Activated Carbon as a Cathode for Water Disinfection through the Electro-Fenton Process

Abstract

Unlike many other water disinfection methods, hydroxyl radicals (HO^•) produced by the Fenton reaction (Fe²⁺/H₂O₂) can inactivate pathogens regardless of taxonomic identity of genetic potential and do not generate halogenated disinfection by-products. Hydrogen peroxide (H₂O₂) required for the process is typically electrogenerated using various carbonaceous materials as cathodes. However, high costs and necessary modifications to the cathodes still present a challenge to large-scale implementation. In this work, we use granular activated carbon (GAC) as a cathode to generate H₂O₂ for water disinfection through the electro-Fenton process. GAC is a low-cost amorphous carbon with abundant oxygen- and carbon-containing groups that are favored for oxygen reduction into H₂O₂. Results indicate that H₂O₂ production at the GAC cathode is higher with more GAC, lower pH, and smaller reactor volume. Through the addition of iron ions, the electrogenerated H₂O₂ is transformed into HO^• that efficiently inactivated model pathogen (Escherichia coli) under various water chemistry conditions. Chick–Watson modeling results further showed the strong lethality of produced HO^• from the electro-Fenton process. This inactivation coupled with high H₂O₂ yield, excellent reusability, and relatively low cost of GAC proves that GAC is a promising cathodic material for large-scale water disinfection.

1. Introduction

The burden of waterborne pathogens is significant and can lead to severe health complications [1,2]. To minimize the likelihood of contamination, drinking water is typically treated using a multi-barrier approach including disinfection and maintenance of a disinfectant residual during distribution of drinking water to the consumers. Most commonly used disinfection methods include ozonation, chlorination, chloramination, UV irradiation, and their combinations [3,4,5]. While each of these methods can be effective, they have key limitations. For instance, the electrochemical production of ozone relies on high voltage [6,7], and storage and transport of gaseous ozone is risky due to its high reactivity, thus it is usually produced onsite. Chlorination can result in formation of carcinogenic disinfection byproducts (DBP) [8], which may require pre-chlorination approaches to minimize DBP formation and additional significant efforts to manage post-chlorination DBP concentrations [9]. UV irradiation kills microorganisms by damaging double-stranded DNA [10,11], but this strategy is ineffective in turbid conditions; moreover, cells might reverse the DNA damage through a repair mechanism [12,13,14].

Hydroxyl radicals generated from H₂O₂ decomposition are effective in inactivating varying cell types, such as bacteria, eukaryotes, and viruses [15,16]. HO^• inactivates microbes primarily through the destruction of vital cellular components, including surface proteins, NAD(P)H, and DNA [17]. Post HO^• exposure, bacteria usually exhibit ruptured morphology [18], which indicates cell surface damage and osmotic shock due to leakage of cellular components. The severe destruction of cell structure is the most plausible reason for strong disinfection potential of HO^•. Despite its effectiveness, the most significant challenge with the large-scale application of the Fenton process is the prohibitive cost of hydrogen peroxide (H₂O₂) (~USD $1.5/kg for bulk purchase) [19].

Recently, electrogeneration of H₂O₂ via two-electron reduction of oxygen has drawn considerable interest. With specific cathodes, dissolved oxygen accepts two electrons and converts to H₂O₂ through an oxygen reduction reaction [Equation (1)] [20,21]. Many types of carbon materials, such as reticulated vitreous carbon foam, graphite felt, carbon felt, and carbon nanotubes, have been used as cathodes [22]. Previous reports suggest that oxygen- and nitrogen-containing functional groups on the surface are particularly beneficial for electron transfer to oxygen [23]. Therefore, a large number of modification methods (e.g., acid/base incubation, pyrolysis of coated resins, etc.) to the abovementioned carbonaceous materials are being tested to enhance H₂O₂ titer [24,25]. However, these costly modifications are not ideally suitable for large-scale applications.

$${\mathrm{O}}_2+2H^{+}+2e^{-}\to H_2{O}_2(0.695\mathrm{V}\mathrm{vs}.\mathrm{SHE},\mathrm{standard}\mathrm{hydrogen}\mathrm{electrode})$$

(1)

