Reuse of Activated Carbon Filter Waste as Filler in Vulcanized Rubber Composites

Abstract

The incorporation of residues into rubber composites has gained attention as a sustainable strategy to address waste management challenges while replacing commercial fillers. In this study, we investigated the potential use of water filter cartridge residue after exhaustion, composed of activated carbon, as a reinforcing filler in vulcanized natural rubber composites. Samples were prepared with 5, 10, 15, and 20 phr (per hundred rubber) of residue and compared to unfilled natural rubber. Stress vs. strain tests reached 13.9 MPa of tension at rupture for composites containing 10 phr of carbon-activated residues, representing a 21.9% increase compared to natural rubber. Interestingly, the tension at rupture for NR/AC_10phr reached values close to those of NR/CB_5phr (with carbon black N330) attaining 14.4 MPa. These results indicate that, even at relatively low concentrations, the carbon filter can offer partial substitution for commercial fillers. Moreover, the use of activated carbon from filter cartridges as filler in rubber composites provides an environmentally favorable alternative to energy-intensive regeneration processes for activated carbon.

1. Introduction

Purification of water contaminated by organic compounds is essential to maintain safe ingestion. Activated carbon filters are usually applied to eliminate these contaminants. Some of these contaminants are impregnated onto the carbon filter and, consequently, become an environmental or health concern, especially heavy metals that, after being absorbed by organisms, are not metabolized, and generate free radicals that hinder organ functions [1]. Some researchers have evaluated the contamination of soil, aquatic environments and even animals exposed to heavy metals, highlighting the importance of mitigating this type of pollution [2,3,4].

In addition, carbon filters must be replaced or regenerated after being saturated. Different techniques have been applied, such as thermal regeneration [5,6], biological [7,8], ozonation [9], chemical [10], by acid treatment [11], or by Fenton reaction using peroxide [12]. Even though activated carbon poorly removes some metals from water compared to reverse osmosis systems [13] or even inorganics [14], carbon filters are still being used due to their lower cost. Despite their effectiveness, regeneration techniques are not always viable on an industrial scale due to the high energy cost, complexity of the processes, and partial degradation of the material structure. Currently, carbon from waste filter cartridges has been used to remove water contaminants, such as hexavalent chromium [15,16], synthetic dyes [17,18], phenol [19], or defluorination of water [20,21].

An alternative to producing activated carbon involves the use of residues as raw material. Some studies have explored the application of tea [22] or agricultural residues as metal absorbents [23], the absorption of drugs using carbon derived from date press cake or agave bagasse [24,25], and Radix Angelica Dahurica residue [26] or Paracetamol applying activated carbon from mango seeds [27]. Additionally, agroforestry residue biomass has been evaluated for gas absorption applications [28,29]. Activated carbon produced from Eucalyptus camaldulensis leaves has been studied for the removal of ascorbic acid contaminants in water treatment processes [30]. Other carbons from eucalyptus or corn (Zea mays) cob residue have demonstrated efficacy in the adsorption of methylene blue [31,32]. Furthermore, activated carbon derived from agroindustry has been investigated for cleaning oil-contaminated soils [33] and as a supercapacitor electrode [34]. Despite their versatility, many of these applications involve complex recovery or regeneration processes. Thus, an alternative is to reuse activated carbon from exhausted filters as a filler in polymer composites.

Rubber composites have demonstrated the ability to encapsulate and retain contaminants from residues used as fillers, as previously reported by our research group [35]. Moreover, Gobetti and co-workers [36] have employed steel slag from electric arc furnaces as filler to replace carbon black in rubber composites. They observed that the incorporation of residue particles into the rubber matrix has reduced the leaching of Cr, V, and Mo. Consequently, there is growing interest in research focused on the incorporation of residues into rubber composites, such as biochar to substitute carbon black [37,38], plant-based [39,40,41], synthetic fiber [42], sugarcane bagasse [43], leather waste [44], cacao shell powder [45], polyurethane [46], coffee and tea waste grounds [47,48] or even waste by-products from filter-tea production [49], waste brick powder (WBP) [50] and by-product waste of electric power generation [51].

This paper proposes the use of activated carbon from exhausted filter cartridges as a reinforcing filler in vulcanized natural rubber composites. Composites were carried out by an open two-roll mill using sulfur as vulcanization agent and pressed at 210 Kgf/cm² and 150 °C, testing 5, 10, 15 and 20 phr of carbon filter waste as fillers compared to composites with a similar amount of commercial carbon black. The aim is to partially replace this commercial filler, and to offer an alternative for reusing this waste without a regeneration process.

