Electro-Oxidation of Humic Acids Using Platinum Electrodes: An Experimental Approach and Kinetic Modelling

Humic acids (HA) are a potential hazard to aquatic ecosystems and human health. Because biological treatment of contaminated water does not satisfactorily remove these pollutants, novel approaches are under evaluation. This work explores electrochemical oxidation of HA in aqueous solution in a lab-scale apparatus using platinum-coated titanium electrodes. We evaluated the effects of HA concentration, current density, chloride concentration and ionic strength on the rate of HA oxidation. The initial reaction rate method was used for determining the rate law of HA degradation. The results showed that the reaction rate was first-order relative to HA concentration, chloride concentration and current density. An appreciable effect of ionic strength was also observed, most likely due to the polyanionic character of HA. We propose a kinetic model that satisfactorily fits the experimental data.


Introduction
The removal of emerging and refractory pollutants from wastewater is a priority research area worldwide [1,2]. Humic compounds are organic macromolecules deriving in nature from biological decomposition of organic matter [3]. They consist of aromatic and aliphatic blocks with carboxylic, phenolic and alkoxy functional groups [4,5]. Humic compounds are generally found in landfill leachate, as the organic fraction of mature leachate is mainly refractory humic substances [6][7][8], domestic sewage [9] and coking wastewater [10]. These are recalcitrant towards conventional biological treatment [11], thereby requiring additional treatment with other methods. Humic acids (HA) are the water-soluble fraction of humic compounds at neutral and basic pH [12]. HA impart undesirable color, taste, and odor to water [13] and are potentially harmful to aquatic ecosystems and human health [14][15][16].

Materials
HA were extracted from a compost sample using an acid/basic procedure [62,63]. The compost was purchased from Gesal S.r.l., Naples (Italy). According to the manufacturer's data, the sample contained 30% organic carbon and 1.7% organic nitrogen. All other chemicals employed in this study were from Sigma-Aldrich and were of analytical grade.

Humic Acids Extraction
One hundred grams of compost were treated with 1.0 L of distilled water and, under stirring, the pH was brought to 12.0 by dissolving solid NaOH. The mixture was stirred for two days and the pH periodically checked, adding a few drops of concentrated NaOH solution if necessary. The mixture was centrifuged at 10,000 rpm for 30 min, paper-filtered, and the filtrate recovered. After that, the pH was brought to 1.5 by adding concentrated HCl. After one night in the fridge, the precipitate was collected by centrifugation at 10,000 rpm, washed with distilled water and oven-dried at 45 • C. Eight grams of dry sample (henceforth denoted as HA) were obtained from 100 g compost, the yield of acid/basic extraction procedure thus being 8%.
For HA purification, a solution of raw HA at neutral pH was dialyzed against 0.1 M NaHCO 3 solution, using a molecular porous membrane tubing with a cut-off of 6-8 kDa (Spectra/Por), until UV-Vis absorbance of the external solution reached a stable value (approximately after one week).

