Approaching easy water disinfection for all: Can in situ electrochlorination outperform conventional chlorination under realistic conditions?

1 Breakpoint electrochlorination in ammonia removal: Unveiling the impact of convective mass transfer Aksana Atrashkevicha, Dominic Vardaa, Kaitlyn Yeagera, Maria Gomez-Mingota,b, Carlos M. SánchezSáncheza,c, Sergi Garcia-Seguraa,* aNanosystems Engineering Research Center for Nanotechnology-Enabled Water Treatment, School of Sustainable Engineering and the Built Environment, Arizona State University, Tempe, AZ bLaboratoire de Chimie des Processus Biologiques, Collège de France, UMR 8229 CNRS, Sorbonne Université, PSL Research University, 11 Place Marcelin Berthelot, 75005 Paris, France cSorbonne Université, CNRS, Laboratoire Interfaces et Systèmes Electrochimiques (LISE), 4 Place Jussieu, 75005 Paris, France Article submitted to be published in Water Research https://doi.org/10.1016/j.watres.2025.123732 Corresponding author: *e-mail: Sergio.garcia.seg[email protected]du (Dr. Sergi Garcia-Segura) © 2025. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/ 2 Abstract Breakpoint chlorination, the point at which ammonia is completely oxidized by chlorine to nitrogen gas, may occur during electrochemical water treatment due to the simultaneous abundance of inorganic nitrogen species and chloride ions in many water matrices. Nevertheless, little is known about the difference between the chemical breakpoint chlorination and electrochemical ammonia abatement as well as the impact of the electrode-electrolyte interface that drives the breakpoint electrochlorination. This study investigates the influence of the interface on ammonia oxidation by comparing indirect breakpoint electrochlorination with chemical approach and by examining the impact of varying convective mass transfer on breakpoint electrochlorination. Our results revealed that, under identical conditions and bulk pH, breakpoint electrochlorination releases much lower residual chlorine species in the bulk solution before ammonia is oxidized, as compared to chemical breakpoint chlorination. It was observed that lower convective mass transfer not only accelerates ammonia removal but also increases the chlorine evolution reaction. Results from a closed divided cell experiment confirmed that chlorine evolution is enhanced under lower convective mass transfer, which suggests a relevant role of species distribution within electrodeelectrolyte interface. We hypothesize that this effect may be due to a more acidic local pH under lower mass transfer conditions, which favors chlorine evolution over oxygen evolution reaction. These findings provide insights into the fundamental differences of chemical breakpoint chlorination and indirect breakpoint electrochlorination. The results can guide operating strategies for electrochemical water treatment that can potentially reduce energy consumption by lowering flow speeds, while achieving higher chlorine yield and faster ammonia removal. Keywords: Breaking point chlorination; electrochlorination; chlorine evolution reaction (CER); dimensionally stable anode (DSA); Ti/IrO2; diffusion layer 3 1. Introduction Chlorine-based disinfectants are widely used in water treatment to eliminate pathogenic microorganisms. Chlorine species' effectiveness comes from both their strong antimicrobial activity and long-lasting residuals, which ensures primary and secondary disinfection in extensive water distribution networks (Bhandari et al., 2010). Water chlorination can be achieved through the addition of chlorine gas (Cl2) or denser chlorine phases (i.e., NaOCl, Ca(OCl)2) (Crittenden et al., 2012). When chlorine gas dissolves in water (eq 1), it rapidly disproportionates (eq 2) to form hypochlorous acid (HOCl), protons (H+) and chloride ions (Cl‒) (Snoeyink and Jenkins, 1980). The distribution of HOCl and its conjugate base, hypochlorite ion (OCl‒), depends on the water pH, with both species being present in equal concentrations at pH 7.5 at 25 ˚C (eq 3). Cl2(g) ⇌Cl2(aq) 𝐾𝐻= 6.2 x 10−2 (eq 1) Cl2(aq)+ H2O ⇌ HOCl + H++Cl− (eq 2) HOCl ⇌ H++ OCl− p𝐾𝑎= 7.5 at 25oC (eq 3) Chlorination is also commonly used in water treatment as a method to remove ammonia species (NH4+/NH3) from water. Ammonia in water causes an unpleasant taste and smell (World Health Organization, 2022) and represents a large portion of total N-pollutants reaching wastewater treatment plants in urban areas (Akkari et al., 2025). Furthermore, ammonia is toxic to aquatic organisms and its removal is encouraged. The presence of ammonia in water increases the chlorine demand, as it consumes added free chlorine forming chloramines: monochloramine (NH2Cl) following (eq 4), dichloramine (NHCl2) according (eq 5), and trichloramine (NCl3) through (eq 6). The sum of chloramine species is also commonly known as combined chlorine. The consumption of free chlorine through these reactions may interfere with the disinfection capabilities since chloramines are weaker disinfectants. Excess of free chlorine can induce the complete oxidation of chloramines to nitrogen gas through the process known as breakpoint chlorination. Achieving complete total NH4+/NH3 oxidation requires a certain chlorine-to- 4 ammonia as nitrogen (Cl2/NH4+ - N) ratio that is based on the reaction stoichiometry (eq 7) (Crittenden et al., 2012). NH3+ HOCl → NH2Cl + H2O (eq 4) NH2Cl + HOCl → NHCl2+ H2O (eq 5) NHCl2+ HOCl → NCl3+ H2O (eq 6) 2NH3+ 3HOCl → N2+3H++ 3Cl−+3H2O (eq 7) Reactions involved in breakpoint chlorination are not only used in centralized water treatment to remove ammonia, but also can be used to generate NH2Cl (eq 4) as disinfectant. Due to its lower reactivity, NH2Cl has a longer lifespan than free Cl2, making it essential for disinfection in extended distribution networks, which enables large-scale water supply over long distances. While the chemistry of homogeneous chemical breakpoint chlorination is well-understood (Crittenden et al., 2012; Saunier and Selleck, 1979), ammonia abatement during chlorine electrogeneration, which is a heterogeneous reaction due to an electrode surface involvement, remains less investigated. Chlorine evolution reaction (CER) or electrochlorination occurs on the anode surface via oxidation of Cl‒ to Cl2 as described by the general expression of (eq 8) (Karlsson and Cornell, 2016; Vanýsek, 1978). 