Granular activated carbon (GAC) has long been used as an adsorbent to remove environmental contaminants [26]. It has been reported that electrochemical polarization, which creates an anode on one side and a cathode on the other side, could regenerate GAC adsorbent beds after saturated adsorption of organic contaminants [27,28]. During this process, H₂O₂ is generated on the cathode side, indicating that GAC has two-electron electrochemical oxygen reduction activity [28]. To date, studies using an entire GAC as a specific cathode to electrogenerate H₂O₂ are relatively lacking. Further, GAC is likely the cheapest functional carbon material, adding further viability towards its large-scale implementation in the future. In this study, we aim to evaluate the feasibility of using pristine GAC for water disinfection with Escherichia coli as a model pathogen. Electrogeneration of H₂O₂ was measured under various conditions, and HO^• from the reaction between iron ion and H₂O₂ was then used to inactivate E. coli. The development of the GAC-based electro-Fenton process for water disinfection is especially desirable in low-income regions and could meet the needs of their sustainable development. Besides, the novel application of amorphous GAC as a cathode has great implication in that other amorphous carbon materials (such as biochar manufactured from biomass precursor) could also be used for hydrogen peroxide generation. It is overall anticipated that this study could advance the application of low-cost carbons for electrochemical water treatment purpose.

2. Results and Discussion

2.1. Generation of Hydrogen Peroxide with GAC Cathode

The used GAC had a specific surface area of 840.5 m²/g, providing sufficient catalytic sites for H₂O₂ generation. Carbon- (C1s) and oxygen-containing (O1s) groups are abundant on the GAC surface [28] and increased the hydrophilicity and mediated the facile generation of H₂O₂ [23]. Typically, functional groups such as carboxyl and etheric groups are particularly favored for oxygen reduction into H₂O₂ [29]. We fabricated the cathode by loading a certain amount of GAC into a stainless-steel mesh. This hybrid cathode was tightly wrapped by a rubber band to ensure sufficient contact between the stainless-steel mesh and the conductive surface of GAC (Figure S1 of Supplementary Material). A Ti/MMO mesh anode was used to produce oxygen instead of air ventilation for H₂O₂ generation. The mechanism for H₂O₂ generation was initially explored in our previous work [28]. The GAC cathode produced a significant amount of H₂O₂ in the rage of 3–20 mg/L after 60 min (

Table 1. Yields of hydrogen peroxide and determined current efficiencies after reaction for 60 min under various conditions.

Granular Activated Carbon (GAC) Amount (g)	pHini	Current (mA)	Volume (mL)	Granularity (mesh)	pHfinal	H2O2 Yield (mg/L)	Current Efficiency
1	7	100	200	4–8	3.54 ± 0.16	4.12 ± 0.19	1.30%
2	7	100	200	4–8	3.51 ± 0.14	7.44 ± 0.30	2.35%
3	7	100	200	4–8	3.42 ± 0.11	14.64 ± 0.65	4.62%
3	3	100	200	4–8	2.85 ± 0.07	20.36 ± 0.68	6.42%
3	5	100	200	4–8	3.44 ± 0.13	14.90 ± 0.68	4.70%
3	9	100	200	4–8	3.46 ± 0.14	13.03 ± 0.52	4.11%
3	11	100	200	4–8	3.77 ± 0.19	12.74 ± 0.46	4.02%
3	7	50	200	4–8	4.21 ± 0.14	3.38 ± 0.11	2.13%
3	7	75	200	4–8	4.04 ± 0.18	5.79 ± 0.26	2.43%
3	7	150	200	4–8	3.49 ± 0.14	10.70 ± 0.43	2.25%
3	7	100	150	4–8	3.41 ± 0.12	16.34 ± 0.59	3.86%
3	7	100	300	4–8	3.51 ± 0.12	12.47 ± 0.42	5.90%
3	7	100	400	4–8	3.49 ± 0.16	10.69 ± 0.49	6.74%
3	7	100	200	4–12	3.42 ± 0.14	14.37 ± 0.57	4.53%
3	7	100	200	4–14	3.47 ± 0.13	11.66 ± 0.42	3.68%
3	7	100	200	< 5 mm	3.54 ± 0.12	16.11 ± 0.54	5.08%

). This is particularly important, because other functional, higher priced carbonaceous materials typically produce 12–36 mg/L H₂O₂ in similar settings [30,31,32].