Electro-Oxidation Experiments
Electro-oxidation experiments were carried out at lab scale at about 25 • C. A batch reactor of 0.5 L was fed with 250 mL of 70-150 mg L −1 HA solution (C HA ). The pH of the samples was recorded using a HI 9017 Hanna Instruments pH meter, calibrated in pH 1, 7 and 10 buffer solutions. The initial pH was brought to 10.0 by adding few drops of dilute NaOH solution. During the EO experiments, the pH was continuously monitored showing only a slight decrease, from 10.0 up to a minimum of about 9.0. These experimental conditions prevented the formation of HA aggregates [64,65].
The effect of reactive (Cl − ) and/or inert (ClO 4 − or NO 3 − ) anions was evaluated by adding 0.01-0.05 mol L −1 NaCl, NaClO 4 or NaNO 3 to the HA solution, which caused an increase of the ionic strength (IS) in the range 10-50 mM. The electrochemical reactor ( Figure 1) consisted of two cylindrical electrodes of platinum-coated titanium 0.80 cm in diameter and 10 cm in length, placed at a distance of 3 cm. The electrodes were connected to a Direct Current (DC) power supply BPS-305 (Lavolta, London, UK Electro-oxidation experiments were carried out at lab scale at about 25 °C. A batch reactor of 0.5 L was fed with 250 mL of 70-150 mg L −1 HA solution (CHA). The pH of the samples was recorded using a HI 9017 Hanna Instruments pH meter, calibrated in pH 1, 7 and 10 buffer solutions. The initial pH was brought to 10.0 by adding few drops of dilute NaOH solution. During the EO experiments, the pH was continuously monitored showing only a slight decrease, from 10.0 up to a minimum of about 9.0. These experimental conditions prevented the formation of HA aggregates [64,65].
The effect of reactive (Cl − ) and/or inert (ClO4 − or NO3 − ) anions was evaluated by adding 0.01-0.05 mol L −1 NaCl, NaClO4 or NaNO3 to the HA solution, which caused an increase of the ionic strength (IS) in the range 10-50 mM. The electrochemical reactor ( Figure 1) consisted of two cylindrical electrodes of platinum-coated titanium 0.80 cm in diameter and 10 cm in length, placed at a distance of 3 cm. The electrodes were connected to a Direct Current (DC) power supply BPS-305 (Lavolta, London, UK). The EO tests were carried out under amperostatic conditions at values of current density (I) in the range 77-230 A m −2 .

UV-Vis Analysis
The starting HA solutions and the aliquots collected at programmed time-intervals during the electro-removal treatment were analyzed by UV-Vis spectroscopy. The UV-Vis spectra were recorded in the range of 200-800 nm on a Lambda 40, spectrometer, optical path = 1.00 cm (Perkin Elmer, Waltham, MA, USA). Quantitative analysis was performed using an absorbanceconcentration calibration curve at 320 nm (molar extinction coefficient = 0.01412 L cm −1 mg −1 ). After each test, the electrodes were washed with a solution containing Na4P2O7 and NaOH, and the rinsing solution was analyzed by UV-Vis spectroscopy to verify the occurrence of HA adsorbed on the electrode surface. In no case was HA detected in measurable amount.

Low-Pressure Size-Exclusion Chromatography (LP-SEC)
In selected experiments, low-pressure size-exclusion chromatography was performed in order to obtain information on the molecular weight distribution of HA and of reaction products, using a Bio-Gel P10 Gel medium with 1.5-20 kDa fractional range (Bio-Rad Laboratories, Hercules, CA, USA), in a column 30 cm in length and 2.5 cm in diameter, eluted with 0.1 mol L −1 NaHCO3 at a flow rate of 2.0 mL h −1 . Vanillic acid was used in LP-SEC experiments as a reference for low molecular weight HA-like compounds.

Total Organic Carbon
The total organic carbon (TOC) was determined at the end of the EO runs using a TOC-L CSN carbon analyzer (Shimadzu, Kyoto, Japan). Samples were heated to 680 °C in an oxygen-rich

UV-Vis Analysis
The starting HA solutions and the aliquots collected at programmed time-intervals during the electro-removal treatment were analyzed by UV-Vis spectroscopy. The UV-Vis spectra were recorded in the range of 200-800 nm on a Lambda 40, spectrometer, optical path = 1.00 cm (Perkin Elmer, Waltham, MA, USA). Quantitative analysis was performed using an absorbance-concentration calibration curve at 320 nm (molar extinction coefficient = 0.01412 L cm −1 mg −1 ). After each test, the electrodes were washed with a solution containing Na 4 P 2 O 7 and NaOH, and the rinsing solution was analyzed by UV-Vis spectroscopy to verify the occurrence of HA adsorbed on the electrode surface. In no case was HA detected in measurable amount.

