Greener Catalytic Oxidation of Azole Fungicides: Coupling EO–O₃ on BDD with Kinetics and Mineralization Targets

Abstract

This study evaluates the abatement of four common azole fungicides—prochloraz, tebuconazole, tetraconazole, and penconazole—using ozonation (O₃), electro-oxidation (EO on boron-doped diamond anode), and their coupling (EO–O₃). A central composite design (CCD) with three coded factors—current (A), electrolyte (B), and ozone concentration in the gas phase (C)—was employed to model three responses: pollutant abatement (%), apparent pseudo-first-order rate constant k (min⁻¹), and TOC removal (%). Quadratic models showed good in-samples (R² ≈ 0.84–0.86). Ozone and current dominate abatement and kinetics (with curvature in current), while the electrolyte penalizes mineralization and narrows the window for TOC removal. Under optimal conditions, 116 mA (current), 0.992 mM (electrolyte), and 7.09 ppm (ozone concentration), the EO–O₃ configuration results in a TOC removal of 33.78%. At a reaction time of 10 min (total abatement of the pollutants), the hybrid EO–O₃ configuration exhibits a specific energy consumption (SEC) of 1.825 kWh·m⁻³. We compare trends with the last decade of literature on ozone-based EAOPs, electro-peroxone variants, and BDD anodic oxidation, and outline practical guidance for its application and scale-up, and model refinement in predictive settings.

1. Introduction

Azole fungicides are pervasive agrochemicals in modern agriculture and horticulture. Their triazole or imidazole rings and halogenated phenyl moieties confer persistence and biological activity at low concentrations, complicating removal by conventional biological treatment. Field and monitoring campaigns have reported azoles in surface waters, wastewater effluents, and occasionally in drinking-water sources at ng·L⁻¹ to µg·L⁻¹ levels [1], with seasonal pulses reflecting application cycles and storm-driven runoff. Beyond persistence, certain azole transformation products (TPs) [2] can retain endocrine activity or antimicrobial properties, motivating advanced treatment that goes beyond primary disappearance to achieve carbon removal [3].

Advanced oxidation processes (AOPs) offer versatile routes to oxidize recalcitrant organics via reactive oxygen species (ROS). In ozonation (O₃) [4], direct electrophilic reactions occur at activated sites, while indirect radical pathways emerge as ozone decomposes to hydroxyl radicals (•OH) [5,6,7]. The balance between direct and radical routes depends on pH, alkalinity, matrix species, and promoters (e.g., H₂O₂ in peroxone schemes) [8,9,10]. Other ozone-based AOPs (coupled with solar radiation, UV, active carbon, or chlorine) have also shown great potential [11,12,13,14].

Electro-oxidation (EO) with boron-doped diamond (BDD) anodes [15,16,17,18] is a robust electrochemical advanced oxidation process (EAOP) that generates surface-bound •OH with a high standard potential and minimal adsorption, enabling non-selective oxidation under mild conditions [19,20]. Integrating O₃ with EO exploits complementary transport and reaction pathways: ozone delivery through the gas–liquid interface and radical generation near the anode, as well as peroxone-like chains when H₂O₂ is electro-generated at the cathode [21,22].

While mechanistic models exist for individual AOPs [4,15], engineering design and optimization of coupled AOPs [23,24,25] require empirical maps of factor influence and curvature across realistic operating windows. In this connection, the response-surface methodology (RSM) [26,27] is a pragmatic framework to achieve this: by using structured designs such as central composite designs (CCDs), one can quantify main effects, two-factor synergies, and curvature (quadratic terms) with relatively few experiments [28,29,30].