2.2. Parameters that Influenced H₂O₂ Yield

We tested various parameters that influenced the H₂O₂ yield (Table 1). Increasing GAC mass from 1 to 3 g in 200 mL solution increased H₂O₂ yield from 4.12 to 14.64 mg/L. The improved yield of H₂O₂ was primarily due to the increase in reactive sites with increasing GAC mass per electrolyte volume. Moreover, the electrogeneration of H₂O₂ by GAC cathode was more effective under acidic pH, as a higher concentration of protons could mediate a faster reaction rate. When the pH was reduced from 5 to 3, the H₂O₂ yield increased by 36.6% from 14.90 to 20.36 mg/L. However, the change of pH from 11 to 5 did not result in such drastic change in H₂O₂ production (i.e., 12.74 to 14.90 mg/L), because low proton concentration under this pH range limited H₂O₂ formation. We further tested H₂O₂ yield at different currents, while other parameters were fixed (3 g GAC, 200 mL volume, pH_ini 7). Increasing current from 50 to 100 mA increased H₂O₂ yield from 3.38 to 14.64 mg/L. Further increase in the current to 150 mA decreased H₂O₂ yield to 10.7 mg/L. This was presumably because of the parasitic H₂O₂ decomposition on the anode [28]. We assessed the effect of reactor volume, which is important for large-scale application in practice. Reaction conditions were 3 g GAC, 100 mA current, and pH_ini 7. Results showed that increasing reactor volume from 200 to 400 mL decreased the H₂O₂ yield from 14.64 to 10.69 mg/L. The current efficiency, however, increased from 4.62% to 6.74%, indicating a higher accumulated H₂O₂ amount, primarily because the consumption pathway of H₂O₂ was weakened at low H₂O₂ concentration. These results suggest that larger reactor volume mediates greater utilization efficiency of the electrons.

The electrogeneration of H₂O₂ tended to acidify the solution, which was desirable for the Fenton reaction. This was because, during the electrogeneration of H₂O₂, the proton generation exceeded hydroxyl ion generation. As a result, solution pH was between 2.8 and 4.2 in assays after 60 min reaction (Table 1), and such acidity was suitable for the subsequent Fenton reaction.

It is important to note that the ability of GAC to generate H₂O₂ is feasible under various granularities. We tested GAC of varying sizes (i.e., 4–8 mesh, 4–12 mesh, 4–14 mesh, and 5 mm) and found that they consistently produced H₂O₂ with a yield of 11–16 mg/L, exhibiting approximately similar oxygen reduction activities. This was likely because oxygen reduction to H₂O₂ occurred on the catalytic sites on both internal pores and the external surface, and the GAC granularity thus did not affect the total number of catalytic sites. This indicated that the electrogeneration of H₂O₂ could occur on a variety of GAC types, suggesting the practical applicability of this method.

2.3. Disinfection of E. coli through Electro-Fenton Process

In this study, electrogenerated H₂O₂ was used for water disinfection of E. coli as a model pathogen. The disinfection efficiency of H₂O₂ alone is relatively low, because the genetically encoded catalases can efficiently decompose H₂O₂ into water and oxygen [33]. In contrast, HO^• from the Fenton reaction could effectively kill a wide variety of microorganisms regardless of their cellular structure or genetic potential [15,16].

Electrogenerated H₂O₂ was transformed into highly oxidizing HO^• in the presence of ferrous iron ions. The process of electrogeneration of H₂O₂ from the GAC cathode and transformation into HO^• by iron ions is known as the electro-Fenton process. The HO^• yield was measured with the benzoic acid quantification method [34]. Iron ions concentrations of 0.05, 0.2, and 0.4 mM resulted in the production of 125.9, 224.53, and 241.8 µM HO^• after 60 min of the electro-Fenton process, respectively (Figure 1a), with no significant difference in HO^• generated between 0.2 mM and 0.4 mM concentrations. Moreover, the GAC cathode at 100 mA mediated the highest HO^• yield compared with other electric currents (Figure 1b), indicating that there is an optimum value for the current.

2.3.1. Mechanism of Electro-Fenton Disinfection

Treatment by either GAC cathode or ferrous ions alone mediated 1.15- and 0.53-log reduction in cell viability, respectively. The electro-Fenton process coupling both GAC cathode and ferrous ions led to a 2.40-log inactivation of E. coli cells, as the generated HO^• was a strong bactericide (Figure 2a,b). The GAC cathode under 100 mA inactivated E. coli mainly via the combination of generated H₂O₂ and acidification (pH decreased to 3.45, Table 1), whereas H₂O₂ at 2 mM or acid pH of 3 individually led to less than 0.46-log of bacterial inactivation (Figure S3 of Supplementary Material). H₂O₂ slightly inactivated bacteria, likely through the oxidation of sulfhydryl groups of functional enzymes [35]. Acidic pH denatured the protein structure and may also have increased the permeability of the cell membrane [36]. In addition, ferrous ions were slightly bactericidal after oxidation into ferric ions (Fe³⁺), which could have denatured the functional enzymes through the oxidation of side chains of amino acids [37].

