Advanced Electro-Oxidation with Boron-Doped Diamond for Acetaminophen Removal from Real Wastewater in a Microfluidic Reactor: Kinetics and Mass-Transfer Studies

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.


TABLE OF CONTENTS
Boron-doped diamond (BDD) anode is implemented in an electrochemical microfluidic cell for electrocatalytic treatment of acetaminophen in synthetic and reclaimed wastewater. An optimal interelectrode distance could be determined at optimal mass transfer and optimal decay rate constant. The low conductivity reclaimed wastewater could be treated with the same range of energy requirement than in synthetic solution. The energy efficiency could be maximized with short interelectrode gap in order to minimize the negative impact of the low conductivity of solutions.

Introduction
Electrochemical processes have been widely studied for the last two decades to answer the challenge of wastewater treatment contaminated by hazardous biorecalcitrant pollutants that are barely removed in conventional wastewater treatment plants (WWTP). [1][2][3][4][5][6] Electrochemical advanced oxidation processes (EAOP) offer the advantages of continuously in situ generated strong oxidizing agents, such as hydroxyl radicals (  OH) (E° = 2.80 V / standard hydrogen electrode (SHE)) responsible for the degradation and mineralization of organic pollution. [7,8] Boron-doped diamond (BDD) anodes have been developed to promote oxidation by heterogeneous catalysis. [9][10][11][12] The high oxygen (O2) evolution overvoltage (approximately 2.3 V / SHE) allows for the generation of physisorbed  OH at the BDD surface from water oxidation (Eq. 1): [9] BDD + H2O  BDD(  OH) + H + + e - (1) This so-called anodic oxidation or advanced electro-oxidation process has been widely developed since the early 2000s as a promising EAOP. [13,14] No pH adjustment is required, and no chemicals are needed except for the supporting electrolyte for the low-conductivity solution.
Effluents at the outlet of municipal WWTP exhibit conductivity of approximately 1 mS cm -1 , which is too low to efficiently drive the electric current in conventional electrolytic systems. The addition of electrolyte would constitute an external contamination, since the salts -usually sulfated or chlorinated compounds -need to be removed after the electrolytic treatment.
Electrochemical microfluidic reactors have therefore been proposed to address this issue by implementing a very short interelectrode distance (50-500 µm). [15][16][17] This process allows for minimizing the ohmic drop and then the cell resistance, i.e., the conductivity requirement. In the meantime, the cell potential is reduced, which permits a decrease in the energy consumption of the cell, [18] another typical drawback in conventional EAOP technologies. A further advantage is the transfer intensification, which increases the oxygen redox cycle and iron redox cycle in undivided microcells. The process then accentuates the rate of H2O2 production and Fe 2+ regeneration in the electro-Fenton system. [17] especially when heterogeneous catalysis is involved with the use of a BDD anode. [19] This study investigates their importance in a BDD-microfluidic reactor for the degradation of acetaminophen as a representative of pharmaceuticals in wastewater (Fig. 1). Also known as paracetamol, it is a widely prescribed analgesic and anti-pyretic drug that has been detected in groundwater [20] and surface water [21,22] in concentrations up to 65 µg L -1 [23] . Even at low concentrations, acetaminophen can cause damage to aquatic life and human health. [24] The influence of main parameters, such as current density, interelectrode gap and solution conductivity, is studied, and the factors are optimized by considering decay rates, mass transfer and specific energy requirements.

Results and discussion
2.1.

