Additively manufactured electrochemical platforms from reclaimed ground tire rubber for environmental monitoring

Abstract

Plastic waste poses a serious threat to ecosystems and human health. Similarly, the increasing use of automobiles has led to a rise in discarded tires, exacerbating environmental concerns. However, repurposing this waste into new materials promotes sustainability and supports the concept of a circular economy. In this study, we explored the combination of reclaimed ground tire rubber (rGTR) and additive manufacturing technology (3D printing) for the development of novel conductive filaments. For this purpose, a filament composed of 25% rGTR, 25% carbon black (CB), and 50% low-density polyethylene (LDPE) was formulated for constructing additively manufactured electrodes. The produced filament exhibited excellent flexibility, with successful incorporation of rGTR into the matrix, as confirmed by TGA, SEM and AFM data. 3D-printed working electrodes prepared using rGTR, CB, and LDPE subjected to an atmospheric air plasma pen (3D-CB/rGTR/LDPE-PT) were applied for detecting a hazardous fungicide, carbendazim (CBZ). Thus, under optimized differential pulse voltammetry (DPV) conditions, the sensor demonstrated a linear response ranging from 1.0 to 40.0 μmol L⁻¹ CBZ, with limits of detection (LOD) and quantification (LOQ) of 79 nmol L⁻¹ and 262 nmol L⁻¹, respectively. The proposed 3D-CB/rGTR/LDPE-PT sensor was successfully applied for detecting CBZ in drinking water and effluent samples, achieving recovery rates between 95 and 105%, highlighting its excellent analytical performance. This work demonstrates that recycled raw materials from the automotive industry can be repurposed to develop innovative devices for detecting environmental contamination caused by indiscriminate pesticide use, while also simultaneously promoting sustainability through the reuse and recycling of waste tire rubber.

---

Green foundation

1. This work aligns with the principles of waste prevention and the use of renewable feedstocks by converting reclaimed ground tire rubber (rGTR) into conductive filaments for additive manufacturing (AM). It enables sustainable, low-cost electrochemical detection of the fungicide carbendazim, supporting circular economy strategies.

2. Valorization of tire waste, a persistent pollutant and microplastic source, is achieved through AM, which reduces production waste. A reagent-free plasma pen activates electrode surfaces, avoiding hazardous chemicals and aligning with the principles of safer synthesis and reduced derivatives.

3. Research could enhance the recyclability and biodegradability of printed devices to reduce environmental impact. Tire rubber's inherent carbon black and plasticizers provide conductivity and flexibility, though eco-friendly alternatives improve sustainability. Expanding rGTR filaments’ use in sensors, batteries, and supercapacitors supports the circular economy and resource efficiency. Process optimization to reduce energy consumption during filament production and AM would further align with green chemistry principles.

1. Introduction

Plastics are made from synthetic organic polymers and are known for their durability, lightness, versatility, and relatively low production costs. These attributes contribute to their status as one of the most widely used materials.¹ Due to their properties, they have become essential to the textile, automotive, manufacturing, and packaging industries.²

However, plastic waste has garnered significant attention in recent years due to poor waste management, which poses a threat to both ecosystems and human health.^1,3,4 Geyer et al.⁵ projected that if current plastic consumption trends persist, the planet will accumulate 12 billion metric tons of plastic waste by 2050. In response, reducing plastic waste and advancing the circular economy (CE) have become priorities for both the European Union and the global community. The CE aims to promote industrial symbiosis, where waste or by-products from one industry are utilized as inputs for another, thereby reducing their carbon footprint, replacing the traditional “end-of-life” concept and moving towards sustainability.^3,6–8

The growing number of waste tires resulting from the increasing use of automobiles has become an environmental problem, with an estimation of one billion tires reaching the end of their useful life each year.⁹ Disposal of these tires is problematic, as sending them to landfills is illegal. Scrap tires can be recycled in two main ways: they can be burned to recover energy or processed by shredding and grinding to produce ground tire rubber (GTR) with desired particle size, which is a valuable material for further recycling and upcycling technologies. GTR can be incorporated into other polymers (both rubber and plastic) to enhance or modify the properties of thermoplastic materials.^10–12 Tires are made from vulcanized rubber and various reinforcing materials. The primary rubber matrix used is a blend of natural rubber (NR) and synthetic rubber (styrene–butadiene rubber (SBR) or butadiene rubber (BR)).¹³ In addition to the rubber compound, tires incorporate carbon black as a reinforcing filler and petroleum oils to enhance flexibility at low temperatures.¹²

The use of additive manufacturing (AM) technologies is a key tool for advancing the circular economy because AM produces minimal or even zero waste when compared to subtractive manufacturing methods.^14–17 AM is a process that involves joining materials to create objects from 3D model data, typically via a layer-by-layer approach, for instance, fused deposition modeling (FDM).^18,19 More commonly known as 3D printing, this approach enables innovation and customization of devices, thereby reducing waste production and facilitating the reuse of materials from other industries as inputs in filament production.^14,16,17,20 However, few biodegradable and recyclable materials have been used in 3D printing.^21,22 Acrylonitrile butadiene styrene (ABS) and polylactic acid (PLA) are the most used thermoplastic materials in non-conductive filaments for 3D-printing.^23,24

Some research groups have employed 3D printing to enhance circularity in supply chains to produce (bio)sensors. Bank's group first introduced the integration of circular economy principles and additive manufacturing into the electroanalytical field to develop advanced sensing platforms.^25,26 For instance, Kalinke et al.²⁷ developed a new additive manufacturing filament by combining recycled polylactic acid (PLA) and carbon black (CB) with carboxylated multi-walled carbon nanotubes (COOH-MWCNTs, 10 wt%) incorporated into the polymer matrix. This filament maintained excellent flexibility and printability while enhancing electrochemical performance, enabling the creation of an electrochemical genosensor for detecting yellow fever virus cDNA. Arantes et al.²⁸ utilized a mixture of graphite and CB combined with recycled PLA and castor oil to create a filament that enabled the electroanalytical detection of oxalate. Sigley et al.²⁶ described the use of a conductive filament, composed of CB and recycled PLA derived from coffee machine capsules in the development of electroanalytical sensors for detecting caffeine. Crapnell et al.²⁹ reported the use of recycled filaments (both conductive and non-conductive) to develop a new electroanalytical detection platform. The sensor, additively manufactured from conductive filaments produced by recycling old electroanalytical sensors and incorporating CB particles into the matrix, demonstrated effective performance for acetaminophen detection. However, it is important to note that repeated reprocessing cycles for recycling (or reuse) old conductive filaments lead to the degradation of the polymer matrix.³⁰ Most of the cited studies focus on incorporating recycled PLA and natural oils into filament matrices; however, to the best of our knowledge, no research group has demonstrated the production of filaments using recycled ground tire rubber, a material that could enhance flexibility and conductivity due to its CB content.

In general, post-printing electrochemical sensors require surface treatment or activation procedures to remove excess polymer and uncover the embedded conductive particles, thereby enhancing their electrochemical performance. Different groups showed various strategies for the surface treatment of 3D-printed electrodes. While these methods often enhance electrochemical performance, many involve complex procedures, utilize toxic solvents, and require costly, bulky equipment, which limit their practicality and increase processing time.^31,32 Siqueira et al.³³ introduced an environmentally friendly, portable, and cost-effective strategy using atmospheric plasma jet pens, capable of removing excess polymer material and enhancing the carbon conductivity of the 3D-printed electrode surfaces. This device excites the gases present in the air by applying a high voltage between the tip (cathode) and the electrode (anode). The air dielectric barrier acts as an insulator, leading to an electrical discharge that typically manifests as sparks or arcs, generating luminous plasma associated with the pen. This method, with a rapid activation time of just 2 minutes, significantly enhances the electroanalytical performance of 3D-printed electrodes.³³ This approach is a promising and versatile alternative, as solvent treatments require specific optimization and most techniques have only been validated for PLA-based compounds.

Pesticides, or plant protection products, are used to keep crops healthy by preventing diseases and infestations.^34,35 They include herbicides, insecticides, fungicides, acaricides, growth regulators, and repellents.³⁴ However, excessive and uncontrolled use of pesticides has resulted in food contamination and pollution of the environment, agriculture, and aquatic systems, where they may pose health risks, such as developmental damage, cancer, or other diseases, as well as harm to ecosystems.^36–38 Carbendazim (CBZ), a primary benzimidazole fungicide, poses health risks due to the stability of the benzimidazole ring. Recently, the European Union (EU) banned the use of CBZ due to its harmful health effects, including carcinogenicity, reproductive toxicity, and endocrine disruption.³⁹ Recently, a study revealed the presence of CBZ in 88% of rainwater samples collected in three Brazilian cities, despite a ban on the use of this substance in Brazil since 2022.⁴⁰ According to the European Food Safety Authority (EFSA), the maximum residue limit (MRL) for carbendazim is 0.01 mg kg⁻¹ in aqueous, acidic, oily and dry matrices.^41,42 Therefore, there is an urgent need to monitor CBZ levels in environmental and food samples.^43,44

Based on these considerations, a novel strategy for the electrochemical detection of the pesticide CBZ has been developed. Integrating circular economy principles, we utilized reclaimed GTR, combined with LDPE and CB particles, to produce conductive filaments. The resulting system offers an effective, sustainable solution for detecting CBZ residues in environmental water samples.