2. Materials and Methods

Vulcanization reagents, stearic acid (95%, Scientific Exotic, São Paulo, São Paulo, Brazil), Zinc oxide (99.8%, Suzano, São Paulo, Brazil), polyethylene glycol 4000 (PEG-4000), vulcanization accelerators tetramethyl thiuram disulfide/TMTD (99%, Basile Química), benzothiazole disulfide/MBTS (99%, Basile Química, São Paulo, Brazil) and sulfur (99.5%, Scientific Exotic) were used as provided by the suppliers.

The waste derived from the activated carbon filters was obtained from the chemistry teaching laboratory at the School of Engineering and Sciences (FEC), São Paulo State University (UNESP), in Rosana, SP, Brazil, using equipment designed to convert water from distillation and deionization. Figure 1 shows the sequence applied to the sample’s preparation, in which activated carbon filters were ground using an industrial knife mill (Figure 1a). After that, the residue was dried in an oven at 300 °C for 2 h to remove all moisture before mixing it with natural rubber. Then, the carbon residue was sieved (Figure 1b) for 24 h and collected when passed through a 325-mesh screen.

Part of the coal was treated with hydrochloric acid solution (1 mol/L), using 10 g for 100 mL of solution, and stirred for one hour at room temperature. After this time, the coal was separated by filtration and washed several times with distilled water, until reaching a pH close to 7. After filtration, the coal was dried in an oven at 80 °C for 24 h, and used to prepare NR/AC_10phr (HCl) composites.

Formulation and vulcanization compounds are presented in

Table 1. Formulations and vulcanization amount of the produced rubber compounds.

Compounds	Amount (phr)
Natural rubber	100
Zinc oxide	5
Stearic acid	2
PEG 4000 1	2
Activated carbon filter waste	5|10|15|20
Sulfur	1.5
MBTS 2	0.7
TMTD 3	0.4

¹ polyethylene glycol 4000 ² Benzothiazyl Disulfide ³ Tetramethylthiuram disulfide.

.

The residues were incorporated into natural rubber using an open two-roll mill (ASTM D3182-21a) with a friction ratio of 1:1.25. The natural rubber (NR) was masticated for 10 min before adding zinc oxide, stearic acid, PEG-4000 (as a plasticizing agent), and the activated carbon (AC) filler at 5, 10, 15, and 20 phr (NR/AC). Samples were subsequently left to rest for 24 h before the addition of the accelerators MBTS and TMTD, and sulfur. After mixing, samples were pressed at 210 Kgf/cm² and 150 °C using a Mastermac press, Vulcan brand, 400/20-1 model. The methodology and characterization were based on previous studies performed by the research group [52].

Rheometric curves were obtained for every composite (ASTM D2084-19a) using a Team Equipamentos (Produced in Brazil) company instrumented with rotational disk rheometer, 1° arc oscillation, with isothermal conditions of 150 °C. Shore A hardness was determined by Digimess company analog durometer (manufactured in Shenzhen, China) with a scale that complies with ASTM D2240-15 (Shore A scale). Tensile strength was examined using a Biopdi company universal testing machine (ASTM D412-16) (produced in Brazil), with speed of 500 mm·min⁻¹, a load cell of 5 kN, and an internal transducer for deformation. Abrasion resistance was determined according to ASTM D5963-22 with MaqTest company equipment (manufactured in Franca, Brazil) and a 40 m test path under 5 N of pressure. X-ray fluorescence (XRF) analyses (Shimadzu EDX-7000) (manufactured in Kyoto, Japan) were employed with a primary Rhodium (Rh) source, with a collimator producing a 5 mm-diameter area. Cross-linked density by swelling in Toluene (Flory-Rehner) was performed as previously described by our group [53].