Low-Pressure Size-Exclusion Chromatography (LP-SEC)
In selected experiments, low-pressure size-exclusion chromatography was performed in order to obtain information on the molecular weight distribution of HA and of reaction products, using a Bio-Gel P10 Gel medium with 1.5-20 kDa fractional range (Bio-Rad Laboratories, Hercules, CA, USA), in a column 30 cm in length and 2.5 cm in diameter, eluted with 0.1 mol L −1 NaHCO 3 at a flow rate of 2.0 mL h −1 . Vanillic acid was used in LP-SEC experiments as a reference for low molecular weight HA-like compounds.

Total Organic Carbon
The total organic carbon (TOC) was determined at the end of the EO runs using a TOC-L CSN carbon analyzer (Shimadzu, Kyoto, Japan). Samples were heated to 680 • C in an oxygen-rich environment in combustion tubes filled with a platinum catalyst. The carbon dioxide generated by oxidation was measured with an infrared gas analyzer.

Elemental Analysis
The C/H ratios in the starting HA solution and two HA solutions obtained by the addition of 0.01 mol L −1 NaCl or NaClO 4 were determined by elemental analysis before and after the electro-removal runs. To this purpose, the solutions were first subjected to LP-SEC. The eluted fractions were recovered, and the pH was brought to about 3 with a dilute HCl solution. This procedure was performed twice. Finally, the samples were dialyzed against water to eliminate the residues of NaHCO 3 (eluent phase of the LP-SEC), lyophilized and analyzed using a CHN628 elemental analyzer (Leco, St. Joseph, Michigan, USA).

Electro-Oxidation of HA
The EO of HA was investigated under several different operational conditions: about one hundred kinetic runs were carried out to evaluate the effect of changes in HA and salt concentrations, type of electrolyte and current density.
UV spectroscopy and TOC measurements demonstrated a marked reduction of HA concentration in all experiments. During EO experiments, the release of bubbles from the electrodes was observed. According to the literature [66], this was probably due to the electrolysis of water leading to the formation of O 2 (anode) and H 2 (cathode) and causing overpotential phenomena. Figure 2 shows, as an example, the time dependence of HA concentration in two kinetic runs performed in the presence of 10 mM NaCl and 10 mM NaClO 4 , respectively. The degradation reaction was faster for the sample containing NaCl. As discussed in detail in Section 3.2, the different degradation rates of HA in the presence of Cl − (reactive) and ClO 4 − (inert) can be ascribed to the different chemical behavior of these anions under current flow.
Water 2020, 12, 2250 4 of 18 environment in combustion tubes filled with a platinum catalyst. The carbon dioxide generated by oxidation was measured with an infrared gas analyzer.

Elemental Analysis
The C/H ratios in the starting HA solution and two HA solutions obtained by the addition of 0.01 mol L −1 NaCl or NaClO4 were determined by elemental analysis before and after the electroremoval runs. To this purpose, the solutions were first subjected to LP-SEC. The eluted fractions were recovered, and the pH was brought to about 3 with a dilute HCl solution. This procedure was performed twice. Finally, the samples were dialyzed against water to eliminate the residues of NaHCO3 (eluent phase of the LP-SEC), lyophilized and analyzed using a CHN628 elemental analyzer (Leco, St. Joseph, Michigan, USA).