2. Experimental section

2.1 Chemicals and solutions

Ground tire rubber (GTR) with a particle size up to 0.6 mm, prepared from a mix of passenger and truck tires, was supplied by Grupa Recykl S.A. (Śrem, Poland).⁴⁵ Low-density polyethylene (LDPE) (Malen E FABS 23-D022) was the matrix of the material. LDPE was obtained from Basell Orlen Polyolefins Sp. z o.o. According to the data sheet from the producer, the used LDPE presents a density of 0.923 g cm⁻³ and a melt flow index (190 °C, 2.16 kg⁻¹) of 1.95 g 10 min⁻¹. Carbon black (CB) (ENSACO 250G) was utilized as an electrically conductive filler. CB was obtained from Imerys Graphite & Carbon Switzerland Ltd (Bironico, Switzerland). According to the information provided by the supplier, the estimated BET nitrogen adsorption surface area was 66 m² g⁻¹, an oil absorption number of 194 mL per 100 g, water conductivity (1 h, 25 °C) of 1.0 μs cm⁻¹ and a total sulfur content of 120 ppm.

Acetic acid (98% w/v) was provided by Vetec (Rio de Janeiro, Brazil), and phosphoric acid (85% w/v), potassium ferricyanide (99% w/w), paracetamol (99% w/w) and dopamine (99% w/w) were sourced from Labsynth (São Paulo, Brazil). Boric acid (99.8% w/w) and sodium hydroxide (98% w/w) were obtained from AppliChem Panreac (Barcelona, Spain). Ethanol (95% v/v) and potassium chloride (98% w/w) were supplied by EasyPath|Diagnostics (São Paulo, Brazil) and Dinâmica® (São Paulo, Brazil), respectively. Ferrocenemethanol (FcMeOH, 97% w/w), carbendazim (98% w/w), caffeine (99% w/w), paracetamol (≤100% w/w), NaNO₂ (≥97.0% w/w), ZnNO₃ (98% w/w), CuNO₃ (99–104% w/w) and nickel(II) chloride (98% w/w) were purchased from Sigma-Aldrich® (St Louis, USA). Ascorbic acid (≥99% w/w), and uric acid (99% w/w) were purchased from Fisher Chemical (Thermo Scientific Chemicals, New Jersey, USA). Standard solutions (10 μg mL⁻¹) of mercury in 5% HNO₃ were purchased from PerkinElmer (Shelton, USA).

Stock solutions of 4.0 mmol L⁻¹ of CBZ were prepared in ethanol. Daily standard solutions of 1.0 mmol L⁻¹ of CBZ in the supporting electrolyte were prepared and diluted in the electrochemical cell before analysis. The Britton–Robinson (BR) buffer was made by mixing 0.04 mol L⁻¹ acetic, boric, and phosphoric acids, with the pH adjusted from 2.0 to 10.0 using 1.0 mol L⁻¹ NaOH. Acetate buffer (pH 4.0) was prepared by partially neutralizing acetic acid with NaOH. Additionally, 0.1 mol L⁻¹ KCl was added to maintain the ionic strength of the solutions.⁴⁶

In this study, all chemicals used were of analytical grade and were utilized as received, without any additional purification. All solutions were prepared with deionized water with a resistivity not less than 18 MΩ cm, obtained from a Milli-Q purification system (Millipore, Bedford, MA, USA).

2.2 Recycled conductive filament production

Ground tire rubber particles have a cross-linked structure. To improve the phase interactions between GTR and the other components of the compound, it is necessary to undergo the process of breaking the S–S and C–S crosslinking bonds, which are formed during the vulcanization process.⁴⁷ GTR was thermo-mechanically treated in a lab-size planetary extruder PLATEX 80 produced by Takimsan Disli Kesici Ltd Sti. (Istanbul, Türkiye). The temperature applied during devulcanization of GTR was set at 200 °C. The screw speed was 15 rpm and throughput was 0.5 kg h⁻¹. The devulcanization level was evaluated using Mooney viscosity determined according to ISO 289-1. The prepared reclaimed GTR was characterized by a Mooney viscosity of ML (1 + 4) 100 °C – 38 MU. The appearance of GTR before and after devulcanization (rGTR – reclaimed ground tire rubber) is shown in Fig. 1.

Fig. 1 (A) Schematic representation of the conductive filament production process and the 3D printing of electrodes. (Inset) Photograph of the 3D-printed electrode (3D-CB/rGTR/LDPE). (B) Image demonstrating the filament's flexibility. (C) Thermogravimetric (TG) and differential thermogravimetric (DTG) curves of the CB/rGTR/LDPE conductive filament.

The filament was prepared by melt-compounding using a laboratory twin-screw co-rotating extruder with conical screws, produced by Łukasiewicz Research Network–Institute of Engineering of Polymer Materials and Dyes (Toruń, Poland), equipped with a 1.75 mm diameter nozzle. The temperatures set on the extruder were as follows: 140 °C, 170 °C, and nozzle 200 °C. With its modified design, the extruder is capable of operating in both closed and continuous modes. This made the mixing process more efficient and could be carried out in one extrusion cycle, limiting the thermal decomposition of the polymer.³⁰ The mixing time lasted up to 8 min, and the screw speed during mixing was 100 rpm. After the material (25 wt% CB, 25 wt% rGTR and 50 wt% LDPE) was mixed, the gears were opened and the material flowed out of the nozzle. The material was then given the desired size and shape using a conveyor belt and then cooled with air using fans. The final step was winding the filament onto a spool.

2.3 Construction of additively manufactured working electrodes

The electrodes manufactured using the CB/rGTR/LDPE filament were named 3D-CB/rGTR/LDPE. The 3D-CB/rGTR/LDPE electrodes were 3D printed using an FDM printer (Flashforge Dreamer NX, China). The following printing parameters were utilized: horizontal orientation, a layer height of 0.05 mm, an infill density of 100%, a nozzle diameter of 0.6 mm, two perimeter shells, a bed temperature of 60 °C, an extruder temperature of 240 °C, and a printing speed of 30 mm s⁻¹. The electrodes were designed in a lollipop shape, as used in previous works in the literature.³³ Their geometry consists of a spherical section with a diameter of 6.4 mm, complemented by a rectangular portion measuring 3.0 mm in width and 9.0 mm in length for electrical contact.

2.4 Treatment of 3D-CB/rGTR/LDPE electrodes using an atmospheric air plasma pen

The 3D-CB/rGTR/LDPE electrodes were treated using an atmospheric air plasma pen (PLASMAX – EHF 2204, KLD Biosistemas, São Paulo), following the protocol outlined by Siqueira et al.³³ The pen was operated in continuous mode, maintaining a distance of less than 1 mm between the tip and the electrode surface. The treatment was carried out in horizontal lines, with an application time of 2 min, plasma power set at 3000 mW, and a 0.2 × 15 mm needle holder. The electrodes treated with the plasma pen are referred to as 3D-CB/rGTR/LDPE-PT. The schematic representation of the procedure and a real image of the electrode after plasma treatment are presented in Fig. S1.

2.5 Physiochemical characterisation

Electrical conductivity was studied using the broadband dielectric spectroscopy (BDS) technique. The DC conductivity values for all samples were measured at a frequency of 100 Hz, which is a good representation of σ_DC, and at a temperature of 25 °C, as close to room temperature as possible. The measurements were conducted using a Novocontrol Concept 40 broadband dielectric spectrometer Alpha-A (Montabaur, Germany), equipped with the ZG4 dielectric interface. For the electrical measurements, gold electrodes were evaporated in a vacuum at a plane parallel to the surfaces of the samples. The samples had a thickness of 2 mm and a diameter of 10 mm.

Thermogravimetric analysis (TGA) was performed using a Netzsch TG 209 (Selb, Germany) apparatus, with a heating rate of 20 °C min⁻¹, ranging from 25 to 800 °C, under a nitrogen atmosphere. Surface morphology was analyzed using a Vega 3 scanning electron microscope (SEM, Tescan, Czech Republic) at 20.0 kV, controlled using Vega TC software. Raman spectra were obtained with a LabRAM HR Evolution spectrophotometer (HORIBA, Japan), using a 532 nm laser at 50 mW over a range of 4000 to 70 cm⁻¹, managed using LabSpec software.

Fourier transform infrared (FT-IR) absorption spectra were collected with an ATR MIR/FIR mode spectrometer (PerkinElmer, USA) covering 500 to 4000 cm⁻¹. Atomic force microscopy (AFM) images were captured with an SPM-9600 scanning probe microscope (Shimadzu, Japan) in dynamic force mode, using silicon probes (PPP-NCHR AFM, NANOSENSORS™, Switzerland) with a resonance frequency of 330 kHz, a force constant of 42 N m⁻¹, a length of 125 μm, an average width of 30 μm, and a thickness of 4 μm.

X-ray photoelectron spectroscopy (XPS) was conducted using an AXIS Supra (Kratos, UK) with a monochromated Al X-ray source (1486.6 eV) at 225 W and a hemispherical sector analyzer. The system operated in fixed transmission mode with a pass energy of 160 eV for survey scans and 20 eV for region scans. The collimator was set to slot mode, analyzing an area of approximately 700 × 300 μm. The FWHM of the Ag 3d5/2 peak at 20 eV pass energy was 0.613 eV. The binding energy scale was calibrated by setting the sp² graphitic C 1s peak at 284.5 eV. Although this approach is admittedly imperfect, it was adopted due to the absence of better alternatives and the limited usefulness of absolute peak positions.²⁸

Contact angle images were captured 10 s after depositing a drop of deionized water onto the electrode surface. The contact angle was measured by tracing a tangent between the liquid phase (water droplet) and the electrode surface using Ossila's Contact Angle Goniometer (Sheffield, UK).