Nonetheless, HO^• likely played a major role in E. coli inactivation. Addition of 50 mM MOPS buffer (3-(N-morpholino)propanesulfonic acid) to attenuate solution acidification of the electro-Fenton process (Figure 3a) decreased H₂O₂ yield from 14.64 to 4.13 mg/L, using the GAC as a cathode after a 60 min reaction. Consistently, HO^• yield was also reduced from 224.53 to 64.15 µM in the electro-Fenton process. Furthermore, the addition of 50 mM MOPS buffer to the electro-Fenton process resulted in 0.53-log E. coli cell inactivation due to reduced HO^• production (Figure 3c), which was significantly less than the electro-Fenton disinfection without the buffer (i.e., 2.40-log). The use of 50 mM thiourea to quench H₂O₂ and HO^• (Figure 3b) resulted in negligible concentrations of H₂O₂ or HO^• during the course of the electro-Fenton reactions, and concomitant decrease in the inactivation of E. coli cells (1.11-log) was observed after 60 min (Figure 3c). Taking these results together, we concluded that the disinfection of E. coli by the electro-Fenton process was primarily due to the HO^• from electrogenerated H₂O₂ under activation of ferrous ions in an acidic pH.

2.3.2. Complete Disinfection of E. coli of Various Concentrations

Continuous monitoring of H₂O₂ and HO^• yields over 300 min (Figure 4a) revealed that H₂O₂ generation was rapid during the first 40 min then slowed down. H₂O₂ concentration reached a plateau of 16.74 mg/L at 100 min. The measured H₂O₂ concentration was the total H₂O₂ generation minus H₂O₂ decomposition [Equation (10)]. Total electrogeneration rate of H₂O₂ was determined by applied electric potential and dissolved oxygen content, while H₂O₂ decomposition was largely affected by H₂O₂ concentration. Therefore, as the H₂O₂ concentration in bulk solution increased, its decomposition rate increased due to the more prominent parasitic pathways [Equations (3)–(6)]. When the H₂O₂ generation rate was equal to the decomposition rate, an equilibrium concentration (plateau) was reached in agreement with electrogeneration of H₂O₂ on other cathodes [30,32].

$${\left[H_2{O}_2\right]}_{measured}={{\displaystyle \int }}_0^t{r}_{{\left[H_2{O}_2\right]}_{generation}}dt-{{\displaystyle \int }}_0^t{r}_{{\left[H_2{O}_2\right]}_{decomposition}}dt$$

(2)

Parasitic H₂O₂ decomposition pathways:

$$H_2{O}_2+2H^{+}+2e^{-}\to 2H_{2}O$$

(3)

$$H_2{O}_2\to {HO}_2^{•}+H^{+}+e^{-}$$

(4)

$${HO}_2^{•}\to O_2+H^{+}+e^{-}$$

(5)

$$2H_2{O}_2\to H_{2}O+O_2$$

(6)

We added 0.2 mM ferrous ions into the solution to transform H₂O₂ into HO^•. The generated Fe³⁺ after the reaction could then be reduced into Fe²⁺ on the cathode and again used for H₂O₂ activation. Hence, the HO^• was constantly generated in the presence of both the GAC cathode and the iron ion, as quantified via the addition of benzoic acid (Figure 4a). The HO^• yield at 100 min was 279.61 µM and reached 379.84 µM after 300 min.

The disinfection of E. coli at 10⁴, 10⁶, and 10⁸ CFU/mL was tested respectively over 300 min (Figure 4b). The electro-Fenton process could completely remove 10⁴ CFU/mL E. coli within 60 min, and 5.02-log E. coli removal was achieved after 180 min at an initial concentration of 10⁶ CFU/mL. The process also inactivated 6.1-log E. coli after 300 min with an initial concentration of 10⁸ CFU/mL. The results indicate that GAC as a cathode could efficiently inactivate E. coli at varying concentrations.