2Cl−⇌Cl2(g) +2e− 𝐸o= 1.36 V vs.SHE (eq 8) So far this reaction has been mainly extensively studied in highly concentrated NaCl solutions (35 M) within the frame of the chlor-alkali process (Trasatti, 1987). In the chlor-alkali industry, divided reactors are commonly used with highly concentrated NaCl solutions to maximize chlorine gas production in the anodic compartment and sodium hydroxide in the cathodic chamber (Bhandari et al., 2010). Electrochlorination has gained recently significant attention for its applicability in decentralized settings for waters with much lower Cl‒ content than used in chlor-alkali industry. Electrochlorination can occur at Cl‒ concentration as low as the usual range of concentrations found in drinking waters of 50-250 mg L-1 (Atrashkevich and Garcia-Segura, 2024). However, previous studies have shown that even extremely low chloride levels, such as those found in stormwater (e.g., 6 mg L-1), can be sufficient to enable electrolocation 5 (Feng et al., 2018). Electrochlorination offers sustainable, off-grid solution for onsite chlorine generation, producing chlorine doses sufficient for contaminant removal, while can be easily powered by renewable energy sources (Atrashkevich et al., 2024; Otter et al., 2020, 2019; Scialdone et al., 2021). Unlike chemical chlorination, electrochlorination does not require storage and handling of hazardous chemicals or even salts, since Cl‒ are ubiquitously present in many water matrices (Baba et al., 2020; Felisardo et al., 2024; Ghaitidak and Yadav, 2013; Olichwer et al., 2013). During electrochemical water treatment, the mechanism of breakpoint chlorination can be induced through the co-existence of NH4+/NH3 and electrogenerated Cl2/HOCl/OCl‒. It has been demonstrated that using a Ti/PtOx-IrO2 anode, the oxidation of NH4+ - N to N2 occurs in close proximity to the anode and this has been attributed to high Cl2 concentration and low pH at the anode (Kapałka et al., 2010). However, the influence of the electrode-electrolyte interface on ammonia oxidation through electrogenerated chlorine species is not yet fully understood. This study investigates the role of the electrode-electrolyte interface during breakpoint electrochlorination. By comparing electrochlorination and chemical chlorination for ammonia oxidation under identical experimental conditions, we aim to elucidate the role of the interface in the reaction. Since mass transfer controls the thickness of the stagnant layer of electrolyte at the electrode surface, we examined the effects of convective mass transfer on the breakpoint electrochlorination and CER alone. Our findings provide unexpected insights that are essential for optimizing treatment processes involving electrochlorination, resulting in increased current efficiency, and improved yields of both ammonia oxidation and chlorine evolution reaction. 2. Material and methods 2.1 Chemicals and solutions Sodium chloride (NaCl, ≥ 99.0%), sodium perchlorate (NaClO4, ≥ 98.0%), sodium nitrate (NaNO3, ≥ 97.0%), and potassium chloride (KCl, ≥ 99.0%) were used as electrolytes. Hydrochloric acid (HCl, 37.0%) and nitric acid (HNO3, ≥ 70.0%) solutions were used as electrolytes for cyclic voltammetry experiments. Bicarbonate buffer was prepared using sodium bicarbonate (NaHCO3, ≥ 99.7%) and sodium 6 carbonate (Na2CO3, ≥ 99.5%). Acetic buffer was prepared using sodium acetate (NaCH3COO, ≥ 99.7%) and acetic acid (CH3COOH, ≥ 99.5%). Ammonium chloride (NH4Cl, ≥ 99.5%) was used as an ammonia source. Ferrocenemethanol or FcMeOH (C11H12FeO, ≥ 97%) was used to approximate mass transfer coefficient and diffusion layer thickness. Chemical chlorination experiments used hypochlorite (NaOCl, 10-15% available chlorine) as an active chlorine species source. Sodium hydroxide (NaOH, ≥ 97.0%) solution was used to collect chlorine from the gas phase. All chemicals were supplied by Sigma-Aldrich. All solutions were prepared with ultrapure water provided from an ultrapure water system Elga Water with resistivity >18.2 MΩ cm at 20 °C. 2.2 Electrochemical experiments 2.2.1. Bulk electrolysis in an undivided open batch reactor Electrochemical experiments in undivided open batch reactor (Figure S1a) were conducted using a power supply TENMA 72-2710 applying a current of 0.12 A. The cell was equipped with two parallel electrodes with a geometric surface area of 6 cm2 (i.e., 2.5 cm x 2.4 cm) each and located maintaining a 1.0 cm distance between them. Titanium plate was used as a cathode and DSA® Ti/IrO2 mesh supplied by DeNora was used as an anode. Solutions of 200 mL containing 84 mM NaCl or NaOCl4 electrolytes were treated under magnetic stirring at various rotations per minute (RPM) (i.e., 0 RPM, 100 RPM, 500 RPM, and 1300 RPM) using a magnetic stirrer with a cylindrical shape with 25 mm length and 8 mm diameter and a stirring plate ThermoScientific. Identifying an optimal operating window free from parasitic reactions with Ti/IrO2 electrodes is challenging, as both the CER (eq 8) and oxygen evolution reaction (OER) (eq 12) occur simultaneously under neutral pH conditions and diluted chloride concentrations. The applied current was chosen so that the resulting anode potential slightly exceeded the standard reduction potential for CER (eq 8), taking into account the chloride concentration in the solution and the low overpotential of CER for Ti/IrO2. The anode potential during the electrolysis was 1.45 ± 0.05 V vs. Ag/AgCl (i.e., 1.65 ± 0.05 V vs. SHE), ensuring CER under used settings. An Ag/AgCl (3 M KCl) reference electrode purchased from BASi 7 was used and located in close proximity to the anode. The duration of electrolysis in the undivided open batch reactor was 60 minutes. 