2.6 Apparatus and electrochemical measurements

A 10 mL electrochemical cell was printed using a GTMax 3D printer (São Paulo, Brazil) with polyethylene terephthalate glycol (PETG) filament. The cell housed a working electrode (3D-CB/rGTR/LDPE-PT), a reference electrode (Ag|AgCl|KCl_(sat.)), and a counter electrode (platinum wire). A rubber ring defined the working electrode's geometric area (0.22 cm²). Further details on the design and arrangement of the electrochemical cell can be accessed in previous works by our research group.^48,49

Cyclic voltammetry (CV), differential pulse voltammetry (DPV), and electrochemical impedance spectroscopy (EIS) were conducted using an Autolab PGSTAT204 potentiostat/galvanostat (Metrohm Autolab BV, Utrecht, Netherlands) equipped with the EIS FRA32M module. The system was connected to a microcomputer and controlled using NOVA 2.1.7 software. EIS measurements were conducted over a frequency range of 0.1 Hz to 50 kHz, with an amplitude of 10 mV, in the presence of 2.0 mmol L⁻¹ [Fe(CN)₆]^4−/3− in 0.1 mol L⁻¹ KCl solution, applying a half-wave potential of +0.25 V (vs. Ag|AgCl|KCl_(sat.)). The Randles equivalent circuit was used to determine the charge transfer resistance (R_ct) between the electrode surface and the redox probe.

The 3D-CB/rGTR/LDPE electrodes were treated chemically and electrochemically (3D-CB/rGTR/LDPE-QET) as per Richter et al.⁵⁰ They were polished with 600 and 1200-grit sandpaper using ultrapure water, then electrochemically activated (+1.4 V for 200 s and −1.0 V for 200 s in 0.5 mol L⁻¹ NaOH).

The double layer capacitance (C_dl) measurements were performed using CV scans in different scan rates (2.5–50 mV s⁻¹), following established procedures.^51,52C_dl values were estimated from the slope of the peak current density (ΔJ), calculated as the difference between the anodic (I_a) and cathodic (I_c) currents divided by the geometric area of the electrodes (0.22 cm²), versus the scan rate. Current peaks were recorded at a potential of +0.05 V (vs. Ag|AgCl|KCl_(sat.)) within a potential window of −0.1 to +0.2 V (vs. Ag|AgCl|KCl_(sat.)) in 0.5 mol L⁻¹ KCl solution.

2.7 Real sample analysis

Water samples from the river, lake, tap, and bottled drinking water were collected in the city of Uberlândia, Minas Gerais, Brazil. The treatment of these samples included the addition of reagents required to prepare 0.04 mol L⁻¹ acetate buffer (pH = 4.0). Subsequently, a specific aliquot of the target analyte was added to the solution to spike the samples with a known amount of CBZ. Next, CBZ was quantified in all samples using the standard addition method, which ensures accurate measurement of analyte concentration. This thorough approach allowed for the reliable assessment of water quality across different sources.

3. Results and discussion

3.1 Production, thermal, morphological and spectroscopic characterization

The production process of the CB/rGTR/LDPE conductive filament, eliminating the need for solvents, and the printing process of 3D-CB/rGTR/LDPE electrodes are illustrated in the schematic representation in Fig. 1A. The resulting filament exhibited excellent flexibility at room temperature, with a resistance (R_s) of 2.84 ± 0.07 kΩ over a 10 cm length, as shown in Fig. 1B. This value is comparable to that of commercial filaments such as Protopasta®, which typically exhibit R_s values around 2.0–3.5 kΩ.⁵³ Furthermore, it demonstrated high print quality, as evidenced by the lollipop-shaped electrodes in Fig. 1B.

The results of electrical conductivity tests clearly indicate that the presence of rGTR in the composite structure significantly enhances the percolation threshold of the materials studied, as shown in Fig. S2. The CB/rGTR/LDPE sample achieved an electrical conductivity of 0.2128 S cm⁻¹. To evaluate the impact of rGTR on the improvement of the filament's electrical conductivity, filament CB/LDPE, composed of 75% LDPE and 25% carbon black, as a reference sample was prepared. The filament without the addition of rGTR shows a conductivity of only 0.0026 S cm⁻¹. For comparison, the commercial Protopasta® filament, commonly used as a conductive material in 3D printing, exhibits a conductivity of 0.1100 S cm⁻¹.³⁰ The observed increase in electrical conductivity following the incorporation of rGTR can be attributed to a synergistic mechanism related to two facts. First, rGTR obtained from end-of-life tires contains residual carbon black, which contributes to lowering the percolation threshold and facilitates the formation of a stable, three-dimensional conductive network. Second, the presence of rGTR may promote a more favorable dispersion of carbon black particles within the polymer matrix, increasing the number of effective interparticle connections.

The thermogravimetric analysis (TGA, solid black line) and differential thermogravimetric (DTG, dashed purple line) curves, derived from the TGA data of the CB/rGTR/LDPE filament, are presented in Fig. 1C. The TGA/DTG curves revealed two distinct thermal events, with peak temperatures at 380 °C and 484 °C, which are associated with its thermal behavior. The first peak corresponds to the decomposition of natural rubber (decomposition of hydrocarbons), and the second peak is related to synthetic rubbers (styrene–butadiene rubber and butadiene rubber, commonly used in the tire industry) and the degradation of LDPE chains, respectively, which corroborated with the literature.^54–56 Additionally, it was observed that the thermal stability, defined as the temperature at which 2% weight loss (T_2%) occurs, was determined to be 307 °C for the CB/rGTR/LDPE filament. Based on the second filament decomposition process, a filler content of 35 wt% was achieved. This is because, in addition to the 25 wt% CB added to the filament composite, the rubber itself contains CB as a reinforcing agent.⁵⁴ Furthermore, according to the TGA/DTG results, temperatures between 220 °C and 280 °C are suitable for the 3D printing process. The TGA parameters, including the temperatures corresponding to 2% (T_2%), 5% (T_5%), 10% (T_10%), and 50% (T_50%) weight loss, the temperatures at which the maximum rate of weight loss (T_max1, and T_max2) occurs, and the carbon filler residual mass at 798 °C, were determined for the CB/rGTR/LDPE filament and are presented in Table S1.

Scanning electron microscopy (SEM) images of the 3D-CB/rGTR/LDPE and 3D-CB/rGTR/LDPE-PT electrodes are shown in Fig. 2A and C, respectively. Additionally, SEM images of the CB/rGTR/LDPE filament at low and high magnifications after fabrication are presented in Fig. S3. In Fig. S3, the surface exhibits a highly heterogeneous morphology, resulting from the integration of the polymer matrix, recycled rubber, and CB particles. Fig. 2A reveals a rougher surface, indicative of the presence of recycled or devulcanized rubber, which imparts a grainy texture to the 3D-CB/rGTR/LDPE electrode. After atmospheric air plasma treatment (Fig. 2C), the surface topography clearly reveals sponge-like structures of CB flakes, along with increased porosity and significant voids on the electrode surface, which may act as active sites for electrochemically active species/analytes.

Fig. 2 SEM images of (A) 3D-CB/rGTR/LDPE and (C) 3D-CB/rGTR/LDPE-PT electrodes. Corresponding AFM images are shown in (B) and (D), respectively.

Atomic force microscopy (AFM) analyses were performed to evaluate the effect of surface treatment on the electrodes (Fig. 2B and D). Topographic images before treatment showed a rough surface, as expected. After plasma treatment, the roughness of the surface increased (RMS = 124 nm) compared to the untreated electrode (rms = 78 nm). These results align with the SEM images, which revealed a flake-like morphology after treatment.

The chemical composition of the 3D-printed electrodes was analyzed using X-ray photoelectron spectroscopy (XPS) before (3D-CB/rGTR/LDPE) and after treatment with the atmospheric air plasma pen (3D-CB/rGTR/LDPE-PT), as shown in Fig. 3. The C 1s spectra for the 3D-CB/rGTR/LDPE and 3D-CB/rGTR/LDPE-PT electrodes are presented in Fig. 3A and B, respectively. In Fig. 3A, the peak at 284.63 eV can be attributed to C–H or C–C bonds, which are present in both the chemical structures of rGTR and LDPE.⁵⁷ Meanwhile, the peak at 288.11 eV corresponds to O–C=O bonds. The presence of oxygen on the surface of the 3D-CB/rGTR/LDPE electrode suggests that it contained some low-level surface oxidation occurring during the 3D printing process or even some contamination.⁵¹ The XPS results revealed that surface modification with atmospheric air plasma significantly increased both the graphitic carbon and oxygen content. As shown in Fig. 3B, the graphitic carbon peak at 284.43 eV became more pronounced after treatment (indicating that treatment effectively removed non-conductive material from the surface, making CB accessible within the XPS detection range), while a new peak emerged at 287.3 eV. This newly formed peak corresponds to C–O groups, which were introduced through surface oxygenation induced by the atmospheric air plasma treatment.

Fig. 3 (A and B) XPS C 1s spectra for the 3D-CB/rGTR/LDPE and 3D-CB/rGTR/LDPE-PT electrodes, respectively. (C) FTIR and (D) Raman spectra of 3D-CB/rGTR/LDPE (black line) and 3D-CB/rGTR/LDPE-PT (green line). (E and F) Contact angle images of the surfaces of 3D-CB/rGTR/LDPE and 3D-CB/rGTR/LDPE-PT electrodes, respectively.