2.3.3. Chick–Watson Model

E. coli inactivation kinetics was fitted with the first-order Chick–Watson model. CT values (disinfectant concentration × contact time) for electric current, H₂O₂, and HO^• as input disinfectant were individually plotted against decimal logarithm of bacterial survival rate [i.e., ln(N/N₀)] (Figure 5). The Chick–Watson coefficients (Λ_CW) for electric current, H₂O₂, and HO^• were determined to be 0.248 A⁻¹ min⁻¹, 1.63 × 10⁻³ mg⁻¹ L min⁻¹, and 6.52 × 10⁵ mg⁻¹ L min⁻¹, respectively. Chick–Watson model fitting with HO^• as disinfectant showed the highest correlation coefficient (R² = 0.984) compared with fitting results of electric current (R² = 0.903) and H₂O₂ (R² = 0.843), presumably because HO^• is the species directly responsible for E. coli inactivation. Interestingly, the Λ_CW of HO^• in this study was 3.9 times greater than the reported value under neutral conditions (1.33 × 10⁵ mg⁻¹ L min⁻¹) [38], possibly because other parameters such as acidic conditions synergistically enhanced the bactericidal role of HO^•. Moreover, the lethality of HO^• (2.70 V) was several orders higher than ozone (50 mg⁻¹ L min⁻¹, 2.07 V), chlorine (15.4 mg⁻¹ L min⁻¹, 1.36 V), or chlorine dioxide (25 mg⁻¹ L min⁻¹, 1.91 V), primarily due to its strong oxidation potential [38].

2.4. Resistance to Water Alkalinity

The generation of HO^• radical bactericide from the Fenton reaction relies on an acidic pH. However, most natural water bodies can maintain buffering capacity, which is called water alkalinity, because of carbonate ions from atmospheric carbon dioxide or sediment rock mineral. The typical alkalinity of waters is below 200 mg CaCO₃/L [39,40]. Thus, large amounts of acid are necessary to overcome such high alkalinity, or the efficiency of the traditional Fenton process for water treatment would be severely compromised. Given that produced protons on the anode exceed released hydroxyl ions on the cathode, the electro-generation of H₂O₂ led to an automatic acidification of the solution, which facilitated the Fenton reaction. However, little is known about solution acidification by the electrogeneration of H₂O₂ in carbonated media.

The effectiveness of the electro-Fenton process toward E. coli disinfection was therefore tested under two representative alkalinity conditions. It was observed that electrogeneration of H₂O₂ tended to acidify the solution regardless of carbonate concentration, and solution pH decreased to 3.45 in the presence of 2 mM carbonate after 60 min. Moreover, the presence of 1–2 mM carbonate ion exhibited negligible impact on H₂O₂ generation (Figure 6). Specifically, the electrogenerated H₂O₂ yields were 14.64, 13.31, and 13.91 mg/L by GAC cathode with 0, 1, and 2 mM carbonate, respectively. As a result, the disinfection efficiency by the electro-Fenton process was not affected by the presence of carbonate ions. This further highlights the practicality of the electro-Fenton process for water disinfection under buffered conditions.

2.5. Reusability of GAC Cathode

The active sites on the GAC surface are under cathodic potential during the electro-Fenton process. This could result in the reductive elimination of functional groups of certain carbonaceous materials, and the potency to produce H₂O₂ might drastically decrease [30]. GAC is known to harbor abundant functional groups that exist both on the external surface and the internal pore surface [41]. Thus, it is important to assess the effect of long-term cathodic current-mediated oxygen reduction on the catalytic activity of GAC.

We tested the durability of the GAC cathode for multiple rounds of application for water disinfection. After each round of the electro-Fenton reaction, the GAC was extensively washed with diluted sulfuric acid and then rinsed with Milli-Q water. The cleaned and oven-dried GAC was then applied to the next round of the electro-Fenton reaction. The H₂O₂ yields were sustained between 11.34 and 14.64 mg/L without notable decrease for a minimum of five rounds of application and regeneration (Figure 7a). Moreover, with the addition of ferrous ions, the stable H₂O₂ generation consistently led to 2.26–2.60 log reduction of E. coli in each round due to the production of HO^• radicals (Figure 7b).

2.6. Disinfection of E. coli with Antibiotic-Resistance Genes

Bacteria hosting antibiotic-resistance genes (ARGs) have been an emerging concern in water treatment, and they are resistant to common antibiotic drugs due to inherent efflux pumps [42]. In addition, ARGs of a bacterial species could be relayed to another bacterial species through horizontal gene transfer [43]. Previous research reported that, among all the water disinfection technologies, only advanced oxidation processes generating oxidative radicals (such as HO^•) could potently damage ARGs and effectively prevent horizontal gene transfer [44].