Electro-Oxidation of HA
The EO of HA was investigated under several different operational conditions: about one hundred kinetic runs were carried out to evaluate the effect of changes in HA and salt concentrations, type of electrolyte and current density.
UV spectroscopy and TOC measurements demonstrated a marked reduction of HA concentration in all experiments. During EO experiments, the release of bubbles from the electrodes was observed. According to the literature [66], this was probably due to the electrolysis of water leading to the formation of O2 (anode) and H2 (cathode) and causing overpotential phenomena. Figure 2 shows, as an example, the time dependence of HA concentration in two kinetic runs performed in the presence of 10 mM NaCl and 10 mM NaClO4, respectively. The degradation reaction was faster for the sample containing NaCl. As discussed in detail in Section 3.2, the different degradation rates of HA in the presence of Cl − (reactive) and ClO4 − (inert) can be ascribed to the different chemical behavior of these anions under current flow. To explore the effect of each investigated parameter on the rate of HA oxidation, the initial reaction rates (v0) of the experimental kinetic curves were estimated and compared. The values of v0 (accessible in Supplementary Material) were obtained from the slope of the tangent to the To explore the effect of each investigated parameter on the rate of HA oxidation, the initial reaction rates (v 0 ) of the experimental kinetic curves were estimated and compared. The values of v 0 (accessible in Supplementary Material) were obtained from the slope of the tangent to the concentration-time curve at t = 0. Figure 3 reports the full set of estimated v 0 values as a function of NaCl concentration at four current densities.
Water 2020, 12, 2250 5 of 18 concentration-time curve at t = 0. Figure 3 reports the full set of estimated v0 values as a function of NaCl concentration at four current densities.  As can be seen, the degradation rate was simultaneously affected by HA concentration, Cl − concentration, current density and ionic strength.
To clarify the role of HA concentration in the kinetics of the process, we investigated the dependence of v0 on HA concentration for different current densities, while keeping the salt concentration constant. The results from experiments carried out in the presence of 10 mM NaClO4 As can be seen, the degradation rate was simultaneously affected by HA concentration, Cl − concentration, current density and ionic strength.
To clarify the role of HA concentration in the kinetics of the process, we investigated the dependence of v 0 on HA concentration for different current densities, while keeping the salt concentration constant. The results from experiments carried out in the presence of 10 mM NaClO 4 and 10 mM NaCl are illustrated in Figure 4A,B, respectively. The curves were obtained by fitting the data to linear equations passing through the origin.
Water 2020, 12, 2250 7 of 18 and 10 mM NaCl are illustrated in Figure 4A,B, respectively. The curves were obtained by fitting the data to linear equations passing through the origin. The reaction rate linearly increased with increasing HA concentration, within experimental error, thus suggesting that the reaction obeys the first-order kinetics relative to HA. As clearly shown The reaction rate linearly increased with increasing HA concentration, within experimental error, thus suggesting that the reaction obeys the first-order kinetics relative to HA. As clearly shown in Figure 4A,B, the slope of the lines proportionally increased with increasing current density. This suggests that the fraction of current loss due to undesirable side-reactions, such as the evolution of O 2 at the platinum anode [67], did not appreciably vary over the investigated range of current density.
The effect of ionic strength on the kinetics of HA EO was studied using NaClO 4 as an inert supporting electrolyte. Interestingly, an increase of the HA removal rate with increasing NaClO 4 concentration for all the investigated experimental conditions was observed (see Figure 5A). It is worth noting that the removal of HA was not strictly proportional to the salt concentration: the initial rate of HA oxidation increased less rapidly with respect to the increase of NaClO 4 (concave downward trend), as can be inferred from the slope of the lines in Figure 5A and from the data points in Figure 5B, which do not lie along a straight line passing from the origin. Experiments with another inert electrolyte, NaNO 3 , gave similar results (see Table S1 in the Supplementary Materials). A similar, although less marked, effect of ionic strength on the oxidation rate of HA was also observed in the presence of chloride ions. Figure 5C shows, as an example, that the rate of HA EO in the presence of 10 mM Cl − slightly increased with an increase in NaClO 4 concentration from 0 to 40 mM.
in Figure 4A,B, the slope of the lines proportionally increased with increasing current density. This suggests that the fraction of current loss due to undesirable side-reactions, such as the evolution of O2 at the platinum anode [67], did not appreciably vary over the investigated range of current density.
The effect of ionic strength on the kinetics of HA EO was studied using NaClO4 as an inert supporting electrolyte. Interestingly, an increase of the HA removal rate with increasing NaClO4 concentration for all the investigated experimental conditions was observed (see Figure 5A). It is worth noting that the removal of HA was not strictly proportional to the salt concentration: the initial rate of HA oxidation increased less rapidly with respect to the increase of NaClO4 (concave downward trend), as can be inferred from the slope of the lines in Figure 5A and from the data points in Figure 5B, which do not lie along a straight line passing from the origin. Experiments with another inert electrolyte, NaNO3, gave similar results (see Table S1 in the Supplementary Materials). A similar, although less marked, effect of ionic strength on the oxidation rate of HA was also observed in the presence of chloride ions. Figure 5C shows, as an example, that the rate of HA EO in the presence of 10 mM Cl − slightly increased with an increase in NaClO4 concentration from 0 to 40 mM.