The surface oxygenation appears to occur not only within the polymer matrix but also on the CB. This interpretation is supported by the shift in the binding energy (BE) of the oxidized CB, which reflects an increased presence of surface oxygen species, either adsorbed or chemically bonded, after air plasma treatment. Notably, the shift toward higher BE values indicates changes in the local chemical environment due to oxygen incorporation. For instance, the C–C peak exhibits a BE shift of approximately 0.36 eV, further confirming the modification of the 3D-CB/rGTR/LDPE-PT electrode surface chemistry. A similar behavior was reported by Glowacki et al.,⁵⁸ who observed comparable surface modifications of CB-PLA electrodes after laser ablation in the presence of air with displacement of the BE due to oxidized CB. The relative concentration and binding energy position of each component assigned to C 1s are summarized in Table S2. Indeed, the literature describes that, during the plasma treatment in the presence of atmospheric air, many ionized species can be generated, including O₂˙⁻, HO₂˙, O˙, and OH˙ ⁵⁹ contributing to the cleavage of the polymer structure and the formation of defects or insertion of oxygenated groups.

Fourier transform infrared (FT-IR) analysis was also performed on the 3D-CB/rGTR/LDPE (black line) and 3D-CB/rGTR/LDPE-PT (green line) electrodes (Fig. 3C). The spectrum of the 3D-CB/rGTR/LDPE electrode exhibits the primary vibrational modes corresponding to its polymer composition (rGTR and LDPE). The bands corresponding to C–H stretching vibrations in the 3121–2800 cm⁻¹ range and CH₂ deformation at 1445 cm⁻¹ were observed.⁵⁴ The absorbance band at 1656 cm⁻¹ corresponds to C=C stretching vibrations, while the band at 1536 cm⁻¹ is associated with the presence of zinc stearate, a compound commonly used in the rubber industry.^54,60 The bands at 1221 cm⁻¹ and 1097 cm⁻¹ correspond to the stretching vibrations of C–O and C–O–C bonds.⁵⁴ The peak at 1740 cm⁻¹ indicated the presence of C=O stretching vibrations of esters, which are absent in LDPE.⁵⁶ The absorbance at 1362 cm⁻¹ corresponds to CH₃ and –CH=CH– vibrations in natural rubber, while the peak at 1015 cm⁻¹ indicates =CH₂ vibrations in polybutadiene and C–H in benzene rings, confirming the presence of butadiene and styrene–butadiene rubber in rGTR. This confirms that used tires were utilized in GTR production.^54,61 The spectrum of the 3D-CB/rGTR/LDPE-PT electrode after plasma treatment (Fig. 3C, green line) showed not only a reduction or removal of certain spectral bands (1656 cm⁻¹, 1097 cm⁻¹, and 1015 cm⁻¹) but also the formation of additional oxygenated functional groups, as also observed in the XPS analysis. Notably, a band at 3456 cm⁻¹, corresponding to the stretching vibration of the OH group, indicates the presence of moisture.⁵⁶ These results suggest the partial removal of rGTR and LDPE from the electrode surface, which consequently reduces electrical resistance by exposing more carbon black particles.

Raman spectra of both electrodes were obtained (Fig. 3D). The peaks of the D band (1350 cm⁻¹), G band (1584 cm⁻¹), and 2D band (2704 cm⁻¹) indicate the presence of carbon black. The D band is linked to defects and sp³ bond formation, while the G band is associated with C=C stretching in sp² species.⁶² The increased intensity of the bands after surface treatment can be attributed to the introduction of defects or reactive species by the plasma, which enhances the intensity of related bands, such as the D band. Additionally, rougher or more porous surfaces with larger contact areas provide more interaction points with the laser, further increasing the intensity of the Raman bands. The greater presence of CB particles on the surface also contributes to this increase in band intensities. The I_D/I_G ratio, used to evaluate surface structural defects,⁶² was 1.03 for 3D-CB/rGTR/LDPE and 1.05 for 3D-CB/rGTR/LDPE-PT. These values suggest that the treatment did not introduce additional defects into the CB crystal structure on the electrode surface.

The wettability of the electrode surfaces was evaluated by means of contact angle measurements, to obtain information about the polarity of the surfaces,^58,63 before and after modification, as shown in Fig. 3. The contact angle value of 121.9° observed for the 3D-CB/rGTR/LDPE electrode indicates a hydrophobic surface. The incorporation of rGTR into the polymer matrix contributes to reduced surface wettability.⁵⁷ This hydrophobic effect becomes even more pronounced following atmospheric air plasma treatment: the 3D-CB/rGTR/LDPE-PT electrode exhibited a contact angle of 135.8°, indicating enhanced surface hydrophobicity. This behavior is further supported by XPS analysis, which revealed an increased proportion of graphitic carbon after treatment, along with 3.7% oxygenated species (see Table S2). Surfaces enriched with graphitic carbon structures tend to exhibit greater hydrophobicity, as reflected by higher contact angle values. A similar effect was observed by Glowacki et al.,⁵⁸ who reported an increase in the hydrophobicity of CB-PLA electrodes, from 45.4° to 72.5°, following laser ablation in the presence of atmospheric air. This enhancement in hydrophobicity can be attributed to the exposure of carbon particles on the surface after the removal of excess surface polymer, despite the concurrent oxidation of the material.

Overall, the data provide strong evidence that treating the electrodes with an atmospheric air plasma pen effectively removes excess rGTR and LDPE from their surface, as demonstrated by SEM, AFM, XPS, FTIR, Raman, and contact angle analyses. This treatment reveals a greater portion of the underlying conductive graphitic material (without altering its crystalline structure, as confirmed by the Raman data) which is expected to significantly enhance the electrochemical performance of the electrodes. The removal of rGTR and LDPE reduces the insulating barrier, while increased exposure of the graphitic layer expands the active surface area available for electrochemical reactions. Furthermore, the formation of oxygenated functional groups may introduce additional active sites, further promoting electrochemical activity.

3.2 Electrochemical behavior of functional recycled filaments and electrodes

Initially, the performance of the electrode in its native form (as printed) was investigated in the presence of 5.0 mmol L⁻¹ [Fe(CN)₆]^3−/4−, as shown in Fig. 4A, black line. No electrochemical response was observed, likely due to the significant amount of non-conductive polymer present on the surface, which hinders charge transfer at the electrode interface. To enhance the conductivity of the electrodes, two different surface treatments were tested to remove excess non-conductive polymer and expose more conductive particles: (1) a chemical/electrochemical treatment in 0.5 mol L⁻¹ NaOH solution, proposed by Richter et al.,⁵⁰ and (2) atmospheric air plasma jet pen treatment, proposed by Siqueira et al.³³ The cyclic voltammograms of the electrodes with different surface treatments in the presence of the redox probe [Fe(CN)₆]^3−/4− are presented in Fig. 4A, blue and green lines, respectively.

Fig. 4 Electrochemical behavior of (A) 5.0 mmol L⁻¹ [Fe(CN)₆]^3−/4− in 0.5 mol L⁻¹ KCl solution at 3D-CB/rGTR/LDPE (black line), 3D-CB/rGTR/LDPE-QET (blue line), and 3D-CB/rGTR/LDPE-PT (green line) electrodes. (A′) Magnified view of the blank signal and 3D-CB/rGTR/LDPE and 3D-CB/rGTR/LDPE-QET responses. (B) 5.0 mmol L⁻¹ FcMeOH in 0.5 mol L⁻¹ KCl solution. (C) 1.0 mmol L⁻¹ paracetamol in 0.1 mol L⁻¹ acetate buffer (pH 4.0) and (D) 1.0 mmol L⁻¹ dopamine in 0.1 mol L⁻¹ HClO₄. The analysis was carried out using 3D-CB/rGTR/LDPE-PT (green line) and 3D-CB/rGTR/LDPE (black line) electrodes. Dashed lines correspond to blank signals. CV conditions: scan rate = 50 mV s⁻¹, step potential = 5 mV. (E) Nyquist diagram of impedance spectra at +0.25 V (vs. Ag|AgCl|KCl_(sat.)) of the 3D-CB/rGTR/LDPE (black dots), 3D-CB/rGTR/LDPE-PT (green dots) and 3D-CB/rGTR/LDPE-QET (blue dots) electrodes in the presence of 2.0 mmol L⁻¹ [Fe(CN)₆]^3−/4− in 0.1 mol L⁻¹ KCl solution. The simplified Randles equivalent circuit is shown in the inset.

The electrode, after undergoing electrochemical treatment in NaOH solution (3D-CB/rGTR/LDPE-QET), exhibits a poor electrochemical response, with only slight differences compared to the profile of the as-printed electrode. This behavior is attributed to the fact that the polymers comprising the electrode's polymer matrix are not susceptible to saponification as they do not contain ester groups in their chemical structure.⁵⁰ In contrast, the electrode surface treated by an atmospheric air plasma pen provided a good reversible electrochemical response for the redox probe [Fe(CN)₆]^3−/4−. The peak-to-peak separation (ΔE_p) values and the ratio between the anodic current (I_pa) and the cathodic current (I_pc) were 121 mV and 1.01, respectively, indicating excellent reversibility after surface treatment. This is close to the optimal ΔE_p of 82 mV for the redox probe [Fe(CN)₆]³⁻/⁴⁻, as proposed by Veloso et al.⁶⁴ for 3D-printed sensors. The performance of the 3D-CB/rGTR/LDPE-PT electrode was also evaluated in the presence of other redox probes commonly used in electrochemistry, such as FcMeOH, paracetamol, and dopamine, as shown in Fig. 4B–D, respectively.