The degree of dispersion of residues in natural rubber compounds can be quantitatively determined by Equation (1) [54] from rheometric parameters:

$$L={\eta }_r-m_r={\displaystyle \frac{{\mathrm{M}}_{\mathrm{L}\mathrm{F}}}{{\mathrm{M}}_{\mathrm{L}\mathrm{G}}}}-{\displaystyle \frac{{\mathrm{M}}_{\mathrm{H}\mathrm{F}}}{{\mathrm{M}}_{\mathrm{H}\mathrm{G}}}}$$

(1)

where

L = degree of particle dispersion in the polymer matrix;

${\eta }_r={\displaystyle \frac{{\mathrm{M}}_{\mathrm{L}\mathrm{F}}}{{\mathrm{M}}_{\mathrm{L}\mathrm{G}}}}$

$m_r={\displaystyle \frac{{\mathrm{M}}_{\mathrm{H}\mathrm{F}}}{{\mathrm{M}}_{\mathrm{H}\mathrm{G}}}}$

M_L represents minimum torque;

M_H represents maximum torque;

F and G are relative to the filler and pure gum, respectively.

3. Results and Discussion

Activated Carbon Residue Characterization

X-ray fluorescence (XRF) analyses (

Table 2. X-ray fluorescence analysis of the activated carbon residue.

Analyte	Result (%)
Ca	0.376
Si	0.227
S	0.055
K	0.031
Fe	0.020
Ag	0.017
Zn	0.012
Cu	0.012
Br	0.004
Ti	0.004
C	99.242

) revealed that the solid content of the activated carbon consisted predominantly of carbon (99.242%). Even after prolonged use (filter exhaustion), the metal content remained at trace levels. The presence of silicon (Si), most likely in the form of silica, may contribute to the reinforcing effect of the residue when incorporated into natural rubber composites.

Recently, our research group [35] has evaluated leather shaves as filler into rubber composites. Leather shaves contain chromium III in their composition, which, if improperly disposed, can react and change to chromium VI, a substance of significant health concern. However, lixiviation tests showed that, after the incorporation into the composites, no significant concentrations of contaminants were released due to the rubber’s wettability. These results support the safe use of carbon filters, indicating minimal risk of residual contamination due to the rubber barrier properties.

The rheometric data obtained for the composites containing residual activated carbon (

Table 3. Rheometric parameters.

Composite Samples	MH (dN.m)	ML (dN.m)	ΔM (dN.m)	t90 (min)
NRpure	18.50	0.80	17.7	03.55
NR/AC5phr	20.20	1.10	19.1	03.55
NR/AC10phr	22.00	1.40	20.6	03.42
NR/AC15phr	22.10	1.45	20.7	03.38
NR/AC20phr	24.00	1.50	22.5	03.35

) exhibited the typical behavior of filler-reinforced systems.

As the filler content increased from 0 to 20 phr, the maximum torque (M_H), minimum torque (M_L), and differential torque (ΔM) increased, indicating greater resistance to shear and a possible higher degree of cross-linking in the vulcanized rubber. M_L value has also risen from 0.80 dN.m to 1.50 dN.m, indicating an increase in the composite viscosity and rigidity due to the presence of the filler. This behavior is typically associated with good particle dispersion and well filler–matrix interaction [35].

M_H value enhanced from 18.50 dN.m (NR_pure) to 24.00 dN.m for the composite with 20 phr residual activated carbon (NR/AC_20phr). The higher torque can be associated with an increased number of cross-links formed during vulcanization, which improved the mechanical strength of the composite. Similar behavior was reported by Santos et al. [55], who observed mechanical properties upon incorporating sugarcane bagasse ash into natural rubber composites.

The ΔM, an indicator of the filler’s reinforcing effect, increased with filler content, rising from 17.7 dN.m (NR_pure) to 22.5 dN.m (NR/AC_20phr). This increase suggests that activated carbon enhances the composite’s 3D structure, resulting in a higher cross-link density.

Regarding the optimum cure time (t₉₀), a decrease from 03:33 to 03:21 min was observed as the filler content increased, indicating a possible acceleration of the vulcanization process due to the filler. Metal impurities and oxides present on the surface of the activated carbon may affect the rate of sulfur cross-link formation and thereby influence t₉₀.

Based on the rheometric property data, the minimum and maximum torques shown in Table 3 and the L values for the dispersion of activated carbon filter residue in the matrix are presented in Figure 2.

A lower L value, at a given filler content compared to the unfilled composite, indicates an improved degree of residue dispersion. The degree of filler dispersion was optimal for composites containing up to 5 phr of residue and is considered adequate up to 20 phr, since L values remain below 1, suggesting good dispersion and low occurrence of filler–filler agglomeration. Lower viscosity tends to reduce the minimum torque, facilitating filler dispersion and enhancing filler–rubber interaction [55].