The results presented above indicate that ionic strength plays an active role in HA EO. In line with transition state theory [68], a possible explanation for this behavior might be in the polyanionic nature of HA. Moreover, a rise in ionic strength causes HA molecules to shift to a more compact conformation, possibly altering their sensitivity to EO [69,70].
UV-Vis spectral analysis, TOC measurement, size-exclusion chromatography and elemental analysis provided useful information concerning the nature of the chemical species generated by the EO process. Figure 6 reports the UV-Vis spectra of untreated HA solution and of two HA solutions containing NaCl or NaClO4 after the EO treatment. The results presented above indicate that ionic strength plays an active role in HA EO. In line with transition state theory [68], a possible explanation for this behavior might be in the polyanionic nature of HA. Moreover, a rise in ionic strength causes HA molecules to shift to a more compact conformation, possibly altering their sensitivity to EO [69,70].
UV-Vis spectral analysis, TOC measurement, size-exclusion chromatography and elemental analysis provided useful information concerning the nature of the chemical species generated by the EO process. Figure 6 reports the UV-Vis spectra of untreated HA solution and of two HA solutions containing NaCl or NaClO 4 after the EO treatment.  The spectrum of the untreated HA solution shows the typical HA absorbance in the UV-Vis region, reflecting the presence of conjugated double bonds systems and aromatic groups [71]. The spectra show a hypsochromic shift and a concomitant hypochromic effect in both treated solutions. These spectroscopic effects can only in part be ascribed to HA mineralization, in line with TOC reduction from 58.2 mg L −1 to 38.0 (≈35% TOC reduction) and 29.0 mg L −1 (≈50% TOC reduction) in the NaClO4 and NaCl solution, respectively. Figure 7 reports LP-SEC chromatograms of HA solution containing 10 mM NaCl, before and after EO and, for comparison, of vanillic acid (a low-molecular-weight HA-like compound). At the end of the EO experiment, the solution shows a reaction product with an intermediate retention time between the pristine HA sample and vanillic acid. This suggests that the treatment induces the formation of compounds with lower hydrodynamic volume (i.e., the volume of the molecules in solution) and lower molecular weight than the original HA molecules, while still retaining a macromolecular character. These findings are supported by dialysis experiments, which showed that the product of HA EO did not diffuse across a membrane with a relatively high cut off (6-8 kDa).
Elemental analysis showed that the C/H ratio decreased from 12.4 in the untreated solution to 9.0 and 8.5 in treated solutions containing NaClO4 and NaCl, respectively. This suggests that the reaction product had a lower degree of unsaturation than the initial HA.
The full characterization and identification of byproducts was not possible. However, the The spectrum of the untreated HA solution shows the typical HA absorbance in the UV-Vis region, reflecting the presence of conjugated double bonds systems and aromatic groups [71]. The spectra show a hypsochromic shift and a concomitant hypochromic effect in both treated solutions. These spectroscopic effects can only in part be ascribed to HA mineralization, in line with TOC reduction from 58.2 mg L −1 to 38.0 (≈35% TOC reduction) and 29.0 mg L −1 (≈50% TOC reduction) in the NaClO 4 and NaCl solution, respectively. Figure 7 reports LP-SEC chromatograms of HA solution containing 10 mM NaCl, before and after EO and, for comparison, of vanillic acid (a low-molecular-weight HA-like compound). At the end of the EO experiment, the solution shows a reaction product with an intermediate retention time between the pristine HA sample and vanillic acid. This suggests that the treatment induces the formation of compounds with lower hydrodynamic volume (i.e., the volume of the molecules in solution) and lower molecular weight than the original HA molecules, while still retaining a macromolecular character. These findings are supported by dialysis experiments, which showed that the product of HA EO did not diffuse across a membrane with a relatively high cut off (6-8 kDa). The TOC reduction obtained in the present work is quite low compared to that reported in other studies on HA EO, the main reason probably being passivation and high overpotential of platinum electrodes [66]. Alternative electrode materials may strongly enhance the dispersion of anions in the solution and, therefore, the removal of HA [74][75][76]. Better HA aqueous removal efficiency was obtained, for example, by Deng et al. [49] using a Ti/RuO2 anode, an electric current density of 100 mA/cm 2 and pH 6.5. Those authors observed a reduction of chemical oxygen demand (COD) by over 80% in biologically pre-treated landfill leachate. Wang et al. [10] reported complete mineralization of HA from an effluent of biotreated coking wastewater after 30-min treatment involving boron-doped diamond anodes. Trellu et al. [59] obtained nearly 99% reduction of TOC in drinking water within 7h treatment at 1000 mA, using a boron doped diamond anode and a stainless-steel cathode.