These compounds were selected as redox probes due to their well-established behavior on carbonaceous surfaces. FcMeOH is a well-known neutral outer-sphere redox probe widely used to assess changes in electrochemically active.^58,65,66 On the other hand, paracetamol and dopamine are inner-sphere probes whose electron transfer kinetics depend not only on changes in surface area but also on surface chemical composition, especially the presence of oxygenated functional groups.^67,68 Indeed, these results are consistent with the characterization described above. For example, XPS analysis revealed that surface modification with atmospheric air plasma significantly increased both graphitic carbon and oxygen content. The increased graphitic carbon peak suggests that plasma treatment effectively removed non-conductive polymer material, exposing the CB particles. Moreover, C–O groups appeared after treatment (Fig. 3B), likely due to the generation of reactive oxygen species during plasma exposure, which contributed to cleavage of the polymer structure and the formation of surface defects and oxygenated functional groups.⁵⁹

Excellent values of ΔE_p and I_pa/I_pc were obtained, as summarized in Table S3. As shown in Fig. 4B, the electrochemical performance of the 3D-CB/rGTR/LDPE-PT-treated electrode outperforms that of the 3D-CB/rGTR/LDPE electrode in the presence of FcMeOH. It exhibits well-defined I_pa and I_pc, with a ΔE_p of 186 mV, indicating enhanced and stable electrode kinetics. In the presence of paracetamol (Fig. 4C), the 3D-CB/rGTR/LDPE-PT electrode exhibited unrestricted electrode kinetics with a ΔE_p of 212 mV, whereas the 3D-CB/rGTR/LDPE electrode showed little to no electrochemical response. The electrochemical response to dopamine (see Fig. 4D) showed a ΔE_p of 106 mV, lower than the 283 mV reported by Pereira et al.⁶⁹ for a carbon black PLA electrode treated with CO₂ plasma.

Fig. 4E displays the Nyquist plot obtained from electrochemical impedance spectroscopy (EIS) measurements in 2.0 mmol L⁻¹ [Fe(CN)₆]^3−/4− (0.1 mol L⁻¹ KCl) for the 3D-CB/rGTR/LDPE, 3D-CB/rGTR/LDPE-PT, and 3D-CB/rGTR/LDPE-QET electrodes. The 3D-CB/rGTR/LDPE-PT electrode demonstrated superior performance, with a charge transfer resistance (R_ct) of 0.073 ± 0.003 kΩ, as determined using a simplified Randles circuit (presented in Fig. 4E), which was significantly lower than the values observed for the 3D-CB/rGTR/LDPE (6.92 ± 0.03 kΩ) and 3D-CB/rGTR/LDPE-QET (10.37 ± 0.41 kΩ) electrodes, which exhibited similar R_ct values. This approximately 95-fold reduction in R_ct for the 3D-CB/rGTR/LDPE-PT electrode, compared to the 3D-CB/rGTR/LDPE electrode, can be attributed to the increased electroactive area resulting from plasma treatment, which enhanced surface roughness and porosity and promoted greater exposure of CB particles on the surface, as confirmed by the SEM, AFM, XPS, FTIR, and Raman data discussed above.

A cyclic voltammetry scan rate study was conducted over a range of 10 to 200 mV s⁻¹ in the presence of 5.0 mmol L⁻¹ FcMeOH in 0.5 mol L⁻¹ KCl solution for the 3D-CB/rGTR/LDPE and 3D-CB/rGTR/LDPE-PT electrodes, as shown in Fig. S4-A and S4-B, respectively. The corresponding plots of I_pa as a function of the square root of the scan rate (Fig. S4-C) displayed a linear trend (R² = 0.997), indicating a diffusion-controlled process at the surface of the 3D-CB/rGTR/LDPE-PT electrode. In contrast, an R² of 0.898 was obtained for the 3D-CB/rGTR/LDPE electrode, likely due to the limited exposure of CB on the electrode surface, caused by the presence of the polymeric material, which may ‘block’ the carbon active sites, resulting in slower electron transfer kinetics.⁷⁰ Based on the diffusion coefficient of FcMeOH (7.6 × 10⁻⁶ cm² s⁻¹), the electrochemically active surface area of the 3D-CB/rGTR/LDPE and 3D-CB/rGTR/LDPE-PT electrodes was determined using the Randles–Ševčík equation.^70–72 The electroactive surface area was 0.05 cm² for the 3D-CB/rGTR/LDPE electrode and 0.12 cm² for the 3D-CB/rGTR/LDPE-PT electrode, showing a 2.4-fold increase due to the modification. The electroactive surface area value for the 3D-CB/rGTR/LDPE-PT electrode, lower than the geometric area of 0.22 cm², reflects the amount of non-conductive material remaining after the modification. Furthermore, this modest increase in electroactive area does not fully account for the surface porosity and roughness, which are not considered in the Randles–Ševčík equation. Therefore, the electrical double layer capacitance (C_dl) values were calculated to better represent the effect of the modification of the electrode surface.

As capacitance is directly proportional to the electroactive area at the electrode surface,⁵² we also estimated C_dl from CV signal data of blank solution (0.5 mol L⁻¹ KCl), as shown in Fig. S5-A and S5-B. C_dl was significantly higher for the 3D-CB/rGTR/LDPE-PT electrode (C_dl = 1830 μF cm⁻²) compared to the 3D-CB/rGTR/LDPE electrode (C_dl = 13.16 μF cm⁻²), indicating a 139-fold increase in the electroactive area, Fig. S5-C.⁵¹ These values corroborate the morphological and spectroscopic characterizations, which revealed partial removal of the polymer matrix (rGTR and LDPE), increased porosity, and greater exposure of CB particles. The C_dl values obtained from CV are consistent with those expected for printed and modified electrodes.^33,51 Thus, it is worth highlighting that the C_dl values obtained better reflect the increase in the electroactive area.

3.3 3D-CB/rGTR/LDPE-PT sensor performance for detecting carbendazim fungicide

As a proof of concept, the electrochemical response of the 3D-CB/rGTR/LDPE-PT electrode for detecting CBZ was assessed. First, we investigated the effect of pH on the electrochemical response of CBZ using differential pulse voltammetry (DPV), at different pH values (2.0–10.0) in BR buffer (Fig. S6-A). The best electrochemical responses were observed at pH 2.0 and 4.0, with similar peak current values. However, at pH 4.0, CBZ oxidation occurred at a lower potential, with an anodic peak around +1.0 V (vs. Ag|AgCl|KCl_(sat.)), making it the optimal pH for subsequent tests. As the pH increased, the oxidation peak potential shifted negatively, suggesting the involvement of protons in the reaction. The peak current decreased in the pH range 6–10, likely due to the lower stability and deprotonation of CBZ in alkaline media, as shown in Fig. S6-B.^43,73 A linear relationship between pH and oxidation peak potential was observed (Fig. S6-B), described using the equation E_p = −0.058 (±0.004) pH + 1.202 (±0.025) (R² = 0.979). The slope (−58 mV per pH) is close to theoretical value (−59.1 mV per pH), indicating an equal transfer of protons and electrons during CBZ electrochemical oxidation.^43,74 In fact, studies indicate a 2H⁺/2e⁻ mechanism, likely occurring at the protonated nitrogen of the imidazole ring (Fig. S6-C).^75,76

The oxidation response of CBZ was evaluated in two different supporting electrolytes at pH 4.0: BR buffer and acetate buffer, in Fig. S6-D. A higher current response was obtained using the acetate buffer, which is also simpler to prepare and requires fewer reagents. For these reasons, the acetate buffer was selected as the supporting electrolyte for further CBZ detection.

The optimization of DPV instrumental parameters was evaluated to enhance the oxidation response of CBZ (Fig. S7). The effect of amplitude (ranging from 10 to 100 mV), shown in Fig. S7-A and S7-B, demonstrates a linear increase in peak current with amplitude. To maximize the CBZ response, an amplitude of 90 mV was selected. The modulation time (10–60 ms), shown in Fig. S7-C and D, revealed an increase in CBZ peak currents up to 30 ms, after which a significant decline in peak currents was observed. Thus, a modulation time of 30 ms was selected. The step potential (1–10 mV) was also evaluated in Fig. S7-E and F. The CBZ peak currents increased up to 4 mV, after which they stabilized. Since the step potential controls the measurement scan rate and higher values (above 6 mV) lead to greater measurement variability (Fig. S7-F), a step potential of 6 mV was selected for CBZ DPV analyses, resulting in a scan rate of 12 mV s⁻¹.

As observed in previous studies,^43,75 allowing CBZ to interact with the electrode surface for a certain period resulted in an enhancement of the electrochemical signal due to adsorption. Therefore, the effect of adsorption or accumulation time (5 to 50 s) of CBZ on the 3D-CB/rGTR/LDPE-PT electrode surface was analyzed, in Fig. S8-A and S8-B. As the adsorption time increased from 0 to 40 s, the peak current increased, indicating greater CBZ adsorption (see Fig. S8-B). However, after 40 s, the current decreased due to the saturation of the electrode's active sites (porous surface). To achieve high peak currents without compromising analytical efficiency, an optimal accumulation time of 30 s was selected for CBZ. Finally, the effect of stirring speed (200–1145 rpm) was also evaluated on the CBZ response (Fig. S8-C and S8-D). Stirring enhances adsorption by increasing analyte mass transfer to the electrode surface, continuously renewing the diffusion layer, and promoting a more uniform analyte distribution—thereby improving adsorption efficiency. Since higher stirring rates (above 767 rpm) resulted in increased measurement errors, this value was chosen for further analyses. An important observation is that pre-accumulation potentials were not required (accumulation performed under open-circuit conditions). Due to the highly porous nature of the surface, the species can easily adsorb onto it through interactions with the active sites. The optimized values for the DPV technique, along with the studied ranges mentioned above, are listed in Table S4.

