic crosslinking network. ⁵⁹ However, it should be noted that such morphological continuity does not directly confirm complete molecular-level healing but but rather reflects good interfacial adhesion and reattachment at the healed fracture surface.

Figure 7.

Healed surface morphology: (a) Unfilled sample before healing, (b) healed unfilled sample (c) CB/Si/MWCNTs hybrid fillers filled sample before healing, (d) healed CB/Si/MWCNTs hybrid fillers filled sample.

In contrast, the filler-loaded NR sample (Figure 7(c) and (d)) displays a more heterogeneous and irregular interface after healing. The presence of visible gaps and discontinuities indicates less uniform interfacial contact between the fractured surfaces. These features suggest that the incorporation of hybrid fillers may limit chain mobility and interfacial rearrangement during the healing process. ⁶⁰

4.7. Filler interactions and dispersion mechanism

The interactions between MWCNTs, CB, and silica play a crucial role in the final network structure of the composite material. The MWCNTs, due to their high aspect ratio and nanoscale dimensions, are particularly effective in improving the dispersion of carbon black and silica within the rubber matrix. This improvement in dispersion arises from the strong interfacial interactions between the MWCNTs and the polymer matrix, which enhances the interfacial compatibility between the different fillers. MWCNTs help reduce the tendency of CB to agglomerate. CB particles are typically prone to aggregation due to their large surface area and hydrophobic nature. However, the high surface area of MWCNTs allow them to interact with CB particles, preventing their aggregation. This interaction enhances the distribution of CB throughout the matrix, resulting in improved mechanical properties such as tensile strength and hardness. Silica particles, which have a polar surface, tend to form agglomerates at higher concentrations. The MWCNTs enhance the dispersion of silica in the matrix by facilitating stronger interfacial interactions through hydrogen bonding or π-π stacking interactions between the surface of MWCNTs and silica. This leads to a more homogeneous distribution of silica within the composite, improving crosslink density and tear strength.

Carbon black (CB), a commonly used filler in rubber composites, is known to provide reinforcement through physical networks. However, silica tends to form agglomerates at higher concentrations due to its polar surface. CB can help mitigate this behavior by inhibiting or reducing agglomeration through van der Waals interactions. The high surface area of CB allows it to interact with silica particles, improving the dispersion of silica in the rubber matrix. This enhanced dispersion of silica allows for better network formation and crosslinking, which improves mechanical properties like tensile strength and tear strength. The synergistic effect of MWCNTs, CB, and silica significantly influences the final network structure of the composite. MWCNTs, being nanoscale fillers, promote the formation of a three-dimensional network by providing crack-bridging and load transfer at low concentrations. Meanwhile, CB and silica contribute to network formation by enhancing the crosslink density and interfacial bonding between the polymer matrix and fillers. This composite network enhances the mechanical properties such as tensile strength, hardness, and tear strength. The synergistic effect of CB, silica, and MWCNTs was observed in their combined influence on both the mechanical performance and self-healing efficiency of the NR composite. The interaction of these three fillers leads to an enhancement that is greater than the sum of their individual contributions.

It should be noted that different response parameters exhibited different optimal filler combinations during the statistical analysis. This variation arises because mechanical properties and healing efficiencies are influenced by multiple competing mechanisms, including reinforcement effects, filler dispersion and polymer chain mobility. As a result, the filler composition that maximizes one property may not necessarily provide the optimal response for another. Therefore, the final optimized formulation represents a balanced compromise between mechanical reinforcement and healing performance rather than a single parameter-specific optimum.

Finally, the trade-off between the healing efficiency and filler loading can be understood by experimental data. The healing efficiency of tensile strength and elongation at break decreases with increasing filler loading, particularly at higher concentrations of MWCNTs. The mechanical properties are enhanced up to a certain filler loading, after which the self-healing efficiency is compromised due to the restricted polymer mobility caused by the filler agglomeration. For tensile strength, the healing efficiency is highest at 5 phr CB, 2 phr MWCNTs, but drops significantly as the filler loading increases to 20 phr CB, suggesting that higher filler content restricts healing efficiency. For elongation at break, the self-healing efficiency is highest at 10 phr CB, 5 phr silica, 6 phr MWCNTs, but decreases as the filler content increases, particularly at higher silica loadings.