Kinetic Modelling
In line with former work [60], the experimental data reported above suggest that the degradation of HA may proceed through one or two pathways, depending on whether chloride is present or not.
In the presence of inert supporting electrolytes (in our protocol, NaClO4 or NaNO3) but without chlorides (or of other salts, such as sulfates, from which oxidant species can be electro-generated), HA EO only depends on the formation of hydroxyl radicals by water discharge at the anode [52]. The first step of this oxidative pathway is water interaction with active sites at the anode surface leading to the formation of adsorbed hydroxyl radicals (Equation (1)): In the second step, hydroxyl radicals oxidize HA (Equation (2)): Pt(OH • ) + HA  Pt + Oxidized HA (2) Oxidation can only occur in close proximity to the anode, because hydroxyl radicals have a very short lifetime [52]. Thus, the rate of reaction (2) most likely controls the overall reaction rate. If so, HA concentration should be a primary determinant of the reaction rate. Our experimental results Elemental analysis showed that the C/H ratio decreased from 12.4 in the untreated solution to 9.0 and 8.5 in treated solutions containing NaClO 4 and NaCl, respectively. This suggests that the reaction product had a lower degree of unsaturation than the initial HA.
The full characterization and identification of byproducts was not possible. However, the formation of chlorinated compounds has been observed by several authors [72,73].
The TOC reduction obtained in the present work is quite low compared to that reported in other studies on HA EO, the main reason probably being passivation and high overpotential of platinum electrodes [66]. Alternative electrode materials may strongly enhance the dispersion of anions in the solution and, therefore, the removal of HA [74][75][76]. Better HA aqueous removal efficiency was obtained, for example, by Deng et al. [49] using a Ti/RuO 2 anode, an electric current density of 100 mA/cm 2 and pH 6.5. Those authors observed a reduction of chemical oxygen demand (COD) by over 80% in biologically pre-treated landfill leachate. Wang et al. [10] reported complete mineralization of HA from an effluent of biotreated coking wastewater after 30-min treatment involving boron-doped diamond anodes. Trellu et al. [59] obtained nearly 99% reduction of TOC in drinking water within 7-h treatment at 1000 mA, using a boron doped diamond anode and a stainless-steel cathode.