2.2.2. Bulk electrolysis in a divided closed batch reactor Electrochemical experiments in the divided closed reactor were conducted using a closed H-cell reactor (Figure S1b). The Nafion 117 membrane supplied by FuelCellStores was set dividing the anodic and cathodic compartments. All experiments in the divided closed reactor were conducted using a potentiostat/galvanostat PGSTAT302N Metrohm under chronopotentiometry mode applying a current of 0.12 A. Identical electrodes with the same dimensions as those used in the undivided open cell were employed in the divided closed cell. Due to the increased distance between the electrodes approaching 9.5 cm, under the applied current of 0.12 A, the potential was not able to reach the value of 1.45 ± 0.05 V vs. Ag/AgCl (i.e., 1.65 ± 0.05 V vs. SHE), which was observed in the undivided open batch cell. Thus, the electrolyte concentrations were increased from 84 mM to 150 mM of NaCl or NaOCl4 to maintain identical anodic potential as the observed in the undivided open cell. The volume of the electrolyte in each compartment was 200 mL, holding a total volume of 400 mL in the closed divided cell. The dimensional volume of each chamber was 500 mL. Thus, an empty head space of 300 mL was left in each compartment. Anolyte of 200 mL was mixed under magnetic stirring at various RPM (i.e., 0 RPM, 100 RPM, 500 RPM, and 1300 RPM) using an identical magnetic stirrer and a stirring plate as in the undivided open reactor. As a reference electrode, which was set up in the anodic chamber, an Ag/AgCl (3 M KCl) reference electrode was used. The closed anodic compartment was purged before experiments using Ar gas. The chlorine from gas phase was trapped at the end of the experiments by purging the headspace of the cell with Ar as carrier gas for 15 minutes towards an alkaline 0.5 M NaOH solution to trap evolved chlorine through the enhanced disproportionation of Cl2 in alkaline media to hypochlorite (OCl‒). Then, trapped chlorine species were quantified. 2.3 Analytical techniques and electrolysis performance parameters 8 The water pH and conductivity were measured with a ThermoScientific Orion Star A221 pH-meter and A322 conductivity meter, respectively. All species concentrations were measured using spectrophotometer HACH DR6000 UV-vis. The concentrations of total (Cl2/HOCl/OCl‒ and NH2Cl/NHCl2/NCl3) and free chlorine (Cl2/HOCl/OCl‒) were quantified using N,N-diethyl-p-phenylene diamine (DPD) reagent with potassium iodide for total chlorine measurement and without potassium iodide present for free chlorine quantification. Absorbance of 530 nm wavelength for both free and total chlorine was used. The method is equivalent to colorimetric Standard Method 4500-Cl G (Baird et al., 2017). Monochloramine was quantified using indophenol method utilizing HACH kits powder pillows containing cyanoferrate and phenol at absorbance wavelength of 655 nm (Lee et al., 2007). The concentrations of NH4+ - N species were quantified using the HACH TNT 830 kits that quantifies total ammonia concentration as the sum of N-containing species (NH4+/NH3 and NH2Cl/NHCl2/NCl3). All reagents were supplied by HACH. Cyclic voltammetry at Ti/IrO2 electrode was conducted at a scan rate of 10 mV s-1 using an Ag/AgCl (3 M KCl) reference electrode and Pt place as a counter electrode in a single-compartment cell. The solution volume was 50 mL and was purged with Ar gas for 15 minutes before the analysis. Faradaic efficiency (FE) provides estimation of the efficiency of charge transfer for the reaction of interest during the electrolysis (eq 9). 𝐹𝐸(%)=𝑄𝑟𝑒𝑎𝑐𝑡𝑖𝑜𝑛 𝑄𝑡𝑜𝑡𝑎𝑙 ×100% = 𝑛 𝐹 𝑁𝑖 𝐼 𝑡 ×100%, (eq 9) where 𝑄𝑟𝑒𝑎𝑐𝑡𝑖𝑜𝑛 is the charge consumed in the reaction of interest (C), 𝑄𝑡𝑜𝑡𝑎𝑙 is the total charge consumed during electrolysis (C), n is the amount of electrons required per mole of product (mol e-/mol), F is the Faraday constant (96 485 C mol-1), 𝑁𝑖 is the amount of product generated during the electrolysis (mol Cl2), I is the applied current (A), t is the time (s). To estimate approximate values of mass transfer coefficient (𝑚) and diffusion layer thickness (𝛿) at the anode under various RPM, we used well-established, one-electron transfer reaction of FcMeOH at a platinum (Pt) electrode with same geometrical area as for Ti/IrO2 of 6 cm2 (i.e., 2.5 cm x 2.4 cm). The 9 diffusion layer thickness at the electrode is time dependent without convection, but an electrode in stirred solution establishes a finite diffusion layer thickness. Thus, if the electrode reaction is mass-transfer controlled, the surface concentration of the electroactive species approaches zero and the current becomes independent of applied potential. In that case, the maximum current reached corresponds to the steady state limiting current (𝑖𝑙), which depends only on the hydrodynamic conditions and allows to estimate the diffusion layer thickness from equation (eq 10). 𝛿 = 𝐷 𝑚 (eq 10) where D is the diffusion coefficient of FcMeOH (7 x 10-10 m2 s-1) (Sánchez-Sánchez et al., 2008), 𝑚 is the mass transfer coefficient at various stirring conditions (m s-1). The mass transfer coefficient is obtained by determining the limiting current, as described by (eq 11) (Bard and Faulkner, 2000). 