For the exhaustive electrolysis, a single compartment in galvanostatic regime and ideally stirred batch reactor was used. The solution flow rate was 2.08 mL∙s⁻¹. The other chemicals used in this work are composed of KClO₄ (Fluka), HClO₄ (Fluka) and NaCl (Fluka). All the chemicals have been used as received without any further treatment for the experiment. All the experiences were carried out under magnetic agitation at ambient temperature of 25˚C. The mass transfer coefficient determined using the potassium ferri/ferrocyanide redox couple (Fluka) was 2.36 × 10⁻⁷ m∙s⁻¹. The anode was Ta₂O₅/Pt-RuO₂-IrO₂ while Zirconium (Zr) plates were used as cathode. During each electrolytic run, samples of 2 mL were taken from the reactor at defined time intervals and analysed in terms of COD and absorbance.

A schematic of experimental mounting is shown in Figure 2.

Chemical oxygen demand (COD) of the samples was determined using the pre-dosed COD tubes of the HACH product. 2 mL of samples are taken and introduced into a HACH tube and heated in a HACH brand digester at 150˚C for 120 minutes. After cooling, the COD is read directly by means of the spectrophotometer DR/6000 of the HACH product with a wavelength of 620 nm.

For the measurement of the absorbance, 2 mL of the sample taken at predefined intervals of time are introduced into a tube. The wavelength of the sample is also determined using the HACH DR 6000 spectrophotometer.

Current density is a very important parameter in electrooxidation [29]. The effect of current density was investigated during the degradation of an amoxicillin solution (1 g/L) prepared in 0.1M potassium perchlorate (0.1 M KClO₄). The current densities used in this work vary from 20 to 100 mA/cm². To follow the degradation of amoxicillin during its electrolysis, the chemical oxygen demand (COD) was measured. In this work, COD was normalized according to Equation (1).

COD * = COD 0 / COD t (1)

where

COD* is the normalized value of the COD.

COD₀ is the chemical oxygen demand at the initial time and COD_t, the oxygen demand at any time t.

Figure 3 shows the effect of the current density on the evolution of COD* at 25˚C as a function of time. COD* decreases rapidly as the current density increases. Thus, the COD abatement rate (insert Figure 2) obtained after 10 h of electrolysis is 0.83%; 10.71% and 40.91% at 20, 50 and 100 mA/cm² respectively. Such an observation has also been made in previous work on BDD used as anode for the degradation of amoxicillin (1 g/L) [30]. However, the COD abatement rates obtained on the PRI are very low compared to those obtained on BDD. This could be explained by the fact that on the BDD, the hydroxyl radicals formed by the decomposition of water are free and therefore very reactive. While

on the PRI, a DSA type electrode, these hydroxyl radicals are chemisorbed and do not contribute effectively to the degradation of the organic compound. Indeed, the initial step in the electrooxidation of amoxicillin on the PRI electrode would be the decomposition or discharge of water leading to the formation of adsorbed hydroxyl radicals (Equation (2)). These hydroxyl radicals have a strong interaction with the surface of the PRI electrode. The next step is the formation of a higher oxide (Equation (3)). This higher oxide is responsible for the degradation of the organic compound (Equation (4)) but that leads to partial and selective oxidation of the organic. However, the oxidation reaction competes with the oxygen evolution reaction (OER) (Equation (5)).

M + H 2 O → M ( • OH ) + H + + e − (2)

M ( • OH ) → MO + H + + e − (3)

MO + R → M + RO (4)

MO → M + 1 2 O 2 (5)

M is the anode material and R is the organic compound.

The initial current density has been calculated using Equation (6).

I l i m = 4 F k d A ⋅ COD (6)

where F is the Faraday constant (F = 96 487 C∙mol⁻¹) C, k_d is the mass transfer coefficient (m∙s⁻¹), A is the electrode surface area in contact with the solution (m²), COD is the chemical oxygen demand (mol∙m⁻³).

According to literature [31] [32], two main regimes can be reached depending on the applied current compared to the initial limiting current.

If I < I_lim, a galvanostatic process governs the overall oxidation process i.e. limited by charge transfer reaction. Under these conditions, all the applied current is used for the oxidation of the organic compound.

If I > I_lim, the oxidation process is mass transfer controlled i.e. limited by diffusion. Under these conditions, part of the applied current, which is equal to the limiting current, is used for the electrooxidation of the organic compound and the rest for the side reaction i.e. oxygen evolution.

The limiting current density calculated in this work is 7.1 mA∙cm⁻². This value of the limiting current density is lower than the applied current density. So the oxidation process of amoxicillin is limited by diffusion. In such a regime, other reactions, such as oxygen evolution reactions, may compete with the oxidation reaction of amoxicillin.

During the electrolysis of 1 g/L of the amoxicillin solution at 20 mA/cm² and 100 mA/cm², the absorbance of the samples withdrawn at different times from the synthetic wastewater tank has been registered. The obtained results are shown in Figure 4. For 20 mA/cm² (Figure 4(a)), the absorption spectra increase with the electrolysis time. The peak recorded at 350 nm was plotted against the electrolysis time (insert of Figure 4(a)). In Figure 4(a), it is observed that the absorbance increases during all the electrolysis duration. This could be due to the production of reaction intermediates that continues to be produced during electrolysis. Indeed, there is a conversion of amoxicillin into

organic compounds that absorb at the same wavelength as the amoxicillin. The degradation of amoxicillin on the PRI electrode at 20 mA/cm² leads to continuous formation of one or more intermediates that absorbed at the same wavelength. The absorbance of the samples withdrawn from the synthetic wastewater tank was also recorded at 100 mA/cm². The obtained result is shown in Figure 4(b). In the insert of Figure 4(b) the peak of the absorbance of the amoxicillin solution was plotted as a function of the electrolysis time for a wavelength of 350 nm. In this figure, the absorbance increases until it reaches a maximum value after 5 hours. After 5 h, there is a decrease of the absorbance up to 10 hours of electrolysis. This shows that the electrolysis of amoxicillin leads to the formation of some reaction intermediates whose concentration increases until 5 h and then decreases. It appears that the degradation of the intermediates produced under this current density occurs during the electrolysis.