Recent reviews have emphasized the rapid progress of ozone-based electrochemical AOPs [31,32], including electro-peroxone (E-peroxone) [21,22] and electrolysis-catalyzed ozonation, highlighting electrode stability, reactor configurations, and energy considerations. Likewise, state-of-the-art reports on EAOPs underline the durability and broad reactivity of BDD anodes [15], but also the sensitivity of selectivity and mineralization to electrolyte identity and current density. For azoles in particular, single-stage ozonation typically yields fast primary decay consistent with pseudo-first-order kinetics in dilute solutions. In contrast, TOC removal lags as aromatic acids and ring-opened products accumulate. Therefore, EO extends oxidation [32] but may divert the pathway in chloride-rich media towards active chlorine and chlorine-radical chemistry, increasing apparent abatement without guaranteeing deeper mineralization.

Coupling O₃ with EO is therefore attractive but non-trivial: mass transfer of ozone, hydrodynamics of bubble dispersion, local radical fluxes at the anode, and electrolyte-dependent side chemistry all interplay. Micro- and nanobubble ozonation [10] can enhance gas–liquid interfacial area and ozone utilization, offering another dimension for process intensification. Yet, aggressive conditions may trigger diminishing returns due to mass-transfer ceilings, scavenging, or parasitic reactions. From a statistical vantage, such behavior manifests as interior optima (negative quadratic terms) and, at times, significant lack-of-fit of quadratic polynomial signals that the surface bends more sharply than the model can capture or that an unmodelled factor exerts leverage.

Although numerous studies have examined electro-peroxone and other ozone-assisted electrochemical advanced oxidation processes (EAOPs) [21,22,31,32,33,34,35,36,37,38,39,40], most have focused on individual pollutants, providing mechanistic insights without quantitative optimization. Recent reviews [15,17,31,32] highlight the lack of integrative studies that couple kinetics, mineralization, and energy efficiency under realistic, multi-contaminant conditions. The present work addresses this gap by coupling electro-oxidation on boron-doped diamond (BDD) with gas-phase ozonation (EO–O₃) and implementing a central composite design (CCD) to develop response-surface models that simultaneously describe pollutant abatement, apparent rate constant (k), and TOC removal for a mixture of four representative azole fungicides. In contrast to previous works that have assessed EO or O₃ separately, or hybrid systems without statistical optimization, this study provides, for the first time, an integrated quantitative framework that identifies the main interaction and curvature effects of current, electrolyte, and ozone dosage. It also elucidates the inhibitory effect of sulfate ions on radical reactivity and defines an operational window that achieves balanced kinetics, mineralization, and energy efficiency (SEC ≈ 1.8 kWh·m⁻³).

Here, we analyze experimental data for the EO–O₃ treatment of a four-azole mixture (prochloraz, tebuconazole, tetraconazole, and penconazole) under a CCD, reconstructing quadratic models and visualizing their response surfaces [27]. In this study, we reconstructed quadratic models for three responses that capture complementary objectives: pollutant abatement (removal percentage, %), the apparent pseudo-first-order kinetic constant k (min⁻¹), and mineralization (TOC removal, %). Three coded factors—the galvanostatic current (A), the supporting electrolyte Na₂SO₄ concentration (B), and the ozone concentration (C) in the feed gas—are the levers most commonly available to practitioners when implementing the EO–O₃ coupled process. We assess model adequacy using ANOVA and R² metrics (both adjusted and predicted). We then interpret the statistics mechanistically, compare with recent literature on ozone-assisted EAOPs, and derive operating guidance. Finally, we discuss implications for predictive use and energy metrics (e.g., electrical energy per order, EE/O) to inform scale-up.

2. Results and Discussion

The design of experiments (DOEs) followed the matrix reported in Table 1. The coded center point corresponded to 101 mA (current), 13 mM (electrolyte concentration), and 5.5 ppm (gas-phase ozone concentration). The factor step sizes were 60 mA, 7.14 mM, and 2.7 ppm, respectively.

Table 1. Coded values for the design of experiments.