𝑚 = 𝑖𝑙 𝑛𝐹𝐴𝐶 (eq 11) where 𝐴 is the surface area of the electrode (m2), and 𝐶 is the bulk concentration of FcMeOH (mol m-3). Figures S4 and S5 show steady-state mass transfer limiting current for FcMeOH oxidation at 0.5 V vs. Ag/AgCl (i.e., 0.7 V vs. SHE) despite the hydrodynamic conditions in the reactor correspond to a turbulent regime. Chronoamperometry experiments were conducted at applying potential of 0.5 V vs. Ag/AgCl (i.e., 0.7 V vs. SHE) at various FcMeOH concentrations (Figures S5a to c) to approximate steadystate limiting current. The average limiting current values over the last 60 seconds of a 120-seconds chronoamperometry (Figure S5d) was used to calculate the mass transfer coefficient for each RPM and each FcMeOH concentration (eq 11). Final mass transfer coefficient value was obtained as an average of 0.5, 1.0, and 2.0 mM FcMeOH concentrations at each RPM (Figure S6a). Thus, Figure S6b shows the estimate values for the diffusion layer thickness under different stirring conditions demonstrating the decrease in the diffusion layer thickness upon increase in stirring speed. Although the diffusion layer 16 total ammonia similar to the observed for the electrochemical treatment (cf. Fig. 2a). A thorough analysis of chlorine containing species showcases relevant differences between breakpoint electrochlorination and chemical breakpoint chlorination at pH 4.0. Despite both treatments show a quick depletion of total ammonia in solution and the no significant accumulation of monochloramine, different trends are depicted by total chlorine. During chemical breakpoint chlorination at pH 4, large concentration of total chlorine residuals were observed and low concentrations of free Cl2 and NH2Cl indicated a significant accumulation of combined chlorine (i.e., NHCl2 and NCl3). Which is even more important, some of these species remain even at elevated Cl2/NH4+ - N weight ratio beyond the point where ammonia is oxidized (i.e., near 14 Cl2/NH4+ - N weight ratio). The low presence of monochloramine at acidic pH can be explained by the combined chlorine species distribution as function of pH. It is well known that at pH 4, only 3% of chloramines are present as NH2Cl, while 61% and 36% are found as NHCl2 and NCl3, respectively (Cimetiere, 2009). The preferred generation NHCl2/NCl3 at acidic pH and their stability can explain the observed increase in combined chlorine (the difference between total and free chlorine) (Jafvert and Valentine, 1992). The significantly lower accumulation of total and combined chlorine during breakpoint electrochlorination can therefore be explained by oxidation of total ammonia at the interface and the natural change of pH after reaching breakpoint that destabilize NHCl2 and NCl3 inducing their complete oxidation to N2 gas. Conversely, the chemical breakpoint chlorination conducted at pH 9.0 (see Fig. 2d and h) represents the other side of the coin where a major accumulation of monochloramine is observed with almost no stabilization of other combined chlorine which remains with no residual after reaching breakpoint (i.e., near 8 Cl2/NH4+ - N weight ratio). In this case, the trend for total ammonia abatement observed does not resemble the electrochlorination profile. These experimental results highlight once more the relevance of the interfacial reaction contribution driven at the electrode surface, as well as the relevance of the localized pH. This comparison provides a perspective that acidic pH within the electro-electrolyte interface would not facilitate ammonia oxidation because NH3 would be protonated and become NH4+ that can slow down the reaction. Thus, the observed more effective NH4+ - N oxidation during electrochlorination 17 compared to chlorination at pH 4.0 might be attributed to the large local Cl2 concentration within electrodeelectrolyte interface that locally elevates Cl2/NH4+ - N ratio. The detailed comparison of breakpoint electrochlorination and chemical breakpoint chlorination demonstrated notable differences. Breakpoint electrochlorination shows significantly lower pre-breakpoint residual total chlorine concentrations under identical conditions, confirming that electrogenerated chlorine was consumed by total NH4+ - N species within the anode interface. This suggests that large free Cl2 concentration facilitates NH4+ - N oxidation due to elevated Cl2/NH4+ - N weight ratio at the electrodeelectrolyte interface. These results suggest that the thickness of the diffusion layer formed at the electrode surface may be an influential factor during breakpoint electrochlorination. 3.2 Elucidating the role of convective mass transfer on breakpoint electrochlorination A concentration gradient is generated during electrolysis between the concentrations of the electroactive species in the bulk solution and at the electrode surface. In addition to diffusion and migration, convection is an important mass transfer phenomenon in electrochemical processes to ensure the concentration of all species uniform and equal within the bulk, except for a certain distance from the electrode surface, which is defined as the diffusion layer. The thickness of this diffusion layer is controlled by the effectiveness of the mass transfer phenomena and can be tuned by the convective conditions in the cell (Bard and Faulkner, 2000). The reaction yield at larger scale is significantly impacted by the effective mass transport from/towards the electrode. Therefore, to better understand the impact of diffusion layer thickness during breakpoint electrochlorination, we conducted experiments to evaluate breakpoint electrochlorination varying convective mass transfer conditions. Figures 3a to d illustrate the abatement of total ammonia during breakpoint electrochlorination in a batch reactor under different stirring velocities from 0 up to 1300 RPM. Note that active chlorine species are continuously electrogenerated during electrolysis and suppose a sustained increase over time of the active chlorine delivered or the Cl2/NH4+ - N weight ratio. As