Influence of current density on degradation efficiency
The current density is an important parameter to optimize electrochemical processes since it controls the kinetics of electrochemical reactions.
The applied current density (Jappl) was varied from 2 to 12 mA cm -2 , and the impact on the kinetics of acetaminophen degradation is plotted in Fig. 2a. A pseudo-first-order kinetic model was considered as a first approach for apparent rate constant determination as usually proposed in the literature. [25] The correlation coefficients (R 2 ) values varied from 0.986 to 0.991, which is close to 1 and validates the model assumption.
An 8-fold increase in the degradation rate was observed from 2 to 4 mA cm -2 , while a plateau from 4 to 9 mA cm -2 was observed (kapp = 0.054 ± 0.002 min -1 ) before a decrease in the degradation rate constant at 12 mA cm -2 (kapp = 0.018 min -1 ) (Fig. 2a). At higher current density values, the secondary reaction, such as the O2 evolution reaction at the anode, becomes increasingly important according to Eq. 2, being in competition with the production of  OH (Eq. 1), while the hydrogen evolution reaction occurs at cathode surface (Eq. 3). [9,26,27] 2H2O  O2(g) + 4H + + 4e - To better assess the oxidation performance in BDD cell system, the electroactivity of acetaminophen has been previously tested in a platinum (Pt) anode / graphite cathode system immersed in 157 mg L -1 of acetaminophen synthetic solution. [28] Pt is an active anode having lower O2 evolution overvoltage (1.7-1.9 V / SHE) than BDD. [4]  OH are therefore chemisorbed at Pt surface indicating that  OH are barely available for oxidation. Thus, direct electrooxidation mainly occurs in presence of Pt anode. At a current density of 100 mA cm -2 , a TOC decrease of 20% was noticed after applying a specific charge of 18 Ah L -1 against 99% of TOC removal with BDD anode. [28] This previous work highlight the ability of acetaminophen to be oxidized by direct electron transfer though the use of BDD clearly depicted its superiority by involvement of advanced oxidation. Acetaminophen's molecular structure presents unsaturated bonds, especially in its aromatic ring.  OH reacts particularly very quickly with double C=C bonds, which makes advanced oxidation mechanism more selective and efficient as compared to direct oxidation mechanism.
Another recent study confirms the electroactivity of acetaminophen using graphite anode that is an active anode such as Pt. [23] Upon these results, it is important to emphasize that the degradation and mineralization efficiencies presented in this paper are also due to direct electro-oxidation, in a lesser extent than advanced oxidation.
The specific energy consumption (Especific), expressed in kWh per gram of pollutant removed (kWh g -1 ) and calculated according to Eq. 4, [1] is a decisive aspect to take into account because it represents an important part of the operating costs. [29] = ( ) where Ecell is the average cell voltage (V), I is the applied current intensity (A), t is the electrolysis time (h), Vs is the solution volume (L), and C0 and C are the initial concentration and the concentration at time t of acetaminophen (mg L -1 ), respectively.
The Especific values are represented as a function of the acetaminophen degradation yield in Fig. 1b for different Jappl values. Jappl of 9 and 12 mA cm -2 clearly appeared as high energydemanding conditions, regardless of the disappearance percentage. 4 mA cm -2 required the lowest specific energy (0.11 -0.50 kWh g -1 ) for up to 83% of decay, while 3 mA cm -2 offered better energy efficiency (0.19 -0.31 kWh g -1 ) for 83% to 87% of acetaminophen degradation.
Knowing that the kinetics rate of degradation was 1.6-fold lower with 3 mA cm -2 conditions compared with 4 mA cm -2 , the latter current density seems optimal and has been selected for further optimization.