Run Nº	Current (A)	Electrolyte Concentration (B)	Ozone Concentration (C)
1	−1	−1	−1
2	0	0	0
3	0	−1.68179	0
4	0	0	1.68179
5	1	1	−1
6	1	1	1
7	0	0	−1.68179
8	0	0	0
9	−1.68179	0	0
10	−1	−1	1
11	0	0	0
12	1.68179	0	0
13	1	−1	1
14	0	1.68179	0
15	0	0	0
16	1	−1	−1
17	−1	1	1
18	−1	1	−1
19	0	0	0
20	0	0	0

The responses were described by a second-order response-surface model defined in terms of the coded factors A (current), B (electrolyte concentration), and C (gas-phase ozone concentration). The polynomial comprised linear (A, B, C), two-factor interaction (AB, AC, BC), and quadratic (A², B², C²) terms and were estimated by ordinary least squares using data from a randomized central composite design (CCD). Model adequacy was established through analysis of variance—reporting the model F-statistic and associated p-value—together with coefficients of determination (R² and adjusted R²) and standard residual diagnostics (assessment of normality, homoscedasticity, and leverage/influence). The fitted surface was subsequently exploited to render response surfaces and contour plots, elucidate main and interaction effects, and locate operating optima within the experimental domain (stationary points and/or targets derived from a multi-response desirability function).

2.1. Pollutants Abatement

2.1.1. ANOVA Test Interpretation

In this case, the target variable was the removal (%) of the sum of concentrations of the four pollutants. As shown in Table 2, the quadratic model adequacy was measured by R² = 0.855 (adjusted R² = 0.725) and a global ANOVA F = 6.570 (p = 0.0035). The small R²–R²(adj) gap confirms that the quadratic model captures curvature effects effectively, with no evidence of overfitting.

Table 2. Model ANOVA (model vs. error) for compound abatement.

Source	SS	df	MS	F	p
Model	2575	9	286.1	6.570	0.0035
Error	435.5	10	43.55
Total	3011	19

In ANOVA regression, the SS (sum of squares) quantifies the variability attributed to each source (model, term, error); dfs (degrees of freedom) indicate how much information underlies each SS; MS (mean square) is the variance estimated for that source (MS = SS/df); the F-statistic compares explained to residual variance (F = MS_source/MS_error) to test whether the effect exceeds random noise; and the p-value is the probability of observing an F at least as extreme under the null hypothesis (coefficients = 0), such that a small p suggests the model or term is statistically significant.

2.1.2. Influence of Operating Variables and Optimization

The combined statistical–mechanistic analysis clarifies the operating levers that matter for treating azole fungicides with the EO–O₃ process. Ozone concentration (C), firstly, and galvanostatic current (A), secondly, are consistently the most influential factors for pollutant abatement (the sum of the four azole concentrations). Their positive effects persist up to curvature-imposed plateaus, beyond which additional input yields diminishing returns. These plateaus reflect either mass-transfer ceilings, radical recombination, or the onset of side pathways—fingerprinted by negative A² and C² terms in the fitted models (see Table 3).

Table 3. Regression coefficients and significance for compound abatement (10 min reaction time).

Term	Coef	SE	t	p
Intercept	87.3580	2.6918	32.4539	0.0000
A	5.5234	1.7859	3.0927	0.0114
B	0.6957	1.7859	0.3896	0.7050
C	11.3093	1.7859	6.3325	0.0001
AB	−1.4563	2.3334	−0.6241	0.5465
AC	−0.7588	2.3334	−0.3252	0.7518
BC	−0.0788	2.3334	−0.0337	0.9737
A2	−4.6469	1.7385	−2.6729	0.0234
B2	−0.1303	1.7385	−0.0749	0.9417
C2	−2.6352	1.7385	−1.5158	0.1605

In regression outputs, SE denotes the coefficient’s standard error (its estimation uncertainty); t is the ratio of the estimate to its SE, quantifying departure from zero; and p is the probability, under the null (actual coefficient = 0), of observing an equal or larger |t|. Consequently, larger |t| (smaller p) indicates a statistically significant term.