previously discussed, the total ammonia decreases linearly with no significant accumulation of monochloramine or other combined chlorine species 18 (i.e., NHCl2, NCl3). Total ammonia follows a linear abatement in agreement with literature (Romano et al., 2021). However, increased convective mass transfer reached at higher RPM resulted in slower NH4+ - N oxidation, despite increased flux of electroactive species. The rate of NH4+ - N oxidation decreased significantly from 0.67 ± 0.01 mg NH4+ - N L-1 min-1 at 0 RPM to 0.161 ± 0.006 mg NH4+ - N L-1 min-1 at 1300 RPM (Table S1). Figures 3e to 3h demonstrate that besides slower NH4+ - N oxidation rate at higher RPM, all residual chlorine species concentrations present were low at all tested RPM until NH4+ - N was completely oxidized. These results confirm that NH4+ - N oxidation occurs precisely within the stagnant layer at the anode surface, which is not impacted by convective mass transfer. An interesting observation in Figures 3e to 3h is the decreased chlorine generation at higher convective mass transfer once NH4+ - N is oxidized. In the absence of NH4+ - N, there is also a tendency towards a decrease in the total Cl2 generation at elevated rotation speeds. The largest electrogenerated chlorine concentrations were detected without convective mass transfer - a trend that appears counterintuitive. The reaction rates for total chlorine generation in the absence of NH4+ - N decreased from 6.6 ± 0.3 mg Cl2 L-1 min-1 at 0 RPM to 2.9 ± 0.2 mg Cl2 L-1 min-1 at 1300 RPM (Table S2). Typically, the electrocatalytic reaction rate is limited by electroactive species mass transfer (Cl‒ in this case), given that ionic adsorption/desorption and electron transfer occur rapidly. According to the Nernst-Plank equation, the total flux of the species consists of diffusion, migration, and advection fluxes (Bard and Faulkner, 2000). Thus, a lower convective flux would generally correspond to a reduced total mass flux to the electrode surface, leading to slower electrochemical reaction rate and consequently decreased chlorine generation. However, our results show the opposite trend: decrease in the convective mass transfer enhances Cl2 generation. Similar results to ours were reported in earlier studies showing increased Cl2 generation at lower RPM when using a DSA anode (Polcaro et al., 2008). This observation was attributed to the enhanced Cl2 reduction at the cathode as parasitic reaction at higher RPM, although no experiments were conducted to support the conclusion. 19 20 Fig. 3 Species concentrations in the bulk as a function of time during breakpoint electrochlorination at different RPM. (a-d) Total ammonia (NH4+/NH3 and NH2Cl/NHCl2/NCl3) concentration and pH changes as a function of the time, and (e-h) variation of total chlorine (Cl2/HOCl/OCl‒ and NH2Cl/NHCl2/NCl3), free chlorine (Cl2/HOCl/OCl‒), and monochloramine (NH2Cl) concentrations during breakpoint electrochlorination for a electrogenerated total dose of chlorine. All experiments were performed in an undivided open batch reactor operated at 0.12 A and containing 84 mM NaCl in the absence or presence of 10 mg NH4+ - N L-1 and under varying stirring rates. Stirring speed: (a, e) 0 RPM, (b, f) 100 RPM, (c, g) 500 RPM, and (d, h) 1300 RPM. Total chlorine and pH without NH4+ - N L-1 present (empty circles), total ammonia (filled squares), total chlorine and pH with NH4+ - N L-1 present (filled circles), free chlorine with NH4+ - N L-1 present (filled rhombi), and monochloramine (empty triangles). To investigate the impact of potential chlorine reduction at the cathode at various RPM, we conducted experiments without Cl‒ ions, but with initially present free chlorine, using 84 mM NaOCl4 as the electrolyte due to similar ionic conductivity to NaCl (Bard and Faulkner, 2000) and adding NaOCl as free chlorine source. Figure 4 shows changes in 200 mg Cl2 L-1 concentration over electrolysis time at various RPM and chemical chlorination at 1300 RPM (blank). As can be observed from Figure 4, there was a mild concentration change over time. The largest convective mass transfer conditions of 1300 RPM resulted in chlorine concentration decline over time, while there was no chlorine decrease for the other experiments (Figure 4). The decrease in Cl2 concentration at 1300 RPM during electrolysis may result from i) Cl2 reduction at the cathode due to H2 bubbles not blocking the cathode surface at 1300 RPM compared to other stirring conditions where noticeable cathode surface hindering by bubbles present (Figure S2); and/or ii) Cl2 gas stripping from solution by electrochemically generated H2 or O2 bubbles at high speeds. Importantly, without applied electricity, Cl2 concentration remained constant at 1300 RPM (Figure 4, blank), eliminating any significant Cl2 leak into the gas-phase due to the convective conditions in solution. Comparison of the instant rates of CER (Table S2) and chlorine reduction reaction (CRR) (Table S3), except rotating speed of 1300 RPM, shows that Cl2 production increases at lower RPM. This trend cannot be 21 attributed to the cathodic reduction of Cl2 or Cl2 stripping by electrogenerated bubbles in the experimental setup due to their small impact over electrolysis time. Our results of the breakpoint electrochlorination across various RPM show that NH4+ - N oxidation is favored under slow mass transfer conditions. This outcome could be due to enhanced CER at lower convective mass transfer. Consequently, this creates a need to investigate the impact of convection exclusively on the CER in a two separated compartments cell, free from any cathode influence. Fig. 4 Electrochemical chlorine reduction in the bulk as a function of time under various RPM. Electrolysis of 200 mg Cl2 L-1 using NaOCl in 84 mM NaClO4 electrolyte at 0 RPM, 100 RPM, 500 RPM, and 1300 RPM stirring rates in undivided open electrochemical cell at current of 0.12 A. The blank experiment corresponds to stirring at 1300 RPM without applying current. 