These investigations show that the current density plays an important role in the process of the amoxicillin degradation on the PRI electrode. For 20 mA/cm², the electrolysis of amoxicillin leads to the formation of intermediates with an increase in their concentration during electrolysis. This result shows that the oxidation of amoxicillin at 20 mA/cm² favors conversion of the amoxicillin to some reaction intermediates. But under 100 mA/cm², the concentration of the intermediates increases until reaching a maximum value at 5 h (100 mA/cm²) before decreasing in the rest of the electrolysis time.

From these results, it could be noted that increasing the applied current density causes an increase of the adsorbed oxidative species such as hydroxyl radicals at the surface of the anode electrode that contributes to the enhancement of the organic compound degradation [33] [34].

The effect of the supporting electrolyte was investigated on the degradation of 1 g/L of amoxicillin at 20 mA/cm² on the PRI electrode used as the anode. The supporting electrolytes used are 0.1 M KClO₄ and 0.1 M HClO₄ at 20 mA/cm². The obtained results are shown in Figure 5.

In this figure, the COD * decreases very slowly in HClO₄ whereas it is almost constant (does not change) in KClO₄. After 10 hours of electrolysis, the obtained COD reduction rates are 0.83% for KClO₄ and 8.15% for HClO₄.

The degradation kinetic constant was determined during the degradation of amoxicillin in 0.1 M KClO₄ and 0.1 M HClO₄ (insert of Figure 5). In both the solutions, the obtained kinetic constants are 0.0007 h⁻¹ and 0.0089 h⁻¹ in potassium chlorate and perchloric acid, respectively. From these results, it appears that the kinetic of the reaction is pseudo-first order and the degradation seems to occur more easily in HClO₄ than in KClO₄. This finding could probably be due to the participation of protons (H⁺) in the degradation process of amoxicillin. Indeed, it has been shown that in acidic solution, DSA type electrode surface is more active and the oxygen evolution reaction (OER) rate diminishes in favor of amoxicillin degradation [33] [35]. Moreover, it’s also reported in some works

that the oxygen evolution reaction (OER) is kinetically favored in a less acidic medium (neutral or an alkaline medium) [35]. We can conclude that the supporting electrolyte medium influence, significantly the degradation of amoxicillin on the PRI electrode.

In order to be closer to real wastewater which pH should be around neutral to be rejected in the environment, we pursue the work in KClO₄.

The degradation of a solution of 1 g/L of amoxicillin in 0.1 M KClO₄ was carried out in the absence and in the presence of 400 mM NaCl at 20 mA/cm² on the PRI electrode. The obtained results are shown in Figure 6. In the presence of NaCl, the degradation rate of amoxicillin is faster than that obtained in the absence of NaCl. After 10 hours of electrolysis, the reduction rates of the chemical oxygen demand is 0.83% and 73.8% respectively in the absence and in the presence of NaCl. The oxidation of amoxicillin is rapid in the presence of chloride ions. This observation is in agreement with previous reports on the degradation of organic pollutants using DSA anodes in the presence of chloride ions [36].

From these findings, it can be indicated that the degradation of amoxicillin can undergo direct oxidation on the surface of the electrode and/or through oxidative species such as chlorine. Similar results were obtained on the RuO₂ electrode [19]. In fact, in this experimental condition, specific adsorption Cl⁻ occurs on the DSA (PRI) surface. Then the adsorbed species react leading to Cl₂ according to Equation (7). Furthermore, Cl₂ passes into solution or is evolved. Moreover, in solution, Cl₂ can be present in various forms depending on the pH of the medium. There can be Cl₂ (pH < 3), HClO (3 < pH < 8) and ClO⁻ (pH > 8) according to Equations (7) to (9):

2Cl − → Cl 2 + 2e − (7)

Cl 2 + H 2 O → HClO + Cl − + H + (8)

HClO ↔ ClO − + H + (9)

During the electrolysis of amoxicillin in the presence of 400 mM NaCl, the pH of the solution was followed (Figure 7). The obtained result showed that Cl₂ is mainly in the form of HClO form predominated during the first 5 hours of the electrolysis. After 5 h of electrolysis, the predominated form is ClO⁻.

That means that at the first stage of the degradation of Amoxicillin in that medium, indirect reaction involving HClO happens. After 5 h of electrolysis, ClO⁻ is responsible for degradation of Amoxicillin or its intermediates.

The color of the amoxicillin solution was followed during the electrolysis in the absence (Figure 8(a)) and in the presence (Figure 8(b)) of NaCl. In the absence of NaCl, the solution goes from colorless to dark yellow. In the presence of NaCl, the solution goes from colorless to a dark yellow for 2 h of electrolysis. After 2 h, it becomes colorless again. These results show that a conversion of the initial amoxicillin is favored on the PRI in the absence of Cl⁻. In the presence of Cl⁻, mineralization of amoxicillin on the PRI occurs.