Mass transfer determination
The acetaminophen concentration decay curve is represented in Fig. 3 for a 250 µm interelectrode distance. It appears that there are two trends during electrolysis, i.e., a linear trend at the beginning of the treatment followed by an exponential trend until the end of the electrolysis, as described previously by Comninellis' team. [13,30] The first trend can be assimilated to a zero-order kinetic model until Jappl reaches the limiting current density (Jlim) value. Then, the diffusion becomes the rate-limiting step, and Jappl is higher than Jlim.
The electrolytic systems are mainly governed by two transfers: (i) the charge transfer at the electrode surface during the electrochemical reaction and (ii) the mass transfer of the target pollutant from the bulk solution to the electrode surface and from the electrode to the bulk. [19,31] Thus, at Jappl < Jlim (zero-order model), the process is under charge transfer control, while at Jappl > Jlim (pseudo-first-order model), it is under mass transfer control. The time at which the kinetic order switches is named the critical time (tcr). [13,30] When t = tcr, Jappl = Jlim, and Jlim (A m -2 ) can be defined as follows (Eq. 5): [13,30,32] = where n is the number of electrons exchanged, F is the Faraday constant (96485 C mol -1 ), km is the mass transfer coefficient (m s -1 ) and Clim is the concentration (mol m -3 ) of the target compound (acetaminophen in this study) at t = tcr.
The number of electrons exchanged during the partial electrochemical combustion of acetaminophen (C8H9O2N) at the BDD anode can be estimated by considering benzoquinone as the first main by-products, [33,34] according to the following reaction (Eq. 6): It can be concluded that n = 2 for acetaminophen degradation. From Fig. 3, the Clim can be deduced from (C/C0)lim (C0 is the initial concentration of acetaminophen) at t = tcr (Table 1), and km can then be calculated for each interelectrode gap experiment using Eq. 5 for Jappl = Jlim = 40 A m -2 . Table 1 gives tcr and (C/C0)lim values for each interelectrode distance. It was shown that (C/C0)lim increases from 500 µm to 1000 µm and to 50 µm because the mass transfer become the limiting factor very quickly as compared to charge transfer in these conditions.   [18,35] The existence of an optimum in this work provides new insight, as it could explain the existence of an optimal degradation efficiency value obtained when varying the interelectrode distance in previous studies dealing with electrochemical microreactors for wastewater treatment. [15,36] This feature could be attributed to the transfer limitation by gas bubbles [37] through the formation of O2(g) at the anode (Eq. 2) and H2(g) at the cathode (Eq. 3) at a low interelectrode gap (50 and 250 µm).
It has been previously emphasized that the current density increases at the edge of a parallelepiped flow-by cell with the contraction of interelectrode distance. [37] As the current density increases the bubble size, [37] it has an impact on the mass transfer at small gaps.
Nevertheless, the maximal km value was higher than in a flow-by reactor with gaps of 1000 µm and 5000 µm, giving km values of 1.5 × 10 -5 m s -1 and 7 × 10 -6 m s -1 respectively, at 5 L min -1 , [35] 830 to 1785 times lower than in our study. The difference is even higher compared with stirred tank reactor configurations (km = 5.9 × 10 -6 m s -1 ). [38][39][40] It remains important to highlight that our km calculation consider the targeted pollutant transfer towards anode for its subsequent degradation and does not take into account the ferrocyanide representative compound usually employed for mass transfer characterization [32] . The difference of method could partly explain the difference of km range.
In parallel to the hydrodynamic aspects, the kinetic rate constants ( value was noticed in this condition, which is also noted in the literature. [15] However, the kinetics rate decreased towards the end of electrolysis, especially at 50 µm and 250 µm interelectrode distances, when looking at acetaminophen concentration decay in Fig. 4c. This phenomenon could be ascribed to the increase of bubble formation with electrolysis time. Indeed, current efficiency tends to decrease after a certain electrolysis time, [19,41] along with the increased occurrence of secondary reactions such as in Eqs. 2 and 3. This feature is also highlighted in energy considerations, as depicted in Fig. 4d. At 50% to 80% of decay, the energy requirements varied from 0.11 to 0.17 kWh g -1 , while they increased dramatically at higher acetaminophen degradation percentages at a 50 µm interelectrode distance. In the range of 50-80% disappearance, the energy consumption equaled 0.17 kWh g -1 at 500 µm of the interelectrode gap and increased from 0.18 to 0.36 kWh g -1 for degradation percentages of 84% to 94%. At the highest oxidative degradation yields, 500 µm became the optimal interelectrode distance. A distance of 1000 µm -corresponding to a more conventional gapclearly showed higher specific energy consumption, varying from 0.18 to 0.56 kWh g -1 with 30% to 83% decay yield, respectively, compared to the shorter intervals. It emphasizes the benefit of decreasing the distance on the energy efficiency due to the decrease of cell potential.
A cell voltage of 4.1 V was measured at 1000 µm against 2.9 V at 500 µm, corresponding to a 30% decrease in cell voltage. This confirmed the trend obtained in a previous work, with a decrease of cell potential of approximately 1 V when decreasing the interelectrode distance from 4000 µm to 120 µm. [15]