The reaction mechanism involves both surface-bound oxidants generated at the BDD anode and solution-phase radical pathways sustained by ozone decomposition. At the anode, water undergoes one-electron oxidation to produce adsorbed hydroxyl radicals (•OH ads), as shown in Equation (1):

H₂O → •OH ads + H⁺ + e⁻ (BDD water oxidation)

(1)

The high oxygen-evolution overpotential of BDD favors the accumulation of •OH ads, which can oxidize organic pollutants directly or initiate further reactions in solution.

Ozone (added in the gas phase) or produced electrochemically at the anode dissolves into the bulk, where its reactivity depends strongly on pH. Under alkaline conditions, ozone enters a radical chain pathway initiated by its reaction with hydroxide ions:

O₃ + OH⁻ → O₂ + HO₂⁻ (ozone chemical decomposition)

(2)

The formation of HO₂⁻ is the initiation step that governs the pH dependence of ozone transformations. The hydroperoxide anion then reacts with ozone, generating both superoxide and hydroxyl radicals in the propagation stage:

O₃ + HO₂⁻ → O₂ + O₂•⁻ + •OH (propagation chain reaction)

(3)

In parallel, ozone may undergo cathodic reduction according to Equation (4):

O₃ + H₂O + 2e⁻ → O₂ + 2OH⁻ (ozone cathodic reduction)

(4)

This pathway increases the concentration of OH⁻ and contributes to the formation of reactive oxygen species that participate in the oxidation sequence.

When electro-oxidation and ozonation coincide, the interaction between anode-generated •OH and dissolved O₃ intensifies the radical environment. The coupled effect is schematically represented by Equation (5):

O₃ + BDD(H₂O) → 2 •OH + O₂ (coupled EO–O₃)

(5)

The radicals produced at the electrode accelerate ozone decomposition, while the intermediates derived from O₃ extend the oxidative lifetime of reactive species in the bulk solution. This mutual reinforcement explains the high efficiency of the combined EO–O₃ system for the degradation of recalcitrant organic contaminants.

As shown in Figure 1, ozone (coded as C) exerts the most significant positive effect, followed by current (A). The negative A² and C² terms imply interior optima in A and C; raising current or ozone concentration improves abatement up to a curvature-limited plateau. Electrolyte (B) shows a negligible (no significant) positive effect on this response. This pattern aligns with ozone-assisted EAOP reports, where synergistic O₃/•OH mechanisms and mass-transfer-limited plateaus are widely documented [33,34,35]. Elevated current densities or high ozone concentrations can negatively impact oxidation efficiency due to radical recombination and deactivation. Excessive ozone may lead to the accumulation of reactive oxygen species, such as hydroxyl radicals (•OH), which at high local concentrations rapidly recombine following reactions like the following (see Equations (6) and (7)):

2 •OH → H₂O₂ (radical recombination)           

(6)

•OH + H₂O₂ → H₂O+ ·HO₂ (less powerful oxidant)     

(7)

Figure 1. Two-dimensional contour map for compound abatement (%) (A vs. C; B = 0).

This recombination reduces the availability of free radicals needed for pollutant degradation. Similarly, high currents can accelerate radical generation beyond optimal levels, favoring termination reactions and dissipating oxidative power. Moreover, excess ozone can act as a radical scavenger, reacting with •OH to form less reactive species, thus decreasing overall oxidation capacity (Equation (8)).

O₃ + •OH → ·HO₂ + O₂

(8)

2.2. Kinetics as Target Variable

2.2.1. ANOVA Test Interpretation

In this case, the target variable was the pseudo-first-order kinetic rate constant, k (min⁻¹). The ANOVA test, R² = 0.842 (adjusted R² = 0.700), and global ANOVA F = 5.91 (p = 0.0052) demonstrate the adequate simulation capacity of the model. The quadratic form captures k satisfactorily in the studied domain, supporting its use for rapid screening and prioritization. The small R²–R²(adj) gap confirms that the quadratic model captures curvature effects effectively, with no evidence of overfitting.