Furthermore, the crosslink density data indicate that increasing filler loadings leads to higher crosslink density, which further limits chain mobility. This phenomenon is explained by the filler-particle interactions that create a more rigid network within the rubber matrix, making it harder for the polymer chains to rearrange and re-bond during the healing process. Additionally, the SEM images of the fracture surfaces show that the presence of fillers (particularly MWCNTs) creates agglomerates that hinder healing at the fracture sites. The gap between the fractured surfaces is visible in the images, indicating that complete healing was not achieved. The poor dispersion of MWCNTs in the rubber matrix leads to stress concentration zones, which may further prevent effective healing.

To better understand the performance of self-healing natural rubber composites, Table 12 presents various studies on self-healing rubber systems. This highlights the unique aspects of hybrid filler systems, including the use of carbon black (CB), silica, and multiwalled carbon nanotubes (MWCNTs).

Table 12.

Studies on self-healing rubber filled with various fillers.

Rubber	Filler	Self-healing properties and conditions	Tensile strength	References
NR	CB	NR in presence of 5 phr CB and zinc thiolate was recycled for 3 times	Tensile strength was increased by more than 430% for unfilled rubber and 520% for NR/5CB composites after the third recycling process	61
NR	Carbon Black	Addition of CB decreased the self-healing efficiency	Tensile strength increased after addition of CB	62
NR	Silica	Healing efficiency is 63.49 at 5 phr silica loading	Tensile strength of filled sample up to 4.22 MPa.	63
NR	Carbon Nanotubes	Self-healing efficiently up to 88% at 2 phr MWCNT	Tensile strength increased by 3 times compared to unfilled samples	64
Natural Rubber	Carbon Nanotubes	Mixer blending and use of Mastermac thermopress with a maximum pressure of 210 kgf cm−2	Mechanical and electrical properties improved after addition of MWCNT	65

5. Conclusion

This study examined the impact of hybrid fillers such as carbon black (CB), silica, and multiwalled carbon nanotubes (MWCNTs), on the mechanical and self-healing properties of NR compounds. The results indicate that the self-healing behaviour of the natural rubber system is primarily governed by reversible zinc–thiolate interactions, whereas the incorporation of hybrid fillers mainly contributes to mechanical reinforcement. The presence of fillers introduces a trade-off between improved mechanical strength and reduced healing efficiency due to restricted polymer chain mobility. The conclusion can be made that CB significantly enhances tensile strength and tear strength due to its reinforcing effect, while silica improves crosslink density and toughness while also aiding in the better dispersion of MWCNTs. MWCNTs, with their high aspect ratio, contribute to nanoscale reinforcement, improving the material’s flexibility and elasticity. Through Taguchi optimization, we identified the optimal filler combination of 15 phr CB, 20 phr silica, and 8 phr MWCNTs. This combination achieves a balance between enhanced mechanical reinforcement and self-healing efficiency, with the synergistic interaction between the fillers leading to overall improved material performance. The incorporation of MWCNTs contributes to mechanical reinforcement and healing behaviour, however, the ANOVA results indicate that their influence varies depending on the response parameter, with other formulation variables also affecting elongation and tear healing performance. The use of Taguchi optimization proved to be an effective tool in identifying the optimal filler composition, providing a robust method for improving both performance characteristics simultaneously. While this study provides significant insights into the self-healing behavior of NR composites, several limitations exist. The study primarily focuses on single-cycle healing and does not explore healing kinetics or multi-cycle durability, which are crucial for evaluating the long-term performance of the material. Future research should address these aspects by incorporating multi-cycle healing tests and exploring the influence of complex noise factors, such as temperature and humidity on the healing process. Additionally, further exploration into alternative nanofillers could yield even more efficient self-healing systems. The developed natural rubber composites reinforced with hybrid nanofillers demonstrate promising potential for applications in automotive components, flexible electronics, vibration-damping systems, and protective elastomeric coatings. The enhanced mechanical performance combined with self-healing capability can contribute to improved durability and extended service life of elastomer-based engineering materials. In conclusion, this work represents a promising step forward in the development of self-healing rubber composites, with the potential for real-world applications that require both durability and repairability.

Supplemental material

Supplemental material - Synergistic hybrid reinforcement effects of carbon black, silica, and multiwalled carbon nanotubes in self-healing natural rubber composites

Supplemental material for Synergistic hybrid reinforcement effects of carbon black, silica, and multiwalled carbon nanotubes in self-healing natural rubber composites by Bhushan Hajare, Subramanian Radhakrishnan, Bazli Hilmi, Mimi Syahira, Noor Faezah Mohd Sani and Raa Khimi Shuib in Polymers and Polymer Composites.