Influence of solution conductivity
The conductivity is another valuable parameter to investigate when the addition of supporting electrolyte needs to be avoided, especially in the context of wastewater treatment. [42]  at the BDD anode surface and persulfate (S2O8 2-) in bulk solution. [43,44] These species are strong oxidizing agents that take part in the whole oxidation process. [2] The influence of solution conductivity on the specific energy is depicted in Fig. 5b. An energy increase from 0.31 to 0.88 kWh g -1 was required for an acetaminophen decay yield varying from 26% to 85% at 0.23 mS cm -1 . In contrast, the energy requirements at 0.86 mS cm -1 and 2.0 mS cm -1 were lower. For instance, from 3.8 times (0.86 mS cm -1 ) to 4.9 times (2.0 mS cm -1 ), lower energy was needed to reach 85% of pollutant degradation. The increase in conductivity decreases the cell resistance and therefore the cell potential. Thus, an average cell potential difference of 3.5 V was noticed between the experiments at 0.23 mS cm -1 and 2.0 mS cm -1 , which mainly explained the contrast of energy demand.
A solution conductivity is still required to drive the electric current, but it can be minimized in microfluidic cells compared to the range of electrolyte concentrations (50-100 mM) used in conventional reactor design. Sulfate-or chlorine-based electrolytes are widely employed in the literature. [45][46][47][48][49] Consequently, high concentrations of sulfate and chlorine ions are released after the electrolytic treatment and are not removed during electrolysis. Hazardous chlorinated transformation products can also be formed during electrolysis, such as organochlorinated compounds (e.g., trihalomethanes (THMs)) and inorganic species that tend to accumulate (chlorate (ClO3 -) and perchlorate (ClO4 -)). [42,50,51] Since a conductivity of 1 mS cm -1 corresponds to the average conductivity of the municipal WWTP outflow, the decrease of degradation efficiency compared to the addition of supporting electrolytes, such as Na2SO4 at 10 mM (2.0 mS cm -1 ) or even higher concentrations, would be negligible considering the nonlinear trend observed in Fig. 5a. Thus, the gain that can be achieved by avoiding the addition of chemicals must be considered. [acetaminophen]0 = 0.1 mM, cathode = carbon felt, anode = BDD, interelectrode distance = 500 µm, Jappl = 4 mA cm -2 , no pH adjustment.

Influence of reclaimed municipal wastewater on degradation and mineralization of acetaminophen
The influence of reclaimed municipal wastewater on the process efficiency has been assessed.
The physico-chemical properties of the real solution are presented in Table 2. The pH was circum-neutral (7.7) and the conductivity equal 0.86 mS cm -1 . The total organic carbon (TOC) was around 4.2 mg L -1 , which represented 7% of the total carbon present in wastewater.
The specific energy remains similar to synthetic solution until 60% of degradation with a quasiconstant value of 0.15 kWh g -1 (Fig. 6b). The energy requirement starts increasing at higher degradation yield with an energy of 0.22 kWh g -1 for 70% of decay. The total organic carbon (TOC) has been monitored in order to take into account the global organic pollution, including the potential toxic organic by-products that could be produced. [25] The TOC removal of reclaimed wastewater spiked with acetaminophen (0.1 mM) is depicted in Fig. 7a. The TOC decreased from 13.7 mg-C L -1 to 0.37 mg-C L -1 in 4 h of treatment, leading to 87% of mineralization. Comparatively, an acetaminophen mineralization of 41% was noticed after 4 h of electrolysis of synthetic solution at 25°C with BDD cell implementing 30,000 µm of interelectrode gap with a treated volume of 100 mL. [34] Therefore, the mineralization yield was two times higher in our work even though our treated volume was 500 mL instead of 100 mL.
Moreover, the applied conditions in the previous study led to an applied specific charge of 12 Ah L -1 against 1.6 Ah L -1 , 7.5 times higher than in the present work. The difference of efficiency could be attributed to the 60-fold shorter electrode distance involved that greatly enhance the mass transfer as above-mentioned.
The mineralization current efficiency (MCE) has been calculated according to Eq. 7: where TOC0 and TOC are the initial TOC concentration and the TOC concentration at time t The MCE values are displayed in Fig. 7b. An increase of MCE until 15.3% is noticed during the first hour of treatment while it decreases until 7.4% after 4 h of electrolysis. This trend is probably due to the more easily oxidizable aromatic intermediates as compared to the carboxylic acids such as oxalic and oxamic acids that are the end-products before CO2 conversion. [34] A MCE of 9.7% was obtained by Brillas et al. [34] after 4 h of treatment. The slightly lower values of MCE in the present study could be explained by the presence of organic matter (OM) that represented around 30% of the TOC. This OM is in competition with the targeted pollutant, since it can also react with  OH. [53] The specific energy towards TOC removal has been determined in the same manner than with acetaminophen degradation (Eq. 4), by considering C0 and C as the initial TOC concentration and the TOC concentration at time t (mg-C L -1 ), respectively. The evolution of Especific is shown in Fig. 7c. From 45% to 87% of TOC removal, the energy increased from 0.23 to 0.49 kWh g-TOC -1 . At similar TOC removal yield (45%), the specific energy in microfluidic reactors was 3.7 times lower than Brillas et al. [34] study. This can be ascribed to the drop of cell potential at very short interelectrode distance, while the transfers are maximized. This study has shown that a current density of 4 mA cm -2 was optimal in BDD-microfluidic electrochemical reactors to degrade acetaminophen in synthetic solution with an optimal interelectrode distance of 500 µm. Since BDD involve heterogeneous oxidation, the reactions should be maximized when the contact between the solution and BDD surface is maximized.
This should be the case when the ratio "BDD surface/cell volume" is maximized, i.e. when the interelectrode gap is shorten, until an optimal gap as highlighted in this study. This is explained by the existence of an optimal mass transfer (12.6 × 10 -3 m s -1 ), being reduced at low electrode intervals (from 50 to 250 µm) by the gas bubble generation at the cathode and anode.
Nevertheless, the mass transfer was 2 to 10 times higher than in conventional EAOP reactor design. In this condition, the specific energy could be reduced to 0.