2.2.2. Influence of Operating Variables and Optimization

The combined statistical analysis (see Table 4) clarifies that ozone concentration (C) (first) and galvanostatic current (A) (second) are the most influential factors for kinetics removal. Their positive effects persist up to curvature-imposed plateaus, beyond which additional input yields diminishing returns. These plateaus reflect either mass-transfer ceilings, radical recombination, or the onset of side pathways (see Equations (6) and (7) of the previous section).

Table 4. Regression coefficients and significance for k (10 min reaction time).

Term	Coef	SE	t	p
Intercept	0.1994	0.0148	13.4403	0.0000
A	0.0250	0.0098	2.5428	0.0292
B	0.0028	0.0098	0.2840	0.7822
C	0.0624	0.0098	6.3438	0.0001
AB	−0.0067	0.0129	−0.5210	0.6137
AC	0.0089	0.0129	0.6959	0.5023
BC	0.0043	0.0129	0.3324	0.7464
A2	−0.0214	0.0096	−2.2346	0.0495
B2	0.0045	0.0096	0.4700	0.6484
C2	0.0004	0.0096	0.0402	0.9688

As shown in Figure 2, both C and A increase k; the curvature in A signals diminishing returns at high current. The A–C synergy is evident as sloped ridges in the response surface, consistent with reports where peroxone-like chains and enhanced radical availability drive kinetic gains in E-peroxone configurations (see Equation (8)).

Figure 2. Three-dimensional response surface for k (min⁻¹) as a function of A and C (B = 0).

Related works [38,39,40] describe the same phenomena: an increase in kinetics with current and ozone, their synergy, diminishing returns at high currents (due to radical recombination and the onset of side pathways), and the crucial role of peroxone-like radical chains in driving the reaction kinetics.

2.3. Mineralization as Target Variable

The TOC removal data were obtained at 60 min, since at 10 min the removal levels were too low to be significant for assessing the influence of the operating variables.

2.3.1. ANOVA Test

For this case, the authors considered mineralization (TOC removal, %) as the target variable. The model adequacy R² = 0.836 (adjusted R² = 0.690) and global ANOVA F = 5.65 (p = 0.0061) confirm the less-predictive ability of the model (compared with previous target variables), in agreement with the common observation that carbon removal involves slower, multi-step pathways not easily approximated by a single quadratic surface.

2.3.2. Influence of Operating Variables and Optimization

Mineralization (TOC removal) is governed by a different balance: electrolyte (B) exerts a statistically significant and negative main effect. As shown in Table 5, electrolyte (B) penalizes TOC removal. The deactivation of hydroxyl radicals (•OH) by sulfate anions (SO₄²⁻) occurs through reactions that reduce the availability of •OH for oxidation processes (see Equation (9)). The primary reactions involve sulfate radicals (SO₄⁻), which are formed via the reaction of sulfate ions with •OH under specific conditions. This process generates sulfate radicals (SO₄^·−), which, although reactive, are less reactive than •OH radicals.

·OH + SO₄²⁻ →·SO₄⁻ + OH⁻

(9)

Table 5. Regression coefficients and significance for TOC removal (60 min of reaction time).

Term	Coef	SE	t	p
Intercept	25.1129	2.8895	8.6912	0.0000
A	3.6614	1.9171	1.9099	0.0852
B	−7.6605	1.9171	−3.9959	0.0025
C	4.2213	1.9171	2.2019	0.0523
AB	0.4688	2.5048	0.1871	0.8553
AC	−0.1088	2.5048	−0.0434	0.9662
BC	−2.7513	2.5048	−1.0984	0.2978
A2	−5.8386	1.8662	−3.1286	0.0107
B2	−2.5029	1.8662	−1.3411	0.2095
C2	−7.7425	1.8662	−4.1487	0.0020

Sulphate radicals can oxidize organic compounds, but typically at lower rate constants compared to •OH. Carbonates and bicarbonates (CO₃²⁻/HCO₃⁻) also act as hydroxyl-radical scavengers, forming carbonate radicals (CO₃⁻) that are less reactive and more selective, thereby reducing overall oxidation efficiency. The overall effect is a reduction in the availability of highly reactive •OH radicals, leading to a “quenching” or deactivation of ·OH oxidative capacity in sulfate-rich environments. This reaction pathway is especially significant in systems where sulfate ions are present as electrolyte support (this case). This explains the empirical observation that treatments delivering nearly complete parent removal may still fall short in TOC, pathways dominated by active •OH radicals.