Under optimized conditions, the analytical performance of the 3D-CB/rGTR/LDPE-PT electrode was evaluated for increasing concentrations of CBZ, ranging from 1.0 to 100.0 μmol L⁻¹, as shown in Fig. 5A. The calibration curve for DPV measurements (n = 3) is presented in Fig. 5B. A linear working range of 1.0 to 40.0 μmol L⁻¹ was established, with the following regression equation: I (μA) = 1.0976C_CBZ (μmol L⁻¹) − 0.2125 (R² = 0.997). The limit of detection (LOD) and limit of quantification (LOQ) were calculated as 0.079 and 0.262 μmol L⁻¹, respectively. The LOD of the proposed sensor is suitable for detecting CBZ in environmental samples. Chen and colleagues analyzed rice field water samples from China, reporting a concentration of 0.58 μmol L⁻¹.⁷⁷ These findings highlight the potential of our sensor for monitoring CBZ in aquatic environments, which is crucial for tracking environmental contaminants.

Fig. 5 (A) DPV responses obtained for successive additions of CBZ at concentrations ranging from 1.0 to 100.0 μmol L⁻¹. (B) Respective calibration plot. (C and E) DPV measurements of water samples (drinking water and river water, respectively) spiked with 2.0 μmol L⁻¹ CBZ (dark blue lines), followed by successive additions of 2.0 μmol L⁻¹ CBZ (red, green, lilac, and pink lines). (D and F) Corresponding calibration plots for CBZ obtained by the standard addition method. The dashed line represents the blank signal (0.04 mol L⁻¹ acetate buffer, pH 4.0). DPV conditions: modulation time = 30 ms; amplitude = 90 mV, step potential = 6 mV, and scan rate = 12 mV s⁻¹. Accumulation time = 30s; stirring speed = 767 rpm.

The intra-electrode precision of DPV measurements (n = 10) was assessed at two CBZ concentration levels (10.0 and 25.0 μmol L⁻¹), as shown in Fig. S9-A and S9-B. The relative standard deviation (RSD) was below 4.6% (Fig. S9-B), indicating that the fabricated sensor exhibits good precision. The analytical parameters used for the determination of CBZ are summarized in Table 1.

Table 1 Analytical parameters for CBZ determination using the 3D-CB/rGTR/LDPE-PT electrode

Analytical parameters	CBZ
a The limit of detection (LOD, 3(Sd/b)) and limit of quantification (LOQ, 10(Sd/b)) were determined using the standard deviation of 10 blank signals (Sb = 2.881 × 10^−8) measurements and the slope (b) of the calibration curve.^81 Values for ^#10.0 μmol L^−1 and ^##25.0 μmol L^−1 CBZ.
Linear range/μmol L^−1	1.0–40.0
R ^2	0.997
Intercept/μmol L^−1	–0.212 ± 0.007
Slope/μA L μmol^−1	1.098 ± 0.003
LOD^a/μmol L^−1	0.079 ± 0.001
LOQ^a/μmol L^−1	0.262 ± 0.006
RSD (intra-electrode, n = 10, 10.0 and 25.0 μmol L^−1 in 0.04 mol L^−1 acetate buffer, pH 4.0)/%	4.6^#/3.1^##

---

The applicability of the 3D-CB/rGTR/LDPE-PT electrode as a sensor for CBZ in environmental samples was verified by quantifying CBZ in four different water sources: drinking water, river water, lake water, and tap water, see Fig. 5C–F and S10. The recovery results from the standard addition method, presented in Table 2, ranged from 95–105%. These results confirmed that the 3D-CB/rGTR/LDPE-PT-based detection method is accurate and suitable for monitoring CBZ contamination in environmental water samples.

Table 2 Determination of CBZ in water samples using the 3D-CB/rGTR/LDPE-PT electrode

Sample	Added (μmol L^−1)	Found (μmol L^−1)	Recovery (%)
Values determined are the average response of three measurements for each sample.
Drinking water	2.0	1.93 (±0.06)	96 (±3)
River water		1.88 (±0.03)	95 (±2)
Lake water		2.11 (±0.08)	105 (±4)
Tap water		2.10 (±0.01)	105 (±1)

---

In addition, the reproducibility of the 3D-CB/rGTR/LDPE-PT sensor was evaluated using three different 3D-printed electrodes modified with a plasma pen. The tests were performed using drinking water as the matrix, as shown in Fig. S11. The RSD values of 5.68% for the peak current and 0.98% for the peak potential of CBZ in the sample demonstrate the good reproducibility and accuracy of the developed 3D-CB/rGTR/LDPE-PT sensor, even in complex sample matrices.

The intraday stability of the sensor for CBZ detection in the presence of a drinking water sample was also evaluated through more than 25 consecutive measurements performed on the same electrode, as shown in Fig. S12. Considering all the conclusions drawn (n = 28), an RSD value of 6.49% was obtained for the peak current, indicating the accuracy of the system and the potential for reusing the 3D-CB/rGTR/LDPE-PT electrode in successive analyses. It was also observed that, between the 25^th and 28^th measurements, the CBZ peak current decreased by approximately 15%, indicating a potential loss of sensitivity due to the electrode surface becoming less sensitive to the sample matrix. This behavior reinforces the need to re-modify the electrode surface after 25 consecutive measurements in the sample medium to maintain analytical performance and sensor reliability. This initially requires polishing the electrode with wet sandpaper for 30 s, followed by a second treatment with a plasma pen. This is particularly noteworthy because it demonstrates the potential for reusing the 3D-CB/rGTR/LDPE-PT sensor through a simple surface regeneration procedure.

The stability of the 3D-CB/rGTR/LDPE-PT sensor was assessed over days (Fig. S13). An RSD value of 1.62% was obtained for the CBZ peak current in the sample medium, indicating the potential for electrode reuse across multiple days, provided that appropriate storage conditions are maintained. In this study, the electrodes were stored in hermetically sealed packaging to prevent moisture ingress, which could compromise the electrode surface and reduce its sensitivity.

A selectivity study of the electrochemical response of CBZ was conducted in the presence of caffeine (CAF), ascorbic acid (AA), paracetamol (PAR), uric acid (UA), nitrite (NO₂⁻), zinc nitrate (ZnNO₃), nickel ions (Ni²⁺), copper nitrate (CuNO₃), and mercury ions (Hg²⁺) in the matrix media. These substances are commonly found in various water sources and may act as interfering agents, potentially posing risks to human health.^78–80 The selectivity study was performed in drinking water to simulate the potential presence of interfering compounds in the matrix, using a 1 : 1 ratio with 5.0 μmol L⁻¹ of CBZ, as shown in Fig. S14. Under the conditions tested, no redox processes associated with these species were observed, and the CBZ electrochemical signal showed a variation of less than 10% (relative response variation = 6.7%). These results indicate that the developed method has good specificity for CBZ and that the tested compounds do not contaminate the sensor surface. This can be explained by the CBZ accumulation step on the surface of the 3D-CB/rGTR/LDPE-PT sensor, which partially blocks the access of interfering species to the sensor's active sites, thus preventing the occurrence of redox reactions. Furthermore, most of the metal species studied require the application of reduction potentials, which were not used under the present experimental conditions. Regarding the organic compounds evaluated, within the optimized potential window, supporting electrolyte, pH, and other parameters optimized for CBZ, the only effect observed of these species was competitive adsorption by the sensor surface, without compromising the CBZ response.

Finally, the electrochemical performance of the proposed sensor was compared to previously reported devices (details in Table S5). The sensor proposed in this work, 3D-CB/rGTR/LDPE-PT, demonstrates a highly satisfactory linear range and LOD. These advantages were achieved through a simple and efficient methodology that enables the rapid production of electrodes via 3D printing, coupled with an environmentally friendly, reagent-free surface treatment that is both rapid (2 min) and effective. Additionally, the sensor benefits from signal enhancement due to the spontaneous interaction of CBZ (accumulation time of 30 s) with the electrode's porous and rough surface, eliminating the need for potential applications or prolonged waiting times, as required in other reported methods.

It is important to mention that the use of waste tire rubber to produce conductive filaments is a promising approach within the context of the circular economy and the development of more sustainable materials.^25,82 The strategy proposed contributes to mitigating a persistent source of solid waste and microplastics. It also helps reduce environmental impact through the use of 3D printing technology, which minimizes material waste compared to subtractive manufacturing processes. Furthermore, the surface treatment of the 3D-printed electrodes using a reagent-free plasma pen eliminates the need for hazardous chemicals.

This work aligns with the European Sustainable Development Goals, such as the European Union's Plastics Strategy for Plastics in a Circular Economy, by converting tire rubber into value-added conductive composites. Furthermore, this research supports several United Nations Sustainable Development Goals (SDGs), including SDG 12 (Responsible Consumption and Production), SDG 9 (Industry, Innovation and Infrastructure), SDG 11 (Sustainable Cities) and SDG 13 (Climate Action).⁸³ As reported in the literature,⁸² exploring open research directions in waste upcycling is important to construct sustainable material systems. Our study makes a practical contribution in this context by demonstrating the scalable and eco-friendly use of recycled carbon-based materials for emerging 3D printing technologies.