Kinetic Modelling
In line with former work [60], the experimental data reported above suggest that the degradation of HA may proceed through one or two pathways, depending on whether chloride is present or not.
In the presence of inert supporting electrolytes (in our protocol, NaClO 4 or NaNO 3 ) but without chlorides (or of other salts, such as sulfates, from which oxidant species can be electro-generated), HA EO only depends on the formation of hydroxyl radicals by water discharge at the anode [52]. The first step of this oxidative pathway is water interaction with active sites at the anode surface leading to the formation of adsorbed hydroxyl radicals (Equation (1)): In the second step, hydroxyl radicals oxidize HA (Equation (2)): Oxidation can only occur in close proximity to the anode, because hydroxyl radicals have a very short lifetime [52]. Thus, the rate of reaction (2) most likely controls the overall reaction rate. If so, HA concentration should be a primary determinant of the reaction rate. Our experimental results support this conclusion by showing a linear relationship between the reaction rate and HA concentration ( Figure 4A).
The rate of HA degradation in the presence of NaCl as electrolyte was significantly higher than with NaClO 4 or NaNO 3 (Figures 2 and 4), indicating that HA oxidation not only occurred via hydroxyl radicals, as was the case with inert electrolytes, but additionally involving active chlorine species originating from chloride ions during the electrochemical process. Active chlorine species such as Cl 2 , HClO and ClO − may be generated at the anode by the following reactions [52]: HClO ↔ H + + ClO − Although active chlorine species have lower reduction potentials than hydroxyl radicals, they may be more efficient oxidants because, after formation at the anode, they are sufficiently stable to diffuse into the solution and bulk-react with HA. The relative abundance of Cl 2 , HClO and ClO − depends on pH. At pH 10 chosen for our experiments, ClO − dominates over the other species and is probably the main driver of HA EO [77].
The reactions of HA oxidation by hydroxyl radicals and active chlorine species probably proceed in parallel as sketched in Figure 8, with k OH and k Cl being the kinetic rate constants of either pathway.
Water 2020, 12,2250 12 of 18 support this conclusion by showing a linear relationship between the reaction rate and HA concentration ( Figure 4A). The rate of HA degradation in the presence of NaCl as electrolyte was significantly higher than with NaClO4 or NaNO3 ( Figure 2; Figure 4), indicating that HA oxidation not only occurred via hydroxyl radicals, as was the case with inert electrolytes, but additionally involving active chlorine species originating from chloride ions during the electrochemical process. Active chlorine species such as Cl2, HClO and ClO − may be generated at the anode by the following reactions [52]: Cl2 + H2O  HClO + H + + Cl − HClO  H + + ClO − Although active chlorine species have lower reduction potentials than hydroxyl radicals, they may be more efficient oxidants because, after formation at the anode, they are sufficiently stable to diffuse into the solution and bulk-react with HA. The relative abundance of Cl2, HClO and ClO − depends on pH. At pH 10 chosen for our experiments, ClO − dominates over the other species and is probably the main driver of HA EO [77].
The reactions of HA oxidation by hydroxyl radicals and active chlorine species probably proceed in parallel as sketched in Figure 8, with kOH and kCl being the kinetic rate constants of either pathway. In order to model the kinetics of HA degradation, we used the initial reaction rate method, a method widely used in enzymatic kinetics [78]. This approach has the advantage of being independent of the time evolution of reactive species and operational conditions (e.g., HA, pH, concentration of chlorine species and OH • ).
Based on the above reasoning and in line with the data presented, we propose the following equation to describe the initial rate of HA EO in the presence of inert supporting electrolytes (v0,OH, upper reaction pathway in Figure 8): v0,OH = −(dCHA/dt)0 = kOH × CHA × IS n × I where CHA, IS and I are the initial HA concentration, ionic strength and current density, respectively, and n is a constant. The rate of HA EO in the presence of Cl − (lower reaction pathway in Figure 8) depends on the formation rate of both hydroxyl radicals and active chlorine species. In line with our experimental data, the initial rate of HA electro-oxidation in the presence of Cl − (v0,Cl) varies with the initial chloride concentration (CCl − ), CHA, IS and I according to the following relation: In order to model the kinetics of HA degradation, we used the initial reaction rate method, a method widely used in enzymatic kinetics [78]. This approach has the advantage of being independent of the time evolution of reactive species and operational conditions (e.g., HA, pH, concentration of chlorine species and OH • ).