Stress vs. strain curve showed the effect of activated carbon on the mechanical behavior of natural rubber (

Table 4. Tensile behavior of the composites.

Composite Samples	Tensile Strength (MPa) at Rupture	Deformation (%) at Rupture
NRpure	11.4 ± 1.0	382 ± 14
NR/AC5phr	13.1 ± 0.6	385 ± 10
NR/AC10phr	13.9 ± 1.0	331 ± 27
NR/AC15phr	11.0 ± 1.0	306 ± 19
NR/AC20phr	10.7 ± 0.5	234 ± 63
NR/AC10phr (HCl)	14.1 ± 0.4	394 ± 37
NR/CB5phr	14.4 ± 2.7	331 ± 14
NR/CB10phr	14.9 ± 1.3	267 ± 25
NR/CB15phr	16.2 ± 1.5	303 ± 22
NR/CB20phr	18.4 ±1.6	269 ± 34

and Figure 3). NR_pure exhibited a relatively high tensile strength of 11.4 MPa and a high elongation at break of approximately 382%, consistent with the characteristics of elastomeric (flexible) and resilient materials.

Tensile strength slightly increased for the NR/AC_5phr sample, reaching 13.1 MPa with the addition of 5 phr of activated carbon, while elongation at break was similar to NR_pure at 385%. Our research group previously reported results using sugarcane bagasse ash as a reinforcing filler and demonstrated that moderate fillers enhanced tensile properties [35]. When the carbon filler was increased to 10 phr, the composites showed the best mechanical performance in tensile strength, reaching 13.9 MPa, an improvement of 21.9% compared to the rubber without waste. Elongation at break was reduced (around 331%) when compared to NR_pure. It was due to the well interaction between filler and polymeric matrix, which corroborates to the effective action of activated carbon as a reinforcing filler, that enhances tensile at break.

This favorable performance can likely be attributed to the fact that at 10 phr, the activated carbon particles smaller than 45 µm or 325 mesh were well-dispersed within the rubber matrix. However, when the filler concentration was increased to 15 phr, a decrease in tensile strength to approximately 11 MPa was observed, along with a significant reduction in elongation at break. At 20 phr, tensile strength and elongation were further reduced to 10.7 MPa and 234%, respectively. At higher filler contents, the activated carbon particles tend to agglomerate, leading to a loss of dispersion homogeneity within the polymer matrix and the formation of localized stress concentration points. This behavior also supports the fact that the filler interacts only physically with the natural rubber matrix [35].

When the residues are well-dispersed within the polymer matrix, particularly in samples at 5 and 10 phr, they can reinforce the material properties through physical interactions. In general, carbon filter waste shows a hydrophobic character which supports the interfacial adhesion between the filler and the matrix (also hydrophobic). As a result, the tensile strength of the composites increased with residue incorporation up to 10 phr. However, for higher residue fillers, the rigidity of carbon particles restricted chain mobility. Furthermore, the stress transfer from the rubber matrix to the filler under stress conditions led to a reduction in tensile strength at rupture.

Composites of natural rubber with carbon filters were compared to rubber composites with commercial carbon black (CB) (N330) at similar proportions. The commercial filler showed high reinforcement characteristics as the amount of filler at rubber matrix was increased. NR/AC_10phr reached 13.9 MPa values close to those of NR/CB_5phr (14.4 MPa). These results indicate that, even at low concentrations, the carbon filter can partially replace commercial fillers.

Considering the presence of metals absorbed by the carbon filter, additional tests were conducted using carbon filter waste after acid treatment (HCl 1M followed by water washing until pH stabilized) to remove surface-bound metals and evaluate their influence on mechanical resistance. The results showed no significant change in tensile strength for composites with 10 phr of treated carbon filter, reaching 14.1 MPa compared to 13.9 MPa for untreated filler. These results corroborate with the lower metal concentration in the composition of carbon filters, as confirmed by XRF analysis.

Table 5. Volumetric abrasion loss and hardness results.

Composite Samples	Filler Amount (phr)	Volumetric Abrasion Loss (mm3/40 m)	Hardness (Shore A)
NRpure	0	145 ± 1	36 ± 1
NR/AC5phr	5	157 ± 8	42 ± 1
NR/AC10phr	10	154 ± 20	44 ± 1
NR/AC15phr	15	164 ± 14	46 ± 1
NR/AC20phr	20	176 ± 16	47 ± 1

presents the hardness values (Shore A) of the composites. Hardness increased with the addition of the filler. The hardness of NR_pure was 36 Shore A, consistent with the characteristics of a relatively soft rubber. Incorporation of NR/AC_5phr increased the hardness to approximately 42 Shore A, while for NR/AC _10phr the hardness increased to 44 Shore A. This indicates that even small amounts of activated carbon could make the rubber stiffer, as the solid particles restrict surface deformation under compression. The hardness values for NR/AC_15phr and NR/AC_20phr were 46 and 47 Shore A, respectively.