Electrogeneration of H₂O₂ with the GAC cathode is a practical method and is affordable by most water treatment plants. We utilized the electro-Fenton process to disinfect E. coli with various ARGs. Antibiotic-resistant E. coli cells were obtained after transferring plasmids encoding corresponding ARGs into competent cells followed by antibiotic selection on a Luria-Bertani (LB)-agar plate. Results indicated that the electro-Fenton process could non-selectively inactivate E. coli with different ARGs (e.g., ampicillin-, kanamycin-, tetracyclin-, chloramphenicol-, zeocin-, and spectinomycin-resistance genes). For instance, 2.18–2.41 log of E. coli cells were reduced after 1 h treatment (Figure S4 of Supplementary Material). Additionally, the complete disinfections of E. coli cells were obtained after 300 min (Figure 8) regardless of the hosted ARGs, highlighting the great promise of the process. However, this was not surprising, because HO^• disinfection mainly occurs through destruction of cellular structure and is not related to genetic potential.

2.7. Techno-Economic Analysis

GAC as a porous carbon material has long been used as an adsorbent for the removal of water contaminants. Owing to the various functional groups on the surface, the GAC cathode is reported to mediate hydrogen peroxide formation via an oxygen reduction reaction. An important advantage of the GAC cathode is the relatively low cost. The market price of GAC is around USD $0.75 per kilogram in a large-scale purchase [45]. Moreover, GAC can be reused for multiple rounds without loss of catalytic activity, further lowering the operation cost. GAC can also be produced from the pyrolysis of biomass, such as coconut shell [46] or other waste materials [47]. The utilization of waste biomass to generate H₂O₂ for water treatment is especially attractive, considering that both water treatment and biomass treatment consume large amounts of cost and energy. The application of GAC as a cathode for water disinfection is thereby promising.

Unlike traditional water treatments, which rely on a single dose of a high concentration of oxidants, the electro-Fenton process continuously generates a low concentration of H₂O₂. The low concentration of H₂O₂ is instantly converted into HO^• in the presence of ferrous ions. In fact, the continuous and mild generation of oxidants is more favored than the single dose of high concentration of oxidants in practice, because the former method holds great promise to address the contaminants rebound phenomena [48,49]. Besides the contaminants rebound, residual water pathogens after a single treatment can also proliferate with a supply of nutrients. The developed electro-Fenton process is anticipated to completely eliminate water pathogens with continuous generation of H₂O₂.

3. Materials and Methods

3.1. Materials

Granular activated carbon (GAC, 4–8 mesh) as a cathode for oxygen reduction was obtained from Calgon Carbon Corporation. GACs of other sizes, i.e., 4–12 mesh, 4–14 mesh, and < 5 mm, were obtained from Calgon Carbon Corporation as well. Sodium sulfate (Na₂SO₄, ≥99%), titanium sulfate (TiSO₄, 99.9%), ferrous sulfate (FeSO₄), hydrogen peroxide (30% wt.), sodium carbonate (Na₂CO₃), and benzoic acid (C₇H₆O₂, >99%) were purchased from Fisher Scientific (Hampon, NH, USA). MOPS buffer (C₇H₁₅NO₄S) was purchased from Corning Inc. (Corning, NY, USA). Thiourea (CH₄N₂S, >99%) as a quencher of H₂O₂ and HO^• was purchased from ACROS Organics Inc. Luria-Bertani (LB) broth and agar for E. coli enumeration were purchased from BD, Difco Inc. Meshed Ti/mixed metal oxide (Ti/MMO, 3N International) was used as anode material.

3.2. Characterizations of Granular Activated Carbon

Nitrogen adsorption/desorption isotherm of the employed GAC was measured at −196 °C (ASAP 2420 V2.05). Then, the Brunauere–Emmete–Teller (BET) specific surface area was calculated from the isotherm curve. The volumes of micropore and mesopore were estimated using the Horvath–Kawazoe method and the Barrett–Joynerand–Halenda method, respectively. The pore size distribution was determined using the nonlocal density functional theory by the adsorption branch, and the total pore volume was calculated from the nitrogen amount adsorbed at a relative pressure of 0.975. Results showed that the GAC was microporous with an average pore width of less than 1 nm, and the specific surface area was 840.5 m²/g.