3.3 Understanding chlorine electrogeneration as a function of convective mass transfer Due to increased Cl2 generation under slower mass transfer conditions observed during electrolysis in an undivided cell, we conducted experiments in a divided cell to assess the impact of mass transfer solely 22 on anodic reactions. Electrochlorination in the anodic compartment resulted in pH drop from 5.74 ± 0.07 to 2.57 ± 0.03 in 15 minutes (Figure S3). The pH shift towards the acid range is attributed to the concomitant OER (eq 12) taking place at the anode and chlorine disproportionation reaction (eq 2) that release H+. As pH in the anodic compartment decreased below pKa value for HOCl/OCl− and approached the pH of equimolar concentrations of the equilibria of Cl2(aq)/HOCl species defined by their apparent pK, there is a need to quantify chlorine released to the gas phase at various RPM to assess any impact of chlorine volatilization. Figure 5 shows that the apparent pK of Cl2(aq)/HOCl species equilibria depends besides pH on Cl− concentration. Changes in the Cl− content from 0.01 M to 0.15 M, and 1.0 M result in a shift of the pH of equimolar Cl2(aq)/HOCl species distribution from 1.22 to 2.4, and 3.22, respectively. Although chlorine gas is soluble in water, the concentration of Cl2(g) at equilibrium is more than one order of magnitude larger than the concentration of Cl2(aq) based on Henry’s constant (eq 1), meaning possible significant losses of Cl2 into the gas phase, with chloride level of 0.15 M. Considering possible losses due to chlorine evolution into gas phase, all further experiments were conducted in a sealed system with a chlorine gas trap to quantify complete mass balances. 23 Fig. 5 Relative distribution of chlorine species in liquid phase as a function of pH at various Cl‒ concentrations. For a) 0.01 M; b) 0.15 M; and c) 1.0 M of Cl‒ content. Chlorine speciation was plotted using formulas (Cherney et al., 2006). Hydrolysis constant of 6 x 10-4 M2 was used (Eigen and Kustin, 1962). Figure 6a shows the electrogenerated free chlorine over time measured in the aqueous phase in the anodic compartment of the closed divided cell. A decrease in convective mass transfer led to an increase in Cl2 generation, in agreement with the experimental results observed in the open undivided cell. The free 24 chlorine generation rate increased from 4.6 ± 0.1 mg Cl2 L-1 min-1 at 1300 RPM to 6.61 ± 0.07 mg Cl2 L-1 min-1 at 0 RPM (Table S4). The difference of free chlorine generation is statistically significant, with pvalues ≤ 0.02. These results further demonstrate that reduction of generated free chlorine at the cathode is not the main driver of such difference on chlorine accumulation. To evaluate the loss of free chlorine to the gas phase under established acidic pH conditions (Figure S3), the release of Cl2 gas was measured for all rotation speeds. It can be seen that aqueous chlorine content and evolved chlorine decreases with increasing rotation speed (Table 5S). However, the average mass fraction of chlorine detected in the gas phase was 24% across all the stirring conditions, while the aqueous phase fraction of free chlorine remained ca. 76%. These results indicate that the relative mass of chlorine evolved into the gas phase was independent of the rotation speed, as no significant differences in mass fractions were observed across various RPM. The relative mass of the chlorine detected in the gas phase was solely dependent on the total mass of free chlorine produced. Thus, these findings suggest that higher free chlorine generation under lower convective mass transfer is not due to increased chlorine stripping at higher RPM, as gas and liquid phase mass fractions remained consistent. In order to assess the impact of applied current on the concentration of pre-existing free chlorine in solution, blank experiment electrolysis was conducted in divided closed cell containing 50 mg Cl2 L-1 added using NaOCl in 150 mM NaClO4 as inert electrolyte. Without Cl‒ concentration present in NaClO4 electrolyte, only OER occurs at the anode, drastically decreasing pH in the anodic compartment (Table S6), which is below measured pH during co-existing CER (Figure S3). Nevertheless, Figure 6b illustrates that free chlorine in solution is not impacted by oxidation at the anode, with the concentration remaining constant over electrolysis time independently of the solution stirring velocity. Therefore, the lower accumulation of free chlorine observed at higher rotating speeds is not driven by the undesired oxidation of HClO/OCl‒ at the anode to other species (e.g., ClO2, ClO3‒, ClO4‒). In addition, no Cl2 was detected in the gas phase during these experiments despite acidic pH. These findings are consistent with earlier studies reporting HOCl as the dominant species even at low pH in the absence of Cl‒, showing that Cl‒ is needed for the reverse reaction to occur (eq 2) (Cherney et al., 2006). 25 Figure 6c illustrates Faradaic efficiencies for free chlorine and chlorine gas evolved as detected in both aqueous and gas phases. The Faradaic efficiencies for free chlorine detected in the liquid phase decreased with increasing convective mass transfer, dropping from 48.0 ± 1.0 % at 0 RPM to 44.4 ± 0.5 %, 38 ± 1.0 %, and 32.9 ± 0.9 % at 100, 500, and 1300 RPM, respectively. A t-test confirmed that these differences are statistically significant, with p-value < 0.05. When considering electrogenerated free chlorine in both liquid and gaseous phases, the Faradaic efficiency similarly decreased from 64.0 ± 2.0 % at 0 RPM to 58.2 ± 0.8 %, 51.0 ± 2.0%, and 44.0 ± 3.0 % at 100, 500, and 1300 RPM, respectively. Statistical analysis confirmed significant differences yielding p-value < 0.05, except for rotations between 0 RPM and 100 RPM, where a p-value of 0.061 was observed, however still near the 0.05 threshold. 