The abrasion resistance of the composites (Table 5) was also evaluated based on the mass loss of the specimens following wear testing against an abrasive surface. NR_pure exhibited a mass loss of 145 mm³/40 m, being a reference for comparative analysis. All composites showed reduced abrasion resistance in relation to NR_pure. For NR/AC_5phr, the abrasion resistance was reduced by 8.3% compared to pure natural rubber (the control value of 145 mm³/40 m from the quantity of 157 mm³/40 m), while NR/AC _10phr showed a slightly lower reduction in performance, 6.2%. For NR/AC_15phr and NR/AC_20phr, the abrasion loss values were considerably higher, recorded at 13.1% and 21.4%, respectively, when compared to NR_pure. These results suggest that, although filler incorporation may contribute to increased hardness, higher AC fillers could impair structural integrity due to poor filler dispersion or the formation of agglomerates, which enhance susceptibility to surface wear. Nevertheless, even though the NR/AC _20phr samples reached 176 mm³/40 m, it is below 200 mm³/40 m required for commercial rubber products.

Cross-link density measures are presented in

Table 6. Cross-link density by swelling in Toluene (Flory-Rehner).

Composites	Cross-link Density (mol/cm3)	Standard Deviation
NRpure	1.43 × 10−4	±0.025 × 10−4
NR/AC5 phr	1.47 × 10−4	±0.021 × 10−4
NR/AC10 phr	1.53 × 10−4	±0.036 × 10−4
NR/AC15 phr	1.44 × 10−4	±0.012 × 10−4
NR/AC20 phr	1.92 × 10−4	±0.078 × 10−4

.

The cross-link density (mol/cm³), determined by Flory–Rehner equation based on swelling tests in toluene, is a key parameter for characterizing the degree of intertwining of polymer chains during the vulcanization process. This parameter is directly related to the structural integrity of vulcanized rubber and significantly influences properties such as mechanical strength, elasticity, and thermal stability [56].

Unfilled natural rubber (NR_pure) exhibits a cross-link density of 1.43 × 10⁻⁴ mol/cm³, which is the reference for the conventional curing system. When different proportions of residual activated carbon (AC) are incorporated, an increase in cross-link density is observed. NR/AC_20phr sample achieved a value of 1.92 × 10⁻⁴ mol/cm³. However, a decrease in tensile strength was recorded. This indicates that although the residue promotes the cross-link formation, higher concentrations may induce filler agglomeration, which negatively impacts the mechanical performance, especially tensile strength.

Additionally, a lower chemical cross-link density allows for greater chain mobility, which facilitates improved molecular orientation under stress, and consequently contributes to increased tensile strength. This behavior was described by Sun et al. [57], who reported enhanced mechanical performance in SBR composites containing carbon black and triazine-based graphdiyne. In their study, improved filler dispersion and effective physical interactions compensated for stable or even reduced cross-link densities, leading to superior mechanical properties.

4. Conclusions

Water decontamination using activated carbon filters is essential for public health; however, large quantities of filter residues have been generated, requiring regeneration techniques to recycle the cartridges. Natural rubber composites containing activated carbon filter residue demonstrated improvements of 21.9% (reaching 13.9 MPa) in tensile strength for 10 phr, when compared to natural rubber and reached values close to those of NR/CB_5phr (with carbon black N330), attaining 14.4 MPa. These results indicate that, even at low concentrations, the carbon filter can partially replace commercial fillers. Abrasion resistance was reduced by only 6.2% (NR/AC _10phr) compared to the pure rubber (from 145 mm³/40 m to 154 mm³/40 m), while hardness increased to 44 Shore A. These results suggest that incorporating activated carbon from filter cartridges into rubber composites seems to be an alternative to avoid the intensive processes of carbon regeneration and reducing environmental impacts from the use of commercial carbon black. Future research should focus on enhancing the interaction between filler and elastomeric matrix to enable the incorporation of higher amounts of residues.