Based on the above reasoning and in line with the data presented, we propose the following equation to describe the initial rate of HA EO in the presence of inert supporting electrolytes (v 0 , OH , upper reaction pathway in Figure 8): v 0,OH = −(dC HA /dt) 0 = k OH × C HA × IS n × I where C HA , IS and I are the initial HA concentration, ionic strength and current density, respectively, and n is a constant. The rate of HA EO in the presence of Cl − (lower reaction pathway in Figure 8) depends on the formation rate of both hydroxyl radicals and active chlorine species. In line with our experimental data, the initial rate of HA electro-oxidation in the presence of Cl − (v 0,Cl ) varies with the initial chloride concentration (C Cl − ), C HA , IS and I according to the following relation: The quantity IS n in Equations (7) and (8) was introduced to take into account the observed effect of ionic strength, which unlike C HA , C Cl − and I cannot be described by a first-kinetic-order reaction (cf. Figure 5).
The overall initial rate (v 0 ) of HA electro-oxidation can be obtained by summing up Equations (7) and (8): To test the validity of the above model and estimate the kinetic parameters, the initial reaction rate (v 0 ) values shown in Figure 3 were fitted to Equation (9) using a multivariable fitting procedure (Table 1).
Our kinetic model fitted the experimental data well (as shown by a high adjusted R 2 value) and produced a reliable estimation of kinetic parameters (because of relatively low associated standard errors). As expected, the results point to a significant effect of ionic strength (n = 0.24). Noticeably, the values of the estimated kinetic parameters suggest that, at low Cl − concentration (e.g., 10 mM), the rate of the oxidation of HA is mainly promoted by the hydroxyl radicals, i.e., the pathway mediated by hydroxyl radicals predominates over the pathway depending on active chlorine species.
The accuracy of the above model was verified by parity plotting model predictions and experimentally measured values of initial v 0 ( Figure 9). Most points fell in a variation range of ±30%, indicating a satisfactory agreement between predicted and experimental data for all operative conditions tested. This supports our kinetic model as a reliable description of HA EO in the experimental conditions investigated. The quantity IS n in Equations (7) and (8) was introduced to take into account the observed effect of ionic strength, which unlike CHA, CCl − and I cannot be described by a first-kinetic-order reaction (cf. Figure 5).
The overall initial rate (v0) of HA electro-oxidation can be obtained by summing up Equations (7) and (8): To test the validity of the above model and estimate the kinetic parameters, the initial reaction rate (v0) values shown in Figure 3 were fitted to Equation (9) using a multivariable fitting procedure (Table 1).  Our kinetic model fitted the experimental data well (as shown by a high adjusted R 2 value) and produced a reliable estimation of kinetic parameters (because of relatively low associated standard errors). As expected, the results point to a significant effect of ionic strength (n = 0.24). Noticeably, the values of the estimated kinetic parameters suggest that, at low Cl − concentration (e.g., 10 mM), the rate of the oxidation of HA is mainly promoted by the hydroxyl radicals, i.e., the pathway mediated by hydroxyl radicals predominates over the pathway depending on active chlorine species.
The accuracy of the above model was verified by parity plotting model predictions and experimentally measured values of initial v0 (Figure 9). Most points fell in a variation range of ±30%, indicating a satisfactory agreement between predicted and experimental data for all operative conditions tested. This supports our kinetic model as a reliable description of HA EO in the experimental conditions investigated.

Conclusions
The electro-treatment of HA solution using Pt electrodes led to rapid oxidation of HA into smaller molecules with lower aromaticity and conjugation degree. The impacts of these products on the ecosystem and human health remain to be determined. Humic acids EO may proceed through two concurrent reactions, one mediated by reactive chlorine species, the other by hydroxyl radicals, the latter being dominant at low chlorine concentration. Kinetic analysis suggests that the overall reaction rate is of a first-order relative to HA and Cl − concentrations and current density. An appreciable effect of ionic strength was observed, probably due to the polyanionic nature of HA molecules. A kinetic model was proposed and successfully applied to predict the rate of HA electro-oxidation. Funding: This work has been supported by VALERE "VAnviteLli pEr la RicErca" PROGRAMME funded by the University of Campania "Luigi Vanvitelli".