32 References Akkari, S., Sánchez-Sánchez, C.M., Hopsort, G., Serrano, K.G., Loubière, K., Tzedakis, T., Benyahia, R., Rebiai, L., Bastide, S., Cachet-Vivier, C., Vivier, V., Lopez-Viveros, M., Azimi, S., 2025. Progress on electrochemical and photoelectrochemical urea and ammonia conversion from urine for sustainable wastewater treatment. Appl Catal B 362. https://doi.org/10.1016/j.apcatb.2024.124718 Atrashkevich, A., Alum, A., Stirling, R., Abbaszadegan, M., Garcia-Segura, S., 2024. Approaching easy water disinfection for all: Can in situ electrochlorination outperform conventional chlorination under realistic conditions? Water Res 250. https://doi.org/10.1016/j.watres.2023.121014 Atrashkevich, A., Garcia-Segura, S., 2024. Engineering decentralized electrodisinfection to sustain consistent chlorine generation under varying drinking water chloride content. Applied Catalysis O: Open 195, 207012. https://doi.org/10.1016/j.apcato.2024.207012 Baba, M. El, Kayastha, P., Huysmans, M., Smedt, F. De, 2020. Evaluation of the groundwater quality using the water quality index and geostatistical analysis in the Dier al-Balah Governorate, Gaza Strip, Palestine. Water (Switzerland) 12. https://doi.org/10.3390/w12010262 Baird, R.B., Eaton, A.D., Rice, E.W., 2017. Standard Methods for the Examination of Water and Wastewater, 23rd ed. American Public Health Association, American Water Works Association, Water Environment Federation, Washington. Bard, A.J., Faulkner, L.R., 2000. Electrochemical methods: fundamentals and applications, Second Edition. ed. John Wiley & Sons, Inc. Bhandari, A., Bond, R., Briggeman, E.R., Burns, N.L., Cooke, T.H., Harms, L.L., Hoehn, R.C., Holakoo, L., Holmes, T.A., Hunter, G.L., Jiang, H., Kobylinski, E., Lewis, K.A., Massart, N.S., Neemann, J., O’Brien, W., Randtke, S., Shorney-Darby, H., Townsend, B.R., Vogt, E.D., 2010. White’s Handbook of Chlorination and Alternative Disinfection, Fifth edition. ed. A John Wiley & Sons, Inc. Cherney, D.P., Duirk, S.E., Tarr, J.C., Collette, T.W., 2006. Monitoring the Speciation of Aqueous Free Chlorine from pH 1 to 12 with Raman Spectroscopy to Determine the Identity of the Potent Low-pH Oxidant. Appl Spectrosc 60, 764–772. 33 Cimetiere, N., 2009. Etude de la décomposition de la monochloramine en milieu aqueux et réactivité avec des composés phénoliques. Crittenden, J.C., Trussell, R.Rhodes., Hand, D.W., Howe, K.J., Tchobanoglou, George., 2012. MWH’s Water Treatment Principles and Design, Third Edition. ed. John Wiley & Sons, Inc., New Jersey. Deborde, M., von Gunten, U., 2008. Reactions of chlorine with inorganic and organic compounds during water treatment-Kinetics and mechanisms: A critical review. Water Res. https://doi.org/10.1016/j.watres.2007.07.025 Eigen, M., Kustin, K., 1962. The Kinetics of Halogen Hydrolysis. J Am Chem Soc 84, 1355–1361. https://doi.org/10.1021/ja00867a005 Felisardo, R.J.A., Brillas, E., Boyer, T.H., Cavalcanti, E.B., Garcia-Segura, S., 2024. Understanding electrochemical treatment of real fresh and hydrolyzed urine matrices to remove trace pharmaceuticals. Sep Purif Technol 342. https://doi.org/10.1016/j.seppur.2024.127016 Feng, W., McCarthy, D.T., Henry, R., Zhang, X., Zhang, K., Deletic, A., 2018. Electrochemical oxidation for stormwater disinfection: How does real stormwater chemistry impact on pathogen removal and disinfection by-products level? Chemosphere 213, 226–234. https://doi.org/10.1016/j.chemosphere.2018.09.038 Fuladpanjeh-Hojaghan, B., Elsutohy, M.M., Kabanov, V., Heyne, B., Trifkovic, M., Roberts, E.P.L., 2019. In-Operando Mapping of pH Distribution in Electrochemical Processes. Angewandte Chemie - International Edition 58, 16815–16819. https://doi.org/10.1002/anie.201909238 Ghaitidak, D.M., Yadav, K.D., 2013. Characteristics and treatment of greywater-a review. Environmental Science and Pollution Research 20, 2795–2809. https://doi.org/10.1007/s11356-013-1533-0 Jafvert, C.T., Valentine, R.L., 1992. Industrial Membrane Processes. Environ Sci Technol 26, 577–586. Kapałka, A., Katsaounis, A., Michels, N.L., Leonidova, A., Souentie, S., Comninellis, C., Udert, K.M., 2010. Ammonia oxidation to nitrogen mediated by electrogenerated active chlorine on Ti/PtOx-IrO2. Electrochem commun 12, 1203–1205. https://doi.org/10.1016/j.elecom.2010.06.019 34 Karlsson, R.K.B., Cornell, A., 2016. Selectivity between Oxygen and Chlorine Evolution in the ChlorAlkali and Chlorate Processes. Chem Rev 116, 2982–3028. https://doi.org/10.1021/acs.chemrev.5b00389 Lee, W., Westerhoff, P., Yang, X., Shang, C., 2007. Comparison of colorimetric and membrane introduction mass spectrometry techniques for chloramine analysis. Water Res 41, 3097–3102. https://doi.org/10.1016/j.watres.2007.04.032 Obata, K., Van De Krol, R., Schwarze, M., Schomäcker, R., Abdi, F.F., 2020. In situ observation of pH change during water splitting in neutral pH conditions: Impact of natural convection driven by buoyancy effects. Energy Environ Sci 13, 5104–5116. https://doi.org/10.1039/d0ee01760d Olichwer, T., Tarka, R., Modelska, M., 2013. Chemical composition of groundwaters in the Hornsund region, southern Spitsbergen. Hydrology Research 44, 117–130. https://doi.org/10.2166/nh.2012.075 Otter, P., Hertel, S., Ansari, J., Lara, E., Cano, R., Arias, C., Gregersen, P., Grischek, T., Benz, F., Goldmaier, A., Alvarez, J.A., 2020. Disinfection for decentralized wastewater reuse in rural areas through wetlands and solar driven onsite chlorination. Science of the Total Environment 721. https://doi.org/10.1016/j.scitotenv.2020.137595 Otter, P., Malakar, P., Sandhu, C., Grischek, T., Sharma, S.K., Kimothi, P.C., Nüske, G., Wagner, M., Goldmaier, A., Benz, F., 