Regarding the curvature in the A and C shapes, a narrow window for mineralization (see Figure 3) was observed. Practically, mineralization benefits from sulfate-based electrolytes at the lowest concentration compatible with conductivity, and from operating A and C near interior optima rather than at extremes. These effects (A and C) were just explained for pollutant abatement and kinetics in previous sections (see Equations (1)–(8)). High ozone or current doses decreased mineralization rates and generated inefficient energy use. Controlling current and ozone concentration is therefore critical to maintain a balance that maximizes radical production while minimizing recombination or deactivation, ensuring efficient pollutant removal and favorable operating costs.

Figure 3. Two-dimensional contour map for TOC removal (%) (A vs. C; B = 0).

Within the CCD design domain (A, B, C ∈ [−1.6818, 1.6818]), the constrained optimum for TOC removal was located at A = +0.2523, B = −1.6818 (boundary), and C = +0.5886 in coded units, corresponding to actual settings of 116.14 mA (current), 0.992 mM (electrolyte), and 7.089 ppm (gas-phase O₃). Under these conditions, the model predicts a TOC removal of ≈33.78%. The placement of B at its lower bound indicates that minimizing electrolyte concentration is beneficial within the explored domain.

In general, the EO–O₃ process remains among the most robust anode technologies, but electrolyte management is pivotal for steering selectivity toward mineralization. Micro-nanobubble ozonation [9,10] emerges as a complementary lever to increase interfacial area and enhance ozone mass transfer, although energy metrics (EE/O) and bubble management must be balanced to avoid parasitic losses.

2.4. Energy Considerations

At reaction times of 10 min (considered in this research for total abatement of the pollutants), the hybrid EO–O₃ configuration exhibits a specific energy consumption (SEC) of 1.825 kWh·m⁻³, apportioned as 1.200 kWh·m⁻³ (≈66%) from electro-oxidation and 0.625 kWh·m⁻³ (≈34%) from ozonation (see Table 6). For the constrained optimum conditions used here (TOC removal 33.78%), the energy per %TOC removed was 0.324 kWh·m⁻³·(%TOC)⁻¹ (considering a reaction time of 60 min).

Table 6. Energy metrics for the EO–O₃ process (10 min of reaction time).

Metric	Formula (Word Equation)	Units	Value
EO power density	${\mathrm{P}}_{\mathrm{E}\mathrm{O}}$	W·L−1	7.20
O3 power density	${\mathrm{P}}_{\mathrm{O}3}$	W·L−1	3.75
Total power density	${\mathrm{P}}_{\mathrm{t}\mathrm{o}\mathrm{t}}={\mathrm{P}}_{\mathrm{E}\mathrm{O}}+{\mathrm{P}}_{\mathrm{O}3}$	W·L−1	10.95
Treatment time	$\mathrm{t}$	s	600
EO energy per liter	${\mathrm{E}}_{\mathrm{E}\mathrm{O}}={\displaystyle \frac{{\mathrm{P}}_{\mathrm{E}\mathrm{O}}·\mathrm{t}}{3.6\times {10}^6}}$	kWh·L−1	0.00120
O3 energy per liter	${\mathrm{E}}_{\mathrm{O}3}={\displaystyle \frac{{\mathrm{P}}_{\mathrm{O}3}·\mathrm{t}}{3.6\times {10}^6}}$	kWh·L−1	0.00062
Total energy per liter	${\mathrm{E}}_{\mathrm{t}\mathrm{o}\mathrm{t}}={\mathrm{E}}_{\mathrm{E}\mathrm{O}}+{\mathrm{E}}_{\mathrm{O}3}$	kWh·L−1	0.00182
Specific energy consumption	$\mathrm{S}\mathrm{E}\mathrm{C}=1000·{\mathrm{E}}_{\mathrm{t}\mathrm{o}\mathrm{t}}$	kWh·m−3	1.825
Electrical energy per order	${\mathrm{E}\mathrm{E}}_{/\mathrm{O}}=\mathrm{S}\mathrm{E}\mathrm{C}/{\mathrm{l}\mathrm{o}\mathrm{g}}_{10}(\mathrm{C}0/\mathrm{C}\mathrm{t})$	kWh·m−3·order−1	10.19