Although the presence of CB particles and plasticisers in tire rubber imparts conductivity and flexibility to the filaments, incorporating bio-based or more eco-friendly additives could further enhance the sustainability profile of these materials. Additionally, process optimization to reduce energy consumption during filament production and 3D printing would enhance the environmental performance of the technology. The conductive reclaimed tire rubber filament developed in this work also holds potential for applications in other electrochemical sensors, wearable devices, batteries, and supercapacitors, which offer cost-effective solutions for constructing sustainable electrochemical systems while supporting circular economy principles and responsible resource use.

4. Conclusion

This study introduces, for the first time, the development and characterization of electrodes fabricated from filaments composed of reclaimed ground tire rubber (rGTR), low-density polyethylene (LDPE) and carbon black (CB). The filament was analyzed using electrical conductivity/BDS, TGA, SEM, AFM, XPS, FT-IR, Raman spectroscopy, contact angle measurements, CV and C_dl, confirming the successful integration of the components and the effectiveness of surface plasma treatment. The conductive filament demonstrated outstanding electrochemical performance in redox probe tests, following treatment with an atmospheric air plasma jet pen, an efficient, reagent-free, and environmentally friendly process. With a LOD of 79 nmol L⁻¹, the 3D-printed sensor showed exceptional sensitivity for detecting the hazardous fungicide carbendazim, achieving recovery rates between 95% and 105% in environmental water samples. By converting waste tire rubber into high-performance, functional electrode materials, this work promotes a more sustainable production cycle, reduces environmental impact, and contributes meaningfully to the circular economy.

Conflicts of interest

There are no conflicts to declare.

Data availability statements

The data supporting this article have been included as part of the SI. See DOI: https://doi.org/10.1039/d5gc02259b. Schematic representation of how the plasma was applied on the 3D-CB/rGTR/LDPE electrode surface; electrical conductivity measurements; SEM images of the CB/rGTR/LDPE filament at low and high magnifications; scan rate dependence study and determination of electroactive surface areas; determination of the electrical double layer capacitance (C_dl); effect of pH and supporting electrolyte composition; optimization of the differential pulse voltammetry technique (DPV) parameters; repeatability study of the proposed method using 3D-CB/rGTR/LDPE-PT electrode; determination of CBZ and recovery studies in water samples; reproducibility study using different 3D-CB/rGTR/LDPE-PT electrodes in the presence of CBZ in drinking water samples; surface stability study of the 3D-CB/rGTR/LDPE-PT electrode during successive measurements, and different days in the presence of CBZ in drinking water samples; selectivity study.

Acknowledgements

The authors wish to express their gratitude for the financial support provided by the Brazilian funding agencies: CNPq (408462/2022-1, 405620/2021-7, 401977/2023-4, 308392/2022-1, 141177/2024-1, and 315838/2021-3); INCT of Bioanalytics (CNPq/INCT 465389/2014-7); CAPES (Financial Code 001), and the FAPEMIG (APQ-02067-23 and RED-00120-23). We also acknowledge the Laboratory Multiuser facilities (RELAM/PROPP) at the Federal University of Uberlândia (grant APQ-02391-22) for providing essential equipment and technical support for the experiments. G. P. S. acknowledges the doctoral scholarship provided by the Fahrenheit Universities program (Gdańsk, Poland). We acknowledge Gdańsk University of Technology's support under the “Excellence Initiative-Research University” program enabling mobility of co-authors [2/1/2024/IDUB/I.2b/Es] (R. M.) [3/1/2025/IDUB/IV.2a/Eu] (J. R.) and [5/1/2024/IDUB/I.1a/No] (T. S.). The authors also thank the MMU Electrochemistry Research Group at Manchester Metropolitan University (Manchester, United Kingdom) for XPS measurements and M. Sc. Ana Clara M. Oliveira for the treatments of XPS data.