Experimental section Chemicals
Sulfuric acid (H2SO4) was provided by Thermo Fisher Scientific (Karlsruhe, Germany).
Acetaminophen was obtained from Sigma-Aldrich (Saint-Quentin-Fallavier, France). Sodium sulfate (Na2SO4) was supplied by VWR International (Fontenay-sous-Bois, France). All the chemicals were of analytical grade. Ultrapure water from a Purelab ® water purification system (Veolia Water, Antony, France) (resistivity > 18 MΩ cm at room temperature) was used to prepare the solutions.

Synthetic solutions and reclaimed wastewater
In synthetic solutions, H2SO4 was supplied to adjust the pH to 3, except in experiments on the influence of electrolyte concentration to avoid the increase of salinity. Although the acidic pH should have no effect on the anodic oxidation experiments, [28] the acidification of solution was performed to compare with future electro-Fenton experiments in which a pH of 3 is necessary to implement Fenton reactions. [1] Na2SO4 was added as the supporting electrolyte by varying This also ensured that no contact between the cathode and anode occurred at very short interelectrode distance. Both the cathode and anode had the same geometric surface area of 50 cm 2 in contact with the solution. The carbon felt was conditioned before use by immersing the material in an ultrasonic bath with an acetone/ultrapure water mixture for 1 h. [54] The material was then thoroughly rinsed with ultrapure water and dried in an oven at 60 °C for 24 h. Each carbon cathode was employed only for one set of experiments to ensure better repeatability of the results. This is to avoid any change of cathode properties after several use (change of porosity, adsorption of organics) and could affect the process efficiency. [42] The interelectrode gap was varied using polytetrafluoroethylene (PTFE) spacers of different thicknesses delivered by Bohlender (Grünsfeld, Germany). The current intensity was applied using a HAMEG 7042-5 (Beauvais, France) power supply, and the current density was calculated as a function of the geometric surface of the BDD anode.

Analytical procedures
Acetaminophen was analyzed with a UV-Vis spectrophotometer (Anthelie Light, Secomam-Aqualabo, Champigny-sur-Marne, France) at a wavelength of 243 nm, giving the maximal absorbance for optimal sensitivity. The amount of acetaminophen could be determined from the calibration curve that gave a molar extinction coefficient of 9767 L mol -1 cm -1 with an R 2 of 0.9998. Thus, it validated the Beer-Lambert law in the considered range of concentration (0.001-0.1 mM).
A VCSH TC/TN analyzer (Shimadzu; Marne-La-Vallée, France) was employed to quantify the TOC, TIC and TN in reclaimed wastewater.
A Hach colorimetric method using salicylate procedure and a DR-2400 spectrophotometer (Hach; Lognes, France) were employed to quantify ammonium ions.