Collectively, these metrics indicate that improving gas–liquid mass transfer of ozone (higher k_La, finer bubbles, optimized contact time, and off-gas management) is a more effective lever for reducing overall energy intensity than merely increasing current density. In practice, operating at a moderate current (to limit parasitic losses and resistive heating), achieving adequate but efficiently dissolved ozone and maintaining minimal electrolytes at the effective conductivity threshold, provides a favorable trade-off between kinetics and power demand, supporting a pragmatic pathway to scale-up with manageable OPEX.

The TOC removal values reported in this work correspond to 60 min of treatment, which represents the plateau region of mineralization. Extending the reaction time beyond this point would provide only marginal additional TOC reduction while significantly increasing specific energy consumption (estimated EE/O > 15 kWh·m⁻³·order⁻¹). Therefore, the EO–O₃ process is more suitable as a pre-oxidation or polishing step to enhance biodegradability with moderate energy demand, rather than as a stand-alone total mineralization technology.

2.5. Practical Implications and Scale-Up

The experimental results indicate that gas-phase ozone dosage (C) is the primary driver of both the apparent rate constant (k) and percent of azole removal, whereas the current (A) exhibits curvature (negative A²), implying diminishing returns beyond a moderate range. Operationally, settings that elevate C while maintaining A at an intermediate level sustain reactivity without unduly increasing electrical demand. Electrolyte concentration (B) does not enhance kinetics or overall removal and displays a negative effect on TOC removal; consequently, B should be minimized to the conductivity threshold required to avoid polarization and excessive ohmic drops, thereby also limiting dissolved-solid loading in the effluent.

The recommended operating window, within the CCD domain and based on the constrained optimum, is as follows: current ≈ 115–135 mA, electrolyte ≈ 1–3 mM, gas-phase O₃ ≈ 7–8 ppm. This region balances kinetic performance and energy cost. For TOC, the constrained optimum occurs at A = +0.252, B = −1.682, C = +0.589 (coded).

Scale-up should preserve key invariants: current density (A·m⁻²), A/V, k_La, and hydraulic/gas residence times comparable to the pilot conditions. Modularization of electrochemical cells and ozone contactors facilitates linear scale-out. Robust anode materials (e.g., BDD) mitigate fouling and sustain oxidative performance. Closed-loop control using DO₃/ORP/UV₂₅₄ and set-points within the recommended window enables adaptation to influent variability.

3. Materials and Methods

3.1. Chemicals

Reference azole fungicides—prochloraz (PCZ), tebuconazole (TBZ), tetraconazole (TCZ), and penconazole (PNZ)—were employed in a mixture (5 ppm for each) as target compounds (≥98% purity; Sigma-Aldrich, Madrid, Spain). Stock solutions were prepared in Type I ultrapure water (Milli-Q) and stored refrigerated; working solutions were freshly diluted on the day of use. The supporting electrolyte (Na₂SO₄ ≥ 99% purity; Panreac Química, Barcelona, Spain) was dissolved in ultrapure water.