3.3. Batch Experiments for Disinfection

Before the experiments, 10 g GAC was pre-treated by vigorously stirring in 300 mL Milli-Q water with a magnetic bar for 10 min. The eluent was discarded, and GAC pellets were re-washed with 300 mL Milli-Q water. The washing steps were repeated until the eluent was clear. The GAC particles were then dried at 80 °C for 2–3 h. Log growth phase E. coli culture [200 mL, 10⁸ CFU/mL (CFU, colony forming unit)] was used for electro-Fenton disinfection experiments. Specifically, a single colony of E. coli K12 from an LB-agar plate was seeded in LB nutrient broth and cultured at 37 °C overnight. Then, 1% of the fully-grown bacteria solution was diluted in a sterile flask containing fresh LB medium and cultured for around 2 h until the log phase (OD₆₀₀ ~ 0.8–1). Bacteria were harvested by centrifuging the E. coli culture at 10,000 g for 10 min, followed by intensive washing with Milli-Q water. The centrifugation and the washing steps were repeated three times, and the washed E. coli culture was used for experiments. During the electro-Fenton process, 3 g GAC held in a stainless-steel mesh was used as a cathode under 100 mA current with 10 mM Na₂SO₄ electrolyte and 0.2 mM FeSO₄ as the catalyst. A schematic configuration of the reactor is shown in Figure S1 of Supplementary Material. Every 20 min, the bacterial solution was withdrawn for subsequent quantification.

3.4. E. coli Quantification Method

E. coli was quantified through a serial dilution method as reported previously [50,51,52]. Briefly, the sampled E. coli solution was put in the first row of a sterilized 96-well plate. The ten-fold serial dilution was performed by taking 20 µL well-mixed bacterial solution into the next row containing 180 µL phosphate buffered saline (PBS) physiological buffer. A total of 5 µL liquid from the serial dilution wells was plated on the LB-agar plate and incubated at 37 °C overnight. The survival rate was calculated from the CFU after treatment divided by that before treatment [Equation (7)].

$$\mathrm{Survival}\mathrm{rate}=\frac{\mathrm{bacteria}\mathrm{after}\mathrm{treatment}\left(\mathrm{CFU}/\mathrm{mL}\right)}{\mathrm{bacteria}\mathrm{before}\mathrm{treatment}\left(\mathrm{CFU}/\mathrm{mL}\right)}\times {\log }_{10}10$$

(7)

3.5. Hydrogen Peroxide and Hydroxyl Radical Quantification Method

The generated H₂O₂ using the GAC cathode without iron ions was quantified by complexing with titanium sulfate [53]. The absorbance of the developed yellow color was measured with a UV-Vis spectrometer at 405 nm. H₂O₂ generation was quantified under various conditions, e.g., GAC amount, solution pH, reactor volume, GAC granularity, and current intensity. During the electro-Fenton process, HO^• was generated from the Fenton (Fe²⁺/H₂O₂) reaction using the GAC cathode in the presence of iron ions and quantified using benzoic acid. Three grams of GAC hosted by stainless-steel mesh were transferred into a 200 mL solution of 10 mM Na₂SO₄ and 10 mM benzoic acid at pH 7 overnight under constant stirring. Benzoic acid saturated GAC was then used as a cathode under 100 mA for the electro-Fenton reaction with the addition of 0.2 mM FeSO₄. The primary products from HO^• oxidizing benzoic acid were mainly o-, m-, and p-hydroxybenzoic acid, and the stoichiometric ratio of HO^• to produced p-hydroxybenzoic acid was determined to be 5.87 ± 0.18 [34]. Hence, HO^• yield was calculated from p-hydroxybenzoic acid concentration multiplied by 5.87. p-Hydroxybenzoic acid concentration was measured with a high-performance liquid chromatography (HPLC, Agilent 1200 Infinity Series) equipped with an Agilent Eclipse AAA C18 column (4.6 × 150 mm) with 0.5 mL/min methanol/1% phosphoric acid (20/80) used as the mobile phase and p-hydroxybenzoic acid with a retention time of 5 min and was detected at 210 nm wavelength using an Agilent 1260 diode array detector.

3.6. Current Efficiency Calculation

The current efficiency (CE) for H₂O₂ generation, defined as the ratio of the electrons consumed by the two-electron oxygen reduction reaction over the total electrons passing through the circuit, was calculated by Equation (8).

$$CE=\frac{nF{C}_{H_2{O}_2}V}{{{\displaystyle \int }}_0^{t}Idt}\times 100\%$$

(8)

where n is the number of electrons transferred for O₂ reduction to H₂O₂, F is the Faraday constant (96486 C/mol), C_H2O2 is the concentration of H₂O₂ (mol/L), V is the electrolyte volume (L), I (A) is the current at time t, and t is the reaction time (s).