2019. Combination of river bank filtration and solar-driven electrochlorination assuring safe drinking water supply for river bound communities in India. Water 11, 1– 17. https://doi.org/10.3390/w11010122 Polcaro, A.M., Vacca, A., Mascia, M., Ferrara, F., 2008. Product and by-product formation in electrolysis of dilute chloride solutions. J Appl Electrochem 38, 979–984. https://doi.org/10.1007/s10800-0089509-3 Qiang, Z., Adams, C.D., 2004. Determination of Monochloramine Formation Rate Constants with StoppedFlow Spectrophotometry. Environ Sci Technol 38, 1435–1444. https://doi.org/10.1021/es0347484 35 Romano, A., Ortiz, I., Urtiaga, A.M., 2021. Comprehensive kinetics of electrochemically assisted ammonia removal in marine aquaculture recirculating systems. Journal of Electroanalytical Chemistry 897. https://doi.org/10.1016/j.jelechem.2021.115619 Sánchez-Sánchez, C.M., Rodríguez-López, J., Bard, A.J., 2008. Scanning electrochemical microscopy. 60. Quantitative calibration of the SECM substrate generation/tip collection mode and its use for the study of the oxygen reduction mechanism. Anal Chem 80, 3254–3260. https://doi.org/10.1021/ac702453n Saunier, B.M., Selleck, R.E., 1979. The Kinetics of Breakpoint Chlorination in Continuous Flow Systems. J Am Water Works Assoc 71, 164–172. Scialdone, O., Proietto, F., Galia, A., 2021. Electrochemical production and use of chlorinated oxidants for the treatment of wastewater contaminated by organic pollutants and disinfection. Curr Opin Electrochem 27. https://doi.org/10.1016/j.coelec.2020.100682 Snoeyink, V.L., Jenkins, D., 1980. Water Chemistery. John Wiley & Sons, Inc. Trasatti, S., 1987. Progress in the understanding of the mechanism of chlorine evolution at oxide electrodes. Electrochim Acta 32, 369–382. Vanýsek, P., 1978. CRC Handbook of Chemistry and Physics, 91th Edition, J. Phys. Chem. Ref. Data. Marcel Dekker. Wagman, D.D., Evans, W.H., Parker, V.B., Schumm, R.H., Halow, I., Bailey, S.M., Churney, K.L., Nuttall, R.L., 1982. The NBS Tables of Chemical Thermodynamic Properties. J Phys Chem Ref Data 11. World Health Organization, 2022. Guidelines for drinking-water quality, 4th ed. WHO. 36 Supplementary information Breakpoint electrochlorination in ammonia removal: Unveiling the impact of convective mass transfer Aksana Atrashkevicha, Dominic Vardaa, Kaitlyn Yeagera, Maria Gomez-Mingota,b, Carlos M. SánchezSáncheza,c, Sergi Garcia-Seguraa,* aNanosystems Engineering Research Center for Nanotechnology-Enabled Water Treatment, School of Sustainable Engineering and the Built Environment, Arizona State University, Tempe, AZ bLaboratoire de Chimie des Processus Biologiques, Collège de France, UMR 8229 CNRS, Sorbonne Université, PSL Research University, 11 Place Marcelin Berthelot, 75005 Paris, France cSorbonne Université, CNRS, Laboratoire Interfaces et Systèmes Electrochimiques (LISE), 4 Place Jussieu, 75005 Paris, France Article submitted to be published in Water Research https://doi.org/10.1016/j.watres.2025.123732 Corresponding author: *e-mail: [email protected] (Dr. Sergi Garcia-Segura) © 2025. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/ 37 Table S1 Reaction rates of 10 mg NH4+ - N L-1 oxidation at various RPM in undivided open cell using 84 mM NaCl electrolyte Mean Standard Deviation mg NH4+ - N L-1 min-1 0 RPM 0.67 0.01 100 RPM 0.51 0.01 500 RPM 0.326 0.008 1300 RPM 0.161 0.006 Table S2 Reaction rates of chlorine evolution at various RPM without NH4+ - N present in undivided open cell using 84 mM NaCl electrolyte Mean Standard Deviation mg Cl2 L-1 min-1 0 RPM 6.6 0.3 100 RPM 5.9 0.3 500 RPM 3.6 0.2 1300 RPM 2.9 0.2 Table S3 Reaction rates of chlorine reduction at various RPM without NH4+ - N present in undivided open cell using 84 mM NaClO4 electrolyte spiked with 200 mg Cl2 L-1 using NaOCl Mean Standard Deviation mg Cl2 L-1 min-1 Blank* No reduction 0 RPM No reduction 100 RPM No reduction 500 RPM No reduction 1300 RPM 0.6 0.1 * - without electricity at 1300 RPM 38 Table S4 Reaction rates of chlorine evolution detected in the aqueous phase in anodic compartment at various RPM in divided closed cell using 150 mM NaCl electrolyte Mean Standard Deviation mg Cl2 L-1 min-1 0 RPM 6.61 0.07 100 RPM 6.15 0.04 500 RPM 5.20 0.05 1300 RPM 4.6 0.1 Table S5 Chlorine mass and fractions detected in liquidous and gaseous phases in anodic compartment at various RPM in divided closed cell using 150 mM NaCl electrolyte in 15 minutes Mean Standard Deviation 0 RPM Mass liquid (mg) 19.2 0.4 Mass gas (mg) 6.2 0.9 Fraction liquid 0.76 0.03 Fraction gas 0.24 0.03 100 RPM Mass liquid (mg) 17.6 0.2 Mass gas (mg) 5.5 0.5 Fraction liquid 0.76 0.02 Fraction gas 0.24 0.02 500 RPM Mass liquid (mg) 15.2 0.4 Mass gas (mg) 5.0 1.0 Fraction liquid 0.76 0.04 Fraction gas 0.24 0.04 1300 RPM Mass liquid (mg) 13.0 0.3 Mass gas (mg) 4.3 1.5 Fraction liquid 0.76 0.07 Fraction gas 0.24 0.07 39 Table S6 Final pH in 15 minutes during chlorine reduction in anodic compartment at various RPM in divided closed cell using 150 mM NaClO4 spiked with 50 mg Cl2 L-1 using NaOCl Mean Standard Deviation 0 RPM 2.35 0.01 100 RPM 2.36 0.03 500 RPM 2.35 0.01 1300 RPM 2.4 0 Figure S1. Schematics of a) undivided open cell, and b) divided closed cell. 40 Figure S2. Cyclic voltammetry of Pt with geometrical area of 6 cm2 (2.5 cm x 2.4 cm) at a scan rate of 20 mV s-1 using 1 mM FcMeOH in 0.1 M KCl electrolyte under various RPM. 41 Figure S3. Chronoamperometry of Pt with geometrical area of 6 cm2 (2.5 cm x 2.4 cm) applying 0.5 V vs. Ag/AgCl (i.e., 0.7 V vs. SHE) under various RPM using 0.1 M KCl electrolyte and a) 0.5 mM FcMeOH, b) 1.0 mM FcMeOH, and c) 2.0 mM FcMeOH. d) Average limiting current value of the last 60 s out of 120 s of chronoamperometry experiments as function of stirrer RPM at various FcMeOH concentrations.