References

1. H. L. Chen, T. K. Nath, S. Chong, V. Foo, C. Gibbins and A. M. Lechner, SN Appl. Sci., 2021, 3, 437.
2. N. P. Ivleva, A. C. Wiesheu and R. Niessner, Angew. Chem., Int. Ed., 2017, 56, 1720–1739.
3. K. Ragaert, L. Delva and K. Van Geem, Waste Manage., 2017, 69, 24–58.
4. M. Cole, P. Lindeque, C. Halsband and T. S. Galloway, Mar. Pollut. Bull., 2011, 62, 2588–2597.
5. R. Geyer, J. R. Jambeck and K. L. Law, Sci. Adv., 2017, 3, e1700782.
6. J. Kirchherr, D. Reike and M. Hekkert, Resour., Conserv. Recycl., 2017, 127, 221–232.
7. I. E. Nikolaou and K. P. Tsagarakis, Sustainable Prod. Consum., 2021, 28, 600–609.
8. W. Lipińska, A. Łepek, S. Wolff, J. Ryl, A. P. Nowak and K. Siuzdak, Sustainable Mater. Technol., 2025, 45, e01434.
9. H. Monteiro, I. Ribeiro, M. Gonçalves, M. Iten and N. S. Caetano, Energy Rep., 2022, 8, 270–276.
10. W. Li, Q. Wang, J. Jin and S. Li, Int. J. Life Cycle Assess., 2014, 19, 1833–1842.
11. R. Feraldi, S. Cashman, M. Huff and L. Raahauge, Int. J. Life Cycle Assess., 2013, 18, 613–625.
12. T. Amari, N. J. Themelis and I. K. Wernick, Resour. Policy, 1999, 25, 179–188.
13. Z. Xiao, A. Pramanik, A. K. Basak, C. Prakash and S. Shankar, Cleaner Mater., 2022, 5, 100115.
14. P. Rosa, C. Sassanelli, A. Urbinati, D. Chiaroni and S. Terzi, Int. J. Prod. Res., 2020, 58, 1662–1687.
15. L. Sun, Y. Wang, G. Hua, T. C. E. Cheng and J. Dong, Resour., Conserv. Recycl., 2020, 162, 105035.
16. B. D. Hettiarachchi, M. Brandenburg and S. Seuring, Int. J. Prod. Econ., 2022, 246, 108414.
17. I. A. Ferreira, R. Godina and H. Carvalho, Sustainability, 2020, 13, 234.
18. A. L. Petsiuk and J. M. Pearce, Addit. Manuf., 2020, 36, 101473.
19. A. Martínez-García, M. Monzón and R. Paz, in Additive Manufacturing, Elsevier, 2021, pp. 395–408.
20. P. Santander, F. A. Cruz Sanchez, H. Boudaoud and M. Camargo, Resour., Conserv. Recycl., 2020, 154, 104531.
21. J. Faludi, C. M. Van Sice, Y. Shi, J. Bower and O. M. K. Brooks, J. Cleaner Prod., 2019, 212, 1580–1590.
22. T. M. Tavares, G. M. D. Ganga, M. Godinho Filho and V. P. Rodrigues, Sustainable Prod. Consum., 2023, 37, 369–388.
23. R. Chen, H. Yin, I. S. Cole, S. Shen, X. Zhou, Y. Wang and S. Tang, Chemosphere, 2020, 259, 127452.
24. T. D. Ngo, A. Kashani, G. Imbalzano, K. T. Q. Nguyen and D. Hui, Composites, Part B, 2018, 143, 172–196.
25. C. Kalinke, R. D. Crapnell, P. R. de Oliveira, B. C. Janegitz, J. A. Bonacin and C. E. Banks, Global Challenges, 2024, 8, 2300408.
26. E. Sigley, C. Kalinke, R. D. Crapnell, M. J. Whittingham, R. J. Williams, E. M. Keefe, B. C. Janegitz, J. A. Bonacin and C. E. Banks, ACS Sustainable Chem. Eng., 2023, 11, 2978–2988.
27. C. Kalinke, R. D. Crapnell, E. Sigley, M. J. Whittingham, P. R. de Oliveira, L. C. Brazaca, B. C. Janegitz, J. A. Bonacin and C. E. Banks, Chem. Eng. J., 2023, 467, 143513.
28. I. V. S. Arantes, R. D. Crapnell, E. Bernalte, M. J. Whittingham, T. R. L. C. Paixão and C. E. Banks, Anal. Chem., 2023, 95, 15086–15093.
29. R. D. Crapnell, E. Sigley, R. J. Williams, T. Brine, A. Garcia-Miranda Ferrari, C. Kalinke, B. C. Janegitz, J. A. Bonacin and C. E. Banks, ACS Sustainable Chem. Eng., 2023, 11, 9183–9193.
30. M. Cieślik, A. Rodak, A. Susik, N. Wójcik, M. Szociński, J. Ryl and K. Formela, Materials, 2023, 16, 1307.
31. R. G. Rocha, D. L. O. Ramos, L. V. de Faria, R. L. Germscheidt, D. P. dos Santos, J. A. Bonacin, R. A. A. Munoz and E. M. Richter, J. Electroanal. Chem., 2022, 925, 116910.
32. A. Koterwa, I. Kaczmarzyk, S. Mania, M. Cieslik, R. Tylingo, T. Ossowski, R. Bogdanowicz, P. Niedziałkowski and J. Ryl, Appl. Surf. Sci., 2022, 574, 151587.
33. G. P. Siqueira, R. G. Rocha, A. B. Nascimento, E. M. Richter and R. A. A. Muñoz, Anal. Chem., 2024, 96, 15852–15858.
34. T. O. Hara and B. Singh, ACS ES&T Water, 2021, 1, 462–478.
35. A. Sharma, V. Kumar, B. Shahzad, M. Tanveer, G. P. S. Sidhu, N. Handa, S. K. Kohli, P. Yadav, A. S. Bali, R. D. Parihar, O. I. Dar, K. Singh, S. Jasrotia, P. Bakshi, M. Ramakrishnan, S. Kumar, R. Bhardwaj and A. K. Thukral, SN Appl. Sci., 2019, 1, 1446.
36. R. Umapathi, S. M. Ghoreishian, S. Sonwal, G. M. Rani and Y. S. Huh, Coord. Chem. Rev., 2022, 453, 214305.
37. K.-H. Kim, E. Kabir and S. A. Jahan, Sci. Total Environ., 2017, 575, 525–535.
38. M. Bilal, H. M. N. Iqbal and D. Barceló, Sci. Total Environ., 2019, 695, 133896.
39. PAN Europe , European Parliament blocks residues of hazardous pesticides in imported food, https://www.pan-europe.info/press-releases/2024/09/european-parliament-blocks-residues-hazardous-pesticides-imported-food, (accessed 9 April 2025).
40. M. A. Dias, V. S. Santos, B. C. Vizioli, B. S. Ferreira and C. C. Montagner, Chemosphere, 2025, 372, 144093.
41. G. Bellisai, G. Bernasconi, A. Brancato, L. Carrasco Cabrera, L. Ferreira, G. Giner, L. Greco, S. Jarrah, A. Kazocina, R. Leuschner, J. O. Magrans, I. Miron, S. Nave, R. Pedersen, H. Reich, S. Ruocco, M. Santos, A. P. Scarlato, A. Theobald, B. Vagenende and A. Verani, EFSA J., 2021, 19, 6777.
42. G. Bellisai, G. Bernasconi, M. Binaglia, L. Carrasco Cabrera, I. Castellan, A. F. Castoldi, A. Chiusolo, F. Crivellente, M. Del Aguila, L. Ferreira, G. Giner Santonja, L. Greco, F. Istace, S. Jarrah, A. Lanzoni, R. Leuschner, I. Mangas, A. Mioč, S. Nave, M. Panzarea, J. M. Parra Morte, R. Pedersen, H. Reich, S. Ruocco, A. P. Scarlato, M. Szot, A. Terron, A. Theobald, M. Tiramani and A. Verani, EFSA J., 2024, 22, e8569.
43. X. Chen, W. Li, C. Lu, J. Chu, R. Lin, P. Wang, G. Xie, Q. Gu, D. Wu and B. Chu, Sci. Total Environ., 2022, 851, 158324.
44. S. Xue, J. Zou, J. Li, J. Xu, H. Chen, L. Wang, Y. Gao, X. Duan and L. Lu, Food Chem., 2024, 456, 140063.
45. A. Susik, A. Rodak, J. Cañavate, X. Colom, S. Wang and K. Formela, Materials, 2023, 16, 1788.
46. P. J. Elving, J. M. Markowitz and I. Rosenthal, Anal. Chem., 1956, 28, 1179–1180.
47. P. Wiśniewska, S. Wang and K. Formela, Waste Manage., 2022, 150, 174–184.
48. S. V. F. Castro, R. M. Cardoso, M. H. P. Santana, E. M. Richter and R. A. A. Munoz, Talanta, 2019, 203, 106–111.
49. G. P. Siqueira, D. A. G. Araújo, L. V. de Faria, D. L. O. Ramos, T. A. Matias, E. M. Richter, T. R. L. C. Paixão and R. A. A. Muñoz, Chemosphere, 2023, 340, 139796.
50. E. M. Richter, D. P. Rocha, R. M. Cardoso, E. M. Keefe, C. W. Foster, R. A. A. Munoz and C. E. Banks, Anal. Chem., 2019, 91, 12844–12851.
51. H. H. Bin Hamzah, O. Keattch, D. Covill and B. A. Patel, Sci. Rep., 2018, 8, 9135.
52. M. Areir, Y. Xu, R. Zhang, D. Harrison, J. Fyson and E. Pei, J. Manuf. Process., 2017, 25, 351–356.
53. Protopasta Filament by Protoplant , Protopasta Conductive PLA, https://proto-pasta.com/, (accessed 6 May 2025).
54. J. Zienkiewicz, M. Karczewska, S. Wang and K. Formela, Polimery, 2024, 69, 357–361.
55. A. Nadal Gisbert, J. E. Crespo Amorós, J. López Martínez and A. M. Garcia, Polym.-Plast. Technol. Eng., 2007, 47, 36–39.
56. M. Sadaf, M. Bragaglia and F. Nanni, J. Manuf. Process., 2021, 67, 141–150.
57. P. Fan and C. Lu, J. Polym. Environ., 2011, 19, 943–949.
58. M. J. Glowacki, M. Cieslik, M. Sawczak, A. Koterwa, I. Kaczmarzyk, R. Jendrzejewski, L. Szynkiewicz, T. Ossowski, R. Bogdanowicz, P. Niedzialkowski and J. Ryl, Appl. Surf. Sci., 2021, 556, 149788.
59. U. Cvelbar, B. Markoli, I. Poberaj, A. Zalar, L. Kosec and S. Spaić, Appl. Surf. Sci., 2006, 253, 1861–1865.
60. X. Colom, M. Marín-Genescà, R. Mujal, K. Formela and J. Cañavate, J. Compos. Mater., 2018, 52, 3099–3108.
61. A. T. Nunes, R. E. dos Santos, J. S. Pereira, R. Barbosa and J. D. Ambrósio, Prog. Rubber, Plast. Recycl. Technol., 2018, 34, 143–157.
62. M. A. Pimenta, G. Dresselhaus, M. S. Dresselhaus, L. G. Cançado, A. Jorio and R. Saito, Phys. Chem. Chem. Phys., 2007, 9, 1276–1291.
63. P. Rochowski, M. Grzegorczyk and S. Pogorzelski, J. Colloid Interface Sci., 2019, 555, 352–360.
64. W. B. Veloso, T. R. L. C. Paixão and G. N. Meloni, Anal. Chem., 2024, 96, 14315–14319.
65. J. C. Zuaznabar-Gardona and A. Fragoso, J. Electroanal. Chem., 2020, 871, 114314.
66. R. Chauhan, R. Fogel, C. Purcarea, G. Necula-Petrareanu, P. Fanjul-Bolado, D. Ibañez, A. Vasilescu, R. M. Banciu and J. Limson, Carbon Trends, 2024, 17, 100408.
67. J. Ghilane, F. Hauquier and J.-C. Lacroix, Anal. Chem., 2013, 85, 11593–11601.
68. M. Stoytcheva, Z. Velkova, V. Gochev, B. Valdez and M. Curiel, Int. J. Electrochem. Sci., 2025, 20, 100924.
69. J. F. S. Pereira, R. G. Rocha, S. V. F. Castro, A. F. João, P. H. S. Borges, D. P. Rocha, A. de Siervo, E. M. Richter, E. Nossol, R. V. Gelamo and R. A. A. Muñoz, Sens. Actuators, B, 2021, 347, 130651.
70. J. S. Stefano, L. R. Guterres e Silva, R. G. Rocha, L. C. Brazaca, E. M. Richter, R. A. Abarza Muñoz and B. C. Janegitz, Anal. Chim. Acta, 2022, 1191, 339372.
71. L. Chen, E. Kätelhön and R. G. Compton, Appl. Mater. Today, 2019, 16, 141–145.
72. C. Amatore, N. Da Mota, C. Sella and L. Thouin, Anal. Chem., 2007, 79, 8502–8510.
73. A. Yamuna, T.-W. Chen, S.-M. Chen and T.-Y. Jiang, Microchim. Acta, 2021, 188, 277.
74. Y. Zhou, R. Cui, Y. Dang, Y. Li and Y. Zou, Sens. Actuators, B, 2019, 296, 126680.
75. M. V. S. Sant'Anna, S. W. M. M. Carvalho, A. Gevaerd, J. O. S. Silva, E. Santos, I. S. C. Carregosa, A. Wisniewski, L. H. Marcolino-Junior, M. F. Bergamini and E. M. Sussuchi, Talanta, 2020, 220, 121334.
76. R. Cui, D. Xu, X. Xie, Y. Yi, Y. Quan, M. Zhou, J. Gong, Z. Han and G. Zhang, Food Chem., 2017, 221, 457–463.
77. X. Chen, W. Li, C. Lu, J. Chu, R. Lin, P. Wang, G. Xie, Q. Gu, D. Wu and B. Chu, Sci. Total Environ., 2022, 851, 158324.
78. C. Wu, X. Huang, J. D. Witter, A. L. Spongberg, K. Wang, D. Wang and J. Liu, Ecotoxicol. Environ. Saf., 2014, 106, 19–26.
79. J. A. Camargo and Á. Alonso, Environ. Int., 2006, 32, 831–849.
80. P. Yuan, X. Wu, Y. Xia, C. Peng, H. Tong, J. Liu, L. Jiang and X. Wang, Environ. Sci. Pollut. Res., 2020, 27, 35556–35566.
81. J. Mocak, S. Mitchell, G. Scollary and A. M. Bond, Pure Appl. Chem., 1997, 69, 297–328.
82. M. Chu, W. Tu, Q. Zhang, W. Huang, J. Chen and D. Ma, Nat. Rev. Clean Technol., 2025, 1, 320–332.
83. United Nations , The 17 Goals, https://sdgs.un.org/goals, (accessed 27 July 2025).

---

Footnote

† These authors contributed equally to this paper.

---