3.7. Reusability Assay of Granular Activated Carbon

After each round of the electro-Fenton reaction, GAC was washed with 0.1 M H₂SO₄ acid solution to dissolve iron precipitates under vigorous stirring for 3 h. The cleaned GAC particles were then collected and rinsed with Milli-Q water to remove residual acid until eluent liquid had neutral pH. The regenerated GAC was further dried in an oven at 80 °C for 2–3 h prior to utilization as a cathode for the subsequent electro-Fenton process.

3.8. Disinfection of Antibiotic-Resistant Bacteria

E. coli K12 strains harboring ampicillin-, kanamycin-, tetracycline-, chloramphenicol-, zeocin-, and spectinomycin-resistance biomarkers were obtained by transferring pET-32a(+), pET24a(+), pTet, pCam, pCC-zeo, and pCC-spc plasmids (Figure S2 of Supplementary Material), respectively, into E. coli K12 competent cells. The bacteria with successful transfers of the biomarker gene were selected on an LB-agar plate with corresponding antibiotics (i.e., 30 μg/mL ampicillin, 50 μg/mL kanamycin, 20 μg/mL tetracycline, 20 μg/mL chloramphenicol, 25 μg/mL zeocin, and 100 μg/mL spectinomycin). Then, 10⁸ CFU/mL of various antibiotic-resistant E. coli cells in the log phase were disinfected by the electro-Fenton process using GAC cathode under 100 mA with 10 mM Na₂SO₄ electrolyte and 0.2 mM FeSO₄ at pH 7.

3.9. Chick–Watson Model

In the Chick–Watson model, the bacterial disinfection efficiency (ln(N/N₀)) is correlated with both disinfectant concentration (C) and contact time (T), which can be expressed as:

ln(N/N₀) = Λ_CWCT

(9)

where Λ_CW is the Chick–Watson coefficient. Electric current, generated H₂O₂, and produced HO^• as input disinfectants were investigated individually. Specifically, CT value for electric current was 0.1 t (A•min) under a constant 100 mA current.

For CT value of generated H₂O₂:

$${{\displaystyle \int }}_0^t{\left[H_2{O}_2\right]}_{t}dt\approx {{\displaystyle \sum }}_0^{t}0.5({\left[H_2{O}_2\right]}_t+{\left[H_2{O}_2\right]}_{t+\Delta t})\Delta t$$

(10)

For CT value of produced HO^•:

$${{\displaystyle \int }}_0^t{\left[{\mathrm{HO}}^{•}\right]}_{t}dt\approx {{\displaystyle \sum }}_0^{t}0.5\left({\left[{\mathrm{HO}}^{•}\right]}_t+{\left[{\mathrm{HO}}^{•}\right]}_{t+\Delta t}\right)\Delta t$$

(11)

During disinfection experiments, the steady-state HO^• concentration was estimated using a similar approach to Cho et al. [38] by adding 0.5 mM benzoic acid as a probe compound. The decay of benzoic acid obeyed the following equations:

HO^• + benzoic acid → products

(12)

$$-\frac{d\left[benzoic acid\right]}{dt}=k_7\left[{\mathrm{HO}}^{•}\right]\left[benzoic acid\right]$$

(13)

$${\left[{\mathrm{HO}}^{•}\right]}_t=-\frac{d\left[benzoic acid\right]}{k_7\left[benzoic acid\right]dt}\approx -\frac{2\Delta \left[benzoic acid\right]}{k_7\left({\left[benzoic acid\right]}_t+{\left[benzoic acid\right]}_{t+\Delta t}\right)\Delta t}$$

(14)

where k₇ is 1.8 × 10⁹ M⁻¹ s⁻¹ [54].

4. Conclusions

In this work, we used activated carbon as a cathode to generate H₂O₂ for the purpose of water disinfection. Ferrous ions were added to transform H₂O₂ into strong bactericide HO^•. Parameters including GAC loading, currents, solution pH, and reactor volume were investigated. The highest concentration of H₂O₂ reached 20.36 mg/L under optimal conditions. With ferrous ions, H₂O₂ was transformed into HO^• to inactivate E. coli. This electro-Fenton process could inactivate 10⁸ CFU/mL E. coli after 300 min regardless of bacterial ARGs. Moreover, the electro-Fenton process automatically acidified the solution and was resistant to a high concentration of buffering carbonates in water. Considering the cost-effectiveness of GAC, this electro-Fenton process is promising in practical water treatment applications. Future work includes the upscaling of this process to test its feasibility in large-scale practice.