3.2. Oxidation Systems

3.2.1. Ozonation (O₃)

Ozonation experiments were carried out in a jacketed glass batch reactor (working volume 0.40 L), thermostated at 20.0 ± 0.5 °C, and operated at a natural pH (6.1). Ozone was generated from oxygen using a Sander GmbH 300.1 (Labor-Ozonisator, Garbsen, Germany) and introduced into the liquid through a porous diffuser to promote homogeneous gas–liquid contact. The inlet gas stream contained a nominal O₃ concentration of approximately 1–10 mg·L⁻¹ (adjustable) at a flow rate of 20 L·h⁻¹.

The off-gas was directed through a 2% KI trap to ensure complete ozone quenching before release, preventing atmospheric emissions. Moreover, based on inlet–outlet concentration measurements, the estimated ozone utilization efficiency was approximately 90–95% under optimized EO–O₃ conditions.

3.2.2. Electro-Oxidation (EO)

The electrochemical system comprised two flat-plate electrodes (5 cm × 8 cm) separated by a 25 mm gap: a boron-doped diamond (BDD) anode (CSEM, Neuchâtel, Switzerland) and a stainless-steel cathode (AISI 304). Smaller electrodes at the same current increase current density and radical flux, which can enhance oxidation kinetics but also promote side reactions (e.g., O₂ evolution, radical recombination). This trade-off is well established in electrochemical advanced oxidation processes and was controlled in our setup by maintaining a constant electrode area of 40 cm². Galvanostatic operation was imposed with a Tektronix PWS2326 (Tektronix, Inc., Beaverton, OR, USA) power supply; reference tests were conducted at I = 0.30 A (V ≈ 24 V). The electrolyte was Na₂SO₄ prepared in Type I water.

3.3. Operating Conditions

All experiments used the same liquid volume (0.40 L) and temperature (20.0 ± 0.5 °C). A reaction time of 10 min was selected to ensure direct comparability among O₃, EO, and EO–O₃ systems under identical energy and hydraulic conditions. Under optimal conditions, practically complete pollutant abatement was achieved (see Supplementary Materials Figures S1 and S2). This duration captures the initial kinetic phase, where pseudo-first-order behavior dominates, and the influence of operational factors is most discernible, while limiting confounding effects associated with ozone depletion and secondary oxidation pathways at longer residence times. However, the TOC removal data were obtained at 60 min, since at 10 min the removal levels were too low to be significant for assessing the influence of the operating variables.

3.4. Analytical Note

Azoles in water were quantified by HPLC-DAD (Agilent 1260 Infinity II, Agilent Technologies, Inc., Santa Clara, CA, USA) using a C18 column (Kinetex, Phenomenex Inc., Torrance, CA, USA, 5 µm, 150 × 4.6 mm). The mobile phase consisted of acetonitrile and an aqueous phase containing phosphoric acid (25 mM) at 1.0 mL·min⁻¹. Detection wavelengths were selected at each compound’s absorption maximum.

4. Conclusions

The combined statistical–mechanistic analysis clarifies the operating levers that matter for treating azole fungicides with the EO–O₃ process. Ozone concentration (C) and galvanostatic current (A) are consistently the most influential factors for pollutant abatement and kinetic acceleration. Their positive effects persist up to curvature-imposed plateaus, beyond which additional input yields diminishing returns. These plateaus reflect either mass-transfer ceilings, radical recombination, or the onset of side pathways—fingerprinted by negative A² terms in the fitted models.

Mineralization (TOC removal) is governed by a different balance: electrolyte (B) exerts a statistically significant and negative main effect, and strong curvature in A and C compresses the feasible window for carbon removal. The deactivation of hydroxyl radicals by sulfate anions occurs through reactions that reduce the availability of •OH for oxidation processes.

From an optimization perspective, the response surfaces recommend increasing ozone and operating at a moderate-to-high current (below curvature-limited optima) while minimizing electrolyte concentration and favoring sulfate over chloride. The recommended operating window, within the CCD domain and based on the constrained optimum, is as follows: current ≈ 115–135 mA, electrolyte ≈ 1–3 mM, gas-phase O₃ ≈ 7–8 ppm. This region balances kinetic performance and energy cost.